Dissipative Optomechanical Coupling

• Dissipative optomechanics is characterized by mechanical modulation of optical loss channels, generating unique interference and energy exchange phenomena.
• It enables tunable regimes where squeezing, cooling beyond the resolved-sideband limit, and backaction suppression are achieved through controlled linewidth modulation.
• Applications span quantum-limited sensing, engineered non-Hermitian dynamics, and transduction in platforms like Fabry–Perot cavities, nanobeams, and interferometers.

Searching arXiv for recent and foundational papers on dissipative optomechanical coupling. Dissipative optomechanical coupling is the class of optomechanical interaction in which mechanical displacement modulates an optical decay channel or linewidth, rather than only shifting an optical resonance frequency. In the standard dispersive case, motion changes the cavity eigenfrequency, typically quantified by derivatives such as $g_\omega=d\omega_c/dx$ or $G_\omega=-d\omega_0/dx$. In the dissipative case, motion changes a leakage, coupling, or loss rate, with representative definitions including $g_\gamma=d\gamma_2/dx$, $G_{\kappa_e}=d\kappa_e/dx$, or $g_\kappa=\partial\kappa/\partial Q$ (Wang et al., 9 Jun 2026, Primo et al., 2022, Qu et al., 2015). This distinction is not merely terminological: dissipative coupling alters the way the cavity exchanges energy with external or internal baths, generates interference between direct bath-mediated and cavity-mediated pathways, and thereby produces phenomena absent or strongly modified relative to purely dispersive optomechanics, including Fano backaction spectra, cooling beyond the resolved-sideband paradigm, on-resonance squeezing mechanisms, nontrivial stability structure, and measurement protocols with suppressed dynamical backaction [(Weiss et al., 2012); (Xuereb et al., 2011); (Tagantsev et al., 2019)].

1. Definition and physical content

In dissipative optomechanics, the mechanical coordinate modulates an optical loss channel. Depending on platform, that loss may be an external waveguide coupling rate, a port-specific decay rate, a taper-induced extraction channel, or an internal absorptive or scattering loss [(Primo et al., 2022); (Hryciw et al., 2014); (Xiao et al., 2014); (Baraillon et al., 2020)]. The defining contrast with dispersive coupling is therefore between displacement-induced changes in $\omega_c$ and displacement-induced changes in $\kappa$, $\gamma$, or related decay constants.

This distinction is explicit in several formalisms. A generic mixed-coupling description uses

$$\omega_c(u_m)=\omega_{c0}+g_{\mathrm{om}}u_m,\qquad \kappa_e(u_m)=\kappa_{e0}+\kappa_{\mathrm{om}}^e u_m,\qquad \kappa_i(u_m)=\kappa_{i0}+\kappa_{\mathrm{om}}^i u_m,$$

so that motion may modulate the cavity resonance frequency, the external coupling rate, and the intrinsic loss rate simultaneously (Baraillon et al., 2020). In Fabry–Perot implementations with a string resonator inserted into the standing-wave field, the two central couplings are written as

$$g_\gamma=\frac{d\gamma_2}{dx},\qquad g_\omega=\frac{d\omega_c}{dx},$$

where the same displacement changes both linewidth and resonance through position-dependent scattering and optical-path perturbation (Wang et al., 9 Jun 2026).

The physical origin depends on geometry. In waveguide-cavity systems, motion can modulate how strongly photons are exchanged with the outside world; in cavity–fiber near-field systems, motion changes the external decay rate $G_\omega=-d\omega_0/dx$0; in graphene microcavities, motion changes absorption loss through interband transitions; in Michelson–Sagnac interferometers (MSIs), motion changes an effective cavity linewidth by modifying interference at a synthetic mirror [(Hryciw et al., 2014); (Xiao et al., 2014); (Xuereb et al., 2011)]. A plausible implication is that “dissipative coupling” is best understood not as a single microscopic mechanism but as a family of loss-modulation mechanisms sharing a common reduced description.

2. Canonical models and interference structure

The minimal theoretical distinction between dispersive and dissipative coupling appears already at the level of linearized equations of motion and input-output relations. In the dissipative case, the coupling enters not only through the intracavity susceptibility but also through the boundary condition to the bath. This is the structural source of the characteristic interference phenomena emphasized across the literature [(Qu et al., 2015); (Weiss et al., 2012); (Tagantsev et al., 2019)].

A representative mixed Hamiltonian is given in the form

$G_\omega=-d\omega_0/dx$1

with $G_\omega=-d\omega_0/dx$2 and $G_\omega=-d\omega_0/dx$3 (Weiss et al., 2012). The dissipative contribution modifies the input-output law directly: $G_\omega=-d\omega_0/dx$4 making the mechanics enter the bath coupling itself rather than only the cavity resonance (Weiss et al., 2012).

This direct bath coupling leads to a Fano force spectrum. In the quantum-noise treatment of mixed coupling,

$G_\omega=-d\omega_0/dx$5

and the force noise vanishes at

$G_\omega=-d\omega_0/dx$6

which is the interference zero responsible for several nonstandard cooling and amplification regimes (Weiss et al., 2012). The same interference picture appears in high-frequency nanobeams, where mechanical spectra become asymmetric because the phase of the intracavity field flips between red and blue detuning while the waveguide field phase does not, producing constructive or destructive interference depending on detuning (Primo et al., 2022).

In squeezing theory, the same structural feature reappears in the modified input-output relation. For a single cavity with linewidth modulation,

$G_\omega=-d\omega_0/dx$7

or in quadratures,

$G_\omega=-d\omega_0/dx$8

This makes input noise act both directly and indirectly through the mechanics, enabling coherent interference between two noise paths (Qu et al., 2015).

3. Experimental realizations and tunability

A major development in the subject is the emergence of platforms in which dispersive and dissipative couplings are independently tunable or continuously reweighted. This has shifted dissipative optomechanics from a niche perturbation to a configurable interaction channel.

Representative platforms

Platform	Dissipative mechanism	Representative result
Fabry–Perot cavity with string resonator	Position-dependent scattering and linewidth broadening	Coupling ratio tuned from 1.3 to 0.6 experimentally; 25 to 0.02 theoretically (Wang et al., 9 Jun 2026)
Michelson–Sagnac interferometer	Motion-dependent effective mirror transmissivity and linewidth	Strong and tunable dissipative coupling with suppressed lower motional sideband (Xuereb et al., 2011)
Split-beam nanocavity with fiber taper	Motion-dependent external decay $G_\omega=-d\omega_0/dx$9 into the taper	Crossover between dissipative and dispersive transduction via fiber position (Hryciw et al., 2014)
High-frequency photonic-crystal nanobeams	Motion-dependent redistribution of supermode losses $g_\gamma=d\gamma_2/dx$0	First dissipative system in the sideband-resolved regime with $g_\gamma=d\gamma_2/dx$1 (Primo et al., 2022)
Graphene in microcavity	Motion-dependent absorptive loss	Strong and tunable dissipative coupling via absorption modulation (Xiao et al., 2014)
Membrane-outside two-sided cavity	Synthetic mirror transmission slope with vanishing phase slope	Regime of strong dissipative coupling with suppressed dispersive coupling (Tagantsev et al., 2021)

In the 2026 Fabry–Perot/string implementation, dissipative coupling means that motion changes how strongly the cavity leaks light as well as its resonance frequency. The optical response is calculated with a transfer-matrix model, and reflection, transmission, and scattering are written as

$g_\gamma=d\gamma_2/dx$2

By varying resonator diameter, material, and placement, the dissipative-to-dispersive coupling ratio is tuned from a dissipation-dominated value of $g_\gamma=d\gamma_2/dx$3 for a $g_\gamma=d\gamma_2/dx$4 iron-wire resonator to a dispersion-dominated value of $g_\gamma=d\gamma_2/dx$5 for a $g_\gamma=d\gamma_2/dx$6 fiber-optic resonator, while the theoretically accessible range is $g_\gamma=d\gamma_2/dx$7 to $g_\gamma=d\gamma_2/dx$8 (Wang et al., 9 Jun 2026).

In the high-frequency nanobeam system, two coupled optical resonators support supermodes whose external loss rates $g_\gamma=d\gamma_2/dx$9 depend on mechanically induced detuning between the resonators. This realizes the first dissipative optomechanical system in the sideband-resolved regime with $G_{\kappa_e}=d\kappa_e/dx$0 and $G_{\kappa_e}=d\kappa_e/dx$1, together with a two-order-of-magnitude increase in mechanical frequency and a tenfold increase in dissipative coupling rate relative to previous work (Primo et al., 2022).

Near-field tuning in split-beam nanocavities shows another route to tunability. A fiber taper introduces a position-dependent decay channel $G_{\kappa_e}=d\kappa_e/dx$2, and the mechanically induced transmission response obeys

$G_{\kappa_e}=d\kappa_e/dx$3

The relative balance is parameterized by

$G_{\kappa_e}=d\kappa_e/dx$4

with $G_{\kappa_e}=d\kappa_e/dx$5 indicating dissipative dominance and $G_{\kappa_e}=d\kappa_e/dx$6 dispersive dominance (Hryciw et al., 2014). The measured crossover near $G_{\kappa_e}=d\kappa_e/dx$7 nm directly established that the two transduction mechanisms can be switched by probe geometry alone.

A related but conceptually sharper architecture is the “membrane-outside” system, where a membrane and adjacent mirror form a synthetic mirror whose transmission $G_{\kappa_e}=d\kappa_e/dx$8 and reflection phase $G_{\kappa_e}=d\kappa_e/dx$9 depend differently on displacement. The dispersive coupling vanishes when $g_\kappa=\partial\kappa/\partial Q$0, while the dissipative coupling survives through $g_\kappa=\partial\kappa/\partial Q$1. Under the condition

$g_\kappa=\partial\kappa/\partial Q$2

the dissipative coupling exceeds the dispersive coupling constant for an optomechanical cavity of the same length (Tagantsev et al., 2021). This suggests a particularly clean route to nominally pure dissipative coupling.

4. Backaction, cooling, and stability

Dissipative coupling changes both the magnitude and topology of optomechanical backaction. Unlike purely dispersive coupling, where cooling is largely organized by sideband resolution and detuning to the red mechanical sideband, dissipative systems exhibit interference-based suppression of heating, multiple cooling and amplification regions, and, in some regimes, elimination of dynamical backaction altogether [(Weiss et al., 2012); (Xuereb et al., 2011); (Tagantsev et al., 2019)].

In the early MSI proposal, quantum interference suppresses the lower motional sideband, allowing strongly enhanced cooling in the non-sideband-resolved regime (Xuereb et al., 2011). In the quantum-noise treatment of mixed coupling, the optical damping is $g_\kappa=\partial\kappa/\partial Q$3, with

$g_\kappa=\partial\kappa/\partial Q$4

Because $g_\kappa=\partial\kappa/\partial Q$5 is Fano-shaped rather than purely Lorentzian, dissipatively coupled systems possess two parameter regions providing amplification and two parameter regions providing cooling (Weiss et al., 2012). The cooling optimum associated with the interference zero occurs near

$g_\kappa=\partial\kappa/\partial Q$6

where $g_\kappa=\partial\kappa/\partial Q$7 and $g_\kappa=\partial\kappa/\partial Q$8 in the weak-coupling theory (Weiss et al., 2012).

The high-frequency sideband-resolved experiment demonstrated that dissipative coupling reshapes both mechanical and optical spectra. The extra damping rate in the resolved-sideband regime is written as

$g_\kappa=\partial\kappa/\partial Q$9

which makes explicit that dispersive and dissipative terms interfere at the amplitude level (Primo et al., 2022). Experimentally, linewidth narrowing for cooling, linewidth broadening for heating, and selective phonon lasing of one mechanical mode were observed (Primo et al., 2022).

A distinct branch of the literature emphasizes that dissipative coupling can remove rather than merely reshape backaction. In a lossy two-port cavity where the oscillator modulates the loss rate $\omega_c$0 of an undetected port, the force in the pure dissipative case is

$\omega_c$1

and there is complete absence of dynamical backaction, hence no optomechanical instability, at critical coupling $\omega_c$2 in the unresolved-sideband regime (Tagantsev et al., 2019). This same setting saturates the Heisenberg imprecision–backaction product,

$\omega_c$3

under critical coupling and $\omega_c$4 (Tagantsev et al., 2019). A plausible implication is that dissipative coupling may interpolate between strongly interference-shaped backaction and effectively backaction-free sensing, depending on port geometry and matching conditions.

Stability theory also differs qualitatively from the dispersive case. For stable optical rigidity in a Fabry–Perot cavity with linewidth-modulating coupling, the detuning and pump must satisfy

$\omega_c$5

together with

$\omega_c$6

yielding a stable optical spring without external feedback (Nazmiev et al., 2018). By contrast, in the bad-cavity limit of a one-sided cavity, purely dissipative coupling suppresses backaction strongly through destructive interference of vacuum-noise pathways, and the squeezing ability is strongly suppressed; the corresponding stability diagram is qualitatively identical to the dispersive one after a detuning-sign reversal, $\omega_c$7 (Tagantsev et al., 2018). These results are not contradictory; they arise in different geometries and operating assumptions.

5. Spectral signatures, squeezing, and quantum correlations

Dissipative coupling leaves identifiable signatures in mechanical spectra, optical transmission, optomechanically induced transparency or absorption, and quantum-noise spectra. Several of these signatures follow directly from the interference structure of the coupling.

In the strong-coupling analysis of dissipatively coupled systems, exact linearized solutions show normal-mode splitting similar to purely dispersive coupling, but the spectra contain an additional feature traced to the Fano line shape of the force spectrum (Weiss et al., 2012). The output spectrum differs from the dispersive case because there is a direct mechanics-to-output contribution not cavity-filtered in the standard way, producing altered sideband asymmetry and exact zeros for certain detunings (Weiss et al., 2012).

OMIT and OMIA provide especially transparent diagnostics. The 2012 theory showed that purely dissipative coupling can lead to optomechanically induced transparency and that this offers an experimentally convenient route to observing normal-mode splitting (Weiss et al., 2012). The 2022 nanobeam experiment reported the first observation of dissipative signatures in OMIT/OMIA, where the depth and width of the transparency window are modified by the presence of dissipative coupling (Primo et al., 2022). The probe response is described by a dressed sideband amplitude,

$\omega_c$8

showing explicitly how dissipative and dispersive couplings reshape the OMIT line (Primo et al., 2022).

Dissipative coupling also supports nonclassical optical states. For on-resonance drive, purely dissipative coupling yields

$\omega_c$9

so the mechanics feeds $\kappa$0 into $\kappa$1, complementary to the standard dispersive ponderomotive mechanism (Qu et al., 2015). The approximate dissipative squeezing spectrum is

$\kappa$2

with optimal variance

$\kappa$3

Under appropriate matching,

$\kappa$4

dissipative and dispersive schemes yield comparable maximal squeezing (Qu et al., 2015).

Mechanical squeezing transfer is another hallmark. In a cavity driven by a coherent laser plus a weaker broadband squeezed vacuum, dissipative coupling allows complete destructive interference of quantum noise in the weak-coupling limit, so that the cavity acts as an effective squeezed reservoir for the mirror (Gu et al., 2013). Under the cooling condition that nulls the heating process, the steady-state quadrature variances become

$\kappa$5

and this transfer is described as irrespective of the ratio between the cavity damping rate and the mechanical frequency (Gu et al., 2013). For moderate coupling, photonic excitation precludes complete destructive interference and the mirror deviates from the ideal squeezed state (Gu et al., 2013).

Quantum entanglement has more recently been explored in a dissipative setting using an MSI with a movable membrane. In that system, purely dissipative coupling yields stronger and more noise-tolerant optomechanical entanglement than purely coherent coupling, with optimal logarithmic negativity near

$\kappa$6

and entanglement remaining robust up to about $\kappa$7 K for $\kappa$8 Hz (Chen et al., 16 Jan 2025). When coherent and dissipative couplings coexist, entanglement is weakened due to quantum interference (Chen et al., 16 Jan 2025). This suggests that mixed coupling is not generically advantageous for every quantum resource.

6. Regimes, applications, and open issues

Dissipative optomechanical coupling is now associated with several application domains: quantum-limited displacement measurement, force sensing, cooling of low-frequency or massive resonators, transduction, tunable mechanical-mode control, squeezed-light generation, and engineered non-Hermitian dynamics (Wang et al., 9 Jun 2026, Primo et al., 2022, Tagantsev et al., 2019, Nazmiev et al., 2018).

In sensing, the photonic-crystal torque sensor demonstrated experimentally observed dispersive and dissipative couplings with $\kappa$9, $\gamma$0, and $\gamma$1, together with thermally limited torque sensitivities of $\gamma$2 in ambient conditions and $\gamma$3 in low vacuum (Wu et al., 2014). The key lesson of that work is that a numerically small external dissipative coupling can dominate transduction because the optical response near resonance is extremely sensitive to $\gamma$4 in an undercoupled device (Wu et al., 2014).

In cooling and transduction, the MSI proposal highlighted that dissipative coupling allows ground-state cooling and low-power quantum-limited position transduction in the non-sideband-resolved regime (Xuereb et al., 2011). The 2026 Fabry–Perot/string system presents continuous tuning between dissipation-dominated and dispersion-dominated regimes on the same experimental platform, which the authors position as useful for exploring quantum effects of massive mechanical resonators and quantum-limited measurements (Wang et al., 9 Jun 2026). This suggests a movement in the field from identifying dissipative effects toward using the dissipative-to-dispersive ratio itself as an experimental control parameter.

The subject also connects to non-Hermitian and chiral physics. In dissipatively coupled cascaded optomechanical systems, adiabatic elimination of fast cavity modes yields an effective directional mechanical coupling

$\gamma$5

with asymmetric steady-state discord and a temperature gradient $\gamma$6 in the parameter sets studied (Pellitteri et al., 2023). In $\gamma$7-symmetric mechanically coupled optomechanical dimers, dissipative coupling shifts the exceptional point to lower driving strength and suppresses chaotic beats in the nonlinear regime (Tchounda et al., 2023). Full-quantum MSI analyses further associate dissipative coupling with an imaginary-Kerr-like nonlinearity,

$\gamma$8

and strong antibunching near the condition $\gamma$9 (Yang et al., 2023).

Several misconceptions recur in the literature. One is that dissipative coupling is simply weaker than dispersive coupling because the vacuum dissipative coupling rate may be numerically smaller. The high-frequency nanobeam work showed that even when $$\omega_c(u_m)=\omega_{c0}+g_{\mathrm{om}}u_m,\qquad \kappa_e(u_m)=\kappa_{e0}+\kappa_{\mathrm{om}}^e u_m,\qquad \kappa_i(u_m)=\kappa_{i0}+\kappa_{\mathrm{om}}^i u_m,$$0 is only about $$\omega_c(u_m)=\omega_{c0}+g_{\mathrm{om}}u_m,\qquad \kappa_e(u_m)=\kappa_{e0}+\kappa_{\mathrm{om}}^e u_m,\qquad \kappa_i(u_m)=\kappa_{i0}+\kappa_{\mathrm{om}}^i u_m,$$1, waveguide-mediated enhancement makes dissipative effects clearly observable (Primo et al., 2022). Another is that dissipative coupling is always beneficial for squeezing because it suppresses some backaction channels. In fact, in the bad-cavity regime of a one-sided cavity, purely dissipative coupling strongly reduces backaction and therefore strongly suppresses squeezing ability relative to the purely dispersive case (Tagantsev et al., 2018). A third is that “dissipative” necessarily means internal absorption; the literature distinguishes external dissipative coupling, intrinsic dissipative coupling, and absorption-mediated coupling, with different signatures and different optimal regimes [(Baraillon et al., 2020); (Xiao et al., 2014)].

Open issues remain centered on controllable regime selection, coexistence with dispersive coupling, and robustness against technical noise. The available evidence indicates that mixed coupling can either enhance readout through interference or weaken targeted quantum correlations through destructive interference, depending on observable and platform [(Wu et al., 2014); (Chen et al., 16 Jan 2025)]. A plausible implication is that future progress will depend less on maximizing a single coupling coefficient than on engineering the full interference landscape among resonance shifts, linewidth modulation, port asymmetry, and bath structure.