A Quantum Memory That Speaks Two Languages

A useful way to think about a quantum network is as an international communication system.

Inside each city, people may speak one language.

Between cities, communication happens in another.

Imagine if every conversation required a human translator standing between every building and every highway.

It would work.

But it would also make the system more complicated, more expensive, and less reliable.

Quantum networks face a surprisingly similar problem.

Atomic quantum memories naturally speak one wavelength.

Optical fibers naturally prefer another.

Every conversion introduces additional opportunities for photons to be lost.

If the memory itself could communicate in both languages, the overall network would become much simpler.

Current quantum memories usually store photons near the wavelength where atoms interact most strongly.

If those photons need to travel through long optical fibers, they are first converted into telecom wavelengths.

Later, they may be converted back again before interacting with another memory.

The workflow looks something like this:

Quantum Memory
  -> Frequency Converter
  -> Telecom Fiber
  -> Frequency Converter
  -> Quantum Memory

Every additional component increases the complexity of the network.

Instead of translating photons before and after the memory, I wondered:

Could the memory itself understand both wavelengths?

The key idea came from an object in quantum optics known as a dark-state polariton.

Normally, a dark-state polariton couples a photon to a collective excitation inside an atomic cloud.

Existing quantum memories use this mechanism to store and retrieve light.

This work extends that idea.

Instead of involving only a single optical wavelength, I developed a theoretical framework where two different colors of light share the same collective atomic excitation.

One wavelength interacts naturally with atoms.

The other lies near the telecom band used for low-loss fiber communication.

Both become part of the same quantum object.

Once both wavelengths share the same atomic excitation, something interesting happens.

The memory can perform four different operations:

• Store a 795 nm photon and retrieve it at 795 nm.
• Store a telecom photon and retrieve it at telecom wavelength.
• Store a 795 nm photon and retrieve it as a telecom photon.
• Store a telecom photon and retrieve it as a 795 nm photon.

Instead of acting only as a memory, it also becomes a bridge between two parts of the quantum network.

To investigate whether this idea could actually work, I constructed a theoretical model based on a six-level rubidium atomic system.

The model extends the classic theory of dark-state polaritons to include two optical modes sharing the same spin-wave excitation.

I then simulated the complete light-matter dynamics during storage and retrieval.

The numerical simulations demonstrated successful operation for all four storage and retrieval pathways predicted by the theory.

Most research on quantum networking focuses on improving individual components.

Better memories.

Better frequency converters.

Better photon sources.

This project asks a different question:

Can we redesign the interface itself so that fewer components are needed?

Sometimes the best engineering solution isn't making every box better.

It's removing boxes altogether.

If future quantum memories can directly communicate using both node-compatible and telecom wavelengths, large-scale quantum networks could become significantly simpler to build.

This work is entirely theoretical.

Whether nature ultimately allows such memories to be realized experimentally remains an exciting question for future research.

More broadly, I think quantum technologies will increasingly benefit from architectures that simplify system integration rather than adding additional layers of hardware.

As quantum networks continue growing, reducing complexity may become just as important as improving individual device performance.

• Proposed a new dual-wavelength dark-state polariton.
• Extended the classic dark-state polariton theory to two optical modes.
• Developed a quantum memory supporting both node-band 795 nm and telecom-band 1324 nm photons.
• Demonstrated bidirectional storage and retrieval through numerical simulation.
• Established a potential route toward frequency-conversion-free quantum network interfaces.

Telecom-Compatible Cross-Band Quantum Memory via Dual Photon Modes Dark-State Polaritons

Authors: Dounan Du, Eden Figueroa

Paper on arXiv