$^{63}$Cu and $^{199}$Hg NMR study of HgBa$_{2}$CuO$_{4+\delta}$ single crystals

Abstract

We report on the temperature dependence of $^{63}$Cu and $^{199}$Hg NMR magnetic shifts and linewidths for an optimally doped and an underdoped HgBa$_{2}$CuO$_{4+\delta}$ single crystal, as well as the quadrupole splitting and its distribution for $^{63}$Cu. From the $^{63}$Cu and $^{199}$Hg \textit{magnetic shifts} we have recently concluded on the existence of two spin components with different temperature dependencies [J. Haase, D. Rybicki, C. P. Slichter, M. Greven, G. Yu, Y. Li, and X. Zhao, Phys. Rev. B 85, 104517 (2012)]. Here we give a comprehensive account of all data and focus on the linewidths and quadrupole splittings. While the $^{63}$Cu quadrupole coupling and its distribution are by and large temperature independent, we identify three regions in temperature for which the magnetic widths differ significantly: at the lowest temperatures the magnetic linewidths are dominated by the rigid fluxoid lattice that seems to have disappeared above about 60 K. In the intermediate temperature region, starting above 60 K, the magnetic linewidth is dominated by the spatial distribution of the magnetic shift due to the pseudogap spin component, and grows linearly with the total shift up to about \textit{$\sim$}170-230 K, depending on sample and nucleus. Above this temperature the third region begins with an sudden narrowing where the second, Fermi-liquid-like spin component becomes homogeneous. We show that all linewidths, quadrupolar as well as magnetic, above the fluxoid dominated region can be understood with a simple model that assumes a coherent charge density variation with concomitant variations of the two spin components. In addition, we find a temperature independent spin based broadening in both samples that is incoherent with the other broadening for the underdoped crystal, but becomes coherent for the optimally doped crystal.