Saturday, August 05, 2017

Highlights from Telluride

Here are a few highlights from the workshop I mentioned.  I'll amend this over the next couple of days as I have time.  There is no question that smaller meetings (this one was about 28 people) can be very good for discussions.
  • I learned that there is a new edition of Cuevas and Scheer that I should pick up.  (The authors are Juan Carlos Cuevas and Elke Scheer, a great theorist/experimentalist team-up.)
  • Apparently it's possible to make a guitar amplifier using tunnel junctions made from self-assembled monolayers.  For more detail, see here.
  • Some folks at Aachen have gotten serious about physics lab experiments you can do with your mobile phone.
  • Richard Berndt gave a very nice talk about light emission from atomic-scale junctions made with a scanning tunneling microscope.  Some of that work has been written about here and here.  A key question is, when a bias of \(eV\) is applied to such a junction, what is the mechanism that leads to the emission of photons of energies \(\hbar \omega > eV\)?  Clearly the processes involve multiple electrons, but exactly how things work is quite complicated, involving both the plasmonic/optical resonances of the junction and the scattering of electrons at the atomic-scale region.  Two relevant theory papers are here and here.
  • Latha Venkataraman showed some intriguing new results indicating room temperature Coulomb blockade-like transport in nanoclusters.  (It's not strictly Coulomb blockade, since the dominant energy scale seems to be set by single-particle level spacing rather than by the electrostatic charging energy of changing the electronic population by one electron).
  • Katharina Franke showed some very pretty data on single porphyrins measured via scanning tunneling microscope, as in here.  Interactions between the tip and the top of the molecule result in mechanical deformation of the molecule, which in turn tunes the electronic coupling between the transition metal in the middle of the porphyrin and the substrate.  This ends up being a nice system for tunable studies of Kondo physics.
  • Uri Peskin explained some interesting recent results that were just the beginning of some discussions about what kind of photoelectric responses one can see in very small junctions.  One recurring challenge:  multiple mechanisms that seem to be rather different physics can lead to similar experimentally measurable outcomes (currents, voltages).
  • Jascha Repp discussed some really interesting experiments combining STM and THz optics, to do true time-resolved measurements in the STM, such as watching a molecule bounce up and down on a metal surface (!).  This result is timely (no pun intended), as this remarkable paper just appeared on the arxiv, looking at on-chip ways of doing THz and faster electronics.
  • Jeff Neaton spoke about the ongoing challenge of using techniques like density functional theory to calculate and predict the energy level alignment between molecules and surfaces to which they're adsorbed or bonded.  This is important for transport, but also for catalysis and surface chemistry broadly.  A relevant recent result is here.
  • Jan van Ruitenbeek talked about their latest approach to measuring shot noise spectra in atomically small structures up to a few MHz, and some interesting things that this technique has revealed to them at high bias.  
  • There were multiple theory talks looking at trying to understand transport, inelastic processes, and dissipation in open, driven quantum systems.  Examples include situations where higher driving biases can actually make cooling processes more efficient; whether it's possible to have experiments in condensed matter systems that "see" many-body localization, an effect most explored in cold atom systems; using ballistic effects in graphene to do unusual imaging experiments or make electronic "beam splitters"; open systems from a quantum information point of view; what we mean by local effective temperature on very small scales; and new techniques for transport calculations. 
  • Pramod Reddy gave a really nice presentation about his group's extremely impressive work measuring thermal conduction at the atomic scale.  Directly related, he also talked about the challenges of measuring radiative heat transfer down to nm separations, where the Stefan-Boltzmann approach should be supplanted by near-field physics.  This was a very convincing lesson in how difficult it is to ensure that surfaces are truly clean, even in ultrahigh vacuum.
  • Joe Subotnik's talk about electronic friction was particularly striking to me, as I'd been previously unaware of some of the critical experiments (1, 2).  When and how do electron-hole excitations in metals lead to big changes in vibrational energy content of molecules, and how to think about this.  These issues are related to these experiments as well.  
  • Ron Naaman spoke about chiral molecules and how electron transfer to and from these objects can have surprising, big effects (see here and here).
  • Gemma Solomon closed out the proceedings with a very interesting talk about whether molecules could be used to make effective insulating layers better at resisting tunneling current than actual vacuum, and a great summary of the whole research area, where it's been, and where it's going.


Anonymous said...


From fundamental principles, is it even possible to have lower conductance than vacuum? If you think of vacuum and the molecule as two resistors placed in parallel, current will through through the most conductive medium, at a given bias, which in this case would be vacuum.

The only way I can conceive of going below vacuum tunneling current would be by lowering the density of electronic states at the metal surfaces, or somehow localizing the surface electrons. It's possible that the molecular layer does just that.

Anonymous said...

Browsing that paper, I am not sure it's "better than vacuum" - the decay of the conductance with increasing distance (length of the chain) may be faster than the decay in vacuum, but that is not shown in the paper. Moreover, even if the decay with length is stronger, the absolute value of the conductance at a certain length does not need to be less than that in vacuum.

So before thinking of explanations, I'd need to see the data on which this statement was based...

Douglas Natelson said...

Anon, you've got the essential idea. The tunneling conductance between two metal electrodes at a fixed separation is set by how rapidly the electronic wavefunction decays into the space between the electrodes. For a given metal, in ultrahigh vacuum conditions, this is set largely by the work function of the metal (though image charge effects lower the effective barrier height from the actual work function). If you can make the wavefunction overlap between the two sides be less than the bare vacuum case, you should get lower tunneling conduction, all other things being equal. There was definitely discussion about whether "orbital engineering" (where you try to get the wavefunctions to be particularly confined; somewhat related to what the magnetic tunnel junction community has done in terms of barrier structure and chemistry to maximize tunneling magnetoresistance) is the really equivalent to effectively raising the work function of the two metal surfaces. In terms of the data, you are right that they can make single-molecule junction structures where conductance decays more rapidly with distance than the pure vacuum, but comparing prefactors is important and not trivial.