- Email: [email protected]
Utilization of sludge from ready-mixed concrete plants as a substitute for limestone fillers
Construction and Building Materials 112 (2016) 790–799
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Utilization of sludge from ready-mixed concrete plants as a substitute for limestone fillers Mariane Audo, Pierre-Yves Mahieux ⇑, Philippe Turcry Laboratoire des Sciences de l’Ingénieur pour l’Environnement, Université de La Rochelle, Avenue Michel Crépeau, 17000 La Rochelle, France
h i g h l i g h t s Environmental impacts of sludge coming from ready-mixed concrete plants. Substitution of limestone fillers by sludge coming-from ready-mixed concrete plants in mortars formulation. Mechanical characterization of mortars made with sludge coming from ready-mixed concrete plants.
a r t i c l e
i n f o
Article history: Received 12 May 2015 Received in revised form 12 January 2016 Accepted 16 February 2016
Keywords: Ready-mixed concrete plant sludge Waste management Environmental impact Mortar design Physical characterization Mechanical characterization
a b s t r a c t This study deals with the incorporation of sludge coming from ready-mixed concrete plants into mortars. Preliminary environmental investigations, made through leaching tests, showed the importance of managing those waste as they can be potentially pollutant regarding to their arsenic and chromium contents. Thus, management of the sludge can be environmentally and economically difficult. Reincorporation of those sludge into a closed loop concrete production is of particular interest. Also, it represents an interesting way to save raw materials (water, sand and limestone fillers). Yet, two main disadvantages were observed when using those sludge as limestone fillers substitute: – a decrease in the workability of the fresh state mortar, calling for a higher superplasticizer content; – a variability in the compressive strength of the hardened state mortars, between 30% and +17% when comparing to a reference mortar. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction European ready-mixed concrete production has notably increased over the last decades. Since 2010, around 370 million cubic meters of ready-mixed concrete are produced each year in Europe, of which around 10% come from the French production [1]. From this high production results a high level of by-products. Indeed, it has been estimated that a 9-m3 truck contains daily around 300 kg of returned concrete [2]. Generally, this leftover concrete is discharged in large containers. The hardened concrete can be then crushed and easily used as recycled aggregates for road construction. After the leftover being discharged, the truck is washed out with huge amounts of water, up to 1300 L per truck, and the suspended matter is allowed to settle in large sedimentation basins. Wash-waters from mixers are also directed to those basins. Several studies have been focused on the reuse of the clarified water of those basins, which presents a high up to 12 pH. This water can be partially reused for trucks washing or in concrete pro⇑ Corresponding author. E-mail address: [email protected] (P.-Y. Mahieux). http://dx.doi.org/10.1016/j.conbuildmat.2016.02.044 0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.
duction [3,4,5]. Su et al. showed that, according to the total suspended solids content, shorter setting times and a lower flowability are obtained [3]. On the opposite, Chatveera et al. highlighted a time lag when using sludge water and no additives nor admixtures, but reported a flowability decrease as well [5]. In another study, the use of the wet sludge as partial cement or sand substitute leads to a decrease in the compressive strength [6]. Recently, a study was led on the use of the sludge as a new raw material for Portland clinker production [7]. The variability in the chemical composition of the sludge leads to an impossibility in their reuse for Portland clinker production, due to their high chemical variations, as well as their high alkali, SO3 and MgO contents. Yet, not many studies have dealt with the utilization of the settled sludge, made of the fines particles (cement, mineral additions, etc.) as well as sand and aggregates. Though, it represents a large available quantity of raw material. As around 1 m3 of wet sludge is created by the production of 90 m3 of concrete, around 4 millions of cubic meters are produced every year in France [8]. Wet sludge is collected several times a year at the bottom of the basins and stored on the ready-mixed concrete plant area so that the water content decreases. When the water content is low enough, the
791
M. Audo et al. / Construction and Building Materials 112 (2016) 790–799
material can disposed of at controlled landfills for inert wastes, regarding to the French decree on passive wastes [9]. Up to now, not many characterizations of this sludge have been led in the literature. However, their hazardous behavior has already been highlighted, due to their high alkalinity and their high heavy metals and metalloid elements content [10]. All these factors lead to a high whole total disposal cost of the sludge. That is why new environmental friendly strategies must be found for sludge managing. We herein propose a new original way to valorize sludge as limestone fillers substitutes in a closed loop concrete production, avoiding any environmental and economical impacts. Nevertheless, for now, the European standard does not allow the utilization of non-standard mineral addition in concrete [11]. Thus, the objective of this study is to prove the feasibility of incorporating sludge into concrete. To reach that aim, the sludge coming from four French ready-mixed concrete plants were firstly characterized from a chemical and physical point of view. The activity index of the dry sludge, as well as their chemical activities were determined. Giving those results, utilization of the studied sludge into mortars compositions was studied using the concrete equivalent mortar concept [12]. Compressive strength and porosity were studied at various due dates, and relationship between mortars composition and compressive strength was stated. 2. Experimental program 2.1. Raw materials The studied sludge were sampled at four ready-mixed concrete plants (C1, C2, C3, C4) located in the Poitou-Charentes area, France (Fig. 1). Those four plants produce basically the same main concrete, whose formulation will be given later. The samples had a high water content, between around 50 and 120%, depending on the weather conditions and the storage duration. The moisture contents are presented in Table 1. The raw materials were then split in two fractions: one finer than 100 lm (fraction A) and one coarser than 100 lm (fraction B). Limestone filler (designed as LF), CEM II/A 42.5 and CEM I 52.5 cements were provided by Carmeuse (Saint Porchaire, France), Calcia (Airvault, France) and Lafarge (Saint Pierre La Cour, France) companies, respectively. Their properties are given in Table 2. 0/4 mm sand from Sablimaris Pallice (La Rochelle, France) was used in the studied mortars. For the determination of activity index, a standard sand from the SNL Company (Leucate, France) was used. Chrysofluid Optima 220 superplasticizer was used for the fabrication of mortars.
2.2. Characterization of raw materials 2.2.1. Leaching tests Leaching tests were performed following the French standard dedicated to sludge characterization [13]. Basically, a test portion of sludge containing 90 g of dry matter is poured into a 1L HDPE flask. Deionized water is added so that the total water content is 900 mL. The whole mixture is kept under continuous mechanical
Fig. 1. Picture of the raw materials coming from the C4 ready-mixed concrete plant.
Table 1 Water contents of the sludge
Water content (%)
C1 47.1
C2 99.5
C3 120.3
C4 155.2
stirring for 24 ± 0.5 h (Fig. 2). The mixtures are then filtered over a 2 lm filter with a Büchner apparatus and the filtrate is recovered for further analysis. Right before the analysis, samples were filtered on 0.45 lm syringe filter. All the leachates are diluted tenth and hundredth in a 5% HNO3 solution (77 mL of HNO3 – FisherScientific, Trace Metal Analysis quality – completed to 1 L with milli-Q water). Ba, Cr, Cu, Mo, Ni, Pb and Zn were analyzed with a Varian Vista Pro ICPOES and As, Cd and Se with a ThermoFischerScientific Xseries2 ICPMS. The quantification limits (lg L1 solution) were 2 (As), 2 (Se), 0.1 (Cd), 50 (Ba), 10(Cr), 10 (Mo), 10 (Ni), 10 (Pb) and 50 (Zn). To check the analytical data precision, all samples were analyzed in duplicate. Ionic chromatography was also performed on the leachates. A Metrohm apparatus with a Metrosep A Supp 5 100/4.0 column and an automatic sample changer was used. A Na2CO3 3.2 mM/NaHCO3 1.0 mM (1:1, v/v) eluent was used, with a 0.7 mL/min flow. F, Cl and SO2 4 were so quantified. Sodium salts from SigmaAldrich company were used as standards. The quantification limits were 0.5 mg L1 for F, Cl and SO2 4 .
2.2.2. Physical and mineralogical characterization Water content of sludge (defined as mass ratio of water over dry matter) was determined by drying at 80 °C until constant weight. Specific surfaces of the dry raw materials were determined with a Blaine apparatus, allowing the measurement of the resistance of the air passing through a porous bed of powder [14]. The densities were determined using a water pyknometer according to the European Standard NF EN 1097-7 [15]. The particle-size distribution of the dry sludge was determined by sieving 500 g of powder at 4, 2, 1.25, 0.5, 0.25, 0.125 and 0.100 mm. The 100 lm passing fraction was characterized by dynamic light scattering with a CILAS 1190 apparatus, used in dry mode. Calculations were performed using the Mie theory [16]. Powder X-Ray Diffraction (XRD) was carried out with a Brücker diffraction instrument, with Cu Ka1 radiation. Measurement range was from 5 to 70° 2b, with a 0.02° step. Identification of the peaks was performed by comparison to a database references. Thermogravimetric analysis (TGA) was performed on a Setaram Setsys Evolution 16/18 apparatus. Around 100 mg of samples were heated from 20 °C to 1000 °C at a 10 °C/min heating rate, under neutral argon atmosphere. To get more accurate results, data were analyzed through the differential thermogravimetric curves. ICP-AES analysis were performed sludge materials after their drying at 80 °C and after being grinded to 80 lm. Basically, around 100 mg of dry materials was digested using 4 mL of a 67–70% HNO3 – 34–37% HCl 2:2 (v/v) solution (FisherScientific, Trace Metal Analysis quality). Acidic digestion of the samples was carried out overnight at room temperature. Each sample was completed to 50 mL with milliQ water. Al, Ca, Fe, S and Si were analyzed with a Varian Vista Pro ICPOES. The quantification limits (lg g1 dry weight) were 500 (Ca), 100 (Al), 50 (Si), 10 (S) and 5 (Fe).
2.2.3. Determination of sludge activity The potential activity of sludge was evaluated through compressive tests on standard mortars. We focused on the sludge fraction lower than 100 lm which should contain the most reactive elements. For this purpose, an activity index was determined according to the French standard dedicated to limestone additions [17]. A mortar made of CEM I cement, deionized water and standard sand was made and used as reference. A second mortar was made by substituting 25% of the cement mass by the 100 lm passing fraction of the sludge. Mortars compositions, as well as the water-to-cement ratio (W/C) are given in Table 3. Compressive strengths were determined on 28-day old mortars kept in water. The activity index (denoted i) of the sludge was defined as the ratio between the compressive strength of the mortar with sludge and the one of the reference mortar. The activity of sludge was also investigated through ionic conductivity measurement. This test is based on the measurement of time-evolution of the ionic concentrations during the cement hydration in aqueous suspension. As done in the case of activity index, only the 100 lm passing fraction of sludge was used. The ionic conductivity time-evolution of the cement and blends (‘‘limestone filler + cement” and ‘‘dry sludge + cement”) were obtained by measurements in dilute medium. In the blends, the ratios between cement and dry sludge or limestone filler were the same as the ones used for the CEM mortars. The same cement as the one used in the concrete formulation was used. The composition of the blends is given in Table 4. Briefly, 300 mL of deionized water are poured into a cell equipped with a conductivity probe. The whole system is kept at 25 °C and under perpetual mechanical stirring. A 1–5 solid-to-liquid ratio is applied, which allows a good sensitivity, as
792
M. Audo et al. / Construction and Building Materials 112 (2016) 790–799
Table 2 Properties of cements and limestone filler. Physical properties
CEM I 52.5 CEM II/A 42.5 LF
Clinker composition (%)
Density (g/cm3)
Blaine (cm2/g)
Clinker (%)
CaO
SiO2
Al2O3
Fe2O3
SO3
CaCO3
MgCO3
3.15 3.12 2.70
3650 3800 3530
98 92 –
64.5 63.0 –
20.1 18.7 –
5.0 4.5 –
3.1 3.2 –
3.3 2.7 –
– 6.0 95.0
– – 3.7
Fig. 2. Picture of the mechanical stirring system.
Table 3 Compositions of mortars used to determine activity index.
Standard sand (g) CEM I 52.5 cement (g) Dry sludge (<100 lm) (g) Deionized water (g) W/C
Standard mortar (reference)
Standard mortar with sludge
1350.0 450.0 – 225.0 0.50
1350.0 337.5 112.5 225.0 0.66
well as a short reaction time. At the zero time, 75 g of cement, cement and filler or cement and dry sludge are added and the conductivity is measured every 5 s during the first 30 min and every two minutes during the last 7h30 [18].
2.3. Study of mortars made with sludge 2.3.1. Mortars mix proportioning Mortars were designed using the concrete equivalent mortar concept [12]. The latter is based on the substitution of the coarse aggregates of concrete by a certain quantity of sand in the mortars, which develops the same area as the coarse aggregates do. This method was firstly developed in order to easily detect compatibility troubles between cements and superplasticizers. It is also a convenient method for comparison studies in laboratory. The mortars compositions were obtained from a reference concrete mixture given in Table 5. The latter corresponds to a ‘‘XC1 C20/25” concrete according to European standard [11], with W/C = 0.6. It is usually produced by the four readymixed concrete plants where the studied sludge come from. The equivalent mortar derived from the reference concrete is denoted CEM in Table 6. Mortars were also
Table 4 Composition of the blends for ionic conductivity measurements.
Cement Limestone filler C1 C2 C3 C4
Carbonates content (%)
CEM II/A (g)
Sludge (fraction A) (g)
75.00 67.53 67.99 67.74 69.15 69.74
– 7.47 7.01 7.26 5.85 5.26
designed with sludge. Firstly, it was chosen to use dry sludge, even if from an industrial point of view, using wet sludge would avoid any costly drying process. At a laboratory scale, using dry sludge allowed us a better mastering of the water content in the mortars. Secondly, the mix proportions were derived from the CEM mortar. Dry sludge was used as a substitute of all the limestone filler and of part of the sand contained in the CEM mortar. The mass ratios between the dry sludge and the sand in the sludge-made mortars were determined by setting the particle-size distribution of those mortars identical as much as possible to the one of the CEM mortar. To determine the masses of dry sand and dry sludge, a second condition was applied to close the equations system: the volume (i.e. sand + mineral addition + cement + water) of both mortars was fixed equal. The dry sludge-made mortars compositions (denoted C1-CEM, C2-CEM, C3-EM and C4-CEM) are given in Table 6. Mortars were also designed with wet sludge to monitor the influence of the water content of the sludge on the properties of the mortars (Table 7). Tap water contents were adjusted following the sludge water contents. The sludge was mixed with cement and sand for 30 s at low speed [11]. Tap water was then added and the blend was mixed for additional 30 s at low speed. After scraping, another 30 s mix were applied at high speed [11]. Some superplasticizer (Chrysofluid Optima 220) was added so that the slump at the Abrams minicone test was the same as the CEM mortar one. A vibration table was used to fill the 4 4 16 cm molds with the mortar. After removal from the moulds (24 hours after their manufacturing), all the mortars were kept into water at 20 °C until their later characterization.
2.3.2. Mortars characterization Fresh mortars were characterized by slump measurement with a cylinder (H = 100 mm, d = 50 mm). Heat releases determination were performed following a protocol adapted from the semi-adiabatic Langavant method [19]. This method consists in the introduction of a reference fresh mortar (denoted CEM’, Table 6) and the studied fresh sludge-based mortar into two separates 800 cm3-Dewar flasks and the measurement of the difference of temperature in the core of the mortars at any time. In the reference CEM’ formulation, no limestone filler was used. Indeed, it had been substituted by the quantity of CEM II cement having the same hydraulic activity. It has to be noted that the mortars were made without any superplasticizer addition. Indeed, it is known that chemical admixtures can modify hydration kinetics [20] and we herein want to monitor the inherent influence of sludge on hydration. Compressive strength was determined at the ages of 2, 7 and 28 days using the 4 4 16 cm prisms, with a Zwick press (loading rate = 2200 N/s and preload = 3000 N). 4 measurements were performed on each mortar formulation. Experimental variability will be indicated as error bars on graphs. Porosity was measured at the age of 28 days by water soaking following the recommended method [21].
3. Results and discussion 3.1. Sludge characterization 3.1.1. Leaching behavior The results of the chemical analysis led on the leachates are presented in Table 8. Concerning the anions contents, fluorides, chlorides and sulfates contents comply with the limit values of the Table 5 Compositions of reference concrete (kg per m3). Weight (kg/m3) Gravel 10/20 Gravel 6/10 Sand 0/4 Cement CEM II/A Limestone filler Water Superplasticizer (%)
823 352 775 253 28 169 0.6
793
M. Audo et al. / Construction and Building Materials 112 (2016) 790–799 Table 6 Compositions of the studied mixtures with dry sludge (kg per m3).
Sand 0/4 Cement CEM II/A Limestone filler Dry sludge Water Superplasticizer (%)
CEM mortar
C1-CEM mortar
C2-CEM mortar
C3-CEM mortar
C4-CEM mortar
CEM’ mortar
1476 424 47 0 283 0.2
1290 424 0 185 283 0.6
1391 424 0 109 283 0.5
1335 424 0 149 283 0.8
1427 424 0 79 283 0.9
1476 436 0 0 283 0
Table 7 Compositions of the studied mixtures with wet sludge (kg per m3).
Sand 0/4 Cement CEM II/A Limestone filler Wet sludge Water Superplasticizer (%)
CEM mortar
C1-CEM mortar
C2-CEM mortar
C3-CEM mortar
C4-CEM mortar
1476 424 47 0 283 0.2
1290 424 0 272 144 1.6
1391 424 0 218 174 2.0
1335 424 0 328 104 1.2
– – – – – –
European decision concerning the acceptance of waste at landfills. Concerning the heavy metals and metalloids content, special attention must be paid to the arsenic and chromium elements, as the detected values are sometimes slightly higher than the limit values for inert waste [22]. Consequently, the sludge cannot be surely stated inert materials. This point is of particular importance from an environmental and economical point of view. Indeed, the price concerning the waste storage is directly depending on the category they belong to: 8 €/t for inert waste over 40 €/t for non-dangerous waste [23]. Nevertheless, the measured suspicious concentrations are right above the inert waste acceptability limit. Thus those results need to be completed by analyzing new sludge. Regarding to those results, finding a way to avoid storage of sludge becomes of particular interest. Thus, incorporation into cement matrix could be a promising, innovative and easy way to valorize them. 3.1.2. Physical and mineralogical properties The particle-size distributions analysis of the four dry sludge showed a continuous material containing between 20 to 40% of particles finer than 100 lm (Fig. 3, sieving). Nevertheless, the particle-size distribution of the fractions finer than 100 lm remains almost the same for all the sludge (Fig. 4, dynamic light scattering). Moreover, this distribution is very closed to the one of limestone filler. Indeed, two of the four modes of the filler are observed on the sludge. The dmax of all sludge being less than
8 mm, it makes them a priori convenient for a use as sand and mineral addition substitute. Measurements of density were led on those two fractions (Table 9). The results showed that both fractions have a lower density (around 2.2) than the limestone fillers (2.7). Compared to fraction B, the fraction A slightly lower density can be explained by a higher cementitious materials content and a lower sand content. Concerning the specific surface, the one measured on the fraction A is up to 5 times higher than the one of the limestone fillers, which could be explained by a higher open porosity or different morphology. XRD patterns were also acquired on each fraction (Fig. 5). Basically, the same peaks were observed for all samples, but with various proportions. The coarsest material is mainly made by aggregates. Indeed, quartz (SiO2) and limestone (CaCO3) were observed, as well as aggregates formed by portlandite (Ca(OH)2) and/or C–S–H conglomeration. The composition of the finest fraction is more complex. Quartz, limestone and portlandite were identified as main constituents, but gypsum (CaSO42H2O), as well as anhydrous (2CaOSiO2; 3CaOSiO2) and hydrated calcium silicates were also observed. The presence of anhydrous calcium silicates show a potential residual hydraulic activity. While no ettringite (6CaOAl2O33SO332H2O) was identified, some hydrated calcium carboaluminates (3CaOAl2O3CaCO311H2O) have been detected. In literature, it has already been shown that this kind of compounds can be formed during the hydration of cement
Table 8 Sludge leachates composition (g/kg of dry matter) and criteria of acceptance of waste at landfills [22] (in bold: inert waste value exceeded). Up: Heavy metals and metalloid elements contents; Down: Anions contents.
C1-sludge C2-sludge C3-sludge C4-sludge Inert waste Non-dangerous waste
C1-sludge C2-sludge C3-sludge C4-sludge Inert waste Non-dangerous waste
As
Ba
Cd
Cr
Cu
Mo
Ni
Pb
Se
Zn
0.54 0.15 <0.02 <0.02 0.5 2
12.53 17.70 13.41 14.16 20 100
<0.001 <0.001 <0.001 <0.001 0.04 1
0.37 0.31 0.64 0.12 0.5 10
0.23 0.32 <0.11 0.31 2 50
<0.11 <0.11 <0.11 <0.11 0.5 10
<0.11 <0.11 <0.11 <0.11 0.4 10
<0.11 <0.11 <0.11 <0.11 0.5 10
<0.02 <0.02 <0.02 <0.02 0.1 0.5
<0.56 <0.56 0.71 <0.56 4 50
F
Cl
SO2 4
<8.37 <4.61 <8.77 <8.76 10 150
39.91 49.91 182.81 39.47 800 15,000
5.65 13.27 5.07 3.84 1000 20,000
794
M. Audo et al. / Construction and Building Materials 112 (2016) 790–799
determined by TGA. Regarding to the total portlandite content, fractions A present lower contents, which can be explained by the conglomeration of portlandite, which forms particles coarser than 100 lm. Global chemical compositions of the sludge are given in Table 10. Weak variability is observed between the four sludge (standard deviation below 20%). Those results can be explained by the homogeneity of the concrete produced in the four studies ready-mixed concrete plants. Indeed, they mainly produced the same concrete (formulation in Table 6). Those compositions are consistent with literature data on sludge chemical composition [4,5,6,7].
Fig. 3. Particle-size distribution of the dry sludge.
containing limestone (such as CEM II type cement), from the timeconversion of hemicarboaluminates [24]. As the studied readymixed concrete plants mainly use a CEM II/A cement with limestone addition, the presence of hydrated calcium carboaluminates in the sludge was expected. Limestone and portlandite contents were both determined with TGA. After differentiation, the typical curve showed four main peaks (Fig. 6). The first two ones, observed between 25 °C and 250 °C result from the free water evaporation, as well as the water loss coming from the dehydration of the C–S–H and calcium carboaluminates. Around 460 °C is observed the signal due to the dehydroxylation of portlandite. The last peak is due to the CO2 loss, corresponding to the decarbonation of the limestone. The integration of those peaks allowed us to determine the mass losses and consequently, the calcium carbonate and portlandite contents (Table 9). The weak peaks near 300 °C and 400 °C may be attributed to polymeric additives decomposition. The latter peak may also be attributed to the quartz dehydroxylation. Fractions A present high content of CaCO3, regarding to the total limestone content of the sludge, which is likely due to the limestone filler and cement being the main CaCO3 inputs. On average, the total limestone content of the sludge is around 11.2%. This mass percentage in the mixture can be explained by considering the limestone coming from a usual concrete formulation (reference concrete, Table 6) produced in the ready-mixed concrete plant. Indeed, considering no gravels in the sludge, a 12.6% content of calcite in the dry sludge (coming from the cement, the limestone fillers and the sand) was expected, which is closed to the content
3.1.3. Reactivity As can be seen in the Table 11 and the Fig. 7, the substitution of 25% of the cement by a mineral addition causes a decrease of the mechanical strength of the standard mortar. Indeed, a compressive strength of 37.52 MPa was obtained on the reference mortar, the compressive strength of mortars with mineral addition were included between 29.20 and 34.73 MPa. However, the French standard [17] stipulates that the activity index of a limestone fillers must be higher than 71%. The activity index of the used filler is around 92%, which is in compliance to the standard. The activity index of the C1, C3 and the C4 sludge (fractions A) are slightly lower than the one of the limestone filler. Nevertheless, they remains higher than the required 71%. The activity index of the C2-sludge (fraction A) reaches a value of 94%, which is even slightly higher than the value obtained with the limestone fillers. Overall, few dispersion is observed when comparing the activity index of the finer fraction of the sludge, which can be explained by a very close mineralogical and chemical composition. The differences may be attributed to a variation in the particle-size distributions, so to a variation in the packing densities. In order to get more accurate results on the chemical activity of the sludge, conductimetric tests were performed on blends between cement and the fraction A of the sludge. The composition of the blends is given in the experimental section. The results highlighted a classical behavior of cementitious materials [18]. During the first minutes, an exponential rising of the conductivity is observed, due to the high dissolution rate of the compounds of the clinker and gypsum. This period is known as the mixing period (I). Then, the dormant period (II) begins. The C–S–H and ettringite precipitation begins, while the conductivity remains rising. Gradually, the solution becomes richer in Ca2+ and HO ions, up to the point where the critical concentrations are reached and the portlandite precipitation begins. At that
Fig. 4. DLS particle-size distribution of the 100 lm passing of the dry sludge and comparison with a limestone filler (left: non-cumulative passing; right: cumulative passing).
795
M. Audo et al. / Construction and Building Materials 112 (2016) 790–799 Table 9 Density, specific surface and CaCO3 and Ca(OH)2 contents of the fraction A and B – Fractions A and B mass content in sludge.
3
Density (g/cm ) Specific surface (m2/kg) Ca(OH)2 content (%) CaCO3 content (%)
Fraction A (%) Fraction B (%)
LF
C1 (A)
C1 (B)
C2 (A)
C2 (B)
C3 (A)
C3 (B)
C4 (A)
C4 (B)
2.7 350 – 98.0
2.2 1100 8.8 17.5
2.2 – 10.8 13.0
2.2 1300 7.5 18.7
2.2 – 13.9 23.5
2.1 1800 6.0 22.2
2.3 – 10.8 13.4
2.2 1350 4.6 19.6
2.3 – 14.7 9.0
C1
C2
C3
C4
20.9 79.1
40.0 60.0
26.9 73.1
38.6 61.4
point, the conductivity starts to decrease: this is the acceleration period (III). The curve obtained on the CEM II/A cement is shown on Fig. 8 for example. The limestone filler, as well as fractions A of the dry sludge, does not show any hydraulic behavior. Indeed, when placed in an aqueous media, the conductivity increases during the first seconds and then remains stable until the end of the experiment (Table 12). But a large difference in the chemical behavior of the sludge and the fillers can be observed, as the maximal conductivity is at least ten times higher when sludge are used over fillers. Nevertheless, when mixed with cement, all studied materials modify the hydration kinetic of the cement. Especially, when the limestone filler delays the portlandite precipitation (30 min for a reference time of 210 min), the fraction A of the sludge speed up the portlandite precipitation (between 10 and 60 min). Those time gaps were higher to the accuracy of the method (estimated to be 4 min). Ten, those results highlight a chemical activity of the sludge within the
cementitious matrix, which is likely due to a high dissolution rate in water. Indeed, the higher is the maximum conductivity in presence of sludge, the quicker is the cement hydration. In concrete, limestone fillers are known to speed up the hydration kinetics [25]. This effect cannot be observed here, as the dilute medium does not allow to highlight the physical behavior of filler additions. 3.2. Mortars made with sludge 3.2.1. Mortars composition In Figs. 9 and 10 are presented the volume proportions of the reference concrete, the reference CEM, and the dry sludge-CEM and wet sludge-CEM. It can be noted that, depending on the particle-size distribution of the sludge, between 4.5 and 11% of the sand of the CEM mortar can be substituted by the sand brought by the sludge. Moreover, the mass of dry sludge with particle size lower than 100 lm in the sludge-based mortar is almost equal to
Fig. 5. XRD patterns of the fractions A and B of the C1- (up, left), C2-sludge (up, right), C3-sludge (down, left) and C4-sludge (down, right) (C: Calcite; CC: Hydrated calcium carbonates; CS: Calcium silicates; CSH: Hydrated calcium silicates; G: Gypsum; P: Portlandite; Q: Quartz).
796
M. Audo et al. / Construction and Building Materials 112 (2016) 790–799
Fig. 6. Typical DTG curve (example on the fraction A of the C3 sludge). Fig. 8. Conductimetric curve obtained with the CEM II/A cement. Table 10 Chemical composition of sludge (%).
C1 C2 C3 C4
Al2O3
Fe2O3
CaO
SiO2
SO3
1.70 2.09 1.65 1.97
1.45 1.72 1.42 1.62
31.79 37.32 32.83 34.85
57.04 50.04 54.85 53.50
0.59 0.98 0.81 0.71
Table 11 Compressive strength of the standard mortars.
fc(28) (MPa)
Reference 37.52
Limestone filler 34.73
C1 33.38
C2 35.13
C3 29.20
C4 30.21
Table 12 Conductimetric characterizations of the CEM II cement mixed with different mineral additions.
Cement Limestone filler C1 dry sludge C2 dry sludge C3 dry sludge C4 dry sludge Cement + limestone filler Cement + C1 dry sludge Cement + C2 dry sludge Cement + C3 dry sludge Cement + C4 dry sludge
rmax (lS/cm)
Slope reversal time (min)
19.8 0.1 5.5 6.5 1.7 1.1 18.3 17.7 18.2 18.1 18.1
210 – – – – – 242 180 148 188 200
Fig. 7. Comparison of the activity of the limestone fillers and the fraction A of the sludge.
Fig. 9. Volume proportions of the reference concrete, the CEM mortar and the C1-, C2-, C3- and C4-CEM mortars made with dry sludge.
the mass of limestone filler in the CEM mortar. Also, using wet sludge allows to substitute tap water, from 37% (C4-CEM) to 56% (C3-CEM). The particle-size distribution of the CEM mortar is presented in Fig. 11(a). In Fig. 11(b) the relative differences between the particle-size distribution of the CEM mortar and the ones of the sludge-made mortars are presented. Those differences are calculated by:
The largest differences between the particle-size distributions of the CEM-mortar and the sludge-made mortars occur between 0 and 10 lm, but are negligible above 10 lm.
Relative difference ¼ 100
CEM cum:passing sludge made CEM cum:passing CEM cum:passing
3.2.2. Slump The measurement of the slump with the Abrams mini-cone highlighted a special behavior of the sludge over the limestone fillers. Indeed, when limestone fillers were used for mortar fabrication, a slump of 2 cm was measured. When the wet sludge was used as limestone fillers substitute, a 0 cm slump was observed. This difference could be explained by a water retention behavior. Either the water of the sludge can be trapped into the porosity of
797
M. Audo et al. / Construction and Building Materials 112 (2016) 790–799
3.2.3. Calorimetric characterization On Fig. 12. are shown typical curve obtained by calorimetric characterizations. Results obtained on several mortars are depicted in Table 13. The incorporation of mineral additions allows to speed up the hydration kinetics, which can be explained by the creation of new nucleation sites for the hydrates precipitation [25] This influence is very variable depending on the kind of addition. While limestone filler, C2- and C4-sludge show only few influence on the hydration kinetics, the influence of C1- and C3-sludge is more distinct. Variations in anhydrous cement content in sludge could explain this phenomena. Also, it can be observed that the heat released is higher when sludge are used (from 29.2 to 38.0 J/g of cement). This could be due to the higher exothermic ions dissolution observed by conductimetric measurements, to the presence of anhydrous cement or also to a filler effect as described previously in the literature [25]. Fig. 10. Volume proportions of the reference concrete, the CEM mortar and the C1-, C2-, C3- and C4-CEM mortars made with wet sludge.
Fig. 11. Particle-size distribution of (a) the CEM mortar and (b) the relative difference between the particle-size distribution of this mortar with the ones of the sludge-made mortars.
the fraction A of the sludge, or the water can be adsorbed on the surface of the particles. This difference can also be explained by a modification of the packing density of the mortar. Thus, some superplasticizer was needed to adjust the rheology when sludge was used. The same slump class (S3) as the reference concrete in ready-mixed concrete plants was targeted for all mortars. While only 0.2% (based on cement weight) of superplasticizer was needed for the reference mortar, a high percentage (from 0.5% to 1%) was need when dry sludge was used. Unfortunately, those percentages are hardly compatible with an industrial application.
3.2.4. Harden state 3.2.4.1. Dry sludge. After 7 and 28 days of curing in water, the compressive strengths of the mortars were measured. The Fig. 13 allows to compare the compressive strengths between the CEM and the dry sludge-made mortars. All mortars present the same hardening kinetic, but different compressive strength. At 7-days old, the C1-, C2- and C3-sludge made mortars present a lower compressive strength compared to the reference CEM (-28%, -21% and 11% respectively). The C4-sludge made mortar is the only one which present a higher compressive strength (+32%). After 28 days, the same tendency is observed: the compressive strength of the C4-CEM remains the highest one, and the only one higher than the reference mortar. Considering the activity index (Fig. 7), the highest compressive strength was expected for the C2-CEM, while the lowest one was expected on the C3-CEM. But no relationship between the compressive strength and the activity index was observed, calling for new explanatory factors. The difference between the CEM and the sludge-made CEM compressive strength could be attributed to a higher porosity in the sludge-made mortars. It is well known that the higher the concrete porosity is, the lower the compressive strength is [26]. Nevertheless, few differences are observed when comparing the CEM porosity with the one of the sludge-made CEM (Table 14). As neither the porosity nor the activity index seem to explain the variability of the compressive strengths, the influence of percentage of fraction A in sludge and in mortars was studied. Calculations were performed with the Statgraphics Centurion software with the input data shown in Table 15, and the results of the multiple linear regressions are detailed in Table 16. Two variables were studied: wA/wS refers to the percentage of fraction A in sludge while wA/wM refers to the percentage of fraction A in mortars. Their influence on the 28-days compressive strength was determined through the p-value. A p-value lower than the defined significance level means that the variables have a significant influence on the measured parameter. On the contrary, a p-value higher than the defined significance level means that there is no significant influence of the variables on the measured parameter. As the p-values in Table 13 for the two studied variables are both below 0.05 (significance level of 95%), a significant relationship exists between wA/wS and wA/wM and the compressive strength. This relationship was then defined as:
rðMPaÞ ¼ 57:6487 þ 0:1844
wA wA 20:3565 2S wM
The validity of the model is proven by a p-value lower than 0.05. Moreover, the R2 coefficient correlation is very closed to 100%, which means that the whole experimental variability can be explained by the variables wA/wS and wA/wM. Given the model
798
M. Audo et al. / Construction and Building Materials 112 (2016) 790–799
Fig. 12. Comparison of the released hydration heats of CEM and CEM’ mortars (left: cumulative; right: non-cumulative).
Table 13 Calorimetric characterizations of the mortars made with limestone and dry sludge.
DQcum (J/g of cement) Dt (min)
CEM mortar
C1-CEM
C2-CEM
C3-CEM
C4-CEM
+12.8 15
+38.0 125
+35.1 15
+36.0 105
+29.2 15
Table 16 Multiple linear regression.
Constant wA/wS wA/wM
Coefficient
p-value
57.6387 0.184367 20.3565
0.0028 0.0121 0.0037
R2 = 99.99%.
Table 17 Compressive strengths of the mortars made with wet sludge.
7 days 28 days
CEM
C1-CEM
C2-CEM
C3-CEM
C4-CEM
23.1 (±0.8) 30.6 (±1.6)
18.7 (±2.0) 19.6 (±3.5)
19.4 (±2.4) 26.3 (±2.3)
20.7 (±0.4) 22.8 (±0.8)
– –
Fig. 13. Compressive strengths of the mortars made with dry sludge.
the compressive strength. Said differently, the compressive strength decreases when the substitution rate of the sand fraction increases.
Table 14 Relationship between compressive strength and porosity.
7-days compressive strength (MPa) 28-days compressive strength (MPa) 28-days porosity (%)
CEM
C1CEM
C2CEM
C3CEM
C4CEM
23.1 (±0.8) 30.6 (±1.6) 23.5
16.6 (±0.5) 21.1 (±2.7) 23.7
18.3 (±0.6) 23.4 (±3.2) 22.8
20.5 (±2.0) 28.8 (±3.7) 24.1
30.5 (±1.9) 35.7 (±1.2) 23.0
Table 15 Model input data.
wA/wS wA/wM Compressive strength (MPa)
C1
C2
C3
C4
0.236 0.0020 21.15
0.416 0.0021 23.43
0.241 0.0016 28.76
0.407 0.0014 35.70
equation, the most influent parameter on the compressive strength is the fraction A content in the mortars. The more fraction A in the mortars, the lower the compressive strength. On the other side, for one given wA/wM ratio, the higher is the wA/wS ratio, the higher is
3.2.4.2. Comparison between wet sludge and dry sludge. In order to industrialize the recycling of sludge into concrete in ready-mixed concrete plants, sludge must be used at their wet state. Consequently, mortars were also made with wet sludge in order to highlight any influence of the drying process on rheological and mechanical resistance. Firstly, wet sludge impacted significantly the fresh state rheological properties of the mortars, as less superplasticizer needed to be added to obtain the same slump. Secondly, the compressive strength obtained on those mortars (C1-, C2- and C3-CEM) are given in Table 17. Given the experimental uncertainties, equivalent compressive strengths were measured on mortars made with dry or wet sludge. Those results validate the dry-sludge approach in laboratory for the prediction of wet-sludge industrial approach. 4. Conclusions and prospects The work presented herein has focused on the effect of the incorporation of sludge coming from ready-mixed concrete plants
M. Audo et al. / Construction and Building Materials 112 (2016) 790–799
into mortars. As the sludge heavy metals and metalloid elements contents are close to the criteria of acceptance of inert wastes at landfills, it seems appropriate to propose a new valorization way: the incorporation in concrete as mineral addition. Firstly, the use of this kind of sludge as limestone fillers substitutes alters the rheology of the fresh state mortar. It becomes less plastic and some superplasticizer must be added in order to obtain the same class of slump as the reference mortar. The superplasticizers added quantities remain acceptable for the industrial. Even though, several options could be applied in order to limit this phenomenon, including the use of binary mixes between sludge and limestone fillers. Then, the 28-days compressive strengths are between 30% lower to 17% higher when the dry sludge are used instead of limestone fillers. This variation may be attributed two main factors: - the fraction A content in the mortars; - the substitution rate of the sand fraction of the mortars. Nevertheless, sludge coming from others ready-mixed concrete plants must be studied in order to confirm those results. Acknowledgments The authors would like to acknowledge the French Agency for Environment and Energy Management for the financial support, as well as all the industrial partners of the project and the trainees of the project. References [1] European Ready Mixed Concrete Organization, 2013, Ready-mixed concrete industry statistics – year 2012. [2] F. Sandrolini, E. Franzoni, Waste wash water recycling in ready-mixed concrete plants, Cem. Concr. Res. 31 (2001) 485–489. [3] N. Su, B. Miao, F.-S. Liu, Effect of wash water and underground water on properties of concrete, Cem. Concr. Res. 32 (2002) 777–782. [4] B. Chatveera, P. Lertwattanaruk, N. Makul, Effect of sludge water from readymixed concrete plant on properties and durability of concrete, Cem. Concr. Compos. 28 (2006) 441–450. [5] B. Chatveera, P. Lertwattanaruk, Use of ready-mixed concrete plant sludge water in concrete containing an additive or admixture, J. Environ. Manage. 90 (2009) 1901–1908.
799
[6] M.F. Pistilli, C.F. Peterson, Properties and possible recycling of solid waste from ready-mixed concrete, Cem. Concr. Res. 5 (1975) 249–259. [7] J. Schoon, K. De Buysser, I. Van Driessche, N. De Belie, Feasibility study of the use of concrete sludge as alternative raw material for Portland clinker production, J. Mater. Civ. Eng. (2014), http://dx.doi.org/10.1061/(ASCE) MT.1943-5533.0001230. [8] P.R. Adavi, A. Biswas, Analysis of sludge formed in R.M.C plant, AKGEC Int. J. Technol. 2 (2011) 60–65. [9] French Official Journal, ministerial order of 28 October 2010 on the storage of inert waste, 2010. [10] B. Sealey, P. Phillips, G. Hill, Waste management issues for the UK ready-mixed concrete industry, Resour. Conserv. Recycl. 32 (2001) 321–331. [11] Standard AFNOR NF EN 206-1, Concrete – part 1: Specification, performance, production and conformity. [12] A. Schwartzentruber, C. Catherine, Method of concrete equivalent mortar (CEM) – a new tool to design concrete containing admixture, Mater. Struct. 33 (2000) 475–482. [13] Standard AFNOR NF EN 12457-2, Characterization of waste – leaching – compliance test for leaching of granular waste materials and sludge – part 2: one stage batch test at a liquid to solid ratio of 10 L/kg for materials with particle size below 4 mm (without or with size reduction). [14] Standard AFNOR NF EN 196-6, Methods of testing cement – part 6: Determination of fineness. [15] Standard AFNOR NF EN 1097-7, Test for mechanical and physical properties of aggregates – part 7: Determination of the particle density of filler pyknometer method. [16] W. Hergert, The Mie Theory – Basics and Applications, Springer, 2012. [17] Standard AFNOR NF P 18-508, Additions for concrete – limestone additions – specifications and conformity criteria. [18] S. Maximilien, J. Péra, M. Chabannet, Study of the reactivity of clinkers by means of the conductometric test, Cem. Concr. Res. 27 (1997) 63–73. [19] Standard AFNOR NF EN 196-9, Methods of testing cement – part 9: Heat of hydration – semi-adiabatic method. [20] P. Siler, I. Kolarova, J. Kratky, J. Havlica, J. Brandstetr, Influence of superplasticizers on the course of hydration of portland cement hydration, Chem. Pap. 68 (2014) 90–97. [21] Standard AFNOR NF P18-459, Concrete – testing hardened concrete -testing porosity and density. [22] Official Journal of the European Union, Council decision 2003/33/EC of 19 December 2002 establishing criteria and procedures for the acceptance of waste at landfills pursuant to Article 16 of and Annex II to Directive 1999/31/ EC, 2003. [23] POYRI SA Étude sur le prix d’élimination des déchets inertes du BTP Technical report, ADEME, 2012. [24] A. Ipavec, R. Gabrovsek, T. Vuk, V. Kaucic, J. Macek, A. Meden, Carboaluminate phase formation during the hydration of calcite-containing Portland cement, J. Am. Ceram. Soc. 94 (2011) 1238–1242. [25] G. Ye, X. Liu, G. De Schutter, A.-M. Poppe, L. Taerwe, Influence of limestone powder used as filler in SCC on hydration and microstructure of cement pastes, Cem. Concr. Res. 29 (2007) 94–102. [26] M. Yudenfreund, K.M. Hanna, J. Skalny, I. Older, S. Brunauer, Hardened Portland cement pastes of low porosity V. Compressive strength, Cem. Concr. Res. 2 (1972) 731–743.
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Utilization of sludge from ready-mixed concrete plants as a substitute for limestone fillers Mariane Audo, Pierre-Yves Mahieux ⇑, Philippe Turcry Laboratoire des Sciences de l’Ingénieur pour l’Environnement, Université de La Rochelle, Avenue Michel Crépeau, 17000 La Rochelle, France
h i g h l i g h t s Environmental impacts of sludge coming from ready-mixed concrete plants. Substitution of limestone fillers by sludge coming-from ready-mixed concrete plants in mortars formulation. Mechanical characterization of mortars made with sludge coming from ready-mixed concrete plants.
a r t i c l e
i n f o
Article history: Received 12 May 2015 Received in revised form 12 January 2016 Accepted 16 February 2016
Keywords: Ready-mixed concrete plant sludge Waste management Environmental impact Mortar design Physical characterization Mechanical characterization
a b s t r a c t This study deals with the incorporation of sludge coming from ready-mixed concrete plants into mortars. Preliminary environmental investigations, made through leaching tests, showed the importance of managing those waste as they can be potentially pollutant regarding to their arsenic and chromium contents. Thus, management of the sludge can be environmentally and economically difficult. Reincorporation of those sludge into a closed loop concrete production is of particular interest. Also, it represents an interesting way to save raw materials (water, sand and limestone fillers). Yet, two main disadvantages were observed when using those sludge as limestone fillers substitute: – a decrease in the workability of the fresh state mortar, calling for a higher superplasticizer content; – a variability in the compressive strength of the hardened state mortars, between 30% and +17% when comparing to a reference mortar. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction European ready-mixed concrete production has notably increased over the last decades. Since 2010, around 370 million cubic meters of ready-mixed concrete are produced each year in Europe, of which around 10% come from the French production [1]. From this high production results a high level of by-products. Indeed, it has been estimated that a 9-m3 truck contains daily around 300 kg of returned concrete [2]. Generally, this leftover concrete is discharged in large containers. The hardened concrete can be then crushed and easily used as recycled aggregates for road construction. After the leftover being discharged, the truck is washed out with huge amounts of water, up to 1300 L per truck, and the suspended matter is allowed to settle in large sedimentation basins. Wash-waters from mixers are also directed to those basins. Several studies have been focused on the reuse of the clarified water of those basins, which presents a high up to 12 pH. This water can be partially reused for trucks washing or in concrete pro⇑ Corresponding author. E-mail address: [email protected] (P.-Y. Mahieux). http://dx.doi.org/10.1016/j.conbuildmat.2016.02.044 0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.
duction [3,4,5]. Su et al. showed that, according to the total suspended solids content, shorter setting times and a lower flowability are obtained [3]. On the opposite, Chatveera et al. highlighted a time lag when using sludge water and no additives nor admixtures, but reported a flowability decrease as well [5]. In another study, the use of the wet sludge as partial cement or sand substitute leads to a decrease in the compressive strength [6]. Recently, a study was led on the use of the sludge as a new raw material for Portland clinker production [7]. The variability in the chemical composition of the sludge leads to an impossibility in their reuse for Portland clinker production, due to their high chemical variations, as well as their high alkali, SO3 and MgO contents. Yet, not many studies have dealt with the utilization of the settled sludge, made of the fines particles (cement, mineral additions, etc.) as well as sand and aggregates. Though, it represents a large available quantity of raw material. As around 1 m3 of wet sludge is created by the production of 90 m3 of concrete, around 4 millions of cubic meters are produced every year in France [8]. Wet sludge is collected several times a year at the bottom of the basins and stored on the ready-mixed concrete plant area so that the water content decreases. When the water content is low enough, the
791
M. Audo et al. / Construction and Building Materials 112 (2016) 790–799
material can disposed of at controlled landfills for inert wastes, regarding to the French decree on passive wastes [9]. Up to now, not many characterizations of this sludge have been led in the literature. However, their hazardous behavior has already been highlighted, due to their high alkalinity and their high heavy metals and metalloid elements content [10]. All these factors lead to a high whole total disposal cost of the sludge. That is why new environmental friendly strategies must be found for sludge managing. We herein propose a new original way to valorize sludge as limestone fillers substitutes in a closed loop concrete production, avoiding any environmental and economical impacts. Nevertheless, for now, the European standard does not allow the utilization of non-standard mineral addition in concrete [11]. Thus, the objective of this study is to prove the feasibility of incorporating sludge into concrete. To reach that aim, the sludge coming from four French ready-mixed concrete plants were firstly characterized from a chemical and physical point of view. The activity index of the dry sludge, as well as their chemical activities were determined. Giving those results, utilization of the studied sludge into mortars compositions was studied using the concrete equivalent mortar concept [12]. Compressive strength and porosity were studied at various due dates, and relationship between mortars composition and compressive strength was stated. 2. Experimental program 2.1. Raw materials The studied sludge were sampled at four ready-mixed concrete plants (C1, C2, C3, C4) located in the Poitou-Charentes area, France (Fig. 1). Those four plants produce basically the same main concrete, whose formulation will be given later. The samples had a high water content, between around 50 and 120%, depending on the weather conditions and the storage duration. The moisture contents are presented in Table 1. The raw materials were then split in two fractions: one finer than 100 lm (fraction A) and one coarser than 100 lm (fraction B). Limestone filler (designed as LF), CEM II/A 42.5 and CEM I 52.5 cements were provided by Carmeuse (Saint Porchaire, France), Calcia (Airvault, France) and Lafarge (Saint Pierre La Cour, France) companies, respectively. Their properties are given in Table 2. 0/4 mm sand from Sablimaris Pallice (La Rochelle, France) was used in the studied mortars. For the determination of activity index, a standard sand from the SNL Company (Leucate, France) was used. Chrysofluid Optima 220 superplasticizer was used for the fabrication of mortars.
2.2. Characterization of raw materials 2.2.1. Leaching tests Leaching tests were performed following the French standard dedicated to sludge characterization [13]. Basically, a test portion of sludge containing 90 g of dry matter is poured into a 1L HDPE flask. Deionized water is added so that the total water content is 900 mL. The whole mixture is kept under continuous mechanical
Fig. 1. Picture of the raw materials coming from the C4 ready-mixed concrete plant.
Table 1 Water contents of the sludge
Water content (%)
C1 47.1
C2 99.5
C3 120.3
C4 155.2
stirring for 24 ± 0.5 h (Fig. 2). The mixtures are then filtered over a 2 lm filter with a Büchner apparatus and the filtrate is recovered for further analysis. Right before the analysis, samples were filtered on 0.45 lm syringe filter. All the leachates are diluted tenth and hundredth in a 5% HNO3 solution (77 mL of HNO3 – FisherScientific, Trace Metal Analysis quality – completed to 1 L with milli-Q water). Ba, Cr, Cu, Mo, Ni, Pb and Zn were analyzed with a Varian Vista Pro ICPOES and As, Cd and Se with a ThermoFischerScientific Xseries2 ICPMS. The quantification limits (lg L1 solution) were 2 (As), 2 (Se), 0.1 (Cd), 50 (Ba), 10(Cr), 10 (Mo), 10 (Ni), 10 (Pb) and 50 (Zn). To check the analytical data precision, all samples were analyzed in duplicate. Ionic chromatography was also performed on the leachates. A Metrohm apparatus with a Metrosep A Supp 5 100/4.0 column and an automatic sample changer was used. A Na2CO3 3.2 mM/NaHCO3 1.0 mM (1:1, v/v) eluent was used, with a 0.7 mL/min flow. F, Cl and SO2 4 were so quantified. Sodium salts from SigmaAldrich company were used as standards. The quantification limits were 0.5 mg L1 for F, Cl and SO2 4 .
2.2.2. Physical and mineralogical characterization Water content of sludge (defined as mass ratio of water over dry matter) was determined by drying at 80 °C until constant weight. Specific surfaces of the dry raw materials were determined with a Blaine apparatus, allowing the measurement of the resistance of the air passing through a porous bed of powder [14]. The densities were determined using a water pyknometer according to the European Standard NF EN 1097-7 [15]. The particle-size distribution of the dry sludge was determined by sieving 500 g of powder at 4, 2, 1.25, 0.5, 0.25, 0.125 and 0.100 mm. The 100 lm passing fraction was characterized by dynamic light scattering with a CILAS 1190 apparatus, used in dry mode. Calculations were performed using the Mie theory [16]. Powder X-Ray Diffraction (XRD) was carried out with a Brücker diffraction instrument, with Cu Ka1 radiation. Measurement range was from 5 to 70° 2b, with a 0.02° step. Identification of the peaks was performed by comparison to a database references. Thermogravimetric analysis (TGA) was performed on a Setaram Setsys Evolution 16/18 apparatus. Around 100 mg of samples were heated from 20 °C to 1000 °C at a 10 °C/min heating rate, under neutral argon atmosphere. To get more accurate results, data were analyzed through the differential thermogravimetric curves. ICP-AES analysis were performed sludge materials after their drying at 80 °C and after being grinded to 80 lm. Basically, around 100 mg of dry materials was digested using 4 mL of a 67–70% HNO3 – 34–37% HCl 2:2 (v/v) solution (FisherScientific, Trace Metal Analysis quality). Acidic digestion of the samples was carried out overnight at room temperature. Each sample was completed to 50 mL with milliQ water. Al, Ca, Fe, S and Si were analyzed with a Varian Vista Pro ICPOES. The quantification limits (lg g1 dry weight) were 500 (Ca), 100 (Al), 50 (Si), 10 (S) and 5 (Fe).
2.2.3. Determination of sludge activity The potential activity of sludge was evaluated through compressive tests on standard mortars. We focused on the sludge fraction lower than 100 lm which should contain the most reactive elements. For this purpose, an activity index was determined according to the French standard dedicated to limestone additions [17]. A mortar made of CEM I cement, deionized water and standard sand was made and used as reference. A second mortar was made by substituting 25% of the cement mass by the 100 lm passing fraction of the sludge. Mortars compositions, as well as the water-to-cement ratio (W/C) are given in Table 3. Compressive strengths were determined on 28-day old mortars kept in water. The activity index (denoted i) of the sludge was defined as the ratio between the compressive strength of the mortar with sludge and the one of the reference mortar. The activity of sludge was also investigated through ionic conductivity measurement. This test is based on the measurement of time-evolution of the ionic concentrations during the cement hydration in aqueous suspension. As done in the case of activity index, only the 100 lm passing fraction of sludge was used. The ionic conductivity time-evolution of the cement and blends (‘‘limestone filler + cement” and ‘‘dry sludge + cement”) were obtained by measurements in dilute medium. In the blends, the ratios between cement and dry sludge or limestone filler were the same as the ones used for the CEM mortars. The same cement as the one used in the concrete formulation was used. The composition of the blends is given in Table 4. Briefly, 300 mL of deionized water are poured into a cell equipped with a conductivity probe. The whole system is kept at 25 °C and under perpetual mechanical stirring. A 1–5 solid-to-liquid ratio is applied, which allows a good sensitivity, as
792
M. Audo et al. / Construction and Building Materials 112 (2016) 790–799
Table 2 Properties of cements and limestone filler. Physical properties
CEM I 52.5 CEM II/A 42.5 LF
Clinker composition (%)
Density (g/cm3)
Blaine (cm2/g)
Clinker (%)
CaO
SiO2
Al2O3
Fe2O3
SO3
CaCO3
MgCO3
3.15 3.12 2.70
3650 3800 3530
98 92 –
64.5 63.0 –
20.1 18.7 –
5.0 4.5 –
3.1 3.2 –
3.3 2.7 –
– 6.0 95.0
– – 3.7
Fig. 2. Picture of the mechanical stirring system.
Table 3 Compositions of mortars used to determine activity index.
Standard sand (g) CEM I 52.5 cement (g) Dry sludge (<100 lm) (g) Deionized water (g) W/C
Standard mortar (reference)
Standard mortar with sludge
1350.0 450.0 – 225.0 0.50
1350.0 337.5 112.5 225.0 0.66
well as a short reaction time. At the zero time, 75 g of cement, cement and filler or cement and dry sludge are added and the conductivity is measured every 5 s during the first 30 min and every two minutes during the last 7h30 [18].
2.3. Study of mortars made with sludge 2.3.1. Mortars mix proportioning Mortars were designed using the concrete equivalent mortar concept [12]. The latter is based on the substitution of the coarse aggregates of concrete by a certain quantity of sand in the mortars, which develops the same area as the coarse aggregates do. This method was firstly developed in order to easily detect compatibility troubles between cements and superplasticizers. It is also a convenient method for comparison studies in laboratory. The mortars compositions were obtained from a reference concrete mixture given in Table 5. The latter corresponds to a ‘‘XC1 C20/25” concrete according to European standard [11], with W/C = 0.6. It is usually produced by the four readymixed concrete plants where the studied sludge come from. The equivalent mortar derived from the reference concrete is denoted CEM in Table 6. Mortars were also
Table 4 Composition of the blends for ionic conductivity measurements.
Cement Limestone filler C1 C2 C3 C4
Carbonates content (%)
CEM II/A (g)
Sludge (fraction A) (g)
75.00 67.53 67.99 67.74 69.15 69.74
– 7.47 7.01 7.26 5.85 5.26
designed with sludge. Firstly, it was chosen to use dry sludge, even if from an industrial point of view, using wet sludge would avoid any costly drying process. At a laboratory scale, using dry sludge allowed us a better mastering of the water content in the mortars. Secondly, the mix proportions were derived from the CEM mortar. Dry sludge was used as a substitute of all the limestone filler and of part of the sand contained in the CEM mortar. The mass ratios between the dry sludge and the sand in the sludge-made mortars were determined by setting the particle-size distribution of those mortars identical as much as possible to the one of the CEM mortar. To determine the masses of dry sand and dry sludge, a second condition was applied to close the equations system: the volume (i.e. sand + mineral addition + cement + water) of both mortars was fixed equal. The dry sludge-made mortars compositions (denoted C1-CEM, C2-CEM, C3-EM and C4-CEM) are given in Table 6. Mortars were also designed with wet sludge to monitor the influence of the water content of the sludge on the properties of the mortars (Table 7). Tap water contents were adjusted following the sludge water contents. The sludge was mixed with cement and sand for 30 s at low speed [11]. Tap water was then added and the blend was mixed for additional 30 s at low speed. After scraping, another 30 s mix were applied at high speed [11]. Some superplasticizer (Chrysofluid Optima 220) was added so that the slump at the Abrams minicone test was the same as the CEM mortar one. A vibration table was used to fill the 4 4 16 cm molds with the mortar. After removal from the moulds (24 hours after their manufacturing), all the mortars were kept into water at 20 °C until their later characterization.
2.3.2. Mortars characterization Fresh mortars were characterized by slump measurement with a cylinder (H = 100 mm, d = 50 mm). Heat releases determination were performed following a protocol adapted from the semi-adiabatic Langavant method [19]. This method consists in the introduction of a reference fresh mortar (denoted CEM’, Table 6) and the studied fresh sludge-based mortar into two separates 800 cm3-Dewar flasks and the measurement of the difference of temperature in the core of the mortars at any time. In the reference CEM’ formulation, no limestone filler was used. Indeed, it had been substituted by the quantity of CEM II cement having the same hydraulic activity. It has to be noted that the mortars were made without any superplasticizer addition. Indeed, it is known that chemical admixtures can modify hydration kinetics [20] and we herein want to monitor the inherent influence of sludge on hydration. Compressive strength was determined at the ages of 2, 7 and 28 days using the 4 4 16 cm prisms, with a Zwick press (loading rate = 2200 N/s and preload = 3000 N). 4 measurements were performed on each mortar formulation. Experimental variability will be indicated as error bars on graphs. Porosity was measured at the age of 28 days by water soaking following the recommended method [21].
3. Results and discussion 3.1. Sludge characterization 3.1.1. Leaching behavior The results of the chemical analysis led on the leachates are presented in Table 8. Concerning the anions contents, fluorides, chlorides and sulfates contents comply with the limit values of the Table 5 Compositions of reference concrete (kg per m3). Weight (kg/m3) Gravel 10/20 Gravel 6/10 Sand 0/4 Cement CEM II/A Limestone filler Water Superplasticizer (%)
823 352 775 253 28 169 0.6
793
M. Audo et al. / Construction and Building Materials 112 (2016) 790–799 Table 6 Compositions of the studied mixtures with dry sludge (kg per m3).
Sand 0/4 Cement CEM II/A Limestone filler Dry sludge Water Superplasticizer (%)
CEM mortar
C1-CEM mortar
C2-CEM mortar
C3-CEM mortar
C4-CEM mortar
CEM’ mortar
1476 424 47 0 283 0.2
1290 424 0 185 283 0.6
1391 424 0 109 283 0.5
1335 424 0 149 283 0.8
1427 424 0 79 283 0.9
1476 436 0 0 283 0
Table 7 Compositions of the studied mixtures with wet sludge (kg per m3).
Sand 0/4 Cement CEM II/A Limestone filler Wet sludge Water Superplasticizer (%)
CEM mortar
C1-CEM mortar
C2-CEM mortar
C3-CEM mortar
C4-CEM mortar
1476 424 47 0 283 0.2
1290 424 0 272 144 1.6
1391 424 0 218 174 2.0
1335 424 0 328 104 1.2
– – – – – –
European decision concerning the acceptance of waste at landfills. Concerning the heavy metals and metalloids content, special attention must be paid to the arsenic and chromium elements, as the detected values are sometimes slightly higher than the limit values for inert waste [22]. Consequently, the sludge cannot be surely stated inert materials. This point is of particular importance from an environmental and economical point of view. Indeed, the price concerning the waste storage is directly depending on the category they belong to: 8 €/t for inert waste over 40 €/t for non-dangerous waste [23]. Nevertheless, the measured suspicious concentrations are right above the inert waste acceptability limit. Thus those results need to be completed by analyzing new sludge. Regarding to those results, finding a way to avoid storage of sludge becomes of particular interest. Thus, incorporation into cement matrix could be a promising, innovative and easy way to valorize them. 3.1.2. Physical and mineralogical properties The particle-size distributions analysis of the four dry sludge showed a continuous material containing between 20 to 40% of particles finer than 100 lm (Fig. 3, sieving). Nevertheless, the particle-size distribution of the fractions finer than 100 lm remains almost the same for all the sludge (Fig. 4, dynamic light scattering). Moreover, this distribution is very closed to the one of limestone filler. Indeed, two of the four modes of the filler are observed on the sludge. The dmax of all sludge being less than
8 mm, it makes them a priori convenient for a use as sand and mineral addition substitute. Measurements of density were led on those two fractions (Table 9). The results showed that both fractions have a lower density (around 2.2) than the limestone fillers (2.7). Compared to fraction B, the fraction A slightly lower density can be explained by a higher cementitious materials content and a lower sand content. Concerning the specific surface, the one measured on the fraction A is up to 5 times higher than the one of the limestone fillers, which could be explained by a higher open porosity or different morphology. XRD patterns were also acquired on each fraction (Fig. 5). Basically, the same peaks were observed for all samples, but with various proportions. The coarsest material is mainly made by aggregates. Indeed, quartz (SiO2) and limestone (CaCO3) were observed, as well as aggregates formed by portlandite (Ca(OH)2) and/or C–S–H conglomeration. The composition of the finest fraction is more complex. Quartz, limestone and portlandite were identified as main constituents, but gypsum (CaSO42H2O), as well as anhydrous (2CaOSiO2; 3CaOSiO2) and hydrated calcium silicates were also observed. The presence of anhydrous calcium silicates show a potential residual hydraulic activity. While no ettringite (6CaOAl2O33SO332H2O) was identified, some hydrated calcium carboaluminates (3CaOAl2O3CaCO311H2O) have been detected. In literature, it has already been shown that this kind of compounds can be formed during the hydration of cement
Table 8 Sludge leachates composition (g/kg of dry matter) and criteria of acceptance of waste at landfills [22] (in bold: inert waste value exceeded). Up: Heavy metals and metalloid elements contents; Down: Anions contents.
C1-sludge C2-sludge C3-sludge C4-sludge Inert waste Non-dangerous waste
C1-sludge C2-sludge C3-sludge C4-sludge Inert waste Non-dangerous waste
As
Ba
Cd
Cr
Cu
Mo
Ni
Pb
Se
Zn
0.54 0.15 <0.02 <0.02 0.5 2
12.53 17.70 13.41 14.16 20 100
<0.001 <0.001 <0.001 <0.001 0.04 1
0.37 0.31 0.64 0.12 0.5 10
0.23 0.32 <0.11 0.31 2 50
<0.11 <0.11 <0.11 <0.11 0.5 10
<0.11 <0.11 <0.11 <0.11 0.4 10
<0.11 <0.11 <0.11 <0.11 0.5 10
<0.02 <0.02 <0.02 <0.02 0.1 0.5
<0.56 <0.56 0.71 <0.56 4 50
F
Cl
SO2 4
<8.37 <4.61 <8.77 <8.76 10 150
39.91 49.91 182.81 39.47 800 15,000
5.65 13.27 5.07 3.84 1000 20,000
794
M. Audo et al. / Construction and Building Materials 112 (2016) 790–799
determined by TGA. Regarding to the total portlandite content, fractions A present lower contents, which can be explained by the conglomeration of portlandite, which forms particles coarser than 100 lm. Global chemical compositions of the sludge are given in Table 10. Weak variability is observed between the four sludge (standard deviation below 20%). Those results can be explained by the homogeneity of the concrete produced in the four studies ready-mixed concrete plants. Indeed, they mainly produced the same concrete (formulation in Table 6). Those compositions are consistent with literature data on sludge chemical composition [4,5,6,7].
Fig. 3. Particle-size distribution of the dry sludge.
containing limestone (such as CEM II type cement), from the timeconversion of hemicarboaluminates [24]. As the studied readymixed concrete plants mainly use a CEM II/A cement with limestone addition, the presence of hydrated calcium carboaluminates in the sludge was expected. Limestone and portlandite contents were both determined with TGA. After differentiation, the typical curve showed four main peaks (Fig. 6). The first two ones, observed between 25 °C and 250 °C result from the free water evaporation, as well as the water loss coming from the dehydration of the C–S–H and calcium carboaluminates. Around 460 °C is observed the signal due to the dehydroxylation of portlandite. The last peak is due to the CO2 loss, corresponding to the decarbonation of the limestone. The integration of those peaks allowed us to determine the mass losses and consequently, the calcium carbonate and portlandite contents (Table 9). The weak peaks near 300 °C and 400 °C may be attributed to polymeric additives decomposition. The latter peak may also be attributed to the quartz dehydroxylation. Fractions A present high content of CaCO3, regarding to the total limestone content of the sludge, which is likely due to the limestone filler and cement being the main CaCO3 inputs. On average, the total limestone content of the sludge is around 11.2%. This mass percentage in the mixture can be explained by considering the limestone coming from a usual concrete formulation (reference concrete, Table 6) produced in the ready-mixed concrete plant. Indeed, considering no gravels in the sludge, a 12.6% content of calcite in the dry sludge (coming from the cement, the limestone fillers and the sand) was expected, which is closed to the content
3.1.3. Reactivity As can be seen in the Table 11 and the Fig. 7, the substitution of 25% of the cement by a mineral addition causes a decrease of the mechanical strength of the standard mortar. Indeed, a compressive strength of 37.52 MPa was obtained on the reference mortar, the compressive strength of mortars with mineral addition were included between 29.20 and 34.73 MPa. However, the French standard [17] stipulates that the activity index of a limestone fillers must be higher than 71%. The activity index of the used filler is around 92%, which is in compliance to the standard. The activity index of the C1, C3 and the C4 sludge (fractions A) are slightly lower than the one of the limestone filler. Nevertheless, they remains higher than the required 71%. The activity index of the C2-sludge (fraction A) reaches a value of 94%, which is even slightly higher than the value obtained with the limestone fillers. Overall, few dispersion is observed when comparing the activity index of the finer fraction of the sludge, which can be explained by a very close mineralogical and chemical composition. The differences may be attributed to a variation in the particle-size distributions, so to a variation in the packing densities. In order to get more accurate results on the chemical activity of the sludge, conductimetric tests were performed on blends between cement and the fraction A of the sludge. The composition of the blends is given in the experimental section. The results highlighted a classical behavior of cementitious materials [18]. During the first minutes, an exponential rising of the conductivity is observed, due to the high dissolution rate of the compounds of the clinker and gypsum. This period is known as the mixing period (I). Then, the dormant period (II) begins. The C–S–H and ettringite precipitation begins, while the conductivity remains rising. Gradually, the solution becomes richer in Ca2+ and HO ions, up to the point where the critical concentrations are reached and the portlandite precipitation begins. At that
Fig. 4. DLS particle-size distribution of the 100 lm passing of the dry sludge and comparison with a limestone filler (left: non-cumulative passing; right: cumulative passing).
795
M. Audo et al. / Construction and Building Materials 112 (2016) 790–799 Table 9 Density, specific surface and CaCO3 and Ca(OH)2 contents of the fraction A and B – Fractions A and B mass content in sludge.
3
Density (g/cm ) Specific surface (m2/kg) Ca(OH)2 content (%) CaCO3 content (%)
Fraction A (%) Fraction B (%)
LF
C1 (A)
C1 (B)
C2 (A)
C2 (B)
C3 (A)
C3 (B)
C4 (A)
C4 (B)
2.7 350 – 98.0
2.2 1100 8.8 17.5
2.2 – 10.8 13.0
2.2 1300 7.5 18.7
2.2 – 13.9 23.5
2.1 1800 6.0 22.2
2.3 – 10.8 13.4
2.2 1350 4.6 19.6
2.3 – 14.7 9.0
C1
C2
C3
C4
20.9 79.1
40.0 60.0
26.9 73.1
38.6 61.4
point, the conductivity starts to decrease: this is the acceleration period (III). The curve obtained on the CEM II/A cement is shown on Fig. 8 for example. The limestone filler, as well as fractions A of the dry sludge, does not show any hydraulic behavior. Indeed, when placed in an aqueous media, the conductivity increases during the first seconds and then remains stable until the end of the experiment (Table 12). But a large difference in the chemical behavior of the sludge and the fillers can be observed, as the maximal conductivity is at least ten times higher when sludge are used over fillers. Nevertheless, when mixed with cement, all studied materials modify the hydration kinetic of the cement. Especially, when the limestone filler delays the portlandite precipitation (30 min for a reference time of 210 min), the fraction A of the sludge speed up the portlandite precipitation (between 10 and 60 min). Those time gaps were higher to the accuracy of the method (estimated to be 4 min). Ten, those results highlight a chemical activity of the sludge within the
cementitious matrix, which is likely due to a high dissolution rate in water. Indeed, the higher is the maximum conductivity in presence of sludge, the quicker is the cement hydration. In concrete, limestone fillers are known to speed up the hydration kinetics [25]. This effect cannot be observed here, as the dilute medium does not allow to highlight the physical behavior of filler additions. 3.2. Mortars made with sludge 3.2.1. Mortars composition In Figs. 9 and 10 are presented the volume proportions of the reference concrete, the reference CEM, and the dry sludge-CEM and wet sludge-CEM. It can be noted that, depending on the particle-size distribution of the sludge, between 4.5 and 11% of the sand of the CEM mortar can be substituted by the sand brought by the sludge. Moreover, the mass of dry sludge with particle size lower than 100 lm in the sludge-based mortar is almost equal to
Fig. 5. XRD patterns of the fractions A and B of the C1- (up, left), C2-sludge (up, right), C3-sludge (down, left) and C4-sludge (down, right) (C: Calcite; CC: Hydrated calcium carbonates; CS: Calcium silicates; CSH: Hydrated calcium silicates; G: Gypsum; P: Portlandite; Q: Quartz).
796
M. Audo et al. / Construction and Building Materials 112 (2016) 790–799
Fig. 6. Typical DTG curve (example on the fraction A of the C3 sludge). Fig. 8. Conductimetric curve obtained with the CEM II/A cement. Table 10 Chemical composition of sludge (%).
C1 C2 C3 C4
Al2O3
Fe2O3
CaO
SiO2
SO3
1.70 2.09 1.65 1.97
1.45 1.72 1.42 1.62
31.79 37.32 32.83 34.85
57.04 50.04 54.85 53.50
0.59 0.98 0.81 0.71
Table 11 Compressive strength of the standard mortars.
fc(28) (MPa)
Reference 37.52
Limestone filler 34.73
C1 33.38
C2 35.13
C3 29.20
C4 30.21
Table 12 Conductimetric characterizations of the CEM II cement mixed with different mineral additions.
Cement Limestone filler C1 dry sludge C2 dry sludge C3 dry sludge C4 dry sludge Cement + limestone filler Cement + C1 dry sludge Cement + C2 dry sludge Cement + C3 dry sludge Cement + C4 dry sludge
rmax (lS/cm)
Slope reversal time (min)
19.8 0.1 5.5 6.5 1.7 1.1 18.3 17.7 18.2 18.1 18.1
210 – – – – – 242 180 148 188 200
Fig. 7. Comparison of the activity of the limestone fillers and the fraction A of the sludge.
Fig. 9. Volume proportions of the reference concrete, the CEM mortar and the C1-, C2-, C3- and C4-CEM mortars made with dry sludge.
the mass of limestone filler in the CEM mortar. Also, using wet sludge allows to substitute tap water, from 37% (C4-CEM) to 56% (C3-CEM). The particle-size distribution of the CEM mortar is presented in Fig. 11(a). In Fig. 11(b) the relative differences between the particle-size distribution of the CEM mortar and the ones of the sludge-made mortars are presented. Those differences are calculated by:
The largest differences between the particle-size distributions of the CEM-mortar and the sludge-made mortars occur between 0 and 10 lm, but are negligible above 10 lm.
Relative difference ¼ 100
CEM cum:passing sludge made CEM cum:passing CEM cum:passing
3.2.2. Slump The measurement of the slump with the Abrams mini-cone highlighted a special behavior of the sludge over the limestone fillers. Indeed, when limestone fillers were used for mortar fabrication, a slump of 2 cm was measured. When the wet sludge was used as limestone fillers substitute, a 0 cm slump was observed. This difference could be explained by a water retention behavior. Either the water of the sludge can be trapped into the porosity of
797
M. Audo et al. / Construction and Building Materials 112 (2016) 790–799
3.2.3. Calorimetric characterization On Fig. 12. are shown typical curve obtained by calorimetric characterizations. Results obtained on several mortars are depicted in Table 13. The incorporation of mineral additions allows to speed up the hydration kinetics, which can be explained by the creation of new nucleation sites for the hydrates precipitation [25] This influence is very variable depending on the kind of addition. While limestone filler, C2- and C4-sludge show only few influence on the hydration kinetics, the influence of C1- and C3-sludge is more distinct. Variations in anhydrous cement content in sludge could explain this phenomena. Also, it can be observed that the heat released is higher when sludge are used (from 29.2 to 38.0 J/g of cement). This could be due to the higher exothermic ions dissolution observed by conductimetric measurements, to the presence of anhydrous cement or also to a filler effect as described previously in the literature [25]. Fig. 10. Volume proportions of the reference concrete, the CEM mortar and the C1-, C2-, C3- and C4-CEM mortars made with wet sludge.
Fig. 11. Particle-size distribution of (a) the CEM mortar and (b) the relative difference between the particle-size distribution of this mortar with the ones of the sludge-made mortars.
the fraction A of the sludge, or the water can be adsorbed on the surface of the particles. This difference can also be explained by a modification of the packing density of the mortar. Thus, some superplasticizer was needed to adjust the rheology when sludge was used. The same slump class (S3) as the reference concrete in ready-mixed concrete plants was targeted for all mortars. While only 0.2% (based on cement weight) of superplasticizer was needed for the reference mortar, a high percentage (from 0.5% to 1%) was need when dry sludge was used. Unfortunately, those percentages are hardly compatible with an industrial application.
3.2.4. Harden state 3.2.4.1. Dry sludge. After 7 and 28 days of curing in water, the compressive strengths of the mortars were measured. The Fig. 13 allows to compare the compressive strengths between the CEM and the dry sludge-made mortars. All mortars present the same hardening kinetic, but different compressive strength. At 7-days old, the C1-, C2- and C3-sludge made mortars present a lower compressive strength compared to the reference CEM (-28%, -21% and 11% respectively). The C4-sludge made mortar is the only one which present a higher compressive strength (+32%). After 28 days, the same tendency is observed: the compressive strength of the C4-CEM remains the highest one, and the only one higher than the reference mortar. Considering the activity index (Fig. 7), the highest compressive strength was expected for the C2-CEM, while the lowest one was expected on the C3-CEM. But no relationship between the compressive strength and the activity index was observed, calling for new explanatory factors. The difference between the CEM and the sludge-made CEM compressive strength could be attributed to a higher porosity in the sludge-made mortars. It is well known that the higher the concrete porosity is, the lower the compressive strength is [26]. Nevertheless, few differences are observed when comparing the CEM porosity with the one of the sludge-made CEM (Table 14). As neither the porosity nor the activity index seem to explain the variability of the compressive strengths, the influence of percentage of fraction A in sludge and in mortars was studied. Calculations were performed with the Statgraphics Centurion software with the input data shown in Table 15, and the results of the multiple linear regressions are detailed in Table 16. Two variables were studied: wA/wS refers to the percentage of fraction A in sludge while wA/wM refers to the percentage of fraction A in mortars. Their influence on the 28-days compressive strength was determined through the p-value. A p-value lower than the defined significance level means that the variables have a significant influence on the measured parameter. On the contrary, a p-value higher than the defined significance level means that there is no significant influence of the variables on the measured parameter. As the p-values in Table 13 for the two studied variables are both below 0.05 (significance level of 95%), a significant relationship exists between wA/wS and wA/wM and the compressive strength. This relationship was then defined as:
rðMPaÞ ¼ 57:6487 þ 0:1844
wA wA 20:3565 2S wM
The validity of the model is proven by a p-value lower than 0.05. Moreover, the R2 coefficient correlation is very closed to 100%, which means that the whole experimental variability can be explained by the variables wA/wS and wA/wM. Given the model
798
M. Audo et al. / Construction and Building Materials 112 (2016) 790–799
Fig. 12. Comparison of the released hydration heats of CEM and CEM’ mortars (left: cumulative; right: non-cumulative).
Table 13 Calorimetric characterizations of the mortars made with limestone and dry sludge.
DQcum (J/g of cement) Dt (min)
CEM mortar
C1-CEM
C2-CEM
C3-CEM
C4-CEM
+12.8 15
+38.0 125
+35.1 15
+36.0 105
+29.2 15
Table 16 Multiple linear regression.
Constant wA/wS wA/wM
Coefficient
p-value
57.6387 0.184367 20.3565
0.0028 0.0121 0.0037
R2 = 99.99%.
Table 17 Compressive strengths of the mortars made with wet sludge.
7 days 28 days
CEM
C1-CEM
C2-CEM
C3-CEM
C4-CEM
23.1 (±0.8) 30.6 (±1.6)
18.7 (±2.0) 19.6 (±3.5)
19.4 (±2.4) 26.3 (±2.3)
20.7 (±0.4) 22.8 (±0.8)
– –
Fig. 13. Compressive strengths of the mortars made with dry sludge.
the compressive strength. Said differently, the compressive strength decreases when the substitution rate of the sand fraction increases.
Table 14 Relationship between compressive strength and porosity.
7-days compressive strength (MPa) 28-days compressive strength (MPa) 28-days porosity (%)
CEM
C1CEM
C2CEM
C3CEM
C4CEM
23.1 (±0.8) 30.6 (±1.6) 23.5
16.6 (±0.5) 21.1 (±2.7) 23.7
18.3 (±0.6) 23.4 (±3.2) 22.8
20.5 (±2.0) 28.8 (±3.7) 24.1
30.5 (±1.9) 35.7 (±1.2) 23.0
Table 15 Model input data.
wA/wS wA/wM Compressive strength (MPa)
C1
C2
C3
C4
0.236 0.0020 21.15
0.416 0.0021 23.43
0.241 0.0016 28.76
0.407 0.0014 35.70
equation, the most influent parameter on the compressive strength is the fraction A content in the mortars. The more fraction A in the mortars, the lower the compressive strength. On the other side, for one given wA/wM ratio, the higher is the wA/wS ratio, the higher is
3.2.4.2. Comparison between wet sludge and dry sludge. In order to industrialize the recycling of sludge into concrete in ready-mixed concrete plants, sludge must be used at their wet state. Consequently, mortars were also made with wet sludge in order to highlight any influence of the drying process on rheological and mechanical resistance. Firstly, wet sludge impacted significantly the fresh state rheological properties of the mortars, as less superplasticizer needed to be added to obtain the same slump. Secondly, the compressive strength obtained on those mortars (C1-, C2- and C3-CEM) are given in Table 17. Given the experimental uncertainties, equivalent compressive strengths were measured on mortars made with dry or wet sludge. Those results validate the dry-sludge approach in laboratory for the prediction of wet-sludge industrial approach. 4. Conclusions and prospects The work presented herein has focused on the effect of the incorporation of sludge coming from ready-mixed concrete plants
M. Audo et al. / Construction and Building Materials 112 (2016) 790–799
into mortars. As the sludge heavy metals and metalloid elements contents are close to the criteria of acceptance of inert wastes at landfills, it seems appropriate to propose a new valorization way: the incorporation in concrete as mineral addition. Firstly, the use of this kind of sludge as limestone fillers substitutes alters the rheology of the fresh state mortar. It becomes less plastic and some superplasticizer must be added in order to obtain the same class of slump as the reference mortar. The superplasticizers added quantities remain acceptable for the industrial. Even though, several options could be applied in order to limit this phenomenon, including the use of binary mixes between sludge and limestone fillers. Then, the 28-days compressive strengths are between 30% lower to 17% higher when the dry sludge are used instead of limestone fillers. This variation may be attributed two main factors: - the fraction A content in the mortars; - the substitution rate of the sand fraction of the mortars. Nevertheless, sludge coming from others ready-mixed concrete plants must be studied in order to confirm those results. Acknowledgments The authors would like to acknowledge the French Agency for Environment and Energy Management for the financial support, as well as all the industrial partners of the project and the trainees of the project. References [1] European Ready Mixed Concrete Organization, 2013, Ready-mixed concrete industry statistics – year 2012. [2] F. Sandrolini, E. Franzoni, Waste wash water recycling in ready-mixed concrete plants, Cem. Concr. Res. 31 (2001) 485–489. [3] N. Su, B. Miao, F.-S. Liu, Effect of wash water and underground water on properties of concrete, Cem. Concr. Res. 32 (2002) 777–782. [4] B. Chatveera, P. Lertwattanaruk, N. Makul, Effect of sludge water from readymixed concrete plant on properties and durability of concrete, Cem. Concr. Compos. 28 (2006) 441–450. [5] B. Chatveera, P. Lertwattanaruk, Use of ready-mixed concrete plant sludge water in concrete containing an additive or admixture, J. Environ. Manage. 90 (2009) 1901–1908.
799
[6] M.F. Pistilli, C.F. Peterson, Properties and possible recycling of solid waste from ready-mixed concrete, Cem. Concr. Res. 5 (1975) 249–259. [7] J. Schoon, K. De Buysser, I. Van Driessche, N. De Belie, Feasibility study of the use of concrete sludge as alternative raw material for Portland clinker production, J. Mater. Civ. Eng. (2014), http://dx.doi.org/10.1061/(ASCE) MT.1943-5533.0001230. [8] P.R. Adavi, A. Biswas, Analysis of sludge formed in R.M.C plant, AKGEC Int. J. Technol. 2 (2011) 60–65. [9] French Official Journal, ministerial order of 28 October 2010 on the storage of inert waste, 2010. [10] B. Sealey, P. Phillips, G. Hill, Waste management issues for the UK ready-mixed concrete industry, Resour. Conserv. Recycl. 32 (2001) 321–331. [11] Standard AFNOR NF EN 206-1, Concrete – part 1: Specification, performance, production and conformity. [12] A. Schwartzentruber, C. Catherine, Method of concrete equivalent mortar (CEM) – a new tool to design concrete containing admixture, Mater. Struct. 33 (2000) 475–482. [13] Standard AFNOR NF EN 12457-2, Characterization of waste – leaching – compliance test for leaching of granular waste materials and sludge – part 2: one stage batch test at a liquid to solid ratio of 10 L/kg for materials with particle size below 4 mm (without or with size reduction). [14] Standard AFNOR NF EN 196-6, Methods of testing cement – part 6: Determination of fineness. [15] Standard AFNOR NF EN 1097-7, Test for mechanical and physical properties of aggregates – part 7: Determination of the particle density of filler pyknometer method. [16] W. Hergert, The Mie Theory – Basics and Applications, Springer, 2012. [17] Standard AFNOR NF P 18-508, Additions for concrete – limestone additions – specifications and conformity criteria. [18] S. Maximilien, J. Péra, M. Chabannet, Study of the reactivity of clinkers by means of the conductometric test, Cem. Concr. Res. 27 (1997) 63–73. [19] Standard AFNOR NF EN 196-9, Methods of testing cement – part 9: Heat of hydration – semi-adiabatic method. [20] P. Siler, I. Kolarova, J. Kratky, J. Havlica, J. Brandstetr, Influence of superplasticizers on the course of hydration of portland cement hydration, Chem. Pap. 68 (2014) 90–97. [21] Standard AFNOR NF P18-459, Concrete – testing hardened concrete -testing porosity and density. [22] Official Journal of the European Union, Council decision 2003/33/EC of 19 December 2002 establishing criteria and procedures for the acceptance of waste at landfills pursuant to Article 16 of and Annex II to Directive 1999/31/ EC, 2003. [23] POYRI SA Étude sur le prix d’élimination des déchets inertes du BTP Technical report, ADEME, 2012. [24] A. Ipavec, R. Gabrovsek, T. Vuk, V. Kaucic, J. Macek, A. Meden, Carboaluminate phase formation during the hydration of calcite-containing Portland cement, J. Am. Ceram. Soc. 94 (2011) 1238–1242. [25] G. Ye, X. Liu, G. De Schutter, A.-M. Poppe, L. Taerwe, Influence of limestone powder used as filler in SCC on hydration and microstructure of cement pastes, Cem. Concr. Res. 29 (2007) 94–102. [26] M. Yudenfreund, K.M. Hanna, J. Skalny, I. Older, S. Brunauer, Hardened Portland cement pastes of low porosity V. Compressive strength, Cem. Concr. Res. 2 (1972) 731–743.