Integrable 3-site, tilted, extended Bose-Hubbard model with nearest-neighbour interactions
Jon LINKS · 2025
Read Integrable 3-site, tilted, extended Bose-Hubbard model with nearest-neighbour interactions by Jon LINKS (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-2506-21789.txt.
Introduction
Jon Links shows in 2025 that a small extended Bose-Hubbard model can be solved exactly for a set choice of its couplings. Such models describe cold atoms with long range dipole forces. The model here has three sites, a tilt, and links between near neighbors.
Exact solvability is rare for these models, so a matched parameter set is a prize. He builds a Bethe ansatz method for the model, a classic tool for exact work. The method yields closed expressions for the full energy spectrum and for the states themselves. Exact cases serve as checks for numeric work on larger cold atom systems.
Catalog record
| CREATOR | Jon LINKS |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 22 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-21789
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Integrable 3-site, tilted, extended Bose-Hubbard model with nearest-neighbour interactions Jon LINKS Abstract Extended Bose-Hubbard models have been employed in the study of cold-atom systems with dipolar interactions. It is shown that, for a certain choice of the coupling parameters, there exists an an integrable extended 3-site Bose-Hubbard model with nearest-neighbour interactions. A Bethe Ansatz procedure is developed to obtain expressions for the energy spectrum and eigenstates. keywords Bose-Hubbard model; quantum integrability; Bethe Ansatz AMS 17B80; 81R12 † † shorttitle: Integrable 3-site, extended Bose-Hubbard model † † runningauthor: J. Links † † address: School of Mathematics and Physics, The University of Queensland, 4072, Australia † † email: [email protected] † † url: https://people.smp.uq.edu.au/JonLinks/ 1 Introduction There is consensus that the Bose-Hubbard model on three or more sites is not an integrable system, due to the display of chaotic behaviours [ 4 , 9 , 15 , 16 , 17 , 23 , 25 ] . Extended Bose-Hubbard models, which accommodate quadratic number operator interaction terms between different sites, have received attention for their role in modelling cold-atom systems with dipolar interactions [ 8 , 18 , 19 , 26 ] . From the mathematical perspective, these extended models also open avenues for constructing integrable generalisations of the Bose-Hubbard model. A broad class of integrable, extended Bose-Hubbard models associated with complete bipartite graphs was formulated in [ 32 ] . Within this class there is a 3-site model, first studied in [ 28 ] with emphasis on the response to an integrability-breaking tilting of the potential. This model has been studied further to characterise the interface between quantum chaos and integrability [ 6 , 7 , 29 , 30 ] , and in relation to entanglement generation [ 27 , 31 ] . One of the distinctive features of the model of [ 28 ] is that it has open boundary conditions with respect to the tunneling terms (i.e, no tunneling between sites 1 and 3), yet closed boundary conditions with respect to the intersite quadratic number operators interactions (i.e., site 1 couples to site 3). While it is feasible to engineer a potential to simulate such a Hamiltonian, as discussed in [ 27 , 28 ] , it is of complementary interest to consider integrable Hamiltonians where the interactions are restricted to being on-site or nearest-neighbour only. In this work it will be demonstrated how this can be achieved for a 3-site system. Remarkably, the construction provides for the inclusion of a tilting of the potential that does not break integrability. However there is a compensation to be paid, viz. the model is not homogeneous with respect to the on-site interactions strength. On the other hand, such a property can in principle be accommodated due to the dipolar properties of the constituent particles [ 8 , 18 , 19 , 26 ] . The integrable Hamiltonian is introduced in Sect. 2. It is demonstrated how the Hamiltonian may be formulated via the o(4)=o(3)\oplus o(3) Lie algebra, realised by canonical boson operators. Through this construction the conserved operators for the Hamiltonian are identified. It is also instructive to view this approach through the lens of the representation theory of o(4) , in order to set up a Bethe Ansatz solution. In Sect. 3 an explicit basis is chosen for the Fock space, in a manner that facilitates the derivation of the Bethe Ansatz equations. The roots of these equations charaterise both the energy spectrum and the eigenstates of the system. Concluding remarks are offered in Sect. 4, including comments regarding the completeness of the Bethe Ansatz solution. The Appendix contains some technical calculations required for deriving the Bethe Ansatz results. 2 Hamiltonian and symmetries Let \{b_{j},\,b_{j}^{\dagger}:j=1,\,2,\,3\} denote bosonic anniihilation and creation operators satisfying the canonical commutation relations [b_{k},b_{k}^{\dagger}]=\delta_{jk}I,\qquad[b_{j},\,b_{k}]=[b^{\dagger}_{j},\,b_{k}^{\dagger}]=0, where I denotes the identity operator. Set N_{j}=b^{\dagger}_{j}b_{j} and \displaystyle N=N_{1}+N_{2}+N_{3}. The Hamiltonian for a tilted, 3-site, extended Bose-Hubbard model with nearest-neighbour interactions has the general form \displaystyle H \displaystyle=U_{1}N_{1}^{2}+U_{2}N_{2}^{2}+U_{3}N_{3}^{2}+U_{12}N_{1}N_{2}+U_{23}N_{2}N_{3} \displaystyle\qquad\quad+\mu_{1}N_{1}+\mu_{3}N_{3}+{\mathcal{E}}_{12}(b_{1}^{\dagger}b_{2}+b_{2}^{\dagger}b_{1})+{\mathcal{E}}_{23}(b_{2}^{\dagger}b_{3}+b_{3}^{\dagger}b_{2}). (2.1) The Hamiltonian acts on the Fock space {\mathcal{F}} spanned by the basis vectors \displaystyle|l,m,n\rangle=(b_{1}^{\dagger})^{l}(b_{2}^{\dagger})^{m}(b_{3}^{\dagger})^{n}|0\rangle, where |0\rangle denotes the Fock vacuum. For the manipulations below it is necessary to work with a particular choice of non-normalised Fock vectors; this is the reason for the omission of normalisation coefficients. Extended Bose-Hubbard Hamiltonians have been been studied in recent times as models for systems comprised of dipolar particles [ 8 , 18 , 19 , 26 ] . Progress in experimental techniques offers a remarkable level of control over dipolar systems via the capacity to tune the Hamitonian coupling parameters through the dipolar interactions. For the 3-site case, a schematic representation is provided in Fig. 1 . While the figure represents the trapping potential as a two-dimensional image, it is important the emphasise that in a physical setup each site is accommodated by an ellipsoidal, three-dimensional, localised potential. The on-site dipole interaction, along with on-site contact interactions, are represented by the couplings U_{1} , U_{2} , U_{3} . These can be varied from negative through to positive values, by changing the shape of the ellipsoid from prolate to oblate [ 14 ] . In addition to influencing interactions on-site, the collective dipole at each site induces interactions between sites, represented by the couplings U_{12} , U_{23} . It is this level of control in adjusting the interaction parameters that renders dipolar systems as candidates for realising integrable systems, whereby the interaction parameters need to be finely tuned to reach integrable limits of the system. It is straightforward to verify that ( 2.1 ) conserves the total particle number N , viz. that [H,\,N]=0 holds. Consequently there is the decomposition \displaystyle{\mathcal{F}}=\bigoplus_{n=0}^{\infty}{\mathcal{F}}_{\mathcal{N}} where {\mathcal{F}}_{\mathcal{N}}={\rm span}\{|l,m,n\rangle:l+m+n={\mathcal{N}}\} . The dimensions of the components in the decomposition are given by the triangular numbers {\mathcal{T}}_{\mathcal{N}} through \displaystyle{\rm dim}({\mathcal{F}}_{\mathcal{N}})={\mathcal{T}}_{{\mathcal{N}}+1}=\frac{({\mathcal{N}}+1)({\mathcal{N}}+2)}{2}. Below it will be shown that the Hamiltonian possesses an additional conserved operator under the constraints \displaystyle{\mathcal{E}}_{12} \displaystyle={\mathcal{E}}_{23}={\mathcal{E}} (2.2) \displaystyle U_{1} \displaystyle=2U_{2}=U_{3}=U_{12}=U_{23}=2U, (2.3) \displaystyle\mu_{3} \displaystyle=-\mu_{1}=\mu. (2.4) \mu_{1} \mu_{3} {\mathcal{E}}_{12} U_{12} {\mathcal{E}}_{23} U_{23} U_{1} U_{2} U_{3} Figure 1: Schematic representation for a system of dipolar bosons tunneling in a tilted 3-site potential. The narrow profile of the potential at site 2 leads to attractive on-site dipole interaction, such that U_{2} may be tuned to be a lower value than U_{1} and U_{3} . In order to expose the symmetries underlying this specialisation of ( 2.1 ) set \displaystyle e_{1} \displaystyle=\sqrt{2}(b_{1}^{\dagger}b_{2}+b_{2}^{\dagger}b_{3}), \displaystyle h_{1} \displaystyle=2(N_{1}-N_{3}), \displaystyle f_{1} \displaystyle=\sqrt{2}(b_{2}^{\dagger}b_{1}+b_{3}^{\dagger}b_{2}), \displaystyle e_{2} \displaystyle=\frac{1}{2}b_{2}^{\dagger}b_{2}^{\dagger}-b_{1}^{\dagger}b_{3}^{\dagger}, \displaystyle h_{2} \displaystyle=\frac{1}{2}(2N+3I), \displaystyle f_{2} \displaystyle=b_{1}b_{3}-\frac{1}{2}b_{2}b_{2}. These operators realise the o(4) Lie algebra by satisfying the commutation relations \displaystyle[e_{j},\,f_{k}] \displaystyle=\delta_{jk}h_{j}, \displaystyle[h_{j},\,e_{k}] \displaystyle=2\delta_{jk}e_{k}, \displaystyle[h_{j},\,f_{k}] \displaystyle=-2\delta_{jk}f_{k}. It is found that the Hamiltonian ( 2.1 ) subject to the constraints ( 2.2 ), ( 2.3 ), ( 2.4 ) is expressible as \displaystyle H \displaystyle=H_{1}+H_{2}, (2.5) \displaystyle H_{1} \displaystyle=\frac{U}{4}h_{1}^{2}-\frac{\mu}{2}h_{1}+\frac{{\mathcal{E}}}{\sqrt{2}}(e_{1}+f_{1}), \displaystyle H_{2} \displaystyle=U\left(h_{2}-\frac{3}{2}I\right)^{2}. Under this realisation, the corresponding o(4) Casimir invariants assume the form \displaystyle C_{1} \displaystyle=\frac{1}{2}h_{1}^{2}+e_{1}f_{1}+f_{1}e_{1} \displaystyle=2N^{2}+2N-8N_{1}N_{3}-2N_{2}^{2}+2N_{2}+4b_{2}^{\dagger}b_{2}^{\dagger}b_{1}b_{3}+4b_{1}^{\dagger}b_{3}^{\dagger}b_{2}b_{2}, \displaystyle C_{2} \displaystyle=\frac{1}{2}h_{2}^{2}+e_{2}f_{2}+f_{2}e_{2} \displaystyle=\frac{1}{4}C_{1}-\frac{3}{8}I. This realisation of the Casimir invariants provides the third independent conserved operator required to claim integrability of ( 2.5 ). 3 Bethe Ansatz solution To derive the Bethe Ansatz solution of the model, differential operator methods are adapted from [ 20 ] . Before proceeding to that task, it is first necessary to completely describe the o(4) action