It’s been a while since I wrote up a ResearchBlogging post, but since a recent paper forced me to update my “What Every Dog Should Know About Quantum Physics” slides with new pictures, I thought I should highlight the work on the blog as well. Not that you could’ve missed it, if you follow physics-y news– it’s been all over, getting almost as much press as rumors that some people whose funding will run out soon saw something intriguing in their data. So, in the usual Q&A format:
OK, what’s this about? Well, the paper title, “Quantum interference of large organic molecules” pretty well says it all. A group in Austria, headed by Markus Arndt, with Stefan Gerlich as the lead author, has demonstrated interference effects using large organic molecules. When they pass a beam of these molecules through an interferometer, they behave exactly like waves, producing an interference pattern analogous to the bright and dark spots you get by sending light into a diffraction grating.
That sounds cool, but didn’t they do this before? They’ve shown interference of organic molecules before– in fact, Arndt did the first demonstration of this using fullerene molecules, way back in 1999. I even wrote up their results in How to Teach Physics to Your Dog, and you can see one of their figures in the preview chapter at dogphysics.com.
So why are you fired up about this new paper? Because they’ve scaled things up to a much bigger system, and still demonstrated clear interference.
What do you mean, “scaled up”? Well, the clearest answer I can give is to copy a figure from their paper, showing the different molecules they have used:
See that little ball labeled “a”? That’s the fullerene, C60 that they used before. The big huge things on the right side of the figure, labeled “b,” “c,” “e,” and “f” are the molecules they used this time around. They contain up to 430 atoms, and are several nanometers across, making them by far the largest molecules objects anybody has ever seen displaying wave behavior.
OK, I’ll admit, that’s pretty cool. So, what, they just made a teeny-tiny little diffraction grating, and shot these molecules at it? Not exactly…
The apparatus they use goes by the slightly unwieldy name “Kapitza-Dirac Talbot-Lau Interferometer” (described in the older paper linked below), which combines three diffraction gratings and two different effects. The “Kapitza-Dirac” part comes from the fact that one of the three gratings is made from a standing wave of light, which Piotr Kapitza and Paul Dirac predicted could be used to diffract electrons way back in 1933. The strong interactions between electrons make this a little tricky in practice, but it works brilliantly for neutral systems like atoms and molecules, and is a common technique in atom interferometry.
The Tablot-Lau part of the name refers to the Talbot effect (I’ve never been clear on the correct pronunciation of this– a number of French-speaking postdocs at NIST used to say it “Tal-BOH,” but the original discoverer was British, so others say “TAL-But”), where waves passed through a diffraction grating interfere are particular distances behind the grating to produce a pattern that is an exact match for the grating itself.
Don’t they call that a “shadow?” You might think that’s all it is, but it’s actually a complex wave phenomenon. The Talbot effect occurs far enough away that even for a physical grating, you wouldn’t expect to see just a shadow. It also works for things like light-wave gratings, where there’s no solid object to cast a shadow.
OK, but then why do you need three gratings? Wouldn’t it be easier to just do one? The problem is that the waves describing the molecules coming out of their source have a wide range of different velocities, and thus wavelengths. They also have a wide range of transverse velocities, which further complicates matters. Finally, if you just use a single material grating, there are interactions between the molecules passing through the slits and the walls of the slits (through “van der Waals forces” that act at very long range) that further distort the pattern. When you start looking at very large molecules, whose quantum wavelengths are very short, these effects make detecting a clean single interference pattern almost impossible.
Yeah, but then, how does adding more gratings help matters? Basically, they allow you to pick out just the right molecules to interact with and detect.
The first grating is a material grating– a nanofabricated silicon nitride structure with slits 90 nm in width separated by 266 nm. This lets only a small fraction of the molecules through, and provides a cleaner source for the second grating to work with.
The second grating is made from a standing wave of light at 532 nm (the Kapitza-Dirac part of the interferometer). This produces a fixed array of bright and dark spots in space, separated by half a wavelength, or 266 nm. This is positioned at a distance behind the first grating where you expect the Talbot effect to create an exact image of the first grating, with bunches of molecules spaced by exactly 266nm.
These serve as, essentially, a bunch of point sources hitting the second grating. The advantage of a light grating in this position is that it doesn’t block anything, so you’re not throwing away any more molecules. All the molecules that make it through the first grating pass through the light, and pick up some transverse momentum by absorbing and emitting photons from the laser. This has exactly the same effect on their velocity distribution as passing through a material grating would, and diffracts the quantum waves out in a way that produces another interference pattern.
The third grating is another silicon nitride structure, identical to the first, and is mounted so that it can be slid from side to side. This grating is placed at the Talbot distance from the second grating, so that there should be a pattern of molecules with 90 nm wide bunches spaced by 266 nm, matching the third grating. The detector is placed behind this grating, and they read out the interference pattern by sliding it back and forth perpendicular to the molecules.
Why use a material grating there? Why not use light again? Because this lets you use an integrating detector. That is, rather than measuring the number of molecules that reach a specific position, you count all the molecules that make it through the third grating. this number is vastly greater than the number at any given point, and makes the experiment much faster.
The third grating acts sort of like a mask for the detector. When it’s positioned exactly right, the bunches of molecules produced by the second grating make it through the slits in the third, and they get a lot of counts in their detector. When they shift it over by half the slit spacing, most of the molecules are blocked, so they get very few counts. Another half-spacing shift, and they get lots of counts again, and so on.
So what they see is an up and down variation in the count rate as they slide this back and forth? Exactly. It looks like this:
The four sub-graphs in this correspond to the four large molecules above, in the same order. The points are their measured count rate, the lines are sine waves fit to the data, and the shaded grey region represents the number of “dark counts” in their detector– that is, the approximate number they would see if they ran the detector without any incoming molecules at all.
As you can see, all four of these molecules show clear sinusoidal interference patterns as the third grating is moved back and forth. This is a direct observation of the wave-like properties of these large molecules.
I don’t know, dude. I mean, it’s not that impressive. Couldn’t you get those little tiny wobbles from some classical effect? It’s possible, barely, that classical particles bouncing through this geometry might give you a similar pattern. As a check on that, they do two things: the first is to calculate the “contrast” of the pattern, which is basically the maximum number of counts at a peak in the pattern minus the minimum number at one of the dips, divided by the maximum number. Contrast is measured as a percentage, so 100% contrast would mean some large number at the peaks, and zero at the dips.
The contrast they observe for these patterns ranges from 16% to 49%. At the upper end, this is in good agreement with the maximum expected contrast using a simulation of the experiment taking into account the quantum wave nature of the molecules. The contrast they expect from a simulation using classical particles bouncing around is less than 1%, so they’re well in excess of that, even for their worst patterns.
I’m not sure I find this all that convincing, dude. What else do you have? The other thing they can do is to vary the laser power used to make the second grating. This basically varies how strongly it acts to diffract the waves hitting it, and while it’s a little complicated to do in detail, you can predict the visibility of the interference pattern you would expect for lots of different laser powers. They go through this calculation for both quantum waves and classical particles, and show that their data agree very nicely with the quantum prediction, but don’t come anywhere near the classical prediction (in a figure I’m not going to copy here, because I don’t want to duplicate all of their data on my blog).
It’s another comparison to simulation, granted, but it’s about as solid as evidence of quantum rather than classical behavior is going to get. This is a good, solid result, showing very clearly that material objects behave like waves, even relatively large material objects.
So, what have we learned that we didn’t already know? After their earlier experiments, there really wasn’t much doubt that heavier molecules would also behave like waves. It’s an extremely difficult experiment to do, though, and the fact that they’ve gotten it to work at all is pretty impressive.
Long term, their goal is to continue to work with heavier and heavier objects– one of the news stories says they’re looking at simple viruses. The ultimate goal of this is to try to push the quantum-classical boundary as far as possible.
The “quantum-classical boundary?” One of the most frustrating features of quantum mechanics is that none of the really cool stuff that it predicts happens in everyday life– macroscopic objects like tennis balls or dog treats or grains of sand appear to follow definite paths, and show no signs of wave-like behavior. Microscopic objects like electrons, atoms, and now molecules, on the other hand, clearly show wave behavior.
Somewhere between the size of one of these molecules and the size of a grain of sand, however, there’s a boundary of some sort where quantum effects stop being observable, and everything begins to look classical. There are a lot of theories about why this is, with some holding that there’s a real physical effect that makes it impossible to ever see quantum effects with large objects, while others feel it’s just a technical issue, that a sufficiently sensitive and clever technique might show large things behaving like waves just fine.
The goal of this work, and other projects along similar lines, is to look for quantum effects in larger systems, to settle the question one way or another.
So, how far have they come? Well, it’s a long way up from electrons, with their heaviest molecule having a mass of just under 7000 times the mass of a proton. They’ve got a long way to go to reach the scale of the Planck mass, though, which is the neighborhood where most theories attempting to explain the boundary seem to operate. That’s around 1019 times the proton mass.
This is an impressive technical achievement, though, and yet more proof (as if it was needed) that the universe we live in is a strange and wonderful place.
Gerlich, S., Eibenberger, S., Tomandl, M., Nimmrichter, S., Hornberger, K., Fagan, P., TÃ¼xen, J., Mayor, M., & Arndt, M. (2011). Quantum interference of large organic molecules Nature Communications, 2 DOI: 10.1038/ncomms1263
Hornberger, K., Gerlich, S., Ulbricht, H., HackermÃ¼ller, L., Nimmrichter, S., V Goldt, I., Boltalina, O., & Arndt, M. (2009). Theory and experimental verification of Kapitza-Dirac-Talbot-Lau interferometry New Journal of Physics, 11 (4) DOI: 10.1088/1367-2630/11/4/043032