文摘
Pyrolysis oils are complex mixtures of hundreds of compounds that can be described as a mixture of few chemical macro-families. This description is highly desirable for modeling bio-oil combustion, cracking, separation, and evaporation. In one of our previous papers we recommended the deconvolution of bio-oil DTG curves as the basis for characterizing bio-oils in chemical families. However, the positioning of the peaks was based on the shape of the DTG curve. More specific guidelines are needed to position the peaks in this deconvolution strategy. In this paper we developed and tested a methodology to identify the position and shape of the peaks in the DTG curves. The first strategy uses the FTIR spectra of the TG evolved pyrolysis oil vapors. The second consists of doping the bio-oil with model compounds and with its fractions. The IR studies provided information on the position and width of the fitting curves describing the behavior of the very light volatile compounds (<150 °C) and of the heavy oligomeric fraction (>400 °C). The range between these two temperatures was fitted based primarily on the shape of the DTG curve and the results from the oil doping studies. Contribution of the chemical families toward formation of coke residue was examined based on the doping studies, and rules for coke assignment were developed. The developed DTG deconvolution method was applied to ten pyrolysis oils produced from compositionally diverse biomass feedstocks and production technologies. The chemical composition of the oils obtained by our new DTG deconvolution strategy was compared with other well accepted characterization methods. Our results show that the method proposed could be used to estimate the changes in the content of water, light volatile compounds, and water-soluble (WS) and water insoluble (WIS) fractions.