All Abstracts | Poster Abstracts | Talk Abstracts

Loophole-free steering for quantum cryptography and for testing the subjectivity of atomic quantum jumps

Howard Wiseman, Griffith University

(Session 9 : Saturday from 10:15 - 11:00)

Abstract. Entanglement is the defining feature of quantum mechanics, and has application in quantum key distribution (QKD) and other quantum technologies. The strongest (and strangest) form of quantum correlation arising from entanglement is Bell-nonlocality [1] (the violation of Bell inequalities at a distance), which lies behind the theoretical possibility of device independent (DI) QKD [2]. The existence of quantum nonlocality of a weaker sort, however, goes back much further than Bell, to the seminal 1935 papers [3] of Einstein, Podolsky and Rosen (EPR) and Schrödinger. The latter coined the term "steering" for the EPR effect, viz. the ability of Alice, by her choice of measurement, to remotely prepare different types of quantum states for Bob, when they share entanglement. EPR-steering has since been formalized as a quantum information task [4], allowing the development of powerful new tests of this phenomenon: EPR-steering inequalities [5]. Just as for Bell inequalities, experimental violations of EPR-steering inequalities may be criticized if they rely upon a fair sampling assumption, as this opens the "detection loophole". For Bell nonlocality, closing this loophole requires both parties' detectors to have high efficiency; for EPR-steering Bob's detector is explicitly trusted [4], so only Alice's efficiency matters. Recently, we have demonstrated for the first time, in three separate experiments, EPR-steering using distant entangled qubits, with no detection loophole [6]. In the first of these, Alice's heralding efficiency was as high as 62% (that is, the probability of Alice seeing a photon given that Bob sees a photon is 62%). This allowed a demonstration using only two settings per side, in the same configuration as standard QKD. The significance of this is that we have shown [7] that there is a more secure version of standard QKD, in which only one detector (Bob's) need be trusted, while Alice's detector and the entanglement source remain untrusted. We call this 1-sided (1s) DI-QKD, as opposed to fully DI-QKD in which no devices need be trusted. For maximally entangled states, 1sDI-QKD is possible with a heralding efficiency for Alice of only 66% [7], while the best known fully DI-QKD scheme [7] requires both Alice and Bob to have heralding efficiencies (if equal) of over 91%. In Ref. [8] we also apply steering to suggest experiments to try to rule out all objective pure-state dynamic models, such as quantum jumps [9] or quantum diffusion [10], for an atom. Our proposed tests, using a strongly driven two-level atom, do not rely upon any special preparation of the atom or field. Our best test (using homodyne detection and a complicated adaptive photo-detection scheme) requires an efficiency of only 58%, and a simpler test (using just homodyne detection) only 73%. REFERENCES [1] J. S. Bell, Physics (N.Y.) 1, 195 (1964). [2] A. Acín et al., Phys. Rev. Lett. 98, 230501 (2007). [3] A. Einstein, B. Podolsky, N. Rosen, Phys. Rev. 47, 777 (1935); E. Schrödinger, Proc. Camb. Phil. Soc. 31, 555 (1935). [4] H. M. Wiseman, S. J. Jones, A. C. Doherty, Phys. Rev. Lett. 98, 140402 (2007). [5] E. G. Cavalcanti, S. J. Jones, H. M. Wiseman, M. D. Reid, Phys. Rev. A. 80, 032112 (2009). [6] D. H. Smith et al., Nature Communications 3, 625 (2012); B. Wittmann, et al., New J. Phys. 14, 053030 (2012); A. J. Bennet et al., Phys. Rev. X (in press, 2012). [7] C. Branciard et al. Phys. Rev. A (Rapid Comm.) 85, 010301(R) (2012). [8] H. M. Wiseman and J. M. Gambetta, Phys. Rev. Lett. 108, 220402 (2012). [9] N. Bohr, Phil. Mag. 26, 1 (1913); A. Einstein, Physikalische Zeitschrift 18, 121 (1917). [10] N. Gisin and I. C. Percival, J. Phys. A 25, 5677 (1992).