Convergence of the Zamolodchikov-Faddeev expansion: local operators affiliated with double-cone algebras in 1+1-dimensional integrable quantum field theories with non-constant scattering function

In a 1+1-dimensional quantum field theory with factorizing S-matrix, local observables in a spacetime double cone are characterized by their Zamolodchikov-Faddeev (ZF) expansion: an infinite series in the interacting creation and annihilation operators whose coefficient functions are boundary values of meromorphic functions F_k constrained by S-symmetry, S-periodicity, a kinematic-pole residue recursion, and Paley-Wiener-type bounds. For the constant scattering function S = -1 (massive Ising model) such expansions are known to converge to closable operators affiliated with the local von Neumann algebra, but for every non-constant S - the sinh-Gordon model, whose form factors have been explicitly known since 1993, being the canonical case - the convergence of the series is an open problem. The problem is to establish (or rigorously refute), for at least one non-constant scattering function, that the formal ZF expansion with an admissible form-factor sequence defines a bona fide local observable: a closable operator on a dense domain whose closure is affiliated with the double-cone algebra. This requires verifying large-order norm bounds on the associated singular integral kernels and the domain-invariance condition needed for affiliation, both of which have been verified to date only for S = -1.

Background

An integrable quantum field theory in 1+1 spacetime dimensions is a model in which scattering is purely elastic and factorized: no particles are created or destroyed, and every multi-particle scattering process decomposes into two-particle exchanges. All information about the interaction is then contained in the two-particle scattering function $S(\theta)$ , a meromorphic function of the rapidity difference $\theta$ , where the rapidity $\theta$ parametrizes a particle's energy-momentum as $p(\theta)=\mu(\cosh\theta,\sinh\theta)$ . The prototype of a nontrivial $S$ is the sinh-Gordon scattering function

S(\theta)=\frac{\sinh\theta-i\sin b}{\sinh\theta+i\sin b},\qquad 0<b<\pi,

while $S=1$ is the free theory and $S=-1$ the massive Ising model.

To construct such a model rigorously, one replaces the usual creation and annihilation operators by Zamolodchikov-Faddeev (ZF) operators $z^\dagger(\theta)$ , $z(\theta)$ , which satisfy an $S$ -deformed version of the canonical commutation relations,

z(\theta_1)z(\theta_2)=S(\theta_1-\theta_2)\,z(\theta_2)z(\theta_1),\qquad z(\theta_1)z^\dagger(\theta_2)=S(\theta_2-\theta_1)\,z^\dagger(\theta_2)z(\theta_1)+\delta(\theta_1-\theta_2)\mathbf 1,

and act on an $S$ -symmetric Fock space. The field $\phi(x)=\int d\theta\,(e^{ip(\theta)\cdot x}z^\dagger(\theta)+e^{-ip(\theta)\cdot x}z(\theta))$ is not local, but it commutes with its spacetime-reflected field $\phi'(x)=U(j)\phi(-x)U(j)$ whenever the two arguments are spacelike separated in the appropriate left/right order; it is therefore said to be localized in an infinitely extended spacelike wedge region. Taking von Neumann algebras (algebras of bounded operators closed in the weak operator topology) generated by bounded functions of the smeared fields $\phi(f)$ with $\mathrm{supp}\,f$ in a wedge, one obtains wedge algebras, and the observable algebra of a bounded region - the standard choice being a double cone $O_r$ , the intersection of a left-opening and a right-opening wedge whose tips are separated by $2r$ - is defined as the intersection $\mathfrak A(O_r)=\mathfrak A(W')\cap\mathfrak A(W)$ . Lechner (2008) proved that for a large class of scattering functions, including sinh-Gordon, this yields a fully consistent local quantum field theory whose local algebras are highly nontrivial; however, his argument uses abstract operator-algebraic machinery (Tomita-Takesaki modular theory and nuclearity conditions) and yields no explicit form for the operators in $\mathfrak A(O_r)$ - only their abstract existence.

To describe local operators explicitly, Bostelmann and Cadamuro (2013, 2015) developed an operator expansion: every operator with a suitable high-energy bound - more precisely, every quadratic form (a sesquilinear expression $\langle\psi,A\chi\rangle$ on a dense domain, which need not derive from an operator) carrying such a bound - has, in the sense of quadratic forms, an expansion in a series of normally ordered ZF monomials,

A=\sum_{m,n\ge 0}\int\frac{d^m\theta\,d^n\eta}{m!\,n!}\,f_{mn}(\theta,\eta)\,z^\dagger(\theta_1)\cdots z^\dagger(\theta_m)\,z(\eta_1)\cdots z(\eta_n),

and $A$ is localized in the double cone $O_r$ precisely when the coefficients are boundary values of meromorphic functions $F_k$ on $\mathbb C^k$ that (i) are $S$ -symmetric under exchange of adjacent arguments, (ii) are $S$ -periodic under shifts of any argument by $2\pi i$ , (iii) have at most first-order poles at the kinematic points $\zeta_n-\zeta_m=i\pi$ whose residues are fixed recursively in terms of $F_{k-2}$ , and (iv) satisfy $r$ -dependent growth bounds of Paley-Wiener type (Paley-Wiener bounds express the localization radius through the exponential growth of the analytic continuation). These conditions are the form-factor equations of the bootstrap form-factor program, whose form factors are the matrix elements of local fields between multiparticle states.

The difficulty is convergence. Because the residue recursion relates $F_k$ to $F_{k-2}$ , the sequence $F_k$ cannot terminate except for very special $S$ , so the operator series has infinitely many terms. For $S=-1$ the construction goes through: even Ising operators reduce to finite sums, and for the odd (order/disorder) operators, whose form factors involve products of $\tanh((\zeta_i-\zeta_j)/2)$ , convergence of the series to a closable operator was proven for test functions in a Jaffe class (smooth functions whose Fourier transforms decay faster than any exponential), and the resulting closed operators are affiliated with $\mathfrak A(O_r)$ - affiliation being the correct sense in which an unbounded operator belongs to a von Neumann algebra: its polar part and all its spectral projections lie in the algebra, equivalently it commutes with the algebra's commutant (the algebra of all bounded operators commuting with it) on its domain. For non-constant $S$ - e.g. sinh-Gordon, where admissible $F_k$ sequences have been known in principle since Fring, Mussardo and Simonetti (1993) - no such convergence result exists: it is not known under which analytic or growth conditions on the $F_k$ or on the spacetime smearings the formal series defines a bona fide operator rather than just a quadratic form. Bostelmann and Cadamuro (2019) reduced the problem to concrete criteria - a summability condition on the weighted norms of the expansion kernels $f_{mn}$ (boundary values of the $F_{m+n}$ , carrying the kinematic-pole singularities), plus a domain-invariance condition needed for affiliation - and verified them only for $S=-1$ . A related open aspect is that even where single operators exist, a common invariant dense domain for several such operators (needed for their operator products and for Wightman $n$ -point functions, i.e. vacuum expectation values of products of fields) is not automatic.

Problem Statement

The problem is the convergence of the Zamolodchikov-Faddeev operator expansion for local observables in 1+1-dimensional integrable quantum field theories with non-constant scattering function, as posed in Bostelmann's summary of the Bostelmann-Cadamuro program (arXiv:1305.1463) and left open by Bostelmann-Cadamuro, Ann. Henri Poincare 20 (2019) 3889.

Let $S$ be a non-constant two-particle scattering function satisfying the standard regularity assumptions of the program: meromorphic continuation, analyticity and boundedness in the physical strip $0<\mathrm{Im}\,\zeta<\pi$ , and the symmetry relations $S(\zeta)^{-1}=S(-\zeta)=\overline{S(\bar\zeta)}=S(\zeta+i\pi)$ , with the sinh-Gordon function $S(\theta)=(\sinh\theta-i\sin b)/(\sinh\theta+i\sin b)$ , $0<b<\pi$ , as the canonical instance. Let $\{F_k\}$ be an admissible sequence of meromorphic functions satisfying the form-factor equations: $S$ -symmetry, $S$ -periodicity, the kinematic-pole residue recursion relating $F_k$ to $F_{k-2}$ , and the $r$ -dependent Paley-Wiener-type bounds. Consider the formal series

A=\sum_{m,n\ge 0}\int\frac{d^m\theta\,d^n\eta}{m!\,n!}\,f_{mn}(\theta,\eta)\,z^\dagger(\theta_1)\cdots z^\dagger(\theta_m)\,z(\eta_1)\cdots z(\eta_n),

where $f_{mn}$ are the boundary values $F_{m+n}(\theta_1+i0,\ldots,\theta_m+i0,\eta_1+i\pi-i0,\ldots,\eta_n+i\pi-i0)$ , with the localization radius $r$ fixed by the Paley-Wiener-type bounds and the smearing supported in the double cone $O_r$ .

Prove or disprove: for at least one such non-constant $S$ , the series defines a bona fide local observable in the precise sense of the program, namely both of the following hold, together with the additional requirement that the resulting operator be nonzero (equivalently, that the form-factor sequence $\{F_k\}$ not be identically zero).

(1) Closability: the series converges to a closable operator on a dense domain - concretely, the weighted norms of the expansion kernels $f_{mn}$ (boundary values of the $F_{m+n}$ , carrying the kinematic-pole singularities) satisfy, at all orders, the summability criterion of Bostelmann-Cadamuro (2019, Proposition 3.3) or an equivalent stated sufficient condition, so that the particle-number cutoff and energy damping can be removed from at least one side of the matrix elements.

(2) Affiliation: the closure of this operator is affiliated with the double-cone von Neumann algebra $\mathfrak A(O_r)=\mathfrak A(W')\cap\mathfrak A(W)$ - in particular, the domain-invariance condition under the commutant of $\mathfrak A(O_r)$ (condition (3.10) of Bostelmann-Cadamuro 2019) holds, established to date only for $S=-1$ .

A positive resolution may use any admissible $F_k$ sequence (the Fring-Mussardo-Simonetti sinh-Gordon sequence being the natural candidate, but newly constructed sequences are admissible), averaged (smeared) fields, and any dense domain consistent with the above, provided locality in $O_r$ and affiliation with $\mathfrak A(O_r)$ are rigorously established. A negative resolution is a rigorous no-go result: a proof that for every non-constant $S$ , no admissible $F_k$ sequence renders the series a nonzero closable operator whose closure is affiliated with $\mathfrak A(O_r)$ ; the failure of a single candidate sequence, or of every admissible sequence for one particular $S$ , does not constitute a refutation, since the positive branch permits newly constructed admissible sequences. Results restricted to constant $S$ ( $S=\pm 1$ ), to wedge (unbounded) localization, or to two-point functions do not constitute a solution. The common objective of the two requested results is to determine whether the formal ZF expansion of local observables converges at all beyond the Ising case; establishing a positive result would additionally address the linked domain question of sufficient conditions for common invariant dense domains on which products of such local operators and their Wightman $n$ -point functions are well-defined.

References

H. Bostelmann, 'Relativistische Materie in zwei Raum-Zeit-Dimensionen', arXiv:1305.1463 (2013, v2 2015), DOI 10.48550/arXiv.1305.1463, https://arxiv.org/abs/1305.1463
G. Lechner, 'Construction of Quantum Field Theories with Factorizing S-Matrices', Commun. Math. Phys. 277 (2008) 821-860, DOI 10.1007/s00220-007-0381-5, https://doi.org/10.1007/s00220-007-0381-5
H. Bostelmann, D. Cadamuro, 'An operator expansion for integrable quantum field theories', J. Phys. A 46 (2013) 095401, DOI 10.1088/1751-8113/46/9/095401, https://arxiv.org/abs/1208.4763
H. Bostelmann, D. Cadamuro, 'Characterization of Local Observables in Integrable Quantum Field Theories', Commun. Math. Phys. 337 (2015) 1199-1240, DOI 10.1007/s00220-015-2294-z, https://arxiv.org/abs/1402.6127
D. Cadamuro, 'A Characterization Theorem for Local Operators in Factorizing Scattering Models', PhD thesis, Universitaet Goettingen (2012), arXiv:1211.3583, https://arxiv.org/abs/1211.3583
H. Bostelmann, D. Cadamuro, 'Towards an Explicit Construction of Local Observables in Integrable Quantum Field Theories', Ann. Henri Poincare 20 (2019) 3889-3926, DOI 10.1007/s00023-019-00847-7, https://arxiv.org/abs/1806.00269
H. Bostelmann, 'On the status of pointlike fields in integrable QFTs', J. Phys. Conf. Ser. 1416 (2019) 012006, DOI 10.1088/1742-6596/1416/1/012006, https://arxiv.org/abs/2001.00538
D. Cadamuro, 'Direct construction of pointlike observables in the Ising model', arXiv:1806.01686 (2018), https://arxiv.org/abs/1806.01686
S. Alazzawi, G. Lechner, 'Inverse Scattering and Local Observable Algebras in Integrable Quantum Field Theories', Commun. Math. Phys. 354 (2017) 913-956, DOI 10.1007/s00220-017-2891-0, https://arxiv.org/abs/1608.02359
D. Cadamuro, Y. Tanimoto, 'Wedge-Local Fields in Integrable Models with Bound States', Commun. Math. Phys. 340 (2015) 661-697, DOI 10.1007/s00220-015-2448-z, https://arxiv.org/abs/1502.01313
A. Fring, G. Mussardo, P. Simonetti, 'Form factors for integrable Lagrangian field theories, the sinh-Gordon model', Nucl. Phys. B 393 (1993) 413-441, DOI 10.1016/0550-3213(93)90252-K, https://doi.org/10.1016/0550-3213(93)90252-K
H. M. Babujian, M. Karowski, 'Towards the construction of Wightman functions of integrable quantum field theories', Int. J. Mod. Phys. A 19 (2004) 34-49, DOI 10.1142/S0217751X04020294, https://arxiv.org/abs/hep-th/0301088
K. K. Kozlowski, 'On convergence of form factor expansions in the infinite volume quantum Sinh-Gordon model in 1+1 dimensions', Invent. Math. 233 (2023) 725-827, DOI 10.1007/s00222-023-01192-7, https://arxiv.org/abs/2007.01740
K. K. Kozlowski, A. Simon, 'On multipoint correlation functions in the Sinh-Gordon 1+1 dimensional quantum field theory', arXiv:2502.03894 (2025), https://arxiv.org/abs/2502.03894
K. K. Kozlowski, A. Simon, 'Wightman axiomatics of the bootstrap construction of the 1+1 dimensional Sinh-Gordon model', arXiv:2511.10800 (2025), https://arxiv.org/abs/2511.10800
K. K. Kozlowski, A. Simon, 'Form factors of field exponentials in the 1+1 dimensional Sine-Gordon model', arXiv:2608.14365 (2026), https://arxiv.org/abs/2608.14365
M. B. Fröb, D. Cadamuro, 'Local Operators in the Sine-Gordon Model: $\partial_\mu\phi\,\partial_\nu\phi$ and the Stress Tensor', Ann. Henri Poincare 27 (2026) 1343-1405, DOI 10.1007/s00023-025-01565-z, https://arxiv.org/abs/2205.09223
D. Buchholz, S. J. Summers, 'String- and brane-localized causal fields in a strongly nonlocal model', J. Phys. A 40 (2007) 2147-2163, DOI 10.1088/1751-8113/40/9/019, https://doi.org/10.1088/1751-8113/40/9/019