Optical physics

A larger quantum alphabet

Quantum objects that can hold more information than quantum bits have been generated and manipulated in an integrated photonic platform, paving the way for advanced protocols in quantum information processing. See Letter p.622

Quantum technologies are expected to introduce revolutionary changes in the ways in which information is processed in the near future. Furthermore, worldwide funding campaigns, such as the Quantum Technologies Flagship of the European Commission, announced last year (see go.nature.com/2mpp4oa), have been set up with the aim of boosting applications that encompass quantum communications, computing and sensing. However, it has yet to be demonstrated experimentally that a quantum technology is appreciably better than its classical counterpart. The main difficulty being faced is the creation and handling of many quantum bits (qubits), which limits the amount of information that can be processed by quantum means. On page 622, Kues et al.1 demonstrate an approach in which photons that have many potential frequencies are generated. This extends the usual two-level encoding capacity of qubits, increasing the amount of information that can be stored in the same number of quantum objects.

In the race to develop a mature quantum technology that has real-world applications, there are many competing platforms, including those that use photons, trapped ions and superconducting circuits2. Quantum photonics is eminently suited to applications in communications and sensing, but it has also proved valuable for quantum computing and simulation3. The photonic platform has become even more valuable in the past few years with the introduction of devices based on photonic integrated circuits, which can be used to generate and manipulate non-classical states of light on a microchip4. This advance has facilitated the impressive miniaturization, cost-effectiveness, scalability and stability of quantum photonic devices.

In Kues and colleagues' approach, pulses of light from a 'pump' laser are converted into pairs of photons using a microresonator — a ring-shaped waveguide, contained in a microchip, that traps light at certain frequencies known as resonances (Fig. 1a). Thanks to a nonlinear optical process called spontaneous four-wave mixing, two photons at the pump frequency are annihilated to create two photons whose frequencies are distributed over the multiple resonance frequencies of the microresonator. The optical spectrum of the resulting photons is a set of evenly spaced frequency lines, known as a frequency comb (Fig. 1b). Crucially, each resonance frequency at which a photon can be generated represents a different letter in an alphabet that extends the 'zero' and 'one' encoding of qubits, enabling multilevel encoding on each photon.

Figure 1: Generation of entangled photons for the multilevel encoding of information.

a, Kues et al.1 demonstrate an approach in which pulses of light from a 'pump' laser are converted into pairs of photons using a microresonator — an optical device, contained in a microchip, that traps light at certain frequencies known as resonances. The red arrows indicate the direction of light propagation through the device. b, Because the microresonator has multiple resonances, the optical spectrum of the resulting photons is seen as a set of evenly spaced frequency lines (green and blue peaks). The authors demonstrate that each of these frequencies can be used to encode information, extending the usual binary alphabet ('zero' and 'one') of quantum bits. Energy conservation implies that the frequencies of the two photons are entangled (correlated in a non-classical way), which imposes spectral symmetry on the photons with respect to the frequency of the pump laser (red peak). The combination of multilevel encoding and entanglement in such an integrated photonic platform is a key requirement for advanced quantum information processing.

As an added benefit, the authors' approach automatically generates pairs of photons whose frequencies are entangled (correlated in a non-classical way) because of the energy conservation of the nonlinear production process. Entanglement is a key requirement for many quantum protocols — from quantum precision measurements5 to quantum communications6 — and was demonstrated by the authors after they propagated their photons through a 24.2-kilometre-long telecommunications fibre.

Previous studies have exploited the paths7 and timing8 of photons, rather than their frequency, to generate multilevel entangled states. However, using the resonance-frequency approach allows Kues et al. to apply simplified circuits and operations — a key aspect of the authors' work is that manipulation of the multilevel-encoded photons is performed by commercial, off-the-shelf telecommunication components. Although they were not designed for quantum operations, it is surprising to see how well these standard devices can operate on quantum-encoded photons, without perturbing the photons' quantum state. As well as their practicality, the components can be easily reconfigured, enabling the implementation of a range of quantum protocols using the same experimental set-up.

One important limitation of integrated quantum photonic platforms is the presence of photon losses. Such losses limit quantum communication through fibres across long distances (hundreds of kilometres), and are a considerable problem in integrated photonic chips — current technologies produce devices that lose a few per cent of the propagating photons for every centimetre travelled. If each photon represents a qubit, and many qubits are used to encode some information, the information is preserved only if none of the photons are lost — something that becomes exponentially more unlikely as the number of qubits required increases. In this respect, Kues and colleagues' use of multilevel encoding, rather than the two-level encoding of qubits, is a smart choice because the same information can be transferred using fewer carriers, which reduces the impact of photon losses.

Another limitation of many quantum photonic systems, including that of Kues et al., concerns the use of a spontaneous nonlinear process to produce the photons used as carriers of the quantum information. This intrinsically probabilistic process causes practical difficulties when synchronizing the operations that comprise a quantum protocol. In addition, the limited rate at which photons are generated, and the low probability that different devices will emit single photons at the same time, make the scaling of these systems rather unfavourable. A promising alternative involves the use of structures called quantum dots9, whose atom-like behaviour enables single photons to be generated on demand and at a relatively high rate. However, the efficiency at which these photons are collected still needs to be improved, and the creation of entanglement between photons is much more complicated than in the nonlinear production processes.

The future success of quantum technologies, in terms of outperforming their classical counterparts, will be crucially dependent on the amount of information that can be processed by a quantum platform. Kues and colleagues estimate that their multilevel encoding of two photons could achieve the equivalent of 13 qubits — still far from the hundreds or more that are required to demonstrate the potential of quantum technologies. This limitation is not specific to quantum photonics, but is common to all such current platforms. The next challenge will be to demonstrate a technical breakthrough in one of these platforms or to combine them in a synergistic way to harness the advantages that each can provide. Footnote 1


  1. 1.

    See all news & views


  1. 1

    Kues, M. et al. Nature 546, 622–626 (2017).

    ADS  CAS  Article  Google Scholar 

  2. 2

    Ladd, T. D. et al. Nature 464, 45–53 (2010).

    ADS  CAS  Article  Google Scholar 

  3. 3

    O'Brien, J. L., Furusawa, A. & Vučković, J. Nature Photon. 3, 687–695 (2009).

    ADS  CAS  Article  Google Scholar 

  4. 4

    Tanzilli, S. et al. Laser Photon. Rev. 6, 115–143 (2012).

    ADS  Article  Google Scholar 

  5. 5

    Giovannetti, V., Lloyd, S. & Maccone, L. Science 306, 1330–1336 (2004).

    ADS  CAS  Article  Google Scholar 

  6. 6

    Ursin, R. et al. Nature Phys. 3, 481–486 (2007).

    ADS  CAS  Article  Google Scholar 

  7. 7

    Schaeff, C., Polster, R., Huber, M., Ramelow, S. & Zeilinger, A. Optica 2, 523–529 (2015).

    ADS  Article  Google Scholar 

  8. 8

    Thew, R., Acin, A., Zbinden, H. & Gisin, N. Quantum Inf. Comput. 4, 93–101 (2004).

    MathSciNet  Google Scholar 

  9. 9

    Somaschi, N. et al. Nature Photon. 10, 340–345 (2016).

    ADS  CAS  Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Roberto Osellame.

Related links

Related links

Related links in Nature Research

Optical physics: One ring to multiplex them all

Atomic physics: A milestone in quantum computing

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Osellame, R. A larger quantum alphabet. Nature 546, 602–603 (2017). https://doi.org/10.1038/546602a

Download citation


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing