The night sky hides more than just stars. It hides a chorus of worlds orbiting those stars, some so close that their light can barely escape the star’s glare. For decades, astronomers have tried to tune that glare down without erasing the faint whispers of distant planets. Now a collaboration led by Yinzi Xin at the California Institute of Technology—with partners at UCLA, JPL, the University of Sydney, the University of Central Florida, the University of Illinois Urbana-Champaign, and UC Irvine—has taken a bold step in that direction. The Photonic Lantern Nuller, or PLN, is not just a clever optical gadget; it’s a new way to shape light so we can listen for planets that sit within the star’s own bright halo. The first demonstrations showed four nulled channels carved into a single beam, but Xin and colleagues have now shown that a data‑driven trick called implicit electric field conjugation can deepen those nulls by orders of magnitude while preserving much of the planet signal.
The project sits at the intersection of astronomy and photonics, a place where physicists and engineers treat starlight like something that can be tamed rather than simply blocked. The PLN builds on the idea that light can be sorted into multiple streams by a mode-selective photonic lantern, a kind of light‑sorting lantern with six ports. Four of those ports are nulled at the star’s center, so the on‑axis starlight essentially cancels out, while off-axis light from a potential planet can couple into the remaining ports. Think of it as a lattice of tiny doors arranged around a bright lantern; twirl the wavefront just so, and the doors shudder shut for the star but yield to the planet. This architecture holds the promise of characterizing exoplanets right where conventional coronagraphs struggle—the diffraction limit, within about one beam width of the host star.
Xin’s team does not claim to have solved the problem of exoplanet imaging on the sky tomorrow. Instead, they’ve demonstrated a powerful proof of concept: a data‑driven method that learns how the deformable mirror in the optical train affects the intensity that leaks into each PLN port, then actively pushes the stellar light down a deeper channel. It’s a bit like teaching a choir of tiny echo chambers to destructively interfere at the star’s core while leaving the planets to sing in the wings. The study makes clear that the payoff is real—deeper nulls, better planet S/N, and a path toward high‑resolution spectroscopy of worlds that are currently just out of reach. The work is a triumph of careful instrument design meeting flexible control theory, and it weighs in at a moment when exoplanet science prizes both precision and breadth: knowing not just that a planet exists, but what it’s made of, what its atmosphere sounds like, and how it moves through its star’s gravity well.
What the PLN is and why it matters
The PLN is built around a six‑port mode‑selective photonic lantern. Light from a telescope’s multimode input is transformed so that each mode corresponds to a specific single‑mode output. Because the lantern is designed with dissimilar cores, the ports map to distinct eigenmodes of the lantern’s optics, effectively sorting light by its spatial structure. Importantly, the symmetries of these ports create a central null: when starlight comes in perfectly aligned with the lantern’s center, it hardly couples into four of the ports. Off‑axis light—potentially from a planet—can couple with higher efficiency into those ports. The result is a set of four nulled outputs that suppress the star in different ways, offering multiple channels for detecting an off‑axis companion and for pinning down its position even when the signal is faint.
In theory, this arrangement is powerful because it doesn’t rely on a single dark hole in the focal plane. Instead, it distributes the critical task of starlight suppression across several ports, which should make it easier to capture a planet’s light with minimal loss. In practice, however, small imperfections—tiny misalignments, wavefront errors, or manufacturing quirks in the lantern—mean the ports do not behave identically. That’s where iEFC comes in. It’s a data‑driven cousin of traditional electric field conjugation: instead of building a detailed physical model of the entire optical train, the method learns how the deformable mirror commands translate into changes in what’s measured at the lantern’s ports. By iteratively probing the system and updating a calibration matrix, iEFC pushes the star’s light into deeper nulls while keeping the planet light from being thrown away.
Crucially, the team’s results reinforce a practical truth of high‑contrast optics: not all ports behave the same way, and sometimes you gain more by focusing on a single channel than by chasing a symmetrical solution across all channels. That insight comes out of the data: when the authors tried to suppress all four nulled ports at once, three of them benefited and deepened substantially, but one port stalled—likely limited by the lab’s stability rather than the lantern itself. When they targeted that stubborn port alone, they achieved a deeper null for that channel, albeit with different DM activity. The upshot is a nuanced, real‑world recipe: multiport nulling offers broad sky coverage, but the deepest starlight rejection in a single place may be the cleanest route when a planet’s location is known.
Raising the bar with implicit electric field conjugation
Implicit electric field conjugation is a clever twist on focal‑plane wavefront control. In brief, the stellar electric field changes when the deformable mirror is nudged with probe patterns. By comparing pairs of images—one with a probe and one with the probe reversed—the scientists build a linear relationship between those probe commands and the field perturbations at the detector. This relationship is summarized in a calibrated response matrix S. The iEFC solution then finds the mirror commands α that minimize the residual perturbation δ in the measured image, while also keeping the commands from running away with energy through a regularization term. The math is familiar to wavefront control veterans, but the data‑driven aspect is especially well suited to a mode‑sorting lantern where the outputs are not simply pixels but mixed into ports by an overlap integral.
The practical payoff is tangible. In a representative run that targeted all four nulled ports, the authors observed that LP 11a and LP 11b—two of the central channels—deepened their nulls substantially, while LP 21a also improved, and LP 21b lagged behind likely due to lab instability rather than a fundamental limitation of the lantern. The reported numbers are eye‑opening: stellar coupling ηs dropped from the 10^−3–10^−2 range to the 10^−5–10^−4 range, and the resulting null depths ηs/ηppeak improved by roughly two orders of magnitude for the ports that responded well. At the same time, the planet throughput ηppeak remained meaningful for the nulled ports, showing that you don’t have to give up planet light to push starlight away.
The study doesn’t gloss over the messy physics that still bites in the real world. The lantern’s ports exhibit cross‑coupling and polarization differences on the order of 10^−3 in intensity, expectations of a few tenths of a percent in reality. Those small mismatches propagate through the mode sorting, producing amplitude errors in the pupil plane that limit how perfectly a single DM can compensate. The team’s simulations—comparing a real‑world lantern to an ideal, perfect lantern—support a hopeful conclusion: the physics isn’t a fundamental barrier. In a hypothetical world with an ideal lantern and perfect wavefront control, nulls could reach unbelievably deep levels. The limiting factor is sensing rather than physics: the current probing scheme is most sensitive to a few low‑order aberrations (Coma and Astigmatism), and higher‑spatial‑frequency errors are not as readily captured. In other words, next‑generation sensing strategies could unlock even deeper nulls without sacrificing planet light.
Deep dives, tradeoffs, and the road to real skies
The paper doesn’t pretend that this is a completed instrument ready for first light on an actual telescope. It is, however, a carefully argued blueprint for turning a clever optical concept into a robust scientific tool. The authors performed parallel simulations to explore how iEFC behaves under different lantern models and aberration regimes. In a perfect lantern with static phase aberrations, the nulls could be pushed to ηs around 10^−9 to 10^−12 in the absence of noise—an extraordinary demonstration of what the architecture could do with ideal sensing and control. When they introduced realistic amplitude and phase errors in the pupil, the best attainable nulls settled at much shallower levels, typically around 10^−5 to 10^−7, and the port‑to‑port tradeoffs reappeared. The signal was consistent: sensing limitations, not light itself, bottlenecked the performance in these more realistic cases.
The authors do not stop at theory. They map out practical implications for ground‑based and space telescopes. On a sky survey, where the planet’s location may not be known ahead of time, using multiple ports—possibly all four—offers broader coverage of the possible planet positions. If the planet’s location is known, it can be advantageous to concentrate on one or two ports to maximize sensitivity at that location. The role of adaptive optics—and whether a second deformable mirror can be slotted into the train for real‑time iEFC or whether daytime calibration followed by open‑loop observations is preferred on sky—are not settled questions but active design decisions for next‑generation instruments. The Keck KPIC experience, which already uses a separate DM to maintain a clean path for single‑mode fiber spectroscopy, hints at how a real instrument might adopt a similar modular approach for iEFC‑driven nulling.
Beyond single‑temporal experiments, the PLN with iEFC also raises questions about broadband operation. The current demonstrations used a monochromatic laser, but prior lab work suggests the PLN can achieve broadband nulls in the 10^−3 regime. The authors acknowledge that chromatic dispersion remains a hurdle to a fully operational spectroscopic instrument. The path forward likely involves dispersing the outputs of the lantern or building broadband control into the iEFC loop so that the nulls stay deep across the wavelengths needed to resolve planetary atmospheres. In the longer term, researchers may explore a hybrid lantern design with a mix of symmetric and asymmetric ports, or even a new class of sensing schemes tailored to the unique mathematics of mode sorters.
So what does this mean for the quest to characterize worlds around other stars? It means a future where the same instrument that finds a planet can also dissect its atmosphere, in some cases from within one beam width of the star. The PLN’s multiport approach, combined with a responsive iEFC loop, could enable high‑resolution spectroscopy at separations that are inaccessible to conventional coronagraphs today. It could also support spectroastrometry, where tiny shifts in the planet’s light reveal its position with exquisite precision, albeit with a 180° degeneracy in the orbital angle. The science value is clear: deeper nulls translate into higher contrast, higher signal‑to‑noise ratios, and the potential to detect faint molecular signatures that tell us whether a world might be habitable or even inhabited.
Lead author Yinzi Xin and a dedicated team at Caltech’s Exoplanet Technology Lab, in collaboration with institutions around the world, have given the field a concrete, testable pathway to sharpen our view of nearby planetary systems. And while the PLN is not yet a turnkey sky instrument, its progress is a reminder that progress in astronomy often travels through the laboratory first: a lantern that teaches us how to coax light into a more honest conversation with the cosmos. If the next generation of telescopes can translate these lab successes to on‑sky demonstrations, we may look back on this moment as a pivotal turn—from passive observation toward active, tunable listening in the faint glow of distant worlds.
