A breakthrough holographic system uses AI and multidimensional light to store vastly more data in less space.

Researchers have created a new method for holographic data storage that captures and retrieves information in three dimensions by combining three key properties of light: amplitude, phase, and polarization. By using these properties together, the technique allows significantly more data to be stored within the same physical space, offering a potential solution to the world’s rapidly increasing demand for data storage.

Unlike traditional storage technologies that write data onto flat surfaces such as hard drives or optical discs, holographic storage embeds information throughout the entire volume of a material using laser light. This volumetric approach enables many overlapping light patterns to exist within the same space, greatly increasing storage density while also improving data transfer speeds.

“In conventional holographic data storage, data encoding typically uses one light dimension such as amplitude or phase alone, or, at most, combines two of these dimensions,” said research team leader Xiaodi Tan from Fujian Normal University in China. “Based on the principle of polarization holography, we used a deep learning architecture known as a convolutional neural network model to enable the use of polarization as an independent information dimension.”

The findings, published today (March 26) in Optica, Optica Publishing Group’s journal for high-impact research, show that this new technique can boost how much information is stored while also making it easier to read the data back.

“With further development and commercialization, this type of multidimensional holographic data storage could enable smaller data centers and more efficient large-scale archival storage, while also enhancing data processing and transmission efficiency,” said Tan. “It could also contribute to safer data transmission, optical encryption, and advanced imaging.”

[embedded content]

This video shows how the researchers were able to retrieve 3D information directly from diffraction intensity images. Credit: Xiaodi Tan, Fujian Normal University in China

Using Polarization as a Data Channel

In holographic storage, information is recorded as page-like images formed by laser light patterns. Encoding transforms digital data into these optical pages, while decoding converts the stored patterns back into usable information.

Although light offers multiple properties that could be used to encode more data, combining them effectively has been a major technical challenge. To overcome this, the researchers refined a method called tensor-based polarization holography, which preserves the polarization state of light when the hologram is reconstructed. This makes polarization a stable and reliable way to carry additional information.

Building on this foundation, the team developed a 3D modulation encoding strategy. By carefully controlling the intensity and phase of two perpendicular polarization states and applying a double-phase hologram technique, they enabled a single phase-only spatial light modulator to encode amplitude, phase, and polarization simultaneously within the optical field.

AI-Powered Decoding of 3D Light Data

Retrieving this multidimensional data presents a challenge because standard sensors can only detect light intensity (amplitude) and cannot directly measure phase or polarization. To address this limitation, the researchers combined tensor-polarization holography with a convolutional neural network capable of reconstructing all three dimensions of information from intensity-based measurements.

The neural network is trained using two types of diffraction images: one captured through a vertical polarizer and another taken without it. By analyzing these complementary images, the system learns to identify patterns related to amplitude, phase, and polarization, allowing it to decode all three simultaneously. This approach increases storage capacity while also improving data transmission speed.

Toward High-Capacity, High-Speed Storage

After confirming the concept, the researchers built a compact experimental setup to record and reconstruct the encoded optical fields within a polarization-sensitive material. During testing, intensity images were examined to extract signatures linked to amplitude, phase, and polarization. These features were then fed into the neural network, enabling full 3D data reconstruction using only intensity measurements.

“Overall, our results showed that multidimensional joint encoding substantially increased the information carried by a single holographic data page, thereby improving storage capacity,” said Tan. “In addition, neural network synchronous decoding reduced the need for complex measurements and step-by-step reconstruction, supporting more efficient readout and decoding. This could enable a practical route toward high-capacity, high-throughput holographic data storage.”

Next Steps Toward Real-World Applications

The researchers emphasize that the technology is still at the experimental stage and requires further refinement before it can be widely used. Future work will focus on increasing the number of gray levels in the encoding process to expand capacity even further, as well as improving the durability, consistency, and reliability of the storage materials.

They also plan to integrate this approach with volumetric holographic multiplexing techniques, which could allow multiple pages and channels of data to be stored simultaneously. Strengthening the integration between optical hardware and AI-based decoding systems will be key to achieving faster and more reliable data retrieval in real-world conditions.

Reference: “Encoding and decoding of multidimensional optical field modulation in holographic data storage” by 19 April 2026, Optica.
DOI: 10.1364/OPTICA.586593

Never miss a breakthrough: Join the SciTechDaily newsletter.
Follow us on Google and Google News.