← Knowledge map
Foundation WaterHealth

Slow Sand Filtration

A barrel or box of graded sand where raw water percolates slowly downward and is purified mainly by a living biological layer on top of the bed, not by the sand alone.

Slow Sand Filtration β€” illustration
Difficultybeginner to build; intermediate to run correctly (patience with the ripening period)
Timehalf a day to a day to build; 2–4 weeks before a new filter reaches full performance
Costlow β€” mostly sand, gravel, and a container

What is it?

A slow sand filter is a container β€” a barrel, box, or masonry tank β€” filled with a deep bed of fine, graded sand, through which raw water trickles downward under gravity at a very slow rate. Unlike a simple sand strainer, most of the purification does not come from the sand grains acting as a sieve. It comes from a thin biologically active layer, the schmutzdecke ("dirty skin"), that grows on top of the sand within the first weeks of operation and consumes organic matter, bacteria, and other pathogens as water passes through it. The sand bed below provides mechanical support and further physical and biological filtration as water percolates through the pore spaces.

What is it good for?

  • Continuous, low-energy treatment of moderately turbid surface water (streams, ponds, rainwater catchment) for drinking or general household use.
  • Village- or household-scale water supply where electricity and chemical dosing are unavailable or unreliable.
  • Substantial reduction of bacteria, protozoa (including Giardia and Cryptosporidium cysts), and turbidity, given a mature, undisturbed biolayer.
  • A stable, passive complement to other steps in a water chain β€” it is not a sterilizer, so it is normally followed by safe storage or a disinfection step (boiling, chlorination, UV) where pathogen risk is high.

The physics behind it

Three overlapping mechanisms do the work: physical straining (sand grains and pore spaces trap suspended particles, especially near the top of the bed); biological action, where the schmutzdecke β€” a mat of algae, bacteria, and protozoa on the top few millimeters of sand β€” metabolizes organic matter and preys on or outcompetes pathogens, acting like a continuously fed biofilm reactor; and adsorption within the bed, where deeper sand grains carry a thin biofilm that keeps trapping fine particles and microorganisms as water percolates through.

Everything depends on keeping the flow slow β€” typically 0.1–0.3 meters per hour of surface loading. At this rate, water stays in contact with the schmutzdecke and biofilm long enough for biological action to happen. Push the flow rate too high and water strains through mechanically, bypassing biological treatment and shortening contact time; it can also physically disturb or wash out the schmutzdecke itself.

History

Slow sand filtration is one of the oldest engineered water-treatment methods still in use. The first documented municipal slow sand filter was built in Scotland in the early 19th century (Paisley, 1804), followed by London's water companies from the 1820s–1850s. It became the standard urban water-treatment method through the second half of the 19th century, credited with major reductions in waterborne disease β€” reinforced by John Snow's 1850s London cholera investigations, which pointed to contaminated water as a disease vector even before germ theory was formalized. Rapid (pressure) sand filtration with chemical coagulation displaced it in most large cities during the 20th century because it treats far more water per unit of land area, but slow sand filtration remains widely used for smaller municipal systems and, in a scaled-down household form (the "biosand filter"), has been promoted since the 1990s as a low-cost household water-treatment technology in rural and developing-world contexts.

Simple version

A single barrel or box, roughly 0.6–1 m deep, layered from bottom to top: a drainage layer of coarse gravel around the outlet, a middle layer of fine gravel, and a top layer of fine sand at least 40–60 cm deep. Raw water is poured in gently at the top, a shallow standing layer of water (5–10 cm) is kept above the sand at all times, and clean water drips out slowly from a pipe near the base. No pump, no power, no chemicals β€” output depends entirely on gravity and the schmutzdecke.

Advanced version

A household biosand filter: a concrete or plastic box with an internal diffuser plate under the inlet (to prevent the incoming stream from cratering the sand surface), a carefully graded sand bed sized to a target flow rate, a "pause period" built into daily use (water is added in batches and allowed to rest, which helps the biolayer's microorganisms starve out pathogens between doses), and an overflow or fixed standing-water height maintained by the outlet pipe geometry so the sand never dries out between uses.

Industrial version

Municipal slow sand filtration plants: large, shallow rectangular basins (often 1,000+ mΒ² per bed) with sand beds roughly 0.6–1.2 m deep over graded gravel and an underdrain network, dosed with pre-settled surface water at controlled loading rates. Multiple beds run in parallel so individual beds can be taken offline for schmutzdecke scraping and resanding without interrupting supply. Some plants add pre-treatment (screening, plain sedimentation) ahead of the sand beds to extend the run time between cleanings, and post-treatment disinfection (chlorination or UV) as a safety barrier after the biological filtration stage.

Building your own

  1. Container: a food-grade plastic drum, cleaned steel drum, or a small brick/concrete box, with an outlet near (but not at) the very bottom, fitted with a pipe that rises back up to just above the sand surface β€” this keeps the sand permanently submerged under a few centimeters of standing water, essential for keeping the schmutzdecke alive.
  2. Drainage layer: 5–10 cm of coarse gravel around the outlet at the base.
  3. Support layer: 5–10 cm of fine gravel above the drainage layer, graded so it doesn't let sand migrate down into the gravel.
  4. Sand bed: 40–60+ cm of washed, graded sand with an effective grain size around 0.15–0.35 mm on top of the support layer β€” the single most important dimension in the whole design.
  5. Diffuser: a flat stone, perforated plate, or a small basin at the inlet so poured water spreads out instead of punching a hole in the sand surface.
  6. Ripening: run the filter with raw water for 2–4 weeks before relying on its output for drinking, discarding or re-treating the output meanwhile. A freshly built filter provides little more than mechanical straining until the schmutzdecke establishes itself.
  7. Flow control: throttle the outlet (a valve, or simply the pipe's diameter and height) so the loading rate stays slow β€” aim for output measured in liters per hour, not liters per minute, on a household-scale unit.

Common mistakes

Mistake Consequence / fix
Running the filter too fast Bypasses biological treatment, shortens contact time, can wash out the schmutzdecke β†’ throttle the outlet, keep loading rate low
Letting the sand bed dry out Kills the schmutzdecke's microorganisms β†’ always keep a standing water layer above the sand
Using dirty or ungraded sand Clogs unevenly or lets fines migrate into the gravel β†’ wash and grade sand before loading
Scraping or disturbing the schmutzdecke unnecessarily Destroys weeks of biological maturation β†’ only scrape the top layer when flow has slowed to a crawl, and expect a new ripening period afterward
Pouring water directly onto the sand surface Gouges a channel, lets water short-circuit through the bed β†’ use a diffuser plate or flat stone
Using the filter's output immediately after building it Little biological treatment has developed yet β†’ wait through the full ripening period
No pre-settling of very turbid raw water Rapid clogging of the schmutzdecke, short run times β†’ let heavily silt-laden water settle first, or add a coarse pre-filter

How to measure

  • Flow rate: time how long it takes to fill a known volume (e.g. a liter) at the outlet; convert to a loading rate (liters per hour per mΒ² of sand surface) and compare against the slow-flow target β€” a rate that is too high is the single most common design fault.
  • Turbidity: compare raw water and filtered output by eye (clarity, settled sediment) or with a simple turbidity tube; filtered water should run visibly clear once the filter has ripened.
  • Ripening progress: track days since construction and watch output clarity and flow-rate stabilization as rough proxies for schmutzdecke maturity β€” 2–4 weeks is the conservative benchmark before trusting the output for drinking.
  • Bed head loss: watch how the standing water above the sand behaves over time β€” a clogging bed (water backing up, flow dropping) signals it's time to scrape the top layer of sand.
  • Water quality testing, where available: basic bacterial indicator tests on the filtered output give the most direct confirmation of performance.

Videos

(TODO)

Downloadable PDF

(TODO)

Sources

  1. General public-health and water-engineering literature on slow sand filtration (municipal treatment history and household biosand-filter design guides)
  2. WHO and rural water-supply technical guidance on household water treatment and safe storage