Dynamical Spin Structure Factor (DSF)

• DSF is the momentum- and frequency-resolved spectral density of spin fluctuations that defines how spin correlations are distributed over energy and momentum.
• It directly connects theoretical models to experiments like inelastic neutron scattering by distinguishing between sharp quasiparticle peaks and broad multiparticle continua.
• Computational frameworks such as DMRG, variational Monte Carlo, and parton mean-field methods provide practical insights into DSF behavior across quantum phase transitions.

The dynamical spin structure factor (DSF), also written as DSSF in parts of the literature, is the momentum- and frequency-resolved spectral density of spin fluctuations. It is the central two-point dynamical observable for quantum magnets because it resolves not only where spin correlations are concentrated in momentum space, but also whether the underlying excitations are sharp quasiparticles, broad continua, gapped collective modes, or critical low-energy states. Across ordered antiferromagnets, spin liquids, Kitaev systems, higher-spin chains, bilayer magnets, and quantum spin nematics, the DSF is the quantity most directly connected to inelastic neutron scattering and related spectroscopies, and it is repeatedly used as a diagnostic of magnons, triplons, spinons, Majorana excitations, visons, and collective mode softening near phase transitions (Zhu et al., 2018, Chen et al., 26 Mar 2026, Lohöfer et al., 2015).

1. Definition, representations, and relation to static correlations

In translationally invariant spin systems, the DSF is the space-time Fourier transform of the spin-spin correlation function. A standard tensor definition is

$$S^{\alpha\beta}(\mathbf q,\omega)=\frac{1}{2\pi N}\sum_{i,j}e^{-i\mathbf q\cdot(\mathbf r_i-\mathbf r_j)}\int_{-\infty}^{\infty}dt\,e^{i\omega t}\langle S_i^\alpha(t)S_j^\beta(0)\rangle,$$

with $\alpha,\beta$ labeling spin components (Burkard et al., 20 May 2025). In resolvent form, widely used in correction-vector calculations, the same object is written as

$$\mathcal S^{\alpha\beta}(\mathbf Q,\omega)=-\frac{1}{\pi}\operatorname{Im}\left\langle S^\alpha(-\mathbf Q)\frac{1}{\omega-(H-E_0)+i\eta}S^\beta(\mathbf Q)\right\rangle,$$

where $E_0$ is the ground-state energy and $\eta\to0^+$ is the broadening (Zhu et al., 2018).

The spectral meaning is clearest in Lehmann form. For example, in isotropic chains one may write

$$S^{zz}(k,\omega)= \frac{2\pi}{N}\sum_{\alpha} \left|\langle \psi_\alpha | S^z_{-k} | \psi_0\rangle\right|^2 \delta\!\left(\omega-\omega_\alpha(k)\right),$$

so the DSF directly weights those excited states that are reachable by the spin operator at fixed momentum (Sharma et al., 16 Jun 2025). The observable therefore depends not only on the Hamiltonian but also on the operator content: single-spin operators, adjacent-spin bond operators, singlet bond operators, and density- or spin-channel operators can emphasize different excitation sectors (Klauser et al., 2012, Lohöfer et al., 2015).

The static structure factor is obtained by integrating over frequency,

$S(\mathbf k)=\int d\omega\,S(\mathbf k,\omega),$

and thus discards the time or energy resolution (Halimeh et al., 2016). This distinction is often decisive. Static correlations may indicate where equal-time weight is concentrated, but the DSF determines whether that weight comes from a sharp pole, a threshold singularity, a multi-particle continuum, or a low-energy instability.

2. Spectral interpretation: poles, continua, and operator-dependent selection

In magnetically ordered systems, the DSF shows the textbook signature of coherent spin-1 quasiparticles. For the $\mathbf q=(0,0)$ ordered phase of the kagome Heisenberg model, it exhibits a sharp gapless mode at the ordering wave vector, the largest spectral weight at $M$, and dispersive magnon-like response; at $\mathbf Q=M$, the spectrum is concentrated at $\alpha,\beta$0, with only a weaker high-energy tail that may be related to multi-magnon effects (Zhu et al., 2018). In the square-lattice bilayer Heisenberg model, the antisymmetric channel contains low-energy Goldstone modes in the ordered phase and evolves continuously into the gapped triplon mode of the quantum disordered phase (Lohöfer et al., 2015).

In fractionalized phases, the same observable acquires qualitatively different support. In kagome spin liquids, chiral spin liquids, algebraic spin liquids, and Kitaev spin liquids, the DSF is repeatedly described as broad, continuum-like, and reduced in coherent weight relative to an ordered magnet; in these cases it is interpreted in terms of two-spinon excitations, fractionalized spinon pairs, matter Majoranas, gauge-flux excitations, or their coexistence with collective modes (Zhu et al., 2018, Halimeh et al., 2016, Vörös et al., 2023, Chen et al., 26 Mar 2026). The basic distinction is therefore between a pole-dominated response and a continuum-dominated response.

One-dimensional integrable and near-integrable systems provide exact benchmarks for this interpretation. In the isotropic spin-$\alpha,\beta$1 Heisenberg antiferromagnet, the DSF decomposes into even-spinon sectors, and corrected normalization yields a two-spinon contribution of about $\alpha,\beta$2 and a four-spinon contribution of $\alpha,\beta$3 rather than the much larger values quoted in earlier literature (Bougourzi, 2014). For adjacent bond operators $\alpha,\beta$4, the DSF carries a large weight of 4-spinon excitations, which are distinguishable from the 2-spinon signal because they are located outside the 2-spinon spectrum (Klauser et al., 2012). These results make clear that the same Hamiltonian can display different apparent excitation content depending on which operator defines the DSF.

This operator dependence persists in nonintegrable frustrated chains. In the spin-$\alpha,\beta$5 and higher-spin $\alpha,\beta$6-$\alpha,\beta$7 chains, the DSF is dominated by spinon continua, whereas magnon-like modes appear as resonant branches embedded in those continua rather than as isolated stable quasiparticles (1803.02359, Sharma et al., 16 Jun 2025). A plausible implication is that “magnon versus spinon” is often not a binary distinction in the DSF; a broad spectral continuum can still contain resonant structures inherited from more conventional modes.

3. Computational and analytical frameworks

Because the DSF is a real-frequency many-body observable, its calculation has generated a large methodological literature spanning exact, variational, tensor-network, partonic, quantum Monte Carlo, and high-temperature techniques.

Framework	Representative formulation	Representative systems
DMRG / tDMRG	Correction-vector DSF and real-time evolution with Fourier transform	Kagome Heisenberg model; spin-$\alpha,\beta$8 and spin-$\alpha,\beta$9 $$\mathcal S^{\alpha\beta}(\mathbf Q,\omega)=-\frac{1}{\pi}\operatorname{Im}\left\langle S^\alpha(-\mathbf Q)\frac{1}{\omega-(H-E_0)+i\eta}S^\beta(\mathbf Q)\right\rangle,$$0-$$\mathcal S^{\alpha\beta}(\mathbf Q,\omega)=-\frac{1}{\pi}\operatorname{Im}\left\langle S^\alpha(-\mathbf Q)\frac{1}{\omega-(H-E_0)+i\eta}S^\beta(\mathbf Q)\right\rangle,$$1 chains
Variational Monte Carlo	Gutzwiller-projected particle-hole or particle-hole-like excited states and generalized eigenvalue problems	Spin-$$\mathcal S^{\alpha\beta}(\mathbf Q,\omega)=-\frac{1}{\pi}\operatorname{Im}\left\langle S^\alpha(-\mathbf Q)\frac{1}{\omega-(H-E_0)+i\eta}S^\beta(\mathbf Q)\right\rangle,$$2 $$\mathcal S^{\alpha\beta}(\mathbf Q,\omega)=-\frac{1}{\pi}\operatorname{Im}\left\langle S^\alpha(-\mathbf Q)\frac{1}{\omega-(H-E_0)+i\eta}S^\beta(\mathbf Q)\right\rangle,$$3-$$\mathcal S^{\alpha\beta}(\mathbf Q,\omega)=-\frac{1}{\pi}\operatorname{Im}\left\langle S^\alpha(-\mathbf Q)\frac{1}{\omega-(H-E_0)+i\eta}S^\beta(\mathbf Q)\right\rangle,$$4 chain; SU(3) chain; SU(4) honeycomb model
Parton mean-field, SBMFT, RPA	Schwinger-boson two-spinon DSF; Majorana parton mean field plus RPA	Chiral kagome spin liquids; Kitaev spin liquids with $$\mathcal S^{\alpha\beta}(\mathbf Q,\omega)=-\frac{1}{\pi}\operatorname{Im}\left\langle S^\alpha(-\mathbf Q)\frac{1}{\omega-(H-E_0)+i\eta}S^\beta(\mathbf Q)\right\rangle,$$5
QMC and analytic continuation	Matsubara correlators continued to real frequency	Square-lattice bilayer Heisenberg model
Exact or integrable form-factor methods	Bethe Ansatz, quantum-group, determinant formulas, exact spinon decompositions	Heisenberg and XXZ chains
High-temperature expansion	Dyn-HTE moments and continued-fraction reconstruction	Frustrated Heisenberg magnets in 1D, 2D, and 3D

In large-scale DMRG for kagome antiferromagnets, the DSF is evaluated directly from the ground state and the correction-vector response on cylinders with open $$\mathcal S^{\alpha\beta}(\mathbf Q,\omega)=-\frac{1}{\pi}\operatorname{Im}\left\langle S^\alpha(-\mathbf Q)\frac{1}{\omega-(H-E_0)+i\eta}S^\beta(\mathbf Q)\right\rangle,$$6- and periodic $$\mathcal S^{\alpha\beta}(\mathbf Q,\omega)=-\frac{1}{\pi}\operatorname{Im}\left\langle S^\alpha(-\mathbf Q)\frac{1}{\omega-(H-E_0)+i\eta}S^\beta(\mathbf Q)\right\rangle,$$7-boundary conditions (Zhu et al., 2018). In higher-spin frustrated chains, the DSF is obtained by time-dependent DMRG: the ground state is computed in MPS form, a local spin operator is applied at the chain center, the state is evolved under $$\mathcal S^{\alpha\beta}(\mathbf Q,\omega)=-\frac{1}{\pi}\operatorname{Im}\left\langle S^\alpha(-\mathbf Q)\frac{1}{\omega-(H-E_0)+i\eta}S^\beta(\mathbf Q)\right\rangle,$$8, and the resulting correlator is Fourier transformed in space and time; to reduce ringing from finite-time evolution, the correlations are multiplied by a Gaussian filter $$\mathcal S^{\alpha\beta}(\mathbf Q,\omega)=-\frac{1}{\pi}\operatorname{Im}\left\langle S^\alpha(-\mathbf Q)\frac{1}{\omega-(H-E_0)+i\eta}S^\beta(\mathbf Q)\right\rangle,$$9 with $E_0$0 (Sharma et al., 16 Jun 2025).

Variational approaches construct an explicit truncated excitation basis. In the spin-$E_0$1 $E_0$2-$E_0$3 chain, projected particle-hole states provide a variational two-spinon basis, and the method avoids analytic continuation from imaginary time while using sign-problem-free Monte Carlo sampling (1803.02359). Closely related projected particle-hole constructions are used for the SU(3) Heisenberg chain and the SU(4) honeycomb algebraic spin liquid, where the DSF is extracted from overlap and Hamiltonian matrices in a Gutzwiller-projected excited-state basis (Vörös et al., 2021, Vörös et al., 2023).

Partonic methods compute the DSF from fractionalized quasiparticles and their collective corrections. Schwinger-boson mean-field theory expresses the kagome DSF as a sum over two-spinon creation channels with matrix elements built from Bogoliubov amplitudes (Halimeh et al., 2016). In the Kitaev $E_0$4–$E_0$5 model, a self-consistent Majorana parton mean field combined with RPA yields a mean-field two-particle continuum corrected by collective paramagnon-like modes (Chen et al., 26 Mar 2026).

For unbiased finite-temperature dynamics, QMC and high-temperature methods are prominent. In the bilayer Heisenberg model, Matsubara-frequency spin correlations are analytically continued by stochastic analytic continuation (Lohöfer et al., 2015). Dyn-HTE instead expands the Matsubara correlator directly in powers of $E_0$6, converts the result into even frequency moments, and reconstructs the spectrum through a continued-fraction representation of the relaxation function in the thermodynamic limit (Burkard et al., 20 May 2025).

4. Frustration, spin liquids, and fractionalization signatures

The kagome spin-$E_0$7 Heisenberg antiferromagnet provides a canonical DSF fingerprint of fractionalization. At the pure nearest-neighbor point, the DSF shows dominant intensity concentrated near the $E_0$8 point, spectral weight spread along the boundary of the extended Brillouin zone, a broad continuum over a wide energy range, strong low-energy weight at $E_0$9 without a sharp magnon pole, and a long high-frequency tail extending to $\eta\to0^+$0 in the paper’s units (Zhu et al., 2018). The same work argues that boundary-condition sensitivity under weak out-of-plane Dzyaloshinskii–Moriya coupling is more consistent with a gapless spin liquid or a very small gap than with a robust gapped phase, and it compares the resulting DSF directly with the low-energy $\eta\to0^+$1-point enhancement and broad high-energy continuum observed in herbertsmithite.

For gapped chiral spin liquids on kagome, the DSF also resolves broken time-reversal symmetry in a way the equal-time structure factor cannot. In Schwinger-boson mean-field theory, fixed-frequency $\eta\to0^+$2 for the cuboc1 state loses $\eta\to0^+$3 symmetry and reduces the apparent six-fold rotational symmetry around $\eta\to0^+$4 to three-fold symmetry, while the onset of the two-spinon continuum is relatively flat along $\eta\to0^+$5-M-K-$\eta\to0^+$6 (Halimeh et al., 2016). The DSF thus functions not merely as a detector of fractionalization, but as a momentum-resolved probe of chirality.

Kitaev materials exhibit a related but distinct structure. In the pure Kitaev spin liquid, the spin DSF is mostly a broad continuum associated with fractionalized Majorana excitations and local flux creation. With a third-nearest-neighbor Heisenberg perturbation $\eta\to0^+$7, however, the response develops low-energy coherent paramagnon-like collective modes below the high-energy Majorana continuum; these modes sharpen and soften with increasing $\eta\to0^+$8, and their gap closing at specific momenta signals the instability of the Kitaev spin liquid toward magnetic order (Chen et al., 26 Mar 2026). In the non-Abelian phase with quenched vortex disorder, the DSF at $\eta\to0^+$9 instead develops a pronounced peak centered at the flux gap, interpreted as a signature of vortex-bound Majorana zero modes whose disorder-averaged spectral line has a non-Lorentzian tail (Otten et al., 2018).

Algebraic and Dirac spin liquids provide yet another DSF morphology. In the SU(4) honeycomb model with a $$S^{zz}(k,\omega)= \frac{2\pi}{N}\sum_{\alpha} \left|\langle \psi_\alpha | S^z_{-k} | \psi_0\rangle\right|^2 \delta\!\left(\omega-\omega_\alpha(k)\right),$$0-flux Dirac spin liquid ground state, both a projected variational treatment and a free-parton mean-field calculation yield a gapless continuum rather than sharp magnons, with low-energy “towers” centered at $$S^{zz}(k,\omega)= \frac{2\pi}{N}\sum_{\alpha} \left|\langle \psi_\alpha | S^z_{-k} | \psi_0\rangle\right|^2 \delta\!\left(\omega-\omega_\alpha(k)\right),$$1 and $$S^{zz}(k,\omega)= \frac{2\pi}{N}\sum_{\alpha} \left|\langle \psi_\alpha | S^z_{-k} | \psi_0\rangle\right|^2 \delta\!\left(\omega-\omega_\alpha(k)\right),$$2, higher-energy towers around $$S^{zz}(k,\omega)= \frac{2\pi}{N}\sum_{\alpha} \left|\langle \psi_\alpha | S^z_{-k} | \psi_0\rangle\right|^2 \delta\!\left(\omega-\omega_\alpha(k)\right),$$3, and a Gutzwiller-induced redistribution of weight from higher to lower energies that emphasizes the lower edge of the continuum (Vörös et al., 2023). The mean-field local response obeys $$S^{zz}(k,\omega)= \frac{2\pi}{N}\sum_{\alpha} \left|\langle \psi_\alpha | S^z_{-k} | \psi_0\rangle\right|^2 \delta\!\left(\omega-\omega_\alpha(k)\right),$$4, while equal-time correlations decay as $$S^{zz}(k,\omega)= \frac{2\pi}{N}\sum_{\alpha} \left|\langle \psi_\alpha | S^z_{-k} | \psi_0\rangle\right|^2 \delta\!\left(\omega-\omega_\alpha(k)\right),$$5, consistent with an algebraic spin liquid.

5. Quantum phase transitions, incommensurability, and collective-mode softening

One of the strongest uses of the DSF is to follow the evolution of excitation content across phase boundaries. In kagome antiferromagnets, tuning from a chiral spin liquid to the $$S^{zz}(k,\omega)= \frac{2\pi}{N}\sum_{\alpha} \left|\langle \psi_\alpha | S^z_{-k} | \psi_0\rangle\right|^2 \delta\!\left(\omega-\omega_\alpha(k)\right),$$6 ordered phase causes the finite-frequency $$S^{zz}(k,\omega)= \frac{2\pi}{N}\sum_{\alpha} \left|\langle \psi_\alpha | S^z_{-k} | \psi_0\rangle\right|^2 \delta\!\left(\omega-\omega_\alpha(k)\right),$$7-point peak of the chiral phase, interpreted as a two-spinon resonance, to move downward in energy and grow in intensity until it lands at $$S^{zz}(k,\omega)= \frac{2\pi}{N}\sum_{\alpha} \left|\langle \psi_\alpha | S^z_{-k} | \psi_0\rangle\right|^2 \delta\!\left(\omega-\omega_\alpha(k)\right),$$8 and becomes the magnon mode of the ordered state (Zhu et al., 2018). This is interpreted there as a spinon-pair condensation mechanism for the quantum phase transition.

In frustrated half-integer spin chains, the DSF resolves a more intricate interplay of dimerization, incommensurability, and fractionalization. For spin-$$S^{zz}(k,\omega)= \frac{2\pi}{N}\sum_{\alpha} \left|\langle \psi_\alpha | S^z_{-k} | \psi_0\rangle\right|^2 \delta\!\left(\omega-\omega_\alpha(k)\right),$$9 and spin-$S(\mathbf k)=\int d\omega\,S(\mathbf k,\omega),$0 $S(\mathbf k)=\int d\omega\,S(\mathbf k,\omega),$1-$S(\mathbf k)=\int d\omega\,S(\mathbf k,\omega),$2 Heisenberg chains, the DSF shows a de Cloizeaux–Pearson-like continuum at low $S(\mathbf k)=\int d\omega\,S(\mathbf k,\omega),$3, develops a gap and structured low-energy boundaries in the partially dimerized phase, and becomes gapless again at larger $S(\mathbf k)=\int d\omega\,S(\mathbf k,\omega),$4 with incommensurate minima; in both cases magnons appear as resonances inside the spinon continuum, and the spinon gap has a nonmonotonic dome shape that closes at the phase boundaries, suggesting a floating phase produced by condensation of incommensurate spinons (Sharma et al., 16 Jun 2025). In the spin-$S(\mathbf k)=\int d\omega\,S(\mathbf k,\omega),$5 $S(\mathbf k)=\int d\omega\,S(\mathbf k,\omega),$6-$S(\mathbf k)=\int d\omega\,S(\mathbf k,\omega),$7 chain, the low-energy weight shifts away from $S(\mathbf k)=\int d\omega\,S(\mathbf k,\omega),$8 toward $S(\mathbf k)=\int d\omega\,S(\mathbf k,\omega),$9 for $\mathbf q=(0,0)$0, directly reflecting frustration-induced rearrangement of low-energy spinon minima (1803.02359).

Soft-mode diagnostics are equally important in higher dimensions. In the square-lattice bilayer Heisenberg model, the antisymmetric DSF provides a continuous spectral connection between ordered-state magnons and disordered-state triplons across the quantum critical point at $\mathbf q=(0,0)$1, whereas the symmetric channel loses most of its low-energy weight in the dimer phase (Lohöfer et al., 2015). The same study finds that the amplitude mode is not cleanly visible as a distinct contribution in the DSF itself, but appears as an only marginally damped mode in the dynamical singlet structure factor of interlayer bond correlations near criticality.

Quantum spin nematics furnish a different collective-mode structure. In the square-lattice spin-$\mathbf q=(0,0)$2 $\mathbf q=(0,0)$3-$\mathbf q=(0,0)$4 model with ferromagnetic $\mathbf q=(0,0)$5 and competing antiferromagnetic $\mathbf q=(0,0)$6, a large-$\mathbf q=(0,0)$7 treatment of the $\mathbf q=(0,0)$8 planar state yields a DSF composed of a high-energy Stoner continuum and low-energy coherent collective modes below that continuum (Shindou et al., 2011). At $\mathbf q=(0,0)$9, the low-energy modes are gapless $M$0-linear director-wave modes whose spectral weight vanishes linearly with momentum; at $M$1, there are gapped gauge-field-like collective modes whose spectral weight vanishes quadratically with momentum. The DSF therefore resolves the coupled director, spin, and gauge sectors of the nematic state rather than ordinary magnons.

6. Experimental scope, finite-temperature dynamics, and non-equilibrium generalizations

The experimental relevance of the DSF is pervasive. It is the quantity directly measured by inelastic neutron scattering in quantum magnets, and exact or quasi-exact formulas are especially useful where anisotropy selects specific channels. In the massive antiferromagnetic XXZ chain, the exact two-spinon longitudinal $M$2 was derived in compact form for $M$3, with explicit relevance to $M$4, $M$5, and $M$6; the formula satisfies the total-intensity and first-moment sum rules and reproduces the isotropic and Ising limits (Castillo, 2020). In two-dimensional noncollinear magnets, such as the triangular-lattice Heisenberg antiferromagnet, nonlinear spin-wave theory predicts broadened quasiparticle peaks, non-Lorentzian lineshapes, spontaneous magnon decays, and substantial spectral-weight transfer to the two-magnon continuum, all directly framed as neutron-scattering fingerprints (Mourigal et al., 2013).

Finite-temperature DSF calculations have become increasingly important for frustrated systems where low-temperature unbiased methods are difficult. Dyn-HTE treats the nearest-neighbor Heisenberg model for arbitrary lattices in the thermodynamic limit and reconstructs the DSF from high-temperature moments via continued fractions (Burkard et al., 20 May 2025). Applied to the triangular lattice, it finds that the dominant spectral peak at the roton-like $M$7 point stays around $M$8 through the anomalous intermediate-temperature regime, suggesting that this regime is not explained by a simple thermal softening of roton-like modes; applied to a pyrochlore $M$9 material, it reproduces the measured “V” shape and the high-frequency dome of spectral weight along $\mathbf Q=M$0.

The DSF has also been generalized beyond equilibrium ground states. In the transverse-field Ising chain, a non-equilibrium DSF defined from quenches out of direct product states reveals two characteristic continuum geometries: a bowtie-shaped continuum associated with two-spinon creation and a shell-like continuum associated with spinon-antispinon processes (Li et al., 2021). Under a finite longitudinal field, the broad continuum splits into sharp branches interpreted as spinon bound states, so the non-equilibrium DSF distinguishes deconfined from confined spinons without requiring preparation of the exact ground state. This suggests that the DSF has become a unifying observable across equilibrium condensed-matter spectroscopy and programmable quantum simulation, while preserving its original role as the most detailed momentum-frequency probe of spin dynamics.