Statistical optimization of H₂, 1,3-propanediol and propionic acid production from crude glycerol using an anaerobic fluidized bed reactor: Interaction effects of substrate concentration and hydraulic retention time

Highlights

• Optimization of H₂, 1,3-propanediol (1,3-PD) and HPr production from glycerol were examined.

• Effects of substrate concentration and hydraulic retention time were evaluated in AFBR.

• The maximum H₂ yield of 0.31 mol H₂.kg COD applied⁻¹ was obtained with 18.0 g L⁻¹ and 7.72 h.

• The maximum yield of 1,3-PD, 0.87 g g glycerol consumed⁻¹, was obtained with 10 g L⁻¹ and 9.24 h.

• The highest yield of HPr, 0.57 g g glycerol consumed⁻¹, was obtained with 15 g L⁻¹ and 2 h.

Abstract

The individual and interactive effects of the influent glycerol, G_in (2.9–17.1 g. L⁻¹), and hydraulic retention time, HRT (0.76–9.24 h), were evaluated to optimize the production of hydrogen (H₂ content and yield), 1,3-propanediol (1,3-PD) and propionic acid (HPr) in a mesophilic (30 °C) anaerobic fluidized bed reactor (AFBR). The maximum H₂ content of 89.6% was obtained under the optimum conditions of 12.6 g. L⁻¹ and 4.58 h, while the maximum H₂ yield of 0.31 mol H₂.kg COD applied⁻¹ was obtained under the conditions of 18.0 g. L⁻¹ and 7.72 h. For both responses, G_in was the more significant individual variable; however, the interactive effects between the G_in and HRT variables also suggest their significance in the process. The influence of the organic loading rate (OLR) on the production of 1,3-PD and HPr was also investigated in the reactor. The maximum yield of 1,3-PD was 0.87 g g glycerol_consumed⁻¹ and was obtained under the conditions of 10 g. L⁻¹ and 9.24 h, which is equivalent to an OLR of 23.62 kg m.⁻³.day⁻¹. In contrast, HPr was produced in the highest yield of 0.57 g g glycerol_consumed⁻¹ under the conditions of 15 g. L⁻¹ and 2 h, which is equivalent to an OLR of 160.60 kg m.⁻³.day⁻¹. The different specific conditions determined to favor each product enhance the mixed fermentation and the existence of competing metabolic pathways in the AFBR.

Introduction

Hydrogen has been identified as a promising energy source due to its high calorific value (122 kJ g⁻¹) and the sole generation of water during its combustion [1]. One process to produce hydrogen that has attracted attention in the scientific community is through the anaerobic digestion of organic matter [[2], [3], [4]]. By means of fermentative pathways, several wastes can serve as the basic raw materials for hydrogen production. Recent studies have shown the technical viability of biological hydrogen production from simple substrates such as glucose [[5], [6], [7]] and sucrose [8,9]. The use of industrial and domestic residues has increased research interest in biohydrogen production from complex substrates, such as sugarcane vinasse [10,11], molasses [12,13], cheese whey [14,15] and glycerol [[16], [17], [18]].

In this context, glycerol, as the main byproduct of biodiesel production (10% w.w⁻¹), has gained attention as a potential substrate in fermentative processes focused on the production of H₂ and other valuable products, such as alcohols and organic acids [19]. In addition, several microorganisms can grow anaerobically in glycerol as the sole carbon and energy source, for example, Citrobacter freundii, Klebsiella pneumoniae, Clostridium pasteurianum, Clostridium butyricum, Enterobacter agglomerans, Entrobacter aerogenes and Lactobacillus reuteri [19].

Because of the metabolic complexity of glycerol fermentation, H₂ production is usually accompanied by the generation of other byproducts that are favored by the microbial community in the reactor, as well as by the reactor hydrodynamics, initial substrate concentration, nutrient medium, pH, hydraulic retention time (HRT), and temperature, among several other factors [20]. According to Biebl et al. (1999) [21], the fermentation of glycerol can occur by a reductive or an oxidative route. By the reductive route, glycerol undergoes a dehydration process producing 1,3-propanediol (1,3-PD), which is highly specific for glycerol fermentation and cannot be obtained by any other anaerobic conversion [22]. The oxidative route consists of the dehydrogenation of glycerol, converting it mainly to propionate but also to 2,3-butanediol, lactate, butyrate, ethanol, formate, acetate, hydrogen and carbon dioxide [[19], [20], [21], [22], [23]]. Because propionic acid has the same degree of reduction as glycerol (for both, γ = 4.67), coproduction is not necessary to maintain the redox balance. This enables a higher production of propionic acid from glycerol than from other sources [24].

Different studies aiming to produce hydrogen, 1,3-PD and propionic acid from glycerol were conducted using various reactor configurations, such as batch reactors [16,[25], [26], [27], [28]], anaerobic packed bed reactors (APBRs) [[29], [30], [31], [32]], continuous stirred tank reactors (CSTRs) [16] and upflow anaerobic sludge blanket (UASB) reactors [17,18]. In recent studies by our research group, the application of anaerobic fluidized bed reactors (AFBRs) was promising for H₂ production from wastewater [10,11,14,15,33] and for the production of 1,3-PD [20] and propionic acid [34] from residual glycerol. Among the advantages of AFBRs are the high concentration of biomass adhered to the support material, lower pressure drops than those in APBRs and low external resistance to mass transfer compared to other reactor configurations [33].

In addition to the configuration and hydrodynamics of the reactor, the anaerobic metabolism of H₂, 1,3-PD or propionic acid production can be shifted by two easily controlled parameters, namely, the HRT and substrate concentration, which results in the organic loading rate (OLR) applied to the reactor [9]. Using two thermophilic (55 °C) AFBRs, Ottaviano et al. [33] observed significant differences in the strategies to adapt H₂ production to higher OLR values. In an AFBR in which the HRT was decreased from 8 h to 0.5 h, the authors observed a maximum hydrogen production rate (HPR) of 4.1 L H₂.h⁻¹. L⁻¹ at an OLR of 240 kg COD.m⁻³.day⁻¹, while an OLR of 60 kg COD.m⁻³.day⁻¹ was considered inhibitory in the AFBR where the substrate concentration was increased from 2800 to 14,600 mg carbohydrates. L⁻¹ [33]. This result shows that the adaptation of the microbial community, the metabolic pathways and the selection of products is not only determined by the operational conditions (HRT and substrate concentration) used to achieve the OLR but also depends on how the OLR was achieved in experimentally. However, the majority of studies emphasize the individual effect of these parameters in a univariate analysis without evaluating their interactive effect in the process or in the strategy to increase the OLR [6,15,38]. The interactive effects of these variables on H₂ production and the influence of the OLR on the 1,3-PD and propionic acid production from crude glycerol in a continuous AFBR have not been previously reported in the literature.

As a valuable tool in this type of analysis, response surface methodology (RSM) allows the exploration of the individual and interactive effects that the independent variables exert on the desired response, in addition to the determination of the optimal conditions that maximize the response [35]. Thus, in the present study, the production of H₂, 1,3-PD and propionic acid from crude glycerol was optimized in a continuous AFBR to evaluate the individual and interactive effects of the influent glycerol concentration (G_in) and HRT using RSM with a central composite design (CCD). Our previous study [20] evaluated the best nutrient medium for each glycerol fermentative pathway, while this study continues the research on evaluating the effects of G_in, HRT and OLR in the operation of a continuous AFBR.