Final Report: Comparative Analysis of Three Sustainable Point of Use Drinking Water Treatment Technologies for Developing Nations

EPA Grant Number: SU831831
Title: Comparative Analysis of Three Sustainable Point of Use Drinking Water Treatment Technologies for Developing Nations
Investigators: Sobsey, Mark D. , Stauber, Christine , Whittington, Dale , Brown, Joe , Casanova, Lisa , Elliott, Mark
Institution: University of North Carolina at Chapel Hill
EPA Project Officer: Page, Angela
Phase: I
Project Period: October 1, 2004 through April 30, 2005
Project Amount: $9,994
RFA: P3 Awards: A National Student Design Competition for Sustainability Focusing on People, Prosperity and the Planet (2004) RFA Text |  Recipients Lists
Research Category: P3 Challenge Area - Water , Pollution Prevention/Sustainable Development , P3 Awards , Sustainability

Objective:

The purpose of this research is to determine the costs, health and economic benefits, and effectiveness in improving water quality of different POU drinking water treatment technologies intended for the developing world. This project focuses on three distinct POU technologies: coagulation-flocculation-chlorination, ceramic microfiltration, and intermittently operating slow sand filtration (IOSSF), also known as the biosand filter. For this project we performed two major analyses of each system: effectiveness and cost-benefit. The first was an analysis of the effectiveness of the technology for improving water quality for removing bacteria and viruses, which are major causes of waterborne disease. Effectiveness was measured by the ability of each system to reduce enteric bacteria, enteric viruses, and coliphage MS-2 (a surrogate for a wide range of pathogenic human viruses) from water in bench-scale testing. Each system was challenged with natural waters spiked with test microorganisms. The log10 reductions of each microorganism by each system were measured, and each was evaluated for its ability to achieve 4 logs (99.99%) of inactivation or removal of these microorganisms, in accordance with U.S. Environmental Protection Agency standards for POU drinking water treatment systems.

The second major analysis in this project, which was begun in Phase I and will be completed in Phase II, will be a detailed analysis of the costs of each technology, including manufacturing, distribution, and time to operate and maintain the POU system, and the relationship of these costs to the benefits of use. The benefits measured will include reduction in diarrheal disease by users as compared to non-users, and the attendant benefits such as reductions in lost work time and resources spent on health care. We will also measure the benefits of health outcomes in terms of Disability Adjusted Life Years avoided (DALYS). This information will make it possible for people in the developing world who could benefit from POU to identify the technology that is the most appropriate, effective and sustainable for their homes and communities.

Summary/Accomplishments (Outputs/Outcomes):

Laboratory testing for ceramic microfiltration was focused primarily on establishing the ability of commercially available ceramic filters to reduce viruses in water, and enhancing this property. Metal oxide coatings have been observed to capture and inactivate viruses under certain conditions. Our laboratory group has developed novel ceramic filters incorporating low cost, naturally occurring iron oxides to enhance the virus reductions by ceramic filter technology. Filter testing was done using bacteriophages, viruses used as models for human pathogenic viruses. Ceramic filters coated with iron compounds achieved 4-7 log10 reduction of phages in spiked groundwater samples, compared with only 1-2 log10 reductions by conventional ceramic filters.

The ability of the PUR (Procter&Gamble) coagulation-flocculation-chlorination system to remove Aeromonas hydrophila, a known waterborne pathogen in both the developing and developed world, and hepatitis A, a cause of waterbome infectious hepatitis, was determined. Test waters were untreated natural surface water to simulate drinking water sources that might be used in the developing world. The average reduction of A. hydrophila after 20 minutes of exposure was 99.9994% (5.2 log10). In order to determine the effects of turbidity from contamination with human fecal waste on disinfectant efficacy, test waters were mixed with primary sewage effluent. In waters contaminated with 10% primary effluent, average reduction after 20 minutes of contact time was 99.98% (3.8 log10), as compared to 99.9994% (5.2 log10) for uncontaminated test waters. Reduction of A. hydrophila in waters contaminated with 20% primary effluent was comparable to that in waters with 10% effluent. Reduction of A. hydrophila ranged from 3.7 to 4.1 log10 (99.98 to 99.992%) after 20 minutes. Average reduction after 20 minutes of contact time was 99.98% (3.8 log10). Overall reduction ranged from 1+ log10 (94.56%) in water contaminated with primary effluent that was stored before treatment, to 6.5 log10 (99.99998%) in uncontaminated lake water. The system was still effective in removing up to 99.999% of A. hydrophila even in water heavily contaminated with fecal matter.

Greater than 3 log10 (>99.9%) reduction of HAV was achieved after 10 minutes contact. In test waters with turbidity ranging from 3.18-4.67 NTU, the coagulation-flocculation-chlorination system removed >3log10 (>99.9%) HAV after 10 minutes of contact time. When treated samples were further concentrated to increase lower limits for virus detection, >5.2 log10 (>99.9993%) reduction of HAV could be observed after 10 minutes. This is exceeds the goal of achieving 4 log10 virus reduction for the technologies tested.

Experiments were conducted with full-scale IOSSF units to simulate as closely as possible the typical field conditions in developing countries, inclusive of sieving the filter media (crushed granite gravel media sieved through mosquito screen and washed manually) and a daily charge of 40 L of surface water, typical of family use. Three extended time, full-scale filter experiments (21, 45 and 53 days of filter operation) were conducted with spiked microorganisms, along with numerous short-term tests related to filter flowrate and hydraulic conditions. A step input form of tracer test with NaCl solution established that the filter operates at conditions very close to plug flow. Perfect plug flow conditions correspond to filter flow with no mixing or dispersion, so these results indicate very little mixing taking place within the filter. Initially, the IOSSF acts like a simple sand filter. But after days or weeks of operation, biological activity begins within the sand bed, especially at the sand surface, and the filter begins to mature. Filter maturation, or ripening, results in a decreasing flowrate as the pores in the upper layer of sand become clogged. Over three-to-five weeks, the flowrate can decrease by 1/4th to 1110th it's the original rate (about one-liter per minute when the influent chamber full). Another phenomenon of filter ripening is increasing effectiveness in reducing microorganism concentrations in the filtered water. As a biological community develops in the filter sand, microbial reductions improved. Results for E. coli indicate that there was a filter ripening time during which the clean filter acquired increased ability for removal. Because the IOSSF nearly simulates plug flow conditions, E. coli removal efficiency was examined in relative to pore volumes filtered. The early grab samples (those representing mostly water that had been sitting in the filter overnight) demonstrate much greater removal efficiency than the later samples, with E. coli reductions that averaged >99.5%. The middle samples, representing a mixture of older and newly loaded water, had less extensive E. coli reductions These data indicate that one of the unique characteristics of intermittent slow sand filtration, the pause time of the filter, is critical to microbial removal efficiency. Microbial reductions increase as the filter matures, with maximum E. coli reductions achieved at the end of the filter run.

In addition to E. coli, the filter was examined for its ability to reduce three viruses: two bacterial viruses (phages MS-2 and PRD-l) and one human pathogenic virus (echovirus type 12). Maximum and average removals for viruses varied between viruses and filter runs but were lower overall than those for E. coli. Average echovirus reductions over a filter run were about 92%, with 98.8% removal in water that remained in the filter overnight. The bacterial viruses were reduced less extensively, with average removals of 78% for PRD-l and 87% for MS-2. As with E. coli, greater microbial reductions were found consistently in the portion of the filtered water that had been retained in the filter overnight than in the subsequent water passing through the filter on the same day.

Conclusions:

All three technologies were able to remove bacteria and viruses from water. The biosand filter showed lower reductions of bacteria and viruses than the other technologies. The basic physical and biological aspects of this filter are still not well characterized, and continued work in Phase II to better understand these mechanisms may lead to filter modifications that can achieve greater microbial reductions. The coagulation-flocculation-chlorination system removed up to 5 log10 of bacteria and viruses, suggesting that the combination of physical and chemical mechanisms can achieve high levels of microbial removal. The F2 ceramic filter, which is a conventional filter modified with an iron coating developed in our laboratory, showed improved virus capture, with >5 log10 reduction compared to 1-2 log10 reduction for conventional filters.

Proposed Phase II objectives and strategies:

The overarching goal of the entire proposed P3 project is to allow for a comprehensive comparison of effectiveness, health benefits, and cost- benefit relationships for these three POU technologies. In order to complete this comprehensive analysis begun in Phase I, we must fill three crucial data gaps: more complete data on the effectiveness of the biosand and ceramic filters in removing microbes from drinking water, rigorous documentation of the health benefits of these technologies under actual use conditions and cost-benefit analysis based on these health data. Phase II will fill these data gaps and allow for a comprehensive comparison of the effectiveness, costs, and benefits of all three of these technologies. Accordingly, the goals of Phase II are to characterize more completely the biosand filter's performance in the laboratory and the field, to assess the ability of the BSF and ceramic filter to reduce diarrheal diseases of users and to perform a cost-benefit analysis of all three technologies using the data on health benefits collected in the field studies. (Health impact data for the coagulation-flocculation-chlorination technology already exist and do not need to be obtained in new field studies.)

Biosand Filter

The first study aim is to isolate the parameters that control filtration performance and investigate mechanisms of microbial removal in the IOSSF. Mechanisms of removal can be investigated by isolating physical-chemical effects from the biological effects of the filter microbial community. This can be accomplished by addition of microbial activity inhibitors such as sodium azide and microbial protease inhibitors. The role of dissolved oxygen (DO) in maintaining an active microbial community is unclear from the literature and personal communication with other researchers (Buzunis, 1995; Weber-Shirk, 2004). We propose to elucidate the hypothesized role of DO by monitoring DO at different levels in the column and by conducting column studies with oxygen-rich or oxygen-poor headspace. We also will do constant flow operated column studies to enable direct comparison with the intermittent mode of operation. Previous studies have been deficient in analysis of the filter media, so morphological examinations by scanning confocal laser microscopy will be used to better understand the physical, chemical biological properties of active biosand filter media.

The second study aim is to determine the ability of the BSF to reduce Cryptosporidium parvum oocysts in the laboratory. We will characterize its removal from seeded raw water under constant daily dosing conditions over the ripening period of the filter and through a typical filter run. Microbial removals of the biosand filter in these studies will be compared to removals of the same or similar microbes by conventional slow sand filters as previously reported in the literature. Filters will be dosed with 40L water/day of lake water containing constant influent concentrations of C. parvum. Influent and effluent concentrations of C. parvum will be measured via direct microscopic techniques using fluorescently labeled antibodies. We will measure the daily microbial reductions periodically until the filter achieves steady state removal or filter operation reaches 42 days (6 weeks).

The third study aim is to measure the ability of the BSF to reduce total coliforms and E. coli in the field. Water samples will be collected and analyzed bi-weekly in 75 households using the BSF technology and 75 control households not using the BSF in a village in the Dominican Republic. We will also measure pH, temperature, turbidity, and chlorine levels. Household water supplies will be sampled over time to account for seasonal variability of water quality and the impacts of periodic rainfall events that typically degrade water quality. The levels of indicator microbes and other water quality parameters will be compared for untreated waters in control and BSF filter households and for BSF-treated household water. We will determine if the quality of untreated water of control and BSF households is the same and we will determine the extent of reduction of microbial and other contaminants by BSF treatment (on a percentage and log10 basis). The quality of household water will be statistically compared in control households and in BSF households. These comparisons will be made using either parametric or non-parametric statistical methods, depending on the statistical distributions of water quality parameters.

The fourth study aim is to determine the ability of the BSF to reduce diarrheal disease incidence rates >15% in children under 5 in user households as compared to non-user households. A longitudinal, prospective cohort study will be performed with consenting members of the community of Jayaco, near the city of Bonao in the Dominican Republic, using questionnaires that were created in Phase 1. The study will have two phases: a background or baseline data collection phase and an intervention collection phase in which randonmly selected households will get biosand filters and other will not (at least until the intervention observation period is completed). Both baseline and intervention periods will include the dry season and the wet season to account for seasonal differences in water sources and quality, seasonal differences in diarrheal illness rates and other seasonal changes.

Ceramic Filter

The first study aim is an assessment of the capability of a novel ceramic filter, the F2 filter developed in our laboratory, to remove microbes in a field intervention setting in Cambodia. The hypothesis of phase is that the ceramic filter is an effective barrier to pathogens potentially present in drinking water. To this end, a community intervention trial will be conducted in rural Cambodia with an NGO partner who has agreed to facilitate a comprehensive field evaluation of the technology as part of their implementation strategy. Resources Development International - Cambodia has established a filter workshop and has begun production, but technological validation is needed before the project can be scaled up. The intervention will take place in the village of Slab Ta Oun, near Phnom Penh, Cambodia.

The second study aim is an assessment of health impacts of ceramic filters in this field intervention setting in the village of Slab Ta Oun. The study design is a triple-blinded, randomized controlled trial, which is the most scientifically rigorous method of hypothesis testing available to assess risks from drinking water. Households will be randomly assigned to one of three groups: those receiving the new and improved filter, those receiving no filter at all, and those receiving a placebo filter. The health impacts of the filter will be based on bi-weekly interviews of all participating households, usually conducted at the time of water quality sampling. These will be conducted by a local health worker with experience in survey data collection.

The third study aim is a sustainability analysis of a ceramic filter in a field intervention setting. Continued effectiveness of the filter as measured by microbiological and health data over the period of one year will be the primary focus of the sustainability of the assessment. In addition to this, data on the durability (number of breakages, number of filters no longer in use for any reason, etc) and economic feasibility will be collected. A cost-benefit analysis will be conducted to evaluate the economic feasibility of the intervention. At the end of the trial, all study households will receive an F2 filter. A community workshop will be held to train users in the use and care of their filters, with the assistance of Resources Development International - Cambodia.

The final overall study aim is a comprehensive cost benefit analysis of the three technologies. Once health data have been obtained, they will be used as the basis for a detailed analysis of the costs of biosand and ceramic filter use, and the relationship of these costs to the benefits of use. The benefits measured will include reduction in diarrheal disease by users as compared to non-users, and the attendant benefits such as reductions in lost work time and resources spent on health care. We will also measure the benefits of health outcomes in terms of Disability Adjusted Live Years avoided (DALYS), using standard DALY weights for diarrheal diseases from the published literature. This analysis can then be integrated with the cost-benefit analyses preformed for the ceramic filter and the coagulation-flocculation-chlorination system from Phase I to provide a comprehensive picture of the comparative costs and benefits of the three technologies. This information will make it possible for people in the developing world who could benefit from POU to identify the technology that is the most appropriate and sustainable for their homes and communities.

Supplemental Keywords:

drinking water, treatment, health effects, point of use technology,, RFA, Health, Scientific Discipline, Water, POLLUTANTS/TOXICS, Environmental Chemistry, Arsenic, Risk Assessments, Environmental Monitoring, Water Pollutants, Drinking Water, monitoring, well water, exposure, arsenic monitoring, arsenic removal, point of use, human exposure, contaminant removal, drinking water treatment, human health, water treatment, other - risk management

Relevant Websites:

http://www.cawst.org/technology/watertreatment/filtration-biosand.php Exit
http://www.samaritanspurse.uk.com/projects/water.asp Exit
http://www.biosandfilter.org/ Exit
http://www.potpaz.org/ Exit
http://www.rdic.org/ Exit

Phase 2 Abstract
Phase 2 Final Report

P3 Phase II:

Comparative Analysis of Three Sustainable Point of Use Drinking Water Treatment Technologies for Developing Nations  | Final Report