A ¹⁷O nuclear magnetic resonance study of high temperature superconductors

• Main Author: Howes, Andrew Paul
• Published: University of Warwick 1992
• Subjects: 530.41; QC Physics
• Online Access: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.332547

17O Nuclear magnetic resonance is an ideal probe of the high temperature superconductors as it can sample both the static and dynamic electronic susceptibilities of the Cu-O planes of these materials. The 17O NMR site assignments and the temperature dependence of the shift and spin-lattice relaxation has been investigated for the n = 1 and 3 phases of Bi2Sr2Can-1CunO4+2n and the n=3 phase of Tl2Ba2Can-1CunO4+2n. The shift of the 17O resonance associated with the Cu-O plane of the non-superconducting Bi2Sr2CuO6 is temperature independent from room temperature to 5K. This is not the situation for Bi2Sr2Ca2Cu3O10 and Tl2Ba2Ca2Cu3O10. For the Bi2Sr2Ca2Cu3O10 superconductor (Tc » 107K) both of the distinct Cu- O planes of the unit cell show a marked but different temperature dependence of the NMR shift above Tc. The shift associated with the central Cu-O plane starts to decrease below - 300K and the spin component of the Knight shift has dropped to -1/2 of its room temperature value by Tc. The outer Cu-O plane shift is almost temperature independent until - 120K when it decreases sharply and the spin component of the Knight shift is -2/3 of its room temperature value by Tc. The difference in these temperature dependencies suggests that a one component susceptibility model cannot describe this system. The Tl2Ba2Ca2Cu3O10 superconductor is structurally very similar to Bi2Sr2Ca2Cu3O10. The 17O NMR shift has been studied for this sample before an annealing process (Tc = 114K) and after (Tc= 124K). The room temperature resonances observed before and after the anneal are essentially the same. Interestingly the resonances observed for the two distinct Cu-O planes of the structure have very similar shift values and cannot be easy resolved (unlike the situation for Bi2Sr2Ca2Cu3O10). Before and after annealing a temperature dependence of the resonance associated with the Cu-O planes is observed but in contrast to Bi2Sr2Ca2Cu3O10 the two different plane resonances show the same temperature dependencies. Before the anneal the spin component of the shift is estimated to be — 2/3 of its room temperature value by Tc whereas after the anneal the shift has drooped to - 1/6 of its room temperature value by Tc. For both Bi2Sr2Ca2Cu3O10 and Tl2Ba2Ca,Cu3O.0 the relaxation behaviour of the Cu-O planes appears to be Korringa like (Kn2T1T is constant) in the normal state despite these temperature dependent shifts. The values obtained for Kn2T1T is similar to the theoretical value for simple s-type metals with no electron-electron or antiferromagnetic interaction. However the data does not completely rule out any other relaxation behaviour.