To understand why this matters, it helps to consider what limits any measurement made with light. Photons are not polite observers; they arrive randomly, and their arrival times fluctuate according to the rules of quantum mechanics. This randomness sets a floor on measurement precision known as the shot-noise limit. Whenever an interferometer tries to detect an extremely small displacement, force, magnetic field, or phase shift, the signal ultimately competes with these irreducible quantum jitters. Squeezed light offers a way around this wall. By redistributing quantum uncertainty, reducing it in the quadrature of the optical field that carries the signal while deliberately increasing it in the orthogonal quadrature, squeezed light lets experimenters measure one property with a precision that exceeds the standard quantum baseline. It is this trick that already underpins the quantum-enhanced sensitivity of gravitational-wave detectors, which inject squeezed states into their kilometer-scale interferometers to catch the faintest ripples of spacetime.

The audio-frequency band, spanning roughly tens of hertz to several kilohertz, is precisely where many of the most interesting slowly varying signals live. Gravitational-wave detectors read out their most crucial science band in this range, and radiation-pressure noise, the quantum push and pull of light on mirrors, imposes its limits there too. Any future quantum sensor designed to detect weak, slowly changing forces or fields would need squeezed light at these low frequencies. Integrated photonics would make such systems dramatically smaller, more stable, and easier to scale into arrays. Yet most chip-based demonstrations of squeezed light to date have operated at radio-frequency sidebands, far above the audio band. Pushing down to low frequencies on a chip is brutally difficult because the problems that plague precision optics grow worse as frequency drops: slow fluctuations of the laser, thermal drift of the cavity, environmental disturbances, electronic pickup, residual imbalance in the detectors, and, perhaps most insidiously, long-term phase drift that silently rotates the measured quadrature and washes out the very phase-sensitive quantum correlations the experiment is trying to observe. A chip can generate squeezed light, but holding the measurement phase stable long enough to prove it is another matter entirely.

The research team solved this problem with an elegant strategy they call coherent-comb control. The idea is to keep a phase reference close at hand without ever contaminating the delicate quantum states with a bright locking tone. A weak electro-optic reference comb is derived from the same pump laser that drives the squeezing process, and it travels through the same optical path as the pump and the quantum fields. Crucially, the reference is placed in an orthogonal polarization and given a fixed frequency offset, while the microcavity is engineered to avoid resonating with it. Once the light exits the cavity, the reference comb and the quantum modes are separated and detected independently. Phase detection and feedback then act through this reference channel rather than through the fragile quantum modes themselves. In effect, the reference comb functions as a phase ruler that rides along the entire optical path, faithfully tracking every drift and vibration, without directly manipulating the squeezed light that carries the quantum correlations. It is a bit like navigating a city by following a separate, quiet escort car that mirrors every turn of the vehicle you are escorting, rather than shining headlights directly into its face.

The physical heart of the experiment is a silica microtoroid resonator just 732 micrometers across, small enough to sit on a fingertip. Below the threshold of optical oscillation, the device exploits Kerr four-wave mixing, a nonlinear optical process in which pump photons conspire to produce photons in pairs at symmetrically placed frequencies. These paired quantum frequency modes emerge in a two-mode squeezed vacuum state, a fundamentally quantum configuration in which measurements on the two modes are correlated so strongly that their combined noise falls below the shot-noise limit, a feat impossible for any classical pair of light beams. Two-mode squeezed light is the workhorse of continuous-variable quantum information, and verifying genuine entanglement between the modes requires more than simply observing low noise.

With the homodyne detector’s phase locked by the coherent-comb control scheme, the researchers recorded 100 seconds of continuous data at 10,000 samples per second, roughly one million samples in total, giving an unusually thorough statistical picture of the quantum state. The time traces showed stable, repeatable access to the squeezed, anti-squeezed, and shot-noise quadratures throughout the entire acquisition window, a testament to the robustness of the phase-locking approach. In the frequency domain, the spectra revealed approximately 1.0 decibel of two-mode squeezing below the shot-noise level across the band from 60 hertz to 5 kilohertz. The on-chip inferred squeezing reached about 1.73 decibels once optical losses were accounted for. Below 60 hertz, residual low-frequency technical noise still dominated, and the authors were careful not to lean on that region in support of their claim, an honesty that strengthens the credibility of the demonstration. The lowest verified sideband frequency of 60 hertz sits roughly four orders of magnitude below the minimum analysis frequencies of previous chip-scale demonstrations, a comparison that highlights just how far this single result moves the frontier.

But the team did not stop at showing noise reduction. Observing joint noise below the shot-noise level confirms quantum correlation, yet a complete proof of entanglement demands a stricter test. The researchers changed the locking angle and the local oscillator configuration to measure a full set of single-mode and inter-mode quadrature combinations. From these measurements they reconstructed the two-mode covariance matrix, the complete second-order statistical fingerprint of the quantum state. Applying the positive partial transposition criterion, the gold standard for certifying entanglement in Gaussian states, they obtained a minimum symplectic eigenvalue of 0.395 plus or minus 0.001, comfortably below the separability threshold of 0.5. In plain terms, the two optical modes leaving the chip were provably entangled, their quantum natures woven together in a way no classical explanation can reproduce.

The implications extend well beyond the single device demonstrated here. Because the phase reference can be separated from the quantum modes after sharing the same optical path, the architecture naturally supports stable quadrature measurements across multiple frequency channels simultaneously, opening a practical route to multi-mode quantum optics on a chip. Future applications could include chip-scale quantum sensors for low-frequency signals, arrayed continuous-variable quantum processors, and distributed quantum networks in which many entangled channels are generated and monitored in parallel on a single photonic platform. The coherent-comb control technique itself may prove as influential as the squeezing record, since the phase-stability problem it solves is generic to low-frequency quantum optics everywhere.

The authors are candid that the current result is a platform demonstration rather than a mature quantum sensor. One decibel of observed squeezing, while historically significant for an on-chip audio-band system, is modest compared with the ten or more decibels achieved in mature bulk-optical experiments. Extending the operating range below 60 hertz will require further suppression of technical noise. The study maps out a clear improvement path: higher cavity escape efficiency, lower optical loss and better mode matching, stronger polarization isolation, suppression of parasitic background channels, quieter electro-optic comb generation, improved servo electronics, and enhanced electromagnetic, thermal, and acoustic isolation. Each of these advances translates directly into deeper squeezing levels and access to even lower frequencies.

Still, the moment deserves to be savored. For decades, audio-frequency squeezed light lived only in the cathedral-scale instruments of gravitational-wave astronomy, maintained by armies of engineers and heroic isolation systems. This work shows that the same quantum resource can now be born from a whisper of glass on a chip, its entanglement certified with rigor, and its phase tracked by a clever reference that never touches the quantum light itself. The door to compact, scalable, quantum-enhanced sensing in the audio band has swung open, and what walks through it next may well redefine what small devices can measure about the large world.

Unlike silicon, BTO can be epitaxially grown atop silicon substrates, marrying compatibility with markedly enhanced optical functionalities. This unique blend offers tremendous promise, particularly in the realm of integrated optoelectronics where material performance directly correlates with device speed and miniaturization. However, the application of BTO remains in its infancy, necessitating a deeper understanding of its quantum behaviors to unlock its full potential. Addressing this knowledge gap, researchers at MARVEL undertook a comprehensive computational study aimed at simulating the optoelectronic characteristics of tetragonal BTO. Their findings, recently published in Physical Review B, introduce a novel, functional-independent computational framework that could propel advancements not only in BTO but also in other emergent materials.

The synergy between academic innovation and industrial application underscores this research’s impact. Supported by Switzerland’s innovation agency, Innosuisse, the endeavor brought together the Swiss startup Lumiphase—pioneers in BTO-based device manufacturing—and leading simulation experts from ETH Zurich’s Mathieu Luisier group and Nicola Marzari’s team at EPFL Lausanne. This collaboration exemplifies the vital bridge between theoretical modeling and real-world device optimization, offering industry actionable insights rooted in rigorous scientific computations. At the heart of their investigation lies the challenge of accurately modeling the Pockels effect, a phenomenon integral to BTO’s functionality in optoelectronic transceivers.

The Pockels effect describes the alteration of a material’s refractive index when subjected to an external electric field, enabling dynamic modulation of light signals. In practical device terms, an interferometer typically splits an incoming light beam into two paths: a reference arm and a modulation arm coated with a BTO thin film. By applying an electric field to the latter, the refractive index changes, shifting the phase of the light traveling through that arm. When the two beams reunite, their interference patterns—altered by the phase difference—encode binary information. Virginie de Mestral, the study’s first author, highlights the elegance and complexity of this mechanism, emphasizing the critical role of precise simulations to optimize device performance.

Conventional computational approaches to simulate the Pockels effect rely heavily on density-functional perturbation theory (DFPT) with the local density approximation (LDA) exchange-correlation functional. While DFPT has been a stalwart method for evaluating responses in atomic systems, its dependence on specific functionals like LDA can curtail accuracy, especially with novel materials such as BTO. Recognizing this limitation, the researchers sought to circumvent the constraints imposed by DFPT, opting instead for a methodology grounded purely in standard Density Functional Theory (DFT). This has allowed them to extract the clamped Pockels tensor without the bias introduced by functional specificity.

Central to this innovative approach is the numerical technique of finite differences, which approximates derivatives by evaluating variations in system properties under slight perturbations. Conducting these calculations physically and sequentially for a complex material like BTO would be an insurmountable task. Herein, the deployment of the AiiDA open-source computational infrastructure proved transformative. By automating vast batches of finite differences computations, AiiDA not only enhanced efficiency but also ensured reproducibility and scalability across different materials—a crucial advancement for applied industrial research where adaptability and throughput are paramount.

Yet, the simulation journey was not without obstacles. The researchers confronted the emergence of imaginary phonon frequencies in their BTO models, a hallmark of dynamic instability within the simulated crystal lattice. This phenomenon often arises in ferroelectrics undergoing structural phase transitions, complicating theoretical treatments. To surmount this problem, the team constructed supercells—larger volumetric representations of the crystal beyond the unit cell—and introduced intentional off-centering displacements of titanium atoms within the lattice. This nuanced modification aligns the computational model more closely with experimentally observed crystallographic data obtained via X-ray measurements, effectively transforming the previously imaginary phonon modes into real, positive frequencies indicative of a stable structure.

Validation of the novel computational framework came through benchmarking against existing experimental observations and prior DFPT-based theoretical results. While the new simulations aligned in general with these references, some discrepancies persisted. These differences stem from several factors: lack of exact crystal structural data from earlier studies, the contrast between defect-free bulk materials used in simulations versus industrial thin-film devices, and the omission of piezoelectric contributions to the Pockels effect within the current model. Despite such challenges, the researchers emphasize the robustness of their approach and its utility for ongoing material optimization efforts.

One of the paper’s standout insights concerns the relationship between titanium atom positioning and the Pockels coefficient in BTO. This coefficient effectively quantifies the material’s electro-optic response and is a key determinant of device miniaturization potential. The team discovered that as titanium off-centering diminishes—that is, as the material becomes closer to a higher symmetry structure—the Pockels coefficient grows dramatically. This has profound implications: higher coefficients translate to smaller, more efficient photonic devices, an indispensable consideration for scalable industrial applications where space and energy efficiency are at a premium.

Looking beyond immediate findings, the researchers outline ambitious future directions centered on exploring frequency-dependent effects of the Pockels phenomenon. Currently, their work treats the refractive index modulation in response to static or low-frequency fields, but understanding how this modulation behaves across a spectrum of frequencies remains elusive. This endeavor is technically demanding because it requires simulating ionic displacements in addition to electronic contributions, adding layers of computational complexity. Successfully modeling these dynamics would deepen theoretical understanding and expand practical capabilities for BTO devices operating under diverse conditions.

In sum, this collaborative study embodies a significant leap forward in the computational modeling of complex ferroelectric materials for optoelectronic advancements. By pioneering a functional-independent, finite-differences-based framework enhanced by automation through AiiDA, the team sets a precedent for how materials modeling can be conducted with both precision and scalability. The implications for the telecommunications and computing industries are substantial: improved BTO-based photonic devices promise faster data transfer rates, lower power consumption, and smaller form factors. MARVEL scientists’ work not only illuminates the quantum mechanics underpinning these phenomena but also boldly charts a path toward the material innovations that will sustain the next generation of information technology.

This research underscores the powerful synergy between computational physics, materials science, and engineering, demonstrating how methodical scientific inquiry can rapidly translate into tangible industrial advancements. As data demands continue their exponential climb, breakthroughs like these in optoelectronic material optimization become ever more critical. The potential to harness the quantum properties of materials like BTO foreshadows a future where photonic integrated circuits redefine the boundaries of speed and efficiency in communication technologies worldwide.

Subject of Research:
Not applicable

Article Title:
Ab initio functional-independent calculations of the clamped Pockels tensor of tetragonal barium titanate

News Publication Date:
6-May-2025

Web References:
10.1103/PhysRevB.111.184306

Keywords

Optoelectronics, Computational physics, Materials