We examined hydrothermal liquefaction (HTL) of simulated food waste over a wide range of temperatures (200–600 °C), pressures (10.2–35.7 MPa), biomass loadings (2–20 wt %), and times (1–33 min). These conditions included water as vapor, saturated liquid, compressed liquid, and supercritical fluid and explored both isothermal and fast HTL. The highest biocrude yields (∼30 wt %) were from HTL near the critical temperature. The most severe reaction conditions (600 °C, 35.3 MPa, 30 min) gave biocrude with the largest heating value (36.5 MJ/kg) and transfer of up to 50% of the nitrogen and 68% of the phosphorus in the food mixture into the aqueous phase. Energy recovery in the biocrude exceeded 65% under multiple reaction conditions. Saturated fatty acids were the most abundant compounds in the light biocrude fraction under all the reaction conditions. Isothermal HTL gave a higher fraction of heavy compounds than fast HTL. A kinetic model for HTL of microalgae predicted 2/3 of the experimental biocrude yields from HTL of food waste to within ±5 wt %, and nearly 90% to within ±10 wt %. This predictive ability supports the hypothesis that biochemical composition of the feedstock is important input for a predictive HTL model.

No abstract

We carbonized simulated food waste (stage I) and then liquefied the biochar produced (stage II) with the goals of producing bio-oil and recovering nitrogen. Both stages used hydrothermal and pyrolytic approaches, so the influence of water during the treatments could be discerned. Pyrolysis produced biochars in the greatest yield (57 wt %) from the biomass feedstock, and it produced biocrudes with the greatest HHV (39.4 MJ/kg) via the liquefaction of biochar from hydrothermal carbonization. Pyrolysis of biochar for stage II gave negligible aqueous-phase product yields, however, which limited the nitrogen recovery with this approach solely to that recovered in the initial carbonization step. The highest N recovery (75%) in the aqueous-phase products occurred with hydrothermal treatment for both carbonization and liquefaction. This N recovery greatly exceeded those (<10%) for single-step hydrothermal liquefaction of this same feedstock. Energy recovery in the biocrude oil produced from this two-step process exceeded 50% in several runs. This two-step approach for food-waste valorization provides an opportunity for comparable energy recovery and much greater N recovery than are available from single-step hydrothermal liquefaction.

We examined the effect of supported metals (Ni/C, Pt/C, Ru/C, Pd/C, Ni/SiO2–Al2O3, Pt/Al2O3, and Ru/Al2O3), bulk metal oxides (CaO, Al2O3, CeO2, La2O3, and SiO2), and a set of salt, acid, and base additives on the hydrothermal liquefaction (HTL) of simulated food waste. Supported metals and the additives did not increase biocrude yields, but three of the metal oxides did lead to higher yields, with the following order: SiO2 > La2O3 > CeO2. The elemental compositions and heating values of the biocrudes were sensitive to the type of potential catalyst used, especially in the presence of high-pressure hydrogen. The higher heating values (HHVs) of the biocrude from HTL were higher with added H2 and supported metal. Of all the potential catalysts tested, K3PO4 produced oil with the greatest HHV (37.5 MJ/kg). Fatty acids were the major GC-elutable compounds in most of the oils, save that produced with added CaO, where amides and N-containing compounds dominated. Thermogravimetric analysis showed that the distribution of the volatilities of the molecules in the biocrude oils is sensitive to the type of metal oxide used. HTL with CaO and no metal oxide recovered the most nitrogen and phosphorus, respectively, in the aqueous phase.