Saturday, December 16, 2017

Finding a quantum phase transition, part 2

See here for part 1.   Recall, we had been studying electrical conduction in V5S8, a funky material that is metallic, but on one type of vanadium site has local magnetic moments that order in a form of antiferromagnetism (AFM) below around 32 K.  We had found a surprising hysteresis in the electrical resistance as a function of applied magnetic field.  That is, at a given temperature, over some magnetic field range, the resistance takes different values depending on whether the magnitude of H is being swept up or back down. 

One possibility that springs to mind when seeing hysteresis in a magnetic material is domains - the idea that the magnetic order in the material has broken up into regions, and that the hysteresis is due to the domains rearranging themselves.  What speaks against that in this case is the fact that the hysteresis happens over the same field range when the field is in the plane of the layered material as when the field is perpendicular to the layers.   That'd be very weird for domain motion, but makes much more sense if the hysteresis is actually a signature of a first-order metamagnetic transition, a field-driven change from one kind of magnetic order to another.   First order phase transitions are the ones that have hysteresis, like when water can be supercooled below zero Celsius.

That's also consistent with the fact that the field scale for the hysteresis starts at low fields just below the onset of antiferromagnetism, and very rapidly goes to higher fields as the temperature falls and the antiferromagnetic state is increasingly stable.   Just at the ordering transition, when the AFM state is just barely favored over the paramagnetic state, it doesn't necessarily take much of a push to destabilize AFM order.... 

There was one more clue lingering in the literature.  In 2000, a paper reported a mysterious hysteresis in the magnetization as a function of H down at 4.2 K and way out near 17-18 T.  Could this be connected to our hysteresis?  Well, in the figure here at each temperature we plot a dot for the field that is at the middle of our hysteresis, and a horizontal bar to show the width of the hysteresis, including data for multiple samples.  The red data point is from the magnetization data of that 2000 paper.  

A couple of things are interesting here.   Notice that the magnetic field apparently required to kill the AFM state extrapolates to a finite value, around 18 T, as T goes to zero.  That means that this system has a quantum phase transition (as promised in the post title).  Moreover, in our experiments we found that the hysteresis seemed to get suppressed as the crystal thickness was reduced toward the few-layer limit.  That may suggest that the transition trends toward second order in thin crystals, though that would require further study.  That would be interesting, if true, since second order quantum phase transitions are the ones that can show quantum criticality.  It would be fun to do more work on this system, looking out there at high fields and thin samples for signatures of quantum fluctuations....

The bottom line:  There is almost certainly a lot of interesting physics to be done with magnetic materials approaching the 2d limit, and there are likely other phases and transitions lurking out there waiting to be found.

5 comments:

Anonymous said...

do you see evidence of (quantum) fluctuations ?
I always thought that extrapolating a phase boundary line to zero Kelvin is not sufficient proof to claim QPTs.
In fact, looking at e.g. Subir Sachdev's work, the general feel I get is that it is specifically the presence of fluctuations above the QCP that "proves" (ahem, more honestly: makes it plausible) that there is a QPT.

This is not criticism, just an invitation to hear your opinion, specifically as it relates to the claims in your own paper.

Douglas Natelson said...

Anon, we haven’t really done the right measurement (which would be out at 18 T or so) to look for quantum fluctuations. Bear in mind, you can certainly have a quantum phase transition (that is, a transition at zero temperature between two different ground states as a function of some tuning parameter in the Hamiltonian) without quantum criticality. You only get critical fluctuations if the quantum phase transition is continuous (analogous to second order). This one might not be.

DanM said...

Hey Doug,
If this material has a Neel temperature of 32K, then it's plausible that it has interesting excitations in the terahertz range. I don't know V5S8, but other similar materials have low-lying magnon modes. Want us to do transmission spectroscopy on it? We can go up to 9 tesla with a field perpendicular to the substrate (but not in plane).

Douglas Natelson said...

Hi Dan - We still have samples, but I don't think they're on the appropriate substrate. AFAIK ours are on doped Si, and I'm not sure of the doping without checking. My memory is that doped Si is bad from the point of view of THz transmission - is that right?

DanM said...

Yes, that's right. If you could get them (or transfer them) to undoped silicon (greater than 10,000 ohm-cm), it would be perfect. We have wafers, if you need one. Also keep in mind that our beam is a few millimeters in diameter, so a large area coverage is required.