Microalgae Commodities Production with a Direct Air Capture Process Award Number: DE-EE0009276 Final Technical Report Prepared by Braden Crowe1, Casper Pratt1, Maria Reyna1, Erik Hoffnagle1, Aubrey Davis1, Kirk Moses1, Cooper Gibson1, Song Gao2, Scott Edmundson2, Michael Huesemann2, Ryan Latta3, Charley O’ Kelly3, Court Warr3, Eric Ping4, Miles Sakwa-Novak4, Brianna Atherton4, Stephanie Didas4, Edward Legere4 and John Benemann1 Recipient 1 MicroBio Engineering Inc. 3988 Short St., Suite 100 San Luis Obispo, CA 93401 Principal Investigator: John Benemann, Ph.D. Project Manager: Braden Crowe Consortium Partners 2 Pacific Northwest National Laboratory 3 Cyanotech Corporation 4 Global Thermostat LLC Prepared for Advanced Algae Systems Program Bioenergy Technologies Office U.S. Department of Energy Washington D.C. Program Manager/Project Officer: Daniel Fishman Project Monitor: Dr. Jamie Meadows Golden, Colorado Report Number: DOE-MBE-EE0009276 1. Direct Air-CO2 Capture (DAC): A physical-chemical process to deliver a concentrated CO2 stream from an external-to-the-cultivation system for capturing and concentrating air-CO2, and 2. Direct Uptake of Air-CO2 by the Algal Cultures: Utilizing the algae's metabolic capacity to absorb CO2 directly from the air. Key Project Objectives and Accomplishments are summarized below: Objective Key Accomplishment/Result Objective 1: DAC Demonstration Near- Over the course of a 2,000-hour campaign, the DAC Pilot term small-scale demonstration of a Unit successfully produced CO2 with greater than 90% technologically feasible and purity. Based on this performance data, the projected cost economically viable process to supply of carbon for a full-scale, nth-plant system is estimated to air-CO2 for microalgae production using be under $150/Mg (metric ton). DAC. Objective 2: Chlorella Production An annual average biomass productivity of 9.3 g ash-free Demonstration of Chlorella production dry-weight (AFDW)/m²-day was measured for the using an air-CO2 source to meet a alkaliphilic Chlorella isolate CT3072 under high Carbon commercial-scale biomass productivity Utilization Efficiency (CUE) conditions (90% projected CUE). goal. This met the project's baseline Go/No-Go criteria. Objective 3: Direct Uptake Proof-of- Exploratory research established the framework for a Concept of an alternative direct air-CO2 proof-of-concept study for biochemically enhanced air-CO2 capture process using the metabolic transfer using the enzyme carbonic anhydrase. This capacity of algal cells. included validating an assay to quantify enhanced air-CO2 transfer. Objective 4: TEA/LCA Techno-economic The final TEA/LCA concluded that the DAC-based scenario analysis (TEA) and life-cycle analyses offered a 27% increase in algae biomass revenue potential (LCA) to assess both processes. under optimized conditions. Final Report | MicroBio Engineering Inc. ii Summary of Main Results and Recommendations The project successfully demonstrated a technologically feasible path for high-productivity microalgae cultivation using a DAC-simulated CO2 source. • DAC is the most promising path: The DAC Pilot Unit's performance data validated a projection of below $150/Mg CO2 for a full-scale system.. • Strain Identified for high CUE cultivation: The alkaliphilic Chlorella isolate CT3072 met the baseline productivity goal (9.3 g AFDW/m²-day) under conditions projecting a 90% CUE. Recommendation: Future work should focus on a demonstration-scale project to fully integrate the DAC unit with a microalgae cultivation facility to validate the projected $150/Mg CO2 cost in a continuous, outdoor system. Further research is also recommended to ensure long-term stability and cost-effective crop protection strategies for the Chlorella isolate CT3072. Final Report | MicroBio Engineering Inc. iii Contents Acknowledgments, Disclaimer, Protected rights ............................................................................... i Executive Summary ......................................................................................................................... ii Introduction .................................................................................................................................... 9 Goals, objectives, accomplishments .............................................................................................. 11 Project activities ............................................................................................................................ 14 5.1. Alkaliphilic strain screening, cultivation, and scale-up .................................................................. 15 5.1.1. Mapping cultivation conditions (pH, alkalinity) to Carbon Utilization Efficiency (CUE) . 15 5.1.2. Identification of candidate strains, testing in outdoor ponds ........................................ 18 5.1.3. Screening candidate strain response to variations in culture pH and media alkalinity .. 24 5.1.4. Endogenous carbonic anhydrase for air-CO2 capture ..................................................... 24 5.2. Design, Construction and Operation of a Direct Air Capture (DAC) System .................................. 36 5.2.1. Design and Construction ................................................................................................. 36 5.2.2. DAC Operations ............................................................................................................... 38 5.2.3. Techno-Economic Analysis .............................................................................................. 40 5.3. DAC cultivation Trial....................................................................................................................... 42 5.3.1. Methods .......................................................................................................................... 42 5.3.2. Results ............................................................................................................................. 57 5.3.3. Conclusion ....................................................................................................................... 72 5.4. Techno-economic and Life-cycle Assessment................................................................................ 72 Conclusions and Recommendations .............................................................................................. 76 Products developed....................................................................................................................... 78 7.1. Peer reviewed papers .................................................................................................................... 78 7.2. Conference presentations ............................................................................................................. 78 Works Cited .................................................................................................................................. 78 Final Report | MicroBio Engineering Inc. iv Tables Table 5-1: Summary of Cyanotech bioprospected strains. ......................................................................... 19 Table 5-3: Modified Tris-minimal recipe with urea as the nitrogen source and alkalinity increase to 10 meq/L. ......................................................................................................................................................... 25 Table 5-4. Carbon burden and corresponding cost of capture per net tonne for a variety of power and heat source combinations. The source of different assumptions to LCA inputs are provided below the table. ........................................................................................................................................................... 41 Table 5-5: Elemental composition of 35/80 media, Azul well water, and the modified recipe for initial Azul operations. Biomass content is for generic algae and not P. cruentum specifically. ......................... 43 Table 5-6: Literature review of elemental concentrations and their mg/g VSS biomass concentration. Sources: Fithriani (2022), Fuentes (2000), and Di Lena (2020; P. cruentum; Cerri (2021) and Kavitha (2016) analyzed P. purpureum; Brown (1999) are average values for a variety of marine microalgae used in aquaculture. ............................................................................................................................................ 44 Table 5-7: Initial and starting densities and their corresponding productivity for all flasks. Flasks denoted with an asterisk indicate potential contamination in those flasks. ............................................................ 71 Table 5-8: Summary of TEA Parameters Varied in the Operational (i.e., Uncertainty) Scenarios for the DAC Case. .................................................................................................................................................... 73 Table 5-9: Summary of DAC Specific Parameters Varied in the Operational (i.e., Uncertainty) Scenarios 73 Table 5-10: Summary of TEA Modeling Results for the DAC Case .............................................................. 73 Figures Figure 5-1: Fraction of daily light intensity over a 24-hr period (yellow), corresponding projected hourly carbon uptake rate (blue), and DAC-C supply rate (black). ........................................................................ 16 Figure 5-2: Example program output for a case with pH 8.0 morning pH, 5 meq/L alkalinity, 4.5 g C/m2- day biomass productivity, and 30 cm depth. TOP: Diel pH (blue, left axis) and DIC (mg C/L, red, right axis) variations under the specified conditions. Air-equilibrium pH is also shown for reference (black x). BOTTOM: Rates of DAC-C supply, assimilation into biomass, and outgassing. .......................................... 17 Figure 5-3: Projected CUE as a function of alkalinity and pH at 8am, in relation to the project Minimum Target (75%; solid black line). ..................................................................................................................... 18 Figure 5-4: Well plate screening assay........................................................................................................ 19 Figure 5-5: Biomass density (as ash free dry weight, calculated from optical density via correlation) of CT3074 from February-April 2022. Gaps in the data set coincide with weekends when no samples were collected. Harvest days are circled in red. Series are coded based on strain ID (3074) and pond identifier (_7, _8, _10, and _11). ................................................................................................................................ 20 Figure 5-6: Growth of CT3067 from 4/6-4/29, with harvests indicated with red circles. Growth lagged after the 4/12 harvest, and the culture was pre-emptively diluted on 4/26, before reaching the 0.5 g/L target density at harvest, in an attempt at revival. After the 4/26 harvest, the culture was flocculated and rapidly settled, little growth was observed, and was terminated on 5/2/22. ..................................... 21 Figure 5-7: CT3072. Ash free dry weight growth from 8/3-9/20. Harvests are circled and occurred every four days from 8/10-9/19. Alkalinity decreased from 200 to 20 meq/L, and the pH setpoint to 7.5-8.0 on 8/29 to attempt to recover culture from diatoms and grazers. ................................................................. 22 Figure 5-8: Pond pH (upper panel) and biomass productivity (lower panel) during the Summer and Fall CT3072 trials. CUE projects in the lower panel are based on the framework outlined in Section 5.1.1. Vertical lines map changes in pond pH setpoint to the biomass productivity time-course. ...................... 23 Final Report | MicroBio Engineering Inc. v Figure 5-9: Cultivation apparatus used for experiments, with one light panel removed for viewing. (1) – water bath, (2) – reactors, 3 – rotameters, (4) – mass flow meter, (5) - chiller recirculation outlet, (6) - chiller recirculation input. The fluorescent lighting array has been replaced by a higher intensity LED array. ........................................................................................................................................................... 26 Figure 5-10: Chlamydomonas Wildtype (WT, Top) and Chlamydomonas cell wall deficient mutant (CWD , Bottom) biomass concentrations are graphed as a function of optical density to generate a correlation relationship to aide in daily biomass tracking. ........................................................................................... 27 Figure 5-11: Chlamydamonas wild type (WT, Left) and Chlamydamonas mutant cell wall deficient (CWD, Right) reactors removed from the water bath (Figure 5.10) for pre-sampling. ......................................... 27 Figure 5-12: Slope productivity (mg VSS/L-day) as function of time for Chlamydamonas CWD mutant (Top) and Chlamydamonas wildtype (Bottom)........................................................................................... 31 Figure 5-13: Average slope productivity of Chlamydamonas reinhardtii wildtype and mutant categorized by monthly trials and overall. April wildtype 237.1 mg VSS/L-day (SD = 88.8, n = 16) and mutant 145.7 mg VSS/L-day (SD = 58., 3 n=20). May wildtype 372.3 mg VSS/L-day (SD = 123.1, n =13 ) and mutant 226.6 mg VSS/L-day (SD = 90.4., 3 n=16). June wildtype 205.6 mg VSS/L-day (SD = 101.1, n =14 ) and mutant 148.6 mg VSS/L-day (SD = 29.3., 3 n=13). Overall average for wild type of 271.6 mg VSS/L-day (SD = 88.5, n=43) and mutant 173.6 VSS/L-day (SD = 45.9, n=49 ). ............................................................ 31 Figure 5-14: Standard Curve of diluted Bovine Serum Albumin (BSA) with measure absorbance at 562 nm; a best-fit polynomial equation was used (R² = 0.9996). ...................................................................... 32 Figure 5-15: Protein concentration of Chlamydamonas reinhardtii (μg/mL) Mutant (top) and wildtype (bottom) during 4 day growth periods calculated with BSA standard curve. ............................................ 33 Figure 5-16. Rendering of the DAC pilot unit – walls hidden. ..................................................................... 36 Figure 5-17. Close up of downstream equipment. ..................................................................................... 37 Figure 5-18. DAC pilot unit container with roll up doors. ........................................................................... 37 Figure 5-19. Photos of the exterior of the DAC pilot unit during the commissioning phase...................... 38 Figure 5-20. Relative humidity (orange line) and ambient temperature (°C, blue line) for duration of DAC pilot unit trial. ............................................................................................................................................. 39 Figure 5-21. CO2 product purity as measured from the vacuum pump discharge from 8/22/24 through the end of the DAC pilot unit trial. .............................................................................................................. 40 Figure 5-22. Average CO2 product purity for K-Series plant running over multiple months at >90%. ....... 40 Figure 5-23: From left to right, 3.4C, 3.4D, and the 22 m2 raceways inoculated with P. cruentum. .......... 45 Figure 5-24: Dry CO2 Output time series from Global Thermostat’s DAC trial. .......................................... 46 Figure 5-25: Example of mass flow controller daily CO2 flowrate to simulate the average DAC daily CO2 output. In this figure, the average CO2 flowrate is set to 290 mL CO2/min. ............................................... 47 Figure 5-26: Example carbon balance in the 3.4 and 22 m2 raceways from 10/8 to 10/21/2024. The left stacked bar in both plots is the net change in VSS carbon (blue), NPOC (orange), and DIC (yellow). The right stacked bar in both plots is the carbon sources: injected carbon (purple) and Azul well water DIC (yellow); and carbon losses: outgassing (orange) and injection losses (blue). The black dots represent the net value of the total stacked bars. ............................................................................................................ 51 Figure 5-27: (LEFT Panel) Measured DIC samples via coulometry and the sample’s pH and alkalinity. (RIGHT Panel) CO2SYS calculated DIC based on the sample’s pH and alkalinity compared to the measured DIC. The 1:1 reference line indicates that CO2SYS calculated DIC perfectly aligns with measured DIC values................................................................................................................................... 53 Figure 5-28: Time series of atmospheric CO2 concentration in the greenhouse throughout the cultivation trial. ............................................................................................................................................................. 54 Figure 5-29: Schematic of air water and flow through an experimental marble equilibrator set up used to measure in situ pCO2 in the water. ............................................................................................................. 56 Final Report | MicroBio Engineering Inc. vi Figure 5-30: As-received DAC water samples. Left is the "DAC unit condensate", and right is the low- volume sample of accumulated water in the compressor. These were amended with nutrients and used in a cultivation trial to assess whether DAC products may contain compounds inhibitory to algal growth. .................................................................................................................................................................... 57 Figure 5-31: Raceway temperature and daily average solar irradiance throughout the cultivation trial. . 58 Figure 5-32: Time series of pH and CO2 injection flowrate throughout the cultivation trial...................... 59 Figure 5-33: Biomass concentration (mg VSS/L; UPPER Panel) and productivity (g VSS/m2-day; LOWER Panel) time series in the 3.4C and 22 m2 raceways. Biomass productivity was calculated from start to end of a given growth cycle, around 1-3 weeks depending on pond operations. Horizontal lines in the lower panel denote time periods over which productivity was calculated. Black circles indicate harvest dates and asterisks indicate raceway re-inoculation. .......................................................................................... 60 Figure 5-34: Microscopy of predatory flagellates (LEFT) and Picochlorum (RIGHT) contamination in the 22 m2 raceway observed on 11/1. ................................................................................................................... 61 Figure 5-35: Time series of NPOC (used as a proxy for dissolved polysaccharides) in the raceways. ........ 62 Figure 5-36: Time series of settling efficiency from Imhoff cones for each raceway. “3.4C+22 Mix” refers to an Imhoff cone containing a composite sample of the total 3.4C raceway volume added to and mixed with the full 22 m2 raceway. ....................................................................................................................... 63 Figure 5-37: Biomass gravity sedimentation after eight minutes, observed on 11/26. The background green color indicates likely Picochlorum contamination. ........................................................................... 63 Figure 5-38: Time series of harvest operations and resulting centrifuge operation flow biomass concentrations. ........................................................................................................................................... 64 Figure 5-39: Water usage breakdown for each growth period throughout the cultivation trial. “Recycled Media” refers to growth media recycled following a harvest whereas “Fresh Media” indicates newly prepared growth media in fresh Azul well water. “Evap Top Off” refers to Azul well water added to raceways to replenish evaporation losses and “Inoculum” refers to media with culture used to re- inoculate the raceways. The percentage above each bar denotes the percentage of water recycled. ..... 65 Figure 5-40: Time series of pH (UPPER LEFT), DIC (UPPER RIGHT), and driving force (LOWER LEFT) throughout the outgassing trial in the 22 m2 raceway. The slope of linear regression between the driving force and flux (LOWER RIGHT) represents the mass transfer coefficient. ................................................. 66 Figure 5-41: Time series of pH (Upper Left), DIC (Upper Right), and driving force (Lower Left) throughout the outgassing trial in 3.4C. The slope of the linear regression between the driving force and flux (Lower Right) represents the mass transfer coefficient. ........................................................................................ 67 Figure 5-42: Outgassing pH and pCO2 time series(Upper Left); pH and pCO2 predicted DIC time series compared to coulometry measured DIC (Upper Right); Driving force time series calculated from measured DIC, pH, and pCO2 (Lower Left); Flux versus driving force linear regression and resulting mass transfer coefficient calculated with pH, pCO2, or DIC data (Lower Right). ................................................. 68 Figure 5-43: Injection efficiency results for the three trials calculated over each tenth decrease in pH. High and low flowrates correspond to 350 mL CO2/min and 175 mL CO2/min, respectively. ................... 69 Figure 5-44: Time series of carbon values throughout the injection period. “C injected” refers mass of carbon sparged into the raceway, “DIC Increase” represents the increase in DIC, and “C Outgassed” refers to the mass of injected carbon that has outgassed. ........................................................................ 69 Figure 5-45: CUE and closure for the 3.4C and 22 m2 raceways and the corresponding period over which they were calculated. The black dashed line indicates the assumed injection efficiency (30%). .............. 70 Figure 5-46: Unknown flocs present in two of the Azul media flasks at the end of the growth trial. ........ 72 Figure 5-47: Comparison of Plant Gate Production Costs for Chlorella Based Nutraceutical Products across Baseline, Intermediate, and Final Operational Scenarios. ............................................................... 74 Figure 5-48: Comparison of Life Cycle GWP Weighted Emissions for Chlorella Based Nutraceutical Products across Baseline, Intermediate, and Final Operational Scenarios (50 meq/L case depicted). ...... 74 Final Report | MicroBio Engineering Inc. vii Figure 5-49: Effect of Increasing Alkalinity on Total Plant Gate Production Costs across Baseline, Intermediate, and Final Operational Scenarios. ......................................................................................... 75 Figure 5-50: Effect of Increasing Alkalinity on Total GWP Emissions across Baseline, Intermediate, and Final Operational Scenarios. ....................................................................................................................... 75 Final Report | MicroBio Engineering Inc. viii Introduction Microalgae production of biofuels and bioproducts is a promising technology to achieve the goal of a renewable economy through the utilization and recycling of waste and atmospheric CO2. The development of low-cost microalgae production systems is still in the Research & Development (R&D) phase, with current commercial production limited to relatively high value products (>$10,000/Mg, plant gate). Significant progress is being made in advancing this technology, and while not yet able to produce biofuels as a target product, current technology could do so in combination with wastewater treatment or other co-products/services. However, for large-scale impacts, stand-alone biofuel production processes will be needed. A major limitation has become apparent: although enormous quantities of waste CO2 are emitted by our industrial economies, they are available to only a limited extent for microalgae production. Locations favorable for algae production, with suitable climate and available land and water resources, rarely have a ready-at-hand nearby source of enriched CO2. Transport of dilute power plant flue gas CO2, the most widely available source, is limited economically to close-by algae cultivation systems (<10 km). Concentration, compression and long-distance transport of flue gas CO2 is also prohibitively costly for algae production. Moreover, such fossil CO2 flue gas sources will not be available in the long-term, nor are they inherently renewable. Thus, there is an imperative to develop processes using air-CO2 in algae cultivation. Algae cultivation on air levels of CO2 is currently limited by the transfer of CO2 from the atmosphere into liquid cultures, even into relatively shallow or well-mixed raceway ponds or attached biofilm cultivation systems. The need is to deliver CO2 from air for algae production at a cost of <$100/Mg (metric ton). Photosynthesis by higher plants globally captures ten-fold more CO2 from the atmosphere than emitted from fossil fuels combustion, but practically all of this is re-emitted. Microalgae fix enormous quantities of air CO2 globally, about equal to that of higher plants, but essentially all this CO2 is returned to the atmosphere in a short time. Microalgae can utilize atmospheric CO2 concentrations once the CO2 diffuses into the culture medium, with only a modest reduction in productivity compared to cultivation on enriched CO2. The limitation is the transfer of CO2 from air into the cultures at a rate that can sustain the high productivities required in large-scale and low-cost microalgae biomass production for commodities, biofuels, animal feeds and other commodities. In nature, such high microalgae productivities are observed only in environments high in nutrients, such as hypereutrophic estuaries or alkaline lakes exhibiting a high carbon (C) turnover. In brief, the relevance and objective of this research project was to overcome the current limitations in supplying and using atmospheric CO2 for high-productivity algae cultivation. The technology advanced in this project is the utilization of CO2 from air to cultivate microalgae and produce biomass for higher value products in the near-term, and commodities, such as feeds, fuels, chemicals and other bioproducts in the longer-term. Two approaches were examined: 1. Direct Air-CO2 Capture (DAC): A physical-chemical DAC process was provided by Global Thermostat (acquired by Zero Carbon Systems in 2024) that can deliver a near 100% concentrated CO2 stream to the algal cultures on a continuous basis. 2. Direct Uptake by Cultures: The use of the algal cultures and cultivation systems themselves, to provide CO2 absorption from air at a rate supporting algal biomass production approaching that with high CO2 sources. Final Report | MicroBio Engineering Inc. 9 Currently, there are no validated baseline technologies for either approach. CO2 capture from air (DAC), by processes such as that developed by Global Thermostat, are not yet commercial. Cultivation of algae on air-levels CO2, at high productivity, has been claimed by some but not widely studied. The main objective of Topic Area 3: Algae Bioproducts and CO2 Direct-Air-Capture Efficiency, in this solicitation DE-FOA-0002203, addressed by this project, was to increase the national resource potential for algal biomass production by decoupling it from enriched sources of CO2, such as power plant flue gases. The U.S. Department of Energy (DOE) supported research had shown that such CO2 sources limit algae biomass production in the US to less than 10% of what would be possible with air-CO2 sources (U.S. DOE 2016; Davis et al. 2018; Venteris et al. 2014). A straightforward approach to this objective would be to use a mechanical/chemical DAC process that delivers high concentration CO2 to an algae production facility. This was the major focus proposed for this project, but as such a process has not yet been demonstrated in practice, the alternative of using the algal cultures directly in absorbing and utilizing air CO2 was also proposed. The project had four key objectives: • A near-term small-scale demonstration of a technologically feasible and potentially economically viable process to supply air-CO2 for microalgae production using DAC. • Demonstration of the production of Chlorella sp. with such an air-CO2 source as a commercial algae bioproduct to be produced in the USA in the near-term, at a scale and market of a thousand tons per year. • Proof-of-concept of an alternative direct air-CO2 capture process which uses the metabolic capacity of algal cells to increase the capture of CO2 from air. • Techno-economic analysis (TEA) and life-cycle analyses (LCA) to assess both processes, considering productivity, energy, and chemical input trade-offs, with respect to their potential scale-up within the United States for biofertilizers, bioplastics, and biofuel production. This report documents the work performed under Award DE-EE0009276 to address these objectives. Final Report | MicroBio Engineering Inc. 10 Goals, objectives, accomplishments The project's overarching goal was to advance the technology for large-scale, low-cost microalgae biomass production in the United States by decoupling it from enriched sources of CO2, such as power plant flue gases. The U.S. Department of Energy (DOE) has identified that reliance on enriched CO2 sources limits the national resource potential for algal biomass production. The project was focused on four key objectives, with the accomplishments summarized below. Objective 1: Near-term small-scale demonstration of a technologically feasible and economically viable process to supply air-CO2 for microalgae production using DAC. Objective Key Accomplishment/Result Installation and operation of a >1 kg The DAC Pilot Unit was designed, constructed, and CO2/hour DAC Pilot Unit. operated for a 2,000-hour operational campaign, producing CO2 at up to 90% (v/v) purity. Operational data, including CO2 output rate and purity, were collected and provided for cultivation simulations and techno-economic analysis, projecting a 75% reduction in CO2 costs for a full- scale system ($600/Mg to $150/Mg). Objective 2: Demonstration of microalgae production with such an air-CO2 source as a commercial algae bioproduct to be produced in the USA in the near-term, at a scale and market of a thousand tons per year. Objective Key Accomplishment/Result Demonstration of alkaliphilic Chlorella An annual average biomass productivity of 9.3 g cultivation in 1 m³ raceway ponds, with AFDW/m²-day was measured for Chlorella isolate CT3072 a biomass productivity matching the under conditions (200 meq/L alkalinity, pH 9.6 setpoint) baseline value (9.3 g AFDW/m²-d, annual that projected a CUE of 90%. Fall productivity was 8.6 g average) but under conditions that AFDW/m²-day at 80% projected CUE. project a DAC carbon utilization efficiency (CUE) of 75%. Demonstration of alkaliphilic microalgae Fall biomass productivity of Porphyridium sp. operated cultivation achieving 10.5 g AFDW/m²-d under conditions amenable to >90% CUE measured 5 – 15 (annual average, a 13% increase from g AFDW/m²-day, with near term applications as a baseline) with simulated air-CO2, under hydrocolloid, protein, and natural food coloring. conditions that project a DAC carbon utilization efficiency of at least 75%. Final Report | MicroBio Engineering Inc. 11 Objective 3: Proof-of-concept of an alternative direct air-CO2 capture process which uses the metabolic capacity of algal cells to increase the capture of CO2 from air. Objective Key Accomplishment/Result Evaluate the potential of microalgae Exploratory research was conducted to establish proof-of- cultures to directly capture CO2 from air. concept for biochemically enhanced air-CO2 transfer using endogenous bioactive compounds (carbonic anhydrase). The task was initiated with efforts to validate a reproducible assay to quantify enhanced air-CO2 transfer, with subsequent steps to assess enhancement via extracted compounds and excretion products. Objective 4: Techno-economic analysis (TEA) and life-cycle analyses (LCA) to assess both processes, considering productivity, energy, and chemical input trade-offs, with respect to their potential scale- up within the United States for biofertilizers, bioplastics, and biofuel production. Objective Key Accomplishment/Result TEA/LCA studies projecting a 15% or Initial TEA/LCA studies supported the progression, with higher increase in algae biomass current TEA revenue potential projections exceeding revenue potential from the baseline $6,000/Mg. The project successfully demonstrated a technologically feasible pathway for microalgae cultivation using a simulated DAC-CO2 source while also advancing the understanding of direct air-CO2 uptake. • DAC Feasibility: The DAC Pilot Unit successfully operated and provided real-world data, enabling projections that significantly reduce the cost of CO2 compared to non-DAC merchant CO2 for a future full-scale system ($150/Mg vs. ~$1,000/Mg). • Alkaliphilic Strain Performance: The project confirmed the ability to meet the baseline annual average biomass productivity of 9.3 g AFDW/m²-day with the alkaliphilic Chlorella isolate CT3072 under high CUE conditions (90% projected CUE). • Direct Uptake Research: The exploratory research into biochemically enhanced air-CO2 transfer established the framework for a future proof-of-concept study utilizing the enzyme carbonic anhydrase. Final Report | MicroBio Engineering Inc. 12 • Economic and Environmental Assessment: The TEA/LCA confirmed that the expansion of resource potential, not cost reduction, is the main benefit of air-CO2 sources for algae production, as the cost of alkalinity in the high-pH direct capture method was comparable to the projected costs of DAC-derived CO2. Final Report | MicroBio Engineering Inc. 13 Project activities Project activities are reported in four chapters: Chapter 5.1 (Alkaliphilic strain screening, cultivation, and scale-up) developed a framework for mapping cultivation conditions (pH, alkalinity, biomass productivity) to Carbon Utilization Efficiency (5.1.1), then applied this in identifying strain candidates (5.1.2) and later screening these candidates to simultaneously optimize biomass productivity and CUE (5.1.3). Sub-chapter 5.1.4 explored the endogenous production of carbonic anhydrase. Chapter 5.1: Alkaliphilic strain screening, cultivation, and scale-up 5.1.1: Mapping cultivation conditions (pH, alkalinity) to Carbon Utilization Efficiency (CUE). 5.1.2: Identification of Candidate Strains - Characterization of top candidates’ response as a function of pH and alkalinity. 5.1.3: Screening candidate strain response to variations in culture pH and media alkalinity 5.1.4. Endogenous carbonic anhydrase for air-CO2 capture Chapter 5.2 describes design, construction, operations, and techno-economic outcomes of a direct-air capture pilot system. 5.2: Design, construction, and operation of a direct-air capture (DAC) system. 5.2.1. Design and Construction 5.2.1. DAC Operations. 5.2.3. Techno-Economic Analysis Chapter 5.3 describes a cultivation trial involving simulated provision of DAC pilot system generated CO2. CO2 output rate (kg CO2/hr) from the 2,000 DAC system trial, incorporating diel variations in output and continuous (24/7) provision to the algal culture, was reproduced in an outdoor cultivation trial. A carbon balance compared CUE for traditionally managed (CO2 provided on-demand based on culture pH) versus DAC-simulated (24/7 continuous provision of DAC-CO2) methods of carbon supply. 5.3. DAC Cultivation Trial 5.3.1. Methods 5.3.2. Results 5.3.3. Conclusions Chapter 5.4 (Techno-economic and Life-cycle Assessment) describes techno-economic (minimum biomass selling price; $/Mg) and life-cycle outcomes. Final Report | MicroBio Engineering Inc. 14 5.1. Alkaliphilic strain screening, cultivation, and scale-up 5.1.1. Mapping cultivation conditions (pH, alkalinity) to Carbon Utilization Efficiency (CUE) To guide the selection and optimization of strains capable of high productivity and high Carbon Utilization Efficiency (CUE), in-silico calculations were performed to estimate CUE as a function of operating pH and media alkalinity. The immediate objective was to establish precise operating parameters for ongoing cultivation trials, verifying that the tested conditions consistently achieved a CUE greater than 75% to support downstream TEA and LCA models. A secondary aim to was to effectively size and pair the DAC system CO2 output with an appropriately sized cultivation system. Furthermore, this effort sought to identify variations in diel pH and carbon speciation associated with a continuous 24/7 DAC-CO2 supply, and to evaluate how these dynamics respond to different alkalinity levels. Methods Inorganic carbon concentration in an outdoor raceway pond mass cultures was modeled over a diel- cycle by accounting for sources and sinks inorganic carbon (IC). Sources and sinks included algal assimilation, CO2 injected via external sources (i.e. DAC system), outgassing (volatilization) losses, or carbon in-gassed via in-pond air-CO2 absorption – applicable if the culture pH reaches above air- saturation. The following simplifying assumptions were also applied: • Diel hourly productivity is estimated based on a typical daily solar irradiance profile, as shown schematically in Figure 5.1, such that the daily integral of biomass productivity reaches a user- selectable value. • 10 g AFDW/m2-day biomass productivity (4.5 g C/m2-day) is used as default, assumed constant regardless of pH or alkalinity. • Constant equilibrium coefficients, assuming no variation in response to diel changes in pond water temperature. • No inorganic carbon is released from biomass at night (dark respiration). • Dissolved inorganic carbon (DIC) returns to a user-specified pH at the end of the 24-hr cycle • User-selectable CO2 injection efficiency, using a default value of 90% assumed for injection via sumps (approximately 1 m depth) • Variations in pH between carbonation stations are not accounted for – the system is modelled as a continuous stirred tank reactor (CSTR), an assumption that is unlikely to hold true for systems larger than 1 m3 • User selectable pond depth • 0.05 m/hr transfer coefficient across the raceway pond air-water interface, as is typical for outdoor open ponds. Culture pH and alkalinity were varied to project CUE at different operating points. Calculations were performed by Matlab v.2018b using the CO2sys Matlab package. The calculation was initiated by the user-specified pH and alkalinity and default values for controlled variables (pond depth, biomass productivity, injection efficiency, mass-transfer coefficient). DIC at the subsequent time-step was calculated from the cumulative effects of DAC-supplied C, with an initial flow rate guess set equal to the biomass productivity (on a carbon basis), DIC incorporated into algal biomass, based on the assumed 24- hr growth profile, as well as DIC lost due to outgassing or gained from ingassing. This calculation was repeated over the 24-hr period in 1-hr time steps, then took an additional step, the ‘25th hour’, to re-calculate the initial DIC value. If the rate of continuously supplied DAC-C perfectly matched carbon demand, the initial and final DIC values should be equal. In the case where ‘25th hour’ Final Report | MicroBio Engineering Inc. 15 DIC was less than the initial DIC, the rate of DAC-C supply was iteratively increased until meeting a user- specified tolerance. Likewise, if ’25th-hour’ DIC was higher than the initial DIC, the DAC-C supply rate was iteratively lowered, again until initial and ’25th-hr’ DIC were within the specified tolerance. The primary calculation output was a prediction of CUE – defined as the amount of carbon fixed into algal biomass relative to the amount of carbon supplied – as a function of the primary variables, including, most importantly, biomass productivity, pH set-point, and culture media alkalinity. A secondary output was the DAC-CO2 flow rate required for a given cultivation area (m2), subsequently used to determine the CO2 flow rate in the simulated DAC-CO2 cultivation trial described in Chapter 5.3. 0.5 1.0 DAC-C supply 0.4 C uptake rate 0.9 Fraction of daily light 0.8 Fraction of daily maximum irradiance 0.4 Carbon supply, uptake, g C/m2-hr 0.7 0.3 0.6 0.3 0.5 0.2 0.4 0.2 0.3 0.1 0.2 0.1 0.1 0.0 0.0 0 2 4 6 8 10 12 14 16 18 20 22 24 Hour of day Figure 5-1: Fraction of daily light intensity over a 24-hr period (yellow), corresponding projected hourly carbon uptake rate (blue), and DAC-C supply rate (black). Results Calculation output for a scenario with constant 24/7 CO2 supply is shown below in Figure 5.2. pH at 8:00 was user-specified at 8.0 (blue triangle, upper panel). pH increases after 8am due to the combined effects of DIC loss via outgassing and algal assimilation exceeding the rate of carbon supply. pH reaches a maximum near 8.8 at 18:00, before decreasing back to the initial value at 8:00 as the rate of carbon supply ‘catches up’ to the rate of carbon losses. Pond pH is super-saturated with CO2 (dCO2bulk > dCO2air- equilibrium) for all but between 14:00 and 22:00 hours, meaning that CO2 is lost to outgassing over a portion of the day, as indicated by the positive outgassing rate during the early morning (blue circles, lower panel). The slight mismatch between the values calculated at 8am, perhaps most obvious with the blue pH triangles, reflects the user specified acceptable tolerance of initial and ’25-th hour’ DIC. CUE is calculated by dividing the amount of biomass grown by the amount of DAC-C supplied, accounting for an assumed 90% injection efficiency – in this case equal to 82% Final Report | MicroBio Engineering Inc. 16 Figure 5-2: Example program output for a case with pH 8.0 morning pH, 5 meq/L alkalinity, 4.5 g C/m2-day biomass productivity, and 30 cm depth. TOP: Diel pH (blue, left axis) and DIC (mg C/L, red, right axis) variations under the specified conditions. Air-equilibrium pH is also shown for reference (black x). BOTTOM: Rates of DAC- C supply, assimilation into biomass, and outgassing. Calculations were performed for scenarios spanning pH set-points from 7.0 to 10.0, for alkalinities of 5, 20, 100, and 200 meq/L, with results summarized in Figure 5.3 for the 30 cm depth and 4.5 g C/m2-day biomass productivity case. As pH is expected to vary throughout the day, calculations are based upon a specified value at a certain time of day, in this case arbitrarily set at 8:00. In general, at low pH setpoint, high relative rates of outgassing result in low CUE. Increasing pH results in higher CUE via reduced outgassing losses. As alkalinity increases, bulk dissolved CO2 concentration increases for a given pH, raising the relative outgassing rate and lowering CUE. For the 5 meq/L (orange), CUE exceeded 75% for a morning pH of 7.8 or higher. Above pH 8.4, CUE exceeded 90%, the theoretical limit based on injection efficiency losses, due to carbon absorbed from the atmosphere. This same effect is seen at the 20 meq/L alkalinity level at pH 9.5, although not seen at the 100 or 200 meq/L levels, likely because the media is so heavily buffered that the pH rarely exceeds air-equilibrium at the selected biomass productivity. Final Report | MicroBio Engineering Inc. 17 30 cm depth; 4.5 g C/m2-day productivity 100 90 80 70 CUE (90% IE) 60 50 5 meq/L 40 20 30 100 20 200 10 Minimum Target 0 7 7.5 8 8.5 9 9.5 10 pH at 8:00 Figure 5-3: Projected CUE as a function of alkalinity and pH at 8am, in relation to the project Minimum Target (75%; solid black line). Conclusions The framework developed here was used to guide selection of experimental parameters in the Cyanotech and PNNL strain screening trials, to inform TEA/LCAs, as well as project CO2 mass flow rates required in the subsequent simulated DAC-CO2 cultivation trial. 5.1.2. Identification of candidate strains, testing in outdoor ponds The aim of this task was to isolate and coarsely down select candidate strains for subsequent optimization of growth conditions at PNNL (5.1.3). Cultivars of the Chlorella genus were identified as near-term targets. The down selection pipeline consisted of an indoor micro-well plate assay, indoor cultivation in carboys (data not shown), and cultivation in outdoor ponds. Methods Indoor screening An indoor strain screening assay was developed to identify candidates for outdoor cultivation. Chlorella isolates, collected from Cyanotech production ponds and maintained on agar plates, were cultivated in a well-plate format in a controlled environment chamber (Figure 5.4). Well plate growth was monitored for five weekdays over a seven day grow cycle, with biomass density estimated from optical density. Lead candidates for outdoor cultivation were identified as those with the largest change in growth. Candidates increasing by at least 300% were tested in outdoor 200 L open ponds. Final Report | MicroBio Engineering Inc. 18 Figure 5-4: Well plate screening assay. Outdoor screening Up-selected strains were tested in outdoor open ponds, and, to a limited extent, outdoor enclosed photobioreactors. Cultures were grown in defined culture media and operated semi-continuously, diluted to approximately 100 mg AFDW/L after reaching a density between 300 – 1,000 mg AFDW/L. Culture density was monitored as the optical density at 680 nm, and converted to AFDW (mg/L) based on a pre-determined OD to AFDW calibration curve. 200 meq/L was used as the baseline alkalinity, and was varied, along with pH, depending on culture performance. Biomass productivity was calculated as the increase in biomass concentration from one sampling day to the next, divided by the time interval. Results Up-selected strains from Indoor screening are summarized in Table 5.1, consisting of one strain from the Parachlorella (CT3074) taxonometric group, and three of the genus Graesiella (CT3069, CT3067, CT3072). All four were shipped to PNNL for additional screening. Table 5-1: Summary of Cyanotech bioprospected strains. Strain ID Taxonometric Sent to Testing summary group PNNL? CT3069 Graesiella Yes, • PNNL-CT MTA in place 12/2021 01/2022 • Down selected Oct-Dec 2021 due to poor growth relative to CT3074 CT3074 Parachlorella Yes, • Tested at 200 L scale Jan-Apr 01/2022 • Tested at 600L photobioreactor scale Jun • Tested at 1000L minipond scale Jun o Downgraded to first alternate status CT3067 Graesiella Yes, • Flask Scale Feb-Mar 06/2022 • Outdoor Scale Mar-May o Down selected due to poor growth in outdoor conditions relative to CT3074 CT3072 Graesiella Yes, • Well plate scale in late June 08/2022 • Outdoor Scale July-current Final Report | MicroBio Engineering Inc. 19 o Lead Candidate Lead candidates from the indoor assay were evaluated for mass cultivation beginning in Winter 2022. CT3074 (Figure 5-5) struggled to survive outdoors for more than three weeks, and it had difficulty reaching the project baseline biomass productivity (9.3 g AFDW m-2 day-1), averaging 5-7 g AFDW m-2 day-1. 74_10 crashed on 3/17 and was reinoculated with culture from 74_7. 74_7 and 74_8 were drained and re-inoculated from lab cultures on 3/28 for a comparison with an alternative chelator (EDDHA iron Chelate). Ponds 74_10 and 74_11 crashed due to pest stress on 3/30. The alternative chelator pond trial continued to 4/5, when ponds were cleaned and drained after the target Chlorella was overtaken by an unknown diatom(s). 74_8 was drained and cleaned 4/5, again due to a large proportion of diatoms. 74_7 and 8 were cultivated at 200 meq/L alkalinity and a pH 9.4-9.6 setpoint from 4/12-4/28, although biomass productivity continued to fall below project targets. 0.38 3/8/22 0.36 0.34 4/12/22 4/19/22 3/1/22 3/17/22 Biomass Density, g/L 0.32 0.3 4/5/22 3/22/22 4/26/22 0.28 0.26 0.24 0.22 0.2 2/18/2022 3/20/2022 4/19/2022 5/19/2022 74_7 74_8 74_10 74_11 Figure 5-5: Biomass density (as ash free dry weight, calculated from optical density via correlation) of CT3074 from February-April 2022. Gaps in the data set coincide with weekends when no samples were collected. Harvest days are circled in red. Series are coded based on strain ID (3074) and pond identifier (_7, _8, _10, and _11). CT3067 was cultured outdoors at 200 meq/L alkalinity at pH 9.5 +/- 0.1 from April-June 2022 (Figure 5.6). Biomass productivity measured 2 g AFDW/m2-day on the first grow-out, and stagnated after dilution. CT3067 was downselected from further trials. Final Report | MicroBio Engineering Inc. 20 1.6 1.4 1.2 Biomass density 1 4/12/22 0.8 4/26/22 0.6 0.4 0.2 0 3/30/2022 4/4/2022 4/9/2022 4/14/2022 4/19/2022 4/24/2022 4/29/2022 5/4/2022 OD(@680) DW (g/L) AFDW Figure 5-6: Growth of CT3067 from 4/6-4/29, with harvests indicated with red circles. Growth lagged after the 4/12 harvest, and the culture was pre-emptively diluted on 4/26, before reaching the 0.5 g/L target density at harvest, in an attempt at revival. After the 4/26 harvest, the culture was flocculated and rapidly settled, little growth was observed, and was terminated on 5/2/22. CT3072 was initially cultured at 200 meq/L alkalinity before it began to auto-flocculate and show signs of stress after six growth cycles from 8/3-8/29 (Figure 5.7). Biomass density at harvest progressively decreased over this period. The alkalinity and pH of the culture was lowered on 8/29 to 20 meq/L and pH 7.5-8, in part to allow for testing repeated daily bleach treatments (10 ppm, added post-harvest) as a means of mitigating pest pressures. At the beginning of the 20 meq/L, pH 7.5-8, with bleach treatment trial, appreciable quantities of diatoms were observed, and CT3072 biomass was heavily flocculated. Within three days, diatom cell pigmentation decreased, as observed via microscopic examination, and the culture recovered (i.e. predominantly vibrant green, non-flocculating CT3072) after four days. The alkalinity set-point was increased to 100 meq/L and pH setpoint to 9.4-9.6 on 9/26. Diatoms and grazers quickly overtook the culture and alternative pest mitigation strategies with hydrogen peroxide proved ineffective at low doses (5-10ppm) and deadly to CT3072 at higher doses (25-50ppm) required to control the pest population. Final Report | MicroBio Engineering Inc. 21 0.5 0.45 0.4 0.35 0.3 g/L 0.25 0.2 0.15 0.1 0.05 0 7/28/2022 8/7/2022 8/17/2022 8/27/2022 9/6/2022 9/16/2022 9/26/2022 72_7 72_9 Figure 5-7: CT3072. Ash free dry weight growth from 8/3-9/20. Harvests are circled and occurred every four days from 8/10-9/19. Alkalinity decreased from 200 to 20 meq/L, and the pH setpoint to 7.5-8.0 on 8/29 to attempt to recover culture from diatoms and grazers. The framework for estimating CUE as a function of pH setpoint and alkalinity (Section 5.1.1) was applied to the CT3072 trial (Figure 5.8) to assess tradeoffs between biomass productivity and resource (CO2) use efficiency. During the initial CT3072 operating period, 8/3 – 8/27, pH was controlled to pH 9.5, and alkalinity to 200 meq/L, projecting a CUE of 90%. Little pest pressure was observed over this time, and duplicate ponds (72_9, 72_7) tracked reasonably well. Decreasing the pH set-point to pH 8.0 (8/31 – 9/27) and alkalinity to 20 meq/L, mapping to a projected CUE of 40%, resulted in an increase in biomass productivity, despite slightly lower irradiance. Poor growth was observed after increasing pH and alkalinity to 9.5 and 60 meq/L respectively (90% CUE, 10/2 – 10/21), partially recovering after reducing pH (9.0) and increasing alkalinity (100 meq/L; 80% CUE). While generally informative, this assessment is limited by short experimental duration – trials longer than approximately two weeks per condition, as tested herein, are required to more definitively understand tradeoffs between biomass productivity and carbon utilization. Regardless, biomass productivity measured 9.3 g AFDW/m2-day, under conditions that exceed 75% CUE, meeting the interim project Go/No-Go Criteria. Final Report | MicroBio Engineering Inc. 22 Figure 5-8: Pond pH (upper panel) and biomass productivity (lower panel) during the Summer and Fall CT3072 trials. CUE projects in the lower panel are based on the framework outlined in Section 5.1.1. Vertical lines map changes in pond pH setpoint to the biomass productivity time-course. Conclusions Chlorella isolate CT3072 was the top performing strain (9.3 g AFDW m-2 day-1 from 8/30 to 9/16, Figure 5.9) and was the anticipated strain for the larger-scale DAC trials. However, further trials should revisit CT3067 and CT3074 when grown at low (20 meq/L or lower) alkalinity, where pest management strategies (10 ppm bleach treatment, applied post-harvest) were previously shown to differentially favor cultivation of Chlorella isolate CT3072. Culture instability limited cultivation campaign duration, often to periods of two weeks or less. Pest pressures, and correspondingly lack of suitable crop protection strategies, are hypothesized to be responsible for the observed instability. Prophylactic bleach treatment was the most reliable management approach tested over the CT3072 campaign, although effectiveness is limited by the near- neutral pKa (7.5) of the hypochlorous acid-hypochlorite ion disinfection system. For example, at the pH set-point (9.0 – 9.5) used in the Summer and Fall CT3072 trials, the minimally effective hypochlorite ion dominates, despite achieving high CUE. While it is difficult to isolate the individual effects of pH, alkalinity, and pest management strategy, CT3072 cultures were most productive and stable at near- Final Report | MicroBio Engineering Inc. 23 neutral pH. Temporary operation at a reduced pH set-point commiserate with higher disinfectant activity may offer a compromise between culture stability (pest management) and biomass productivity, although may not be feasible at high (100 – 200 meq/L) alkalinity, when outgassing losses incurred while approach a near-neutral pH may be cost prohibitive. 5.1.3. Screening candidate strain response to variations in culture pH and media alkalinity Project activities in this section are described in the following in-preparation manuscript: Improving biomass productivity of the alkaliphilic microalga Graesiella emersonii Song Gao a, Braden Crowe b, Charley O’Kelly c, Scott Edmundson a, John Benemannb, Michael Huesemann a Pacific Northwest National Laboratory, 1529 W Sequim Bay Rd., Sequim, WA 98382, USA b MicroBio Engineering Inc., PO Box 15821, San Luis Obispo, CA 93406, USA c Cyanotech Corporation, 73-4460 Queen Kaahumanu Hwy., Suite #102, Kailua-Kona, HI 96740 Abstract We explore optimization of two culture management alternatives for the cultivation of Graesiella emersonii – inorganic carbon regime (alkalinity, pH) and standing biomass concentration (dilution). Cultivation of alkaliphilic microalgae allows for efficient CO2 utilization but providing a high alkalinity growth medium can incur a significant cost. Biomass productivity of the alkaliphilic Graesiella emersonii strain was measured at varying pH (7, 8, 9, 10) and alkalinity (5, 25, and 200 meq L⁻¹) levels. When cultivated in a photobioreactor under batch cultivation conditions that replicated an outdoor pond light and temperature environment, biomass productivity was marginally higher at the optimal pH of 9 and an alkalinity of 25 meq L⁻¹, 9.5 versus 8.3 g m⁻² day⁻¹ (pH 8, 200 meq/L), indicating the feasibility of reducing cultivation cost by lowering alkalinity. Optimal dilution of the culture maintained the culture within a concentration range that maximized the net biomass accumulation over the light and dark cycles, which increased biomass productivity by 36% over the baseline batch cultivation. The results showed that reducing alkalinity requirements in the cultivation of such alkaliphilic microalgae algae is possible and appropriate diluting operation could significantly improve biomass productivity, although further studies are needed to validate these results in outdoor ponds under varying seasonal and climatic conditions. 5.1.4. Endogenous carbonic anhydrase for air-CO2 capture In addition to exploring provision of CO2 captured via an external direct air capture systems, a second air-CO2 capture approach - utilizing endogenous bioactive compounds, aimed to accelerate the rate of air-CO2 absorption across the pond air-water interface. The experimental approach involved growth of microalgae reported or suspected to produce carbonic anhydrase specifically localized external to the cell wall, or released into the culture media, where it could then play an active role in air-CO2 mass- transfer. Work by Kimpel et al. (1983) demonstrated that the majority of CA activity in the green alga Chlamydomonas reinhardtii is located outside the plasmalemma, suggesting localization in the periplasmic space or cell wall association. Further, they demonstrated that CA activity was released into the medium by cell-wall deficient mutants or wild type cells exposed to wall-lysing enzymes. For these reasons, a wild type C. reinhardtii strain (CC-3269) and a cell wall-less mutant (CC-400 (cw15 mt+) were chosen for exploration. The investigative approach consisted of: (1) Growth of external or external CA producing strains; Final Report | MicroBio Engineering Inc. 24 (2) Analysis of the culture media and/or cell lysates for CA; and (3) Measurement of air-CO2 transfer rates of the media and/or lysate. Combining CA identification (2) and activity (3) measurements allow attribution of increases in air-CO2 transfer relative to non-biochemically enhanced abiotic controls to carbonic anhydrase facilitated air- CO2 transfer. Progress towards each assay component is summarized below, although was focused on the first two areas. 5.1.4.1 Methods 5.1.4.1.1 Cultivation Two strains of Chlamydomonas reinhardtii were utilized, a wild-type (CC-3269, WT) and a cell wall deficient mutant CC-400, CWD). Literature has shown that the mutant strain has the capacity to excrete carbonic anhydrase (CA) directly into the media more readily due to lack of a cell wall (Kimpel 1983). The wild type was used as a control during these trials. A modified Tris-Minimal recipe with urea as the N- source was used for both Chlamydomonas strains (Table 5.3). Additionally, media alkalinity was increased to 10 meq/L. Table 5-2: Modified Tris-minimal recipe with urea as the nitrogen source and alkalinity increase to 10 meq/L. Concentration in Final Concentration in Compound Trace Metals Medium, mM Final Medium, mM Urea 17 Boric acid 0.046 Manganese chloride Dipotassium phosphate 0.17 0.009 tetrahydrate Magnesium sulfate Zinc sulfate 0.30 0.0008 heptahydrate heptahydrate Calcium chloride, Copper (II) sulfate 0.24 0.0003 dihydrate pentahydrate Sodium Molybdate Citric acid 0.03 0.0016 Dihydrate Ferric ammonium Cobalt chloride 0.02 0.0002 citrate hexahydrate EDTA (dinatrium salt) 0.002 EDTA (disodium salt) 0.002686439 Sodium Bicarbonate 0.01 Sodium carbonate 0.19 The cultivation apparatus (Figure 5.10) consisted of a 122 cm (L) x 30 cm (W) x 30 cm (H) glass aquarium, filled to a depth of 12 cm with distilled water to maintain a constant temperature between reactors. Light was supplied continuously (24 hours/day). Water bath temperature was regulated to 22oC ± 3°C. Reactors consisted of a cylindrical glass jar with a 7.1 cm outer diameter, 16 cm height, 480 mL total volume and 300 mL working volume. Reactors were continuously mixed with a length 51 mm and diameter 8 mm octahedral stir bar at 160 RPM. Rubber stoppers capped the reactors, with ports for a pH probe, an inverted 1 mL serological pipette extending to just above the stir bar for air bubbling, and a vent. Air bubbling was controlled with rotameters (Dwyer RMA-12, 0.1 – 0.5 LPM) connected to an outside air manifold and regulated to 400 +/-20 mL/min. Syringe filters (0.2 µm pore size) were placed downstream of the rotameters to remove contaminating organisms. Final Report | MicroBio Engineering Inc. 25 Figure 5-9: Cultivation apparatus used for experiments, with one light panel removed for viewing. (1) – water bath, (2) – reactors, 3 – rotameters, (4) – mass flow meter, (5) - chiller recirculation outlet, (6) - chiller recirculation input. The fluorescent lighting array has been replaced by a higher intensity LED array. Optical density (OD) was used to track growth throughout a dilution to harvest cycle. VSS measurements were taken at every harvest to develop an OD:VSS correlation curve (Figure 5.11) and the correlation curves were used to calculate VSS based on OD readings during sampling. 350 y = 422.44x 300 R² = 0.9535 Biomass Concentration (mg VSS/L) 250 200 150 100 50 0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Optical Density (750 nm) Final Report | MicroBio Engineering Inc. 26 300 Biomass Concentration (mg VSS/L) 250 y = 303.22x 200 R² = 0.9644 150 100 50 0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Optical Density ( 750 nm) Figure 5-10: Chlamydomonas Wildtype (WT, Top) and Chlamydomonas cell wall deficient mutant (CWD , Bottom) biomass concentrations are graphed as a function of optical density to generate a correlation relationship to aide in daily biomass tracking. Reactors were removed from the cultivation apparatus for sampling (Figure 5.12). Evaporative losses were replenished with autoclaved DI water back to operating volume. Reactors were homogenized via vigorous magnetic stir-bar mixing 15 seconds. Once homogenous, about 3 ml were removed for pH and OD measurement. If handheld pH differed by more than 0.2 pH from the in-situ pH sensor value, in-situ pH sensors were recalibrated. To reduce risk of cross-contamination reactors were removed one media type at a time and one strain type at a time. After sampling, an equal volume of fresh media was added to maintain a consistent reactor volume. Reactors were diluted approximately weekly, replacing roughly 2/3’s of the reactor volume with fresh culture media. Figure 5-11: Chlamydamonas wild type (WT, Left) and Chlamydamonas mutant cell wall deficient (CWD, Right) reactors removed from the water bath (Figure 5.10) for pre-sampling. 5.1.4.1.2 Analysis of culture effluent 5.1.4.1.2.1 Dissolved proteins A SpectraMAX 340 96-well plate reader was used to measure the concentration of bovine CA (Novozymes) using the Pierce BCA assay (Thermo Scientific No. 23235). Following manufacturer guidelines, 150μL of each standard or unknown sample was pipetted into triplicate microplate wells. Final Report | MicroBio Engineering Inc. 27 Fifty microliters of the working reagent was added to each well and the plate was mixed thoroughly on a plate shaker for 30 seconds. The plate was covered using sealing tape for a 96-Well plates and incubated at 37°C for 2 hours. The plate was allowed to cool to room temperature (20° C - 25° C). The micro plate reader was set to read the absorbance at 562nm. The average readings of the blank standard were subtracted from the 562nm readings of all the individual standards and unknown sample triplicates. A standard curve was plotted with the average blank-corrected 562nm reading for each BSA standard vs. its concentration in μg/mL. The standard curve was used to determine the protein concentration of each unknown sample. 5.1.4.1.2.2 SDS-Page The goal of this assay is to concentrate proteins, including carbonic anhydrases, produced and released into the culture medium and to apply sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS- PAGE) to separate proteins based on molecular weight. Molecular weight comparison to bands of commercially available bovine CA (Novozymes) facilitates the identification of algal counterparts, and comparison to known CA concentrations allows the method to be semi-quantitative. The assays were applied in tandem to measure CA concentration of recycled algal culture media. Several trials were conducted to develop appropriate methods of whole cell and spent media preparation to release and/or concentrate proteins while preserving activity. For whole cell material, samples ranging between 0.5 – 5ml, depending on the cell density, of the three strains were pelleted by centrifugation and resuspended in appropriate loading buffer (see below). Proteins from spent media were concentrated 50-fold in a 10K MWCO Pierce Protein Concentrator (ThermoFisher Scientific) prior to PAGE analysis. Commercially available bovine carbonic anhydrase from Novozymes at a concentration of 1053 µg/ml was used at varying volumes to aide in validation of the SDS-PAGE and subsequent protonography methods. Initial PAGE trials were conducted under denaturing and reducing conditions to ensure sufficient protein separation for method validation. Commercially available CA was diluted and wet cell pellets were resuspended in a total of 80µl of 1X Bolt LDS Sample Buffer containing 1X Bolt Sample Reducing Agent per manufacturer’s recommendations (NW0412B, ThermoFisher Scientific). Samples were boiled for 2 minutes, allowed to cool and then centrifuged for 1 minute to removed residual cellular debris. Subsequent PAGE trials were conducted under denaturing but non-reducing conditions to preserve CA activity by omitting the reducing agent and boiling steps. Samples were loaded in the wells of the pre- made 4 to 12% Bis-Tris, 1.0 mm, protein gel cassettes and a PageRuler Plus pre-stained ladder (10 to 250 kDa) (NW0412C, ThermoFisher Scientific) was used for size comparison. The expected size of CA is approximately 30kD, although can vary depending on the type of CA or the organism from which it originated. Gels were submerged in 1X MES SDS running buffer and electrophoresis was performed at 200 volts for 22 minutes. Two different staining techniques were used in tandem. Proteins were Coomassie stained using Simply Blue Safestain in order to visualize all proteins separated on the gel. A replicate gel was run and subsequently subjected to protonography to detect the presence of CAs via a pH dependent color change following the methods of De Luca et al. (2015), described below. SimplyBlue SafeStain (#LC6060) is a premixed ready to use, propriety Coomassie G-250 stain from ThermoFisher Scientific that does not require the use of methanol or acetic acid, thus eliminating the need to dispose of hazardous waste. After electrophoresis concluded, the gel was rinsed 3 times for 5 minutes with 100 mL of distilled water to remove SDS and buffer salts, which interfere with binding of the dye to the protein. Then about 20 mL of SimplyBlue SafeStain was used to cover the mini gel, then incubated for 1 hour at room temperature with gentle shaking. One final wash with about 100 ml of distilled water was completed for Final Report | MicroBio Engineering Inc. 28 1 – 3 hours (dependent on band development) with gentle shaking. After the wash, the gel was imaged under a fluorescent light to observe any bands that may have developed. 5.1.4.1.2.3 Protonography Upon completion of electrophoresis, the gels were soaked in 2.5% Triton X-100 for 1 hr under gentle shaking, then washed twice in 100mM Tris, pH 8.2 containing 10% isopropanol for 10 minutes under gentle shaking. After the wash, the gel was incubated in 0.1% bromothymol blue in 100mM Tris, pH 8.2 for 30 min. Finally, the gel was immersed in CO2-saturated distilled water to visualize the hydratase activity of the enzyme. CO2-saturated water was sourced from either over-the-counter sparkling water (pH 4.8) or by bubbling distilled water with gaseous CO2 for several hours, reaching a pH near 5.0. The local decrease in pH due to the presence of CA activity was visualized by the formation of yellow bands due to the change of the indicator color from blue (alkaline pH) to yellow (acidic pH). Figure 6.1.3: Staining of SDS-PAGE with 0.1% bromothymol blue in 100mM Tris (Left). Submersion of stained SDS-PAGE with sparkling water(right). 5.1.4.1.3 Western Blot Methods Development To advance CA estimation during cultivation under various conditions, a western blotting approach was developed to detect inducible, cell-surface carbonic anhydrase. Six carbonic anhydrases have been previously identified in Chlamydomonas reinhardtii, including three alpha-CAs and three beta-CAs and the range of 20–30 kDa (SDS-PAGE); furthermore, the periplasmic alpha-CA, CAH1 converts HCO3- to CO2 and facilitates entry of carbon dioxide into the algal cell (Ynalvez et al. 2008, De Luca et al. 2015). CAH1 is the major periplasmic CA and its transcription is highly induced under low-CO2 condition (Fujiwara et al. 1990). Previous publications have shown essentially no CA were present in 5% CO2-grown (CO2 rich) cultures but after a 5-h induction period (air-only CO2), the level of activity was 50% of the maximal activity reached after 24 h of exposure to air (Coleman et al 1984). Several lines of evidence helped Coleman and colleagues establish the identity of the extracellular form of the C. reinhardtii carbonic anhydrase. The monomeric species for this enzyme is a diffuse band at 37,000 D which appears following the transfer of high CO2 grown cells to air levels of CO2. Most of the CA activity was excreted into the culture medium by the cell wall-less mutant (Coleman et al 1984). It is the major component of a nondenatured protein which binds dansylamide, a specific inhibitor of CA (Coleman et al 1984). And finally, antibodies prepared against the 37,000-D polypeptide inhibit CA activity, and although two species sometimes can be discerned from this diffuse band on a gel, the antibody reacts with only one antigenic component (Coleman et al. 1984). Final Report | MicroBio Engineering Inc. 29 Algal culture samples were centrifuged and the supernatant removed to be heated (80o C) and denatured for 2 minutes. A Bolt™ Bis-Tris Plus Mini Protein Gel, 4‐12%, 1.0 mm gel was used with NuPAGE™ Sample Reducing Agent (NP0004) to separate the protein (22 minutes at 200V) in MES running buffer. Transfer from gel to PVDF membrane (LC2002) was conducted in a Mini-Blot module (B1000) for 1 hour at 20V in Bolt™ Transfer Buffer (BT0006). The membrane was incubated in blocking solution (Thermo Scientific SuperBlock™ Blocking Buffer (37516)) at room temperature for 1 hour. The primary antibody consisted of rabbit antiserum against (CAH1) diluted 1:500 in SuperBlock™. The antiserum was kindly provided by Dr. David Husic (Husic et al. 1989). The membrane was washed 3 times in 15 minute intervals with Tween 20 blocking buffer (AAJ61544K2) and the secondary antibody was a Goat anti-Rabbit IgG (H+L) conjugated to Horse Radish Peroxidase (656120) was diluted to 1:3,000 in SuperBlock™ and incubated for 1 hour at room temperature on a shaker table. A second round of washing was conducted, 3 changes 15 min on a shaker table. HRP was detected colorimetrically using 1- Step™ Ultra TMB-Blotting Solution (PI37574). 5.1.4.2 Results 5.1.4.2.1 Cultivation Chlamydomonas wild type (WT) and cell wall deficient mutant (CWD) cultures were maintained as a source of culture effluent for CA analysis. Time-course biomass productivity is reported in Figure 5-12, and also summarized by month (Figure 5-13). Culture pH was slowly increased of this time, an an effort to induce CA production, although cultures exhibited instability at a pH setpoint of 9.0. Cultures were mostly stable at a pH setpoint of 8.5, slightly below the estimated air-equilibrium pH (9.0) for this culture media amended to 10 meq/L alkalinity. We hypothesized that if the wall-less mutant were successful in producing, and releasing CA into the culture media, higher biomass productivity may result. However, at pH 8.5, it is unlikely there was any driving force for air-CO2 mass transfer. The wildtype strain (CC-3269) was moderately more productive than the wall-less mutant (CC-400). Effluent samples were retained for analysis. 800 700 Slope Productivity (mg VSS/L-Day) 600 500 400 300 200 100 0 3/25 4/4 4/14 4/24 5/4 5/14 5/24 6/3 6/13 6/23 7/3 7/13 -100 -200 -300 CC-400 (Mutant) - Rep 1 - (Position 9) CC-400 (Mutant ) - Rep 2 - (Position 10) CC-400 (Mutant) - Rep 1 - (Postion 1) CC-400 (Mutant) - Rep 2 - (Postion 2) Final Report | MicroBio Engineering Inc. 30 1400 Slope Productivity (mg VSS/L-Day) 1200 1000 800 600 400 200 0 3/15 4/4 4/24 5/14 6/3 6/23 7/13 -200 CC-3269 (Wildtype) - Rep 1 - (Position7) CC-3269 (Wildtype) - Rep 2 - (Postion 8) CC-3269 (Wildtype) - Rep 1 - (Postion 5) CC-3269 (Wildtype) - Rep 2 - (Postion 6) Figure 5-12: Slope productivity (mg VSS/L-day) as function of time for Chlamydamonas CWD mutant (Top) and Chlamydamonas wildtype (Bottom). 600 500 Average Slope Productivity (mg VSS/L-day) 400 300 200 100 0 April Average May Average June Average Overall Average Wildtype Mutant Figure 5-13: Average slope productivity of Chlamydamonas reinhardtii wildtype and mutant categorized by monthly trials and overall. April wildtype 237.1 mg VSS/L-day (SD = 88.8, n = 16) and mutant 145.7 mg VSS/L-day (SD = 58., 3 n=20). May wildtype 372.3 mg VSS/L-day (SD = 123.1, n =13 ) and mutant 226.6 mg VSS/L-day (SD = 90.4., 3 n=16). June wildtype 205.6 mg VSS/L-day (SD = 101.1, n =14 ) and mutant 148.6 mg VSS/L-day (SD = 29.3., 3 n=13). Overall average for wild type of 271.6 mg VSS/L-day (SD = 88.5, n=43) and mutant 173.6 VSS/L- day (SD = 45.9, n=49 ). Final Report | MicroBio Engineering Inc. 31 5.1.4.2.2 Dissolved proteins Dissolved protein levels in effluent media were monitored as a broad, readily-measured surrogate for the relative levels of carbonic anhydrase, and were quantified against a Bovine Serum Albumin (BSA) standard curve (Figure 5.16). Dissolved protein concentration was measured in both Chlamydomonas wildtype (CC-3269) and mutant cell wall deficient (CC-400), increasing throughout batch cultivation over four days of cultivation. The cell wall deficient mutant (CC-400) had an average protein concentration increase of about 2.2-fold and the wildtype (CC-3269) had about a 2.7-fold increase (Figure 5.17). Specific rates of dissolved protein formation (µg BSA protein equivalent/mg volatile suspended solids) were similar between the two strains, indicating the wall-less mutant, under the cultivation conditions studied, did not ‘leak’ a significantly greater relative percentage of proteins into the culture media. Even so, the assay confirmed a relative increase in organic carbon, external to biomass as suspended solids, of which carbonic anhydrase may be a partial component. Subsequent exploration aimed to characterize such biomass exudates. 250 200 Final BSA Concentration (μg/mL) 150 y = 10.153x2 + 48.777x - 4.1556 R² = 0.9996 100 50 0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 Abs562nm Figure 5-14: Standard Curve of diluted Bovine Serum Albumin (BSA) with measure absorbance at 562 nm; a best- fit polynomial equation was used (R² = 0.9996). Final Report | MicroBio Engineering Inc. 32 80 70 67.1 Protein Concnetration (μg/mL) 60 50 44.4 40 36.7 30 18.0 16.3 20 11.7 11.2 12.7 10.1 10.4 10 0 Cycle 1 Cycle 2 Cycle 3 Cycle 4 Cycle 5 Day 0 Day4 60 49.1 Protein Concnetration (μg/mL) 50 42.8 43.0 40 33.1 30 21.2 20 15.3 11.8 8.5 9.1 10 6.7 0 Cycle 1 Cycle 2 Cycle 3 Cycle 4 Cycle 5 Day 0 Day4 Figure 5-15: Protein concentration of Chlamydamonas reinhardtii (μg/mL) Mutant (top) and wildtype (bottom) during 4 day growth periods calculated with BSA standard curve. 5.1.4.2.3 Protonography, SDS-PAGE Culture effluents were further examined for carbonic anhydrase activity via protonography, and the relative distribution of molecular weights and relative concentration by SDS-PAGE. An initial gel was run under denaturing and reducing conditions and Coomassie stained in order to validate PAGE techniques. Commercially available Novozyme CA bands can be observed in the expected molecular weight size range of ~30 kDa; lanes 2 – 7 (Figure 6.2.4). Band intensity was directly proportional to the amount of enzyme loaded into each well for all the Novozyme samples. The secondary goal was to observe whole cell material on the gel. Lanes 9 – 11 contained whole cell material and appeared smeared due to being overloaded, guiding future campaigns. Final Report | MicroBio Engineering Inc. 33 Figure 6.1.4: PAGE run of Novozymes CA at varying concentrations along with whole cell visualized with Coomassie stain. A second run was conducted under denaturing but non-reducing conditions to test the protonography protocol. A replicate gel was stained with Coomassie for comparison. Bands were not as clear as the first gel due to the non-reducing conditions and the fact that the samples were not boiled prior to loading in order to preserve enzyme activity. Although difficult to image effectively, faint yellow bands were visible in lanes 10 – 12 (Figure 6.2.5) indicating a localized pH shift suggesting that the Novozyme CA was present and active. Lanes 8 and 9 may not have had high enough concentration of CAs to trigger the color change. De Luca et al. (2015) observed CA activity when 30µg of protein was loaded in each well, which is in agreement with our results that the higher concentrations in our trial improved signal to background contrast, facilitating visualization. It should be noted that it is challenging to capture the localized color change since the reaction reaches completion in under 1 minute. Longer incubations cause in a pH shift throughout the entire gel resulting in a homogenous yellow color. Figure 6.1.5: PAGE run of Novozymes CA at varying concentrations visualized with Coomassie stain protocol (left) and protonography protocol (right). A trial was conducted with the goal of detecting differences in whole cell material between air sparged and non-air sparged cultures. Again, too much whole cell material was loaded as can be seen by the smearing of bands; the concentration of the diatoms in lanes 3 and 9 (Figure 6.1.6) seemed especially high. In the protonography gel, there was a reactive band between 25 – 35 kDa visible in lane 2, which contained the Novozyme CA used as a positive control, validating the protonography method. Some reactive bands between 25 – 35 kDa were also visible in lanes 5 and 6 containing C. reinhardtii wild type samples (Figure 6.1.6). The amount of whole cell material sampled from the Chlamydomonas culture appears to be with in a good range, based on Coomassie stain, facilitating future trials. The amount of whole cell material used of diatom culture needs to be further optimized. Final Report | MicroBio Engineering Inc. 34 Figure 6.1.6: PAGE run of whole cell material that were separated into a air sparged group and a non-air sparged group, visualized with Coomassie stain protocol (right) and protonography protocol (left). Novozymes CA to act as a positive control. A subsequent trial analyzed culture effluent media for the presence of CA activity. Strains were cultivated in both air-sparged and non-air-sparged conditions for 48 hours. Samples from the non-air- sparged condition were centrifuged to separate cells from spent media. The resulting supernatant was concentrated 50-fold with a Pierce Protein concentrator and samples were run under denaturing but non-reducing conditions. In this trial, the CO2 enriched water used for protonography consisted of ~200 ml of distilled water sparged for about 3 hours. While the positive control consisting of Novozyme CA is visible with coomassie stain in lane 2, protonography results were not as easily visualized as in previous trials (Figures 6.1.6 and 6.1.7). Perhaps a combination of not enough material in the spent media lanes and the CO2 enriched water not being at a high enough concentration contributed to the lack of bands and color change. In future experiments, spent media will be further concentrated and we will return to using carbonated seltzer water for the protonography analysis. Figure 6.1.7: Spent media concentrated 50-fold and subjected to PAGE analysis. Samples consist of an air- sparged group and a non-air-sparged group visualized with Coomassie stain protocol (left) and protonography protocol (right). Novozymes CA to act as a positive control. 5.1.4.2.4 Western blot Transfer and detection methods require further optimization, as no bands were visible in the tested samples. A more sensitive detection reagent may be investigated. Further, a positive control is needed to validate the western blot methods, specifically, a polyclonal antibody against bovine CA, both of which are commercially available. 5.1.4.3 Conclusions Despite evidence from the oceanographic community (Matthews 1999) suggesting that dissolved carbonic anhydrase can increase the rate of air-CO2 mass-transfer, trials conducted herein were inconclusive. Chlamydomonas walled (CC-3269) and wall-less (CC-400) strains were used as a model Final Report | MicroBio Engineering Inc. 35 organism, based on previous characterization of inducible external CA formation. Both strains struggled to increase culture media pH to levels above air-saturation, where air-CO2 ingassing would be observed, at least in the semi-continuous regime tested here. Higher pH was reached in ad-hoc batch growth, although could not be sustained via dilution. Dissolved protein levels increased throughout the cultivation period, reaching on the order of 50 µg/mL BSA-equivalent, and bands within the expected molecular weight range (25 – 35 kDa) were observed in some culture effluents. Thus, while CA was likely detected in culture media effluent herein, we did not establish whether this impacted the rate of air-CO2 ingassing, and any changes to biomass productivity which may result. 5.2. Design, Construction and Operation of a Direct Air Capture (DAC) System 5.2.1. Design and Construction The project included the design and construction of a containerized DAC unit by the subcontractor Global Thermostat Operations LLC, and was originally intended to be shipped to a project partner’s Cyanotech site for onsite CO2 generation. However, due to logistical and other limitations it was decided to carry out a virtual DAC operation, as described below. This subtask included design of the adsorption and regeneration assemblies, process equipment, piping, electrical panels, equipment stands, process skid and container, and process controls. Figure 5-16-Figure 5-18 provide some renderings of the final assembly from the design phase: Figure 5-16. Rendering of the DAC pilot unit – walls hidden. Final Report | MicroBio Engineering Inc. 36 Figure 5-17. Close up of downstream equipment. Figure 5-18. DAC pilot unit container with roll up doors. Detailed drawing packages were created to complete the skid design and allow us to generate bids for fabrication of the unit. Springs Fab was the vendor selected for fabricating the adsorption and regeneration assemblies, equipment stands, process skid and container. Piping was performed both by Springs Fab and our own staff for different components, electrical wiring was performed by a subcontractor, and programming for the process control system was completed in-house by our staff. Figure 5-19 provides photos of the completed DAC pilot unit installed in Brighton, Colorado. Final Report | MicroBio Engineering Inc. 37 Figure 5-19. Photos of the exterior of the DAC pilot unit during the commissioning phase. Before initializing the startup processes, we conducted a comprehensive pre-startup safety review. This included verifying the functionality of all safety systems to prevent any potential mishaps during operation as well as any deficiencies that could lead to operational hazards which were addressed. Subsequently, we were able to move forward with commissioning the pilot unit. This involved a cold commissioning phase (without process fluids like steam and no contactors), followed by a hot commissioning phase (including steam but still no contactors) to ensure proper operation of the system. Once commissioning was complete, we were able to install contactors and operate the unit to produce CO2. 5.2.2. DAC Operations The DAC trial began on Jul. 8, 2024 and reached 2000 h of operation on Oct. 12, 2024. This equates to a 97-day trial, with 86% uptime. The major causes for downtime could have been ameliorated with better preventative maintenance and having critical spares on hand. In addition, the fact that the skid was not staffed 24/7 meant that if something happened overnight or during the weekend there was additional down time before operations staff were available to bring the unit back online. Relative humidity and ambient temperature over the course of the trial are reported in Figure 5-20. Initial tests provided insight on the process kinetics when running at pseudo equilibrium adsorption and desorption times which were not intended to yield maximum CO2 production. Equilibrium tests were run through Aug. 21, 2024. Subsequent to that, we cycled through eight (8) different adsorption (ads) and desorption (des) operating conditions to determine the optimal process conditions for maximizing hourly CO2 production. The three best process conditions evaluated were: • Ads 1 / Des 1 • Ads 1/ Des 2 • Ads 2 / Des 2 • NOTE: Six (6) other ads/des combinations were tested and yielded poor CO2 production (Ads 3 / Des 3; Ads 4 / Des 1; Ads 5 / Des 1; Ads 5 / Des 2; Ads 6 / Des 2; Ads 2 / Des 1) The optimal cycle to maximize hourly CO2 production was Ads 1/Des 1. Final Report | MicroBio Engineering Inc. 38 Figure 5-20. Relative humidity (orange line) and ambient temperature (°C, blue line) for duration of DAC pilot unit trial. Daily CO2 production for the duration of the DAC pilot unit campaign is provided in Figure 5-21, with the different marker and line colors corresponding to the various adsorption and desorption times used in the different process conditions evaluated. The dashed lines indicate when contactors were switched out in the campaign. Overall, process conditions with shorter duration adsorption and desorption times led to higher daily production amounts. The first two sets of contactors were washed out from extended steam sweep desorption durations while the third set of contactors inserted had been partially used in a different test rig, and then an issue with the boiler on October 1st resulted in wet steam exposure that soaked/washed out the panel. The final set of contactors inserted appear as though they may be losing productivity from the figure, however on the final day the unit did not operate the full 24 hours (~20 hours), and on the previous day it can be seen in Figure 5-20 that there was higher RH and ambient temperature that day which can lead to reductions in productivity. Product purity data for the DAC pilot unit campaign is provided in Figure 5-21 from Aug. 22, 2024 through the end of the trial. Similar to the circumstances for plant run time, this is when a tag was added to pull product purity from our database into our data visualization tool. There are two places where CO2 product purity can be read, at the vacuum pump (VP) discharge and from the high-pressure storage tank. While normally we would monitor this output from the high-pressure (HP) tank, there is a leak on the high-pressure compressor that draws air into the system; thereby diluting the CO2 in the high-pressure tank. In order to avoid additional capital costs, the high-pressure compressor has not been replaced and the reported CO2 purity is taken at the discharge of vacuum pumps which represents the expected purity if the high-pressure compressor was not leaking. Note that there are spikes in the data due to the sampling port switching back and forth taking readings from the VP discharge and HP tank, and so the product purity should be viewed as the top trace on the noise. As can be seen in the figure in reference to the daily CO2 production, the product purity starts to decline when the contactors start to degrade, as the dips in purity align with loss in productivity that led to contactor swap out. Additionally, due to the small scale of the DAC pilot unit (producing ~tons of CO2 per year), the system is not designed to reach higher product purities. However, we have demonstrated on our K-Series plant that product purity above 90% is achievable, as shown in Figure 5-22, which were used in the accompanying algal growth trials. Final Report | MicroBio Engineering Inc. 39 Figure 5-21. CO2 product purity as measured from the vacuum pump discharge from 8/22/24 through the end of the DAC pilot unit trial. Figure 5-22. Average CO2 product purity for K-Series plant running over multiple months at >90%. 5.2.3. Techno-Economic Analysis Scale-up of DAC technology enables lower cost of capture. The demonstration unit constructed for this project is capable of producing on the order of tonnes of CO2 per year which yields capture costs on the order of thousands of dollars per tonne. Global Thermostat has built and operated a kilotonne scale DAC unit (K-Series) at the Brighton, Colorado facility demonstrating that this DAC technology can be successfully scaled. The total CO2 burden for the K-Series (~2,000 tonne/yr)is shown in Table 5-3 for a variety of power and heat source combinations. Final Report | MicroBio Engineering Inc. 40 Table 5-3. Carbon burden and corresponding cost of capture per net tonne for a variety of power and heat source combinations. The source of different assumptions to LCA inputs are provided below the table. Total Burden (t CO2e / t CO2) Power Source Heat Source K-Series US Grid Mix Natural Gas Boiler 1.01 US Grid Mix E-Boiler 1.35 US Grid Mix Heat Pump 0.99 US Grid Mix Waste Heat 0.65 Global Average 2030 Natural Gas Boiler 0.85 Global Average 2030 E-Boiler 1.00 Global Average 2030 Heat Pump 0.73 Global Average 2030 Waste Heat 0.48 Global Average 2050 Natural Gas Boiler 0.52 Global Average 2050 E-Boiler 0.31 Global Average 2050 Heat Pump 0.23 Global Average 2050 Waste Heat 0.16 Wind Natural Gas Boiler 0.40 Wind E-Boiler 0.06 Wind Heat Pump 0.05 Wind Waste Heat 0.04 PV Natural Gas Boiler 0.53 PV E-Boiler 0.31 PV Heat Pump 0.23 PV Waste Heat 0.16 Notes: GaBi. GaBi 9.2.0.58. Software-System and Database for Life Cycle Engineering, DB 8.7 - Wind 10 g/kWh SP 39 (thinkstep AG,Leinfelden-Echterdingen, Germany, 2019). GaBi. GaBi 9.2.0.58. Software-System and Database for Life Cycle Engineering, DB 8.7 - PV 83 g/kWh SP 39 (thinkstep AG,Leinfelden-Echterdingen, Germany, 2019). US grid mix 376 g/kWh eia GaBi. GaBi 9.2.0.58. Software-System and Database for Life Cycle Engineering, DB 8.7 - Global Average 2030 277 g/kWh SP 39 (thinkstep AG,Leinfelden-Echterdingen, Germany, 2019). GaBi. GaBi 9.2.0.58. Software-System and Database for Life Cycle Engineering, DB 8.7 - Global Average 2050 81 g/kWh SP 39 (thinkstep AG,Leinfelden-Echterdingen, Germany, 2019). Waste Heat 0 g/kWh Assumed burdened free Natural Gas Boiler 201 g/kWh eia, 90% eff E-Boiler 0 g/kWh Dependent on Power Source Heat Pump 0 g/kWh Dependent on Power Source https://www.iea.org/data-and-statistics/charts/direct-co2-intensity-of-the-iron-and- Steel 1.3 kg CO2 / kg steel-sector-in-the-net-zero-scenario-2010-2030 Concrete 0.9 kg CO2 / kg Standard (not low carbon) concrete Sorbent 11 kg CO2 / kg https://www.nature.com/articles/s41560-020-00771-9#Sec23 Substrate 0.25 kg CO2 / kg https://www.sciencedirect.com/science/article/pii/S2666790821003190 Final Report | MicroBio Engineering Inc. 41 5.3. DAC cultivation Trial With the transition to a virtual direct air capture (DAC) demonstration, this task compared microalgal growth under simulated DAC-carbon delivery to that of baseline, on-demand based carbon supply. These cultivation trials took place at MBE’s ‘Azul Farms’ location in Paso Robles, CA. As described above in Task 3, Global Thermostat initiated their 2,000 hr DAC pilot trial on 7/8/2024, subsequently providing a representative 7-day CO2 production time course as the basis for simulation in the MBE cultivation trials. This CO2 production profile was averaged into a representative 24-hr day of DAC-CO2 supply, then scaled to match the predicted CO2 requirement in the MBE cultivation systems, described in more detail below. The ultimate goal of this task and project was to identify implications of cultivation with continuous CO2 supply via DAC, taking advantage of the culture media carbonate buffering system to store CO2 during the night, reducing DAC system capacity approximately 3-fold relative to if sized based on peak summer algal carbon demand. The consequence of such CO2 supply is a diel change in culture pH, as described previously. The section below describes the methodology and initial results of the simulated-DAC cultivation trial. The experimental setup consisted of cultivating Porphyridium cruentum in two raceways, one with a surface area of 3.4 m2 and the other with a surface area of 22 m2. For the DAC trial, the 22 m2 raceway is supplied with continuous flow of CO2 into the raceway to simulate consistent CO2 feed from a DAC unit. The 3.4 m2 raceway, on the other hand, is only dosed with CO2 to maintain a pH near 9.0, which serves as the control raceway. Both ponds undergo regular biomass harvests. Previously, this project focused on production of a Chlorella-like microalga, a commercial interest of Cyanotech. With a lack of previous successful cultivation history sufficient to justify a 3-month (2,000 hr) cultivation trial at moderate scale, MBE instead cultivated an alkaliphilic red microalga, Porphyridium sp., with potential commercial applications as a red food coloring, hydrocolloid in either food or cosmetics, and/or multi-functional alternative protein. Porphyridium cruentum was identified as a top summer strain candidate during the indoor DISCOVR trials – an ongoing, multi-year, multi-national lab effort to decrease algal biomass production cost (Huesemann et al. 2024). Temperature and salinity tolerance testing under CO2-replete conditions identified optimum growth between 23-33 °C, characterizing it as a warm season strain, with a relatively wide salinity tolerance, spanning brackish (5 ppt) to slightly hypersaline (40 ppt) conditions. P. cruentum growth under carbon deplete conditions was characterized in the joint MBE-PNNL Air Carbon for Algae Production (AirCAP) DOE-BETO project, again emerging as top candidate after two-tiers of indoor screening. Summer biomass productivity in CO2-replete ponds (pH 9.5, 80 meq/L) averaged between 15 – 20 g AFDW/m2-day. Additionally, MBE had previously successfully scaled P. cruentum from 3.4 m2 to 22 m2 raceways, with stable cultivation over an 83-day period. Thus, sufficient cultivation history and experience, as well as commercial potential, justified transition to this microalga. 5.3.1. Methods 5.3.1.1 Culture Media Cultivation trials were carried out in PNNL 35/80 media (Table 5-4), based on earlier optimization trials at PNNL, which measured highest biomass productivity for P. cruentum at this salinity (35 ppt) and alkalinity (80 meq/L). Nutrient concentrations were adjusted based on anticipated biomass carrying capacity during either indoor or outdoor cultivation. Carrying capacity is an estimate of the maximum biomass density a given media recipe could support, calculated based on literature general alga elemental composition and corresponding medial elemental concentration. The purpose is to ensure that laboratory cultures, which tend to reach higher biomass concentration than in outdoor cultures, remain nutrient replete. Conversely, this avoids wasting nutrients in outdoor raceway pond cultures, Final Report | MicroBio Engineering Inc. 42 which generally transition into stationary phase based on light limitation rather than macro-nutrient levels and cultivated with a lower initial nutrient load. Indoor culture media was prepared to 2,000 mg VSS/L (Volatile Suspended Solids) carrying capacity using an Azul water base (water collected from Azul Farms, the outdoor raceway location), while outdoor raceway pond media was prepared to a biomass carrying capacity of 700 mg VSS/L. Outdoor reactor cultures were adjusted to account for minerals present in the Azul well water. Modification included reduction of supplemental magnesium by half (20 to 10 mg Mg/L) to account for the 10 mg/L already present in Azul well water, as well as the complete omission of Ca; Azul water already contained 130% of the recipe requirement. Table 5-4: Elemental composition of 35/80 media, Azul well water, and the modified recipe for initial Azul operations. Biomass content is for generic algae and not P. cruentum specifically. Azul 35/80 Biomass Components Azul 35/80 Carrying Cap (mg Element Units Well Media, content (g/g added for Azul Concentration VSS/L) Water Original VSS) Na mg/L 130 12581 12451 12581 ‐ ‐ K mg/L 2.4 775 775 777 0.025 31082 Mg mg/L 10 19.7 9.7 19.7 0.005 3944 Ca mg/L 17 10.9 0.0 17.0 0.008 2125 Cl mg/L 32 16680 16648 16680 ‐ ‐ S mg/L 34 211 177 211 0.016 13198 N mg/L ‐ 70 70 70.1 0.085 824 P mg/L ‐ 9.6 9.6 9.6 0.012 800 C mg/L ‐ ‐ ‐ ‐ ‐ ‐ Fe mg/L ND 2.0 2.0 2.0 0.000160 13 Cu µg/L ND 0.001 0.001 0.001 9.20E‐06 0 Mo µg/L ND 0.396 0.396 0.396 1.21E‐06 328 Zn µg/L ND 0.227 0.227 0.227 2.00E‐05 11 Co µg/L ‐ 0.244 0.244 0.244 2.03E‐06 120 Mn µg/L ND 0.397 0.397 0.397 7.96E‐05 5 Sr µg/L 260 ‐ ‐ 260 1.67E‐04 1557 Cd µg/L ND ‐ ‐ ‐ 8.97E‐06 ‐ B µg/L 350 1.751 1.75 352 ‐ ‐ Throughout the outdoor cultivation trial, media was recycled (rather than preparing fresh media) for each growth period. Following a harvest, alkalinity and salinity levels were returned to their original 35/80 concentrations. Nutrients, trace metal, and vitamin additions, however, were not as straightforward and on 9/24, algae color in the raceway transitioned from P. cruentum’s characteristic red hue and became browner and paler in color, suggesting a nutrient deficiency. Up until this point, nutrients were added based on elemental uptake during the prior growth period and 1 and 0.5 L of trace metals and vitamin solutions, respectively, were added to the raceways. This event inspired a literature review of nutrient, trace metal, and vitamin concentrations in P. cruentum and to compare literature values to initial concentrations presented for generic alga (Table 5-5). For the three papers presented analyzing P. cruentum, ash content was consistently measured at approximately 20%, therefore allowing direct comparison across elemental concentrations in the biomass. Based on this, updated concentrations were established for phosphorus, nitrogen, sulfur, manganese, and copper using values reported in the literature for P. cruentum. Boron, cobalt, and molybdenum concentrations were Final Report | MicroBio Engineering Inc. 43 maintained at their original values due to a lack of literature data. Vitamin concentrations were averaged from values reported by Brown (1999) across multiple algae species. Starting 10/1/2024, nutrients were added after each harvest based on the increase in TSS (Total Suspended Solids) mass during the growth period and the corresponding element uptake based on the updated concentrations. VSS was used initially, however, TSS provides a higher estimate of elemental uptake, which serves as a precaution to prevent future nutrient depletion. Additionally, new trace metal and vitamin solutions were prepared using the updated concentrations so that 1 and 0.5 L of each, respectively, would be added following a harvest. The new solutions had concentrations significantly higher than the original solutions, and precipitation was observed in the trace metal solution. A small experiment to test the effectiveness of these new solutions was performed using indoor culture flasks inoculated with P. cruentum. No noticeable impacts were observed in the flasks, however, overall hesitancy of the effectiveness of the new solutions led to the decision to continue using the original solutions rather than the new solutions. Therefore, for the rest of the cultivation trial, 1 and 0.5 L of the original trace metal and vitamin solutions were added, respectively, following a harvest. Table 5-5: Literature review of elemental concentrations and their mg/g VSS biomass concentration. Sources: Fithriani (2022), Fuentes (2000), and Di Lena (2020; P. cruentum; Cerri (2021) and Kavitha (2016) analyzed P. purpureum; Brown (1999) are average values for a variety of marine microalgae used in aquaculture. Before Fithriani Fuentes Di Lena Kavitha Cerri Brown Element After 10/1 9/24 (2022) (2000) (2020) (2016) (2021) (1999) P 12 2.2 ‐ 3.167 ‐ 13.9 ‐ 3 N 85 ‐ 48 40.9 47.84 ‐ ‐ 45 S 16 ‐ 15.3 ‐ ‐ ‐ ‐ 16 Mg 5 14 5.92 4.737 7.21 13.1 ‐ 5 Fe 0.16 0.0382 4.22 0.377 1.102 0.595 ‐ 0.7 Zn 0.02 0.00345 2.83 0.115 4.27 0.056 ‐ 2 Mn 0.008 ‐ 0.37 ‐ ‐ 0.021 ‐ 0.2 B 0.01 ‐ ‐ ‐ ‐ ‐ ‐ 0.01 Co 0.02 ‐ ‐ ‐ ‐ ‐ ‐ 0.02 Mo 0.0012 ‐ ‐ ‐ ‐ ‐ ‐ 0.0012 Cu 0.0092 ‐ ‐ 0.012 0.0985 0.059 ‐ 0.06 Vitamin B1 ‐ ‐ ‐ ‐ ‐ ‐ 0.059 0.059 Vitamin H ‐ ‐ ‐ ‐ ‐ ‐ 0.00132 0.00132 Vitamin B12 ‐ ‐ ‐ ‐ ‐ ‐ 0.00166 0.00166 5.3.1.2 Pond Inoculation Prior to raceway inoculation, raceways were scrubbed to remove any leftover biomass or debris. The raceways were then filled with Azul well water to near their maximum operational depth (c.a. 30 cm) and growth media components were added. The raceways were then disinfected with 1 mL of 12% hypochlorite (commercial bleach) per L of media. After at least one hour of contact time, the hypochlorite was neutralized using a 784 g/L sodium thiosulfate pentahydrate solution, confirmed using a Hach Free % Total Chlorine Test Strip. Raceways were then drained to their desired depths and inoculated. In the event of a pond crash requiring re-inoculation, the same procedure was performed. Final Report | MicroBio Engineering Inc. 44 5.3.1.3 Pond Operation Two 3.4 m2 (labeled as 3.4C and 3.4D) and one 22 m2 raceway (Figure 5-23), were operated under carbon replete conditions. 3.4C and 3.4D were supplemented with CO2 to maintain a pH of 9.0 0 ± 0.1 whereas the 22 m2 raceway had a consistent supply of CO2 as part of the 2,000-hour DAC trial. CO2 flowrate to the 22 m2 raceway was controlled with a mass flow controller (Omega Mass Flow Controller FMA5512A). On 9/24/2024, the CO2 flow rate regulation in 3.4C was switched from a rotameter to a mass flow controller, matching the one used in the 22 m2 raceway. CO2 was injected via a bubble diffuser (Aquacave 1.25” dia. X 2.75” Long Cylinder Air Stone) anchored to the raceway floor. pH (Sensorex S200CD/10/BNC – Process pH Sensor, Light Duty) and temperature (Neptune Temperature Probe) probes recorded in-situ measurements every ten minutes via a Neptune monitoring system. If the measured pH in 3.4C/D exceeded 9.1, a solenoid valve automatically opened, supplying CO2 into the ponds until the pH reached 8.9, when the solenoid valve closes. The 22 m2 and 3.4C raceways were maintained at an operational depth of 25 cm (786 L and 5,500 L in the 3.4m2 and 22 m2 raceways, respectively). Raceway 3.4D’s depth and biomass concentration on the other hand, was not maintained and allowed to grow without intervention as a backup culture in the event of a pond crash. Paddle wheel speed was set to 7 rpm (VFD = 25 Hz) and 8.4 rpm (VFD = 35 Hz) in the 3.4 m2 and 22 m2 ponds, respectively. A foam catch (window screen) was hung downstream of the paddle wheel to prevent light attenuation. Raceway samples were collected approximately three times a week (before and after a harvest, and once during the growth period between harvests). The sampling procedure entailed turning off the paddlewheels and allowing the water level to be steady prior to measuring pond depth, then evaporative losses were replenished with Azul well water. Paddles are then turned back on, and a sample was collected. Discrete salinity (YSI ProSOLO Optical Dissolved Oxygen meter), pH, and water temperature (Oakton pH 150) were measured via handheld meters. The handheld pH measurement was compared to the in-situ Neptune pH measurement, and if the discrepancy was greater than 0.2 +/- units, the raceways probes were recalibrated in pH 7 and 10 buffers or replaced. Prior to the DAC-simulation trial, the 3.4C and 22 m2 raceways grew in batch until reaching a biomass concentration of approximately 500 mg VSS/L at which point sufficient volume was removed and replaced with new culture media to initiate the next growth cycle. A post-dilution concentration of 100 VSS/L was targeted. 3.4D did not undergo these dilutions unless it was used to inoculate another pond. Figure 5-23: From left to right, 3.4C, 3.4D, and the 22 m2 raceways inoculated with P. cruentum. Final Report | MicroBio Engineering Inc. 45 5.3.1.4 Direct Air Capture (DAC) Trial The 2,000-hour DAC trial in the 22 m2 raceway began on 9/3/2024. To determine CO2 flowrate, the expected algal productivity and its corresponding carbon demand was used to back calculate the amount of CO2 needed to be provided to the system and the corresponding continuous flowrate required throughout the day. Leading up to the start of the DAC trial, outdoor culture productivity has been upwards of 20 g/m2-day, therefore 290 mL CO2/min is the required continuous flowrate to support this productivity. On 10/25/2024 the CO2 flowrate decreased to 157 mL CO2/min to account for seasonal decrease in productivity and then decreased again to 120 mL CO2/min on 11/15. It remained at this flowrate until 12/3/2024, which marked the end of the 2,000-hour DAC trial. Up until 10/25/2024, the set CO2 flowrate was constant throughout the day. As shown in data from Global Thermostat’s (GT) DAC Unit (Figure 5-24), CO2 production varied throughout the day. To account for this daily variation, the CO2 production from GT’s trial was averaged across all days for each 20-minute period throughout the day. This provided a daily average CO2 output (Figure 5-25) time series which was then programmed into the mass flow controllers to simulate. The average CO2 flowrate is still set based on productivity and carbon demand; however, it more realistically simulates small variations in DAC supplied CO2 flow into a raceway 0.15 Dry CO2 Per Cycle (kg) 0.10 0.05 0.00 8/21/24 8/22/24 8/23/24 8/24/24 8/25/24 8/26/24 8/27/24 8/28/24 8/29/24 Figure 5-24: Dry CO2 Output time series from Global Thermostat’s DAC trial. Final Report | MicroBio Engineering Inc. 46 170 0.10 0.09 150 Average Dry CO2 Per Cycle (kg) 0.08 CO2 Flow (mL CO2/min) 0.07 130 0.06 110 0.05 0.04 90 0.03 0.02 70 0.01 50 0.00 0:00 4:00 8:00 12:00 16:00 20:00 Figure 5-25: Example of mass flow controller daily CO2 flowrate to simulate the average DAC daily CO2 output. In this figure, the average CO2 flowrate is set to 290 mL CO2/min. 5.3.1.5 Raceway Harvest Raceways were harvested after reaching a density of approximately 400-500 mg VSS/L, diluted to a post- harvest density between 100 – 200 mg VSS/L. Culture productivity in the 3.4 m2-C and 22 m2 raceways dictated harvest frequency, ranging from weekly (Summer) to once per three weeks (Fall). In the afternoon prior to a harvest day, samples were collected from both raceways (‘Pre-harvest’), then 3.4 m2-C was drained into the 22 m2 raceway, with the mixed culture transferred into an 8,300 L cone bottom settling tank via a flexible impeller pump (Liverani FP-Mini 1”1/2) for overnight settling. The following morning, a portion of the supernatant in the cone bottom tank was drained into the 22 m2, the relative volume calculated to achieve a 100-200 mg VSS/L post-harvest biomass concentration. The settled fraction was fed to a disk-stack centrifuge (Alfa Laval Clara 20) via either a peristaltic (Cole- Parmer Masterflex L/S 7553-71 drive, Masterflex I/P 56C motor adapter with 3 to 1 speed reduction, Masterflex I/P Model 77601-10 pump head, ¼” ID tubing) or a flexible impeller pump (Liverani Pompa Minor 1”½ Flexible Impeller Pump) at a flow rate ranging from 8-16 L/min. The centrifuge allows for programmed or manual discharges of the algal sludge. Most times, the discharges were performed manually based on the feed and centrate clarity. The discharge interval ranged from 2 to 60 minutes, depending on feed density. Centrate was returned to the 22 m2 raceway. Once the harvest process was complete, make-up water and nutrients were added to the combined 3.4 m2 and 22 m2 volume, then a portion was transferred into 3.4 m2-C to initiate the next culture cycle. Harvested solids were frozen and saved for future analysis. Raceway samples were taken from both raceways before and after a harvest. Settling tank supernatant samples were composited over three evenly spaced samples. Slurry return samples from the settling tank used to reinoculated the raceways were collected beginning 9/10/2024. Composite samples were additionally collected of the centrifuge feed and centrate. Due to variable feed and centrate biomass concentration, five 300 mL samples for the centrate and feed flows were taken at evenly spaced time intervals during the centrifugation. The five samples for each flow were then mixed in a pitcher and collected as a composite sample representative of the feed and centrate. Final Report | MicroBio Engineering Inc. 47 On 11/12/2024, the 3.4 m2-C and 22 m2 raceways were contaminated with Picochlorum and as an attempt to address this, the supernatant in the cone bottom tank was centrifuged rather than immediately decanted. Picochlorum has much worse settling efficiency than Porphyridium, therefore the majority of the Picochlorum was expected to remain in the supernatant with the expectation that it would be removed during centrifugation. The centrate was then returned to freshly scrubbed raceways and sterilized with hypochlorite. The settled Porphyridium was then centrifuged, and the harvested cake was then used to reinoculate the raceways after the hypochlorite was neutralized. 5.3.1.6 Analytical and Calculations 5.3.1.6.1 Biomass Concentration Biomass concentration was estimated via total, volatile, and non-volatile suspended solids (TSS, VSS, and NVSS). TSS includes both biomass and ash, VSS is an estimate of biomass, and NVSS approximates ash content. Samples were measured in triplicate using pre-rinsed, ashed, then weighed glass fiber filters (4.7 cm diameter, 1.2 µm pore size). Culture volume was dispensed through the filter until the filtration rate decreased to a slow drip (approximately 1 drop per second), which generally would correspond to 5-10 mg of VSS collected. Replicates were rinsed with 20 ml of 0.3 M ammonium formate to remove culture media salts. Following the rinsing step, filters were dried in a 105°C oven overnight, weighed after 24 h, and subsequently ashed in a muffle furnace at 550 °C for 15 minutes prior to having their last weight collected. Biomass ash content was estimated as NVSS, the portion of the total suspended solids not lost during the combustion process. Results are expressed in mass of TSS, VSS, or NVSS per unit volume (mg/L). Centrifuge thickened biomass concentration was measured via total, volatile, and non-volatile solids (TS, VS, NVS). Approximately 1 g of wet material was loaded onto a pre-weighed aluminum foil tray, then processed analogously to TSS, VSS, and NVSS samples. Results were expressed in terms of g of TS, VS, or NVSS per unit volume (g/L) or percent solids (g/g wet material * 100%). 5.3.1.6.2 Biomass Productivity Biomass productivity was calculated from the starting and ending biomass concentration during the growth cycle between harvests. Once a harvest was complete and the ponds are at their operating depth (25cm), a sample was taken, marking the start of the growth cycle. The end of the growth cycle is when ponds were pumped into the cone bottom tank to settle overnight prior to a harvest. Productivity for each pond was calculated as the increase in biomass concentration (𝛥𝑉𝑆𝑆𝑝𝑜𝑛𝑑 ) from the start of the growth cycle until its end (equation below). D is the depth of the raceway at the time of sampling and Δt is the time interval between harvests. 𝑔 𝑉𝑆𝑆 (𝑉𝑆𝑆𝑓 ∙ 𝐷𝑓 ) − (𝑉𝑆𝑆𝑖 ∙ 𝐷𝑖 ) )= 𝑚 − 𝑑𝑎𝑦 ∆𝑡 5.3.1.6.3 Alkalinity Alkalinity was measured on pre and post samples on days of raceway dilution or harvest via titration with standardized 0.2 N sulfuric acid to an end point pH of 4.5. The normality of the acid was checked against a verification standard prepared from dry, analytical grade potassium bicarbonate and sodium bicarbonate with a total alkalinity of 125 meq/L. Based on verification standard titration, the normality was recalculated and used for sample alkalinity calculations (normality was typically measured as 0.19 N). Final Report | MicroBio Engineering Inc. 48 5.3.1.6.4 Microscopy Microcopy checks were performed on an “as needed” basis when culture health prompted closer examination i.e., culture discoloration, foul odor, changes in colloidal dispersion, or abrupt fluctuations in growth curves. Wet mounts were examined via brightfield microscopy at 100x, 400x, and 1,000x (with oil immersion) total magnification. 5.3.1.6.5 Settling Efficiency Biomass harvestability was assessed via gravity sedimentation. Pond grab samples were poured in 1 L Imhoff cones (𝑡𝑖 ) the afternoon before the harvest. A sample from the cone supernatant was then extracted approximately 2 cm from the air-water surface the following day (𝑡𝑓 ), allowing the Imhoff cone to settle overnight. Biomass concentration was measured as VSS, described earlier. Settling efficiency was calculated as per the formula below and reflects the percentage of initial biomass which can be removed with settling overnight. 𝑉𝑆𝑆𝑡𝑓 − 𝑉𝑆𝑆𝑡𝑖 𝑉𝑆𝑆𝑡𝑖 5.3.1.6.6 Dissolved Polysaccharides (DPS) Dissolved polysaccharide (DPS) levels were estimated by quantification of dissolved organic carbon (DOC). Particulate was removed via centrifugation at 3,000 rotations per minute (RPM) for 10 minutes and then filtered through a 1.2 µm glass fiber filter prior to analysis. These centrifuged and filtered samples were then preserved by acidification to a pH of less than 2 using 0.5mL of 9.38N sulfuric acid. DOC was quantified via non-purgeable organic carbon (NPOC) via a Shimadzu TOC-L-CPN analyzer. Briefly, the samples were sparged to remove inorganic carbon, then organic carbon catalytically oxidized to carbon dioxide at 680°C and quantified via a non-dispersive infrared (NDIR) sensor. Note that in the process of measuring DOC via NPOC, an unquantified amount of volatile organic carbon (VOC) may be inadvertently removed during inorganic carbon removal (sparging). VOCs are assumed to be negligible relative to the non-volatile carbon pool. These samples were oxidized via a high temperature furnace and the evolved CO2 quantified by coulometry. 5.3.1.6.7 Dissolved Inorganic Carbon (DIC) Dissolved Inorganic Carbon (DIC) was measured via the UIUC CM5130 acidification module and CM5014 Coulometer. Analysis involved acidification with 3N sulfuric acid, evolving DIC into CO2. CO2-free air carries the evolved CO2 into the coulometer where it is absorbed by a cathode solution and forms a titratable acid. A pH indicator reacts with the acid and as the color changes, the change in transmittance is measured. Based on change in transmittance, a titration current is developed in the cell to return the transmittance to its original value (29.5%). Using Faraday’s Law, the amount of carbon in the system is calculated based on the titration current required to restore the transmittance. Prior to sample analysis, a blank (4 mL distilled water) corrects for background CO2. Accuracy was checked with analytical grade sodium carbonate, with an acceptance threshold of 10% of the theoretical mass of carbon. Sample volume was chosen to load 1 to 3 mg of carbon, generally requiring 3 mL at the experimental alkalinity and pH (80 meq/L, pH ~9). Samples were analyzed in duplicate, and the reported values are the average of the two measurements. Inorganic carbon of microalgal culture samples can be analyzed either before or after removal of suspended solids – i.e. algal biomass particulate and perhaps trace precipitated inorganic carbon. A side- study compared IC concentration of filtered (1.2 µm nominal pore size, glass fiber filters) and unfiltered samples, finding similar concentrations for either preparation. Thus, samples were analyzed without Final Report | MicroBio Engineering Inc. 49 prior particulate removal to avoid unnecessary sample gas-liquid exchange, which could otherwise augment DIC. Values are reported as “dissolved inorganic carbon (DIC)”, although may contain a small amount of particulate inorganic carbon. A further assumption is that minimal inorganic carbon is contributed from cellular respiration during the sample hold time or sequence of analysis. 5.3.1.6.8 Mass Flow Controller (MFC) CO2 Flowrate The flowrate displayed on the MFC is expressed as mL of N2 equivalents per minute at standard temperature and pressure (𝑄𝑁2 ; 273 K and 101,325 Pa). The user’s guide provides a table of K factors to convert between N2 equivalent flow rate and various gases. The K factors for N2 and CO2 are 1.000 and 0.739, respectively (FMA 5400A/FMA 5500A Mass Flow Controllers, n.d.). Additionally, the density of CO2 ( 𝜑𝐶𝑂2 = 1.964 mg/mL) is provided and from this, the mass flow rate of CO2 is calculated as shown in the equation below. 𝑚𝑔 𝐾𝐶𝑂2 𝑀̇𝐶𝑂2 ( ) = 𝑄𝑁2 ∗ ( ) ∗ 𝜑𝐶𝑂2 𝑚𝑖𝑛 𝐾𝑁2 5.3.1.6.9 Weather Conditions Air temperature and solar radiation were downloaded from the California Irrigation Management Information System (CIMIS) from Station 265, approximately 6.5 miles north-northwest of the experimental location. Measurements were recorded hourly. Air temperature measurements were collected at a height of 1.5 m above the ground using a thermistor. Total incoming solar radiation was measured with a pyranometer (W/m2) at a height of 2.0 m above the ground. 5.3.1.7 Carbon Modeling Calculations 5.3.1.7.1 Carbon Utilization Efficiency, Mass closure Carbon Utilization Efficiency (CUE) is defined as the ratio between the amount of biological carbon accumulated over a growth period and the total carbon supplied through the mass flow controller during the growth period, as shown in the equation below. ∆𝑁𝑃𝑂𝐶 ∗ 𝑉 + ∆𝑉𝑆𝑆 ∗ 𝛼 ∗ 𝑉 𝐶𝑈𝐸 = 𝑀̇𝐶𝑂2 ∗ Δ𝑡 NPOC (mg/L) represents the concentration of non-purgeable organic carbon (equal to the dissolved organic carbon), VSS (mg/L) represents algal biomass as volatile suspended solids, and V (L) is the raceway’s volume. 𝛼 is the mass fraction of carbon in TSS (40%), derived as the average value of total carbohydrates in P. cruentum presented by Fuentes (2000) and Di Lena (2020), using a 1:1 ratio of carbon to carbohydrates (Albalasmeh et al., 2013). 𝑀̇𝐶𝑂2 is the mass flowrate of CO2 through the mass flow controller, on a g C basis, and Δ𝑡 is the length of time CO2 was injected. Carbon introduced during evaporative make-up, which generally accounted for less than 5% of the total carbon provided, was excluded. Additionally, with nitrate as the nitrogen source, urea derived carbon is not included. Following from the definition above, the theoretical maximum CUE is equal to the injection efficiency, in an idealized case with no outgassing losses and all injected (and dissolved) CO2 being assimilated into algal biomass. Relative accuracy of the calculated CUE was assessed by calculation of the carbon mass-closure for a given growth period. Mass closure was determined by comparing the net measured change in carbon (DIC, NPOC, and carbon contained within algal biomass) to the net carbon input and output. Carbon inputs included gaseous CO2 injected via the sparger and inorganic carbon introduced from evaporative make-up water. Carbon outputs included CO2 lost either during injection (bubble ebullition) or via outgassing. Injection losses were calculated from the CO2 injection time-course and measured injection Final Report | MicroBio Engineering Inc. 50 efficiency, while outgassing losses were calculated as described below (Section 1.2.5.3). Mass closure was then calculated as follows: 𝐶𝑙𝑜𝑠𝑢𝑟𝑒 % = 𝑁𝑒𝑡 𝐶𝑎𝑟𝑏𝑜𝑛 𝐼𝑛𝑝𝑢𝑡 An example mass-closure assessment is shown schematically in Figure 5-26, corresponding to the 10/8 – 10/21/2024 growth period. For each pond, the left-hand bar represents the net carbon increase in the raceway, based on the measured change in DIC, NPOC, and algal-carbon, while right-hand stacked bar represents the net of carbon sources and losses. Net values are represented by black circles on either bar. In this example, the net change in carbon exceeded the sum of carbon inputs/outputs, in each the 3.4C and 22 m2 ponds, thus yielding a mass closure greater than 100%. The mass closure is used to interpret the CUE reported over the given period. In this case, possible interpretations include that the measured change in carbon was overestimated, the amount of carbon injected was underestimated, or injection losses/carbon lost to outgassing over-estimated. For the former two scenarios, as these directly relate to the CUE definition above, the net effect, in this example, is that the reported CUE is likely overreported. Figure 5-26: Example carbon balance in the 3.4 and 22 m2 raceways from 10/8 to 10/21/2024. The left stacked bar in both plots is the net change in VSS carbon (blue), NPOC (orange), and DIC (yellow). The right stacked bar in both plots is the carbon sources: injected carbon (purple) and Azul well water DIC (yellow); and carbon losses: outgassing (orange) and injection losses (blue). The black dots represent the net value of the total stacked bars. Final Report | MicroBio Engineering Inc. 51 5.3.1.7.2 CO2SYS Modeling Inputs All calculations to determine the speciation of carbon in the raceway’s media was performed using CO2SYS (Sharp, 2019/2024), a MATLAB CO2 equilibrium software used to calculate carbon speciation in marine environments. The user inputs two carbon parameters (either alkalinity, DIC, pH, pCO2, fCO2, HCO3-, CO32-, or CO2) and the rest of the parameters are calculated. The two most common inputs for this study are pH and alkalinity. pH was the measured value of the in-situ probe during the time of sampling and the measured alkalinity via titration (consistently measured near 80 meq/L, the alkalinity of the 35/80 media). Additionally, the sample salinity, temperature, and pressure must be inputted into the software. Salinity was inputted as 35 ppt (salinity of the 35/80 media), temperature was what the in- situ probe measured during the time of sampling, and pressure was set to 0 dbar (samples were not taken at a significant depth for hydrostatic pressure to have an impact). The concentrations of silicon, phosphate, ammonium, and hydrogen sulfide are also included as these contribute to alkalinity. Due to the high alkalinity of media and the bulk of it being contributed from bicarbonate and carbonate, these concentrations are set at 0 µmol/L. The pH scale (total, seawater, free, and NBS) must be defined to convert dissociation constants to that scale. The in-situ probes measure pH based on the NBS scale, which may lead to some error within the calculations as the NBS scales are standardized for low ion activities, whereas seawater and the 35/80 media have significantly higher activities. CO2SYS does correct for NBS pH based on the salinity, however, this correction isn’t unique to our special media, and it may introduce errors into the calculations. The dissociation constants for carbonic acid must also be defined by the user. Embedded within CO2SYS are the dissociation constants published from literature for different salinities, pH scales, and type of water (artificial seawater, real seawater, or fresh water). To determine the most accurate set of constants to use for the 35/80, raceway DIC coulometry measurements were compared to CO2SYS predicted DIC values for each set of dissociation constants. Alkalinity and pH were measured along with each DIC sample (Figure 5-27, LEFT). These values were input into CO2SYS and DIC was calculated using each set of dissociation constants (Figure 5-27, RIGHT). Unexpectedly, the freshwater constants reported by (Millero, 1979) produced the smallest difference when compared to coulometry measured DIC. The rest of the other dissociation constants proposed for saltwater, calculated DIC values that were significantly less than measured values. Therefore, the freshwater constants were used for all CO2SYS calculations. Final Report | MicroBio Engineering Inc. 52 Figure 5-27: (LEFT Panel) Measured DIC samples via coulometry and the sample’s pH and alkalinity. (RIGHT Panel) CO2SYS calculated DIC based on the sample’s pH and alkalinity compared to the measured DIC. The 1:1 reference line indicates that CO2SYS calculated DIC perfectly aligns with measured DIC values. 5.3.1.7.3 Outgassing Rates As CO2 is continuously injected into the pond, the bulk CO2 concentration increases and is greater than the air-water CO2 equilibrium concentration. This results in CO2 outgassing from the raceway. The outgassed CO2 flux is the product between the mass transfer coefficient, k, and the driving force, which is the difference between the bulk and equilibrium aqueous CO2 concentration and is calculated as shown in the equation below. 𝑔 𝑚 𝑔 𝑔 𝑚2 ∙ ℎ𝑟 ℎ𝑟 𝑚 𝑚 Outgassing is calculated during a growth period (immediately after a harvest up until being transferred to the cone bottom settling tank) using MATLAB program and CO2SYS (Sharp 2024) software. To determine bulk CO2 concentrations throughout the growth period, Neptune pH measurements and starting/ending alkalinity measurements (which are linearly interpolated to align with pH measurements) are used as CO2SYS inputs to calculate bulk CO2. For equilibrium CO2 concentrations, equilibrium pCO2 and alkalinity are used as the inputs. Equilibrium pCO2 in the water is equal to air pCO2. From this, the CO2 flux is calculated every ten minutes within the growth period. The corresponding total mass of CO2 outgassed is the product between the flux, raceway’s surface area (22 or 3.4 m2), and the length of time (10 minutes). 5.3.1.7.4 Equilibrium Air/Water CO2 Concentration Atmospheric CO2 concentration in the greenhouse was occasionally measured (Figure 5-28) using an infrared gas analyzer system (LI-840/A) that was continuously pumping air from the greenhouse. The average CO2 concentration was 425 ppm (standard deviation = 41), with a minimum value near 400 ppm and a maximum value never exceeding 1,000 ppm. The large spikes in concentration represent times when the greenhouse was not vented. The average concentration (425 ppm) was used, when relevant, Final Report | MicroBio Engineering Inc. 53 in all carbon modeling calculations. 425 ppm in the atmosphere corresponds to a pCO2 of 425 µatm in water, which was used in all CO2SYS inputs as equilibrium dissolved CO2. 1100 1000 900 Atmospheric CO2 (ppm) 800 700 600 500 400 300 10/5 10/10 10/15 10/20 10/25 10/30 11/4 11/9 11/14 11/19 11/24 Figure 5-28: Time series of atmospheric CO2 concentration in the greenhouse throughout the cultivation trial. 5.3.1.8 Abiotic Testing 5.3.1.8.1.Pond Preparation Once the 2,000-hour DAC trial was completed, the 3.4C and 22 m2 raceways were centrifuged to remove any remaining biomass and the centrate was returned to the raceways. Azul well water was then added to return the raceways to their operational depth of 25 cm and alkalinity and salinity components were added to ensure the raceways were at an alkalinity and salinity of 80 meq/L and 35 ppt, respectively. The raceways were then disinfected with sodium hypochlorite to minimize biotic effects. Prior to any abiotic testing, the raceways were neutralized with sodium thiosulfate. 5.3.1.8.2. Mass Transfer Coefficient In or order to determine CO2 outgassing rates from the raceways, the mass transfer coefficient, k, needed to be determined for the 3.4 and 22 m2 raceways. The mass transfer coefficient can be calculated as the slope between the flux and driving force. Flux is the rate of CO2 outgassing from the raceway whereas the driving force is the difference between the bulk CO2 concentration in the raceway and theoretical equilibrium concentration of CO2 in the raceway. To calculate the flux and driving force, the raceways were injected with pure CO2 to decrease pH. For the 3.4 and 22 m2 raceways, the pH was brought down to 7.2 and 8.0, respectively. The CO2 supply was then stopped, and CO2 was allowed to outgas from the raceways until the equilibrium pH (approximately 9.0) was nearly reached. In situ probes provided time series of pH and temperature every ten minutes. DIC, bulk CO2 concentration, and theoretical equilibrium CO2 concentration were calculated at every tenth of a pH. These values were calculated using CO2SYS for MATLAB with pH and alkalinity as the calculation inputs. Flux was calculated based on the difference in the DIC for every tenth of a pH increase using the following equation. Final Report | MicroBio Engineering Inc. 54 𝑔𝐶 (𝐷𝐼𝐶𝑡𝑓 − 𝐷𝐼𝐶𝑡𝑖 ) ∙ 𝑑𝑒𝑝𝑡ℎ 𝐹𝑙𝑢𝑥 ( 2 )= 𝑚 ∙ ℎ𝑟 𝑡𝑓 − 𝑡𝑖 Driving force was calculated as the difference in bulk dissolved CO2 concentration and the theoretical equilibrium dissolved CO2 concentration, shown in the equation below. Like DIC, the bulk dissolved CO2 concentration was calculated for every tenth of a pH using CO2SYS with pH and alkalinity as the inputs. Equilibrium dissolved CO2 was calculated with pCO2 and alkalinity as the inputs, with pCO2 equal to atmospheric pCO2 (425 µatm). 𝑔 𝐷𝑟𝑖𝑣𝑖𝑛𝑔 𝐹𝑜𝑟𝑐𝑒 ( ) = 𝐶𝑂2𝑏𝑢𝑙𝑘 − 𝐶𝑂2𝐸𝑞. 𝑚3 Once the flux and driving force were calculated for every tenth of a pH, the flux was plotted against the driving force. A linear fit was then fitted to the data with the intercept forced through the origin. The mass transfer coefficient is the slope of the linear fit. For the 3.4 m2-C pond, a novel technique was additionally used to calculate the mass transfer coefficient. A marble equilibrator, based off the design of Frankignoulle (2001), was used to determine the partial pressure of CO2 in the raceway throughout the outgassing time-course. The marble equilibrator functions by pumping the raceway water counter current to a closed air stream. The water and air streams are pumped through an equilibration chamber filled with small marbles, promoting turbulence in the system and increasing the exchange surface area between the water and air (Figure 5-29). This process forces the air stream to equilibrate with the water stream and an Infrared Gas Analyzer (LI-840/A) measures the resulting CO2 concentration in the air stream, which represents the water’s pCO2. One potential effect of the system is that the water effluent from the system slightly agitates the raceway’s flow, however, this effect is minimal compared to the raceway’s surface area. The mass-transfer coefficient is calculated analogously to the pH/alkalinity approach described earlier, replacing pH with the marble-equilibrator measured pCO2 as a calculation input in determining carbon system parameters, and by extension the driving force and air-water flux. The benefit of using the marble equilibrator is to avoid the use of pH as a calculation input, which can give erroneous readings for waters with seawater level salinity and isn’t as sensitive to changes in CO2. Final Report | MicroBio Engineering Inc. 55 Figure 5-29: Schematic of air water and flow through an experimental marble equilibrator set up used to measure in situ pCO2 in the water. 5.3.1.8.3 Injection Efficiency Each raceway contains a sparger to inject pure CO2 into the raceway. A portion of this CO2 is dissolved and hydrated; however, the remaining is lost to the atmosphere in the form of CO2 bubbles that aren’t dissolved which are considered as injection losses. To determine injection efficiency, CO2 is sparged into the raceways to drive the pH down (same minimum pH as the mass transfer coefficient study) where in situ probes record the pH and temperature time series. For 3.4 m2-C, injection efficiency was calculated for one flow rate, 350 mL CO2/min, which is the raceway’s operational flow rate. The 22 m2 injection efficiency, however, was determined for two flowrates since flow rate in the 22 m2 raceway varied throughout the quarter. The two flow rates tested in the 22 m2 were 350 mL CO2/min and 175 mL CO2/min. DIC and dissolved CO2 is calculated using CO2SYS with pH and alkalinity as the inputs. DIC is used to quantify the amount of CO2 injected into the system. Dissolved CO2 is used to calculate outgassing rates of CO2 as this represents a portion of CO2 injected that was hydrated but eventually lost to outgassing. The outgassed carbon is multiplied by the ratio of injected DIC to total DIC to represent the portion of outgassed carbon that was injected during that period. From the change in the DIC and outgassed carbon, injection efficiency can be calculated from the following equation, where 𝑀̇ is the mass flowrate of CO2, V is the pond’s volume and ∆𝑡 is the time. The injection efficiency is calculated for every tenth of a pH unit decrease. ∆𝐷𝐼𝐶 ∆𝐷𝐼𝐶 ∙ 𝑉 + ∙ 𝑂𝑢𝑡𝑔𝑎𝑠𝑠𝑒𝑑 𝐶 𝐷𝐼𝐶𝑓 𝐼𝑛𝑗𝑒𝑐𝑡𝑖𝑜𝑛 𝐸𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑐𝑦 % = ∙ 100% 𝑀̇ ∙ ∆𝑡 5.3.1.8.4. DAC water biocompatibility test An indoor cultivation trial was designed to address possible concerns with DAC-microalgae integration. One potential scenario is contamination of the DAC product CO2 with adsorbent materials, with unknown effects on microalgal growth once injected into the culture. DAC adsorbents may contain ethanolamines which others have previously shown to inhibit algal growth, for example. In this effort, the project team identified two water byproducts from the DAC unit as probable locations where such contaminants may accumulate. The first water sample was collected after desorption of CO2 from the Final Report | MicroBio Engineering Inc. 56 adsorbent, after removing water vapor by condensation, referred to as the “DAC unit condensate”. The second was collected from the compressor, where a small amount of residual water, not removed during primary dehydration, accumulated over time. (Figure 5-35). Insufficient volume of the second sample was available for a cultivation trial, although an ad-hoc test confirmed algal growth once supplemented with nutrients. The section below describes the methodology and results of a cultivation trial using the DAC unit condensate. No water quality analyses were conducted to measure possible contamination directly. Figure 5-30: As-received DAC water samples. Left is the "DAC unit condensate", and right is the low-volume sample of accumulated water in the compressor. These were amended with nutrients and used in a cultivation trial to assess whether DAC products may contain compounds inhibitory to algal growth. 5.3.1.8.5. Methods One liter of 35.80 media was prepared using either Azul well water or DAC unit condensate as the base water. Three 250 mL flasks were then autoclaved and prepared for each media type, and each was inoculated with P. cruentum, the same alga as used during the Summer 2024 cultivation trial. These flasks were left undisturbed for two weeks to acclimate cultures to each media type. Cultures were subsequently pooled into two groups for each media type, an aliquot of each used to initiate the trial, consisting of triplicate 250 mL flasks for each media type. Cultures were continuous illuminated and cultivated at room temperature. Each day, the flask positions were rotated to ensure even light exposure. The trial concluded after eight days, when biomass concentration was estimated as total and volatile suspended solids (TSS, VSS), measured in triplicate for each flask. 5.3.2. Results 5.3.2.1 Biomass Production Weather conditions during the cultivation campaign were typical of late summer and fall conditions for inland-coastal Central California, with raceway temperatures following seasonal trends in temperature (Figure 5-31). Up until the end of September, ambient air temperatures were hot (upwards of 40ºC) Final Report | MicroBio Engineering Inc. 57 resulting in raceway temperatures ranging from 15 to 35ºC in the raceways. Temperature probes in the 3.4 m2 raceways provided likely erroneous measurements until 9/21/2024 as maximum raceway temperatures greater than 40ºC are unreasonable. Throughout the end of fall, raceway temperatures and average daily solar irradiance decreased with some days in November having minimum raceway temperatures near freezing. Temperature ranges within 3.4 m2-C were larger than the 22 m2 as its smaller volume allowed quicker heat transfer rates. In general, however, raceway temperatures were within 10-20ºC throughout the end of the trial. Singular large temperature spikes are due to probes being left in a small pitcher when raceways were empty. Figure 5-31: Raceway temperature and daily average solar irradiance throughout the cultivation trial. The raceway’s pH was well managed and kept near a pH of 9.0 (Figure 5-32). pH in 3.4 m2-C was normally within 8.9 – 9.1, however, a few lapses in CO2 supply allowed for the pH to climb as high as 9.25. The pH in the 22 m2 raceway was also at a reasonable level, however, a leak in the CO2 line allowed the pH to climb up to 9.3 on 9/17/2024. At the start of November, the pH was increased by turning off CO2 supply in both raceways as a method of addressing Picochlorum contamination (Picochlorum grows at a pH near 7). We assumed this to also be analogous to DAC unit repairs and servicing. The 22 m2 raceway CO2 supply was totally shut off and the pH climbed as high as 9.5. 3.4 m2-C, on the other hand, still had CO2 dosing, but pH thresholds set to maintained at a higher pH of 9.2. On 11/16/2024, the CO2 flow was mistakenly set on in 3.4 m2-C which caused the pH to decrease down to 8.6, however, this was resolved and the pH returned to its operational value of 9.2. Salinity was consistently measured near its operational value, with the average salinity measured being 34.7 ppt. Alkalinity at the start of each growth period was consistently measured near 80 meq/L, however, evaporative top off and respiration allowed the alkalinity to climb as high as 95 meq/L towards the end of the growth period. Final Report | MicroBio Engineering Inc. 58 Figure 5-32: Time series of pH and CO2 injection flowrate throughout the cultivation trial. Biomass concentration time series through the 2,000-hour DAC of the 22 m2 and 3.4 m2-C raceways are presented in Figure 5-33 (Top). At the start of the trial, productivity was high, and two harvests occurred within one week of each other in mid-September. On 9/24/2024, both raceway’s color was observed to be pallid brown rather than P. cruentum’s distinct deep maroon color. This was suspected to be due to a lack of phosphorous availability. Therefore, a full dose of phosphorus nutrients was added to the 22 m2 raceway, however, as a precaution, 3.4 m2-C was reinoculated with culture grown in 3.4D. Following the phosphorus addition to the 22 m2, growth was observed, and a harvest occurred on 10/1/2024. Following the harvest, both raceways crashed on 10/4/2024, suspected to be due to Picochlorum contamination, which was being grown in neighboring raceways. Observation of the media revealed a green hue to the water indicating its presence. 3.4 m2-C was reinoculated from 3.4D and 3.4 m2-C was allowed to reach a high enough density to inoculate the 22 m2 raceway. By 10/8/2024, both ponds were inoculated, and a successful harvest occurred on 10/22/2024. Discoloration continued, and microscopy of both raceways revealed contamination of not only Picochlorum but also predatory flagellates in the media (Figure 5-34). Ammonium bicarbonate has been cited (He et al., 2021) as a form of managing flagellates growth in ponds. Therefore, on 11/1 ammonium bicarbonate was added to both raceways at a concentration of 100 mg/L. In an attempt to address Picochlorum contamination, the pH in both ponds were raised to over 9, based on an unsupported hypothesis that Porphyridium may exhibit faster relative growth under low-CO2 conditions. Picochlorum, however, remained present in the media. As a novel technique to remove Picochlorum, the raceways were allowed to settle and harvested on 11/12/2024. Porphyridium biomass at this time exhibited high settleability, whereas Picochlorum was poorly settleable. The supernatant was centrifuged to remove any non-settleable cells (such as Picochlorum) and the centrate was returned to the raceways. Unfortunately, following the harvest, pest contamination remained. Ammonium bicarbonate was again added on 11/26/2024, although Porphyridium growth continued to decline, with 12/3 marking the official end of the cultivation trial. Final Report | MicroBio Engineering Inc. 59 Productivity in both raceways decreased throughout the trial (Figure 5-33, Bottom). This is due to a combination of seasonal temperature changes and pest pressure. Initially, productivity reached 11 - 14 g VSS/m2-day in the 22 m2 and 3.4C raceways, respectively. However, by the end of fall, productivity was approximately 5 g VSS/m2-day for both raceways. Throughout the trial, 3.4 m2-C exhibited higher productivity when compared to the 22 m2, with the gap being more prominent at the start of the trial and decreasing towards the end. Dissolved oxygen measurements revealed that the 22 m2 raceway had higher dissolved oxygen levels than 3.4 m2-C. This corroborates the decrease in productivity difference throughout the trial as solar irradiance decreases throughout the end of fall. Figure 5-33: Biomass concentration (mg VSS/L; UPPER Panel) and productivity (g VSS/m2-day; LOWER Panel) time series in the 3.4C and 22 m2 raceways. Biomass productivity was calculated from start to end of a given growth cycle, around 1-3 weeks depending on pond operations. Horizontal lines in the lower panel denote time periods over which productivity was calculated. Black circles indicate harvest dates and asterisks indicate raceway re-inoculation. Final Report | MicroBio Engineering Inc. 60 Figure 5-34: Microscopy of predatory flagellates (LEFT) and Picochlorum (RIGHT) contamination in the 22 m2 raceway observed on 11/1. 5.3.2.2 NPOC and Settleability Rather than a traditional cell wall, P. cruentum cells are encapsulated in a polysaccharide sheath. Solubilization of this sheath and accumulation in the culture media is widely documented in literature. Solubilization is generally initiated by a nutrient depletion or onset of the stationary phase. NPOC, used as a surrogate for dissolved polysaccharides (DPS), levels in the 3.4 m2-C and 22 m2 raceways ranged between 40 – 70 mg C/L, as shown in Figure 5-35. Since 3.4 m2-D was not maintained, DPS accumulated in the raceway, reaching nearly 100 mg C/L. Prior to the initial inoculation of the 22 m2 raceway on 8/15, 3.4 m2-C was allowed to grow without dilution and DPS levels climbed as high as 80 mg C/L. The 40 mg C/L baseline concentration is suspected to potentially be contributed from carbon impurities present in the NaCl component of the 35/80 growth media. The large concentration spike on 8/21/2024 (1,400 mg C/L) was initially suspected to be from the shearing and removal of polysaccharides attached to the cell wall from the centrifuge and therefore increasing dissolved concentrations. NPOC analysis of the centrate, however, demonstrated no increase in NPOC concentrations and therefore the measurement on 8/21/2024 is likely due to analytical error. During each growth period in between harvests, DPS levels increased, and then brought down to the 40 mg C/L threshold after the harvest dilution. From the start of September, NPOC gradually increased from 40 mg C/L reaching upwards of 70 mg C/L in both raceways. Due to the reinoculation of both raceways on 9/24/2024 and 10/8/2024, NPOC levels fell to approximately 15 mg/L, and a d The trend of DPS accumulating and then being diluted from inoculation occurred throughout the rest of the cultivation trial. The NPOC never returned to its original value of 40 mg/L, perhaps due to limited polysaccharide production because of decreased productivity (resulted from pest contamination and colder temperatures). Settling efficiency of the raceways was initially very high (upwards of 75%), however, after the first harvest, settling efficiency decreased, reaching a low of 20% during the harvest on 9/2/2024, shown in Figure 5-36. It was suspected that centrifuging the supernatant and returning the centrate to the raceway was unintentionally inoculating the raceway with the non-settleable algal phenotype. To overcome this, during the 9/27 harvest, a portion of the settled solids in the settling tank was returned to the raceway prior to centrifugation. The remaining supernatant was then centrifuged and the centrate was directed to the raceway. This resulted in increased settleability for the next harvest and this process continued for the rest of the harvests and settling efficiency. Settling efficiency continued to increase for both the 22 m2 and 3.4C raceways. The last harvest’s settling efficiency for the 3.4C Final Report | MicroBio Engineering Inc. 61 raceway, however, drastically dropped to below 10%, most likely an analytical error or unnoticed agitation of the Imhoff cone. The following two harvests, however, demonstrated remarkable settleability, with cells visibly settling within minutes (Figure 5-37). On 11/12/2024, settling efficiency was upwards of 100%. Observation of the raceways during this time revealed high flocculation of the Porphyridium cells, perhaps due to seasonal trends or a contamination response. This flocculation contributed to its high settling efficiency velocity, with noticeable changes in density occurring within minutes. 200 180 160 140 NPOC (mg C/L) 120 3.4D 100 3.4C 80 22 60 40 20 0 7/7 7/27 8/16 9/5 9/25 10/15 11/4 11/24 12/14 Figure 5-35: Time series of NPOC (used as a proxy for dissolved polysaccharides) in the raceways. Final Report | MicroBio Engineering Inc. 62 100% 3.4C 22 3.4C+22 Mix 90% 80% Settling efficiency (% removal) 70% 60% 50% 40% 30% 20% 10% 0% 8/6 8/26 9/15 10/5 10/25 11/14 Figure 5-36: Time series of settling efficiency from Imhoff cones for each raceway. “3.4C+22 Mix” refers to an Imhoff cone containing a composite sample of the total 3.4C raceway volume added to and mixed with the full 22 m2 raceway. Figure 5-37: Biomass gravity sedimentation after eight minutes, observed on 11/26. The background green color indicates likely Picochlorum contamination. 5.3.2.3 Harvest Operations Throughout the cultivation trial, 9 harvests were performed with 5 occurring within the 2,000 DAC trial. Harvest flow concentrations are presented in Figure 5-38. Despite harvests consistently occurring when raceway densities reached between 400 – 500 mg VSS/L, settling efficiency impacted centrifuge feed concentrations, ranging from 500 mg VSS/L to upwards of 4,000 mg VSS/L. A common issue resulting in variable feed concentration was that the settled P. cruentum would attach to the walls of the cone bottom settling tank, requiring it to be rinsed off. Mixing and recirculation and mixing of the settled solids in the cone bottom did not resolve this issue and a more effective technique that does not dilute the settled solids needs to be developed. This resulted in variable feed concentration and is the reason the 11/12 harvest has two feed concentrations with a large range (70 to 4,000 mg VSS/L). Despite the Final Report | MicroBio Engineering Inc. 63 variable feed concentration, centrate quality was consistent at 50 mg VSS/L, however, on 10/1 its density was 150 mg VSS/L due to exceptionally poor settling efficiency and low feed concentration. Harvested wet cake yields ranged from 0.8 kg to over 5 kg, with largest yield being on 11/12 (5.4 kg) and the lowest being on 9/30 (0.83 kg). Most harvests required refining centrifugation parameters and if the discharge consistency was not ideal, it was returned to the centrifuge feed. Harvest operations also affected cake quality, ranging from 5.7% (9/30) to 12% VS (11/12). Similarly, cake ash content ranged from 25% to 60%. On 12/4, harvested cake was washed with Azul well water and recentrifuged with a bench centrifuge (9,000 rpm for 10 minutes) and the Alfa Laval disc stack centrifuge to determine washing’s effect on cake ash content. Prior to any washing, the ash content was 58%. Following washing and centrifuging with the disc stack centrifuge, the ash content was reduced to 41%. Centrifuging with the benchtop centrifuge reduced ash content to 29%, half of the original ash content. 4,500 70% 4,000 Biomass concentration, mg VSS/L 60% 3,500 Feed 50% 3,000 Centrate Cake - % VS 2,500 40% Supernatant cone decant 2,000 Slurry return 30% Cake - %VS 1,500 20% Cake - %Ash 1,000 10% 500 0 0% 6/7 7/27 9/15 11/4 12/24 Figure 5-38: Time series of harvest operations and resulting centrifuge operation flow biomass concentrations. 5.3.2.4 Water Usage and Recycling A unique component of the cultivation trial was the use of media recycling following each harvest. Instead of preparing fresh growth media at the start of each growth period, media from the previous growth period was recycled. As shown in Figure 5-39, each growth period required on average 6,700 L of water total. The volume of this water that was recycled was dependent on harvest operations and raceway operations. For the first three, fifth, and sixth growth periods, water recycling is high, however, its variability is due to water recovery operations during the harvest as some may have been lost. Additionally, evaporation losses were not replenished prior to a harvest, therefore the centrifuged volume and volume of media recycled was less. For the “9/25-9/30” and “10/5-10/21” growth periods, raceways were contaminated or were nutrient depleted, therefore reinoculation and fresh media preparation was required. This drastically reduces water recycling as the only recycled water came from inoculum, which was typically half of the volume of 3.4D (approximately 500 L). Despite not being reinoculated, the last growth period had a significantly less percentage of water recycling as the media was contaminated with Picochlorum and therefore, a portion of the recovered media was discarded. Overall, pond contamination and algae health had the largest impact on the extent to which media could be recycled. Reinoculation requires preparation of fresh media, therefore, raceway operations should be optimized to prevent contamination. Another conflicting factor is optimizing media Final Report | MicroBio Engineering Inc. 64 component additions following a harvest to ensure concentrations at their original values. Despite these factors, media was successfully recycled throughout the growth period and across the whole cultivation trial, water demand was 2,400 L/kg VSS grown. If fresh media was prepared for every growth period and assuming a fifth of the media was recycled to reinoculated the raceways, the water demand would be 3,100 L/kg VSS. These values, however, include raceway reinoculation events and if raceways operated at ideal conditions with no need to re-inoculate, water demand could be as low as 1,600 L/kg VSS. Recycled Media Evap Top Off Fresh Media Innoculum 9000 66% 8000 72% 7000 90% 8% 46% 62% 77% 6000 9% Volume (L) 5000 4000 3000 2000 1000 0 9/3-9/9 9/10-9/16 9/17-9/24 9/25-9/30 10/1-10/4 10/5-10/21 10/22-11/11 11/12-11/26 Figure 5-39: Water usage breakdown for each growth period throughout the cultivation trial. “Recycled Media” refers to growth media recycled following a harvest whereas “Fresh Media” indicates newly prepared growth media in fresh Azul well water. “Evap Top Off” refers to Azul well water added to raceways to replenish evaporation losses and “Inoculum” refers to media with culture used to re-inoculate the raceways. The percentage above each bar denotes the percentage of water recycled. 5.3.2.5 Abiotic Study Trials 5.3.2.5.1 Mass Transfer Coefficient For the outgassing trial, both 3.4 m2-C and the 22 m2 were sparged with CO2 with the raceways reaching a minimum pH of 7.2 and 8.0, respectively (the 22 m2 took significantly longer to decrease its pH, therefore a starting pH of 8 was deemed satisfactory given time constraints). The CO2 supply was then turned off and allowed to outgas until they approached their equilibrium pH of approximately 9.0. From the pH and alkalinity, the DIC time series was calculated for both raceways as well as the dissolved bulk CO2 saturated CO2 assuming an equilibrium pCO2 of 420 µatm, shown in Figure 5-40 (22 m2 trial) and Figure 5-41 (3.4C trial). For every tenth of a pH increase, the flux and driving were calculated. A linear regression was then performed for the driving force and flux and the intercept was forced through the origin. The slope of the linear regression represents the mass transfer coefficient, and it was calculated to be 0.05 and 0.01 m/hr for 3.4 m2-C and the 22 m2 raceways, respectively. The linear regression for 3.4C had a significantly higher R2 of 0.91 compared to the 22 m2 regression’s R2 of 0.68. Since 3.4C has a smaller volume than the 22 m2, driving the pH down to a lower value was much quicker than the 22 m2, allowing for 3.4C to have a much larger measured pH range of 7.2 to 9.0 compared to the 22 m2’s 8.2 to 9.0. This led to more flux and driving force measurements which in turn increases the sample size for the linear regression. Additionally, the outgassing period for 3.4 m2-C was Final Report | MicroBio Engineering Inc. 65 shorter than the 22 m2’s, requiring a little over 6 and 10 days for both, respectively. The quicker outgassing time for 3.4 m2-C allowed for quicker changes in pH whereas for the 22 m2, hours would pass before a change in pH was measured (depending on the temperature, this could result in an increase in DIC as the pH probe resolution couldn’t measure subtle increases in pH). This required filtering the 22 m2’s data so that a consistent decrease in DIC was observed. Therefore, when calculating the flux, the uncertainty in time between an increase in pH increased the fluxes error. Figure 5-40: Time series of pH (UPPER LEFT), DIC (UPPER RIGHT), and driving force (LOWER LEFT) throughout the outgassing trial in the 22 m2 raceway. The slope of linear regression between the driving force and flux (LOWER RIGHT) represents the mass transfer coefficient. Final Report | MicroBio Engineering Inc. 66 Figure 5-41: Time series of pH (Upper Left), DIC (Upper Right), and driving force (Lower Left) throughout the outgassing trial in 3.4C. The slope of the linear regression between the driving force and flux (Lower Right) represents the mass transfer coefficient. During 3.4C’s outgassing trial, a marble equilibrator was used to monitor pCO2 throughout the first six hours of the outgassing period (Figure 5-42, Upper Left). DIC calculated with pH and pCO2 were compared to measured coulometry DIC values (Figure 5-42, Upper Right). pCO2 calculated DIC values consistent with measured DIC values, whereas pH predicted DIC values were much higher (initially greater by near 150 mg/L), however, this difference decreases as the pH rises to equilibrium. The same phenomenon is observed when calculating bulk CO2 concentrations and the corresponding driving force (Figure 5-42, Lower Left). The discrepancy between pH and pCO2 predicted DIC hints at how NBS scale pH measurements isn’t perfectly representative of the media’s pH. The high ion activity of the media influences the pH measurement which CO2SYS can’t properly correct, which leads to erroneously calculated DIC and bulk CO2 concentrations. The ionic activity of water, however, doesn’t impair pCO2 measurements and when using pCO2 and alkalinity as an input into CO2SYS, pH isn’t required to calculate carbonic acid speciation. Therefore, the error associated with pH is eliminated. Interestingly, however, when calculating the mass transfer coefficient using pH or pCO2 data, they yield eerily similar results of 0.06 m/hr and 0.05 m/hr, respectively (Figure 5-42, Lower Right). The mass transfer coefficient is the ratio between the flux and driving force, which happens to “cancel out” the error associated with the pH approach. Therefore, the pH approach to determine the mass transfer in the coefficient in the 22 m2 raceway is assumed to be justified due to this phenomenon. The mass transfer coefficient predicted from coulometry DIC, however, is significantly less (0.008 m/hr). This is due to the error associated with the coulometry measurements and how the coulometer isn’t sensitive enough to measure changes in DIC at this small of a scale. Final Report | MicroBio Engineering Inc. 67 Figure 5-42: Outgassing pH and pCO2 time series(Upper Left); pH and pCO2 predicted DIC time series compared to coulometry measured DIC (Upper Right); Driving force time series calculated from measured DIC, pH, and pCO2 (Lower Left); Flux versus driving force linear regression and resulting mass transfer coefficient calculated with pH, pCO2, or DIC data (Lower Right). 5.3.2.5.2 Injection Efficiency Injection efficiency was performed in both 3.4 m2-C and the 22 m2 raceways by injecting the raceways with CO2. For both 3.4 m2-C and the 22 m2 raceways, a high flowrate of 350 mL CO2/min was tested, however, a low flowrate of 175 mL CO2/min was additionally tested in the 22 m2. For the high flowrates, the pH was driven down from 9.0 to 8.1, whereas the low flowrate ranged from a pH of 8.85 to 8.35 due to time limitations. For every tenth of a decrease in pH, the change in DIC, mass of CO2 injected, outgassed carbon, and the resulting injection efficiency (Figure 5-43) was calculated. As shown in Figure 5-44, the increase in DIC was linear, as expected from the constant CO2 injection rate, however, the increase in DIC in 3.4 m2-C had more of an exponential shape towards the end of the injection period, which corresponds to the increase in injection efficiency with time. Why 3.4 m2-C had a non linear increase in DIC towards the end of the injection is unresolved. If it was a CO2SYS error from pH input, then the same phenomenom would be observed in the 22 m2 trials. For all three experiments, outgassed carbon had minimal impact on the injection efficiency as it only represents a small portion of carbon injected. Final Report | MicroBio Engineering Inc. 68 Figure 5-43: Injection efficiency results for the three trials calculated over each tenth decrease in pH. High and low flowrates correspond to 350 mL CO2/min and 175 mL CO2/min, respectively. Figure 5-44: Time series of carbon values throughout the injection period. “C injected” refers mass of carbon sparged into the raceway, “DIC Increase” represents the increase in DIC, and “C Outgassed” refers to the mass of injected carbon that has outgassed. For each of the three experiments, the total injection efficiency from the start to the end of the injection was calculated as 20% and 30% for the high flow rate in the 22 m2 and 3.4 m2-C, respectively. The low Final Report | MicroBio Engineering Inc. 69 flow rate in the 22 m2, however, produced a lower total injection efficiency of 14%. We would assume, however, that the injection efficiency across all three trials would be identical as the sparger type were identical and placed in similar locations (on the bottom of the raceway along the outer wall) and each raceway had the same surface velocity, water composition, depth, etc. Therefore, we assumed that 3.4C’s injection efficiency result was more representative of the injection efficiency for both flow rates in the 22 m2. The injection trials in the 22 m2 took several days whereas 3.4 m2-C only required half a day. The longer time as well as its larger volume, most likely introduced more errors into the 22 m2’s injection efficiency calculations. Additionally, when performing carbon mass balances across cultivation trial, using an injection efficiency value of 30% yielded more complete closures of the mass balances Therefore, injection efficiency was assumed to be 30% for both raceways regardless of the CO2 flowrate. 5.3.2.5.3 Carbon Utilization Efficiency, Mass Closure CUE and carbon mass-closure were calculated over six growth periods during the DAC cultivation trial (Figure 5-45). Two were excluded from analysis due to culture instability. In either pond, CUE regularly exceeded the assumed 30% injection efficiency, the theoretical maximum. Correspondingly, mass closure, for either raceway, consistently exceeded 100%, ranging from 120% to 170% for 3.4C and 110% to almost 300% in the 22 m2. Closures greater than 100% indicate the net carbon input is likely underestimated. Conversely, the net carbon increase in the raceway may be overestimated, however, this isn’t a suspected error as these values are less sensitive to measurement error. Therefore, it assumed that the main error is due inaccuracies in flowrate from the mass flow controller which can have flowrate errors upwards of 10% (FMA 5400A/FMA 5500A Mass Flow Controllers, n.d.). Figure 5-45: CUE and closure for the 3.4C and 22 m2 raceways and the corresponding period over which they were calculated. The black dashed line indicates the assumed injection efficiency (30%). Higher than theoretical CUE and poor mass closure (>200 %) complicate comparisons between the two ponds and modes of CO2 supply (3.4 C: on-demand injection based on pH; 22 m2: continuous, DAC- simulated CO2 delivery). Notwithstanding, CUE was comparable between the two modes of carbon supply. Final Report | MicroBio Engineering Inc. 70 In the 22 m2 raceway, CO2 was dosed at night, periods where no photosynthetic activity assimilates the injected carbon into biomass and expected to be lost. This phenomenon would lead to a poor CUE; however, CUE for the last three analyzed growth periods were near 30% (along with decent closures within 100 - 130%) indicating that carbon injected at night was successfully stored in the raceway media and eventually assimilated rather than outgassed. 3.4C similarly had CUE near 30%, indicating that the extra carbon injected throughout the night in the 22 m2 was ultimately assimilated in the biology. Choosing the continuous CO2 injection flowrate for the 22 m2 raceway also had impacts on its CUE. Flowrate was determined based on estimates of anticipated productivity and its corresponding carbon demand. Overestimated injection rates would lead to lower CUE and, conversely, underestimated rates would lead to higher CUE. For example, the carbon balance period in Figure 5-26, CUE in the 22 m2 was near 30%, indicating that the set CO2 flowrate was near the actual carbon demand, which was quite fortunate as injection efficiency was estimated to be 50% and expected future productivity was somewhat uncertain. CUE in 3.4 m2-C during that period also had a CUE near 30%; however, more CO2 was injected per unit surface area than in the 22 m2 to account for greater outgassing losses per unit surface area. Therefore continuous CO2 flow can produce high CUE with minimal outgassing losses, compared to the on demand approach which requires additional CO2 injection to account for outgassing losses. It is important, however, to make accurate predictions of anticipated productivity and carbon demand to determine injection rates that do not exceed the demand. 5.3.2.6 DAC biocompatibility trial The experiment aimed to compare the growth of P. cruentum in 35.80 media prepared with two different water sources: Azul well water and DAC unit condensate. While efforts were made to achieve uniform initial culture density, the Azul media flasks had an initial density of 140 mg VSS/L, compared to 70 mg VSS/L in the DAC unit condensate flasks. After eight days, the Azul media demonstrated a higher average final concentration of 600 mg VSS/L, compared to 430 mg VSS/L for the DAC media. Notably, despite having a higher initial concentration, the Azul media exhibited approximately 25% greater average productivity than the DAC media (Table 5-6). This difference is not significant at the 95% confidence interval in an unpaired t-test. We hypothesize that additional acclimation to the DAC amended media type may further bring the resulting productivities into alignment. While not conclusive that the DAC was has no significant effect, the results suggest that DAC condensate is not strongly inhibitory. Note that inhibitory compounds partition primarily to the gas phase would not necessarily have been accounted for in this trial. Table 5-6: Initial and starting densities and their corresponding productivity for all flasks. Flasks denoted with an asterisk indicate potential contamination in those flasks. Initial Density Final Density Productivity (mg Avg. Productivity ± Media – Flask # (mg VSS/L) (mg VSS/L) VSS/L-day) STD Azul – 1* 142.5 520.0 45.7 Azul – 2 142.5 646.7 61.0 54.8 ± 8.1 Azul – 3* 142.5 620.0 57.8 DAC – 1 69.2 413.3 41.7 DAC – 2 69.2 406.7 40.9 43.8 ± 4.4 DAC – 3* 69.2 473.3 48.9 Final Report | MicroBio Engineering Inc. 71 Some variability in productivity was observed among the triplicate flasks within each media type. The Azul media flasks had productivities ranging from 46 to 61 mg VSS/L-day while the DAC media flasks ranged from 41 to 49 mg VSS/L-day. The greater variability in the Azul media flasks may be attributed to uneven light distribution or potential biological contamination. In two of the three Azul media flasks and in one of the DAC media flasks, white spheres, suspected to be flocs (Figure 5-46), were observed, which could have contributed to variation in those flasks’ productivities. The exact composition of these flocs remains to be determined. Figure 5-46: Unknown flocs present in two of the Azul media flasks at the end of the growth trial. 5.3.3. Conclusion The simulated DAC cultivation trial successfully demonstrated cultivation of P. cruentum in a 22 m2 raceway with continuous CO2 supply, as would be the case with a direct air capture (DAC) system. The 22 m2 raceway was grown in tangent with a 3.4 m2 raceway that was dosed with on demand CO2 based on pH. CUE near 30% throughout the cultivation trial was observed in both raceways, suggesting that continuous CO2 supply is not detrimental to carbon utilization efficiency and the continuous supply case may actually minimize relative outgassing losses and injection requirements. However, it is necessary to make an informed decision on what the continuous CO2 flowrate should be. Overall, however, continuous CO2 supply allowed for the successful cultivation of P. cruentum. A major challenge throughout the cultivation trial was pest contamination. The presence of Picochlorum and predatory flagellates had a significant impact on P. cruentum’s productivity, and ultimately led to the end of the cultivation trial as no mitigation techniques had a significant impact. Not only did this effect harvest operations, but also the extent of water recycling as fresh media was required for re- inoculation. Disregarding re-inoculation events, up to 90% of growth media volume was successfully recycled between growth periods. Nutrient additions during media recycling remain a challenge as identifiers of nutrient deficiencies were observed. Overall, six successful biomass harvests occurred during the cultivation trial using the DAC simulation system, demonstrating the potential for success of microalgae cultivation DAC systems for bioproduct production. The ability to maintain high carbon utilization efficiency with continuous CO2 supply, and the potential for water savings through media recycling, highlight the economic and environmental benefits of this approach. Addressing the challenges associated with pest contamination and harvest optimization will be crucial for future commercial-scale implementation of integrated DAC and algae cultivation systems. 5.4. Techno-economic and Life-cycle Assessment Final Report | MicroBio Engineering Inc. 72 TEA modeling results for the DAC case project a ~46% increase in the algae biomass revenue potential when comparing the baseline to intermediate scenarios (Table 4 based on Table 5, also, Figure 32). This is assuming that the intermediate and final scenarios utilize a pond alkalinity of 50 meq/L (Table 5). Total life cycle global warming potential (GWP) weighted emissions have decreased by ~17% when comparing the baseline to intermediate scenarios (Figure 33). Figures 34 and 35 show the effect of alkalinity on production costs and GWP. Table 5-7: Summary of TEA Parameters Varied in the Operational (i.e., Uncertainty) Scenarios for the DAC Case. TEA Parameter Baseline (Low) Intermediate (Expected) Final (Best) CO2 Source Merchant DAC DAC Chlorella productivity (gmd) 9.2 9.2 10.4 CO2 transfer efficiency 50% 75% 75% Alkalinity (meq/L)* N/A 50 50 *Based on previous verification meeting, all scenarios were standardized at 50 meq/L Table 5-8: Summary of DAC Specific Parameters Varied in the Operational (i.e., Uncertainty) Scenarios TEA Parameter Baseline (Low) Intermediate (Expected) Final (Best) CO2 Supplied (Mgs/yr) 4,676 3,118 3,524 DAC Total CAPEX ($/Mg CO2) 145 92.5 40 DAC Total OPEX ($/Mg CO2) 101 77 52 Electricity ($/Mg CO2) 60 48 35 Natural Gas ($/Mg CO2) 11 10 9 Other ($/Mg CO2) 30 19 8 Table 5-9: Summary of TEA Modeling Results for the DAC Case TEA Result Baseline (Low) Intermediate (Expected) Final (Best) Biomass Production (Mgs/yr) 1,340 1,340 1,520 Plant Gate Production Costs ($2020/Mg) 6,688 3,821 3,343 Net Revenue Potential ($2020/yr) 3,332 6,179 6,657 Final Report | MicroBio Engineering Inc. 73 Figure 5-47: Comparison of Plant Gate Production Costs for Chlorella Based Nutraceutical Products across Baseline, Intermediate, and Final Operational Scenarios. Figure 5-48: Comparison of Life Cycle GWP Weighted Emissions for Chlorella Based Nutraceutical Products across Baseline, Intermediate, and Final Operational Scenarios (50 meq/L case depicted). Final Report | MicroBio Engineering Inc. 74 Figure 5-49: Effect of Increasing Alkalinity on Total Plant Gate Production Costs across Baseline, Intermediate, and Final Operational Scenarios. Figure 5-50: Effect of Increasing Alkalinity on Total GWP Emissions across Baseline, Intermediate, and Final Operational Scenarios. Final Report | MicroBio Engineering Inc. 75 Conclusions and Recommendations This project, "Microalgae Commodities Production with a Direct Air Capture Process", advanced the technology required to decouple large-scale microalgae biomass production from enriched sources of CO2, a critical limitation for expanding the national resource potential for algal biomass. The work addressed two primary pathways for supplying air-CO2 to cultivation systems: a physical-chemical Direct Air Capture (DAC) process and a direct uptake process utilizing the metabolic capacity of the algae itself. Key Findings and Achievements Direct Air Capture (DAC) Feasibility and Cost Reduction The project achieved a near-term small-scale demonstration of a DAC unit's technological feasibility. The DAC Pilot Unit, designed, constructed, and operated for a 2,000-hour operational campaign, successfully produced CO2 with a product purity of over 90% (v/v) during the trial. • Cost Projection: Operational data from the DAC Pilot Unit informed projections for a full-scale system, indicating a cost of CO2 at $150/Mg. This projection represents a significant reduction compared to non-DAC merchant CO2 (estimated at ~$1,000/Mg). • Techno-Economic Analysis (TEA) and Life-Cycle Analysis (LCA): The final TEA/LCA confirmed the value of the DAC-based scenario by projecting a future 27% increase in algae biomass revenue potential under optimized conditions. It also demonstrated an environmental benefit, projecting a 22% reduction in greenhouse gas emissions compared to the chemically enhanced air-CO2 transfer approach. Alkaliphilic Strain Performance The project successfully demonstrated the cultivation of alkaliphilic Chlorella isolate CT3072 under conditions amenable to high CUE. • Biomass Productivity: An annual average biomass productivity of 9.3 g AFDW/m²-day was achieved for Chlorella isolate CT3072 under conditions (200 meq/L alkalinity, pH 9.6 setpoint) that projected a high Carbon Utilization Efficiency (CUE) of 90%. Proof-of-Concept for Direct Uptake Exploratory research was conducted to establish a proof-of-concept for biochemically enhanced air-CO2 transfer. The work established the framework to quantify enhanced air-CO2 transfer using the enzyme carbonic anhydrase and validated an assay to measure the protein concentration increase in the culture media. Recommendations The results provide a strong technological foundation for utilizing DAC-derived CO2 for microalgae commodity production. The following areas are recommended for future work to transition the technology from the R&D phase to commercial viability: Final Report | MicroBio Engineering Inc. 76 1. Scale-Up and Integration: Proceed with a demonstration-scale project to fully integrate the DAC unit with a microalgae cultivation facility, using the optimized operational data to validate the projected $150/Mg CO2 cost in a continuous, outdoor system. 2. Strain and Cultivation Optimization: Focus further research on the high-performing Chlorella isolate CT3072 to ensure culture stability and develop robust, cost-effective crop protection strategies, especially under high-CUE, high-alkalinity conditions. 3. Direct Uptake Continuation: Continue the proof-of-concept for biochemically enhanced air-CO2 transfer by advancing the Western Blot method for detecting carbonic anhydrase and assessing the full enhancement potential of extracted and excreted bioactive compounds. Refined TEA/LCA: Incorporate detailed costs of crop protection strategies and long-term alkalinity management into the TEA/LCA to fully capture production costs and validate the projected 27% increase in revenue potential. Final Report | MicroBio Engineering Inc. 77 Products developed 7.1. Peer reviewed papers Gao, S., Crowe, B., O’Kelly, C., Edmundson, S., Benemann, J., Huesemann, M. Improving biomass productivity of the alkaliphilic microalga Graesiella emersonii. In preparation. 7.2. Conference presentations Oral Presentation at ABLC Next October 2024, San Francisco, CA, USA “Algae, Wastewater Treatment and SAF.” Presented by John Benemann. October 2024. DOI: https://doi.org/10.2172/2507043 OSTI ID:2507043 Oral Presentation at the TechConnectWorld Conference 2023, National Harbor, MD, USA “Decarbonization of Wastewater Treatment with Microalgae Processes.” Presented by Juergen Polle. June 2023. DOI: https://doi.org/10.2172/1985913 OSTI ID:1985913 Works Cited Albalasmeh, A. A., Berhe, A. A., & Ghezzehei, T. A. (2013). A new method for rapid determination of carbohydrate and total carbon concentrations using UV spectrophotometry. Carbohydrate Polymers, 97(2), 253–261. https://doi.org/10.1016/j.carbpol.2013.04.072 Brown, M. R., Mular, M., Miller, I., Farmer, C., & Trenerry, C. (1999). The vitamin content of microalgae used in aquaculture. Journal of Applied Phycology, 11(3), 247–255. https://doi.org/10.1023/A:1008075903578 Cerri, R., Niccolai, A., Cardinaletti, G., Tulli, F., Mina, F., Daniso, E., Bongiorno, T., Chini Zittelli, G., Biondi, N., Tredici, M. R., & Tibaldi, E. (2021). Chemical composition and apparent digestibility of a panel of dried microalgae and cyanobacteria biomasses in rainbow trout (Oncorhynchus mykiss). Aquaculture, 544, 737075. https://doi.org/10.1016/j.aquaculture.2021.737075 CIMIS. (n.d.). Retrieved January 17, 2025, from https://cimis.water.ca.gov/Default.aspx Coleman JR, Berry JA, Togasaki RK, Grossman AR. Identification of Extracellular Carbonic Anhydrase of Chlamydomonas reinhardtii. Plant Physiol. 1984 Oct;76(2):472‐7. Davis, R. E., Markham, J. N., Kinchin, C. M., Canter, C., Han, J., Li, Q., ... & Zhu, Y. (2018). 2017 Algae harmonization study: evaluating the potential for future algal biofuel costs, sustainability, and resource assessment from harmonized modeling (No. NREL/TP‐5100‐70715). National Renewable Energy Laboratory (NREL), Golden, CO (United States). De Luca V, Del Prete S, Supuran CT, Capasso C. Protonography, a new technique for the analysis of carbonic anhydrase activity. J Enzyme Inhib Med Chem. 2015 Apr;30(2):277‐82. Di Lena, G., Casini, I., Lucarini, M., Sanchez del Pulgar, J., Aguzzi, A., Caproni, R., Gabrielli, P., & Lombardi‐ Boccia, G. (2020). Chemical characterization and nutritional evaluation of microalgal biomass from large‐scale production: A comparative study of five species. European Food Research and Technology, 246(2), 323–332. https://doi.org/10.1007/s00217‐019‐03346‐5 Ebeling, J. M., Timmons, M. B., & Bisogni, J. J. (2006). Engineering analysis of the stoichiometry of photoautotrophic, autotrophic, and heterotrophic removal of ammonia–nitrogen in aquaculture systems. Aquaculture, 257(1‐4), 346‐358. Final Report | MicroBio Engineering Inc. 78 Fithriani, D., & Melanie, S. (2022). Vitamin and mineral content of microalgae Phorpyridium and Chlorella and development prospects for food raw materials in Indonesia. IOP Conference Series: Earth and Environmental Science, 1034(1), 012043. https://doi.org/10.1088/1755‐1315/1034/1/012043 FMA 5400A/FMA 5500A Mass Flow Controllers. (n.d.). Retrieved October 25, 2024, from https://assets.omega.com/manuals/test‐and‐measurement‐equipment/flow/mass‐flow‐ meters/M5372.pdf Frankignoulle, M., Borges, A., & Biondo, R. (2001). A new design of equilibrator to monitor carbon dioxide in highly dynamic and turbid environments. Water Research, 35(5), 1344–1347. https://doi.org/10.1016/S0043‐1354(00)00369‐9 Fujiwara S, Fukuzawa H, Tachiki A, Miyachi A (1990) Structure and differential expression of two genes encoding carbonic anhydrase in Chlamydomonas reinhardtii. Proc Natl Acad Sci USA 24: 9779–9783 He, Y., Ma, M., Hu, Q., & Gong, Y. (2021). Assessment of NH4HCO3 for the control of the predator flagellate Poterioochromonas malhamensis in pilot‐scale culture of Chlorella sorokiniana. Algal Research, 60, 102481. https://doi.org/10.1016/j.algal.2021.102481 Huesemann, M. H., Knoshaug, E. P., Laurens, L. M., Dale, T., Lane, T. W., & McGowen, J. (2024). Development of integrated screening, cultivar optimization, and verification research (DISCOVR): A coordinated research‐driven approach to improve microalgal productivity, composition, and culture stability for commercially viable biofuels production. Algal Research, 79, 102961. Husic, D. et al., Identification of Intracellular Carbonic Anhydrase in Chlamydomonas reinhardtii which Is Distinct from the Periplasmic Form of the Enzyme, Plant Physiology, Volume 89, Issue 3, March 1989, Pages 904–909 Kavitha, M. D., Seema Shree, M. H., Vidyashankar, S., & Sarada, R. (2016). Acute and subchronic safety assessment of Porphyridium purpureum biomass in the rat model. Journal of Applied Phycology, 28(2), 1071–1083. https://doi.org/10.1007/s10811‐015‐0655‐9 Kimpel, D. L., Togasaki, R. K., & Miyachi, S. (1983). Carbonic anhydrase in Chlamydomonas reinhardtii I. Localization. Plant and cell physiology, 24(2), 255‐259. Matthews, B. J. (1999). The Rate of Air‐Sea CO2 Exchange: Chemical Enhancement and Catalysis by Marine Microalgae. Ph. D. Thesis, University of East Anglia. Millero, F. J. (1979). The thermodynamics of the carbonate system in seawater. Geochimica et Cosmochimica Acta, 43(10), 1651–1661. https://doi.org/10.1016/0016‐7037(79)90184‐4 Rebolloso Fuentes, M. (2000). Biomass nutrient profiles of the microalga Porphyridium cruentum. Food Chemistry, 70(3), 345–353. https://doi.org/10.1016/S0308‐8146(00)00101‐1 Sharp, J. (2024). CO2SYSv3 for MATLAB [MATLAB]. https://github.com/jonathansharp/CO2‐System‐Extd (Original work published 2019) U.S. Department of Energy. 2016. 2016 Billion-Ton Report: Advancing Domestic Resources for a Thriving Bioeconomy, Volume 1: Economic Availability of Feedstocks. M. H. Langholtz, B.J. Stokes, and L.M. Eaton (Leads), ORNL/TM‐2016/160. Oak Ridge National Laboratory, Oak Ridge, TN. 448p. doi: 10.2172/1271651. http://energy.gov/eere/bioenergy/2016‐billion‐ton‐report. Venteris, E. R., McBride, R. C., Coleman, A. M., Skaggs, R. L., & Wigmosta, M. S. (2014). Siting algae cultivation facilities for biofuel production in the United States: trade‐offs between growth rate, site constructability, water availability, and infrastructure. Environmental science & technology, 48(6), 3559‐3566. Final Report | MicroBio Engineering Inc. 79 Xinyi E, Czarena Crofcheck, Mark Crocker; Application of recycled media and algae‐based anaerobic digestate in Scenedesmus cultivation. Journal of Renewable and Sustainable Energy 1 January 2016; 8 (1): 013116. Ynalvez RA, Xiao Y, Ward AS, Cunnusamy K, Moroney JV. Identification and characterization of two closely related beta‐carbonic anhydrases from Chlamydomonas reinhardtii. Physiol Plant. 2008 May;133(1):15‐26. Final Report | MicroBio Engineering Inc. 80