Understanding of metal-insulator transition in VO2 based on experimental and theoretical investigations of magnetic features

Abstract

The metal-insulator transition temperature Tc in VO2 is experimentally shown to be almost the same as a magnetic transition temperature Tm characterized by an abrupt decrease in susceptibility, suggesting the evidence of the same underlying origin for both transitions. The measurement of susceptibility shows that it weakly increases on cooling for temperature range of T > Tm, sharply decreases near Tm and then unusually increases on further cooling. A theoretical approach for such unusual observations in susceptibility near Tm or below is performed by modeling electrons from each two adjacent V4+ ions distributed along V-chains as a two-electron system, which indicates that the spin exchange between electrons could cause a level splitting into a singlet (S = 0) level of lower energy and a triplet (S = 1) level of higher energy. The observed abrupt decrease in susceptibility near Tm is explained to be due to that the sample enters the singlet state in which two electrons from adjacent V4+ ions are paired into dimers in spin antiparallel. By considering paramagnetic contribution of unpaired electrons created by the thermal activation from singlet to triplet levels, an expression for susceptibility is proposed to quantitatively explain the unusual temperature-dependent susceptibility observed at low temperatures. Based on the approach to magnetic features, the observed metal-insulator transition is explained to be due to a transition from high-temperature Pauli paramagnetic metallic state of V4+ions to low-temperature dimerized state of strong electronic localization.

Magnetic phase and magneto-resistive effects in vanadium oxide epitaxial nanoclusters

“…This phase could explain the magnetic behavior observed in some vanadium oxide thin films. [6][7][8] Finally, we observe that the magnetization in these oxide clusters leads to MR effects at low temperatures. Applying an in-plane magnetic field from À5 to þ5 T noticeably modifies the conductance, which follows a bell-shaped curve (Fig.…”

Toward Room-Temperature Thermochromism of VO2 by Nb Doping: Magnetic Investigations

“…Also, it can be noticed that they interpreted the absence of the magnetic moment associated to V-V pairs from spin pairing of the electrons from adjacent V 4+ ions along the chain direction at an angle θ close to 180°, on the contrary to us. Other tests, by the same group, (27) led to the proposition of a spin gap system due to the V-V dimerization state with an antiferromagnetic interaction. The magnetic susceptibility would result from the thermal activation from singlet to triplet levels, but these considerations were unsuccessfully giving gap values in the range of 2 K which can be assumed irrelevant.…”

“…Because each dimer contains two electrons, the number of near-free electrons would change from Np to (Np -2Nd0) in the M phase and (Np -2Nd0)/Np is the percentage of unpaired V4+ ions. (27) The results are presented in Figure 12b. It can be seen that the major V 4+ ions are paired into V-V dimers, and ∼94% V 4+ are paired into dimers for the undoped VO 2 .…”

“…It is characterized by Pauli and Curie-like paramagnetic (PM) behaviors above and below the T c . Through experimental and theoretical investigations on pure VO 2 , Zhang et al (27) concluded that the abrupt decrease in susceptibility is due to that the VO 2 sample enters the singlet state in which the two electrons from adjacent V 4+ ions are paired into dimers, which are formed due to the spin exchange between electrons. However, the discussions were only based on undoped VO 2 .…”

Unraveling the Mott-Peierls intrigue in vanadium dioxide

“…However, GW does not give easy access to the total energy, and therefore it does not explain why low temperatures should favour the M1 distorted phase against the rutile undistorted one. In turns, LDA+U or GGA+U calculations, known to overemphasise the exchange splitting, predict the existence of local moments even in the rutile phase [61][62][63], not observed in experiments [64]. Relatively recent calculations based on HSE hybrid functionals bring even worst results: both rutile and M1 phases are predicted to be magnetically ordered insulators, with the former lower in energy [65,66], even though earlier calculations were claimed to be more in accordance with experiments [25].…”