We report extensive measurements of the Knight shift K, the nuclear spin-lattice relaxation rate 1/${\mathit{T}}_1$, and the Gaussian spin-echo decay rate 1/${\mathit{T}}_{2\mathit{G}}$ of ${}_{}{}^{63}{}_{}{}^{}\mathrm{Cu}$ in overdoped ${\mathrm{TlSr}}_2$${\mathrm{CaCu}}_2$${\mathrm{O}}_{7\mathrm{-}\mathrm{\delta }}$ (Tl1212) with ${\mathit{T}}_{\mathit{c}}$ = 70 K, 52 K, and 10 K, in order to elucidate the origin of the reduction in ${\mathit{T}}_{\mathit{c}}$ with increasing holes and to identify the symmetry of the order parameter. In the normal state, it is shown that 1/${\mathit{T}}_1$T obeys the Curie-Weiss law, pointing to the presence of the antiferromagnetic (AF) spin correlation. From the analyses of 1/${\mathit{T}}_1$ and 1/${\mathit{T}}_{2\mathit{G}}$, it is found that the increase of the hole content in Tl1212 compounds makes the characteristic energy of the AF spin fluctuation around a zone boundary, Q=(π/a,π/a), ${\mathrm{\Gamma }}_{\mathit{Q}}$, transfer to a higher-energy region and concomitantly reduces the magnetic correlation length ${\xi }_{\mathit{m}}$ significantly. The AF spin correlation is concluded to become less distinct in going from the optimum-doped to the overdoped regime. In the superconducting state, the T dependences of K and 1/${\mathit{T}}_1$ have revealed that the superconductivity is in the gapless regime with a finite density of states at the Fermi level. The NMR results are consistently interpreted in the d-wave model in which the impurity scattering is incorporated in terms of the unitarity limit as demonstrated in most of the high-${\mathit{T}}_{\mathit{c}}$ cuprates so far. Eventually, the reduction in ${\mathit{T}}_{\mathit{c}}$ from 70 K to 52 K in Tl1212 is concluded to be not due to the impurity effect associated with the oxygen content. In the previous works, the enhancement of ${\mathit{T}}_{\mathit{c}}$ from 93 K in ${\mathrm{YBa}}_2$${\mathrm{Cu}}_3$${\mathrm{O}}_7$ with double ${\mathrm{CuO}}_2$ layers to 115–135 K in ${\mathrm{Tl}}_2$${\mathrm{Ba}}_2$${\mathrm{Ca}}_2$${\mathrm{Cu}}_3$${\mathrm{O}}_{10}$ and ${\mathrm{HgBa}}_2$${\mathrm{Ca}}_2$${\mathrm{Cu}}_3$${\mathrm{O}}_{8+\mathrm{\delta }}$ with triple ${\mathrm{CuO}}_2$ layers was shown to be due to the increase in ${\mathrm{\Gamma }}_{\mathit{Q}}$ with ${\xi }_{\mathit{m}}$ unchanged appreciably. This finding was compatible with the relationship of ${\mathit{T}}_{\mathit{c}}$∝${\mathrm{\Gamma }}_{\mathit{Q}}$${\xi }_{\mathit{m}}^2$exp(-1/λ) based on the spin-fluctuation-induced mechanism for the d-wave superconductivity. Within the same scheme, the origin of the marked decrease in ${\mathit{T}}_{\mathit{c}}$ irrespective of increasing ${\mathrm{\Gamma }}_{\mathit{Q}}$ in Tl1212 is proposed to be due to the significant reduction in ${\xi }_{\mathit{m}}$ which makes the pairing interaction weaken and λ in the above formula reduced. © 1996 The American Physical Society.

• Received 15 May 1996

DOI:https://doi.org/10.1103/PhysRevB.54.10131