Abstract
7
accumulating organisms (PAO) was not observed and phosphorous content in biomass reached
a maximal concentration of 2.8%. This value is probably too low to claim the presence of PAO in
activated sludge.
A provisional simplified mathematical model for the simulation of the nutrients (COD and
nitrogen) removal processes has been developed. It satisfactory simulated the experimental data
(real process). This first form will be developed with more detailed differential equations in
further studies.
Removal of total coliforms was very effective: a 4.7-5.1 log reduction was observed. Escherichia
coli was never found in the permeate. The variation of biomass concentration in the bioreactor
did not affected the removal efficiency of pathogens
The values of the kinetic constants for the MBR process were experimentally determined based
on data collected during the five (pseudo) steady-state conditions reached in the five periods,
from 1 to 5. The biomass yield, Y, and the decay coefficient kd, were found to be equal to 0.24
gMLVSS/gCODremoved and 0.01 day
-1
, respectively. These values are slightly lower than those
generally found in conventional activated sludge. Yobs value showed to decrease exponentially
with increasing SRTs.
The application of high SRTs determined a decrease in the volatile matter (bacteria) content in
activated sludge. The MLVSS/MLSS ratio passed from 80% to 53% and the biomass activity
accordingly decreased: the ammonium utilisation rate doubled when passing from some 1000
down to 200 days or less of SRT. This means that working with high biomass concentrations in
the bioreactor, let say over 15 g/l, determines the oxidation of sludge in the bioreactor and
therefore useless expenses for aeration. Moreover, a decrease in biomass viability was observed.
Therefore, it is more convenient to operate in an interval of biomass concentration of some 8-12
g/l. This determines also other advantages in treatment process since this biomass concentration
makes it much easier to keep the membranes clean, improve the oxygen transfer efficiency and
mix and handles the activated sludge.
The removal efficiencies and the fate of both metals and organic compounds in the MBR and in
a conventional activated sludge process were studied. The removal efficiency from treated
wastewaters for the MBR system was generally high for metals (> 75%). This efficiency was
generally 10-15% higher than that of a conventional activated sludge process. The increase in
metals removal efficiency was due to the capability of the membrane to perfectly retain the
suspended solids from the bioreactor effluent. These results were nearly equivalent when
considering periods at 9 or 16 – 18 g/l of biomass in the MBR, so it seems sensible to operate at
the lower concentration of MLSS to reduce the treatment process expenses (aeration, mixing,
and stabilisation). The MBR allows improving the exploiting of the bio-sorption capability of
the activated sludge for metals (especially heavy metals) and can be considered a suitable
technology for the removal of this class of pollutants from wastewaters.
According to the removal efficiencies found in the MBR experimentation, three different groups
of metals can be individuated: the easily removable, the metals which showed removal yields >
75%, like Al, Ag, Ba, Cd, Cr, Cu, Fe, Hg, Sn, the metals only partially removable, with removal
yields in the range 40 - 60%, like Co, Mn, Ni, Pb, V and Zn and the metals hard removable, like,
As, B and Se, which showed removal yields < 25%. Arsenic in particular showed the worst
removal efficiency and was the main problem encountered in the MBR experimentation since it
exceeded the discharge limits imposed for very sensitive areas like the lagoon of Venice.
The metals concentrations in sludge were equal or higher in the MBR experimentation at 9 g/l of
biomass compared to the conventional process (CASP), but those concentrations decreased
increasing the solid retention time (SRT) of the MBR system, probably because a steady state
condition was not really reached during the experimental periods 2, 3 and 4, when very high
SRTs were applied to the MBR (i.e., up to 1000 days).
Nutrients and micropollutants removal by a membrane bioreactor (MBR) and activated carbons
8
The removal of organic substances was very effective for both the conventional and the MBR
process, generally higher than 90%. However, the MBR, thanks to the membrane separation
system, showed the capability of better removing some classes of compounds, like non-ionic
surfactants (BIAS) and some chlorinated priority pollutants: hexachlorobenzene and
polychlorinated biphenyls (PCB). Some priority pollutants, like the organic aromatic solvents,
the polycyclic aromatic hydrocarbons (PAH), dioxin, and polychlorinated biphenyls (PCB)
showed efficiency removal near or equal to 100% in the MBR process. Only the nitrogen
herbicides, probably because of their chemical structure, showed removal efficiencies in the
range 20% - 60%.
The ultrafiltration module performances have been studies with particular concern for flux and
transmembrane pressure variation with varying biomass concentration within the reactor. The
resistance-in series- model has been applied for resistances evaluation to investigate the
importance of fouling and concentration polarization on the MBR performances.
Sommario
9
SOMMARIO
Lo scopo del lavoro sperimentale condotto nel corso dei tre anni di dottorato di ricerca è stato
quello di verificare l’affidabilità e le rese di un processo biotecnologico avanzato per il
trattamento delle acque reflue: il bioreattore a membrana (MBR). La necessità di imporre
standard di qualità di alto livello per le acque ed il possibile riutilizzo delle acque di scarico
hanno determinato la necessità di adottare processi di trattamento di tipo avanzato per le acque
reflue. Tra le migliori tecnologie in campo biotecnologico che permettano il raggiungimento di
standard così elevati, emerge il bioreattore a membrana. Al fine di ottenere maggiori
informazioni in merito all’applicazione dei bioreattori a membrana soprattutto nell’ottica di un
utilizzo in piena scala, a partire dal 1999 è stato condotto uno studio a scala pilota tramite un
bioreattore a membrana con modulo filtrante da ultrafiltrazione (dimensione nominale dei pori
0.02 Πm). I principali obbiettivi della ricerca sono stati la determinazione delle efficienze e lo
studio dei meccanismi alla base di esse per la rimozione dei nutrienti (C, N, P) e dei
microinquinanti, oltre alla determinazione delle condizioni operative ottimali per il processo a
fanghi attivi MBR. La sperimentazione è stata organizzata in cinque periodi sperimentali
durante i quali la concentrazione di biomassa in reattore è stata incrementata da 3.7 g/l (Run 1),
a 9.2 g/l (Run 2), a 16.7 g/l (Run 3), a 18.2 g/l (Run 4) e successivamente diminuita a 10 g/l (Run
5). Durante il Periodo 4 si è proceduto all’aggiunta, direttamente in reattore, di carbone attivo
granulare (GAC) in ragione dell’1% in peso secco per valutarne gli effetti sulla rimozione dei
microinquinanti.
L’applicazione della tecnologia a membrana ha portato ad ottimi risultati in termini di
rimozione dei macroinquinanti (solidi sospesi e nutrienti). In particolare, la rimozione dei solidi
sospesi ha portato ad un effluente completamente privo di essi. Questo aspetto è di
fondamentale importanza quando si considera che a questa frazione sono legati la maggior
parte di macro e microinquinanti che possono pregiudicare il riutilizzo dell’acqua trattata.
La rimozione del COD è stata sempre ottima, generalmente nell’intervallo 84-94% e le rese
dell’MBR sono state sempre maggiori del 30% rispetto ad un processo convenzionale. Ancora
una volta questo risultato è legato alla ritenzione delle macromolecole organiche da parte del
modulo da ultrafiltrazione. La respirazione, ovvero la conversine delle sostanze organiche in
CO2, è stato il principale meccanismo di rimozione del COD quando la biomassa in reattore era
in concentrazione uguale o superiore a 9 g/l.
L’azoto è stato rimosso in modo soddisfacente, con rese nell’intervallo 61-90%, e con una
concentrazione di azoto totale nell’effluente finale tra 4.5 e 11 mgN/l. La denitrificazione ha
rappresentato la principale via di rimozione dell’azoto in corrispondenza di elevate
concentrazioni di biomassa in reattore, quando la resa osservata di crescita era estremamente
limitata (0.1 gMLVSS/gCODrimosso o meno). Inoltre, la concentrazione di azoto ammoniacale
nell’effluente è sempre stata al di sotto di 1 mgN/l. Questo risultato è dovuto alla perfetta
ritenzione della biomassa nitrificante all’interno del reattore che ha comportato un
miglioramento nel processo di nitrificazione biologica. Il tasso di utilizzo dell’azoto
ammoniacale (AUR) è stato di 2.5-3 mgN/gVSS h, valori che corrispondono ai più alti valori
riportati in letteratura per processi convenzionali a fanghi attivi.
La rimozione del fosforo è variata tra efficienze del 73 e del 77%, grazie all’azione contenitiva
del modulo da ultrafiltrazione. Non si sono osservati fenomeni di “luxury uptake” da parte dei
Nutrients and micropollutants removal by a membrane bioreactor (MBR) and activated carbons
10
microrganismi fosforo-accumulanti (PAO) dal momento che il contenuto di fosforo nel fango
non ha superato il 2.8% (su base TS).
E’ stato sviluppato un modello semplificato e provvisorio per la simulazione dei processi di
rimozione di azoto e fosforo che ha permesso una buona simulazione dei dati sperimentali.
Nonostante ciò, sono necessari ulteriori sviluppi del modello, soprattutto in termini di
differenziazione delle costanti cinetiche.
La rimozione dei coliformi totali è stata molto efficace (riduzione di 4.7-5.1 log) mentre la E. coli
non è mai stata trovata nel permeato. La variazione di concentrazione in reattore non ha influito
sull’efficienza di rimozione dei patogeni.
Sono stati determinati sperimentalmente I valori delle costanti cinetiche per il processo MBR. La
resa di crescita della biomassa Y è risultata pari a 0.24 gMLVSS/gCODrimosso mentre il coefficiente
di decadimento endogeno è risultato pari a 0.01 day
-1
. IL valore della crescita osservata ha
mostrato di diminuire con l’aumento dell’SRT.
L’applicazione di elevate età del fango ha determinato una diminuzione nel contenuto di
sostanza volatile in reattore (biomassa). Il rapporto MLVSS/MLSS è passato dall’80% al 53%,
come anche l’attività della biomassa. Ciò significa che operare ad età del fango così elevate
(corrispondenti a MLSS > 15 g/l) comporta solo inutili costi per l’aerazione senza portare ad
effettivi benefici.
Sono stati studiate le efficienze di rimozione ed il destino finale sia di metalli che di
microinquinanti organici da parte del processo MBR, in confronto anche con un sistema
convenzionale a fanghi attivi. L’efficienza di rimozione dei metalli per l’MBR è stata in generale
superiore al 75%, valore a sua volta superiore del 10-15% rispetto al processo convenzionale.
Questa discrepanza è dovuta essenzialmente alla capacità della membrana di trattenere i solidi
sospesi ai quali i metalli sono legati. I risultati sono stati ottimi anche operando a valori di
concentrazione di biomassa inferiori (9 g/l).
Sulla base delle efficienze di rimozione si possono individuare tre gruppi di metalli: un primo
gruppo, facilmente rimovibile (efficienza > 75%), al quale appartengono Al, Ag, Ba, Cd, Cr, Cu,
Fe, Hg, Sn, un secondo gruppo a media efficienza di rimozione (40 - 60%), al quale
appartengono Co, Mn, Ni, Pb, V and Zn ed un terzo gruppo di metalli difficili da rimuovere
(efficienza < 25%) quali As, B e Se. L’arsenico ha rappresentato il problema principale dal
momento che la sua concentrazione nell’effluente era sempre superiore a quella permessa
attualmente in laguna di Venezia.
La rimozione delle sostanze organiche è stata elevata sia nel caso dell’MBR che del processo
convenzionale sebbene l’MBR abbia mostrato efficienza migliore per alcuni composti quali
tensioattivi non ionici (BIAS), esaclorobenzene e PCB.
Le rese del modulo da ultrafiltrazione sono state studiate ponendo particolare attenzione alle
variazioni di flusso e pressione transmembrana rispetto alla concentrazione di biomassa in
reattore. E’ stato applicato il modello delle resistenze in serie per la valutazione del valore della
resistenza offerta alla filtrazione, al fine di verificare la relativa importanza dei fenomeni di
fouling e di polarizzazione di concentrazione.
Introduction
11
INTRODUCTION
With the increased worldwide pressure on water resources, effluent recycle and reuse are
developing for irrigation and agriculture as well as for indirect and even direct potable water
supply. The interest on wastewater reclamation and reuse has increased considerably over the
past several years, especially for arid countries even though its implementation is still limited.
The obvious need for increased reuse of water significantly impacts both source-water quality
and product water value. Water reuse can be of two primary forms: the first of these, cascading
reuse, is a form practiced to one degree or another since humankind first used water. The
practice involves sequential use of water of deteriorating quality for purposes having
decreasingly stringent requirements. Fundamentally a product water having the most stringent
use standards command the highest quality source water available. After its designated use it
becomes the source for product water having the next most stringent requirement, etc. in many
applications, particularly industry and agriculture, such cascading reuse may require only
minimal intermediate treatments and the lowest quality water is then disposed of as an
unusable waste.
The second type of water reuse involves repeated recycling for the same use. The ultimate
configuration of a recycling scheme might be envisaged as closing the loop on the cascading
reuse scheme; i.e. the lowest quality product becomes the source for the use having the highest
quality requirement. The recycling reuse scenario is relatively new compared to simple
cascading reuse with ultimate wastage of the lowest quality water. It is the scenario that
presents the greatest technology challenge and it is the scenario that is rapidly becoming most
prevalent as the sizes and per capita water demands of human populations escalate
simultaneously in the face of a fixed global water resource.
In this sphere, water quality standards are becoming increasingly stringent and place an
emphasis on turbidity, organic content and micropollutants. A high degree of tertiary treatment
is needed to satisfy the constraints of receiving streams and water quality standards for reuse.
In both cases, maximum removal of suspended solids and colloids appears to be a necessary
step. Deep bed filtration, which is the conventional answer to this problem, has some obvious
limitations. A sand filter cannot be an absolute filter: the residual turbidity is related to
operating conditions but it would be quite impossible to ensure continuous operation at very
low turbidity levels. In the relation between efficiency and particle size, a minimum efficiency
appears for a size in the micrometer range which corresponds to the size of microorganisms.
These points provide an explanation for the recent development of membrane processes in the
field of water and wastewater treatment.
The combination of membrane separation technology and bioreactors has led to a new focus on
wastewater treatment. The application of membranes contributes to very compact wastewater
treatment systems with an excellent effluent quality. Till now, membrane bioreactors have been
applied full scale mainly on fairly concentrated wastewater streams due to the relatively
expensive method of membrane separation and a wide spread application of membrane
bioreactors has been hindered due to the relatively high costs for membrane separation. At the
moment, new low energy types of membrane separation, makes energy requirements
comparable to conventional wastewater systems. This opens possibilities for far going reuse of
wastewater, both industrial and municipal, decrease in sludge production and small foot print
bioreactors for less concentrated wastewater streams.
Nutrients and micropollutants removal by a membrane bioreactor (MBR) and activated carbons
12
OBJECT AND STRUCTURE OF THE THESIS
Nutrients biological removal from wastewater is a well established process and it is generally
agreed that the availability of organic biodegradable matter, process configuration and
activated sludge characteristics determine predominantly the efficiency and reliability of the
process. Nevertheless, regarding the production of excess sludge and the removal of particular
types of pollutants conventional activated sludge systems still need to be improved.
The aim of the experimental work carried out during the three-years PhD was to verify the
reliability of and advanced biotechnological process for wastewater treatment: the membrane
biological reactor (MBR).
Besides typical aspects of wastewater treatment such as suspended solids, nutrients and
pathogens removal, one of the main goals of this thesis was to study and validate a process to
reduce excess sludge production and to uncouple the hydraulic characteristics of the system
(i.e. hydraulic retention time) from the biological ones (i.e. mean cell residence time) in order to
attain a better control of the process, avoiding the typical problems of conventional activated
sludge systems linked to the sludge final sedimentation. Furthermore, the removal of peculiar
micropollutants outfits (organics and metals) was faced with particular concern to the process
modelling.
According to these general themes, every chapter of this thesis is a self-sustaining section.
In Chapter I membranes and membrane bioreactors are described. A first part is dedicated to
the membrane itself and to the description of the theoretical aspects linked to its configuration,
fabrics and use. A second part is then dedicated to the state of the art in the field of membrane
bioreactors applied to wastewater treatment.
Chapter II is about the materials used and the methods applied for the experimentation. The
pilot plant is described, together with the explanation of the research structure.
In Chapter III the theme of nutrients biological removal from wastewater is faced with single
chapters dedicated to carbon, nitrogen and phosphorus. In each section the fundamentals of the
process are reported and then discussed on the basis of experimental data. A provisional
simplified model for carbon and nitrogen removal has been outlined. Conclusions are then
drawn for each component analysed.
Chapter IV is about organic micropollutants removal from wastewater. The experimental data
are discussed and compared with literature and with a conventional wastewater treatment
process with particular concern for the influence of sludge age on removal efficiencies.
Chapter V regards metals removal from wastewater. Particular attention has been given to the
mechanism of biosorption and to the effect of sludge age on metals uptake by activated sludge,
both from the MBR and from a conventional activated sludge system.
In Chapter VI pathogens removal is described. Removal efficiencies are discussed for total
coliforms and Escherichia coli.
Chapter VII is dedicated to sludge production. In this section the relationship existing between
excess biomass production and SRT has been investigated together with the determination of
sludge growth and typical kinetic parameters.
Chapter VIII describes the ultrafiltration module performances with particular concern for flux
and suction pressure. The effect of biomass concentration on the filtration module has been
investigated and the resistance-in-series model has been applied to determine the relative
Object and structure of the thesis
13
importance of fouling concentration polarization on resistance to filtration. A final part is then
dedicated to the description of some methods to reduce fouling.
Chapter IX is conclusive to the whole experimental work and summarizes the main evidences
that occurred during the discussion.
Nutrients and micropollutants removal by a membrane bioreactor (MBR) and activated carbons
14
CHAPTER 1
MEMBRANES AND MEMBRANE BIOREACTORS
1.1 Introduction
Conventional aerobic treatment methods, such as activated sludge processes, are generally used
to treat wastewater flows containing oxidizable substances (e.g. organics, reduced nitrogen). A
drawback of the conventional activated sludge process is that sludge concentration limits
volumetric load, with a sludge concentration of 5-8 kg/m
3
being the maximum (Stephenson et
al., 2000). In spite of the fact that wastewater flow is often low, relatively large specific volumes
in conventional plant design are required. In some cases the space available is not sufficient,
such as for specific industrial wastewaters. In these cases compact biological systems are often
applicable. The main biological compact system is the membrane bioreactor.
In this chapter the main features of membranes and membranes bioreactors for wastewater
treatment are described.
1.2 The membrane
A separation membrane is a thin barrier layer through which fluids or solutes are selectively
transported under the influence of a pressure gradient, a chemical concentration gradient or a
difference in electric potential called the driving force. Separation of a mixture occurs if there is
a significant difference in the transport coefficients through the membrane for the components
of the mixture. Membranes can be classified by:
Nature of the membrane (i.e. natural versus synthetic);
Structure of the membrane (i.e. porous versus non porous, morphological characteristics);
Application of the membrane (e.g. gaseous phase separation, gas-liquid, liquid-liquid, etc.);
Mechanism of membrane action (i.e. adsorptive versus diffusive, ion exchange, osmotic or
non-selective/inert membranes.
1.2.1 Materials
The principal objective in membrane manufacture is to produce a material of reasonable
mechanical strength and which can maintain a high throughput of a desired permeate with a
high degree of selectivity. These last two parameters are mutually counteractive since a high
degree of selectivity s normally only achievable using a membrane having small pores and thus
an inherently high hydraulic resistance (or low permeability).
The permeability increases with increasing density of pores, implying that a high material
porosity is desirable. The overall membrane resistance is directly proportional to its thickness.
Finally, selectivity will be compromised by a broad pore size distribution. It stands to reason,
therefore, that the optimum physical structure for any membrane material is based on a thin
layer of material with a narrow range of pore size and a high surface porosity.
Chapter 1 – Membranes and membrane bioreactors
15
The traditional materials used in the fabrication of membranes included cellulose acetate,
polyamide and polysulfone. Other polymers are now used in the manufacturing of membranes
and include polypropylene, nylon, polyacronitrile (PAN), polycarbonate, polyvinyl alcohol
(PVA) and polyvinylidene fluoride (PVDF). In addition, ceramic and metallic membranes are
used for MF and UF applications. Membrane materials are typically polymer- or ceramic-based
(Tab.1).
Table 1. Materials used for the manufacture of membranes.
Material MF UF Material MF UF RO
Alumina X Ceramic composites X X
Carbon-carbon composites X Polyacrylonitrile (PAN) X X
Cellulose esters (mixed) X Polyvinyl alcohol (PVA) X X
Cellulose nitrate X Polysulfone (PS) X X
Polyamide, aliphatic X Polyethersulfone (PES) X X
Polycarbonate (track-etch) X Cellulose acetate (CA) X X X
Polyester (track-etch) X Cellulose triacetate (CTA) X X X
Polypropylene X Polyamide, aromatic (PA) X X X
Polytetrafluoroethylene (PTFE) X Polyimide (PI) X X
Polyvinyl chloride (PVC) X Composites X
Polyvinylidene fluoride (PVDF) X Composites, polymeric thin film X
Sintered stainless steel X Polybenzimidazole (PBI) X
Cellulose (regenerated) X X Polyetherimide (PEI) X
Polymeric spiral membranes are generally used when a high throughput is required, while
polymeric tubular membranes, which can often be mechanically cleaned, are more suited for
low-maintenance operations, highly viscous products or fluids with suspended material.
Hostile environments, high levels of solvents, wide pH ranges and other process considerations
may dictate the use of ceramic membranes. This technology is normally adopted for
ultrafiltration and microfiltration applications and typically uses an alumina or zirconium
coating that is applied to the inside surface of a ceramic support. The capital cost of ceramic
membranes is much higher than conventional polymeric membranes but in some applications
they are the only viable proposition. Ceramic membranes often provide a longer operational
lifetime, offsetting the high initial cost. Ceramics are not resistant to abrasion although
polymerics may be.
Organic membranes
Cellulose acetate (CA) is the classic membrane material used by the pioneers of modern
membrane technology to create skinned membranes. There are several advantages to the use of
CA and its derivatives as membrane materials such as hydrophilicity (to minimize fouling of
the membrane), wide range of pore sizes that can be manufactured (from RO to MF) with
reasonably high fluxes, low cost. Among the disadvantages of CA membranes are a fairly
narrow temperature range (maximum temperature 30-35°C), a rather narrow pH range
(preferably 3-6), poor resistance of CA to chlorine (less than 1 mg/l of free chlorine under
continuous exposure and 50 mg/l in a shock dose), risk of creep or compaction phenomenon to
a slightly greater extent than other materials, highly biodegradability.
Polyamide membranes (PA) overcame some of the problems associated with CA membranes, e.g.
the pH tolerances are wider. However, PA membranes are much worse with regard to chlorine
tolerance and Biofouling tendencies. Polyamides for the contact skin layer in many composite
membranes.
Nutrients and micropollutants removal by a membrane bioreactor (MBR) and activated carbons
16
The family of polysulfone membranes are widely used in MF and UF. Polysulfone (PS) and
Polyethersulfone (PES) are considered breakthroughs for MF and UF applications due
principally to the following favourable characteristics: wide temperature limits (up to 125°C),
wide pH tolerances (1-13), fairly good chlorine resistance (up to 200 ppm chlorine for cleaning
and up to 50 ppm chlorine for short-term storage of the membrane), easy to fabricate
membranes in a wide variety of configurations and modules, wide range of pore sizes available
for UF and MF applications (from 1000 MWCO to 0.2 µm), good chemical resistance to aliphatic
hydrocarbons, fully halogenated hydrocarbons, alcohols and acids. The main disadvantages of
PS and PES are the apparent low pressure limits (typically 7 bar with flat sheet membranes and
1.7 bar with PS hollow fibres) and hydrophobicity which leads to an apparent tendency to
interact strongly with a variety of solutes, making it prone to fouling in comparison to the more
hydrophilic polymers such as cellulose and regenerated cellulose.
Other polymeric materials. available industrially are:
Nylon: these membranes are naturally hydrophilic with fluxes of the same order of
magnitude as cellulosic membranes. They are also autoclavable; however they strongly
bind biological solutes such as nucleic acids and proteins.
Polyvinylidene fluoride (PVDF): it can be autoclaved and its chemical resistance to common
solvents is quite good. A very popular material for MF and UF. Is has better resistance to
chlorine than the PS family.
Polytetrafluoroethylene (PTFE): is also very stable to strong acids, alkalis and solvents and
can be used in a wide range of temperatures (from -100°C to 260°C). It is extremely
hydrophobic and finds many uses in the treatment of organic feed solutions, vapours and
gases. PTFE membranes are available only in MF pore sizes.
Polypropylene (PP): it is widely available in the form of hollow fibres. It is hydrophobic,
relatively inert and can withstand moderately high temperatures.
Regenerated cellulose (RC): these membranes are very hydrophilic and have exceptional non-
specific protein-binding properties. RC has also good resistance to some common solvents
such as 70% butanol and 70% ethanol and can tolerate temperatures up to 75°C.
Polycarbonate: is one of two polymers (polyester is the other one) that are used to make
track-etch membranes.
Inorganic membranes
Inorganic membranes are also generically referred to as ceramic or mineral. With a couple of
exceptions, inorganic membranes are available in tubular form, either as a single-channel tube
or multi-channel element that may contain 7 to 37 individual circular channels depending on
the relative diameters of the channel and the element. The inner diameter of individual channel
varies from 2-6 mm and lengths from 0.8 – 1.2 m. in all inorganic modules the feed flows
through the inside of the channels, while the permeate flows through the support layer around
the lumens in the monolith and to the outside of the element. Several individual elements are
assembled in one housing and two to four housings are placed in series in a stack, with several
stacks in parallel. The properties of inorganic membranes are: inert to common chemicals and
solvents, wide temperature limits, wide pH limits, pressure limits, extended operating lifetimes,
back flushing capability. Some limitations of ceramic membranes are also: they are brittle, pore
sizes available are mostly MF and UF, a large pumping capacity is required, expensive.
Chapter 1 – Membranes and membrane bioreactors
17
1.2.2 Membrane separation processes
The major pressure-driven processes are microfiltration (MF), ultrafiltration (UF), nanofiltration
(NF) and reverse osmosis (RO). In addition, important membrane processes include dialysis,
electrodialysis (ED) and electrodialysis reversal (EDR), gas separation and pervaporation (PV).
Pressure-driven processes
The primary difference between the four pressure-driven processes is the size of particles that
the membrane rejects, which is determined in part by the pore size of the membrane. The
relationship of these major membrane processes and conventional particle filtration in terms of
rejection of particle sizes is provided in Table 2.
Table 2. Particle size range of membrane separation processes
Separation
process
Reverse
osmosis
Nanofiltration Ultrafiltration Microfiltration
Particle
filtration
Size (µm) 0.0001 0.001 0.01 0.1 1 10
MWCO < 100 200 20000 500000
Microfiltration (MF) refers to membranes that have pore diameters from 0.1 to 10 µm (Cheryan,
1998). MF systems are used to filter suspended particulates, large colloids, bacteria and oil from
solutions. MF systems are used in water purification (sterile water for industry and drinking
water), in the food industry for clarification of fruit juices and wine and concentration of
gelatine and sugars, for recycling the fluids in aqueous parts washers by removing the oil and
solids contamination, and for metals removal in the mining and metal finishing industries.
Operating at low pressures, MF technology can separate suspended and colloidal materials
down to approximately 0.1 µm.
Ultrafiltration (UF) membranes have the ability to separate molecules in solution based on size
down to approximately 0.005 µm and have a MWCO in the low 1000s. UF is used to separate
proteins from saline solutions by concentrating the larger molecules on the feed side of the
membrane while allowing water and dissolved salts to pass through the barrier. The largest
single use of UF membranes is to remove proteins with a molecular weight greater than 8000
from cheese whey.
Nanofiltration (NF) membranes can be visualized as UF membranes with smaller pore sizes
(0.001 µm). NF systems are sued to provide higher quality water than UF. NF membranes reject
dyes and some large molecular weight compounds such as sugars. NF membranes with smaller
pore sizes can separate divalent cations from monovalent cations.
Reverse osmosis (RO) was the first membrane process to be widely commercialized. RO
membranes are used to separate salts and low molecular weight compounds from water
because they are highly permeable to water and highly impermeable to microorganisms,
colloids, salts and organic molecules. RO is a reversal of the natural process of osmosis; a
process by which a dilute solution passes through a semi-permeable membrane into a more
concentrated solution. In reverse osmosis the process is reversed by applying external pressure
greater than the osmotic pressure. Typically, RO technology is used to remove inorganic salts
form wastewater. However RO can also be used to treat wastewater containing some organic
solvents. RO systems can be used to separate pure water from contaminated matrices, such as
Nutrients and micropollutants removal by a membrane bioreactor (MBR) and activated carbons
18
the treatment of some hazardous wastes through concentration of hazardous chemical
constituents, where pure water can be recovered on the other side of the membrane. There high-
pressure devices require pre-treatment for effective operation.
Other important membrane technologies include dialysis, electrodialysis and pervaporation.
The largest market for membranes is still haemodialysis where a co-current or counter-current
membrane system is used to clean the blood. Electrodialysis (ED) involves setting up an electric
cell across membranes so that cations and anions are attracted to the anode and cathode and
travel through the membranes to the waste stream. The difference in electrical potential is the
driving force. ED is used primarily in desalination of seawater or deionization (including
softening) of water, as well as the removal of heavy metals form water and wastewater.
Electrodialysis reversal (EDR) is an ED process in which the polarity of the electric cell is
reversed during the process to reduce fouling and maintain flux through the system.
Pervaporation (PV) is a process that essentially evaporates solvents through a membrane. PV is
used to separate water from solvents, either to remove a small amount of solvents from a clean
water stream or to remove a small amount of water to purify a solvent stream and to separate
one organic solvent from another.
1.2.3 Membrane configurations
Figure 1 presents schematic of the various types of ideal continuous-flow membrane separation
process.
Figure 1. Types of ideal continuous flows used in membrane separation processes.
Co-current flow (Fig.1a) is the flow pattern in a membrane module in which the fluids on the
upstream and downstream sides of the membrane move parallel to the membrane surface and
in the same direction. Co-current flow is used in certain dialysis units.
Counter-current flow (Fig.1b) is the flow pattern in a membrane module in which the fluids on
the upstream and downstream sides of the membrane move parallel to the membrane surface
but in opposite directions. Counter-current flow is also used in certain dialysis units.
Cross-flow or tangential flow (Fig.1c) is the flow pattern in a membrane module in which the
fluid on the upstream side of the membrane moves parallel to the membrane surface and the
Concentrate
Permeate
Feed
Sweep
(a) Co-current flow
Concentrate
Permeate
Feed
Sweep
(b) Counter-current flow
Concentrate
Permeate
Feed
(c) Cross flow (d) Dead-end flow
Permeate
Feed
Chapter 1 – Membranes and membrane bioreactors
19
fluid on the downstream side of the membrane moves away from the membrane surface in a
direction perpendicular to the surface.
Dead-end flow (Fig.1d) is the flow pattern in a membrane module in which the only outlet for
the upstream fluid is through the membrane.
The geometry of the membrane, i.e. the way it is shaped, is crucial in determining the overall
process performance. Other practical considerations concern the way in which the individual
membrane elements, that is the membranes themselves, are housed to produce modules. The
optimum geometry, or configuration, for an individual membrane element is one that has high
membrane area to module bulk volume ratio, high degree of turbulence for mass transfer
promotion on the feed side, low energy expenditure per unit product water volume, low cost
per unit membrane area, design that facilitates cleaning and that permits modularisation.
Commercial membrane separation systems are cross-flow or tangential flow devices in which
the inlet fluid flows parallel to the membrane filter surface with a fraction of the flow passing
through the membrane on a pass. The common arrangements for membranes include pleated
filter cartridge, plate and frame, tubular, spiral wound and hollow fine fibre (Tab.3).
Table 3. Membrane configurations.
Configuration Pleated
cartridge
Plate-and-frame Spiral-wound Tubular Hollow fibre
Area/volume ratio
(m
2
/m
3
)
800 – 1000 400 – 600 800 – 1000 20 – 30 5000 – 40000
Cost Low High Low Very
high
Very
low
Turbulence
promotion
Very poor Fair Poor Very good Very poor
Advantages Robust
construction,
compact
design
Can be
dismantled for
cleaning
Low energy
costs, robust
and compact
Easily
mechanically
cleaned,
tolerates high
TSS waters
Can be
backflushed,
compact
design,
tolerates high
colloid levels
Disadvantages Easily
fouled,
cannot be
cleaned
Complicated
design, cannot
be backflushed
Not easily
cleaned,
cannot
backflush
Replacement
cost
Sensitive to
pressure
shocks
Applications Dead end
MF
ED, UF, RO RO, UF Cross-flow
filtration,
high TSS
waters
MF, UF, RO
Pleated filter cartridge is the lowest cost flat plate geometry, used exclusively in microfiltration
and generally designed as a disposable unit.
Plate and frame membranes were among the earliest configurations in the market. Plate and frame
devices use flat membrane sheets with permeate collection between the sheets. The sheets are
sealed around the edges but with a provision for permeate removal (usually by a tube). Several
of these plates are stacked on top of each other and clamped together o form a module or
cartridge. Many plate and frame systems are based on dead-end flow and are more subject to
plugging. The most commercially significant application of the plate and frame design is in
electrodialysis modules (or stacks), although some microfiltration units and one reverse osmosis
module design are also based on this configuration.