Nature 413, 716–718 (2001)

In this Letter, there was a mistake in the presentation of the synthesis conditions of the reported samples. The actual range of the temperatures of synthesis for the four rhombohedral samples was 975–1,025 K. One of the five reported samples was wrongly characterized in relation to the polymerization type: the sample was actually prepared at 2.5 GPa (synthesis temperature, 1,125 K), representing a mixture of the rhombohedral and tetragonal phases with some hard carbon. The error in characterization of this sample weakens our attribution of the ferromagnetism to defects in the rhombohedral phase (Rh-C60) but does not influence our main conclusion concerning the observation of magnetism in a carbon solid based on polymerized fullerenes, although its origin and the actual magnitude remain an open question. Also, we were unaware of earlier work on magnetism in polymerized fullerenes1, that should have been cited.

 T.L.M. takes full responsibility for the misidentification of the sample preparation conditions. We thank A. V. Talyzin for alerting us to this mistake.