Enabling secure quantum communications with photonic components and integrated circuits

Fraunhofer HHI is advancing quantum key distribution systems by developing photonic components such as single-photon detectors and transmitters based on InP, and creating hybrid PICs that can generate photon pairs and control their polarisation states.

By Nino Walenta, Patrick Runge and Moritz Kleinert, Fraunhofer HHI

In an era marked by growing concerns over cybersecurity and data privacy, quantum communication is emerging as a promising solution for secure information exchange. Although classical encryption methods have so far been robust, they are increasingly vulnerable to advances in computing – especially the anticipated capabilities of quantum computers.

Quantum communication systems aim to ensure information-theoretic security – meaning it cannot be broken even with unlimited computational resources and time. These systems leverage principles such as quantum entanglement, which links pairs of particles, and the no-cloning theorem, which prevents quantum information from being perfectly copied. Such principles can be seen at work in quantum key distribution (QKD), a well-known cryptographic task that enables two parties to share encryption keys with security rooted in the laws of physics.

However, realising practical quantum networks requires more than just theoretical foundations; it depends critically on scalable, stable, and high-performance photonic systems. Photons are ideal quantum information carriers, since they do not interact much with the environment and they are compatible with existing optical fibre networks. To support quantum communication, advanced photonic components such as low-loss waveguides, integrated sources of entangled photons, and high-speed single-photon detectors are vital.

Moreover, PICs enable the miniaturisation and mass production of quantum devices, paving the way for cost-effective deployment. This synergy between quantum principles and photonic technologies will form the backbone of next-generation communication infrastructure.

Figure 1. SPAD module with the key performance parameters measured at 1550 nm wavelength for room temperature and internally TEC cooled operation.

Photonic components
One of the main challenges in the quantum industry is detecting photons in the short-wave infrared (SWIR) wavelength range (1000-1600 nm). For shorter wavelengths up to 900 nm, single-photon avalanche diodes (SPADs) fabricated on silicon material platforms work well. However, when targeting the use of low-loss single-mode fibres for QKD applications, detectors must operate at telecom-wavelengths from the O- to L-band. In this case, indium phosphide (InP) is the material platform of choice.

For quantum applications requiring detectors that can pick up just a few photons, there are photodiodes available with high detection efficiencies of 99 percent [1]. This is sufficient for certain applications, such as those using squeezed light. However, other applications require yet higher-sensitivity detectors that can even catch individual photons with high accuracy.

To achieve this performance, systems need to incorporate either internal amplification to boost the detection signal, or superconductivity to reduce background noise from electrical resistance. A drawback of the latter is its need for cryogenic cooling, which is impractical for most non-academic settings.

SPADs are a preferred choice for industrial applications, thanks to their compact size, high reliability, and low power consumption. However, these devices suffer from dark counts, meaning they occasionally generate a false detection signal when no photon is present. Since dark count rates (DCR) can significantly impact the system performance, for example by reducing the signal-to-noise ratio and decreasing the secret key rate, the system design must consider this parameter. However, moderate cooling with a thermoelectric cooler (TEC) reduces the DCR significantly, meaning that the SPAD’s packaging becomes as important as the design of the chip itself.

Figure 1 shows a fibre-coupled SPAD module with a system application-oriented housing (fibre in-plane with printed circuit board (PCB)) and its measurement results [2]. Besides the excellent photon detection efficiency (PDE) and DCR, the module performs well on several other characteristics. It has an afterpulsing probability – or likelihood of generating a false secondary signal shortly after a genuine detection – of under 0.5 percent. This is measured at a hold-off time of 10 µs, during which it is intentionally switched off to “reset” after a detection.

Additionally, the timing jitter, or uncertainty about precisely when each photon triggers the detection signal, is just 162 picoseconds.

The SPAD module also has a compact footprint of less than a thumbnail and includes a TEC that can cool the device to 70 degrees C below the temperature of its environment. With this cooling, the module can achieve a DCR of 1500 counts per second and 25 percent PDE, which represents state-of-the-art performance at 1550 nm. Besides the SPAD design itself, the fabrication technology, including the high-quality epitaxial growth of the semiconductor layer stack, contributes to its excellent performance.

Since the SPAD operates by triggering an electrical current when it detects a photon, it also needs an inbuilt process to stop this current, “resetting” it so it is ready for the next photon. This is called quenching and there are multiple ways of achieving it.

Typically, QKD systems use one of two techniques: high-speed gating or free-running passive quenching. In the former setup, a voltage source sets the SPAD above the breakdown voltage – the threshold value at which the voltage is high enough to set the photodetector to single-photon detection mode. The voltage source repeatedly drops below this value and returns to it at a defined time interval, continually turning the SPAD off and resetting it.

Meanwhile, the other quenching technique sets the SPAD into a free-running passive quenching mode, in which the voltage source remains constantly above the breakdown voltage. If a detection event generates a photocurrent, an internal resistor reduces the voltage, which drops across the SPAD, until the current stops. The voltage then recovers so the SPAD can detect another photon. SPADs with internal resistors are called negative feedback avalanche diodes.

Figure 2. Thin-film filter as a polarising beam splitter/combiner (left) and as a pump light suppression filter (right).

On the transmitter side, InP-based PICs are excellent candidates for QKD signal generation, since they enable monolithic integration of lasers and electro-absorption modulators. Moreover, InP PICs are already a mature technology, thanks to well established component manufacturing for the telecom industry. With a few minor design changes, telecom PICs can be readily adapted for QKD transmitters.

Since QKD systems rely on single-photon interactions, the PIC design needs to include variable optical attenuation, to reduce the emitted light intensity. Additionally, to provide the random numbers that QKD systems use to ensure secure encryption, the design incorporates quantum random number generators built to introduce non-deterministic processes such as laser beating or laser switching.

Researchers have already demonstrated high-performance transmitter QKD components on InP PICs, showing the potential of this approach [3] [4]. As a cost-efficient low entry level for InP PIC development, engineers can use multi-project wafers to get experience with PIC design and characterisation.