The unexpected - the fun part of experimental science

There's an old quote from Isaac Asimov that is very true:  "The most exciting phrase to hear in science, the one that heralds new discoveries, is not 'Eureka!' (I found it!) but 'That's funny...'."  In my group, we recently had an experience that supports this, and it highlights what I think is some of the most fun you can have as an experimental scientist:  trying to use the tools at your disposal to learn as much as you can about what's behind some unexpected and surprising phenomenon.

The story starts out several years ago.  We'd been having some nice success making single-molecule electronic junctions and using them as model systems to study a particular piece of physics, the Kondo effect.  Our theorist colleagues pointed out that these molecular devices, unlike typical semiconductor quantum dots, might give us an opportunity to study a particularly rich and interesting piece of physics called a quantum phase transition, because molecular devices are easier to attach to ferromagnetic electrodes.  The idea, not directly germane to this story, is that an unpaired electron on the molecule is torn between two competing "baths" of excitations.  On the one hand, the unpaired electron can undergo Kondo processes with the conduction electrons of the electrodes.  On the other hand, spin waves in the ferromagnetic electrodes can also talk to the unpaired electron.  By varying a gate voltage appropriately, the hope was to tune from one limit (the Kondo regime) into the other (expected to be a more exotic "non-Fermi liquid" state).  

Anyway, for various reasons, it became clear that working with palladium electrodes might be a good place to start.  Palladium is almost ferromagnetic.  That means that it has long-lived spinwave-like excitations (paramagnons).  At the same time, it's chemically friendlier (less prone to forming magnetically complicated oxides) than common ferromagnetic metals like iron, nickel, or cobalt.  So, step zero of this project would be to make some bare palladium tunnel junctions (just two pointy palladium electrodes, without any molecule bridging them) and make sure that they're simple and boring, as expected.  After all, we'd looked at literally thousands of gold tunnel junctions like this, and if properly made (so that you don't have extra metal nanoparticles around), they are dull as dirt:  current-voltage (I-V) curves that are nearly linear and essentially temperature-independent.

Surprise!  My postdoc, Gavin Scott, found that Pd tunnel junctions are very much not boring.  While they look dull at, say, 10 K, if they are cooled down to lower temperatures, all kinds of sharp features appear in their differential conductance (dI/dV as a function of V).  The features appear at voltages symmetric around V = 0, and they evolve with temperature in a very interesting way.  In fact, if you look at the temperature dependence of those features, it looks very much like what you see for the temperature dependence of the order parameter in a "mean field" phase transition.  We spent months trying various things, turning all the easily turned "knobs" like temperature, magnetic field, gate voltage, etc.  One striking trend is the observation that, looking at all of our devices on one set of axes, the voltages where the conductance features appear extrapolate to zero when the conductance of the junction approaches e²/h.  In other words, when the metal tips touch, the whole effect goes away.   

It's been science-as-puzzle-solving, trying to figure out what could be going on here.  We came up with many possible explanations, and tried to come up with ways to test the possibilities, eliminating the ones that didn't fit.   For example, the data look (qualitatively) rather like superconductor tunnel junctions, but the quantitative values (specifically, the relationship between the voltage scale and the temperature scale) are far, far away from numbers that would make sense for a superconductor.  In the end, we think that the most likely physics ingredient is the onset of magnetic order at the tips, though that's not a perfect explanation by any means.  It is clear, though, that Pd is special - other metals (Au, Ni, Pt) just don't seem to show this.  The paper is out here (email me if you want a copy).  Hopefully others will get interested.  Suggestions are always appreciated.  In the meantime, it's a good example of how sometimes systems that you think are dull can surprise you, and how science is supposed to work.

Nanotechnology-enabled sensing

Last May, I participated in a very well run and informative workshop on nanotechnology for sensing purposes.  The workshop was run by the National Nanotechnology Initiative (and associated Federal agencies).  The report that we ended up producing is now available for download.  Here is a direct link to the pdf.  Other NNI reports and information may be found here.

A word of caution re: departmental rankings

The US News rankings of US graduate programs are out again, and I've heard a fair bit of discussion about them.  My department, for example, went up slightly in the rankings, while the chemistry department here slipped a few spots.  I want to point out something that US News makes no effort to broadcast:  The US News graduate rankings are a popularity contest.  What I mean is, the US News rankings are the result of an opinion survey taken of department chairs, not the result of actual quantitative metrics like publication rates, citation rates, research funding, major awards, graduation rates, or the like.  Essentially the rankings give you a snapshot of the perception of the community of department chairs in a discipline, not an actual real ranking of some defined quality.  This has some consequences.  For example, perceptions are very hard to change, so it's unlikely that there will be lots of movement on these rankings unless there are exceptional circumstances (e.g., a particular department wins a couple of Nobel prizes out of the blue).  It is distressing to me to see how much importance some people (prospective students on the one hand, administrators on the other) place on rankings that measure reputation rather than something truly quantitative.

The NRC, by contrast, does survey real data like those mentioned above.  Unfortunately, they seem to be in a mode of continuously delaying the release of their "decadal survey".  Anyone have any more insight into why that is taking so long? 

An ethical dilemma

Here is a scientific ethical dilemma that came up in conversation recently.  What do you do if you get a positive referee report, but it's clear from the comments that the referee completely misunderstood your manuscript, or basically had no clue at all?  Do you point this out to the editors?  I'm actually a bit surprised that this never seems to come up, given how often people complain about the complementary case of a clueless referee that bashes a paper because of a lack of understanding....

This week in cond-mat

Three interesting pieces of physics (among many) from the arxiv this week:

arxiv:1004.0546 - Mak et al., Atomically thin MoS₂:  A new direct-gap semiconductor
The electronic structure of ordinary semiconductors is usually presented, in textbooks, in the context of band theory, neglecting electron-electron interactions and assuming infinitely large crystals.  In the case of the layered dichalcogenide, MoS₂, the structure of the bulk is that of an indirect gap semiconductor.  The highest filled (single-particle) electronic states (at the top of the valence band) are labeled by wavevectors k that are near zero.  That is, the wavelike electronic states have very long wavelengths.  The (energetically) lowest empty states happen to have k values that are away from zero.  That means, for example, that the energetically cheapest way to kick an electron from the valence into the conduction band requires enough momentum (probably made up via a phonon) to make up the difference in k vectors.  The authors of this paper have observed something interesting:  as this layered material is made thinner and thinner, down toward the atomic limit, the band structure changes.  The finite-k conduction states go up in energy relative to the k = 0 conduction band states, so that the material becomes instead a direct gap system.  Neat stuff.


arxiv:1004.1233 - Cabrera et al., Oblique propagation of electrons in crystals of germanium and silicon at sub-Kelvin temperature in low electric fields
This paper also involves the concept of indirect-gap semiconductors.  Silicon and germanium are both indirect gap systems, so that the lowest energy electronic states in the conduction band live in "valleys" far from k = 0.  That means that if processes that allow intervalley scattering (such as inelastic interactions w/ phonons) are turned off, the motion of conduction electrons in real space has to reflect these valleys.  Of course, the structure of the valence band in k space is quite different than the conduction band.  That means that holes will propagate differently in real space under the same circumstances.  These folks, part of the CDMS collaboration who have been using cryogenic Si and Ge crystals as dark matter detectors, have completed a clean study of this.  Great stuff.

arxiv:1004.1202 - Cheng and Robbins, Defining contact at the atomic scale
As my students and I have encountered, particularly in a paper currently out for review, sometimes it can be very challenging to define what we mean when we say two pieces of material are touching at the one atom or two atom level, or what me mean when we say they're not touching, but instead are a certain distance apart.  These authors take a hard look at this topic in terms of the forces between the two sides.

Medical physics FTW

Blogging will be on hold for a few days, as I'm going to have an appendectomy very soon. Thank you, medical physicists, for developing the wicked cool CT machine (and blech-tasting contrast agents) key to confirming my diagnosis! Thanks, too, to the scientists and engineers that gave us wireless Internet.....

UPDATE! I had the surgery this morning and it seemed to go smoothly. Barring weirdness, I should be back up to some kind of speed in a couple of more days. Thanks for the sentiments!

Superconductivity induced in a ferromagnet

Four years ago, this paper caught my attention.  The authors had made a structure with superconducting NbTiN contacts on top of a CrO₂ film, with the intent of studying how superconductivity leaks into the chromium dioxide.  The "leakage" of superconductivity into a non-superconducting metal is called the proximity effect.  In a normal metal, the proximity effect extends over a spatial scale comparable to the coherence length, the distance that the electrons can travel before their quantum mechanical phases become scrambled due to inelastic processes (such as electron-electron scattering, or spin-flip scattering from magnetic impurities).  The coherence length in a normal metal can be quite long at low temperatures - say a micron in a clean normal metal at 1 K.  

Now, CrO₂ is not a normal metal.  Rather, it is a half-metal, an extreme limit of an itinerant ferromagnet, where all of the mobile charge carriers have the same spin polarization.  This is important, because ordinary ("s-wave") superconducting correlations rely on pairing up electrons with opposite spins and momenta.  If only one spin polarization is allowed, that should preclude any s-wave superconductivity.  Practically speaking, in a typical ferromagnet with some magnetic exchange characterized by an exchange energy U, one can define an exchange length (in a diffusive material, given by sqrt(\hbar D/U), where D is the diffusion constant for the electrons) over which these correlations should die.  For a strong ferromagnet, one finds that the exchange length is very short - a few nanometers or less.  Knowing this, one would not expect to see any proximity induced superconductivity in a ferromagnet over longer distances.  That's why this paper was surprising - the authors did see evidence of long-range (hundreds of nm) superconductivity in the ferromagnetic oxide.  This implies some kind of unusual superconductivity in the ferromagnet - either p-wave pairing (when each pair of electrons in the superconducting material has one quantum of orbital angular momentum), or some more exotic state ("odd-frequency pairing", for the experts).

Several years passed, and no one reproduced this result.  Until now.  The authors of this new paper see the same sort of thing, and they try to explain in detail why this has been so hard to reproduce.  The short version:  CrO₂ is a pain to work with.

Interestingly, there have been other signs of similar effects within the last year.  For example, the Birge group at Michigan State has reported long-ranged proximity superconductivity induced in cobalt layers, though careful engineering of the contacts was required.  Likewise, a Penn State collaboration has seen proximity superconductivity in Co nanowires hundreds of nm long.  It's nice to see so much progress in this area lately.

Slightly missing the point

I am, of course, pleased that Nature Materials ran a news item about our recent paper.  However, they appear to have missed the point a bit.  What we were trying to point out was that plotting data in scaled coordinates (current normalized by temperature to some power vs. voltage normalized by temperature, in this case) can be misleading.  In this particular case, plotting temperature-independent data in this way on a log-log plot can make it look like the data all collapse onto some universal curve (with deep implications).  In fact, the data themselves aren't doing any such thing - the apparent collapse is due to a flawed plotting procedure.  Ross McKenzie got this point immediately.  Ahh well.  Bottom line: be very careful when plotting "scaled" quantities, to make sure that you're not biasing yourself toward a particular conclusion.

Three years of hindsight

Back in February, 2007, I mentioned that I thought that making and measuring nanostructures directly out of strongly correlated materials was a promising research direction.  I still think so, in part because of the experiences my lab and others have had in the mean time.  As I've mentioned before, strongly correlated materials are those where we can no longer get away with our (often miraculously good) simple single-particle description of electronic structure that ignores the electron-electron interaction.  Examples of correlated materials include the high temperature superconductors and various transition metal oxides.  In the high T_c case, single-particle band structure says that the undoped parent compound should be a metal, when it's found instead to be an antiferromagnetic insulator.  Strongly correlated materials often display a rich variety of interesting electronic states as well as phase transitions between them.

Why nanostructures?  Three main reasons.  First, when some strongly correlated materials go through electronic (and structural, sometimes) transitions, they display inhomogeneities.  For example, when vanadium dioxide goes through its metal-insulator transition (from below) near 341 K, metallic domains nucleate and grow, eventually encompassing the whole material.  Nanostructures can allow you to access such materials on the scale of individual domains, as was done here and here in VO₂, and here in a colossal magnetoresistance oxide.  

Second, nanospaced electrodes allow you to use electric field as an interesting perturbation, and to tell the difference between effects driven by voltage (or energy) and those actually driven by field.  For example, in this paper (very influential and highly cited), the authors found that a charge-ordered oxide could be kicked out of an insulating state and into a low resistance state when around 1000 V was applied across a 1mm crystal of the material.  That's cool, and prompted much work, but as an experimentalist one always wonders whether the 10⁶ V/m electric field is responsible (which would be particularly interesting), or whether the energy available to the electrons is actually doing some kind of chemistry or damage to the material.  In a nanostructured system, one could get the same electric field by applying 100 mV across electrodes separated by 100 nm, getting big fields without big available energies.  We've done work along these lines in Fe₃O₄, and have had quite a bit of fun with it.

Third, nanospaced electrodes are a way of driving systems far from equilibrium and potentially measuring them under those conditions.  When an ordinary electron is injected into a correlated material, it can be thought of as a superposition of the "natural" low-energy excitations of the correlated system.  For example, in a Luttinger liquid, the injected carrier "fractionalizes" into a spinon (spin-1, chargeless mode) and a holon (charged, spinless mode).  Truly nano-spaced measurements may be a means of catching this process in action, though it won't be easy!

I still think that this is a fun and exciting area, and interest appears to be growing.

March Meeting wrap-up

The March Meeting is over, and overall it was good, as usual, and far too large, as usual.  Clearly the favorite topics this year were topological insulators, graphene, and iron pnictide superconductors.  On Thursday I did see a very good invited session on scanned probe microscopy, including examinations of vortices in the iron pnictides (among other things) and careful measurements of spin excitations at the atomic scale (that last being a substitute talk by Prof. Hla of Ohio University).  I had good discussions with colleagues from other places, got some good ideas for different experimental techniques, and looked at all the gadgets being hawked by vendors.  Clearly the era of the cryogen-free dilution refrigerator is upon us, if you can actually get any ³He, have a spare couple of hundred thousand dollars, and can support a 6-10 kW compressor.  I was disappointed by turnout at an invited session that I'd helped organize (more in a separate post), but it was up against a session with talks by three Nobel laureates.  On the return flight, I had a fun time talking with the neighboring passenger, the drummer for Ra Ra Riot, on his way to a gig in Austin.  Good to be home, though.

Physics for Everyone

This was the title of an invited session yesterday at the APS meeting.  Ivan Schuller put together a very interesting collection of talks, with subjects of broad interest to a large audience.  Ray Orbach gave an updated version of his presentation about the global energy challenge, including a strong plea that the US re-start reprocessing spent nuclear fuel (an idea I've held for a long time, so clearly it has merit :-) ).  Ted Postol spoke about his ideas on ballistic missile defense.  Charles Falco presented his recent work on converting digital cameras from conventional visible light photography to infrared (and recently ultraviolet) wavelengths, and using those cameras to look at art works.  Fascinating stuff.  Eugenie Reich gave a half-hour talk to a packed room about Hendrik Schön (see this pdf for the whole megillah), based on her book.  No revelations, and surprisingly little discussion of co-author/collaborator responsibilities.  She did have some amusing graphs, like the one displaying Schön's publications vs. time next to the Lucent stock price vs. time.  Finally, the session closed with a fun talk by Alan Nathan about the physics of baseball.  These kinds of sessions are a great feature of large meetings, though they must be a lot of work to put together.

March Meeting - minor update

I'm at the APS March Meeting, and as it is every year, it's really too big.  I've largely been spending much of my time at sessions where my students are speaking or where I'm chairing.  One major point of the meeting is the networking that takes place away from the talks.  Even in these days of modern communications technology, there is still no substitute for sitting down at a table with someone and hashing out a problem with a pen and a pad of paper.  

There are two types of talks:  invited talks (30 min + 6 min for questions) and contributed talks (10 min + 2 min for questions).  Contributed talks are usually most useful to specialists in a field, since with that kind of time constraint it's almost impossible to give an intro to something new.  Occasionally, however, it is possible to learn something new from a contributed talk.  I just saw an excellent one by Kieron Burke of UC Irvine, speaking about how one can get from density functional theory (which sometimes feels like a mysterious black art) and get back to simple Thomas-Fermi theory.  The relevant paper is here, and I really do feel like I learned something.  More later....

Peer review idiocy

In the Wall Street Journal over this past weekend, Peter Berkowitz argues that peer review is a corrupt system with no objectivity and little value.  He says this in connection with the climate science kerfluffle, claiming that peer review makes it easy for scholars to reward friends and punish enemies.  He argues by analogy:  we don't let athletes referee their own games, so why should we allow scholars to do the equivalent?  He does somehow seem aware that there is a real fundamental problem with ditching the system, though - the ability to competently evaluate intellectual work requires actual expertise.  I think Berkowitz gravely underestimates the qualification/expertise problem when it comes to actual physical science.   Sometimes the particular technical areas are incredibly challenging, requiring years of study to appreciate the subtle problems and issues.  Your choice is to let the people who have done such legwork do your evaluations, or to let people without the proper background make decisions.  Those of us who do science know that peer review has its set of problems, but, like democracy, it's the worst system except for all the alternatives.  Harping on its flaws without having a real discussion of these difficulties or offering alternatives is just intellectually lazy.

Narrowcasting.

I know that this will appeal to a differentially small fraction of my readership, but as a native of Pittsburgh of a certain age I am compelled to share this video.  An actual science post will follow tomorrow.  (For those readers that have never spoken with me, no, I do not talk with this kind of accent. But I can.)

Self-promotion, travel, and talking heads

I'm going to be doing some travel this week, and next week is the APS March Meeting in Portland, OR, so blog posts are likely to be thin.  I will try to write a bit about the APS.

In the meantime, I wanted to point out our new paper about shot noise measurements in atomic-scale junctions.  This is the paper I was talking about when posting about (good!) referees.  The referee reports were very helpful in making our presentation much more clear here.

Lastly, with my previous post about being a nanoscale science "talking head", I wanted to point out this terrific video that shows many of the things wrong with TV journalism today.  The dry British sense of humor is tough to beat.

Comedy Central: better science than the Science Channel?

Zapperz wrote an interesting post that links to an article in USA Today about whether Comedy Central (specifically The Daily Show and The Colbert Report) give some of the best real science coverage on television.  They bring on actual scientists (Neil Degrasse Tyson, Sean Carroll, Brian Greene, Lisa Randall, Steven Chu, Bob Park (!)) and have conversations with them that last more than a 30 second sound bite.  I notice that there are no nano folks on the guest lists for either show over the last couple of years.  Clearly this is a clarion call to write a general audience book and try to go on there to promote it; either that, or to try and be their go-to person to debunk outrageous claims about nanotechnological dystopias.

Topological insulators

A very big story in recent years in condensed matter physics is that of topological insulators, and it's a great tale of finding something new "in plain sight".  For something like 70 or 80 years, physicists thought that they had a handle on the insulating state.  Take a large crystalline solid, and ignore electron-electron interactions for the moment.  The allowed electronic states for such a material come in bands, when you look at how they're distributed as a function of energy.  That is, there are many electronic states clustered so close together in energy, separated by energy gaps where there are no allowed electronic states.  Now consider filling up those states with some number of electrons, counting two electrons (one spin up, one spin down) per state.  (This is short-hand for something more sophisticated, but it gets the point across, just as filling up atomic orbitals does a pretty good job describing the periodic table.)  If you end up in the middle of a band, with lots of empty states right next to the filled states in energy, then you have a metal.  If you end up exactly filling a band, then you have either a band insulator (if the energy gap next to the most energetic filled state is several eV) or a semiconductor (if the energy gap is more like 1-3 eV).   Turning on electron-electron interactions can change things a bit, but not too much.  (Interactions can lead to one more kind of insulator, a Mott insulator, in which interactions open up a gap in what would otherwise have been a metallic system.)

Until recently, we thought that this was all there was to it, as far as band insulators go.  It turns out that this is not the case, because of what happens at the boundary of the material (which we have so far been ignoring).  In some band insulators, the surface states (in 3d) or edge states (in 2d) can have special properties.  For example, one could have a situation where (because of spin-orbit coupling + band structure) the right-moving charge carriers have to have their spin pointed in one direction, while the left-moving charge carriers have to have their spin pointed the opposite way.  The result is that these surface states with unusual Dirac-like dispersion are thought to be able to resist back-scattering very effectively.  This means that these surface states may be very interesting for electronics applications, having ballistic properties over long distances.  Moreover, these properties are expected to be rather robust, because they come from the topology of the states, which is not easily disturbed by disorder.  For great reviews of this, see this paper (soon to appear in RMP), this article in Physics Today, and this video.

There is evidence that these states do exist, particularly from surface scattering techniques such as ARPES.  Transport experiments have faced a challenge, however, since many of the candidate materials (Bi₂Se₃, for example) are difficult to grow in sufficient purity that the bulk is actually insulating.  Still, this is exciting stuff, and a new paper on the arxiv (1003.0155) reveals that there may be a whole new class of other materials to play with.  Surprises from nature are always fun.   

Amazing technology.

It's not nano-related, but I found this video fascinating.  Clearly technology has gotten to the point where Hollywood can fake just about anything, often on a TV budget.

Claims of priority

I was looking at this week's Phys Rev Letters, and I saw this paper being highlighted as an editor's suggestion.  Now, I don't know anything about this work, but I was struck by the title, which says explicitly that this measurement is the first of its kind.  This is repeated a few paragraphs into the paper at the end of their introduction.  I thought that claims of "first"s were very strongly discouraged by the editors, as mentioned here, let alone being included in the title, regardless of how well founded the statement.  Was this an oversight, or am I missing something?

High Tc, pseudogaps, broken symmetries

What distinguishes one phase of matter from another?  A physicist would probably say that different phases possess different symmetries.  More specifically, transitions between phases can be described (when going in the right direction) by the breaking of a symmetry.  For example, when water freezes, the continuous rotational and translational symmetry of the liquid (liquid water looks, on average, the same in every direction and at different points within the liquid) are broken, because crystalline ice has a specific lattice (and therefore certain preferred lattice directions, as well as a spatial periodicity).  Solid ice instead has discrete rotational and translational symmetries, rather than continuous ones.

High temperature superconductors have been confounding physicists for 24 years now.  Progress has been made in understanding these complicated materials (typically layered, multicomponent copper oxides with weird oxygen stoichiometries to control the number of mobile charge carriers), but the situation is still a mess.  These compounds have a complicated phase diagram as a function of, e.g., temperature and chemical doping.  The undoped parent compounds are antiferromagnetic insulators.  Over a range of chemical compositions, the ground state is a d-wave superconductor.  Within a good part of that range of composition, at temperatures above the superconducting transition, these materials show a "pseudogap" below some higher temperature, T*.  That is, the number of available electronic states near the Fermi level is depressed compared to what you'd expect for a metal, but not vanishing as you'd expect for a superconductor.  People have been arguing for years about what the pseudogap is - is this a distinct phase?  Is it a precursor to superconductivity (e.g., pairing of electrons w/o long-range coherence), or does it compete with superconductivity?

This recent paper by Louis Taillefer reports the observation of broken rotational symmetry in the pseudogap phase (mainly via the Nernst effect).  The claim is that below T*, the four-fold rotational symmetry (because it's a square lattice) of the electronic properties of the CuO2 planes is broken, and the system becomes electronically anisotropic.  This is important, because it firmly argues that the pseudogap state is a real thermodynamic phase of some kind, and that kind of broken symmetry apparently places strong constraints on possible theories of high Tc.  Not my direct area of expertise, but it looks very interesting.  I'll admit, though, I was surprised by the strong statements made here.  Unless there's way more to this than meets the eye, it's not clear to me why it's justified to claim that we're now much closer to room temperature superconductivity....