In previous studies, the production of ions in an APCI source without any corona discharge was observed, and the intensity of the ion signals showed significant increases on placing a metallic surface at 45° inside an orthogonal ion source. This method was named surface-activated chemical ionization (SACI). The present study was performed to investigate the mechanisms of ion production with or without the presence of the metallic surface, by varying instrumental parameters and the geometrical configuration. Approximate calculations show that, in the absence of corona discharge and of any additional surfaces, ions cannot be produced by collisional phenomena, because of their low kinetic energy, in the 10−2 to 10−3 eV range. Two alternative possibilities have been considered: the first takes into account that ions may originate by collision of neutral clusters of polar solvent molecules with the APCI source surfaces through clusterelectric effect. The second takes into account that the water dissociation constant kw is temperature dependent, passing from 10−14.1669 at 20 °C to 10−12.4318 at 90 °C. It means that the [H+] varies from 8.3 × 10−8 to 6.1 × 10−7M going from 20 to 90 °C. Hence, at the high temperatures experimented in the APCI vaporizer, H+ becomes available in solution in molar quantities analogous to those of analyte, and the protonation of the analyte itself can consequently occur. The activation of further ionization processes in the presence of the metallic surface can be reasonably attributed to interactions between gas-phase analyte molecules and solvent molecules adsorbed on the surface. Experiments performed with a thin layer of deuterated glycerol on the surface led to unequivocal results, i.e. the production of [M + D]+ ions of the analyte. Copyright © 2005 John Wiley & Sons, Ltd.