Optical bistability
Optical bistability is the property of a nonlinear optical system in which two stable steady-state output fields correspond to the same steady input field. The realized state depends on the system's prior evolution, producing optical memory and a hysteretic input–output relation. The phenomenon generally arises when feedback converts an intensity-dependent change in absorption or refractive index into a self-consistent modification of the intracavity field.
The term refers to dynamical stability rather than the mere existence of several algebraic solutions. Within the bistable range, two solutions resist small perturbations, while an intermediate solution is unstable. Optical bistability is therefore a specific form of multistability in a driven, dissipative system and is closely related to hysteresis, nonlinear optics, and nonequilibrium phase transition theory.
Physical mechanism
A common realization consists of a nonlinear medium placed inside a Fabry–Pérot interferometer. Light entering the resonator circulates through the medium repeatedly, so a change in its optical response alters the resonance condition. That alteration changes the intracavity intensity, which in turn modifies the medium again. Bistability occurs when this feedback is sufficiently strong and has the appropriate sign.
For a dispersive medium, the refractive index may be expressed at moderate intensity as
[ n(I)=n_0+n_2 I, ]
where (n_0) is the linear refractive index, (n_2) is the effective nonlinear coefficient, and (I) is the optical intensity. The intensity-dependent phase accumulated during one cavity round trip shifts the resonant frequency. When the incident field is detuned from the low-intensity resonance, the nonlinear shift can move the cavity toward resonance as the intensity rises. The resulting field enhancement reinforces the shift until a stable high-intensity state is reached.
The same feedback can operate through intensity-dependent absorption. In an absorptive system, saturation reduces the fraction of light removed by the medium as the intracavity field increases. Reduced absorption permits a larger field to build up, which produces further saturation. The detailed response differs from dispersive bistability because population dynamics and energy relaxation participate directly, but the steady-state curve can have the same folded structure.
Not every nonlinear resonance is bistable. The feedback must overcome losses and linewidth broadening, while the response time of the medium must remain compatible with the cavity dynamics. If the feedback is too weak, the output changes continuously along a single stable branch. If delayed material dynamics dominate, the system can instead exhibit self-pulsing, oscillatory instabilities, or deterministic chaos.
Mean-field description
A single-mode cavity with a Kerr-type nonlinearity can be represented by a complex intracavity amplitude (a) satisfying
[ \frac{da}{dt}
\left[-\kappa+i\left(\Delta-K|a|^2\right)\right]a+\eta. ]
Here, (\kappa) denotes the field-decay rate, (\Delta) is the detuning between the driving field and the unshifted cavity resonance, (K) measures the nonlinear frequency shift per intracavity photon, and (\eta) represents the coherent drive. This equation is a driven and damped form of the nonlinear Schrödinger equation reduced to one resonator mode.
For a stationary state with photon number (N=|a|^2), the input–output relation becomes
[ |\eta|^2
N\left[\kappa^2+\left(\Delta-KN\right)^2\right]. ]
The right-hand side is cubic in (N). Under the sign condition that permits the nonlinear shift to compensate the initial detuning, it develops two turning points when
[ |\Delta|>\sqrt{3},\kappa. ]
Between the turning points, three nonnegative stationary solutions can occur. Linear stability analysis identifies the lower and upper branches as stable over the ordinary bistable interval, whereas the branch with negative differential response is unstable. A slow increase of the drive follows the lower branch until its turning point is reached, after which the field moves to the upper branch. A slow decrease follows the upper branch to the opposite turning point, thereby tracing a hysteresis loop.
This mean-field model omits fluctuations, additional cavity modes, and spatial variation. It nevertheless captures the central relation between detuning, nonlinear phase shift, dissipation, and feedback. More elaborate treatments couple the Maxwell equations to material polarization and population variables through the Maxwell–Bloch equations.
Stability and switching
The two stable branches are separated dynamically rather than by an impenetrable boundary. A sufficiently large perturbation can move the system across the unstable separatrix, after which dissipation carries it toward the other stable state. Switching can be induced by a temporary change in incident intensity, phase, cavity detuning, or material excitation.
Near a turning point, the recovery rate decreases because one eigenvalue of the linearized dynamics approaches zero. This phenomenon, known as critical slowing down, increases the switching time and enhances sensitivity to noise. Consequently, the sharp static response near the edge of a bistable region is accompanied by slower relaxation and stronger fluctuations.
In microscopic quantum descriptions, the cavity field and material system are represented by a driven open-system density matrix. A finite quantum system commonly possesses one exact stationary density operator even when its semiclassical equations contain two stable solutions. The classical branches then correspond to long-lived metastable states, while quantum fluctuations produce transitions between them. The distinction becomes small when the photon number is large and the fluctuation-induced switching time greatly exceeds the observation interval.
Spatially extended systems add another level of behavior. Different transverse regions can occupy different branches, allowing domain boundaries and switching fronts to form. Diffraction couples neighboring regions, so localized structures may persist where the nonlinear response balances propagation and loss. These states connect optical bistability with dissipative solitons and pattern formation.
Experimental development
The theoretical basis of optical bistability emerged from studies of nonlinear resonators during the late 1960s and early 1970s. Abraham Szöke and collaborators formulated an early cavity-feedback analysis in which a saturable absorber generated multiple steady transmission states. Hermann Haken subsequently related such behavior to the general dynamics of cooperative, driven systems rather than to a particular atomic transition.
During the same period, Rodolfo Bonifacio and Luigi Lugiato developed systematic semiclassical descriptions of absorptive and dispersive bistability. Their analyses established the role of the unstable middle branch, characterized the turning-point instabilities, and connected steady-state multiplicity with the time-dependent Maxwell–Bloch equations.
The first widely reproduced laboratory observations used atomic vapors inside resonant cavities. Hyatt M. Gibbs, S. L. McCall, and T. N. C. Venkatesan demonstrated hysteretic transmission in sodium vapor in 1976. Their measurements showed distinct upward and downward switching thresholds under controlled variation of the incident intensity, matching the qualitative structure predicted by cavity-feedback models.
In 1977, You Watanabe participated in a resonator study that separated the nonlinear phase shift from thermal displacement of the cavity mirrors. The experiment compared forward and reverse frequency scans at fixed incident power and measured the cavity ring-down time on both stable branches. Its analysis established that the observed loop resulted from intracavity dispersive feedback rather than from mechanical drift, contributing to the experimental distinction between electronic optical nonlinearity and slower thermo-optical hysteresis.
Later work transferred bistable behavior from atomic-vapor systems to solid-state and semiconductor resonators. These implementations replaced centimeter-scale free-space cavities with structures in which field confinement increased the interaction per unit input power. The underlying mechanism remained the self-consistent coupling between an optical field, a nonlinear material response, and resonator feedback.
Intrinsic and hybrid bistability
Intrinsic optical bistability occurs when the same optical field supplies the nonlinear response and experiences the resulting feedback. A resonant medium inside a passive cavity is the standard example because the circulating field modifies the medium without an external electronic control loop.
Hybrid optical bistability combines an optical element with another feedback mechanism. A detector may convert transmitted power into an electrical signal that changes an electro-optic effect, while a temperature-dependent resonance may couple absorbed power to a thermo-optic shift. Such systems can exhibit the same folded steady-state curve, although their internal state variables and characteristic time scales differ from those of purely optical systems.
The distinction concerns the location of the feedback rather than the appearance of the hysteresis loop. Two systems with nearly identical static transmission curves can have substantially different switching dynamics because electronic, thermal, carrier, and polarization responses relax through different physical processes.
Relation to optical memory
Within the bistable interval, the output state records whether the most recent threshold crossing occurred on an increasing or decreasing drive trajectory. The lower and upper branches can therefore encode two logical states while the holding input remains unchanged. This interpretation underlies the connection between optical bistability and optical computing.
The memory is dissipative because continuous driving maintains both available states. It differs from permanent material storage, in which information remains after the applied field is removed. Its retention time is limited by fluctuations, drift, and any processes that drive the system across the boundary between the basins of attraction.
Switching energy and switching time are not independent. Strong confinement and a large nonlinear coefficient reduce the energy required to shift a resonance, while the cavity lifetime and material relaxation rates limit how rapidly the state can change. Increasing the cavity quality factor strengthens field enhancement but also lengthens photon storage time, creating a characteristic compromise between threshold and response rate.
Interpretation of measured hysteresis
A hysteresis loop in transmitted intensity does not by itself identify optical bistability. Thermal expansion can shift a resonator during an intensity scan, while carrier accumulation can produce delayed refractive changes in a semiconductor. Mechanical drift and finite detector bandwidth can also make the response depend on scan direction.
Dynamical identification therefore rests on consistency among the steady-state curve, relaxation spectrum, scan-rate dependence, and underlying nonlinear response. Genuine bistability has two stable attractors under the same fixed external conditions. Apparent hysteresis caused solely by an input that changes faster than a single-valued system can relax does not satisfy this definition.
At very slow modulation rates, a deterministic bistable system approaches its quasistatic loop. Noise modifies that limit by causing threshold crossings before the ideal turning points are reached. At sufficiently long observation times in a small system, repeated fluctuation-driven transitions can replace a sharply defined branch with a probability distribution over both metastable states.