Time-reversal symmetry breaking? (reply)

Kaminski et al. reply- There are two components of the circular dichroism (CD) signal in angle-resolved photoemission (ARPES) measurements. One is always present in crystals, regardless of any time-reversal symmetry considerations. This component, which we refer to as ‘geometric’, is antisymmetric about any symmetry plane of the crystal, and is therefore zero at that plane. But in underdoped samples of the high-temperature superconductor Bi2212, we find another component, which is non-zero at the symmetry plane below the pseudogap temperature. We attribute that component to time-reversal symmetry breaking. The objections of Borisenko et al.1 comprise three main points: the circular dichroism that we observe at the mirror plane2 is due to the superstructure of the Bi–O layer; our momentum accuracy is not as we stated; and the absence of dichroism in overdoped samples is due to a weaker influence of the superstructure because of an increased Fermi surface volume compared with underdoped samples.