It Really Is Just Dust

The Planck collaboration has an inimitable way of releasing important new results, they like to do it in French (see here for instance). Tonight a French Planck website contains the long-awaited news of the results from the BICEP2/Keck/Planck collaboration to reanalyze the BICEP2 data on polarized B-modes, in a way that allows proper estimation of the contribution of dust. The bottom line is that the BICEP2 claims of seeing a primordial r=.16-.20 that got a huge amount of media attention last year have been shot down. The new analysis says that r is less than .13. I don’t see a paper yet, rumor is that the paper will be on the arXiv Monday night.

My understanding is that full polarization results from Planck’s own data are still not quite ready, as has been the case for quite a while now.

Hat tip for this news to Steve (retired LLNL physicist). Thanks Steve!

Update: The Planck collaboration now has information available about this (in English!). An NSF version is here, NBC News here. It still looks like we’ll have to wait until next week for the paper, which has been submitted to PRL.

Update: The paper is now available here.

This entry was posted in Uncategorized. Bookmark the permalink.

31 Responses to It Really Is Just Dust

  1. martibal says:

    We are not so far from the 200th anniversary of the Waterloo’s battle. This is a mark of resistance against the linguistic domination of the Perfide Albion !

  2. JG says:

    Google Chrome browser it offers automatic translation

    It turns out that the part of the dust had been significantly underestimated.

    Once the part of the galactic emission subtracted correctly, there is always an excess but it is too small for this to be considered a detection and could be the result of simple changes associated with experimental sounds.

    Not a perfect translation, but you get the gist

  3. Sakura-chan says:

    Off-topic, but pretty big piece on Zhang in the New Yorker.

  4. BMc says:

    ” D’après les modèles à leur disposition, la collaboration BICEP estimait que cette part était négligeable, tout au plus très faible. Avec une seule fréquence d’observation, ils n’avaient pas d’autre possibilité pour évaluer le signal d’origine galactique…”

    A tiny bit of sarcasme there?

    I’m trying hard to laugh about this whole situation. But I can’t.

  5. Peter Woit says:

    My own translation of the last part:

    The spatial distribution of the Planck 353 GHz signal shows up in the 150 Ghz map of BICEP2 and of KECK: these last two thus contain a non-negligeable contribution from galactic dust. Once the galactic dust emission is correctly subtracted there is still an excess, but at the present time too small to count as a detection and could be the result of simple fluctuations due to experimental noise.

    The parameter r measures the size of the signal produced by primordial gravitational waves. Today’s limit is a bit less strong because there is a small excess: too small to be a measurement of anything, but sufficient to weaken the limit [from r=.11 without the BICEP2/KECK data to the new r=.13]

  6. West says:

    Looking at the likelihood(?) plots at the bottom of the page, the K+P and B+P curves have some pretty dramatic separation. As the Keck array was supposed to be noticeably more sensitive than BICEP2, I would go put money on a small value of r as suggested by the red curve.

  7. Peter Woit says:

    Just noticed that I hadn’t clicked on the “Expert” level to see the plots that West is referring to. One needs to be sure to do that.

  8. JG says:

    The “expert” page makes it clear that we haven’t really got that much more certainty on this than what was known since last year.

    What is certain today is that the primordial origin signal is not at the level previously estimated – at most about half. Given the instrumental uncertainties (remember that the signal is incredibly weak!), We can not say more today. It will take further measures of extreme sensitivity at other frequencies to better understand the question …

    So 700 million euros wasn’t enough for the Planck mission to decide this one – maybe that’s why they prefer to release updates on obscure french wep pages rather than youtube and a harvard-like press conferences

    Results from POLARBEAR, Spider, BICEP3 etc will decide this I guess

  9. Physicsphile says:


    Planck may not have been able to significantly constrain primordial gravitational waves but they have had some other very useful results. For example they have shown that the non-gaussianity parameter is very small which rules out many proposed inflation models.

  10. Pingback: Allgemeines Live-Blog ab dem 26. Januar 2014 | Skyweek Zwei Punkt Null

  11. AW says:

    Dear Peter,

    I think the heading of your blog post is a bit strong:

    “It really is just dust”

    You yourself just quoted from the Planck website in your own translation:

    “Once the galactic dust emission is correctly subtracted there is still an excess, but at the present time too small to count as a detection and could be the result of simple fluctuations due to experimental noise.

    The parameter r measures the size of the signal produced by primordial gravitational waves. Today’s limit is a bit less strong because there is a small excess: too small to be a measurement of anything, but sufficient to weaken the limit [from r=.11 without the BICEP2/KECK data to the new r=.13]”

    This statement is inconsistent with saying “it really is just dust”. For that I would require the analysis to show the absence of any residual excess at a level of r < 0.01 with a least 3 sigma significance.

    This not the case. On the contrary, the statement on the Planck website might very well be construed to imply a residual excess (with definitely r > 0.01 given their current sensitivity) at 1-2 sigma level (this would fit with their notion, that the excess, while there, might still very well be a fluke). So I think your title is misstating their results and hence too strong.

  12. I went to a colloquium on Wednesday given by Paolo de Bernardis at LMU-Munich, and Slava Mukhanov worked hard to get him to say that BICEP2 was wrong. He finally did. But he also said that Planck would publish something in February (he didn’t say when in February), and he wouldn’t say what those conclusions would be. I guess that might be the arXiv paper “due out” on Monday.

  13. chris says:

    The link points to an empty page. Have they taken the results down again?

  14. Planck people I spoke to this morning were surprised to hear that the results had been announced already. I suspect it was unauthorised jumping of the gun, which is why they have been taken down. The paper has already been submitted to a journal a couple of weeks ago, and will be on the arXiv early next week, followed a few days after by the Planck 2014/2015 release papers (or so I’m told).

  15. stringph says:

    So if ‘It’ = ‘what caused the BICEP collaboration to claim a detection’, then the new results show ‘it’ was exactly dust. If the dust had not been there (or had been accurately subtracted) the measurements would be a lot smaller and they could not possibly have claimed detection.

  16. martibal says:

    According to a cosmologist friend, it seems that someone has given to people outside the collaboration the “secret” URL of the non-yet public and still in preparation Planck page on the Planck/BICEP analysis, in violation of all ethical rules. Unless the page has been cracked. In any case it seems that the page was not “searchable” and the URL complicated enough not to be randomly found. This friend ask people not to sprea
    d the google cache page.

  17. Peter Woit says:

    I don’t think it’s misleading. It’s a null result, definitively showing that the original BICEP2 claims were wrong, and that’s the obvious meaning of the title.

    Your .01 criterion is randomly chosen. Of course all one can do here is put bounds (unless you have a real observation). You can’t claim evidence for something based on 1-2 sigma, especially something extraordinary (“extraordinary claims require extraordinary evidence”). Claims from BICEP2 that “maybe we really did see something and weren’t wrong” are what would be misleading.

  18. Peter Woit says:

    One of many odd things here is that if this was a page not intended for the public you would normally make that clear by putting something at the top indicating that. It looked exactly like an intended official announcement, and that site is exactly the one used last year by Planck as the only place for official news from the collaboration.

    If it had been clear that was not meant to be public, I would have not linked to it (although I might have mentioned “rumors about r <.13"…..)

  19. AS says:

    I have a paper comparing my predictions for r and ns to the global fit of Planck, Bicep and Keck. Should I cite a web page disappeared even from webcache or wait?

  20. Peter Woit says:

    You could cite various tweets on Twitter, a BBC news story, or maybe this blog…

  21. Sasha says:


    What other projects associated with the primordial gravitational waves will produce results this year?

  22. Peter Woit says:

    I know little about the current state of other CMB experiments, perhaps someone better informed can comment about what to expect.

  23. JG says:

    Paper released

    (and to clarfiy my above comment, I meant that the ESA public announcement implies that the entire signal can be attributed to dust – which is what I interpreted Peter’s headline to mean – since obviously he isn’t making a prediction about the entire future of observations in this field)

  24. West says:

    @Sasha: SPIDER just finished its flight down on Antarctica. The collaboration has a nice set of blog entries about the balloon flight.

  25. Mark says:

    Has there been any commentary from the BICEP2 team about their flawed initial analysis? In other words, is anyone eating crow and acknowledging the mistake in their – shall we say “over exuberant” – analysis?

  26. West says:

    @Mark: At least from the quotation from Kovac given to Nature News, not really.

    I find the quotation “These [original dust maps] seemed to indicate that the region of the sky chosen for our observations had dust polarisation much lower than the detected signal” rather problematic. After the March announcement there were a number of papers showing that a careful analysis using the preliminary dust maps (Planck’s?) meant there was a sizable microwave foreground from dust. Did anyone go back and dig through the other noise models and quantify their actual uncertainty?

  27. B says:

    The New York Times now has an article “A Speck of Interstellar Dust Rebuts a Big Bang Theory” quoting Dr. Clem Pryke of BICEP “We can’t say with any certainty whether any gravity wave signals remain… Obviously, we’re not exactly thrilled, but we are scientists and our job is to try and uncover the truth. In the scientific process, the truth will emerge.”
    [ ]

  28. Sasha says:

    I know about SPIDER. But I wonder when the results will be available? Anyone heard the news or rumors about Bicep3? Will there be more results from Keck and Polarbear?And whether or not to hope for a positive result?

  29. Jim Akerlund says:

    Here is the arXiv version 1502.00612.

Comments are closed.