For bicep/ Keck and SPTpol because the observations are being made at the South Pole with scans along the azimuth direction, scanwise filtering leads to modes along the ℓ y axis being suppressed. The color stretch in all four panels is identical. The axis scales for the bicep/ Keck Fourier plane are zoomed in compared with the rest of the panels to focus on the modes accessible by bicep/ Keck’s small apertures. Simulated 2D E-mode signal power spectra of bicep/ Keck, SPTpol, and Planck. This is the first demonstration of improvement in an r constraint through delensing.

Applying the technique to the real data, the constraint on r is improved from r 0.05 < 0.090 to r 0.05 < 0.082 (95% C.L.). From simulations, we find that the uncertainty on r is reduced by ∼ 10 %, from σ ( r ) = 0.024 to 0.022, which can be compared with a ∼ 26 % reduction obtained when using a perfect lensing template or if there were zero lensing B-modes. We expand the bicep/ Keck likelihood analysis framework to accept a lensing template and apply it to the bicep/ Keck dataset collected through 2014 using the same parametric foreground modeling as in the previous analysis. The large-scale-structure tracer used is a map of the cosmic infrared background derived from Planck satellite data, while the polarized CMB map comes from a combination of South Pole Telescope, bicep/ Keck, and Planck data.

This template is constructed by combining an estimate of the polarized CMB with a tracer of the projected large-scale structure. We present a constraint on the tensor-to-scalar ratio, r, derived from measurements of cosmic microwave background (CMB) polarization B-modes with “delensing,” whereby the uncertainty on r contributed by the sample variance of the gravitational lensing B-modes is reduced by cross-correlating against a lensing B-mode template.

Template Label 103