---
title: Rotor-Spin-Wave Theory in Quantum Magnets
url: https://www.emergentmind.com/topics/rotor-spin-wave-theory
type: topic
---

# Rotor-Spin-Wave Theory in Quantum Magnets

Searching arXiv for the cited rotor/spin-wave and time-dependent spin-wave papers to ground the article in the relevant literature.
arXiv search query: "Rotor spin-wave theory U(1) symmetry quantum spin models"
Rotor-spin-wave theory denotes a family of analytical constructions that separate collective rotational degrees of freedom from finite-momentum spin-wave fluctuations in quantum spin systems, or, in explicitly driven magnets, formulate spin-wave expansions in a co-rotating reference frame aligned with the instantaneous magnetization. In both usages, the central objective is to regularize pathologies of naive linear spin-wave theory associated with collective modes: in finite-size $U(1)$-symmetric systems the problematic mode is the Goldstone zero mode, while in time-dependent magnets it is the failure of laboratory-frame bosonization to track the true nonequilibrium order parameter. The theory therefore combines an exact or nonperturbative treatment of a rotor sector with a quadratic treatment of dilute spin waves, yielding controlled descriptions of low-energy spectra, finite-size dynamics, driven ferromagnets, and magnon condensation phenomena [2303.00380], [1111.2052].

## 1. Conceptual scope and historical setting

Rotor-spin-wave theory appears in two closely related but technically distinct settings. In the first, developed for finite-size lattice spin models with spontaneous $U(1)$ symmetry breaking only in the thermodynamic limit, the zero-momentum sector is isolated and resummed into a quantum rotor associated with the Anderson tower of states, while the $k\neq 0$ sector is treated by spin-wave theory [2303.00380], [2302.09271]. In the second, formulated for explicitly time-dependent quantum spin Hamiltonians, spin operators are first transformed into properly chosen rotating reference frames and only then bosonized, so that the spin-wave expansion follows the instantaneous direction of the nonequilibrium magnetization [1111.2052].

These two strands share a common structural principle: the collective coordinate cannot be handled as an ordinary linear bosonic fluctuation without generating unphysical divergences or singularities. In finite-size $U(1)$ systems, linear-spin-wave theory produces a spurious zero-energy oscillator at $\mathbf q=0$, whereas the exact collective sector is a nonlinear rotor [2303.00380]. In rotating-field problems, naive laboratory-frame perturbation theory develops artificial singularities near resonance, whereas rotating-frame bosonization yields a regular spectrum and a transparent interpretation of the drive as a Coriolis-like shift [1111.2052].

A plausible unifying interpretation is that rotor-spin-wave theory is best understood as a symmetry-adapted or frame-adapted reorganization of the semiclassical $1/S$ expansion. The collective coordinate is promoted from a linearized fluctuation to a genuine dynamical variable, and only the remaining weakly occupied modes are linearized.

## 2. Zero mode, quantum rotor, and Anderson tower of states

For $U(1)$-symmetric quantum spin models, the canonical example is the spin-$1/2$ XX model on a Bravais lattice with power-law couplings
\[
\mathcal H_{\alpha\text{-XX}}=-\sum_{i<j}J_{ij}\bigl(S_i^xS_j^x+S_i^yS_j^y\bigr),
\]
with $J_{ij}=J/|\mathbf r_i-\mathbf r_j|^\alpha$ [2303.00380]. Because this Hamiltonian commutes with $J^z=\sum_i S_i^z$, it has continuous $U(1)$ symmetry. In the thermodynamic limit it develops ferromagnetic order in the $xy$ plane, but on any finite lattice the symmetry remains exact [2303.00380].

If one performs a conventional Holstein-Primakoff expansion around the $x$-polarized coherent state, the quadratic Hamiltonian yields a spin-wave dispersion
\[
\varepsilon_{\mathbf q}=S\sqrt{(J_0-J_{\mathbf q})(J_0+J_{\mathbf q})},
\]
which vanishes at $\mathbf q=0$ [2303.00380]. The resulting Bogolyubov coefficient $v_{\mathbf q=0}$ diverges, giving a divergent boson number $\langle b_0^\dagger b_0\rangle$; this is the zero-mode pathology of naive linearization [2303.00380].

Rotor-spin-wave theory resolves this by collecting all Hamiltonian terms involving only $b_0,b_0^\dagger$. Under the Holstein-Primakoff map, the uniform components become those of a “giant spin” $\mathbf K$ of length $K=NS$, obeying an SU(2) algebra [2303.00380]. Projecting onto the maximal-spin Dicke sector yields exactly
\[
\mathcal H_{\rm R}=E_{0,\rm R}+\frac{(K^z)^2}{2I},
\qquad
E_{0,\rm R}=-\frac{NS^2J_0}{2},
\qquad
I=\frac{N}{J_0},
\]
with $[\phi,K^z]=i$ for the conjugate phase $\phi$ [2303.00380]. The spectrum
\[
\mathcal H_{\rm R}|K,M\rangle
=
\left(E_{0,\rm R}+\frac{M^2}{2I}\right)|K,M\rangle
\]
reproduces the Anderson tower of states, with each $M$ labeling a different $J^z=M$ sector [2303.00380].

In the more general XXZ formulation,
\[
H_{\rm XXZ}
=
-\sum_{i<j}J_{ij}\bigl(S_i^xS_j^x+S_i^yS_j^y+\Delta\,S_i^zS_j^z\bigr),
\]
the extracted rotor Hamiltonian reads
\[
H_{\rm R}
=
E_{0,R}+\frac{J_0(1-\Delta)}{2N}(K^z)^2
=
\frac{\chi}{2N}(K^z)^2
\]
after dropping the constant, with $\chi\equiv J_0(1-\Delta)$ and moment of inertia $I=N/\chi$ [2302.09271]. In that form the rotor sector is exactly the one-axis-twisting Hamiltonian.

The significance of this construction is not merely regularization. It identifies the finite-size precursor of spontaneous symmetry breaking as a rotor with nonlinear spectrum rather than as a harmonic mode. This explains why the collective sector produces the Anderson tower rather than a single Goldstone oscillator [2303.00380].

## 3. Separation from finite-momentum spin waves

Once the zero mode is extracted, the remaining finite-momentum modes can be treated within a dilute-gas approximation. In the XX model, the Hamiltonian for $\mathbf q\neq 0$ is truncated to the quadratic piece
\[
\mathcal H_{\rm SW}
=
\tfrac12\sum_{\mathbf q\neq0}
(b_{\mathbf q}^\dagger,b_{-\mathbf q})
\begin{pmatrix}
A_{\mathbf q}&B_{\mathbf q}\\
B_{\mathbf q}&A_{\mathbf q}
\end{pmatrix}
\begin{pmatrix}
b_{\mathbf q}\\
b^\dagger_{-\mathbf q}
\end{pmatrix}
-\tfrac12\sum_{\mathbf q\neq0}A_{\mathbf q},
\]
with $A_{\mathbf q}=S(J_0-J_{\mathbf q})$ and $B_{\mathbf q}=-SJ_{\mathbf q}$, and is diagonalized by a Bogolyubov transform [2303.00380]. In the XXZ formulation one obtains
\[
H_{\rm SW}
=
\sum_{q\neq0}\Bigl[
A_q\,b_q^\dagger b_q+\tfrac12 B_q(b_qb_{-q}+b_q^\dagger b_{-q}^\dagger)
\Bigr],
\]
with
\[
A_q=S\bigl[J_0-\tfrac{1+\Delta}{2}J_q\bigr],
\qquad
B_q=-S\tfrac{1-\Delta}{2}J_q,
\qquad
\omega_q=\sqrt{A_q^2-B_q^2}
\]
after Bogoliubov diagonalization [2302.09271].

The central approximation is the neglect of residual couplings between rotor and spin-wave sectors, described as terms of order $\mathcal O(n_0 n_q)$ or $\mathcal O(b_0^2 b_{\mathbf q\neq0}^2)$ [2302.09271], [2303.00380]. Under this approximation,
\[
\mathcal H\approx \mathcal H_{\rm R}+\mathcal H_{\rm SW},
\]
and the eigenstates factorize as rotor states times spin-wave occupation states [2303.00380].

The validity regime is stated explicitly. The finite-momentum spin waves must remain weakly populated, $\langle b_{\mathbf q\neq0}^\dagger b_{\mathbf q\neq0}\rangle\ll 2S$ [2303.00380]. The $k=0$ mode should dominate the total boson number, $\langle n_0\rangle\gg \sum_{q\neq 0}\langle n_q\rangle$, and the neglected rotor-spin-wave coupling then only renormalizes the moment of inertia by $\mathcal O(\langle n_q\rangle/\langle n_0\rangle)$ [2302.09271]. For power-law interactions in $d$ dimensions, spin-wave excitations remain dilute for $\alpha<\alpha_c\approx 3d/2$, so that OAT-like dynamics persists; above $\alpha_c$ the spin-wave sector depolarizes the spin too quickly, spoiling scalable squeezing [2302.09271].

This separation yields several immediate consequences. The ground-state energy and long-distance correlators split into rotor and spin-wave contributions [2303.00380]. The finite-size order parameter vanishes,
\[
\langle J^x\rangle =0,
\]
because the rotor sector respects the exact $U(1)$ symmetry and the spin-wave sector supplies the compensating population [2303.00380]. This clarifies a common misconception: rotor-spin-wave theory does not restore spontaneous symmetry breaking on a finite system; rather, it encodes how symmetry restoration coexists with ordered correlations through the Anderson tower plus finite-momentum fluctuations.

## 4. Nonequilibrium dynamics and entangling evolution in $U(1)$ systems

A major application is the nonequilibrium evolution from a coherent spin state. For the quench from
\[
\ket{\!\rightarrow_x}^{\otimes N},
\]
the rotor sector evolves under the exactly solvable one-axis-twisting Hamiltonian
\[
i\frac{d}{dt}\ket{\psi_{\rm R}(t)}
=
\Bigl(\frac{(K^z)^2}{2I}\Bigr)\ket{\psi_{\rm R}(t)},
\]
while the finite-momentum spin waves evolve independently as Gaussian modes with covariance matrices
\[
G_{\mathbf q}(t)=\langle b_{\mathbf q}^\dagger b_{\mathbf q}\rangle,
\qquad
F_{\mathbf q}(t)=\langle b_{\mathbf q}b_{-\mathbf q}\rangle,
\]
obeying
\[
\dot G_{\mathbf q}=-2B_{\mathbf q}\,\mathrm{Im}\,F_{\mathbf q},
\qquad
\dot F_{\mathbf q}=-i[2A_{\mathbf q}F_{\mathbf q}+B_{\mathbf q}(1+2G_{\mathbf q})]
\]
with analytic Bogolyubov solutions [2303.00380].

The essential improvement over naive linear-spin-wave theory lies in the zero mode. In linearized treatment, the $\mathbf q=0$ mode is handled as a quadratically squeezed oscillator and one finds $\langle b_0^\dagger b_0\rangle\propto t^2\to\infty$, leading to an unphysically fast collapse of $\langle J^x\rangle$ [2303.00380]. Retaining the full $(K^z)^2$ nonlinearity keeps the zero mode finite and reproduces OAT-like dynamics for times up to $\mathcal O(N)$ [2303.00380].

In the XXZ treatment, this factorized dynamics explains entanglement generation. Since $H\approx H_R+H_{SW}$, one has to leading order
\[
\langle J^x\rangle=\langle K^x\rangle-N_{FM},
\qquad
\mathrm{Var}(J^\perp)\simeq \mathrm{Var}(K^\perp),
\]
where $N_{FM}=\sum_{q\neq0}\langle b_q^\dagger b_q\rangle$ [2302.09271]. The squeezing parameter satisfies
\[
\xi_R^2(t)\simeq \xi_{\rm OAT}^2(t)+O(N_{FM}/N),
\]
and the minimal squeezing scales as $\xi_{\min}^2\sim N^{-2/3}$ at $t_{\rm sq}\sim N^{1/3}$, up to small renormalizations from finite spin-wave population [2302.09271]. At the “cat-times”
\[
t_q=2\pi I/q,
\]
the rotor wave packet splits into $q$ coherent lobes on the Bloch sphere, producing ideal $q$-headed cats, while the spin-wave sector adds small fluctuations and reduces the purity by $O(N_{FM})$ [2302.09271].

The same framework was used to explain the persistence of OAT-like entangling dynamics in power-law models, including dipolar XX systems in $d=2$, through an effective separation between zero-momentum degrees of freedom associated with the Anderson tower and finite-momentum spin waves [2302.09271].

## 5. Time-dependent rotating-frame formulation

In explicitly time-dependent magnets, rotor-spin-wave theory takes the form of a rotating-frame spin-wave expansion. The starting point is the time-dependent Heisenberg Hamiltonian
\[
H(t)=-\frac12\sum_{ij}J_{ij}\,S_i\!\cdot\!S_j-\sum_i h_i(t)\!\cdot\!S_i,
\]
with, in the uniform rotating-field case,
\[
h(t)=h_\perp[\cos(\omega t)\,\hat e_x-\sin(\omega t)\,\hat e_y]+h_z\,\hat e_z
\]
[1111.2052]. An equivalent interpretation is a ferromagnet on a rotating cylinder in a static field [1111.2052].

The defining step is to construct a time-dependent unitary rotation
\[
U_0(t)=\exp\!\left[-i\sum_i \alpha_i(t)\!\cdot\!S_i\right]
\]
that aligns the local $z'$ axis with the actual nonequilibrium magnetization direction
\[
\hat m_i(t)=\frac{\langle S_i(t)\rangle}{|\langle S_i(t)\rangle|}.
\]
For uniform rotation, one may choose Euler angles with $\psi_i(t)=0$, $\phi_i(t)=-\omega t$, and constant nutation angle $\theta_i\equiv\theta$ [1111.2052]. The rotated spins are
\[
\tilde S_i(t)=U_0^\dagger(t)S_iU_0(t)=e^{\alpha_i(t)\times}S_i,
\]
and the explicit time dependence of $U_0$ generates a Berry-phase or Coriolis Hamiltonian
\[
H_B(t)=-iU_0^\dagger(t)\partial_tU_0(t).
\]
In the Euler-angle parametrization with $\psi_i=0$,
\[
H_B(t)=
-\sum_i\Bigl[
\dot\theta_i\,\tilde S_i^{(1)}
+\dot\phi_i\sin\theta_i\,\tilde S_i^{(2)}
+\dot\phi_i\cos\theta_i\,\tilde S_i^{(\parallel)}
\Bigr]
\]
[1111.2052].

Holstein-Primakoff bosonization is then performed in the rotated local basis $\{\hat e_i^{(1)},\hat e_i^{(2)},\hat m_i\}$, with
\[
\tilde S_i^{(\parallel)}=S-c_i^\dagger c_i,
\qquad
\tilde S_i^+\simeq \sqrt{2S}\,c_i,
\qquad
\tilde S_i^-\simeq \sqrt{2S}\,c_i^\dagger
\]
to leading order in $1/S$ [1111.2052]. The Hamiltonian
\[
H_A(t)=U_0^\dagger(t)H(t)U_0(t)+H_B(t)
\]
is then expanded to quadratic order in $c_i,c_i^\dagger$ [1111.2052].

For the uniformly rotating field, linear terms vanish when the tilt angle satisfies
\[
\cos\theta=\frac{h_z-\omega}{\tilde h_\omega},
\qquad
\tilde h_\omega=\sqrt{h_\perp^2+(h_z-\omega)^2}.
\]
The quadratic spin-wave Hamiltonian becomes
\[
H_{\rm sw}=\sum_k E_k(\omega)c_k^\dagger c_k,
\qquad
E_k(\omega)=\varepsilon_k+\tilde h_\omega,
\]
with $\varepsilon_k=S(J_0-J_k)$ [1111.2052].

The interpretation is explicit: the rotation enters through $H_B$, effectively shifting the longitudinal field by $-\omega$ and mixing transverse components. In the spectrum, the combination $(h_z-\omega)$ replaces $h_z$, so resonance at $k=0$ would occur as $\omega$ approaches $h_z$ [1111.2052]. Proper rotating-frame bosonization thereby removes the artificial singularities of naive laboratory-frame perturbation theory and captures the dynamic reorganization of the magnetic state [1111.2052].

## 6. Magnon Bose-Einstein condensation in YIG and broader significance

An important application of the time-dependent formulation is microwave-pumped yttrium-iron garnet. There one obtains an effective spin Hamiltonian with a time-dependent pair-creation term
\[
H(t)\simeq
\sum_k\Bigl[
(\varepsilon_k+h)b_k^\dagger b_k
+\frac{h_c}{2}\bigl(e^{2i\omega t}b_{-k}b_k+\mathrm{h.c.}\bigr)
\Bigr],
\]
where $h_c\sim$ pump amplitude [1111.2052]. In the rotating frame this maps onto a tilted ferromagnet with magnon gap
\[
E_k=\sqrt{[\varepsilon_k+(h-\omega)]^2-h_c^2}.
\]
At $|h-\omega|=h_c$, the $k=0$ gap closes, signaling an Ising-type quantum phase transition in which the static magnetization tilts away from the field axis [1111.2052].

Within linear spin-wave theory the magnetization vector takes the form
\[
M(\omega,T)=\hat m(\omega)\Bigl[S+\tfrac12-\frac1N\sum_k\{\ldots\}\Bigr],
\]
and exhibits a characteristic dip of order one percent near the condensation threshold, in agreement with observed BEC of magnons in YIG [1111.2052]. The details state that this dip should be measurable in experiments and interpret the onset of magnon BEC as a genuine quantum phase transition in the Ising class [1111.2052].

The broader significance follows directly from the formalism. Rotor-spin-wave theory provides a systematic $1/S$ expansion for explicitly time-dependent spin models, goes beyond the adiabatic approximation, and is applicable to nonequilibrium dynamics in ordered magnets including antiferromagnets and frustrated systems [1111.2052]. In the finite-size $U(1)$ context, it supplies a nearly parameter-free description of equilibrium and quench physics across interaction ranges, benchmarked against quantum Monte Carlo, exact diagonalization, and time-dependent variational Monte Carlo [2303.00380].

## 7. Benchmarks, misconceptions, and limits of applicability

The available benchmarks are unusually explicit. For equilibrium properties of the power-law XX model, rotor-spin-wave ground-state energies and long-distance correlators agree with quantum Monte Carlo to better than a few percent already at $L\sim 20$ for all $\alpha$ [2303.00380]. Exact diagonalization on $4\times 4$ dipolar XX exhibits Anderson towers linear in $(J^z)^2$ plus spin-wave bands; rotor-spin-wave theory reproduces both the rotor part with moment of inertia $I$ and the finite-$\mathbf q$ excitations, whereas linear spin-wave theory with a small symmetry-breaking field treats $\mathbf q=0$ as a harmonic oscillator and fails to capture the nonlinear tower beyond the first level [2303.00380]. For quenches, time-dependent variational Monte Carlo shows that $\langle J^x(t)\rangle$ depolarizes on an $\mathcal O(1)$ timescale and that the minimum spin-squeezing parameter attains the OAT scaling $N^{-2/3}$ at $t\sim N^{1/3}$; rotor-spin-wave theory matches these trends quantitatively up to long times, while naive linear-spin-wave theory breaks down once $\langle b_0^\dagger b_0\rangle$ grows without control [2303.00380].

Several misconceptions can be addressed precisely. First, the rotor is not an ad hoc infrared regulator. In the finite-size $U(1)$ problem it is the exact nonlinear form of the zero mode after resumming all terms involving $b_0,b_0^\dagger$ [2303.00380]. Second, the theory does not assert exact decoupling in the full Hilbert space. It neglects quartic and higher couplings between collective and finite-momentum sectors, justified only when the spin-wave gas is dilute and the collective mode dominates [2302.09271], [2303.00380]. Third, the time-dependent formulation is not merely a change of coordinates: the Berry-phase term
\[
H_B=-iU_0^\dagger\partial_tU_0
\]
is dynamically essential and encodes the Coriolis-like effects of the rotating frame [1111.2052].

The limitations are also explicit. For large $\alpha$, finite-momentum fluctuations grow, and although rotor-spin-wave theory still outperforms linear-spin-wave theory with a symmetry-breaking field by capturing the tower-of-states nonlinearity, the separation becomes less accurate [2303.00380]. In entangling dynamics, scalable OAT-like behavior persists only when the spin-wave sector remains sufficiently dilute, quantified by $\alpha<\alpha_c\approx 3d/2$ in the power-law setting [2302.09271]. In the time-dependent case, the formalism remains a $1/S$ expansion and therefore presupposes an ordered reference state with controllable spin-wave density [1111.2052].

Taken together, these results establish rotor-spin-wave theory as a unified strategy for treating collective modes that are intrinsically nonlinear or frame dependent. In finite-size symmetry-breaking problems it resolves the zero-mode divergence by identifying a quantum rotor behind the Anderson tower of states; in driven magnets it resolves laboratory-frame singularities by aligning the bosonization basis with the instantaneous magnetization. The resulting framework connects finite-size symmetry restoration, spin squeezing, Schrödinger-cat formation, rotating-frame magnons, and magnon Bose-Einstein condensation within a single methodological lineage [2302.09271], [2303.00380], [1111.2052].

Source: https://www.emergentmind.com/topics/rotor-spin-wave-theory