On the quantum information processing in nuclear magnetic resonance quantum computing experiments

Full text: Nuclear Magnetic Resonance appeared in the late nineties to be the most promising candidate to run quantum computing algorithms. An impressive number of experiments demonstrating the implementation of all logic gates and quantum algorithms in systems with a small number of qubits stimulated the general excitement about the technique, and greatly promoted the field. Particularly important were those experiments where entanglement of particles were aimed at. Entanglement is the most fundamental (and weird !) aspect of quantum systems, and is at the basis of quantum teleportation and quantum cryptography, yet impossible to prove in NMR experiments. The hardcore of NMR quantum computing are the so-called pseudo-pure states, upon which radiofrequency (RF) pulses act to implement quantum mechanical unitary transformations, promoting changes in both, Zeeman level populations and coherences in the density matrix. Whereas pseudo-pure states are special non-equilibrium diagonal states, coherences encode information about superposition states. Now, one could safely say that the whole business of quantum computing goes about controlling relative ket phases. In spite of the impossibility to univocally associating a given quantum state to a NMR spectrum, it is possible to demonstrate the phase action of RF pulses over relative ket phases, even if no population changes take place. In this talk these issues will be addressed, and we will show experimental results of our own where this is done in the two-qubit quadrupole nuclei 23Na in C10H21NaO4S liquid crystal. We demonstrate the reversibility of the Hadamard gate, and of a quantum circuit which generates pseudo-Bell states. The success of the operation reaches almost 100% in the case of the state |01+|10, 80% in the cases of |00> + |01> and |10> + |11>, and 65% for the cat-state |00> + |11>. (author)