Journal of Membrane Science 442 (2013) 177–186

Contents lists available at SciVerse ScienceDirect

Journal of Membrane Science

journal homepage: www.elsevier.com/locate/memsci

Separation of polyphenols and proteins from flaxseed hull extracts

by coagulation and ultrafiltration
M. Loginov n, N. Boussetta, N. Lebovka, E. Vorobiev
Unité Transformations Intégrées de la MatièreRenouvelable, Université de Technologie de Compiègne, Centre de Recherche de Royallieu, B.P. 20529-60205
CompiègneCedex, France

art ic l e i nf o a b s t r a c t

Article history: This work is devoted to the separation of polyphenols and proteins, extracted from flaxseed hulls. The
Received 26 February 2013 separation comprised two steps: (i) coagulation of proteins by adjustment of pH and subsequent
Received in revised form centrifugation, and (ii) ultrafiltration of supernatants. The value of pH, used for extract coagulation,
12 April 2013
varied from 13.2 to 3.0. Ultrafiltration experiments were carried out in the concentration mode (up to the
Accepted 13 April 2013
final volume reduction factor of 7.5) using a batch dead-end stirred filtration cell and hydrophilic
Available online 24 April 2013
polyethersulfone membranes with nominal molecular weight cut-off 30 kDa. The impact of pH on the
Keywords: efficiency of protein coagulation and filtration steps of polyphenol purification was investigated and
Flaxseed hull optimal conditions of purification were determined. The purity of polyphenols in filtrate increased from
Extraction
33.6% to 56.0% after coagulation and centrifugation at pH¼4.4, and it was additionally increased to 76.6%
Separation of polyphenols/proteins
after the ultrafiltration. The impact of pH on filtration rate decline and rejection of polyphenols and
Ultrafiltration
Fouling control proteins was demonstrated. Filtration rate decreased and filtrate purity increased as pH decreased from
13.2 to 4.4. The decrease of filtration rate at lower values of pH was explained by formation of a less
permeable filter cake of residual proteins. It was demonstrated that filter cake formation significantly
increased the selectivity and resulted in a better filtrate quality.
& 2013 Elsevier B.V. All rights reserved.

1. Introduction field treatment [2], use of microwave-assisted extraction [11],

ohmic heating [12], and low polarity water extraction [13]. How-
Membrane filtration is a powerful tool for clarification, purifi- ever, along with polyphenols, the obtained extracts usually contain
cation and concentration of various solutions (e.g., juices, extracts, other valuable compounds, such as proteins and polysaccharides.
whey) and separation of valuable compounds from by-products of Traditional method of polyphenol purification comprises coa-
food industry and agriculture. Nowadays, these by-products are gulation and precipitation of impurities (use of chemicals, sol-
considered as a perspective source of high-added value nutrient vents), adsorption of polyphenols on resins, elution of purified
compounds (proteins, polyphenols, fibers, etc.) [1]. Particular polyphenols, and their concentration by solvent evaporation [14].
attention is paid to the extraction of polyphenols from various These operations account for significant part of the total costs of
renewable sources, considered as by-products or wastes: grape polyphenol production [14].
seeds [2], grape pomace [3–6], flaxseed by-products [7] and olive During the last decade, the possibility of purification and
mill wastewater [1]. The elaboration of economically efficient concentration of polyphenol containing extracts by ultra- and
methods of polyphenol recovery from these resources is very nanofiltration was demonstrated [3,9,15–18]. However, compari-
attractive because of a high nutritive value and relatively high cost son and systematization of the results, reported in different
of polyphenols. studies for different extracts, is rather difficult. For example, it
The optimization of the extraction of polyphenols from by- was reported that filtration of different extracts and juices through
products was intensively studied, and a number of methods was the membranes with molecular weight cut offs (MWCO) of 5–
proposed for increasing the polyphenol yield, reducing the che- 70 kDa increases the purity of polyphenols by approximately 20%
micals and energy consumptions, and reducing extraction dura- at the low level of polyphenols rejection (less than 10%) [15–18].
tion: optimization of traditional solvent extraction [8–10], The other studies have shown that phenolic compounds may be
application of high voltage electrical discharges and pulsed electric concentrated by the filtration of extracts through membranes with
MWCO ranging from 1 kDa to 20–150 kDa [19–21].Vaillant et al.
n
[22] and Mirsaeedghazi et al. [23] have demonstrated that sig-
Corresponding author. Tel.: +33 3 065 25 25 227.
E-mail addresses: maksym.loginov@gmail.com,
nificant part of valuable polyphenols may be retained in filtration
maksym.loginov@utc.fr (M. Loginov). cell even during microfiltration (data for pomegranate and melon

0376-7388/$ - see front matter & 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.memsci.2013.04.036
178 M. Loginov et al. / Journal of Membrane Science 442 (2013) 177–186

juices). At the same time, it was reported that microfiltration of [29,30], especially for extracts of flax by-products. It was reported
pineapple juice did not result in rejection of polyphenols [24]. that filtration of flax shive extract through the membrane with
Such difference in MWCO values, required for retention of poly- 30 kDa resulted in the removal of impurities [7]; however, the
phenols, can be the result of difference in the molecular weight details of filtration experiment and filtration mechanism were not
and aggregation state of polyphenols in extracts or interaction of discussed.
polyphenols with fouled membrane or filter cake [25,27,31,33]. It may be concluded that ultrafiltration is a promising method
The type and the quantity of impurities in the extract also may for purification of polyphenols from the other solutes, contained
affect the filtration mechanism and retention of polyphenols. by extracts and juices, and purification efficiency is determined by
Filtration rate is negatively influenced by all the components of the operating conditions and extract quality. Optimal conditions of
natural extracts (proteins, polyphenols, pectins, and polysacchar- purification may vary for extracts of different origin, therefore,
ides) [25–31,36]. Different fouling mechanisms (adsorption of mechanism of filtration and solute rejection must be determined
solutes by membrane, internal pore blocking, cake formation and in every particular case in order to optimize the efficiency of
compression) may be responsible for the permeate flux decline purification.
during microfiltration of extracts [25–27,31,36]. Formation of The present work is devoted to the separation of proteins and
polyphenol–protein and polyphenol–polysaccharide aggregates in polyphenols, extracted from the flax hulls. Separation was per-
the real extracts results in more severe membrane fouling and formed in two steps: (i) coagulation of proteins at different values
sharper permeate flux decline as compared to pure solutions of of pH and subsequent centrifugation, and (ii) ultrafiltration of
proteins and polyphenols [29,30,36]. It is interesting to note that supernatants through the polyethersulfone membrane. The aim of
removal of aggregates (e.g., by centrifugation) may have little this work is to compare the efficiency of coagulation and ultra-
influence on the improvement of microfiltration rate of some filtration for polyphenols purification, to determine filtration and
complex extracts [31]. This may be attributed to the significance of solute rejection mechanisms, and to determine the optimal con-
irreversible membrane fouling (adsorption, pore blocking) by ditions of extract purification.
residual solutes [25,31]. It was shown by Ulbricht et al. [35],
Vernhet and Moutounet [36], El Rayess et al. [25], and Susanto
et al. [34] that adsorption of each component of the extract 2. Materials and methods
(especially, polyphenols) during the initial stage of filtration
reduces the membrane permeability, thus increasing rejection of 2.1. Raw material
the solutes. Preliminary removal of solutes from extracts (e.g.,
removal of polyphenols by adsorption) significantly improves the Flaxseed (Linumusitatissimum, cultivar Baladin) hulls (seed
rates of microfiltration [25] and ultrafiltration [28]. However, this teguments) were provided by Lasalle Beauvais (Beauvais, France).
method is inappropriate when filtration of extract is aimed at The dry matter content of dried seeds was 96.3%. Flaxseed hulls
purification of polyphenols. were crushed for 40 s in a laboratory coffee grinder (SEB, Ecully,
Several studies have demonstrated the possibility of improve- France) (180 W). The powdered samples were sieved to select
ment of the filtration rate and reduction of polyphenol rejection by particles smaller than 1 mm.
reasonable choice of filtration parameters [5,25]. Also, it was
shown that adjustment of pH prior to ultrafiltration of 2.2. Extraction
polyphenol-containing grape must extracts [32,34] and model
polyphenol solutions may be used as a tool for reduction of the Preliminary extraction experiments were carried out at differ-
fouling resistance and improvement of transmission of polyphe- ent compositions of extracting medium (water to ethanol ratio and
nols to the filtrate. The positive effect of pH adjustment on the concentration of sodium hydroxide) in order to estimate the
filterability was explained by the reduction of polyphenol adsorp- impact of extracting medium on the extract quality (concentration
tion on the membrane surface in an alkaline medium and of proteins and polyphenols and extract viscosity).
corresponding reduction of membrane fouling [34]. The grinded flaxseed hulls (10 g) were added to a mixture of
Much more information is available regarding the impact of water and ethanol (200 mL). The solid to liquid ratio was 20. The
proteins on the membrane filterability and rejection of solutes water–ethanol mixtures with different concentrations of ethanol
during the micro- and ultrafiltrations [26,37–46]. Filtrate flux (0, 50 and 70 wt%) were supplemented by 0–0.3 M sodium
decline during filtration of proteins is explained by adsorption of hydroxide (NaOH) for alkaline extraction. The suspension was
proteins on the membrane surface and pore blocking by protein introduced in a beaker under agitation at 100 rpm for 2 h. The
aggregate, with the following formation of the filter cake [26,37– temperature of the mixture was 20 1C. Then the extract was
39,41,46,53,54]. The intensity of adsorption and pore blocking, as separated from the exhausted hulls by centrifugation (5 min at
well as the specific cake resistance, are determined by the value of 4000 rpm). The obtained supernatant was frozen and stored at
pH of the protein solution: this value increases and filterability of −5 1C before purification experiments.
the model protein solution decreases as pH approaches the The properties of extracts, obtained from extracting media with
isoelectric point of protein [55–58]. It was explained by increasing different compositions, are listed in Table 1.
attraction between uncharged proteins and membrane surface and
increase of compactness of the filter cakes formed from uncharged Table 1
proteins [37–39,41,43,44,46,55–58]. The membrane fouling also Concentration of proteins and polyphenols and purity of extracts, obtained at
has negative impact on transmission of proteins to the filtrate. different concentrations of ethanol and NaOH in preliminary extraction
Van Reis et al. [47] have demonstrated that adjustment of pH of experiments.
mixed protein solutions may be used for the regulation of filter-
Ethanol (wt%) NaOH (M) Polyphenols (g/l) Proteins (g/l) Purity (%)
ability and transmission of proteins to the filtrate, and it is an
effective method for improvement of the filtration-separation of 0 0.3 2.5 3.0 45
proteins with similar molecular weights and different isoelectric 50 0.3 2.8 3.5 44
points. However, the impact of operational parameters and pH of 50 0 1.0 0.7 61
50 0.1 2.6 3.9 40
extract on the filtration rate and efficiency of separation of the 70 0.1 2.5 2.8 47
polyphenols and proteins is rarely discussed in the literature
 M. Loginov et al. / Journal of Membrane Science 442 (2013) 177–186 179

Extraction of flaxseed hulls in a non-alkalinized ethanol solu- Filtrate flux, J, and the values of clean membrane resistance
tion (0 M NaOH, 50 wt% ethanol) resulted in low concentration of (rm), membrane resistance during filtration of centrifugate (r), and
polyphenols and proteins. Aqueous extraction (0.3 M NaOH, 0 wt% fouled membrane resistance (rf) were determined using equation
ethanol) resulted in sufficiently high yield of proteins and poly-
phenols. However, this extract was very viscous due to simulta- J ¼ −dV=ðAdtÞ ¼ Δp=ðmrÞ ð1Þ
neous extraction of water-soluble polysaccharides (mucilage). The
where J is the filtrate flux, V is the volume of filtrate, t is the
following composition of extracting medium was chosen for
filtration time, A is the membrane surface area, Δp is the filtration
filtration-purification experiments: 50 wt% of ethanol, 0.1 M NaOH
pressure and μ is the viscosity of filtrate.
(since it yielded sufficiently high concentration of proteins and
In each experiment, 9 g of the supernatant was used and 7.8 g
polyphenols, and low viscosity of extract at moderate consump-
of filtrate was obtained during filtration. The weight of filtrate was
tion of ethanol and NaOH).
continuously registered by means of electronic balance. The
For the purification study, extractions (liquid to solid ratio of
experiments were carried out in the concentration mode with
10) were performed with higher amount of flax product in order to
removing of the filtrate. The volume reduction factor, VRF, was
have homogenous samples for all the experiments. The suspension
calculated as
was introduced into a beaker under agitation at 100 rpm for 2 h.
The temperature of the mixture was 20 1C. The extract was then VRF ¼ V o =ðV o −VÞ ð2Þ
separated from the exhausted hulls by centrifugation (5 min at
4000 rpm). The obtained supernatant was frozen and stored at where Vo is the initial volume. The final volume concentration
−5 1C before purification experiments. factor was equal to 7.5. Filtrate samples with the weight of 1.3 g
The composition of extract, obtained at optimal extraction were collected and analyzed separately. Filtration experiments
conditions, was the following: polyphenol concentration were carried out at the constant pressure of 4  105 Pa and the
Cph,o ¼(2.98 7 0.05) g/l, protein concentration Cpr,o ¼(6.09 70.36) ambient temperature of 20 1C.
g/l, purity of polyphenols in extract Po ¼(32.9 70.9)%. Several experiments (they are referred to below as unstirred
dead-end filtration) were carried out without stirring of retentate
2.3. Purification of extract for determination of specific resistance of the filter cake. The
simplified form of Ruth equation was used for estimation of the
Purification of flaxseed hull extract comprised two steps: filter-cake resistance in unstirred filtration experiments [48]
(i) coagulation of proteins by pH adjustment with a following
t=V ¼ αcρmV=ð2A2 ΔpÞ þ r m m=ðAΔpÞ ð3Þ
centrifugation, and (ii) stirred dead-end ultrafiltration of
supernatants. were α is the specific filtration resistance of the filter-cake, c is the
weight fraction of cake-forming solute in the supernatant, rm is the
2.3.1. First step: coagulation and centrifugation membrane resistance, Δp is the filtration pressure, ρ and μ are the
Initial pH of flaxseed hull extract was equal to 13.2. It was density and the viscosity of filtrate, respectively, A is the filter
adjusted by slow addition of concentrated hydrochloric acid under medium surface area, t and V are the filtration time and the filtrate
the continuous stirring. The following values of pH were obtained: volume, respectively. The filter cake resistance is proportional to
13.2, 9.0, 7.0, 4.4, and 3.0. After pH adjustment, the extracts were the value of αc, which may serves as a measure of membrane
stirred during 20 min at room temperature in order to attain fouling due to deposit formation (the higher is the value of αc, the
maximal coagulation of proteins. Then extracts were centrifuged higher is the fouling).
during 10 min at 2000 rpm. Centrifugation at these conditions In some experiments (both stirred and unstirred), the inter-
resulted in complete sedimentation of suspended particles and rupted filtration was applied for estimation of fouling reversibility.
clarification of supernatant for all the studied values of pH After 1.0 g of filtrate was obtained, filtration pressure was imme-
(pH¼ 13.2–3.0). Chemical composition of supernatants was deter- diately decreased from 4  105 Pa to 0 Pa for the period of 60 s.
mined, and supernatants were subjected to ultrafiltration. In some During this period retentate was stirred with the help of magnetic
experiments, pH of supernatant was adjusted from 4.4 to 7.0 prior mixer in order to remove the filter cake from the membrane
to ultrafiltration with the help of the concentrated NaOH solution. surface. Then, filtration pressure increased to 4  105 Pa and
filtration was continued (with or without stirring in stirred and
unstirred experiments, respectively).
2.3.2. Second step: dead-end ultrafiltration
Filtration of the supernatants was carried out in the batch
dead-end filtration plastic cell with maximal capacity of 15 ml and
effective filtration surface area A¼3.1  10−4 m2. Filtration purifi- 2.4. Analysis of extracts, supernatants and filtrates
cation experiments were carried with constant stirring of the
retentate by means of magnetic stirrer, fixed over the membrane 2.4.1. Total polyphenol content
surface and rotating at the constant rate of ω¼ 500 rpm (these Amount of total polyphenols was assayed colorimetrically by
experiments are referred to below as the stirred dead-end means of Folin–Ciocalteu method, based on oxidation/reduction
ultrafiltration). reactions of phenols [49]. The portions of 0.2 ml of diluted extract
A hydrophilic polyethersulfone ultrafiltration membrane and 1 ml of ten-fold diluted Folin–Ciocalteu reagent (Sigma-
(Microdyn-Nadir GmbH, Germany) with nominal molecular Aldrich, St-Quentin Fallavier, France) were mixed. Then, 0.8 ml of
weight cut-offs (MWCO) of 30 kDa was used for filtration. New Na2CO3 (75 g/l) (VWR, Fontenay-sous-Bois, France) was added. The
membrane was used for every filtration test. sample was incubated for 10 min at 50 1C and then cooled at room
Prior to filtration, the membranes were washed by distilled temperature. For the control sample, 0.2 ml of phosphate citrate
water, and then filtration of supernatant was started. After the buffer (0.2 M of Na2HPO4, 0.1 M of citric acid) at pH ¼ 4.0 was
filtration, the filter-cake and retentate were removed from the taken. The absorbance was measured at 750 nm by the UV/vis
membrane surface by rinsing, and distilled water was filtered spectrophotometer (Libra S32, Biochrom, Lagny-sur-Marne,
through the fouled membrane for the determination of fouled France). Gallic acid (Sigma-Aldrich, St-Quentin Fallavier, France)
membrane resistance. was used for the calibration curve.
180 M. Loginov et al. / Journal of Membrane Science 442 (2013) 177–186

2.4.2. Total protein content Centrifugation of extract without pH modification (pH¼ 13.2),
The concentration of proteins was determined by means of practically, had no influence on its chemical composition (Rph and
Bradford method [50]. A volume of 0.2 ml of diluted extract and Rpr≈0 at pH ¼13.2). Decrease of pH from 13.2 to 9.0 resulted in
1.8 ml of three-fold diluted Bradford Reagent (Sigma-Aldrich, St- formation of aggregates that were precipitating during centrifuga-
Quentin Fallavier, France) were mixed. The sample was kept for tion. The relative removal of proteins increased with decrease of
5 min at room temperature. The absorbance was measured at pH (enhancing of aggregation) and reached its maximal value of
595 nm by the UV/vis spectrophotometer (Libra S32, Biochrom, Rpr≈0.67 at pH ¼4.4. This value of pH corresponded to the value of
Lagny-sur-Marne, France). Bovine Serum Albumin (Sigma-Aldrich, isoelectric point, previously reported for flax seed proteins
St-Quentin Fallavier, France) was used for the calibration curve. (pH¼4.4) [8]. Coagulation was most efficient in this point due to
the absence of electrical repulsion between uncharged protein
2.5. Estimation of purification efficiency and data analysis molecules. However, even in isoelectric point, centrifugation of
extracts did not result in complete removal of proteins (Rpr o1).
Concentrations of polyphenols, Cph, and proteins, Cpr, were This may be explained by the presence of proteins with different
measured in extracts, supernatants and filtrate samples. The purity values of isoelectric point in the extracts.
towards polyphenols, P, was estimated as Further reduction of pH (from pH ¼4.4 to 3.0) provoked a sharp
decrease of Rpr (Fig. 1a). This may be explained by recharging of
P ¼ C ph =ðC ph þ C pr Þ ð4Þ amphoteric proteins in the acidic medium, which increased their
The purity towards proteins was estimated as electrical repulsion.
The pH adjustment also resulted in losses of polyphenols
P* ¼ 1–P ð5Þ during the centrifugation (Fig. 1a). The relative removal of poly-
Efficiency of purification of the supernatants and filtrates was phenols increased from Rph ≈0 to Rph ≈0.12 as pH was decreasing
estimated using the value of relative removal of proteins, Rpr, and from 13.2 to 9.0, and insignificantly increased with further
polyphenols,Rph decrease of pH. Polyphenol losses during the centrifugation of
acidified extracts may be explained by the impact of pH on their
Rpr ¼ C pr =C pr;o ð6Þ dissociation degree. According to the literature, polyphenols are
weakly acidic compounds (with pKa ∼8–10) and form anions in
Rph ¼ C ph =C ph;o ð7Þ
strongly basic medium, while they are neutral in acidic medium
where Cpr,o and Cph,o are the initial concentrations of polyphenols [51]. Decrease of pH suppresses their dissociation that can provoke
and proteins in the untreated extract, respectively. their hydrophobic interaction with proteins present in the extract.
Coefficients of rejection of proteins, RCpr, and polyphenols, RCph, Therefore, partial removal of polyphenols during the centrifuga-
were estimated at different values of VRF using the equation tion of coagulated extracts (at pH ≤9.0) may be explained by their
adsorption on the surface of protein aggregates. Nevertheless,
RC ¼ 1–C f =C r ð8Þ
impact of pH on the relative removal of polyphenols during the
where Cf and Cr are solute (polyphenols or proteins) concentra- centrifugation was significantly lower as compared to the relative
tions in the filtrate and retentate, respectively. The values of removal of proteins (Fig. 1a). It resulted in the increase of super-
protein and polyphenol concentrations in the filtrate (Cf) were natant purity from P ≈34% to 56% with decrease of pH from 13.2 to
measured continuously. The values of proteins and polyphenols 4.4 (Fig. 1b, open symbols).
concentration in the retentate were estimated as The dead-end stirred ultrafiltration of supernatants resulted in
additional removal of proteins and polyphenols at all the studied
C r ¼ ðC o V o −ΣC f ΔV Þ=ðV o –VÞ ð9Þ
values of pH (Fig. 1a, solid symbols). The relative contribution of
where Co is the initial concentration of polyphenols (or proteins) filtration into removal of proteins decreased as pH decreased from
in unfiltered supernatant. 13.2 to 4.4. The application of filtration step resulted in a notable
Transmission coefficients of polyphenols, TCph, or proteins, TCpr, supplementary purification of polyphenols: the maximal average
were estimated as purity of filtrate (P≈77% at pH ¼ 4.4) was significantly higher as
compared to the maximal purity obtained for supernatant (P≈56%
TC ¼ 1–RC ð10Þ
at pH ¼4.4).
The selectivity, S, of filtration-purification was calculated as The variation of filtrate purity, P, during filtration of super-
natants was insignificant: P changed from 47.6% to 42.8% (7 3.5%)
S ¼ TC ph =TC pr ð11Þ
at pH ¼13.0, from 72.5% to 61.4% (7 4.7%) at pH ¼ 7.0, and
All the experiments were repeated, at least, four times, and the remained constant for pH ¼4.4 (76.8 7 3.4%). The retentate purity,
mean values and standard deviations were calculated. In the P*, increased with increase of VRF at all the studied values of pH.
figures presented below, the values of error bars are equal to Filtration also resulted in near fivefold increase of the relative
means standard deviations. Differentiation of filtration curves concentration of proteins in the retentate.
(calculation of filtrate flux from filtrate volume dependence on The dependence of filtrate purity on VRF at pH¼4.4 did not show
filtration time) and least square fitting of dependencies were done a tendency to decrease with the increase of VRF from 1 to 4.8.
with the help of software Table Curve 2D (Jandel Scientific, USA). Therefore, it may be assumed that at pH¼ 4.4 filtration may be
continued in order to reach higher values of VRF and higher degree of
separation of polyphenols from proteins without losses in filtrate
3. Results and discussion quality. It suggests that coagulation of the flaxseed hull extracts at
pH¼4.4 with consequent centrifugation and ultrafiltration may be a
3.1. Impact of pH on supernatant and filtrate quality promising method for separation of polyphenols from proteins and
for simultaneous concentration of non-coagulated proteins.
Relative decrease of protein, Rpr, and polyphenol, Rph, concen- Fig. 2a shows that coefficient of rejection of polyphenols was the
trations and purity, P, of supernatants, obtained after centrifuga- highest during filtration of supernatant wih pH¼13.2, while the
tion of the extracts, coagulated at different values of pH, are decrease of pH from 13.2 to 7.0 and 4.4 reduced the polyphenol
presented in Fig. 1a and b (open symbols). rejection. Filtration of acidified supernatant (pH¼4.4) also resulted in
 M. Loginov et al. / Journal of Membrane Science 442 (2013) 177–186 181

1 100

0.8 80
Pr, UF
UF

RPr, RPh [−]

0.6 60

P [%]
CF
0.4 Pr, CF 40
Ph, UF
0.2 20
Ph, CF
0 0
4 6 8 10 12 4 6 8 10 12
pH pH
Fig. 1. (a) Relative removal of polyphenols (Ph), Rph, and proteins (Pr), Rpr, and (b) average purity, P, obtained during centrifugation (CF) of extracts (open symbols) and
ultrafiltration (UF) of supernatants (filled symbols) at different values of pH.

1 1
pH = 4.4

0.8 0.8
7.0
13.2
0.6 0.6 0.5
pH = 13.2
RCPh [−]
RCPh [−]

0.4
4.4 0.4 0.3

TCPr
0.4 13.2
0.2
7.0
0.1
0.2 0.2 4.4
0
1 2 3 4 5
VRF
0 0
1 2 3 4 5 1 2 3 4 5
VRF [−] VRF [−]
Fig. 2. Coefficient of rejection of (a) polyphenols, RCph, and (b) proteins, RCpr, in filtrates versus VRF during filtration of supernatants, obtained at different values of pH. Inset
shows dependence of the transmission coefficient of proteins, TCpr, on VRF.

the highest rejection of proteins (Fig. 1b): reduction of pH from 13.2 to In our case, the impact of pH reduction from 13.2 to 4.4 on
4.4 reduced the transmission coefficient of proteins, approximately, by rejection of solutes was opposite for polyphenols (Fig. 2a) and proteins
the factor of 1.5 (Insert in Fig. 2b). Consequently, reduction of the (Fig. 2b). Different impact of pH on RCph and RCpr may be explained by
extract pH from 13.2 to 4.4 resulted not only in a more efficient difference in chemical properties of proteins and polyphenols (differ-
purification by centrifugation (Fig. 1b) but also in a more efficient ent dependence of their charge, structure, and intermolecular inter-
removal of proteins and lower rejection of polyphenols during action on pH). However, simultaneous increase of RCpr and RCph during
ultrafiltration (Fig. 2a and b). Both RCph and RCpr monotonously the filtration suggests that formation of fouled layer played an
increased during filtration at all the studied values of pH (Fig. 2a important role in purification of flaxseed hull extracts. In order to
and b). clarify the role of membrane fouling in purification, filterability of
Similar decrease of polyphenol rejection with decrease of pH (as it supernatants, obtained at different values of pH, was analysed.
is presented in Fig. 2a) was previously observed during the cross-
flow ultrafiltration of bergamot juice with the help of fluoropoly- 3.2. Impact of pH on filterability of supernatants. Analysis
meric membranes with MWCO¼1 kDa [32]. The authors assumed of filtration mechanism
that higher rejection of polyphenols at basic pH may be explained by
formation of a filter cake on the membrane surface, which retains Fig. 3 presents the dependence of filtrate flux, J, versus VRF for
polyphenols [32]. It was shown that rejection of the polyphenols supernatants with different initial values of pH, obtained during
increased with the increase of filtration pressure, which promoted the stirred dead-end filtration. The experiments were performed
the membrane fouling (cake formation) [32]. Negative impact of filter in a concentration mode (without recycling of filtrate).
cake formation on transmission of polyphenols into filtrate was The similar behaviour of filtrate flux was observed at all the studied
observed during the ultrafiltration of grape must [5], roselle extract values of pH: steep decrease of J at the initial stage of filtration (VRF
[19] and grape seed extracts [21], microfiltration of pomegranate ≤1.1) was followed by its monotonous decrease at the later stage
juice [23] and melon juice [22]. Ultrafiltration experiments with (VRF¼1.1–5.0). Similar dependence of J on VRF is commonly observed
model solutions of different proteins [38,45,47] had shown that for micro- and ultrafiltration of various juices [17,20,28,32], extracts
maximal rejection of proteins was observed at pH value correspond- [5,19], model solutions of proteins [37,38,43,44,52], polyphenols [34],
ing to their isoelectric point. This was explained by formation of a and their mixtures [29,35]. Usually, it is explained by fast irreversible
thicker, or denser, (less permeable for solutes) fouling layer and membrane fouling (adsorption of proteins or polyphenols on the
enhanced attraction of uncharged proteins to filtration membrane. membrane surface [34,35,43], and pore blocking [37,44]) followed by
182 M. Loginov et al. / Journal of Membrane Science 442 (2013) 177–186

0.4
10 VRF = 1.1
4 15
0.3

J/Jm [−]
8 0.2 5.0
3 ·c

·c [1014 m/kg]

10

 [1017 m/kg]
0.1
J [10−6 m3 / m2·s]

pH = 13.2
6 0 3
2 pH = 4.4
4 7 10 13

t/V [s/ml]
pH 2 13.2 
5
4 7.0 1
1
0
0 0.5 1
4.4 V [ml]
2
2 4 6 8 10 12 14
pH
Fig. 4. Cake resistance, α  c, and specific cake resistance, α, versus pH of super-
0 natants. The data were obtained from unstirred dead-end filtration experiments.
1 2 3 4 5
Filtration pressure was equal to 4  105 Pa. Inset shows examples of filtration
VRF [−] curves obtained during the unstirred dead-end filtration experiments and pre-
sented in coordinates of Eq. (3).
Fig. 3. Dependence of filtrate flux, J, on VRF at different values of pH. Inset shows
the impact of pH on the relative filtrate flux, J/Jm, at initial (VRF¼ 1.1) and final
(VRF¼5.0) stages of filtration. Filtration pressure was equal to 4  105 Pa. absence of the electrostatic repulsion between the uncharged
protein molecules [57].
formation of cake layer or polarized layer on the membrane surface In order to verify our assumption about the role of cake layer
[34]. Also, both internal and external fouling may control the flux formation during the ultrafiltration of flaxseed hull extracts,
decline simultaneously from the beginning of cross-flow ultrafiltration unstirred dead-end filtration experiments were performed at
of proteins [52]. different values of pH.
Filtrate flux was decreasing as pH of supernatants decreased from At all the studied values of pH, the obtained filtration curves
13.2 to 4.4 (Fig. 3). The impact of pH reduction on filterability was were linear in the coordinates t/V vs. V (Insert in Fig. 4), which
negative both at initial (VRF¼1.1) and final (VRF¼ 5.0) stages of confirms that cake formation was the fouling mechanism during
filtration (Insert in Fig. 3). Similar trend was repeatedly reported for the unstirred filtration. The filtration curves were fitted by Eq. (3)
ultrafiltration of model protein solutions [37–45]. However, the effect with the correlation coefficient ρ2 4 0.99, and the values of α∙c
of pH on filterability of complex polyphenol-containing liquids (juices, were determined. Fig. 4 demonstrates that filter cake resistance,
extracts or complex mixtures of proteins and polyphenols) was rarely α∙c, increased with decrease of pH from 13.2 to 7.0 and remained
discussed in the literature. Conidi et al. [32] had shown that reduction approximately constant during the further decrease of pH. How-
of pH from 8.5 to 5.4 and 2.5 enhanced filterability of polyphenol- ever, the values of α∙c, presented in Fig. 4, are compared for
containing bergamot juice. This was explained by better solubility of supernatants with different initial concentrations of proteins and
polyphenols in acidic medium and reduced cake formation [32]. In polyphenols (Fig. 1a). If we assume that the obtained filter cake is
contrast, Susanto et al. [34] have demonstrated that filtration rate of predominantly composed of protein molecules, then the specific
the model solution of polyphenols through the PES membrane with cake resistance, α, may be estimated by division of α∙c on the
MWCO¼10 kDa decreased as pH of solution decreased from 9.0 to 4.0. concentration of proteins in the supernatant. Fig. 4 shows that
This was explained by higher adsorptive fouling of PES membrane by specific resistance of the filter cakes, formed during the filtration
polyphenols at lower pH. of flaxseed hull extracts, is relatively high (α∼1017 m/kg) and
Comparison of literature data with the results presented in increases significantly with pH decrease from 13.2 to 4.4. A similar
Fig. 3 suggests that decrease of filterability of the clarified flaxseed dependence of α on pH was reported by Iritani et al. [57,58] for
hull extracts (Fig. 3) at lower pH may be explained by the pivotal pure proteins (BSA, lysozyme). These data support the assumption
role of proteins in the membrane fouling. Though concentration of that lower filtrate flux at the final stage of stirred ultrafiltration of
proteins is lower at pH ¼4.4 than at pH ¼13.2 (Fig. 1a), the the flaxseed hull extracts with pH ¼4.4 may be explained by
presence of residual proteins in supernatants may cause both the formation of a proteins-containing filter cake with significantly
membrane fouling and cake formation. higher specific resistance.
At the following stage the filtrate flux remained lower for lower In order to study the reversibility of filter cake formation and
values of pH (Insert in Fig. 3, VRF¼ 5.0). This may be attributed to role of stirring in the reduction of fouling during the ultrafiltration
appearance of the filter cake with lower hydraulic permeability. It of flaxseed hull extracts, the interrupted filtration experiments
was reported previously that during ultrafiltration of model protein were performed. Fig. 5 demonstrates filtration curves, obtained
solutions, the long-term flux was decreasing as pH approached the during the interrupted dead-end filtration with and without
isoelectric point [55,56], and this behaviour was explained by continuous stirring at different values of pH.
formation of a denser cake layer on the membrane surface [56]. Continuous stirring of retentate resulted in significant improve-
Also, Iritani et al. [57,58] have demonstrated that specific resistance ment of filtration rate as compared to unstirred dead-end filtration
of the filter cake, formed during the dead-end ultrafiltration of at all the studied values of pH. In experiments with unstirred
protein solutions, was maximal, and the cake porosity was minimal, dead-end filtration (filled symbols in Fig. 5), the interruption of
in the isoelectric point of proteins. This was explained by the filtration for 60 s (reduction of filtration pressure to 0 with
 M. Loginov et al. / Journal of Membrane Science 442 (2013) 177–186 183

(transmission of proteins and polyphenols) through this layer was

higher at lower values of pH. The drag force, associated with the
101 filtrate flow, is constant at the constant values of filtrate flux (or
pH = 13.2
fouling resistance) regardless of pH value. A possibility of trans-
mission of the protein molecules into the filtrate must be deter-
mined by the balance between the drag force of filtrate flow
through the cake and intermolecular interaction between dis-
solved proteins and deposited proteins (filter cake). Therefore, the
J [10−6 m3/m2·s]

4.4 decrease of the rejection coefficient of proteins as pH approaches

the isoelectric point (pH ¼4.4, Fig. 6a) can be explained by
reduction of electrical repulsion between the bulk proteins and
100 proteins deposited on the surface of membrane. Weaker rejection
of polyphenols by the fouling layer, formed at lower values of pH,
13.2 probably, was determined by weaker interaction between the
solved polyphenols and uncharged proteins.
Fig. 6b presents dependence of selectivity on relative fouling
resistance determined at different values of pH. It is interesting to
4.4 note that selectivity was practically independent of pH, and it
increased with increase of fouling resistance. Dependence of S on
r/rm may be fitted by linear equation
1 1.1 1.2
VRF [−] S ¼ 0:83⋅ðr=r m Þ–0:68 ð12Þ
Fig. 5. Dependence of filtrate flux, J, on VRF during the unstirred dead-end with correlation coefficient 0.96 (dependence, calculated by least
filtration (filled symbols) and stirred dead-end filtration (open symbols) of super-
natants with different values of pH. Filtration pressure was equal to 4∙105 Pa. Arrow
square fitting, is presented in Fig. 6b by dashed line). According to
shows the moment of deposited layer removal (decreasing of filtration pressure to this equation, S ¼1 when r/rm ≈2. It follows that filtration through
0 with simultaneous stirring during 60 s). the non-fouled membrane (when r/rm ¼ 1) does not result in
purification of filtrate, since S ¼1 when r/rm ≤2. Therefore, forma-
simultaneous application of retentate stirring) resulted in signifi- tion of a fouling layer with a certain minimal resistance (which is
cant recovery of filtrate flux. This may be explained by partial equal or superior to resistance of clean PES membrane, rm, used in
removal of the filter cake during the stirring without pressure this study) is necessary for purification of the extracts. According
application. Nevertheless, recovery of the filtrate flux was incom- to the data, presented in Fig. 3, this value of fouling resistance was
plete: the value of filtrate flux, observed after the pressure attained practically immediately after the beginning of filtration
restoration, was notably lower than the initial filtrate flux (espe- (at VRF ≈1). The growth of the fouling layer (increase of r/rm)
cially at pH ¼ 4.4). In experiments with continuous stirring (open resulted not only in the flux decline, but also in the increase of
symbols in Fig. 5) the filtration interruption resulted in partial selectivity (Fig. 6b). It caused efficient separation of polyphenols
recovery of filtrate flux for extract with pH ¼ 13.2 and had from proteins despite of constant increase of protein concentration
negligible effect on filtration at pH ¼4.4. It may be concluded that in the retentate. The importance of fouling layer role in purifica-
fouling during the stirred ultrafiltration of flaxseed hull extracts tion (rejection of impurities and target solutes) has been already
was partially reversible at high values of pH, but practically noted in several studies [48,59–63].
irreversible in the isoelectric point (pH ¼4.4), at least, at studied Also, it may be supposed that the size of residual aggregates of
hydrodynamic conditions of cake removal. This may be explained proteins and polyphenols that remain in extracts after the coagu-
either by higher affinity of deposited layer of uncharged proteins lation and centrifugation are different at different values of pH.
to the membrane surface [38,39,41,43,46] or by more significant This may explain the difference in filtration behaviour (rejection of
role of internal membrane fouling during filtration of supernatants solutes and filter cake resistance) of the studied centrifugates.
with pH ¼ 4.4 as compared to supernatants with pH ¼13.2 [37,44]. It may be deduced from the data, presented in Fig. 6, that
It should be noted that rinsing of the used membrane by fouling layer removal from the membrane surface (or reduction of
distilled water after the stirred filtration experiments resulted in fouling) may cause undesirable decrease of the filtrate purity. In
practically complete recovery of the membrane permeability (i.e., order to verify this assumption, pH of supernatant with pH ¼4.4
filtration rate of water through the rinsed membrane was practi- was adjusted to pH ¼7.0 by addition of concentrated sodium
cally equal to the initial filtration rate of supernatants through the hydroxide solution, and the obtained solution was subjected to
new membrane at pH ¼ 4.4–13.2). stirred ultrafiltration. Initial composition of this solution (concen-
tration of proteins and polyphenols) corresponded to the compo-
3.3. Impact of deposited layer on filtrate quality sition of the supernatant with pH ¼ 4.4.
Fig. 7a compares the filtrate flux and the rejection of proteins
Since filtration efficiency is controlled by both filtration rate and polyphenols during filtration of supernatants without pH
and filtrate purity, relation between filtration rate (fouling resis- adjustment (pH ¼4.4) and with pH adjustment (pH ¼4.4-7.0).
tance) and rejection of proteins and polyphenols at different Increase of pH resulted in the increase of filtrate flux, J (Fig. 7a).
values of pH was determined (Fig. 6a). According to Fig. 4, this may be explained by formation of a fouling
Increase of the fouling resistance (that may be associated with layer with lower specific resistance at pH¼7.0. At the same time,
a growth of filter cake layer) resulted in a monotonous increase of adjustment of pH from 4.4 to 7.0 reduced rejection of proteins, RCpr,
rejection coefficients of proteins and polyphenols at all the studied and polyphenols, RCph, as it was expected from the data, presented in
values of pH (Fig. 6a). At constant value of r/rm (constant value of Fig. 2. Impact of pH adjustment and corresponding reduction of
filtrate flux), rejection of polyphenols and proteins was lower at fouling resistance on filtrate quality is demonstrated in Fig. 7b.
lower values of pH. Therefore, though reduction of pH of the Increase of pH from 4.4 to 7.0 after centrifugation resulted in a
flaxseed hull extracts resulted in formation of the filter cake layer notable decrease of selectivity. The selectivity of alkalinized super-
with lower hydraulic permeability, the permeability of solutes natant (pH¼4.4-7.0) declined during filtration from high values,
184 M. Loginov et al. / Journal of Membrane Science 442 (2013) 177–186

1 10
pH = 13.2 7.0 pH = 13.2
7.0
0.8 4.4 8
4.4

RCPr, RCPh [−] 0.6 6

S [−]
13.2
0.4 4
7.0
0.2 4.4 Pr 2
Ph

0 0
2 4 6 8 10 2 4 6 8 10
r/rm [−] r/rm [−]

Fig. 6. Dependences of the (a) coefficients of rejection of proteins (Pr), RCpr, and polyphenols (Ph), RCph, and (b) selectivity, S, on relative resistance of fouled membrane, r/rm,
at different pH.

1 6
pH = 4.4, 4.4 →7.0
10 4

S [−]
6 0.8 2
RCpr pH = 4.4
8
0
J [10−6 m3/m2·s]

RCPr, RCPh [−]

4 0.6 2 3 4 5 6
r/rm [−]
S [−]

6
4.4 →7.0
J 0.4
2
4
RCph 7.0
0.2
0
2
0
1 2 3 4 5 1 2 3 4 5
VRF [−] VRF [−]
Fig. 7. (a) Dependence of filtrate flux, J, and coefficients of rejection of proteins, RCpr (filled symbols), and polyphenols, RCph (open symbols), on VRF (filtration pressure was
equal to 4∙105 Pa.). (b) Dependence of selectivity, S, on VRF and relative resistance of fouled membrane, r/rm(Inset). Filtration of extracts without pH adjustment after
centrifugation (pH ¼ 4.4 and 7.0, solid lines) and with pH adjustment (pH ¼ 4.4-7.0, dashed lines).

observed for supernatant with pH¼ 4.4, to lower values that were steps: (i) coagulation of proteins by adjustment of pH and subsequent
practically equal to those observed for supernatant with pH¼ 7.0. centrifugation, and (ii) ultrafiltration of supernatants. The value of pH,
However, increase of the fouling layer resistance during filtration used for extract coagulation, varied from 13.2 to 3.0.
(with increasing of VRF) resulted in the increase of selectivity. As a The ultrafiltration of flaxseed hull extract may be a promising
consequence, dependence of filtration selectivity, S, on the relative method for separation of polyphenols and proteins, since it resulted
resistance of fouled membrane, r/rm, of alkalinized supernatant in high degree of purification, which cannot be attained by simple
(Insert in Fig. 7b, filled symbols) was not significantly different from coagulation and centrifugation. Adjustment of pH prior to ultrafil-
those of supernatants with pH¼4.4 and 7.0 (Insert in Fig. 7b, open tration had a dual impact on filtration efficiency: decrease of pH
symbols). The results of this experiment suggest that the factor, from 13.2 to 4.4 resulted in more severe filtrate flux decline, but also
which positively influences filtration rate (reduction of fouling due to in better purity and lower rejection of polyphenols. This was
increase of pH), had negative effect on the filtrate purity. This must explained by pivotal role of residual proteins in the membrane
be taken into consideration for optimization of filtration purification fouling. Despite of lower concentration of proteins in supernatants
of the flaxseed hull extracts. If formation of a fouling layer is with pH¼4.4, their negative effect on filterability was more
undesirable, filtration membrane with higher permeability and pronounced as compared to supernatants with pH¼13.2. This
higher selectivity should be used for filtration in order to assure may be explained by (i) formation of the filter cake with signifi-
efficient purification at high filtration rate. However, it is necessary to cantly higher specific resistance, and (ii) irreversibility of fouling,
verify this conclusion by filtration experiments with fouling reduced caused by stronger attraction of uncharged protein molecules to PES
by more convenient method (for example, by intense shearing membrane at lower values of pH. The membrane fouling signifi-
during the dynamic or tangential filtration [64–66]). cantly increased the selectivity of separation of polyphenols and
proteins and resulted in better filtrate quality.
The results of this study suggest that the factor, which posi-
4. Conclusions tively influences filtration rate (reduction of fouling due to
increase of pH), had a negative effect on filtrate purity. This must
The possibility of separation of polyphenols and proteins, extracted be taken into consideration for the optimization of filtration
from flaxseed hulls, was discussed. Their separation comprised two purification of the flaxseed hull extracts.
 M. Loginov et al. / Journal of Membrane Science 442 (2013) 177–186 185

Acknowledgements [4] B. Díaz-Reinoso, N. González-López, A. Moure, H. Domínguez, J.C. Parajó,

Recovery of antioxidants from industrial waste liquors using membranes and
polymeric resins, J. Food Eng. 96 (2010) 127–133.
Authors thank Lasalle Beauvais (Beauvais, France) for provision [5] A. Cassano, A. Mecchia, E. Drioli, Analyses of hydrodynamic resistances and
of flaxseed hulls, Dr. Zhenzhou Zhu for provision of filtration operating parameters in the ultrafiltration of grape must, J. Food Eng. 89
membrane and Dr. Yuri Boiko for provision of filtration cell. (2008) 171–177.
[6] N. Boussetta, E. Vorobiev, V. Deloison, F. Pochez, A. Falcimaigne-Cordin, J.-
Authors gratefully thank N. S. Pivovarova for her kind help with L. Lanoisellé, Valorisation of grape pomace by the extraction of phenolic
preparation of the manuscript. antioxidants: application of high voltage electrical discharges, Food Chem. 128
(2011) 364–370.
[7] A.U. Buranov, G. Mazza, Extraction and purification of ferulic acid from flax
shives, wheat and corn bran by alkaline hydrolysis and pressurised solvents,
Food Chem. 115 (2009) 1542–1548.
Nomenclature [8] C. Gutiérrez, M. Rubilar, C. Jara, M. Verdugo, J. Sineiro, C. Shene, Flaxseed and
flaxseed cake as a source of compounds for food industry, J. Soil Sci. Plant
Nutrition 10 (2010) 454–463.
A membrane surface area, m2 [9] B. Díaz-Reinoso, A. Moure, H. Domínguez, J.C. Parajó, Ultra- and nanofiltration
Cph concentration of polyphenols, g/l of aqueous extracts from distilled fermented grape pomace, J. Food Eng. 91
(2009) 587–593.
Cpr concentration of proteins, g/l
[10] H. Nawaz, J. Shi, G.S. Mittal, Y. Kakuda, Extraction of polyphenols from grape
Cpr,o initial concentration of proteins, g/l seeds and concentration by ultrafiltration, Sep. Pur. Technol. 48 (2006)
c weight fraction of fouling compounds in super- 176–181.
[11] V. Beejmohun, O. Fliniaux, E. Grand, F. Lambalin, L. Bensaddek, P. Christen,
natant, dimensionless
J. Kovensky, M.-A. Fliniaux, F. Mesnard, Microwave-assisted extraction of the
J filtrate flux, m3/(m2 s) main phenolic compounds in flaxseed, Phytochem. Analys. 18 (2007) 275–282.
Jm filtrate flux through the clean membrane, [12] N. El Darra, N. Grimi, E. Vorobiev, N. Louka, R. Maroun, Extraction of
m3/(m2 s) polyphenols from red grape pomace assisted by pulsed ohmic heating, Food
Bioproc. Technol. (2012) 1–9, article in press.
P purity towards polyphenols, % [13] C.H.L. Ho, J.E. Cacace, G. Mazza, Extraction of lignans, proteins and carbohy-
P* purity towards proteins, % drates from flaxseed meal with pressurized low polarity water, LWT—Food
Δp filtration pressure, Pa Sci. Technol. 40 (2007) 1637–1647.
[14] N. Boussetta, Intensification de l'extraction des polyphénols par électro-
Rph relative decrease of polyphenol concentration, technologies pour la valorisation des marcs de champagne, PhD Thesis,
dimensionless Université de Technologie de Compiègne, 2010.
Rpr relative decrease of protein concentration, [15] Z. Borneman, V. Gökmen, H.H. Nijhuis, Selective removal of polyphenols and
brown colour in apple juices using PES/PVP membranes in a single-
dimensionless
ultrafiltration process, J.Membr. Sci. 134 (1997) 191–197.
RCph coefficient of rejection of polyphenols, dimensionless [16] B. Díaz-Reinoso, A. Moure, H. Domínguez, J.C. Parajó, Membrane concentration
RCpr coefficient of rejection of proteins, dimensionless of antioxidants from Castanea sativa leaves aqueous extracts, Chem. Eng. J. 175
r resistance of fouled filtration membrane, m−1 (2011) 95–102.
[17] A. Cassano, L. Donato, C. Conidi, E. Drioli, Recovery of bioactive compounds in
rm resistance of clean membrane, m−1 kiwifruit juice by ultrafiltration, Innov. Food Sci. Emerg. Technol. 9 (2008)
S selectivity, dimensionless 556–562.
TCpr coefficient of transmission of proteins, dimensionless [18] A. Cassano, C. Conidi, E. Drioli, Clarification and concentration of pomegranate
juice (Punicagranatum L.) using membrane processes, J. Food Eng. 107 (2011)
t filtration time, s 366–373.
V volume of filtrate, ml [19] M. Cissé, F. Vaillant, D. Pallet, M. Dornier, Selecting ultrafiltration and
VRF volume reduction factor, dimensionless nanofiltration membranes to concentrate anthocyanins from roselle extract
(Hibiscus sabdariffa L.), Food Res. Int. 44 (2011) 2607–2614.
[20] C. Conidi, A. Cassano, E. Drioli, Recovery of phenolic compounds from orange
Greek letters press liquor by nanofiltration, Food Bioprod. Proc. 90 (2012) 867–874.
[21] D. Liu, E. Vorobiev, R. Savoire, J.-L. Lanoisellé, Intensification of polyphenols
extraction from grape seeds by high voltage electrical discharges and extract
α specific cake resistance, m/kg concentration by dead-end ultrafiltration, Sep. Pur. Technol. 81 (2011)
μ viscosity of filtrate, Pa s 134–140.
[22] F. Vaillant, M. Cisse, M. Chaverri, A. Perez, M. Dornier, F. Viquez, C. Dhuique-
ρ density of filtrate, kg/m3
Mayer, Clarification and concentration of melon juice using membrane
ω stirring rate, rpm processes, Innov. Food Sci. Emerg. Technol. 6 (2005) 213–220.
[23] H. Mirsaeedghazi, Z. Emam-Djomeh, S.M. Mousavi, R. Ahmadkhaniha,
A. Shafiee, Effect of membrane clarification on the physicochemical properties
Abbreviations of pomegranate juice, Int. J. Food Sci. Technol. 45 (2010) 1457–1463.
[24] A. Laorko, Z. Li, S. Tongchitpakdee, S. Chantachum, W. Youravong, Effect of
membrane property and operating conditions on phytochemical properties
CF centrifugation
and permeate flux during clarification of pineapple juice, J. Food Eng. 100
MWCO molecular weight cut-off (2010) 514–521.
PES polyethersulfone [25] Y.E. Rayess, C. Albasi, P. Bacchin, P. Taillandier, M. Mietton-Peuchot,
Ph polyphenols A. Devatine, Cross-flow microfiltration of wine: effect of colloids on critical
fouling conditions, J. Membr. Sci. 385–386 (2011) 177–186.
Pr proteins [26] A. El Rayess, Y. Albasi, C. Bacchin, P. Taillandier, P. Raynal, J. Mietton-Peuchot,
UF ultrafiltration M. Devatine, Cross-flow microfiltration applied to oenology: a review, J.
Membr. Sci. 382 (2011) 1–19.
[27] A. El Rayess, Y. Albasi, C. Bacchin, P. Taillandier, P. Mietton-Peuchot,
M. Devatine, Analysis of membrane fouling during cross-flow microfiltration
of wine, Innov. Food Sci. Emerg. Technol. 16 (2012) 398–408.
References [28] K.S. Bahçeci, Effects of pretreatment and various operating parameters on
permeate flux and quality during ultrafiltration of apple juice, Int. J. Food Sci.
Technol. 47 (2012) 315–324.
[1] C.M. Galanakis, Recovery of high added-value components from food wastes: [29] M. Zator, M. Ferrando, F. López, C. Güell, Microfiltration of protein/dextran/
conventional, emerging technologies and commercialized applications, Trends polyphenol solutions: characterization of fouling and chemical cleaning
Food Sci. Technol. 26 (2012) 68–87. efficiency using confocal microscopy, J. Membr. Sci. 344 (2009) 82–91.
[2] N. Boussetta, E. Vorobiev, L.H. Le, A. Cordin-Falcimaigne, J.-L. Lanoisellé, [30] P. Czekaj, F. López, C. Güell, Membrane fouling by turbidity constituents of
Application of electrical treatments in alcoholic solvent for polyphenols beer and wine: characterization and prevention by means of infrasonic
extraction from grape seeds, LWT—Food Sci. Technol. 46 (2012) 127–134. pulsing, J. Food Eng. 49 (2001) 25–36.
[3] B. Santamaría, G. Salazar, S. Beltrán, J.L. Cabezas, Membrane sequences for [31] A. Vernhet, D. Cartalade, M. Moutounet, Contribution to the understanding of
fractionation of polyphenolic extracts from defatted milled grape seeds, fouling build-up during microfiltration of wines, J. Membr. Sci. 211 (2003)
Desalination 148 (2002) 103–109. 357–370.
186 M. Loginov et al. / Journal of Membrane Science 442 (2013) 177–186

[32] C. Conidi, A. Cassano, E. Drioli, A membrane-based study for the recovery of [49] V.L. Singleton, R. Orthofer, R.M. Lamuela-Raventos, Analysis of total phenols
polyphenols from bergamot juice, J. Membr. Sci. 375 (2011) 182–190. and other oxidation substrates and antioxidants by means of Folin-Ciocalteu
[33] Z.S. Saleh, J.M. Farr, S.A. Friedrich, Mitigating and modelling membrane fouling reagent, oxidants antioxidants, Methods Enzymol. 299 (1999) 152–178.
by bioactive mixtures, Filtration 11 (2011) 52–59. [50] M.M. Bradford, A rapid and sensitive method for the quantification of
[34] H. Susanto, Y. Feng, M. Ulbricht, Fouling behavior of aqueous solutions of microgram quantities of protein utilizing the principle of protein dye binding,
polyphenolic compounds during ultrafiltration, J. Food Eng. 91 (2009) Analytical Biochem. 72 (1976) 248–254.
333–340. [51] O. Dangles, Propriétéschimiques des polyphénols, in: P. Sarni-Manchando,
[35] M. Ulbricht, W. Ansorge, I. Danielzik, M. König, O. Schuster, Fouling in V. Cheynier (Eds.), Les polyphénols en agroalimentaire, Lavoisier, Paris, 2006,
microfiltration of wine: the influence of the membrane polymer on adsorption pp. 29–54.
of polyphenols and polysaccharides, Sep. Purif. Technol. 68 (2009) 335–342. [52] T.-A. Nguyen, S. Yoshikawa, K. Karasu, S. Ookawara, A simple combination
[36] A. Vernhet, M. Moutounet, Fouling of organic microfiltration membranes by model for filtrate flux in cross-flow ultrafiltration of protein suspension, J.
wine constituents: mportance, relative impact of wine polysccharides and Membr. Sci. 403–404 (2012) 84–93.
polyphenols and incidence of membrane properties, J. Membr. Sci. 201 (2002) [53] E. Iritani, S. Tachi, T. Murase, Influence of protein adsorption on flow resistance
103–122. of microfiltration membrane, Coll. Surf. A Physicochem. Eng. Aspects 89 (1994)
[37] M. Ouammou, N. Tijani, J.I. Calvo, C. Velasco, A. Martin, F. Martinez, F. Tejerina, 15–22.
A. Hernández, Flux decay in protein microfiltration through charged mem- [54] E. Iritani, S. Nakatsuka, H. Aoki, T. Murase, Effect of solution environment on
branes as a function of pH, Coll. Surf. A: Physicochem. Eng. Aspects 298 (2007) unstirred dead-end ultrafiltration characteristics of proteinaceous solutions, J.
267–273. Chem. Eng. Jpn 24 (1991) 177–183.
[38] I.H. Huisman, P. Prádanos, A. Hernández, The effect of protein-protein and [55] Y.-N. Wang, C.Y. Tang, Protein fouling of nanofiltration, reverse osmosis, and
protein-membrane interactions on membrane fouling in ultrafiltration, J. ultrafiltration membranes. The role of hydrodynamic conditions, solution
chemistry, and membrane properties, J. Membr. Sci. 376 (2011) 275–282.
Membr. Sci. 179 (2000) 79–90.
[56] Y.P. Lim, A.W. Mohammad, Influence of pH and ionic strength during food
[39] S.P. Palecek, A.L. Zydney, Intermolecular electrostatic interactions and their
protein ultrafiltration: elucidation of permeate flux behavior, fouling resis-
effect on flux and protein deposition during protein filtration, Biotechnol.
tance, and mechanism, Sep. Sci. Technol. 47 (2012) 446–454.
Prog. 10 (1994) 207–213.
[57] Y. Mukai, E. Iritani, T. Murase, Effect of protein charge on cake properties in
[40] G. Capar, Separation of silkworm proteins in cocoon cooking wastewaters via
dead-end ultrafiltration of protein solutions, J. Membr. Sci. 137 (1997)
nanofiltration: effect of solution pH on enrichment of sericin, J. Membr. Sci.
271–275.
389 (2012) 509–521.
[58] E. Iritani, Y. Mukai, E. Hagihara, Measurements and evaluation of concentra-
[41] J.A. Koehler, M. Ulbricht, G. Belfort, Intermolecular forces between proteins
tion distributions in filter cake formed in dead-end ultrafiltration of protein
and polymer films with relevance to filtration, Langmuir 13 (1997) 4162–4171.
solutions, Chem. Eng. Sci. 57 (2002) 53–62.
[42] R. Chan, V. Chen, The effects of electrolyte concentration and pH on protein
[59] H. Susanto, I.N. Widiasa, Ultrafiltration fouling of amylose solution: behavior,
aggregation and deposition: critical flux and constant flux membrane filtra-
characterization and mechanism, J. Food Eng. 95 (2009) 423–431.
tion, J. Membr. Sci. 185 (2001) 177–192. [60] S. Kishihara, H. Tamaki, S. Fujii, M. Komoto, Clarification of technical sugar
[43] C. Herrero, P. Prádanos, J.I. Calvo, F. Tejerina, A. Hernández, Flux decline in solutions through a dynamic membrane formed on a porous ceramic tube, J.
protein microfiltration: influence of operative parameters, J. Coll. Int. Sci. 187 Membr. Sci. 41 (1989) 103–114.
(1997) 344–351. [61] R. Jiraratananon, D. Uttapap, C. Tangamornsuksun, Self-forming dynamic
[44] S.-H. Lin, C.-L. Hung, R.-S. Juang, Applicability of the exponential time membrane for ultrafiltration of pineapple juice, J. Membr. Sci. 129 (1997)
dependence of flux decline during dead-end ultrafiltration of binary protein 135–143.
solutions, Chem. Eng. J. 145 (2008) 211–217. [62] N.K. Saha, M. Balakrishnan, M. Ulbricht, Sugarcane juice ultrafiltration: FTIR
[45] P. Sarkar, S. Ghosh, S. Dutta, D. Sen, C. Bhattacharjee, Effect of different and SEM analysis of polysaccharide fouling, J. Membr. Sci. 306 (2007)
operating parameters on the recovery of proteins from casein whey using a 287–297.
rotating disc membrane ultrafiltration cell, Desalination 249 (2009) 5–11. [63] N.K. Saha, M. Balakrishnan, M. Ulbricht, Polymeric membrane fouling in
[46] L. Ricq, S. Narçon, J.-C. Reggiani, J. Pagetti, Streaming potential and protein sugarcane juice ultrafiltration: role of juice polysaccharides, Desalination
transmission ultrafiltration of single proteins and proteins in mixture: β- 189 (2006) 59–70.
lactoglobulin and lysozyme, J. Membr. Sci. 156 (1999) 81–96. [64] Z. Zhu, J. Luo, L. Ding, O. Bals, M.Y. Jaffrin, E. Vorobiev, Chicory juice
[47] R. van Reis, J.M. Brake, J. Charkoudian, D.B. Burns, A.L. Zydney, High- clarification by membrane filtration using rotating disk module, J. Food Eng.
performance tangential flow filtration using charged membranes, J. Membr. 115 (2013) 264–271.
Sci. 159 (1999) 133–142. [65] A. Massé, H.N. Thi, P. Legentilhomme, P. Jaouen, Dead-end and tangential
[48] M. Loginov, K. Loginova, N. Lebovka, E. Vorobiev, Comparison of dead-end ultrafiltration of natural salted water: influence of operating parameters on
ultrafiltration behaviour and filtrate quality of sugar beet juices obtained by specific energy consumption, J. Membr. Sci. 380 (2011) 192–198.
conventional and ‘cold’ PEF-assisted diffusion, J. Membr. Sci. 377 (2011) [66] V. Espina, M.Y. Jaffrin, L. Ding, B. Cancino, Fractionation of pasteurized skim
273–283. milk proteins by dynamic filtration, Food Res. Int. 43 (2010) 1335–1346.