17O nuclear spin-lattice relaxation in YBa2Cu3O6+x single crystals

From underdoped to overdoped regime in YBa2Cu3O6+x, an NMR investigation of single crystals

“…8. However, if we now mrn to data measured with H 0 II Z, we see that such a relationship no longer holds, which has the important consequence that the anisotropy of the NSLRR is now temperature dependent [16], in contrast with the behavior observed in YBal.92Sr0.08Cu30 7 [16,31]. Such a temperature dependence of the anisotropy of the NSLRR for H 0 in the a,b plane was first reported by Barriquand et al [33] in a YBa2Cu306.…”

“…It can be seen in Figs.5 and 6 that this anisotropy is constant in sample 1 [16,31,32], whereas it is varying in sample 3 [16], where the ratio 17Tl(Z)/17Tl(Y) is close to 1.5 from room temperature down to 100 K, but decreases down to 1.15 when the temperature decreases to T c. Such a variation of the anisotropy, which was first observed by Bar¡ et al [33], is extensively discussed by Horvati~ et al [31] and we shall only mention here the conclusion of the analysis: both 17( T1 ( y)T)-l and 17( TI(X)T)-I are proportional to Zs(T), the deviation from this behaviour appearing only for H 0 II Z axis, an additional contribution to the NSLRR occurs at low temperature, and it is larger for this peculiar orientation than for the others; it happens that orbital relaxation by spinless particles fits this type of anisotropy. Again, this point favours the phase diagram mentioned above, but new experiments are needed to specify in which range of concentration x this temperature dependence of tTTl(Z)/17Tl(Y) is present, and how it is related to the decrease of 63( TI 7)-I.…”

“…We know from Takigawa [9}, Yoshinari et al [12], and Horvati~ et al [16,31], that x7( Ta 7)-] (for H0 II X, Y axes) is accurately proportional to the static susceptibility Zs (7). Thus in order to reconcile neutron data and Cu(2) and 0(2,3) NSLRR results, we need to find an expression for Z"(q, co) which is compatible with the neutron data as lar as the width of the peaks around q --QAV (about 0.1 to 0.14 in reduced lattice unit for the concentration range x--0.5--0.7) and the characteristic energy coN are concerned, and which leads to a negligible contribution for the 0(2,3) site; i.e., we want the contribution of the peaks around QAV to oxygen NSLRR to be an order of magnitude smaller that the experimental value of the NSLRR.…”

“…iv) The NSLRR divided by the temperature 17( T1T)-l at the 0(2,3) site has been found with great accuracy to scale affinely with the MHS tensor K(7) [9,12,16] (K(T)TIT= const. ), a behavior which is quite different from the well-known Korringa law (KZTI T= const.)…”

“…v) Finally, it has been shown by Horvati~ et al [2,3] that the temperature dependence of the Cu(2) NSLRR could switch between two different behaviors: i) For the so-called regular YBa2Cu307 a continuous increase of (T17) -l from high temperature down to Te, followed by a sudden decrease just below the phase transition; at the same time, the static spin susceptibility Z s remains constant (or rather slightly increasing [16,18,19]. ii) For sub-stoichiometric samples, for either value of Te, 90 of 60 K, the 63( T1T)-i exhibits a maximum at a temperature T* of the order of 130 K, which is well above Te.…”