Planck intermediate results. XLVII. Planck constraints on reionization history
Planck Collaboration, R. Adam, N. Aghanim, M. Ashdown, J. Aumont, C. Baccigalupi, M. Ballardini, A. J. Banday, R. B. Barreiro, N. Bartolo, S. Basak, R. Battye, K. Benabed, J.-P. Bernard, M. Bersanelli, P. Bielewicz, J. J. Bock, A. Bonaldi, L. Bonavera, J. R. Bond, J. Borrill, F. R. Bouchet, M. Bucher, C. Burigana, E. Calabrese, J.-F. Cardoso, J. Carron, H. C. Chiang, L. P. L. Colombo, C. Combet, B. Comis, A. Coulais, B. P. Crill, A. Curto, F. Cuttaia, R. J. Davis, P. de Bernardis, A. de Rosa, G. de Zotti, J. Delabrouille, E. Di Valentino, C. Dickinson, J. M. Diego, O. Doré, M. Douspis, A. Ducout, X. Dupac, F. Elsner, T. A. Enßlin, H. K. Eriksen, E. Falgarone, Y. Fantaye, F. Finelli, F. Forastieri, M. Frailis, A. A. Fraisse, E. Franceschi, A. Frolov, S. Galeotta, S. Galli, K. Ganga, R. T. Génova-Santos, M. Gerbino, T. Ghosh, J. González-Nuevo, K. M. Górski, A. Gruppuso, J. E. Gudmundsson, F. K. Hansen, G. Helou, S. Henrot-Versillé, D. Herranz, E. Hivon, Z. Huang, S. Ili_, A. H. Jaffe, W. C. Jones, E. Keihänen, R. Keskitalo, T. S. Kisner, L. Knox, N. Krachmalnicoff, M. Kunz, H. Kurki-Suonio, G. Lagache, A. Lähteenmäki, J.-M. Lamarre, M. Langer, A. Lasenby, M. Lattanzi, C. R. Lawrence, M. Le Jeune, F. Levrier, A. Lewis, M. Liguori, P. B. Lilje, M. López-Caniego, Y.-Z. Ma, J. F. Macías-Pérez, G. Maggio, et al. (66 additional authors not shown)
AAstronomy & Astrophysics manuscript no. reionisation c (cid:13)
ESO 2016September 6, 2016
Planck intermediate results. XLVII.Planck constraints on reionization history
Planck Collaboration: R. Adam , N. Aghanim , M. Ashdown , , J. Aumont , C. Baccigalupi , M. Ballardini , , , A. J. Banday , ,R. B. Barreiro , N. Bartolo , , S. Basak , R. Battye , K. Benabed , , J.-P. Bernard , , M. Bersanelli , , P. Bielewicz , , , J. J. Bock , ,A. Bonaldi , L. Bonavera , J. R. Bond , J. Borrill , , F. R. Bouchet , , F. Boulanger , M. Bucher , C. Burigana , , , E. Calabrese ,J.-F. Cardoso , , , J. Carron , H. C. Chiang , , L. P. L. Colombo , , C. Combet , B. Comis , F. Couchot , A. Coulais , B. P. Crill , ,A. Curto , , , F. Cuttaia , R. J. Davis , P. de Bernardis , A. de Rosa , G. de Zotti , , J. Delabrouille , E. Di Valentino , , C. Dickinson ,J. M. Diego , O. Dor´e , , M. Douspis , A. Ducout , , X. Dupac , F. Elsner , , , T. A. Enßlin , H. K. Eriksen , E. Falgarone ,Y. Fantaye , , F. Finelli , , F. Forastieri , , M. Frailis , A. A. Fraisse , E. Franceschi , A. Frolov , S. Galeotta , S. Galli , K. Ganga ,R. T. G´enova-Santos , , M. Gerbino , , , T. Ghosh , J. Gonz´alez-Nuevo , , K. M. G´orski , , A. Gruppuso , , J. E. Gudmundsson , , ,F. K. Hansen , G. Helou , S. Henrot-Versill´e , D. Herranz , E. Hivon , , Z. Huang , S. Ili´c , , , A. H. Ja ff e , W. C. Jones , E. Keih¨anen ,R. Keskitalo , T. S. Kisner , L. Knox , N. Krachmalnico ff , M. Kunz , , , H. Kurki-Suonio , , G. Lagache , , A. L¨ahteenm¨aki , ,J.-M. Lamarre , M. Langer , A. Lasenby , , M. Lattanzi , , C. R. Lawrence , M. Le Jeune , F. Levrier , A. Lewis , M. Liguori , ,P. B. Lilje , M. L´opez-Caniego , Y.-Z. Ma , , J. F. Mac´ıas-P´erez , G. Maggio , A. Mangilli , , M. Maris , P. G. Martin ,E. Mart´ınez-Gonz´alez , S. Matarrese , , , N. Mauri , J. D. McEwen , P. R. Meinhold , A. Melchiorri , , A. Mennella , ,M. Migliaccio , , M.-A. Miville-Deschˆenes , , D. Molinari , , , A. Moneti , L. Montier , , G. Morgante , A. Moss , P. Naselsky , ,P. Natoli , , , C. A. Oxborrow , L. Pagano , , D. Paoletti , , B. Partridge , G. Patanchon , L. Patrizii , O. Perdereau , L. Perotto ,V. Pettorino , F. Piacentini , S. Plaszczynski , L. Polastri , , G. Polenta , , J.-L. Puget , J. P. Rachen , , B. Racine , M. Reinecke ,M. Remazeilles , , , A. Renzi , , G. Rocha , , M. Rossetti , , G. Roudier , , , J. A. Rubi˜no-Mart´ın , , B. Ruiz-Granados , L. Salvati ,M. Sandri , M. Savelainen , , D. Scott , G. Sirri , R. Sunyaev , , A.-S. Suur-Uski , , J. A. Tauber , M. Tenti , L. To ff olatti , , ,M. Tomasi , , M. Tristram ∗ , T. Trombetti , , J. Valiviita , , F. Van Tent , P. Vielva , F. Villa , N. Vittorio , B. D. Wandelt , , ,I. K. Wehus , , M. White , A. Zacchei , A. Zonca (A ffi liations can be found after the references) Accepted 23 July 2016
ABSTRACT
We investigate constraints on cosmic reionization extracted from the
Planck cosmic microwave background (CMB) data. We combine the
Planck
CMB anisotropy data in temperature with the low-multipole polarization data to fit Λ CDM models with various parameterizations of the reioniza-tion history. We obtain a Thomson optical depth τ = . ± .
012 for the commonly adopted instantaneous reionization model. This confirms,with data solely from CMB anisotropies, the low value suggested by combining
Planck ff ect using additional information from the high-resolution Atacama Cosmology Telescopeand South Pole Telescope experiments. The average redshift at which reionization occurs is found to lie between z = . ∆ z < .
8. In all cases, we find that the Universe is ionized at less than the 10 % level at redshifts above z (cid:39)
10. This suggeststhat an early onset of reionization is strongly disfavoured by the
Planck data. We show that this result also reduces the tension between CMB-basedanalyses and constraints from other astrophysical sources.
Key words.
Cosmology – cosmic background radiation – Polarization – dark ages, reionization, first stars
1. Introduction
The process of cosmological recombination happened aroundredshift z (cid:39) ff ect (Gunn & Peterson 1965) in quasarspectra (Becker et al. 2001; Fan et al. 2006b; Venemans et al.2013; Becker et al. 2015) indicate that intergalactic gas had be-come almost fully reionized by redshift z (cid:39)
6. Reionization is ∗ Corresponding authors:M. Tristram [email protected],M. Douspis [email protected] thus the second major change in the ionization state of hydrogenin the Universe. Details of the transition from the neutral to ion-ized Universe are still the subject of intense investigations (for arecent review, see the book by Mesinger 2016). In the currentlyconventional picture, early galaxies reionize hydrogen progres-sively throughout the entire Universe between z (cid:39)
12 and z (cid:39) z (cid:39) (cid:39) a r X i v : . [ a s t r o - ph . C O ] S e p lanck Collaboration: Planck constraints on reionization history The CMB is a ff ected by the total column density of free elec-trons along each line of sight, parameterized by its Thomsonscattering optical depth τ . This is one of the six parameters ofthe baseline Λ CDM cosmological model and is the key mea-surement for constraining reionization. Large-scale anisotropiesin polarization are particularly sensitive to the value of τ . TheWMAP mission was the first to extract a τ measurement throughthe correlation between the temperature field and the E -modepolarization (i.e., the T E power spectrum) over a large fractionof the sky. This measurement is very demanding, since the ex-pected level of the E -mode polarization power spectrum at lowmultipoles ( (cid:96) <
10) is only a few times 10 − µ K , lower by morethan two orders of magnitude than the level of the temperatureanisotropy power spectrum. For such weak signals the di ffi cultyis not only to have enough detector sensitivity, but also to reduceand control both instrumental systematic e ff ects and foregroundresiduals to a very low level. This di ffi culty is illustrated by theimprovements over time in the WMAP-derived τ estimates. The1-year results gave a value of τ = . ± .
04, based on thetemperature-polarization
T E cross-power spectrum (Kogut et al.2003). In the 3-year release, this was revised down to 0 . ± . E -modes alone, whereas the combined T T , T E , and EE power spectra gave 0 . ± .
03 (Page et al. 2007). Error bars im-proved in further WMAP analyses, ending up with 0 . ± . Planck satellite cosmological results were based on Planck temperature powerspectra combined with the polarized WMAP data and gave thesame value τ = . ± .
014 (Planck Collaboration XVI 2014).However, using a preliminary version of the
Planck
353 GHzpolarization maps to clean the dust emission (in place of theWMAP dust model), the optical depth was reduced by approx-imately 1 σ to τ = . ± .
013 (Planck Collaboration XV2014).In the 2015 Planck analysis (Planck Collaboration XIII2016), the Low Frequency Instrument (LFI) low-resolution mapspolarization at 70 GHz were used. Foreground cleaning was per-formed using the LFI 30 GHz and High Frequency Instrument(HFI) 353 GHz maps, operating e ff ectively as polarized syn-chrotron and dust templates, respectively. The optical depthwas found to be τ = . ± . . ± .
016 when adding CMB lensing data. This value isalso in agreement with the constraints from the combination“PlanckTT + lensing + BAO,” yielding τ = . ± . (cid:96) polarization.In this paper and its companion(Planck Collaboration Int. XLVI 2016), we derive the firstestimate of τ from the Planck -HFI polarization data at largescales. For the astrophysical interpretation, the power spectraare estimated using a PCL estimate which is more conservative.Indeed, it gives a slightly larger distribution on τ than the QMLestimator used in Planck Collaboration Int. XLVI (2016) butis less sensitive to the limited number of simulations availablefor the analysis. Using only E -mode polarization, the Planck lollipop likelihood gives τ = . + . − . for a standard instan-taneous reionization model, when all other Λ CDM parametersare fixed to their
Planck -2015 best-fit values. We show that in Planck ( ) is a project of theEuropean Space Agency (ESA) with instruments provided by two sci-entific consortia funded by ESA member states and led by PrincipalInvestigators from France and Italy, telescope reflectors providedthrough a collaboration between ESA and a scientific consortium ledand funded by Denmark, and additional contributions from NASA(USA). combination with the Planck temperature data the error bars areimproved and we find τ = . ± . Λ CDM model, improved accuracy on the reioniza-tion optical depth helps to reduce the degeneracies with otherparameters. In particular, the measurement of τ reduces the cor-relation with the normalization of the initial power spectrum A s and its spectral index n s . In addition to this τ is a particularlyimportant source of information for constraining the history ofreionization, which is the main subject of this paper. When com-bined with direct probes at low redshift, a better knowledge ofthe value of the CMB optical depth parameter may help to char-acterize the duration of the EoR, and thus tell us when it started.In addition to the e ff ect of reionization on the po-larized large-scale CMB anisotropies, reionization generatesCMB temperature anisotropies through the kinetic Sunyaev-Zeldovich (kSZ) e ff ect (Sunyaev & Zeldovich 1980), caused bythe Doppler shift of photons scattering o ff electrons moving withbulk velocities. Simulations have shown that early homogeneousand patchy reionization scenarious di ff erently a ff ect the shape ofthe kSZ power spectrum, allowing us to place constraints on thereionization history (e.g., McQuinn et al. 2005; Aghanim et al.2008). Zahn et al. (2012) derived the first constraints on theepoch of reionization from the combination of kSZ and low- (cid:96) CMB polarization, specifically using the low- (cid:96) polarizationpower spectrum from WMAP and the very high multipoles ofthe temperature angular power spectrum from the South PoleTelescope (SPT, Reichardt et al. 2012). However one shouldkeep in mind that kSZ signal is complicated to predict and de-pends on detailed astrophysics which makes the constraints onreionization di ffi cult to interpret (Mesinger et al. 2012).In this paper, we investigate constraints on the epoch ofreionization coming from Planck . Section 2 first briefly de-scribes the pre-2016 data and likelihood used in this paper,which are presented in detail in Planck Collaboration Int. XLVI(2016). In Sect. 3 we then present the parameterizations weadopt for the ionization fraction, describing the reionization his-tory as a function of redshift. In Sect. 4, we show the results ob-tained from the CMB observables (i.e., the optical depth τ andthe amplitude of the kSZ e ff ect) in the case of “instantaneous”reionization. Section 5 presents results based on the CMB mea-surements by considering di ff erent models for the ionization his-tory. In particular, we derive limits on the reionization redshiftand duration. Finally, in Sect. 6, we derive the ionization frac-tion as a function of redshift and discuss how our results relateto other astrophysical constraints.
2. Data and likelihood
The present analysis is based on the pre-2016 full mis-sion intensity and polarization
Planck -HFI maps at 100 and143 GHz. The data processing and the beam description arethe same as in the
Planck
Planck -HFI polariza-tion maps are constructed from the combination of polarized de-tectors that have fixed polarization direction. The
Planck scan-ning strategy produces a relatively low level of polarization an-gle measurement redundancy on the sky, resulting in a high levelof I - Q - U mixing, as shown in Planck Collaboration VIII (2016).As a consequence, any instrumental mismatch between detec-tors from the same frequency channel produces leakage from in-tensity to polarization. This temperature-to-polarization leakagewas at one point the main systematic e ff ect present in the Planck - HFI data, and prevented robust low- (cid:96) polarization measurementsfrom being included in the previous
Planck data releases.The maps that we use here di ff er in some respects from thosedata released in 2015. The updated mapmaking procedure, pre-sented in Planck Collaboration Int. XLVI (2016), now allows fora significant reduction of the systematic e ff ects in the maps. Inparticular, the relative calibration within a channel is now accu-rate to better than 0.001 %, which ensures a very low level ofgain-mismatch between detectors. The major systematic e ff ectthat remains in the pre-2016 maps is due to imperfections of thecorrection for nonlinearity in the analogue-to-digital converters(ADCs) but produces very low level of residuals in the maps. Inaddition to the 100 GHz and 143 GHz maps, we also make useof 30 GHz LFI data (Planck Collaboration II 2016) and 353 GHzHFI data to remove polarized foregrounds.Using the pre-2016 end-to-end simulations, we show that thepower spectrum bias induced by the remaining nonlinearities isvery small and properly accounted for in the likelihood. Figure 1shows the bias (in the quantity D (cid:96) ≡ (cid:96) ( (cid:96) + C (cid:96) / π , where C (cid:96) isthe conventional power spectrum) computed as the mean of the EE cross-power spectra from simulated maps, including realis-tic noise and systematic e ff ects without and with Galactic fore-grounds. In the latter case, the foregrounds are removed for eachsimulation using the 30 GHz and 353 GHz maps as templates forsynchrotron and dust, respectively. The resulting bias in the EE ×
143 cross-power spectrum can be used to correct the mea-sured cross-spectrum, but in fact has very little impact on thelikelihood.
Multipole ‘ D ‘ [ µ K ] With foregroundsWithout foregrounds
Fig. 1.
Bias in the 100 ×
143 cross-power spectrum computedfrom simulations, including instrumental noise and systematice ff ects, with or without foregrounds (dark blue and light blue),compared to the cosmic variance level (in grey).Furthermore, we use end-to-end simulations to propagate thesystematic uncertainties to the cross-power spectra and all theway to the cosmological parameters. Figure 2 shows the impacton the variance due to the inclusion of the main ADC nonlin-earity systematic e ff ect, compared to realistic noise and cosmicvariance. The resulting C (cid:96) covariance matrix is estimated fromthese Monte Carlos. In the presence of such systematic e ff ects,the variance of the C (cid:96) is shown to be higher by roughly a factorof 2 compared to the pure noise case.Polarized foregrounds at Planck -HFI frequencies are es-sentially dominated by Galactic dust emission, but also in-
Multipole ‘ -5 -4 -3 -2 -1 ∆ C ‘ [ µ K ] Cosmic varianceNoiseSystematic+noise
Fig. 2.
Variance of the 100 × EE cross-power spectrum forsimulations, including instrumental noise and noise plus system-atic e ff ects, compared to cosmic variance.clude a small contribution from synchrotron emission. Weuse the 353 GHz and 30 GHz Planck maps as templates tosubtract dust and synchrotron, respectively, using a singlecoe ffi cient for each component over 95 % of the sky (seePlanck Collaboration IX 2016; Planck Collaboration X 2016).However, foreground residuals in the maps are still dominantover the CMB polarized signal near the Galactic plane. We there-fore apply a very conservative mask, based on the amplitude ofthe polarized dust emission, which retains 50 % of the sky for thecosmological analysis. Outside this mask, the foreground residu-als are found to be lower than 0.3 and 0.4 µ K in Q and U Stokespolarization maps at 100 and 143 GHz, respectively. We havechecked that our results are very stable when using a larger skyfraction of 60 %.In this paper, we also make use of the constraints de-rived from the observation of the Gunn-Peterson e ff ect on high-redshift quasars. As suggested by Fan et al. (2006a), these mea-surements show that the Universe was almost fully reionized atredshift z (cid:39)
6. We later discuss the results obtained with andwithout imposing a prior on the redshift of the end of reioniza-tion.
For temperature anisotropies, we use the combined
Planck likelihood (hereafter “PlanckTT”), which includes the
T T power spectrum likelihood at multipoles (cid:96) >
30 (using the
Plik code) and the low- (cid:96) temperature-only likelihood basedon the CMB map recovered from the component-separationprocedure (specifically
Commander ) described in detail inPlanck Collaboration XI (2016).For polarization, we use the
Planck low- (cid:96) EE polarizationlikelihood (hereafter lollipop ), a cross-spectra-based likeli-hood approach described in detail in Mangilli et al. (2015) andapplied to Planck data as discussed here in Appendix A. Themultipole range used is (cid:96) = C (cid:96) estimator Xpol (a generalization to polar-ization of the algorithm presented in Tristram et al. 2005). For afull-sky analysis, the statistics of the reconstructed C (cid:96) are givenby a χ distribution that is uncorrelated between multipoles.For a cut-sky analysis, the distribution is more complex and Multipole ‘ ‘ ( ‘ + ) C EE ‘ / π [ µ K ] τ Multipole ‘ ‘ ( ‘ + ) C T E ‘ / π [ µ K ] τ Fig. 3. EE and T E power spectra for various τ values ranging from 0.04 to 0.08. The ionization fraction is modelled using aredshift-symmetric tanh function with δ z = .
5. Grey bands represent the cosmic variance (full-sky) associated with the τ = . (cid:96) -to- (cid:96) correlations. Hamimeche & Lewis (2008) pro-posed an approximation of the likelihood for cut-sky auto-powerspectra that was adapted by Mangilli et al. (2015) to be suitablefor cross-spectra. Cross-spectra between independent data setsshow common sky signal, but are not biased by the noise be-cause this should be uncorrelated. This approximation assumesthat any systematic residuals are not correlated between the dif-ferent data sets; We have shown using realistic simulations (in-cluding Planck -HFI noise characteristics and systematic e ff ectresiduals), that the bias in the cross-spectra is very small and canbe corrected for at the power-spectrum level. Nevertheless, wechoose to remove the first two multipoles ( (cid:96) = (cid:96) = C (cid:96) covariance matrixused in the likelihood, which propagates both the noise and thesystematic uncertainties. For the astrophysical interpretation, thepower-spectra are estimated with a PCL estimate which is moreconservative. Indeed, it gives a slightly larger distribution on τ than a QML estimator but is less sensitive to the limited numberof simulations available for the analysis.With Planck sensitivity in polarization, the results from thelow- (cid:96) EE power spectrum dominate the constraints compared tothe T E power spectrum, as can be seen in Fig. 3. This is becauseof the relatively larger cosmic variance for
T E (arising from thetemperature term) and the intrinsically weaker dependence on τ ( ∝ τ compared with τ for EE ), as well as the fact that thereis only partial correlation between T and E . As a consequence,we do not consider the T E data in this analysis. Furthermore,we do not make use of the high- (cid:96) likelihoods in EE and T E from
Planck , since they do not carry additional information onreionization parameters.
Planck temperature observations are complemented atsmaller angular scales by measurements from the ground-based Atacama Cosmology Telescope (ACT) and South PoleTelescope (SPT). As explained in Planck Collaboration XI(2016), the high- (cid:96) likelihood (hereafter VHL) includes ACTpower spectra at 148 and 218 GHz (Das et al. 2014), with a re-vised binning (described in Calabrese et al. 2013) and final beamestimates (Hasselfield et al. 2013), together with SPT measure-ments in the range 2000 < (cid:96) <
13 000 from the 2540 deg SPT-SZ survey at 95, 150, and 220 GHz (George et al. 2015).To assess the consistency between these data sets, we extend the
Planck foreground models up to (cid:96) =
13 000, with addi-tional nuisance parameters for ACT and SPT (as described inPlanck Collaboration XIII 2016). We use the same models forcosmic infrared background (CIB) fluctuations, the thermal SZ(tSZ) e ff ect, kSZ e ff ect, and CIB × tSZ components. The kSZtemplate used in the Planck A s and τ (as discussed inPlanck Collaboration XIII 2016). Despite this potential forimprovement, we show in Sect. 4.1 that Planck ’s low- (cid:96) polar-ization signal-to-noise ratio is su ffi ciently high that the lensingdoes not bring much additional information for the reionizationconstraints.The Planck reference cosmology used in this paper corre-sponds to the PlanckTT + lowP + lensing best fit, as described intable 4, column 2 of Planck Collaboration XIII (2016), namely Ω b h = . Ω c h = . Ω m = . n s = . H = .
81 km s − Mpc − , for which Y P = . Planck likelihoods:the temperature power spectrum likelihood at high (cid:96) ; the “lowP”temperature + polarization likelihood, based on the foreground-cleaned LFI 70 GHz polarization maps, together with the tem-perature map from the Commander component-separation algo-rithm; and the power spectrum of the lensing potential as mea-sured by
Planck .
3. Parametrization of reionization history
The epoch of reionization (EoR) is the period during whichthe cosmic gas transformed from a neutral to ionized stateat the onset of the first sources. Details of the transition arethus strongly connected to many fundamental questions in cos-mology, such as what were the properties of the first galax-ies and the first (mini-)quasars, how did the formation ofvery metal-poor stars proceed, etc. We certainly know that, atsome point, luminous sources started emitting ultraviolet ra- diation that reionized the neutral regions around them. Aftera su ffi cient number of ionizing sources had formed, the av-erage ionized fraction of the gas in the Universe rapidly in-creased until hydrogen became fully ionized. Empirical, ana-lytic, and numerical models of the reionization process havehighlighted many pieces of the essential physics that led to thebirth to the ionized intergalactic medium (IGM) at late times(Couchman & Rees 1986; Miralda-Escude & Ostriker 1990;Meiksin & Madau 1993; Aghanim et al. 1996; Gruzinov & Hu1998; Madau et al. 1999; Gnedin 2000; Barkana & Loeb 2001;Ciardi et al. 2003; Furlanetto et al. 2004; Pritchard et al. 2010;Pandolfi et al. 2011; Mitra et al. 2011; Iliev et al. 2014). Suchstudies provide predictions on the various reionization observ-ables, including those associated with the CMB.The most common physical quantity used to characterizereionization is the Thomson scattering optical depth defined as τ ( z ) = (cid:90) t t ( z ) n e σ T c d t (cid:48) , (1)where n e is the number density of free electrons at time t (cid:48) , σ T is the Thomson scattering cross-section, t is the time today, t ( z ) is the time at redshift z , and we can use the Friedmannequation to convert d t to d z . The reionization history is con-veniently expressed in terms of the ionized fraction x e ( z ) ≡ n e ( z ) / n H ( z ) where n H ( z ) is the hydrogen number density. In prac-tice, the CMB is sensitive to the average over all sky directionsof x e (1 + δ b ) (where δ b denotes the baryon overdensity). TheIGM is likely to be very inhomogeneous during reionization pro-cess, with ionized bubbles embedded in neutral surroundings,which would impact the relation between the optical depth andthe reionisation parameters (see Liu et al. 2016) at a level whichis neglected in this paper.In this study, we define the redshift of reionization, z re ≡ z
50 % , as the redshift at which x e = . × f . Here the normaliza-tion, f = + f He = + n He / n H , takes into account electrons in-jected into the IGM by the first ionization of helium (correspond-ing to 25 eV), which is assumed to happen roughly at the sametime as hydrogen reionization. We define the beginning and theend of the EoR by the redshifts z beg ≡ z
10 % and z end ≡ z
99 % atwhich x e = . × f and 0 . × f , respectively. The duration of theEoR is then defined as ∆ z = z
10 % − z
99 % . Moreover, to ensurethat the Universe is fully reionized at low redshift, we imposethe condition that the EoR is completed before the second he-lium reionization phase (corresponding to 54 eV), noting that itis commonly assumed that quasars are necessary to produce thehard photons needed to ionize helium. To be explicit about howwe treat the lowest redshifts we assume that the full reioniza-tion of helium happens fairly sharply at z He = . δ z = .
5. While there is still some debate on whether he-lium reionization could be inhomogeneous and extended (andthus have an early start, Worseck et al. 2014), we have checkedthat varying the helium reionization redshift between 2.5 and 4.5changes the total optical depth by less than 1 %.The simplest and most widely-used parameterizations de-scribes the EoR as a step-like transition between an essentiallyvanishing ionized fraction x e at early times, to a value of unityat low redshifts. When calculating the e ff ect on anisotropies it is The reason this is not defined symmetrically is that in practice wehave tighter constraints on the end of reionization than on the beginning. The ionized fraction is actually matched to the relic free electrondensity from recombination, calculated using recfast
Seager et al.(2000). necessary to give a non-zero width to the transition, and it canbe modelled using a tanh function (Lewis 2008): x e ( z ) = f (cid:34) + tanh (cid:32) y − y re δ y (cid:33)(cid:35) , (2)where y = (1 + z ) / and δ y = (1 + z ) / δ z . The key parametersare thus z re , which measures the redshift at which the ionizedfraction reaches half its maximum and a width δ z . The tanh pa-rameterization of the EoR transition allows us to compute the op-tical depth of Eq. (1) for a one-stage almost redshift-symmetric reionization transition, where the redshift interval between theonset of the reionization process and its half completion is (byconstruction) equal to the interval between half completion andfull completion. In this parameterization, the optical depth ismainly determined by z re and almost degenerate with the width δ z . This is the model used in the Planck δ z = . ∆ z = . α emitters, or γ -ray bursts(Faisst et al. 2014; Chornock et al. 2014; Ishigaki et al. 2015;Robertson et al. 2015; Bouwens et al. 2015). The two simplestchoices of redshift-asymmetric parameterizations are polyno-mial or exponential functions of redshift (Douspis et al. 2015).These two parameterizations are in fact very similar, and weadopt here a power law defined by two parameters: the redshift atwhich reionization ends ( z end ); and the exponent α . Specificallywe have x e ( z ) = f for z < z end , f (cid:18) z early − zz early − z end (cid:19) α for z > z end . (3)In the following, we fix z early =
20, the redshift around whichthe first emitting sources form, and at which we smoothly match x e ( z ) to the ionized fraction left over from recombination. Wechecked that our results are not sensitive to the precise value of z early , as long as it is not dramatically di ff erent.Non-parametric reconstructions of the ionization fractionhave also been proposed to probe the reionization history. Suchmethods are based on exploring reionization parameters in binsof redshift (Lewis et al. 2006). They should be particularly use-ful for investigating exotic reionization histories, e.g., doublereionization (Cen 2003). However, the CMB large-scale ( (cid:96) < ∼
10) polarization anisotropies are mainly sensitive to the over-all value of the optical depth, which determines the ampli-tude of the reionization bump in the EE power spectrum (seeFig. 3). We have estimated the impact on C EE (cid:96) for the two dif-ferent models (tanh and power law) having the same τ = . ff erences of less than 4 % for (cid:96) <
10. Even fora double reionization model, Fig. 4 shows that the impact on C EE (cid:96) is quite weak, given the actual measured value of τ , andcannot be distinguished relative to the cosmic variance spread For convenience, we refer to this parameterization as “redshiftsymmetric” in the rest of the paper, even although it is actually symmet-ric in y rather than z . The asymmetry is maximum in the instantaneouscase, but the di ff erence in x e values around, for example, z re = ±
1, isless than 1 %. 5lanck Collaboration: Planck constraints on reionization history
Redshift z x e Multipole ‘ -3 -2 -1 ‘ ( ‘ + ) C EE ‘ / π [ µ K ] Fig. 4.
Left : Evolution of the ionization fraction for several functions, all having the same optical depth, τ = .
06: green and blue arefor redshift-symmetric instantaneous ( δ z = .
05) and extended reionization ( δ z = . Right : corresponding EE power spectra with cosmic variance in grey. All models havethe same optical depth τ = .
06 and are essentially indistinguishable at the reionization bump scale.(i.e., even for a full-sky experiment). We also checked that
Planck data do not allow for model-independent reconstruc-tion of x e in redshift bins. Principal component analysis hasbeen proposed as an explicit approach to try to capture the de-tails of the reionization history in a small set of parameters(Hu & Holder 2003; Mortonson & Hu 2008). Although thesemethods are generally considered to be non-parametric, they arein fact based on a description of x e ( z ) in bins of redshift, ex-panded around a given fiducial model for C EE (cid:96) . Moreover, the po-tential bias on the τ measurement when analysing a more com-plex reionization history using a simple sharp transition model(Holder et al. 2003; Colombo & Pierpaoli 2009) is considerablyreduced for the (lower) τ values as suggested by the Planck re-sults. Consequently, we do not consider the non-parametric ap-proach further.
4. Measuring reionization observables
Reionization leaves imprints in the CMB power spectra, bothin polarization at very large scales and in intensity via the sup-pression of
T T power at higher (cid:96) . Reionization also a ff ects thekSZ e ff ect, due to the re-scattering of photons o ff newly liberatedelectrons. We sample from the space of possible parameters withMCMC exploration using CAMEL . This uses an adaptative-Metropolis algorithm to generate chains of samples for a set ofparameters. Thomson scattering between the CMB photons and free elec-trons generates linear polarization from the quadrupole momentof the CMB radiation field at the scattering epoch. This occursat recombination and also during the epoch of reionization. Re-scattering of the CMB photons at reionization generates an ad-ditional polarization anisotropy at large angular scales, becausethe horizon size at this epoch subtends a much larger angularsize. The multipole location of this additional anisotropy (essen- available at camel.in2p3.fr tially a bump) in the EE and T E angular power spectra relates tothe horizon size at the new “last-rescattering surface” and thusdepends on the redshift of reionization. The height of the bumpis a function of the optical depth or, in other words, of the historyof the reionization process. Such a signature (i.e., a polarizationbump at large scales) was first observed by WMAP, initially inthe
T E angular power spectrum (Kogut et al. 2003), and later incombination with all power spectra (Hinshaw et al. 2013).In Fig. 3 we show for the “instantaneous” reionization case(specifically the redshift-symmetric parameterization with δ z = .
5) power spectra for the E -mode polarization power spec-trum C EE (cid:96) and the temperature-polarization cross-power spec-trum C TE (cid:96) . The curves are computed with the CLASS
Boltzmannsolver (Lesgourgues 2011) using τ values ranging from 0.04 to0.08. For the range of optical depth considered here and giventhe amount of cosmic variance, the T E spectrum has only amarginal sensitivity to τ , while in EE the ability to distinguishdi ff erent values of τ is considerably stronger.In Fig. 4 (left panel), the evolution of the ionized fraction x e during the EoR is shown for five di ff erent parameterizationsof the reionization history, all yielding the same optical depth τ = .
06. Despite the di ff erences in the evolution of the ioniza-tion fraction, the associated C EE (cid:96) curves (Fig. 4, right panel) arealmost indistinguishable. This illustrates that while CMB large-scale anisotropies in polarization are only weakly sensitive to thedetails of the reionization history, they can nevertheless be usedto measure the reionization optical depth, which is directly re-lated to the amplitude of the low- (cid:96) bump in the E -mode powerspectrum.We use the Planck data to provide constraints on theThomson scattering optical depth for “instantaneous” reioniza-tion. Figure 5 shows the posterior distributions for τ obtainedwith the di ff erent data sets described in Sect. 2 and comparedto the 2015 PlanckTT + lowP results (Planck Collaboration XIII2016). We show the posterior distribution for the low- (cid:96) Planck polarized likelihood ( lollipop ) and in combination with thehigh- (cid:96)
Planck likelihood in temperature (PlanckTT). We alsoconsider the e ff ect of adding the SPT and ACT likelihoods (VHL) and the Planck lensing likelihood, as described inPlanck Collaboration XV (2016).The di ff erent data sets show compatible constraints on theoptical depth τ . The comparison between posteriors indicatesthat the optical depth measurement is driven by the low- (cid:96) like-lihood in polarization (i.e., lollipop ). The Planck constraintson τ for a Λ CDM model when considering the standard “instan-taneous” reionization assumption (symmetric model with fixed δ z = . τ = . + . − . , lollipop ; (4) τ = . + . − . , lollipop + PlanckTT ; (5) τ = . + . − . , lollipop + PlanckTT + lensing ; (6) τ = . + . − . , lollipop + PlanckTT + VHL . (7)We can see an improvement of the posterior width when addingtemperature anisotropy data to the lollipop likelihood. Thiscomes from the fact that the temperature anisotropies help to fixother Λ CDM parameters, in particular the normalization of theinitial power spectrum A s , and its spectral index, n s . CMB lens-ing also helps to reduce the degeneracy with A s , while gettingrid of the tension with the phenomenological lensing parameter A L when using PlanckTT only (see Planck Collaboration XIII2016), even if the impact on the error bars is small. Comparingthe posteriors in Fig. 6 with the constraints from PlanckTT alone(see figure 45 in Planck Collaboration XI 2016) shows that in-deed, the polarization likelihood is su ffi ciently powerful that itbreaks the degeneracy between n s and τ . The impact on other Λ CDM parameters is small, typically below 0 . σ (as shownmore explicitly in Appendix B). The largest changes are for τ and A s , where the lollipop likelihood dominates the con-straint. The parameter σ shifts towards slightly smaller val-ues by about 1 σ . This is in the right direction to help resolvesome of the tension with cluster abundances and weak galaxylensing measurements, discussed in Planck Collaboration XX(2014) and Planck Collaboration XIII (2016); however, sometension still remains.Combining with VHL data gives compatible results, withconsistent error bars. The slight shift toward lower τ value (by0 . σ ) is related to the fact that the PlanckTT likelihood alonepushes towards higher τ values (see Planck Collaboration XIII2016), while the addition of VHL data helps to some extent inreducing the tension on τ between high- (cid:96) and low- (cid:96) polarization.As mentioned earlier, astrophysics constraints from mea-surements of the Gunn-Peterson e ff ect provide strong evidencethat the IGM was highly ionized by a redshift of z (cid:39)
6. Thisplaces a lower limit on the optical depth (using Eq. 1), whichin the case of instantaneous reionization in the standard Λ CDMcosmology corresponds to τ = . The Thomson scattering of CMB photons o ff ionized elec-trons induces secondary anisotropies at di ff erent stages of thereionization process. In particular, we are interested here inthe e ff ect of photons scattering o ff electrons moving with bulkvelocity, which is called the “kinetic Sunyaev Zeldovich” orkSZ e ff ect. It is common to distinguish between the “homoge-neous” kSZ e ff ect, arising when the reionization is complete(e.g., Ostriker & Vishniac 1986), and “patchy” (or inhomoge-neous) reionization (e.g., Aghanim et al. 1996), which arises In this case only, other Λ CDM parameters are held fixed, including A s exp ( − τ ). Optical depth τ P / P m a x lowP+PlanckTTlowP+PlanckTT+lensinglollipoplollipop+PlanckTTlollipop+PlanckTT+lensinglollipop+PlanckTT+VHL Fig. 5.
Posterior distribution for τ from the various combinationsof Planck data. The grey band shows the lower limit on τ fromthe Gunn-Peterson e ff ect. ln(10 A s ) τ n s σ lowP + Planck TT Lollipop + Planck TT3.003.053.103.15 ln(10 A s ) τ n s σ lowP + Planck TT + lensing Lollipop + Planck TT + lensing Fig. 6.
Constraints on τ , A s , n s , and σ for the Λ CDM cosmol-ogy from PlanckTT, showing the impact of replacing the lowPlikelihood from
Planck lollipop likelihood. The top panels show results without lensing, whilethe bottom panels are with lensing.during the process of reionization, from the proper motion ofionized bubbles around emitting sources. These two compo-nents can be described by their power spectra, which can becomputed analytically or derived from numerical simulations. InPlanck Collaboration XI (2016), we used a kSZ template basedon homogeneous simulations, as described in Trac et al. (2011).In the following, we assume that the kSZ power spectrum isgiven by D kSZ (cid:96) = D h − kSZ (cid:96) + D p − kSZ (cid:96) , (8)where D (cid:96) = (cid:96) ( (cid:96) + C (cid:96) / π and the superscripts “h-kSZ” and “p-kSZ” stand for “homogeneous” and “patchy” reionization, re-spectively. For the homogeneous reionization, we use the kSZtemplate power spectrum given by Shaw et al. (2012) calibratedwith a simulation that includes the e ff ects of cooling and star-formation (which we label “CSF”). For the patchy reionizationkSZ e ff ect we use the fiducial model of Battaglia et al. (2013). Multipole ‘ D k S Z ‘ / D k S Z Homogeneous (ref)Homogeneous (CSF)PatchyCSF + patchy
Fig. 7.
Power spectrum templates for the kSZ e ff ect. The dif-ferent lines correspond to: homogeneous reionization as used inPlanck Collaboration XI (2016) (dark blue), based on Trac et al.(2011); “CSF” (light blue), which is a homogeneous reionizationmodel from Shaw et al. (2012); Patchy (green dashed) based onpatchy reionization model from Battaglia et al. (2013); and thesum of CSF and patchy (red).In the range (cid:96) = (cid:96) = (cid:96) = A kSZ ≡ D kSZ (cid:96) = : A kSZ ∝ (cid:32) h . (cid:33) . (cid:18) σ . (cid:19) . (cid:32) Ω b . (cid:33) . (cid:32) . Ω m (cid:33) . (cid:32) . n s (cid:33) . . (9)The amplitude of the kSZ power spectrum at (cid:96) = A kSZ is another observable of the reion-ization history that can be probed by CMB data. Its scalingswith the reionization redshift and the duration of the EoR canbe extracted from simulations. We assume for the patchy andhomogeneous kSZ e ff ect, the scalings of Battaglia et al. (2013)and Shaw et al. (2012), respectively. For the Planck base Λ CDMcosmology given in Sect. 2.2, we find (in µ K ): A hkSZ = . × (cid:18) τ . (cid:19) . ; (10) A pkSZ = . × (cid:34)(cid:32) + z re (cid:33) − . (cid:35) (cid:18) z
25 % − z
75 % . (cid:19) . . (11)For the measured value τ = . ± . A hkSZ = . µ K and A pkSZ = . µ K ,respectively.For the multipole range of Planck , the amplitude of thekSZ spectrum is dominated by other foregrounds, includingGalactic dust, point sources, CIB fluctuations, and the tSZ e ff ect.Moreover, the Planck signal-to-noise ratio decreases rapidlyabove (cid:96) = Planck data alone. Combining with additional data at higher multipoles helpsto disentangle the di ff erent foregrounds. We explicitly use theband powers from SPT (George et al. 2015) and ACT (Das et al.2014), covering the multipole range up to (cid:96) =
13 000.Despite our best e ff orts to account for the details, the kSZamplitude is weak and there are large uncertainties in the mod-els (both homogeneous and patchy). Moreover, there are cor-relations between the di ff erent foreground components, comingboth from the astrophysics (we use the same halo model to de-rive the power spectra for the CIB and for CIB × tSZ as the oneused for the kSZ e ff ect) and from the adjustments in the data.We carried out several tests to check the robustness of the con-straints on A kSZ with respect to the template used for the CIB,CIB × tSZ, and kSZ contributions. In particular, the CIB × tSZpower spectrum amplitude is strongly anti-correlated with thekSZ amplitude and poorly constrained by the CMB data. As aconsequence, if we neglect the CIB × tSZ contribution, the kSZamplitude measured in CMB data is substantially reduced, lead-ing to an upper limit much lower than the one derived when in-cluding the CIB × tSZ correlation. In the following discussion weconsider only the more realistic case (and thus more conservativein terms of constraints on A kSZ ) where the CIB × tSZ correlationcontributes to the high- (cid:96) signal. τ A k S Z Fig. 8.
68 % and 95 % confidence intervals on the reionizationoptical depth, τ , and the amplitude of the kinetic SZ e ff ect, A kSZ ,from the CMB ( lollipop + PlanckTT + VHL).We combine the
Planck likelihoods in
T T (PlanckTT) andfrom low- (cid:96) EE polarization ( lollipop ) with the very high- (cid:96) data from ACT and SPT (VHL), assuming a redshift-symmetricparameterization of the reionization. Figure 8 shows the 2D pos-terior distribution for τ and A kSZ after marginalization over theother cosmological and nuisance parameters.Figure 9 compares the constraints on the kSZ power at (cid:96) = A kSZ , obtained for three di ff erent kSZ templates: the“homogeneous” reionization template from Trac et al. (2011),which neglects contributions from inhomogeneous reionization;a more complex model “CSF & patchy,” including both homo-geneous and patchy contributions; and a pure “patchy” templatefrom Battaglia et al. (2013). We find very similar upper limits on A kSZ , even in the case of pure patchy reionization.Using the “CSF & patchy” model, the upper limit is A kSZ < . µ K (95 % CL) . (12) A kSZ P / P m a x Homogeneous (CSF)PatchyCSF + patchy
Fig. 9.
Constraints on the kSZ amplitude at (cid:96) = lollipop + PlanckTT + VHL likelihoods. The three cases corre-spond to di ff erent kSZ templates.Compared to Planck A kSZ = . + . − . µ K (PlanckTT + WP + highL,Planck Collaboration XVI 2014) is reduced to an upper limit inthis new analysis. The data presented here provide the best con-straint to date on the kSZ power and is a factor of 2 lower thanthe limit reported in George et al. (2015). Our limit is certainlynot in tension with the homogeneous kSZ template, which pre-dicts A kSZ = . µ K . However, it does not leave much roomfor any additional kSZ power coming from patchy reionization.Consistent with George et al. (2015), we find the total kSZpower to be stable against varying tSZ and CIB templates. Wealso find very little dependence on the choice of the kSZ tem-plate (Fig. 9). This confirms that there is only a modest amountof information in the angular shape of the kSZ signal with thecurrent data.
5. Constraints on the reionization history
We now interpret our measurements of the reionization observ-ables in terms of constraint on the reionization history. Wemainly focus on the determination of the reionization redshift z re and its duration ∆ z = z beg − z end . We show only the resultsfor ∆ z greater than unity, which corresponds to approximatively90 Myr at redshift z =
8. We first begin by looking at constraintson the EoR for symmetric and asymmetric models using
Planck data only ( lollipop + PlanckTT). Then we introduce the VHLdata and discuss additional constraints from the kSZ amplitude.In each case, we also derive the constraints that follow from pos-tulating that reionization should be completed at a redshift of 6(see Sect. 2.1), i.e., when imposing the prior z end > We use the
Planck
CMB likelihoods in temperature (PlanckTT)and polarization ( lollipop ) to derive constraints on Λ CDMparameters, including the reionization redshift z re and width ∆ z for a redshift-symmetric parameterization. Figure 10 shows (inblue) the posterior on z re and ∆ z after marginalization over theother cosmological and nuisance parameters. As discussed inSect. 3, the large-scale polarized CMB anisotropies are almost insensitive to the width δ z of the tanh function. We thus recoverthe degeneracy in the direction of ∆ z . Imposing an additionalGunn-Peterson constraint on the ionization fraction at very lowredshift can break this degeneracy. This is illustrated in Fig. 10,where we show (in green) the results of the same analysis withan additional prior z end >
6. In this case, we find δ z < . z beg − z end ) of ∆ z < . z re ∆ z ∆ z Fig. 10.
Posterior distributions (in blue) of z re and ∆ z for aredshift-symmetric parameterization using the CMB likelihoodsin polarization and temperature ( lollipop + PlanckTT). Thegreen contours and lines show the distribution after imposingthe additional prior z end > z re is shown in Fig. 10 aftermarginalizing over ∆ z , with and without the additional constraint z end >
6. This suggests that the reionization process occurred atredshift z re = . + . − . (uniform prior) , (14) z re = . + . − . (prior z end > . (15)This redshift is lower than the values derived previouslyfrom WMAP-9 data, in combination with ACT and SPT(Hinshaw et al. 2013), namely z re = . ± .
1. It isalso lower than the value z re = . ± . Planck
Planck + lowP alone, z re = . + . − . or in combina-tion with other data sets, z re = . + . − . (specifically forPlanckTT + lowP + lensing + BAO) estimated with δ z fixed to 0.5.The constraint from lollipop + PlanckTT when fixing δ z to 0.5is z re = . + . − . . This slightly lower value (compared to the oneobtained when letting the reionization width be free) is explained by the shape of the degeneracy surface. Allowing for larger du-ration when keeping the same value of τ pushes towards higherreionization redshifts; marginalizing over ∆ z thus shifts the pos-terior distribution to slightly larger z re values. z end P / P m a x z beg P / P m a x Fig. 11.
Posterior distributions on the end and beginning ofreionization, i.e., z end and z beg , using the redshift-symmetric pa-rameterization without (blue) and with (green) the prior z end > z re and δ z using the redshift-symmetric parameterization, the distributions of the end andbeginning of reionization, z end (i.e., z
99 % ) and z beg (i.e., z
10 % ),are plotted in Fig. 11. In such a model, the end of reionizationstrongly depends on the constraint at low redshift. On the otherhand, the constraints on z beg depend only slightly on the low-redshift prior. These results show that the Universe is ionized atless than the 10 % level above z = . ± . We now explore more complex reionization histories usingthe redshift-asymmetric parameterization of x e ( z ) described inSect. 3. In the same manner as in Sect. 5.1, also examine thee ff ect of imposing the additional constraint from the Gunn-Peterson e ff ect.The distributions of the two parameters, z end and z beg , areplotted in Fig. 12. With the redshift-asymmetric parameteriza-tion, we obtain z beg = . + . − . (imposing the prior on z end ),which disfavours any major contribution to the ionized fractionfrom sources that could form as early as z > ∼ z end P / P m a x z beg P / P m a x Fig. 12.
Posterior distributions of z end and z beg using the redshift-asymmetric parameterization without (blue) and with (green) theprior z end >
6. In Fig. 13, we interpret the results in terms of reionizationredshift and duration of the EoR, finding z re = . + . − . (uniform prior) , (16) z re = . + . − . (prior z end > . (17)These values are within 0 . σ of the results for the redshift-symmetric model. For the duration of the EoR, the upper limitson ∆ z are ∆ z < . , (18) ∆ z < . z end > . (19) z re ∆ z ∆ z Fig. 13.
Posterior distributions for z re and ∆ z using the redshift-asymmetric parameterization without (blue) and with (green) theprior z end > In order to try to obtain better constraints on the reionizationwidth, we now make use of the additional information comingfrom the amplitude of the kinetic SZ e ff ect. Since Planck aloneis not able to provide accurate limits on the kSZ amplitude, wecombine the
Planck likelihoods in temperature and polarizationwith the measurements of the CMB
T T power spectrum at high-resolution from the ACT and SPT experiments, “VHL.”Using the redshift-symmetric model, when adding the VHLdata, we recover essentially the same results as in Sect. 5.1. Thereionization redshift is slightly lower, as suggested by the resultson τ (see Eq. 7 and the discussion in Sect. 4.1). We also see thesame degeneracy along the ∆ z direction.With the addition of kSZ information, we are able to breakthe degeneracy with ∆ z . This might allow us to determine howmuch kSZ power originated during reionization (i.e., patchykSZ) and how much at later times, when the Universe becamefully ionized (i.e., homogeneous kSZ). We use the templatesfrom Shaw et al. (2012) and Battaglia et al. (2013) for the ho-mogeneous and patchy kSZ contributions, respectively, with the dependency on Λ CDM cosmological parameters as described inSect. 4.2. Those specific relations rely on a redshift-symmetricmodel for the description of the EoR. Note, however, that theresults presented here are derived from specific simulations ofthe reionization process, and so explicit scalings need to beassumed, as discussed by Zahn et al. (2012) and George et al.(2015).As described in Sect. 4.2, the amplitude of the kSZ powerprimarily depends on the duration of reionization, while theepoch is essentially constrained by the optical depth. Using the2D distribution for τ and A kSZ , as measured by Planck in combi-nation with very high- (cid:96) temperature data (Fig. 8), we derive a 2Dlikelihood function for z re and ∆ z . We can then sample the reion-ization parameters (the epoch z re and duration ∆ z of the EoR),compute the associated optical depth and kSZ power and deriveconstraints based on the 2D likelihood. The allowed models interms of z re and ∆ z are shown in Fig. 14 (in blue). We also plot(in green) the same constraints with the additional prior z end > z re ∆ z ∆ z Fig. 14.
Posterior distributions on the duration ∆ z and the red-shift z re of reionization from the combination of CMB polariza-tion and kSZ e ff ect constraints using the redshift-symmetric pa-rameterization without (blue) and with (green) the prior z end > ∆ z < . , (20) ∆ z < . z end > . (21)This is compatible with the constraints from George et al.(2015), where an upper limit was quoted of z
20 % − z
99 % < . . . z end , respectively.For the reionization redshift, we find z re = . + . − . (uniform prior) , (22) z re = . + . − . (prior z end > , (23) which is compatible within 1 σ with the results from CMB Planck data alone without the kSZ constraint (Sect. 5.1).The distributions of z end and z beg are plotted in Fig. 15.Within the redshift-symmetric parameterization, we obtain z beg = . + . − . (with the prior on z end ). z end P / P m a x z beg P / P m a x Fig. 15.
Posterior distributions of z end and z beg using the redshift-symmetric parameterization, combining Planck and VHL data,and using information from the kSZ amplitude, without (blue)and with (green) the prior z end > ∆ z and the beginning of reionization (corresponding to the10 % ionization limit) z beg . However, as discussed in Sect. 4.2,those results are very sensitive to details of the simulations usedto predict both the shape and the parameter dependences of thekSZ template in the di ff erent reionization scenarios (patchy orhomogeneous).
6. Discussion
The CMB has long held the promise of measuring the Thomsonoptical depth in order to derive constraints on the reionizationhistory of the Universe. Despite its importance, this constraint isfundamentally limited by cosmic variance in polarization and isfurther challenged by foregrounds and systematic e ff ects. Thefirst results, from WMAP, gave τ = . ± .
04, suggestinga reionization redshift between 11 and 30 (Kogut et al. 2003).This was revised in the final 9-year WMAP results to a cen-tral value of τ = .
084 (Hinshaw et al. 2013), which, in the in-stantaneous reionization model, implies z re = .
4. However,with the context of the same model, the
Planck z re = . + . − . )or in combination with other data sets ( z re = . + . − . ), showedthat the reionization redshift was smaller. The main result wepresent here, z re = . + . − . , further confirms that reionization oc-curred rather late, leaving little room for any significant ioniza-tion at z > ∼
15. This is consistent with what is suggested by otherreionization probes, which we now discuss (for reviews, see e.g.,Becker et al. 2015; McQuinn 2015).The transition from neutral to ionized gas is constrained byabsorption spectra of very distant quasars and gamma ray bursts(GRBs), revealing neutral hydrogen in intergalactic clouds. Theyshow, through the Gunn-Peterson e ff ect, that the di ff use gasin the Universe is mostly ionized up to a redshift of about6 (Fan et al. 2006a). Given the decline in their abundance be-yond redshift z (cid:39)
6, quasars and other active galactic nu-clei (AGN) cannot be major contributors to the early stagesof reionization (e.g., Willott et al. 2010; Fontanot et al. 2012, but see Madau & Haardt 2015; Khaire et al. 2016, for alterna-tive AGN-only models). A faint AGN population can producesignificant photoionization rates at redshifts of 4–6.5, consis-tent with the observed highly ionized IGM in the Ly- α forestof high- z quasar spectra (Giallongo et al. 2015). Star-forminggalaxies at redshifts z > ∼ z galaxies, it has been sug-gested that either the UV luminosity density flattens, or phys-ical parameters such as the escape fraction and the clumpingfactor evolved significantly, or alternatively, additional, unde-tected sources (such as X-ray binaries and faint AGN) musthave existed at z > ∼
11 (e.g., Kuhlen & Faucher-Gigu`ere 2012;Ellis et al. 2013; Cai et al. 2014; Ishigaki et al. 2015).The
Planck results, both from the 2015 data release andthose presented here, strongly reduce the need for a signifi-cant contribution of Lyman continuum emission at early times.Indeed, as shown in Fig. 16, the present CMB results on theThomson optical depth, τ = . ± . τ , if we maintain a UV-luminosity density at the maxi-mum level allowed by the luminosity density constraints at red-shifts z <
9, then the currently observed galaxy population at M UV < −
17 seems to be su ffi cient to comply with all the obser-vational constraints without the need for high-redshift ( z = Planck data are certainly consistent with a fully reion-ized Universe at z (cid:39)
6. Moreover, they seem to be in good agree-ment with recent observational constraints on reionization in thedirection of particular objects. The H i absorption along the lineof sight to a distant γ -ray burst, GRB-140515A (Chornock et al.2014), suggests a Universe containing about a 10 % fractionof neutral hydrogen at z = z (cid:39)
7, observation of Ly- α emitters suggests that at least70 % of the IGM is neutral (Tilvi et al. 2014; Schenker et al.2014; Faisst et al. 2014). Similarly, quasar near-zone detectionand analysis (including sizes, and Ly- α and β transmissionproperties) have been used to place constraints on z end fromsignatures of the ionization state of the IGM around individ-ual sources (Wyithe & Loeb 2004; Mesinger & Haiman 2004;Wyithe et al. 2005; Mesinger & Haiman 2007; Carilli et al.2010; Mortlock et al. 2011; Schroeder et al. 2013). However, in-terpretation of the observed evolution of the near-zone sizesmay be complicated by the opacity caused by absorptionsystems within the ionized IGM (e.g., Bolton et al. 2011;Bolton & Haehnelt 2013; Becker et al. 2015). Similarly, it is dif-ficult to completely exclude the possibility that damped Ly- α systems contribute to the damping wings of quasar spectrablueward of the Ly- α line (e.g., Mesinger & Furlanetto 2008;Schroeder et al. 2013). Nevertheless, most such studies, indicatethat the IGM is significantly neutral at redshifts between 6 and z τ Fig. 16.
Evolution of the integrated optical depth for the tanhfunctional form (with δ z = .
5, blue shaded area). Thetwo envelopes mark the 68 % and 95 % confidence inter-vals. The red, black, and orange dashed lines are the mod-els from Bouwens et al. (2015), Robertson et al. (2015), andIshigaki et al. (2015), respectively, using high-redshift galaxyUV and IR fluxes and / or direct measurements.7 (see also Keating et al. 2015), in agreement with the current Planck results, as shown in Fig. 17. z -5 -4 -3 -2 − Q H II Fig. 17.
Reionization history for the redshift-symmetric parame-terization compared with other observational constraints comingfrom quasars, Ly- α emitters, and the Ly- α forest (compiled byBouwens et al. 2015). The red points are measurements of ion-ized fraction, while black arrows mark upper and lower limits.The dark and light blue shaded areas show the 68 % and 95 %allowed intervals, respectively.Although there are already all the constraints describedabove, understanding the formation of the first luminous sourcesin the Universe is still very much a work in progress. Our new(and lower) value of the optical depth leads to better agreementbetween the CMB and other astrophysical probes of reioniza-tion; however, the fundamental questions remain regarding howreionization actually proceeded. z x e z z Fig. 18.
Constraints on ionization fraction during reionization. The allowed models, in terms of z re and ∆ z , translate into an allowedregion in x e ( z ) (68 % and 95 % in dark blue and light blue, respectively), including the z end > Left : Constraints fromCMB data using a redshift-symmetric function ( x e ( z ) as a hyperbolic tangent with δ z = . Centre : Constraints from CMB datausing a redshift-asymmetric parameterization ( x e ( z ) as a power law). Right : Constraints from CMB data using a redshift-symmetricparameterization with additional constraints from the kSZ e ff ect.
7. Conclusions
We have derived constraints on cosmic reionization using
Planck data. The CMB
Planck power spectra, combining the EE polar-ization at low- (cid:96) with the temperature data, give, for a so-called“instantaneous” reionization history (a redshift-symmetric tanhfunction x e ( z ) with δ z = . τ = . ± .
012 ( lollipop + PlanckTT), (24)which is significantly more accurate than previous measure-ments. Thanks to the relatively high signal-to-noise ratio of thelow- (cid:96) polarization signal, the combination with lensing or datafrom high-resolution CMB anisotropy experiments (ACT andSPT) does not bring much additional constraining power. Theimpact on other Λ CDM parameters is only significant for theamplitude of the initial scalar power spectrum A s and (to a lesserextent) on its tilt n s . Other parameters are very stable comparedto the Planck
Planck data, we have derived constraints on two mod-els for the reionization history x e ( z ) that are commonly used inthe literature: a redshift-symmetric form using a hyperbolic tan-gent transition function; and a redshift-asymmetric form param-eterized by a power law. We have also investigated the e ff ectof imposing the condition that the reionization is completed by z = z re than in the case of instantaneous reioniza-tion. This can be understood through the shape of the degeneracysurface between the reionization parameters. For an asymmetricparameterization, z re is smaller, due to the fact that x e ( z ) changesmore rapidly at the end of reionization than the beginning. Wespecifically find: z re = . ± . , (25) z re = . ± . ff erent parameterizations of the reionizationhistory shows how much results on e ff ective parameters (likethe redshift of reionization or its duration) are sensitive to theassumption of the reionization history shape. The best models ofsymmetric and asymmetric parameterization give similar valuesfor τ , and provide reionization redshifts which di ff er by less than0 . σ . Constraints on the limits of possible early reionization aresimilar, leading to 10 % reionization levels at around z = ff ect. In the case of a redshift-symmetric model,we found ∆ z < . , (27)using the additional constraint that the Universe is entirely reion-ized at redshift 6 (i.e., z end > + kSZ (all with prior z end > τ as measurednow by Planck , the CMB is not able to give tight constraintson details of the reionization history. However, the
Planck datasuggest that an early onset of reionization is disfavoured. In par-ticular, in all cases, we found that the Universe was less than10 % ionized for redshift z >
10. Furthermore, comparisons withother tracers of the ionization history show that our new result onthe optical depth eliminates most of the tension between CMB-
Table 1.
Constraints on reionization parameters for the di ff erentmodels presented in this paper when including the z end > z re and z beg , while we quote 95% upperlimit for ∆ z and z end . model z re ∆ z z end z beg redshift-symmetric . . 8 . ± . < . < . . ± . . ± . < . < . . ± . . ± . < . < . . ± . based analyses and constraints from other astrophysical data.Additional sources of reionization, non-standard early galaxies,or significantly evolving escape fractions or clumping factors,are thus not needed.Ongoing and future experiments like LOFAR, MWA, andSKA, aimed at measuring the redshifted 21-cm signal from neu-tral hydrogen during the EoR, should be able to probe reioniza-tion directly and measure its redshift and duration to high ac-curacy. Moreover, since reionization appears to happen at red-shifts below 10, experiments measuring the global emission ofthe 21-m line over the sky (e.g., EDGES, Bowman & Rogers2010, LEDA, Greenhill & Bernardi 2012, DARE, Burns et al.2012), NenuFAR, Zarka et al. 2012, SARAS, Patra et al. 2013,SCI-HI, Voytek et al. 2014, ZEBRA, Mahesh et al. 2014, andBIGHORNS, Sokolowski et al. 2015) will also be able to derivevery competitive constraints on the models (e.g., Liu et al. 2016;Fialkov & Loeb 2016). Acknowledgements.
The Planck Collaboration acknowledges the support of:ESA; CNES, and CNRS / INSU-IN2P3-INP (France); ASI, CNR, and INAF(Italy); NASA and DoE (USA); STFC and UKSA (UK); CSIC, MINECO,JA, and RES (Spain); Tekes, AoF, and CSC (Finland); DLR and MPG(Germany); CSA (Canada); DTU Space (Denmark); SER / SSO (Switzerland);RCN (Norway); SFI (Ireland); FCT / MCTES (Portugal); ERC and PRACE (EU).A description of the Planck Collaboration and a list of its members, indicatingwhich technical or scientific activities they have been involved in, can be foundat http: // / web / planck / planck-collaboration. References
Aghanim, N., Desert, F. X., Puget, J. L., & Gispert, R., Ionization by earlyquasars and cosmic microwave background anisotropies. 1996, A&A, 311,1, arXiv:astro-ph / / = / α emitters at redshift z (cid:38) = + ∆ z > .
06 for the durationof the reionization epoch. 2010, Nature, 468, 796Burns, J. O., Lazio, J., Bale, S., et al., Probing the first stars and black holes in theearly Universe with the Dark Ages Radio Explorer (DARE). 2012, Advancesin Space Research, 49, 433, arXiv:1106.5194Cai, Z.-Y., Lapi, A., Bressan, A., et al., A Physical Model for theEvolving Ultraviolet Luminosity Function of High Redshift Galaxies andtheir Contribution to the Cosmic Reionization. 2014, ApJ, 785, 65,arXiv:1403.0055 Calabrese, E., Hlozek, R. A., Battaglia, N., et al., Cosmological parame-ters from pre-planck cosmic microwave background measurements. 2013,Phys. Rev. D, 87, 103012, arXiv:1302.1841Carilli, C. L., Wang, R., Fan, X., et al., Ionization Near Zones Associated withQuasars at z ˜ 6. 2010, ApJ, 714, 834, arXiv:1003.0016Cen, R., The Universe Was Reionized Twice. 2003, ApJ, 591, 12, arXiv:astro-ph / = / α Emitters at z ˜ 7.7 and Implications onRe-ionization. 2014, ApJ, 788, 87Fan, X., Strauss, M. A., Becker, R. H., et al., Constraining the Evolution of theIonizing Background and the Epoch of Reionization with z˜6 Quasars. II. ASample of 19 Quasars. 2006a, AJ, 132, 117, arXiv:astro-ph / > / / / / / Functions and Cosmic Reionization. 2015, ApJ, 799, 12, arXiv:1408.6903Keating, L. C., Haehnelt, M. G., Cantalupo, S., & Puchwein, E., Probing the endof reionization with the near zones of z (cid:38) / / / ff ect from Reionization. 2005, ApJ, 630, 643,arXiv:astro-ph / α damping wing constraints on inhomoge-neous reionization. 2008, MNRAS, 385, 1348, arXiv:0710.0371Mesinger, A. & Haiman, Z., Evidence of a Cosmological Str¨omgren Surface andof Significant Neutral Hydrogen Surrounding the Quasar SDSS J1030 + / z > . / = / ff ects on cosmological parameters. 2011,Phys. Rev. D, 84, 123522, arXiv:1111.3570Park, H., Shapiro, P. R., Komatsu, E., et al., The Kinetic Sunyaev-Zel’dovichE ff ect as a Probe of the Physics of Cosmic Reionization: The E ff ect of Self- regulated Reionization. 2013, ApJ, 769, 93, arXiv:1301.3607Patra, N., Subrahmanyan, R., Raghunathan, A., & Udaya Shankar, N., SARAS:a precision system for measurement of the cosmic radio background and sig-natures from the epoch of reionization. 2013, Experimental Astronomy, 36,319, arXiv:1211.3800Peebles, P. J. E., Recombination of the Primeval Plasma. 1968, ApJ, 153, 1Planck Collaboration XV, Planck
Planck
Planck
Planck
Planck
Planck
Planck ff use component separation:CMB maps. 2016, A&A, in press, arXiv:1502.05956Planck Collaboration X, Planck ff use component separation:Foreground maps. 2016, A&A, in press, arXiv:1502.01588Planck Collaboration XI, Planck
Planck
Planck
Planck intermediate results. XLVI. Reductionof large-scale systematic e ff ects in HFI polarization maps and estimation ofthe reionization optical depth. 2016, A&A, submitted, arXiv:1605.02985Pritchard, J. R., Loeb, A., & Wyithe, J. S. B., Constraining reionization using21-cm observations in combination with CMB and Ly α forest data. 2010,MNRAS, 408, 57, arXiv:0908.3891Reichardt, C. L., Shaw, L., Zahn, O., et al., A Measurement of SecondaryCosmic Microwave Background Anisotropies with Two Years of South PoleTelescope Observations. 2012, ApJ, 755, 70, arXiv:1111.0932Robertson, B. E., Ellis, R. S., Dunlop, J. S., McLure, R. J., & Stark, D. P., Earlystar-forming galaxies and the reionization of the Universe. 2010, Nature, 468,49, arXiv:1011.0727Robertson, B. E., Ellis, R. S., Furlanetto, S. R., & Dunlop, J. S., CosmicReionization and Early Star-forming Galaxies: A Joint Analysis of NewConstraints from Planck and the Hubble Space Telescope. 2015, ApJ, 802,L19, arXiv:1502.02024Robertson, B. E., Furlanetto, S. R., Schneider, E., et al., New Constraints onCosmic Reionization from the 2012 Hubble Ultra Deep Field Campaign.2013, ApJ, 768, 71, arXiv:1301.1228Schenker, M. A., Ellis, R. S., Konidaris, N. P., & Stark, D. P., Line-emittingGalaxies beyond a Redshift of 7: An Improved Method for Estimatingthe Evolving Neutrality of the Intergalactic Medium. 2014, ApJ, 795, 20,arXiv:1404.4632Schroeder, J., Mesinger, A., & Haiman, Z., Evidence of Gunn-Peterson dampingwings in high-z quasar spectra: strengthening the case for incomplete reion-ization at z ∼ / α Emissiontoward the Reionization Era. 2014, ApJ, 794, 5, arXiv:1405.4869Trac, H., Bode, P., & Ostriker, J. P., Templates for the Sunyaev-Zel’dovichAngular Power Spectrum. 2011, ApJ, 727, 94, arXiv:1006.2828Tristram, M., Mac´ıas-P´erez, J. F., Renault, C., & Santos, D., XSPECT, estima-tion of the angular power spectrum by computing cross-power spectra withanalytical error bars. 2005, MNRAS, 358, 833, arXiv:astro-ph / > Experiment. 2014, The Astrophysical Journal Letters, 782, L9Willott, C. J., Delorme, P., Reyl´e, C., et al., The Canada-France High-z QuasarSurvey: Nine New Quasars and the Luminosity Function at Redshift 6. 2010,AJ, 139, 906, arXiv:0912.0281Worseck, G., Prochaska, J. X., Hennawi, J. F., & McQuinn, M., Early andExtended Helium Reionization Over More Than 600 Million Years of CosmicTime. 2014, ArXiv e-prints, arXiv:1405.7405Wyithe, J. S. B. & Loeb, A., A characteristic size of ˜10Mpc for the ionized bub-bles at the end of cosmic reionization. 2004, Nature, 432, 194, arXiv:astro-ph / z >
6. 2005, ApJ,628, 575, arXiv:astro-ph / Appendix A: the Lollipop likelihood lollipop , the LOw- (cid:96)
LIkelihood on POlarized Power-spectra,is a likelihood function based on cross-power spectra for the lowmultipoles. The idea behind this approach is that the noise can beconsidered as uncorrelated between maps and that systematicswill be considerably reduced in cross-correlation compared toauto-correlation.At low multipoles and for incomplete sky coverage, the C (cid:96) s are not Gaussian distributed and are correlated be-tween multipoles. lollipop uses the approximation pre-sented in Hamimeche & Lewis (2008), modified as described inMangilli et al. (2015) to apply to cross-power spectra. The ideais to apply a change of variable C (cid:96) → X (cid:96) so that the new vari-able X (cid:96) is Gaussian. Similarly to Hamimeche & Lewis (2008),we define X (cid:96) = (cid:113) C f (cid:96) + O (cid:96) g (cid:101) C (cid:96) + O (cid:96) C (cid:96) + O (cid:96) (cid:113) C f (cid:96) + O (cid:96) , (A.1)where g ( x ) = √ x − ln( x ) − (cid:101) C (cid:96) are the measured cross-power spectra, C (cid:96) are the power-spectra of the model to evaluate, C f (cid:96) is a fiducial model, and O (cid:96) are the o ff sets needed in the caseof cross-spectra. For multi-dimensional CMB modes (i.e., T , E ,and B ), C (cid:96) is a 3 × C (cid:96) = C TT C TE C T B C ET C EE C EB C BT C BE C BB (cid:96) , (A.2)and the g function is applied to the eigenvalues of C − / (cid:96) (cid:101) C (cid:96) C − / (cid:96) .In the case of auto-spectra, the o ff sets are replaced by thenoise bias e ff ectively present in the measured power-spectra. Forcross-power spectra, the noise bias is null and here we use thee ff ective o ff sets defined from the C (cid:96) noise variance: ∆ C (cid:96) ≡ (cid:114) (cid:96) + O (cid:96) . (A.3)The distribution of the new variable X can be approximatedas Gaussian, with a covariance given by the covariance of the C (cid:96) s. The likelihood function of the C (cid:96) given the data (cid:101) C (cid:96) is then − P ( C (cid:96) | (cid:101) C (cid:96) ) = (cid:88) (cid:96)(cid:96) (cid:48) X T (cid:96) M − (cid:96)(cid:96) (cid:48) X (cid:96) (cid:48) , (A.4) where the C (cid:96) covariance matrix M (cid:96)(cid:96) (cid:48) is estimated via MonteCarlo simulations.In this paper, we restrict ourselves to the one-field approx-imation in order to derive a likelihood function based onlyon the EE power spectrum at very low multipoles. We usea conservative sky fraction including 50 % of the sky, with aGalactic mask based on a threshold on the polarisation ampli-tude measured in the 353 GHz Planck channel, further apodizedusing a 4 ◦ Gaussian taper (see Fig. A.1). We use
Xpol (apseudo- C (cid:96) estimator described in Tristram et al. 2005 extendedto polarisation) to derive cross-power spectra between the 100and 143 GHz channel maps from Planck . We also reject thefirst two multipoles ( (cid:96) = ff ects (seePlanck Collaboration Int. XLVI 2016). Fig. A.1.
Galactic mask used for the lollipop likelihood, cov-ering 50 % of the sky.This likelihood has been tested on Monte Carlo simulationsincluding signal (CMB and foregrounds), realistic noise, andsystematic e ff ects. The simulated maps are then foreground-subtracted, using the same procedure as for the data. We con-structed the C (cid:96) covariance matrix M (cid:96)(cid:96) (cid:48) using those simulations.Figure A.2 shows the distribution of the recovered τ values foran input model with τ = .
06, fixing all other cosmological pa-rameters to the
Planck A s e − τ ).To validate the choice of multipole and the stability of the re-sult on τ , we performed several consistency checks on the Planck data. Among them, we varied the minimum multipole used (from (cid:96) = (cid:96) =
4) and allowed for larger sky coverage (increasingto 60 % of the sky). The results are summarized in Fig. A.3.
Appendix B: Impact on Λ CDM parameters
In addition to the restricted parameter set shown in Fig. 6,we describe here the impact of the lollipop likelihood on Λ CDM parameters in general. Figure B.1 compares results from lollipop + PlanckTT with the lowP + PlanckTT 2015. The newlow- (cid:96) polarization results are su ffi ciently powerful that theybreak the degeneracy between n s and τ . The contours for τ and A s , where the lollipop likelihood dominates the constraint, aresignificantly reduced. The impact on other Λ CDM parametersare small, typically below 0 . σ . APC, AstroParticule et Cosmologie, Universit´e Paris Diderot,CNRS / IN2P3, CEA / lrfu, Observatoire de Paris, Sorbonne ParisCit´e, 10, rue Alice Domon et L´eonie Duquet, 75205 Paris Cedex13, France16lanck Collaboration: Planck constraints on reionization history n s Ω c h θ M C τ l n ( A s ) Ω b h n s Ω c h θ MC τ ln(10 A s ) lowP + Planck TTLollipop + Planck TT Fig. B.1. Λ CDM parameters for PlanckTT combined with the low- (cid:96) polarization likelihood from the
Planck lollipop , in red). Aalto University Mets¨ahovi Radio Observatory and Dept of RadioScience and Engineering, P.O. Box 13000, FI-00076 AALTO,Finland African Institute for Mathematical Sciences, 6-8 Melrose Road,Muizenberg, Cape Town, South Africa Agenzia Spaziale Italiana Science Data Center, Via del Politecnicosnc, 00133, Roma, Italy Aix Marseille Universit´e, CNRS, LAM (Laboratoired’Astrophysique de Marseille) UMR 7326, 13388, Marseille,France Aix Marseille Universit´e, Centre de Physique Th´eorique, 163Avenue de Luminy, 13288, Marseille, France Astrophysics Group, Cavendish Laboratory, University ofCambridge, J J Thomson Avenue, Cambridge CB3 0HE, U.K. Astrophysics & Cosmology Research Unit, School of Mathematics,Statistics & Computer Science, University of KwaZulu-Natal,Westville Campus, Private Bag X54001, Durban 4000, South Africa CITA, University of Toronto, 60 St. George St., Toronto, ON M5S3H8, Canada CNRS, IRAP, 9 Av. colonel Roche, BP 44346, F-31028 Toulousecedex 4, France California Institute of Technology, Pasadena, California, U.S.A. Computational Cosmology Center, Lawrence Berkeley NationalLaboratory, Berkeley, California, U.S.A. DTU Space, National Space Institute, Technical University ofDenmark, Elektrovej 327, DK-2800 Kgs. Lyngby, Denmark D´epartement de Physique Th´eorique, Universit´e de Gen`eve, 24,Quai E. Ansermet,1211 Gen`eve 4, Switzerland 17lanck Collaboration: Planck constraints on reionization history τ r e a li z a t i o n s Fig. A.2.
Distribution of the peak value of the posterior distri-bution for optical depth from end-to-end simulations includingnoise, systematic e ff ects, Galactic dust signal, and CMB with thefiducial value of τ = . τ P / P m a x ‘ min =2 ‘ min =3 ‘ min =4 τ P / P m a x f sky = 70% f sky = 60% f sky = 50% Fig. A.3.
Posterior distributions for optical depth, showing thee ff ect of changing two of the choices made in our analysis. Top : Di ff erent choices of minimum multipole. Bottom : Di ff erentchoices of sky fraction used. Departamento de Astrof´ısica, Universidad de La Laguna (ULL), E-38206 La Laguna, Tenerife, Spain Departamento de F´ısica, Universidad de Oviedo, Avda. Calvo Sotelos / n, Oviedo, Spain Department of Astrophysics / IMAPP, Radboud UniversityNijmegen, P.O. Box 9010, 6500 GL Nijmegen, The Netherlands Department of Physics & Astronomy, University of BritishColumbia, 6224 Agricultural Road, Vancouver, British Columbia,Canada Department of Physics and Astronomy, Dana and David DornsifeCollege of Letter, Arts and Sciences, University of SouthernCalifornia, Los Angeles, CA 90089, U.S.A. Department of Physics and Astronomy, University College London,London WC1E 6BT, U.K. Department of Physics and Astronomy, University of Sussex,Brighton BN1 9QH, U.K. Department of Physics, Gustaf H¨allstr¨omin katu 2a, University ofHelsinki, Helsinki, Finland Department of Physics, Princeton University, Princeton, New Jersey,U.S.A. Department of Physics, University of California, Berkeley,California, U.S.A. Department of Physics, University of California, One ShieldsAvenue, Davis, California, U.S.A. Department of Physics, University of California, Santa Barbara,California, U.S.A. Department of Physics, University of Illinois at Urbana-Champaign,1110 West Green Street, Urbana, Illinois, U.S.A. Dipartimento di Fisica e Astronomia G. Galilei, Universit`a degliStudi di Padova, via Marzolo 8, 35131 Padova, Italy Dipartimento di Fisica e Astronomia, Alma Mater Studiorum,Universit`a degli Studi di Bologna, Viale Berti Pichat 6 /
2, I-40127,Bologna, Italy Dipartimento di Fisica e Scienze della Terra, Universit`a di Ferrara,Via Saragat 1, 44122 Ferrara, Italy Dipartimento di Fisica, Universit`a La Sapienza, P. le A. Moro 2,Roma, Italy Dipartimento di Fisica, Universit`a degli Studi di Milano, ViaCeloria, 16, Milano, Italy Dipartimento di Fisica, Universit`a di Roma Tor Vergata, Via dellaRicerca Scientifica, 1, Roma, Italy Dipartimento di Matematica, Universit`a di Roma Tor Vergata, Viadella Ricerca Scientifica, 1, Roma, Italy Discovery Center, Niels Bohr Institute, Copenhagen University,Blegdamsvej 17, Copenhagen, Denmark European Space Agency, ESAC, Planck Science O ffi ce, Caminobajo del Castillo, s / n, Urbanizaci´on Villafranca del Castillo,Villanueva de la Ca˜nada, Madrid, Spain European Space Agency, ESTEC, Keplerlaan 1, 2201 AZNoordwijk, The Netherlands Gran Sasso Science Institute, INFN, viale F. Crispi 7, 67100L’Aquila, Italy HGSFP and University of Heidelberg, Theoretical PhysicsDepartment, Philosophenweg 16, 69120, Heidelberg, Germany Haverford College Astronomy Department, 370 Lancaster Avenue,Haverford, Pennsylvania, U.S.A. Helsinki Institute of Physics, Gustaf H¨allstr¨omin katu 2, Universityof Helsinki, Helsinki, Finland INAF - Osservatorio Astronomico di Padova, Vicolodell’Osservatorio 5, Padova, Italy INAF - Osservatorio Astronomico di Roma, via di Frascati 33,Monte Porzio Catone, Italy INAF - Osservatorio Astronomico di Trieste, Via G.B. Tiepolo 11,Trieste, Italy INAF / IASF Bologna, Via Gobetti 101, Bologna, Italy INAF / IASF Milano, Via E. Bassini 15, Milano, Italy INFN - CNAF, viale Berti Pichat 6 /
2, 40127 Bologna, Italy INFN, Sezione di Bologna, viale Berti Pichat 6 /
2, 40127 Bologna,Italy INFN, Sezione di Ferrara, Via Saragat 1, 44122 Ferrara, Italy18lanck Collaboration: Planck constraints on reionization history INFN, Sezione di Roma 1, Universit`a di Roma Sapienza, PiazzaleAldo Moro 2, 00185, Roma, Italy INFN, Sezione di Roma 2, Universit`a di Roma Tor Vergata, Via dellaRicerca Scientifica, 1, Roma, Italy Imperial College London, Astrophysics group, Blackett Laboratory,Prince Consort Road, London, SW7 2AZ, U.K. Institut d’Astrophysique Spatiale, CNRS, Univ. Paris-Sud,Universit´e Paris-Saclay, Bˆat. 121, 91405 Orsay cedex, France Institut d’Astrophysique de Paris, CNRS (UMR7095), 98 bisBoulevard Arago, F-75014, Paris, France Institute of Astronomy, University of Cambridge, Madingley Road,Cambridge CB3 0HA, U.K. Institute of Theoretical Astrophysics, University of Oslo, Blindern,Oslo, Norway Instituto de Astrof´ısica de Canarias, C / V´ıa L´actea s / n, La Laguna,Tenerife, Spain Instituto de F´ısica de Cantabria (CSIC-Universidad de Cantabria),Avda. de los Castros s / n, Santander, Spain Istituto Nazionale di Fisica Nucleare, Sezione di Padova, viaMarzolo 8, I-35131 Padova, Italy Jet Propulsion Laboratory, California Institute of Technology, 4800Oak Grove Drive, Pasadena, California, U.S.A. Jodrell Bank Centre for Astrophysics, Alan Turing Building, Schoolof Physics and Astronomy, The University of Manchester, OxfordRoad, Manchester, M13 9PL, U.K. Kavli Institute for Cosmological Physics, University of Chicago,Chicago, IL 60637, USA Kavli Institute for Cosmology Cambridge, Madingley Road,Cambridge, CB3 0HA, U.K. LAL, Universit´e Paris-Sud, CNRS / IN2P3, Orsay, France LERMA, CNRS, Observatoire de Paris, 61 Avenue del’Observatoire, Paris, France Laboratoire Traitement et Communication de l’Information, CNRS(UMR 5141) and T´el´ecom ParisTech, 46 rue Barrault F-75634 ParisCedex 13, France Laboratoire de Physique Subatomique et Cosmologie, Universit´eGrenoble-Alpes, CNRS / IN2P3, 53, rue des Martyrs, 38026Grenoble Cedex, France Laboratoire de Physique Th´eorique, Universit´e Paris-Sud 11 &CNRS, Bˆatiment 210, 91405 Orsay, France Lawrence Berkeley National Laboratory, Berkeley, California,U.S.A. Max-Planck-Institut f¨ur Astrophysik, Karl-Schwarzschild-Str. 1,85741 Garching, Germany Mullard Space Science Laboratory, University College London,Surrey RH5 6NT, U.K. Nicolaus Copernicus Astronomical Center, Bartycka 18, 00-716Warsaw, Poland Niels Bohr Institute, Copenhagen University, Blegdamsvej 17,Copenhagen, Denmark Nordita (Nordic Institute for Theoretical Physics),Roslagstullsbacken 23, SE-106 91 Stockholm, Sweden SISSA, Astrophysics Sector, via Bonomea 265, 34136, Trieste, Italy School of Chemistry and Physics, University of KwaZulu-Natal,Westville Campus, Private Bag X54001, Durban, 4000, SouthAfrica School of Physics and Astronomy, University of Nottingham,Nottingham NG7 2RD, U.K. Simon Fraser University, Department of Physics, 8888 UniversityDrive, Burnaby BC, Canada Sorbonne Universit´e-UPMC, UMR7095, Institut d’Astrophysiquede Paris, 98 bis Boulevard Arago, F-75014, Paris, France Space Research Institute (IKI), Russian Academy of Sciences,Profsoyuznaya Str, 84 /
32, Moscow, 117997, Russia Space Sciences Laboratory, University of California, Berkeley,California, U.S.A. Sub-Department of Astrophysics, University of Oxford, KebleRoad, Oxford OX1 3RH, U.K. The Oskar Klein Centre for Cosmoparticle Physics, Department ofPhysics,Stockholm University, AlbaNova, SE-106 91 Stockholm,Sweden UPMC Univ Paris 06, UMR7095, 98 bis Boulevard Arago, F-75014,Paris, France Universit´e de Toulouse, UPS-OMP, IRAP, F-31028 Toulouse cedex4, France University of Granada, Departamento de F´ısica Te´orica y delCosmos, Facultad de Ciencias, Granada, Spain87