The Planck Satellite and the Cosmic Microwave Background

Home , Fritz Zwicky, Planck (spacecraft)

The Cosmic Microwave Background, Dark Matter

and Dark Energy Anthony Lasenby, Astrophysics Group, Cavendish Laboratory and Kavli Institute for Cosmology, Cambridge Overview

The Cosmic Microwave Background — exciting new results from the Planck Satellite Context of the CMB =⇒ addressing key questions about the Big Bang and the Universe, including Dark Matter and Dark Energy Planck Satellite and planning for its observations have been a long time in preparation — ﬁrst meetings in 1993! UK has been intimately involved Two instruments — the LFI (Low — e.g. Cambridge is the Frequency Instrument) and the HFI scientiﬁc data processing (High Frequency Instrument) centre for the HFI — RAL provided the 4K Cooler The Cosmic Microwave Background (CMB)

So what is the CMB? Anywhere in empty space at the moment there is radiation present corresponding to what a blackbody would emit at a temperature of ∼ 2.74 K (‘Blackbody’ being a perfect emitter/absorber — furnace with a small opening is a good example - needs perfect thermodynamic equilibrium) CMB spectrum is incredibly accurately black body — best known in nature! COBE result on this showed CMB better than its own reference b.b. within about 9 minutes of data! Universe History

Radiation was emitted in the early universe (hot, dense conditions) Hot means matter was ionised Therefore photons scattered frequently off the free electrons As universe expands it cools — eventually not enough energy to keep the protons and electrons apart — they History of the Universe: superluminal inflation, particle plasma, ‘recombine’ to form atoms of Hydrogen atomic plasma, recombination, Suddenly the photons are able to structure formation free-stream away crossing the entire universe without interruption Universe History

History of the Universe: superluminal inflation, particle plasma, atomic plasma, recombination, structure formation Universe History

Can see directly today the imprints present at recombination Good evidence these were created by amplification of quantum-generated irregularities during period of inflation, −35 s History of the Universe: taking place about 10 after the superluminal inflation, particle plasma, atomic plasma, Big-Bang! recombination, structure formation http://www.sdss3.org/surveys/boss.php (Chris Blake and Sam Moorfield)

Development of these (initially quantum) ﬂuctuations from inﬂation until recombination, imprints characteristic scales on the universe (basically how far ‘sound’ could have travelled by then) — should see this in both matter and CMB The Power Spectrum

Look at power in fluctuations as a function of angular scale on the sky Shown is a theoretical curve Series of coherent peaks is crucial — if can observe them, then fluctuations must have been ‘phased up’ Inflation is only known mechanism for achieving this! Details of peak location and height depend on the cosmological parameters such as density and age of the Universe The ‘Omegas’ refer to densities in various components, and H0 is Hubble’s constant (linked to age) The Power Spectrum

THETWOFURTHERINGREDIENTS:

We know there are big problems with understanding the dynamics of galaxies and clusters of galaxies There appears to be a large amount of ‘missing mass’ — i.e. inferred dynamically, but not visible Very obvious in the ‘rotation curves’ of galaxies From mv 2 GMm Instead rotational velocity is = ﬂat or even increasing with r r 2 distance! p expect v ∝ 1/r outside galaxy Dark Matter (contd.)

For clusters of galaxies, the visible matter is only about 1/10th of that needed to explain the dynamics we see (First pointed out by Fritz Zwicky in 1933 — so this problem has been round a long time!) General consensus is that the A cluster showing lensing ‘missing mass’ is provided by a hitherto undetected particle, which only interacts gravitationally (Though particularly for the galactic rotation curve problem, many attempts also to explain in terms of modiﬁcations to the laws of gravity, e.g. MOND theories.) Fritz Zwicky Dark Energy

On the largest scales in the universe we see not extra attraction, but ‘repulsion’ The universe is accelerating, as measured by the brightness of distant supernovae Is this Λ? Einstein introduced this into his ﬁeld equations for General Relativity to try to get a static universe When he realised the universe was expanding, he discarded this term — we ﬁnally knew that it was necessary in about 1998 A source term or geometry?

Schematically, Einstein’s equations are:

G = 8πT

a geometrical object derived the stress-energy tensor of from the metricg of spacetime sources of matter and radiation

Where does the cosmological constant enter?

G − Λg = 8πT Modiﬁes gravity itself or G = 8πT + Λg A new source of energy

More generally, should we interpret the late-time acceleration of the universe in terms of a modiﬁed gravity theory? — or as the action of e.g. a new form of matter, such as a new scalar ﬁeld (like the Higgs, recently discovered)? Lambda CDM

Putting the two together, we get ΛCDM

This is now the ‘standard model of cosmology’ (in analogy with the Standard Model of particle physics) Here dark matter particle is ‘cold’ — basically moving slowly and non-relativistically today Suitable candidates could be e.g. large mass WIMPS And what provides the repulsion for the accelerating universe is a simple cosmological constant Λ This has a constant ratio of pressure to energy density = −1 Other possibilities like scalar ﬁelds, this changes with time Key tests come from the CMB power spectrum 4. Flight traj e c to r y of HERSCHEl & PLANCK

The launcher’s attitude and trajectory are totally controlled by the two onboard computers, located in the Ariane 5 vehicle equipment bay (VEB). 7.05 seconds after ignition of the main stage cryogenic engine at T-0, the two solid-propellant boosters are ignited, enabling liftoff. The launcher first climbs vertically for 6 seconds, then rotates towards the East. It maintains an attitude that ensures the axis of the launcher remains parallel to its velocity vector, in order to minimize aerodynamic loads throughout the entire atmospheric phase, until the solid boosters are jettisoned. Once this first part of the flight is completed, the onboard computers optimize the trajectory in real time, The Planck Satelliteminimizing propellant consumption to bring the launcher first to the intermediate orbit targeted at the end of the main stage propulsion phase, and then the final orbit at the end of the flight of the cryogenic upper stage. The main stage falls back off the coast of Africa in the Atlantic Ocean (in the Gulf of Guinea). On orbital injection, the launcher will have attained a velocity of approximately 9967 meters/second, and will be at an altitude of about 852 kilometers. The fairing protecting the HERSCHEL, PLANCK spacecraft is jettisoned shortly after the boosters are jettisoned at about T+243 seconds.

Planck has been called ‘the coolest spacecraft ever built’! Certainly payload is one of the most complex scientiﬁc mission ever put into space Cost 700M euros, and mass at launch 1.9 tonnes It ﬂew out to the Second Lagrangian point (L2) of the

Earth/Sun system For more information, visit us on www.arianespace.com 5 Scanning strategy (1 rpm, Semi-stable — ﬂies in a Lissajous plus 1 degree advance per orbit about L2 day) leads to2 × 7 month surveys, each covering entire sky once Planck Science

So what did Planck see, and why is it such a big advance? The key is much improved resolution and sensitivity compared to the previous missions At the higher frequencies, each Planck sky map gives about 50 million pixels at each frequency — compare ∼ 3 million for WMAP Sensitivity about 10 times higher per beam Frequency coverage much improved compared to previously as well — can better discriminate the CMB from Galactic and other foregrounds Planck Science

28 papers plus associated data products released Mar 21 Made headlines around the world, including front page of the NY Times Release based on ﬁrst 15 months of data rest of data (another 15 months) + crucial polarisation data, due in 1 year HFI cryogens ran out in early 2012 — LFI observations ﬁnished recently and Planck now ‘de-orbited’ Planck Cosmology Results

Broad overview of results wouldPlanck Collaboration: be: Cosmological parameters

Table 8. Approximate constraints with 68% errors on Ωm and 1 1 SpectacularH0 (in units of kmoverall s− Mpc− agreement) from BAO, with ωm withand ωb fixed to the best-fit Planck+WP+highL values for the base ΛCDM ΛCMDcosmology. cosmology But with some hints of departures in Sample Ωm H0 places +0.032 +3.2 6dF ...... 0.305 0.026 68.3 3.2 +−0.019 +−2.2 SDSS ...... 0.295 0.017 69.5 2.1 And some tensions with other+−0. results015 +−1.7 SDSS(R) ...... 0.293 0.013 69.6 1.5 +−0.041 −+4.1 WiggleZ ...... 0.309 0.035 67.8 2.8 −+0.015 +−1.6 For exampleBOSS ...... rate ...... of . . . universe. . . . . 0.315 0.015 67.2 1.5 −+0.010 +−1.1 6dF+SDSS+BOSS+WiggleZ ...... 0.307 0.011 68.1 1.1 expansion (H0) from CMB now−+0.009 +−1.0 6dF+SDSS(R)+BOSS ...... 0.305 0.010 68.4 1.0 −+0.009 +−1.0 6dF+SDSS(R)+BOSS+WiggleZ . . . . 0.305 0.008 68.4 1.0 discrepant with recent optical− and IR− determinations at about2 .5σ level surements constrain parameters in the base ΛCDM model, we (Universeform χ2, has got slightly older 2 ΛCDM T 1 ΛCDM χ = (x x ) C− (x x ), (50) PlanckBAO about− 40 MyrBAO −> WMAP9 ΛCDM value.)where x is the data vector, x denotes the theoretical pre- Fig. 16. Comparison of H0 measurements, with estimates of diction for the ΛCDM model and C 1 is the inverse covari- 1 σ errors, from a number of techniques (see text for details). BAO− ± ance matrix for the data vector x. The data vector is as fol- These are compared with the spatially-flat ΛCDM model con- lows: D (0.106) = (457 27) Mpc (6dF); r /D (0.20) = straints from Planck and WMAP-9. V ± s V 0.1905 0.0061, rs/DV(0.35) = 0.1097 0.0036 (SDSS); A(0.44)±= 0.474 0.034, A(0.60) = 0.442 ±0.020, A(0.73) = The results of this section show that BAO measurements are 0.424 0.021 (WiggleZ);± D (0.35)/r = 8.88± 0.17 (SDSS(R)); an extremely valuable complementary data set to Planck. The ± V s ± and DV(0.57)/rs = 13.67 0.22, (BOSS). The off-diagonal com- measurements are basically geometrical and free from complex 1 ± ff ponents of CBAO− for the SDSS and WiggleZ results are given systematic e ects that plague many other types of astrophysical in Percival et al.(2010) and Blake et al.(2011). We ignore any measurements. The results are consistent from survey to survey covariances between surveys. Since the SDSS and SDSS(R) re- and are of comparable precision to Planck. In addition, BAO sults are based on the same survey, we include either one set of measurements can be used to break parameter degeneracies that results or the other in the analysis described below, but not both limit analyses based purely on CMB data. For example, from together. the excellent agreement with the base ΛCDM model evident in The Eisenstein-Hu values of rs for the Planck and WMAP-9 Fig. 15, we can infer that the combination of Planck and BAO base ΛCDM parameters differ by only 0.9%, significantly measurements will lead to tight constraints favouring ΩK = 0 smaller than the errors in the BAO measurements. We can obtain (Sect. 6.2) and a dark energy equation-of-state parameter, w = an approximate idea of the complementary information provided 1 (Sect. 6.5). − by BAO measurements by minimizing Eq. (50) with respect to Finally, we note that we choose to use the either Ωm or H0, fixing ωm and ωb to the CMB best-fit parame- 6dF+SDSS(R)+BOSS data combination in the likelihood ters. (We use the Planck+WP+highL parameters from Table5.) analysis of Sect.6. This choice includes the two most accu- The results are listed in Table8 19. rate BAO measurements and, since the effective redshifts of As can be seen, the results are very stable from survey to these samples are widely separated, it should be a very good survey and are in excellent agreement with the base ΛCDM approximation to neglect correlations between the surveys. 2 parameters listed in Tables2 and5. The values of χBAO are also reasonable. For example, for the six data points of the 2 5.3. The Hubble constant 6dF+SDSS(R)+BOSS+WiggleZ combination, we find χBAO = 4.3, evaluated for the Planck+WP+highL best-fit ΛCDM param- A striking result from the fits of the base ΛCDM model to Planck eters. power spectra is the low value of the Hubble constant, which is The high value of Ωm is consistent with the parameter anal- tightly constrained by CMB data alone in this model. From the ysis described by Blake et al.(2011) and with the “tension” dis- Planck+WP+highL analysis we find cussed by Anderson et al.(2013) between BAO distance mea- 1 1 surements and direct determinations of H (Riess et al. 2011; H0 = (67.3 1.2) km s− Mpc− (68%; Planck+WP+highL).(51) 0 ± Freedman et al. 2012). Furthermore, if the errors on the BAO A low value of H has been found in other CMB experi- measurements are accurate, the constraints on Ωm and H0 (for 0 ments, most notably from the recent WMAP-9 analysis. Fitting fixed ωm and ωb) are of comparable accuracy to those from Planck. the base ΛCDM model, Hinshaw et al.(2012) find

19 1 1 As an indication of the accuracy of Table8, the full likelihood H0 = (70.0 2.2) km s− Mpc− (68%; WMAP-9), (52) results for the Planck+WP+6dF+SDSS(R)+BOSS BAO data sets give ± 1 1 Ωm = 0.308 0.010 and H0 = 67.8 0.8 km s− Mpc− , for the base consistent with Eq. (51) to within 1 σ. We emphasize here that ΛCDM model.± ± the CMB estimates are highly model dependent. It is important

30 Planck Cosmology Results

Planck collaboration: CMB power spectra & likelihood

Angular scale 90◦ 18◦ 1◦ 0.2◦ 0.1◦ 0.07◦ Planck has produced a 6000 wonderful power spectrum of 5000

the ﬂuctuations in the CMB ] 4000 2 K

sky µ

[ 3000 ` D Very big increase in accuracy 2000

— can now deﬁnitely say Dark 1000

Energy and Dark Matter exist, 0 2 10 50 500 1000 1500 2000 2500 just from primordial CMB Multipole moment, ` alone Figure 37. The 2013 Planck CMB temperature angular power spectrum. The error bars include cosmic variance, whose magnitude is indicated by the green shaded area around the best ﬁt model. The low-` values are plotted at 2, 3, 4, 5, 6, 7, 8, 9.5, 11.5, 13.5, 16, 19, 22.5, 27, 34.5, and 44.5.

Proportions of DE and DM Table 8. Constraints on the basic six-parameter ΛCDM model using Planck data. The top section contains constraints on the six now slightly different: instead primary parameters included directly in the estimation process, and the bottom section contains constraints on derived parameters. Planck Planck+WP of what’s shown in Pie chart Parameter Best ﬁt 68% limits Best ﬁt 68% limits Ω h2 ...... 0.022068 0.02207 0.00033 0.022032 0.02205 0.00028 b ± ± 2 (based on previous space Ωch ...... 0.12029 0.1196 0.0031 0.12038 0.1199 0.0027 ± ± 100θ ...... 1.04122 1.04132 0.00068 1.04119 1.04131 0.00063 MC ± ± +0.012 τ ...... 0.0925 0.097 0.038 0.0925 0.089 0.014 experiment (WMAP) values), ± WMAP− n ...... 0.9624 0.9616 0.0094 0.9619 0.9603 0.0073 s ± ± 10 +0.024 ln(10 As) . . . . . 3.098 3.103 0.072 3.0980 3.089 0.027 Planck now has 69% for DE, ± values− +0.018 ΩΛ ...... 0.6825 0.686 0.020 0.6817 0.685 0.016 ± − Ω ...... 0.3175 0.314 0.020 0.3183 0.315+0.016 26% for DM and 5% for m 0.018 ± − σ ...... 0.8344 0.834 0.027 0.8347 0.829 0.012 8 ± ± +4.0 zre ...... 11.35 11.4 2.8 11.37 11.1 1.1 ordinary matter − ± H ...... 67.11 67.4 1.4 67.04 67.3 1.2 0 ± ± 9 +0.051 10 As ...... 2.215 2.23 0.16 2.215 2.196 0.060 ± − Ω h2 ...... 0.14300 0.1423 0.0029 0.14305 0.1426 0.0025 m ± ± Age/Gyr ...... 13.819 13.813 0.058 13.8242 13.817 0.048 ± ± z ...... 1090.43 1090.37 0.65 1090.48 1090.43 0.54 ∗ ± ± 100θ ...... 1.04139 1.04148 0.00066 1.04136 1.04147 0.00062 ∗ ± ± z ...... 3402 3386 69 3403 3391 60 eq ± ±

33 Planck Cosmology Results (contd.)

Many other interesting results A key result for inflation is the slope of the primordial power spectrum of perturbations Generic inflation models predict a primordial power spectrum slope of about 0.96 whereas pre-inflation theory expected value was1 Planck gets0 .9603 ± 0.0073 Incredible that something first predicted about 30 years ago, concerning the first 10−35 seconds of the universe, we are now starting to get confirmation of Planck Cosmology Results — still to come

Polarisation results will be key over next year – potentially can tell us directly energy scale of inﬂation (which is currently constrained to 1012 times larger than LHC can probe) Detecting this mode of polarisation (the B-mode) is LIGO equivalent to detecting gravitational waves in early universe! This may give ﬁrst point of contact with String Theory, since this component predicted to be generically small in string-based cosmologies

String Theory Further space missions

PRISM (Polarized Radiation Imaging and Spectroscopy Mission) is a proposal for an L-class mission to be the ‘ultimate’ mapper of both temperature and polarisation for the CMB Unfortunately has now lost out to Athena (X-ray) and eLISA (gravitatonal waves)

For Dark Matter and Dark Energy, the future is brighter(!) Euclid is an M-Class mission already selected, due for launch 2020 Important for both DE and DM — DM via lensing, and DE via mapping the distribution and redshift of galaxies, and seeing how characteristic scales Euclid evolve with time Planck Results — still to come

Returning to Planck, quality of polarisation data on small angular scales already extremely impressive Line shown is not a ﬁt, but predicted from Temperature data Also Planck, with its high resolution and large frequency coverage, is a very impressive instrument for Galactic studies First release, with about 1000 pages total, has just scratched the surface — deﬁnitely many mysteries remaining! Planck image of dust in the Galaxy