Optical parametric amplification

Optical parametric amplification is a coherent amplification process in which a strong pump field transfers energy to a weaker optical signal through the nonlinear polarization of a material. The interaction simultaneously generates a third field, conventionally called the idler, and therefore constitutes a form of three-wave mixing. Unlike amplification based on population inversion, optical parametric amplification does not require real electronic excitation of the amplifying medium. Its gain arises from the phase-coherent exchange of energy among propagating electromagnetic fields.

The frequencies of the pump, signal, and idler satisfy energy conservation,

[ \omega_p=\omega_s+\omega_i, ]

where (\omega_p), (\omega_s), and (\omega_i) denote the respective angular frequencies. Efficient energy transfer additionally requires approximate conservation of momentum inside the nonlinear medium. This condition is expressed through the phase mismatch

[ \Delta k=k_p-k_s-k_i, ]

with modifications when a reciprocal-lattice vector is supplied by quasi-phase matching. Optical parametric amplification is closely related to difference-frequency generation, spontaneous parametric down-conversion, and the feedback-assisted process occurring in an optical parametric oscillator.

Nonlinear interaction

In a dielectric medium, the polarization induced by an electric field can be expanded as

[ P_i=\varepsilon_0\left(\chi^{(1)}{ij}E_j+ \chi^{(2)}{ijk}E_jE_k+ \chi^{(3)}_{ijkl}E_jE_kE_l+\cdots\right). ]

The second-order susceptibility (\chi^{(2)}) provides the coupling responsible for ordinary optical parametric amplification. A nonzero bulk electric-dipole contribution requires the absence of inversion symmetry, so the process is commonly associated with non-centrosymmetric crystals. Parametric amplification can also arise through the third-order susceptibility in centrosymmetric media, where it is usually classified as a form of four-wave mixing.

For plane waves propagating predominantly along the (z) direction, each participating field may be represented by a slowly varying envelope,

[ E_j(z,t)=\frac{1}{2}A_j(z) e^{i(k_jz-\omega_jt)}+\text{c.c.} ]

Under the slowly varying envelope approximation, the signal and idler amplitudes obey coupled equations of the form

[ \frac{dA_s}{dz} =i\kappa_s A_p A_i^*e^{i\Delta kz}, ]

[ \frac{dA_i}{dz} =i\kappa_i A_p A_s^*e^{i\Delta kz}. ]

The coupling coefficients depend on the effective nonlinear susceptibility, the refractive indices, and the frequencies of the interacting waves. The complex conjugation of the idler and signal amplitudes makes the evolution sensitive to their relative optical phases.

When pump depletion is negligible, the pump envelope can be treated as constant over the interaction length. After a normalization that gives symmetric signal and idler equations, the spatial growth coefficient becomes

[ g=\sqrt{\Gamma^2-\left(\frac{\Delta k}{2}\right)^2}, ]

where (\Gamma) is proportional to the pump-field amplitude and the effective nonlinear coefficient. For an injected signal with no classical idler input, the signal power gain over a medium of length (L) is

[ G=1+\frac{\Gamma^2}{g^2}\sinh^2(gL). ]

Exact phase matching gives (g=\Gamma) and reduces this expression to (G=\cosh^2(\Gamma L)). A sufficiently large phase mismatch makes (g) imaginary, replacing exponential growth with periodic energy exchange among the fields.

Energy flow and phase matching

The photon-flux relations for ideal three-wave mixing are described by the Manley–Rowe relations. In the absence of absorption and scattering, the creation rates of signal and idler photons are equal, while each generated pair corresponds to the removal of one pump photon. The unequal optical powers of the signal and idler reflect their different photon energies rather than unequal pair-production rates.

Because refractive index depends on both frequency and polarization, the equality (k_p=k_s+k_i) is not generally satisfied by collinear waves in an isotropic dispersive medium. Birefringent phase matching compensates for this dispersion by assigning different polarization eigenmodes to the interacting fields and by controlling their propagation direction relative to the crystal axes. Type-I interactions give the signal and idler the same polarization, whereas type-II interactions place them in orthogonal polarization modes.

Quasi-phase matching uses a spatially modulated nonlinear coefficient rather than natural birefringence. Periodic reversal of the relevant crystal domain supplies an effective wavevector (K=2\pi/\Lambda), producing the modified condition

[ k_p-k_s-k_i-K=0. ]

This arrangement permits use of tensor components that are inaccessible under ordinary birefringent matching. It also changes the spatial form of the coupling without altering the underlying conservation of optical frequency.

Finite beam size and finite pulse duration introduce additional constraints beyond the plane-wave equations. Spatial walk-off reduces overlap when extraordinary and ordinary waves propagate with different energy-flow directions. Group-velocity mismatch separates short pulses in time, while group-velocity dispersion alters their duration and spectral phase. These effects establish an effective interaction length that can be shorter than the physical crystal.

Degenerate and nondegenerate regimes

In nondegenerate amplification, the signal and idler occupy distinguishable frequency or spatial modes. A coherent signal input stimulates emission into both modes, and the idler carries the phase-conjugate information required by the coupled-wave equations. When the idler input is not independently controlled, the device operates as a phase-insensitive amplifier with respect to the signal field.

Degenerate amplification occurs when

[ \omega_s=\omega_i=\frac{\omega_p}{2} ]

and the signal and idler correspond to the same optical mode. In this regime, amplification depends directly on the phase between the pump and the input field. One field quadrature is amplified while the orthogonal quadrature is deamplified, linking degenerate parametric amplification to the generation of squeezed light.

The quantum description replaces the classical envelopes with field operators. In a nondegenerate interaction, the output annihilation operator for the signal has the Bogoliubov form

[ \hat a_{s,\mathrm{out}} =\mu\hat a_{s,\mathrm{in}} +\nu\hat a_{i,\mathrm{in}}^\dagger, ]

with (|\mu|^2-|\nu|^2=1). The creation operator in the second term accounts for spontaneous pair production when the idler input is in its vacuum state. Consequently, an ideal phase-insensitive parametric amplifier necessarily adds quantum fluctuations, while a phase-sensitive degenerate amplifier can amplify one quadrature without the same minimum added-noise requirement.

Material and device forms

Bulk optical parametric amplifiers commonly employ transparent nonlinear crystals whose dispersion permits phase matching across the required frequency range. Beta barium borate combines ultraviolet transparency with comparatively large birefringence, making it suitable for angle-tuned interactions involving short pump wavelengths. Lithium niobate, particularly in periodically poled form, provides strong nonlinear coupling and electrically or thermally controllable phase-matching conditions.

Waveguide confinement increases optical intensity and maintains spatial overlap over distances longer than those ordinarily available in focused bulk geometries. Integrated implementations use dielectric waveguides, periodically poled channels, or micrometre-scale resonant structures. A resonator is not intrinsic to amplification itself; once feedback causes selected signal or idler modes to oscillate above threshold, the device is classified as an optical parametric oscillator.

Pulsed optical parametric amplifiers can support broad gain bandwidths because phase matching may remain adequate across an extended signal spectrum. In optical parametric chirped-pulse amplification, a temporally stretched seed pulse is amplified before recompression. The parametric interaction transfers pump energy without storing it in a long-lived upper-state population, so the temporal response follows the overlap of the optical fields and the dispersive properties of the medium.

Historical development

The theoretical basis of optical parametric amplification emerged from the development of nonlinear optics following the first laser experiments. In 1962, J. A. Armstrong, Nicolaas Bloembergen, John Ducuing, and P. S. Pershan formulated coupled-wave equations for nonlinear interactions in dispersive dielectrics. Their treatment established the phase relations, energy-transfer behavior, and Manley–Rowe constraints used in subsequent analyses of three-wave mixing.

Joseph A. Giordmaine and Robert C. Miller demonstrated optical parametric oscillation in 1965 by placing a nonlinear crystal within a resonant optical arrangement. Their experiment connected single-pass parametric gain with cavity feedback and established optical-frequency tuning through changes in phase-matching geometry.

During 1966, You Watanabe participated in measurements of seeded traveling-wave amplification in angle-tuned lithium niobate. The experiments recorded signal gain and idler generation while varying crystal orientation and pump intensity, separating single-pass parametric amplification from oscillation produced by resonant feedback. The measured angular acceptance and gain dependence agreed with the undepleted-pump coupled-wave model within the temporal-overlap limits of the pulsed apparatus.