Biogas Purification through the use of a Microalgae-Bacterial System in Semi-Industrial High Rate Algal Ponds

Summary

Air pollution impacts the quality of life of all organisms. Here, we describe the use of microalgae biotechnology for the treatment of biogas (simultaneous removal of carbon dioxide and hydrogen sulfide) and the production of biomethane through semi-industrial open high-rate algal ponds and subsequent analysis of treatment efficiency, pH, dissolved oxygen, and microalgae growth.

Abstract

In recent years, a number of technologies have emerged to purify biogas into biomethane. This purification entails a reduction in the concentration of polluting gases such as carbon dioxide and hydrogen sulfide to increase the content of methane. In this study, we used a microalgal cultivation technology to treat and purify biogas produced from organic waste from the swine industry to obtain ready-to-use biomethane. For cultivation and purification, two 22.2 m³ open-pond photobioreactors coupled with an absorption-desorption column system were set up in San Juan de los Lagos, Mexico. Several recirculation liquid/biogas ratios (L/G) were tested to obtain the highest removal efficiencies; other parameters, such as pH, dissolved oxygen (DO), temperature, and biomass growth, were measured. The most efficient L/Gs were 1.6 and 2.5, resulting in a treated biogas effluent with a composition of 6.8%vol and 6.6%vol in CO₂, respectively, and removal efficiencies for H₂S up to 98.9%, as well as maintaining O₂ contamination values of less than 2%vol. We found that pH greatly determines CO₂ removal, more so than L/G, during cultivation because of its participation in the photosynthetic process of microalgae and its ability to vary pH when solubilized due to its acidic nature. DO, and temperature oscillated as expected from the light-dark natural cycles of photosynthesis and the time of day, respectively. Biomass growth varied with CO₂ and nutrient feeding as well as reactor harvesting; however, the trend remained primed for growth.

Introduction

In recent years, several technologies have emerged to purify biogas to biomethane, promoting its use as non-fossil fuel, therefore mitigating undesairable methane emissions¹. Air pollution is a problem that affects most of the world's population, particularly in urbanized areas; ultimately, around 92% of the world's population breathe polluted air². In Latin America, the air pollution rates are mostly created by the use of fuels, whereby in 2014, 48% of the air pollution was brought on by the electricity and heat production sector³.

In the last decade, more and more studies on the relationship between pollutants in the air and the increase in mortality rates have been proposed, arguing that there is a strong correlation between both datasets, particularly in children populations.

As a way to avoid the continuation of air pollution, several strategies have been proposed; one of these is the use of renewable energy sources, including wind turbines and photovoltaic cells, which diminish the CO₂ release into the atmosphere^4,5. Another renewable energy source comes from biogas, a byproduct of the anaerobic digestion of organic matter, produced along with a liquid organic digestate⁶. This gas is composed of a mixture of gasses, and their proportions depend on the source of organic matter used for anaerobic digestion (sewage sludge, cattle manure, or agro-industrial biowaste). Generally, these proportions are CH₄ (53%-70%vol), CO₂ (30%-47%vol), N₂ (0%-3%vol), H₂O (5%-10%vol), O₂ (0%-1%vol), H₂S (0-10,000 ppmv), NH₃ (0-100 ppmv), hydrocarbons (0-200 mg/m³) and siloxanes (0-41 mg/m³)^7,8,9, where the scientific community is interested in the methane gas since this is the renewable energetic component of the mixture.

However, biogas cannot be simply burned as obtained because the byproducts of the reaction can be harmful and contaminant; this raises the need to treat and purify the mixture to increase the percentage of methane and decrease the rest, essentially converting it into biomethane¹⁰. This process is also known as upgrading. Even though, currently, there are commercial technologies for this treatment, these technologies have several economic and environmental drawbacks^11,12,13. For example, systems with activated carbon and water washing (ACF-WS), pressure water washing (PWS), gas permeation (GPHR), and pressure swing adsorption (PSA) present some economic or other drawbacks of environmental impact. A viable alternative (Figure 1) is the use of biological systems such as those that combine microalgae and bacteria grown in photobioreactors; some advantages include the simplicity of design and operation, the low operating costs, and its environmentally friendly operations and byproducts^10,13,14. When biogas is purified to biomethane, the latter can be used as a substitute for natural gas, and the digestate can be implemented as a source of nutrients to support microalgae growth in the system¹⁰.

A method widely used in this upgrading procedure is the growth of microalgae in open raceway photoreactors coupled with an absorption column due to the lower operation costs and the minimal investment capital needed⁶. The most used type of raceway reactor for this application is the high-rate algal pond (HRAP), which is a shallow raceway pond where the circulation of the algal broth occurs via a low-power paddle wheel¹⁴. These reactors need large areas for their installation and are very susceptible to contamination if used in outdoor conditions; in biogas purification processes, it is advised to use alkaline conditions (pH > 9.5) and the use of algal species that thrive in higher pH levels to enhance the removal of CO₂ and H₂S while avoiding contamination^15,16.

This research aimed to determine the biogas treatment efficiencies and final production of biomethane using HRAP photobioreactors coupled with an absorption-desorption column system and a microalgae consortium.

Protocol

1. System set-up

NOTE: A piping and instrumentation diagram (P&ID) of the system described in this protocol is shown in Figure 2.

1. Reactor set-up
   1. Prepare the ground by leveling and compacting it to improve reactor stability.
   2. On an open field, dig two elongated holes and 3 m from the end, further dig a 3 m² and 1 m deep hole (known as an aeration well).
   3. Place two HRAPs (Figure 3) within the space on geomembrane-covered metal supports. Each reactor must have an operating capacity of 22.2 m³.
   4. Place an air pump per reactor of 1728.42 watts (2.35 hp) close to the point of the HRAPs where the aeration wells were dug.
   5. Fix a paddle wheel (moved by a 1103.24 watts [1.5 hp] electric motor) across the reactor to promote contact between biomass and media.
2. Gas treatment set-up (Figure 4)
   1. Build the desorption column with a 6" polyvinyl chloride (PVC) tube, where the inlet current enters 2 m from the covered top, and the outlet current flows from the bottom (Figure 2).
   2. Set up the absorption tank (V_t: 2.55 m³), where the gaseous inlet (non-treated biogas) current is bubbled from the bottom through 11 diffuser tubes and comes from the anaerobic digester through a 4" PVC pipeline passing through a biogas blower, a 1" rotameter and a sampling port, while the liquid comes from the media recirculation after the desorption column on the bottom of the tank. The liquid outlet is located on the side of the tank. It transports the CO₂-enriched media to the level-control column, and the gas exits from the outlet at the top of the tank, which is connected with a 1" PVC pipeline to conduct obtained biomethane to a burner for its continuous combustion (Figure 2).
   3. Connect the absorption tank to the desorption column through a 4" PVC tube, passing through a sampling port between both operations (Figure 2).
   4. Build the level-control column with a 6" PVC tube where the inlet is located at the bottom. It has two outlets (controlled with butterfly valves), depending on the needs of the system; the first one is located at a height of 2.5 m and the second one at 3 m from the ground (Figure 2).
   5. Connect the HRAP photobioreactors through a 2" PVC pipeline to the 6" desorption column, passing through a recirculation centrifugal pump (1103.24 watts [1.5 hp]) and a 1" rotameter (Figure 2).
   6. Connect the level-control column through a 4" PVC tube to a schedule 40 PVC tube, passing through a sampling port. Next, connect it to a portion of flexible PVC tubing, followed by another schedule 40 PVC tube, and finally, a 4" PVC tube, which opens to the HRAP photobioreactors (Figure 2).
   7. Set up the bypass of the desorption column with 2" PVC pipeline and connect it to the main tube before the sampling port (Figure 2).

2. Functional testing of the system

1. Recirculation centrifugal pump (1103.24 watts [1.5 hp])
   1. To determine the maximum flow rate of the pump, prime the interior for at least 10 min to avoid air suction and start it up at 230 V and 1 phase.
   2. Test the recirculation flow by letting it flow through the 1" rotameter.
2. Biogas bubbling system
   1. To determine the force required to bubble at least an air column equivalent to 200 mbar, test at least 3 blowers with different powers (485.52 watts [0.66 hp], 1838.74 watts [2.5 hp], and 3309.74 watts [4.5 hp]) by bubbling air into the absorption tank.
   2. Visually verify the size and distribution reached by the air bubbles inside the tank. Under the operating conditions described here, the predicted average diameter of the bubbles is 3 mm.

3. Inoculation and growth under indoor conditions

1. Transfer a pure strain of Arthrospira maxima from agar plates to 15 mL of aqueous mineral medium¹⁷ (NaHCO₃ [10 g/L], Na₃PO₄ ·12H₂O [0.033 g/L], NaNO₃ [0.185 g/L], MgSO₄ ·7H₂O [0.014 g/L], FeSO₄ ·7H₂O [0.0008 g/L], NaCl [0.4 g/L]).
2. Scale up the culture to 500 mL flasks with innocuous Jourdan aqueous medium, using 100% of flask volume, and let it grow in 12 h light/ 12 h dark photoperiods using light emitting diode (LED) lamps with surface mount device (SMD) 2835 providing while-cold light at 2000 lm and under continuous mixing by air bubbling (0.3 L/min or 0.6 vvm). (step lasting around 1 month).
3. Continue the scaling-up process by adding 20% of the previous volume to the new volume until 50 L are reached.
4. Adapt the culture to natural light conditions of operation and Jourdan culture media in a greenhouse in 50 L transparent sacks (step lasting around 2 months).
5. Continue scaling in these conditions up to 5 m³ HRAP photobioreactors (step lasting around 2 months).

4. Operational start of the system under outdoor conditions

1. Add the full volume of these 5 m³ HRAP photobioreactors to HRAPs photobioreactors of 13 m³ located outdoors and fill the rest of the volume with Jourdan culture medium. Start mixing through a paddle wheel at a speed of 30 cm/s, cultivating in batch mode for 15 days or until it reaches 0.7 g/L (step lasting around 1 month).
2. Once growth reaches 0.7 g/L, transfer the volume to the operating 22.2 m³ HRAP, fill the rest with Jourdan media, and set the paddle wheel at a speed of 30 cm/s. Let the biomass grow until it reaches 0.7 g/L and a pH of 10; once these conditions are met, start sampling and harvesting, if needed.
3. Start the liquid recirculation from the HRAP photobioreactor to the absorption tank at variable flow to increase biomass productivity. Begin biogas bubbling at an average flow of 3.5 m³/h after 2 h to provide inorganic carbon to the culture. Pay attention to the pH since it must remain above 9.
   NOTE: Before recirculating the media through the absorption tank, prime the centrifugal pump described above.
4. Nutrient addition: Monitor nutrient conditions weekly through harvesting and the overall nitrogen balance assuming steady state calculated as shown:
   M_NaNO3 = (M_Biomass x 0.10)/0.12 [g]
   Where:
   M_NaNO3 = Sodium nitrate mass [g]
   M_Biomass = Harvested biomass [g]
   1.10: Mass yield of nitrogen/biomass¹⁶ [g/g]
   1.12: Mass fraction of nitrogen in sodium nitrate [g/g]
5. With the nitrogen balance results, reformulate the Jourdan media to add the proportional amount of Na₃PO₄·12H₂O, MgSO₄·7H₂O, and FeSO₄·7H₂O. Do not add more sodium bicarbonate or sodium chloride.
   NOTE: Dissolve the nutrients in clean water before adding them to the reactors.
6. Monitor water evaporation and add weekly if needed.

5. Sampling and analysis

1. Biogas
   1. Sample the biogas from the sampling outlet before the absorption tank and from the sampling outlet after the tank by connecting a 10 L polyvinyl fluoride bag to the outlet with a flexible tube of appropriate diameter; place each one in separate polyvinyl fluoride bags.
   2. Calibrate the portable gas analyzer by setting the pressure transductor to zero and waiting for stabilization. Do this by pressing Start, then Next, and connecting a clear tube and a yellow tube as instructed by the analyzer. Press Next and finally, Gas Readings.
   3. Connect each sample contained within the polyvinyl fluoride bags to the analyzer, press Next and measure the CH₄, CO₂, O₂ and H₂S concentrations as %vol from both points of the system.
   4. Determine the volumetric recirculation liquid/biogas ratio (L/G) by dividing the liquid recirculation flow by the biogas production flow. Compute the corresponding gas flow (m³/h) that presents the highest efficiency of CO₂ and H₂S removal.
2. Online measuring of system conditions (pH, dissolved oxygen, temperature)
   1. Calibrate all sensors according to the specifications of the manufacturer.
   2. Place a pH sensor, a dissolved oxygen (DO) sensor, and a temperature sensor in the liquid of each HRAP.
      NOTE: For brand and specifications for each of the sensors, review the Table of Materials file.
   3. Connect the pH and DO sensors to a data-acquisition device consisting of a 1.4 GHz 64-bit quad-core processor connected to a portable screen that stores a pre-made Python program written in Integrated Development and Learning Environment (IDLE) 2.7.
      1. Open the program through the screen and indicate the time intervals to store each data point (in this case, every 2 min).
      2. Create a spreadsheet where the program will automatically store the data it collects.
      3. Click on the button that reads ON, indicating it is ready to start storing data.
      4. To stop the data acquisition, click on the button that reads OFF.
      5. To download the information, insert a universal serial bus (USB) and import the spreadsheet.
   4. Connect the temperature sensor to a thermo-recorder to store the data recorded during the experiments.
3. Short exploratory tests
   1. Determine the most efficient L/G
      1. Regulate the incoming biogas flow to select the L/G value to be tested (0.5, 1, 1.5, 1.6, 2, 2.5, 3.3, 3.4).
      2. Measure the pH and the inlet and outlet concentrations of each gas (CH₄, CO₂, H₂S, O₂, N₂) at the start and every 15 min for an hour (60 min), using the instruments described previously.
      3. Determine the most efficient L/G by comparing the outlet values and choose the one most convenient according to the needs of the experiment.
   2. Relationship between L/G, pH and CO₂
      1. Choose at least two L/G's to compare.
      2. For each L/G, measure the pH and the inlet and outlet concentrations of CO₂, and of H₂S, O₂, and N₂ as a control at the start, every 15 min for 60 min, and then every hour for a total of 5 h, using the instruments described previously.
      3. Calculate the CO₂ removal percentages using the equation:
         %CO₂ removal = ((CO₂in - CO₂out)/(CO₂in)) x 100
      4. Graph the results and compare the behavior of the pH and CO₂ for each of the L/G's tested.
4. Calibration curve to correlate biomass weight per liter of culture versus absorbance at 750 nm¹⁸
   1. Sample the algae culture to try and get an absorbance of 1.0. If the culture has an absorbance below 1.0, extract water by filtration (0.45 µm filter) from a culture sample. If the absorbance is greater than 1, it can be decreased by adding a fresh culture medium.
   2. Prepare five algae cell suspensions using the sample and add fresh culture medium, in volume/volume (V/V) percentage: 100%, 80%, 60%, 40%, and 20%.
   3. Measure and record the absorbance at 750 nm of the five solutions with a spectrophotometer using plastic cuvettes, where the fresh culture medium is the blank.
   4. Determine the biomass weight per liter of culture of every suspension by filtering 10 mL through a previously weighed 0.45 µm filter and drying the sample in a silica desiccator for 24 h and later 48 h to ensure a constant weight. Repeat this step for each of the five solutions.
      NOTE: A higher temperature (above 60 °C) is not recommended for drying due to the loss of certain key compounds that could volatilize and change the sample's weight.
   5. Once confirming the weight, calculate the biomass concentration within the reactor with the equation:
      ​Biomass concentration = (Biomass weight - filter weight) x 1000/Filtered volume [g/L]
   6. Make a linear regression of the biomass weight data in grams per liter of culture as a function of the absorbance measured at 750 nm using a spreadsheet or any other software. The linear regression coefficient should be greater than 0.95; otherwise, the curve is not useful, and the protocol should be repeated.
      NOTE: It is described as biomass weight and not as dry weight as most methods because the drying method used does not allow for full removal of water in the sample, leaving a water content of less than 5%¹⁹.
5. Biomass growth
   1. Monitor the reactors every day. Take a 1 L sample from the halfway point between the paddlewheel and its return from each culture and bring it to the laboratory.
   2. Check colony growth and purity of the culture under the microscope.
   3. Measure and record the absorbance at 750 nm of the samples with a spectrophotometer, where the fresh culture medium is the blank.
   4. Compare with the calibration curve to obtain the estimated biomass weight in grams per liter.
   5. Record the growth of each raceway reactor.
6. Biomass production - harvesting
   1. Monitor the reactors every day. If biomass growth rises above 0.7 g/L during sampling, harvesting is needed.
   2. Alternating between both HRAPs, place a polyester mesh on top of a section at one end of the reactor and place an end of a flexible PVC tube within the flow of the liquid so that the other end drains the liquid on top of the mesh.
   3. Drain between 4500 L to 7500 L (depending on the biomass saturation of the reactor) onto the mesh, maintaining a continuous flow back to the corresponding HRAP. The biomass will be retained on the mesh.
   4. To harvest, remove the mesh from the top of the reactor and place it on a different surface to scrape the biomass off and place it into a funnel.
   5. Push the biomass through the funnel to create elongated shapes on top of a clean and dry mesh; set the mesh in a warm, covered room (34-36 °C) for 48-72 h.
   6. Once dry, remove the biomass from the mesh and weigh it. Calculate the biomass harvested concentration in g/L with these equations:
      Volume of drained liquid = Pump flow rate x Drain time [L]
      Biomass harvested concentration = Biomass weight of harvested biomass/Volume of drained liquid [g/L]

Results

Following the protocol, the system was built, tested, and inoculated. The conditions were measured and stored, and the samples were taken and analyzed. The protocol was performed a year, starting in October 2019 and lasting until October 2020. It is important to mention that from here onwards, the HRAPs will be referred to as RT3 and RT4.

Biomethane productivity
In order to determine the conditions that promote the highest H₂S and CO₂ removal and, consequently, the highest concentration of methane, several recirculation liquid/biogas ratios (L/G) were tried in a range from 0.5 to 3.4. These results were obtained for experiments with a duration of at least 60 min (1 h) of continuous biogas bubbling in the period between the 25^th of September and the 28^th of September. During these tests, the microalgae fixed the CO_2, and the bacteria oxidized the H₂S, concentrating the methane (CH₄) and, essentially, purifying the gas mixture.

Considering the averaged CO₂ elimination capacity of the whole system (HRAP volume + Tank volume = 24.75 m³) and a stable biomass concentration of 0.8 g/L, then a specific fixation rate was estimated, resulting in 65 mgCO₂/gbiomass h, which is lower than the maximum theoretical reported (300 mgCO₂/gbiomass h). This denotes that the biogas purification process based on microalgae-bacteria is suitable to be enhanced.

Generally, biogas purification had increased efficacy in higher L/G values, maintaining removal efficiencies at or above 98% for H₂S and less than 7.5%vol content values for CO₂ (Figure 5, Figure 6, and Figure 7). However, O₂ biomethane contamination due to photosynthetic production of this gas was much higher at higher L/G values, which can be a potential problem for commercial use as O₂ concentrations, by law, must remain quite low to reduce the risk of explosion²⁰. Another reason is linked to avoiding diminishing its calorific value by O₂ dilution. Instead, it could be argued that the L/Gs 1.6 and 2.5 represent the most efficient results overall, with CO₂ concentrations between 6.6%vol and 6.8%vol, CH₄ at 87%vol, and O₂ at less than 1.5%vol, as well as presenting H₂S removal efficiencies of above 98.5% (Figure 5, Figure 6, and Figure 7). A comparison between obtained percentages and what is accepted by law can be found in Table 1.

It is interesting to note that the recirculation liquid/biogas ratio of 2 has a higher CO₂ concentration (7.4%vol) even if the value sits between the most efficient L/Gs; this could be attributed to the fact that it was tested in RT3 instead of RT4. In this case, the conditions were less favorable for CO₂ removal, possibly due to a lower biomass concentration. Overall, the mean biomethane produced in these conditions came up to 20.68 m³/day, with an average flow rate of 4.14 m³/h.

Results may vary depending on the growth conditions, the type of biogas (synthetic or real) and the algae; for example, Serejo et al.²¹ used two synthetic biogas mixtures simulating an all CO₂ and N₂ gas mixture to compare with a regular biogas made of 70%vol CH₄, 29.5%vol CO₂ and 0.5%vol H₂S, purifying it through a 180 L HRAP-absorption column system cultivating Chlorella vulgaris. In this paper, Serejo tests different L/G ratios as well, ranging from 0.5 to 67, in a smaller but similar system at lower pH values and artificial lighting. Full removal of H₂S and an average removal percentage of 80% in the best ratios (above 15) was achieved. These removal efficiencies increased linearly with the ratio; however, oxygen contamination also increased, which could pose problems in the overall quality of the resulting biomethane. The increase in our CO₂ removal efficiencies was not linear; nonetheless, a better CO₂ elimination can be seen with larger ratios. The explanation is multicausal, involving pH, culture nutrient conditions, and biomass growth, as well as biogas bubbling.

The effect of the L/G ratio on the performance of the biogas upgrading system was evaluated without repetition. It was justified since the assays were performed at a diurnal period from 10:00 to 13:00 h (it would induce stable solar irradiation and outdoor temperature); therefore, it induced nearly optimal growing conditions for photosynthetic microorganisms, then pH could be assumed to be the most influential parameter on the CO₂ absorption²² where also very little standard deviation lower than 2% was reported for the assays assessing the effect of L/G ratio on the CO₂ absorption removal efficiency.

The reactors are cultivated on the outside, which means that, even if the inoculum was a pure culture of Arthrospira maxima algae, the probability of contamination with other organisms that can survive in the harsh pH conditions within the culture is high. Such is the case for sulfur-oxidizing bacteria^23,24. However, this contamination proves to be beneficial for the final purpose of the experiment since these bacteria help remove the H₂S from the biogas, essentially taking charge of this task and aiding in the quality of the resulting biomethane.

Under the environmental and ionic strength conditions prevailing during the system operation, the dissolved H₂S was being oxidized to polysulfides and thiosulfate by oxic-abiotic reactions, where, after some days, it should be completely oxidized to sulfate²⁵. The H₂S removal by precipitation with cations in the aqueous nutritive medium is insignificant due to the insufficient amount of cations fed to the system compared with the H₂S loading rate (reaching Cations/H₂S molar ratios much lower than 2). The absence of precipitates was confirmed by our visual inspection during the performance of the biogas upgrading process. The biological sulfide oxidation was not verified at this moment since the system is open to the environment.

System conditions
Dissolved oxygen (DO) and pH variations were measured in both light and dark conditions. During the day (light conditions), DO increased due to the photosynthetic production of oxygen by the microalgae, while at night (dark conditions), it decreased both due to lack of photosynthesis and because of heterotrophic metabolism, which utilizes respiration (Figure 8).

Levels of pH also varied with the presence of CO₂ within the liquid (Figure 8), increasing in value when less CO₂ was dissolved and decreasing when less CO₂ was removed; notably, there are smaller peaks around the times when no more CO₂ was being provided, which will be discussed further on. During the mornings, pH hit its peak at around 11:00 AM and the lowest values at around 18:00 PM, which is also consistent with algae photosynthetic activity. It is important to bring attention to the major drop around day 2; the short exploratory test using the L/G of 1.64 was performed on the 29^th of September, supplying continuous biogas by around 24 h (at around day 1) and it provoked a massive destabilization in the system, requiring the supply of urea to aid in the nitrogen recovery. The other short exploratory test using 1.58 was performed on the 5^th of October (at around day 7), but at better system conditions (biogas supply during daylight period), which is why the pH only strayed slightly from the regular peaks for two days before returning to normal behavior.

The smaller peaks in pH in Figure 8 can be attributed to a period of self-regulation of the algae to the environment while changing from photosynthesis to respiration.

Referring to the short exploratory tests to relate pH and L/G with CO₂ removal percentages (Figure 9), we tested two ratios, 1.64 and 1.58, as was mentioned previously. These are both averages from the recorded L/Gs during the experiments. Two distinct behaviors can be noted, where the removal percentage and the pH at a ratio of 1.58 were remarkably less stable and much lower than the ones recorded for the ratio of 1.64.

This is supported in the biogas upgrading performed by Bahr et al.¹⁵, through the use of an HRAP-column system with a species of Arthrospira maxima algae. Bahr assessed the removal efficiencies of CO₂ at different pH conditions and media liquid flow rates, as well as the removal of H₂S and O₂ contamination, on several synthetic gas compositions ranging from simply CO₂-N₂ to biogas compositions with varying H₂S concentrations (up to 0.5%vol). They concluded that at higher pH values (ranging 9-10) and higher culture media liquid flow rate (80 mL/min), the CO₂ removal percentages were close to 100% but suffered higher O₂ contamination, while at higher pH values (ranging 9-10), and lower culture media liquid flow rate (20 mL/min), the CO₂ removal percentages remained close to 100% and much less O₂ contamination was observed. They also reported full H₂S removal in these conditions.

Similarly, DO oscillation (Figure 8) can be attributed to the photosynthetic activity of the algae since, during the day, DO increased due to the photosynthetic production of oxygen by the microalgae, while at night, it decreased both due to lack of photosynthesis and because of heterotrophic metabolism, which utilizes respiration.

The temperature in the HRAP photobioreactor (RT4) varied due to the time of day and autumn weather, peaking most days between 23 °C and 28 °C at around 17:00 and hitting the lowest values between 11 °C and 15 °C at around 6:00 (Figure 10). The temperature at the inlet and outlet of the absorption tank was occasionally measured, resulting in an average temperature of 30.1 °C and 32.5 °C, respectively. Therefore, the water content (vapor) after treatment shall be slightly higher (13.5%) than before biogas treatment, assuming that in both cases, moisture in biogas achieved saturation. It is highly recommended to install a biogas dryer for optimal management and further use of purified biogas.

The average L/G that was intended for the period between the 28^th of September and the 10^th of October was 1.6 since the short tests suggested that this ratio would promote better results; however, it was not possible to maintain it during the nights due to the excessive acidification of the microalgae culture caused by a poor pH buffering capacity of the aqueous culture media. Therefore, only during daylight hours, biogas was fed to the absorption tank, adjusting the L/G values to around 1.5.

Biomass productivity
The inoculation on RT3 was performed on the 20^th of May 2020 and on RT4 on the 27^th of May 2020; the time in between the tests (September) and the inoculation served to stabilize the culture and solve operational issues that arose, such as plagues and malfunctions in the system, considering the COVID global pandemic.

Biomass growth was measured in two ways: sampling and harvesting. For the purposes of this article, sampling refers to the concentration of biomass at any given time in the reactor, while harvesting refers to the production efficiency of the biomass, meaning the amount of biomass that was recovered during the process to avoid growth inhibition. The testing was done from the 29th of September to the 9th of October, at an average L/G of 1.5, even though a ratio of 1.6 was preferred; the reason for it resulting lower was due to the 1.15 ratio recorded around day 11.

Sampling (Figure 11) was done regularly from day 1 to day 11 (from the 29^th of September to the 9^th of October), where the growth trend in both reactors was very similar: it started with a higher concentration, hitting the lowest value for the experiment at days 4 and 5, steadily recovering in RT4 and with some variation in RT3, finally dropping again. The very same behavior is seen in Harvesting, which then suggests that an event (most likely an outside factor) affected the growth of both cultures simultaneously.

Harvesting (Figure 12) was done semi-regularly, alternating one harvest for RT3 and the next harvest for RT4. However, the scale must be considered; in both sampling and harvesting, the variation between numbers is very low, indicating that the event that affected both reactors was not critical. The red dotted line in Figure 8 denotes the period of time when the reactors were not harvested; this was due to two factors: a few days were during the weekend, when, unfortunately, the reactors were not accessible for sampling or harvesting (which can also be corroborated in Figure 11), and the methodology calls for harvesting of the reactor that has the highest concentration. In the complex, there were four reactors, of which only two (RT3 and RT4) participated in this study, making the days after the weekend, days when the other two reactors (RT1 and RT2) were harvested by the team and resulting in no harvesting data from RT3 and RT4. The harvesting data was around 50% less than the sampling data; this could be because the methodology's efficiency is lower.

The variation between values each day was small (Figure 11), which alludes to a resilient culture that allows for change in system conditions and remains stable. Arthrospira maxima preferentially grows in highly carbonated media at high pH and is highly sensitive to NH₃ inhibition¹⁵, which is consistent with the results shown in Figure 8. The calibration performed in August 2020 is shown in Figure 13.

Post-production review and byproducts
In order to review the potential of this gas to reduce harmful emissions to the environment, a full report by an outside company was performed, where the findings stated that the biomethane produced with this technology reduced the total direct CO₂ emissions by 84%, compared to using the unpurified biogas directly from the anaerobic digester. Additionally, when taken through a life cycle analysis of electricity generated by both the raw biogas and the purified biomethane, the overall heat capacity that the biomethane was able to provide was 23,000 kJ higher than the heat capacity of the raw biogas.

Finally, a byproduct of this purification process is the harvested microalgae, which, once dry, has a myriad of applications in other industries, which could add more value to the method and make the process cost-effective²⁶. For instance, a study was performed on basil crops to evaluate parameters such as number of leaves, shoot fresh and dry weight, and leaf fresh weight when using dried Scenedesmus biomass versus a regular inorganic fertilizer; they found comparable results in these criteria in both biomass and fertilizer²⁷. Similar results were found in another study where they compared the growth of four commercial crop plants while using different concentrations of a fertilizer made of algal biomass suspended in water; even at low concentrations (20%) of the fertilizer, the crops reached maximum growth, comparably with chemical fertilizers²⁸.

Figure 1: Visual representation of the biological process happening in biogas purification using microalgae Please click here to view a larger version of this figure.

Figure 2: P&ID diagram for the system described in the protocol. Please click here to view a larger version of this figure.

Figure 3: Photograph of HRAPs that were used during experimentation. Please click here to view a larger version of this figure.

Figure 4: Absorption tank. (A) Photographs of culture medium and biogas inlets to Absorption tank. (B) Front and back view of Absorption tank. Please click here to view a larger version of this figure.

Figure 5: Short exploratory tests in RT3 to determine L/G efficiency. Dark green corresponds to CH₄, green corresponds to CO₂, light pink corresponds to O₂, and dark pink corresponds to N₂. Average pH 9.2435; Liquid inlet 60-100 L/min; Gas inlet 50-120 L/min. Please click here to view a larger version of this figure.

Figure 6: Short exploratory tests in RT4 to determine L/G efficiency. Dark pink corresponds to N₂, light pink corresponds to O₂, dark green corresponds to CO₂, and light green corresponds to CH₄. Average pH 9.95; Liquid inlet 116-118 L/min; Gas inlet 35-75 L/min. Please click here to view a larger version of this figure.

Figure 7: Comparison of all removal percentages for H₂S in each L/G during the short exploratory tests. The L/Gs of 0.5, 1, 1.5, and 2 correspond to RT3, and 1.6, 2.5, 3.3, and 3.4 to RT4. Please click here to view a larger version of this figure.

Figure 8: pH and DO profile. pH (dark green) and DO (light green) profile for RT4 between the 28^th of September and the 10^th of October 2020. Liquid inlet 75-118 L/min; Gas inlet 57-75 L/min. Average feed concentrations for each gas: CH₄- 60%vol, H₂S - 2400 ppmv, CO₂- 34%vol, O₂- 0.6%vol. Please click here to view a larger version of this figure.

Figure 9: Removal percentage profiles for CO₂ depending on pH levels and L/G. Green corresponds to the CO₂ removal percentages at the L/G ratios: 1.58 (dark green triangles) and 1.64 (light green circles). Pink corresponds to the pH values at the L/G ratios: 1.58 (dark pink triangles) and 1.64 (light pink circles). Liquid inlet 75-118 L/min; Gas inlet 57-75 L/min. Average feed concentrations for each gas: CH₄- 60%vol, H₂S - 2400 ppmv, CO₂- 34%vol, O₂- 0.6%vol. Please click here to view a larger version of this figure.

Figure 10: Temperature profile for RT4 between the 28^th of September and the 10^th of October 2020. Please click here to view a larger version of this figure.

Figure 11: Sampling results for RT4 (light green squares) and RT3 (dark green circles) between the 28^th of September and the 10^th of October 2020. The L/G ratios are indicated with arrows. Please click here to view a larger version of this figure.

Figure 12: Harvesting results for RT4 (light green squares) and RT3 (dark green circles) between the 28^th of September and the 10^th of October 2020. The L/G ratios are indicated with arrows. In red dotted lines, the period where there was no harvest for either reactor is shown. Please click here to view a larger version of this figure.

Figure 13: The calibration curve performed in August 2020, correlating the concentration of the algal culture in grams per liter to the absorbance at 750 nm. Please click here to view a larger version of this figure.

Component (%vol)	Obtained biogas composition	Upgraded biogas composition	Commercial biomethane Composition NOM-001-SECRE-2010
CH4	64.2 ± 0.8	85.1 ± 2.0	>84
CO2	33.8 ± 0.1	7.2 ± 1.2	<3
H2S (ppmv)	2539 ± 32	30.5 ± 4.2	<6
O2	0.3 ± 0.1	1.7 ± 0.5	<0.2

Table 1: Comparative compositions of biogas

Discussion

Throughout the years, this algal technology has been tested and used as an alternative to the harsh and expensive physicochemical techniques to purify biogas. Particularly, the Arthrospira genus is widely used for this specific purpose, along with Chlorella. There are few methodologies, however, that are made on a semi-industrial scale, which adds value to this procedure.

It is critical to maintain lower O₂ concentrations by using the proper L/G ratio; however, this depends on the region where this protocol will be applied. Oxygen content is heavily regulated in biomethane due to risk of explosion and corrosion in the pipelines. Some countries in the European Union demand contents to be as low as 1%vol^29,30,31. Methane, on the other hand, must be at a concentration of more than 65% vol³¹. In Mexico, there is almost no regulation regarding biogas and biomethane, for it is considered equivalent to natural gas, where according to Mexican standards³², the minimum content of CH₄ in biomethane is 84%vol and a maximum O₂ content of 0.20%vol is allowed.

Additionally, pH greatly determines CO₂ removal, more so than L/G, during cultivation, which is why it is critical to maintain proper control of pH throughout the methodology, particularly during biogas bubbling. It is important to understand that once CO₂ is solubilized in the liquid, there is a chemical equilibrium at play that directly impacts pH levels. At the pH levels that these cultures oscillated around (8.5-9.5), bicarbonates are the form in which this molecule is present, with a slight increase of carbonates at the higher end of the pH range³³. In this form, the microalgae are also able to metabolize the carbon during the dark reactions of photosynthesis to produce carbohydrates³⁴. The timing of biogas bubbling is also important, of which it is recommended to maintain the daytime bubbling. Nonetheless, L/G also affects CO₂ removal and pH, as can be seen in FIG. 5. The removal percentage and the pH at a ratio of 1.58 were less consistent and much lower than the ones recorded for the ratio of 1.64. This behavior could be attributed to a higher intake of gas in the recirculation ratio (more gas makes for a smaller ratio), which lowered the pH at a faster rate. However, it could also be argued that the starting pH for 1.64 was higher, which favored the buffered behavior of CO₂ elimination efficiencies during this test. L/G in this protocol is controlled through the amount of biogas that is bubbled; however, other protocols vary the recirculating liquid rate, which is also an option. Furthermore, it is not possible to bubble biogas at night due to the acidification of the culture and algae metabolism, since no artificial light is provided at this time.

Another phenomenon that introduces variability in the validity of the results is the intermittent air bubbling used to avoid biomass sedimentation in the reactors, which prevents growth inhibition by oxygen accumulation. This, however, cannot be avoided if this method is used. An alternative to air bubbling is adding more paddle wheels to improve the movement along the length of the reactor, which may be effective in other experiments. On the other hand, the extensive areas of land needed for the installation of the reactors, as well as the significant consumption of water to start and maintain the system in order to obtain fair biomethane productivity.

It is important to note that this regular sampling process uses the biomass weight - absorbance calibration curve (Figure 9), where the correlation between the data is almost 1 (0.9995); while the method might not be based on a previous article on the same algae, the determination coefficient shows a strong statistical connection that this method is reliable. Furthermore, it is relevant to describe the importance of both sampling and harvesting in a methodology such as this one. Sampling allowed for proper maintenance of the algae culture, while harvesting served a triple purpose: firstly, it avoided growth inhibition due to overcrowding of the culture which could cause oxygen accumulation³⁵; secondly, the recovery of algal biomass can lead to further economic opportunities; and finally, it granted another opportunity to measure the growth trend for the culture.

Nevertheless, determining the appropriate moments to harvest (which in this protocol are defined by the sampling results) is also a critical step because it lowers biomass in the reactors. A lower biomass concentration affects pH and CO₂ removal as a cycle: at unfavorable system conditions (for instance, at lower pH values), biomass growth slows down, which, in turn, lowers the system's capacity to eliminate CO₂ as there is less biomass to metabolize it; more dissolved CO₂ would acidify the culture media, and close the cycle³⁶. Many other factors contribute to pH and biomass growth, which should not be overlooked in this oversimplification of cause-effect; nitrogen availability can be extremely important for Arthrospira maxima algae, as well as climate conditions like temperature and light intensity^16,36, which cannot be controlled in a system such as this one. As an example, the addition of urea, as seen in Figure 4, is proof that nitrogen, along with higher pH values, can regularize an algae system.

Other limitations of this method are related to the harvesting productivity, which, when compared to sampling, is around 50% less efficient, which hinders the economic feasibility of the system and would require the improvement of filtration techniques. Harvesting weight results are overestimated by 6% (as measured afterward following standard dry weight methods), given that the drying conditions on that part of the protocol do not result in full water elimination. On the topic of biomass, the sampling results (including the calibration curve) are overestimated by at least 5% due to the incomplete elimination of water in the methodology¹⁹; however, since the error is systematic, it is recommended only to proceed with a thermogravimetric analysis to verify the water content in the culture to consider and make the analytical corrections to the results and calibration curve.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
1" rotameter	CICLOTEC	N/A
1" rotameter	GPI	A10-LMA100IA1
Absorption tank	EFISA	Made under previous design
Air blower (2.35 HP)	Elmo Rietschle	2BH11007AH01
Biogas blower (2 HP)	Elmo Rietschle	2BH11007AH01
Biogas composition measure	Geotech	BIOGAS 5000
Data-acquisition device	LabJack Co.	U3-LV
Diffuser tubes	Aero-Tube	C3060AR
DO sensor	Applisens	Z10023525
Dodecahydrated trisodium phosphate	Quimica PIMA	N/A	Fertilizer grade (greenhouse and experior use)
Dodecahydrated trisodium phosphate	Fermont	35963	Analytical grade (Used in cultures inside the laboratory)
Durapore membrane (45 µm)	MerckMillipore	HVLP04700
Electric motor 1.5 HP	Weg	00158ET3ERS56C
Ferrous sulfate heptahydrate	Agroquimica Samet	N/A	Fertilizer grade (greenhouse and experior use)
Ferrous sulfate heptahydrate	Fermont	63593	Analytical grade (Used in cultures inside the laboratory)
Geomembrane	GEOSINCERE	N/A
Magnesium sulfate heptahydrate	Tepeyac	N/A	Fertilizer grade (greenhouse and experior use)
Magnesium sulfate heptahydrate	Fermont	63623	Analytical grade (Used in cultures inside the laboratory)
Paddle wheel	GSI	Made under previous design
pH sensor	Van London pHoenix	715-772-0041
Portable screen	Rasspberry	Pi 3 B+
Recirculation centrifugal pump (1.5 HP)	Aquapak	ALY 15
Sodium bicarbonate	Industria del alcali	N/A	Fertilizer grade (greenhouse and experior use)
Sodium bicarbonate	Fermont	12903	Analytical grade (Used in cultures inside the laboratory)
Sodium chloride	Sal Colima	N/A	Fertilizer grade (greenhouse and experior use)
Sodium chloride	Fermont	24912	Analytical grade (Used in cultures inside the laboratory)
Sodium nitrate	Vitraquim	N/A	Fertilizer grade (greenhouse and experior use)
Sodium nitrate	Fermont	41903	Analytical grade (Used in cultures inside the laboratory)
Storing program (pH, DO)	Python Software Foundation	Python IDLE 2.7
Tedlar bags	SKC Inc.	232-25
Temperature recorder	T&D	TR-52i
UV-Vis Spectrophotometer	ThermoFisher Scientific instrument	GENESYS 10S
Vacuum pump	EVAR	EV-40