V51NMR study of the quasi-one-dimensional alternating chain compoundBaCu2V2O8

Abstract

$^{51}\mathrm{V}$ nuclear magnetic resonance (NMR) studies in polycrystalline sample of ${\mathrm{BaCu}}_{2}{\mathrm{V}}_{2}{\mathrm{O}}_{8}$ reveal the existence of nonzero spin densities, with unequal magnitudes, at the two vanadium sites, viz., ${\mathrm{V}}_{1}$ and ${\mathrm{V}}_{2}$ and hence the participation of ${\mathrm{VO}}_{4}$ tetrahedra to the intrachain and interchain exchange interaction. The behavior of ${K}_{\mathrm{iso}}$ below 200 K provide a clear signature of the reduction of ${\mathrm{Cu}}^{2+}$ spin susceptibility within the chain. Furthermore, there exist unique hyperfine coupling constants for ${\mathrm{V}}_{1}$ and ${\mathrm{V}}_{2}$ sites. This data also confirms the existence of only one spin component and the ground state corresponds to a nonmagnetic spin singlet. The $T$ dependence of ${K}_{\mathrm{iso}}$ above 120 K follows the nature of ${\ensuremath{\chi}}_{\text{spin}}(T)$ proposed by Hatfield for alternating chain model. While below 80 K, it follows the expression for ${\ensuremath{\chi}}_{\mathrm{spin}}$ of a one-dimensional chain with a spin gap $\ensuremath{\Delta}$. From the ${K}_{\mathrm{iso}}$ data we have obtained $\ensuremath{\Delta}=215\ifmmode\pm\else\textpm\fi{}5\phantom{\rule{0.3em}{0ex}}\mathrm{K}$, which is smaller than that determined from the magnetic susceptibility $(\ensuremath{\Delta}=230\phantom{\rule{0.3em}{0ex}}\mathrm{K})$ data. The hyperfine field for ${\mathrm{V}}_{1}$ and ${\mathrm{V}}_{2}$ sites are 24.6 and $14.4\ifmmode\pm\else\textpm\fi{}1.0\phantom{\rule{0.3em}{0ex}}\mathrm{kOe}∕{\ensuremath{\mu}}_{B}$, respectively, with a ratio of 1.7 between them. Thus the exchange path $\mathrm{Cu}\ensuremath{-}\mathrm{O}\ensuremath{-}{\mathrm{V}}_{2}\ensuremath{-}\mathrm{O}\ensuremath{-}\mathrm{Cu}$ in ${\mathrm{BaCu}}_{2}{\mathrm{V}}_{2}{\mathrm{O}}_{8}$ is not negligible. The spin-lattice relaxation time, ${T}_{1}$, for both the sites are identical in the range 300--40 K and vary from 3 ms to 10 s. $1∕{T}_{1}$ follows an activated behavior and provides a gap parameter $\ensuremath{\Delta}=380\phantom{\rule{0.3em}{0ex}}\mathrm{K}$, which is much higher than that obtained from shift data. Moreover, the magnitude of $1∕{T}_{1}T$ decreases more rapidly than that of ${K}_{\mathrm{iso}}$ below 75 K.