Prediction of Properties and Elemental Composition of Biomass Pyrolysis Oils by NMR and Partial Least Squares Analysis

Abstract

As with many substances derived from natural products, pyrolysis bio-oils are complex chemical mixtures and are extremely challenging to chemically characterize, requiring multiple separation and pretreatment steps followed by several different analytical techniques that need tedious adjustments and modifications when sample properties change. In this study, we present a way to simplify this analysis by using 13C NMR to characterize such substances as a whole without modification. Using partial least-squares (PLS) regression, we report what we believe to be the first reported use of 13C NMR to derive elemental composition information (mass fractions of C, H, N, and O) as well as the enthalpy of combustion (higher heating value), phenol and cresols concentrations, and the total acid number. Several PLS models were created correlating these various properties with the binned intensities of the 1H and 13C NMR spectra of 73 different samples consisting of pyrolysis bio-oils from various biomass sources and treatment protocols as well as finished fuels (gasoline, diesel, and biodiesel) and small molecule standards. Two models based exclusively on 13C NMR data demonstrated the best overall ability to predict these same properties for unknown samples. The R2 and RMSE of the predicted values are discussed in detail and are acceptable for many biofuel-related applications. That such properties and compositional measurements may be extracted from 13C NMR spectra is a direct result of the detailed chemical structural information influencing the chemical shifts and resonance patterns. Because these models were built using a wide range of samples and conditions, they are expected to also be useful for a wider range of applications.