bacteria purpleSand filters are often used prior to RO and UF membranes, when sand filters are operated properly,  they will give excellent results. The filters also have the advantage that they are low cost and easy to maintain,  however there is always room for improvement as new technology is developed.

Silica sand is a perfect substrate for the growth of bacterium, indeed sand filters make very good biological filters for the treatment of drinking water in the form of slow bed filters.  The filters operate at a water flow in the region of 0.1m/hr as a velocity.  However when the flow velocity is increased to 5 or 10m/hr a bio-dynamics of the microbiology start to change.

Slow bed sand filters are in a state of endogenous respiration, the filter could be considered as a stable ecosystem that is self maintaining. However when the flow rates are increased there will be bio-instablity in the system.  Initially sand filters can give a very good water quality.  Initially, the filter will be colonised by heterotrophic bacteria that use organic carbon as a carbon source. During the growth phase the bacteria improve the performance of the sand filter, this period may last for a few weeks to perhaps 12 months, the time frame is dictated by the temperature of the water and the concentration of dissolved organics.

After a given period the heterotrophic bacteria will have used up all of the available living space in the filter,  the thickness of the biofilm layer now starts to develop and species diversity of bacteria begins to increase.  Around the same time bacterium will be sloughing off the surface of the sand due to the velocity of water passing through the filter.  However the bacteria levels and dissolved organic level in the product water will probably be lower than the influent water.   

Biofilm could be compared to a plant and the planktonic bacteria as their seed.  It is natural for biofilm to discharge bacterium into the water,  indeed the discharge of bacterium into the product water may be synchronised by the bacteria.  This is a perfectly natural process,  so the performance of the filter can be very good,  but it is interspaced by periods when the sand is discharging high concentration of bacteria into the product water.

biofilmThe species diversity of the biofilm is increasing all of the time, and after a few months autotrophic bacteria start to become significant. These bacteria can manufacture organic matter from inorganic carbon.  They will normally appear in the lower reaches of the filter bed, they are much slower growing than the heterotrophs but their impact on water quality is significant. The sand now starts to mineralise,  the bacteria produce an alginate based biofilm matrix,  but they also start laying down a honeycomb mineral structure which gives the biofilm added protection. The shape of the sand grains are changing due to the bio-mineralisation, the production of alginates is increasing and sand is become more unstable. The sand filter has now turned into a factory with the sole purpose to produce as many bacterium as possible.

By this stage nitrogen fixing bacteria start to take up residence in the filter along side the autotrophs near the base of the filter.  As the ecological diversity of bacteria increases, the production of organic mater increases.  The sand filters are now producing more solids and organic matter than are entering the filter,  the problem is that it is coming in sudden discharges that may go un-noticed unless there is continual turbidity monitors on all the filters.

Back-washing the filter bed and using air scour will remove some of the biofilm, but as the sand ages the biofilm becomes more stabilized. If the water temperature is over 25 deg C the doubling time of the heterotrophs can be as short as 15 minutes, so after a few hours  the biofilm is restored. At Dryden Aqua we used to manufacture fluidised be sand biofilters,  if you have a sand filter in continual 50% bed expansion on back-wash mode it actually makes a great biofilter, so we know that irrespective how vigorously a sand filter is back-washed, you will never remove all the bacterium.

Clinoptilolite a zeolitic sand

zeolite 300x204At Dryden Aqua,  Dr.Howard Dryden did a PhD on the natural zeloite clinoptilolite for the removal of ammonium from freshwater systems.  Zeolites do not work well as ion exchange minerals in marine systems due to the competing cations.  However in freshwater Clinoptilolite will absorb ammonium which then acts as a food source for the growth of autotrophic bacteria.  Zeolites therefore tend to make very poor mechanical filtration media in warm freshwater systems.

For marine applications there are fewer issues with biofouling, however the electron microscope photo shows the crystalline nature of clinoptilolite which makes a perfect habitat for the growth and protection of bacteria. Clinoptilolite also contains free silica, so as a pretreatment prior to membranes, a food quality silica sand is actually a better media, and of course AFM is much better than either media.

AFM and Bio-instability in comparison to sand

table bacteria afm

The biofouling of sand can be measured by taking samples of media after a back-wash, shaking in water to remove (using ultrasonics) them and then measuring the bacterial concentrations.  The following was data provided by a laboratory in New Zealand from testing sand and AFM taken from two swimming pool filters in the same facility.

AFM does not biofoul, 100% removed on back-wash

afm back washBacteria will grow in an AFM filter bed,  however they can not stay in the filter.  The back-wash performance is close to 100% efficiency confirmed by IFTS.  download IFTS report The IFTS data was conduced under controlled conditions in a laboratory in accordance to ISO procedures. Sand also gave a very similar result,  however it was new sand not subjected to biofouling. When sand ages the sticky alginate biofilm layer reduces back-wash efficiency as demonstrated by the results from Lyonaise des Eaux.

The data was collected from two parallel filters operating on the same water supply at the same time. The two points of merit to note are;

1. The back-wash profile from AFM is a SIN curve that is repeatable, you can lay one back-wash curve on the next. However the sand filter was a variable. Performance is not predictable and  every back-wash (and run phase) was different reflecting the biological variability or change of bacterium species diversity as well as structure in the filter.

2. The area under the AFM curve is 30% greater, this tells you that 30% more solids were removed by AFM during the run phase in comparison to sand.

Wormhole channelling and bio-instability

It is also interesting to look at the product water quality and to repeat the back-wash profile over a longer time frame.  This work was conducted by a Spanish Water company and the results published in the Spanish Water Journals (download report). The two key graphs from the data are presented below,  the first graph is a sand filter run and back-wash series and the second is for AFM.


sand cycle

afm cycle

The AFM profile  is like clockwork,  the run phase and backwash sequence follow a very precise pattern,  where as the sand filter, operated in parralel and under identical conditions demonstrated a variable performance and channelling.  The differences in the performance of the two filters is down to the quality of AFM and the lack of biofouling.

biofouling graphAFM and sand run phase performance & bacterium discharge

The stability of AFM is reflected in performance,  another example is a report by Arkal (Amiad) in Israel on the treatment of waste water for irrigation.  Tests were conducted, prior to full scale installations, again two filters were operated one with sand and the second with AFM a copy of the report may be down loaded (Revadim).  The data confirms the superiority of AFM over sand filters.

The biofouling sequence has been investigated in more detail  by companies such as Scottish Water,  the following data is from trial conducted by Scottish Water,  Dryden Aqua were just observers and took no part in the tests.  The data has been published by the European Commission and appears in the  UK Journal WWT.  The results are from two identical pressure filters, one with sand and the second with AFM. The parameters presented below include;

  1. Bacterial levels before and after filtration
  2. Indirect measurement of autotrophic bacterium colonisation of the filter beds
  3. The removal of dissolved organic carbon

Bacterial levels before and after filtration

Total aerobic bacterium were counted before and after a sand and an AFM filter at 37degC and 22degC, average trend lines were also added.

Both the sand and the AFM filters tracked the influent bacterium concentrations, however periodically the sand filter would dump bacterium into the water, this behaviour was not observed with the AFM filter.

There was a low concentration of ammonium in the product water, autotrophic bacteria will convert the ammonium to nitrite and then to nitrate.  This can be indirectly observed by a reduction of ammonium concentration and an increase in nitrite.  The two graphs below show this happening for the sand filter, while there was no change with the AFM filter.  The result clearly demonstrations the autotrophic bacterial colonisation of the sand.  While this could be beneficial for sand filtration during the early stages,  it soon results in the production of bacterial cell biomass and DOC (dissolved organic carbon) levels start to increase.



doc productionDissolved Organic Carbon

AFM is a selective molecular sieve adsorber,  dissolved organic matter will be adsorbed onto the surface of AFM and then released during the back-wash.  The graph below confirms that AFM can reduce DOC.  During the first few weeks or months of a sand filter,  the biofilm is very good at reducing DOC by conversion into bacterial cell biomass.  However rapid gravity or pressure sand filters are not in a state of endogenous respiration, so they have to discharge organics back into the product water.  

Once the autotrophic bacteria become established the level of DOC in the product water will often be higher than the influent concentration.


The data presented in this report, is all by independent organisations, water companies and accredited laboratories from across the world.

The data confirms that AFM does not become a biofilter,  this is a really easy parameter to check, simply open up a filter after a back-wash and measure the bacterial levels, they will be close to zero.  The French NGO also wanted to check the performance of AFM,  this time against a glass product called Garo filter media manufactured in France.  The results (download a report)  confirm that Garo crushed glass became a biofilter,  yet AFM remained clear.  All of the data over a 15 year period confirms AFM does not biofoul, nor does it experience worm-hole channelling.  Independent verifiable data confirms AFM to be at least twice the performance of sand and more than three times the performance of crushed glass.

AFM is the perfect filtration media  for drinking water as well as waste water treatment, and when used as a pre-treatment it will out perform membranes and diatomaceous earth to protect RO membranes from particulate matter down to sub-micron levels as well as many dissolved chemicals.

Latest Articles