Construction and Building Materials 53 (2014) 392–402

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier .com/locate /conbui ldmat

Influences of superplasticizer, polymer latexes and asphalt emulsionson the pore structure and impermeability of hardened cementitiousmaterials

0950-0618/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.conbuildmat.2013.11.104

⇑ Corresponding author. Tel.: +86 10 62783703; fax: +86 10 62785836.E-mail address: kxm@tsinghua.edu.cn (X. Kong).

Yanrong Zhang, Xiangming Kong ⇑Department of Civil Engineering, Tsinghua University, Beijing 100084, China

h i g h l i g h t s

� The chemical admixtures affect the pore structure due to their plasticizing effects.� Superplasticizer enhances the impermeability of mortar due to its plasticizing effect.� Polymers raise the impermeability due to combined effects of filling and plasticizing.� Filling effect is dominating the enhancement in impermeability at high P/C value.� Mortars with latex L1 and anionic asphalt emulsion are superior in impermeability.

a r t i c l e i n f o

Article history:Received 6 March 2013Received in revised form 15 July 2013Accepted 27 November 2013Available online 25 December 2013

Keywords:SuperplasticizerPolyacrylate latexAsphalt emulsionPore structureAC impedance

a b s t r a c t

Three types of polymers commonly used in concrete and mortar (polycarboxylate superplasticizer, poly-acrylate latexes and asphalt emulsions) which differ in molecular/particle size from nanometer to micronwere employed to investigate their effects on the pore structure of hardened cement pastes and theimpermeability of hardened mortars. The pore structure and the impermeability of the hardened onescured for 7 days and 28 days were measured by mercury intrusion porosimetry and alternating currentimpedance, respectively. Results show the incorporation of superplasticizer obviously reduces the aver-age pore size and enhances the impermeability. The polyacrylate latexes also lead to the decline in poresize and consequently the enhanced impermeability at dosage higher than 3%. At the same dosage, latexwith smaller polymer particle size is more effective in reducing the average pore size and enhancing theimpermeability than that with larger particle size due to its better plasticizing effect. Similarly, asphaltemulsions also facilitate the enhancement in impermeability, and the anionic asphalt emulsion with bet-ter plasticizing effects brings about stronger impermeability than the cationic one. It is believed that forsuperplasticizer, the plasticizing effect is the main controlling factor for the finer pore structure and theenhanced impermeability. In the case of polymer latexes and asphalt emulsions, the plasticizing effectcontributes actively at low dosage and the filling effect is dominant at high dosage in terms of decliningpore size and augmenting impermeability.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Concrete, being the most widely used construction material, hasdrawn increasing research interests and attentions on its durability[1–3]. Recently, many investigations have been conducted to im-prove the durability of concrete [1–7]. It has been well acceptedthat most issues regarding concrete durability are related to thepermeability of concrete, such as freeze–thaw deterioration,chloride ingress, sulfate attack, and carbonation [8]. In general,

increasing concrete impermeability is the key to increase thedurability [9]. The impermeability of concrete is primarily deter-mined by the pore structure of cement pastes, and is also affectedby cracks and interfaces between the cement pastes and aggre-gates [10,11].

Various polymers have been incorporated in modern concretein order to achieve desired properties. For example, adding superp-lasticizers into fresh cementitious materials can improve their rhe-ological properties and thus in the premise of satisfying theconstruction requirements, lower water to cement ratio (W/C)could be achieved. As well known, the lower W/C is required toproduce concrete with higher strength, lower permeability, andhigher durability [12,13]. Khatib and Mangat [14] reported that

Table 1Composition of Portland cement (wt%).

SiO2 Fe2O3 Al2O3 SO3 MgO CaO Na2Oeq

21.56 2.78 4.44 3.14 2.32 62.83 0.6f-CaO Cl� IL C3S C2S C4AF C3A0.79 0.007 2.04 46.00 27.14 8.45 7.05

Table 2Properties of PC superplasticizer.

Solidcontent(%)

Hydrodynamicradius (nm)

Component Mn Mw/Mn

Density(g/cm3)

40 8 Poly (AA-co-MEPEGMA-coAMPS)

3.662 � 104 2.482 1.04

Note: Hydrodynamic radius of polymer chain in aqueous was measured by dynamiclaser scattering. Molecular weight of polymer was measured by gel permeationchromatograph.

Y. Zhang, X. Kong / Construction and Building Materials 53 (2014) 392–402 393

superplasticizers were beneficial to the refinement of pore struc-ture at a constant W/C. Sakai et al. [15] discussed the influenceof various types of superplasticizers on the pore structure andfound that the size of the cluster of aggregated cement particlesbecame smaller when superplasticizer with a higher dispersingability was added. Polymer latexes are often used as cementmortars and concrete modifiers to improve mortars and concreteproperties such as adhesion, fracture toughness, flexural strengths,crack resistance and waterproof [16,17]. Ohama and Demura [18]revealed oxygen diffusion resistance of the polymer-modifiedmortars is larger than that of unmodified mortars, and is markedlyincreased with an increase in polymer to cement ratio. Moreover,Gao et al. [19] found that the pore volume and the pore size oflatex-modified cement pastes tended to become smaller with anincrease in latex to cement ratio because the capillary pores ofthe hardened cement pastes were filled in with the polymer parti-cles or the polymer membranes formed by agglomeration of thepolymer particles. Meanwhile, the polymer latex film could effec-tively compact the interfacial zones between fine aggregates andcement pastes [20–22], enhancing the impermeability of hardenedmortars [23]. Cement asphalt mortar (CAM) is one kind ofinorganic–organic composite which is composed of Portlandcement, asphalt emulsions, water, fine aggregates, and otheradmixtures. With high elasticity and desired toughness, CAMserves as a vibration-absorbing layer in the slab track system ofrailroad structures. The properties of CAM have significant effectson slab track performances [24,25]. As a necessary component, as-phalt emulsions greatly influence the properties of CAM, such asmechanical properties, temperature sensitivity, and microstruc-ture. [26–28]. Especially, the covering of hydrophobic asphaltemulsions on cement grains could increase the impermeability ofCAM to some extent [29].

These polymers, which are widely used in cementitious materi-als, have different particle sizes. Polycarboxylate (PC) superplasti-cizer usually has a hydraulic radius of 10–100 nm in aqueoussolution. The particle size of polymer latexes ranges from 100 to1000 nm, while the particle size of asphalt emulsions is usuallyin the range of 1–10 lm. Although much research on the cementmortar with superplasticizer, polymer latex and asphalt emulsionhave been conducted, few studies dwell on their impacts on thepore structure and the impermeability from the viewpoints ofmicrostructure in the fresh state of cement paste, especially theirdifferent impacts originating from the particle size. Specifically,the formation of the pores may be affected by addition of thesepolymers due to their impacts on the rheological properties offresh pastes, cement hydration, and the shrinkage of hardenedpastes. Furthermore, the type of polymers with varied particlesizes also plays an important role in changing the pore structureand the impermeability. It is supposed that these polymers may af-fect the pore structure and the impermeability from three perspec-tives: (1) changing the flocculation microstructure of cementgrains, which is demonstrated by the variations of the fluidity offresh pastes in macro scales; (2) altering cement hydration pro-cess; (3) filling the pores and the cracks in the transition zonesand forming films in many cases.

In this study, PC superplasticizer, polyacrylate latexes andasphalt emulsions were incorporated into cement pastes andmortars to investigate their impacts on the pore structure andthe impermeability. The pore structure of the cement pastesand the impermeability of the mortars cured for 7 and 28 dayswere tested by mercury intrusion porosimetry (MIP) and alter-nating current (AC) impedance, respectively. By analyzing thechanges in the pore structures and the impermeability with var-ied polymer dosages and types, the working mechanisms ofsuperplasticizers, polyacrylate latexes and asphalt emulsionswere discussed.

2. Experimental

2.1. Materials

Portland cement P�I 42.5 which complies with the Chinese standard GB8076-2008 from Lafarge Shui On Cement Co., Ltd. was used, whose compositions arelisted in Table 1. The fineness of the cement is 0.5 and the density is 3.10 g cm�3.A self-synthesized PC superplasticizer was employed, whose properties are shownin Table 2. Two different styrene–acrylate copolymer latexes (latex L1 and latex L2with different particle sizes and Tg) produced by BASF (China) Co., Ltd. were used,the properties of which are listed in Table 3. Two types of anionic and cationic as-phalt emulsions (whose properties can be seen in Table 4) were provided by ChinaPetrochemical Corporation. Both standard sand (complying with GB178) and 40–70mesh quartz sand were used for the preparation of mortar specimens. RHODOLINEDF 642, antifoaming agent, was provided by Rhodia (China) Co., Ltd.

2.2. Specimens preparation and measurements

Cement pastes and mortars were prepared according to the formulations pre-sented in Table 5. P/C is defined as the mass solid/solid ratio of polymer to cement.W/C is the water to cement ratio. In the formulations of cement mortars, the massratio of sand to the sum of cement and polymer is expressed as S/(C + P). The anti-foaming agent to polymer ratio in all specimens was set at 0.0005. All specimenswere prepared in a mixer equipped with paddles rotating helicoidally at successivespeeds. The mixing procedure of cement pastes and mortars followed the Chinesestandards GB/T8077 and GB/T17671-1999, respectively. Three different curing con-ditions were used: the standard curing condition (moist curing at 20 �C and 95% rel-ative humidity (R.H.)), the dry curing condition (1-day moist curing and 27-daycuring at 20 �C and 65% R.H.) and the mix curing condition (21-day moist curingand 7-day dry curing).

2.2.1. MIP measurementThe pore structure of the hardened cement pastes (HCPs) was determined by

MIP. After being cured for 7 days or 28 days, the HCPs were cut into small piecesand placed into an alcohol bath. After 3-day storage in an oven with a controlledtemperature of 60 ± 2 �C, they were subjected to MIP tests to determine the porestructure characteristics by using an Hg-porosimetry (Autopore, IV 9510, USA).The difference in mercury volume was determined by re-intruding mercury intopastes after the first intrusion-depressurization cycle was completed. A part ofthe intruded mercury was entrapped in the pores with certain geometric shapes,which were called inkbottle pores [30]. As well known, the addition of superplast-icizers brings significant effects on the rheological properties of fresh cement pastes(FCPs) by modifying their microstructures. Information on the pore structure ofHCPs especially at early age could provide understanding on the microstructuresof FCPs. It has been reported that the entrapped mercury pores in FCPs is relatedto the flocculation structure of cement grains in FCPs [15]. Therefore, two cyclesof intrusion-extrusion were performed and the inkbottle pores volume was calcu-lated as follows [15]: Vib = V1–V2, where V1 is the volume of mercury intruded dur-ing the first cycle, and V2 is the volume of mercury intruded during the second cycle.

2.2.2. AC impedance measurementAC impedance spectroscopy technique has been widely used to investigate the

microstructure and the permeability of hardened cementitious materials based onthe relationships between the parameters of AC impedance and the structure of

Table 3Properties of acrylate latexes.

Polyacrylate latexes Solid content (%) Particle size (nm) Tg (�C) Viscosity (mPa�s) pH MFFT (�C) Density (g/cm3)

L1 50 200 10 500–2000 7.0–9.0 12 1.03L2 57 300 -6 140–200 7.0–8.5 <1 1.04

MFFT: minimum film formation temperature.

Table 4Properties of asphalt emulsions.

Type of asphaltemulsions

Solid content(%)

Particle size(lm)

Storage stability1 day (%)

Residue on 1.18 mmsieve (%)

Penetration depth at 25 �C(0.1 mm)

Density(g/cm3)

Anionic 61.16 2.9 0.30 0.02 73 1.02Cationic 59.29 3.1 0.31 0.02 98 1.02

Note: These properties were measured according to Chinese Industry Standard JTJ 052-2000.

Table 5Mix proportions of pastes and mortars specimens.

Polymers Specimen no. W/C P/C S/(C + P) AN/P

Polycarboxylate superplasticizer PC-1-(S1)-B 0.29 0 1.33 0.0005PC-2-(S1)-B 0.29 0.1% 1.33 0.0005PC-3-(S1)-B 0.29 0.3% 1.33 0.0005PC-4-(S1)-B 0.29 0.5% 1.33 0.0005

Polyacrylate latex L1-1-(S1)-M 0.50 0 3.00 0.0005L1-2-(S1)-M 0.50 1% 3.00 0.0005L1-3-(S1)-M 0.50 3% 3.00 0.0005L1-4-(S1)-M 0.50 6% 3.00 0.0005L1-5-(S1)-M 0.50 12% 3.00 0.0005L2-1-(S1)-M 0.50 0 3.00 0.0005L2-2-(S1)-M 0.50 1% 3.00 0.0005L2-3-(S1)-M 0.50 3% 3.00 0.0005L2-4-(S1)-M 0.50 6% 3.00 0.0005L2-5-(S1)-M 0.50 12% 3.00 0.0005

Asphaltc emulsions AAE-1-(S2)-G 0.50 0 3.00 0.0005AAE-2-(S2)-G 0.50 20% 3.00 0.0005AAE-3-(S2)-G 0.50 40% 3.00 0.0005AAE-4-(S2)-G 0.50 60% 3.00 0.0005AAE-5-(S2)-G 0.72 80% 3.00 0.0005CAE-1-(S2)-G 0.50 0 3.00 0.0005CAE-2-(S2)-G 0.50 20% 3.00 0.0005CAE-3-(S2)-G 0.50 40% 3.00 0.0005CAE-4-(S2)-G 0.50 60% 3.00 0.0005CAE-5-(S2)-G 0.72 80% 3.00 0.0005

Note: PC-series, L1-series, L2-series, AAE-series and CAE-series refer to the specimens with PC superplastizer, styrene–acrylate copolymer latex L1, latex L2, anionic asphaltemulsion and cationic asphalt emulsion respectively; S1 and S2 represent the standard sand and quartz sand, respectively; B-series, M-series and G-series respectively referto the specimens cured under standard curing, mix curing and dry curing conditions.

394 Y. Zhang, X. Kong / Construction and Building Materials 53 (2014) 392–402

porous cementitious systems [31–35]. AC impedance concerns the application of alow-amplitude AC excitation by surface electrodes over a range of frequencies [36].The current response (i.e., gain and phase angle) could be measured by the imped-ance analyzer. Similar to the chloride permeability measurement according toASTM C1202, after being cured, all mortars were vacuum-saturated for 2 h priorto being soaked in a saturated Ca(OH)2 aqueous solution for 18 h. Then, the mortarswere ready for the AC impedance measurements which were performed using animpedance gain/phase analyzer (Agilent 4294A, Palo Alto, CA, USA) in high-fre-quency range from 40 kHz to 100 MHz and a potentiostat/galvanostat (PARSTAT2263, Amtek, USA) in low-frequency range from 0.1 Hz to 100 kHz.

2.2.3. Observation of the microstructureEnvironmental scanning electron microscopy (ESEM) and SEM (FEI, Quanta 200,

USA) were used to observe the microscopic morphology of FCPs and HCPs respec-tively. PC-1-B and PC-3-B were observed in a low vacuum mode by ESEM right afterwell-mixing. Some internal parts of L1-5-M and L2-5-M cured for 28 days were ta-ken and dried at 50 ± 2 �C for 24 h. The parts were processed by carbon-plating be-fore observed by SEM.

2.2.4. XRD measurementCa(OH)2 is one of the main phases in cement hydrates, whose amount is in lin-

ear proportion to the hydration degree of cement. X-ray diffraction (XRD) is aneffective method to analyze the amount of Ca(OH)2 in cement hydrates. XRD was

performed with a graphite-mono-chromatized Cu Ka radiation generated at40 kV and 200 mA in an X-ray diffracto-meter (Rigaku, D/max 2550, Japan). Step-scanning was carried out at the 2h range of 17–19�, in which the diffraction peakof Ca(OH)2 was located (approximately at 18�). The step length was 0.02� and thesettle time of every step was 3 s.

3. Results and discussion

3.1. Superplasticizer

3.1.1. Effects on the pore structureIt is well known that, cement grains in fresh pastes tend to form

flocculates due to the van der Waals force or electrostatic interac-tions among the cement grains, rather than existing in the form ofsingle cement grain. The adsorption of superplasticizer moleculeson the surface of cement grains disassembles the flocculates byinducing electrostatic and/or steric repulsion interaction amongthe cement grains [37], which enables better dispersion of the ce-ment grains, as schematically illustrated in Fig. 1(a and b). Fig. 2shows the ESEM images of the FCPs with and without superplast-

Y. Zhang, X. Kong / Construction and Building Materials 53 (2014) 392–402 395

icizer. For the FCP without superplasticizer, as shown in Fig. 2(a),many flocculates are formed in the fresh paste as marked by awhite circle. For the FCP with superplasticizer, as shown inFig. 2(b), the cement grains are well dispersed. With the develop-ment of cement hydration, most of the superplasticizer added inthe FCP may be consumed by being embedded or integrated intohydration products and only a small portion of superplasticizermay stay in the capillary pores of the hardened pastes [38]. Ithas been reported that when superplasticizer is added in FCPs,the difference in the pore structure of HCPs is related to the disper-sion of cement grains in the fresh pastes [15]. In other words, thedifferent organization structure of cement grains in FCPs may leadto different pore structure formed in HCPs. The cement pastes withwell dispersed cement grains may generate finer pores after thepastes are hardened, as described in Fig. 1(c and d).

Pore size distribution of HCPs with varied superplasticizer dos-ages is presented in Fig. 3 and the characteristics parameters arelisted in Table 6. It is clearly seen that for HCPs cured for 7 days,the inclusion of superplasticizers brings about an increment inthe volume of pores smaller than 10 nm and a decline in the vol-

Fig. 1. Schematic illustration of the effects of superplasticizer on the microstructureof FCPs and the pore structure of HCPs. (a) Blank FCP; (b) FCP with superplasticizer;(c) blank HCP; and (d) HCP with superplasticizer; ( : cement grain; : flocculentstructure; : free water; : entrapped water; : hardened cement paste; :capillary pore; : inkbottle pore).

Fig. 2. ESEM photographs of the FCPs at 5 min after mixi

ume of pores larger than 100 nm. With the growth in curing age,more hydrates filling in the pores result in that the total pore vol-ume of the HCPs cured for 28 days is notably lower than thosecured for 7 days, as presented in Fig. 3. Besides, compared to thepore structure of HCPs at 7 days, the volume of the pore smallerthan 10 nm is less affected by the different dosages of superplast-icizer. Slight reductions in volume of the pores with size rangingfrom 10 to 100 nm and the pores larger than 100 nm are foundin Fig. 3 as superplasticizer is added. Table 6 exhibits that pore vol-ume, porosity, average pore size and inkbottle pores volume in theHCPs gradually decrease as superplasticizer dosage increases. Inaddition, the cement pastes cured for 28 days have less pore vol-ume, lower porosity, and smaller average pore size than thosecured for 7 days.

It is reported that the pores smaller than 10 nm in hardened ce-ment paste are attributed to gel pores, whose volume is directlyproportional to cement hydration degree. Thus, for the 7 daysHCPs, the increase in gel pore volume (<10 nm) with P/C is believedto result from the enhanced hydration degree by addition ofsuperplasticizer. The XRD results shown in Fig. 4 validate thatthe hydration degree of the HCPs cured for 7 days increases withsuperplasticizer dosage, which stems from the better dispersionof cement grains in FCPs as superplasticizer is added. At 28 days,the stable gel pore volume (<10 nm) indicates that the hydrationdegree of the HCPs cured for 28 days is little affected by superp-lasticizer dosage, as presented in Fig. 4. On the other hand, the de-

ng. (a) Blank FCP; and (b) FCP with superplasticizer.

Fig. 3. Pore size distribution of the cement pastes with varied superplasticizersdosages.

Fig. 4. XRD peaks of Ca(OH)2 in hydrating cement pastes.

Fig. 5. Nyquist plots of the mortars with different superplasticizer dosages. Z0: realimpedance; Z00: imaginary impedance. (a) High frequency; and (b) low frequency.

396 Y. Zhang, X. Kong / Construction and Building Materials 53 (2014) 392–402

cline in the volume of pores larger than 100 nm is supposed to re-sult from the impact of better dispersion of cement grains in FCPson the pore structure, which is consistent with the early finding inreference [15]. The impact is also found on the basis of the obviousdrop in the inkbottle pore volume. Overall, it is the better disper-sion of cement grains in FCPs that results in the increase in gelpores (<10 nm) at early ages and the decline in large pores(>100 nm), which consequently leads to the lowered average poresize. These results in Fig. 3 and Table 6 primarily verify the infer-ence in Fig. 1 that FCPs with superplasticizer may develop finerand more homogenously distributed pores after hardened due towell dispersed cement grains in the fresh pastes.

For one cement paste, the pore volume and the porosity aredetermined by W/C and hydration degree. The pore volume andthe porosity linearly decrease with the hydration degree of the gi-ven cement paste and increase with the W/C in completely hy-drated cement pastes. In this study, it is found that the additionof PC superplasticizer undoubtedly affects the pore volume andthe porosity of the HCPs although the W/C is fixed. The increasein the hydration degree with P/C indicated in Fig. 4 explains theaddition of superplasticizer leads to the decrease in the pore vol-ume and the porosity of the HCPs at 7 days. For the cement pastescured for 28 days, the hydration degree is little affected by superp-lasticizer dosage (Fig. 4). There must exist another reason why theaddition of superplasticizer influences the pore volume and theporosity. It has been known that the shrinkage of a hardening ce-ment paste, including autogenuous shrinkage and drying shrink-age, is highly related to its pore structure [39]. The decreasedpore size could enhance the shrinkage during curing [40]. A possi-ble reason for the lowered pore volume and porosity by addingsuperplasticizer in cement pastes cured for 28 days correlates withthe shrinkage of the pastes during curing. The smaller pore sizeoriginating from the inclusion of superplasticizer leads to moreshrinkage of the pastes during hardening and consequently a slight

Table 6Pore characteristics of the HCPs with different superplasticizer dosages.

Specimen no. Pore volume (mL/g) Porosity (%)

7 d 28 d 7 d 28 d

PC-1-B 0.1299 0.0981 25.50 18.9PC-2-B 0.1188 0.0957 22.52 18.3PC-3-B 0.1175 0.0948 22.35 18.3PC-4-B 0.1119 0.0931 21.60 18.1

decline in the pore volume and the porosity with the growth ofsuperplasticizer dosage is observed.

3.1.2. Effects on the impermeabilityAC impedance spectra are often plotted in real ± imaginary

complex plane and are parameterized by frequency. This type ofparametric plot showing the current response generated at eachfrequency is known as Nyquist plot. The Nyquist plot for hardenedmortar is of the Randles type, which is characterized by a semicir-cle in high-frequency range and a straight line in low-frequencyrange. The diameter of the semicircle is closely related to themicrostructure of porous materials and very sensitive to thechanges in the microstructure, which increases significantly withthe hydration time and compactness [41]. The diameter of thesemicircle is also found to be inversely proportional to the porosityand the average pore size of the porous materials [42]. The leftinterception of the semicircle on the real axis in high-frequencyrange is determined by the resistance of the electrolytic solutions

Average pore size (nm) Vib (mL/g)

7 d 28 d 7 d 28 d

1 25.6 16.5 0.0769 0.04884 21.1 16.3 0.0527 0.04547 18.1 15.9 0.0474 0.04156 16.7 15.4 0.0419 0.0377

Fig. 6. Different equivalent circuits of mortars.

Fig. 7. Fluidity of the FCPs with addition of polymer latexes.

Y. Zhang, X. Kong / Construction and Building Materials 53 (2014) 392–402 397

in the pores. The resistance is inversely proportional to the porosity[43,44]. The coefficient of diffusion impedance can be calculatedfrom the slope in the resistance vs. w�1/2 plots in the low-frequency curve, which reflects the development of capillary poresstructure and is in inverse proportion to diffusion coefficient [45].

The Nyquist impedance spectra of the hardened mortars withdifferent P/Cs are shown in Fig. 5. The impedance response of thecement mortars is described by the equivalent circuit illustratedin Fig. 6(a) [46], where R0 is the resistance of the electrolytic solu-tion, R1 is the resistance of the charge transfer, and C1 is the capac-itance of the electrode/mortar specimen interface. All of theseelectric properties (as shown in Table 7) could be simulated usingthe software Zsimpwin. 3.1.

The Nyquist plots of AC impedance spectra for hardened mor-tars in high-frequency range exhibit typical semicircle shape andthe diameters of the semicircles grow as P/C and curing age in-crease (Fig. 5(a)). Correspondingly, R1 increases with the additionof superplasticizer and longer curing age, which indicates declinesin porosity and average pore size. This result is consistent with theresult measured by MIP. Based on the results of AC impedancespectra in high-frequency region, it is concluded that the additionof superplasticizer and longer curing age lead to a denser micro-structure of the hardened mortars.

In low-frequency region, the Nyquist plots of AC impedancespectra for hardened mortars are straight lines (Fig. 5(b)) and thecoefficient of diffusion impedance r are calculated as listed in Ta-ble 7. r increases with the rise in superplasticizer dosage, indicat-ing that the impermeability of the mortars is enhanced in thepresence of superplasticizer. In addition, r of the mortars curedfor 28 days is higher than that cured for 7 days. It has been well ac-cepted that the impermeability of cement mortars is primarilydetermined by the pore structure of HCPs as well as the interfacialtransition zones (ITZs) between aggregates and cement pastes. Asdescribed above, the addition of superplasticizer results in reduced

Table 7Calculated electric properties of the impedance spectroscopy for the mortars with differen

Specimen no. R0 (X) C1 (F)

7 d 28 d 7 d 28 d

PC-1-S1-B 137.0 126.0 2.290E�11 1.55PC-2-S1-B 118.8 111.8 1.765E�11 1.41PC-3-S1-B 107.8 101.1 1.634E�11 1.49PC-4-S1-B 123.4 123.1 1.677E�11 1.27

pore volume, porosity, average pore size and denser structure inHCPs. All of these are the factors leading to lower permeability ofthe hardened cement mortars.

3.2. Polyacrylate latexes

It has been often reported that some polymer latexes also haveplasticizing effect in fresh cementitious mixtures due to theadsorption of polymer particles on the surface of cement grainsand the consequently better dispersion of cement grains [17,47].As seen from Fig. 7, the incorporation of both polymer latexes L1and L2 has significant influences on the fluidity of the FCPs. For la-tex L1, the fluidity is markedly increased at high dosage of polymerlatexes (P/C > 3%). The addition of latex L2 obviously declines thefluidity at P/C lower than 3% and then the fluidity begins to in-crease with further addition of polymer latex. Similar to superp-lasticizer, the influences of polymer latexes on the pore structureof HCPs that originates from the changes in the coagulated struc-ture of cement grains should be expected.

On the other hand, different from PC superplasticizer, the poly-mers in latexes exist in a condensed state with particle size of 50–1000 nm, which are impossible to be integrated into cement hy-drates during cement hydration. During cement pastes hardening,the polymer particles or the polymer films formed by the agglom-eration of polymer particles fill in the capillary pores (as schemat-ically described in Fig. 8). In this way, the impermeability of HCPsmay be more remarkably enhanced by adding polymer latexesthan adding superplasticizer due to the combination of plasticizingeffect and filling effect [22,48].

3.2.1. Effects on the pore structureThe pore size distribution of the cement pastes with different

latex contents is shown in Fig. 9. Four peaks corresponding to fourfamilies of pores, marked as peak I, II, III and IV, can be recognizedin the MIP curves of the HCPs cured for 7 days and 28 days.

t superplasticizer dosages.

R1 (X) r (X)

7 d 28 d 7 d 28 d

7E�11 678.0 940.2 1454.51 1958.849E�11 781.2 1060.0 1467.93 1982.929E�11 898.8 1115.0 1535.93 1995.414E�11 925.0 1281.0 1647.60 2052.33

Fig. 9. Pore size distribution of the cement pastes with varied latex contents. (a)HCPs with latex L1; and (b) HCPs with latex L2.

398 Y. Zhang, X. Kong / Construction and Building Materials 53 (2014) 392–402

Traditionally, the pores in cement pastes vary in size ranging froma few nanometers to hundreds of microns. From their sizes and ori-gins, the pores in HCPs can be subdivided into the following threeclasses: hydration products pores, capillary pores and air voids[49]. Hydration products pores consist of inner C–S–H gel pores(typical size of 2–4 nm) and outer C–S–H gel pores (also includingthe pores among various hydrates, typical size of 20–30 nm). TheC–S–H gel pores are assumed to be the intrinsic properties of hy-drates and should not evolve with W/C in a fully hydrated cementpaste. Capillary pores, usually larger than 50 nm are generallyformed by the volume occupied by water that is not consumedin the hydration process. The capillary pores are believed to exertmajor effects on transport processes of attacking species in HCPs[8]. Air voids are formed by air entrainment during the preparationof cement pastes and usually have larger pore size than 1 lm. Rad-mi Vocka et al. [49] pointed out that the pore size distribution indi-cated by MIP results depended not only on the pore sizes, but alsoon the connectivity of pore space; moreover, lower connectivityled to the peaks of pore size distribution shifting to smaller size.

Based on the theory above, it can be concluded that in Fig. 9,peak I and peak II represent the C–S–H gel pores; peak III reflectsthe capillary pores and peak IV indicates the air voids. For the HCPscured for 7 days, the C–S–H gel pores (peak I and peak II) are verysmall due to the low hydration degree. The incorporation of poly-mer latex in HCPs leads to lowered peak I and left shifting of peakIII at P/C larger than 3%. The lowered peak I must result from thereduced hydration degree by the addition of polymer latexes,which is known as the retardation effect of polymer latexes[17,50]. At P/C larger than 3%, the left shifting of peak III impliesthe decrease in connectivity of capillary pores and the reduced por-tion of large pores due to the filling effect of polymer particles orfilms. For 28 days HCPs, the C–S–H gel pores (peak I and peak II)grow significantly with the development of cement hydration. Itis noted that high polymer content (P/C > 3%) results in lower peakI and more pronounced peak II compared with the blank HCP, sug-gesting that more outer C–S–H gel pores are produced by the incor-poration of polymer. Similar to 7 days HCPs, left shifting of peak III(the capillary pores) is also observed in 28 days HCPs with theaddition of polymer. It is known that the capillary pores(>50 nm) is the dominating factor for the permeability of HCPs,

Fig. 8. Impact of polymer latexes with different particle sizes on the microstructure of ceL2; (d) blank HCP; (e) HCP with latex L1; and (f) HCP with latex L2. ( : cement grain;paste; : agglomerated polymer particles).

while the pores with size less than 50 nm play minor role on thepermeability of HCPs. Therefore, it is expected that the left shiftingof peak III is an indicator for lower permeability of HCPs when

ment pastes at the same P/C. (a) blank FCP; (b) FCP with latex L1; (c) FCP with latex: free water; : entrapped water; : latex L1; : latex L2; : hardened cement

Fig. 10. Nyquist plots of the hardened mortars with different latex contents. (a) L1-high frequency; (b) L1-low frequency; (c) L2-high frequency; and (d) L2-low frequency.

Table 8Calculated electric properties of the impedance spectroscopy for the mortars with different latex contents.

Specimens no. R0 (X) C1 (F) R1 (X) C2 (F) R2 (X) r (X)

7 d 28 d 7 d 28 d 7 d 28 d 7 d 28 d 7 d 28 d 7 d 28 d

L1-1-S1-M 34.63 162.6 2.433E�11 1.743E�11 403.6 703.4 – – – – 1005.41 1027.73L1-2-S1-M 45.68 160.3 2.571E�11 1.725E�11 383.3 693.8 – 3.500E�10 – 276.0 991.57 1188.01L1-3-S1-M 46.72 162.6 2.547E�11 1.565E�11 390.3 806.5 – 3.070E�10 – 356.2 1042.90 1262.55L1-4-S1-M 64.71 165.3 2.402E�11 1.279E�11 453.2 946.8 – 1.429E�10 – 553.4 1071.86 1421.02L1-5-S1-M 67.19 187.7 2.239E�11 9.772E�12 541.0 1466.0 – 7.168E�11 – 1251.0 1151.72 1613.13L2-1-S1-M 34.63 162.6 2.433E�11 1.743E�11 403.6 703.4 – – – – 1005.41 1027.73L2-2-S1-M 37.86 162.9 2.949E�11 1.755E�11 335.7 662.0 – 4.120E�10 – 265.6 953.09 1022.85L2-3-S1-M 37.80 166.3 2.649E�11 1.658E�11 380.0 732.4 – 3.233E�10 – 331.9 995.79 1208.32L2-4-S1-M 55.25 166.9 2.466E�11 1.493E�11 436.2 818.6 – 2.305E�10 – 436.3 1043.14 1382.83L2-5-S1-M 76.54 184.8 1.728E�11 8.627E�12 653.4 2501.0 – 4.123E�11 – 2123.0 1155.75 1626.69

Y. Zhang, X. Kong / Construction and Building Materials 53 (2014) 392–402 399

polymer latexes are incorporated. The impacts of polymer on thepore structure of HCPs are similar for both polymer latex L1 andL2 (Fig. 9(a and b)).

3.2.2. Effects on the impermeabilityThe Nyquist impedance spectra of the mortars with different

latexes contents are shown in Fig. 10. The impedance response ofthe latex modified mortars cured for 7 days and 28 days can be de-scribed by the equivalent circuits illustrated in Fig. 6(a and b) respec-tively [51]. In Fig. 6(b), R2 and C2 are the resistance and thecapacitance of polymer, respectively. It is seen in Fig. 10, at verylow P/C (<3%), the diameter of the semicircle is slightly affected bypolymer. With more incorporation of polymer latexes in the pastes,the diameter of the semicircle becomes larger. The left interceptionsof the semicircles on real axis in high-frequency curves almost re-main constant regardless of the addition of polymer latexes. These

observations indicate that the incorporation of polymer remarkablyincreases the compactness of the hardened mortars when P/C ishigh, and has little effect on the whole porosity. It is also obvious thatthe diameter of the semicircle grows with the curing age because ofthe development of cement hydration. Correspondingly, Table 8 pre-sents that R0 makes a trivial fluctuation, R1 increases constantly aftera slight drop, and R2 significantly increases with the growth in poly-mer content. The increase in R1 suggests average pore size declineswith the addition of polymer latexes, which is in very good agree-ment with the left shifting of peak III in MIP curves (Fig. 9). The slightdrop in R1 implies average pore size increases when P/C is low, whichmay result from the poorer fluidity of FCPs or the poorer dispersionof cement grains in FCPs as seen in Fig. 7. The mortars cured for28 days possesses higher R0 and R1, which indicates that the averagepores size decreases with longer curing ages due to the enhance-ments in hydration degree.

Fig. 11. SEM photographs of the HCPs cured for 28 days. (a) HCP with 12% latex L1; and (b) HCP with 12% latex L2.

Fig. 12. Nyquist plots of the CAMs with different asphalt contents. (a) CAMs with anionic asphalt emulsion; and (b) CAMs with cationic asphalt emulsion.

400 Y. Zhang, X. Kong / Construction and Building Materials 53 (2014) 392–402

Table 9Coefficient of diffusion impedance r (X) of the CAMs with different asphalt contents.

Specimensno.

AE-1-S2-G

AE-2-S2-G

AE-3-S2-G

AE-4-S2-G

AE-5-S2-G

AAE 7 d 956.57 1003.97 1011.50 1091.11 1072.7928 d 963.96 1713.98 1896.27 2665.19 1903.17

CAE 7 d 956.57 957.63 971.25 1021.98 1009.1628 d 963.96 1122.56 1813.60 1948.94 1846.16

Y. Zhang, X. Kong / Construction and Building Materials 53 (2014) 392–402 401

The variation of r presents that low latex contents couldslightly degrade the impermeability of mortars while high latexcontent is beneficial to enhance the impermeability of mortars,which is consistent with the results of AC impedance in high-fre-quency range.

3.2.3. Comparison between different polyacrylate latexesL1 and L2 are two types of latexes and different in particle size

and glass transition temperature (Tg). They may lead to differentbehaviors of polymer latexes in cement pastes. In comparison withL2, L1 more significantly decreases the average pore size of the HCPs(indicated by R1 in AC impedance), and enhances the impermeabilityof the mortars when P/C is less than 12%. This phenomenon can beexplained by the fact that latex L1 leads to better dispersion of ce-ment grains in FCPs than L2 due to its smaller polymer particle sizeas suggested in Figs. 7 and 8. At the same P/C, the number of smallpolymer particles in L1 is much more than that in L2. Therefore, L1reduces the flocculates more effectively than L2. The schematic illus-tration of the adsorption of polymer particles with different sizesonto the surface of cement grains is shown in Fig. 8.

When P/C is 12%, the pore structure is significantly affected bythe filling effect of latex particles in the pores and the transitionzones, and also the film formation of polymer particles in theseplaces. The lower impermeability of HCP with L1 might be due tothe poorer film formation ability of L1 because it has higher Tg thanL2 [17]. L2 is more effective to compact the interfacial zones be-tween phases, so that the impermeability of the HCP is more en-hanced. As presented in Fig. 11, it is observed that themicrostructure of the HCP with L2 is more compact than that withL1. All in all, L1 with smaller particle size and poorer film formationability results in higher impermeability when P/C is low. By con-trast, L2 with larger particle size and better film formation abilityproduces higher impermeability when P/C is higher.

3.3. Asphalt emulsions

Asphalt emulsions are usually used to prepare CAMs in order toadjust its mechanical properties. They can also affect the fluidity offresh CAMs and the pore structure of hardened CAMs due to theiradsorption on cement grains and filling effects. A previous study[52] proved that, at the same W/C, the fluidity of cement asphaltpaste can be increased by adding a certain amount of asphalt emul-sions (asphalt to cement ratio >20%). In other words, the additionof asphalt emulsions causes a plasticizing effect on the cementpaste to some extent. Along with cement hydration, the agglomer-ation of asphalt emulsion particles and film formation adsorbed oncement occur, which effectively changes the microstructure bycompacting the interfacial zones among phases. Finally, a structureof dense interpenetrating network is formed, involving cement hy-drates, fine aggregates, and asphalt membranes [53]. In addition,with the increase in asphalt content (P/C > 60%), the hydrophobicasphalt emulsion gradually becomes the main phase in CAMs.The covering of asphalt membrane on cement grains, hydrationproducts and aggregates resists the penetration of water [29],

which ensures enhancement in the impermeability of CAMs to acertain extent.

Based upon the above discussions, it is convinced that both theincreased fluidity of fresh pastes and the filling effect of polymerlatexes and asphalt emulsion have positive effects on the declinesin the average pore size and the enhancement in impermeability.The pore structure and the impermeability are primarily deter-mined by the filling effect when their contents are high. In practice,the P/C of a typical CAM is ranging from 0.2 to 1.0, much higherthan those of superplasticizer and polymer latexes. As a result,the massive addition of asphalt emulsions results in a sharp de-crease in the permeability of CAMs. Therefore, the semicircle diam-eter of the CAMs in high-frequency curves increases with the risein P/C, as shown in Fig. 12 and Table 9. It is noted that, at P/C of80%, the reduced compactness and impermeability of the CAMsare caused by the increased W/C resulting from the large amountof water associated with the asphalt emulsions.

Zhang et al. [52] reported that anionic asphalt emulsion en-hances the fluidity of cement asphalt (CA) pastes more effectivelythan cationic asphalt emulsion because the former causes a betterdispersion of cement grains in the fresh pastes. Consequently, theCAM containing anionic asphalt emulsion should possess morecompact structure. The AC impedance results show that the CAMswith anionic asphalt emulsion has relatively denser structure andstronger impermeability than those with cationic asphalt emul-sion, as shown in Fig. 12(a and b).

4. Conclusion

Three types of polymers (superplasticizer, polyacrylate latexes,and asphalt emulsions) were used to investigate their effects onthe pore structure of the hardened pastes and the impermeabilityof the mortars cured for 7 and 28 days. Based on the aforemen-tioned results, the following conclusions can be drawn:

(1) Superplasticizer significantly alters the dispersion of cementgrains by disassembling the flocculates in FCPs, thusdecreasing the average pore size and connectivity in HCPs.Moreover, the pore volume and the porosity decline slightlywith the addition of superplasticizer, which is ascribed tothe impacts of superplasticizer on cement hydration andthe shrinkage of HCPs. The addition of polymer latexes leadsto the peak of capillary pores in HCPs shifting to smaller size,and reduces the connectivity of capillary pores due to bothplasticizing effect and filling effect. C–S–H gel pores volumepresents an obvious drop due to the retardation effect ofpolymer latexes, especially at early age.

(2) The impermeability of the hardened mortar is enhanced bythe addition of superplasticizer due to the better dispersionof cement grains in FCPs and the correspondingly decreasedaverage pore size in HCPs. For the cement mortars withpolymer latexes and asphalt emulsions, the impermeabilityis modified due to the combined effects of plasticizing andfilling. At very low P/C (<3%), adding polymer latex has littleeffect on the impermeability. Further addition of polymerlatexes leads to a significant increase in the impermeability.It is believed that the filling effect plays a dominant role inincreasing the impermeability when their dosages are high.

(3) At the same P/C, latex L1 is more effective in reducing theaverage pore size and enhancing the impermeability thanL2. This is because latex L1 possesses better plasticizingeffect than L2 at the same dosage due to its smaller particlesize. Similarly, at the same asphalt content, anionic asphaltemulsion with better plasticizing effect leads to strongerimpermeability of CAMs.

402 Y. Zhang, X. Kong / Construction and Building Materials 53 (2014) 392–402

(4) For PC superplasticizers, which usually have hydraulicradius of 10–100 nm in aqueous solution, the plasticizingeffect in FCPs is the main controlling factor for the finer porestructure of HCPs and the enhanced impermeability of mor-tars. On the other hand, for polymer latexes with particlesize in the range of 100–1000 nm, the plasticizing effect con-tributes a lot at low dosage and the filling effect is dominantat high dosage in terms of declining pore size and augment-ing impermeability. Regarding asphalt emulsions, fillingeffect plays essential role in reducing pore size and enhanc-ing impermeability due to larger particle size ranging from 1to 10 lm.

Acknowledgement

The supports from National Natural Science Foundation ofChina (Grant Nos. 51173094, U1262107) are appreciated.

References

[1] Dhir RK, Hewlett PC, Chan YN. Near surface characteristics of concrete:intrinsic permeability. Mag Concr Res 1989;41(147):87–97.

[2] Aitcin PC. Cements of yesterday and today: concrete of tomorrow. Cem ConcrRes 2000;30(9):1349–59.

[3] Czarneki L, Justnes H. Sustainable and durable concrete. Cem Wapno Beton2012;6:341–62.

[4] Levy SM, Helene P. Durability of recycled aggregates concrete: a safe way tosustainable development. Cem Concr Res 2004;34(11):1975–80.

[5] Singh M, Garg M. Durability of cementing binders based on fly ash and otherwastes. Constr Build Mater 2007;21(11):2012–6.

[6] Hassan AAA, Lachemi M, Hossain KMA. Effect of meta-kaolin and silica fume onthe durability of self-consolidating concrete. Cem Concr Compos2012;34(6):801–7.

[7] Somna R, Jaturapitakkul C, Amde AM. Effect of ground fly ash and groundbagasse ash on the durability of recycled aggregate concrete. Cem ConcrCompos 2012;34(7):848–54.

[8] Basheer L, Kropp J, Cleland DJ. Assessment of the durability of concrete from itspermeation properties: a review. Constr Build Mater 2001;15(2–3):93–103.

[9] Aitcin PC. The durability characteristics of high performance concrete: areview. Cem Concr Compos 2003;25(4–5):409–20.

[10] Zhang MH, Li H. Pore structure and chloride permeability of concretecontaining nano-particles for pavement. Constr Build Mater2011;25(2):608–16.

[11] Nilsen AU, Monteiro PJM. Concrete: a three phase material. Cem Concr Res1993;23(1):147–51.

[12] Malhotra VM. Innovative applications of superplasticizers in concrete: areview. Role Admix High Performa Concr 1999;5:421–60.

[13] Gagne R, Boisvert A, Pigeon M. Effect of superplasticizer dosage on mechanicalproperties, permeability, and freeze-thaw durability of high-strengthconcretes with and without silica fume. ACI Mater J 1996;93(2):111–20.

[14] Khatib JM, Mangat PS. Influence of superplasticizer and curing on porosity andpore structure of cement paste. Cem Concr Compos 1999;21(5–6):431–7.

[15] Sakai E, Kasuga T, Sugiyama T, Asaga K, Daimon M. Influence ofsuperplasticizers on the hydration of cement and the pore structure ofhardened cement. Cem Concr Res 2006;36(11):2049–53.

[16] Jenni A, Holzer L, Zurbriggen R, Herwegh M. Influence of polymers onmicrostructure and adhesive strength of cementitious tile adhesive mortars.Cem Concr Res 2005;35(1):35–50.

[17] Kong XM, Li QH. Properties and microstructure of polymer modified mortarbased on different acrylate latexes. J Chin Ceram Soc 2009;37(1):107–14.

[18] Ohama Y, Demura K. Pore size distribution and oxygen diffusion resistance ofpolymer-modified mortars. Cem Concr Res 1991;21(2–3):309–15.

[19] Gao JM, Qian CX, Wang B, Morino K. Experimental study on propertiesof polymer-modified cement mortars with silica fume. Cem Concr Res2002;32(1):41–5.

[20] Beeldens A, Gemert DV, Schorn H, Ohama Y, Czarnecki L. From microstructureto macrostructure: an integrated model of structure formation in polymer-modified concrete. Mater Struct 2005;38(6):601–7.

[21] Czarnecki L, Schorn H. Nanomonitoring of polymer cement concretemicrostructure. Restor Build Monu 2007;13(3):141–52.

[22] Knapen E, Gemert DV. Cement hydration and microstructure formation in thepresence of water-soluble polymers. Cem Concr Res 2007;39(1):6–13.

[23] Beeldens A, Monteny J, Vincke E, Belie ND, Gemert VD, Taerwe L, et al.Resistance to biogenic sulphuric acid corrosion of polymer-modified mortars.Cem Concr Compos 2001;23(1):47–56.

[24] Zuo JQ, Jiang QB, Fu DZ. Research on CA mortars used as cushion layer in slabtrack. Railway Eng 2005;9:96–8 [Chinese].

[25] Zhao DT, Wang TC, Liu XY, Wu L. Configuration and performance of CA mortarsfor ballastless slab track. J Tianjin Univ 2008;41(7):794–800 [Chinese].

[26] Liu YL, Kong XM, Zou Y, Yan PY. Static and dynamic mechanical behavior ofcement asphalt mortars. J Railway Eng Soc 2009;6(3):1–7 [Chinese].

[27] Wang FZ, Liu ZC. Research on the fatigue behavior of CA mortar used inballastless slab track of high speed railway. J Wuhan Univ Technol2008;30(11):7–8.

[28] Kong XM, Liu YL, Zhang YR, Yan PY. Static and dynamic mechanical propertiesof cement-asphalt composites. J Mater Civ Eng 2013;25(10):1489–97.

[29] Song H, Do J, Soh Y. Feasibility study of asphalt-modified mortars using asphaltemulsion. Constr Build Mater 2006;20(5):332–7.

[30] Aligizaki KK. Pore structure of cement-based materials: testing, interpretationand requirements. 1st ed. Abingdon: Taylor & Francis; 2006.

[31] Mccarter WJ, Brousseau R. The A.C. responses of hardened cement paste. CemConcr Res 1990;20(6):891–900.

[32] Gu P, Xie P, Fu Y, Beaudoin JJ. A.C impedance phenomena in hydrating cementsystems: frequency dispersion angle and pore size distribution. Cem Concr Res1994;24(1):86–8.

[33] Cabez M, Merino P, Miranda A, Novoa XR, Sanchez I. Impedance spectroscopystudy of hardened Portland cement paste. Cem Concr Res 2002;32(6):881–91.

[34] Cabez M, Merino P, Miranda A, Novoa XR, Sanchez I. Impedance spectroscopyto characterize the pore structure during the hardening process of Portlandcement paste. Electrochim Acta 2006;51(8–9):1831–41.

[35] Cabez M, Merino P, Miranda A, Novoa XR, Sanchez I. Electrical effectsgenerated by mechanical loading of hardened Portland cement paste. CemConcr Compos 2003;25(3):351–6.

[36] Woo LY, Wansom S, Ozyurt N, Mu B, Shah SP, Mason TO. Characterizing fiberdispersion in cement composites using AC – Impedance Spectroscopy. CemConcr Compos 2005;27(6):627–36.

[37] Ferrari L, Kaufmann J, Winnefeld F, Plank J. Interaction of cement modelsystems with superplasticizers investigated by atomic force microscopy, zetapotential, and adsorption measurements. J Colloid Interface Sci2010;347(1):15–24.

[38] Flatt RJ, Houst YF. A simplified view on chemical effects perturbing the actionof superplasticizers. Cem Concr Res 2001;31:1169–76.

[39] Bentur A, Igarashi S, Kovler K. Prevention of autogenous shrinkagein high-strength concrete by internal curing using wet lightweight aggregates. CemConcr Res 2001;31(11):1587–91.

[40] Loukili A, Khelidj A, Richard P. Hydration kinetics, change of relative humidity,and autogenous shrinkage of ultra-high-strength concrete. Cem Concr Res1999;29(4):577–84.

[41] Xu ZZ, Gu P, Xie P, Beaudoin JJ. Application of A.C. impedance techniques instudies of porous cementitious materials (III): ACIS behavior of very lowporosity cementitious systems. Cem Concr Res 1993;23(5):1007–15.

[42] Gu P, Xu ZZ, Xie P, Beaudoin JJ. Application of A.C. impedance techniques instudies of porous cementitious materials (I): influence of solid phase and poresolution on high frequency resistance. Cem Concr Res 1993;23(3):531–40.

[43] Xu ZZ, Gu P, Xie P, Beaudoin JJ. Application of A.C. impedance techniques instudies of porous cementitious materials (II): relationship between ACISbehavior and the porous microstructure. Cem Concr Res 1993;23(4):853–62.

[44] Song GL. Equivalent circuit model for AC electrochemical impedancespectroscopy of concrete. Cem Concr Res 2000;30(11):1723–30.

[45] Wei DF, Chatterjee I, Jones DA. Evaluation of corrosive degradation in coatedsteel using alternating current impedance spectroscopy. Corrosion1995;51(2):97–104.

[46] Xie P, Gu P, Xu ZZ, Beaudoin JJ. A rationalized A.C. impedance model formicrostructural characterization of hydrating cement systems. Cem Concr Res1993;23(2):359–67.

[47] Ohama Y. Polymer-based admixtures. Cem Concr Compos 1998;20:189–212.[48] Huang B, Wu H, Shu X, Burdette EG. Laboratory evaluation of permeability and

strength of polymer-modified pervious concrete. Constr Build Mater2010;24(5):818–23.

[49] Vocka R, Galle C, Dubois M, Loverab P. Mercury intrusion porosimetry andhierarchical structure of cement pastes theory and experiment. Cem Concr Res2000;30(4):521–7.

[50] Silva DA, Monteiro PJM. The influence of polymers on the hydration ofPortland cement phases analyzed by soft X-ray transmission microscopy. CemConcr Res 2006;36(8):1501–7.

[51] Zhong SY, Shi ML, Chen ZY. A Study of Polymer-modified mortar by the ACimpedance technique. Cem Concr Res 2002;32(6):979–82.

[52] Zhang YR, Kong XM, Hou SS, Liu YL, Han S. Study on the rheological propertiesof fresh cement asphalt paste. Constr Build Mater 2012;37(1):534–44.

[53] Yang JB, Yan PY, Kong XM, Li X. Study on the hardening mechanism of cementasphalt binder. Sci China Ser E 2010;53(5):1406–12.