The figure shows a cubic double perovskite in rocksalt structure with two magnetic ions occupying the B/B'-site alternatively along [111] direction (A2BB'O6). Assuming that this is a ideal structure without any distortions, is it possible that the Heisenberg exchange parameter Jin \neq Jout ? What may give rise to such a phenomenon? Strong electronic correlations or what else?
Any comments and discussions are appreciated here! In particular, welcome Prof. Tapan Chatterji to give insightful views and critical comments.
Thank you very much!
Caption for the figure: Schematic view of A2BB'O6 structure. B and B' are magnetic ions. A is a non-magnetic ion. Not that all of O atoms and most of A atoms are removed for clarity.
Sorry, somehow I missed your question. The double perovskites are quite interesting and have many interesting properties. The question you have asked is quite pertinent. Blow the magnetic phase transition even apparently cubic substance can show magnetic anisotropy in a singe domain sample depending its magnetic structure. At the ordering temperature considerable magnetostriction may modify the crystalline symmetry (See our recent publication on MnS2) and the bond distances may vary in the plane and out-of-plane directions. So cubic substances below its magnetic transition may show even quasi-two-dimensional behavor. The famous example is LaMnO3 below its Neel temperature. The spin wave dispersion shows that in plane and out of the plane exchange interactions are different, in plane being much stronger than the out-of-plane exchange interaction. So this is even true for the single perovskite LaMnO3. I shall not be surprised if you find that in your double perovskite system. In fact it s quite likely. Unfortunately the double perovskites are often disordered or only partially ordered. That makes the thing more complicated. Also it is difficult to grow large single crystal. So we can study spin waves and dertermine the exchange interactions and show that the in-plane and out-of-plane nteractions are differnt. If you can grow a large single crystal of a ordered double perovskite then we can settle this by doing inelastic neutron investigations.
You can of course get a very prounced two-dimensional behavior if can make the related Ruddelsden-Popper phase.
If there is no distortion then you will not see the anisotropy in exchane interaction but a very small distertion (which you cannot measure often) can lead to anisotropy in exchange interaction. In fact magnetostriction is always there; you cannot measure it always due to the limited resolution of your instrument. Read our recent paper on MnS2 then you will appreciate this fact.
Dear Prof. Tapan Chatterji,
Thanks very much for your fruitful information!
1. Actually we have grown some films (50~250 nm) with structural order. How large is it needed to perform inelastic neutron investigations? For the details, we can discuss in the email.
2. At present, I have conducted many first-principles calculations in order to find out the possible magnetic structures. It's strange that in the cubic models J_in \neq J_out without distortion. Do you think the orbital order can bring the anisotropy? I found one of the B-site atoms shows orbital order in some certain magnetic structures.
Besides, thanks for introducing your recent work. I'll read your paper on MnS2 to get more.
I guess one has to take Khomski-Kugel charge-orbital model. The charge-orbital interaction is also responsible for the exchange anisotropy in LaMnO3. For double perovskite things are more complex. First of all you have to know the ordering of B site ions. It is really difficult to get perfect order. In our polycrystalline materials we achieved only parial order. I know nothing about the degree of order in your thin films. It can however estimated by looking at the superlattice reflection like 011 (if I remember correctly). You can do X-ray diffraction from your thin films to check the degree of order.
To do inelastic neutron scattering from thin films is very difficult due to the weak signals involved.
Hi, Prof. Tapan Chatterji,
Thanks for your prompt answer!
It seems that it would be more intricate if Kugel-Khomski model is involved in. I need to further inspect it's validity for our case. As for our films, there is B-B' ordering along [111] , which extends into over 50 nm and has been confirmed by our XRD and TEM data. At the surface and the interface contacting the substrate, they are not highly ordered. After proposing a reliable magnetic structures (I'm still working on this) which can explain the colleagues‘ observations, we'll upload it to arXiv first. The manuscript is almost ready. I'll keep you posted.
Besides, I also read your latest study on MnS2 in JPCM. It has a tiny structure transition below 48 K, have you got out-of-plane filed-cooled and zero-filed-cooled temperature dependent magnetization (M-T) curves for MnS2? I'm curious about how the structure transition affects the behaviors of these curves.
Can you please tell me the exach formula of your sample?
If you have multilayers then you may see spin wave dispersion in inelastic neutron scattering experiments but the interpretation is tricky.
What you see below 48 K in MnS2 is just the magnetoelastic effect. The structural transition is purely spin driven. I do not have magnetization data so far. But we can measure it.
Do you know about our work on the spin wave dispersion in the bilayer CMR manganite La1.2Sr1.8Mn2O7. This is a Ruddeleden-Pooper phase with n = 2 and is really quasi two dimensional. We however modelled it by a simple Heisenber Hamiltonian.
Since it is metallic we took charge suscptibilty into account and calculated magnon damping. My collaborator Nic Shanon did these quantum correction calculations.
The samples in our experiments have ferroelectricity, so are insulators. However, due to the limitation of first-principles calculations within density of functional theory or some unexpected causes, in some magnetic structures the insulating features cannot be reproduced well. This is one of the reasons why I need inspect my data, the exchange parameters are very sensitive to those properties.
Can you synthesize 3-5 g of powder sample that is ordered? In that one can do detailed neutron diffraction as a function of temperature. That may clarify a lot of things.
Dear Prof. Tapan Chatterji,
Thanks for your comments. Unfortunately, at present we can only prepare film samples. In our coming manuscript, we will combine SQUID, XAS, XMCD data with first-principles calculations to demonstrate the possible magnetic structures with the absence of Neutron diffraction data. Actually we and our friendly colleagues don't have these experimental facilities, I think, it is rare in China.
- Badges
- Science topic
- Similar topics
- Physics
- Condensed Matter Physics