Modelling of Amino Acid Fermentations and Stabilization of Anaerobic Digestates by Extracting Ammonium Bicarbonate

: With the current increase in demand for animal and agricultural products, management of agrowaste has become critical to avoid greenhouse gas emissions. The present article investigates the applicability of ammonium bicarbonate synthesis via ﬂash distillation to valorize and stabilize several types of anaerobic digestates which are produced from individual fermentations of amino acids. The content of CO 2 in the digestate was found to be responsible for the OH alkalinity (0.4 equivalents of acid/kg digestate), while the partial and total alkalinities (0.8 eq/kg digestate) were essentially derived from the content of NH 3 . The most suitable conditions for the ﬂash distillation were 95 ◦ C and 1 bar with the condensation occurring at 25 ◦ C. However, in order to attain the precipitation of NH 4 HCO 3 in the distillate, it was necessary to consider digestates with a moisture content of 50 wt.%, since saturation levels of inorganic nitrogen and inorganic carbon were not attained otherwise. Even under these conditions, few amino acids (i.e., arginine, glycine, and histidine) were able to provide an anaerobic digestate upon fermentation that would be suitable for NH 4 HCO 3 stabilization. The process of stabilization with a capacity of a t of digestate per h was improved by adding hydrochloric acid or sodium hydroxide at a rate of 44 kg/h, leading to production of 34 kg NH 4 HCO 3 /h. Given the role of the volatile elements of the biogas as endogenous stripping agents, it is recommended to use a fresh and saturated digestate as feed for the ﬂash distillation.


Introduction
Anaerobic digestion (AD) is one the most promising technologies for reliable production of clean energy [1]. According to Patel et al. [2], AD has the potential to convert up to 95 wt.% of organic matter to biogas. The valorization of waste by means of this fermentative process enhances the circular economy because most of the nutrients required for cultivation of crops end up in the organic residue (i.e., anaerobic digestate), which is mainly applied to land as organic soil amendment and reduces the consumption of industry-based inorganic fertilizers [3]. The mineralization of organic nutrients during AD implies that these become more readily available to plants but also that they can be easily lost via volatilization and leaching during the management of organic manures. Current regulations in force prescribe closing slurry stores and the use of low-emission spreading techniques at the time of land application, which should be done at specific locations and during the right season [4]. It is also possible to carry out the isolation of the nutrients, for example, by precipitating the struvite (i.e., magnesium ammonium phosphate). This open-loop management strategy requires prior solid-liquid separation of the anaerobic digestate, an external source of magnesium [5], and often additional phosphate to be able to deplete the ammoniacal nitrogen (NH 4 + -N) initially contained in the aqueous solution [6,7]. The formation of ammonium carbonate in the anaerobic digestate has been traditionally reported as one of the phenomena that increase and regulate the pH during anaerobic fermentation [8]; therefore, this could be among the most efficient routes for the isolation of NH 4 + -N. The ammonium bicarbonate process (known as ABC) [9][10][11] is, therefore, considered a more sustainable alternative for recovering ammonia, compared with other processes involving the use of exogenous scrubbing agents, such as sulfuric acid [12], and the operation of expensive equipment with high energy demand, such as an air blower [13]. Hence, the stabilization of anaerobic digestate by means of NH 4 HCO 3 synthesis has gained attention as a promising profitable technology [13]. Unlike open-loop processes, the patented ABC [9] offers a synergistic approach in which biogas with a composition of 65 vol.% CH 4 and 35 vol.% CO 2 [14] is efficiently used as a scrubbing agent of the stripped gas. In this process, the pH of the liquid fraction of the anaerobic digestate is adjusted to 11.5 via addition of NaOH to reduce 10 times its content of total ammonia (from 800 to 80 mg/L), using the NH 3 &CO 2 -depleted biogas (i.e., biomethane grade or 99 vol.% CH 4 ) as stripping gas. Subsequently, the resulting stripped gas is mixed with 58 vol.% of the fresh biogas, originally produced in the AD, to promote the precipitation of NH 4 HCO 3 and produce the biomethane stream. In addition to the NH 3 &CO 2 -depleted/stabilized liquid digestate and the solid crystals of NH 4 HCO 3 , the purges of 42 vol.% fresh biogas and 1 vol.% biomethane (these percentages refer to the total volume of the streams generated in the steps of anaerobic digestion and NH 4 HCO 3 precipitation, respectively) could be regarded as valuable outputs of the ABC. The 99 vol.% of the biomethane generated in the step of NH 4 HCO 3 precipitation is recirculated and employed as stripping gas. The role of the CH 4 in the ABC is not clear [9], nor whether the use of the biomethane as stripping gas affects (a) the volatilization of NH 3 from the alkalinized liquid digestate and (b) the subsequent NH 4 HCO 3 precipitation during the contact of the stripped gas with the fresh biogas.
In the distillation process engineered by Drapanauskaite et al. [13] for the production of NH 4 HCO 3 from the liquid fraction of an anaerobic digestate, the release of CH 4 was not accounted in the liquid distillate stream obtained at the top of the column at 3 bar and 49 • C (89.0 vol.% H 2 O, 5.0 vol.% HCO 3 − , 5.0 vol.% NH 4 + , 0.3 vol.% CO 2 , 0.1 vol.% NH 3 , 0.3 vol.% NH 2 COO − , and 0.1 vol.% CO 3 2− ). It is noteworthy that the Henry's law volatility constant of methane (71.43·10 3 Pa·m 3 ·mol −1 ) is greater than those of ammonia (1.69 Pa·m 3 ·mol −1 ) and carbon dioxide (3.03·10 3 Pa·m 3 ·mol −1 ) in an aqueous solution [15]. It should be noted that since NH 3 and CO 2 behave like non-condensable gases under the normal operating conditions of the distillation, Henry's law is preferred to Raoult's law to describe this two-step system: degasification and absorption. This agrees with the fact that less than 1% of the NH 4 + -N is volatilized as part of the biogas release during the AD [8]. In the condensed distillate at 3 bar and 49 • C and in the uncondensed purge there were shares of 0.2 vol.% and 0.3 vol.%, respectively, that were not identified and could be attributed to methane [13]. It is important to account for the release of any gas during the NH 4 HCO 3 synthesis because this can have a significant effect on the residual biogas potential of the stabilized digestate, for which an upper limit of 250-450 mL/g volatile solids is considered [16][17][18].
The present article investigates whether NH 4 HCO 3 manufacturing technology via flash distillation is suitable for any composition of anaerobic digestate or whether some requirements need to be introduced at the stage of selecting the feedstock for AD. It is proposed to further expand the process simulation model by Rajendran et al. [19] for AD while minimizing the heat requirements for the flash distillation, in agreement with the approach followed by Centorcelli et al. [20,21]. Therefore, this investigation supports and advances the 7th (affordable and clean energy) and 13th (climate action) Sustainable Development Goals of the United Nations. The closed-loop process, where the CO 2 is valorized as absorbent of the NH 3 for synthesis of NH 4 HCO 3 , is relevant for agroindustry and the chemical fertilizer industry.

Materials and Methods
Following the methodology of Drapanauskaite et al. [13], prior to the development of the models in Aspen Plus ® v12, the simulation was validated by comparing the empirical data available in the literature to the values provided by the software for the partial pressures of NH 3 and CO 2 in an aqueous system at the bubble point. Similarly, the supersaturation of the aqueous system in NH 4 HCO 3 , NH 4 COONH 2 , NaHCO 3 , and NH 4 Cl was also tested to ensure that the formation of solid phases was closely monitored in the flash distillation process. A model digestate was elaborated (Table A1) with the descriptions by Rajendran et al. [19] and Akhiar et al. [22] to verify the simulation of adding acids or bases to shift the chemical equilibria, following the titration methodology of Moure Abelenda et al. [23]. When it was confirmed that the system CO 2 -NH 3 -H 2 O was modelled correctly, the minimum conditions for the production of NH 4 HCO 3 were investigated with the calculation block of the flash tank and using the electrolyte non-random two-liquid (ELECNRTL) property method. The ELECNRT method is defined by Aspen Plus ® as versatile and capable of handling both very low and high concentrations of solutes in aqueous systems and other solvents. The first step was to determine the optimum temperatures for flash separation and condensation, to handle a tonne of digestate per hour with typical composition of 2 g/L of NH 3 and 3 g/L of CO 2 [22]. This processing capacity was selected based on the subsidies available for covering 40% of the total cost of the equipment dealing with slurry separation [24]; hence, a wider adoption by the stakeholders of agroindustry can be expected. The synthesis of NH 4 HCO 3 was subsequently confirmed by a parametric study and the minimum feasible conditions (i.e., concentration of NH 3 and CO 2 and temperature of the flash distillation) were readjusted. In order to select the most suitable types of anaerobic digestates for the manufacturing of NH 4 HCO 3 , the minimum concentrations of NH 3 and CO 2 were matched to streams coming out of the anaerobic digestion of amino acids. These amino acids were set as the lower limit for the composition of the feedstock employed for the anaerobic digestion, as the presence of any other type of molecule would result in a diluted stream with lower concentrations of inorganic nitrogen and carbon, which would constrain the production of NH 4 HCO 3 by the flash distillation process. The stoichiometry and the kinetics for amino acid anaerobic fermentation were mostly taken from the previous investigation by Rajendran et al. [19], except for the fermentation of glutamine (Equation (1)) which was proposed in a similar fashion to the degradation of the other amino acids. The production of oxygen in some of these reactions implies that they are only spontaneous under strongly reducing conditions and would be very limited under aerobic conditions (Le Chatelier's principle). It is important to mention that the type of model that was used for the AD was unstructured and unsegregated, which implies that it did not involve the metabolism of the microorganisms nor differentiate the species that degraded the biomass [25].
As in the simulation by Rajendran et al. [19], thermophilic conditions (55 • C) were considered to simulate the AD of amino acids. The hydrolysis of all amino acids followed the first-order reaction, with the kinetic constant (s −1 ) described in Equation (23) The subsequent methanogenic stage [26] was also included as part of the anaerobic fermentation of all amino acids by considering the stoichiometry of Equations (24) and (25) and the first-order kinetic constant (s −1 ) following Equation (26) with the temperature in K [19]. Equation (25) was only considered if there was production of hydrogen in the previous acetogenic stage of degradation of amino acids. The modelling of the fermentation was confirmed by monitoring a constant value for Henry's parameter of each volatile compound.
Finally, a comprehensive model was developed to confirm the synthesis of NH 4 HCO 3 with particular types of feedstocks for AD. The model included the stages of amino acid fermentation and extraction of NH 4 HCO 3 for stabilization of the resulting anaerobic digestate ( Figure 1). The process engineering software employed to develop the model allows consideration of a single stream with liquid and gas phases coming out of the fermenter. This concept is representative of real conditions when treating fresh anaerobic digestate saturated with biogas. The titrations of amino acid digestates were investigated by adding NaOH or HCl before the flash distillation, with the purpose of tuning the ratio of NH 4 + to HCO 3 − in the distillate. The equilibrium-driven reactions that were considered for the modelling of the multiphase system NH 3 -CO 2 -H 2 O are shown in Table 1. A discussion is offered on which other parameters would need to be considered, in addition to the molar ratio of inorganic nitrogen (NH 4 + +NH 3 +NH 2 COO − ) to inorganic carbon (CO 2 +HCO 3 − +CO 3 2− +NH 2 COO − ) of the anaerobic digestate, to facilitate the stabilization of this organic material by means of isolating the NH 4 HCO 3 . The 3 streams coming out of this synergistic process ( Figure 1) are the upgraded biogas, the distillate (i.e., saturated NH 4 HCO 3 aqueous solution), and the stabilized digestate. Finally, a comprehensive model was developed to confirm the synthesis of NH4HCO3 with particular types of feedstocks for AD. The model included the stages of amino acid fermentation and extraction of NH4HCO3 for stabilization of the resulting anaerobic digestate ( Figure 1). The process engineering software employed to develop the model allows consideration of a single stream with liquid and gas phases coming out of the fermenter. This concept is representative of real conditions when treating fresh anaerobic digestate saturated with biogas. The titrations of amino acid digestates were investigated by adding NaOH or HCl before the flash distillation, with the purpose of tuning the ratio of NH4 + to HCO3 − in the distillate. The equilibrium-driven reactions that were considered for the modelling of the multiphase system NH3-CO2-H2O are shown in Table 1. A discussion is offered on which other parameters would need to be considered, in addition to the molar ratio of inorganic nitrogen (NH4 + +NH3+NH2COO − ) to inorganic carbon (CO2+HCO3 − +CO3 2− +NH2COO − ) of the anaerobic digestate, to facilitate the stabilization of this organic material by means of isolating the NH4HCO3. The 3 streams coming out of this synergistic process ( Figure 1) are the upgraded biogas, the distillate (i.e., saturated NH4HCO3 aqueous solution), and the stabilized digestate.   Figure 2a shows that the results of the simulation follow the trends of the experimental data of Goppert and Maurer (1988), which were reported by Darde [27] in Figure   Figure 1. Process flow diagram with capacity to handle 1 t of feedstock (blend of amino acid and water) per h, to evaluate the feasibility of producing NH 4 HCO 3 from different types of anaerobic digestates.  Figure 2a shows that the results of the simulation follow the trends of the experimental data of Goppert and Maurer (1988), which were reported by Darde [27] in Figures 5-12 and Figures 5-13. The greatest volatility of CO 2 corresponds to the greater partial pressure (at least two orders of magnitude greater) than that exerted by NH 3 for any of the compositions of the anaerobic digestate investigated at the bubble point, as shown in Figure 2a. The chemistry of this blend also plays a crucial role in determining the volatility of these compounds, as partial pressure of the CO 2 started to increase at a greater rate when the moles of this compound in the anaerobic digestate were greater than the moles of NH 3 (Figure 2a), which was set to 0.6 mol/kg H 2 O for the whole test to comply with the experimental data available in Figures 5-12 and Figures 5-13 of Darde [27]. Typical concentrations of CO 2 and NH 3 in the anaerobic digestate are around 0.12 mol NH 3 /kg H 2 O and 0.07 mol CO 2 /kg H 2 O [22], which are milder concentrations than those tested by Darde et al. [28] for the development of a carbon capture process. However, given the wide scope of the ELECNRTL property method, the validation could be performed at the lower end of the concentration investigated by Darde et al. [28]. Figure 2b shows the calibration carried out with a concentration of 0.13 mol NH 3 /kg H 2 O with the experimental data of Pexton and Badger (1938), which is reported in Figures 2 and 3 of Darde et al. [28]. Figure 2c validates the precipitation of the NH 4 HCO 3 , which is less soluble than the NH 4 COONH 2 and has similar solubility to NH 4 Cl and higher solubility than NaHCO 3 . Contrary to the explanation by Möller and Müller [8] on the formation of ammonium carbonate in the anaerobic digestate, Aspen Plus ® v12 ELECNRTL does not consider the formation of this compound. Modelling the solubilities of NH 4 Cl and NaHCO 3 was considered because the HCl and NaOH were tested as titrants of the anaerobic digestate to assist the flash distillation, tune the ratio of inorganic nitrogen to inorganic carbon in the distillate, and promote the precipitation of NH 4 HCO 3 . Therefore, NH 4 Cl and NaHCO 3 might precipitate in the residual stabilized digestate and they are not expected in the distillate despite having solubility similar to or lower than NH 4 HCO 3 . The big difference in the solubility of NH 4 HCO 3 modelled above 60 • C with regard to the experimental data available in the literature [29,30] is because this compound is not very stable and easily undergoes decomposition [31]. Figure 2d shows the modelling of the titration of the manure digestate resulting from the process described by Rajendran et al. [19] and considering some of the alkaline elements identified in the comprehensive characterization of various anaerobic digestates performed by Akhiar et al. [22]: NaHCO 3 , CaCl 2 , NaCl, and KHCO 3 (Table A1). The validation of the simulation of the effect of HCl and NaOH on the proposed anaerobic digestate was achieved by comparison with experimental data of titrations available in the literature, particularly the study by Vandré and Clemens [32] of raw animal slurry and the previous work of Moure Abelenda et al. [23] with agrowaste digestate and food waste digestate. The results showed that the amount of CO 2 and carbonates in the anaerobic digestate was responsible for the OH alkalinity at pH > 10. The content of ammonia was found to be the main cause of the partial (P) alkalinity and the total (M) alkalinity. This means that the volatile fatty acids (i.e., acetic acid and propionic acid) and other components (e.g., hydrogen sulfide; Table A1) previously reported with a role in regulating the pH of the anaerobic digestate [33] did not have a significant buffer effect in the present model. As can be seen in Figure A1, simply tuning of the ratio CO 2 to NH 3 in the digestate affects the pH, which shows the how the pH depends on the composition of the anaerobic digestates and the remaining stabilized digestates (Figure 1) after the flash separation indicated in Figure 2b.  [28]. (c) Validation of the solid formation by comparing the results of the simulation with the experimental data reported by Mullin [29], Green and Perry [34], and Haynes [29]. (d) Validation of the processing of manure digestate [19,22] with the addition of NaOH and HCl by comparison with the empirical titration of agrowaste digestate, food waste digestate [23], and animal slurry [32].

Optimization of the Conditions of the Flash Distillation to Produce NH4HCO3
In the initial development stage of the flash distillation process for stabilization of anaerobic digestate and manufacturing of NH4HCO3, the target processing capacity was set at 1 t per h (i.e., proposed organic slurry with a composition of 55 kmol H2O/h, 0.1 kmol NH3/h, and 0.1 kmol CO2/h). The optimum temperatures for evaporation and condensation were investigated with this simplified composition of anaerobic digestate. Figure 3a shows the volatilization of NH3, CO2, and H2O as a function of the temperature. In order to appreciate significant release of CO2 from the anaerobic digestate, the temperature of the flash tank should be greater than 75 °C. There is not a big difference (i.e., around 0.2 °C) between the minimum temperature to obtain the maximum volatilization of NH3 and that of H2O (Figure 3a). However, it is this small difference that needs to be taken into account in the next stages of the design of the process to maximize the overall heat integration and profitability of the flash distillation process ( Figure A2). Figure 3bpresents the concentrations of the different species in the distillate as a function of the condensation temperature. The condensation at a temperature below 70 °C maximizes the concentration of NH4 + and HCO3 − (Figure 2b), which is necessary for the precipitation of NH4HCO3 when supersaturation is attained in the aqueous solution. In order to precipitate NH4HCO3 in  [28]. (c) Validation of the solid formation by comparing the results of the simulation with the experimental data reported by Mullin [29], Green and Perry [34], and Haynes [29]. (d) Validation of the processing of manure digestate [19,22] with the addition of NaOH and HCl by comparison with the empirical titration of agrowaste digestate, food waste digestate [23], and animal slurry [32].

Optimization of the Conditions of the Flash Distillation to Produce NH 4 HCO 3
In the initial development stage of the flash distillation process for stabilization of anaerobic digestate and manufacturing of NH 4 HCO 3 , the target processing capacity was set at 1 t per h (i.e., proposed organic slurry with a composition of 55 kmol H 2 O/h, 0.1 kmol NH 3 /h, and 0.1 kmol CO 2 /h). The optimum temperatures for evaporation and condensation were investigated with this simplified composition of anaerobic digestate. Figure 3a shows the volatilization of NH 3 , CO 2 , and H 2 O as a function of the temperature. In order to appreciate significant release of CO 2 from the anaerobic digestate, the temperature of the flash tank should be greater than 75 • C. There is not a big difference (i.e., around 0.2 • C) between the minimum temperature to obtain the maximum volatilization of NH 3 and that of H 2 O (Figure 3a). However, it is this small difference that needs to be taken into account in the next stages of the design of the process to maximize the overall heat integration and profitability of the flash distillation process ( Figure A2). Figure 3bpresents the concentrations of the different species in the distillate as a function of the condensation temperature. The condensation at a temperature below 70 • C maximizes the concentration of NH 4 + and HCO 3 − (Figure 2b), which is necessary for the precipitation of NH 4 HCO 3 when supersaturation is attained in the aqueous solution. In order to precipitate NH 4 HCO 3 in the distillate cooled at 1 bar and 25 • C, the minimum temperature of the flash tank should be between 85 and 95 • C and the composition of the anaerobic digestate was found to be around 10 g NH 3 /L and 13 g CO 2 /L. This corresponds to a stream of anaerobic digestate fed to the flash distillation process at a rate of 55 kmol H 2 O/h, 0.6 kmol NH 3 /h, and 0.3 kmol CO 2 /h. Under these conditions, the maximum production of NH 4 HCO 3 at a rate of 35.4 mol/h or 2.8 kg/h was found when the temperature of the flash tank was 95 • C ( Figure A2). the distillate cooled at 1 bar and 25 °C, the minimum temperature of the flash tank should be between 85 and 95 °C and the composition of the anaerobic digestate was found to be around 10 g NH3/L and 13 g CO2/L. This corresponds to a stream of anaerobic digestate fed to the flash distillation process at a rate of 55 kmol H2O/h, 0.6 kmol NH3/h, and 0.3 kmol CO2/h. Under these conditions, the maximum production of NH4HCO3 at a rate of 35.4 mol/h or 2.8 kg/h was found when the temperature of the flash tank was 95 °C ( Figure  A2).

Anaerobic Fermentation of Amino Acids
As this minimum composition of digestate to attain stabilization via manufacturing of NH4HCO3 is not in agreement with the typical composition of anaerobic digestate [22], the next step in defining the specifications of the flash process was to find a source of anaerobic digestate that provides a concentration of 10 g NH3/L and 13 g CO2/L. Since amino acids are ultimately responsible for the content of NH4 + -N in the anaerobic digestate, the composition of the feedstock of AD was investigated by starting with these organic molecules. Although this is not a realistic approach because AD is a technology for handling complex matrices of organic waste materials, this investigation with process engineering software could give an idea of the type of residues that are more suitable for stabilizing the resulting anaerobic digestate with NH4HCO3 extraction. The concentration profile of all chemical species was elucidated for the fermentation of 18 amino acids (Figure A3), considering the stoichiometry and kinetics defined by Rajendran et al. [19], which are shown in Equations (1)- (26). In all cases, the feed of the anaerobic digester was considered to have 90 wt.% moisture [35] and the remaining 10 wt.% of the feedstock corresponded to the mass of amino acids. The discontinuous fermenters were operated under thermophilic conditions (i.e., 55 °C and 1 bar) for 1000 hrs (i.e., 42 days) in order to be able to observe the constant profiles of all chemical species, with the exception of glutamic acid that required a longer residence time (Figures A3h and S8). For the present theoretical work, the hydrolysis reactions that would break down proteins into amino acids were omitted and the model only included the acidogenic, acetogenic, and methanogenic reactions. As the fermentation kinetics proposed by Rajendran et al. [19] were implemented, it was not necessary to define a highly active inoculum for the calculation block of the bioreactor to function properly. Most amino acids required at least 500 hrs in the bioreactor, with the exception of cysteine ( Figure A3n) that was rapidly converted in less than 100 hrs. This simplified model corresponds with the first steps in elucidating the extent to which the feasibility of the NH4HCO3 stabilization of anaerobic digestate is limited by the

Anaerobic Fermentation of Amino Acids
As this minimum composition of digestate to attain stabilization via manufacturing of NH 4 HCO 3 is not in agreement with the typical composition of anaerobic digestate [22], the next step in defining the specifications of the flash process was to find a source of anaerobic digestate that provides a concentration of 10 g NH 3 /L and 13 g CO 2 /L. Since amino acids are ultimately responsible for the content of NH 4 + -N in the anaerobic digestate, the composition of the feedstock of AD was investigated by starting with these organic molecules. Although this is not a realistic approach because AD is a technology for handling complex matrices of organic waste materials, this investigation with process engineering software could give an idea of the type of residues that are more suitable for stabilizing the resulting anaerobic digestate with NH 4 HCO 3 extraction. The concentration profile of all chemical species was elucidated for the fermentation of 18 amino acids ( Figure A3), considering the stoichiometry and kinetics defined by Rajendran et al. [19], which are shown in Equations (1)- (26). In all cases, the feed of the anaerobic digester was considered to have 90 wt.% moisture [35] and the remaining 10 wt.% of the feedstock corresponded to the mass of amino acids. The discontinuous fermenters were operated under thermophilic conditions (i.e., 55 • C and 1 bar) for 1000 h (i.e., 42 days) in order to be able to observe the constant profiles of all chemical species, with the exception of glutamic acid that required a longer residence time ( Figure A3h and Figure S8). For the present theoretical work, the hydrolysis reactions that would break down proteins into amino acids were omitted and the model only included the acidogenic, acetogenic, and methanogenic reactions. As the fermentation kinetics proposed by Rajendran et al. [19] were implemented, it was not necessary to define a highly active inoculum for the calculation block of the bioreactor to function properly. Most amino acids required at least 500 h in the bioreactor, with the exception of cysteine ( Figure A3n) that was rapidly converted in less than 100 h. This simplified model corresponds with the first steps in elucidating the extent to which the feasibility of the NH 4 HCO 3 stabilization of anaerobic digestate is limited by the type of feedstock for AD. The concentrations of inorganic nitrogen and inorganic carbon that were found in the anaerobic digestates of amino acids ( Figure 4) were similar to those reported by Akhiar et al. [22]. Fermentation 2023, 9, x FOR PEER REVIEW 9 of 20 type of feedstock for AD. The concentrations of inorganic nitrogen and inorganic carbon that were found in the anaerobic digestates of amino acids ( Figure 4) were similar to those reported by Akhiar et al. [22].

Suitable Conditions for NH4HCO3-Stabilization of Anaerobic Digestate
As was expected according to the molecular formulas and stoichiometries of the amino acids' fermentations, arginine provided the greatest amount of mineralized nitrogen (Figure 4), but this corresponded to only 0.6 g/L. The arginine digestate was also considered in the study by Drapanauskaite et al. [13]. This means that it would be necessary to consider at least a feedstock for AD with 3.5 times less water in relation to arginine for the development of the process. In this way, the overall process of fermentation and stabilization via synthesis of NH4HCO3 (Figure 1) was found to be feasible for arginine (C6H14N4O), glycine (C2H5NO2), and histidine (C6H9N3O2) only when the feedstocks of these amino acids were prepared with a moisture content lower than 90 wt.%. For

Suitable Conditions for NH 4 HCO 3 -Stabilization of Anaerobic Digestate
As was expected according to the molecular formulas and stoichiometries of the amino acids' fermentations, arginine provided the greatest amount of mineralized nitrogen (Figure 4), but this corresponded to only 0.6 g/L. The arginine digestate was also considered in the study by Drapanauskaite et al. [13]. This means that it would be necessary to consider at least a feedstock for AD with 3.5 times less water in relation to arginine for the development of the process. In this way, the overall process of fermentation and stabilization via synthesis of NH 4 HCO 3 (Figure 1) was found to be feasible for arginine (C 6 H 14 N 4 O), glycine (C 2 H 5 NO 2 ), and histidine (C 6 H 9 N 3 O 2 ) only when the feedstocks of these amino acids were prepared with a moisture content lower than 90 wt.%. For example, NH 4 HCO 3 production rates were 15.16, 3.00, and 1.82 kg/h when processing 1 tonne/h of digestates resulting from the individual AD of arginine, glycine, and histidine with 50 wt.% moisture. The moisture content of the digestate was estimated considering the initial composition of the feedstock to be fermented. The results of the titrations showed that the extent of stabilization of the anaerobic digestates coming from these amino acids can be greatly improved via acidification (in the case of arginine digestate; Figure 5a) and alkalization (for glycine and histidine; Figure 5b). The rate of addition of HCl was not enough to deplete the partial and total alkalinities of arginine, glycine, and histidine digestates (Figure 5a), as the pH never dropped below a level of 7 (Figure 2d). Similarly, the addition of NaOH was not enough to overcome the OH alkalinity (Figure 2d), except for the alkalization of histidine digestate which reached a pH above 9.5 (Figure 5b). The greatest NH 4 HCO 3 extraction rates were 22.71 kg/h from arginine digestate (50 wt.% moisture) with 20 kg HCl/h (Figure 5a), 22.08 kg/h from glycine digestate (50 wt.% moisture) with 31 kg NaOH/h, and 34.18 kg/h from histidine digestate (50 wt.% moisture) with 44 kg NaOH/h (Figure 5b). The detrimental effect that acidification of glycine digestate and histidine digestate have on the production rate of NH 4 HCO 3 should be noted (Figure 5a). Similarly, any addition of NaOH to the arginine digestate reduces the amount of NH 4 HCO 3 that can be recovered. This is important to consider when elaborating a more complex digestate (i.e., coming from the blend of several amino acids). The maximum tolerances observed for the anaerobic digestates with 50 wt.% moisture were 27 kg NaOH/h applied to arginine digestate yielding 327.96 g NH 4 HCO 3 /h, 4 kg HCl/h applied to glycine digestate yielding 245.54 g NH 4 HCO 3 /h, and 2 kg HCl/h applied to histidine digestate yielding 146.34 g NH 4 HCO 3 /h. However, these trends changed when anaerobic digestates with greater moisture content were considered for the NH 4 HCO 3 stabilization. The highest moisture content of amino acid digestates that was suitable for production of NH 4 HCO 3 was dependent on the dose of acid and base. NaOH was a more effective titrant than HCl in the case of arginine digestate (>50 wt.% moisture), and HCl was a more effective titrant than NaOH for cysteine and histidine digestates (>50 wt.% moisture). The models developed in Aspen Plus ® v12 ELECNRTL show the feasibility of (a) extraction of 142. 30  example, NH4HCO3 production rates were 15.16, 3.00, and 1.82 kg/h when processing 1 tonne/h of digestates resulting from the individual AD of arginine, glycine, and histidine with 50 wt.% moisture. The moisture content of the digestate was estimated considering the initial composition of the feedstock to be fermented. The results of the titrations showed that the extent of stabilization of the anaerobic digestates coming from these amino acids can be greatly improved via acidification (in the case of arginine digestate; Figure 5a) and alkalization (for glycine and histidine; Figure 5b). The rate of addition of HCl was not enough to deplete the partial and total alkalinities of arginine, glycine, and histidine digestates (Figure 5a), as the pH never dropped below a level of 7 (Figure 2d).
Similarly, the addition of NaOH was not enough to overcome the OH alkalinity ( Figure  2d), except for the alkalization of histidine digestate which reached a pH above 9.5 ( Figure  5b). The greatest NH4HCO3 extraction rates were 22.71 kg/h from arginine digestate (50 wt.% moisture) with 20 kg HCl/h (Figure 5a) (Figure 5b). The detrimental effect that acidification of glycine digestate and histidine digestate have on the production rate of NH4HCO3 should be noted (Figure 5a). Similarly, any addition of NaOH to the arginine digestate reduces the amount of NH4HCO3 that can be recovered. This is important to consider when elaborating a more complex digestate (i.e., coming from the blend of several amino acids). The maximum tolerances observed for the anaerobic digestates with 50 wt.% moisture were 27 kg NaOH/h applied to arginine digestate yielding 327.96 g NH4HCO3/h, 4 kg HCl/h applied to glycine digestate yielding 245.54 g NH4HCO3/h, and 2 kg HCl/h applied to histidine digestate yielding 146.34 g NH4HCO3/h. However, these trends changed when anaerobic digestates with greater moisture content were considered for the NH4HCO3 stabilization. The highest moisture content of amino acid digestates that was suitable for production of NH4HCO3 was dependent on the dose of acid and base. NaOH was a more effective titrant than HCl in the case of arginine digestate (>50 wt.% moisture), and HCl was a more effective titrant than NaOH for cysteine and histidine digestates (>50 wt.% moisture). The models developed in Aspen Plus ® v12 ELECNRTL show the feasibility of (a) extraction of 142. 30    The square symbols represent the pH of the stabilized digestates that leave the bottom of the flash tank ( Figure 1).

Discussion
The present study questions the reliability of measurements of nitrogen in anaerobic digestate [22,36,37], since stoichiometric calculations show that the nitrogen content is expected to be lower than 0.63 g/L when moisture is 90 wt.% ( Figure A3). Modeling of AD resulted in almost the complete consumption of amino acids by the end of the 42 days of residence time ( Figure A3), and this could be regarded as a very favorable scenario for the production of white ammonia [38]. However, the content of organic nitrogen in anaerobic digestate is never less than 30% of the total nitrogen [22]. Still considering all the conversion of organic nitrogen to inorganic forms, the amino acid digestate with 90 wt.% moisture was found not to be enough for stabilization via synthesis of NH 4 HCO 3 . It was necessary to reduce the moisture content of the feedstock of AD to 50 wt.% to enable the synthesis of NH 4 HCO 3 from arginine digestate, glycine digestate, and histidine digestate. The fermentation of alanine (C 3 H 7 NO 2 ; Figure A3o) and that of glutamine (C 5 H 10 N 2 O 3 ; Figure A3r) also provided anaerobic digestates (50 wt.% moisture) with sufficient concentration of inorganic nitrogen and inorganic carbon (Figure 4), but the manufacturing of NH 4 HCO 3 was not possible in these cases. The threshold values 0.02 mol N/L and 0.005 mol C/L, which were established based on the profile of glycine (Figure 4), explain why NH 4 HCO 3 valorization was not attained with proline (C 5 H 9 NO 2 ) and lysine (C 6 H 14 N 2 O 2 ), despite their anaerobic digestates having the greatest inorganic nitrogen to inorganic carbon ratio. The inorganic forms of nitrogen are more readily available for plants, since these compounds do not need to be metabolized by the microbes in the rhizosphere, unlike organic nitrogen. These inorganic forms are also more prone to be lost via leaching and volatilization, hence the digestate is less stable and does not behave as controlled-release fertilizer. Traditionally, acidification has been used to prevent ammonia emissions from organic manure and slurry during storage and land application, because it reduces the pH and the NH 3 can be kept in the aqueous solution as NH 4 + [39]. However, in the approach of stabilizing the anaerobic digestate by NH 4 HCO 3 extraction, the only purpose of using acids or alkalis is to promote the formation of a supersaturated distillate from where the crystals of the chemical grade fertilizer can be easily harvested. Hence, the doses of these endogenous agents to tune the pH of the anaerobic digestate need to be optimized accordingly, as long as NH 4 HCO 3 is isolated.
A difference between the alanine digestate and glutamine digestate and the arginine digestate, glycine digestate, and histidine digestate is the presence of oxygen among the volatile compounds ( Figure 6). Looking at the volatility of these gases, the use of the components of the residual biogas as stripping agents is appropriate because the volatility of H 2 (129.87·10 3 Pa·m 3 ·mol −1 ) and O 2 (76.92·10 3 Pa·m 3 ·mol −1 ) is greater than that of NH 3 (1.69 Pa·m 3 ·mol −1 ) and CO 2 (3.03·10 3 Pa·m 3 ·mol −1 ). The effect of these endogenous stripping agents could be the reason that the manufacturing of NH 4 HCO 3 was possible with arginine, glycine, and histidine digestates, even when these had lower concentrations than 0.5 mol NH 3 /L and 0.3 mol CO 2 /L ( Figure A2). The most intuitive reason for the lack of formation of NH 4 HCO 3 in the distillate of the alanine and glutamine digestates (50 wt.% moisture) could be the low flowrate of NH 3 in the stream of volatiles leaving the top of the flash tank at 95 • C (0.06 kmol NH 3 /h; Figure 6a). From a broader perspective, there are resemblances among the H 2 O and NH 3 trends of the streams of volatile elements leaving the top of the flash tank during the processing of the five untreated amino acid digestates with 50% moisture (Figure 6a). Together with the pH of the digestate, this could be an explanation for why the volatilization of NH 3 was the greatest in the arginine digestate ( Figure 6a) and more NH 4 + left the bottom of the flash tank when processing glycine, histidine, and glutamine digestates (Figure 6b). The pHs that were found in the stabilized digestates (50 wt.% moisture) leaving the bottom of the flash tank ( Figure 1) were: 7.98 ± 0.52 (arginine); 7.53 ± 0.65 (glycine); 7.61 ± 0.59 (histidine); 7.12 ± 0.23 (alanine); and 7.14 ± 1.42 (glutamine). According to Ukwuani and Tao [40], at the high pH employed for NH3 stripping, H2S and VFAs are largely dissociated in the aqueous phase and render little volatilization. This could explain the fact that Ukwuani and Tao [40] did not find acetic acid in the sulfuric acid solution upon absorption of the stripped ammonia under vacuum pressure at different boiling-point temperatures. However, other volatile organic compounds (VOCs) can end up in the distillate stream depending on the operating conditions. For example, working at low pressures (i.e., under vacuum conditions) contributes more than operating at high temperatures to increasing emissions of VOCs [40]. Ukwuani and Tao [40] reported cyclohexene, 4-methylphenol, 4-ethylphenol, trichloromethane, and (p-hydroxyphenyl)phosphoric acid as the most common VOCs in the acid NH3 absorbent solution. The analysis of volatility that Ukwuani and Tao [40] carried out was based primarily on the concentration of these compounds in the absorbent solution at the different boiling points of the anaerobic digestate (ranging from 50 to 100 °C) under vacuum conditions, and secondly on the vapor pressure of the compound at 25 °C. Only for cyclohexene and chloroform are Henry's law volatility constants greater than that of ammonia. It is important to highlight that Henry's volatility constants are often reported as the ratio of vapor pressure and aqueous solubility [15]. Without considering the much greater solubility of NH3 in the aqueous solution, since the vapor pressure of NH3 (1002 kPa at 25 °C) is much greater than that of cyclohexane and chloroform (12 and 26 kPa, respectively), the VOC mass transfer is much slower. Cyclohexene (3.45·10 3 Pa·m 3 ·mol −1 ) has a volatility even greater than that of carbon dioxide but lower than that of methane. Ukwuani and Tao [40] explained the presence of (p-hydroxyphenyl)-phosphoric acid in the absorbent solution by the excessive foam formation in the anaerobic digestate at a pH 9 in such a way that the foam reached the sulfuric acid solution. This was the reason that Drapanauskaite et al. [13] used defoaming agents based on silicon for the operation of distillation and stripping processes. Hence, in addition to the titrant, the need for an antifoaming agent needs to be assessed experimentally [35,[41][42][43][44]. Alternatively, the use of exogenous stripping agents, such as nitrogen (156.25·10 3 Pa·m 3 ·mol −1 ) of air, which has even greater volatility than oxygen, is interesting but makes more difficult the subsequent upgrading of the biogas [45], as this gaseous stream becomes diluted. Ideally, the synthesis of NH4HCO3 will be coupled with biogas upgrading [46], but the purity of the crystals according to the specifications of the chemical fertilizers needs to be confirmed [40]. The kinetic reactions detailed in Table 2, which were proposed by AspenTech [47] to develop a rate-based model of CO2 According to Ukwuani and Tao [40], at the high pH employed for NH 3 stripping, H 2 S and VFAs are largely dissociated in the aqueous phase and render little volatilization. This could explain the fact that Ukwuani and Tao [40] did not find acetic acid in the sulfuric acid solution upon absorption of the stripped ammonia under vacuum pressure at different boiling-point temperatures. However, other volatile organic compounds (VOCs) can end up in the distillate stream depending on the operating conditions. For example, working at low pressures (i.e., under vacuum conditions) contributes more than operating at high temperatures to increasing emissions of VOCs [40]. Ukwuani and Tao [40] reported cyclohexene, 4-methylphenol, 4-ethylphenol, trichloromethane, and (p-hydroxyphenyl)phosphoric acid as the most common VOCs in the acid NH 3 absorbent solution. The analysis of volatility that Ukwuani and Tao [40] carried out was based primarily on the concentration of these compounds in the absorbent solution at the different boiling points of the anaerobic digestate (ranging from 50 to 100 • C) under vacuum conditions, and secondly on the vapor pressure of the compound at 25 • C. Only for cyclohexene and chloroform are Henry's law volatility constants greater than that of ammonia. It is important to highlight that Henry's volatility constants are often reported as the ratio of vapor pressure and aqueous solubility [15]. Without considering the much greater solubility of NH 3 in the aqueous solution, since the vapor pressure of NH 3 (1002 kPa at 25 • C) is much greater than that of cyclohexane and chloroform (12 and 26 kPa, respectively), the VOC mass transfer is much slower. Cyclohexene (3.45·10 3 Pa·m 3 ·mol −1 ) has a volatility even greater than that of carbon dioxide but lower than that of methane. Ukwuani and Tao [40] explained the presence of (p-hydroxyphenyl)-phosphoric acid in the absorbent solution by the excessive foam formation in the anaerobic digestate at a pH 9 in such a way that the foam reached the sulfuric acid solution. This was the reason that Drapanauskaite et al. [13] used defoaming agents based on silicon for the operation of distillation and stripping processes. Hence, in addition to the titrant, the need for an antifoaming agent needs to be assessed experimentally [35,[41][42][43][44]. Alternatively, the use of exogenous stripping agents, such as nitrogen (156.25·10 3 Pa·m 3 ·mol −1 ) of air, which has even greater volatility than oxygen, is interesting but makes more difficult the subsequent upgrading of the biogas [45], as this gaseous stream becomes diluted. Ideally, the synthesis of NH 4 HCO 3 will be coupled with biogas upgrading [46], but the purity of the crystals according to the specifications of the chemical fertilizers needs to be confirmed [40]. The kinetic reactions detailed in Table 2, which were proposed by AspenTech [47] to develop a rate-based model of CO 2 capture process with NH 3 , could also be implemented in the present model to enhance the predictions of speciation in the NH 3 -CO 2 -H 2 O system.

Conclusions
The present study confirms the suitability of the conditions of the flash distillation (95 • C at 1 bar) for extracting NH 4 HCO 3 from anaerobic digestates produced by the fermentation of arginine, glycine, and histidine. The use of HCl and NaOH was found necessary to maximize the stabilization of anaerobic digestates, although the rates of application of these titrants need to be further assessed from an economic point of view as well. This investigation elucidated an important role carried out by the most volatile compounds of biogas in the process of stabilization of anaerobic digestate. Therefore, it is recommended to carry out NH 4 HCO 3 synthesis with anaerobic digestates saturated in biogas to maximize the endogenous stripping effect during the flash distillation. This process needs to be performed at the beginning of the storage of anerobic digestate to minimize the loss of the residual biogas.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/fermentation9080750/s1, Figure S1: Concentration profile of the species of inorganic nitrogen and inorganic carbon during the fermentation of glycine; Figure S2: Concentration profile of the species of inorganic nitrogen and inorganic carbon during the fermentation of threonine; Figure S3: Concentration profile of the species of inorganic nitrogen and inorganic carbon during the fermentation of arginine; Figure S4: Concentration profile of the species of inorganic nitrogen and inorganic carbon during the fermentation of proline; Figure S5: Concentration profile of the species of inorganic nitrogen and inorganic carbon during the fermentation of methionine; Figure S6: Concentration profile of the species of inorganic nitrogen and inorganic carbon during the fermentation of serine; Figure S7: Concentration profile of the species of inorganic nitrogen and inorganic carbon during the fermentation of aspartic acid; Figure S8: Concentration profile of the species of inorganic nitrogen and inorganic carbon during the fermentation of glutamic acid; Figure S9: Concentration profile of the species of inorganic nitrogen and inorganic carbon during the fermentation of lysine; Figure S10: Concentration profile of the species of inorganic nitrogen and inorganic carbon during the fermentation of leucine; Figure S11: Concentration profile of the species of inorganic nitrogen and inorganic carbon during the fermentation of valine; Figure S12: Concentration profile of the species of inorganic nitrogen and inorganic carbon during the fermentation of phenylalanine; Figure S13: Concentration profile of the species of inorganic nitrogen and inorganic carbon during the fermentation of tyrosine; Figure S14: Concentration profile of the species of inorganic nitrogen and inorganic carbon during the fermentation of cysteine; Figure S15: Concentration profile of the species of inorganic nitrogen and inorganic carbon during the fermentation of alanine; Figure S16: Concentration profile of the species of inorganic nitrogen and inorganic carbon during the fermentation of histidine; Figure S17: Concentration profile of the species of inorganic nitrogen and inorganic carbon during the fermentation of tryptophan; Figure S18: Concentration profile of the species of inorganic nitrogen and inorganic carbon during the fermentation of glutamine. The models of titration, anaerobic digestion, and subsequent flash distillation process for amino acids can be found at: https://zenodo.org/record/7738947 (accessed on 29 July 2023).       (Figures S1-S18).