Beyond Vision Restoration: Broadband Retinal Nanoprosthetics with Tellurium Nanowire Networks — open-access commentary

Citation: Zhang S, Rong Z, Chu L. “Beyond Vision Restoration: Broadband Retinal Nanoprosthetics with Tellurium Nanowire Networks.” Research (Wash DC) 8: 0857 (published 9 Sep 2025). DOI: 10.34133/research.0857. PMID: 40933234. PMCID: PMC12417632. CC BY 4.0. (Hangzhou Dianzi University.)

Note on capture: Open-access commentary (CC BY 4.0) on Wang et al. 2025 (Science); captured from PMC full text on 2026-09-14. Provides technical detail not present in the main paper’s public abstract.

Abstract

Retinal degenerative disorders pose critical challenges for vision restoration, with traditional prosthetics limited by spectral sensitivity and invasive surgeries. In recent work published in Science, revolutionary tellurium nanowire networks converted broadband visible to near-infrared-II (up to 1,550 nm) light into electrical signals without external bias, restoring functional vision in blind animal models.

Technical details (from full text)

• Material: Te narrow bandgap 0.3 eV; high optical absorption coefficient (~9% per layer) enables low-energy IR photon detection. The centrosymmetric crystal structure otherwise limits the photovoltaic effect.
• Asymmetry engineering: during chemical vapor deposition, the SnTe2 precursor created substitutional Sn defects and Te vacancies within the helical Te lattice, breaking structural symmetry and generating localized electronic states near the Fermi level → charge separation and transport. Heterointerfaces between TeNWNs and retinal cells (simulated with Au electrodes) create additional asymmetry → built-in electric fields.
• Network morphology: randomly oriented nanowires form a porous, interconnected network ~150 nm thick, “simulating the degenerated photoreceptor layer”; vertical alignment promotes perpendicular contact with bipolar cells and RGCs; minimizes mechanical stress; high surface area → efficient light absorption and charge transfer.
• No external bias → no intraocular power supplies or extraocular goggles. vs. polymer nanoparticle systems: broader spectral sensitivity (VIS–NIR-II vs. primarily VIS), intrinsic zero-bias operation, lower retinal stress (porous network vs. bulk injections); polymers may scale better in some fabrication processes. Aligns with the “analog compute-in-memory” design paradigm.
• Mouse results (Pde6brd1/rd1/cDTA; complete photoreceptor loss): robust neural responses to VIS (470–635 nm) and NIR-II (1,550 nm); RGC action-potential activation thresholds as low as 18.98 mW mm⁻² (NIR-II); response latency 200–600 ms; frequency tuning up to 5 Hz (mirroring natural photoreceptor signalling); occipital-cortex VEPs confirm propagation to higher visual centers. Behavior: robust pupil constriction to 635-nm VIS and 1,550-nm NIR-II light (untreated blind mice: no IR reflex).
• Primate results (Macaca fascicularis): fundus imaging + OCT: stable subretinal integration for 112 days, no signs of retinal detachment or inflammation; flash ERGs: robust a-wave (photoreceptor origin) and b-wave (bipolar-cell origin) responses to NIR-I 940 nm, amplitudes significantly higher in implanted eyes vs. unimplanted controls.
• Stated challenges: spatial resolution — mouse receptive field 135.57 deg² vs. 163.02 deg² for normal RGCs (limited spatial discrimination; density vs. biocompatibility trade-offs); long-term stability: TeNWN degradation/glial scarring unresolved beyond 112 days; spectral tunability: isolating specific bands (e.g., colour) unaddressed; human translation barriers: subretinal surgery risk (injectable nanowire suspensions floated as a mitigation), chronic toxicity, Te-ion (degradation product) clearance, cortical re-education/neurofeedback.
• Outlook: “paradigm shift in vision restoration”; beyond therapy, “augmented vision systems that have the potential to redefine human interaction with the electromagnetic spectrum”.
• Competing interests: none declared.