What is nanofiltration?

Ken Sutherland

A relatively recent development in membrane processes, Ken Sutherland looks at the rapidly expanding field of nanofiltration, its characteristics and its applications.

The background to nanofiltration

The burst of filtration and filtration-related activity that followed the development of the phase-inversion process for the manufacture of polymeric membranes, in the early 1960s, led to the establishment of three membrane separation processes: reverse osmosis, ultrafiltration and, more recently, microfiltration. These processes took the separation spectrum from the traditional cut point limit of standard filtration of around 0.01 mm (10 μm) to the very finest distinct solids, a few nanometres in size, and enabled the separation of large molecules from solution. The actual size ranges vary somewhat from source to source, but there is general agreement that microfiltration covers the range 10 μm down to 0.1 μm, while ultrafiltration covered 0.1 μm down to 0.005 μm (5 nm) in terms of discrete particles or Molecular Weight Cut-Off (MWCO) figures of 300,000 down to around 300 Daltons for dissolved materials. Reverse osmosis, of course, was designed to retain the very small sodium chloride molecule, which meant passing nothing else but water.


These intended size ranges actually still left a gap in their coverage at the lower end of that for ultrafiltration (at around 100 to 300 Daltons). Membrane development was fairly rapid during the 1970s and 1980s, leading to a “loose RO” membrane process, which was given the name “nanofiltration” at the end of the 1980s.
In this sense, then, nanofiltration is a fairly recent development in the range of membrane separation processes, which takes in the upper end (in separation size terms) of reverse osmosis, and the lower end of ultrafiltration, covering MWCO values of 100 to 1000 Daltons. It deals with materials that are dissolved in a liquid, and not with distinct particles suspended in the liquid. The separation between solute and solvent occurs by diffusion of the molecules of the solvent through the mass of the membrane material, driven mainly by a high transmembrane pressure, and not through any physical hole (pore) in the membrane. Some of the solute molecules may also diffuse through the membrane, either by the process designer's intent, or because the solute has a finite (although very small) diffusion coefficient in the membrane material.
The key difference between nanofiltration and reverse osmosis is that the latter retains monovalent salts (such as sodium chloride), whereas nanofiltration allows them to pass, and then retains divalent salts such as sodium sulphate. Robert Peterson, in his Foreword to Elsevier's Nanofiltration – principles and applications, describes reverse osmosis (especially in the water treatment business) as the main course, the steak perhaps, of a meal, whereas nanofiltration “is like the wine menu … an opportunity for creativity and exploration”.
How nanofiltration has developed
The preceding paragraphs have described the origins and nature of the generally accepted process called nanofiltration, which is a liquid-phase separation removing dissolved solids, carried out by means of membranes, with a relatively high transmembrane pressure. However, the progress of much of the filtration business is being driven by demands for finer and finer cutpoints, in both liquid and gas filtration, and these demands are now being met by the use of correspondingly finer fibres to make the filter media. Increasingly, these fibres have diameters of significantly less than one micrometre, and are therefore measured in nanometres, and are becoming commonly known as nanofibres. These are used to make composite filter media, with a web of nanofibres supported on a coarser substrate.
The very fine filtration that can be achieved with these nanoweb media is taking the separation process that is effectively microfiltration to much lower cutpoints. The materials are also being referred to as membranes, even though they are very different in format from the semipermeable plastic sheet still most commonly thought of when membranes are mentioned. It is worthy of note that, at the 10th World Filtration Congress (in 2007), out of a total of almost 250 separate papers and 85 poster presentations, there were 12 that featured nanofiltration, and 14 concerning nanofibres as filter media.
Whilst it is hoped that the two systems – nanofiltration and filtration with nanofibres – are sufficiently different so as to avoid their confusion, both are covered in the rest of this article.
It is worth noting that the term “nanotechnology” is now very widely used, referring to a whole range of scientific, engineering and manufacturing activities involving very small things. Unfortunately the term has entered the public consciousness with a component of “fear of the unknown” attached to it. This does not concern nanofiltration, since the media involved in it are mostly continuous and indistinguishable from RO or UF membranes. It does concern nanofibre production and use, however, and the makers and users of nanofibres will have to take care not to magnify the concern.
The membrane separation process known as nanofiltration is essentially a liquid phase one, because it separates a range of inorganic and organic substances from solution in a liquid – mainly, but by no means entirely, water. This is done by diffusion through a membrane, under pressure differentials that are considerable less than those for reverse osmosis, but still significantly greater than those for ultrafiltration. It was the development of a thin film composite membrane that gave the real impetus to nanofiltration as a recognised process, and its remarkable growth since then is largely because of its unique ability to separate and fractionate ionic and relatively low molecular weight organic species.
The membranes are key to the performance of nanofiltration systems. They are produced in plate and frame form, spiral wound, tubular, capillary and hollow fibre formats, from a range of materials, including cellulose derivatives and synthetic polymers, from inorganic materials, ceramics especially, and from organic/inorganic hybrids.

Recent developments of membranes for NF have greatly extended their capabilities in very high or low pH environments, and in their application to non-aqueous liquids. The plastic media are highly cross-linked, to give long-term stability and a practical lifetime in more aggressive environments. NF membranes tend to have a slightly charged surface, with a negative charge at neutral pH. This surface charge plays an important role in the transportation mechanism and separation properties of the membrane.
As with any other membrane process, nanofiltration is susceptible to fouling, and so nanofiltration systems must be designed to minimise its likelihood – with proper pretreatment, with the right membrane material, with adequate cross-flow velocities to scour the membrane surface clear of accumulated slime, and by use of rotating or vibrating membrane holders.
Industrial applications of nanofiltration are quite common in the food and dairy sector, in chemical processing, in the pulp and paper industry, and in textiles, although the chief application continues to be in the treatment of fresh, process and waste waters.
In the treatment of water, NF finds use in the polishing at the end of conventional processes. It cannot be used for water desalination, but it is an effective means of water softening, as the main hardness chemicals are divalent. At first sight, NF would not seem to have much place in MBR processes, because the higher transmembrane pressure differentials needed for NF are not available in most bioreactor systems, but there are some specialised uses for MBRs in which NF is finding a place. Smith's review covers the whole field of nanotechnology well, including reference to Argonide's NanoCeram fibres of 2 nm alumina, used for the filtration of 99.9999% of bacteria, viruses and protozoan cysts (now available as Ahlstrom's Disruptor technology).
NF membranes are also used for the removal of natural organic matter from water, especially tastes, odours and colours, and in the removal of trace herbicides from large water flows. They can also be used for the removal of residual quantities of disinfectants in drinking water.
Food industry applications are quite numerous. In the dairy sector, NF is used to concentrate whey, and permeates from other whey treatments, and in the recycle of clean-in-place solutions. In the processing of sugar, dextrose syrup and thin sugar juice are concentrated by NF, while ion exchange brines are demineralised. NF is used for degumming of solutions in the edible oil processing sector, for continuous cheese production, and in the production of alternative sweeteners.
There are probably as many different applications in the whole chemical sector (including petrochemicals and pharmaceuticals) as in the rest of industry put together. Many more are still at the conceptual stage than are in plant use, but NF is a valuable contributor to the totality of the chemicals industry. The production of salt from natural brines uses NF as a purification process, while most chemical processes produce quite vicious wastes, from which valuable chemicals can usually be recovered by processes including NF. The high value of many of the products of the pharmaceutical and biotechnical sectors allows the use of NF in their purification processes.
The paper pulp industry uses a very great quantity of water in its production processes, a quantity that the industry is striving to reduce, mainly by “closing the water cycle” – a system in which the purification properties of NF have a major role.
All of these specifically mentioned applications have been water-based, but nanofiltration is not restricted to the treatment of aqueous suspensions. Indeed one of the largest NF plants was installed at a petroleum refinery for the dewaxing of oils. Boam and Nozari, in their review of organic solvent nanofiltration, point out that many organic system separation processes are quite highly energy intensive, and that, by contrast, OSN can be quite an energy saving alternative (for example, by comparison with distillation).
In aqueous systems, nanofiltration uses hydrophilic polymeric materials, such as polyether-sulphone, polyamides and cellulose derivatives. These materials, in contact with organic solvents, quickly lose their stability. Special membranes have therefore been developed to provide the same kind of performance as in aqueous systems, and they are now used for solvent exchange, solvent recovery and separation, for catalyst recovery and for heavy metal removal.
Nanofibre media
The synthetic materials, both organic and inorganic, that are nowadays being spun from the molten state into ever finer fibres, are no different from the materials that have been used for decades for this purpose (except for the ever widening range of thermoplastic polymers that are available). What has changed has been the equipment downstream of the spinaret, which enables a wide range of fibre diameters to be produced. Starting over 40 years ago with spun bonded media, whose fibre diameter was 10 μm or more, the list runs through flash spun and melt blown (at just over 1 μm), to electrospun materials, which are approaching 100 nm in fibre diameter capability.
Each of these materials can be produced as a random array of fibres as a web, which, in itself, makes a very good filter medium, so long as it is adequately supported on a stronger substrate. Tucker gives a good review of these materials in introducing du Pont's new HMT media for liquid or gas filtration. United Air Specialists has developed nanofibres for dust removal, as has Donaldson with its Ultra-Web media.
The fine spinning techniques have proved amenable to the production of carbon and ceramic fibres, and are obviously going to result in a major segment of the filter media business, especially for air filtration. Because these media are capable of removing contaminants to below 0.1 μm, they are going to be counted as membranes – and certainly known as nanomembranes.
For references, contact the editor.

Share this article

More services


This article is featured in:
Drinking water  •  Electronics/process water  •  Food and beverage  •  Pharma and biotech  •  Pulp and paper  •  Water and Wastewater



Anonymous said

19 January 2012
Ken, I realise this article is old, but it may still be hitting people's screens with the right search, so I thought I'd add my 10 cent's worth.... Ken, to continue with the kitchen theme, I would use the analogy of steak knives; the RO membrane is a big crude meat cleaver, capable of a big glorious chop with no subtlety - all RO membranes do essentially the same thing, separating water from all salts; whereas nanofiltration membranes have a delicate beauty akin to a good set of sushi chef's knives, very sharp but delicate, capable of very subtle separations, particularly when allied to chemistry adjustments that can change both the membrane characteristics and the dissolved species behaviours. Separations are achieved in part on charge, in part on ionic radius, and in part on counter-ions, each being ipacted by pH and associated dissociation / speciation. "Softening" would be the most common example, but is by no means the only one. I have used various NF membranes, from 200 Mwco to 1000 Mwco, to separate ammonium sulphate from sodium chloride (including the Donnan effect), free cyanide from copper cyanide, to concentrate and purify uranium sulphate (by lowering chloride and silica) and even to remove high molecular weight organics in 60C bayer liquor. Maik, from an inorganic chemistry perspective at least, I'd see NF being defined in the range of 200-1000 Mwco for separating dissolved ionic salts, and not as a virus removal tool. The membranes I used were all spiral wound, akin to RO, and as such not particularly tolerant of TSS, so I was careful to pre-treat with UF beforehand. I also saw precipitation of salts on the permeate side, which was and remains a challenge in some applications. To me, the world of NF is easily more exciting than RO, particularly SW-RO, but the complexity (read 'cost'), their small size typically, and limited 'repeat sales' of such plant, probably ensures that NF will remain an interesting hobby for most engineers, which is a pity as it offers potentially huge benefits. Thanks for the article Ken, hope to read more in time. Regards to all readers, Paul Forder, Perth, Western Australia.

MWJornitz said

16 June 2010
As I always, I read this article with interest and enjoyed the wealth of information within. The author writings always impress and are highly educational.
However, I would like to take the opportunity to clarify some minor oversights or differences in definition.
For example the article states that phase inversion processes created microfiltration membranes recently, more recent than ultrafiltration, which is not the case. Phase inversion processes in membrane production cover vapour phase precipitation, precipitation in a non-solvent (quenching) and potentially also polymere cooling. Vapour phase precipitation (casting) exists since the early 1930s and is one of the oldest microfiltration membrane processes.
Now, nanofiltration (NF) as described within the article shall bridge the gap between RO and UF and separates divalent salts, but not partciles. Any virus filtration expert would reject this statement, as the most commonly used term for nanofiltration (NF) can be found in the world of virus removal filtration, the separation of virus particles. The retention rating of these filters lies within a range of 10 - 100 nm, most commonly used though are 20 and 50 nm filter ratings for parvo and retro viruses. The question to be posted, what would be the right definition, as the term nanofiltration is in use for virus removal application since the late '80s.Maybe one cannot just define nanofiltration within a chemical or food & beverage setting, but also requires to contemplate all biopharmaceutical applications.
As mentioned, filtration with nanofibres is within a complete different spectrum. As nanofibres progress, one sees a multitude of possible nanofibre filters come to use within prefiltration applications, for example to protect nanofilters, like 20 nm virus filters. The wealth of the technology is possibly untapped so far, as these fibrous materials are gaining lower fibre diameters. Although, as stated in the article, these fibrous materials will never be called membrane, as it is not the retention rating which defines a membrane or a randomly spun fleece material, it is the defined structure and solid state of membrane, which defines it as such.
Sincere regards
Maik W. Jornitz

Note: The majority of comments posted are created by members of the public. The views expressed are theirs and unless specifically stated are not those Elsevier Ltd. We are not responsible for any content posted by members of the public or content of any third party sites that are accessible through this site. Any links to third party websites from this website do not amount to any endorsement of that site by the Elsevier Ltd and any use of that site by you is at your own risk. For further information, please refer to our Terms & Conditions.

Comment on this article

You must be registered and logged in to leave a comment about this article.