fundamental aspects of sludge characterization
TRANSCRIPT
Fundamental aspects of sludge characterization
Herwijn, A.J.M.
DOI:10.6100/IR458461
Published: 01/01/1996
Document VersionPublisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)
Please check the document version of this publication:
• A submitted manuscript is the author's version of the article upon submission and before peer-review. There can be important differencesbetween the submitted version and the official published version of record. People interested in the research are advised to contact theauthor for the final version of the publication, or visit the DOI to the publisher's website.• The final author version and the galley proof are versions of the publication after peer review.• The final published version features the final layout of the paper including the volume, issue and page numbers.
Link to publication
Citation for published version (APA):Herwijn, A. J. M. (1996). Fundamental aspects of sludge characterization Eindhoven: Technische UniversiteitEindhoven DOI: 10.6100/IR458461
General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.
• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal ?
Take down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.
Download date: 29. Mar. 2018
Fundamental aspects of sludge characterization
A.J.M. Herwijn
\
FUNDAMENTAL ASPECTSOF SLUDGE CHARACTERIZATION
FUNDAMENTAL ASPECTS OF SLUDGE CHARACTERIZATION
PROEFSCHRIFT
ter verkrijging van de graad van doctor aan de Technische Universiteit Eindhoven, op gezag van de Rector Magnificus, prof. dr. J .H. van Lint, voor een commissie aangewezen door het College van Dekanen in het openbaar te verdedigen op
woensdag 17 april1996 om 16:00 uur
door
AREND JOHANNES MARIA HERWIJN
Geboren te Eindhoven
Dit proefschrift is goedgekeurd door de promotoren:
prof.dr.ir. P.J.A.M. Kerkhof
prof.dr. W.G.M. Agterof
en de copromotor:
dr.ir. W.J. Coumans
ter herinnering aan mijn vader,
aan mijn moeder
DANKWOORD
Vele personen hebben een belangrijke bijdrage geleverd aan de totstandkoming van het
proefschrift. Ik wil ten eerste mijn promotor,· Piet Kerkhof en copromotor Jan
Coumans bedanken voor het in mij gestelde vertrouwen en de aangeboden kans om het
verrichtte onderzoek vast te leggen in een proefschrift. De financiers van het onder
zoek en de leden van de begeleidingscommissie dank ik voor de ondersteuning.
Mijn naaste collega's Erik La Heij en Paul Janssen bedank ik voor de prettige en
uitstekende samenwerking. Paul Janssen bedank ik speciaal voor het uitvoeren van een
groot aantal experimenten en het vervaardigen van vele tekeningen die zijn opgenomen
in dit proefschrift.
Mijn waardering gaat ook uit naar de afstudeerders Paul Dohmen, Lotte Boon, Juul
Dzermans, Diederic van Dijke, Annemiek van derZandeen Moshe van Berlo, die alle
een essentiële bijdrage hebben geleverd aan het onderzoek.
De technici wil ik bedanken voor het ontwerpen en fabriceren van meetopstellingen.
Mijn kamergenoten, Gerben Mooiweer en Ton van der Zanden, bedank ik voor de
altijd prettige werksfeer.
De samenwerking met de vakgroep Colloïdchemie en Thermodynamica en speciaal
met de heer van Diemen heb ik altijd als zeer positief ervaren.
Mijn moeder en broer bedank ik tenslotte voor de morele ondersteuning in de
afgelopen jaren.
SUMMARY
Sewage sludge, which is a suspension with a dry solids content of 3 to 4 wt%,
originates from the purification of waste water. In the Netherlands the annual produc
tion of communal sewage sludge amounts to about 300 million kilograms of dry
solids. Before depositing the sludge, mechanica! dewatering is used. Chamber filter
presses, belt presses, and centrifuges are utilized to dewater sewage sludges. The
importance of mechanical dewatering will only increase. The deposit of sewage sludge
in agriculture and the preparation of black earth and compost will strongly decrease in
the near future to meet the requirements of the limitations of heavy roetal concentrati
ons and the allowed dosages. Moreover, one strives to reduce the deposition of
sewage sludge on tipping sites owing to insufficiently available space in the Nether
lands. The sewage sludge production will increase in the near future due to the growth
of the population, the higher degree of purification, and the introduetion of chemical
dephosphatation.
A better understanding of the mechanical dewatering process can lead to a better
preparation for these technological and social developments. A better insight into the
fundamental aspects of sewage sludge dewatering can be obtained by characterizing
sewage sludge. Characterization methods must be used which are of relevanee for a
better understanding of the mechanical dewatering behaviour.
In this thesis the determination of a great number of dewatering properties of sludge
and sludge cake determined with both existing and newly developed measuring
methods are discussed. Moreover, an attempt bas been made to interrelate the various
dewatering characteristics. Four different sludges originating from four different waste
water treatment plants have been studied.
Water is bound to sludge particles in different ways. Two methods were used to study
the solid-to-water bond strength: thermal analysis to determine isothermal drying
curves and the measuring of water vapour sorption isotherms at different temperatures.
The isotherms measured at different temperatures can be described very well with the
S-shaped temperature-dependent G.A.B. equation. With both methods it is possible to
determine the water bond enthalpy as a function of the sludge cake moisture content.
Both methods show that at decreasing moisture content the bond enthalpy is initially
zero ('free water') and differs significantly from zero at sample moisture contents
smaller than 0.3 to 0.6 kg water per kg dry solids ('bound water'). The sludge origin,
type of flocculant, and flocculant dosage do not appear to influence this critical
moisture content. On the basis of these measurements, one may conclude that the
maximum feasible dry solids content in a mechanica! dewatering process amounts to
about 65 to 75 wt%. Obviously, the amount of free water which has been entrapped
during filter cake formation dictates the attainable fina1 dry solids content in practice.
A great number of sludge dewatering characteristics at microscale and macroscale has
been determined. In order to determine these characteristics both existing and newly
developed measuring techniques have been used. Microproperties that have been
determined are composition (cluster properties like dry solids content, ash content, pH
and electrical conductivity), zeta potential, partiele size distribution, and rheological
properties. Macroproperties that have been determined are the specific cake resistance,
end dry solids content, porosity, permeability, capillary suction time, concentration
ferric ions and polyelectrolyte in the filtrate. A lot of dewatering characteristics of
sludge and sludge cake have been measured with a newly developed apparatus: the
filtration-expression cell. Micro- and macroproperties were determined as a function
of the flocculant dosage. The addition of flocculants improves the dewatering process.
Three types of flocculants per sludge type have been studied: the polyelectrolyte Röhm
KF975, the polyelectrolyte used at the sludge treatment plant, and ferric chloride in
combination with lime. It appears that the flocculant dosage has a large impact on the
dewaterability of sewage sludges. At the optimum flocculation condition (dosage and
mixing intensity) some characterization parameters show a maximum or a minimum:
minimum specific cake resistance, minimum vacuum suction time, minimum CST
value, minimum iron content in the filtrate, maximum dry solids content of the cake,
maximum permeability, maximum median floc diameter, and maximum degree of
thixotropy. At the optimum flocculation conditions, sewage sludges can ·be dewatered
at the highest rate and the highest dry solids content is reached.
The conventional Capillary Suction Time apparatus is often used in modem practice to
determine the sludge dewaterability. However, this apparatus has some disadvantages.
Filter paper, which is used as capillary medium, often differs in structure which
results in different porosities and permeabilities. These differences in filter paper lead
to a bad reproducibility of the measurements. Another disadvantage is the determinati
on of the position of the liquid front at only two times. Consequently, only little
information on the dynamic dewatering process can be obtained. A physical-mathema
tical model has been developed for the apparatus. With the model the position of the
liquid front can be calculated as a function of time. The influence of the various
process parameters on this relationship has been investigated. A modified CST
apparatus which registers the position of the liqoid front as a function of time has been
developed. Ceramics is osed as capillary medium. Ceramics is an isotropie materiaL
As a resolt the reprodocibility of the measorement was improved. The resolt of fitting
the model on the experimental data yields the specific cake resistance of the sludge
cake. Both flocculated and unflocculated slodges, can be studied. The specific cake
resistance is an intrinsic valoe, as opposed to the capillary soction time which on
among other things depends on the dry solids concentration of the suspension.
Before mechanical dewatering, sewage slodge is conditioned with flocculants.
Flocculants that are applied in practice are polyelectrolytes and ferric chloride in
combination with lime. Insight into the varioos destabilization mechanisms involved in
slodge flocculation may lead to a better fundamental onderstanding of the dewatering
process. An important parameter is the zeta potenrial which has been determined with
electroacoustophorese. The dominant floccolation mechanism indoeed by adding ferric
chloride to a slodge suspension is specific adsorption of mainly monovalent and
divalent positively charged hydrolysed ferric ions at active sites of the surface of
negatively charged slodge particles. Specific adsorption results in partiele charge
reversal from negative to positive. Typically, the charge reversal is sharp and
discontinoous. The discontinuity is attriboted to the presence of an undesired electroly
te in the suspension. Specific adsorption is energetically favoorable within a eertaio
pH range. Electroacoostophoresis seems not very suitable to study the mechanism of
flocculation indoeed by polyelectrolytes. The measuring probe cannot detect the
formed flocs. However, charge reversal was observed. The main floccolation mecha
nism indoeed by adding cationic polyelectrolytes is charge neotralization.
SAMENVATTING
Zuiveringsslib, bestaande uit een waterige suspensie met 3 à 4 gewichtsprocenten
drogestof, ontstaat bij de zuivering van afvalwater. De jaarlijkse produktie aan
communaal slib in Nederland bedraagt circa 300.000 ton op drogestofbasis. Alvorens
het slib af te zetten, vindt er mechanische ontwatering plaats. De mechanische
ontwatering gebeurt in Nederland met behulp van kamerfilterpersen, zeefbandpersen
en centrifuges. Het belang van de mechanische ontwatering zal naar verwachting
alleen maar toenemen. De afzet van slib in de landbouw en de mogelijkheden het slib
te verwerken tot compost of zwarte grond zal in de toekomst sterk verminderen ten
gevolge van de normen die er gesteld worden aan de toegestane verontreinigingen met
zware metalen enerzijds en aan de slibdosering anderzijds. Bovendien wordt vanwege
ruimtebeslag zoveel mogelijk gestreefd naar het reduceren van het te storten volume.
De produktie van zuiveringsslib zal onder andere door de groei van de bevolking, een
verdergaande zuiveringsgraad en de invoering van chemische defosfatering verder
toenemen. Door een beter begrip van het mechanisch ontwateringsproces van slib kan
op adequate wijze worden ingespeeld op deze maatschappelijke en technologische
ontwikkelingen. Meer inzicht in de fundamentele aspecten van het slibontwateringspro
ces kan worden verkregen door het zuiveringsslib te karakteriseren. Hierbij dienen
karakteriseringsmetboden te worden gebruikt die relevant worden geacht voor het
ontwateringsgedrag.
Dit proefschrift gaat in op het vastleggen van een groot aantal karakteriseringsparame
ters van slib en slibkoek die is bepaald met zowel bestaande als nieuw ontwikkelde
meetmethodieken. Tevens is getracht de verschillende parameters met elkaar in
verband te brengen. Er zijn vier verschillende slibben afkomstig van vier verschillende
rioolwaterzuiveringsinrichtingen onderzocht.
Water is op verschillende manieren gebonden aan slibdeeltjes. De slib-water binding is
onderzocht met twee methoden: thermische analyse voor de bepaling van isotherme
droogcurven en het meten van waterdamp-isothermen bij verschillende temperaturen.
De waterdamp-isothermen gemeten bij verschillende temperaturen blijken goed te
kunnen worden beschreven met de S-vorrnige temperatuurafhankelijke GAB vergelij
king. Met beide methoden is het mogelijk om de bindingsenthalpie als functie van het
vochtgehalte van een slibmonster te bepalen. Beide methoden tonen aan dat bij
afnemend vochtgehalte de bindingsenthalpie van slib/water aanvankelijk nul is ("vrij
water") om vervolgens sterk toe te nemen. bij vochtgehaltes lager dan ca. 0,3-0,6 kg
water per kg drogestof ("gebonden water"). De slibsoort, het flocculanttype en de
tlocculantdosering lijken geen invloed te hebben op dit kritieke vochtgehalte. Op grond
van deze metingen kan worden geconcludeerd dat in een mechanisch ontwateringspro
ces drogestofgehalten kunnen worden bereikt van 65 tot 75 gewichtsprocenten. In de
praktijk worden echter drogestofgehalten bereikt van 20 tot 30 gewichtsprocenten.
Blijkbaar is de hoeveelheid vrij water dat wordt ingesloten tijdens de filterkoekvor
ming ("interstitieel water") bepalend voor het te bereiken drogestofgehalte in de
praktijk.
Een groot aantal ontwateringskarakteristieken zowel op micro- als macroschaal is
bepaald van vier slibben. Voor het bepalen van deze karakteristieken zijn zowel
bestaande als nieuw ontwikkelde meetmethodieken gebruikt. Micro-eigenschappen die
zijn bepaald zijn samenstelling (clnstergrootheden zoals drogestofgehalte, gloeirest, pH
en electrische geleidbaarheid), zêta-potentiaal, deeltjesgrootteverdeling en reologische
eigenschaPpen. Ontwateringseigenschappen (macro-eigenschappen) die zijn gemeten
zijn de specifieke filtratieweerstand, einddrogestofgehalte, porositeit, permeabiliteit,
capillaire afzuigtijd (CST), concentratie ijzerionen en polyelectrolyt in het filtraat.
Vele ontwateringskarakteristieken van een slibkoek zijn bepaald met behulp van een
nieuw ontwikkeld meetapparaat: de filtratie-expressieceL De micro- en macroeigen
schappen zijn bepaald als functie van de dosering flocculant. Toevoeging van floccu
lanten heeft tot gevolg dat het ontwateringsproces sterk wordt verbeterd. Er zijn drie
typen flocculant per slibsoort getest: het polyelectrolyt Röhm KF975, het polyelectro
lyt dat wordt gebruikt bij de betreffende slibverwerkingsinstallatie en ijzerchloride in
combinatie met kalk. Het blijkt dat de flocculantdoseriitg een grote invloed heeft op de
ontwaterbaarbeid van zuiveringsslib. Er blijkt een optimale flocculatieconditie
(dosering en mengintensiteit) te bestaan waarbij een aantal parameters een maximum
of minimum vertoont: minimale specifieke filtratieweerstand, minimale afzuigtijd,
minimale CST waarde, minimale ijzerionenconcentratie in het filtraat, maximale
einddrogestofgehalte, maximale permeabiliteit, maximale mediaan vlokdiameter en
maximale thixotropie. Bij de optimale flocculatiecondities wordt het slib met de
hoogste snelheid ontwaterd en wordt het hoogste drogestofgehalte bereikt.
Het conventionele capillaire afzuigtijd apparaat wordt veel toegepast in de huidige
praktijk voor de bepaling van slibontwateringseigenschappen. Echter dit apparaat kent
enkele nadelen. Het filtreerpapier, dat gebruikt wordt als capillair medium, verschilt
vaak in structuur hetgeen resulteert in een verschillende porositeit eu permeabiliteit.
Dit verschil in structuureigenschappen leidt tot een slechte reproduceerbaarbeid van de
metingen. Een ander nadeel is dat de positie van het vloeistoffront slechts wordt
gemeten op twee verschillende tijdstippen, waardoor er slechts geringe informatie
wordt verkregen over het dynamische ontwateringsproces. Er is een fysisch-mathema
tisch model ontwikkeld voor het CST -apparaat. Met dit model kan de positie van het
vloeistoffront worden berekend als functie van de tijd. De invloed van de verschillende
procesparameters op deze relatie kan worden onderzocht. Een gemodificeerd eST
apparaat, waarin de positie van het vloeistoffront als functie van de tijd wordt
geregistreerd, is ontwikkeld. Als capillair medium is keramiek gebruikt. Het gebruikte
keramiek is een isotroop materiaal, hetgeen de reproduceerbaarbeid van de metingen
verbetert. Door het model toe te passen op het experimentele resultaat kan de
specifieke filtratieweerstand van de slibkoek worden berekend. Zowel geflocculeerde
als ongeflocculeerde monsters kunnen worden onderzocht De specifieke filtratieweer
stand is een intrinsieke ontwateringskarakteristiek, dit in tegenstelling tot de capillaire
afzuigtijd die afhangt van onder andere de drogestofconcentratie in de snspensie.
Vóór de mechanische ontwatering wordt het slib geconditioneerd met behulp van
flocculanten. Flocculanten die worden toegepast in de praktijk zijn ijzerchloride in
combinatie met kalk en polyelectrolyten. Inzicht in de verschillende mechanismen die
een rol spelen bij de flocculatie van zuiveringsslib leidt mede tot een beter fundamen
teel begrip van het slibontwateringsproces. Een belangrijke parameter hierbij is de
zêta-potentiaal, die bepaald is met behulp van electroakoestoforese. Het dominante
flocculatiemechanisme geïnduceerd door toevoeging van ijzerchloride aan een
slibsuspensie is specifieke adsorptie van voornamelijk monovalent en divalent positief
geladen ijzerhydroxydecomplexen aan actieve plaatsen op het oppervlak van de
negatief geladen slibdeeltjes. Specifieke adsorptie leidt tot ladingsomkeer van de
slibdeeltjes van negatief naar positief. Uit het verloop van de zêta-potentiaal als functie
van de pH van de slibsuspensie blijkt dat de ladingsomkeer scherp en discontinu is en
plaatsvindt bij een pH gelijk aan 6, het iso-electrisch punt van het systeem slib/ijzer
chloride. De discontinuïteit is te wijten aan de aanwezigheid van een ongewenst
electrolyt in de suspensie. Specifieke adsorptie is energetisch gunstig binnen een
bepaald pH-gebied. Electroakoestoforese lijkt niet erg geschikt voor de bestodering
van het flocculatiemechanisme geïnduceerd door polyelectrolyten. De meetsonde kan
de gevormde vlokken niet goed detecteren. Ladingsomkeer is echter wel waarge
nomen. Het belangrijkste flocculatiemechanisme geïnduceerd door toevoeging van
katione polyelectrolyten is ladingsneutralisatie.
CONTENTS
1 INTRODUCTION
2 WASTE WATER PROCESSING 7
2.1 Introduetion 7
2.2 The activated sludge process 8
2.3 Sludge stabilization 10
2.4 Thickening 12
2.5 Mechanica! dewatering 13
2.6 Description of various types of sewage s1udge 17
2.7 History and origin of studges investigated 19
2. 7.1 The Eindhoven waste water treatment plant 19
2.7.2 Waste water treatment plant 'Amsterdam-Oost' 20
2. 7.3 Oxidation ditch system 'Veghel-Uden' 21
2.7.4 The oxidation ditch system 'De Hooge en Lage Zwaluwe' 22
3 WATER BINDING IN SEW AGE SLUDGE 25
3.1 Introduetion 25
3.2 The presence of water in sewage sludge 26
3.3 Isothermal drying curves (TGA/DTA) 27
3.3.1 The TGA-DTA drying model 29
3.3 .2 Evaporation of pure water and the calibration of the DT A probe 31
3.3.3 Isothermal drying of a sludge cake 35
3.3.4 Experimental results 37
3.4 Water vapour sorption isotherms 41
3 .4. 1 Introduetion 41
3 .4. 2 Sorption roodels 43
3 .4. 3 Metbod of saturated salt solutions 46
3.4.4 Coulter Omnisorp 100 54
3.5 Conclusions 59
4 SLUDGE DEW A TERING CHARACTERISTICS 61
4.1 Introduetion 61
4.2 Plan of cbaracterization research 62
4.3 Composition 63
4.4 Piltration and expression 66
4.4.1 The filtration-expression cell 67
4.4.2 Modified Piltration Test 75
4.4.3 Conventional CST apparatus 78
4.4.4 The compression-permeability cell 80
4.5 Amount of iron in filtrate 84
4.6 Polyelectrolyte concentration in filtrate 86
4.7 Partiele size distribution 87
4.8 Rheological properties 95
4.9 Conelusions 100
5 MODIFIED CAPILLARY SUCTION TIME (CST) APPARATUS 103
5.1 Introduetion 103
5.2 A theoretical model descrihing the liquid flow in a CST apparatus 104
5.3 Parameter studies 107
5.4 Modified CST apparatus 110
5.5 Experimental results and discussion 113
5.6 Conelusions 119
6 FLOCCULATION BEHA VIOUR OF SEWAGE SLUDGE 121
6.1 Introduetion 121
6.2 The electrical double layer around a spherical sludge partiele 122
6.3 Colloidal stability in terms ofthe electrical double layer 128
6.4 Specific adsorption flocculation by metal coagulants 131
6.5 Polymerie adsorption flocculation 133
6.6 Experimental technique to determine the ESA signal and zeta
potential 137
6.7 Results and discussion 142
6.8 Model to describe hydrolyzable metal ion adsorption at the sludge
solid-water interface 154
6.9 Conelusions 159
7 CONCLUDING REMARKS AND PERSPECTIVES 161
NOTATION 165
LITERATURE 173
Appendix 1 : Process schemes and/or maps of the waste water and sludge 185
treatment plants
Appendix 2: Output ofMAPLE program 191
Appendix 3 : W orking scheme of sludge characterization 193
Appendix 4 : Shift of the absorption maximum of cobaltphtalocyanine due to 195
increasing added amounts of polyelectrolyte
Chapter 5 has been publisbed in Ind. Eng. Chem. Res., vol. 34, no. 4, pp 1310-1319,
1995.
1 INTRODUCTION
Sewage sludge dewatering is an important step in waste water treatment. Sewage
Slûdge is a rest product of the waste water treatment process, and consists of settleable
solids with the addition of the waste activated sludge generated in the biologica!
treatment stage.
In 1991, about 330 million kilograms of sewage sludge dry solids were produced in
the Netherlands. 25 percent of the total sludge solids removed was used as fertilizer in
agriculture, 20 percent in the preparation of compost and black earth, 51 percent for
landfills and 4 percent was incinerated [CBS, 1993]. In the near future, an increase of
the sludge production in the Netherlands is to be expected due to the increase in
population in the Netherlands, the completion of new waste water treatment plants,
and the introduetion of dephosphatation in 1995.
At the beginning of the seventies undewatered sewage sludge was nsed as a useful
fertilizer in agriculture. Due to the surplus of animal fertilizers and the more severe
legislation on the use of sewage sludge as fertilizer in agriculture at the end of the
seventies, mechanica! dewatering became more important. Nowadays, before its
ultimate disposal a large part of the total sewage sludge produced in the Netherlands is
dewatered. Different dewatering methods are used: mechanical dewatering by belt
presses, chamber filter presses and centrifuges (see section 2.5), and natural dewa
tering in lagoons, buffers and drying beds. The amounts of sewage sludge dewatered
mechanically and naturally in 1991 were 196 million kg and 11 million kg of dry
solids, respectively. The remaining part of the total sewage sludge produced (122
miltion kg of dry solids) was not dewatered at the waste water treatment plant [CBS,
1993].
Disposal of (dewatered) sewage sludge in agriculture and in the preparation of
compost and black earth have to meet the requirements of the Dutch Ministries of
Agriculture & Fisheries, and for Housing, Regionat Development, and the Environ
ment (V.R.O.M. 1). This implies on the one hand limitation of the heavy metal
concentrations in sewage sludge, and restrietion of the sludge dosage per hectare on
the other. The most occurring heavy metals which are present in sewage sludge are
copper, chromium, zinc, lead, cadmium, nickel, mercury and arsenic. The maximum
sewage sludge dosage on arabie land is equal to 2 tons of dry solids per hectare per
1 V.R.O.M. Volkshuisvesting Ruimtelijke Ordening en Milieubeheer
2 Cbapter 1
year. The maximum dosage on grassland is restricted to 1 ton of dry solids per hectare
per year. Disposal of sewage sludge in the North Sea is forbidden.
The amount of sludge disposed on tipping sites will be restricted before long. A dry
solids content of 40 wt% is required to dump sewage sludge on tips. Deposition of
sewage sludge on dumping sites bas to be restricted due to insufficiently available
space in the Netherlands, and the presence of heavy metals in sewage sludge.
The operaring costs of sewage treatment in 1991 amounted to 887 million guilders for
the plants and 229 million guilders for the transport systems [CBS, 1993]. Sewage
sludge handling is responsible for 20 to 50% of the total operaring costs. Sludge
disposal costs amount to 15% of the total running costs.
The current policy is to strive for 1) better controlling the mechanica! dewatering
process, 2) reducing volume and mass of sewage sludge and/or 3) more rapid
dewatering in smaller equipment, and 4) lowering annual costs of the total sludge
processing. Because of the increasing urbanization in the Netherlands, smaller waste
water treatment plants are required. As a consequence, new technologies (e.g. smaller
dewatering systems) have to be developed.
Sludge volume rednetion can be reached in different ways :
1. Rednetion of the production of sewage sludge [Eikelboom, 1993].
2. Impravement of the dewatering characteristics of existing dewatering techniques.
The study presented bere is focused on the second option. An important dewatering
parameter is the dry solids content of the sludge. The basic aim is to achleve higher
dry solids content of the dewatered sludges. Removal of water is accompanied with
volume reduction. In tigure 1.1 the sludge volume is given as a function of the dry
solids fraction for two different initial dry solids contents (2 and 5 wt%). The sludge
volume is inversely proportional to the dry solids content. An increase of the dry
solids fraction from 0.02 to 0.20 reduces the sludge volume by a factor of 10. Another
consequence of increasing the dry solids content of the dewatered sludge is the
rednetion of the energy needed for incineration. Above a dry solids content of 30 to
35 wt% the combustion process is self-sustaining and does not need additional fuel.
Incineration is used as post-treatment after mechanical dewatering. In 1991, only 4%
of the total sewage sludge produced (12 million kg of dry solids) was incinerated.
However, in the near future three big incineration plants for sewage sludge will be in
operation for about 35 % of the total sludge production in the Netherlands [Mars
kamp, 1993].
Introduetion 3
100
90
80
~ 70 ._,
~ 60
50
<I) 40 bi)
] 30 ... 20
10
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
dry solid.s fraction (-)
2% S%
Fig. 1.1 Sludge volume as a julletion of dry solids fraction for two initial dry solids contents
[Koot, 1974/.
The mecbanical dewatering of sludges by filter presses, belt presses and centrifuges
appears to be a unit operation wbicb is very poorly understood. The main reason for
tbis is the complexity of the sludge material. The cbemical composition of the sludge
material is caused by the biomass (bacteria, protozoa, fungi) and by organic and inor
ganic materials. External factors, sucb as the sewer system, weather conditions, and
climatological circumstances, also influence its composition.
Moreover, in a flitration process the sludge cake formed appears to be bigbly
deformable as welt. This means that when mechanical forces are exerted on the filter
cake a lot of water remains entrapped within the cake. lt is believed that tbis mecha
nism is to a high extent responsible for the attainable final dry solids content in a
mechanical dewatering process.
In the period 1979-1983 the 'Foundation for Applied Waste Water Research' (Dutcb
abbreviation: STOW A) has carried out a stndy on mechanica! dewatering of sewage
sludges. The study included a literature survey on the nature of the solid/water bond
strength in sewage, sludge properties, and dewatering results of mechanical dewatering
4 Cbapter 1
equipment. Moreover, studies on optimization of sewage sludge conditioning and the
operation of cbamber filter presses and belt presses were carried out. The results have
been presented in eight reports [STOW A, 1979 .. 1983]. The studies were principally
empirical by nature and much progress in fundamentally understanding the mechanica!
dewatering process could not be expected. A better fundamental understanding of the
sewage sludge dewatering process can help to improve the dewatering characteristics
(high dry solids contentand more rapid dewatering) of existing techniques.
The aim of the Dutch research program 'Future treatment technique for municipal
waste water (RWZI 2000)' is to develop new technologies for waste water treatment.
This research program is a co-operation between the 'Foundation for Applied Waste
Water Research' and the 'lnstitute of Inland Water Management and Waste Water
Treatment' (Dutch abbreviation: RIZA). Various research projects were initiated by
RWZI 2000 with respect to purification of waste water and sewage sludge processing,
such as fundamental aspects of sewage sludge dewatering, anaerobic prepurification,
magnetic dephospbatation, rednetion of sewage sludge production, and the Vertech wet
ox.idation process.
A research project entitled 'Fundamental aspects of sewage sludge dewatering' was
carried out at the Laboratory of Separation Technology of the Eindhoven University of
Technology in the period 1990-1994 [Herwijn et al., 1994]. Before the start of this
research project, the state of the art of mecbanical dewatering was investigated [van
Dijck et al., 1989]. Based on this preliminary study a proposal for a research project
'Fundamental aspects of sewage sludge dewatering' was written [Coumans and
Kerkhof, 1989] and accepted.
The global objectives of this research project were:
l. Characterization of sewage sludges. This implied determining sewage sJudge
properties which are expected to be of relevanee for a better understanding of
mecbanical dewatering processes. Characterization methods had to be developed
and tested on the applicability to diagnose and optimize the mecbanical dewatering
process. With suitable characterization methods the attainable dry solids content of
sewage sludge under certain process conditions can be determined.
The result of this study is presenled in this thesis.
3. To develop a physical-mathematical model for the description of the solid-liquid
separation of sewage sludge. The dewatering bebaviour of sewage sludge depends
not only on the sludge properties but also on the way in which the solid-liquid
separation process is performed. In practice the dewatering capacity of mecbanical
Introduetion 5
dewatering equipment is often smaller than the design capacity. Also, the dewa
tering results may fluctnate. Models which describe the operation of dewatering
equipment as a function of process and machine parameters can help to solve these
problems. A physical-mathematical model for the description of the dewatering
behaviour of sewage sludge is a basis for equipment models. The result of this
work was presented in another thesis [La Heij, 1994].
As a matter of course, both studies can not be considered as two different parts, but
are si:rongly interrelated.
Contents of this thesis
In this thesis various sewage sludge characteristics which are of relevanee for a better
understanding of the mechanica! dewatering process will be discussed.
In chapter 2 the various types of waste water processing and mechanica! dewatering
equipment will be dealt with. Emphasis is laid on four waste water treatment plants in
the Netherlands. Sewage sludges from these plants were characterized in this research.
In chapter 3 the solid/water bond strength in sewage sludge is discussed. The
solid/water bond enthalpy as a function of the water content is a fundamental charac
teristic which enables the prediction of the theoretically maximum dry solids content in
a certain dewatering process. The measuring methods used to deterrnine the bond
enthalpy are theemal analysis teclmiques and water vapour sorption isotherrns. In
chapter 4 various sludge dewatering characteristics and their interrelationships are
discussed. Macroproperties (e.g. specitic cake resistance, Capillary Suction Time,
permeability) as well as floc microproperties (e.g. composition, partiele size distribu
tion, rheological properties) are deterrnined in well-defined laboratory tests. Newly
developed characterization tests and tests used in previous studies are evaluated. The
sludge properties are investigated as a function of the flocculant dosage. In chapter 5 a
modified 'Capillary Suction Time (CST)' apparatus is discussed. The conventional
CST instrument provides an indication of the dewatering rate of sewage sludges. The
modified CST apparatus enables to predict the average specific cake resistance of
flocculated as well as non-flocculated sludges. In order to promote the dewatering
behaviour of sewage studges flocculants (iron chloride/time and polyelectrolytes) are
added. In chapter 6 the results of a preliminary study of the destabilization phenomena
occurring in sludge flocculation are presented. The zeta potential of the sludge
suspension is hereby a characteristic parameter. On the basis of the measurements a
model is developed to describe the hydrolysed metal ion adsorption at the sludge solid surface.
6 Chapter 1
Finally, in chapter 7 the main conclusions and perspectives for future research are
presented.
2 WASTE WATER PROCESSING
2.1 Introduetion
One of the major steps in waste water treatment is the removal of dissolved and
undissolved solids that otherwise might damage the effluent quality, and subsequently
the concentration of removed solids into a much smaller vohune for ease of handling
and disposal. The waste suspension generated in waste water treatment processes is
generally referred to as sewage sludge and exists in many forms and quantities. The
amount and quality of sludge depend on the origin of waste water, type of treatment
plant, and on the metbod of plant operation. Waste water entering the sewer system
originates from householcts (domestic sewage), andJor from industries (industrial
sewage). In the past, various methods were developed to remove organic matter from
sewage: sedimentation in tanks, the trickling filter and the activared sludge process.
The earliest metbod of waste water treatment was sedimentation in septic tanks.
Sedimentation of municipal waste water has a limited effectiveness, since only a part
of the waste organics is settleable. The sedimentation tank is a primary treatment in
which no deliberate attempt is made to remove oxygen-demanding materials.
The trickling filter, also called biological bed, was the first major breakthrough in
secondary waste water treatment. A secondary treatment plant was designed to remove
solids and to rednee biologica! oxygen demand (BOD). BOD is, by defmition, the
quantity of oxygen utilized during a certain time at a certain temperature by a mixed
population of micro-organisms in the aerobic oxidation process, and is a measure of
the biodegradable organic content of waste water. The oxygen demand at a tempera
ture of 20 oe during a period of 5 days is used as a standard, and abbreviated as
B0~0•
The trickling filter process occurring in a biologica! bed is based on slow movement
of waste water through a bed of rocks covered by biological mass, and results in rapid
rednetion of organic matter. Excess microbial growth is removed from the filter
effluent by a final clarifier. The waste sludge of the final clarifier is called filter
humus. In the Netherlands biologica! beds are still used for waste water purification
(total capacity in 1991: 1.638·106 inbabitant equivalents).
Another secondary treatment is the activated sludge process, which is the most
occurring waste water process in the Netherlands (total capacity in 1991: 1.7263·107
inbabitant equivalents [CBS, 1993]. The inbabitant equivalent is defined as the oxygen
8 Chapter 2
demand of sewage discharged by one persou in one day (see also section 2.7).
2.2 The activated sludge process
A major advancement in biological treatment took place when it was observed that
biological solids, developed in polluted water, flocculated organic solids. These
masses of micro-organisms (mainly bacteria), referred to as activated sludge, rapidly
metabolize pollutants from a solution and can subsequently be removed by gravity
settling.
In tigure 2.1 a typical process scheme of municipal waste water treatment is given.
The influent is stored in a buffer tank (1) to intercept high sewage flows and is
screened to separate large solids (2). The grit chamber (3) protects the mechanical
equipment and sludge pumps against abrasive wear. In the primary settling tank (4),
settleable solids are removed. The solids withdrawu from the bottorn of this tank by a
seraper are knowu as primary sludge (9). Primary settling produces a sludge of coarse
organic and inorganic solids. The organic part is actively decomposed by bacteria and
diffuses offensive odours.
influent--.-i2J:1/ i /
/ 2 1
7 Sludge processing
4
Fig. 2.1 Process scheme of a typical activated sludge processjor municipal waste water
treatment. l=bujfer tank, 2=screen, 3=grit chamber, 4=primary settling tank, 5=aeration
tank, 6=final settling tank, 7=sludge processing, 8=grit, 9=primary sludge, IO=secondary
sludge.
Waste water processing 9
Primary sludges thicken and dewater readily because of their fibrous and coarse
nature. Solids concentrations in primary sludge are in the 4 to 6 percent range.
The basic cbaracteristic of the whole system is the use of a mixed bacterial culture for
the conversion of pollutants. Conversion of organic materials to oxidized end products
(C02 and H20) takes place in the aeration tank (5). Bacteria metabolize dissolved
waste solids, produce new growth while taking in dissolved oxygen and releasing
carbon dioxide and water. The organic material serves not only as an energy source,
but also as a carbon souree for cell synthesis. In order to meet the required oxygen
demand, air is driven into the aeration tank by porous diffusers, by surface aerators,
or by some other means, such as brushes or aspirators.
The biomass produced in the aeration tank must be settled out in the secondary or
final settling tank (6), and partially recycled to the head of the aeration tank. The
excess of biomass mnst be wasted. This biological waste material is known as waste
activaled sludge or secondary sludge (10). Waste-activated sludge from the aeration
tank consists of flocculated microbial growths with entrained nonbiodegradable,
noncolloidal, and colloidal solids. It is relatively odour-free because of biological
oxidation, but the dispersed particles make it difficult to dewater. Secondary biological
sludge from aeration processes is less concentrated in solids than primary sludge.
So in activated sludge systems a part of the activated sludge is recycled and the other
part is withdrawn from the final settling tank. The primary and/or secondary sludge
must be thickened and/or mecbankally dewatered (7) before ulrimate disposal. The
purified efflnent is discbarged into surface waters.
Complete mixing aeration without primary sedimentation is popular for treatment of
small waste water flows. This type of waste water treatment, named oxidation ditch
[Pasveer, 1957], is often used to serve a town with a population of several thousands.
Typical ditierences between the normal activated sludge plant and the oxidation ditch
are:
The grit chamber and primary settling tank are missing in an oxidation ditch
system.
The greater size of the aeration basin in an oxidation ditch system. The purifica
tion system consists of a large ring-shaped aeration basin in wbich the liquor is
circulated and aerated by rotaring brushes. The sludge needs sufficient time to take
up oxygen for stabilization (see section 2.3).
The organic removal rate in an oxidation ditch is relatively small compared with
the regular activated sludge process. The organic removal rate is expressed in
10 Chapter 2
units of kg BOD per kg dry solids (ds) per day. Typical value of the organic
removal rate in an meidation ditch is 0.05 kg BOD/kg ds per day,whereas typical
values for an activated sludge plant are in the range of 0.1 to 0.2 kg BOD/kg ds
per day.
When a shortage of biological food is present in a system, bacteria liquidate
themselves and organic cells are oxidized. This process is called mineralization
and is typical for an oxidation ditch. In the case of an abondance of nutrients,
bacteria only metabolize a part of the dissolved and undissolved solids applied.
This is a typical characteristic for a high-loaded activated sludge plant. As a
consequence, the amount of surplus sludge produced is relatively higher in a
regular activated sludge system.
Elimination of primary settling dramatically affects the character of waste sludge. The
dry solids content of oxidation sludge is in the 0.5 to 2 percent range.
In tigure 2.2 a scheme of the successive steps in sludge processing is presented.
2.3 Sludge stabilization
Sewage sludge is a good growth substrate for bacteria and therefore putrefies very
rapidly, usually requiring stabilization prior to further use or disposal. The content of
organic matter must be reduced such that intensive putrefaction processes are no
longer able to proceed. Sludge stabilization may be of a biological, chemical or
thermal nature. In this section only biological sludge stabilization is dealt with.
Biological sludge stabilization can take place in the presence of oxygen (aerobic
stabilization) or in the absence of oxygen (anaerobic stabilization or digestion). More
than half (52 %) of the total sludge produced in the Netherlands in 1991 was
anaerobically digested, 37 % was aerobically stabilized, and the remainder was not
stabilized [CBS, 1993].
Allaerobic stabilization
In the conventional activated sludge plant anaerobic processes are generally used for
the purpose of sludge stabilization. The bacterial process consists of two successive
processes · that occur simultaneously in digesting sludge. In the fust stage large
components are broken down and converted into organic acids. This step is performed
by a variety of bacteria operating in an environment devoid of oxygen. In the second
stage, gasification is needed to convert the organic acids into 65 wt% methane, 35
purified water
Waste water processing
polluted
water
activated sludge process
primary
sludge
sludge stabilization
aerobic or anaerobic
thickening
mechanica! dewatering
incineration
secondary sludge
Fig. 2.2 Scheme of the consecutive steps in sludge processing.
11
12 Chapter 2
wt% carbondioxide (biogas), and trace amounts of nitrogen, hydrogen, and hydrogen
sulfide. Methane-forming bacteria are strictly anaerobic and are very sensitive to the
environmental conditions of pH and temperature. Digesters operate normally at 30 to
35 oe. The biogas produced in anaerobic digestion is an excellent fuel and is almost
always used to heat the digester tank, and/or buildings situated on the waste water
treatment plant.
Other advantages of the anaerobic stabilization process are:
Rednetion of the dry solids mass.
Increase of ash content due to conversion of organic matter. As a consequence
less odour probieros and a better dewaterability (less compressible cake) are to be
expected.
A disadvantage of anaerobic stabilization is the rednetion of the sludge heat content,
which might adversely affect a possible incineration process.
Aerobic stabilization
Aerobic stabilization may occur simultaneously in the aeration tank, or separately in
an additional tank. In small sewage works (e.g. the oxidation ditch system), simulta
neons aerobic sludge stabilization is applied. An important parameter in this stabiliz
ation process is the required sludge retention time needed for excessive mineralization.
The sludge retention time is strongly dependent on temperature. The organic removal
rate in such systems is relatively low.
In the separate aerobic sludge stabilization process, sludge undergoes aeration in a
separate tank, and the organic materials are decomposed by aerobic roetabolie pro
cesses. The organisms only experience an abundant nutrient supply during the initial
stages, and over a period 5 to 10 day period they reach the condition of stable,
underfed sludge.
2.4 Thickening
Thickening may be used as the first step in sludge processing for volrune reduction.
Different methods are utilized to thicken sewage sludges:
1. Thickening in separate tanks. Figure 2.3 illustrates a typical gravity thickener.
Sludge flow enters from behind an inlet well in the centre of the tank and is
directed downwards. The supernatant overflows a peripheral weir, while an
underflow of thickened sludge is drawn down from a bottorn sump in the tank.
Waste water processing 13
Piekets attached to the collector arms stir through the sludge providing cavities for
the release of entrapped water.
2. Flotation thickeners. Waste sludge enters the bottorn of the flotation tank, where it
is merged with recirculated flow containing compressed air. Dissolved air flotation
is achieved by releasing fme air bubbles that attach to sludge particles and cause
them to float. The overflow, discharged by a mechanica! skimming device, is the
thickened sludge. Dry solids concentrations of between 3 and 7 percent are
reached.
3. Mechanica! thickening by centrifuges. Centrifuges used for sludge thickening
produce a solids concentration normally varying between 4 and 12 percent. In the
next section a description is given of the working principle of this dewatering
device.
4. Lagoons. Sedimentarlon in lagoons is a natura! dewatering process. Both the sun
and the wind influence the dewatering result. This metbod requires a large surface
area. Soil and ground water polintion are disadvantages of natura! dewatering.
Fig. 2.3 Schematic drawing qf a gravity thickener.
2.5 Mechanica! dewatering
In the Netherlands various methods are used to dewater sewage sludges. In table 2.1
14 Chapter 2
an overview is given of the distinct dewatering techniques used, and their typical
dewatering results achieved in 1991.
Table 2.1 Overview of dewatering results gained in 199l[CBS, 1993].
Dry solids Volume (1 03 m3) Dry solids fracnon
mass (106 kg) (%)
in out in out ===
Drying beds 4.4 127 17 3.5 31.6
Lagoons 6.7 222 69 4.0 20.9
Belt presses 112.9 3515 553 3.5 21.3
Filter presses i 65.1 1957 253 4.0 34.1
Centrifuges 17.9 544 104 3.5 21.0
Incineranon 10.4 64 4 99.9
Sewage sludge dewatering in filter presses and belt presses is performed by means of
filtranon and expression (see secnon 4.4). In filtranon and expression the water moves
relanve to the solids under the influence of a liquid pressure gradient. Solids are
deposited in tbe form of a cake on tbe up-stream side of a filter medium while tbe
clear liquid passes tbrough. A porous plate, filter paper, or textile fabric acts as a
filter medium. In order to improve tbe mechanica! dewatering behaviour of sewage
sludge, tlocculants are added to sewage sludge before dewatering.
The chamber filter press, a batchwise operaring filter, is an important dewatering
technique in tbe Netberlands. Figure 2.4 shows tbe basic layout of a chamber filter
press. The basic unit is constructed of a sequence of plates and frames mounted on
suitable supports, e.g. a pair of rails. The hollow frame is separated from the plate by
a filter clotb to create a series of clotbwalled ebarobers into which slurry can be forced
under pressure. The plates and frames are held togetber by hydraulic pressure, or by
means of a screw. Two cakes are formed simultaneously in each cbarober. The liquid
passes through tbe cloth, runs down the corrugated surface of tbe plates, and is
discharged at drain points. When the two cakes join, tbe ebarobers are full of cake and
the dewatering process is stopped. Subsequently tbe ebarobers are opened and tbe cake
is removed.
Slurry In let
Waste water processing
Frame Plate
• ~w I I I 1 1 1
I ~~ i 1 l 1
I I I w ~--! ! !1
I I I I I l I t I I I i I i I I I I 1 1 1 1
I I
~'-."'-L-"'.J./-bw/%1
Flitrata outleta
Fig. 2.4 Sche1111ltic diagram of a chamher filter press.
15
In the course of the dewatering process the mechanica! pressure increases until a
maximum of about 15 bars is reached. Cake depths are commonly up to 50 mm. The
dewatering time takes about several hours. In order to promote the mechanica!
dewatering behaviour of sewage sludge, flocculants are added to sewage sludge before
dewatering. The type of flocculants usually used in chamber filter presses are iron
chloride in combination with time.
The belt press is a continuously operated machine and is nowadays the most applied
dewatering technique in the Netherlands (see table 2.1). The operating principle is
based on mechanically squeezing sewage sludge cake between two beits or filter
fabrics (see figure 2.5). The feeder delivers conditioned sludge to the porous lower
moving belt. The sludge successively passes the 'pre-dewatering zone', the 'pressing
zone', and the 'friction zone'. In the 'pre-dewatering zone', water is only removed by
gravitational force. The two moving beits apply pressure on the sludge cake in the
'pressing zone'. Further squeezing of the cake is applied in the 'friction zone', where
16 Chapter 2
shear stresses are exerted on the cake. The cake structure is disrupted and in this way
more dewatering is achieved. The cake is finally discharged and filtrate passing
through the lower filter belt is collected. The type of flocculants used to dewater
sewage studges in belt presses are organic polymers, also called polyelectrolytes
(p.e.). Compared to the chamber filter press, the dewatering results of belt presses are
worse (see table 2.1). This is due tothesmaller residence times of sludge particles in
belt presses (8 to 10 minutes), and the excess amount of dry solids (flocculants) added
prior to the dewatering in chamber filter presses.
pre dewaterlng zone
pressing zone
Fig. 2.5 Schematic diagram of the belt press.
friction zone
In the Netherlands centrifuges are getting more popular to dewater sewage sludges.
Centrifugal sedimentation is based on a density difference between solids and liquids.
The particles are subjected to centrifugal forces, which makes them to move radially
through the liquid either outwards or inwards, depending on whether they are heavier
or lighter than the liquid.
In practice, the scroll-type centrifuge (also called decanter centrifuge) is utilized to
dewater sewage sludges. lts operaring principle is shown in tigure 2.6. A charac
teristic feature of the scroll-type centrifuge is the horizontal conical bowl, containing a
screw conveyor that rotates in the same direction but at a slightly higher or lower
speed.
The sludge suspension enters through an axial stationary feed pipe at the centre of the
rotor and passes through the distributor into the rotaring bowl. On their way from the
entrance to the cylindrical end of the bowl the solids are separated from the carrier
liquid by the centrifugal force. Deposited solids are moved by the screw conveyor
towards the conical end of the bowl and are discharged. The supernatant is freely
discharged over a ring dam at the other end. Polyelectrolytes are used to flocculate
Waste water 17
sewage sludges to be dewatered in centrifuges. A major advantage of this type of
centrifuge is the operational flexibility. Bowl speed affects centrifugal forces on the
settling particles, and the difference between bowl speed and conveyor speed controls
the solids retention time.
Overflow
Cake
Fig. 2.6 Schematic diagram of the scrolt-type centrifuge.
2.6 Description of various types of sewage sludge
~ tinderflow (solids)
Feed
Depending on their point of origin and nature of treatment, different types of sewage
sludge may be distinguished. In the previous sections, several types of sludge have
already been mentioned. In this section, a general overview of sludge types occurriug
in the activated sludge process will be given.
1. Primary sludge.
Primary sludge is separated from the incoming sewage in the primary sedimen
tation stage. lts composition is made up of the settleable solids contained in the
municipal raw sewage. It excludes those materials collected on the sereens and the
grit separated out in the grit chamber. It undergoes thickening in the primary
settling tank to a solids concentration of 4 to 6 wt%. Primary sludge is of a
course, non-homogeneaus and variabie nature.
2. Activated sludge, (secondary sludge).
The activated sludge or sludge biomass present in the aeration tank is the natural
vehicle for the biological treatment process. Within its biomass the enzymatic
power is manifested which is required for the conversion of high-molecular sub
stances into the desired inorganic endproducts of treatment, or for incorporation
into the cell biomass. When sewage is aerated, the micro-organisms form colonies
(activated sludge flocs) and function as adsorptive surfaces to which biologically
18 Chapter 2
inert matcrials may be attached. The chemica! composition of activated sludge is
determined by the biomass itself (bacteria, protozoa, fungi), and by organic and
inorganic matcrials which are present either as deposits or inclusions. The
proportion of the individual constituents is dependent on the sludge loading (high
sludge loading=high organic fraction), the nature of the sewage, and the effi
ciency of the primary settling. In plants without prior settling (e.g. mddation
ditches) the activated sludge will contain high proportions of inorganic matter.
The combination of an aerated reaction chamber and the succeeding fiual settling
tank with parrial recycling ofthe separated biomass (recycled sludge) is the basis
of the activated sludge process. Separation and recycling of the sludge biomass
enable a high biomass concentration in the aeration tank. Sludge thickened and
separated from the final sedimentation stage is called secondary sludge, surplus
sludge, or waste activated sludge. Sludge flocs are required to settie rapidly with
no undesirable metabolic processes taking place, i.e. the flocs must be compact,
heavy, and inactive. The tenns •activated sludge', 'secondary sludge', 'surplus
sludge', 'waste activated sludge' refer to physically identical sludges.
Sludge withdrawn from the settling tank in the oxidation ditch system is called
oxidation ditch sludge.
4. Stabilized sludge.
Stabilized sludge refers to sludge which is treated such that putrefaction processes
do no longer proceed and offensive odours are no longer given off. During
biologica! sludge stabilization, the organic fraction is reduced by controlled
metabolic reaction processes until the desired degree of stabilization is achieved.
Two types of biologica! stabilization processes may be distinguished: aerobic and
anaerobic stabilization (see section 2.3). Sludge produced in the anaerobic
stabilization process (digestion) is called digested sludge.
5. Conditioned sludge.
Prior to mechanica! dewatering sewage sludge is conditioned by flocculation in
order to achleve a better dewatering result (see chapter 6). Flocculants used in
waste water treatment are inorganic salts and polyelectrolytes.
Types of sludge to be dewatered are primary sludge, secondary sludge (or a mixture
of both), aerobically stabilized sludge, digested sludge, and oxidation ditch sludge.
Waste water processing 19
2. 7 History and origin of sludges investigated
In the scope of this study four different sludges originating from four waste water
treatment plants in the Netherlands were characterized:
1. A mixture of prirnary and secondary sludge originating from the activated sludge
plant in Eindhoven.
2. Digested sludge coming from the activated s1udge plant 'Amsterdam-Oost'.
3. Oxidation ditch sludge from the waste water treatment plant 'Veghel-Uden'.
4. Oxidation ditch sludge originating from the waste water treatment plant 'De
Hooge en Lage Zwaluwe'.
In the next sections some typical characteristics of the above-mentioned waste water
purification plants are given. First some definitions which are typical for waste water
processing will be introduced.
The capacity of a waste water treatment plant is expressed in inhabitant equivalents
(i.e.). One inhabitant equivalent is, by defmition, the biologica! oxygen demand of
sewage produced by one persou per day (unit: BOD/day). This oxygen demand is
equal to 54 g BOD~0 per day. Oxygen consumption of industrial sewage is also
expressed in inhabitant equivalents.
Sludge age is defmed as the ratio between the rate of sludge wastage or withdrawal
and the quantity of sludge in the aeration tank, and corresponds with the meao
retention time of sludge in the aeration tank (unit: days).
The organic removal rate is the amount of high-energy organic substrates (pollutants)
converted into low-energy mineral endproducts per kilogram dry sludge solids per
day, and is expressed in units of kg BOD/kg ds per day.
2.7.1 The Eindhoven waste water treatment plant
The map of this activated sludge plant is given in Appendix 1. The incoming sewage
originates from about 25 cities and villages situated in the 'Dommeldal'. The mixture
of prirnary and secondary sludge is stored in a buffer tank and transported to the
sludge treatment plant in Mierlo. A process scheme of this plant is also presented in
Appendix 1. The ratio between the amount of primary sludge and secondary sludge is
strongly dependent on the weather conditions. During heavy raio showers the supply
of prirnary sludge is high, due to sloicing of the sewer system.
20 Chapter 2
Characteristics
Waste water purification
Sewage composition: 60% domestic, 40% industrial
Design capacity: 750,000 i.e.
Influent flow: 5·107 m3/year
Organic removal rate: 0.20 kg BOD/kg ds per day
Sludge age: 6 days
Sewage sludge processing
Sludge flow: 5.6·105 m3/year
Meebankal dewatering equipment: 5 belt presses, 1 centrifuge, 2 chamber filter
presses
Dry solids content incoming sludge: 2.5 wt%
Dry solids content dewatered sludge: 22 wt% (belt presses), 24 wt% (centrifuge), 35
wt% (chamber filter presses)
Flocculant dose: 50-70 g/kg ds FeC13, 400-600 g/kg ds Ca(OH)2 (chamber filter
presses)
4-5 g p.e./kg ds (centrifuge and belt presses)
Flocculant type: Nalco 41162 from Nalco Company
Ultimate disposal: Dewatered sludge produced in the centrifuge and belt presses is
composted. Dewatered sludge from the chamber filter presses is
deposited on dumping grounds.
2.7.2 Waste water treatment plant 'Amsterdam-Oost'
The city of Amsterdam possesses four waste water treatment plants. Sewage origina
ting from the north and centre of the city is treated in the activated sludge plant 'Am
sterdam-Oost' (map of plant is given in Appendix 1). Half the total sewage produced
in Amsterdam is supplied to this plant. Since January 1993 treatment of sewage
sludges produced in the four plants bas been centralized in 'Amsterdam-Oost'. The
mixture of primary and secondary sludge is thickened in four gravity thickeners, and
is subsequently anaerobically stabilized in seven digesters, which operate at a. tempera
ture of 30 °C. The sludge retention time in the digesters is about 20 to 30 days. The
Waste water processing 21
daily biogas production is about 25,000 m3• The biogas is bumt in gas engines to
generate electricity. The released heat is used to heat up both the digester tanks and
water for heating the buildings on the plantside.
Characteristics
Was te water purification
Sewage composition: exclusively dornestic sewage
Design capacity: 750,000 i.e.
Influent flow: 4.8·107 m3/year
Organic removal rate: 0.10 kg BOD/kg ds per day
Sludge age: 9 days
Sewage sludge processing
Sludge flow: 8.4·105 m3/year
Mechanica! dewatering equipment: 3 centrifuges (3·1,200 kg ds/hour), 4 chamber
filter presses (1,120 m3 sludge/day)
Dry solids content supplied sludge: 2.0-2.5 wt%
Dry solids content dewatered sludge: 25 wt% (centrifuges and chamber filter presses)
Flocculant dose: 5 g p.e./kg ds (chamber filter press), 7-8 g p.e./kg ds (centrifuge)
Flocculant type: Zetag 63 and Zetag 73 from Allled Colloids
illtimate disposal: After composting the material is dumped on tips.
2.7.3 Oxidation ditch system 'Veghel-Uden'
Sewage from the municipalities Heeswijk-Dinther, Uden, Volkel, Eerde, Zijtaart,
Vorstenbosch, Mariaheide, Erp, Keldonk, Boekel and Venhorst is supplied to the
oxidation ditch system 'Veghel-Uden' (map in Appendix 1). Differences with the
conventional oxidation ditch system are:
Sewage treatment includes a grit chamber.
Only a small part (2.5 %) of the sludge produced and thickened in two tanks is
further dried in lagoons. Sludge from lagoons (5 wt% dry solids) is incidentally
used as a fertilizer.
22 Cbapter 2
Cbaracteristics
Waste water puri:fication
Sewage composition: 35 % domestic, 65 % industial
Design capacity: 250,000 i.e.
Influent flow: 1.3·107 m3/year
Organic removal rate: 0.06 kg BOD/kg ds per day
S1udge age: 13 days
Sewage sludge processing
Sludge flow: l. 4·1 05 m3/year. A bout 10 % of the total sludge supplied consists of
digested sludge which was produced extemally.
Mechanica! dewatering equipment: 4 belt presses (4·350 kg dslhour)
Dry solids content supplied sludge: 3.5-4.0 wt%
Dry solids content dewatered sludge: 17 wt%
F1occulant use: 5-6 g p.e./kg ds
Flocculant type: Superfloc C496 from Cyanamid B.V.
Ultimate disposal: Dewatered sludge is composted and subsequently deposited on
dumping grounds.
2. 7.4 The oxidation ditch system 'De Hooge en Lage Zwaluwe'
The municipality 'De Hooge en Lage Zwaluwe' bas its own waste water purification
system. The process scheme of the oxidation ditch system is presented in Appendix 1.
Surplus sludge withdrawn from the fina1 settling tank is stored in buffer tanks. The
thickened sludge (dry solids content 3 wt%) is supplied to the waste water treatment
plant in Rijen (see Appendix 1) for further treatment. Sewage sludge originating from
the 'De Hooge and Lage Zwaluwe' plant is only a sma1l fraction of the total sewage
sludge supplied to the 'Rijen' plant. In 1992, a total amount of 122,215 m3 sewage
sludge bas been dewatered with two belt presses, 2, 765 m3 was coming from the 'De
Hooge en Lage Zwaluwe' plant, 43,640 m3 from the 'Rijen' plantand 75,810 m3 from
nine other waste water treatment plants.
Waste water processing
Characteristics
Waste water purification
Sewage composition: 100 % dornestic sewage
Design capacity: 6,000 i.e.
Influent flow: 7 .1·105 m3/year
Organic removal rate: 0.06 kg BOD/kg ds per day
Sludge age: 20 days
Sewage sludge processing in 'Rijen'
23
Sludge flow: 1.2·105 m3/year; 2.8·103 m3/year from the 'de Hooge en Lage Zwa
luwe' plant
Mechanical dewatering equipment: 2 belt presses (2·30 m3 sludge/hour)
Dry solicts content supplied sludge: 3.5 wt%
Dry solicts content dewatered sludge: 20 wt%
Flocculant use: 6 g p.e./kg ds
Flocculant type: Zetag 87 from Allied Colloids
Ultimate disposal: Dewatered sludge is dumped on tips.
3 WATER BINDING IN SEWAGE SLUDGE
3.1 Introduetion
A clear onderstanding of the sludge solid-to-water bond strength is of importance to
get a fundamental insight into the dewatering behaviour of sewage sludges. Sewage
sludges typically have moisture contents of 95 to 99 wt%. The amount and type of
water present in sewage sludges can play an important role in defining their dewater
ing characteristics. Three models have been proposed to describe the types of water
present in sludges [Smollen, 1986; Vesilind, 1974; STOWA, 1981]. All three models
refer to water that is 'bound' in some fashion to sludge solids, whether the binding is
obtained chemically or physically. Water binding is expressed - with respect to pure
water - in lower values for vapour pressure, water activity, entropy and enthalpy of
water molecules.
Different techniques have been proposed for estimating the bound water content of
s1udges. 'Bound water' is defmed operationally by the measuring metbod used.
Smollen [1988 and 1990] reported on the use of the drying curve metbod for quanti
fying the bound water content of biologica! sludges. The hypothesis of the metbod is
that physically bound water evaporates from the sludge at a slower rate than free
water.
The dilatometric metbod is based on the hypothesis that bound water does not freeze
below the freezing point of pure and free water. The volume of freezable water could
be determined from the net expansion of the fluid level in the dilatometric unit. The
difference between this freezable ( or free) water and the total water was defined as the
bound water content. Dilatometric tests were used by Barber and Veenstra [1986] and
Robinson and Knocke [1992].
Katsiris and Kouzeli-Katsiri [1987] used differential thermal analysis (DTA) to
determine freezing curves and quantify the free and bound water fractions in waste
activated and digested sludge samples.
In this study two measuring methods are used: isothermal drying curves and water
vapour sorption isotherms. A defmition of 'bound water' used in this research study is
given in section 3.3.4. The sludge solid-water bond enthalpy can be obtained as a
function of the sample moisture content by the two methods mentioned. Know1edge
about the water bond enthalpy in a sewage sludge cake as a function of the moisture
26 Chapter 3
content enables the predierion of the theoretically maximum feasible dry solids content
in a certain dewatering process.
In this chapter a model is presented to describe the different types of water present in sewage sludge. Subsequeutly, the measuring teclmiques used and experiments carried
out to determine the solid-to-water bond strength are dealt with.
3.2 The presence of water in sewage sludge
Figure 3.1 schematically represents the way in which water may be present in a
sewage sludge suspension and a sludge cake.
a
ocs
• • r in particles
erstitial
Hydration layer Additive particles with
water
Fig. 3.1 Schematical representation of types of water present in sewage sludge. a) bulk water between jlocs; b) interstitial water inside floc; c) hydrated water to surface of jloc particles; d) incorporated water, e.g. intracellular water within floc particles [Kerkhof, 1991].
Water binding in sewage sludge 27
In a suspension or in a filter cake one can distinguish a bulk water and a floc pbase. A
three-dimensional floc network is formed by processes tbat aggregate the colloidal
sludge particles. Flocculants are used to promote the aggregation of the basic sludge
particles (see cbapter 6). The floc consists of a skeleton in which interstitial water is
present.
The basic sludge particles, like microbial cells, pieces of wood, etc., may contain
water inside (incorporated water). Moreover, additives like tlocculants and filter aids
may possess incorporated water. Water enclosed in organic cells is called intracellular
water. Further hydration layers may be present at the surface of floc particles, for
instanee bound to the ionogenic groups of the partiele surface. Hydrated water may
also surround the added flocculants (e.g. FeC13.6H20).
3.3 Isothermal drying curves (TGAIDT A)
In this study thermal analysis is used to determine isothermal drying curves of sewage
sludge cakes. Thermal analysis is a group of techuiques in which a physical property
of a substance and/or its reaction products are measured as a function of temperature
whilst the substance is subjected to a controlled temperature program. W ell-known
techuiques are thermogravimetrie analysis (TGA) and differential thermal analysis
(DTA).
In thermogravimetrie analysis (TGA) the change in sample mass (mass loss or gain) is
determined as a function of temperature and/or time. The resulting mass change
versus temperature curve (thermogram) generally provides information on the thermal
stability and the composition of the sample, the intermediate compounds, and the
residue. Thermogravimetry is universally applied to a large number of analytical
probierus in the fields of metallurgy, paint and ink science, ceramics, mineralogy,
food technology, inorganic and organic chemistry, polymer chemistry, biochemistry,
and others.
In differential thermal analysis (DTA) the temperature difference between a sample
and a reference material is measured as a function of the sample, inert material, or
furnace temperature when the sample is subjected to a controlled temperature pro
gram. Temperature differences between the sample and the reference material are due
to enthalpie transitions occurring in the sample material, such as those caused by
pbase changes, fusion, boiling, sublimation and vaporization, dehydration reactions,
dissociation and decomposition reactions, and other chemica! reactions. Differential
28 Cbapter 3
thermal analysis is generally used to determine the heat of enthalpie transition ( or
reaction), or the mass of the reactive sample.
The available thermal analysis equipment (SETARAM, TGA 92 with DTA probe)
provides the possibility to carry out thermogravimetry and differential thermal analysis
simultaneously. The equipment consists of: a cylindrically shaped furnace (4>=21.9
mm), a thermobalance, a temperatore programmer and controller, and a computer.
The filmace may operate with temperatores up to 1000 °C and is externally cooled by
water circulation. The filmace may employ a carrier gas, such as air, inert gases, or
reactive gases. In the experiments carried out nitrogen (inert gas) was used as carrier
gas to remove the vapours given off by the sample. The thermobalance permits
continuous weighing of a sample as a function of time and/or temperatore. The
controller contains an acqnisition and amplification card for the various signals and
transfers digitized signals to the computer. The typical measuring probe which is connected to the microbalance is presented in fignre 3.2. The measuring probe bas
been developed by Boersma [1955]. Two aluminium cups (4>=3.7 mm; height=3.9
mm) are positioned on the probe.
TO MICROBALANCE
HEFEREN CE
FURNACE WALL
Fig. 3.2 The typical measuring probe positioned in a cylindrically shaped jurnace (<!>=21.9
mm). The probeis connected toa microbalance.
Water binding in sewage sludge 29
The reference sample is an inert material in the temperature domain in wbich experi
ments are carried out; however, in tbis study an empty aluminium sample holder is
taken as reference. The sample material is a piece of sludge cake (weight 40 to 60
mg) obtained from a fUtration experiment (see secdon 4.4.1 ). Thermocouple A (Pt
Pt/Rh 10%) measures the temperature of the sludge cake sample. Either the sample or
furnace temperature is regulated by the temperature controller. In the experiments
carried out the sample temperature is kept constant.
During an isothermal drying experiment, the sludge cake sample will take up heat
needed to evaporate water. The temperature of the sample will then be lower than the
temperature of the reference material. The temperature difference is registered with
two thermocouples, Br and B, (Pt-Pt/Rh 10%), wbich are positioned just below the
sample bolders. The temperature difference is proportional to the heat flow to the
sample (see next section). Due to the position of the thermocouples (see tigure 3.2),
the measured temperature difference is not equal to the real temperature difference
between reference and sample. Because of non-quantitied heat resistances, the relation
between the temperature difference and the heat flow bas to be calibrated. A suitable
calibration procedure is based on the evaporation of pure water at constant temperat
ures.
3.3.1 The TGA-DTA drying model
In tbis section a model is presented that describes the isothermal drying of samples
with the 'Boersma' measuring probe (see tigure 3.2). The model enables the calcula
tion of the water bond enthalpy. The model is based on mass and energy balances for
both the reference and the sample cup. The following assumptions are made:
1. The aluminium cup possesses a high thermal conductivity; consequently both the
sample material and cup have the same temperature.
2. The convective heat transfer coefficients for sample material and cup are differ
ent. However, an average convective heat transfer coefficient is assumed for both.
3. The emissivity E for the reference cup and the sample cup are the same (taken to
be 0.2 for unpolished aluminium).
30 Chapter 3
Energy balance reference cup
where Ar Hr m.-Tgas Tr Twa11 ar E (J
= heat transferring surface area of reference cup [m2]
enthalpy of reference cup [J.kg-1]
mass of reference cup [kg] temperature bulk gas in furnace tube [KJ reference temperature [KJ furnace wall temperature [KJ convective heat transfer coefficient for reference cup [W .m·2 .K 1]
emissivity [-] Stefan-Boltzmann constant = 5. 8·10·8 W .m·2 .K4
Equation (3 .1) means tbat lhe sum of lhe convective and radiative heat flow equals lhe
rate of heat accumulation in lhe reference cup. By definition, the change in mass of
lhe reference is zero: d.m,./dt=O. It will be shown tbat during a drying experiment lhe
reference temperature appears to be virtnally constant: dT/dt""O.
Consequently dH/dt=Cp,/dT/dt""O. If it is assumed tbat lhe wall temperature equals
lhe gas temperature, equation (3.1) is rewritten as:
(3.2)
Thns:
(3.3)
Energy balance sample cup
where
d ~. dH 4 4 dt(m.HJ = H.-d.t + m, dt. = a,A.(Tg .. -T,) + EUA,(Twau-T.) (3.4)
A, A .. H, Miv jw m. T, a,
=
j~ssdllv
heat transferring surface area of sample [m2]
sample surface area for moisture transport [m2]
enthalpy of sample [J.kg-1]
enthalpy of evaporation of pure water [J.kg-1]
moisture flux 1hrough surface area [kg.m·2 .s'1]
mass of sample [kg] sample cup temperature [KJ convective heat transfer coefficient for sample [W.m·2.K1]
Water binding in sewage sludge 31
Equation (3.4) means that the rate of heat accumulation in the sample cup equals the
sum of the convective and radiative heat flow minus the heat flow needed to evaporate
water.
The pure water evaporation enthalpy Miv is a function of the water temperature T w (in
K):
where pure water evaporation enthalpy at 273.15 K and 1 bar = 2504·103 J.kg-l specific heat of water vapour = 1. 87·103 J.kg-1• °C 1
specific heat of water = 4.18·103 J.kg-1• °C1
Mass balance sample cup
(3.5)
(3.6)
Substitution of equations (3.3) and (3.6) in (3.4) yields the ruling equation for the
drying of samples with the combined TGA-DTA technique:
(3.7)
The first and secoud term on the right-hand side of equation (3.7) equals Q:
(3.8)
where 01eff is the 'effective heat transfer coefficient':
(3.9)
3.3.2 Evaporation of pure water and the calibration of the DTA probe
The result of an experiment in which pure water (initial mass 50 mg) was evaporated
at a constant sample temperature of 60 oe is presented in figure 3.3. Water mass IDw,
evaporation rate diDw/dt, and the signal S produced by the thermocouples Br and B, (in
p, V) are given as functions of time. The derivative diDw/dt is digitally calculated by the
computer and specifies the evaporation rate. The temperature difference (which is
32 Cbapter 3
proportional to S) and the evaporation rate are virtually constant during the experiment
(see figure 3.3). From equation (3.7) it follows that the evaporation of pure water with
the 'Boersma' probe is given by:
(3.10)
where HW = cp.w(T.- 273.15) = cp,w(). enthalpy of water [J.kg'1]
50 1.000
0 0.001
0 1000 2000 3000 4000 5000 6000 7000 8000 9000
time (sec)
Fig. 3.3 Evaporation of pure water (initia/ mass 50 mg) at a constant sample temperature of
60 oe. Water mass 1nw (-), evaporation rate dmjdt (-·-·-), and the thermocouple signa/ S
(---) are registered as juncrions of time.
The water temperature is virtually constant during the experiment, thus
~/dt = Cp,w*dT.fdt = 0. The left-hand side of equation (3.10) equals zero, so
equation (3 .10) can be rewritten as:
(3.11)
The thermocouples Br and B. do not measure the real temperatures of the reference
cup and sample cup. Due to the position of these thermocouples, a measuring error is
Water binding in sewage sludge 33
introduced. It is assumed that both thermocouples are positioned in the temperatnre
boundary layer at a fractional distance f=(x/or) from the cup surface (see tigure 3.4).
i
i I Tgas
Fig. 3.4 Position of the thermocouple Bs in the temperature boundary la:yer (thickness br) at a
distance x from the sample cup suiface.
B· .. f.tT -T) = (T -T) \ gas s s,exp s (3.12)
B· r· (3.13)
Subtraction of equation (3.13) from equation (3.12) yields:
T-T = r s (Tr,exp (3.14)
The constant CroA ( = 11(1-f) > 1) is a correction factor wbich takes into account the
measuring error due to the position of the thermocouples. From equation (3.14) it can
be concluded that the position of the thermocouples B, and Br weakens the sensitivity
of the thermocouple signal. The measured temperatnre difference between the
reference cup and sample cup is calculated from:
34
s T -T =-r,exp s,exp ctc
Chapter 3
(3.15)
The constant ctc is a conversion factor which equals 6.625 p. V .K1• Substitntion of
equation (3.14) in (3.15) yields:
(3.16)
The sample temperatnre T, is maintained at 60 oe during the whole experiment. The
thermoconple signal S is virtnally constant in the experiment (see figure 3.3). As a
result the reference temperatnre T., which can be calculated from equation (3.16),
appears to be virtnally constant as well. Substitntion of equation (3 .16) in equation
(3.11) gives:
Q = aerrA-sCTGAS =- <imw(~-Cpw0,) ctc dt ·
The calibration factor Cr is defined as follows:
and thus from equation (3.17) it follows that:
c = Q f s
<imw 1 --(AU - C 0)·-dt ~-'v p,w ' S
(3.17)
(3.18)
(3.19)
The evaporation rate dffiw/dt is derived from the experiment and appears to be
virtnally constant. The sample temperatnre T., the thermoconple signa! S, and the
evaporation rate are k:nown as functions of time. Now, the calibration factor Cr (in
Wip. V) can be calculated as a function of time with equation (3.19). In figure 3.5 er is
depicted as a tunetion of the water mass.
The calibration factor appears to be virtnally constant during the experiment:
er 1.63 ± 0.05 mW/p.V. The resolution of the calibration factor is about 3%.
Consequently, the error in determining the heat of evaporation of water with the TGA
DT A technique is about 3%.
Water binding in sewage sludge 35
0.0018
·"'\ 0.0016 A .. r 0.0014
>' 0.0012
~ • 0.0010 I ~ 0.0008 4-o • u 0.0006 • • 0.0004 • •
0.0002 I 0.0000 r
0 10 20 30 40 50
mass mw (mg)
Fig. 3.5 The calibration factor c1 (in Wl~t V) as a fitnetion of the water rru2ss fnw.
3.3.3 Isothermal drying of a sludge cake
Sewage sludge cakes were driedat a constant sample temperature of 60 oe. The mass
of a sludge cake sample m. equals the sum of the mass of sludge dry solids ll\ts aud
the mass of water lllw:
where u = sample moisture content [kg w.(kg dsY1]
The derivative dm,/dt is given by:
dm,
dt du
mdsdt
The total heat accumulated in the sludge cake sample is given by:
where enthalpy of sludge solids [J. oe1]
enthalpy of pure water [J. 0 e 1]
(3.20)
(3.21)
(3.22)
36 Chapter 3
The derivative of m,;H, to time is represented as:
The enthalpy of the sJudge solids is given by:
Where Cp,ds = SpecifiC heat Of SlUdge SOlids [J.kg-t. oe-I]
0, = temperature [ 0 C]
The sample temperature remains constant during the experiment, thus
(3.23)
(3.24)
dHds/dt Cp.ds *dO,/dt = 0. As a result, the first term on the right-hand side of
equation (3.23) equals zero.
The enthalpy of water is given by:
where Cp,w = specific heat of pure water [J.kg-1• °C1
]
.Mlb = water bond enthalpy [J.kg-1]
(3.25)
The water bond enthalpy .Mlb is defined as the excess enthalpy to evaporale water out
of the sample. The derivative of llw to time is given as follows:
(3.26)
The drying process is isothermal, thus the first term on the right-hand side of equation
(3.26) is equal to zero. Substitution of equations (3.26) and (3.25) in equation (3.23)
yields:
d(m)I.)
dt
(3.27)
Substitution of equation (3.27) in the ruling equation for the drying of a sample with
the TGA-DTA technique (equation (3.7)) yields:
Water binding in sewage sludge
du [ iMHb m- -u--+ C 0 ds dt àu p.w s
Equation (3.28) cao be rewritten as:
Mlbl = Q + dm, LUi dt V
Q dm,
dt
37
(3.28)
(3.29)
The sample mass m., the evaporation rate dm./dt, and the heat flow to the sample Q
( =cr'S) are known at any given moment. The sample moisture content u is related to
the sample mass m. according to:
u(t) m,(t) -mds
(3.30)
Equation (3.29) provides the possibility to calculate the water bond enthalpy as a
tunetion of the sample moisture content.
3.3.4 Experimental results
The moisture-to-sludge solid bond strength was studied with four types of sludges
originating from four waste water treatment plants (see section 2.7). Sludge cake
samples obtained from a tiltration experiment carried out with the filtration-expression
cell (see section 4.4.1) were dried. Typical initial moisture contents varied between 4
and 7 kg w/kg ds. Different flocculants were used to condition the sludge sample prior
to filtration: FeC13/Ca(OH)2 , polyelectrolyte Röhm KF975, and the polymer applied in
practice. Three different dosages per flocculant type were used. The dosage Ca(OH)2
was maintained constant if the sample was flocculated with FeCl3/Ca(OH)2 •
In total 36 isothermal drying experiments were carried out. In all drying experiments
the sample temperature was kept at 60 oe. A drying experiment took about 2 to 3
hours, depending on the mass and initial moisture content of the sludge cake sample.
At lower drying temperatures the drying time increases and the sludge cake composi
tion may change due to biologica! activity. A uittogen flow of 1 liter/hour was
adjusted to remove the water vapour given off by the sample. The uittogen flow had
an upward direction.
38 Chapter 3
80011 1.000
7000
J ---.----:
-~-----:--80011 . /
[ '/"
i t·
5000 I: o.mo c:l I I
I ' 4000 I
I I
iS: 8000
I --- :1! liOOO --- ___",__--. --.--- O.Olll
~"' ] 111110
0
-111110 0.001
o.o 0.11 1.0 :Lil 2.(1 2.5 8.0
moistu:re content u [kg waterika ds)
Fig. 3.6 Result of the isothermal drying experiment (60 oe) with a Veghel sludge cake sample
flocculated with 3 g Röhm KF975/kg ds. The drying rate dm/dt (---), the heat flow Q and the bond enthalpy Mlb (-) arepresentedas functions of the sample moisture content u.
80011 1.000
7000
I 80011 -----:---
~ 11000
ê:t
' 4000
iS: 8000
1 liOOO
I ~..:.----
] 11100
(I
./ -11100
0.0
m.oistu:re content u [kg water/kg ds)
Fig. 3.7 Result of the isothermal drying experiment (60 oC) with a Veghel sludge cake sample
flocculated with 1 g Superflocfkg ds. The drying rate dm/dt (---), the heat flow Q(---),
and the bond enthalpy illlb (-) are presented as junctions of the sample moisture content u.
Water binding in sewage sludge 39
8000 :LOOO
7000'
i 1
.---~ 8000 /
~ /
i I 5000 I o.wo Cl
I
~ 4000 I/ i <l
>. 11000
~ I I ------- ~ 2000 I --- 0.010
lä -- .f' ] 1000
0
( -1000 0.1101
0.0 G.5 :LO L5 z.o 2.5 3.0
moisture content u [kg water/kg ds]
Fig. 3.8 Result of an isothennal drying experiment (60 oq with a Veghel sludge cake sample
jlocculated with 3 g Supeifloclkg ds. The drying rate dmjdt (---), the heat flow Q ,-.·-··J. and the bond enthalpy Mfb (-) arepresentedas functions of the sample moisture content u.
In figures 3.6, 3.7 and 3.8 the results of three drying experiment.<; are presented. In
the figures the drying rate dm./dt, the heat flow to the sample Q, and the water bond
enthalpy &Ib are depicted as functions of the sample moisture content. The drying
rate as a function of moisture content is also called the drying curve.
At the start of an experiment the evaporation rate and heat flow are relatively high. In
the first drying stage the evaporation rate remains virtually constant and free water is
transported. The drying rate during this period (the 'constant rate period') is determi
ned by the conditions in the continuons phase: temperature, humidity, and mass
transfer coefficient. The heat of evaporation in the sample equals the heat of evapora
tion of pure water. The type and added amount of polyelectrolyte do not cause a
marked difference between the drying rates of the constant rate periods. Halde [ 1979]
investigated the influence of an added concentration of Praestol 444K on the vacuum
drying rate of digested sludges. He concluded that the added concentration of poly
electrolyte had a minor influence on the sludge drying rate in the initial period.
40 Chapter 3
The figures show that at a sample moisture content of about 0.4 to 0.6 kg wlkg ds the
drying rate and heat flow start to decrease and the bond enthalpy starts to deviate
significantly from zero. Passing this critical moisture content the resistance to moisture
transfer inside the drying body starts to grow rapidly and mainly controts the mass
transfer. This second drying stage is called the 'falling rate period'. The movement of
water inside the sludge cake specimen may occur by various mechanisms, e.g.
capillary flow, liquid and vapour diffusion [Whitaker, 1977].
The critical moistnre content of 0.4 to 0.6 kg wlkg ds is found in every drying
experiment carried out. The 'bound water' content in a sludge cake was not affected
by the sludge and flocculant type and flocculant dosage. All graphs showing the bond
enthalpy as a function of sample moistnre content were similar. Moistnre having a
zero bond enthalpy is here by definition 'free water', and is mainly removed by liquid
flow in the filtration-expression process. Moisture having a bond enthalpy larger than
1 kJ/kg is called 'bound water', and cannot be removed in a mechanical dewatering
process at applied pressures smaller than 10 bar. 'Bound water' can only be released
in e.g. a thermal drying process. It is unlikely that a sharp physical boundary exists
that separates 'bonnd' water from 'free' water. The transition proceeds fairly continu
ously over the mentioned range in moistnre contents.
The condusion can be drawn that the maximum feasible dry solids content in a
mechanical dewatering process is about 65 to 75 wt%. The remaining parts of liquid
in the cake are bound to the sludge particles andlor represent incorporated water.
There seems to be different types of attachment for water bonding in sludges. Keey
[1972] distinguished three types of water binding, i.e. chemical (ionic and molecular), ·
physical (osmotic and adsorptive) and mechanical (capillary). Typical bond energies
for molecular, adsorptive, and capillary water are 5000, 3000 and 100 kJ/kmol,
respectively.
La Heij [1994] studied the effect of applied mechanical pressure on the equilibrium
dry solids content with the filtration-expression cell (see section 4.4.1). Dry solids
contents of 35 to 45 wt% were reached at low expression pressures (5 to 7 k:Pa) and at
optimal laboratory conditions. The considerable d:ifference in the theoretically
maximum dry solids content and the dry solids contents reached in laboratory
experiments indicates that most of the water remaining in the sludge cake is not
'bound' to particles. Clearly, other fractions of water have more impact on the final
cake solids concentration than the 'bound water' fraction. Probably, it is the water
present in the interstitial spaces (interstitial or interfloc water) that determines the final
Water binding in sewage sludge 41
cake dry solids content. In a mechanica! dewatering process the porous structure of the
cake material collapses and a lot of interstitial water remains entrapped within the cake
(immobilized water). The porous structure of the cake depends on the type of
flocculant used. The compressibility of sludge flocculated with polyelectrolyte is
higher than that of sludge flocculated with iron chloride/lime. The compressibility
influences the structure of the formed cake in terms of permeability [La Heij, 1994].
In mechanica! dewatering equipment the moisture content in sludge fllter cakes
amounts about 2 to 4 kg w/kg dry solids ( = 20 to 35 wt% dry solids). A simple
calculation learns that 90% of the water present in the fllter cake with a dry solids
content of 20 wt% is present as 'free water' (bond enthalpy equals 0) and the remai
ning 10% is present as 'bound water'.
The dry solids contents reached in practice are much smaller than the dry solids
contents reached at laboratory scale (35 to 45 wt%). This discrepancy indicates that
other factors, like bad flocculation conditions, floc break-up during transport, filling
problems of chambers in fllter presses, negatively affect the sludge dewatering process
in practice. More research is needed into the flocculation process and transport of
(flocculated) sludge in practice.
Higher mechanica! pressures are needed to remove a larger amount of interfloc water.
La Heij [1994] showed in laboratory experiments that by applying high mechanica!
pressures (6 to 10 MPa) in a hydraulic frame press a fllter cake solids content of 60
wt% could be reached. The remaining parts of water in these experiments were only
'bound' to the sludge particles!flocs or represented incorporated water.
3.4 Water vapour sorption isotherms
3.4.1 Introduetion
The secoud experimental technique to characterize the solid to water bond strength in
sewage sludge is measuring water vapour sorption isotherms. The water vapour
sorption isotherm of a substance is the constant temperature relation of the water
content in the substance and its thermadynamie water activity. The water activity <lw is
defined as a ratio of fugacities:
42 Chapter 3
(3.31)
where fw is the fugacity (or 'escaping tendency') of water in the mixtnre at equilibrium
and ~ is the fugacity of pure water at standani temperatnre and pressure. The fugacity
becomes equal to pressure as the system approaches ideal gas behaviour. Equation
(3.31) is then rewritten as:
(3.32)
where Pw is the partia1 vapour pressure of water in the system and p~ is the satnration
vapour pressure of pure water at the same temperatnre. The ratio PwiP~ is also called
the equilibrium relative humidity (RH) or relative vapour pressure of the system. For
a certain system, the water activity is a function of temperatnre and moistnre content,
and lies in the range 0 s aw s 1. In this context, bound water is defined as water ha ving
a water activity below the bu1k water activity (aw= 1).
Since sorption of water by a sludge cake is a spontaneous process, it is accompanied
by a decrease in the thermadynamie Gibbs free energy G. During sorption the entropy
S is decreased, caused by trapping vapour molecules at active sorption sites or in a
thin surface layer. The enthalpie change .:UI during the sorption process reads:
AH= ÄG + TÄS (3.33)
Thus the enthalpie change during this process must be negative (exothermic process).
At high water activities the molar heat evolved from sorption AH..,r from the vapour
phase equals the molar enthalpy of condensation ~nd ( = 44 kJ/mol at 20 °C).
Under these conditions water vapour sorption is comparable with condensation. The
enthalpy of sorption increases with decreasing water activity. The excess or net
enthalpy of sorption AHexc is defmed as:
(3.34)
and corresponds thermodynamically with the bond enthalpy (section 3.3.3). The excess
sorption enthalpy is indirectly obtained from sorption isotherms at different temper
atnres by applying the equation of Clausius-Clapeyron:
Water binding in sewage sludge 43
R (3.35)
where Llllw ( < 0) corresponds to the differential enthalpy of wetting, and R is the gas
constant. The contribution of the sorbent to the enthalpy effect may be neglected, so
Llllw ""' Llll.,xc · Gal [1967] gave an overview of available tecbniques for the deterrniuation of iso
therms for water vapour sorption. In the present study, water vapour sorption iso
therms were determined with two tecbniques: firstly the conventional tecbnique of
vacuum exsiccators with saturated salt solutions to control the water activity, and
second1y the Cou1ter Omnisorp 100, a fully-automatic measuring instrument.
In order todetermine the differential enthalpy of wetting, water vapour isotherms have
to be measured at different temperatures.
3.4.2 Sorption models
In literature about 80 equations to describe sorption isotberms are known [lglesias and
Chirife, 1982; v.d. Berg and Bruin, 1978], because the interactions between water
(sorbate) and dry substance (sorbent) are very complex. A sorption model which takes
into account all kinds of interactions will lead to a complex isotherm equation witb
many constants. However, a more practical approach is an isotherm equation contain
ing only a few parameters, which must have a physical meaning. It should be realized
that it will remain a simplification insofar as not all structural and interaction details
can be covered. In the present study the following two sorption models were studied.
B.E.T. model (0<1!,.,<0.35)
Brunauer, Emmet and Teller [1938] derived an isotherm equation for localized
multimolecu1ar adsorption onto independent sites, assuming the adsorbed molecules
beyoud the first one to have bulk liquid properties. The B.E. T. equation reads:
(3.36)
44 Chapter 3
where Cs is a constant depending on the interaction of the first adsorbed molecule
with the adsorption site and the temperature.
The temperature dependenee of Cs is given by:
[MI -MI l c "" c exp 1 cond
s s,o RT (3.37)
The adsorption enthalpy of water molecules in the first layer (MI1) is larger than the
condensation enthalpy of pore water (LUiconJ. The adsorption enthalpy in the second,
third, .... mlh layer equals the condensation enthalpy (LUI2 =LUI3 = ... =Mlm=MJ.,onJ·
G.A.B. model (0<~<0.9)
Andersou [1946] proposed that the molar sorption enthalpies of the secoud and follo
wing molecules on a sorption site are not the same for condensation of the bulk liquid.
The influence of the sorption site on the second, third,. .. and mlh layer is noticeable.
Anderson multiplied the water activity by a constant k, less than unity, to take this
inflnence into account.
The isotherm eqnation is then represented as:
(3.38)
where Cg is the Guggenheim constant. De Boer [1953] and Guggenheim [1966]
derived the same isotherm equation on a thermodynamic base. All three authors
worked independently. lsotherm eqnation (3.38) is referred to as the Guggenheim,
Andersou and de Boer (G.A.B.) model of adsorption.
The temperature dependenee of the Guggenheim constant is given by:
(3.39)
The factor k, which corrects for differences in properties of water molecules present
in the multilayer with respect to pure water, is also a function of the temperature:
Water binding in sewage sludge 45
[ (.6-H -.6-H ) l P k = ka exp m cond = ka exp(~) (3.40)
The G.A.B. model provides energetic infonnation on the averaged multilayer and
must be considered to be a more general model for localized homogeneons sorption
than the B.E.T. model.
In this stndy the temperature dependent G.A.B. model is preferred for experimental
isotherm analysis. The experimental data can be fitted with a S-shaped curve. The
G.A.B. equation represents a S-shaped curve. However, the G.A.B. model cannot be
used as a physical model for water sorption in sludge cakes. Multilayer adsorption
does not play an important role in sludge cakes. Water absorption in sludges particles
is the most important sorption mechanism in sludge cakes (see figure 3.1).
lnsulation
Sludge cake sample
Wheels
Thermostata
Water bath
Fig. 3.9 Schematic diagram of the whole expertmental set-up.
Metalplate
Glass vessel
Chapter 3
3.4.3 Metbod of saturated salt solutions
Equimnent and procedure
In this study a new experimental apparatus bas been developed to measure water
vapour sorption isotherms, based on the couventional technique of vacuum exsiccators
with saturated salt solutions [van Dijke, 1992]. The data for the water activity of these
solutions were taken from the tables of Greenspan [1977]. These tables relate the type
of salt solution and its temperature to the value of the water activity. Por keeping a
constant temperature during equilibration, the exsiccators were positioned in a
thermostated water bath(± 0.1 K).
A schematic diagram of the equipment is shown in tigure 3.9. Twelve glass vessels,
positioned in a revolving metal frame, were partially filled with twelve different
saturated salt solutions to control the water activity. Sludge cake samples having initial weights of about 1 gram were obtained from a tiltration experiment. The samples were
brought into small glass vessels which were positioned on small perspex tables. The
samples were pretreated with thymol, which is an effective fungicide for a relatively
long period (2-3 months) and does not influence the water vapour sorption behaviour
[v.d. Zande, 1993]. Prior to the start of an experiment the exsiccators were evacuated
to reduce the equilibration time. The samples will desorp water dependent on the
water activity. Regularly the samples were weighed. Attainment of equilibrium was
assumed if two subsequent weighings within 24 hours gave the same results. When
equilibrium was reached, the temperature of the water bath was changed and subse
quently the same samples were equilibrated again. At the end of the experiment the
equilibrium moisture content was determined by drying the sample at 105 oe. In this
way twelve points of each sorption isotherm were determined.
With the vacuum exsiccator method, water vapour desorption isotherms were measu
red of three different sludge cakes (Mierlo, Amsterdam and Veghel). The flocculant
used in the experimeuts was FeClsfCa(OH)2 • Each sludge cake sample was flocculated
with three different dosages of FeC13• The dosage Ca(OH)2 was kept constant for each
sludge type. In this way 9 different sludge cake samples were studied.
Equilibration time
Figures 3 .1 0 and 3 .11 show desorption isotherms measured at three temperatures of
two sludge cakes: Amsterdam sludge cake flocculated with 100 g FeC!lkg dry solicts
and 400 g Ca(OH)2/kg dry solicts and Veghel sludge cake flocculated with 50 g
0.32
0.28
0.24
~ ..!f 0.20
lit O.:IB
~ 0.22
~ '0.08
0.04.
0.0
Water binding in sewage sludge
0.~ 0.2 0.3
* * * 298..15
298.15
0.4 0.15 0.6
aW' (-)
0 0 0 313.15
- - 313..15
47
0.7 0.8 0.9 1.0
A A A 328..15
328.15
Fig. 3.10 Desorption isotherms of Amsterdam sludge cake flocculated with 100 g FeCf/kg ds
anti 400 g Ca(OH)/kg ds at three different temperatures. Symbols: experimental results.
Lines: model.
0.32
0.28
-0.24
.a .Jf 0.20
lit O.:IB
~ o.u ~ 0.08
0.04.
o.o o.~ 0.2 0.8
* * * 303 .. 15
803.15
0.4 0.15 0.8
aW' (-)
0 0 0 818.15
818..15
0.7 0.8 0.9 1.0
A A A 328..15
328.15
Fig. 3.11 Desorption isotherms of Veghel sludge cake flocculated with 50 g FeCf/kg ds anti
200 g Ca(OH)/kg ds at three different temperatures. Symbols: e.xperimental results. Lines:
model.
48 Chapter 3
FeC13/kg dry solids and 200 g Ca(OH)2/kg dry solids. The symbols show the experi
mental resnlts. One experiment took about 2-3 months. The graphs show isotherms
with only eight to ten data points. At high water activities (a_.> 0.8) the samples did
not equilibrate. Wolf et al. [1985] related the long equilibration time of food products
at high a_.-values to the problem of microbial deterioration of the samples.
GAB-fit
Upon increase of the isothermal temperature, water sorption equilibria shift towards a
lower moisture content at a given water activity. The temperature-dependent G.A.B.
equations (3.38 to 40) were fitted to the measured isotherms, using the sum of least
squares metbod for minimizing the absolute differences between measured and
calculated moisture contents. This was efficiently done by nsing the Statistica!
Analysis System package (SAS). The lines in the graph represent the temperature
dependent G.A.B. equation. The S-shaped G.A.B. equation describes the experimental
isotherms very well. The isotherms for these two samples are almast identical over the
entire range of water activities, especially at low water activities (a_. <0.3). It indicates
that the two sludge cakes, which originated from different waste water treatment
plants and were flocculated with different dosages of FeCl3 and lime, have about the
same water sorption isotherm. The small differences in sorption capacity may be
attributed to biologica! variations of the slndge cake samples. The same sorption
behaviour bas also been found for the other sludge cakes investigated [v.d. Zande,
1993].
Fit parameters of the temperature dependent G.A.B. model are given in table 3.1.
Van den Berg [1981] measured sorption isothermsof potato starch samples, which are
also organic products, at different temperatures. The sorption equilibria were analysed
in terms of the B.E.T. equation and G.A.B. equation. The typical monolayer value u1
for s1udge cake samples of about 0.1 kg wlkg ds was also found by van den Berg
[1981] for potato starch samples.
Water binding in sewage sludge 49
Table 3.1 Results of desorption analysis for sludge cake samples in termsof the temperature
dependent G.A.B. equation. Veghel 5 wt% FeCl3 means Veghel sludge flocculated with 50 g
FeC[/kg ds.
Sample cg.o(-) ko(-) Et E2 llt
(kJ/mol) (kJ/mol) (kg w/kg ds)
Veghel 5 wt% FeCI3 7.1E-10 0.081 64.52 5.525 0.089
Veghel 10 wt% FeC13 7.5E-7 0.080 46.97 5.628 0.087
Veghel 30 wt% FeCI3 2.2E-8 0.166 56.86 4.091 0.072
A'dam 5 wt% FeCI3 1.7E-8 0.015 56.07 9.45 0.096
A'dam 10 wt% FeCI3 1.2E-6 0.037 44.05 7.34 0.106
Mierlo 5 wt% FeCI3 3.4E-4 0.033 30.55 7.56 0.085
Mierlo 9 wt% FeCI3 3.2E-4 0.056 30.70 6.45 0.083
Mierlo 27 wt% FeCI3 l.lE-9 0.016 63.88 9~ 0.113
The GAB parameters Gg and k can now be calculated as functions of the temperature
with the equations (3.39) and (3.40), respectively. In figures 3.12 and 3.13, Cg apd k
are plotted as functions of the absolute temperature for the Amsterdam sludge cake
sample flocculated with 100 g FeClikg ds.
The Guggenheim constant Cg is a measure of the binding energy of tbe first adsorbed
layer. The binding energy of the first adsorbed layer decreases with increasing temper
ature. Van den Berg [1981] observed a shorter residence time of adsorbed molecules
in the fust layer when the temperature is increased. The sorption process becomes less
strongly localized. As a consequence, at constant water activity the sample moisture
content reduces with increasing temperature. This is confirmed by the experimental
results (see figures 3.10 and 3.11).
All parameters of the temperature-dependent G.A.B. equation are known, so it is
possible to calculate the sorption isotherm of a sludge cake at any given temperature,
which therefore needs not be measured.
Figure 3.14 shows desorption isotherms at eight temperatures for the Veghel sludge
cake sample flocculated with 100 g FeCl,lkg ds obtained by this analysis.
50 Chapter 3
350
300
260 -I 200 -ct
0 160
100
60
0 270 280 290 300 310 320 330 340 360
Tempersture (K)
Fig. 3.12 The Guggenheim constant Cg as a function of the absolute temperature for the
Amsterdam sludge cake sample jlocculated with 100 g FeClikg ds.
1.00
0.90
0.80
0.70
- 0.60 I - 0.60
...:.:::: 0.40
0.30
0.20
0.10
0.00 270 280 290 300 310 320 330 340 350
Tempersture (K)
Fig. 3.13 The G.A.B. parameterkas a function of the absolute temperature for the Amster
dam sludge cake sample flocculated with 100 g FeC[/kg ds.
0.8
- 0.7 ~ 0.6 ~ 0.5 ~ 0.4 ~ 0.3 ~ 0.2
Water binding in sewage sludge
/ /
51
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 aw (-)
Model -- 283.15 . 293.15 - - . 303.15 - - 313.15 - 323.15 - 333.15 - 343.15 ·------- 353.15
Fig. 3.14 Desorption isotherms of the Veghel sludge cake sample flocculated with 100 g
FeCl/kg ds at eight temperatures.
The temperature-dependent G.A.B. equation relates the sample moisture content u to
the water activity <lw and the absolute temperature T: u=u(Clw,T). The inverse function
<lw=<lw(u,T) is given by:
where
and
e 0·5(e u2 -2uu (e -2)+u 2e )05 +u(e -2)-u e g g 1 g 1 g g lg
El egO exp(-)
· RT
2uk(eg-1) (3.41)
(3.42)
52
Ez k =ko exp(-) RT
Entbalpy of wetting
Chapter 3
(3.43)
Application of the Clausius-Clapeyron equation (3.35) and the G.A.B. equation (3.41)
yields an analytical expression for the differential entbalpy of wetting Mlw. Unfortu
nately this expression, calculated by using the MAPLE"' package, is very large and
therefore presented in Appendix 2. Figures 3.15 and 3.16 show the differential
enthalpy of wetting, -Mlw, as a function of the sample moisture content u for two
sludge cake samples as derived from the procedure mentioned. The entbalpy of
wetting increases with decreasing sample moisture content and flatteus at very small
moisture contents. The shape of the bond entbalpy curve corresponds well with that
measured of other products containing organic cells, such as potatoes [Görling, 1955]
and wood [Pidgeon and Maass, 1930].
Figure 3 .17 presents the bond entbalpy as a function of the moisture content of the
Veghel sludge cake sample conditioned with 50 g FeC13/kg ds as obtained from
sorption isotherms and isothermal drying curves. Both curves are not identical. The
boud entbalpy obtained from drying curves does not flatten, but increases sharply
when the sample moisture content is reduced. At very small moisture contents the
evaporation rate dm,/dt reduces stronger than the heat flow Q. As a result, the ratio
Q.(dm,/dt)"1 used to calculate the bond entbalpy increases strongly with decreasing
moisture content.
At a eertaio bond entbalpy, the moisture content determined in an isothermal drying
experiment is larger than the moisture content obtained from the sorption isotherm.
During drying of the sludge cake, a moisture content profile is present in the cake.
The water content at the surface is smaller than the moisture content in the cake. In
the sorption experiment, the moisture content was measured in an equilibrium state.
As a result, the moisture concentradon profile in the sludge cake is flat. The average
moisture content of the sludge cake in the drying process is thus larger than the
average moisture content measured in the sorption experiment.
With the sorption analysis the bond entbalpy was only calculated over a small moisture
content range (u<0.3 kg w!kg ds), whereas the drying analysis yields valnes of the
bond entbalpy over the whole working range of sample moisture contents.
Water binding in sewage sludge 53
3000
2000
1000
0 ~----.---~----~-----.----,-----,-----,---~
o.oo 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40
u (kg w/kg ds)
Fig. 3.15 Dif.ferential enthalpy of wetting as a junction of the sample moisture content.
Sample: Amsterdam sludge cake flocculated with 100 g FeCl/kg ds and 400 g Ca(OH)/kg ds.
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40
u (kg w/kg ds)
Fig. 3.16 Dif.ferential enthalpy of wetting as a function of the sample moisture content.
Sample: Veghel sludge cakeflocculated with 50 g FeCl/kg ds and 200 g Ca(OH)/kg ds.
54
8000
~ 5000
~ 4000
~
f 3000
2000
i 1000 ] 2 0
-1000
0.0
Chapter 3
i
l I
. \. .I
. \ ~ \
o.t
\ \. '. ""-..
0.2 0.3
-~-.---·- ·,
0.4 0.5 0.6
moisture content (kg w/kg ds)
Fig. 3.17 Comparison of bond enthalpy as a function of water content for the Veghel sludge
cake sample flocculated with 50 g FeC[/kg ds as derived from isotherm analysis (-) and
drying analysis (- -).
The transition from 'free' to 'bound' water is difficult to determine in both graphs.
However, it seems that this transition predicted by the sorption analysis occurs at
somewhat smaller moisture contents. The condusion based on the drying analysis, that
only a small part of water (about 10%) present in a sludge filter cake is chemically or
physically bound to the sludge particles, is confirmed by the sorption analysis.
3.4.4 Coulter Omnisorp 100
The Coulter Omnisorp 100 instrument is an automated gas sorption analyser to
measure both adsorption and desorption isotherms. Adsorptive gases that are common
ly used are nitrogen, oxygen, hydrogen, argon, krypton, carbon dioxide, carbon
monoxide, and water vapour.
Water binding in sewage sludge 55
helium nitrogen
V5
Pil
furnace vacuum pump
outgassing section physisorption section
Fig. 3.18 Schematic diagram of the Omnisorp 100.
Figure 3 .18 represents a schematical diagram of the Omnisorp 100. The instrument
may be considered to be split into two sections: the ontgassing section and the
physisorption section. The ontgassing section is fitted with a furnace for heatiug
samples under vacuum up to a temperature of 450 °C. The physisorption section is
used for measuring the vapour pressure of the adsorbate, and comprises of a manifold
(black area in tigure 3.18), a sample port A, and a reference port B.
The Omnisorp 100 instrument may operate in two different modes: the contiuuous
volumetrie and the static volumetrie mode. In the contiuuous volumetrie mode, the gas
is continuously dosed to the sample at a very slow rate (typically at 0.3 ml/min) by
means of a mass flow controller. In the static flow method, the adsorbate is supplied
from the reference bulb into the manifold and subsequently to the sample. The
different valves (V) are operated by the software of the Omnisorp. To performa water
vapour sorption measurement, the static flow metbod is used. A small volume of
deionized water is added to the reference bulb. Reference bulb and sample bulb should
be kept at the analysis temperature. The prevailing pressure in the reference bulb is
the saturation pressure of water at the analysis temperature (23 torr at room tempera
ture).
56 Chapter 3
At the start of an adsorption measurement the sample bulb is evacuated. The vacuum
system comprises of a rotary vacuum pump and a diffusion pump. After evacuation a
gas dose of known volume and pressure is supplied to the sample and this dose is
equilibrated with the sample. Manifold or sample pressure and saturation pressure of
the adsorbate are measured by three capacitance-type pressure transducers fi.tted to the
manifold. PO is the reference pressure transducer, which continuously measures the
satnration pressure of the adsorbate. The low-range transducer P1 provides high
resolution pressure measurements within the range of 0 to 10 torr. P2 is a high-range
transducer, with a full-scale reading of 1000 torr. Equilibrium is reached when a
number of consecutive sample pressure readings are within the equilibration tolerance.
The number of consecutive pressure readings and the toleranee are speci:fied by the
operator. The total volume of gas present in the sample bulb is calculated from the
difference between dose pressure and equilibrated pressure. The 'dead volume', which
is defined as the volume of the sample bulb where no adsorption takes place, is
substracted from the calculated volume to determine the volume of gas adsorbed by
the sample. The determination of the dead space is performed with helium, which is a
non-adsorbing gas.
When the equilibration criterium bas been met, the same vapour dose is again supplied
to the sample and a new equilibrium sample pressure will adjust. In this way a new
data point of the adsorption isotherm (relation between relative vapour pressure and
volume of gas adsorbed) is determined. Afterwards the sample bulb is dosed again
with water vapour. Thus the whole adsorption isotherm is determined. The sample
measurement time is generally 24 to 72 hours, and thus this technique is less time
consurning than the vacuum exsiccator method. Water vapour adsorption can be
followed by vapour desorption. If a desorption experiment is required, first the
reference bulb is equilibrated with the sample before the run is started. After equili
bration a certain gas dose is removed from the sample bulb, the sample pressure is
suddenly decreased, and a new equilibrium will adjust, etc.
In the scope of this research, some preliminary experiments were carried out with
sludge cakes. Figure 3.19 presents the result of a desorption experiment. The sludge
cake used bad a mass of 0.03 g and originated from the Mierlo sludge treatment plant.
This experimental result is compared with the results obtained with the vacuum
exsiccator method. Both results agree. The amount of data points determined by the
Omnisorp (99) is much higher than the amount of data points obtained by the vacuum
exsiccator method (12).
Water binding in sewage sludge 57
0.22
1).20
ll.UI -~ 0.16
g: étlll.UI ~om ::som
liJM
11.82
0.00
0.0 0.1 OJI 0.8 O.<l D.5 OJI 11.7 0.8 0.9 10
aw(-)
LI!GBND -- GAB • • • Omm * * * &*
Fig. 3.19 Desorption isothenns of Mierlo sludge cake jlocculated with 45 g FeC[/kg ds and
200 g Ca(OH)/kg ds determined with the Omnisorp and the vacuum exsiccator method. lhe
line represems the G.A.B. equation fitted on the results of the Omnisorp experiment.
Consequently, the accuracy of the fitted parameters is much smaller in tenns of the
95% confidence interval (see table 3.2).
Table 3.2 Parameters of the G.A.B. equation fïtted on the Omnisorp experiment and the
results obtained by the vacuum exsiccator method.
Omnisorp
k (-) 0.752 0.605
k 95% confidence interval 0.745 <k<0.759 0.454<k<0.756
27.896 45.87
26.806 <cg< 28.986 Cg<79.37
0.081 0.095
0.0801 <u1 <0.0821 0.077 <u1 <0.1138
58 Chapter 3
Especially in the low activity range (aw<O.l) a lot of data points are available. This
low activity range is very interesting for tbe purpose of tbis stndy. For 3w < 0.1 the
bond enthalpy differs significantly from zero.
With the Omnisorp tbe water vapour isotherm is measured of only one sample, where
as with the vacuum exsiccator metbod the isotherm is determined by measuring at
twelve samples simultaneously. The sludge cake samples may differ in composition
due to biologica! variations.
It was only possible to determine the water vapour sorptilln isotherm at one temperat
ure: room temperature. The Coulter Omnisorp bas to be modified. The physisorption
section may be air-thermostated so that sorption experiments can be measured at
different temperatnres (working range: 25 to 70 °C). In the near future more experi
ments have to be carried out at different temperatnres to delermine the bond enthalpy
accurately and to aim at a better onderstanding of the working principle of the
apparatns.
Calibration of somtion apparatus
Sorption properties of biological or food produelS may differ from each other due to
biological variations of tbe substrate, differences in equipment design, and different
procedures in handling. A collaborative research program named 'COST 90 project'
was initiated in 1985 to study possible influences of the measuring metbod applied on
the adjustment of equilibrium conditions over a wide range of water activities [Wolf et
al., 1985]. Microcrystalline cellulose (MCC) was chosen as reference material,
because this material shows well-defined and stabie sorption properties. The sorption
isotherm of MCC was measured in 32 laboratories, which all took part in tbis study.
The mean isotherm of MCC was fitted with the G.A.B. model. The results of tbe
COST 90 project were summarized in recommendations for the measurement of
sorption isotherms [Wolf et al. , 1985].
To investigate how far the vacuum exsiccator metbod and tbe Omnisorp 100 meet the
requirements of the COST 90 project, experiments were carried out with MCC. The
results are shown in figure 3.20. Botb experimental isotherms show acceptable ag
reement with tbe mean sorption isotherm determined in the scope of tbe COST 90
project. The vacuum exsiccator and Omnisorp 100 adsorption run gave data which
show a slightly lower moisture content. This may be attributed to differences in the
structnre of MCC (particle size distribution, porosity) used.
8.18
0.17 0.18 o.u o.H
IL18
:i'0.:12
f~ JfiUW :::to.oe
o.olll
o.oc o.oa o.oz o.m
Water binding in sewage sludge
o.oo.,~~~~~~~~~~~~~~~~~~~~~~~~~T
0.0 0.1 0.8
--- OOSTilO
O.<l O.IS
aw
• • • OMNI80
0.7 0.9
***Salt
59
Fig. 3.20 Adsorption isothermsof microcrystalline cellulose at 25 oe_ (-) COST 90 project;
(···) Omnisorp 100; (***) Vacuum exsiccator method.
Moreover, the suppliers of MCC used in the COST 90 project and the MCC used in
our experiments are not the same.
3.5 Conclusions
In order to study tbe solid-to-water bond strengtb in sewage sludge, two different
methods were used: tbermal analysis to determine isotbermal drying curves, and the
measuring of water vapour desorption isotberms at different temperatures. Witb botb
metbods it is possible to determine the water bond enthalpy as a function of tbe sludge
cake moisture content. Botb metbods show tbat tbe bond entbalpy differs significantly
from zero at sample moisture contents smaller than 0.3 to 0.6 kg water/kg ds. The
sludge origin, type of flocculant and flocculant dosage do not influence tbis critical
moisture content.
Moisture having a bond entbalpy smaller than 1 kJ/kg is classified as 'free water' and
moisture ha ving a bond entbalpy larger than 1 kJ/kg is called 'bound water'. 'Bound
water' is not removable in a mechanical dewatering process. So the maximum feasible
60 Chapter 3
dry solids content in a mechanical dewatering process amOWlts to about 65 to 75 wt%.
However, in practice dry solids contents of about 20 to 30 wt% are reached. The
'bound' water fraction does not contribute significantly to the sludge moisture
retention capacity. Probably the interstitial or interfloc water which bas been
entrapped during filter cake formation mainly dictates the final cake solids concentra
tion.
The experimental isotherms measured at different temperatures can be described very
well with the S-shaped temperature-dependent G.A.B. equation. Sewage sludge cakes
from different waste water treatment plants show minor differences in sorption
capacity. Moreover, the different additives and their dosages influence the sorption
isotherm to a minor extent.
4 SLUDGE DEWATERING CHARACTERISTICS
4.1 Introduetion
Sewage sludges produced by municipal waste water treatment works exhibit wide
variatious in their physical, chemica! and biologica! properties due to differences in
the types of waste waters and the design and operation of waste water treatment
plants. Seasonal and weather conditious can often complicate matters even further. By
no meaus can sewage studges beregardedas well-defined systems with 'excellent' and
constant properties.
In academie research it is common practice to perform studies with well-defined
model systems. However, with respect to sewage sludges this is not an easy approach,
because the ruling properties are not sufficiently known. In the past, model sludge
studies were carried out sporadically. Haustveit et al. [1977] used an Arthrobacter
pure culture as model sludge. The pure culture was insulated from an activated sludge
sample. An important property of these bacteria was their ability to form aggregates.
Over the years many types of tests have been developed to measure the specific
properties of studges in relation to partienlar forms of treatment, such as activated
sludge processing, stabilization, thickening, dewatering, tipping, and incineration.
Vesilind [1985] gave an overview of tests on sludges. The tests were categorized as
being either physical, chemica! or biological.
It is the objective of this study to establish a set of sludge dewatering characteristics.
A set that may be cousidered to be a fingerprint of the sludge. Successively four
sewage studges originating from four different waste water treatment plants have been
characterized (see section 2.7). Eropbasis was placed on evaluating characterization
tests used in previous studies, and on developing and reviewing potenrial new
methods. The practical relevanee of characterization tests to monitor and control the
performance of the operational sludge treatment plant was studied. Up to now the
determination of sludge dewatering characteristics at sludge treatment plants is still
rather ambiguous.
Moreover, an attempt was made to establish (cor)relatious between floc micropropert
ies and filtration behaviour of studges (macroproperties) in well-defmed laboratory
tests. The sludge microproperties that were studied are composition, zeta potential,
partiele size distribution, and rheological properties. The dewatering behaviour of
studges was studied in terros of average specific cake resistance, cake solids concen-
62 Chapter 4
tration, porosity, permeability, vacuum snction time, eapillary Suction Time,
concentration ferric ions and polyelectrolyte in filtrate. The macro- and micropropert
ies were studied by using different types and dosages of flocculants. The flocculation
process appears to be a very critica! operation which affects the dewatering properties
of the sludge to a great extent.
4.2 Plan of characterization research
As pointed out in the previous section, four different sewage sludges were characteriz
ed. In Appendix 3 the woricing scheme of the sewage sludge characterization research
is presented. Sewage sludge samples were collected from the various treatment plants
at positions in the dewatering system just before the sludge is mixed with flocculants
and transported to the dewatering equipment. The sludge samples were storedat 5 oe.
La Heij and Janssen [1990] measured filtration properties of a sewage sludge sample
as a function of the storage time. Their condusion was that the major changes in
filtration characteristics (increase in specific cake resistance) occurred during the first
day, when the sewage sludge sample is collected at the sludge treatment plant. Eikurn
and Paulsmd [1974] measured the specitic cake resistance of primary and primary
chemical sludges as a function of the storage time at 10 oe. During 18 days of storage
the specific resistance increased to four to nine times its original value. Therefore the
characterization research on one sewage sludge sample must be carried out in a
relatively short time (one week).
The following characterization tests were daily applied to unconditioned sludge
samples to check possible changes in composition: A TP content and esT value.
The next characterization parameters were determined as functions of the flocculant
dosage: electrical conductivity, pH, zeta potential, concentration of ferric ions and
polyelectrolyte in filtrate, eST value, vacuum suction time (VST), dry solids content,
average specitic cake resistance, permeability, partiele size distribution, thixotropy,
and bond enthalpy. Three different flocculants were used per sludge type: ferric chlo
ride/lime, the polyelectrolyte Röhm KF975, and the type of cationic polyelectrolyte
used at the waste water treatment plant concemed. Flocculation was carried out
according to the conditions prescribed intheSTORA manual [1982, 1983].
The determination of the bond enthalpy has already been discussed in chapter 3. The
determination of the zeta potenrial will be discussed in chapter 6.
Sludge dewatering characteristics 63
In the next sections, methods to determine sludge dewatering characteristics are
discussed and some typical results are shown.
4.3 Composition
The dry solicts content of a sewage sludge sample or a filter cake is determined by
drying it in a furnace at 105 oe for 24 hours [NEN 6620]. The dry solids content is
expressed in termsof weight percentage of solids.
The ash content is related to the inorganic fraction of the sludge solids and is meas
ured by burning the dried sample at 600 oe for 30 minutes [NEN 6620]. The 'loss of
ignition' equals the organic fraction.
The pH and electrical conductivity are ordinary Iabaratory routines and do not need
any further introduetion here.
Dry solids and ash content, pH, and electrical conductivity of unconditioned sewage
sludge samples collected from different sludge treatment plants and used in the
characterization research are presented in table 4 .1.
Table 4.1 Overview of pH, electrical conductivity, dry solids and ash content of the investig
ated unflocculated sludges.
sludge type dry solids con- ash content pH electrical
tent (% of sludge solids) conductivity
(wt%) co·lm•l)
Mierlo 2.6 25 6.35 ; Veghel 3.2 33 6.50 0.21
Amsterdam 2.4 35 7.38 0.54
Lage Zwaluwe 0.9 36 6.90 0.25
Sewage studges from Mierlo, Veghel, and Amsterdam pass a thickener prior to
dewatering. Thickening accounts for relatively higher initia! dry solids contents.
The ash content of the Mierlo sludge sample is smaller compared to the other sludges.
Mierlo sludge is not stabilized, whereas at the other plants biological stabilization
processes occur. During sludge stabilization the organic fraction of the sludge solids is
reduced (see section 2.3).
64 Chapter 4
The pH of Amsterdam sludge is higher than the pH of the other sludges. Amsterdam
sludge bas been anaerobically digested. Aerobic bacteria are virtually not present in
the sludge sample. These bacteria are responsible for the following reaction:
Hydrogen carbonate is weakly acidic in nature and as a result the pH decreases.
The electrical conductivity of Amsterdam sludge is two to three times the conductivity
of the other sludges. Apparently a larger amount of ions or small charged particles are
present in the sludge sample.
Addition of lime to sludge increases the pH to about 12 (see figure 6.14). An increas
ing amount of added ferric chloride decreases the pH (see figure 6.10). Ferric ions are
acidic in a neutral pH medium (see secdon 6.4).
The increase of the ferric chloride dosage raises the electrical conductivity of the
sludge suspension. At relatively high ferric chloride dosages (250 to 300 glkg ds) the
electrical conductivity is about 1 to 2 0·1m·1 [van Berlo, 1993].
The type and added amount of polyelectrolyte do not influence the electrical conduc
tivity and pH of the sewage sludge sample [van Berlo, 1993].
ATP (Adenosine-5'-Triphosphate) is a specific indicator of cell viability because it
exists only in living cells. A TP acts as an energy buffer for living cells. The quantity
of ATP can thus be used as an indicator of the amount of living biomass in sewage
sludge samples [Patterson et al., 1970]. The technique to determine the A TP content is
based on the light-producing (luminescent) reaction with the Iuciferase enzyme derived
from fireflies. Since the Iuciferase enzyme will not penetrate bacterial cells, the ATP
must be extracted and presented in a solution for measurement. The total light output
(photons) is directly proportional to the amount of ATP present in a reaction mixture.
The light intensity is measured with a spectrophotometer.
Figure 4.1 presents measured ATP contentsas a function of the storage time (in days)
of two different sludges. The storage temperature was 5 oe. For both studges the A TP
content slightly increased with storage time.
Sludge dewatering charaderistics 65
25
20 -fl.) "CS bO 15 El -bO ::s. ._
10 ~ E-4 <
5
0
________________ .,.. _____________________________ .,.. -------------
1 2 3
day
Fig. 4.1 Measured ATP content of two sludge samples as a junction of the starage time (in
days). (•) Mierlo sludge, ( •) Amsterdam sludge. Starage temperature was 5 oe.
However, a decrease of the ATP content would be expected. During storage at 5 oe living cells die off. Hanstveit et al. [1977] measured ATP contents of activated sludge
samples as a function of the storage time under aerobic and anaerobic circumstances at
15 oe. In most cases the ATP content decreased during storage time.
It can be concluded that the use of this technique to measure the A TP content of
sewage sludge samples is questionable [Eikel boom, 1994]. Many technique factors like
temperature, ionic concentration of the sample, and presence of enzymatic inhibitors
in the extracted samples influence the light emission. Moreover, there are some objec
tions to this technique when it is applied to sewage sludge samples. The biochemica!
reactions occurring during storage of the samples are unknown. Possibly A TP is
produced or ruptured in these reactions.
66 Chapter 4
4.4 Filtration and expression
The dewatering of sewage studges in mechanical dewatering equipment can be
subdivided into two pbases: the tiltration and subsequently the expression pbase [Yeh,
1985; La Heij, 1994]. At the start of the solid-liquid separation process, a pressure is
exerted on the sludge slurry. Solid particles and/or flocs are retained on the filter
medium which acts as the separating agent. A cake is built up from the filter medium
and the clear liquid (filtrate) passes through. The cake thickness L increases in time.
Because of the frictional losses arising from the liquid flow through the cake, there
will be a hydraulic pressure gradient. In the tiltration phase the fluid pressure near the
filter medium approaches zero and the hydraulic pressure at the slurry-cake interface
remains constant (eqnal to the applied pressure). The drag on each partiele is commu
nicated to the next particle. Consequently the (so-called) solid compressive pressure
increases as the medium is approached. At every position in the cake the sum of the
hydraulic and compressive drag pressure equals the applied pressure. The tiltration
pbase appears to be described by a model in which non-linear elastic material
behaviour was assumed [La Heij, 1994].
The expression pbase starts when the slurry has disappeared and the cake is com
pressed. A material is said to be compressible if the volume of the material reduces
when pressure is exerted on it. Sewage sludges are highly compressible materials. As
a result, the average liquid content of a sewage sludge cake changes with changing
pressure. The cake thickness decreases in time in the expression phase. At the end of
this phase the liquid content and hydraulic pressure are uniform throughout the cake
and the liquid flow is zero. The expression phase can be described with a non-linear
visco-elastic model [La Heij, 1994].
In order to simulate the tiltration and/or expression pbase at laboratory scale and to
characterize the dewatering behaviour of sludges, three measuring devices were used:
the filtration-expression cell (FE-cell), the Modified Piltration Test (MFT), and the
Capillary Suction Time (CST) apparatns. The compression-permeability cell (CP-cell)
was used to determine dewatering properties of filter cakes under equilibrium
conditions.
Sludge dewatering charaderistics
r---·-------------- sludge sample
---~air escape ~~--- gas pressure
------~ perspex cylinder
--~piston
-----------~ sludge cake ----~filter medium
Jr---f-·-~ filtrate
D~:l::::::--~ balance
clasp
Fig. 4.2 Schematic drawing of the filtration-expression equipment.
4.4.1 The filtration-expression cell
67
A schematic drawing of the developed filtration-expression cell (FE-cell) is presented
in tigure 4.2. The cell consists of a perspex cylinder (inner diameter 70 mm, height
100 mm) with a porons roetal plate which is covered with a filter paper (Schleicher &
Schuell, 5893, ref.no. 300209). At the beginning of the experiment a gas pressure
(range 0 to 1500 kPa) is suddenly exerted on the sludge sample. A filter cake is built
up and the liquid is collected in a beaker glass positioned on a balance. The balance
continuously registers the weight of the filtrate, from which the filtrate volume can be
derived. If expression is also carried out, a closed piston moves downwards and the
cake is compressed.
With this experimental set up it is possible to carry out experiments under different
process conditions, such as gas (or expression) pressure, type of flocculant, flocculant
68 Chapter 4
dosage, amount of sewage sludge (or final cake thickness) and mixing energies during
flocculation. In this characterization research only tiltration experiments were carried
out with this device.
Expression experiments, which take a few hours, were carried out by La Heij [1994],
who used the experimental data to verify the modelling of the dewatering behaviour of
sewage sludges; he also studied the influence of pressure, cake thickness and slurry
concentration on the expression behaviour. Dewatering parameters that may be
determined from the expression phase are equilibrium dry solids content, expression
time and visco-elastic parameters. The FE-cell may also be used to measure hydraulic
pressure and porosity profiles [La Heij, 1994].
The liquid flow through a compressible porous medium has been extensively described
by La Heij [1994]. The constitutive equations, relating porosity e and solidosity e. to
the compressive pressure p., and the relation between permeability K and compressive
pressure p, are one of the basic concepts of the model. These equations can be
determined by carrying out experiments with the compression-permeability cell (see
section 4.4.4). Since the CP-cell experiments are time-consuming (only one experi
ment per day) this measuring device was not incorporated in the working scheme of
characterization.
The compressibility of sewage sludges is very small in the tiltration phase. The
tiltration phase can therefore be described by the integrated form of Darcy's law for
incompressible media, in which tiltration time t is related to filtrate volume V
[Svarovsky, 1990]:
where A = cross-sectional area of the filter medium [m2]
<;, = concentration of solids in the suspension [kg.m-3]
LW= applied pressure difference [Pa] R",= filter medium resistance [m-1
]
aav = average specific cake resistance [m.kg-1]
11 = viscosity of filtrate [Pa.s]
(4.1)
In deriving equation (4.1) it is assumed that the pressure difference across the cake LW
is constant. The filter medium resistance R", can be obtained by doing an experiment
with pure liquid (cv=O). Since filter medium blinding may be regarded as a minor
Sludge dewatering characteristics 69
problem, Rro may be assumed to be constant in the dewatering period [La Heij, 1994].
The average specifïc cake resistance am a characteristic of the sludge cake, follows
from a regression of the experimental data with equation ( 4.1). This resistance was
used as a characterization parameter in this study. All tiltration experiments were
carried out with 100 ml sludge (exclusive flocculants), at a constant applied gas
pressure of 200 kPa. Flocculation was carried out according to the conditions
prescribed in the STORA manual [1982, 1983]. Mixing of flocculant with sludge was
carried out with a standard stirrer positioned in a 250 ml beaker. The distance from
the stirring rod to the bottorn was 15 mm. In the case of flocculation with iron chlo
ride/lime, fust sludge was mixed with iron chloride by stirring at a speed of 1000 rpm
for 15 s. The suspension pH decreased (see tigure 6.10). Afterwards lime was mixed
with the sludge sample with by stirring at a speed of 500 rpm for 60 s. Mixing of lime
with sludge led to an increase of the sludge suspension pH to 12 (see tigure 6.14).
Precipitation of insoluble ferric hydroxide (Fe(OH)3) occurred. In the case of condi
tioning with po1yelectrolyte, s1udge samples were mixed with a stirring speed of 1000
rpm. In the experiments an optimum stirring time, which gives the highest dry solids
content, was used. The optimum stirring time depends on the type of sludge, type of
polyelectrolyte, and the dosage used, and was determined for every combination of
sludge type-polyelectrolyte dosage with the Modified Piltration Test (see section
4.4.2). In the figures 4.3 and 4.4 results are shown of tiltration experiments with
Mierlo sludge flocculated with iron chloride/lime and polyelectrolyte (Röhm KF 975),
respectively.
From these figures it can be concluded that the average liquid rate for the sludge
conditioned with polyelectrolyte is much larger than the liquid rate for the sludge
conditioned with iron chloride/lime.
In the figures Darcy's equation (equation (4.1)) is also presented. Although incom
pressible cake tiltration was assumed, the model describes the experimental data very
wen.
In both experiments a sudden increase in filtrate volume can be observed. This can be
attributed to formation of cracks in the filter cake. The experimental data just before
the moment the cake cracked were used to fit with Darcy's equation.
The average specific cake resistance strongly depends on the amount of flocculant
used.
Fig. 4.3 Filtrate volume as a function of time according to experiment (• • •) and model (-).
Mierlo sludge jlocculated with 100 g FeCVkg ds and 200 g Ca(OH)zlkg ds.
110
100
~ 90
- 80
j 70
80
~ 50
J 40
30 20
lO
0
• • • • • • •
o ~ 2 s 4 s a 7 a 9 20 n u u H u m ~ time (s)
••• __., __ tel --- Mod!al
Fig. 4.4 Filtrate volume as a function of time according to experiment (•••) and model (-).
Mierlo sludge jlocculated with 9 g polyelectrolytelkg ds.
Sludge dewatering characteristics 71
In tigure 4.5 the specifïc cake resistance is depicted as a function of the dosage of
ferric chloride. At low iron chloride dosages, the specific cake resistance is relatively
high due topoor flocculation conditions. Only a small part of the sludge solid surface
was occupied with ferric hydrolysis products. As a result polymerization of the
hydrolysis products did virtually not occur. With increasing flocculant dosage the
average specific cake resistance strongly decreased. Polymerization of the ferric
hydralysis produelS occurred, causing the formation of flocs. After a certain dosage
the specific cake resistance slightly increased, due to restabilization of the particles. At
excess flocculant dosages, partiele surfaces became saturated with specially adsorbed
iron hydrolysis products. The sludge particles had a high and positive charge and
electrostatic repulsion prevented further flocculation. A minimum average speci:fic
cake resistance and thus an optimum flocculant dosage (or small range) were found.
At tbis optimum dosage the sludge can be dewatered at the highest rate. An optimum
iron chloride dosage (or small range) was found for every sludge investigated in tbis
study [van Berlo, 1993]. Optimum ferric chloride dosages were found in the range of
90 to 120 g FeClikg ds and were independent of the sludge type (and thus slurry
concentration).
0 20 40 60 80 100 120 140 160 180 200
dosage of iron chloride (g/kg ds)
Fig. 4.5 Average specific cake resistance aav as a function of the dosage of FeCl3 (Mierlo
sludge). The lime concentration was kept constant at 200 glkg ds.
72 Chapter 4
-i -~
1013 ~----------------------------------------~
• tS 0 g • -lli'J • .... ! • • -! •
<.> <.> ;::: .... <.> 0 c:::a.. lli'J
2 4 6 8 10 12
dosage of Röhm K.F97S (g/kg ds)
Fig. 4.6 Average specific cake resistance cxav as a function of the dosage of polyelectrolyte
Röhm KF975 (Veghel sludge). Optimum mixing conditions.
The dosage of iron chloride used in the sludge treatment plant in Mierlo was about 40
to 60 g/kg ds.
The same trends were found for sewage sludges floccnlated with different types of
polyelectrolyte [van Berlo, 1993]. In fignre 4.6 the specific cake resistance is depicted
as a function of the dosage of polyelectrolyte Rölnn KF975. The curve can be
explained by the consecutive phenomena: poor floccnlation, charge neutralization,
bridging mechanism and restabilization (see section 6.7). An optimum polyelectrolyte
dosage was also found: 10 glkg ds. The optimum polyelectrolyte dosage depends,
among other things, on the type of polyelectrolyte (molecnlar weight, charge density)
used. In the characterization research Rölnn KF975 was nsed as a standard polyelectr
olyte for which optimum dosages were found in the range of 9 to 13 glkg ds.
A rnininnJm specific sludge cake resistance at a certain coagulant (FeCl3 and cationic
polyelectrolyte Zetag 63) dosage was also found by Katsiris and Kouzeli-Katsiri
[1987]. Lotito and Spinosa [1990] used a cationic polyelectrolyte in their study and
found a minimum specific cake resistance at a certain polyelectrolyte dosage.
Typical differences in experimental results between sludges floccnlated with ferric
chloride/lime and polyelectrolytes are:
Sludge dewatering characteristics 73
• The average specific cake resistance cx.v of studges conditioned with FeC13/Ca(OH)2
at optimum flocculation conditions is five to ten times larger than cx.v for sludges
flocculated with polyelectrolyte (at optimum conditions). As a result the average
dewatering rate in the flitration phase is larger for sludges flocculated with polyelectr
olyte compared to sludge flocculated with ferric chloride/lime. The type of polyelec
trolyte (trademark) has a considerable influence on the average specific cake resis
tance. It was found that a dosage of 10 g of Röhm KF975/kg ds to Veghel sludge
yielded a specific cake resistance cx.v of 0.22·1012 mlkg, whereas the same dosage of
Superfloc C496 resulted in cx.v equal to 7 .3·1012 m/kg.
• The amount of added ferric chloride to obtain the optimum laboratory dosage is
about five to ten times larger than the amount of added polyelectrolyte.
In an additional research the effects of other variables like time dosage and mixing
intensity were studied [Janssen et al., 1994]. Expression experiments, which took 15
minutes, were conducted at 300 k:Pa. Increasing the lime dosage deercases the average
specific cake resistance cxav. Lime which acts as an ordinary filling material reduces
the compressibility of the cake. Consequently the three-dimensional sludge matrix will
get stronger, less pores will collaps and this way the dewatering rate will be prom
oted. However, a high lime dosage is disadvantageous because it introduces an excess
amount of dry solids (increase of volume and mass). This leads to an increase of
transport, dumping, and incineration costs. Moreover, the cake can be expressed less.
At a lime dosage of 200 g/kg ds, the dewateriug efficiency1 is at a maximum.
However, at this optimum dosage filter cakes will stick to the filter cloth in chamber
filter presses. For this reason, practical lime dosages are higher than the optimum
dosage. This problem needs more investigatiou.
The mixing energy largely influences the dewatering behaviour if polyelectrolyte is
used as conditioner. In a series of expression experiments, Mierlo sludge samples
were mixed with 4 g of polyelectrolyte Nalco 41/62 per kg dry solids with a stirring
speed of 1000 rpm. The stirring time was varied. The results of the series of experi
ments are shown in figure 4.7. The average specific cake resistance is depicted as a
function of the stirring time.
1 The dewatering efficiency is defined as the percentage of the initia! water mass removed in a filtration-expression process.
74 Chapter 4
stirring time (s)
Fig. 4. 7 aav as a function of the stirring time. Mierlo sludge flocculated with 4 g of polyelectr
olyte Nalco 41/62 per kg dry solids. Stirring speed was kept constant at 1000 rpm.
An optimum stirring time of 5 to 10 s was found. Mixing of polyelectrolyte Nalco
41/62 with sludge initially leads to large aggregates. A higher mixing energy accom
plishes floc breakup due to the high shear forces. A smaller floc/aggregate effective
diameter reduces the dewatering rate (see section 4. 7). Mixing energy does virtually
not influence the dewatering rate when iron chloride/lime is used as flocculant. Small
aggregates (20-200 /Lm) are formed which are more resistant against high shear forces
(see section 4.8).
The filtration-expression cell is an automated instrument to study the dynamic
dewatering behaviour of sewage sludges. The influence of different parameters
(flocculant type and dosage, pressure, slurry concentration) can be investigated. The
FE cellis not labour-intensive and can rather easily be used at sludge treatment plants,
for instanee to control and optimize the flocculation process. More research is
necessary to find out the added value of the FE-cell at waste water treatment plants.
Sludge dewatering characteristics 75
4.4.2. Modified Filtration Test
The Modified Piltration Test (MFT) bas been developed by TNO (Heide aod Kampf,
1983). ln figure 4.8 a schematic drawing of the MFT equipment, consisting of three
parallel tubes, is given. This test is a modification of the Büchner fiJtration test. ln the
set up shown, three experiments cao be carried out simultaneously. The test provides
insight both into the dewatering rate aod the final dry solids content after dewatering
[NEN 6691].
The test comprises a vacuum expression of conditioned sludge, which is carried out
with a 7 cm diameter Büchner funnel at 50 k:Pa for 10 minutes. Cracking of the sludge
filter cake is prevented in the following way. After the tiltration phase a plastic foil is
laid loosely over the Büchner funnel. On the foil a 2 cm layer of water is brought,
causing the plastic foil to fit closely at the upper side of the sludge aod the inside of
the funnel. After 10 minutes of expression the test is end ed.
Fig. 4.8 Schematic diagram of the MFT equipment.
The weight of the filter cake before aod after drying at 105 oe is determined. From
this test the following dewatering characteristics cao be determined:
76 Chapter 4
• 'Vacuum suction time (VST)', which is defined as the time needed to collect 60 mi
of filtrate. The VST is a measure of the average filtradon rate.
• Final dry solids content (MFTds) of the sludge cake based on total solids (sludge
solids plus flocculant plus additives).
• Water content of the sludge cake (MFTid) expressed in kg water per kg initial dry
solids in the sludge sample. This moisture content is corrected for the dry solids from
additives and provides a better judgement of the effect of e.g. the nature and amount
of flocculant on the dewatering process. Another parameter, which does not take into
account the amount of dry solids from additives, is the dewatering efficiency.
Results of MFT experiments carried out with Mierlo sludge flocculated with different
dosages of iron chloride are presented in figures 4.9 and 4.10. The lime dosage was
kept constant. Both the vacuum suction time and the MFTid show a minimum at a
dosage of 100 to 120 g of FeC13/kg ds. At this optimum dosage the sludges are dewat
ered fustest and the highest dry solids content is attained. Both the average specific
cake resistance (figure 4.5) and vacuum suction time (figure 4.9) show the sametrend
as a function of the ferric chloride dosage. Optimum flocculant dosages (always to be
regarded as a small range of dosages) were found with the MFT for other sludge
flocculant combinations [Heide and Kampf, 1983, van Berlo, 1993]. The VST of
studges flocculated with iron chloride/lime is typically ten times larger than the VST
of sludges flocculated with polyelectrolyte (at optimum stirring conditions).
The MFT is an inexpensive and suitable procedure for the characterization of sludge
for dewatering with belt presses, but less appropriate for cbamber filter presses.
Dewatering times in belt presses are about 10 minutes. Unfortunately, the dry solids
content attained in a MFT bas no predictive meaning for cbamber filter presses. This
can be explained by the rather different conditions in a cbamber filter press, viz. long
expressiou times and high cake thickness. The cake thickness in a MFT experiment
equals 5 mm and in the cbamber filter press a few centimeters. The cake thickness
influences the dewatering time: the dewatering time increases with the square of the
cake thickness [La Heij, 1994]. A disadvantage of the use of the MFT is that the
metbod provides no clear insight into the dynamic expression behaviour. Moreover,
only pressure differences in a small range (0 to 100 k:Pa) can be exerted.
600
500
400
E-t 300 1':1.2
> 200
100
0
Sludge dewatering charaderistics 77
0 20 40 60 80 100 120 140 160 180 200
dosage of iron chloride (g/kg ds)
Fig. 4.9 Vacuum suction time as a function of the dosage of FeCl3 (Mierlo sludge). The lime
dosage was kept constant at 200 glkg ds.
20
15
10
5
0 0 20 40 60 80 100 120 140 160 180 200
dosage of iron chloride (g/kg ds)
Fig. 4.10 MFTid as a tunetion of the dosage of FeCl3 (Mierlo sludge). The lime dosage was
kept constant at 200 glkg ds.
78 Cbapter4
4.4.3 Conventional CST apparatus
The conventional CST apparatus is a simple automatic instrument for determining the
dewaterability of sewage sludges in a quick way [NEN 6690]. CST was presented for
the first time by Baskerville and Gale [1968] and is still used as a dewatering test in
sewage sludge treatment plants. The conventional CST apparatus ( tigure 4.11) consists
of a cylindrical tube (inner radius 10 mm) resting on a reetangolar piece of filter paper
(Whatman no. 17 or Schleicher & Schuell ref. no. 382455). The filter paper is
sandwiched between two reetangolar plates of perspex. When sludge, or any other
suspension, is poured into the stainless steel cylinder, liquid will be sucked into the
paper nnder the influence of the capillary suction pressure and the sludge head. The
capillary suction pressure of the filter paper used is about 10 to 15 kPa [Baskerville
and Gale, 1968, Leu, 1981.]. Thus, with this apparatus only dewatering in the low
pressure range is studied. The liquid front will move in a radial direction, forming
more or less an ellipse due to the grain of the paper. At two different positions (r=6
and 13 mm) two electrodes are fi.xed in the perspex plates. When the liquid front
arrives at the first electrode, an electrical signal will be given to a chronometer,
whereupon time measurement will start. When the liquid front arrives at the second
electrode, the time measurement will end. The time needed to move the liquid front
from the first to the second electrode is called the 'Capillary Suction Time', abbrevi
ated CST. From this CST the dewaterability of the suspension can be estimated; a
small CST implies a good dewaterability. The CST value is, however, oot an intrinsic
dewatering parameter. The CST value depends on the slurry concentration and paper
properties (see section 5.5).
The CST value was used as a parameter in this characterization research. The result of
a series of experiments is presented in tigure 4.12. The CST value is depicted as a
function of the ferric chloride dosage added to Mierlo sludge. AB expected, a mini
mum CST value was fonnd. A minimum CST value at a certain inorganic flocculant
was also fonnd for other sewage studges investigated [Baskerville and Gale, 1968, van
Berlo, 1993].
Measurements of specific cake resistance aav (figure 4.5), VST (figure 4.9), and CST
(figure 4.11) as functions of the dosage of FeC13 gave the same results. lnorder to
correlate the specific cake resistance aav with the CST parameter one needs to specify
the slurry concentration.
Sludge dewatering characteristics
1: cylinder 2: sludge 3: sludge cake 4: electrode 5: timer 8: filter paper
Fig. 4.11 Schematic drawing ofthe conventional CST apparatus.
79
Eikurn and Paulsrud [1974] showed a fairly good relationship between the specific
cake resistance and the CST value divided by the total suspended solids concentration.
From our experiments it appeared that the conventional CST apparatus is not suitable
for unflocculated sludges and sludges flocculated with polyelectrolyte. The reproduc
ibility of CST measurements carried out with unflocculated studges is bad [van Berlo,
1993]. The dewatering process is very slow. Different cake structures are formed due
to the large inhomogeneity of the unflocculated sludge sample, resulting in large
deviations in the CST value.
Large sludge flocs/aggregates are formed when studges are flocculated with polyelec
trolyte (see section 4.7). Due to the small inner radius of the CST tube, cake forma
tion bardly occurs. Since hardly any resistance against liquid flow occurs in the tube,
small CST values are measured. In the scope of the study presented here a modified
CST instrument has been developed (see cbapter 5). The instrument provides the
possibility to continuously measure the position of the liquid front as a tunetion of
time and to calculate a specific cake resistance from the experimental data.
80 Chapter 4
200
lSO
-en -E-t 100 ti')
u
so
0 0 20 40 60 80 100 120 140 160 180 200
dosage of iron chloride (g/kg ds)
Fig. 4.12 CST value as ajunetion of the dosage of FeC/3 (Mierlo sludge). The lime concentra
tion was kept constant. Initia! dry solids content is 2.6 glkg ds.
4.4.4 The compression-permeability cell
The compression-permeability cell (CP-cell) was developed by Ruth [1946] and is
used to determine both the permeability and porosity as functions of the compressive
pressure. A schematic drawing of this cell is given in tigure 4.13. The cell consists of
a perspex cylinder with a porous metal bottom plate covered with a filter paper
(Schleicher & Schuell, 5893, ref. no. 300209). The double plated piston, which is
positioned in the cell, bas an upper solid plate and a lower porous plate. A sludge
(cake) s~ple is situated between the bottomsupport plate and the lower porous plate.
A gas pressure is exerted on the upper closed piston and in this way the sludge will be
expressed. The filtrate is collected in a beaker glass positioned on a balance, which
registers the weight of the water obtained. When the equilibrium situation bas been
reached, the mechanical pressure equals the compressive pressure p.. After equilib
rium the mechanical pressure is increased and the cake will again equilibrate. This is
repeated a few times.
Sludge dewatering characteristics 81
,. ... ______ ··---IE·····)o-displacement transducer
Ps ~='IME-- mechanica! pressure
closed ..... ~plate
···•• PL hydraulic pressure
E porous
........ plate ................ sludge cake
••·•••· .... botlom support plate
Fig. 4.13 Schematic diagram of the compression-penneability cell.
The cake porosity E or solidosity E,
followiug equation:
1-E) at any pressure can be calculated from the
where
u U1 + -- d - +-E [ 1 u']
(1-E) w d, dw
d,= density of dry solids [kg solids.m·3 solids] dw= density of pure water [kg water.m·3 water] u moisture content [kg water.(kg solids)"1
]
u' bound water content [kg bound water.(kg solidsY1]
(4.2)
Bound and/or iutracellu1ar water is not responsible for the free liquid flow iu the cake
and thus does not contribute to the filliug up of the porous space. The bound water
content u' is taken as 0.4 kg bound water/kg ds (see chapter 3). The average pure dry
solids density d, of Eindhoven/Mierlo sludge is 1280 to 1300 kg/m3• The cake
moisture content u at any pressure can be determiued from the experimental data:
82
where
Chapter 4
U = nt.v,oo + (~,oo ~) md.
ruw= mass of dry solids [kg] m..v . ." = mass of water in the cake at the end of the experiment [kg] ám..v = loss of water mass at any pressure [kg] ám..v,,., = totalloss of water in the experiment [kg]
After the experiment the cake is dried, so ruw and m..v."" are known.
(4.3)
Following this experimental procedure both equilibrium porosity € 00 and equilibrium
solidosity e,,,., as :functions of the compressive pressure p, are obtained [Tiller et al.,
1980, La Heij, 1994]:
(4.4)
and
e =e0
[1 + P,ltJ '·"'
8 Pa (4.5)
where 'A and fi are compressibility (or compactibility) coefficients. e0 and €80 are the
porosity and solidosity respectively, when the compressive pressure p, equals zero. p.
is a constant making the equations dimensionless. Equations (4.4) and (4.5) are called
constitutive relations.
The permeability of the porous sludge cake is measured by passing water through the
cake. The liquid flow Q1 through the cake is measured by using a balance which is
connected to a computer. The cake thickness L can be measured with a displacement
transducer (see tigure 4.13). When the cake thickness L and the hydraulic pressure
drop .6.P1 across the cake are known, the equilibrium permeability K." can be calcu
lated with the integrated Darcy's law:
(4.6)
where 1J is the viscosity of the filtrate.
Sludge dewatering characteristics 83
The dependenee of the equilibrium permeability K"' on the compressive pressure p, is
represented as:
(4.7)
Some of the typical results presented here are a part of the research carried out by La
Heij [1994]. The CP-cell was not incorporated in the working scheme of the sludge
characterization. The main reason was the rather long time it takes to carry out a
single experiment, i.e. one day. Only one sludge type was studied: Eindhoven sludge.
The power law functions as given by the equations (4.4), (4.5) and (4.7) showed to he
the most appropriate equations to describe the experimental data. Both the permeabil
ity and porosity decrease with increasing compressive pressure p,. The solidosity E,
increases with increasing compressive pressure. Permeabilities and solidosities versus
compressive pressure were measured for Eindhoven sludge flocculated with different
dosages of ferric chloride. Both the compressibility coefficients ó and B did not change
with the dosage. lt indicates that the dosage of iron chloride has a minor influence on
the compressibility of the cake. However, the dosage of iron chloride has a major
influence on the absolute permeability of sludge cake. An optimum flocculant dosage
of 100 g/kg ds was found.
CP-cell experiments with Eindhoven sludge flocculated with polyelectrolyte showed
that the compressibility (o=2.3) is higher than the compressibility of sludges
flocculated with iron chloride/lime ( ó = 1. 6). The higher compressibility is causcd by
the large weak flocs which are formed during polymerie flocculation (see section 4.7);
a higher value of Kv and lower value of E,0 were found. The initia! permeability Kv is
high, which results in quick dewatering during the initia! stage of the filtradon phase.
The average permeability in the filtradon phase K,.v is inversely proportional to the
average specific cake resistance aav according to [Tiller, 1975]:
(4.8)
A higher initia! permeability for studges flocculatcd with polyelectrolytes corresponds
with a smaller average specific cake resistance (see section 4.4.1).
84 Chapter 4
The polyelectrolyte dosage hardly influences the compressibility and the equilibrium
solidosity. After expression times of about a few hours, the sludge cakes are com
pacted to the same structure.
The CP-cell provides information on the dewatering behaviour of sludge cakes under
static conditions. It is a valuable instrument to determine the compressibility of sewage
sludges and the optimum flocculant dosages (in terms of permeability). However, the
time needed to carry out a single experiment is about one day.
4.5 Amount of iron in the filtrate
Knowledge about the iron concentration in the flltrate is important, as it possibly
provides insight into the flocculation mechanism induced by ferric chloride/lime.
Besides, from an environmental point of view, the amount of iron in flltrate must be
as small as possible.
The amount of iron in the flltrate can be determined by measuring the light absorption
of an iron complex with a spectrophotometer. First all the Fe3+ was reduced to Fel+.
Subsequently the flltrate was treated with 2,2 '-bi pyridine which formed a red coloured
iron-2,2' -bipyridine complex.
0 ..... CIO .::: 2.00 0.15 -.;::: 0 -s 0.12 .6 1.50 g • .... 0.09 ·-'"0 1.00 -8 6 0.06 '"0
CIO ..... 0.50 0
0 • 0.03 toO CIO i:i • ~ ~ 0 0.00 0.00 u .... 0 50 100 150 200 250 300 0 0.
dosage of FeCl3 (g/kg ds)
Fig. 4.14 Percentage of added iron in the filtrate (•) and iron concentration in filtrate (A) as
a function of the dosage of ferric chloride. Mierlo sludge jlocculated with 200 g of time/kg ds.
Sludge dewateriug characteristics 85
Cl) .... CIS .... 100 0.50 ...... -~ 90 -Cl) • ---5 -80 0.40 0 !:= 6. El ·- 70 • • lilt • -g = 60 • • 0.30 0 .... -·-·- 6. ............
"'0 50 ' CIS Cl) ~:=
"'0 40 0.20 - = "'0 6 Cl)
CIS 30 6. ä ..... 0
0 u Cl) 20 0.10 g t>Q 6 CIS 10 6 .... d ·-Cl) 0 0.00 u .... 0 50 100 150 200 250 300 350 Cl) Q"
dosage of FeCl3 (g/kg ds)
Fig. 4.15 Percentage of added iron in the filtrate (•) and iron concentration in the filtrate ( t:.)
as a function of the dosage of ferric chloride added to Mierlo sludge. No addition of lime.
The light absorption was measured at 523 nm, which is the optimum wavelength for
Fe2 + absorption. The iron concentration in the filtrate follows from the Lambert-Beer's
law.
Examples of two series of measurements are presented in figures 4 .14 and 4 .15. Both
the percentage of added iron in the filtrate and the iron concentration in the filtrate are
depicted as a function of the iron chloride dosage added to Mierlo sludge.
In one series the time dosage was kept constant at 200 glk:g ds ( tigure 4 .14). The
lowest percentage of added iron in the filtrate corresponded with the smallest average
specific cake resistance. At the optimum dosage of 150 g of ferric chloride/kg ds, the
absolute iron concentratien in the filtrate was minimal (10-6 mol/I). The smallest iron
concentration in the filtrate was also found at optimum ftltration conditions for the
other sludges investigated [van Berlo, 1993].
The different flocculation mechanisrns are responsible for the results found. Ferric
hydrolysis products are formed which specifically adsorb at the surface of the sludge
particles. Li.kewise precipitation of insoluble ferric hydroxide occurs, which stays in
the filter cake rather than passing through the filter medium. At a relatively small
ferric chloride dosage (25 g/kg ds) the percentage of added iron in the filtrate is
realtively high. This is probably caused by colloidal and small particles in the f:tltrate
86 Chapter 4
which contain adsorbed iron and disturb the measurement. At high ferric chloride
dosages ( > 200 glkg ds) a slight increase of the percentage of added iron in the filtrate
has been measured. All active sites on the surface of the sludge particles/flocs are
likely occupied by ferric hydrolyis products. Moreover, there may oot be enough lime
to precipitate all the FeH. The excess amount of added iron dissolves in the filtrate.
Figure 4.15 shows the measured iron concentation in the filtrate resulting from
experiments with sludges that were only conditioned with ferric chloride. The amount
of iron in the filtrate is much higher without lime than with lime being added to the
sludge. This is due to the :tàct that Fe(OH)3 is precipitated when lime is added (an
increase of pH toabout 12).
The condusion cao be drawn that the determination of the iron concentradon in the
filtrate seems to provide useful information on the flocculation mechanisms and
optimum dewatering conditious when sewage sludges are conditioned with ferric
chloride/lime. Experiments with this characterization test cao be performed simply and
quickly.
4.6 Polyelectrolyte concentration in the filtrate
Several techniques for determining the polyelectrolyte concentradon in the filtrate
were investigated [van Berlo, 1993]. The most promising method is the formation of
an ionene/cobaltphthalocyanine complex [van Welzen, 1989]. It is well known that
cobaltphthalocyanine (CoPc(NaS03) 4) in a neutral aqueous solution exhibits two
absorption maxima in the visible region, duetopartlal dimerisation according to:
2M~D
The dimeric complex D shows a maximum at a wavelength of 628 om, whereas the
monomer M has its maximum at 662 om. The presence of increasing amounts of
ionenes causes a shift of the absorption maximum from 662 om to 628 om (see
Appendix 4). The shift is thought to be the result of the formation of polynuclear
CoPc(NaS03) 4 aggregates. Phthalocyanine absorbances were measured with visible
light spectroscopy.
A certain amount of ionene (positively charged polymer molecules) is needed to shift
the absorption maximum. This amount depends on the type (trademark) of polyelectro-
Sludge dewatering characteristics 87
lyte used (molecular weight, charge density). The minimum amount of polyelectrolyte
per g CoPc(NaS03) 4 was determined for every polyelectrolyte used in this study:
0.74 g of Röhm KF975/g CoPc(NaS03) 4 ,
9.7 g of Zetag 63/g CoPc(NaS03) 4 ,
71 g of Nalco/g CoPc(NaS03) 4•
Filtrate samples were analysed with the technique described. A certain volume of
filtrate is needed to shift the absorption maximum from 662 nm to 628 nm and
corresponds with the predetermined amount of polyelectrolyte. The ratio between the
amount of polyelectrolyte and the filtrate volume equals the concentration of polyelect
rolyte in the filtrate. Unfortunately, it appeared that a quantitative determination of the
polyelectrolyte concentration in the filtrate is not possible. The shift of the absorption
maximum was not clear (high noise level), because of the disturbance of the measure
ment by colloidal particles. However, a trend could be observed that an increase of
the amount of polyelectrolyte added to sewage sludge increases the polyelectrolyte
concentration in the filtrate. The amount of colloidal particles in the filtrate may be
minimized by centrifugation of the filtrate samples prior to the absorption measure
ment. More investigation into this problem is needed.
4. 7 Partiele size distribution
The size of sludge aggregates (flocs) is of major interest in thickening and dewatering
operations. Karr and Keinath [1978] experimentally determined the significanee of
partiele size distribution of sewage sludges. Sludge solids were fractionated into the
following size ranges: settleable (> 100 p.m), supracolloidal (1 to 100 p.m), colloidal
(0.001 to 1 p.m) and dissolved ( <0.001 p.m). Sludge samples with high supracolloidal
solids concentrations had high specific cake resistances.
The microscopie appearance of different sludge samples under investigation are shown
in tigure 4.16. The appearance of unflocculated sludge (figure 4.16 a) is not clearly
defined, owing to a great number of dispersed particles in the liquid phase. Studges
flocculated with iron chloride and polyelectrolyte Röhm KF975 display flocs and a
few dispersed particles ean be seen in the liquid. The nature of the sludge flocs
produced by conditioning with ferric chloride is different compared to that of sludge
flocs obtained by flocculation with polyelectrolyte (see tigure 4.16b and 4.16c).
88
1000 J.tffi (a)
Chapter 4
1000 JLffi
(c)
1000 J.tffi
(b)
Fig. 4.16 Microscopie photographs of sludge: a) uriflocculated, b) flocculated with iron chlo
ride, and c) flocculated with polyelectrolyte Röhm KF 975.
The 'image analysing technique' was used in this research to examine the partiele size
distribution of a sewage sludge sample. The experimental equipment consists of an
optical microscope, a camera and a computer (figure 4.17). A picture of a sludge
sample was registered and digitalized and shown on the screen of the computer
system. The projected surface area of the particles/flocs was determined with the TIM"
package and subsequently mathematically converred to an 'effective diameter' of
Sludge dewatering characteristics 89
spherical particles tbat have the same projected surface area as those of the measured
particleslflocs. Results were compiled to produce a partiele size distribution.
Different types of partiele size distributions can be defined: partiele size distribution
by number, length, surface or mass (volume). In this study the partiele size distribu
tion was based on the number of particles within a size interval.
camera
microscope
sludge sample
fl.~·. /. -.'. =
'image analysing'
Fig. 4.17 Schematic diagram of the image analysing equipment.
The distributions are given as either frequencies f(x) or undersize cumulative fre
quencies (expressed as fractions or percentages) F(x). The measured undersize
cumulative partiele size distribution is approximated by the Harris three-parameter
equation:
F(x) 1 _ [ 1 -(~)·f (4.9)
where Xo is the maximum diameter in the distribution and a1 and b1 are constants. The
frequency distribution is obtained by differentiation of equation (4.9).
90 Chapter 4
(4.10)
There is a great number of definitions for averaged partiele sizes. In this study, the
median diameter is defined as the partiele size for which half of the particles is larger
and half is smaller, i.e. the size which divides the area under the distribution fre
quency curve into two equal parts. Thus, at the median diameter F(x)=0.5.
The partiele size distribution of sewage sludge samples conditioned with different
types and dosages of flocculants were determined and evaluated according to the above
method. Figures 4.18 and 4.19 show the calculated partiele size distributions of
studges conditioned with different dosages of iron chloride and polyelectrolyte,
respectively. The calculated distributions were obtained by determining the three
parameters of the Harris equation. Sludge samples were flocculated at optimum
mixing conditions. It appeared that deviating the stirring and mixing conditions will
change the partiele size distribution, especially for sludges conditioned with polyelect
rolyte [van Berlo, 1993].
The effect of an increasing amount of flocculant is a shifting of the distribution curves
towards greater diameters. However, at sufficiently high dosages shift in size to larger
particles does virtually not occur any more (see tigure 4.20). This is attributed to
restabilization of the particles.
Typical values for the median diameter are:
- 5-10 pm for particles in unflocculated sludges,
- 8-100 pm for particles/flocs in sludges flocculated with ferric chloride,
- 500-2000 pm for particleslflocs in studges flocculated with polyelectrolyte.
In tigure 4.20, the median diameter is depicted as a function of the polyelectrolyte
dosage. Larger-sized particles (in terms of the median diameter) result in correspon
ding reductions in the sludge specific resistauce (see tigure 4.6). This empirical
relation was also found for other sludge-flocculant combinations [van Berlo, 1993;
Knocke and Wakeland, 1983]. Knocke and Wakeland [1983] examined the impact of
the floc size distribution on the flitration rates of metal hydroxide sludges. Any
treatment varlation (e.g. polymer addition or mixing intensity) that resulted in a shift
in partiele size distribution had a corresponding effect on the specitic resistauce of
metal hydroxide sludges.
Sludge dewateriug cbaracteristics 91
0.045
e 0.040
Sr 0.085 120 .......,
~ 0.030 40
t 0.025 - 80 0.020
ts 60 0.015 = :8 0.0:10
~ 0.005
0.000 ---- -..:-_::.:_-.::.:.::::. -::::...::-- - -
0 25 50 75 100 :125
diameter (J.tm)
Fig. 4.18a Partiele size distributions of Veghel sludge flocculated with different dosages of
iron chloride. The numbers in the graph represent the iron chloride dosage (in glkg ds).
1.0
--- 0.9 I
-o.a t ~7 0.8
~ o.5
j 0.4
0.3
0.2
0.1
0 25 50 75 1.00 1.25
diameter (}..t,m)
Fig. 4.18b Cumulative partiele size distributions (Harris equation) of Veghel sludge flocculat
ed with different dosages of ferric chloride. The numbers in the graph represent the iron
chloride dosage (in g!kg ds).
92
0.001.5
e ~ 0.0012 ........, til
~
i 0.0009
'êS 0.0006
@ û 0.0003
~ 0.0000
I
0
f I I,
I I
' I
Chapter 4
500 1000 1500 2000
diameter (,.Lm)
Fig. 4.19a Partiele size distriburions of Veghel sludge jlocculated with different dosages of
polyelectrolyte Röhm KF 975. 1he numbers in the graph represent the polyelectrolyte dosage
(in g!kg ds).
1.0
..-0.9 I
-o.s
j 0.7
! 0.6
(3o.5 j:: j 0.2
0.1 ..... --.--T""""""T""..,.......--.--..-.-...... --.--.,....,....,.......-.-T""""""T""..,-.-....-r..,.......-.-..-.-.,....-.-..-.-..,.......-.-.......... .,...
0 500 1000 1500 2000
diameter (J.Lm.)
Fig. 4.19b Cumulative partiele size distributions (Harris equation) of Veghel sludge flocculat
ed with different dosages of polyelectrolyte Röhm KF 975. 1he numbers in the graph represent
the polyelectrolyte dosage (in glkg ds).
Sludge dewatering characteristics 93
800 • - • •
El 600 ::s. • - • ~ -Q.)
f1 400 ;a
f;1 ·-"Ct 200 ä • • 0 0 2 4 6 8 10 12
dosage of KF975 (g/kg ds)
Fig. 4.20 Median diameter versus the dosage of polyelectrolyte Röhm KF975.
CampbeU and Crescuolo [1982] investigated the impact of the sludge partiele size
distribution on the rheological behaviour. They determined partiele size distributions
using a Coulter Counter and found that a raise of the polymer dosage from 0 to 4 kg/t
dry solids increases the mean partiele size from 19 p.m to 14 7 p.m.
The specific surface <1v of a partiele is defined as the surface of the partiele divided by
its volume. Por a sphere, the specitic surface is inversely proportional to the partiele
diameter x:
6 (4.11) x
The specific surface distribution by number can be obtained by substitution of equation
(4.11) in equation (4.10):
f(a,.) (4.12)
In tigure 4.21, the frequency distributions of the specific surface of Veghel sludge
conditioned with different dosages of polyelectrolyte KF975, is given.
94
a ~ ......... til Q)
i ~
'ïS 8 t3 J)
0.0015
0.0012
0.0009
0.0006
0.0003
o.oooo 0.000
Chapter 4
'. f
. '
-2
0.005 0.010 0.015 0.020 0.025 0.030 0.035
specific surface (1/ /Lm)
Fig. 4.21 Specific suiface distributtons of Veghel sludge flocculated with different dosages of
polyelectrolyte Röhm KF975. The numbers in the graph represent the polyelectrolyte dosage
(in glkg ds).
Any shift in size to larger particles will rednee the specific surface area per unit
weight of slndge solids in the slndge matrix with a corresponding rednetion in surface
shear stresses encountered during dewatering. Conseqnently, the specifïc cake
resistance is smaller. Surface shear stress results from the difference in the partiele
and fluid velocity and is obtained by dividing the fluid drag force by the surface area
ofthe floc.
Theoretica! equations re lating the permeability K ( or specific cake resistance) of a
porons bed to the specitic surface area (or partiele diameter) like the Kozeny-Carman,
Happel-Breuner and Brinkman equations cannot be used for sludge filter cakes [van
Veldhuizen, 1991]. For iustance, the Kozeny-Carman equation is given by:
Sludge dewatering characteristics 95
K (4.13)
where E is the porosity, 3v is the specitïc surface area, <4 is the median particle!floc
diameter, and kKc the so-called Kozeny constant, which is specitic for every porons
system. The Kozeny-Carman equation was derived for Poiseuille flow in an incom
pressible bed, whereas sludge is a highly compressible materiaL Moreover, the
specitic surface area of sludge particles!flocs is not constant during compression, but
will change. Because of this time-dependent complex structure and extremely compli
cated liquid flow, in this case empirica! relations should be used. The measured
overall permeability of a sludge cake is determined by liquid flow alongside the sludge
flocs as wellas through the flocs. Combination of equations (4.4) and (4.7) yields an
experimental relationship between permeability K"' and porosity E00 of a sewage sludge
cake, which takes into account all the sludge dewatering properties and conditions
mentioned:
(4.14)
Allthough it is difficult to obtain a representative and objective measurement of the
size of flocs/particles due to the irregular floc shapes, changes in measured partiele
size distribution due to different flocculant dosages can be determined and correlated
with the average specific cake resistance. One should also notice that the measured
particle!floc size distribution is a distribution of unstressed particles. Another disad
vantage is the labour-intensiveness of the measuring technique, which makes this
characterization testnotpractical for sludge treatment plants.
5.8 Rheological properties
The influence of chemica! conditioning on the rheological properties of sewage sludge
was exarnined by using a Searle type coaxial rheometer (Contraves Rheomat 115). A
schematic diagram of the rheometer is given in tigure 4.22.
A sewage sludge sample is introduced into the small gap between two cylinders. The
inner cylinder (Mooney-Ewart spindle) rotates and the outer cup, which holds the
sample, remains stationary. In an experiment the angular velocity w of the spindie is
adjust and read
Cbapter 4
motor
measurenutnt of torque
Fig. 4.22 Schematic diagram of the Searle type coaxial cylinder rheometer.
increased in 15 small time-steps (5 seconds) until tbe maximum speed is reached and
is subsequently decreased to zero, also in 15 steps.
The shear rate i' is related to the angu1ar velocity w according to:
. dw R.vw 'Y = r- = -- 13.36w
dr Rb-Rs (4.15)
where Rt, is tbe radius of tbe outer cup (2.425·10-2 m), R. the radius of the inner
spindie (2.25·10-2 m) and Rav = (Rt, + R.)/2. The maximum shear rate reached in an
experiment is 1008 s·1• The torque T needed to rnaintaio tbe angu1ar velocity is
measured and converted to a shear stress r:
T T T=-----
211T2l 2'11'R!l (4.16)
where lis the lengthof the iuner cylinder. Equations (4.15) and (4.16) are valid when
the gap widtb (Rt,-R.) is much smaller than the radius of tbe outer cylinder. The appar
ent viscosity rJ of tbe suspension is defined as:
Sludge dewatering characteristics 97
T (4.17) .y
The result of an experiment is called a rheogram. The flow in a rheometer must be
laminar. Turbulence occurs in a Searle type rheometer above a critical Reynolds
number Recrit:
Re " ~R,(R" -R,)p > 41.3 J R" Re"' 11 Rb -R,
(4.18)
The turbulent flow consists of so-called Taylor vortices. Inequality (4.18) means that
Taylor vortices hold out above an apparent viscosity rJ > ')r/50 mPas.
A typical rheogram of a sewage sludge flocculated with iron chloride is presented in
tigure 4.23. From this rheogram the following information is obtained:
• The rheological behaviour of sewage sludge can be interpreted as pseudoplastic
flow, also called shear thinning, which is presented with the following equation:
(4.19)
where T0 is the yield stress and 'flv(.Y,t) the plastic viscosity, which depends on the
shear rate and time. The yield stress is the minimum stress which must be overcome
before true flow occurs. Below this yield stress, molecular bindings and interaction
between particles (London-van der Waals force, electrastatic repulsion) result in a
three-dimensional floc network. Above To the floc networkis ruptured and sludge flow
may occur. There is no strong and extensive floc network in sewage sludge, so the
yield stress To is very small. The plastic viscosity k decreases with increasing shear
rate, which may be the result of the rearrangement of the particles/flocs in the flow
direction.
At each shear rate an equilibrium exists between the aggregation of particleslflocs and
the degradation of flocs. The equilibrium floc size is determined by the various forces
acting on the aggregates: London-van der Waals attraction, electrostatic repulsion,
hydrodynamic forces, and thermodynamic forces like the Brownian motion. Both the
London-van der Waals force and electrostatic repulsion increase with increasing floc
diameter ( equations ( 6. 18) and ( 6.14)). The repulsive hydrodynamic force arises from
the distartion of the fluid flow due to the presence of particles, and increases with
increasing shear rate. The equilibrium size also depends on the flow history.
98 Chapter 4
13
12
11
~ 10 9
1:- 8
I 7 6
l 5 4
3
z 1
0
0
shear rate j (1/s)
Fig. 4.23 Shear stress as a function of shear rate for sludge flocculated with 90 g iron
chloride/kg ds; (-) Ascending curve, (---) Deseending curve.
Differences in equilibrium median floc diameter due to different constant shear rates
could, however, not be observed with the image analysing technique. CampbeU and
Crescuolo [1982] measured only a slight increase in mean partiele diameter from 50 to
60 p.m after the particles had been sheared to 113 s·1•
The plastic viscosity 'lip can be determined as a function of time in a stationary shear
experiment. It appeared that for flocculated studges the viscosity decreases with time
[van Berlo, 1993]. The decline in viscosity is stronger for higher dosages of
flocculant. Unflocculated sludges show no rednetion of viscosity as a function of time.
• The change of the slope in the curve at a shear rate of 700 s·1 indicates the onset of
turbulence. The slope change depends, among other things, upon the type of sludge
used. CampbeU and Crescuolo [1982] determined a break in the rheogram of anaerob
ically digested sewage studges at a shear rate of 400 to 500 s·1•
• 1t is notabie that the deseending curve lies below the ascending curve. This loss in
shear strength of the sludge flocs is called thixotropy and is a typical sludge property.
The rheology of the floccnlated sludge suspension has been altered during the initia]
phase of the test. At increasing rates the floc network breaks down, whereas at
Sludge dewateriug characteristics 99
decreasing rates the flocs are allowed to regrow. However, the end stress does not
coincide with the yield stress 7 0• lt indicates that no full bnilt-up of the original floc
structure occurs. This irreversible loss in viscosity is called rheodestruction.
Thixotropie behaviour of anaerobically digested sewage sludges was also found by
Campbelland Crescuolo [1982].
The surface area between the two curves can be interpreted as the dissipated power
per unit volume and is a measure of both the degree of thixotropy exhibited by the
sludge, and floc strength of the aggregates. In tigure 4.24 the degree of thixotropy is
depicted as a function of the dosage of ferric chloride added to Veghel sludge.
Unconditioned sludges do not show thixotropie behaviour. The particles in unfloccu
lated sludges appear to be stabie and are relatively unaffected by shear. As the ferric
chloride dosage increases, the sludge sample becomes more susceptible to shear and
some breakdown in floc structure may be occurring. Studges flocculated with ferric
chloride/lime and polyelectrolyte exhibit a maximum value of the degree of thixotropy
(or dissipated power per unit volume) at a certain flocculant dosage. Different sludges
investigated exhibit different valnes of viscosity and degree of thixotropy [van Berlo,
1993].
Some problems arise when studges conditioned with polyelectrolyte are investigated
with the rheometer used. Relative large aggregates are formed (median diameter 500
to 2000 /Lm), which may be ruptured as a result of the small gap width (1750 /Lm). A
spindie having a smaller diameter should be used. It might be advantageous to study
this problem in more detail.
Partiele size analyses were conducted in conjunction with the rheological studies. The
partiele size distribution determined from the same sewage sludge sample has been
depicted as a function of the ferric chloride dosage in tigure 4 .18a. The partiele size
distribution and the changes that occur in the distribution as a result of chemical
conditioning will have a significant impact on the rheological behaviour. A maximum
degree of thixotropy (minimum floc strength of aggregates) corresponds with a
maximum median partiele size of the initially flocculated sludge sample. The maxi
mum degree of thixotropy is reached at the optimum flocculant dosage for dewatering.
This empirica! relationship has also been found for other sludge-flocculant combina
tions. The more fragile structure of larger flocs is due to their formation process in
which smaller more dense aggregates have collided which remained linked at their
points of contact, resulting in a more porons arrangement of lower overall density and
hence in a weaker structure [François and Haute, 1985].
100 Chapter 4
600 • •
ç;" a
400 ~ • t • 0 • b ~ 200 • :s
• 0 0 20 40 60 80 100 120
dosage of iron chloride (g/kg ds)
Fig. 4.24 Degree of thixotropy as a junction of the dosage of ferric chloride added to Veghel
sludge.
The general condusion can be drawn that the Searle type rheometer is a proper and
fust instrument to determine the degree of thixotropy (measure of floc strength) of
unflocculated sludges and sludges flocculated with ferric chloride/lime. The degree of
thixotropy is a maximum at the optimum flocculant dosage for dewatering. CampbeU
and Crescuolo [1989] developed an on-line measurement of rheological properties
( especially a rheogram) that could he used to control chemical conditioning at belt
presses.
4.9 Conclusions
In order to simulate the filtration-expression process at laboratory scale, four charac
terization tests for dewatering were used: Filtration-Expression cell, Modified
Piltration Test, Capillary Suction Time apparatus, and Compression-Permeability cell.
These tests provide information on the dewatering rate (in terros of average specific
cake resistance, permeability, vacuum suction time, and CST value) and final dry
solids content. The following sludge microproperties wbich were considered to he of
Sludge dewateriug characteristics 101
relevanee for a better understanding of the sludge dewatering process were measured:
composition (pH, electrical conductivity, and ATP content), partiele size distribution,
and rheological properties (floc strength, degree of thixotropy). The sludge macro- and
microproperties mentioned were determined as a function of the flocculant dosage.
Three types of flocculant were added to the sewage sludge samples: ferric chlor
ide/lime, the cationic polyelectrolyte KF975, and the cationic polyelectrolyte used at
the sludge treatment plant. It appeared that the flocculant dosage has a large impact on
the dewaterability of sewage sludges. The type of sludge and thus the design of the
plant and the operation of the waste water treatment have a minor influence on the
dewatering behaviour of sludges. At the optimum flocculation conditions (flocculant
dosage and mixing conditions) some characterization parameters show a minimum or a
maximum:
• mimimum specific cake resistance,
• minimum vacuum suction time,
• minimum CST value,
• minimum MFfid,
• minimum iron content in filtrate,
• maximum permeability,
• maximum median floc diameter,
• maximum degree of thixotropy.
At the optimum flocculation conditions, sewage sludges can be dewatered at the
highest rate and also the highest dry solids content is reached. A large median floc
diameter is accompanied by a small dewatering rate. Large-size particles!flocs
decrease the specific surface area with a rednetion in shear stresses during dewatering.
Larger aggregates are weaker in nature (high degree of thixotropy). This is due to the
level organization of the floc aggregates: primary particles, dense aggregates
(flocculi), flocs and floc aggregates. The minimum iron content in the filtrate at the
optimum dosage indicates a maximum occupation of ferric hydrolysis products which
specifically actsorb at the sludge solid particles.
Typical differences in dewatering behaviour of sludges flocculated at optimum
conditions between ferric chloride/lime and polyelectrolytes are:
• Both the average specific cake resistance and vacuum suction time of sludges
conditioned with ferric chloride/lime are five to ten times larger than those of sludges
flocculated with polyelectrolyte. Typical values for the specific cake resistance are in
102 Chapter 4
the range of 1013 to 1014 m/kg for s1udges flocculated with iron chloride/lime and in
the range of 1011 to 1013 mlkg for sludges flocculated with polyelectrolyte.
• The compressibility coefficient of sludges flocculated with polyelectrolyte (o=2.3)
is higher than that of sludges conditioned with ferric chloride/lime ( o = 1. 6).
• The median floc diameter of sludges flocculated with polyelectrolyte is 5 to 20
times higher than that of s1udges flocculated with ferric chloride/lime. Typical median
floc diameters are 8 to 100 J.tiD for flocs in s1udges flocculated with iron chloride, and
500 to 2000 J.tiD for sludges flocculated with polyelectrolyte.
The type (trademark) of polyelectrolyte may have a considerable influence on the
dewatering results.
Some of the characterization tests are considered to be useful tools in sludge treatment
plants. On the basis of the characterization research carried out, the following ranking
of dewatering tests recommended for application at sludge treatment plants can be
presented:
1. Filtration-Expression cell
2. Modified Piltration Test
3. Spectrophotometer to determine iron content in filtrate
4. Image analysing technique
5. Modified CST apparatus
6. Rheometer
7. Compression-Permeability cell
By making choices for these tests one should also consider:
• The significanee and usefulness of the information obtained for the mechanical
dewatering process. Tests may be used for diagnostic properties, for adjustment of the
flocculant dosage and for regulation of pressure control inchamber filter presses.
• The rapidity and simplicity of the method.
• Investment and maintenance costs.
The determination of dry solids content and ash content are useful characterization
tests, but are not incorporated in the order of tests. The pH, electrical conductivity,
and A TP content do not give useful information on the mechanical dewatering of
sludges.
5 MODIFIED CAPILLARY SUCTION TIME (CS1) APPARATUS
5.1 Introduetion
In section 4.4.3 the conventional CST apparatus is described. The conventional CST
apparatus bas, however, some disadvantages. Firstly, filter paper often differs in
structure, resulting in a different permeability and suction pressure. These differences in
paper negati vely influence the reproducibility of the measurement. Secondly, the validity
of a theoretical model descrihing the filtration process for a CST apparatus is difficult to
check when the position of the liquid front is only known at two times. Thirdly, the
CST value is dependent on the slurry concentration. CST valnes reported without the
slurry concentration could therefore be misinterpreted. The slurry concentration
influences the thickness of the formed cake and thus its resistance to flow. Since the
proposal of Baskerville and Gale [1968], several investigators have discussed CST
problems [Leu, 1981; Unno et al., 1981; Vesilind, 1988; Vesilind et al., 1988;
Dohänyos et al., 1988; Tilier et al., 1990; Lee and Hsu, 1992, 1993]. A theoretica!
model descrihing liquid flow in a CST apparatus, and some experimental results were
presentedinall of these papers.
The purpose of the research presented in this chapter can be subdivided into three parts:
• to deduce a theoretica! model that describes the position of the liquid front in a CST
apparatus. With the model an average specific cake resistance can be calculated from
experimental data;
• to develop a continuons CST apparatus with which reliable reproducible data can be
obtained;
• to carry out experiments to verify the model calculations and to be able to calculate
the average specific cake resistance of flocculated and unflocculated sewage sludges.
The advantage of a speci:fic cake resistance is that it is an intrinsic value, while the
conventional CST valnes changes, e.g. with the slurry concentration.
104 Cbapter 5
5.2 A theoretical model descrihing the liquid flow in a CST apparatus
The dewatering process in a CST apparatus is in fact a tiltration process in which the
capillary pressure of the filter paper is the main deiving force. The dewatering in a CST
apparatus consists of two consecutive processes, namely the tiltration of sludge in the
cylindrical tube, and the peneteation of filtrate into the filter medium. It is assumed that
the structnre of the filter medium is isotropie so that the shape of the liquid front will be
circular.
The model is based on four equations. Two equations describe the pressure difference
across the sludge layer and two describe the pressure difference across the filter
medium.
1. According to Darcy's law the following equation is obtained for the pressure
difference across the sludge cake.
1 dV f!U."C dV2
L\P. =A K"" dt = 2A 2 dt I
(5.1)
where L\P is the pressure difference across the sludge cake, A the area of the cross-s
section of the inner CST tube, 11 the viscosity of the filtrate, L the thickness of the
sludge cake, K the average permeability of the sludge cake, V the filtrate volume, C av
the cake mass deposited per unit flltrate volume, and a the average speci:fic cake av
resistance. The initial condition is: t=O; V =0.
2. The pressure difference across the sludge cake is the sum of the head exerted by the
sludge layer and the suction pressure exerted by the filter medium.
F' L\P = p gH+-• • A (5.2)
where p is the sludge density, g the gravitational acceleration, H the height of the s
sludge layer, and F' the suction force exerted by the filter medium under the CST tube.
3. As in the filter medium liquid flows in radial direction, Darcy's law must be
expressed in the following manner:
dP dr
dV r=r0 P=P0 (5.3)
Modilied capillary soction time (CST) apparatus 105
where P is the hydraolie pressore, P the hydraolie pressure at position r , r the position 0 0
of the liquid front at time t, r the position of the liquid front at time t=O (i.e. the 0
internal radius of CST tube), h the thickness of the filter medium, and ~ the
permeability of the filter medium.
Solving this differential equation gives:
p -P= .1.P = _11-.-1 ( !:._)dV ° F 2nhKF UCr0 dt
(5.4)
In this equation .1P F expresses the pressure difference across the filter medium.
4. The last ofthe foor equations is analogons to equation (5.2).
2ruh.1PF = 2nrhf3y cose- F' (5.5)
where y is the interfacial tension of the ftltrate, and f3 the reciprocal hydraolie radius.
The product j3ycose is equal to the capillary suction pressure P . The reciprocal cap
hydraolie radius f3 is introduced since the driving force for the flow of liquid in a
capillary medium depends on the measure of wetring of the surface by the liquid.
Combination of equations (5.1), (5.2), (5.4) and (5.5) , and where the area of the cross
section of the CST tube is introducedas nr02
, leads to:
21tlbf3y cose (5.6)
The liquid flow in the capillary medium is treated as a displacement process. The liquid
volume in the capillary medium is equal to the porosity times the wetted ceramic
volume. This is equal to:
(5.7)
Equation (5.6) can now be written as:
106 Chapter5
r~~}eh d(rz roz) p,gHro 2 = -'--=-.,---- --'------'-- + --'---...:....
KF dt 2mi3Ycose (5.8)
with the initial condition: t=O: r=r0.
If equation (5.8) is rewritten, the basic equation, describing the movement of liquid in a
CST apparatus, is obtained:
2r2
In(r0 }1,eh Îjdr KF dt
(5.9)
with the initial condition: t=O : r=r0
•
Equation (5.9) can only be solved numerically. A result of this numerical solution is
presented in tigure 5. 1.
The constauts in this numerical solution were chosen to be physically reasonable for
sludge and ceramics used in the modified CST apparatus (see section 5.5), i.e. p = 1030 ·3 -3 13 ·1 s
kg.m ,TJ=lO Pas,s=0.46,H=0.06m,a =10 m.kg ,P =50kPa,C=18 -3 ·15 2 av cap
kg.m , h=0.0017 m, ~=9.6•10 m, r0
=0.006 m.
The models presented by Unno et al. [1983] and Leu [1981] arebasedon the samebasic
equations. However, Unno et al. failed to give the right solution of the model, and Leu
never solved his model showing the position of the liquid front as a function of time.
0.024
:g 0.018 r:J'l ;:3
~ 0.006 -
0·000 ol----....,I;-;!oo;-;;:----:;;2±oo~---:;;3:\;-oo~--7.4oo!;-;;-----,s~oo
Time (s)
Figure 5.1 Example of a numerical solution of the CST model. Wetted radius as a junction of
time.
Modified capillary suction time (CST) apparatus 107
5.3 Parameter studies
The influence of the capillary pressure P , the filter medium porosity e, and the cap
average specific cake resistance a on the model calculations were investigated with the av
presented CST model. The constants used in these parameter studies were given in the
previous section.
The influence of the capillary pressure. The capillary pressure is a driving force for
liquid flow in the capillary medium. Therefore one should expect an increase of the
liquid front velocity with an increasing capillary pressure. This is shown in figure 5.2.
The capillary pressure was varled from 10 to 100 kPa.
The injluence of the porosity of the filter medium. The consequences of a variation of the
porosity of the filter medium E for the calculated radial position of the liquid front as a
0.030
k:Pa 70k:Pa 30k:Pa lOkPa
0.024
,-..., s 0.018 '-" rl:l ;:I ·-'"0 ro 0.012 ~
0.006
0.000 0 300 600 900 1200 1500
Time (s) Figore 5.2 The influence of the capillary pressure. Calculated wetted radius as a tunetion of
time.
function of time are not evident, because the capillary pressure P and the filter cap
medium permeability ~ both depend on the porosity E. The capillary pressure depends
on the redprocal hydraulic radius [3, the interfacial tension of the filtrate y, and the
contact angle e capillary medium/filtrate, according to the following formula:
108 Chapter 5
P •• p = l3y cose (5.10)
The redprocal hydraulic radius 13 is given by:
(5.11)
where a represents the specitic surface area of the ceramics. Substitution of equation V
(5 .11) into equation (5 .1 0) yields the dependenee of the capillary suction pressure on the
porosity:
a (1-e) P = v ycose
oap E (5.12)
The permeability can be calcnlated with the Kozeny-Carman equation. This equation has
been experimentally proved valid for the capillary medium used in the modified CST
apparatus [Uze~, 1992]. Results ofthe calculations are shown in tigure 5.3.
The porosity was varled from 0.3 to 0.8. The liquid front velocity is a maximumfora
capillary medium porosity of 0.5, when it is taken into account that the variation of the
0.030 r---------------------
0.8
0.006
0.000 L__ ___ ..__ ___ ..__ ___ _J__ ___ _J__ __ _
0 200 400 600 800 1000
Time (s)
Figure 5.3 The influence of the porosity & of the filter medium. Calculated wetted radius as a
junction of time.
Modilied capillary suction time (CST) apparatus 109
porosity bas consequences for the valnes of the capillary pressure and the filter medium
permeability. In the modified CST apparatus, ceramics is used as a capillary medium
(see section 5.4). The ceramic slab is characterized by the structure parameters of
porosity, permeability, and capillary suction pressure. The porosity can be determined
with mercury porosimetry, permeability with a permeability cell, and the capillary
suction pressure by conductinga CST experiment with demiwater (see section 5.5).
The irifluence of the average specific cake reststance a . The cake resistance a av av
represents the liquid flow resistance in the CST rube, since there has been a build-up of
a sludge cake. Therefore one should expect a decrease of the distance covered by the
liquid front at a certain time with an increase of the average specîfic cake resistance
a . Results of calculations shown in figure 5.4 conflfiD this expectation. av
The specific cake resistance a was varled from 1012 m.kg-1 to 1013 m.kg-1
. The av
calculated radial position of the liquid front at a time t decreases with an increasing cake
resistance aav· The value of~ used in this parameter study is 9.6•10-15
m2
• As can be
seen in figure 5.4, the differences between the curves are relatively small. To increase
this difference a filter medium with a higher permeability should be used. The
consequence of a smaller permeability of the filter medium is that the influence of the
specific cake resistance on the radial liquid front position as a function of time is
negligible.
0.020
~ s 0.015 . ..._, rn ;:I
:.0 ~ 0.010
0.005
0.000 0 60 120 180
Time (s)
12 --uo -----3.1012
12 : -----7.10 .
1.1013
240 300
Figure 5.4 The influence of the specifïc cake resistance aav . Calculated wetted radius versus
time.
110 Chapter 5
From these parameter examinations several conclusions cao be drawn. Firstly, in order
to calculate ao average specific cake resistance, the permeability of the sludge cake must
be small in comparison with the permeability of the filter medium. Secondly, it is
obvious that the capillary pressure is ao important driving force for the liquid flow in a
capillary medium. The hydrostatic driving force induced by the sludge head is negligibly 5
small (± 600 Pa) in comparison with the capillary driving pressure (± 10 Pa; see section
5.5). Finally, it cao be concluded that the porosity of the capillary medium has great
influence on the results of the model calculations aod experiments, because the porosity
influences the redprocal hydraulic radius 13 aod therefore the capillary pressure P , cap
aod at the same time the permeability of the filter medium Kp.
5.4 Modified CST apparatus
Several apparatuses to measure the CST are described in literature. The first apparatus,
developed by Baskerville aod Gale [1968] has been commercialized by Triton
Electronics. Alternative CST apparatuses are known with which e.g. CST must be
determined visually by a radial scale drawnon the capillary medium [Leu, 1981], or by
use of several electrodes [Lee aod Hsu, 1992, Unno et al. ,1983]. These instruments
have two important properties in common. Firstly, filter paper is used as a capillary
medium. Secondly, the position of the liquid front is only measured discontinuously.
Both properties imply several problems.
Filter paper is ao anisotropic material, which implies that the liquid front will move with
different veloeities in different directions. As a result no circular liquid front, on which
the model is based, will develop. Further, the reproducibility of the measurement will be
small, because of the differences between the filter papers.
To check the presented CST model on the physical reality more than two datapoints are
required. A continuous method of measurement is the best option.
Two typical differences between the conventional aod the modified CST apparatus cao
be distinguished.
Filter medium. Instead of filter paper, ceramics will be used as a capillary medium
(pressed Al20 3, initial partiele size ± 50 Jtm, pressing force 300 MPa, sintering
temperature 1500 °C). Since this capillary medium is isotropic, a circular liquid front
will develop during a CST experiment. One cao also expect the experiments to be
reproducible. The required thickness of the ceramic slab is 1-2 mm; this thickness
avoids slow saturation of the ceramic slab in the axial direction.
Modified capillary suction time (CST) apparatus 111
metal plate
Figure 5.5 Schematic diagram of the modified CST apparatus.
Method of measurement. The new continuous metbod of measurement is based on the
fact that the electrical resistance of the cernmies will decrease when the pores are fitled
up with water or any other liquid. The modified CST apparatus is schematically shown
in tigure 5.5. The ceramic slab is tightened between two copper plates. The copper
plates act as electrodes. A multimeter (measuring frequency 10 kHz), connected to the
electrodes, continuously measures the electrical resistance of the ceramic slice. The
experimental data are transferred to a computer. The sampling time of the computer is 1
second. When liquid is sucked into the ceramic slab, a decrease of electrical resistance
of the capillary medium will be measured. At the end of the experiment a constant
resistance has been reached, which depends on the type of ceramics used and the ionic
strength of the liquid. From the measured resistance as a function of time the radial
position of the liquid front as a function of time can be calculated. The situation at a
certain momentduringa CST experiment is also shown in tigure 5.5. The shaded area
represents the part of the ceramic slice that has been ftlled up with water and the white
area represents the dry ceramics. r0 equals the position of the liquid front at time t=O, r
the position at timet, and re represents the radius of the ceramic slab. This situation at a
time t can be interpreted as a parallel conneetion of the electrical resistances of the dry
part of the slice R d and the wetted part of the ceramic slice R , respectively. The ~ e~
total resistance of this system R can therefore be represented as: e,tot
R R R = e,d e,w e,tot R +R
e,VY' e,d
(5.13)
112 ChapterS
The position of the liquid front at an arbitrary time t can now be calculated from the
measured resistance R according to the following equation: e,tot
(5.14)
where p d and p w represent the specific electrical resistance of the dry and wetted
ceramics, respectively. It is essential that the whole experimental set-up is tightened in a
'clamping-screw', see tigure 5.6. This set-up assures good contact both between the
wetted ceramics and the copper plates and between the tube and the ceramics (no
leakage). To avoid pollution of the ceramics with small sludge particles, a small filter
paper (0 6 mm) with negligible flow resistance is placed between the tube and the
ceramic slab.
Figure 5.6 Schematic diagram for the whole experimental set-up for the modified CST
apparatus. 1) rubber, 2) copper electrode, 3) ceramic slab, 4) CST tube.
Modified capillary suction time (CST) apparatus 113
5.5 Experimental results and discussion
Detennination of capillary suction pressure. The capillary suction pressure P, of the -15 2 cap
circular ceramic slab (re =29 mm, h=l mm, E=0.46, ~=9.6•10 m) used in the
experimentsis an important system parameter. The capillary suction pressure is assumed
to be uniform in all directions of the ceramic slab and to be constant in time. 1n an
experiment a colonred water salution is poured into the CST tube and the position of the
wet front radius is measured by a ruler. 1n this way, the position of the wet front is
recorded directly. If only pure liquid is used in the CST experiment (i.e. there is no
sludge cake formation), the first two terms on the right-hand side of equation (5.9) can
b~ deleted. The mathernatical model is then used to obtain the capillary pressure. The
result of this experiment is shown in tigure 5. 7. A model calculation yields a value for 5 -2
the capillary suction pressure of 1.08•10 N.m . Another way todetermine the capillary
pressure is to carry out CST experiments with demiwater.
0.1130 ..
0.1128 :.
0.026
(1.024 .
11.022
c;;:? 0.020 .
6 0.018 .
Cf.) 0.018 ..
E3 o.m~ •.
~ ::: 0Jl06
O.OM
0.002 :·
* *
* * . '•·.
* . .. ·•
*
ODOO~~~~~~~~~~~~~~~~~~~~~~~
0 lOO 200
TIME (S)
Figure 5. 7 Wetted radius versus time for a CST experiment with water to determine the
capillary suction pressure; the position of the liquid front has been measured visually.
114 Chapter 5
O.OY
0.01.3
0.012
O.Ol1
0.010
~0.000 o.ooa
ril ~O.D07
~o.ooa * O.oo&
0.1103
D.002
0.001
0.000
0 10 20 30 40 50
TIME (S)
Figure 5.8 Wetted radius as a function of time for a CST experiment to determine the capillary
pressure; position of the liquid front determined indirectly by use of the electrical resistance of
ceramics.
The radial position of the liquid front as a function of time is determined indirectly by
measuring the electrical resistance of the ceramic slab as a function of time according to 11
equation (5.14). The specitic electrical resistance ofthe dry ceramics pd equals 10
nm. The specitic resistance of the wetted ceramics is calculated from the measured
resistance at the end of the experiment, which depends on the electrical conductivity of
the filtrate and the geometry of the slab. It is assumed that the ceramic slab is fully
saturated. The experimental result is fitted with the theoretica! model to calculate the 5
capillary suction pressure (fig. 5.8). The result of the calculation is Pcap=0.884•10 -2
N .m . lt can be concluded that both experiments showed satisfactory agreement. 5 -2
Henceforth in model calculations a value of the capillary suction pressure of 10 N.m
is assumed.
Reproducibility. In tigure 5. 9 three CST experiments carried out under the same
process conditions with unflocculated sludge are shown. In these experiments the same
ceramic slab was used. It can be seen that the reproducibility is acceptable. Model
Modified capillary suction time (CST) apparatus 115
15 -1 15 calculations yield values of tbe specific cake resistance of 4.39•10 m.kg , 3.64•10
-1 15 -1 m.kg , and 3.01•10 m.kg . One should notice tbat sewage sludges are biologica! in
O.o10
0.009
0.008
0.007
(/) ::J 0.005
0 <( 0.004
Cl:
0 100 200 300 400 500
TIME (S)
Figure 5.9 Results of three identical experiments carried out with unflocculated Eindhoven
sludge.
nature. Due to microbial activity tbe sludge composîtion may change witb time and
negatively înfluence tbe reproducibility. However, CST tests witb non-biological sludges
have not been carried out.
Determination of the specific cake resistance. CST experiments have been carried out
wîtb Eindhoven sludge. In these experiments a cylindrical perspex column witb an imler
radius of 3 mm was used. A larger column radius will increase tbe filtration area and
tberefore also tbe liquid flow through tbe cake and tbe capillary medium. Consequently,
differences in specific cake resistance will be hard to distinguish. At tbe start of an
experiment 3.5 ml of a sewage sludge sample (dry solids content 1.9 wt%) was poured
116 s
into the tube. The experiment ended when the ceramic slab was fully saturated. In tigure
5.10 two experimental results are shown.
0.020 re--:---:--:--:--:----,-;:--:--:--:--:--:--:--:-:--:--:--:-:--:--:-~
0.019 .
0.018 .· 0.017 ..
0,016 .
0.015 .
0.014 :
......... o.ou . ~ 0.012 : ......... . (/) 0.011 :
:::::> 0.010
Cl 0.009 <( 0.008
0::: 0.007
0.003
0.002
0.001
0.000 'r:-.'-r..,....,...,.;..,;..~,.;..":--rr;.,.;..,;...;....;...;,...",;-...:.,..:..r..,...:....,.;..,;.....,.;,.....;...--rr..,...:....,.;..,;..r-'~--rr..,...:....,-,' 0 100 200 300 500
TIME (S)
Figure 5.10 Experimental results of CST experiments, wetred radius as ajunetion qftime. Dots:
uriflocculated sludge; Stars: sludge flocculated with 100 g FeC/3/kg ds and 200 g Ca(OH)2/kg
ds. lines: results of the model calculations.
The stars represent a result of an experiment carried out with a sample flocculated with
100 g FeC~Ikg ds and 200 g Ca(OH)21kg ds. The dots represent the result of an
experiment carried out with uuflocculated sludge. It is clear that flocculation of a sludge
sample accomplishes a decrease of the time needed for the liquid front to move a certain
distance. The lines in the graph represent the results of the model calculations. A
specific cake resistance is determined from fitting the experimental results with the
theoretical model. Results of model calculations are: 15 -1
aav, uuflocculated = 1.24•10 m.kg 13 -1
<Xav, flocculated = 2.56•10 m.kg
As expected, the specific cake resistance of unflocculated sludge is much higher than the
cake resistance of a flocculated sample. At the beginning of the experiment, the model
overestimates the velocity with respect to the experiment. This is possibly caused by the
resistance of the small filter paper positioned under the perspex column. The modified
Modilied capillary suction time (CST) apparatus 117
CST device provides the possibility to determine a specitic cake resistance of flocculated
as wellas unflocculated sludges. Unflocculated sludges are hard-to-filter suspensions in
a batch filtration experiment.
0.012
0.009
0.006
0.003
1000 2000 3000
TIME (S)
Fig. 5.11 Wetted radius as a function of time in two CST experiments carried out with
unjlocculated sludge with two different slurry concentrations.
Effect of the slurry concentration. In figure 5 .11 the movement of the liquid front as a
function of time for two different slurry concentrations is shown. The movement of the
liquid front is slower when the slurry concentration increases. This is caused by the
formation of a thicker cake. The calculated average specific cake resistance was in both -15 -1
cases 2.7•10 m.kg . If only conventional CST values were reponed, a large
difference was found. However, since for the calculation of the specific cake resistance
the slurry concentration is needed, the cake resistance is equal for both slurry
concentrations. The intrinsic dewatering property bas not been changed. It may be
concluded that effects of slurry concentrations can also be investigated with the modified
CST apparatus.
118 Chapter5
To corre1ate the calculated specitic cake resistance from a CST experiment with normal
batch filtration cake resistances, it is useful to carry out filtration experimentsnot only at
100 kPa (capil1ary suction pressure of ceramics used) but also at other pressures. As
· sludge is very compressible, the average specific cake resistance depends on the
pressure drop over the sludge cake. In tigure 5.12 results of normal filtration and CST
experiments are shown. From tigure 5.12 it can be conclnded that there is an acceptable
corre1ation between normal filtration and CST experiments. The difference may be
caused by the varianee of the determined capil1ary pressure, possible sedimentation
effects, or wall friction effects which have more influence on a CST experiment than on
normal filtration experiments, because of the smaller dimensions of the CST tube.
However, only a few of these experiments have been carried out. To get a good idea
about the corre1ation between normal filtration and CST experiments more experiments
are needed.
70
- 60 'b Jol: • .s • 8 50 c • • flitration at --_CII
40 ..... el!!. 0 CST B • 8 30 Q
0 ;: "6 ID 20 Cl. liD •
10 0 100 200 300 400 500 600
(Thousands) pressure [Pa]
Fig. 5.12 Average specijic cake resistance versus pressW'e for normal flitration experiments
(filled dots) and a CST experiment (open dot).
Modified capillary suction time (CST) appara~_u_s _______ 1_19_
5.6 Conclusions
Witb tbe model presented and tbe modified type of CST apparatus average specific cake
resistances can be obtained for unflocculated as well as flocculated sludges. The
modified CST instrument is based on continuously measuring tbe electrical resistance of
a ceramic slab during filtrate penetration. The measured electrical resistance as a
function of time is converted to a radial position of tbe liquid front as a function of time.
Before any calculations can be made, several parameters of tbe ceramic porous medium
have to be k.nown. This implies tbe permeability, the capillary pressure, and tbe specific
electrical resistance of tbe ceramics. It is of great importance that tbe permeability of tbe
ceramics is high enough (10-14
to 10-15
m2) in comparison witb tbe permeability of tbe
-15 -17 2 s1udge cake (10 to 10 m ) in order to calculate a correct average specific resistance.
The required thick.ness of tbe ceramic slab must be equal to 1 to 2 mm to avoid slow
saturation in tbe axial direction. The slurry concentration does not influence tbe average
specific resistance of tbe cake formed in tbe CST tube because of tbe low
compressibility of tbe sludge at low applied pressures. However, tbe conventional CST
value depends on tbe initial solids concentration of the sludge suspension.
6 FLOCCULATION BEHA VIOUR OF SEWAGE SLUDGE
6.1 Introduetion
Material in waste water may comprise suspended and/or dissolved organic and
inorganic matter and numerous biologica! forms such as bacteria, algae and viruses.
Particles in the dimeosion range of 10·9 to 10·6 mare referred to as colloids.
Particles of colloidal or lesser dimensions are able to retain a dispersed state because
of eertam inherent characteristics which promote their stability. The term stability
describes the ability of individual particles to remain separate entities. The stability of
colloidal material arises from the predominanee of forces associated with the solid~
liquid interface.
If the size of particles becomes progressively smaller for a given total partiele mass,
the total surface area becomes extremely large. Colloidal material possesses a colossal
surface area to mass ratio. lt is apparent that for a given total mass, the smaller the
particles the more predominant the influence of phenomena associated with interfaces
becomes, and the lesser the influence of gravity effects associated with mass becomes.
Partienlate material of colloidal size may be removed from dispersions by methods
other than those relying on gravity effects. Finely dispersed colloidal material bas to
be converted into a form whereby separation from the dispersion is practicable.
Conversion of the stabie state of a given dispersion to an unstable state is termed
destabilization. Conversion processes could either alter the surface properties of
partienlate material, thereby increasing the adsorptivity of the particles to a given filter
medium, or generating a tendency for aggregation of small particles into larger units
or precipitate dissolved material, thereby creating partienlate material for which
separation by sedimentation and/or filtration is feasible.
A process which is able to accomplish destabilization of sludge particles is
flocculation. Flocculation is the process whereby small particles dump together and
thereby form larger agglomerates or flocs as a result of destabilization, and is a
generic term nsed in the waste water treatment indnstry. Chemical coagulants are
added to the dispersion to induce flocculation. In literature, a number of defmitions of
flocculation bas been presented. According to the La Mer system [1964], which is the
generally accepted standard, aggregation is used as the general term. The term
coagulation is reserved for those processes in which the London~van der Waals force
(see section 6.3) is the primary driving force causing aggregation.
122 Chapter 6
Destabilization reactions of colloids in an aqneons dispersion by chemical coagu]ants
are complex and arise from several mechanisms. Doderstanding of varions
destabilization phenomena involved in sludge tlocculation is essenrial to get insight
into the dewatering behaviour. In order to stndy sludge destabilization phenomena the
so-called electroacoustic technique (MA TEC) is used. With this technique the ESA
signal of a given sludge suspension is determined. The ESA signal is related to the
zeta potential. Zeta potentlal measurements are useful in stndying the colloidal stability
of dispersions, ion adsorption studies and characterization of partiele surfaces.
Before introducing this technique and carrying out experiments with it, the fust steps
are to consider the surface charge carried by sludge particles, the electrical double
layer around the charged surface, the effect of specific adsorption on both the effective
surface charge and its zone of intluence and destabilization phenomena of colloids
induced by metal coagulants and polyelectrolytes.
6.2 Tbe electrical double layer around a spberical sludge partiele
The charge of a partiele is predominantly produced by ionization of functional groups,
which are attached to the particle. Sewage sludge particles mainly consist of organic
matter. Sludge partiele surfaces are polysaccharide in nature, composed of neutral
sugars and glucuronic acid (monosaccharide CJI100 7, a sugar acid). Glucuronic acid is
one of the main ionogenic compounds at the sludge surface due to the presence of
ionized carboxyl groups [Steiner et al., 1976]. Carboxyl groups cao act as a proton
donor and proton acceptor group. looization of a carboxyl group is represented by:
As a consequence sludge particles are negatively charged. loos which arise from
dissociation of surface molecules and in this way causing the partiele surface charge
are called potential-determining ions. H+ and (OH)" ions are potential-determining ions
of sludge particles.
The partiele surface charge intluences the distribution of nearby ions in the liquid.
Due to electrostatle forces, ions of opposite charge (counter-ions) are attracted towards
the surface and ions of like charge (co-ions) are repelled away from the surface. This,
together with the random thermal motion and mutnal ionic repulsion or attraction,
leads to the formation of an electrical double layer. The electrical double layer
Flocculation behaviour of sewage sludge 123
consists of the charged surface and a neutralizing excess of counter-ions over co-ions
distributed in a diffuse manner in the nearby liquid.
The electrical double layer can be viewed schemarically made up of three regions:
1. A surface layer, having a electrical surface potenrial if;0 and a surface charge a0 •
2. A Stern layer with thickness ö.
3. A diffusedregion of ions having a charge ad.
In figure 6.1 a schematic illustrarion is given of the electrical double layer and the
variarion of the electrical potenrial with distance from a negarively charged wall in the
absence of specific adsorprion.
·~· ••• • .I •• . ' .. , .. stern diffuse ioyer tayer
• •
<llotance
Fig. 6.1 Electrical potenrial (absolute) as a tunetion of the distance from a negatively charged
wall. Due to electrastatic and van der Waals forces counter-ion adsorption predominafes over
co-ion adsorption. No specific adsorption [Shaw, 1969].
Ions cannot be assumed as point charges, but possess a finite size. This prevents the
eentres of the counter-ions approaching the surface closer than within a distance ö.
124 Chapter 6
Counter-ions are not easily dehydrated so that they retain their hydration shells in the
adsorbed state. The so-called Stern plane is located at about a hydrated ion radius
(distance o) from the partiele surface. The region between the eentres of the surface
charge and the distance o (Stern layer) is charge-free. The potenrial decay over the
Stem layer is linear. This is a consequence of Poisson's law (equation 6.4). The
electrical potenrial at the Stem plane is called the Stem potenrial 1/;0• Due to overall
electrical neutrality of the whole double layer, the charge of the diffuse double layer
ud is opposite to the charge of the partiele u0:
(6.1)
Ions or molecules can be attracted to the solid surfaces not only by repulsive forces
but also by van der Waals forces. Another mechanism is specific adsorption. All
interactions at the particle-solution interface which are not coulombic are designated as
specific. They may be of chemical or physical nature, such as dipolar, hydrogen,
entropie or covalent. Specific adsorption of co-ions or counter-ions at the surface
infl.uences the sign and magnitude of the Stern potenrial, in this way affecting
flocculation. The plane through the eentres of the specifically adsorbed ions inside the
Stem layer is called the 'inner Hehnholtz plane (iHp)'. The plane parallel to the
surface at distance ó is called the 'outer Hehnholtz plane (oHp)'.
Specific adsorption of counter-ions may result in a reversal of charge to take place
within the Stem layer, i.e. 1/10 and 1/lo have opposite signs, and is called superequivalent
adsorption (:figure 6.2). Specific adsorption of co-ions could create a situation in which
1/;6 bas the same sign as 1/;0 and is largerinmagnitude (see tigure 6.3). If the charge in
the Stern layer is called u., then because of electroneutrality in the electrical double
layer:
(6.2)
The diffuse part of the double layer plays an important role in colloidal stability of
dispersions. When two particles approach one other, the electrical double layers will
overlap and repulsive forces become operative. The electrical potenrial in the diffuse
layer decays exponentially from 1/;8 at the Stem plane to zero at infinity (see section
6.3).
Flocculation behaviour of sewage sludge 125
distallee
l
Fig. 6.2 Reversal of charge may take place within the Stem layer due to specific adsorption of
counter-ions. Superequivalent adsorption of counter-ions.
Fig. 6.3 Specific adsorption of co-lons could increase the ejfective surface charge:
I u, I > I 11o I .
The Stem potenrial cannot be measured. A close approximation may be obtained by
measuring the potential at the plane of shear between a moving partiele and the
surrounding liquid. This electrical potential is known as the zeta m potential. The
126 Chapter 6
exact location of the shear plane is not known since it depends on the adsorbed ions in
the Stern layer and the degree of hydration. It is assumed that the shear plane is
located a small distance further away from the surface than the 'outer Helmholtz
plane' and that the zeta potentlal is, in general, marginally smaller in magnitude than
Vtö· The electroneutrality condition is then presented by:
(6.3)
where <rb is the charge between the surface and slipping plane and <rek the charge at the
slipping plane.
The description of the diffuse part of the double layer proposed by Gouy and Chap
man is based on an assumption of point charges in the electrolyte solution.
The electrical potential 1/t and the density p of the space charge in the solution
surrounding one single rigid sphere are related by Poisson's equation:
(6.4)
where Eo is the permittivity of vacuum and €r the dielectric constant of the medium.
The space charge is considered to be built up by un unequal distribution of pointlike
positive and negative ions:
(6.5)
where n+ and n_ are the local concentrations (number/m3) of positive and negative
ions, respectively, z the valency and e the elementary charge. Only one type of
positive and one type of negative ions are considered.
The statistical equilibrium of co-ions and counter-ions in the diffuse double layer sur
rounding a charged partiele is described by Boltzmann's law:
z e'ljt(r) n_(r) = n-Q exp(~)
(6.6)
(6.7)
Flocculation behaviour of sewage sludge 127
where n+o and n_0 are the concentradons of co~ions and counter-ions at inftnity,
respectively.
Electrical neutrality of the bulk solution yields:
(6.8)
Substitution of equations (6.6) and (6.7) in (6.5) and the subsequent substitution of
(6.5) in (6.4) gives:
en+0z+( ( -z+el/t(r)) (z~el/t(r))) --- exp -exp ---
(V>r kT kT (6.9)
The boundary conditions are:
(6.10)
r=a 1/t = r (6.11)
a represents the sum of the sphere radius and the thickness of the Stem layer. An
analytical solution of the differentlal equation (6.9) is only possible when one assumes
that ezl/t < <kT. When the surface potenrial is sufficiently small ( if; 0 < 25 m V), this
inequality is valid. In that case, the Poisson-Boltzmann equation can be linearized by
expanding the exponential terms in the right-hand side and retaining only two terms.
This estimate is called the 'Debye-Hückel approximation'.
The solution of equation (6.9) is then given by:
a if; = t - exp( -K(r-a)) (6.12) r
where K·1 represents the double layer 'thickness' and is expressed by:
(6.13)
The double layer thickness is dependent on the electrolyte concentration. Increasing
the ionic strengthof the solution compresses the double layer.
128 Cbapter 6
6.3 Colloidal stability in terms of the electrical double layer
The Derjaguin-Landau-Verwey-Overbeek (DLVO) theory [Verwey and Overbeek,
1948] bas made it possible to study the stability of colloids quantitatively. The theory
involves estimations of the energy of attraction (London-van der Waals forces) and the
energy of repulsion (overlapping of electrical double layers) in terms of the distance
between particles.
The potenrial energy of interaction between two spherical double layers can be
calculated using Derjaguin's method, which is based on the interaction between two
flat double layers. The magnitude of the potenrial energy of repulsion between two
equal spheres of radius a is given by:
-lo in this equation is represented as:
expzet -1 2 2kT
'Yo = ---,---expze! +1
2kT
(6.14)
(6.15)
H0 is the shortest distance between spheres, a the partiele radius and llo represents the
concentration of ions in the bulk solution.
The London-van der Waals force between two atoms or molecules is a short-range
force and its potenrial energy varles inversely with the sixth power of the inter
molecular distance. Because of the additivity of this force, the London-van der Waals
force between macroscopie bodies becomes a long-range force.
For the interaction of two equal spheres with radius a and a distance R between their
centres, the London-van der Waals attraction potenrial is given by:
(6.16)
where
s R a
Flocculation behaviour of sewage sludge
H 2+-0
a
For H0 < < a equation (6.16) is approximated as
V = A
129
(6.17)
(6.18)
where AH is the Hamaker constant, which takes into account that particles of material
1 are embedded in a medium 2.
The stability of colloidal particles in aqueous media depends on the total potenrial
energy of interaction VT, that is
(6.19)
Figure 6.4 shows the total potenrial energy as a function of the separation distance H0•
Negative values of the potenrial energy correspond to attraction and positive values to
repuls ion.
vf
0
Fig. 6.4 London-van der Waals attraction potenrial VA (negative), potential energy of
repulsion VR (positive), and the sum of both V z; as a junction of the interpartiele distance H0•
The height of the potential harrier depends on the Stern potential 1/;ó and the electrolyte
concentration.
130 Chapter 6
Since V R varles exponentially with the distance and V A varles inversely with the
interpartiele distance, V A surpasses VR at short and long distances, thus producing
attraction between the particles. At intennediate distances, the plot of V1 shows
varlous shapes depending on the Stem potential I/la and electrolyte concentration. For
large values of I/la there is a potential barrier and for small values of I/la there is an
attractive potential. If the potential energy maximum A is large compared with the
thennal kinetic energy (kT), the system will be stabie and the potential barrier (curve
V n) will hinder two particles to stick together into the deep primary minimum,
appearing at a small separation distance Ho. The electrolyte concentration also has an effect on the potential barrier. Increase of the
electrolyte concentration (and thus decrease of the double layer thickness K-1) canses
the disappearance of the potenrial barrier. V 1 shows that there is attraction between the
particles and destabilization of the suspension takes place (curve V12). The potential
energy curve V 12 can be taken to require an expression for the critical coagulation
concentration. The critical coagulation concentration is defined as the electrolyte
concentration at which coagulation commences. From curve V 12 there is a certain
value of H0 at which
(6.20)
and
(6.21)
The equations mean that the maximum in the potenrial energy curve touches the
horizontal axis. Solution of equations (6.20} and (6.21) yields the critical coagulation
concentration n" (number of ions/m3):
(6.22)
The critical coagulation concentration C0 expressed in units of moles per liter is
related ton" according to:
Flocculation behaviour of sewage sJudge 131
(6.23)
where NA represems Avogadro's number.
The critical coagulation concentration is inversely dependent to z6• It states that the
higher the counter-ion valency, the lower the critica! coagulation concentration.
Theoretica! critica! coagulation concentrations of indifferent electrolytes, yielding
counter-ions with a valency z = 1 ,2,3 should be in the ratio of 729: 11: 1. This phenom
enon is known as the rule of Schulze and Hardy. In practice it is found that a trivalent
ion is 700 to 1 000 times as effective as a monovalent ion in destabilizing a colloid of
opposite charge. The rule of Schulze and Hardy is in reasonable agreement with
experimental evidence on the coagulation of colloids by non-specifically adsorbable
ions. Typical valnes for the critical coagulant concentration are very small. For
example, to flocculate negative As2S3 sols, about 9·10·5 moles.(liter)-1 of trivalent
cation and 7·10·4 moles.(liter) 1 of divalent cation are neerled [Hiemensz, 1986].
lt should be noted that the expression for the concentration of indifferent electrolyte
relies on a model of double layer depression according to the Gouy-Chapman treat
ment. It does not take the specific adsorption of roetal coagulants into account. In the
next chapters we wi11 deal with the destabilization of a dispersion due to specific
adsorption of metal coagulants and polyelectrolytes.
6.4 Specific adsorption flocculation by metal coagulants
Inorganic salts, e.g. ferric chloride, are commonly used as a flocculant in the
treatment of waste waters. Ferric salts, when in solution, immediately dissociate to
form hydrated reaction products. The ferric ions form coordination compounds with
water molecules to give Fe(H20)63+.
The high positive charge on the central metal ion canses some polarization of the 0-H
honds and there is a tendency for protons to dissociate, giving one or more hydrolyzed
species, thus
Fe(H20)6.0(0H)
0
3-n + H20 ;;::± Fe(H20)6-n-t(OH)n+t3·n-t + H30+ (n=O up to and
including 5)
132 Chapter 6
In this way, six mononuclear hydrolysis products are present in the aqueous solution.
The equilibrium reacrions are characterized by a particular equilibrium constant, which
depends on the nature of the metal ion. Small and highly charged ions have a great
tendency to release protons and hence are acidic in a neutral pH medium, e.g. an
unconditioned sludge suspension. As the pH of a sludge suspension is increased, the
equilibria are driven to the right. Figure 6.5 presents a distribution diagram for the
various mononuclear ferric hydrolysis products in an equilibrium state at different pH.
1.0
I f'e(OH);"'
0.6 I
·o l 0.6'
i: .. ~ .. ~ c ·§ 0.4
2
0.2 I i
Fe(OHl, Ft(O~);-
I
I 0
0 2 10 pH
Fig. 6.5 Distribution diagram for iron hydralysis products for iron concentration of 1()5 M [Singley and Sullivan, 1969].
The diagram is valid for a molar iron concentration of 10-s M. The distribution
diagram is basedon calculations carried out by Singley and Suilivan [1969].
Besides mononuclear hydrolysis products also multinuclear hydrolysis products,
collectively given by FexOH/x-y, can be formed. Multinuclear hydrolysis products are
not incorporated in calculating the distribution diagram presented in tigure 6.5.
Mononuclear and multinuclear iron hydrolysis products show enhanced adsorption
characteristics. Specific adsorption of hydrolysis products is possibly another way to
destabilize a sludge suspension. Through adsorption of charged coagulant species of
opposite sign to the partiele surface, the effective charge is reduced and, as a conse
quence, the extent of the double layer repulsive interaction between adjacent particles
Flocculation behaviour of sewage sludge 133
is reduced. When treated with an excess of counter-ions, a charge reversal of particles
may even occur (see tigure 6.2). A secoud mechanism considered, again as aresult of
adsorption of coagulant species at the particle-solution interface, is that described as
the bridging mechanism. During hydrolysis reactions, metal coagulants have a
pronounced tendency to polymerization. On adsorption of such polymerie species to
particles, a coagulant bridge spanning between adjacent particles is formed, thereby
promoting destabilization.
Another mechanism of destabilization by ferric salts is that of precipitate entrapment.
Under appropiate conditions of the coagulant concentration and pH, ferric coagulants
in an aqueous solution form insoluble ferric hydroxide precipitates Fe(OH)3 (see tigure
6.5), initially as a fine colloidal dispersion. These particles then aggregate to form
hydroxide flocs which enmesh the colloidal particles originally present in the water.
This destabilization process is called sweep flocculation [ Gregory, 197 8].
In practice, after an addition of ferric chloride, calcium hydroxide is mixed with the
s1udge suspension. As a consequence, the suspension pH increases to 12. The formed
ferric hydroxide flocs precipitate (sweep flocculation process). Since lime bas a low
solubility it stays undissolved in the sludge cake and can be regarded as an ordinary
filling material. The formed fllter cake will be less compressible and the dewatering
properties of the sludge cakes are improved [Saunders, Holmes, 1987; Janssen et al.,
1994]. Other effects of lime addition are:
• The elimination of offensive odours.
• The removal of sulphide, sulphate, bi-carbonate, and ammonium by precipitation
(calcium sulphide, calcium sulphate, calcium carbonate) and gas escape (ammonia).
• Desinfection: dying of (pathogenie) rnicroorganisms, bacteria, and viruses.
• An increase of the amount of dry solids, which is regarded as a disadvantage. The
amount of lime to condition sewage sludge is high in practice: 10 to 20 kg per m3
sludge.
6.5 Polymerie adsorption flocculation
Polyelectrolytes are widely used as sewage sludge conditioners. The term 'polyelectro
lyte' is often used to describe all polymerie flocculants. Polyelectrolytes contain
functional groups which may or may not carry a charge. If the polyelectrolyte is
charged, the groups may be such as to give an anionic character, a cationic character
or an ampholytic character to the chain, where both anionic and cationic charged sites
134 Chapter 6
are present. The intensity of the charge carried by the polyelectrolyte is dependent on
the degree of ionization of the functional groups and on the degree of
copolymerization. Besides the possibility of functional groups carrying a charge, also
sites are present along the polyelectrolyte cbain which possess the property of being
adsorbed. Destabilization by polyelectrolytes could involve a mechanism combining
both charge effects and effects due to adsorption.
The extent of polymerization of the polyelectrolyte is characterized by the molecular
weight. High molecular weights signi:fy long chains, whereas low molecular weights
indicate short polyelectrolyte chains.
Especially in the fields of water and waste water treatment, it appears that the only
effective polymerie flocculants are those of opposite charge to the negatively charged
particles: cationic polyelectrolytes.
The basic elements (monomers) of cationic polyelectrolytes are derivatives of poly
(meth)acrylamides. Dissociation of the derivatives in water yields positively charged
polymerie ions. Monomer molecule structures of these derivatives are presented as
[Röhm, 1991]:
The monomeric ions are characterized by the presence of tertiary (RNH+) or
quarternary (RN+) ammonium groups. The polymerization is started by the reaction of
the monomer with a radical.
For polymers which are neutral or similarly charged to the partiele surfaces, one or
more of the following mechanisms of interaction can lead to adsorption: a) hydrogen
bonding, b) dipolar and van der Waals interactions, c) linkage of similarly charged
polymerie groups and the partiele surface of a divalent and trivalent ion of opposite
sign. When the polymer and the surface are oppositely charged, general electrostatic
forces act in addition to one or more of the above-mentioned mechanisms,
Flocculation behaviour of sewage sludge 135
In this section we will discuss in detail three different mechanisms of destabilization of
charged particles by polyelectrolytes of opposite charge [Levine and Friesen, 1987]:
charge neutralization, bridging mechanism and electrostatic patch mechanism.
Charge neutralization
Cationic flocculants interact strongly with surfaces of opposite charge and are
adsorbed, at least up to the point of charge neutralization (iso-electric point). In this
way electrostatic repulsion between particles is eliminated, attractive forces become
effective and flocculation may occur. Because of the strong electrostatic interaction,
polyelectrolyte ebains should adopt a rather flat contiguration on oppositely charged
surfaces. At excess counter charged polymer concentrations, surfaces become
saturated which may result in charge reversal. This phenomenon is called restabiliz
ation of particles by the adsorbed polymer.
Bridging mechanism
Bridging flocculation is dependent on the adsorption of polymer segments onto
colloidal particles. The adsorption should not be too strong, since a fair proportion of
segments must remain unattached and available for adsorption on other particles. So
bridging flocculation requires the attachment of the adsorbed polymer to vacant sites
on other particles, thereby creating increasingly larger flocs.
The contiguration of the adsorbed polymer depends on the degree of electrostatic
attraction between the ebains and the surfaces. The amount of electrostatic attraction is
influenced by:
• The charge density of the polyelectrolyte.
At high charge densities the polymer adopts a nearly flat conformation on the
surface. At low charge densities the polymer-particle attraction is weak and the
contiguration of the polymer consists of more loops and tails.
• The ionic strength of the solution.
Bridging is possible when the adsorbed polymer spans the distance over which
double layer repulsion operates k 1). At higher indifferent electrolyte concentration,
the diffuse layers are less extensive and bridging is more pronounced. When more
loops and tails tend to extend further into solution, more segments of the chain will
permitadsorption.
The contiguration of the adsorbed polymer also depends on the flocculation rate. In
bridging flocculation the following rate processes are involved:
136 Chapter 6
• Mixing of polymer solution and sludge dispersion.
• Collisions between polymer and particles, leading to attaclnnent. The rate of this
process is determined by the concentration of polymer and particles and by the
migration speed of both species in solution. Encounters between particles and
polymer molecules may be brought about by a diffusion-controlled process (peri
kinetic flocculation) and by fluid motion (orthokinetic flocculation).
• Reconformation of polymer molecules at the surface of the particles.
• Partiele collisions during which bridges may be formed.
We consider two possibilities: (1) the adsorption is slow with respect to the reconfor
mation process and (2) the adsorption is faster than this reconformation (figure 6.6).
In case (1) the attached polymer ebains are within the bounds of the double layer.
Flocculation can be induced by salt addition (depression of the double layer) and
effective bridges are created. This type of bridging flocculation is called 'equilibrium
flocculation'.
__ .--~ .. altachment
·-~.
low altachment rate
1
no jlocculation
I low collision rate
~
high co Dision
rate
non-equilibrium jlocculoiion
M-.·> salt addiiÎ~~./~
···· ....
equilibrium floccukJtion
Fig. 6.6 Schematical representation of the mechanisms of bridging flocculation in charged
systems [Pelssers, 1988].
Flocculation behaviour of sewage sludge 137
In case (2) enough polymer is attached and still available in the extended state so that
effective bridging is possible. The partiele collision rate determines whether or not a
stabie bridge is formed. At slow partiele collision rates the polymer ebains will have
flattened before a collision occurs. Equilibrium flocculation may only occur if salt is
added. In the case of a high attachment rate and high partiele collision rate the
polymers stay in the random-coil configuration and enough polymer is available for
bridging to take place during collisions. This phenomenon is called 'non-equilibrium
flocculation'.
Electrastatic patch model
This model arises from a consideration of unlikely charged densities for partiele
surfaces and polyelectrolyte chains. Consequently, it would not be possible for each
charged site on a partiele surface to be neutralized individually by a charged polymer
segment. Even though suftkient polymer may be adsorbed on particles to give zero
net charge, regions of positive and negative charge would still remain. A collision
between two such particles could occur so that positive and negative 'patches' come
into contact and adhere as a result of electrostatic attraction. In this way flocs are
formed.
6.6 An experimental technique to determine the ESA signal and zeta potential
In flocculation behaviour the electrical potenrial of the partiele plays a crucial role.
The zeta potential can be determined, for instance, by the standard techniques electro
phoresis, electroosmosis and streaming potential. Electrophoresis is the most widely
used of the three classica! teclmiques. In a specific experimental teclmique, called
microelectrophoresis, the colloidal suspension is contained in an enclosed cell of small
dimensions and the movement of the particles in an applied electric field is observed
directly. The velocity of the partiele divided by the electric field strength is the
electrophoretic mobility of the particle. The magnitude of the electrophoretic mobility
is a function of the zeta potential. In this stndy, a new electroacoustic teclmique for
the application of electrokinetic measurements of colloidal suspensions is used.
Let us consider a typical probe which can perform electroacoustic measurements (see
figure 6.7). The colloidal system is placed between the electrodes. The delay line,
which is made from a solid nonconductive material, separates the piezocrystal trans
ducer from the electrodes.
138 Chapter 6
U2
ele<l1rodes
Fig. 6. 7 The principles of Ultrasound Vibration Potenfiat (UVP) and Electrokinetic Sonic
Amplitude (ESA) measurements.
When a voltage U2 is applied at the transducer, a sound wave of the same frequency
propagates through the delay line and electrode into the colloid. For particles more
dense than the continuous phase, the motion of the particles will lag bebind the motion
of the liquid. Tbis leads to a relative motion between the particles and the liquid. If
the colloidal particles are charged, the resulting motion creates a periodic polarization
of the electtic double layers and an alternaring dipole moment at the frequency of the
applied field. The alternating dipoles sum up to a potential U1 that can be detected by
placing a pair of electcodes in the suspension.
Tbis effect is termed the Ultrasonic Vibration Potential (UVP) and was first predicted
for electrolyte solutions by Debye [1933]. UVP is measured in units of volts per unit
velocity amplitude of the applied acoustic field. In 1938, Rutgers and Hermans pointed
out that the effect would also be present in colloidal suspensions.
In the case of an applied alternating electtic field, the relative motion between the
charged particles and the surrounding liquid generates a sound wave at the frequency
of the applied field. Each partiele vibrating in the electric field radiates sound which
Flocculation behaviour of sewage sludge 139
sums up to a coherent sound wave when many particles are present. The sound wave
is detected as a voltage U2 •
This effect was discovered by Matec [1985] and has been termed the Electrokinetic
Sonic Amplitude or ESA of the colloid. The magnitude of ESA is the pressure
amplitude per unit electric field generated by the colloid and has SI units of pascals
per volt per meter. Both the ESA and UVP effects can be used to determine an
electrophoretic mobility of the particles, where the mobility in this case is a dynamic
or AC mobility.
The electroacoustic technique for application of electrokinetic measurements is
marketed by Matec lnstruments Inc. The MATEC ESA system uses O'Brien's theory
[1988] for electroacoustic effects in a dilute suspension of spherical particles to
calculate the suspension zeta potenrial from the measured ESA. O'Brien's equation
relating the dynamic mobility P-iw) to the zeta potenrial r of particles suspended in
aqueous systems is given by:
where
and
P-iw) = 2ef (1 +t) G(a) 3'7
G(a)
(X =
-I
where: w = angular frequency (s-1)
fJ.p = density difference between the particles and the liquid (kg.m-3)
a = partiele radius (m)
E = dielectric permittivity of the suspension (C. v-'.m-1)
" = kinematic viscosity of the liquid (m2.s-1)
'7 = viscosity of the liquid (Pa. s)
f = constant
(6.24)
(6.25)
(6.26)
140 Chapter 6
The factor fin equation (6.24) can be assumed to be equal to 0.5 for most cases in
ESA measurements (frequency 1 MHz) where the ionic strength is at least 10'3
moles.(liter)·1 and the zeta potentialis less than 75 mV. This then yields the following
equation for the dynamic mobility:
(6.27)
The equation given by O'Brien for G(cx) is expressed as a complex quantity. When
converting ESA magnitudes to dynamic mobility, it is necessary to calculate the
magnitude of G(cx). The equation for converting the ESA amplitude to the dynamic
mobility is given by:
ESA = ~td<fiLlQCGf = re I ~(ex) I cfiáQCGf (6.28)
where cfi equals the volume fraction of particles, Gr represents a correction factor for a
given electrode geometry and c represents the velocity of sonnd in the suspension. For
the case of a parallel plate electrode geometry, Gr equals one.
This calculation is only valid nnder the condition that the electrical donbie layer is thin
relative to the partiele radins (Ka> 50). In this case ESA is linearly dependent on the
volume fraction. At high volume fractions (cfi > 0.1), hydrodynamic and electrical
double layer interactions lead to a non-linear dependenee on the volume fraction. Gen
erally, non-linear behaviour can be expected when the double layer thickness is
comparable to the interpartiele spacing.
The lntrasonic Vibration Potential (UVP) is related to the Electrokinetic Sonic
Amplitude (ESA) of the colloid by:
ESA(w) <fiLlQCG~d(w) UVP(w) = = ----
K* K• (6.29)
where K* represents the magnitude of the complex conductivity.
The complex conductivity is used to characterize electrical properties of colloids,
which are neither dielectrics nor conductors. The magnitude of the complex conductiv
ity needs to be known for conversion of the UVP signal to the dynamic mobility. 1t is
not easy to measure the complex conductivity and this is one of the limitations of the
UVP mode analysis. In this study, experiments with the MA TEC ESA system have
been carried out in the ESA mode.
Flocculation behaviour of sewage sludge 141
1. Sladie 1ample l. Titraa.t 3. Stlrrer 4. Platla.D.IIl ltTD temp. probe 5. Coada.c:ttrity electrode 6. pH eleetrode 7. ESA ultrUG.D.lc probe 8. Tefloa. 1euor .._bly
Fig. 6.8 Schematic diagram of the MATEC sample eelt.
In figure 6.8 a schematic diagram is presenred of the sample cell. The vessel is stirred
by a paddie type agitator and includes sensors for pH, temperature, electrical conduct
ivity and electroacoustic measurements. Electroacoustic measurements are carried out
with Matec's ESA probe sensor which fits directly in the vessel. The electroacoustic
properties that are determined are the ESA amplitude and the phase angle between the
applied electric field and the response signal. The system includes a syringe pump
burette for carrying out volumetrie titrations with simultaneons electrokinetic measure
ments (concentration series measurements). The user specifies the total volume to be
titrated, the volume increment, and a time delay between increments. Tbe concentra
tion series module is suitable to study destabilization phenomena of a sludge suspen
sion induced by flocculants.
142 Chapter 6
6. 7 Results and discussion
Experiments with inorganic jlocculants
In order to study the flocculation behaviour of sewage sludge in detail, only experi
ments were carried out with secondary sludge suspensions originating from the
Eindhoven waste water treatment plant. Experiments with other sludges were carried
out incidentally.
Figures 6.9 and 6.10 showtheresult of a duplex experiment in which a ferric chloride
solution was added to 250 m1 of an Eindhoven sludge suspension. The dry solids
content of the sludge suspension was 2 wt%. The volume increment of the ferric
chloride solution used was 1 ml, the time between two additions 30 seconds and the
total of titrated volume 25 ml. The sludge suspension was agitated with a propellor
typed stirrer having a speed of 500 rpm. The total of added volume (25 ml) of the
ferric chloride solution corresponded with a dosage of 300 g FeC13/kg ds. In tigure
6.9 the ESA amplitude and the phase angle are given. as functions of the ferric
chloride dosage.
0.10
0.09
0.08
0.07
~ 0.06 0.05
i 0.04
0.03
~ 0.02
j::l::l 0.01
0.00 -0.01 -
-0.02
-0.03
0
180 ------ ~-----~: :~-,--
r-"""""'"" . ".r .. '
160
140
120
100
80
60
40
20
0
-20
30 60 90 120 150 180 210 240 270 300 g FeC13 /kg els
.......... ~ ~ ~ -1
Q)
j ~
Fig. 6.9 ESA amplitude (-;---) and phase angle (- -;- --) as functions of the FeC/3 dosage.
Result of a duplex experiment.
0.10 O.o9 0.08 0.07
~ 0.06 0.05
1 0.04 0.03
~ 0.02 0.01 0.00
-0.01 -0.02 -0.03
3
Flocculation behaviour of sewage sludge
. "" ' I \ . \•
\ .. '
'•
'· ·"...:_...,:,_
4
.,, ;~
5 6
pH
Fig. 6.10 ESA signa/ as a function of the sludge suspension pH.
143
7 8
In figure 6.10 the ESA amplitude is depicted as a function of the suspension pH.
Conversion of the ESA signal to the zeta potential is not practicabie in the sewage
sludge/FeC13 system. During an experiment carried out with a sludge suspension, the
zeta potential, partiele radius, and volume fraction of particles will change and all
influence the magnitude of the ESA signal (equation 6.28). The exact changes of
partiele radius and volume fraction in time are unknown, so it is not possible to
convert the ESA signal to zeta potential. Moreover, we are not interested in the
absolute magnitude of the zeta potential but in the course of the ESA signal in time.
At the beginning of the experiment (FeC13 was not yet added to the sludge suspen
sion), the ESA signal sign was negative, which confirms the negative charge of the
sewage sludge particles (see section 6.2).
In an experiment, the suspension pH decreases due to both the weak acidic nature of
ferric ions (see section 6.4) and the increase of ferric ion concentration in the
suspension. At a certain FeC13 dosage and pH, ESA reversal and consequently charge
reversal (- to +) occur in both experiments. The ESA reversal is sharp and discon
tinuous and occurs in a narrow pH range. Marlow and Fairhorst [1988] attributed this
144 Chapter 6
discontinuity to the presence of an undesired electrolyte signal, which is appreciably
comparable to the colloid signal.
Debye [1933] predicted that an ultrasonic sound wave passing through an electrolyte
solution would result in different displacement amplitudes and phases between anions
and kations. The relative displacement of anions and kations produces a separation of
charge accompanying the sound wave resulting in potential differences. This mechan
ism is called the 'Debye effect' and the resulting alternaring potential is termed the ion
vibration potential (IVP). The displacement of a charged partiele in a colloidal system
from its surrounding 'ion atmosphere' induces a similar alternaring potential termed
the colloid vibration potential (CVP) [Hermans, 1938; Rutgers, 1938]. The IVP is
generally out of phase with the CVP. CVPs are normally orders of magnitude larger
than ion vibration potentials (IVPs).
Bruil [1983] measured the ESA signal of a titanium dioxide suspension (density 4000
kg/m3) as a function of the solution pH. The suspension was titrated with HCl. ESA
amplitudes were in the order of magnitude of 0.1 mPa.m/V. At the isoelecttic point
(pH=7.8), the phase angle sudddenly changed from oo to 180°, which means that
partiele charge reversal occurred. It can be concluded that for heavily dispersed
materials and in the absence of salt in the system the acoustopboretic mobility agrees
with the electrophoretic mobility.
However, when the particles have a low zeta potential, volume fraction, and density
relative to the medium and the continuons phase has a high ionic strength, the
magnitude of the CVP approaches that of IVP [Marlow and Fairhurst, 1988]. This
also occurs when the colloid is near its isoelecttic point.
The partiele concentration (2 wt%), the density difference (30 kg/m3), and the ESA
amplitude (0.01 mPa.m!V) are relatively small in the sludge/FeCl3 system. Possibly,
salt is present in the sewage sludge suspension. Under these conditions the resultant
measured signal is the vector sum of the CVP and the IVP.
In the experiments the phase angle gradually changes from oo to 180° (see figure
6.9). This is also due to the presence of the electrolyte signal [Marlow and Fairhurst,
1988]. However, for most colloids this results in errors in determining the isoelecttic
point that are insignificant [Marlow and Fairhurst, 1988]. At a phase angle equal to
90° the polarity of the CVP changes as well. As a consequence, the polarity of the
sludge partiele changes as well. The charge reversal is attributed to an abrupt increase
of specific adsorption of positively charged hydrolysed ferric ions on active surface
sites of sludge particles. These hydrolysis products react to produce polymers. These
Flocculation behaviour of sewage sludge 145
polymerie species span bridges between adjacent particles, thus forming flocs. After
charge reversal the ESA signal increases due to the continuing polymerization of
hydralysis products and, as a consequence, both the zeta potenrial and partiele radius
increase as well. Obviously, positively charged hydralysis products are present in the
sludge suspension for pH valnes smaller than 6. Moreover, at this critica! pH value
electronentral ferric hydroxide species are formed (see figure 6.5) and precipitation of
the insoluble Fe(OH)3 takes place. Precipitation of Fe(OH)3 only occurs over a small
pH range. At pH valnes larger than the critica! value, negatively charged polymerie
hydralysis complexes are predominant species present in the sludge suspension.
Electrastatic repulsion between sludge particles and complexes binders destabilization
of the sludge suspension.
In the two experiments charge reversal occurs at FeC13 dosages of 34 and 50 g/kg ds.
These dosages are much higher than the critica! coagulation concentration of non
adsorbable trivalent ions: 0.12 g/kg ds. The amount of metal coagulant needed for
polymerization and interpartiele bridging is much higher than the amount needed for
coagulation by double layer depression. lt can be concluded that coagulation does not
play a role in the destabilization of a sludge suspension. The dominant mechanism of
destabilization is specific adsorption of positively charged hydrolysed ferric ions.
Theoretically, the dosage at which ESA equals zero (t-potential equals zero) corres
ponds to some suspension properties: maximum dewaterability, minimum colloidal
stability and minimum sediment volume. Wakeman et al. [1992] showed that the most
rapid filtration of anastase (moderately compressible material) occurred at the point of
zero charge, whereas at greater (either positive or negative) zeta potentials the
filtration rate was reduced.
Charge reversal was also observed in sludges originating from other waste water
treatment plants: Mierlo, Amsterdam, and Lage Zwaluwe.
Determination of the point-of-zero-charge of sludge particles
H+ and (OHY ions are potential-determining ions, because they are the only species
through which the solid could be charged. At a certain concentration of potential
determining ions the sludge surface charge equals zero. Kations and anions adsorb to
sludge particles to an equivalent amount. This situation is known as the point-of-zero
charge (PZC).
146 Cbapter 6
H the conditions in the solution are such that t=O, the system is at its isoelectric point
(IEP). In the absence of chemical adsorption, the point-of-zero-charge and isoelecttic
point coincide. Due to specific adsorption, the charge between surface and slipping
plane ub is higher than it would be if no such specific effects were present. H some
cations adsorb specifically at the surface of the sludge particles, the new IEP is
attained by compensating the positive charge between surface and slipping plane ob by
(OHt adsorption, i.e. by increasing the pH. So specific adsorption of kations shifts the
IEP to higher pH values. In this way the occurrence of specific adsorption is tested.
o.oa O.o&
I 0.04
0.02
0.00
~ -0.02
-0.04
-0.08
1
't\' 1· .. I' . ','
' .. '
I I I
.I. l l l
'
2 3 4
pH
5 6 7
Fig. 6.11 Determination of the point-of-zero-charge of a sewage sludge suspension. Results of three experiments in which sludge suspensions were titrated with a HCl solution. The
continuons line (-) is the result of an experiment carried out with a sludge suspension adjusted with ]()'4 M KCl. The dotted lines (--) and (· · ·) are the results of experiments
carried out with a pure sewage sludge suspension.
In order to determine the point-of-zero-charge of an Eindhoven sludge suspension,
three experiments were carried oot with the MA TBC ESA system. A sludge suspen
sion was titrated with a HCl solution in two experiments (duplex experiment). In the
third experiment a HCI solution was added to a sewage sludge suspension adjusted
with KCI toa concentration of 1(}'4 M. KCI is assumed to be an indifferent electrolyte.
The point-of-zero-charge is found as the intersection of the three ESA versus pH
curves. At the intersection point the surface charge density is independent of the
Flocculation behaviour of sewage slndge 147
indifferent electrolyte concentration and equals zero. In tigure 6.11 the results of the
experiments are shown. The suspension pH decreases during the experiments. Charge
reversal (- to +) is observed in the experiments. The pH at which ESA equals zero (t
potential 0) corresponds to the isoelectric point of sewage sludge particles.
Because of the absence of specitic adsorption in the experiments, the isoelecttic point
coincides with the point-of-zero-charge.
Theoretically the three curves must interseet at the point-of-zero-charge. This is not
the case, although the differences are very small. It can be concluded that the pH of
zero partiele charge (PZC) and thus the pH at the isoelectric point (IEP) both equal 2.
The determined PZC is in good agreement with the PZC of the Si02 , which is the
main inorganic material present in sewage sludge solids (about 30 wt%). The
isoelecttic point of the sludge/FeCl3 system equals 6 (see tigure 6.10). The isoelectric
point of the sludge particles shifts from 2 to 6 due to the presence of ferric chloride in
the sludge suspension. It physically means that specific adsorption of positively
charged hydrolysed ferric ions takes place. The results of the experiments presented in
tigure 6.11 are an additional proof of the occurrence of chemica! adsorption in the
system.
+
CR 1 CR2 CR3
pH
Fig. 6.12 Schematic illustration of the general electrophoretic mobility behaviour of colloid
systems in the presence of hydrolysable metal ions.
148 Chapter 6
James and Healey [1972] studied the adsorption of hydrolysable metal ions at the
oxide-water interface. They showed that hydrolysable metal ions are able to reverse
the charge of anionic colloidal substrates. The study was focused on the adsorption of
Co(ll) by Si02 and Ti~ colloids. The electrophoretic mobility behaviour, supple
mented by streaming potentlal data, was determined as a function of the pH.
0.10 9
0.05 ,... ""': ..... 8 I I I
i ' 0.00 I 7 I I 1', I I
' I I
' -6.05 I' ' ' 8 i ' ' ' .\
~ ' ' ' -0.10 ., ' 5
" "' -~ ' --~
~
-0.15 "', _",
4 >-· .... .:....:. ~ .,. .... '
-0.20 3
0 10 20 30
do8ap (ml)
Fig. 6.13 ESA signal as a junction of the dosages of FeCl3 and KOH solutions. lnitially, the
suspension was titrated with a ferric chloride solution ) until the occurrence of charge
reversal (- to + ). The suspension pH ( .. .) decreased. Afterwards, the suspension was titrated
with a KOH solution (--). Charge reversal occurred (+ to -). The suspension pH ( ... )
increased.
The various charge reversals observed are, in order of increasing pH (see figure
6.12), shown to represent the point-of-zero-charge on the Si02 substrate (CRI),
surface precipitation of hydrolysable metal ions (CR2) and the point-of-zero-charge of
the metal hydroxide solid itself {CR3).
CRI and CR2 were also observed in experiments carried out with a sludge suspension.
An experiment in which a sludge suspension was used as a colloidal substrate was set
up to check the occurrence of CR3. Initially, a ferric chloride solution was added to
the sludge sample until the occurrence of charge reversal CRl (- to + ).
Flocculation behaviour of sewage slndge 149
Subsequently, the sludge suspension was titrated with a KOH solution. The experimen
tal result is shown in tigure 6.13. Charge reversal CR3 ( + to -) occurred at a pH of
4.5. If sufficient metal ion is adsorbed to yield a complete coating of adsorbed metal
hydroxide, the charge reversal corresponds to the point-of-zero-charge of the metal
hydroxide. The point-of-zero-charge of ferric hydroxide is presented by the reaction:
The PZC of ferric hydroxide occurs at pH=8.5 [Parks and de Bruyn, 1962]. In the
experiment CR3 occurred at a lower pH value than the PZC of ferric hydroxide. This
is possibly caused by incomplete coating of adsorbed metal hydroxide on particles.
Simulation of the flocculation process in practice
0.14 14.0
0.12 12.5
I 0.10 11.0
0.08 9.5
0.06 8.0 a ~
0.04 ' ., . 6.5 ' '
r.:.::l 0.02 5.0
0.00 --- 3.5
-0.02 2.0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
felTic chloride dosage (ml)
Fig. 6.14 Simu/ation of the jlocculation process induced by ferric chloride/time in practice.
ESA signal (- ) and suspension pH (---) as a function of the ferric chloride dosage. After
addition of 20 ml of the FeCl3 solution, JO ml of time dispersion (350 glkg ds) was dosed in
one step.
150 Chapter 6
Lower concentradons of metal or higher concentradon of colloidal substrate will
reflect surface-coated and uncoated areas. The formed insoluble ferric hydroxides will precipitate and eumesh the sewage sludge particles (sweep flocculation).
An experiment was carried out to simulate the flocculation process induced by ferric
chloride/lime in practice. In figure 6.14, the result of the experiment is shown. First,
250 m1 of a sludge dispersion (dry solids content 2 wt%) was titrated with a ferric
chloride salution up to a total dosage of 25 ml, which corresponded with a dosage of
375 g!kg ds. During the addition of ferric chloride the suspension pH decreased. At a
pH of 6.3 the expected charge reversal (- to +) occurred. After 20 m1 of the ferric
chloride salution (300 g!kg ds) was dosed, 10 m1 of a lime dispersion (350 glkg ds)
was added to the sludge suspension in one step. As a consequence both the ESA signal
and pH sharply increased (figure 6.14). The sudden increase in ESA signal is possibly
attributed to specific adsorption of calcium ions at sludge particles covered by ferric
hydrolysis complexes. Other processes may occur simultaneously, such as the
formation of negatively charged ferric hydralysis products and the precipitation of iron
hydroxide due to the strong increase in pH.
Experiments with organic jlocculants
In order to study the mechanism of cationic polyelectrolyte flocculation of a sewage
sludge dispersion, two different polyelectrolytes were used in experiments carried out
with the MATEC ESA system: Röhm KF 975, which is the successar of Praestol
444K and Röhm KF 945. Both polymers are widely used as sewage sludge condi
tioners. Röhm KF 975 is a very strongly cationic polyelectrolyte, having a high
molecular weight (viscosity 2000-5000 mPas). Röhm KF 945 is a strongly cationic
polyelectrolyte, having a viscosity of 50-200 mPas.
Figures 6.15 and 6.16 show the results of two experiments in which a 250 m1 sludge
dispersion (dry solids concentradon 1.75 wt%) was titrated with the polyelectrolyte
solution Röhm KF 945 and Röhm KF 975, respectively.
The ESA signal (and thus the t potential) of the flocculated sludge sample changed
sign from negative to positive with an increase in the amount of cationic polyelectro
lyte, after which it approached a constant positive value. It means that positively
charged polymers were adsorbed at the sludge partiele surface. The polymerie ebains
adopted a nearly flat conformation on the partiele surface due to the strong electra
static attractive forces between polymer and particle.
0.012 0.010 0.008 0.006
0.002
Flocculation bebaviour of sewage sludge 151
i 0.004
0.000 +-+---'---.;.__--,----'----------'------'----'-------1
-0.002 ............. -0.004
~· -0.006 ~ -0.008
-0.010 -0.012 -0.014
-0.~ ~~~~~~~ .. ~~~~~~~~~~~~~~~~
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
polyelectrolyte dosage (g KF 945/kg ds)
Fig. 6.15 Measured ESA signal (related to the zeta potential) as a function of the dosage of
polyelectrolyte KF 945. Charge reversal occurred at a dosage of 0.03 g KF945/kg ds.
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0
polyelectrolyte dosage (g KF 975/kg ds)
Fig. 6.16 Measured ESA signal (related to the zeta potential) as a function af the dosage af
polyelectrolyte KF 975. Charge reversal accurred at a dosage of 0.12 g KF 975/kg ds.
152 Cbapter 6
At the cationic polyelectrolyte dosage wbereby the zeta potential becomes zero, the
repulsive forces between sludge particles decrease and tlocculation due to charge
neutralization may occur. The pbenomenon of charge reversal caused by adsorption of
polyelectrolytes was found by many investigators. Gregory [1969] made a detailed
analysis of the flocculation of anionic polystyrene latex particles by cationic polyelec
trolytes. The electrophoretic mobility of the particles cbanged sigmoidally from
negative to positive with an increase in polymer dosage. Black et al. [1965] investig
ated the use of polyelectrolytes in tbe destabilization of a kaolin clay suspension. The
isoelecttic point corresponded witb tbe optimum condition for destabilization. At the
optimum dosage the lowest supernatant turbidity was found. Igarashi and Nishizawa
[1992] measured tbe zeta potentlal of digested sewage sludge samples witb a new
streaming potentlal measurement device. The zeta potential cbanged sigmoidally from
negative to positive when tbe cationic polyelectrolyte was increased. When tbe
colloidal charge was zero, tbe gravity tiltration speed reached a maximnm and the
moisture content in dewatered cakes reached a minimum. Roberts [1978] observed that
the optimal cationic polyelectrolyte (Zetag 92) dosage for dewatering of an activated
sludge sample occurred at zero electropboretic mobility.
Charge neutralization is not necessarily the main mechanism teading to destabilization
wben sludge particles and polyelectrolytes are of opposite signs. The polyelectrolytes
used possess a high molecular weight. Bridging is more efficient witb polyelectrolytes
of a high molecular weight since loops will tend to extend further into solution.
Moreover, tbe high charge density of the polyelectrolytes used results in strong
electrostatle repulsion between adjacent ebains and consequently the polyelectrolyte
ebains are stretched. This also leads to a more effective bridging mechanism.
Destabilization of a sewage sludge suspension by the bridging mechanism is possible
at polymerie concentrations higher than tbe concentration at which charge reversal
occurs (point-of-zero-charge).
The degree to which a sludge suspension will be stabilized is difficult to determine.
Many factors are involved in the destabilization process by polyelectrolytes: partiele
concentration, inter-partiele interactive forces, ionic strengtb, polymer contiguration in
tbe adsorbed state and intensity and duration of mixing. The influence of these
parameters on the destabilization process of a sewage sludge dispersion was not
investigated in tbis study.
The phenomenon of charge reversal was only observed in a few experiments. In many
experiments the expected charge reversal did not occur. Possible explanations are:
Flocculation behaviour of sewage sludge 153
• The sludge flocs formed could not be detected by the ESA probe. The sludge flocs
could not pass the small gap between the two electrodes.
• The time available for reconformation was large and equal to the time between two
dosages (30 s). The fust attached polymers could flatten to an extension smaller
than the 'Debye length' K-1 before further adsorption took place. Electrostalie
repulsion between the particles hindered flocculation.
Discussion about the experiments carried out
In some experiments carried out with various sewage sludge samples, the expected
charge reversal due to specific adsorption of hydrolysed ferric ions on the surface of
sludge particles did not occur. The real reason is unknown. However, several factors
which are inherent in the sewage sludge/FeC13 system could disturb the measuring
result:
• The volume fraction of particles il> in the sewage sludge suspension is small (1 to 4
wt%). Consequently the ESA signalis small (equation (6.28)).
• The density difference tJ.p between sludge particles and the liquid phase is small
(!J.p ::::::30 kg/m3). A small density difference yields a small ESA signa!, too (equa
tion (6.28)).
• As already stated in this section, an undesired electrolyte signal plays an important
role in the sewage sludge/FeC13 system. A few experiments were carried out to
determine the relative importallee of the electrolyte signa! in the sewage
sludge/FeC13 system.
Two sewage sludge samples and demi-water were titrated with a ferric chloride
solution. The results of the experiments are presented in tigure 6.17. It can be
concluded that the ESA signal measured in the demi-water/FeCl3 system is of the
same order of magnitude as that measured in the sewage sludge/FeC13 system. The
electrolyte signalis caused by growing ferric hydrolysis complexes.
154 Cbapter 6
0.12 .---------------------------------------------~
0.08
0.04
-0.04 L..L->---c.__._......._,..._._......___._._..J......J.__,_.i..-L_._ ......... _,__~.....~.. .......... ._._ ............ __._ .......... _.__.___._.._.~..-~....J
0 50 100 150 200 250 300 350
dosage of FeCl3 (g/k.g ds)
Fig. 6.17 ESA signa! as a function of the dosage af FeC/3• The continuous line represents the
ESA signal of the demi-water-FeCl3 system. The dotted lines are the results of experiments carried out with the sludge-FeCl9 system.
6.8 A model to describe hydrolysable metal ion adsorption at the sludge solid
lValer interface
James and Healey have presented a model for the adsorption of hydrolysable metal
ions, which provides a general onderstanding of the phenomena observed in Co(ll)
adsorption by SiO:z or Ti02 colloids. The model is used as a base for the description of
adsorption of hydrolysable ferric ions at the surface of sludge solid particles.
James and Healey measured adsorption densirles of co2+ on Si02 colloids and
concluded that these metal ions characteristically do not adsorb until a critical pH is
obtained. At pH valnes smaller than this critical pH the adsorption is close to zero.
The most likely interaction that could prevent adsorption of (hydrolysed) cations on
negative surfaces against electrostatic attraction below the critical pH is a solvation
term [James and Healey, 1972]. Changes in solvation energy areexpressedas changes
in primary and secondary hydration. James and Healey concluded that the adsorbed
species are separated from the solid surface by at least one layer of water molecules so
that direct chemical bonding is precluded. Consequently the meta1 ions are not
required to lose their primary hydration shells. The solvation energy term refers to the
Flocculation behaviour of sewage sludge 155
removal of the second hydration layer of a cation. The total work of cation adsorption
is separated into a simple coulombic term and a secondary solvation term. The
coulombic term will be corrected with a 'chemical' free energy term where necessary.
The adsorption of metal ions at the sludge solid-liquid interface is treated in terms of
competing energy changes as the ion approaches the interface. The change in free
energy of adsorption ..1Gads [J.mot-1] equals the sum of the change in coulombic energy
..1Gcoul [J.mot'], the change in secondary solvation energy .dGsoiv [J.mot'] and a
specitic adsorption energy contribution ..1Gchem [J.mol-1], i.e.
(6.30)
The energy terms .dGcoui and ..1Gchem are favourable to adsorption, whereas .dGsoiv is
unfavourable. Adsorption may occur when:
(6.31)
The change in coulombic energy by brioging an ion charge ze up to a surface where
the potenrial is t/lx is zet/;x and equals the binding energy. The change in specitic
adsorption is equal to ze4>, where 4> is termed a superequivalent adsorption potenriaL
An expression for the change in secondary solvation energy was derived by James and
Healey [1972].
In the scope of this thesis we present some results which are of relevanee in this
study.
In the region adjacent to charged interfaces electtic fields of considerable magnitude
exist. Water molecules adsorbed on the interfaces will undergo parrial or complete
electrical saturation, in which case the dielectric of these molecules is reduced from
the bulk value of 78.5 to the value of 6.
156
.tl 80 '"" iS
;l ., l " 60 '"' .....
j 40
20
0
~ /
/ (A) I f I I I 13 I I I
I
8
Distance 1 Ä
Cbapter 6
1
2 -
12 0
I I I I I I
Distance 1 Ä
(B)
12
Fig. 6.18 Actual continuous (A) and approximate discontinuous (B) representations of the
variation af the dielectric constant as a tunetion of distance for: (1) a surface of low electric
field; (2) a surface of high electric field; (3) a hydrated ion or a very high electric field
[James and Healey, 1972].
The varlation of the dielectric of water er,x as a function of elistance from the interface
is expressed by [Anderson and Bockris, 1964]:
€bulk -6 6 er.x = [ dY, 2] +
1 + B (-) dx
(6.32)
where the constant B bas a value of 1.2·10-17 m2.V'2 and (dY,/dx)x represents the
electtic field strength estimated from the Gouy-Chapman model of the double layer. It
is possible to make a good approximation to the system by dividing the interface into
three regions and using a constant value of the dielectric in each of the regions (see
figure 6.18).
The regions are:
1. The solid ( e, = Eson.J.
2. The layer of adsorbed water at the interface (er=eintr= 6).
3. Water outside the adsorbed layer (er=ebulk=78.5).
Flocculation behaviour of sewage sludge
(a)
te---: r·"': 1on
157
(b)
Fig. 6.19 Representation of the solid-liquid intelface showing two of the possible locations of
an adsorbed ion: (a) in the diffuse double layer; and (b) near the inner Helmholtz plane
[James and Healey, 1972].
There will be a discontinuity in the dielectric of the medium at each boundary of the
three regions. Becanse of this artificial division in the model there are two general
locations for the hydrated metal ion ( radins rion + 2r w) with respect to the interface. The
fust is for the case where the primary hydration sheath of the ion and the water
adsorbed on the solid do not overlap, as is shown in tigure 6 .19a. The secoud case
covers the possibility that the primary hydration shell of the ion may include the
adsorbed water ( tigure 6.19b). For case 1, .:lGsoiv [J.ion-1] is given by [James and
Healey, 1972]:
(6.33)
For case 2 the change in secondary solvation energy is represented as [James and
Healey, 1972]:
158 Chapter 6
r. ) 2(r. +2r \2
ton wl
1 1 1 - _1) Eintf
(6.34)
Both equations represent the change in secondary solvation free energy in moving a
hydrated metal ion from the bulk solution to the inner Helmholtz plane at the solid
liquid interface. The dielectric constant of sludge solids is assumed to be the dielectric
constant of SiÜ:l, the main inorganic material present in sewage sludge solids. The
dielectric constant of Si02 equals 4.3. From equations (6.33) and (6.34) it can be
concluded that .1.Gso1v > 0, so work must be done to remove the second hydration
layer of a cation and reptace it by interfacial water with a very low dielectric constant.
At the start of the continuons series experiment, presenled in figure 6.9, in which a
sludge suspension is titrated with a ferric chloride solution, only negatively charged
hydrolysed ferric ions are present. For suspension pH valnes between 7 and 6, no
adsorption occurs due to electrostatle repulsion between particles and negatively
charged hydrolysed complexes:
(6.35)
During the experiment the suspension pH decreases until the predominant form is
Fe(HzOMOH)2 + (at pH equals 6). Adsorption of these monovalent complexes is
energetically favourable, becanse of the validity of the inequality:
(6.36)
If more salt is added, higher positively charged (twovalent and trivalent) hydrolysed
complexes are formed. The equilibrium as presenled in section 6.4 is driven to the
left. As a result, the average ionic charge increases and the suspension pH decreases.
The increase of the average ionic charge ze increases the coulombic, the specific
adsorption, and the solvation energy. However, because both the conlombic and the
specific adsorption energy vary with the valency z, and the solvation energy change
varles with z2, the increase of charge due to hydrolysis increases the magnitude of the
Flocculation behaviour of sewage sludge 159
solvation term much more than the som of the conlombic and specifïc adsorption term.
As a consequence, above a certain average ionic charge and thus below a certain pH:
(6.37)
Thus, adsorption of higher charged hydrolysed products is energetically unfavourable.
The general condusion can be that within a certain pH range chemica! adsorption of
hydrolysed ferric ions at active sites of sludge particles occurs. Outside this pH range
speciiic adsorption does not occur. Since total ion concentrations are different no exact
calculations of ~Gads can be made. However, in tigure 6.20 a schematic presentation
of the free energy of adsorption as a function of the suspension pH for Eindhoven
sludge is given.
+
4 5
pH
Figure 6.20 Free energy of adsorption as a function of the suspension pH for Eindhoven
sludge.
6.9 Conclusions
The flocculation behaviour of sewage sludge was studied with the MA TEC ESA sys
tem, an electroacoustic technique for application of electrokinetic measurements. With
this system the Electrokinetic Sonic Amplitude (ESA) of a (flocculated) sewage sludge
suspension was measured, which is related to the zeta potenrial of sludge particles.
160 Chapter 6
The study was focussed on the various destabilization phenomena of sludge suspen
sions indoeed by the flocculants ferric chloride and polyelectrolytes.
The dominant mechanism of destabilization of a sludge suspension induced by ferric
chloride is specific adsorption of mainly monovalent positively charged hydrolysed
ferric ions at active sites of the surface of sludge particles. Specific adsorption of
positively charged hydrolysis products results in partiele charge reversal (- to + ).
Specific adsorption of these monovalent complexes on sludge particles is energetically
favourable within a certain pH range. As the hydrolysed ferric ion approaches the
sludge particle, the sum of change in conlombic energy and specific adsorption energy
(both favourable to adsorption) is larger than the solvation energy (unfavourable to
adsorption) and adsorption may occur. The solvation energy refers to the removal of
the secoud hydration layer of a ferric ion.
The electrok:inetic behaviour of a sewage s1udge colloid was determ.itled in the
presence of various hydrolysable ferric ions as a function of the pH. Three charge
reversals were observed and listed as CRI, CR2 and CR3. CRI (- to +)is the point
of-zero-charge (PZC) of the colloidal sludge particles and occurred at pH equal to 2.
The measured PZC corresponds with the PZC of Si02 , which is the main inorganic
material present in sewage sludge solids. CR2 (- to + ), occurring at pH equal to 6,
indicated specific adsorption of principally monovalent hydrolysed ferric ions. Charge
reversal CR3 ( + to -) reflected a partial coating of ferric hydroxide on the sewage
sludge particles and occurred at a pH equal to 4.5. Typically, the charge reversals are
sharp and discontinuous, occurring within a very small pH range. This can be
attributed to the presence of an undesired electrolyte signal in the sludge/FeCl3
system. However, errors in detennining the isoelectric point due the electrolyte signal
are insignificant.
The MA TBC ESA system is not suitable to study the destabilization phenomena of
sewage sludge particles induced by cationic polyelectrolytes. Only in a few experi
ments the expected charge reversal occurred. Due to the high charge densities of the
cationic polyelectrolytes used (Röhm KF975 and KF945), the zeta potential becomes
zero at very small polyelectrolyte dosages. The main destabilization mechanism is
charge neutralization. At polymerie dosages higher than the concentration at which
charge reversal occurred, the bridging mechanism is responsible for destabilization.
Partiele bridging is efficient because the polyelectrolytes used possess a high molecu
lar weight and loops tend to extend further into solution.
7 CONCLUDING REMARKS AND PERSPECTIVES
The relevant dewatering properties have been determined of four sewage studges
originating from four different waste water purification plants in the Netherlands. The
design and operation of these waste water treatment plants are different, including
different dewatering techniques (chapter 2). In this way, a great variety of sewage
studges could be studied.
The vacuum exsiccator metbod for water vapour isotherms and thermal analysis
techniques (TGA and DT A) for isothermal drying curves are suitable measuring
methods to determine the water bond enthalpy (chapter 3). The TGA-DTA drying
model can be regarded as an acceptable model for the drying behaviour of sewage
sludges and the determination of the water bond enthalpy as a function of the sample
moisture content. Both measuring methods show that the water bond enthalpy differs
significantly from zero at sample moisture contents smaller than 0.3 to 0.6 kg water
per kg dry sollds. The sludge type, and type and dosage of flocculant do not influence
this critica! moisture content. Water having a bond enthalpy larger than l kJ/kg is
categorized as 'bound water' . and cannot be removed in a mechanica! dewatering
process at applied pressures smaller than 10 bar. So the theoretica( maximum dry
solids content that can be reached is about 65 to 75 wt%. In practice dry solids
contents of about 15 to 35 wt% are obtained. A lot of 'free' water remains enclosed in
the sludge filter cake during the mechanica! dewatering process. Higher mechanica!
pressures are needed to remove more 'free' water out of the sludge filter cake. Water
vapour sorption isotherms measured at different temperatures with the vacuum
exsiccator technique can bedescribed very well with the S-shaped G.A.B. equation.
The dewatering behaviour of sewage sludges was tested with different techniques: the
filtration-expression cell, the conventional Capillary Suction Time test, the compres
sion-permeability cell, and the Modified Piltration Test (chapter 4). The most appro
priate instrument to study the dynarnic dewatering behaviour of sludges is the
fûtration-expression cell. The influence of various process parameters, such as
pressure, slurry concentration (cake thickness), flocculant type and dosage can be
investigated. In the characterization research, floc microproperties, which are of
relevanee for a better understanding of the dewatering process, were determined too.
Sludge floc microproperties of interest are composition (dry solids content, ash
content, A TP content, pH, electrical conductivity), zeta potential, partiele size
distribution and rheological properties (thixotropy). At the optimum flocculation
162 Chapter 7
conditions (flocculant dosage and mixing conditions) some characterization parameters
show a minimum or maximum: minimum specific cake resistance, minimum vacuum
suction time, minimum CST value, minimum iron content in the filtrate, maximum
dry solids content, maximum permeability, maximum median floc diameter, maximum
degree of thixotropy. By definition, the dewaterability of sewage sludges shows a
maximum at the optimum flocculation conditions.
The modified CST test, which continuously measures the position of the liquid front in
a ceramic slab, provides information on the dynamic dewatering behaviour of both
non-flocculated and flocculated sludges (chapter 5). The model, which describes the
position of the liquid front as a function of time, enables the possibility to calculate an
average specific cake resistance from the experimental data. The modified CST test is
an improvement of the conventional CST apparatus, which only measures the position
of the üquid front at two different times.
Prior to dewatering, flocculants are added to the sludge suspension in order to
improve the dewatering behaviour. The addition of flocculants induces the destabiliza
tion of the suspension (chapter 6). The so-called electroacoustic technique is suitable
to study the destabilization mechanisms of sewage sludges induced by ferric chloride.
The dominant destabilization phenomenon is specific adsorption of mainly monovalent
positively charged hydrolysed ferric ions at the active sites of the sludge particles.
Specific adsorption manifests in the reversal of the partiele charge from negative to
positive and occurs within a certain pH range. The pH range can he determined with a
model that presents an energetic consideration of the hydrolysable metal ion adsorption
at the sludge-solid-water interface.
Future areas of research
It is obvious to study the practical applicability of some characterization tests at sludge
treatment plants. The filtration-expression cell and the compression-permeability cell
are to he considered for this study. Nowadays, there is no question of an unambiguous
determination of sewage sludge dewatering properties at sludge treatment plants.
Another important consumer market is the industries which produce sludges as a part
of their waste products. The filtration-expression cell can he nsed to diagnose and
optimize the flocculation process, which has a large influence on the dewatering
hehaviour. The introduetion of the filtration-expression cell may lead to a better
insight into the dewatering process, a better cantrolling mechanical dewatering, and
Concluding remarks and perspectives 163
possibly a smaller use of the amount of flocculants. Achieving higher dry solids
contents yields a rednetion of expenses for the total sludge processing. In the Nether
lands, the costs of the dewatering of sewage sludge rougWy amount to 300 to 400
guilders per ton of dry solids. The annual production of sewage sludge dry solids is
about 300,000 tons of dry solids. Consequently, the total costs of sludge dewatering is
approximately 100 million guilders. A saving in one percent on the sludge dewatering
costs means a saving of one million guilders per year. Mechanica! dewatering is
foliowed by other processing steps, such as drying, incineration, and dumping. The
rate for sludge processing amounts to 150 guilders per ton of sludge cake. The
average dry solids content is 23 wt% , so the rate equals 650 guilders per ton of dry
solids. Consequently, the total expenses for sludge processing annually amounts to 200
million guilders. The increase of the average dry solids content from 23 to 24 wt%
annually yields a cost sa ving of about 7.5 million guilders.
NOTATION
a sphere radius plus thickness of Stem layer [m]
al constant H <ly specitic surface area [mz.m-s]
~ water activity H A cross-sectional area of the filter medium [mz]
Ac area of the cross-section of the inner CST tube [mz]
AH Hamaker constant [J]
Ar heat transferring surface area of reference cup [mz]
A, heat transferring surface area of sample cup [mz]
A .. sample surface area for moisture transport [mz]
ATP = Adenosine-triphosphate
bi constant [-]
B constant in equation (6.32) [mz.v-1]
BOD= biologica! oxygen demand
c velocity of sound in suspension [m.s-1]
Cr calibration factor [W."'v-lJ
Ctc conversion factor [/kV.K1]
Cr GA correction factor [-]
Cv concentration of solids in suspension [k:g.m-3]
c cake mass deposited per unit filtrate volume [k:g.m-3]
Co critica! coagulation concentration [mol.lite11]
Cs BET adsorption constant [-]
CB,O BET adsorption constant [-]
cg Guggenheim constant [-]
Cg,O Guggenheim constant H cp.d specific heat of water vapour [J.kg·l.KI]
cp,ds specitïc heat of dry sludge solids [J .kg·l.Kl]
cp.r specific heat of reference cup [J .kg·l.K-1]
cp,w specitic heat of water [J.kg·l.Kt]
CR charge reversal
CST Capillary Suction Time
CVP colloid vibration potentlal
~ mean partiele diameter [m]
166 Notation
d. = density of dry solids [kg.m-3]
ds dry solids
d,. density of water [kg.m-3]
DTA= differential thermal analysis
e elementary charge [C]
Et difference between molar sorption entbalpy in fust layer and in the mth layer [J.mol1
]
~ difference between molar sorption entbalpy in mth layer
and condensation entbalpy [J.mot-1]
ESA= electrokinetic sonic amplitude
f = parameter defined by equation (6.24) [-]
fractional distance in the temperature boundary layer [-]
fw fugacity of water [-]
e fugacity of pure water at standard temperature (-]
p' suction force exerted by filter medium under CST tube [N]
g gra vitational acceleration [m.s-2]
.6.G Gibbs free energy [J]
.6.Gads = change in adsorption energy [J.mot1]
.6.Gchem= change in chemica! adsorption energy [J.mol1]
.6.GcouF change in conlombic energy [J.mot-1]
Gr = correction factor for a given electrode geometry [-]
.6.Gso1v= change in secondary solvation energy [J.mol-1]
h thickness of cernmie slab [m]
H = height of sludge layer in CST tube [m]
.6.H = enthalpie change [J]
Ho distance between two spheres [m]
.6.Hb bond entbalpy of water in slndge cake [J.kg-1]
.6.Hcow:F condensation enthalpy [J.mot1]
Hds entbalpy of sludge dry solids [J.kg-1]
.6.Hexc= excess entbalpy of sorption [J.mot1]
.6.H = m adsorption entbalpy in the mth layer [J.mot1]
H,. = enthalpy of reference cup [J.kg-1]
H, enthalpy of sample cup [J.kg-1]
.6.Hsor = sorption entbalpy [J.mol1]
.6.H= V entbalpy of evaporation of pure water [J.kg-1]
Notation 167
.ó.Hv,o= enthalpy of evaporation of water at 273.25 K and 1 bar [J.kg-1]
HW enthalpy of water [J.kg-1]
.ó.Hw differential enthalpy of wetring [J.mot1]
i.e. inhabitant equivalent
IEP isoelectric point
iHp inner Helmholtz plane
IVP ion vibration potenrial
jw moisture flux [kg.m·2 .s 1]
k Boltzmann constant [J.Kl]
GAB constant [-]
ko GAB constant [-]
kKc Kozeny constant [-]
K permeability [mz]
Ko permeability at p,=O [mz]
K,.v average permeability [m2]
KP permeability of filter medium [m2]
Koo equilibrium permeability [m2]
K• complex conductivity of suspension [-]
length of inner cylinder of rheometer [m]
L cake thick:ness [m]
11\is mass of dry solids [kg]
m,. mass of reference cup [kg)
m. mass of sample cup [kg]
illw water mass [kg]
illw,oo = water mass at the end of a CP-cell experiment [kg]
.:lm., = loss of water mass [kg]
.:lm.,,.,.;= total loss of water mass [kg]
MCC= microcrystalline cellulose
MFT= Modified Piltration Test
llo concentration of ions in bulk solution [number.m 3]
n+ local concentration of positive ions [number.m 3]
n+o concentration of counter-ions at infmity [number.m-3]
n. local concentration of negative ions [number.m-3]
ll.o concentration of co-ions at infinity [number.m 3]
llç critica! coagulation concentration [number.m 3]
168 Notation
NA Avogadro's number [-]
oHp outer Helmholtz plane
Pa constant [Pa]
p.e. polyelectrolyte
p, compressive pressure [Pa]
Pw = partial vapour pressure of water [Pa]
p~ vapour pressure of pure water [Pa]
p pressure [Pa]
äP = applied pressure difference [Pa]
Po = hydraulic pressure at position r0 in ceramic slab [Pa]
pcap capillary suction pressure [Pa]
.ó.Pp pressure difference across the filter medium [Pa]
.ö.Pt = hydraulic pressure drop [Pa]
äP, pressure difference across the sludge cake [Pa]
PZC = point of zero charge
Q heat flow [J.s-1]
Q = liquid flow [m3.s-l]
r radial distance [m]
= position of liquid front at time t [m]
ro position of liquid front at time t=O [m]
re radius of ceramic slab [m]
re distance from centre of ion to adsorbed water layer [m]
rion radius of ion [m]
fw radius of water molecule [m]
R general gas law constant [J.mol1.K1]
R.v average radius [m]
~ = radius of outer cup of rheometer [m]
Re distance between eentres of two spheres [m]
Re = Reynolds number [-]
Reent= critical Reynolds number [-]
Re electrical resistance [0]
Re.d = electrical resistance of dry part of ceramic slab [0]
R.o.tot = total electrical resistance of ceramic slab [0] R.,,w electrical resistance of wetted part of ceramic slab [0] RH = relative humidity [-]
Notation 169
R", ftlter medium resistance [m-I]
R. radius of inner spindie of rheometer [m]
s ratio between distance between eentres of two spheres (R.,)
and sphere radius (a) [-]
s thermocouple signa! [t-tV]
AS entropy change [J.Kl]
t time [s]
T temperature [K]
torque [N.m]
Tgas gas temperature [K]
TGA thermogravimetrie analysis
Tr reference temperature [K]
Tr,exp measured reference temperature [K]
T, sample temperature [KJ
Ts,exp measured sample temperature [KJ
Tw water temperature [K]
Twall furnace wall temperature [K]
u solid based moisture content [kg w. kg ds-1]
u' bound water content [kg w. kg M 1]
UI solid based moisture content in monolayer [kg w. kg ds-1]
UVP ultrasound vibration potential
Vw molar volume of water [m3.mol-1]
V ftltrate volume [m3]
VA London-van der Waals attraction potential [J]
VR potential energy of repulsion [J]
VST vacuum suction time
VT total potential energy of interaction [J]
x mole fraction [-]
diameter [m]
distance co-ordinate [m]
Xo maximum diameter [m]
z valency [-]
170 Notation
Greek symbols
(l' = parameter defined by equation (6.26) [-]
Cl'av average specitic cake resistance [m.kg-1]
Cl'eff = effective heat transfer coeffident [J.s·1.m·2.K"1]
Cl'r convective heat transfer coeffident for reference cup [J.s·1.m·2.K1]
a, convective heat transfer coefficient for sample cup [J .s·1 .m-2 .K1]
6 compressibility coefficient [-]
redprocal hydraulic radius [m-1]
'Y interfacial tension of water or filtrate [N.m-1]
i' shear rate [s-t]
'Yo = parameter defined by equation (6.15) [-]
0 Stem layer thickness [m]
compressibility coefficient [-]
ÓT thickness of the temperature boundary layer [m]
€ porosity [-]
emissivity [-]
dielectric permittivity of suspension [C. v-•.m-1]
Eo = permittivity of vacuum [N-1 .m-2. C2]
porosity at p.=O [-]
€." = equilibrium porosity [-]
€bulk = dielectric constant of bulk material [-]
Eintf dielectric constant of adsorbed water [-]
Er dielectric constant of medium [-]
Es solidosity [-]
Esolid = dielectric constant of solid [-]
E,_o = solidosity at p, = 0 [-]
Es,oo equilibrium solidosity [-]
r zeta potenrial [V]
fJ viscosity of filtrate [Pa.s]
fla apparent viscosity of suspension [Pa.s]
'f/p plastic viscosity [Pa.s]
() = contact angle (filter medium/filtrate) [rad]
(), sample temperature [OC]
K. = redprocal double layer thickness [m-1]
Notation 171
À compressibility coefficient [-]
f.'d dynarnic mobility [mz. y-t.s-1]
V kinematic viscosity of the liquid [nt.s-1]
p density of space charge [C.m-3]
Äp density difference between particles and liquid [kg.m 3]
Pd specific electrical resistance of dry cerarnics [O.m}
Pn specific electrical resistance of wetted cerarnics [O.m]
p, sludge density [kg.m-3]
(J Stefan-Boltzmann constant [W.m-2.K4]
(Jo surface charge [C]
(Jb charge between surface and slipping plane [C]
(Jd charge of diffuse double layer [C]
(Jek charge at slipping plane [C]
(J, charge in Stem layer [C]
7 shear stress [Pa]
7o yield stress [Pa]
q, superequivalent adsorption potential [V] .p volume fraction of particles [-]
"" electrical potential [V]
% surface potentlal [V]
""~ Stem potentlal [V]
w angular velocity [rad.s 1]
angular frequency [s-t]
LITERATURE
Anderson, R.B., J. Am. Chem. Soc., vol. 68, p 686, 1946.
Anderson, T.N., Bockris, J.O.M., Electrochem. Acta, vol. 9, pp 347-371, 1964.
Barber, J., Veenstra, J., Evaluation of biologica! sludge properties influencing volume
reduction, J. Water Pol/ut. Control Fed., vol. 58, pp 149-156, 1986.
Baskerville, R.C., Gale, R.S., A simple instrument for determiniug the ftltrability of
sewage sludges, J. Water Pollut. Control Fed., vol. 67, pp 233-241, 1968.
Berg, van den C., Vapour sorption equilibria and other water-starch interactions; a
physico-chemical approach, Ph.D. thesis, Agricultnral University Wageningen, 1981.
Berg, van den C., Bruin, S., Water activity and its estimation rn food systerns:
theoretical aspects, Second international symposium on properties of water in relation
to food quality and stability (ISOPOW-11), Osaka, Japan, 1978.
Berlo, van M., Karakterisering van fysische eigenschappen van zuiveringsslib; studie
van parameters die van invloed z~in op de ontwaterbaarheid, ir-thesis in Dutch,
Eindhoven University of Technology, the Netherlands, August 1993.
Black, A.P., Birkner, F.B., Morgan, J.J., Destabilization of dilute clay suspensions
with labelled polymers, J. Am. Water Works Association, vol. 57, pp 1547-1560,
1965.
174 Literature
Boer, de J.H., The dynamical character of absorption, CJarendon Press, Oxford (2nd
edition 1968), 1953.
Boersma, S.L., A theory of differential thermal analysis and new methods of measure
ment and interpretation, J. Am. Ceram. Soc., vol. 38, no. 8, pp 281-284, 1955.
Bruil, H. G., Personal communication, 1993.
Brunauer, S., Emmett, P.H., Teller, E., Absorption of gases in multimolecular layers,
J. Am. Chem. Soc., vol. 60, p 309, 1938.
Campbell, H.W., Crescuolo, P.J., The use of rheology for sludge characterization,
Water Science Technology, vol. 14, pp 475-489, 1982.
Campbell, H.W., Crescuolo, P.J., Control of polymer addition for sludge conditio
ning: a demonstration study, Water Science Technology, vol. 21, pp 1309-1317, 1989.
C~S, Zuivering van afvalwater, deel b, 1991, 's Gravenhage, 1993.
Coumans, W.J., Kerkhof, P.J.A.M., research proposal in Dutch 'Fundamenteel
onderzoek slibontwatering', 1989.
Debye, P., A metbod for the determination of the mass of electrolytic ions, J. Chem.
Phys., vol. 1, no. 1, pp 13-16, 1933.
Literature 175
Dohanyos, M., Grau, P., Sedlacek, M., lnterpretation of dewaterability measurements
by capillary suction time (CST), Water Science Technology, vol. 20, no. 1, pp 265-
267, 1988.
Dijck, van F.W.H.M., Coumans, W.J., Kerkhof, P.J.A.M., Boulan, R.P., Breemen,
van A.N., Kop, J.H., Slibontwatering, een voorstudie, TU Eindhoven, TU Delft (in
Dutch), RWZI 2000 89-01.
Dijke, van D., Ontwerp en ontwikkeling van een meetopstelling ter bepaling van
waterdampsorptie-isothermen, ir-thesis in Dntch, Eindhoven University of Techno
logy, the Netherlands, June 1992.
Eikelboom, D.H., Stouthamer, A.H., Verseveld, van H.B., Muller, E.B., High
performance bioreactor: a physiological approach to wastewater treatment with zero
sludge production by complete sludge retention, Proc. of the secoud Japanese-Dntch
Workshop on the treatment of municipal waste water, Miyazaki, Japan, 1993.
Eikelboom, D.H., Personal communication, 1994.
Eikum, A.S., Paulsrud, B., Piltration properties of aerobic stabilized primary and
mixed primary-chernical (alum) sludge, Water Research, vol. 8, pp 203-209, 1974.
François, R.J., Haute, van A.A., Structure of hydroxide flocs, Water Research, vol.
19, pp 1249-1254, 1985.
Gál, S., Die Methodik der Wasserdampf-Sorptionsmessungen (in German), Springer
Verlag, Berlin/New York, 1967.
176 Literature
Görling, P., Untersuchungen zur Aufklärung des Trocknungsverhaltens pflanzicher
Stoffe, insbesoudere von Kartoffelstücken, Ph.D. thesis in German, Technische
Universität Darmstadt, Germany, 1955.
Greenspan, L., Humidity fixed points of binary saturated aqueous solutions, J. Res.
Nat. Bur. Standards, vol. Sla, pp 89-96, 1977.
Gregory, J., Flocculation of polystyrene particles with cationic polyelectrolytes,
Trans. Faraday Soc., vol. 65, pp 2260, 1969.
Gregory, J., Flocculation by inorganic salts, in 'The scientific basis of flocculation',
editor: Ives, K.J., Noordhof, Leiden, 1978.
Guggenheim, E.A., Applications of statistical mechanics, Ciarendon Press, Oxford,
1966.
Halde, RE., Sewage sludge characterization by vacuum drying, Filtration and
Separation, vol. 16, pp 238-242, 1979.
Hanstveit, A.O., Visser, M.A., Eikelboom, D.H., Biologisch-biochemisch onderzoek,
part 4, in 'Onderzoek naar de waterbinding van surplusslib van rioolwaterzuiveringsin
stallaties', report (in Dutch) working group fundamental sludge research, Centraal
Technisch Instituut TNO, 1977.
Heide, B.A., Kampf, R., Selection and use of polyelectrolytes in dewatering of sludge
with belt presses. Proc. of workshop 'Methods of characterization of sewage sludge',
Dublin, 1983.
Literature 177
Hermans, J., Phil. Mag., vol. 25, p 426, 1938.
Herwijn, A.J.M., La Heij, E.J., Janssen, P.M.H., Coumans, W.J., Kerkhof,
P.J.A.M., Fundamentele aspecten van slibontwatering, fina1 report RWZI 2000 94-02
(in Dutch), 1994.
Hiemensz, P.C., Principlesof col1oid and surface chemistry, 2nd ed., Marcel Dekker,
lnc., New Y ork, 1986.
Igarasbi, C., Nishizawa, K., A new technique for measuring the f-potential of organic
fioc fonned by cationic polyelectrolyte addition: description and application, Water
Environment Research, vol. 64, no. 1., pp 44-52, 1992.
Iglesias, H.A., Cbirife, J., Handbook for food isothenns, water sorption parameters
for food components, Academie Press Inc., New York, 1982.
James, R.O., Healy, T.W., Adsorption of hydrolyzable metal ions at the oxide-water
interface.
I Co(II) adsorption on Si02 and Ti02 as model systems, J. Colloid Interface Science,
vol. 40, no. 1, pp 42-52, 1972.
ll Charge reversal of Si02 and Ti02 colloids by adsorbed Co(ll), La(Ill) and Th(IV) as
model systems. J. Colloid Interface Science, vol. 40, no. 1, pp 53-64, 1972.
Ill Thermadynamie model of adorption. J. Colloid Interface Science, vol. 40, no. 1,
pp 65-81, 1972.
Janssen, P.M.H., Herwijn, A.J.M., La Heij, E.J., Coumans, W.J., Kerkhof,
P.J.A.M., Het uittesten van de Filtratie-Expressiecel in de praktijk (in Dutch), report
RWZI 2000 94-07, 1995.
178 Literature
Karr, P.R., Keinath, T.M., Intluence of partiele size on sludge dewaterability, J.
Water Pollut. Control Fed., vol. 50, pp 1911-1930, 1978.
Katsiris, N., Kouzeli-Katsiri, A., Bound water content of biological sludges in relation
to filtration and dewatering, Water Research, vol. 21, no. 11, pp 1319-1327, 1987.
Keey, R.B., Drying: Principles and Practice, Pergamon Press, 1972.
Kerkhof, P.J.A.M., Some fundamental aspects of sludge dewatering, fust Dutch
Japanese workshop on the treatment of municipal waste water, Heelsum, 1991.
Knocke, W.R., Wakeland, D.L., Fundamental characteristics of water treatment plant
sludges, J. Am. Water Works Association, vol. 75, no. 10, pp 516-523, 1983.
Koot, Behandeling van afvalwater (in Dutch), 1974.
La Heij, E.J., Janssen, P.M.H., internal report, Eindhoven University of Technology,
1990.
La Heij, E.J., An analysis of sludge flitration and expression, Ph.D. thesis, Faculty of
Chem. Eng. and Chemistry, Eindhoven University of Technology, 1994.
La Mer, V.K., J. Colloid Science, vol. 19, p 291, 1964.
Langmuir, I., J. Am. Chem. Soc., vol. 40, p 1361, 1918.
Literature 179
Lee, D.J., Hsu, Y.H., Fluid flow in Capillary Suction Apparatus, Ind. Eng. Chem.
Res., vol. 31, no. 10, pp 2379-2385, 1992.
Lee, D.J., Hsu, Y.H., Cake formation in Capillary Suction Apparatus, Ind. Eng.
Chem. Res., vol. 32, no. 6, pp 1180-1185, 1993.
Leu, W.F., Cake ftltration, Ph.D. thesis, Depart. Chem. Eng., University of Houston,
USA, 1981.
Levine, S., Friesen, W.I., Flocculation of colloid particles by water-soluble polymers,
in 'Flocculation in biotechnology and separation systems', editor: Attia, Y.A.,
Elsevier, Amsterdam, 1987.
Lotito, V., Spinosa, L., Sewage sludge conditioning by polyelectrolytes, Filtration
and Separation, vol. 27, no. 11, pp 1319-1327, 1987.
Marlow, B.J., Fairhurst, D., Colloid vibration potenrial and the electrokinetic
characterization of concentrated colloids, Langmuir, vol. 4, no. 3, pp 611-626, 1988.
Marskamp, M., Trends in sludge incineration in the Netherlands, Proceedings of the
second Japanese-Dutch workshop on the treatment of municipal waste water, Miya
zaki, Japan, 1993.
Matec, United StatesPatent # 4,497,207, 1985.
180 Literature
NEN 6620, Bepaling van de indamprest en gloeirest, Dutch Standardization Institute,
1988.
NEN 6690, Bepaling van de capillaire zuigtijd (CST waarde), Dutch Standardization
Institute, 1988.
NEN 6691, Bepaling van de ontwaterbaarbeid met de MFT-methode, Dutch Standar
dization Institute, 1988.
Novak, J.T., O'Brien, J.H., Polymer conditioning of chemical sludges, J. Water
Pollut. Control Fed., vol. 47, no. 10, pp 2397-2410, 1975.
0 'Brien, R. W., Electro-acoustic effects in a dilute suspension of spherical particles,
J. Fluid Mech., vol. 190, pp 71-86.
Parlcs, G.A., Bruyn, de, P.L., The zero point of charge of oxides, J. Phys. Chem.,
vol. 66, pp 967-973, 1962.
Pasveer, A., Eenvoudige afvalwaterzuivering (in Dutch), De Ingenieur, vol. 69, no.
17, pp G1-G6, 1957.
Patterson, J.W., Brezonik, P.L., Putman, H.D., Measurement and significanee of
Adenosine Triphosphate in activated sludge, Environmental Science & Technology,
vol. 4, no. 7, pp 569-575, 1970.
Pidgeon, L.M., Maass, 0., The adsorption of water by wood, J. Am. Chem. Soc.,
vol. 52, pp 1053-1069, 1930.
Literature 181
Roberts, K., Flocculation and dewatering of sludges, in 'Scientific basis of floccula
tion', editor: lves, K.J., Noordhof, Leiden, 1978.
Robins on, J. , Knocke, W. R. , U se of dilatometric and drying techniques for assessing
sludge dewatering characteristics, Water Environment Research, vol. 64, no. 1, pp 60-
68.
Rutgers, A., Physica, vol. 5, p 46, 1938.
Ruth, B.F., Correlating ftltration theory with industrial practice, Ind. Eng. Chem.,
vol. 38, no. 6, pp 564-571, 1946.
Saunders, F.M., Holmes, M., Chemistry of lime conditioning of water treatruent
sludges, in 'Fundamental aspects of sludge characterization and dewatering', editor:
Vesilind, P.A., Engineering Foundation, New York, 1987.
Shaw, D., Electroforesis, Academie Press, 1969.
Singley, J.E., Sullivan, J.H., Reactions of metal ionsin dilute solutions: Recalculation
of hydrolysis of iron (ill) data, J. Am. Water Works Association, voL 61, pp 190-192,
1969.
Smollen, M., Categones of moisture content and dewatering characteristics of
biological sludges. Proc. of the 4th World flitration Congress, Ostend (Belgium),
1986.
182 Literstore
Smollen, M., Evaluation of municipal sludge drying and dewatering with respect to
sludge volume reduction. Proc. IA WPRC Specialty Conference Sludge Management,
Los Angeles, 1990.
Steiner, A.E., Me Laren, D.A., Forster, C.F., The nature of activated sludge flocs,
Water Research, vol. 10, no.1, pp 25-30, 1976.
STORA, part I, Slibontwatering, 1. Aard van de waterbinding in slib (literatuur),
1981.
part ll, Slibeigenschappen en resultaten van slibverwerkingsapparatuur,
1981.
part m, Optimalisering van slibontwatering met polyelectrolyt (zeefband
persen-uitgegist slib), 1982.
part IV, Zeefhandpersen-slibkarakterisering en optimaal gebruik van
polyelectrolyt (handleiding), 1982.
part V, Filterpersen-slibkarakterisering en optimaal gebruik van organische
chemicaliën (handleiding), 1983.
part VI, Filterpersen-slibkarakterisering en optimaal gebruik van organische
chemicaliën (onderzoek), 1983.
STORA, Rijswijk, the Netherlands.
Svarovsky, L., Solid-liquid separation, third edition, Butterworths, London, 1990.
Tiller, F .M., Compressible ~e filtration, in 'The scientific basis of filtration', editor:
lves, K.J., Noordhof, Leiden, 1975.
Tiller, F.M., Crump, J.R., Ville, F., A revised approach to the theory of cake
filtration, Proc. Int. Symp. on Fine Partiele Processing, editor: Somassundarem P.,
Am. inst. mining, metallurgical and petto. eng., pp 1550-1582, 1980.
Literature 183
Tiller, F.M., Shen, Y.L., Adin, A., Capillary suction theory for rectangular cells, J.
Wat. Pollut. Control Fed., vol. 62, no. 2, pp 130-136, 1990.
Unno H., Muraiso, H., Akehata, T., Theoretica! and experimental study of factors
affecting Capillary Suction Time, Water Research, vol. 17, no. 2, pp 149-156, 1983.
Veldhuizen, van, A.J. W., Compressiegedrag van zuiveringsslib, ir-thesis in Dutch,
Eindhoven University of Technology, the Netherlands, 1991.
Verwey, E.J. W., Overbeek, J. T. G., Theory of the stability of lyophobic colloids,
Elsevier, Amsterdam, 1983.
Vesilind, P.A., Treatment and disposal of wastewater sludges, Ann Arbor Science
Publishers lnc., Ann Arbor, 1974.
Vesilind, P.A., Test on sludges: Status of standard test procedures for characterizing
sludge, in 'Fundamental aspects of sludge characterization and dewatering', editor:
Vesilind, P.A., Engineering Foundation, New York, 1987.
Vesilind, P.A., Capillary suction time as a fundamental measure of sludge dewaterabi
lity, J. Wat. Pollut. Control Fed., vol. 60, no. 2, pp 215-220.
V esilind, P.A. , Da vis, H.A. , U sing the capillary suction time device for characteri
zing sludge dewaterability, Water Science Technology, vol. 20, no. 1, pp 203-205.
184 Literature
Wakeman, R.J., Tarleton, E.S., Sabri, M.N., Effects of the solid-liquid interface in
filter cake fonnation and consolidation, Proc. European symposium on solid-Iiquid
separation, 474th eventof EFCE, GVC, Cologne, 1992.
W elzen, van J. , Effects of polycations on structure and catalytic activity of cobalt
phthalocyanine: mechanistic study of the ionene-promoted thoil autoxidation, Ph.D.
thesis, Faculty of Chem. Eng. and Chemistry, Eindhoven University of Technology,
1989.
Whitaker, S., Simultaneons heat, mass and momenturn transfer in porons media: a
theory of drying, Adv. heat mass transfer, vol. 13, no. 1, pp 110-203, 1977.
Wolf, W., Spiess, W.E.L., Jung, G., Standardization of isothermmeasurements in:
Properties of water in foods, editors: Simatos, D., Multon, J.L., Nato Asi Series,
Martinus Nijhoff Publishers, Dordrecht, pp 661-677, 1985.
Yeh, S., Cake deliquoring and radial filtration, Ph.D. thesis, Depart. Chem. Eng.,
University of Houston, 1985.
Uzermans, J.J., Capillary suction time: theorie en praktijk, ir-thesis in Dutch,
Eindhoven University of Technology, the Netherlands, 1992.
Zande, van der A., Karakterisering van de slib-water binding met behulp van
thermische analyse technieken en sorptie-isothermen, ir-thesis in dutch, Eindhoven
University of Technology, the Netherlands, 1992.
::sl IJ.CI
> .... ~ (")
~ "" "" g.
~ ~ s. "" ~ §=
~ ~ ~
~ ~ ... ;;-; ;:,
I l
,/·
bo ~ !ooo6
INGANG·-·-·-·-·-·-· ·-·--·-·-
i I ! ~ 9 i s. S' 1'1)
l 1'1)
~ a ~
[ "" = ~ ~ i 1'1) a "=
I
~ ~ ~ ....
186 Appendix 1
"chenu.f::al
tond<:tám<'hf -
e~ ta.w..t.,. r'hovt9>
/k:e.; - s-che~ne o/ ihe rÛ<1"" P.~a/me-;/ ;0/anl
1 m(er-6.
Fig. A2 Process scheme of the Mierlo sludge treatment plant.
Appendix 1 187
JJJ
Fig. A3 Map of the Amsterdam-Oost waste water treatment plant.
1 Ontvangput 2 Roosterinstallat1a 3 Hoo!dgemaal-Bedrijtsgebouv 4 Zandvanger 5 Mangcontaottank 6 Oxydatietank 7 Habednktank
-
8-8a Retour- en surplusslibgemaal BY. Bio-filter 9 Ringaloot
10 ludikkèr r 11 Sliblagunea ( 6+ ""- 1,~dt.-l v.tRa.J;c jwec&c..,/ J} 12 Slibmengtank I 1' Slibverwerking 14 Laadplaats slihcontainers
RIJ.z.T
Appendix 1 189
\ I ,
..... i;l ...
::::: i <->
"" ç=> ...., ;;1!
"' ' "" !~ i 'o i ,o
'""
Fig. A5 Process scheme of the oxidation ditch system De Hooge en Lage Zwaluwe.
~ ~ "'cl d <')
~ "" g.
~ ~ s. "'
I <::>' ;::
ti: g.
~ ~ ::tl
<.::::: ~
!lli.lL l=~====--AANVOER' VREEM05tl6~
HOOGHEEMRAADSCHAP WES]'_:!\:~~NT ~-~!::!· ,~··!J .• §: .. JBo.•!:!-:J·~-"!:!-"";:?CC:==l hslMUil Projwtl_..,
RWZI RIJEN
411f((ltHl ..... ..,..,...., .. , ludtfAO fllWtt n trltkitll: tn-•HMt "*~'
:i
(AfiROUSfl
~
1 ~ ....
1?1.150
(ffLIJtiH
....
.....CAO td .... n ""*--dil ·-- •l!ltrM do ~•t.
APPENDIX2
Output of MAPLE® program
Determination of the analytica! expression for the differential enthalpy of wetting i\Hw.
cg: Cg; c:=Cg,o; e: = f:J; f. = E1 ;
kO:=ko; m: = u1 ;
r: R; y: 1/T;
cg:= c*exp(f*y/r) ;
k: kO*exp(e*y!r) ;
a : ln (aw);
a:= ln((cgA(0.5)*(cg*uA2-2*m*n*(cg-2) +mA2*cgn0.5) +n*(cg-2)-m*cg)/ ((2*u*k)*(cg-1 )));
a:= ln(l/2 (
.5 f y .5 I f y 2 I f y \ 2 f y \.5 c exp(---) Ie exp(---) u 2 m u I c exp(---)- 21 + m c exp(---) I
r \ \ r I rl
I fy \ fy I ey I fy \ + n I c exp(---) 21 - m c exp(---)) I (n kO exp(---) I c exp(---) 11))
\ I rl r\ r I
b : (à In <Iw )1(8 1/T) AHw /R
b: =diff(a,y);
.5 fy.5 .5 c exp(---) %2 f
r b 2 ( 1/2 (. 5 ------------------------
r
192 Appendix2
I fy 2 fy 2 fy \ Ie fexp(---) u m u e fexp(---) m e fexp(---) I
.5 f y .5 I r r r I c exp( --) 1-------- -2 ---------- + -------------1
r \ r r r I + .5 -----------------------------------------------------------
fy ue fexp(-)
fy me f exp(---)
.5 %2
r r I ey I fy \ + --------- - --------)I (u kQ exp(-) Ie exp(---)- 1 D
r r I r\ r/
%3e - 112 ---------------------
ey I fy \ u kQ exp(-) Ie exp(---)- ll r
r \ r I
fy %3 e fexp(-)
r ey I fy \ -1/2 -----------------------)u kQ exp(-) Ie exp(--)- ll/%3
ey I fy \2 r \ I ukQexp(--) leexp(-) -11 r
r \ r I
fy %1:= qexp(---)-2
r
fy 2 2 fy %2:= cexp(--)u 2mu%l+meexp(--)
r
.5 fy .5 .5 fy %3:= c exp(--) %2 +u%1-mcexp(--)
r r
APPENDIX 3
Working scheme of sludge characterization
1. Sludge dewatering properties
Used flocculants: ferric chloride/lime, polyelectrolyte KF 975 and the polyelectrolyte
used at the sludge dewatering plant.
First day
Coneetion of sewage sludge sample at the sludge treatment plant
Measurements:
dry solicts content
- ash content
- loss of ignition
- ATP content
CST </>10
(unconditioned sludge)
(unconditioned sludge)
(unconditioned sludge)
(unconditioned sludge)
(unconditioned sludge)
44 MFT experiments to determine optimal flocculation conditions; 16 MFT experi
ments per combination sludge-polyelectrolyte and 12 experiments for the conditioning
with ferric chloride.
Secoud day
ATP content
- CST </>10
- MFT&CST </>10
- constant pressure filtration
pH
electrical conductivity
Third day
- A TP content
CST </>10
partiele size distribution
rheological properties
Fourth day
(unconditioned sludge)
(unconditioned sludge)
( conditioned sludge)
( conditioned sludge)
(conditioned sludge)
(conditioned sludge)
(unconditioned sludge)
(unconditioned sludge)
( conditioned and unconditioned sludge)
(conditioned and unconditioned sludge)
- ferric ion concentration in filtrate (conditioned sludge)
polyelectrolyte concentration in filtrate (conditioned sludge)
194 Appendix 3
2. Sludge solid-to-water bond strengtb properties
First day
- Constant pressure :filtration; use of effective fungicide
- Water vapour sorption isotherms
Secoud day
-Constant pressure :filtration
- TGA/DTA: isothermal drying curves (3x)
Tbird day
- TGA/DTA: isothermal drying curves (3x)
Fourth day
TGA/DTA: isothermal drying curves (3x)
APPENDIX4
Shift of the absorption maximum of cobaltphthalocyanine due to increasing added
amounts of polyelectrolyte.
In the eight figures presented, the absorbance of a 212 mg/1 CoPc(NaS03) 4 solution is
given as a function of the wavelengtb. A solution of 100 mg KF975/l was stepwise
added to the cobaltphthalocyanine solution. The absorption maximum shifts from 662
nm to 628 nm. A volume of 56 J.Ll polyelectrolyte solution is needed to shift the
absorption maximum. A further increase of the added amount of polyelectrolyte does
not shift the absorption maximum .
... .. t: ~~
,IS
i .I
.os
VAVElEUGTH CrwJ)
... :1: llo;J
.12
.I
I ·"" ... ... ·""
IIAVEt EIIGlH in.)
196 Appendix 4
... 3: s""/''(
.!Z
..
I .oa
...
...
... 't: n_rJ
.I
...
I .... ...
• !l ll ~ il ~ IL D i I lil f!
WAVELEIIGlU lmt)
.u --------r~-----r-
.12 >:>.z~
..
I ... • 1111
••• • oz
0
~ H ~ M M a D i I i I! V4VElEHiïnt (na)
Appendix 4 197
... .12
t; s~r
.I
I .OB
·"" .o•
.oz
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ WAYELEJICHI (rw)
·" 7' '~
.I
.DB
I ,06 .
. o•
.oz
~ ~ ~ ~ ~ ~ ~ ~ ~ Ij ~ WAVEtEilGlit (n~~)
·" --------.----,.-
8: fPtJ/.J
.I
.OB
i • DO
.Dl
.oz
0
!l---~---t---~- --~--~ ~ ~ ~ i ~
VN.IElfUGTii (o•)
CURRICULUM VITAE
De auteur werd geboren op 19 februari 1958 te Eindhoven. In 1976 behaalde hij het
Atheneum-B diploma aan het Bisschop Bekkers College te Eindhoven. Aansluitend
studeerde hij Technische Natuurkunde aan de Technische Universiteit Eindhoven. Het
afstudeerwerk werd verricht bij Prof.dr.ir. G. Vossers in de vakgroep transportfysica.
In 1984 behaalde hij zijn doctoraal examen. Van 1985 tot medio 1987 was hij
werkzaam als procestechnoloog bij Vitrite te Middelburg. De Vitrite fabriek vormt
onderdeel van Philips Ligbting. Gedurende de periode van medio 1987 tot 1990 was
hij als procesontwikkelaar in dienst van Silenka te Hoogezand. In de periode 1990 -
1994 was hij werkzaam als senior-onderzoeker in het Laboratorium voor Scheidings
technologie, vakgroep Chemische Proceskunde van de Technische Universiteit
Eindhoven. Het in dit proefschrift beschreven onderzoek werd uitgevoerd in de groep
van Prof.dr.ir. P.J.A.M. Kerkhof. Vanaf januari 1995 is hij vennoot van het bedrijf
Herwijn & Janssen sludge technology vof en sedert februari 1996 directeur van het
bedrijf Herwijn & Janssen dewatering bv i.o. Het laastgenoemde bedrijf is een
vennootschap onder firma aangegaan met Witteveen +Bos Raadgevende ingenieurs bv
te Deventer. Deze vennootschap draagt de naam Sludge Consultants.
Stellingen
behorende bij het proefschrift van A.J.M. Herwijn
1. Het fysisch/chemisch gebonden watergehalte in slib draagt slechts in geringe
mate bij tot de vochtretentie bij het mechanisch ontwateringsproces.
dit proefschrift, hoofdstuk 3
2. De compressie-permeabiliteitscel kan ook gebruikt worden voor de bepaling van
de hoeveelheid gebonden water in slibkoeken. dit proefschrift, hoofdstuk 4
3. De resultaten van een "Capillary Suction Time" test worden niet alleen bepaald
door de ontwateringseigenschappen van de slibkoek, maar ook door de structuur
eigenschappen van het filtreerpapier, waardoor meetwaarden verkeerd kmmen
worden geïnterpreteerd. dit proefschrift, hoofdstuk 5
4. In de literatuur wordt ten onrechte aangenomen dat de diffusiecoëfficiënt van
vochttransport in poreuze materialen tijdens een droogproces een eenduidige
functie is van het vochtgehalte.
IsotheJJllal vapour and liquid transport inside clay during drying, Zanden, van der, A.J.J. et al.,
accepted for pubHeation in Drying Technology, vol. 14, no. 3, 1996.
5. De oorzaak van een slechte slibontwatering in de praktijk wordt vaak ten
onrechte toegewezen aan het purificatieproces.
6. Omdat de chemische industrie Nederland dreigt te verlaten, dienen de hoorcolle
ges voor het scheikunde-curriculum in de Engelse taal te worden gegeven.
7. De kostenbesparing die bedrijven kmmen boeken door reductie van de slibstroom
wordt vaak onderschat of in het geheel niet onderkend.
8. Door het uitschrijven van een toenemend· aantal prijsvragen en competities voor
afstudeerscripties of andere wetenschappelijke resultaten, worden studeren en
wetenschap bedrijven steeds meer een topsport.
9. De steeds verdergaande commercialisering van de Nederlandse televisie pleit
voor afschaffing van het luister- en kijkgeld.
10. Loonmatiging is slechts gericht op behoud van bestaande banen en niet op het
creëren van nieuwe banen zodat het gangbare standpunt dat loonmatiging goed is
voor de werkgelegenheid in twijfel kan worden getrokken.
11. Het doel om door een nieuw aan te leggen woon-werk-recreatiegebied in Oost
Groningen genaamd "Blauwe Stad" het imago en de leefbaarheid van dit deel van
Nederland te verbeteren, wordt door de naamgeving enigszins voorbij gestreefd.
12. Om de tijd, die verloren gegaan is met het gekrakeel rond de aanleg van de
tracés van de hogesnelheidslijn, weer in te halen moet de TGV wel heel hard
kunnen rijden.