Lithium niobate: a crystal that makes relations physical

The useful claim

Lithium niobate does not reveal a new fundamental force. It does something more immediately valuable: it makes several normally separated relations controllable in one physical object. The same non-centrosymmetric ferroelectric crystal can couple optical fields to optical fields, microwave fields to optical phase, optical excitation to lattice motion and temperature history to a persistent device response.

That makes lithium niobate an unusually complete workbench for the corpus's ordinary-language grammar:

The carrier is the thing that can vary. The selector is the intervention that chooses which relation is effective. The detector is the surviving record that distinguishes the alternatives.

The important point is not that the crystal has three ingredients. It is that the experimenter can change a domain orientation, pump phase, modulation frequency, boundary or poling period while keeping much of the carrier fixed, and then ask which phase-sensitive output changes. The material turns a relation into an engineering variable.

One material, several real channels

At room temperature lithium niobate is a polar ferroelectric. Its broken inversion symmetry permits a second-order optical susceptibility, a linear electro-optic response and piezoelectric coupling. These are not three names for one effect. They are different tensor contractions supported by the same ordered crystal.

Its optical polarization may 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 carrier is therefore not merely “light in a crystal.” Depending on the experiment it includes optical modes, microwave modes, phonons, the ferroelectric order parameter and their allowed tensor couplings. A claim about one channel does not automatically transfer to the others.

First-principles work traces the ferroelectric instability to a zone-centre polar mode and the balance between long-range dipolar and short-range forces. This gives the domain orientation a material origin: it is a stable lattice configuration, not an abstract sign inserted into an equation.

The product really does create a sum and a difference

Suppose two real optical fields enter the quadratic polarization:

\[ E_1(t)=A_1\cos(\omega_1t+\phi_1),\qquad E_2(t)=A_2\cos(\omega_2t+\phi_2). \]

Their product is

\[ E_1E_2=\frac{A_1A_2}{2} \left[ \cos((\omega_1-\omega_2)t+\phi_1-\phi_2) +\cos((\omega_1+\omega_2)t+\phi_1+\phi_2) \right]. \]

The quadratic material polarization can therefore source fields at the sum and difference frequencies. Setting the two inputs equal gives a static term and the second harmonic. With a strong pump, the reverse process supports parametric amplification, squeezing and correlated photon-pair production.

This resolves an important ambiguity in the wider corpus:

in thermodynamics, \(\log Z\) changes a product of partition functions into an additive free-energy quantity;
in the Xi construction, \(\partial\log\) changes a conjugate analytic product into an additive response;
in lithium niobate, a bilinear field product creates sum- and difference-frequency components.

These operations rhyme, but they are not the same map.

Phase matching is an admissibility condition

Energy conservation is not sufficient for efficient conversion. The generated contributions must also accumulate coherently along the device. For sum-frequency generation the mismatch is conventionally written

\[ \Delta k=k_3-k_2-k_1, \qquad \omega_3=\omega_2+\omega_1. \]

If \(\Delta k\neq0\), contributions generated at different positions rotate out of phase and cancel. Periodically poled lithium niobate deliberately reverses the ferroelectric domains, and hence the sign of the effective quadratic coefficient. A reciprocal vector supplied by the poling pattern changes the condition to

\[ \Delta k-\frac{2\pi m}{\Lambda}\simeq0, \]

for an appropriate integer \(m\) and domain period \(\Lambda\).

This is a particularly literal form of admissibility. The frequencies may be energetically allowed and the local coupling may be nonzero, yet the global geometry can forbid a large outgoing record through destructive accumulation. The domain pattern does not add energy. It selects which locally generated amplitudes are allowed to close coherently over the device.

Corpus role	Lithium-niobate realization	What must remain distinct
carrier	guided optical, microwave or acoustic modes and the polar lattice	a mode is not its measured intensity
selector	crystal cut, polarization, pump phase, electrode drive, domain sign and poling period	an imposed selector is not spontaneous selection
dynamics	electro-optic modulation, three-wave mixing, phonon coupling and propagation	shared material does not make the mechanisms identical
detector	output spectrum, heterodyne quadrature, conversion efficiency, coincidence or squeezing record	a phase-blind power measurement erases part of the relation
memory	domains, trapped charge, photorefraction and pyroelectric history	drift is not automatically fundamental memory

The “dual face” is physically visible

Reversing a ferroelectric domain reverses the sign of selected odd-rank response coefficients. The two domains are therefore not merely two labels for the same laboratory configuration. They present opposite coupling faces to the same field geometry.

But a sign flip is not yet a detector. If the measurement retains only output power, some phase reversals can remain invisible or become confused with loss. A reference phase, interferometric quadrature, closed conversion loop or correlation measurement is needed to retain orientation. This is the same discipline imposed by the corpus's three-resonator result: two local couplings can often be rephased, while a closed loop can carry a gauge-resistant phase.

In a three-mode description the interaction has the schematic form

\[ H_{\mathrm{int}}=\hbar g\, a_3^\dagger a_2a_1+ \hbar g^*a_3a_2^\dagger a_1^\dagger. \]

The complex coupling \(g\) contains tensor overlap, domain sign and phase matching. Local phase conventions can move phase among the three modal amplitudes and \(g\); an experimentally closed sequence or external reference is what makes a residual phase observable. The triangle is useful because it forces the claimed relation to survive re-labelling.

The Pockels channel is a controlled selector

In the linear electro-optic effect, an applied low-frequency electric field changes the optical impermeability tensor:

\[ \Delta(n^{-2})_{ij}=r_{ijk}E_k^{\mathrm{drive}}. \]

A static drive changes optical phase or resonance frequency. A time-dependent drive creates sidebands and couples frequency modes. Integrated lithium- niobate resonators have used this to build synthetic frequency lattices in which resonator modes behave like lattice sites and the microwave modulation sets the link amplitudes and phases.

This is close to the proposed relation-emulator hardware, but with a strict boundary. A programmable frequency lattice demonstrates that the grammar can be engineered. It does not show that the programmed graph is nature's microscopic substrate or that its symmetry is fundamental.

The lattice is not passive background

Lithium niobate also supports polar phonons, acoustic waves, piezoelectric conversion and phonon-polaritons. Experiments have directly mapped strain waves generated when a photoinduced shift current changes polarization and the piezoelectric coupling drives the lattice. Here an optical intervention becomes a mechanical record.

Temperature enters in a similarly concrete but non-gravitational way. It can change refractive indices, phase mismatch, resonance positions, domain kinetics and charge transport. Pyroelectric fields and photorefractive charge redistribution can leave slow, history-dependent shifts in thin-film devices. The device therefore has a genuine configuration state beyond the instantaneous optical amplitudes.

That state is a useful model of memory, hysteresis and metastability. It is not evidence that thermodynamics is a second time direction, that configuration space generates gravity or that lithium niobate supplies an aether. Each of those stronger claims would require an additional invariant prediction and a measurement that excludes conventional material response.

A practical experiment for the corpus

The best next use is a same-device relation-closure experiment rather than an attempt to reproduce every known nonlinear-optical effect.

Fabricate or select a thin-film lithium-niobate ring with three resolved optical or synthetic-frequency modes.
Independently calibrate every pairwise electro-optic or quadratic coupling, including loss, detuning, dispersion and detector response.
Program two configurations with the same coupling magnitudes and opposite closed-loop phase \(\Phi=\phi_{12}+\phi_{23}+\phi_{31}\).
Retain a coherent reference and measure directional transfer, output quadratures and the full pole spectrum, not intensity alone.
Reverse domain orientation, pump phase and modulation order separately so that material parity, drive convention and detector labelling cannot mimic one another.
Predict a held-out closure record from the pairwise calibrations with one shared nuisance model.

The conventional payoff would already be substantial: a phase-stable frequency converter, synthetic-frequency circuit or quantum transducer whose orientation record is robust against gauge choice and drift. A new-physics claim would require a reproducible residual that survives the complete electro-optic, nonlinear, thermal, photorefractive, acoustic and detector model, then repeats on an independent carrier.

Regulus-frame extension

The completed regulus programme supplies a larger but still finite target for frequency-bin hardware. Forty explicitly listed two-qutrit commuting classes form 36 complete ten-basis MUB frames. Their frame-overlap algebra has only two relations: pairs share either four bases or one. A frequency processor with nine coherent modes can therefore be tested with a separate calibration checklist:

Calibration gate	Required record
commuting classes	compile and calibrate the 40 Weyl measurements
control costs	assign measured costs to the six logical Clifford generators and pre-register the cheapest predicted frame or tied set
MUB geometry	verify the ten-basis relations frame by frame
leakage	apply frame-preserving Clifford control noise and quantify probability outside the 36 frames
invariant channel	fit the conditional kernel to \(\operatorname{span}\{I,A_4,A_1\}\)
held-out prediction	predict one-overlap and four-overlap transitions

The distinction between leakage and relation residual is essential. Generic loss, dispersion or non-Clifford mode mixing need not map one full frame to another; those effects must not be disguised as a failed exceptional transition law. See the operator/noise completion and the experimental resolution.

What the material changes

Lithium niobate does not change the equations of known physics. It changes which combinations of known physics can be prepared, reversed and read out in one place.

That matters for this corpus. “Carrier,” “selector” and “detector” no longer need to remain metaphors attached to remote mathematical objects. In lithium niobate they can be assigned to independently calibrated pieces of an actual device. The crystal supplies a bridge from relation language to laboratory grammar:

\[ \boxed{ \text{ordered material} +\text{sourced selector} +\text{coherent closure} \longrightarrow \text{phase-bearing record}. } \]

The scientific value lies precisely in the restraint. If ordinary nonlinear and condensed-matter physics closes the record, the result is useful technology and a validated experimental language. If it does not, the same device provides the reversals, nuisance channels and independent records needed to say exactly what failed.