Spin and charge dynamics in the hole-doped one-dimensional-chain-ladder composite material Sr 14 Cu 24 O 41 : Cu NMR/NQR studies
暂无分享,去创建一个
Comprehensive ${}^{63,65}$Cu NMR/NQR measurements have been performed on single crystals of Sr${}_{14}{\mathrm{Cu}}_{24}{\mathrm{O}}_{41},$ a hole-doped material containing alternating layers of one-dimensional CuO${}_{2}$ chains and Cu${}_{2}$O${}_{3}$ ladders. While the ladder sites show a unique resonance, two distinct resonance spectra are obtained for the chain sites. They are assigned to the magnetic Cu sites with spin-1/2 and the nonmagnetic Cu sites, which form the Zhang-Rice (ZR) singlet with holes on the oxygen sites. The NMR spectrum at the ZR chain sites shows sharp multipeak structure at low temperatures, indicating a long period of superstructure. The structure becomes obscure and peaks merge into a single broad line with increasing temperature due to thermally induced disorder or motion. A giant oscillation of the spin-echo intensity was observed at the magnetic chain sites as a function of the time separation between $\ensuremath{\pi}/2$ and $\ensuremath{\pi}$ rf pulses. This is well explained if these sites form spin-singlet dimers, which interact very weakly with each other. The nuclear spin-lattice relaxation rate ${(1/T}_{1})$ at both chain sites shows an activated temperature dependence below $T=50$ K with a gap of 125 K, corresponding to the singlet-triplet splitting of the dimers. The ZR chain sites show an anomalous increase of ${1/T}_{1}$ above 200 K. The ladder Cu sites also show an activated temperature dependence of ${1/T}_{1}$ with a gap of 650 K above 200 K, indicating a spin-gap in the ladders. However, ${1/T}_{1}$ at the ladder sites measured by zero-field NQR is dominantly caused by fluctuations of the electric-field gradient (EFG) in the temperature range 30\char21{}150 K and shows a peak near $T=100$ K. This is most likely caused by slow motion of doped holes and/or lattice distortion. The inverse correlation time of the EFG fluctuations is estimated using a simple model of motional effects. It shows an activated temperature dependence with a gap of 230 K, which is an order of magnitude smaller than the activation energy for the electrical conductivity (2200 K).