Journal of Water Process Engineering: Sciencedirect
Journal of Water Process Engineering: Sciencedirect
Journal of Water Process Engineering: Sciencedirect
A R T I C LE I N FO A B S T R A C T
Keywords: The presence in natural watercourses of different types of emerging contaminants (including stimulants and
Drugs antibiotics) from different sources constitutes a serious environmental issue. The problem is aggravated in re-
Membrane gions where water is scarce and its reuse should form part of its integrated cycle. The removal of these con-
Wastewater treatment plant and reusing taminants, so that the water may be reused, is currently one of the most widely studied issues in the field of
wastewater treatment. Within this context, the present paper addresses the removal of different contaminants
(caffeine, theobromine, theophylline, amoxicillin, and penicillin G). The chosen treatment was the use of reverse
osmosis membranes, which were installed at the outlet of the secondary settling tank of the wastewater treat-
ment plant (WWTP) in Medina Sidonia (SW Spain), carrying out sampling in the feed, reject and permeate
streams over a period of 72 h. DOW-FILMTEC BW30-2540 membranes were used for this purpose, operating at a
pressure of 15 bar. The emerging contaminants were analysed by Ultra Performance Liquid Chromatography-
Mass Spectrometry (UPLC-MS). The applied treatments allowed a contaminant-free permeate to be obtained
with the required quality for its reuse. This technique achieved 100% removal of the contaminants without signs
of membrane deterioration after running the 72 h test.
⁎
Corresponding author.
E-mail address: [email protected] (A. Egea-Corbacho Lopera).
https://doi.org/10.1016/j.jwpe.2019.100800
Received 19 November 2018; Received in revised form 2 March 2019; Accepted 9 March 2019
Available online 27 March 2019
2214-7144/ © 2019 Elsevier Ltd. All rights reserved.
A. Egea-Corbacho Lopera, et al. Journal of Water Process Engineering 29 (2019) 100800
2
A. Egea-Corbacho Lopera, et al. Journal of Water Process Engineering 29 (2019) 100800
treatment plant. A further aim was to study the removal capacity of this necessary to concentrate them by means of solid phase extraction. To
technology at concentrations higher than or equal to those usually optimize the operating conditions and the type of columns used, a series
observed at the WWTP. of trials were carried out following the procedures as used on a previous
For this purpose, two pressure pumps located in the secondary study [26]. Acetonitrile (ACN) and methanol (MeOH) were tested as
settling tanks of the WWTP were used to pump the water through a eluting agents. The best results were obtained with mini-columns
200 μm self-cleaning filter (COPERSA, model 85102M-p-DC) for its pre- equipped with 60 mg solid phase Oasis HLB cartridges, supplied by
treatment. Once pre-treated, the water is sent (750 L·h−1 flow rate) to Waters, using methanol as the eluyent.
the feed tank (350 L) of the reverse osmosis unit, where it is mixed with The column was activated using 10 mL methanol and 10 mL water
a solution of the pharmaceutical drugs under study. The feed tank at a flow rate of 1 mL·min−1. Then, 100 mL of sample was passed
worked at all times in open circuit. This solution was prepared by di- through the mini-column at the same flow rate with the aim of retaining
luting 56 g of each of the contaminants in a mixing tank with 400 L of the studied compounds. The retained compounds were desorbed and
wastewater. The flow rate of the solution of pharmaceuticals into the eluted again using 10 mL methanol in a test tube at a flow rate of
reverse osmosis unit feed tank was 5 L·h−1. This high concentration in 1 mL·min−1. Finally, the extract was evaporated with liquid nitrogen at
the mixing tank is due to the fact that through the small flow of the 40 °C and the dry residue dissolved in 1 mL water. The samples were
peristaltic pump it is possible to contaminate the water for the study of filtered through a 0.22 μm filter prior to the entire extraction analysis,
the feed water. Fig. 2 shows the flow chart in open-circuit working Nylonflo™47 mm, provided by Pall Corporation, Michigan.
conditions
As a preventative measure, the feed line from the feed tank is
2.6. Analytical techniques
equipped with a pre-filter consisting of a 5 μm mesh size polypropylene
thread cartridge. This ensures a sufficient quality of water before it
For the analysis of emerging contaminants an ultra-performance
enters the reverse osmosis unit, thereby extending the useful life of the
liquid chromatography (UPLC) equipment was used. The UPLC was
unit. The necessary pressure to pump the feed stream through the
coupled to time-of-flight mass spectrometry was used for the identifi-
membrane unit is achieved by means of a 075 kW (1HP) centrifugal
cation quantification of analytes (Xevo, Waters Synapt G2). To identify
pump that provides a flow rate of 700-750 L·h−1. Furthermore, the
emerging contaminats was compared retention times and accurate mass
plant is equipped with a 1.49 kW (2 HP) pistons pump capable of
measurements (allowing an error of less than 5 mg L-1) to those for
raising the pressure to 50 bars.
commercially available pure standards.
The concentration of each contaminant in the influent mixture to
MassLynx 4.1 software was employed to analize the compounds.
the reverse osmosis unit varied, ranging from 0.177-0.557 mg L−1 for
The noise type selected was root mean square (RMS). Limits of detec-
caffeine, 0.101-0.534 mg L−1 for theobromine, 0.135-0.641 mg L−1 for
tion were calculated for the quantifier transition. Instrumental limits of
theophylline, < LODs (Limit of detection)-0.164 mg L-1 for amoxicillin
detection (iLOD) and quantitation (iLOQ) for each target compound
and < LODs-0.213 mg L−1 for penicillin G. Similar amounts have been
were calculated based on the signal to noise ratio of 3 (iLODs) and a
found for the family of antibiotics in influents to the Gulf of Cadiz
signal-to-noise ratio of 10 (iLOQs) near the target peak by using the
(Spain) by other authors [24,25]. The open-circuit process operated for
lowest standard solution.
72 h, with samples being taken at 1, 2, 4, 12, 24, 48 and 72 h. An au-
Quantification of target compounds was performed using calibra-
topsy of the membrane was performed to determine its integrity, the
tion curves (from 1·10−3 to 1 mg L-1), prepared in water in 1 mL vials. A
accumulated fouling and the pollutants that may have been retained in
dilution factor 1:2 of the greatest concentration samples was performed
it, after carrying out the experiments.
to calculate the levels of the analytes. The reproducibility and repeat-
ability of the methods were evaluated by performing three successive
2.5. Solid-phase extraction (SPE) extractions and injections of the same sample and by re-analyzing a
batch of standards two weeks after its first analysis.
The low concentrations of the pollutants under study in the reverse The mass spectrometer equipped with a 50 mm BEH C18 column
osmosis unit influent, effluent and permeate streams meant that it was and 1.7 μm pore size, both from Waters (Milford, MA). A volume of
3
A. Egea-Corbacho Lopera, et al. Journal of Water Process Engineering 29 (2019) 100800
10 μL was injected into the column, using methanol as the eluent. The maximum values detected in most of the influent parameters
The mass spectrum ranged from 50 to 1200 amu. Positive and ne- correspond to a sample taken after one of the rainiest weeks in March,
gative electrosprays (ESI +/-) were used to identify the peaks re- which led to the increase in the flow of water entering the treatment
presenting the studied contaminants. plant. The parameters with the highest standard deviations are che-
The total organic carbon (TOC) and total nitrogen (TN) chromato- mical demand of oxygen, (COD), (416 mgO2·L−1), total solids in sus-
graphy and analyser samples were filtered with Merck 0.2 μm syringe. pension, (TSS), (525 mg L−1), turbidity (374.32 NTU), conductivity
They were measured on a Shimadzu TOC-L (CPH) analyser with an ASI- (2031 mS·cm−1) and hardness (747 mg CaCO3·L−1).
L autosampler and a TN unit. Thermo Elemental IRIS Intrepid ICP-AES The BOD5/COD ratio presents an average value of 0.15. This is
spectrophotometer was used to the meausures of the concentrations of lower than the values usually found in municipal wastewater, which
cations and anions in the fouling. The deposits accumulated on the vary between 0.2-0.5 [31], thus indicating that there are a fraction of
membrane were visualized using a Nova NanoSEM 450 scanning elec- not easily biodegradable constituents [32].
tron microscope (SEM). Microanalysis of the majority chemical was In the case of the pollutants under study, a general decrease is ob-
analize in an energy-dispersive spectrometers (EDS). served with respect to the values found in previous studies. The family
Determination of the other parameters (COD, BOD5, SST, Turbidity, of stimulants was always detected in the WWTP influent, with average
Nitrogen and Total Phosphorus, Conductivity, Hardness, Ammonium, values of 19, 3 and 19 μg·L−1 for caffeine, theobromine and theophyl-
Chlorides, Total Coliforms and Escherichia coli) was carried out in ac- line, respectively. The values for the antibiotics were below 2 μg·L−1
cordance with the Standard Methods for the Examination of Waters and (limit of quantification for the family of antibiotics) or they were below
Wastewater [27]. Samples (not UPLC-MS) were analysed in duplicate or the limit of detection, findings in line with the results provided by other
triplicate, having a maximum standard deviation of ± 10%. authors in the Gulf of Cadiz [25].
The values of all the parameters in the effluent are lower than those
3. Results and discussion obtained in the influent, except for nitrates, which increased from 13.63
to 21.34 mg L−1. The average removal rates obtained for COD, BOD5
3.1. Quality assurance/quality control and TSS were 69%, 91% and 99%, respectively. The increase in nitrates
in the effluent may be due to the oxidation of ammonia (NH3) and
The intra-day variation was evaluated by continuously injecting two nitrite (NO2−) to nitrate (NO3−) by nitrite oxidizing bacteria [33].
replicates within a day (differences observed ranged from 1% to 8%), These findings show that the WWTP achieves a quality effluent that
whereas the inter-day variation was performed by analysing some re- complies with Directive 91/271/EEC [1] regarding discharges.
plicates on different days (differences observed were between 3% and Average values of 4 ( ± 0), 4 ( ± 0) and 30 ( ± 2) μg·L−1 were
10%). The MS response of all compounds was linear between 1·10−3 to quantification for caffeine, theobromine and theophylline in the ef-
1 mg L−1 and coefficients of determination (R2) for calibration curves fluent, respectively. The high concentration of theophylline is worth
were always above 0.9 for target compounds. The limits of detection highlighting, possibly due to the fact that theophylline is a metabolite
(LODs) ranged were 0.001, 0.0003, 0.001, 0.001 and 0.0006 mg L−1 to of caffeine. Hence, a decrease in the concentration of caffeine may in-
caffeine, theobromine theophylline, amoxicillin and penicillin G, re- crease the concentration of theophylline [34]. The family of antibiotics
spectively, and limits of quantitation (LOQs) were 0.004, 0.001, 0.004, was rarely quantified in the samples from Medina Sidonia WWTP, but
0.0014 and 0.002 mg L−1 respectively. they were detected, contrary to what occurs in other larger cities in
The retention time of the compounds was 0.68, 0.44, 0.56, 0.42 and other European countries (France: 0.33-0.51 μg·L−1, Italy: 0.29-
3.38 for caffeine (MW = 195.0882), theobromine (MW = 181.0726), 0.58 μg·L−1, Greece: 0.46 μg·L−1, Sweden: 0.12 μg·L−1 and Switzerland:
theophylline (MW = 181.0726), amoxicillin (MW = 366.1124) and 0.25-0.41 μg·L−1) [35]. Although the concentrations of these sub-
penicillin G (MW = 357.0890) respectively. stances reaching aquatic ecosystems are low, they may pose environ-
The aqueous phase consisted of a 0.1% solution of formic acid (A) mental and human health problems in the receiving waters due to the
and methanol (B) as organic solvent. The flow rate was 0.40 mL·min-1. plant’s limited capacity to degrade them [36,37] or to problems of
The best results were obtained in ESI + mode, with an operating gra- accumulation in the environment.
dient that began with a 70/30% mix of water/methanol for 3 min; the On the other hand, application of the water reuse regulations
water concentration was then decreased until 4:20 min until obtaining a (Table 3) to the WWTP effluent shows that the values for E. coli in the
ratio of 10% A/90% B; subsequently returning to the initial conditions effluent (4.15 CFU 100 mL−1, u.log10) are higher than those of this
(70%/30%) at 4:50 min, and maintaining this ratio until 5 min (overall same parameter for less restrictive uses (4 CFU 100 mL−1, u.log10),
run time = 5 min). thus indicating that its reuse is not possible. This finding, together with
The recovery efficiencies of the mini-columns using methanol to the fact that the emerging contaminants are not removed by the con-
activate the columns and as the eluent were respectively 49%, 43% and ventional treatments employed at the plant, means that more specific
42% for caffeine, theobromine and theophylline, and 10% and 82% for technologies need to be used for their treatment if the aim is to reuse
amoxicillin and penicillin G. Although these values are low, except for this wastewater.
penicillin G, the results coincide with those obtained by the authors of
the reference method [26]. 3.3. Removal of pharmaceutical drugs and stimulants by reverse osmosis
3.2. Characterization of the influent and effluent at Medina sidonia WWTP Before carrying the experiments to separate the contaminants using
membrane technology, it was necessary to check whether the solution
Table 2 shows the average values of the parameters analysed in the of the contaminants underwent a process of biodegradation, photo-
influent and effluent at Medina Sidonia WWTP, as well as those cor- oxidation, adsorption to the walls of the tank, etc., during its time of
responding to the four samplings carried out in the months in which the residence in the mixing tank (72 h) that might result in variations in the
present study was conducted (January-March 2018). During this concentration of the feed to the reverse osmosis unit.
period, appreciable rainfall was registered on 25 days, a rate much Two types of experiments were accordingly performed. In the first, a
higher than the 6.7 days of recent years (1981–2015) [28]. co-solution in ultrapure water of 1 mg L−1 of the contaminants under
In general, if we compare these results with those obtained by the study was prepared and placed in an amber glass flask to ensure com-
same authors in previous studies [29], the wastewater reaching the plete darkness. This experiment would serve as a reference blank ex-
Medina Sidonia treatment plant can still be considered medium-pol- periment. The flask was placed in the same room as the mixing tank to
luted wastewater [30]. ensure the same environmental conditions. In the second experiment,
4
A. Egea-Corbacho Lopera, et al. Journal of Water Process Engineering 29 (2019) 100800
Table 2
Characterization of the influent and effluent of the WWTP of Medina Sidonia. During the months it is the ones that were carried out the experiments (January to
March 2018).
INFLUENT EFFLUENT
MIN MAX AVERAGE S.D. MIN MAX AVERAGE S.D. Real Efficiency (%)
the necessary amount of each antibiotic and stimulant was added to the In the mixing tank, the percentage degradation for the stimulants on
400 L of the mixing tank to achieve a concentration of approximately the third day was respectively 20% and 18% for caffeine and theo-
1 mg L−1 for each pollutant, the maximum concentration of caffeine bromine, and 7% for theophylline. This may be because theophylline is
found in the environment by other authors [38]. The temperature in the a metabolite of caffeine and hence a decrease in the concentration of
glass flask varied between 15 and 25 °C, while the measured tempera- caffeine may lead to an increase in the concentration of theophylline
ture in the mixing tank ranged between 10 and 25 °C. A daily sample [34]. The greatest difference occurred in the family of antibiotics. From
was taken from both the flask and the tank for analysis during 3 days. the first day in the mixing tank, percentage degradations of 52% and
In the glass flask, the degradation percentage on the third day was 46% were obtained for amoxicillin and penicillin G, respectively,
similar for caffeine, theobromine and theophylline (23%, 24% and reaching 89% and 70% on the third day of testing.
23%, respectively, under the best conditions). The results for amox- The results of both experiments confirmed that increasing the re-
icillin and penicillin G were 46% and 37%, respectively. These findings sidence time generally led to an increase in the percentage degradation
show that degradation exists, even under the best conditions, and needs of the solution of contaminants, and that the antibiotics degraded ra-
to be taken into account. pidly due to some type of process of hydrolysis of the drug itself,
Table 3
Summary for uses, parameters to be analysed and maximum permitted. Summary of Spanish Royal Decree 1620/2007.
USES MAXIMUM ADMISSIBLE VALUE (MAV)
MUNICIPAL 1.1 1 0 10 2
1.2 1 2.3 20 10
AGRICULTURAL 2.1 1 2 20 10
2.2 1 3 35 Not established
2.3 1 4 35 Not established
INDUSTRIAL 3.1 A and B Not established 4 35 15
3.1. C 1 3 35 Not established
3.2 1 Absence 5 1
RECREATIONAL 4.1 1 2.3 20 10
4.2 Not established 4 35 Not established
ENVIRONMENTAL 5.1 Not established 3 35 Not established
5.2 1 Absence 10 2
5.3 Not established Not established 35 Not established
5.4 The required minimum quality is to be studied case by case
5
A. Egea-Corbacho Lopera, et al. Journal of Water Process Engineering 29 (2019) 100800
Table 4
Evolution of the concentration of the studied contaminants in the three streams (feed, reject and permeate).
Caffeine [mg·L−1]
Hours 1 2 4 12 24 48 72
Feed 0.456 0.557 0.495 0.213 0.177 0.179 0.272
Reject 0.626 0.767 0.706 0.219 0.248 0.209 0.296
Permeate < LODs < LODs < LODs < LODs < LODs < LODs < LODs
Theobromine [mg·L−1]
Hours 1 2 4 12 24 48 72
Feed 0.433 0.534 0.439 0.202 0.107 0.104 0.345
Reject 0.667 0.745 0.537 0.230 0.147 0.187 0.756
Permeate < LODs < LODs < LODs < LODs < LODs < LODs < LODs
Theophylline [mg·L−1]
Hours 1 2 4 12 24 48 72
Feed 0.429 0.531 0.458 0.266 0.135 0.157 0.641
Reject 0.636 0.701 0.618 0.278 0.190 0.172 0.742
Permeate < LODs < LODs < LODs < LODs < LODs < LODs < LODs
Amoxicillin [mg·L−1]
Hours 1 2 4 12 24 48 72
Feed 0.164 0.163 0.164 0.162 0.062 < LODs < LODs
Reject 0.177 0.164 0.169 0.164 0.103 < LODs < LODs
Permeate < LODs < LODs < LODs < LODs < LODs < LODs < LODs
Penicillin G [mg·L−1]
Hours 1 2 4 12 24 48 72
Feed 0.213 0.155 0.130 0.125 0.038 < LODs < LODs
Reject 0.296 0.208 0.204 0.129 0.052 < LODs < LODs
Permeate < LODs < LODs < LODs < LODs < LODs < LODs < LODs
adsorption to the solids in the water or to the tank walls, biodegrada- separation was obtained for all the compounds under study. These re-
tion processes, etc. [10]. All this led to a decrease in the concentration sults are in agreement with those reported by other authors [40], who
of contaminants reaching the reverse osmosis plant. detected removal rate of over 99% for caffeine in wastewater when
These results also indicate that the compounds under study, espe- using reverse osmosis.
cially the antibiotics, might be degraded during the treatment processes On the other hand, the concentrations in the reject stream are at all
employed at the plant. However, it should be recalled that the hydraulic times higher than the starting concentrations in the feed stream. The
residence times of wastewater in the biological treatment units of lowest concentrations found in the reject stream for caffeine, theo-
WWTPs are considerably lower than those during which the wastewater bromine and theophylline were respectively 0.209 mg L−1 (48 h),
remained in the mixing tank in the present study. In addition, the total 0.230 mg L−1 (12 h) and 0.190 mg L-1 (24 h) and the highest,
mineralization of the compounds could not be confirmed, since inter- 0.706 mg L−1 (4 h), 0.756 mg L−1 (72 h) and 0.742 mg L−1 (72 h). For
mediate metabolites were not quantified, which could become more amoxicillin and penicillin G, the lowest reject concentrations were seen
dangerous than the contaminants studied [39]. This uncertainty is to occur at 24 h (0.103 and 0.052 mg L−1) and the highest, in the first
further accompanied by the higher energy consumption of biological hour of operation (0.177 and 0.296 mg L−1), respectively. Higher
processes. The definitive solution thus lies in the implementation of concentrations in the reject stream for the family of antibiotics are
advanced treatment systems that are capable of removing these con- normal during the first hours of operation due to the decrease in con-
taminants from wastewater and enable its reuse [40]. centration in the feed stream.
After determining the way that the contaminant solutions might In view of these results, it can be stated that the DOW FILMTEC
behave during the 72 h that the test lasted, a series of experiments were BW30-2540 membrane provides very good results for the removal of
carried out to test the performance of reverse osmosis membranes in the the contaminants under study at the WWTP, even at high concentra-
removal of stimulants and antibiotics. Table 4 shows the results thus tions. Hence, the discharge or use of an effluent treated by means of this
obtained. type of membrane would not pose any risk to the environment or
The concentrations in the feed stream were the same as those found human health.
by other authors in WWTP effluents and in effluents from hospital As has been demonstrated, it removes all possible risks due to the
wastewaters [24,41]. It can be seen that the concentration of each presence of the contaminants under study. It is necessary to know the
pollutant generally decreases with time due to the previously discussed associated costs. The theoretical power for a centrifugal pump is ob-
processes. The deviations from this behaviour are due to variations in tained according to the following equation:
the influent flow rate to the membrane feed tank from the WWTP itself,
Pb=ρ·g·Q·hb (1)
which means that the concentration of each pollutant can vary.
Thus, caffeine varies from 0.557 mg L−1 (2 h) to 0.177 mg L−1 where:
(24 h) and theobromine from 0.534 mg L−1 (2 h) to 0.104 mg L-1 Pb is the theoretical power of the pump (watts)
(24 h). The highest concentration of theophylline in the feed stream was P is the density of the fluid (1000 kg·m−3 for water)
observed at 72 h (0.641 mg L−1), possibly due to the degradation of g is the acceleration of gravity (9.81 m·s−2)
caffeine in the mixing tank, with the subsequent increase in theophyl- Q is the flow rate (0.0002 m3·s−1 = 720 L·h−1 for both membranes)
line, as previously discussed. The minimum concentration of theo- Hb is the working pressure (15 bar)
phylline was detected at 2 h (0.104 mg L−1). Amoxicillin and penicillin Hence, the theoretical power of the centrifugal pump was
G were not detected in the feed after 48 h of testing due to their de- 300.186 W (0.32 kW) for RO. Then one day will consume
gradation or complete hydrolysis in the mixing tank that feeds the unit’s 7.68 kW·day−1.
feed tank. In the permeate stream, it can be observed that no con- Taking into account the serious problems, such as environmental,
taminants were detected in any case whatsoever during the 72 h the social and bacterial resistance and the low energy cost, it would be
study lasted (Table 4). It can therefore be concluded that 100% plausible to use these technologies as tertiary treatment in large
6
A. Egea-Corbacho Lopera, et al. Journal of Water Process Engineering 29 (2019) 100800
Fig. 3. Concentration in the feed, reject and permeate streams applying reverse osmosis. BDO5 (A), COD (B), TSS (C), phosphates (D), ammonium (E), nitrates (F), TN
(G), TP (H) Key: ◆Feed ◼Reject ▴ Permeat.
7
A. Egea-Corbacho Lopera, et al. Journal of Water Process Engineering 29 (2019) 100800
Fig. 4. Concentration in the feed, reject and permeate streams applying reverse osmosis. Turbidity (A), TOC (B), E. Coli (C), total coliforms (D), chlorides (E),
hardness (F), conductivity (G), pH (H). Key: ◆Feed ◼Reject ▴ Permeate.
polyphosphates from cleaning agents, orthophosphates, pyropho- machinery, formation of calcareous deposits in pipes, etc.
sphates, triphosphates and anions, that are difficult to remove from Concentrations below the limit of detection (1 mg L−1) were ob-
wastewater, and show the good performance of the membrane against tained for TOC in the permeate stream, except for the first 4 h of op-
these compounds. eration (1.69 mg L−1, 1.24 mg L−1 and 1.09 mg L−1 for 1, 2 and 4 h,
Nitrogenous compounds (ammonium and nitrates, Fig. 3E and F) respectively). The pH does not show significant changes in any of the
present values below the limits of detection in the permeated stream, three streams, the lowest values being for the reject stream (Fig. 4H).
except for the case of ammonium at 2 h. Total nitrogen, with an 88% The results obtained in Sections 3.2 and 3.3 allow us to conclude
removal rate, presents values in the feed stream ranging from that the water obtained after the tertiary treatment by reverse osmosis
11.63 mg L−1 (4 h) to 20.07 mg L−1 (72 h), and in the permeate stream may be used for any of the uses stipulated in the regulations, including
from 1.57 mg L−1 (4 h) to 2.01 mg L−1 (72 h). the most demanding.
E. coli, and total coliforms (except for the first 4 h) are absent in the
permeate (Fig. 4C and D). Turbidity, chlorides and hardness are below
3.5. Autopsy of the membrane
the limits of detection in the permeate stream (Fig. 4A, E and F). These
results mean that this stream is suitable for any of the uses allowed
One of the main aspects to be taken into account in membrane
under current Spanish regulations, without any risk of fouling of
processes is the decrease in performance due to fouling [42]. A lower
8
A. Egea-Corbacho Lopera, et al. Journal of Water Process Engineering 29 (2019) 100800
Acknowledgments
9
A. Egea-Corbacho Lopera, et al. Journal of Water Process Engineering 29 (2019) 100800
10