Quantum key distribution (QKD) allows two remote parties to grow a shared secret key. Its security is founded on the principles of quantum mechanics, but in reality it significantly relies on the physical implementation. Technological imperfections of QKD systems have been previously explored, but no attack on an established QKD connection has been realized so far. Here we show the first full-field implementation of a complete attack on a running QKD connection. An installed eavesdropper obtains the entire 'secret' key, while none of the parameters monitored by the legitimate parties indicate a security breach. This confirms that non-idealities in physical implementations of QKD can be fully practically exploitable, and must be given increased scrutiny if quantum cryptography is to become highly secure.

Although perfect copying of unknown quantum systems is forbidden by the laws of quantum mechanics, approximate cloning is possible. A natural way of realizing quantum cloning of photons is by stimulated emission. In this context the fundamental quantum limit to the quality of the clones is imposed by the unavoidable presence of spontaneous emission. In our experiment a single input photon stimulates the emission of additional photons from a source based on parametric downconversion. This leads to the production of quantum clones with near optimal fidelity. We also demonstrate universality of the copying procedure by showing that the same fidelity is achieved for arbitrary input states.No device is capable of producing perfect copies of an unknown quantum system. This statement, known as the "no-cloning theorem" [1,2], is a direct consequence of the linearity of quantum mechanics, and constitutes one of the most significant differences between classical and quantum information. The impossibility of copying quantum information without errors is at the heart of the security of quantum cryptography [3]. If one could perfectly copy arbitrary quantum states, this would make it possible to exactly determine the state of an individual quantum system, which -in combination with quantum entanglement -would even lead to superluminal communication [4]. Thus the no-cloning principle also ensures the peaceful coexistence of quantum mechanics and special relativity.Given that perfect cloning is impossible, it is natural to ask how well one can clone. This question was first addressed in [5], and initiated a large amount of theoretical work. In particular, bounds on the maximum possible fidelity of the clones produced by universal cloning machines were derived [6]. A universal cloning machine produces copies of equal quality for all possible input states. Following the work of [5], quantum cloning was discussed mainly in the language of quantum computing, where its realization was envisaged in the form of a certain quantum logical network, consisting of a sequence of elementary quantum gates. An implementation of the cloning network based on NMR has recently been reported [7], but uses ensemble techniques and thus does not constitute true cloning of individual quantum systems. In another experiment the polarization degree of freedom of a single photon was approximately copied onto an external degree of freedom of the same photon [8]. Although formally this is a realization of a quantum cloning network, only a single particle is involved in the whole process.One might look for more natural ways of realizing quantum cloning. In the first papers on the topic a connection to the process of stimulated emission was made and it was suggested that stimulated emission might allow perfect copying [4]. It was subsequently pointed out [9,10] that perfect cloning is frustrated by spontaneous emission. Recently it was proposed [11] that optimal quantum cloning, where the quality of the copies saturates the fundamental quantum bounds, could b...

Quantum theory predicts and experiments confirm that nature can produce correlations between distant events that are nonlocal in the sense of violating a Bell inequality 1 . Nevertheless, Bell's strong sentence 'Correlations cry out for explanations' (ref. 2) remains relevant. The maturing of quantum information science and the discovery of the power of non-local correlations, for example for cryptographic key distribution beyond the standard quantum key distribution schemes 3-5 , strengthen Bell's wish and make it even more timely. In 2003, Leggett proposed an alternative model for non-local correlations 6 that he proved to be incompatible with quantum predictions. We present here a new approach to this model, along with new inequalities for testing it. These inequalities can be derived in a very simple way, assuming only the non-negativity of probability distributions; they are also stronger than previously published and experimentally tested Leggett-type inequalities 6-9 . The simplest of the new inequalities is experimentally violated. Then we go beyond Leggett's model, and show that we cannot ascribe even partially defined individual properties to the components of a maximally entangled pair.Formally, a correlation is a conditional probability distribution P(α, β|a, b), where α, β are the outcomes observed by two partners, Alice and Bob, when they make measurements labelled by a and b, respectively. On the abstract level, a and b are merely inputs, freely and independently chosen by Alice and Bob. On a more physical level, Alice and Bob hold two subsystems of a quantum state; in the simple case of qubits, the inputs are naturally characterized by vectors on the Poincaré sphere, hence the notation a,b.How should we understand non-local correlations, in particular those corresponding to entangled quantum states? A natural approach consists in decomposing P(α, β|a, b) into a statistical mixture of hopefully simpler correlations:Bell's locality assumption is P l (α, β|a, b) = P A l (α|a)P B l (β|b), admittedly the simplest choice, but an inadequate one as it turns out: quantum correlations violate Bell's locality 1 . Setting out to explore other choices, it is natural to require first that the P l fulfil the so-called no-signalling condition, that is, that none of the correlations P l results from a communication between Alice and Bob. This can be guaranteed by ensuring spacelike separation between Alice and Bob. Non-signalling correlations happen without any time ordering: there is not a first event, let us say on Alice's side, that causes the second event via some spooky action at a distance. We may phrase it differently: nonsignalling correlations happen from outside space-time, in the sense that there is no story in space-time that tells us how they happen. This is the case in orthodox quantum physics, or in some illuminating toy models such as the non-local box of Popescu and Rohrlich (PR box) 10 . Mathematically, the no-signalling condition reads P l (α|a,b) = P l (α|a) and P l (β|a,b) = P l (β|b): Alice'...