As aquarists, we are all aware of the need for biological filtration and most of us also understand the principles involved as well. Biological filtration is the term used for the consumption of fish waste by biological means, but as aquarists we usually limit the term biological filtration to the oxidization of ammonia by aerobic bacteria to produce nitrate. The process of oxidization of ammonia is called nitrification. Ammonia is excreted by fish through their gills and in their urine, but it is also produced by heterotrophic (food-eating) bacteria that consume solid fish wastes. Nitrification is a two step process in which ammonia and ammonium (NH3 and NH4+, respectively) are oxidized by bacteria to produce the nitrite ion (N02-), then the nitrite is oxidized by other bacteria to produce nitrate (N03-), thus completing the process of nitrification. It is often stated that Nitrobacter and Nitrosomonas bacteria accomplish these tasks, but it has recently discovered that this is not the case, at least in freshwater aquaria. We don’t know yet actually what bacteria are involved in nitrification, but they are not Nitrobacter and Nitrosomonas. Regardless of which bacteria are involved, it is however known that they require a solid surface on which to grow.
The traditional nitrification filter for home aquaria is the under gravel filter (UGF). These are relatively simple devices in which water is drawn from below a false bottom under a layer of gravel. Water thus flows down through the gravel so nitrification takes place on the surfaces of the gravel grains. However, UGF is not entirely satisfactory in large aquaria or aquaria with high fish densities because the fishes’ solid wastes are trapped in the gravel. This results in the gravel clogging up. Under gravel filters are also generally considered to be detrimental to plants, although a lot of debate still goes on as to how much harm to plants UGF actually does.
A vast improvement over UGF is wet/dry filtration (WDF). In WDF, mechanically filtered tank water is allowed to trickle over emersed (“dry”) filter media. Since the filter media are exposed to the air, the growth of bacteria in these filters is never oxygen-limited. These filters therefore have tremendous biological capacity. Wet/dry filters also serve to oxygenate the water, unlike UGF, which depletes water of oxygen.
The primary difficulty with WDF is that the filters are physically large. They require a volume that is significant fraction of the tank volume in order to handle the ammonia produced. A much more space efficient design is a fluidized bed filter (FBF).
Fluidized bed filters consist of a vertical contact column in which water is pumped upward. Sand placed in the contact column is thus suspended by the upward flush of water. The trick is to supply a sufficiently rapid water flow to suspend the sand, but not so rapid that the sand is blown out of the top of the filter all together. The sand supplies a surface on which the desired bacteria may grow. A FBF is also self-cleaning since the sand is in constant motion and so no detritus can settle on it. This is of limited benefit however, since the detritus is simply flushed out of the filter and ends up back in the tank, where you don’t want it either. Well, at least it doesn’t clog the filter up. But the greatest advantage of FBF is the great surface area provided by the sand. Since sand has a small particle size the surface area available per unit volume of sand is very large. Spherical sand with a grain diameter of one mm (i.e. medium coarse sand) has a surface area of 3141 square meters for every cubic meter of sand. So what? Well it means a couple of handfuls of sand have as much surface area as a typical wet/dry filter.
FBF systems are not without their own problems however. A wet/dry filter makes better use of its relatively limited surface area by growing a thicker coating of bacteria than is possible with FBF. The mutual abrasion between moving sand particles in FBF limits the depth to which bacteria may grow on the sand. FBF also share a disadvantage of UGF in that they deplete the water of oxygen. Fluidized bed filters are also potentially dangerous. When the water pump stops because of power or mechanical failure the sand will settle to the bottom. The packed sand can quickly become anaerobic, resulting in the death of the desired aerobic bacteria. Anaerobic sand may release highly toxic hydrogen sulfide, sometimes within hours of pump failure. However, you can prevent the filter from dying in a power outage if you have a backup generator, or even a battery-operated air pump with which to feed air to the bottom of the sand filter. This will keep the sand moving around and oxygenated enough to prevent problems from developing, at least temporarily.
Fluidized bed filters are now commercially available in sizes appropriate to home aquaria. However, if you are like me, after seeing the commercial FBF units pictured, you would have thought to yourself, “They look pretty simple: why not build one myself?”. Well, the good news is that you can, but the little devils are trickier than you might think. Here are a few things to consider when building your own FBF.
There are many designs that can be used in the construction of an FBF. The simplest design is a simple vertical column. Sand in these cylinders flows in a sort of circulatory motion, up in the center and down along the walls. This circulation quickly breaks down if the water flow gets too great, as sand will then simply blow out the top of the cylinder. Therefore, simple cylinders require careful flow adjustments to provide a suspended sand bed. They are therefore the most difficult to tune, even though they are simplest to build. Another design, which is fairly easy to construct, is the V-shaped trough filter (shown below). A cone design can also be used. And another of the many possible designs is a wide topped cylinder, which is similar in shape to those short ice cream cones with the flat bottoms.
All successful FBF must provide a sufficiently high upward water velocity to fluidize the sand, as well as a region of sufficiently low upward water velocity for the sand to settle back into the filter. A useful rule of thumb is that the water velocity required to fluidize quartz sand in sea water, measured in meters per minute, is equal to the sand grain diameter measured in millimeters. Therefore a filter with sand of grain size one mm requires a water velocity of at least one meter per minute to suspend the particles.
The ability of water to fluidize sand is however also dependent on the viscosity and density of the water, so the rule of thumb has to be modified for water of different temperatures or salinities. The rule is valid for seawater at a temperature of 20C, but the temperature correction is small and can be ignored within the narrow temperature range permissible for a tropical aquarium. The differences in density and viscosity between seawater and fresh water are however more important than those created by temperature changes. The velocities required for fresh water are 13% larger than those needed for seawater. A fresh water filter with a sand grain diameter of one mm therefore requires 1.13 m/min.
So now lets design a FBF for a large tank. This will illustrate some of the design problems that have to be considered when constructing your own filter.
First of all, a design type must be chosen. Let’s reject the simple cylinder design since they are too prone to losing their sand out of the top. The conical and topped-cylinder designs are not easily constructed from commercially available plumbing fixtures, so they are rejected as possible designs also. That leaves the V shaped trough design. They are not nearly as compact as cylinder designs, but (being flat-sided) can be easily constructed with plate glass, acrylic, or even suitably water-proofed plywood.
Lets put the filter behind the aquarium and have it sitting level with the tank bottom. We don’t want to have to drill any holes in the aquarium, so the influent water has to be supplied by a siphon. For simplicity and reliability, a simple overflow pipe will be used to return the water to the aquarium. The height of the filter must therefore be about the same as, put slightly higher than, the aquarium itself to allow the overflow pipe to be positioned over the aquarium top.
Assume that the filter is to be serviced by a water pump with a flow rate of 2000 liters per hour, which is 0.033 cubic meters per minute. The filter is be filled with 1 mm diameter quartz sand. Recall that the water velocity in the contact column must therefore be at least 1.13 meters per minute to suspend the sand in a fresh water, tropical aquarium. So to suspend the sand the trough must have a cross sectional area near the bottom of less than 0.04 m2 (= 0.033 m3/min * 1.13 m/min). Conversely, the top of the trough must have a cross-sectional area considerably larger than 0.04 m2. So for this filter, we could use a V with an open top of 20cm wide by 40cm long (giving an area 0. 08 m2). Pumping 2000 liter per hour, this filter can comfortably service a 500 to 1000 liter aquarium.
We could not fill this filter with sand to any more than the height where the width of the V is 10cm (where the area is 0.04 m2), because filling the V with any more sand than that will prevent the sand from fluidizing. How much sand must we put in? As a general rule, the fluidized sand will have about twice the volume as does the same sand when it has settled down to the bottom, so you should about half fill the filter with sand (or about to where the V is 10cm wide in this case).
One difficult part of building your fluidized bed filter is finding suitable filter sand. Graded filter sand is commercially available for water treatment plants and other purposes, and sandblasting sand is also graded for size and is suitable. However such commercial sources will usually sell their sand in 100-pound bags, or by the cubic yard or cubic meter. Graded filter sand is also expensive, usually costing about $200 per cubic meter, although the cost varies greatly with the uniformity of the sand particles. It is however not necessary to buy graded filter sand as you can grade your own. Quartz river sand is best since it is rounded, but beach sand can also be used if it is relatively coarse and is free of shells. A bug screen in a wooden frame can easily be used to separate sand particles larger than 1 mm from your sand. These large grains must be removed since they won’t fluidize properly. Removing sand particles that are too small is however more difficult, since suitable screens are not readily available. Fortunately, it is only necessary to remove those particles that are too small to remain in the filter under normal operations. Therefore you can grade the sand adequately by simply running the filter and discarding those sand particles that are pumped out of it. A useful method is to fill a bare aquarium with water out on the lawn. The FBF is then attached so that water flows up through it as normal, but then simply flows out the overflow pipe onto the grass, rather than going back into the tank where the fine sand can damage the pump. This process drains the aquarium rapidly, so keep the aquarium full with a continuously running garden hose. Keep it up until the water runs clear. Not only does this water your grass, it adds sand to your lawn (usually a good thing)
Once constructed, the filter must be run continuously, even if the filter is unhooked from the tank. When not in use, a hose connecting the inlet and outlet can be put in place, thus allowing the filter to cycle water through itself. Or, as mentioned, an airstone buried in the sand will keep it alive for the duration of most emergencies
I have deliberately cut short of giving detailed plans for the construction of a FBF. I did this for a few reasons: First of all, you might build a filter to be exact specifications and then you’ll blame me if you are unhappy with it. A selfish reason to be sure, but I also think its a lot more fun to design your own filter using parts you find yourself. Remember that when designing your own filter, the most important thing to determine is the water velocity through the contact column and the water velocity required to fluidize sand of your grain size.
Levich, V.G., 1962, Physiochemical Hydrodynamics, Prentice-Hall Inc.
Spotte, S.H., 1979, Sea Water Aquariums, Wiley-Interscience
Wheaton, F.W., 1977, Aquacultural Engineering, Wiley-Interscience?