Susceptibilities of the vanadium Magnéli phasesVnO2n−1at low temperature

Abstract

The magnetic susceptibilities of the insulating Magnéli phases ${\mathrm{V}}_3$ ${\mathrm{O}}_5$, ${\mathrm{V}}_4$ ${\mathrm{O}}_7$, ${\mathrm{V}}_6$ ${\mathrm{O}}_{11}$, and ${\mathrm{V}}_8$ ${\mathrm{O}}_{15}$ have been measured at low temperatures. Due to a short-range-order effect ${\mathrm{V}}_3$ ${\mathrm{O}}_5$ exhibits a broad maximum at 125 K, which is substantially higher than the Néel temperature, 76 K, of this compound. For ${\mathrm{V}}_4$ ${\mathrm{O}}_7$ and ${\mathrm{V}}_6$ ${\mathrm{O}}_{11}$ the difference between the Néel temperature and the temperature of the maximum in the susceptibility is small. This excludes the possibility that a linear-chain model discussed by Khattak et al. can explain the specific heat of these two compounds. The temperature derivative of the susceptibility for these two compounds has a maximum at the Néel temperature which corresponds to Fisher's relation. The susceptibility for ${\mathrm{V}}_8$ ${\mathrm{O}}_{15}$ is nearly constant between 5.6 and 68 K, and depends on past history below 5.6 K, i.e., whether the sample is being cooled or heated. However this temperature hysteresis is sample dependent. These experimental data cannot be accounted for by an ordinary antiferromagnetic transition in the insulating state and suggest another type of transition. The possibility of the existence of a mixed charge-density and spin-density wave is discussed, since this can explain qualitatively the experimental features found for ${\mathrm{V}}_8$ ${\mathrm{O}}_{15}$.