Mass-renormalized electronic excitations at Ñp,0Ö in the superconducting state

A critical goal in the study of high-temperature superconductors ~HTSCs! is an understanding of the interactions or correlation effects which ‘‘dress’’ the electrons near the Fermi energy EF . This is important ~1! for its own right, ~2! because these interactions have been considered to be so extreme that even the concept of a quasiparticle may breakdown in these systems, 1,2 and ~3! because any particularly strong interactions may serve as candidates for mediating the pairing of electrons within a Cooper pair, just as the interactions of the electrons with phonons is responsible for the pairing in conventional superconductors. In the many-body language of solid-state physics, the electron self-energy S(k ,v) contains the information of these interactions or correlation effects. This dressing renormalizes the dispersion of electrons near the Fermi energy, giving them an enhanced mass or flatter E vs k dispersion. At large binding energies ~greater than the energy of the boson being coupled to!, the dispersion returns to its bare value, giving the dispersion a ‘‘kink.’’ The energy scale and strength of the kink are thus related to the boson energy and coupling strength, respectively. In the HTSCs, particular attention should be payed to the electrons near the ( p,0) region of the Brillouin zone, which is the region in which both the ‘‘nonquasiparticle like’’ effects and superconducting pairing fluctuations are largest. In this paper, we present what we believe to be the first unambiguous evidence of mass renormalizations or kinks in the electronic structure near ( p,0) of a high-temperature superconductor. In particular, we used angle-resolved photoemission spectroscopy ~ARPES! to make the first observations of a relatively low-energy kink of 40 meV or less ~dependent on doping! which is distinct from the higher-energy ~60-70 meV! kink which has been observed along the nodal or (p,p) direction of the Brillouin zone where the superconducting gap and other interaction effects are weakest. 3‐7 In optimal and underdoped samples these energy scales merge and it is harder to deconvolve the two types of kinks. Previous ARPES efforts at measuring kink or renormalization effects near the ( p,0) region have had great disparity. Lanzara 6 argued that the observed kink effects are essentially k independent, while Valla 4 and Kaminski 5 argued that the kink continually evolved, growing stronger as ( p,0) approached. Kaminski additionally argued that the kink at (p,0) was the origin of the well-known peak-dip-hump ~PDH! line shape. 9‐11 We believe that the main reason for the disparity of these results was an inability to properly deconvolve the various features, especially the bilayer splitting. This deconvolution is especially difficult near (p,0) as the features are numerous, overlapping, and typically quite broad. By overdoping high quality single-crystalline Bi2Sr2CaCu2O81d samples, we have obtained very sharp spectral features near ( p,0) and have been able to accurately deconvolve the bilayer splitting as well as superstructure effects 12,13 with similar work done by Feng et al. 14 and more recently, other studies as well. 15,16 This was a necessary prerequisite for this work. Our experiments give qualitatively different results than any of these previous studies, and additionally indicate that the PDH line shape previously observed near (p,0) 8‐1 1 has major complications from the bilayer splitting, which obscures much of the true interaction effects. Other recent studies have indicated similar concerns of bilayer splitting on the PDH line shape. 15