Lappeenranta University of Technology LUT School of Engineering Science
Double Degree Programme in Chemical and Process Engineering
Gafiullina Anastasia
ELECTROCHEMICAL WATER TREATMENT AND FILTRATION OF GENERATED SLURRIES
Examiners: Professor Antti Häkkinen Doctoral student Maria Mamelkina
Doctoral student Paula Vehmaanperä
2 ABSTRACT
Gafiullina Anastasia
Electrochemical water treatment and filtration of generated slurries Master’s Thesis, 2018
Lappeenranta University of Technology LUT School of Engineering Science
Double Degree Programme in Chemical and Process Engineering 98 pages, 39 figures, 16 tables and 19 appendices
Examiners: Professor
Doctoral student Doctoral student
Antti Häkkinen Paula Vehmaanperä Maria Mamelkina
Keywords: Electrochemical water treatment, electrocoagulation, pressure filtration, particle properties, particle aggregates.
This work is focused on the electrocoagulation (EC) as an electrochemical water treatment process followed by solid-liquid separation of the solid species, which form during the treatment. In the theory part of the thesis, the mechanisms of electrolyte coagulation are described. The mechanisms of pollutants removal in the EC and the factors affecting the EC efficiency are listed. That is followed by the theory of the cake filtration and the description of parameters that have an influence on the process performance. The description of the equipment used for performing the filtration tests are given.
In the experimental part of the paper, batch EC-experiments were performed at the laboratory scale with iron and aluminum electrodes used in different combinations. The removal efficiencies of nitrate, sulphate, chloride, ammonium and total nitrogen under the established conditions were determined. The measurements of pH, reduction-oxidation potential and conductivity were performed to investigate the processes occurred during the treatment and their effects on the formed solids. For determination of the solid species X-ray diffraction analyses were made. The flocculated nature of the particles was confirmed by the scanning electron microscope analyzes where the aggregates were observed. Based on the SEM images, shapes of the aggregates were estimated as very irregular. The size distribution data for the aggregates were obtained based on the laser diffraction method.
Suspensions generated during the EC procedure consisted of mineral solid matters but their filtration behavior differed from the typical behavior of mineral-based slurries due to the properties of the particles contained in the slurry. To understand this phenomenon, the filtration experiments were conducted. The effects of operating pressure and slurry volume on the process performance were estimated based on the constant pressure filtration tests. The shape of the particle aggregates was considered as a main parameter that determines the cake structure with such a high specific
3
resistance and moisture content. To facilitate the filtration the addition of filter aids was also studied. In addition to this, preliminary tests were made to estimate the applicability of vacuum filtration for the separation of solids.
EC was found to be an effective method for removal of nitrate and sulphate from the water solutions. The highest removal efficiency was achieved by using Al/Fe electrode combinations.
Vacuum filtration has shown a poor efficiency in terms of the purity of filtrate whereas pressure filtration was considered as an appropriate method for SLS of the studied electrochemically generated slurries. Application of filter aids was estimated to be reasonable and recommended in order to facilitate and shortened the pressure filtration procedure.
4
ACKNOWLEDGEMENTS
This Master’s Thesis was done as part of the EWT-CYNCOR (Electrochemical water treatment for cyanide and nitrogen compounds removal) project funded by European Institute of Innovation and Technology (EIT RawMaterials). The project aims at upscaling of electrochemical process of water treatment for mining industry.
Hereby I would like to thank everyone who helped me during my studies and work on the Master’s Thesis at Lappeenranta University of Technology. I express my thanks to the Department of Separation and Purification Technology, LUT, and EIT RawMaterials for the engrossing and diversified research topic that was provided to me. I am grateful to my examiner Professor Antti Häkkinen for making me interested in the solid-liquid separation processes and for his help and guidance on conducting the research.
I owe my deepest gratitude to my supervisors Paula Vehmaanperä and Maria Mamelkina for their professional attitude, friendly advices and support that encouraged me to perform at my best. I would like to express my thanks to Mikko Huhtanen and Dmitry Safonov, who spent long hours teaching me to work with laboratory equipment. I would like to thank Toni Väkiparta for his help with SEM and XRD analyses and all members of the solid-liquid separation group for creating a positive mood during all these months.
I am grateful to my parents and my sister Natalia for their understanding and advices, and to my friends Lisa, Ania and Misha, on whose support I can always count. I fail words to express my thanks to Vitalii, who makes this year a wonderful time and without whom this work could not be finished.
I hope readers will find this work interesting and helpful.
Best regards, Anastasia
This thesis has been supported by EIT RawMaterials
5
TABLE OF CONTENTS
INTRODUCTION ... 9
LITERATURE REVIEW ... 11
1 THEORY OF ELECTROLYTE COAGULATION ... 11
2 PRINCIPLES OF ELECTROCOAGULATION ... 16
3 SOLID-LIQUID SEPARATION OF THE ELECTROCHEMICALLY GENERATED SLUDGE ... 18
3.1 Properties of the electrochemically generated sludge ... 18
3.2 Settling and filtration ... 19
4 CAKE FILTRATION THEORY ... 20
4.1 Cake formation ... 21
4.2 Fluid flow through a porous media ... 21
4.3 Filter cake characteristics... 23
4.3.1 Average specific cake resistance ... 23
4.3.2 Porosity, saturation and moisture content of compressible cakes ... 24
4.4 Cake filtration in the constant pressure mode ... 26
5 FACTORS AFFECTING THE CAKE FILTRATION PERFORMANCE ... 29
5.1 Chemical pre-treatment: coagulants and flocculants ... 29
5.2 Addition of filter aids ... 29
5.3 Filter media ... 31
5.4 Properties of the slurry ... 32
5.5 Particle properties ... 32
5.5.1 Particle size and particle size distribution ... 32
5.5.2 Particle shape... 34
5.5.3 The interactions between particles and the surrounding liquid ... 34
5.6 Operating conditions ... 35
5.6.1 pH ... 35
5.6.2 Pressure ... 35
6 CHARACTERISATION OF PARTICLE PROPERTIES ... 36
6.1 Laser diffraction technique ... 39
6.2 Automated imaging analysis ... 39
7 FILTRATION EQUIPMENT... 40
7.1 Vacuum filters ... 40
7.2 Industrial-scale Nutsche filter ... 41
7.3 Laboratory-scale Nutsche filter ... 42
EXPERIMENTAL PART ... 43
8 AIMS AND CONTENT OF THE WORK ... 43
9 MATERIALS AND METHODS ... 44
9.1 Materials ... 44
9.1.1 Chemicals ... 44
6
9.1.2 Filter aids... 44
9.2 Experimental setup ... 46
9.2.1 Electrocoagulation ... 46
9.2.2 Pressure filtration... 47
9.2.3 Vacuum filtration... 49
9.3 Experimental procedure ... 49
9.3.1 Electrocoagulation experiments and sampling ... 50
9.3.2 Pressure filtration... 50
9.3.3 Filtration with filter aids ... 52
9.3.4 Vacuum filtration... 53
9.4 Analyses ... 53
RESULTS AND DISCUSSIONS... 56
10.1 Electrocoagulation ... 56
10.1.1 Removal efficiency of anions and total nitrogen ... 56
10.1.2 Conductivity of the samples ... 62
10.1.3 Determination of solid species ... 63
10.1.4 SEM analysis of the solid species... 68
10.1.5 Particle size distributions of the solids... 70
10.2 Pressure filtration ... 74
10.2.1 Properties of the formed slurries and densities of solid species ... 74
10.2.2 Average specific cake resistance and filter medium resistance ... 76
10.2.3 Properties of filter cakes: moisture content, moisture ratio, porosity and thickness 80 10.3 Filtration with filter aids ... 84
10.3.1 Body-feed filtration mode ... 85
10.3.2 Pre-coat filtration ... 90
10.4 Vacuum filtration ... 93
CONCLUSION ... 95
REFERENCES ... 99
APPENDICIES ... 107
7
NOMENCLATURE
A filtration area [m2]
a, b coefficients [-]
Ach cross-section of the streaming channel [m2]
AH Hamaker constant [J]
C attractive Van der Waals constant [-]
c distance between particles [m]
c0 initial concentration of the solution [mg/L]
ci concentration of a sample taken at the i-period of time [mg/L]
cw solid content by mass fraction [%]
d particle diameter [m]
di mean diameter in the examined size interval [m]
E strength of applied electrical field [V/m]
GA Van der Waals attraction forces [J]
Gel electrostatic repulsion forces [J]
h distance between particle surfaces [m]
J current density [A/m2]
j total amount of size intervals in the particle size distribution [-]
k Debye length [m-1]
K permeability constant of a porous material [m2]
kB electrolyte conductivity [S/m]
L material thickness [m]
Lc cake thickness [m]
Lch length of the streaming channel [m]
m mass of the solids [kg]
mc mass of the cake [kg]
MC moisture content [%]
mds mass of dry solids in the cake [kg]
mfl mass of fluid retained in the cake [kg]
ml mass of the liquid retained in the cake [kg]
mp+w mass of the pycnometer fully filled with pure water [kg]
mp+w-s mass of the pycnometer filled with solids and pure water [kg]
mps mass of the solids added into the pycnometer [kg]
MR moisture ratio [%]
ms mass of dry solids in the sample of slurry [kg]
msl mass of the slurry [kg]
mwc mass of the wet cake [kg]
n compressibility index [-]
ni number of particles in the examined size interval [-]
nr refraction index of the solvent [-]
∆" applied pressure [Pa]
∆"# pressure drop through the filter cake [Pa]
Q volumetric flow rate of the liquid [m3/h]
R1, R2, R12 radii of the partcles [m]
Rc cake resistance [1/m]
RE removal efficiency [%]
Rm medium resistance [1/m]
S saturation of the cake [%]
SSA specific surface area [m2/kg]
8
ts time at the beginning of the constant pressure period [s]
V volume of filtrate [m3]
Vc volume of the cake [m3]
Vs filtrate volume at the beginning of the constant pressure period [m3]
Vvoid volume of voids in the cake [m3]
x cake moisture ratio [%]
z valence of couter-ions [-]
$ specific cake resistance [m/kg]
$% resistance at unit applied pressure drop [m/kg]
$&' average specific cake resistance [m/kg]
( cake porosity [%]
ε0 electric permittivity of vacuum 8.85·10-12 [F/m]
(&' average cake porosity [%]
(+ dielectric constant of the solvent [F/m]
(, relative dielectric permittivity of the medium containing particles [-]
- zeta-potential [V]
. angle of scattering [˚]
/ compressibility index [-]
/’ wavelength of the light [m]
0 liquid viscosity [Pa·s]
01 electrophoretic mobility [m2/V·s]
23 electron densities of the particles [-]
256 density of fluid retained in the cake [kg/m3]
2+ density of solids which form the cake [kg/m3]
27 density of the pure water [kg/m3]
8 Heywood shape factor of the particles [-]
89 Stern potential [V]
CCC critical coagulation concentration EAA electroacoustic attenuation EC electrocoagulation
EDL electrical double layer EF electroflotation
ELS electrophoretic light scattering EO electrooxidation
FA filter aids
PSD particle size distribution RC regenerated cellulose RedOx reduction-oxidation RHA rice hull ashes
SEM scanning electron microscope SVI sludge volume index
TN total nitrogen TOC total organic carbon
US ultrasound
XRD x-ray diffraction ZP zeta-potential
9
INTRODUCTION
Mining industry consumes large amount of fresh water. It is also a source of huge volumes of polluted water which treatment is challenging due to uniqueness of water composition at each particular mine. However, such compounds as suspended solids, sulphates, nitrates and other anions, and metals can be found at any mine water (Lottermoser, 2010). Once being discharged, these compounds become a treat for the environment. Namely, the contamination of surface and ground water, impossibility of consumption of downstream water, degradation of soil are considered as the main problems (Cidu et al., 2009). Enforcement in the field of environmental legislation indicates that new solutions are needed. Implementation of new effective techniques can allow not only to meet requirements of the standard, but also allow to recycle the treated effluent thus solving the problem with fresh water withdrawing.
Current methods for mine water treatment include conventional precipitation with lime or addition of chemicals and advanced biological degradation, adsorption, ion exchange, and membrane filtration (Al-Shannag et al., 2015; Guimarães and Leao, 2014). Conventional techniques have become more widespread, however, in many cases, it is not possible to achieve high removal efficiencies of compounds that are presented in low concentrations but are considered as dangerous (Gatsios et al., 2015; Al-Shannag et al., 2015). For advanced methods, volume and high concentrations in the process streams may become an issue (Guimarães and Leao, 2014). In terms of the composition of mine effluents, electrocoagulation (EC) seems to be a promising technique.
It is suitable for removal of various problematic (biologically/chemically stable, toxic) pollutants since EC combines several treatment mechanisms, namely, electrocoagulation, flocculation, sorption, electrooxidation, electroreduction, electroflotation (Liu et al. 2010). Among the key benefits of EC is simplicity, high performance, chemical-free “nature” and lower costs (Mollah et al., 2001).
EC has been successfully implemented in many industries (Kabdaşlı et al., 2012), but mining is an exception from this fact. It represents a new area which should be investigated. Only a few studies have been made to verify the applicability of electrocoagulation and to estimate its performance for purification of mine water (Mamelkina et al., 2017; Nariyan et al., 2017; Touahria et al., 2016). Among them, a work made by researchers from Lappeenranta University of Technology conducted in the “Water Conscious mining” project framework (2017) seems to be the most detailed. The obtained results demonstrate the high efficiency of EC in the removal of
10
sulphates, nitrates and metals. Studies of the fundamentals of the electrocoagulation and interplay of operating parameters would allow to determine the most optimal conditions thus increasing overall process efficiency. However, EC does not remove all of the contaminants from the system.
Some of them become transformed into non-harmful components but they are still presented in the solution in the form of insoluble compounds. Further step is required to separate the formed solid matter from the treated solution.
Filtration seems to be an appropriate solution to perform the solid-liquid separation of the EC suspension. Pressure filters are widely used in the mining industry since they are easy to operate, provide a clear filtrate and cakes with relatively low moisture content, can deal with concentrated slurries, suitable for cake drying operations and can operate with large volumes if automated (Townsend, 2003). Prior to upscaling of the method process parameters need to be determined.
The main parameter is the average specific cake resistance which value can only be found experimentally since the combination of a particular slurry, a certain filter media and operation conditions gives the unique value of the resistance. Mineral slurries are in general more easily filtered that biological ones (Henrikson, 2000 Wetterling et al., 2014). However, according to the preliminary tests, the filtration of the slurry obtained after EC treatment is challenging since it behaves more in a bio-slurry manner forming cakes with the high specific resistance (Vainio, 2017).
In the current Master’s Thesis the efficiency of electrocoagulation for mine water treatment was estimated. Experiments were conducted with a synthetic solution at the pilot scale. Iron and aluminum electrodes were tested at three different combinations and the most effective one was determined. The suspension that was forming within the treatment was collected to analyse the concentrations of total nitrogen and anions and changes in the formed solid species. The collected data were used to investigate the processes that took place in the reactor and their effect on the generated matter.
The pressure filtration performance was determined for the slurries obtained in the pilot-scale electrocoagulation treatment made with different electrode materials. The effect of volume of slurry and operating pressure were investigated. The obtained results were explained by the properties of the particles which constitute the solid phase. Effectiveness of the method of filter aid addition was estimated in the pre-coating and body-feed tests and the better solution was provide after comparing the results. Additionally, applicability of a vacuum filter to perform the separation was studied.
11
LITERATURE REVIEW
1 THEORY OF ELECTROLYTE COAGULATION
The electrolyte coagulation mechanism was described by the scientists Derjaguin and Landau, Vermey and Overbeek who developed the theory named after them (DLVO theory). According to the theory, the energy of the particle interaction in the solution is a sum of the Van der Waals attraction and electrostatic repulsion forces. Van der Waals forces determines particle attraction, whereas the electrostatic repulsion seeks to distance particles. Electrostatic repulsion occurs when the electrical double layers (EDL) of particles overlap resulting in increase of the concentration of co-ions (Figure 1.1) which repel each other. (Derjaguin and Landau, 1941)
Figure 1.1. The schematic view of the overlapping of the EDLs of the negatively charged particles (Sillanpää and Shestakova, 2017).
The mentioned forces can be described quantitatively. Van der Waals attractive force for the two particles with a spherical shape is calculated according to Eq. (1.1) (Hamaker, 1937):
:; ==>2?2>
6 A B 2D?D>
E>− (D?+ D>)>+ 2D?D>
E>− (D?− D>)> + JK LE>− (D?+ D>)>
E>− (D?− D>)>MN
= −OP
12A B 2D?D>
E>− (D?+ D>)>+ 2D?D>
E>− (D?− D>)>+ JK LE>− (D?+ D>)>
E>− (D?− D>)>MN
(1.1)
Where: ρ?,ρ> - electron densities of the particles; C – the attractive Van der Waals constant;
c – interparticle distance; R?, R> - radii of the particles; AV – the Hamaker constant.
12
If two particles interact with each other at a distance much smaller that their radii and these particles are identical, then Eq. (1.1) takes the form of Eq. (1.2):
:; = −OPD?>
12ℎ (1.2)
Where: D?> = D? = D>; ℎ = E − (R?+ R>) - the distance between particle surfaces.
As the AV value is positive, the GY will be negative that indicates the attraction between particles.
The second law of thermodynamics postulates that spontaneous process’ direction leads to the free energy reduction. Overlapping of particles’ EDLs leads to the increase of the free energy. To compensate this, the repulsive forces start to act and the value of it can be found with Eq. (1.3) (Sillanpää and Shestakova, 2017):
:Z6 = 2=(,(%D?>89>[\"(−]ℎ) (1.3) Where: ε^ - the relative dielectric permittivity of the medium containing particles; ε% - the electric permittivity of vacuum; k – Debye length or inverse of the thickness of diffuse EDL; 89 - the Stern potential.
With the increase of the thickness of the diffuse EDL, the particles repulsion is stronger. Hence, for making coagulation easier, the thickness of EDL should be decreased which can be achieved to some extent by temperature decrease and with the rise of the concentration of electrolyte and the charge of ions.
Both of the mentioned forces depend on the distance. Attraction action is inversely proportional to the interparticle distance in a power of two, which makes its influence to become significant on the short distances. Repulsion forces is manifested on the distances of the order of EDL thickness and it exponentially decreases while moving away from the particle (Vepsäläinen, 2012).
Figure 1.2 illustrates the relationship between the distance between particles and the forces impacting on them.
13
Figure 1.2. The schematic view of the dependence of the EDL repulsion and Van der Waals attraction forces on the distance between particles (Sillanpää and Shestakova, 2017).
It should be mentioned, that the electrostatic repulsion is not the only force that prevents the particles to approach each other. Another reason for this is a solvation shell of solvent molecules surrounding the particles, which results in maintaining the stability of colloids. Its influence is considered much smaller than EDL repulsion, but it still affects the system. When the particles come too close, the overlapping not only of the electrical double layers but also the water boundary layers take place, providing the increase in the repulsion force (Salopek et al., 1992).
After the attraction and repulsion forces are combined, the two minimums can be seen in Figure 1.2. The primary minimum, which exist when the distance is short, represents the state of the system when the coagulation took place and the rest is presented as a result of the process passed. The secondary minimum has the influence at longer distance and its value is related to the flocculation process. The potential energy at the primary minimum is significantly smaller than at the secondary minimum and, hence, the particles at the primary minimum are more tightly attached. The particles need to posses a defined value of the kinetic energy in order to overcome the energy barrier before coagulation occur (Sillanpää and Shestakova, 2017).
The state of the dispersed system can be described in general via taking three different relative values of the energy barrier and the secondary minimum depth. They can be (1) both insignificant, (2) high barrier/small minimum depth, and (3) high value of the minimum depth (high value is the determining factor in this case and, hence, amount of energy of the barrier is not important). The first and the third cases are the opposite situations. In the first case system can be characterized as
14
instable and the coagulation occurs easily and does not reverse. The kinetic energy of the particles makes them move resulting in their convergence and interactions, which lead to the decreasing of the interface of colloids relative to the suspension, and coagulation occurs (Sillanpää and Shestakova, 2017). The third case illustrates a stable system, where particles can move in pairs or in more complex units, they form by the attaching some other particles to the existing pair.
Nevertheless, the system in general retain its dispersed state in equilibrium (Sillanpää and Shestakova, 2017). A high barrier/small minimum depth values – the second case, – correspond to the situation when no coagulation occurs because the energy of the particles that does not reach the value required to overcome the energy barrier. To destabilize the system, additional kinetic energy should be given to the particles in order to make the collision between them more frequent.
Simple ways to provide the additional energy are temperature increase and mechanical stirring (Vepsäläinen, 2012).
Another option of the system destabilization is related not to the properties of the particles but to the energy barrier. An introduction of the indifferent electrolytes to the solution leads to the compression of the EDL (increasing its thickness) of the particles. Once it happens, the repulsive forces become weaker, Van der Waals forces start to prevail and coagulation can be observed (Vepsäläinen, 2012). As an example, in the work of Goncharuk et al. (2017) effects of indifferent electrolytes (NaCl, KCl, LiCl, NaI, NaNO3, CaCl2, and MgCl2) on the EDL and process of coagulative structuring (aggregation) for aqueous dispersions of nanosilica were investigated.
Addition of all investigated electrolytes resulted in an increase in the size of silica aggregates after the EDL was compressed. Among the studied materials, the strongest effect was given by electrolytes that contained calcium and magnesium salts due to their higher charge.
The two modes of coagulation can be distinguished, slow and rapid form. If not all collisions between particles are effective, coagulation proceeds slowly, and vice versa. Rapid coagulation obeys the Shulze-Hardy rule that postulates that the interaction between lyophobic colloidal particles and the oppositely charged ions leads to coagulation. The effect of coagulation increases with the increasing of the valence of the counter-ions. Hence, the process of coagulation can be intensified by introducing the counter-ions with higher valences to the system (e.g., through addition of salts of aluminum or iron among which the most widely used are AlCl3, FeCl3, Al2(SO4)3, FeSO4, Fe2(SO4)3) in the sufficient concentrations called critical coagulation concentration (CCC). CCC can be easily estimated by Eq. (1.4) (Sillanpää and Shestakova, 2017).
15 AAA = A
_` (1.4)
Where: C – the coagulating capacity which is constant for the particular system; z – the counter-ion valence.
Obviously, the addition of multivalent ions will result in a more pronounced effect and trivalent ions are preferable than divalent ones. The ratio between the amount of mono-, di-, and trivalent ions required to start the rapid coagulation process can be demonstrated quantitatively by Eq. (1.5).
In some cases, the abnormalities are presented, e.g. due to the high-energy barrier existence.
Among the ions with the same charge value, the most effective ions are those whose ionic radii is larger.
abKc: abK>c: abKec = 1 1`: 1
2`: 1
3`≈ 730: 65: 1 (1.5)
Coagulation can be initiated via the introduction of ions that, once being adsorbed by the particle surface, neutralize the charge of the particle and destabilize them. Examples of such type of ions are ferrous (FeSO4) and ferric (Fe2(SO4)3) salts and aluminum salts (AlCl3, Al2(SO4)3, etc.), which have the greatest applications in the coagulation/flocculation processes. More precisely, their products of hydrolysis are more effective in the process and, hence, the pH of the solution becomes an important parameter and that is why the optimal pH-value needs to be determined and maintained at the required level. Along with it, highly charged polynuclear hydroxo complexes of Al and Fe, which can be formed during hydrolyses, can interact with negatively charged colloidal pollutants. This results in the pollutants neutralization and the coagulation enhancement. The potential value at which destabilization occurs is close to 0 mV (Mollah, 2001). As the reactions occur, iron and aluminum precipitate in a form of Fe(OH)3 and Al(OH)3. Negatively charged pollutants, attracted by positively charged aggregates, are neutralized, destabilized and precipitated. The attention should be given to the coagulant dosage to avoid the reversal of the charge and restabilization of the colloids.
To sum the given information, the mechanisms, which act to destabilise colloidal particles, are (Crittenden et al., 2005; Bratby, 2006):
1) EDL compression: adding of the ions to the solution increases their concentrations leading to the approach of the particles, decreasing EDL length and its compression.
16
2) Adsorption of the counter-ions or counter-charged polymers on the particle surfaces. Surface charge is neutralized resulting in decreasing of repulsive forces.
3) Interparticle bridging between polymerised organic chemicals or metal coagulants and the particles in which one polymer chain adsorbs on multiple particles leading to the increase of a molecular weight via bridging.
4) Precipitation of insoluble products of salt (Al, Fe) hydrolysis and enmeshment of colloidal particles into sweep flocs.
2 PRINCIPLES OF ELECTROCOAGULATION
Technologies based on the coagulation/flocculation mechanism described above are widely used for water purification. Conventionally, during the coagulation process the coagulants (chemicals that make the particle repulsive energy decrease) are introduced to the treated wastewater to destabilize particles and make them agglomerate more easily. Then the coagulants (long chain synthetic polymers) are added for making the agglomerated particles attached to each other with a weak bond resulting in enhancement of the settling, thus, decreasing the separation time (Kuokkanen et al., 2013).
Another method of the treatment of water based on the mentioned mechanism is considered as one of the advanced methods and is related to the electrochemical water treatment technologies. The electrocoagulation (EC) allows obtaining the same results and proposing some significant benefits.
The main advantages of the method are: simplicity of equipment and operation; possibility to vary many parameters in a wide range; high quality of effluent; no necessity to neutralize water after the treatment, thus, no pollution caused by pH adjusting chemicals; no added chemicals, hence, less amount of sludge (Siringi et al., 2012). EC effectively treats various pollutants that cannot be removed via conventional treatment methods due to their biologically and/or chemically stability, toxicity, etc. (Zodi et al., 2011).
In the electrocoagulation, the positive electrode undergoes the anodic reactions while on the negative side the cathodic reactions occur. Highly positively charged ions are produced by the consumable (“sacrificial”) electrodes that are the metal plates. Cathode is a negative electrode and anode is positive. The release of cations and the formation of metal hydroxides during electrolytic wastewater treatment is achieved by the dissolution of anode (Emamjomeh and Sivakumar, 2009).
Highly adsorptive complexes of polyvalent polyhydroxides form aggregates with colloidal
17
particles. The destabilization mechanism of the contaminants follows the rules described in the previous section. Destabilization, which occurs between the electrode space, is expressed by breaking up of the suspension and creation of a sludge that bridges colloidal particles in the treated solution and, as a result, formation of flocs (Sahu et al., 2014).
During the electrocoagulation the processes of electroflotation (EF) and electrooxidation (EO) are often take place (Mickova, 2015). In the EF process, the removal of contaminants occurs via their movement to the surface of solution once being entrapped by the H2 and O2 gas bubbles which are generated in the electrolysis. The result of EO is a decomposition of organic pollutants to water, CO2 and other oxides. The illustration of these mechanisms is presented in Figure 2.1. The dominant mechanism of the reaction depends on the treated pollutants and the values of operating process parameters (Sahu et al., 2014).
Figure 2.1. The main mechanisms of pollutants removal during the EC process (Mollah et al., 2004).
Many factors have an influence on the electrocoagulation performance. Sometimes they interact with each other, and it also might happen that changes in the same parameter cause a positive as well as a negative effect on the overall process depending mainly of the type of pollutant. Thus, the optimal values/configurations should be determined for each particular solution for the following parameters (Mollah et al., 2001; Liu et al., 2017):
18 v Electrode material
v Electrode potential v Electrode distance v Electrode arrangements
v Hydrodynamical conditions: water v flow rate between electrodes v Current density
v Conductivity v pH value v Temperature
v Reaction time v Reactor design:
§ Physical design: batch/
continuous mode of operation;
reactor assembly
§ Chemical interaction: electrode material; electrode passivation;
solution pH
v Concentration of anions
v Concentration of the pollutants
Optimisation is also important, as the process parameters are one of the factors, which determine the properties of the formed slurry. However, the investigation of the influence (synergistic, additive and antagonistic) of all the parameters involved in the process in a one research is very difficult, and the process optimisation for each particular case should be done.
3 SOLID-LIQUID SEPARATION OF THE ELECTROCHEMICALLY GENERATED SLUDGE
Separation of sludge generated during the treatment of the solution is an important step of the process. It can be done by settling or by filtration of the slurry.
3.1 Properties of the electrochemically generated sludge
The sludge properties are required to be determined as its further treatment and disposal are an ecological and economical issue for many sludge-generating processes. As it was previously mentioned, the sludge formed during the EC has a good settleability and dewatering properties due to its composition (Mollah et al., 2001). Such sludge characteristics as easy dewatering and low sludge volume index cause the attractiveness of the sludge management (Zodi et al., 2009).
SVI, i.e. the volume in millilitres occupied by 1 g of a suspension after 30 min settling that is typically used to monitor settling characteristics of sludge, was equal 73 mL/g for aluminium electrodes and 52 mL/g for iron ones (Zodi et al., 2009).
According to the article of Gomes et al. (2007), the sludge composition varied with different electrode materials. With iron electrodes the crystalline (magnetite) and poorly crystalline
19
(oxyhydroxides) phases were formed; by using the aluminium electrodes the amorphous (aluminium hydroxide and oxyhydroxide) phases were formed; by using combined electrodes (iron-aluminium) the substances mentioned above and some other products were obtained. Sludge generated when using aluminum electrodes had higher SVI than generated in the treatment with iron electrodes and worse settleability mainly due to the structure of the sludge which was «fluffy»
for the Al-formed sludge, whereas Fe-formed sludge was heavy enough to produce the compact layer (Zodi et al., 2009).
3.2 Settling and filtration
Sahu et al. (2014) have studied the settling of EC-treated effluents from pulp and paper and textile industries and chemical mechanic polishing processes of semiconductor fabrication. They have determined that the addition of sodium chloride leads to the higher anode consumption, which negatively affects the sludge settling rate and increases the volume of the sludge. According to (Mahesh et al., 2006(a)), the addition of alkalinity agent in a purification of pulp and paper industry effluents in the EC, made with iron electrodes, enhance the settlability of the sludge. When studying the filtration characteristics of the slurry and the formed sludge pressure filters can be used (Mahesh et al., 2006(b)). The main results, which can be obtained from the tests, are the values of the average specific cake resistance and filter medium resistance. The data on the cake compressibility can be obtained as well (Mahesh et al., 2006(b)).
Solid-liquid separation of the EC slurry can be performed in the cake filtration mode. In this procedure, the solids are retained on the surface of the filter medium and a porous layer called cake is formed. The formed cake allows to retain the particles smaller than medium pores (Tien, 2006). Cake filtration is applied for the treatment of slurries with concentration of solids higher than 1%-v (Svarovsky, 2000). In a recent work (Vainio, 2017), cake filtration was done by the laboratory scale pressure Nutsche filter at constant pressure filtration mode and some difficulties were found. The surprisingly high moisture content was determined in the formed cakes. Namely, the cake moisture content obtained after the filtration of the slurry generated by using iron electrodes in the EC process was 76.2 % by weight, whereas for the slurry generated by using aluminium electrodes it was 94.7 %. This could be an issue when applying the method on the industrial scale. Particle shape and zeta-potential value were suggested as possible reasons for it.
However, these parameters were not determined to prove the suggestion.
20 4 CAKE FILTRATION THEORY
One of the most used type of the filtration procedures is the cake filtration (Ripperger et al., 2012).
For performing the cake filtration, the two main components should be presented: a driving force, which is a pressure difference (or a centrifugal acceleration if filtration is performed by a centrifuge), and a separation material called “filter media”. During the cake filtration the solids are retained mainly by the cake layers earlier accumulated on the surface of a filter media. Because of this reason the cake filtration mode is considered as a surface filtration (Tien, 2006). Cake filtration can be performed with gravitational, centrifugal, vacuum or pressure filters. Higher pressure difference is achieved in the pressure filters.
There are several factors, which have an effect on performance of the pressure filtration:
1) Operating conditions: pressure difference, pH and temperature
2) Particle properties: particle size and shape, particle size distribution, particle interactions and zeta-potential
3) Filter cake characteristics: porosity, compressibility and specific cake resistance
4) Other parameters: filter medium resistance, concentration of solid phase, liquid viscosity.
Some of the mentioned factors have significantly higher influence than the others, however, all of them should be determined and optimised to achieve the effective separation. Modes of the cake filtration are derived on the constant flow rate, constant pressure, constant flow followed by constant pressure and variable flow and pressure filtration. Among them, the constant pressure filtration has become the most spread because of the ease of operation under the fixed pressure value.
The numerical cake filtration theory started to develop with the Darcy’s work (1856) where the Darcy's law was postulated. In the 1930s, Ruth’s studies showed that the occurring filtration resistance is a sum of resistance offered by the filter medium and resistance that arises from the cake growth. The following research conducted by Tiller, Huang, Cooper, Shirato, Okamura and others formed the classical theory on the cake filtration (Shirato et al., 1987; Tiller, 1953; Tiller and Cooper, 1960). Particle interactions, their size/shape and charge, process modelling, methods which allow to enhance filtration (e.g., filter aids) are a subject of nowadays studies. Cake filtration is widely applied in many industrial processes in the chemical, mineral, pulp and paper, petroleum, food and pharmaceuticals fields (Tien, Hsiau, 2009).
21 4.1 Cake formation
In the cake filtration, the solids are retained on the filter medium surface and form the porous layer called “filter cake”. During the first seconds, the filter medium is involved in the process as a material to capture the first particles to initiate their accumulation (Tien and Bai, 2003). Not all of these solids are retained by the medium as their size is in general smaller than the filter medium pores. Passage of the smallest particles through the filter medium takes place until the formation of the first cake layer (Ripperger et al., 2009). Clusters of particles cover the filter medium pores via the “bridging” mechanism, and the cake formation occurs. Once the cake layer is formed, it acts as a primary filtration medium. Due to the formed cake, the retention of the particles that are smaller than the medium pores is possible (Tien, 2006).
To form the cake, solid matter should be presented in sufficient amount and, thus, the cake filtration is applied for the treatment of slurries characterized by a sufficiently high (higher than 1% by volume) concentration of solids (Svarovsky, 2000; Tien, 2006). In addition, the ratio of the size of solids to size of the pores in the filtration material is important. It was empirically determined that the initial (it can be higher but no lower) ratio is close to 1:3 (particles: pores) (Ripperger et al., 2009). The cake formation and growth are crucial in the process of packing of particles. Generally, particles tend to orient randomly or to form loose structures so the packing is not ordered (Ni et al., 2006).
4.2 Fluid flow through a porous media
In the conventional filtration theory, a number of assumptions have been made to simplify the investigation process. Thus, such parameters as the velocity of solid phase and the moving effect, observed on the boarder between the surfaces of the suspension and the cake, are considered to have very small influence and, hence, they can be neglected. Fluid passes through the cake with the constant velocity. The permeability, porosity and average cake resistance are assumed to be affected by the applied compressive stress. In addition, according to the theory, a total pressure drop value is the sum of the pore liquid pressure and compressive stress (Tien, Bai 2003). The simple scheme of the cake filtration process is presented in Figure 4.1.
22
Figure 4.1. A schematic view of the cake filtration process (Svarovsky, 2000).
Mathematical description of the process of filtration started with the Darcy’s law (see Eq. (4.1)), which postulates that fluid volumetric flow rate is directly proportional to the pressure (which is the driving force of the process) applied to the filtration area and it is inversely proportional to the thickness of the porous material (Darcy, 1856):
j = kO∆"
0l (4.1)
Where: j – the volumetric flow rate of the liquid, k – the permeability constant of a porous material, A – a filtration area, ∆" – an applied pressure, 0 – the viscosity of the liquid, L – a material thickness.
Permeability K, [m2], is the measure of the ability of porous material to allow fluids to pass through it. In addition to the cake permeability, the permeability of the filter medium should also be considered. The ratio of the thickness to the filter medium permeability is called the filter medium resistance, Rm, and thus, Eq. (4.1) takes a form of Eq. 4.2 (Tarleton, Willmer, 1997):
j = O∆"
0Dm (4.2)
As it was mentioned above, the total pressure difference consists of the pressure drop through the filter medium, ∆"m, and of the pressure drop through the filter cake, ∆"#. Similar to the resistance offered by the filtration material, the filter cake resistance acts to prevent the fluid flow through it and Eq. (4.2) is again modified to Eq. (4.3) (Tarleton and Willmer, 1997):
23 j = O∆"
0(Dm+ D#) (4.3)
Where: D# – the cake resistance.
If the solid particles form an incompressible cake, the resistance depends on the mass of the cake, n#, multiplied by the specific cake resistance, $, and Eq. (4.4) can be used for the calculations:
j = O∆"
0$n# + 0Dm (4.4)
The mass of the cake can be calculated by Eq. (4.5):
n# = 2o(1 − ()Ol# (4.5)
Where: 2o – the solid density, ( – the porosity of the cake, l# – cake thickness.
The main filter cake characteristics, which are resistance, porosity and moisture content, are discussed in more details below.
4.3 Filter cake characteristics
The filter cake properties are considered as the main factors affecting the filtration performance, thus, they should be estimated correctly in order to make more accurate predictions of the filtration behaviour and to select the most suitable method and equipment for the filtration process.
4.3.1 Average specific cake resistance
As the cake builds up, the resistance offered by the accumulating cake increases. The specific cake resistance, $, is a local and time-dependent quantity and it is considered as a function of the compressive stress. The specific cake resistance should be constant for incompressible cakes but it may change with time because of the possible flow consolidation of the cake and also, in the case of variable rate filtration, because of variable approach velocity (Chen, Hsiau, 2009). Most of the cakes, however, are compressible and their specific resistance changes with the pressure
24
drop across the cake, ∆"# (Svarovsky, 2000). The value of the specific cake resistance, $, can be calculated by Eq. (4.6). Svarovsky (2000) noticed that using of this equation is limited due to its empirical nature.
$ = $%(∆"#)p (4.6)
Where: $% – the resistance at unit applied pressure drop, ∆"# – the pressure drop across the cake, n – a compressibility index which expresses the influence of applied filtration pressure on the average cake resistance.
The average specific cake resistance, $&', is considered as a stress-averaged value of the specific cake resistance across which the liquid pressure drop takes such values that the cake compressive stress varies from zero at the cake-suspension interface to the pressure equal to pressure drop across filter cake at the cake-medium interface (Tien, 2002) and can be determined with Eq. (4.7):
$&' = $%(1 − K)(∆"#)p (4.7)
The total resistance that acts to prevent the free liquid flow includes not only the cake resistance, but also the filter medium resistance, Rm. In practical applications, the medium resistance value is considered as constant. However, it varies due to the particles penetration into the medium and its blocking by them. Another reason of its changing might be the applied pressure resulted in the compression of the medium fibres (Svarovsky, 2000). As the overall pressure drop across the installed filter includes losses not only in the medium but also in the associated piping and in the inlet and outlet ports, it is convenient in practice to include all these extra resistances in the value of the medium resistance (Svarovsky, 2000).
4.3.2 Porosity, saturation and moisture content of compressible cakes
As it was mentioned previously, filter cakes can be incompressible or compressible. The permeability of an incompressible cake remains constant during the filtration procedure and is not affected by the pressure value, but only by the particle size and porosity (Rushton et al., 2000).
The formation of incompressible cakes mainly occurs during the filtration of large particles that do not have the tendency for aggregation (Wakeman and Tarleton, 1999). If the porosity of cake
25
changes with the change in the applied pressure, the cake is compressible (Rushton et al., 2000).
The compressibility of cakes can be tested by several methods, e.g., by using a permeability cell in which the mechanical action of the piston simulates the influence of the hydraulic pressure (Svarovsky, 2000). It was found empirically, that compressible materials have a value of the compressibility index n in the range between 0 and 1. The extreme zero value is related to the incompressible solids, whereas compressibility index higher than one indicates supercompressible materials. The intermediate values from 0 to 0.5 and from 0.5 to 1 are characterized slightly and moderately compressible materials respectively (Rushton et al., 2000).
With the higher pressure, the average cake porosity, (&', which is a percentage of voids in the volume of cake, reduces. The relationship between the average cake porosity and the applied pressure is demonstrated by Eq. (4.8). In practice, the cake porosity can be found by measuring the weights of the wet and dry cakes for the determination of the volumes of these cakes (Tien, 2006).
(&' = (%∆"qr (4.8)
Where: / – a compressibility index which expresses the effect of the applied pressure on the average cake porosity
The voids (pores) in the filter cake can contain liquid or gas. If the liquid occupies all the pore volume, the cake is considered as saturated (S, [%], saturation of the cake). However, the fully saturated cakes (S=100%) are rare and the moisture content is used to express the fraction of pores in which liquid is present. The moisture content can be calculated by Eq. (4.9) (Svarovsky, 2000) and moisture ratio can be found from Eq. 4.10:
sbtuvwx[ EbKv[Kv (sA) = n6
n7#100% =(n7#− nz+)
n7# 100% (4.9)
Where: ml – a mass of the liquid retained in the cake, mwc – a mass of the wet cake, mds – a mass of the dry solids in the cake which is a mass of the dry cake
sbtuvwx[ x{vtb (sD) =n7#
nz+ (4.10)
26
Concentration of solid matter (solid content by mass fraction) is calculated with Eq. 4.11:
E| = nz+
n+6 (4.11)
Where: E| – solid content by mass fraction, %; n+6− mass of the slurry, kg.
Both characteristics, MC and cW, can be easily determined since the masses of slurry and wet and dry cakes can be measured directly when conducting laboratory tests.
4.4 Cake filtration in the constant pressure mode
According to Eq. (4.12) (Rushton et al., 2000), where the filtration area, viscosity, filtration concentration and specific cake resistance have constant values for incompressible cakes, the volumetric filtrate flow rate (dV/dt) and pressure difference can be variables depending on the filtration mode:
åç
åv =O>∆é
0Eç$ (4.12)
The solution of Eq. (4.11) can be found when either one of these variable parameters remains constant. Thus, tree possible modes of the filtration procedure are used: the constant rate and constant pressure filtration and the mode in which both mentioned characteristics are varied.
Constant pressure filtration is widely used to determine the values of average specific cake resistance and filter medium resistance with laboratory filters to further upscaling of the process.
For the transition from the general filtration equation (Eq. (4.13)) to the constant pressure filtration mode, the general equation is integrated Eq. (4.14) (Rushton et al., 2000):
åv
åç= 0$&'E
O>∆é ç +0Dm O∆é
(4.13)
v
ç = 0$&'E
2O>∆éç +0Dm
OƎ (4.14)
where at the given (constant) pressure, the volumetric filtrate flow rate (dV/dt) varies and all the other parameters remain constant. Under the constant pressure the decrease in the flow rate occurs due to the growth of the cake (Rushton et al., 2000).
27
By plotting the filtration time over the filtrate volume, t/V, against the collected filtrate volume, V, a straight line (Figure 4.2) is obtained. Cake resistance, αav, and filter medium resistance, Rm, can be calculated by using the slope, a, and intercept, b, of the trend line. The following Eq. (4.15) and Eq. (4.16) are used for the calculations (Rushton et al., 2000):
$&' = 2{O>è"
µE (4.15)
Dm= Oëè"
µ
(4.16)
Figure 4.2. A simplified example of the results obtained during the constant pressure filtration mode; t/V is plotted against V for determination of the a and b coefficients (Svarovsky, 2000).
However, in the constant pressure filtration tests some initial time is spent before the operating pressure difference is achieved. This period, as well as the volume of filtrate which is collected during the period, should be taken into account. Eq. (4.17) contains the correction for these values (Rushton et al., 2000):
v − v+
ç − ç+ = 0$&'E
2O>∆é(ç + ç+) +0Dm
OƎ (4.17)
Where: ts and Vs – time and collected filtrate volume at the beginning of the constant pressure period, respectively
In this case, the graphical representation of the experimental data, which is a straight line obtained by plotting v/ç against ç, is shown in Figure 4.3. Determination of the average specific cake resistance and filter medium resistance is made as described above.
28
For compressible cakes, the compressibility index n can be determined. Series of tests should be done under different pressure differences to obtain enough data. The compressibility index can be calculated from Eq. (4.6) (Rushton et al., 2000) by plotting results of all experiments in series as
$&' = ì(∆") and using the corresponding power-type trend-line (Figure 4.4):
Figure 4.3. Determination of average specific cake resistance and filter medium resistance by using data obtained during the constant pressure filtration tests (Kinnarinen, 2014).
Figure 4.4. Determination of compressibility index for a compressible cake by using data obtained during the constant pressure filtration tests (Rushton et al., 2000).
The constant pressure filtration is the most widely used mode of filtration operation and it is applied both in laboratory and industrial scales (Smiles D., 1970). The laboratory tests are performed to investigate the properties of different slurries, formed filter cakes (porosity, cake resistance, moisture content) and filter media performance with the particular cakes (permeability/resistance). The obtained data can be used for equipment selection and in the scale- up calculations.
29
5 FACTORS AFFECTING THE CAKE FILTRATION PERFORMANCE
Pre-treatment operations, properties of the substances (fluid, solid particles, slurry and the filter media material) and operating parameters all have an influence on the pressure filtration procedure.
The properties of the formed filter cake and their influence on the process are described in the previous section while the other factors are discussed in this chapter.
5.1 Chemical pre-treatment: coagulants and flocculants
During the coagulation, particles come closer to each other to form agglomerates. Destabilisation of suspension occurs via the addition of special inorganic chemicals called «coagulants». The formed agglomerated particles have a size up to 1 mm. The coagulants act in two ways: neutralize the surface charges of the particles present in the suspension (AlCl3, Al2(SO4)3, FeSO4, etc.) or suppress the EDL (NaCl, MgSO4) (Svarovsky, 2000). Flocculants can be added to the suspension to interconnect aggregated particles and make them to form flocs with larger size (up to 10 mm) which positively affect the filtration (Moody, 1995). The non-ionic (polyacrylamides) or anionic (acrylamide-acrylate copolymers) flocculants are most widely used.
The dosage of flocculants is one of the most important parameter to be adjusted as too high dose may lead to the inhibition of flocculation since solids become surrounded by the polymer and no free space for addition of the other particles is remained. In addition, it can result in the higher SVI, filter medium blinding and rise of the treatment cost. Other factors on which one should pay attention are: solid phase concentration, type of coagulant, surface chemistry, pH, mixing energy, etc. The complex effect of these parameters obstructs the determination of the optimal chemical dosage (Svarovsky, 2000).
5.2 Addition of filter aids
In some cases, the introduction of chemicals to the system is not possible and the physical pre- treatment methods can be used instead. However, in general they are limited by several specific applications and are more expensive but can provide some improvements to the slurry or the process. Physical pre-treatment methods consist of heating, freezing, precipitation/crystallization, elutriation/thickening, solvent or surfactant addition, addition of filter aids, ultrasound, irradiation (Tarleton and Wakeman, 2007). Among all of the listed, filter aids are the most widespread. They can be introduced to the system to enhance the filtration of “problematic” suspensions that contain fine, slimy particles or low concentration of solids (Svarovsky, 2000).
30
Filter aids are made from rigid, porous and highly permeable powdered materials which can be both inorganic and organic. The former are well studied and widespread and contain such materials as diatomaceous earth (diatomite), perlite and asbestos. The latter are relatively new and represents cheap and abundant materials, which are sometimes considered as the wastes (biowaste) of the industries (Gerdes E., 1997). Cellulose, calcined rice hull, ashes, wood and plant fibres and non- activated carbon are typical examples of them (Tarleton and Wakeman, 2007).
Filter aids can be applied on the surface of the filter medium (“pre-coating” mode) or can be mixed with the feed slurry before the filtration starts (“body feed” mode) (Svarovsky, 2000). Pre-coat (surface application) allows filtering very fine particles avoiding the penetration of fines to the filter media and its blinding. Pre-coat technique is used, e.g., in rotary drum pre-coat filters. The body feed mode allows to improve the permeability of the cake formed during filtration. The required amount of the aids can be determined experimentally with laboratory filters or by modelling of the corresponding process. However, in most cases the amount of filter aids added is in the same order as the solid concentration in the suspension and can be expressed as a percentage of the particle concentration or as a ratio of the mass of filter aids to the mass of solids. The effect of mass of filter aids added to the slurry body feed on the overall throughput is schematically shown in Figure 5.1. As it is shown, too low concentration of the filter aids only decreases the filtrate recovery since the thickness of the cake gets higher whereas permeability does not increase.
With increasing the filter aids to solid content ratio the throughput increases achieving the maximum at the point that indicates the optimal ratio. After this point further increase of aid amount will lead to decrease of the throughput (Rushton et al., 2000).
Figure 5.1. Determination of the optimal concentration of filter aids in the body-feed filtration tests in terms of obtaining the maximum filtrate recovery (Rushton et al., 2000).
31
With the presence of filter aids both on the surface of the filter medium and in the slurry, it becomes possible to filter very dilute suspensions effectively, even with the concentration less than a 0.1%-v. This combining technic is mostly used when the filtration is done by the pressure filter, but some other options can exist. Whichever technique is chosen, the adjustment of many parameters is required to optimize the process. The filter aids recovery does not usually occur, although it can be profitable for the processes with the large capacity (Tarleton and Wakeman, 2007). However, the addition of filter aids increases the cost of the process and requires extra process equipment. Filter aids should not be added in cases where cakes are the product of the process since it is impossible to remove the aids from the cake (Rushton et al., 1996).
5.3 Filter media
All the filtration techniques require the filter media. According to the (Svarovsky, 2000): “The media can be considered as the heart of the filtration device or machine”. Filter medium is the material used in the filtration process that retains the substances that should be separated and allows the other components to pass. Filter media can be made from natural, synthetic, metal, glass and ceramic materials and be woven or non-woven. There are a plenty of different filtration materials which should be chosen very carefully for each specific application.
The tree main characteristics of a filter medium are its productivity, lifetime and cost. (Tarleton and Wakeman, 2007). There are numerous criteria, which determine the selection of the particular filter medium. However, the main characteristics of the media are (1) the ability to retain particles with the particular size with the required efficiency, (2) initial permeability/resistance of the media, and (3) the permeability/resistance of the used media (Tarleton and Wakeman, 2007).
The correct choice of filter medium results in high filter productivity. Some general information, such as the fact that multifilament yarns provide clearer filtrates but are more subjected to the blinding when compared to the monofilament yarns, is available. Nevertheless, the materials should be tested at the laboratory to obtain preliminary data on the filtration time and quality of the filtrate, with the pilot scale filter, or to test the whole filtration system on the industrial scale (Svarovsky, 2000).
32 5.4 Properties of the slurry
Slurry can be characterised through the liquid and solid phases and the formed suspension. Since the particle properties have a significant influence on the filtration performance, they are described in more detail (see Section 5.4). Among the liquid properties, viscosity of the fluid is the main variable and viscosity depends on the fluid nature and temperature. As it affects the filtration rate and the rate of dewatering, viscosity can be reduced by increasing the temperature. Liquid density has an impact on the separation if sedimentation or centrifuges and hydrocyclones are used (Svarovsky, 2000). The concentration of the solid phase in the feed slurry affects the rates of filtration and cake formation. The higher the feed concentration, the quicker the cake will be formed and the longer time will be required to obtain the certain amount of filtrate (Svarovsky, 2000).
5.5 Particle properties
Particle properties, namely, the particle size, shape and size distribution and interactions between the particles and surrounding liquid can vary in a wide range. These parameters are estimated to be able to make a preliminary selection of a technique of filtration.
5.5.1 Particle size and particle size distribution
A suspension that contains only large particles that have equal size and spherical shape is the perfect option to be separated by filtration. However, in practice the particles in the suspension vary in size from very thin to very large, coarse solids. According to their origin, the suspension may be homogeneous (all the solid particles are made of the same material) or heterogeneous (the particles made from different materials are represented in the composition of suspension). The size of the particles in the suspension is important to be determined as it has ф significant influence on the separation process. Whether the liquid (filtrate) or the solid is the targeted product, the information on the particle size is used to estimate the filterability and the filtrate and/or the cake quality (Bourcier et al., 2016).
Firstly, particle size is needed for the preliminary choice of the method of separation and the filter medium to match the particle and filter medium pore sizes. The slurry that contains particles smaller than the pores of the filtration material in the low concentration will undergo “depth filtration” where the fine solids are retained in the medium structure. If the particle size exceeds the size of pores of the media, then the solids retain on the surface of the media and form a filter
