Date of this Version
Physical Review Letters, 105, 249902 (2010); DOI: 10.1103/PhysRevLett.105.249902
The Letter  should have acknowledged and cited the work by Andreev , which was inadvertently overlooked. This latter work introduced a phenomenological surface magnetization and concluded, by analyzing exchange invariants, that it may be finite for all antiferromagnets and that those with unbroken macroscopic time-reversal symmetry can exhibit surface magnetization domains. These arguments are highly relevant to Ref. , which I happily acknowledge. The work  treats the problem of (otherwise poorly defined) boundary magnetization as a special case of a general, microscopically definable probe functional, explicitly taking into account boundary roughness and allowing for relativistic interactions. It also spells out the implications for electrically controlled magnetism using magnetoelectric and multiferroic materials.