51�������

Abstract:

In Fig. 1 of this Letter, there should have been two nitrogen (N) atoms at the 1,3-positions of all the blue chemical structures (next to the oxygen atoms), rather than one at the 2-position. The figure has been corrected online, and the original incorrect figure is shown as Supplementary Information to the accompanying Amendment.

Abstract:

Graphene, a single-layer network of carbon atoms, shows outstanding electrical and mechanical properties, and graphene ribbons with nanometer-scale widths, should exhibit half-metallicity, quantum confinement and edge effects. Magnetic edges in graphene nanoribbons have undergone intense theoretical scrutiny, because their coherent manipulation would be a milestone for spintronic and quantum computing devices. Experimental investigations are however hampered by the fact that most nanoribbons do not have the required atomic control of the edges, and that the proposed graphene terminations are chemically unstable. Here we solve both of these problems, by using molecular graphene nanoribbons functionalized with stable spin-bearing radical groups. We observe the predicted delocalized magnetic edge states, and test present theoretical models about the spin dynamics and the spin-environment interactions. Comparison with a non graphitized reference material allows clear identification of fingerprint behaviours. We quantify the spin-orbit coupling parameters, define the interaction patterns, and unravel the spin decoherence channels. Even without any optimization, the spin coherence time is in the μs range at room temperature, and we perform quantum inversion operations between edge and radical spins. This new approach to problem of spins in well-defined electronic nanostructures offers a long awaited experimental testbed for the theory of magnetism in graphene nanoribbons. The observed coherence times open up encouraging perspectives for the use of magnetic nanoribbons in quantum spintronic devices.