1 Influence of co digestion and water content on biodigester gas production Madeleine S. Gregory Department of Molecular Environmental Biology University of California, Berkeley EAP Tropical Biology and Conservation Spring 2019 7 June 2019 Abstract Excess livestock wast e can pollute water sources. Limited access to clean energy leads to burning of fossil fuels such as propane. Biodigesters treat animal manure, converting it into a nutrient rich fertilizer and methane gas for fuel. I tested different substrate mixtures (i ncluding pig manure co digested with goat manure, chicken manure, and whey) in a lab setting, measuring methanogenic potential of each treatment in order to test one in the biodigester. I also tested the effect of water to waste ratio on gas production. Tr eatments containing pig manure, either alone or co digested with other manure types, produced the most gas. Mixtures containing whey produced the least gas. In the field, I tested a mixture of pig and goat manure against the control conditions of 6:1 water to pig manure ratio. The experimental treatment significantly increased flame height but did not significantly increase the total burn time. Influencia de codigestiÂ—n y la cantidad de agua en la producciÂ—n de gas en un biodigestor Resumen Demasiada boÂ– iga de animales puede contaminar fuentes de agua. Sin energÂ’a limpia, las personas usarÂ‡n combustibles fÂ—siles. En los biodigestores se utiliza boÂ–iga de animales para obtener gas metano y fertilizantes. En mi experimento, dispuse tratamientos en botellas en las que mezclÂŽ diferentes proporciones de boÂ–iga y agua (boÂ–iga de cerdo con boÂ–iga de cabra, cuita de gallina, o suero). Los globos que coloquÂŽ sobre las botellas se fueron llenando con gas metano, y medÂ’ cuÂ‡nto se inflaron cada dÂ’a. Los tratamientos c on boÂ–iga de cerdo solo, o con otras boÂ–igas, produjeron la mayor cantidad de gas. Los tratamientos con boÂ–iga y suero produjeron menos gas. DespuÂŽs del experimento en el laboratorio, yo puse boÂ–iga de cabra en el biodigestor con boÂ–iga de cerdo. Este trat amiento tiene un menor proporciÂ—n entre agua y boÂ–iga, y tambiÂŽn una mezcla de boÂ–iga diferente de el control. En el tratamiento experimental, el gas producido aumentÂ— la altura de la llama en una cocina, y ÂŽste se quemÂ— durante mÂ‡s tiempo. Sin embargo, es e tiempo de quemarse no fue significativamente mayor. Introduction Waste disposal and energy access are constant problems to rural communities worldwide. Biodigesters present a unique solution to the problem of waste treatment: they use anaerobic methods to break down waste, such as animal manure or food waste, to produce methane gas, solid digestate, and liquid supernatant. This provides both waste disposal and renewable energy, while creating a nutrient rich effluent that can be used as fertilizer (Lv et al., 2018). Biodigesters vary widely in their scale, influent
Biodigester efficiency Gregory 2 composition, and efficiency. This study will focus on biogas production (methanogenesis) and select effluent qualities. Life Monteverde, a family run coffee farm in CaÂ–itas de Abangares, Costa Rica, uses a "salchicha" biodigester, a small polyethylene tube. This is the most commonly used biodigester in Central America because of its low cost and easy implementation (Hojnacki et al., 2011). Life Monteverde uses pig manure in their biodigester, th ough the farm also has goats and chickens. Pig manure has been shown to have higher methanogenic potential than goat manure, but the effects of mixing pig manure with other manure types Ã‘ such as goat or chicken Ã‘ is less well studied (Achinas et al., 2017 , Kafle et al., 2016). However, this process of co digestion (or inputting multiple substrates) has been shown to increase biogas production by balancing some chemical properties, including pH and C/N ratio (Kavuma, 2013). Currently, Life Monteverde mostly uses the biodigester for educational purposes. Though some gas is generated, it is a relatively small amount, and the flame produced is small. In an interview with Jerson SantamarÂ’a from Life Monteverde, he attributed this to limited storage capacity and problems with the water to waste ratio. The ideal ratio, he said, was approximately 3 L water : 1 kg manure. Jerson SantamarÂ’a said that they should input 21 L of water, as the pigs produce around 7 kg of manure each day. However, Life does not explicitly measure the amount of water added each day. Instead, they wash the pig pens and send that water and waste into the biodigester. This leads to a much higher volume of water than the theoretical ideal. The effects of this dilution on the methanogenic potenti al is not well known. Additionally, the composition of the biodigester input will affect many qualities in the effluent, which will ultimately be used as fertilizer. The ideal pH for most crops is slightly acidic, from 5.5 6.5 depending on the crop (Cropnu trition, 2019). Both pig and goat manure are acidic, but goat manure has a lower pH than pig manure (Ano and Ubochi, 2007). From an efficiency perspective, the pH of the biodigester can affect the gas production, the ideal pH range for anaerobic digestion is 6.8 to 7.2 (Cioabla et al., 2012). Another parameter of interest is biochemical oxygen demand (BOD), which is an indication of the rate of organic decomposition of microbes. It is an important indicator of how much organic load is still in the effluent prior to discharge. Adding more manure to the biodigester could increase the organic load, increasing the microbial activity in the effluent. Besides pig manure, Monteverde has many biodigester substrates at their disposal, including goat manure, chicken manure, and the cheese by product whey. Co digesting pig manure with one or more of these substrates could lead to equal or increased gas production. This can help reaffirm Life's commitment to their biodigester, incentivizing maintenance and investment. D uring the lab phase, I seek to understand the effects of waste composition and water to waste ratio on biodigester efficiency. I will use these results to inform the field experiment, determining how these treatments affect biogas production at Life. Ideal ly, I can provide suggestions for future biodigester use at Life and other local farms. Materials and Methods The study was conducted at Finca Life Monteverde, CaÂ–itas, Guanacaste, Costa Rica (N 10.32 ! , W 84.84 ! ) from 10 May 2019 through 24 May 2019. L ife Monteverde is a 42 acre coffee farm in a tropical pre montane region.
Biodigester efficiency Gregory 3 Lab experiment: My lab experiment compared the methanogenic potential of different co digestion treatments. I tested 7 mixtures, each with 5 replicates ( Table 1 ). Because these wer e designed to replicate the existing plug flow system of the biodigester, the ratio of input (volume of substrates) to inoculate (bacteria rich liquid content of the digester) mimics the ratio of influent to the current biodigester. Thus, all replicates ha ve the same quantity of inoculate and varying inputs ( Table 1 ). All of the following tests are based on volume ratios. They are named with the first letter of their inputs. The first two treatments tested the effects of water to waste ratio, with the con trol including a 6:1 water to pig manure dilution and Treatment P including a 3:1 water to pig manure dilution. Assuming that the 3:1 water to manure ratio was more effective, the rest of the tests follow this dilution, changing the manure composition inst ead. Treatment PG had 1:1 pig to goat manure, Treatment PC 1:1 pig to chicken, Treatment PW 1:1 pig manure to whey, Treatment PGW 1:1:1 pig to goat to whey, and one Treatment GCW 1:1:1 goat to chicken to whey ( Table 1 ). I first calculated the densities of each manure type by weighing 100 mL of the substance. I used this conversion to determine the weight of each manure type needed, using a digital scale to measure these out. These mixtures were placed in equal volume glass bottles with 315.6 mL inoculate. I placed balloons on top of each vial, sealing these balloons with electric tape. To measure the inflation of these balloons, I created a scale that measures balloon fill from 0 to 10 ( Fig. 1 ). Each day, I measured the ambient temperature while I performed these measurements at 9am. Each day, I calculated an average fill for each treatment type. I then compared these averages for the first 5 days of every treatment type, as the balloons began deflating around this point. I performed an ANOVA test to compare the differences in fill. I also performed a Tukey test to identify which treatments are statistically different than others. To measure the effects of C/N ratio, I performed a regression analysis. Field (control): Throughout the first week of this lab portion, I also collected baseline data on the functioning of the biodigester. To evaluate baseline biogas production, I measured the daily burn time and initial height of the flame. I used salt to color the flame, measuring the initial height with a tape measure. I measured burn time three times a day, at 8 am, 12 pm, and 3 pm. I began timing when I turned the flame on, stopping when it blew out. I then added the burn times together to get a total burn time each day, a proxy for total gas production. I av eraged the initial flame height of each of the three burns. I measured the pH of the biodigester influent and effluent. I conducted a BOD 5 test by measuring the dissolved oxygen in the influent and effluent respectively, sealing these samples in tubes, an d then measured the dissolved oxygen again after 5 days. I created 3 replicates for both the influent and effluent treatment. I used a t test to compare the average flame height and average daily burn time of the treatments. Field (experimental)
Biodigester efficiency Gregory 4 After o ne week of the lab testing, I determined that I would implement Treatment B (goat and pig manure) to test the effects of co digestion at a larger scale. I chose this treatment to test the merits of co digestion at a larger scale, as it was the best perform ing co digestion treatment. Each morning at 6am, I weighed out 1.4 kilograms (7 liters) of goat manure and placed in the biodigester intake prior to the pig stalls be washed upstream by the caretaker. By doing so, the desired pig manure, goat manure and wa ter ratio and mixing was achieved. I measured the same parameters (flame height and burn time) three times a day, at 8 am, 12 pm, and 3 pm. After a week of this treatment, I measured the effluent and influent pH and performed another BOD 5 test. I used a t test to compare the average flame height and average daily burn time of the treatments. I also used a t test to compare the BOD 5 results from the influent and effluent. Figure 1 : Balloon fill scale. This hand drawn scale was used each day to measure th e numerical fill level of each balloon. The radius of each concentric circle is approximately .5 cm larger than the previous one (not to scale).
Biodigester efficiency Gregory 5 Figure 2 : The general flow of Life's biodigester. The influent comes from the pig pens, the effluent goes to the garden, and the gas goes to a stove. Table 1 : Experimental setup. The treatments are named with acronyms based on their inputs. Input Ratio Inoculate (ml) Water (mL) Pig manure (mL) Goat manure (mL) Chicken manure (mL) Whey (mL) Control 6:1 water: pig manure, bacteria* 315.6 5.16 .84 0 0 0 Treatment P 3:1 water: pig manure, bacteria 315.6 4.5 1.5 0 0 0 Treatment PG 3:1 water: waste, 1:1 pig: goat manure, bacteria 315.6 4.5 .75 .75 0 0 Treatment PC 3:1 water: waste, 1:1 pig: chicken manure, bacteria 315.6 4.5 .75 0 .75 0 Treatment PW 3:1 water: waste, 1:1 pig: whey, bacteria 315.6 4.5 .75 0 0 1.5 Treatment PGW 3:1 water: waste, 1:1:1 pig: goat: whey, bacteria 315.6 4.5 .5 .5 0 1 Treatment GCW 3:1 water: waste, 1:1:1 goat: 315.6 4.5 0 .5 .5 1
Biodigester efficiency Gregory 6 chicken: whey, bacteria *bacteria is from the biodigester, and will serve as an inoculate for the tests to begin producing methane Results Lab experiment: All of the treatments produced some amount of gas, with the three treatments that included whey as a substrate produced less gas than the treatments with only animal manure. All 7 treatments produced the most gas the first few days, tapering off substantially after day 3 or 4 ( Fig. 3 ). Those with whey took longer to reach their maximum total gas production. Diameter measurements were taken of each replicate each day, and these closely matched the rating measurements in their relative sizes and increases. Treatments with higher C/N ratios yielded higher gas production (R 2 = .68, p=.023) ( Fig. 4 ). Figure 3 : Production of gas in each of the 7 treatments, stopping at the maximum gas production for each treatment. Production of gas is measured by average size each day, as measured by the balloon scale ( Fig. 1 ).
Biodigester efficiency Gregory 7 Figure 4 : The relationship between the C/N ratio of the treatments and their gas production. Each point is marked with the corresponding treatment. As C/N ratio increases, the gas production also increases (R 2 =.67). This relationship is statistically significant (p=.o23 ). Though results differed in their average gas production, with Treatments P and PG showing higher average gas production, only Treatment PGW was significantly different from the Control. The treatments were sorted into groups of statistically similar re sults (Tukey test). Group "A" includes treatments P, PG, PC, and Control. Group "B" includes treatments PC, PW, GCW, and Control. Group "C" includes PW, PGW, and GCW ( Fig. 5 ).
Biodigester efficiency Gregory 8 Figure 5 : The difference in average gas production of each lab treatment type . The first 5 days of each treatment are considered in this graph. Treatments marked with the same letter are statistically similar. Field experiment: Adding goat manure to the biodigester increased the daily average flame height. Flame height was signifi cantly increased (p= .0037). The flame height was also more consistent in the experimental treatment, with less variation between days. Though the average flame burn time increased with the experimental treatment, it was not a statistically significant inc rease (p=.159). The experimental week was colder and rainier than the control week.
Biodigester efficiency Gregory 9 Figure 6: The average flame heights of the control and experimental treatments. These are statistically distinct results (p=.0037), showing that the experimental treatmen t produced a larger average flame.
Biodigester efficiency Gregory 10 Figure 7 : The daily burn times of the control and experimental treatments. Though the averages are different, there is no statistical difference between these results (p=.16). The first BOD 5 test did not work, as some samples had higher dissolved oxygen after the 5 day wait period. The seal most likely broke, so this test was discarded. The results of the second BOD 5 test were mixed. The three influent samples displayed positive BOD 5 values. The three effluent samples displayed negative BOD 5 values ( Fig. 8 ).
Biodigester efficiency Gregory 11 Figure 8: The average BOD in the effluent and influent of the biodigester. They are significantly different (p=.0185). The average pH of the the effluent was less than the pH of the influent, in both the control and experimental treatment. However, the experimental treatment was more basic, as both the influent and the effluent had higher pH levels ( Table 2 ). The differences in pH between the control and experimental treatments were s tatistically significant (p < .05). Table 2 : Average pH over 3 replicates of influent and effluent Treatment pH Influent (average) pH Effluent (average) Control 8.22 7.98 Experimental 9.28 8.21 Discussion Lab experiment: Animal manures, digested alone or with other animal manures, produced the most gas. Though some animal manure treatments produced more gas than others, there was no significant difference between the different animal manure treatments in my experiment. This suggests that co digesting pig manure with goat or chicken can create significant methane for fuel, but that it does not outperform pig manure alone in total gas production. The water to waste ratio did not significantly affect total gas production in the lab setting ei ther, as the differences between the control and Treatment P were not statistically significant. This result was refined in the field experiment, as reducing water content affected flame height rather than burn time.
Biodigester efficiency Gregory 12 Many chemical properties influence the gas production of a given treatment, including the carbon to nitrogen ratio. The C/N ratio of each treatment was calculated based on existing research (Carlini et al., 2015). The ideal C/N ratio for anaerobic digestion is estimated to be around 20 30 (Kav uma, 2013). No treatment reached this ideal C/N ratio, but treatments with higher ratios yielded higher gas production ( Fig. 5 ). Co digestion with whey, conversely, produced significantly less gas than the control. Although other studies have found that whey produces gas, it did not show great methanogenic potential in this experiment (Gelegenis et al., 2007). One potential explanation for this finding was the low C/N ratio of whey, which could be lowering the gas production (Carlini et al., 2015). Whey t ypically takes longer to produce gas than manure, but the experiment ran for 11 days without any significant gas production from whey containing mixtures. These lab results can be adapted to other farms considering implementing or refining a biodigester. Field experiment: The flame height significantly increased from the control to the experimental treatment. The experimental treatment lowered the water to waste ratio, which could have increased the concentration of methane in the biogas produced , leading to a stronger flame. This stronger flame could reduce the time needed for cooking and heating. The flame burn time did increase with the implementation of a goat and pig co digestion treatment, but this change was not statistically significant. T his could indicate the relative similarity of pig manure and co digestion in terms of gas production. There are, however, other confounding variables, including storage size and temperature. Life Monteverde has limited storage for the biogas produced, and it is possible that no adjustment could have significantly increased the biogas use, as they already go over capacity. The average temperatures were decreased the second week, which could have decreased the gas production. From my daily temperature calcul ations, the average temperature the first week was 24 Â¡C, while the second week was 20.8Â¡C. It is hard to correlate these values directly with gas production, as burn time did increase in the experimental week ( Fig. 7 ). However, lower temperatures have been shown to decrease biodigester gas production, so it is possible that burn time could have increased more under controlled temperature conditions (Bogich, 2014, Sheffield, 2010, Hojnacki et al., 2011). Regardless, these results indicate that co digestion with goat manure does produce sufficient gas, and easily solves the water to waste ratio without decreasing the water input. However, it also introduces a substantial amount of solid waste to the system, which could present other operation and maintenance problems. Life Monteverde is considering reducing their pig stock. My results show that, while co digestion can produce adequate gas, it does not significantly increase gas production. Pig manure, with its comparatively high C/N ratio, was essenti al for each mixture's gas production. Thus, Life will likely get the best results by focusing on reducing the water content. This can be achieved by adding goat manure (as I did), or by reducing water input each day. If the stock of pigs decreases, both ad ding manure and reducing water may be needed to achieve the desired ratio. The results of the BOD 5 tests are difficult to interpret. The control BOD test was discarded, as the values were inconsistent, and many had higher DO values after 5 days.
Biodigester efficiency Gregory 13 The seco nd BOD test, which was sealed twice, was more mixed. The influent had positive BOD values, indicating that microorganisms were actively consuming oxygen. However, the average BOD (less than 1) indicates a much cleaner water source than was present (Hocking , 2005). The effluent, meanwhile, had negative BOD values. This result implies an issue with the testing. It is possible that algae grew in the bottles and produced oxygen. More algae, in this case, grew in the effluent tube. The pH of the experimental t reatment was higher in both the influent and effluent. As stated in the introduction, the ideal pH range for anaerobic digestion is 6.8 to 7.2 (Cioabla et al., 2012). The control conditions were closer to this ideal ( Table 2 ). This could be a reason that t he gas production did not significantly increase. However, the relatively high gas production demonstrates that this pH did not significantly limit gas production. This experiment did not address a number of important issues of the Life biodigest er. This biodigester is already 7 years old, installed in 2012. Polyethylene biodigesters are only built to last 8 or 9 years. This age can lead to sedimentation, reducing the space for gas storage and slowing the efficiency of the biodigester. Another iss ue is storage space, as the biodigester lacks a large enough container to store the gas produced daily. The third is insulation, as the biodigester efficiency depends partly on temperature. On days when it is colder and rainier, the biodigester produces le ss gas. In some places, especially with greater temperature variation, biodigesters are insulated using wood chips or other methods. Future studies could explore the precise effects of temperature variation on biodigester efficiency in Monteverde. If Life wants to replace and update its biodigester, this study provides information on proper inputs and maintenance to maximize their gas production, thus maximizing its usefulness. Acknowledgements To Guillermo Vargas and Jerson SantamarÂ’a of Life Monteverde: thank you for your trust, early morning company, invaluable knowledge, and free flowing fresh coffee. To Justin Welch: thank you for your enthusiasm, endless understanding of waste water systems, quick humor, and suga ry treats. To SofÂ’a Arce Flores: thank you for inspiring me to do an agroecological project, and for your determined advising in the face of a very important due date. To Alex Reep: thank you for your unwavering support, sharing your tortillas, and keeping me company each day on the farm. To Katie Sanko: thank you for your careful reading, scientific prowess, and constant kindness. To my homestay family: Gracias por su amabilidad y por toda la comida rica. To everyone at Life: thank you for your kindness an d your patience with my slow Spanish. To Frank Joyce, Federico Chinchilla, Emilia Triana, and FÂŽlix Salazar: thank you for making me laugh, grow, and see the natural world differently. Finally, to my classmates at UCEAP: you are an inspiring, brilliant bun ch. Thanks for making Hotel Cacts, 2 person tents, and the E staciÂ—n feel like home. Literature Cited Achinas, Spyridon; Achinas, Vasileios; and Euverink, Gerrit Jan Willem. "A Technological Overview of Biogas Production from Biowaste." Engineering , vol. 3 , no. 3, 2017, pp. 299 Ã 307., doi:10.1016/j.eng.2017.03.002.
Biodigester efficiency Gregory 14 Ano, A, and Ubochi, C. "Neutralization of soil acidity by animal manures: mechanism of reaction." African Journal of Biotechnology Vol. 6 (4). January, 2007. http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.866.5858&rep=rep1& type=pdf Bogich, John. "Comparisons of biodigestion of pig manure versus coffee skins and pulp: methanogenic poten tial and effluent content." UCEAP . Fall 2014. Carlini M., Castellucci S., Moneti M. "Biogas production from poultry manure and cheese whey wastewater under mesophilic conditions in batch reactor." Elsevier. 2015 doi: 10.1016/j.egypro.2015.11.817. Cioa bla, Adrian; Ioana Ionel, Gabriela Alina Dumitrel, and Francisc Popescu. "Comparative study on factors affecting anaerobic digestion of agricultural vegetal residue" Biotechnol Biofuels. 2012. doi: 10.1186/1754 6834 5 39 Gelegenis, John; Georgakakis, Dimitris; Angelidaki, Irini; and Mavris, Vassilis. "Optimization of biogas production by co digesting whey with diluted poultry manure." Renewable Energy , vol. 32, issue 13. 2007. Hocking, Martin. "Biochemical Oxygen Demand." Handbook of Chemical Technology and Pollution Control . Third Edition. 2005. https://www.sciencedirect.com/topics/earth and planetary sciences/biochemical oxygen demand Hojnacki, Angela; Li, Luyao; Kim, Nancy; Markgraf, Claire; and Pierson, Drew. "Biodigester Global Case Studies.pdf." MIT , 2011. https://www.build a biogas plant.com/PDF/D_Lab_Waste_Biodigester_Case_Studies_Report.pdf . Kafle, Gopi Krishna, and Lide Chen. "Comparison on Batch Anaerobic Digestion of Five Different Livestock Manures and Prediction of Biochemical Methane Potential (BM P) Using Different Statistical Models." Waste Management , vol. 48, 2016, pp. 492 Ã 502., doi:10.1016/j.wasman.2015.10.021. Kavuma, Chrish. "Variation of Methane and Carbon dioxide Yield in a biogas plant" Department of Energy Technology, Royal Institute of Technology, Sweden. 2013. http://kth.diva portal.org/smash/get/diva2:604559/FULLTEXT02.pdf Lv, Zongyan; Feng, Lei; Shao, Lijie; Kou, Wei; Liu, Peihan; Gao, Peng; Dong, Xia oying; Yu, Meiling; Wang, Jiuzhang; and Zhang, Dalei. "The Effect of Digested Manure on Biogas Productivity and Microstructure Evolution of Corn Stalks in Anaerobic Cofermentation." BioMed Research International , vol. 2018, 2018, pp. 1 Ã 10., doi:10.1155/201 8/5214369. Sheffield, Jordan. "Influents and Effluents: An analysis of Biodigesters in San Rafael." UCEAP . Fall 2010.
Biodigester efficiency Gregory 15 "Soil Acidity." Crop Nutrition. https://www.cropnutrition.com/efu soil ph 2019.