Abstract Linear quantum optics is advancing quickly, driven by sources of correlated photon pairs. Multi-photon sources beyond pairs would be a powerful resource, but are a difficult technology to implement. We have discovered a way in which we can combine multiple pair-sources to act analogous to sources of four, six or even eight correlated photons for the creation of highly entangled quantum states and other quantum information tasks. The existence of such setups is interesting from a conceptual perspective, but also offers a useful abstraction for the construction of more complicated photonic experiments, ranging from state generation to complex quantum networks. We show that even just going from probabilistic two-photon sources to effective four-photon sources allows conceptually new experiments for which no other building principles were known before. The setups which inspired the formulation of these abstract building blocks were discovered by a computer algorithm that can efficiently design quantum optics experiments. Our manuscript demonstrates how artificial intelligence can act as a source of inspiration for the scientific discoveries of new ideas and concepts in physics.
Abstract Linear quantum optics is advancing quickly, driven by sources of correlated photon pairs. Multi-photon sources beyond pairs would be a powerful resource, but are a difficult technology to implement. We have discovered a way in which we can combine multiple pair-sources to act analogous to sources of four, six or even eight correlated photons for the creation of highly entangled quantum states and other quantum information tasks. The existence of such setups is interesting from a conceptual perspective, but also offers a useful abstraction for the construction of more complicated photonic experiments, ranging from state generation to complex quantum networks. We show that even just going from probabilistic two-photon sources to effective four-photon sources allows conceptually new experiments for which no other building principles were known before. The setups which inspired the formulation of these abstract building blocks were discovered by a computer algorithm that can efficiently design quantum optics experiments. Our manuscript demonstrates how artificial intelligence can act as a source of inspiration for the scientific discoveries of new ideas and concepts in physics.
In the emerging field of quantum information technology the two basic subfields are quantum communication and quantum computation. Photonic qubits are considered as most promising information carriers for this new technology due to the immense advantage of suffering negligible decoherence. Additionally, the very small photon-photon interactions can be replaced by inducing effective nonlinearities via measurements which allow for the implementation of crucial two-qubit gate operations. Although the spontaneous parametric down-conversion gives access to the generation of highly entangled few-photon states, such as four-qubit cluster states which allow to demonstrate the new concept of the one-way quantum computer, its applicability is highly limited due to the poor scaling of the simultaneous emission of more than one-entangled photon pair. Therefore of particular interest is the reversible mapping of qubits from photon states to atomic states. This might allow the implementation of photonic quantum repeaters for long-distance quantum communication or the generation of arbitrary multi-photon states as required for linear-optics quantum computing. Thus for the realization of such a quantum network several approaches for achieving the required quantum control between matter and photons have been studied during the past few years. Recent experiments demonstrating the generation of narrow-bandwidth single photons using a room-temperature ensemble of 87Rb atoms and electromagnetically induced transparency should emphasize the progress towards such a quantum network.
We introduce the concept of hypergraphs to describe quantum optical experiments with probabilistic multiphoton sources. Every hyperedge represents a correlated photon source, and every vertex stands for an optical output path. Such a general graph description provides new insights for producing complex high-dimensional multiphoton quantum entangled states, which go beyond limitations imposed by pair creation via spontaneous parametric down-conversion. Furthermore, the properties of hypergraphs can be investigated experimentally. For example, the NP-complete problem of deciding whether a hypergraph has a perfect matching can be answered by experimentally detecting multiphoton events in quantum experiments. By introducing complex weights in hypergraphs, we show a general many-particle quantum interference and manipulating entanglement in a pictorial way. Our work paves the path for the development of multiphoton high-dimensional state generation and might inspire new applications of quantum computations using hypergraph mappings.
Advances in photonics will lead to breakthroughs in technology with a similar impact to that resulted from the development of electronics. At the quantum level, photons exhibit some unique features, such as non-classical correlations and bosonic bunching. The manipulation of these quanta of light foresee the construction of quantum devices capable of outperforming their classical counterparts in the purposes for which they are built. Despite a significant progress in developing quantum technologies, several unsolved issues certainly remain: Efficient quantification of quantum correlations in a system, and manipulation of multiple photons at once, are two of such hurdles to be overcome. In this thesis, we tackle these two issues and present experimental progress towards their solution. We first explore quantum entanglement between two single-photons and its effect on phases appearing on the system's wavefunction. The observed phases revealed an intimate connection between entanglement––a purely quantum phenomena––and geometric and topological structures of abstract manifolds: A special case of geometric phases was found to increase monotonically with the amount of entanglement in the system, thus serving as an alternative measure of quantum correlations. We then employed a system of three single-photons, and two concatenated entangling gates, as the basis of a quantum simulation protocol: A three-particle quantum simulator, mimicking the entangling evolution of two particles, is used to experimentally demonstrate that the overhead in measuring concurrence––a measure of two-particle entanglement––can be significantly reduced down to measuring only two observables. At this point, increasing the number of photons being handled proved extremely difficult; the low efficiency of photon sources based on parametric downconversion––the gold-standard and only source at-the-time used for multi-photon experiments––imposed a hard limit on countrates and hence the number of photons to be manipulated. We thus then focused on overcoming this obstacle: We studied a solid-state emitter and its performance as a multi-photon source. We employed a quantum dot-cavity system to show that streams of tens of indistinguishable single-photons are produced, whose temporal-to-spatial demultiplexing serves as a novel kind of multi-photon source. A three-photon source was then built and employed in the first demonstration of multi-photon interference from a solid-state source: We implemented a BosonSampling device with a source that is between one and two orders-of-magnitude more efficient than its downconversion counterpart. The main limiting factors for scaling to higher photon numbers were identified to be not intrinsic to the solid-state emitter, but determined by low detector efficiencies and the probabilistic nature of the demultiplexing scheme. Future improvements on these limitations are expected to allow the generation of sources with a higher number of single-photons, a task that has to date remained otherwise impossible.
Entangled photons are at the heart of experimental quantum physics. They were used for the first fundamental tests of quantum theory, and became a basic building block for many novel quantum protocols, such as quantum cryptography, dense coding or teleportation. Therefore, the efficient generation of entangled photons, as well as their distribution and accurate analysis are of paramount importance, particularly with regard to the practicability of many applications of quantum communication. This thesis deals largely with the problem of efficient generation of photonic entanglement with the principal aim of developing a bright source of polarization-entangled photon pairs, which meets the requirements for reliable and economic operation of quantum communication prototypes and demonstrators. Our approach uses a correlated photon-pair emission in nonlinear process of spontaneous parametric down-conversion pumped by light coming from a compact and cheap blue laser diode. Two alternative source configurations are examined within the thesis. The first makes use of a well established concept of degenerate non-collinear emission from a single type-II nonlinear crystal and the second relies on a novel method where the emissions from two adjacent type-I phase-matched nonlinear crystals operated in collinear non-degenerate regime are coherently overlapped. The latter approach showed to be more effective, yielding a total detected rate of almost 10^6 pairs/s at >98 % quantum interference visibility of polarization correlations. This performance, together with the almost free of alignment operation of the system, suggest that it is an especially promising candidate for many future practical applications, including quantum cryptography, detector calibration or use in undergraduate lab courses. The second issue addressed within the thesis is the simplification and practical implementation of quantum-assisted solutions to multiparty communication tasks. While the recent rapid progress in the development of bright entangled photon-pair sources has been followed with ample experimental reports on two-party quantum communication tasks, the practical implementations of tasks for more than two parties have been held back, so far. This is mainly due to the requirement of multiparty entangled states, which are very difficult to be produced with current methods and moreover suffer from a high noise. We show that entanglement is not the only non-classical resource endowing the quantum multiparty information processing its power. Instead, only the sequential communication and transformation of a single qubit can be sufficient to accomplish certain tasks. This we prove for two distinct communication tasks, secret sharing and communication complexity. Whereas the goal of the first is to split a cryptographic key among several parties in a way that its reconstruction requires their collaboration, the latter aims at reducing the amount of communication during distributed computational tasks. Importantly, our qubit-assisted solutions to the problems are feasible with state-of-the-art technology. This we clearly demonstrate in the laboratory implementation for 6 and 5 parties, respectively, which is to the best of our knowledge the highest number of actively performing parties in a quantum protocol ever implemented. Thus, by successfully solving and implementing a cryptographic task as well as a task originating in computer science, we clearly illustrate the potential to introduce multiparty communication problems into real life.
Major efforts in quantum information science are devoted to the development of methods that are superior to the one of classical information processing, for example the quantum computer or quantum simulations. For these purposes, superposition and entangled states are considered a decisive resource. Furthermore, since the early days of quantum mechanics, entanglement has revealed the discrepancy between the quantum mechanical and the everyday life perception of the physical world. This combination of fundamental science and application-oriented research makes the realization, characterization, and application of entanglement a challenge pursued by many researchers. In this work, the observation of entangled states of polarization encoded photonic qubits is pushed forward in two directions: flexibility in state observation and increase in photon rate. To achieve flexibility two different schemes are developed: setup-based and entanglement-based observation of inequivalent multi-photon states. The setup-based scheme relies on multi-photon interference at a polarizing beam splitter with prior polarization manipulations. It allows the observation of a family of important four-photon entangled states. The entanglement-based scheme exploits the rich properties of Dicke states under particle projections or loss in order to obtain inequivalent multi-photon entangled states. The observed states are characterized using the fidelity and entanglement witnesses. An increase in photon rate is crucial to achieve entanglement of higher photon numbers. This holds especially, when photon sources are utilized that emit photons spontaneously. To this end, a new photon source is presented based on a femtosecond ultraviolet enhancement cavity and applied to the observation of the six-photon Dicke state with three excitations. The implemented schemes not only allow the observation of inequivalent types of entanglement, but also the realization of various quantum information tasks. In this work, the four-photon GHZ state has been used in a quantum simulation of a minimal instance of the toric code. This code enables the demonstration of basic properties of anyons, which are quasiparticles distinct from bosons and fermions. Further, the six-photon Dicke state has been applied for quantum metrology: It allows one to estimate a phase shift with a higher precision than by using only classical resources. Altogether, a whole series of experiments for observing inequivalent multi-photon entangled states can now be substituted by a single experimental setup based on the designs developed in this work. In addition to this new approach of photon processing, a novel photon source has been implemented, paving the way to realizations of applications requiring higher photon numbers.
Two important challenges in quantum photonics are to generate useful states with high fidelity, and to detect them and verify their properties. Particularly valuable states are single photons and entangled photon pairs in well-defined optical modes, as they can be used in many quantum information protocols or used to build up more complex states. For sources, we employ integrated nonlinear optics (waveguides in lithium niobite and potassium titanyl phosphate) to maximize brightness and go beyond what is possible in bulk optics, showing simultaneously high state fidelity, heralding efficiency, and spectral purity across three experiments: first we show record heralding efficiency in a fully-fibered heralded single-photon source, and use it to probe the tradeoff between spectral purity and heralding efficiency in non-engineered sources. With an engineered source, we then herald up to 50 photons in a nonclassical state. The last source is for polarization-entangled photon pairs, with brightness of 3.5 million pairs/s·mW, fidelity to a Bell state of 96%, heralding efficiency of 43%, and HOM interference visibility of 82%. Once a complex state is constructed, it must also be verified. For this we employ a time-multiplexed detector consisting of a fibre loop and a single-photon detector. Surprisingly, we are able to extract information even in the saturation regime of the detector. We use the click statistics of the time-multiplexed detector to verify the non-classicality of quantum light, and we use its extremely high dynamic range (123 dB) to measure a macroscopic power level with a single-photon detector. Eliminating calibrated attenuators with this approach will allow direct standardization of quantum and classical optical power levels.