Driven Jaynes-Cummings System Dynamics

• Driven Jaynes–Cummings system is a quantum light–matter model enhanced by classical drives that precisely control bosonic fields and two-level excitations.
• It exhibits phenomena such as dressed energy ladders, Autler–Townes splitting, bistability, and multiphoton tunneling, highlighting its complex nonlinear dynamics.
• Its versatile formulation across platforms enables advanced state preparation, spectral control, and tailored quantum protocols for experimental and theoretical applications.

A driven Jaynes–Cummings system is a Jaynes–Cummings (JC) light–matter model augmented by a classical drive acting on either the bosonic mode or the two-level subsystem. In its canonical rotating-wave form, it describes a qubit of transition frequency $\omega_q$ coherently exchanging excitations with a single bosonic mode of frequency $\omega_c$, while an external field injects amplitude, phase, and detuning control. Across cavity QED, circuit QED, neutral-atom, trapped-ion, and semiconductor platforms, this driven setting supports laser-dressed JC ladders, Autler–Townes structure, bistability, multiphoton tunneling, spectral collapse at critical drive, and control protocols ranging from counterdiabatic state preparation to universal oscillator logic (Lee et al., 2016, Gutiérrez-Jáuregui et al., 2018, Hopfmann et al., 2016).

1. Canonical formulation and drive representations

In the rotating-wave approximation, the canonical JC Hamiltonian is

$$H_{JC}=\hbar \omega_c\,a^\dagger a+\frac{\hbar \omega_q}{2}\,\sigma_z+\hbar g\left(a\,\sigma_+ + a^\dagger\,\sigma_-\right),$$

with detuning $\Delta=\omega_q-\omega_c$. The two standard classical drives are a bosonic-mode drive,

$$H_{\text{drive}}=\hbar\big(\varepsilon\,a\,e^{-i\omega_d t}+\varepsilon^* a^\dagger e^{i\omega_d t}\big),$$

and a qubit drive,

$$H_{\text{drive}}=\hbar \Omega \cos(\omega_d t)\,\sigma_x.$$

On exact resonance, the driven JC model can be written in the resonant interaction picture as

$$\mathcal{H}_0=i\hbar \lambda\left(\hat a\,\hat \sigma_+ - \hat a^\dagger \hat \sigma_-\right)+\hbar \epsilon(\hat a+\hat a^\dagger),$$

while an extension with counter-rotating interactions is

$$\mathcal{H}_\eta=i\hbar\lambda'\left[(\hat a+\eta \hat a^\dagger)\hat \sigma_+ - (\hat a^\dagger+\eta \hat a)\hat \sigma_-\right]+\hbar \epsilon'(\hat a+\hat a^\dagger),$$

with $\eta\in[0,1]$ tuning the counter-rotating sector (Gutiérrez-Jáuregui et al., 2018).

For coherent-state driving, cavity drive and atom drive are dynamically equivalent up to a trivial coherent offset in the emitted field. The equivalence becomes explicit after a displacement transformation $a(t)\to b(t)+\alpha(t)$, where $\omega_c$0 solves the classical empty-cavity response,

$\omega_c$1

so that the internal JC dynamics under cavity driving matches atom driving with coherent envelope

$\omega_c$2

The only difference is the coherent background

$\omega_c$3

which can be removed interferometrically by homodyne cancellation (Fischer et al., 2018).

A distinct exact mapping arises when both atom and field are driven. With

$\omega_c$4

an invariant construction yields an exact reduction to the standard JC Hamiltonian provided

$\omega_c$5

After the corresponding displacement and rotating-frame transformations, the effective interaction becomes

$\omega_c$6

so the driven problem inherits the exact JC solution structure (Bocanegra et al., 2023).

2. Dressed ladders, Autler–Townes structure, and phase-space organization

The undriven JC ladder is anharmonic, with doublets

$\omega_c$7

and the hallmark $\omega_c$8 scaling at resonance. Under coherent driving, these doublets are themselves dressed, producing regimes that interpolate between a quantum JC ladder and a semiclassical Autler–Townes ladder. In a strongly coupled quantum dot–microcavity, matter-driven excitation yields a rotating-frame Hamiltonian

$\omega_c$9

and, at exact resonance in the single-excitation subspace,

$$H_{JC}=\hbar \omega_c\,a^\dagger a+\frac{\hbar \omega_q}{2}\,\sigma_z+\hbar g\left(a\,\sigma_+ + a^\dagger\,\sigma_-\right),$$0

The weak-drive JC regime $$H_{JC}=\hbar \omega_c\,a^\dagger a+\frac{\hbar \omega_q}{2}\,\sigma_z+\hbar g\left(a\,\sigma_+ + a^\dagger\,\sigma_-\right),$$1, intermediate regime $$H_{JC}=\hbar \omega_c\,a^\dagger a+\frac{\hbar \omega_q}{2}\,\sigma_z+\hbar g\left(a\,\sigma_+ + a^\dagger\,\sigma_-\right),$$2, and Autler–Townes regime $$H_{JC}=\hbar \omega_c\,a^\dagger a+\frac{\hbar \omega_q}{2}\,\sigma_z+\hbar g\left(a\,\sigma_+ + a^\dagger\,\sigma_-\right),$$3 are experimentally distinguishable. In that platform, a pronounced resonance-fluorescence maximum at $$H_{JC}=\hbar \omega_c\,a^\dagger a+\frac{\hbar \omega_q}{2}\,\sigma_z+\hbar g\left(a\,\sigma_+ + a^\dagger\,\sigma_-\right),$$4 serves as a robust fingerprint of the intermediate regime, and injection pulling of the polariton branches is observed with slopes $$H_{JC}=\hbar \omega_c\,a^\dagger a+\frac{\hbar \omega_q}{2}\,\sigma_z+\hbar g\left(a\,\sigma_+ + a^\dagger\,\sigma_-\right),$$5 for the upper polariton and $$H_{JC}=\hbar \omega_c\,a^\dagger a+\frac{\hbar \omega_q}{2}\,\sigma_z+\hbar g\left(a\,\sigma_+ + a^\dagger\,\sigma_-\right),$$6 for the lower polariton at $$H_{JC}=\hbar \omega_c\,a^\dagger a+\frac{\hbar \omega_q}{2}\,\sigma_z+\hbar g\left(a\,\sigma_+ + a^\dagger\,\sigma_-\right),$$7 (Hopfmann et al., 2016).

Near cavity resonance, monochromatic bosonic driving also produces two branches of multiphoton excitation that can be described semiclassically in the rotating frame. After displacing the driven cavity mode, one obtains

$$H_{JC}=\hbar \omega_c\,a^\dagger a+\frac{\hbar \omega_q}{2}\,\sigma_z+\hbar g\left(a\,\sigma_+ + a^\dagger\,\sigma_-\right),$$8

and the branch energies obey

$$H_{JC}=\hbar \omega_c\,a^\dagger a+\frac{\hbar \omega_q}{2}\,\sigma_z+\hbar g\left(a\,\sigma_+ + a^\dagger\,\sigma_-\right),$$9

This yields upper and lower driven-polariton ladders and two corresponding families of multiphoton resonances (Ermann et al., 2020).

A complementary quasienergy construction describes the driven resonant JC system as motion on a phase-space quasienergy surface $\Delta=\omega_q-\omega_c$0 with effective Planck constant

$\Delta=\omega_q-\omega_c$1

For one dressed-spin branch, the surface develops a nonperturbative well with minimum

$\Delta=\omega_q-\omega_c$2

saddle

$\Delta=\omega_q-\omega_c$3

and barrier height $\Delta=\omega_q-\omega_c$4. The state localized at the well bottom is a displaced squeezed state with squeezing parameter

$\Delta=\omega_q-\omega_c$5

sub-Poissonian statistics, and a local harmonic level spacing governed by the small-oscillation frequency

$\Delta=\omega_q-\omega_c$6

Within this picture, multiphoton transitions are reinterpreted as resonant tunneling transitions from the local maximum of the quasienergy surface (Peano et al., 2010).

3. Critical drive, spectral collapse, and geometric criticality

At exact resonance, the driven JC model exhibits a critical drive amplitude

$\Delta=\omega_q-\omega_c$7

below which the quasienergy spectrum is discrete and above which it becomes continuous. Approaching $\Delta=\omega_q-\omega_c$8 from below, the level spacings collapse according to

$\Delta=\omega_q-\omega_c$9

The same transition persists in the driven Jaynes–Cummings–Rabi model, where counter-rotating terms shift the critical point to

$$H_{\text{drive}}=\hbar\big(\varepsilon\,a\,e^{-i\omega_d t}+\varepsilon^* a^\dagger e^{i\omega_d t}\big),$$0

A Dirac-particle correspondence identifies the critical condition with the Lorentz invariant $$H_{\text{drive}}=\hbar\big(\varepsilon\,a\,e^{-i\omega_d t}+\varepsilon^* a^\dagger e^{i\omega_d t}\big),$$1, so that spectral collapse occurs precisely at $$H_{\text{drive}}=\hbar\big(\varepsilon\,a\,e^{-i\omega_d t}+\varepsilon^* a^\dagger e^{i\omega_d t}\big),$$2, or equivalently $$H_{\text{drive}}=\hbar\big(\varepsilon\,a\,e^{-i\omega_d t}+\varepsilon^* a^\dagger e^{i\omega_d t}\big),$$3 (Gutiérrez-Jáuregui et al., 2018).

An independent resonant cavity-driven analysis yields the same threshold structure in the steady-state spectrum. For the interaction-picture Hamiltonian

$$H_{\text{drive}}=\hbar\big(\varepsilon\,a\,e^{-i\omega_d t}+\varepsilon^* a^\dagger e^{i\omega_d t}\big),$$4

the exact quasienergies are

$$H_{\text{drive}}=\hbar\big(\varepsilon\,a\,e^{-i\omega_d t}+\varepsilon^* a^\dagger e^{i\omega_d t}\big),$$5

Normalizable steady states exist only for $$H_{\text{drive}}=\hbar\big(\varepsilon\,a\,e^{-i\omega_d t}+\varepsilon^* a^\dagger e^{i\omega_d t}\big),$$6; for $$H_{\text{drive}}=\hbar\big(\varepsilon\,a\,e^{-i\omega_d t}+\varepsilon^* a^\dagger e^{i\omega_d t}\big),$$7, no normalizable steady states exist. In the weak-drive regime, the field splits into two counterrotating coherent components whose Husimi peaks move on circles centered at $$H_{\text{drive}}=\hbar\big(\varepsilon\,a\,e^{-i\omega_d t}+\varepsilon^* a^\dagger e^{i\omega_d t}\big),$$8, and revival structure is governed by the overlap of these peaks in phase space (Tuguldur et al., 2012).

A more recent eigenstate-level analysis formulates the critical point in terms of the dimensionless drive $$H_{\text{drive}}=\hbar\big(\varepsilon\,a\,e^{-i\omega_d t}+\varepsilon^* a^\dagger e^{i\omega_d t}\big),$$9, with $$H_{\text{drive}}=\hbar \Omega \cos(\omega_d t)\,\sigma_x.$$0 corresponding to $$H_{\text{drive}}=\hbar \Omega \cos(\omega_d t)\,\sigma_x.$$1. The driven resonant JC eigenenergies take the exact form

$$H_{\text{drive}}=\hbar \Omega \cos(\omega_d t)\,\sigma_x.$$2

so all bright-state splittings collapse onto a unique dark state at criticality. The corresponding eigenstates are displaced-and-squeezed superpositions, and the quantum metric components $$H_{\text{drive}}=\hbar \Omega \cos(\omega_d t)\,\sigma_x.$$3, $$H_{\text{drive}}=\hbar \Omega \cos(\omega_d t)\,\sigma_x.$$4 and Berry curvature $$H_{\text{drive}}=\hbar \Omega \cos(\omega_d t)\,\sigma_x.$$5 all diverge in the critical region. The divergence is markedly stronger for bright states than for the unique dark state, and it increases with excitation index $$H_{\text{drive}}=\hbar \Omega \cos(\omega_d t)\,\sigma_x.$$6 (Chen et al., 8 Feb 2026).

4. Dissipation, linewidths, and open-system formulations

Driven JC dynamics is typically modeled with Lindblad dissipation. For a cavity mode and a two-level system, the standard master equation is

$$H_{\text{drive}}=\hbar \Omega \cos(\omega_d t)\,\sigma_x.$$7

and, when pure dephasing is included,

$$H_{\text{drive}}=\hbar \Omega \cos(\omega_d t)\,\sigma_x.$$8

This structure underlies driven microcavity spectroscopy, driven cavity-versus-atom input–output theory, and the quantum regression treatment of emission spectra and $$H_{\text{drive}}=\hbar \Omega \cos(\omega_d t)\,\sigma_x.$$9 (Hopfmann et al., 2016, Fischer et al., 2018).

In the weakly driven regime, a microscopic dressed-basis treatment modifies both the vacuum Rabi splitting and the decoherence dynamics. For a resonant cavity-driven Hamiltonian

$$\mathcal{H}_0=i\hbar \lambda\left(\hat a\,\hat \sigma_+ - \hat a^\dagger \hat \sigma_-\right)+\hbar \epsilon(\hat a+\hat a^\dagger),$$0

the perturbatively corrected dressed energies are

$$\mathcal{H}_0=i\hbar \lambda\left(\hat a\,\hat \sigma_+ - \hat a^\dagger \hat \sigma_-\right)+\hbar \epsilon(\hat a+\hat a^\dagger),$$1

and the vacuum Rabi splitting becomes

$$\mathcal{H}_0=i\hbar \lambda\left(\hat a\,\hat \sigma_+ - \hat a^\dagger \hat \sigma_-\right)+\hbar \epsilon(\hat a+\hat a^\dagger),$$2

The same microscopic Liouvillian analysis predicts that, for highly inverted initial states in the weak-drive limit, the dominant decoherence oscillation frequency approaches

$$\mathcal{H}_0=i\hbar \lambda\left(\hat a\,\hat \sigma_+ - \hat a^\dagger \hat \sigma_-\right)+\hbar \epsilon(\hat a+\hat a^\dagger),$$3

twice the non-driven benchmark (Yu et al., 2016).

For damped driven JC and Rabi-type equations with polynomial pumping and damping, recent functional-analytic work establishes global well-posedness in the Hilbert space of Hermitian Hilbert–Schmidt operators. With time-independent pumping $$\mathcal{H}_0=i\hbar \lambda\left(\hat a\,\hat \sigma_+ - \hat a^\dagger \hat \sigma_-\right)+\hbar \epsilon(\hat a+\hat a^\dagger),$$4 and dissipation $$\mathcal{H}_0=i\hbar \lambda\left(\hat a\,\hat \sigma_+ - \hat a^\dagger \hat \sigma_-\right)+\hbar \epsilon(\hat a+\hat a^\dagger),$$5, the closed generator $$\mathcal{H}_0=i\hbar \lambda\left(\hat a\,\hat \sigma_+ - \hat a^\dagger \hat \sigma_-\right)+\hbar \epsilon(\hat a+\hat a^\dagger),$$6 produces a strongly continuous contraction semigroup

$$\mathcal{H}_0=i\hbar \lambda\left(\hat a\,\hat \sigma_+ - \hat a^\dagger \hat \sigma_-\right)+\hbar \epsilon(\hat a+\hat a^\dagger),$$7

provided both $$\mathcal{H}_0=i\hbar \lambda\left(\hat a\,\hat \sigma_+ - \hat a^\dagger \hat \sigma_-\right)+\hbar \epsilon(\hat a+\hat a^\dagger),$$8 and $$\mathcal{H}_0=i\hbar \lambda\left(\hat a\,\hat \sigma_+ - \hat a^\dagger \hat \sigma_-\right)+\hbar \epsilon(\hat a+\hat a^\dagger),$$9 are nonpositive on a dense finite-rank core. For time-dependent pumping and a polynomial GKSL-compatible dissipator, global generalized solutions exist in the weak matrix-entry sense, positivity is preserved for nonnegative initial data, and the Hilbert–Schmidt norm is nonincreasing. In that nonautonomous setting, positivity and contraction are established, whereas trace conservation is not proved (Komech et al., 18 Mar 2026, Komech et al., 18 Mar 2026).

5. Physical realizations and platform-specific mappings

A neutral-atom realization maps the JC ladder onto a Rydberg-dressed, blockaded ensemble. For $$\mathcal{H}_\eta=i\hbar\lambda'\left[(\hat a+\eta \hat a^\dagger)\hat \sigma_+ - (\hat a^\dagger+\eta \hat a)\hat \sigma_-\right]+\hbar \epsilon'(\hat a+\hat a^\dagger),$$0 atoms in the symmetric subspace with at most one collective Rydberg excitation,

$$\mathcal{H}_\eta=i\hbar\lambda'\left[(\hat a+\eta \hat a^\dagger)\hat \sigma_+ - (\hat a^\dagger+\eta \hat a)\hat \sigma_-\right]+\hbar \epsilon'(\hat a+\hat a^\dagger),$$1

and the rotating-frame Hamiltonian is

$$\mathcal{H}_\eta=i\hbar\lambda'\left[(\hat a+\eta \hat a^\dagger)\hat \sigma_+ - (\hat a^\dagger+\eta \hat a)\hat \sigma_-\right]+\hbar \epsilon'(\hat a+\hat a^\dagger),$$2

The identifications are

$$\mathcal{H}_\eta=i\hbar\lambda'\left[(\hat a+\eta \hat a^\dagger)\hat \sigma_+ - (\hat a^\dagger+\eta \hat a)\hat \sigma_-\right]+\hbar \epsilon'(\hat a+\hat a^\dagger),$$3

Autler–Townes spectroscopy then measures

$$\mathcal{H}_\eta=i\hbar\lambda'\left[(\hat a+\eta \hat a^\dagger)\hat \sigma_+ - (\hat a^\dagger+\eta \hat a)\hat \sigma_-\right]+\hbar \epsilon'(\hat a+\hat a^\dagger),$$4

and on resonance the splitting becomes $$\mathcal{H}_\eta=i\hbar\lambda'\left[(\hat a+\eta \hat a^\dagger)\hat \sigma_+ - (\hat a^\dagger+\eta \hat a)\hat \sigma_-\right]+\hbar \epsilon'(\hat a+\hat a^\dagger),$$5. For $$\mathcal{H}_\eta=i\hbar\lambda'\left[(\hat a+\eta \hat a^\dagger)\hat \sigma_+ - (\hat a^\dagger+\eta \hat a)\hat \sigma_-\right]+\hbar \epsilon'(\hat a+\hat a^\dagger),$$6, the measured ratio of two-atom to one-atom splitting is $$\mathcal{H}_\eta=i\hbar\lambda'\left[(\hat a+\eta \hat a^\dagger)\hat \sigma_+ - (\hat a^\dagger+\eta \hat a)\hat \sigma_-\right]+\hbar \epsilon'(\hat a+\hat a^\dagger),$$7, establishing the $$\mathcal{H}_\eta=i\hbar\lambda'\left[(\hat a+\eta \hat a^\dagger)\hat \sigma_+ - (\hat a^\dagger+\eta \hat a)\hat \sigma_-\right]+\hbar \epsilon'(\hat a+\hat a^\dagger),$$8 nonlinearity in a finite ladder $$\mathcal{H}_\eta=i\hbar\lambda'\left[(\hat a+\eta \hat a^\dagger)\hat \sigma_+ - (\hat a^\dagger+\eta \hat a)\hat \sigma_-\right]+\hbar \epsilon'(\hat a+\hat a^\dagger),$$9 (Lee et al., 2016).

A distinct collective-mode realization replaces the bosonic cavity mode by a weakly excited symmetric atomic ensemble. A control atom, driven by an auxiliary classical field and coupled dispersively to a nonresonant cavity, interacts with the ensemble mode

$\eta\in[0,1]$0

in the low-excitation limit. After cavity elimination, the effective Hamiltonian is JC-like,

$\eta\in[0,1]$1

with collective enhancement

$\eta\in[0,1]$2

In the Stark-tuned regime, resonance is set by

$\eta\in[0,1]$3

allowing the drive to switch between resonant and dispersive JC dynamics for state engineering and tomography of the collective atomic mode (Zheng, 2012).

In trapped ions, the driven JC interaction arises on motional sidebands of a classically driven two-level ion. Beyond the Lamb–Dicke regime, the interaction-picture Hamiltonian is

$\eta\in[0,1]$4

and, in each coupled block $\eta\in[0,1]$5, the effective Rabi frequency is

$\eta\in[0,1]$6

with $\eta\in[0,1]$7 determined by generalized Laguerre-polynomial matrix elements. The exact Magnus convergence radius depends explicitly on detuning, and operator time ordering materially affects the nonclassicality of the motional state at longer times (Lipfert et al., 2018).

In strongly dispersive circuit QED, a driven JC oscillator behaves as a nonlinear cavity with an amplitude-dependent dispersive shift

$\eta\in[0,1]$8

Its semiclassical steady-state response exhibits a finite bistable region bounded by

$\eta\in[0,1]$9

and an upper critical point

$a(t)\to b(t)+\alpha(t)$0

beyond which the response returns toward the bare cavity line. This bounded bistability is a direct consequence of the saturability of the JC nonlinearity and differs from Kerr/Duffing behavior (Bishop et al., 2010).

6. Control protocols, engineered interactions, and many-body extensions

Driven JC interactions have become a control primitive rather than only a spectroscopic probe. In one-dimensional JC lattices restricted to the one-excitation manifold, exact counterdiabatic driving is nonlocal in real space, but symmetry allows an exactly equivalent local implementation. For uniform couplings $a(t)\to b(t)+\alpha(t)$1, $a(t)\to b(t)+\alpha(t)$2, and detuning $a(t)\to b(t)+\alpha(t)$3, the local counterdiabatic Hamiltonian can be written as

$a(t)\to b(t)+\alpha(t)$4

with

$a(t)\to b(t)+\alpha(t)$5

For the ground-state manifold, this local construction reproduces the exact counterdiabatic dynamics with numerical agreement better than $a(t)\to b(t)+\alpha(t)$6. In the reported $a(t)\to b(t)+\alpha(t)$7 example, local counterdiabatic driving also enables $a(t)\to b(t)+\alpha(t)$8-state preparation with infidelity below $a(t)\to b(t)+\alpha(t)$9 at $\omega_c$00 even when $\omega_c$01 and $\omega_c$02 are included (Govindarajan et al., 2024).

Another route engineers an ultrastrong JC Hamiltonian by synchronously modulating the qubit and cavity frequencies in the quantum Rabi model,

$\omega_c$03

Because the modulation cancels from the rotating-wave phase but doubles in the counter-rotating phase, the effective JC interaction survives with

$\omega_c$04

while counter-rotating sidebands are suppressed. In the high-frequency regime, choosing $\omega_c$05 at a zero of $\omega_c$06 removes the central counter-rotating component; in the low-frequency regime, a short-time approximation detunes counter-rotating processes by $\omega_c$07. The resulting effective Hamiltonian is

$\omega_c$08

and the scheme opens access to JC physics across the ultrastrong and deep-strong regimes (Huang et al., 2019).

At the gate level, sideband-driven JC control now supports universal oscillator logic. In a superconducting cavity–transmon platform, a drive near the $\omega_c$09 sideband generates

$\omega_c$10

and, within each doublet $\omega_c$11,

$\omega_c$12

Compiling arbitrary oscillator gates into alternating ancilla rotations and closed JC layers produces native unitaries that are leakage-protected below a chosen photon cutoff. In the reported experiment, universal qutrit control is demonstrated with a mean post-selected process fidelity of $\omega_c$13, along with ququart and ququint shift gates (Huang et al., 18 May 2026).

Taken together, these developments show that the driven Jaynes–Cummings system is no longer a single model with a weak probe attached to it. It is a family of controllable, platform-dependent Hamiltonians whose driven structure governs spectroscopy, criticality, dissipation, and compilation. The same basic ingredients—exchange $\omega_c$14, classical drive, and frame engineering—support Autler–Townes ladders, finite and infinite quasienergy structures, bistable nonlinear response, counterdiabatic state preparation, and programmable bosonic quantum control across neutral atoms, semiconductors, trapped ions, and superconducting hardware.