An extended overview of the organic and inorganic phase composition of biomass was conducted. Some general considerations and problems related to phase composition of biomass as a solid fuel were discussed initially. Then, reference peer-reviewed data including contents of cellulose, hemicellulose, lignin and bulk extractives of 93 varieties of biomass were used and grouped for their comparison and classification. Additionally, reference peer-reviewed data and own investigations for various minor organic components and minerals, and modes of element occurrence identified in biomass were also applied and organised to describe the biomass systematically. It was found that the phase distinctions among the specified natural and anthropogenic (technogenic) biomass groups, sub-groups and varieties are significant and relate to different biomass sources and origin. The phase composition of biomass is highly variable due to the extremely high variations of structural components and different genetic types (authigenic, detrital and technogenic) of inorganic matter. The technogenic biomass group is quite complicated as a result of incorporation of various non-biomass materials during biomass processing. It was identified that the biomass phase composition is significantly different from that of coal. Correlations and associations among phase and chemical characteristics were studied to find some major trends and important relationships occurring in the natural biomass system. Certain leading associations related to the occurrence, content and origin of elements and phases in biomass were identified and discussed, namely: (1) C-H (mainly as authigenic cellulose, hemicellulose, lignin and organic extractives); (2) Si-Al-Fe-Na-Ti (mostly as detrital silicates and oxyhydroxides, excluding authigenic opal); (3) Ca-Mg-Mn (commonly as authigenic oxalates and carbonates); and (4) N-K-S-P-Cl (normally as authigenic phosphates, sulphates, chlorides and nitrates). Finally, it was emphasised that these important associations have potential applications and can be used for initial classifications or prediction and indicator purposes connected with future advanced and sustainable processing of biomass to biofuels and chemical feedstock. (C) 2011 Elsevier Ltd. All rights reserved.