No evidence for spacetime foam

It goes under the name spacetime-foam or fuzz, or sometimes graininess or, more neutral, fluctuations: The general expectation that, due to quantum gravitational effects, spacetime on very short distance scales is not nice and smooth but, well, fuzzy or grainy or foamy.

If that seems somewhat fuzzy to you, it's because it is. Absent an experimentally verified and generally accepted theory of quantum gravity, nobody really knows what exactly spacetime foam looks like. A case for the phenomenlogists then! And, indeed, over the decades several models have been suggested to describe spacetime's foamy, fuzzy grains, based on a few simple assumptions. The idea is not that these models are fundamentally the correct description of spacetime but that they can, ideally, be tested against data, so that we can learn something about the general properties that we should expect the fundamental theory to deliver.

One example for such a model, going back to Amelino-Camelia in 1999, is that spacetime foam causes a particle to make a random walk in the direction of propagation. For each step of distance of the Planck length, l_P, the particle randomly gains or loses another step. This is a useful model because random walks in one dimension are very well understood. Over a total distance L that consists of N =L/l_P steps, the average deviation from the mean is the length of the step times the square root of the number of steps. Thus, over a distance L, a particle deviates a distance

where I have put in a dimensionless constant α - for a quantum gravitational effect, we would expect it to be of order one. See also the below figure for illustration and keep in mind that c=1 on this blog

While simple, this model is not, and probably was never meant to be, particularly compelling. Leaving aside that it's not Lorentz-invariant, there is no good reason why the steps should be discrete or be aligned in one direction. One might have hoped that this general idea would be worked out to become somewhat more plausible. Yet that never happened because this model was ruled out pretty much as soon as it was proposed. The reason is that if you consider the detection of lightrays from some source, the deviation from the normal propagation on the light-cone will have the effect of allowing different phases of the emitted light to arrive at once. The average phase blur is

If the phase blur is comparable to 2π, interferences would be washed out. See below figure for illustration

As it happens however, interference patterns, known as Airy rings, can be observed on objects as far as some Gpc away from earth. If you put in the numbers, for such a large distance and α ≈ 1 the phase should be entirely smeared out. To the right you see an actual image of such a distant quasar from the Hubble Space Telescope (Figure 3 from astro-ph/0610422). You can clearly see the interference rings. And there goes the random walk model.

There is a similar model going back to Wheeler and later Ng and Van Dam, that the authors have called the "holographic foam" model. (Probably because everything holographic was chic in the late 1990s. Except for the general scaling it has little to do with holography.) In any case, the main feature of this model is that the deviation from the mean goes with the 3rd root, rather than the square root, of N. Thus, the effects are smaller.

It is amazing though how quickly smart people can find ways to punch holes in your models. Already in 2003 it was pointed out, that with some classical optic formulas from the late 19th century, modern telescopes allow to set much tighter bounds. Roughly speaking, the reason is that a telescope with diameter D focuses a much larger part of the light's wavefront than just one wavelength λ. The telescope is very sensitive to phase-smearing all over its opening. Telescopes are for example sensitive to air turbulences, a problem that the Hubble Space Telescope does not have.

The sensitivity of a telescope to such phase distortions can be quantified by a pure number known as the "Strehl ratio." The closer the Strehl ratio is to 1, the closer the telescope's images are to those of an ideal telescope, showing a point-like sources as a perfect Airy patterns. A non-ideal telescope will cause an image degradation, most importantly a smearing of the intensity. The same effect would be caused by the holographic space-time fuzz. Thus, up to the telescope's limit on image quality, the additional phase distortion would be observable: it lowers the Strehl ratio of images of very far-away objects such as quasars. (Though, if it was observed, one wouldn't know exactly what its origin is.)

The relevant point is that, using the telescope's sensitivity to image degradation, one gains an additional factor of D/λ ≈ 10⁸. In their paper:

No quantum gravity signature from the farthest quasars
By Fabrizio Tamburini, Carmine Cuofano, Massimo Della Valle, Roberto Gilmozzi

the authors have presented an analysis of the images of 157 high-redshift (z > 4) quasi-stellar objects. They found no blurring. With that, also the holographic foam model is ruled out. Or, to be precise, the parameter α is constrained into a range that is implausible for quantum gravitational effects.

As it is often the case in the phenomenology of quantum gravity, the plausible models are difficult, if not impossible, to constrain by data. And the implausible ones nobody misses when they are ruled out. This is a case of the latter.

Thanks to Neil for reminding me of that paper.

---

PS: We were not able to find a derivation for the exact expression for the phase blurring as a function of the Strehl ratio, Eq. (5), that is used in the paper. We got so far that it's called the Marechal approximation. If you know of a useful reference, we'd be interested!