Construction and Building Materials 68 (2014) 434–443

Contents lists available at ScienceDirect

Construction and Building Materials

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

Synthesis of novel polymer nano-particles and their interaction withcement

http://dx.doi.org/10.1016/j.conbuildmat.2014.06.0860950-0618/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Tel.: +86 10 6278 3708.E-mail address: kxm@tsinghua.edu.cn (X. Kong).

Kong Xiangming a,⇑, Shi Zhihua b, Lu Zichen a

a Key Laboratory of Safety and Durability of China Education Ministry, Department of Civil Engineering, Tsinghua University, Beijing 100084, Chinab Chemical and Environmental Engineering College, China University of Mining and Technology, Beijing 100083, China

h i g h l i g h t s

� A novel superplasticizer, polymernano-particle (PNP) was put forwardand synthesized.� PNP can adsorb on cement surface

and improve the fluidity of freshcement pastes.� PNPs’ retardation effect on cement

hydration is less significant thantraditional PCEs.� PNPs reduce the connectivity of

micro-pores in hardened cementpastes.

g r a p h i c a l a b s t r a c t

a r t i c l e i n f o

Article history:Received 24 October 2013Received in revised form 26 May 2014Accepted 30 June 2014

Keywords:AdmixtureWorkabilityAdsorptionHydrationPore size distribution

a b s t r a c t

Polymer nano-particles (PNPs) with particle size range of 29.4–52.7 nm were synthesized via emulsionpolymerization. The mini-cone tests were conducted to evaluate the dispersion capability of PNPs in freshcement pastes (fcps). Interactions of PNPs with cement were studied by measurements of total organiccarbon, zeta potential, transmission electron microscopy, calorimetry and mercury intrusion porosime-try. Results show that the prepared PNPs can be adsorbed on to cement surface and improve fluidityof fcps effectively. The addition of PNPs leads to lesser retardation effect on cement hydration than pop-ularly used polycarboxylate superplasticizers and reduces pore connectivity of micro-pores in hardenedcement pastes.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Concrete technology has undergone rapid advancement in thelast decades. Workability and durability of concrete are gainingmore and more emphasis [1]. With the intensive development ofchemical admixtures, a large variety of polymers are being usedin concrete formulations in forms of water soluble polymers, poly-mer fibers, polymer dispersions and redispersible powders, etc. toachieve desired properties [2].

The use of superplasticizers has drastically changed the proper-ties of concrete nearly in all aspects including workability, strength,impermeability and durability due to the lowered water to cementratio. The development of superplasticizers experiences the firstgeneration of lignosulfate, the second generation of polycondensate(poly (melamine sulfonate), poly (naphthalene sulfonate)) to thelate generation of polycarboxylate (PCE) type. PCE superplasticizersare comb-like water soluble polymers, which are usually composedof polycarboxylate main chains and polyethylene oxide as sidechains. The main chains usually contain anionic anchor groups suchas –COO�, –SO3

� to allow the molecules to adsorb onto surfaces ofcement and hydrates. Adsorption of PCE on the cement surface

X. Kong et al. / Construction and Building Materials 68 (2014) 434–443 435

generates the electrostatic repulsion force and steric hindrance, andconsequently leads to better dispersion of cement grains in watermedium. This unique mechanism makes PCE have more robust dis-persing power than polycondensates [3,4].

On the other hand, polymer latexes have been widely applied incementitious systems to improve important properties of fresh andhardened mortars and concretes such as adhesion, cohesion, flex-ural strength, impermeability, water-proofing and durability [5–8]. Since many years, it has been a field of intense research to dem-onstrate the properties of the polymer modified mortar (PMM) andconcrete (PMC), to disclose the mechanism of PMM and PMC andto describe the microstructural formation and evolution of PMMand PMC. Several models have been put forward to illustrate themechanism of PMM and PMC including the most famous Ohamatheory [2] and Puterman model [9].

PCE superplasticizers are usually fully water soluble polymersand the hydraulic radius of PCE molecule is about 10 nm in aque-ous solution [10]. Casein, a biopolymer has been known as cementdispersant for a long time [11–13]. Thanks to its excellent plasticiz-ing effect and self-healing effect on the surface of the grout, caseinis widely used in self-leveling underlayments. Casein is composedof protein molecules and calcium phosphate clusters. When caseinis dispersed in water, it forms colloidal particles usually with diam-eter of 90–140 nm, which are built of submicelles with size in therange between 12 nm and 15 nm in diameter. Plank et al. havestudied the working mechanism of casein as superplasticizer forcement. They suggested that casein particles dissociate into submi-celles with size of 10–20 nm in alkaline solution. The smaller sub-micelles show higher adsorption on cement surface and producebetter superplasticizing effect [11].

Due to the special bio-resource and fluctuant quality for differentbatches, the application of casein in concrete is limited compared tothe industrially synthesized products such as PCE superplasticizers.Our question is then, ‘Is it possible to synthesize nano-particle dis-persions that may serve as superplasticizer for cement?’ If yes, wewill have a chance to obtain similar superplasticizing effect ascasein, and even to combine the functions of water soluble superp-lasticizers like PCE and the polymer latex dispersions.

Some researchers have studied the effects of polymer latexes onproperties of fresh cement mortars and the interaction betweenlatex particles and cement. Plank found that both cationic andanionic polymer latex particles can adsorb onto the inorganic bin-der surface due to the heterogeneous charge distribution devel-oped by the mineral phases in cement and the cement hydrates[14]. Some literatures reported that addition of polymer latexesinfluences the flowability as well as workability of cement mortars.Some are beneficial while some are detrimental to the workabilityof fresh cementitious mixtures [15–17]. Rheological study on freshcement asphalt paste conducted by Zhang found that both anionicand cationic asphalt emulsion improve the flowability of the pastesand the anionic asphalt emulsion is more effective than the cat-ionic type due to its favorable adsorption [18].

However, all literatures above mentioned are dealing with poly-mer latexes with relatively big particle size of several hundreds ofnanometers or with asphalt emulsions having particles size as bigas several microns. Emulsion polymerizations including mini-emulsion polymerization, traditional emulsion polymerizationand micro-emulsion polymerization enable to produce polymerdispersions with finely tuned particle size from 20 nm to hundredsof nanometers. In this paper, we used semi-batch emulsion poly-merization to synthesize anionic polymer nano-particle (PNP) dis-persions with particle size around 30–50 nm. We have chosenpolystyrene dispersion as a model system to explore the possibilitythat the PNPs can be functioning as cement dispersant similar tothe PCE superplasticizers. Influences of the structural parametersof the synthesized PNP dispersions on the fluidity of fresh cement

pastes (fcps) were firstly investigated. After that, interaction of aselected PNP dispersion with cement was specifically studied bymeans of adsorption measurement, zeta-potential measurement,isothermal calorimetry, transmission electron microscopy (TEM)as well as mercury intrusion porosimetry (MIP). Adsorption ofthe PNPs on the cement grains was investigated by total organiccarbon (TOC) tests on the supernatant that was separated by cen-trifuging the cement pastes. Zeta potentials of cement pastes withwater to cement ratio (w/c) of 0.5 were measured at various dos-ages of PNP dispersions. Isothermal calorimetry was adopted tostudy the impacts of the PNPs on cement hydration. Pore structureof the hardened cement pastes (hcps) with varied contents of PNPwas characterized by MIP with the expectation that addition ofPNPs can modify the pore structure of hcps and consequentlyenhance the impermeability and durability of cementitiousmaterials.

2. Materials and methods

2.1. Materials

Analytical grade of styrene (St), acrylic acid (AA), sodium dodecyl sulfate (SDS),sodium methyl acryl sulfonate (SMAS) and sodium persulfate (SPS) were used asreceived (all >98% purity). Marco-monomer, methoxy polyethylene glycol methac-rylate (MPEGMA, 85%), was prepared via esterification reaction of methacrylic acidand polyethylene glycol whose weight average molecular weight is about 1300 at130 �C. Emulsifier MS-1 (C8H17C6H4O(CH2CH2O)10COC3H6-SO3Na, 40 wt% aqueoussolution) and emulsifier OP-10 (C8H17C6H4O(CH2CH2O)10H, 40 wt% aqueous solu-tion) were provided by Haian petrochemical factory. A home-made PCE superplast-icizer was used to benchmark the plasticizing effect of the PNPs in cement pastes.The PCE used is a co-polymer of AA, MPEGMA and SMAS with monomer molar ratioof 2.12:1.00:0.29. The weight average molecular weight of the PCE used is9.08 � 104 which was measured by gel permeation chromatograph.

Ordinary Portland cement classified by P.O.42.5 and compliant with the ChineseNational Standard GB8076-1997 was used to prepare the cement pastes. Chemicalcomposition and mineral composition of this cement are presented in Table 1.

2.2. Preparation and characterization of the polymer nano-particle dispersions

PNP dispersions were prepared by micro-emulsion polymerization in a 1000 mLthree-neck glass flask equipped with a mechanical stirrer, dosing units for both thewater-soluble monomers and the oil-soluble monomer. A water bath with tunedtemperature of 85 �C was used to ensure the constant temperature during polymer-ization. Firstly, 10% of the total amount of St, 80% of the total emulsifiers, 10% of ini-tiator (SPS) and �190 g de-ionized (DI) water were well mixed by a high-shearmixer and then charged into the flask. The pre-charge was heated up to 85 �C understirring and pre-polymerized for 10 min. The pre-emulsion, which was prepared bymixing all the rest of monomers, emulsifiers and 100 g DI water, and the aqueoussolution of the rest initiator SPS with concentration of 4.45% were then separatelydosed into the flask at constant dosing rates. The dosing time of pre-emulsion andthe initiator solution were 2.5 h and 3.0 h respectively. After polymerization, theobtained polymer dispersion was cooled down to the room temperature. Recipesfor preparation of PNP dispersions are shown in Table 2.

Solid content of the prepared polymer dispersion was measured by drying aportion of the dispersion at 80 �C until constant weight was reached. The solid con-tent of all the prepared polymer dispersions was in a range of 18–20%. The driedpolymer was then dissolved in tetrahydrofuran (THF) and closely kept under stir-ring for 24 h. Afterwards, the insoluble part of the polymer was separated by usinga copper mesh filter and the filtrate was used to determine the molecular weight ofpolymer by gel permeation chromatograph (GPC) (Shimadzu, LC-20AD, Japan). Tet-rahydrofuran (THF) was used as the eluent and the flow rate was 1 mL/min. Duringemulsion polymerization, cross-linked polymers may be produced due to chaintransfer reaction and they are not any more soluble in solvent [19]. This insolublepart of polymer is called gel. Gel content as well as the molecular weight distribu-tion of the soluble part of polymer was listed in Table 3. The results confirm the suc-cessful polymerization for those samples studied in this paper. Particle size of thepolymer particles in PNP dispersions was determined by dynamic light scattering(DLS) with a Malvern Zetasizer 3000hs (UK). The characterization results of the syn-thesized PNP dispersions are presented in Table 3.

2.3. Mini-cone tests

The mini-cone test was conducted according to the Chinese National StandardGB/T 8077-2000 to evaluate the fluidity of fcps. The mixing procedure of the cementpaste was as follows: Water and PNP dispersions were first added into the mixer.300 g cement was gradually introduced over a time span of 2 min into the mixer

Table 1Chemical composition and mineral composition of the standard Portland cement.

Chemical composition w/% Mineral composition w/%

SiO2 Al2O3 Fe2O3 CaO MgO SO3 Na2Oeq f-CaO C3S C2S C3A C4AF

22.10 4.04 3.38 61.91 2.66 2.87 0.56 0.79 57.34 18.90 6.47 11.25

Table 2Recipes for preparation of PNP dispersions.

No. Anionic emulsifier (g) Nonionic emulsifier (40%) (g) St (g) MPEGMA (g) AA (g) SMAS (g) SPS (g)

Series 1 PNP-E1 SDS 5.98 OP-10 7.48 50.00 15.00 3.00 1.00 1.04PNP-E2 MS-1(40%) 14.95 OP-10 7.48 50.00 15.00 3.00 1.00 1.04

Series 2 PNP-R1 SDS 4.49 OP-10 11.13 50.00 15.00 3.00 1.00 1.04PNP-R2 SDS 5.98 OP-10 7.48 50.00 15.00 3.00 1.00 1.04PNP-R3 SDS 7.18 OP-10 4.49 50.00 15.00 3.00 1.00 1.04

Series 3 PNP-P1 SDS 1.73 OP-10 4.31 50.00 15.00 3.00 1.00 1.04PNP-P2 SDS 2.42 OP-10 6.04 50.00 15.00 3.00 1.00 1.04PNP-P3 SDS 3.80 OP-10 9.49 50.00 15.00 3.00 1.00 1.04PNP-P4 SDS 4.49 OP-10 11.13 50.00 15.00 3.00 1.00 1.04PNP-P5 SDS 5.18 OP-10 12.94 50.00 15.00 3.00 1.00 1.04

Series 4 PNP-A1 SDS 4.36 OP-10 10.89 50.00 15.00 1.00 1.00 1.01PNP-A2 SDS 4.42 OP-10 11.05 50.00 15.00 2.00 1.00 1.02PNP-A3 SDS 4.49 OP-10 11.13 50.00 15.00 3.00 1.00 1.04PNP-A4 SDS 4.55 OP-10 11.38 50.00 15.00 4.00 1.00 1.05

Series 5 PNP-M1 SDS 4.16 OP-10 10.40 50.00 10.00 3.00 1.00 0.96PNP-M2 SDS 4.49 OP-10 11.13 50.00 15.00 3.00 1.00 1.04PNP-M3 SDS 4.81 OP-10 12.03 50.00 20.00 3.00 1.00 1.11PNP-M4 SDS 5.14 OP-10 12.84 50.00 25.00 3.00 1.00 1.19

Table 3Characteristics of the synthesized PNP dispersions.

No. Gel content (%) Mw Mn Polydispersibilty Mw/Mn Particle size (nm) Particle size distribution index

Series 1 PNP-E1 29.85 148,500 55,390 2.681 32.5 0.11PNP-E2 53.73 110,100 43,840 2.511 52.7 0.32

Series 2 PNP-R1 40.81 129,500 51,060 2.536 30.7 0.13PNP-R2 29.85 148,500 55,390 2.681 32.5 0.11PNP-R3 40.38 195,400 72,560 2.707 32.2 0.13

Series 3 PNP-P1 89.32 187,300 79,130 2.367 47.2 0.08PNP-P2 0 213,400 69,370 3.007 39.8 0.12PNP-P3 40.38 195,400 72,560 2.707 35.8 0.14PNP-P4 40.81 129,500 51,060 2.536 30.7 0.13PNP-P5 20.00 157,000 62,360 2.518 29.4 0.10

Series 4 PNP-A1 0 125,400 55,660 2.253 31.2 0.11PNP-A2 16.41 204,000 70,420 2.897 33.3 0.11PNP-A3 40.81 129,500 51,060 2.536 30.7 0.13PNP-A4 60.19 155,200 66,330 2.340 33.7 0.10

Series 5 PNP-M1 62.69 139,200 63,170 2.204 35.9 0.14PNP-M2 40.81 129,500 51,060 2.536 30.7 0.13PNP-M3 33.33 213,400 72,530 2.943 31.4 0.09PNP-M4 9.09 185,700 66,310 2.801 32.2 0.23

436 X. Kong et al. / Construction and Building Materials 68 (2014) 434–443

at rotating rate of 62 rpm. After a 15 s interval, mixing was resumed for another2 min at rotating rate of 125 rpm. The fluidity of the fcps was represented by thespread flow in the mini-cone test. A cone with an upper diameter of 36 mm, lowerdiameter of 60 mm, and height of 60 mm was used in the test. The spread diameterwas recorded as the average of two perpendicularly crossing diameters.

In the fluidity tests, the w/c of the cement pastes was set at 0.35 and the dosageof PNPs was varied in the range of 0.2–2.0% by weight of cement (bwoc). For eachcement paste, the spread diameter was respectively recorded at time point of 5, 30,60 and 120 min after water–cement contact. The spread diameter at 5 min wasdefined as the initial fluidity, while the latter values could reflect the fluidity reten-tion of fcps.

2.4. Adsorption amount of polymer nano-particles on cement surface

As well known, the dispersing function of superplasticizers in cement pastes isfacilitated by the adsorption of superplasticizer molecules on the surface of cementgrains. Similarly, the adsorption amount of PNPs on cement grains was measured inorder to understand their effects on fluidity of cement pastes by using a TOC Ana-lyzer (Shimadzu, TOC-VCPH, Japan).

Cement pastes were prepared with a w/c of 0.35 and PNP dosages in the rangeof 0–2.0% bwoc. The freshly prepared fcps were then centrifuged at 4000 rpm for5 min, and the clear supernatant solutions were collected by using a syringe filterwith pore diameter of 0.22 lm. TOC tests were conducted to determine the concen-tration of PNPs in aqueous solution. Thus, the adsorbed amount of PNPs on cementsurface was then calculated by deducting the amount of PNPs remaining in theaqueous phase from the initially added PNP amount in the cement pastes. Theadsorption ratio is defined as the adsorbed amount of PNPs divided by the totaladded amount of PNPs.

2.5. Transmission electron microscopy (TEM)

To directly observe morphology of the synthesized PNPs and the adsorption ofPNPs on cement grains, transmission electron microscope (Tecnai G2 F20-200 kV)with EDX was used. PNP dispersion was diluted to solid content of 0.5% using DIwater. A drop of the diluted PNP dispersion was then placed on the copper meshfor TEM observation. Moreover, a cement paste with w/c of 0.35 and PNP dosage1.0% bwoc was prepared to observation the adsorption of PNPs on surface of cementgrains. The freshly mixed cement paste was diluted to solid content of 1.0% by

X. Kong et al. / Construction and Building Materials 68 (2014) 434–443 437

adding alcohol and ultrasonic oscillation was exerted for 10 min. The sample wasthen dropped on the copper mesh. The TEM observation was performed after thesample dried out.

2.6. Zeta potential measurement

Electroacoustic method enables to measure the zeta potential of a highly solidsloaded suspensions [14,20]. Only such suspensions truly represent the conditionsexisting in a mortar or concrete. Zeta potentials of the cement pastes were mea-sured with a DT 1201 instrument produced by Dispersion Technology (USA). Thew/c of cement pastes was set as 0.50 with various additions of PNPs (0–2.0% bwoc)or PCE (0–0.5% bwoc) for the zeta potential measurement.

2.7. Isothermal calorimetry

Isothermal calorimetry tests were conducted on the cement pastes at 25 �C byusing a TAM Air calorimeter (Thermometric AB, Sweden) to investigate the influ-ences of the PNPs on cement hydration. Prior to the calorimetry tests, the calorim-eter was regulated at 25 �C and then equilibrated for 24 h. Thereafter, 7 g of thefreshly well mixed cement pastes with w/c of 0.35 and various dosages of PNPsor PCE were placed in 20 mL ampoule bottles and then introduced into the channelsof the micro-calorimeter. The heat evolution was recorded for 60 h after water–cement contact.

2.8. Mercury intrusion porosimetry (MIP)

MIP has been a mature method to characterize the pore structure of the hcps.MIP measurements were conducted to investigate the effects of PNPs on the porestructure of the hcps. The w/c of cement pastes was set as 0.35 and various dosagesof PNPs (0.0%, 0.2%, 0.6%, 1.0% and 2.0% bwoc) or dosage of PCE (0.15% bwoc) wereapplied. After being cured for 28 days, the specimens were cut into small pieceswith thickness of about 3 mm, and placed into an alcohol bath. After 3 days storageat 60 ± 2 �C for completely drying, they were subjected to MIP tests to determinepore structure characteristics by using an Hg-porosimetry (Autopore, IV 9510, USA).

3. Results and discussion

3.1. Synthesis and characterization of polymer nano-particles

Via micro-emulsion polymerization, semi-transparent disper-sions (Fig. 1a) with solid content of 18–20 wt.% and particle sizeranging from 29.4 nm to 52.7 nm were obtained under varied syn-thesis conditions, including varied amount of water soluble poly-mers and emulsifier combinations as synthesis parameters (seeTable 2). According to the theory of emulsion polymerization[19,21], after copolymerization of styrene and water soluble mono-mers such like AA and MPEGMA, the polystyrene rich part formsthe major part of the nano-particles because of its high hydropho-bicity, while polyacrylic acid (PAA) and MPEGMA rich polymersstay at the interface between the nano-particle and the aqueousphase due to their high hydrophilicity. Thus, PNPs, which are com-posed of polystyrene as the major component and a minor amountof water soluble copolymers chemically attached on the PNPs’ sur-face as hairy layer, were obtained as illustrated in Fig. 1b. TEMtechnique was employed for direct observation of the polymernano-particles. As seen from Fig. 2, PNP-R1 has particle size of�30 nm and exhibits relatively narrow particle size distribution,which is in good agreement with the particle size measurementby DLS.

Fig. 1. Image of the synthesized PNP dispersion (a) an

Emulsifiers are the essential components in emulsion polymer-ization, which play important roles during polymerization by pro-viding polymerization location (nucleation stage) on the one hand,and are responsible for the stabilization of the polymer particlesafterward on the other hand. A combination of an anionic emulsi-fier and a non-ionic emulsifier is usually used, in which the non-ionic emulsifier contributes to the better stability of the polymerparticles by introducing steric hindrance between particles. In thisstudy, a combination of anionic and non-ionic emulsifiers was usedas Series 1 shown in Table 2. The particle size of PNPs is mainlycontrolled by the number of the micelles during the nucleationstage of the emulsion polymerization. Knowing that the criticalmicelle concentration (CMC) of the SDS (0.008 mol/L) is muchlower than that of MS-1 (0.015–0.025 mol/L), the emulsifier com-bination of SDS and OP-10 produces much smaller particle sizethan the combination of MS-1 and OP-10 at the same usage ofemulsifier amount (PNP-E1 and PNP-E2). In Series 2, the ratio ofthe anionic emulsifier SDS to the non-ionic emulsifier OP-10 wasvaried from 1:1 to 4:1. It is seen that the particle size of PNPs isnot largely influenced by this factor. Therefore, the ratio of SDSto OP-10 was fixed at 1:1 for further investigation. In emulsionpolymerization, the amount of emulsifiers is a very effective toolto finely tune the particle size and usually more emulsifier leadsto smaller particle size. In experiment Series 3, the amount ofemulsifiers was changed in order to tune the particle size of PNPs.With increasing amount of emulsifiers, the particle size propor-tionally decreases. In Series 4 and Series 5, the amount of watersoluble monomers AA, MPEGMA were varied respectively. Thewater soluble monomers mainly polymerize in the aqueous phaseand the produced water soluble polymers are majorly attached onthe surface of polymer particles to form a so called hairy layerwhen the water soluble monomers are copolymerized with sty-rene. The group of –COOH from monomer AA located on the sur-face of the polymer particles makes the polymer particlesnegatively charged due to its deprotonation in aqueous medium.By changing the amount of AA, the amount of negative charge onthe particle surface can be controlled (Series 4). The existence ofcopolymerized MPEGMA on the particle surface may bring sterichindrance both to the polymer particles and to the cement grainsonce PNPs adsorb onto the cement surface. Therefore, the amountof MPEGMA monomer was varied to study the effect of the sterichindrance (Series 5). It is seen from Table 3 that the variations inamounts of AA and MPEGMA do not lead to large changes in parti-cle size of PNPs. Increases in amounts of AA or MPEGMA mainlyresult in increased density of hairy layer on the surface of PNPs.

3.2. Fluidity of the fcps with polymer nano-particle dispersions

The influences of synthesis variables of PNP dispersions on thefluidity of the fcps are shown in Fig. 3. The fluidity of the fcps withw/c of 0.35 and without addition of superplasticizers is very poorand the spread diameter is only about 60 mm. As seen in Fig. 3,at dosage of 1.0% bwoc, the addition of various PNPs greatlyincreases the fluidity of the fcps. This certainly proves the concept

(b)

d a schematic drawing of the PNP structure (b).

Fig. 2. TEM images of the synthesized polymer nano-particles (PNP-R1).

Fig. 3. Fluidity of the fcps with various PNP dispersions (a. Series 1; b. Series 2; c. Series 3; d. Series 4; e. Series 5).

438 X. Kong et al. / Construction and Building Materials 68 (2014) 434–443

that the synthetic PNPs can also provide plasticizing effect tocement mixtures similar as the traditional superplasticizers suchlike PCE. The working mechanism of PNPs is proposed as schematicillustration in Fig. 4. As well understood, the dispersing capabilityof PCE in cement mixtures is provided by the adsorbed PCE mole-cules on the cement surface, which generate electrostatic repulsionforce and steric hindrance between cement grains. Consequently,this leads to the better dispersion of cement grains in water med-ium. This unique mechanism makes PCE have much more robustdispersing force than other types of superplasticizers [3,4]. Simi-larly to PCE molecules, the anionic PNPs are able to be adsorbedon to the surface of cement grains and hydrates due to the electro-static interaction between PNPs and cement surface. The adsorbedPNPs provide electrostatic repulsion as well as steric hindrancebetween cement grains. The adsorbed PCE molecules provide thesteric hindrance by the PEO (polyethylene oxide) sides, while in

case of PNP, the steric hindrance may be generated by the volumeeffect of the PNPs as well as the hairy layer attached on their sur-face. Therefore, even stronger steric hindrance effect than PCE mol-ecules could be expected for PNPs. This way, the synthesized PNPsare expected to work as a superplasticizer in cement. As shown inFig. 4, the thickness of adsorption layer of PNPs with particle size of30 nm will be notably larger than that of PCE, which is reported as�3–10 nm [10]. The thicker adsorption layer will on the one handbring stronger steric hindrance effect, which is beneficial to thedispersing effect. On the other hand, it will require higher dosageto cover the cement surface, which is not favored in the practicalapplication due to the increased material demand.

Fig. 3a indicates that the fluidity of the fcp with PNP-E1 is largerthan that of the fcp with PNP-E2 over the first two hours. This isbelieved to be the result of the smaller particle size of PNP-E1 thatleads to higher surface coverage of cement grains by adsorption on

Fig. 4. The proposed dispersing mechanisms of PCE superplasticizer (a) [10] and PNPs (b).

X. Kong et al. / Construction and Building Materials 68 (2014) 434–443 439

cement surface than PNP-E2 at the same dosage. In Series 2, thevariation in the ratio of anionic to non-ionic emulsifiers in nano-particle dispersions does not bring great influence on the initial flu-idity of the fcps, while the fluidity retention over time is obviouslyreduced by the higher fraction of anionic emulsifier in the emulsi-fier combination (Fig. 3b). As known in case of PCE, the more PEOside chains allow PCE having higher fluidity retention capability forthe fcp [22]. It is seen the similar effect here that the higher frac-tion of non-ionic emulsifier attached on the surface of PNPs canhelp to provide stronger steric hindrance between cement grainsand consequently higher fluidity retention. From Fig. 3c, it isclearly seen that at the same dosage, PNPs with smaller particlesize lead to higher initial fluidity as well as the fluidity at 2 h. Thismust result from the higher surface coverage ratio when the smal-ler PNPs are adsorbed on cement surface, suggesting that PNPswith smaller particle size have higher plasticizing efficiency inthe fcps. As introduced before, by changing the amount of mono-mer AA, the surface charge density of PNPs can be modified. Asseen from Fig. 3d, PNPs produced with more fraction of monomerAA show stronger plasticizing power. However, too large amountof AA may lead to a drop in fluidity retention capability of thefcp as the case of PNP-A4. It has been frequently reported thatthe higher fraction of macro-monomer MPEGMA in PCE leads tohigher fluidity retention for the fcp, because the higher density ofside chain in PCE molecules provide stronger steric hindrancebetween cement grains [23]. Similar effect is seen in Fig. 3e.PNP-M1 shows poor capability on improving the fluidity retentionof fcp although it has strongest initial plasticizing power amongthe four PNPs with varied amount of MPEGMA. With increase inmacro-monomer MPEGMA during synthesis, the fluidity retentioneffect of PNPs is increasing, which is believed to be the result of thehigher density of hairy layers on the surface of nano-particles pro-viding stronger steric hindrance between cement grains once theyare adsorbed on to cement surface. In summary from Fig. 3, PNP

Fig. 5. Initial spread diameter of PNP and PCE wi

can work as a superplasticizer in cementitious system and the syn-thesis parameters strongly influence their performances.

The curves of the fluidity of the fcps versus dosages of PCEsuperplasticizer usually exhibit a typical S shape [24]. Namely,the fluidity of fcps reaches a maximum at a certain dosage ofPCE, which can be defined as the critical dosage. And furtherincrease in PCE dosage does not increase the fluidity of the fcps.Fig. 5 presents the performance comparison between a typicalPCE superplasticizer and PNPs aforementioned. It is clearly seenthat the fluidity of the fcps certainly increases with increase in dos-age of PNPs and some of them also present S shape curves such asPNP-A3, PNP-A4, PNP-M1 and PNP-M2. However, the critical dos-ages of PNPs are generally much higher than that of PCE and themaximum fluidity of the fcps added with PNPs is lower than thatwith PCE. This suggests that the plasticizing efficiency of PNPs islower than that of PCE, which should be mainly because of theirlarger particle size and thicker adsorption layer on cement surface.

3.3. Adsorption of polymer nano-particles on the cement grains

3.3.1. Measurement of adsorption amount by TOCIt has been well accepted that the superplasticizer works as a

dispersant in cement pastes by firstly adsorbing on to the cementsurface and then the adsorbed superplasticizer molecules intro-ducing electrostatic repulsion and/or steric hindrance betweencement grains. The adsorption amount is one of the key factorsdetermining the efficiency of superplasticizers. The adsorptionamount of superplasticizer in the fcps can be measured by deduct-ing the amount of superplasticizer remaining in aqueous phasefrom the total added superplasticizer. TOC is the often used tech-nique to determine the residual concentration of superplasticizerin the aqueous phase of the fcps. The adsorption amount of bothPCE and the synthesized PNP-R1 are compared as shown inFig. 6. It is seen that PNPs are able to be effectively adsorbed on

th different dosages (a. AA%; b. MPEGMA%).

Fig. 6. Comparison of the adsorption of PNPs and PCE in cement pastes at variousdosages.

Fig. 8. Zeta potentials of the fcps containing PCE and PNPs at various dosages.

440 X. Kong et al. / Construction and Building Materials 68 (2014) 434–443

to cement surface and the adsorption amount of PNP-R1 is muchlarger than that of PCE at the same dosage. The driving force ofadsorption on cement surface for both PCE and PNPs should bethe electrostatic interaction [25], as parts of cement surface arepositively charged while PCE and PNPs are negatively charged[14,26]. In addition, as described in Fig. 4, the adsorption layerfor PNPs must be much thicker than that of PCE due to its largerparticles size. Therefore, the saturated adsorption amount of PNPsis much higher than that of PCE in mass.

3.3.2. Transmission electron microscopy observationIt is hardly possible to direct observe the adsorbed layer of PCE

on cement surface due to the very small size of PCE molecules,whereas PNPs are usually observable by either SEM or TEM tech-niques. The morphology of adsorption layer of PNPs is directlyobserved by TEM as shown in Fig. 7. A mono-layer of nano-parti-cles adsorbed on cement surface is clearly seen in Fig. 7. EDX isapplied to determine the elemental composition in the selectedranges during TEM observation. It is interesting to note that theadsorption density of PNPs on the surface of aluminates is higherthan that on the surface of silicates. This finding is in well agree-ment with the results previously reported in the literatures[14,26]. When cement is dispersed in water, a heterogeneouscharge distribution is developed on the surface of hydratingcement grains. The surfaces of aluminate phases (C3A, ettringiteand monosulfate) are usually positively charged, whereas the sur-faces of silicate phases (C3S and C–S–H) are negatively charged. Thepositively charged aluminate surfaces can directly adsorb the anio-nic PNPs by electrostatic interaction. It should be noted that for thenegatively charged surfaces of silicates, a layer of counterions(mainly Ca2+) is formed and hence this enables the adsorption ofanionic PNPs. That is to say that the anionic PNPs are also able to

Fig. 7. TEM images of PNPs adsorbed on cement grains

adsorb on to the negatively charged surfaces of silicates via a jointCa2+ layer but with a lower adsorption density than on the surfacesof aluminates. The TEM observation (Fig. 7b) exactly confirms thededuction above.

3.4. Effect of polymer nano-particles on zeta potential of the freshcement pastes

It has been well documented that the addition of PCE in the fcpssignificantly reduces the zeta potential of the fcps [27]. The zetapotentials of fcps containing either PCE or PNPs were measuredon a cement pastes with w/c of 0.50 by using the electroacousticinstrument DT-1201 and the results are shown in Fig. 8. The mea-sured zeta potential of pure cement paste is about 4.2 mV. Theaddition of PCE in the fcps leads to a sharp decrease of zeta poten-tial. At PCE dosage of 0.3% bwoc, the zeta potential of fcp reachesthe lowest value of �10 mV and stops to change with furtherincreasing dosage of PCE. It can be seen from Fig. 8, the additionof PNP-R1 also reduces the zeta potential of the fcps but with muchlower effectiveness than PCE in reducing the zeta potential of fcps.This again confirms the adsorption of PNPs on cement surface. Thelower effectiveness of the anionic PNPs in decreasing zeta potentialof fcps than PCE is believed to result from the larger particle sizeand hence the thicker adsorption layer. The thicker adsorptionlayer extends the shear plane of cement grains outwards and con-sequently leads to a lowered absolute value of zeta potential.

3.5. Impact of polymer nano-particles on cement hydration kinetics

The influences of PCE on the cement hydration kinetics have beenintensively investigated [28,29]. Retardation on cement hydration isa characteristic behavior for most PCE superplasticizers. Isothermal

and the EDX spectrum for different mineral phases.

Fig. 9. Calorimetry curves of cement pastes with PCE and PNPs at different dosages.

X. Kong et al. / Construction and Building Materials 68 (2014) 434–443 441

calorimetry was employed to compare the impacts of PCE and PNPson cement hydration kinetics in a cement paste with w/c of 0.35. Theresults are presented in Fig. 9 and Table 4. It is found that the addi-tion of PNPs in fcps also causes so called retardation effect but to a

Table 4Cement hydration parameters of cement pastes with PCE and PNPs at different dosages.

No. tisa (h) (dQ/dt)is

b (*10�4 w/

Blank 2.98 5.14PNP-R1 0.2% 3.42 4.66PNP-R1 0.6% 4.77 4.18PNP-R1 1% 6.37 3.42PNP-R1 2% 13.47 2.52PCE 0.15% 9.04 2.66PCE 0.3% 12.35 2.10PCE 0.6% 22.26 1.57

a tis denotes the time point at the end of induction period or the start of the acceleratioas initial setting.

b (dQ/dt)is is the heat flow rate at the time point tis. Since the heat generation rate durinheat generation rate during induction period of cement hydration.

c tpeak is the time when the maximum heat flow rate is reached.d (dQ/dt)max is the maximum heat generation rate during the acceleration period.

Fig. 10. Pore structure of the hcps (at age of 28

Table 5Pore size distribution in the hcps (at age of 28 d) with PCE and PNPs at different dosages.

No. <10 nm(mL/g)

10–100 nm(mL/g)

>100 nm(mL/g)

Blank 0.02 0.07 0.05PNP-R1 0.2% 0.02 0.08 0.05PNP-R1 0.6% 0.02 0.09 0.06PNP-R1 1.0% 0.02 0.08 0.06PNP-R1 2.0% 0.03 0.09 0.05PCE 0.15% 0.03 0.10 0.03

lesser degree than PCE, although the dosage and the adsorptionamount of PNPs are much higher than those of PCE. Looking at moredetails, when PNPs are added into the fcps, the induction period (tis)and tpeak increase with dosages in a rough linear fashion; the heatgeneration rate during induction period (dQ/dt)is and the maximumheat generation rate (dQ/dt)max decrease linearly with dosages. ForPCE, the delaying effect on cement hydration is much stronger thanPNPs, whereas the (dQ/dt)is and the (dQ/dt)max of cement hydrationare less affected when the dosage of PCE increases. The retardationeffect of PCE has been well understood to be closely related to thegroup of –COOH contained in the molecules of PCE [26], which hasstrong complexation interaction with Ca2+ions in the pore solutionof hydrating cement paste. However, the retardation mechanismof PNPs must be different from that of PCE. We suppose that theretardation of PNPs on cement hydration results from the adsorptionof PNPs on cement surface that strongly interrupts the diffusion pro-cess of cement hydration, because it is found that the retardationeffect closely depends on the adsorption amount of the PNPs oncement surface, i.e. with the surface coverage ratio.

g) tpeakc (h) (dQ/dt)max

d (*10�3 w/g)

9.59 2.8110.29 2.7812.20 2.6813.9 2.6321.07 2.3015.71 2.7419.32 2.6630.11 2.52

n period in calorimetry curve of cement hydration. This time point is often regarded

g induction period is almost constant in most cases, the (dQ/dt)is can also reflect the

d) with PCE and PNPs at different dosages.

Center diameterof Peak 1 (nm)

Center diameterof Peak 2 (nm)

Center diameterof Peak 3 (nm)

3.99 32.04 772.143.99 32.45 364.364.51 32.26 284.284.02 28.83 747.774.54 26.41 928.184.51 40.39 No peak

442 X. Kong et al. / Construction and Building Materials 68 (2014) 434–443

3.6. Effect of polymer nano-particles on pore structure of the hardenedcement particles

Pore structure is one of the most important properties ofcementitious materials determining many properties such as thestrength development, shrinkage and permeability. The impactsof chemical admixtures on the pore structure of hardened cemen-titious materials have been intensively investigated [30–32]. Sakaifound that the addition of PCE superplasticizers reduces the vol-ume of large pores (>0.1 lm) in hcps, which is explained to berelated to better dispersion of cement grains in fcps [33]. Therefore,the pore structure of hcps partly reflects the flocculated structureof cement grains in fcps. MIP is the often used technique to charac-terize the pore structure of hardened cementitious materials.RadmiVocka has developed a MIP simulator and interpreted MIPexperimental data for the hcps with w/c of 0.30, 0.40 and 0.50.Two types of porosity are distinguished in hcps: the porosity ofhydration product (C–S–H porosity) and the macroporosity (alsocalled capillary porosity) that is usually larger than 30 nm [34].The C–S–H porosity consists of nanopores (typical diameter of 2–4 nm) representing the porosity of inner C–S–H and the micropores(typical diameter of 20–30 nm) standing for the porosity of outerC–S–H. The C–S–H porosity is assumed to be the intrinsic proper-ties of hydrates and should not evolve with w/c in a fully hydratedcement pastes. In pore size distribution curves measured by MIP,three peaks peak I, peak II and peak III are observed, which charac-terize the clusters of nanopores, micropores and macroporesrespectively. With decreasing w/c, peak III shifts to smaller diame-ters. They pointed out that the pore sizes indicated by MIP resultsdepend not only on the pore size distribution, but also on the con-nectivity of pore space.

In order to disclose the influences of PNPs on the pore struc-ture of the hcps, MIP was employed to characterize the porestructure of hcps containing PNPs or PCE at various dosages. Thepore size distributions are shown in Fig. 10 and Table 5. TheMIP results are interpreted by referring the theory proposed byRadmiVocka. It is seen that the position of peak I stays little chan-ged despite types and contents of chemical admixtures. This is inaccordance with the conclusion that peak I characterizes theintrinsic morphology of C–S–H. The most important observationfrom Fig. 10 is that peak II shifts to smaller pore size with increas-ing PNP dosage in the hcps. As explained in the theory of Radm-iVocka, peak II indicates the porosity of outer C–S–H, whichshould be the intrinsic property of C–S–H. Therefore the left shift-ing of peak II as PNP dosage increases must be a result of the low-ered the connectivity of the micropores. This implies that theaddition of PNPs in the hcp exerts strong influences on the micro-porosity due to the perfect match between the particle size ofPNPs and the pore diameters of micropores. When the PNPs andPCE are compared, the pore volume of micropores (peak II) ismuch lower in the hcps containing with PNPs than that in thehcp where PCE is added. As well known, the pores with size rangeof 30–100 nm are highly detrimental to many properties of hcpsincluding shrinkage, and permeability [35]. Therefore the modifi-cation of PNPs on the pore structure of micropores may providepotential advantages when PNPs is used as chemical admixturefor cementitious materials instead of PCE. For macroporosity, itis seen that peak III moves toward the bigger diameter withincreasing dosage of PNPs, which is believed to be the result ofthe increase in connectivity of macropores since the w/c is keptconstant. When PCE is used as chemical admixture, instead of apeak, a step appears at the location of peak III, which suggests avery low connectivity of macropores. This observation is fully inagreement with the findings of Sakai that PCE significantlyreduces the pore volume of large pores (>100 nm) due to betterdispersion of cement grains in fcp [33].

4. Conclusions

A novel type of PNP dispersions, which compose of polystyreneas the major component of the PNP dispersions and a certainamount of water soluble copolymers attached on the PNP surfaceas hairy layer, have been synthesized by micro-emulsion polymer-ization. The particle size of PNPs ranges from 29.4 nm to 52.7 nm.According to our findings, PNPs can be adsorbed on the surface ofcement grains and improve the fluidity of the fcps effectively. PNPswith smaller particle size have higher plasticizing efficiency in thefcps, because the adsorption of smaller nano-particles on cementcauses higher surface coverage ratio at the same dosage. Higherfraction of macro-monomer MPEGMA or non-ionic emulsifier insynthesis of PNP dispersions leads to higher fluidity retention overelapsed time for the fcps, because they provide stronger steric hin-drance between cement grains. The fluidity of fcps increases withincrease in dosage of PNPs and for some of PNPs, the curve of flu-idity of fcps versus dosage of PNPs also present S shape similar totypical PCE superplasticizers. The plasticizing efficiency of PNPs islower than that of PCE, which should be due to their larger particlesize and thicker adsorption layer on cement surface. Upon adsorp-tion of PNPs on cement surface, the addition of PNPs decreases thezeta potential of cement paste. The lower effectiveness of the anio-nic PNPs in decreasing zeta potential of fcps than that of PCE isbelieved to also result from its larger particle size and hence thethicker adsorption layer. It is found that the addition of PNPs infcps also causes so called retardation effect but to a lesser extentthan PCE. We suppose that the retardation of PNPs on cementhydration is a result of the adsorption of PNPs on cement surfacethat interrupts the diffusion process of cement hydration, becausethe retardation effect is closely related to the adsorption of nano-particles on cement surface, i.e. to the surface coverage ratio. PNPsexert strong influences on the pore structure of micropores in hcpsdue to the perfect match between the particle size of PNPs and thepore size of micropores in hardened cement pastes. The addition ofPNPs leads to the peak of micropores shifting to smaller size due tothe lowered connectivity. The pore volume of micropores in hcpscontaining with PNPs is much smaller than that of hcps wherePCE is used as plasticizer. The peak of macropores move towardbigger size as PNP dosage increases. The modification of PNPs onthe pore structure of micropores may provide potential advantageswhen PNP dispersion is used as chemical admixture for cementi-tious materials instead of PCE.

Acknowledgement

The supports from the National Natural Science Foundation ofChina (Grant Nos. 51173094 and U1262107) and the CollaborativeInnovation Center for Advanced Civil Engineering Materials(Southeast University, Nanjing, 211189, China) are appreciated.

References

[1] AõÈtcin PC. Cements of yesterday and today: concrete of tomorrow. Cem ConcrRes 2000;30:1349–59.

[2] Ohama Y. Polymer-based admixtures. Cem Concr Compos 1998;20:189–212.[3] Giraudeau C, Jean-Baptiste d’Espinose de Lacaillerie, Souguir Z. Surface and

intercalation chemistry of polycarboxylate copolymers in cementitioussystems. J Am Ceram Soc 2009;92:2471–88.

[4] Miao CW, Ran QP, Hong JX, Xu J, Zhou DL. Present situation and developmentaltrends of polycarboxylate-type superplasticizers. Mater Chin 2009;28:36–45.

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

[6] Jenni A, Zurbriggen R, Holzer L, Herwegh M. Changes in microstructures andphysical properties of polymer-modified mortars during wet storage. CemConcr Res 2006;36:79–90.

[7] Beeldens A, Monteny J, Vincke E, Belied ND, Van Gemerta D, et al. Resistance tobiogenic sulphuric acid corrosion of polymer-modified mortars. Cem ConcrCompos 2001;23:47–56.

X. Kong et al. / Construction and Building Materials 68 (2014) 434–443 443

[8] Wang R, Wang PM, Li XG. Physical and mechanical properties of styrene–butadiene rubber emulsion modified cement mortars. Cem Concr Res 2005;35:900–6.

[9] Puterman M, Malorny W. Some doubts and ideas in the microstructureformation of PCC. In: Proceedings of IX international congress on polymers inconcrete, Bologna; 1998.

[10] Houst YF, Bowen P, Perche F. Design and function of novel superplasticizers formore durable high performance concrete (superplast project). Cem Concr Res2008;38:1197–209.

[11] Plank J, Bian H. Method to assess the quality of casein used as superplasticizerin self-levelling compounds. Cem Concr Res 2010;40:710–5.

[12] Yu B, Bian H, Plank J. Self-assembly and characterization of Ca-Al-LDHnanohybrids containing casein proteins as guest anions. J Physchem Sol2010;71:468–72.

[13] Bian H, Plank J. Re-association behavior of casein submicelles in highly alkalineenvironments. ZNaturforsch 2012;67b:621–30.

[14] Plank J, Gretz M. Study on the interaction between anionic and cationic latexparticles and Portland cement. Colloids Surf A: Physicochem Eng Asp 2008;330:227–33.

[15] Shi HX, Kong XM, Fan DK. Study on the microstructure and the properties ofstyrene-acrylic emulsion modified concrete. Chin Concr 2010;09:60–5.

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

[17] Boutti S, Urvoy M, Dubois-Brugger I. Influence of low fractions of styrene/butylacrylate polymer latexes on some properties of ordinary Portland cementmortars. Macromol Mater Eng 2007;292:33–45.

[18] Zhang YR, Kong XM, Hou SS, Liu YL, Han S. Study on the rheological propertiesof fresh cement asphalt paste. Constr Build Mater 2012;27:534–44.

[19] Dieter U, Koichi T. Polymer dispersions and their industrial applications.Weinheim: Wiley-VCH Verlag GmbH; 2002.

[20] Dukhin AS, Goetz PJ. Characterization of liquids, nano- and microparticulates,and porous bodies using ultrasound. 2nd ed. Oxford: Elsevier; 2010.

[21] Van Herk AM. Chemistry and technology of emulsion polymerization. JohnWiley & Sons; 2013.

[22] Wang ZM, Lu ZC, Lu F, Li HQ. Effect of side chain density of comb-shapedstructure on performance of polycarboxylate superplasticizer. J Chin CeramSoc 2012;11:1570–5.

[23] Winnefeld F, Becker S, Pakusch J, Götz T. Effects of the molecular architectureof comb-shaped superplasticizers on their performance in cementitioussystems. Cem Concr Compos 2007;29:251–62.

[24] Jayasree C, Gettu R. Experimental study of the flow behaviour ofsuperplasticized cement paste. Mater Struct 2008;41:1581–93.

[25] Mollah MYA. Review of cement superplasticizer interactions and their models.Adv Cem Res 2000;4:153–61.

[26] Zingg A, Winnefeld F, Holzer L, Pakusch J, Becker S, et al. Adsorption ofpolyelectrolytes and its influence on the rheology, zeta potential, andmicrostructure of various cement and hydrate phases. J Colloid Interface Sci2008;323:301–12.

[27] Yoshioka K, Tazawa E, Kawai K, Enohata T. Adsorption characteristics ofsuperplasticizers on cement component minerals. Cem Concr Res 2002;32:1507–13.

[28] Shin JY, Hong JS, Suh JK. Effects of polycarboxylate-type superplasticizer onfluidity and hydration behavior of cement paste. Korean J Chem Eng2008;25:1553–61.

[29] Rieger J, Thieme J, Schmidt C. Study of precipitation reactions by X-raymicroscopy: CaCO3 precipitation and the effect of polycarboxylates. Langmuir2000;16:8300–5.

[30] Puertas F, Santos H, Palacios M, Martínez-Ramírez S. Polycarboxylatesuperplasticiser admixtures: effect on hydration, microstructure andrheological behaviour in cement pastes. Adv Cem Res 2005;17:77–89.

[31] Ribeiro AB, Goncalves A, Carrajola A. Effect of shrinkage reducing admixtureson the pore structure properties of mortars. Mater Struct 2006;39:179–87.

[32] Zhang YR, Kong XM. Influences of superplasticizer, polymer latexes andasphalt emulsions on the pore structure and impermeability of hardenedcementitious materials. Constr Build Mater 2014;58:392–402.

[33] 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;36:2049–53.

[34] VocÏka R, Galle ÂC, Dubois M, Lovera P. Mercury intrusion porosimetry andhierarchical structure of cement pastes theory and experiment. Cem Concr Res2000;30:521–7.

[35] Metha PK, Monteiro PJM. Concrete: microstructure, properties, and materials.third ed. New York: McGraw-Hill; 2006.