Italian quantum backbone

1 Introduction

Optical fibers have been a disruptive innovation in many fields, in particular in telecommunications, opening the possibility of broadband and ultra-broadband communications, today exceeding transmission rate of 100 Gbit/s. This allows new services, like video streaming, and boosts scientific areas where there is the need of exchanging a large amount of data, like radioastronomy.

Here, we describe that fiber-optic infrastructures are a breakthrough not only in communications, but also for Time and Frequency (T/F) metrology and Quantum Technologies (QT).

From historical perspective, the first idea to use optical fibers to transfer T/F signals comes from the U.S., with the proposal of John Hall, Long Shen Ma and co-workers in 1994. The concept is simple: by transmitting an ultrastable laser radiation over a fiber, we remotely transfer information about its frequency. Measuring this frequency at the transmitting end with respect to a primary frequency standard, i.e. an atomic clock, we realize the dissemination of the primary frequency reference. Following the same argumentation, we can also broadcast time using fibers and lasers.

Today, the dissemination of T/F reference signals over fiber is the most performing solution up to distances of thousands of kilometres. Moreover, fiber distribution is today the only available method to transfer the accuracy of the new generation of atomic clocks, called optical frequency standards or optical clocks. Optical clocks now achieve an accuracy of 10^-18 in terms of relative frequency, but this accuracy cannot be disseminated to a remote user by the commonly used microwave satellite techniques like GPS and in general Global Navigation Satellite Systems, since their noise limits the transfer. In other words, satellite T/F dissemination is today quite common but does not allow a remote user to get the best available time and frequency reference.

T/F distribution over fiber is also faster, since it allows to achieve the target uncertainty with measurements that are at least 1000 times shorter than using satellites. In fact, satellite comparisons offer a best instability of 10^-10 at 1 second of measurement, while fiber-based dissemination has an instability of 10^-15 at 1 s over 1000 km, i.e. 5 orders of magnitude lower. This instability averages down with the inverse of the measurement time: optical techniques allow a comparison of optical clocks in few hours, whilst satellite techniques are not suited to this task. Figure 1 depicts the evolution of clocks accuracy and the possibility offered by the best comparison techniques.

If we consider the fiber-optic technique from the point of view of the National Metrological Institutes (NMIs), and the International Metre Convention, an effective comparison of remote optical clocks is a prerequisite for considering the redefinition of the second in the International System of units. In fact, today the definition of the unit of time is based on a quantum property of the Caesium atom, and it is realized by Caesium clocks. Optical clocks are based on different atoms, such as Ytterbium, Strontium, Indium, Mercury, Aluminium, etc., and currently outperform primary standards based on Caesium. A redefinition of the second based on optical clocks is a breakthrough in metrology and must be prepared by comparing different realizations in different laboratories to ensure the absence of major errors. Up to now, it was unpractical to think about a possible redefinition, as satellite techniques were limiting optical clocks comparisons. Today, fiber-optic dissemination overcomes the problem, and a rigorous roadmap towards a possible redefinition is possible.

The development of fiber-optic links for T/F is particularly intense in Europe and in Japan, even if seminal works were performed in the U.S. Pioneering activities are also being performed using free-space optical links. Albeit nowadays free-space links suffer from performance degradation over hauls longer than few tens of kilometres because of air turbulence, in the future they could cope for cases where a fiber-optic infrastructure is not present.

Optical clocks comparison for the redefinition of the second is not the only application of fiber-optic links. In the last decade, INRIM developed a fiber backbone in Italy, and starting from T/F distribution, presented seminal works in different fields of science. Some of the demonstrated technologies over fiber are now available for industrial cases.

In particular, we have used our ultra-accurate T/F distribution to study possible advances in radioastronomy, and in those geodesy techniques that are based on radioastronomical observations. Moreover, we enabled new measurements in atomic and molecular physics. Lastly, interesting applications concern quantum technologies research. Using our fiber quantum backbone we demonstrated the first proof of principle of relativistic geodesy, where we realized a gravity measurement by the comparison of two remote atomic clocks. We then focused our research on quantum simulation and quantum communication, in particular on the transmission of entanglement pairs and Quantum Key Distribution (QKD) experiments, in real field environment and with different techniques.

In addition, there is a relevant knowledge transfer activity from the scientific infrastructure to the industrial stakeholders. For T/F over fiber, the adoption of the White Rabbit-Precision Time Protocol (PTP-WR) method, powerful but still cost effective, offers high robustness, high integrity levels and a very accurate traceability to UTC ( IT ), with sub-nanosecond accuracy.

In this paper, we describe the quantum backbone infrastructure, the T/F experiments, and the scientific highlights achieved in recent experiments.

2 The development of the Italian Quantum Backbone

The Italian quantum fiber-optic backbone uses commercial fibers, and is deployed all along the peninsula for a total length of 1850 km. It connects INRIM’s premises in Turin to Milan (after about 280 km fiber haul), Bologna (550 km), Florence (640 km), Rome (1000 km), Naples and Pozzuoli (1300 km) and Matera (1685 km). From Turin, another 150 km fiber link crosses the Italian-French border, reaching Modane: from here we will connect to the rest of Europe, via Grenoble, Lyon and Paris. Another cross-border link will possibly be established towards Geneve via Milan. Those European connections are part of a pan-European network; with respect to primary metrology, this network will link together the National Metrology Institutes of UK, France, Germany, Poland, The Netherlands, and Switzerland. Moreover, INRIM’s quantum backbone can be linked to fiber infrastructures devoted to quantum technologies only, thanks to its architecture, which will be discussed hereafter. Figure 2 shows a map of the Italian Quantum Backbone and of the developments at the European level.

We realized, or are realizing, other connections with and to industrial partners, which presently exploit T/F distribution only. Partners include Telespazio, Thales Alenia Space Italia, Leonardo and the Consorzio TOP-IX. The latter offers a calibration service (certified time stamping, traceable to UTC) in collaboration with INRIM to the financial district in Milan.

The quantum backbone is based on a dedicated fiber pair (dark fibers); one fiber is dedicated to T/F dissemination and equipped with optical amplifiers; the other one is dedicated to quantum technologies requiring single-photon broadcasting. This fiber is not equipped with optical amplifiers, which would destroy the single-photon condition. Twenty-six intermediate housing locations are present along the backbone throughout the Country, offering a rack to install optical equipment. They are generally easy to reach, with 24/7 access to our staff.

Considering T/F atomic standards distribution, we use two different techniques. First, a coherent method with Doppler cancellation of the phase noise introduced by the fiber, achieving the highest performances. Second, a T/F transfer based on WR-PTP, with reduced performances but cost effective and easier installation and maintenance at the user premises. Presently, we have implemented the coherent frequency transfer on the whole 1830 km link, and we are upgrading the optical infrastructure to embed the WR-PTP, too.

Both techniques rely on the consideration that the fiber itself is a source of degradation for the stability and the accuracy of the distributed atomic standards. In both techniques, we transfer an optical carrier on the fiber. The fiber changes its optical length due to acoustic noise and temperature variations, introducing a time-varying delay which affects the phase of the transferred carrier. Such delay must be measured and compensated. We do this using a two-way approach in which the information about the noise or about the delay introduced by the fiber is continuously retrieved by sending back a signal from the remote end and assuming that the two paths (from INRIM to the user and back) are reciprocal. This assumption is justified since we use a single fiber for the two directions, which then share the same physical medium. This is in contrast to what happens in data transmission, where separate fibers are used.

In the coherent technique, we distribute a laser at 1542 nm that is ultrastable, i.e. its frequency does not change in time by more than $5 \times 10^{-15}$ in terms of relative frequency, with linewidths <10 Hz and long-term drifts around 0.1 Hz/s. The laser frequency is locked on the long term to the best atomic clocks present at INRIM: the Hydrogen Masers used to generate the international timescale UTC ( IT ) and/or the Cs fountain clock, the national primary frequency standard, with an accuracy of $2 \times 10^{-16}$, and/or the Ytterbium optical clock, with a present accuracy of $4 \times 10^{-17}$.

In this approach, the round-trip signal for the phase-noise detection and cancellation is obtained by reflecting back part of the radiation that arrives at the remote fiber end. Comparing at INRIM the original and the reflected signal, we detect the noise of the fiber itself, and we cancel it using an optoelectronic circuit. Custom bidirectional amplifiers are used to cope with fiber attenuation. Figure 3 shows a schematic of the noise cancellation concept.

The second method, the WR-PTP, is much easier and cost effective, but performances are 1000 times worst: they are still better than satellite transfer, and they are adequate for a lot of applications, but not to spread or compare optical clocks. White Rabbit technique was invented at CERN. It is an improvement of the PTP method, already well known in telecommunications. A simple telecom laser is impinged on a fiber and amplitude-modulated to carry a T/F information. At the end of the fiber, a device is synchronized with the code carried by the laser, and transmits a T/F signal back to the origin. After receiving the remote signal, the optoelectronic terminal at the starting point combines digital phase measurement and an algorithm to calculate the delay introduced by the fiber and compensate it.

This ensures a relevant stability and accuracy to the time distribution.

3 Quantum technologies

The quantum backbone is relevant to the implementation of in-field quantum technologies in two different ways. First, it can be a classical tool to enable quantum measurements. Second, the backbone can directly carry quantum light, i.e. single photons or light quanta.

Examples of the first kind of applications are the demonstration of relativistic geodesy by atomic clocks and the novel possibilities offered in quantum simulation experiments.

In 2018, we published the results of the first in-field demonstration of relativist geodesy using atomic clocks. Because of Einstein’s theory of general relativity, two clocks placed at different gravity potentials, are affected by a frequency offset. This is known as the gravitational redshift, and all the primary atomic clocks in the world which contribute to the Universal Time are corrected for this bias.

Having very accurate clocks in different locations, and a means to compare them in real-time, we can reverse the paradigm, and use the clocks as quantum sensors for gravity differences between two locations.

We coordinated an international effort involving INRIM and colleagues of the NMIs in Germany and UK, the University of Hannover and the Underground Laboratory of Modane to detect the gravity difference in real time using two atomic clocks. Within a European funded project, INRIM realised a coherent fiber link between its clocks (Cs fountain and Yb optical clock) located in Turin, at 240 m height on the Geoid, and a transportable Strontium lattice clock, realized by the German colleagues and placed inside the Frejus tunnel, on the Italy-France border, 1000 m higher than Turin. We measured the gravity potential difference between the two locations with traditional techniques, and in those conditions General Relativity predicts a relative frequency difference of about $1\times 10^{-13}$ between the clocks. This frequency difference was independently measured using the fiber link and was in agreement with the predicted results. This demonstrated that General Relativity and quantum clocks can be used to measure the gravity potential difference instead of the traditional techniques. This paves the way for new campaigns devoted to the computation of a better map of gravity potential and possibly monitor its stability in time, exploiting the network of clocks/fiber links. In fact, clock measurements of the gravity field difference can be much more accurate, faster and with higher spatial resolution, whilst the adequate accuracy offered by present techniques comes at the cost of averaging on large areas (tens of km²) and longer measurement times.

A second experiment relying on the use of the fiber link as a “classical tool” for quantum technologies exploited the T/F connection between Turin and Florence. At the University of Florence, CNR and LENS laboratories, an ultra-cold Ytterbium experiment has been developed to study tiny quantum effects, in particular, the spin orbit coupling. This experiment is trying to use an ultra-cold ¹⁷³Yb quantum gas as a quantum simulator of complex quantum systems like magnetic states of matter, in a more controlled way. INRIM set a metrological chain composed of the Cs fountain clock, the optical link and an optical terminal in Florence. With this chain, we directly compared the frequency of the laser at 578 nm used in the local Yb experiment to the Cs fountain in Turin. First, with respect to previous values, we improved by 4 orders of magnitudes the accuracy of the 578 nm transition of ¹⁷³Yb (uncertainty $2\times 10^{-14}$ in relative frequency). Figure 4 shows the typical result of these measurements, where we can clearly appreciate the difference having a GPS as a frequency reference and the fiber link.

Then, INRIM’s set-up enabled the study of a component of the quantum simulator, i.e. the Spin-Orbit Coupling on the cold Yb sample.

Another class of experiments we started to investigate using the quantum backbone infrastructure is related to the transmission of real “quantum” light. In particular, we are working, in collaboration with CNR, on entangled photons and Quantum Key Distribution (QKD) in real-field conditions.

QKD is a relevant outer reach of quantum technologies for different reasons. First, it is related to quantum communication and is a physical method to implement a quantum layer on cryptography, hence on secure communications. This quantum layer is seen as one of the two possible answers to the threats in cyber security coming from the realization of quantum computers. The two possible envisaged answers to this innovation are an algorithmic enforcement (the so-called post-quantum cryptography), or the physical application of quantum principles of measurement to the distribution of crypto graphical keys. This latter is known as QKD.

Probably, the two possibilities will be complementary. Another point that makes QKD a relevant outer reach is the availability of some commercial solutions: hence, QKD is seen as the first quantum technology mature enough to be implemented in the following years in real applications. Nonetheless, there are still some challenges. Attenuation is a major issue in view of long-haul transmission. Optical amplifiers cannot be used, since QKD relies on single photons, and amplifiers will break the quantum condition; on the other hand, quantum repeaters are yet to come.

Other major challenges are the sharing of the fibers between QKD and data traffic; the study of attacks and their detection in field; the standardisation of techniques and devices.

INRIM is studying those issues in laboratory since many years, now we started to look at the challenges of real field using our quantum backbone.

The first published results came from an international collaboration with Austria and Malta, regarding the possibility of distributing entangled photons over 100 km of fibers in a highly stable environment, that of a submarine fiber. This experiment also offered a unique opportunity to compare the challenges of fibers on land with a privileged, yet real environment, the fiber in submarine cables.

4 Submarine experiments

When implementing a quantum backbone, there are three infrastructural challenges. The first is the compatibility with data traffic. For some applications, scalability matters, and this can be achieved only if the application (quantum technology or classical new service) is compatible with data traffic, since in this case economical synergies are relevant. The second challenge is attenuation: for some applications amplification is allowed but, when working with single photons, we cannot use common amplification. The last challenge is phase noise: when transmitting coherent ultrastable radiation, the mechanical noise on the fiber (vibrations, temperature fluctuations) reduces performances.

We studied a particularly quiet environment, yet in real field and in real applications: submarine fiber-optic cables, using two testbeds, between the islands of Sicily and Malta, each about 100 km long, 200 m below sea level.

Three experiments were performed, investigating different aspects. We already mentioned the entanglemnt distribution, but we have also tested the distribution of T/F signals. In particular, we investigated the feasibility of ultra-long coherent fiber links for transcontinental optical clocks comparisons. In transoceanic fiber links, the architecture does not allow the optimal cancellation of phase noise as it is not possible to bypass standard optical amplifiers. As a result, T/F distribution must rely on separate fibers for the two transmitting directions, reducing the reciprocity of the optical path and the amount of rejected noise. On the other hand, submarine fibers are exposed to much lower noise than terrestrial fibers. Our results show that the noise is even 4 orders of magnitude smaller (fig. 5), and demonstrate the feasibility of a transcontinental atomic clock comparison using submarine fibers with higher accuracy than achievable today with satellites. From this, the importance is clear of setting up links between Europe, America and Asia in view of optical clocks comparisons that are necessary for the possible redefinition of the second.

The last experiment on submarine cables investigated seismology applications. The basic concept reverses the paradigm of phase noise cancellation. A seismic event shakes the fiber and this vibration is imprinted on the phase of the ultra-stable laser. Instead of cancelling it, as we usually do with environmental noise on the fiber, we exploited it to detect earthquakes. We detected earthquakes with magnitudes as small as 3.4 and demonstrated an adequate sensitivity, both on close and tele-seismic events with signals comparable to those of classical seismometers. The complete analysis of the signals, and the absolute determination of magnitude are still under investigation; on the other hand, the localization of the seismic event appeared to be clearly feasible. The main potential of this technique relies on the possibility of building a submarine network for earthquake detection: in fact, currently a network of classical seismometers on the sea-floor does not exist. Detection is today mostly performed by seismometers on land, with the consequent delay in detection and the impossibility of detecting small events. Now, we can imagine using the network of submarine cables for data traffic to develop a network for seismic detection, with a sustainable cost.

5 Radioastronomy and geodesy

In Bologna and Matera, INRIM’s quantum backbone is used to study possible improvements in space geodesy and VLBI radioastronomy. Those techniques are based on the correlation of measurements by a network of distant radio antennas equipped with high-quality local clocks. Using fibers it is possible to replace the local frequency reference with better, fiber-disseminated standards. This also allows the establishment of a common clock architecture to multiple antennas, with the full rejection of clocks’ offsets and noise. We demonstrated pioneering results in European VLBI campaigns involving the Medicina Radiotelescope of the Italian Astrophysical Institute (INAF); now, we reached also the radio-antenna at the Space Geodesy Centre of the Italian Space Agency (ASI), in Matera, and first observations are ongoing.

The geodetic VLBI technique is highly demanding in terms of reliability of the frequency transfer. This required improving the remote controlling capabilities of all the components of the quantum backbone, and now we have achieved an uptime > 90-95%. We consider this a relevant result for an infrastructure at this stage of development, and further improvements are feasible.

6 Conclusions

The Italian Quantum Backbone offers today an infrastructure with unique features. Connecting several important research laboratories in the Country, the link sets the basis for innovative experiments, and paves the way for new services. Time and Frequency Distribution is already a recognized reality over this infrastructure; we demonstrated the first results with quantum technologies, and certainly more has to come. The implementation of a European fiber Network for Time and Frequency is advancing and the possibility of comparing remote optical clocks is now possible at least in our continent. This will be relevant to the roadmap towards a possible redefinition of the Unit of Time in the International System.

The possibility to boost a Quantum Technology Network based on fibers is now a concrete possibility, which deserves serious consideration.

Acknowledgments

The Italian Quantum Backbone is a project of the Istituto Nazionale di Ricerca Metrologica INRIM. The authors acknowledge the invaluable role in its development of Alberto Mura, Filippo Levi, Claudio Calosso, Elio Bertacco from INRIM and Giovanni Antonio Costanzo from Politecnico di Torino. The link development enabled fruitful collaborations with several laboratories of INAF, CNR-INO, ASI, LENS, Università di Firenze, NPL and University of Malta and we would like to thank all the colleagues of those institutes: together we have obtained interesting scientific results, documented in several papers on international journals. The extension from Florence to Matera has been strongly encouraged by Massimo Inguscio, former president of INRIM, now President of CNR. On the extension, we acknowledge the relevant role of ASI, in particular the input from Roberto Battiston, Alberto Tuozzi, and Giuseppe Bianco. We thank Consortium GARR and Consortium TOP-IX for their assistance in providing the optical fibers, and Consortium TOP-IX for the collaboration to establish a timing service to the Milan finance district.

The authors acknowledge for funding: the project 17IND14 WRITE, that has received funding from the EMPIR programme co-financed by the Participating States and from the European Union’s Horizon 2020 research and innovation programme; the CLONETS project, that receives funding from the European Union’s Horizon 2020 research and innovation programme (2014-2020) under grant agreement no. 731107; ASI funding within the DTF-Matera project; the Italian Ministry of Research MIUR (projects LIFT, and Metgesp).