Recently, Gnahm et al. [Electrochim. Acta 55 (2010)6212] analyzed and discussed capacitive processes at the interface between the ionic liquid 1-butyl-3-methylimidazolium hexafluorophosphate and a single-crystalline Au(1 1 1) electrode. In order to derive interfacial capacitance spectra, they subtracted the bulk resistance R-s from the experimental impedance spectra Z(nu) and used the following expression (C) over cap(nu) = (i2 pi nu x ((Z) over cap(nu) - R-s))(-1). The subtraction procedure was based on the assumption that lim(nu ->infinity)Z'(nu) = R-s. In this commentary we argue that this assumption neglects the high-frequency capacitance C, acting in parallel to R-s. In a three-electrode setup, the high-frequency capacitance is often governed by the input capacitance of the reference electrode. The parallel action of C-s and R-s results in lim(nu ->infinity)Z'(nu) = 0. Consequently, the subtraction of R-s leads, in general, to negative -C ''(sub)(nu) values at high frequencies which have no physical meaning. Moreover, the subtraction procedure causes an artificial distortion of the double layer semicircle in the complex capacitance plane. (C) 2011 Elsevier Ltd. All rights reserved.