Bounding the Biohydrogen Potential of Organic Market Waste: Fruit-Fraction-Dependent Acidogenesis, Non-Additive Mixture Response, and Overestimation of Avoided-Methane Credits

Dark fermentation of market waste is repeatedly proposed as a decentralized biohydrogen pathway for Latin American cities, and its climate benefit is almost invariably credited using the IPCC Tier-1 term for avoided methane applied to the full diverted mass. This study evaluates how much hydrogen this waste stream can actually deliver and whether such a credit survives a stoichiometric balance. A seven-day gravimetric survey of 10 stalls at El Pino Market (Carabayllo, Lima, Peru) yielded 754.6 kg week−1, distributed as 41.04% fruit, 22.82% vegetables, and 36.14% cooked food.Four mixtures reproducing the recorded generation ratios of fruit (F), vegetables (V), and cooked food (A) were fermented in batch reactors in quintuplicate for 72 h: (F + V, V + A, F + A, and F + V + A). were fermented in batch reactors, with cumulative gas recorded every 6 h, yielding 240 observations. The modified Gompertz model fitted all curves (R2=0.982–0.994). Specific yields were 26.45 mL g−1 VS for F + A, 18.89 for F + V, 18.41 for F + V + A, and only 0.63 for the fruit-free V + A mixture (Welch’s ANOVA F3,7.32=8.65×104, p<0.001; Games–Howell separated all pairs except F + V versus F + V + A, p=0.713). A linear Scheffé mixture model formulated on volatile-solids fractions and calibrated on the three fruit-containing mixtures predicted 5.48 mL g−1 VS for V + A, whereas the observed value was 11.6% of that prediction (95% bootstrap CI 9.8–13.7%), an 88.4% suppression of total fermentative gas that persists under every admissible additive null (2.5–13.7%) and cannot be explained by substrate concentration. Because the gas phase was not speciated, hydrogen production is bounded between an upper bound in which all gas is H2 and a reference case was measured at 45% v/v, yielding 11.90–26.45 mL H2 g−1 VS for the best-performing mixture. Even at the upper bound, stoichiometric closure relative to the acetate pathway reaches only 9.71%, and energy recovery amounts to 0.76–1.68% of the substrate chemical energy, compared with 73.8–84.3% for a methanogenic reference. A carbon and energy balance shows that more than 98% of the substrate energy and more than 95% of the carbohydrate carbon leave the process in the effluent. Acidogenesis therefore leaves most of the degradable organic carbon in the effluent, and the uncorrected Tier-1 credit of 29.44 t CO2eq yr−1 overestimates the attainable benefit by a factor of 10 to 23 when scaled by the H2-based stoichiometric proxy (1.29–2.86 t CO2eq yr−1), and by a factor of 15 to 69 when scaled by the pathway-consistent carbon-based factor (0.43–1.90 t CO2eq yr−1). The biohydrogen potential of market waste is therefore bounded by its fruit fraction; fruit-free mixtures are not merely poorer but fermentatively suppressed; and avoided-methane credits must be constrained by measured process conversion and by the fate of residual carbon rather than by diverted waste mass alone.