Symmetric mass generation is the interaction-induced opening of a fermion gap without spontaneous symmetry breaking. The anomaly-free 3-4-5-0 model of Wang and Wen provides a minimal one-dimensional setting for this phenomenon, but a direct lattice realization faces two obstacles: fermion doubling for local chiral discretizations and perturbative irrelevance of the six-fermion gapping interaction. We address both obstacles. First, we formulate the model on a strictly one-dimensional tangent-fermion lattice, where a nonlocal hopping produces a single chiral branch without a mirror partner while retaining an efficient tensor-network representation. Second, we add a Hubbard-type density-density interaction (Luttinger parameter
K) that reduces the scaling dimension of the 3-4-5-0 interaction from
5 to
5K, making it relevant for
K<2/5. Density-matrix renormalization group calculations show the opening of an excitation gap in this regime without the appearance of a degenerate ground state, the hallmark of symmetric mass generation.
Semiclassical approach has been developed for the one-dimensional interacting fermion systems. Starting from the incommensurate spin density wave (SDW) mean field state for the repulsive Hubbard model in 1D, the non-Abelian bosonized Lagrangian describing the spin-charge separation is obtained. The Berry phase term is derived from the chiral anomaly, and we obtain the massless Tomonaga-Luttinger liquid in the single chain case while the spin gap opens in the double-chain system. This approach offers a new method to identify the strong-coupling fixed point, and its relation to the Abelian bosonization formalism is discussed on the spin gap state. The generalization to higher dimensions is also discussed.
We critically examine the issue raised by Nagaosa and Oshikawa [J. Phys. Soc. Jpn. 65 (1996) 2241] on how Fujikawa's Jacobian arising from chiral anomaly contributes a topological term to the derivative expansion of the 1d Hubbard model at half filling, thus governing the behaviour of the spin gap. We find that when the nonabelian nature of the system is properly accounted for, the scheme requires modification. The modified result is shown to agree with nonabelian bosonization. Extensions to spin-Peierls systems are outlined.