CMB component separation and the physics of foregrounds

1
Latest Results of the COSMOSOMAS Experiment

• CMB component separation and the physics of
  foregrounds
• 2008 July 14-18 Pasadena, California

Sergi R. Hildebrandt, IAC
2
COSMOSOMAS TEAM ISRafael Rebolo (PI)Sergi
Hildebrandt (PS)Carlos M. Gutiérrez de la Cruz
    José Alberto Rubiño Martín     Roger J.
Hoyland     Robert A. Watson (Jodrell Bank)    
Former members        Silvia Fernández Cerezo
    Julio E. Gallegos    Juan F. Macías-Pérez
(LPSC, Grenoble) Elia S. Battistelli (University
British Columbia)Roger Oliva Balagué (ESA/ESTEC)

• SUMMARY
• Main features of the COSMOSOMAS Experiment
• Correlation method
• CMB detection
• free-free, synchrotron and radio sources
  detection
• Signal correlated with far infrared data
• Conclusions and the QUIJOTE CMB experiment

The COSMOSOMAS experiment is placed at the Teide
Observatory in Tenerife island, Spain. Location
16º 30 35 W, 28º 18 00" N, Altitude 2390 m.
Reference Hildebrandt et al., 2007, MNRAS, 382,
594. Web Page http//www.iac.es/project/cmb/cosm
osomas/
3
Main features of the COSMOSOMAS Experiment
CYGNU S A
C
4
Correlation method
Data for different channels (d ) (aprox. 150
good observing days)
X
Our data
Mask
X
X
Templates (t )
Other data
Window function
Mask
d S i a i t i n , where i 0, , N
templates and n is the noise of the data
channel. a is found by a standard Least Square
Fitting. The as are then used to derive the
contributions of the templates that are to be
found in the data and/or mean spectral indexes.
5
CMB detection
COSMOSOMAS
WMAP
C16
C11
W
K
Q
Ka
V
C13
C15

• Detection of CMB in all COSMOSOMAS with an
  amplitude around 27/- 2 mK.
• In agreement with the expected CMB fluctuations
  under our window function (26-27 mK).
• We observe a systematic difference between the
  two channels at 11 GHz, of order 3, compatible
  with the expected polarization of the CMB.
  However, more sensitive is needed to call for a
  detection.

6
Free-free, synchrotron and

• Free-free emission
• The template used is the Ha map provided by
  Finkbeiner 2003.
• The correlation values ranges from 3-6 mK at 11
  GHz to compatible with 0 up to 33 GHz.
• The conversion from Rayleighs to mK is 40-60
  mK/R, compatible with the theoretical value of
  51-55 mK/R for an electron temperature of
  7000-8000 K.
• The spectral index is consistent with -2.

• Synchrotron emission
• The template used is the 408 MHz map provided
  by Haslam et al 1982.
• At first sight one obtains significant
  correlations with a flatter spectral index of
  -2.5 between 11 GHz and 22.5 GHz.
• However at the COSMOSOMAS angular scales the
  template is dominated by extragalactic sources
  (see next slide).
• After a correlation including the NVSS map,
  spectral index of Galactic synchrotron turns out
  to be compatible with -3.
• This result is confirmed with the latest data at
  1420 MHz (Burigana et al. 2006).

7
radio sources detection
408 MHz
?
?
NVSS data
8
radio sources detection
408 MHz
NVSS data
We examine the excess of correlation in 11-22 GHz
after taking into account the CMB, free-free and
Galactic synchrotron emission found before. We
get
9
radio sources detection
408 MHz
NVSS data
We examine the excess of correlation in 11-22 GHz
after taking into account the CMB, free-free and
Galactic synchrotron emission found before. We
get
Unresolved radio sources model by de Zotti et
al. 2005.
N. B. same level as the CMB emission.

Observed by COSMOSOMAS and WMAP K band.
Effective spectral index between 11-22.5 GHz
-1.35
10
Signal correlated with far infrared data

• The templates used in the analysis consist of
  the DIRBE maps from 25 to 240 mm (DIRBE06, 07,
  08, 09 and 10) and the dust map in Schlegel et
  al. 1998.
• All COSMOSOMAS channels detect correlation with
  dust templates at Galactic latitude bgt30º.
• The amplitude of the signal ranges from 10-12 mK
  at 11 GHz, down to 4-7 mK in the 12-17 GHz and
  2.1-2. mK at 22.5 GHz.
• The Galactic latitude dependence supports a
  Galactic origin.

b - 3
Results for the correlation with DIRBE08 where it
turns out to be more significant
K
b - 2
11
Signal correlated with far infrared data

• Removing few, very intense, regions of 1-5
  squared degrees where the spectra is clearly
  free-free dominated (but not traced by Ha), we
  get

• We fit three models
• A single power-law model. The result is b
  0.1 /- 0.2 with a model probability of 46.9 .
• The standard spinning dust model of Draine
  Lazarian where CNM, WNM and WIM have relative
  amplitudes of 0.43, 0.43 and 0.14, respectively.
  Model probability is 39.1 .
• A phenomenological model with a simple Gaussian
  curve. The one plotted. Model probability is in
  this case 63.0 .

b - 1
b - 0.1
12
Signal correlated with far infrared data

• Removing a few, very intense, regions of 1-5
  squared degrees where the spectra is clearly
  free-free dominated (but not traced by Ha), we
  get

• We fit three models
• A single power-law model. The result is b 0.1
  /- 0.2 with a model probability of 46.9 .
• The standard spinning dust model of Draine
  Lazarian where CNM, WNM and WIM have relative
  amplitudes of 0.43, 0.43 and 0.14, respectively.
  Model probability is 39.1
• A phenomenological model with a simple Gaussian
  curve. The one plotted. Model probability is in
  this case 63.0 .

b - 1
b - 0.1

• Model 1 rules out an spectral index of -1 (5 s
  level).
• Model 3 implies a peak frequency around 21.7 /-
  3.5 GHz.
• Correlations with infrared data with shorter
  wavelengths (e.g., 25 mm) show less significant
  results than 100 mm results.

13
Conclusions

• For the scales allowed by our window function
  (1º-5º)
• Presence of CMB component of amplitude 27 /- 2
  mK in agreement with CMB fluctuations after
  taking into account window function.
• Unresolved extragalactic sources dominate at 11
  GHz. The observations are in very good agreement
  with the predictions of de Zotti et al. (2005).
• Correlations with the Ha map that obeys a power
  law compatible with a -2 temperature spectral
  index.
• We only detect diffuse Galactic synchrotron at 11
  GHz. In the mean, the amplitudes follow a power
  law with temperature spectral index of -3.
• A dust correlated emission is detected in each of
  the COSMOSOMAS channels at bgt30º. The Galactic
  latitude dependence of this signal supports a
  Galactic origin. On the other hand, an important
  fraction of the correlation comes from regions
  with high dust emission where free-free emission
  is not well traced by Ha map, probably due to
  extinction.

14
Do not miss the presentation on Friday of the
Quijote CMB experiment by José Alberto
Rubiño-Martín.
Conclusions

• The remaining emission shows a clear flattening
  which is compatible with a free-free index.
  However a better description is obtained with
  models that resemble spinning dust emission (or
  turn-over like models).

The future
Reference Hildebrandt et al., 2007, MNRAS, 382,
594. Web Page http//www.iac.es/project/cmb/cosm
osomas/