Almost Strong Zero Modes at Finite Temperature

Niklas Tausendpfund, Aditi Mitra, Matteo Rizzi · 2025

Read Almost Strong Zero Modes at Finite Temperature by Niklas Tausendpfund, Aditi Mitra, Matteo Rizzi (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-11121.txt.

Introduction

Tausendpfund, Mitra and Rizzi study in 2025 what happens to almost strong zero modes at temperatures between the two known extremes. At zero temperature, interacting fermion chains hold edge zero modes that make ground states degenerate. At infinite temperature, related spin chains host long-lived edge modes instead. The middle range was largely unexplored. The team blended Lanczos expansion with tensor networks, lifted to operator level.

That let them simulate large systems over long times and extract decay rates. For the Kitaev-Hubbard model, the edge mode's decay rate falls exponentially as inverse temperature grows. The controlling energy scale exceeds the system's thermodynamic gap.

Catalog record

CREATORNiklas Tausendpfund, Aditi Mitra, Matteo Rizzi
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE80 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-11121

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Almost Strong Zero Modes at Finite Temperature Niklas Tausendpfund, Aditi Mitra, Matteo Rizzi Abstract Interacting fermionic chains exhibit extended regions of topological degeneracy of their ground states as a result of the presence of Majorana or parafermionic zero modes localized at the edges. In the opposite limit of infinite temperature, the corresponding non-integrable spin chains, obtained via generalized Jordan-Wigner mapping, are known to host so-called Almost Strong Zero Modes, which are long-lived with respect to any bulk excitations. Here, we study the fairly unexplored territory that bridges these two extreme cases of zero and infinite temperature. We blend two established techniques for states, the Lanczos series expansion and a tensor network ansatz, uplifting them to the level of operator algebra. This allows us to efficiently simulate large system sizes for arbitrarily long timescales and to extract the temperature-dependent decay rates. We observe that for the Kitaev-Hubbard model, the decay rate of the edge mode depends exponentially on the inverse temperature β \beta , and on an effective energy scale Δ eff \Delta_{\rm eff} that is greater than the thermodynamic gap of the system Δ \Delta . I Introduction Fractionalization of low energy excitations is one of the most interesting properties of topological many-body systems, with the simplest example of this being the Majorana zero modes ( MZM s) [ 1 , 2 , 3 , 4 , 5 ] that appear in fermionic chains protected only by the fermion parity P . These MZMs correspond to fractionalized fermions exponentially localized at the edges of a finite-size system. Because of this non-local nature, a (topological) ground-state degeneracy arises, as there is no local physical operator that couples to these fractionalized fermions. The appearance of edge modes can easily be understood in the non-interacting limit as the model formally belongs to the BDI symmetry class in the Altland-Zirnbauer classification [ 6 , 7 ] . For finite interaction strengths, it has been shown that the topological region – with the appearance of fractionalized edge modes within the degenerate ground-state manifold – persists for a wide range of parameters [ 8 , 9 , 10 , 11 ] . Away from the ground-state manifold, little is known about the spectral properties for generic interactions, and whether stable MZMs exist even for excited states. However, this is important because it strongly affects the finite temperature lifetime of local edge excitations that have overlap with the MZMs, and has ramifications on practical realization of topologically protected qubits. In the zero-temperature limit, where only the ground-state manifold contributes to the dynamics, the topological degeneracy leads to an infinite lifetime of these edge excitations. On the other hand, at infinite temperature, the same edge excitations were shown to have an unusually long lifetime compared to generic bulk excitations [ 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 ] . In integrable limits, this behavior can be explained by the appearance of a Strong Zero Mode ( SZM ) [ 1 , 20 , 21 , 22 , 23 ] , a generalization of the Majorana zero mode to the full spectrum. Here, the existence of the SZM implies a protected degeneracy of the whole spectrum and not only of the ground-state manifold, and thus one recovers an infinite lifetime. Away from these special limits, the lifetime becomes strictly finite, with the edge mode often referred to as an Almost Strong Zero Mode ( ASZM ) [ 12 ] . Besides some phenomenological arguments [ 13 ] , it is still an open question how the lifetime of the edge excitations behaves at finite temperatures. In particular, it is not known how the infinite lifetime emerges as the temperature is lowered from infinity to zero. In this work, we explore this rather unchartered territory by employing a tensor network ansatz [ 24 ] to approximate the sequence of operators generated by the Lanczos algorithm for Heisenberg time evolution [ 25 , 26 ] . This procedure maps the operator dynamics to the time evolution of a single-particle problem on an artificial one-dimensional chain. Originally used to study the complexity growth of time-evolved operators, this method has been shown to be a useful tool for understanding the emergence of these long-lived ASZMs [ 16 , 17 ] . However, these earlier studies were highly limited in system size and, most importantly, to infinite temperatures. The formulation of the Lanczos algorithm for tensor networks removes the limitation of small system sizes by introducing a controlled approximation given by the bond dimension \chi of the tensor network. Moreover, an intrinsic tensor network formulation allows for the efficient inclusion of arbitrary temperatures T=1/\beta , since the density matrix \rho(\beta) can also be approximated by a tensor network [ 27 ] . We emphasize that in contrast to direct integration schemes such as the time-dependent variational principle (TDVP) [ 28 , 29 ] , the Lanczos method developed here converges quickly with the bond dimension in the case of a long-lived ASZM. In the former case, the bond dimension necessarily has to grow exponentially with time, while this is not the case for our Lanczos approach, see Appendix F.3 for details. As an application, we explicitly calculate the lifetime of the Majorana edge mode in the non-integrable Kitaev-Hubbard chain [ 30 ] . We find an exponential dependence of the lifetime on the inverse temperature \tau(\beta)=\exp(\Delta_{\rm{eff}}\beta) . However, in contrast to generic bulk excitations where \Delta_{\rm{eff}} is expected to be the energy gap in the many-body spectrum [ 31 ] , we observe a non-trivial dependence of \Delta_{\rm{eff}} on the interaction strength. In particular, we find that \Delta_{\rm{eff}} is consistently larger than the many-body gap, hinting at a degenerate structure of the low-energy portion of the spectrum and not exclusively of the ground state. The paper is organized as follows: In Section II , we introduce the main concepts such as the finite temperature autocorrelation function from which the lifetime of an excitation can be extracted, and the Lanczos iteration for calculating the autocorrelation function. We also introduce our algorithm for evaluating the Lanczos series approximately using tensor networks. We close this section by reviewing the concept of ASZMs, which are the central object studied in this paper. In Section III , we introduce the Kitaev-Hubbard chain: a toy model exhibiting an extended topological phase in its ground-state phase-diagram. The Majorana edge modes appearing in this topological phase serve as a perfect test for our algorithm. Finally in Section IV , we present the numerical findings for the lifetime of these Majorana edge modes at various parameter points in the topological phase of the Kitaev-Hubbard chain. We close this paper with Section V that summarizes our findings and comments on possible extensions to other systems such as parafermions [ 32 , 33 , 34 ] , Floquet circuits [ 20 , 22 , 35 , 36 , 37 ] and number conserving realizations of MZMs [ 38 , 39 , 40 , 34 , 41 , 42 , 43 , 44 , 45 ] . II Definitions II.1 Autocorrelation functions at finite temperatures We define the lifetime of an excitation generated by the operator \hat{O} by

Read the complete text