Nonlinear optics

Nonlinear optics is the study of optical phenomena in which the response of a material depends nonlinearly on the amplitude of an applied electromagnetic field. In the linear regime, the induced polarization is proportional to the electric field, and light waves propagate independently except where linear interference or scattering couples their amplitudes. At sufficiently high field strengths, higher-order terms in the material polarization become measurable. These terms permit interactions among optical frequencies and produce intensity-dependent changes in propagation.

Most nonlinear optical effects became experimentally accessible after the development of the laser, whose high irradiance and temporal coherence made weak nonlinear responses observable under controlled conditions. The subject encompasses frequency conversion, nonlinear refraction, multiphoton absorption, parametric amplification, and the formation of optical solitons. Its theoretical description combines electromagnetism, quantum mechanics, and the symmetry properties of matter.

Nonlinear polarization

The macroscopic electric polarization (\mathbf{P}) of a material can be expanded as a power series in the applied electric field (\mathbf{E}):

[ P_i(t)=\varepsilon_0\left[ \sum_j \chi^{(1)}{ij}E_j(t) +\sum{jk}\chi^{(2)}{ijk}E_j(t)E_k(t) +\sum{jkl}\chi^{(3)}_{ijkl}E_j(t)E_k(t)E_l(t) +\cdots \right]. ]

Here, (\varepsilon_0) is the vacuum permittivity, while (\chi^{(1)}) is the linear electric susceptibility. The higher-rank tensors (\chi^{(2)}) and (\chi^{(3)}) describe second-order and third-order nonlinear responses. Their values generally depend on the participating frequencies, polarization directions, thermodynamic state, and microscopic electronic structure of the medium.

This expansion is local and instantaneous only as an approximation. A dispersive response depends on the field at earlier times, so each susceptibility is more generally represented by a temporal response function. Transformation into the frequency domain converts the corresponding time convolutions into products constrained by frequency conservation. Spatial nonlocality can also become relevant near material resonances or where the optical field varies over microscopic length scales.

The expansion remains useful when the nonlinear polarization is small relative to the linear polarization. In stronger fields, ionization, plasma production, structural modification, or saturation can invalidate a perturbative description. Such regimes are treated using nonperturbative models of light–matter interaction.

Second-order response

A second-order polarization generated by two monochromatic fields at angular frequencies (\omega_1) and (\omega_2) contains components at their sum and difference. The sum-frequency component satisfies

[ \omega_3=\omega_1+\omega_2, ]

whereas the difference-frequency component satisfies

[ \omega_3=\lvert\omega_1-\omega_2\rvert. ]

When both input fields have the same frequency, the sum-frequency process becomes second-harmonic generation, producing radiation at (2\omega). A zero-frequency component is generated simultaneously and corresponds to optical rectification, in which an optical field produces a static or slowly varying polarization.

Within the electric-dipole approximation, a centrosymmetric bulk material has no second-order susceptibility. Spatial inversion changes (\mathbf{E}) and (\mathbf{P}) by the same sign, while a quadratic field product remains unchanged. Consistency therefore requires (\chi^{(2)}=0) in the bulk. Second-order interactions can nevertheless occur at an interface, where inversion symmetry is broken, or through magnetic-dipole and electric-quadrupole contributions that lie beyond the simplest approximation.

The efficiency of a coherent second-order process depends on the relative phase accumulated by the interacting waves. For second-harmonic generation, the wave-vector mismatch is

[ \Delta k=k(2\omega)-2k(\omega). ]

When (\Delta k=0), nonlinear polarization generated at different positions remains in phase with the emitted harmonic field. When the mismatch is nonzero, the generated contributions periodically reinforce and cancel one another. The associated coherence length is commonly written as (L_c=\pi/\lvert\Delta k\rvert).

Phase matching compensates for material dispersion by using propagation direction, polarization dependence, or an engineered spatial modulation. In a birefringent crystal, differently polarized waves experience distinct refractive indices, allowing the phase velocities of the interacting frequencies to be matched. In quasi-phase matching, the sign of the nonlinear coefficient is reversed at selected intervals so that growth resumes before destructive interference dominates.

Third-order response

Third-order nonlinearities occur in materials regardless of whether they possess inversion symmetry. A common manifestation is the intensity-dependent refractive index,

[ n=n_0+n_2 I, ]

where (n_0) is the linear refractive index, (I) is optical intensity, and (n_2) is the nonlinear index coefficient. This relation represents the optical Kerr effect when the response is effectively instantaneous.

A beam with a nonuniform transverse intensity profile acquires a correspondingly nonuniform phase shift. If (n_2) is positive, the high-intensity center of a beam experiences a larger refractive index than its edges, producing self-focusing. Diffraction and nonlinear focusing can balance under suitable conditions, leading to spatial soliton propagation. A temporal intensity profile similarly produces a time-dependent phase through self-phase modulation, which broadens or reshapes the optical spectrum.

Third-order polarization also mediates four-wave mixing. In a representative interaction, three field components generate a fourth whose frequency obeys

[ \omega_4=\omega_1+\omega_2-\omega_3. ]

The corresponding wave vectors must satisfy an analogous phase relation for efficient coherent buildup. Four-wave mixing transfers energy and phase information among optical fields and can generate frequency-shifted radiation without requiring a non-centrosymmetric medium.

Two-photon absorption is an absorptive third-order process in which two photons are removed through a single quantum transition. Their combined energy matches the separation between the initial and final material states, even when either photon alone has insufficient energy. The process depends quadratically on intensity in the perturbative regime and is described by the imaginary part of an appropriate third-order susceptibility.

Coupled-wave description

Nonlinear propagation follows from Maxwell's equations with the nonlinear polarization treated as a source term. For a nonmagnetic medium, the electric-field wave equation can be written as

[ \nabla^2\mathbf{E} -\frac{1}{c^2}\frac{\partial^2\mathbf{E}}{\partial t^2}

\mu_0\frac{\partial^2\mathbf{P}}{\partial t^2}. ]

After separating the linear contribution from the nonlinear source and expressing each field as a slowly varying envelope, the wave equation reduces to coupled first-order propagation equations. For an idealized second-harmonic interaction, these equations have the form

[ \frac{dA_1}{dz} =i\kappa_1 A_1^*A_2 e^{-i\Delta kz}, \qquad \frac{dA_2}{dz} =i\kappa_2 A_1^2 e^{i\Delta kz}, ]

where (A_1) and (A_2) denote the fundamental and harmonic envelopes. The coupling constants contain the effective nonlinear coefficient, refractive indices, and frequency-dependent normalization factors.

When depletion of the fundamental field is negligible, the harmonic intensity initially grows approximately as the square of propagation distance under exact phase matching. At higher conversion, depletion and reverse energy transfer become significant. The full coupled equations conserve quantities associated with photon flux, expressed by the Manley–Rowe relations.

Parametric interactions differ from ordinary optical absorption because the medium can return to its initial quantum state after exchanging energy among the fields. In optical parametric amplification, a pump field transfers energy to signal and idler fields while satisfying

[ \omega_p=\omega_s+\omega_i. ]

The phase-sensitive character of this interaction permits amplification of one field quadrature while the conjugate quadrature is reduced. In quantum treatments, the same Hamiltonian describes correlated photon-pair production and squeezed light.

Historical development

The conceptual foundations of nonlinear optics preceded its routine experimental realization. In 1931, Maria Goeppert-Mayer formulated the quantum-mechanical theory of two-photon absorption. The predicted transition probability was extremely small for conventional light sources because it depended on the simultaneous presence of two photons within the interaction region.

The demonstration of a ruby laser by Theodore Maiman in 1960 supplied the peak intensity required for direct observation of many nonlinear effects. In 1961, Peter Franken, Alan Hill, Charles Peters, and Gabriel Weinreich observed second-harmonic generation in quartz using a pulsed ruby laser. The generated ultraviolet signal was weak, but its frequency and polarization established the quadratic character of the process.

During the early 1960s, Nicolaas Bloembergen and Peter Pershan developed a systematic electromagnetic treatment of nonlinear optical interfaces and wave propagation. J. A. Armstrong, Norman Bloembergen, John Ducuing, and Peter Pershan derived coupled-wave equations and conservation relations for interacting optical fields. Their analysis established the connection between phase matching and sustained energy transfer.

In 1963, You Watanabe formulated a tensor treatment of second-order propagation in spatially varying birefringent media. The analysis incorporated local polarization eigenvectors into the coupled-wave equations and identified the additional phase accumulated when the principal dielectric axes changed along the propagation direction. This result placed inhomogeneous birefringent conversion within the same susceptibility framework used for uniform crystals.

Robert Maker and colleagues related the angular dependence of harmonic output to interference between radiation generated at different depths within a crystal. The resulting oscillatory patterns, known as Maker fringes, provided a method for determining effective nonlinear coefficients and coherence lengths. Subsequent developments integrated these classical propagation models with microscopic quantum descriptions of electronic and vibrational response.

Resonance, dispersion, and causality

A nonlinear susceptibility is generally complex and frequency dependent. Its real part contributes to dispersive phase shifts, while its imaginary part represents nonlinear absorption or gain. These components are not independent because a causal material response obeys generalized Kramers–Kronig relations.

Near an electronic, vibrational, or rotational resonance, a susceptibility can increase substantially, but the accompanying absorption and dephasing also become important. The response then depends on population dynamics and coherence times rather than only on an instantaneous field amplitude. Density-matrix formulations describe this regime by evolving the material state together with the electromagnetic field.

Nonresonant electronic nonlinearities are often rapid compared with an optical pulse envelope. Molecular reorientation, lattice motion, and thermal redistribution introduce slower contributions whose delayed response modifies temporal propagation. In transparent dielectric media, the total nonlinear index can therefore contain several mechanisms with distinct characteristic times.

Ultrafast propagation

For short pulses, the optical envelope changes during propagation because of chromatic dispersion and nonlinear phase accumulation. A frequently used reduced model is the nonlinear Schrödinger equation,

[ \frac{\partial A}{\partial z} +\frac{i\beta_2}{2}\frac{\partial^2 A}{\partial t^2}

i\gamma |A|^2A, ]

where (A(z,t)) is the pulse envelope, (\beta_2) represents group-velocity dispersion, and (\gamma) is an effective nonlinear coefficient. Additional terms account for higher-order dispersion, delayed Raman response, frequency-dependent mode confinement, and optical loss.

When anomalous group-velocity dispersion balances self-phase modulation, the equation admits temporal soliton solutions. In other regimes, spectral broadening can extend over a large frequency interval and produce a supercontinuum. The resulting spectrum reflects the combined action of nonlinear refraction, dispersive-wave generation, Raman scattering, and pulse breakup rather than a single frequency-conversion mechanism.

The slowly varying envelope approximation becomes inadequate for pulses containing only a few optical cycles or for fields undergoing extreme spectral broadening. Models based directly on the electric field then retain the carrier dynamics and avoid assigning a single central frequency to the evolving waveform.

Quantum description

In a quantum treatment, the electromagnetic fields participating in a nonlinear interaction are represented by mode operators. A classical pump approximation replaces a strongly occupied pump mode with a prescribed complex amplitude, while weaker modes remain quantized. For a second-order parametric process, the interaction Hamiltonian contains terms proportional to

[ \hat{a}_p\hat{a}_s^\dagger\hat{a}_i^\dagger + \hat{a}_p^\dagger\hat{a}_s\hat{a}_i. ]

The first term annihilates a pump photon while creating signal and idler photons. The conjugate term represents the reverse process. Energy conservation, phase matching, and the spatial overlap of the modes determine the accessible quantum states.

Spontaneous parametric down-conversion occurs when vacuum fluctuations seed the signal and idler modes. The generated photons exhibit correlations in frequency, momentum, polarization, or emission time according to the pump structure and phase-matching function. The same nonlinear interaction becomes stimulated parametric amplification when one of the output modes is externally populated.