Defluoridation of drinking water by hybrid coagulation and

Defluoridationofdrinkingwaterbyhybridcoagulationandfiltrationprocess

Dissertation

zurErlangungdesakademischenGradeseinesDoktorsderNaturwissenschaften

–Dr.rer.nat.–

vorgelegtvon

RatnaningtyasBudhiLestari

geboreninJember,Indonesien

InstitutfürInstrumentelleAnalytischeChemieder

UniversitätDuisburg-Essen

2016

DievorliegendeArbeitwurdeimZeitraumvonOktober2008bisNovember2016im

ArbeitskreisvonProf.Dr.TorstenC.SchmidtamInstitutfürInstrumentelleAnalytische

ChemiederUniversitätDuisburg-EssenundIWWRheinisch-WestfälischenInstitutfür

Wasser,einemAn-InstitutederUniversitätDuisburg-Essen,durchgeführt.

TagderDisputation:30.März2017

Gutachter: Prof.Dr.TorstenC.Schmidt

Prof.Dr.MathiasUlbricht

Vorsitzender: Prof.Dr.EckartHasselbrink

Acknowledgements

1

Acknowledgements

Firstly, Iwould like toexpressmy sinceregratitude tomyadvisorProf.Dr. TorstenC.

Schmidt for the continuous support of my Ph.D study, for his patience, motivation, and

immenseknowledge.Hisguidancehelpedmeinallthetimeofresearchandwritingofthis

thesis.IcouldnothaveimaginedhavingabetteradvisorandmentorformyPh.Dstudy.

Iwould like to thankmany people in the IWW, especially. Dr. AndreasNahrstedt for

providingmethenecessary infrastructure formyresearch,RenéHerzog forhisgreathelp

andverykindsupport,especiallyforthepracticalaspectofmyexperimentinthelaboratory

andmanydrivingtopickupwaterfromriverRuhr,withouthimIwouldnothavebeenable

todotheexperimentalwork.

I am very thankful to Prof.MathiasUlbricht for givingme the opportunity to join his

group during my pre-Ph.D study, gave access to research facilities and his effort as the

secondassessorofthisthesis. Iwouldalso liketothankDr.HeruSusantoforhisguidance

andmyfellowlabmatesFengYuforallthefunweworktogether.

Iwouldalso liketothankthewholememberof IACteam,especiallyMr.GerdFischer

andPD.Dr.UrsulaTelgheder,Dr.HolgerKrohn,Dr.BerndWermeckes,Mrs.ClaudiaUlrich,

Mrs.LydiaVassen,Mr.RobertKnierim,Ms.AlexandraBeermanfortheirsupport.Ithankmy

mastercolleaguesMs.YalingLiu,Ms.YunyunNie,Mr.Thomaspeulenfortheircontribution

insomeofthestudy.IsincerelythankProf.Dr.Heinz-MartinKußwhointroducesmetoProf.

Schmidt.

IgratefullyacknowledgedtheBundesministeriumfürBildungundForschung(BMBF)for

theirfinancialsupportinformofInternationalPostgraduateStudiesinWaterTechnologies

(IPSWaT)scholarships,whichenablemetoattaindoctoraldegreesinGermany.

MyhusbandRomiandmysonGalih thankyou foryourendlessly loveandpatience. I

thankmyparentsfortheirmoralandfinancialsupport,mybigfamilyespeciallyRosi,Rina,

Dewi,WawanandChandra,allmy friendsespecially Iin,Uli,Frieda,Osmond,MbakDiana,

Rabee,Myint,Lena,Dea,Vania,Puti,SannyoandDamifortheirhelpandsupporttofinish

this work. Most of all, I am grateful to Allah SWT, Alhamdulillahirabbilalamiin…

Abstract

2

Abstract

Fluoride contamination in drinking water can cause severe health problems, namely

fluorosis.Defluoridationofdrinkingwaterisapracticaloptiontoovercometheproblemof

excessive fluoride indrinkingwater.Considering thatmostaffected regionsare located in

less developed countries, it is necessary to find a safe and inexpensive defluoridation

techniqueinordertoremovetheexcessfluoridefromdrinkingwater.Thisstudyproposes

twohybridmethodsthathavenotbeeninvestigatedwithemphasisnotonlyonthefluoride

removalbutalsoontheremovalofthealuminumresidueintheproductwater.

Thefirststepofthehybridprocessisbasedoncoagulationandco-precipitationbased

onNalgondatechnique.Alum(Al2(SO4)3)isusedascoagulant,whilelime(Ca(OH)2)accounts

formaintaining the systempH and as precipitation agent.However, the drawback of this

processisasignificantaluminumlevelintheproductwater.

Asystematicprocessstudyhasshownthattheremovaloffluorideoccurredveryfast.It

wasbest carriedoutatneutralpHandwithanexcessiveamountofaluminumcoagulant.

Fluoride ions were adsorbed by precipitated aluminum hydroxide. After certain time the

precipitated aluminum hydroxide collided and enmeshed fine particles and later settled

down.Itissuggestedthattheremovalreactionfollowsasweepmechanism.

Furthermore, to decrease fluoride concentration to the desired concentration, the

optimum alum dosing was successfully determined. Reduction of fluoride concentration

fromaninitialconcentrationof10mg/Ltobelow1.5mg/Lwasbestatanaluminumdosing

of100mg/Lthat iscorrespondingtoanAl3+ toF-molarratio≥7.Meanwhile, for fluoride

with initial concentration of 4 mg/L, an Al3+ to F- molar ratio ≥4, equal to 17 mg/L

aluminum, achieved the same purpose. This amount of aluminum is clearly lower than

needed in the Nalgonda technique which is 16 to 181 mg/L or treating raw water with

fluoridelevelsof2to8mg/L.Loweramountsofaluminumarepreferredtoavoidexcessof

aluminumresidueintheproductwaterandtominimizethesludgeformation.

A species diagram shows that pH plays the most important role on the process

especially incontrollingthequalityofproductwater.Theremovaloffluorideinrawwater

fromtheinitialconcentrationof10mg/Ltobelow1.5mg/LwasachievedinthepHrange6-

Abstract

3

8.AtthispHrangeAl(OH)3haslowsolubilityandeasilyprecipitates.Besides,atthispHthe

pHPZCofAl(OH)3indicatesthattheprecipitateisneutraltopositivelycharged.Inaddition,by

maintaining the pH on this level, the amount ofOH- ions as competing ion to fluoride to

occupyAl(OH)3precipitateisalsosmaller.

As second step in the hybrid process a sand filter has been investigated to dealwith

excess residueof aluminium in theproducewater. The insertionof a sand filter after the

coagulation and co-precipitation step is proposed considering that the high level of

aluminium in the productwater is caused by suspended aluminium that still remained in

productwaterafterlongtermofsettlingdown.Inthelaboratoryscale,sandfiltrationaspart

ofthehybridprocessshowedsuccessfulremovalofaluminiumtoaconcentrationthatwill

notleadtoariskforconsumerhealthbasedonWHOstandard(0.2mg/L).

Finally, a hybrid process of coagulation and co-precipitation with membrane

ultrafiltration has been investigated. The hybrid process also successfully achieved 90%

reductionofaluminumconcentrationinthewaterafterthecoagulationandco-precipitation

stepbytheUFmembraneoperation.Thealuminumconcentrationafterthehybridprocess

fulfilledtheWHOstandardof0.2mg/L.

In conclusion, the study on the hybrid proces coagulation and co-precipitation with

filtration fordefluoridationofdrinkingwaterhasbroughtanew insightonhowtocontrol

the coagulation and co-precipitation process. Extention to this work by using additional

parameter gives a newopportunity on its development. The investigationhas also shown

that the hybrids defluoridation process is considerable technique that can be applied as

alternativestotheexistingone.Investigationtodifferentmaterialofmembranewillopena

newchallengeinitsapplication.

Kurzsfassung

4

Kurzfassung

Fluorid verunreinigung im Trinkwasser kann schwere gesundheitliche Probleme

verursachen, nämlich Fluorosis. Fluoridentfernung von Trinkwasser ist eine praktische

Möglichkeit,umdasProblemzubeseitigen.Aufgrunddessen,dassdiemeistenbetroffenen

Regionen weniger entwickelten Ländern sind, ist es notwendig, eine sichere und

kostengünstigeTechnikzufinden,umdasüberflüssigeFluoridausTrinkwasserzuentfernen.

Diese Studie bittet zwei Hybridmethoden, die nicht nur auf die Fluoridentfernung

konzentrieren, sondern auch auf die Entfernung des Aluminiumrücktstand im

Produktwasser.

Die ersten Maßnahmen des Hybridsverfahrens enstanden durch Koagulation und

Mitfällung, eine Technik, die auf sogenannte Nalgonda Technik basiert. Als

Koagulationsmittel wird Alaun (Al2(SO4)3) verwendet, während Kalkmilch (Ca(OH)2) für die

Pflege des pH-Wertes des Systems und als Fällungsmittel verantwortlich ist. Der Nachteil

diesesVerfahrensistjedocheinsignifikanterhoheAluminiumgehaltimProduktwasser.

DieseStudiezeigte,dassdieEntfernungvonFluoridsehrschnellauftrat.Eswurdeam

besten bei neutralem pH-Wert und mit einer übermäßigen Menge an Aluminium als

Koagulationsmitteldurchgeführt.FluoridionenwurdendurchausgefällteAluminiumhydroxid

absorbiert. Nach einer gewissen Zeit, kollidierte das ausgefällte Aluminiumhydroxid

zusammen und verästelte feine Teilchen und setzte sich später nieder. Es wird

vorgeschlagen,dassdieEntfernungsreaktioneinemSweep-Mechanismusfolgt.

Weiterhin wurde in dieser Studie die optimale Aluminiumdosierung erfolgreich

ermitteln,umdieFluoridkonzentrationaufdiegewünschteKonzentrationzureduzieren.Die

Reduktion der Fluoridkonzentration von einer Initialskonzentration von 10mg/L bis unter

1,5mg/LwarambestenbeieinerAluminiumdosierungvon100mg/L,dieeinemAl3+ zuF-

Molverhältnis ≥ 7 entspricht. Dieses Molverhältnis sichert eine hohe Möglichkeit zur

Fluoridadsorption an Aluminiumhydroxidpräzipitaten als Adsorptionsstellen. Diese Menge

anAluminiumistdeutlichniedrigeralsdie,dieinderNalgonda-Technikverbrauchtist.Diese

entspricht 16 bis 181 mg/L für zur Behandlung für eine von 2 bis 8 mg/L

Fluoridekonzentration in Rohwasser. Selbsverständlich, geringere Mengen an Aluminium

Kurzsfassung

5

werdenbevorzugt,umeinenÜberschussanAluminiumrestimProduktwasserzuvermeiden

unddieSchlammbildungzuminimieren.

DieinderStudieausgebildetSpeziesdiagrammenzeigtediedeutlich,dassderpH-Wert

diewichtigstenRollebeidemProzessspielt,insbesonderebeiderKontrollederQualitätdes

Produktwassers.DieEntfernungvonFluoridinRohwasserausderInitialskonzentrationvon

10mg/L bis unter 1,5mg/L erreichte pH-Bereich von 6 bis 8. Bei diesempH-Bereich hat

Aluminiumhydroxid eine geringe Löslichkeit und fällt leicht aus. Darüber hinaus ist die

Menge an OH-Ionen als konkurrierende Ionen zu Fluorid, um eine Al(OH)3 Präzipität zu

besetzen,ebenfallsgeringer,wennderpHaufdiesemNiveaugehaltenwird.

Als zweiter Schritt im Hybridverfahren wurde ein Sandfilter untersucht, um mit

überschüssigem Aluminiumrückstand im Produktwasser umzugehen. Das Einfügen eines

SandfiltersnachdemKoagulations-undMitfällungsschrittwirdvorgeschlagen,daderhohe

Aluminiumgehalt im Produktwasser durch suspendiertes Aluminium verursacht wird, das

jedoch nach längerem Absetzen in Produktwasser verblieb. Im Labormaßstab führte die

SandfiltrationalsTeildesHybridverfahrenszueinererfolgreichenEntfernungvonAluminium

zu einer Konzentration, die nicht zu einem Risiko für die Verbrauchergesundheit auf der

Grundlage des WHO-Standards (0,2 mg/L) führt. Schließlich wurde noch ein hybrides

VerfahrennämlichderKoagulationundMitfällungmitMembran-Ultrafiltrationuntersucht.

Das Hybridverfahren, durch den UF-Membranbetrieb, führte zum einen erfolgreichener

Betrieb,woerbiszueiner90%ReduktionderAluminiumkonzentration,dieimWassernach

dem Koagulations- und Mitfällung verbleibt, erfolgte. Die Aluminiumkonzentration nach

demHybridverfahrenerfülltedenWHO-Standardvon0,2mg/L.

Schlussendlich die Hybridsverfahren Studiemit dem Einsatz von Koagulation und Co-

Präzipitation mit Filtration zur Defluoridierung von Trinkwasser gewährte einen neuen

EinblickindieSteuerungdesKoagulations-undCo-Präzipitationsprozesses.DieErweiterung

dieserArbeitdurchdieVerwendungvonzusätzlichenParameternbieteteineneueChance

für ihre Entwicklung. Die Untersuchung zeigte auch, dass das Hybridsverfahren eine

bedeutungsvolle Technik ist, die alsAlternative zudembestehenden angewendetwerden

kann. Die Untersuchung verschiedener Membranmaterialien eröffnet eine neue

HerausforderunginihrerAnwendung.

Contents

6

Contents

Acknowledgements...................................................................................................................1

Abstract.....................................................................................................................................2

Kurzfassung...............................................................................................................................4

Contents....................................................................................................................................6

IndexofFigures.......................................................................................................................10

IndexofTables........................................................................................................................14

Chapter1 GeneralIntroduction...........................................................................................15

1.1 Fluorideandfluorosis.....................................................................................................15

1.2 Waterdefluoridation......................................................................................................19

1.2.1 Nalgondatechnique....................................................................................................20

1.2.2 Coagulationandprecipitationusingcalciumchloride................................................22

1.2.3 Contactprecipitation..................................................................................................23

1.2.4 Electrocoagulation......................................................................................................24

1.2.5 Activatedalumina.......................................................................................................26

1.2.6 Ionexchanger..............................................................................................................27

1.2.7 Bonechar....................................................................................................................28

1.2.8 BioSandfilter(BSF)......................................................................................................29

1.2.9 Membraneclasification...............................................................................................29

1.2.10 Reverseosmosisandnanofiltration..........................................................................31

1.2.11 Diffusiondialysis........................................................................................................32

1.2.12 Electrodialysis............................................................................................................33

1.2.13 Summary...................................................................................................................34

1.3 References......................................................................................................................40

Contents

7

Chapter2 GeneralObjectiveandScopeofStudy................................................................45

2.1 Generalobjective...........................................................................................................45

2.2 Scopeofstudy................................................................................................................45

Chapter3 AnalyticalInstruments.........................................................................................48

3.1 Ionselectiveelectrodes(ISE)..........................................................................................48

3.2 Ionchromatography(IC)................................................................................................50

3.3 Inductivelycoupledplasma–opticalemissionspectroscopy(ICP-OES)........................52

3.4 Colorimetry(Filterphotometry).....................................................................................53

3.5 Turbidityanalysis............................................................................................................56

3.6 Elementquantification...................................................................................................57

3.7 References......................................................................................................................58

Chapter4 CoagulationandCo-precipitationforFluorideRemovalusingAluminumSulfate

asCoagulant............................................................................................................................60

4.1 Background....................................................................................................................60

4.2 Theoreticalconsiderations.............................................................................................61

4.3 Experimental..................................................................................................................65

4.3.1 Materials.....................................................................................................................65

4.3.2 Experimentalmethods................................................................................................67

4.3.3 Analyticalmethods......................................................................................................69

4.4 Resultsanddiscussion....................................................................................................70

4.4.1 Effectofcalciumhydroxideonfluorideremoval........................................................70

4.4.2 Effectofsodiumhydroxideonfluorideremoval.........................................................71

4.4.3 Reactiontimeinfluorideremoval...............................................................................73

4.4.4 Effectofaluminumtofluoridemolarratioonfluorideremoval................................75

4.4.5 EffectofpHonfluorideremoval.................................................................................77

4.4.6 Summary.....................................................................................................................80

Contents

8

4.5 References......................................................................................................................82

Chapter5 HybridProcessestoRemoveFluorideandExcessofAluminum.........................84

5.1 Background....................................................................................................................84

5.2 Sandfilterasaseparationmethod................................................................................84

5.2.1 Previousstudyon“BioSand”filter..............................................................................84

5.2.2 Sandasfiltermediainwaterworks.............................................................................85

5.2.3 Removalmechanismoffiltermedia...........................................................................85

5.3 MembraneProcessasseparationmethod....................................................................86

5.3.1 Ultrafiltrationmembranescharacteristic....................................................................87

5.3.2 Membranefoulingandconcentrationpolarization....................................................88

5.3.3 Operationmodeofmembraneprocess......................................................................89

5.3.4 Foulingmechanism.....................................................................................................89

5.4 Aimofstudy...................................................................................................................90

5.5 Experimental..................................................................................................................91

5.5.1 Materials.....................................................................................................................91

5.5.2 Experimentalset-up....................................................................................................93

5.5.3 Analyticalmethods......................................................................................................95

5.6 Resultsanddiscussion....................................................................................................97

5.6.1 Aluminumremovaloftheproductwaterbythesandfilter.......................................97

5.6.2 Furtherfluorideremovalofproductwaterbysandfilter...........................................98

5.6.3 Turbidityremovalofproductwaterbythefiltrationstep..........................................98

5.6.4 FluorideandAluminumremovalbyUFmembrane..................................................100

5.6.5 Permeate-fluxdeclineofultrafiltrationmembranes................................................101

5.6.6 Foulantanalysisandmaterialquantificationofsubstanceretainedfrommembrane

surface...................................................................................................................................103

5.6.7 Summary...................................................................................................................105

Contents

9

5.7 References....................................................................................................................107

Chapter6 GeneralConclusionandOutlook.......................................................................109

6.1 References....................................................................................................................111

ListofAbbreviations..............................................................................................................112

ListofPublications................................................................................................................116

CurriculumVitae...................................................................................................................117

Erklärung...............................................................................................................................118

IndexofFigures

10

IndexofFigures

Figure1-1Reportedendemiccasesoffluorosis,adapted5....................................................16

Figure1-2Illustratedhydrogeochemicalcycleoffluorine,adapted8....................................16

Figure1-3Overviewoffluorideremovalmethods.................................................................20

Figure1-4Solubilitydiagramofα-Al(OH)3(s)27........................................................................21

Figure1-5Illustrationofinteractionsoccurringwithinanelectrocoagulationreactor42.......25

Figure1-6Porediameterofmembrane67..............................................................................30

Figure1-7SchematicdrawingillustratingtheprincipleofdiffusiondialysisusingNaOHas

strippingsolutioninastackofanion-exchangemembranes(ownwork).......................32

Figure1-8Schematicdiagramillustratingtheprincipleofdesalinationbyelectrodialysisina

stackwithcation-andanion-exchangemembranesinalternatingseriesbetweentwo

electrodes,whereA=anodeandC=cathode80............................................................33

Figure2-1Illustrationandtherelationshipsbetweenthechapterinthestudy.....................47

Figure3-1Typicalcalibrationcurveofionselectiveelectrode(METROHM781ph/Ionmeter)1

.........................................................................................................................................50

Figure3-2TypicalillustrationofICsystemwithsuppressor4.................................................51

Figure3-3Schematicoperationofasuppressorfor(a)anion-exchangeand(b)cation-

exchangeseparations5...................................................................................................52

F 54

Figure3-5Chromazurol-Sor3-sulfo-2,6-dichloro-3,3-dimethyl-4-hydroxyfuchson-5,5-

dicarboxylicacid11..........................................................................................................55

Figure3-6Aluminonor5-[(3-carboxy-4-hydroxyphenyl)(3-carboxy-4-oxo-2,5-cyclohexadien-

ylidene)methyl]-2-hydroxybenzoicacidtriammoniumsalt13........................................55

IndexofFigures

11

Figure3-7Basicprincipleofnephelometry14.........................................................................56

Figure4-1SpeciesdistributionofaluminuminsolutionasfunctionofpHwiththepresence

of[F-]=10mg/Land[Al3+]=100mg/L...........................................................................64

Figure4-2ComparisonoffluorideconcentrationsdeterminedbyIonChromatography(IC)

andanIonSelectiveElectrode(ISE)standarddeviation:0.2298mg/L..........................70

Figure4-3EffectofadditionalCaOH2withinitialfluorideconcentrationof10mg/L,without

additionalaluminumsulfatesolutiontofluorideconcentration,pH=7.5.....................71

Figure4-4EndpHofthesystembyadditionofNaOHasbackgroundalkalinity,[F-]=10mg/L

andAl3+toF-molarratio=7............................................................................................72

Figure4-5EffectofNaOHonfluorideremovalwithF0=10mg/LandAl3+=100mg/L,Al3+to

F-molarratio7.................................................................................................................72

Figure4-6CoagulationofspikedRuhrwaterwithF-=10mg/L,withadditionof20mLNaOH

1Masbackgroundalkalinity(Al3+toF-molarratio=7.EndpH=7.5)............................74

Figure4-7CoagulationofspikedRuhrwaterwithF-=10mg/L,withadditionof3mLCa(OH)2

asbackgroundalkalinity(Al3+toF-molarratio=7.EndpH=7.5)..................................75

Figure4-8EffectofaluminumsulfatesolutiondosingonfluorideremovalwithF0=10mg/L,

atpH7.5..........................................................................................................................76

Figure4-9EffectofaluminumsulfatesolutiondosingonfluorideremovalwithF0=4mg/L...77

Figure4-10EffectofpHonfluorideremovalwithinitialfluorideconcentrationof10mg/L.

Al3+toF-molarratio=7.....................................................................................................78

Figure4-11Speciesdistributionofaluminum-fluoridecomplexinsolutionasfunctionofpH

byusingthefluorideandaluminumconcentrationsfromtheexperimentsasinputdata

forcalculationusingChemEQLV3.0................................................................................80

Figure5-1Fluxillustrationtodistinguishmembranefoulingandconcentrationpolarization;

thesolidblacklinedescribesthefluxwithconcentrationpolarization,thesolidlight

greylinedescribesfluxwithfouling................................................................................89

IndexofFigures

12

Figure5-2IllustrationofDead-end(left)andCross-flowfiltration(right)..............................89

Figure5-3Foulingmechanismsofamembrane31..................................................................90

Figure5-4Illustrationofchemicalstructureofpolyethersulfone33.......................................92

Figure5-5SEMimageofunusedPESUF100kDamembranecross-section..........................92

Figure5-6SEMimageofunusedPESUF100kDamembranetopsurface..............................92

Figure5-7Experimentalsetupofhybridcoagulation-precipitationandsandfilter,whereM=

Motor,Fl=Flowmeter,PI=Pressuremeter,..................................................................94

Figure5-8Experimentalsetupillustrationofhybridcoagulationandco-precipitationwithUF

membrane.M=MotorizedStirrer,F=Flowmeter,P=Pressuregauge,.......................94

Figure5-9Aluminuminterferencediagram34........................................................................95

Figure5-10Totaldissolvedaluminumconcentrationofproductwaterbeforeandafter

filtrationstep...................................................................................................................97

Figure5-11Fluorideconcentrationofproductwaterbeforeandafterfiltrationstep...........98

Figure5-12Turbidityremovalbythesandfilter,withv=0.57m/h.Thepointat0minutes

representstheturbidityofwaterbeforeenteringtheinletofsandfiltermodule.........98

Figure5-13Turbidityremovalbythesandfilter,withflowrate=0.69L/h.Thepointat0

minutesrepresentstheturbidityofwaterbeforeenteringtheinletofsandfilter

module............................................................................................................................99

Figure5-14Permeatefluxprofileoftheproductwaterwithdifferentpressureoperationin

theUFfiltrationstepswithdead-endmodebyPESUFmembrane100kDa................101

Figure5-15Rateofpermeatefluxdeclineinthefirst100secondsaftermaximumfluxhad

beenachievedintheUFfiltrationstepswithdeadendmodebyPESUFmembrane100

kDa................................................................................................................................102

IndexofFigures

13

Figure5-16SEMimageofPESUF100kDamembranecross-sectionwithdepositionof

particlesoveristsurfaceafterbeingemployedfor30minutesfiltrationwith2.0bars

withdead-endmode.....................................................................................................102

Figure5-17SEMimageofPESUF100kDamembranesurfaceafterbeingemployedfor30

minutesfiltrationwith1.0barwithdead-endmodetotallycoveredwithretainedwhite

solidsubstanceswith5000timesmagnification...........................................................103

Figure5-18EDXZAFanalysisspectraofretainedsubstancesfromcoagulationand

precipitationstepovertheUFPESmembrane100kDasurface,afterbeingemployed

for30minuteswith2.0bars.........................................................................................104

Figure5-19SEMimageoftheretainedsubstancesovertheUFPESmembrane100kDa

surfaceafterbeingemployedfor30minuteswith2.0barswith50,000times

magnification.................................................................................................................105

IndexofTables

14

IndexofTables

Table1-1Reportedelevatedfluorideconcentrationsinwaterinsomeendemicareasinthe

world...............................................................................................................................17

Table1-2Fluorideconcentrationsinsomefoodsandbeverages...........................................18

Table1-3Selectedtypeofmembraneseparationprocessesbasedondrivingforceand

separationmechanism66................................................................................................30

Table4-1Listofmaterials.......................................................................................................65

Table4-2Characteristicoftapwaterusedinmostexperiments(translatedfromRWW

Rheinisch-WestfälicheWasserwerksgesellschaftmbH:www.rww.de)..........................66

Table4-3TypicalcompositionofriverRuhrwater[internaldataobtainedfromIWWWater

Centre].............................................................................................................................67

Table4-4Summaryofexperimentalparametervariation......................................................68

Table5-1Listofmaterials.......................................................................................................91

Table5-2IonconcentrationsinproductwaterafterfiltrationwithUFmembrane100kDa

withdead-endmode.....................................................................................................101

Table5-3EDAXZAFanalysiscompositionpercentageofretainedsubstancesovertheUFPES

membrane100kDasurface,afterbeingoperatedfor30minuteswith2bar.............103

15

Chapter1 GeneralIntroduction

1.1 Fluorideandfluorosis

Fluoride,asacommondissolvedconstituentofdrinkingwater,causeseitherbeneficial

oradversehealtheffectsontheconsumerdependingonitsconcentration.Intakeoffluoride

inexcessamountsmostcommonlythroughdrinkingwatercancause fluorosis thataffects

teeth and bones health.Millions of people are exposed to excessive amounts of fluoride

throughcontaminateddrinkingwatercausedbynaturalgeologicalsources.

Fluorosis is a chronic disease manifested by mottling of teeth in mild cases (dental

fluorosis) and softening of bones, ossification of tendons and ligaments, and neurological

damage in severe cases (skeletal fluorosis) 1. Therefore, when the levels of fluoride in

drinkingwaterareabovethestandardlimitsetbytheWorldHealthOrganisation(WHO)to

1.5mg/L, theremight be an increased risk of fluorosis. Children aremost at risk because

they are still in the growing age. It is estimated that around 200million people, from 25

nations over the world, suffer from fluorosis 2. India and China, the two most populous

countriesoftheworld,aretheworstaffected.Figure1-1highlightscountrieswithreported

casesoffluorosisincidents3.

Fluoride is the ionic form of fluorine (F2), which in aqueous solution forms F– ions.

Fluorineisthemostelectronegativeandthemostreactivechemicalelement.Ithasastrong

tendencytoacquireanegativecharge.Therefore,itisalmostneverfoundasfluorineinthe

environment but as fluoride. In earth’s crust, fluoride is significantly abundant with an

averageof625mg/kgandmostlyretainedinmineralsbecauseofitslowmobilityinaqueous

environments4.Fluorideformsmineralswithanumberofcationsandsomemineralspecies

oflowsolubilitycontainfluoride,forinstancefluorite(CaF2).

GeneralIntroduction

16

Figure1-1Reportedendemiccasesoffluorosis,adapted5

Intheenvironment,fluorideiscommonlyfoundinthelithospheremainlyasfluorspar,

rock phosphate, cryolite, apatite, mica and hornblende 6. In seawater, fluoride

concentrations reacharound1.2–1.4mg/L, ingroundwatersup to67mg/Land inmost

surfacewaterslessthan0.1mg/L7.Thehydrogeochemicalcycleoffluorineisillustratedin

Figure1-2.

Figure1-2Illustratedhydrogeochemicalcycleoffluorine,adapted8.

GeneralIntroduction

17

Waterwithhighfluorideconcentration isusuallyfoundatthefootofhighmountains,

and inareaswithgeologicaldepositsofmarineorigin.Typicalexamplesare thegeological

beltfromSyriathroughJordan,Egypt,Libya,Algeria,MoroccoandtheRiftValley.Another

beltistheonestretchingoverIndia,northernThailandandChina.Similarareascanbefound

intheAmericasandinChinaandJapan3.

Fluorideisalsofoundatelevatedconcentrationsneartoplaceswherevolcanicactivity

exists, especially with the emission of fumarolic gases through fumaroles. Fumaroles are

ventsintheearthcrustthatemitsteamandgases.Theinteractionofvolcanicgaswithhigh

temperature(600–700°C)rockwallformsacidiccondensate.ThiscondensatehaslowpH

andisrichinhalogencontentafterescapethroughfumaroleswithasuddenpressuredrop.

After condensation and oxidation, the rock becomes a halogen rich mineral, including

fluoride9.Thermalwatersoftenfoundinvolcanicareasarealsorich influoride,especially

those with high pH 4. Some fluoride minerals are categorized as industrial raw material

includingcryolite(Na3AlF6),whichisusedfortheproductionofaluminumandpesticides6.

Sodiumfluoride(NaF)isasourceoffluoridethatisaddedtotoothpasteinordertoprotect

teethagainstdentalcaries.ThefollowingTable1-1summarizesreportedelevatedfluoride

concentrationsinseveralareasintheworld.

Table1-1Reportedelevatedfluorideconcentrationsinwaterinsomeendemicareasinthe

world

Location(areain) DetectedFluorideConcentration(mg/L)

Typeofwater References

NorthernCape,SouthAfrica 30 Groundwater 10

Agra,India 210 Groundwater 11

RiftValley,Ethiopia 26 Groundwater 12

UnitedRep.Tanzania 95 Groundwater 13

LakeNakuru,Kenya 2800 Surfacewater 14

Ijen(crater)Lake,Indonesia ~1500 Surfacewater 15

Chihuahua,Mexico 5.9 Groundwater 16

ElPaso,TX,USA 4.8 Groundwater 16

GeneralIntroduction

18

Fluoride present in drinking water is considered an important source for the human

body.Fluoride isalsopresent insomefoods insignificantconcentrations.Table1-2shows

somefoodswithsubstantialfluoridecontents.

Table1-2Fluorideconcentrationsinsomefoodsandbeverages

Food/beverage Fluorideconcentration References

BarleyandRice ~2mg/kg 17

Meatingeneral 0.2–1mg/kg 17

Fish 2–5mg/kg 17

Fishproteinconcentrate ~370mg/kg 17

Vegetableandfruit 0.1–0.4mg/kg 17

HumanbreastMilk <0.02mg/L 6

Cow’smilk 0.02–0.05mg/L 18,18in6

Tealeaves(dryweight) 400mg/kg 6

Tea(infusion) 0.5–1.5mg/L 6

Freshfruitjuice 0.3–0.5mg/L 6

Beer 0.3–0.8mg/L 6

Wine 6–8mg/L 6

Cereal 4.2mg/kg 13

Fluorideplaysanimportantroleinthedevelopmentofteethenamelinyoungchildren

and strengthening the bone matrix. Bones and teeth consist mainly of apatite, that is

composedof amixtureofhydroxyapatite (HAP), (Ca10(PO4)6 (OH)2) and fluorapatite (FAP),

(Ca10(PO4)6F2). F2 and (OH)2 in these compounds are interchangeable. Due to the dietary

behaviourinfluorideconsumption,theratioofFAPtoHAPinhumanteethandbonescanbe

verylow.Thisleadstoeasilysolubleteethmaterialunderacidicconditionthatinthedental

practice is called dental caries. On the contrary, a high ratio of FAP to HAP can cause

fluorosis19.

WHOsupportsthehypothesisthattheprevalenceofdentalcariesisantiproportionally

relatedtotheconcentrationoffluorideindrinkingwater3Sincethediscoveryofthecaries-

preventive effect of fluorides in the 1930’s, different forms of fluoride administration

programs have been implemented. Fluoride has been added to different media such as

water,salt,toothpaste,andmilk20,21.

GeneralIntroduction

19

Later,itwassuggestedthatexcessinfluorideuptakemayleadtoharmfuleffectsonthe

human body and that there is a relationship between the concentration of fluoride in

drinkingwater and the prevalence of dental fluorosis 15. In the range of 0.8 to 1.2mg/L,

fluorideisconsideredtobeeffectiveinpreventingdentalcaries,butconcentrationshigher

than1.5mg/Lare reported tobeharmful 22.A study inAsembagus, Indonesia, suggested

that the risk for dental fluorosis is already increased at only 0.5mg/L fluoride in drinking

water and that 1.1mg/L fluoridemay cause skeletal fluorosis 15. This shows that climate,

drinking and dietary behaviour may determine the need or danger of fluoride uptake

throughdrinkingwater.

Thedentaleffectsoffluorosisdevelopmuchearlierthantheskeletaleffects inhuman

beingswhen they are exposed to large amounts of fluoride. Symptoms can developwith

regular consumption ofwater containing fluoride concentrations as low as 1 to 2mg/L 1.

Dental fluorosis typically can be seen as staining and pitting of the teeth. Inmore severe

cases,alltheenamelmaybedamaged.Childrenareparticularlysensitiveduringtheyearsof

teeth formation.During the growthphaseof the skeleton, a relatively high fractionof an

ingestedfluoridedosewillbedepositedintheskeleton.Long-termintakeoflargeamounts

andchronichigh-levelexposuretofluoridecan leadtopotentiallysevereskeletal fluorosis

becausefluorideaccumulatesinthebonesprogressivelyovermanyyears.

1.2 Waterdefluoridation

Water defluoridation is one of the options available to prevent fluorosis endemic in

areas that have high concentrations of fluoride in their water sources. From previous

studies, water defluoridation methods can be categorized into three groups, namely

coagulationandprecipitation,adsorption,andmembraneprocessesasshowninFigure1-3.

The main purpose in water defluoridation is to reduce the fluoride concentration to the

permissiblelevelsetbyWHOto1.5mg/L.Thefollowingsectiondiscussesselectedmethods

andtechniquesondefluoridation.

GeneralIntroduction

20

Figure1-3Overviewoffluorideremovalmethods

1.2.1 Nalgondatechnique

Oneoftheapplicabletechniquestoremoveexcessivefluoridefromwateriscoagulation

and precipitation. Lime (Ca(OH)2) and alum (Al2(SO4)3) are the most commonly used

coagulants.TheNalgondatechniqueisoneofthemoststudiedinthiscategory.Itisnamed

after the Nalgonda area in India where fluorosis is widespread. This technique was

developed by the National Environment Engineering Research Institute (NEERI) Nagpur,

Indiaafterextensivetestingofmanymaterialsandprocesses23,24.TheNalgondatechnique

involves addition of alum and lime followed by mixing, flocculation, precipitation,

sedimentationanddisinfection.

Thefirstremovalprocessoccursthroughprecipitationfollowinglimedosing

2NaF(aq)+Ca(OH)2(aq)→2Na(OH)(aq)+CaF2(s) (1)

ThisprocesswillleadtotheprecipitationofcalciumfluorideandusuallyraisethewaterpH

up to 12. Some reports show that by only lime dosing it was possible to reduce fluoride

concentrationsfrom10mg/Lto8mg/L1,23,25.Byaddingalumintothewater,coagulationis

facilitated. When alum is added to water, essentially two reactions occur. In the first

reaction,alumreactswithsomeofthehydroxidetoproduceinsolublealuminumhydroxide

GeneralIntroduction

21

[Al(OH)3]astheactivesurfaceoffluorideadsorption.Inthesecondreaction,aluminumions

react with fluoride that is still present in the water to form fluoride complexes and

aluminumhydroxylfluoride24.Thereactionscanbedescribedasfollows:

Aluminumprecipitation:

2Al3++6H2O→2Al(OH)3+6H+ (2)

Co-precipitation:

F–+Al(OH)3→Al–Fcomplex+undefinedproduct (3)

Compared to commondrinking-water coagulationand flocculation, a largerdosageof

alum is normally required in the defluoridation process. In practice, 16 – 181mg/L Al is

neededtotreatrawwaterwithfluoridelevelsof2–8mg/L24.Coagulationiscarriedoutby

rapid stirring, while flocculation is facilitated by slow stirring. The process continueswith

sedimentation.Thebest fluoride removal isaccomplishedbyusingalumascoagulantata

neutralpHrange6.5–7.526.AsitisshowninFigure1-4solubilityofaluminumisminimized

atneutralpHbasedonthesolubilitydiagramofα-Al(OH)3(s)27.

Figure1-4Solubilitydiagramofα-Al(OH)3(s)27

The involved chemicals in this technique are fairly regular. After years of experience,

someimprovementshavebeenmadetoadaptwiththelocalcommunityfactor.Theneeded

chemicals are already prepared in particular packaging with properly dosage to avoid

GeneralIntroduction

22

miscalculation.TheNalgondatechnique isalsoreportedtobeadaptabletodrinkingwater

equipmentsuchashandpumps23,28.

However,aluminumandlimedosagedependontheinitialconcentrationoffluoridein

the water. Thus a concentration test has to be done before treating the water. Wrong

dosageofalumwillleadtoinappropriatepHofthewater29.InappropriatepHofthewater

willleadtoineffectivetreatmentandcancausetheformationofAl3+,Al(OH)2+,Al(OH)2+that

aredissolvedinwaterandtoxictohumans,thuseliminatingthebenefitofthetechnique30.

Previousstudiessuggestedthattheresidualaluminum,whichmayreachconcentrations

above0.2mg/L, the threshold limit value setbyWHO,under certain circumstances could

cause several potential health risks to humans. Excess of aluminum concentrations in

drinkingwater is suspected to have an effect on cognitive decline 31. Excessive aluminum

uptakeisalsobelievedtoincreaseprevalenceofAlzheimerdisease30,32,33.Besidesthat,due

to the use of aluminum sulfate as coagulant, the sulfate ion concentration increases

tremendously, and in few cases, it exceeds the maximum permissible limit of 400 mg/L,

whichmaycausecatharticeffectsinhumanbeings23.

Another disadvantage is that treated water has a specific taste due to chemicals

addition.Sincethewatermatrixmaychangeovertimeandseason,regularanalysisoffeed

andtreatedwaterisrequiredtocalculatethecorrectdoseofchemicalstobeadded.Asthe

process involves an excessive amount of aluminum sulfate as coagulant, this technique

resultsinformationoflargevolumesofsludge.Therefore,largespaceisrequiredfordrying

thesludgeanditsdisposalthatleadtoadditionalcostforit.

1.2.2 Coagulationandprecipitationusingcalciumchloride

Recently,calciumchloride(CC)hasbecomeanalternativetolimeadditionpriortothe

coagulationwithaluminumcoagulant inthefluorideremovalprocess.CCispreferredover

limebecause itproduces lesssludgethan lime34.CCwillprecipitatewith fluoride forming

CaF2.Butasdiscussedintheprevioussection,precipitationaloneisnotsufficienttoremove

fluoride to concentrations below 1.5 mg/L. Higher CC dosage leads to increased water

hardness.CC isusuallybeingapplied inwastewater treatmentplants treating fluoriderich

wastewatersuchas fromglass,ceramicandsemiconductor industries.However, thereare

GeneralIntroduction

23

stilldrawbacksonitsapplicationsuchasveryfineprecipitationofCaF2requiringlongertime

forsedimentation.

1.2.3 Contactprecipitation

Contact precipitation is a practicable technique for defluoridation in less developed

countries. It ismostly applied in Africa. This technique could remove fluoride fromwater

through the principle of adsorption of fluoride by bone char and precipitation by the

addition of calcium chloride andmonosodiumphosphate (MSP) in a columnbed 35. Bone

charconsistsmostlyofhydroxyapatite(Ca5(PO4)3OH)andthepercentageofhydroxyapatite

variesdependingon itscalcinationprocess.At first, thecolumn issaturatedwithwater to

avoidtheformationofairbubblesthatwillreducetheremovallevel.

Theremovaloffluorideworksintwoways.First,adsorptionoffluoridetakesplaceon

theactivesurfaceofhydroxyapatite,afterthereleaseofOH-ions.Second,calciumchloride

andMSPdissolveastheyarecontactedwithwater.Inthosetwoprocessescalciumfluoride

andfluorapatiteareformed36.Incaseoftheuseofcalciumphosphatepellets,precipitation

occurs as thepellets slowly release calciumandphosphate ionswhen they are contacted

withwater.Thereactionsissuggestedtooccurasfollows17:

Dissolutionofcalciumchloride:

CaCl2.2H2O(s)→Ca2++2Cl–+2H2O (4)

DissolutionofMSP

NaH2PO4.H2O(s)→PO43-+Na++2H++H2O (5)

Precipitationofcalciumfluoride:

Ca2++2F–→CaF2(s) (6)

Precipitationoffluorapatite:

10Ca2++6PO43–+2F–→Ca10(PO4)6F2(s) (7)

It has been concluded that the contact bed also acts as deep bed filter for the

suspension of fine precipitate 37. The optimum contact time is the critical point for this

GeneralIntroduction

24

technique. Though it is reported that 20 to 30 minutes of contact time had shown an

excellent result, previous studies failed to determined the optimum contact time 17. Too

short contact time leads to reduction in removal capacity and too long contact timewill

minimizeremovalefficiency.Themajor issueforthescaleupapplication isthe lifetimeof

the filter and maintenance due to the continuous supply of calcium and phosphate. As

mentionedabove,thepotentialutilizationofbonecharalsoneedstotakeintoaccountlocal

beliefsinspecificregion.

1.2.4 Electrocoagulation

Electrocoagulation (EC) is commonly used to remove fluoride from industrial

wastewater 38. Some industries, e.g. coke manufacture, electronics and semiconductor

manufacture, electroplating operations, steel, glass and aluminum manufacture, metal

etching,woodpreservatives, andpesticideand fertilizermanufacturedischarge significant

quantities of fluoride containingwastewater in formof hydrogen fluoride (HF) or fluoride

ions(F−),dependinguponthepHofthewaste25,39.Besidefluoride,thewastewatertypically

alsocontainsheavymetals,phenol,andotherorganicchemicals.

EC was also applied to remove fluoride from drinking water 40. In EC the aluminum

sorbent isgeneratedelectrochemicallyusingasacrificialelectrodeasanalternativeto the

conventional alum addition or treatmentwith activated alumina. Themethod principle is

shown in Figure 1-5. As the electric current passes through the anode, the aluminum

electrode is oxidized to aluminum ions. After that, aluminum ions are transformed to

polymericspeciesorAl(OH)3flocs,whichcanco-precipitateoradsorbthefluorideions41.

GeneralIntroduction

25

Figure1-5Illustrationofinteractionsoccurringwithinanelectrocoagulationreactor42

Nevertheless, the kinetic of this method is still under debate with two opposite

opinions.Eachopinionsupportstheirfindingwithlaboratorydata.Onesidesuggestedthat

the defluoridation rate of the EC process is a first-order reactionwith respect to fluoride

concentration43,44.Thereactionkineticsisdescribedasfollows:

Ft−=[F0−].e(−k1t),

where:

Ft- =fluorideconcentrationattimet

F0- =initialfluorideconcentration

k1 =first-orderrateconstant

t =reactiontime.

On the other hand, there is a different interpretation of the experimental results 41.

Because their results demonstrated that k1 declines as the initial fluoride concentration

rises,theyconcludedthatthereactioncouldnotbeafirstorderreaction.Thek1 inafirst

orderreactionmustnotchangewiththeinitialfluorideconcentration.Thedefluoridationof

theECprocess,therefore,shouldbeapseudo-first-orderreaction.

Another study reported that freshly generated Al-sorbent is able to reduce fluoride

concentration from 16 to 2 mg/L in 2 minutes 25. However, in that experiment the

GeneralIntroduction

26

adsorption of fluoride ions on Al-sorbent reaches equilibrium at 2 mg/L of fluoride

concentrationthroughoutthe12minutestreatment.Thispreventedtheadsorptionprocess

tofurtherreducethefluorideconcentrationtolessthan1.5mg/L.Besides,thepresenceof

sulfateionsledtolowerdefluoridationefficiency45.

1.2.5 Activatedalumina

Activated alumina is one of themost frequently used adsorbents to remove fluoride

fromdrinkingwater.Activatedaluminathatcommonlyreferstoγ-Al2O3iscreatedfromthe

dehydroxylationof aluminumhydroxides.Activatedaluminahas veryhigh surface area to

weightratioduetomanyporessthatisformedduringtheformation.TheBETsurfaceofγ-

Al2O3 is known to be 174 - 192 m2/g, depending on its calcination temperature 46. The

mechanism of fluoride removal, although not fully clear, is based on the electrostatic

interactionbetweenthefluorideionandthepositivesurfacechargeofalumina(pHpzc9.5)27.

Furthermore, its removal efficiency is affected by total surface area, surface charge, pH

condition,andalsothenumberofavailableadsorptionsitesonthemediasurfaceaswellas

crystallinestructure.Differentcrystallinestructuresofaluminawillleadtodifferentbinding

energycapacitytointeractwithfluorideions27.

ThepHconditionplaysaroleondissolutioneffectofthecompoundsinwaterthatwill

leadtothereleaseofbothaluminumandfluorideionintothewater.Studiesshowthatthe

pointofzerocharge(pHpzc)isinthepHrangeof6.2to9.647-49.WhenpHisabovethepHpzc,

fluoridesorptiondecreasesduetotheelectrostaticrepulsionbetweenthealumina’ssurface

andfluorideanionsduetothenetnegativesurfacechargeofalumina.Inaddition,hydroxide

ions become stronger competitors to the fluoride ions for occupying the active sites on

activatedalumina.AtpHbelowpHpzc,itisreportedthatfluorideremovalmightreachupto

94% 50. Several studies give a good agreement on the optimum pH range for fluoride

adsorption on alumina between pH 5 to 6 47 48. At pH below 5, the solubility of alumina

increasesandwhenalumina isdissolved,thesorbedfluoridewillbereleased51.Therefore

thismethodissensitivelypHdependent.

Thoughadsorptionwithactivatedaluminawassuggestedasrelativelysaveandefective

method, still there is another challenge on its application namely regeneration. Different

GeneralIntroduction

27

grade of activated alumina has different lifetime. Monitoring of water quality has to be

performedinordertocontroliftheadsorbentissaturatedandhastoberegenerated.

1.2.6 Ionexchanger

It is reported that fluoride removal by using anion-exchange resins is very complex

considering the order of selectivity for anionic species by comercially available anion

exchangeresins.Theselectivityisasfollows:citrate>SO42−>oxalate>I−>NO3

−>CrO42−>

Br−>SCN−>Cl−>formate>acetate>F−52.

Fromthelistabove,itcanbeseenthatfluoridehastheleastselectivitywhilecitratehas

the highest selectivity. Due to the low selectivity for fluoride, anion exchangers are less

considered to remove fluoride from drinking water. Therefore, several studies have

examinedmodified cationexchangers for fluoride removal (seebelow).Researchworkon

fluoride removal has been carried out using metal doped cation exchangers such as

lanthanum(III)53. Inaddition, inorganiccationexchangerssuchassilicagel54andalumina

gelhasalsobeenstudied55.

Byanion-exchange resin, the removal canbedescribedas the following ionexchange

mechanism23,56:

Resin-NR3+Cl−+F−(aq)→Resin-NR3+F−+Cl− (8)

Thefluorideionsreplacethechlorideionsoftheresinuntiltheequilibriumisreached.

The resin is regeneratedbybackwashingwithwater that is supersaturatedwithdissolved

sodium chloride salt. Regenerated chloride ionswill replace the fluoride ions and lead to

rechargeoftheresin.

Cation-exchanger resin dopped with metal ions e.g. Aluminum, Iron, Lanthanum, etc

removesfluoridewithsimilarmechanismasadsorptionmechanism.Themechanismcanbe

describedasfollow:

Resin-SO3H+F−(aq)→Resin-SO3Hδ+----Fδ- (9)

Due toH-bonding, the selectivityof fluoride in cationexchangers ishigher than theother

ion.Thepreferenceofadsorptionofanionsissupposedtobeinthefollowingorder:F−>Cl−

>NO3−>SO4

2−.

GeneralIntroduction

28

Nevertheless, ion exchanger has drawbacks in its utilisation. First, it requires trained

operatorsduetoitsadvancedtechnique,maintenance,anddatainterpretation.Second,the

concentratedfluoridewastefromresinsregenerationneedstobetreatedseparatelybefore

finaldisposal.Third,thetreatedwaterhasalowpHandhighlevelofchloridethatneedsto

bepolishedbeforebeingconsumed.Furthermore,syntheticresinsarenotaneconomically

viableoptionforlessdevelopedcountries1,57.

1.2.7 Bonechar

The application of bone char in thewater treatment process to remove fluoride has

been known for decades 57. Thismethod is based on the sorption of fluoride by charred

bone.Modificationoftheprocessbyaddingbrushite(CaHPO4·2H2O)andcalciumhydroxide

to improve the efficiency of the bone char method was investigated 58. The addition of

brushite leads the increase of bone char adsorption to the maximal until it reach

supersaturated condition and formed fluorapatite. Itwas found that the fluoride removal

increases 20 times compared to the process without addition of brushite. Many

investigations have been done on the sorption capacity of bone char which was mainly

obtainedfrompyrolysisofcattlesbone57,59andfromfishbone60.Anotherstudyproposed

the mechanism of the fluoride removal by bone char is caused by exchange between

phosphate ions in the charred bone with fluoride ions in the treated water and it is

endothermic61.

Ithasbeenreportedthatmorethan700Lofwaterwithinitialfluorideconcentrationof

12mg/LcanbefilteredtobelowtheWHOstandardof1.5mg/Lbyabonecharcolumnof

height50cmanddiameter15cm(equivalentto>80filterbedvolumes)62.However,bone

charefficiencymaydecreaseiftheproductwateriswithdrawnatthehighrate.Inaddition,

as an adsorbingmedium, renewal of bonechar is needed when it is saturated. However,

attempts topredict the saturationpoint failed so far.Therefore,monitoringneeds special

attention17.Besides,theoriginofthebonecouldbeofinterestforsomegroupsofpeople

considering their faith. For instance, Muslims will not use charred bone that came from

swine,aswellasHindusthatwillnotusecharredbonefromcattle.

GeneralIntroduction

29

1.2.8 BioSandfilter(BSF)

TheBioSandFilter (BSF) isabiological filter thatutilisesthebiologicalactivitycoupled

withcontinuouslyoperatedslowsandfiltrationprocessestothehouseholdlevel.TheBSF’s

biologicalfiltration,appliedphysicalfiltration,adsorptionbythesandparticlesaswellasthe

formationofalivinglayerintheuppermostlayersofsand.Thismicroorganismlayerthatis

particularlydenseinaboutthetop1-3cmisknownasthebio-layer.Theoriginalpurposeof

BSFwastoremoveturbidityandbacteria.PreviousstudiessuggestedthatBSFoperatedat

flowrateof15andwithoutanymodificationforfluorideremovalindicatedthatnofluoride

removal and or no consistent result in the experiments 63 or 20 L/h are able to remove

turbidityupto1NTU43,44.

Modification with zeolites were investigated to remove fluoride, and the result

indicated54%offluoridehasbeenremovedfromwater64.WhileanadaptationofBSFusing

lateriticclay,localbrickandbonecharwereinvestigated,onlyadaptationwithbonecharas

filtermediawasabletoremovefluoridefrom8.6mg/LtomeettheWHOlimitof1.5mg/L.

However,itwasconcludedthatthefiltermediaareeasilyexhaustedandpretreatmenthas

tobeperformed65.

1.2.9 Membraneclasification

Nowadays,membrane processes have emerged as a preferred alternative to provide

safe drinkingwaterwithminimumproblems associatedwith other conventionalmethods

suchasexcessivemineralcontent,hygienicproblemsfromthepresenceofmicroorganisms,

etc.Thereareseveraltypesofmembranesthatareclassifiedbytheirstructure,whichcan

behomogeneous,heterogeneous, symmetricor asymmetric; by their nature,biological or

synthetic; and according to the nature of the transport, active or passive. Synthetic

membranescanalsobedividedintoorganic(polymericorliquid)andinorganic(ceramicor

metal).Table1-3showsageneralclassificationofmembranesandexemplaryapplications.

GeneralIntroduction

30

Table1-3Selectedtypeofmembraneseparationprocessesbasedondrivingforceand

separationmechanism66

Membraneprocess MembranebarrierDrivingforce

Separationmechanism Example

Microfiltration(MF) Porous(d>100nm) P Sizeexclusion(sieving) Sterilefiltrationofaqueoussolutions

Ultrafiltration(UF) Porous(d<100nm) P Sizeexclusion(sieving) Concentrationofbiomacromoleculesolutions

Nanofiltration(NF),aqueous

Nonporousormicroporous,charged P

Solution-diffusionDonnanexclusion

Removalofbi-/multivalentionsfromwater

Reverseosmosis(RO) Nonporouspolymer P Solution-diffusion Ultrapurewaterproduction

Electrodialysis(ED) Nonporousorporous E Solution-diffusion,sizeexclusionorchargeexclusionWaterdesalination,

Porousmembranes are used forMF andUF processes, and their principle is physical

separation.NFmembraneshaveporesinthenanometerscalerange,betweentherangeof

UF andRO, andROmembranes theoretically havenopores, though in practice there are

smallmanufacture defects that result in pore formation. In RO, process separation takes

place by the principle of diffusion through the membrane. Figure 1-6 illustrates the

comparativesizeofthemembraneporeanddiameterofsomesubstances.

Figure1-6Porediameterofmembrane67

Theefficiencyofamembrane isdeterminedby itsselectivityandthefluxthroughthe

membrane. Selectivity is the capacity of the membrane to retain some particles and let

otherspass.ThisisnormallyexpressedasRetention(R)thatisgivenby:

GeneralIntroduction

31

where:

R =retention,dimensionless[-]

$% =soluteconcentrationinthepermeate[w/v]

$& =soluteconcentrationinthefeed[w/v]

1.2.10 Reverseosmosisandnanofiltration

Not all membrane types that were mentioned in the previous section are able to

remove fluoride directly. There are two types of membranes that can remove fluoride

directly fromwater:nanofiltration (NF)and reverseosmosis (RO).NF isoperatedat lower

pressurecomparedtoROandremovesprimarilythelargersuspendedsolids.ROisoperated

at higher pressures with greater rejection of ions. Many researchers have documented

fluorideremovalefficienciesupto98%bymembraneprocesses68-73.ROisusuallyutilisedin

desalination processes to treat saltwater to drinkingwater. RO is the reverse process of

natural osmosis. Pressure is applied to overcome the natural osmotic pressure. The RO

membranerejectsionsbasedonsizeandelectricalcharge.

Beside fluoride, RO and NF can also remove suspended solids, organic and inorganic

pollutants,pesticidesandmicroorganisms,etc. If thewatercharacteristic isalreadyknown

andpre-treatmentisprovided,thelifeofmembranemightbelongercomparedtoaprocess

without pre-treatment. In addition, membranes work at a wide pH range. RO is applied

extensively in desalination industry. However, in general water treatment process, NF is

suggested to bemore suitable over RO for future applications in producing larger water

quantitiesregardingtheenergyconsumption74-76.

However,membraneprocessesrequireskilledpersonneltooperateit.ROmembranes

removeall ionspresent inwater,thoughsomemineralsareessentialforthehumanbody.

Therefore,mineraladjustmentissometimesrequiredaftertreatment,dependingontheraw

water composition and purpose of water treatment. Fluoride-rich concentrate may also

require additional treatment. Membrane application is quite expensive in comparison to

otheroptions.

f

p

f

pf

CC

CCC

R -=-

= 1

GeneralIntroduction

32

1.2.11 Diffusiondialysis

The diffusion dialysis (DD) technique is based on the application of an ion-exchange

membranetotransportionsthroughamembrane.Thetechniqueemploysonlyonekindof

membrane, cation or anion exchange membrane. The driving force in this technique is

concentration difference. It works by supplying a stripping solution on one side of the

membrane, while the other side is supplied with the solution to be treated. In fluoride

removal, anion exchange membrane (AEM-positively charged membrane) and alkaline

stripping solutionareusually employed, as it is shown in Figure1-7.DiffusionofOH- ions

fromthestrippingsolutionthroughtheAEMoccursbecauseofthedifferentconcentration

betweenthetwosolutions.Then,as theOH- ions fromstrippingsolutionpermeateto the

treated solution, electrical potential is generated by the two solutions and act as driving

forcefortheF-ionsinsolutiontopermeatethroughthemembranetothestrippingsolution.

Figure1-7SchematicdrawingillustratingtheprincipleofdiffusiondialysisusingNaOHas

strippingsolutioninastackofanion-exchangemembranes(ownwork)

Ahybridofdiffusiondialysiswithadsorptionwasproposedtoinvestigateitscapability

to eliminate excess fluoride in drinkingwater 77. Though removing fluoride byDD can be

consideredasapotentialtorecover ionicformoffluoridefromtheconcentratedsolution,

strippingsolutionintheoutletmayleadtoanothertreatment78.Thistreatmentisneeded

GeneralIntroduction

33

toavoidsideproblemscausedbyhighfluorideconcentratedstrippingsolution.Theremoval

of fluoride from feedwaterwill be less efficientwhen thedifference in fluoride activities

betweenthetwocompartments istoo low77.Toavoidthebackflowof fluoride iontothe

feeding solution, complexation of fluorides with aluminum is necessary if the receiving

solutioncirculatesinabatchmode79.

1.2.12 Electrodialysis

Electrodialysis(ED)canbedescribedasmembrane-basedtechniqueforionseparation,

which employs concentration difference as driving force for ion transport under the

influenceofanappliedelectricalpotentialdifference.This isdoneinaconfigurationcalled

an electrodialysis cell. The cell consists of a feed (diluate) compartment, a concentrate

(brine) compartmentandanEDUnit.TheEDunit is formedbyconfigurationof staplesof

anionandcationexchangemembranesplacedbetween twoelectrodes.TheEDprocess is

illustratedinFigure1-8.

Figure1-8Schematicdiagramillustratingtheprincipleofdesalinationbyelectrodialysisin

astackwithcation-andanion-exchangemembranesinalternatingseriesbetweentwo

electrodes,whereA=anodeandC=cathode80

The performance of ED as method to reduce fluoride concentration from brackish

water and the effect of process parameters and ionic species for fluoride removal using

electrodialysis were investigated in previous studies 81,82. It was concluded that initial

fluoride concentration and applied potential are the decisive factors for the increase of

fluorideremoval.Incaseofwatersourcesrichinotherionsthanfluoride,itmaycausehigh

GeneralIntroduction

34

concentrationof salt in theconcentratedcompartment.Thisbecomesan important issue,

becauseitcouldgeneratefoulinganddamagetothemembrane82. It isalsoreportedthat

theremovaloffluoridewashigherintheabsenceofmono-andbi-valentions65.

Duetotherelativelyhighoperationcost,thisprocessisnotusedextensively.Theanion

and cation exchangemembrane are considered as expensivemembranes. High operating

costhastobeconsideredifelectrodialysis isemployedforhighrateof ionremovaldueto

intensive use of current. Besides, operators have to be trained for operation and

maintenanceofthesystemappropriately.

1.2.13 Summary

This review on existing techniques to remove fluoride from drinking water is

summarizedinthefollowingtableasanoverviewofalltechniques.

GeneralIntroduction

35

Method Technique Mechanism Simplicity Risk

Coagulationand

precipitation

Nalgonda Fluorideisremovedthrough

coagulationandco-

precipitationofaluminum-

fluoridecomplex

• Establishedtechnique

• Simpleoperation

• Commonchemicalsare

involved(limeandalum)

• Aluminumandsulfate

presenceintheproduct

water

• Turbidityoffineprecipitate

• Formationofhighvolume

ofsludge

• Difficulttocontrolthe

dosageincaseofwater

sourceswithdifferent

fluorideconcentration

Coagulationand

precipitationusing

CaCl2

Fluorideisremovedthrough

precipitationofcalciumfluoride

• Simpletomedium

operation

• Lesssludgecomparedto

lime

• Minimizeriskincaseof

wrongdosageofchemicals

comparedtoNalgonda

• Moresuitableinindustrial

application

• Chemicalsresidue,sludge

formation

• Turbidityduetofine

precipitate

• Longersedimentationtime

• Couldnotmeetstandard

limit1.5mg/LF-,while

higherdosageleadsto

increasedhardness

GeneralIntroduction

36


Contact

precipitation


adsorptionoffluorideiononto

bonecharandco-precipitation

ofcalciumfluorideand

fluorapatite

• Simpleoperation

• Commonchemicals(MSP

andcalciumchloride)

• Bonecharcanbeproduced

usinglocalmaterial(cattle

bone)

• Socio-culturalacceptance

(Cattlebone,pigboneare

unacceptedinseveral

areas)

• Chemicalsresidueneed

specialtreatment

(phosphate,saturated

bonechar)

• Experimentsfailedto

concludeoptimumcontact

time,soremovalcapacity

andefficiencycouldnotbe

determined

Electrocoagulation Fluorideisremovedthrough

precipitationofaluminum-

fluoridecomplex.Aluminumis

generatedthroughasacrificial

electrode

• Advancedoperation

• Aluminumneededcanbe

calculated

• Passivation

• Lowrateofremovalwhen

initialfluoride

concentrationislow

• Highenergyconsumption

• Experimentfailedto

concludereaction,soitis

difficulttocontrolthe

GeneralIntroduction

37


process

• Presenceofsulfateion

decreasestheremoval

efficiency

Adsorption ActivatedAlumina Fluorideisremovedthrough


activatedaluminasurface

• Wellknownandwell

established

• Mediumlevelofsimplicity

• Nochemicalsaddition

• Foulingandclogging

• Backwashsolution

IonExchanger Fluorideisremovedthrough


resin



• Regenerationsolution

• Chemicalsresidue

• LowpH

• Expensiveresin

Bonechar Fluorideisremovedthrough


bonechar

• Simpleoperation

• Bonecharcanbeproduced

usinglocalmaterial(cattle

bone)

• Socio-culturalacceptance

• Chemicalsresidue

BioSandFiltration

(BSF)

Couldnotbedetermined

• Localfiltermediumcanbe

used

• Needpre-treatment

• Lowornoremoval

efficiency

GeneralIntroduction

38


Membrane

Technology

ReverseOsmosis

andNanofiltration


sizeexclusionbytheporous

membrane;different

concentrationfornonporous

membrane(RO)



• Processcanbecontrolled

• Membranefouling

shortensthemembrane

life

• Regularbackwashrequired

• Needspretreatment

• Becausemorethan90%of

ionsarerejected,product

waterneedsfurther

treatmentbefore

consumption(polishing)

DiffusionDialysis Fluorideisremovedfollowing

permeationoffluorideionin

thetreatedwaterthroughAEM

tostrippingsolutionby

differentconcentration


• Fluoriderecoveryis

possible

• Nochemicaladdition

• Membranefouling

shortensmembranelife

• Regenerationsolution

needed

• Chemicalsresidueneedsto

betreated

Electrodialysis Fluorideisremovedfollowing

permeationoffluorideionin

thetreatedwaterthroughAEM

tostrippingsolutionby


• Fluoriderecoveryis

possible

• Membranefouling

• Highenergyconsumption

• Needsintensivepre-

treatmenttoavoid

GeneralIntroduction

39


differentconcentration

influencedbyelectrical

potentialdifference

membranefoulingbecause

ofprecipitation

GeneralIntroduction

40

1.3 References

(1) Wang,Y.;Reardon,E.J.ActivationandRegenerationofaSoilSorbentforDefluoridationofDrinkingWater.AppliedGeochemistry2001,16(5),531–539.

(2) Ayoob,S.;Gupta,A.K.FluorideinDrinkingWater:aReviewontheStatusandStressEffects.2006,36(6),433–487.

(3) Gordon,B.;Mackay,R.;Rehfuess,E.InheritingtheWorld:theAtlasofChildren'sHealthandtheEnvironment;WorldHealthOrganization:Geneva,2004.

(4) Smedley,P.FluorideinNaturalWaters;Selinus,O.,Alloway,B.,Centeno,J.A.,Finkelman,R.B.,Fuge,R.,Lindh,U.,Smedley,P.,Eds.;AcademicPress:Burlington,2005;pp301–329.

(5) WHO.InheritingtheWorld:theAtlasofChildren'sHealthandtheEnvironment;Gordon,B.,Mackay,R.,Rehfuess,E.,Eds.;WorldHealthOrganization:Geneva,2004.

(6) Murray,J.J.AppropriateUseofFluoridesforHumanHealth;WorldHealthOrganization:Geneva,1986.

(7) Lennon,M.;Whelton,H.;O'Mullane,D.;Ekstrand,J.Flouride;WorldHealthOrganization:Geneva,2005.

(8) Selinus,O.;Alloway,B.;Centeno,J.A.;Finkelman,R.B.;Fuge,R.;Lindh,U.;Smedley,P.EssentialsofMedicalGeology:ImpactsoftheNaturalEnvironmentonPublicHealth;AcademicPress:Burlington,2005.

(9) Africano,F.;Bernard,A.AcidAlterationintheFumarolicEnvironmentofUsuVolcano,Hokkaido,Japan.JournalofVolcanologyandGeothermalResearch2000,97(1-4),475–495.

(10) Coetzee,P.P.;Coetzee,L.L.;Puka,R.;Mubenga,S.CharacterisationofSelectedSouthAfricanClaysforDefluoridationofNaturalWaters.2004,29(3),331–338.

(11) Pal,B.PotentialHazardsofNitrateandFluorideinUndergroundWaters.WaterResearch1983,17(3),353–354.

(12) Mekonen,A.;Kumar,P.;Kumar,A.IntegratedBiologicalandPhysiochemicalTreatmentProcessforNitrateandFluorideRemoval.WaterResearch2001,35(13),3127–3136.

(13) WorldHealthOrganization.FluoridesandOralHealth:ReportofaWHOExpertCommitteeonOralHealthStatusandFluorideUse;WorldHealthOrganization:Geneva,1994;Vol.846,pp1–44.

(14) Kloos,H.;Tekle-Haimanot,R.Fluorosis.InTheEcologyofHealthandDiseaseinEthiopia;Kloos,H.,Ed.;WestviewPress,1993;pp455–461.

(15) Heikens,A.;Sumarti,S.;vanBergen,M.;Widianarko,B.;Fokkert,L.;vanLeeuwen,K.;Seinen,W.TheImpactoftheHyperacidIjenCraterLake:RisksofExcessFluoridetoHumanHealth.ScienceofTheTotalEnvironment2005,346(1-3),56–69.

(16) Piñón-Miramontes,M.;Bautista-Margulis,R.G.RemovalofArsenicandFluorideFromDrinkingWaterwithCakeAlumandaPolymericAnionicFlocculent.Fluoride2003,36(2).

(17) Fawell,J.K.;Bailey,K.FluorideinDrinking-Water;WorldHealthOrganization,2006;pp1–144.

(18) Tinanoff,N.;Mueller,B.FluorideContentinMilkandFormulaforInfants.ASDCjournalofdentistryforchildren1978,45(1),53.

(19) Kaseva,M.E.OptimizationofRegeneratedBoneCharforFluorideRemovalinDrinkingWater:aCaseStudyinTanzania.JournalofWaterandHealth2006,4(1),139–147.

GeneralIntroduction

41

(20) Lopez,A.;Mathers,C.;Ezzati,M.;Jamison,D.T.;Murray,C.J.L.GlobalBurdenofDiseaseandRiskFactors;Lopez,A.D.,Mathers,C.D.,Ezzati,M.,Jamison,D.T.,Murray,C.J.L.,Eds.;OxfordUniversityPress:NewYork,2006.

(21) Mulyani,D.;McIntyre,J.CariesInhibitoryEffectofFluoridatedSugarinaTrialinIndonesia.2002,47(4),314–320.

(22) Banuchandra,C.;Selvapathy,P.AHouseholdDefluoridationTechnique;2005.(23) Meenakshi;Maheshwari,R.C.FluorideinDrinkingWaterandItsRemoval.Journalof

HazardousMaterials2006,137(1),456–463.(24) George,S.;Pandit,P.;Gupta,A.B.;Agarwal,M.ModelingandSimulationStudiesfor

Aluminium-FluorideInteractionsinNalgondaDefluoridationProcess.ChemicalProductandProcessModeling2009,4(1),27.

(25) Yang,C.-L.;Dluhy,R.ElectrochemicalGenerationofAluminumSorbentforFluorideAdsorption.JournalofHazardousMaterials2002,94(3),239–252.

(26) Sollo,F.W.;Larson,T.E.;Mueller,H.F.FluorideRemovalFromPotableWaterSupplies;UniversityofIllinois,WaterResourcesCenter,1978;pp1–38.

(27) Stumm,W.;Morgan,J.J.AquaticChemistry:ChemicalEquilibriaandRatesinNaturalWaters;1996.

(28) Jagtap,S.;Yenkie,M.K.;Labhsetwar,N.;Rayalu,S.FluorideinDrinkingWaterandDefluoridationofWater.Chem.Rev.2012,112(4),2454–2466.

(29) Bregnhøj,H.CriticalSustainabilityParametersinDefluoridationofDrinkingWater;Dahi,E.,Nielsen,J.M.,Eds.;Nazreth,1997;pp191–197.

(30) M,J.,V;Mahajan,Y.S.ConsequencesofResidualAluminiumFromAlumCoagulantonHumans.InternationalJournalofResearchinChemistryandEnvironment2011,1(2),22–27.

(31) Rondeau,V.;Jacqmin-Gadda,H.;Commenges,D.;Dartigues,J.-F.Re:AluminuminDrinkingWaterandCognitiveDeclineinElderlySubjects:thePaquidCohort.Am.J.Epidemiol.2001,154(3),288–290.

(32) Flaten,T.P.AluminiumasaRiskFactorinAlzheimer’sDisease,withEmphasisonDrinkingWater.BrainResearchBulletin2001,55(2),187–196.

(33) Rao,S.M.;Mamatha,P.WaterQualityinSustainableWaterManagement.Currentscience2004.

(34) Chang,M.;Liu,J.PrecipitationRemovalofFluorideFromSemiconductorWastewater.J.Environ.Eng.2007,133(4),419–425.

(35) Albertus,J.;Bregnhøj,H.;Kongpun,M.BoneCharQualityandDefluoridationCapacityinContactPrecipitation;Dahi,E.,Rajchagool,S.,Osiriphan,N.,Eds.;ChiangMai,2000;pp61–72.

(36) Johnson,A.;Abbaspour,K.C.;Amini,M.;Bader,H.P.;Berg,M.;Hoehn,E.;Hug,S.;Mosler,H.J.;Mülller,K.;Rosenberg,T.;etal.GeogenicContaminants.EawagNews.Dübendorf,Switzer-landDecember2008,pp16–19.

(37) Dahi,E.ContactPrecipitationforDefluoridationofWater;NewDelhi,1996.(38) Hu,C.-Y.;Lo,S.-L.;Kuan,W.-H.;Lee,Y.-D.TreatmentofHighFluoride-Content

WastewaterbyContinuousElectrocoagulation–FlotationSystemwithBipolarAluminumElectrodes.SeparationandPurificationTechnology2008,60(1),1–5.

(39) Drouiche,N.;Aoudj,S.;Hecini,M.;Ghaffour,N.;Lounici,H.;Mameri,N.StudyontheTreatmentofPhotovoltaicWastewaterUsingElectrocoagulation:FluorideRemovalwithAluminiumElectrodes—CharacteristicsofProducts.JournalofHazardousMaterials2009,169(1–3),65–69.

(40) Ün,U.T.;Koparal,A.S.;Öğütveren,Ü.B.ElectrochemicalProcessfortheTreatment

GeneralIntroduction

42

ofDrinkingWater;Hurghada,Egypt,2009;pp1–9.(41) Hu,C.-Y.;Lo,S.-L.;Kuan,W.-H.SimulationtheKineticsofFluorideRemovalby

Electrocoagulation(EC)ProcessUsingAluminumElectrodes.JournalofHazardousMaterials2007,145(1–2),180–185.

(42) Holt,P.K.;Barton,G.W.;Wark,M.;Mitchell,C.A.AQuantitativeComparisonBetweenChemicalDosingandElectrocoagulation.ColloidsandSurfacesA:PhysicochemicalandEngineeringAspects2002,211(2–3),233–248.

(43) Mameri,N.;Yeddou,A.R.;Lounici,H.;Belhocine,D.;Grib,H.;Bariou,B.DefluoridationofSeptentrionalSaharaWaterofNorthAfricabyElectrocoagulationProcessUsingBipolarAluminiumElectrodes.WaterResearch1998,32(5),1604–1612.

(44) EMAMJOMEH,M.;SIVAKUMAR,M.AnEmpiricalModelforDefluoridationbyBatchMonopolarElectrocoagulation/Flotation(ECF)Process.JournalofHazardousMaterials2006,131(1-3),118–125.

(45) Hu,C.-Y.;Lo,S.-L.;Kuan,W.-H.EffectsofCo-ExistingAnionsonFluorideRemovalinElectrocoagulation(EC)ProcessUsingAluminumElectrodes.WaterResearch2003,37(18),4513–4523.

(46) Chary,K.V.R.;RamanaRao,P.V.;VenkatRao,V.CatalyticFunctionalitiesofNickelSupportedonDifferentPolymorphsofAlumina.2008,9(5),886–893.

(47) LópezValdivieso,A.;ReyesBahena,J.L.;Song,S.;HerreraUrbina,R.TemperatureEffectontheZetaPotentialandFluorideAdsorptionattheΑ-Al2O3/AqueousSolutionInterface.JournalofColloidandInterfaceScience2006,298(1),1–5.

(48) Ku,Y.;Chiou,H.-M.TheAdsorptionofFluorideIonFromAqueousSolutionbyActivatedAlumina.Water,Air,&SoilPollution2002,133(1-4),349–361–361.

(49) Choi,W.-W.;Chen,K.Y.TheRemovalofFluorideFromWatersbyAdsorption.Journal(AmericanWaterWorksAssociation)1979,71(10),562–570.

(50) Srimurali,M.;Karthikeyan,J.ActivatedAlumina:DefluoridationofWaterandHouseholdApplication–aStudy;Alexandria,Egypt,2008;pp153–165.

(51) ReyesBahena,J.L.;RobledoCabrera,A.;LópezValdivieso,A.;HerreraUrbina,R.FluorideAdsorptionOntoΑ-Al2O3andItsEffectontheZetaPotentialattheAlumina–AqueousElectrolyteInterface.SeparationScienceandTechnology2002,37(8),1973–1987.

(52) Clifford,D.A.IonExchangeandInorganicAdsorption.InWaterQualityandTreatment:AHandbookofCommunityWaterSupplies;McGraw-Hill:NewYork,1999.

(53) Fang,L.;Ghimire,K.N.;Kuriyama,M.;Inoue,K.;Makino,K.RemovalofFluorideUsingSomeLanthanum(III)-LoadedAdsorbentswithDifferentFunctionalGroupsandPolymerMatrices.J.Chem.Technol.Biotechnol.2003,78(10),1038–1047.

(54) Wasay,S.A.;Haran,M.J.;Tokunaga,S.AdsorptionofFluoride,Phosphate,andArsenateIonsonLanthanum-ImpregnatedSilicaGel.waterenvironres1996,68(3),295–300.

(55) Wasay,S.A.;Tokunaga,S.;Park,S.-W.RemovalofHazardousAnionsFromAqueousSolutionsbyLa(Lll)-andY(Lll)-LmpregnatedAlumina.SeparationScienceandTechnology1996,31(10),1501–1514.

(56) Meenakshi,S.;Viswanathan,N.IdentificationofSelectiveIon-ExchangeResinforFluorideSorption.JournalofColloidandInterfaceScience2007,308(2),438–450.

(57) Medellin-Castillo,N.A.;Leyva-Ramos,R.;Ocampo-Perez,R.;GarciadelaCruz,R.F.;Aragon-Piña,A.;Martinez-Rosales,J.M.;Guerrero-Coronado,R.M.;Fuentes-Rubio,

GeneralIntroduction

43

L.AdsorptionofFluorideFromWaterSolutiononBoneChar.Ind.Eng.Chem.Res.2007,46(26),9205–9212.

(58) Larsen,M.J.;Pearce,E.I.F.;Jensen,S.J.DefluoridationofWateratHighpHwithUseofBrushite,CalciumHydroxide,andBoneChar.JournalofDentalResearch1993,72(11),1519–1525.

(59) Mwaniki,D.L.FluorideSorptionCharacteristicsofDifferentGradesofBoneCharcoal,BasedonBatchTests.JournalofDentalResearch1992,71(6),1310–1315.

(60) Bhargava,D.S.NomographforDefluoridationofWaterinBatchUsingFishBoneChar;Dahi,E.,Rajchagool,S.,Osiriphan,N.,Eds.;ChiangMai,2000;pp73–79.

(61) Abe,I.;Iwasaki,S.;Tokimoto,T.;Kawasaki,N.;Nakamura,T.;Tanada,S.AdsorptionofFluorideIonsOntoCarbonaceousMaterials.JournalofColloidandInterfaceScience2004,275(1),35–39.

(62) Mjengera,H.;Mkongo,G.AppropriateDeflouridationTechnologyforUseinFlouroticAreasinTanzania;DaresSalaam,2002;pp1–10.

(63) Murphy,H.M.;McBean,E.A.;Farahbakhsh,K.ACriticalEvaluationofTwoPoint-of-UseWaterTreatmentTechnologies:CanTheyProvideWaterThatMeetsWHODrinkingWaterGuidelines?JournalofWaterandHealth2010,8(4),611–620.

(64) ThembaOMahlangu.ASimplifiedCost-EffectiveBiosandFilter(BSFZ)forRemovalofChemicalContaminantsFromWater.J.Chem.Eng.Mater.Sci.2011,2(10),1–12.

(65) Hillman,A.AdaptingtheBiosandFiltertoRemoveFluoride:anInvestigationIntoAlternativeFilterMedia;Hillman,A.,Ed.;2007.

(66) He,D.;Susanto,H.;Ulbricht,M.Photo-IrradiationforPreparation,ModificationandStimulationofPolymericMembranes.ProgressinPolymerScience2009,34(1),62–98.

(67) Baker,R.W.MembraneTechnologyandApplications,2nded.;Wiley,2004;pp1–545.

(68) Kettunen,R.;Keskitalo,P.CombinationofMembraneTechnologyandLimestoneFiltrationtoControlDrinkingWaterQuality.Desalination2000,131(1–3),271–283.

(69) Hilal,N.;Abri,Al,M.;Hinai,Al,H.EnhancedMembranePre-TreatmentProcessesUsingMacromolecularAdsorptionandCoagulationinDesalinationPlants:aReview.SeparationScienceandTechnology2006,41(3),403–453.

(70) Goosen,M.F.A.;Sablani,S.S.;Al-Hinai,H.;Al-Obeidani,S.;Al-Belushi,R.;Jackson,D.FoulingofReverseOsmosisandUltrafiltrationMembranes:aCriticalReview.SeparationScienceandTechnology2005,39(10),2261–2297.

(71) Hu,K.;Dickson,J.M.NanofiltrationMembranePerformanceonFluorideRemovalFromWater.JournalofMembraneScience2006,279(1–2),529–538.

(72) Lhassani,A.;Rumeau,M.;Benjelloun,D.;Pontié,M.SelectiveDemineralizationofWaterbyNanofiltrationApplicationtotheDefluorinationofBrackishWater.WaterResearch2001,35(13),3260–3264.

(73) Dach,H.ComparaisonDesOpérationsDeNanofiltrationEtD’OsmoseInversePourLeDessalementSelectifDesEauxSaumatres:DeL’ÉchelleDuLaboratoireAuPiloteIndustriel,Universitéd'Angers,2008.

(74) Pontié,M.;Diawara,C.;Lhassani,A.;Dach,H.;Rumeau,M.;Buisson,H.;Schrotter,J.C.Chapter2WaterDefluoridationProcesses:aReview.Application:Nanofiltration(NF)forFutureLarge-ScalePilotPlants.InFluorineandtheEnvironment—Agrochemicals,Archaeology,GreenChemistry&Water;Tressaudd,A.,Ed.;FluorineandtheEnvironment—Agrochemicals,Archaeology,GreenChemistry&Water;Elsevier,2006;Vol.2,pp49–80.

GeneralIntroduction

44

(75) Tahaikt,M.;AitHaddou,A.;Habbani,El,R.;Amor,Z.;Elhannouni,F.;Taky,M.;Kharif,M.;Boughriba,A.;Hafsi,M.;Elmidaoui,A.ComparisonofthePerformancesofThreeCommercialMembranesinFluorideRemovalbyNanofiltration.ContinuousOperations.Desalination2008,225(1–3),209–219.

(76) Tahaikt,M.;Habbani,El,R.;AitHaddou,A.;Achary,I.;Amor,Z.;Taky,M.;Alami,A.;Boughriba,A.;Hafsi,M.;Elmidaoui,A.FluorideRemovalFromGroundwaterbyNanofiltration.Desalination2007,212(1-3),46–53.

(77) Garmes,H.;Persin,F.;Sandeaux,J.;Pourcelly,G.;Mountadar,M.DefluoridationofGroundwaterbyaHybridProcessCombiningAdsorptionandDonnanDialysis.Desalination2002,145(1–3),287–291.

(78) Durmaz,F.;Kara,H.;Cengeloglu,Y.;Ersoz,M.FluorideRemovalbyDonnanDialysiswithAnionExchangeMembranes.Desalination2005,177(1–3),51–57.

(79) Wiśniewski,J.;Różańska,A.;Winnicki,T.RemovalofTroublesomeAnionsFromWaterbyMeansofDonnanDialysis.Desalination2005,182(1-3),339–346.

(80) Strathmann,H.Chapter6Ion-ExchangeMembraneProcessesinWaterTreatment.InSustainabilityScienceandEngineering;Isabel,C.E.,Andrea,I.S.,Eds.;SustainabilityScienceandEngineering;Elsevier,2010;Vol.Volume2,pp141–199.

(81) Amor,Z.;Bariou,B.;Mameri,N.;Taky,M.;Nicolas,S.;Elmidaoui,A.FluorideRemovalFromBrackishWaterbyElectrodialysis.Desalination2001,133(3),215–223.

(82) Kabay,N.;Arar,Ö.;Samatya,S.;Yüksel,Ü.;Yüksel,M.SeparationofFluorideFromAqueousSolutionbyElectrodialysis:EffectofProcessParametersandOtherIonicSpecies.JournalofHazardousMaterials2008,153(1–2),107–113.

45

Chapter2 GeneralObjectiveandScopeofStudy

2.1 Generalobjective

The importance to prevent endemic fluorosis cannot be argued. Reducing fluoride

concentrations to the permissible level in drinkingwater is one option to avoid excessive

intakeoffluoridethroughdrinkingwater.However,conventionalwatertreatmentmethods

donotremovefluorideions.

Basedonpreviousworksasitisalreadydiscussedinchapter1,itisconcludedthateach

method has its own advantage and drawback. Therefore, further investigations on

alternativemethods that combine the advantageof eachmethodwhile at the same time

eliminatingitsdrawbackareneeded.Infact,theproblemconcerningfluoride-contaminated

drinking water is muchmore relevant in developing countries. Therefore, the alternative

methods should allow for a simple operation, be affordable and effective in removing

fluoride without causing other problems such as excess of aluminium concentration in

drinkingwaterandhighturbidity.

2.2 Scopeofstudy

This study is focusing on the investigation of several techniques/hybrids for water

defluoridation that have not been proposed before or have not yet been studied

systematically.Aimofthework is toprovide inexpensiveandsaferalternativemethodsto

theexistingones.Toachievetheobjectivesmentionedabove,areviewandassessmentof

fluorideremovalmethodswascarriedoutasabasisforfurtherwork.Eachtechniqueineach

method category is reviewed based on its principle and its advantages as well as

disadvantages. A matrix providing an overview of all techniques is presented as a quick

referenceattheendofChapter1.

Based on the review and assessment, a defluoridation technique via coagulation and

precipitationbymeansofaluminumsulfateand lime is selectedas the firstmethodtobe

investigatedinChapter4.Thoughit isalreadyappliedinseveralendemicareasbecauseof

its simplicity in operation, especially in India and Tanzania known asNalgonda technique,

GeneralObjectiveandScope

46

thedrawbackof anexcessivealuminumcontent in thedrinkingwater that leads toother

healthriskshasnotbeeneliminated.Meanwhile,althoughthisprocessisalreadyknownfor

alongtime,adetailedstudyoftheprocesschemistryisrarelyfound.Thesefactsleadtothe

limitation with regard to its development for a combination with other methods. The

experiments were carried out in laboratory scale by using spiked water with fluoride

concentrationadjustment.Theprimaryresearchinterestswereintheeffectsofoperational

factorssuchassolutionpH,aluminumtoinitialfluorideratioandpresenceofotherionson

fluoride removal from aqueous solutions. The fluoride species distribution diagrams of

fluorideinaqueoussolutionswerealsoestablished.Theoptimumprocessparametersfrom

Chapter 4 are then used as the standard parameters for the development of proposed

hybridmethodsinChapter5.

Investigation of the hybrid process of coagulation and co-precipitation and sand

filtration,withemphasisonthepossibilityfornotonlyremovingfluoridebutalsoremoving

aluminum residue is carried out in as the first method in Chapter 5. This technique is

proposed to exploit the advantage of a combination of coagulation-precipitation with

conventional filtration. Since a filtration step is not included in the general Nalgonda

domestic unit, filtration using low cost sand filter is proposed to eliminate turbidity and

excessaluminumintheproductwateraftercoagulationandprecipitationstep.Thoughthis

combination issimilartotheclassicalmethodforwatertreatment inmostwaterworks, its

development for removingexcessive fluorideconcentrationhasnotyetbeen investigated.

Performanceof thehybridmethod is investigatedbystudying its removalmechanismand

interactionbetweenaluminum-fluoridecomplexesandfiltermedia.

ThesecondmethodproposedinChapter5 isanotherhybriddefluoridationtechnique,

namely coagulation andmembrane ultrafiltration. This hybrid is developed based on the

finding and development from the first method and can improve the filtration step by

replacingtheconventionalsandfilterwithamembranefilter.Ultrafiltrationmembranesare

expectednotonlytobesufficienttoremovetheexcessaluminuminproductwaterbutalso

to enhance the fluoride removal. The study is focussing on the investigation of the

interactionofwater and themembrane itself. This study is very importantbecause itwill

definethereliabilityofthishybridtobeoperatedforlong-termfluoridewatertreatment.

GeneralObjectiveandScope

47

Finally,somegeneralconcludingremarksarepresentedinChapter6.Theoutlookand

improvementintheexistingandhybridtechniquesaswellasanalyticalmethodforanalyzing

fluoride isproposed.The followingFigure2-1 illustrates the flowandconnectionbetween

thechapters.

Figure2-1Illustrationandtherelationshipsbetweenthechapterinthestudy

48

Chapter3 AnalyticalInstruments

3.1 Ionselectiveelectrodes(ISE)

Mostof the fluoridedata in this studywereanalysedusingan ion selectiveelectrode

(ISEMETROHM,781pH/IonMeter).An ISE isused fordirectdeterminationof fluoride ion

concentration in water samples. ISE is a potentiometric sensor based on an ion selective

membrane. The measured potential refers to the activity of the ion that has to be

determinedinthesample.

The electrode in ISE uses a crystal of lanthanum fluoride (LaF3) as the selective

membrane,wherefluorideionswillbemobile.Thecrystalisdopedwitheuropiumfluoride

(EuF2)toincreaseitsconductivity.Whentheelectrodeisimmersedinasolutioncontaining

fluorideions,apotentialdifferencedevelopsacrossthemembraneandismeasuredagainst

aconstantreferencepotentialwithastandardpH/mVmeter.TheNernstequationdescribes

themeasuredpotentialasafunctionoftheactivityoffluorideionsinsolution.

! = !# − !% +'()* ln -

where:

! =measuredelectrodepotentialinV

!# =referencepotential(aconstant)inV!% =junctionpotentialinV

' =8.3145J/mol.K

( =temperatureinK

) =charge(-1forfluoride)

* =Faraday’sconstant=96,485C/mol

- =activityoffluorideionsinsolution

The ISE responds to the fluoride ion activity, not to the fluoride concentration.

According to Nernst’s equation, the measurement of equilibrium potential indicates the

activityoffluorideions.Theelectroderespondsonlytofreeions1. It is importanttoavoid

AnalyticalInstruments

49

the formation of complexes that are to be measured because complexation, e.g with

aluminium,willdecreasetheactivity.Bufferingthesamplehelpstomaintainaconstanttotal

ionicstrength.Constanttotalionicstrengthpreventsafluctuationintheactivitycoefficient

ofthe ionbeingmeasuredbecause ionshavedifferentactivityatdifferent ionicstrengths,

e.g., the activity of F- at an ionic strength of 0.001, 0.01, 0.1 is 0.975, 0.926, 0.810

respectively2.ForfluorideanalysisaTotalIonicStrengthAdjustmentBuffer(TISAB)isused.

TISABdoesnotonlymaintainthe ionicstrengthbutalsoadjustsandbuffersthepHvalue.

ThesolutionpHismaintainedat5whereF-isthepredominantspecies,atlowerpHHF(pKa

= 3.14) will be dominant. TISAB is usually prepared based on German Institute for

Standardization No. 38405-4 (DIN) by dissolving 300 g tri-sodium citrate dihydrate

(C6H5Na3O7.2H2O); 22 g Titriplex IV or 1,2-cyclohexylenedinitrilotetraacetic acid

monohydrate (C14H22N2O8.2H2O) and 60 g sodium chloride (NaCl) in 1000 mL deionized

water3.

In order to determine the relation between measured electrical potential and

concentrationoffluorideions,acalibrationisdonebyvaryingthefluorideconcentrationsin

standard solutions and measuring the corresponding electrical potential. The standard

solutionswerepreparedof fivedifferent concentrations, 0.2,0.5,1,5, and10mg/L. The

calibrationcurveisillustratedinFigure3-1.


50

Figure3-1Typicalcalibrationcurveofionselectiveelectrode(METROHM781ph/Ion

meter)1

3.2 Ionchromatography(IC)

In this research, ion chromatography (IC-DX500, Dionex with Automated Sampler AS

3500) was used to analyse water samples in order to determine fluoride, chloride and

sulfate.TheanalysiswasdonebyIWWWaterCentreGermany.ICisananalyticaltechnique

that uses theprinciple of ion exchange for analyzing anions and cations. It is one typeof

liquidchromatographythatusesanionexchangeresinasstationaryphase.Thedetectorin

ICwillrespondanddetecttheconductivityofthesamples.

Thefirststepofionchromatographyistheintroductionofsampleintoamobilephase.

Thismixturepasses intoacolumnthat isuniformlypackedwithresinfixedwithfunctional

group of an opposite charge to the analyte. Fluoride analysis in IC uses a column that

containspositivelychargedactivesites (anionexchanger).Themost frequentlyusedanion

exchangerisbasedonalkylchainswithammonium-basedfunctionalgroups.Thetwomost


51

important functional groups in anion chromatography are derived from trimethylamine

(TMA)and2-dimethylaminoethanol,(DMEA).Themobilephase,oreluent,ismadeupofan

aqueoussolutionofsalts,e.g.,carbonateorbicarbonatesalt.Figure3-2showstheset-upof

stypicalICsystem.

Figure3-2TypicalillustrationofICsystemwithsuppressor4

A suppressor is technically an anion or cation membrane used to minimize the

contribution of mobile phase ions to conductivity by selectively removing mobile-phase

electrolyteionswithoutremovingsoluteions.Thecolumnisplacedbetweentheanalytical

column and the detector. With this suppressor, the IC detector will show only strong

conductivity of the solute ion. Figure 3-3 describes typical suppressor operation for anion

exchangewithNa2CO3/NaHCO3aseluentandcationexchangewithHClaseluent.

Chapter 3: Ion chromatography

Figure 3.1. Typical IC analysis system with conductometric detection.

Figure 3.2. Principle of the Dionex EG40 anion eluent generator.

Various types of pumps are used in IC. They include single and dual piston pumps

and isocratic and gradient pumps, e.g., quaternary gradient pumps. The purpose of the

pump is to maintain the eluent flow through the IC system and, in the case of gradient

52


52

Figure3-3Schematicoperationofasuppressorfor(a)anion-exchangeand(b)cation-

exchangeseparations5

3.3 Inductivelycoupledplasma–opticalemissionspectroscopy(ICP-OES)

The aluminium analysis for the water samples was mostly conducted also by IWW

WaterCentreGermany,using InductivelyCoupledPlasma–OpticalEmissionSpectroscopy

(ICP-OESVarianVistaProwithaxialPlasmaandAutosamplerSPS3).ICP-OESisananalytical

technique for determining a wide range of elements based on atomic emission. Atomic

emission occurs when a valence electron in a higher-energy atomic orbital returns to a

lower-energyatomicorbital.Becauseeveryelementhasdifferentenergylevels,theemitted

wavelengthisspecificfortherespectiveion.

ThebasicoperationofICPisbasedonapartiallyionizedgas,typicallyArgon(Ar),which

islessthan1%ionizedintheplasma.Theplasmaitselfisgeneratedinaquartztorchusinga

1-2.5-kWradiofrequencypowersupply.Samplesaretypicallyintroducedintothecenterof

theplasmaasaerosols6.

Theatomsandionspresentintheplasmaarethermallyexcited.Whenthoseatomsand

ions are returning to the ground state, photons are emitted. The emitted radiation is

spectrally dispersed in a spectrometer and the emission of individual wavelengths is

measuredwithasuitabledetector.Ascanningmonochromatorcanbeprogrammedtomove

the monochromator rapidly to an analyte’s desired wavelength, pausing to record its

P.R. Haddad et al. / J. Chromatogr. A 1000 (2003) 725–742 731

scavenger) solution passes over the exterior of the support, altering the shape of the packing beads andfibre, usually in a countercurrent direction. The first application of an ultrasonic field to the system. Thehollow-fibre suppressor was reported by Stevens et effects of these approaches have been discussed inal. [9] and consisted of a collection of sulfonated some detail by Dasgupta [13].cation-exchange fibres, with which sulfuric acid was The regenerant employed with fibre suppressorsused as the regenerant. The mode of operation of this must supply the ion required for effective eluent

1 2suppressor with a bicarbonate /carbonate eluent is suppression (e.g. H or OH ), but must not contami-illustrated schematically in Fig. 1. The overall results nate the eluent with any other ion. The chiefof these processes are identical to those achieved by potential contaminant is the regenerant ion havingthe column suppressor, but the hollow-fibre design the same charge sign as that of the solute. Whilst thishad the chief advantages of greatly reduced band- ion is theoretically prevented from entering thebroadening and continuous regeneration [10,11]. It eluent stream as a result of Donnan exclusion by thehas also been noted that suppression efficiency is ion-exchange functionality on the fibre, this repulsiveincreased at elevated temperatures because of im- effect may not totally prevent penetration of theproved diffusion of ions both in solution and through forbidden ion, especially when the regenerant con-the membrane [10,12]. centration is high. For this reason it was common to

Transfer of ions through the fibre was found to be use regenerants containing large co-ions, such asenhanced if some type of inert packing was inserted dodecylbenzenesulfonate.into the fibre. Typical packings were nylon filamentsor polystyrene beads and these served to also de- 4 .3. Micromembrane suppressorcrease the dead volume inside the suppressor. Suchpacked-fibre suppressors were developed extensively The chief drawback of fibre suppressors was thatusing such approaches as replacing the inert beads the small internal diameter of the fibre meant that thewith ion-exchange resin beads, packing beads around surface area of the fibre was low, and this, in turn,the exterior of the fibre to provide mechanical led to low ion-exchange and thereby low suppression

Fig. 1. Schematic operation of a hollow fibre suppressor for (a) anion-exchange and (b) cation-exchange separations.


53

emissionintensitybeforemovingtothenextanalyte’swavelength.Theconcentrationofthe

elements is determined by comparing emission of sample with the emission of external

calibrationstandards.

In this research, ICP-OESwas used to analyse aluminium, calcium andmagnesium in

watersamples.Beforetheanalysis,watersampleswerepre-treatedwithacid7.

3.4 Colorimetry(Filterphotometry)

In comparison to ICP-OESmeasurements, aluminiumconcentrations inwater samples

werealsoanalysedusingtheHach-LangeSchnelltestcolorimeterDR/890method.Everyset

of SchnelltestHach-Lange consists of ascorbic acidpowder,AluVer 3 reagentpowder and

bleachingpowder.Colorimetryisaphotometricanalysistechniqueusingacontinuoussingle

white light source of which light is passed through a filter for alternative wavelength

selection, usually between 400 – 800 nm. When incident light passes through a cuvette

containing solution, the intensity of the light leaving the sample will be less than the

intensity of light entering the cuvette. In this method, the loss of light is due to the

absorptionby thecompound.Figure3-4 illustrates thecolorimetryprinciple.Bouguerand

Lambertstatedthatabsorptionisdependentproportionallyonthethicknessofthesamples

iftheconcentrationisheldconstant.Wherek1isproportionalitycoefficientanddissample

thickness,absorbancecanbeformulatedasfollows(Lambert-Bouguerlaw):

A=k1d

Beerdeterminedtheproportionallyoftheabsorbancewithconstantlayerthicknessto

theconcentrationoftheabsorbingmaterial.ThisisknownasBeerlaw:

A=k2c

where k2 is a proportionality coefficient and c is the concentration of absorbingmaterial.

Combination of these two relations later became known as Lambert-Beer law. The

expressioncanbewrittenasfollows:

A=ε(λ)cd

where ε(λ) is the molar absorption coefficient at λ wavelength (in L/mol cm), c is the

concentration of absorbing material (in mol/L) and d is the sample thickness (in cm) 8.


54

Absorbancecanbealsoexpressedastransmittanceasafunctionoflightintensity.IfI0isthe

incidentlightintensityenteringthecuvetteandIistheoneleavingthecuvettethefollowing

relationapplies9:

. = log 1( = −log( = −log 22#

Theinstrumentiscalibratedto0%transmittance(T)usingashuttertoblockthesource

radiation from the detector. After removing the shutter, the instrument is calibrated to

100%Tusinganappropriateblank.Sincethesource’sincidentpowerandthesensitivityof

thedetectorvarywithwavelength,thephotometermustberecalibratedwheneverthefilter

ischanged.

Figure3-4Basicprincipleofcolorimetry9

Because each compound in the visible wavelength range in solution has a typical

absorption spectrum, the compound in solution should show predominant colour. For

colorimetricanalysisofaluminiuminthisresearch,Chromazurol-S(C23H13Cl2O9SNa3)(Figure

3-5) isused to forma coloured complexwithaluminium 10,11. This reagentwill show light

pink to dark purple colourwhen it forms a complexwith aluminium. The complex has its

maximumabsorbancenear540nm11.


55

Figure3-5Chromazurol-Sor3-sulfo-2,6-dichloro-3,3-dimethyl-4-hydroxyfuchson-5,5-

dicarboxylicacid11.

Anothercommonreagentforaluminiumdeterminationisaluminon(C22H23N3O9)(Figure

3-6).Aluminonshowslightorangetoredcolourascomplexwithaluminiumwithmaximum

absorbancenear530nm12.

Figure3-6Aluminonor5-[(3-carboxy-4-hydroxyphenyl)(3-carboxy-4-oxo-2,5-

cyclohexadien-ylidene)methyl]-2-hydroxybenzoicacidtriammoniumsalt13


56

3.5 Turbidityanalysis

ANephelometer or turbidimeterwas used to analyze the correlationof turbidity and

thefluorideremoval.Nephelometrymeasurestheamountoflightscatteredat90°anglesto

anincidentlightbeambyparticlespresentinafluidsample.Onlylightthatpassesthrough

absolutelypurewater,hasrelativelyundisturbedpaths. Ina fluidthatcontainssuspended

solids, light isscatteredbyparticlespresent.Figure3-7showstheschematicprincipleofa

turbidimeter.

Figure3-7Basicprincipleofnephelometry14

The pattern of interaction between the sample and light transmitted in a sample

containingsuspendedsoliddependsonseveralfactors:size,shapeandcompositionofthe

particles in the solution and to the wavelength of the incident light. The relationship

betweenthemcanbedescribedbythefollowingequation:

Is=ksI0C

where:

I =Intensityofscatteredlight

I0 =Intensityofsourcelight

ks = System constant (has to be determinedwith a calibration curve prepared by usingknownstandardconcentrationofparticles)

C =Solutionconcentrationscatteringparticles

7

StablCal™ Stabilized Formazin Turbidity StandardsA relatively new turbidity standard has been developedfor use in calibrating or verifying the performance of anyturbidimeter. StablCal™ Turbidity Standards contain thesame light scattering polymer as traditional formazinprimary turbidity standards. By using a different matrix,the formazin polymer in StablCal™ Standards is stabilized,and will not deteriorate over time as is the case withtraditional low turbidity formazin standards. Due to thisenhanced stability, StablCal™ Standards of any concen-tration ranging up to 4000 NTU can be manufacturedand packaged in ready-to-use formats.

StablCal™ Turbidity Standards have many advantages over traditional formazin and other secondary turbiditystandards. First, StablCal™ Standards are stable for aminimum of two years. Figure 5 (p. 8) displays thestability of StablCal™ Standards of three differentconcentrations — 2.0, 10.0, and 20.0 NTU. The stabilityof these standards is independent of concentration.Second, StablCal™ Standards are prepared at specificconcentrations, eliminating the tedious and technique-sensitive preparation through volumetric dilutions.Third, StablCal™ Standards have the same particle sizedistribution as formazin and they can be directlysubstituted for formazin. Thus a StablCal™ Standard has a defined concentration that is independent of anyinstrumentation. Figure 6 (p. 8) demonstrates thiscomparable performance of the StablCal™ Standards to traditional formazin standards in the 1 to 5 NTU range on a wide array of turbidimeters. Last, StablCal™Standards can be repeatably prepared from traceable raw materials, and can be considered primary standards.

The nature of the matrix of StablCal™ Standards has also helped to reduce the potential health risks that are associated with traditional formazin standards.Components in this matrix effectively scavenge any tracehydrazine from the standard. The hydrazine concentrationis reduced to levels that are below analytical detectionlimits. Hydrazine levels in StablCal™ Standards havebeen reduced by at least three orders of magnitude overthose in traditional formazin standards of equal turbidity.

Since the StablCal™ Standards are pre-made, the onlyuser preparation required is to thoroughly mix thestandards before use. This eliminates exposure to thestandard, reduces potential to contaminate the standard,and saves time that would otherwise be spent inpreparing these standards by volumetric dilution.

NephelometryHistorically, the need for precise measurements of verylow turbidity in samples containing fine solids demandedadvancements in turbidimeter performance. TheJackson Candle Turbidimeter presented serious practicallimitations because it could not measure turbidity lowerthan 25 JTU, was somewhat cumbersome, and wasdependent on human judgment to determine the exact

extinction point. In addition, because the light source in the Jackson instrument was a candle flame, incidentlight emitted was in the longer wavelength end of thevisible spectrum (yellow-red) where wavelengths are notscattered as effectively by small particles. For this reason,the instrument was not sensitive to very fine particlesuspensions. (Very fine silica will not produce a flameimage extinction in a Jackson Candle Turbidimeter.) TheJackson Candle Turbidimeter was also incapable ofmeasuring turbidity due to black particles such as charcoalbecause light absorption was so much greater than lightscattering that the field of view became dark beforeenough sample could be poured into the tube to reach animage extinction point.

Several visual extinction turbidimeters were developedwith improved light sources and comparison techniques,but human judgment errors contributed to a lack of preci-sion. Photoelectric detectors, sensitive to very smallchanges in light intensity, became popular to measure theattenuation of transmitted light through a fixed-volumesample. The instruments provided much better precisionunder certain conditions, but still were limited in theirability to measure high or extremely low turbidity. Atlow scattering intensities, the change in transmitted light,viewed from a coincident view, was so small that it isvirtually undetectable by any means. Typically, the signalwas lost in the electronic noise. At higher concentrations,multiple scattering interfered with direct scattering.

The solution to this problem was to measure the lightscattered at an angle to the incident light beam and thenrelate this angle-scattered light to the sample’s actualturbidity. A detection angle of 90° is considered to bevery sensitive to particle scatter. Most modern instrumentsmeasure 90° scatter (Figure 4); these instruments arecalled nephelometers, or nephelometric turbidimeters,to distinguish them from generic turbidimeters, whichmeasure the ratio of transmitted to absorbed light.

Figure 4. In nephelometric measurement, turbidity isdetermined by the light scattered at an angle of 90°from the incident beam.

GlassSample Cell

TransmittedLight

90° ScatteredLight

Detector

Aperture

Lamp

Lens


57

3.6 Elementquantification

EnergyDispersiveX-rayspectroscopy(EDX)wasusedtoquantifyelementsretainedon

thesurfaceofthemembranethatwasusedtoenhancefluorideandaluminiumremovalin

thisstudy.TheanalysesweredonebyCentralLaboratoryforScanningElectronMicroscopy,

UniversityDuisburg-Essen.EDXspectroscopy isananalytical techniqueusedforqualitative

and quantitative elemental analysis of solids and liquids that need only small amount of

sample.ThistechniqueiscategorizedasX-rayfluorescence(XRF)spectroscopy.

Thefundamentalprincipleofthistechniqueisbasedonthefactthateachelementhas

different X-ray characteristics.When an incident photonwith high-energy radiation hits a

sample, inner shell electrons of elements in the sample are excited and ejected from the

shell. Following the electron ejection, an electron hole is created in the placewhere the

electronescaped.Then,anelectronfromanoutershellthathashigherenergyfillsthehole.

Theenergydifferencebetween thehigherenergy shell and the lowerenergy shellwill be

emittedasX-ray radiation.Because theenergydifferencebetween the two shellsofeach

atomic structure is different, the energy of the X-ray radiation is characteristic for each

element. Therefore, the elemental composition of the sample can be measured. As

electromagneticspectrum,characteristicofX-raycanbedescribeasthefollowingequation:

! = ℎ. 56

where:

ℎ =Planck’sconstant,6.6254x10-34J.s

5 =Velocityofpropagationoflight,3.00x108m/sinvacuum

while the wavelength of radiation depends on the atomic number (Z) of the particular

elementasdescribeinthefollowingequation:

16 = 7() − 9);

where:

7 ≈ '(Rydbergconstant,10973731,43m-1)15

9~1 (screeningcoefficient)


58

InEDX,theemittedradiationisdispersedaccordingtoitsenergyusingasemiconductor

detector.TheX-rayspectrumacquiredduringtheprocessshowsanumberofcharacteristic

peaks. The energy of the peaks leads to the identification of the elements present in the

sample (qualitative analysis), while the peak intensity provides absolute elemental

concentration(quantitativeanalysis).

3.7 References

(1) Chiba,K.;Tsunoda,K.;Haraguchi,H.;Fuwa,K.DeterminationofFluorineinUrineandBloodSerumbyAluminumMonofluorideMolecularAbsorptionSpectrometryandwithaFluorideIon-SelectiveElectrode.AnalyticalChemistry1980,52(11),1582–1585.

(2) Kielland,J.IndividualActivityCoefficientsofIonsinAqueousSolutions.JournaloftheAmericanChemicalSociety1937,59(9),1675–1678.

(3) DIN.DIN38405-4:DeutscheEinheitsverfahrenZurWasser-,Abwasser-UndSchlammuntersuchung;Anionen(GruppeD);BestimmungVonFluorid(D4);BeuthVerlag,1985.

(4) Wang,W.InorganicandOrganicSpeciationofAtmosphericAerosolsbyIonChromatographyandAerosolChemicalMassClosure,GhentUniversity.FacultyofSciences:Ghent,Belgium,2010,pp1–395.

(5) Haddad,P.R.;Jackson,P.E.;Shaw,M.J.DevelopmentsinSuppressorTechnologyforInorganicIonAnalysisbyIonChromatographyUsingConductivityDetection.JournalofChromatographyA2003,1000(1-2),725–742.

(6) Olesik,J.W.ElementalAnalysisUsingICP-OESandICP/MS.AnalyticalChemistry1991,63(1),12A–21A.

(7) DIN.DINenISO11885:Wasserbeschaffenheit-BestimmungVonAusgewähltenElementenDurchInduktivGekoppeltePlasma-Atom-Emissionsspektrometrie(ICP-OES)(ISO11885:2007);DeutscheFassungenISO11885:2009;BeuthVerlag,2009.

(8) HandbookofInstrumentalTechniquesforAnalyticalChemistry;Settle,F.A.,Ed.;2005;pp1–728.

(9) Schwedt,G.TheEssentialGuidetoAnalyticalChemistry(Schwedt,G.);Wiley,1997.(10) Bhargava,O.P.;Hines,W.G.ChromazurolSSpectrophotometricDeterminationof

AluminainIronOres,Sinters,andOpenHearthSlags.AnalyticalChemistry1968,40(2),413–415.

(11) Amelin,V.G.;Chernova,O.B.FeaturesofTestReactionsofMetalIonswithChromazurolSImmobilizedonThin-LayerMatrices.JAnalChem2008,63(8),799–804.

(12) Clark,R.A.;Krueger,G.L.Aluminon:ItsLimitedApplicationasaReagentfortheDetectionofAluminumSpecies.JournalofHistochemistry&Cytochemistry1985,33(7),729–732.

(13) AmericanChemicalSociety.CommitteeonAnalyticalReagents.ReagentChemicals:SpecificationsandProcedures,10ed.;OxfordUniversityPress,2006.

(14) Sadar,M.J.TURBIDITYSCIENCE,TechnicalInformationSeries—BookletNo.11;HachCompany,1998.

(15) Hänsch,T.W.;Nayfeh,M.H.;Lee,S.A.;Curry,S.M.;Shahin,I.S.PrecisionMeasurementoftheRydbergConstantbyLaserSaturationSpectroscopyofthe


59

BalmerΑLineinHydrogenandDeuterium.Phys.Rev.Lett.1974,32(24),1336–1340.

60

Chapter4 CoagulationandCo-precipitationforFluorideRemoval

usingAluminumSulfateasCoagulant

4.1 Background

Aluminumsaltsarecommoncoagulantsindrinkingwatertreatment.Thereareseveral

aluminumcompoundsusedforthatpurposeincludingaluminumsulfate,aluminumchloride,

polyaluminum chloride (PAC), and sodium aluminate. Aluminum sulfate and lime addition

wasproposedforfluorideremovalindrinkingwaterintheUSAwhenfluoridewassuspected

of causing teeth mottling 1. As already discussed in chapter 3.2.1, because fluoride

concentration in thesourceofdrinkingwater ishigh, thedosageofcoagulant for fluoride

removalishigherthanincommondrinkingwatertreatment.Laterthisprocesswasadopted

by National Environmental Engineering Research Institute (NEERI) as the Nalgonda

techniqueanddevelopedforlowcosttreatmentusedatthecommunityandhouseholdlevel

in India. This technique is claimed as easy to be handled and able to remove fluoride in

contaminatedwaterupto90%2-4.

The Nalgonda technique uses alum, lime and bleaching agent to be added to source

water with elevated fluoride concentration. It was suggested that addition of lime is

required tomaintain thepH 5. Thealum, in this casehydratedaluminumsalts, is usedas

coagulantandableachingagentsuchascalciumhypochloritepowderasdisinfectant.Based

ontheguidelineofNEERIthedosagesofalumarevaryingfrom145to1600mg/L(16to181

mg/LasAl)fortreatingrawwaterwithfluoridelevelsof2to8mg/Latvaryingalkalinity6.

Becauseexcessivedosageofaluminumisinvolvedinthisprocess,thepHoftheprocesshas

tobecontrolledtoavoidelevatedaluminumconcentrationinthewaterproduct.

Although it has beenwidely applied, the removal process in the Nalgonda technique

itself isnotfullyunderstoodbecauseofitscomplexity3,7.Thisstudywasdesignedtofoster

the understanding of the aqueous chemistry for fluoride removal by coagulation and co-

precipitation with aluminum sulfate solution as coagulant. The study included the

investigationofreactiontimeoffluorideremoval,theeffectofaluminumtofluoridemolar

CoagulationandCo-precipitation

61

ratio, the effect of pH, the effect of the addition of Ca(OH)2 and NaOH to the fluoride

removal.

4.2 Theoreticalconsiderations

Aluminumasdestabilizingagent

As coagulant, aluminum salt is a destabilization agent. The destabilization process

usuallyoccursintwoways,namelychargeneutralizationandsweepflocculation.Bycharge

neutralization, positively charged aluminum species from hydrolysis products neutralise

negatively charged particles followed by the aggregation of destabilized particles and

adsorptionofparticle. In sweep flocculation, the formationof aluminumhydroxide is the

mainfactor,wherethegrowingaluminumhydroxideprecipitateenmeshedtheparticlesand

colloidal contaminant in the solution. For this reason, excessive dosage of coagulant is

usuallyneeded.

Aluminumcomplexandco-precipitationoffluoride

The aluminum ions form a hexaquo complex with water (Al(H2O)63+), which is often

abbreviatedasAl3+.Thewater inthehexaquocomplex(hydratedaluminum)mayundergo

exchangereactionswithligandsthatcouldbeanionsorneutralmoleculesinbulksolution.

Hence,insuchacomplextheanionsandneutralmoleculesarebounddirectlytoaluminum

via charge transfer without the presence of water molecule between them. This

complexationmechanism isknownas innerspherecomplex.Ontheotherhand,anouter

spherecomplex is formedduetoelectrostatic interactionbetweenthealuminumandthe

ligandwiththepresenceofwatermoleculebetweenthem.

Previous studies suggest that the removal of fluoride by aluminum coagulant is a co-

precipitationprocess rather thanprecipitation 8. In thecaseofco-precipitation, fluoride is

adsorbedonthesurfaceofAl(OH)3precipitate.Otherstudyassumedthataluminum-fluoride

complexes in the aqueous medium are transformed to form dissolved, colloidal and

precipitated formsdependingonthesolutionconditionssuchasdosageofaluminumsalt,

pHandfluorideinitialconcentration6.Therefore,thesefactorsplayanimportantroleinthe

fluorideremovalbutarenotindependentofeachother.


62

pHofpointofzerocharge

ThepHofpointofzerocharge(pHPZC)isanimportantparameterinfluoridesorption.If

thepH isequaltothepHPZCthesorbenthasanoverallneutralcharge,abovethepHPZCthe

sorbentbecomesnegatively charged,andbelow thepHPZC the sorbentbecomespositively

charged.StudieshaveshownthatwhenpH is increasedabovethepHPZC, fluoridesorption

decreasesduetoelectrostaticrepulsionbetweenthesurfaceandfluorideanionsaswellasa

result of competition with hydroxide ions in solution 9-11. Reported pHPZC values for

aluminumhydroxideareintherangefrom6.2to9.6dependingupontype,treatment,and

hydrationofthesorbent12.

EffectofpHonaluminumsaltinaqueoussolution

Most of the dissolved aluminum is supposed to be present as one of the hydroxo

complexes innaturalwater 13.Aluminum iseasilyhydrolyzed inwateratpHabove3.5 14.

When the pH of solution that contains Al3+ is raised, dissolved species of Al3+ and (OH)-

complexare formed. Incaseofwatercontainingahighconcentrationofaluminum,atpH

above 6.3 (Ks0=31.5) Al(OH)3 may start to precipitate. Hydrolysis of aluminum ions is

representedassequenceofreplacementofthewatermoleculesbyhydroxyl ionsthatcan

bedescribedasfollows:Al3+↔Al(OH)2+↔Al(OH)2+↔Al(OH)3↔Al(OH)4-.

Whenthefluorideconcentrationinwaterishigh,aluminum-fluoridecomplexeswillbe

formed.Watermolecules are replaced by fluoride ions as follows: Al3+, AlF2+, AlF2+, AlF30,

AlF4-, AlF52- and AlF63-. In aqueous environments there are various ligands beside F-, e.g.,

SO42-,NOM,PO4

3-thatcouldaffectperformanceofaluminumionsascoagulant.Theycould

substitute stepwise thewatermolecules to formboth solubleand insolubleproducts that

willaffectthedoseofthecoagulant15.

By using the ChemEQL V3.0 program the aluminum-hydroxyl-fluoride species

distributionwascalculated.Withtheinputof3.7M(or100mg/L)aluminumand0.53M(or

10mg/L)fluorideaspeciationdiagramasfunctionofpHcanbegeneratedthatisshownin

Figure4-1.Both values are insertedas total concentration in theprogramcalculation, i.e.

absence of precipitation is assumed. Then, the program calculates each species

concentration as well as the ionic strength of each system at different pH. Finally, these

valuesareusedinmanualcalculationtoobtainthespeciesfraction.Complexformationof


63

aluminum ions with hydroxyl ions and fluoride in aqueous solution is described by the

followingformationconstants:

Al3++OH-⇌Al(OH)2+logKf=9.00

Al3++2OH-⇌Al(OH)2+logKf=17.9

Al3++3OH-⇌Al(OH)3logKf=25.2

Al3++4OH-⇌Al(OH)4−logKf=33.3

Al3++F-⇌AlF2+logKf=7.00

Al3++2F-⇌AlF2+logKf=12.6

Al3++3F-⇌AlF3logKf=16.7

Al3++4F-⇌AlF4−logKf=19.4




64

Figure4-1SpeciesdistributionofaluminuminsolutionasfunctionofpHwiththepresence

of[F-]=10mg/Land[Al3+]=100mg/L

EffectofCa(OH)2onfluorideremoval

Ca(OH)2 is usually used to treat fluoride containing wastewater in the industrial sector

because it isconsideredthatproducingCaF2 isthe leastexpensivewaytoremovefluoride

fromwastewater.Thisisbecausefluoridehashigheraffinitytocalciumwithbondingenergy

of 527 kJ/mol compared to other ions, for example to magnesium with 462 kJ/mol 16.

FluoridereactswithcalciumformingCaF2withasolubilityof0.016g/Linwater.Basedona

solubility calculation, it was suggested that in solution where both ions are present in

stoichiometricamounts, the fluorideconcentrationwill remainat8.18mg/L insolution 17.

Howeverlowerfluorideconcentrationmightbeachievedincaseofexcesscalciumamount

duetocommonioneffect17.


65

4.3 Experimental

4.3.1 Materials

Table4-1Listofmaterials

Chemicals Suppliers Purities

NaF Merck >99.5%

Al2(SO4)3.18H2O Merck 51%-59%

Ca(OH)2 IWW 88.5%

Sodiumcitrate Merck 99.0%-101.0%

TitriplexIV Merck >99.0%

NaCl MerckGermany >99.5%

NaOH1M MerckGermany Standardsolution

HCl SigmaAldrich 37%

CaCl2.2H2O Merck >98.0%

Preparationofsolutions

• Fluoridestocksolutionwith10g/LofF-concentrationwaspreparedbydissolving

22.10gNaFindeionizedwaterandfillingupto1000mL.

• Aluminumsulfatestocksolutionwith10g/LofAl3+concentrationwaspreparedby

dissolving123.50gAl2(SO4)3.18H2Ointodeionizedwaterandfillingupto1000mL.

• Ca(OH)2solutionwaspreparedbyIWWwithaconcentrationof1.7g/mLofCa(OH)2.

• TISAB(TotalIonicStrengthAdjustmentBuffer)waspreparedbydissolving300g

sodiumcitrate;22gTitriplexIVand60gNaClintodeionizedwaterandfillingupto

1000mL18.

• CaCl2stocksolutionwith10g/LofCa2+concentrationwaspreparedbydissolving

27.69gCaCl2.2H2Ointodeionizedwaterandfillingupto1000mL.

• HCl0.1Mwaspreparedbydiluting6.2mLHCl37%intodeionizedwaterandfillingup

to1000mL.

AllsolutionswerestoredinPETcontainersformaximum1month.

MostoftheexperimentswereusingtapwaterwithtypicalqualityaspresentedinTable

4-2.


66

Table4-2Characteristicoftapwaterusedinmostexperiments(translatedfromRWW

Rheinisch-WestfälicheWasserwerksgesellschaftmbH:www.rww.de)

ParameterDescription Dimension Median LowValue Highvalue LimitbyGermandrinkingwaterordinance

Turbidity FTU <0.1 <0.1 0.2 1.0

Specificelectricalconductivity

µS/cm 457 290 582 2500

pHvalueat20oC 7.86 7.16 8.22 6.5-9.5

Calcium mg/L 43.7 33.1 61.9

Magnesium mg/L 7.5 5.9 9.4

Sodium mg/L 43.3 26.0 64.7 200

Potassium mg/L 4.2 2.8 5.8

Iron mg/L <0.01 <0.01 <0.01 0.2

Manganese mg/L <0.01 <0.01 <0.01 0.05

Ammonium mg/L <0.01 <0.01 0.02 0.5

Nitrite mg/L <0.01 <0.01 <0.01 0.5

Nitrate mg/L 12.9 9.4 18.9 50

Chloride mg/L 49 27 73 250

Sulfate mg/L 46 32 66 240

Phosphate mg/L 0.14 <0.10 0.34 6.7

Fluoride mg/l 0.12 0.09 0.15 1.5

Oxygen mg/L 7.3 3.0 13.1

TotalOrganicCarbon mg/L 0.64 0.41 0.94

Arsenic µg/L 0.5 0.2 0.9 1.0

Lead µg/L <0.5 <0.5 0.5 25(1)

Cadmium µg/L <0.1 <0.1 <0.1 5

Chromium µg/L <0.5 <0.5 0.6 50

Nickel µg/L 2.2 1.4 3.3 50

Mercury µg/L <0.05 <0.05 <0.05 1

Aluminum µg/L 2 <1 9 200

Boron mg/L <0.10 <0.10 0.12 1

Copper µg/L 1.6 <0.5 3.7 2000

Cyanide mg/L <0.01 <0.01 <0.01 0.05

During some experiments, rawwater obtained from river Ruhrwas also used and spiked

with 10 mg/L F-. Afterwards this water will be denoted as Ruhr water. The typical

compositionofRuhrwaterissummarizedintable4-3.


67

Table4-3TypicalcompositionofriverRuhrwater[internaldataobtainedfromIWWWater

Centre]

SubstanceConcentrationRange[mg/L]

Calcium(Ca) 35.5–45.3

Fluoride(F) <0.2

Magnesium(Mg) 6.5–8.6

Phosphate(PO4)total 0.13–0.21

Aluminum(Al) 0.017–0.035

Iron(Fe) 0.062–0.093

Manganese(Mn) 0.025–0.037

DissolvedOrganicCarbon(DOC) 1.9–2.1

4.3.2 Experimentalmethods

SpikedRuhrwater

Spikedwaterwas prepared by adding 1.8mL of fluoride stock solution to 1800mL Ruhr

water and stirred to assure the mixture contained 10 mg/L of F-. Each experiment was

performed in 2000-mL jars. To obtain 4mg/L of fluoride concentration in thewater, the

sameprocedurewasdonewithadditionof0.72mLstocksolution.

Effectofcalciumhydroxideonfluorideremoval

To investigate the effect of calcium hydroxide, 4.5 g of Ca(OH)2 solution was added into

spikedRuhrwaterwith10mg/Lfluorideconcentrationandwasstirredwith250rpmfor5

seconds.Then,thestirringratewasloweredto100rpmfor1minuteandcontinuedby30

rpmfor4minutes.Afterthat,thestirrerwasturnedoffandthesolutionallowedtostandfor

another45minutestocompleteprecipitationandthen20mLofsupernatantwastakenas

sampleofproductwaterandanalysed.Thisprocedurewasrepeatedforexperimentswith

1.5,2.25,3.0,and3.75gofCa(OH)2.

Reactiontimeforfluorideremoval

To investigate the reaction time, two experiments were done. The first experiment was

prepared by addition of 20mLNaOH 1M into spiked Ruhrwater. Then aluminum sulfate


68

solutioncontaining100mg/LAl3+wasaddedintothesystemandwasstirredwith250rpm

for5seconds.Thenthestirringratewasloweredto100rpmfor1minuteandcontinuedby

30rpmfor4minutes.Afterthat,thestirrerwasturnedoffandthesolutionallowedtostand

foranother45minutes to completeprecipitation.20mLof samplewere taken5 seconds

afteradditionofaluminumsulfatesolution,andsamplingrepeatedafter30seconds,1min,

2min,3min,4min,5min,10min,20min,30minand1hour.Thesecondexperimentwas

donebyadditionof4.5gCa(OH)2.Thefurtherprocedurewasthesameasbeforebutinthe

secondexperiment,pHwasadjusted tonearpH7.5byaddingconcentratedHClafter the

additionofaluminumsulfatesolution.

Effectofdifferentworkingparametersoncoagulationandprecipitation

The coagulation and precipitation performance could be affected by several working

parameters such asAl3+ to F-molar ratio, pH, Ca(OH)2, andNaOH. In order to investigate

these factors, severalexperimentswereconductedwithsimilarworkingconditions.At the

beginning, the spiked Ruhr water was mixed with 4.5 g of Ca(OH)2 (except for NaOH

concentrationparameterexperiment)solutionandstirredwith100rpmfor2minutes.Then

aluminumsulfatesolutionwasaddedaccordingtoTable4-4andstirringcontinuedwith250

rpmfor5seconds.Thenthestirringratewasloweredto100rpmfor1minutewhilepHwas

adjustedtothedesignatedvaluebyaddingHClorNaOHandstirringwascontinuedby30

rpmfor4minutes.Afterthat,thestirrerwasturnedoffandthesolutionallowedtostandfor

another 45 minutes to complete precipitation. Then 20 mL of supernatant was taken as

sampleofproductwaterandanalysed.

ThefollowingTable4-4summarizesthevariedexperimentalparameters.

Table4-4Summaryofexperimentalparametervariation

[F-]mg/L

pH [Al3+]to[F-]molarratio

Ca(OH)2g

4 7.5 1;3;4;6;7;10 4.5

10 7.5 5;6;7;8;11;13;14 4.5

10 4;5;6;6.5;7;7.5;8;8.5;9;9.5;10

7 4.5

10 7.5 7 1.5;2.25;3;3.75;4.5;


69

Reactiontimeexperiments

10 7.5 7 NoadditionofCa(OH)2butNaOH20mL

10 7.5 7 4.5

NaOHexperiments

10 ThepHbecomesthevariablesinceitisaffectedbyadditionofdifferentvolumesof1MNaOH

7 NoadditionofCa(OH)2butreplacedbyNaOH1Mwith0;5;10;15;20;25;30mL

4.3.3 Analyticalmethods

All samples were filtered through 0.45 µm membrane filters at the end of each

experiment.Filtrates forcationanalysiswereacidifiedwith0.1MHNO3tonearpH3.The

filtrateswereanalyzedby IWWfortotalaluminum,calciumandmagnesiumconcentration

simultaneouslyusing ICP-OES (VarianVistaProwithaxial PlasmaandAutosampler SPS3).

SulfateandChloridewerealsoanalyzedbyIWWusingionchromatographysimultaneously

withfluoride(IC-DX500DionexwithAutomatedSamplerAS3500).

To determine fluoride concentration, each sample wasmeasuredwith twomethods.

The first was ion selective electrode (ISE; METROHM, 781pH/Ion Meter). 50 mL of the

sample was taken after each experiment and the fluoride concentration was measured

accordingtoDIN38405(4)withanISE.Topreventinterferencesfromotherions(Al3+,Fe3+,

etc.),20mLofTISABsolutionwasaddedtosamples.Thesecondmeasurementmethodwas

by ion chromatography (IC-DX500, Dionex with Automated Sampler AS 3500) which was

done by IWW. A comparison of fluoride concentrations determined with both methods

showed good agreement, indicating that both methods were suitable for fluoride

concentrationanalysisaspresentedinFigure4-2.


70

Figure4-2ComparisonoffluorideconcentrationsdeterminedbyIonChromatography(IC)

andanIonSelectiveElectrode(ISE)standarddeviation:0.2298mg/L

4.4 Resultsanddiscussion

4.4.1 Effectofcalciumhydroxideonfluorideremoval

Theoriginalpurposeofthelime(Ca(OH)2)additionintowatercontainingfluorideduring

the“Nalgonda”processwastomaintainthepH4,19.WithoutmaintainingthepH,additionof

aluminumsulfatesolutionintothesystemleadstoadecreaseofpHintotheacidicrange.In

this condition, aluminumhydroxide is formed in lowamount, so there isnoprecipitation.

Thus,co-precipitationoffluorideisalsohindered.

In this experiment, the effect of calcium ionwas observed by adding Ca(OH)2 to the

system.TheadditionofCa(OH)2resultsinanincreaseofpH.DifferentCa(OH)2dosageswere

addedtothesystemasitisshowninTable4-4.Thefluorideconcentrationdecreasedfrom

10mg/Ltoapproximately8mg/L.Thisisbecausefluoridehasverystrongaffinitytocalcium

ionsformingfineCaF2precipitatethathaslowsolubilityinwater20.However,thesolubility

offluorite(CaF2)inCa(OH)2saturatedwaterisonly2mg/Landthispreventfurtherremoval

offluoride21,4.

0"

1"

2"

3"

4"

5"

6"

0" 1" 2" 3" 4" 5" 6"

F"#(m

g/L)#by#IC#

F"#(mg/L)#by#ISE#


71

Figure4-3EffectofadditionalCaOH2withinitialfluorideconcentrationof10mg/L,

withoutadditionalaluminumsulfatesolutiontofluorideconcentration,pH=7.5

4.4.2 Effectofsodiumhydroxideonfluorideremoval

Toinvestigateitseffectonfluorideremoval,NaOHwasaddedinthesystemtoreplace

Ca(OH)2. Different dosage of NaOH was added. Without addition of NaOH, system pH

decreasedtonear4.pHincreasedalongwithhigherdosageofNaOH.Overall,theresultof

theexperimentswassimilartoexperimentswithCa(OH)2.Fluorideremovalwasbestatthe

addition of 20 mL NaOH, which corresponds to pH 7.5 as shown in Kesalahan! Sumber

referensitidakditemukan..Fluorideconcentrationcanbedecreasedfrom10mg/Ltobelow

1mg/L.


72

Figure4-4EndpHofthesystembyadditionofNaOHasbackgroundalkalinity,[F-]=10

mg/LandAl3+toF-molarratio=7

Figure4-5EffectofNaOHonfluorideremovalwithF0=10mg/LandAl3+=100mg/L,Al3+to

F-molarratio7


73

4.4.3 Reactiontimeinfluorideremoval

Thereactiontimeexperimentsweredonetoensurethesystemwasatequilibriumstate

at the end of the experiments. After the addition of NaOH or Ca(OH)2, the system pH

increasedfromneutralpHtopHnear12.Then,Al2(SO4)3wasadded.Thereactionstarttime

wasmeasuredwhenAl2(SO4)3wasadded.LaterpHdecreasedrapidlybecausealuminumsalt

was hydrolyzed as it entered the system and liberated H+. However, the hydroxide ions

providedbyNaOHmaintainedthepHat7.5evenaftertheadditionofAl2(SO4)3.While,inthe

experimentswithCa(OH)2,afterAl2(SO4)3wasaddedintothesystem,HClhadtobeaddedto

adjustpHnear7.5.

Theresultoftheexperiment,asitisshowninFigure4-6,indicatedthatthedecreaseof

fluorideconcentrationoccursveryfast. InbothsystemswithadditionofNaOHorCa(OH)2,

the concentration of fluoride in the first 5 seconds decreased rapidly. After stirring was

stopped, the fluoride concentration reached the equilibrium state as it is indicated in the

resultthattherewasnofurthersignificantfluorideremoval.Fluorideendconcentration in

thesystemwithNaOHadditionwas1.4mg/LwhereasinthesystemwithCa(OH)2additionit

was0.9mg/LasitisseeninFigure4-7.


74

Figure4-6CoagulationofspikedRuhrwaterwithF-=10mg/L,withadditionof20mL

NaOH1Masbackgroundalkalinity(Al3+toF-molarratio=7.EndpH=7.5)

ThedecreaseofaluminuminthesystemwithNaOHoccurredveryfast.Thisisbecause

thesystemwerereachingequilibriumatpH7.5fasterthanthesystemwithoverdosageof

Ca(OH)2.TheslightincreaseofaluminumconcentrationinthesystemwithNaOHduringthe

stirringtimeissupposedtobetheeffectofmixing.Meanwhiletheincreaseofthealuminum

concentrationshowninFigure4-7issupposedtobetheeffectofadditionofconcentrated

HCl in a very short time that causes sudden decrease of the pH. At acidic pH aluminum

coordinationwithwater isdominant,especiallyatpHbelow3.5.Afterwards,pH increased

alongwithfurthermixing.AspHraised,thealuminumcoordinationwithhydroxylionsand

fluoride aremore dominant and the corresponding complexeswere formed. In this case,

aluminum hydroxide precipitate were formed an adsorb fluoride ion which led to the

decrease of both the concentration of fluoride and aluminum ion. Considering that the

coagulationwasdoneinexcessamountofcoagulant,sothepossibleformationofaluminum

hydroxide precipitate is very high, the fluoride concentration decrease was very fast and


75

tookplaceintheneutralpH,itcanbeconcludedthattheremovaloffluoridewasfollowing

sweepmechanism22. In addition, Figure 4-1 shows the dominant dissolved species at pH

between 6 and 8 is Al(OH)4-. Thismeans that the possibility of removing fluoride ions by

chargeneutralizationislow.

Figure4-7CoagulationofspikedRuhrwaterwithF-=10mg/L,withadditionof3mL

Ca(OH)2asbackgroundalkalinity(Al3+toF-molarratio=7.0,endpH=7.5)

4.4.4 Effectofaluminumtofluoridemolarratioonfluorideremoval

To investigate theeffectofaluminumto fluoridemolar ratioon the fluoride removal,

different ratioswere applied in the experiments.Additionof lime at thebeginningof the

experimentmadethepHtobecome12.ThenpHwasadjustedto7.5byaddingHClintothe

systemwhile itwas stirred. Figure 4-8 shows the result of different aluminum to fluoride

molarratiosonfluorideremoval.

Withaninitialconcentrationof10mg/L,lowfluorideremovalcanbeobservedinFigure

4-8withAl3+toF-molarratio=4,whileAl3+toF-molarratio>7thatisequalto100mg/L


76

aluminum succeeded to remove 85-96% fluoride, i.e., to achieve the desired fluoride

concentrationbelow1.5mg/L.Aloweramountofaluminumashydrolysisproductleadsto

lower possibility of complexation with fluoride. A higher molar ratio ensures the higher

possibilityforfluoridetocoordinatewithaluminum.However,lowerAl3+toF-molarratiois

favorable because it will decrease the risk of high residual aluminum dissolved and

suspendedintheproductwater.

Figure4-8EffectofaluminumsulfatesolutiondosingonfluorideremovalwithF0=10mg/L,

atpH7.5

Theresultoftheexperimentwithinitialfluorideconcentrationof4mg/Lispresentedin

Figure4-9.Theresultshowsthatfluorideremovaltobelowaconcentrationof1.5mg/Lwas

achieved at a molar ratio > 4 that is equal to 17 mg/L aluminum. This means, in such

environmentwherefluorideconcentrationinwaterisnottremendouslyhigh,lessaluminum

isneededtoremovefluoride.So,thismeansalso,thatindividualpre-analysisonparticular

watersourcehastobedonebeforeaddingaluminumtoremovefluorideinwaterinorder

tohaveanoptimumdosage.


77

Figure4-9EffectofaluminumsulfatesolutiondosingonfluorideremovalwithF0=4mg/L

4.4.5 EffectofpHonfluorideremoval

ThisexperimentaimedtoinvestigatetheeffectofpHonfluorideremovalbyaseriesof

experiments with different end pH. In the experiments, it was observed that when the

Ca(OH)2enteredthesystempHraisedto12.Thereactionofcalciumandfluorideionswas

supposedtooccurinthesystem,formingveryfinecolloidalprecipitate.Thisreactionleads

to the decrease of the fluoride concentration in the water from 10 to 8 mg/L, which

corroboratestheresultsofpreviousstudies4,8,21.

Asaluminumenteredthesystem,pHdecreasedfastandwasadjustedto7.5byadding

dropsofconcentratedHClwhilestillunderstirring.Figure4-10showsatpHbetween6to8,

fluoride removal to concentrations below 1.5 mg/L were achieved. However, at pH > 8

fluorideconcentrationincreasedagaintoabove1.5mg/L.

At pH below 6, the solubility of Al(OH)3 is higher, and the possibility of Al(OH)3

precipitate to be formed is lower than at higher pH. Without formation of Al(OH)3

precipitate,co-precipitationwillnotoccur.


78

Figure4-10EffectofpHonfluorideremovalwithinitialfluorideconcentrationof10mg/L.

Al3+toF-molarratio=7

AlongwithincreasingpHbetween6tonear8,hydroxylionswerepresentinquitelarge

concentration, which ensures the formation of Al(OH)3 precipitation. In addition, at pH

between 6 to 8, Al(OH)3 is below its pHPZC, thus it is positively charged, which supports

sorptionof thenegativelycharged fluoride ionsbyelectrostatic interactions.Therefore,at

this range of pH between 6 – 8, fluoride removal to concentrations below1.5mg/Lwas

achieved.

At pH above 8, the formation of Al(OH)4- predominates the system. In addition, the

smallamountofAl(OH)3precipitateisnegativelychargedatpHaboveitspHPZC,sothisleads

torepulsionoffluorideions.Therefore,atpH>8fluorideremovalisverylowwithfluoride

concentrationsincreasingwithincreasingpHandresidualfluorideconcentrationsattheend

oftheprocessabove1.5mg/L.

ApreviousstudyfoundthatattheneutralpHbelowthepHPZCtheflocsformedduring

flocculation are more stable23. This result is supported by a previous study on fluoride

removalusingcakealum,whichconcludedthatthesludgeflocsformedaroundpH7.1were

muchmorestableandpresentedbettersettlingcharacteristicsthanintheacidicrangeofof


79

6.1–6.423.Themorestabletheflocsare,thelowertheconcentrationoffluorideionsinthe

producedwater.

Byusingfluorideandaluminumconcentrationfromtheexperimentsasinputdatainthe

massbalancecalculationforspeciescalculation,adiagramofspeciesdistributionasfunction

ofpHcanbeestablishedasshowninFigure4-11.Thealuminumandfluorideconcentrations

were inserted as free component in the system, while in Figure 4-1 those values were

insertedastotalconcentrationofcomponents.AtpHbetween4to6AlF3predominatesin

the system. At pH near 7 AlF2+ is the predominant species. This is different from the

theoreticalcalculationinFigure4-1.AtpH>7,Al(OH)4-isthedominantspecies.

Since fluoride ions were removed mainly by co-precipitation, formation of Al(OH)3

precipitate is very important. At pH between 4-6 there was no formation of Al(OH)3

precipitate.Thus,fluorideremovalwasverylittleandfluorideconcentrationislargeinthis

pHrange.Thepresenceoffluorideandaluminum-fluoridecomplexesinthesystemchanges

theequilibriumofthesystem,thusthespeciesfractiondoesalsochangeatthespecificpH.


80

Figure4-11 Speciesdistributionof aluminum-fluoride complex in solutionas functionof

pHbyusingthefluorideandaluminumconcentrationsfromtheexperimentsasinputdata

forcalculationusingChemEQLV3.0

4.4.6 Summary

The result of the study on the coagulation and co-precipitation method for fluoride

removalusingaluminumsulfatesolutionascoagulantcanbesummarizedasfollows:

1. Theremovaloffluorideoccurredveryfast.ItwasbestcarriedoutatneutralpHand

under excessive amount of aluminum coagulant. Fluoride ions were adsorbed by

precipitated aluminum hydroxide. After certain time the precipitated aluminum

hydroxidecollidedandenmeshedtheparticlesandlatersettleddown.Itissuggested

thattheremovalreactionfollowsasweepmechanism.

2. Reduction of fluoride concentration from an initial concentration of 10 mg/L to

below1.5mg/Lwasbestatanaluminumdosingof100mg/Lthatiscorrespondingto

anAl3+toF-molarratio≥7.Thismolarratioassuresahighpossibilityforfluorideto


81

occupy the aluminum hydroxide precipitate as adsorption sites. Meanwhile, for

fluoridewithinitialconcentrationof4mg/L,anAl3+toF-molarratio≥4,equalto17mg/L aluminum, achieved that purpose. This amount of aluminum is clearly lower

thanneededintheNalgondatechniquewhichis16to181mg/LasAlfortreatingraw

waterwithfluoridelevelsof2to8mg/L.Loweramountsofaluminumarepreferred

toavoidexcessofaluminumresidueintheproductwaterandtominimizethesludge

formation.

3. The removal of fluoride in rawwater from the initial concentration of 10mg/L to

below1.5mg/LwasachievedinthepHrange6-8.AtthispHrangeAl(OH)3haslow

solubilityandeasilyprecipitates.Besides,at thispH thepHPZCc ofAl(OH)3 indicates

thatthesolidisneutraltopositivelycharged.Inaddition,bymaintainingthepHon

this level, the amount of OH- ions as competing ion to fluoride to occupy Al(OH)3

precipitateisalsosmaller.

4. ItcanbealsoconcludedthatadditionalCa(OH)2servesnotonlyformaintainingpH

butalsoforprecipitatingfluorideion.Inaddition,Ca(OH)2mightalsoactastheseed

ofcoagulationnucleus.


82

4.5 References

(1) Fawell,J.K.;Bailey,K.FluorideinDrinking-Water;WorldHealthOrganization,2006;pp1–144.

(2) Jacks,G.;Bhattacharya,P.;Chaudhary,V.;Singh,K.P.ControlsontheGenesisofSomeHigh-FluorideGroundwatersinIndia.AppliedGeochemistry2005,20(2),221–228.

(3) Njau,B.;Dahi,E.;Dahi,E.;Mtalo,F.;Mtalo,F.;Njau,B.;Bregnhøj,H.DefluoridationUsingtheNalgondaTechniqueinTanzania;22ndWEDC…:NewDelhi,1996.

(4) Meenakshi;Maheshwari,R.C.FluorideinDrinkingWaterandItsRemoval.JournalofHazardousMaterials2006,137(1),456–463.

(5) Agarwal,K.C.;Gupta,S.K.;Gupta,A.B.DevelopmentofNewLowCostDefluoridationTechnology(Krass).WaterScienceandTechnology1999,40,167–173.

(6) George,S.;Pandit,P.;Gupta,A.B.;Agarwal,M.ModelingandSimulationStudiesforAluminium-FluorideInteractionsinNalgondaDefluoridationProcess.ChemicalProductandProcessModeling2009,4(1),27.

(7) Ayoob,S.;Gupta,A.K.;Bhat,V.T.AConceptualOverviewonSustainableTechnologiesfortheDefluoridationofDrinkingWater.CriticalReviewsinEnvironmentalScienceandTechnology2008,38(6),401–470.

(8) Yang,C.-L.;Dluhy,R.ElectrochemicalGenerationofAluminumSorbentforFluorideAdsorption.JournalofHazardousMaterials2002,94(3),239–252.

(9) Ku,Y.;Chiou,H.-M.TheAdsorptionofFluorideIonFromAqueousSolutionbyActivatedAlumina.Water,Air,&SoilPollution2002,133(1-4),349–361–361.

(10) Mahramanlioglu,M.;Kizilcikli,I.;Bicer,I.O.AdsorptionofFluorideFromAqueousSolutionbyAcidTreatedSpentBleachingEarth.JournalofFluorineChemistry2002,115(1),41–47.

(11) Mekonen,A.;Kumar,P.;Kumar,A.IntegratedBiologicalandPhysiochemicalTreatmentProcessforNitrateandFluorideRemoval.WaterResearch2001,35(13),3127–3136.

(12) LópezValdivieso,A.;ReyesBahena,J.L.;Song,S.;HerreraUrbina,R.TemperatureEffectontheZetaPotentialandFluorideAdsorptionattheΑ-Al2O3/AqueousSolutionInterface.JournalofColloidandInterfaceScience2006,298(1),1–5.

(13) Weng,L.;Temminghoff,E.J.M.;VanRiemsdijk,W.H.AluminumSpeciationinNaturalWaters:MeasurementUsingDonnanMembraneTechniqueandModelingUsingNICA-Donnan.WaterResearch2002,36(17),4215–4226.

(14) Hem,J.D.;Roberson,C.E.FormandStabilityofAluminiumHydroxideComplexesinDiluteSolution.Geol.Surv.Water-SupplyPap.(U.S.);(UnitedStates)1967,1827-A.

(15) Crittenden,J.;Trussell,R.R.;Hand,D.W.;Howe,K.J.;Tchobanoglous,G.WaterTreatmentPrinciplesandDesign,2nded.;JohnWiley&Sons,Inc:NewYork,2005.

(16) Lide,D.R.;Frederikse,H.P.R.CRCHandbookofChemistryandPhysics,1993-1994:aReady-ReferenceBookofChemicalandPhysicalData;CRCPress:BocaRaton(Florida),1993.

(17) Parthasarathy,N.CombinedUseofCalciumSaltsandPolymericAluminiumHydroxideforDefluoridationofWasteWaters.WaterResearch1986,20(4),443–448.

(18) DIN.DIN38405-4:DeutscheEinheitsverfahrenZurWasser-,Abwasser-UndSchlammuntersuchung;Anionen(GruppeD);BestimmungVonFluorid(D4);Beuth


83

Verlag,1985.(19) Parks,G.A.AqueousSurfaceChemistryofOxidesandComplexOxideMinerals.In

EquilibriumConceptsinNaturalWaterSystems;Stumm,W.,Ed.;IsoelectricPointandZeroPointofCharge;AMERICANCHEMICALSOCIETY:Washington,D.C.,1967;Vol.67,pp121–160.

(20) Lawler,D.F.;Williams,D.H.Equalization/NeutralizationModelinganApplicationtoFluorideRemoval.WaterResearch1984,18(11),1411–1419.

(21) Wang,Y.;Reardon,E.J.ActivationandRegenerationofaSoilSorbentforDefluoridationofDrinkingWater.AppliedGeochemistry2001,16(5),531–539.

(22) Lee,J.-D.;Lee,S.-H.;Jo,M.-H.;Park,P.-K.;Lee,C.-H.;Kwak,J.-W.EffectofCoagulationConditionsonMembraneFiltrationCharacteristicsinCoagulation−MicrofiltrationProcessforWaterTreatment.EnvironmentalScience&Technology2000,34(17),3780–3788.

(23) Piñón-Miramontes,M.;Bautista-Margulis,R.G.RemovalofArsenicandFluorideFromDrinkingWaterwithCakeAlumandaPolymericAnionicFlocculent.Fluoride2003,36(2),122-128

84

Chapter5 HybridProcessestoRemoveFluorideandExcessof

Aluminum

5.1 Background

Afteryearsofexperience,theNalgondaprocesshasbeenfoundtohaveitsdrawbacks

namely elevated aluminum concentration and turbidity in the treatedwater. In domestic

defluoridationdesign,separationofproductwaterfromtheprecipitateiscompletedonlyby

thesedimentationprocess. Insomehouseholdoperation,screenandclothfiltersareused

afterthecoagulationsteptoremoveprecipitate1,2.However,thisstepcannoteliminatethe

excessofaluminumbecausealuminumprecipitateisveryfinewithadiameterofabout10 μm3.Apreviousstudy foundthat theresidualaluminumcontent inproductwaterwas in

therangeof2to7mg/L4.Thisaluminumconcentrationinproductwaterobviouslyexceeds

theWorldHealthOrganization(WHO)drinkingwaterlimit,whichisonly0.2mg/L5.

A high concentration of dissolved aluminum in drinking water is considered to be

harmfultothehumanhealth6.Itispresumedthataluminumisapotentialneuro-toxicant.

Medicalresearchandepidemiologicalsurveyssuggestthatdissolvedaluminumenteringthe

bloodstreammay cause Alzheimer disease 7. It has been reported that Al3+, Al(OH)2+ and

Al(OH)2+arethemosttoxicspeciesofaluminumwhilecomplexeswithfluorideandorganic

acidsuchascitricandoxalicacidarelessharmful8,9.Turbidityinthefinishedwater,another

drawbackof theNalgondatechnique,occursbecausenotallprecipitatesettlesdown.The

very fine precipitates are left in the water as suspended particle. Suspended precipitate

cannoteasilyberemovedwithoutadequateseparationtechnique.

5.2 Sandfilterasaseparationmethod

5.2.1 Previousstudyon“BioSand”filter

ApreviousstudyinthatcontextwasdonetoinvestigatetheuseofaBioSandFilter(BSF)

to remove fluoride as single-step treatment and in combination with the Nalgonda

technique. BSF is amodification of slow sand filtrationwith an active layer on topof the

filter.WhenonlyaBSFwasapplied to treat the rawwaternosignificant fluoride removal

HybridProcesses

85

occurred.Inaddition,alternativelocalmaterialsusedasfiltermediasuchaslocalbonechar,

locallateriticclay,andlocalbrickwererapidlylostitsperformancethatarenotsuitablefor

longtermapplication.Thestudyrecommendedthatfluorideremovalshouldbeperformed

asaseparateprocesspriortoBSFandconductedpreferablybytheNalgondatechniquefor

fluoride concentrations less than 10mg/L 10. In some communitywaterworks, Nalgonda

techniquewas equippedwith sand filter 11. However, aluminum concentrations after the

filtrationstephavenotbeeninvestigated.

5.2.2 Sandasfiltermediainwaterworks

Inmostwaterworks,coagulationandprecipitationisfollowedbyfiltration,usuallybya

rapidsandfilter 6.This is thesimplesthybridsystem inwater treatment.Thesandfilter is

ensuring that particles not settling in the sedimentation step can be removed effectively.

Typically,simplefiltersuseasingletypemedia,whichismostlysand.

Slow sand filter work on low filtration rate between 0.05-0.2m/h 12. This process is

treatingwater by twoprocesses. The first is physical treatment such as screening, setling

andadsorptionandthesecondisbiologicaltreatment.Thebiologicaltreatmentisdoneby

biologicalgrowthontopofthesand.Rapidsandfilterworkwithagreaterfiltrationrateof

5-15m/handuseacoarsermedium,whichhashigherpermeability12.Thisfiltertreatsthe

inflowwaterwithonlyphysical treatment. It is reportedthat rapidsand filtersareable to

removeheavymetalsbutalso,undercertaincircumstances,phosphorousthroughchemical

precipitation12.Pressurefiltersareaformoffiltersthatapplyingpressuretoenhanceflow

thewater.Itisnormallyusedinthetreatmentofgroundwaterthatispumpeddirectlyfrom

theboreholeintothedistributionnetworkfortreatment.Flowratesareoftenintheregion

of0.3–0.6m/h

5.2.3 Removalmechanismoffiltermedia

There are two major mechanisms of the particle removal in a sand filter, namely

transport andattachment. Transportmechanisms are responsible for providing forces to

moveparticlesoutoftheirstreamlinesontothegrainsurface.Thesuspendedparticlesare

transported to the filter material in different ways. This transport mechanism includes

screening when colloidal or suspended particles are retained on the filter bed and this

applies to large particles that will not fit to pass the space between filter materials.

HybridProcesses

86

Interceptionoccurswhenwaterstreamlineapproachthefiltergrainsothatlightparticleare

intercepted and become attached to filter grain. Heavy particles are usually subject of

inertia forces andgravitational settling,which canbeneglected for smallparticles in the

micron size. Small particles are dominantly affected by diffusion and hydrodynamic

turbulencebecausetheirinertiaissmallerthanhydrodynamicforcesinthesystem.

After being transported, particles that reached the surface of the filter media with

diameter of < 1 μm are subjects of surface force 3. These particles are retained on the

surfaceofthemediabasedonthreeattachmentmechanisms: vanderWaals force isthe

major attraction force that supports the attraction between filter grain and suspended

particle. Electrostatic interaction occurs when the particles and media surface carry

electrical charges, which could be attraction or repulsion forces. The last is chemical

bridging when the attachment is supported by destabilizing agents that lead to the

formationofheavyweightpolymerthatforma‘bridge’betweenfiltergrainandsuspended

particleorbetweenparticles13

5.3 MembraneProcessasseparationmethod

Aspreviouslydiscussedinchapter1,thereareseveralmembraneprocessesthatcould

remove fluoride directly. Reverse osmosis (RO) and nanofiltration (NF) are the most

commonandwidelyusedmembraneprocessesindefluoridationofwaterwithhighfluoride

concentrationand/ordesalination14-20.GenerallyROisusedtodesalinateseawaterinareas

with limited availability of fresh water or on ships. Some researchers 14,17 reported the

applicationofnanofiltration(NF)indefluoridationasanalternativetoRO.

ThedrawbackofROandNFintreatingfluoride-richwatersisthatupto99%ofthesalts

inthewaterarerejectedbythemembrane,whichmeansnotonlyfluoridewillberemoved

butalsootheressentialsaltsareeliminated.Besides,thehighpressureappliedtobothtypes

ofmembrane results ina less costeffectivemeasure, especially if it isdesignated for less

developed areas, which actually comprises most areas of fluoride contaminated drinking

watersources.

Ontheotherhand,ultrafiltration(UF)membranesareincreasinglyappliedinadvanced

drinkingwatertreatmentprocesses,particularlyto improvethewaterqualitywithrespect

to organic and microbiological parameters. UF membranes can deliver constant water

HybridProcesses

87

quality, show good removal efficiency towardsmicroorganisms, less production of sludge

andsmallsizeplantcomparedtoconventionaltreatmentplant21-23.However,duetoitsvery

small size, ca. 0.0001 μm ionic radius, dissolved fluoride ion cannot be removed by UF

membraneswhichusuallyhasporesizesintherangeof0.1–0.01μm.

Considering that aluminum floc size is 10 μm 3, coupling coagulation and co-

precipitationwithUF filtrationwas proposed to be an appropriate treatment not only to

remove fluoride in drinkingwater but also the aluminum residue due to large dosage of

coagulant.Apreviousstudyonthecombinationoftheprecoagulatedwaterusingaluminum

based coagulant prior to UF membrane process succeeded in minimize the membrane

fouling24.Otherstudiesonthiscombinationreportedthatflocsretainedonthemembrane

surfaceascakelayercouldimprovemembraneperformance25.However,thereisnostudy

yetonthisparticularcombinationtoremovefluoride.

5.3.1 Ultrafiltrationmembranescharacteristic

Ultrafiltrationmembranes arenormally applied inorder to separatewater from large

particlesandarecharacterizedbytheirabilitytoremovesuspendedandcolloidalparticles.

Theyarealsoknownaslowpressuredrivensystems,sincehighfluxcanbereachedwitha

pressuredifferenceof0.2–2bar.UFmembraneshaveanominalporesizeof0.01to0.1μm26. UF is normally used for treating water and wastewater, purifying, concentrating and

fractionating macromolecules. UF membranes are also able to remove a large range of

materials fromwater, including fine particle suspensions, colloids, and to retain bacteria,

viruses and parasites. Nowadays, UF membranes have become a cost competitive and

feasible alternative to conventionalmethods due to a remarkable decrease inmembrane

filtrationcostsenabledbyinnovationsinmembranemanufacturingandprocessconditions27.

UFmembranes show usually anisotropic structures generated in the phase inversion

method.Theyarenormallycharacterizedintermsofmolarmass(molecularweight)cut-off

(limit) instead of particle size (pore size in microns). Nominal molecular weight cut-off

(NMWCO)ornominalcut-off isusuallydefinedas themolarmassofa testmolecule that

wouldberetainedto>90%bythemembrane.TypicalNMWCOofUFmembranesareinthe

rangeof1to200kg/mol..ThepressureusedinUFisusuallywithintherangeof50kPa-1

HybridProcesses

88

MPa. Among several polymer materials, polysulfone (PS), polyethersulfone (PES), and

cellulose-based polymers are usually used for UF membranes. Due to their mechanical

strength,thermalandchemicalstabilityaswellasexcellentfilmformingproperties,PSand

PESarefrequentlyusedasmaterialsforhighperformanceUFmembranes28.

5.3.2 Membranefoulingandconcentrationpolarization

Membrane fouling is one factor that limits the spread of membrane technologies in

water treatment plants. Fouling is a deposition of suspended or dissolved substances on

externalsurfaces,attheporeopeningsorwithintheporesofamembrane,resultinginloss

of performance 29. Fouling is a result of the interactions between the membrane and

solute(s) in the feed and perhaps between the adsorbed solutes and other solutes in the

feedstream.Fouling isgenerally influencedbythreefactors:membraneproperties,solute

(solution)properties,andoperatingparameters30.Thesefactorsmightinteractandleadto

different effects in different combinations. The presence of inorganic compounds and

suspendedparticlesmay leadtomorecomplexmembrane fouling.Fouling isalsostrongly

influencedbypressureandfeedconcentrationaswellasoverallequipmentdesignsuchas

temperatureandconfigurationofequipment.

Concentration polarization (CP) is a natural consequence of the selectivity of a

membrane that leads to the build-up of the solute concentration near and/or on the

membrane’ssurface.ThefluxduringUFwill increaseastrans-membranepressure(TMP)is

increased.CPaffectsthefluxbyreducingtheeffectiveTMPdrivingforceduetotheosmotic

pressuredifferencebetweenfiltrateandfeedsolutionatthemembranesurface.However,

afterCPreachesmaximalpointwhereagellayerhasbeenformed,thefluxwillnotdecrease

anymore.Thisphenomenonisinevitable,butitisreversiblesothatTMP,whichalsoleads

to the reduction of fluxes, can be reduced again 31. Figure 5-1 illustrates the distinction

betweenfoulingandCP.

HybridProcesses

89

Figure5-1Fluxillustrationtodistinguishmembranefoulingandconcentration

polarization;thesolidblacklinedescribesthefluxwithconcentrationpolarization,the

solidlightgreylinedescribesfluxwithfouling.

5.3.3 Operationmodeofmembraneprocess

Therearetwodifferentoperationmodesofthemembrane:dead-endandcross-flowas

shown inFigure5-2. Indead-endmode,all feedwaterpassesthroughthemembraneand

fouling materials (foulants) are carried to it. It means deposition of material on the

membranesurfacealreadybeginswhenthefiltrationprocessstarts.Incross-flowmode,the

major part of the feedwater flows tangentially across themembrane surface and only a

small part of the flow passes themembrane. Generally cross-flow operation is preferred

becausetheturbulencegeneratedduringtheoperationprovidesthinnerdepositslayerand

henceminimizesfouling26.

Figure5-2IllustrationofDead-end(left)and

Cross-flowfiltration(right)

5.3.4 Foulingmechanism

Fouling mechanisms can be classified in the

followingformsthatareillustratedinFigure5-3:

Membrane

Suspensionsflux

Clogginglayer

Permeateflux

Membrane

Suspensionflux Clogginglayer

Permeateflux

HybridProcesses

90

• Adsorption: occurs when attractive interactions between the membrane and the

solutesorparticlesexist.Amonolayerofparticlesandsolutescangroweveninthe

absence of permeation flux leading to an additional hydraulic resistance. If the

degree of adsorption is concentration dependent, concentration polarization will

exacerbatetheamountofadsorption.

• Poreblockage:whenfiltering,poreblockagecanoccurleadingtoareductionofflux

duetotheclosure(orpartialclosure)ofpores.

• Deposit:adepositoftheparticlecangrowlayerby layeratthemembranesurface

leading to an additional hydraulic resistance, which is often referred to as cake

resistance.

• Gel: the level of concentration polarization may lead to gel formation for certain

macromolecules.

Figure5-3Foulingmechanismsofamembrane31

Solute charge, density, hydrophobicity, ionic strength and pH are solution properties

influencingtheextentandthebehavioroffouling.Forexample,mineralsaltscanprecipitate

on themembrane surfacebecauseofpoor solubilityorbind to themembranedirectlyby

chargeinteractions30.Inaddition,saltscanincreasetheionicstrength,whichinturnaffects

solute-membraneinteractions.

5.4 Aimofstudy

This study focused on the removal of aluminum residue in the product water after

coagulationandco-precipitationprocessandproposedtwoalternativesmethods.Thefirst

methodishybridprocessofcoagulation-precipitationandsandfilterwithfocusontheability

ofthesandfiltertominimizeturbidityandtoremovetheexcessofaluminumconcentration

toaconcentrationbelowthelimitsetbyWHOindrinkingwater,whichis0.2mg/L,andfor

turbidity below 1 NTU. The second method is a hybrid process of coagulation-co-

precipitation andUFmembrane processwith focus on the overall aluminum and fluoride

removalabilityofthehybridsystem.Besides,sincefoulingisthemajorfactorthatlimitsthe

application of membrane processes in water treatment, the interaction between the

HybridProcesses

91

membraneandproductwaterafterthecoagulationprocesswasalsostudied.Experiments

wereperformedinlaboratoryscale.

5.5 Experimental

5.5.1 Materials

Table5-1Listofmaterials

Chemicals Suppliers Purities

NaF Merck >99.5%

Al2(SO4)3.18H2O Merck 51%-59%

Ca(OH)2 n.a. 88.5%

Sodiumcitrate Merck 99.0%-101.0%

TitriplexIV Merck >99.0%

NaCl MerckGermany >99.5%

NaOH1M MerckGermany Standardsolution

HCl SigmaAldrich 37%

CaCl2.2H2O Merck >98.0%

Preparationofsolutions

• Fluoridestocksolutionwith10g/LofF-waspreparedbydissolving22.10gNaFin

deionizedwaterandfillingupto1000mL.

• Aluminumsulfatesolutionstockwith10g/LofAl3+waspreparedbydissolving123.50

gAl2(SO4)3.18H2Oindeionizedwaterandfillingupto1000mL.

• Ca(OH)2solutionwaspreparedbyIWWwithaconcentrationof1.7g/mLofCa(OH)2.

• TISAB(TotalIonicStrengthAdjustmentBuffer)waspreparedbydissolving300g

sodiumcitrate;22gTitriplexIVand60gNaClindeionizedwaterandfillingupto1000

mL32.

• HCl0.1Mwaspreparedbydiluting6.2mLHCl37%indeionizedwaterandfillingupto

1000mL.

AllsolutionswerestoredinPETcontainersforamaximumofonemonth.

SpikedriverRuhrwater

HybridProcesses

92

Spikedwaterwaspreparedbyadding4.8mLoffluoridestocksolutionto4800mLRuhr

water and stirred to assure the mixture contained 10 mg/L of F-. Each experiment was

performedin5000-mLjars.

Filtermediaforsandfilter

Thefiltermediumusedinthisstudywassand(Europeanquartz)withaneffectivesize

(d10,10thpercentileofgraindiameter)of0.82mmanduniformitycoefficient(d60/d10,with

d60meaning60thpercentileofgraindiameter)=1.256.Thesandfilterwassetupinaglass

columnwithdimension4.7cmdiameterand15cmlength.Thesandbeddepthwas8.5cm

Ultrafiltrationmembrane

Commercial polyethersulfone (PES) UF membranes with a nominal cut-off of 100

kg/mol, obtained from Sartorius, Germany were used. The molecular structure of PES is

shown in Figure 5-4. By scanning electron microscope (SEM) analysis of the membrane

surfaceand its cross-section, it is shown that thepolyethersulfonemembrane is aporous

asymmetricmembraneas itcanbeobserved inFigure5-5.PES isathermoplasticpolymer

thatcanbeusedasamembraneforitstolerancetochlorideandawiderangeofsolvents.

Figure5-6showstheporesandtheporesizeofthemembrane.Variationsinporesizesand

poredensitiesareclearlyobservedfromsurfaceimagemorphology.Theporeandporesize

isnotuniformlydistributed.Regardingitsinteractionwithpurewater,thisPESmembraneis

a hydrophilic material, which is easily wettable as seen from its contact angle of 51° for

water.Thismembraneisnegativelycharged.

Figure5-4Illustrationofchemicalstructureofpolyethersulfone33

Figure5-5SEMimageofunusedPESUF100kDamembranecross-section

Figure5-6SEMimageofunusedPESUF100kDamembranetopsurface

HybridProcesses

93

5.5.2 Experimentalset-up

Coagulation–co-precipitationstep

The hybrid process was based on a two steps operation set-up, first coagulation–co-

precipitation and second filtration and 8 experiments had been performed. The first step

parameterswerebasedonthefindingsdiscussedinchapter4withanoptimalaluminumto

fluoridemolarratiosetto7whilethepHwasadjustednear7.5.

Thefirststepwasconductedin2jarswith4800mLofspikedRuhrwater.Intoeachjar,

first,3mLCa(OH)2solutionthatcontained4.5gCa(OH)2wereaddedintospikedRuhrwater

and stirredwith100 rpm for2minutes. Thiswas followedbyadditionof48mlof10g/L

aluminumsulfatesolutiontothesystemandwasstirringwith250rpmfor5seconds.Then

thestirringratewasloweredto100rpmfor1minutewhilepHwasadjustedtonear7.5by

adding HCl or NaOH and continued by 30 rpm for 4 minutes. After that, the stirrer was

turned off and the solution allowed to stand for another 45 minutes to complete

precipitation before 100 mL of supernatant was taken as sample of product water and

analyzed.

Filtrationstepforhybridcoagulation-coprecipitationandsandfilter

Thenextstepwasthefiltrationprocess.Theproductwaterofthepreviousprocesswas

fed into the sand filter. Figure 5-7 shows the diagram of the experimental set up. The

filtration flow rate was set to 10 L/h and 12 L/h or equal to 0.57m/h and 0.69m/h by

applyingpump.Whenthewaterinbothjarsreachedabout1cmabovetheuppertopofthe

precipitationlayer,filtrationwasstopped.Basedonthepre-experimentalresultthatafter1

hoursofoperation, turbidityvaluehad increased,thefiltrationprocesswasrepeateduntil

totalproductwatervolumethatpassedtothefilterequalto1hourofoperationforeach

experimental batch or ca. 14 L. 20mL samples for turbidity analysis were taken every 5

minutes.Aftereverybatch,thesandinthefilterwasremovedandwashedwithdeionized

wateruntil thewater thatwereuse toclean thesandshowturbidityvalueof<0.3.After

properdrying,thecleanedsandwasusedagain.Thecleaningstepwasimportanttoassure

thereproducibilityofdata.

HybridProcesses

94

Figure5-7Experimentalsetupofhybridcoagulation-precipitationandsandfilter,whereM

=Motor,Fl=Flowmeter,PI=Pressuremeter,

Filtrationwithdead-endmodeUFflatmembrane

After twocontainerswere filled,productwaterwas flowedto themembranemodule

andthesystemwasoperatedfor30minutesforeachexperiment.TheUFmembranewith

100kDafromSartorius(PESD=47mm)wasusedtogetherwithafilterholderSM16249,Ø

47mm,Anom=13cm2.Withlongbodyand200mLcapacity,thisholdermodelpermitsthe

flux to flowwith a constant profile. Two PTFE screens and a silicon O-ring were used to

supportthemembrane.Insteadofcompressedair,nitrogenwasused,asitisaninertgasto

buildup thepressure.Nitrogenwaspressurized inside thepressurevesselscontaining the

product water by using hoses connected to the filter holder. A pressure gauge was

connectedtothehosetoascertainthatthepressurewassetat0.5,1.0,1.5and2.0barsand

eachexperimentwasdone in triplicate.The filtratewas collected inapre-cleanedbeaker

glass and weighted by a digital balance (Kern Compact Balance EW serial number

(3000±0.5g).Thebalancereadthedata4-6timespersecondandtheresultwassentdirectly

to the computer. 20 mL sample of filtrate was taken from the beaker glass after the

experiment for fluoride and aluminum analysis. The following Figure 5-8 illustrates the

experimentalsetupofthecoagulationandUFunit.

Figure5-8Experimentalsetupillustrationofhybridcoagulationandco-precipitationwith

UFmembrane.M=MotorizedStirrer,F=Flowmeter,P=Pressuregauge,

HybridProcesses

95

5.5.3 Analyticalmethods

Fluorideconcentration

To determine fluoride concentration, each sample was measured with ion selective

electrode (ISE;METROHM, 781pH/IonMeter). 20mL of the samplewas taken after each

experimentandthefluorideconcentrationwasmeasuredaccordingtoDIN38405(4)withan

ISE.Topreventinterferencesfromotherions(Al3+,Fe3+,etc.),20mLofTISABsolutionwas

addedtosamples.

Aluminumconcentration

AluminumconcentrationwasanalyzedbyusingSchnelltestHACHLange foraluminum

testwithcolorimeterDR/890.Thisinstrumenthasdetectionrangeof0to0.80anddetection

limitof0.013mg/LforAl3+.First,50mLsamplewastakenandthepHbeingadjustedtopH=

3.5–4.5beforebeinganalyzed.Becausefluoridecomplexesandinterferesaluminumatall

concentration levels, it is necessary toderive theactual aluminumconcentrationbyusing

thealuminuminterferencegraphshowninFigure5-9.

Figure5-9Aluminuminterferencediagram34

HybridProcesses

96

Aluminum,calcium,magnesiumandsulfateanalysis

All samples were filtered through 0.45 µm membrane filters at the end of each

experiment.Filtrates forcationanalysiswereacidifiedwith0.1MHNO3tonearpH3.The

filtrateswereanalyzedby IWWfortotalaluminum,calciumandmagnesiumconcentration

simultaneouslyusing ICP-OES(VarianVistaProwithaxialPlasmaandAutosamplerSPS3).

Sulfate was also analyzed by IWW using ion chromatography (IC-DX500 Dionex with

AutomatedSamplerAS3500).

Turbidity

Turbidity wasmeasuredwith Dr. Lange Turbidimeter (Nephla). 20mL of samplewas

taken and filled into a cuvette. The turbidity of the samples was read on the display as

FormazinNephelometricUnit(FNU),whichisalmostequivalenttoNephelometricTurbidity

Unit (NTU)with slight differences. Both techniquesmeasure scattered light at 90degrees

from the incident light beam by suspended solid in water sample. However, the FNU is

measuredwithaninfraredlightsourcewhereastheNTUismeasuredwithwhitelight.

Characterizationofmembranebyscanningelectronmicroscopy

Thetopsurfaceandcross-sectionmorphologiesofthemembranewereobservedusing

scanningelectronmicroscope(SEM)byCentralLaboratoryofScanningElectronMicroscopy,

UniversityDuisburgEssen.Anenvironmentalscanningelectronmicroscope(ESEM)Quanta

400FEGatstandardhigh-vacuumconditionswasused.AsputtercoaterK550(Emitech,UK)

wasusedforcoatingoftheoutersurfaceofthescansamplewithgold/palladium.Forcross-

sectionanalysis,tomaintainthestructure,themembranewasbrokeninliquidnitrogenand

sputtered for 1.5minutes, while for analysis of outer membrane surface, sputtering was

donefor0.5minutes.

Foulantanalysisandmaterialquantification

Foulantanalysiswaspreparedbydrying1mgoffoulantlayerin120°Covenfor1hour

beforebeing analyzed. Foulant analysis andmaterial quantificationwasperformedduring

theSEManalysisusingDSM962withEDXThermoElectronEDSanalysissystemVoyagerII.

QuantificationwasdoneusingEDX–ZAF,anenergydispersiveX-raysquantificationanalysis

HybridProcesses

97

method with ZAF correction that combines the atomic number (Z), absorbance (A) and

fluorescence(F)oftheelementintothecalculation.

5.6 Resultsanddiscussion

5.6.1 Aluminumremovaloftheproductwaterbythesandfilter

After coagulation and precipitation, aluminum concentrations were inconsistent

betweeneach experiment. The total aluminumconcentrationwas in the rangeof 0.01 to

0.74mg/L.AsshowninthespeciesdistributiondiagraminthepreviousChapter,onfigure

4.1and4.9,Al(OH)4-concentrationisverymuchaffectedbypH,especiallyatpHbetween6

and 8. In this pH range, a slight alteration in pH, showed significantly different species

distribution. Considering that in daily practical use, feed water will not be buffered, this

inconsistentresultshowedthatafurtherstepisobviouslyneededtoassurethataluminum

concentrationsmeetthelimitcriteria.

Theexperimentalresultsshowedthatexcessofaluminumweresuccessfullydecreased

fromthewaterafterpassingthesandfiltertobelow0.2mg/LasshowninFigure5-10.Itis

suggestedthatthedecreaseofaluminumconcentrationbypassageofthesandfilterdoes

not have a linear correlation to the initial aluminum concentration before filtration. It is

presumed that aluminum concentration detected in the filtrate water is in the form of

aluminumionwithdominationofAl(OH)4-.Becausethisaluminumspeciessizeisverysmall,

screening is the last possibility in the removalmechanism. Therefore, it is proposed that

aluminum removal by sand filter was facilitated by chemical bridging where aluminum

coagulant formed heavyweight polymer and acted as a ‘bridge’ between filter grain and

suspended particle or between particles. Overall, the concentration of aluminum in the

waterafterfiltrationmeetsthedrinkingwaterstandard.

Figure5-10Totaldissolvedaluminumconcentrationofproductwaterbeforeandafter

filtrationstep

HybridProcesses

98

5.6.2 Furtherfluorideremovalofproductwaterbysandfilter

Fluorideconcentrationattheoutletofthesandfiltrationprocesswaslowerthanatthe

inlet though itwas not significant. The result in Figure 5-11 shows that there is no linear

correlationbetweenfluorideconcentrationdecreaseanditsinitialconcentrationinthesand

filter process. However, further fluoride removal in the sand filter was supported by

aluminumasdestabilizingagent. Sincealuminumsulfatewasused in theprevious stepas

coagulantordestabilizingagent,itcanbeexpectedthatitfacilitatedfluorideionattachment

tothefiltergrainthroughchemicalbridgingmechanism.

Figure5-11Fluorideconcentrationofproductwaterbeforeandafterfiltrationstep

5.6.3 Turbidityremovalofproductwaterbythefiltrationstep

In theexperiments, itwasclearlyobservedthataftercoagulationandco-precipitation

thereweresuspendedparticlesinproductwater.Figure5-12showstheturbiditytrendsof

the sand filterwith flow rate of 10 L/h. The result shows that the sand filterwas able to

removeturbidity from>1FNUtobelow0.4FNU.Anaverageof78%ofturbidityhasbeen

removed in the system during 60minutes filtration. After 70minutes of filtration, water

turbiditywasincreasing.

Meanwhile,withflowrateof12L/hturbidityremovalswereinthebestperformancein

thefirst30minutes,achievingavaluebelow0.2FNU,asshowninFigure5-13.Thesystem

was able to remove 72% of turbidity. Overall the results show that the turbidity of the

filtrate is <1 FNU foroperation timeof75–90minutes. This result shows that the sand

filterhaseffectivelyremovedthesuspendedsolid.

Figure5-12Turbidityremovalbythesandfilter,withv=0.57m/h.Thepointat0minutes

representstheturbidityofwaterbeforeenteringtheinletofsandfiltermodule.

HybridProcesses

99

During filtration, removal of suspended particles likely occurs inside the filter bed.

Regarding that the aluminum floc size is very small compared to grain diameter of sand

filter, interceptionanddiffusionarethedominanttransportmechanismsintheremovalof

turbidity. It is suggested that Van der Waals interaction and charge attraction between

suspended solid and the surface of sand grains support the attachment. The positively

chargedsuspendedparticlesareboundtonegativelychargedquartzsand.Besidesthat,the

coagulationprocessmightstillbeongoingbeforefiltrationwhilewaterstreamwasflowing

inthepipebeforeitreachedthefilter.Therestofaluminumcoagulantinthewatermight

coagulate the smaller particle facilitated by the shear stress provided by the pump and

furtherencounterchemicalbridging.

Figure5-13Turbidityremovalbythesandfilter,withflowrate=0.69L/h.Thepointat0

minutesrepresentstheturbidityofwaterbeforeenteringtheinletofsandfiltermodule.

Theincreaseofturbidityinsomeofthedatapoints,whichisvisibleinFigure5-12and

Figure 5-13, might be caused by detachment. At the beginning of filtration, during the

attachment,particleswereretainedandaccumulatedontheuppergrainsurface.Thislayer

mightcollapsebecauseof thehydrodynamic forcesduetocontinuouswater flowthrough

themedium. In addition, continuous shear stress generated by the pump also facilitated

detachment of particles from filtermedia. This pressure breaks the attachments and the

flocssotheflocsizebecomesveryfine.Thosefineparticleswerenotretainedbythefilter

andflowingbacktothewaterstream.

Another possibility includes the phenomenon of sedimentation on the bottomof the

filter due to full occupation of upper bed surface by the suspended solid. Because of

continuouswaterflow,thissedimentwasdriftedtotheoutletofwaterflow.This leadsto

theincreaseofturbidityafteraperiodtimeofoperation.However,thisscenariohastobe

investigated furtherbecause in thisexperimental setup, itwasnotpossible tocollect the

suspendedsolidonthesandfiltertocompareitwiththecapacityofthefilter.

HybridProcesses

100

5.6.4 FluorideandAluminumremovalbyUFmembrane

Theremovaloffluoridefromthespikedwaterbycoagulationandco-precipitationwith

aluminumcoagulantsandUFwasfirstinvestigated.After45minutesofsettlingtime,large

particles settled down and could be completely removed from the supernatant. Smaller

particles remained unsettled, resulting in a stable suspension. The average fluoride

concentrationwas1.2mg/L,whiletheaverageconcentrationoftotalaluminumresidue in

thewaterwas1.6mg/L.Asalreadydiscussedinchapter4,coagulationandco-precipitation

process succeeded to remove fluoride to the desired level, however, aluminum

concentrationisabovethedrinkingwaterstandard.Thiswaterwasthenusedasfeedinthe

UFmembraneprocess.

Afterwards, the performance of the ultrafiltration membrane was investigated.

Ultrafiltration experiments were conducted in a dead-end batch unit. Collected product

water in the container was flown to the UF membrane module and was filtered by UF

membrane for30minutes.Experimentsweredonewithworkingpressureof0.5;1.0;1.5;

and2.0barsandflowrateof12L/h.AnalysisofthefiltrateisshowninTable5-2.

Comparedto thealuminumconcentration in the feed,aluminumconcentration in the

membrane outlet is very low. The result shows that 90%of aluminum concentrationwas

removedbythemembranewithaverageconcentrationof<0.2mg/L.Inthecaseoffluoride,

there was further removal by themembrane with working pressure of 1.5 and 2.0 bars.

Comparedtothefluorideconcentrationafterthecoagulationstep,therewas25%fluoride

removalbythemembraneprocess.Therewasnoremarkabledecreaseforother ionssuch

asmagnesium,calciumandsulfate.Amongthe investigatedconditions,ultrafiltrationwith

1.5barsgivesthebesteffect.

Themechanismof fluoride removalduring themembraneoperationwith1.5and2.0

barswassuggestedastheeffectofthesolidlayerformedonthemembranesurface.Dueto

pressureand thedead-endmodeofoperation, thesolid layerwas formed intensivelyand

thissolidlayerformedadenselayeronthemembranesurfaced.Since90%ofthealuminum

wasretainedonthemembrane,thissolidlayerisabletoadsorbfreefluorideinthewater.

Hencefluorideconcentrationinthewaterdecreased.

HybridProcesses

101

Table5-2IonconcentrationsinproductwaterafterfiltrationwithUFmembrane100kDa

withdead-endmode

Calcium(mg/L)

Magnesium(mg/L)

Fluoride(mg/L)

Sulfate(mg/L)

TotalAluminum(mg/L)

Feed* 229 8.3 1.2 538 1.6

0.5Bar 229 8.3 1.2(SD=0.06) 537 0.15(SD=0.01)

1.0Bar 228 8.3 1.2(SD=0) 536 0.17(SD=0.01)

1.5Bar 227 8.3 0.9(SD=0.06) 536 0.17(SD=0)

2.0Bar 230 8.3 0.93(SD=0.06) 542 0.18(SD=0.01)

*Feed=wateraftercoagulationandprecipitation

5.6.5 Permeate-fluxdeclineofultrafiltrationmembranes

Experimentalresultsshowtherearesmalldeviationsineachrepetition,howeveritwas

the consequence of different surface porosity of each membrane. Such observation is

commoninmembranetechnologyparticularlyforexperimentswithasmallmembranearea.

Similarobservationshavebeenreportedinapreviousstudy35.TheresultsshowninFigure

5-14areaveragevalueswithrelativestandarddeviation(RSD)foreachfluxof0.0619barfor

0.5bar;0.0611barfor1.0bar;0.0525barfor1.5barand0.0604barfor2bar.

It is evident from results presented in Figure 5-14 that as the working pressure

increased,thefluxaverageswerealsoincreasingproportionally.Figure5-14alsoshowsthat

ineachoperationthefluxstartedatthelowerrateandafterfewseconds,itincreaseduntil

reachingthemaximum.Theshapeofthemembraneholderplaysanimportantroleonthis

profile.With its longbodyshape,pressurizedwaterneedsfewsecondstoreachuniformly

thesurfaceofthemembrane.Maximumfluxwasachievedwhenpressurizedwaterreached

themembranesurfaceuniformly.

Figure5-14Permeatefluxprofileoftheproductwaterwithdifferentpressureoperationin

theUFfiltrationstepswithdead-endmodebyPESUFmembrane100kDa

Aftermaximumfluxwasachieved,fluxdecreasedslightlyovertime.Thedecreaseofflux

indicatesthattherewasanaddedfiltrationresistancebythefoulantlayer.Thismeans,due

to its greater size compared to themembranepore size, the foulantwas retainedby the

HybridProcesses

102

membraneandformedalayeratthemembranesurfacethat,bytime,decreasedthefluxof

thewater.Besides,atpHabove7, themembraneporesbecomemorenegativelycharged

duetopreferentialanionadsorption24.Inthiscase,anionislesshydratedthancation,and

thismakeanionabletoapproachmembranesurfacecloselybeyonditsplaneofshear.Thus,

the membrane surface acquires negative electro kinetic potential as effect of the anion

presencebeyongitsplaneshear36.Thisreducesthemembraneporesize,whichinturnleads

tofluxdecrease.

Figure5-15Rateofpermeatefluxdeclineinthefirst100secondsaftermaximumfluxhad

beenachievedintheUFfiltrationstepswithdeadendmodebyPESUFmembrane100kDa

The experimental results obtained during UF through PES membranes show that

membrane foulingwasstrongly influencedbyoperationalpressure.Figure5-15showsthe

rateof thepermeate flux decline aftermaximum flux hadbeen achieved.Operationwith

pressureof1.5and2.0barsgeneratefastfluxdecline,whilelowerpressurewith0.5and1.0

barsproducedmoderate fluxdecline.After50 seconds, the fluxwas relatively stable.The

higher pressure resulted in a higher flow rate of filtration. Therefore, more flocs were

rejectedonthesurfaceofthemembraneovertime.

A foulant layer supposedly formed faster at the higherworking pressure than at the

lowerworkingpressure.Inaddition,thesmallerfoulantparticlesgeneratedbyshearstress

of the higher pressure are able to penetrate into themembrane porewhich also caused

fouling of the membrane and decrease of flux. This conclusion is supported by the SEM

image shown in Figure 5-16. It is shown that the membrane pores cannot be observed

becausetheyarecoveredandfilledwithwhiteparticles.Inaddition,thepolymermembrane

willchangeitsstructurewhenitisexposedunderpressure.Thischangeinstructureleadsto

fluxdeclineduetolowervolumeporosityandincreasedmembraneresistance37.

Figure5-16SEMimageofPESUF100kDamembranecross-sectionwithdepositionof

particlesoveristsurfaceafterbeingemployedfor30minutesfiltrationwith2.0barswith

dead-endmode

HybridProcesses

103

Figure 5-16 shows the cross-section of the membrane after being employed for 30

minutes with 2.0 bars of working pressure. Due to the filtration process, the porous

membranesupportingsectionwasdeformedheavilyandblockedbywhiteparticles. Itcan

beclearlyseenalsoinFigure5-17,thattherewasdepositionofwhiteparticlesontopofthe

surface of themembrane. This layer leads to thewater flux decline during the filtration.

Besides, it can be concluded that 2.0 bars of pressure results in a deterioration of the

membranestructure.

Figure5-17SEMimageofPESUF100kDamembranesurfaceafterbeingemployedfor30

minutesfiltrationwith1.0barwithdead-endmodetotallycoveredwithretainedwhite

solidsubstanceswith5000timesmagnification

5.6.6 Foulantanalysisandmaterialquantificationofsubstanceretainedfrommembrane

surface

FoulantanalysisresultsperformedusingEDX–ZAFmethodsareshowninTable5-3and

Figure 5-18. The largest components byweight in the foulant are oxygen and aluminum.

Aluminumandoxygenfoundintheanalysismaycomefromthealuminumspeciesthatwere

suspended in thewater,andaluminumhydroxide fineprecipitate thatalsoactsas further

adsorptionbedof fluoride.This findingmaysupport theproposed formationofaluminum

hydroxideasthemechanismofthefluorideremoval38.

Table5-3EDAXZAFanalysiscompositionpercentageof retainedsubstancesover theUF

PESmembrane100kDasurface,afterbeingoperatedfor30minuteswith2bar

Element Weight%

C 11.9

O 56.14

F 1.44

Mg 0.22

Al 24.46

HybridProcesses

104

Si 1.68

S 1.95

Ca 2.21

Total 100

Figure5-18EDXZAFanalysisspectraofretainedsubstancesfromcoagulationand

precipitationstepovertheUFPESmembrane100kDasurface,afterbeingemployedfor

30minuteswith2.0bars

The appearance of carbon as a significant component in the analysis is probably

originatingfromthemembraneitselfsincethemembraneisformedfrompolyethersulfone

whichcontainscarbonas it isshown inFigure5-4.Fluoridewasalso found inasignificant

amountinthefoulantanalysis.ThisshowsthatAl(OH)3astheprimarylayerformedonthe

surfaceofthemembraneisscavengingfluorideionandwasretainedbythemembrane.It

canbealsosuggestedthatthiswhitesubstanceswasthefinecalciumfluorideprecipitate,as

calcium was also found in significant value in the analysis. In addition, the presence of

calciummightcomefromtheoverstoichiometricadditionofcalciumduringthecoagulation

process. Other components such as silicate and magnesium presumably come from the

wateritself,whilesulfurispresumablyderivedfromtherestofhydrolyzedaluminumsulfate

salt as coagulant and from the raw water. Though, sulfur could also originate from the

membraneitselfsincethePESmembranealsocontainssulfurinitspolymerchains.Overall,

thisfoulantanalysisresultsupportstheresultofwateranalysisthatshowed90%reduction

ofaluminumconcentrationfromtheproductwater.

Figure5-19showsthephysicalformoftheretainedsubstanceontopofthemembrane

surface. The image shows porous white fine substances with irregular shape that totally

coverthemembranesurface.Comparedtothemembraneporesizewichisintherangeof

0.01 to0.1μm(Figure5-6), thewhitesubstance isvery small<0.01μm.Therefore it can

infiltrate the membrane pore and cause an irreversible fouling. Since aluminum-fluoride

compound is known to be colorless, the white color is supposed to be coming from the

calciumcontentintheprecipitate.

HybridProcesses

105

Figure5-19SEMimageoftheretainedsubstancesovertheUFPESmembrane100kDa

surfaceafterbeingemployedfor30minuteswith2.0barswith50,000timesmagnification

5.6.7 Summary

Based on the result of the investigation, hybrid coagulation and co-precipitationwith

sand filter shows a meaningful outcome in the laboratory scale. The hybrid process of

coagulation and co-precipitationwith a subsequent sand filterwas successful to decrease

aluminumconcentrationtoacceptableconcentrationwithaverageof70%.Inadditionthere

isunsubstantialfurtherfluorideconcentrationdecreasewithaverageof13%comparedwith

theconcentrationbeforethefiltrationstep.Moreover,thehybridwasalsoabletoremove

turbidity with average of 70%. The water turbidity, aluminum concentration and fluoride

concentrationintheoutletofthesandfilterfulfilledthestandardsetbytheWHO.However,

applicationofthehybridinthefield,needtobeinvestigated.Upscalingofsandfilteraswell

asuseofhighervolumeofproductwaterfromcoagulationstepisrecommendedinfuture

investigations.

Meanwhile,theinvestigationofhybridcoagulationandprecipitationwithultrafiltration

membranewassuccessfullyachievedwith90%reductionofaluminumconcentrationinthe

waterafter thecoagulationandco-precipitationstepbytheUFmembraneoperation.The

aluminumconcentrationafterthehybridprocessfulfilledtheWHOstandardof0.2mg/L.In

the membrane operation with 1.5 and 2.0 bars, there was 25% further fluoride removal

compared to the fluoride concentration after leaving the coagulation and co-precipitation

process.Overall, thehybridprocessassuresthefulfillmentof thedrinkingwaterstandard,

with fluoride concentrationsof<1.5mg/L.After30minutesof filtrationprocess, thePES

membrane structure was already heavily deformed. The surface of the membrane was

covered and the pores of the membrane were blocked by the white porous aluminum

precipitate.Thisledtofluxdecline.Filtrationprocesswithoperationpressureof1.5and2.0

bars generated strongerpermeate-fluxdecline,while lowerpressurewith0.5 and1.0bar

producedmoderatefluxdecline.

HybridProcesses

106

HybridProcesses

107

5.7 References

(1) Fawell,J.K.;Bailey,K.FluorideinDrinking-Water;WorldHealthOrganization,2006;

pp1–144.(2) Njau,B.;Dahi,E.;Dahi,E.;Mtalo,F.;Mtalo,F.;Njau,B.;Bregnhøj,H.Defluoridation

UsingtheNalgondaTechniqueinTanzania;22ndWEDC…:NewDelhi,1996.(3) Ives,K.J.FiltrationofClaySuspensionsThroughSand.ClayMinerals1987,22(1),

49–61.(4) Agarwal,K.C.;Gupta,S.K.;Gupta,A.B.DevelopmentofNewLowCost

DefluoridationTechnology(KRASS).1999,40(2),167–173.(5) Meenakshi;Maheshwari,R.C.FluorideinDrinkingWaterandItsRemoval.Journalof

HazardousMaterials2006,137(1),456–463.(6) Zouboulis,A.;Traskas,G.;Samaras,P.ComparisonofEfficiencyBetweenPoly-

AluminiumChlorideandAluminiumSulphateCoagulantsDuringFull-ScaleExperimentsinaDrinkingWaterTreatmentPlant.SeparationScienceandTechnology2008,43(6),1507–1519.


(8) Bi,S.;An,S.;Tang,W.;Yang,M.;Qian,H.;Wang,J.ModelingtheDistributionofAluminumSpeciationinAcidSoilSolutionEquilibriawiththeMineralPhaseAlunite.EnvGeol2001,41(1-2),25–36.

(9) Ma,J.F.RoleofOrganicAcidsinDetoxificationofAluminuminHigherPlants.PlantCellPhysiol2000,41(4),383–390.

(10) Hillman,A.AdaptingtheBiosandFiltertoRemoveFluoride:anInvestigationIntoAlternativeFilterMedia;Hillman,A.,Ed.;2007.

(11) Mjengera,H.;Mkongo,G.AppropriateDeflouridationTechnologyforUseinFlouroticAreasinTanzania;DaresSalaam,2002;pp1–10.

(12) Crittenden,J.;Trussell,R.R.;Hand,D.W.;Howe,K.J.;Tchobanoglous,G.WaterTreatmentPrinciplesandDesign,2nded.;JohnWiley&Sons,Inc:NewYork,2005.

(13) Faust,S.D.;Aly,O.M.ChemistryofWaterTreatment;CRC,1998.(14) Kettunen,R.;Keskitalo,P.CombinationofMembraneTechnologyandLimestone

FiltrationtoControlDrinkingWaterQuality.Desalination2000,131(1–3),271–283.(15) Hilal,N.;Al-Abri,M.;Al-Hinai,H.EnhancedMembranePre-TreatmentProcesses

UsingMacromolecularAdsorptionandCoagulationinDesalinationPlants:aReview.SeparationScienceandTechnology2006,41(3),403–453.

(16) Goosen,M.F.A.;Sablani,S.S.;Al-Hinai,H.;Al-Obeidani,S.;Al-Belushi,R.;Jackson,D.FoulingofReverseOsmosisandUltrafiltrationMembranes:aCriticalReview.SeparationScienceandTechnology2005,39(10),2261–2297.

(17) Hu,K.;Dickson,J.M.NanofiltrationMembranePerformanceonFluorideRemovalFromWater.JournalofMembraneScience2006,279(1–2),529–538.

(18) Lhassani,A.;Rumeau,M.;Benjelloun,D.;Pontié,M.SelectiveDemineralizationofWaterbyNanofiltrationApplicationtotheDefluorinationofBrackishWater.WaterResearch2001,35(13),3260–3264.

(19) Dach,H.ComparaisonDesOpérationsDeNanofiltrationEtD’OsmoseInversePourLeDessalementSelectifDesEauxSaumatres:DeL’ÉchelleDuLaboratoireAuPiloteIndustriel,Universitéd'Angers,2008.

HybridProcesses

108

(20) Pontié,M.;Diawara,C.;Lhassani,A.;Dach,H.;Rumeau,M.;Buisson,H.;Schrotter,J.C.Chapter2WaterDefluoridationProcesses:aReview.Application:Nanofiltration(NF)forFutureLarge-ScalePilotPlants.InFluorineandtheEnvironment—Agrochemicals,Archaeology,GreenChemistry&Water;Tressaudd,A.,Ed.;FluorineandtheEnvironment—Agrochemicals,Archaeology,GreenChemistry&Water;Elsevier,2006;Vol.2,pp49–80.

(21) Hagen,K.RemovalofParticles,BacteriaandParasiteswithUltrafiltrationforDrinkingWaterTreatment.Desalination1998,119(1–3),85–91.

(22) Brehant,A.;Bonnelye,V.;Perez,M.ComparisonofMF/UFPretreatmentwithConventionalFiltrationPriortoROMembranesforSurfaceSeawaterDesalination.Desalination2002,144(1–3),353–360.

(23) Lerch,A.FoulingLayerFormationbyFlocsinInside-OutDrivenCapillaryUltrafiltrationMembranes,UniversitätDuisburg-Essen,2008.

(24) Kabsch-Korbutowicz,M.ImpactofPre-CoagulationonUltrafiltrationProcessPerformance.Desalination2006,194(1–3),232–238.

(25) Feng,L.;Wang,W.;Feng,R.;Zhao,S.;Dong,H.;Sun,S.;Gao,B.;Yue,Q.CoagulationPerformanceandMembraneFoulingofDifferentAluminumSpeciesDuringCoagulation/UltrafiltrationCombinedProcess.ChemicalEngineeringJournal2015,262IS-,1161–1167.

(26) Baker,R.W.MembraneTechnologyandApplications,2nded.;Wiley,2004;pp1–545.

(27) AGARWAL,K.;GUPTA,S.;Gupta,A.DevelopmentofNewLowCostDefluoridationTechnology(Krass).WaterScienceandTechnology1999,40(2),167–173.

(28) Ho,W.S.W.;Sirkar,K.K.MembraneHandbook;Chapman&Hall,1992.(29) TerminologyforMembranesandMembraneProcesses(IUPACRecommendation

1996).JournalofMembraneScience1996,120(2),149–159.(30) Cheryan,M.UltrafiltrationandMicrofiltrationHandbook;CRCPress,1998.(31) Bacchin,P.;Aimar,P.;FIELD,R.CriticalandSustainableFluxes:Theory,Experiments

andApplications.JournalofMembraneScience2006,281(1-2),42–69.(32) DIN.DIN38405-4:DeutscheEinheitsverfahrenZurWasser-,Abwasser-Und

Schlammuntersuchung;Anionen(GruppeD);BestimmungVonFluorid(D4);BeuthVerlag,1985.

(33) Mulder,M.BasicPrinciplesofMembraneTechnology;Springer,1996.(34) Hach.DR/890ColorimeterProceduresManual;DR/890;HachCompany,2009;pp1–

614.(35) Susanto,H.;Ulbricht,M.InfluenceofUltrafiltrationMembraneCharacteristicson

AdsorptiveFoulingwithDextrans.JournalofMembraneScience2005,266(1-2),132–142.

(36) Elimelech,M.;Chen,W.H.;Waypa,J.J.MeasuringtheZeta(Electrokinetic)PotentialofReverseOsmosisMembranesbyaStreamingPotentialAnalyzer.Desalination1994,95(3),269–286.

(37) Persson,K.M.;Gekas,V.;Trägårdh,G.StudyofMembraneCompactionandItsInfluenceonUltrafiltrationWaterPermeability.JournalofMembraneScience1995,100(2),155–162.

(38) Choi,W.-W.;Chen,K.Y.TheRemovalofFluorideFromWatersbyAdsorption.Journal(AmericanWaterWorksAssociation)1979,71(10),562–570.

109

Chapter6 GeneralConclusionandOutlook

In this work, hybrid techniques as alternativemethods in removing elevated fluoride

concentration in drinking water resources without causing excess of aluminum

concentration and high turbidity in treatedwater have been evaluated. The study on the

chemistry of coagulation and co-precipitation of fluoride with aluminum sulfate as the

basicmethodforthehybridshassucceededtoachieveitsgoaltofostertheunderstanding

of the aqueous chemistry for fluoride removal. Beyond that, this understanding provides

usefulinformationtocontrolthequalityofproductwaterespeciallywithintentiontoavoid

elevated aluminum concentration in the product water and to minimize the use of

aluminumcoagulantintheprocess.Theanalysisoftheproductwatershowedthat100mg/L

aluminumsulfatewereneededtodecrease fluorideconcentration from10mg/L tobelow

1.5 mg/L, which corresponds to a molar ratio of Al:F of 7:1. For a lower initial fluoride

concentration of4mg/L17mg/Laluminumsulfatewereneeded toachieve reduction to

below 1.5mg/L, thus a lower molar ratio of 4:1. Both aluminum concentrations are far

below previously reported studies, where 16 to 181 mg/L for treating raw water with

fluoride levels of 2 to 8 mg/L had been used 1. Lower amounts of added aluminum are

definitelypreferablebecause thiswill reduce theriskofelevatedaluminum in the treated

water.Inadditionitisalsobetterfromaneconomicalpointofview.However,moredifferent

initialfluorideconcentrationshavetobeexamined.Thesedatacouldcompletethecurrent

findingsoalineofcorrespondingvaluescouldbepreciselydescribedinagraph.Thisgraph

would help in the quick estimation in aluminum dosage to treat different fluoride

concentrationsinrawwater.

Fluorideremoval istakingplacebyco-precipitation.Becausethisprocess isverymuch

pHdependent,pHisthemostimportantparametertogovernthisprocess.Theremovalwas

mostlydependingontheformationofAl(OH)3thatfluoridewillco-precipitatewith.Neutral

pHwas the ideal condition for this process, because acidic pH hindered the formation of

Al(OH)3,whilebasicpHenhancedtheformationofAl(OH)4-species.Speciesdiagramsfora

system that contains 10 mg/L fluoride and 100 mg/L aluminum calculated in this study

support this. These diagrams also explain the extremely important role of pH on fluoride

GeneralConclusionandOutlook

110

concentrationsaftercoagulationandco-precipitation.ControlingthepHduringtheprocess

wasthemostimportantfindingtocontrolthequalityofproductwater.Subsequenttothis

workinthefuture,additionalbuffertothesystemmightprovidenewinformationforbetter

systemunderstandingbecausebufferingthesystemwilleliminatetheeffectofotherwater

parametersonpH.

Inthecoreofthisthesis,hybridprocesscoagulationandco-precipitationwithasand

filterhasbeenstudiedtodealwithexcessresidueofaluminumintheproductwater.Inthe

laboratory scale, the hybrid coagulation and precipitation with a sand filter shows that

turbiditycanberemovedsuccesfully,whilealuminumconcentrationcanbedecreasedwith

no linearcorrelation to its initial concentration.Additionofaluminumsulfatecoagulant in

excessamountinthissystemassuredtheformatlonofchemicalbridgingasthemechanism

oftheremovalofaluminumresiduebythesandfilter2.Thepresentedstudyshowedthatthe

overall hybrid performance strongly depends on the coagulation - co-precipitation step.

Maintaining stable and good coagulation - co-precipitation step will ensure the optimum

resultofthehybrid.Utilizationofactivatedcarbonasfiltermaterialincombinationwiththe

quartz sandwouldbe an alternativeextension to theworkbecause activated carbonwas

reported to be able to remove dissolved organic aluminum3. Furthermore, up-scaling the

sizeoffilterandinvestigatingtheeffectofdifferentwaterflowratewillberequiredinthe

viewoftechnicalapplicationinthefield.

Finally, hybrid coagulation and precipitation with ultrafiltration membrane using a

deadendmodeandPESasmembranematerialwasinvestigated.Ultrafiltrationaloneisnot

abletoremovefluorideinwater.However,theultrafiltrationmembranewasthoughttobe

sufficient to remove colloidal aluminum-fluoride complexes that could not be removed

during sedimentation after coagulation and co-precipitation process. This hypothesis was

developedbasedonthealuminum-fluoridecomplexsize4.Thishybridisthoughttobemore

cost effective rather than using energy-intensive-RO or even NF that is able to remove

fluoridedirectlybecausecoagulation-UFenergyconsumptionwasreportedtobelowerthan

NF5.

Analysis of theproductwater of thehybrid at laboratory scale confirmed90%of the

aluminum residues from coagulation and co-precipitation process were retained in the

GeneralConclusionandOutlook

111

membranethatwasoperatedatseveraldifferentlowpressures(atmax.2.0bar).Thestudy

also demonstrated that lower pressure provided smaller flux declination. It was also

concluded that theuseof anultrafiltrationmembrane assured that removal of aluminum

andturbiditymetthelimitsetbytheWHO.

Usingdifferentmembranematerialsuchasceramicmembranewouldbeanalternative

in the future investigation to findamoredurablemembranematerial for sustainableuse.

Afterall,despitethishybridisquitehardtobepracticedinhouseholdlevel,itsapplicationin

thewaterworksatsmallcommunitylevelorinsmall-scaleindustrythatdealswithfluoride

containingwastewaterwouldbeaninterestingtaskinfuturestudies.

6.1 References


(2) Faust,S.D.;Aly,O.M.ChemistryofWaterTreatment;CRC,1998.(3) Polasek,P.;Muti,S.CationicPolymersinWaterTreatment:Part1:Treatabilityof

WaterwithCationicPolymers.WaterSA2002,28(1),69–82.(4) Ives,K.J.FiltrationofClaySuspensionsThroughSand.ClayMinerals1987,22(1),

49–61.(5) Machenbach,I.DrinkingWaterProductionbyCoagulationandMembraneFiltration,

NorwegianUniversityofScienceandTechnology,2007,pp1–250.

ListofAbbreviations

112

ListofAbbreviations

Å Ångström

AEM AnionExchangeMembrane

Anom Nominalfiltrationarea

aq Aqua

BET Brunauer–Emmett–Teller

BSF BioSandfilter

C Celcius

CC Calciumchloride

cm Centimeter

CP ConcentrationPolarization

D Diameter

DD DiffusionDialysis

DIN DeutschesInstitutfürNormung

DMEA Dimethylaminoethanol

DOC DissolvedOrganicCarbon

E Energy

EC Electrocoagulation

ED Electrodialysis

EDX EnergyDispersiveX-ray

ESEM EnvironmentalScanningElectronMicroscope

ListofAbbreviations

113

FAP Fluorapatite

FNU FormazinNephelometricUnit

FTU FormazinTurbidityUnit

g Gram

h Hour

HAP Hydroxyapatite

HF Hydrogenfluoride

IC IonChromatography

ICP-OES Inductivelycoupledplasma–opticalemissionspectroscopy

ISE IonSelectiveElectrode

K Kelvin

kDa Kilodalton

kg Kilogram

KJ KiloJoule

kW KiloWatt

L Litre

M Molar

Max. Maximal

MF MicroFiltration

mg Milligram

µg Microgram

min Minutes

ListofAbbreviations

114

mL Mililitre

µµ Micrometer

MSP MonosodiumPhosphate

mV Millivolt

n.a. Notapplicable

NEERI NationalEnvironmentalEngineeringResearchInstitute

NF NanoFiltration

nm Nanometer

NMWCO NominalMolecularWeightCut-Off

NTU NephelometricTurbidityUnit

PAC Polyaluminumchloride

PES Polyethersulfone

PET Polyethyleneterephthalate

PS Polysulfone

PTFE Polytetrafluoroethylene

pzc PointofZeroCharge

RO ReverseOsmosis

rpm RotationperMinute

RSD RelativeStandardDeviation

S Second

SD Standarddeviation

SEM ScanningElectronMicroscope

ListofAbbreviations

115

TISAB TotalIonicStrengthAdjustmentBuffer

TMA Trimethylamine

TMP Trans-MembranePressure

TX Texas

UF UltraFiltration

USA UnitedStatesofAmerica

V Volt

WHO WorldHealthOrganization

XRF X-RayFluorescence

116

ListofPublications

Publicationsinpeer-reviewedjournals

-

Posterpresentations

• Lestari,R.B.;Schmidt,T.C.ObservationofFluorosisCases inSeveralVolcanicAreas in

JavaIsland;127.GDNÄVersammlung,Göttingen,2012

• Lestari, R. B.; Liu, Y.; Nie, Y.; Wermeckes, B.; Schmidt, T. C. Fluoride Removal From

Drinking Water with Electrocoagulation by Employing Al-Pt Electrodes, 243rd ACS

NationalMeeting&Exposition,SanDiego,2012

• Lestari, R. B.; Herzog, R.; Nahrstedt, A.; Schmidt, T. C. Insight in the Coagulation and

Precipitation Process in the Nalgonda Technique for Drinking Water Defluoridation,

International Conference and Exhibition on Water in the Environment, Stellenbosch,

2009

• Lestari, R. B. Laboratory-scale studyonhybridmembrane system for defluoridationof

drinkingwater,GDCHWasser2009,Stralsund,2009

• Susanto,H.,Lestari,R.B.,Ulbricht,M.Mechanismandcontrolofultrafiltrationfoulingby

hydrophilicorganicmixtures,12thAachenMembraneColloquium,Aachen,2008

Oralpresentations

• Lestari, R.B.,“Raising Awareness Concept on Water Pollution Issue in Upper Citarum

River”,presentationatDAAD–ITTAlumniSeminar,4-10.Nov.2007inVientiane,Laos

117

CurriculumVitae

DerLebenslaufistinderOnline-VersionausGründendesDatenschutzesnichtenthalten.

118

Erklärung

Hiermitversichereich,dassichdievorliegendeArbeitmitdemTitel

„Defluoridationofdrinkingwaterbyhybridcoagulationandfiltrationprocess”

selbstverfasstundkeineaußerdenangegebenenHilfsmittelnundQuellenbenutzthabe,

unddassdieArbeitindieseroderähnlicherFormnochbeikeineranderenUniversität

eingereichtwurde.

Bremen,08December2016

Documents

Defluoridation of drinking water by hybrid coagulation and