Full Floquet-Space Formalism

• Full Floquet-space formalism is an extended Hilbert/Sambe-space approach that formulates periodic quantum problems as stationary eigenvalue problems with controlled truncation.
• It employs composite Fourier lattices and perturbative methods to analyze multimode drives, scattering, and Green’s-function formulations.
• The method improves predictions in quantum and solid-state systems by incorporating micromotion, gauge corrections, and mode hybridizations beyond low-order approximations.

Searching arXiv for the cited papers and closely related "Floquet-space" formulations to ground the article in current literature. In the literature considered here, “full Floquet-space formalism” designates an enlarged-space treatment in which the periodic problem is not immediately replaced by a low-order effective static Hamiltonian, but is formulated in an extended Hilbert/Floquet or Sambe space and then truncated in a controlled way when necessary. For a $T$-periodic Hamiltonian, the canonical eigenproblem is $$(H(t)-i\partial_t)|u_\alpha(t)\rangle=\varepsilon_\alpha|u_\alpha(t)\rangle$$, or equivalently a block matrix on $\mathcal H\otimes\mathcal T$ whose diagonal sectors are shifted by multiples of the drive frequency. Across later developments, the same enlarged-space logic is adapted to comb-like multimode drives, open-system Green’s functions and scattering, operator/Liouville superoperators, basis choices that count pump photons, and geometric or symmetry-refined reformulations of quasienergy structure (Simion et al., 25 Jun 2026, Tiwari et al., 26 Jan 2025, Seibold et al., 2024, Schindler et al., 2024).

1. Canonical extended-space construction

The canonical formulation starts from a periodic Hamiltonian $H(t+T)=H(t)$ and the Floquet ansatz

$$|\psi_\alpha(t)\rangle=e^{-i\varepsilon_\alpha t}|u_\alpha(t)\rangle, \qquad |u_\alpha(t+T)\rangle=|u_\alpha(t)\rangle .$$

The crucial conceptual step is that

$\hat H_F=\hat H(t)-i\frac{d}{dt}$

is treated as a time-independent operator in an enlarged space: the ordinary Hilbert space tensored with the space of $T$-periodic functions. This is exactly the Sambe-space construction. In generic notation, with $H(t)=\sum_l H_l e^{il\omega t}$, the Floquet Hamiltonian has matrix elements

$$\langle \alpha,l|\mathcal H_F|\beta,l'\rangle=(H_{l-l'})_{\alpha\beta}+l\hbar\omega\,\delta_{ll'}\delta_{\alpha\beta},$$

so the extended-space problem becomes a stationary eigenvalue problem with quasienergies defined modulo the drive frequency (Simion et al., 25 Jun 2026, Seibold et al., 2024).

This extended-space viewpoint is also the basis of a general perturbative formalism in which high- and low-frequency approximations are put on the same footing. In the extended Floquet Hilbert space $\mathscr F=\mathscr H\otimes\mathscr I$, the periodic problem is rewritten as

$$(H(t)-i\partial_t)|u_\alpha(t)\rangle=\varepsilon_\alpha|u_\alpha(t)\rangle$$0

with $$(H(t)-i\partial_t)|u_\alpha(t)\rangle=\varepsilon_\alpha|u_\alpha(t)\rangle$$1 the harmonic ladder operator and $$(H(t)-i\partial_t)|u_\alpha(t)\rangle=\varepsilon_\alpha|u_\alpha(t)\rangle$$2. Ordinary time-independent perturbation theory can then be applied in $$(H(t)-i\partial_t)|u_\alpha(t)\rangle=\varepsilon_\alpha|u_\alpha(t)\rangle$$3, so that high-frequency expansions, adiabatic perturbation theory, and diabatic corrections near Floquet resonances arise from different splittings of the same extended-space operator (Rodriguez-Vega et al., 2018).

2. Multimode and generalized harmonic lattices

A major extension replaces the single harmonic ladder by a multimode or composite Fourier lattice. For a comb spectrum

$$(H(t)-i\partial_t)|u_\alpha(t)\rangle=\varepsilon_\alpha|u_\alpha(t)\rangle$$4

the many-mode Floquet theory of polychromatically driven rf atoms does not introduce one Fourier index for every mode. Instead it uses the composite basis

$$(H(t)-i\partial_t)|u_\alpha(t)\rangle=\varepsilon_\alpha|u_\alpha(t)\rangle$$5

where $$(H(t)-i\partial_t)|u_\alpha(t)\rangle=\varepsilon_\alpha|u_\alpha(t)\rangle$$6 is associated with the base frequency $$(H(t)-i\partial_t)|u_\alpha(t)\rangle=\varepsilon_\alpha|u_\alpha(t)\rangle$$7 and $$(H(t)-i\partial_t)|u_\alpha(t)\rangle=\varepsilon_\alpha|u_\alpha(t)\rangle$$8 with the comb spacing $$(H(t)-i\partial_t)|u_\alpha(t)\rangle=\varepsilon_\alpha|u_\alpha(t)\rangle$$9. The Floquet Hamiltonian then has diagonal offsets

$\mathcal H\otimes\mathcal T$0

and off-diagonal couplings

$\mathcal H\otimes\mathcal T$1

so that the multimode problem becomes a sparse two-index Fourier lattice rather than a tensor product over all physical modes. In block form, the resulting operator is a block Toeplitz-like matrix in the spacing index $\mathcal H\otimes\mathcal T$2, with each diagonal block given by the single-frequency Floquet Hamiltonian in the index $\mathcal H\otimes\mathcal T$3 (Chakraborty et al., 2017).

A different generalization treats drives whose Fourier amplitudes vary slowly in time. There the Hamiltonian is written as $\mathcal H\otimes\mathcal T$4, periodic in the fast phase but slowly modulated in the second argument, and the extended space is $\mathcal H\otimes\mathcal T$5. After the unitary shift that removes the explicit fast phase, the quasienergy operator becomes

$\mathcal H\otimes\mathcal T$6

A time-dependent unitary $\mathcal H\otimes\mathcal T$7 then block-diagonalizes $\mathcal H\otimes\mathcal T$8 to

$\mathcal H\otimes\mathcal T$9

while the extra term $H(t+T)=H(t)$0 generates geometric corrections. The physical evolution factorizes into a slowly varying effective Floquet Hamiltonian and rapidly oscillating micromotion operators with slow temporal dependence (Novičenko et al., 2016).

3. Basis choice, truncation, and numerical realization

Within the enlarged-space formulation, the choice of basis in the physical Hilbert space can itself reorganize the Floquet sectors. In interacting bosonic driven systems, one proposal is to quantize not in the bare-oscillator basis $H(t+T)=H(t)$1, but in a basis $H(t+T)=H(t)$2 with $H(t+T)=H(t)$3, so that one “counts the photons of the drive.” The rotating-frame Hamiltonian is still decomposed into Fourier components $H(t+T)=H(t)$4, and a van Vleck expansion still produces

$H(t+T)=H(t)$5

but the sectors are organized around pump-photon counting rather than bare excitations. The transformation between the two descriptions is a Bogoliubov or squeezing transformation, and the paper’s central claim is that the $H(t+T)=H(t)$6-basis aligns the Floquet blocks with the actual response at the drive frequency and restores the classical Krylov–Bogolyubov correspondence order by order (Seibold et al., 2024).

For realistic solids in the velocity gauge, the formal difficulty is different: Floquet theory itself is exact in the enlarged time-periodic space, but a naively truncated electronic Hilbert space destroys gauge equivalence and convergence. The truncation-aware remedy is a nested-commutator Hamiltonian,

$H(t+T)=H(t)$7

which is then embedded in the standard Floquet-Bloch/Sambe construction. In the resulting matrix

$H(t+T)=H(t)$8

replicas separated by more than one photon can couple directly, so the Floquet matrix is no longer block tridiagonal. In a separate spintronics implementation, the periodic spin Hamiltonian is expanded only in harmonics $H(t+T)=H(t)$9, giving a block-tridiagonal Sambe operator that is truncated to $$|\psi_\alpha(t)\rangle=e^{-i\varepsilon_\alpha t}|u_\alpha(t)\rangle, \qquad |u_\alpha(t+T)\rangle=|u_\alpha(t)\rangle .$$0; convergence is then checked directly on quasienergies, spin expectation values, and Bloch-sphere trajectories. This is the sense in which “full Floquet-space” means: keep the Fourier structure explicitly in Sambe space, diagonalize the truncated Floquet Hamiltonian itself, and verify convergence of observables (Tiwari et al., 26 Jan 2025, Simion et al., 25 Jun 2026).

4. Open systems, scattering, and operator-space extensions

For open systems, full Floquet space appears either as an exact Sambe-space scattering problem or as a Floquet-resolved Green’s-function problem. In Floquet scattering theory, the exact flux-normalized scattering matrix is written as

$$|\psi_\alpha(t)\rangle=e^{-i\varepsilon_\alpha t}|u_\alpha(t)\rangle, \qquad |u_\alpha(t+T)\rangle=|u_\alpha(t)\rangle .$$1

with $$|\psi_\alpha(t)\rangle=e^{-i\varepsilon_\alpha t}|u_\alpha(t)\rangle, \qquad |u_\alpha(t+T)\rangle=|u_\alpha(t)\rangle .$$2 the quasi-energy operator in Floquet-Sambe space. High-frequency reduction proceeds by block-diagonalizing $$|\psi_\alpha(t)\rangle=e^{-i\varepsilon_\alpha t}|u_\alpha(t)\rangle, \qquad |u_\alpha(t+T)\rangle=|u_\alpha(t)\rangle .$$3 with the micromotion unitary $$|\psi_\alpha(t)\rangle=e^{-i\varepsilon_\alpha t}|u_\alpha(t)\rangle, \qquad |u_\alpha(t+T)\rangle=|u_\alpha(t)\rangle .$$4, but the resulting scattering matrix depends not only on the effective Hamiltonian $$|\psi_\alpha(t)\rangle=e^{-i\varepsilon_\alpha t}|u_\alpha(t)\rangle, \qquad |u_\alpha(t+T)\rangle=|u_\alpha(t)\rangle .$$5 but also on the micromotion block $$|\psi_\alpha(t)\rangle=e^{-i\varepsilon_\alpha t}|u_\alpha(t)\rangle, \qquad |u_\alpha(t+T)\rangle=|u_\alpha(t)\rangle .$$6; the paper’s explicit conclusion is that micromotion modifies the coupling of the driven system to the leads. In the Green’s-function formulation for periodically driven open fermionic systems, the two-time propagator is expanded as

$$|\psi_\alpha(t)\rangle=e^{-i\varepsilon_\alpha t}|u_\alpha(t)\rangle, \qquad |u_\alpha(t+T)\rangle=|u_\alpha(t)\rangle .$$7

and the local voltage-noise spectrum becomes a sum over all particle-hole fluctuation processes through Floquet sidebands (Li et al., 2018, Rodriguez-Vega et al., 2018).

A second line of development replaces Sambe space by operator/Liouville space. For statistically periodic noisy drives, the central object is the noise-averaged Floquet superoperator

$$|\psi_\alpha(t)\rangle=e^{-i\varepsilon_\alpha t}|u_\alpha(t)\rangle, \qquad |u_\alpha(t+T)\rangle=|u_\alpha(t)\rangle .$$8

whose eigenoperators and complex eigenphases generalize Floquet states and quasienergies. In a different operator-space construction, the stroboscopic adjoint action of a Floquet unitary is compressed into a Krylov/Arnoldi basis generated by a chosen Hermitian operator; the resulting Krylov-space matrix is exactly represented by the edge operator of a one-dimensional noninteracting Floquet transverse-field Ising model with inhomogeneous couplings. Both works are explicit that this is not a standard Sambe-space quasienergy formalism, but an enlarged operator-space treatment of Floquet dynamics (Sieberer et al., 2018, Yeh et al., 2023).

5. Geometric, spontaneous, and symmetry-adapted reformulations

Not all “full Floquet-space” work is written as an explicit block matrix on $$|\psi_\alpha(t)\rangle=e^{-i\varepsilon_\alpha t}|u_\alpha(t)\rangle, \qquad |u_\alpha(t+T)\rangle=|u_\alpha(t)\rangle .$$9. In geometric Floquet theory, the standard quasienergy redundancy is reinterpreted as a gauge phenomenon: quasienergy folding is derived from a broken gauge group $\hat H_F=\hat H(t)-i\frac{d}{dt}$0. The periodic Hamiltonian is decomposed as

$\hat H_F=\hat H(t)-i\frac{d}{dt}$1

and the Kato operator

$\hat H_F=\hat H(t)-i\frac{d}{dt}$2

provides a gauge-invariant unfolded average-energy spectrum that sorts Floquet states without quasienergy-zone ambiguity. For spontaneous Floquet states, an explicit $\hat H_F=\hat H(t)-i\frac{d}{dt}$3 extended-space formalism introduces a second variable $\hat H_F=\hat H(t)-i\frac{d}{dt}$4 and the full harmonic basis

$\hat H_F=\hat H(t)-i\frac{d}{dt}$5

so that spontaneous symmetry breaking yields Floquet-Nambu-Goldstone modes with zero quasi-energy, including a temporal mode associated with broken continuous time translation (Schindler et al., 2024, Nova et al., 2024).

A symmetry-adapted alternative appears in space-time crystals. There the physically relevant stroboscopic step is not the full period $\hat H_F=\hat H(t)-i\frac{d}{dt}$6 but the smaller interval $\hat H_F=\hat H(t)-i\frac{d}{dt}$7, and the central exact object is the space-time Floquet operator

$\hat H_F=\hat H(t)-i\frac{d}{dt}$8

Its spectrum unfolds conventional Floquet bands into an oblique reciprocal wavevector-frequency lattice. By contrast, a Lie-algebraic alternative exact Floquet construction bypasses explicit Sambe-space diagonalization altogether: the micromotion is parameterized by a Wei–Norman product, and the drive is inverse-engineered by fixing the gauge of the micromotion. Taken together, these works suggest that “full Floquet-space” can also mean a symmetry-refined or gauge-refined realization of the same underlying periodic structure, rather than only an explicit harmonic block matrix (Melkani et al., 18 Oct 2025, Bandyopadhyay et al., 2021).

6. Physical consequences, applications, and recurring limitations

The practical output of full Floquet-space methods is the ability to capture structures that low-order effective Hamiltonians miss. In rf-dressed cold atoms, the multimode construction predicts lattice-like periodic variation in the eigen-energies, a lattice type atom trapping potential, and large two-photon transition probabilities generated by secondary avoided crossings in the two-index Fourier lattice. In driven coupled-spin systems, the full Hilbert–Floquet formalism shows that exchange alone does not modify the collective spin expectation values under the chosen initial condition, whereas increasing DMI generates a finite expectation value of $\hat H_F=\hat H(t)-i\frac{d}{dt}$9, suppresses $T$0, and produces tilted, elliptical Bloch-sphere trajectories. In realistic solids, the truncation-aware velocity-gauge Floquet-Bloch formalism reproduces the converged 140-band velocity-gauge spectrum of laser-dressed trans-polyacetylene with only a few Wannier bands and substantially reduced computation time (Chakraborty et al., 2017, Simion et al., 25 Jun 2026, Tiwari et al., 26 Jan 2025).

The same enlarged-space logic also supports topological and spectroscopic applications. The magnonic Floquet Hofstadter problem is formulated by combining a static Hofstadter-like Aharonov–Casher phase with a time-periodic electric field and then applying the Floquet formalism in the magnetic Bloch basis, yielding a Floquet Hofstadter butterfly and frequency-dependent deformation of the fractal structure. In periodically driven open fermionic systems, local voltage-noise spectra detect regular and anomalous Floquet topological bound states through peaks around zero-, half-, and full-drive frequency. In measurement-only Floquet codes, a full-Floquet-space reading reconstructs the relevant structure from a periodic sequence of noncommuting two-body Pauli measurements, a round-dependent instantaneous stabilizer group, and syndrome information that becomes complete only over the full period, although that work is explicit that it does not yet provide a fully abstract Floquet-space formalism (Owerre, 2018, Rodriguez-Vega et al., 2018, Watanabe et al., 13 Feb 2026).

Taken together, these works suggest that “full Floquet-space formalism” is not a single universally fixed construction. In some papers it means the standard Hilbert–Floquet or Sambe-space eigenproblem and its controlled truncation; in others it means a two-index or symmetry-adapted harmonic lattice; in still others it means operator/Liouville space, a noise-averaged Floquet superoperator, or a geometric unfolding of quasienergy redundancy. The common thread is more stable than the terminology: periodic dynamics is embedded into an enlarged space in which the problem becomes stationary, exact before truncation, and capable of retaining mode hybridization, counter-rotating terms, multiphoton processes, micromotion, and gauge structure that simpler rotating-wave, Magnus, or low-order effective-static descriptions neglect (Rodriguez-Vega et al., 2018, Schindler et al., 2024, Simion et al., 25 Jun 2026).