Evaluation of the removal of heavy metals using the Biosand Filter Lesly J. Mamani Paco Laboratorios Analíticos del Sur, Peru CAWST Learning Exchange - June 2011

Reason & Purpose for Study The drinking water sources for the peri-urban communities surrounding Lima, Peru are contaminated with heavy metals from closed mines that had not been abandoned properly. This study was undertaken to evaluate if the biosand filter*, with and without amendment, could remove these heavy metal contaminants from the drinking water. * A sand filter sized for a household’s water needs www.cawst.org CAWST Learning Exchange - June 2011

CAWST Learning Exchange - June 2011 Objective To evaluate the removal of heavy metals* from drinking water; using the Biosand filter (BSF) amended with copper-zinc granules and activated carbon. * chromium (Cr), cadmium (Cd), lead (Pb) and iron (Fe) CAWST Learning Exchange - June 2011

Experimental procedure 10 biosand filters for each metal studied (40 in total) For 3 metals; Chromium (Cr), Cadmium (Cd) and Iron (Fe): 6 filters were amended with differing amounts of Cu-Zn and activated carbon (accessory) and 4 were not amended. In the experiments for Lead (Pb); all 10 filters were amended. Targeted 3 initial concentrations (Ci) of each metal: Ci ≈ LMP; Ci ≈ 10 · LMP; Ci ≈ 100 · LMP Where LMP is maximum allowable limit for drinking water; Cr = 50 µg/L; Cd = 3 µg/L; Pb = 10 µg/L; Fe = 300 µg/L (Peruvian Legislation) LMP: límite máximo permitido (Maximum limit allowed) CAWST Learning Exchange - June 2011

CAWST Learning Exchange - June 2011 Chromium (Cr) CAWST Learning Exchange - June 2011

Results: Percentage removal of Cr Chromium average removal (final concentration / initial concentration) in percent Removal % Ci (low): 20.6 µg/L Ci (intermediate): 192.2 µg/L Ci (high). 2,530 µg/L -For an initial low chrome concentration (near to LMP), the removal percentage is between 86-93 %. Maximum removal was obtained with the filter of 1000g of grains of Cu-Zn and 200 grams of activated carbon. The four filters without accessories have lower removals. For intermediate initial concentrations ( of approximately 10 times the LMP) the removal % oscillates between 97 and 99 %. The highest removal is for filters with grains Cu-Zn of 1000g de Cu-Zn, but it is not observed the influence of active carbon. The removal is greater in a maximum of 3%, that in the case of filters without accessories. -For high initial concentrations (100 times LMP), the % of removal is always higher than 99%. It is observed that when initial concentration of chrome increases, the removal percentage of chrome also increases. Filter Activated carbon (g) Cu-Zn granules (g)

Results: Percent removal with and without accessory Chromium removal percentage - High initial conc. (Ci) (final concentration/ initial concentration) in percent Removal WITH accessory Ci = 2,530 µg/L Removal WITHOUT accessory Ci = 1,545 µg/L Removal % Removal percentages for filters with accessories (grains of Cu-Zn) with initial high concentrations of chrome are always between 99,9 % In the case of the same initial concentration condition but for filters without accessory, the removal percentages are above 99%. Consequently, the difference between including or not accessory is 0,5% -1 % in Cr removal. Filter CAWST Learning Exchange - June 2011

CAWST Learning Exchange - June 2011 Results for Chromium Removal Effectiveness: In all tests, with and without accessory, Cr concentrations were reduced by 85% - 99.9% Higher removals were found for higher initial concentrations Improvement in removal WITH accessory versus WITHOUT accessory ranged from 1 -5% depending on initial concentration Little benefit apparent from greater amounts of Cu-Zn granules Output Concentration of Chromium: No result exceeded the maximum allowed limit: 50 µg/L pH Influence: In general, initial pH has no influence in chromium removal The increase of pH might be due to contact of water with the limestone in the filter bed. This is favorable for the removal process due to precipitation of chrome due to increase of pH. The basis of chrome precipitation theory in aqueous solutions is based on the solubility curve where appears the theoretical performance of chrome (that presents a minimum solubility of approximately pH 8). CAWST Learning Exchange - June 2011

CAWST Learning Exchange - June 2011 Cadmium (Cd) CAWST Learning Exchange - June 2011

Results : Percentage removal of Cadmium Cadmium average removal (final concentration / initial concentration) in percent Removal % Ci (low): 4.1 µg/L Ci (intermediate): 31µg/L Ci (high). 280.2 µg/L -For an initial low concentration of cadmio (next to LMP) the removal percentage is 88-91 %. No significant influence of Cu-Zn grains is noted. The same for activated carbon. For intermediate initial concentrations (of approximately 10 times LMP), the removal % is over 99 %. No significant influence is appreciated from Cu-Zn grains nor of activated carbon. For high initial concentrations (100 times LMP), removal % is always over 99,9%. No significant influence is appreciated from Cu-Zn grains nor of activated carbon. It is noted that as the initial concentration of cadmio increases, the percentage of cadmio removal also increase; in all the filters (independently that they have accessories or not) there are removal % higher than 90% Filter Activated carbon (g) Cu-Zn granules (g) CAWST Learning Exchange - June 2011

Results: Cadmium removal with and without accessory Cadmium removal percentage (High Ci) (final concentration/ initial concentration) in percent AVERAGE WITH accessory Ci = 280 µg/L AVERAGE WITHOUT accessory Ci = 253 µg/L Removal % It can be observed how the filters with and without accessories have no significant differences in the removal percentage Filter CAWST Learning Exchange - June 2011

CAWST Learning Exchange - June 2011 Results for Cadmium Removal Effectiveness: In all tests, with and without accessory, Cd concentrations were reduced by 88% - 99.9% Higher removals were found for higher Ci No improvement in removal WITH accessory versus WITHOUT accessory is apparent at any initial concentrations No benefit from any amount of Cu-Zn or activated carbon Output Concentration of Cadmium: No result exceeded the maximum allowed limit: 3 µg/L pH Influence: In general, initial pH has no influence in chromium removal The increase of pH might be due to contact of water with the limestone in the filter bed. This is favorable for the removal process due to precipitation of chrome due to increase of pH. The basis of chrome precipitation theory in aqueous solutions is based on the solubility curve where appears the theoretical performance of chrome (that presents a minimum solubility of approximately pH 8). CAWST Learning Exchange - June 2011

CAWST Learning Exchange - June 2011 Lead (Pb) CAWST Learning Exchange - June 2011

Results: Percent Removal of Lead Lead average removal Ci (low): 21.5 µg/L Ci (intermediate): 101.4 µg/L Ci (high). 890.2 µg/L Removal % In this case all the filters had accessories. -For an initial low concentration of lead (next to LMP), the removal percentage is between 27- 40 %. Apparently there is no influence of activated carbon and the highest removals are obtained with 750 and 1000 g of Cu – Zn grains. For intermediate initial concentrations (of approximately 10 times the LMP), the removal % is over 75 %. The higher removal is observed for 1000 g of Cu-Zn grains and 200 g of activated carbon. For initial high concentrations (100 times LMP), the removal % is always over 98%. No significant influence is noted from the Cu-Zn grains or activated carbon. It is observed that as the initial lead concentration increases, the lead removal percentage also increases. The difference in lead removal percentages might be explained by the affinity of some microorganisms for various heavy metals, included lead. Filter Activated carbon (g) Cu-Zn granules (g) CAWST Learning Exchange - June 2011

Results: Lead removal with and without accessory Lead removal percentage (High Ci) (final concentration/ initial concentration) in percent AVERAGE WITH accessory Ci = 890 µg/L AVERAGE WITHOUT accessory Ci = 879 µg/L Removal % Studies of filters without accessories were made to study their influence It can be observed how the filters with accessories have removal percentages slightly higher than the filters without accessories (between 0,3-1,6 %). Even though, in all the cases, the removal percentage is 97% or higher. Filter CAWST Learning Exchange - June 2011

CAWST Learning Exchange - June 2011 Results for Lead Removal Effectiveness: 20 – 40% removal for Low Ci (21.5 µg/L) 70 – 90% for Intermediate Ci (104.5 µg/L) 97 – 99% for High Ci (890.2 µg/L) No apparent change from differing amounts of Cu-Zn or activated carbon Output Concentration of Lead: Some results exceeded the maximum allowed limit of 10 µg/L especially at the higher Ci pH Influence: In general, initial pH has no influence in lead removal The increase of pH might be due to contact of water with the limestone in the filter bed. This is favorable for the removal process due to precipitation of chrome due to increase of pH. The basis of chrome precipitation theory in aqueous solutions is based on the solubility curve where appears the theoretical performance of chrome (that presents a minimum solubility of approximately pH 8). CAWST Learning Exchange - June 2011

CAWST Learning Exchange - June 2011 Iron (Fe) CAWST Learning Exchange - June 2011

Results: Average % of Iron removal Iron removal; average percentage Ci (low) = 315.3 µg/L Ci (intermediate) = 553.1 µg/L Ci (high) = 3146 µg/L Removal % -For an initial low concentration of iron (next to LMP) the removal percentage is between 77-89%. Apparently there is no direct influence of the accessory on the removal % as F39 had the best results lacking it of accessories. For intermediate initial concentrations (of approximately 20 times LMP) the removal % is between 93-97%. Once again it is not observed the influence of the accessory in the iron removal. For initial high concentrations (100 times LMP) the removal % is always over 98%. No significant influences is noted from Cu –ZN grains nor of activated carbon. It is observed that as the initial iron concentration increase the removal percentage also increase. The absence of influence of accessories in the removal or iron might be due to oxidation and later precipitation of ions soluble in iron II as hydroxides. - Filter Activated carbon (g) Cu-Zn granules (g) CAWST Learning Exchange - June 2011

Results: Iron removal with and without accessory Iron removal percentage (High Ci) (final concentration/ initial concentration) in percent AVERAGE WITH accessory Ci = 3,146 µg/L AVERAGE WITHOUT accessory Ci = 3,712 µg/L Removal % The comparative analysis of the results with or without accessory are not conclusive. In general, no influence of the accessory in the removal of iron can be observed; it might be due to the precipitation process and not to the presence of accessory. Filter CAWST Learning Exchange - June 2011

CAWST Learning Exchange - June 2011 Results for Iron Removal Effectiveness: 75 – 90% removal for Low Ci (315 µg/L) 90 – 97% for Intermediate Ci (553 µg/L) 98 – 99% for High Ci (3146 µg/L) No improvement in removal WITH accessory versus WITHOUT accessory is apparent at any initial concentrations No apparent change from differing amounts of Cu-Zn or activated carbon Output Concentration of Lead: No results exceeded the maximum allowed limit of 300 µg/L pH Influence: In general, initial pH has no influence in iron removal The increase of pH might be due to contact of water with the limestone in the filter bed. This is favorable for the removal process due to precipitation of chrome due to increase of pH. The basis of chrome precipitation theory in aqueous solutions is based on the solubility curve where appears the theoretical performance of chrome (that presents a minimum solubility of approximately pH 8). CAWST Learning Exchange - June 2011

Conclusions (Laboratory) Based on the results from these experiments, the biosand filter demonstrated good removal effectiveness with and without the use of Cu- Zn granules and activated carbon. Removal effectiveness increased with higher initial concentrations for all metals. No influence of the accessory, Cu-Zn granules and activated carbon, was observed in the cadmium and iron removal. For the case of chromium, the presence of the accessory improved removal a maximum of 5%. pH input had no influence in the removal effectiveness. pH increased in the effluent in all cases as a result of the water contacting the calcium carbonate in the concrete body of the filter. For all the metals (except the lead) all the output samples were under the guidelines determined by WHO. Conclusion of study: Conclusions that can be obtained from the study are as follows: Additional analysis should be made on this theme. BSF seems to be quite efficient in the removal of metals studied. The presence of accessories (Cu-Zn grains and activated carbon) do not improve removal results. CAWST Learning Exchange - June 2011

CAWST Learning Exchange - June 2011 Research Issues This study was short-term (approx. 2 months) with limited repeated sampling. Additional long-term testing is needed to determine if the removal effectiveness decreases with time. Detailed information regarding the filters (sand size, ripening time, daily throughput) was not available. The source water chemistry was not characterized. Further testing with different water sources is necessary to determine if the results are applicable generally. The biosand filter did remarkably well in removing these heavy metal contaminants. Further research of this topic is recommended to validate the results and address the research issues. CAWST Learning Exchange - June 2011