Hobbyists moving to reef tanks from the freshwater or even saltwater fish only tanks are frequently puzzled over the apparent lack of massive power filters that they are familiar with in the other hobbies.  Reef tanks require a degree of finesse rather than sheer brute force when it comes to filtration.  Whereas the other hobbies typically are trying to achieve crystal clear water via removal of all particulate matter and rely upon water changes to dilute pollution, reef tanks are trying to achieve crystal CLEAN water through removal of all pollutants at a level that water changes alone could not accomplish.

What does this mean?  In simple terms:

There are a number of ways to try to achieve this goal which are outlined below.

Wet / Dry Filter

Old school was that a filtration device called a wet / dry filter was necessary for success with reef tanks.  In fact, there are many still being sold today.  Wet/dry filters are large acrylic or grass filters that become the sump for the tank and include a tower area in which return water from the tank is trickled or sprayed over a plastic media.  The logic was that oxygen loving bacteria would colonize the media and convert the waste products generated by the tank into relatively harmless nitrates.  This was the 'dry' portion of the filter.  The 'wet' portion typically included a sponge for mechanical filtration and space for heater, carbon and possibly a protein skimmer.  They in fact accomplished this mission, but have several important downsides.  Tanks with wet/dry filters frequently have problems with nitrates continually increasing, only to have the nitrates drop to zero when the wet/dry filter is removed. Why this might be, is not fully understood, but it is a well established phenomenon.  The conversion of wastes to nitrates is accomplished efficiently, but the final step to convert it to harmless gases is not and therefore the nitrates accumulate in the system.  Also, the mechanical filtration that is sometimes employed in wet/dry filters is counter productive.  Reef tanks house countless organisms whose lives depend on filtering the water.  Mechanical filtration is redundant and counterproductive in the long-term.

Long story short, do not spend money on a wet/dry filter.   If you already have one and are having problems with nitrate accumulation, slowly remove the biomedia until the unit serves as simply a sump.  

Protein Skimmer

The first step in clean water is removal of pollutants before they enter the nitrogen cycle.  Mechanical filtration of visible particulate matter in the water may seem like the best way to achieve this, but it is not.  Much of the visible matter is harmless flocculent while the harmful stuff is invisibly dissolved in the water column.  This is where protein skimming comes into play.

A protein skimmer operates on the principal that protein and other organic wastes tend to collect at the water surface due to the affinity to having half their molecule in the air and half in the water.  This is why scum forms on the water surface.  A bubble floating in the water tends to collect these wastes for the same reason.

A protein skimmer consists of a reaction chamber through which tank water is mixed with fine air bubbles.  These bubbles collect the wastes onto their surface through natural processes as the bubbles rise through the chamber.  The bubbles collect and carry the waste products up toward to the top of the reaction chamber.  As these waste products rise in the chamber, a foam is created which is eventually pushed up through a tube and into a collection vessel.  These organic wastes are mostly fish wastes and the protein skimmer does a good job of removing them before they have to go through the biological breakdown cycle.  A person only has to see and smell the output from the protein skimmer to realize how much gunk these things do remove from the water. 

Protein skimmers are constructed in several different ways.  

Originally,  skimmers were constructed of a tall acrylic tube with a wooden air stone in the bottom.  Wood was used to provide the fine air bubbles necessary to make the skimmer work properly.  The water was forced to travel down the tube while the air bubbles were trying to rise up through the water.  These skimmers worked well and are still in use in small quantities today.  Their downside is primarily the need to occasionally replace the airstone and the need for a fairly high pressure air pump.

Most skimmers sold today are called venturi skimmers.  These skimmers remove the need for the air stone and air pump by creating the bubbles through a venturi effect.  Water is forced via a water pump under pressure through a restriction which has a small air hole in it called a venturi valve.  The restriction causes a vacuum to be created and air is sucked in through the small hole and mixes with the water.  Venturi skimmers can create a more turbulent water column than airstone based skimmers and do not require as tall a reaction chamber to achieve the same affect and therefore can be easier to fit under the tank stand.

Protein skimmers have been in the hobby for quite some time and have proven their value in almost all situations.  It is highly recommended that protein skimming be provided on any reef.  Although it is possible to setup a reef tank with a skimmer, there is no downside to providing one unless you are trying to setup a tank which houses specimens from more turbid waters, such as Goniopora which might benefit from increased nutrient levels in the water.

Protein skimmers can either be purchased as in-sump models or hang-on-tank (HOT) models.  In-sump models are generally more powerful and should be used if a sump is used on the tank.  If no sump is used, then a HOT is the only other option.  Some of these hang on models are ineffective enough to make them mostly worthless unfortunately.

Protein skimmers also have the incidental, but very valuable side benefit of providing a high degree of gas exchange and oxygenation of the water.

Algae Filtration

A method of filtration that has found favor with those who prefer the natural approach.  The logic behind this approach is that algae will consume nitrates and other water pollutants as they grow and subsequent frequent pruning of the algae will export those nutrients outside of the tank.  The algae employed is usually one or more of the Caulerpa species, although any algae, including hair algae, will perform the same function.  Unless the tank is a lagoon tank, the algae is normally grown outside the main display tank either in the sump or a refugium if one is used.  Simply adding a small light over the sump and and throwing some Caulerpa in gets this process started.

Downside to this method includes the fact that Caulerpa algae can go sexual and disintegrate almost overnight, dumping a heavy biological load back into the system.  Frequent pruning seems to help minimize the chance of this occurrence.  Algae growth can also leach chemicals back into the water that tend to yellow the water.  Carbon can be used to minimize this problem.

Deep Sand Beds

Sand beds with medium to find grain particles that are over about 1" in depth create oxygen poor conditions in their deeper layers that are conducive to growing the type of bacteria needed to convert nitrates to harmless gases.  DSB's take this strategy to extremes through the use of sand beds that are generally 4" to 6" in depth.  At this point, they enter the realm of intentional (vs. incidental) filters and so are included here.  Old school is that deep substrates will allow noxious gases, like hydrogen sulfide, to form  in the anaerobic areas  that can cause problems in the tank.  Current thinking is that this is not a major concern and that this setup closely approximates what occurs in the substrate in the ocean.  DSB's, and for that matter any substrate, should not be disturbed more than necessary.  At one point, vacuuming the substrate was a common practice to keep it clean.  Current thinking is that the sand bed should not be disturbed more than necessary to allow it to create and maintain the anaerobic regions.  The sand bed also becomes populated with worms and other organisms that further help to recycle wastes within the tank.

The downside to this method is primarily that the DSB can consume large amounts of the tank vertical depth and be somewhat unsightly where it is in view along the front glass of the tank.  For this reason, some DSBs are being setup in sumps where it's work can be done out of view of the main display tank.

Carbon

Carbon is frequently employed in reef tanks as an adjunct to other filtration methods.  Carbon can be from a natural mined source similar to coal, or can be produced through extremely high temperature processing of coconut shells or similar.  Carbon absorbs compounds in the water that can discolor the water improving the look and light penetration.  It can also absorb other compounds that might be otherwise harmful to the tank.  Some reef keepers are concerned that it might also remove useful components as well, but the benefits seem to outweigh any likely negatives.  Carbon is recommended to be used continually in small quantities.  For instance, a quantity of about 1/2 cup for a 55 gallon tank would be appropriate.  The carbon should be replaced about once a month to keep it effective.   Some people only use GAC a few days a month to minimize it absorption of any good compounds, but the sudden change in water clarity and composition this method imposes on the tank does not make much sense and is counter to the goal of maximizing water stability within the reef environment.

When using carbon, it is very important to chose a high grade version which is typically called GAC (Granulated Activated Carbon).  GAC is fairly soft and has small rice grain particle sizes that increase absorption rate.  High quality versions also minimize any phosphates or other contaminates that they might otherwise contain since they are organic in nature.  The small grains of GAC require that it be contained in a mesh bag or similar to prevent the particles from migrating around.  It is also important to not put loose GAC in direct water flow as there is evidence that tumbling GAC can create small carbon particles that enter the tank and can be somewhat detrimental to the tank inhabitants.  If used in a high water flow area, ensure the GAC is tightly packed to prevent its movement, otherwise a quieter part of the sump where the water can circulate gently around the GAC is the best place to locate it.

The carbon typically sold in LFS for freshwater use that is very hard and has very large particle size should not be used in reef tanks.  They are fairly ineffectual and may have unwanted contaminates.

Other Contraptions

Reef Sock - A large fine woven sock is sometimes placed over the overflow tube where it enters the sump.  The sock can catch large to fairly fine particles that go over the overflow and provide a form of mechanical filtration.  The other benefit of a sock is that it can help to keep bubbles from the overflow from making it back to the return pump and to the tank.  I currently use a sock on my 225 gallon tank.  I clean it every 2-3 days by turning it inside out and spraying it with a high pressure spray from a garden hose from the inside out.  it is surprising how much stuff the sock can catch.

Hang-on style power filters have little value in a reef tank and should generally not be used, although on a small tank an outside power filter can sometimes be used to house a small heater and carbon outside the tank and provide water flow in the tank.  In that case, the filter portion of the outside filter is left unused.  There are some outside filters which are designed specifically for reef tanks that include a small protein skimmer and these are generally ineffectual enough to be mostly useless.

Under tank power filters are also generally not used, although they sometimes are used for holding GAC, or other filter media that might be used on occasion.

UV sterilizers were quite popular for a while and are still used by some hobbyists.  They work on the principal that a stream of water is passed through an enclosed tube and over a intense UV light source which kills any organisms that it comes into contact with, effectively sterilizing the water.  The value is dubious at best although a case may be made for it when the water has parasites or similar that need to be dealt with.  There is some evidence that they may be useful in combating dinoflagellate outbreaks.

Utilizing ozone to sterilize water was popular at one time, but is less used today.  Ozone is a gas which is formed in the presence of an electrical field and has a very noticeable odor to it.  The smell associated with lighting or other electrical arcs is ozone.  Ozone, unlike UV, it is not limited to a small reaction chamber and has far reaching side affects.  Ozone is a strong oxidant and it can easily damage delicate membranes such as fish gills.

Current reef keeping theory is that a healthy tank automatically helps to fight bad organisms in the water and such preventative measures such as UV sterilizers and ozone are usually unnecessary and potentially bad.  Unlike fish only tanks that seem to frequently have massive fish infections from parasites and bacteria which must be controlled via medication or mechanical means, reef tanks that are in good health rarely have significant disease problems and when they do occur, they frequently seem to resolve themselves on their own.

Further Reading:

Build Your Own Foam Fractionator
By L. Jackson

Granular Activated Carbon: Part 1 
By R. Harker

Granular Activated Carbon: Part 2 
By R. Harker