Archive for BICEP


Posted in Astrohype, The Universe and Stuff with tags , , , , on March 17, 2014 by telescoper

Well, it’s official that this afternoon’s announcement of a “major discovery” is going to be from the BICEP team, and it specifically concerns the BICEP2 CMB telescope experiment. I’ve just got back to Sussex (after a weekend in Cardiff) and will be following the events in among other things I have to do before going off to give a lecture at 5pm GMT.

The schedule of events is as follows: there will be a special webcast presenting the first results from the BICEP2 CMB telescope. The webcast will begin with a presentation for scientists 10:45-11:30 EDT, followed by a news conference 12:00-1:00 EDT.

You can join the webcast from the link at

Papers and data products will be available at 10:45 EDT from

EDT is four hours behind Greenwich Mean Time so the webcast will begin at 14:45 GMT, i.e. in about half an hour.

In the mean time, for those of you wondering what these BICEPS are all about, here is a useful graphic in which a Harvard astrophysicist demonstrates the possibilities:



14:36 The press conference server has gone down. There’s no truth in the rumour that ex-members of the Clover collaboration have sabotaged it.

14:42 There’s a grave danger that this press conference will run into tea time.

14:45 The BICEP2 papers are now live at

14:48 Straight to the headline: R=0.2 (+0.07, -0.05) with R=0 rejected at about 7 sigma, if you like things stated in such terms…

14:53 Here’s the crucial graph. Results a bit higher than the expected  signal at l in range 200-300?


15:06 The news avalanche has started, e.g. here at the BBC, but there is some concern about the shape of the spectrum.

15:10 I’m not getting anything from the press conference, so may have missed important details. It seems to me though that there’s a significant possibility of some of the polarization signal in E and B not being cosmological. This is a very interesting result, but I’d prefer to reserve judgement until it is confirmed by other experiments.

15:35 Despite the press hype there’s still some scepticism among cosmologists arising from the strange-looking shape of the spectrum. I’m not convinced myself. Anyway, I have to sign off now in order to prepare a lecture..

16:20 Back-of-the-envelope time: if the result is correct then the inflationary energy scale is about 2×1016 GeV. That’s just two orders of magnitude below the Planck scale…

18:19 Returned from my 5pm Theoretical Physics lecture. Couldn’t resist spending 30 minutes talking about BICEP2, though I did tell them it’s not in the examination.

18:25 Main points of controversy:

  1. there seems to be evidence of leakage of temperature into polarization (lines in Fig. 5);
  2. there’s an excess in the B-B spectrum at l~250 shown above;
  3. there’s an excess at low l in the E-E spectrum
  4. there’s a deficit at low l in the cross-correlation with Keck

There may be a connection between 1. and 2.-4. If 2.-4 are real then they may be evidence of something interesting that requires more than a straightforward modification of inflation (such as might include just a running of the spectral index).

18:35 Other controversy: why has this result been announced before the paper has been published or even peer-reviewed?

Some B-Mode Background

Posted in Astrohype, Science Politics, The Universe and Stuff with tags , , , , , , , , , , , on March 15, 2014 by telescoper

Well, in case you hadn’t noticed, the cosmology rumour mill has gone into overdrive this weekend primarily concerning the possibility that an experiment known as BICEP (an acronym formed from Background Imaging of Cosmic Extragalactic Polarization). These rumours have been circulating since it was announced last week that the Harvard-Smithsonian Center for Astrophysics (CfA) will host a press conference  on Monday, March 17th, to announce “a major discovery”. The grapevine is full of possibilities, but it seems fairly clear that the “major discovery” is related to one of the most exciting challenges facing the current generation of cosmologists, namely to locate in the pattern of fluctuations in the cosmic microwave background evidence for the primordial gravitational waves predicted by models of the Universe that involve inflation.

Anyway, I thought I’d add a bit of background on here to help those interested make sense of whatever is announced on Monday evening.

Looking only at the temperature variation across the sky, it is not possible to distinguish between tensor  (gravitational wave) and scalar (density wave) contributions  (both of which are predicted to be excited during the inflationary epoch).  However, scattering of photons off electrons is expected to leave the radiation slightly polarized (at the level of a few percent). This gives us additional information in the form of the  polarization angle at each point on the sky and this extra clue should, in principle, enable us to disentangle the tensor and scalar components.

The polarization signal can be decomposed into two basic types depending on whether the pattern has  odd or even parity, as shown in the nice diagram (from a paper by James Bartlett)

The top row shows the E-mode (which look the same when reflected in a mirror and can be produced by either scalar or tensor modes) and the bottom shows the B-mode (which have a definite handedness that changes when mirror-reflected and which can’t be generated by scalar modes because they can’t have odd parity).

The B-mode is therefore (at least in principle)  a clean diagnostic of the presence of gravitational waves in the early Universe. Unfortunately, however, the B-mode is predicted to be very small, about 100 times smaller than the E-mode, and foreground contamination is likely to be a very serious issue for any experiment trying to detect it. To be convinced that what is being measured is cosmological rather than some sort of contaminant one would have to see the signal repeated across a range of different wavelengths.

Moreover, primordial gravitational waves are not the only way that a cosmological B-mode signal could be generated. Less than a year ago, a paper appeared on the arXiv by Hanson et al. from SPTpol, an experiment which aims to measure the polarization of the cosmic microwave background using the South Pole Telescope. The principal result of this paper was to demonstrate a convincing detection of the so-called “B-mode” of polarization from gravitational lensing of the microwave background photons as they pass through the gravitational field generated by the matter distributed through the Universe. Gravitational lensing can produce the same kind of shearing effect that gravitational waves generate, so it’s important to separate this “line-of-sight” effect from truly primordial signals.

So we wait with bated breath to see exactly what is announced on Monday. In particular, it will be extremely interesting to see whether the new results from BICEP are consistent with the recently published conclusions from Planck. Although Planck has not yet released the analysis of its own polarization data, analysis of the temperature fluctuations yields a (somewhat model-dependent) conclusion that the ratio of tensor to scalar contributions to the CMB pattern is no more than about 11 per cent, usually phrased in the terms, i.e. R<0.11. A quick (and possibly inaccurate) back-of-the-envelope calculation using the published expected sensitivity of BICEP suggests that if they have made a detection it might be above that limit. That would be really interesting because it might indicate that something is going on which is not consistent with the standard framework. The limits on R arising from temperature studies alone assume that both scalar and tensor perturbations are generated by a relatively simple inflationary model belonging to a class in which there is a direct relationship between the relative amplitudes of the two modes (and the shape of the perturbation spectrum). So far everything we have learned from CMB analysis is broadly consistent with this simplifying assumption being correct. Are we about to see evidence that the early Universe was more complex than we thought? We'll just have to wait and see…

Incidentally, once upon a time there was a British experiment called Clover (involving the Universities of  Cardiff, Oxford, Cambridge and Manchester) which was designed to detect the primordial B-mode signal from its vantage point in Chile. I won’t describe it in more detail here, for reasons which will become obvious.

The chance to get involved in a high-profile cosmological experiment was one of the reasons I moved to Cardiff in 2007, and I was looking forward to seeing the data arriving for analysis. Although I’m primarily a theorist, I have some experience in advanced statistical methods that might have been useful in analysing the output.  Unfortunately, however, none of that actually happened. Because of its budget crisis, and despite the fact that it had spent a large amount (£4.5M) on it already,  STFC decided to withdraw the funding needed to complete it (£2.5M)  and cancelled the Clover experiment. Had it gone ahead it would probably have had two years’ data in the bag by now.

It wasn’t clear that Clover would have won the race to detect the B-mode cosmological polarization, but it’s a real shame it was withdrawn as a non-starter. C’est la vie.