- Email: [email protected]
Electrical and rheological properties of sewage sludge – Impact of the solid content
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/watres
Electrical and rheological properties of sewage sludge e Impact of the solid content C. Segalen, E. Dieude-Fauvel*, J.C. Baudez Irstea, UR TSCF, Domaine des Palaquins, F-03150 Montoldre, France
article info
abstract
Article history:
Sludge treatment is a multistep process during which sludge is mixed, pumped, thickened
Received 29 November 2014
and dewatered. The total solid content (TSC) increases from a few grams to more than a
Received in revised form
hundred grams per liter and as underlined by the existing literature, rheological charac-
18 March 2015
teristics are key criteria for sludge management. However, these characteristics remain
Accepted 6 April 2015
difficult to be determined in-situ and professionals are looking for alternative techniques to evaluate them. In that context, the potential of electrical measurements has been high-Fauvel et al., 2009, 2014). This paper investigates the additional benefits of lighted (Dieude
Keywords:
correlating both rheological and electrical properties for sludge characterization within the
Sludge
range of 1e23%TSC. On a rheological point of view, results are consistent with previous
Total solid content
literature. In parallel, electrical impedance spectroscopy allowed us to define an equivalent
Storage modulus
electrical circuit to model the sludge electrical signature. Results highlight that the circuit
Infinite viscosity
parameters follow two regimes according to the range of solid content, similarly to rheo-
Electrical impedance spectroscopy
logical properties. This work opens new insights about sludge characterization and treat-
Conductivity
ment monitoring. © 2015 Elsevier Ltd. All rights reserved.
1.
Introduction
Due to the increasing efficiency of wastewater treatment plants and the development of international water treatment policies, sludge management is becoming a major concern. The pioneering work of Dick and Ewing (1967) and O'Neil (1985) had highlighted that rheological measurements are key criteria to manage sludge treatment. The rheological behavior of sludge is affected, among other parameters (Baudez and -Fauvel et al., 2009), by the solid conCoussot, 2001; Dieude centration and has been well described in the literature, though different models are used (Seyssiecq et al., 2003; Chaari et al., 2003; Ratkovich et al., 2013; Eshtiaghi et al., 2013). It was found that sludge rheological parameters
* Corresponding author. Tel.: þ33 470 47 74 29; fax: þ33 470 47 74 11. -Fauvel). E-mail address: [email protected] (E. Dieude http://dx.doi.org/10.1016/j.watres.2015.04.008 0043-1354/© 2015 Elsevier Ltd. All rights reserved.
(infinite viscosity, yield stress, model parameters of the flow curve) increase with the concentration following either an exponential law (Dick and Ewing, 1967; Forster, 2002; Guibaud et al., 2004; Mori et al., 2006; Pevere, 2009) or a power law (Forster, 2002; Baudez, 2008; Markis et al., 2014). Baudez et al. (2011) merged these results by showing that liquid-like characteristics follow an exponential law while solid-like characteristics follow a power-law model which can only be defined above a critical concentration (Lolito et al., 1997; Forster, 2002; Baudez, 2008), when a solid network can be noticed. However, because sludge remains the residue of wastewater treatment with unpredictable composition and fluctuating concentration, its rheology cannot be summarized as a basic relationship with total solid content (Spinosa and Wichmann, 2004): a single parameter does not capture the
26
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
changes in the overall sludge characteristics and cannot be used to monitor and predict sludge behavior. Therefore, there is a need for descriptors that encompass the physical, chemical and biological parameters, reflect the changes in sludge characteristics, and predict sludge structural parameters. -Fauvel et al. (2009, 2014) explored In that purpose, Dieude an innovative approach based on the coupling of rheology and electrical impedance spectroscopy (EIS), which is a powerful tool to characterize structural properties of materials (Bonanos et al., 1987). Focusing on sludge, electrical and rheological properties are driven by common parameters: water, salts, and temperature (Forster, 1982, 2002; Seyssiecq -Fauvel et al., 2009; Baudez et al., 2013). et al., 2003; Dieude -Fauvel et al. (2009) demonstrated that the same moDieude lecular movements and interactions are probably involved in both viscous flow and charges mobility. With a simple electrolytic solution, the relationship between resistivity and viscosity is linear and straightforward. Only free ions contribute to the solution conductivity and the only source of energy dissipation is through viscous friction in the solvent. Thus, the resistivity is described by the following equation: rz6ph0
!1 X 1 zi ci Fe ri i
Where r and h0 are the solution resistivity and viscosity, respectively. zi, ci and ri are the ionic charge, concentration and radius, respectively, for ionic species i. F and e are the Faraday and unit electronic charge, respectively. This equation is basically the StokeseEinstein equation for the viscous friction coefficient of spherical species in a uniform medium. It can easily be modified in order to take into account the presence of an electrically insulating phase of a given volume fraction. For instance, in the high dilution limit, using the Maxwell mixture model for conductivity and the Einstein model for the viscosity of dilute suspensions, resistivity r and viscosity h are related to the solid content f by the following simple relationships: r ¼ r0 =ð1 fÞ h ¼ h0 ð1 þ 2:5fÞ In that case, resistivity and viscosity are still expected to be approximately linearly related to each other. The relationship between resistivity and viscosity is much less obvious in complex fluids where the interactions between non-solvent species, that are “solid matter” (neutral or charged macromolecules, colloidal particles, micellar assemblies, lipids, etc.), leads to non-Newtonian behavior and strong departure from Einstein's law, even at very low volume fraction of nonsolvent species. In spite of this, as will be shown in the following, rather simple and reliable empirical relationships between resistivity and viscosity can be obtained. It must be specified that we do not claim that the use of electrical measurements would allow the determination of quantitative rheological parameters but we assert that the evolution of a rheological parameter can be effectively followed by the evolution of an electric characteristic related to the structure. Indeed, EIS is also commonly used to characterize structure evolution phenomena (Keddam et al., 1997; Song, 2000;
-Fauvel et al. (2014) Assifaoui, 2002). In this way, Dieude showed that apparent viscosity and electrical resistivity of sludge sample during anaerobic digestion can be represented as vectors of the original sludge and the inoculum with exactly the same coordinates: hdigestate ¼ a:hsludge þ b:hinoculum rdigestate ¼ a:rsludge þ b:rinoculum Where hdigestate is the viscosity of the sludge during anaerobic digestion, hsludge the viscosity of raw sludge, hinoculum the viscosity of the inoculum, rdigestate the resistivity of the sludge during anaerobic digestion, rsludge the resistivity of raw sludge and rinoculum the resistivity of the inoculum. Thus, used in accurate conditions, this technique appears to provide good indicators of the evolution of sludge rheologgalen et al. (2015) delved this ical properties. More recently, Se approach by analyzing the impact of temperature on sludge at a single concentration. It was evidenced that pasty sludge can be summarized by an idealized electrical circuit made of insulators and capacitors, each element being closely linked with solid-like and liquid-like rheological characteristics. However, these results were only defined through the temperature dependence of a single sludge with a given solid concentration: they may not be directly extrapolated to a wide range of concentrations because solid content clearly impacts sludge rheological properties (Baudez, 2008). The aim of this paper was to go deeper in that coupling by analyzing the electrical signature of sludge at different solid contents regarding the concentration dependence of their rheological characteristics. A new electrical model was built, including both liquid-like and solid-like characteristics. This model was discussed regarding the brand new publications galen et al., 2015). devoted to sludge electrical signature (Se
2.
Materials and methods
2.1.
Materials
Sludge was sampled at the outlet of the dewatering stage at Vichy wastewater treatment plant (Allier, Centre of France) which is an activated sludge plant equipped with a draining table and centrifuge. Its initial solid content was 17.6%. First, the initial sludge was deflocculated and homogenized with a mechanical stirrer (VMI Rayneri) at 300 rpm during 2 h. Then samples at different solid contents were prepared by diluting sludge with demineralized water. Specific attention was previously paid to the choice of the diluting substrate. The use of demineralized water or supernatant did affect neither rheology nor electrical measurement (data not shown). Then, demineralized water was chosen in order not to add external charges and only focus on the initial sludge content impact. Samples were stored at 4 C for 30 days before experiments, to ensure no temporal variability, allowing us to use the same material over several days. This technique was successfully used by Curvers et al. (2009). The final exact total solid content (TSC) was determined by drying at 105 C during 24 h (ASAE standard, 1999). The
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
different samples had the following total solid contents (TSC): 0.8, 1.5, 2.2, 3, 3.7, 6.8, 8.7, 10.3, 12.9, 15.4, and 17.6%. It is well-known that sludge bacteria form extra polymeric substances (EPS), presenting a three-dimensional gel-like biofilm matrix (Wingender et al., 1999). In order to evaluate the impact of this 3D structure on the electrical properties, sludge was also dried at 60 C during 72 h to eliminate bacteria without removing any organic compounds (Derikx et al., 1994). Then, a second series of dilutions was prepared by mixing this dried sludge with water, at the same solid contents as previously (Fig. 1). This series of reconstituted samples was called ‘unstructured” samples.
2.2.
Rheological measurements
Rheological measurements were performed using a MCR301 rheometer (Anton Paar, GmbH, Austria) piloted by RheoPlus software. Coaxial cylinders (internal radius 12.5 mm, external radius 13.5 mm, height 35 mm) were used to determine both liquid-like characteristics and solid-like viscoelastic properties. Rough surfaces were used to avoid wall slip and temperature was set to 20 C. The following experimental procedure was applied: 10 min of intense preshear (800 s1) to reach a steady state. This procedure allowed us to erase material memory and to have reproducible measurements (Baudez et al., 2011); Starting from a steady state, a logarithmic decreasing ramp of shear rate, from 800 s1 to 0.1 s1 in 500 s, with 10 points per decade. This step allows us to determine the flow curve; 5 min of rest to ensure the mobile velocity is effectively equals to zero at the beginning of the next step; Strain sweep from 0.01% to 1000% at 1 Hz (780 s duration), preceded by a short plateau of a constant 0.01% strain at 1 Hz (30 s duration), to ensure an initial steady state at the initial strain and to neglect inertia effects. This step allows us to determine the linear viscoelastic region where the -Fauvel and storage modulus G' (Pa) is constant (Dieude Dentel, 2011). Measurements were done in duplicate to highlight the reproducibility of the measurements.
2.3.
Electrical measurements
Sludge electrical properties were determined using a Hioki IM3570 impedance analyzer, combined with a conductivity
27
probe WTW Tetracon 325 (cell constant 0.475 cm1). The apparatus was piloted by a specific Matlab routine which ensures the material stability by checking measurements standard deviation is smaller than 0.01% at specific frequencies (10 Hz, 10 kHz, and 1 MHz) prior to the complete measurement. Temperature (set at 20 C) was also checked to remain constant. Similarly to rheological measurements, samples were submitted to 10 min of intense preshear (800 s1) before the measurements to ensure the same initial structural state. A 50 mV sinusoidal voltage was imposed with a logarithmic frequency sweep from 4 Hz to 5 MHz (201 points). Impedance modulus jZj (Ohm) and phase q ( ) were extracted from this electrical signal by the analyzer based upon the Ohm law applied to impedance measurements. Measurements were triplicated and a standard deviation smaller than 1% was found. (Error bars were not plotted on the graphs as they appeared too small). From these data, the complex impedance diagram is obtained, representing the absolute value of the impedance imaginary part Im(Z) (Ohm) as a function of its real part Re(Z) (Ohm) with: ImðZÞ ¼ jZjsin q ReðZÞ ¼ jZjcos q
3.
Results and discussion
3.1.
Rheology
As already found in the literature (Lolito et al., 1997; Slatter, 1997; Seyssiecq et al., 2003; Mori et al., 2006; Baudez et al., 2013; Markis et al., 2014), above a critical solid content, sewage sludge is a viscoelastic shear-thinning material: it highlights both solid-like and liquid-like properties according to the shear stress it is subjected to. In the liquid regime, flow curves show that above 3% TSC, sludge are shear-thinning yield stress fluids (Fig. 2) as demonstrated by Baudez et al. (2011). The flowing behavior is well modeled by a modified HerscheleBulkley relationship of the form: t ¼ tc þ Kg_ n þ mg_ Where tc represents the extrapolated yield stress (the lowest shear stress to apply to initiate flow), K the consistency, n the consistency index and m the (infinite) Bingham viscosity.
Fig. 1 e Structured sludge (a), dried sludge obtained by drying at 60 C during 72 h (b), reconstituted non-structured sludge prepared by mixing dried sludge with water (c). Samples (a) and (c) have a solid content equal to 14%.
28
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
Fig. 2 e Dimensionless flow curve (determined for samples above the critical solid concentration).
Table 1 e Value of the modified HerscheleBulkley model parameters. Total solid content [%]
Yield stress [Pa]
Consistency [Pa.sn]
Consistency index [-]
Infinite viscosity [Pa.s]
1.2 2.4 15.4 57.4 88.2 99.3 110.3 220.6
0.72 1.34 6.47 24.04 33.5 58.9 80.3 163
0.32 0.32 0.32 0.32 0.32 0.32 0.32 0.32
0.0084 0.0134 0.0355 0.132 0.1486 0.686 1.451 3.047
3 3.7 6.8 8.7 10.3 12.9 15.4 17.6
By reducing shear stress and shear rate by respectively the yield stress and a characteristic time defined by m/tc (Baudez et al., 2011), a master-curve can be obtained (Fig. 2). This master-curve is modeled by the following equation (Table 1): t K m ¼ 1 þ g_ n þ g_ tc tc tc In agreement with the existing literature, yield stress and consistency index, which are solid-like parameters, follow a power-law with the solid content (Table 2) while the infinite viscosity follows an exponential relationship: tc ¼ aðf f0 Þb and K ¼ pðf f0 Þq m ¼ AexpðBfÞ
Table 2 e Parameters of the power law A(f¡f0)B describing the evolution of the rheological characteristics with the total solid content. Parameters Constant A [Pa ou Pa.sn] Critical solid content f0 [%] Index B [-]
Elastic modulus G'
Yield stress tc
Consistency K
0.07
0.72
0.7
2.92
2.9
2.9
4.79
2.11
1.93
Where tc is the extrapolated yield stress, K the consistency, m the infinite viscosity, f the solid content, f0 the critical solid content below which some parameters cannot be determined, and a, b, p, q, A and B the model parameters. For solid concentration smaller than 3%, the model gave no yield stress, a low consistency associated to a power-law index close to zero but an infinite viscosity higher than the one obtained for the 3% sludge (Table 3). We ascribe this latest observation to the high inertia of the measuring geometry which may bias the results (at the lowest shear rates, the corresponding shear stress was recorded negative). When neglecting the inertia by only considering only the high shear rates, a Newtonian behavior is found (Fig. 3). As we will focus on samples with a yield stress, we basically considered that below 3%, the rheological behavior can be assumed Newtonian. In parallel, as already shown by Ayol et al. (2006) and -Fauvel and Dentel (2011) when samples are submitted Dieude to small amplitude sweep, their storage modulus and loss
Table 3 e Value of the rheological parameters of the flow curve for low total solid contents. Total solid content [%] Yield stress [Pa] Consistency [Pa.sn] Power law index Infinite viscosity [Pa.s]
2.2% 0 0.661 0.055 0.018
1.5% 0 0.447 0.055 0.016
0.8% 0 0.103 0.055 0.012
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
29
Fig. 3 e Flow curve of the 1.5%TSC sample (viscosity ¼ 0.0152 Pa s).
modulus remain constant upon a plateau identified as the linear viscoelastic domain (Fig. 4). As expected and in agree-Fauvel, 2007; ment with the existing literature (Dieude Baudez, 2008; Charlou, 2014), both moduli increases with the total solid content, following a power-law above a critical value f0 from which a solid structure exists (Table 2): 0
n
G ¼ aðf f0 Þ
G00 ¼ bðf f0 Þm Experimentally, the critical value f0 can be observed from the G ¼ f(g) curves: as soon as low solid contents are reached, sludge samples show no linear regime anymore (and the
curves become noisy). Thus, from Fig. 4, we may observe that f0 is close to 3%.
3.2.
Electrical measurements
Sludge electrical properties are clearly impacted by the total solid content. Whichever the total solid content, electrical impedance diagrams look qualitatively the same (semicircle) but the higher the solid content, the smaller the amplitude of the electrical impedance diagram (Fig. 5). As commonly used in the literature, the samples electrical behavior can be modeled by equivalent circuits (Macdonald and Garber, 1977; Macdonald, 1992; Mason et al., 2002): the
Fig. 4 e Evolution of the storage modulus during strain sweep.
30
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
Fig. 5 e Complex electrical impedance diagrams in the Nyquist form.
circuit elements and connectivity are selected as far as possible to represent the physical processes thought to be of importance for the system. Focusing on how the temperature impacts the sludge galen et al. electrical signature at a single concentration, Se (2015) modeled sludge as an ideal electrical circuit, composed of two resistances and one capacitor (Fig. 6), similar to a Debye equivalent circuit (Fru¨bing, 2011) (Table 4). The R1C1 branch was assumed to be representative of the solid-like characteristics while the R2 branch was assumed to be representative of the liquid-like characteristics. The transfer function (illustrated by the impedance Z) corresponding to this circuit is the following: Z¼
" # R2 R2 R2 ðR1 þ R2 ÞC1 u R1 þ j R1 þ R2 1 þ ðR1 þ R2 Þ2 ðC1 uÞ2 1 þ ðR1 þ R2 Þ2 ðC1 uÞ2
Quite surprisingly, this model also perfectly fits the electrical signature of sludge samples for all the solid concentrations (Table 4), even in the diluted regime for which there are no solid-like characteristics. However, a deeper analysis of the circuit elements highlighted some changes between diluted and concentrated regimes. For a given composition, resistance R1 remains constant between 0.8 and 3.7%TSC then increases with the solid content (Fig. 7), the very first point being overestimated due to
experimental uncertainties linked to the fast settling of solid particles. On the contrary, the capacity C1 does not capture two distinct regimes and constantly decreases with the solid content (Fig. 8). At first sight, this result seems paradoxical galen et al. (2015) as the regarding to results obtained by Se capacitance was assumed to be representative of the solid-like characteristics. To explain this apparent deviation, let us consider a slice of cross-sectional area A and thickness x. Its capacity is given by the relationship (Coster et al., 1996): C ¼ ε0 εr
A x
Where εr is the dielectric permittivity and ε0 the permittivity of free space. According to the ClausiuseMossotti relationship (Fourie and Roland, 2000), the permittivity is closely related to the polarisability a given by: a ¼ 3ε0
εr 1 M $ ε r þ 2 Na d
With Na the Avogadro number, M the molecular mass and d the density.
Table 4 e Equivalent circuit elements for the different solid contents. Total solid content [%]
Fig. 6 e Equivalent circuit modeling sludge electrical signature.
0.8 1.5 2.2 3 3.7 6.8 8.7 10.3 12.9 15.4
R1 [Ohm]
C1 [1010.F]
R2 [Ohm]
18.96 16.65 16.16 16.43 16.26 22.34 26.74 30.35 38.51 44.93
4.07 4.05 3.96 3.98 3.81 3.56 3.45 3.53 3.24 3.26
255.8 151 111.4 89.3 85.4 57.5 48.9 45.6 37.6 34.9
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
31
Fig. 7 e Resistance R1 vs. total solid content. The two regimes are highlighted by different plots.
Fig. 8 e Capacitance as a function of the total solid content.
In that case, because we added deionized water, sludge composition remains constant whichever the solid concentration. Thus, we may consider that M/(Nad) is constant, which leads to the expression of the polarizability which appears to be quite stable for the experimental range of total solid content (Fig. 9). On a physical point of view, that means that the final position of the dipoles within the electrical field remains stable with the total solid content. From Table 4, it is also possible to calculate the dipolar relaxation frequency of the samples: f2 ¼
1 2pR1 C1
Looking at this parameter, two regimes clearly appear on the curve (Fig. 10) and they correspond to the behavior evidenced for R1. At low solid contents, below the critical value,
the frequency increases, which means that the higher the solid content of sludge, the faster the polarization. Above the critical value, the relaxation frequency clearly decreases with the solid content: the structure which becomes stronger when the solid content increases acts like a brake on particles relaxation. On a more fundamental point of view, the polarizability and the relaxation time are the key factors governing the parameters R1 and C1. Thus, it is possible to consider that: C1 finally represents the polarization via the ClausiuseMossotti relationship, which is quite stable with the total solid content. Because the nature of the particles remains the same, so does their polarizability. R1 represents how the material resists to this polarization. It is quite stable below a critical total solid content f0elec
32
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
Fig. 9 e Ratio (εr ¡ 1)/(εr þ 2), equivalent to the polarisability a as a function of the solid content.
Fig. 10 e Evolution of the relaxation frequency as a function of the total solid content.
because there is no structure and consequently no percolation threshold. Above it, solid components are entangled and R1 increases as the polarizability decreases. From the above-mentioned observations and assumptions, the increase of R1 could be connected to the structure by involving stronger and stronger interactions between the solid compounds, inducing a slowdown of the dipoles rotation. In parallel, for the same range of solid content, the polarizability of the dipoles, connected to C1, appears to be slightly impacted.
On the other branch of the electrical circuit, the resistance R2 decreases with the solid content (i.e. for a given composition, sludge conductance S2 ¼ 1/R2 increases with the total solid content). Two zones can be defined on the curves (Fig. 11): Below a given solid fraction f0elec, the conductance increases linearly with the solid content, similarly to what it is observed for an electrolyte, Above this critical value, a departure from linearity is noticed.
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
33
Fig. 11 e Conductance S2 ¼ 1/R2 vs. the total solid content (diamonds) and corresponding electrolyte behavior (straight line).
To explain this change in the conductance, conductivity measurements were performed in order to compare the usual structured sludge (Fig. 1a) with the ‘unstructured’ sludge (Fig. 1c) and to determine if sludge structure really impacts its electrical properties (Fig. 12) (Remember that the ‘structured’ sludge is activated sludge whereas the ‘unstructured’ sludge was obtained by mixing dried sludge with water: thus, due to previous drying, for a given solid content, the reconstituted sludge sample has no structure). It was found that ‘unstructured’ sludge highlights an electrolyte-like behavior through the whole range of concentration: the conductivity is a linear relationship of the solid content. Moreover, at the lowest solid content, i.e. below a critical solid content, the conductivity of ‘unstructured’ and initial sludge are similar.
This result suggests that the departure from linearity may be attributed to the sludge structure. This structure impact can be modeled by considering the total resistance R2 is the sum of the structure resistance and a bulk resistance which decreases with the solid concentration (Fig. 13). This bulk resistance can be written as the inverse of the bulk conductivity which is found to be a linear relationship with the total solid content. Thus, the total resistance R2 can be written as: R2 ¼
1 þ RS sD þ fsMS
with f the total solid content, R2 the sludge resistance, sD the conductivity of the liquid phase, sMS the conductivity of the solid content and RS the resistivity of the structure.
Fig. 12 e Conductivity as a function of the total solid content for both structured (diamonds) and non-structured (squares) sludge samples with the electrolyte behavior (dotted line). Experimentally, the critical solid content is close to 3.7%.
34
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
Fig. 13 e Evolution of the conductivities of sludge (straight line), sludge structure (large dotted line) and liquid phase (narrow dotted line).
To summarize, as for rheological properties, electrical properties also describe two different regimes: the liquid and the solid regimes, separated by a critical solid content. This critical fraction appears to be close to the critical fraction found from rheological measurements and we may assume that f0elec ¼ f0. -Fauvel et al. (2009, 2014), this result As shown by Dieude endorses the assumption of similar molecular phenomena involved in both electrical and rheological properties of sludge: Below f0, the material shows a rheological liquid-like behavior, the electrical conductivity is a linear function of the solid content as for a electrolyte and the polarizability of the material remains stable Above f0, the material shows a solid-like behavior, due to the solid network and the circuit parameters of the equivalent circuit abruptly evolve with the solid content. The conductivity experiments also highlight that the charges are not the only parameters impacting sludge electrical signature. Thus, the structure explains why the conductivity does not evolve linearly with the solid content anymore.
galen et al. (2015), the However, in the paper written by Se material used was quite concentrated (11% total solid content): as consequence this concentrated sample belongs to the concentrated regime evidenced above. We may also emphasize that when the temperature increases, despite some changes, global sludge composition remains quite stable and the charges movements could mainly related to molecules conformation (Rees, 1969), which is not the case when the solid content changes. That is why, when the solid content is sufficiently high, the impact of the liquid phase can be neglected. Thus, the variations that were observed were mainly related to the structured phase. Eventually, whichever the context, sludge electrical signature can be modeled by an equivalent circuit composed by a resistance and a capacity in parallel with another resistance. These parameters are global parameters and the resistance R2 can be decomposed into two resistances respectively related to the liquid-like properties of the material and to its solid-like properties, i.e. to the structure. Nevertheless, more work is needed to delve the related physical phenomena.
4. galen et al., 2015) focusing on the In a previous paper (Se impact of the temperature on sludge electrical and rheological properties, it was shown that sludge becomes more fluid and more conductive when the temperature increases. It was then demonstrated that there is proportionality between dissipative characteristics (electrical resistance and viscous properties) on the one hand and ‘accumulative’ characteristics (elastic modulus and capacitance) on the other hand. At first sight, those previous results do not appear consistent with the tendencies observed when the solid content is modified as a capacitance is measured even when there is no solid structure.
Conclusion
This paper investigated how sludge total solid content impacts its rheological and electrical properties in order to determine the relationship between electrical properties and sludge structure. Previous work has shown that the sludge electrical properties could be an indicator of sludge rheological properties -Fauvel et al., 2014). It was also shown that similar (Dieude molecular phenomena should be involved in these both types of physical properties. This work went further and showed that sludge has not only a dual rheological behavior but also a
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
dual electrical behavior separated by a critical solid content f0. Below this critical value sludge exhibits liquid-like properties whereas above it shows solid-like properties which are related to the structure. Future work will consist in combining this work and the galen et al. (2015) and to decompose results illustrated by Se sludge electrical structure and solid content impact at different temperatures in order to validate our hypothesis. A thermodynamic approach would also be suitable to complete this phenomenological approach. Thus, it will allow us to delve how sludge structure is related to its electrical properties.
Acknowledgments The authors would like to thank the French Agency for the Environment and the Energy Management (ADEME) for its supports.
references
de barbotines a base Assifaoui, A., 2002. Etude de la stabilite ramiques d’argiles locales e Application aux formulations ce industrielles (PhD thesis). University Hassan II Ain-Chock, Casablanca, Morocco, 134 p. Ayol, A., Dentel, S.K., Filibeli, A., 2006. Toward efficient sludge processing using novel rheological parameters: dynamic rheological testing. Water Sci. Technol. 54, 17e22. Baudez, J.C., 2008. Physical aging and thixotropy in sludge rheology. Appl. Rheol. 18 (1), 13495e13491-13495-13498. Baudez, J.C., Coussot, P., 2001. Rheology of aging, concentrated, polymeric suspensions: application to pasty sewage sludges. J. Rheol. 45 (5), 1123e1140. Baudez, J.C., Markis, F., Eshtiaghi, N., Slatter, P., 2011. The rheological behaviour of anaerobic digested sludge. Water Res. 45 (17), 5675e5680. Baudez, J.C., Slatter, P., Eshtiaghi, N., 2013. The impact of temperature on the rheological behaviour of anaerobic digested sludge. Chem. Eng. J. 215e216, 182e187. Bonanos, N., Steele, B.C.H., Butler, E.P., 1987. Applications of Impedance Spectroscopy - Characterization of Materials. Impedance Spectroscopy. John Wiley & Sons, pp. 191e237. Chaari, F., Racineux, G., Poitou, A., Chaouche, M., 2003. Rheological behavior of sewage sludge and strain-induced dewatering. Rheol. Acta 42 (3), 273e279. risation et mode lisation de Charlou, C., 2014. Caracte coulement de boues re siduaires dans un se cheur a palettes l’e se de doctorat). Universite de Toulouse. (The Coster, H.G.L., Chilcott, T.C., Coster, A.C.F., 1996. Impedance spectroscopy of interfaces, membranes and ultrastructures. Bioelectrochemistry Bioenerg. 40 (2), 79e98. Curvers, D., Saveyn, H., Scales, P.J., Van der Meeren, P., 2009. A centrifugation method for the assessment of low pressure compressibility of particulate suspensions. Chem. Eng. J. 148 (2e3), 405e413. Derikx, P.J.L., Willers, H.C., Ten Have, P.J.W., 1994. Effect of pH on the behaviour of volatile compounds in organic manures during dry-matter determination. Bioresour. Technol. 49, 41e45. Dick, R.I., Ewing, B.B., 1967. The rheology of activated sludge. J. Water Pollut. Control Fed. 39 (4), 543e560.
35
-Fauvel, E., 2007. Apport couple des proprie te s rhe ologiques Dieude lectriques pour l’optimisation du traitement secondaire et die siduaires (The se de doctorat). et de la valorisation des boues re Pierre et Marie Curie. Universite -Fauvel, E., Van Damme, H., Baudez, J.C., 2009. Dieude Improving rheological sludge characterization with electrical measurements. Chem. Eng. Res. Des. 87 (7), 982e986. -Fauvel, E., Dentel, S.K., 2011. Sludge conditioning: impact Dieude of polymers on floc structure. J. Residuals Sci. Technol. 8 (3), 101e108. -Fauvel, E., He ritier, P., Chanet, M., Girault, R., Dieude Pastorelli, D., Guibelin, E., Baudez, J.C., 2014. Modelling the rheological properties of sludge during anaerobic digestion in a batch reactor by using electrical measurements. Water Res. 51, 104e112. Eshtiaghi, N., Markis, F., Yap, S.D., Baudez, J.C., Slatter, P., 2013. Rheological characterisation of municipal sludge: a review. Water Res. 47 (15), 5493e5510. Forster, C.F., 1982. Sludge surfaces and their relation to the rheology of sewage sludge suspensions. J. Chem. Technol. Biotechnol. 32 (8), 799e807. Forster, C.F., 2002. The rheological and physico-chemical characte ristics of sewage sludges. Enzyme Microb. Technol. 30 (3), 340e345. , R., Roland, R., 2000. Die lectriques- Bases The oriques. Fourie nieur D2300. Techniques de l'Inge Fru¨bing, P., 2011. Dielectric Spectroscopy. University of Postdam. Institute of Physics. Advanced lab experiments. Guibaud, G., Dollet, P., Tixier, N., Dagot, C., Baudu, M., 2004. Characterisation of the evolution of activated sludges using rheological measurements. Process Biochem. 39 (11), 1803e1810. Keddam, M., Takenouti, H., Novoa, X.R., Andrade, C., Alonso, C., 1997. Impedance measurements on cement paste. Cem. Concr. Res. 27 (8), 1191e1201. Lolito, V., Spinosa, L., Mininni, G., Antonacci, R., 1997. The rheology of sewage sludge at different steps of treatment. Water Sci. Technol. 36 (11), 79e85. Macdonald, J.R., 1992. Impedance spectroscopy. Ann. Biomed. Eng. 20, 289e305. Macdonald, J.R., Garber, J.A., 1977. Analysis of impedance and admittance data for solids and liquids. Electrochem. Sci. Technol. 124 (7), 1022e1030. Markis, F., Baudez, J.-C., Parthasarathy, R., Slatter, P., Eshtiaghi, N., 2014. Rheological characterisation of primary and secondary sludge: impact of solids concentration. Chem. Eng. J. 253 (0), 526e537. Mason, T.O., Campo, M.A., Hixson, A.D., Woo, L.Y., 2002. Impedance spectroscopy of fiber-reinforced cement composites. Cem. Concr. Compos. 24 (5), 457e465. Mori, M., Seyssiecq, I., Roche, N., 2006. Rheological measurements of sewage sludge for various solids concentrations and geometry. Process Biochem. 41 (7), 1656e1662. O'Neil, D.J., 1985. Rheology and mass/heat transfer aspects of anaerobic reactor design. Biomass 8 (3), 205e216. Pevere, A., Guibaud, G., Goin, E., van Hullebusch, E., Lens, P., 2009. Effects of physico-chemical factors on the viscosity evolution of anaerobic granular sludge. Biochem. Eng. J. 43 (3), 231e238. Ratkovich, N., Horn, W., Helmus, F.P., Rosenberger, S., Naessens, W., Nopens, I., Bentzen, T.R., 2013. Activated sludge rheology: a critical review on data collection and modelling. Water Res. 47 (2), 463e482. Rees, D.A., 1969. Structure, conformation, and mechanism in the formation of polysaccharide gels and networks. Adv. Carbohydr. Chem. Biochem. 24, 267e332.
36
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
galen, C., Dieude -Fauvel, E., Cle ment, J., Baudez, J.C., 2015. Se Relationship between electrical and rheological properties of sewage sludge e impact of the temperature. Water Res. 73, 1e8. Seyssiecq, I., Ferrasse, J.-H., Roche, N., 2003. State-of-the-art : rheological characterisation of wastewater treatment sludge. Biochem. Eng. J. 16, 41e56. Slatter, P.T., 1997. The rheological characterisation of sludges. Water Sci. Technol. 36 (11), 9e18.
Song, G., 2000. Equivalent circuit model for AC electrochemical impedance spectroscopy of concrete. Cem. Concr. Res. 30 (11), 1723e1730. Spinosa, L., Wichmann, K., 2004. Sludge Characterization: the ^ le of Physical Consistency, pp. 59e65. Ro Wingender, J., Neu, T.R., Flemming, H.C., 1999. Microbial Extracellular Polymeric Substances. Springer, Berlin.
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/watres
Electrical and rheological properties of sewage sludge e Impact of the solid content C. Segalen, E. Dieude-Fauvel*, J.C. Baudez Irstea, UR TSCF, Domaine des Palaquins, F-03150 Montoldre, France
article info
abstract
Article history:
Sludge treatment is a multistep process during which sludge is mixed, pumped, thickened
Received 29 November 2014
and dewatered. The total solid content (TSC) increases from a few grams to more than a
Received in revised form
hundred grams per liter and as underlined by the existing literature, rheological charac-
18 March 2015
teristics are key criteria for sludge management. However, these characteristics remain
Accepted 6 April 2015
difficult to be determined in-situ and professionals are looking for alternative techniques to evaluate them. In that context, the potential of electrical measurements has been high-Fauvel et al., 2009, 2014). This paper investigates the additional benefits of lighted (Dieude
Keywords:
correlating both rheological and electrical properties for sludge characterization within the
Sludge
range of 1e23%TSC. On a rheological point of view, results are consistent with previous
Total solid content
literature. In parallel, electrical impedance spectroscopy allowed us to define an equivalent
Storage modulus
electrical circuit to model the sludge electrical signature. Results highlight that the circuit
Infinite viscosity
parameters follow two regimes according to the range of solid content, similarly to rheo-
Electrical impedance spectroscopy
logical properties. This work opens new insights about sludge characterization and treat-
Conductivity
ment monitoring. © 2015 Elsevier Ltd. All rights reserved.
1.
Introduction
Due to the increasing efficiency of wastewater treatment plants and the development of international water treatment policies, sludge management is becoming a major concern. The pioneering work of Dick and Ewing (1967) and O'Neil (1985) had highlighted that rheological measurements are key criteria to manage sludge treatment. The rheological behavior of sludge is affected, among other parameters (Baudez and -Fauvel et al., 2009), by the solid conCoussot, 2001; Dieude centration and has been well described in the literature, though different models are used (Seyssiecq et al., 2003; Chaari et al., 2003; Ratkovich et al., 2013; Eshtiaghi et al., 2013). It was found that sludge rheological parameters
* Corresponding author. Tel.: þ33 470 47 74 29; fax: þ33 470 47 74 11. -Fauvel). E-mail address: [email protected] (E. Dieude http://dx.doi.org/10.1016/j.watres.2015.04.008 0043-1354/© 2015 Elsevier Ltd. All rights reserved.
(infinite viscosity, yield stress, model parameters of the flow curve) increase with the concentration following either an exponential law (Dick and Ewing, 1967; Forster, 2002; Guibaud et al., 2004; Mori et al., 2006; Pevere, 2009) or a power law (Forster, 2002; Baudez, 2008; Markis et al., 2014). Baudez et al. (2011) merged these results by showing that liquid-like characteristics follow an exponential law while solid-like characteristics follow a power-law model which can only be defined above a critical concentration (Lolito et al., 1997; Forster, 2002; Baudez, 2008), when a solid network can be noticed. However, because sludge remains the residue of wastewater treatment with unpredictable composition and fluctuating concentration, its rheology cannot be summarized as a basic relationship with total solid content (Spinosa and Wichmann, 2004): a single parameter does not capture the
26
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
changes in the overall sludge characteristics and cannot be used to monitor and predict sludge behavior. Therefore, there is a need for descriptors that encompass the physical, chemical and biological parameters, reflect the changes in sludge characteristics, and predict sludge structural parameters. -Fauvel et al. (2009, 2014) explored In that purpose, Dieude an innovative approach based on the coupling of rheology and electrical impedance spectroscopy (EIS), which is a powerful tool to characterize structural properties of materials (Bonanos et al., 1987). Focusing on sludge, electrical and rheological properties are driven by common parameters: water, salts, and temperature (Forster, 1982, 2002; Seyssiecq -Fauvel et al., 2009; Baudez et al., 2013). et al., 2003; Dieude -Fauvel et al. (2009) demonstrated that the same moDieude lecular movements and interactions are probably involved in both viscous flow and charges mobility. With a simple electrolytic solution, the relationship between resistivity and viscosity is linear and straightforward. Only free ions contribute to the solution conductivity and the only source of energy dissipation is through viscous friction in the solvent. Thus, the resistivity is described by the following equation: rz6ph0
!1 X 1 zi ci Fe ri i
Where r and h0 are the solution resistivity and viscosity, respectively. zi, ci and ri are the ionic charge, concentration and radius, respectively, for ionic species i. F and e are the Faraday and unit electronic charge, respectively. This equation is basically the StokeseEinstein equation for the viscous friction coefficient of spherical species in a uniform medium. It can easily be modified in order to take into account the presence of an electrically insulating phase of a given volume fraction. For instance, in the high dilution limit, using the Maxwell mixture model for conductivity and the Einstein model for the viscosity of dilute suspensions, resistivity r and viscosity h are related to the solid content f by the following simple relationships: r ¼ r0 =ð1 fÞ h ¼ h0 ð1 þ 2:5fÞ In that case, resistivity and viscosity are still expected to be approximately linearly related to each other. The relationship between resistivity and viscosity is much less obvious in complex fluids where the interactions between non-solvent species, that are “solid matter” (neutral or charged macromolecules, colloidal particles, micellar assemblies, lipids, etc.), leads to non-Newtonian behavior and strong departure from Einstein's law, even at very low volume fraction of nonsolvent species. In spite of this, as will be shown in the following, rather simple and reliable empirical relationships between resistivity and viscosity can be obtained. It must be specified that we do not claim that the use of electrical measurements would allow the determination of quantitative rheological parameters but we assert that the evolution of a rheological parameter can be effectively followed by the evolution of an electric characteristic related to the structure. Indeed, EIS is also commonly used to characterize structure evolution phenomena (Keddam et al., 1997; Song, 2000;
-Fauvel et al. (2014) Assifaoui, 2002). In this way, Dieude showed that apparent viscosity and electrical resistivity of sludge sample during anaerobic digestion can be represented as vectors of the original sludge and the inoculum with exactly the same coordinates: hdigestate ¼ a:hsludge þ b:hinoculum rdigestate ¼ a:rsludge þ b:rinoculum Where hdigestate is the viscosity of the sludge during anaerobic digestion, hsludge the viscosity of raw sludge, hinoculum the viscosity of the inoculum, rdigestate the resistivity of the sludge during anaerobic digestion, rsludge the resistivity of raw sludge and rinoculum the resistivity of the inoculum. Thus, used in accurate conditions, this technique appears to provide good indicators of the evolution of sludge rheologgalen et al. (2015) delved this ical properties. More recently, Se approach by analyzing the impact of temperature on sludge at a single concentration. It was evidenced that pasty sludge can be summarized by an idealized electrical circuit made of insulators and capacitors, each element being closely linked with solid-like and liquid-like rheological characteristics. However, these results were only defined through the temperature dependence of a single sludge with a given solid concentration: they may not be directly extrapolated to a wide range of concentrations because solid content clearly impacts sludge rheological properties (Baudez, 2008). The aim of this paper was to go deeper in that coupling by analyzing the electrical signature of sludge at different solid contents regarding the concentration dependence of their rheological characteristics. A new electrical model was built, including both liquid-like and solid-like characteristics. This model was discussed regarding the brand new publications galen et al., 2015). devoted to sludge electrical signature (Se
2.
Materials and methods
2.1.
Materials
Sludge was sampled at the outlet of the dewatering stage at Vichy wastewater treatment plant (Allier, Centre of France) which is an activated sludge plant equipped with a draining table and centrifuge. Its initial solid content was 17.6%. First, the initial sludge was deflocculated and homogenized with a mechanical stirrer (VMI Rayneri) at 300 rpm during 2 h. Then samples at different solid contents were prepared by diluting sludge with demineralized water. Specific attention was previously paid to the choice of the diluting substrate. The use of demineralized water or supernatant did affect neither rheology nor electrical measurement (data not shown). Then, demineralized water was chosen in order not to add external charges and only focus on the initial sludge content impact. Samples were stored at 4 C for 30 days before experiments, to ensure no temporal variability, allowing us to use the same material over several days. This technique was successfully used by Curvers et al. (2009). The final exact total solid content (TSC) was determined by drying at 105 C during 24 h (ASAE standard, 1999). The
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
different samples had the following total solid contents (TSC): 0.8, 1.5, 2.2, 3, 3.7, 6.8, 8.7, 10.3, 12.9, 15.4, and 17.6%. It is well-known that sludge bacteria form extra polymeric substances (EPS), presenting a three-dimensional gel-like biofilm matrix (Wingender et al., 1999). In order to evaluate the impact of this 3D structure on the electrical properties, sludge was also dried at 60 C during 72 h to eliminate bacteria without removing any organic compounds (Derikx et al., 1994). Then, a second series of dilutions was prepared by mixing this dried sludge with water, at the same solid contents as previously (Fig. 1). This series of reconstituted samples was called ‘unstructured” samples.
2.2.
Rheological measurements
Rheological measurements were performed using a MCR301 rheometer (Anton Paar, GmbH, Austria) piloted by RheoPlus software. Coaxial cylinders (internal radius 12.5 mm, external radius 13.5 mm, height 35 mm) were used to determine both liquid-like characteristics and solid-like viscoelastic properties. Rough surfaces were used to avoid wall slip and temperature was set to 20 C. The following experimental procedure was applied: 10 min of intense preshear (800 s1) to reach a steady state. This procedure allowed us to erase material memory and to have reproducible measurements (Baudez et al., 2011); Starting from a steady state, a logarithmic decreasing ramp of shear rate, from 800 s1 to 0.1 s1 in 500 s, with 10 points per decade. This step allows us to determine the flow curve; 5 min of rest to ensure the mobile velocity is effectively equals to zero at the beginning of the next step; Strain sweep from 0.01% to 1000% at 1 Hz (780 s duration), preceded by a short plateau of a constant 0.01% strain at 1 Hz (30 s duration), to ensure an initial steady state at the initial strain and to neglect inertia effects. This step allows us to determine the linear viscoelastic region where the -Fauvel and storage modulus G' (Pa) is constant (Dieude Dentel, 2011). Measurements were done in duplicate to highlight the reproducibility of the measurements.
2.3.
Electrical measurements
Sludge electrical properties were determined using a Hioki IM3570 impedance analyzer, combined with a conductivity
27
probe WTW Tetracon 325 (cell constant 0.475 cm1). The apparatus was piloted by a specific Matlab routine which ensures the material stability by checking measurements standard deviation is smaller than 0.01% at specific frequencies (10 Hz, 10 kHz, and 1 MHz) prior to the complete measurement. Temperature (set at 20 C) was also checked to remain constant. Similarly to rheological measurements, samples were submitted to 10 min of intense preshear (800 s1) before the measurements to ensure the same initial structural state. A 50 mV sinusoidal voltage was imposed with a logarithmic frequency sweep from 4 Hz to 5 MHz (201 points). Impedance modulus jZj (Ohm) and phase q ( ) were extracted from this electrical signal by the analyzer based upon the Ohm law applied to impedance measurements. Measurements were triplicated and a standard deviation smaller than 1% was found. (Error bars were not plotted on the graphs as they appeared too small). From these data, the complex impedance diagram is obtained, representing the absolute value of the impedance imaginary part Im(Z) (Ohm) as a function of its real part Re(Z) (Ohm) with: ImðZÞ ¼ jZjsin q ReðZÞ ¼ jZjcos q
3.
Results and discussion
3.1.
Rheology
As already found in the literature (Lolito et al., 1997; Slatter, 1997; Seyssiecq et al., 2003; Mori et al., 2006; Baudez et al., 2013; Markis et al., 2014), above a critical solid content, sewage sludge is a viscoelastic shear-thinning material: it highlights both solid-like and liquid-like properties according to the shear stress it is subjected to. In the liquid regime, flow curves show that above 3% TSC, sludge are shear-thinning yield stress fluids (Fig. 2) as demonstrated by Baudez et al. (2011). The flowing behavior is well modeled by a modified HerscheleBulkley relationship of the form: t ¼ tc þ Kg_ n þ mg_ Where tc represents the extrapolated yield stress (the lowest shear stress to apply to initiate flow), K the consistency, n the consistency index and m the (infinite) Bingham viscosity.
Fig. 1 e Structured sludge (a), dried sludge obtained by drying at 60 C during 72 h (b), reconstituted non-structured sludge prepared by mixing dried sludge with water (c). Samples (a) and (c) have a solid content equal to 14%.
28
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
Fig. 2 e Dimensionless flow curve (determined for samples above the critical solid concentration).
Table 1 e Value of the modified HerscheleBulkley model parameters. Total solid content [%]
Yield stress [Pa]
Consistency [Pa.sn]
Consistency index [-]
Infinite viscosity [Pa.s]
1.2 2.4 15.4 57.4 88.2 99.3 110.3 220.6
0.72 1.34 6.47 24.04 33.5 58.9 80.3 163
0.32 0.32 0.32 0.32 0.32 0.32 0.32 0.32
0.0084 0.0134 0.0355 0.132 0.1486 0.686 1.451 3.047
3 3.7 6.8 8.7 10.3 12.9 15.4 17.6
By reducing shear stress and shear rate by respectively the yield stress and a characteristic time defined by m/tc (Baudez et al., 2011), a master-curve can be obtained (Fig. 2). This master-curve is modeled by the following equation (Table 1): t K m ¼ 1 þ g_ n þ g_ tc tc tc In agreement with the existing literature, yield stress and consistency index, which are solid-like parameters, follow a power-law with the solid content (Table 2) while the infinite viscosity follows an exponential relationship: tc ¼ aðf f0 Þb and K ¼ pðf f0 Þq m ¼ AexpðBfÞ
Table 2 e Parameters of the power law A(f¡f0)B describing the evolution of the rheological characteristics with the total solid content. Parameters Constant A [Pa ou Pa.sn] Critical solid content f0 [%] Index B [-]
Elastic modulus G'
Yield stress tc
Consistency K
0.07
0.72
0.7
2.92
2.9
2.9
4.79
2.11
1.93
Where tc is the extrapolated yield stress, K the consistency, m the infinite viscosity, f the solid content, f0 the critical solid content below which some parameters cannot be determined, and a, b, p, q, A and B the model parameters. For solid concentration smaller than 3%, the model gave no yield stress, a low consistency associated to a power-law index close to zero but an infinite viscosity higher than the one obtained for the 3% sludge (Table 3). We ascribe this latest observation to the high inertia of the measuring geometry which may bias the results (at the lowest shear rates, the corresponding shear stress was recorded negative). When neglecting the inertia by only considering only the high shear rates, a Newtonian behavior is found (Fig. 3). As we will focus on samples with a yield stress, we basically considered that below 3%, the rheological behavior can be assumed Newtonian. In parallel, as already shown by Ayol et al. (2006) and -Fauvel and Dentel (2011) when samples are submitted Dieude to small amplitude sweep, their storage modulus and loss
Table 3 e Value of the rheological parameters of the flow curve for low total solid contents. Total solid content [%] Yield stress [Pa] Consistency [Pa.sn] Power law index Infinite viscosity [Pa.s]
2.2% 0 0.661 0.055 0.018
1.5% 0 0.447 0.055 0.016
0.8% 0 0.103 0.055 0.012
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
29
Fig. 3 e Flow curve of the 1.5%TSC sample (viscosity ¼ 0.0152 Pa s).
modulus remain constant upon a plateau identified as the linear viscoelastic domain (Fig. 4). As expected and in agree-Fauvel, 2007; ment with the existing literature (Dieude Baudez, 2008; Charlou, 2014), both moduli increases with the total solid content, following a power-law above a critical value f0 from which a solid structure exists (Table 2): 0
n
G ¼ aðf f0 Þ
G00 ¼ bðf f0 Þm Experimentally, the critical value f0 can be observed from the G ¼ f(g) curves: as soon as low solid contents are reached, sludge samples show no linear regime anymore (and the
curves become noisy). Thus, from Fig. 4, we may observe that f0 is close to 3%.
3.2.
Electrical measurements
Sludge electrical properties are clearly impacted by the total solid content. Whichever the total solid content, electrical impedance diagrams look qualitatively the same (semicircle) but the higher the solid content, the smaller the amplitude of the electrical impedance diagram (Fig. 5). As commonly used in the literature, the samples electrical behavior can be modeled by equivalent circuits (Macdonald and Garber, 1977; Macdonald, 1992; Mason et al., 2002): the
Fig. 4 e Evolution of the storage modulus during strain sweep.
30
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
Fig. 5 e Complex electrical impedance diagrams in the Nyquist form.
circuit elements and connectivity are selected as far as possible to represent the physical processes thought to be of importance for the system. Focusing on how the temperature impacts the sludge galen et al. electrical signature at a single concentration, Se (2015) modeled sludge as an ideal electrical circuit, composed of two resistances and one capacitor (Fig. 6), similar to a Debye equivalent circuit (Fru¨bing, 2011) (Table 4). The R1C1 branch was assumed to be representative of the solid-like characteristics while the R2 branch was assumed to be representative of the liquid-like characteristics. The transfer function (illustrated by the impedance Z) corresponding to this circuit is the following: Z¼
" # R2 R2 R2 ðR1 þ R2 ÞC1 u R1 þ j R1 þ R2 1 þ ðR1 þ R2 Þ2 ðC1 uÞ2 1 þ ðR1 þ R2 Þ2 ðC1 uÞ2
Quite surprisingly, this model also perfectly fits the electrical signature of sludge samples for all the solid concentrations (Table 4), even in the diluted regime for which there are no solid-like characteristics. However, a deeper analysis of the circuit elements highlighted some changes between diluted and concentrated regimes. For a given composition, resistance R1 remains constant between 0.8 and 3.7%TSC then increases with the solid content (Fig. 7), the very first point being overestimated due to
experimental uncertainties linked to the fast settling of solid particles. On the contrary, the capacity C1 does not capture two distinct regimes and constantly decreases with the solid content (Fig. 8). At first sight, this result seems paradoxical galen et al. (2015) as the regarding to results obtained by Se capacitance was assumed to be representative of the solid-like characteristics. To explain this apparent deviation, let us consider a slice of cross-sectional area A and thickness x. Its capacity is given by the relationship (Coster et al., 1996): C ¼ ε0 εr
A x
Where εr is the dielectric permittivity and ε0 the permittivity of free space. According to the ClausiuseMossotti relationship (Fourie and Roland, 2000), the permittivity is closely related to the polarisability a given by: a ¼ 3ε0
εr 1 M $ ε r þ 2 Na d
With Na the Avogadro number, M the molecular mass and d the density.
Table 4 e Equivalent circuit elements for the different solid contents. Total solid content [%]
Fig. 6 e Equivalent circuit modeling sludge electrical signature.
0.8 1.5 2.2 3 3.7 6.8 8.7 10.3 12.9 15.4
R1 [Ohm]
C1 [1010.F]
R2 [Ohm]
18.96 16.65 16.16 16.43 16.26 22.34 26.74 30.35 38.51 44.93
4.07 4.05 3.96 3.98 3.81 3.56 3.45 3.53 3.24 3.26
255.8 151 111.4 89.3 85.4 57.5 48.9 45.6 37.6 34.9
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
31
Fig. 7 e Resistance R1 vs. total solid content. The two regimes are highlighted by different plots.
Fig. 8 e Capacitance as a function of the total solid content.
In that case, because we added deionized water, sludge composition remains constant whichever the solid concentration. Thus, we may consider that M/(Nad) is constant, which leads to the expression of the polarizability which appears to be quite stable for the experimental range of total solid content (Fig. 9). On a physical point of view, that means that the final position of the dipoles within the electrical field remains stable with the total solid content. From Table 4, it is also possible to calculate the dipolar relaxation frequency of the samples: f2 ¼
1 2pR1 C1
Looking at this parameter, two regimes clearly appear on the curve (Fig. 10) and they correspond to the behavior evidenced for R1. At low solid contents, below the critical value,
the frequency increases, which means that the higher the solid content of sludge, the faster the polarization. Above the critical value, the relaxation frequency clearly decreases with the solid content: the structure which becomes stronger when the solid content increases acts like a brake on particles relaxation. On a more fundamental point of view, the polarizability and the relaxation time are the key factors governing the parameters R1 and C1. Thus, it is possible to consider that: C1 finally represents the polarization via the ClausiuseMossotti relationship, which is quite stable with the total solid content. Because the nature of the particles remains the same, so does their polarizability. R1 represents how the material resists to this polarization. It is quite stable below a critical total solid content f0elec
32
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
Fig. 9 e Ratio (εr ¡ 1)/(εr þ 2), equivalent to the polarisability a as a function of the solid content.
Fig. 10 e Evolution of the relaxation frequency as a function of the total solid content.
because there is no structure and consequently no percolation threshold. Above it, solid components are entangled and R1 increases as the polarizability decreases. From the above-mentioned observations and assumptions, the increase of R1 could be connected to the structure by involving stronger and stronger interactions between the solid compounds, inducing a slowdown of the dipoles rotation. In parallel, for the same range of solid content, the polarizability of the dipoles, connected to C1, appears to be slightly impacted.
On the other branch of the electrical circuit, the resistance R2 decreases with the solid content (i.e. for a given composition, sludge conductance S2 ¼ 1/R2 increases with the total solid content). Two zones can be defined on the curves (Fig. 11): Below a given solid fraction f0elec, the conductance increases linearly with the solid content, similarly to what it is observed for an electrolyte, Above this critical value, a departure from linearity is noticed.
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
33
Fig. 11 e Conductance S2 ¼ 1/R2 vs. the total solid content (diamonds) and corresponding electrolyte behavior (straight line).
To explain this change in the conductance, conductivity measurements were performed in order to compare the usual structured sludge (Fig. 1a) with the ‘unstructured’ sludge (Fig. 1c) and to determine if sludge structure really impacts its electrical properties (Fig. 12) (Remember that the ‘structured’ sludge is activated sludge whereas the ‘unstructured’ sludge was obtained by mixing dried sludge with water: thus, due to previous drying, for a given solid content, the reconstituted sludge sample has no structure). It was found that ‘unstructured’ sludge highlights an electrolyte-like behavior through the whole range of concentration: the conductivity is a linear relationship of the solid content. Moreover, at the lowest solid content, i.e. below a critical solid content, the conductivity of ‘unstructured’ and initial sludge are similar.
This result suggests that the departure from linearity may be attributed to the sludge structure. This structure impact can be modeled by considering the total resistance R2 is the sum of the structure resistance and a bulk resistance which decreases with the solid concentration (Fig. 13). This bulk resistance can be written as the inverse of the bulk conductivity which is found to be a linear relationship with the total solid content. Thus, the total resistance R2 can be written as: R2 ¼
1 þ RS sD þ fsMS
with f the total solid content, R2 the sludge resistance, sD the conductivity of the liquid phase, sMS the conductivity of the solid content and RS the resistivity of the structure.
Fig. 12 e Conductivity as a function of the total solid content for both structured (diamonds) and non-structured (squares) sludge samples with the electrolyte behavior (dotted line). Experimentally, the critical solid content is close to 3.7%.
34
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
Fig. 13 e Evolution of the conductivities of sludge (straight line), sludge structure (large dotted line) and liquid phase (narrow dotted line).
To summarize, as for rheological properties, electrical properties also describe two different regimes: the liquid and the solid regimes, separated by a critical solid content. This critical fraction appears to be close to the critical fraction found from rheological measurements and we may assume that f0elec ¼ f0. -Fauvel et al. (2009, 2014), this result As shown by Dieude endorses the assumption of similar molecular phenomena involved in both electrical and rheological properties of sludge: Below f0, the material shows a rheological liquid-like behavior, the electrical conductivity is a linear function of the solid content as for a electrolyte and the polarizability of the material remains stable Above f0, the material shows a solid-like behavior, due to the solid network and the circuit parameters of the equivalent circuit abruptly evolve with the solid content. The conductivity experiments also highlight that the charges are not the only parameters impacting sludge electrical signature. Thus, the structure explains why the conductivity does not evolve linearly with the solid content anymore.
galen et al. (2015), the However, in the paper written by Se material used was quite concentrated (11% total solid content): as consequence this concentrated sample belongs to the concentrated regime evidenced above. We may also emphasize that when the temperature increases, despite some changes, global sludge composition remains quite stable and the charges movements could mainly related to molecules conformation (Rees, 1969), which is not the case when the solid content changes. That is why, when the solid content is sufficiently high, the impact of the liquid phase can be neglected. Thus, the variations that were observed were mainly related to the structured phase. Eventually, whichever the context, sludge electrical signature can be modeled by an equivalent circuit composed by a resistance and a capacity in parallel with another resistance. These parameters are global parameters and the resistance R2 can be decomposed into two resistances respectively related to the liquid-like properties of the material and to its solid-like properties, i.e. to the structure. Nevertheless, more work is needed to delve the related physical phenomena.
4. galen et al., 2015) focusing on the In a previous paper (Se impact of the temperature on sludge electrical and rheological properties, it was shown that sludge becomes more fluid and more conductive when the temperature increases. It was then demonstrated that there is proportionality between dissipative characteristics (electrical resistance and viscous properties) on the one hand and ‘accumulative’ characteristics (elastic modulus and capacitance) on the other hand. At first sight, those previous results do not appear consistent with the tendencies observed when the solid content is modified as a capacitance is measured even when there is no solid structure.
Conclusion
This paper investigated how sludge total solid content impacts its rheological and electrical properties in order to determine the relationship between electrical properties and sludge structure. Previous work has shown that the sludge electrical properties could be an indicator of sludge rheological properties -Fauvel et al., 2014). It was also shown that similar (Dieude molecular phenomena should be involved in these both types of physical properties. This work went further and showed that sludge has not only a dual rheological behavior but also a
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
dual electrical behavior separated by a critical solid content f0. Below this critical value sludge exhibits liquid-like properties whereas above it shows solid-like properties which are related to the structure. Future work will consist in combining this work and the galen et al. (2015) and to decompose results illustrated by Se sludge electrical structure and solid content impact at different temperatures in order to validate our hypothesis. A thermodynamic approach would also be suitable to complete this phenomenological approach. Thus, it will allow us to delve how sludge structure is related to its electrical properties.
Acknowledgments The authors would like to thank the French Agency for the Environment and the Energy Management (ADEME) for its supports.
references
de barbotines a base Assifaoui, A., 2002. Etude de la stabilite ramiques d’argiles locales e Application aux formulations ce industrielles (PhD thesis). University Hassan II Ain-Chock, Casablanca, Morocco, 134 p. Ayol, A., Dentel, S.K., Filibeli, A., 2006. Toward efficient sludge processing using novel rheological parameters: dynamic rheological testing. Water Sci. Technol. 54, 17e22. Baudez, J.C., 2008. Physical aging and thixotropy in sludge rheology. Appl. Rheol. 18 (1), 13495e13491-13495-13498. Baudez, J.C., Coussot, P., 2001. Rheology of aging, concentrated, polymeric suspensions: application to pasty sewage sludges. J. Rheol. 45 (5), 1123e1140. Baudez, J.C., Markis, F., Eshtiaghi, N., Slatter, P., 2011. The rheological behaviour of anaerobic digested sludge. Water Res. 45 (17), 5675e5680. Baudez, J.C., Slatter, P., Eshtiaghi, N., 2013. The impact of temperature on the rheological behaviour of anaerobic digested sludge. Chem. Eng. J. 215e216, 182e187. Bonanos, N., Steele, B.C.H., Butler, E.P., 1987. Applications of Impedance Spectroscopy - Characterization of Materials. Impedance Spectroscopy. John Wiley & Sons, pp. 191e237. Chaari, F., Racineux, G., Poitou, A., Chaouche, M., 2003. Rheological behavior of sewage sludge and strain-induced dewatering. Rheol. Acta 42 (3), 273e279. risation et mode lisation de Charlou, C., 2014. Caracte coulement de boues re siduaires dans un se cheur a palettes l’e se de doctorat). Universite de Toulouse. (The Coster, H.G.L., Chilcott, T.C., Coster, A.C.F., 1996. Impedance spectroscopy of interfaces, membranes and ultrastructures. Bioelectrochemistry Bioenerg. 40 (2), 79e98. Curvers, D., Saveyn, H., Scales, P.J., Van der Meeren, P., 2009. A centrifugation method for the assessment of low pressure compressibility of particulate suspensions. Chem. Eng. J. 148 (2e3), 405e413. Derikx, P.J.L., Willers, H.C., Ten Have, P.J.W., 1994. Effect of pH on the behaviour of volatile compounds in organic manures during dry-matter determination. Bioresour. Technol. 49, 41e45. Dick, R.I., Ewing, B.B., 1967. The rheology of activated sludge. J. Water Pollut. Control Fed. 39 (4), 543e560.
35
-Fauvel, E., 2007. Apport couple des proprie te s rhe ologiques Dieude lectriques pour l’optimisation du traitement secondaire et die siduaires (The se de doctorat). et de la valorisation des boues re Pierre et Marie Curie. Universite -Fauvel, E., Van Damme, H., Baudez, J.C., 2009. Dieude Improving rheological sludge characterization with electrical measurements. Chem. Eng. Res. Des. 87 (7), 982e986. -Fauvel, E., Dentel, S.K., 2011. Sludge conditioning: impact Dieude of polymers on floc structure. J. Residuals Sci. Technol. 8 (3), 101e108. -Fauvel, E., He ritier, P., Chanet, M., Girault, R., Dieude Pastorelli, D., Guibelin, E., Baudez, J.C., 2014. Modelling the rheological properties of sludge during anaerobic digestion in a batch reactor by using electrical measurements. Water Res. 51, 104e112. Eshtiaghi, N., Markis, F., Yap, S.D., Baudez, J.C., Slatter, P., 2013. Rheological characterisation of municipal sludge: a review. Water Res. 47 (15), 5493e5510. Forster, C.F., 1982. Sludge surfaces and their relation to the rheology of sewage sludge suspensions. J. Chem. Technol. Biotechnol. 32 (8), 799e807. Forster, C.F., 2002. The rheological and physico-chemical characte ristics of sewage sludges. Enzyme Microb. Technol. 30 (3), 340e345. , R., Roland, R., 2000. Die lectriques- Bases The oriques. Fourie nieur D2300. Techniques de l'Inge Fru¨bing, P., 2011. Dielectric Spectroscopy. University of Postdam. Institute of Physics. Advanced lab experiments. Guibaud, G., Dollet, P., Tixier, N., Dagot, C., Baudu, M., 2004. Characterisation of the evolution of activated sludges using rheological measurements. Process Biochem. 39 (11), 1803e1810. Keddam, M., Takenouti, H., Novoa, X.R., Andrade, C., Alonso, C., 1997. Impedance measurements on cement paste. Cem. Concr. Res. 27 (8), 1191e1201. Lolito, V., Spinosa, L., Mininni, G., Antonacci, R., 1997. The rheology of sewage sludge at different steps of treatment. Water Sci. Technol. 36 (11), 79e85. Macdonald, J.R., 1992. Impedance spectroscopy. Ann. Biomed. Eng. 20, 289e305. Macdonald, J.R., Garber, J.A., 1977. Analysis of impedance and admittance data for solids and liquids. Electrochem. Sci. Technol. 124 (7), 1022e1030. Markis, F., Baudez, J.-C., Parthasarathy, R., Slatter, P., Eshtiaghi, N., 2014. Rheological characterisation of primary and secondary sludge: impact of solids concentration. Chem. Eng. J. 253 (0), 526e537. Mason, T.O., Campo, M.A., Hixson, A.D., Woo, L.Y., 2002. Impedance spectroscopy of fiber-reinforced cement composites. Cem. Concr. Compos. 24 (5), 457e465. Mori, M., Seyssiecq, I., Roche, N., 2006. Rheological measurements of sewage sludge for various solids concentrations and geometry. Process Biochem. 41 (7), 1656e1662. O'Neil, D.J., 1985. Rheology and mass/heat transfer aspects of anaerobic reactor design. Biomass 8 (3), 205e216. Pevere, A., Guibaud, G., Goin, E., van Hullebusch, E., Lens, P., 2009. Effects of physico-chemical factors on the viscosity evolution of anaerobic granular sludge. Biochem. Eng. J. 43 (3), 231e238. Ratkovich, N., Horn, W., Helmus, F.P., Rosenberger, S., Naessens, W., Nopens, I., Bentzen, T.R., 2013. Activated sludge rheology: a critical review on data collection and modelling. Water Res. 47 (2), 463e482. Rees, D.A., 1969. Structure, conformation, and mechanism in the formation of polysaccharide gels and networks. Adv. Carbohydr. Chem. Biochem. 24, 267e332.
36
w a t e r r e s e a r c h 8 2 ( 2 0 1 5 ) 2 5 e3 6
galen, C., Dieude -Fauvel, E., Cle ment, J., Baudez, J.C., 2015. Se Relationship between electrical and rheological properties of sewage sludge e impact of the temperature. Water Res. 73, 1e8. Seyssiecq, I., Ferrasse, J.-H., Roche, N., 2003. State-of-the-art : rheological characterisation of wastewater treatment sludge. Biochem. Eng. J. 16, 41e56. Slatter, P.T., 1997. The rheological characterisation of sludges. Water Sci. Technol. 36 (11), 9e18.
Song, G., 2000. Equivalent circuit model for AC electrochemical impedance spectroscopy of concrete. Cem. Concr. Res. 30 (11), 1723e1730. Spinosa, L., Wichmann, K., 2004. Sludge Characterization: the ^ le of Physical Consistency, pp. 59e65. Ro Wingender, J., Neu, T.R., Flemming, H.C., 1999. Microbial Extracellular Polymeric Substances. Springer, Berlin.