Magnetic susceptibility measurements on metakaolin admixtured cement hydrated with ground water and

International Journal of Minerals, Metallurgy and MaterialsVolume 16, Number 3, June 2009, Page 349MaterialsEL SEVIERMagnetic susceptibility measurements on metakaolin admixturedcement hydrated with ground water and sea waterD. Govindarajan and R. GopalakrishnanDepartment of Physics, Annamalai University, Annamalai Nagar, 608002 Tamilnadu, India(eceived 2008-06-27)Abstract: The role of metakaolin in the properties of Porland cement hydrated with ground water and sea water was described bymagnetic susceptibility study. Cement pastes containing 0wt%, 10wt%, 20wt% and 30wt% replacement of metakaolin and in a wa-ter/cement (W/C) ratio of 0.4 were prepared. The susceptibility at different hydration periods was determined by Faraday Curie bal-ance and it was related to the changes in setting time and compressive strength of admixtured cement. Compared with sea wa-ter-treated cement paste, the magnetic susceptibility of ground water-treated cement paste is higher in value. The observed resultshows that, irrespective of water, the magnetic susceptibility increases with increasing metakaolin percentage replacement level incement.Key words: magnetic susceptibility; admixtured cement; hydration; pozzolanic activity; sea water1. Introductionral calamities (floods, delay of monsoon, tsunami andearthquake etc,) the potable water quality and theCement is a finely ground, usually grey coloredavailability of potable water have decreased. Undermineral powder. Portland cement is mainly composedthese circumstances, using other alternatives like rainof tricalcium silicate, dicalcium silicate, tricalciumwater, river water and one of the natural potential ie,aluminate and tetracalcium aluminoferrite, which reactsea water can also be thought of, provided it is harm-with water in the presence of gypsum, producing a va-less. The influence of different waters on cement hy-riety of hydration products. In general tricalcium sili-dration was studied by Barathan et al. [6] with thecate and dicalcium silicate react with water to formhelp of X-band microwave technique. They have re-calcium silicate hydrate (C-S- H) and calcium hydrox-ported that the cement mixing with sea water has theide (CH). Tricalcium aluminate and tetracalcium alu-highest value of dielectric constant (&) and electricalminoferrite react with gypsum and water producingconductivity (), and then ground water, and then dis-ettringite (AF) and monosulphate (AFm).tilled water. According to Manu Santhanam et al. [7],The replacement of cement by pozzolanic materialsthe high chloride concentration of sea water couldis advantageous such as improving the cementitioushave played an important role in the C3A reaction.properties, reducing the total energy, production cosPalomo et al. [8] reported that prisms of mortar madeand CO2 emission. One such pozzolan is metakaolinof sand and alkali-activated metakaolin were im-obtained under a carefully controlled process of ther-mersed in deionized water, sea water, sodium sulphatemal activation of kaolinite clay [1]. The kaolinite issolution and sulphuric acid solution. It was observeddehydrated at 600-700°C, which causes major struc-that the nature of the aggressive solution had littletural disorganization and significant increase in poznegative effect on the evolution of microstructure andzolanic activity. The effects of metakaolin on the mi-the strength of these materials.crostructure, thermal properties and mechanical prop-Cement notations: C=CaO; S =SiO2; H=H2O; .erties of pastes and mortar have been widely reported C3S= 3CaO.SiO2; C2S= =2CaO SiO2; C3A=3CaOAl2O3;[2-5].C4AF=4CaOA中国煤化工In India, because of the climatic changes and natu-As the cem(| YHCNMHGthereactionCorresponding author: D. Govindarajan, E-mail: degerajan@redifmail.comavailable online at www.sciencedirect.com◎2009 University of Science and Technology Beijing. AIl rights reserved..350International Journal of Minerals, Metallurgy and Materials, Vol.16, No.3, Jun 2009products vary with time from one phase to another andof Physics, Annamalai University, India.hence there is every possibility of change in magneticsusceptibility. Susceptibility (x) measurements ar2. Experimentalused to detect the solubility limit of iron in different2.1. Materials and methodsforms in CzA [9]. According to David et al. [10], clayminerals have a higher amount of oxides and henceCommercial metakaolin (MK) and ordinary Port-this factor enhanced the susceptibility values. Sagnottiand cement (OPC) were used in the present work.et al. [11] reported the inter-laboratory calibration ofBoth samples were subjected to chemical analysis us-low-field magnetic and anhysteretic susceptibilitying standard procedure and the components (wt%)measurements of paramagnetic rare earth salt, Gd2O3were as follows. Cement: CaO 63.32, SiO2 21.70,and commercial cement. Bidegain et al. [12] have re-Al2O3 5.40, Fe2O3 3.40, MgO 2.69, SO3 2.70, Loss onported that the magnetic constituents of Portland ce-ignition 0.79. Metakaolin: CaO 0.50, SiO2 68.54,ment paste and mortar acquire a magnetic remanenceAl2O3 23.55, Fe2O3 0.58, MgO 0.60, MnO 0.02, Lossbecause of their alignment with the earth's magneticon ignition 3.98, insoluble residue 1.20. Ground waterfield at the casting place. The measurement of mag-(GW) and sea water (SW) were analyzed using stan-netic susceptibility clearly shows the variation of ce-lard procedure recommended by Tamilnadu W atermentitious properties with time.Supply and Drainage Board, Cuddalore, Tamilnadu,India (Table 1).As far as we are aware, magnetic studies on hydra-tion of cement are limited and no magnetic studiesIn this investigation, pastes were prepared with ahave been reported so far on hydration of metakaolinwater/cement ratio of 0.4 using ground water and seaadmixtured cement hydrated with ground water andwater. Cement was partially replaced with differentsea water. Several magnetic methods are available focontent of MK (0wt%, 10wt%, 20wt% and 30wt%).measuring magnetic susceptibility but Faraday-CurieSamples hydrated for different periods, namely, im-balance is a very sensitive and suitable method tomediate, every 30 min for 10 h after mixing and then 1measure the susceptibility of even a very small quan-d, 2 d, 4d, 1 week, 2 weeks and 4 weeks were sub-tity of sample. In the present study, the authors prejected to acetone and last oven-dried at 105°C for 1 h .sented magnetic susceptibility measurements on ce[13]. The dried samples were powdered and used forment and metakaolin admixtured cement paste hymagnetic susceptibility measurements. Setting time,drated with ground water and sea water at differentcompressive strength on metakaolin admixtured ce-hydrated time intervals using Faraday-Curie balancement are measured [14] and shown in Table 2.available in Material Science Laboratory, DepartmentTable 1. Contents of solvents of sea water and ground water (ug"g )WaterCl_MgCa____ sNaSea water1640010802725.809100Ground water70682.5055Table 2. Setting time, compressive strength of metakaolin admixtured cement treated with ground water and sea waterMK content /Setting timeCompressive strength 1 MPawt%StartEnd11 week4 weeks4:45:006:35:000.634.348.54:15:006:15:001.242.358.2GW203:15:005:30:0024.554.571.5__32:15:004:30:0032.458.404:10:006:20:0013.134.433.11(3:55:006:00:0023.743.441.722:40:005:05:0025.55.531:55:0034.59.257.52.2. Experimental procedureend is soldered to a torsion head (T). The sampleFig.1 shows the Faraday-Curie balance along withholder is conne中国煤化IThe glass tubeHeyding's type pole pieces [15]. It consists of a light(g) suspendedaluminiumaluminium beam (AB) with a length of 25 cm sus-beam. A smallTMYHCNMHGthecenterofpended by means of a fine copper wire and its upperthe beam. A circular scale (S) graduated in degree.D. Govindarajan et al, Magnetic susceptibility measurements on metakaolin admixtured...351provided at the upper end of the glass tube facilitatesNiSO4, CaSO4:7H2O, FeSO47H2O, MnSO4, Gd2O3,the rotation of the sample from 0° to 360°. The rightFe(OH)3, HOCl3, Er2O3) and the corresponding annul-end of the arm carries a damping vane (V) made ofling voltage per unit mass.aluminium and is dipped in a dash-pot of light viscous40 -oil. This arrangement effectively reduces all externalvibrations. Above the aluminium beam there is a bal-ancing coil (C) comprising of 500 turns of copper wirewound over a hollow rectangular frame. The powerful20 Fhorse shoe-shaped permanent magnet (MM) produc-ing a strong radial field encloses the coil‘C'. The coilmoving in the radial field constitutes a galvanometer.By passing a suitable current through the coil usingpotential divider arrangement, the deflection can be :0200nullified. A digital millivoltmeter measures the currentSusceptibility/(10-6 emug -.m-)flowing in the balancing coil and directly displays theFig. 2. Calibration graph.value in terms of voltage [16].The magnetic susceptibility of different time inter-?Tvals of hydrated cement is obtained from the calibra-tion graph, by measuring the annulling voltage. Thesensitivity of the instrument is +0.001. All the meas-MoMurements were taken at 28+2°C.3. Results and discussionThe observed susceptibility values for anhydrous上81cement and metakaolin a51.05x10- and123.02x10 6 emu/(g:m) respectively. x>0 shows thatthe cement and metakaolin samples are paramagnetic[17-18]. The positive paramagnetic susceptibility ofthe anhydrous cement may be considered as a resul-Fig. 1. Faraday-Curie balance.tant of the contribution from all the impurity ions andespecially Fe2O3 contained in C3A and C4AF phases2.3. Calibration[19]. Graphs are drawn between the observed suscep-The Curie balance has to be calibrated for varioustibility values and different hydrated time intervals forsamples of known susceptibility values. The sample incement and metakaolin admixtured cement treatedpowder form is flld in sample holder and is sus-with GW (Fig. 3) and SW (Fig. 4) in a non-uniformpended in the predetermined region where magneticscale. For ground water hydrated cement paste (GC inforce is uniform and maximum. When the magneticFig. 3), the observed susceptibility values start to in-field is switched on, the sample experiences a mag-crease and reach the first maximum value during thenetic force. This produces a deflection of the beamfirst 2 h. When cement is mixed with water, becauseand the deflection is noted by mirror (m) and the lampof initial dissolution, the concentrations of hydroxyland scale arrangement. The force exerted on the sam-(OH ) and calcium (Cat ) ions increase and the firstple is counter-balanced passing a suitable current, byreaction product ettringite (AF) forms with a layer ofadjusting the resistance in the potential divider ar-hydrous alumina from the reactions between C3A,rangement and hence the voltage required to compen-C4AF and gypsum by the following equations [20]:sate the magnetic force on the sample is noted. Thesample is rotated in a step of 20° and at each positionC3A+ 3CSH2+ 26H-→C,AS;H32 .(1the procedure is repeated and finally the average an-C4AF + 3CSH2 +30H -→C。AS;H32 +CH+ FH3(2nulling voltage Vs is calculated. The experiment is re-peated for empty sample holder to determine its an-where C3A is tricalcium aluminate， 3CSH2 isnoulling voltage Vo. The difference of voltage (Vs-V)gypsum, H. is water, CoAS,H3 is ettringite,is a measure of annulling voltage for the sample. ThC4AF is t中国煤化工FH3 is fer-calibration graph (Fig. 2) is drawn between the knownric-aluminium IHCNM HGsusceptibility value of different samples (K2Cr2O7,Initially, C3A phase plays a dominant role in the.352International Journal of Minerals, Metallurgy and Materials, Vol.16, No.3, Jun 2009setting of cement. The variation of magnetic suscepti-calcium silicate hydrate gel (C-S-H), Ca(OH)2 andbility with time may be attributed to the type and con-hence the cement paste becomes condensed andcentration of iron added to CzA. The solubility of ironthickened [17] according to the following equations:in C3A is the main cause for increase in x values ob-2C2S+4H→CzS2H3 +CH(4)served up to 2 h. Moreover, Mohamed Ahmed et al,[9] have pointed out that the Fe2O3 has a tendency to2CgS+ 6H→C3S2H3 +3CH(5)dissociate and liberate Fe ion in the form of Fe' anFe'+t. This tendency may increase the solubility limitwhere C2S is dicalcium silicate, CzS is tricalcium sili-cate，C3S2H3 is C-S-H, and CH is calcium hydrox-of Fe in C3A. Initially, the active substitution of Fe'ide.for Al+ in C3A lttices causes a steep increase in xvalues. This replacement was confirmed by MajumdarThe C-S-H gel is one of the major strength render-[21], Moore [22], and Schlaudt and Roy [23].ng components of the hydrated cement and it pos-sesses sparse dispersibility, numerous interior struc-300tural effects and coarser particle surfaces. These char-250acteristics decrease the magnetic susceptibility but in-,GMK 20wt%crease the strength of cement.GMK 30w%The magnetic susceptibility of MK admixtured ce-150个GMK 10w%-ment treated with GW (GMK in Fig. 3) also follows100the same trend as that of the ground water-treated ce-ment paste. From Fig. 3, as the replacement of MK in50cement increases, the x values also increase. There isa clear trend observed that both first minimum and0h1h2h4h6h8h9h1d7d28dstarting of final setting time was shortened as the re-Hydration time (non-linear scale)placement of MK increased. This accelerating effectFig. 3. Magnetic susceptibility vs. hydration time of meta-of MK is probably owing to the dispersing effect ofkaolin admixtured cement paste treated with ground waterMK on the agglomerated cement grains that act as nu-After the first maximum, x values begin to decreasecleating agent. Further, MK contains a higher amountgradually with time and reach a minimum at a par-of silicon and aluminium oxides, they can also accel-ticular time interval. Because the reaction betweenerate the reaction with hydration of cement, increaseC3A, C4AF and gypsum was slightly retarded by hy-the formation rate of AF and hence shorten the settingdrous alumina, the formations of hydration productstime [18]. The increase and decrease of x values fromare less paramagnetic than reactants [24]. This is anstart to the first minimum are faster when compared toindication that realignment of hydroxyl group is tak-cement pastes (Fig. 3-GC). This result accords with Le-ing place where paramagnetic particles are interlocked.coanet et al. [25] reporting that quickly setting cementAs hydration proceeds CzA contain more Fe + thanhas high susceptibility value.Fe'+ and Fe2+ substutes Ca2+ in C3A lttices. ThisMK reacts with calcium hydroxide produced in thepartial reduction ofFef and Fe' leads to the decrease hydration of Portland cement. Ca(OH)2 cannot di-of X values after the first maximum. The time takenrectly increases the strength of the cement paste. It canfor this paramagnetic susceptibility value reaching ado so only after it was translated into C-S-H gel byminimum coincides well with the starting of final set-pozzolanic reaction with active minerals [18]. Theting time of that particular sample measured usingmore of Ca(OH)2 was consumed, the more C-S-H wasVicat apparatus (Table 2). After the first minimum theformed and it increases the strength of cement but de-x values increase slightly and then decrease formingcreases the susceptibility values at a higher rate afterthe second maximum. This may be due to a suddenthe second maximum.increase in the reactivity of unreacted C3A. At theSea water-treated pastes (Fig. 4) also show thestarting of final setting time, the rapid conversion ofsame trend as that of the ground water-treated pasteAF into AFm takes place and increase the iron content.(Fig.3) but the susceptibility values are lower. ThisC6AS;H32 + 2C3A +4H→3C4ASH12(3)may be attributed to sea water containing moreamounts of diamagnetic impurities and it decreaseswhere 3C4ASH12 is monosulphate.the magnetic中国煤化r1agnetic suS-After the second maximum, the susceptibility val-ceptibility sh:ter with lessues decreases gradually with hydration time. This canparamagnetismlMYHCNMHTmpurities inbe due to the formation of hydration products such asthe solid solution made up of seawater. In addition to.D. Govindarajan et al, Magnetic susceptibility measurements on metakaolin admixtured...353MK, sea water also accelerates the cement hydration300due to the mineral content available in the water. TheSMK 20wt%acceleration depends on the chloride, sodium and sul-phate ions [6, 26]. This effect reduces the thickening, SMK 30wt%5 200time, enhances the hydration reaction and the con-secutive strength development. At the 4th week com-150SMK 10wt%pressive strength is slightly decreased compared toground water. This may be attributed to the reductionofCr, Nat and Ca2+ ions and the reaction of C2S withsea water, which occurred earlier and hence decreasedstrength in the later stage.oh 1h2h 4h 6h 8h 9h1d 7d 28dTo have a better view in the changes of x values,Hydration time (non-linear scale)the X values Vs. strength and MK replacement perFig. 4.Magnetic susceptibility vs. hydration time of met-centage have been plotted and shown in Figs. 5-6.akaolin admixtured cement paste treated with sea water.a)-.◆..GC|(b).. SMK 10wt%E 160..■. . GMK 20w%t 160a .. SMK 20wt%- - GMK 30wt%气官120-.. SMK 30wt%SMK 10w% (x)- 120. GMK 10wt% (). SMK 20wt% (x)120一一GMK 20wt%(2)-120 2, SMK 30wt% (x)80:140-417d28 d028dHydration timeFig. 5. Relationship of hydration time vs. magnetic Susceptibility (x) and compressive strength for different MK replace-ments in cement: (a) treated with ground water; (b) treated with sea water.Anhydrousatic increase in compressive strength and decrease in★-GW2hmagnetic susceptibility are due to high pozzolanic ac-一一- GW first minimumtivity of metakaolin. As MK percentage replacementGW 4 weekslevel increases in cement, the susceptibility values-▲--SW 2h言.sw fist mnimumalso increase irrespective of water and at the 4th week三-. -SW 4 weeksx values fall below the anhydrous (Fig. 6). .4. Conclusions(1) Before the setting time, the Portland cement hasdissolved and reacted well, hydration products dis-100perse easily and result in high magnetic susceptibilityvalues. After the setting time, the cement gel particles .condense and thicken, x values start decreasing andstrength effect becomes starting. This trend is the3(same irrespective of samples mixed with ground waterMetakaolin replacement / wt%or sea water.Fig. 6. Variation of magnetic susceptibility with MK re-(2) The qua中国煤化工: susceptibil-placement.ity shows also:CHCNMHGity of water.After 1 d, the graph clearly depicts that the system-Because sea water contans more amounts of diamag-.354International Journal of Minerals, Matallury and Materials, Vol.16, No.3, Jun 2009netic impurities, the magnetic susceptibility valuesApplication of magnetic susceptibility to study low irondecrease compared with that of ground water-treatedsubstitution in tricalcium aluminate, J. Am. Ceram. Soc,78( 1995), No.7, p.1958.samples.10] C. David, L. Louis, V. Metz, P. Robion, B. Menendez,(3) Irrespective of water, the first minimum valuesand C. Kissel, Microstructural control on the anisotropy ofof all samples coincide with the starting of final set-elastic and transport properties in undeformed sandstone,ting time.Int. J. Rock Mech. Min. Sci, 42(2005), p.911.11] L. Sagnotti, P. Rochette, M. Jackson, F. Vadeboin, J. Di-(4) At all ages, the magnetic susceptibility valuesnares-Turell, and A. W inker, Inter-laboratory calibrationincrease with an increase in MK replacement level.of low-field magnetic and anhysteretic susceptibilitymeasurements，Phys. Earth Planet. Inter., 138(2003),(5) In addition to MK, sea water also acceleratesp.25.cement hydration, hence shortens the setting time and12] J.C. Bidegain, A.M. Sinito, and M.E. Rodriguez, Rema-increases the compressive strength compared withnent magnetization in Portland cement-based materials,ground water. It is evidenced that the mineral contentStud. Geophys. Geod, 43(1999), No.3, p.289.in water plays a major role in setting time, compres-13] N. Voglis, G. Kakali, S. Chaniotaki, and S. Tsivilis, Port-sive strength and magnetic susceptibility.land-limestone cements: their properties and hydrationcompared to those of other composite cements, Cem.(6) Magnetic susceptibility study is a confirmationConcr. Res, 27(2005), p.191.of the well-known accelerating effect of MK and sea14] M.S. Shetty, Concrete Technology, S. Chand and Co.water in the hydration of the Portland cement.Ltd, New Delhi, 2004, p.62.15] R.D. Heyding, J.B. Tavlor, and M.L. Hair, Four-inchAcknowledgementshaped pole caps for susceptibility measurements by theCurie method, Rev. Sci. Instrum, 32(1961), No.2, p.161.The authors wish to acknowledge M/S.20 Microns[16] L.F. Bates, Modern Magnetism, Cambridge UniversityLimited, Vadodara, India, for supplying the metakao-Press, New York, 1961, p.109.lin used in this investigation.17] Z. Wang, L. Sun, Y. Hou, and E. Tang, In situ monitoringC-S-H gel processing by magnetic susceptibility and tur-Referencesbidity measurements, J. Am. Ceram. Soc., 88(2005), No.7,p.1799.[1] N.J. Saika, P. Sengupta, P.K. Gogoi, and P.C. Borthakur,18] Z. Li and Z. Ding, Property improvement of PortlandHydration behaviour of lime co-calcined kaolin-petroleumcement by incorporating with metakaolin and slag, Ceneffluent treatment plant sludge， Cem. Concr. Res.,Concr. Res., 33(2003), p.579.32(2002), p.297.19] H.W.F. Taylor, Cement Chemistry, Academic Press Inc.,[2] M.F. Rojas， Study of hydrated phases present in aNew York, 1990, p.31.MK-lime system cured at 60°C and 60 months of reaction,20] M.Y.A. Mollah, W. Yu, R. Schennach, and D.L. Cocke,Cem. Concr. Res, 36(2006) p.827.Fourier transform infrared spectroscopic investigation of[3] C.A. Love, L.G. Richardson, and A.R. Brough, Composi-the early hydration of Portland cement and the influencetion and structure of C-S-H in white Portland cement 20%of sodium lignosulfonate, Cem. Concr. Res., 30(2000),metakaolin pastes hydrated at 25°C, Cem. Concr. Res.,p.267.37(2007), No. 2, p.109.[4] NJ. Coleman and W.R. Mcwhinnie, The solid state[21] A.J Majumdar, The frrite phase in cements, Trans. Br.Ceram. Soc., 64(1965), p.105.chemistry of metakaolin-blended ordinary Portland C[22] A.E. Moore, Binary system 3CaOAl2O3-Na2O-ment, J. Mater: Sci, 35(2000, p.2701.8CaO:3Al2O3 with in the system 3CaO Al2O3-Na2O, Mag.[5] C. Fortes-Revilla, S. Martinez Ramirez, and M. TeresaConcr. Res, 18(1966), p.59.Blanco-Varela， Modelling of slaked lime-metakaolin[23] C.M. Schlaudt and D.M. Roy, Crytalline solutions inmortar engineering characteristics in terms of process vari-3CaO-Al2O3 on the join CagAl2O6-Ca4Fe2O6， Nature,ables, Cem. Concr.206(1965), p.919.[6] S. Barathan, D. Govindarajan, G. Sivakumar, and K.[24] P.W. Brown, Early hydration of tetracalcium aluminofer-Raghu, Microwave study of hydration of cement with dif-rite in gypsum and lime-gypsum solutions, J. Am. Ceram.ferent water, IndianJ. Pure Appl. Phys, 44(2006), p.334.Soc., 70(1987), No.7, p.493.[7] Manu Santhanam, M. Cohen, and J. Olek, Differentiating[25] H. Lecoanet, F. Leveque, and S. Seguva, Magnetic sus-seawater and groundwater sulfate attack in Portland ce-ceptibility in environmental applications: comparison ofment mortars, Cem. Concr. Res, 3620060, No.12, p.2132.field probes, Phys. Earth Planet. Inter, 1999), p.191.[8] A. Palomo, V. Blanco, M.T. Granizo, F. Puertas, T.26] HY. Ghorab, M.S. Hilal, and E.A. Kishar, Effect ofVazquez, and M.W. Grutzeck, Chemical stability of ce-mixing and curing waters on the behaviour of cementmentitious materials based on metakaolin, Cem. Concr:pastes and concrpte nart I: Mirnetrurthre nf cement pastes,Res, 291(999), No.7, p.997.Cem. Concr.中国煤化工[9] Mohamed A. Ahmed, Mohamed Y. Hassaan, SalahMahrous, Habib Mikhail, and Mohamed M. ElI-Desoky,YHCNMH G.