The primordial E-mode fluctuations are lensed by large scale structure in the Universe and produce both E- and B-modes. The B-modes from lensing have a well predicted amplitude and should be detectable by the next generation of CMB polarization experiments. CMB polarization lensing has a fundamental advantage over galaxy lensing measurements. The redshift where the lensing occurs is much higher for CMB polarization than for galaxies, and therefore structure formation is still in the linear regime where theory is more accurate. Dark Energy changes the evolution of structure, and therefore the lensing power spectrum. For Polarbear-II's sensitivity, the optimal sky area for characterization of lensing is ~10% Such an observation would give a measurement of the Dark Energy equation of state w to 30% accuracy. A future satellite that measured the entire sky could reach 10% errors on w.

Neutrinos also affect the formation of structure since they are hot and do not readily clump. The CMB polarization lensing power spectrum is sensitive to the amount of hot dark matter and therefore to the sum of the neutrino masses. Optical weak lensing is also sensitive to the sum of neutrino masses, but CMB polarization lensing has a higher potential accuracy since the structure formation is in the linear regime.

Currently, the sum of neutrino masses is limited by large scale structure and the WMAP CMB temperature measurement to <0.7 eV. A CMB polarization map of 10% of the sky with Polarbear-II would give an error of ~0.2 eV on the sum of neutrino masses. An eventual all-sky satellite measurement could have an error as low as 0.03 eV. Polarbear will be able to map 10% of the sky to the depth required to approach 10% errors on w and 0.1 eV error on the sum of the neutrino masses.