The past few weeks, I've had my home lab running near 100% utilization running https://github.com/ourofoundation/ggen, searching for an ideal rare-earth-free permanent magnet. Many decent candidates have emerged, but I'm going to keep the search going. Ideal has not been found. What I'm focused on now is finding candidates as low energy as possible and watching out for "dangerous" competing phases. We want to ensure the time in lab is spent well. I also want to already be thinking about easy of synthesis. Some exotic material that's hard to produce just really isn't that interesting when the goal is to replace an incumbent materials that has a demand of thousands of tons every year.
While I'm generally running ggen unconstrained (except for choosing chemical system), in my own hand picking I'm generally looking for P4/mmm with a high fraction of Fe/Co, and a low fraction of the heavy/large SOC contributor, like Bi or W.
So far, my local ggen database has produced:
173,842 structures
10,691 unique formulas
101 unique chemical systems
Let me share a bit what I've been learning about some of the systems at a high level and see if any new guiding streams emerge.
I initially had a lot of hope for this kind of system because I almost always found the Bi brought really strong SOC and high MAE. MAE is notoriously one of the hardest parts in a good rare-earth-free permanent magnet, so I I could rely on a heavy element I could be a bit more lenient on needing some fancy structure, metastable phase, or hard-to-model doping.
What I learned is that there is nothing stable near the right composition. I say nothing but of course I want to be careful since you never really know. But I've searched plenty and it looks like the energy landscape is very high all over that part of the map.
On this page
I've mostly focused on ternaries here, and have searched through a handful of hand picked candidates like Ga, Nb, S, Al, Sn, Se, and so far nothing has shown much promise in stabilizing Fe-Bi.
The Bi-Fe-Ga Phase Diagram provides an illustration of the relationships between the phases of bismuth (Bi), iron (Fe), and gallium (Ga) at varying compositions. It is crucial for materials scientists and researchers studying ternary alloy systems, as it outlines the stability ranges and phase transformations that occur in this specific alloy.
The story is very clear from a fully explored system. The hole of materials in the middle of the PD, where Fe and Bi come close to equal fractions, wasn't skipped in the search - the materials there are so high energy they don't make the 150 meV/atom cutoff.
Then we look closer at the high Fe content materials and we see a lot of red. Everything there is high energy, generally >100 meV above the hull.
I was talking with and about this issue and the conversation went in the direction of a more thorough search across elements. Right now I'm hand picking the elements to test, and testing them thoroughly. Perhaps we can be smarter about this. Search across Fe-Bi-X and change X more quickly when we find out it may not be a good candidate. One we have the 3 corners of the PD (which we can get from Materials Project easily), just testing a few in our desired composition area should at least give us an idea. Since that area of high Fe, low Bi, low X (trying to roughly mirror Nd2Fe14B) has always shown high energy in the systems we've searched before, even just finding on good candidate is a good sign.
It's going to require ggen updates but I'll start working in this direction and see what we find.
The other issue was that the lowest compounds I did find were likely poor in MAE. Good MAE comes when the magnetic contributor is close by, almost propped up by the SOC contributor. And you want as much of this as possible. Of course structure matters too, but that's the most basic requirement. Most of these low energy compounds of Fe-Bi-X found the Fe as far from Bi as they could be, generally separated by X.
28 meV above the hull
See where Ga is in this compound is ideally where we want Bi to be. Instead, Bi is pushed out and then is surrounded by fewer Fe. The example we have below with tungsten is the better outcome.
This has been my latest interest. Because of all the trouble Bi has been causing, I figured it was time to give it a bit of a break and try another approach.
I haven't really seen much interest or prior work looking into tungsten as an element in rare-earth-free permanent magnets, and I'm not exactly sure why. If someone knows and can save me the time searching down a path I shouldn't be, let me know.
What I've noticed so far structurally/thermodynamically is that it may have an issue with dynamical stability. I haven't been running the full check within ggen just so that I can explore more, but when I bring things to Ouro to check a good candidate, they've always been unstable. We'll see if that's a real issue or just unlucky.
The other issue has generally been lower Ms too, but with the right composition I feel like we should always be able to do something about that.
Otherwise, I've been finding decent promise.
73 meV above the hull
Now comparing to the bismuth example we had before, the tungsten is getting cozy with as much iron as it can. This is just a vibe I've picked up from NdFeB, but it seems that's the role Nd played so we're looking to see if tungsten can essentially play the same role.
Still more work to be done. The candidate I'm showing above isn't the lowest energy, neither does it have good Ms or dynamical stability. With the right third element, I think we may find something good here.
So the idea is again the same.
Add ggen functionality to efficiently explore across different chemical systems, looking for low energy domains.
I imagine there are some other use cases where this kind of functionality might be helpful.
Back to building!