Absence of true critical exponents in relaxor ferroelectrics: the case for nanodomain freezing

Scott's (2006 J. Phys.: Condens. Matter 18 7123) recent review of Kleemann et al' s (2002 Europhys. Lett. 57 14) critical exponents of strontium barium niobate is shown to be misled by erroneous input parameters. Although the observed set of exponents reflects, indeed, the absence of true three-dimensional (3D) random-field Ising model critical behaviour, it cannot be compatible with the proposed domain wall model in d = 2.5 dimensions or with Levanyuk and Sigov's (1988 Defects and Structural Phase Transitions (London: Gordon and Breach)) defect model. As was argued independently by Kleemann et al (2006 Phys. Rev. Lett. 97 065702), it is rather in agreement with the pure two-dimensional (2D) Ising model. (letter to the editor)