Inﬂuence of Leachate and Nitrifying Bacteria on Photosynthetic Biogas Upgrading in a Two-Stage System

: Photosynthetic biogas upgrading using two-stage systems allows the absorption of carbon dioxide (CO 2 ) in an absorption unit and its subsequent assimilation by microalgae. The production of microalgae requires large amounts of nutrients, thus making scale-up difﬁcult and reducing economic feasibility. The photosynthetic process produces oxygen (O 2 ) (1 mol per mol of CO 2 consumed), which can be desorbed into puriﬁed biogas. Two-stage systems reduce its impact but do not eliminate it. In this study, we test the use of landﬁll leachate as a nutrient source and propose a viable and economical strategy for reducing the O 2 concentration. First, the liquid/gas ( L / G ) ratio and ﬂow mode of the absorber were optimized for 20% and 40% CO 2 with COMBO medium, then landﬁll leachate was used as a nutrient source. Finally, the system was inoculated with nitrifying bacteria. Leachate was found to be suitable as a nutrient source and to result in a signiﬁcant improvement in CO 2 absorption, with outlet concentrations of 0.01% and 0.6% for 20% and 40% CO 2 , respectively, being obtained. The use of nitrifying bacteria allowed a reduction in dissolved oxygen (DO) concentration, although it also resulted in a lower pH, thus making CO 2 uptake slightly more difﬁcult.


Introduction
The use of biogas as a renewable energy source is strongly encouraged by international organizations and states. In this sense, the European Union has established an objective of reducing total greenhouse gas emissions by between 80% and 95% compared to 1990 [1]. Within the intermediate objectives in the framework on climate and energy for 2030, an estimated reduction in these emissions of at least 40% has been proposed [1]. In its report "A roadmap for moving to a competitive low carbon economy in 2050", the EU has established a target of an 80% reduction in emissions by 2050 [2]. Methane (CH 4 ) and carbon dioxide (CO 2 ) are the main gases present in biogas, with the potential effect of CH 4 on global warming being 24.5times higher than that of CO 2 . To obtain biomethane, according to the standard specifications of each region or country, biogas must be purified to eliminate minor compounds and then upgraded to vary the methane content [2].
The upgrading of biogas mainly reduces the CO 2 content, which is performed using physical, chemical and/or biological processes [3,4]. Of these, CO 2 assimilation by microalgae is a rapidly growing technology. The photosynthetic upgrading of biogas can be achieved byusing one-stage [5,6] or two-stage systems [7][8][9][10]. In one-stage systems, the biogas is fed into a photobioreactor, where CO 2 is absorbed and assimilated by microalgae as a carbon source. In these systems, the oxygen (O 2 ) produced during photosynthesis can increase in concentration in the outlet biogas stream, with the CO 2 consumed mostly vvm), when required, for growth or maintenance of the microalgae. A mass flow controller (F-201 CV, Bronkhorst High-Tech B.V., Ruurlo, TheNetherlands) was used to fix the inlet CO2 concentration and a variable area flow meter (FR2A12BVBN, Key Instruments, Tevose, PA, USA) was used to measure the air flow rate. The pH and DO were monitored (Multimeter M44, CrisonIntruments S.A., Barcelona, Spain). The photobioreactor was illuminated with 2 LED panels (120 × 60 cm, 72W, 2880-3200, Lifud, Shenzhen, China) with a photoperiod of 24:0 light:dark cycles and an average surface irradiance of 126 μmol m −2 s −1 . The temperature was kept constant (20 ± 1 °C) by recirculating the culture through a thermostatic bath (RA-8 alpha, LAUDA, Lauda-Königshofen, Germany). The absorption unit was made of transparent PVC (inlet diameter 2.84 cm) (Agruquero Thermoplastics S.L., Pinto, Spain) packed with Rasching glass rings (diameter 5 mm). The working volume was 0.8 L, with a height:diameter (H:D) ratio of 44. The absorption unit was fed with substitute biogas (mixture of CO2 and N2) at a constant flow rate of 0.6 L h -1 . The substitute biogas was fed from Tedlar ® Air Sample Bags (50 L, 232-50, SKC, Eighty Four, PA, USA) using a peristaltic pump (7544-30, Cole Parmer Instruments Company LLC., Vernon Hills, IL, USA). The system was controlled and monitored using LabVIEW TM 2015 (v.15.0f2, National Instruments™, Austin, TX, USA) with cDAQ Chassis (NO-9184) and modules for analog input (NI-9208) and a digital input-output interface (NI-9375).

Experimental Conditions
Two nutrient solutions were used: COMBO medium [26] enriched in phosphorus (5 mM) and nitrogen (5 mM NaNO3), and landfill leachate. The COMBO medium was fed

Experimental Conditions
Two nutrient solutions were used: COMBO medium [26] enriched in phosphorus (5 mM) and nitrogen (5 mM NaNO 3 ), and landfill leachate. The COMBO medium was fed semi-continuously, with between 2 and 5 L of culture being removed in order to maintain the biomass concentration at between 1.2 and 1.4 g TSS L −1 . The nitrate concentration was  Table 1. The non-axenic microalgae consortium was isolated [27] from landfill leachate obtained from the "Miramundo-Los Hardales" landfill (Cadiz, Spain), location: 36 • 28 42.5 N, 6 • 00 56.1 W. The microalgae were spherical and had a homogenous size (3.67 ± 0.6 µm) ( Figure S1), with a total protein content of 39.5%, and was able to store lipids under nitrogen and phosphorus limitation up to 53% after 9 d with COMBO medium (2 mM NaNO3) [27]. The predominant species based on size and protein and lipid concentration could belong to Nannochloropsis sp. or Chlorella sp. [27]. The experimental conditions are summarized in Table 2, and each experimental condition was performed in duplicate. Experiment 1 allowed us to establish what flow mode (co-current or counter-current) was most suitable for the maximum removal of CO 2 . Three L/G ratios were used at an inlet CO 2 concentration of 40%. In experiment 2, the effect of L/G ratio on the CO 2 and O 2 outlet biogas concentrations was analyzed. The absorption column flow mode was counter-current and L/G was 1, 1.5, 2 and 4. The inlet CO 2 concentration was 20% and 40%. In experiment 3, landfill leachate was used as a nutrient source, with the L/G ratio being fixed at 1.5 and an inlet CO 2 concentration of 20% and 40%. The photobioreactor was adapted to leachate gradually over 18 days. Initially, 3 L of the liquid medium was removed and replaced with diluted leachate, reaching an approximate concentration of 1 mM N-NH 4 + . Additional periodic replacements were performed when the N-NH 4 + concentration dropped below 0.4 mM. Experiment 3 considers the combined use of leachate and nitrifying bacteria. The inlet CO 2 concentration and L/G were 20% and 1.5, respectively. Leachate was fed into the photobioreactor to obtain a maximum nitrogen concentration of 2 mM N-NH 4 + . In experiment 4, a nitrifying bacterial culture was used. This culture was obtained from a laboratory continuous stirred tank bioreactor (CSTBR) operated for 354 days with a synthetic eluent (ammonium-rich water) [28].Two inoculation procedures were carried out:

•
Inoculation of the photobioreactor: 2 L of nitrifying bacterial culture were added directly to the photobioreactor. To avoid light inhibition of the nitrifying bacteria, the lower third of the bubble column was covered.

•
Inoculation of the absorption column: the absorption column recirculated the nitrifying culture for 15 days, thus allowing biofilm formation on the Rasching rings.

Analytical Methods
The inlet CO 2 concentration was measured by gas chromatography (GC-450, BRUKER, Berlin, Germany) with a Thermal Conductivity Detector and Poraplot Q plot FS 25 m × 0.53 mm column. The outlet CO 2 concentration was measured using an infrared CO 2 transmitter (2112BC4-V, Euro-Gas, Devon, UK). Total suspended solids (TSS) was determined according to Standard Method 2540-C [29]. Nitrate and ammonium concentrations were determined by ion chromatography (Metrohm 930 Compact IC Flex, Herisau, Switzerland).

Fitting to Empirical Model
The experimental results were fitted with an empirical model. A second-order polynomial model was used to predict the outlet CO 2 concentration as the response variable. The independent variables were the L/G ratio and the inlet CO 2 concentration. The levels of the L/G ratio were 1, 1.5, 2 and 4 and the levels for the inlet CO 2 concentration were 20% and 40%. The data were analyzed using Statgraphics ® Centurion 19 (v.19.1.3)

Optimization of the Two-Stage System
The photobioreactor was first operated with the COMBO synthetic medium for 55 days. Figure 2 shows the evolution of biomass and nitrogen concentration in the form of nitrate. The nitrate concentration was 24.95 ± 3.41 mg N-NO 3 − L −1 , the DO was 9.69 ± 0.14 mg O 2 L −1 and pH was 8.7 ± 0.1. Under these operating conditions, the average biomass concentration and biomass productivity were 1.28 ± 0.05 g TSS L −1 and 31.0 ± 11.7 g m −3 d −1 , respectively. Other authors have reported similar productivities. For example, Chiu et al. [30] described a semi-continuous operation with Nannochloropsisoculata and obtained a productivity of between 37 and 48 g m −3 d −1 , and a biomass concentration of between 0.75 and 0.92 g TSS L −1 for a CO 2 concentration in the range 2-15%. In another example, Ruiz et al. [31] obtained a productivity of 17 g m −3 d −1 with a culture of Chlorella vulgaris.  The pH is of great importance in the absorption of acidic gases such as CO2. Thus, the gas-liquid equilibrium changes as a function of pH, and this change in equilibrium can be described by a coefficient relating Henry's law constant, dissociation constants, and pH [25] ( Figure S2). This coefficient relates the concentration in the gas and in the liquid at equilibrium. For example, at a pH of 8.5, the value is 7•10 −3 , while for pH 9.5, it is 6.2•10 −4 , thus indicating that an increase of 1 pH point causes the gas to be 11 times more soluble. Figure 3a shows the concentration of CO2 and O2 at the outlet stream of the absorption column for the three L/G ratios studied (1, 2 and 4), for both counter-current and cocurrent flow. The stabilization period for CO2 and O2 outlet concentrations was between 4 and 6 h (example in Figure 3b). A statistically significant difference (Multifactor ANOVA, L/G p-value < 0.0001 and flow mode p-value = 0.0020) between CO2 concentration and both factors was found. It can be seen how an increase in L/G causes a greater absorption of CO2 and, therefore, a lower concentration in the output gas stream. On the other hand, an increase in L/G also causes a greater O2 desorption and, therefore, an increase in the outlet O2 concentration. Figure 3a also shows that CO2 absorption was higher when the flow was in counter-current, with an outlet concentration of 0.4% being obtained when L/G was 2, compared to the value of 10.9% in the co-current experiment. In contrast, O2 desorption was slightly lower when the flows were co-current. For an L/G of 4, the values were 2.1% and 2.5% for co-current and counter-current flows, respectively. In view of these results, a counter-current flow and an L/G ratio of 2 resulted in the lowest outlet The pH is of great importance in the absorption of acidic gases such as CO 2 . Thus, the gas-liquid equilibrium changes as a function of pH, and this change in equilibrium can be described by a coefficient relating Henry's law constant, dissociation constants, and pH [25] ( Figure S2). This coefficient relates the concentration in the gas and in the liquid at equilibrium. For example, at a pH of 8.5, the value is 7·10 −3 , while for pH 9.5, it is 6.2·10 −4 , thus indicating that an increase of 1 pH point causes the gas to be 11 times more soluble. Figure 3a shows the concentration of CO 2 and O 2 at the outlet stream of the absorption column for the three L/G ratios studied (1, 2 and 4), for both counter-current and co-current flow. The stabilization period for CO 2 and O 2 outlet concentrations was between 4 and 6 h (example in Figure 3b). A statistically significant difference (Multifactor ANOVA, L/G p-value < 0.0001 and flow mode p-value = 0.0020) between CO 2 concentration and both factors was found. It can be seen how an increase in L/G causes a greater absorption of CO 2 and, therefore, a lower concentration in the output gas stream. On the other hand, an increase in L/G also causes a greater O 2 desorption and, therefore, an increase in the outlet O 2 concentration. Figure 3a also shows that CO 2 absorption was higher when the flow was in counter-current, with an outlet concentration of 0.4% being obtained when L/G was 2, compared to the value of 10.9% in the co-current experiment. In contrast, O 2 desorption was slightly lower when the flows were co-current. For an L/G of 4, the values were 2.1% and 2.5% for co-current and counter-current flows, respectively. In view of these results, a counter-current flow and an L/G ratio of 2 resulted in the lowest outlet concentration in the absorption column, giving a combined CO 2 and O 2 concentration of 2.2%. The average DO concentrations in the photobioreactor were 9.60 ± 0.03 and 9.70 ± 0.04 mg O 2 L −1 for the co-current and counter-current experiments, respectively. concentration in the absorption column, giving a combined CO2 and O2 concentr 2.2%. The average DO concentrations in the photobioreactor were 9.60 ± 0.03 an 0.04 mg O2 L −1 for the co-current and counter-current experiments, respectively.  Most of the literature reports use an absorption column in co-current mode. Cervantes et al. [8], for example, evaluated the effect of gas-liquid flow configur absorption column performance in a co-current configuration and obtained a productivity of 15 g m −2 d −1 , whereas biomass productivity decreased to 8.7 ± 0.5 g in counter-current due to a limitation of trace metals. This limitation was cause precipitation of metal sulfides due to the low DO concentration in the lower pa absorption column, where the liquid stream is brought into contact with bioga Most of the literature reports use an absorption column in co-current mode. Toledo-Cervantes et al. [8], for example, evaluated the effect of gas-liquid flow configuration on absorption column performance in a co-current configuration and obtained a biomass productivity of 15 g m −2 d −1 , whereas biomass productivity decreased to 8.7 ± 0.5 g m −2 d −1 in counter-current due to a limitation of trace metals. This limitation was caused by the precipitation of metal sulfides due to the low DO concentration in the lower part of the absorption column, where the liquid stream is brought into contact with biogas with a higher concentration of hydrogen sulfide. These authors observed a lower CO 2 concentration when the operation was co-current, whereas O 2 and N 2 concentrations did not differ significantly. The best configuration was obtained at an L/G ratio of 0.5 and co-current operation. Similarly, Serejo et al. [32] obtained a CO 2 removal efficiency of 80%, and less Processes 2021, 9, 1503 8 of 13 than 2% O 2 , using an L/G ratio of 10 in co-current mode together with synthetic biogas with a CO 2 concentration of 30%. Toledo-Cervantes et al. [33] obtained a removal efficiency of 98.6% using an absorption column fed with alga-bacterial broth at a pH of 10 ± 0.3. These authors also observed that part of the O 2 and N 2 is desorbed on the absorption column in proportion to the L/G ratio. For an L/G of 5, Franco-Mortado et al. [34] found an outlet CO 2 concentration of between 1.8% and 3.3% and an outlet O 2 concentration of 2.6% in a system operating at pH 9.5, and Rodero et al. [35] used a counter-current configuration in a semi-industrial scale system. The maximum biomethane concentration of 90% was limited by desorption of N 2 and O 2 . Finally, Marin et al. [36] used an absorption unit in which gas and liquid flows were co-current (L/G ratio of 0.5). These authors used various operating strategies: with no aeration of the photobioreactor, the CO 2 concentration of purified biogas was up to 6.1% and the pH was 9.1, whereas with aeration (1 vvm), they obtained a biogas CO 2 concentration of 0.3-0.4% and pH of 9.8.

Influence of L/G and Inlet Concentration
The L/G ratio mainly affects two aspects of CO 2 absorption, namely the superficial liquid velocity, which can affect the mass-transfer coefficient, and the concentration of inorganic carbon along the absorption column. Thus, at a liquid velocity of between 0.001 and 0.005 m s −1 , the mass transfer coefficients were similar to that obtained for a flow rate equal to 0 (3.46 ± 0.05 h −1 ). When operating at a higher ratio, the inorganic carbon concentration will be lower, thus causing a higher driving force for the absorption of CO 2 contained in the biogas. This phenomenon can be seen in Figure 4. Statistical analysis shows dependence between CO 2 concentration and L/G ration and inlet CO 2 concentration (Multifactor ANOVA, L/G p-value =0.0042 and inlet CO 2 p-value = 0.0082). For instance, when the L/G ratio was increased from 1 to 4, the outlet CO 2 concentration decreased from 1.5% to 0.1% (inlet CO 2 of 20%), or from 4.2% to 0.4% (inlet CO 2 of 40%). It is interesting to note that an increase in L/G ratio from 2 to 4 did not result in an increase in CO 2 uptake. The specific CO 2 removal rate ranged between 0.25 and 0.27 g L −1 h −1 for 20% CO 2 and between 0.51 and 0.54 g L −1 h −1 at a value of 40%.
higher concentration of hydrogen sulfide. These authors observed a lower CO2 c tion when the operation was co-current, whereas O2 and N2 concentrations did significantly. The best configuration was obtained at an L/G ratio of 0.5 and c operation. Similarly, Serejo et al. [32] obtained a CO2 removal efficiency of 80% than 2% O2, using an L/G ratio of 10 in co-current mode together with synthe with a CO2 concentration of 30%. Toledo-Cervantes et al. [33] obtained a rem ciency of 98.6% using an absorption column fed with alga-bacterial broth at a p 0.3. These authors also observed that part of the O2 and N2 is desorbed on the a column in proportion to the L/G ratio. For an L/G of 5, Franco-Mortado et al. [ an outlet CO2 concentration of between 1.8% and 3.3% and an outlet O2 concen 2.6% in a system operating at pH 9.5, and Rodero et al. [35] used a counter-curre uration in a semi-industrial scale system. The maximum biomethane concentrati was limited by desorption of N2 and O2. Finally, Marin et al. [36] used an absor in which gas and liquid flows were co-current (L/G ratio of 0.5). These authors u ous operating strategies: with no aeration of the photobioreactor, the CO2 conc of purified biogas was up to 6.1% and the pH was 9.1, whereas with aeration (1 v obtained a biogas CO2 concentration of 0.3-0.4% and pH of 9.8.

Influence of L/G and Inlet Concentration
The L/G ratio mainly affects two aspects of CO2 absorption, namely the s liquid velocity, which can affect the mass-transfer coefficient, and the concen inorganic carbon along the absorption column. Thus, at a liquid velocity of betw and 0.005 m s −1 , the mass transfer coefficients were similar to that obtained for a equal to 0 (3.46 ± 0.05 h −1 ). When operating at a higher ratio, the inorganic carbo tration will be lower, thus causing a higher driving force for the absorption of tained in the biogas. This phenomenon can be seen in Figure 4. Statistical analy dependence between CO2 concentration and L/G ration and inlet CO2 concentrat tifactor ANOVA, L/G p-value =0.0042 and inlet CO2p-value = 0.0082). For instan the L/G ratio was increased from 1 to 4, the outlet CO2 concentration decreased f to 0.1% (inlet CO2 of 20%), or from 4.2% to 0.4% (inlet CO2 of 40%). It is interesti that an increase in L/G ratio from 2 to 4 did not result in an increase in CO2 up specific CO2 removal rate ranged between 0.25 and 0.27 g L −1 h −1 for 20% CO2 and 0.51 and 0.54 g L −1 h −1 at a value of 40%.  The pH in the absorption column depends on the quantity of CO 2 absorbed, the liquid flow rate and the concentration of inorganic carbon in the liquid stream from the photobioreactor. At a CO 2 concentration of 20%, the pH in the photobioreactor remained constant at 8.3. In contrast, at a CO 2 concentration of 40%, the pH at the end of the experiments varied as a function of the liquid flow rate, decreasing from a value of 8.05 when the L/G ratio was 1 to a value of 7.79 when the L/G ratio was 4. With regard to the pH in the absorption column at the end of each experiment, as can be seen in Figure 4, this value was proportional to the L/G ratio, and an increase in the inlet CO 2 concentration caused a decrease in pH. Indeed, a lower decrease in pH between the inlet and outlet of the absorption column was observed with increasing L/G ratio (∆pH of 2.45, 2.2, 2.06 and 1.79 for 20% CO 2 and a ∆pH of 2.36, 2.22, 1.82 and 1.46 for 40% CO 2 ). In this regard, Rodero et al. [35] observed that the highest CO 2 uptake occurred at the highest L/G ratio evaluated (3.5). The absorption column inlet CO 2 concentration in that study was 32.7 ± 2.8%, and a removal efficiency of 88.9 ± 1.5% was obtained at a biogas flow rate of 274 L h −1 . These authors also observed a pH decrease of 1.7, 1.5 and 1.2 for L/G ratios of 1.2, 2.1 and 3.5, respectively.The removal efficiency obtained in this study, for similar operating conditions (40% CO 2 and L/G equal to 4) was 99.39%. Marin et al. [37] found that the maximum CO 2 absorption was obtained for an L/G ratio of two, with removal efficiencies in the range 90.4-99.9%. These authors also found that the concentration of N 2 and O 2 increased from 3.4% for an L/G ratio of 0.5 to 11.9% at a ratio of five due to desorption processes.

Empirical Model
The statistical results show the significance and high predictability of the regression model. The R-squared was 86.64%, the residual standard deviation was 0.5990, and the mean absolute error was 0.4166. The second-order polynomial model fitted with calibration data is represented by Equation (1).
The most influential factor on the outlet CO 2 concentration was the L/G ( Figure S3) with a negative effect. The model can be used to predict the optimum L/G ratio to achieve the minimum outlet CO 2 concentration for the specified inlet CO 2 . Therefore, for an inlet concentration of 20%, the optimum L/G would be 2.76, for an inlet concentration of 30%, the L/G ratio would be 3.00 and for 40%, the L/G would be 3.24

Use of Leachate as Culture Medium
Leachate was used as culture medium for 20 days. As can be seen in Figure 5, the biomass concentration remained in the same range: 1.51 ± 0.08 g TSS L −1 when the photobioreactor was operated with COMBO medium, and 1.52 ± 0.09 g TSS L −1 when operated with leachate. DO was maintained at 8.88 ± 0.20 mg O 2 L −1 , whereas the pH decreased from 8.2 ± 0.2 to 6.9 ± 0.1. In order to use pH conditions similar to those used with COMBO medium, the pH was increased prior to the absorption column experiments.
The L/G used (1.5) was lower than the optimum found in order to observe any possible improvement in the removal efficiency of the absorption column. Figure 6 shows the outlet concentrations of CO 2 and O 2 and the outlet pH of the absorption column. When COMBO medium was used, the CO 2 concentration at the outlet was 0.6% and 1.7% for inlet concentrations of 20% and 40%, respectively. The DO concentrations in the photobioreactor were 9.79 and 9.65 mg O 2 L −1 , and the estimated O 2 concentration in the output gas was 1.21% and 1.47%, respectively. On the other hand, when leachate was used as the culture medium, the concentrations at the outlet were 0.01% and 0.6% for CO 2 and 1.09% and 1.37% for O 2 for inlet CO 2 concentrations of 20% and 40%, respectively. A lower DO of 8.81 and 8.89 mg O 2 L −1 , respectively, was measured when leachate was used. Multifactorial ANOVA analysis showed a correlation between the outlet CO 2 and nutrient solution and inlet CO 2 with a p-value equal to 0.0406 for the nutrient solution and 0.0390 for inlet CO 2 . The L/G used (1.5) was lower than the optimum found in order to observe any possible improvement in the removal efficiency of the absorption column. Figure 6 shows the outlet concentrations of CO2 and O2 and the outlet pH of the absorption column. When COMBO medium was used, the CO2 concentration at the outlet was 0.6% and 1.7% for inlet concentrations of 20% and 40%, respectively. The DO concentrations in the photobioreactor were 9.79 and 9.65 mg O2 L −1 , and the estimated O2 concentration in the output gas was 1.21% and 1.47%, respectively. On the other hand, when leachate was used as the culture medium, the concentrations at the outlet were 0.01% and 0.6% for CO2 and 1.09% and 1.37% for O2 for inlet CO2 concentrations of 20% and 40%, respectively. A lower DO of 8.81 and 8.89 mg O2 L −1 , respectively, was measured when leachate was used. Multifactorial ANOVA analysis showed a correlation between the outlet CO2 and nutrient solution and inlet CO2 with a p-value equal to 0.0406 for the nutrient solution and 0.0390 for inlet CO2. A smaller pH drop in the absorption column was observed when leachate was (ΔpH of 2.20 vs. 1.34 for 20% CO2 and ΔpH of 2.22 vs. 1.54 for 40% CO2). This result a higher pH at the outlet of the absorption column when leachate was used, as can be in Figure 6, and thus higher CO2 solubility and better absorption overall. This behav due to the higher alkalinity found in the culture medium from leachate. The bene influence of alkalinity on CO2 removal has been reported by various authors. For exam Marin et al. [37] found that an increase in alkalinity resulted in a higher CO2 remov pacity, and these authors found a clear decrease in the absorption column output co tration with an increase in alkalinity from 42 ± 1 to 1557 ± 26 mg L −1 associated wit A smaller pH drop in the absorption column was observed when leachate was used (∆pH of 2.20 vs. 1.34 for 20% CO 2 and ∆pH of 2.22 vs. 1.54 for 40% CO 2 ). This resulted in a higher pH at the outlet of the absorption column when leachate was used, as can be seen in Figure 6, and thus higher CO 2 solubility and better absorption overall. This behavior is due to the higher alkalinity found in the culture medium from leachate. The beneficial influence of alkalinity on CO 2 removal has been reported by various authors. For example, Marin et al. [37] found that an increase in alkalinity resulted in a higher CO 2 removal capacity, and these authors found a clear decrease in the absorption column output concentration with an increase in alkalinity from 42 ± 1 to 1557 ± 26 mg L −1 associated with the increase in pH of the culture medium (from 6.5 ± 0.1 to 9.3 ± 0.0).

Inoculation with Nitrifying Bacteria
Two strategies were employed. The first strategy allowed the formation of a biofilm of nitrifying bacteria on the support material of the absorption column by recirculating a suspension culture from a CSTBR, whereas in the second strategy, 2 L of culture medium with nitrifying bacteria was added directly to the photobioreactor, one-third of which was covered to provide darkness. The pH in the photobioreactor was 7.54. Both nitrosation [38] and NH 4 + consumption by the microalgae release protons into the medium, thus contributing to its acidification, whereas NO 3 assimilation causes a slight increase in pH [39]. The inoculation period for the nitrifying bacterial consortium in the rings lasted 15 days. Once this time had elapsed, the absorption column was placed in contact with the photobioreactor and bacterial aggregates were found to form, thus indicating possible detachment of the bacteria that formed the biofilm in the Rasching rings. The presence of nitrifying bacteria decreased the O 2 concentration in the photobioreactor from 8.81 to 8.22 and 8.17 mg O 2 L −1 when attached to the support and in suspension, respectively. The consumption of ammonium by the bacteria to generate nitrate decreases the pH, whereas the consumption of nitrate by the microalgae consumes protons. The pH in the photobioreactor was 7.54, lower than that found without bacteria under similar conditions (8.26). The decrease in pH caused less-efficient CO 2 absorption, with a CO 2 concentration at the outlet of the absorption column of 1%.
The presence of NH 4 + in the culture medium inhibits the consumption of nitrate by the microalgae [27]. During the experimental period, the ammonium concentration decreased from 30.5 to 3.0 mg N-NH 4 + L −1 and the final nitrate concentration was 19.7 mg N-NO 3 − L −1 . Saldarriaga et al. [27] reported an ammonium inhibition constant for specific nitrate uptake of 0.75 mg NH 4 + L −1 , at a concentration of 3.0 mg NH 4 + L −1 , thus meaning that the specific rate of nitrite uptake is inhibited by 80%. This fact explains the competitive consumption of ammonium by nitrifying bacteria and microalgae and the accumulation of nitrate in the culture medium. It is therefore necessary to look for an alternative strategy to consume the O 2 produced in the photobioreactor, preferably involving the combined and symbiotic action of microorganisms. These strategies may concern the use of the oxygen content in biogas as an electron acceptor. An example could be the use of an aerobic desulphurization unit, adding sulfur or thiosulphate as the electron donor; a possible handicap is the production of hydrogen sulfide in reductive ambient. Low oxygen concentration and solubility will require the use of gas transfer enhancements, as can be the use of oxygen vectors, such as n-dodecane.

Conclusions
The use of a two-stage system comprising an absorption column and a photobioreactor has been successfully implemented. Landfill leachate has been found to be a feasible nutrient source, and it has also been demonstrated that the CO 2 contained in the biogas can be more efficiently removed. In addition, the O 2 concentration in the biogas leaving the absorber was lower. When leachate was used, the pH of the photobioreactor was similar to that recorded when COMBO medium was used, whereas the pH in the absorber was 0.77 ± 0.12 higher. As such, we can conclude that the greater buffer capacity of the medium containing landfill leachate allows operation under conditions in which the solubility of CO 2 was higher. A reduction in DO in the photobioreactor of 0.87 ± 0.15 mg O 2 L −1 was also observed.
The inoculation of nitrifying bacteria had two effects. Firstly, the DO decreased in the whole system, therefore the outlet O 2 concentration was lower, and secondly, there was simultaneous consumption of NH 4 + by both the nitrifying bacteria and the microalgae, thus favoring an acidification of the medium. A higher outlet CO 2 concentration was observed as a result of the lower pH than that found before the inoculation of bacteria.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/pr9091503/s1, Figure S1: Optical photomicrograph of the consortium, Figure S2: Variation in gas-liquid equilibrium constant as a function of pH, Figure S3: Standardized Pareto Chart for outlet CO 2 empirical model. Funding: This research was funded by the "Ministerio de Economía y Competitividad", grant number CTM2016-79089-R "Enhancement of landfill gas by an integrated biological system".