2017ApJ...847L..15O

We constrain the quasar contribution to the cosmic reionization based on our deep optical survey of z ∼ 6 quasars down to z_R = 24.15 using Subaru/Suprime-Cam in three UKIDSS-DXS fields covering 6.5 deg². In Kashikawa et al. (2015), we select 17 quasar candidates and report our initial discovery of two low-luminosity quasars (M₁₄₅₀∼-23) from seven targets, one of which might be a Lyα-emitting galaxy. From an additional optical spectroscopy, none of the four candidates out of the remaining 10 turn out to be genuine quasars. Moreover, the deeper optical photometry provided by the Hyper Suprime-Cam Subaru Strategic Program (HSC-SSP) shows that, unlike the two already-known quasars, the i - z and z - y colors of the last six candidates are consistent with M- or L-type brown dwarfs. Therefore, the quasar luminosity function (QLF) measurement in the previous paper is confirmed. Compiling the QLF measurements from the literature over a wide magnitude range, including an extremely faint AGN candidate from Parsa et al. (2017), to fit them with a double power law, we find that the best-fit faint-end slope is α=-2.04_–0.18^+0.33 (-1.98_–0.21^+0.48) and characteristic magnitude is M₁₄₅₀^* =-25.8_–1.9^+1.1 (-25.7_–1.8^+1.0) in the case of two (one) quasar detection. Our result suggests that, if the QLF is integrated down to M₁₄₅₀=-18, quasars produce ∼1%-12% of the ionizing photons required to fully ionize the universe at z ∼ 6 with a 2σ confidence level, assuming that the escape fraction is f_esc=1 and the intergalactic medium clumpy factor is C = 3. Even when the systematic uncertainties are taken into account, our result supports the scenario that quasars are the minor contributors of the reionization.