Electrical Measurement to Assess Hydration Process and the Porosity Formation

Journal of Wuhan University of Technology-Mater. Sci. Ed. Oct 2008761DOI 10.1007/s1 1595-007-5761-3Eleetrical Measurement t0 Assess HydrationProcess and the Porosity FormationWEI Xiaosheng', XIA0 Lianzhen', LI Zongjin2(1 Department of Civil Engineering, Huazhong University of Science and Technology, Wuhan 430074, China;2 Department of Civil Engieering. The Hong Kong University of Science & Technology, Kowloon, Hong Kong)Abstraet: The change of electrical resistivity with time at early ages was used to investigatethe hydration process and the porosity development. Porosity reduction process of cement-based ma-terials hydration was developed by a proposed method. The porosity reduction is fast at the sttingperiod. The results find that the pore discontinuity occurs faster at lower water/cement ratios than athigher waterlcement ratios which is similar to the results of the Percolation method.Keywords: electrical resistivity; hydration; seting time; percolation1 Introductionhydration of cement in certain mixture and curing condi-tions. Reversely, the changing microstructure determinesA variety of techniques have been applied in the the evolution of the material engineering properties.study of cementitious systems. The purpose is to deter- When the relationship between the microstructure andmine the hydrate change such as SEIM2, X-ray diffrac- electrical resistivity is obtained, it will be possible totion, DTA, and small angle X-ray scattering or to deter- obtain the cement-based materials strength and durabilitymine the porosity by mercury intrusion porosimetry by monitoring the electrical resistivity of fresh cement(MIP). The microstructure of hydrated cement paste can and concrete. When the mineral admixtures are incor-also studied by back-scattered electron imaging in a SEM. porated, the pozzolantic activity can be represented. TheSEM and XRD techniques have led considerable pro- electrical properties of cement pastes have been provedgress, but SEM cannot quantitatively identify differences to be rather sensitive to microstructure changes of ce-between nominally similar systems. TGA/DTA can only mentitious materials'10-13)]Electrical resistivity meas-very roughly quantify the amount of hydrate product CH urement has been applied to determine the time of settingand the amount of chemically combined water in the and hardening, and the influence of different admix-paste. A considerable quantity of the hydrates is poorly turest 16. The advanced work includes the prediction ofcrstalline, so cannot be measured accurately by XRD. hydration and conductivity by computer modeling'",There are still many problems to be solved, even if there which is shown in Tumidajski et alls. Li et all foundare some fundamental and general theories about the that the relationship between porosity and electrical re-hydration process that have been accepted by cement sistivity in cementitious systems follows Archie's law"9].scientists and engineers. One of the theories is the hy- The different methods for measuring pore structures aredration process by calorimetric curve.compared in Fig.1. Electrical resistivity measurementElectrochemical methods have also been proved to (ERM) is selected to detect the hydration from thehave distinct advantages in the hydration process inves- methods. It has many advantages, which cover a longtigation. Zeta potential3o, conductivity7.8l and AC im- range.pedance'T measurements proved that the electrical resis-The eletrical resistivity ρ of concrete is affected bytivity change with time is directly related to change in the the water/cement ratio, cement content, cement type,pores and the liquid phase composition that monitor the chemil中国煤化工degree (includingcuringMHCNMHGThe total porosity(ecceived: Sept.23, 2007: Acepted July 17, 2008)and the wuluclualvll UI IUIIsS ac unfluenced by the wa-WEI Xiashneg(魏小胜): Asse. Prof; Email: weixiaosheng@, gmail.com ter/cement ratio. Hughes et al' 20) studied the electricalFunded by the National Naturl Science Foundation of China (No. resistivities of four different cement pastes using AC and50778078)762Vol.23 No.5 WEI Xiaosheng et al: Electrical Measurement to Assess Hydration...DC measurement. The electrical resistivities have beenfound to display very different depending on age. These3 Results and Discussionsdifferences cannot be readily explained in terms of theircomposition and water/cement ratios. They found that the3.1 The hydration process based on thedevelopment of electrical resistivity in cement pastes isdifferential eurveFig.3 demonstrates the change in electrical resistiv-similar to the development of the compressive strength.ity with time up to 24 hours for the Portland cementPore size distribution/mpastes with w/c=0.3 and 0.4 at curing temperature 20 C.10-1010~” 10”10~”10-* 10~5 10~+ 10~3 10210~P0.3|Special technique具10SorptionMIPSEMEyes5PjPO.4Electrical resistivityFig.1 The comparison of the dfferent methods1152025In this paper, the principle of the non-contactingTime/hoursmethod is briefly introduced. The hydration processFig.3 Development of rsistivity duringbased the electrical resistivity is described and the prop-first 24 hours after mixingerty of the porosity reduction is compared the modelingT11TITIVof percolation.p2 ExperimentalIt can be seen that the problems affecting test re-sults mainly from electrodes. To overcome the problems,豆十0.5a non-contacting electrical resistivity measurement de-vice without any electrode was invented by Z Li and WLi in test system!2n. It successfully solves the contact10problems and can accurately investigate the electricalTimc/hourresistivity p() of cement-based materials.Fig.4 The electrical resistivity curve and the rateTwo tests with the same water/cement ratio wereof change in resistivity developmentcarried out with the same equipment at curing tempera-The difference in resistivity between the two sam-ture 30 C as shown in Fig.2. The measured electricalples appears that resistivity decreases with increase ofresistivity varies only approximately 1%. This meansw/c. It reveals that the sample with lower w/c has morethat the equipment's reproducibility is very high.initial cement particles and more hydration productsformed later periods per unit volume, and the channels121for ions to go through have higher tortuosity. It is rea-sonable that the paste with lower the w/c ratio shows the一P0.4_ Ihigher the resistivity.P0.4 2Fig.4 displays derivative curves [(d p()d1)-1] which ,indicates the rate of resistivity development. Based on thefirst pe中国煤化工derivative curve, the1000 2000 3000 4000 5 000hydratEssolving period ()Timc/mins(befordYHCN M H G(before Pa), aceFig.2 Reproduibility of non-contacting resistivityeration period (II) (before Pi), deceleration and steadymeasurement systemstate development (IV) (after Pi) of hydration. Fig.5Journal of Wuhan University of Technology-Mater. Sci. Ed. Oct 2008763shows the schematic representation of cement paste0.structure with time.0.55time= 0, hydration degree=0.5.P0.40.45P0.30.35P0)0.3.1052:Time/hourtime- t, hydration degree aFig.7 Porosity change with hydration timeIt shows that there is a rapid drop during the settingtime. The lower water/cement ratio paste drops morerapidly.3.3 Summary of the hydration processlow w/c.high wle,based on the eleetrieal measurementlou porosily,high porosily.high re sistivity P(1)low resistivity(0)The main characteristic features of the microstruc-Fig.5 Schematic representation of cement pasteture development in cement paste are the reduction inporosity (pore phase) accompanied by the formation of3.2 The formation of porosityA simple two-component model is proposed, as athe hydration products (solid phase). The mainparallel model like Equation (1) and Equation (2), inthree-hydration products are calcium-silica-hydrate ge(C-S-H), calcium hydroxide (CH) and etringite (AFt).Fig.6 to evaluate the hydration process.The pores include gel pores and capillary pores. Theb(1)development of the reaction products and microstructureRc Rucis schematically shown in Fig.8. The electrical resistivitycurve is also shown in Fig.8.1_ CiVc . CzVuc_ CiVc_ V_ p"(2)ρ ρcpucPc0c poporosityThe method is proposed by Wei and Li221 to ob-CSH geltain the porosity development curve (t)-t) from (p(t)-1)curve. The result is shown in Fig.7.Ca(OI)2pastc becomes denserC(AF)H,electrical resistivityhydration timeMonosulphatc、上tringiteinsulator(hydrates and unhydrated cement)Structure of typical cemcnt hydration with timeFig.8 Pore-formation, hydrate formationand resistivity change with timedecrcascsFor the curves in Fig.3, five periods can be identi-fied based on the critical points Pm, Pa, Pi , which arelabeled for sample P0.3. These four periods comprise adissolution period (I), a induction perid(I), an accel-cration perid(I), a deceleration and diffusion-controlledconductor( pore solution with ions)period中国煤化工. hydration sequenceElectrical conduction modelproposies are summarizedFig.6 Two-component model of cement-based materialand lisYHCNMHGand time-dependent764Vol.23 No.5 WEI Xiaosheng et al: Electrical Measurement to Assess Hydraton..Table 1 Charateristics of five periods of hydration processPeriod nameRangeCharacteristicsA(1) and中(t)dpldt≤0, dissolution of cnmented-based φ(t)changes slightlyDissolution period (I)r≤tmmaterials, chemical controlA(t) changes significantlydp/dr≈0, coating formation and retarding 0(1)changes slighlyInduction period(I)m1mation of hydrates, chemical and diffusion Po(1) changes slightlysion-controlled period (IV)control3.4 Coneept of pereolation theoryFig.9 shows random site percolation on a 10X 10Percolation theory was developed to mathemati- lttice for site percolation of 0.3 and 0.6, and is gener-cally deal with disordered media, in which the disorder ated in EXCEL using the random function, and demon-is defined by a random variation in the degree of con- strates site percolation on a two-dimensional square 10nectivityl's. The main concept of percolation theory is X 10 lttice. Occupied sites are shown by colored back-the existence of a percolation threshold, defined in theground, and empty sites do not have colored background.following way. Suppose p is a parameter that defines theA site which belongs to a non-spanning (finite) cluster isaverage degree of connectivity between variouscolored yellow or red, and sites belonging to a spanningsub-units of some arbitrary system. When p=0, allcluster are colored green.sub-units are totally isolated from every other sub-unit.p=0.3.clusters- 14.no-spanningclusters- 14.no spanningclustersWhen p=1, all sub-units are connected to some maxi-mum number of neighboring sub-units. At this point, thesystem is connected from one side to the other, sincethere are paths that go completely across the system,linking one sub-unit to the next, along the spanningcluster. Now suppose, starting at p=1, connections arerandomly broken, so that P, the measure of averageinfiny-49hoccupied-43/63connectivity, decreases. The percolation threshold is thatvalue of p, usually denoted Pe at which there is nolonger an unbroken path from one side of the system tothe other. Alternatively, starting out at p=0, and ran-domly creating connections so that p increases, then Pe isdefined as the point at which a spanning cluster firstappears. For p less than Pc, only isolated, non-spanningFig.9 Random site percolation on 10X 10 lttice for siteclusters can exist. For p greater than Pe, there is always apercolation =0.3 and 0.6spanning cluster, although some isolated, non-spanningAfter generating the site matrix on the computer,clusters can still be present.A typical lttice example of a percolation problemanother task is to identify the clusters that belong to eachis that of site percolation on a simple two-dimensionalother and label them with distinct numbers. Hoshen andsquare lttice. The lattice starts off empty, with all sitesKopelman introduced an alorithem that revolutionizedunoccupied. The sites of the lttie are then randomlyspatial cluster analysis in percolation theory that is nowoccupied, one at a time. If two occupied sites are nearcalled中国煤化工gorithm.neighbors, a connection is made between them. When a3.cement pasieCNMHGn theory tocritical fraction p-=0.593 of the sites are present!24, aDispersion and displacement processes in porousspanning cluster will come into existence, and the oCcu-media arise in many diverse fields of science and engi-pied sites will percolate.Jourmal of Wuhan University of Technology-Mater. Sci. Ed. Oct 2008765neering. In construction these processes play a key factorratios, the percolation threshold of pore fluid phase oc-in respect to the durability of materials. The classical curred when its volume fraction fell slightly below 20%.theory describes the kinetics of water absorption of po-This is consistent with the 18% value determined fromrous media (e 8,concrete, mortar or cement paste) with a the same model previously using a different cerment par-capillary or potential theory. Both theories result in the ticle size distributionl20l, implying that the capillary po-well known square-oot equation,e g, water uptake is rosity percolation threshold is fairly insensitive to ce-proportional to the square root of time. Laboratory ex- ment particle size distribution.periments in different laboratories world-wide and in theGarboczi and Bentzl27) simulated the hydrationUST laboratory showed that this square root equation process and showed that the major reduction in the per-does not fit the experimental data. The major reason iscolation through capillary porosity occurs at a capillarythat these theories do not consider the complex pore volume below 40% and total discontinuity was achievedstructure of the medium. Research work has shown that at 16% (Fig. 10). They also simulated discontinuity of thethe pore structure of the cement paste can be described pore structure at various levels of hydration for dffentwith a discrete model as a random network of bonds and w/c ratios versus degree of hydration as in Fig11.sites. The percolation theory describes the connectivity,conductivity and diffusivity effcts in these models.og4The percolation properties of the cement paste mi-。0.8-CrO structural model can be generated by computer,0 w/c=0.35since the digital-image-based model has an underlying0.6口w/c=0.45lattice structure. This means that any algorithm designed。w/c=0.50for simple percolation problems can be applied to this0 w/c=0.60more complex digital-image lattice. In particular, the●w/c-0.70。percolation of the pore space, or any other phase of in-terest, may be determined by the use of a“buming”algorithm [24I. This algorithm is a simple way of identi-0.0.40.81.0fying all the pixels that are part of a spanning cluster, ifCapillary porositysuch a cluster exists, and works as follows. Conceptually,Fig.10 Computer simulation of discontinuity of the pore structureall the pixels belonging to the phase of interest are clas-at various levels of hydration for different w/c ratiossified to be“combustible" .A“fire" is started on oneversus capillary porosityside of the model's unit cell, and is allowed to propagateonly along these combustible pixels. If any pixels on theopposite side of the mode! cell are found to have been“burmed" , then a spanning cluster of the phase of in-terest must exist. The number of“burned” pixels are宫0.6-0 w/e-0.35counted to determine the fraction of the phase of interest口w/c-0.45that is a part of the spanning cluster. This connectivitycan be assessed at any stage of hydration, and that has0 w/c-0.60， 0.2● @/c-0.70been modelled by E J Garboczi of NIST21.cA measure of the connectivity of the pore phase日✧0.2(and other phases) is readily achieved in the simulatedDegree of hyrttonmicrostructures by using a“burning” algorithm. Thebuming algorithm propagatesa“ fire”throughout the Fig.11 Compuer simulaion of discontinuity of the pore structuremicro structural phase being asssed. If the fire reachesat various levels of hydration for different w/c ratios versusthe opposite side from which it was started, the phase isdegree of hydrationpercolated in that particular direction. This technique中国煤化工he idea that the porewas employed on all the microstructures, which is used discon:MYHCNMHGrw/cratiosthanatin this study to detemine if the pore fluid phase was higher w/c ratios, and is similar to the calculated resutsconnected in the direction of the applied voltage. For based on the resistivity development in Fig.7.systems of both sizes and several different water-cement766Vol.23 No.5 WEI Xiaosheng et al: Electrical Measurement to Assess Hydration.. .4 Conclusions[12] McCarter W J, Tran D. Monitoring Pozzolanic Activity byDirect Activation with Calcium Hydroxide [J. Constructiona) The curve of resistivity vs time can be dividedand Building Materials, 1996, 10(3), 179-184into dissolving period, induction period, acceleration [13] McCarter W J, Curran P N. The Electrical Response Char-period, deceleration period and steady state developmentacteristics of Setting Cement Paste []. Magazine of Concreteof hydration process.Research, 1984, 36(126):42-49b) The porosity reduction is fast at the setting pe- [14] Paya J. Enbanced Conductivity Measurement Techniquesriod by the electrical resistivity.for Evaluation of Fly Ash Pozzolantic Activity [J]. Cementc) The porosity formation derived from the electricaland Concrete Research, 2001, 31(1):41-49resistivity is similar to the results of the percolation theory.[15] Torrents J M, Roncero J, Getu R. Utilization of ImpecdanceSpectroscopy for Studying the Retarding Effect of a Super-Referencesplasticizer on the Setting of Cement [J]. Cement ConcreteResearch, 1998, 28 (9),1 325-1 333[1] Scrivener K L. Development of the MicrostructuDuring the [16] Li Z, We X, Li W. Preliminary Interpretaion of PortlandHydration of Porland Cement (D]. London: University ofCement Hydration Process Using Resistivity MeasurementsLondon, 1984[I. ACI Material Jourmal,2003, 100(3):253-2572] Scrivener K L. The Microstructure of Concrete, Materials[17] Coverdale R T, Christensen B J, Mason T 0, et al. Inter-Science of Concrete I [M]. In: Skalny J P, ed. Amercanpretation of the Impedance Spectroscopy of Cement Paste viaCeramic Society, Westervile OH, 1989, 127-161Computer Modelling I: Bulk Conductivity and Offset Re-3] Suzuki K, Nichikawa T， Kato K, et al. Approach bysistance []. J. Mater. Sci, 1995, 30:712-719Zeta-Potential Measurement on the Surface Change of Hy-[18] Tumidajski P J. Electrical Conductivity of Portland Cermentdraing C;S ([]. Cement and Concrete Research, 1981Mortars [0]. Cement and Concrete Research, 1996, 26(4);:11(5-6):759-764529-5344] Krstulovic R, Feric T, and Krolo T. Study of Kinetics of [19] Archie G E. The Electrical Resistivity Log as an Aid inIndustrial Cements Hydration [C]. In: 7th Int. Congress onDetermining Some Reservoir Characterstics []. Trans.the Chemistry of Cement, Vol. II, Paris. 1980, 153AIME ,1942, 146:54-625] Singh N B, Singh A K, Singh P. Effect of Citric Acid on the[201 Hughs B P, Soleit A K 0, Brierley R w. New Technique forHydration of Portland Cement []. Cement and ConcreteDetermining the Electric Resistivity of Concrete [小. Maga-Research, 1986, 16);911-920zine of Concrete Research, 1985,37 (133):243-2486] Whittington H w, McCarter J, Forde M C. The Conduction of[21] Li z, Li W. Conatles,Transformer-based MeasurementElectricity through Concrete []. Magazine of Concrete Re-of the Resistivity of Materials IP]. United States Patent 2003,search, 1981, 33(1 14):48-606639401[7] Xie P, Beaudoin J J, and Brousseau R. Flat aggregate Port-[22] Wei X, Li z. Study on Hydration of Portland Cement withland Cement Paste Interfaces: I Electrical ConductivityFly Ash Using Electrical Measurement小Materials andModels [U]. Cement Concrete Research, 1991， 21(4);Strucures, 2005, 38(277):411-417515-522[23] Zallen R. The Physics of Amorphous Solids [M]. New York:[8] Xie P, Gu P, Xu z, et al.J J. A Rationalized A.C. ImpedenceJ. Wiley and Sons,1983Model for Microstructural Characterization of Hydrating[241 Stauffer D. Introduction t0 Percolation Theory [M]. London:Cement Systems []. Cement Concerete Research, 1993, 23Taylor and Francis, 1985():359-367[25] Garboczi E J, Bentz D P. Fundamental Computer Simula-19] Gu P, Xie P, Beaudoin J, et al. Impedance Spectroscopy (ID):ion Models for Cement Based Materials [M]. In: Skalny」,Microstructural Characterization of Hydrating Cement-silicaMindess s,Eds. Materials Science of Concrete. Cincinati,Fume Systems [小. Cement and Concrete Research, 1993,OH: American Ceramics Society, 1991:249-27723(1);157-168[26] Powers T C, Brownyard T L. Ssudies on the Physical Prop-[10] McCarter w」, Forde M C, H w Whtington. Resistivityeries of Hardened Porland Cement Paste M]. PortlandCharacteristics of Concrete|C]. In: Proc. Instn Civ. Engrs,中国煤化工ago, lioi 1948Part 2, 1981:107-117(2]fYHFimulation of the Ag-[11] McCarter w J, Chrisp T M, Starrs G. The Early Hydration ofCNMH G& in Conce 11Alkali activated Slag: Developments in Monitoring Tech-Mater. Res, 1991, 6:19niques []. Cement and Concrete Composites, 1991, 21 (4):277-283