ENHANCING CARBON EFFICIENCY OF ANAEROBIC DIGESTION THROUGH FORMATE METHANOGENESIS By John Grivins A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Biosystems Engineering – Master of Science 2025 ABSTRACT Anaerobic digestion is an important technology for waste treatment and renewable energy production. To pursue avenues of improvement in digester technology and foster a circular bioeconomy, this study investigated the impact of activated carbon, formic acid, and sodium bicarbonate addition on the performance, parameters, and ecology of bench-scale anaerobic reactors. The study found that the addition of formate in combination with activated carbon was able to significantly increase the production of biogas from digestate collected from the South Campus Anaerobic Digester (SCAD). Reactors treated with formic acid also showed evidence of superior formic acid utilization, with formic treated reactors having similar or lesser formic acid concentration than non-formic treated reactors. Additionally, formic acid treated reactors showed lower concentrations of volatile fatty acids (VFAs), including propionic and butyric acid, indicating enhanced reactor performance. DNA analysis revealed an increase in formate-scavenging methanogens and syntrophic bacteria. Thes analyses suggests that the formate treatment led to enhanced syntrophic interactions among the microbial consortia. Following these promising results, a life-cycle analysis (LCA) and were conducted on a hypothetical two-stage treatment system utilizing electrocatalysis to convert waste CO2 into formic acid for treatment of a secondary digester. The mass balance and LCA suggest that formate-treated AD has the potential to greatly enhance energy eWiciency, carbon utilization, and environmental outcomes of the AD process. To Jennifer and Petra Grivins I couldn’t have made it without you iii ACKNOWLEDGEMENTS This thesis was researched and written at the Department of Biosystems and Agricultural Engineering of Michigan State University. I would like to thank the BAE Department for being my home and family during my time and MSU, and for providing me with so many great friends, classes, teachers, and mentors. Without the support of our department’s scholarly community, I would never have become the person and the engineer that I am today. My time here will always be precious to me, and I will never forget how welcomed I felt here. I would like to thank Dr. Wei Liao, my major professor, for guiding me through my master’s program. Your guidance and tireless spirit (something) Thank you to my committee members, Dr. Yan Susie Liu and Dr. James Wallace. Thank you to Jim Wallace, Bob Hickey, Mike Hickey, and everyone else at SustainRNG for providing funding for the majority of my master’s program. I would like to thank Dr. Sibel Uludag-Demirer for advocating for my hiring to the research group, and for being a great manager and mentor during my time working at ADREC. Thank you to my fellow grad students, Meicai Xu and Emelia Emerson, for providing guidance on certain procedures and assistance with data analysis for this thesis. Thank you to the experts at the MSU Genomics Core and the MSU NMR Laboratory for lending technical services that made the gathering of critical data possible. iv Thank you to my parents, Petra and Jennifer Grivins, for your causal relationship to my existence and following development. It was your support and inspiration that most of all made these accomplishments possible. v TABLE OF CONTENTS LIST OF ABBREVIATIONS ............................................................................................... viii CHAPTER 1: INTRODUCTION .......................................................................................... 1 1.1 Significance ........................................................................................................ 1 1.2 Goals and Objectives .......................................................................................... 3 1.3 LITERATURE REVIEW ........................................................................................... 4 1.3.1 Anaerobic Digestion ................................................................................. 4 1.3.2 Formate Addition ...................................................................................... 5 1.3.3 Bicarbonate Addition ................................................................................ 7 1.3.4 Activated Carbon Addition ........................................................................ 7 1.3.5 Catalytic Hydrogenation of CO2 ................................................................. 8 1.3.6 Other Applications of Formic Acid ............................................................. 8 1.3.7 Life Cycle Assessment .............................................................................. 8 CHAPTER 2: MATERIALS AND METHODS ....................................................................... 10 2.1 Bench-Scale Formate Digesters ......................................................................... 10 2.2 Bench-Scale Analysis ........................................................................................ 12 2.2.1 Total and Volatile Solids (TS/VS) .............................................................. 12 2.2.2 Chemical Oxygen Demand ..................................................................... 12 2.2.3 pH ......................................................................................................... 13 2.3 Gas Production and Quality ............................................................................... 13 2.4 VFA Analysis ..................................................................................................... 15 2.5 DNA Analysis .................................................................................................... 16 2.6 Statistical Analysis ............................................................................................ 16 2.7 Mass And Energy Balance Analysis ..................................................................... 18 2.8 Life Cycle Impact Assessment (LCIA) ................................................................. 19 CHAPTER 3: FORMATE-ENHANCED AD ......................................................................... 21 3.1 Gas Production ................................................................................................. 21 3.2 Gas Quality ....................................................................................................... 24 3.3 pH, TSVS, and COD ........................................................................................... 25 3.3.1 Chemical Oxygen Demand ..................................................................... 25 3.3.2 pH ......................................................................................................... 26 3.3.3 Total and Volatile Solids .......................................................................... 28 3.4 Volatile Fatty Acids ............................................................................................ 31 3.5 Formic Acid ....................................................................................................... 31 3.6 Acetic Acid, Propionic Acid, Butyric Acid, Valeric Acid ......................................... 32 3.6.1 Acetic Acid ............................................................................................. 32 3.6.2 Propionic Acid ........................................................................................ 34 3.6.3 Butyric and Isobutyric Acid ..................................................................... 35 3.6.4 Isovaleric Acid ........................................................................................ 37 3.7 Microbial DNA ................................................................................................... 38 vi 3.7.1 Alpha Diversity Shannon, Simpson, and Sobs Indices .............................. 38 3.7.2 Beta Diversity ......................................................................................... 40 3.7.3 Bacterial Community Analysis ................................................................ 43 3.7.4 Archaeal Community Analysis ................................................................ 45 CHAPTER 4: LIFE CYCLE ASSESSMENT .......................................................................... 47 4.1 Introduction ...................................................................................................... 47 4.1.1 Studied System ...................................................................................... 47 4.1.2 Goal and Scope ...................................................................................... 49 4.2 Mass and Energy Balance .................................................................................. 50 4.3 Life Cycle Impact Assessment ........................................................................... 56 4.4 Carbon EWiciency .............................................................................................. 59 CHAPTER 5: CONCLUSIONS ......................................................................................... 61 5.1 Findings ............................................................................................................ 61 5.2 Recommendations for Future Work .................................................................... 62 WORKS CITED .............................................................................................................. 65 APPENDIX A: DETAILED STATISTICAL RESULTS ............................................................... 69 APPENDIX B: PYTHON AND R CODE .............................................................................. 74 APPENDIX C: DETAILED DATA ....................................................................................... 80 APPENDIX D: SUPPLEMENTARY FIGURES ...................................................................... 95 vii LIST OF ABBREVIATIONS AD Anaerobic Digestion ADREC Anaerobic Digestion Research and Education Center CH4 CHP CO2 Methane Combined Heat and Power Carbon Dioxide CO2-e Carbon Dioxide Equivalent COD Chemical Oxygen Demand CSTR Continuous Stirred Tank Reactor DI Deionized DIET Direct Interspecies Electron Transfer FU Functional Unit GHG Greenhouse Gas GWP Global Warming Potential IET HRT H2S LCA LCI LCIA MSU Interspecies Electron Transfer Hydraulic Retention Time Hydrogen Sulfide Life Cycle Assessment Life Cycle Inventory Life Cycle Impact Assessment Michigan State University NMDS Non-Metric Multidimensional Scale viii SCAD South Campus Anaerobic Digester TEA TS VFA VS Techno-Economic Analysis Total Solids Volatile Fatty Acid Volatile Solids WEP Water Eutrophication Potential ix CHAPTER 1: INTRODUCTION 1.1 Significance Anaerobic Digestion (AD) is a waste treatment, energy generation, and waste-to- resources technology implemented at many dairy farms and water treatment facilities. AD takes in organic wastes, particularly animal manure and food waste, and converts their organic solids into biogas, a mixture of methane (CH4) and carbon dioxide (CO2). This provides a dual benefit of treating organic waste and producing renewable natural gas (RNG). As of June 2024, there were 400 manure-based AD systems operating in the United States (US EPA, 2024). In 2023, US manure-based AD operations avoided 14.8 million tonnes CO2-e GHG emissions and generated 3.9 million MWh energy equivalent. However, the technology still faces barriers to a wider potential implementation. Although there were 400 operations in the US, the EPA AgSTAR program estimated that in June 2024 there were 8,000 large dairy and hog operations where an AD operation would be technically viable (US EPA, 2024), meaning that just 5% of the nation’s potential manure-based operations were active. One barrier to the wider implementation of AD operations is the lack of economic viability for electricity and RNG for many small to medium operations. The natural gas market is often unavailable to small and medium facilities; either due to the cost of biogas upgrading systems, which are prohibitively expensive for their smaller scale; or other complications, particularly in places like the US where there is a lack of incentives and support for RNG markets (Edwards et al., 2015). For this reason, it is often the best option for AD operations to generate electricity, using it first to oWset their own energy costs before selling excess to the grid (Hjort-Gregersen et al., 2011). Even with this approach, the 1 payback period for new digester operations based on sale of electricity to the grid remains lengthy. It is therefore desirable to create innovations in AD operation technology that can increase economic viability, and therefore implementation, through an increase in electrical yields or the production of valuable co-products. A potential improvement in AD is greater potential GHG emission oWsets and carbon utilization. The CO2 released from AD biogas activities is biogenic; because it’s produced from plant material, it will be recaptured in the next crop growing cycle. Therefore, this gas is part of the short-term carbon cycle and isn’t considered as contributing to GHG emissions. However, if the CO2 were to be captured and reused, improving its carbon eWiciency, the AD operation could potentially reach net negative carbon emissions by replacing additional fossil gas usage without contributing to GHG emissions. One use for CO2 is conversion into formic acid via electrocatalysis, which is both a valuable co-product of digestion and a promising candidate for digester addition. A relatively new avenue of anaerobic digestion research is the eWect of formate and bicarbonate addition on digester microbial cultures. Formic acid has been shown to have a positive impact on digester performance in terms of biogas yield (Li et al., 2024). This is thought to occur through the enhancement of interspecies electron transfer (IET), which improves the consumption of intermediary products, specifically volatile fatty acids (VFAs) and enhances methanation. Combining the concepts of CO2 to formic acid conversion and formic acid enhancement of AD performance, it becomes clear that a study to evaluate the impact of formate addition and the potential of an operation which converts waste CO2 to 2 formic acid for digester treatment is valuable to the improvement of the environmental, technical, and economic viability of AD as a technology. 1.2 Goals and Objectives The primary aim of this thesis is to assess a potential system for recycling CO₂ through catalytic conversion to formate. This innovative approach seeks to enhance the eWiciency and sustainability of anaerobic digestion (AD) processes. The specific objectives supporting this goal are: • Evaluate the impact of formate and bicarbonate addition on lab-scale bioreactors: This involves conducting controlled experiments to determine how these additives influence the performance and eWiciency of AD systems. The focus will be on key metrics such as biogas production, volatile fatty acid (VFA) concentrations, and microbial community dynamics. • Design a process flow for a potential formate-enhanced anaerobic digestion (AD) biorefinery: This objective aims to create a detailed blueprint for integrating formate enhancement into existing AD systems. The process flow will outline the necessary steps, equipment, and operational parameters required to implement this technology at a larger scale. • Conduct a life-cycle assessment of the proposed formate-enhanced AD biorefinery: This involves evaluating the environmental impact of the proposed biorefinery. The life-cycle assessment will provide a comprehensive evaluation of the environmental impacts from cradle to grave. 3 1.3 LITERATURE REVIEW 1.3.1 Anaerobic Digestion Anaerobic Digestion (AD) is a well-characterized method of waste treatment and energy production that has been in use for decades. It is a process in which organic wastes are converted into biogas through the cooperation of diWerent microbes using syntrophy. Syntrophy is the process by which the products of one microbe’s metabolism are used as the nutrients of another (Marietou, 2021). There are four steps involved in AD: hydrolysis, acidogenesis, acetogenesis, and methanogenesis. Hydrolysis is the process by which large molecules are broken down into smaller components that may be accessed by other microorganisms. Hydrolytic bacteria convert carbohydrates into sugars, lipids into long-chain fatty acids, and proteins into amino acids. Hydrolysis can be a rate-limiting step in the AD process (Meegoda et al., 2018). The products of hydrolysis can be diWused through the membranes of acidogenic microbes in the second step of AD, acidogenesis. These microbes produce volatile fatty acids (VFAs) such as acetic, formic, propionic, and butyric acid. Acetic and formic acid can be utilized directly by methanogens. However, butyric and propionic acid must be converted into acetic acid through acetogenesis (F. Shen et al., 2018). Acidogenesis has been observed to be the fastest step in AD. It is important to consider this fast rate with regards to reactor stability, as VFA acidification is a common cause of reactor failure (Meegoda et al., 2018). VFA acidification is the accumulation of fatty acids in the digester to an extreme extent that causes the metabolism of microbes in the environment to fail. 4 Acetogenesis is the process of converting larger VFAs into acetate. (Meegoda et al., 2018) Larger VFAs in this case means VFAs with a longer carbon chain than acetate (i.e. propionate, butyrate, and valerate). This step is considered rate-limiting in the overall process due to the high Gibbs free energy of VFA oxidation, especially propionate (Mu et al., 2023). To become favorable, the propionate reaction must be coupled with a reaction that consumes hydrogen. Therefore, a successful reactor should pair propionate degradation with the conversion of H2 to CH4, aiming to keep the partial pressure of H2 in the reactor below 10-4 atm (Mu et al., 2023). Methanogenesis is the final step of AD, where biogas is produced by a class of obligate anaerobic archaea. These microbes can produce CH4 primarily from acetate or hydrogen (Meegoda et al., 2018). Formate has also been observed as a possible source for methanogenesis (Belay et al., 1986). 1.3.2 Formate Addition A major challenge in the process of anaerobic digestion is that the oxidation of fatty acids to acetate, hydrogen, and formate, i.e., acetogenesis, is thermodynamically unfavorable unless the concentration of those products is extremely low. However, the overall process of digestion can be made thermodynamically favorable due to the process of interspecies electron transfer (IET). The metabolism of fermentative organisms eWectively produces surplus electrons due to reduction; meanwhile, methanogens require additional electrons and thus serve as the electron sink (L. Shen et al., 2016). To accomplish this, H2 and formate serve as electron shuttles. An organism can release its electrons for use by other species using enzymes called hydrogenases, if using hydrogen, 5 or formate dehydrogenases, if using formate as the carrier. These enzymes take protons or CO2, along with electrons, and convert them to H2 and formate respectively. These compounds are then transferred to methanogenic organisms, which metabolize them to produce CH4 (L. Shen et al., 2016). Using IET to couple these reactions results in an overall set of thermodynamically favorable reactions. Table 1.3.1 summarizes the reactions involved in AD and their respective Gibb’s Free Energy values. Table 1.3.1. Anaerobic Digestion Substrate Reactions (Zhang et al., 2023) Substrate Propionate Reaction DG°’(kJ/mol) CH3CH2COOH + H2O ® 2CH3COOH + 3H2 + CO2 Butyrate CH3CH2CH2COOH + 2H2O ® 2CH3COOH + 2H2 Lactate CH3CH2OH + H2O ® CH3COOH + 2H2 Ethanol CH3CH2OH + H2O ® CH3COOH + 2H2 Formate 4HCOOH ® CH4 + 3CO2 + Acetate Hydrogen 2H2O CH3COOH ® CH4 +CO2 4H2 + CO2 -> CH4 + 2H2O +76.1 +48.1 -4.2 +9.6 -130 -33 -135 6 Research suggests that reactors with formate added demonstrate increased CH4 production. At concentrations ranging from 5 mM to 30 mM of formate, CH4 yield from propionate was shown to be enhanced (Li et al., 2024). 1.3.3 Bicarbonate Addition Food waste has a low C/N ratio compared to other digester feedstocks (Akindele & Sartaj, 2018). This can result in the rapid buildup of VFAs in digesters where food waste is a significant component (Gao et al., 2020). Because digesters require both neutral and stable pH due to the sensitivity of culture microbes, rapid VFA accumulation can cause digester failure (Deublein & Steinhauser, 2011). In order to counter fluctuating pH due to changes in VFA concentration, digesters require a source of alkalinity to serve as a pH buWer (Valença et al., 2021). Sodium bicarbonate (NaHCO3) has been identified as an eWective buWer in AD cultures, improving biogas quality, culture stability, and CH4 yield (Valença et al., 2021). In addition to its buWering capability, bicarbonate is also a possible carbon source. It has been suggested that bicarbonate may be used as an external carbon source for biomass growth (Mokashi et al., 2016). This could make bicarbonate an eWective carbon capture route through carbon storage via biomass. 1.3.4 Activated Carbon Addition Activated carbon (AC) is a conductive material that has been shown to enhance biogas yields in AD (Tiwari et al., 2021). Conductive materials have been suggested to facilitate direct interspecies electron transfer (DIET) (Feng et al., 2023). The addition of carbonaceous materials has also been shown to shorten or eliminate reactor start-up times (Xu et al., 2021). 7 1.3.5 Catalytic Hydrogenation of CO2 One reason that formate is interesting as a digester additive is that it can be made from CO2. Using a catalyst, CO2 may be hydrogenated to formic acid. In the general CO2 to formate pathway, CO2 is adsorbed on the cathode of the catalyst surface. This forms a radical CO2 intermediate, which is a negative CO2 ion. The exact intermediate step that follows depends on whether CO2 bound to the catalyst with a carbon atom or an oxygen atom. Whether the reduction product that follows this step is protonated depends on the pH of the reaction medium. Finally, formic acid is formed via desorption (Ma et al., 2024). Commonly, this is accomplished using an acidic solution to put the CO2 into an aqueous phase, and a metal catalyst, such as ruthenium or rhodium, to catalyze the reaction (Moret et al., 2014). 1.3.6 Other Applications of Formic Acid Even in the case that formic acid/formate proves non-useful as an AD enhancer, it is still a useful chemical for industrial applications and renewable energy systems. Notably, formic acid is dense in hydrogen, with an H2 density of 53 g L-1 (Joó, 2008). This property, which is the primary reason formate is useful for methanation, also makes it a promising candidate for hydrogen storage in fuel cells. Formic acid has the added benefit of being a non-volatile liquid at ambient conditions. This mitigates many of the risks associated with the use, storage, and transportation of H2 gas. 1.3.7 Life Cycle Assessment Life cycle assessment (LCA) evaluates a product or process’s environmental impact based on its entire life cycle. It includes such factors as manufacturing, distribution, use, 8 reuse, raw material extraction, recycling, and product or process end-of-life. LCA integrates the management of the framework, impact assessment, and data quality. The term “cradle- to-grave”, often used to describe LCA, refers to this complete accounting of all steps in the product or process’s lifetime (Dwilaksono, 2022, Odey et al., 2021). This more holistic method of environmental assessment, where all steps are considered together, makes LCA more robust compared to other metrics of environmental impact, such as water footprint or carbon footprint, which consider only a single dimension of environmental impact. LCAs are able to consider trade-oWs between diWerent steps and factors, which isn’t possible in less holistic methods of evaluation (Odey et al., 2021). The United Nations Environment Programme (UNEP), Society of Environmental Toxicology and Chemistry (SETAC) Life Cycle Initiative, International Reference Life Cycle Database System (ILCD), and others have promoted the use of LCA in pursuit of economically viable and sustainable societies (Odey et al., 2021). LCA is divided into four phases: goal and scope definition, life cycle inventory analysis, life cycle impact assessment, and life cycle interpretation (Reap et al., 2008). 9 CHAPTER 2: MATERIALS AND METHODS 2.1 Bench-Scale Formate Digesters To evaluate the eWect of formate and bicarbonate addition on a potential digester, a bench-scale experiment was conducted. Five conditions were tested, using digestate from SCAD as a feedstock. Each condition was tested with a biological duplicate. An overview of the operating conditions is shown in Table 2.1.1. Table 2.1. Experimental Conditions Reactor Digestate Control (C) 40 mL Activated Carbon (A) Bicarbonate (B) 40 mL 40 mL Formate (F) 40 mL Bicarbonate + Formate (CF) 40 mL Activated Carbon No 4% w/v at start- up 4% w/v at start- up 4% w/v at start- up 4% w/v at start- up Bicarbonate Formate No (5 mL DI water) No (5 mL DI water) 5 mL solution, total conc. 0.5 g/L No (5 mL DI water) 5 mL solution, total conc. 0.5 g/L No (5 mL DI water) No (5 mL DI water) No (5 mL DI water) 5 mL solution, total conc. 0.1 g/L 5 mL solution, total conc. 0.1 g/L Reactors were fed Monday, Wednesday, and Friday. The working volume of the reactors was 500 mL. Feedstock was a mixture of 40 mL of digestate, 5 mL of formate solution, and 5 mL of DI (deionized) water for the Formate reactors; 40 mL of digestate 5 mL of bicarbonate solution, and 5 mL of DI water for the Bicarbonate reactors; 40 mL of digestate, 5 mL of bicarbonate solution, and 5 mL of formate solution for the Bicarbonate/Formate reactors; and 40 mL of digestate and 10 mL of DI water for the 10 Control and Activated Carbon reactors. Figure 2.1.1 presents a graphical representation of these feeding conditions. The formate solution was prepared using a stock solution of formic acid that was diluted to a 1 g/L concentration. The bicarbonate solution was prepared using dry sodium bicarbonate stirred into DI water. Reactors that were treated with activated carbon were given 20 g (4% w/v) activated carbon powder at the beginning of the experiment and were not treated further with activated carbon. The reactors were operated with an HRT of 20 days for 4 HRTs, totaling 80 days. Reactors were operated at 50 C, which is in the thermophilic range. This range was chosen to reduce start-up time, and because thermophilic digestion works well at low solids content (Yu et al., 2017). Relative to a digester fed directly with manure and food waste, these digestate-fed digesters had low solids content. Table 2.1.2 lists key properties of the digestate used to feed the reactors. Composition of the digestate was measured on 5/17/2021. Table 2.2. Properties of Feedstock Digestate Measurement Value Units Total Solids Volatile Solids COD pH 40.81 g/L 29.29 g/L 50800 mg/L 8.22 - 11 2.2 Bench-Scale Analysis 2.2.1 Total and Volatile Solids (TS/VS) Total and Volatile Solid (TS/VS) analysis was performed to give information on the organic and inorganic solid content and moisture content of samples. TS/VS analysis was based on Hach Methods 8276 and 8271. Drying time for TS samples was increased to overnight, rather than 6 hours. Furnace time for VS samples was increased to 6 hours, rather than one hour. These changes help to ensure complete drying and combustion considering the higher solid content of digestate as compared to wastewater, which the methods were developed for. Samples were mixed via manual shaking. Materials for this test included digestate samples taken from the reactors (approximately 10 mL per test), ceramic crucibles or glass beakers (2 per test), analytical balance, oven, furnace, desiccator, desiccant, 10 mL micropipette, 10 mL micropipette tips, and a marker and whiteboard (for labeling purposes). 2.2.2 Chemical Oxygen Demand Chemical Oxygen Demand (COD) is used as a measure of organic pollutants in wastewater and is equivalent to the amount of oxygen needed to reduce compounds within the sample. Higher COD corresponds to a higher concentration of organics. COD testing was performed according to Hach method 8000. Sample COD concentrations were consistently high enough that they required dilution to fall within the test range. Materials for COD testing included digestate samples, DI water, heated reactor, HACH spectrophotometer 5000, HACH COD test vials, delicate wipes, 1 mL micropipette and tips, and markers. 12 2.2.3 pH Correct pH is critical for digester health and proper operation. pH values close to neutral are best, with the ideal range lying between 6.4 and 8.2. The pH probe used was periodically calibrated using standards with pH values of 4.0, 7.0, and 10.0. The probe was rinsed with DI water and cleaned with a delicate wipe after calibration and between each use. The pH probe was left in a storage solution when not in use. Materials and equipment for pH measurement were digestate samples, the pH meter and probe, DI water, calibration standard solutions at pH 4.0, 7.0, and 10.0, delicate wipes, and beakers. 2.3 Gas Production and Quality A modified shaker system, depicted in Figure 2.3.1, was used to maintain reactor temperature and monitor biogas production. In this gas collection system, each reactor is connected via two needles and a tube to a water bottle. This water bottle is then connected to an empty bottle, which is open to the atmosphere, via a tube open on one end and with a long needle on the other. As biogas is produced in the reactor, it pushes down on the water in the water bottle. This causes the water to be displaced and flow into the empty bottle at the same volumetric rate that gas is produced. Gas production was measured by measuring the displaced water volume with a graduated cylinder. 13 Figure 2.3.1. Biogas Volume Measurement System The required materials for this measurement were two bottles, two caps and rubber stoppers, two short needles, one long needles and two tubes per reactor, in addition to a Thermo-Scientific heated shaker, a graduated cylinder, and tap water. Gas composition was determined using gas chromatography using an SRI Instruments 8610C gas chromatographer with an equipped thermal conductivity detector (TCD). The GC program used measures only nitrogen, CO₂, CH4, and hydrogen sulfide. To measure gas composition, 5 mL of gas were injected into the GC, and the resulting values were recorded. Because the bacteria do not produce any nitrogen gas, and hydrogen 14 sulfide is not significant in the overall composition, the gas percentages can be adjusted to find the true CH4/ CO₂ ratio of the gas using Equation 2.3.1. 𝐴𝑑𝑢𝑠𝑡𝑒𝑑 %𝐶𝐻4 = % 𝐶𝐻4 % 𝐶𝐻4 + %𝐶𝑂2 ∗ 100 Equation 2.3.1. Biogas Quality 2.4 VFA Analysis To prepare samples for VFA analysis, digestate was diluted to 1/10th concentration before being centrifuged for five minutes at 6000 rpm. The samples were then passed through 0.22-micron filters before being stored for further analysis. The primary VFA of concern in this analysis is formic acid. To measure the quantity of formic acid, nuclear-magnetic resonance spectrometry (NMR) was used. To prepare the NMR samples, a solution of 4 mg TSP-d4 in 100 mL of D2O was prepared. Each sample was prepared in an NMR tube with 550 µL of sample and 50 µL of TSP solution. Samples were analyzed using the 600 MHz Bruker Avance NEO located in room B8 of the MSU Chemistry Building. Materials required for NMR analysis of formic acid include plastic and glass vials, micropipettes, 0.22-micron filters, 50-mL vials, centrifuge, NMR tubes, TSP-d4, D2O, and a 600 MHz Bruker Avance NEO. Gas chromatography was used to measure acetic, propionic, butyric, and valeric acid concentrations. For this analysis, a Shimadzu GC-2010 with an Agilent Technologies capillary column and a Shimadzu flame ionization detector (FID) was used. The test was conducted under isothermal conditions. 15 2.5 DNA Analysis The first step in DNA analysis was extraction of microbial DNA. This was accomplished with the use of a QIAGEN DNeasy Powersoil Pro Kit, using QIAGEN method HB-2494-003. This method includes the required materials. After the DNA was extracted, the DNA samples were stored at -80 C while awaiting analysis. DNA amplification was performed according to the methods detailed by Xu et al., 2025. The samples were multiplied via PCR (Polymerase Chain Reaction) to increase the concentration of DNA. The PCR used the universal primers; Pro 341 F (5′- CCTACGGGNBGCASCAG-3′) for the forward primer and Pro 805 R (3′- GACTACNVGGGTATCTAATCC-5′) for the reverse primer. The amplification target was the 16S rRNA gene in the V3-V4 region. A reaction solution containing 12.5 µL of GoTaq Green Master Mix, 0.5 µL of forward primer, 0.5 µL of reverse primer, 1 µL of extracted DNA, and 10.5 µL of DNase- and RNase-free water (total 25 µL) was used for PCR. An Eppendorf Mastercycler Pro Thermal Cycler Sample validity was confirmed using gel electrophoresis. The DNA samples were then submitted to the Michigan State University Research Technology Support Facility Genomics Core for sequencing on an Illumina MiSeq flow cell (v2) using a 500-cycle reagent kit. The methods and technical documents can be found at the RTSF Genomics Core website at https://rtsf.natsci.msu.edu/genomics/. 2.6 Statistical Analysis The recording processing of raw data was performed using Microsoft Excel. Statistical analysis was performed using R statistical software. The correct statistical test 16 to use is based on the normality and variance of the data (Emerson, Emilia Maria, 2024). Table 2.6.1 summarizes the diWerent available statistical tests. Table 2.3. Statistical Test Methods (adapted from Emerson, 2024) Test t-test or ANOVA Assumed Normal Assumed Equal Distribution? Variance? Mann-Whitney t-test Welch’s Standard t-test t-test Kruskal-Wallis ANOVA Brown-Forsythe and ANOVA No Yes Yes No Yes Welch Ordinary ANOVA Yes - No Yes - No Yes This study was conducted using biological duplicates; therefore, n=2. It is impossible to determine normality with the preferred statistical test for small sample sizes, the Shapiro-Wilk test, which requires at least 3 data points. It is then necessary to make assumptions about the data. Ordinary and Kruskal-Wallis ANOVA tests were both conducted, providing results both for and against the assumption of normally distributed data. Equal variance was tested using a Bartlett’s test. Following ANOVA testing, a Tukey’s Honest Significant DiWerence test conducted. The HSD test determines which specific means are significantly diWerent from one another. 17 For DNA analysis, the R packages Vegan, MASS, Phyloseq, and ggplot2 were used to analyze taxonomic data. Mnova software (Mestrelab Research) was used to analyze the output of the NMR analysis of formic acid. 2.7 Mass And Energy Balance Analysis Mass and energy balance analysis were conducted based on the laboratory data generated from the lab-scale digester experiment. The mass and energy balances are based on a system in which the eWluent from the main digester is used as the feedstock for a secondary digester, matching the conditions of the lab-scale study. The energy inputs for the digestion operation include both thermal energy (Wheat, kWh-e/year), to maintain the digestion temperature, and electricity energy (Welectricity, kWh-e/year), to operate auxiliary equipment such as pumps, mixers, and the control unit. The energy inputs were estimated using the Equation 2.7.1 and Equation 2.7.2, which were modified from a previous study (Bustamante & Liao, 2017). 𝑊!"#$ = 𝑚 × 𝐶% × (𝑇& − 𝑇’) × (1 + 20%) × 0.0002778 Equation 2.7.1. Estimated Required Heat Input 𝑊"(")$*+)+$, = 𝑚 × 0.00788 Equation 2.7.2. Estimated Required Electricity Input In these equations, m is the amount of the wet weight of the feedstock per year (kg); Cp is the heat capacity of the wet feedstock (4.12 kJ/(kg K)); TR is the reactor temperature, which was 313 K for the main digester and 323 K for the formate-enhanced digester, respectively; TO is the temperature of the wet feedstock, which was 313 K, the temperature of the filtrate from the previous digester; 20% is the percentage of the additional heat 18 needed to maintain the digestion temperature; 0.0002778 is the conversion factor of KJ to kWh; and 0.00788 is the average electricity demand of the MSU SCAD digester operation (kWh/kg wet feedstock). 2.8 Life Cycle Impact Assessment (LCIA) The life cycle assessment (LCA) methodology was developed to quantify and compare the environmental impacts associated with a formate-enhanced anaerobic digester and two control configurations, AD with activated carbon and AD alone, based on the mass and energy balance. Two environmental impact categories were assessed: Global Warming Potential (GWP), expressed in CO₂-equivalents (g CO₂-e), and Water Eutrophication Potential (WEP), expressed in nitrogen equivalents (g N-e). Greenhouse gas emissions were calculated based on CH4 and nitrous oxide (N₂O) emissions from land application of digestate. The emission factors for land application were 0.001 g CH₄/g volatile solids (VS) and 0.005 g N₂O/g total nitrogen (TN) in the digestate. Global warming potentials were taken as 25 g CO₂-e/g CH₄ and 265 g CO₂-e/g N₂O (US EPA, 2016). Electricity usage was evaluated based on system energy demand and its associated GWP. A GWP of 491 g CO₂-e/kWh-e was used for natural gas-based electricity generation, consistent with U.S. EPA emission factors. Eutrophication impacts were assessed based on total nitrogen (TN), total phosphorus (TP), and chemical oxygen demand (COD) in the eWluent. Characterization factors were 0.9864 g N-e/kg TN, 7.29 g N-e/kg TP, and 0.05 g N- e/kg COD. In alignment with standard LCA practice, CO₂ emissions from the degradation of biogenic materials in the feedstock and eWluent were not included in the GWP 19 calculations. These biogenic CO₂ emissions are considered to be oWset by atmospheric CO₂ capture through plant regrowth due to the biological origin of the feedstock. The key parameters and impact factors used in the analysis are summarized in Table 2.4. Table 2.4. Parameters of Life Cycle Assessment Value Unit Source CH4 emission of digestate – land application 0.001 g CH4/g VS in the waste (Owen & Silver, 2015) GWP of CH4 25 g CO2/g CH4/ (Turnbull, 2004) N2O emission of digestate – land application GWP of N2O GWP of natural gas energy Water eutrophication potential (WEP) of TN Water eutrophication potential (WEP) of TP Water eutrophication potential (WEP) of COD 0.005 g N2O/g TN in the waste (US EPA, 2016) 265 491 g CO2-e/g N2O (US EPA, 2016) g/kWh-e (US EPA, 2016) 0.9864 g N-e/kg TN in the waste (US EPA, 2016) 7.29 g N-e/kg TP in the waste (US EPA, 2016) 0.05 g N-e/kg COD in the waste (US EPA, 2016) 20 CHAPTER 3: FORMATE-ENHANCED AD 3.1 Gas Production Figure 3.1.1 shows the total gas production for each reactor over the course of the experiment. This number is the entire biogas volume, including both CH4 and CO2. The data presented is for the individual reactors. Figure 3.1.1. Cumulative Biogas Production An ANOVA analysis of the gas production data indicated that there were significant diWerences between the mean gas production of diWerent reactor conditions (a = 0.05). A following Tukey’s HSD test indicated means between which a significant diWerence occurred. 21 With a = 0.05, entries with adjusted p-values below 0.05 qualify as significantly diWerent from the compared mean. There is a significant diWerence between the formate reactors and the control reactors, as well as a significant diWerence between the bicarbonate/formate and control reactors. There is not a significant diWerence between other reactors, including between control and activated carbon reactors or formate and formate/bicarbonate reactors. Taken together, these factors suggest that formate was the determining factor in whether a reactor would produce more biogas. Activated carbon did not produce significantly more biogas on its own, and the addition of bicarbonate with formate did not show a significant increase in production either. Production was in fact lower, but not to a statistically significant degree. A Kruskal-Wallis test on total gas production did not reveal a significant diWerence between the means. However, this may be because non-parametric tests require larger sample sizes to achieve the same power as parametric tests. Figure 3.1.2 shows the average daily CH4 production rate for each set of reactors in each HRT. In this figure, Ctrl-AD is Control, Ctrl-AC is Activated Carbon, R1 is Bicarbonate, R2 is Formate, and R3 is Bicarbonate and Formate. The data is the average of all data points for both replicates in the respective HRT. 22 Figure 3.1.2. Daily Methane Production per Reactor and HRT The CH4 production of the formate-enhanced reactors is notably higher than that of other reactors across all HRTs. The diWerence is particularly notable in the fourth, stable HRT. The bicarbonate/formate reactor had CH4 production between formate reactors and bicarbonate reactors, suggesting that bicarbonate was detrimental to biogas production during formate treatment. Gas production numbers were not normalized to added volatile solids. This approach was chosen because all reactors received the same feedstock, and therefore the same volatile solids amount. In the case of the formate-added reactors, the total extra solids added was considered to be negligeable in the overall experimental results. 23 3.2 Gas Quality The average gas quality of each reactor is shown in Figure 3.2.1. The quality is given in terms of the decimal fraction of the biogas that is CH4, with 1 indicating that the gas is 100% CH4. The data is the average of all data points for both replicates in the respective HRT. Figure 3.2.1. Mean Gas Quality Per HRT An ANOVA test indicated that there were significant diWerences between means, with an extremely low p-value (p = 1.65x10^-5). A following Tukey’s HSD test indicated that there were significant diWerences between most mean groups. Notably, there is no significant diWerence between formate and activated carbon reactors, nor between control 24 and carbonate-formate reactors. There was a significant diWerence between control and bicarbonate reactors, with bicarbonate reactors being worse. These results indicate that formate addition did not significantly improve biogas quality. The results also indicate that there is a significant negative impact on biogas quality exerted by bicarbonate addition. 3.3 pH, TSVS, and COD 3.3.1 Chemical Oxygen Demand Figure 3.3.1 shows the mean COD values in mg/L for each reactor in each HRT. The data is the average of all data points for both replicates in the respective HRT. The dilution factor was then applied; reported values are undiluted COD. Figure 3.3.1. Mean COD per HRT 25 COD is an indicator of the amount of digestible solids available in the digester culture. All the reactors were fed with the same feedstock, so variations in the COD indicate diWerences in the digestion. Higher COD suggests that less of the organic solids available were consumed, while lower numbers indicate higher usage. Overall, COD destruction increases in the first and second HRTs, appearing to stabilize in the third and fourth. In the fourth HRT, reactors that were treated with bicarbonate show higher COD values, indicating that less COD reduction was achieved in those reactors. Statistical tests were conducted on the fourth HRT, with the average value of all measurements for a particular reactor taken in that period used as the value for that reactor. The Kruskal-Wallis test did not indicate a significant diWerence between means. The ordinary ANOVA on COD did indicate a significant diWerence. A following Tukey’s HSD test indicated that significant diWerences between means existed between all means except for Control-Activated Carbon and Bicarbonate/Formate-Bicarbonate. This indicates that bicarbonate addition is associated with a significant increase in COD, which in turn indicates a lower level of COD destruction and therefore a decrease in reactor eWiciency. The formate reactors also showed a significant increase in COD as compared to the activated carbon reactors, indicating that COD destruction was decreased when either formate or bicarbonate was added. 3.3.2 pH Figure 3.3.2 displays the mean pH measured for each reactor condition in each HRT. The data is the average of all data points for both replicates in the respective HRT. 26 Figure 3.3.2. Mean pH per HRT pH remained within the expected range for healthy digester operation. Every pH measurement was above 7.0, indicating some alkalinity in all reactors. The pH in the Activated Carbon and Formate reactors shows a trend of decreasing over the course of the experiment as compared to the other reactors. A Kruskal-Wallis test did not indicate any significant diWerence between means; however, the ordinary ANOVA test did indicate a significant diWerence between means. In particular, the activated carbon and formic acid treated reactors were shown to have a significantly lower pH compared to the control reactor or either reactor that received 27 bicarbonate. The higher pH of the bicarbonate reactors is expected, as bicarbonate serves as a buWer to stabilize pH. However, in this case, it seems that it was not necessary to have that buWer, as none of the other reactors acidified to a problematic degree. In fact, the buWer may have inhibited the reactor activity by bringing it further outside of the ideal pH range for methanogenic activity. 3.3.3 Total and Volatile Solids Figure 3.3.3 shows the mean total solids per HRT in g/L concnetration. The data is the average of all data points for both replicates in the respective HRT. Figure 3.3.4 is the equivalent figure for volatile solids. Both measurements are discussed together, as they follow similar trends and the diWerences in total solids are largely attributable to changes in volatile solids, as the degradation of volatile solids is the primary way in which the microbial culture can influence total solids. 28 Figure 3.3.3. Mean Total Solids per HRT 29 Figure 3.3.4. Mean Volatile Solids per HRT Total and volatile solids indicate the amount of material in the digester. Most total solids are volatile solids, as this is the fraction of solids that is organic and can be degraded or burned, which makes up most of the solids in the feedstock. Greater solids concentration indicates lower performance, as biogas is produced from the degradation of volatile solids and all digesters were fed with the same feedstock with the same solids concentration. A Kruskal-Wallis test did not detect a significant diWerence in either total or volatile solids between reactors; an ordinary ANOVA detected a significant diWerence in both. Significant diWerences were found between both bicarbonate-added conditions and the 30 non-bicarbonate added conditions, with bicarbonate-added reactors showing significantly higher concentrations of total and volatile solids. This indicates a decrease in the metabolism of organic materials in the bicarbonate treated reactors. 3.4 Volatile Fatty Acids 3.5 Formic Acid The addition of formate was shown to have significant impacts on biogas production and quality. The average concentration of formic acid is shown at each HRT in Figure 3.5.1. For VFA data, values are the average for samples from that HRT, but from one replicate each. Figure 3.5.1. Mean Formic Acid per HRT 31 All samples, with one exception, lie in the 9-30 mg/L range for formic acid concentration. However, the formate and carbonate/formate reactors had feedstock with a 100 mg/L concentration of formate. This means that the formate reactors were able to metabolize all of the formic acid in the feedstock. In the case of the HRT 4, the formate only reactor has lower formic acid than any of the other reactors, indicating that the formic acid produced in the culture was also metabolized more eWiciently. The activated carbon only reactors also demonstrated a reduced formic acid concentration, indicating that the presence of the conductive material aided in the metabolism of formic acid. In the first HRT, the bicarbonate-only reactors indicated a much larger formic acid concentration than any other reactors. This is due to two measurements in that HRT that are much larger than the other measurements. However, these values fell to values in the normal range by the next HRT. Formic acid is produced and consumed quickly, so it is plausible that these reactors were able to change concentration relatively quickly compared to the other feeding conditions. While there may have been temporary formic acid build-up in those reactors, by the time reactor stability was reached, bicarbonate addition did not seem to have caused continuing formic acid accumulation. 3.6 Acetic Acid, Propionic Acid, Butyric Acid, Valeric Acid 3.6.1 Acetic Acid Acetic acid is a main intermediary product in AD. A diWerence in acetic acid concentrations would suggest a diWerence in the rate of acetogenesis or acetate- consuming methanogenesis. The concentration of acetic acid decreased over the course 32 of the experiment as the reactors approached the stable phase. Figure 3.6.1 shows the mean concentration of acetic acid in each feeding condition in each HRT. For VFA data, values are the average for samples from that HRT, but from one replicate each. Figure 3.6.1. Mean Acetic Acid per HRT In the fourth HRT, by which time the reactors had stabilized, the concentration of acetic acid was even across all feeding conditions. This suggests that the presence or absence of activated carbon, bicarbonate, or formate did not have a large impact on the rate of acetate metabolism or acetate production. 33 3.6.2 Propionic Acid Figure 3.6.2 shows the mean concentration of propionic acid in each reactor condition during each HRT, obtained by averaging the propionic acid measurements for each condition in that HRT. For VFA data, values are the average for samples from that HRT, but from one replicate each. Figure 3.6.2. Mean Propionic Acid per HRT As with acetic acid, the concentration of propionic acid fell over the course of the experiment. In the fourth HRT, there is a notable diWerence in the concentration of propionic acid between the reactors treated with formic and the reactors that lack formic acid; that is, reactors with formic acid show a decrease in the concentration of propionic acid. This suggests that formate addition plays a role in enhancing the metabolism of 34 propionic acid. This is consistent with the notion discussed in the literature that formic acid addition serves to enhance IET, as that process is essential to the eWicient degradation of propionic acid. 3.6.3 Butyric and Isobutyric Acid Figure 3.6.3 shows the mean concentration of butyric acid in each HRT, while Figure 3.6.4 shows the mean concentration of isobutyric acid. For VFA data, values are the average for samples from that HRT, but from one replicate each. Figure 3.6.3. Mean Butyric Acid per HRT 35 Figure 3.6.3. Mean Isobutyric Acid per HRT Butyric acid and isobutyric acid follow similar trends, both to one another and other VFAs. There is a decrease in concentration as the experiment progresses. In the fourth HRT, there is again a decreased concentration of the acid in reactors that received formic acid. There is no bar shown for isobutyric acid for formate and bicarbonate/formate in the fourth HRT because the concentration was below the detection threshold, and is therefore listed as 0 mg/L. Like propionic acid, butyric acid is thermodynamically unfavorable to degrade and therefore benefits from the enhancement of IET. 36 3.6.4 Isovaleric Acid Figure 3.6.5 shows the mean concentration of isovaleric acid in each HRT. For VFA data, values are the average for samples from that HRT, but from one replicate each. Figure 3.6.4. Mean Isovaleric Acid per HRT The last VFA with a detectable quantity was isovaleric acid. The concentration follows a similar trend to propionic, butyric, and isobutyric acid, with the concentration decreasing as the experiment progresses. In the fourth HRT, the reactors treated with formic acid show decreased acid concentration compared to reactors without formic acid addition. Notably, the bicarbonate/formic acid combined treatment shows a greater reduction in isovaleric acid concentration as compared to formate alone. 37 3.7 Microbial DNA In the microbial DNA analysis section, the reactors are labeled Ctrl-AD, Ctrl-AC, R1, R2, and R3. These correspond to Control (C), Activated Carbon (A), Bicarbonate (B), Formate (F), and Bicarbonate and Formate (CF) respectively. Table 3.7.1 provides these equivalents for ease of reference. The analysis of the microbial communities focuses on HRT 3 and 4, which are the stable periods of digestion. Table 3.1. Reactor Name Equivalencies Reactor Experimental Designation Control Activated Carbon Bicarbonate Formate Bicarbonate plus Formate C A B F CF Microbial Analysis Designation Ctrl-AD Ctrl-AC R1 R2 R3 3.7.1 Alpha Diversity Shannon, Simpson, and Sobs Indices A statistical analysis of the microbial community was performed to determine the Simpson, Shannon, and Sobs indices of alpha diversity. Alpha diversity measures the diversity of a single community or sample. The Simpson Index of Diversity quantifies species dominance. Higher values indicate greater diversity, as lower values indicated domination by a single species. The Shannon Index indicates community diversity and evenness. Higher values indicate a more diverse ecology with a more even distribution of species. The Sobs Index measures community abundance. Figure 3.7.1 shows the distribution of Shannon, Simpson, and Sobs indices for culture bacteria. 38 Figure 3.7.1. Shannon, Simpson, and Sobs Indices of Alpha Diversity for Bacteria Between HRT 3 and HRT 4, both the Shannon and Simpson index increased, while the Sobs index decreased, particularly in formate-only and bicarbonate-only reactors. This indicates that while a more balanced microbial community emerged as stabilization approached, the selective pressures created by the formate or bicarbonate rich environments of the digesters may have reduced the richness of bacterial species present. Figure 3.7.2 shows the distribution of Shannon, Simpson, and Sobs indices for culture archaea. 39 Figure 3.7.2. Shannon, Simpson, and Sobs Indices of Alpha Diversity for Archaea The diversity and richness of the archaea community were much lower compared to the bacterial community, indicating more similar community members were present. Bicarbonate-only and formate-only reactors had lower Simpson indices, indicating that both formate and bicarbonate, when added separately, contributed to a more even archaea community. 3.7.2 Beta Diversity Beta diversity measures diWerences in community between multiple groups. A non- metric multidimensional scale (NMDS) analysis was conducted to evaluate the beta diversity. The stress value of the model indicates reliability, with values below 0.1 indicating 40 a more reliable model. Figure 3.7.3 shows the beta diversity plot for the bacterial community. Figure 3.7.3. Beta Diversity of Bacteria The stress value for the bacterial community model was 0.06. A significant diWerence (P < 0.05) was indicated between reactors, suggesting that the diWerent feedstock conditions had a significant impact on the composition of the microbial community. The archaea model yielded similar results, with a stress value of 0.04. That analysis is displayed in Figure 3.7.4. 41 Figure 3.7.4. Beta Diversity of Archaea 42 3.7.3 Bacterial Community Analysis Figure 3.7.5. Phylum-Level Distribution of Bacteria The primarily dominant bacterial phyla were Fimicutes, Bacteroidota, Proteobacteria, and Synergistota. In particular, the Firmicutes phylum dominated in all reactors, ranging from 57.56% abundance in bicarbonate reactors to 67.01% in bicarbonate and formate reactors. Its abundance in all reactors increased from HRT 3 to HRT 4. This increase was particularly large in the formate-treated reactors, increasing from 44.6% to 70.86%. 43 Figure 3.7.6. Genus-Level Distribution of Firmicutes Phylum In contrast to the phylum’s dominance, there was no single dominant genus within Firmicutes. Only the Fastidiosipila, MBA03, Romboutsia, and Sedimentibacter genera exhibited an average realitve abundance greater than 5% across all reactors. Another notable phylum-level change was the increase in abundance of Synergistota, which increased in all reactors from HRT 3 to HRT 4, but particularly increased in the formate reactors. 44 3.7.4 Archaeal Community Analysis Figure 3.7.7. Phylum-Level Distribution of Archaea The dominant and in fact only archaeal phylum observed in all reactors was Euryarchaeota. In the fourth HRT, where the CH4 production diWerences were the most pronounced, there is a notable diWerence in the abundance of archaea in the formate- enhanced reactors as compared to the other reactors in that HRT. This suggests that formate enhancement of methanogens contributes to increased CH4 yield. 45 Figure 3.7.8. Genus-Level Distribution of Euryarchaeota Phylum At the genus level, Methanobacterium, and Methanobrevibacter dominated, with a relative abundance above 95%. In HRT 4, 26.73% of archaea in formate reactors were Methanobrevibacter, as compared to 22.99% in the bicarbonate reactors or 25.80% in the formate and bicarbonate reactors. Methanobrevibacter contains species capable of formate utilization for methanogenesis. This finding indicates that the presence of formic acid in the reactor enhances the abundance of microbes adapted for the utilization of formate in methanogenesis. 46 CHAPTER 4: LIFE CYCLE ASSESSMENT 4.1 Introduction Building on the findings presented in Chapter 3, a comprehensive LCA was conducted to evaluate the environmental performance of a formate-enhanced AD system. In this system, a secondary digestion is performed using the liquid digestate of a primary digester to utilize the residual carbon in the digestate and CO2 from the primary digester. The LCA aims to assess the environmental impacts associated with implementing formate- enhanced AD to provide insights into its potential as a sustainable, value-added pathway to improve digester performance. 4.1.1 Studied System Figure 4.1.2 depicts a diagram of the studied system. SCAD produces, on average, 17,081 metric tons/year of liquid digestate, which was used as the feedstock for this study. 47 Figure 4.1.1. Diagram of the Studied Formate Digester System First, the feedstock enters a formate-enhanced AD from SCAD. The digestion produces biogas, a mixture consisting mainly of CH4 and CO₂. The biogas is directed to a biogas utilization unit that combusts CH4 to generate electricity. The flue gas from biogas combustion is then fed into an electrocatalysis unit that converts CO2 into formic acid using electricity generated biogas consumption. The formic acid is recycled back into the formate-enhanced digester to stimulate methanogenesis, enhancing the carbon conversion eWiciency. Activated carbon is added during digester start-up based on the results from Chapter 3. The system produces a liquid eWluent which is used in land application for nutrient recycling. 48 4.1.2 Goal and Scope The goal of this LCA is to determine the environmental impacts of formate- enhanced anaerobic digester operation. To assess the potential environmental benefits and/or trade-oWs of this enhanced configuration, two control scenarios were selected for comparison based on experimental setups described in Chapter 3. AD alone served as the baseline system, while AD with activated carbon served as an enhancement control for comparison to the formate-enhanced configuration. The bicarbonate and bicarbonate- formate reactors were not evaluated. While the digestion was not unsuccessful, the performance of the bicarbonate-treated reactors was not deemed to exceed the activated carbon reactors enough to be considered for full-scale estimation. Similarly, the performance of the bicarbonate-formate reactor was lower than formate alone, and therefore not considered for full-scale. The scope of the LCA includes all major processes directly related to digestion, biogas production and utilization, formate production via electrocatalysis, and digestate management. Environmental burdens from upstream or unrelated processes, such as feedstock generation, infrastructure construction, and unrelated agricultural activities, are excluded from the system boundary. This LCA aims to provide a system-level understanding of how formate addition as a strategy for CO₂ valorization impacts the mass and energy balances, greenhouse gas emissions, and environmental sustainability as compared to conventional AD configurations. 49 4.2 Mass and Energy Balance The mass balance analysis compares the formate-enhanced AD system to the control, a conventional AD system, and the enhanced control, a conventional AD system using activated carbon. All three systems use liquid digestate eWluent from MSU’s SCAD as feedstock, with a total amount of 17,081 tons/year. The characteristics of AD filtrate are 4.1% TS, 3.0% VS, 26,625 mg/kg TC, 53,380 mg/kg COD, 3,293 mg/kg TN, and 509 mg/kg TP, with those values based on SCAD operational data. The outputs of the mass balance were biogas production, energy usage and production, and eWluent characteristics, which were quantified to evaluate digester performance and implications for environmental impacts. Figure 4.2.1 displays the mass balance for the formate-enhanced system. Figure 4.2.1. Mass Balance of the Formate-Enhanced AD System 50 a. The data used for the mass balance is from the 4th HRT. The formate-enhanced AD has a retention time of 20 days. b. The electrolyzer runs 10 hours/day with a flow rate of 46 m3/hr. The formic acid concentration generated from electrolyzer is 150 g/kg solution (Complete 5 Cm2 Electrolyzer to Convert CO2 to Formic Acid | Dioxide Materials, n.d.). The AD has a retention time of 20 days. c. The lambda value for the CHP engine is 1. In the formate-enhanced AD system model, 150 g additional formate per kg digestate was introduced into the digester via electrocatalytic conversion of captured CO₂ from flue gas. A total of 115 tons/year of formate solution was generated using 113.5 tons/year of water and 48,400 m³/year of flue gas containing 21% CO₂. This system achieved the highest biogas production among all three systems, producing 107,647 m³ biogas/year, consisting of 63% CH₄ and 37% CO₂. The eWluent volume was 16,769 tons/year, with reduced TS (1.6%), VS (1.1%) and COD (7,393 mg/kg), indicating eWective solids reduction. The flue gas output was 314,370 m³/year, from which 48,400 m³/year was diverted to the electrolyzer. Overall, the formate-enhanced AD model demonstrated improved carbon utilization and enhanced CH4 yield. Figure 4.2.2 shows the mass balance on a system with added activated carbon, but without formate addition. The activated carbon addition in this model slightly improved digestion performance relative to the control. 51 Figure 4.2.2. Mass Balance of the Activated Carbon AD System b. The lambda value for the CHP engine is 1. d. The data used for the mass balance is from the 4th HRT. The AD has a retention time of 20 days. Activated carbon was added at the beginning of the AD. The biogas yield reached 78,333 m³/year, with a CH₄ concentration of 60%, similar to the control configuration. EWluent output was 16,637 tons/year, with TS of 1.5% and VS of 1.1%. COD was reduced to 6,870 mg/kg. Flue gas release totaled 255,143 m³/year, lower than in the formate-enhanced system, due to lower total biogas production. Although activated carbon facilitated improved CH4 production, the enhancement was limited compared to the formate-enhanced digestion. The final mass balance was performed on an AD with no additives using digester eWluent as feedstock, which is shown in Figure 4.2.3. 52 Figure 4.2.3. Mass Balance of the Conventional AD System b. The lambda value for the CHP engine is 1. e. The data used for the mass balance is from the 4th HRT. The AD has a retention time of 20 days. No additives were used. The standard AD system produced the lowest biogas volume at 65,352 m³/year, with a CH4 content of 60%. The eWluent profile was comparable to that of the AD with activated carbon, with TS of 1.5%, VS of 1.2%, and COD of 6,967 mg/kg, indicating relatively less eWicient organic degradation compared to formate-added digestion. Flue gas emissions were the lowest among the three systems at 212,860 m³/year, attributed to lower CH4 production and combustion activity. The mass balance model data show that the formate-enhanced system outperformed the other configurations in terms of CH4 yield and carbon recovery, demonstrating the value of recirculating CO₂ as a substrate via electrolyzation to formate. The integration of electrocatalysis enables a partial carbon loop closure, improving energy eWiciency and potentially providing a beneficial eWect on emissions. Activated carbon 53 showed moderate benefits but did not significantly impact eWluent quality or CH4 content compared to the control. Following the mass balance results, an energy balance analysis was conducted to evaluate the energy performance of the formate-enhanced AD, AD with activated carbon, and conventional AD systems. Table 4.2.1 summarizes the energy inputs and outputs associated with each system, including electricity and heat requirements for system operation and energy recovery from combined heat and power (CHP) units. Table 4.2.1. Summary of Energy Balances Unit operation Electrolyzer energy input Formate- enhanced AD Control AD with activated carbon Control AD Electricity input (kWh-e/year) -8,083 - - AD energy input Heat input (Wheat, kWh-e/year) b Electricity input (Welectricity, kWh- e/year) c -236,159 -236,159 -135,504 -135,504 -236,159 -135,504 163,586 280,434 233,960 236,043 404,645 CHP energy output from the formate AD Energy output as heat (Eheat, kWh-e/year) d Energy output as electricity (Eelectricity, kWh-e/year) e Net energy output Net heat output (kWh-e/year) f Net electricity output (kWh- e/year) g a. b. c. d. content of the biogas. The LHV of CH4 is 35.8 MJ/m3 CH4. The thermal conversion eXiciency of the CHP unit is 65%. e. f. g. Negative numbers indicate energy inputs; positive numbers indicate energy outputs. Eq. 2.7.1 was used to calculate the heat input. Eq. 2.7.2 was used to calculate the electricity input. An annual biogas production of 1,323,757 m3 with 65% (v/v) of CH4 was used to calculate the energy The electricity output is the metered number of the digestion operation. Net heat output = Eheat - Wheat Net electricity output = Eelectricity - Welectricity 168,487 136,477 28,082 44,275 92,456 -2,199 972 54 In the formate-enhanced AD system, additional energy input is required for the electrolyzer to perform CO2 to formic acid conversion. This unit consumed 8,083 kWh- e/year, which is the only diWerence in energy demand among the three systems. All three systems required the same operational energy inputs for digestion, which was 236,159 kWh-e/year of thermal energy to maintain thermophilic conditions and 135,504 kWh- e/year of electricity to power the operation, including pumps, mixers, and other auxiliary equipment. While the formate-enhanced system had slightly higher total energy input due to the electrolyzer, it generated significantly more recoverable energy than either of the control systems because of its higher biogas yield. Specifically, it produced 404,645 kWh- e/year of heat and 236,043 kWh-e/year of electricity from the CHP unit. Comparatively, the AD with activated carbon yielded 280,434 kWh-e/year of heat and 163,586 kWh-e/year of electricity, while the conventional AD produced the lowest outputs, at 233,960 kWh-e/year of heat and 136,477 kWh-e/year of electricity. These data reflect the proportional increase in CH4 production seen in the mass balance data, as increased CH4 production leads directly to increased energy production from combustion. The most meaningful metric of the system’s energy eWiciency is the net energy output, which is calculated by subtracting energy inputs from energy outputs. The formate- enhanced AD achieved the highest net heat output at 168,487 kWh-e/year, and the highest net electricity output, at 92,456 kWh-e/year. The AD-activated carbon system showed moderate improvement at 44,275 kWh-e/year of net heat and 28,082 kWh-e/year of net electricity. The control AD system had a negative net heat output, at −2,199 kWh-e/year, meaning that the heat recovered from the CHP was insuWicient to oWset the heating energy 55 required for thermophilic operation. Its net electricity output was marginal, at only 972 kWh-e/year. These results demonstrate that only the formate-enhanced AD system achieves a substantially positive energy balance, confirming its technical advantage in surplus energy generation and self-suWiciency. The superior CH4 yield from formate supplementation not only enhances carbon conversion eWiciency but also significantly boosts energy recovery. The improved energy balance of the formate-enhanced system suggests a viable pathway for increasing the sustainability of AD operations. Despite the additional electricity demand of the electrolyzer, the increase in energy output—especially electricity—makes the system less dependent on external energy inputs. This may allow excess energy to be exported or used to support adjacent agricultural or industrial operations. Furthermore, the activated carbon system, while oWering moderate improvement over the baseline, did not achieve the same level of net energy gains and remained largely dependent on external heating sources. 4.3 Life Cycle Impact Assessment To evaluate the environmental impacts of the three anaerobic digestion configurations studied, life cycle impact assessment (LCIA) was conducted focusing on two key categories: global warming potential (GWP) and water eutrophication potential (WEP). The systems assessed included formate-enhanced AD, AD with activated carbon, conventional AD, and a reference case involving direct land application of AD filtrate without secondary digestion. 56 The GWP results are presented in Figure 4.3.1, which breaks down the contributions from CH₄ and nitrous oxide (N₂O) emissions, as well as the GHG oWset from energy recovery. Figure 4.3.1. GWP Contribution Analysis All three digestion systems resulted in comparable direct CH₄ emissions of approximately 5 metric tons CO₂-e per year, primarily from residual volatile solids in the digestate applied to land. In contrast, the land application of untreated digestate produced the highest CH₄ emissions, reaching 13 metric tons CO₂-e/year due to the absence of organic matter stabilization prior to land application. N₂O emissions were uniformly high across all scenarios involving land application—117 metric tons CO₂-e/year due to the consistent TN content in the eWluent and the use of the same emission factor. These emissions are significant and often dominate the climate impact from digestate management. The most notable diWerences between systems arise from the CH₄ energy credit, which accounts for avoided emissions resulting from renewable electricity and heat 57 generation. The formate-enhanced AD system provided the largest GHG oWset, with an energy credit of –128 metric tons CO₂-e/year due to its enhanced biogas yield and greater CH4 recovery. The AD with activated carbon yielded a smaller credit of –36 metric tons CO₂- e/year, while the conventional AD system contributed minimally, with just -1 metric tons CO₂-e/year. As a result, the formate-enhanced system was the only one to achieve net- negative GWP, eWectively compensating for all CH₄ and N₂O emissions through natural gas energy replacement. In contrast, the conventional AD system and direct land application scenario had net-positive GHG emissions, emphasizing the superior climate impacts of the formate-enhanced system. Figure 4.3.2 presents the results for WEP, again comparing formate added-AD, activated carbon AD, untreated AD, and direct application of digestate without treatment. Figure 4.3.2. WEP Contribution Analysis Across all digestion systems, the contributions from TN and TP remained consistent at 55 kg N-e/year and 63 kg N-e/year respectively, due to similar nutrient profiles in the final 58 eWluent. The key diWerence was observed in the contribution of COD, which reflects the organic matter content of the eWluent. The formate-enhanced AD system exhibited the lowest COD-related eutrophication potential at 6.2 kg N-e/year, followed by AD with activated carbon at 5.7 kg N-e/year, and conventional AD at 5.8 kg N-e/year. Land application of untreated digestate resulted in a significantly higher COD contribution of 46 kg N-e/year due to the lack of prior stabilization and degradation of organic matter. These findings confirm that all anaerobic digestion systems oWer environmental advantages over direct land application, particularly in reducing eutrophication risks. Among them, the formate-enhanced AD system demonstrates the best overall environmental performance, achieving both the lowest net GWP and WEP. This improvement is primarily attributed to the dual benefits of CO₂ utilization via formate recycling and enhanced CH4 production. The activated carbon system, while oWering some benefit over the control, does not deliver the same reduction of environmental impacts. In summary, the LCIA results support the conclusion that formate-enhanced AD not only improves energy recovery but also oWers clear climate and water quality benefits, making it a promising strategy for sustainable waste-to-energy conversion. 4.4 Carbon EZiciency To further compare the three operating conditions, the carbon eWiciency of the reactors was evaluated. The carbon eWiciency is expressed as the percentage of total carbon loaded that was converted into biogas. Figure 4.4.1 shows the carbon eWiciency of the three situations based on the mass balance. 59 Figure 4.4.1. Carbon Ebiciencies Based on Mass Balance The figure shows both the eWiciency for the entire biogas and the eWiciency for methane only. The carbon eWiciency is higher in the activated carbon AD compared to the control and further increased with formate enhancement. Comparing the control and the activated carbon treatment, the carbon eWiciency is increased by 21% when comparing methane content. The formate-enhanced digester, when comparing methane content, shows a 76% carbon eWiciency increase as compared to control. This result has significant implications for the viability of digester operations, as the eWective number of cattle needed to ensure digester viability would be lowered by this increase in carbon eWiciency. 60 CHAPTER 5: CONCLUSIONS 5.1 Findings The study found that there was evidence to support a significant increase in biogas production when reactors were treated with formic acid. The mean biogas production between reactors with activated carbon and formic acid was approximately 21% greater than reactors treated only with activated carbon, and ~47% greater than reactors treated with neither activated carbon or formic acid. This was determined to be a statistically significant increase when equal variances and normality were assumed. The addition of sodium bicarbonate resulted in a 4% increase in biogas yield, which was not determined to be statistically significant under any assumptions. The combination of formate and bicarbonate resulted in an 8.8% increase in total gas production as compared to activated carbon only reactors and a 10% decrease in total gas production compared to the formate reactors. The addition of formic acid had a significant enhancement eWect on gas production, while bicarbonate had a minimal or even detrimental eWect on the ability of the anerobic microbes to metabolize the feedstock. The concentration of formic acid measured largely did not vary in the diWerent feeding conditions, and most samples had concentrations in the 1-3 mg/L range. However, two feeding conditions were receiving feedstock with a 100 mg/L formic acid concentration. This means that the reactors that received This means that the reactors that received formic acid were able to fully metabolize the formic acid from the feedstock. Further, reactors treated with formic acid showed lower concentrations of propionic, 61 butyric, isobutyric, and isovaleric acid, indicating a more eWicient metabolism of VFAs when formate was present. The microbial community analysis indicated a higher abundance of formate- scavenging organisms in formic acid reactors. This suggests that the microbial community could adapt under selective pressure to more eWectively utilize formate, which is further reinforced by the lack of increased formic acid concentration in reactors treated with formic acid. Archaea abundance was shown to be enhanced by the addition of formic acid. The LCA demonstrated that a system using formate-enhanced digestion would be capable of significantly reducing both GWP and WEP impacts of AD operation. Further, the energy eWiciency of such a system was shown to be superior to a standard two-stage digester operation, as well as improving carbon eWiciency and utilization. This suggests that formate-enhanced digestion as a viable strategy for improving anaerobic digester operation. 5.2 Recommendations for Future Work Although the formate study presented in Chapter 2 of this thesis found that there was evidence to support the positive impact of formate addition on digester performance, it is unlikely that said findings would constitute suWicient evidence to motivate implementation of formate addition to full-scale digester operations. Further, existing data on formate addition remains sparse at the time of writing. Additional studies on the eWects of formate addition to digesters, incorporating a variety of feedstocks, reactor conditions, additives, and formate concentrations, are a much-needed step to bring this technology to full-scale fruition. In specific, the author recommends that future studies perform 62 experiments with a greater number of biological replicates (n >= 3). A study that includes data from three or more reactors will have greater statistical power and be able to utilize tests for normality and variance more eWectively. More detailed monitoring of eWluent and influent COD and total solids data should be performed in future studies, as the suggested COD and TS reductions used in the mass balance are overly eWicient, and evoke some skepticism. Additionally, studies on feedstocks other than liquid food waste/manure co- digestate would provide vital insight into the eWects of formic acid on other cultures. Of particular utility would be a study on formic acid treatment for reactors fed with manure and/or food waste directly. Provided that future bench-scale studies support the viability of formate addition, pilot-scale studies should be performed to ensure that the technology remains viable as it approaches full-scale. This work was conducted in 500 mL bottles, and is in no way guaranteed to perform similarly at increased volumes. The LCA portion would benefit from future work involving results from pilot-scale studies on codigestion operations treated with formic acid. Such results would be an important step in determining potential full-scale viability. Additionally, it may be possible that a more nutrient-rich environment could take on more formic acid addition, allowing further reductions GHG emissions and low- or negative-carbon energy production. A full review of the process’s economic viability and optimization of that viability should also be conducted to determine the technology’s fitness for commercial implementation. A comprehensive review of available electrocatalysis technology and other details of the formate-enhanced system could bring the technology into economic viability, providing 63 great environmental and financial benefits to the AD waste treatment and energy production system. 64 WORKS CITED Akindele, A. A., & Sartaj, M. (2018). The toxicity eWects of ammonia on anaerobic digestion of organic fraction of municipal solid waste. Waste Management, 71, 757–766. https://doi.org/10.1016/j.wasman.2017.07.026 Belay, N., Sparling, R., & Daniels, L. (1986). Relationship of formate to growth and methanogenesis by Methanococcus thermolithotrophicus. Applied and Environmental Microbiology, 52(5), 1080–1085. https://doi.org/10.1128/aem.52.5.1080-1085.1986 Bustamante, M., & Liao, W. (2017). A self-sustaining high-strength wastewater treatment system using solar-bio-hybrid power generation. Bioresource Technology, 234, 415– 423. https://doi.org/10.1016/j.biortech.2017.03.065 Complete 5 cm2 electrolyzer to convert CO2 to Formic Acid | Dioxide Materials. (n.d.). Retrieved April 13, 2025, from https://dioxidematerials.com/product/complete-5- cm2-fa- electrolyzer/?bundle_quantity_275=1&bundle_quantity_280=1&bundle_quantity_33 8=1&bundle_quantity_339=1&bundle_quantity_340=1&bundle_quantity_341=1&bu ndle_quantity_342=1 Deublein, D., & Steinhauser, A. (2011). Biogas from Waste and Renewable Resources: An Introduction. John Wiley & Sons. Dwilaksono, F. (2022). Evaluation of operational performance and environmental impact of a commercial scale anaerobic digester utilizing multiple feedstocks. Michigan State University. Edwards, J., Othman, M., & Burn, S. (2015). A review of policy drivers and barriers for the use of anaerobic digestion in Europe, the United States and Australia. Renewable and Sustainable Energy Reviews, 52, 815–828. https://doi.org/10.1016/j.rser.2015.07.112 Emerson, Emilia Maria. (2024). THE EFFECT OF HYDRAULIC RETENTION TIME ON RECOVERABLE AMMONIA, VIRUS-PARTICLE ASSOCIATION, AND BACTERIAL, FUNGAL, AND VIRAL POPULATIONS IN A BENCH SCALE ACTIVATED SLUDGE MUNICIPAL WASTEWATER TREATMENT SYSTEM [Michigan State University]. https://doi.org/10.25335/X460-HG08 Feng, L., He, S., Gao, Z., Zhao, W., Jiang, J., Zhao, Q., & Wei, L. (2023). Mechanisms, performance, and the impact on microbial structure of direct interspecies electron transfer for enhancing anaerobic digestion-A review. Science of The Total Environment, 862, 160813. https://doi.org/10.1016/j.scitotenv.2022.160813 65 Gao, M., Zhang, S., Ma, X., Guan, W., Song, N., Wang, Q., & Wu, C. (2020). EWect of yeast addition on the biogas production performance of a food waste anaerobic digestion system. Royal Society Open Science, 7(8), 200443. https://doi.org/10.1098/rsos.200443 Hjort-Gregersen, K., Blandford, D., & Gooch, C. A. (2011). Biogas from Farm-based Biomass Sources Developments in Europe and the US Le biogaz dérivé de la biomasse d’origine agricole Développements en Europe et aux États-Unis Biogas aus Biomasse von landwirtschaftlichen Betrieben Entwicklungen in Europa und den USA. EuroChoices, 10(3), 18–23. https://doi.org/10.1111/j.1746-692X.2011.00211.x Joó, F. (2008). Breakthroughs in Hydrogen Storage—Formic Acid as a Sustainable Storage Material for Hydrogen. ChemSusChem, 1(10), 805–808. https://doi.org/10.1002/cssc.200800133 Li, Y., Cai, G., Pan, X., Lv, N., Feng, L., Zhu, G., Lv, Z., & Ye, Z. (2024). Promoting or Inhibiting: New Insights into the Role of Formate in Syntrophic Propionate Metabolism. Water, 16(24), Article 24. https://doi.org/10.3390/w16243551 Ma, Y., Xu, R., Wu, X., Wu, Y., Zhao, L., Wang, G., Li, F., & Shi, Z. (2024). Progress in Catalysts for Formic Acid Production by Electrochemical Reduction of Carbon Dioxide. Topics in Current Chemistry, 383(1), 2. https://doi.org/10.1007/s41061-024-00487-4 Marietou, A. (2021). Chapter Two—Sulfate reducing microorganisms in high temperature oil reservoirs. In G. M. Gadd & S. Sariaslani (Eds.), Advances in Applied Microbiology (Vol. 116, pp. 99–131). Academic Press. https://doi.org/10.1016/bs.aambs.2021.03.004 Meegoda, J. N., Li, B., Patel, K., & Wang, L. B. (2018). A Review of the Processes, Parameters, and Optimization of Anaerobic Digestion. International Journal of Environmental Research and Public Health, 15(10), Article 10. https://doi.org/10.3390/ijerph15102224 Mokashi, K., Shetty, V., George, S. A., & Sibi, G. (2016). Sodium Bicarbonate as Inorganic Carbon Source for Higher Biomass and Lipid Production Integrated Carbon Capture in Chlorella vulgaris. Achievements in the Life Sciences, 10(1), 111–117. https://doi.org/10.1016/j.als.2016.05.011 Moret, S., Dyson, P. J., & Laurenczy, G. (2014). Direct synthesis of formic acid from carbon dioxide by hydrogenation in acidic media. Nature Communications, 5(1), 4017. https://doi.org/10.1038/ncomms5017 66 Mu, L., Wang, Y., Xu, F., Li, J., Tao, J., Sun, Y., Song, Y., Duan, Z., Li, S., & Chen, G. (2023). Emerging Strategies for Enhancing Propionate Conversion in Anaerobic Digestion: A Review. Molecules, 28(9), 3883. https://doi.org/10.3390/molecules28093883 Odey, G., Adelodun, B., Kim, S.-H., & Choi, K.-S. (2021). Status of Environmental Life Cycle Assessment (LCA): A Case Study of South Korea. Sustainability, 13(11), Article 11. https://doi.org/10.3390/su13116234 Owen, J. J., & Silver, W. L. (2015). Greenhouse gas emissions from dairy manure management: A review of field-based studies. Global Change Biology, 21(2), 550– 565. https://doi.org/10.1111/gcb.12687 Reap, J., Roman, F., Duncan, S., & Bras, B. (2008). A survey of unresolved problems in life cycle assessment. The International Journal of Life Cycle Assessment, 13(4), 290– 300. https://doi.org/10.1007/s11367-008-0008-x Shen, F., Li, H., Wu, X., Wang, Y., & Zhang, Q. (2018). EWect of organic loading rate on anaerobic co-digestion of rice straw and pig manure with or without biological pretreatment. Bioresource Technology, 250, 155–162. https://doi.org/10.1016/j.biortech.2017.11.037 Shen, L., Zhao, Q., Wu, X., Li, X., Li, Q., & Wang, Y. (2016). Interspecies electron transfer in syntrophic methanogenic consortia: From cultures to bioreactors. Renewable and Sustainable Energy Reviews, 54, 1358–1367. https://doi.org/10.1016/j.rser.2015.10.102 Tiwari, S. B., Dubey, M., Ahmed, B., Gahlot, P., Khan, A. A., Rajpal, A., Kazmi, A. A., & Tyagi, V. K. (2021). Carbon-based conductive materials facilitated anaerobic co-digestion of agro waste under thermophilic conditions. Waste Management (New York, N.Y.), 124, 17–25. https://doi.org/10.1016/j.wasman.2021.01.032 Turnbull, J. (2004). Greenhouse Gas Emission Reduction Associated with Livestock Waste Management Systems: A Case Study for the Langerwerf Dairy Waste Management System. https://www.semanticscholar.org/paper/Greenhouse-Gas-Emission- Reduction-Associated-with-A- Turnbull/017d98626ba303c987b0bd05441783d1e80944e0/figure/19 US EPA, O. (2016, September 8). Greenhouse Gas Emissions Estimation Methodologies for Biogenic Emissions from Selected Source Categories [Overviews and Factsheets]. https://www.epa.gov/air-emissions-factors-and-quantification/greenhouse-gas- emissions-estimation-methodologies-biogenic 67 US EPA, O. (2024). AgSTAR Data and Trends [Data and Tools]. https://www.epa.gov/agstar/agstar-data-and-trends Valença, R. B., Santos, L. A. dos, Firmo, A. L. B., Silva, L. C. S. da, Lucena, T. V. de, Santos, A. F. de M. S., & Jucá, J. F. T. (2021). Influence of sodium bicarbonate (NaHCO3) on the methane generation potential of organic food waste. Journal of Cleaner Production, 317, 128390. https://doi.org/10.1016/j.jclepro.2021.128390 Xu, M., Uludag-Demirer, S., Fang, D., Zhou, L., Liu, Y., & Liao, W. (2021). EWects of Pyrogenic Carbonaceous Materials on Anaerobic Digestion of a Nitrogen-Rich Organic Waste—Swine Manure. Energy & Fuels, 35(3), 2282–2292. https://doi.org/10.1021/acs.energyfuels.0c03390 Xu, M., Uludag-Demirer, S., Liu, Y., & Liao, W. (2025). Improving Anaerobic Digestion EWiciency of Animal Manure Through Ball Milling Pretreatment. Agronomy, 15(2), Article 2. https://doi.org/10.3390/agronomy15020305 Yu, J., Zhao, Y., Zhang, H., Hua, B., Yuan, X., Zhu, W., Wang, X., & Cui, Z. (2017). Hydrolysis and acidification of agricultural waste in a non-airtight system: EWect of solid content, temperature, and mixing mode. Waste Management, 59, 487–497. https://doi.org/10.1016/j.wasman.2016.10.019 Zhang, Y., Li, C., Yuan, Z., Wang, R., Angelidaki, I., & Zhu, G. (2023). Syntrophy mechanism, microbial population, and process optimization for volatile fatty acids metabolism in anaerobic digestion. Chemical Engineering Journal, 452, 139137. https://doi.org/10.1016/j.cej.2022.139137 68 APPENDIX A: DETAILED STATISTICAL RESULTS Table 6.1. ANOVA Results Summary Comparison ANOVA Test P-value Significant? Total gas production Ordinary Gas quality Ordinary 0.0114 Yes 1.65e-5 Yes Total gas production Kruskal-Wallis 0.08148 No Gas quality Kruskal-Wallis 0.06829 No pH, HRT 4 Ordinary 0.00134 Yes pH, HRT 4 Kruskal-Wallis 0.07798 No COD, HRT 4 Ordinary 6.1e-08 Yes COD, HRT 4 Kruskal-Wallis 0.06829 No Total Solids, HRT 4 Ordinary 2.98e-05 Yes Total Solids, HRT 4 Kruskal-Wallis 0.09705 No Volatile Solids, HRT Ordinary 0.000312 Yes 4 Volatile Solids, HRT Kruskal-Wallis 0.09292 No 4 69 Table 6.2. Tukey's HSD on Total Biogas Production Ordinary ANOVA Results Summary Comparison P-value Significant? Bicarbonate vs AC C+F vs AC Control vs AC Formate vs AC C+F vs Bicarbonate Control vs Bicarbonate Formate vs Bicarbonate Control vs C+F Formate vs C+F Formate vs Control 0.96 No 0.65 No 0.15 No 0.084 No 0.94 No 0.077 No 0.17 No 0.038 Yes 0.36 No 0.007 Yes Table 6.3. Tukey's HSD on Biogas Quality Ordinary ANOVA Results Summary Comparison P-value Significant? Bicarbonate vs AC 0.000022 Yes C+F vs AC Control vs AC Formate vs AC C+F vs Bicarbonate Control vs Bicarbonate 0.00043 Yes 0.00018 Yes 0.48 No 0.00078 Yes 0.0026 Yes Formate vs Bicarbonate 0.000029 Yes 70 Table 6.3 (cont’d) Control vs C+F Formate vs C+F Formate vs Control 0.27 No 0.00093 Yes 0.00034 Yes Table 6.4. Tukey's HSD on pH Ordinary ANOVA Results Summary Comparison P-value Significant? Bicarbonate vs AC C+F vs AC Control vs AC Formate vs AC C+F vs Bicarbonate Control vs Bicarbonate Formate vs Bicarbonate Control vs C+F Formate vs C+F Formate vs Control 0.0044 Yes 0.038 Yes 0.017 Yes 0.52 No 0.16 No 0.42 No 0.0017 Yes 0.87 No 0.0098 Yes 0.0051 Yes Table 6.5. Tukey's HSD on COD Ordinary ANOVA Results Summary Comparison P-value Significant? Bicarbonate vs AC C+F vs AC 0.0000005 Yes 0.0000005 Yes 71 Table 6.5 (cont’d) Control vs AC Formate vs AC C+F vs Bicarbonate 0.16 No 0.00011 Yes 0.107 No Control vs Bicarbonate 0.0000005 Yes Formate vs Bicarbonate 0.0000011 Yes Control vs C+F Formate vs C+F Formate vs Control 0.0000006 Yes 0.0000018 Yes 0.00032 Yes Table 6.6. Tukey's HSD on Total Solids Ordinary ANOVA Results Summary Comparison P-value Significant? Bicarbonate vs AC C+F vs AC Control vs AC Formate vs AC C+F vs Bicarbonate Control vs Bicarbonate Formate vs Bicarbonate Control vs C+F Formate vs C+F 0.000109 Yes 0.00012 Yes 0.997 No 0.797 No 0.995 No 0.00012 Yes 0.00016 Yes 0.00014 Yes 0.00018 Yes 72 Table 6.6 (cont’d) Formate vs Control 0.92 No Table 6.7. Tukey's HSD on Volatile Solids Ordinary ANOVA Results Summary Comparison P-value Significant? Bicarbonate vs AC C+F vs AC Control vs AC Formate vs AC C+F vs Bicarbonate Control vs Bicarbonate Formate vs Bicarbonate Control vs C+F Formate vs C+F Formate vs Control 0.0012 Yes 0.0014 Yes 0.49 No 0.94 No 0.994 No 0.00296 Yes 0.0008 Yes 0.0037 Yes 0.00097 Yes 0.23 No 73 B.1 Grouped Bar Charts Generation (COD, TS, VS, GC) APPENDIX B: PYTHON AND R CODE #plot to generate COD data import numpy as np import matplotlib.pyplot as plt def generate_grouped_bar_chart(): # Define parameters num_groups = 4 num_bars_per_group = 5 bar_width = 0.15 # Width of individual bars # Manually set all data values to 1 without using np.ones or a loop data = [14860, 9213.4, 11236.7, 9750,10916.7],[9610,8415,10075,10770,10315],[6566.7,5243.4,6493.4,5280,6666.7],[696 6.666667,6870,8846.666667,7393.333333,8736.666667] # Define the positions of the groups on the x-axis group_positions = np.arange(num_groups) # Define colors and labels for the bars colors = ['b', 'g', 'r', 'c', 'm'] labels = ['Control', 'Activated Carbon', 'Bicarbonate', 'Formate', 'Bicarbonate + Formate'] # Create figure and axis fig, ax = plt.subplots(figsize=(8, 6)) # Plot each bar in the group 74 for i in range(num_bars_per_group): ax.bar(group_positions + i * bar_width, [row[i] for row in data], width=bar_width, color=colors[i], label=labels[i]) # Formatting ax.set_xlabel('HRT') ax.set_ylabel('COD (mg/L)') ax.set_title('Mean COD per HRT') ax.set_xticks(group_positions + (num_bars_per_group - 1) * bar_width / 2) ax.set_xticklabels(['HRT 1', 'HRT 2', 'HRT 3', 'HRT 4']) ax.legend() # Show plot plt.show() # Run the function generate_grouped_bar_chart() Line Chart Generation (Cumulative Gas Production) import numpy as np import matplotlib.pyplot as plt import pandas as pd # Load data from CSV file data = pd.read_csv('formate_gas_prod.csv',header=None) # Generate x values x = data.iloc[0, :] 75 # Define y values for each line y1 = data.iloc[1, :] # Adjust column indices as needed y2 = data.iloc[2, :] y3 = data.iloc[3, :] y4 = data.iloc[4, :] y5 = data.iloc[5, :] y6 = data.iloc[6, :] y7 = data.iloc[7, :] y8 = data.iloc[8, :] y9 = data.iloc[9, :] y10 = data.iloc[10, :] plt.figure(figsize=(15, 10), dpi=600) # Plot each line separately plt.plot(x, y1, label='Control 1') plt.plot(x, y2, label='Control 2') plt.plot(x, y3, label='AC 1') plt.plot(x, y4, label='AC 2') plt.plot(x, y5, label='Bicarbonate 1') plt.plot(x, y6, label='Bicarbonate 2') plt.plot(x, y7, label='Formate 1') plt.plot(x, y8, label='Formate 2') plt.plot(x, y9, label='C+F 1') 76 plt.plot(x, y10, label='C+F 2') # Add title and labels plt.title('Cumulative Gas Production') plt.xlabel('Days of Operation') plt.ylabel('Total Biogas (mL)') plt.legend() plt.grid(True) # Show the plot plt.show() 5.2.1 B.2 Ordinary ANOVA and Tukey’s HSD Test R Code pH <- read.csv(file = "ph_data.csv", header = TRUE, sep = ",") pH reactor <- c(rep('Control', 2),rep('AC', 2),rep('Bicarb', 2),rep('Formate', 2),rep('C+F', 2)) pH_4 <-c(pH$Control, pH$AC, pH$Bicarb, pH$Formate, pH$C.F) df_pH <- data.frame(reactor, pH_4) ph.aov <- aov(pH_4 ~ reactor, data = df_pH) summary(ph.aov) cod_4 <- c(cod$Control, cod$AC, cod$Bicarb, cod$Formate, cod$C.F) df_cod <- data.frame(reactor, cod_4) cod.aov <- aov(cod_4 ~ reactor, data = df_cod) summary(cod.aov) ts_4 <- c(ts$Control, ts$AC, ts$Bicarb, ts$Formate, ts$C.F) 77 df_ts <- data.frame(reactor, ts_4) ts.aov <- aov(ts_4 ~ reactor, data = df_ts) summary(ts.aov) vs_4 <- c(vs$Control, vs$AC, vs$Bicarb, vs$Formate, vs$C.F) df_vs <- data.frame(reactor, vs_4) vs.aov <- aov(vs_4 ~ reactor, data = df_vs) summary(vs.aov) TukeyHSD(ph.aov) TukeyHSD(cod.aov) TukeyHSD(ts.aov) TukeyHSD(vs.aov) 5.2.2 B.3 Kruskal-Wallis Test R Code #arranging the data for each measurement gasProd <- read.csv(file = "gas_prod_data.csv", header = TRUE, sep = ",") print(gasProd) gasConc <- read.csv(file = "gas_conc_data.csv", header = TRUE, sep = ",") print(gasConc) pH <- read.csv(file = "ph_data.csv", header = TRUE, sep = ",") print(pH) # cod <- read.csv(file = "cod_data.csv", header = TRUE, sep = ",") print(cod) 78 ts <- read.csv(file = "ts_data.csv", header = TRUE, sep = ",") print(ts) vs <- read.csv(file = "vs_data.csv", header = TRUE, sep = ",") print(vs) # run KW test for each - if p < a, reject the null kwTest <- kruskal.test(gasProd) kwTest kwTest_conc <- kruskal.test(gasConc) kwTest_conc kwTest_pH <-kruskal.test(pH) kwTest_pH kwTest_cod <-kruskal.test(cod) kwTest_cod kwTest_ts <-kruskal.test(ts) kwTest_ts kwTest_vs <-kruskal.test(vs) kwTest_vs 79 APPENDIX C: DETAILED DATA Table 6.8. Biogas Production Values Final Biogas Production (mL) Reactor Control 1 Control 2 Activated Carbon 1 Activated Cabon 2 Bicarbonate 1 Bicarbonate 2 Formate 1 Formate 2 Bicarbonate/Formate 1 Bicarbonate/Formate 2 Table 6.9. Mean Biogas Quality Values HRT Reactor 1 Control Mean Value Activated Carbon Bicarbonate Formate Bicarbonate + Formate 2 Control 80 9365 9600 11105 11890 11800 12125 14470 13350 11635 13310 0.58 0.59 0.49 0.61 0.57 0.52 Table 6.9 (cont’d) Activated Carbon Bicarbonate Formate Bicarbonate + Formate 3 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 4 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 0.56 0.56 0.58 0.54 0.51 0.53 0.50 0.54 0.51 0.60 0.63 0.59 0.60 0.60 Table 6.10. Methane Production per HRT (mL) Control AC Bicarbonate Formate HRT-1 HRT-2 HRT-3 HRT-4 1221 1369 1434 1149 1603 1806 1443 1507 2173 1930 1939 1983 1654 1890 1739 1383 81 Formate + Bicarbonate 1847 1708 1706 1662 Table 6.10 (cont’d) Total CH4 5172 6666 6358 8025 6923 Percent Increase vs Control 28.88% 22.94% 55.17% 33.86% Table 6.11. Complete Table of pH Results Control AC Bicarbonate Formate Formate + Bicarbonate 8.06 8.15 7.87 7.95 7.97 8.53 8.38 8.60 8.43 8.55 8.66 8.50 8.69 8.46 8.58 8.43 8.32 8.50 8.38 8.26 8.26 8.14 8.37 8.06 8.24 8.22 8.07 8.35 8.09 8.23 8.07 7.97 8.20 7.95 8.10 8.17 8.07 8.30 8.01 8.20 8.18 8.05 8.28 7.97 8.18 Day 6 Day 13 Day 20 Day 27 Day 34 Day 40 Day 48 Day 55 Day 64 Day 69 8.13 7.90 8.14 7.86 8.04 82 Table 6.11 (cont’d) Day 76 8.19 8.01 8.27 7.96 8.18 Table 6.12. Mean pH Measurement Values per HRT HRT Reactor Mean Value 1 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 2 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 3 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 4 Control 83 8.42 8.34 8.38 8.28 8.37 8.35 8.23 8.44 8.22 8.25 8.15 8.04 8.28 8.02 8.18 8.17 Table 6.12 (cont’d) Activated Carbon Bicarbonate Formate Bicarbonate + Formate 7.98 8.23 7.93 8.13 Table 6.13. Complete Table of COD Results (mg/L) Control AC Bicarbonate Formate 14490 12360 8040 7000 10030 8420 9420 7500 Formate + Bicarbonate 9730 8500 17730 12600 15260 12330 14520 9820 9400 6910 7110 5680 7010 7430 6460 7430 9400 5120 5650 4960 6540 7280 6790 8850 9740 7730 11300 11800 12900 6450 6730 6300 7350 9800 9390 5230 5800 4810 7120 7740 7320 6010 7630 6360 8300 9260 8650 Day 6 Day 13 Day 20 Day 27 Day 34 Day 40 Day 48 Day 55 Day 64 Day 69 Day 76 Table 6.14. Mean COD Measurement Values HRT Reactor Mean Value 84 Table 6.14 (cont’d) 1 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 2 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 3 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 4 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 85 14860 9210 11240 9750 10920 9610 8420 10080 10770 10320 6570 5340 6490 5280 6670 6970 6870 8850 7390 8740 Table 6.15. Complete Table of Total Solids Results (mg/L) Control AC Bicarbonate Formate 19155 16045 18355 17890 Formate + Bicarbonate 19150 15950 15170 18995 16540 17370 14915 14020 16200 14415 16105 14720 13970 17345 16925 18120 14035 13310 17210 14730 17295 12350 12640 15375 13305 16285 14155 12440 16560 12330 17000 14725 13630 18070 13795 18675 19985 17465 28270 18460 27855 11910 13355 17840 14180 17795 14700 15395 19405 14960 19400 Day 6 Day 13 Day 20 Day 27 Day 34 Day 40 Day 48 Day 55 Day 64 Day 69 Day 76 Table 6.16. Mean Total Solids Value Measurements per HRT HRT Reactor Mean Value 1 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 2 Control 86 16.67 15.08 17.85 16.28 17.54 14.38 Table 6.16 (cont’d) Activated Carbon Bicarbonate Formate Bicarbonate + Formate 3 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 4 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 13.64 17.28 15.83 17.71 13.74 12.90 16.67 23.24 27.32 15.53 15.41 21.84 15.87 21.68 Table 6.17. Complete Table of Volatile Solids Results Control AC Bicarbonate Formate 19155 16045 18355 17890 Formate + Bicarbonate 19150 15950 15170 18995 16540 17370 14915 14020 16200 14415 16105 14720 13970 17345 16925 18120 Day 6 Day 13 Day 20 Day 27 87 Table 6.17 (cont’d) Day 34 Day 40 Day 48 Day 55 Day 64 Day 69 Day 76 14035 13310 17210 14730 17295 12350 12640 15375 13305 16285 14155 12440 16560 12330 17000 14725 13630 18070 13795 18675 19985 17465 28270 18460 27855 11910 13355 17840 14180 17795 14700 15395 19405 14960 19400 Table 6.18. Mean VS Measurement Values per HRT HRT Reactor Mean Value 1 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 2 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 3 Control 88 10.45 9.25 10.87 10.33 10.84 8.23 7.70 9.19 8.61 8.53 7.55 Table 6.18 (cont’d) Activated Carbon Bicarbonate Formate Bicarbonate + Formate 4 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate Table 6.19. Average Formic Acid Concentrations HRT Reactor 2 Control Measurement (mg/L) Activated Carbon Bicarbonate Formate Bicarbonate + Formate 3 Control Activated Carbon Bicarbonate Formate 89 6.12 7.43 6.67 7.29 11.63 11.05 14.27 10.80 14.14 2.58 2.48 8.01 2.55 2.34 2.57 1.55 1.28 2.50 Table 6.19 (cont’d) Bicarbonate + Formate 4 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate Table 6.20. Average Acetic Acid Concentration HRT Reactor 2 Control Measurement (mg/L) Activated Carbon Bicarbonate Formate Bicarbonate + Formate 3 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 4 Control 90 2.39 2.46 1.14 1.88 0.930 2.11 963 851 778 735 718 456 445 433 416 403 240 Table 6.20 (cont’d) Activated Carbon Bicarbonate Formate Bicarbonate + Formate Table 6.21. Mean Propionic Acid Concentrations HRT Reactor 2 Control Measurement (mg/L) Activated Carbon Bicarbonate Formate Bicarbonate + Formate 3 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 4 Control Activated Carbon Bicarbonate Formate 91 265 252 253 236 324 276 276 247 219 96.8 103 104 94.8 80.0 60.2 55.3 68.2 20.7 Table 6.21 (cont’d) Bicarbonate + Formate 20.0 Table 6.22. Mean Butyric Acid Concentrations HRT Reactor 2 Control Measurement (mg/L) Activated Carbon Bicarbonate Formate Bicarbonate + Formate 3 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 4 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 92 576 527 490 451 458 294 290 285 271 274 237 224 227 164 155 Table 6.23. Mean Isobutyric Acid Concentrations HRT Reactor 2 Control Measurement (mg/L) Activated Carbon Bicarbonate Formate Bicarbonate + Formate 3 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 4 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate Table 6.24. Mean Isovaleric Acid Concentration Measurement (mg/L) HRT Reactor 2 Control Activated Carbon 93 198 174 164 150 138 59.2 43.4 38.7 44.6 33.4 29.9 18.2 19.4 0 0 256 226 212 191 196 83.0 86.4 72.8 76.8 75.5 45.6 39.4 45.6 10.2 2.04 Table 6.24 (cont’d) Bicarbonate Formate Bicarbonate + Formate 3 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 4 Control Activated Carbon Bicarbonate Formate Bicarbonate + Formate 94 APPENDIX D: SUPPLEMENTARY FIGURES Figure 5.2.1. Average Daily Biogas Production Rate 95 Figure 5.2.2. Average Daily COD Measurements 96 Figure 5.2.3. Average Daily pH Measurements 97 Figure 5.2.4. Average Daily Total Solids Measurements 98 Figure 5.2.5. Average Daily Volatile Solids Measurements 99