Batch and continuous-flow thermophilic hydrogen production from molasses and cheese whey by suspended and immobilized cells of Thermotoga neapolitana

BACKGROUND. Bio-hydrogen production from organic residues is an attractive process that combines energy generation with waste treatment. This work represents the continuation of a previous study of biological H2 production from food industry wastes (Cappelleti et al., J. Chem. Technol. Biotechnol. 87:1291-1301), where the comparison of four thermophilic Thermotoga strains and four carriers for biomass immobilization led to the identification of T. neapolitana as the best-performing strain and of a porous ceramic material named Biomax as the best-performing support for biofilm formation. AIMS. The goals of this work were: a) to develop and calibrate a kinetic model of H2 production from glucose, molasses and cheese whey by suspended and attached cells of T. neapolitana, and b) to set-up and optimize a continuous-flow process of H2 production from cheese whey by T. neapolitana in a 20-L stirred reactor. METHODS. The experimental H2 rates used to develop the kinetic model were obtained from suspended- and attached-cell batch tests conducted in 120-mL bioreactors, exposed to different initial concentrations of glucose, molasses or cheese whey (0.1-95 g L-1, as total sugars). The process was scaled-up in a 20-L fermentor, fed only with cheese whey in consideration of the significantly higher production of this waste in Northern Italy, in comparison with molasses. The fermentor was operated initially with suspended-cells, and subsequently with attached-cells. To immobilize T. neapolitana in the fermentor, 5 L of the previously selected carrier were introduced in pockets made with steel net and fixed to the top and bottom of the vessel. RESULTS. The H2 specific production rates obtained with suspended and attached cell tests of T. neapolitana are shown in Figure 1, together with the best fitting simulations performed with the Haldane substrate inhibition model. The corresponding values of the kinetic parameters are reported in Table 1. For all the substrates studied, a marked substrate inhibition was noticed. The attached-cell tests resulted in higher rates than the suspended-cell ones, in particular in the case of glucose and molasses. Further tests indicated the absence of oxygen inhibition (studied range: 0-1% in the headspace, or 0-0.21 mgdissolved O2 L-1) and of product (H2) inhibition (studied range: 0-60% in the headspace, or 0-0.73 mgdissolved H2 L-1). The batch tests of H2 production from cheese whey conducted in the 20-L fermentor led to a productivity of 6.8 mmolH2 L-1 day-1 with suspended cells, and 22.5 mmolH2 L-1 day-1 with attached cells. The fermentor was also operated with attached cells in continuous-flow mode, with hydraulic retention times (HRT) in the 0.8-2.9 day range. The resulting H2 productivity showed an increasing trend with decreasing HRT, with a maximum value of 125 mmolH2 L-1 day-1. CONCLUSIONS. The calibration of a model of thermophilic H2 production from different wastes and the successful scale-up of the process of cheese whey conversion into H2 in a continuous-flow 20-L bioreactor pose the basis for a model-based optimization of the attached-cell process, and for its further scale-up to a pilot plant.