Ask the Membrane Expert

Ask a Question

Please leave this field empty.

In the case of high purity applications, the main criterion for clean vs. replace is salt passage. Any increase in salt passage to the demineralizers can be enough financial justification for replacement. As an example, even a slight decrease in salt rejection from 99.6% to 99.2% may not seem troublesome to someone treating brackish water for a potable application. But in a high purity application that generally employs mixed bed demineralizers to polish the RO permeate, the amount of salts that have to be removed by the demineralizer system would double, which in turn doubles the number of regeneration cycles. This corresponds to doubling of acid, caustic, and neutralization costs would drive the decision of whether to replace or clean.

Increased salt passage accompanied by increased differential pressure, and/or reduction in normalized permeate flow would suggest that the decline in salt rejection is a result of scaling/fouling, and that cleaning may improve permeate quality. However, if permeate quality is not recovered after cleaning, it can be assumed that the loss of salt rejection is permanent and that membrane replacement is necessary.

The differential pressure (dP) is the pressure lost due to friction as water passes through the system. You take a reading before the pressure vessels and after the pressure vessels of each stage, and that tells you how much pressure was lost across that stage.

A very high concentrate flow rate will result in a high dP because of the high friction. From a fouling perspective, scale, suspended solids, and biological growth, etc… build up in the feed channels and interfere with flow, thereby increasing friction and resulting in pressure losses.

If you are concerned about scale, then you should be monitoring your normalized permeate flow, not your dP. Monitoring normalized permeate flow allows you to detect scale formation and address it well before it is detectable by an increasing dP. Once your dP has increased, the scale will have already become very thick and will become more difficult to clean. The opposite is true for suspended solids, where you may see a significant increase in dP with minimal impact on normalized permeate flow.

In simple terms, pH is the concentration of acid protons [H+]. On the other hand, the alkalinity of a solution is its ability to neutralize acids. Alkalinity consists of ions that incorporate acid protons into their molecules so that they are not available as a free acid that can lower the pH. This is known as buffering.

For example acid reacts with CO3– to make HCO3-, and converts PO4— to HPO4–. This makes it so that a significantly higher quantity of acid is required to lower the pH compared to a solution that does not contain these ions. Deionized water can drop from pH 7 to pH 2 with just one drop of acid, while natural well water may require 200 – 300 ppm of acid just to lower pH from 7 to 6.

Alkalinity consists primarily of carbonate, bicarbonate, phosphate, borate, orthosilicate, sulfides, and organic acids. Most people refer to alkalinity as the concentrations of carbonate (CO3–) and bicarbonate (HCO3-) ions, which are the buffers that are typically present in the highest concentrations in natural waters. Bicarbonate in particular, is the strongest buffer (largest Ka value) and the effect of other buffers becomes insignificant in its presence.

At very high pH, like pH 12, the hydroxide ion [OH-] concentration is so high that it takes a significant amount of acid to neutralize enough of them before the pH drops. For that reason, hydroxides (OH) are considered as contributors to alkalinity above ~pH 10.5. At very low pH, hydronium ion [H+] concentrations are very high, and as a result, a much higher concentration of acid is required to further lower the pH.

Higher temperature shifts the equation to the right, slightly increasing the carbonate to bicarbonate ratio. At the same time, the acid (H+) concentration increases slightly which causes a slight drop in pH. This means that a warmer solution can have better buffering capacity despite a lower pH.

HCO3- <–> CO3– + H+

The opposite happens at lower temperature.

In reverse osmosis (RO) and nanofiltration (NF) membrane systems, bicarbonate is the primary driver behind the increased pH in the concentrate relative to the feed. Bicarbonates increase in concentration as they are rejected by the membrane, while CO2 concentrations remain constant since gases are not rejected. The increased ratio between bicarbonate and carbonic acid results in a shift to the left, absorbing more free acids and thereby increasing the pH.

H2CO3 <–> HCO3- + H+

**H2CO3 and CO2 are used interchangeably because H2CO3 converts instantly to CO2.

More on alkalinity can be found at the following blog post:

There are different types of antiscalants. Some are phosphonate based,and some are polymer based. Many are blends to provide synergistic effects of the two chemistries. Antiscalants have to be non-toxic, or they would not be approved by NSF for potable water applications. Both phosphonate and polymer based antiscalants are biodegradable over time. Ultraviolet light increases the rate of degradation as the chemical bonds are broken, and certain types of bacteria produce enzymes that cleave the bonds.

Recently some “green” antiscalants have been introduced into the market, with the claim that they biodegrade at a faster rate. In reverse osmosis membrane systems, anything that biodegrades too quickly will be a carbon source for bacteria and contribute to biofouling. It is best to avoid such chemistries from an operational standpoint.

H2S acts as a weak acid, much like carbonic acid:

H2S <–> HS- + H+
HS- <–> S– + H+

If your pH is above 5.5 and oxygen is allowed to contact with air, then sulfide ions will oxidize to element sulfur. We have recently come accross a case where a plant in South America made the mistake of using peroxide to oxidize the H2S and then attempting to remove the resulting elemental sulfur by ultrafiltration prior to the RO. We just performed an RO membrane autopsy and their membrane was completely coated with elemental sulfur.

The fact that you have H2S implies that your source is anaerobic. You should maintain the feed water under pressure from the well directly into the RO. If you operate at pH 5.5, your H2S will be 100% gas and pass to the permeate side. If you operate at a higher pH, then more of your H2S will be in sulfide ion form and be rejected, ending up in the RO concentrate (as long as your water is free of oxygen or other oxidizers).  If you have iron in your water, the sulfide ion can form a black ferrous sulfide scale on your membranes (again assuming anaerobic conditions).  This can be removed with low pH membrane cleaners.

In all cases, some H2S and/or sulfide ions will pass through to the permeate side.  Acid should be added to the permeate prior to a degasifier to maintain all sulfides as H2S, because air stripping does not remove ions, only gases. Any ions will react with the oxygen and precipitate as elemental sulfur in your degasifier.

Some plants with  low  concentrations of sulfides use strong oxidizers such as chlorine or ozone to oxidize hydrogen sulfide (H2S) to sulfate (SO4).  Such a treatment is too expensive with water containing high H2S  levels.

It is not possible to have colloidal silica in your feedwater unless your silica concentration is above the solubility limit under those conditions, resulting in polymerization and hence colloidal silica formation.

The difference between results from ICP-AA vs molybdate reactive silica is often mistakenly attributed to the presence of colloidal silica – but 99% of the time, the difference is due to calibration issues.

Silica leaks through demineralizers because it only partially deprotonates at neutral pH and is not fully charged unless the pH is very high. That has no relevance to its existence in colloidal or silicic acid form. Surface groups on colloidal silica will also deprotonate and become anionic at high pH, and can be removed no differently than silicic acid by ion exchange.

H4SiO4 = H3SiO4- + H+
H3SiO4- = H2SiO42- + H+

At neutral pH, the calculations show that no measurable amount of silica is charged.  Even at pH 10, about 40% of the silica remains unionized, and therefore cannot be removed by ion exchange resin.

As  an example, here are some calculations from a water analysis where reactive silica was 33 ppm as SiO2 (52.793 ppm as H2SiO4):

pH 7.3, Temp 25 deg C, Ionic Strength 0.0554, reactive silica 52.793 as silicic acid

mols/L         mg/L
[H2SiO42-]   8.80E-11       0.000
[H3SiO4-]     1.53E-06       0.146
[H4SiO4]     5.48E-04      52.646

pH 10, Temp 25 deg C, Ionic Strength 0.0918 (increased due to caustic addition), reactive silica 52.793 as silicic acid

mols/L           mg/L
[H2SiO42-]     9.33E-06        0.878
[H3SiO4-]        3.17E-04     30.105
[H4SiO4]         2.23E-04     21.472

Chlorine dioxide is a highly potent biocide that has a higher efficacy than hypochlorite in destroying protozoa, bacteria and viruses, without allowing them to build a tolerance.  It is effective over a wide pH range as both a biocide and metal precipitant and has gained popularity because it creates significantly lower concentrations of disinfection byproducts such as trihalomethanes (THM’s) and haloacetic acids (HAA’s) when compared to chlorine. However, at least one study has shown that THM formation is still significant when naturally occuring organic matter (NOM) such as humic acids are present (A.A.Stevens, “Reaction Products of Chlorine Dioxide”,1982).

Chlorine dioxide is a powerful oxidizer and can damage polyamide reverse osmosis (RO) membranes, so a reducing agent such as sodium bisulfite is usually dosed as a precautionary measure.

The use of chlorine dioxide at a low residual that would deplete prior to contact with the RO membrane surface  has been a hot topic.  But as with any other biocide, when the residual depletes to sub-lethal doses prior to membrane contact, biofouling will be enhanced.  This was confirmed in a 2010 publication by Shemesh et al, “The biocide chlorine dioxide stimulates biofilm formation in Bacillus subtilis by activation of the histidine kinase KinC.”

Chlorine dioxide can be used in the permeate but only at low dosages because it is reduced to chlorite and chlorate ions which can cause anemia and affect the nervous system.  In the United States, the EPA limits the allowable residual of chlorine dioxide in drinking water to 0.8 ppm.

Citric acid is an organic acid that is often used for removal of calcium carbonate scale and iron hydroxide.  Citric acid is not very effective at removing phosphate salts such as calcium and iron phosphates.  It cannot be used to dissolve sulfate scales or silica, and is ineffective for biofilm removal.

Although it is often recommended for organics removal, it is not very effective due to the protonation of the carboxylic acid functional groups in natural organic matter (NOM) in the low pH environment created by citric acid.  For this reason, attempting to clean at low pH when organics are present can compact the NOM into the membrane, resulting in a foulant that is more difficult to clean, or even irreversible.

Certain anti-caking agents present in citric acid can act as membrane foulants, and it is therefore always safer to use specialty RO cleaning chemicals.   Furthermore, in the United States, citric acid prices have been inflated due to anti-dumping duties imposed on imports.  For this reason, it makes little sense to use citric acid in place of specialty cleaning chemicals which are more effective over a broader spectrum and are of a similar cost.

Antiscalant dosage cannot be determined by TDS (Total Dissolved Solids).  The reason is that TDS can be made up entirely of sodium and chloride which do not have a scaling potential, or it could be made up entirely of calcium and sulfate which have a very high scaling potential.

In order to determine the correct antiscalant and dosage, the ion species in the water must be identified, and their concentrations have to be calculated based on the concentration factor.  The concentration factor is variable with the type of membrane (Reverse Osmosis vs. Nanofiltration) and the % salt rejection (NF membranes can vary from a 50% to 90% divalent salt rejection.)

We can then calculate the potential of the ions in the water to form a scale, and the antiscalant will be selected based on the type of scale most likely to form.  For example, if there is a potential for both calcium carbonate and calcium sulfate to form, an antiscalant that can control both these scales would be selected.  On the other hand, if there is a potential for calcium phosphate to form, a more exotic antiscalant would be required.

The dosage is then calculated based on the driving force for scale formation, which is dependent on the concentration of each ionic species, pH, temperature and ionic strength.

However, there are also other parameters that have to be taken into account when calculating dosage.  For example, antiscalants have a greater affinity for certain surfaces, like ferric hydroxide salts,  so when iron is present in the ferric state, a higher dosage is usually required.

So in order to obtain a recommendation for antiscalant selection and dosage, it is very important to perform a complete and accurate water analysis, and to measure pH and temperature immediately upon collection of the water sample.

The parameters that are required for an accurate recommendation are:

Cations: Ca, Mg, Na, Ba, Sr, Fe (specify ferrous or ferric), Mn, Al

Anions: Alkalinity, SO4, Cl, PO4, SiO2

pH, Temperature (both measured immediately upon sample collection), TDS (estimated from conductivity) and %Recovery.

Soluble iron and manganese can be removed by using a Manganese Greensand Filter. However, in order to separate salt from water, either Reverse Osmosis or Evaporation have to be used.

Events Calendar

  • AMTA/SEDA Technology Transfer Workshop: “Industrial and Municipal Water Reuse”

    April 4, 2017 to April 6, 2017 | Atlanta, GA
  • SWMOA 2017 Annual Symposium

    May 1, 2017 to May 3, 2017 | Henderson, NV
  • Southwest Chemistry Workshop

    June 19, 2017 to June 22, 2017 | Santa Fe, NM
  • SEDA Spring Symposium

    June 25, 2017 to February 28, 2017 | Delray Beach, FL
  • AMTA/NWMOA Technology Transfer Workshop: “Brackish Groundwater”

    July 11, 2017 to July 13, 2017 | Bozeman, MT

Case Studies

The Impact of Ferrous Ion Oxidation on Silica Scaling in RO Systems

Controlling Aluminum Silicate Formation in Membrane Separation Processes

An Evaluation of Corrosion Control Additives for Potable RO Permeate

A Relationship Between Phosphate Scales and Silica Fouling in WasteWater RO Membrane Systems

Excessive Sulfuric Acid Dosing Resulting in Irreversible Scale Formation

Optimizing Scale Inhibition Costs in Reverse Osmosis Desalination Plants