Determination of the Effect of Manganese(II) on the Removal of Arsenic in Contaminated Water (Spiked Solution) using Electrochemical Arsenic Remediation (ECAR)

Journal of Human Ecology and Sustainability Citation Rubin, A. K. B., Austero, S. B., Orozco, C. R., & Resurrecion, A. C. (2024). Determination of the Effect of Manganese(II) on the Removal of Arsenic in Contaminated Water (Spiked Solution) using Electrochemical Arsenic Remediation (ECAR). Journal of Human Ecology and Sustainability, 2(3), 4. doi: 10.56237/jhes24ichspd06 Corresponding Author Austin Kinn B. Rubin Email [email protected] Academic Editor Casper B. Agaton Received: 31 July 2024 Revised: 19 November 2024 Accepted: 24 November 2024 Published: 28 November 2024 Funding Information This study is under the Philippine Electrochemical Arsenic Remediation (PHIL-ECAR-I) Project funded by CHED-PCARI and implemented by the University of the Philippines Diliman. ©The Author(s) 2024. This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution(CC BY) license (https://creativecommons.org/ licenses/by-nc-nd/4.0/). Original Research Article Determination of the Effect of Manganese(II) on the Removal of Arsenic in Contaminated Water (Spiked Solution) using Electrochemical Arsenic Remediation (ECAR) Austin Kinn B. Rubin 1, Sheila B. Austero 2, Christian R. Orozco 1, and Augustus C. Resurreccion 1 1Institute of Civil Engineering, University of the Philippines Diliman, Quezon City, 1101, Philippines 2Department of Community and Environmental Resource Planning, College of Human Ecology, University of the Philippines Los Baños, College 4031, Laguna, Philippines Abstract Sustainable Development Goals (SDGs) 3, 6, and 11 emphasize the need for safe and clean water to create livable and sustainable human settlements. Arsenic contamination poses a significant challenge to these goals, threatening water safety and community health. Electrochemical Arsenic Remediation (ECAR) is an effective method for arsenic removal in water but is dependent on factors such as pH, dissolved oxygen, and ions such as calcium and magnesium. Electrocoagulation experiments were performed to determine the effect of manganese onthearsenicremovalofECAR at a lowchargeloadingrate(5C/L)usingarsenicandarsenic-manganesesolutions. Results show that the percent arsenic removal for the control setup ranges from 61-96% and approximately 100% for the arsenic-manganese setup. The presence of 500 ppb manganese in the solution reduced arsenic concentration below the 10 ppb maximum limit set by WHO and DOH for drinking water for all arsenic-manganese setup which indicates that manganese has the potential to aid arsenic removal in water even at a low charge loading rate. Complete removal of manganese from the solution after ECAR was also observed. Statistical analysis shows that manganese and initial arsenic concentration significantly affect arsenic removal with an average estimated marginal effect of 0.4078 and 1.778, respectively. Results consequently illustrate that ECAR is a viable and effective technology in helping achieve SDGs 3, 6, and 11. Keywords— arsenic, electrochemical, manganese, groundwater, remediation 1 1 Introduction Accesstosafeandcleanwaterisafundamentalhumanrightandapillarofsustainabledevelopment. The United Nations has emphasized this through Sustainable Development Goal (SDG) 6, which aims to “ensure availability and sustainable management of water and sanitation for all” and SDG 3 Target 3.9, which aims to “substantially reduce the number of deaths and illness from hazardous chemicals and air, water and soil pollution and contamination” by 2030 [1]. This goal is directly linked to SDG 11, which seeks to “make cities and human settlements inclusive, safe, resilient and sustainable” [1]. The link between these goals highlights the critical role of access to safe and clean water in creating livable and sustainable human settlements. Without reliable access to safe and clean water, communities and settlements face significant health risks, economic challenges, and barriers to social development [2]. As global urbanization increases rapidly, the pressure on water resources and infrastructure in human settlements continues to rise [3]. To adapt to the growing demand for water resources, groundwater is increasingly preferred over surface water as the source of drinking water due to its availability, practicality, and affordability [4]. Natural groundwater contains cations and anions. According to the study conducted by Husana & Yamamuro [5], the primary ions found in Philippine groundwater are calcium(II), magnesium(II), sodium, potassium, ammonium ion, chloride, nitrite and nitrate, bromide, and sulfate. However, groundwater contamination of heavy metals such as arsenic presents a significant obstacle in achieving SDG 3, SDG 6 and SDG 11 since it threatens the safety of groundwater supply and, thus, the overall health of human settlements and communities. In 2020, Solis et al. [6] found that the groundwater samples in the municipality of Guagua in the province of Pampanga were found to contain arsenic(As) concentrations as high as 95 parts per billion (ppb), which is above the 10 ppb limit set by the World Health Organization (WHO) [7] and by the Department of Health (DOH) [8] under the Philippine National Standards for Drinking Water of 2017. Similarly, Apostol et al. [9] also reported groundwater contamination of arsenic in the province of Batangas. They found out that 12 out of 14 (85.71%) wells used as a source of drinking water found in the municipalities surrounding the Taal Volcano in 2020 and 2021 have arsenic concentrations between 10 ppb to 39 ppb and arsenic concentrations between 10 ppb and 46 ppb, respectively. When non-drinking water sources are considered, they found that 20 out of 26 wells (76.92%) they studied had persistent elevated arsenic concentrations in 2021 based on their statistical analyses. They have recommended further continuous monitoring of water quality and arsenic mapping to be conducted in the province of Batangas to determine the extent of arsenic contamination of the groundwater. They have also recommended to provide alternative water supplyfortheaffectedcommunities. Inadditiontothese, 9outof17(52.94%)watertreatmentplant outlets and water sources were found to have arsenic concentrations ranging from 20 ppb to 70 ppb in the municipalities of Calauan, Bay, and Los Baños in the province of Laguna last September 2020, according to the 2020 Annual Audit Report on the Laguna Water District by the Commission on Audit (COA) [10]. Alarmingly, arsenic concentrations in the province of Laguna, especially in the municipalities of Bay and Los Baños, still have arsenic concentrations above the 10 ppb limit after two years, according to the 2022 Annual Audit Report on the Laguna Water District by COA [11]. On the same annual audit report, a 30-day trial was conducted to reduce the arsenic concentrations at Lopez Heights Pumping Station below the 10 ppb limit last October 10, 2022; however, the trial failed, and all groundwater sources contaminated with high amounts of arsenic is being considered to be shut down since the proposed solution of using reverse osmosis to remove arsenic incurs high capital expenditure. Within the three provinces, arsenic contamination of groundwater hinders the availability of cheap and affordable water supply. It threatens human health, which impedes the achievement of SDG 3, SDG 6, and SDG 11 for these communities. Arsenic is a naturally occurring element found in highconcentrationsincertainwatersourcesandisclassifiedasaGroup1humancarcinogenbyWHO Rubin et al. (2024) | Journal of Human Ecology and Sustainability 2 [12]. In addition, arsenic is also considered by the US Environmental Protection Agency (USEPA) and the Council of European Communities (CEC) as a prime pollutant [13]. Arsenic in drinking water can lead to arsenic poisoning or arsenicosis [12,14]. Generally, arsenic contamination in surface water is less than in groundwater, and surface water is contaminated mainly due to anthropogenic activities, while groundwater contamination is primarily by natural sources [12]. ElectrochemicalArsenicRemediation (ECAR) is onemethodthatremovesarsenic in water. ECAR produces coagulants in situ through electrocoagulation to remove arsenic in groundwater [12, 14,15]. Electrocoagulation is becoming more prominent due to its rapid and effective removal of arsenic from water [13]. It is also deemed environmentally friendly compared with chemical coagulation and produces lower quantity, stable sludge [12,13,14,16]. Unlike conventional methods, electrocoagulation is low-cost and is easy to implement, operate, and maintain with locally available materials. Electrocoagulation also does not introduce undesirable anions like chemical coagulation [17]. Thearsenicremovalefficiencyofelectrocoagulationishigherthanthatofconventionalchemical coagulation. Iron electrocoagulation can remove greater than 95%-99% of arsenic from the water [12]. This range of arsenic removal efficiencies is proven by a study done by the University of California Berkeley (UC Berkeley) in Bangladesh, where they showed that iron electrocoagulation technology was able to reduce the arsenic content of both synthetic and natural groundwater from 3000 ppb to 10 ppb, which is within the WHO acceptable level [14]. The arsenic removal efficiency of ECAR depends on a lot of factors. These factors include pH [18], dissolved oxygen [16,19], and ions such as phosphates [12,13,20], calcium, and magnesium [15,17]. These factors should be considered for the efficient removal of arsenic in water. According to Amrose et al. [21], Dutta & Gupta [22], and Kobya et al. [23], charge loading is a crucial factor in ECAR since it controls the production rate of coagulants which affects arsenic removalrateandincreasingchargeloadinggenerallyresultsinhigherarsenicremovalrate. Itshould be noted, however, that excessive charge loading can hinder floc flotation and separation due to high charge density and poor affinity between hydroxide flocs and gas bubbles [23]. Additionally, excessive charge loading results in higher energy costs and, with prolonged use and high charge loading cycles, results in the formation of a passivation layer on the electrodes which can increase power requirements over time, impacting the long-term operational costs of ECAR [23,24]. Van Genuchten et al. [15] found that bivalent cations (e.g., calcium(II) and magnesium(II)) improve arsenic removal by (1) enhancing oxyanion adsorption, (2) promoting crystallite aggregation and (3) altering precipitate mineral phase and primary crystallite size. This contrasts oxyanions that reduce precipitate crystallite size, resulting in stable colloidal suspensions. Their study also suggests calcium(II) has stronger interactions with oxyanions than magnesium(II). Oxyanion interaction of both calcium(II) and magnesium(II) are arranged in the following from strongest to weakest: phosphates> arsenic(V) > silicates. Van Genuchten et al. [25] also reported that calcium(II) specifically acts as a catalyst for interparticle bridging of iron-arsenic complexes and iron-phosphate complexes. Catrouillet et al. [26] investigated the mechanism of arsenic removal in the presence of manganese(Mn) in water of different pH. At pH 4.5, they found out that arsenic(III) was rapidly oxidized by hydroxyl radicals(OH•), and manganese-enhanced the aggregation of arsenic(V)-iron(III) polymers. On the other hand, they found that manganese(II) and arsenic(III) compete for iron(IV) at pH 8.5, which results in arsenic(III) remaining in the solution. Arsenic(V) that formed at pH 8.5 is incorporated into the structure of arsenic(V)-iron(III) polymers and ferrihydrite-like phases that contained 8% manganese(III) and adsorbed some arsenic(III). Catrouillet et al. [26] also noted that arsenic(III) and manganese(II) compete for the oxidants at intermediate pH values ( ∼ pH 6.5); however, manganese(III) behaved as a reactive intermediate that reacted with iron(II) or arsenic(III) which explains the presence of arsenic(V) in the precipitates. Rubin et al. (2024) | Journal of Human Ecology and Sustainability 3 Understanding the factors affecting arsenic removal in electrocoagulation, especially those with limitedstudiessuch as cations, isessential as this helpspredict the performanceof ECAR andallows better optimization of parameters for operational implementation. Both studies conducted by Catrouillet et al. [26] and Van Genuchten et al. [15] are among the few studies investigating cations’ effects on arsenic removal during electrocoagulation. Both studies examined the mechanisms and reactions involved in arsenic removal in the presence of calcium or manganese. However, neither of these studies incorporated rigorous statistical analyses that could quantitatively demonstrate whether the observed effects of cations in arsenic removal in their experiments were statistically significant. Additionally, low charge loading has been used by different studies such as those conducted by Kobya et al. [23] where they investigated the effects of 1-130 C/L charge loading on 75-500 ppb arsenic, by Dutta et al. [22] where they find the optimal charge loading from 5-30 C/L with 60-110 ppb arsenic, and by Dutta and Gupta [27] where they find the optimal charge loading from 5-120 C/L with 200 ppb arsenic and activated alumina after ECAR. However, these studies use a direct power source with no reported reference electrode to ensure constant charge loading and mostly used high charge loading on high arsenic concentrations. In this study, the effect of manganese(II) on arsenic removal in electrocoagulation at a low charge loading rate (5 C/L) and low arsenic concentrations mimicking real-world concentrations was investigated and quantified using regression analysis. This research aims to support SDGs 3, 6, and 11 by promoting technologies that could help in producing safe and clean water to protect the health and well-being of everyone residing in cities and communities with the following objectives: 1. determine the removal of arsenic using Electrochemical Arsenic Remediation (ECAR) in contaminated water (spiked solution) with and without manganese(II); 2. investigate the effect of manganese(II) on the removal of arsenic in contaminated water (spiked solution) using ECAR; and 3. verifythesignificantfactorsandestimatemarginaleffectsinECARusing theGeneralizedAdditive Model (GAM). 2 Methodology 2.1 Preparation of Materials Thepreparationof materialsinthisstudy wasbasedonvan Genuchtenetal. [25]. Allglasswareused in the experiment was washed using a 1% nitric acid ( H N O3 ) bath and then washed with ultrapure 18.2 M Ω deionized water before air drying. Iron electrodes were soaked in 10% hydrochloric acid (HCl) solution for 20 minutes. Iron electrodes were then washed with copious running water. Rust and other particles were removed using steel wool and then with fine sandpaper. Iron electrodes were then stored in a desiccator with silica gel. The appearance of iron electrodes after cleaning is shown in Figure 1. Analytical grade sodium arsenite ( N aAsO2 ), manganese(II) sulfate ( M nSO4 ), and ultrapure 18.2 Ω deionized water were used to prepare 250 mL of 20 parts per million (ppm) arsenic(III) and 250 mL of 500 ppm manganese(II) stock solution. Two liters of 20 ppb, 40 ppb, and 100 ppb arsenic(III)solutionwerepreparedforthecontrolsetup. Twolitersofsolutionwith20ppbarsenic(III) + 500 ppb manganese(II) and 120 ppb arsenic(III) + 500 ppb manganese(II) were prepared for the arsenic-manganese combination setup. 2.2 ECAR Experiment The electrocoagulation setup of the experiment was also based on van Genuchten et al. [25]. Figure 2shows the configuration of the electrocoagulation setup used in this experiment. Before each run, Fe electrodes were washed with ultrapure 18.2 Ω deionized water and were only used once. Silver/silver chloride (Ag/AgCl) reference electrode was used and washed with ultrapure 18.2 Ω Rubin et al. (2024) | Journal of Human Ecology and Sustainability 4 Figure 1. Fe electrodes after electrode cleaning deionized water before each run. An Ag/AgCl electrode was used as a reference electrode to ensure a fixed charge loading of 5 C/L was applied to the solutions. A reference electrode provides a well-defined and stable reference potential. Electrode spacing was ∼ 5 cm. The anode and cathode were arranged to have approximately the same submerged area (∼13-14 cm2). Each setup has 500 mL of the solution and was run with a charge loading of 5 C/L for 30 minutes (0.17 C/L-min dosage rate) using potentiostat. Each experiment setup is run twice and sampled thrice for measurement. Samples were filtered using 0.22 µ m filter paper and analyzed using inductively coupled plasma atomic emission spectroscopy (ICP-AES) from Teledyne. The pH of each setup was also measured before and after electrocoagulation using pH paper. 2.3 Statistical Analysis The statistical significance of the initial arsenic and manganese concentrations on the final arsenic concentration was determined using R and R Studio for regression analysis. Correlation analysis was conducted using the “ggpairs” function from the “GGally" package developed by Schloerke et al. [28]. Generalized additive models (GAM) were used by using the “gam” function from the “mgcv” package developed by Wood [29] to model the relationships of the variables in the data. In many studies, GAMs have already been used to investigate relationships between variables. For instance, Liu et al. [30] used GAMs to analyze the interactive effects of different water quality Rubin et al. (2024) | Journal of Human Ecology and Sustainability 5 Figure 2. ECAR setup. The anode is the red terminal, the cathode is the black terminal, and the Ag/AgC l electrode is the blue terminal. parameters on the water purification of a deep oxidation pond with horizontal subsurface flow constructedwetland. Ontheotherhand,Liuetal. [31], ontheotherhand, examinedtheinteractions between different water quality parameters and the removal efficiency of sulfamethoxazole in an electrolysis-integrated tidal flow-constructed wetland system. Further, Amin et al. [32] used GAMs to determine the impact of various factors on carbon dioxide emissions in Bangladesh, Ye et al. [33] applied GAMs to assess the impact of multiple factors on heavy metal leaching from nickel tailings, and Marques et al. [34] assessed potential biomarkers of contaminants in estuaries using biochemical and physiological parameters in croakers (Micropogonias furnieri) with GAMs. The following is the formulation of the GAM used in this study: arsenic_final_concentration ~ s(manganese_initial_concentration, k = 3) + s(arsenic_initial_concentration, k = 3) + ti(manganese_initial\_concentration, arsenic_initial_concentration, k = 3) The “s()” wrapped around the variables means that the variables are modeled using smooth terms and the “ti()” wrapped around the two variables means that the variables are modeled as tensor product interactions. The basis dimensions (k) for all terms are set to equal three to avoid overfitting due to the limited number of measurement data used in the GAM (n=27). Basis functions are additively combined to create smooth terms, as discussed by Wood [35]. Setting the basis dimension to three means that the basis functions can only have degree two (k-1) complexity, i.e., the basis functions are only limited to quadratic functions at most. The average estimated marginal effects of initial arsenic concentration and initial manganese concentration to the final arsenic concentration were also calculated using the “slopes()” function of the “marginal effects” package developed by Arel-Bundock et al. [36]. Marginal effects were estimated at 0 ppb, 400 ppb, 450 ppb, 500 ppb, and 550 ppb of initial manganese concentrations while marginal effects of initial arsenic concentration were estimated at 10 ppb, 20 ppb, 30 ppb, 40 ppb, 80 ppb, 100 ppb, and 120 ppb. Intermediate concentration values were also included in the estimation of marginal effects to capture the inherent variability between raw concentration values Rubin et al. (2024) | Journal of Human Ecology and Sustainability 6 of arsenic and manganese and to help quantify how the effects of the initial arsenic and manganese concentration on the final arsenic concentration changes as their concentration increases. 3 Results and Discussion 3.1 ECAR Experiments Arsenic is removed during the electrocoagulation via the following mechanisms. Electric current is passed through iron electrodes and iron(II) is then dissolved, which in the presence of dissolved oxygen, gets oxidized into iron(III) [19,20]. Iron(III) ions polymerize to produce reactive iron(III) (oxyhydr)oxides that bind to arsenic(V) [19]. If arsenic(V) is present, iron(III)-arsenic precipitates and complexes are formed; thus, arsenic gets removed by the separation of the precipitates and complexes from the solution through gravitational settling and/or filtration [19,20]. During oxidation by dissolved oxygen, iron(II) can also form a highly reactive iron(IV) oxidant, which can oxidize arsenic(III) into arsenic(IV) [20]. If enough iron(IV) is present, iron(IV) may further oxidize arsenic(IV) to arsenic(V), which then adsorbs iron(III) and precipitates from the solution [16, 20]. Hydrogen gas also forms in the cathode during electrocoagulation [12,13,19]. Hydrogen gas may remove flocs that never settle through gravitation by taking the flocs at the top of the solution (also called electro-floatation), forming a foam-like phase at the top of the solution, which can be removed by skimming [12,13]. The flocs formed are usually negatively charged [12]. Figure 3. pH of control setup before and after ECAR The pH value of both the control and arsenic-manganese setup increased after ECAR from acidic to basic, as shown in Figure 3. This increase in pH is in agreement with Nidheesh and Singh [12] and Song et al. [13], as hydroxyl ions accumulate during electrolysis. All ECAR setups have also been observed to have yellowish to orangish color during and after electrocoagulation, as shown in Figure 4. The colors indicate the presence of lepidocrocite ( γ -FeOOH) and/or goethite ( α -FeOOH) as a predominant species of coagulant in electrocoagulation, as explained by Dubrawski et al. [37], Song et al. [13], and Wan et al. [18]. Rubin et al. (2024) | Journal of Human Ecology and Sustainability 7 Figure 4. Color of contaminated water (spiked solution) after ECAR The mean arsenic percent removal of the control setup across concentration levels is about 76.78% while the mean arsenic percent removal of the arsenic-manganese setup is 97.59% across the concentration levels as shown in Figure 5. The raincloud plot – a combination of point plot, box plot, and violin plot – also illustrates that the measurements for arsenic percent removal greatly vary with a median of about 73.59%, while the arsenic percent removal of the arsenic-manganese setup aggregate around a single value (100%) which causes the box plot to be not visible and only to look a single line. Based on the aggregation of values in the arsenic-manganese setup, it could be seen that one measurement is an outlier which can be attributed to imprecise measurement as seen in Table 1where the standard errors are large. Table 1. Arsenic Removal of the Control Setup and Arsenic-Manganese Setup. Each trial consists of three measurements Arsenicremoval in the control setupis between61%-96% and large variabilityof arsenic percent removal for each trial, especially for the second trial, is also observed as seen in Figure 6. This large variability in the measurement of arsenic percent removal can be attributed to the low concentrations used in the experiments. Despite the variability in measurements, arsenic percent removal generally decreases as the initial concentration of arsenic increases. This trend is expected since the charge loading of the setup was held constant at 5 C/L. Thus, the generation rate of coagulants Rubin et al. (2024) | Journal of Human Ecology and Sustainability 8 Figure 5. Distribution of Arsenic Percent Removal of Control and As-Mn. The green diamond point indicates the mean is constant. This means that the constant amount of coagulants is not enough to precipitate arsenic at higher concentrations. Overall, ECAR at a charge loading of 5 C/L of the control setup is only compliant with the maximum 10 ppb arsenic concentration set by WHO and DOH when the initial arsenic concentration is 20 ppb. For the arsenic-manganese setup, arsenic removal is mostly 100% with no variability aside from a single outlier as shown in Figure 7. As discussed previously, the outlier can be attributed to imprecise measurement. Despite the outlier with a final arsenic concentration of 6.66 ppb, final arsenic concentrations in all setups of arsenic-manganese solutions are below the maximum 10 ppb arsenic concentration set by WHO and DOH. Results of these experiments suggest that manganese(II) helps remove arsenic in water at pH 5 to pH 6, which agrees with the results of Catrouillet et al. [26] at pH 4.5. The charge loading rate is given by the following equation from Amrose et al. [22]: q=I∗te V(1) Rubin et al. (2024) | Journal of Human Ecology and Sustainability 9 Figure 11. Partial Effects of Initial Arsenic Concentration on Final Arsenic Concentration partial dependence plot illustrated in Figure 10. By taking the average of the significant estimated marginal effect, the average estimated marginal effect of initial manganese concentrations on the final arsenic concentration is -0.4078, which means that we expect a decrease in the final arsenic concentration by 0.4078 ppb for each unit increase of initial manganese concentration as long as the unit increase is small. Theestimated marginaleffectof initial arsenic concentrationsonthefinalarsenicconcentration is generally positive except at 100 ppb and 120 ppb arsenic, in which the estimated marginal effects are negative and insignificant. These exceptions can be explained by the lack of data for the GAM, as shown in the wide confidence band in the 100 ppb and 120 ppb range of the partial dependence plot shown in Figure 11. Nevertheless, by only considering the significant estimated marginal effects, we can see that the estimated marginal effect of initial arsenic concentrations on the final arsenic concentration is in accordance with the partial dependence plot of initial arsenic concentration on the final arsenic concentration. By taking the average of the significant estimated marginal effect, the average estimated marginal effect of initial arsenic concentration on the final arsenic concentration is 1.778, which means that we expect an increase in the final arsenic concentration by 1.778 ppb for each unit increase of the initial arsenic concentration as long as the unit increase is small. Based on the average estimated marginal effects of initial manganese concentrations and initial arsenic concentrations on the final arsenic concentration, we can see that the presence of Rubin et al. (2024) | Journal of Human Ecology and Sustainability 16 Table 3. Estimated Marginal Effect of Initial Manganese Concentration on Final Arsenic Concentration Table 4. Estimated Marginal Effect of Initial Arsenic Concentration on Final Arsenic Concentration Rubin et al. (2024) | Journal of Human Ecology and Sustainability 17 Figure 12. Partial Effects of Initial Arsenic and Manganese Concentrations on the Final Arsenic Concentration. Red and green dashed lines indicate standard errors while the solid black line indicates the interaction effects estimated by the GAM. manganese helps remove arsenic in ECAR. However, the effect of manganese is still small (about 4 times less) compared to the effect of initial arsenic concentration in arsenic removal. 4 Conclusion and Recommendations Manganese was found to aid significantly arsenic removal mostly by 100% during ECAR with approximately 2:145:50 and 12:145:50 arsenic(III):iron(II):manganese(II) ratio at pH 5 to pH 6 and at a low charge loading rate of 5 C/L. Regardless of the arsenic-manganese solution setup, 500 ppb manganese reduced the arsenic concentration below the 10 ppb maximum arsenic concentration limit set by WHO and DOH for drinking water. Total removal of manganese after ECAR was also observed on all arsenic-manganese solutions, which may be due to the electrosynthesis of excess iron(II) and manganese(II) into manganese ferrite (M nF e2O4) nanoparticles. Statistical analysis shows that the effects of initial arsenic and manganese concentrations and their interaction on the final arsenic concentration are significant. However, it should be noted that the significance of the interaction effects may be unreliable due to the concurvity between the interaction effects and the main effects. The average estimated marginal effect of initialarsenic and manganese concentrations on the final arsenic concentration is -0.40778 and 1.778 respectively. The study shows that ECAR effectively removes arsenic from groundwater, particularly in the Rubin et al. (2024) | Journal of Human Ecology and Sustainability 18 Figure 13. Partial Effects Heatmap of Initial Arsenic and Manganese Concentrations on the Final Arsenic Concentration. The color ranges from red to yellow, corresponding to negative effect (red) to positive effect (yellow). presence of manganese. This technology supports clean water access and aligns with three UN Sustainable Development Goals: health (SDG 3), clean water (SDG 6), and sustainable communities (SDG 11). However, it should be noted that the study is limited by the following: 1. the study used contaminated water (spiked solution) to determine the effect of manganese(II). 2. the study used only a fixed 5 C/L charge loading. 3. the study was conducted during the height of the pandemic which resulted in lockdowns; thus, not all planned trials were conducted. Toaddresstheselimitations,furtherinvestigationofthemechanismoftotalmanganeseremoval after ECAR and further experiments with different arsenic(III):iron(II):manganese(II) ratios at various pH levels are recommended to elucidate further the interaction of arsenic, iron, and manganese in ECAR. For future research, the impact of additional factors, such as different charging times and charge loadings, diverse chemical ratios, and various metal impurities can be explored to simulate actual groundwater conditions. Finally, given the marginal effects observed in this study, future experimentsshouldbedesignedandconductedwithsufficientexperimenttrialsandmeasurements to achieve appropriate statistical power. Rubin et al. (2024) | Journal of Human Ecology and Sustainability 19 Figure 14. Effects of Different Initial Concentrations of Arsenic and Manganese on Final Arsenic Concentration Statements and Declarations Funding Information This study is under the Philippine Electrochemical Arsenic Remediation (PHIL-ECAR-I) Project funded by CHED-PCARI and implemented by the University of the Philippines Diliman. Conflicts of Interest The authors declare no conflict of interest. Ethical Considerations Not Applicable Data Availability The data in this study are available upon request from the authors. Authors Contribution A.K.B.R.: conceptualization, methodology, experimentation, statistical analysis, writing-major contribution and original draft preparation, S.B.A.: conceptualization, supervision, mentoring, writing-minor contribution, reviewing and editing of manuscript; C.R.O. and A.C.R.: conceptualization, funding, academic advising. All authors have read and agreed to the published version of the manuscript. Rubin et al. (2024) | Journal of Human Ecology and Sustainability 20 References [1] United Nations. (2015). Transforming our world: The 2030 agenda for sustainable development. https://sdgs.un.org/sites/default/files/publications/ 21252030 % 20 Agenda% 20 for% 20Sustainable%20Development%20web.pdf [2] Tortajada, C., & Biswas, A. K. (2018). Achieving universal access to clean water and sanitation in an era of water scarcity: Strengthening contributions from academia. Current opinion in environmental sustainability,34, 21–25. https://doi.org/10.1016/j.cosust.2018.08.001 [3] McDonald, R. I., Weber, K., Padowski, J., Flörke, M., Schneider, C., Green, P. A., Gleeson, T., Eckman, S., Lehner, B., Balk, D., et al. (2014). Water on an urban planet: Urbanization and the reach of urban water infrastructure. Global environmental change,27, 96–105. https: //doi.org/10.1016/j.gloenvcha.2014.04.022 [4] Margat, J., & Van der Gun, J. (2013). Groundwater around the world: A geographic synopsis. Crc Press. https://doi.org/10.1201/b13977 [5] Husana, D. E., & Yamamuro, M. (2013). Groundwater quality in karst regions in the Philippines. Limnology,14, 293–299. https://doi.org/10.1007/s10201-013-0398-8 [6] Solis, K. L. B., Macasieb, R. Q., Parangat Jr, R. C., Resurreccion, A. C., & Ocon, J. D. (2020). Spatiotemporal variation of groundwater arsenic in Pampanga, Philippines. Water,12(9), 2366. https://doi.org/10.3390/w12092366 [7] World Health Organization. (2022). Guidelines for drinking-water quality: Incorporating the first and second addenda.https://www.who.int/publications/i/item/9789240045064 [8] Department of Health. (2017). DOH Administrative Order No. 2017-0010: Philippine National Standards for Drinking Water.https://www.fda.gov.ph/wp-content/uploads/ 2020 / 10 / Administrative-Order-No.-2017-0010.pdf [9] Apostol, G. L. C., Valenzuela, S., & Seposo, X. (2022). Arsenic in groundwater sources from selected communities surrounding Taal Volcano, Philippines: An exploratory study. Earth, 3(1), 448–459. https://doi.org/10.3390/earth3010027 [10] Commission on Audit. (2021). Laguna Water District, Laguna Annual Audit Report 2020.https: //www.coa.gov.ph/wpfd_file/laguna-water-district-laguna-annual-audit-report-2020/ [11] Commission on Audit. (2023). Laguna Water District, Laguna Annual Audit Report 2022.https: //www.coa.gov.ph/wpfd_file/laguna-water-district-laguna-annual-audit-report-2022/ [12] Nidheesh, P., & Singh, T. A. (2017). Arsenic removal by electrocoagulation process: Recent trends and removal mechanism. Chemosphere,181, 418–432. https://doi.org/ 10 . 1016 /j. chemosphere.2017.04.082 [13] Song, P., Yang, Z., Zeng, G., Yang, X., Xu, H., Wang, L., Xu, R., Xiong, W., & Ahmad, K. (2017). Electrocoagulation treatment of arsenic in wastewaters: A comprehensive review. Chemical Engineering Journal,317, 707–725. https://doi.org/10.1016/j.cej.2017.02.086 [14] Amrose, S. E., Bandaru, S. R., Delaire, C., van Genuchten, C. M., Dutta, A., DebSarkar, A., Orr, C., Roy, J., Das, A., & Gadgil, A. J. (2014). Electro-chemical arsenic remediation: Field trials in West Bengal. Science of the Total Environment,488, 539–546. https://doi.org/ 10 . 1016 /j. scitotenv.2013.11.074 [15] van Genuchten, C. M., Peña, J., Amrose, S. E., & Gadgil, A. J. (2014). Structure of Fe (III) precipitates generated by the electrolytic dissolution of Fe (0) in the presence of groundwater ions. Geochimica et Cosmochimica Acta,127, 285–304. https://doi.org/ 10 . 1016 /j.gca. 2013 . 11 . 044 [16] Delaire, C., Amrose, S., Zhang, M., Hake, J., & Gadgil, A. (2017). How do operating conditions affect As (III) removal by iron electrocoagulation? Water research,112, 185–194. https://doi. org/10.1016/j.watres.2017.01.030 Rubin et al. (2024) | Journal of Human Ecology and Sustainability 21 [17] Li, L., Li, J., Shao, C., Zhang, K., Yu, S., Gao, N., Deng, Y., & Yin, D. (2014). Arsenic removal in synthetic ground water using iron electrolysis. Separation and Purification Technology,122, 225–230. https://doi.org/10.1016/j.seppur.2013.11.012 [18] Wan, W., Pepping, T. J., Banerji, T., Chaudhari, S., & Giammar, D. E. (2011). Effects of water chemistry on arsenic removal from drinking water by electrocoagulation. Water research, 45(1), 384–392. https://doi.org/10.1016/j.watres.2010.08.016 [19] VanGenuchten,C.M.,Bandaru,S. R., Surorova, E.,Amrose,S. E.,Gadgil,A.J.,&Pena,J.(2016). Formation of macroscopic surface layers on Fe (0) electrocoagulation electrodes during an extended field trial of arsenic treatment. Chemosphere,153, 270–279. https://doi.org/ 10 . 1016/j.chemosphere.2016.03.027 [20] Banerji, T., & Chaudhari, S. (2016). Arsenic removal from drinking water by electrocoagulation using iron electrodes-an understanding of the process parameters. Journal of Environmental Chemical Engineering,4(4), 3990–4000. https://doi.org/10.1016/j.jece.2016.09.007 [21] Amrose, S., Gadgil, A., Srinivasan, V., Kowolik, K., Muller, M., Huang, J., & Kostecki, R. (2013). Arsenic removal from groundwater using iron electrocoagulation: Effect of charge dosage rate. Journal of Environmental Science and Health, Part A,48(9), 1019–1030. https://doi.org/ 10.1080/10934529.2013.773215 [22] Dutta, N., Haldar, A., & Gupta, A. (2021). Electrocoagulation for arsenic removal: Field trials in rural West Bengal. Archives of environmental contamination and toxicology,80(1), 248–258. https://doi.org/10.1007/s00244-020-00799-8 [23] Kobya, M., Demirbas, E., & Ulu, F. (2016). Evaluation of operating parameters with respect to charge loading on the removal efficiency of arsenic from potable water by electrocoagulation. Journal of environmental chemical engineering,4(2), 1484–1494. https://doi.org/ 10 . 1016 /j. jece.2016.02.016 [24] Balares, K. L., Nuevo, J., Chiong, M., Macasieb, R., Resurreccion, A., & Orozco, C. (2020). Estimating Energy Consumption and Cost for the Electrocoagulation of Arsenic-laden Water (ECAR) Using Iron Electrodes. E3S Web of Conferences,148, 01004. https://doi.org/ 10 . 1051 / e3sconf/202014801004 [25] van Genuchten, C. M., Gadgil, A. J., & Peña, J. (2014). Fe (III) nucleation in the presence of bivalent cations and oxyanions leads to subnanoscale 7 Å polymers. Environmental science & technology,48(20), 11828–11836. https://doi.org/10.1021/es503281a [26] Catrouillet,C.,Hirosue,S.,Manetti,N., Boureau,V.,&Peña,J.(2020).Coupled As andMnredox transformations in an Fe (0) electrocoagulation system: Competition for reactive oxidants and sorption sites. Environmental science & technology,54(12), 7165–7174. https://doi.org/ 10.1021/acs.est.9b07099 [27] Dutta, N., & Gupta, A. (2022). Development of arsenic removal unit with electrocoagulation andactivatedalumina sorption:Fieldtrialatruralwestbengal,india.Journal of Water Process Engineering,49, 103013. https://doi.org/10.1016/j.jwpe.2022.103013 [28] Schloerke, B. e. a. (2024). GGally: Extension to “ggplot2". cran.r-project. https://cran.rproject.org/web/packages/GGally/GGally.pdf [29] Wood, S. N. (2011). Fast stable restricted maximum likelihood and marginal likelihood estimationofsemiparametricgeneralizedlinearmodels.Journalof theRoyalStatisticalSocietySeries B: Statistical Methodology,73(1), 3–36. https://doi.org/10.1111/j.1467-9868.2010.00749.x [30] Liu, Y., Li, Y., Yin, W., Wang, H. - c., Zhao, X., Liu, X., Lu, S., & Wang, A. - j. (2024). Long-term performance of a deep oxidation pond with horizontal subsurface flow constructed wetland for purification of rural polluted river water. Environmental Research,240, 117498. https: //doi.org/10.1016/j.envres.2023.117498 Rubin et al. (2024) | Journal of Human Ecology and Sustainability 22 [31] Liu, X., Wang, Y., Lu, S., Liu, Y., Zhao, B., Xi, B., Guo, X., Guo, W., & Zhang, J. (2020). Intensified sulfamethoxazoleremovalinanelectrolysis-integratedtidalflowconstructedwetlandsystem. Chemical Engineering Journal,390, 124545. https://doi.org/10.1016/j.cej.2020.124545 [32] Amin, R., Salan, M. S. A., & Hossain, M. M. (2024). Measuring the impact of responsible factors on CO2 emission using generalized additive model (GAM). Heliyon,10(4), e25416. https: //doi.org/10.1016/j.heliyon.2024.e25416 [33] Ye, Z., Hong, S., He, C., Zhang, Y., Wang, Y., Zhu, H., & Hou, H. (2022). Evaluation of different factorsonmetalleachingfromnickeltailingsusinggeneralizedadditivemodel(GAM).Ecotoxicology and Environmental Safety,236, 113488. https://doi.org/ 10 . 1016 /j.ecoenv. 2022 . 113488 [34] da Silva Marques, D., Costa, P. G., Souza, G. M., Cardozo, J. G., Barcarolli, I. F., & Bianchini, A. (2019). Selection of biochemical and physiological parameters in the croaker Micropogonias furnieri as biomarkers of chemical contamination in estuaries using a generalized additive model (GAM). Science of the Total Environment,647, 1456–1467. https://doi.org/ 10 . 1016 /j. scitotenv.2018.08.049 [35] Wood, S. N. (2017). Generalized additive models: An introduction with R (Texts in Statistical Science). Chapman Hall/CRC. [36] Arel-Bundock, V., Greifer, N., & Heiss, A. (2024). How to interpret statistical models using marginaleffectsinRandPython.Journalof StatisticalSoftware,55(2),1–31.https://marginaleffects. com/assets/marginaleffects_arel-bundock_greifer_heiss_jss5115.pdf [37] Dubrawski, K. L., van Genuchten, C. M., Delaire, C., Amrose, S. E., Gadgil, A. J., & Mohseni, M. (2015). Production and transformation of mixed-valent nanoparticles generated by Fe (0) electrocoagulation. Environmental science & technology,49(4), 2171–2179. https://doi.org/ 10.1021/es505059d [38] Mosivand, S., & Kazeminezhad, I. (2015). A novel synthesis method for manganese ferrite nanopowders: The effect of manganese salt as inorganic additive in electrosynthesis cell. Ceramics International,41(7), 8637–8642. https://doi.org/10.1016/j.ceramint.2015.03.074 [39] Dancey, C., & Reidy, J. (2020). Statistics without maths for psychology. Pearson Education. Rubin et al. (2024) | Journal of Human Ecology and Sustainability 23