Inconstant Constants: "Probing fundamental constant evolution with redshifted conjugate-satellite OH lines"

Via Jennifer Ouellette on Twitter, I ran across a Discovery News story touting a recent arxiv preprint claiming to see variation in the fine-structure constant. It's a basically OK story, but garbles a few details, so I thought it would be worth giving it the ResearchBlogging treatment, in the now-traditional Q&A format.

What did they do? The paper looks at some spectral lines in radio emission from a moderately distant galaxy with the poetic name "PKS1413+135." These lines are produced by OH molecules in interstellar gas clouds, and the frequencies they see suggest that there may have been changes in some dimensionless constants during the not quite three billion years since the light was emitted.

Dimensionless constants? What are those? Most of the things we tend to think of as fundamental constants-- particle masses, Planck's constant, and that sort of thing-- are numbers with units. The only way to really measure one of these things, though, is by measuring it relative to one or more of the others. As a result, people who think about hard-core particle physics and cosmology and that sort of thing prefer to talk about "dimensionless constants," which are ratios of these things arranged so that all the units cancel. The most famous is the "fine structure constant" α which is the ratio of the electron charge squared to Planck's Constant times the speed of light, helpfully shown in an image lifted from Wikipedia. Other dimensionless constants of interest are the ratio of the proton and electron masses, and the "gyromagnetic ratio" which relates the spin of a fundamental particle to its magnetic moment. These ratios are the things that really matter if you want to look for changes in the fundamental constants.

Changes in the fundamental constants? Aren't they, you know, constant? You'd like to think that, but they don't necessarily have to be. And a lot of the tricks particle theorists pull when they're trying to explain fundamental forces end up giving you fundamental "constants" that change in time. This is something that you can look for experimentally, and that's what the current paper is about.

How do you do that? It's not like you can get a three billion year old proton and weigh it, can you? No, but you can look at the light emitted by really old atoms and molecules. The frequencies at which atoms and molecules emit light depend on the exact values of those dimensionless constants, so if the constants change, then the frequencies change.

How can you tell, though? Doesn't the Doppler shift from the expanding universe shift all the frequencies we see, anyway? The trick is to compare the light emitted by different transitions in the same atoms or molecules. Some states will shift up in frequency with a change in the fine structure constant (for example), while other states will shift down. Doppler shifts due to the motion of the universe or objects in it will always go in the same direction, depending on the velocity of the source. If you look at the relative frequencies of light from these different states, then, you can take out the Doppler shift, and still see if there's been a change in the relative frequencies.

Isn't that really difficult? It is, and while several attempts have been made to do this sort of analysis, it's hard to do well, and the results are not without controversy. This paper reports on a cleaner way of doing this basic measurement, using two different transitions in OH molecules.

How does that work? It's a really clever trick. They use a pair of radio-frequency lines emitted by OH molecules, one at 1720 MHz, the other at 1612 MHz (for reference, the FM radio station I listen to around here has a frequency of 102.7 MHz). These lines shift in different directions in response to changes in the constants. And, better yet, they behave differently in response to light, thanks to a maser effect.

A who what? Under the right conditions, you can get an astrophysical maser, in which gas clouds out in space absorb energy from radio waves at one frequency, and are stimulated to re-emit it at a different frequency, in much the same way that a laser takes energy from a pump, and uses it to produce photons at a different output wavelength. In particular, you can get a situation where OH molecules will absorb radio waves at 1720 MHz, and emit more radio waves at 1612 MHz.

So that's what this graph is? Right. It's actually half of a figure from the paper, but the other half is more or less identical, and just confuses matters. The blue line at the top is the absorption feature they see at 1720 MHz, flipped over so it forms an upward-going peak (so the peak is actually a place where they see less light). The red line is the emission feature at 1612 MHz, where they see increased light. To make this plot, the two lines have been shifted so they line up at the same point. The black curve is the sum of the two, showing that they really have the same shape, which confirms that they're coming from the same gas cloud.

So how does this tell us about changes in the constants? Well, careful analysis of the amount they need to correct each line to get the two to line up suggests that the values of the fundamental constants were different when these molecules were emitting light three billion years ago than they are now. The change isn't very big-- the change in the relevant parameter divided by the current value of that parameter is about -1.18 ± 0.48 × 10^-5, which is right at the edge of being trustworthy, statistically speaking (a bit more than twice the uncertainty). If you attributed all the change to changes in the fine structure comment, the fractional change would be -3.1 ± 1.2 × 10^-6, which is a little smaller than the other big result claiming a change (which looked over a longer time period), and bigger than other measurements that claim to see no change.

Well, that's pretty good. Yeah, it's very good for this sort of thing. It's also a very clean measurement-- the maser effect ensures that the lines they're looking at are from the same cloud of gas, and there's nothing else along the line of sight in the relevant frequency range that would mess up their signal. They even did it with two different radio telescopes, the big one at Arecibo, and the Westerbork Synthesis Radio Telescope in the Netherlands. Granted, this is not my exact field, but it looks like a fairly solid number, though at only 2.6 standard deviations, I wouldn't bet huge money on it yet.

So, they just need to do this a bunch more times, right? That's the catch. They can certainly do more observations of this particular source, for longer periods of time, which might get them an even better result, but the maser effect they're looking at it pretty rare, and there aren't a lot of known sources.

How many is "not a lot?" According to the paper, there's one other, PMNJ0134â0931. Which means they're not really going to be able to fight off claims that there's something anomalous about the source they're looking at, unless somebody finds a whole bunch more of these gas clouds to look at. (The earlier optical measurements used something like 140 different sources, and are still highly controversial.)

So, it's a cute technique and a suggestive result, but nothing that's going to rock particle astrophysics to its core? Pretty much. It's a very clever technique, and a nice analysis, but not conclusive by a long shot. If your Theory of Everything predicts that the fine-structure constant was a few parts per million smaller three billion years ago, it's reason to keep your hopes up (if you think it used to be bigger, you're screwed), but you shouldn't be expecting a call from the Nobel committee any time soon on the basis of this observation.

Nissim Kanekar, Jayaram N. Chengalur, & Tapasi Ghosh (2010). Probing fundamental constant evolution with redshifted
conjugate-satellite OH lines Astrophysical Journal Letters arXiv: 1004.5383v1