1 Constraining the Dark Universe Rachel Beana, Steen H. Hansenb, Alessandro Melchiorrib aTheoretical Physics, The Blackett Laboratory,Imperial College, Prince Consort Road, London, U.K. bNAPL, University of Oxford, Keble road, OX1 3RH, Oxford, UK We combine complementary datasets to constrain dark energy. Using standard Big Bang Nucleosynthesis and the observed abundances of primordial nuclides to put constraints on ΩQ at temperatures near T ∼ 1MeV, we find the strong constraint ΩQ(MeV)<0.045 at 2σ c.l.. Under the assumption of flatness, using results from Cosmic Microwave Background(CMB)anisotropymeasurements,highredshiftsupernovae(SN-Ia)observationsanddatafromlocalcluster 2 abundanceswe put a new constraint on the equation of state parameter wQ <−0.85 at 68% c.l.. 0 0 2 Introduction. The discovery that the universe’s Ia) to put constraints on the dark energy equation n evolution may be dominated by an effective cosmo- of state parameterized by a redshift independent a logical constant [1], is one of the most remarkable quintessence-field pressure-to-density ratio w . We Q J cosmologicalfindingsofrecentyears. Anexceptional also made use of the Hubble Space Telescope (HST) 0 opportunity is now opening up to decipher the na- constraint on the Hubble parameter h=0.72±0.08 1 tureofdarkmatter[2],totesttheveracityoftheories [19]. 2 andreconstructthedarkmattersprofileusingawide Wewillbrieflyreviewtheresultsobtainedinthose v variety of observations over a broad redshift range. paper in the next sections. 7 One candidate that could possibly explainthe ob- 2 Early-Universe Constraints from BBN. Inthe last servationsisadynamicalscalar“quintessence”field. 1 few years important experimental progress has been One of the strong aspects of quintessence theories 1 made in the measurement of light element primor- 0 is that they go some way explaining the fine-tuning dial abundances. For the 4He mass fraction, Y , 2 problem, why the energy density producing the ac- He 0 celerationis ∼10−120M4. A vastrangeof“tracker” two marginally compatible measurements have been / pl obtained from regression against zero metallicity in h (see for example [3,11]) and “scaling” (for example blue compact galaxies. A low value Y = 0.234± p [12]-[15]) quintessence models exist which approach He - attractor solutions, giving the required energy den- 0.003 [5] and a high one YHe = 0.244± 0.002 [6] o give realistic bounds. We use the high value in our r sity, independent of initial conditions. The common t characteristic of quintessence models is that their analysis;if one instead consideredthe low value, the s bounds obtained would be even stronger. a equationsofstate,w =p/ρ,varywithtimewhilsta Q : Observations in different quasar absorption line v cosmological constant remains fixed at wQ=Λ = −1. systems give a relative abundance of deuterium, i Observationally distinguishing a time variation in X which is critical in fixing the baryon fraction, of the equation of state or finding w different from r Q D/H =(3.0±0.4)·10−5 [7]. a −1 will therefore be a success for the quintessential In the standard BBN scenario, the primordial scenario. abundances are a function of the baryon density We will here discuss observational constraints on η ∼ Ω h2 only. In order to put constraints on the general quintessence models. b energy density of a primordial field a T ∼ MeV, we In [23] we used standard big bang nucleosynthe- modifiedthestandardBBNcode [4]byincludingthe sis and the observed abundances of primordial nu- quintessence energy component Ω . We then per- clides to put constraints on the amplitude of the en- Q formed a likelihood analysis in the parameter space ergy density, Ω , at temperatures near T ∼ 1MeV. Q (Ω h2,ΩBBN) using the observed abundances Y The inclusion of a scaling field increases the expan- b Q He and D/H. In Fig. 1 we plot the 1,2 and 3σ likeli- sion rate of the universe, and changes the ratio of hood contours in the (Ω h2,ΩBBN) plane. neutrons to protons at freeze-outandhence the pre- b Q Our main result is that the experimental data for dicted abundances of light elements. 4He and D does not favour the presence of a dark In [24] we have then combined the latest observa- energy component, providing the strong constraint tions of the Cosmic Microwave Background (CMB) Ω (MeV)<0.045at2σ (corresponding to λ>9 for anisotropies provided by the Boomerang [16], DASI Q the exponential potential scenario), strengthening [17] and Maxima [18] experiments and the informa- significantlythepreviouslimitof [8,15]Ω (MeV)< tion from Large Scale Structure (LSS) with the lu- Q 0.2. The reason for the difference is due to the minosity distance of high redshift supernovae (SN- improvement in the measurements of the observed 2 degenerate in the interplay between w and Ω . Q Q Furthermore, it does not add any new features be- yond those produced by the presence of a cosmolog- ical constant [25], and it is not particularly sensitive to further time dependencies of w . Q CMB Degeneracy 6000 Ω=0.30 Ω=0.050 h=0.65 w= -1.0 m b Q 5000 Ω=0.44 Ω=0.073 h=0.54 w= -0.6 m b Q 2K Ωk=0 µ 4000 π 2 C/l Figure 1. 1,2 and 3σ likelihood contours in the 1) 3000 (Ωbh2,ΩQ(1MeV)) parameter space derived from ll(+ 4He and D abundances. 2000 1000 abundances,especiallyforthedeuterium,whichnow 0 corresponds to approximately∆N <0.2−0.3 ad- 10 100 1000 eff ditional effective neutrinos (see, e.g. [9]), whereas Multipole l Ref. [8,15] used the conservative value ∆N < 1.5. eff One could worry about the effect of any underesti- matedsystematicerrors,andwethereforemultiplied the error-bars of the observed abundances by a fac- Figure2. CMBpowerspectraandtheangulardiam- tor of 2. Even taking this into account, there is still eter distance degeneracy. The models are computed a strong constraint ΩQ(MeV) < 0.09 (λ > 6.5) at assuming flatness, Ωk = 1−ΩM −ΩQ = 0). The 2σ. Integrated Sachs Wolfe effect on large angular scale Constraints on the Dark energy equation of state. slightly breaks the degeneracy. The degeneracy can be brokenwith a strong prior on h, in this paper we Theimportanceofcombiningdifferentdatasetsin use the results from the HST. order to obtain reliable constraints on w has been Q stressed by many authors (see e.g. [20], [21],[22]), since eachdatasetsuffersfromdegeneraciesbetween the various cosmological parameters and w . Even Secondly, the time-varying Newtonian potential Q if one restricts consideration to flat universes and after decoupling will produce anisotropies at large to a value of w constant in time then the SN-Ia angular scales through the Integrated Sachs-Wolfe Q luminosity distance and position of the first CMB (ISW) effect. The curve in the CMB angular spec- peakarehighlydegenerateinw andΩ ,theenergy trumonlargeangularscalesdependsnotonlyonthe Q Q density in quintessence. valueofwQbutalsoitsvariationwithredshift. How- The effects of varying w on the angular power ever, this effect will be difficult to disentangle from Q spectrumoftheCMBanisotropiescanbereducedto thesameeffectgeneratedbyacosmologicalconstant, just two . Firstly, since the inclusion of quintessence especially in view of the affect of cosmic variance changes the overall content of matter and energy, and/orgravitywavesonthe largescale anisotropies. theangulardiameterdistanceoftheacoustichorizon In order to emphasize the importance of degen- size at recombination will be altered. In flat models eraciesbetweenalltheseparameterswhile analyzing (i.e. where the energy density in matter is equal the CMB data, we plot in Figure 2 some degener- to Ω = 1−Ω ), this creates a shift in the peaks ate spectra,obtainedkeepingthe physicaldensity in M Q positions of the angular spectrum as matter ΩMh2, the physical density in baryons Ωbh2 andRfixed. Aswecanseefromtheplot,modelsde- R = (1−Ω )y, (1) generate in w can easily be constructed. However Q Q q the combination of CMB data with other different zdec y = [(1−Ω )(1+z)3 datasets can break the mentioned degeneracies. Z Q 0 +Ω (1+z)3(1+wQ)]−1/2dz. Type Ia supernovae, in particular, can be ex- Q tremely useful in this. Evidence that the universe’s It is important to note that the effect is completely expansionratewasacceleratingwasfirstprovidedby 3 -0.5 Table 1 Constraints on w and Ω = 1−Ω using differ- Q M Q ent priors and datasets. We always assume flatness B -0.6 M and t > 10 Gyr. The 1σ limits are found from C 0 S the 16% and 84% integrals of the marginalized like- N -Ia lihood. The HST prior is h = 0.72 ± 0.08 while -0.7 for the BBN prior we use the conservative bound Q W Ω h2 =0.020±0.005. b -0.8 CMB+HST w <−0.62 Q -0.9 0.15<ΩM <0.45 CMB+HST+BBN −0.95<w <−0.62 Q 0.15<Ω <0.42 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 M Ω M CMB+HST+SN-Ia −0.94<wQ <−0.74 0.16<Ω <0.34 M CMB+HST+SN-Ia+LSS w <−0.85 Q Figure 3. Contours of constant R (CMB) and 0.28<Ω <0.43 M SN −Ia luminosity distance in the w -Ω plane. Q M The degeneracy between the two distance measures can be broken by combining the two sets of com- plementary information. The luminosity distance is chosentobe equalto dl atz =1forafiducialmodel ter marginalizing over all remaining nuisance pa- withΩ =0.7,Ω =0.3,h=0.65. (Wenotethatas rameters. The analysis is restricted to flat uni- Λ M ΩQ =1−ΩM goes to zero the dependence of R and verses. One can see that wQ is poorly constrained d upon w also become zero, as there is no dark from CMB data alone, even when the strong HST L Q energy present.) Figure taken from [24] prior on the Hubble parameter, h = 0.72 ± 0.08, is assumed. Adding a Big Bang Nucleosynthesis prior, Ω h2 = 0.020 ± 0.005, has small effect on b the CMB+HST result. Adding SN-Ia breaks the two groups, the SCP and High-Z Search Team([1]) CMB Ω −w degeneracy and improves the upper Q Q usingtypeIasupernovae(SN-Ia)toprobethenearby limit on w , with w < −0.74. Finally, including Q Q expansion dynamics. SN-Ia seem to be good stan- information from local cluster abundances through dard candles, as they exhibit a strong phenomeno- σ = (0.55±0.1)Ω−0.5, where σ is the rms mass logicalcorrelationbetweenthedeclinerateandpeak flu8ctuation in spherMes of 8h−1 M8pc, further breaks magnitude of the luminosity. The observed appar- the quintessential-degeneracy,giving w <−0.85 at Q ent bolometric luminosity is related to the lumi- 1-σ. Also reportedin Table I, are the constraints on nosity distance, measured in Mpc, by m = M + B Ω . As we can see, the combined data suggests the M 5logd (z) + 25, where M is the absolute bolomet- L presence of dark energy with high significance, even ric magnitude. The luminosity distance is sensi- in the case CMB+HST. It is interesting to project tive to the cosmological evolution through an inte- ourlikelihoodintheΩ −w plane. Proceedingasin Q Q gral dependence on the Hubble factor d = (1 + l [27],we attribute alikelihoodto apointinthe (Ω , z ′ ′ M z) (dz /H(z ,Ω ,w ), and can therefore be used 0 Q q wQ) plane by finding the remaining parametersthat toRconstrain the scalar equation of state. As can be maximize it. We then define our 68%, 95%and 99% seen in Figure 3 there is an inherent degeneracy in contourstobewherethelikelihoodfallsto0.32,0.05 theluminositydistanceintheΩ /w plane;onecan M Q and0.01ofits peakvalue,aswouldbe thecasefora seethatlittlecanbesaidabouttheequationofstate two dimensional multivariate Gaussian. In Figure 4 from luminosity distance data alone. However, the weplotlikelihoodcontoursinthe(Ω ,w )planefor M Q degeneraciesofCMBandSN1adatacomplementone the joint analyses of CMB+SN-Ia+HST+LSS data another so that together they offer a more powerful together with the contours from the SN-Ia dataset approachfor constraining w . Q only. As we can see, the combination of the data breaks the luminosity distance degeneracy. Table I shows the 1-σ constraints on w for dif- Q ferent combinations of priors, obtained in [24] af- Conclusions. 4 and datasets), our analysis is compatible with other -0.6 recent analysis on w ([30]). Our final result is per- Q fectly in agreement with the w = −1 cosmological Q constant case and gives no support to a quintessen- -0.7 tial field scenario with w > −1. A frustrated net- Q wQ-0.8 Supernovae 95% c.l.99% c.l. wfiaeonlrdukmaorbfeedreoxmocfaluqindueiwndtaelalstsseohnritgaihaplsumigreonldiyfiecelsaxnpacoreen.ehnIitgnihaallydsdcdaiitlsiifonang- vored, e.g. power law potentials with p ≥ 1 and the 66 -0.9 88%% cc..ll.. CMB+SN-Ia+LSS+HST oscIitllwatiollrybeptohteendtiuatly, toofnhaigmheerarfeedws.hift datasets, for examplefromclusteringobservations[29]topointto -1.0 a variation in w that might place quintessence in a 0.2 0.4 0.6 0.8 1.0 more favorable light. Ω The result obtained here, however, could be M plagued by some of the theoretical assumptions we made. The CMB and LSS constraints can be weakened by the inclusion of additional relativis- tic degrees of freedom [33], by a background of gravity waves, of isocurvature perturbations or by Figure 4. The likelihood contours in the (Ω , w ) M Q adding features in the primordialperturbationspec- plane, with the remaining parameters taking their tra. These modifications are not expected in the best-fitting values for the joint CMB+SN-Ia+LSS most basic and simplified inflationary scenario but analysis described in the text. The contours corre- they are still compatible with the present data. The spondto 0.32,0.05and0.01ofthe peak value of the SN-Ia result has been obtained under the assump- likelihood, which are the 68%, 95% and 99% confi- tion of a constant-with-time w > −1. Inclusion of Q dence levels respectively. theregionwithw <−1couldaffectourconstraints Q ([32]). In[24]wehaveshownthatingeneralw isa eff rathergoodapproximationfordynamicalquintessen- tialmodelssincethe luminositydistancedepends on We haveprovidednew constraintsonthe darken- wQ through a multiple integral that smears its red- ergy by combining different cosmological data. We shift dependence, and our results are therefore valid have examined BBN abundances in a cosmological forawideclassofquintessentialmodels. This‘numb- scenariowith a scaling field. We have quantitatively ing’ofsensitivitytowQ,firstnoticedby[31],implies discussed how large values of the fractional density thatmaybeaneffective equationofstateis themost inthescalingfieldΩ atT ∼1MeVcanbeinagree- tangible parameter able to be extracted from super- Q mentwiththeobservedvaluesof4HeandD,assum- novae. However with the promise of large data sets ingstandardBigBangNucleosynthesis. The2σlimit fromPlanckandSNAPsatellites,opportunitiesmay Ω (1MeV) < 0.045 severely constrains a wide class yet still be open to reconstructa time varying equa- Q ofquintessentialscenarios,likethosebasedonanex- tion of state [22]. ponentialpotential. For example,for the pure expo- Acknowledgements It is a pleasure to thank Ruth nential potential the total energy today is restricted Durrer, Pedro Ferreira, Massimo Hansen and Matts to Ω = 3Ω (1MeV)≤0.04. This result put strong Q 4 Q Roosforcommentsandsuggestions. RBandAMare constraints on the presence of quintessence during supportedby PPARC.SHH is supportedby a Marie recombination. CurieFellowshipofthe EuropeanCommunityunder We have then provided constraints on the equa- thecontractHPMFCT-2000-00607. Weacknowledge tion of state parameter w . The new CMB results Q the use of CMBFAST [26]. provided by Boomerang and DASI improve the con- straints from previous and similar analysis (see e.g., REFERENCES [20],[28]),withw <−0.85at68%c.l. (w <−0.76 Q Q at 95% c.l.). We have also demonstrated how the 1. P.M. Garnavich et al, Ap.J. Letters 493, L53- combination of CMB data with other datasets is 57 (1998); S. Perlmutter et al, Ap. J. 483, crucial in order to break the Ω −w degeneracy. 565 (1997); S. Perlmutter et al (The Super- Q Q The constraintsfrom eachsingle datasets are,as ex- novaCosmologyProject),Nature39151(1998); pected, quite broadbut compatible with eachother, A.G. Riess et al, Ap. J. 116, 1009 (1998); B.P. providinganimportantconsistencytest. Whencom- Schmidt, Ap. J. 507, 46-63 (1998). parison is possible (i.e. restricting to similar priors 2. N. Bahcall, J. P. Ostriker, S. Perlmutter and 5 P.J.Steinhardt,Science284,1481(1999)[astro- [astro-ph/0 107029], J. Newman, C. Marinoni, ph/9906463]; A. H. Jaffe et al,Phys. Rev. Lett. A. Coil & M. Davis [astro-ph/0109131],T. Mat- 86 (2001) [astro-ph/0007333]. subara & A. Szalay [astro-ph/0105493]. 3. I. Zlatev, L. Wang, & P. Steinhardt, Phys. Rev. 30. P.S. Corasaniti & E.J. Copeland, astro- Lett. 82 896-899(1999). ph/0107378; T. Saini, S. Raychaudhury, V. 4. L. Kawano,Fermilab-Pub-92/04-A(1992). Sahni and A.A. Starobinsky, Phys. Rev. Lett. 5. K.A. Olive and G. Steigman, Astrophys. J. 85, 1162 - 1165 (2000); Y. Wang, G. Lovelace, Suppl. Ser. 97, 49 (1995). astro-ph/0109233;D.Marfatia,V.Barger,astro- 6. Y.I. Izotov and T.X. Thuan, ApJ, 500 188 ph/0009256.C. Baccigalupi,A. Balbi, S. Matar- (1998); rese, F. Perrotta,N. Vittorio, astro-ph/0109907. 7. S. Burles and D. Tytler, ApJ, 499 689 (1998). 31. I. Maor, R. Brustein, P.J. Steinhardt, 8. P. G. Ferreira and M. Joyce, Phys.Rev.Lett. 79 Phys.Rev.Lett. 86 (2001) 6 (1997) 4740-4743[astro-ph/9707286]. 32. I. Maor, R. Brustein, J. McMahon and P. J. 9. S. Burles, K. M. Nollett, J. N. Truran and Steinhardt, astro-ph/0112526 M. S. Turner, Phys. Rev. Lett. 82, 4176 (1999) 33. R. Bowen et al., arXiv:astro-ph/0110636. [astro-ph/9901157]; S. Esposito, G. Mangano, A. Melchiorri, G. Miele and O. Pisanti, Phys. Rev. D 63 (2001) 043004 [arXiv:astro- ph/0007419]; 10. S. H. Hansen, G. Mangano, A. Melchiorri, G. Miele and O. Pisanti, arXiv:astro- ph/0105385. 11. P. Brax, J. Martin & A. Riazuelo, Phys. Rev. D.,62 103505 (2000). 12. C Wetterich, Nucl. Phys B. 302 668 (1988) 13. B. Ratra and J. Peebles, Phys. Rev D37 (1988) 321. 14. J. Frieman, C. Hill, A. Stebbins, I. Waga, Phys. Rev. Lett, 75 2077 (1995). 15. P. Ferreira and M. Joyce, Phys.Rev. D58 (1998) 023503. 16. C.B. Netterfield et al., [astro-ph/0104460]. 17. C. Pryke et al., [astro-ph/0104489]. 18. A. Lee et al., [astro-ph/0104459]. 19. Freedman W. L. et al, 2000, ApJ in press, preprint astro-ph/0012376. 20. S. Perlmutter, M.S. Turner, M. White, Phys.Rev.Lett. 83 670-673 (1999). 21. W. Hu, astro-ph/9801234. 22. J. Weller, A. Albrecht, Phys.Rev.Lett. 86 1939 (2001) [astro-ph/0008314]; D. Huterer and M. S. Turner, [astro-ph/0012510]; M. Tegmark, [astro-ph/0101354]. 23. R. Bean, S. H. Hansen and A. Melchiorri, Phys. Rev. D 64 (2001) 103508 [arXiv:astro- ph/0104162]. 24. R. Bean and A. Melchiorri, arXiv:astro- ph/0110472,Phys. Rev. D in press. 25. G. Efstathiou & J.R. Bond [astro-ph/9807103]; A. Melchiorri&L.M.Griffiths, New Astronomy Reviews, 45, 4-5, 2001, [astro-ph/0011197]. 26. M.Zaldarriaga&U.Seljak,ApJ.469437(1996). 27. A. Melchiorri et al., Astrophys.J. 536 (2000) L63-L66, astro-ph/9911445. 28. J.R.Bondetal.[TheMaxiBoomCollaboration], astro-ph/0011379. 29. M.O. Calv˜ao, J.R.T De Mello Neto& I. Waga