Journal of Environmental Management 222 (2018) 447–453

Contents lists available at ScienceDirect

Journal of Environmental Management

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

Research article

Novel micro/ultra/nanocentrifugation membrane process assessment for T

revalorization and reclamation of agricultural wastewater
J.M. Ochando-Pulido∗, A. Martinez-Ferez
Department of Chemical Engineering, University of Granada, Avenida Fuentenueva s/n, 18071, Granada, Spain

A R T I C LE I N FO A B S T R A C T

Keywords: The concentration and recovery of the high-added value phenolic fraction from two-phase olive mill wastewater
Olive mill wastewater and the simultaneous effluent treatment by a novel micro/ultra/nanocentrifugation membrane process assess-
Membranes ment is addressed, permitting to gather information for a correct and effective screening procedure for the
Phenolic compounds adequate membrane election (MF-UF-loose NF) for the target. Phenolic compounds are the major factor of
Wastewater reclamation
phytotoxicity of these effluents, but on the other hand they present high antioxidant properties that makes them
Sustainability
very relevant for food, cosmetic, pharmaceutical and biotechnological industries. The selection of a membrane
MWCO between 100 kDa and 0.45 μm permitted the complete transfer of the phenolic fraction to the permeate,
whereas below 3 kDa they would be transferred to the concentrate stream instead, with ∼60% COD reduction
and EC lowered to 551–662 μS cm−1 in the final treated stream ensured, sensibly improving the effluent quality.
This would provide a purified effluent with good salinity standards according to the indications given by the FAO
for irrigation reuse. This procedure could be quick and reliable for the assessment of the adequate membrane
needed for a particular purification process, in contrast with long-term, time consuming common bench-scale
procedures.

1. Introduction and geographically dispersed factories called mills (Niaounakis and

Agro-industrial wastewater is a problem increasingly affecting the
current availability of water resources worldwide, which are becoming
insufficient to satisfy the rising demand of fresh water. This specially
concerns agricultural irrigation, which copes with more than three
quarters of the total water consumption worldwide (International Olive
Oil Council, 2015). The intensive and unsustainable development of
industry is causing deterioration of the environmental quality.
One solution in this scenario is the use of regenerated wastewater
for irrigation purposes, which could be very positive environmentally,
as well as result in economic return. Moreover, industrial effluents are
being subjected to increasingly tougher environmental regulations,
especially in European countries (https://ec.europa.eu/prog, 2020).
Hence, the adequate management of the by-produced effluents in in-
dustries is becoming a key task to be accounted for.
In that regard, the treatment of agro-industrial wastewater is an
issue that has joined much research effort in the last years. However,
the medium to low size of these industries and their dispersity in most
cases poses a handicap for the cost-efficient management of their ef-
fluents. This is the case of olive mill wastewater (OMW), a deleterious
effluent generated during the production process of olive oil in small
Halvadakis, 2006). Olive oil, a product that has withdrawn much in-
terest in the market in the last years owed to its nutritional, antioxidant
and health-promoting properties, is obtained throughout a process
based on physical operations, without implication of chemicals. In this
line, this industry is committed to meet the conditions to make the
whole production process environmentally friendly, and this comprises
the treatment of the residual effluents by-produced (Paraskeva and
Diamadopoulos, 2006). Moreover, the important growth of this agro-
industry, ancestrally circumscribed in the Mediterranean Basin but now
in expansion to very different regions in Europe, Australia, the USA, the
middle East and China, is rendering the management of these effluents
a task of global concern (International Olive Oil Council, 2015).
A feasible solution for the treatment of OMW is urgent: 10–15 m3 of
OMW are generated daily in average-sized olive mills, which implies
the by-production of millions of cubic meters of these phytotoxic ef-
fluents yearly and the equivalent potable water consumption
(Paraskeva and Diamadopoulos, 2006). A plethora of technologies can
be found in the scientific literature addressing the problem of OMW
treatment. However, conventional methods imply a series of drawbacks
and limitations such as high cost and energy consumption, complicated
treatment processes, and secondary pollution. Most processes proposed

∗
Corresponding author.
E-mail address: jmochandop@ugr.es (J.M. Ochando-Pulido).

https://doi.org/10.1016/j.jenvman.2018.05.092
Received 7 November 2017; Received in revised form 18 December 2017; Accepted 27 May 2018
Available online 14 June 2018
0301-4797/ © 2018 Elsevier Ltd. All rights reserved.
J.M. Ochando-Pulido, A. Martinez-Ferez Journal of Environmental Management 222 (2018) 447–453

up to the moment for the treatment of OMW are rather cost-inefficient, 2. Experimental
and olive oil industry, in its present status, cannot assume those costs
(Paraskeva and Diamadopoulos, 2006; Ochando-Pulido et al., 2017a; 2.1. The effluent stream: olive mill wastewater
Hodaifa et al., 2013). Among them, advanced oxidation processes
comprising the use of novel nanoparticles (Ochando-Pulido et al., Samples of OMW were collected from various olive oil mills in the
2017a; Hodaifa et al., 2013; Vilardi et al., 2017; Vilardi and Di Palma, Andalusian provinces of Jaén and Granada (Spain), operating with the
2017), tailored biological treatments (Sampaio et al., 2011; Di Palma most up-to-date two-phase olive oil production technology (hereafter
et al., 2003; Stoller et al., 2010) and membrane technologies (Ochando- referred as OMW2). The raw OMW2 was taken directly from the ver-
Pulido and Martinez-Ferez, 2012; Cassano et al., 2011; Di Lecce et al., tical centrifuges in-situ during the production process of olive oil.
2014; Garcia-Castello et al., 2010; Ochando-Pulido et al., 2013; Stoller, After gathering, the raw OMW2 was readily used for the experi-
2011) seem the most promising solutions. ments, with the intention to use the effluent as fresh as it is by-pro-
In this situation, a change of paradigm regarding the economy is duced, to avoid modifying its natural characteristics. On another hand,
being currently experienced, implying that the linear vision is now samples of the raw effluent were kept well refrigerated (- 5 °C) to
being replaced by a circular vision, where waste is regarded as a po- maintain them in good condition for further research experiments. The
tential resource to be reintroduced in the production cycle. Agro-food main physico-chemical characteristics of the raw OMW2 are hereafter
production chain is one of the main waste producers. reported in the experimental results section.
In this line, OMW has been highlighted as an important potential
source of natural antioxidants, mainly by virtue of its rich content in
2.2. Assessment of MF-UF-loose NF membranes
polyphenolic compounds (Obied et al., 2005; Boskou, 2006), which
exhibit excellent antioxidant, radical scavenging, antimicrobial and
The equipment used for the micro/ultra/nanocentrifugation mem-
anticarcinogenic properties (Boskou, 2006). However, the vast majority
brane process assessment, comprised a vertical centrifuge (Thermo
of the processes proposed in the last years aimed for the abatement of
Fischer Scientific, Sorval ST40 series), Falcon tubes (50 mL) provided
the phenolic fraction, which is otherwise the main responsible of the
with membranes of molecular weight cut-off (MWCO) ranging from 3 to
highly phytotoxic and recalcitrant characteristics of these effluents
100 kDa, nitrate cellulose filters (111306-047N Sartorius, 47 mm dia-
(Paraskeva and Diamadopoulos, 2006; Ochando-Pulido et al., 2017a;
meter, Ø, and 0.45 μm mean pore size, Dp), a volumetric flask and a
Hodaifa et al., 2013; Sampaio et al., 2011). Nevertheless, the recovery
diaphragm pump for vacuum, as well as a precision electronic mass
of this fraction from OMW can provide not only economic profit but
balance (AX -120 Cobos, 0.1 mg accuracy).
specifically make OMW less toxic and thus easier to be ulteriorly
The characteristics of the used membranes, purchased from Merck
treated, promoting the overall sustainability of OMW management.
Millipore, are reported in Table 1.
Among recent separation and purification technologies effectively
The proposed micro/ultra/nanocentrifugation membrane process
transferable at industrial scale, membranes offer a series of advantages
aimed to gather information for a correct, effective and reliable
if compared to classic separation techniques, mainly the facts that they
screening procedure for the election of the adequate membrane (MF-
are modular, compact and less energy-intensive, and capable to offer
UF-loose NF) for the target purpose, sensibly quicker than bench-scale
high separation standards (Ochando-Pulido and Martinez-Ferez, 2012;
membrane experiments (Ochando-Pulido et al., 2013).
Cassano et al., 2011; Di Lecce et al., 2014; Garcia-Castello et al., 2010;
The procedure performed consisted, first, in filtering 200 mL OMW2
Ochando-Pulido et al., 2013; Stoller, 2011). This has promoted the use
samples through the 0.45 μm nitrate cellulose filters. After this, 15 mL
of membrane technologies for the recovery of antioxidants from OMW
of the filtered samples were poured into Falcon tubes provided with
(Russo, 2007; Zagklis et al., 2015; Conidi et al., 2014; Bazzarelli et al.,
membranes of different mean pore diameters (MWCO ranging from 100
2016; Abdel-Shafy et al., 2015; Janangiri et al., 2016; Ochando-Pulido
to 3 kDa: 100, 50, 30, 10 and 3 kDa). The micro/ultra/nanocen-
et al., 2017b).
trifugation membrane system is shown in Fig. 1. Subsequently, the
In this work, the concentration and recovery of high-added value
Falcon tubes were centrifuged at 4000 rpm for 3 min. After this pro-
compounds (polyphenolic fraction) from two-phase OMW and the si-
cedure, the permeate and concentrate volumes were taken from the
multaneous treatment of the effluent by a novel micro/ultra/nanocen-
Falcons and weighed on the precision electronic mass balance. The
trifugation membrane process assessment is proposed, aimed to gather
sequence was performed in triplicate.
information for a correct, effective and reliable screening procedure for
Then, samples of permeate and concentrate collected were taken at
the election of the adequate membrane (MF-UF-loose NF) for the target
the end of operation for final and readily analysis, comprising the total
purpose, sensibly quicker than bench-scale membrane experiments. The
polyphenolic fraction (TPhs) concentration, COD, electroconductivity
minor specific energy consumption (SEC) of membranes, if compared
(EC) and pH of all permeate and concentrated volumes of the cen-
with conventional separation processes, have paved the way for their
trifuged-filtered samples. The observed rejection efficiencies (Ri, %)
implementation at industrial scale facilities in multiple applications,
were calculated with the following expression:
and particularly as tertiary treatments in wastewater plants (Ochando-
Pulido and Martinez-Ferez, 2012; Cassano et al., 2011; Di Lecce et al., Cp, i ⎞
2014; Garcia-Castello et al., 2010; Ochando-Pulido et al., 2013; Stoller, Ri (%) = ⎜⎛1 − ⎟ x 100
⎝ Cf , i ⎠ (1)
2011). Membrane processes can be technically and economically effi-
cient if the productivity (flux) is enhanced and fouling is strongly in- where cp,i is the concentration of the solute i (mg L−1) in the permeate,
hibited. Moreover, given that the product, that is, purified water, is not
a priori of high added value, the recovery of added-value compounds Table 1
can counter-balance the economic feasibility of the effluent reclamation Specifications of the used membranes.
process. In this regard, the selection of the adequate membrane upon
Feature Parametric value
the proposed procedure also puts a focus on membrane fouling. To the
authors' knowledge, the combination of membrane filtration and cen- Membrane type MF-UF-loose NF
trifugation is a novel methodology not yet investigated until now for Provider Millipore
Model Amicon Ultra-15
purification and revalorization of these agro-industrial effluents, tar-
Material Regenerated cellulose
geted to membrane treatment screening for recovery of added-value MWCO, kDa 3 - 10–30 - 50 - 100
compounds.
∗
MWCO: molecular weight cut off.

448
J.M. Ochando-Pulido, A. Martinez-Ferez
Fig. 1. Membranes system contained in Falcons used.
and cf,i is the concentration of the solute i (mg L−1) in the feed.
At the end of each experimental run, the membranes were subjected
to a cleaning-in-place procedure, adapted from former research by the
authors (Ochando-Pulido et al., 2015a), with NaOH and sodium do-
decyl sulfate (SDS) (Panreac S.A.), to recover them for the next ex-
periments.
The cleaning sequence comprised the following steps: i) the corre-
sponding membrane, contained in its Falcon tube, was primary rinsed
to remove superficial fouling, by filling the tube with deionized water
and putting it into the centrifuge, at 4000 rpm for 3 min; ii) then,
deionized water was removed and 0.075–0.1 w/w % SDS + NaOH
solution was poured into the Falcon, then introduced in the centrifuge
for soft chemical cleaning of the membrane, at 4000 rpm for 3 min; iii)
finally, secondary rinsing of the corresponding membrane was carried
out, under the same conditions as the primary rinsing, to remove the
rests of the cleaning reagents used.
The COD of the permeate was measured at the end of the sequence,
to assure the effectiveness of the membranes cleaning procedure.
2.3. Estimation of membrane fouling
To evaluate the quantity of organic fouling on the used membranes
as a function of the membrane MWCO, a global mass balance to the
COD was carried out, as expressed in the following equation:
Mfouling = vc ·CODc − vp·CODp (2)
where vc is the volume of concentrate (mL), vp is the volume of
permeate, CODc is concentrate COD (mg L−1) and CODp is the permeate
COD (mg L−1). The quantity of organic fouling on the membrane
(Mfouling) corresponds to the difference of the organic matter quantity in
the permeate and that in the concentrate, which means that the organic
matter which is nor in the concentrate nor in the permeate volume is
that built-up on the membrane as fouling.
2.4. Analytical methods
Analyses of chemical oxygen demand (COD), total suspended solids
(TSS), total phenolic compounds (TPhs), electrical conductivity (EC)
449
Journal of Environmental Management 222 (2018) 447–453
and pH were carried out in both the raw effluent (OMW2) as well as in
the permeate and concentrate streams of the membranes, following
standard methods (Greenberg et al., 2005). All analytical methods were
performed with analytical grade reagents (minimum 99% purity) in
triplicate.
A Helios Gamma UV–-visible spectrophotometer (Thermo Fisher
Scientific) was used for the COD and TPhs analysis. EC and pH were
measured with a Crison GLP31 conductivity-meter and a Crison GLP21
pH-meter, provided with autocorrection of temperature (25 °C), and
calibrated with buffer standard solutions (Crison) for EC (1413 μS cm−1
and 12.88 mS cm−1) and pH (4.01, 7.00 and 9.21).
The total phenolic fraction concentration (TPhs, mg/L of equivalent
gallic acid, GAE) was measured following Folin-Cicolteau method
(Greenberg et al., 2005), with Folin-Cicolteau reagent (Sigma-Aldrich).
This reagent contains sodic molybdate and tungstate, which at basic pH
react with any type of phenol: the transfer of electrons reduces the
phosphomolybdic-phosphotungstic complexes in oxides of tungsten
(W8O23) and of molybdenum (Mo8O23), of intense blue color, which is
proportional to the number of hydroxyl groups, then measured at
765 nm in the UV–-visible spectrophotometer (Thermo Fisher Scien-
tific).
3. Results and discussion
The physicochemical characterization of the raw OMW2 is reported
in Table 2. A considerable and phytotoxic but potentially valuable
concentration of phenolic compounds, 770.0 ± 21.2 mg L−1, was
quantified in the effluent from the vertical centrifuges of the two-phase
olive oil production process. On another hand, the pH of the effluent
was confirmed to be slightly acid - mainly owed to the content of or-
ganic acids in the raw OMW2 effluent (Niaounakis and Halvadakis,
2006; Hodaifa et al., 2013) - and presented a high COD, in the order of
13.9 ± 0.6 g L−1.
The olive fruit is very rich in phenolic compounds, but only 2% of its
total phenolic content is transferred to the oil phase, while the re-
maining amount is lost in the wastewater stream (Obied et al., 2005).
Because of their excellent antioxidant properties, phenolic compounds
are greatly demanded in cosmetic, food, pharmaceutical and bio-
technological industrial sectors, being this reflected by their very high
market price. In addition, up to now these compounds have been
commonly synthesized by chemical methods, which has contributed to
considerably rise their price.
Consequently, the development of methods to concentrate and
purify the polyphenolic content of these effluents has arisen much in-
terest in the recent years. What is more, if phenolic compounds are left
untreated in these wastewaters, they become gradually oxidized and/or
polymerized, rendering them highly phytotoxic and recalcitrant to
biological degradation (Paraskeva and Diamadopoulos, 2006; Ochando-
Pulido et al., 2017a; Hodaifa et al., 2013; Sampaio et al., 2011). Hence,
their recovery from these effluents can provide not only economic
benefits but also help reduce their phytotoxicity and the consequent
sustainability of their reclamation management.
Subsequently, OMW2 samples were conducted to the membrane
screening process proposed, with membranes exhibiting MWCO
Table 2
Raw OMW2 physicochemical characterization.
Parameter Value
pH 5.0 ± 0.1
EC, mS cm−1 1.9 ± 0.2
TSS, g L−1 3.6 ± 0.2
COD, g L−1 13.9 ± 0.6
Total phenols, mg L−1 770.0 ± 21.2
∗
EC: electrical conductivity; TSS: total suspended solids; COD:
chemical oxygen demand.
J.M. Ochando-Pulido, A. Martinez-Ferez
Fig. 2. Rejection of total phenolic compounds as a function of the membrane
MWCO (100 - 3 kDa); % standard deviation σs (I).
ranging from 100 down to 3 kDa, following the procedure formerly
described. After this, TPhs concentration, COD, EC and pH of the
permeate and concentrate centrifuged-filtered samples obtained were
analyzed.
The design and operation of membrane purification plants is a key
task, and the development of quick and reliable methods for the as-
sessment of the feasibility of the proposed membrane processes is a
need for this industry (Espinasse et al., 2002; Stoller et al., 2013). The
rejection of the tested membranes towards OMW2 phenolic fraction as
a function of the membrane MWCO is shown in Fig. 2.
The rejection performance of the tested membranes was found to be
linear (RTPhs = - 0.2624·MWCO + 56.189; R2 = 0.987) as a function of
the membrane MWCO. This implied that the rejection of the total
phenols concentration could be increased from around 30.3% for a
membrane with a MWCO of 100 kDa, up to 57.2% for 3 kDa (see Fig. 2).
A focus on the obtained results permits noting that the selection of a
membrane in the range of UF (50 - 10 kDa) would lead to the con-
centration of 43.2% (50 kDa) and up to 52.6% (10 kDa) of OMW2
phenolic fraction, whereas the selection of a membrane in the range of
MF (100 kDa) would permit concentrating minimum 30% of the phe-
nolic content. The selection of a membrane in the near range of loose
NF (3 kDa) led to 57.2% concentration of the phenolic fraction.
On the contrary, the selection of a 0.45 μm MF membrane led to
negative rejection towards the phenolic fraction (−-19.7%). This
means that the phenolic compounds are entirely transferred into the
permeate stream, where their concentration becomes incremented as a
result of the volume reduction, given that certain feed volume is kept in
the concentrate side (Fig. 3). Otherwise as observed for the rest of the
membranes pore sizes examined, the selection of this Dp would permit
the transference of the totality of the phenolic fraction to the permeate
instead of the concentrate stream, another possible option layout,
leading to 19.7% concentration increase.
In Fig. 3, the resulting concentrate and permeate samples are
shown. The concentrate was a dark stream, disregarding the MWCO of
the membrane selected, whereas the permeate was increasingly clearer
as the MWCO of the membrane was narrowed from 100 to 3 kDa. The
recovery of the phenolic fraction from these effluents by this membrane
procedure, a green and environmentally friendly technology which
does not imply the use of reagents, contrasts with extraction methods
based on solvents, widely proposed to concentrate and purify the
polyphenolic content of these effluents (Galanakis et al., 2011; Al-
Asheh et al., 2003).
Summing up, taking in regard that the segregation of the phenolic
content between the permeate and concentrate streams is not of interest
beforehand, but to concentrate it as much as possible in one stream or
the other, the obtained results pin-point for the selection of a membrane
450
Journal of Environmental Management 222 (2018) 447–453
Fig. 3. Concentrate (dark) and permeate (brownie-light) fractions of OMW2
samples subjected to membrane screening (MWCO 100 - 3 kDa).
of MWCO between 100 kDa and 0.45 μm to transfer most phenolic
fraction in the permeate, or below 3 kDa to concentrate them in the
concentrate stream instead.
On the other hand, the results of the quantification of the permeate
vs. concentrate volume recovery as a function of the membrane MWCO
is given in Fig. 4. The obtained results showed an increment of the
permeate volume collected with the membrane MWCO, resulting from
the major mean pore Dp of the membranes, obtaining minimum 16.3%
w/w of permeate (3 kDa) and maximum 28.9% upon MWCO of
100 kDa. It was found that the quantity of the permeate stream could be
augmented from 4.9 g to 8.7 g, that is, a 43.7% increment, when the
membrane MWCO increased from 3 up to 100 kDa.
The volume concentration factor (VCF), that is, the ratio of final
concentrate stream volume to initial feed effluent volume, was calcu-
lated as follows:
Vinitial
VCF =
Vconcentrate (3)
To sum up, the selection of a membrane in the range of UF
(10–50 kDa) or loose NF (3 kDa) would lead to VCF of ∼1.28–1.37 and
1.21, respectively, whereas in the range of MF (100 kDa) a VCF of 1.50
could be provided.
In other words, the permeate volume upon the selection of a
membrane of 10–50 kDa (4.96–6.98 g), in the range of UF, could be up
to 29.9% higher (4.89 g) in contrast with the membrane of the lowest
size, 3 kDa (loose NF), and the selection of a membrane of 100 kDa (MF
range) would lead to 19.4–42.7% higher permeate volume than those in
the range of UF (10–50 kDa).
Fig. 4. Permeate quantity recovered as a function of the membrane MWCO
(100 - 3 kDa); % standard deviation σs (I).
J.M. Ochando-Pulido, A. Martinez-Ferez
Fig. 5. COD rejection behaviour of the selected membranes as function of
MWCO (100 - 3 kDa); % standard deviation σs (I).
Otherwise, the behaviour of the selected micro/ultra/nanocen-
trifugation membranes with respect to the organic matter in the OMW2
feedstream, measured as COD, is shown in Fig. 5.
Similar as observed for the phenolic fraction rejection pattern of the
examined membranes, the COD rejection (RCOD) was found to increase
linearly as a function of the membrane MWCO (RCOD = -
0.5525·MWCO + 60.3; R2 = 0.994). Hence, the RCOD efficiency could
be incremented from the minimum yielded by the 100 kDa membrane
(6.23%) up to the maximum provided by the tightest membrane, 3 kDa
(59.95%), which means a RCOD enhancement of up to 53.7% (Fig. 5).
On another hand, according to the results, the selection of a mem-
brane ranging from 100 kDa to 0.45 μm would enable the recovery of
the phenolic fraction without loss. The use of a membrane in the range
of MF, with a mean pore size of 0.45 μm, permitted a concentration
increase of 19.7% of the phenolic fraction in the permeate stream, a
reduction of 3.3% of the COD and the complete rejection (100%) of the
suspended solids present in the effluent. This would mean the obtention
of a permeate stream exempt of suspended solids and phenolic com-
pounds. Otherwise, the selection of a 100 kDa membrane would lead to
∼70% recovery of TPhs in the permeate, whereas, on the contrary,
upon choosing a membrane in the range of 3 kDa, ∼57% TPhs could be
recovered in the concentrate stream.
In addition, the pH of the permeate and concentrate fractions of all
membranes were measured (Table 3). The permeate pH was found to be
incremented as the membrane MWCO decreased, as well as in the
concentrate volume. Moreover, the pH values in the permeate samples
were in the range or above those in the raw OMW2 stream, whereas in
the concentrate samples the pH values were consistently below the raw
stream pH. This fact can be explained by the content in organic acids of
the raw OMW2 effluent, among which there are long-chain ones, not
being transferred across the membranes to the permeate side, but being
retained in the concentrated and therefore leading to its acidification.
The measured EC values in the OMW2 effluent, 1.9 ± 0.2 mS cm−1
(Table 2), presented risk of high salinity according to the standards
established by the Food and Agricultural Association (FAO) for reusing
Table 3
Permeate and concentrate pH as a function of membrane pore size.
Membrane size Permeate pH Concentrate pH
0.45 μm 4.95 4.60
100 kDa 4.96 4.68
50 kDa 4.98 4.68
30 kDa 5.09 4.69
10 kDa 5.09 4.86
3 kDa 5.15 4.90
451
Journal of Environmental Management 222 (2018) 447–453
Fig. 6. EC rejection (REC, %) as a function of the membrane MWCO; % standard
deviation σs (I).
the regenerated water for irrigation (Sancho Cierva and Giner, 2000).
Among them, inorganic compounds including chloride, sulfate and
phosphoric salts of potassium, calcium, iron, magnesium, sodium,
copper and traces of other elements are common traits of OMW2
(Niaounakis and Halvadakis, 2006; Paraskeva and Diamadopoulos,
2006). In Fig. 6, the EC rejection (REC, %) as a function of the mem-
brane MWCO is reported.
As it can be seen, REC increased upon lower MWCO, up to a max-
imum of 66% ensured by the 3 kDa membrane. However, as observed in
Fig. 6, there is a key point that divides the REC behaviour as a function
of the MWCO of the membranes: whereas upon a MWCO equal to or
above 50 kDa the measured REC was practically negligible, there was a
MWCO (30 kDa) from which the REC efficiency began to be significant
(5.9%). REC further increased considerably, up to more than ten times,
upon a MWCO beyond the latter, that is, 59.1% for a MWCO of 10 kDa,
which could be incremented up to a maximum of 66% for a 3 kDa
membrane MWCO. Otherwise, negative REC (- 5.2%) was observed for a
membrane Dp of 0.45 μm (Table 4), which means the selection of this
membrane mean pore size would enable the passage of the salts in the
raw OMW2 to the permeate side.
By using 3–10 kDa membranes, the EC in the permeate stream could
be lowered to values ranging 551–662 μS cm−1, thus sensibly im-
proving the quality of the effluent with respect to the saline con-
centration. This would permit the obtention of a purified effluent with
good salinity standards according to the indications given by the FAO to
reuse it for irrigation (Sancho Cierva and Giner, 2000).
Finally, estimation of the organic fouling build-up on the selected
membranes was carried out. Inhibition or minimization of membrane
fouling, as far as it does make the process feasible, determines the
success of membrane plants design and operation control.
A range of factors affect significantly the performance of mem-
branes, which can be gathered in hydrodynamics (operating condi-
tions), feed stream composition, temperature, and membrane surface
characteristics. The latter, concretely the membrane mean pore size,
has been highlighted as a critical feature for MF and UF membranes
performance (Field and Pearce, 2011; Le-Clech et al., 2006; Stoller
Table 4
EC rejection (REC, %) and permeate EC as a function of membrane pore size.
Membrane size Permeate EC, μS cm−1 REC, %
0.45 μm 1618 - 5.2
100 kDa 1606 0.8
50 kDa 1589 1.8
30 kDa 1522 5.9
10 kDa 662 59.1
3 kDa 551 66.0
J.M. Ochando-Pulido, A. Martinez-Ferez
Fig. 7. Organic fouling (Mfouling) measured as a function of the membrane
MWCO.
et al., 2015). Commonly, as stated by some authors, MF and UF
membranes are very prone to suffer from fouling, which reduces their
effectiveness, selectivity and productivity, stabilisingstablishing a han-
dicap for the technical and economic feasibility (Field and Pearce,
2011; Le-Clech et al., 2006; Stoller et al., 2015; de Barros et al., 2003;
Vincent-Vela et al., 2011; Wang and Tarabara, 2008; Akdemir and Özer,
2009).
The results of the estimation of the organic fouling (Mfouling) built-
up on the selected membranes as a function of the MWCO is reported in
Fig. 7. According to the results, no significant fouling was attained on
the used membranes of 3–10 kDa MWCO, and also no considerable
Mfouling was measured on the 100 kDa membrane. The latter can be
explained by the fact of the pores (Dp) for this membrane (100 kDa)
being bigger than the mean size of the particles (dp) present in the ef-
fluent, leading them to pass across the membrane without causing
measurable fouling (see Fig. 8). These is supported by the rejection
efficiencies (RCOD and REC) formerly measured. On the other hand,
Mfouling was negligible for 3–10 kDa MWCO, given that the minor Dp of
these membranes permits the effective rejection of particles of dp > Dp,
also confirmed by the rejection efficiencies previously reported, such
that these solutes can be easily swept from the membrane surface
during filtration without causing fouling (see Fig. 8), being kept in the
concentrate volume.
Nevertheless, certain Mfouling was measured for the membranes of
30–50 kDa MWCO (Fig. 7), similar in both cases. This pin-points for a
critical MWCO value, from the point of view of the selected membrane
and feedstream binomial, comprehended between these two values, due
to the similarity between the membrane mean Dp and the average
particles dp in the OMW2 feedstream (Fig. 8). This was pin-pointed by
Stoller's studies on particles size distribution effect after flocculation on
different membranes performances (Stoller, 2009; Stoller and Serrão
Fig. 8. Pore blocking mechanisms as a function of the membrane mean dia-
meter and particle size (adapted from (Stoller, 2009)): dp/Dp < 0.1 (left cap-
tion), 0.1 < dp/Dp < 10 (central caption) and dp/Dp > 0.1 (right caption).
452
Journal of Environmental Management 222 (2018) 447–453
Mendes, 2017).
Stoller et al. analyzed how fouling, among other factors, is a func-
tion of particle size and concentration (Stoller, 2009; Stoller and Serrão
Mendes, 2017), relating the particle size distribution to the boundary
flux (Stoller and Ochando Pulido, 2015). A rule of thumb was given,
that is particles with a size of 1/10 to 10 times the pore size, are those
affecting significantly pore blocking and therefore fouling (Fig. 8). MF
pores are in the size range of many macromolecules in many industries
concerning biotech, food and manufacturing. NF membranes have
smaller pores, thus even if characterized by lower permeability be-
forehand, they may over-perform UF if concentration and size of the
molecules in the feedstream are outside the dangerous range. This is
especially true if molecules agglomerate. As soon as the agglomerates
are formed, they will grow in size and therefore may reach the dan-
gerous size for the membrane, whereas in case of membranes ap-
proaching tight UF and NF range, bigger molecules may hinder gel
formation by steric means, since larger particles therefore will not affect
the membrane pores. In other words, plugging or sealing of the mem-
branes pores is promoted when the effluent has a significant con-
centration of particles of dp of similar range to that of the membrane Dp
(de Barros et al., 2003; Vincent-Vela et al., 2011; Wang and Tarabara,
2008; Akdemir and Özer, 2009; Stoller, 2009; Stoller and Serrão
Mendes, 2017; Stoller and Ochando Pulido, 2015; Ochando-Pulido
et al., 2015b, 2016).
For UF membranes, as the ones used in this work, too high permeate
fluxes crossing the membrane, resulting in yield of local flux recovery
values so high that lead the concentrate stream on top of the membrane
to overcome the gel concentration, can cause severe gelification. The
gel starts to deposit over the surface and to cover the pores, resulting in
deleterious fouling. The passage from the gel state to irreversible
fouling is short, thus in the long run the membrane will definitively
loose its performance. This phenomenon is more likely to happen on UF
(Stoller, 2009, 2011).
This is graphically depicted in Fig. 8. Molecules presenting a size in
the near range of the membrane's pores (10 < dp/Dp < 10) have the
probability whether to settle on the pores or membrane defects, or on
the membrane layer, such that the pores may be blocked (Fig. 8, central
caption), and it may be possible as well that some solutes settle over
others previously deposited (Field and Pearce, 2011; Le-Clech et al.,
2006; Stoller et al., 2015; Stoller, 2009; Stoller and Serrão Mendes,
2017; Stoller and Ochando Pulido, 2015). Every position of the mem-
brane surface presents the same probability to be fouled, determined by
the ratio of free spaces on the membrane surface to the ones that have
been fouled already (Field and Pearce, 2011; Vincent-Vela et al., 2011;
Wang and Tarabara, 2008).
This novel micro/ultra/nanocentrifugation membrane process as-
sessment proposed, aimed to gather information for a correct, effective
and reliable screening procedure for the election of the adequate
membrane for the target purification purpose, in this case the revalor-
ization and reclamation of OMW2, can be reliable and sensibly quicker
than time-consuming common bench-scale membrane experiments.
4. Conclusions
In this work, concentration and recovery of high-added value
compounds (polyphenolic fraction) from two-phase OMW and its si-
multaneous treatment by a novel micro/ultra/nanocentrifugation
membrane process assessment is addressed, gathering information for a
correct, effective and reliable screening procedure for the adequate
membrane selection for the target purpose. To the authors' knowledge,
the combination of membrane filtration and centrifugation is a novel
methodology not yet investigated until now for purification and re-
valorization of these agro-industrial effluents, targeted to membrane
treatment screening for recovery of added-value compounds.
Phenolic compounds are one of the major factors of the environ-
mental problems caused by these effluents, but on the other hand
J.M. Ochando-Pulido, A. Martinez-Ferez
present antioxidant properties that makes them very relevant for food,
cosmetic, pharmaceutical and biotechnological industries.
The obtained results pin-point for the selection of a membrane of
MWCO between 100 kDa and 0.45 μm to transfer most of the phenolic
fraction in the permeate, whereas below 3 kDa to confine them in the
concentrate stream instead, whilst the EC in the final permeate would
be lowered to 551–662 μS cm−1, sensibly improving the quality of the
feedstream saline concentration, as well as up to ∼60% COD reduction.
This would permit the obtention of a purified effluent with good salinity
standards according to the indications given by the FAO for irrigation
reuse.
The proposed procedure can be quick and reliable for a primary
assessment of the adequate membrane needed for a particular pur-
ification process. This fact is quite significant, given that it can aid in
the selection of the key membrane pore size by quick experiments, to
make a first screening analysis, in contrast with the long-term, time
consuming bench-scale membrane experiments.
Acknowledgements
Spanish Ministry of Economy and Competitiveness is acknowledged
for having funded the project CTM 2014-61105-JIN ‘Design and de-
velopment of an integral process for revalorization and treatment of the
effluents from olive oil industry’, as well as the University of Granada
and J.R. Corpas for his support during the performance of the experi-
ments.
References
Abdel-Shafy, H.I., Schories, G., Mohamed-Mansour, M.S., Bordei, V., 2015. Integrated
membranes for the recovery and concentration of antioxidant from olive mill was-
tewater. Desal. Water Treat. 56 (2), 305–314.
Akdemir, E.O., Özer, A., 2009. Investigation of two ultrafiltration membranes for treat-
ment of olive oil mill wastewater. Desalination 249, 660–666.
Al-Asheh, S., Banat, F., Abu-Aitah, L., 2003. Adsorption of phenol using different types of
activated bentonites. Separ. Purif. Technol. 33, 1–10.
Bazzarelli, F., Piacentini, E., Poerio, T., Mazzei, R., Cassano, A., Giorno, L., 2016.
Advances in membrane operations for water purification and biophenols recovery/
valorization from OMWWs. J. Membr. Sci. 497, 402–409.
Boskou, D., 2006. Sources of natural phenolic antioxidants. Trends Food Sci. Technol. 17,
505–512.
Cassano, A., Conidi, C., Drioli, E., 2011. Comparison of the performance of UF membranes
in olive mill wastewaters treatment. Water Res. 45, 3197–3204.
Conidi, C., Mazzei, R., Cassano, A., Giorno, L., 2014. Integrated membrane system for
production of phytotherapics from olive mill wastewaters. J. Membr. Sci. 454,
322–329.
de Barros, S.T.D., Andrade, C.M.G., Mendes, E.S., Peres, L., 2003. Study of fouling me-
chanism in pineapple juice clarification by ultrafiltration. J. Membr. Sci. 215,
213–224.
Di Lecce, G., Cassano, A., Bendini, A., Conidi, C., Giorno, L., Gallina-Toschi, T., 2014.
Characterization of olivemill wastewater fractions treatment by integrated membrane
process. J. Sci. Food Agric. 94, 2935–2942.
Di Palma, L., Merli, C., Paris, M., Petrucci, E., 2003. A steady-state model for the eva-
luation of disk rotational speed influence on RBC kinetic: model presentation. Bio
Technol. 86 (2), 193–200. https://doi.org/10.1016/S0960-8524(02)00137-2.
Espinasse, B., Bacchin, P., Aimar, P., 2002. On an experimental method to measure cri-
tical flux in ultrafiltration. Desalination 146, 91–96.
ec.europa.eu 2020 https://ec.europa.eu/programmes/horizon2020/.
Field, R.W., Pearce, G.K., 2011. Critical, sustainable and threshold fluxes for membrane
filtration with water industry applications. Adv. Colloid Interface Sci. 164, 38–44.
Galanakis, C.M., Goulas, V., Tsakona, S., Manganaris, G.A., Gekas, V., 2011. A Knowledge
base for the recovery of natural phenols with different solvents. Int. J. Food Prop. 16,
382–396.
Garcia-Castello, E., Cassano, A., Criscuoli, A., Conidi, C., Drioli, E., 2010. Recovery and
concentration of polyphenols from olive mill wastewaters by integrated membrane
system. Water Res. 44, 3883–3892.
Greenberg, A.E., Clesceri, L.S., Eaton, A.D., 2005. Standard Methods for the Examination
of Water and Wastewater, twenty-second ed. APHA/AWWA/WEF, Washington DC.
Cabs.
Hodaifa, G., Ochando-Pulido, J.M., Rodriguez-Vives, S., Martinez-Ferez, A., 2013.
453
Journal of Environmental Management 222 (2018) 447–453
Optimization of continuous reactor at pilot scale for olive-oil mill wastewater treat-
ment by Fenton-like process. Chem. Eng. J. 220, 117–124.
International Olive Oil Council, IOOC, 2015-2016. http://www.internationaloliveoil.org.
Janangiri, M., Rahimpour, A., Nemati, S., Alimohammady, M., 2016. Recovery of poly-
phenols from olive mill wastewater by nanofiltration. Cellul. Chem. Technol. 50
(9–10), 961–966.
Le-Clech, P., Chen, V., Fane, T.A.G., 2006. Fouling in membrane bioreactors used in
wastewater treatment. J. Membr. Sci. 284 (1–2), 17–53.
Niaounakis, M., Halvadakis, C.P., 2006. second ed. Olive Processing Waste Management
Literature Review and Patent Survey, vol. 5 Elsevier: Waste Management Series.
Obied, H.K., Allen, M.S., Bedgood, D.R., Prenzler, P.D., Robards, K., Stockmann, R., 2005.
Bioactivity and analysis of biophenols recovered from olive mill waste. J. Agric. Food
Chem. 53, 823–837.
Ochando-Pulido, J.M., Martinez-Ferez, A., 2012. A focus on pressure-driven membrane
technology in olive mill wastewater reclamation: state of the art. Water Sci. Technol.
66 (12), 2505–2516.
Ochando-Pulido, J.M., Hodaifa, G., Victor-Ortega, M.D., Rodriguez-Vives, S., Martinez-
Ferez, A., 2013. Reuse of olive mill effluents from two-phase extraction process by
integrated advanced oxidation and reverse osmosis treatment. J. Hazard Mater. 263
(1), 158–167.
Ochando-Pulido, J.M., Victor-Ortega, M.D., Martinez-Ferez, A., 2015a. On the cleaning
procedure of a hydrophilic reverse osmosis membrane fouled by secondary-treated
olive mill wastewater. Chem. Eng. J. 260, 142–151.
Ochando-Pulido, J.M., Victor-Ortega, M.D., Martinez-Ferez, A., 2015b. Physicochemical
analysis and adequation of OMW after AOP for reclamation by pressure-driven
membrane technology. Sci. Total Environ. 503–504, 113–121.
Ochando-Pulido, J.M., Víctor-Ortega, M.D., Martínez-Ferez, A., 2016. Membrane fouling
insight during reverse osmosis purification of pretreated olive mill wastewater. Separ.
Purif. Technol. 168, 177–187.
Ochando-Pulido, J.M., Pimentel-Moral, S., Verardo, V., Martinez-Ferez, A., 2017a. A
focus on advanced physico-chemical processes for olive mill wastewater treatment.
Separ. Purif. Technol. 179, 161–174.
Ochando-Pulido, J.M., Corpas-Martínez, J.R., Martinez-Ferez, A., 2017b. About two-
phase olive oil washing wastewater simultaneous phenols recovery and treatment by
nanofiltration. Process Saf. Environ. Protect (in press). https://doi.org/10.1016/j.
psep. 2017. 12.005.
Paraskeva, P., Diamadopoulos, E., 2006. Technologies for olive mill wastewater (OMW)
treatment: a review. J. Chem. Technol. Biotechnol. 81, 475–1485.
Russo, C., 2007. A new membrane process for the selective fractionation and total re-
covery of polyphenols, water and organic substances from vegetation waters (VW). J.
Membr. Sci. 288, 239–246.
Sampaio, M.A., Gonçalves, M.R., Marques, I.P., 2011. Anaerobic digestion challenge of
raw olive mill wastewater. Bio Technol. 102, 10810–10818.
Sancho Cierva, J., Giner, J.F., 2000. Water Quality for Irrigation Use. Universidad
Politécnica de Valencia-Generalitat Valenciana-Phytoma, Valencia.
Stoller, M., 2009. On the effect of flocculation as pretreatment process and particle size
distribution for membrane fouling reduction. Desalination 240, 209–217.
Stoller, M., 2011. Effective fouling inhibition by critical flux based optimization methods
on a NF membrane module for olive mill wastewater treatment. Chem. Eng. J. 168,
1140–1148.
Stoller, M., Ochando Pulido, J.M., 2015. The Boundary Flux Handbook: a Comprehensive
Database of Critical and Threshold Flux Values for Membrane Practitioners. Elsevier,
Netherlands.
Stoller, M., Serrão Mendes, R., 2017. Advanced control system for membrane processes
based on the boundary flux model. Separ. Purif. Technol. 175, 527–535.
Stoller, M., Gonzini, O., Di Palma, L., 2010. Enhancement of biomass speciation in Step
Sludge Recirculation activated sludge system by the use of submerged membranes.
Desalination 250 (2), 583–586.
Stoller, M., De Caprariis, B., Cicci, A., Verdone, N., Bravi, M., Chianese, A., 2013. About
proper membrane process design affected by fouling by means of the analysis of
measured threshold flux data. Separ. Purif. Technol. 114, 83–89.
Stoller, M., Ochando-Pulido, J.M., Di Palma, L., Martínez-Férez, A., 2015. Membrane
process enhancement of 2-phase and 3-phase olive mill wastewater treatment plants
by photocatalysis with magnetic-core titanium dioxide nanoparticles. J. Ind. Eng.
Chem. 30, 147–152.
Vilardi, G., Di Palma, L., 2017. Kinetic study of nitrate removal from aqueous solutions
using copper-coated iron nanoparticles. Bull. Environ. Contam. Toxicol. 359–365.
http://dx.doi.org/10.1007/s00128-016-1865-9.
Vilardi, G., Stoller, M., Verdone, N., Di Palma, L., 2017. Production of nano Zero Valent
Iron particles by means of a spinning disk reactor. Chem. Eng. Trans. 57, 751–756.
http://dx.doi.org/10.3303/CET1757126.
Vincent-Vela, M.C., Cuartas-Uribe, B., Álvarez-Blanco, S., Lora-García, J., 2011. Analysis
of fouling resistances under dynamic membrane filtration. Chem. Eng. Process 50,
404–408.
Wang, F., Tarabara, V.V., 2008. Pore blocking mechanisms during early stages of mem-
brane fouling by colloids. J. Colloid Interface Sci. 328, 464–469.
Zagklis, D., Vavouraki, A., Kornaros, M., Paraskeva, C.A., 2015. Purification of olive mill
wastewater phenols through membrane filtration and resin adsorption/desorption. J.
Hazard Mater. 285, 69–76.