Seediscussions,stats,andauthorprofilesforthispublicationat:http://www.researchgate.net/publication/259678901

Influenceofactivateddrinking-watertreatmentwasteonbinarycement-basedcompositebehavior:Characterizationandproperties

ARTICLEinCOMPOSITESPARTBENGINEERING·DECEMBER2013

ImpactFactor:2.6·DOI:10.1016/j.compositesb.2013.12.020

CITATION

1

DOWNLOADS

13

VIEWS

56

5AUTHORS,INCLUDING:

MoisésFrías

SpanishNationalResearchCouncil

163PUBLICATIONS1,283CITATIONS

SEEPROFILE

IsabelS.DeSotoGarcía

UniversidadPúblicadeNavarra

31PUBLICATIONS25CITATIONS

SEEPROFILE

RosarioGarcía

UniversidadAutónomadeMadrid

117PUBLICATIONS579CITATIONS

SEEPROFILE

TrinoBaloa

CentralUniversityofVenezuela

12PUBLICATIONS9CITATIONS

SEEPROFILE

Availablefrom:TrinoBaloa

Retrievedon:24June2015

Composites: Part B 60 (2014) 14–20

Contents lists available at ScienceDirect

Composites: Part B

journal homepage: www.elsevier .com/locate /composi tesb

Influence of activated drinking-water treatment waste on binarycement-based composite behavior: Characterization and properties

1359-8368/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.compositesb.2013.12.020

⇑ Corresponding author. Tel.: +34 913020440; fax: +34 913020700.E-mail address: mfrias@ietcc.csic.es (M. Frías).

M. Frías a,⇑, R. Vigil de la Villa b, I. de Soto c, R. García b, T.A. Baloa d

a Eduardo Torroja Institute (CSIC), 28033 Madrid, Spainb Dpto de Geología y Geoquímica, Unidad Asociada CSIC-UAM, Universidad Autónoma de Madrid, 28049 Madrid, Spainc Dpto de Ciencias del Medio Natural, Universidad Pública de Navarra, 31006 Pamplona, Spaind División de Estructuras y Materiales, IMME-Facultad de Ingeniería, Universidad de Central de Venezuela, Venezuela

a r t i c l e i n f o

Article history:Received 28 October 2013Received in revised form 15 November 2013Accepted 12 December 2013Available online 23 December 2013

Keywords:A. RecyclingB. Chemical propertiesB. Mechanical propertiesB. MicrostructuresBlended cement

a b s t r a c t

Drinking water treatment plants regularly dispose of large volumes of industrial sludge in landfill sites,which often has negative environmental consequences. The calcination products of these kaolinite-basedsludges have properties that could make them appropriate supplementary cementing materials in theproduction of blended binary cements.

This research analyses the pozzolanic and thermodynamic properties of a Venezuelan drinking watersludge activated at 600 �C for 2 h and its behavior in blended cement matrices prepared with 15% Acti-vated Waste (AW) and 85% Ordinary Portland Cement (OPC). Our results show that this activated drink-ing water sludge presents high pozzolanic properties, mainly during the first 24 h of reaction. The XRD,SEM/EDX and thermodynamic studies confirm the formation of C2ASH8, C–S–H gels and C4AH13 as thehydration products from the pozzolanic reaction. The binary mixture of 15% AW/85% OPC complied withthe physical and mechanical specifications contained in current European cement standards.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Over recent years, fewer high-quality limestone quarries andclays pits have been supplying raw materials to the cement indus-try, due fundamentally to the environmental policies introduced indeveloped countries. This reduction coupled with an imperativeneed to limit the CO2 emissions of the cement industry, hasprompted the search for supplementary cementing materialsextracted from industrial waste and its related sub-products[1–10]. In consequence, research into binary and ternary cement-based composites is a priority research line in the cement sectorat present, due to energetic, economic and technical advantages,stoked mainly by the present global economic crisis [11–14].

In this area, scientific interest has recently been focused on kao-linite-based industrial waste, because it can be converted into ahighly pozzolanic product (metakaolinite-based pozzolan), havingundergone a process of thermal activation [15–22]. The sludgesgenerated in drinking water treatment plants are among this classof waste, generating significant quantities of drinking watersludges, which are normally taken to landfills for disposal, withthe ensuing environmental problems [23]. In general, a drinking-

water treatment plant of 1 m3/s of volume generates about8300 kg/day of sludge [24], which means that the material is avail-able in abundance, bearing in mind that there are 160,000 publicwater systems in the United States alone. In some European coun-tries, only 25% of drinking water sludges are re-used, mainly as rawmaterials in the manufacture of Portland clinker as well as in var-ious other industrial sectors [25–28].

In a previous work by Frías et al. [29], the scientific bases wereestablished for the use of these siliceous–aluminous sludges as apozzolanic material in activated waste systems/Ca(OH)2, whichhighlighted suitable activation conditions at 600 �C and 2 h reten-tion in a laboratory furnace to transform this inert kaolinite-basedsludge into a metakaolinite based pozzolan.

However, there is a large gap in scientific and technical knowl-edge of the properties of binary cement-based composites pre-pared with these activated sludges. Hence, this present studyreports a novel study and a full characterization of industrialsludge activated at 600 �C, its pozzolanic behavior, the evolutionof its mineralogical phases and their influence on the propertiesof the new eco-cement composites. The results obtained from thisresearch are fundamental for the establishment of scientific-technical knowledge bases for the manufacture of future commer-cial composites.

Table 1Chemical and mineralogical composition of OPC.

Compositions Chemical (%) Mineralogical (%)

SiO2 20.16Al2O3 4.36Fe2O3 2.52CaO 63.41 C2S = 8.95MgO 2.21 C3S = 64.75Na2O 0.35 C3A = 7.29K2O 0.91 C4AF = 7.65TiO2 0.21P2O5 0.14S03 3.57LOI 1.99

M. Frías et al. / Composites: Part B 60 (2014) 14–20 15

2. Materials and methods

2.1. Materials

The Starting Waste (SW) for this research came from a Venezu-elan drinking water treatment plant (Embalse La Mariposa), lo-cated at 8 km from Caracas. Its water content was initiallyaround 92%. After drying, the SW was activated at 600 �C for 2 hof retention in an electric laboratory furnace at a heating rate of20 �C/min, following the indications of a previous paper, [29] inwhich these conditions were selected as the suitable activationconditions from an economic and energetic point of view. The acti-vated waste (AW), when cooled to room temperature in a desicca-tor, showed a reddish coloration (Fig. 1). The sample was ground inan agate mortar and pestle and then sieved through a 63 lm sizedmesh. According to the laser granulometry analysis [30], 50% of theparticles were of a size below 9.8 lm, 60% below 12 lm and 97%passed by 45 lm.

2.2. Blended cements

CEM I 52.5 R Ordinary Portland Cement (OPC) supplied from theLafarge Cement Company’s plant at Villaluenga de la Sagra (Toledo,Spain) was used. All the cement particles were under 63 lm and47.41% passed through a 12 lm sieve mesh. The chemical compo-sition and mineralogical of the OPC is given in Table 1.

The blended cements were prepared in a high-speed powdermixer to guarantee homogeneity. The blends were calculated byweight, with AW/OPC ratios of 0/100 and 15/85. This replacementlevel corresponds to the standardized ratios for type II/A cements(6–20%) [31].

2.3. Pozzolanic and thermodynamic methods

Pozzolanic activity method: the pozzolanic behavior in a pozzo-lan/calcium hydroxide (lime) system was studied using the solidsludge waste after applying an accelerated chemical method. Aftereach period of 1, 7, 28 and 90 days of reaction, the sludge waswashed with acetone and dried in an electric oven at 60 �C for24 h, in order to stop the pozzolanic reaction. The content of fixedlime was calculated as the difference between the CaO concentra-tion (mmol/l) in the original saturated lime solution (17.68 mmol/L) and the content of this compound in the solution at each estab-lished time. Extra pure calcium hydroxide Ph Eur, USP, BP chemicalreagent was used.

PHREEQC geochemical software program version 2.18 was usedto evaluate the evolution of the hydrated phases formed in this

Fig. 1. Appearance of AW activated at 600 �C for 2 h.

study, in terms of their thermodynamic behavior [33]. The samplesentered an aqueous species concentration -AW/Ca(OH)2- at differ-ent ages (1, 7, 28 and 90 days of reaction). Even though the pH val-ues were optimized with the geochemical PHREEQC code, due tothe various uncertainties of experimental analyses, rather thanusing the PHREEQC database, our simulation used the THERMOD-DEM database [33] that lists all the minerals presented in thisstudy.

The saturation index of the strätlingite (Ca2Al2SiO2(OH)10�2.5H2O),C–S–H phases with Ca/Si ratios 0.8, 1.2 and 1.6 (Ca0.8SiO2.8�1.54H2O,Ca1.2SiO3.2�2.06H2O and Ca1.60SiO3.6�2.58H2O), C3AH6 (Ca3Al2(OH)12)and C4AH13 [Ca4Al2(OH)14�6H2O], C3AH6 and portlandite Ca(OH)2 werecalculated with the simulation at different ages (1, 7, 28 and 90 days ofreaction), in order to study the stability and evolution of the minerals(dissolution/precipitation) over time. The simulation also allowed usto calculate the activities of the aqueous species. The concentrationsof aqueous species measured in the solution at different times wereintroduced in the model at pH = 12. The aqueous carbonate (not exper-imentally determined in solution) was equilibrated with calcite in themodel.

2.4. Standardized methods

The rheological behavior of the blended cement pastes was as-sessed as described in standard EN 196-3 [34], using a Vicat appa-ratus to determine normal consistency and setting times. Thesoundness of the blended cement pastes was tested with Le Chate-lier apparatus, following the procedure specified in the currentEuropean standard [34]. The effect of this addition on the mechan-ical behavior of new cements was ascertained as per standard EN196–1[35], which describes the methodology for testing mortarcomponents, and their preparation, curing and strength. Blendedcement mortar specimens measuring 4 � 4 � 16 cm were preparedwith a sand/binder ratio of 3/1 and a water/binder ratio of 0.5.

2.5. Characterization techniques

Different techniques were used for chemical, physical, mineral-ogical, morphological and microporosity characterization of thesample. Its chemical composition was studied with a Philips PW-1404 780 X-ray fluorescence analyzer fitted with a Sc–Mo anti-cathode tube. Particle size was analyzed by laser ray diffraction(LRD) on a Sympatec Helos 12 KA spectrometer using isopropylalcohol as the non-reactive liquid. Material mineralogy was deter-mined with X-ray diffraction (XRD) using random powder mountsfor the bulk sample and oriented slides for the <2-lm fraction. Thesamples were analyzed with a SIEMENS D-500 Cu anode diffrac-tometer operating at 30 mA and 40 kV (2-mm divergence slit;0.6-mm reception slit; 2h goniometer; step size: 0.04; count time:3 s). The morphological observations and microanalysis of the sam-ples were performed with a SEM/EDX, by using an INSPECT FEI

Fig. 2. Evolution of pozzolanic activity vs hydration time: activated sludge paper(APS), carbon mining waste (CMW) and fly ash (FA).

16 M. Frías et al. / Composites: Part B 60 (2014) 14–20

COMPANY electron microscope, equipped with an energy disper-sive X-ray analyzer (W source, DX4i analyzer and Si/Li detector).The chemical composition was obtained by an average value often analyses for each sample, in this case the value was the jointstandard deviation. These semi-quantitative analyses wereperformed on clean surfaces to avoid any source of possiblecontamination, such as high calcium oxide concentrations, whichmight have perturbed the assignation of the mineral formula fromthe EDX analysis. The results were expressed in oxides (wt%),adjusted to 100%. The structural formulas were calculated fromthese data, by considering 14 negative charges for C2ASH8

[Ca2Al2(SiO2)(OH)10].An aqueous solution was used to measure the concentration of

aqueous species of interest for the simulation (Na, K, Mg, Ca, Al andSi). These concentrations were determined by inductively coupledplasma mass spectrometry (ICP-MS) with an Elan 6000 Perkin–El-mer Sciex analyzer.

3. Results and discussion

3.1. Characterization of activated drinking water treatment waste(AW)

Table 2 shows the chemical compositions of SW and AW. TheAW waste was mainly composed of SiO2 (41.5%), Al2O3 (33.6%)and Fe2O3 (11.1%), followed by K2O (3.7%), MgO (1.5%) and TiO2

(1.4%), and the other oxides were below 1%. The Starting Waste(SW) presented a high Loss On Ignition (LOI) of 15.32% with respectto that shown by AW (1.99%), which was due, mainly, to the pres-ence of colloidal or dissolved organic matter in water (TOC = 2.3%),kaolinite (11%) and water (1.3%) in the starting waste (SW) [29].These chemical values differ from those values recently publishedfor a Spanish atomized sludge [24], the major oxide contents ofSiO2 (29.63%), Al2O3 (17.57%), and Fe2O3 (5.18%) of which werelower than those reported in the present paper.

This kind of waste also presents traces of heavy metals in vari-able concentrations, mainly chlorides, barium, chromium, andvanadium (Table 2), which could have a direct influence on the ini-tial setting behavior (accelerating or delaying the hydration reac-tion), once mixed with cements. Some of these minor elementsdiffered substantially from those quantified in a Portuguese sludge,mainly in zinc, copper and chromium [36]. This finding impliesthat the sludge characteristics depend on the source water type(surface or groundwater) and the treatments that are employed.

The mineralogical composition obtained by XRD of the AW(Fig. 2) shows the presence of phyllosilicates 2:1 (micas 47%) (M)localised at 10.08 Å (8.76� 2h), 4.48 Å (19.80� 2h), 3.66 Å (24.28�2h), 3.35 Å (26.58� 2h), 2.56 Å (35.02� 2h) and 2.45 Å (36.64� 2h),quartz (17%) (Q) at 4.26 Å (20.82� 2h), 3.34 Å (26.66� 2h) and1.81 Å (50.36� 2h), cation-disordered Ca–Mg carbonates (14%) (C)in the d-spacing range 3.02 Å (29.54� 2h) – 2.99 Å (29.84� 2h),potassium-rich feldspars (Feld-K) (13%) at 13.25 Å (27.44� 2h), pla-gioclases (2%) (Pl) in the d-spacing range 3.19 Å (27.96� 2h) –

Table 2Chemical composition of the starting and activated waste.

Main oxides (%)

SiO2 Al2O3 Fe2O3 CaO MgO Na2O

SW 36.24 29.46 10.05 0.98 1.23 0.83AW 41.54 33.60 11.09 0.42 1.51 0.98

Minor elements (ppm)

Cl Sr Cu Ni Co

SW 663 64 26 76 25AW 211 65 51 81 27

3.21 Å (27.78� 2h) and hematite (7%) (He) localized at 2.69 Å(33.30� 2h)[35].

3.2. Pozzolanic activity

The results of pozzolanic activity for this industrial waste acti-vated at 600 �C and two hrs of retention, expressed as a fixed limepercentage (%), are given in Fig. 2. The AW showed a high reactioncapacity for fixing the available lime (70%), mainly after 24 h ofreaction time. Over longer times, the fixed lime increased from79.8% at 7 days of reaction to 84.1% at the end of test (90 days).This pozzolanic activity of the AW was comparable, mainly after7 days, to other thermally activated kaolinite-based wastes suchas coal mining wastes (CMW) and paper sludge (APS), and itshowed a higher activity than fly ash (FA), a traditional pozzolanin the cement industry, over the first 28 days of reaction [19,38].

3.3. Reaction kinetics and thermodynamic modelling

XRD analysis of the AW/ Ca(OH)2 system (Fig. 3) at 1, 7, 28 and90 days of reaction showed that, in all cases, the main crystallinereaction product detected by XRD was the formation of strätlingite(C2ASH8)(St), with reflections at 12.61 Å (7.0� 2h), 6.28 Å (14.08�2h), 4.15 Å (21.38� 2h) and 2.87 Å (31.14� 2h); followed by tetracal-cium aluminate hydrate (C4AH13) in minor proportions (5%), local-ized at 7.9 Å (11.18� 2h), 2.88 Å (31.02� 2h) and 2.86 Å (31.24� 2h).The chemical reactions involved when calcined MK based clays areused as pozzolans for cements and concretes have been discussedpreviously [39,40]. This reaction between MK and calcium hydrox-ide produced during the cement hydration forms additionalphases: aluminum containing CSH gel, together with crystallineproducts, which include metastable phase such as: calcium alumi-nate hydrates and alumina-silicate hydrates (C4AH13, C2ASH8,), andstable phases (C3AH6–C3ASH6), belonging to the hydrogarnet fam-ily. The crystalline products formed depend principally on the MK/CH ratio and the reaction temperature [41–44].

In alkaline solutions, Al and Si solubilities increase with pH. Dis-solved aluminum can be existed in several forms in solution. In the

K2O TiO2 P2O5 SO3 MnO LOI

3.31 1.23 0.67 0.29 0.38 15.323.67 1.36 0.74 0.31 0.42 3.31

As Ba Cr V Zn Zr

20 233 137 275 82 -21 226 130 263 77 -

Fig. 3. Evolution of mineralogical phases during pozzolanic reaction.

M. Frías et al. / Composites: Part B 60 (2014) 14–20 17

absence of other ligands, the most important are Al3+, AlðOHÞþ2 ,Al(OH)2+, Al(OH)3, and Al(OH)4�. At high pH (pH > 10), Al(OH)4�

is the dominant specie in solution, these ions start to combine withthe readily available Ca2+ ions via metastable phases formingC3AH6. The total dissolved silica concentration in equilibrium withquartz or amorphous silica will increase at high pH values(pH > 10). At these pH values, the total dissolved silica concentra-tion will be the sum of the ionized and un-ionized species(H4SiO4, H3SiO�4 and H2SiO2�

4 ). If the total silica concentration inthe solution becomes supersaturated with respect to amorphoussilica, begins to form polymers which combine with the Ca2+ ionsforming CSH gels [45,46].

These ions react with the C3AH6 initially formed, formingC2ASH8. High MK contents, relative to those with low MK contents,was attributed, in part, to the formation of increased amounts ofC2ASH8 and reduced amounts of C4AH13 as the MK content in-creased [47]. In this study, after activation process of AW at600 �C 2 h, the presence of high metakaolinite (MK) content fa-vored the formation of stratlingite C2ASH8 as the main reactionproduct at all times of the pozzolanic reaction.

The appearance of C4AH13 is attributed to supersaturation of theaqueous phase with respect to calcium hydroxide and low MK con-tents. High concentrations of Ca2+ and OH� in the solution, createdby the Ca(OH)2 dissolution, maintain a composition that enablesprecipitation of C4AH13 [45,46]. In this study, 1 and 7 days of thepozzolanic reaction enables precipitation of C4AH13.

The laminar phases corresponding to C2ASH8 may be observedin the AW/CaOH)2 system by means of SEM. The existence of lam-inar mica and laminar microaggregates of silica enriched composi-tion may also be observed and material deposits on the surface(Fig. 4).

The study of the evolution of the aqueous chemistry of the sam-ples provides useful information on the new hydrated phasesformed in this study. Ca and Si concentrations decreased through-out the reaction and Al concentration decreased at 28 days reac-tion, due to the incorporation of Ca, Si and Al formed in the mainphase (strätlingite and C4AH13) (Fig. 5).

In contrast, an increase of Na and K concentrations over time inthe aqueous solution was studied after the pozzolanic reaction.This constant increase showed a continuous dissolution of thephyllosilicates in the pozzolanic reaction over time (Fig. 5). Finally,Mg concentrations were very low throughout the reaction (Fig. 5).

Saturation indexes of the phases formed in the pozzolanic reac-tion were obtained with PHREEQC [32]. Positive saturation indexvalues (oversaturation conditions) were obtained for stratlingite

C2ASH8 during the whole reaction, which agreed with the experi-mental observations obtained by SEM–EDX and XRD. The simula-tion also confirmed that the precipitation of C4AH13 due to thepositive saturation index calculated at 1 and 7 days reaction(Fig. 6). The thermodynamic analysis predicted under-saturationconditions (negative saturation indexes) for portlandite, C–S–Hgels and C3AH6 phases, according to the experimental data.

3.4. Chemical composition of the blended cement

Table 3 shows the chemical composition by XRF analysis of theblended cement prepared with 15% AW replacement. This blendedcement consisted primarily of CaO (54.1%), SiO2 (23.7%), Al2O3

(8.5%), Fe2O3 (3.9%) and SO3 (3.7–4.2%); followed by MgO (1.9%)and K2O (1.22%). The content of the other oxides was below 0.5%.

Table 3 also compares the sulfate and chloride contents in theAW blended cement analyzed according to the specifications inthe current European standard for type CEM II/A cements [31].The cements prepared with 15% activated drinking water treat-ment waste proved to be standard-compliant.

3.5. Physical properties of the blended cements

Tests on physical properties in line with European standards fo-cused on water content at normal consistency, the determinationof setting times and volume stability [34]. The incorporation ofAW substantially modified the water content needed to reach anormal consistency in the blended cement paste with respect tothe control paste (Table 4). Defining OPC water content as 100%,water content demand increased by 50% following the additionof 15% AW. This fact will be related to a higher fineness of theAW than the OPC (as mentioned above), and to the clayey natureof this activated waste which presents strong water retentionpowers.

With respect to the results obtained for setting times (±10 min),it was clearly detected that the incorporation of AW in the cementpastes neither substantially modified the initial nor the settingtimes. Compared to the control paste, the values found were sim-ilar in both cements and the slight variations that were observedwould be related to testing error. This behavior is contrary to thefindings reported by Frías et al. [48,49] for the cements elaboratedwith 20% activated coal mining wastes, 10% rice husk ash, 10–20%sugar cane bagasse ash, which delayed the initial setting times.

Hence, it is important to note that the presence of low heavymetals concentrations such a barium, chromium, vanadium andzinc in AW ash (Table 2) did not affect the setting times of blendedcement, possibly due to very low Zn content in the sample, ele-ment well known to delay the start of initial setting (<82 ppm).

The expansion test was conducted as specified in EN 196-3 [34].The findings for the three blended cements (average of three mea-surements with a 0.5-mm test error) given in Table 4 show that theaddition of AW had no effect on cement paste soundness.

3.6. Mechanical properties of blended cements

The effect of the addition of AW on the compressive strength ofblended cement mortar (average of six specimens), including thestandard deviation intervals, is depicted in Fig. 7. The AW blendedcement mortars exhibited lower compressive strength than theOPC mortar, a similar tendency to that shown for most pozzolansused in the manufacture of blended cements (fly ash, natural poz-zolan, SiMn slag, ceramic waste) [48,50]. Fig. 8 shows the relativelosses in compressive strength in greater detail.

The blended cement mortars containing 15% AW maintainedrelative strength losses of about 12% during the first 24 h of reac-tion, and subsequently, this loss was reduced by up to 9% with re-

Fig. 4. Morphological aspects of laminar stratlingite, mica aggregates and mica with surface depositation.

Fig. 5. Evolution of the aqueous ions as a function of time.

Fig. 6. Saturation indices of the AW/Ca(OH)2 system over time.

Table 4Values of water content (g), setting times (±10 min) and volume stability (mm).

W. content Initial set Final set Soundness

OPC 150 150 202 0OPC + 15% AW 225 130 216 0Requirements – P45 – 610

18 M. Frías et al. / Composites: Part B 60 (2014) 14–20

spect to the control specimen, at 7 days of curing. For longer hydra-tion times, 15% ASW blended cement mortars increased their rela-tive losses, reaching 16% and 18% with respect to the controlmortar at 28 and 60 days of reaction, respectively. Strength lossin AW blended mortars with rising hydration time may be re-

Table 3Chemical composition of blended cement.

Chemical composition (%)

SiO2 Al2O3 Fe2O3 CaO MgO N

15% OPC 23.70 8.48 3.90 54.10 1.89 0

European standard requirements (%)

SO3: 63.5

garded as an exception among other activated kaolinite basedwastes (paper sludge, coal mining) [48,49] that showed similarand/or slightly higher compressive strengths than the control mor-tar from 7 to 28 days of curing. However, this same tendency wasalso observed in a previous work when drinking water treatmentsludge (activated at 700 �C) was used as an active addition forthe blended mortars manufacture [38]. The addition of 10% Portu-guese activated sludge produced a compressive strength loss of the14.4% at 7 days of reaction, although no justification was given toexplain it.

According to the results shown in the present paper (chemicalcomposition, mineralogy, pozzolanic activity, setting times), no sci-entific explanation for this unfavorable behavior can presently beput forward because of the novelty of the research and the absenceof prior studies. A possible explanation of the important compressivestrength losses of 15% AW blended cement mortars could be relatedto the high water demand needed to reach a normal consistency. Asmentioned above, a 15% AW replacement increased the water con-tent of blended pastes by 50%. However, the blended cement mortars

a2O K2O TiO2 P2O5 SO3 Cl� LOI

.41 1.22 0.39 0.24 3.01 0.01 15.32

Cl�: 60.1

Fig. 7. Evolution of compressive strength up to 60 days of curing.

Fig. 8. Relative compressive strength losses (%).

Table 5Porosity values at 60 days of curing.

P. total (%) Av. pore diam. (lm)

OPC 11.80 0.067OPC + 15% AW 12.93 0.084

Fig. 9. Comparison of pore-size distribution of OPC and 15% AW blended cementmortars at 60 days of reaction.

M. Frías et al. / Composites: Part B 60 (2014) 14–20 19

had a consistency (tested in flow table) of 15% prepared with a con-stant water/binder ratio (w/b = 0.5) [37]. Notwithstanding the above,all the AW blended mortars had 28-day compressive strength valuesthat were above the requirements of the existing European standard(P52.5 MPa) for class 52.5 R cements.

3.7. Microporosity of cements

Total porosity, average pore diameter and pore size distributionwere analysed by mercury porosimetry, in order to determine the

effect of addition of 15% of AW on microprorosity of mortars curedat 60 days. The total porosity results (Table 5) showed that theaddition of 15% AW did not affect practically to the microporosityof the cement mortars; the 15% AW blended mortar (12.9%) obtain-ing a slightly higher value than the control mortar (11.8%) inmortar cured at 60 days. The distribution density curves (Fig. 9)confirmed this tendency, in which there is a well defined maxi-mum in each case, localised at 0.11 and 0.13 lm for the OPC and15% AW mortars, respectively. These findings are totally in agree-ment with the average pore diameter results (4 V/A), showing val-ues of 0.067 and 0.084 lm for percentages of 0 and 15% ofactivated waste.

4. Conclusions

The conclusions drawn from the experimental results are asfollows.

1. The drinking water treatment waste activated at 600 �C for 2 hshowed a siliceous–aluminous nature with a Fe2O3 content ofabout 11% and minor elemental contents of over 100 ppm (Cl,Ba, Cr, V). Mineralogically, the AW is formed by phyllosilicates2:1 (micas), quartz, cation-disordered Ca–Mg carbonates,potassium-rich feldspars, plagioclases and hematite.

2. The AW activated at 600 �C and 2 h of retention showedmaximum pozzolanic activity during the first 24 h of reaction,taking the fixed lime percentages (70% of available lime) intoaccount.

3. Strätlingite (C2ASH8) was the main crystalline reaction productdetected through XRD and tetracalcium aluminate hydrate(C4AH13) in minor proportions (5%) throughout the pozzolanicreaction. The laminar phases corresponded to C2ASH8 andSEM analysis revealed the existence of laminar mica and lami-nar micro-aggregates of silica enriched composition with super-ficial deposits.

4. The precipitation of strätlingite during the pozzolanic reactionwas calculated with PHREEQC. The results indicate that stra-tlingite was the main mineral phase produced during the poz-zolanic reaction, according with the experimental XRD andSEM data.

5. The addition of 15% AW to the cement paste raised the waterdemand by 50% and a consistency reduction of about 38% inblended mortars.

6. The 15% AW cement pastes complied with standardized chem-ical requirements in terms of sulfate and chloride content of thecements type CEM II/A, and with physical requirements (initialsetting time and volume stability).

7. The addition of 15% AW in the blended cement mortars manu-facture produced compressive strength losses of between 12%and 18% in function of the cured age. In spite of this, all 15%AW blended cement mortars complied with the mechanicalrequirements specified in the existing standards.

8. In general, the incorporation of AW affected slightly to the totalmicroporosity and the average size diameter values of theblended cement mortars. A small increase in the average diam-eter was observed for the 15% AW cement mortar with respectto the control mortar, passing from 0.067 lm for the OPC to0.084 lm for the 15% AW blended mortar.

In summary, the use of activated drinking water treatmentwaste as an active pozzolanic addition to blended cements is a so-lid scientific option that is technically viable. Further research (intodurability, higher AW percentages, and the use of water reduceragents) is necessary before this knowledge may be transferred tothe industrial sector.

20 M. Frías et al. / Composites: Part B 60 (2014) 14–20

References

[1] Siddique A. Waste materials and by-products in concrete. Berlin: Springer;2008.

[2] Frías M, Vegas I, Vigil R, García R. Recycling of paper sludge in cements:characterization and behaviour of new eco-efficient matrices. In: Kumar S,editor. Integrated Waste Management. vol II. Croatia: 2011, p. 301–318.

[3] Ibrahim R, Hamid R, Taha MR. Fire resistance of high volume fly ash mortarswith nanosilica addition. Const Build Mater 2012;36:779–86.

[4] Arizzi A, Cultrone G. Aerial lime-based mortars blended with a pozzolanicadditive abd different admixture: a mineralogical, textural and physical–mechanical study. Constr Build Mater 2012;31:135–43.

[5] Frías M, Sánchez de Rojas MI. Microstructural alterations in fly ash mortars:study on phenomena affecting particle and pore size. Cem Concr Res1997;27(4):619–28.

[6] de Lomas M, Sánchez de Rojas MI, Frías M. Pozzolanic reaction of a spent fluidcatalytic cracking catalyst in FCC cement mortars. J Therm Anal Calorim2007;35(11):2137–42.

[7] Vigil R, Fernández R, Rodriguez R, García R, Villar E, Frías M. Evolution of thepozzolanic activity of a thermally treated zeolite. J Mater Sci2013;48:3213–24.

[8] Najigivi A, Khaloo A, Iraji A, Abdul S. Investigating the effects of using differenttypes of SiO2 nanoparticles on the mechanical properties of binary blendedconcrete. Composites Part B 2013;54:52–8.

[9] Sanjuan MA, Argiz C. The new European standard on common cementsspecifications EN 197-1:2011. Construcc Mater 2012;62(307):425–30.

[10] Argiz C, Menéndez E, Sanjuan MA. Effect of mixes made of coal bottom ash andfly ash on the mechanical strength and porosity of Portland cement. Construccmater 2013;63(309):49–64.

[11] Frías M, Goñi S. Accelerated carbonation effect on behaviour of ternaryPortland cements. Composites Part B 2013;48(may):122–8.

[12] Goñi S, Frías M, Vegas I, García R, Vigil R. Quantitative correlations amongtextural characteristics of CSH gel and mechanical properties: case of ternaryPortland cements containing activated paper sludge and fly ash. Cem ConcrComp 2012;34(8):911–6.

[13] Medina C, Banfill PFG, Sánchez de Rojas MI, Frías M. Rhelogical andcolorimetric behaviour of cements blended with containing ceramic sanitaryware and construction/demolition waste. Constr Build Mater2013;40(march):822–31.

[14] Madani M, Shao Y, Whalen JK. Biocement production from silicon-rich plantresidues: perspectives and future potential in Canada. Biosyst Eng2011;110:351–62.

[15] Lin DF, Huang LS, Luo HL, Weng RS. Effects on cement after partial replacementwith burned joss paper ash. Environ Technol 2012;33(22):2595–601.

[16] Yan S, Sagoe-Crentsil K, Shapiro G. Reuse of de-inkg sludge from wastepaperrecycling in cement mortar products. J Environ Manage 2011;92:2085–90.

[17] Banfill PFG, Frías M. Rheology and conduction calorimetry of cement modifiedwith clacined paper sludge. Cem Concr Res 2007;37(2):184–90.

[18] Vegas I, Urreta J, Frías M, García R. Freeze–thaw resistance of blended cementscontaining calcined paper sludge. Constr Build Mater 2009;23(8):2862–8.

[19] Frías M, Vigil R, Sanchez de Rojas MI, Medina C, Juan A. Scientific aspects ofkaolinite based coal mining wastes in pozzolan/Ca(OH)2 system. J Am CeramSoc 2012;95(1):386–91.

[20] Frías M, Martínez S, Blasco T, Frías-Rodríguez M. Evolution of mineral phasesby 27Al and 29Si NMR in MK-Ca(OH)2 system cured at 60 �C. J Am Ceram Soc2013;96(7):2306–10.

[21] Frías M, Villar E, Savastano H. Brazilian sugar cane bagasse ashes from thecogeneration industry as active pozzolans for cement manufacture. Cem ConcrComp 2011;33(4):490–6.

[22] Resmini C, Anglioletto E, Gracher H, Peterson M, Rocha MR, Resmini A, et al.Production of MK from industrial cellulose waste. J Therm Anal Calorm2012;109:1341–5.

[23] Kyncl M. Opportunities for water treatment sludge re-use. GeoSci Eng2008;54(1):11–22.

[24] Verlicchi P, Masotti L. Reuse of drinking water treatment plants sludges inagriculture: Problems, perspectives and limitations. Google website:www.ramiran.net/doc; 2012.

[25] Chen H, Ma X, dai H. Reuse of water purification sludge as raw material incement production. Cem Concr Comp 2010;32(6):436–9.

[26] Husillos N, Martínez S, Blanco MT, Guillem M, Puig J, Larrotcha E, et al.Evaluation of spary dried sludge from drinking wáter treatment plants as aprime material for clinker manufacture. Cem Concr Comp 2011;33(2):267–75.

[27] Liang Ch, Tseng DH, lin TT. Characterization of eco-cement paste producedfrom waste sludges. Chenosphere 2011;84(2):220–6.

[28] Lin KL, Lin CY. Hydration characteristics of waste sludge ash utilized as rawcement material. Cem Concr Res 2005;35:1999–2007.

[29] Frias M, Vigil R, García R, Sánchez de Rojas MI, Baloa TA. Mineralogicalevolution of kaolin based drinking water treatment waste for use aspozzolanic material. The effect of activation conditions. J Am Ceram Soc2013;96(10):3188–95.

[30] Frías M, Sánchez de Rojas MI, Lúxan MP, García N. Determination of specificsurface area by the laser diffraction technique. Comparison with the Blainepermeability method. Cem Concr Res 1991;21(5):709–17.

[31] UNE-EN 197-1. Cement Part 1: composition, specifications and conformitycriteria for common cements, 2011.

[32] Parkhurst DL, Appelo CAJ. User’s guide to PHREEQC (Version 2) – a computerprogram for speciation, batch-reaction, one-dimensional transport, andinverse geochemical calculations.US Geological Survey, Reston, VA. 1999.

[33] Blanc P, Lassin A, Piantone P. Thermoddem a database devoted to wasteminerals. Orléans, France: BRGM. <http://thermoddem.brgr.fr>. 2007.

[34] UNE EN 196-3. Methods of testing cement. Part 3: Determination of settingtime and sounness; 2005.

[35] UNE EN 196-1. Methods of testing cement. Part 1: Determination of strength;2005.

[36] Goncalves A, Esteves A, Carvalho M. Incorporation of sludges from a watertreatment plant in cement mortars. In: Proceeding of Inter Conference on theUse of Recycled Materials in Building and Structure, vol. 1, Edited by RILEM,Barcelona, 2004, p. 347–350.

[37] Brindley GW, Brown G. Crystal structures of clay minerals and their X-rayidentification. Mineralogical Society Monograph no. 5. London, 1980.

[38] Rodríguez O, Frías M, Sánchez de rojas MI. Influence of the calcined papersludge on the development of hydration heat in blended cement mortars. JTherm Anal Calorim 2008;92(3):865–71.

[39] de Silva PS, Glasser FP. Hydration of cements based on metakaolin:thermochemistry. Adv Cem Chem 1990;3(12):167–77.

[40] Dunste AM, Parsonage JR, Thomas MJK. The pozzolanic reaction of metakaolinand its effects on PC hydration. J Mater Sci 1993;28:1345–50.

[41] Frías M, Sanchez de Rojas MI. The effect of high curing temperature on thereaction kinetics in MK/lime and MK-blended cement matrices at 60�C. CemConcr Res 2003;33:643–9.

[42] Frías M. Study of hydrated phases present in a MK-lime system cured at 60 �Cand 60 months of reaction. Cem Concr Res 2006;36(5):827–31.

[43] Wild S, Khatib J, Roose LJ. Chemical and autogenous shrinkage of Portlandcement-metakaolin pastes. Adv Cem Res 1998;10(3):109–19.

[44] Khatib JM, Sabir BB, Wild S. Some properties of metakaolin paste and mortar.In: Dhir RK, Dyer TD, editors. International congress, concrete in the service ofMankind-concrete for environmental enhancement and protection. Theme 6waste materials and alternative product. London: E & F N Spon; 1996. p.637–44.

[45] McBride MB. Environmental chemistry of soils. New York, Oxford: OxfordUniversity Press; 1994.

[46] Drever JI. The geochemistry of natural waters. 2nd ed. New Jersey07632: Prenyice Hall. Englewood Cliffs; 1988.

[47] Sabir BB, Wild S, Bai J. Metakaolin and calcined clays as pozzolans for concrete:a review. Cem Concr Comp 2001;23:441–54.

[48] Frías M, Sánchez de Rojas MI. Natural and artificial pozzolans in eco-efficientconcrete. In: Pacheco F, Jalali S, Labrincha J, John VM, editors. Eco-efficientconcrete. Cambridge: Woodhead Publishing Limited; 2013. p. 105–22 [chapter5].

[49] Frías M, Sánchez de Rohas MI, García R, Juan A, Medina C. Effect of activatedcoal mining wastes on the properties of blended cement. Cem Concr Comp2012;34(5):678–83.

[50] Taylor HFW. In: Cement chemistry. Thomas Telford Publishing. ThomasTelford Services Ltd.: London; 1997.