Document [original]

Direct Contact Membrane Distillation of Saline

Wastewater for Recycling

vorgelegt von

Sana Abdelkader

geb. Tunis, Tunesien

von der Fakultät III – Prozesswissenschaften

der Technischen Universität Berlin

und der

Höheren Schule für Lebensmittelindustrie Tunis der Carthage

Universität, Tunesien

(im Rahmen des Doppel-Promotionsabkommens)

zur Erlangung des akademischen Grades

Doktorin der Ingenieurwissenschaften

- Dr.-Ing. –

genehmigte Dissertation

Promotionsausschuss:

Vorsitzender: Prof. Dr. Hamdi Mokhtar (INSAT, Tunesien)

Gutachter: Prof. Dr.-Ing. Mongi Seffen (ESSTHS, Tunesien)

Gutachter: Prof. Dr. Martin Denecke (Universität Duisburg-Essen)

Gutachter: Dr. Joachim Koschikowski (Fraunhofer ISE)

Gutachter: Prof. Dr.-Ing. Sven-Uwe Geissen (TU Berlin)

Gutachter: Prof. Dr.-Ing. Latifa Bousselmi (CERTE, Tunesien)

Tag der wissenschaftlichen Aussprache: 11. Dezember 2018

an der Höheren Schule für Lebensmittelindustrie Tunis, Tunesien

Berlin 2018

Dissertation Report

Direct Contact Membrane Distillation of Saline

Wastewater for Recycling

Presented by

Sana Abdelkader

In the frame of an international joint supervision convention of a doctoral thesis

Submitted to the:

National Higher School of Food Industries (ESIAT), Tunisia

and

Technical University Berlin (TUB), Germany

Faculty III - Process Sciences

Department of Environmental Technology, Chair of Environmental Process Engineering

For the attainment of the academic degree of:

DOCTOR (Dr.-Ing.)

Disciplines:

Food Industries (ESIAT)

Environmental Science and Technology (TUB)

Jury members:

Chairman: Prof. Dr. Hamdi Mokhtar (INSAT, Tunisia)

Evaluator: Prof. Dr.-Ing. Mongi Seffen (ESSTHS, Tunisia)

Evaluator: Prof. Dr. Martin Denecke (University of Duisburg-Essen, Germany)

Examinator: Dr. Joachim Koschikowski (Fraunhofer ISE, Germany)

Supervisor: Prof. Dr.-Ing. Sven-Uwe Geissen (TU Berlin)

Supervisor: Prof. Dr.-Ing. Latifa Bousselmi (CERTE, Tunisia)

Centre of Water Research and Technologies (CERTE), Tunisia

Defense date: 11. December 2018

To my precious and dearest family

To my beloved parents

Ali and Aida

This work would not have been possible without your unconditional love and

support, your guidance and your constant belief in me.

A heartfelt Thank You to each and every one of you.

Acknowledgements

I would like to thank all those who helped me and contributed to the preparation and completion

of this thesis.

I would like to express my sincere gratitude to my supervisor Professor Latifa Bousselmi, head

of the research laboratory "Wastewater and Environment (LabEaUE)" at the Center for Water

Research and Technologies (CERTE), for her valuable guidance and support throughout the

elaboration of this thesis. Her great mentorship, positive outlook and encouragements inspired

me and enabled me to carry out this thesis successfully.

I would like also to thank my supervisor Professor Sven-Uwe Geissen, head of the “Chair of

Environmental Process Engineering” at the Technische Universität Berlin (TUB), for his valuable

support and for his constructive suggestions and recommendations during the planning and

development of this research work.

I extend my sincere thanks to all members of the jury for the honor they do me by agreeing to

read my manuscript and to participate in the defense of this thesis.

My sincere thanks to Dr. Ali Boubakri, assistant professor at CERTE, for his help, interest and

kind support during the realization of this work.

My thanks also go to my colleagues at the “Chair of Environmental Process Engineering (TUB)”

for their support. My special thanks go to Dr. Gesine Götz for her constant encouragement and

motivation and for the relevance of her advices.

I would like to express my gratitude to Dr. Joachim Koschikowski, head of the "Water

Treatment and Separation" research group at Fraunhofer-Institute for Solar Energy Systems

(ISE), for his support, scientific advice and valuable collaboration. My sincere thanks go also to

all the members of the "Water Treatment and Separation" research group at Fraunhofer ISE, for

their generosity, their time, resources and expertise which helped me greatly in the elaboration of

this work.

III

I would like also to thank Dr. Malini Balakrishnan, senior fellow at The Energy and Resources

Institute (TERI), and her research group for giving me the opportunity to work in collaboration

with their team and for their interest in my work.

Last but not least, I would like to express my gratitude to all the colleagues and friends who have

granted me their moral and intellectual support throughout my thesis. I express my gratitude to

all my colleagues at the CERTE, those who kindly agreed to help and to provide their valuable

support.

Tunis, August 2018

Sana Abdelkader

Abstract

Freshwater availability is suffering from an increasing pressure created by the growing demand,

depleting resources and the environmental pollution. Saline wastewater and saline industrial

effluents are creating a complex issue due to their contribution in the pollution and the alteration

of both surface water and groundwater properties. Thus, attention is drawn to investigating

complete treatment systems for water recovery, especially in the industry. Desalination of saline

wastewater presents a possibility to supply households, industry and agriculture with water.

Hence, the development of cost effective and efficient treatment processes has become an

obligation.

In this work, the first step was to test two biological processes, namely aerobic and anaerobic

processes, in combination with two membrane treatments (microfiltration and ultrafiltration) for

saline wastewater treatment. It was observed during the tests that both the biological treatments

and the membrane filtration performances were significantly affected by the NaCl concentration

increase in the feed solutions.

Membrane Distillation (MD) is a competitive technology for water desalination which forms an

appealing solution for industrial wastewater treatment. It is getting an increasing attention thanks

to its advantages in terms of energy consumption and final permeate quality. In this study, the

second step was to apply Response Surface Methodology (RSM) to optimize the Direct Contact

Membrane Distillation (DCMD) treatment of synthetic saline wastewater. The aim was to

enhance the process performance and the permeate flux by optimizing the operating parameters.

Despite its advantages, one of the most challenging issues in DCMD is membrane fouling and

wetting. Therefore, the DCMD treatment of real industrial effluent was a focus in the present

research work. Saline dairy effluent discharged from hard cheese industry was first pretreated by

macrofiltration (MAF) and ultrafiltration (UF) and then processed by DCMD to investigate the

extent of the aforementioned issues. Effluents pretreated by UF have led to the best process

performance with stable DCMD permeate flux values at different operating conditions. Fouling

has occurred in all the experiments, though their effect on the flux behavior and membrane

wetting was different from one feed to the other. The utility theory was applied on the different

scenarios that have been experimented using the studied system in order to identify the best

demonstrated improvements in terms of process performance and cost-effectiveness.

Zusammenfassung

Die Verfügbarkeit von Süßwasser leidet unter einem zunehmenden Druck, der durch die

wachsende Nachfrage, die Erschöpfung der Ressourcen und die Umweltverschmutzung entsteht.

Salzwasser und salzhaltige Industrieabwässer stellen ein komplexes Problem dar. So wird die

Aufmerksamkeit auf die Untersuchung kompletter Aufbereitungssysteme zur

Wasserrückgewinnung, insbesondere in der Industrie, gelenkt. Die Entsalzung von salzhaltigem

Abwasser bietet die Möglichkeit, Haushalte, Industrie und Landwirtschaft mit Wasser zu

versorgen. Daher ist die Entwicklung kostengünstiger und effizienter Behandlungsverfahren zu

einer Verpflichtung geworden.

In dieser Arbeit war der erste Schritt, zwei biologische Prozesse zu testen, nämlich aerobe und

anaerobe Prozesse, in Kombination mit zwei Membranbehandlungen (Mikrofiltration und

Ultrafiltration) zur Behandlung von Salzwasser. Bei den Tests wurde festgestellt, dass sowohl die

biologischen Behandlungen als auch die Membranfiltrationsleistungen durch den Anstieg der

NaCl-Konzentration in den Feed-Lösungen signifikant beeinflusst wurden. Die Membran-

Destillation (MD) ist eine wettbewerbsfähige Technologie zur Wasserentsalzung, die eine

attraktive Lösung für die industrielle Abwasserbehandlung darstellt. Dank seiner Vorteile in

Bezug auf Energieverbrauch und Endpermeatqualität erfährt es eine zunehmende

Aufmerksamkeit. In dieser Studie wurde im zweiten Schritt die Response Surface Methodology

(RSM) zur Optimierung der Direktkontakt Membrandestillation (DCMD) Behandlung von

synthetischem Salzwasser angewendet. Ziel war es, die Prozessleistung und den Permeatfluss

durch Optimierung der Betriebsparameter zu verbessern. Trotz seiner Vorteile ist

Membranverschmutzung und Benetzung eines der schwierigsten Themen bei der DCMD. Daher

war die DCMD-Behandlung von realem industriellem Abwasser ein Schwerpunkt der

vorliegenden Forschungsarbeiten. Das aus der Hartkäseindustrie eingeleitete salzhaltige

Milchabwasser wurde zunächst durch Makrofiltration (MAF) und Ultrafiltration (UF)

vorbehandelt und anschließend von der DCMD verarbeitet, um das Ausmaß der oben genannten

Probleme zu untersuchen. Die mit UF vorbehandelten Abwässer haben zu bester Prozessleistung

mit stabilen DCMD-Permeatfluss bei unterschiedlichen Betriebsbedingungen geführt. Fouling ist

in allen Experimenten aufgetreten, obwohl ihr Einfluss auf das Flussverhalten und die Benetzung

der Membranen von einem Feed zum anderen unterschiedlich war. Die Utility-Theorie wurde auf

die verschiedenen Szenarien angewendet, die mit dem untersuchten System experimentiert

wurden, um die an den besten nachgewiesenen Verbesserungen in Bezug auf Prozessleistung und

Kosteneffizienz zu identifizieren.

Résumé

La disponibilité de l'eau douce souffre d'une pression croissante créée par la demande progressive,

l'épuisement des ressources et la pollution de l'environnement. Les eaux usées salines et les

effluents industriels salins créent un problème complexe en raison de leur contribution à la

pollution et à l'altération des propriétés des eaux de surface et des eaux souterraines. Ainsi,

l'attention est attirée sur l'étude de systèmes complets de traitement pour la récupération de l'eau,

en particulier dans l'industrie. Le dessalement des eaux usées salines offre la possibilité

d'approvisionner en eau les ménages, l'industrie et l'agriculture. Par conséquent, la mise au point

de procédés de traitement rentables et efficaces est devenue une obligation.

Dans ce travail, la première étape consiste à tester deux procédés biologiques, à savoir les

procédés aérobie et anaérobie, en combinaison avec deux traitements membranaires

(microfiltration et ultrafiltration) pour le traitement des eaux usées salines. Il a été observé au

cours des essais que les traitements biologiques et les performances de filtration ont été

significativement affectées par l'augmentation de la concentration en NaCl dans les effluents.

La distillation à membrane (MD) est une technologie compétitive qui constitue une solution

attrayante pour le traitement des eaux usées industrielles salines. Elle fait ainsi l'objet d'une

attention croissante grâce à ses avantages en termes de consommation d'énergie et de qualité de

perméat final. Dans cette étude, la deuxième étape consiste à appliquer la méthode d’optimisation

par les plans à surfaces de réponses (RSM) pour optimiser le traitement des eaux usées salines

synthétiques par distillation à membrane à contact direct (DCMD). L'objectif est d'améliorer les

performances du procédé et le flux de perméat en optimisant les paramètres opératoires.

Malgré ses avantages, l'un des problèmes les plus difficiles à résoudre dans la DCMD est

l'encrassement et le mouillage des membranes. Par conséquent, la DCMD des effluents industriels

réels a été étudié dans ce travail. Les effluents laitiers salins rejetés par une industrie fromagère

ont d'abord été prétraités par macrofiltration (MAF) et ultrafiltration (UF), puis traités par DCMD

pour étudier l'ampleur des problèmes susmentionnés. Les effluents prétraités par l'UF ont permis

d'obtenir les meilleures performances avec des valeurs de flux de perméat stables dans différentes

conditions de fonctionnement. L'encrassement s'est produit dans toutes les expériences, bien que

leur effet sur le comportement du flux et le mouillage de la membrane ait été différent selon la

qualité de l’effluent initial. La théorie de l'utilité a été appliquée aux différents scénarios

expérimentaux en utilisant le système de traitement étudié afin d'identifier les meilleures

améliorations en termes de performance du processus et de rentabilité.

VII 
 
TABLE OF CONTENTS 
Literature survey .............................................................................................................. 9 
1 Introduction and objectives ............................................................................................... 9 
2 Saline wastewater ............................................................................................................. 11 
2.1 Saline effluents treatment ..................................................................................................................... 12 
2.2 Cheese industry wastewater .................................................................................................................. 17 
3 Membrane distillation ...................................................................................................... 19 
3.1 Process description ............................................................................................................................... 19 
3.2 Membrane distillation configurations .................................................................................................. 21 
3.3 Transport phenomena in Direct Contact Membrane Distillation (DCMD) ........................................ 23 
3.4 Membranes in DCMD .......................................................................................................................... 24 
3.5 Fouling and wetting in DCMD ............................................................................................................. 25 
3.6 Pretreatment and membrane cleaning .................................................................................................. 26 
3.7 Fields of application .............................................................................................................................. 26 
3.8 Utility theory ......................................................................................................................................... 28 
Material and methods ..................................................................................................... 29 
1 Studied saline effluents ..................................................................................................... 29 
1.1 Synthetic saline wastewater .................................................................................................................. 29 
1.2 Saline cheese whey wastewater ............................................................................................................ 29 
2 Biological treatments of the saline synthetic wastewater ................................................ 30 
2.1 Aerobic bioreactors ............................................................................................................................... 30 
2.2 Anaerobic bioreactors ........................................................................................................................... 31 
2.3 Microfiltration and ultrafiltration membranes ...................................................................................... 32 
3 Direct Contact Membrane Distillation (DCMD) ............................................................. 34 
3.1 Benchscale DCMD setup...................................................................................................................... 34 
3.2 Laboratory scale DCMD setup ............................................................................................................. 36 
3.3 DCMD membranes ............................................................................................................................... 38 
4 DCMD pretreatments ...................................................................................................... 39 
4.1 Macrofiltration (MAF) .......................................................................................................................... 39 
4.2 Ultrafiltration (UF) ................................................................................................................................ 39 
5 Analytical methods ........................................................................................................... 40 
5.1 Pollution parameters ............................................................................................................................. 40 
5.2 Membrane characterization .................................................................................................................. 41 
5.3 Permeate flux and permeability ............................................................................................................ 42 

 
VIII 
 
6 Response Surface Methodology (RSM) ........................................................................... 43 
7 Single-Node model simulation.......................................................................................... 44 
8 Experimental performance analysis ................................................................................ 45 
Results and discussion .................................................................................................... 51 
1 Biological and membrane processes performance for saline effluents treatment .......... 51 
1.1 Aerobic bioreactors performance ......................................................................................................... 51 
1.2 Anaerobic bioreactors performance ..................................................................................................... 53 
1.3 Membrane microfiltration and ultrafiltration ....................................................................................... 58 
1.4 Conclusions ........................................................................................................................................... 66 
2 DCMD optimization for saline synthetic effluent treatment ........................................... 68 
2.1 Permeate flux stability .......................................................................................................................... 68 
2.2 Response Surface Methodology (RSM) application in DCMD .......................................................... 69 
2.3 Analysis of Variance ............................................................................................................................. 73 
2.4 Response surface and contour-line plots .............................................................................................. 74 
2.5 DCMD response optimization and model verification ........................................................................ 77 
2.6 Conclusions ........................................................................................................................................... 79 
3 DCMD application for cheese whey wastewater treatment ............................................ 80 
3.1 Raw dairy effluent DCMD treatment ................................................................................................... 80 
3.2 Pretreated dairy effluents treatment ...................................................................................................... 82 
3.3 Parameters affecting the permeate flux ................................................................................................ 85 
3.4 Membrane fouling and wetting ............................................................................................................ 93 
3.5 Experimental performance and cost-effectiveness analysis .............................................................. 102 
3.6 Conclusions ......................................................................................................................................... 114 
Conclusions and perspectives ....................................................................................... 116 
References ............................................................................................................................ 120 
 
 

1.Literature survey

1.1.Introduction and objectives

Sana Abdelkader

1 LITERATURE SURVEY

1.1 INTRODUCTION AND OBJECTIVES

The increasing domestic and industrial demand for clean and fresh water is creating an ever-

growing pressure on the global security since water has numerous interlinkages with all the

aspects in our life in terms of economic development, energetic demand, environmental

security and industrial growth which is getting even more critical with the actual witnessed

climate change (UN report, 2015). Water scarcity is an urgent problem that needs innovative

and adapted solutions in order to satisfy the world’s current and future population. According

to the UN report “Water and Energy” (2014), it is expected that the growing global water

demand will contribute in the aggravation of the actual issue of water scarcity and

unavailability all over the world since it is expected, by 2050, to push 40% of the global

population under water scarcity level.

Water desalination is considered as one of the most prevalent solution to overcome the

aforementioned problem. However, it presents one of the most energy-consuming processes

with 75.2 Terawatt hour per year, which presents 0.4% of the global electricity demand. The

high energy consumption characterizes the common desalination technologies: membrane

separation processes (Reverse Osmosis: RO) and thermally driven processes (Multistage Flush:

MSF, Adsorption Desalination: AD, etc.) (Shahzad et al. 2017). Water desalination and more

precisely wastewater treatment and desalination is providing an interesting alternative to

partially overcome the need for more water resources by creating an alternative water resource

through the reuse of the discharged effluents and to integrate them in the industrial or domestic

water cycles via multiple desalination technologies like Membrane Distillation (MD) which is

offering the production of high quality permeates by the means of lower energy consumption

in comparison to the conventional and well established desalination processes (Miller, 2003;

Samblebe, 2006).

In this context, MD comes as an emerging water treatment technology that is having a growing

interest since it is associating the advantages of both membrane and thermal processes to avoid

large amounts of energy consumption without compromising the final treated water quality and

with an interesting ability to process multiple saline feed water qualities. MD has the

particularity of operating under low hydrostatic pressure and low operating temperature (below

the feed boiling temperature) without overlooking the particular nature of the driving force

which is the transmembrane water vapor pressure difference (El-Bourawi et al. 2006).

1.Literature survey

1.1.Introduction and objectives

Sana Abdelkader

In this framework comes the present work that is focusing on the saline wastewater treatment:

• The first part of this work is focusing on the treatment of synthetic saline wastewater

using different combinations of biological and filtration processes: aerobic biological

treatment and anaerobic biological treatment coupled with both microfiltration and

ultrafiltration membrane processes by ceramic and polymeric membranes. With a

particular emphasis on the salt concentration, the salt content was varied in all of the

aforementioned treatments and by using new and a particular kind of membranes made

in TERI University laboratories, India. In this section, different treatment processes

and wastewater qualities were analysed and investigated to show the effect of each

parameter on the processes performance, the treated water quality and on the studied

membrane filtration.

• To deal with the efflunts desalination, the second part of this study is focusing on

investigating the optimization of synthetic saline wastewater treatment by Direct

Contact Membrane Distillation (DCMD) by maximizing the process response in terms

of permeate flux through the application of Response Surface Methodology (RSM).

This objective will be achieved by the identification of the optimum levels of the

operating independent variables, namely, the temperature difference, feed velocity, salt

concentration and glucose concentration.

• Considering on one hand the importance of the food industry effluents treatment as

well as the possibility of water recovery and on the other hand the lack of researches

treating those effluents by membrane distillation, this present work investigates also

the capability of Direct Contact Membrane Distillation (DCMD) of treating raw and

pretreated saline dairy effluent. The emphasis of these investigations was on the

fouling and wetting phenomena that occur during the processing of three feed qualities:

i) raw effluent ii) pretreated by macrofiltration (MAF) and iii) followed by

ultrafiltration (UF). For the sake of accuracy, the DCMD experiments were performed

under various operating conditions. This allows to reveal new insights related to the

impact of such conditions on the DCMD process. The major constitutions of this work

are listed in the following: An experimental environment is setup to assess the

efficiency of DCMD for treating saline dairy effluent.The membranes pore sizes and

the effluent pretreatment type are varied in order to analyze their effect on the

performance of The DCMD process. The feed velocity and temperature difference are

also varied for this purpose and the obtained results are discussed in order to explore

the efficiency of DCMD process in this type of industry.The novelty of this research

1.Literature survey

1.2.Saline wastewater

Sana Abdelkader

work consists in investigating the conception of a complete treatment system based on

DCMD for a specific real industrial effluent (dairy wastewater). This system takes into

consideration the effluent pretreatment with two different technologies.

• The performance analysis of saline dairy wastewater DCMD process is extended so as

to consider its cost-effectiveness when in it is associated with two types of

pretreatment. To this purpose, we take into consideration a set of parameters (i.e.

criteria) that impact the applicability of the DCMD process in industry in order to

assess its performance. Having introduced these criteria, a set of indicators are defined

and are used in our analysis through the application of the utility theory. Then, a

mathematical model is developped to assess the performance of the scenarios that have

been experimented using our setup. Finally, the results are discussed based on our case

study in order to analyze the cost-effectiveness of the DCMD process with respect to

the two pretreatment options.

1.2 SALINE WASTEWATER

Saline wastewater has generally two major origins, namely, sewage and industrial effluents.

On one hand, the contamination of surface water and groundwater with salts affects the

domestic and municipal wastewater quality since this used resources will end up in the sewage

system , this contamination could be attributed to the salinization of the groundwater bodies in

some coastal areas (i.e. Mediterranean basin) by seawater and aquifers recharge that is

aggravated by the overexploitation of the natural water resources for irrigation and/or domestic

activities (Demir et al., 1999; Panno et al., 2002; Khaska et al., 2013).Moreover, in some

cases the salinity of the groundwater is caused by evaporates dissolution in the groundwater

that is driven by gravity (Bethke et al., 1990).

On the other hand, some industrial activities are important contributors in the generation of

saline effluents because of the high salt concentration consumption (mainly NaCl) in their

processes that induces the discharge of large amounts of saline effluents in the environment.

Some of the most influencing industries that could be listed in this context are: the chemical

industry and agro-food industry as well as leather, textile and petroleum industries (Antileo et

al., 1997; Diaz et al., 2002; Lefebvre et al., 2006.a). In addition to the aforementioned saline

effluents sources, the water softening (desalination) stations have their share in the generation

of non-negligible amounts of hyper saline brines following the application of water desalination

technologies (generally, Reverse Osmosis RO) (Lattemann et al., 2008).

1.Literature survey

1.2.Saline wastewater

Sana Abdelkader

According to the UN report (2017), untreated and poorly treated wastewater discharge in the

environment is still a common practice in some countries leading to direct negative impact.

This influence affects not only the human health but also the fauna and flora since we cannot

dissociate the discharged effluents from the water cycle (Figure 1). This dilemma gets even

more complex and critical when the wastewater contains high salt levels in addition to the

organic pollutants (Lefebvre et al., 2006.a).

Figure 1: Wastewater in the water cycle (source: UN report 2017).

1.2.1 Saline effluents treatment

1.2.1.1 Biological processes

Aerobic activated sludge process and anaerobic digestion are generally the most applied

processes for wastewater treatment that can lead to good effectiveness in terms of Chemical

Organic Carbon COD reduction levels (Sundaresan et al., 2008; Saleh et al., 2004).

1.Literature survey

1.2.Saline wastewater

Sana Abdelkader

Membrane bioreactors (MBRs) which are the combination of membrane separation and the

activated sludge process, are now widely used in wastewater treatment with well recognized

advantages (Judd and Judd, 2006). Meanwhile, their performance is also known to be limited

by fouling, which constrains both operating flux and ultimately, the membrane lifetime (Ji et

al., 2006) due to soluble and bound colloidal/macromolecular substances in the sludge biomass

which is referred to as soluble and bound extracellular polymeric substances (EPS),

respectively (Bura et al., 1998; Nuengjamnong et al., 2005; Ji et al., 2006).

There are two fundamental strategies for cells to survive under osmotic stress:

• The increase of the intracellular ion concentration in order to balance the external

osmotic pressure, and all intracellular enzymes have to adapt to the new conditions.

This strategy is followed by the anaerobic halophilic bacteria, whose entire physiology

has been adapted to high saline environments (Kempf et al., 1998).

• Many microorganisms accumulate osmotically active solutes called “compatible

solutes”. The high external osmotic pressure is balanced within the cells by organic

compatible solutes to protect and enhance the bacterial growth in high osmolality

environments (Kempf et al., 1998).

The complexity of saline wastewater treatment derives mainly from the fact that the usual

biological processes has shown in some cases limited performance mainly linked to the

effluents salt concentration (Wang et al., 2005). Moreover, it is well known that in high salt

content levels, the microorganism’s osmotic stress is increased which could have inhibitory or

toxic effects on the sludge bacteria. This could lead to the cells plasmolysis and/or the inhibition

of their activity, thereby reducing the effectiveness of biological treatment in terms of chemical

oxygen demand (COD) removal (Reid et al., 2006; Luo et al., 2015).Hence, it was found in

previous research that in MBR treatment the microbial community is altered due to salinity

changes and the effluent salinity significantly affects the physical and biochemical properties

of the activated sludge, leading to several changes in the MBR treatment performance: the

settlement phase, the bio-flocculation, the membrane surface charge, hydrophobicity and

filterability (Kargi and Dincer, 1996; Dincer and Kargi, 2001).

Reid et al. (2006), revealed in their study that effluents witnessing salinity shocks over a certain

range (0.1–4 g/L) are likely to become even more challenging to an MBR than effluents having

readily stable salinity levels.

Therefore, and according to them, coastal-based installations that are struggling with seawater

intrusion and other treatment plants that are dealing with the treatment of certain saline

1.Literature survey

1.2.Saline wastewater

Sana Abdelkader

effluents may benefit from the use of buffer tanks, or possibly pre-sedimentation which would

enhance the MBR treatment by reducing the effluents’ organic load.

Several studies have examined the anaerobic treatment of different saline industrial effluents

such as those generated by seafood, tanning and textile industries in addition to saline sewage

wastewater (Lefebvre et al., 2006.a; Moharram et al., 2015).

Ozalp et al. (2003) proposed the use of an up-flow anaerobic sludge bed reactor (UASBR) and

they measured an 85% total organic carbon (TOC) removal in synthetic sewage wastewater

with NaCl concentration of 15 g/l. In their study, the hydraulic retention time (HRT) was 24h

and the reactor gradually reached 15 g NaCl/L over a period of 109 days. Besides, it is worth

mentioning that TOC analysis was used due to difficulties encountered in carrying out COD

analyses because of the interference of chlorides in highly saline samples.

Rovirosa et al. (2004) studied the treatment of saline wastewater by using a laboratory down-

flow anaerobic fixed bed reactor. The results showed that an HRT of 24 h was required to

obtain total COD, organic-N, total-P and fecal coliform concentration reduction efficiencies

higher than 72%, 51%,39% and 98%, respectively, with sea salts concentrations in the range

from 5 g/L to 15 g/L.

Gebauer (2004) studied anaerobic treatment of sludge from saline fish farm effluents and

followed the COD removal in undiluted sludge with 35%salinity varied between 40% and 54%

depending on the operating condition. Another research focusing on the treatment of saline

wastewater of tannery soak liquor in an up-flow anaerobic sludge blanket reactor (UASB)

showed that 78% COD removal can be obtained at 5 days HRT and 71 g/L total dissolved

solids (TDS) as salinity (Lefebvre et al., 2006.b).

To overcome the limitations observed in the anaerobic digestion, some studies focused on

investigating the influence of the operating conditions on the process efficiency. For instance,

Vyrides and Stuckey (2009.b) studied the possibility of compatible solutes addition to the

anaerobic saline medium and found that glycine betaine had the best effect on its performance

and that under high salinity conditions (2% and 4% NaCl) there has been more production of

extracellular polymeric substances (EPS) and the mean floc size was relatively bigger.

Moreover, Shi et al. (2014) reported in their study that sequential anaerobic–aerobic

treatments of saline pharmaceutical wastewater achieved high COD removal rates that reached

94.7% and 91.8%.The use of salt-tolerant (halophilic) microorganisms was also investigated

to study how effective the biological treatment could be when treating saline effluents with the

selected microbial groups through the enhancement of the COD removal rates (70% to 84%

1.Literature survey

1.2.Saline wastewater

Sana Abdelkader

with effluents having an NaCl concentration of 3%) with emphasis on the microorganism’s

survival capability in osmotic stress (Kapdan et al., 2007; Zhang et al., 2014).

1.2.1.2 Membrane processes

Membrane treatment includes multiple separation classes according to the substances rejected

by the membranes which is mainly controlled by the membranes pore size (Figure 2).

Macrofiltration, microfiltration and ultrafiltration could be considered as pretreatments to the

saline effluents desalination since they have no effect on the salt concentration reduction, but

they could provide good suspended solids and colloids removal rates which improve the saline

effluents quality for further desalination (Lefebvre et al., 2006.a).

Figure 2: Schematic illustration of the different membrane filtration types according to the

retained particle’s size. (Source: lee and al. 2016)

1.Literature survey

1.2.Saline wastewater

Sana Abdelkader

Nanofiltration, as a membrane separation technology using both electric charge (Donnan

effect) and pore size (sieving effect), could separate low molecular weight solutes (e.g. glucose,

saccharides, amino acid, and peptide) from inorganic salt solutions, and simultaneously

concentrate organic solutes and remove inorganic salt, showing a good potential in desalination

and/or the recovery of valuable organic substances (permeation of monovalent salts)

(Mohammad et al., 2010). Moreover, the salt ions could easily pass through the membrane at

high salt concentrations; even the nominal monovalent salt rejection was often negative in

mixtures of salts and large charged organic molecules or mixed monovalent–multivalent salts,

thus greatly decreasing osmotic pressure difference across membrane (Luo et al., 2009;

Mohammad et al., 2010).

The most commonly applied desalination technology is Reverse Osmosis (RO) which showed

good performance but has an important disadvantage which is the process high energy

consumption which reflects on the process cost, since for a typical reverse osmosis seawater

desalination plant, it is required to provide 3 to 10 kWh of electric energy in order to produce

one cubic meter of freshwater and most of the required energy is used to pressurize the feed

streams and such operating pressure is related to feed salinity degree (Dashtpour et al., 2012).

For instance, the applied pressure could vary between 15 to 30 bars for brackish water and

could range between 55 to 70 bars for seawater desalination (Dashtpour et al., 2012). Another

issue that is common in membrane treatments, and especially in RO, is the membrane fouling

which highly affects the water recovery rates that can drop to 25% to 45% when treating

seawater (Charcosset, 2009).

When observing the membrane processes performance, it appears to be a complex function of

the feed water characteristics and the chemical and physical properties of the membrane itself.

These complex interactions make process control difficult as the feed water mixture can change

in time (Tchobanoglous and al., 2003).

1.2.1.3 Other processes

Different physico-chemical technologies were employed for the treatment of saline effluents

such as thermal processes which include the solar evaporation as the simplest technique and

the multi-effect evaporation which could have a competitive cost in comparison to membrane

processes (mainly RO) in some regions where there is high availability of energy with low cost

(Lefebvre and al. 2006.a).

Chemical oxidation was also applied as one of the viable options for the treatment of saline

effluents, for example, Boron-Doped Diamond anodes were utilized for electrochemical

1.Literature survey

1.2.Saline wastewater

Sana Abdelkader

oxidation of industrial saline streams which has led to good removal rates in terms of TOC and

ammonia and consequently increased the effluent’s biodegradability and demonstrated that

electro-oxidation could be considered as a good alternative for pretreatment before biological

processes application (Anglada et al., 2010).

Ion exchangers are also used to desalinate water by removing the effluents ionic content

through anionic and cationic resins that catches the ions and then the resins are subjected to a

regeneration phase (backwashing) which leads both to cost increase of the process and the

generation of high concentration streams (McGhee, 1991).

Another possible treatment is the coagulation/flocculation process which could be a

pretreatment in the case of high salinity effluents in order to reduce the colloidal COD by the

means of coagulant agents. However, this process doesn’t decrease the salt concentration which

requires further treatments for salt removal (Lefebvre et al., 2006.a).

Nevertheless, the existing alternative desalination treatment approaches, such as crystallization,

extraction, ion-exchange and electrodialysis, are usually too expensive to be effectively put in

industrial applications and the cost of treatment can eventually undermine the economic

viability of the whole process with progressing tightening of environmental regulations

(Jianquan et al., 2013).

1.2.2 Cheese industry wastewater

Treating real industrial wastewater, more precisely food industry effluents, presents a

challenging issue when it comes to water recovery and reuse due to their complex and generally

fluctuating compositions and characteristics (Blaschek et al., 2007). Among the food

processing industries, we find dairy industries which consume variable water quantities and

generate effluents that have variable characteristics and qualities depending on the final product

and the steps of the process design (Rivas et al., 2010).

1.2.2.1 Cheese whey wastewater composition

Cheese making procedures generate effluents that are mainly composed of the water volumes

that are utilized for cheese manufacture and the cleaning process in addition to cheese whey

that could have different qualities depending on the processed raw milk characteristics and on

whether there has been a valorization step of the cheese whey in the industry (Table 1)

(Carvalho et al., 2013).

Saline cheese whey effluents are often characterized, in addition to their salt content, by a

variable organic load due to the presence of whey proteins and traces of other milk fat and

1.Literature survey

1.2.Saline wastewater

Sana Abdelkader

proteins which are found with variable proportions in the discharged wastewater after hard

cheese manufacture (Blaschek et al., 2007).Whey proteins are mainly composed of: β-

lactoglobulin (60%), α-lactalbumin (20%), Immunoglobulin G (10%), Bovine Serum

Albumin: BSA (3%) and Lactoferrin (0.1%) (Singh et al., 2014). Their molecular masses are

respectively: 18.3 kDa, 14.2 kDa, 161 kDa, 66.4 kDa and 76.1 kDa. Concerning the other milk

proteins, they are casein micelles (molecular weight ranging from 106 to >109 Da) and they are

consisting in aggregates formed through hydrophobic and electrostatic repulsive interactions

of casein proteins and micellar calcium phosphates. Among the caseins, which are responsible

of the whitish color of cheese whey, we find αS1-casein, αS2-casein and β-casein (Singh et al.,

2014).

Table 1: Compilation of cheese whey wastewater characteristics (adapted from Carvalho et

al., 2013 and Kezia et al., 2015).

pH/EC

Lactose

Proteins

Fats

COD

TOC

Total

solids

Reference

7±2/-

1.62

±0.56

0.55

±0.09

Fang et

al.,1991

4.3-8.7/-

2.3-33.5

0.4-

5.7

5.4-77.3

3.9-58.9

Kalyuzhnyi

et al., 1997

5.92

44.37

±0.88

9.06

71.41

35.4

62.44±2.9

Yang et al.,

2003

7.83-7.9

0.33-

0.95

11.8-

17.6

Janczukowicz

et al., 2008

-/ 110-

127

4.5±1

0.3±0.05

114±5

Kezia et al.,

2015

4-4.6/

11.3-

13.5

0.178-

0.182

(939-

947) 10-6

1.83-

3.76

8.8-25.6

7-8.3

Rivas et al.,

2010

0.18

±0.003

(943±6)

10-6

18.5±1.4

7.7±0.6

Rivas et al.,

2011

-/ 66-74

8±0.6

3.6±0.3

77±7

Kezia et al.,

2015

All parameters are expressed in (g/L) except EC (mS/cm).

1.Literature survey

1.3.Membrane distillation

Sana Abdelkader

As illustrated in Table 1, the cheese whey wastewater has a fat content that varies in a wide

range and this fraction comes originally from the raw milk lipids which are present in the milk

serum under a dispersed form of globules that have different diameters varying generally

between <1µm to almost 20µm with a mean size of 3 to 4µm and the fat globules size

distribution is closely linked to the medium (milk or whey) homogenization, since it induces

the breakdown of the fat globules and the change of their native structure (Argov et al., 2008;

Singh et al., 2014).

1.2.2.2 Cheese whey wastewater treatment

Generally, cheese whey wastewater is conventionally treated chemically (coagulation/

flocculation) and/or via biological processes (aerobic and anaerobic digestion) to reduce its

Chemical Oxygen Demand (COD) significantly with removal rates that can reach 99% in some

cases (Carvalho et al., 2013). However, in the presence of high salt content, biological effluent

treatment showed limited performance due to the alteration of the biomass properties and

growth which causes the instability and reduction of the organic removal process especially

with the presence of salt built-up (Jang et al., 2013).

1.3 MEMBRANE DISTILLATION

1.3.1 Process description

MD is an emerging water treatment technology that focuses mainly on water streams

desalination. This technology is known as a thermally driven process that includes membrane

separation via the transport of water vapor molecules through a hydrophobic microporous

membrane (Figure 3). MD has interesting advantages thanks to the fact that it summarizes both

benefits of membrane and thermal processes through its relatively low energy consumption

and the generation of water with high quality (El-Bourawi et al., 2006). Additionally to its

high salt rejection rates, MD is characterized by good rejection rates of non-volatile and low

adsorptive organic compounds (Carnevale et al., 2016; Plattner et al., 2017).

1.Literature survey

1.3.Membrane distillation

Sana Abdelkader

Figure 3: Schematic of MD process

As shown in Figure 3, MD process is based on the presence of two separate hydraulic loops:

first, one evaporation side where the feed is being heated (TF) and circulated continuously and

has to enter in direct contact with the hydrophobic membrane and second, one condensation

side where the permeate is being cooled (TP) and circulated continuously which creates a

temperature gradient between the two membrane’s sides following which the water vapor

molecules start passing from the first side to the second one through the microporous barrier

that prevents the liquid streams from being transferred to the opposite side due to surface

tension forces and consequently liquid/vapor interfaces are created on the membrane(Khayat

et al., 2011).

One of the most important characteristics of MD is the ability of operating under low

hydrostatic pressure and low operating temperature since the feed stream temperature could be

inferior to its boiling temperature, which makes MD an attractive option for water desalination,

and another particularity of MD process is the nature of the driving force that is based on the

transmembrane water vapor pressure difference between the two streams (Alkhudhiri et al.,

2012). Since the feed in MD process doesn’t require to be heated up to high temperature levels,

the use of the waste heat in the plants (via heat recovery heat exchangers) or alternative energy

Feed

(evaporation)

side

Permeate

(condensation)

side

Hydrophobic microporous membrane

1.Literature survey

1.3.Membrane distillation

Sana Abdelkader

resources (i.e. solar energy, etc…) for this purpose is a possible option that helps in dwindling

the MD process cost (this could include low amount of additional heating if needed) especially

when the thermal energy resources are limited (Guan et al., 2015). One of the possibilities to

recover heat for MD process operation is the concentration cascades design that is formed of

multiple MD units with countercurrent cross-flow continuous water circulation where the heat

is recycled for each cascade in order to achieve maximum heat recovery with good water

recovery levels that can reach 50% to 60% without additional heat resources (Lee et al. 2011;

He et al., 2013).

1.3.2 Membrane distillation configurations

Depending on the methodology following which the water vapor is recovered and collected in

the cool side of the porous hydrophobic membrane (the permeate side) we can distinguish

multiple MD configurations where the way of inducing the vapor pressure gradient across the

membrane is different (Figures 4 and 5) (Drioli et al., 2014).

Figure 4: Common types of membrane distillation configurations (adapted from El-Bourawi

et al., 2006)

1.3.2.1 Direct Contact Membrane Distillation (DCMD)

This is the simplest configuration of MD, where the permeate stream is kept in direct contact

with the opposite side of the hydrophobic membrane (Figure 4). In this case, both feed and

permeate streams are being circulated using pumps and the created vapor pressure across the

membrane induces the vapor molecules to pass through the membrane and to be condensed

directly in the cooler water stream (Khayat et al., 2011).

1.Literature survey

1.3.Membrane distillation

Sana Abdelkader

1.3.2.2 Air-gap Membrane Distillation (AGMD)

In this configuration, a condensing surface (or plate) is placed after an air gap between the

membrane surface and coolant stream (Figure 4). The vapor molecules which pass through the

membrane cross the air gap and condensate on the cold waterproof surface to be collected

afterwards as the MD permeate and consequently the membrane has no wetting risks on the

permeate side (Drioli et al., 2014; Ashoor et al., 2016).

1.3.2.3 Sweeping-gas Membrane Distillation (SGMD)

The liquid stream in the permeate side is replaced in this configuration by an inert gas which

sweeps the membrane on the permeate side and the condensation of water vapor molecules is

performed outside the membrane module (Figure 4), such configuration is useful when treating

feed streams that contains volatile compounds and the obtained permeate quality will be

independent of the membrane wetting (El-Bourawi et al., 2006; Alkhudhiri et al., 2012;

Drioli et al., 2014).

1.3.2.4 Vacuum Membrane Distillation (VMD)

In this particular configuration vacuum is created on the permeate side using a vacuum pump

which creates on this side a pressure lower than the saturation pressure of the volatile molecules

in the feed stream and afterwards the vapor condensation is made outside the membrane

module like in the SGMD configuration (Figure 4). VMD configuration provides high

permeate fluxes but it can present higher risks of membrane wetting and fouling in some cases

(El-Bourawi et al., 2006; Khayat et al., 2011; Drioli et al., 2014).

1.3.2.5 Osmotic Membrane Distillation (OMD)

The major difference between OMD and the previous MD configuration is that OMD is a non-

thermal process that operates at constant and relatively low temperature and the driving force

(vapor pressure gradient) is created by the water activity difference between the two liquid

streams (Figure 5) which enter in direct contact with the membrane surface on both sides (on

one side the feed stream and on the other typically a brine stream) which is realized under

atmospheric pressure.

This configuration could be interesting when treating viscous and/or thermally sensitive feed

streams (Babu et al., 2006; Warczok et al., 2007).

1.Literature survey

1.3.Membrane distillation

Sana Abdelkader

Figure 5: Schematic of Osmotic Membrane Distillation (OMD) (source: Hogan et al., 1998)

1.3.3 Transport phenomena in Direct Contact Membrane Distillation (DCMD)

In Membrane distillation, the water vapor molecules transport through the microporous

hydrophobic membrane involves two types of transport mechanisms, namely, heat transfer and

mass transfer which are governing the process performance.

1.3.3.1 Heat transfer

According to El-Bourawi et al. (2006), the heat transfer could be divided into four distinct

steps: the first step is the heat transfer from the circulating feed stream to the membrane surface

which is made through the membrane thermal boundary layer on the feed side; the second step

is realized by conductive heat transfer across the membrane matrix and its pores that are

generally filled with gas; the third step is the heat transfer related to the vaporization latent heat

which is considered as the efficient heat in the membrane distillation process; the fourth step

of heat transfer does occur from the permeate side membrane surface to the circulating

permeate stream and this through membrane’s thermal boundary layer.

The first and fourth heat transfer steps are associated to the temperature polarization effect

which is equivalent to a heat transfer resistance on both membrane sides resulting in a decrease

in the MD driving force that could even reach 80% and this due to the lower temperature

difference on the membrane surface in comparison to the bulk temperature difference (Lawson

et al., 1996; Martinez-Diez et al., 1999).

1.3.3.2 Mass transfer

In direct contact membrane distillation, the mass transfer, which is expressed generally in mass

flux (J), is created through the vapor pressure gradient between the feed and permeate streams

1.Literature survey

1.3.Membrane distillation

Sana Abdelkader

that enter in direct contact with the two membrane surfaces. During the mass transfer

mechanism, three different steps could be distinguished (Onsekizoglu, 2012): the water that

enters in contact with the membrane surface on the feed side is evaporated to transfer afterwards

across the hydrophobic membrane through the pores and finally the water vapor condensates

at the permeate side on the membrane vapor/liquid interface.

1.3.4 Membranes in DCMD

One of the most important characterizations of MD process is the hydrophobic microporous

membrane through which water vapor is transferred due to a temperature gradient created

between the feed and the permeate streams. (Camacho et al., 2013; Ashoor et al., 2016).

The membranes utilized in MD process are generally made from polytetrafluoroethylene

(PTFE), polypropylene (PP) or polyvinylidene fluoride (PVDF) with low resistance to mass

transfer and in order to ovoid heat losses across the membrane, it should also have low thermal

conductivity with good resistance to high temperatures and chemicals (Alkhudhiri et al.,

2012).

In addition to the aforementioned specifications, supplementary characteristics should be

available in the hydrophobic membrane in order to have good performance of the membrane

distillation process. Among the important membrane properties, we can distinguish: membrane

thickness, porosity and nominal pore size as well as liquid entry pressure (wetting pressure)

(Drioli et al., 2009). Membrane thickness has an inversely proportional relationship with the

permeate flux since increasing the thickness has direct positive effect on the membrane mass

transfer resistance increase (Alkhudhiri et al., 2012). The membrane porosity in MD

membranes varies between 30% and 85% and it is equivalent to the percentage of the pores

volume divided by the total membrane volume which means that when increasing the

membrane’s porosity the permeate flux increases (Alkhudhiri et al., 2012; El-Bourawi et al.,

2006).Membrne nominal pore size is a sensitive parameters since a choice has to be made to

operate with the appropriate pore size (having dimensions that are large enough to have good

permeate fuxes) and, at the same time, maintaining a good permeate quality since membranes

with large pore sizes are more vulnerable to wetting (Drioli et al., 2014). Thus, an optimum

pore size should be considered depending on the feed quality to avoid the pores wettability (El-

Bourawi et al., 2006). As for the membrane Liquid Entry Pressure (LEP), this parameter has

significant effect on the process performance and presents an important indication of the

membrane quality and the MD process performance, since the feed pressure should not exceed

the membrane’s LEP in order to maintain the its hydrophobic character (Alkhudhiri et al.,

1.Literature survey

1.3.Membrane distillation

Sana Abdelkader

2012). Moreover, the membrane’s LEP is significantly affected by the feed concentration and

organic content and other solutes in the feed stream such us ethanol which usually reduces the

membranes hydrophobicity (Alkhudhiri et al., 2012; Gostoli, 1989).

The enhancement of MD process has drawn attention to the improvement of the membrane

hydrophobicity using multiple techniques such as the utilization of hydrophobic surface

modifying macromolecules on poly(vinylidene fluoride) hydrophobic composite membranes

which has led to promising results in sea water MD treatment (Prince et al., 2014.a). Moreover,

multiple studies have focused on the enhancement of the driving force for the water vapor

transport in MD process through the development of hydrophobic membranes that have high

resistance to pore wetting which did lead to significantly higher flux levels along with high

rejection rates (Prince et al., 2012; Prince et al., 2014.b; Efome et al., 2016).

1.3.5 Fouling and wetting in DCMD

One of the challenges that faces MD is the capability of treating real industrial effluents that

could contribute in the creation of new alternative water resources for the industry, knowing

that the feed quality has an important impact on the membrane fouling, by the means of the

variable organic and/or mineral components that are present in the feed streams (Ashoor et al.,

2016).

According to Tijing et al. (2015),there are different parameters that affect and govern the

membrane fouling in MD process which could be classified into four different groups, namely:

the feed water characteristics (organic load, ionic strength, pH, etc.), the operational conditions

(flow rate, liquid streams temperature, etc.) the membrane properties (pore size,

hydrophobicity, etc.) and the foulant characteristics (such as its concentration, molecular size,

solubility, hydrophobicity, etc.). The effluent’s organic content is one of the most affecting

factors responsible of the membrane alteration in membrane processes, especially in MD, by

its deposition on the hydrophobic membrane surface during the process and which increases

the possibilities of fouling phenomena (Naidu et al., 2014).

In addition to the organic fouling, mineral fouling (scaling) could also occur during MD process

and has significant effect on the flux decline such as in the case of CaCO3 precipitation on the

membrane surface which reduces the membrane permeability and rises the temperature

polarization effect in addition to its influence on the membrane wettability after the formation

of a deposit on the membrane surface (Gryta, 2008).

1.Literature survey

1.3.Membrane distillation

Sana Abdelkader

1.3.6 Pretreatment and membrane cleaning

Multiple strategies could be considered to reduce membrane fouling in MD among which we

find effluent pretreatment using multiple methods that include sedimentation, acidification and

degasification, coagulation/flocculation, thermal processes and membrane pretreatments which

have shown an undeniable positive effect in the MD process enhancement and such filtration

pretreatments are gaining more attention due to their ability to efficiently remove of different

particles and large macromolecules in the feed stream (Tijing et al., 2015).

The membrane chemical cleaning could also be considered as a solution to remove the

deposited fouling layer, periodically. Nevertheless, cleaning could also contribute in the

membrane wettability and such risk imposes a limited membrane cleaning without complete

dissolution of the deposit (Gryta, 2008). The chemical cleaning involves generally the

circulation of heated acids and basic solutions to address the both the inorganic (rinsing with

acid solution) and the organic fouling (rinsing with basic solutions) which could relatively

allow to regain the initial flux in the case where the fouling layer is deposited on the membrane

surface but if the fouling is located inside the membrane pores, the flux recovery would be very

difficult to achieve and the membrane will lose in hydrophobicity with the repeated cleaning

(Tijing et al., 2015).

1.3.7 Fields of application

Based on its high separation performance, MD is being investigated and applied for water, and

in some cases, nutrients recovery from various types of effluents, taking into consideration the

effect of the experimental parameters (Izquierdo-Gil et al., 1999; Singh et al., 2013; Jia et

al., 2017). In addition to its high salt rejection rates, MD is characterized by good rejection

rates of non-volatile and low adsorptive organic compounds (Carnevale et al., 2016; Plattner

et al., 2017).

With respect to MD applications, MD is showing promising performances when it comes to

water desalination and some food processing applications. There have been multiple

investigations to separate different types of components from water: non-volatile compounds

(ions, colloids and macromolecules) or traces volatile organics for environmental purposes

(wastewater treatment and reuse, desalination, etc.) also MD is tested for food processing (juice

and milk concentration) and biomedical applications (El-Bourawi et al., 2006; Onsekizoglu,

2013).

1.Literature survey

1.3.Membrane distillation

Sana Abdelkader

1.3.7.1 Water desalination

With respect to water desalination, some mathematical theoretic models were developed to

simulate the membrane distillation process for multiple membrane module configurations in

order to enhance their performances (Laganà et al., 2000; Boukhriss et al., 2015; Deshpande

et al., 2017). According to Khayat et al. (Khayet et al., 2007), the application of response

surface methodology to optimize salt aqueous solutions treatment by direct contact membrane

distillation is suitable to assess the permeate fluxes for both commercial- and laboratory-made

membranes. Moreover, they had proposed an algorithm which can help in the searching step

for optimum localization considering as factors the stirring rate, the feed temperature and the

salt concentration. Another optimization study carried by Cheng et al. (2016) using response

surface methodology (via the application of quadratic rotation-orthogonal composite design)

has proven that both operating conditions and membrane module parameters optimization can

lead to a remarkable improvement of the process performance. The recorded enhancement was

in terms of average permeate flux, water productivity per unit volume of module, water

production per unit energy consumption, and comprehensive index.

1.3.7.2 Wastewater treatment

The interest in membrane distillation systems has led to more emphasis on the process

optimization in order to achieve the highest possible permeate fluxes, in other terms higher

recovery rates. In their study, Zhang et al. (2016) found that an optimization of saline

wastewater treatment by vacuum membrane distillation using PP hollow fiber membrane had

increased the water recovery percentage to 88.6% with an improvement of the permeate quality

in terms of electrical conductivity and total suspended solids removals in comparison to one-

stage vacuum membrane distillation.

The MD process has been successfully studied for purification of waste waters of

pharmaceutical and textile industries as well as underground waters contaminated with heavy

metals. More recently, the feasibility of applying membrane distillation process for recovering

potable water from arsenic, uranium and fluoride contaminated brackish waters has been

proposed. High quality permeates with dissolved solids concentrations less than 20 ppm (more

than 99% rejection of salts) along with arsenic, fluoride and uranium contaminant reductions

in the range of 96.5– 99.9% were reported (Yarlagadda et al., 2011).

1.Literature survey

1.3.Membrane distillation

Sana Abdelkader

1.3.8 Utility theory

Utility theory is a well-known and well-established notion that is applied in different fields for

decision making purposes through the analyses of an individual's preference-indifference

relation (Kaplinski, 2013).

On the practical level, utility theory takes into consideration assumptions about a person's

preferences which gives the opportunity to present them numerically in useful ways and to help

the decision maker in choosing his preferences among multiple complicated alternatives

(Fishburn, 1968). Utility theory could be utilized to analyze the customer’s attitudes towards

taking the risk and this can explain the mechanism behind the decision-making process, with

emphasis on the merge of the economic and psychological aspects (Kaplinski, 2013).

In order to fulfill this objective, a utility function, which is a mathematical function, could be

developed to show the preference of an individual who is facing various complex alternatives

(Cohon, 1978).

The utility theory has been applied in this work to assess the performance of the saline dairy

effluent DCMD treatment with respect to two types of pretreatments in order to evaluate the

two pretreatment scenarios in term of cost-effectiveness.

2.Material and methods

2.1.Studied saline effluents

Sana Abdelkader

2 MATERIAL AND METHODS

2.1 STUDIED SALINE EFFLUENTS

2.1.1 Synthetic saline wastewater

The Synthetic wastewater used throughout the studies was composed of 1 g/L C6H12O6, 1 g/L

NH4Cl, 0.3 g/L KH2PO4, 2 g/L MgCl.6H2O, 0.2 g/L CaCl2.2H2O, 1 g/L C2H3NaO2.3H2O, and

various concentrations of salt (0–3% NaCl).

Trace elements solution was composed of 3 g/L MgSO4.7H2O, 0.5 g/L MnSO4.2H2O, 1 g/L

NaCl, 0.1 g/L FeSO4.7H2O, 0.1 g/L CaCl2.2H2O, 0.18 g/L ZnSO4.7H2O, 0.01 g/L

CuSO4.5H2O, 0.01 g/L H3BO3, 0.01 g/L Na2MoO4.2H2O, 0.25 g/L NiCl2.6H2O. One milliliter

of stock trace element solution was added to 1 L of synthetic feed wastewater. The pH was

adjusted to pH=7 by adding required amounts of 5% phosphoric acid and/or 5% KOH

(Kapdan and Erten, 2007). All chemicals were analytical grade and were used as received.

In section (3.2.), a new feed solution was prepared for each experiment in the optimization

study and various concentrations of NaCl and glucose (C6H12O6) were added to simulate

different saline wastewater compositions. The chosen ranges of variation of COD and [NaCl]

have average levels and are as follows: the COD varies from 0.3 to 10 gO2/L and the [NaCl]

from 10 to 30 g/L. The COD and [NaCl] variation ranges are including different effluent

qualities in terms of salinities and organic loads (Kapdan and Erten, 2007).

2.1.2 Saline cheese whey wastewater

The studied salty dairy effluent was kindly provided by the cheese industry Bergpracht

Milchwerk GmbH & Co, Germany. This industry is specialized in the manufacturing of four

types of cheese: Camembert, Feta, Mozzarella and White cheese where the salt addition is

made in different steps. The added salt ends up in the final wastewater, which includes,

essentially, effluents from the cleaning process, the cheese whey and the cheese salt baths. The

company discharges 100 m3 of wastewater per day that are continuously collected in

neutralization tanks before sending it to the wastewater treatment plant. The samples were

collected from the neutralization tanks and stored at 2°C before the experiments and analyses.

Raw wastewater quality is presented in Table 3.

The studied raw effluent is characterized by high electrical conductivity and total organic

carbon concentration (TOC). According to previous studies, the selected cheese whey

2.Material and methods

2.2.Biological treatments of the saline synthetic wastewater

Sana Abdelkader

wastewater has a fluctuating composition that is closely linked to the cheese making and the

cleaning processes (Carvalho and al., 2013).

2.2 BIOLOGICAL TREATMENTS OF THE SALINE SYNTHETIC

WASTEWATER

2.2.1 Aerobic bioreactors

Four aerobic reactors, each having a capacity of 1.3 liters, were fed with the synthetic

wastewater as feed solution (pH=7± 0.5) with addition a different NaCl concentration for each

reactor: 0%, 0.5%, 1% and 2%. The reactors were inoculated with activated sludge from a

municipal wastewater treatment plant (Okhla, New Delhi, India) and the sludge was

acclimatized for one week. Coarse air diffusers were used to supply the required oxygen in the

reactor. The aerobic reactors were operated in a continuous mode at 24°C ± 1°C with a fixed

Hydraulic Retention Time (HRT) of 6 hours (Figure 6).

Figure 6: Aerobic biological reactors

The Mixed Liquor Suspended Solids (MLSS) of the aerobic reactors was measured every week.

Excess sludge was removed, as required, to maintain the MLSS in the reactors between 8 and

9 g/L. After biological treatment, the sludge was allowed to settle for 30 min and the

supernatant was subjected to analysis and membrane filtrations. After the supernatant’s

removal, and before running the next set of experiments, the sludge was washed with distilled

water to avoid any salt accumulation from the previous treatment.

2.Material and methods

2.2.Biological treatments of the saline synthetic wastewater

Sana Abdelkader

2.2.2 Anaerobic bioreactors

In order to obtain an anaerobic media, 450 ml of the synthetic wastewater was put in 1 L

incubation bottle which served as an anaerobic reactor, then heated at 60-70°C for 2-3 hours.

The oxygen content in the reactors is removed by sparging the bottle with nitrogen for 20 min

and the addition of 0.25 g of a reducing agent (L-cysteine) before sealing and autoclaving the

bottle (Figure 7). After autoclaving, the media was inoculated with 50 ml of anaerobic sludge

previously enriched in the same media for two weeks.

Figure 7: Anaerobic biological reactors during the nitrogen sparging step

The anaerobic sludge was collected from a 2-stage anaerobic reactor treating food waste

located at TERI Gual Pahari Campus, Gurgaon, India. Five identically set-up anaerobic

reactors were preparedwith the synthetic feed (pH=7± 0.5), each reactor is containing a

constant NaCl concentration: 0%, 0.5%, 1%, 2% and 3%, respectively. The reactors were

incubated at 37°C for 14 days.

2.Material and methods

2.2.Biological treatments of the saline synthetic wastewater

Sana Abdelkader

2.2.3 Microfiltration and ultrafiltration membranes

All filtration experiments were conducted at 24 ± 1°C. The microfiltration tests of the reactors’

supernatants were carried out using ceramic membranes made of sugarcane bagasse fly ash in

TERI were assembled into a flat-sheet module and were used to filter the biologically treated

wastewater. The ceramic membrane filters, (Figure 9), have surface porosity of 30% and

average pore diameter of 1.6 µm, as determined by image J (NIH, USA) (Marel et al., 2010).

The filtration setup is composed of a peristaltic pump (Electrolab peristaltic pump PP-VT)

operating at 13 rpm corresponding to 0. 6 L/h, linked to a 5.1 cm x 5.1 cm ceramic filter. Two

ceramic filters were joined in a polyacrylic support frame, so the effective filtration area was

50 cm2 and the set was submerged in a 500 ml glass beaker that contains the reactors

supernatant. The pressure was determined by a mercury manometer connected to the filter

(Figure 8).

Figure 8:microfiltration membrane and set-up

Sampling was done every 10 min. The conductivity of obtained permeate was measured

immediately before the sample was stored at in the refrigerator for further analysis.

2.Material and methods

2.2.Biological treatments of the saline synthetic wastewater

Sana Abdelkader

Figure 9: SEM image of the ceramic membrane (left: membrane cross section, right: top

view of separation layer)

The ultrafiltration tests were performed by the means of an asymmetric polymeric membrane

(Polysuphone PSF 1700), with porous support layer overlain by thin separating layer (Figure

10), which was prepared by phase inversion technique with 18% Polysulphone (PSF). The

porosity and the pore size of the membrane were 42±1% (determined by gravimetric method)

and 0.01-0.0095 µm (determined by filtration velocity method) respectively.

Figure 10: SEM image of the cross-section of PSF (18%) membrane

Coupons with 47mm diameter and an effective filtration area of 0.00134 m2 were used for

batch filtration in a stirred cell formed of metal base and cover, in between a glass cylinder

maintains the coupon with a plastic ring and the whole set is pressurized using nitrogen gas

(Figure 11).

2.Material and methods

2.3.Direct Contact Membrane Distillation (DCMD)

Sana Abdelkader

Figure 11:Ultrafiltration set-up

Pure water flux was measured before filtration of the biologically treated wastewater. 50 ml

wastewater was filtered at 1 bar applied pressure. The filtration was done for 20 minutes before

starting to take samples every 10 minutes.

2.3 DIRECT CONTACT MEMBRANE DISTILLATION (DCMD)

2.3.1 Benchscale DCMD setup

DCMD apparatus used in the section (3.2.) is composed of a bench scale direct contact

membrane distillation set-up composed mainly of one heating and one cooling thermostatic

water baths connected to the membrane cell via two peristaltic pumps. The two thermostatic

cycles (heating closed loop for feed and cooling closed loop for permeate) are connected to a

membrane cell made of Plexiglas that has an effective membrane surface of 0.0032 m2 (Figure

12).

2.Material and methods

2.3.Direct Contact Membrane Distillation (DCMD)

Sana Abdelkader

Figure 12:(a) Bench scale DCMD set-up and (b) schematic of the DCMD set-up [(1)

membrane test cell, (2) circulation pump, (3) heat exchanger, (4) feed water tank (5)permeate

water tank (6)electronic balance].

New Hydrophobic PTFE membrane sample, with a nominal pore size of 0.45μm (Table 2),

was placed inside the module for each experiment. Permeate electrical conductivity and

temperatures of the feed and permeate solutions were measured continuously during the

experiments with variation ranges of [26-80°C] and [18-23°C], respectively.

(a)

(b)

2.Material and methods

2.3.Direct Contact Membrane Distillation (DCMD)

Sana Abdelkader

All experiments were conducted during 90 minutes in a counter-courant configuration to

maintain the temperature gradient.

The permeate flux (Jp) is by definition the flux of water vapor which is crossing the

hydrophobic membrane through its pores and it is proportional to the vapor pressure difference

across the membrane following the law of Darcy:

𝐽𝑝=𝐵𝑚 (𝑃𝑓− 𝑃𝑝)

Where Bm is the membrane coefficient, Pf is the vapor pressure (bar) at the feed side

water/vapor interface and Pp is the vapor pressure at the permeate side water/vapor interface.

Besides, it is noteworthy to indicate that the water vapor is an exponential function of the

operating temperature according to Antoine equation:

𝑃0=exp(23.238−3.841

𝑇−45)

Such that, P0 is the pure water vapor (Pa) and T is the water temperature (K).

The volumetric permeate flux (L/m2h) is expressed as follows:

𝐽𝑝=∆𝑉

𝑆.∆𝑡

Such that, ∆V is the permeate volume (L), S is the effective membrane surface (m2) and ∆t is

the DCMD operating time (h).

2.3.2 Laboratory scale DCMD setup

Direct Contact Membrane Distillation (DCMD) experiments were performed using a

laboratory scale unit developed and built by Fraunhofer ISE, Freiburg, Germany. The

experimental setup is formed of a 150*250 mm2 test cell; a detailed description of the setup

(Figure 13) is listed in Winter et al. (2013). For each DCMD experiment, a new membrane

sample was placed in the test cell. All experiments were run for several hours (at least 5 hours)

and the process evaluation was made after 3 hours. The DCMD plant is equipped with sensors

and controllers that record multiple parameters in the systems every 5 seconds: inlet and outlet

temperatures of both, the feed and permeate (°C), feed and permeate flow rates (L/h), feed inlet

pressure (bar) and feed tank weight (kg). The feed inlet and permeate outlet temperatures (Tei

and Tco, respectively) were controlled automatically to set a constant temperature difference

2.Material and methods

2.3.Direct Contact Membrane Distillation (DCMD)

Sana Abdelkader

(∆T=Tei-Tco). When varying the temperature difference in the experiments, only the permeate

flow temperature was varied (from 25 to 40°C) and the feed flow has had a constant

temperature of 55°C through all experiments. The temperature range was set between 25°C

and 55°C for two main reasons: to preserve the proteins’ quality for possible valorization and

recovery of the concentrate, on one hand, and on the other hand, to reduce the energy

consumption during the DCMD treatment. Besides, the feed and permeate velocities (Ve and

Vc, respectively) were controlled and were maintained at equal values during the experiments.

For all experiments, the tests started with a feed tank volume of around 8 liters and ended with

1.5 to 2 liters of concentrated feed.

Reference tests were conducted, using tap water as the feed stream, before each experiment at

the same operational conditions to control the setup stability and functioning. The obtained

reference flux served as a reference to study the fouling and wetting phenomena.

(a)

2.Material and methods

2.3.Direct Contact Membrane Distillation (DCMD)

Sana Abdelkader

Figure 13: (a) Laboratory scale DCMD set-up and (b) Schematic diagram of experimental

laboratory scale DCMD set-up; (1) membrane test cell, (2) feed water tank, (3) feed pump,

(4) heat exchanger, (5) support heat exchanger, (6) electric heater, (7) heat exchanger, (8)

motor valve, (9) laboratory cooling circuit, (10) filter, (11) electronic balance (Winter et al.,

2013).

Concerning both the membrane and setup cleaning, it was performed using Cleaning in Place

(CIP) method by circulating three different aqueous solutions at 70°C for 20 minutes for each

one of them: 0.5% Sodium Hydroxide (NaOH) solution followed by water circulation and then

0.5% nitric acid (HNO3) solution, respectively. One final rinsing step with water at room

temperature was performed in the end. The setup cleaning procedure was carried after each

DCMD test.

2.3.3 DCMD membranes

Two commercial flat sheet polytetrafluoroethylene (PTFE) membranes were investigated. The

utilized membranes were of two different brands (Membrane Solution (M1) GORE membrane

(M2)) with two different pore sizes and backing structures (Table 2): the first membrane (M1)

has a polypropylene (PP) nonwoven backing and the second membrane (M2) has a PP scrim

backing. In addition to the difference of the nominal pore sizes between M1 and M2 and the

different hydrophobicity (see LEP and contact angle) the backing structure of M2 gives a higher

resistance to the membrane which can be identified by its thickness (Table 2). During all

experiments, the PTFE surface was orientated towards the feed flow.

(b)

2.Material and methods

2.4.DCMD pretreatments

Sana Abdelkader

Table 2: DCMD membranes characterization (source: membranes data sheets, *parameters

measured in the laboratory)

Membrane

Material

Nominal

Pore Size

（μm）

Thickness

(μm）

Porosity

(%)

Tortuosity

LEP*

(bar)

Contact

Angle*

(°)

PTFE

0.45

160±40

2.3

123.4

backing

PTFE

0.2

1.3

3.9

131.2

backing

280

0.5

1.8

For each experiment, a new membrane was applied which finally was used for further analyses

and fouling diagnostic after finishing the test.

2.4 DCMD PRETREATMENTS

In order to study the influence of the dairy effluent pre-treatment on the DCMD performance,

two pretreatments were adopted to process the raw cheese whey wastewater, namely

microfiltration and ultrafiltration processes.

2.4.1 Macrofiltration (MAF)

The raw effluent was first subjected to a macrofiltration (MAF) step using a 30µm pleated

nonwoven Polyester filter cartridge (Pentek Filtration R30). Macrofiltration pretreatment will

help to decrease the organic load in the feed, (Table 3), by the elimination of all suspended

solids, colloids and macromolecules that have bigger size than 30µm.

2.4.2 Ultrafiltration (UF)

For further processing, after the MAF step an ultrafiltration (UF) step is performed using a

tubular PES Multibore® membrane 0.9 module (Inge GmbH) with a pore size of approximately

0.02 µm and a molecular weight cut-off (MWCO) = 100 KDa. The UF was applied in cross-

flow filtration. And the tubular membrane module has capillaries with an internal diameter =

0.9 mm, external diameter = 4 mm and pores diameter ≈ 0.02 µm.

2.Material and methods

2.5.Analytical methods

Sana Abdelkader

In this investigation, all traces of casein micelles (large aggregates = 105KDa) which are

causing the effluent’s whitish color are eliminated by the UF step, in addition to the traces of

fat globules. However, the UF pretreated feed still contains the majority of the whey proteins

(β-lactoglobulin and α-lactalbumin) (Singh and al., 2014).

Table 3: Characterization of the raw and pretreated effluents

Parameters

Raw effluent

MAF

permeate

MAF-UF

permeate

pH (± 0.1)

6.8

Conductivity (± 0.4 mS/cm) at 25°C

9.9

TOC (mg/L)

1,730 (± 45)

784 (± 27)

293 (± 13)

TC (mg/L)

1,860 (± 45)

851 (± 27)

344 (± 13)

TIC (mg/L)

130 (± 45)

67 (± 27)

51 (± 13)

Both obtained permeates, by MAF and UF pre-treatments, were characterized before being

used for further DCMD experiments (Table 3). The measures illustrated in Table 3 are the

mean average of the results of three samples characterization.

2.5 ANALYTICAL METHODS

2.5.1 Pollution parameters

2.5.1.1 Mixed Liquor Suspended Solids (MLSS)

The sludge’s MLSS was determined according to the “Standard Methods for the Examination

of Water and Wastewater”.

2.5.1.2 Gas Chromatography analysis (GC)

The composition of the gas formed during anaerobic process was analyzed using gas

chromatography (GC). This analysis was performed immediately after taking the anaerobic

reactors out of the incubator and before unsealing them for further analysis. Agilent

Technologies system number 7890A was used for this analysis. The GC analysis were

2.Material and methods

2.5.Analytical methods

Sana Abdelkader

conducted at a total flow rate= 6ml/min using a Nucon SS packed column with Argan as

reference gas (flow rate=10ml/min) and Nitrogen for the makeup flow (flow rate=3 ml/min).

2.5.1.3 Chemical Oxygen Demand (COD)

The Chemical Oxygen Demand (COD) was measured depending on the samples’ salinity: the

permeate quality was tested according to the “Standard Methods for the Examination of Water

and Wastewater” and for the samples with high electrical conductivity the COD was

determined using the modified method described by Vyrides and Stuckey (2009.a).

2.5.1.4 Microbiological observation

The bacterial Gram staining was realized using a Gram stain kit and the followed by a

microscopic observation were conducted at 100 times magnification.

2.5.1.5 Turbidity

The turbidity was measured using a “Bionics” digital Nephelometer, model No: BST/DN-341.

2.5.1.6 Total Organic Carbon (TOC)

Permeate samples were analyzed to measure their Total Organic Carbon (TOC) using the

difference method by cuvette tests (LCK 380 and LCK 381) purchased from Hach Lange,

Germany.

2.5.1.7 Electrical conductivity (EC)

In section 3.1., EC was measured directly by a “Bionics” conductivity meter: Digital

conductivity meter model NO: BST/CM-601.

During all DCMD experiments, the permeate quality was monitored by measuring its electrical

conductivity in the permeate tank every hour using a digital conductivity meter (Kobold, HND-

C) and using an S823 pH/Conductivity meter.

2.5.2 Membrane characterization

2.5.2.1 Liquid Entry Pressure (LEP)

The LEP was measured using 0.5 wt. % NaCl solution to which a gradually increasing pressure

perpendicularly to the PTFE membrane surface with a fixed pressure slope (dp/dt= 0.01 bar/s)

was applied. Instantly, when the liquid solution breaks through the membrane the

corresponding pressure (LEP) is recorded. The LEP setup and experimental procedure are

developed and designed at Fraunhofer ISE.

2.Material and methods

2.5.Analytical methods

Sana Abdelkader

2.5.2.2 Contact Angle (CA)

Contact angle measurements were performed to analyze the membranes hydrophobicity degree

using an automated method of the DataPhysics Contact Angle System OCA and the

DataPhysics SCA20 software. The contact angle measurements were performed by an

automatic generation of water drops using 5 μm dispenser. Pure water was utilized as reference

at 22°C.

The measured contact angles in Table 2 show that the membrane has a good hydrophobic

nature that forms a physical barrier in front of the feed water stream (CA > 90°). According to

the literature, the utilized membranes has good contact angle values in comparison to other

synthesized and modified hydrophobic membranes in which its value varies between 112.7°

and 154.2° (Prince et al., 2012; Prince et al., 2014.b; Efome et al., 2016). Moreover,

according to the aforementioned references, this parameter could contribute in the

improvement of the DCMD performance.

2.5.2.3 Scanning Electron Microscopy (SEM)

The SEM analysis was used to determinate the morphology of the fouling layers after DCMD

experiments by a scanning electron microscope «Auriga 60" from the company Zeiss and for

the detection and quantification of the fouling layer, an Energy Dispersive X-ray (EDX)

Detector "XFlash Detector with internal FET" from the company Bruker was used for the

detection and quantification of the fouling layer.

For the membrane samples coating, a Cressington Sputter Coater 108 auto was used at 30 mA

and 0.06 mbar for 30s to 40s. A gold layer of ~10nm thickness was necessary to get an adequate

conductivity to the undercuts of the membranes. For cross-section SEM images, an additional

step was added in which the samples were frozen using liquid nitrogen.

2.5.3 Permeate flux and permeability

In section (3.1.), the flux was measured volumetrically (permeate volume collected over a

known time). Applied pressure was measured using a mercury manometer (mm Hg) for the

ceramic membrane filtration. For the polymeric membrane, pressure was measured using a

pressure gauge fixed to the nitrogen cylinder used for pressurizing the filtration cell.

The flux (J) is calculated using the following formula where flux is in LMH (L/m2.h):

𝐽=𝑉𝑜𝑙 (𝐴×𝑡)

⁄

2.Material and methods

2.6.Response Surface Methodology (RSM)

Sana Abdelkader

Permeability (P)is expressed in LMH/bar [which is the abbreviation of the permeability unit

(L/m2.h.bar)] and it was calculated using the following equation:

𝑃 =𝑉𝑜𝑙 (𝐴×𝑡×∆𝑝)

⁄

Where: Vol = Permeate volume (liter)

A = Membrane filtration area (m2)

∆p = Transmembrane pressure (bar)

t = Filtration time (hour)

2.6 RESPONSE SURFACE METHODOLOGY (RSM)

Response Surface Methodology is applied in this study to optimize the DCMD process for

saline wastewater treatment through one of the most used forms of RSM: The Central

Composite Design (CCD). This method, which is fit for a quadratic surface, is generally

utilized to optimize the effective parameters of a process and to identify their existing

interactions and their extent with a minimum number of experiments. The CCD is a second-

order design which is based on adding a number of (2k) axial-points experiments and a number

of center-points replications (n0) to a simple first order design (2k) where k is the factors

number. The added experiments serve in helping to get more information about the response

surface and in the optimization of the process through a second-order model (Khuri &

Mukhopadhyay, 2010).

The CCD model is a mathematical polynomial function expressed as follows, in the case of

four independent variables:

𝒀= 𝑏0+ 𝑏1𝑿𝟏+𝑏2𝑿𝟐+𝑏3𝑿𝟑+𝑏4𝑿𝟒+𝑏11𝑿𝟏

𝟐+𝑏22𝑿𝟐

𝟐+𝑏33𝑿𝟑

𝟐+𝑏44𝑿𝟒

𝟐+ 𝑏12𝑿𝟏𝑿𝟐

+𝑏13𝑿𝟏𝑿𝟑+𝑏14𝑿𝟏𝑿𝟒+𝑏23𝑿𝟐𝑿𝟑+𝑏24𝑿𝟐𝑿𝟒+𝑏34𝑿𝟑𝑿𝟒

In our study, with a factors number k=4, we have carried a total number of 28 experiments that

included 24 orthogonal design points, 8 axial points (α=±1.682) and 4 replications of the center-

points (Table 4).

2.Material and methods

2.7.Single-Node model simulation

Sana Abdelkader

The analysis of variance (ANOVA) was applied to determine the statistical significance of the

obtained model and to study the quality of the model’s fit. The statistical and graphical analysis

tool used in this work is Minitab® 17 software by Minitab Inc.

In our study, the volumetric permeate flux Jp (L/m2.h) is considered as the main response (Y)

to be optimized by CCD through the optimization of four different factors [X1: temperature

difference ΔT (°C), X2: feed velocity Vf (m/s), X3: NaCl concentration [NaCl] (g/L) and X4:

glucose concentration [Gluc] (g/L)]. The first step to apply the CCD is to identify the factors

and their different levels. Table 4 illustrates the four chosen independent factors in our study

with their respective coded levels and actual values. The choice of the variation ranges of each

parameter was made according to preliminary tests with respect to the volumetric permeate

flux (Jp) as the main response monitored during this study.

Table 4: Central Composite Design (CCD) variables levels in their coded and actual values.

2.7 SINGLE-NODE MODEL SIMULATION

As an evaluation tool for the DCMD process performance, one of the parametric node models

developed by Winter (2015) was utilized in order to predict the transmembrane vapor flux as

well as the thermal efficiency (based on calculations of mass and heat transfers) in the studied

laboratory scale DCMD configuration where the operating conditions, materials and channel

configuration were taken into consideration. The C programming language was used for the

model implementation and the different mass and heat transfer calculations were carried out

via an iterative algorithm. Since the model conception is made according to a steady-state

scenario and homogeneous conditions, its application in our study could present an interesting

indicator of the fouling and wetting phenomena effects and the extent of their influence on the

process performance in comparison to the ideal DCMD treatment outcome.

Factors

Symbols

Actual values of the coded levels

-α

-1

+α

Temperature difference ΔT (°C)

4.77

55.23

Feed velocity Vf (m/s)

0.027

0.039

0.057

0.075

0.086

Initial salt concentration [NaCl] (g/L)

10.07

14.1

25.9

29.92

Initial glucose concentration [Gluc] (g/L)

0.49

4.45

10.25

16.05

2.Material and methods

2.8.Experimental performance analysis

Sana Abdelkader

2.8 EXPERIMENTAL PERFORMANCE ANALYSIS

The main purpose of this section is to extend the process performance analysis by incorporating

the cost-effectiveness analysis of the saline dairy wastewater DCMD process associated with

the two different pretreatments. To this purpose, the set of parameters (i.e. criteria) is first

introduced. These parameters have been shown in the literature to affect the applicability and

the performance of the DCMD process in the industry. Based on these parameters, a set of

indicators is defined that will be used in our cost-effectiveness analysis relying on the utility

theory. Then, a mathematical model is presented to assess the performance of the scenarios that

have been experimented using our setup. Finally, the results are discussed based on the present

case study in order to analyze the cost-effectiveness of the DCMD process with respect to the

two pretreatment options. In addition to the indicators used to assess the efficiency of the

filtration process, the time needed to reach a specific level of performance is considered as a

measurement of the cost-effectiveness of each of our scenarios. In fact, time is strongly

correlated to the cost in the context of the DCMD process since the components of the cost

(e.g., energy consumption, membrane lifetime) vary according to time.

The criteria as well as their corresponding indicators used to assess the performance of DCMD

in the context of the saline diary effluents treatment are displayed in Table 5. It is noteworthy

that these criteria have been already used in the literature to evaluate the performance of the

DCMD process.:

1. Rejection factors: One of the most important advantages of DCMD is its ability to

provide a better quality of the output water compared to other technologies. Production

of high-quality water is well-established with rejection factors of almost 100% of non-

volatile compounds (Khayat and Matsuura, 2011). We use in our work the electrical

conductivity of the DCMD permeate and its total organic carbon (TOC) to evaluate the

performance of the DCMD process with regard to this criterion.

2. Permeate flux: it refers to the ability of the DCMD process to achieve good

performance in terms of mass transfer. From the practical point of view, the flux varies

following different parameters including the thickness and the morphology of the

membrane. Therefore, it is trivial to state that this characteristic is substantially affected

by the DCMD process since during the effluent treatment the surface of the membrane,

its porosity and its thickness could be affected (Drioli et al., 2009; Alkhudhiri et al.,

2012).

2.Material and methods

2.8.Experimental performance analysis

Sana Abdelkader

3. Energy consumption: As for any other distillation process, this technology also requires

energy for water evaporation. Henceforth, we consider the amount of energy spent to

conduct the water treatment process as a comparison criterion between the

implementation scenarios that will be described in the sequel. Mainly, intel cold

temperature, intel hot temperature, flow rate and pretreatment energy are the four

indicators used to assess the energy consumption. The pretreatment energy is the energy

spent to conduct the pretreatment process varies according to the characteristics of the

equipment used to this purpose. The interest of considering this parameter stems from

the fact that it is one of the fundamental components of the Capital Expenditure

(CAPEX) and the Operational Expenditure (OPEX) of the global effluent treatment

solution.

4. Void volume: This criterion expresses the void volume fraction of the membrane

(defined as the volume of the pores divided by the total volume of the membrane).

Higher porosity membranes have a larger evaporation surface area. In our experimental

setup, we use the Scanning Electron Microscopy (SEM) as an indicator to estimate the

membrane fouling. This indicator is very important since, beyond a low porosity

threshold, the performance of the DCMD process becomes unacceptable and the

membrane must be either cleaned or changed.

The void ratio in the SEM images can be estimated if the regions corresponding to the

membrane are known. In fact, the void volume fraction of the membrane, defined as

the volume of the pores divided by the total volume of the membrane, is proportional

to the density of the membrane material (Khayet and Matsuura, 2001). In fact, the

Smolder-Franken equation provides an expression of the void ratio:

𝜚 =1− 𝜚𝑚

𝜚𝑝𝑜𝑙,

where𝜚𝑚 and 𝜚𝑝𝑜𝑙 refer to the densities of membrane and polymer materials, respectively.

In order to estimate the void ratio, we rely on the following three-step process:

a. The SEM image is segmented in order to identify the regions corresponding to the

membrane region.

b. The segmented image is binarized in order to convert the values of the pixels to black

(non-membrane regions) and white (membrane regions).

2.Material and methods

2.8.Experimental performance analysis

Sana Abdelkader

c. The void ratio is computed based on the following equation:

𝑉𝑅 = #𝑊

#𝑊+#𝐵

where #𝑊 and #𝐵 are the number of white pixels and the number of black pixels,

respectively. This process is illustrated in Figure 14.

(a)

2.Material and methods

2.8.Experimental performance analysis

Sana Abdelkader

(b)

(c)

Figure 14: Illustration of the steps of the void ratio estimation: original SEM image (a),

segmented image (b), binarized image (c).

2.Material and methods

2.8.Experimental performance analysis

Sana Abdelkader

In addition to the volume of the void region, the geometric distribution of the pores consistently

affects the efficiency of the membrane. Therefore, we consider the entropy of the membrane

regions (i.e., white regions provided by the previous segmentation process) as an indicator of

the distribution of the pores in this region.

5. Liquid entry pressure: Liquid entry pressure (LEP) is a significant membrane

characteristic. The liquid feed must not penetrate the membrane pores in other terms

the applied pressure should not exceed the limit (LEP) above which the feed (i.e.

aqueous solution) penetrates the hydrophobic membrane. LEP depends on the

maximum pore size and the membrane hydrophobicity and it is directly related to the

feed concentration and the presence of organic solutes, which usually reduce the LEP

(Gostoli, 1989; Alkhudhiri et al., 2012; Abdelkader et al., 2018).

6. Time: The time needed to achieve a level of mass transfer is an important parameter

when applying the DCMD process in an industrial environment. In fact, it is inversely

proportional to the cost of the process. Therefore, reducing the execution time can be

considered as one of the objectives to be reached. The indicators for this criterion are

the distillation time (DCMD time) and the pretreatment time. The DCMD time

corresponds to the time spent to reach a given mass transfer performance. In our

context, we set the termination condition of the experiment to a concentrate weight of

2kg.

2.Material and methods

2.8.Experimental performance analysis

Sana Abdelkader

Table 5: Numerical indicators and criteria

Criterion

Indicator

Index

Rejection factors

TOC

𝐼1

Conductivity

𝐼2

Flux

Permeate flux

𝐼3

Energy consumption

Inlet cold temperature

𝐼4

Inlet hot temperature

𝐼5

Flow rate

𝐼6

Pretreatment energy

𝐼7

Void volume

Void ratio in SEM images

𝐼8

Entropy of SEM images

𝐼9

LEP

𝐼10

Time

Distillation time

𝐼11

Pretreatment time

𝐼12

3.Results and discussion

3.1.Biological and membrane processes performance for saline effluents treatment

Sana Abdelkader

3 RESULTS AND DISCUSSION

3.1 BIOLOGICAL AND MEMBRANE PROCESSES PERFORMANCE FOR

SALINE EFFLUENTS TREATMENT

3.1.1 Aerobic bioreactors performance

The microorganism’s growth in the aerobic reactors was monitored by measuring the MLSS

content. The MLSS profile for all the reactors under different salt mass percentage is shown in

Figure 15, where for each aerobic reactor, we have the almost the same initial sludge quantity

expressed in MLSS (8±0.4 g/L).

Figure 15 depicts also the effect of salt concentration on the aerobic microbial growth. As

shown in Figure 15, the increase of the salt concentration in the synthetic effluent has leaded

to a gradual decrease in the MLSS growth rate to obtain the lowest growth with 2% NaCl

bioreactor. In terms of MLSS increase, the growth rates observed after 8 days incubation in the

different bioreactors in comparison with the initial one equal to 8 g/L, were as follows: 101%,

62%, 54% and 53% for the NaCl concentration of 0%, 0.5%, 1% and 2%, respectively. The

observed MLSS drop could be attributed to the increasing osmotic stress endured by the

different microbial communities causing the microorganisms’ plasmolysis and/or their activity

inhibition which affect significantly the physical and biochemical properties of the activated

sludge (Kargi and Dincer, 1996.a; Dincer and Kargi, 2001; Reid et al., 2006; Luo et al.,

2015).

3.Results and discussion

3.1.Biological and membrane processes performance for saline effluents treatment

Sana Abdelkader

Figure 15: MLSS evolution after 8 days in the aerobic bioreactors at different NaCl

concentrations (%).

With respect to the aerobic treatment efficiency and performance, COD removals were

measured for each bioreactor after an HRT of 6 hours (Figure 16). It is noteworthy that the

addition of 0.5% of NaCl didn’t affect significantly the COD removal in comparison to the

control bioreactor (0% NaCl) where the rates reached 93%. This result could be explained by

the adaptation ability of microorganisms at low and moderate salt concentration (<1%) which

allows the aerobic process to relatively maintain its good efficiency through the adjustment of

the microorganisms’ metabolism (Kargi and Dincer, 1996.a; Dincer and Kargi, 2001; Wang

et al., 2005). In another side, the 40% decrease of microorganism’s growth (Figure 15) after

increasing NaCl concentration to 0.5% did not have any impact on the COD performances of

the reactor.

0 0.5 1 2

MLSS (g/L)

[NaCl] (%)

Initial MLSS (g/L) MLSS after 8 days (g/L)

3.Results and discussion

3.1.Biological and membrane processes performance for saline effluents treatment

Sana Abdelkader

Figure 16: COD removal in aerobic bioreactors at different NaCl concentrations (%)

As shown in Figure 16, the aerobic treatment of 1% and 2% NaCl effluents has led to lower

process efficiency in terms of COD removal (65% and 70%, respectively). This result confirms

the suspected effect of high salt content on the aerobic treatment effectiveness. The recorded

performance reduction could be related on one hand to the notable bacteria’s growth decrease

(Figure 15) and on the other hand to the reduction of the effluents biodegradability engendered

by its toxic effect on the bacterial community which causes the inhibition of its activity in

addition to its limited and slowed down growth since most of halophilic bacteria are moderate

or extremely sensitive to oxygen (Kargi and Dincer, 1996.a; Dincer and Kargi, 2001; Wang

et al., 2005).

3.1.2 Anaerobic bioreactors performance

In order to study the increasing salt concentration effect on the anaerobic process performance,

both the methane production and the COD removal after 13 days incubation time were

measured in the anaerobic bioreactors with NaCl concentrations up to 3%. The methane

production, expressed in ml/l, is shown in Figure 17. The illustrated results indicate that though

the methane production in all bioreactors is relatively low, there has been a visible reduction

of the anaerobic microorganisms’ activity in terms of methane production especially when the

NaCl concentration is 1%. This observation could be explained by the metabolism changes

accompanying the salinity increase. As it was previously reported, under high saline conditions

100

0 1 2 3

COD removal (%)

[NaCl] (%)

3.Results and discussion

3.1.Biological and membrane processes performance for saline effluents treatment

Sana Abdelkader

the anaerobic biomass tries to acclimate to extreme environmental conditions by producing the

compatible solutes and consequently there are lower substrates for methane synthesis (Kempf

et al., 1998; Vyrides and Stuckey, 2009.b).

Figure 17: Methane production in the anaerobic bioreactors (13 days incubation time)

Figure 17 shows also an increase in the methane production at 2% and 3% NaCl probably due

to the development of moderate halophiles at 2% and 3% NaCl concentrations. It is previously

reported that some of the halophiles maintain their osmotic balance in the presence of high salt

concentration using intracellular accumulation which is enabled only in this saline medium and

this makes the survival of this kind of bacteria not possible in low salt concentration medium

(Vijay et al., 2018). Moreover, among anaerobic bacteria we find halophilic methanogens

which tolerate up to 12% NaCl concentration and which could be responsible of the increase

in methane production in the presence of 2% and 3% NaCl concentrations (Lai et al., 2002).

With respect to the COD reduction, the anaerobic treatment of saline effluents has shown too

lower pollution reduction in terms of COD removal for all the bioreactors. Figure 18 illustrates

the COD reduction for the different salt concentrations after 13 days incubation.

0,002

0,004

0,006

0,008

0,01

0,012

0 0,5 1 2 3

[CH4] (ml/L)

[NaCl] (%)

3.Results and discussion

3.1.Biological and membrane processes performance for saline effluents treatment

Sana Abdelkader

Figure 18: COD removal in the different anaerobic bioreactors at different NaCl

concentrations (%)

As shown in Figure 18, when no NaCl is added to the medium the COD removal reached 57%.

When adding 0.5%, 1%, 2% and 3% of NaCl to the medium the COD removal had shown a

relatively small change to reach 46%, 40%, 55% and 62%, respectively. The increase in COD

removal in the 2% and 3% NaCl bioreactors could be attributed to a change in the biomass

composition through the development of halotolerant and moderately halophilic anaerobic

bacteria that decreases the supernatants’ COD (Lai et al., 2002; Vijay et al., 2018).

Moreover, the observed variations in the COD removal in comparison to the 0% NaCl

anaerobic bioreactor is lower than those observed in the aerobic treatment which could be

explained by a possible lower sensitivity that anaerobic biomass has towards saline

environments (Vyrides and Stuckey, 2009.c; Ozalp et al., 2003).

Figure 19 illustrates the microbiological observations of the aerobic and anaerobic sludge

samples after Gram staining at different NaCl concentrations mass percentage.

As it could be shown in Figure 19, in anaerobiose conditions, fermentative halophilic

bacteria could be a Gram (-) or Gram (+) bacteria with a slight difference under 1% and 2 %

of NaCl. The increase of the NaCl concentration leaded to the selection of Gram-negative

bacteria in the anaerobic sludge especially at 3% NaCl where almost all the bacteria didn’t

retain crystal violet dye (Gam-negative bacteria). The halophilic fermentative bacteria belong

mostly to proteobacteria or actinobacteria (Kivistö and Karp, 2011). This observation

100

0 1 2 3 4

COD removal (%)

[NaCl] (%)

3.Results and discussion

3.1.Biological and membrane processes performance for saline effluents treatment

Sana Abdelkader

confirms the results shown by Figure 17 and Figure 18, where the increase of the methane

production and COD removals at 2% and 3% NaCl suggested the development of moderately

halophilic Gram-negative anaerobic strains. In addition, it has been previously proved in the

literature that the majority of moderate halophiles are Gram-negative rods which confirms

the obtained results (Quesada et al., 1983; Ventosa, 1988; Prado et al., 1991; Kivistö and

Karp, 2011).

However, under aerobic conditions, it is notable that most of bacteria are a Gram (-) bacteria

under 1 and 2% NaCl. With respect to the aerobic sludge, The Gram staining in Figure 19

shows that the NaCl increase from 1% to 2% has led to the reduction of the bacterial

population density that could be induced by the increased osmotic stress of the aerobic

bacteria at high NaCl concentrations which could have inhibitory or toxic effects and lead to

cell plasmolysis (Reid et al., 2006; Luo et al., 2015). Besides, most of halophilic bacteria

are a moderate or extremely sensitive to oxygen.

The observed population density decrease confirms the observed decrease in the sludge

MLSS and the COD removal at 2% and 3% NaCl concentrations (Figure 15 and Figure 16).

3.Results and discussion

3.1.Biological and membrane processes performance for saline effluents treatment

Sana Abdelkader

Figure 19:Microbiological observations of sludge samples after Gram staining: anaerobic

sludge at 1% NaCl (a), aerobic sludge at 1% NaCl (b), anaerobic sludge at 2% NaCl (c),

aerobic sludge at 2% NaCl (d) and anaerobic sludge at 3% NaCl (e)

(b)

(c)

(d)

(e)

(a)

3.Results and discussion

3.1.Biological and membrane processes performance for saline effluents treatment

Sana Abdelkader

3.1.3 Membrane microfiltration and ultrafiltration

3.1.3.1 Biological treatments coupled with Microfiltration

Following the synthetic saline wastewater biological treatments, further filtration tests were

performed as an additional step to improve the effluents’ quality. First, microfiltration

treatment was applied on the obtained aerobic bioreactors supernatants and the permeability

was calculated and plotted as a function of the filtration time (Figure 20).

Figure 20: Microfiltration membrane permeability of the different aerobic bioreactors’

supernatants (control is reverse osmosis water)

As shown in Figure 20, the membrane permeability has maintained comparable values for the

supernatants of the 0% and 0.5% NaCl aerobic bioreactors. However, a notable reduction in

membrane permeability values in the case of the 1% and 2% NaCl supernatants was observed.

This is accompanied with a relatively stationary phase over the operation time.

The aforementioned permeability drop reached 50% and 64%, respectively for 1% and 2%

NaCl, in comparison to the 0% NaCl microfiltration which could be related to the cells’

metabolism changes occurring under high salinity conditions. The obtained results are

confirmed by the observations of She et al. (2017) which found that the increase of NaCl to

1% and 2% led to a significant increase of the total extracellular polymeric substances (EPS)

contents in the sludge, especially humic substances. The production of the EPS becomes

100

200

300

400

500

600

700

800

900

10 20 30 40 50 60 70 80

Permeability (LMH/bar)

Time (min)

Control

0% NaCl

0,5% NaCl

1% NaCl

2% NaCl

3.Results and discussion

3.1.Biological and membrane processes performance for saline effluents treatment

Sana Abdelkader

remarkably higher due to the cells’ plasmolysis and the secretion of polymers and intracellular

constituents which effects significantly the filtration process and induces more severe

membrane fouling (Reid et al., 2006; Vyrides and Stuckey, 2011; Jang et al., 2013).

Moreover, the sludge properties and its settlement ability are significantly affected by the

treated effluents salinity since the produced EPS can influence the hydrophobicity of the

biomass aggregates and the flocs formation in the bioreactors (Guo et al., 2012).

Figure 21: COD removal after microfiltration of the aerobic bioreactors’ supernatants and

permeate.

Following the microfiltration of the aerobic bioreactors’ supernatants, COD removal in

permeates was measured as illustrated in Figure 21. The improvement of the COD removal

was insignificant after filtration of the 0% and 0.5% NaCl supernatants which has led to

removal of 95% and 96% COD, respectively. Meanwhile, after the microfiltration of the 1%

and 2% NaCl supernatants, the COD removal showed higher difference with a decrease to

reach 76% and 81%, respectively.

Regarding the anaerobic treatment, the microfiltration of the bioreactors supernatants has

shown lower permeability values for all treated effluents in comparison to the aerobic

bioreactors’ supernatants microfiltration (Figure 22).

When comparing the membrane permeability levels depicted by Figure 22, we can see that the

best permeability was obtained when treating the 0% NaCl supernatant followed by the 0.5%,

100

0 0,5 1 1,5 2 2,5

COD removal (%)

[NaCl] (%)

Supernatant

permeate

3.Results and discussion

3.1.Biological and membrane processes performance for saline effluents treatment

Sana Abdelkader

1%, 2% and 3% NaCl as the lowest obtained permeability level, the permeability reduction in

comparison to the 0% NaCl was as follows: about 28%, 54%, 62% and 77%, respectively. The

observed results could be explained by the fouling layer formation due to the released anaerobic

biomass’ EPS which increase gradually with the increasing effluent salinity and builds up the

membrane resistance during the filtration process due to biofilm formation over operation time

(Vyrides and Stuckey, 2011). Moreover, the salinity increase contributes in the change of the

EPS characteristics including their compositions and contents (Zhao and al., 2016).

Figure 22: Microfiltration membrane permeability of the different anaerobic bioreactors’

supernatants (control is reverse osmosis water)

Figure 23 illustrates the COD removal measured after both anaerobic treatment and

microfiltration process at different NaCl concentrations. The microfiltration process has

enhanced the effluents quality and had led to final COD removal of 77%, 59%, 51%, 72% and

77% which correspond to the 0%, 0.5%, 1%, 2%, and 3%NaCl concentrations, respectively.

100

200

300

400

500

600

700

800

900

1000

10 30 50 70

Permeability (LMH/bar)

Time (min)

Control

0% NaCl

0,5% NaCl

1% NaCl

2% NaCl

3% NaCl

3.Results and discussion

3.1.Biological and membrane processes performance for saline effluents treatment

Sana Abdelkader

Figure 23: COD removal after microfiltration of the anaerobic bioreactors’ supernatants

The turbidity removals measured following microfiltration of the aerobic and anaerobic

supernatants has shown relatively good rates with slight variation between the different salt

concentrations (Table 6). It is important to mention that the supernatant turbidity is affected by

the increase of the NaCl concentration since it could induce more production of extracellular

polymeric substances and the mean floc size is relatively bigger which influence the sludge

settlement ability (Vyrides and Stuckey, 2009.b). Nevertheless, the turbidity removals by

microfiltration did not drop significantly during the operation time when increasing the salt

concentration (Table 6).

Table 6: Turbidity removals following microfiltration of aerobic and anaerobic supernatants

at different salinities

Aerobic supernatant

Anaerobic supernatant

0% NaCl

78%

68%

0.5% NaCl

96%

74%

1% NaCl

79%

88%

2% NaCl

86%

79%

3% NaCl

72%

100

0 1 2 3 4

COD removal (%)

[NaCl] (%)

Supernatant

permeate

3.Results and discussion

3.1.Biological and membrane processes performance for saline effluents treatment

Sana Abdelkader

Even though the microfiltration process has improved the final COD removal in the anaerobic

process, the observed increase didn’t exceed the results obtained after the

aerobic/microfiltration treatment. Consequently, we can conclude that, in our study, when the

aerobic digestion is coupled with microfiltration process higher treatment performance and

effluent quality could be reached for all the studied NaCl concentrations.

3.1.3.2 Biological treatments coupled with Ultrafiltration

Ultrafiltration tests were performed to treat the biological aerobic bioreactors’ supernatants at

different NaCl concentrations. The measured permeability has shown similar behavior with a

relatively stable evolution over the studied filtration time and permeability reduction rates that

are lower than the ones obtained during the microfiltration process when compared to the 0%

NaCl supernatant (Figure 24).

The membrane permeability levels reduced by around 25% with 0.5% NaCl supernatant and

by 25% and 19% with 1% and 2% NaCl supernatants, respectively. As shown previously, the

increasing NaCl concentration could lead to more membrane fouling due to the released

biomass’ EPS which becomes remarkably higher due to cell plasmolysis at high NaCl

concentration in addition to the secretion of polymers and intracellular constituents which

participate in the biofilm formation over filtration time (Reid et al., 2006; Vyrides and

Stuckey, 2011; Jang et al., 2013). This influences directly the permeability which decreases

gradually with the increasing salt content. In the case of ultrafiltration tests, the fouling layer

formation could have higher impact on the treatment performance due to the smaller pore size

(0.01 µm) in comparison to the microfiltration’s membrane (1.6 µm).

3.Results and discussion

3.1.Biological and membrane processes performance for saline effluents treatment

Sana Abdelkader

Figure 24: Membrane permeability during ultrafiltration of the aerobic bioreactors’

supernatants (control is reverse osmosis water)

In terms of COD removal, the ultrafiltration treatment has improved the effluent quality

especially for the 2% and 3% NaCl supernatants where the improvement is of 11% and 18%,

respectively (Figure 25).

The COD removal have reached the following percentages: 97%, 96%, 74% and 85% in the

0%, 0.5%, 1% and 2% NaCl ultrafiltration permeates, respectively. The obtained results are

comparable to the ones illustrated in Figure 23 after microfiltration process, with a slight

increase in the COD reduction after ultrafiltration treatment.

10 20 30 40 50

Permeability (LMH/bar)

Time (min)

Control

0% NaCl

0,5% NaCl

1% NaCl

2% NaCl

3.Results and discussion

3.1.Biological and membrane processes performance for saline effluents treatment

Sana Abdelkader

Figure 25: COD removal after ultrafiltration of the aerobic bioreactors’ supernatants

Figure 26: Membrane permeability during ultrafiltration of the anaerobic bioreactors’

supernatants (control is reverse osmosis water)

100

0 1 2

COD removal (%)

[NaCl] (%)

Supernatant

Permeate

10 20 30 40 50

Permeability (LMH/bar)

Time (min)

Control

0% NaCl

0,5% NaCl

1% NaCl

2% NaCl

3% NaCl

3.Results and discussion

3.1.Biological and membrane processes performance for saline effluents treatment

Sana Abdelkader

The ultrafiltration of anaerobic bioreactors’ supernatants led to similar membrane permeability

behavior and lower permeability levels which ranged between 10 and 20 LMH/bar for all salt

concentrations (Figure 26). The illustrated membrane permeability levels didn’t show a

marked variation between the different treated NaCl concentrations which is similar to the

results obtained by the ultrafiltration of the aerobic bioreactors’ supernatants.

Table 7: Turbidity removals following ultrafiltration of aerobic and anaerobic supernatants

at different salinities

Aerobic supernatant

Anaerobic supernatant

0% NaCl

84%

0.5% NaCl

85%

82%

1% NaCl

83%

84%

2% NaCl

81%

86%

3% NaCl

87%

Table 7 shows the turbidity removals measured following ultrafiltration of the aerobic and

anaerobic supernatants. The ultrafiltration at the different NaCl concentrations has led to good

removal rates that are comparable to the ones obtained by microfiltration in the case of the

aerobic reactors supernatants (Table 6). Concerning the anaerobic reactors supernatants, the

turbidity removals obtained by ultrafiltration are slightly higher and this could be explained by

the tight pore size of the utilized membrane in addition to possible differences between the

supernatants composition (EPS, flocs settlement, etc...) that contribute in the fouling layer

formation.

3.Results and discussion

3.1.Biological and membrane processes performance for saline effluents treatment

Sana Abdelkader

Figure 27: COD removal after ultrafiltration of the anaerobic bioreactors’ supernatants

The COD removal obtained after the anaerobic supernatants’ ultrafiltration are shown in

Figure 27 which depicts the increase of the 0%, 0.5%, 1% and 2% NaCl permeate qualities

which reached 68%, 73%, 60% and 65% of COD reduction, respectively. For the 3% NaCl

ultrafiltration permeate, the COD removal didn’t show significant difference in comparison to

the anaerobic digestion supernatant.

3.1.4 Conclusions

In this first section, different treatment processes and wastewater qualities were analyzed and

investigated to show the effect of the NaCl mass percentage and the treatment characteristics

on the final water quality and the processes performances.

According to this investigation, the aerobic treatment was more efficient when combined to the

ceramic membrane microfiltration since we have obtained higher water quality with good

overall performance: the microfiltration led to COD removal of 96%, 76% and 81% for

aerobically treated feed of 0.5%, 1% and 2% NaCl, respectively. Concerning the ultrafiltration,

the COD removal reached 96 %, 74%, 85% for saline feed concentrations of 0.5%, 1% and 2%

NaCl, respectively. In our study, the anaerobic biological treatment didn’t lead to sufficiently

good COD removal when compared to the aerobic process which didn’t improve significantly

after the filtration processes.

100

0 1 2 3

COD removal (%)

[NaCl] (%)

Supernatant

Permeate

3.Results and discussion

3.1.Biological and membrane processes performance for saline effluents treatment

Sana Abdelkader

Regarding the studied membranes, the ceramic microfiltration membrane was more cost-

effective since it was regenerated during the study and required only a peristaltic pump to run

the different tests. However, the polymeric membrane required higher pressure and its study

could be developed into investigating the membrane regeneration and use for longer periods.

In all the studied processes, the conductivity and the salt concentration were not affected by the

filtration procedures and the main COD was removed mainly by the biological treatments.

Besides, the final water qualities were very similar in the case of low salt concentrations (0%

NaCl and 0.5% NaCl) in comparison to the higher concentrations (1%, 2% and 3% NaCl).

In order to solve the treated effluents salinity issue, good performance desalination process has

to be considered for this purpose. Direct Contact Membrane Distillation (DCMD) as emerging

process and one of the promising desalination technologies will be investigated and discussed

in the next sections of this research work.

3.Results and discussion

3.2.DCMD optimization for saline synthetic effluent treatment

Sana Abdelkader

3.2 DCMD OPTIMIZATION FOR SALINE SYNTHETIC EFFLUENT

TREATMENT

3.2.1 Permeate flux stability

The stability of the DCMD process during operation time for synthetic saline wastewater

treatment is checked through preliminary tests to evaluate the permeate flux evolution and the

obtained permeate quality (Figure28). The volumetric permeate flux maintained an almost

constant level with an average value of 9.5 L/m2.h during 5 hours treatment under the following

operating conditions: ΔT= 30°C, Vf = 0.051 m/s, [NaCl]= 10g/L and [Gluc]= 5 g/L.

Figure 28: Permeate flux evolution during operation time.

Concerning the permeate quality, the electrical conductivity reduction from the feed side to the

permeate side is 99.9% and in terms of COD the reduction was higher than 99.9 %. These

reduction rates confirm the ability of the DCMD process with the PTFE membrane (pore size=

0.45 μm) to treat and desalinate the synthetic saline wastewater with a stable performance over

time (5 hours treatment).

Following this observation, an operation time of one hour has been selected for all optimization

experiments via response surface methodology. Following each experiment performed in this

study, both COD and permeate conductivity were measured and the removal rates have been

0 1 2 3 4 5 6

Permeate flux (L/m2.h)

Time (h)

3.Results and discussion

3.2.DCMD optimization for saline synthetic effluent treatment

Sana Abdelkader

all higher than 99.9% with very low variations. Therefore, only the permeate flux was

considered as the main response in our work.

3.2.2 Response Surface Methodology (RSM) application in DCMD

The results of CCD experiments in terms of experimental and predicted permeate flux values

are shown in Table 8.

As it could be observed in this table, the predicted response values are highly comparable to

the experimental set of data which indicates that the developed regression model provides a

good fit to the experimental results.

The predicted flux is calculated by the means of the following regression equation that presents

the empirical model written in terms of actual variables. As illustrated in Table 8, the predicted

permeate flux values have a large range of variation following the changes in the studied

parameters from 3 to 23 L/m2.h which indicates the important effect of the chosen parameters

on the DCMD process. In the following, the regression model written with the coded symbols:

𝑱𝒑=11.38+5.706𝑿𝟏+2.21𝑿𝟐−0.549𝑿𝟑+1.209𝑿𝟒+ 0.313𝑿𝟏

𝟐−0.321𝑿𝟐

𝟐−

0.405𝑿𝟑

𝟐−0.127𝑿𝟒

𝟐+0.646𝑿𝟏𝑿𝟐−0.417𝑿𝟏𝑿𝟑−0.098𝑿𝟏𝑿𝟒−0.074𝑿𝟐𝑿𝟑+

1.292𝑿𝟐𝑿𝟒−0.503𝑿𝟑𝑿𝟒

The regression model could be also written in terms of actual variables as follows:

𝑱𝒑=11.38+5.706𝜟𝑻+2.21𝑽𝒇−0.549[𝑵𝒂𝑪𝒍]+1.209[𝑮𝒍𝒖𝒄]+ 0.313𝜟𝑻2−

0.321𝐕𝐟𝟐−0.405[𝑵𝒂𝑪𝒍]𝟐−0.127[𝐆𝐥𝐮𝐜]𝟐+0.646𝜟𝑻𝑽𝒇−0.417𝜟𝑻[𝑵𝒂𝑪𝒍]−

0.098𝜟𝑻[𝑮𝒍𝒖𝒄]−0.074𝑽𝒇[𝑵𝒂𝑪𝒍]+1.292𝑽𝒇[𝑮𝒍𝒖𝒄]−0.503[𝑵𝒂𝑪𝒍][𝑮𝒍𝒖𝒄]

When studying the coefficients given in the previous equation, we can conclude that the

temperature difference ΔT has the most significant influence on the mass transfer in DCMD

followed by the feed velocity with coefficient values of b1=5.706 and b2=2.21, respectively.

Both studied variables did affect positively the permeate flux (Jp) since when increasing one

or both of them we did get an increase in the permeate flux.

3.Results and discussion

3.2.DCMD optimization for saline synthetic effluent treatment

Sana Abdelkader

Previously, it was proven the significant effect of both temperature difference and feed flow

rate, and consequently the feed velocity, as part of the major parameters on which DCMD

process performance is depending. Following the temperature difference and feed flow rate

variations, significant high or low permeate fluxes could be reached and this could be attributed

to the nature of the process driving force (Hou et al., 2010; He et al., 2011; Singh et al., 2013

Ashoor et al., 2016;).

The summarized main effects of the investigated operating conditions, which are illustrated in

Figure 29, confirm the previously revealed results which indicate that the water vapor, and

consequently the permeate flux, are an exponential function of the temperature.

Moreover, Figure 29 shows that the temperature difference (ΔT) has the greatest effect on

DCMD treatment of the saline synthetic wastewater in comparison to the rest of the considered

and studied variables, followed by the feed velocity (Vf) as proven previously. Meanwhile, the

NaCl and glucose concentrations are showing lower influence on the permeate flux in the

respective ranges of variation.

Figure 29: Main effects plot for the permeate flux (Jp).

However, the NaCl concentration has a low negative influence (b3=-0.549) on the response

value (Jp). Regarding its influence on the DCMD performance, the NaCl concentration

increase affects slightly the water activity in the feed side and consequently, the driving force

3.Results and discussion

3.2.DCMD optimization for saline synthetic effluent treatment

Sana Abdelkader

which leads to a small decrease in the permeate flux (Martinez & Rodriguez-Maroto, 2007;

Boubakri et al., 2014).

The glucose concentration has shown low influence on the studied response with a coefficient

value of b4=1.209. In previous studies on DCMD with sucrose aqueous solutions, the increase

of the feed concentration has led to a small drop of the permeate flux that is linked to change

in the feed water activity. However, this influence was lower than the one obtained with salt

feed solutions at the same concentration.

It is important to indicate the close link between the feed viscosity and its water activity since

when adding sucrose as a solute to the feed stream viscosity increases. More precisely when

the feed solution temperature is relatively low: ≤35°C and at high glucose concentrations (from

100 g/L to 500 g/L) (Schofield et al., 1990; Izquierdo-Gil et al., 1999; Martinez &

Rodriguez-Marot, 2007).

However, in our case of study the glucose concentration didn’t exceed 20 g/L as a maximum

value which is much lower than the concentrations used in the aforementioned investigations

realized on sucrose aqueous solutions that started with 100 g/L as the lowest sucrose

concentration. This difference in glucose concentration could be the reason behind the different

observed impact on the permeate flux. According to our model, at high temperature differences

the positive effect of glucose concentration gets lower.

3.Results and discussion

3.2.DCMD optimization for saline synthetic effluent treatment

Sana Abdelkader

Table 8: CCD with predicted and experimental DCMD results.

Experiment

number

Predicted flux

(L/m2.h)

Experimental

flux (L/m2.h)

-1

3.1

3.1 ±0.4

-1

14.3

12.8±0.3

-1

3.8

4.4±0.7

-1

17.6

17.5±0.1

-1

4.0

2.8±0.2

-1

13.5

14.1±0.4

-1

4.4

4.1±0.1

-1

16.5

15.9±0.4

-1

4.2

5.6±0.8

-1

14.9

14.1±0.5

-1

6.6±0.9

-1

23.6

27.5±0.3

-1

3.0

1.9±0.2

-1

12.1

12.5±0.3

-1

8.6

13.0±0.5

20.3

19.1±0.4

-α

2.7

2.3±0.2

21.9

22.5±0.2

- α

6.8

8.1±0.3

14.2

13.1±0.2

-α

11.2

10.8±0.3

9.3

10.0±0.2

-α

9.0

10.5±0.1

13.1

11.9±0.4

11.4

11.3±0.1

11.4

11.4±0.1

11.4

11.3±0.1

11.4

11.3±0.1

3.Results and discussion

3.2.DCMD optimization for saline synthetic effluent treatment

Sana Abdelkader

3.2.3 Analysis of Variance

Variance analysis (ANOVA) was applied to investigate the statistical significance of the

created regression model as well the effects significance of the various studied parameters and

their interactions. The statistical significance is mainly reflected by both the p-value that has

usually a significance level of 0.05 (p-value <0.05 for significant influence) and the R2 which

indicates the proportion of variation in the regression model response with a value close to 1

for a model having good effectiveness and well predictive ability (Boubakri et al., 2014).

With respect to the different operating variables and their interactions, the results of ANOVA

showed that the temperature difference and the feed velocity had most significant positive

influence on the DCMD response with a low probability values for both of them (p-value=

0.000). Concerning the NaCl concentration, the variance analysis indicated that it has an

insignificant effect on the permeate flux response (p-value= 0.119). However, the glucose

concentration has a relatively significant influence on the process response with p-value=

0.003. According to the ANOVA results, the only significant variables interaction in this case

is the one between the feed velocity and the glucose concentration (p-value= 0.005). The

increase of the feed flow rate, and consequently its velocity, decreases the temperature

difference between the inlet and outlet of the DCMD module. This decrease leads to higher ΔT

which affects the feed stream viscosity and consequently causes the increase of the permeate

flux (Schofield et al. 1990; Gryta, 2012).

In our study, the developed model has a high R2 level of 0.967, which indicates that the

empirical model could explain more than 96.7% of the data deviation and therefore has a good

statistical significance. This observation is strengthened by the obtained R2Adj value: 0.932

which means that, significant terms have been included in the empirical model (Khayet et al.,

2007). In addition, the regression model has a very low p-value (0.000) which reflects its high

significance and its ability to provide a good prediction of the response permeate flux (Jp). All

the results obtained by the statistical estimators of ANOVA lead to confirm the effectiveness

and significance of the model.

By considering both the good effectiveness of the developed model and its high fit to the

experimental data by the comparison between the predicted and the experimental responses,

we can conclude that its statistical validation is well established. The regression model could

be applied for the description and optimization of the DCMD permeate flux response through

3.Results and discussion

3.2.DCMD optimization for saline synthetic effluent treatment

Sana Abdelkader

the optimization of the chosen variables (ΔT (°C), Vf (m/s), [NaCl] (g/L) and [Gluc] (g/L)) in

their aforementioned respective ranges of variation.

3.2.4 Response surface and contour-line plots

In order to study the interactions between the different studied variables, response surface and

contour-line plots are presented in Figure 30. These graphics and plots will lead to the

identification of the variables optimum levels which help to achieve the higher possible

permeate flux value. For each plot, two of the four variables are maintained constant and the

regression model is used to calculate the permeate flux and follow its evolution as a function

of the two other variables.

Figure 30 (a) and (b) show the simultaneous effect of ΔT and Vf on Jp at constant glucose and

NaCl concentrations (maintained at their center points). As both variables have the most

significant effects with a more pronounced influence of ΔT (b1=5.706 > b2=2.21), high

permeate fluxes (>25 L/m2.h) are obtained at their highest levels (ΔT ≥ 50°C and Vf ≥ 0.081

m/s). At low temperature difference the increase of the feed velocity does not improve

significantly the permeate flux which shows the importance of providing higher driving force

through increasing ΔT (Nakoa et al., 2014). It is important to indicate that there is no significant

interaction between the temperature difference and the feed velocity (p-value> 0.05).

The effect of the temperature difference and the glucose concentration is illustrated by Figure

30 (c) and (d) where the feed velocity and the NaCl concentration have constant values (0.057

m/s and 20 g/L, respectively). As expected, the temperature difference has the greater influence

on the flux response which achieves the highest values (> 15 L/m2.h) at ΔT > 40°C for all

glucose concentrations varying from 0 to 20 g/L. The pronounced effect of ΔT in comparison

to [Gluc] is confirmed by both their coefficients in the regression model (b1=5.706 > b4=1.209)

and their calculated p-values (0.000 and 0.003, respectively). In addition, the plotted curve

presents a parallel aspect that could lead to the assumption that there is no significant interaction

between the two studied variables.

At constant feed velocity (0.057 m/s) and glucose concentration (10.25 g/L), the permeate flux

in Figure 30 (e) and (f) increases significantly when increasing the temperature difference. A

slight decrease could be identified when increasing the NaCl concentration. The observed

results could be attributed to the induced reduction of the partial vapor pressure over the

membrane surface as it was proven previously (Schofield et al., 1990; Kamrani et al., 2014).

The plotted response surface suggests that the interaction between ΔT and [NaCl] is negligible.

3.Results and discussion

3.2.DCMD optimization for saline synthetic effluent treatment

Sana Abdelkader

Vf (m/s) 0,057

[Gluc] (g/L) 10,25

Hold Values

ΔT (°C)

[NaCl] (g/L)

5040302010

–

< 5

5 10

10 15

15 20

(L/m2.h)

Vf (m/s) 0,057

[NaCl] (g/L) 20

Hold Values

ΔT (°C)

[Gluc] (g/L)

5040302010

–

< 0

0 5

5 10

10 15

15 20

(L/m2.h)

[NaCl] (g/L) 20

[Gluc] (g/L) 10,25

Hold Values

ΔT (°C)

Vf (m/s)

5040302010

0,08

0,07

0,06

0,05

0,04

0,03

–

< 0

0 5

5 10

10 15

15 20

20 25

(L/m2.h)

[NaCl] (g/L) 20

[Gluc] (g/L) 10,25

Hold Values

020 04

20 0,

60,0

0,04

20 0,

80,0

)h.2m/L( pJ

)s/m( fV

)C°( TΔ

Vf (m/s) 0,057

[NaCl] (g/L) 20

Hold Values

002 04

)h.2m/L( pJ

)L/g( ]culG[

)C°( TΔ

Vf (m/s) 0,057

[Gluc] (g/L) 10,25

Hold Values

002 04

)h.2m/L( pJ

)L/g( ]lCaN[

)C°( TΔ

(e)

(c)

(b)

(d)

(f)

(a)

3.Results and discussion

3.2.DCMD optimization for saline synthetic effluent treatment

Sana Abdelkader

Figure 30: Response surface plots (a, c, e, g, i and k) and contour-line plots (b, d, f, h, g and

l) of the predicted DCMD permeate flux (Jp) as function of: temperature difference and feed

velocity [(a) and (b)], temperature difference and glucose concentration [(c) and (d)],

temperature difference and NaCl concentration [(e) and (f)], feed velocity and glucose

concentration [(g) and (h)], NaCl concentration and glucose concentration [(i) and (j)] and

feed velocity and NaCl concentration [(k) and (l)]

ΔT (°C) 30

[Gluc] (g/L) 10,25

Hold Values

0,04 60,0

0,0

20 0,04 60,0

0,08

)h.2m/L( pJ

)L/g( ]lCaN[

)s/m( fV

ΔT (°C) 30

[NaCl] (g/L) 20

Hold Values

400, 60,0

0,0

02 400, 60,0

,080

)h.2m/L( pJ

)L/g( ]culG[

)s/m( fV

ΔT (°C) 30

[NaCl] (g/L) 20

Hold Values

Vf (m/s)

[Gluc] (g/L)

0,080,070,060,050,040,03

–

< 5,0

5,0 7,5

7,5 10,0

10,0 12,5

12,5 15,0

15,0 17,5

17,5

Jp (L/m2.h)

ΔT (°C) 30

[Gluc] (g/L) 10,25

Hold Values

Vf (m/s)

[NaCl] (g/L)

0,080,070,060,050,040,03

–

< 6

6 8

8 10

10 12

12 14

(L/m2.h)

ΔT (°C) 30

Vf (m/s) 0,057

Hold Values

[NaCl] (g/L)

[Gluc] (g/L)

282624222018161412

–

< 8

8 9

9 10

10 11

11 12

12 13

13 14

(L/m2.h)

ΔT (°C) 30

Vf (m/s) 0,057

Hold Values

/m2.h)L( pJ

]culG[ / )Lg(

)L/g( Na[ Cl]

(h)

(j)

(l)

(g)

(i)

(k)

3.Results and discussion

3.2.DCMD optimization for saline synthetic effluent treatment

Sana Abdelkader

As illustrated in Figure 30 (g) and (h), the plotted responses show the variation of the DCMD

studied response Jp when varying both the feed velocity and glucose concentration with

constant ΔT and [NaCl] of 30°C and 20 g/L, respectively. At low feed velocities (Vf <0.046

m/s) the permeate flux doesn’t exceed 10 L/m2.h for all glucose concentrations. A simultaneous

increase of both variables leads to Jp augmentation that could exceed 17.5 L/m2h. This increase

might be induced by the decrease of the temperature polarization effect caused by the increase

in the heat transfer on the membrane surface that is led by the feed flow rate increase (Al-

Asheha et al., 2006). The interaction surface plot between Vf and glucose concentration could

indicate a significant interaction between these variables which is confirmed with the

probability p-value obtained by ANOVA (<0.05).

The combined variation effect of [NaCl] and [Gluc] on the permeate flux at constant ΔT (30°C)

and feed velocity (0.057 m/s) is shown in Figure 30 (i) and (j). At low levels of both studied

variables, the permeate flux has a minimum of 8 L/m2.h and it increases gradually with the

glucose concentration to reach 14 L/m2.h at low NaCl concentration and it ranges between 10

L/m2.h and 11 L/m2.h for the highest [NaCl]. We could assume from the response plot that

there is no significant interaction between the two variables.

Figure 30 (k) and (l) illustrate the influence of feed velocity and salt concentration on the

permeate flux variation when maintaining the rest of the variables constant (ΔT= 30°C and

[Gluc] = 10.25 g/L). The plotted response reveals the effect of Vf increase which leads to the

augmentation of the Jp that exceeds 14 L/m2.h when Vf is around 0.081 m/s and [NaCl] <

21g/L. At low feed velocities (<0.035 m/s) the calculated flux is under 8 L/m2.h for all salt

concentrations in the studied range. As it could be seen in the plotted responses, the interaction

between Vf and [NaCl] is negligible.

3.2.5 DCMD response optimization and model verification

As a result of the response surface methodology application on the studied saline synthetic

wastewater, a maximum response (Jp) should be predicted. In order to identify the optimum

operating conditions leading to achieve the aforementioned response, all the studied effects and

interactions are taken into consideration using Minitab software (Table 9).

A maximum permeate flux of 34.14 L/m2.h is predicted to be achieved at the following

optimum operating conditions: temperature difference ΔT= 55.23°C, feed velocity Vf= 0.086

m/s, NaCl concentration [NaCl]=10.08 g/L and glucose concentration [Gluc]= 20.01 g/L.

3.Results and discussion

3.2.DCMD optimization for saline synthetic effluent treatment

Sana Abdelkader

The composite desirability is calculated to assess how good the combination of the defined

optimum variables does meet the desired goal (maximizing the permeate flux Jp). In this study

we have a good desirability value of 0.975 (1 is the maximum).

Table 9: Optimum operating conditions and their correspondent predicted and experimental

responses.

In order to verify the optimization procedure and results, experiments were conducted at the

obtained optimum conditions and the responses were compared to the predicted maximum

response.

Table 9 illustrates both the experimental permeate flux mean value after three repeated

experiments at optimum operating conditions: 35.69± 0.6 L/m2.h and the maximum permeate

flux predicted by the developed regression model at the same conditions: 34.14 L/m2.h. The

deviation of 4.3 % indicates and confirms the validity of the optimization procedure and the

developed DCMD optimization model in addition to high performance in terms of COD

removal and conductivity rejection.

Variable

ΔT

(°C)

(m/s)

[NaCl

(g/L)

[Gluc]

(g/L)

(L/m2.h)

predicted

(L/m2.h)

experimental

Composite

Desirability

Value

55.23

0.086

10.08

20.01

34.14

35.69

0.975

3.Results and discussion

3.2.DCMD optimization for saline synthetic effluent treatment

Sana Abdelkader

3.2.6 Conclusions

In this second section, the response surface methodology was applied for the optimization of

synthetic saline wastewater DCMD treatment and desalination via the optimization of the

permeate flux (Jp) with a remarkable high permeate quality in terms of electrical conductivity

reduction (99.9%) and COD removal rate (>99.9%). In this study, four different variables were

considered to create the optimization model, namely: temperature difference ΔT, feed velocity

Vf, NaCl concentration [NaCl] and glucose concentration [Gluc]. The developed regression

model demonstrated a good level of effectiveness and ability to predict the process response

with R2= 0.967 and R2Adj = 0.932. The statistical evaluation and analysis of the model showed

that it has a high significance and that the most influencing variable in the studied ranges are

the temperature difference and the feed flow rate followed by the glucose concentration with

lower significant effect. The optimization procedure has led to the prediction of the maximum

permeate flux (34.14 L/m2.h) that could be achieved under the following optimum

experimental conditions: ΔT= 55.23°C, Vf = 0.086 m/s, [NaCl] = 10.08 g/L and [Gluc] = 20.01

g/L. As a final step, the experimental verification of the developed regression model showed

its validity and suitability within the studied range of variables.

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

3.3 DCMD APPLICATION FOR CHEESE WHEY WASTEWATER TREATMENT

3.3.1 Raw dairy effluent DCMD treatment

The first step in testing DCMD process on the raw saline dairy effluent (R-WW) was using M1

for the treatment. The feed inlet temperature and the permeate outlet temperature were

controlled at 55 and 25°C, respectively. Both feed and permeate velocities were maintained at

0.08 m/s.

Figure 31 shows the fast decline of the permeate flux during 7 hours operation that is

characterized by a drop from 11 kg/m2.h to around 4 kg/m2.h. Simultaneously, the permeate

electrical conductivity continuously increased and reached 1.95 mS/cm after only two hours of

treatment which corresponds to the pores breakthrough. This increase reached 5.28 mS/cm at

the end of the treatment with a TOC value that is higher than 100 mg/L. This observation shows

that M1 had lost its hydrophobicity by letting salts and other organic components crossing

through to the permeate stream (Kezia et al., 2015).

Figure 31: Permeate flux and electrical conductivity of raw effluent (R-WW), DCMD

treatment with M1, Tei=55°C, Tco=25°C and Ve=Vc=0.08 m/s

Consequently, the other membrane (M2), with lower pore size (0.2 μm) and higher thickness

and resistance was tested with raw saline effluent.

1000

2000

3000

4000

5000

6000

1 2 3 4 5 6 7

Electrical conductivity (µS/cm)

Flux (kg/m2.h)

Time (h)

Reference Flux Flux Conductivity

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

Figure 32: Permeate flux and electrical conductivity of raw effluent (R-WW), DCMD

treatment with M2, Tei=55°C, Tco=25°C and Ve=Vc=0.08 m/s

Figure 32 shows the evolution of the permeate flux and conductivity when operating with M2.

The permeate flux is showing a slower decrease in comparison to M1. Nevertheless, it is

decreasing during 10 hours to a value of 10.2 kg/m2h. The electrical conductivity remains stable

during 5 hours and starts increasing afterwards to 44 µS/cm after 10 hours. According to Ge et

al. (2014), this increase in the case of R-WW is related to the initiation of membrane wetting

since the feed stream contains a significant quantity of organic substances (Table 3) which

interact with the membrane surface and change its properties (De La Fuente and al., 2002). It

is important to note that the permeate TOC did not exceed 2 mg/L until the end of the treatment.

This shows that M2 has a higher heat transfer resistance than M1 at the same operating

conditions and feed quality in addition to its lower pore size. This can be explained by the fact

that it is characterized, in addition to the lower pore size, by a higher thickness, which indicates

longer diffusion distance and higher heat transfer resistance. Therefore, as it was proven

previously, with higher isolation against heat transfer through the membrane higher

temperature difference could be maintained leading to relative flux increase (Winter,

2015).The SEM observation of the membranes surfaces, M1 (7 hours) and M2 (10 hours), at

the same operating conditions show that the fouling layers cover the majority of the surfaces

in both cases with the presence of more cracks in M2 that could possibly be formed following

the membrane drying (Figure 33).

1 3 5 7 9 11

Electrical conductivity (µS/cm)

Flux (kg/m2.h)

Time (h)

Reference Flux Flux Conductivity

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

Figure 33: SEM pictures of membrane samples: fouling layers after raw effluent DCMD

treatment with M1 after 7 hours (a) and M2 after 10 hours (b). Operating conditions: ∆T= 30

°C and V=0.08m/s

3.3.2 Pretreated dairy effluents treatment

The feed inlet temperature and the permeate outlet temperature were controlled at 55 and 25°C,

respectively during the treatment of MAF and MAF-UF permeates as feed streams. Both feed

and permeate velocities were controlled to maintain a constant value of 0.08 m/s. Figure 34

shows the flux variation obtained by the treatment of: tap water as reference and two pretreated

dairy streams by MAF and MAF-UF (MAF-WW and MAF-UF-WW, respectively). The

permeate fluxes are illustrated starting from DCMD process stabilization after one hour. The

difference in the flux levels at the starting point could be linked to the quick formation of the

fouling layer during the first hour.

(a)

(b)

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

Figure 34: Permeate flux variation during MAF-WW and MAF-UF-WW DCMD treatment

with: (a) 0.45µm pore size membrane (M1) and (b) 0.2µm pore size membrane (M2).

Operating conditions: Tei=55°C, Tco=25°C and Ve=Vc=0.08 m/s

Figure 34 shows differences in the permeate flux during the operation time between the two

PTFE membranes M1 and M2 for the MAF pretreated effluent (MAF-WW) and the

additionally UF pretreated effluent (MAF-UF-WW). At the beginning, the permeate fluxes of

M1 of MAF-UF-WW were significantly higher than those of M2 which could be attributed to

the larger pore size of M1. For the experiments using MAF-WW as feed solutions with M1,

the flux dropped rapidly during the first 5 hours. This fast decline was accompanied by a fast

increase in permeate electrical conductivity, equivalent to a pore breakthrough (Figure 35).

1 3 5 7 9

Flux (kg/m2.h)

Time (h)

Reference MAF-WW MAF-UF-WW

(a)

1 3 5 7 9

Flux (kg/m2.h)

Time (h)

Reference MAF-WW MAF-UF-WW

(b)

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

Additionally, the permeate TOC exceeded 100 mg/L. These findings are similar to previous

results where it was found that there is higher decrease in the permeate flux when using a

hydrophobic membrane with a pore size of 0.45 µm than with 0.2 µm during DCMD process

of dairy effluents which is comparable to the results of this study (Kezia et al., 2015). However,

when processing the UF permeate with M1, the flux remains stable at 17 kg/m2.h but decreases

slightly after 7 hours to reach 15 kg/m2.h after 10 hours (Figure 34). This flux decrease will

induce the reduction of the process productivity. The permeate maintained good quality after

the same time with a conductivity of 60 µS/cm and a TOC below 2 mg/L (Figure 35).

Figure 35: Permeate electric conductivity during DCMD treatment with: (a) 0.45µm pore

size membrane (M1) and (b) 0.2µm pore size membrane (M2) of different saline dairy

streams. Operating conditions: Tei=55°C, Tco=25°C and Ve=Vc=0.08 m/s

0 2 4 6 8 10

Electrical Conductivity (µS/cm)

Time (h)

MAF-WW MAF-UF-WW

1000

2000

3000

4000

0 2 4 6 8 10

Electrical Conductivity (µS/cm)

Time (h)

MAF-WW MAF-UF-WW

(a)

(b)

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

With respect to the DCMD process using the second membrane (M2) with a pore size of 0.2µm,

Figure 34 shows that the permeate flux remains relatively stable during the first hours of

treatment for both effluents. This follows the typical flux variation in the membrane distillation

(MD) process (Laganà et al., 2000). The feed quality appears to have an influence on the flux

behavior even for M2, since for MAF-WW, the flux started to decline gradually after almost 5

hours of treatment. Additionally, the permeate electrical conductivity showed an increase after

the same duration from 20 to 27 µS/cm after 10 hours for MAF-WW (Figure 35). The observed

electrical conductivity increase corresponds to the start of the membrane wetting phenomena,

though the measured permeate TOC is lower than 2 mg/L (Ge et al., 2014).

The observed difference between the reference flux and the evolution of the permeate fluxes

obtained by MAF-WW and MAF-UF-WW confirm that the flux decrease over time is linked

to the formation of a fouling layer on the membrane surface.

The previous experiments showed that the DCMD treatment with M2 can lead to higher

permeate quality and more stable flux values and variations. The initial flux values are lower

in comparison to M1 but the sensitivity against fouling is smaller. Therefore, the following

investigations were performed with M2 and MAF/UF permeates as a feed solution.

3.3.3 Parameters affecting the permeate flux

Since the DCMD process could be affected by multiple operating parameters such as the feed

velocity and the temperature difference (Schofield et al., 1990). Therefore, the influence of

these two parameters and the dairy effluent quality on the formation of the fouling layer, the

permeate flux and quality was carried out. The thermal efficiency (ƞth) calculation is based on

the single Node simulation developed by Winter (2015). The thermal efficiency, which is the

ratio of the latent heat to the total heat transferring through the membrane is calculated using

the monitored mean temperature and permeate flux during the DCMD process. This parameter

serves as an indicator of the undesirable heat transfer in the system (heat transferred by

conduction).

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

3.3.3.1 Feed velocity

The thermal efficiency calculated after 5 hours treatment of MAF-WW and MAF-UF-WW at

∆T= 30°C and two different feed flow rates), showed that the treatment of MAF-UF-WW had

led to higher thermal efficiencies at both studied velocities (ƞth = 56% and 52%, at 0.08 m/s

and 0.04 m/s respectively). The permeate fluxes reached also higher values of 14 kg/m2.h and

10 kg/m2.h at 0.08 m/s and 0.04 m/s, respectively after 5 hours DCMD treatment of MAF-UF-

WW. In comparison, after 5 hours MAF-WW treatment, the thermal efficiency did not exceed

50% at 0.08 m/s and 0.04 m/s (ƞth = 47% and 35%, respectively) and the obtained permeate

fluxes were lower (12 kg/m2.h and 6 kg/m2.h, respectively) (Figure 36). In terms of permeate

flux, applying higher feed flow rates, and consequently higher feed velocity, increases the heat

transfer coefficient in the fluid boundary layer which declines the temperature polarization

effect and consequently increases the permeate flux (Gryta, 2012).

The difference between MAF-WW and MAF-UF-WW treatment in terms thermal efficiency

could be explained by the presence of higher organic matter concentration in MAF-WW. More

precisely, the higher proteins concentration in MAF-WW contributes to the formation of the

fouling layer since the beginning of the experiments (Table 3). The growth of the fouling layer

thickness over time is directly linked to the presence of more proteins and organic matter in the

MAF-WW feed stream (in comparison to MAF-UF-WW) that are known to deposit and adsorb

on hydrophobic materials and consequently, form a higher resistance in front of the mass and

heat transfer phenomena (Bottino et al., 2000; Hausmann et al., 2013.a).

Figure 36 shows also that the treatment of MAF-UF-WW had led to results that are close to

the reference tests in terms of thermal efficiency and permeate flux. This confirms that an

ultrafiltration pretreatment improves the feed quality in such a way that the DCMD efficiency

is enhanced significantly.

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

Figure 36: Thermal efficiency ƞth and flux J of MAF-WW and MAF-UF-WW DCMD

treatment in comparison to normal tap water test (Reference). Operating conditions: ∆T=

30°C and (a) V= 0.08 m/s and (b) V=0.04 m/s

100

J(kg/m2.h)

ƞth(%)

ƞth (%)

J (Kg/m2.h)

(a)

100

J(kg/m2.h)

ƞth(%)

(b)

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

A comparison of the results illustrated by Figure 36 shows that both the thermal efficiency and

the permeate flux diminished after 10 hours treatment of MAF-WW slightly (flux reduction of

3.5% at V=0.04 m/s and 9.3% at V=0.08 m/s). In contrast to the UF-pretreated effluent which

resulted in almost stable values of ƞth and the flux.

Figure 37: Flux behavior during operation time of DCMD treatment with MAF-UF-WW at

∆T=30°C and three velocities (0.08 m/s, 0.06 m/s and 0.04 m/s)

A study of the permeate flux during MAF-UF-WW treatment at a constant temperature

difference and multiple velocities (0.08 m/s, 0.06 m/s and 0.04 m/s) shows an interesting

difference between the flux profiles at the two higher velocities and the lowest velocity (V=0.04

m/s) (Figure 37). At 0.08 and 0.06 m/s feed velocities, the flux kept stable in the range of 13

kg/m2h. However, at a feed velocity of 0.04 m/s, the permeate flux did not exceed 10 kg/m2.h

and it started to decrease slightly after 10 hours to 9 kg/m2h.

Following the different tests, the liquid entry pressure analyses of the used membranes showed

also that at 0.08 m/s and 0.06 m/s the LEP had very close values (1.54 bar and 1.53 bar,

respectively) in comparison to 0.87 bar at 0.04 m/s which indicates a stronger wetting under

the latter operating conditions. In despite of the LEP drop at 0.04 m/s, the permeate

conductivity at the end of the experiment was 68.2 µS/cm and the TOC <2 mg/L. At 0.08 m/s

and 0.06 m/s, the electrical conductivity was 25 and 45.3 µS/cm, respectively. These results

confirm that a relatively low feed flow rate and consequently low feed velocity declines the

1 6 11 16

J(Kg/m2h)

Time (h)

V=0,08m/s

V=0,06m/s

V=0,04m/s

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

shear stress that influences the fouling layer deposition. The fouling layer alters both the

membrane and the permeate quality due to the wetting phenomenon.

3.3.3.2 Temperature difference

Researchers have found that when the permeate temperature is constant the flux increases

monotonically with the increase of the bulk temperature difference. However, when the feed

temperature is maintained constant and temperature difference is increased by decreasing the

permeate stream temperature, as it is the case in this study, the flux tends to follow an

asymptotic behavior at high ∆T (Lawson and Lloyd 1996; Laganà et al. 2000; El-Bourawi

et al. 2006).

DCMD experiments with MAF-WW and MAF-UF-WW as feed solutions were carried at 0.08

m/s and at two temperature differences, ∆T=30°C and ∆T=15°C (Figure 38). ∆T was changed

by keeping the feed stream temperature constant and varying the permeate stream temperature

which changes also the mean temperature. After 5 hours treatment, a significant decrease in

the permeate flux was observed for both experiments: MAF-WW (from 11.63 kg/m2.h to 7.2

kg/m2.h at ∆T=30°C and 15°C, respectively) and MAF-UF-WW (from 13.6 kg/m2.h to 7.96

kg/m2.h at ∆T=30°C and 15°C, respectively). This observation is in line with the general

finding that permeate flux rises with increasing temperature difference. This increase is due to

the growing driving force by the increase of vapor pressure difference that has an exponential

effect on the permeate flux (Schofield et al., 1987).

Figure 38 depicts also the influence of the operating time on the flux evolution. It is interesting

to note that for MAF-UF-WW, permeate flux reached almost stable values after 10 hours

treatment at ∆T=30°C and ∆T=15°C. This could be due to the lower organic content in MAF-

UF-WW which influences the fouling layer composition and deposition and contributes in the

stability of water vapor diffusion over time. In the case of MAF-WW, a decline in the flux after

10 hours was observed that could be attributed to the buildup of the fouling layer. This decrease

was more important at ∆T=30°C than at the lower temperature difference: from 11.63 kg/m2.h

to 10.55 kg/m2.h at ∆T=30°C and from 7.2 kg/m2.h to 6.84 kg/m2.h at ∆T=15°C, respectively.

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

Figure 38: Thermal efficiency ƞth and flux J of MAF-WW and MAF-UF-WW DCMD

treatment in comparison to normal tap water test (Reference). Operating conditions: V=0.08

m/s and (a) ∆T= 30°C and (b) ∆T= 15°C

The thermal efficiency (ƞth), decreased by increasing ∆T as shown in Figure 38. This could be

explained by a possible difference in the fouling layer thickness formed at the two studied

100

J(Kg/m2.h)

ƞth(%)

(b)

100

J(kg/m2.h)

ƞth(%)

ƞth (%)

J (Kg/m2.h)

(a)

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

temperature differences. The presence of the fouling layer on the membrane surface adds to its

thickness and forms an additional separation layer that increases the thermal resistance of the

system (membrane/fouling layer) and thereby, decreases the thermal efficiency. This

observation is supported by previous results which prove that the thermal efficiency gets higher

when the membrane wall is thinner (Gryta, 2012). It is important to note that at both studied

temperature differences the thermal efficiencies of MAF-UF-WW maintained more or less

stable and were close to the ones calculated from the reference tests with tap water at the same

operating conditions. Concerning MAF-WW, lower ƞth values were recorded at both

temperature differences in comparison to MAF-UF-WW and during operation time (after 10

hours) the thermal efficiency decreased by 6% at ∆T=30°C and by 3.8% at ∆T=15°C.

Figure 39: Flux behavior during operation time of DCMD treatment with MAF-UF-WW at

V=0.08 m/s and two temperature differences (∆T=30°C and 15°C)

With respect to the influence of temperature difference variation on the permeate flux during

MAF-UF-WW treatment, Figure 39 illustrates the flux over operating time at two temperature

differences. The flux is characterized by relatively stable behavior during operation time at

both temperature differences (∆T=30°C and 15°C).

The flux level difference observed in Figure 39 could be induced by both the variation of ∆T

and the way how this variation was conducted.

1 6 11 16

J(Kg/m2.h)

Time (h)

∆T= 30°C

ΔT= 15°C

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

3.3.3.3 Process evaluation with Single-Node simulation model

In order to evaluate the saline effluents DCMD treatment, calculations with the single-node

simulation model were carried to predict the process thermal efficiency and permeate flux in

the steady-state conditions (Figure 40). The predicted flux thermal efficiency at V= 0.04 m/s

(∆T=30°C) was 17.92 kg/m2.h and 65.8%, respectively, whereas at 0.08 m/s and at the same

temperature difference the predicted flux increased to reach 23.72 kg/m2.h with no change in

the thermal efficiency. As expected, both predicted parameters had higher values in comparison

to the ones obtained after MAF-WW and MAF-UF-WW permeates treatment at multiple

velocities (0.08 m/s and 0.04 m/s) (∆T=30°C). The lower thermal efficiency and permeate flux

values could be attributed to the formation of fouling layers in both cases, with higher

performance at the highest studied flow rate (0.08 m/s) which confirms our previous findings

(Gryta, 2012).

Figure 40: Permeate flux and thermal efficiency of treated effluents (MAF-WW and MAF-

UF-WW) compared to theoretical prediction at V= 0.04 m/s and 0.08 m/s (∆T=30°C).

J (Kg/m2.h)

ƞth (%)

ƞth (%) J (Kg/m2.h)

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

Figure 41: Permeate flux and thermal efficiency of treated effluents (MAF-WW and MAF-

UF-WW) compared to theoretical prediction at ∆T=15°C and 30°C (V= 0.08 m/s).

At the same flow rate (V= 0.08 m/s), the predicted permeate fluxes and thermal efficiencies at

∆T=15°C were as follows: 72.7% and 14.92 kg/m2.h while the predicted values at ∆T=30°C

were: 65.7% and 23.21 kg/m2.h, respectively (Figure 41). These calculated values are, as well,

higher than the ones calculated after MAF-WW and MAF-UF-WW treatments where

permeates fluxes were lower by 51.7% and 46.7% at ∆T=15°C, respectively. Besides, at

∆T=30°C, the difference between the predicted permeate flux and the ones obtained by S2 and

S3 was about 51% and 42.7%, respectively. The highlighted performance difference confirms

the aforementioned influence of the feed quality on the DCMD performance and the significant

affect that fouling has on this process.

3.3.4 Membrane fouling and wetting

3.3.4.1 SEM analysis

SEM pictures of the fouling layers of the tested PTFE membrane samples as well as the new

membrane sample (Reference) are shown in Figure 42. The SEM pictures of the membrane

samples showed that the fouling layer (observed after the same operation time and at the same

operating conditions) had different aspects.

It can be seen that the membrane surface is increasingly covered by the fouling layer with

increasing organic content of the treated feed solutions. The fouling layer formed after treating

J (Kg/m2.h)

ƞth (%)

J (Kg/m2.h)

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

R-WW is characterized by a relative uniformity and a more or less compact aspect that shows

less membrane cracks formed after the drying process (in comparison to the other samples).

This observation indicates that the fouling layer in R-WW membrane sample could be denser

due to the presence of higher organic content in the raw effluent. EDX analysis of the

membrane sample (R-WW) at 10 kV showed the levels of chloride (Cl), sodium (Na) and

potassium (K) which were around the following values: 13.06 %, 10.64 % and 1.58 %,

respectively (Table 10).

The results obtained from Figure 42 and Table 10 are supported by the literature. In their study

of PTFE membrane fouling during DCMD treatment of synthetic whey solutions, Hausmann

et al. (2013.a) showed that during 6 hours, the formation of the fouling layer is mainly

depending on the proteins’ deposition on the membrane surface. This has a direct influence on

the flux due the increased diffusion resistance. Since proteins tend to aggregate usually at a

high concentration (present on the membrane surface), the formation of the fouling layer

appears to involve proteins aggregation. This is on one hand via their interaction with the

membrane material (PTFE) and on the other hand via the presence of salts that promote with

their ionic strength the reduction of intermolecular repulsions to create the interaction between

the proteins causing their aggregation (De La Fuente et al., 2002).

(a)

(b)

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

Figure 42: SEM pictures of membrane samples: Reference (a) and fouling layers after 10h

DCMD process of different feed solutions: raw effluent R-WW (b), cross section with R-WW

(c), macrofiltration permeate MAF-WW (d) and macrofiltration-ultrafiltration permeate

MAF-UF-WW (e). Operating conditions: ∆T= 30°C and V= 0.8 m/s, during 10 hours

treatment

A close observation of the SEM pictures of both membrane samples MAF-WW and MAF-UF-

WW (Figure 42) shows that at the same operating conditions and time, the formed layers are

uneven and that their density is less when the feed is pretreated by macrofiltration (MAF-WW)

and even less after an ultrafiltration step (MAF-UF-WW). The reduced density is shown by the

increasing visibility of the membrane surface through the fouling layer which is confirmed by

the presence of fluorine (F), since F is one of the most important components of the PTFE

membrane material (Table 10). This observation could be related to the removal of any traces

of lipids by UF. Such lipids interact by adsorption with whey proteins to form a film that is

adsorbed on the hydrophobic membrane surface since the early stages of the DCMD process

(c)

(d)

(e)

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

and grows thicker and more stable with time and this result is consistent with previous

observations (Waninge et al., 2005).

The observed low fouling layer density in S3 confirms the higher flux values that we obtained

by DCMD and this shows that the fouling layer is playing an important role in the decrease of

mass transfer resistance through the membrane.

Table 10 shows that the ultrafiltration step that removed all traces of whey proteins and other

whey components larger than 100 kDa had led to a less dense fouling layer (Figure 42). This

layer is mainly characterized by the presence of the highest content of calcium (Ca) and

phosphorus (P) and the lowest of sodium (Na) and chloride (Cl) among the samples analyzed

by EDX. This result is consistent with a previous work (Hausmann and al., 2013.a) with

DCMD process of synthetic whey solution in which it was found that those ions were dominant

in the fouling layer. However, in our study this finding is not applicable in the case of R-WW

and MAF-WW, where the predominant ions present in the fouling layers were sodium and

chloride. This result could be linked back to the presence of micellar casein in both R-WW and

MAF-WW (which is only removed by a 0.2 µm pore size filtration). It was showed in previous

researches that the existence of micellar casein in solution with whey proteins and salts (NaCl)

could affect the proteins aggregation process through the enhancement of the attractive forces

between proteins in high ionic strength solutions (Kehoe and Foegeding, 2011; Mounsey and

O’Kennedy, 2009; Pierre et al., 1992).

Table 10: Chemical elements mass percentage by EDX analysis at 10Kv of multiple

membrane samples: without any treatment (Reference) and with 10 hours DCMD treatment

of R-WW, MAF-WW and MAF-UF-WW at the following operating conditions: ∆T= 30°C and

V= 0.08 m/s

Mass percentage (%)

Reference

10.64

20.54

R-WW 10kV

23.21

2.44

10.64

13.06

1.58

3.61

MAF-WW 10kv

42.43

3.31

3,38

3.07

6.20

2.47

MAF-UF-WW 10kv

10.51

4.36

1.31

13,33

1.09

12.76

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

According to Rice et al. (2009), calcium and phosphorus salts solubility is highly influenced

by the solution acidity (through an inversely proportional relationship) and it decreases at high

temperature values. Therefore, since the treated feed solutions had a neutral pH (pH=7), the

presence of both calcium and phosphorus salt in the fouling layer will contribute both to the

organic fouling and the mineral scaling. This could explain the strong presence of those salts

in the fouling layer treating the MAF-UF-WW. Nevertheless, the effluent acidity could affect

the mineral scaling. It has an insignificant influence on the final flux and on the membrane

wetting (Kezia et al., 2015).

3.3.4.2 Liquid entry pressure and membrane resistance

Liquid entry pressure tests have shown that the pre-treatment of the salty wastewater by MAF-

UF improved significantly the membrane quality in comparison to the raw and the MAF

pretreated effluent (Figure 43). However, since the LEP indicates directly the degree of

membrane wettability, it is important to note that the measured LEP of a new membrane is

more than the double of that achieved by MAF-UF pre-treatment (3.88 bar and 1.54 bar,

respectively).

The effluent organic load and the composition have a great influence on the pores’ wettability.

Since the raw saline whey effluent is mainly composed of whey proteins, minerals and sugars,

all these components contribute in the fouling layer formation which alters the membrane

resistance, especially with the presence of casein micelles and fat traces (Hausmann et al.,

2013.b).

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

Figure 43: Liquid entry pressure (LEP) of membrane samples after 10h DCMD process of

different feed solutions: R-WW, MAF-WW and MAF-UF-WW. Operating conditions: ∆T=

30°C and V= 0.08 m/s

Kezia et al. (2015), found that the presence of a protein concentration superior to 2 g/L in the

salty whey effluent could have an influence on the membrane hydrophobicity along with the

presence of minerals which enhances the pores wetting. Besides, they found that for 0.2 µm

PTFE membrane and for an effluent pretreated by MAF-UF, the LEP values decreased by

almost the same rate as those in this study.

The change of the operating conditions (feed velocity and temperature difference) for R-WW

and MAF-UF-WW had an influence on the membrane’s LEP. At a feed velocity of 0.04 m/s,

the LEP dropped to 0.24 bar and to 0.28 bar for R-WW and MAF-WW, respectively, which

indicates a lower membrane hydrophobicity in comparison to a velocity of 0.08 m/s. Similar

LEP reduction was observed for MAF-UF-WW (section 3.3.3.1). These results show that low

feed velocity induces changes in the fouling layer and reduces the membrane hydrophobicity.

However, when decreasing the temperature difference to 15°C an increase in the LEP values

was observed (0.55 bar and 0.58 bar for R-WW and MAF-WW, respectively). This could be

linked to the influence of the temperature on the interactions between proteins with the

hydrophobic membrane via the salts’ presence (De La Fuente et al., 2002).

0,44 0,45

1,54

3,88

R-WW MAF-WW MAF-UF-WW Reference

LEP (bar)

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

Figure 44: Electrical conductivity (µS/cm) at p=0.3 bar during DCMD process of different

feed solutions: R-WW, MAF-WW and MAF-UF-WW. Operating conditions: ∆T= 30°C and

V= 0.08 m/s

In order to investigate the influence of the feed pressure on the membrane resistance and

wetting, DCMD experiments at ∆p= (pevaporator-pcondensor)= 0.3 bar were conducted during 5

hours with the three different salty dairy effluents (Figure 44). By monitoring the electrical

conductivity during the experiments, it could be confirmed that the raised feed pressure which

was close to the LEP of R-WW and MAF-WW, the permeate conductivity increased until it

reached after 5 hours 853 µS/cm and 682 µS/cm, respectively. This result confirms that at a

pressure close to the LEP, the liquid breaks through the pores to the permeate side. As expected,

the rise in the feed pressure didn’t influence the permeate quality in the case of MAF-UF-WW

where the conductivity didn’t exceed 29 µS/cm.

3.3.4.3 Contact angle and membrane cleaning effect

Another parameter that reveals the membrane hydrophobicity in DCMD is the contact angle

which is inverse to the wettability. It is known that for membranes with hydrophobic character

the contact angle formed by a water drop on the membrane surface has to be higher than 90°

(Zisman, 1964). The results illustrated in Figure 45 confirm the previous findings and show

that by pretreating the raw salty dairy effluent first by MAF-UF, higher membrane

200

400

600

800

1000

0 1 2 3 4 5 6

Electrical Conductivity (µS/cm)

Time (h)

R-WW MAF-WW MAF-UF-WW

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

100

hydrophobicity can be achieved. As shown above, this influences directly the permeate quality

and its stability over operating time. It is important to mention that the contact angle decreases

at higher temperatures since the PTFE membrane evolves with temperature variation and the

mean pore size get larger with increasing the temperature (Saffarini et al., 2013; Ge et al.,

2014).

Figure 45: Contact angle of new and cleaned membrane samples with different feed

solutions: R-WW, MAF-WW, MAF-UF-WW at T= 22°C. Operating conditions: ∆T= 30°C

and V= 0.08 m/s

By cleaning the membrane with basic and acid solutions the membrane loses part of its

hydrophobicity measured by a lower contact angle in all the experiments (Figure 45). This

could be explained by the effect of acidic solution during the removal of the fouling layer which

induced the irreversible wetting of a fraction of the membrane pores (Gryta, 2008).

The lowest hydrophobicity loss was measured in MAF-UF-WW where the contact angle

decreased by 3.9% in comparison to R-WW and MAF-WW where the reduction was of 8.1%

and 9.5%, respectively.

100

120

140

Raw WW MAF WW MAF-UF

Reference

Contact Angle (°)

New

membrane

Cleaned

Membrane

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

101

Figure 46: SEM picture of a cleaned membrane sample (following the DCMD process of

MAF-UF-WW). Operating conditions: ∆T= 30°C and V= 0.08 m/s

SEM picture of the cleaned membrane sample shows that the membrane is still covered partly

by the fouling layer (Figure 46). In order to identify the general composition of the different

observed zones, an EDX analysis was performed on three different areas of the membrane at

20 kV.

Table 11: Chemical elements’ mass percentage obtained by EDX analysis at 20kV (the

marked areas in Figure 46). ND: Not Detectable

Mass percentage (%)

Area1 20kV

28,95

3,63

46,66

0,56

0,76

Area2 20kV

20,05

60,56

Area3 20kV

18,35

55,72

0,50

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

102

The main chemical elements found in the darkest shaded area 1 on the membrane surface were

nitrogen (N) 3.63%, sodium (Na) 0.56% and calcium (Ca) 0.76% which shows that some

remaining parts of the fouling layer containing proteins and salts are still deposited and adhered

to the membrane which could contribute in the recorded hydrophobicity decrease (Table 11).

Concerning the other analyzed areas only membrane material, mainly fluorine (F), was found

in addition to small mass percentage of calcium (0.5%).

3.3.5 Experimental performance and cost-effectiveness analysis

3.3.5.1 Analytical model

In this section, we study the cost-effectiveness of the DCMD process with respect to the two

pretreatment options, namely the MAF and the UF-MAF processes. Based on the results

reached in the foregoing sections, we analyze the performance of the DCMD process not only

from the perspective of water quality, but also based on the various investments needed to

conduct the process. To this end, we represent the DCMD process using a set of inputs and

outputs for which the cost and the benefit are assessed using the indicators introduced in Table

5. From the practical point of view, this allows an industry to design a cost-effective DCMD

process based on a convenient choice of the options available for the pretreatment process. The

cost and the benefit of the pretreatment process depend of the following indicators:

• The cost is impacted by the energy consumption, void volume, LEP and time

(indicators I4-I12). The energy consumption and time correspond to direct costs that are

relatively easy to evaluate. Nonetheless, the void volume and LEP affect the lifetime of the

membranes. Hence, the budget spent to acquire the membranes varies accordingly;

• The benefit varies according to the rejection factors and the permeate flux (indicators

I1-I3). These indicators are considered as the benefit variables because they affect the volume

of water collected posteriori, which is the main objective of the membrane filtration process.

Since the respective expressions of the cost and benefit functions with regard to the underlying

indictors are not straightforward, we use the utility theory (Cohon, 1978) to evaluate the two

pretreatment scenarios with respect to the options (experimental setup) considered in the

treatment phase. Utility theory has been applied in different fields for decision-making

purposes through the analyses of an individual's preference-indifference relation (Kaplinski,

2013).

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

103

In the sequel, we define the demand function x as a mapping between a cost value, denoted by

c, and the indicators defined in Table 5. In other terms, for a cost value c, the output of the

demand function is represented as follows:

𝑥(𝑐)={𝐼1

𝑐,𝐼2

𝑐,…,𝐼12

𝑐}

In order to apply utility theory to our context, we define the utility function, denoted by 𝑢(𝑥,𝑐),

as the projection of x on the subspace corresponding to the indicators that affect the benefit.

Hence, the expression of 𝑢(𝑥) is defined below.

𝑢(𝑥,𝑐)={𝐼1

𝑐,𝐼2

𝑐,𝐼3

𝑐}

We use this function in order to express the indifference between the alternatives to implement

the MAF and UF-MAF processes. It is noteworthy that, in our experiments, we only consider

the indicators {𝐼3,𝐼8}. This choice is motivated by the fact that these indicators strongly affect

the cost of the process and the unavailability of data regarding the evolution of other indicators

(e.g., I7 and I9).

We use the Slutsky equation to model the substitution relationship between the MAF and the

UF-MAF pretreatment processes. For a demand x, the Slutsky equation is expressed as:

(∆𝑥

∆𝑐)MAF = (∆𝑥

∆𝑐)UF−MAF − 𝑥 ×∆𝑥

∆𝑏𝑖

where bi stands for one of the benefit indicators (i.e., {𝐼1

𝑐,𝐼2

𝑐,𝐼3

𝑐}) , (∆𝑥

∆𝑐)UF−MAF denotes the

substitution effect (which must be negative as long as the indifference curve between MAF and

UF-MAF exhibits a diminishing marginal rate of substitution) and (− 𝑥1 ×∆𝑥

∆𝑏) denotes the

compensating effect.

The indifference corresponds to a situation in which the demand functions of the MAF and the

UF-MAF processes satisfy the following condition:

𝑐1𝑥MAF(𝑐1)+𝑐2𝑥UF−MAF(𝑐2)−𝑏MAF −𝑏UF−MAF =0

where 𝑏MAF =𝑢(𝑥MAF,𝑐1) and 𝑏UF−MAF =𝑢(𝑥UF−MAF,𝑐2)

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

104

Since the time dimension is important in the DCMD process, we express the utility function as

𝑢(𝑥,𝑐,𝑡), where x expresses the demand function, c expresses the corresponding cost, and t

expresses the time to conduct the process satisfying the demand x. Under these assumptions,

the indirect utility function, denoted by 𝑉(𝑐1,𝑐2,𝑡1,𝑡2) is given by:

𝑉(𝑐1,𝑐2,𝑡1,𝑡2)=max{𝑢(𝑥MAF,𝑐1,𝑡1),𝑢(𝑥UF−MAF,𝑐2,𝑡2)}

We also define the parameters z and l such as that:

𝑧+𝑐1𝑥MAF =𝑥UF−MAF 𝑎𝑛𝑑 𝑙+ 𝑐2𝑡1=𝑡2

The utility function is assumed to be differentiable and quasi-concave in

, and the constraints

are differentiable and linear in

and in both budget, cost and time (Cohon, 1978).

The Lagrangian function of V is represented by:

𝑉(𝑐1,𝑐2,𝑡1,𝑡2)max{𝑢(𝑥MAF,𝑐1,𝑡1),𝑢(𝑥UF−MAF,𝑐2,𝑡2)}+𝜆(𝑥UF−MAF −𝑧−𝑐1𝑥MAF)

+𝜇(𝑡2−𝑙−𝑡1𝑐2)

The Lagrange multipliers λ and μ represent the shadow values of cost/budget and time,

respectively. Intuitively, λ is the marginal utility of benefit (i.e., optimization of the indicators)

and μ is the marginal utility of time.

Through the application of the envelope theorem (Cohon, 1978), we define the Value of Time

(VoT) as the marginal substitution rate for utility between the cost parameter c and the duration

time t. The value of time can be expressed as:

𝑉𝑜𝑇 =−𝑑𝑐

𝑑𝑡 |𝑉=𝑐𝑜𝑛𝑠𝑡 =𝜇

𝜆.

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

105

Therefore, it follows from the previous equation that VoT as a resource can be expressed in

terms of observable demand x as:

𝑉𝑜𝑇 =𝑑𝑉 𝑑𝑡2

⁄

𝑑𝑉 𝑑𝑦

⁄=𝑑𝑥MAF 𝑑𝑡1

⁄+𝑥𝑑𝑥UF−MAF 𝑑𝑡2

⁄

𝑑𝑥MAF 𝑑𝑐

⁄+𝑥𝑑𝑥UF−MAF 𝑑𝑦

⁄.

This equation establishes a link between the marginal rate of substitution of the demand

functions (relative to the MAF and UF-MAF pretreatment processes) in terms of time and

cost/budget, respectively.

In the following section, we illustrate our analytical model through the analysis of a set of

experimental results. Through the indifference curves, we show how a relation can be found

between the cost/benefit/time triple for specific executions of the MAF and the UF-MAF

processes. In concrete industrial situations, this relation is useful as a decision-support

mechanism based on the comparison of the costs needed for each of the pretreatment

techniques to achieve the same time performance or, in general, the same benefit.

3.3.5.1 Cost-effectiveness analysis

As it has been stated in the previous subsection, the comparison criterion used in this simulation

is the DCMD operation time. In other terms, we generate (from the experimental data sets) the

indifference curves allowing the assessment of the amount of time that is gained/lost if one

pretreatment technology is utilized rather than the other, for the same output parameter. To this

end, we apply the mathematical model detailed previously to estimate the overhead in terms of

time for a fixed value of one of the two output parameters (permeate flux and membrane void

ratio). Our experiments encompass six scenarios (three scenarios for MAF and three scenarios

for UF-MAF) as illustrated in the following table (Table 12).

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

106

Table 12: Experimental scenarios.

Scenario

Pretreatment

Q (L/h)

∆T (°C)

MAF

100

MAF

100

MAF

UF-MAF

100

UF-MAF

100

UF-MAF

To illustrate the potential brought by the utility theory, and more precisely indifference curves,

to evaluate the two pretreatment scenarios in terms of cost-effectiveness, we first consider the

evolution of the flux with respect to time as a parameter to discuss an illustrative case study of

the approach. Figure 47 illustrates the evolution of the flux with respect to time for each of the

scenarios specified in Table 12.

It is worthy to underline that the termination criterion corresponds to a mass of 2 kg of the

concentrated feed tank. From this figure, it can be deduced that, as a DCMD pretreatment, the

UF-MAF technique outperforms the MAF technique in terms of amount of collected water per

unit of time and DCMD membrane surface.

A comparison of the performances of the two pretreatment techniques in the same experimental

conditions is performed by associating each scenario in which MAF is used to its homologous

in which UF- MAF is used.

To this purpose, and using Table 12, the scenarios are grouped into three couples (1,4), (2,5),

and (3,6). From Figure 47, it comes that the average improvement reached by UF- MAF with

regard to MAF in terms of permeate flux is 20%.

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

107

Figure 47: Flux evolution during time for the 6 studied scenarios.

Such statement provides a view on the performances of the MAF and UF-MAF pretreatment

techniques of the DCMD process from the experimental perspective but it obfuscates the issues

related to the overhead that would affect the use of these techniques in concrete industrial

environments. For instance, an important parameter that would be interesting to investigate

using the results depicted in Figure 47 is the time overhead, which is the additional amount of

time spent for one technique to reach the same performance using the other technique. This

parameter is important from the practical point of view since most of the components of the

Operational Expenditure (OpEx) are connected to time, such as energy and manpower.

To cope with this issue, we apply the utility theory formalism described above to the functions

representing the evolution of flux with respect to time for each of the six scenarios of Table

12. Three main steps are followed:

1. Draw the indifference curve in which the two axes represent the DCMD permeate flux

obtained following MAF and UF pretreatment techniques, respectively. Such curve is

used to model the hypothetic situation in which a user would have access to both

pretreatment techniques and expresses his preference (in terms of permeate flux for this

case) between them.

2. The equation of the indifference curve elaborated in Step 1 is used to find a mapping

between specific values of flux reached through the MAF technique and their

1 6 11 16 21

Flux (kg/m2h)

Time (h)

Scenario 1

Scenario 2

Scenario 3

Scenario 4

Scenario 5

Scenario 6

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

108

corresponding values reached through the UF-MAF technique and leading to the same

preference.

3. We identify, using the Splines interpolation technique applied to the values of Figure

47, the amount of time corresponding to the values of flux identified in Step 2.

Throughout this process, it is possible to model the evolution of time with respect to the flux

(the dual behavior of Figure 47) by considering some values of flux that were not reached

experimentally but reconstructed using the indifference curves used in Steps 1 and 2. In the

following, we illustrate the execution of this process in the case where the flux is used to draw

the indifference. Then, we extend this reasoning to the case where the void percentage is used

to this purpose. Obviously, the process is generic enough to encompass all the indicators

considered in Table 12. However, more experiments must be performed to address the rest of

the indicators.

Based on the values expressing the evolution of the flux with respect to time (Figure 47), we

obtain the indifference curve represented in Figure 48 where the two axes represent the DCMD

permeate flux expressed in (kg/m2.h) following the UF-MAF (X axis) and MAF (Y axis)

application. In other terms, this figure illustrates the relationship between the performances, in

terms of permeate flux, of the two pretreatment scenarios. It is obtained through the following

process:

a. The average permeate flux is computed with respect to time for each of the six scenarios

considered above.

b. The averages obtained in step (a) are mapped so as to plot 3 points (plotted in red in

Figure 48) so as to represent the equivalence (in terms of percentage of flux allocated

to each technique) of the pretreatment techniques in terms of flux. From the

experimental point of view, this mapping is performed by associating the couple of flux

values that result in the least variation of the sum of fluxes. This extends the analogy

between baskets (customarily used in utility theory) and our context.

c. The plots are interpolated for the values identifying the aforementioned scenarios using

the cubic Spline interpolation to draw a hyperbola which represents the indifference

curve. This curve is constituted by the set of points for which we suppose that the

performance of the DCMD process for which a user would be indifferent in terms of

permeate flux.

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

109

The indifference curve plotted in Figure 48 has the following hyperbolic equation:

−1.29 𝑥2+23.63 𝑥𝑦+4.64 𝑦2−102.05 𝑥− 247.28 𝑦= −1022.57

Figure 48: Indifference curve between the MAF and UF-MAF pretreatment technologies in

terms of DCMD permeate flux.

The plotted indifference curve allows the calculation of the global improvement, in terms of

permeate flux, which can be reached through the use of the UF-MAF technique as a DCMD

pretreatment. This is, by definition, the average of the different local improvements computed

through the mapping between unitary decrease in the MAF axis and the corresponding increase

in the UF axis in terms of permeate flux (Figure 49).

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

110

According to Figure 49, we can conclude that the average global improvement (taking both

positive and negative performances of UF-MAF with respect to MAF into consideration) is

1.5%. In terms of neat improvement (taking only the positive performances into consideration),

the permeate flux improvement ratio that we can obtain by UF-MAF pretreatment have a

minimum of 25% and a maximum of 71%.

Figure 49: Indifference curve between the MAF and UF-MAF pretreatment technologies in

terms of DCMD permeate flux and the improvement provided by UF-MAF.

Hence, we reach the practical use of this indifference curve which consists in mapping every

point in Figure 47 for scenarios 1, 2, and 3 to the corresponding point as if it was obtained

from scenarios 4, 5, and 6; respectively (this is to preserve the same experimental condition in

this mapping). This is achieved using the hyperbolic equation of Figure 48. Through this

mapping, we reconstruct the values of time corresponding to the flux values that were not

experimentally reached in scenarios 4, 5, and 6.

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

111

Using the results of Figure 47, we interpolate the evolution of time with respect to one of the

six scenarios of Table 12.

Figure 50 shows the result of such interpolation in the case where the values of Scenario 1 are

considered. For example, it underlined that it is possible to estimate the time corresponding to

a flux equal to 7 kg/m2.h.

Figure 50: Interpolation of time with respect to flux.

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

112

By combining the results of Figure 49 and Figure 50, it is possible to identify the time

overhead needed to achieve the same performance as UF-MAF, in terms of flux, using MAF

under the assumption of an indifferent user (this means that the MAF and UF-MAF techniques

are not compared in similar experimental conditions). It is found that this overhead is equal in

average to 43%. It reaches as a maximum 72 % and as minimum 28%.

This result is important from the cost-effectiveness perspective since it provides information

about the cost overhead corresponding to the same value of flux. It gives a complementary

view on the performances of the two pretreatment techniques (UF-MAF and MAF) in terms of

the additional cost that would be spent for a similar performance and regardless of the

requirement to have similar experimental conditions. Obviously, we considered the operation

time as an indicator of cost because most of the OPEX (Operational Expenditure) are often

function of time (e.g., energy, manpower).

The next evaluated indicator that impacts the DCMD treatment applicability is the membrane’s

void ratio that describes the available membrane surface on the utilized samples which could

reflect its lifetime. The calculated void ration is also closely linked to the membrane fouling

and consequently, to the process performance in terms of permeate flux.

Similarly, the indifference curve illustrating the relationship between the effects of the two

pretreatment scenarios on DCMD performance, in terms of membrane void ration, is plotted in

Figure 51.

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

113

Figure 51: Indifference curve between the MAF and UF-MAF pretreatment technologies in

terms of membrane void ratio.

According to the mapped local improvements, we calculate the average global improvement

which is 12.6%, taking into consideration both positive and negative performances of UF-MAF

with respect to MAF. In terms of neat improvement (taking only the positive performances into

consideration), the permeate flux improvement ratio that we can obtain by UF-MAF

pretreatment have a minimum of 66.3% and a maximum of 94.8%.

3.Results and discussion

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

114

3.3.6 Conclusions

In this third section, DCMD treatment of saline dairy effluent has been investigated. The

treatment with two pore sized PTFE membranes showed that using tighter pore sizes could lead

to higher process performances for all feed qualities. Raw effluent pretreatment by

ultrafiltration has a positive effect on both the final permeate flux and the permeate quality

which indicates that the feed content and organic load has a direct and significant effect on the

DCMD performance which is mainly linked to the fouling layer deposition on the membrane

surface and therefore to the membrane wetting that becomes more severe when treating the raw

or the macrofiltrated effluent. Besides, it was found that the fouling layer composition has been

different after treating the three feed qualities with respect mainly to the salts involved in the

fouling layer protein interactions.

Reducing the flow rate during the experiment compromised more the membrane

hydrophobicity in all treated feeds. However, when lower temperature difference was applied

by increasing the permeate stream temperature, the membrane hydrophobicity improved

though lower flux values have been obtained. In all tested operating conditions, DCMD of the

pretreated effluent by macrofiltration-ultrafiltration has led to the best permeate quality with

stable flux levels and better thermal efficiencies.

The membrane cleaning in our study did contribute in the hydrophobicity loss which could be

further investigated to determinate the most suitable cleaning procedure relatively to the fouling

layer nature and composition.

In our study, some observations have raised questions and need further investigation such as

the impact of the membrane pore size, feed velocity and temperature difference on the fouling

layer formation and thickness during dairy effluents DCMD treatment.

The utility theory has been applied in this work to assess the performance and cost-

effectiveness of the saline dairy effluent DCMD treatment with respect to two types of

pretreatments: MAF and UF-MAF. The obtained results have shown that, in terms of DCMD

permeate flux, the neat improvement ratio that we can obtain by UF-MAF pretreatment have a

minimum of 25% and a maximum of 71% in comparison to MAF.

This improvement is associated with an average time overhead of 43% that is needed when

utilizing MAF to achieve the same performance of UF-MAF, under the assumption of an

indifferent user.

3.3.DCMD application for cheese whey wastewater treatment

Sana Abdelkader

115

With respect to the process cost-effectiveness, this time overhead, which is directly linked to

the additional cost that has to be provided when utilizing MAF, have as maximum 72% and as

minimum 28%. Regarding the membrane void ratio, the permeate flux improvement ratio that

we can obtain by UF-MAF pretreatment have a minimum of 66.3% and a maximum of 94.8%,

in terms of neat improvement. This analysis could be extended using the same procedure in

order to evaluate and investigate all the listed indicators that impact the DCMD treatment

applicability, which requires more precise experimental data.

4.Conclusions and perspectives

Sana Abdelkader

116

4 CONCLUSIONS AND PERSPECTIVES

Saline wastewater treatment is a current and growing issue that is getting an increasing attention

due to the water resources scarcity worldwide. These effluents have the particularity of

containing important salt and organic contents which creates the complexity of their treatment.

Besides, the conventional wastewater treatment processes have shown in some cases limited

performances that mainly linked to the effluents salt concentration. In order to address these

issues, new alternatives for saline effluents treatment and water recovery have to be

investigated and analyzed to overcome the present challenges.

To achieve this objective, this research work is focusing on the treatment of saline wastewater

and saline dairy effluents obtained from a cheese industry. First, the limitations of biological

and membrane processes saline wastewater were studied. Then, comes the main objective of

this thesis which is the examination of the feasibility, performance and application of direct

contact membrane distillation (DCMD) of the aforementioned effluents. For this purpose,

DCMD is applied at different experimental conditions and combined with various

pretreatments, namely macrofiltration and ultrafiltration. The cost-effectiveness of the

industrial effluent DCMD treatment is also considered in this work.

In the first section, different treatment processes and wastewater qualities were analyzed and

investigated to show the effect of the NaCl mass percentage and the treatment characteristics

on the final water quality and the processes performances. According to this investigation, the

aerobic treatment was more efficient when combined to the ceramic membrane microfiltration

since we have obtained higher water quality with good overall performance: the microfiltration

led to COD removal of 96%, 76% and 81% for aerobically treated feed of 0.5%, 1% and 2%

NaCl, respectively. Concerning the ultrafiltration, the COD removal reached 96 %, 74%, 85%

for saline feed concentrations of 0.5%, 1% and 2% NaCl, respectively. In our study, the

anaerobic biological treatment didn’t lead to sufficiently good COD removal when compared

to the aerobic process which didn’t improve significantly after the filtration processes. In all

the studied processes, the main COD removal rates were achieved mainly by the biological

treatments. Besides, the final water qualities were very similar in the case of low salt

concentrations (0% NaCl and 0.5% NaCl) in comparison to the higher concentrations (1%, 2%

and 3% NaCl).

4.Conclusions and perspectives

Sana Abdelkader

117

In the second section, the issue of saline effluents treatments is addressed using DCMD. For

this purpose, the response surface methodology was applied to optimize the DCMD treatment

of saline synthetic wastewater via the optimization of the permeate flux (Jp). The studied

process has led to the obtention of a remarkable high quality permeate in terms of electrical

conductivity reduction (99.9%) and COD removal rate (>99.9%). The DCMD optimization

was provided by the application of the response surface methodology in which four different

variables were considered to create the optimization model. The four different variables are

namely: temperature difference ΔT, feed velocity Vf, NaCl concentration [NaCl] and glucose

concentration [Gluc]. The developed regression model demonstrated a good level of

effectiveness and ability to predict the process response with R2= 0.967 and R2Adj = 0.932. The

statistical evaluation and analysis of the model showed that it has a high significance and that

the most influencing variable in the studied ranges are the temperature difference and the feed

flow rate followed by the glucose concentration with lower significant effect. The optimization

procedure has led to the prediction of the maximum permeate flux (34.14 L/m2.h) that could

be achieved under the following optimum experimental conditions: ΔT= 55.23°C, Vf = 0.086

m/s, [NaCl] = 10.08 g/L and [Gluc] = 20.01 g/L. As a final step, the experimental verification

of the developed regression model showed its validity and suitability within the studied range

of variables.

The dairy industries are one of the food processing industrial activities that contribute in the

pressure applied on the environment. This contribution starts from the reception of raw

materials and milk processing to the packaging of the finished products. The dairy industries

are characterized by consuming and discharging massive quantities of water and wastewater,

respectively. Among them, we find the cheese industry that is also characterized by the

production of a particular by-product, the cheese whey. This by-product is one of the most

important components of the discharged saline effluents which increases our need to treat them.

In the third section of this thesis, DCMD treatment of saline dairy effluent has been

investigated. The treatment with two PTFE membranes with two different pore sizes showed

that using tighter pore sizes could lead to higher process performances for all feed qualities.

The raw effluent pretreatment by ultrafiltration has showed a positive effect on both the final

permeate flux and the permeate quality which indicates that the feed content and organic load

has a direct and significant effect on the DCMD performance that is mainly linked to the fouling

layer deposition on the membrane surface. Besides, it was found that the fouling layer

composition was different following the treatment of three different feed qualities, mainly with

respect to the salts involved in the fouling layer protein interactions.

4.Conclusions and perspectives

Sana Abdelkader

118

The variation of the experimental parameters has led to different outcomes in terms of permeate

flux and membrane quality. Reducing the flow rate during the experiment compromised more

the membrane hydrophobicity in all treated feeds. However, when lower temperature

difference was applied by increasing the permeate stream temperature, the membrane

hydrophobicity improved though lower flux values have been obtained. In all tested operating

conditions, DCMD of the pretreated effluent by macrofiltration-ultrafiltration has led to the

best permeate quality with stable flux levels and better thermal efficiencies.

The performance analysis of saline dairy wastewater DCMD process is extended in this work

so as to consider its cost-effectiveness when in it is associated with two types of pretreatment.

To this purpose, a set of parameters (i.e. criteria) that impact the applicability of the DCMD

process in industry are taken into consideration in order to assess its performance. The utility

theory has been applied to assess the performance of the saline dairy effluent DCMD treatment

with respect to two types of pretreatments: MAF and UF-MAF. The obtained results have

shown that, in terms of DCMD permeate flux, the neat improvement ratio that we can obtain

by UF-MAF pretreatment have a minimum of 25% and a maximum of 71% in comparison to

MAF. This improvement is associated with an average time overhead of 43% that is needed to

achieve by MAF the same performance using UF-MAF, under the assumption of an indifferent

user. With respect to the cost-effectiveness, this time overhead, which is directly linked to the

additional cost that has to be provided when utilizing MAF, have as maximum 72% and as

minimum 28%. Regarding the membrane void ratio, the permeate flux improvement ratio that

we can obtain by UF-MAF pretreatment have a minimum of 66.3% and a maximum of 94.8%,

in terms of neat improvement.

This analysis could be extended using the same procedure in order to evaluate and investigate

all the listed indicators that impact the DCMD treatment applicability, which requires more

precise experimental data.

In this research work, some observations have raised questions and need further investigation

such as the impact of the membrane pore size, feed velocity and temperature difference on the

fouling layer formation and thickness during dairy effluents DCMD treatment. Additionally,

the membrane cleaning in our study did contribute in the hydrophobicity loss which could be

further investigated to determinate the most suitable cleaning procedure relatively to the fouling

layer nature and composition.

4.Conclusions and perspectives

Sana Abdelkader

119

The tested treatment concept presents a possible alternative to the conventional treatment

processes of saline dairy wastewater. This alternative is interesting when the valorization of the

concentrated proteins is possible in addition to the water recovery.

In terms of energy consumption, further research could be also dedicated to investigate the

possibility of using waste-heat in the DCMD treatment which could increase its cost-

effectiveness and create an appealing alternative for industrial saline effluents treatment.

References

Sana Abdelkader

120

REFERENCES

A. Alkhudhiri, N. Darwish, et N. Hilal. «Membrane distillation: A comprehensive review.»

Desalination 287 (2012): 2–18.

A. Anglada , R. Ibanez, A. Urtiaga, et I. Ortiz. «Electrochemical oxidation of saline industrial

wastewaters using boron-doped diamond anodes.» Catalysis Today 151 (2010): 178–

184.

A. Bottino , G. Capannelli, O. Monticelli , et P. Piaggio . «Poly(vinylidene fluoride) with

improved functionalization for membrane production.» Journal of Membrane Science

166, 2000: 23–29.

A. Boubakri , A. Hafiane , et S.A. Bouguecha . «Application of response surface

methodology for modeling and optimization of membrane distillation desalination

process.» Journal of Industrial and Engineering Chemistry 20 (2014): 3163–3169.

A. Hausmann , et al. «Fouling mechanisms of dairy streams during membrane distillation.»

Journal of Membrane Science 441 (2013.a): 102–111.

A. Hausmann, et al. «Fouling of dairy components on hydrophobic polytetrafluoroethylene

(PTFE) membranes for membrane distillation.» Journal of Membrane Science 442

(2013.b.): 149–159.

A. Kaminska . Development and Implementation of a Liquid-Entry-Pressure Test Facility for

Assessment of the Wettability of Hydrophobic Membranes. University of Applied

Sciences in Offenburg, 2014.

A. Lee , J.W. Elam , et S.B. Darling. «Membrane materials for water purification: design,

development, and application.» Environ. Sci.: Water Res. Technol. 2 (2016): 17-42.

A. Pierre, J. Fauquant, Y. Le Graat, M. Piot , et J. L. Maubo. «Préparation de

phosphocaséinate natif par microfiltration sur membrane.» Lait 72 75 (1992): 461–

474.

A. Ventosa . «Taxonomy of moderately halophilic heterotrophic eubacteria». In: Halophilic

bacteria.» Boca Raton JFL CRC Press: 1, 1988: 77-84.

References

Sana Abdelkader

121

A. Vijay, S. Aroraa , S. Guptab , et M. Chhabra. «Halophilic starch degrading bacteria

isolated from Sambhar Lake, India, as potential anode catalyst in microbial fuel cell:

A promising process for saline water treatment.» Bioresource Technology 256, 2018:

391-398.

A.I. Khuri , et S. Mukhopadhyay. «Response surface methodology.» John Wiley & Sons, Inc.

2 (2010): 128-149.

A.R. Dincer, et F. Kargi. «Performance of rotating biological disc system treating saline

wastewater.» Process Biochem 36 (2001): 901–906.

A.T. Kivistö , et M.T. Karp. «Halophilic anaerobic fermentative bacteria.» Arch Microbiol

170(5), 2011: 319–330.

A.W. Mohammad, R.K. Basha, et C.P. Leo,. «Nanofiltration of glucose solution containing

salts: effects of membrane characteristics, organic component and salts on retention.»

J. Food Eng. 97, 2010: 510–518.

B. Kempf, et E. Bremer. «Uptake and synthesis of compatible solutes as microbial stress

responses to high-osmolality environments.» Arch Microbiol 170(5) (1998): 319–330.

B. Prado , et al. «Numerical Taxonomy of Moderately Halophilic Gram-negative Rods

Isolated· from the Salar de Atacama, Chile.» System. Appl. Microbiol. 14, 1991: 275-

281.

B.B. Ashoor, S. Mansour, A. Giwa, V. Dufour, et S.W. Hasan. «Principles and applications

of direct contact membrane distillation (DCMD): A comprehensive review.»

Desalination 398 (2016): 222–246.

B.R. Babu, N.K. Rastogi, et K.S.M.S. Raghavarao. «Mass transfer in osmotic membrane

distillation of phycocyanin colorant and sweet-lime juice.» Journal of Membrane

Science, 2006.

C. Antileo , E. Aspe , C. Zaror, M. Roeckel , H. Urrutia , et M.C. Marti. «Differential

bacterial growth kinetic and nitrification of fisheries wastewaters containing high

ammonium and organic matter concentration by using pure oxygen.» Biotechnol. Lett.

19 (3), 1997: 241-244.

References

Sana Abdelkader

122

C. Charcosset. A review of membrane processes and renewable energies for desalination.

Elsevier, 2009.

C. Gostoli, et G.C. Sarti. «Separation of liquid mixtures by membrane distillation.» J. Membr.

Sci. 41 (1989): 211–224.

C. M. Bethke, et S. Marshak. «Brine migration across North America – The plate tectonics of

groundwater.» Annual Review of Earth Planetary Science 18 (1990): 287–315.

C. Nuengjamnong, J.H. Kweon, J. Cho, C. Polprasert, et K.-H. Ahn. «Membrane fouling

caused by extracellular polymeric substances during microfiltration processes.»

Desalination 179 (2005): 117–124.

D. Cheng, W. Gong, et N. Li. «Response surface modeling and optimization of direct contact

membrane distillation for water desalination.» Desalination 394 (2016): 108–122.

D. Jang, Y. Hwang, H. Shin, et W. Lee. «Effects of salinity on the characteristics of biomass

and membrane fouling in membrane bioreactors.» Bioresource Technology 141

(2013): 50–56.

D. Singh, P. Prakash, et K.K. Sirkar. «Deoiled produced water treatment using direct contact

membrane distillation.» Ind. Eng. Chem. Res. 52 (2013): 13439–13448.

D. Winter. Membrane Distillation - a Thermodynamic, Technological and Economic

Analysis. Aachen, Germany.: Schriftenreihe der Reiner Lemoine-Stiftung. D 386.

ISBN 978-3-8440-3706-7., 2015.

D. Winter, J. Koschikowski, D. Düver, P. Hertel, et U. Beuscher. «Evaluation of MD process

performance: Effect of backing structures and membrane properties under different

operating conditions.» Desalination 323 (2013): 120–133.

D.Y. Hou, J. Wang, B.Q. Wang, Z.K. Luan, X.C. Sun, et X.J. Ren. «Fluoride removal from

brackish groundwater by direct contact membrane distillation.» Water Sci. Technol.

61 (2010): 3178–3187.

E. Corcoran, C. Nellemann, E. Baker, R. Bos, D. Osborn, et H. Sav. Sick Water? The central

role of wastewater management in sustainable development. A Rapid Response

Assessment. United Nations Environment Programme,. UN-HABITAT GRID-

Arendal, 2010.

References

Sana Abdelkader

123

E. Drioli, A. Ali, et F. Macedonio. «Membrane distillation: Recent developments and

perspectives.» Desalination, 2014.

E. Drioli, A. Criscuoli, et L.P. Molero. Water And Wastewater Treatment Technologies Vol.

III - Membrane Distillation. Encyclopedia of Life Support Systems (EOLSS), 2009.

E. Quesada , A. Ventosa, F. Rodriguez-Valera, et A. Megias and Ramos-Cormenzana.

«Numerical Taxonomy of Moderately Halophilic Gram-negative Bacteria from

Hypersaline Soils.» Journal of General Microbiology 129, 1983: 2649-2657.

E. Reid, X. Liu, et S.J. Judd. «Effect of high salinity on activated sludge characteristics and

membrane permeability in an immersed membrane bioreactor.» J. Membr. Sci. 283

(2006): 164–171.

F. Carvalho, A.R. Prazeres, et J. Rivas. «Cheese whey wastewater: Characterization and

treatment.» Science of the Total Environment 445–446 (2013): 385–396.

F. He, J. Gilron, et K.K. Sirkar. «High water recovery in direct contact membrane distillation

using a series of cascades.» Desalination 323 (2013): 48.

F. Jia, J. Li, J. Wang, et Y. Sun. «Removal of strontium ions from simulated radioactive

wastewater by vacuum membrane distillation.» Annals of Nuclear Energy 103 (2017):

363–368.

F. Kargi, et A.R. Dincer. «Effect of salt concentration on biological treatment of saline

wastewater by fed-batch operation.» Enzyme Microb. Technol. 19 (1996): 529–537.

F. Kargi, et A.R. Dincer. «Saline wastewater treatment by halophile-supplemented activated

sludge culture in an aerated rotating biodisc contactor.» Enzyme Microbiol Technol

22(6) (1998): 427-433.

F. Laganà, G. Barbieri, et E. Drioli. «Direct contact membrane distillation: modelling and

concentration experiments.» Journal of Membrane Science 166 (2000): 1–11.

G. Guan, X. Yang, R. Wang, et A.G. Fane. «Evaluation of heat utilization in membrane

distillation desalination system integrated with heat recovery.» Desalination 366

(2015): 80–93.

G. Naidu, S. Jeong, S.-J. Kim, I. S. Kim, et S. Vigneswaran. «Organic fouling behavior in

direct contact membrane distillation.» Desalination 347 (2014): 230–239.

References

Sana Abdelkader

124

G. Ozalp, C.Y. Gomec, I.Ozturk, S. Gonuldinc, et M. Altinbas. «Effect of high salinity on

anaerobic treatment of low strength effluents.» Water Science and Technology 48

(2003): 207–212.

G. Rice, S.E. Kentish, A. O'Connor, A. Barber, et A. Philajama. «Analysis of separation and

fouling behavior during nanofiltration of dairy ultrafiltration permeate.» Desalination

236 (2009): 23-29.

G. Tchobanoglous, F.F. Burton, et H.D. Stensel. Wastewater engineering: Treatment and

Reuse. 4th ed. New York: Metcalf & Eddy, McGraw-Hill, 2003.

H. Lee, F. He, L. Song, J. Gilron,, et K. K. Sirkar. «Desalination with a Cascade of Cross-

flow Hollow Fiber Membrane Distillation Devices Integrated with a Hollow Fiber

Heat Exchanger.» AIChE J. 57 (7) (2011): 1780.

H. Singh, M. Boland, et A. Thompson. Milk Proteins: from expression to food, second

edition. New Zealand: Food Science and Technology International Series, 2014.

H.H.P. Fang. «Treatment of wastewater from a whey processing plant using activated sludge

and anaerobic processes.» J Dairy Sci 74 (6) (1991): 2015–2019.

I. Demir, et B. Seyler. «Chemical Composition and Geologic History of Saline Waters in Aux

Vases and Cypress Formations, Illinois Basin.» Aquatic Geochemistry 5 (1999): 281–

311.

I. Vyrides, et D.C. Stuckey. «Adaptation of anaerobic biomass to saline conditions: Role of

compatible solutes and extracellular polysaccharides.» Enzyme and Microbial

Technology 44 (2009.b): 46–51.

I. Vyrides, et D.C. Stuckey. «A modified method for the determination of chemical oxygen

demand (COD) for samples with high salinity and low organics.» Bioresource

Technology 100 (2009.a): 979–982.

I. Vyrides, et D.C. Stuckey. «Fouling cake layer in a submerged anaerobic membrane

bioreactor treating saline wastewaters: curse or a blessing?» Water Sci. Technol. 63

(12) (2011): 2902–2908.

References

Sana Abdelkader

125

I. Vyrides, et D.C. Stuckey. «Saline sewage treatment using a submerged anaerobic

membrane reactor (SAMBR): Effects of activated carbon addition and biogas-

sparging time.» Water Research 43 (2009.c): 933-942.

I.K. Kapdan , et B. Erten. «Anaerobic treatment of saline wastewater by Halanaerobium

lacusrosei.» Process Biochemistry 42, 2007: 449–453.

J. Deshpande, K. Nithyanandam, et R. Pitchumani. «Analysis and design of direct contact

membrane distillation.» Journal of Membrane Science 523 (2017): 301–316.

J. Ge, Y. Peng, Z. Li, P. Chen, et S. Wang. «Membrane fouling and wetting in a DCMD

process for RO brine concentration.» Desalination 344 (2014): 97–107.

J. J. Kehoe, et E. A. Foegeding. «Interaction between β-Casein and Whey Proteins as a

Function of pH and Salt Concentration.» J. Agric. Food Chem. 59 (2011): 349–355.

J. Luo, S. Wei, Y. Su, X. Chen , et Y. Wan. «Desalination and recovery of iminodiacetic acid

(IDA) from its sodium chloride mixtures by nanofiltration.» J. Membr. Sci. 342,

2009: 35–41.

J. Rivas, A.R. Prazeres, F. Carvalho, et F. Beltrán. «Treatment of cheese whey wastewater:

combined coagulation–flocculation and aerobic biodegradation.» J Agric Food Chem

58 (13) ((2010)): 7871–7877.

J. Rivas, A.R. Prazeres, et F. Carvalho. «Aerobic biodegradation of precoagulated cheese

whey wastewater.» J Agric Food Chem 59 (6) (2011): 2511–2517.

J. S. Mounsey, et B. T. O’Kennedy. «Stability of β-lactoglobulin/ micellar casein mixtures on

heating in simulated milk ultrafiltrate at pH 6.0.» Int. J. Dairy Technol. 62 (4) (2009):

493-499.

J. Warczok, M. Gierszewskab, W. Kujawski, et C. Guell. «Application of osmotic membrane

distillation for reconcentration of sugar solutions from osmotic dehydration.»

Separation and Purification Technology 57 (2007): 425–429.

J.A. Prince , D. Rana , T. Matsuura, N. Ayyanar, T.S. Shanmugasundaram , et G. Singh.

«Nanofiber based triple layer hydro-philic/-phobic membrane - a solution for pore

wetting in membrane distillation.» Sci. Rep. 4, 2014.b: 6949.

References

Sana Abdelkader

126

J.A. Prince , D. Rana, G. Singh, T. Matsuura, T. Jun Kai , et T.S. Shanmugasundaram .

«Effect of Hydrophobic Surface Modifying Macromolecules on Differently Produced

PVDF Membranes for Direct Contact Membrane Distillation.» Chemical Engineering

Journal 242, 2014.a: 387-396.

J.A. Prince, G. Singh , D. Rana , T. Matsuura, V. Anbharasi , et T.S. Shanmugasundaram.

«Preparation and characterization of highly hydrophobic poly(vinylidene fluoride)–

Clay nanocomposite nanofiber membranes (PVDF–clay NNMs) for desalination

using direct contact membrane distillation.» Journal of Membrane Science 397– 398,

2012: 80– 86.

J.E. Efome , D. Rana, T. Matsuura , et C.Q. Lan. «Enhanced performance of PVDF

nanocomposite membrane by nanofiber coating: A membrane for sustainable

desalination through MD.» Water Research, 89, 2016: 39-49.

J.E. Miller. Review of Water Resources and Desalination Technologies. SAND 0800, 2003.

J.L. Cohon. Multiobjective Programming and Planning. California, United States: Volume

140, 1st Edition. ISBN: 9780080956497., 1978.

J.-L. Wang, X.-M. Zhan, Y.-C. Feng, et Y. Qian. «Effect of Salinity Variations on the

Performance of Activated Sludge System.» Biomedical and Environmental Sciences

18 (2005): 5-8.

J.-S. Kim, C.-H. Lee, et I.-S. Chang. «Effect of pump shear on the performance of a

crossflow membrane bioreactor.» Water Research 35 (2001): 2137–2144.

K. He , H.J. Hwang , M.W. Woo, et I.S. Moon. «Production of drinking water from saline

water by direct contact membrane.» Journal of Industrial and Engineering Chemistry

17 (2011): 41–48.

K. Kezia, L. Judy, W. Mike, et S. Kentish. «Direct contact membrane distillation for the

concentration of saline dairy effluent.» Water Research 81 (2015): 167-177.

K. Nakoa, A. Date , et A. Akbarzadeh. «A research on water desalination using membrane

distillation.» Desalination and Water Treatment., 2014: 1–13.

References

Sana Abdelkader

127

K. Yang, Y. Yu, et S. Hwang. «Selective optimization in thermophilic acidogenesis of

cheese-whey wastewater to acetic and butyric acids: partial acidification and

methanation.» Water Research 37(10) (2003): 2467–2477.

K.M. Blaschek, W.L. Wendorff, et S.A. Rankin. «Survey of salty and sweet whey

composition from various cheese plants in Wisconsin.» J. Dairy Sci. 90 (4) (2007):

2029-2034.

K.W. Lawson, et D.R. Lloyd. «Direct contact membrane distillation.» J. Membr. Sci. 120

(1996): 123–133.

L. Jianquan, et W. Yinhua. «Desalination of effluents with highly concentrated salt by

nanofiltration: From laboratory to pilot-plant.» Desalination 315, 2013: 91–99.

L. Martinez, et J.M. Rodriguez-Maroto. «On transport resistances in direct contact membrane

distillation.» Journal of Membrane Science 295 (2007): 28–39.

L. Martinez-Diez, et M.I. Vazquez-Gonzalez. «Temperature and concentration polarization in

membrane distillation of aqueous salt solutions.» J. Membr. Sci. 156, 1999: 265–273.

L.D. Tijing, Y. C. Woo, J.-S. Choi, S. Lee, S. H. Kim, et H. K. Shon. «Fouling and its control

in membrane distillation—A review.» Journal of Membrane Science 475 (2015): 215-

24.

L.M. Camacho, L. Dumée, , J. Zhang, J. De Li, M. Duke, et J. Gomez. «Advances in

membrane distillation for water desalination and purification applications.» Water

(Water) 5 (2013): 94–196.

M. A. Izquierdo-Gil, M. C. Garca-Payo , et C. Fernndez-Pineda. «Direct Contact Membrane

Distillation of Sugar Aqueous Solutions.» Separation Science and Technology 349

(1999): 1773-1801.

M. Boukhriss, H. Ben Bacha, K Zarzoum, et k. Zhani. «Study of Modeling and Simulation of

Direct Con-tact Membrane Distillation.» International Journal of Scientific &

Engineering Research 6 (2015): 1317-1325.

M. Gryta. «Effectiveness of Water Desalination by Membrane Distillation Process.» Journal

of Membranes 2, 2012: 415-429.

References

Sana Abdelkader

128

M. Gryta. «Fouling in direct contact membrane distillation process.» Journal of Membrane

Science 325 (2008): 383–394.

M. Khaska, et al. «Origin of groundwater salinity (current seawater vs. saline deep water) in a

coastal karst aquifer based on Sr and Cl isotopes. Case study of the La Clape. Case

study of the La Clape massif (southern France).» Applied Geochemistry 37 (2013):

212–227.

M. Khayet, C. Cojocaru, et C. Garcia-Payo. «Application of Response Surface Methodology

and Experimental Design in Direct Contact Membrane Distillation.» Ind. Eng. Chem.

Res. 46 (2007): 5673-5685.

M. Khayet, et T. Matsuura. «Preparation and characterization of polyvinylidene

fluoridemembranes for membrane distillation.» Ind. Eng. Chem. Res. 40 (24), 2001:

5710–5718.

M. M. A. Saleh, et U. F. Mahmood. «Anaerobic digestion technology for industrial

wastewater treatment.» Eighth International Water Technology Conference, IWTC8.

Alexandria, Egypt, 2004.

M. Samblebe. «Wastewater re-use and desalination a summary of the drivers for, and

technology evolution to satisfy the global push for sustainable water use.» 69th

Annual Water Industry Engineers and Operators’ Conference. 2006. 52-58.

M. Wybult. Characterization and Investigation of Conditions Influencing the Aging of

Hydrophobic, Microporous Membranes for Membrane Distillation. University of

Applied Sciences in Offenburg, 2015.

M.A. De La Fuente, H. Singh, et Y. Hemar. «Recent advances in the characterization of heat

induced aggregates and intermediates of whey proteins,.» Trends Food Sci. Technol.

13(8) (2002): 262–274.

M.A. Izquierdo-Gil, M.C. Garca-Payo, et C. Fernndez-Pineda. «Direct Contact Membrane

Distillation of Sugar Aqueous Solutions.» Separation Science and Technology 34:9

(1999): 1773-1801.

M.C. Lai , et al. «Methanocalculus taiwanensis sp. nov., isolated from an estuarine

environment.» Int J Syst Evol Microbiol. 52, 2002: 1799-806.

References

Sana Abdelkader

129

M.P.Diaz , K.G. Boyd, S.J.W. Grigson , et J.G. Burgess. «Biodegradation of crude oil across

a wide range of salinities by an extremely halotolerant bacterial consortium MPD-M,

immobilized onto polypropylene fibers.» Biotechnol. Bioeng. 79 (2), 2002: 145–153.

M.S. El-Bourawi, Z. Ding, R. Ma, et M. Khayet. «A framework for better understanding

membrane distillation separation process.» Journal of Membrane Science 285 (2006):

4–29.

M.S. Khayet, et T. Matsuura. Membrane Distillation : Principles and Applications. Elsevier

1st edition, 2011.

M.W. Shahzad, M. Burhan, L. Ang, et K. Choon Ng. «Energy-water-environment nexus

underpinning future desalination sustainability.» Desalination 413 (2017): 52–64.

N. Argova, D.G. Lemaya, et J.B. German. «Milk Fat Globule structure & function;

nanosciece comes to milk production.» Trends Food Sci Technol 19(12) (2008).

N. Rovirosa, E. Sanchez, M. Cruz, MC. Veiga, et R. Borja. «Coliform concentration

reduction and related performance evaluation of a down-flow anaerobic fixed bed

reactor treating low-strength saline wastewater.» Bioresour Technol 94 (2004): 119–

1227.

N. Sundaresan, et L. Philip. «Performance evaluation of various aerobic biological systems

for the treatment of domestic wastewater at low temperatures.» Water Science &

Technology 58.4 (2008): 819-830.

O. Kaplinski. «The Utility Theory in Maintenance and Repair Strategy.» Procedia

Engineering 54, 2013: 604 – 614.

O. Lefebvre, et R. Moletta. «Treatment of organic pollution in industrial saline wastewater: A

literature review.» Water Research 40 (20) (2006.a): 3671–3682.

O. Lefebvre, N. Vasudevan, M. Torrijos, K. Thanasekaran, et R. Moletta. «Anaerobic

digestion of tannery soak liquor with an aerobic post treatment.» Water Research 40

(2006.b): 1492–1500.

P. M. Kamrani, O. Bakhtiari, P. Kazemi, et T. Mohammadi. «Theoretical modeling of direct

contact membrane distillation (DCMD): effects of operation parameters on flux.»

Desalination and Water Treatment, 2014: 1-10.

References

Sana Abdelkader

130

P. Onsekizoglu. «Membrane Distillation: Principle, Advances, Limitations and Future

Prospects in Food Industry.» Dans Distillation - Advances from Modeling to

Applications, de Sina Zereshki (Ed.), 234-257. InTech, 2012.

P. Onsekizoglu. «Potential of membrane distillation for production of high quality fruit juice

concentrate- A comprehensive review.» Food Science and Nutrition, 2013.

P.A. Hogan, R.P. Canning, P.A. Peterson, R.A. Johnson, et A.S. Mi. «A new option: Osmotic

Distillation.» Chemical Engineering Progress., 1998.

P.C. Fishburn. Utility Theory. Management Science. USA.: Vol. 14, No. 5, 335-378., 1968.

P.M. Kamrani, O. Bakhtiari , P. Kazemi , et T. Mohammadi . «Theoretical modeling of direct

contact membrane distillation (DCMD): effects of operation parameters on flux.»

Desalination and Water Treatment, 2014: 1-10.

R. Bura, et al. «Composition of extracellular polymeric substances in the activated sludge

floc matrix.» Water Sci. Technol. 37 (1998): 325–333.

R. Dashtpour, et S.N. Al-Zubaidy. «Energy Efficient Reverse Osmosis Desalination

Process.» International Journal of Environmental Science and Development, Vol. 3,

No. 4, 2012: 339-345.

R. Gebauer. «Meosphilic anaerobic treatment of sludge from saline fish farm effluents with

biogas production.» Bioresour Technol 93 (2004): 155–167.

R. Waninge, P. Walstra, J. Bastiaans, H. Nieuwenhuijse, et T. Nyl. «Competitive Adsorption

between â-Casein or â-Lactoglobulin and Model Milk Membrane Lipids at Oil-Water

Interfaces.» J. Agric. Food Chem. 53 (2005): 716-724.

R.B. Saffarini, B. Mansoor, R. Thomas, et H.A. Arafat. «Effect of temperature-dependent

microstructure evolution on pore wetting in PTFE membranes under membrane

distillation conditions.» J. Membr. Sci. 429 (2013): 282–294.

R.W. Schofield, A.G. Fane , C.J.D. Fell , et R. Macoun. «Factors affecting flux in membrane

distillation.» Desalination 77 (0) (1990): 279-294.

R.W. Schofield, A.G. Fane, et C.J.D. Fell. «Heat and mass transfer in membrane distillation.»

Journal of Membrane Science 33 (1987): 299-313.

References

Sana Abdelkader

131

S. Abdelkader , et al. «Application of direct contact membrane distillation for saline dairy

effluent treatment: performance and fouling analysis.» Environ Sci Pollut Res, 2018:

1–14.

S. Al-Asheha, F. Banatb, M. Qtaishatc, et M. Al-Khateeb. «Concentration of sucrose

solutions via vacuum membrane distillation.» Desalination 195 (2006): 60–68.

S. Lattemann, et T. Höpner. «Environmental impact and impact assessment of seawater

desalination.» Desalination 220 (2008): 1-15.

S. Yarlagadda, V.G. Gude, L.M. Camacho, S. Pinappu, et S. Deng. «Potable water recovery

from As, U, and F contaminated ground waters by direct contact membrane

distillation process.» Journal of Hazardous Materials 192, 2011 : 1388–1394.

S.V. Kalyuzhnyi, E.P. Martinez, et J.R. Martinez. «Anaerobic treatment of high-strength

cheese-whey wastewater in laboratory and pilot UASB-reactors.» Bioresource

Technologies 60 (1) (1997): 59–65.

S.V. Panno, et al. «Source Identification of Sodium and Chloride Contamination in Natural

Waters: Preliminary Results.» The 12th Annual Conference of the Illinois

Groundwater Conference of the Illinois Groundwater Consortium. Makanda, 2002.

T.J. McGhee. Treatment of Brackish and Saline Waters. Water Supply and Sewerage, 6th ed.

New York: McGraw-Hill, Inc., 1991.

V.P. Marel, A. Zwijnenburg, A.J.B. Kemperman, M. Wessling, H.Temmink, et W.G.J. van

derMeer. «Influence of membrane properties on fouling in submerged membrane

bioreactors.» J. Membr. Sci. 348 (2010): 66-74.

W. A. Zisman. Relation of the Equilibrium Contact Angle to Liquid and Solid Constitution.

Advances in Chemistry. Washington, DC: American Chemical Society, 1964.

W. Guo, H.-H. Ngo, et J. Li. «A mini-review on membrane fouling.» Bioresource

Technology 122 (2012): 27–34.

W. Janczukowicz, M. S. Zielin, et D. M. bowski . «Biodegradability evaluation of dairy

effluents originated in selected sections of dairy production.» Bioresource

Technologies 99 (10) (2008): 4199–41205.

References

Sana Abdelkader

132

W. Luo, et al. «Effects of salinity build-up on biomass characteristics and trace organic

chemical removal: Implications on the development of high retention membrane

bioreactors.» Bioresource Technology 177 (2015): 274-281.

Wastewater the Untapped Resource. France: The United Nations World Water Development,

2017.

Water and Energy. The United Nations World Water Development, 2014.

Water for Sustainable World. The United Nations Educational Scientific and Cultural

Organization: The United Nations World Water Development, 2015.

X. Shi, O. Lefebvre, K. Kwang Ng, et H. Yong Ng. «Sequential anaerobic–aerobic treatment

of pharmaceutical wastewater with high salinity.» Bioresource Technology 153

(2014): 79–86.

X. Zhang, J. Gao, F. Zhao, Y. Zhao, et Z. Li. «Characterization of a salt-tolerant bacterium

Bacillus sp. from a membrane bioreactor for saline wastewater treatment.» Journal of

Environmental Sciences 26 (2014): 1369–1374.

X. Zhang, Z. Guo, C. Zhang, et J. Luan. «Exploration and optimization of two-stage vacuum

membrane distillation process for the treatment of saline wastewater produced by

natural gas exploitation.» Desalination 385 (2016): 117–125.

List of figures

Sana Abdelkader

133

LIST OF FIGURES

Figure 1: Wastewater in the water cycle (source: UN report 2017). ....................................... 12

Figure 2: Schematic illustration of the different membrane filtration types according to the

retained particle’s size. (Source: lee and al. 2016) ................................................................... 15

Figure 3: Schematic of MD process ......................................................................................... 20

Figure 4: Common types of membrane distillation configurations (adapted from El-Bourawi

et al., 2006)................................................................................................................................. 21

Figure 5: Schematic of Osmotic Membrane Distillation (OMD) (source: Hogan et al., 1998)

.................................................................................................................................................... 23

Figure 6: Aerobic biological reactors ....................................................................................... 30

Figure 7: Anaerobic biological reactors during the nitrogen sparging step ............................ 31

Figure 8:microfiltration membrane and set-up ........................................................................ 32

Figure 9: SEM image of the ceramic membrane (left: membrane cross section, right: top view

of separation layer) .................................................................................................................... 33

Figure 10: SEM image of the cross-section of PSF (18%) membrane ................................... 33

Figure 11:Ultrafiltration set-up ................................................................................................. 34

Figure 12:(a) Bench scale DCMD set-up and (b) schematic of the DCMD set-up [(1)

membrane test cell, (2) circulation pump, (3) heat exchanger, (4) feed water tank (5)permeate

water tank (6)electronic balance]. ............................................................................................. 35

Figure 13: (a) Laboratory scale DCMD set-up and (b) Schematic diagram of experimental

laboratory scale DCMD set-up; (1) membrane test cell, (2) feed water tank, (3) feed pump, (4)

heat exchanger, (5) support heat exchanger, (6) electric heater, (7) heat exchanger, (8) motor

valve, (9) laboratory cooling circuit, (10) filter, (11) electronic balance (Winter et al., 2013).

.................................................................................................................................................... 38

Figure 14: Illustration of the steps of the void ratio estimation: original SEM image (a),

segmented image (b), binarized image (c). ............................................................................... 48

List of figures

Sana Abdelkader

134

Figure 15: MLSS evolution after 8 days in the aerobic bioreactors at different NaCl

concentrations (%). .................................................................................................................... 52

Figure 16: COD removal in aerobic bioreactors at different NaCl concentrations (%) ......... 53

Figure 17: Methane production in the anaerobic bioreactors (13 days incubation time) ....... 54

Figure 18: COD removal in the different anaerobic bioreactors at different NaCl

concentrations (%) ..................................................................................................................... 55

Figure 19:Microbiological observations of sludge samples after Gram staining: anaerobic

sludge at 1% NaCl (a), aerobic sludge at 1% NaCl (b), anaerobic sludge at 2% NaCl (c), aerobic

sludge at 2% NaCl (d) and anaerobic sludge at 3% NaCl (e) .................................................. 57

Figure 20: Microfiltration membrane permeability of the different aerobic bioreactors’

supernatants (control is reverse osmosis water) ....................................................................... 58

Figure 21: COD removal after microfiltration of the aerobic bioreactors’ supernatants and

permeate. .................................................................................................................................... 59

Figure 22: Microfiltration membrane permeability of the different anaerobic bioreactors’

supernatants (control is reverse osmosis water) ....................................................................... 60

Figure 23: COD removal after microfiltration of the anaerobic bioreactors’ supernatants .... 61

Figure 24: Membrane permeability during ultrafiltration of the aerobic bioreactors’

supernatants (control is reverse osmosis water) ....................................................................... 63

Figure 25: COD removal after ultrafiltration of the aerobic bioreactors’ supernatants ......... 64

Figure 26: Membrane permeability during ultrafiltration of the anaerobic bioreactors’

supernatants (control is reverse osmosis water) ....................................................................... 64

Figure 27: COD removal after ultrafiltration of the anaerobic bioreactors’ supernatants ..... 66

Figure 28: Permeate flux evolution during operation time. .................................................... 68

Figure 29: Main effects plot for the permeate flux (Jp). .......................................................... 70

Figure 30: Response surface plots (a, c, e, g, i and k) and contour-line plots (b, d, f, h, g and

l) of the predicted DCMD permeate flux (Jp) as function of: temperature difference and feed

velocity [(a) and (b)], temperature difference and glucose concentration [(c) and (d)],

temperature difference and NaCl concentration [(e) and (f)], feed velocity and glucose

List of figures

Sana Abdelkader

135

concentration [(g) and (h)], NaCl concentration and glucose concentration [(i) and (j)] and feed

velocity and NaCl concentration [(k) and (l)] ........................................................................... 76

Figure 31: Permeate flux and electrical conductivity of raw effluent (R-WW), DCMD

treatment with M1, Tei=55°C, Tco=25°C and Ve=Vc=0.08 m/s ............................................ 80

Figure 32: Permeate flux and electrical conductivity of raw effluent (R-WW), DCMD

treatment with M2, Tei=55°C, Tco=25°C and Ve=Vc=0.08 m/s ............................................ 81

Figure 33: SEM pictures of membrane samples: fouling layers after raw effluent DCMD

treatment with M1 after 7 hours (a) and M2 after 10 hours (b). Operating conditions: ∆T= 30

°C and V=0.08m/s ..................................................................................................................... 82

Figure 34: Permeate flux variation during MAF-WW and MAF-UF-WW DCMD treatment

with: (a) 0.45µm pore size membrane (M1) and (b) 0.2µm pore size membrane (M2).

Operating conditions: Tei=55°C, Tco=25°C and Ve=Vc=0.08 m/s ........................................ 83

Figure 35: Permeate electric conductivity during DCMD treatment with: (a) 0.45µm pore size

membrane (M1) and (b) 0.2µm pore size membrane (M2) of different saline dairy streams.

Operating conditions: Tei=55°C, Tco=25°C and Ve=Vc=0.08 m/s ........................................ 84

Figure 36: Thermal efficiency ƞth and flux J of MAF-WW and MAF-UF-WW DCMD

treatment in comparison to normal tap water test (Reference). Operating conditions: ∆T= 30°C

and (a) V= 0.08 m/s and (b) V=0.04 m/s .................................................................................. 87

Figure 37: Flux behavior during operation time of DCMD treatment with MAF-UF-WW at

∆T=30°C and three velocities (0.08 m/s, 0.06 m/s and 0.04 m/s) ............................................ 88

Figure 38: Thermal efficiency ƞth and flux J of MAF-WW and MAF-UF-WW DCMD

treatment in comparison to normal tap water test (Reference). Operating conditions: V=0.08

m/s and (a) ∆T= 30°C and (b) ∆T= 15°C ................................................................................. 90

Figure 39: Flux behavior during operation time of DCMD treatment with MAF-UF-WW at

V=0.08 m/s and two temperature differences (∆T=30°C and 15°C) ....................................... 91

Figure 40: Permeate flux and thermal efficiency of treated effluents (MAF-WW and MAF-

UF-WW) compared to theoretical prediction at V= 0.04 m/s and 0.08 m/s (∆T=30°C). ....... 92

Figure 41: Permeate flux and thermal efficiency of treated effluents (MAF-WW and MAF-

UF-WW) compared to theoretical prediction at ∆T=15°C and 30°C (V= 0.08 m/s). ............. 93

List of figures

Sana Abdelkader

136

Figure 42: SEM pictures of membrane samples: Reference (a) and fouling layers after 10h

DCMD process of different feed solutions: raw effluent R-WW (b), cross section with R-WW

(c), macrofiltration permeate MAF-WW (d) and macrofiltration-ultrafiltration permeate MAF-

UF-WW (e). Operating conditions: ∆T= 30°C and V= 0.8 m/s, during 10 hours treatment .. 95

Figure 43: Liquid entry pressure (LEP) of membrane samples after 10h DCMD process of

different feed solutions: R-WW, MAF-WW and MAF-UF-WW. Operating conditions: ∆T=

30°C and V= 0.08 m/s ............................................................................................................... 98

Figure 44: Electrical conductivity (µS/cm) at p=0.3 bar during DCMD process of different

feed solutions: R-WW, MAF-WW and MAF-UF-WW. Operating conditions: ∆T= 30°C and

V= 0.08 m/s ................................................................................................................................ 99

Figure 45: Contact angle of new and cleaned membrane samples with different feed solutions:

R-WW, MAF-WW, MAF-UF-WW at T= 22°C. Operating conditions: ∆T= 30°C and V= 0.08

m/s ............................................................................................................................................ 100

Figure 46: SEM picture of a cleaned membrane sample (following the DCMD process of

MAF-UF-WW). Operating conditions: ∆T= 30°C and V= 0.08 m/s .................................... 101

Figure 47: Flux evolution during time for the 6 studied scenarios. ...................................... 107

Figure 48: Indifference curve between the MAF and UF-MAF pretreatment technologies in

terms of DCMD permeate flux. ............................................................................................... 109

Figure 49: Indifference curve between the MAF and UF-MAF pretreatment technologies in

terms of DCMD permeate flux and the improvement provided by UF-MAF. ....................... 110

Figure 50: Interpolation of time with respect to flux. ............................................................ 111

Figure 51: Indifference curve between the MAF and UF-MAF pretreatment technologies in

terms of membrane void ratio. ................................................................................................ 113

List of tables

Sana Abdelkader

137

LIST OF TABLES

Table 1: Compilation of cheese whey wastewater characteristics (adapted from Carvalho et

al., 2013 and Kezia et al., 2015). ............................................................................................... 18

Table 2: DCMD membranes characterization (source: membranes data sheets, *parameters

measured in the laboratory) ....................................................................................................... 39

Table 3: Characterization of the raw and pretreated effluents ................................................. 40

Table 4: Central Composite Design (CCD) variables levels in their coded and actual values.

.................................................................................................................................................... 44

Table 5: Numerical indicators and criteria ............................................................................... 50

Table 6: Turbidity removals following microfiltration of aerobic and anaerobic supernatants

at different salinities ................................................................................................................... 61

Table 7: Turbidity removals following ultrafiltration of aerobic and anaerobic supernatants at

different salinities ...................................................................................................................... 65

Table 8: CCD with predicted and experimental DCMD results. ............................................ 72

Table 9: Optimum operating conditions and their correspondent predicted and experimental

responses. ................................................................................................................................... 78

Table 10: Chemical elements mass percentage by EDX analysis at 10Kv of multiple membrane

samples: without any treatment (Reference) and with 10 hours DCMD treatment of R-WW,

MAF-WW and MAF-UF-WW at the following operating conditions: ∆T= 30°C and V= 0.08

m/s .............................................................................................................................................. 96

Table 11: Chemical elements’ mass percentage obtained by EDX analysis at 20kV (the marked

areas in Figure 46). ND: Not Detectable ................................................................................ 101

Table 12: Experimental scenarios........................................................................................... 106

List of abbreviations

Sana Abdelkader

138

List of abbreviations

Abbreviation

Designation

Contact angle.

CAPEX

Capital Expenditure.

CIP

Cleaning in place.

COD

Chemical oxygen demand.

DCMD

Direct contact membrane distillation.

EDX

Energy dispersive X-ray.

EPS

Extracellular polymeric substances.

LEP

Liquid entry pressure.

0.45µm pore size membrane.

0.2µm pore size membrane.

MAF

Macrofiltration.

MAF-UF-WW

Pretreated effluent by macrofiltration and ultrafiltration.

MAF-WW

Pretreated effluent by macrofiltration.

MLSS

Mixed liquor suspended solids.

OPEX

Operational Expenditure.

Polypropylene.

PTFE

Polytetrafluoroethylene.

R-WW

Raw effluent.

SEM

Scanning electron microscopy.

Total carbon.

TIC

Total inorganic carbon.

TOC

Total organic carbon.

Ultrafiltration.

VoT

Value of time.

List of symbols

Sana Abdelkader

139

List of symbols

Symbol

Unit

Designation

[CH4]

ml/L

Methane concentration.

[Gluc]

g/L

Glucose concentration.

[NaCl]

g/L; %

Nacl concentration.

∆p

bar

Transmembrane pressure.

∆t

hour (h)

DCMD operating time.

Membrane filtration area.

The cost corresponding to the demand function x.

µS/cm

Electrical conductivity.

Indicator.

kg/m2.h

Permeate flux.

Jp,

L/m2.h

Volumetric permeate flux.

LEP

bar

Liquid entry pressure.

ƞth

Thermal efficiency.

bar

Pure water vapor.

L/h

Flow rate.

Effective membrane surface.

°C

Water temperature.

Filtration time.

Time to conduct the process satisfying the demand

Tco

°C

Permeate outlet temperature.

Tei

°C

Feed inlet temperature.

TOC

mg/L

Total organic carbon.

m/s

Velocity.

List of symbols

Sana Abdelkader

140

m/s

Permeate velocity in the laboratory scale DCMD

test cell.

m/s

Feed velocity in the laboratory scale DCMD test

cell.

m/s

Feed velocity in the bench scale DCMD test cell

Vol

Permeate volume.

Demand function.

ΔT

°C

Temperature difference in the DCMD test cell.