Clover: What Might Have Been

Quite a few people have been asking me whether the UK’s cancelled B-mode experiment, Clover, could have detected what BICEP2 may  have found; I’m still not convinced, by the way. It therefore seemed apt to do a quick post in order to direct you to relevant sources of information. If you’re interested in Clover’s capabilities you can find a nice summary on the ArXiv here. The abstract reads:

We describe the objectives, design and predicted performance of Clover, which is a ground-based experiment to measure the faint “B-mode” polarisation pattern in the cosmic microwave background (CMB). To achieve this goal, clover will make polarimetric observations of approximately 1000 deg^2 of the sky in spectral bands centred on 97, 150 and 225 GHz. The observations will be made with a two-mirror compact range antenna fed by profiled corrugated horns. The telescope beam sizes for each band are 7.5, 5.5 and 5.5 arcmin, respectively. The polarisation of the sky will be measured with a rotating half-wave plate and stationary analyser, which will be an orthomode transducer. The sky coverage combined with the angular resolution will allow us to measure the angular power spectra between 20 < l < 1000. Each frequency band will employ 192 single polarisation, photon noise limited TES bolometers cooled to 100 mK. The background-limited sensitivity of these detector arrays will allow us to constrain the tensor-to-scalar ratio to 0.026 at 3sigma, assuming any polarised foreground signals can be subtracted with minimal degradation to the 150 GHz sensitivity. Systematic errors will be mitigated by modulating the polarisation of the sky signals with the rotating half-wave plate, fast azimuth scans and periodic telescope rotations about its boresight. The three spectral bands will be divided into two separate but nearly identical instruments – one for 97 GHz and another for 150 and 225 GHz. The two instruments will be sited on identical three-axis mounts in the Atacama Desert in Chile near Pampa la Bola. Observations are expected to begin in late 2009.

The following points, gleaned from a very quick skimming of the above paper, are worth noting (but please note the important corrections and clarifications in the comments below from the first author of the above paper and also bear in mind that the Clover numbers are estimated rather than based on actual measurements):

  1. The sky coverage of Clover would have been 1000 square degrees, compared with 380 square degrees of BICEP2
  2. Clover measurements would have been made at three frequencies, 97 GHz, 150 GHz and 225 GHz. This would have enabled the possibility of foreground contamination to be rejected with much greater confidence than in BICEP2 (which only operates at 150 GHz)
  3. The sensitivity of Clover at 150 GHz (the frequency at which BICEP2 operates) would have been about 1.4 times better than BICEP2
  4. Had it gone ahead, Clover would have started taking data at around the same time as BICEP2 (perhaps even a bit earlier).
  5. Clover was originally intended to be positioned at the South Pole, where observing conditions are better than in Chile and where BICEP2 is placed, but this was in the middle of STFC’s financial crisis and running costs would have been much higher than the alternative location in Chile. This might have had a negative impact on its sensitivity.

Here’s a plot from the above paper showing a the anticipated measurement if the tensor scalar ratio had been 0.1; BICEP2 detection (if real) corresponds to a signal twice this amplitude:


In other words, we don’t know whether Clover would have hit its target sensitivity and there are many other imponderables, but  it’s a very great shame it never got the chance to try…


6 Responses to “Clover: What Might Have Been”

  1. Some factual corrections and clarifications.

    ***Disclosure: I was a member of the Clover team and did my PhD (sorry, DPhil) on the project while at Oxford University. It’s cancellation meant a significant shift in my career, though I was lucky enough to move on to Planck and then Herschel and stay at Cardiff (where I was working when Clover was cancelled on 2009). Other colleagues were not so lucky and either moved country or institution, or left astronomy all together.

    First of all, congratulations to the BICEP2 team on what, at first glance (to me at least), seems to be a convincing result, though the complete cross-examination and peer-review process will establish that on due course. Whatever the outcome, this is a remarkable result from what seems to be a very well designed and run experiment.

    – BICEP2 did observe 1000deg^2; the 380deg^2 is the effective area after apodisation to limit E->B leakage from partial sky coverage. The (nominal) Clover observation strategy was to observe for fields of about 250deg^2 each. The reasoning was to avoid Galactic foregrounds and optimise observing efficiency (since no one field is above the horizon at all times), though It looks now as if the polarised foregrounds are not as big an issue as predicted at high Galactic latitudes.
    – my (pretty simple/basic/naive) calculations (which is where I assume you got the factor 1.4 from) show that the Clover polarisation sensitivity would have been 1.4x BICEP2 if *both* 97 and 150 GHz were combined, giving 192 detectors (half the number BICEP2 had at 150GHz), and using detectors that are ~2x more sensitive. The polarisation sensitivity of 150 GHz alone is roughly comparable to the BICEP2 sensitivity.
    – the Clover numbers are specs/design values, whereas the BICEP2 numbers are what they achieved in practise.
    – The predicted integration time for Clover from a 2 year campaign is about the same as what BICEP2 achieved (also from a 2 year campaign), though I think the Atacama may have worse weather than the South Pole. The advantage of Atacama (latitude -23°) over the South Pole is that field rotation would allow cross linking (scanning across the field(s) at multiple angles)
    – Clover’s original site was actually ‘Dome C’, to be provided and operated (pretty much for free) by French and Italian collaborators. When that fell through (largely, if I recall, because the site was not going to be ready in time), the Atacama desert was selected as an alternative. Since that site was not free, It put significant pressure on an already overstretched budget.
    – it must also be remembered that the budgets were significantly different. If the numbers I’ve heard are correct, then the BICEP2 budget was about 4-5x the Clover budget.
    – another important difference is that the BICEP2 team had experience with a very similar experiment (BICEP1), and so had ironed out many of the problems.

    I’m clearly biased by my involvement in Clover, but I wholeheartedly agree that (despite its problems, financial and otherwise) its cancellation was a mistake for many and varied reasons. it is by no means certain, however, that Clover would have obtained this result ahead of BICEP2, it is clear that there would have been a very good chance for a UK team to have been in the game, providing one of the first, if not the first, detection of B-modes.

    But, what is done is done. In summary, many congratulations to the BICEP2 team. I wish we could have been there too.

    P.S. Can I count this as a citation of my paper? 😉

    • telescoper Says:

      Thanks for the clarifications!

    • My understanding is that no-one has yet deployed a telescope to Dome C, beyond just doing seeing measurements.

      • I think BRAIN may have been intended for there, though may be mistaken. Though I’m not sure what’s happened to BRAIN, or if they solved the difficulties of doing bolometric interferometry.

    • “Cenorth”: Really nice and i thought gracious way to fill in details about the relative capabilities for Clover and Bicep2. I remember being disappointed by Clover’s cancellation, and much else other damage done when STFC did its budget-slashing. Of course, scientists are never promised indefinite amounts of money and carte blanche to pursue everything, and decisions and tradeoffs have to be made relative to other areas of science *and* the rest of a country’s budget, but in hindsight this one seems to have been an especially bad decision.. :-/


  2. We’ll never know! There are a number of experiments which may detect the B-modes. I am not yet convinced by the BICEP detection: it seems to me that the paper in its current form would (should) not make it past a referee. Confirmation by a second experiment would help. Clover might have done that. With the time lines and the underfunding, I doubt that Clover would have been there first. STFC is blamed for the withdrawal of funding, but did the team bid for sufficient money in the first place? Or did they bid for what they thought they might get?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: