Previous Page Table of Contents Next Page

2.5 Filtration systems

Filtration systems are primarily used to purify water for domestic consumption. Several types of filtration systems have been used extensively in developing countries throughout the world, particularly in Latin America and the Caribbean. These include residential filters, slow and rapid sand filters, and dual media filters. Vertical flow pre-filters with gravel media tested in Guatemala and up-flow solids contact filters used in Brazil have potential for future use.

The design and application of different types of filters depend on the volume, flow rate, and quality of the inflowing water; the desired degree of water purification; and the use of the filtered water. The availabilities of filtering materials and skilled personnel are also factors to be considered in the selection of an appropriate filtration system.

Normally, the quality of the product water can be improved by mechanical straining through a porous material, such as sand or gravel. Depending on the size of the pores and the nature of the filter material, straining, or filtering, may remove a significant portion of the undesirable contents of the feedwater: suspended and colloidal matter, bacteria and other microorganisms, and, sometimes, certain chemicals. The filter material may be any porous, chemically stable material, but sand (silica and garnet) is used most often. Sand is cheap, inert, durable, and widely available. It has been extensively tested and has been found to give excellent results. (Other materials have been used, some of which are described below; others, such as the reverse osmosis technologies described previously, are also a specialized form of filtration.)

Technical Description

· Residential Filters

Residential filters are a common form of filtration. They can be either homemade or purchased commercially. The homemade filters usually consist of a sand- or gravel-filled pipe or tub, while the commercial systems usually have a stainless steel frame, with appropriate connections that make installation and operation relatively simple. Many commercial filters contain filtration media other than sand or gravel.

The basic form of residential filter, used in rural areas with no public water supply, is the tub filter. The tub filter consists of two tubs made of mud or clay, pottery or plastic, and joined together. The upper tub contains the filter medium (sand, gravel, coal, stone, etc.), into which the water to be treated is poured. It moves through the filter medium, through holes in the base of the upper tub, to the lower tub, where it is stored until used. A faucet is usually installed in the lower tub for convenient access. Homemade filters, such as the tub filter, are usually constructed of locally available materials. For example, in El Salvador, they are constructed of a concrete pipe, approximately 0.5 m in diameter and 1 m in length, fitted with a perforated pipe, which is placed at the bottom of the filter in a 10 cm layer of gravel and connected to a pipe with a 3/4-inch internal diameter from which the filtered water is extracted. The gravel is overlain by 60 cm of sand. Both the gravel and the sand are cleaned and dried in the sun, before use. In Mexico, residential filters are constructed of porous volcanic rock assembled in a wooden frame and protected by a screen. In the Dominican Republic, residential filters are installed at the point of discharge of storage cisterns, or at the point where water enters the houses. The frame of these filters is usually made of stainless steel, with layers of sand, quartzitic gravel, anthracite, and activated carbon as the filtration media.

· Slow Sand Filters

A slow sand filter consists of a watertight box, fitted with an underdrain, which supports the filtering material and distributes the flow evenly through the filter. Many different media have been used for the underdrain system. Bricks, stone, and even bamboo have been used for this purpose; bamboo, however, requires frequent replacement because it is organic and subject to decomposition. The effective size of the sand used in slow sand filters is about 0.2 mm, and may range between 0.15 mm and 0.35 mm, with a coefficient of uniformity of between 1.5 and 3.0. In a mature bed, a layer of algae, plankton, and bacteria forms on the surface of the sand. The walls of the filter can be made of concrete or stone. Sloping walls, dug into the earth and supported or protected by chicken wire reinforcement and a sand or sand-bitumen coating, could be a cost-effective alternative to concrete. Some Latin American countries, such as Ecuador and El Salvador, use concrete reinforced with a minimal amount of iron (ferrocement). Inlets and outlets should be provided with controllers to keep the raw water level and the filtration rate constant. Lateral pipes range from 2 to 8 in, while the bottom drains are normally between 10 and 30 in. Bottom drains consist of a system of manifold and lateral pipes. Figure 26 is a diagram of a typical slow sand filter.

The successful performance of a slow sand filter depends mainly on the retention of inorganic suspended matter by the straining action of the sand. Filtration rates usually employed in developing countries range between 2.5 and 6.0 m3/m2/day. Higher rates may be used after a series of tests demonstrates that the effluents are of good quality. The system should be designed for flexibility, and should consist of a number of separate units to enable maintenance to be performed without interruption of the water service. The suggested number of units for a given population size ranges from two units for a population of 2 000 up to six units for a population of 200 000.

· Rapid Sand Filters

Rapid sand filters differ from slow sand filters in the size of the media employed. Media in rapid sand filters may range in size from 0.35 to 1.0 mm, with a coefficient of uniformity of 1.2 to 1.7. A typical size might be 0.5 mm, with an effective size of 1.3 to 1.7 mm. This range of media size has demonstrated the ability to handle turbidities in the range of 5 to 10 NTU at rates of up to 4.88 m3/m2/h. Filtration rates for rapid filters may be as high as 100 to 300 m3/m2/day, or about 50 times the rate of a slow sand filter. The number of filters used for a specific plant ranges from 3 filters for a plant capacity of 50 l/s to 10 filters for a plant capacity of 1 500 l/s.

A typical rapid sand filter consists of an open watertight basin containing a layer of sand 60 to 80 cm thick, supported on a layer of gravel. The gravel, in turn, is supported by an underdrain system. In contrast to a slow sand filter, the sand is graded in a rapid rate filter configuration. The sand is regraded each time the filter is backwashed, with the finest sand at the top of the bed. The underdrain system, in addition to performing the same functions served in the slow rate filter, serves to distribute the backwash water uniformly to the bed. The underdrain system may be made of perforated pipes, a pipe and strainer, vitrified tile blocks with orifices, porous plates, etc. A clear well is usually located beneath the filters (or in a separate structure), to provide consistent output quantity. The minimum number of filter units in a system is two. The surface area of a unit is normally less than 150 m2. The ratio of length to width is 1.25 to 1.35.

· Dual- or Multi-Media Filters

Dual-media filtration uses two layers, a top one of anthracite and a bottom one of sand, to remove the residual biological floe contained in settled, secondary-treated wastewater effluents and residual chemical-biological floe after alum, iron, or lime precipitation in potable water treatment plants. It is also used for tertiary or independent physical-chemical waste treatment in the United States and other countries. Gravity filters operate by using either the available head from the previous treatment unit or the head developed by pumping the feedwater to a flow cell above the filter cells. A filter unit consists of an open watertight basin; filter media; structures to support the media; distribution and collection devices for influent, effluent, and backwash water flows; supplemental cleaning devices; and the necessary controls to sequence water flows, levels, and backwashing.

Figure 26: Slow Sand Filtration System.

Source: Edward J. Martin, Handbook for Appropriate Water and Wastewater Technology for Latin America and the Caribbean, Washington, D.C., PAHO and IDB, 1988.

· Upflow Solids Contact Filter

These units eliminate the need for separate flocculators and settling tanks, since they perform liquid-solid separation, filtration, and sludge removal in a single unit process. Coagulation and flocculation are performed in a granular medium (such as a layer of gravel under a sand bed). The use of flocculent aids improves filtration results. This process should be restricted to raw waters of low turbidity (up to 50 JTU) and no more than 150 mg/l of suspended solids. It is widely used, especially in Brazil. These filters are designed for rates of filtration between 120 and 150 m3/m2/day.

Extent of Use

Both homemade and commercially purchased residential filters are commonly used in developing countries where the quality of water for domestic use is poor. El Salvador, Dominican Republic, and Mexico have promoted the use of these types of filters. In general, most Latin American countries use residential filters for water purification, particularly in rural areas.

Slow and rapid sand filters have been used in the rural community of La Pinera, El Salvador. In Ecuador, slow sand filters are used extensively for both surface and groundwaters. Filtration systems using vertical reactors with gravel beds have been tested as a means of pre-filtration in a water treatment plant in the municipalities of Cabañas and Zacapa, Guatemala. Rapid sand filters are more complex to operate than then-slow sand filter counterparts, but they are widely used, especially in areas with high turbidity and where land requirements may be an important design consideration. Conventional rapid sand filtration plants are widely available and widely used in Latin America and other developing countries throughout the world.

Dual or multimedia filters are limited to developing countries that can inexpensively acquire anthracite. The higher skill level and energy requirements for the operation of these high rate systems may limit their application.

Upflow solids contact filters, because of their simplicity and low cost, could be an effective technology in many developing countries. Brazil has successfully used this type of filtration system.

Operation and Maintenance

The filter media of homemade residential filters must be periodically changed to maintain the filter's effectiveness. Most of the residential filters acquired commercially can be purchased with a maintenance contract, which will prolong their operational life.

A number of factors affect the operation and maintenance of slow sand filters. The initial resistance (loss of head) of a clean filter bed is about 6 cm. During filtration, impurities are deposited in and on the surface layer of the sand bed, and the loss of head increases. At a predetermined limit (the head loss is usually not allowed to exceed the depth of water over the sand, or about 1 m to 1.5 m), the filter is taken out of service and cleaned. The period between cleaning is typically 20 to 60 days. The filter can be cleaned by either scraping off the surface layer of sand and replacing it with washed sand stored after previous cleanings (periodic re-sanding of the bed), or washing the sand in place with a washer that travels over the sand bed. If sand is readily available, the former method is favored; workers with wide, flat shovels do the scraping, removing 1 to 2 cm of the topmost material. The amount of time this takes depends on the area of the filter bed, but it can usually be completed in one or two days. After washing, the sand is stored and replaced on the bed when, after successive cleanings, the thickness of the sand bed has been reduced to about 50 to 80 cm. A sand and gravel filter needs to be replaced every two years or so. When using the method of washing in place, about 0.2% to 0.6% of the water filtered is required for washing purposes. The bacteriological layer, which is the most important layer in the filtration process, needs to be reactivated in the new filter. Reactivation usually lasts two months.

Rapid sand filtration plants are complicated to operate, requiring operator training in order for the plant to produce a product water of consistent quality and quantity. The filters require frequent backwashing to maintain satisfactory operating heads in the system (filter runs may vary from only a few hours to as many as 24 to 72 hours, depending on the suspended solids in the influent). Backwashing rates are typically 0.6 m3/min or higher, for a period of several minutes. In addition, the initial production following backwashing is channeled to waste for several minutes. Thus, the water backwashing uses can be as much as 10% to 15% of the total plant output. On the other hand, rapid sand filtration plants (including chemical treatment) can effectively treat higher solids loadings and produce higher outputs than slow sand filters. The land area requirements are significantly lower.

Dual-media filters, like rapid sand filters, are cleaned by hydraulic backwashing (upflow) with potable water. Thorough cleaning of the bed makes it advisable in the case of single medium filters, and mandatory in the case of dual- or mixed-media filters, to use auxiliary scour or so-called surface wash devices before or during the backwash cycle. In dual-media and mixed-media beds, such additional effort is needed to remove accumulated floe, which is stored throughout the bed depth to within a few inches of the bottom of the fine media. Backwashing is generally carried out every 24 to 72 hours. The optimum rate of washwater application is a direct function of water temperature, as expansion of the bed varies inversely with the viscosity of the washwater. For example, a backwash rate of 18 gpm/ft2 at 20°C equates to 15.7 gpm/ft2 at 5°C, and to 20 gpm/ft2 at 35°C. The time required for backwashing varies from 3 to 15 minutes. After the washing process, water should be discharged to waste until the turbidity drops to an acceptable value. Few data are available on the operation and maintenance of the vertical reactor pre-filters tested in Guatemala, which remain in the experimental stage.

Other operational considerations relating to the use of filtration technologies include the use of flocculent aids. Coagulants such as alum, ferrous sulfate, and lime may be added to aid in the flocculation and sedimentation of particulates. The coagulant dosage is generally determined from jar tests, and the chemicals are almost always added with rapid mix systems. In the case of water treatment plants, flocculation is usually performed ahead of the settling process to improve the effectiveness of this process.

Maintenance considerations include the resolution of a number of problems which can interfere with the consistent operation of sand filters. These problems often are due to poor design or operation of the filtration systems. The problems most often encountered and their possible solutions are as follows:

· Surface clogging and cracking: This problem, caused by an overload of solids at the thin filter layer in sand filters, can be alleviated by using dual or multiple media, which allows deeper penetration of solids into the bed, and, generally, longer run times.

· Gravel displacement or mounding: This problem can be alleviated by placing a 76 mm layer of coarse garnet between the gravel supporting the media and the fine bed material.

· "Mudball" formation: This problem can be reduced by increasing the backwash flow rate (e.g., up to 20 gpm/ft2), and by providing for auxiliary water or air scouring of the washed surface.

· Sand leakage: This problem may be alleviated by adding the garnet layer.

· Accumulation of air bubbles in the bed: This problem, which causes a significantly increased resistance to flow through the filter, can be minimized by maintaining adequate water depths in the clear well and filters; frequent backwashing may help.

Level of Involvement

In many developing countries, filtration methods are introduced and promoted by both governmental agencies and NGOs, with the full participation of the community. This is the case in El Salvador, where the Centro Salvadoreño de Tecnología Apropiada (CESTA) builds and installs residential filters for rural communities at a minimum cost. In Dominican Republic, the private sector, particularly the companies which manufacture filtration systems, promotes the technology. In Ecuador, NGOs like Plan Internacional and CARE actively participate in the implementation of these technologies in order to reduce the use of contaminated water. In Brazil, the government and the private sector are actively involved in the development and implementation of filtration systems.


Homemade residential filters were constructed in El Salvador at a cost of $23. Operation and maintenance costs are about $6/year. The cost of residential filters manufactured and commercially distributed in Dominican Republic varies with the flow capacity of the filter. It ranges from $382 for 1 gpm to $588 for 6 gpm; this price includes installation and maintenance. Commercially manufactured tub filters are sold in Dominican Republic hardware stores at a price ranging from $26 to $45. The cost of quarry filters used in Mexico was $50, with little or no operation and maintenance cost. Figure 27 shows the construction cost of an upflow solids contact filter a function of the filtration area.

The unit filtering cost of slow and rapid sand filters in Ecuador ranges between $0.13/m3 and $0.20/m3. Slow sand filters were constructed in Ecuador at a cost of $132.30 with an estimated operation and maintenance cost of 25% of the construction cost. Table 12 shows estimated per capita costs of construction and of operation and maintenance for slow and rapid sand filters.

Effectiveness of the Technology

Homemade residential filters can adequately reduce the level of contaminants in water, but, because quality control tests are usually not performed on the product water, there is a risk of some contamination remaining after filtration. For example, quarry filters used in Mexico reduce bacteriological contaminants by up to 90%. However, quarry filters must be covered and protected with a screen, and a faucet at the outlet is recommended. This filter needs to be cleaned every 3 to 4 months, depending on the quality of the water treated.

Commercially available residential filters are usually more effective at producing a good quality product water since quality control is performed during the manufacturing process and a level of efficiency is initially guaranteed. Quality product water can be further ensured through the regular inspections performed by technicians from the supplier in the case of systems sold with a service contract.

Slow sand filters are very effective in removing solids and turbidity when the raw water has low turbidity and color (turbidity up to 50 NTU and color up to 30 Pt units). Taste and odor are also improved. However, if the raw water quality is poor, filtration is often less effective. In such situations, roughing filters, or pre-filters, are often used before the feedwater enters the slow sand filters. The slow sand filters are very effective in removing bacteria; in general, their effectiveness in removing bacteriological contaminants ranges between 80% and 99%, depending on the initial level of contaminants and the number and design of the filtration units. In many regions of Ecuador, the effectiveness is close to 100%. In El Salvador, they are estimated to remove 84% and 99% of total and fecal coliform bacteria, respectively. Reductions in the levels of iron, manganese, and nitrate concentrations and turbidity are also observed. Chemicals are typically not used. The flow rates for slow sand filters are many times slower than for rapid sand and roughing filters, and the operating filter bed is not stratified.

Multimedia filters are usually more effective, since the filtration media combine the filtration properties of several materials. In the system of vertical flow pre-filters used in Guatemala, turbidity reduction ranged from 23% to 45%, and color reduction between 34% and 56%.


Filtration technologies are suitable for use throughout the region. Homemade residential filters are better suited to rural areas where the equipment, skills, and infrastructure necessary to provide piped domestic water supplies are lacking. The other, more complex filtration systems are best used at water treatment plants and are generally located in urban areas.


· Filtration systems have a low construction cost, especially when built using manual labor.

· These systems are simple to design, install, operate, and maintain, which makes them ideal for use in areas where skilled personnel are few.

· No chemicals are required, although flocculent aids are sometimes used in conjunction with large-scale filtration systems; supplies of sand can usually be found locally.

· Power is not required.

· Large quantities of washwater are not required.

· Use of filtration to pretreat water and wastewaters results in fewer sludge disposal problems because fewer contaminants are left to be removed during the treatment process.

· Residential filters provide adequate treatment of water for average-sized households, particularly in rural areas.

· Filters are environmentally friendly.


· In some areas, there is a lack of locally available filtration media.

· There may be a lack of skilled personnel to operate the more sophisticated filtration systems, particularly in rural areas.

· Use of filtration alone is recommended only for source waters with low levels of contamination.

· Pretreatment may be required for many applications.

· Provision must be made for washing and storing used sand from sand filters, either permanently or temporarily, and for moving sand from the filters to the wash site and from the wash site to the storage site and back, as needed.

· If sand from slow sand filters is to be washed, a separate backwash facility and washwater supply may be required; treated water must often be used for washing, which could reduce the available supply of treated water, especially in water-poor areas.

· Precise operational control of the rate of head loss is required to prevent air bubbles from entering and binding the system; this type of interference is a potential problem in all types of filters.

· There may be a lack of quality control of the product water in rural areas.

· To obtain good results from slow sand filters, the raw feedwater must not generally have a suspended solids content of less than 50 mg/1.

Cultural Acceptability

Filtration is a well-accepted technology when applied in the treatment of industrial and public water supplies. It has limited acceptance in other applications, and at the household level in rural areas.

Further Development of the Technology

Additional research is needed to develop more efficient filtration media that can remove both bacteriological and chemical contaminants. Education is needed, particularly in the rural areas, to encourage the use of homemade filtration systems and disinfection of household water supplies.

Figure 27: Construction Cost ofUpflow Solids Contact Filter.

Source: Edward J. Martin. Handbook for Appropriate Water and Wastewater Technology for Latin America and the Caribbean, Washington, D.C., PAHO and IDB, 1988.

Table 12 Per Capita Costs of Construction, and of Operation and Maintenance for Slow Sand Filters and Rapid Sand Filters (S)

Population Scale


Cost Range

Slow Sand Filter

Rapid Sand Filter

500 - 2499


17.08 - 27.00

12.84 - 15.12

Operation and Maintenance

1.80 - 6.75

2.43 - 5.40

2500 - 14999


12.19 - 19.28

10.08 - 11.88

Operation and Maintenance

0.81 - 3.04

1.22 - 2.70

15000 - 49999


8.55 - 13.5

5.73 - 6.75

Operation and Maintenance

0.45 - 1.69

5.72 - 2.36

50000 - 100000


5.33 - 8.44

3.04 - 3.58

Operation and Maintenance

0.27 - 1.01

0.91 - 2.03

Source: G. Reid and K. Coffey. Appropriate Methods of Treating Water and Wastewater in Developing Countries, Stil-water, Oklahoma, University of Oklahoma, Bureau of Water and Environmental Resources Research, 1978.

Information Sources


Omar Fonseca Moreno, Instituto Mexicano de Tecnología del Agua (IMTA), Subordinación de Comunicación Rural, Paseo Cuauhnáhuac No. 8532, Progreso, Jiutepec, Morelos 62550, México. Tel. (52-73)19-3544,19-3567 and 19-4000 ext. 355. Fax (52-73)19-4341.

Carlos Cisneros E. and Osvaldo Encalada, Instituto de Investigaciones de Ciencias Tecnicas (IICT), Plan Internacional/Care International, Cuenca, Ecuador. Tel. (593-7)840-073. Fax (593-7)840-183.

Yolanda López, Centro Salvadoreño de Tecnología Apropriada (CESTA), Dirección 17 calle Oriente N° 285, Colonia Santa Eugenia, Barrio San Miguelito, San Salvador, El Salvador. Tel. (503)220-0046.

Luis Ochoa Marroquín, Instituto Nacional de Sismología, Vulcanología, Meteorología e Hidrología (INSIVUMEH), 7 Avenida 13, Guatemala, Guatemala. Tel. (502-2)31-4967. Fax (502-2)31-5005.


Bello, J.D., and M. Acosta. 1993. Análisis de la Aceptación de las Empresas Purificadoras de Agua en la Ciudad de Santiago. Santiago, República Dominicana, Pontificia Universidad Católica Madre y Maestra. (Tesis)

Man, H.T., and D. Williamson. 1986. Water Treatment and Sanitation: Simple Methods for Rural Areas. London, Intermedia Technology Publications.

Martin, Edward J. 1988. Handbook for Appropriate Water and Wastewater Technology for Latin America and the Caribbean. Washington, D.C., PAHO and IDB.

----, and E.T. Martin. 1985. Water and Wastewater Cost Analysis Handbook for Latin America and the Caribbean. Washington, D.C., PAHO.

Nuñez, R.D., et al. 1992. Dominican Republic Natural Resource Policy Inventory: vol. II, The Inventory. Santo Domingo, USAID.

Reid, G., and K. Coffey. 1978. Appropriate Methods of Treating Water and Wastewater in Developing Countries. Stilwater, Oklahoma, University of Oklahoma, Bureau of Water and Environmental Resources Research.

Rodríguez, R., F. Saguez, and RE. Yunén. 1993. Boletín Diagnostico General de Los Problemas Ambientales de la Ciudad de Santiago. Santiago, República Dominicana, Pontificia Universidad Católica Madre y Maestra.

Tropical Research and Development, Inc. 1992. Intensive Survey of Rural and Urban Activities Impacting Water and Coastal Resources. Santo Domingo, USAID.

USEPA. 1980. Innovative and Alternative Technology Assessment Manual. Washington, D.C. (Report No. EPA-430/9-78-009)

Previous Page Top of Page Next Page