We demonstrate the difference between local, single-particle dynamics and global dynamics of entangled quantum systems coupled to independent environments. Using an all-optical experimental setup, we showed that, even when the environment-induced decay of each system is asymptotic, quantum entanglement may suddenly disappear. This "sudden death" constitutes yet another distinct and counterintuitive trait of entanglement.

Steering is a form of quantum nonlocality that is intimately related to the famous Einstein-Podolsky-Rosen (EPR) paradox that ignited the ongoing discussion of quantum correlations. Within the hierarchy of nonlocal correlations appearing in nature, EPR steering occupies an intermediate position between Bell nonlocality and entanglement. In continuous variable systems, EPR steering correlations have been observed by violation of Reid's EPR inequality, which is based on inferred variances of complementary observables. Here we propose and experimentally test a new criterion based on entropy functions, and show that it is more powerful than the variance inequality for identifying EPR steering. Using the entropic criterion our experimental results show EPR steering, while the variance criterion does not. Our results open up the possibility of observing this type of nonlocality in a wider variety of quantum states.

We report on an experimental investigation of the dynamics of entanglement between a single qubit and its environment, as well as for pairs of qubits interacting independently with individual environments, using photons obtained from parametric down-conversion. The qubits are encoded in the polarizations of single photons, while the interaction with the environment is implemented by coupling the polarization of each photon with its momentum. A convenient Sagnac interferometer allows for the implementation of several decoherence channels and for the continuous monitoring of the environment. For an initially-entangled photon pair, one observes the vanishing of entanglement before coherence disappears. For a single qubit interacting with an environment, the dynamics of complementarity relations connecting single-qubit properties and its entanglement with the environment is experimentally determined. The evolution of a single qubit under continuous monitoring of the environment is investigated, demonstrating that a qubit may decay even when the environment is found in the unexcited state. This implies that entanglement can be increased by local continuous monitoring, which is equivalent to entanglement distillation. We also present a detailed analysis of the transfer of entanglement from the two-qubit system to the two corresponding environments, between which entanglement may suddenly appear, and show instances for which no entanglement is created between dephasing environments, nor between each of them and the corresponding qubit: the initial two-qubit entanglement gets transformed into legitimate multiqubit entanglement of the Greenberger-Horne-Zeilinger (GHZ) type.

We derive several entanglement criteria for bipartite continuous variable quantum systems based on the Shannon entropy. These criteria are more sensitive than those involving only second-order moments, and are equivalent to well-known variance product tests in the case of Gaussian states. Furthermore, they involve only a pair of quadrature measurements, and will thus prove extremely useful in the experimental identification of entanglement.

When separated measurements on entangled quantum systems are performed, the theory predicts correlations that cannot be explained by any classical mechanism: communication is excluded because the signal should travel faster than light; preestablished agreement is excluded because Bell inequalities are violated. All optical demonstrations of such violations have involved discrete degrees of freedom and are plagued by the detectionefficiency loophole. A promising alternative is to use continuous variables combined with highly efficient homodyne measurements. However, all the schemes proposed so far use states or measurements that are extremely difficult to achieve, or they produce very weak violations. We present a simple method to generate large violations for feasible states using both photon counting and homodyne detections. The present scheme can also be used to obtain nonlocality from easy-to-prepare Gaussian states (e.g., two-mode squeezed state).

Most of the attention given to continuous variable systems for quantum information processing has traditionally been focused on Gaussian states. However, non-Gaussianity is an essential requirement for universal quantum computation and entanglement distillation, and can improve the efficiency of other quantum information tasks. Here we report the experimental observation of genuine non-Gaussian entanglement using spatially entangled photon pairs. The quantum correlations are invisible to all second-order tests, which identify only Gaussian entanglement, and are revealed only under application of a higher-order entanglement criterion. Thus, the photons exhibit a variety of entanglement that cannot be reproduced by Gaussian states.continuous variables | quantum information | non-Gaussian states | transverse spatial modes S ince the realization that quantum systems could prove useful for the efficient processing of information, there have appeared many proposals designed to exploit their power. From database searching and factoring algorithms to secure communication, these protocols achieve an impressive range of tasks. Typically, these quantum protocols have been first designed for discrete variable systems, described by finite dimensional Hilbert spaces, and hence are mathematically simpler than continuous variable (CV) systems with infinite dimensional spaces (1). CV systems provide a very interesting ground for quantum information processing, precisely due to their large and rich Hilbert space structure, which, however, makes characterization and detection of CV entanglement a difficult task.One way to tackle the difficulty in dealing with infinite dimensions is to confine the allowed states to a sector of Hilbert space defined by a finite number of parameters. Among the infinite ways one could do this, the most common is to restrict oneself to Gaussian states. For zero mean values, these states can be completely characterized by their covariance matrix containing moments up to second order (e.g. x 2 , xp , etc.) (1). This simplification has allowed for many results from the realm of finite-dimensional discrete-variable entanglement to find their CV analog (2, 3).It has recently become apparent, however, that the use of non-Gaussian states and operations is not only advantageous but at times necessary to successfully perform quantum information tasks. Non-Gaussian operations are required for entanglement distillation or swapping (4-8), and their use along with non-Gaussian states has been shown to improve quantum teleportation (9, 10) and cloning (11). It is also particularly remarkable that non-Gaussianity (either states or operations) is necessary for universal quantum computation with CVs (12), and proof of quantum nonlocality through violation of Bell's inequality (13). These advantages have motivated the experimental de-Gaussification of Gaussian states (14-17) and the idea that non-Gaussianity is a resource that can be quantified (18).The most popular criteria for identifying CV entanglement deal with mom...

We investigate nonlocality distillation using measures of nonlocality based on the Elitzur-Popescu-Rohrlich decomposition. For a certain number of copies of a given nonlocal box, we define two quantities of interest: (i) the nonlocal cost and (ii) the distillable nonlocality. We find that there exist boxes whose distillable nonlocality is strictly smaller than their nonlocal cost. Thus nonlocality displays a form of irreversibility which we term "bound nonlocality." Finally, we show that nonlocal distillability can be activated.

We derive analytical upper bounds for the entanglement of generalized Greenberger-Horne-Zeilinger states coupled to locally depolarizing and dephasing environments, and for local thermal baths of arbitrary temperature. These bounds apply for any convex quantifier of entanglement, and exponential entanglement decay with the number of constituent particles is found. The bounds are tight for depolarizing and dephasing channels. We also show that randomly generated initial states tend to violate these bounds, and that this discrepancy grows with the number of particles.