In ethanol wet bonding, water is replaced by ethanol to maintain dehydrated collagen matrices in an extended state to facilitate resin infiltration. Since short ethanol dehydration protocols may be ineffective, this study tested the null hypothesis that there are no differences in ethanol dehydration protocols for maintaining the surface roughness, fibril diameter, and interfibrillar spaces of acidetched dentin. Polished human dentin surfaces were etched with phosphoric acid and waterrinsed. Tested protocols were: (1) water-rinse (control); (2) 100% ethanol-rinse (1-min); (3) 100% ethanol-rinse (5-min); and (4) progressive ethanol replacement (50-100%). Surface roughness, fibril diameter, and interfibrillar spaces were determined with atomic force microscopy and analyzed by one-way analysis of variance and the Student-Newman-Keuls test (α = 0.05). Dentin roughness and fibril diameter significantly decreased when 100% ethanol (1-5 min) was used for rinsing (p < 0.001). Absolute ethanol produced collapse and shrinkage of collagen fibrils. Ascending ethanol concentrations did not collapse the matrix and shrank the fibrils less than absolute ethanol-rinses.KEY WORDS: atomic force microscopy, acidetched dentin, ethanol-saturated, surface topography, water replacement.