Membrane fouling is a loss of permeability, not one deposit
Membrane fouling in wastewater treatment is the build-up of material on the surface of a membrane or inside its pores, and it is measured as a cost in pressure. Run the membrane at a fixed permeate flow and the transmembrane pressure climbs; run it at a fixed pressure and the flow falls away. MBR membrane fouling and the fouling of an ultrafiltration train are the same phenomenon on different membranes, and what matters in both is the rate at which the pressure moves rather than the pressure itself.
The word suggests one material, and the first useful step is to stop treating it as one. Mixed liquor arriving at a membrane carries suspended solids, colloids, dissolved organic matter, multivalent ions and a living microbial community, and any of them can settle on the same square metre of membrane. What the plant sees is usually a combination: cake layer and pore blocking at the surface, an adsorbed gel of biopolymers over it, and a mineral scale crystallising where the concentration is highest.
The combination matters because the shape of the transmembrane pressure trend is a diagnosis. A deposit a backwash lifts is not the same deposit as one that needs a hypochlorite soak, and the wrong answer spends membrane life. The four mechanisms below are the ones the literature and operating experience keep returning to.
The four mechanisms of membrane fouling
The four overlap in time and share one symptom. Each is answered by a different action, and an action aimed at the wrong mechanism makes the deposit worse rather than better.
Cake layer
A cake layer is the deposit that builds on the membrane surface from material too large to enter the pores: floc, colloids, cell debris and the polymers they carry. It is the most common of the four and the most forgiving, because it is held on the surface rather than inside the pore. A fresh cake is what a relaxation, a backwash or the shear from an air scour is designed to lift. A cake left to compact, and then to grow a biofilm over itself, stops behaving like a loose deposit and starts behaving like a second membrane of unknown resistance.
Pore blocking
Pore blocking happens when particles small enough to enter a pore do so, and either narrow it or seal it. The classic description divides it into complete blocking, where a particle seals a pore mouth, and standard blocking, where material deposits on the pore wall and constricts it. It is the mechanism behind the fastest part of a pressure curve: a sharp rise in the first hours or days after a membrane is cleaned or commissioned, then a much flatter slope once the accessible pores are taken. Only part of that rise comes back with a physical clean, which is why an early rise a backwash does not recover is worth recording rather than dismissing.
Organic adsorption and the gel layer
Dissolved organic matter — the soluble microbial products and extracellular polymeric substances that a biological process produces — adsorbs directly onto the membrane material and forms a gel at the interface. This is a surface chemistry event rather than a sieving one: the molecules are far smaller than the pores, and they arrive whether or not solids are being retained. The gel is thin, dense and resistant, and it is the mechanism most often responsible for fouling that a physical clean cannot touch and a chemical clean can. It is fed by the biology: a long sludge age, a high mixed liquor concentration and a low food-to-microorganism ratio all push soluble microbial products into the water the membrane has to pass, so a change in the biological operating point shows up at the membrane first.
Inorganic scaling and biofouling
Inorganic fouling is precipitation. Calcium carbonate, calcium sulphate, struvite and the oxides and hydroxides of iron, manganese and aluminium crystallise on the membrane where concentration and pH make them insoluble, and they arrive with a signature none of the other three share: a sudden step, no recovery from a physical clean, and a chemistry of their own. Biofouling is the part usually folded into this group rather than listed beside it, because a biofilm is a community held in a matrix of the same extracellular polymers, so it binds particulate and organic material to the surface while acid and alkali attack the mineral deposit around it. In a membrane bioreactor the biology is not a contamination from outside but the process itself, which is why membrane bioreactor fouling behaves differently from fouling in a filtration train that receives no biology at all.
Reading the four mechanisms from the TMP curve
A plant that logs transmembrane pressure and permeate flow has a diagnosis in the record, provided the record is read as a shape rather than as a number. Four shapes recur.
- A steady climb that a relaxation or a backwash pulls back towards the starting value is a cake layer behaving as a cake layer. It is a maintenance signal rather than a failure, and it is the signal the routine cycle is designed to answer.
- A fast rise in the first hours or days after a clean, flattening afterwards, points at pore blocking. The steeper that early phase, the smaller the particles reaching the pores, and the more the pre-treatment in front of the membrane has to be looked at.
- A rise that survives every physical clean, and that is worse after a period of high sludge age or high mixed liquor concentration, is the adsorbed gel. It is answered by chemistry and by the biological operating point together.
- A sudden step at a point where nothing else changed is scaling. It is often preceded by a change in the feed chemistry, a pH excursion, or a chemical clean that used a reagent the deposit does not dissolve in.
One more shape has to be separated from all four, because it is not fouling. Water viscosity rises as it cools — by roughly half again between 25 °C and 10 °C — and resistance to flow rises with it, so a constant-flux plant needs a higher transmembrane pressure to hold the same permeate flow in winter than in summer. A winter rise that reverses as the water warms, and that leaves no change in the slope of the recovery after a clean, is temperature rather than deposit. Judgements about fouling rate belong on temperature-corrected pressure.
Where membrane bioreactor fouling is decided before the membrane sees it
Most of the fouling potential of a stream is fixed upstream of the membrane. Five operating decisions carry most of the effect.
- Screening. A membrane bioreactor is normally protected by fine screening, commonly quoted at 1 to 3 mm, because fibres, hair and grit that pass a coarse screen arrive at the membrane and cannot be removed from it by any chemistry.
- Sludge age and mixed liquor concentration. A longer sludge age lowers surplus sludge and raises the soluble organic fraction; a higher mixed liquor suspended solids concentration, commonly quoted at 8 to 12 g/L for a submerged MBR, raises viscosity and the load on the air scour. Past a point the extra concentration costs more in fouling than it saves in tank volume.
- Aeration at the membrane. The air added under the modules is there to scour the surface, not only to supply oxygen, and a plant that trims it to save energy finds the saving in the transmembrane pressure trend a few weeks later.
- Flux. Operating below the critical flux — the flux above which the fouling rate stops being modest and starts being steep — is the largest single lever on cake formation.
- Temperature. Cold water raises viscosity, slows biological kinetics and changes the composition of what the biology excretes, so a plant designed on a warm-season sample can meet all four mechanisms at once in its first winter.
Bands of the kind quoted above are the ones usually given in the literature and in operating experience, and they should be read as a starting method rather than as a specification, because the balance between them belongs to the water and the biology of a particular plant.
Reversible and irreversible fouling, and the rate that matters
The distinction between reversible and irreversible fouling is decided by what removes the deposit, not by what it is made of. Material a relaxation or a backwash removes is reversible; material that comes back only with a chemical clean is chemically reversible; material that no clean recovers is irreversible, and it accumulates until the membrane is replaced. Every clean moves some fraction of the deposit from the second group into the first, and that fraction is the measure of whether the cleaning regime is working.
The rate matters more than the level. A plant that records permeability after each clean and plots the slope between them can tell a stable fouling rate from an accelerating one months before a pressure ceiling is reached, and an accelerating rate is the signal that something upstream has changed: a new industrial discharge, a screen bypassing, a sludge age that has drifted. Reading only the absolute pressure postpones the diagnosis to the point where the only answer left is a recovery clean.
What to record so that a diagnosis is possible
- Transmembrane pressure and permeate flow at the same moment, so permeability can be calculated rather than estimated.
- Temperature, so the viscosity correction can be applied before any conclusion is drawn.
- The time and the result of every relaxation, backwash and chemical clean, with the permeability recovered by each one.
- Mixed liquor suspended solids, sludge age and dissolved oxygen in the membrane tank.
- Permeate turbidity or suspended solids, which separate a membrane problem from a biological one.
- Every change in the influent: a new industrial discharge, a seasonal load, a screen that has been bypassed.
From a diagnosis to a membrane duty
The four mechanisms are why the running cost of a membrane plant is a maintenance question as much as a design question. The companion article on membrane cleaning chemicals and frequency takes the pressure trend described here and turns it into a cleaning regime. The integrated MBR equipment page sets out how the membrane, the biology and the air scour are packaged together, and the flux, the cycle and the cleaning chemistry that suit your membrane module are confirmed with the factory for your membrane module and your water rather than assumed from a table.