Seeds for the CMB anisotropies and for all the structure of Universe
can be either adiabatic (curvature) or entropy (isocurvature)
perturbations or a (possibly correlated) mixture of these. The CMB and
large-scale structure (LSS) data constrain the nature of these
primordial perturbations on super-Hubble scales, deep at the radiation
dominated era, but well after inflation, reheating, and big bang
nucleosynthesis. The constraints can be converted into constraints on
the fundamental physics and on inflationary models (if one believes in
inflation). For example, single-field inflation produces ONLY adiabatic
perturbations, while multi-field inflation can produce both adiabatic
and isocurvature perturbations. Thus, a detection of even a tiny
primordial isocurvature contribution in the CMB or LSS would
automatically rule out all single-field models.
I'll give a long introduction to "isocurvature" and try to define this
rather technical concept in an INTUITIVE WAY, since I have realised
that, while most of cosmologists have heard about adiabatic and
isocurvature primordial perturbations, they don't necessarily remember
anything more. After the definitions, I talk about the most recent
constraints [1,2], if there is time left! In 2000, when the Boomerang
and Maxima balloon borne CMB experiments detected of the first acoustic
peak, it became clear that the peak position was fit well by an almost
purely adiabatic model, while a subdominant isocurvature contribution is
still allowed by the data. Exact constraints on isocurvature are highly
model and parametrization dependent: Using a phenomenological
parametrization we find that the current CMB data allow for a 15% cold
dark matter (CDM) isocurvature contribution to the primordial
perturbation (while 85% of the primordial perturbation power is in the
adiabatic mode) at scale k = 0.01 Mpc^{-1}. However, assuming generic
slow-roll two-field inflation as a source of perturbations leads to a
much tighter constraint, 3%. Instead of reporting the primordial
isocurvature fraction which depends on the pivot scale, one can also
study how much of the observed total CMB temperature variance today
could be due to the primordial CDM isocurvature mode. We find that the
nonadiabatic (aka isocurvature + correlation) contribution to the
temperature variance is |\alpha_T| < 5% at 95% C.L. in both
phenomenological and slow-roll parametrizations [1].
Finally, the determination of primordial tensor-to-scalar ratio, r,
might be degenerate with the determination of the isocurvature
contribution. However, we show that this is not the case in the
phenomenological parametrization which leads to a stable constraint r <
0.26 at k=0.01 Mpc^{-1}. In the contrary, in the slow-roll case the
adiabatic result 0.26 changes to r < 0.18, if the CDM isocurvature mode
is taken into account. There are also strong degeneracies between the
isocurvature contribution and the determination of \Omega_\Lambda and
the Hubble parameter H_0 that I hopefully have time to explain in the
talk.
[1] http://arxiv.org/abs/arXiv:1202.2852
J. Valiviita, M. Savelainen, M. Talvitie, H. Kurki-Suonio and S.
Rusak,
``Constraints on scalar and tensor perturbations in
phenomenological
and two-field inflation models:
Bayesian evidences for primordial isocurvature and tensor
modes.''
[2] http://arxiv.org/abs/arXiv:0909.5190
J. Valiviita and T. Giannantonio,
``Constraints on primordial isocurvature perturbations and
spatial curvature by Bayesian model selection,''
Phys. Rev. D80, 123516 (2009).