Very brief end of the year round-up

It's hard to believe that it's already the end of 2023.  It's been a busy year for condensed matter; it's unfortunate that two of the biggest stories (problems with high pressure superconductivity papers; the brief excitement about LK99, the not-actually-a-superconductor) were probably the field's highest profile events.  Still, hopefully the latter at least had the effect of bringing to the public a little bit of the excitement and potential of how condensed matter and materials physics affects our lives.  Physics World summarizes some of their picks for big materials-related stories of 2023 here.  Similarly, here are Quanta's choices for biggest physics stories of the year, and these are the choices from the editors of APS's Physics.  

It's been a busy year personally, with lots going on and too much proposal writing, but at least my blog posting was more frequent than in 2022.  It's still surprising to me that I've been writing this since mid-2005, enough to see almost the entire lifecycle of blogging.  Happy New Year to my readers, and if there are any particular topics about which you think I should write, please let me know in the comments.  I'm always looking for CM/materials concepts that I can try to explain on a non-specialist accessible level.  Still looking for the time and appealing perspective to write that popular book....

Anyway, I hope you have a very happy new year, and best wishes for a great 2024.

New paper - plasmons, excitons, and steering energy

We have a new paper out in Nano Letters (arxiv version here), and I wanted to explain a bit about it and why I think it's a really cool result.   

I've written before about the Purcell Effect.  When we study quantum mechanics, we learn that the rates of processes, like the spontaneous emission of light from an atom, are actually malleable.  The rate of a particular process is usually proportional to the number of ways that process can happen - this is quantified in something called Fermi's Golden Rule.  When we are talking about something like emission of light from an atom, the rate is proportional to the number of possible final states of the photon.  We know how to count those states in a given energy range in free space, and Purcell pointed out that by placing that atom in an optical cavity, we alter the density of final states as a function of frequency, \(\rho(\omega)\) from its empty space value, and hence can change the rate of emission.  Pretty wild that placing a system in a cavity can alter the flow of energy in that system away from what it would otherwise be.

I've also written before about what happens we take two resonators and couple them together - we get "hybridization" or "new normal modes".  If you take a mass on a spring (natural frequency \(\omega_0 = \sqrt{k/m}\)) and couple it mechanically to another identical mass on an identical spring, the coupled system will now have two resonances, one above and one below \(\omega_{0}\).  The chemistry analog of this is, bonding two hydrogen atoms (each with 1s orbitals) together leads to two \(\sigma\) orbitals, one bonding and one antibonding.  

In the new paper, we start with a little metal tunnel junction that hosts plasmonic resonances, like the junctions I wrote about here.  We showed in that paper and subsequent work that it is possible to use an applied voltage and current to get some of the electrons, right near where the electrodes almost touch, to become effectively so hot that they glow (emitting light at energies larger than the applied voltage), while the atomic lattice itself remains cold.  The light emission process here is the radiative recombination of hot electrons and holes in the metal, where an electron drops down in energy to fill in a hole and spit out a photon.  The plasmon resonances of the bare metal act like a sort of cavity, shaping the density of photon states \(\rho(\omega)\), as we also showed here.  The plasmons, set by the metal shape and electronic properties, actually affect the rate at which the electrons and holes in that same metal radiatively combine.

Left: A thin flake of WSe2 is placed on a plasmonic
Au junction.  Right: Overbias light emission from the
device at a particular emitted polarization shows a big
peak splitting right around where the exciton resonance
is of the WSe2 (orange curve).  Adapted from the
SI of this paper.

The wrinkle in the new paper is that we couple that metal plasmonic junction with a thin (few nm) layer of 2D semiconductor by placing the semiconductor on top of the metal.  The semiconductor can host excitons, bound electron-hole pairs, and if the semiconductor is excited with enough energy to create them, the excitons can radiatively annihilate, leading to a comparatively narrow resonance at an energy that overlaps the plasmon resonances of the metal junction.  Thanks to hybridization between the plasmons in the metal and the excitons in the semiconductor, the photon density of states now has a split peak structure ("upper and lower plexciton polariton resonances" if you are an expert).  Light emission in this device is still due to recombination of electrons and holes in the metal, but now the recombination dynamics of those electrons "feels" the strong coupling between the excitons and plasmons.  (The polarization of the emitted light is rather complicated because of the polarization properties of the plasmon resonances).  

There are a lot of interesting possibilities on where to go from here, but it's always amazing to me to see how this physics comes together.  In this case, by changing the optical environment of a metal structure, we can alter the fate of energy stored in the electrons of that metal.  Really neat.

AI/ML and condensed matter + materials science

Materials define the way we live.  That may sound like an exaggeration that I like to spout because I'm a condensed matter physicist, but it's demonstrably true.  Remember, past historians have given us terms like "Stone Age", "Bronze Age", and "Iron Age", and the "Information Age" has also been called the "Silicon Age".  (And who could forget plastics.)

Perhaps it's not surprising, then, that some of the biggest, most wealthy companies in the world are turning their attention to materials and the possibility that AI approaches could lead to disruptive changes.  As I mentioned last week, there have been recent papers (back to back in Nature) by the Google Deep Mind group on this topic.  The idea is to use their particular flavor of AI/machine learning to identify potential new compounds/solids that should be thermodynamically stable and synthesizable, and make predictions about their structures and properties.  This is not a new idea, in that the Materials Genome Initiative (started in 2011) has been working in this direction, compiling large amounts of data about solid materials and their properties, and the Materials Project has been pushing on efficient computational methods with the modest goal of computing "the properties of all inorganic materials and provid[ing] the data and associated analysis algorithms for every materials researcher free of charge".

In addition to the Google work, Microsoft has released on the arxiv their effort, MatterGen, which uses a generative AI approach to try to predict new stable materials with desirable properties, such as a target symmetry or chemical composition or mechanical/electronic/magnetic response.  An example from their paper is to try to find new magnetic materials that have industrially useful properties but do not involve rare earths.  

There is a long way to go on any of these projects, but it's easy to see why the approach is enticing.  Imagine saying, I want a material that's as electrically conductive and mechanically strong and workable as aluminum, but transparent in the visible, and having software give you a credible approach likely to succeed (rather than having to rely on a time-traveling Mr. Scott).  

I'd be curious to know readers' opinions of what constitute the biggest obstacles on this path.  Is it the reliability of computational methods at predicting formation energies and structures?  Is it the lack of rapid yet robust experimental screening approaches?  Is it that the way generative AI and related tools work is just not well-suited to finding truly new systems beyond their training sets?

Intriguing papers - exquisite thermal measurements + automated materials discovery/synthesis

It's a busy time, but I wanted to point out a couple of papers from this past week.

First, I want to point to this preprint on the arxiv, where the Weizmann folks do an incredibly technically impressive thing.  I'd written recently about the thermal Hall effect, when a longitudinal heat current (and temperature gradient) in the presence of a magnetic field results in a transverse temperature gradient as well as the usual longitudinal one.  One of the most interesting ways this can happen is if there are edge modes, excitations that propagate around the perimeter of a 2D system and can carry heat (even if they are neutral and don't carry charge).  Unsurprisingly, to measure thermal transport requires putting thermometers at different places on the sample and carefully measuring temperature differences.  Well, these folks have done just exquisitely nice measurements of Johnson-Nyquist noise in particular contacts for thermometry, and they can see the incredibly tiny heat currents carried by rather exotic edge modes in some unusual fractional quantum Hall states.  It's just a technical tour de force.

Second, on a completely unrelated note, there are back to back papers in Nature this week from the Google deep mind folks - their own write-up is here.  The first paper uses their methods to predict a large number of what are expected to be new stable crystal structures.  The second paper talks about how they used an automated/robot-driven lab to try to synthesize a bunch of these in an automated way and characterize the resulting material.  This is certainly thought-provoking.  It is worth noting that detailed characterization (including confirming that you've made what you were trying to make) and optimized synthesis of new materials is very challenging and of concern here.  Update:  there is further discussion of the characterization here (on LinkedIn by the authors) as well, and more on Twitter here and here.

Third, this paper looks extremely interesting.  It’s long been a staple of condensed matter theory to try to capture complex materials with effective low energy models, like suggesting the Hubbard model as a treatment of the essential physics of the cuprate superconductors.  The authors here report that they’ve done a more orbital-based/ab initio version of this, solved these models numerically, and state that they can reproduce details of the phase diagram of four of the cuprates spanning a big range of superconducting transition temperatures.  Seems like this may bode well for gaining insights into these systems.