Spin Peierls transition of $J_{1}-J_{2}$ and extended models with ferromagnetic $J_{1}$.Sublattice dimerization and thermodynamics of zigzag chains in $\beta$-TeVO$_{4}$

TL;DR

This study investigates the spin-Peierls transition in ferromagnetic-antiferromagnetic zigzag chains, showing that sublattice dimerization explains thermodynamic properties of $eta$-TeVO$_{4}$ better than previous models, with detailed numerical analysis.

Contribution

It introduces a model with sublattice dimerization for ferromagnetic $J_1$ chains, extending understanding of spin-Peierls instabilities and thermodynamics in $eta$-TeVO$_{4}$.

Findings

Sublattice dimerization occurs for $ ext{J}_2/ ext{J}_1 > 0.65$.

Models reproduce $eta$-TeVO$_{4}$ susceptibility and specific heat above 8 K.

Lower temperature data suggest a gapped chain rather than a pure $J_1-J_2$ model.

Abstract

The spin$-1/2$ chain with ferromagnetic exchange $J_1 < 0$ between first neighbors and antiferromagnetic $J_2 > 0$ between second neighbors supports two spin-Peierls (SP) instabilities depending on the frustration $\alpha = J_2/\vert J_1\vert$. Instead of chain dimerization with two spins per unit cell, $J_1-J_2$ models with $\alpha > 0.65$ and linear spin-phonon coupling are unconditionally unstable to sublattice dimerization with four spins per unit cell. Unequal $J_1$ to neighbors to the right and left extends the model to gapped ($\gamma > 0$) chains with conditional SP transitions at $T_{SP}$ to dimerized sublattices and a weaker specific heat $C(T)$ anomaly. The spin susceptibility $\chi(T)$ and $C(T)$ are obtained in the thermodynamic limit by a combination of exact diagonalization of small systems with $\alpha > 0.65$ and density matrix renormalization group (DMRG) calculations of systems up to $N \sim 100$ spins. Both $J_1-J_2$ and $\gamma > 0$ models account quantitatively for $\chi(T)$ and $C(T)$ in the paramagnetic phase of $\beta$-TeVO$_{4}$ for $T > 8$ K, but lower $T$ indicates a gapped chain instead of a $J_1-J_2$ model as previously thought. The same parameters and $T_{SP} = 4.6$ K generate a $C(T)/T$ anomaly that reproduces the anomaly at the $4.6$ K transition of $\beta$-TeVO$_{4}$, but not the weak $\chi(T)$ signature.