Detailed compositional analysis of a bio-oil (BO) from pyrolysis of rice husk was carried out. The BO was extracted sequentially with n-hexane, CCl4, CS2, benzene and CH2Cl2. In total, 167 organic species were identified with GC/MS in the extracts and classified into alkanes, alcohols, hydroxybenzenes, alkoxybenzenes, dioxolanes, aldehydes, ketones, carboxylic acids, esters, nitrogen-containing organic compounds and other species. The benzene ring-containing species (BRCCs) were attributed to the degradation of lignin while most of the rests were derived from the degradation of cellulose and hemicellulose. Along with guaiacyl and p-hydroxyphenyl units as the main components, a new type of linkage was suggested, i.e.. C-ar-CH2-C-ar in 4,4'-methylenebis(2,6-dimethoxyphenol). Based on the species identified, a possible macromolecular structure of the lignin and the mechanism for its pyrolysis are proposed. The BO was also extracted with petroleum ether in ca. 17.8% of the extract yield and about 82.1% of the extracted components are BRCCs. (C) 2012 Elsevier Ltd. All rights reserved.