In this study, we report on field testing of ceramic water filters (CWFs) fabricated using a new method of silver application (using silver nitrate as a raw material) compared to conventionally manufactured CWFs (fabricated with silver nanoparticles). Both types of filters were manufactured at the PureMadi ceramic filter production facility in Dertig, South Africa. Thirty households received filters fabricated with silver nitrate (AgNO3), and ten of those households were given an extra filter fabricated with silver nanoparticles. Filter performance was quantified by measurement of total coliform and Escherichia coli (E. coli) removal and silver residual concentration in the effluent. Silver-nitrate CWFs had removal efficiencies for total coliforms and E. coli of 95% and 99%, respectively. A comparison of the performance of silver-nitrate and silver-nanoparticle filters showed that the different filters had similar levels of total coliform and E. coli removal, although the silver nitrate filters produced the highest average removal of 97% while silver nanoparticles filters recorded an average removal of 85%. Average effluent silver levels were below 10 ppb for the silver-nitrate and silver-nanoparticle filters, which was significantly below the Environmental Protection Agencies of the United States (EPA) and World Health Organization (WHO) secondary guidelines of 100 ppb. Silver-nitrate filters resulted in the lowest effluent silver concentrations, which could potentially increase the effective life span of the filter. A cost analysis shows that it is more economical to produce CWFs using silver nitrate due to a reduction in raw-material costs and reduced labor costs for production. Furthermore, the production of silver-nitrate filters reduces inhalation exposure of silver by workers. The results obtained from this study will be applied to improve the ceramic filtration technology as a point-of-use (POU) water treatment device and hence reduce health problems associated with microbial contamination of water stored at the household level.
Keywords: ceramic water filter (CWF); point-of-use (POU) water treatment technologies; silver nitrate; silver nanoparticles; waterborne diseases; public health
Access to clean, safe, and adequate amounts of water is a fundamental human need and, therefore, a basic human right. Microbes such as viruses, bacteria and protozoa are easily transported through drinking water. Ingestion of such pathogens in water leads to the greatest water-related health risks and is a major cause of waterborne diseases [[
Developing countries report the highest number of deaths related to waterborne diseases, particularly in rural areas [[
- A technology that is produced using locally available materials and labor;
- A technology that is cost-effective and easy to manufacture;
- A technology that is easy to operate and socially acceptable so that users are willing to maintain it to enable it to last for its maximum possible lifetime [[
8 ]].
Thus, more practical options for low-income countries could include using proven point-of-use (POU) water treatment methods. In rural areas of South Africa, POU water treatment technologies are being widely promoted by the government and other organizations as an appropriate intervention for reducing the burden of waterborne diseases. Ceramic water filters (CWFs) are examples of POU water treatment technologies that are being used by rural communities in South Africa. CWFs are usually produced by firing a mixture of locally available materials, which include suitable clays, burn-out materials (e.g., rice husks and sawdust) and water. Silver nanoparticles (AgNPs) are the major disinfectants that are usually added to CWF to aid microbial inactivation and to prevent the recontamination of treated water and the growth of biofilms on the surface of the filters [[
Various methods have been reported on the mode of silver nanoparticles (AgNPs) in addition to CWF. This includes the painting of colloidal silver on the surface of the filter, dipping of the filters into AgNPs solution and the adding of AgNPs solution as part of the material mixture to make the filter before firing [[
With the advancement of material development, silver nanoparticles can be easily applied to solid materials for the inactivation of microorganisms in contaminated water [[
Regardless of the success that has been recorded in attempts aimed at the provision of potable water using AgNPs CWFs, research is still ongoing to improve on previous concepts and discover more cost-effective methods that could be applicable, particularly in marginalized areas with low economic status.
PureMadi is one organization that runs a filter manufacturing facility in rural Dertig, North West Province of South Africa and Mukondeni, Limpopo Province of South Africa. CWFs are manufactured using local labor and materials (clay, sawdust and water). Silver nanoparticles are painted to these CWFs to act as a disinfectant towards pathogens, and this method of silver application is used in most filter-making facilities globally.
However, the use and painting method for AgNPs application has several disadvantages:
- AgNPs are not locally available in South Africa and other developing world markets and therefore are imported by filter production facilities;
- Nanoparticles may be released from the filter, particularly in the early-stages of filter use, which can potentially result in silver concentrations in the treated water that are greater than the World Health Organization (WHO) and Environmental Protection Agencies of the United States (EPA) secondary drinking water guideline value of 100 ppb based on health effects [[
18 ]]; - Application of the aqueous nanoparticle suspension is labor-intensive, requiring facility workers to manually paint the solution on the surfaces of every filter;
- Using nanoparticles during the manufacturing process may also constitute a health risk for workers manufacturing the filters, as some research suggests that inhalation of silver particles may result in genotoxic effects [[
20 ]].
The use of silver nitrate in CWF can reduce the risk of inhalation exposure by workers manufacturing CWFs [[
The aim of this study is to evaluate the comparative microbiological effectiveness of silver-impregnated CWFs made from AgNO
The study area is located in Dertig, Bojanala District, North West Province of South Africa (Figure 1). Dertig is governed by the Moretele Municipality, and its geographical coordinates are 25°16′45″ South, 28°13′21″ East [[
A baseline census of the Dertig area was conducted in its two wards (14 and 22) to identify households in which there was at least water interruption that lasted for more than two days and practiced water storage. Thirty households (15 from each ward) were randomly selected based on the baseline census and willingness to participate in the study. Baseline data were collected from each household, including the assessment of water quality. A sample of drinking water was taken every month during the sampling period, which lasted for 13 months. Ten of the households with five or more residents were given two filters (one with AgNPs and the other with either 1 g or 2 g of Ag prepared from AgNO
The main water source in the Dertig area is a piped water system from Magalies Water Treatment Works. Due to water rationing, residents in the Dertig area suffer from 3–7 day water interruptions every week. Water rationing has, therefore, forced families to store enough water for usage during times when there is no running water from the taps. Secondary water sources in the Dertig area include groundwater (from boreholes, some with protected hand pumps), rainwater harvesting and water from trucks. Tshwane River water is only used for other domestic purposes except drinking.
In this study, three kinds of CWF were used. The first kind of CWFs (
Visual inspections took place before each major step of the production process so that defective filters could be removed from the production line. Formal visual inspections were carried out before surface finishing, loading the kiln, flow rate testing, silver application and packaging. Filters were examined for cracks, warping, inconsistent filter walls, large pieces of burn-out material and consistent surface finish. In fired filters, filters were examined for discoloration, including blackened areas indicating incomplete combustion of the burn-out; warping; cracks; holes or spaces from large pieces of burn-out material; charring; crumbling; and that the base and rim of the filter was at the proper angle to the wall of the filter. The filter rim of fired filters was checked for size and warping by placing a receptacle lid on each filter element. The lid was turned slowly, and it was checked that the filter rim meets the lid evenly. If the lid did not fully cover the filter rim, the filter element was ground using caution not to damage the body of the filter or grind any more material than necessary [[
Health and safety measures were ensured to reduce any risk associated with filter-making, such as inhalation of dust particles during sieving of the sawdust and throughout the production of the CWFs. Personal protecting equipment such as gloves, goggles and dust masks were provided and used at all times.
Drinking water samples, regardless of the source, were collected monthly from the 30 households into sterile containers. The physicochemical and microbiological water quality parameters (total coliform and E. coli) from the source water were measured monthly for 13 months. Similarly, the filtered water was also collected using the spigot in the receptacle of the bucket containing the CWF into a sterile container, and the levels of total coliform and E. coli were also enumerated. This was done monthly for 13 months for the different kinds of filters. Distilled water from the Hydrology and Water Resources laboratory of the University of Venda was used as a control sample for physicochemical and microbiological analyses.
Physicochemical parameters of source water samples were measured in the field by a YSI Professional Plus meter (YSI Inc., Yellow Springs, OH, USA) for pH and conductivity. The probes and meter were calibrated according to the manufacturer's instructions. Turbidity was measured in the field with an Orbeco-Hellige portable turbidimeter (Orbeco-Hellige, Sarasota, FL, USA). The turbidimeter was calibrated according to the manufacturer's instructions. Measured levels were compared to the South African National water-quality standards.
Analysis of the samples was carried out within 24 h of sample collection. Water samples from ceramic filters made with silver nitrate and silver nanoparticles were evaluated for total coliforms and E. coli. Systematic measurement and observation of the two microbial parameters were carried out at the PureMadi Ceramic Filter Facility. The membrane filtration technique was used to detect the presence of microorganisms in water. As a disinfection measure, manifold sample cups were placed in a boiling water bath set to 100 °C for 15 min. Filter paper disks of 47 mm diameter and 0.45 micropore size (4.5 × 10
Graphite furnace atomic absorption spectrophotometry (AA2100; PerkinElmer, Waltham, MA, USA) (GFAA) was used for the quantification of silver in the filtered water. Ten millimeter samples from the filtered water were collected from each participating household monthly and stored in a refrigerator. The samples were analyzed at the Department of Engineering Systems and Environment at the University of Virginia. Before analysis, the samples were prepared with nitric acid (1%) to reduce the chelation of ions [[
An analysis of the economics involved in the production was carried out to assess the economic benefit of using silver nitrate instead of silver nanoparticles in filter making. Since the labor cost, cost of clay and other production materials is the same for all kinds of filters, only the cost of silver nitrate and AgNPs were used to compute this. In addition, the shipping cost was included in the cost of the AgNPs. The unit cost analysis was employed to ascertain the cost involved in producing 1000 units of each kind of filter.
First, ethical authorization was obtained from the ethics committee of the University of Venda. Consent to carry out the study was then requested from PureMadi, the implementers of the Ceramic water filter technology in Dertig, South Africa.
Prior to sample collection, permission was requested from the Moretele Municipality and Dertig community leaders. Consent was then requested from the volunteering households where they were also informed about the purpose of the research, details of their participation, how collected data will be used as well as the benefits of the study.
The research involved using different laboratory chemicals in microbial water analysis. Hence there was safe handling and safe disposal of cultures, reagents, and materials and while operating sterilization equipment to protect the health of the individuals working at the filter facility as well as safeguarding the environment at large.
Excel version 26 was used to analyze the samples statistically using One-way Analysis of Variance (ANOVA). Delta graph was used for some of the plots.
Thirty households enrolled in the study and responded to a survey data questionnaire. The highest range of people per household was between 4 to 6 (n = 19, 63%) as shown in Table 1.
Adult women are most often responsible for water management (n = 23, 77%) at home, while adult men are least often responsible for managing water (n = 7, 23%). One hundred percent of households have their primary water source piped to their yards, with all households reporting the origins of their water to be a municipal treated source (n = 30). One hundred percent of the households suffer 3–7-days water interruptions weekly. Because of prevailing water supply interruptions in the Dertig area, most enrolled households store their drinking water in plastic buckets (n = 24, 81%), while a few households store their water in plastic bottles (n = 6, 19%).
Most households (n = 24, 80%) fill their storage containers directly from the tap, while only a fifth of the households (n = 6, 20%) use hosepipes to fill their storage containers. Collection of water from storage containers is through cups with handle (n = 30, 100%). All households cover their stored water with a lid. When the stored water is used up, most households (n = 17, 57%) have their secondary water source from tanker trucks (delivered at a central community point every 2 days), while some (n = 7, 23%) households use rain harvested water from their JoJo tanks. Only a few households (n = 5, 17%) have their secondary water source from nearby boreholes.
Forty-seven percent of the respondents describe their drinking water quality as average as it is sometimes cloudy and smells bad. Some households describe their drinking water as poor (n = 12, 39%), while a few households describe their drinking water as very good quality (n = 4, 14%) (Figure 2).
Raw water in this context refers to the water obtained from the homes of the respondents, which could be municipal stored water and water from rainwater harvesting and boreholes.
Physicochemical tests carried out on all raw water samples at the beginning of the study included conductivity, turbidity, total dissolved solids, color and pH. The test results showed that raw water from households in the Dertig area had conductivity, total dissolved solids and pH within the recommended South Africa National Standards (SANS) for drinking water quality limits (Table 2) [[
In this study, high color levels in raw water could be due to the frequent water interruptions in the Dertig area. A policy position statement issued by The Chartered Institution of Water and Environmental Management (CIWEM) [[
On the other hand, WHO [[
POU water treatment technologies such as CWFs may be affected by high turbidity levels in the source of water, and this may strongly affect their lifespan and effectiveness in water purification [[
A summary of the mean of total coliform and E. coli of the raw water (before filtration) from different water sources in the Dertig area as a function of time is presented in Figure 3 and Figure 4. Some household raw water samples tested positive for both total coliform and E. coli. The majority of the households, however, tested negative for E. coli. The average levels of E. coli recorded for some sampling months was high due to the value recorded in a few of the households. Generally, low levels of E. coli were determined in the household source water. The presence of both indicator organisms in some of the drinking water implies that the water is not safe for drinking and could possibly put the consumers at risk of waterborne diseases hence the need for a point-of-use water treatment system.
At the beginning of the study, 100% of households reported their primary source of drinking water is municipally treated water. However, during the period of rationing, they either use their stored water or get water from secondary sources, which are often untreated. Therefore, the presence of indicator organisms in some of the household water could be due to inadequate municipal treatment of water, the poor microbiological water quality of the secondary source of water and possible recontamination of treated water during storage and use.
One-way ANOVA was used to ascertain if the microbial level of the raw water from the households that receive CWF of 1 g and 2 g were statistically different. The results obtained showed that the households that receive both kinds of filters vary significantly with the levels of total coliforms (p < 0.05), but the levels of E. coli in the household samples did not vary significantly (p > 0.05).
Results show that CWFs made with silver nitrate recorded a high removal efficiency for total coliforms (95%). This implies that incorporating silver nitrate before firing the filters is effective in inactivating total coliform in contaminated water. A previous study carried out by Mwabi et al. [[
Figure 5 gives a summary of the distribution of total coliform in both raw and filtered water. The total coliform median value for raw water decreases from 160 CFU/100 mL to 2 CFU/100 mL in the treated water after filtration. Ceramic water filters are therefore effective in the reduction of total coliforms in drinking water. Although a 0 CFU/100 mL is recommended by SANS drinking water standards, the presence of total coliform does not indicate the health risk of the water to the consumer [[
CWFs made with silver nitrate were effective in reducing total coliform throughout the 13-month period, and their long disinfection capacity could be due to the release of the silver ion into the water. A similar study conducted by Nangmenyi et al. [[
Water samples from the 30 participating households were collected and tested for E. coli prior to and after treatment by CWFs made with silver nitrate. The removal efficiency of E. coli proved to be very high, at 99%. A similar study assessing the effect of activated metallic silver on water quality in a laboratory setup was reported by Meierhofer et al. [[
Proper water handling, hygiene practices and safe storage are essential in the provision of good water quality as they prevent recontamination and offer long-time inactivation of bacteria by silver during storage. Whenever the researchers visited households for sample collection, water was continuously stored in the ceramic filter receptacles. A possible factor contributing to the high removal efficiency of E. coli by the CWFs made with silver nitrate could also be linked to the contact time with silver during storage [[
Out of 30 households participating in the study, 15 households had CWFs with 1 g of silver nitrate added during the manufacturing process, while 15 households had CWFs with 2 g of silver nitrate added during the manufacturing process. Ten of those 30 households were randomly selected and given an extra filter with painted AgNPs. A comparison of the microbiological quality of the three types of CWFs within a 13-month period was carried out.
The effectiveness of both filters in improving microbial water quality could be because the filters were still new and had only been used for 13 months (they have a life span of 3 years). New filters with fresh silver coating have been found to be very effective in water purification. A removal efficiency of 99–100% for E. coli was established in new filters [[
Figure 7 highlights that both methods of silver application (i.e., incorporating silver nitrate before the firing stage and painting-on silver nanoparticles after the firing stage) are effective in total coliform and E. coli removal. Silver ions are highly effective in the disinfection of a wide range of waterborne microorganisms [[
A calculation of the percentage total coliform and E. coli removal by 1 g, 2 g and silver nanoparticles filters over 13 months period was carried out, and results show that the different filters result in similar levels of total coliform and E. coli removal (Table 3). It found that silver nanoparticles filters had a slightly lower removal efficiency of total coliform (72%) compared to 1 g and 2 g silver nitrate filters.
The results are in line with Jackson et al. [[
Therefore, it can be concluded that CWFs made with both 1 g and 2 g silver nitrate and silver nanoparticles had comparable efficiency for bacteria inactivation. It is also noted that there is no significant improvement in performance for filters made with 2 g silver nitrate relative to filters made with 1 g silver nitrate. The performance of the three kinds of filters was also tested using One-way ANOVA, and the bacteria removal efficiency of both total coliform and E. coli did not vary significantly (p > 0.05), implying that the filters perform comparably though the 1 g Ag CWF recorded a marginal higher bacteria inactivation.
This study has proved that applying both silver nitrate and AgNPs to CWFs improves microbiological efficacy in household water treatment. The application of silver to CWFs also prevents stored water from recontamination. Lyon-Marion et al. [[
Average effluent silver levels in this study were 0.07 ± 0.04 µg/L (1 g Ag
It was observed that silver levels in all filters decrease throughout the study. A possible explanation for this could be the frequency of use of CWFs by households. A study on silver nitrate filters by Kendarto et al. [[
For silver nanoparticle filters, the nanoparticles may be transported out of the filter, resulting in nanoparticles in the drinking water. Ren and Smith [[
Figure 8 shows that silver nitrate filters release extremely low levels of silver ions, and it has been reported in previous studies that ionic silver is not genotoxic at different concentrations [[
A similar study by Jackson et al. [[
Brown et al. [[
Ryner et al. [[
All the households in this study reported that their main source of drinking water is municipally treated water, which is often disinfected using chlorine. The storage of the water for days creates the possibility of recontamination. A study conducted by Lyon-Marion et al. [[
In conclusion, since the 1 g, ionic silver from silver nitrate filters release extremely low silver levels compared to 2 g and silver nanoparticles filters during water treatment, and both attain similar levels of bacteria inactivation; hence 1 g ionic silver filters are the best option to adopt. CWFs made with silver nitrate do not only make water safe for human consumption (microbiologically and in terms of silver release) but also reduce the risks of occupational exposure (to silver) to workers involved in filter making. Silver nitrate is mixed with clay, sawdust, grog and water during the manufacturing process, while AgNPs are painted to the filters after firing. Using silver nitrate, therefore, eliminates the possibility of inhalation exposure to the filter manufacturing workers.
A cost analysis of the economic benefit of substituting AgNPs with silver nitrate in the production of CWFs was carried out. Overall, the silver nitrate chemical costs less than silver nanoparticles in terms of the purchasing price per kilogram and considering that silver nitrate is locally available in South Africa, no shipping costs are incurred (2019 pricing: 1 South African Rand approximately equals 0.066 United States Dollars). AgNPs were purchased and shipped from Spain, implying that there was additional shipping cost associated with the purchase of AgNPs, as shown in Table 4.
Besides the above-mentioned purchasing costs, using silver nitrate eliminates one stage labor costs as the painting step is removed because silver nitrate is added during the manufacturing process. Ryner et al. [[
However, one drawback, as noted by Jackson et al. [[
During the manufacturing of different types of filters under study (1 g Ag
Silver nitrate impregnated CWF slightly performs slightly better in removing microorganisms from drinking water compared to the conventional AgNPs CWF. Therefore, silver nitrate impregnated CWFs can be adopted in the provision of safe drinking water at the household level. Both silver nitrate and silver nanoparticles CWF release silver concentrations that are below the recommended drinking water guideline (100 ppb) for silver levels. In terms of silver released, the 1 g ionic silver filter recorded the lowest levels with excellent bacteria inactivation.
This method reduces the risk of inhalation of the chemical by workers, thereby improving their occupational health and safety. One gram ionic silver from silver nitrate impregnated CWFs is a viable option to adopt because they are cheaper to produce. Silver nitrate can be purchased locally in South Africa; hence there are no importing costs associated with its use. It is therefore economical to substitute AgNPs with silver nitrate in the production of CWFs.
In summary, this method of silver application could potentially improve performance, reduce production costs, and increase the safety of production for workers as well as consumers drinking ceramic filtered water.
Graph: Figure 1 Study area: South Africa.
Graph: Figure 2 Baseline perception of water quality by households.
Graph: Figure 3 Average values of total coliform of raw water (n = 390).
Graph: Figure 4 Average values of E. coli of raw water (n = 390).
Graph: Figure 5 Box and whisker plot highlighting differences of inflow (n = 390) and outflow (n = 390) total coliform count for the silver nitrate filters.
Graph: Figure 6 Box and whisker plot highlighting differences in inflow (n = 390) and outflow (n = 390) of E. coli count. for the silver nitrate filters.
Graph: Figure 7 Total coliform (left) and E. coli (right) inactivation using the various filter types (n = 1040).
Graph: Figure 8 Results of experiments showing silver concentration in effluent over a 13 month period (n = 42).
Graph: Figure 9 Costs associated with producing 1000 filters.
Table 1 Number of people per household.
Range of Number of People per Household Number of Households Percentage 1–3 7 23% 4–6 19 63% 7–9 4 14%
Table 2 Results showing average values of physicochemical determinants of raw water (n = 780) used in filter performance tests with reference to SANS for drinking water quality.
Water Quality Parameter Average Raw Water Concentration St. Dev for Raw Water Concentration Risk SANS Drinking Water Standards [ Conductivity 120 mS/m 21.89 Esthetic ≤170 Total dissolved solids 1150 mg/L 67.72 Esthetic ≤1200 Color 16 mg/L as Pt-Co 1.145 Esthetic ≤15 Turbidity 2 NTU 0.695 Operational ≤1 and esthetic pH 8 1.083 Operational ≥5 and ≤9.7
Table 3 Percentage coliform removal for total coliform and E. coli by 3 filter types.
Filter Type Total Coliform Removal 1 g silver nitrate 96% 99% 2 g silver nitrate 89% 100% AgNPs 72% 99%
Table 4 Summary of costs related to the purchase of silver nitrate and AgNPs.
Silver Nitrate AgNPs R9,085 R28,062 R0 R4,502 R9,085 R32,564
Conceived and designed the experiments: N.N., J.N.E., J.A.S. and J.O.O.; performed the experiments: N.N.; contributed reagents/materials/analysis tools: N.N., J.N.E. and J.A.S.; analyzed the data: N.N. and J.N.E.; wrote the paper: N.N. and J.N.E.; participated in the editing of the manuscript: J.N.E., J.A.S. and J.O.O. All authors have read and agreed to the published version of the manuscript.
This project was partially funded by the University of Virginia's Jefferson Public Fellows (JPC) program. The content is solely the responsibility of the authors and does not represent the official views of the funders.
The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of the University of Venda (protocol code: SES/18/HWR/07/2905 and date of approval: 29/05/2018).
Informed consent was obtained from all subjects involved in the study.
The main data obtained in this work is contained in the manuscript. Others can be made available upon request from the corresponding author.
The authors declare no conflict of interest.
The authors thank workers from the PureMadi Dertig Filter Facility for their assistance in making filters, as well as A. Estrella from the University of Virginia, who did much of the analysis for silver levels in the effluent. The authors also acknowledge the work of T. Beddow, E. Taylor-Fishwick and M. Sutton from UVA, who were involved in survey data collection, sample collection and microbiological analysis at the beginning of the study.
By Nkosinobubelo Ndebele; Joshua N. Edokpayi; John O. Odiyo and James A. Smith
Reported by Author; Author; Author; Author