CERN Accelerating science

001993893 001__ 1993893
001993893 003__ SzGeCERN
001993893 005__ 20211014230328.0
001993893 0247_ $$2DOI$$a10.1051/0004-6361/201525936
001993893 0248_ $$aoai:cds.cern.ch:1993893$$pcerncds:FULLTEXT$$pcerncds:CERN:FULLTEXT$$pcerncds:CERN
001993893 035__ $$9arXiv$$aoai:arXiv.org:1502.05956
001993893 035__ $$9Inspire$$a1345438
001993893 037__ $$9arXiv$$aarXiv:1502.05956$$castro-ph.CO
001993893 041__ $$aeng
001993893 100__ $$aAdam, R.$$uLPSC, Grenoble$$vLaboratoire de Physique Subatomique et Cosmologie, Université Grenoble-Alpes, CNRS/IN2P3, 53 rue des Martyrs, 38026 Grenoble Cedex, France
001993893 245__ $$aPlanck 2015 results. IX. Diffuse component separation: CMB maps
001993893 260__ $$c2016-09-20
001993893 269__ $$c20 Feb 2015
001993893 300__ $$a42 p
001993893 500__ $$aComments: 42 pages, 36 figures. Submitted to A&A
001993893 500__ $$9arXiv$$a42 pages, 36 figures. Submitted to A&A
001993893 520__ $$aWe present foreground-reduced CMB maps derived from the full Planck data set in both temperature and polarization. Compared to the corresponding Planck 2013 temperature sky maps, the total data volume is larger by a factor of 3.2 for frequencies between 30 and 70 GHz, and by 1.9 for frequencies between 100 and 857 GHz. In addition, systematic errors in the forms of temperature-to-polarization leakage, analogue-to-digital conversion uncertainties, and very long time constant errors have been dramatically reduced, to the extent that the cosmological polarization signal may now be robustly recovered on angular scales $\ell\gtrsim40$. On the very largest scales, instrumental systematic residuals are still non-negligible compared to the expected cosmological signal, and modes with $\ell < 20$ are accordingly suppressed in the current polarization maps by high-pass filtering. As in 2013, four different CMB component separation algorithms are applied to these observations, providing a measure of stability with respect to algorithmic and modelling choices. The resulting polarization maps have rms instrumental noise ranging between 0.21 and 0.27$\,\mu\textrm{K}$ averaged over 55 arcmin pixels, and between 4.5 and 6.1$\,\mu\textrm{K}$ averaged over 3.4 arcmin pixels. The cosmological parameters derived from the analysis of temperature power spectra are in agreement at the $1\sigma$ level with the Planck 2015 likelihood. Unresolved mismatches between the noise properties of the data and simulations prevent a satisfactory description of the higher-order statistical properties of the polarization maps. Thus, the primary applications of these polarization maps are those that do not require massive simulations for accurate estimation of uncertainties, for instance estimation of cross-spectra and cross-correlations, or stacking analyses.
001993893 520__ $$9EDP Sciences$$aWe present foreground-reduced cosmic microwave background (CMB) maps derived from the full Planck data set in both temperature and polarization. Compared to the corresponding Planck 2013 temperature sky maps, the total data volume is larger by a factor of 3.2 for frequencies between 30 and 70GHz, and by 1.9 for frequencies between 100 and 857GHz. In addition, systematic errors in the forms of temperature-to-polarization leakage, analogue-to-digital conversion uncertainties, and very long time constant errors have been dramatically reduced, to the extent that the cosmological polarization signal may now be robustly recovered on angular scales l ≳ 40. On the very largest scales, instrumental systematic residuals are still non-negligible compared to the expected cosmological signal, and modes with l are accordingly suppressed in the current polarization maps by high-pass filtering. As in 2013, four different CMB component separation algorithms are applied to these observations, providing a measure of stability with respect to algorithmic and modelling choices. The resulting polarization maps have rms instrumental noise ranging between 0.21 and 0.27muK averaged over 55' pixels, and between 4.5 and 6.1muK averaged over [??] pixels. The cosmological parameters derived from the analysis of temperature power spectra are in agreement at the 1sigma level with the Planck 2015 likelihood. Unresolved mismatches between the noise properties of the data and simulations prevent a satisfactory description of the higher-order statistical properties of the polarization maps. Thus, the primary applications of these polarization maps are those that do not require massive simulations for accurate estimation of uncertainties, for instance estimation of cross-spectra and cross-correlations, or stacking analyses. However, the amplitude of primordial non-Gaussianity is consistent with zero within 2sigma for all local, equilateral, and orthogonal configurations of the bispectrum, including for polarization E-modes. Moreover, excellent agreement is found regarding the lensing B-mode power spectrum, both internally among the various component separation codes and with the best-fit Planck 2015 Lambda cold dark matter model.
001993893 520__ $$9arXiv$$aWe present foreground-reduced CMB maps derived from the full Planck data set in both temperature and polarization. Compared to the corresponding Planck 2013 temperature sky maps, the total data volume is larger by a factor of 3.2 for frequencies between 30 and 70 GHz, and by 1.9 for frequencies between 100 and 857 GHz. In addition, systematic errors in the forms of temperature-to-polarization leakage, analogue-to-digital conversion uncertainties, and very long time constant errors have been dramatically reduced, to the extent that the cosmological polarization signal may now be robustly recovered on angular scales $\ell\gtrsim40$. On the very largest scales, instrumental systematic residuals are still non-negligible compared to the expected cosmological signal, and modes with $\ell < 20$ are accordingly suppressed in the current polarization maps by high-pass filtering. As in 2013, four different CMB component separation algorithms are applied to these observations, providing a measure of stability with respect to algorithmic and modelling choices. The resulting polarization maps have rms instrumental noise ranging between 0.21 and 0.27$\,\mu\textrm{K}$ averaged over 55 arcmin pixels, and between 4.5 and 6.1$\,\mu\textrm{K}$ averaged over 3.4 arcmin pixels. The cosmological parameters derived from the analysis of temperature power spectra are in agreement at the $1\sigma$ level with the Planck 2015 likelihood. Unresolved mismatches between the noise properties of the data and simulations prevent a satisfactory description of the higher-order statistical properties of the polarization maps. Thus, the primary applications of these polarization maps are those that do not require massive simulations for accurate estimation of uncertainties, for instance estimation of cross-spectra and cross-correlations, or stacking analyses.
001993893 540__ $$aarXiv nonexclusive-distrib. 1.0$$barXiv$$uhttp://arxiv.org/licenses/nonexclusive-distrib/1.0/
001993893 595__ $$aLANL EDS
001993893 595__ $$aNot CERN
001993893 595__ $$aCDS
001993893 595__ $$aNot for annual report
001993893 65017 $$2arXiv$$aAstrophysics and Astronomy
001993893 695__ $$9LANL EDS$$aastro-ph.CO
001993893 690C_ $$aARTICLE
001993893 690C_ $$aCERN
001993893 700__ $$aAde, P.A.R.$$uCardiff U.
001993893 700__ $$aAghanim, N.$$uOrsay
001993893 700__ $$aArnaud, M.$$uDAPNIA, Saclay
001993893 700__ $$aAshdown, M.$$uCambridge U., KICC$$uCambridge U.
001993893 700__ $$aAumont, J.$$uOrsay
001993893 700__ $$aBaccigalupi, C.$$uSISSA, Trieste
001993893 700__ $$aBanday, A.J.$$uToulouse III U.$$uIRAP, Toulouse
001993893 700__ $$aBarreiro, R.B.$$uCantabria Inst. of Phys.
001993893 700__ $$aBartlett, J.G.$$uAPC, Paris$$uCaltech
001993893 700__ $$aBartolo, N.$$uPadua U.$$uINFN, Padua
001993893 700__ $$aBasak, S.$$uSISSA, Trieste
001993893 700__ $$aBattaner, E.$$uGranada U.
001993893 700__ $$aBenabed, K.$$uParis, Inst. Astrophys.$$uParis U., VI-VII
001993893 700__ $$aBenoit, A.$$uNeel Lab, Grenoble
001993893 700__ $$aBenoit-Levy, A.$$uUniversity Coll. London$$uParis, Inst. Astrophys.$$uParis U., VI-VII
001993893 700__ $$aBernard, J.P.$$uToulouse III U.$$uIRAP, Toulouse
001993893 700__ $$aBersanelli, M.$$uMilan U.$$uBrera Observ.
001993893 700__ $$aBielewicz, P.$$uToulouse III U.$$uIRAP, Toulouse$$uSISSA, Trieste
001993893 700__ $$aBonaldi, A.$$uManchester U.
001993893 700__ $$aBonavera, L.$$uCantabria Inst. of Phys.
001993893 700__ $$aBond, J.R.$$uCanadian Inst. Theor. Astrophys.
001993893 700__ $$aBorrill, J.$$uLBL, Berkeley$$uUC, Berkeley
001993893 700__ $$aBouchet, F.R.$$uParis, Inst. Astrophys.
001993893 700__ $$aBoulanger, F.$$uOrsay
001993893 700__ $$aBucher, M.$$uAPC, Paris
001993893 700__ $$aBurigana, C.$$uBologna Observ.$$uFerrara U.$$uINFN, Bologna
001993893 700__ $$aButler, R.C.$$uBologna Observ.
001993893 700__ $$aCalabrese, E.$$uOxford U.
001993893 700__ $$aCardoso, J.F.$$uLTCI, Paris$$uAPC, Paris$$uParis, Inst. Astrophys.
001993893 700__ $$aCasaponsa, B.$$uCantabria Inst. of Phys.
001993893 700__ $$aCastex, G.$$uAPC, Paris
001993893 700__ $$aCatalano, A.$$uLPSC, Grenoble$$uLERMA, Ivry
001993893 700__ $$aChallinor, A.$$uCambridge U., Inst. of Astron.$$uCambridge U., KICC$$uCambridge U., DAMTP
001993893 700__ $$aChamballu, A.$$uDAPNIA, Saclay$$uOrsay
001993893 700__ $$aChary, R.R.$$uCaltech
001993893 700__ $$aChiang, H.C.$$uPrinceton U.$$uKwaZulu Natal U.
001993893 700__ $$aChristensen, P.R.$$uBohr Inst.
001993893 700__ $$aClements, D.L.$$uImperial Coll., London
001993893 700__ $$aColombi, S.$$uParis, Inst. Astrophys.$$uParis U., VI-VII
001993893 700__ $$aColombo, L.P.L.$$uUCLA$$uCaltech
001993893 700__ $$aCombet, C.$$uLPSC, Grenoble
001993893 700__ $$aCouchot, F.$$uOrsay
001993893 700__ $$aCoulais, A.$$uLERMA, Ivry
001993893 700__ $$aCrill, B.P.$$uCaltech
001993893 700__ $$aCurto, A.$$uCambridge U.$$uCantabria Inst. of Phys.
001993893 700__ $$aCuttaia, F.$$uBologna Observ.
001993893 700__ $$aDanese, L.$$uSISSA, Trieste
001993893 700__ $$aDavies, R.D.$$uManchester U.
001993893 700__ $$aDavis, R.J.$$uManchester U.
001993893 700__ $$ade Bernardis, P.$$uRome U.
001993893 700__ $$ade Rosa, A.$$uBologna Observ.
001993893 700__ $$ade Zotti, G.$$uPadua Observ.$$uSISSA, Trieste
001993893 700__ $$aDelabrouille, J.$$uAPC, Paris
001993893 700__ $$aDesert, F.X.$$uGrenoble Observ.
001993893 700__ $$aDickinson, C.$$uManchester U.
001993893 700__ $$aDiego, J.M.$$uCantabria Inst. of Phys.
001993893 700__ $$aDole, H.$$uOrsay$$uIUF, Paris
001993893 700__ $$aDonzelli, S.$$uBrera Observ.
001993893 700__ $$aDore, O.$$uCaltech
001993893 700__ $$aDouspis, M.$$uOrsay
001993893 700__ $$aDucout, A.$$uParis, Inst. Astrophys.$$uImperial Coll., London
001993893 700__ $$aDupac, X.$$uESA, Madrid
001993893 700__ $$aEfstathiou, G.$$uCambridge U., Inst. of Astron.
001993893 700__ $$aElsner, F.$$uUniversity Coll. London$$uParis, Inst. Astrophys.$$uParis U., VI-VII
001993893 700__ $$aEnsslin, T.A.$$uGarching, Max Planck Inst.
001993893 700__ $$aEriksen, H.K.$$uOslo U.
001993893 700__ $$aFalgarone, E.$$uLERMA, Ivry
001993893 700__ $$aFantaye, Y.$$uOslo U.
001993893 700__ $$aFergusson, J.$$uCambridge U., DAMTP
001993893 700__ $$aFinelli, F.$$uBologna Observ.$$uINFN, Bologna
001993893 700__ $$aForni, O.$$uToulouse III U.$$uIRAP, Toulouse
001993893 700__ $$aFrailis, M.$$uTrieste Observ.
001993893 700__ $$aFraisse, A.A.$$uPrinceton U.
001993893 700__ $$aFranceschi, E.$$uBologna Observ.
001993893 700__ $$aFrejsel, A.$$uBohr Inst.
001993893 700__ $$aGaleotta, S.$$uTrieste Observ.
001993893 700__ $$aGalli, S.$$uParis, Inst. Astrophys.
001993893 700__ $$aGanga, K.$$uAPC, Paris
001993893 700__ $$aGhosh, T.$$uOrsay
001993893 700__ $$aGiard, M.$$uToulouse III U.$$uIRAP, Toulouse
001993893 700__ $$aGiraud-Heraud, Y.$$uAPC, Paris
001993893 700__ $$aGjerlow, E.$$uOslo U.
001993893 700__ $$aGonzalez-Nuevo, J.$$uCantabria Inst. of Phys.$$uSISSA, Trieste
001993893 700__ $$aGorski, K.M.$$uCaltech$$uWarsaw U.
001993893 700__ $$aGratton, S.$$uCambridge U., KICC$$uCambridge U., Inst. of Astron.
001993893 700__ $$aGregorio, A.$$uTrieste U.$$uTrieste Observ.$$uINFN, Trieste
001993893 700__ $$aGruppuso, A.$$uBologna Observ.
001993893 700__ $$aGudmundsson, J.E.$$uPrinceton U.
001993893 700__ $$aHansen, F.K.$$uOslo U.
001993893 700__ $$aHanson, D.$$uRutherford$$uCaltech$$uCanadian Inst. Theor. Astrophys.
001993893 700__ $$aHarrison, D.L.$$uCambridge U., Inst. of Astron.$$uCambridge U., KICC
001993893 700__ $$aHelou, G.$$uCaltech
001993893 700__ $$aHenrot-Versille, S.$$uOrsay
001993893 700__ $$aHernandez-Monteagudo, C.$$uCEFCA, Teruel$$uGarching, Max Planck Inst.
001993893 700__ $$aHerranz, D.$$uCantabria Inst. of Phys.
001993893 700__ $$aHildebrandt, S.R.$$uCaltech
001993893 700__ $$aHivon, E.$$uParis, Inst. Astrophys.$$uParis U., VI-VII
001993893 700__ $$aHobson, M.$$uCambridge U.
001993893 700__ $$aHolmes, W.A.$$uCaltech
001993893 700__ $$aHornstrup, A.$$uDenmark, Tech. U.
001993893 700__ $$aHovest, W.$$uGarching, Max Planck Inst.
001993893 700__ $$aHuffenberger, K.M.$$uFlorida State U.
001993893 700__ $$aHurier, G.$$uOrsay
001993893 700__ $$aJaffe, A.H.$$uImperial Coll., London
001993893 700__ $$aJaffe, T.R.$$uToulouse III U.$$uIRAP, Toulouse
001993893 700__ $$aJones, W.C.$$uPrinceton U.
001993893 700__ $$aJuvela, M.$$uHelsinki U.
001993893 700__ $$aKeihanen, E.$$uHelsinki U.
001993893 700__ $$aKeskitalo, R.$$uLBL, Berkeley
001993893 700__ $$aKisner, T.S.$$uLBL, Berkeley
001993893 700__ $$aKneissl, R.$$uEuropean Southern Obs., Chile
001993893 700__ $$aKnoche, J.$$uGarching, Max Planck Inst.
001993893 700__ $$aKrachmalnicoff, N.$$uMilan U.
001993893 700__ $$aKunz, M.$$uGeneva U.$$uOrsay$$uAfrican Inst. Math. Sci., Cape Town
001993893 700__ $$aKurki-Suonio, H.$$uHelsinki U.$$uHelsinki Inst. of Phys.
001993893 700__ $$aLagache, G.$$uMarseille, Lab. Astrophys.$$uOrsay
001993893 700__ $$aLamarre, J.M.$$uLERMA, Ivry
001993893 700__ $$aLasenby, A.$$uCambridge U.$$uCambridge U., KICC
001993893 700__ $$aLattanzi, M.$$uFerrara U.
001993893 700__ $$aLawrence, C.R.$$uCaltech
001993893 700__ $$aLe Jeune, M.$$uAPC, Paris
001993893 700__ $$aLeonardi, R.$$uESA, Madrid
001993893 700__ $$aLesgourgues, J.$$uCERN$$uLPHE, Lausanne$$uSavoie U.
001993893 700__ $$aLevrier, F.$$uLERMA, Ivry
001993893 700__ $$aLiguori, M.$$uPadua U.$$uINFN, Padua
001993893 700__ $$aLilje, P.B.$$uOslo U.
001993893 700__ $$aLinden-Vornle, M.$$uDenmark, Tech. U.
001993893 700__ $$aLopez-Caniego, M.$$uESA, Madrid$$uCantabria Inst. of Phys.
001993893 700__ $$aLubin, P.M.$$uUC, Santa Barbara
001993893 700__ $$aMacias-Perez, J.F.$$uLPSC, Grenoble
001993893 700__ $$aMaggio, G.$$uTrieste Observ.
001993893 700__ $$aMaino, D.$$uMilan U.$$uBrera Observ.
001993893 700__ $$aMandolesi, N.$$uBologna Observ.$$uFerrara U.
001993893 700__ $$aMangilli, A.$$uOrsay
001993893 700__ $$aMarshall, D.J.$$uDAPNIA, Saclay
001993893 700__ $$aMartin, P.G.$$uCanadian Inst. Theor. Astrophys.
001993893 700__ $$aMartinez-Gonzalez, E.$$uCantabria Inst. of Phys.
001993893 700__ $$aMasi, S.$$uRome U.
001993893 700__ $$aMatarrese, S.$$uPadua U.$$uINFN, Padua$$uGran Sasso
001993893 700__ $$aMazzotta, P.$$uRome U.
001993893 700__ $$aMcGehee, P.$$uCaltech
001993893 700__ $$aMeinhold, P.R.$$uUC, Santa Barbara
001993893 700__ $$aMelchiorri, A.$$uRome U.$$uINFN, Rome
001993893 700__ $$aMendes, L.$$uESA, Madrid
001993893 700__ $$aMennella, A.$$uMilan U.$$uBrera Observ.
001993893 700__ $$aMigliaccio, M.$$uCambridge U., Inst. of Astron.$$uCambridge U., KICC
001993893 700__ $$aMitra, S.$$uPune U.$$uCaltech
001993893 700__ $$aMiville-Deschenes, M.A.$$uOrsay$$uCanadian Inst. Theor. Astrophys.
001993893 700__ $$aMolinari, D.$$uCantabria Inst. of Phys.$$uBologna Observ.
001993893 700__ $$aMoneti, A.$$uParis, Inst. Astrophys.
001993893 700__ $$aMontier, L.$$uToulouse III U.$$uIRAP, Toulouse
001993893 700__ $$aMorgante, G.$$uBologna Observ.
001993893 700__ $$aMortlock, D.$$uImperial Coll., London
001993893 700__ $$aMoss, A.$$uNottingham U.
001993893 700__ $$aMunshi, D.$$uCardiff U.
001993893 700__ $$aMurphy, J.A.$$uNUIM, Maynooth
001993893 700__ $$aNaselsky, P.$$uBohr Inst.
001993893 700__ $$aNati, F.$$uPrinceton U.
001993893 700__ $$aNatoli, P.$$uFerrara U.$$uINFN, Rome2$$uBologna Observ.
001993893 700__ $$aNetterfield, C.B.$$uToronto U.
001993893 700__ $$aNorgaard-Nielsen, H.U.$$uDenmark, Tech. U.
001993893 700__ $$aNoviello, F.$$uManchester U.
001993893 700__ $$aNovikov, D.$$uLebedev Inst.
001993893 700__ $$aNovikov, I.$$uBohr Inst.$$uLebedev Inst.
001993893 700__ $$aOxborrow, C.A.$$uDenmark, Tech. U.
001993893 700__ $$aPaci, F.$$uSISSA, Trieste
001993893 700__ $$aPagano, L.$$uRome U.$$uINFN, Rome
001993893 700__ $$aPajot, F.$$uOrsay
001993893 700__ $$aPaladini, R.$$uCaltech
001993893 700__ $$aPaoletti, D.$$uBologna Observ.$$uINFN, Bologna
001993893 700__ $$aPasian, F.$$uTrieste Observ.
001993893 700__ $$aPatanchon, G.$$uAPC, Paris
001993893 700__ $$aPearson, T.J.$$uCaltech
001993893 700__ $$aPerdereau, O.$$uOrsay
001993893 700__ $$aPerotto, L.$$uLPSC, Grenoble
001993893 700__ $$aPerrotta, F.$$uSISSA, Trieste
001993893 700__ $$aPettorino, V.$$uHeidelberg U.
001993893 700__ $$aPiacentini, F.$$uRome U.
001993893 700__ $$aPiat, M.$$uAPC, Paris
001993893 700__ $$aPierpaoli, E.$$uUCLA
001993893 700__ $$aPietrobon, D.$$uCaltech
001993893 700__ $$aPlaszczynski, S.$$uOrsay
001993893 700__ $$aPointecouteau, E.$$uToulouse III U.$$uIRAP, Toulouse
001993893 700__ $$aPolenta, G.$$uINFN, Rome2$$uRome Observ.
001993893 700__ $$aPratt, G.W.$$uDAPNIA, Saclay
001993893 700__ $$aPrezeau, G.$$uCaltech
001993893 700__ $$aPrunet, S.$$uParis, Inst. Astrophys.$$uParis U., VI-VII
001993893 700__ $$aPuget, J.L.$$uOrsay
001993893 700__ $$aRachen, J.P.$$uNijmegen U.$$uGarching, Max Planck Inst.
001993893 700__ $$aRacine, B.$$uAPC, Paris
001993893 700__ $$aReach, W.T.$$uUniversities Space Research Assoc.
001993893 700__ $$aRebolo, R.$$uIAC, La Laguna$$uCSIC, Madrid$$uLaguna U., Tenerife
001993893 700__ $$aReinecke, M.$$uGarching, Max Planck Inst.
001993893 700__ $$aRemazeilles, M.$$uManchester U.$$uOrsay$$uAPC, Paris
001993893 700__ $$aRenault, C.$$uLPSC, Grenoble
001993893 700__ $$aRenzi, A.$$uRome U.$$uINFN, Rome
001993893 700__ $$aRistorcelli, I.$$uToulouse III U.$$uIRAP, Toulouse
001993893 700__ $$aRocha, G.$$uCaltech
001993893 700__ $$aRosset, C.$$uAPC, Paris
001993893 700__ $$aRossetti, M.$$uMilan U.$$uBrera Observ.
001993893 700__ $$aRoudier, G.$$uAPC, Paris$$uLERMA, Ivry$$uCaltech
001993893 700__ $$aRubino-Martin, J.A.$$uIAC, La Laguna$$uLaguna U., Tenerife
001993893 700__ $$aRusholme, B.$$uCaltech
001993893 700__ $$aSandri, M.$$uBologna Observ.
001993893 700__ $$aSantos, D.$$uLPSC, Grenoble
001993893 700__ $$aSavelainen, M.$$uHelsinki U.$$uHelsinki Inst. of Phys.
001993893 700__ $$aSavini, G.$$uUniversity Coll. London
001993893 700__ $$aScott, D.$$uBritish Columbia U.
001993893 700__ $$aSeiffert, M.D.$$uCaltech
001993893 700__ $$aShellard, E.P.S.$$uCambridge U., DAMTP
001993893 700__ $$aSpencer, L.D.$$uCardiff U.
001993893 700__ $$aStolyarov, V.$$uCambridge U.$$uCambridge U., KICC$$uStavropol, Astrophys. Observ.
001993893 700__ $$aStompor, R.$$uAPC, Paris
001993893 700__ $$aSudiwala, R.$$uCardiff U.
001993893 700__ $$aSunyaev, R.$$uGarching, Max Planck Inst.$$uMoscow, Space Res. Inst.
001993893 700__ $$aSutton, D.$$uCambridge U., Inst. of Astron.$$uCambridge U., KICC
001993893 700__ $$aSuur-Uski, A.S.$$uHelsinki U.$$uHelsinki Inst. of Phys.
001993893 700__ $$aSygnet, J.F.$$uParis, Inst. Astrophys.
001993893 700__ $$aTauber, J.A.$$uESTEC, Noordwijk
001993893 700__ $$aTerenzi, L.$$uFacolta di Ingegneria, Rome$$uBologna Observ.
001993893 700__ $$aToffolatti, L.$$uOviedo U.$$uCantabria Inst. of Phys.$$uBologna Observ.
001993893 700__ $$aTomasi, M.$$uMilan U.$$uBrera Observ.
001993893 700__ $$aTristram, M.$$uOrsay
001993893 700__ $$aTrombetti, T.$$uBologna Observ.
001993893 700__ $$aTucci, M.$$uGeneva U.
001993893 700__ $$aTuovinen, J.$$uTrinity Coll., Dublin
001993893 700__ $$aValenziano, L.$$uBologna Observ.
001993893 700__ $$aValiviita, J.$$uHelsinki U.$$uHelsinki Inst. of Phys.
001993893 700__ $$aVan Tent, B.$$uOrsay
001993893 700__ $$aVielva, P.$$uCantabria Inst. of Phys.
001993893 700__ $$aVilla, F.$$uBologna Observ.
001993893 700__ $$aWade, L.A.$$uCaltech
001993893 700__ $$aWandelt, B.D.$$uParis, Inst. Astrophys.$$uParis U., VI-VII$$uIllinois U., Urbana
001993893 700__ $$aWehus, I.K.$$uCaltech
001993893 700__ $$aYvon, D.$$uDAPNIA, Saclay
001993893 700__ $$aZacchei, A.$$uTrieste Observ.
001993893 700__ $$aZonca, A.$$uUC, Santa Barbara
001993893 710__ $$gPlanck Collaboration
001993893 773__ $$cA9$$pAstron. Astrophys.$$v594$$y2016
001993893 8564_ $$uhttp://arxiv.org/pdf/1502.05956.pdf$$yPreprint
001993893 8564_ $$81284097$$s28832425$$uhttp://cds.cern.ch/record/1993893/files/arXiv:1502.05956.pdf
001993893 8564_ $$81284076$$s2096$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_spec_raw_TT.png$$y00129 Power spectra of the foreground cleaned CMB maps from FFP8 simulations. \emph{Top:} $TT$ power spectra evaluated using the \texttt{FFP8-UT74} mask. \emph{Bottom:} $EE$ power spectra evaluated using the \texttt{FFP8-UP76} mask. Thick lines show the spectra of signal plus noise estimated from the half-mission half-sum maps; thin lines show the noise levels from half-mission half-difference maps. The black line shows the input spectrum.
001993893 8564_ $$81284077$$s5343$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_int_components_paper_sevem_preliminary.png$$y00133 $TT$ angular power spectra of residuals from the indicated FFP8 components in the \Planck\ 2015 CMB maps, compared with the predicted signal from the best fit cosmology. ``Other'' is the sum of CO, free-free, thermal and kinetic SZ, spinning dust, and synchrotron emission. The horizontal axis is linear in $\ell^{\,0.5}$.
001993893 8564_ $$81284078$$s11868$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_Q_nilc_sevem_cmb_hp_20_40_080a_0128.png$$y00105 Pairwise difference maps between $Q$ maps obtained on FFP8 siulations. Smoothing and degrading as in Fig.~\ref{fig:dx11_diff_I}.
001993893 8564_ $$81284079$$s10797$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_int_commander_smica_cmb_080a_0128.png$$y00098 Pairwise difference maps between CMB temperature maps obtained on FFP8 simulations. Prior to differencing, the maps have been smoothed to 80 arcminutes FWHM and downgraded to N$_{side}$ = 128.
001993893 8564_ $$81284080$$s922$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_smica_pol_mask_005a_2048.png$$y00095 \smica\ masks in temperature (\emph{top}) and polarization (\emph{bottom}).
001993893 8564_ $$81284081$$s78761$$uhttp://cds.cern.ch/record/1993893/files/figs_threept_eq_qqu_dx11v2_hp_diff_n64.png$$y00074 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ \Planck\ 2015 CMB estimates and the corresponding means estimated from 1000 Monte Carlo simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equialteral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$ and $U_rU_rU_r$. The red solid, orange dot dot dot-dashed, green dashed and blue dot-dashed lines correspond to the \commander, \nilc, \sevem, and \smica\ maps, respectively. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284082$$s11274$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_int_sevem_smica_cmb_080a_0128.png$$y00101 Pairwise difference maps between CMB temperature maps obtained on FFP8 simulations. Prior to differencing, the maps have been smoothed to 80 arcminutes FWHM and downgraded to N$_{side}$ = 128.
001993893 8564_ $$81284083$$s12262$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_diff_Q_commander_nilc_case1_cmb_hp_20_40_080a_0128.png$$y00037 Pairwise differences between CMB $Q$ maps, after smoothing to FWHM 80\arcm\ and downgrading to $\nside = 128$.
001993893 8564_ $$81284084$$s25967$$uhttp://cds.cern.ch/record/1993893/files/figs_threept_coll_ttt_ffp8_hp_diff_n64.png$$y00115 The difference between the $N$-point functions and the corresponding means estimated from 1000 MC simulations. From left to right, results for the 2-point, pseudo-collapsed 3-point, equilateral 3-point and connected rhombic 4-point functions for the $N_{\rm side}=64$ FFP8 CMB temperature estimates. The black solid, red dot dot dot-dashed, orange dashed, green dot-dashed, and blue long dashed lines correspond to the true, {\tt Commander}, {\tt NILC}, {\tt SEVEM}, and {\tt SMICA} maps, respectively. The true CMB map was analysed with added noise corresponding to the {\tt SMICA} component separation method. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284085$$s10595$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_U_nilc_smica_cmb_hp_20_40_080a_0128.png$$y00112 Pairwise difference maps between $U$ maps obtained on FFP8 siulations. Smoothing and degrading as in Figs.~\ref{fig:dx11_diff_I},\ref{fig:dx11_diff_Q}.
001993893 8564_ $$81284086$$s15249$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_spec_TT_cropped.png$$y00060 Caption not extracted
001993893 8564_ $$81284087$$s5748$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_par_TT_EE_cropped.png$$y00062 Comparison of cosmological parameters estimated from the $TT$ and $EE$ spectra computed from the foreground-cleaned CMB maps. Within each group, the three left-most points show results for $TT$ with $\ell_{\mathrm{max}} = 1000$, $1500$, and $2000$ the two right-most points show results for $EE$ with $\ell_{\mathrm{max}} = 1000$ and $1500$. For comparison, we also show the corresponding parameters obtained with the \Planck\ 2015 likelihood including multipoles up to $\ell_{\mathrm{max}} = 2500$ as the horizontal line surrounded by a grey band giving the uncertainties. The foreground model used for the cleaned CMB maps is the method-tailored full-sky model from the FFP8 simulations.
001993893 8564_ $$81284088$$s29284$$uhttp://cds.cern.ch/record/1993893/files/figs_needlet_bands-crop.png$$y00084 Needlet bands used in the analysis. The solid black line shows the normalization of the needlet bands, that is, the total filter applied to the original map after needlet decomposition and synthesis of the output map from needlet coefficients.
001993893 8564_ $$81284089$$s29847$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_sevem_U_case1_cmb_hp_20_40_010a_1024.png$$y00029 Caption not extracted
001993893 8564_ $$81284090$$s11996$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_diff_int_nilc_smica_cmb_080a_0128.png$$y00016 Caption not extracted
001993893 8564_ $$81284091$$s6078$$uhttp://cds.cern.ch/record/1993893/files/figs_variance_P_maps_new_common2_cropped.png$$y00064 Polarized intensity variance evaluated from the FFP8 Monte Carlo simulations (histogram) and from the \Planck\ 2015 maps (vertical red lines) outside the \texttt{UPB77} mask. Columns from left to right show different resolutions ($\nside = 1024$, $256$, and $64$), while rows show results for the four component separation methods. Unlike in Fig.~\ref{Fig:onepointsinglecomp}, the variance distributions are not normalized.
001993893 8564_ $$81284092$$s4429$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_pol_components_paper_nilc_preliminary.png$$y00136 $EE$ angular power spectra of residuals from the indicated FFP8 components in the \Planck\ 2015 CMB maps, compared with the predicted signal from the best fit cosmology. ``Other'' is the sum of CO, free-free, thermal and kinetic SZ, spinning dust, far-infrared background, and radio and infrared unresolved sources. The horizontal axis is linear in $\ell^{\,0.5}$.
001993893 8564_ $$81284093$$s115512$$uhttp://cds.cern.ch/record/1993893/files/figs_comm_zoom_NEP_full_Q-crop.png$$y00031 $20\deg\times20\deg$ patch of the high-pass filtered \commander\ CMB polarization map, centered on the North Ecliptic Pole, $(l,b)=(96\deg,30\deg)$. Each map is pixelized with a \healpix\ resolution of $\nside=1024$, and has an angular resolution of $10\arcm$ FWHM. The top row shows $Q$ and $U$ maps derived from the full-mission data set, the middle row shows the corresponding $E$ and $B$ maps, and the bottom row shows the $E$ and $B$ maps of the half-ring half-difference (HRHD) map. Note the characteristic $+$ and $\times$ patterns in the $Q$ and $U$ maps, and the clear asymmetry between $E$ and $B$ in the full data set. Also note that the HRHD $E$ map is consistent with both the full and HRHD $B$ maps.
001993893 8564_ $$81284094$$s31241$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_diff_Q_nilc_sevem_case1_cmb_hp_20_40_080a_0128.png$$y00040 Pairwise differences between CMB $Q$ maps, after smoothing to FWHM 80\arcm\ and downgrading to $\nside = 128$.
001993893 8564_ $$81284095$$s5312$$uhttp://cds.cern.ch/record/1993893/files/figs_smica_filter_dx11.png$$y00092 \smica\ weights for temperature (\emph{top}) and polarization (\emph{bottom}). For readibility, the values are shown for input maps in units of antenna temperature. The plot goes up to $\ell\sim3600$, but the output maps are synthesized uses all multipoles up to $\ell=4000$. For polarization, the thick solid lines show the contribution of input $E$ modes to the CMB $E$ modes and the thick dashed lines show the same for the $B$ modes. The thin lines, all close to zero, show ``cross-contributions'' of input $E$ modes to the CMB $B$ modes and vice versa.
001993893 8564_ $$81284096$$s61752$$uhttp://cds.cern.ch/record/1993893/files/figs_threept_coll_uuu_dx11v2_hp_diff_n64.png$$y00071 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ \Planck\ 2015 CMB estimates and the corresponding means estimated from 1000 Monte Carlo simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equialteral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$ and $U_rU_rU_r$. The red solid, orange dot dot dot-dashed, green dashed and blue dot-dashed lines correspond to the \commander, \nilc, \sevem, and \smica\ maps, respectively. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284098$$s1105$$uhttp://cds.cern.ch/record/1993893/files/figs_colourbar_15uK.png$$y00011 Caption not extracted
001993893 8564_ $$81284099$$s12095$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_diff_int_commander_nilc_cmb_080a_0128.png$$y00012 Pairwise difference maps between CMB temperature maps. As in the previous Fig.~\ref{fig:2015_2013_diff_I}, the maps have been smoothed to FWHM 80\arcm\ and downgraded to $\nside =128$.
001993893 8564_ $$81284100$$s4184$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_pol_components_paper_commander_preliminary.png$$y00135 $EE$ angular power spectra of residuals from the indicated FFP8 components in the \Planck\ 2015 CMB maps, compared with the predicted signal from the best fit cosmology. ``Other'' is the sum of CO, free-free, thermal and kinetic SZ, spinning dust, far-infrared background, and radio and infrared unresolved sources. The horizontal axis is linear in $\ell^{\,0.5}$.
001993893 8564_ $$81284101$$s4428$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_pol_components_paper_smica_preliminary.png$$y00138 $EE$ angular power spectra of residuals from the indicated FFP8 components in the \Planck\ 2015 CMB maps, compared with the predicted signal from the best fit cosmology. ``Other'' is the sum of CO, free-free, thermal and kinetic SZ, spinning dust, far-infrared background, and radio and infrared unresolved sources. The horizontal axis is linear in $\ell^{\,0.5}$.
001993893 8564_ $$81284102$$s27699$$uhttp://cds.cern.ch/record/1993893/files/figs_twopt_tt_ffp8_hp_diff_n64.png$$y00114 The difference between the $N$-point functions and the corresponding means estimated from 1000 MC simulations. From left to right, results for the 2-point, pseudo-collapsed 3-point, equilateral 3-point and connected rhombic 4-point functions for the $N_{\rm side}=64$ FFP8 CMB temperature estimates. The black solid, red dot dot dot-dashed, orange dashed, green dot-dashed, and blue long dashed lines correspond to the true, {\tt Commander}, {\tt NILC}, {\tt SEVEM}, and {\tt SMICA} maps, respectively. The true CMB map was analysed with added noise corresponding to the {\tt SMICA} component separation method. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284103$$s1471$$uhttp://cds.cern.ch/record/1993893/files/figs_commander_hybrid_overview.png$$y00079 Multipole moment weights used for multi-resolution hybridization in the \commander\ CMB map, as described by Eq.~\ref{eq:comm_hybrid}.
001993893 8564_ $$81284104$$s12378$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_diff_U_commander_sevem_case1_cmb_hp_20_40_080a_0128.png$$y00045 Pairwise differences between CMB $U$ maps, after smoothing and downgrading as in Fig.~\ref{fig:dx11_diff_Q}.
001993893 8564_ $$81284105$$s1280$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_sevem_int_mask_005a_2048.png$$y00090 \sevem\ masks in temperature (\emph{top}) and polarization (\emph{bottom}).
001993893 8564_ $$81284106$$s10993$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_spec_EE_cropped.png$$y00061 Caption not extracted
001993893 8564_ $$81284107$$s11661$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_diff_Q_commander_smica_case1_cmb_hp_20_40_080a_0128.png$$y00039 Pairwise differences between CMB $Q$ maps, after smoothing to FWHM 80\arcm\ and downgrading to $\nside = 128$.
001993893 8564_ $$81284108$$s10206$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_U_commander_smica_cmb_hp_20_40_080a_0128.png$$y00110 Pairwise difference maps between $U$ maps obtained on FFP8 siulations. Smoothing and degrading as in Figs.~\ref{fig:dx11_diff_I},\ref{fig:dx11_diff_Q}.
001993893 8564_ $$81284109$$s47071$$uhttp://cds.cern.ch/record/1993893/files/figs_threept_eq_ttq_ffp8_hp_diff_n64.png$$y00125 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ FFP8 CMB estimates and the corresponding means estimated from 1000 MC simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equilateral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$. The black solid, red dot dot dot-dashed, orange dashed, green dot-dashed, and blue long dashed lines correspond to the true, {\tt Commander}, {\tt NILC}, {\tt SEVEM}, and {\tt SMICA} maps, respectively. The true CMB map was analysed with added noise corresponding to the {\tt SMICA} component separation method. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284110$$s11086$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_int_commander_input_cmb_080a_0128.png$$y00019 Caption not extracted
001993893 8564_ $$81284111$$s38027$$uhttp://cds.cern.ch/record/1993893/files/figs_threept_coll_tqq_dx11v2_hp_diff_n64.png$$y00069 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ \Planck\ 2015 CMB estimates and the corresponding means estimated from 1000 Monte Carlo simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equialteral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$ and $U_rU_rU_r$. The red solid, orange dot dot dot-dashed, green dashed and blue dot-dashed lines correspond to the \commander, \nilc, \sevem, and \smica\ maps, respectively. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284112$$s29911$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_commander_U_case1_cmb_hp_20_40_010a_1024.png$$y00027 Caption not extracted
001993893 8564_ $$81284113$$s11415$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_diff_int_commander_sevem_cmb_080a_0128.png$$y00013 Difference between output and input CMB temperature maps from FFP8 simulations. Smoothing and downgrading as in Fig.~\ref{fig:dx11_diff_I}.
001993893 8564_ $$81284114$$s10988$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_int_commander_sevem_cmb_080a_0128.png$$y00097 Pairwise difference maps between CMB temperature maps obtained on FFP8 simulations. Prior to differencing, the maps have been smoothed to 80 arcminutes FWHM and downgraded to N$_{side}$ = 128.
001993893 8564_ $$81284115$$s117217$$uhttp://cds.cern.ch/record/1993893/files/figs_comm_zoom_NEP_full_E-crop.png$$y00033 $20\deg\times20\deg$ patch of the high-pass filtered \commander\ CMB polarization map, centered on the North Ecliptic Pole, $(l,b)=(96\deg,30\deg)$. Each map is pixelized with a \healpix\ resolution of $\nside=1024$, and has an angular resolution of $10\arcm$ FWHM. The top row shows $Q$ and $U$ maps derived from the full-mission data set, the middle row shows the corresponding $E$ and $B$ maps, and the bottom row shows the $E$ and $B$ maps of the half-ring half-difference (HRHD) map. Note the characteristic $+$ and $\times$ patterns in the $Q$ and $U$ maps, and the clear asymmetry between $E$ and $B$ in the full data set. Also note that the HRHD $E$ map is consistent with both the full and HRHD $B$ maps.
001993893 8564_ $$81284116$$s11971$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_diff_int_sevem_smica_cmb_080a_0128.png$$y00017 Caption not extracted
001993893 8564_ $$81284117$$s892$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_nilc_pol_mask_005a_2048.png$$y00089 \nilc\ masks for temperature (\emph{top}) and polarization (\emph{bottom}).
001993893 8564_ $$81284118$$s63159$$uhttp://cds.cern.ch/record/1993893/files/figs_threept_eq_uuu_ffp8_hp_diff_n64.png$$y00128 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ FFP8 CMB estimates and the corresponding means estimated from 1000 MC simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equilateral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$. The black solid, red dot dot dot-dashed, orange dashed, green dot-dashed, and blue long dashed lines correspond to the true, {\tt Commander}, {\tt NILC}, {\tt SEVEM}, and {\tt SMICA} maps, respectively. The true CMB map was analysed with added noise corresponding to the {\tt SMICA} component separation method. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284119$$s11253$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_U_commander_nilc_cmb_hp_20_40_080a_0128.png$$y00108 Pairwise difference maps between $U$ maps obtained on FFP8 siulations. Smoothing and degrading as in Figs.~\ref{fig:dx11_diff_I},\ref{fig:dx11_diff_Q}.
001993893 8564_ $$81284120$$s12547$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_smica_U_case1_cmb_hp_20_40_010a_1024.png$$y00030 Caption not extracted
001993893 8564_ $$81284121$$s101606$$uhttp://cds.cern.ch/record/1993893/files/figs_comm_zoom_NEP_hrhd_E-crop.png$$y00035 $20\deg\times20\deg$ patch of the high-pass filtered \commander\ CMB polarization map, centered on the North Ecliptic Pole, $(l,b)=(96\deg,30\deg)$. Each map is pixelized with a \healpix\ resolution of $\nside=1024$, and has an angular resolution of $10\arcm$ FWHM. The top row shows $Q$ and $U$ maps derived from the full-mission data set, the middle row shows the corresponding $E$ and $B$ maps, and the bottom row shows the $E$ and $B$ maps of the half-ring half-difference (HRHD) map. Note the characteristic $+$ and $\times$ patterns in the $Q$ and $U$ maps, and the clear asymmetry between $E$ and $B$ in the full data set. Also note that the HRHD $E$ map is consistent with both the full and HRHD $B$ maps.
001993893 8564_ $$81284122$$s61622$$uhttp://cds.cern.ch/record/1993893/files/figs_twopt_qu_dx11v2_hp_diff_n64.png$$y00067 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ \Planck\ 2015 CMB estimates and the corresponding means estimated from 1000 Monte Carlo simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equialteral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$ and $U_rU_rU_r$. The red solid, orange dot dot dot-dashed, green dashed and blue dot-dashed lines correspond to the \commander, \nilc, \sevem, and \smica\ maps, respectively. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284123$$s12543$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_smica_Q_case1_cmb_hp_20_40_010a_1024.png$$y00026 Caption not extracted
001993893 8564_ $$81284124$$s10865$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_int_nilc_input_cmb_080a_0128.png$$y00020 Caption not extracted
001993893 8564_ $$81284125$$s53392$$uhttp://cds.cern.ch/record/1993893/files/figs_threept_coll_qqu_ffp8_hp_diff_n64.png$$y00123 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ FFP8 CMB estimates and the corresponding means estimated from 1000 MC simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equilateral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$. The black solid, red dot dot dot-dashed, orange dashed, green dot-dashed, and blue long dashed lines correspond to the true, {\tt Commander}, {\tt NILC}, {\tt SEVEM}, and {\tt SMICA} maps, respectively. The true CMB map was analysed with added noise corresponding to the {\tt SMICA} component separation method. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284126$$s12252$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_U_commander_sevem_cmb_hp_20_40_080a_0128.png$$y00109 Pairwise difference maps between $U$ maps obtained on FFP8 siulations. Smoothing and degrading as in Figs.~\ref{fig:dx11_diff_I},\ref{fig:dx11_diff_Q}.
001993893 8564_ $$81284127$$s1177$$uhttp://cds.cern.ch/record/1993893/files/figs_colourbar_300uK.png$$y00006 Caption not extracted
001993893 8564_ $$81284128$$s11264$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_int_commander_nilc_cmb_080a_0128.png$$y00096 Pairwise difference maps between CMB temperature maps obtained on FFP8 simulations. Prior to differencing, the maps have been smoothed to 80 arcminutes FWHM and downgraded to N$_{side}$ = 128.
001993893 8564_ $$81284129$$s31349$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_nilc_U_case1_cmb_hp_20_40_010a_1024.png$$y00028 Caption not extracted
001993893 8564_ $$81284130$$s32329$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_diff_U_sevem_smica_case1_cmb_hp_20_40_080a_0128.png$$y00049 Pairwise differences between CMB $U$ maps, after smoothing and downgrading as in Fig.~\ref{fig:dx11_diff_Q}.
001993893 8564_ $$81284131$$s2041$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_spec_raw_EE.png$$y00130 Power spectra of the foreground cleaned CMB maps from FFP8 simulations. \emph{Top:} $TT$ power spectra evaluated using the \texttt{FFP8-UT74} mask. \emph{Bottom:} $EE$ power spectra evaluated using the \texttt{FFP8-UP76} mask. Thick lines show the spectra of signal plus noise estimated from the half-mission half-sum maps; thin lines show the noise levels from half-mission half-difference maps. The black line shows the input spectrum.
001993893 8564_ $$81284132$$s109864$$uhttp://cds.cern.ch/record/1993893/files/figs_comm_zoom_NEP_full_B-crop.png$$y00034 $20\deg\times20\deg$ patch of the high-pass filtered \commander\ CMB polarization map, centered on the North Ecliptic Pole, $(l,b)=(96\deg,30\deg)$. Each map is pixelized with a \healpix\ resolution of $\nside=1024$, and has an angular resolution of $10\arcm$ FWHM. The top row shows $Q$ and $U$ maps derived from the full-mission data set, the middle row shows the corresponding $E$ and $B$ maps, and the bottom row shows the $E$ and $B$ maps of the half-ring half-difference (HRHD) map. Note the characteristic $+$ and $\times$ patterns in the $Q$ and $U$ maps, and the clear asymmetry between $E$ and $B$ in the full data set. Also note that the HRHD $E$ map is consistent with both the full and HRHD $B$ maps.
001993893 8564_ $$81284133$$s12341$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_Q_commander_sevem_cmb_hp_20_40_080a_0128.png$$y00103 Pairwise difference maps between $Q$ maps obtained on FFP8 siulations. Smoothing and degrading as in Fig.~\ref{fig:dx11_diff_I}.
001993893 8564_ $$81284134$$s1049$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_nilc_int_mask_005a_2048.png$$y00088 \nilc\ masks for temperature (\emph{top}) and polarization (\emph{bottom}).
001993893 8564_ $$81284135$$s50435$$uhttp://cds.cern.ch/record/1993893/files/figs_twopt_tq_ffp8_hp_diff_n64.png$$y00118 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ FFP8 CMB estimates and the corresponding means estimated from 1000 MC simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equilateral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$. The black solid, red dot dot dot-dashed, orange dashed, green dot-dashed, and blue long dashed lines correspond to the true, {\tt Commander}, {\tt NILC}, {\tt SEVEM}, and {\tt SMICA} maps, respectively. The true CMB map was analysed with added noise corresponding to the {\tt SMICA} component separation method. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284136$$s10700$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_Q_nilc_smica_cmb_hp_20_40_080a_0128.png$$y00106 Pairwise difference maps between $Q$ maps obtained on FFP8 siulations. Smoothing and degrading as in Fig.~\ref{fig:dx11_diff_I}.
001993893 8564_ $$81284137$$s30051$$uhttp://cds.cern.ch/record/1993893/files/figs_threept_eq_ttt_ffp8_hp_diff_n64.png$$y00116 The difference between the $N$-point functions and the corresponding means estimated from 1000 MC simulations. From left to right, results for the 2-point, pseudo-collapsed 3-point, equilateral 3-point and connected rhombic 4-point functions for the $N_{\rm side}=64$ FFP8 CMB temperature estimates. The black solid, red dot dot dot-dashed, orange dashed, green dot-dashed, and blue long dashed lines correspond to the true, {\tt Commander}, {\tt NILC}, {\tt SEVEM}, and {\tt SMICA} maps, respectively. The true CMB map was analysed with added noise corresponding to the {\tt SMICA} component separation method. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284138$$s10853$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_int_smica_input_cmb_080a_0128.png$$y00022 Caption not extracted
001993893 8564_ $$81284139$$s11913$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_U_sevem_smica_cmb_hp_20_40_080a_0128.png$$y00113 Pairwise difference maps between $U$ maps obtained on FFP8 siulations. Smoothing and degrading as in Figs.~\ref{fig:dx11_diff_I},\ref{fig:dx11_diff_Q}.
001993893 8564_ $$81284140$$s11339$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_spec_raw_EE_cropped.png$$y00059 CMB $TT$ (\emph{left}) and $EE$ (\emph{right}) power spectra for each of the four foreground-cleaned maps. Top panels show raw bandpowers with no subtraction of extragalactic foregrounds; the grey lines show the best-fit \LCDM\ model from the \Planck\ 2015 likelihood. The bottom panels show residual bandpowers after subtracting the best-fit \LCDM\ model showing the residual extragalactic foreground contribution. The foregrounds are modelled and marginalized over when estimating parameters, see Figure~\ref{fig:dx11_params_TT_EE}.
001993893 8564_ $$81284141$$s10453$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_Q_commander_smica_cmb_hp_20_40_080a_0128.png$$y00104 Pairwise difference maps between $Q$ maps obtained on FFP8 siulations. Smoothing and degrading as in Fig.~\ref{fig:dx11_diff_I}.
001993893 8564_ $$81284142$$s9270$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_int_components_paper_commander_preliminary.png$$y00131 $TT$ angular power spectra of residuals from the indicated FFP8 components in the \Planck\ 2015 CMB maps, compared with the predicted signal from the best fit cosmology. ``Other'' is the sum of CO, free-free, thermal and kinetic SZ, spinning dust, and synchrotron emission. The horizontal axis is linear in $\ell^{\,0.5}$.
001993893 8564_ $$81284143$$s31813$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_int_nilc_sevem_cmb_080a_0128.png$$y00099 Pairwise difference maps between CMB temperature maps obtained on FFP8 simulations. Prior to differencing, the maps have been smoothed to 80 arcminutes FWHM and downgraded to N$_{side}$ = 128.
001993893 8564_ $$81284144$$s13031$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_smica_int_cmb_005a_2048.png$$y00005 Caption not extracted
001993893 8564_ $$81284145$$s1744$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_smica_int_mask_005a_2048.png$$y00094 \smica\ masks in temperature (\emph{top}) and polarization (\emph{bottom}).
001993893 8564_ $$81284146$$s30800$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_sevem_Q_case1_cmb_hp_20_40_010a_1024.png$$y00025 Caption not extracted
001993893 8564_ $$81284147$$s1014$$uhttp://cds.cern.ch/record/1993893/files/figs_colourbar_1uK.png$$y00043 Pairwise differences between CMB $Q$ maps, after smoothing to FWHM 80\arcm\ and downgrading to $\nside = 128$.Pairwise differences between CMB $U$ maps, after smoothing and downgrading as in Fig.~\ref{fig:dx11_diff_Q}.Caption not extractedPairwise difference maps between $Q$ maps obtained on FFP8 siulations. Smoothing and degrading as in Fig.~\ref{fig:dx11_diff_I}.Pairwise difference maps between $U$ maps obtained on FFP8 siulations. Smoothing and degrading as in Figs.~\ref{fig:dx11_diff_I},\ref{fig:dx11_diff_Q}.
001993893 8564_ $$81284148$$s12086$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_diff_int_commander_smica_cmb_080a_0128.png$$y00014 Caption not extracted
001993893 8564_ $$81284149$$s113735$$uhttp://cds.cern.ch/record/1993893/files/figs_comm_zoom_l1.png$$y00080 $5\deg\times5\deg$ zoom-in of the multi-resolution contributions to the \commander\ hybrid CMB map from the $40\arcm$ (\emph{top left}), $7\parcm5$ (\emph{top right}) and $\approx
001993893 8564_ $$81284150$$s177809$$uhttp://cds.cern.ch/record/1993893/files/figs_comm_zoom_l2.png$$y00081 $5\deg\times5\deg$ zoom-in of the multi-resolution contributions to the \commander\ hybrid CMB map from the $40\arcm$ (\emph{top left}), $7\parcm5$ (\emph{top right}) and $\approx
001993893 8564_ $$81284151$$s283616$$uhttp://cds.cern.ch/record/1993893/files/figs_comm_zoom_l3.png$$y00082 $5\deg\times5\deg$ zoom-in of the multi-resolution contributions to the \commander\ hybrid CMB map from the $40\arcm$ (\emph{top left}), $7\parcm5$ (\emph{top right}) and $\approx
001993893 8564_ $$81284152$$s66441$$uhttp://cds.cern.ch/record/1993893/files/figs_threept_eq_qqu_ffp8_hp_diff_n64.png$$y00127 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ FFP8 CMB estimates and the corresponding means estimated from 1000 MC simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equilateral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$. The black solid, red dot dot dot-dashed, orange dashed, green dot-dashed, and blue long dashed lines correspond to the true, {\tt Commander}, {\tt NILC}, {\tt SEVEM}, and {\tt SMICA} maps, respectively. The true CMB map was analysed with added noise corresponding to the {\tt SMICA} component separation method. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284153$$s33860$$uhttp://cds.cern.ch/record/1993893/files/figs_iweight_mean_plot_xlog_ylin-crop.png$$y00085 Full-sky average of needlet weights for different frequency channels and needlet bands. From top to bottom, the panels show results for temperature, $E$, and $B$ modes.
001993893 8564_ $$81284154$$s57067$$uhttp://cds.cern.ch/record/1993893/files/figs_threept_coll_qqu_dx11v2_hp_diff_n64.png$$y00070 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ \Planck\ 2015 CMB estimates and the corresponding means estimated from 1000 Monte Carlo simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equialteral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$ and $U_rU_rU_r$. The red solid, orange dot dot dot-dashed, green dashed and blue dot-dashed lines correspond to the \commander, \nilc, \sevem, and \smica\ maps, respectively. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284155$$s31718$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_nilc_int_cmb_005a_2048.png$$y00003 Differences between the component-separated CMB temperature maps from the 2013 and the 2015 releases. The maps have been smoothed to FWHM 80\arcm\ and downgrading to $\nside = 128$.
001993893 8564_ $$81284156$$s10575$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_int_nilc_smica_cmb_080a_0128.png$$y00100 Pairwise difference maps between CMB temperature maps obtained on FFP8 simulations. Prior to differencing, the maps have been smoothed to 80 arcminutes FWHM and downgraded to N$_{side}$ = 128.
001993893 8564_ $$81284157$$s13136$$uhttp://cds.cern.ch/record/1993893/files/figs_VarSkeKur_common_new_components4_cropped.png$$y00063 Polarised intensity variance (\emph{left column}), skewness (\emph{middle column}), and kurtosis (\emph{right column}) evaluated from the FFP8 Monte Carlo simulations (histogram) and from components of the fiducial FFP8 map at $\nside = 1024$ outside the \texttt{FFP8-UPA76} mask. The variance distributions have been normalized to the mean value of the Monte Carlo distributions for visualization purposes. Coloured vertical lines correspond to different combinations of components: the sum of CMB and noise is shown in blue; the sum of CMB, noise, and thermal dust is shown in green; the sum of CMB, noise, and radio point sources is shown in orange; the sum of CMB, noise, and all foregrounds is shown in red.
001993893 8564_ $$81284158$$s55585$$uhttp://cds.cern.ch/record/1993893/files/figs_threept_eq_tqq_dx11v2_hp_diff_n64.png$$y00073 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ \Planck\ 2015 CMB estimates and the corresponding means estimated from 1000 Monte Carlo simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equialteral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$ and $U_rU_rU_r$. The red solid, orange dot dot dot-dashed, green dashed and blue dot-dashed lines correspond to the \commander, \nilc, \sevem, and \smica\ maps, respectively. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284159$$s12036$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_diff_int_nilc_sevem_cmb_080a_0128.png$$y00015 Caption not extracted
001993893 8564_ $$81284160$$s11113$$uhttp://cds.cern.ch/record/1993893/files/figs_sevem_diff_2013-2014_80arcmin_nomono.png$$y00009 Caption not extracted
001993893 8564_ $$81284161$$s57042$$uhttp://cds.cern.ch/record/1993893/files/figs_threept_eq_tqq_ffp8_hp_diff_n64.png$$y00126 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ FFP8 CMB estimates and the corresponding means estimated from 1000 MC simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equilateral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$. The black solid, red dot dot dot-dashed, orange dashed, green dot-dashed, and blue long dashed lines correspond to the true, {\tt Commander}, {\tt NILC}, {\tt SEVEM}, and {\tt SMICA} maps, respectively. The true CMB map was analysed with added noise corresponding to the {\tt SMICA} component separation method. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284162$$s11557$$uhttp://cds.cern.ch/record/1993893/files/figs_smica_diff_2013-2014_80arcmin_nomono.png$$y00010 Caption not extracted
001993893 8564_ $$81284163$$s56366$$uhttp://cds.cern.ch/record/1993893/files/figs_twopt_qu_ffp8_hp_diff_n64.png$$y00120 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ FFP8 CMB estimates and the corresponding means estimated from 1000 MC simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equilateral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$. The black solid, red dot dot dot-dashed, orange dashed, green dot-dashed, and blue long dashed lines correspond to the true, {\tt Commander}, {\tt NILC}, {\tt SEVEM}, and {\tt SMICA} maps, respectively. The true CMB map was analysed with added noise corresponding to the {\tt SMICA} component separation method. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284164$$s30111$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_nilc_Q_case1_cmb_hp_20_40_010a_1024.png$$y00024 Component-separated CMB $U$ maps at resolution FWHM 10\arcm, $\nside = 1024$.
001993893 8564_ $$81284165$$s11312$$uhttp://cds.cern.ch/record/1993893/files/figs_commander_diff_2013-2014_80arcmin_nomono.png$$y00007 Caption not extracted
001993893 8564_ $$81284166$$s34589$$uhttp://cds.cern.ch/record/1993893/files/figs_threept_coll_ttq_dx11v2_hp_diff_n64.png$$y00068 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ \Planck\ 2015 CMB estimates and the corresponding means estimated from 1000 Monte Carlo simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equialteral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$ and $U_rU_rU_r$. The red solid, orange dot dot dot-dashed, green dashed and blue dot-dashed lines correspond to the \commander, \nilc, \sevem, and \smica\ maps, respectively. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284167$$s36782$$uhttp://cds.cern.ch/record/1993893/files/figs_threept_eq_ttq_dx11v2_hp_diff_n64.png$$y00072 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ \Planck\ 2015 CMB estimates and the corresponding means estimated from 1000 Monte Carlo simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equialteral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$ and $U_rU_rU_r$. The red solid, orange dot dot dot-dashed, green dashed and blue dot-dashed lines correspond to the \commander, \nilc, \sevem, and \smica\ maps, respectively. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284168$$s75475$$uhttp://cds.cern.ch/record/1993893/files/figs_threept_eq_uuu_dx11v2_hp_diff_n64.png$$y00075 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ \Planck\ 2015 CMB estimates and the corresponding means estimated from 1000 Monte Carlo simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equialteral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$ and $U_rU_rU_r$. The red solid, orange dot dot dot-dashed, green dashed and blue dot-dashed lines correspond to the \commander, \nilc, \sevem, and \smica\ maps, respectively. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284169$$s1620$$uhttp://cds.cern.ch/record/1993893/files/figs_commander_cmb_masks_pol.png$$y00078 \commander\ processing masks for temperature (\emph{top}) and polarization (\emph{bottom}). For temperature, the different shades of grey correspond to different angular resolutions, ranging from $5\arcm$ (light grey) through $7.5\arcm$ (dark grey) to 40\arcm\ FWHM (black). For polarization, the same mask is used for both 10\arcm\ and 40\arcm\ FWHM resolution.
001993893 8564_ $$81284170$$s11384$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_Q_commander_nilc_cmb_hp_20_40_080a_0128.png$$y00102 Pairwise difference maps between $Q$ maps obtained on FFP8 siulations. Smoothing and degrading as in Fig.~\ref{fig:dx11_diff_I}.
001993893 8564_ $$81284171$$s39793$$uhttp://cds.cern.ch/record/1993893/files/figs_threept_coll_ttq_ffp8_hp_diff_n64.png$$y00121 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ FFP8 CMB estimates and the corresponding means estimated from 1000 MC simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equilateral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$. The black solid, red dot dot dot-dashed, orange dashed, green dot-dashed, and blue long dashed lines correspond to the true, {\tt Commander}, {\tt NILC}, {\tt SEVEM}, and {\tt SMICA} maps, respectively. The true CMB map was analysed with added noise corresponding to the {\tt SMICA} component separation method. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284173$$s52095$$uhttp://cds.cern.ch/record/1993893/files/figs_twopt_qq_ffp8_hp_diff_n64.png$$y00119 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ FFP8 CMB estimates and the corresponding means estimated from 1000 MC simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equilateral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$. The black solid, red dot dot dot-dashed, orange dashed, green dot-dashed, and blue long dashed lines correspond to the true, {\tt Commander}, {\tt NILC}, {\tt SEVEM}, and {\tt SMICA} maps, respectively. The true CMB map was analysed with added noise corresponding to the {\tt SMICA} component separation method. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284174$$s10960$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_diff_U_nilc_smica_case1_cmb_hp_20_40_080a_0128.png$$y00048 Pairwise differences between CMB $U$ maps, after smoothing and downgrading as in Fig.~\ref{fig:dx11_diff_Q}.
001993893 8564_ $$81284175$$s115838$$uhttp://cds.cern.ch/record/1993893/files/figs_comm_zoom_NEP_full_U-crop.png$$y00032 $20\deg\times20\deg$ patch of the high-pass filtered \commander\ CMB polarization map, centered on the North Ecliptic Pole, $(l,b)=(96\deg,30\deg)$. Each map is pixelized with a \healpix\ resolution of $\nside=1024$, and has an angular resolution of $10\arcm$ FWHM. The top row shows $Q$ and $U$ maps derived from the full-mission data set, the middle row shows the corresponding $E$ and $B$ maps, and the bottom row shows the $E$ and $B$ maps of the half-ring half-difference (HRHD) map. Note the characteristic $+$ and $\times$ patterns in the $Q$ and $U$ maps, and the clear asymmetry between $E$ and $B$ in the full data set. Also note that the HRHD $E$ map is consistent with both the full and HRHD $B$ maps.
001993893 8564_ $$81284176$$s12998$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_commander_int_cmb_005a_2048.png$$y00002 Component-separated CMB temperature maps at full resolution, FWHM 5\arcm, $\nside = 2048$.
001993893 8564_ $$81284177$$s11329$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_spec_raw_TT_cropped.png$$y00058 Power spectra of the the foreground-cleaned CMB maps. \textit{Left}: $TT$ power spectra evaluated using the \texttt{UT78} mask. \textit{Right}: $EE$ power spectra evaluated using the \texttt{UP78} mask. The thick lines show the spectra of the half-mission half-sum maps containing signal and noise. The thin lines show the spectra of the half-mission half-difference maps, which given an estimate of the noise. The black line shows the \Planck\ 2015 best-fit CMB spectrum for comparison.
001993893 8564_ $$81284178$$s10905$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_diff_Q_nilc_smica_case1_cmb_hp_20_40_080a_0128.png$$y00041 Pairwise differences between CMB $Q$ maps, after smoothing to FWHM 80\arcm\ and downgrading to $\nside = 128$.
001993893 8564_ $$81284179$$s41721$$uhttp://cds.cern.ch/record/1993893/files/figs_threept_coll_tqq_ffp8_hp_diff_n64.png$$y00122 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ FFP8 CMB estimates and the corresponding means estimated from 1000 MC simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equilateral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$. The black solid, red dot dot dot-dashed, orange dashed, green dot-dashed, and blue long dashed lines correspond to the true, {\tt Commander}, {\tt NILC}, {\tt SEVEM}, and {\tt SMICA} maps, respectively. The true CMB map was analysed with added noise corresponding to the {\tt SMICA} component separation method. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284180$$s29949$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_commander_Q_case1_cmb_hp_20_40_010a_1024.png$$y00023 Component-separated CMB $Q$ maps at resolution FWHM 10\arcm, $\nside = 1024$.
001993893 8564_ $$81284181$$s6652$$uhttp://cds.cern.ch/record/1993893/files/figs_Fig7_new_compsep_robustness_test_88mm.png$$y00076 Lensing-induced $B$-mode power spectra in the component-separated polarization CMB maps. The solid line represents the best fit cosmology from the \Planck\ data release in 2015. Error bars were evaluated using a semi-analytical approximation validated over the FFP8 simulations as described in \citet{planck2014-pip116}.
001993893 8564_ $$81284182$$s5230$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_int_components_paper_smica_preliminary.png$$y00134 $TT$ angular power spectra of residuals from the indicated FFP8 components in the \Planck\ 2015 CMB maps, compared with the predicted signal from the best fit cosmology. ``Other'' is the sum of CO, free-free, thermal and kinetic SZ, spinning dust, and synchrotron emission. The horizontal axis is linear in $\ell^{\,0.5}$.
001993893 8564_ $$81284183$$s46933$$uhttp://cds.cern.ch/record/1993893/files/figs_twopt_tq_dx11v2_hp_diff_n64.png$$y00065 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ \Planck\ 2015 CMB estimates and the corresponding means estimated from 1000 Monte Carlo simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equialteral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$ and $U_rU_rU_r$. The red solid, orange dot dot dot-dashed, green dashed and blue dot-dashed lines correspond to the \commander, \nilc, \sevem, and \smica\ maps, respectively. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284184$$s34788$$uhttp://cds.cern.ch/record/1993893/files/figs_eweight_mean_plot_xlog_ylin-crop.png$$y00086 Full-sky average of needlet weights for different frequency channels and needlet bands. From top to bottom, the panels show results for temperature, $E$, and $B$ modes.
001993893 8564_ $$81284185$$s12044$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_Q_sevem_smica_cmb_hp_20_40_080a_0128.png$$y00107 Pairwise difference maps between $Q$ maps obtained on FFP8 siulations. Smoothing and degrading as in Fig.~\ref{fig:dx11_diff_I}.
001993893 8564_ $$81284186$$s1843$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_common_int_mask_005a_2048.png$$y00000 Preferred masks for analysing component-separated CMB maps in temperature (\emph{left}) and polarization (\emph{right}).
001993893 8564_ $$81284187$$s22794$$uhttp://cds.cern.ch/record/1993893/files/figs_fourpt_red_tttt_ffp8_hp_diff_n64.png$$y00117 The difference between the $N$-point functions and the corresponding means estimated from 1000 MC simulations. From left to right, results for the 2-point, pseudo-collapsed 3-point, equilateral 3-point and connected rhombic 4-point functions for the $N_{\rm side}=64$ FFP8 CMB temperature estimates. The black solid, red dot dot dot-dashed, orange dashed, green dot-dashed, and blue long dashed lines correspond to the true, {\tt Commander}, {\tt NILC}, {\tt SEVEM}, and {\tt SMICA} maps, respectively. The true CMB map was analysed with added noise corresponding to the {\tt SMICA} component separation method. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284188$$s30514$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_Q_smica_input_cmb_hp_20_40_080a_0128.png$$y00053 Caption not extracted
001993893 8564_ $$81284189$$s242908$$uhttp://cds.cern.ch/record/1993893/files/figs_comm_zoom_hybrid.png$$y00083 $5\deg\times5\deg$ zoom-in of the multi-resolution contributions to the \commander\ hybrid CMB map from the $40\arcm$ (\emph{top left}), $7\parcm5$ (\emph{top right}) and $\approx
001993893 8564_ $$81284190$$s30497$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_U_smica_input_cmb_hp_20_40_080a_0128.png$$y00057 Caption not extracted
001993893 8564_ $$81284191$$s1253$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_common_pol_mask_new_010a_1024.png$$y00001 Preferred masks for analysing component-separated CMB maps in temperature (\emph{left}) and polarization (\emph{right}).
001993893 8564_ $$81284192$$s30118$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_U_nilc_input_cmb_hp_20_40_080a_0128.png$$y00055 Caption not extracted
001993893 8564_ $$81284193$$s29405$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_Q_sevem_input_cmb_hp_20_40_080a_0128.png$$y00052 Caption not extracted
001993893 8564_ $$81284194$$s11733$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_U_nilc_sevem_cmb_hp_20_40_080a_0128.png$$y00111 Pairwise difference maps between $U$ maps obtained on FFP8 siulations. Smoothing and degrading as in Figs.~\ref{fig:dx11_diff_I},\ref{fig:dx11_diff_Q}.
001993893 8564_ $$81284195$$s12367$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_diff_Q_commander_sevem_case1_cmb_hp_20_40_080a_0128.png$$y00038 Pairwise differences between CMB $Q$ maps, after smoothing to FWHM 80\arcm\ and downgrading to $\nside = 128$.
001993893 8564_ $$81284196$$s5242$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_int_components_paper_nilc_preliminary.png$$y00132 $TT$ angular power spectra of residuals from the indicated FFP8 components in the \Planck\ 2015 CMB maps, compared with the predicted signal from the best fit cosmology. ``Other'' is the sum of CO, free-free, thermal and kinetic SZ, spinning dust, and synchrotron emission. The horizontal axis is linear in $\ell^{\,0.5}$.
001993893 8564_ $$81284197$$s1164$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_sevem_pol_mask_010a_1024.png$$y00091 \sevem\ masks in temperature (\emph{top}) and polarization (\emph{bottom}).
001993893 8564_ $$81284198$$s29987$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_Q_nilc_input_cmb_hp_20_40_080a_0128.png$$y00051 Difference between output and input CMB $U$ maps from FFP8 simulations. Smoothing and downgrading as in Fig.~\ref{fig:ffp8_res_Q}.
001993893 8564_ $$81284199$$s4456$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_pol_components_paper_sevem_preliminary.png$$y00137 $EE$ angular power spectra of residuals from the indicated FFP8 components in the \Planck\ 2015 CMB maps, compared with the predicted signal from the best fit cosmology. ``Other'' is the sum of CO, free-free, thermal and kinetic SZ, spinning dust, far-infrared background, and radio and infrared unresolved sources. The horizontal axis is linear in $\ell^{\,0.5}$.
001993893 8564_ $$81284200$$s30942$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_U_commander_input_cmb_hp_20_40_080a_0128.png$$y00054 Caption not extracted
001993893 8564_ $$81284201$$s1228$$uhttp://cds.cern.ch/record/1993893/files/figs_colourbar_7_5uK.png$$y00018 Caption not extractedPairwise difference maps between CMB temperature maps obtained on FFP8 simulations. Prior to differencing, the maps have been smoothed to 80 arcminutes FWHM and downgraded to N$_{side}$ = 128.
001993893 8564_ $$81284202$$s11642$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_diff_U_commander_smica_case1_cmb_hp_20_40_080a_0128.png$$y00046 Pairwise differences between CMB $U$ maps, after smoothing and downgrading as in Fig.~\ref{fig:dx11_diff_Q}.
001993893 8564_ $$81284203$$s12381$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_diff_Q_sevem_smica_case1_cmb_hp_20_40_080a_0128.png$$y00042 Pairwise differences between CMB $Q$ maps, after smoothing to FWHM 80\arcm\ and downgrading to $\nside = 128$.
001993893 8564_ $$81284204$$s34562$$uhttp://cds.cern.ch/record/1993893/files/figs_bweight_mean_plot_xlog_ylin-crop.png$$y00087 Full-sky average of needlet weights for different frequency channels and needlet bands. From top to bottom, the panels show results for temperature, $E$, and $B$ modes.
001993893 8564_ $$81284205$$s2403$$uhttp://cds.cern.ch/record/1993893/files/figs_commander_cmb_masks.png$$y00077 \commander\ processing masks for temperature (\emph{top}) and polarization (\emph{bottom}). For temperature, the different shades of grey correspond to different angular resolutions, ranging from $5\arcm$ (light grey) through $7.5\arcm$ (dark grey) to 40\arcm\ FWHM (black). For polarization, the same mask is used for both 10\arcm\ and 40\arcm\ FWHM resolution.
001993893 8564_ $$81284206$$s54259$$uhttp://cds.cern.ch/record/1993893/files/figs_twopt_qq_dx11v2_hp_diff_n64.png$$y00066 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ \Planck\ 2015 CMB estimates and the corresponding means estimated from 1000 Monte Carlo simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equialteral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$ and $U_rU_rU_r$. The red solid, orange dot dot dot-dashed, green dashed and blue dot-dashed lines correspond to the \commander, \nilc, \sevem, and \smica\ maps, respectively. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284207$$s29872$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_U_sevem_input_cmb_hp_20_40_080a_0128.png$$y00056 Caption not extracted
001993893 8564_ $$81284208$$s11705$$uhttp://cds.cern.ch/record/1993893/files/figs_nilc_diff_2013-2014_80arcmin_nomono.png$$y00008 Caption not extracted
001993893 8564_ $$81284209$$s30716$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_Q_commander_input_cmb_hp_20_40_080a_0128.png$$y00050 Difference between output and input CMB $Q$ maps from FFP8 simulations. Smoothing and downgrading as in Figs.~\ref{fig:dx11_diff_Q} and \ref{fig:dx11_diff_U}.
001993893 8564_ $$81284210$$s53738$$uhttp://cds.cern.ch/record/1993893/files/figs_threept_coll_uuu_ffp8_hp_diff_n64.png$$y00124 The difference between the $N$-point functions for the high-pass filtered $N_{\rm side}=64$ FFP8 CMB estimates and the corresponding means estimated from 1000 MC simulations. The Stokes parameters $Q_r$ and $U_r$ were locally rotated so that the correlation functions are independent of coordinate frame. The first row shows results for the 2-point function, from left to right, $TQ_r$, $Q_rQ_r$, and $Q_rU_r$. The second row shows results for the pseudo-collapsed 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$, and the third row shows results for the equilateral 3-point function, from left to right, $TTQ_r$, $TQ_rQ_r$, $Q_rQ_rU_r$, and $U_rU_rU_r$. The black solid, red dot dot dot-dashed, orange dashed, green dot-dashed, and blue long dashed lines correspond to the true, {\tt Commander}, {\tt NILC}, {\tt SEVEM}, and {\tt SMICA} maps, respectively. The true CMB map was analysed with added noise corresponding to the {\tt SMICA} component separation method. The shaded dark and light grey regions indicate the 68\% and 95\% confidence regions, respectively, estimated using \smica\ simulations. See Sect.~\ref{sec:npoint_correlation} for the definition of the separation angle $\theta$.
001993893 8564_ $$81284211$$s12272$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_diff_U_commander_nilc_case1_cmb_hp_20_40_080a_0128.png$$y00044 Pairwise differences between CMB $U$ maps, after smoothing and downgrading as in Fig.~\ref{fig:dx11_diff_Q}.
001993893 8564_ $$81284212$$s31264$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_diff_U_nilc_sevem_case1_cmb_hp_20_40_080a_0128.png$$y00047 Pairwise differences between CMB $U$ maps, after smoothing and downgrading as in Fig.~\ref{fig:dx11_diff_Q}.
001993893 8564_ $$81284213$$s31794$$uhttp://cds.cern.ch/record/1993893/files/figs_dx11_v2_sevem_int_cmb_005a_2048.png$$y00004 Caption not extracted
001993893 8564_ $$81284214$$s11562$$uhttp://cds.cern.ch/record/1993893/files/figs_ffp8_diff_int_sevem_input_cmb_080a_0128.png$$y00021 Caption not extracted
001993893 8564_ $$81284215$$s101524$$uhttp://cds.cern.ch/record/1993893/files/figs_comm_zoom_NEP_hrhd_B-crop.png$$y00036 $20\deg\times20\deg$ patch of the high-pass filtered \commander\ CMB polarization map, centered on the North Ecliptic Pole, $(l,b)=(96\deg,30\deg)$. Each map is pixelized with a \healpix\ resolution of $\nside=1024$, and has an angular resolution of $10\arcm$ FWHM. The top row shows $Q$ and $U$ maps derived from the full-mission data set, the middle row shows the corresponding $E$ and $B$ maps, and the bottom row shows the $E$ and $B$ maps of the half-ring half-difference (HRHD) map. Note the characteristic $+$ and $\times$ patterns in the $Q$ and $U$ maps, and the clear asymmetry between $E$ and $B$ in the full data set. Also note that the HRHD $E$ map is consistent with both the full and HRHD $B$ maps.
001993893 8564_ $$81284216$$s5375$$uhttp://cds.cern.ch/record/1993893/files/figs_smica_filterEB_dx11.png$$y00093 \smica\ weights for temperature (\emph{top}) and polarization (\emph{bottom}). For readibility, the values are shown for input maps in units of antenna temperature. The plot goes up to $\ell\sim3600$, but the output maps are synthesized uses all multipoles up to $\ell=4000$. For polarization, the thick solid lines show the contribution of input $E$ modes to the CMB $E$ modes and the thick dashed lines show the same for the $B$ modes. The thin lines, all close to zero, show ``cross-contributions'' of input $E$ modes to the CMB $B$ modes and vice versa.
001993893 916__ $$sn$$w201508
001993893 960__ $$a13
001993893 980__ $$aARTICLE