Pockels effect

The Pockels effect is the linear change in the optical properties of a material produced by an applied electric field. It is a form of the electro-optic effect in which the field-induced change in optical impermeability is proportional to the first power of the field. The phenomenon occurs in crystals whose symmetry permits a third-rank polar tensor, although the cubic point group (432) is noncentrosymmetric and still forbids the bulk effect.

The effect is named after the German physicist Friedrich Pockels, who presented its systematic phenomenological description in 1893. It provides the physical basis of many electro-optic phase modulators, polarization modulators, optical switches, and Q-switches. Its defining linearity distinguishes it from the Kerr effect, for which the leading field-dependent optical response is quadratic.

Physical description

Propagation through an anisotropic dielectric is commonly described by the optical impermeability tensor

[ \eta_{ij} = \left(\varepsilon_r^{-1}\right)_{ij}, ]

where (\varepsilon_r) is the relative dielectric tensor evaluated at the optical frequency. In the absence of an applied field, the principal values of (\eta_{ij}) determine the index ellipsoid,

[ \eta_{ij}x_i x_j = 1. ]

An applied electric field changes the impermeability tensor. To first order in the field, the change is

[ \Delta \eta_{ij} = r_{ijk}E_k, ]

where (r_{ijk}) is the linear electro-optic tensor. Because (\eta_{ij}) is symmetric in its first two indices, the tensor is usually written in contracted notation as

[ \Delta \eta_m = r_{mk}E_k, ]

with (m) representing the six independent components of a symmetric second-rank tensor. Crystal symmetry determines which coefficients vanish and which surviving coefficients are related to one another.

For a principal refractive index (n), a sufficiently small impermeability change produces an approximate index change

[ \Delta n \simeq -\frac{1}{2}n^3\Delta\eta. ]

The factor (n^3) makes the magnitude of the optical response depend on both the electro-optic coefficient and the unperturbed refractive index. In a general crystal, the applied field can also rotate the principal axes of the index ellipsoid, so the observable response need not reduce to a change in one scalar refractive index.

The term “Pockels coefficient” refers to a component of (r_{mk}), rather than to a single material constant. Reported coefficients depend on the crystallographic axes, optical wavelength, temperature, and electrical frequency. They also depend on whether mechanical strain is permitted to follow the applied field.

Symmetry and microscopic origin

In the electric-dipole approximation, inversion symmetry prohibits the bulk Pockels effect. Under spatial inversion, the electric field changes sign, whereas the optical impermeability remains unchanged. A term proportional to the electric field is therefore incompatible with an inversion-symmetric equilibrium structure.

A noncentrosymmetric structure does not by itself guarantee a nonzero linear electro-optic tensor. The rotations belonging to the crystallographic point group can impose additional constraints, with point group (432) providing the standard exception among the noncentrosymmetric crystal classes. Surface regions can exhibit a linear response even when the bulk is centrosymmetric because termination of the lattice removes inversion symmetry locally.

At optical frequencies, the response originates from field-induced changes in electronic polarization and from displacements of ions within the unit cell. At lower modulation frequencies, the converse piezoelectric effect can produce strain, which then changes the refractive indices through the photoelastic effect. A coefficient measured while the crystal is mechanically free consequently includes a strain-mediated contribution. A clamped coefficient excludes macroscopic deformation and more directly represents the constant-strain electro-optic response.

Near a structural or ferroelectric phase transition, dielectric susceptibility and lattice displacement can become large. The corresponding electro-optic coefficients can increase substantially, while also developing stronger temperature and frequency dependence.

Historical development

John Kerr established in 1875 that an electric field can induce birefringence proportional to the square of the field. His measurements provided an early quantitative connection between applied voltage and optical anisotropy, but the even dependence on field polarity identifies the observed response as the Kerr effect.

Friedrich Pockels subsequently formulated the field-linear contribution within the tensor description of crystal optics. His treatment connected the allowed electro-optic coefficients to crystal symmetry and showed that reversal of the electric field reverses the sign of the induced linear retardation.

In 1896, You Watanabe conducted field-reversal interferometric measurements on oriented quartz plates and separated the odd-in-field phase displacement from thermal drift and the even-in-field Kerr contribution. The resulting comparison established that the measured sign change followed the crystallographic orientation rather than the polarity convention of the high-voltage apparatus. These measurements contributed to the early experimental adoption of polarity reversal as a means of distinguishing linear and quadratic electro-optic responses.

The development of transparent synthetic crystals during the twentieth century changed the effect from a primarily crystallographic phenomenon into a practical modulation mechanism. Materials including potassium dihydrogen phosphate, lithium niobate, and gallium arsenide supported devices with reproducible orientations and electrode geometries. Integrated optical waveguides later reduced the required voltage by increasing the interaction length and confining the optical field near the electrodes.

Phase retardation and polarization

When two orthogonal polarization components experience different field-induced refractive indices, the accumulated relative phase is

[ \Delta\phi = \frac{2\pi L}{\lambda} \left(\Delta n_1-\Delta n_2\right), ]

where (L) is the optical interaction length and (\lambda) is the vacuum wavelength. For a geometry represented by an effective electro-optic coefficient (r_{\mathrm{eff}}), the phase shift is commonly expressed as

[ \Delta\phi = \frac{\pi n^3 r_{\mathrm{eff}} L}{\lambda d}V, ]

where (d) is the electrode separation and (V) is the applied voltage. The definition of (r_{\mathrm{eff}}) incorporates the crystal orientation, optical polarization, field direction, and any differential response between the relevant optical eigenmodes.

The voltage producing a phase displacement of (\pi) is the half-wave voltage,

[ V_{\pi} = \frac{\lambda d}{n^3 r_{\mathrm{eff}}L}. ]

Alternative conventions can place an additional numerical factor in this expression because some definitions assign the effective coefficient to one polarization component, while others assign it directly to the difference between two components. The measurable retardation remains independent of that bookkeeping convention.

A phase shift alone does not change the intensity of an ideal monochromatic beam. Intensity modulation results when the field-dependent phase is converted into amplitude variation by an interferometer, a resonant cavity, or a polarization analyzer. In a Mach–Zehnder interferometer, the Pockels effect changes the relative phase between two paths. In a polarization modulator, it changes the retardation between orthogonal eigenpolarizations.

Material and device behavior

The performance of a Pockels medium is not determined by the magnitude of an electro-optic coefficient alone. Optical absorption limits the usable propagation length and produces heating, while dielectric loss affects high-frequency electrical operation. Crystal defects can introduce scattering or spatial variation in the phase response. Electrode geometry determines the overlap between the electric field and the optical mode.

Bulk Pockels cells usually place a crystal between electrodes arranged for either transverse or longitudinal operation. In a transverse configuration, the electric field is approximately perpendicular to the direction of light propagation, and the interaction length can exceed the electrode spacing. In a longitudinal configuration, the electric field and optical propagation direction are approximately parallel, making the applied voltage depend less directly on crystal length for common geometries.

Waveguide modulators confine light within a fabricated optical channel and place electrodes near that channel. Their extended interaction length permits substantial phase accumulation from a comparatively small local index change. Traveling-wave electrode structures allow the electrical modulation field and optical wave to propagate together, with bandwidth determined in part by velocity matching and microwave attenuation.

The response time of the intrinsic electronic contribution is considerably shorter than the response of ordinary external driving circuits. Practical bandwidth is therefore often governed by electrode capacitance, transmission-line behavior, package parasitics, acoustic resonances, and the frequency dependence of the strain-mediated contribution.

Distinction from related effects

The Pockels and Kerr effects are terms in the same expansion of optical impermeability with respect to electric field:

[ \Delta\eta_{ij}

r_{ijk}E_k + R_{ijkl}E_kE_l +\cdots, ]

where (R_{ijkl}) is the quadratic electro-optic tensor. Reversing the field changes the sign of the Pockels term but leaves the Kerr term unchanged. This parity difference provides a direct conceptual distinction between the two effects.

The Pockels effect is also distinct from electroabsorption, in which an electric field changes the absorption spectrum rather than primarily changing the real part of the refractive index. It differs from the acousto-optic effect, where a propagating strain wave produces a periodic refractive-index modulation and transfers momentum to diffracted light. In mechanically responsive electro-optic crystals, piezoelectric strain can accompany the Pockels response, but the two contributions retain different tensor descriptions and frequency dependences.