CMB constraints on cosmic strings

1
CMB constraints on cosmic strings
UK cosmology meeting Ambleside August 2006
Neil Bevis
Mark Hindmarsh (Sussex) Martin Kunz (Sussex,
then Geneva) Jon Urrestilla (Sussex, then Tufts,
shortly Sussex again)
in preparation this work
astro-ph/0605018 CMB calculations
astro-ph/0403029 global textures
2
Introduction
CMB observations discount strings (topological
defects in general) from being primarily
responsible for anisotropies... ... but not from
playing a sub-dominant effect. Adiab
atic inflation component and defect component are
uncorrelated in linear theory such that
individual power spectra simply
add. (Perturbations in cosmic fluids are 10-5 so
do not affect defect evolution)
3
Introduction
The normalization of the defect contribution is
a free parameter relating to the energy-scale of
the defects. For Abelian-Higgs
strings Perturbations proportional to
CMB power spectrum proportional to

4
Bouchet et al. 2001

string contribution
Ad2
5
Bevis et al. - Global textures
For case of global O(4) textures calculated via
UETC approach Using WMAP first year data (and
ACBAR, CBI, VSA) Considered also the changes in
cosmology and the primordial power
spectra via MCMC approach using modified
CosmoMC How much can the inflationary
contribution change to accommodate defects?
6
Bevis et al. - Global textures
CAMB calculation of inflationary CMB TT and TE
takes 3 s (on Cosmos) MCMC using 100,000
iterations takes 3 days (on Cosmos) In defect
CMB calculation, cosmology enters only in
UETC-enabled version of CMBEASY the very costly
simulations need not be repeated! But
UETC-CMBEASY takes 20 hours (on 2.4 GHz 32-bit
Intel chips) MCMC is not feasible, at least at
first sight
7
Bevis et al. - Global textures
Suppose that (i) the cosmology is quite
tightly constrained
(ii) the defect contribution is small Then,
changes in cosmology have small effect on small
defect contribution, i.e. they are very
small. Smaller than possible numerical errors in
defect calculation and smaller than WMAP error
bars (so would not affect the likelihood)
Method is to fully vary the inflationary
contribution, but keep the cosmology fixed for
the defect part only change the defect
normalization. MCMC then takes only 3 days.
8
Bevis et al. - Global textures
But... WMAP 1yr VSA CBI ACBAR data does
not constrain global defect contribution to be
small Data allows for degeneracy involving
Ad2, As2 (obviously) and ?bh2, h, ns allowing
high defect fractions. fd fractional defect
contribution at l10 But large fd
incompatible with Kirkman et al. value of ?bh2
and Hubble Key Project value of h
degeneracy
9
Bevis et al. - Global textures

68
WMAP 1yr
95
WMAP 1yr BBN HKP
68
95
Apply this independent data and suppositions are
valid but detection of textures is removed
10
Wyman et al.
Wyman et al. 2005 have performed a
multi-parameter analysis for cosmic strings using
the moving segment model approach (Albrecht et
al. 1997) Simple model allows for rapid CMB
calculations and they do include the variation of
the defect contribution with the cosmological
parameters. They do not see a
significant degeneracy, but their analysis
includes SDSS data. However strings do not
contribute greatly to Pk and SDSS data is less
precise than CMB data. Or it is due to a number
of human-decided priors, which are reasonable,
but from one perspective a little tight.
strings contribute lt 0.07 at l10
11
Present work
Our previous MCMC method for cosmic strings
using field evolution simulations and
corresponding UETC-based CMB calculations of
astro-ph/0605018. And using 3
year WMAP data and new BOOMERanG data
Cosmic strings (Wyman et al.) Cosmic strings
(Bevis et al.) Global textures (Bevis et al.)
12
MCMC with all CMB data
WMAP 3rd year (astro-ph/0603451) BOOMERanG
(astro-ph/0507494) CBI (astro-ph/0402359) VSA
(astro-ph/0402498) ACBAR (astro-ph/0212289)

13
MCMC with all CMB data
Strings are favoured by the
data - 2 sigma detection!
68
95
14
MCMC with WMAP3 data
Strings are favoured by the
WMAP3 data, at between 1 and 2 sigma level
68
95
15
Present work

16
Present work

17
Present work

18
All CMB BBN HKP
WMAP 3rd year (astro-ph/0603451) BOOMERanG
(astro-ph/0507494) CBI (astro-ph/0402359) VSA
(astro-ph/0402498) ACBAR (astro-ph/0212289) BBN
(astro-ph/0302006) HKP(astro-ph/0012376)

19
All CMB BBN HKP

all CMB
all CMB BBN HKP
20
Numerical results
WMAP normalization at l10 gave 0 2.0
x 10-6 (astro-ph/0604018)
Compare to Wyman et al. who obtained
lt 0.27 x 10-6 (astro-ph/0604141) (moving
segment model, WMAP-1 and SDSS, 0 1.1 x
10-6)
Or Fraisse 2006 who obtained lt 0.26 x
10-6 (astro-ph/0603589) (moving segment model,
WMAP-3)
21
Polarization
22
Polarization
Tensors _at_ r0.3
EE lensed
23
Conclusion
First likelihood analysis for string CMB results
from field theoretic simulations. CMB data
deviates from best-fit inflation model in a
series of wiggles which gives a moderate
preference for strings, and in fact a 2-sigma
detection. However, detection utilizes slightly
unfavourable values of ?bh2 and h such that, upon
inclusion of BBN and HKP data, the significance
is reduced. Data gives an upper bound of 10
contribution to TT from strings at l10 New
data is on the horizon, particularly the B-mode
polarization Use of SDSS data remains to be
explored
24
Polarization

TT
TE
EE
BB
r0.3
EE lensed
WMAP figure from lambda.nasa.gov
25
Polarization

TT
fd 0.1
TE
EE
BB
WMAP figure from lambda.nasa.gov