Introduction To Diafiltration and Ultrafiltration
Diafiltration is a membrane separation process which often forms part of a series of unit operations in downstream processing. Diafiltration is distinguished from ultrafiltration because it involves the addition of water to replenish the volume lost during the filtration process. Ultrafiltration traditionally involves the separation of components combined with concentration. Diafiltration is usually exploited to remove and clean out impurities such as toxins and undesirable compounds from a batch. It is quite common for diafiltration and ultrafiltration to be combined so that there is both a purification and cleansing of a biochemical material alongside its concentration.
Incidentally, nanofiltration should not be ignored but is best discussed separately and references in our article on membrane processing in the dairy industry.
Terms Used In Membrane Processing
The two main fractions are the retentate and the permeate. The retentate is that volume of the original batch which is retained and does not pass through the semi-permeable membrane. The permeate is that portion of a batch which includes water and solutes which can pass through the semi-permeable membrane.
Diafiltration is commonly encountered in the food and pharmaceutical industries. It is highly useful for separating smaller molecules especially solutes from their much larger counterparts such as proteins and carbohydrate polymers. Notable examples in the research literature exist where this technique has been usefully applied. Good examples include the production of protein isolates from soy, legumes like beans and peas, and milk. Here, ultrafiltration coupled with diafiltration is practised on a very large scale.
Factors Influencing Ultrafiltration and Diafiltration
The performance in both ultrafiltration (UF) and diafiltration (DF) depends on a large number of factors. These are:-
- The type of molecules to be retained or separated using the semi-permeable membrane
- The amount of fouling material in the starting batch including levels of insoluble materials
- The type of membrane used
- The salt content of the feedstock and the osmotic pressure it can exert
- The flow-rate of the retentate if a cross-flow filtration system is employed.
- The applied driving pressures that force liquid and solutes across the membranes
- The shape and charge of the molecules
- The ability of the solutes to bind other compounds such as metal ions or interact with other molecules through hydrogen bonding.
Various types of filtration geometry are used. There are flat sheets, hollow fibres, tubular and spiral wound systems. The examples below have used all four types in their development.
Membranes used must have high performance in their mechanical strength, show good chemical resistance, a high level of hydraulic permeability, have an average definable porosity and pore size distribution.
Materials used in Membrane Construction
Membranes are either made from organic polymers or inorganic materials. The organic membranes are usually prepared by phase inversion. The organic polymers include:
- polysulphone (PS)
- cellulose acetate (CA)
- polyethersulphone (PES)
- regenerated cellulose
- polyamides (PA)
- polyvinylidenedifluoride (PVDF)
- polyacrylonitrile (PAN)
The inorganic membranes use the following:
- aluminium oxides
- borosilicate glass
- zirconia and stainless steel
- zirconia carbon
- pyrolysed carbon
Polyethersulphone (PES) has increasingly increased in popularity. It has very low protein binding, is hydrophilic and easily wetted. It shows good mechanical strength with a wide range of resistance to chemicals. A typical pore size is 25,000 molecular weight cut off.
Spiral Wound Systems
These have the advantage of having a high membrane packing density whilst being relatively low cost to produce. The main issues are their capacity for membrane deformation and the difficulty of cleaning them.
The hollow fiber membrane (HFM) is more acceptable than other modules, due to its high membrane packing density, structural integrity, construction and thus it can withstand high permeate backpressure (Kuriyel, 2000, Cheryan, 1998 and Nielsen,
2000). . Hollow fiber membrane modules have several significant advantages over other conventional membrane modules because you can recover permeate flux after cleaning of the membranes (Cheryan, 1998 and Nielsen, 2000). Moreover, being a compact module, its structural configuration allows a high membrane surface area resulting in the increase of the output through the process, while utilizing minimal space, with low power consumption.
Examination Of Membrane Behaviour
The variation of permeate flux with transmembrane pressure is a useful characteristic to monitor of any system. A comparison is made with water flux just to eliminate issues associated with equipment configuration. There is with an ideal solution, a linear relationship between the flux across the membrane and the transmembrane pressure (TMP) which is expressed as:
J = TMP/( Rt*μ)
where the flux flow rate per unit membrane area is directly proportional to the transmembrane pressure TMP. The factors which alter with concentration and lead to loss of linearity are the viscosity of the feed solution and the resistance which is a sum of both membrane and fouling resistance.
The critical flux is the lowest flux across the membrane that exists with irreversible fouling on a filtration membrane. It is obtained when a deviation from linearity is observed in the TMP to flux relationship. The limiting flux is the highest flux obtained as a function of TMP applied.
A guard filter is needed to ensure particulates greater than 250 microns in diameter are removed so that there is no physical damage to the membrane.
From the slope of the water flux, the hydraulic membrane permeability Lp is estimated. The relationship to the hydraulic resistance, Rm is:
Lp = 1/μ*Rm
The symbol of μ is viscosity.
An aspect of all membrane operations is gradual flux decline. Ideally, the operator would like volumetric permeate flux rates to be constant and without variation. That is often not the case!
Various models have been suggested as a means to analyse and predict flux decline behaviour during the filtration of macromolecular solutions (Rai et al., 2006).
All of these models can be classified into three broad categories: (a) osmotic pressure controlled, (b) cake or gel layer controlled and (c) resistance in series models. According to the resistance in series model, flux decline is due to the combined effects of irreversible membrane fouling and reversible fouling (a.k.a. concentration polarization) on the membrane surface. Most irreversible loss is due to fouling which we discuss later.
Others have applied models to the experimental data of crossflow ultrafiltrations of water–oil emulsions and reported that the model proposed by Field et al. (1995) was the most accurate for predicting the permeate flux behaviour (Koltuniewicz et al., 2000; Arnot et al. 2000).
It is possible to improve permeate flux by applying a direct current electric field, about 800 V/m which influences and reduces the formation of the gel layer (Sarkar et al., 2008).
Classic Examples Of Diafiltration
 Antibiotic removal
A combination of ultrafiltration and diafiltration was used to remove antibiotics from milk. Antibiotics are given to cows suffering mastitis but the antibiotic usually appears in the milk (Kosikowski & Jimemez-Flores, 1985; 1987). Here, penicillin G was deliberately added in varying concentrations to raw milk to see if it could be cleared. The milk processed this way was concentrated 3 fold. Antibiotic free permeate from milk was added back to reconstitute the whole milk. Unfortunately, the flavour of this treated milk was flat but it did mean that the milk could also be used for cheese manufacture. It appears a viable alternative to the use of activated carbon which has been used to treat milk but with less success on the flavour although the antibiotic was removed.
 Removal Of Toxins
Aflatoxins often appear in milk and many approaches have been tried to remove them. A non-chemical method has often been sought in the dairy industry.
An ultrafiltration-diafiltration unit was used to remove aflatoxin M1 (AFM1) from raw whole, homogenised and acidified milk. The milk had been spiked with the aflatoxin to demonstrate the degree of separation that could be achieved. The degree of removal was influenced by the initial concentration ranging from 0.5 to 3.5 micrograms/litre.
The type of unit used was a Romicon pilot-scale hollow-fibre UF unit (Model HFLAB-5, Romicon, Woburn, MA) which had a membrane cartridge of about 4,500 cm2 filtering surface with a molecular weight cut-off of 10,000 daltons. It was operated with an inlet pressure of 1.7kg/cm2 and an outlet pressure of 0.7 kg/cm2. The concentrate was ultrafiltered at constant temperature at 55ºC to generate two-thirds (3:2) and one-third (3:1) of its original volume.
To ensure a clean solution was produced the 3:1 retentate was further washed using distilled water at the same temperature and ultrafiltered again.
 Processing of Gelatin
Ultrafiltration coupled to diafiltration is used to clean up animal and fish wastes for gelatin production for the food industry (Simon et al., 2002). In the conventional production process, after
acid or alkaline extraction from the raw materials, the gelatin broths are clarified, demineralised by ion-exchange and then concentrated in vacuum evaporators up to 25-35 wt% gelatin.
Dewatering costs are lower compared to conventional evaporation technology, there is lower thermal degradation and a higher purity level for the same concentration ratio because impurities are removed during the diafiltration phase (Chakravorty & Singh, 1990).
In addition, UF can offer further product improvement by the removal of salts from the gelatin liquors by operating in the diafiltration mode (Dutre and Tragårdh, 1995; Greenlaw et al.,1997) Desalting is usually achieved through diafiltration and monitored by checking conductivity.
Permeate fluxes are improved 8 fold by adjusting the pH in the case of gelatin to below pH 4. The isoelectric point (pI) of this protein is between 4 and 5. It’s likely that proteins at their pI find membrane fouling more easily. The interaction is ionic. It seems for good permeate flux, the pI of alumina is close to pH 8-9, but below pH 4 the membrane surface is positively charged as are the protein molecules. There is rejection of the protein from the surface. It is a good example of manipulating retentate conditions to help with concentration and diafiltration.
In the Simon paper, diafiltration was operated continuously where the permeate flux was compensated by an equal input of de-ionised water.
Assuming a constant transmission of the microsolute, Ti, through the membrane, the mass balance on the i-component:
- V*dCr,i/dt = Jv*S*Cp,i
predicts an exponential decrease in the concentration of i with respect to V*:
Cr,i = Co,i exp (-Ti * V*)
where Cr,i, Co,i and Cp,i are, respectively, the concentration of microsolutes in the retentate, in the feed and in the permeate, and V* is the diafiltration volume (ratio of volume of pure water added to initial volume of solution to be washed).
 Treatment of Emulsions such as Milk and Cream
The most common type of emulsion which is ultrafiltered and diafiltered is milk. Buttermilk is the fraction formed when butter is being churned. It is rich in lipids and proteins. These components are often separated by microfiltration and ultrafiltration in particular.
In large-scale MF/UF, buttermilk factions were generated using devices such as the Tetra Alcross Pilot M1 hollow-fibre system (Tetra Pak Filtrations Systems, Lund. Sweden). In one example, a 0.5 μm ceramic Membralox P35-37 membrane was used (Pall Corp., Mississauga, Ontario. Canada) with a 0.35 m2 filtration surface. The TMP was 100 kPa generating a distilled water flux rate of 2,900 L/hr.m2.
The same type of unit was used to prepare concentrated (Tetra Alcross Pilot M1) pasteurised skim milk with a 50 kDa MWCO (Meena et al., 2016; 2018). The skim milk fractions can be diafiltered against reverse osmosis water or salt solutions to increase the ionic strength of the fraction.
Halloumi cheese has been prepared from ultrafiltered goats milk (Deshwal et al., 2020). The membrane is conditioned using alkaline-chlorine washing and the Ultrasil 25 (0.01%, pH 13) and XY-12 (200 ppm Cl−) cleaning solution (Ecolab, Mississauga, Ontario, Canada).
Why proteins can be fractionated. In a model system, lactalbumin and lactoglobulin in a NaCl solution were separated using an Amicon stirred-cell with a 30-kDa cellulose membrane (Cheang and Zydney, 2003; 2004).
 Ultrafiltration and Diafiltration of Fruit Juices
Most fruit juices have been concentrated using ultrafiltration but diafiltration is not as common unless a specific component is required.
 Removal of Sugars
The impact of sugar with polyphenols on diafiltration has been explored for its effects on fouling (Wei et al., 2008).
The removal of sugars such as glucose and fructose from honey is feasible by diafiltration so that it can then be spray dried to a powder. It’s normally produced then by spray drying. The diafiltered honey had improved agglomeration properties (Samborska et al., 2017). The MWCO 15kDa membrane used was a ceramic tubular type.
In some cases sugar is removed from monoclonal antibody preparations prior to analysis (Lee et al., 2021) and also from beer (Sansome-Smith, private communication, 2011) for similar studies.
Water has been recovered from emulsions and oily wastewater using membranes of differing degrees of hydrophilicity. In some instances these are blends of polycarbonate with polyvinylidene fluoride.
Typical devices used for small-scale laboratory studies include the Sartocon Slice 200 bench-top system (Sartorius) which allows for a TMP of 1.1 bar and flow-rates around 250ml/min. Types of membrane to explore include regenerated cellulose and polyethersulphone with a 200 cm2 filter area. Some researchers cast their own membranes using polystyrene-block-poly(4-vinylpyridine) (PS-b-P4VP) isoporous diblock copolymer (PSBC), polyacrylonitrile (PAN) or polyvinylidene difluoride (PVDF) with 50% (m/m) titan dioxide.
With operation of these membrane systems, following each concentration cycle, a buffer with the same pH and conductivity as the feed is added to restore the feed starting volume.
Treatment of Membrane Fouling Materials
One of the issues of ultrafiltration and diafiltration, as with any filtration technology is fouling. Numerous examples occur of fouling due to proteins and carbohydrates physically clogging up pores as well as developing stable gel layers.
The main types of fouling are:-
 particulate deposition
- standard blocking where macromolecules are uniformly deposited on pore walls
- complete blocking where the membrane pore is totally blocked by macromolecules
- cake formation – accumulated particles and macromolecules form a fouling layer on the membrane surface.
- intermediate blocking – macromolecules deposit into pores or previously blocked pores and add to cake formation.
 Scale formation
Scale formation as in any application occurs because of concentration polarisation at the membrane surface. The ion concentrations of metal ions increases to such a point that it exceeds the solubility threshold of the salt in solution and precipitates on the membrane surface.
The presence of precipitated inorganic salts deposits blocks pores and stops permeate flux along with damage to the membrane and overall loss of performance.
Scale formation is dependent on solubility and concentration polarisation including permeation rate, velocity of the retentate across the membrane, pH, temperature and the types of chemical species in solution.
 Microbial stability: If the feedstock is not microbiologically stable then microorganisms can grow on the membrane surface to form a biofilm which is a gel. This film reduces resistance to flow and acts as a barrier to permeation.
The Impact of Various Forms of Fouling on TMP
Pectin is problematic in the membrane filtration of fruit juices. The technology has been used to clarify apple juice, passion fruit juice, citrus juices etc. One of the reasons it is preferred to dead-end filtration is the removal of diatomaceous earths which improves productivity and health sand safety issues. Pectin is also a gelling agent and in combination with sugar readily forms sugar-rich gels.
As well as seeing a reduction in transmembrane flow there is also an increase in transmembrane pressure (TMP). Studies on pectin solutions show that formation of gel layers is one of the biggest issues alongside particle interactions due to electrostatic forces (Sulaiman et al., 1991).
Fruit juices offer useful insights into the performance of membrane fluxes altered by concentration polarisation layers. Initially during the run, at low operating TMP, most juice components such as pectin, sugars, organic acid, etc. pass through the membrane pores. However, at higher TMPs, the permeate flux reaches a plateau due to a large accumulation of juice components in the polarization layer. This induces membrane fouling and increases the resistance to permeate flow (De Barros et al., 2003).
Reversible fouling which means it is easily removed by treatment is usually explained away because of the presence of pectin in a juice. However, in addition to pectin, polyphenolic compounds and proteins combine to form soluble complexes that also cause membrane fouling. This has been observed in the clarification of cranberry juice and pomegranate juice (Baklouti et al., 2012; Cassano et al., 2015).
The application of electric fields to reduce gel layer formation has already been mentioned but physical methods such as the use of turbulence promoters are also popular. It is also a good idea to adjust the pH and ionic strength of the feed solution to ensure the chemistry of solutes are optimised and so minimise membrane adsorption as well as deposition of precipitates.
Pre-filtration techniques included guard filters and other agents to remove large molecules. It’s a a good idea to alter processing conditions such as increasing transmembrane pressure to maximise flux.
The membranes themselves can be modified to try and reduce the fouling layers. Polysulphone membranes are modified by adding polymeric additives such as polyvinylpyrrolidone (PVP), polyetherimide (PEI), polyethylene glycol (PEG) and polyethersulfone (PES). These are most effective in the sepration of oil and water mixtures. An addition of an additive here optimises oil retention whilst preventing a concentration polarisation phenomenon from developing (Pagidi et al., 2014).
Cleaning needs to be conducted regularly to prevent the accumulation of fouling material. It is done between batch runs. Some operators conduct regular and periodic backwashing every 10 or 15 minutes to dislodge cake layers formed on the membrane surface.
Fouling from pectin is restored by circulating a caustic solution of 0.5% NaOH solution for 1 hour. A small amount of polyethylene oxide (0.003%) is added to improve cleaning times further. It is thought the polymer possible sours the membrane but most likely forms a complex with pectin which removes it from the membrane. This is rinsed from the membrane by the turbulent action of the pumping system (Tzeng & Zall, 1990).
Another approach is to add enzyme which degrade proteins and carbohydrates. For example, it was claimed that ultrafiltration of an apple juice containing pectin could be improved if the pectins were hydrolysed using pectinesterases (Echavarria et al., 2011). This might not be beneficial for those developers aiming for a soluble fibre claim but it might considered a treatment. A pectinesterase solution in its own right has been used to treat pectin fouled membranes after their use for ultrafiltering fruit juices (Sansome-Smith, 1992). In one instance it is feasible to immobilise a hydrolysing enzyme to the membrane to encourage depectinisation on each occasion (Echavarria et al., 2012).
Nowadays, a typical commercial enzyme cocktail used for the manufacture of fruit juices contains the three main but different pectinases: pectinlyase, polygalacturonase and pectinesterase to make pectic complexes soluble to complete the sedimentation and clarification of the juice irrespective of whether a membrane concentration process is required or not.
Another approach is to reversibly direct permeate flow back across the membrane so that it forces accumulated layers off the membrane.
Suppliers of Membranes, Ultrafiltration and Diafiltration Kits
- Alfa Laval (ex. DDS/DSS) Denmark- DSS GR 61 membranes for diafiltration and ultrafiltration used in food and pharmaceutical processing. Most flat sheets rely on a polypropylene support material with good pH and temperature range. Polysulphone is one of their most popular types The pore sizes range from 1,000 MWCO up to 100,000 MWCO. A number of systems available as plate & frame systems.
- China Blue Star Membrane Technologies, Beijing, China – hollow fibre membranes.
- Dow Chemical Co., Minneapolis, MN – NF200 membranes for nanofiltration.
- DuPont – ultrafiltration membranes for water purification and water recovery from waste streams (FilmTec™ RO-390).
- GE-Desal Osmonics (part of SUEZ), Herentals, Belgium & Minnetonka, MN, USA – produce equipment for their membranes – Sepa® CF II Membrane Cell System: manufacturers of nanofiltration membranes such as the DS-5 DL. This membrane is a thin-film composite (TFC) with a polyamide surface on a polysulfone support and polyester matrix. Its cut-off ranges is between 150 and 300 Da; 98% of lactose retention was measured at 22ºC for a feed solution containing 50 g/L of lactose concentration (Cuartas-Uribe et al., 2009).
- Graver Technologies (USA) -stainless steel tubular, Scepter metallic membranes.
- Lenntech – do SUEZ membranes for water desalination. Desal GK membranes for nanofiltration.
- Merck [Amicon (Lexington, MA) Hollow-fibre ultrafiltration units are useful for small-scale to larger-scale studies below pilot-plant. A good model that was available as the bench standard was a Model CH4. The hollow-fibre cartridges had the brand name DIAFLO® e.g. H1P30-20 and had molecular weight cut-offs from 10 kDaltons through to 30kDaltons MWCO.
- Merck Millipore (Darmstadt, Germany) produce lab-scale ultrafiltration membranes such as the 30kD Ultracel® disc-type for lab-scale work. The diafiltration units can be run using a Pellicon 3 Cassette system with Ultracel membrane filters.
- Microdyn-NADIR Filtration (Wiesbaden, Germany) – membranes especially polyethersulphone and cellulose acetate in a hollow fibre set-up. PES 10 membrane with a MWCO of 1000 Da for nanofiltration
- Pall systems – solid all-round performer with a range of membrane systems.
- PNAP Process Kersep ceramic UF systems – tubular
- Paterson Candy International (PCI) – no longer operating but some machines for scale-up work still exist.
- Romicon – a pilot-scale hollow-fibre UF unit
- SANI Membranes
- Sartorius – Sartocon® single-use filtration systems for concentrating and diafiltering extremely sensitive molecules including recombinant proteins, toxic ADC production and vaccines.
- SCT (France) – suppliers of Al2O3 membranes branded as MEMBRALOX® Tl-70 with a nominal MWCO of 10kDa.
- Steriltech, Auburn/Kent, WA, USA – producers of crossflow tangential filtration equipment for large-scale and pilot-scale use. Rely on flat membrane systems. Good example is the HP4750 for ultrafiltration.
- Synder Filtration Inc. Vacaville, CA, USA – producers of flat-sheet UF membranes.
- Tami Industries, Nyons, France. – ceramic tubular membranes from 100 to 300 kDa cut-off
- Tetra Pak (Tech-Sep (France)) – a good pilot scale unit is the Tetra Alcross M1 which is used for the preparation for large volumes of diafiltered protein solutions
- TIA, France – suppliers of small-scale ultrafiltration units with a tubular configuration.
A good laboratory membrane system would consist of a minifilter made up of a support plate with a stem, which holds the membrane module with a feed and retentate outlet in splined ends, a lower and upper hydraulic box to feed the product to be treated and collect the concentrated product (outlet), and inlet and outlet pressure gauges, two pressure gauges with separators with their gaskets and collars, a valve to regulate the trans-membrane flow pressure in the 350–400 kPa range and a housing for a tubular membranes or a plate for flat membranes.
All membrane need to be in compliance with EU Regulation (EC)
1935/2004, EU Regulation 10/2011, EU Regulation (EC)
2023/2006 and FDA regulations (CFR) Title 21
The economic operation of a diafiltration unit has been explored (Sharma et al., 2017).
Arnot, T. C., Field, R. W., & Koltuniewicz, A. B. (2000). Cross-flow and dead-end microfiltration of oily-water emulsions: Part II. Mechanisms and modelling of flux decline. Journal of Membrane Science, 169(1), pp. 1-15.
Baklouti, S., Ellouze-Ghorbel, R., Mokni, A. (2012) Clarification of pomegranate juice by ultrafiltration: Study of juice quality and of
the fouling mechanism. Fruits, 67, pp. 215–225
Cassano, A.; Conidi, C.; Tasselli, F. (2015) Clarification of pomegranate juice (Punica granatum L.) by hollow fibre membranes: Analyses of membrane fouling and performance. J. Chem. Technol. Biotechnol. 90, pp. 859–866 (Article).
Chakravorty, B. and Singh, D.P. (1990) Concentration and purification of gelatin liquor by ultrafiltration. Desalination, 78 pp.279-286
Cheang, B., & Zydney, A. L. (2003). Separation of α‐lactalbumin and β‐lactoglobulin using membrane ultrafiltration. Biotechnology and Bioengineering, 83(2), pp. 201-209.
___________________(2004). A two-stage ultrafiltration process for fractionation of whey protein isolate. Journal of Membrane Science, 231(1-2), pp. 159-167
Cheryan, M. (1998) Ultrafiltration and Microfiltration Handbook
Technomic Pub. Co, Lancaster, p. 527
Cuartas-Uribe, B., Alcaina-Miranda, M. I., Soriano-Costa, E., Mendoza-Roca, J. A., Iborra-Clar, M. I., & Lora-García, J. (2009). A study of the separation of lactose from whey ultrafiltration permeate using nanofiltration. Desalination, 241(1-3), pp. 244-255.
De Barros, S., Andrade, C., Mendes, E., Peres, L. (2003) Study of fouling mechanism in pineapple juice clarification by ultrafiltration. J. Membr. Sci. 215, pp. 213–224 .
Deshwal, G. K., Ameta, R., Sharma, H., Singh, A. K., Panjagari, N. R., & Baria, B. (2020). Effect of ultrafiltration and fat content on chemical, functional, textural and sensory characteristics of goat milk-based Halloumi type cheese. LWT, 126, 109341 (Article).
Domingues, R. C. C., Ramos, A. A., Cardoso, V. L., & Reis, M. H. M. (2014). Microfiltration of passion fruit juice using hollow fibre membranes and evaluation of fouling mechanisms. Journal of Food Engineering, 121, pp. 73-79.
Dutré, B. and G. Tragårdh, G. (1995) Purification ofgelatin with
a forced solvent stream along the membrane permeate side: an experimental approach, J. Food Engineering, 25 pp. 233-244.
Echavarría, A., Pagán, J., & Ibarz, A. (2011). Effect of previous enzymatic recirculation treatment through a tubular ceramic membrane on ultrafiltration of model solution and apple juice. Journal of Food Engineering, 102(4), pp. 334-339 (Article)
Echavarría, A. P., Torras, C., Pagán, J., & Ibarz, A. (2011). Fruit juice processing and membrane technology application. Food Engineering Reviews, 3(3-4), pp. 136-158.
Echavarría, A. P., García-Valls, R., Torras, C., Pagan, J., & Ibarz, A. (2012). Effect of pectinase immobilization in a polymeric membrane on ultrafiltration of fluid foods. Separation Science and Technology, 47(6), pp. 796-801. (Article)
Jönsson, A. S., & Trägårdh, G. (1990). Ultrafiltration applications. Desalination, 77, pp. 135-179 (Article).
Greenlaw, D.B., Flinn, C., Skelton, R. (1997) The use of spiral wound ultrafiltration technology in gelatin processing- an overview. Imaging Sci. J., 45 pp. 256-259.
Koltuniewicz, A. B., Field, R. W., & Arnot, T. C. (1995). Cross-flow and dead-end microfiltration of oily-water emulsion. Part I: Experimental study and analysis of flux decline. Journal of Membrane Science, 102, pp. 193-207.
Kuriyel, R. (2000) Ultrafiltration, Kirk-Othmer Encyclopedia of Chemical Technology. John Wiley & Sons, Inc, New Jersey (2000), pp. 1–22
Lee, J., Na, J., & Baek, Y. (2021). Effects of impurities from sugar excipient on filtrate flux during ultrafiltration and diafiltration process. Membranes, 11(10), pp. 775.
Madsen, R.F. (2001) Design of sanitary and sterile UF- and diafiltration plants. Sep. Purif. Technol. 22–23 pp. 79–87
Meena, G.S., Singh, A.K., Borad, S., Raju, P.N. (2016) Effect of concentration, homogenization and stabilizing salts on heat stability and rheological properties of cow skim milk ultrafiltered retentate. J. Food Sci. Technol. 53(11) pp. 3960–3968. doi: 10.1007/s13197-016-2388-4. (Article).
Meena, G. S., Singh, A. K., Gupta, V. K., Borad, S., Arora, S., & Tomar, S. K. (2018). Effect of pH adjustment, homogenization and diafiltration on physicochemical, reconstitution, functional and rheological properties of medium protein milk protein concentrates (MPC70). Journal of Food Science and Technology, 55(4), pp.1376-1386.
Pagidi, A., Saranya, R., Arthanareeswaran, G., Ismail, A. F., & Matsuura, T. (2014). Enhanced oil–water separation using polysulfone membranes modified with polymeric additives. Desalination, 344, pp. 280-288.
Pal, S., Bharihoke, R., Chakraborty, S., Ghatak, S. K., De, S., & DasGupta, S. (2008). An experimental and theoretical analysis of turbulence promoter assisted ultrafiltration of synthetic fruit juice. Separation and Purification Technology, 62(3), pp. 659-667 (Article).
Perreault, V., Gouin, N., Bérubé, A., Villeneuve, W., Pouliot, Y., & Doyen, A. (2021). Effect of Pectinolytic Enzyme Pretreatment on the Clarification of Cranberry Juice by Ultrafiltration. Membranes, 11(1), pp. 55.
Rai, P., Majumdar, G., Dasgupta, S., Deb, S. (2005a) Understanding
ultrafiltration performance with mosambi juice in an unstirred
batch cell. J Food Eng. 28(2) pp. 166–180
of flux decline of depectinized mosambi juice using unstirred
batch ultrafiltration. J Food Process Eng. 28 pp. 359–377
______________________________________ (2007) Modeling of
permeate flux of synthetic fruit juice and mosambi juice (Citrus
sinensis) in stirred continuous ultrafiltration. LWT 40 pp. 1765–
Rai, P., Raia, C., Majumdara, G., DasGupta, S., Deb, S. (2006)
Resistance in series model for ultrafiltration of mosambi (Citrus
sinensis (L.) Osbeck) juice in a stirred continuous mode. J Memb. Sci 283 pp. 116–122
Samborska, K., Sokołowska, P., & Szulc, K. (2017). Diafiltration and agglomeration as methods to improve the properties of honey powder obtained by spray drying. Innovative Food Science & Emerging Technologies, 39, pp. 33-41.
Sansome-Smith, A.W. (1992) Fruit juice membrane studies. Unpublished communication. SmithKline Beecham. Pilot Plant 1992 Report.
Sarkar, B., DasGupta, S., & De, S. (2008). Effect of electric field during gel-layer controlled ultrafiltration of synthetic and fruit juice. Journal of Membrane Science, 307(2), pp. 268-276.
Simon, A., Vandanjon, L., Levesque, G., & Bourseau, P. (2002). Concentration and desalination of fish gelatin by ultrafiltration and continuous diafiltration processes. Desalination, 144(1-3), pp. 313-318.
Sulaiman, M. Z., Sulaiman, N. M., & Shamel, M. (2001). Ultrafiltration studies on solutions of pectin, glucose and their mixtures in a pilot scale crossflow membrane unit. Chemical Engineering Journal, 84(3), pp. 557-563.
Tzeng, W.C., Zall, R.R. (1990) Polymers decrease cleaning time of an ultrafiltration membrane fouled with pectin. J. Food Sci., May (Article).
Urošević, T., Povrenović, D., Vukosavljević, P., Urošević, I., & Stevanović, S. (2017). Recent developments in microfiltration and ultrafiltration of fruit juices. Food and Bioproducts Processing, 106, pp. 147-161
Wei, D. S., Hossain, M., & Saleh, Z. S. (2008). Separation of polyphenolics and sugar by ultrafiltration: Effects of operating conditions on fouling and diafiltration. International Journal of Chemical and Biomolecular Engineering, 1(1), pp. 14-23