Sunday, March 25, 2012

Responsibilities, rational and otherwise

Professors have many responsibilities - to their students and postdocs, to their departments and colleagues, to their university, to the scientific community, and to the public. When on a doctoral committee, for example, a professor's duty is to make sure that the candidate's thesis is rigorous and careful, and that the student actually knows what they're talking about. Obviously primary responsibility for supervision of the student lies with the advisor(s), but the committee members are not window dressing; they're supposed to serve a valuable role in upholding the quality of the work.

I have a colleague at another institution (names and circumstances have been changed here; I'll say no more about specifics) who really had to put his foot down several years ago, as a committee member, to make sure that a student (the last one of a just-retired professor) didn't hand in a thesis sufficiently fringe that it bordered on pseudoscience. It was pretty clear that the advisor would have been willing to let this slide (!) for the sake of getting the last student out the door. My colleague (junior faculty at the time) had to push hard to make sure that this got resolved. Eventually the student did complete an acceptable thesis (on a much more mainstream topic) and got the degree. This colleague just recently came across the former student again, and was disappointed and sad to see that the fringe aspects of science are back in what he's doing. My colleague is now feeling (irrational) guilt about this (that the former student is now credentialed and pushing this stuff), even though the actual thesis was fine in the end. This does raise the question, though: how much of a gatekeeper should a committee member be?

Sunday, March 18, 2012

Paranormal activity edition

Two items, oddly about parapsychology (as a means to raise points about science and the public).  First, this article from The Guardian last week is both unsurprising and disappointing.  It is not at all surprising that careful attempts to reproduce almost-certainly-spurious results implying precognitive phenomena have shown that those effects apparently to not really exist.  What is worth pondering and discussion, however, is the fact that the authors who tried to check the original results had such a hard time publishing their work, because the major journals dismiss attempts to reproduce controversial results as unoriginal or derivative.  This is a problem.  Sure, you don't want to take up premiere journal space with lots of confirmations or repetitions of previous work.  However, if a journal is willing to hype controversial results to boost circulation, then surely there is some burden on them to follow up on whether those extraordinary claims withstand the test of time.  

Second, this morning's Dear Abby column (yes, I still read a newspaper on Sundays) had a letter from a woman seeking advice about how to use her "psychic gifts".   It's very depressing that the response said "Many people have psychic abilities to a greater or lesser degree than you do, and those "vibes" can be invaluable."  Really?  Many people have psychic abilities?  How's this for advice:  if you really have psychic abilities, go to the James Randi Foundation and take their Million Dollar Challenge.  Once you pass, you can use the money to make peoples' lives better.  I know it's stupid to get annoyed by this, just as it's pointless to complain about the horoscopes that run in the paper.  Still, if someone has an audience as large as Dear Abby, they should think a little bit about spreading this silliness.

Friday, March 16, 2012

Tidbits

Some interesting and thought-provoking things have come up in the last week or so. For instance, here is an article from the IEEE that discusses the decline in science and engineering jobs in the US. Figure 2 is particularly thought-provoking, showing that the number of US undergrad STEM degrees is very strongly correlated with the number of non-medical US federal research dollars spent, from 1955-2000. My personal take is, if you really want Americans to become scientists, engineers, and more broadly supportive of technical education, you need to create a culture where those professions are (more) respected and valued, not viewed as nerdy, geeky, asocial, elitist, or otherwise unacceptable.

On this same theme, there was this op-ed in the New York Times about why so few American political figures are scientists. Accurate (in my opinion) and depressing. I'm not saying we should live in a society run by technocrats, but surely we can be better than this. As a culture, do we really need more lawyers and undergrad "business" majors?

On a more technical note, the ICARUS collaboration, another group in Gran Sasso in Italy working with neutrinos produced by CERN, has announced (paper here) that their measurements show neutrinos traveling at a speed consistent w/ c. Not surprising, and only truly independent measurements can really pin down the issues w/ the OPERA work.

Here is a beautiful new paper by the Manoharan group at Stanford. By arranging spatially ordered arrays of CO molecules on a copper surface, they can manipulate surface states in a way that produces dispersion relations (the relationship between energy and momentum for electrons) with the same kinds of features seen in graphene. While I haven't had a chance to read this in detail yet, it is very slick, and makes explicit the connection between real-space distortions of the graphene structure and how these are mathematically equivalent to electric and magnetic fields for the charge carriers confined to that 2d environment. It's also a great demonstration of how the motion of charge carriers in a condensed matter environment depends on the potential energy's distribution as a function of position, rather than the details. Here, the electrons are not carbon p electrons feeling the "chickenwire" potential energy of the carbon atom lattice in graphene. Rather, the electrons are those that live in the copper surface state, and they feel a designer "chickenwire" potential energy due to the arrangement of CO molecules on the copper surface. However, the net effect is the same. Very pretty. (Still makes me wonder a bit about the details, though.... At the end of the day, electrons have to scatter out of that surface state and into the bulk for the STM measurement to work, and yet that process has to be sufficiently weak that it doesn't screw up the surface state much. Very fortunate that the numbers happen to work!)

Finally, here is a cool, fun project, using nanofab tools to make art (too small to see with the unaided eye). Sameer Walavalkar did his PhD with the well known nano group of Axel Sherer at CalTech. This kind of creative outlet is another way to do outreach, and it's a heck of a lot cooler than many other approaches.

Saturday, March 10, 2012

Mini update

I am out on a brief break, but I wold be remiss if I didn't point out this exciting result. The investigators have managed to make a light emitting diode with greater than 100% electrical efficiency when operated just right. The trick is, the LED gets the energy for the "extra" photons from the temperature difference between the LED and it's surroundings. Basically it's a combination LED and heat engine. Very clever. I wonder if there are some entropy restrictions that come into play, particularly if the final photon state is, e.g., the macroscopically occupied state of a laser cavity.

Tuesday, March 06, 2012

NSF - proposal compliance

This is for everyone out there who submits to proposals to the Division of Materials Research, and more broadly, to the National Science Foundation. Here's some context for those who don't know the story. The NSF has a Grant Proposal Guide that spells out, in detail, the proper content and formatting for proposals. You can understand why they do this, particularly with regard to things like font size. There's a 15 page limit on the "Project Description" part of a proposal, and if they didn't specify a font size and margins, there would be people trying to game the system by submitting proposals in 6-pt unreadable font with 1cm margins. Historically, NSF has erred on the side of latitude about the minutiae, however. For example, they have never really been aggressive about policing whether the bibliographic references are perfectly formatted.

That's why this news came as a surprise: As part of a new policy, starting this past fall, DMR is taking basically a zero-tolerance approach regarding compliance with the Grant Proposal Guide. That means, for example, that any letter of collaboration included with a proposal can only say, in effect, "I agree to do the tasks listed in the Project Description". Anything more (e.g., context about what the collaborator's expertise is, or mentioning that this continues an existing collaboration) is no longer allowed, and would be cause for either deletion of the letter or outright rejection of the proposal without review. This new policy also means, and this is scary, that your references have to be perfectly formatted - leaving out titles, or leaving out the second page number, or using "et al." instead of long author lists - all of these can lead to a proposal being rejected without review. I heard this first hand from a program officer. Imagine spending weeks writing a proposal, and having it get bounced because you used the wrong setting in bibTeX or EndNote.

We can have a vigorous discussion in the comments about whether this policy makes much sense. In the meantime, though, I think it's very important that people be aware of this change. The bottom line: Scrupulously follow the Grant Proposal Guide. Cross every "t" and dot every "i".

Please spread this information - if one division of NSF is doing this, you can bet that it will spread, and you don't want to be the one whose proposal gets bounced.

Sunday, March 04, 2012

March Meeting last day and wrap-up

Not too much to report from the final day of the March Meeting. Lots of good conversations with colleagues, though I never did get a chance to sit down with a couple of folks I'd wanted to see. Ahh well.

I split most of my time between two invited sessions. The first of these was on the unusual properties of the nu=5/2 fractional quantum Hall state. This may sound very narrow and esoteric, but it is actually quite profound. A good review of the whole topic in more generality is here. At a very particular value of perpendicular magnetic field (related to the number of charge carriers per square centimeter), the electrons in a 2d layer in GaAs/AlGaAs semiconductor structures apparently condense into a really weird state. The lowest energy excitations of this state, its quasiparticles, have very strange properties. First, they have an effective electronic charge of 1/4 e. Second, when two of these fractionally charged quasiparticles are moved around each other to swap positions, the whole quantum mechanical state of the system changes (to another state with the same energy as the original), in a way much more complex than just picking up a phase factor (which would be -1 if the quasiparticles acted like ordinary electrons). Somehow the detailed history of winding the particles around each other is supposedly encoded in the many-body state itself. Quasiparticles with this bizarre property are said to obey "non-Abelian statistics". To date, there has not been an experimental "smoking gun" demonstrating these weird properties unambiguously. My postdoc mentor, Bob Willett, gave a very data-heavy talk showing persuasive evidence for consistency with a number of the relevant theory predictions in this system. Following him, Woowon Kang of the University of Chicago showed other data that also looks consistent with some of these ideas (though I'm no expert).

The other invited session dealt with the theory behind the transport of electrons and ions in nanoscale systems. Unfortunately I missed the beginning (since I was seeing the other talks above), but I did get to hear a neat discussion by Kirk Bevan of McGill University about the physics of electromigration. Electromigration is the mechanism by which flowing electrons can scatter off defects and grain boundaries, dumping momentum into atoms and pushing them around.

Final suggestions for the APS:
1) Don't have the small rooms arranged so that getting to seats in the front requires blocking the projector. The result of that is that the front 6 rows or so remain almost completely empty, while people pile up in the back of the rooms.

2) Would it really be that hard to have wireless internet access that doesn't suck? Are there no convention centers that can really support this?

3) Having a big bio presence at the meeting and then scheduling it directly opposite the Biophysical Society meeting seems odd.

4) Every year, there is an electronic letter-writing or petition campaign to support federal funding of research. That's fine and dandy, but is there any way we could try to get some representative Congress-critters to come hear a session, perhaps one of the fun, general invited sessions, or one about industrially relevant research? Remember, next year in Baltimore is quite close to DC....

Friday, March 02, 2012

March Meeting day 3 (for me)

Yesterday I spent a fair bit of time seeing specialized talks related to my group's research. In the contributed session in the morning, I saw a couple of talks by the theory group of Kevin Ingersent at the University of Florida. When describing electronic transport through a molecule, there are two basic theory approaches. One way to tackle this problem is to try to do realistic quantum chemistry calculations about specific molecular orbitals and how a molecule couples electronically to metal electrodes. A complementary tactic is to construct a mathematical model that you think contains the essential physics (e.g., treat the molecule as a "dot" with two electronic levels, each coupled to generic conduction electrons in the leads; then add in a single, local harmonic vibrational mode with some coupling between the level populations and the amplitude of the vibration, etc.). These two schemes correspond well with approaches to bulk materials: realistic electronic structure calculations vs. construction of model Hamiltonians. Ingersent's group takes the latter approach, and it looks like there is even more rich physics buried in single-impurity junctions than I'd previously appreciated.

In the same session, there were some nice experimental talks from Latha Venkataraman's group at Columbia. Recently, her students have seen that it's possible to create comparatively good contacts between molecules and metals, with a single quantum channel being transmitted through the system with a transmission of about 90%. This is in contrast to the more common situation, where transmission is more like 0.1%. She's also started doing single-molecule measurements of thermopower and Seebeck coefficient, where you apply a temperature gradient across a molecule and look at the resulting voltage difference that shows up. Cool data, though thermal transport at these scales is very challenging.

Later in the day I heard some nice invited talks. Jean-Marc Triscone gave a nice presentation of the properties of the two-dimensional electron gas that shows up at the interface between strontium titanate and lanthanum aluminate (STO/LAO). This field of oxide heterostructures has become very popular, and includes all sorts of rich physics, including coexistent superconductivity and magnetism. Any topic that gets to talk about a "polarization catastrophe" has to be good.

In another invited session, Cyrus Hirbijehedin talked in detail about Kondo physics in single magnetic atoms on very thin insulating layers, as probed by STM. Dan Ralph gave an extremely clear talk (via iphone from Cornell, due to inclement weather) on Kondo physics in single-molecule junctions, with their particular experimental twist of being able to stretch or squish the junctions in situ. Very neat. There are some lingering technical points in such structures that need further examination by the community.

I did not, unfortunately, see Leo Kouwenhoven's ballyhooed (here and here) talk about Majorana fermions. I need to read more about the particular work before I can offer any intelligent commentary.