This plot show the result of a 2 parameter analysis of the DASI polarization data. This is a "shaped band power" analysis assuming the E mode power spectrum shape as predicted for the Concordance Model favored by last year's temperature data - the units are relative to this model. The same shape is assumed for the B mode spectrum. (right panel) The point shows the maximum likelihood value with Fisher matrix error bars, and the contours are at the likelihood fall-offs relative to the peak which would correspond to 1 sigma increments for a 1D Gaussian distribution (i.e. exp(-n^2/2) where n is the number of sigma). (left panel) Shows the 2D distribution at right marginalized onto the axes. The grey lines enclose 68% of the total likelihood. The green band shows the distribution of E mode expectation values for a large grid of cosmological models weighting by the likelihood of those models given last year's temperature data.
The likelihood ratio between the point E=0, B=0 and the maximum value is 1e-6 corresponding to 5 sigma.
This plot shows that DASI sees E mode polarization at a level consistent with the Concordance Model, and highly inconsistent with zero, and does not detect B mode polarization.
This plots shows results from several likelihood analyses - an E5/B5 joint analysis, a T5 analysis, and the TE results from a T/E/TE5 analysis. These are flat band power analyses. The blue line shows the maximum likelihood band power values with the grey error bars indicating the 68% central region of the likelihood marginalizing over the other parameter values (analogous to the grey lines in the first plot above). In each case the green line is the canonical Concordance Model.
This plot shows that the new temperature data is consistent with last year's result (although not nearly as good - these new observations are optimized for polarization detection). Note also the extremely small error bars for the first E/B bins - we are clearly showing that there is tiny E mode power in the region of the spectrum where the model says there should not be any. The TE cross correlation is also an important result.
This plot shows a 2 parameter T single shaped band power versus temperature spectral index analysis. Spectral index is relative to thermal - on this plot synchrotron emission would be expected to have an index of approx. -3. The constraint on the spectral index is very tight, and any synchrotron contribution to the temperature data is clearly very small.
This plot shows a 2 parameter E single shaped band power versus spectral index analysis. Spectral index is relative to thermal - on this plot synchrotron emission would be expected to have an index of approx. -3. The constraint is not terribly strong, but is nearly a 2 sigma result against synchrotron.
This plot shows the result of a 3 parameter shaped band power analysis T/E/TE. Again marginal distributions for each parameter are shown at left, and at right the 2D E/TE distribution is shown marginalized over the T dimension. Each of the parameters is seen to be consistent with the concordance value, and TE cross correlation is detected at around the 2 sigma level.
These maps are constructed from polarized datasets that have been split by epoch, and formed into sum or difference data vectors. In order to isolate the most significant signal in our data, only the subset of 34 eigenmodes which, within the concordance model, are expected to have signal/noise >1 have been used. These maps visually illustrate the result that these individual modes in our polarized dataset do indeed show a significant signal, which has been measured repeatably. The residual in the difference map is consistent with instrumental noise in the included eigenmodes.
The upper panel shows the E (solid blue) and B (solid red, much lower curve) parameter window functions that indicate the response to the E power spectrum of the two parameters in our E/B analysis. The blue dashed curve shows the result of multiplying the E parameter window function by the concordance E spectrum, illustrating that for this CMB spectrum, most of the response of our single band E parameter comes from the region of the second peak (250<l<450), with a substantial contribution also from the third peak and a smaller contribution from the first. The lower panel shows E1-E5 (blue) and B1-B5 (red, again much lower) parameter window functions for the E power spectrum for our E5/B5 analysis. DASI's response to E and B is very symmetric, so that the corresponding plots that show these parameters response to the B power spectrum are nearly identical to these, with the E and B parameters reversed.