Embedded Files
Control of a new degree of freedom of light
Control of a new degree of freedom of light
Published in Nature
August 12, 2026
Experimental polarization-coloured image. Colours correspond to distinct polarization states rather than optical wavelengths.
Spatial light modulation is a cornerstone of modern photonics. Crucial advances in photonic information processing rely on the spatial manipulation of the optical phase and state of polarization (SOP), capabilities driving emergent fields such as neuromorphic photonics. The degree of polarization (DOP) is a fundamental property that, if controlled, provides an extra resource. However, no technology exists at present that can spatially modulate the DOP in a programmable manner.
In our work published in Nature, we demonstrate spatial DOP modulation, unlocking a new degree of freedom of light for information processing. We introduce a statistical approach to control the DOP: by engineering the statistics of light at the micrometre scale with phase-only spatial light modulators (SLMs), we achieve full programmability of the DOP and SOP across the beam wavefront, creating structured light sculpted in its polarization content to arbitrary shapes. This enables information encoding in a high-dimensional space — the volume of the Poincaré sphere. We encode colour images in a single-wavelength laser by creating what we call polarization colours and exploit the DOP modulator to expand the dimensionality of various photonic neural networks, demonstrating a general route to high-dimensional optical computing and multidimensional optical encryption. The finding establishes DOP as a powerful degree of freedom that can play a key role in emerging photonic technologies.
Read the full article in Nature:
PhD position - ERC LOOP project
PhD position - ERC LOOP project
PhD in Physics at Sapienza funded by LOOP
June 18, 2026
Experimental setup in the LOOP laboratories
PhD position at Sapienza University of Rome within the project ERC Starting Grant 2024 LOOP Optical Polarization for Ultrafast Computing.
Fully-funded 3-year scholarship to work on nonlinear and complex photonics within the PhD program in Physics.
Research topic: Experimental photonics for developing novel optical computing machines based on the polarization of light
Location: Department of Physics, Sapienza University of Rome, Rome, Italy
Supervisor: Prof. Davide Pierangeli
Application deadline: July 17, 2026
Info: davide.pierangeli@uniroma1.it
Note: This position is now closed. Interested candidates looking for future opportunities are encouraged to contact Prof. Davide Pierangeli.
The first observation of the lump soliton
The first observation of the lump soliton
Published in Physical Review Letters and Featured in Physics
February 6, 2026
Solitons — nonlinear waves that behave like particles — are fundamental entities in nonlinear physics. They are ubiquitous in one-dimensional systems, but the question of whether they exist in higher dimensions without collapsing has remained unanswered for a long time. In the 1970s, a two-dimensional soliton known as lump soliton — a bump of energy localized in space — was theoretically predicted in hydrodynamics, yet never experimentally identified in any physical system.
Within a collaboration coordinated by Prof. Claudio Conti (Sapienza), we report the first experimental evidence of lump solitons. In our nonlinear optics experiment, the wavefunction of a fluid of light in a photorefractive crystal is engineered by spatial light modulation to excite and control an exact lump, which propagates conserving its wave shape and collides elastically. The result reveals that lumps exist in nature, opening new scenarios in classical and quantum wave physics.
The article, published in Physical Review Letters, has been selected as an Editors' Suggestion and highlighted in Physics.
Highlight in Physics: Solitons Take Their Lumps in Two Dimensions
Photonic Neural Network Polarimeter
Photonic Neural Network Polarimeter
Published in Laser & Photonics Reviews
December 26, 2025
Concept of the photonic neural network polarimeter and its implementation using a stack of scattering media and a camera sensor.
Optical neural networks are emerging as a powerful tool for processing optical signals directly in the optical domain with superior speed and integrability, achieving functionalities beyond current devices. While their application in polarimetry has enabled novel neuromorphic polarization sensors, their operation has so far been restricted to fully polarized light.
In our work published in Laser & Photonics Reviews, we introduce a photonic random neural network (PRNN) that achieves single-shot full-Stokes polarimetry of partially polarized light. The PRNN consists of optical random layers (ground-glass diffusers) connected through an intensity camera sensor to a few trainable digital nodes. By exploiting multiple light scattering, the optical network projects the input polarization components into intensity speckle patterns, enabling simultaneous measurement of the degree of polarization and the Stokes parameters with precision comparable to conventional polarimeters. The device operates at high speed across multiple visible wavelengths without requiring polarization optics. This fast, compact, and broadband neuromorphic polarimeter offers a widely applicable route for high-speed polarization sensing, imaging, and optical communications.
Read the full article: Alessandro Petrini, Claudio Conti, and Davide Pierangeli, Single-Shot Full-Stokes Analysis of Partially-Polarized Light with a Photonic Deep Random Neural Network, Laser & Photonics Reviews, e01467 (2026).
High-dimensional Ising machine by polarization oscillators
High-dimensional Ising machine by polarization oscillators
Published in Physical Review Letters
August 4, 2025
Scheme of the polarization Ising machine based on the dimensional collapse of nonlinear polarization oscillators in a nonlinear crystal.
Photonic Ising machines have emerged as promising hardware for rapidly solving NP-hard optimization problems by mapping spin models onto interacting optical states. However, current paradigms based on low-dimensional spaces — such as optical phase or amplitude — frequently get trapped in local energy minima, severely restricting their computational efficiency and scalability.
To overcome these limitations, our recent work published in Physical Review Letters introduces the Polarization Ising Machine (PIM), a photonic architecture that exploits the full three-dimensional state space of light polarization. Through a novel computing mechanism termed dimensional collapse, the system operates using nonlinear polarization oscillators in third-order nonlinear media that evolve within the volume of the Poincaré sphere and spontaneously settle into binary fixed points, naturally encoding Ising spins. By exploiting the extra dimensions, phase-space trajectories can dynamically bypass the energy barriers that trap conventional Ising machines. Benchmarks on Max-Cut problems and dense graphs demonstrate an exponential performance scaling advantage. The polarization Ising machine paves the way for a new class of ultrafast and scalable photonic hardware with enhanced computing capabilities.
Read the full article: Salvatore Chiavazzo, Marcello Calvanese Strinati, Claudio Conti, and Davide Pierangeli, Ising Machine by Dimensional Collapse of Nonlinear Polarization Oscillators, Phys. Rev. Lett. 135, 063801 (2025).
Page updated
Report abuse