Phonon suppression of coherence peak in nuclear spin relaxation rate of superconductors

Abstract

THE temperature dependence of the nuclear spin relaxation rate l/T₁ peaks sharply (the 'Hebel-Slichter' or 'coherence' peak) just below the transition temperature T_c of a superconductor¹. Because the observation² of this peak definitively confirmed BCS theory³, its absence^4–6 in the high-T_c oxide superconductors is the best evidence against a BCS picture. Here we show that, to the contrary, an extended form of BCS theory gives a natural explanation. Using the extension by Migdal⁷ and Eliashberg⁸ to retarded interactions (for phonon coupling, the attraction between electrons propagates at the speed of sound, slower than the Fermi velocity of electrons), we predict unexpectedly strong damping effects in all dynamical properties when the temperature T is close to T_c. The origin of the damping (which suppresses the coherence peak) is numerous electron–phonon decay channels open to excitations because of the high T_c itself. This process still works even if another source of attraction beyond electron–phonon coupling causes the highT_c. Thus our observations remove a barrier inhibiting conventional descriptions of high-T_c materials, but by no means force such an interpretation.