The central obstacle the team set out to overcome is one that has haunted optical storage for decades. While solid-state drives have multiplied their capacity through multilevel encoding schemes, stacking several bits of information onto each memory cell, optical disks have struggled to do the same. The photo-responsive materials used in conventional disks have a limited dynamic range, meaning the difference between the brightest and dimmest signals they can reliably produce is narrow. Layer on top of that the inherent noise of optical readout, and the signal-to-noise ratio collapses long before the number of distinguishable levels becomes useful. The result is that optical encoding fidelity has remained stuck, preventing optical platforms from keeping pace with the explosive growth of electronic storage.

The Shanghai team’s answer lies in a phenomenon called anti-Stokes photoluminescence, and more specifically in a giant, phonon-assisted version of it. In ordinary Stokes photoluminescence, a material absorbs a high-energy photon, say blue light, and re-emits a lower-energy photon, say green or red, with the lost energy dissipated as heat. Anti-Stokes emission flips that logic: the material absorbs low-energy light and emits higher-energy light, seemingly violating intuition about energy flow. In practice, the missing energy is borrowed from lattice vibrations, the phonons of the material, which donate small quanta of thermal energy to push the excited electron up to a higher emissive state. When this phonon assistance is strong enough, the researchers call it giant, the upconverted emission becomes bright and stable enough to serve as a practical data signal.

The material at the heart of the system is a class of nanoparticles known as carbonized polymer dots, or CPDs. These are nanoscale clusters formed when polymer precursors undergo cross-linking-induced nucleation and partial carbonization, producing hybrid organic-inorganic structures with rich, tunable photoluminescence. Rather than synthesizing the dots in a flask and dispersing them into the disk beforehand, the team uses the writing laser to carbonize the polymer medium in situ. A focused beam converts the polymer locally into upconversion carbonized polymer dots, or u-CPDs, precisely where a data bit should reside. This in situ approach has a subtle but crucial advantage: the dots are born embedded in their final environment, with their optical properties shaped by the writing conditions themselves, which the researchers show can be tuned through the laser power, exposure and photoinitiator chemistry.

The physics of the readout process is where the paper makes its most striking claims. When near-infrared light illuminates the laser-written u-CPDs, the dots emit upconverted photoluminescence at higher energy, a process the team attributes to phonon-assisted anti-Stokes excitation. Because the writing and reading both exploit near-infrared light, the approach maximizes the utilization of that wavelength band while achieving high spatial resolution with ultralow noise. Conventional Stokes-based media suffer from background fluorescence and crosstalk between layers, which erodes the contrast between adjacent greyscale levels. The upconversion route sidesteps much of that noise floor, allowing the dynamic range of the material to be divided into far more distinguishable steps. In their demonstration, the researchers resolved 1,024 distinct greyscale levels in three-dimensional multilayer encoding, an enormous leap over the binary or few-level schemes of traditional optical disks.

Those numbers translate into headline capacity. By combining multilayer three-dimensional recording with 1,024-level greyscale encoding, the team estimates a storage capacity of approximately 400 terabits on a disk the size of a standard DVD. To put that in perspective, a dual-layer Blu-ray disk holds roughly 50 gigabytes, so the new figure represents a capacity gain of nearly a thousandfold on the same physical footprint. Equally important is the reliability: the demonstration maintained greater than 99.99 percent readout fidelity, a threshold that matters enormously for archival applications where a single misread bit can corrupt a file. The researchers also report a scalable pathway toward extending greyscale encoding beyond 10,000 levels through optimized phonon-photon coupling, which they suggest could surpass current petabit-scale optical storage capabilities.

The experimental system behind these results is a sophisticated combination of laser writing and optical readout architectures. The team developed confocal readout for high-fidelity reconstruction of complex patterns, as well as galvo-scanning-based dynamic writing for high-throughput fabrication. Supplementary demonstrations show large-area laser writing of optical data-bit arrays, including fully populated arrays of 1,089 sites with precise voxel placement and uniformity, and wide-field readout that reconstructs sequential frames at high speed. In one striking demonstration, the researchers encoded a high-definition animation of a running cat as twelve sequential frames of patterned data bits, then read it back through the phonon-assisted upconversion channel with sufficient fidelity for HD movie playback. These demonstrations underline that the platform is not merely a materials curiosity but a functioning storage pipeline with programmable patterning and parallel fabrication capability.

The mechanism underlying the giant upconversion response connects this work to a broader body of research on anti-Stokes processes in low-dimensional materials. Phonon-assisted upconversion has been observed in carbon nanotubes, two-dimensional semiconductors such as monolayer tungsten disulfide, quantum dots and perovskites, where lattice vibrations bridge energy gaps that single photons cannot. Carbonized polymer dots, with their confined carbon domains and cross-link-enhanced emission, now join that family with what the authors describe as a giant response. The in situ laser carbonization step is also part of a growing toolkit of laser-based fabrication methods for patterned carbon materials, and the team’s characterization of how photoinitiator concentration, laser power and exposure govern bit formation provides a practical recipe for controlling the process. The researchers suggest that further optimization of the phonon-photon coupling could push the level count past 10,000, opening a route to encoding densities that would have seemed implausible for optical media only a few years ago.

The significance of the work becomes clearer when set against the looming data storage crisis. Global data volumes have been growing relentlessly, and archival storage, the long-term preservation of data that is written once and read rarely, is an acute bottleneck. Solid-state drives excel at speed but are costly and have limited endurance for cold data. Glass-based storage projects, including Microsoft’s Project Silica, have demonstrated laser writing in fused silica for dense, durable archival storage, and petabit-capacity three-dimensional nanoscale optical disks have been reported using other nanophotonic approaches. The Shanghai team’s contribution adds a new material platform to this landscape, one that leverages cheap polymer chemistry and near-infrared optics rather than exotic nanocrystals or ultrafast glass machining. Because the dots form in situ, the medium could in principle be a simple polymer film, keeping material costs low and fabrication scalable.

Challenges remain before 400-terabit disks appear on store shelves. The reported capacity is an estimate based on demonstrated encoding density, and translating laboratory-scale writing and readout into the fast, parallel, error-corrected systems required for commercial archival products is a formidable engineering task. Writing speed, laser power budgets, media longevity and compatibility with existing optical pickup architectures will all need to be proven at scale. Nevertheless, the demonstration of 1,024 greyscale levels with greater than 99.99 percent reliability, achieved through a mechanism as elegant as borrowing energy from the vibrations of the material itself, marks a genuine advance in the quest to keep optical storage relevant in the age of big data. If the promised extension beyond 10,000 greyscale levels can be realized, the humble optical disk, long written off as a relic of the DVD era, may yet become the workhorse of next-generation archival storage, preserving humanity’s digital record with unprecedented density and fidelity.

Subject of Research: Phonon-assisted upconversion photoluminescence in in situ photo-carbonized polymer dots for high-capacity three-dimensional optical data storage

Article Title: Giant phonon-assisted upconversion photoluminescence of in situ photo-carbonized polymer dots for 3D high-definition optical disks

Article References: Zhang, H., Lamon, S., Yu, H., Gu, W., Ding, H., Liu, L., Zhang, Q., & Gu, M. (2026). Giant phonon-assisted upconversion photoluminescence of in situ photo-carbonized polymer dots for 3D high-definition optical disks. Nature Photonics, 20(10), 1214-1223. https://doi.org/10.1038/s41566-026-02004-z

Image Credits: AI Generated

DOI: 10.1038/s41566-026-02004-z

Keywords: optical data storage, upconversion photoluminescence, carbonized polymer dots, phonon-assisted anti-Stokes emission, three-dimensional recording, greyscale encoding, near-infrared laser writing, archival storage, nanophotonics, laser carbonization, petabit capacity, data storage materials