Comparative analysis of two low cost point-of-use water treatment systems

Water treatment systems, particularly, for developing nations, do not always need to be sophisticated or automated to be effective and useful, but should be able to produce bacteria free and aesthetically acceptable water. Hence, for point-of-use (PoU) water treatment systems, the safety of the water is of utmost concern. According to McAllister (2005), viruses and chemical pollutants cause far fewer problems as a result of drinking untreated water than bacteriological agents. The first and most important step in the battle against consumption of poor quality water is thus the elimination of bacteria (McAllister 2005) and particles, so that users do not opt for water that looks aesthetically better but is actually contaminated (Center for Affordable Water & Sanitation 2011). PoU technologies have been proposed for providing safe water in developing countries (Sobsey 2002), as opposed to centralized water supply systems, since they minimize the risk of contamination between the water treatment plants and users. Many design guidelines and criteria exist for conventional water treatment systems (Kawamura 2000; Davis 2010), whereas PoU systems have varying guidelines making them vulnerable to quality and performance variability. There are very few low cost water systems that are well designed and produced, and give excellent sustainable performance. Comparative analysis on two, low-cost PoU water treatment systems was carried out in the water quality lab at Stellenbosch University, South Africa. The systems studied were the Gift of Water System (GWS) (Gift of Water, Inc. USA) and the Drip Filter System (DFS) Model-JW-PD-1-70 (Headstream Water Holdings, South Africa). Both are relatively well produced and affordable (DrinC 2017; Gift of Water Inc. 2017).

A number of PoU systems worldwide can treat various types of contaminated water (McAllister 2005). However, most are expensive and fail to meet the specific needs of poor communities (McAllister 2005). The quality of many affordable PoU systems depends largely on the materials used and the fabricator's ability. In this study two, low-cost PoU technologies, developed and produced respectively in the USA and South Africa, were assessed by exposure to a single surface water source for three months. The comparison was based mainly on the efficiency of bacterial and particulate removal from poor quality urban stream water. Bacterial removal efficiency was estimated using E. coli and thermo-tolerant (fecal) coliforms as indicator organisms (Ritter 2010). Particulate reduction efficiency was tested using turbidity and total suspended solids (TSS) (Shammas & Wang 2015).

The cost of the PoU systems considered is relatively low compared to other modern PoU systems based on, e.g., ion exchange, reverse osmosis and other advanced technologies (de Moel et al. 2007; Ritter 2010; WHO 2011). An attempt was also made to determine the systems’ effectiveness and how their treated outputs compare to good quality potable water produced by high-tech, or excellently maintained and operated, conventional water treatment systems (Howe et al. 2012). The systems’ treated water quality was therefore compared to the good quality tap water municipal supply at Stellenbosch University. It is noted in this context that the municipal supply at Stellenbosch comprises: (i) screening at the reservoir, to remove suspended matter and floating debris; (ii) pre-chlorination; (iii) cascade aeration; (iv) pH correction with hydrated lime; (v) coagulation and flocculation with aluminum sulfate or sodium aluminate; (vi) sedimentation; (vii) rapid gravity sand filtration; (viii) stabilization with lime; and, (ix) chlorination.

The potable water supplied to the university is obtained from surface water in the same catchment as Kromrivier stream, the raw water source for the PoUs. The reservoir supplying water to the treatment plant receives some run-off from agricultural land and has recently recorded increases in algae and turbidity (Enviro Metsi (Pty) Ltd 2017), as does the Kromrivier stream.

The study was designed to compare the PoUs ability to improve both particulate and bacterial quality of water. Most PoU studies focus on bacterial removal and neglect the removal of particulates. According to the Center for Affordable Water & Sanitation (2011) PoUs must provide clear water (with little or no turbidity) so that users do not opt for water that looks aesthetically better but is actually contaminated.

Raw surface water samples were obtained from Kromrivier, a small stream in Stellenbosch, South Africa, at 33°55′34.68″S and 18°51′40.56″E, next to the bridge between Ryneveld Street and Kromrivier Road, Stellenbosch, see Figure 1.

GWS (Figure 2) comprises low-cost water treatment technology for developing countries. It was developed initially for use in Haiti to combat water-borne diseases and complications from malnourishment arising from drinking unsafe water (Gift of Water Inc. 2017). It is a two-bucket system that uses a 1 micron (μm) polypropylene string filter, a granular activated carbon (GAC) filter and chlorine tablets. The chlorine tablets are made of Sodium Dichloroisocyanurate (NaDCC) which dissociates in water to release hypochlorous acid (HClO) that kills microorganisms through oxidization (Center for Affordable Water & Sanitation 2011; WHO 2016, 2003). Raw water is put into a 20-liter top bucket, with a 67 mg NaDCC tablet, and left for 30 minutes. A 17 mg NaDCC tablet is then added to the bottom bucket for post-chlorination, to ensure that the chlorine concentration remains high enough to prevent recolonization by (most) bacteria. The top bucket is placed on the bottom bucket, activating a check-valve enabling water to flow into the bottom bucket, passing in transit through the string and GAC filters. The former removes suspended solids and larger organisms like protozoa, the latter – the GAC filter – removes organic compounds and excess chlorine (Gift of Water Inc. 2017). Gift of Water Inc. (2017) recommend replacement of the GAC filter every 6 months. Treated water is available through a tap in the bottom bucket. The average flow through the GWS is estimated at 46.8 L/h (Gift of Water Inc. 2017), and the system costs 25 USD in the USA.

DFS (Figure 3) is a low-cost, two-bucket, ceramic candle, filter system distributed under the name DrinC. It costs about South African Rand (ZAR) 600 (44 USD) in South Africa. The DFS candle filter, normally wedged between two 20-liter buckets, consists of a 0.2 μm, silver-impregnated ceramic shell containing activated carbon (charcoal) (DrinC 2017). The silver serves as a disinfectant. According to DrinC (2017), the ceramic shell sometimes has a fabric cover (filter sock) to remove larger debris (e.g. leaves and insects) from the source water. As water drips through the filter, suspended solids are removed, followed by bacteria and micro-organisms down to 0.2 μm. Raw water is put into the top bucket and drips through the filter into the bottom bucket, which is mainly for storage and is fitted with a tap. The GAC lasts for about 6 to 8 months and the filter must be replaced after one year's use (DrinC 2017), but it is advisable to shake it every 3 months to dislodge debris and extend its life, and ensure that the carbon stays loose. The DFS flow rate can be up to 13.26 L/h, when the system is new. During the study it was observed that the flow rate falls over time.

The tap water used during the study was collected daily from the Civil Engineering Department water quality laboratory, Stellenbosch University. Samples were analyzed immediately after collection. The key treatment steps are outlined in the introduction, above, and shown in Figure 4.

To ensure that the study's source water was contaminated with bacteria and particulate matter, raw water samples were collected and tested for at least two weeks before the evaluation tests started. In both weeks fecal coliforms reported more than 500 Colony Forming Units (CFU)/100 ml and E. coli more than 400 CFU/100 ml. TSS and turbidity were consistently above 14 mg/l and 10 Nephelometric Turbidity Units (NTU), respectively. The concentrations of fecal coliforms, E. coli, TSS, and turbidity were quantified before and after treatment by each system. Other parameters measured were electrical conductivity (EC), total dissolved solids (TDS), pH, and dissolved oxygen (DO).

The evaluation was done over a period of 3 months and 2 weeks. The raw water was passed through the PoUs five days a week for 2 to 3 hours each day, to mimic normal daily use as closely as possible. Treated water was collected fortnightly for bacteriological tests and daily for physico-chemical tests. Tests for E. coli and fecal coliforms were done by Water Analytical Laboratory (WALAB) accredited to the South African National Accreditation System (SANAS), No: T0375 for microbiological analysis. Physico-chemical tests were done in the Water Quality Laboratory at Stellenbosch University. All tests were performed in accordance with the Standard Methods for the Examination of Water and Wastewater (APHA/AWWA/WEF 2012).

where

• C_O = concentration of contaminant in untreated water

• C_e = concentration of contaminant in treated water

The source water was characterized initially and on every sampling day during the study for the selected parameters. The raw water characteristics were compared to the South African National Standards (SANS) 241 and the World Health Organization (WHO) guidelines for domestic and drinking water – see Table 1. In addition to high turbidity and suspended solids, the raw water was highly contaminated with fecal coliforms and E. coli.

The GWS and DFS were both very effective in bacterial removal, recording 100% in terms of both E. coli and fecal coliforms (Figures 5 and 6). The DFS gave similar results to those in a study by Adeyemo et al. (2015) and Center for Affordable Water & Sanitation (2011) on bacteria removal by silver-coated ceramic candle filters (bacterial removals >99% and >99.95% of laboratory and field treatment efficiency, respectively). The GWS gave results similar to Lantagne et al. (2006) and Nath et al. (2006), who reported bacterial removal efficiencies for systems using combined filtration and chlorination >99.99%. It is therefore clear that both systems can meet WHO and SANS 241 standards. The authors believe that the two PoUs may often offer advantages over centralized water treatment systems by minimizing the risk of contamination between the source and the point-of-use, particularly in poor communities. In many countries, centralized systems commonly suffer from recontamination between water treatment plants and point-of-use, e.g., because of infrastructure failures connected with water storage and/or distribution. The bacteriological quality of the raw water improved by a factor of between about 2 and 3 during the course of the tests, and there was a noticeable surge in suspended solids concentrations around the middle of the tests, but this did not coincide with the bacterial count peaks, which were earlier. This was because the authors stirred the streambed gently for 3 weeks (31 July to 22 August) to collect water with higher particulate content to test the PoUs. Most suspended solids comprise inorganic materials (clay, silt, sand, etc.), although bacteria and algae also contribute to suspended solids concentrations (Ritter 2010; Howe et al. 2012). The rather higher bacterial counts recorded earlier could be attributed to increased bacterial inflow from the storm drain that discharges about 20 m upstream of the sampling point, especially as the study area experiences winter rains at that time of the year. The storm drain is thought to collect fecal material from underneath the bridge (on Jan Celliers Road close by) where some homeless people find shelter. Evidence for this claim was noted on the day of the highest bacterial count (17 July) when the storm drain discharge looked and smelt like sewage.

There was substantial removal of TSS and turbidity from the raw water by both GWS and DFS – see Figures 7–10. DFS was slightly better than GWS in TSS and turbidity removal, but particulate removal was highly significant by both. On average TSS reductions were 89 and 95% for GWS and DFS, respectively, and 87 and 94% for turbidity. The higher particulate removals by DFS could be attributed to the smaller pore size of its filter, which is 0.2 μm vs 1 μm for GWS. While neither GWS nor DFS met the WHO guideline level (0.1 mg-TSS/l) consistently, they always removed a very large fraction of the particles from the water and their effluent TSS values were little higher than those in the tap water – see Figure 7. The turbidity of the treated waters from both systems consistently met the WHO and SANS 241 level of 5 NTU, and compared well to that of the tap water.

There was a noticeable difference in performance between the two systems until around 19 July, after which, similarity was observed until almost the study end. The DFS performed better than GWS at first, until, the coarser filter of the latter began to clog, when their performances became fairly similar. Additionally, comparative performance of GWS and DFS in relation to turbidity was almost identical to that of the pair when removing TSS, which is not particularly surprising because turbidity and TSS complement each other, and are similar in the sense that both are measures of water clarity although they reflect different issues (Ritter 2010). Even though they cannot be directly correlated, turbidity and TSS overlap in the measurement of some particles such as bacteria, algae, clay, silt and non-settleable solids (Ritter 2010; Howe et al. 2012).

Since both systems contain a disinfection step and produce relatively clear water, they are good options for improving water security in poor communities, especially if produced locally and promoted by Non-governmental organizations (NGOs), who should ensure adequate user motivation and training.

Although DO levels in the DFS water were relatively low (Figure 11), both systems consistently met the SANS 241 and WHO guidelines– see Table 1 – in terms of pH. It was clear that the PoUs had little if any effect on the raw water's pH (Figure 12), EC or TDS, but they were not designed to do so. It is a good idea, therefore, to obtain raw water whose chemical content is reasonably close to the potable water guidelines when using these PoUs.

Table 2 gives key comparisons between GWS and DFS. The PoUs can both meet basic water needs of about 15 to 20 liters/capita/day (WHO 2016), particularly for poor communities. The main drawback with respect to GWS is the potential for the production of disinfection-by-products – e.g., trihalomethanes – due to the use of NaDCC tablets in both the top and bottom buckets, especially if the GAC, which removes excess chlorine, fails during use. The major drawback with DFS is the slow filtration flow rate and regular filter cleaning to remove clogging.

Table 3 below gives average raw water quality for the PoUs vs treated water quality in comparison with tap water quality over the study period. There was no significant difference between the tap water quality and PoUs treated water quality in terms of pH, TSS or turbidity.

The results show that both GWS and DFS can treat the urban stream water used, and produce water that meets the SANS 241 standards and WHO guidelines with respect to the parameters measured. The treated water from the PoUs compared well with good quality tap water supplied to Stellenbosch University with respect to bacterial, turbidity and suspended solids content. Both systems are relatively low cost water treatment solutions, with their own benefits and drawbacks. Both can improve the quality of the raw surface water in terms of bacterial counts and clarity. The study also included seasonal variations in water quality to some extent, as it ran from winter through spring to summer. Since PoUs (such as these) often offer advantages over centralized water treatment systems by minimizing the risk of contamination between the source and the point-of-use, they are a good option to help improve water security in many communities over the world.