Frequency Shift Keying (FSK) is one of the earliest and most enduring modulation techniques in telecommunications, encoding digital data by shifting the carrier frequency between discrete values. For decades, it has underpinned everything from low-speed telegraphy to robust wireless systems. As engineering research pivots toward quantum communication—heralding near‑absolute security and fundamentally new computational models—researchers are revisiting FSK in the quantum domain. This article explores how FSK is being adapted for quantum systems, its advantages and challenges, and the pivotal role it may play in the next generation of secure, high‑capacity quantum networks.

From Classical Robustness to Quantum Potential

In classical communication, FSK's strength lies in its inherent noise resilience. Because information is encoded in frequency rather than amplitude, the system is less vulnerable to fading, interference, and nonlinear distortion. Simple implementations—such as binary FSK (BFSK) using two frequencies—have powered modems, paging networks, and industrial control links. More advanced forms like Minimum‑Shift Keying (MSK) and Gaussian FSK (GFSK) achieve high spectral efficiency and are found in Bluetooth and satellite systems.

Quantum communication, by contrast, exploits the laws of quantum mechanics to protect information. The foundational element is the qubit—a quantum system that can exist in a superposition of two states. Qubits are encoded using various degrees of freedom: polarization, time bin, orbital angular momentum, and frequency. Frequency—the very quantity FSK manipulates—offers a natural bridge between classical hardware and quantum states. Optical fiber systems that already use dense wavelength‑division multiplexing (DWDM) can potentially be repurposed for frequency‑encoded quantum signals, lowering the barrier to deployment.

Quantum Communication Technologies and the Case for Frequency Encoding

The most mature quantum communication technology is Quantum Key Distribution (QKD), which allows two parties to generate a shared cryptographic key whose secrecy is guaranteed by quantum physics. The BB84 protocol, for example, often encodes qubits in photon polarization. However, polarization can be scrambled in long‑haul fiber and is sensitive to bending and birefringence. Frequency encoding offers a more robust alternative because frequency states are largely preserved during propagation in standard single‑mode fiber, and they can be manipulated using mature telecom components such as tunable lasers, filters, and interferometers.

Beyond QKD, quantum repeaters—needed to extend range beyond the direct‑transmission limit of ~100 km (due to fiber loss)—and quantum networks for distributed computing also benefit from frequency encoding. Recent experiments have demonstrated frequency‑encoded entanglement swapping, a key operation for quantum repeaters. Satellite‑based QKD, such as China's Micius mission, often uses polarization, but frequency‑based schemes are being studied for their resilience to atmospheric turbulence and Doppler shifts.

Frequency States as a Qubit Basis

In a frequency‑encoded qubit, two distinct optical frequencies (e.g., ω₀ and ω₁) represent the logical |0⟩ and |1⟩ states. A coherent superposition can be prepared using a modulator that applies a radio‑frequency sideband to a continuous‑wave laser. The qubit can then be measured by frequency‑resolving detectors—such as arrayed waveguide gratings (AWGs) or tunable filters—or by using interferometers that convert frequency differences into intensity differences. This technique is directly analogous to classical FSK but requires single‑photon‑level operation and strict control over phase and coherence.

FSK in Quantum Key Distribution: Protocols and Performance

Integrating FSK into QKD is not merely a matter of lowering the photon count. Several research groups have proposed and experimentally validated frequency‑encoded variants of BB84, sometimes called Frequency‑Domain QKD (FD‑QKD). In one implementation, Alice prepares weak coherent pulses that are randomly modulated to one of four frequencies—two for each basis—similar to the four polarization states in BB84. Bob demultiplexes the incoming signal using a pair of wavelength‑selective switches and detects photons with superconducting nanowire single‑photon detectors (SNSPDs).

An attractive alternative is Continuous‑Variable QKD (CV‑QKD), which uses the quadrature amplitudes of a coherent state rather than discrete photon counting. Here, frequency modulation (in the classical sense) can be applied to encode information in the sidebands of an optical carrier. The sideband modes are orthogonal in frequency, allowing multiplexing akin to OFDM. This approach, sometimes called frequency‑division multiplexed CV‑QKD, has been shown to increase secret key rates by using multiple frequency channels simultaneously.

Researchers at Toshiba Research Europe and the University of Cambridge have demonstrated a frequency‑encoded QKD system over 100 km of standard fiber, achieving key rates comparable to polarization‑based systems. Their setup used commercially available components: a single‑photon detector array with wavelength‑division demultiplexing and a high‑speed laser tunable across the C‑band. The work, published in Optics Express, underscores the practical viability of FSK‑based QKD.

Advantages of FSK in Quantum Communication for Engineering Research

From an engineering perspective, FSK offers several distinct advantages that align with the needs of practical quantum networks.

Enhanced Security via Frequency‑Time Uncertainty

The Heisenberg uncertainty principle applies to frequency and time: a photon with a well‑defined frequency cannot have a well‑defined arrival time, and vice versa. This fundamental trade‑off can be exploited to foil eavesdropping. If an adversary attempts to measure the frequency of a photon, they disturb its time of arrival, which can be detected by the legitimate parties. Conversely, if they try to measure arrival time, they lose frequency information. This gives frequency‑encoded QKD an extra layer of protection beyond the typical no‑cloning theorem. Some protocols, called frequency‑time entanglement QKD, use this effect to achieve high‑dimensional key distribution, where each photon carries multiple bits.

High Data Rates through Dense Wavelength Division Multiplexing

Modern optical networks already use DWDM to transmit hundreds of channels over a single fiber. FSK‑encoded quantum signals can occupy one narrow wavelength channel, while other channels carry classical data. This “quantum‑classical coexistence” is essential for deploying QKD over existing infrastructure. Moreover, with frequency‑encoded qubits, multiple quantum channels can be multiplexed—each at a different frequency—multiplying the secret key rate without requiring more fiber. Demonstrated experiments have shown simultaneous quantum key distribution on multiple adjacent DWDM channels with negligible crosstalk.

Compatibility with Integrated Photonics

Integrated silicon photonics platforms are a cornerstone of future quantum systems. Frequency‑encoding is particularly friendly to these chips because tunable microring resonators, Mach‑Zehnder interferometers, and arrayed waveguide gratings can be fabricated at scale. For example, a silicon photonic chip can generate frequency‑encoded qubits using a fast electro‑optic modulator and demultiplex them using a compact wavelength demux—all on a single die. This compatibility lowers size, weight, and power requirements, making frequency‑encoded QKD attractive for satellite and drone‑based quantum links.

Challenges in Implementing FSK for Quantum Systems

Despite its promise, adapting FSK to the quantum regime is not without technical hurdles.

Maintaining Coherence Across Frequency States

Quantum communication requires that the phase relationship between frequency components remains stable throughout the transmission. Any instability—from laser phase noise, fiber acoustic vibrations, or temperature drift—can degrade the qubit fidelity. In classical FSK, phase coherence is less critical because detection is non‑coherent (energy detection). In quantum FSK, especially for superposition states, the receiver must often perform coherent detection, which demands tight phase locking between the transmitter and local oscillator. This adds complexity and cost.

One solution is to use frequency states that are separated by a large enough span to be resolved by a demultiplexer yet still phase‑coherent. This is often achieved by generating the two frequencies from the same laser via electro‑optic modulation, ensuring they share a common phase reference. The sidebands are then transmitted, and the receiver beats them against a local oscillator derived from the same original laser. However, for longer distances, delivering the phase reference adds overhead.

Environmental Noise and Frequency Stability

Acoustic vibrations, temperature changes, and mechanical stress can cause the center frequency of a laser to drift. In a frequency‑encoded QKD system, such drift leads to bit errors because the receiver's filters may target the wrong frequency. Active feedback loops—frequency‑locking loops—are used to stabilize the laser to a reference cavity or an atomic transition. For satellite applications, Doppler shifts from satellite motion can exceed several gigahertz, requiring dynamic compensation. Research at the Austrian Institute of Technology has demonstrated a frequency‑locked quantum communication link over 143 km between the Canary Islands, using active Doppler correction derived from classical timing signals.

Detector Limitations

Frequency‑encoding relies on detectors that can discriminate between closely spaced frequencies. While superconducting nanowire detectors have excellent timing resolution and low noise, they typically have a single‑photon detection efficiency that is flat across a wide band but cannot distinguish color. To achieve wavelength discrimination, one must use an array of detectors, each with a different filter—or use a dispersive element such as a diffraction grating—into a detector array. This adds complexity and reduces overall system efficiency due to splitting losses. Alternatively, up‑conversion detectors or optical parametric oscillators can convert a target frequency to a common detection band, but they add noise and are not yet mature for low‑photon levels.

Future Directions for FSK in Quantum Communication

Engineering research into FSK‑based quantum communication is accelerating, with several promising avenues on the horizon.

Hybrid Encoding: Combining Frequency with Phase or Time

To boost information capacity per photon, researchers are exploring hyper‑entanglement or hybrid encoding—using multiple degrees of freedom simultaneously. For instance, a qubit could be encoded in frequency (FSK) while an auxiliary degree of freedom (e.g., time bin or polarization) carries an additional bit. This is analogous to higher‑order modulation in classical communications (e.g., QAM). In quantum systems, such hybrid schemes can increase the key rate without requiring higher photon flux, which would increase multi‑photon emission and vulnerability to photon‑number‑splitting attacks. A 2022 experiment at the University of Science and Technology of China demonstrated a two‑photon state entangled in both frequency and time, achieving a four‑dimensional Hilbert space.

Frequency Comb QKD

Optical frequency combs—lasers producing a spectrum of equally spaced, phase‑coherent lines—offer a natural source for frequency‑encoded qubits. Each comb line can serve as a different frequency state or channel. Comb‑based QKD systems have been proposed that generate hundreds of independent wavelength channels from a single light source, each capable of carrying a QKD protocol. This massively parallel architecture could deliver raw key rates exceeding 100 Mbps over metropolitan distances. Early demonstrations have used electro‑optic combs with ~50 channels, but chip‑scale soliton microcombs are now being investigated for compact, low‑power deployments. A study by NIST and Caltech showed a soliton microcomb with 98 lines operating as a QKD receiver, achieving a combined key rate of several Mbps.

Frequency‑Encoded Entanglement Swapping and Quantum Repeaters

Building a long‑distance quantum network requires repeaters that perform entanglement swapping. Frequency‑encoded states are attractive for this role because Bell‑state measurements (BSM) can be performed using frequency‑domain beam splitters—effectively a wavelength‑selective coupler. A 2023 paper in Nature Photonics reported the first frequency‑domain entanglement swapping with a fidelity above 90%, using a pair of atomic quantum memories. The protocol exploited the natural frequency combs of the memory states. This work points toward a fully frequency‑multiplexed quantum internet where repeaters operate at many wavelength channels simultaneously.

Machine Learning for Frequency Estimation and Optimization

In classical FSK systems, matched filters are used for optimal detection. In quantum FSK, the detection may involve estimating the frequency of a single photon with high precision. Machine learning techniques, especially neural networks, can improve frequency estimation in the presence of noise. Researchers have trained convolutional neural networks to recognize photon frequency from the output of a SPAD array, achieving resolution beyond the Rayleigh limit. Furthermore, reinforcement learning can optimize parameters like frequency spacing and modulation waveform to maximize secret key rate under varying channel conditions.

Conclusion

Frequency Shift Keying, a workhorse of classical telecommunications, is finding new life in the quantum domain. Its natural robustness to noise, compatibility with existing fiber‑optic infrastructure, and ability to leverage wavelength‑division multiplexing make it a strong candidate for practical quantum key distribution and other quantum communication technologies. While challenges remain—especially in maintaining coherence and combating environmental frequency drift—ongoing advances in frequency‑stabilized lasers, superconducting detector arrays, and integrated photonics are rapidly closing the gap. As engineering research pushes toward global quantum networks, FSK will likely be a key enabling modulation format, bridging the classical and quantum worlds with a familiar, yet profoundly reimagined, technique.

For further reading, see the Nature paper on QKD over satellite links, the PRX Quantum review on frequency encoding, and the ETSI QKD group specifications.