ze

P. S116vans

Received 13 May 2014

Available online 23 May 2015

Keywords:Nano-kaolinite clayConcreteFreezethawChloride permeability

effeior of concrete. In our experiments, we substituted NKC for 0%, 1%, 3%, and 5% of mixtures of ordinary

freezethaw Cabinet was then used to measure the resistance of ordinary Portland cement concrete,as opposed to the concrete/NKC mixture, to examine deterioration caused by repeated FT actions. We

concrete depends on the structure of the material, namely itsporosity, the size of its pores and capillaries, their distributionacross the material, and the type of pores (open or closed) [26].To protect concrete from FT damage, a number of researchershave studied the factors affecting the material performance of con-crete exposed to FT actions. Experimental data from both

omaterialsrol calcium

ing [1935]. The results show that adding kaolinite claymass of cement) is effective way to improve the chloride resof concrete [3640]. Research indicates that the introductio3% nanoclay results in even higher compressive strength, lowerpermeability, and higher acid resistance within concrete struc-tures. It is anticipated that using nanoclay may increase the FTresistance of cement concrete. Despite recent attention to the per-formance of nanoparticle additions in cementitious materials, littleinformation exists related to the behavior of nanoclay-modiedconcrete, when exposed to FT actions.

Corresponding author.E-mail address: fanyf72@aliyun.com (Y. Fan).

Cement & Concrete Composites 62 (2015) 112

Contents lists availab

Cement & Concre

ev60 years [1].Concrete is inherently a porous material. The FT resistance of

eral studies reported that the addition of nanreduce concretes permeability to uids and conthttp://dx.doi.org/10.1016/j.cemconcomp.2015.05.0010958-9465/ 2015 Elsevier Ltd. All rights reserved.rs, sev-couldleach-(1% byistancen of 1Worldwide, innumerable concrete structures are in need ofrepair, due to exposure to severe winter conditions. Freezethaw(FT) durability of concrete is a crucial factor that affects the dura-bility of these concrete structures in cold regions. Many theoriesrelated to the action of frost on concrete (including osmotic pres-sure theory, classic hydraulic pressure theory, crystallization pres-sure theory, and others) have been put forward over the last

resistance [7]. Therefore, the life of concrete structures will begreatly increased once less permeable concrete is produced.Some researches reported that the addition of additives, such asy ash, silica fume, ground granulated blast furnace slag, rice huskash, and polypropylene bers, in concrete can improve both itspermeability and freezethaw resistance [815]. However, somestudies gave the contradictory results [1618]. In recent yeaRelative dynamic modulus of elasticityPoresElectrical resistivity

1. Introductionregularly measured the properties of the concrete specimens, including the pore structure, mass, electri-cal resistivity, chloride diffusion coefcient, compressive strength and dynamic modulus of elasticity. Acomputed tomography scan test evaluated the porosity characteristics of the concrete. This paper alsoapplied scanning electron microscopy and X-ray diffraction tests in order to investigate the micro mor-phology and chemical element distributions inside of the concrete. The experimental results and visualcomparisons revealed that the introduction of NKC improves the FT resistivity values, as compared tothe control concrete. The samples with 5% NKC exhibited the highest compressive strength, chloride dif-fusion resistivity, relative dynamic modulus of elasticity, and the most electrical resistivity after 125 FTcycles. We designated the anti-freezing durability coefcient (DF) as the index to assess the FT resistiv-ity of concrete. The following research discusses the relationship between the concretes DF and the num-ber of FT cycles, compressive strength, chloride diffusion coefcient, and the electrical resistivity of theconcrete samples.

2015 Elsevier Ltd. All rights reserved.

laboratory and eld tests has shown that well-distributed air voidscan provide pressure release and improve freezing and thawingReceived in revised form 30 March 2015Accepted 2 May 2015

Portland, cement, by weight. The blended concrete was prepared using w/c ratio as 0.5. A rapidEffects of nano-kaolinite clay on the free

Yingfang Fan a,, Shiyi Zhang a, Qi Wang a, Surendraa Institute of Road and Bridge Engineering, Dalian Maritime University, Dalian, LiaoningbCivil and Environmental Engineering, Northwestern University, 2145 Sheridan Road, E

a r t i c l e i n f o

Article history:

a b s t r a c t

This paper investigates the

journal homepage: www.elsthaw resistance of concrete

hah b

026, Chinaton, IL 60208, USA

cts of nano-kaolinite clay (NKC) on the freezing and thawing (FT) behav-

le at ScienceDirect

te Composites

ier .com/locate /cemconcomp

The objective of the study is to analyze the effects ofnanokaolinite clay (NKC) in concrete subjected to FT actions.We conducted experiments to determine the FT durability of con-

ture and contains silicon, whose theoretical formula is Al2Si2O5(OH)4 [41]. The chemical compositions of the cement and clay

dispersed in water using an ultrasonic dispersion method [37].Then, the dispersed clays were mixed with ne and coarse aggre-

in Fig. 2(b).

To better understand the effects of NKC on the microstructure

Furthermore, a XRD analysis was conducted to discover the

within the concrete. This investigation uses a Siemens somatomsensation 16-slice spiral computed tomography scanner, whichwas made in Germany. The pore characteristics in the concretesamples (with the dimensions of 100 mm 100 mm 100 mm)were examined using a CT test after exposure to 0, 50, and100 FT cycles. The samples were scanned with a xed X-raysource, at 140 kV, 200 mA, and 22.60 mGy CTD. The samples were

2 Y. Fan et al. / Cement & Concretegate following the JTG E30-2005 [42]. The mixed concrete was thenpoured into oiled molds to form prisms, sized100 mm 100 mm 400 mm, which were used for the FT tests.The fabricated samples were demolded after 24 h and were thencured using standard curing conditions (the temperature is20 C 2 C and the relative humidity is over 95% RH). Three spec-imens were created in the control group (ordinary Portland cementconcrete with no clay added) and the clay-modied concrete. Toinvestigate the effects of clay additives on concretes propertieswe studied clay-modied concrete mixtures with 1%, 3%, and 5%NKC by mass. The control concrete specimen and the concrete con-taining 1%, 3%, and 5% NKC additives were denoted as NC0, NC1,NC3, and NC5, respectively.

Table 1Chemical composition of cement.are listed in Table 1. The properties of NKC used in this study arelisted in Table 2. To characterize the chemical and microstructureof the NKC studied in this paper, X-ray diffraction (XRD) analysisand Transmission electron microscopy (TEM) techniques were car-ried out on the clay powder. The resulting TEM and XRD images ofthe clay powder samples are shown in Fig. 1. Based on the micro-graph of NKC powder, the particle size distribution of the powder isanalyzed and plotted in Fig. 1c. We obtained the elemental compo-sition of the NKC power samples from the EDS spectra, as listed inTable 3.

2.2. Specimen preparation

The mass ratio of cement: water: sand: aggregate in this studysconcrete mixture is 350: 175: 619: 1256, respectively. The effectivedispersion of NKC throughout the concrete mixture is critical toachieving full benets. Based on previous studies, the NKC was rstcrete containing 1%, 3%, and 5% NKC, using the rapid FT method.Tests on physical pore characteristic, mechanical property, perme-ability, and electrical conductivity were conducted on concretewith and without NKC particles that were subjected to FT cycles.Furthermore, we characterized the microstructure of cement con-crete using SEM, EDS, and XRD in order to evaluate the effects ofNKC particles on the improved strength and decreased permeabil-ity of concrete. By knowing the concretes base behavior during thefreezing and thawing process, it is possible to identify the benetscaused by the addition of NKC particles. These results can help tobetter inform architectural designs and maintenance for concretestructures, by taking the FT durability of NKC-modied concreteinto consideration.

2. Experimental study

2.1. Material properties

This study utilizes ordinary Portland cement, type 42.5R, andcommercially available NKC powder. NKC has a crystalline struc-Chemical components CaO SiO2 Al2O3 Fe2O3 MgO SO3

Content (%) 59.30 21.91 6.27 3.78 1.64 2.41effects of the NKC addition to the growth of crystals in the con-crete. The cubic samples with the dimensions of1 cm 1 cm 1 cm were extracted from the central part of theconcrete specimen. The prepared samples were examined using aRigaku D/max-Ultima+ Powder XRD system. A measurable2h-range is from 10 to 90, and the scan rate is 4/min.

2.3.3. Pore characteristicIt is known that the FT durability of concrete has close rela-

tionship with its pore structure. Within a certain temperatureinterval, concrete that contains more frozen pores induces greaterinternal hydraulic pressure and, consequently, more severe frostdamage [6]. Therefore, in studying the behavior of NKC-modiedconcrete subjected to FT actions, it is critical to assess the poresof the concrete samples, we conducted a microstructural morphol-ogy and elemental composition analysis on the plain concrete andNKC-modied concrete samples, using scanning electron micro-scopy (SEM) and energy dispersive spectroscopy (EDS). Weextracted an ordinary Portland cement-based concrete specimen,with the dimension of 5 mm 5 mm 5 mm, from the centralpart of the concrete specimen. The prepared samples were thenobserved using the JSM-6360LV SEM system, and the intensity ofthe applied voltage is 20 kV. To make the samples conductive,the surface of the samples was coated with 10 nm thick gold. Thesame equipment was used to determine the morphology of theNKC-modied concrete, in order to compare it to the control con-crete. 20 images were taken per samples.

2.3.2. X-ray power diffractionAt regular intervals of 25 FT cycles, the samples are removedfrom the apparatus. We allowed the removed specimens to dryon the surface, and then performed physical, dynamic modulus,electrical resistivity, compressive strength, CT, SEM/EDS, and XRDtests. We compared our experimental results in order to explorethe attributions of NKC on the properties of concrete.

2.3.1. SEM/EDS2.3. Methods

The Rapid FreezeThaw Cabinet (see Fig. 2a), which satises theGBJ82-85 procedure requirements, was used to produce FT cyclesin water [43]. The FT cycle consisted of alternatively lowering thetemperature of the specimens from 4 to 18 C and raising it from18 to 4 C in 3 h. The temperature curve of the FT cycle is shown

Table 2Physical index of nano-kaolinite clay.

Average akediameter (nm)

Average akethickness (nm)

Specic surfacearea (m2/g)

Density(g/cm3)

300500 2050 30 0.6

Composites 62 (2015) 112scanned at 1 mm spacing, and 100 slices were obtained. The outputscanning section of concrete is documented using the DCM format,and the CT number is stored in 12 digital capacities. The obtained

(a) (b)

20 40 60 800

500

1000

1500

2000

2500

Fe2O3

Al2O3

Al2O3

SiO2

Inte

nsity

2

SiO2

40

Y. Fan et al. / Cement & Concrete Composites 62 (2015) 112 330 / %50image is of 512 512 pixels, and anything above 100 lm in theconcrete can be identied.

2.3.4. Mass lossFor each specimen, the mass was recorded using an electronic

scale with an accuracy of 0.1 g during the FT actions.

(c)

0 200 400 600 800 1000 1200 1400 16000

10

20Con

tent

Partical size / nm

Fig. 1. TEM and XRD spectra of neat nano-kaolinite powder. (a) TEM

(a)

3

Fig. 2. Schematic diagram of rapid freezethaw testing equipment and temperatureThermocouples; 4 Water; 5 Antifreeze uid; 6 Rubber box).

Table 3Chemical composition of nano-kaolinite clay.

Chemicalcomponents

SiO2 CaO Al2O3 Fe2O3 MgO K2O TiO2 Na2O

Content (%) 47.80 0.28 41.80 0.30 0.03 0.58 0.02 0.062.3.5. Chloride diffusion coefcient (Dcl)

We used the rapid chloride migration (RCM) method, rst pro-posed by Tang and Nilsson [44], to determine the chloride diffusionproperty of the concrete samples. A migration cell is set up with aspecimen 50 mm thick and 100 mm in diameter. The testing proce-dure and calculation method of the chloride diffusion coefcienthave been described in full detail in our previous study [45].

2.3.6. Electrical resistivityMany factors may inuence the electrical resistivity of concrete,

including water/cement ratio, cement type, pozzolanic admixtures,degree of hydration, porosity, the moisture content, the composi-tion of the pore solution, pore size, transport property and

micrograph, (b) XRD results and (c) particle size distribution.

(b)

change curve. (1 Specimen; 2 Specimen for temperature measurement; 3

reteNNC0

4 Y. Fan et al. / Cement & Concconnectivity [46]. Evaluating concrete properties is possible withelectrical resistivity, and the electrical resistivity can be used forcondition surveying of concrete structures [47]. It is revealed thata strong relationship existed between chloride diffusivity and elec-trical resistivity [48]. To investigate the transport property of theconcrete experiencing repeated FT cycles, the electrical propertiesof the concrete samples was examined. A four-point probeResitest-400 type resistivity meter was applied to measure theelectrical resistivity of the concrete specimens.

2.3.7. Compressive strengthThe compressive strength of the concrete samples was mea-

sured following JTG E30-2005 [31] using the 2000 kNelectro-hydraulic servo compressive testing system (YAW-YAW2000A). The loading rate on the concrete samples was0.5 MPa/s. For each mixture at each age, three samples per batch

NC3

Fig. 3. SEM image of the concrete sam

10 20 30 40 50 602-theta (degrees)

Ca(OH)2

NC0

NC3

NC5

NC1

C-S-H

C-S-H

C-S-H

C-S-H

Ca(OH)2

Ca(OH)2

Ca(OH)2

Fig. 4. X-ray diffraction patterns of concrete samples.were tested; the average value was taken to be the representativestrength.

2.3.8. Dynamic modulus of elasticity (E)

NC1

NC5

ples before freezethaw actions.

Composites 62 (2015) 112Based on the theory of resonance, the base frequency of the con-crete specimens before and after various FT cycles was measuredusing a model DT-12 testing instrument. The output frequency ofthe instrument is between 0.1 and 2 kHz. Three samples per batchwere tested and the average base frequency was reported. The rel-ative dynamic modulus (E/E0) was then calculated by Eq. (1) toassess the internal damage of the concrete samples during FTcycles.

P f2n

f 20 100 1

where E/E0 is the relative dynamic elastic modulus of the specimenafter FT cycles for N times, %; fn is the base frequency of the spec-imen after FT cycles for N times, Hz; f0 is the base frequency of thespecimen before FT cycles, Hz.

3. Results and discussions

3.1. Microstructure of the mortars before FT actions

3.1.1. SEMFig. 3 shows the SEM micrographs of the concrete specimen. In

the gure, we can distinguish CSH gel, Ca(OH)2 crystals, andmicrocracks found in the concrete.

From the SEM images of the NC0 samples, we observed that,during the cement hydration process, the Ca(OH)2 crystals occupythe voids between the cement clinkers, the growth of the CSH islimited, and the microstructure becomes porous, which lessens thedurability of the concrete. It is revealed that the hydrate productswere denser, more compact, and had a more uniform microstruc-ture in the samples with the NKC additives. Therefore, the SEM

rete100 Table 4Failure behavior of concrete corroded by acid solution.

Cycles NKC additives0% 1% 3%

0

5

25

Y. Fan et al. / Cement & Concresults conrmed that evenly dispersing clay in the cement pasteimproves the cements porosity. In addition, the NKC activatesthe hydration process in the cement. Moreover, the more NKCadded, the more compact the concrete becomes.

3.1.2. XRDA XRD analysis was also undertaken on the concrete samples

with various NKC additions. The XRD diagrams for the controlspecimens and specimens containing 1%, 3%, and 5% NKC areshown in Fig. 4.

From the XRD results X-ray diffraction patterns of concrete withand without NKC, it can be observed that with the addition of NKC,the intensity of the X-ray diffraction peak for Ca(OH)2 crystals has aslight reduction, while the intensity of X-ray diffraction peak for CSH gel has a slight increase. This implies that the addition of NKCgenerally improves the hydration process of cement, because ahigher amount of CSH gel are produced in the hybrid clay/con-crete specimen, which results in a denser and more compactmicrostructure within the concrete. Although it is hard to deter-mine the difference quantitatively through X-ray diffraction pat-terns, the qualitative result coincides with the SEM analysis. It isalso noted that a 5% NKC addition shows a slightly lower effectthan that of 3% NKC addition. Combined with the SEM imagesshown in Fig. 3, it can be deduced that the decrease may be attrib-uted to the poorer dispersion of NKC within the concrete. Betterresults can be achieved when the NKC is evenly dispersed through-out the concrete.

125 5%

Composites 62 (2015) 112 53.2. Macro-, meso- and microscopic appearance

3.2.1. Macro-appearanceThe surface of the concrete was observed at regular intervals

during the FT actions. Using the Canadian StandardCAN3-A231.2-M85 [1], the surface conditions of the concrete sam-ples after FT actions were plotted in Table 4. It can be seen thatafter ve FT cycles, surface spalling and aggregate exposed wereobserved on NC0 and NC1 samples and the surface became moreporous, however, no damage was observed on NC3 and NC5 sam-ples. After 25 FT cycles, both NC0 and NC1 samples displayed sev-ere scaling and microcracking, while light scaling was observed onNC3 and NC5 samples. After 100 FT cycles, both NC0 and NC1samples showed extensive cracking, while moderate scaling andsome coarse aggregate manifested in NC3 and NC5 samples. After125 FT cycles, concrete on the edge of the NC0 and NC1 sampleswere broken; while the NC3 and NC5 samples became more por-ous, showing microcracking, and the damage state of NC3 is a littlemore severe than that of NC5.

From the surface condition of the concrete samples observedduring the FT cycles, the damage evolution was obtained. Dueto the repeated FT actions, the water penetrates into the concrete,and the swelling effect resulted in microcracks, which includeddamages including scaling and spalling. Since the addition ofNKC leads to lower permeability, the samples with NKC sustainedless surface damage, compared to the control specimen during theFT cycles. The concrete samples with more NKC additives showeddecreased damage rates throughout the FT cycles.

rete6 Y. Fan et al. / Cement & Conc3.2.2. Pore characteristic3.2.2.1. CT digital image. Fig. 5 provides the CT scanning images ofconcrete before and after 50 and 100 FT cycles. In this gure,we can distinguish the matrix, aggregate, pores, and microcracksin the concrete.

From Fig. 5, it is clear that the darker regions become wider inthe specimen after exposure to 50 FT cycles, which means thatthe voids in the specimen become larger. Consequently, the densityof the concrete specimen decreases with additional FT cycles. Up

(a) CT images of concrete samples before F-T cycles.

(b) CT images of concrete samples after 50 F-T cycles.

(c) CT images of concrete samples after 100 F-T cycles. NC0 NC1

cracks

cracks cracks

cracks

NC0 NC1

NC0 NC1

Fig. 5. CT scanning images of concrete subjected to FT cycles. (a) CT images of concrete(c) CT images of concrete samples after 100 FT cycles.

0 10 20 30 40 50 60 70 80 90 100

Slice number

0

0.5

1

1.5

2

2.5

3

3.5

Poro

sity

/ %

Poro

sity

/ %

0 cycleNC0NC1NC3NC5

0 10 20 30 40 5

Slice n

0

1

2

3

4

5

6

7

Fig. 6. Porosity distribution iComposites 62 (2015) 112until 100 FT cycles, several cracks appeared around the aggregate,and some cracks ran across the cement paste in both the NC0 andNC1 samples. However, no obvious cracks were observed in boththe NC3 and NC5 samples. This result provides an explanationfor the visual observations made of the specimen during the FTcycles.

3.2.2.2. Porosity evolution of concrete. To identify the pore distribu-tion inside of the concrete specimen, we executed a binary

NC3 NC5

NC3 NC5

NC3 NC5

samples before FT cycles, (b) CT images of concrete samples after 50 FT cycles and

Poro

sity

/ %

0 60 70 80 90 100

umber

50 cyclesNC0NC1NC3NC5

0 10 20 30 40 50 60 70 80 90

Slice number

0

5

10

15

20100 cycles

NC0NC1NC3NC5

n the concrete samples.

images. Fig. 6 presents the results of the porosity of 100two-dimensional slice images for each concrete specimen fromCT, before and after 50 and 100 FT cycles, respectively.

Based on the theory of probabilistic statistics, the probabilisticcharacteristics of porosity distribution inside of the concrete spec-imens were studied further. Fig. 7 exhibits the porosity distributioninside of the concrete specimens before and after FT cycles.

From Fig. 7, it is obvious that the distribution of the porosityinside the concrete specimens obeys normal distribution.Therefore, the distribution density function of the porosity in con-

0 0.5 1 1.5 2 2.5 3

Porosity / %

0

5

10

15

20

25

300 cycle

N0N1N3N5

0 1 2 3 4 5 6

Porosity / %

0

5

10

15

20

2550 cycles

N0N1N3N5

0 2 4 6 8 10 12 14 16

Porosity / %

0

2

4

6

8

10

12100 cycles

N0N1N5N3

Fig. 7. Probabilistic analysis of the porosity in the concrete samples exposed to freezethaw cycles.

4

6

8

10

12

14

rosi

ty,m

ean

valu

e/ %

0 cycle50 cycles100 cycles

Y. Fan et al. / Cement & Concrete Composites 62 (2015) 112 70 1 2 3 4 5Nanoclay additive / %

0

2

Poprocessing when scanning the digital images. The digital imageswere rst processed using a programming calculation withPro-Plus, and to identify the pores inside of the concrete. We devel-oped a pore recognition graph of the concrete specimens beforeand after the FT cycles, before computing the porosity of the slice

Fig. 8. Relation between the average value of porosity and nanoclay additivesduring FT cycles.

NC0

NC3

Microcrack

Fig. 9. SEM images of concrete screte specimens can be expressed as,

pq 12p

p exl2

2r2 0 < q

25 50 75 100 125Freeze-thaw cycles

-1

0

1

2

3

4

5

6

Mas

s lo

ss /%

0%1%3%5%

8 Y. Fan et al. / Cement & Concretecreate a more porous microstructure in the concrete. For NC5 sam-ples, the porosity increased 1.2 times after 100 FT cycles. In addi-tion, the mean value of the NKC-modied concretes porositydecreased linearly with the amount of NKC additives after sustain-ing 50 and 100 FT cycles.

3.2.3. Micro-appearanceFig. 9 shows the SEM micrographs of the concrete samples after

30 FT cycles. We observed extensive cracking in the ordinary con-crete after 30 FT cycles. Meanwhile, the NC1 concrete sampleexhibits a denser microstructure and less extensive cracks thanthe NC0 sample. On the contrary, no crack was observed in bothNC3 and NC5 concrete samples.

3.3. Mass loss

Based on the visual observations described above, it was obvi-ous that honeycomb voids were formed as FT cycles continued,causing the mass to change. The mass for all of the specimens dur-ing the FT cycles were obtained. The damage sustained by theconcrete specimens subjected to the FT cycles was evaluatedusing a mass loss ratio, which is dened as,

Dc 1mcnmc0

100% 3

Fig. 10. Mass loss of concrete versus FT cycles.where Dc is the mass loss ratio of the specimen after suffering FTcycles; mcn is the mass of the specimen after exposure to the NthFT cycles; and mc0 is the mass of the specimen before the FT test.

0 25 50 75Freeze-thaw cycles

0

1E-011

2E-011

3E-011

4E-011

5E-011

6E-011

7E-011

8E-011

Chl

orid

e di

ffusi

on c

oeffi

cien

t / m

2 /s

0%1%3%5%

Fig. 11. Chloride diffusion coefcient of concrete during the FT cycles.The relation between the mass loss ratio and FT cycles is plotted inFig. 10.

It is illustrated that the mass loss goes up gradually for the NC0,NC1, and NC5 concrete specimens as the FT cycles continues.However, the concrete specimens with 3% NKC experienced aslight mass gain throughout the FT actions. After 125 FT cycles,the highest mass loss is 5.13% for NC0, followed by 2.83% forNC1, and 0.85% for NC5. However, the mass gain of the NC3 is0.51% after 125 FT cycles. This may be attributed to a densermicrostructure in the NKC, due to the smaller size of NKC particles,which may lead to less damage and less mass loss within the con-crete. The evolution of the mass corroborates the results providedfrom the porosity results.

3.4. Chloride diffusion property

The FT damage inuenced the transport properties and accel-erated the deterioration of the concrete. Fig. 11 summarizes theresults of the chloride diffusion coefcient for the control andNKC-modied concrete specimens after 25 FT cycles.

As indicated in Fig. 11, the chloride diffusion coefcient of NC0,NC1, and NC3 increased throughout the FT tests. The control sam-ples chloride diffusion coefcient increased by 200% after 75 FTcycles, as compared to the value before the FT tests. As expected,the addition of NKC does improve the chloride resistivity duringthe FT actions. Yet, the chloride diffusion coefcient of the speci-

0 25 50 75 100 125Freeze-thaw cycles

0

50

100

150

200

250

300

350

Elec

tric

al re

sisi

tivity

/ k

cm 0%1%3%5%

Fig. 12. Electrical resistivity of concrete exposed to FT cycles.

Composites 62 (2015) 112men with 5% NKC additive just decreased 26% after 75 FT cycles.The increase in chloride resistance can be attributed to the rene-ment of pore structures in the NKC-modied concrete. It wasobserved that the 5% NKC addition resulted in a lower chloride dif-fusion coefcient.

3.5. Effects of NKC on electrical resistivity in concrete

It is well known that the corrosion of steel rebar in concrete is achemo-electrical process in practical civil engineering. Concretewith a higher electrical resistivity is better able to resist electricalcurrents, which decreases the amount of corrosion of the steelrebar, thus creating a more durable concrete structure. Therefore,civil engineers always use concrete with a high electrical resistiv-ity, to withstand severe environments. Since any changes in thevolume fraction of the matrix will affect the electrical resistivityof concrete, we investigated the effect of clay on the electricalresistivity of concrete during the FT cycles. For each specimen,the electrical resistivity on ve locations was measured. The aver-age value of the electrical resistivity obtained from the three spec-imens in each group was taken as the representative value. Fig. 12

the electrical resistivity of concrete. During the FT actions, thedegree of improvement found in the electrical resistivity is propor-

5

25

30

35

40

45C

otr

engt

h /M

Pa

0%1%3%5%

Y. Fan et al. / Cement & Concretetional to the amount of NKC added. Therefore, the NKC additiveinuences the electrical properties of the composite. Also, it isnoticed that repeated FT actions have a negative effect on theelectrical resistivity of concrete. As the number of FT cyclesincrease, the electrical resistivity of the different mixturesdecreases. Moreover, the specimens containing 5% NKC demon-strated the best electrical resistivity throughout the FT tests, fol-lowed by the specimens containing 3% NKC. It is noted that theelectrical resistivity in the concrete samples containing 5% NKCachieved a 64% decrease, as compared to the control samples after75 FT cycles. After 125 FT cycles, this value decreased to 29%.The reduction of the electrical resistivity after repeated FT cyclesshows a good agreement of the apparent porosity increaseobtained by CT test.

3.6. Compressive strength

Fig. 13 shows the variations in compressive strength withdisplays the tested electrical resistivity of the different mixturesafter 25 FT cycles.

Fig. 12 indicates that the addition of NKC obviously improves

0 25 50 100 125Freeze-thaw cycles

0

Fig. 13. Compressive strength of control and NKC-modied concrete exposed to FT cycles.10

15

20

mpr

essi

ve srespect to the number of FT cycles. The mix with 3% and 5%NKC showed a lower decrease in compressive strength throughoutthe FT testing. The ordinary cement concrete displayed a decreaseof 46% in compressive strength after 125 FT cycles. In the case ofthe 5% NKC addition, this value decreased to 12%. The evolution ofthe strength corroborates the results provided from the porosityresults.

3.7. Dynamic modulus of elasticity

Changes in the dynamic modulus of the elasticity of the differ-ent mixture samples were measured at regular intervals up to

Table 5Dynamic modulus of concrete after various FT cycles (unit: GPa).

FT cycles NC0 NC1 NC3 NC5

25 38.19 37.54 40.34 37.9350 32.68 33.59 39.60 36.1375 30.20 33.32 38.39 33.54

100 30.20 31.70 38.18 33.01125 26.43 27.38 37.15 30.86125 FT cycles. Table 5 summarizes the test data on the dynamicmodulus of elasticity.

As seen in Table 6, the dynamic modulus was obviously affectedby 125 FT cycles. The mixtures with 3% and 5% NKC showed alower decrease in the dynamic modulus of elasticity throughoutthe FT testing. Comparing to the concrete before FT cycles, theordinary cement concrete displayed a decrease up to 31% in thedynamic modulus of elasticity at 125 FT cycles. Whereas, the con-crete specimen with the additives of 3% and 5% NKC showed 8%and 18% decrease in dynamic modulus of elasticity, respectively.

3.8. The relative dynamic modulus of elasticity E/E0

The relative dynamic modulus of elasticity is the ratio of thedynamic modulus when determined at a particular test interval,relative to the initial dynamic modulus at the start of the test.Fig. 14 shows the experimental data concerning the relativedynamic modulus of elasticity of all of the tested concrete speci-mens during the FT cycles. The relative dynamic modulus of elas-ticity data for all of the tested concrete specimens under various FT actions have been tted to the potential function of the numberof FT cycles.

The best ts for all of the four kinds of concrete specimens havebeen obtained with a unied equation; the correlation coefcientsare listed in Table 5.

E=E0 aNb 4where N is the number of FT cycles and a and b are the coefcientsthat are related to the NKC addition, which are shown in Table 6.

3.9. Anti-freezing durability of concrete

The anti-freezing durability of concrete with and without NKCwas assessed using the anti-freezing durability coefcient, whichcan be calculated as follows:

DF E=E0 N125

5

where DF is the anti-freezing durability coefcient of concrete; N isthe number of FT cycle; and E/E0 is the relative dynamic elasticmodulus of concrete after N times of FT cycle. Fig. 15 shows theexperimental data of the DF of all of the tested concrete specimensduring the FT cycles. It is indicated that concrete containing 3%NKC addition shows the best improvements for FT resistivity.

Fig. 16 shows the relation between the DF and porosity in the

Table 6Fitting parameters and correlation coefcients to Eq. (2).

Sample no. a b Correlation coefcient

NC0 192.67 0.20 0.94NC1 162.71 0.15 0.80NC3 121.00 0.05 0.92NC5 151.78 0.12 0.92

Composites 62 (2015) 112 9control specimen and specimens containing NKC, before and afterthe FT cycles. It is shown that a linear relation exists between theDF and the porosity value of the control and NKC-modied con-crete during the FT action. It implies that the porosity in the con-crete increases linearly with the anti-freezing durabilitycoefcient. In addition, a 5% NKC addition shows the lowest poros-ity during the FT actions.

Fig. 17 shows the relation between the DF and chloride diffu-sion coefcient of the control specimen and specimens containingNKC before and after the FT cycles. It shows that a linear relationexists between the DF and the chloride diffusion coefcient of thecontrol and NKC-modied concrete during the FT actions. It

rete0.95

1.05

ulus

/E/E

010 Y. Fan et al. / Cement & Concimplies that the chloride diffusion coefcient of the concretedecreases linearly along with the anti-freezing durability coef-cient. In addition, a 5% NKC addition shows the highest chlorideresistivity during the FT actions.

Fig. 18 shows the relation between the DF and compressivestrength of all of the tested concrete specimens before and afterFT cycles. It can be seen that a linear relation exists between

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9DF

0

2

4

6

8

10

12

14

NC0NC1NC3NC5

Poro

sity

mea

n va

lue

/ %

Fig. 16. DF versus porosity value in the control specimen and specimens containingNKC before and after the FT cycles.

20 40 60 80 100 120 140Freeze/thaw cycles / N

0.65

0.75

0.85

Rel

ativ

e dy

nam

ic m

od

0%1%3%5%

Fig. 14. Evolution of dynamic modulus of elasticity of concrete exposed to FTcycles.

0 20 40 60 80 100 120 140Freeze-thaw cycles /N

0

0.2

0.4

0.6

0.8

1

1.2

Ant

i-fre

ezin

g du

rabi

lity

coef

ficie

nt

NC0NC1NC3NC5

Fig. 15. Evolution of DF with the FT cycles.6E-011

7E-011

8E-011

9E-011NC0NC1NC3NC5

effic

ient

/ m

2 /s

Composites 62 (2015) 112the DF and the compressive strength of the concrete during theFT actions. In addition, the specimen containing a 3% NKC addi-tion shows the highest compressive strength before 50 FT actions,while the specimen with a 5% NKC addition shows the highestcompressive strength after 50 FT actions.

Fig. 19 shows the relation between the DF and electrical resis-tivity of all of the tested concrete specimens during the FT cycles.In Fig. 19, it shows that an exponential decrease relation exists

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7DF

1E-011

2E-011

3E-011

4E-011

5E-011

Chl

orid

e di

ffusi

on c

o

Fig. 17. DF versus porosity value in the control specimen and specimens containingNKC before and after the FT cycles.

0 0.2 0.4 0.6 0.8 1DF

0

10

20

30

40

50

Com

pres

sive

str

engt

h /M

Pa

NC0NC1NC3NC5

Fig. 18. DF versus compressive strength.

0 0.2 0.4 0.6 0.8 1DF

0

100

200

300

400NC0NC1NC3NC5

Elec

tric

al re

sisi

tivity

/ k

m

Fig. 19. DF versus electrical resistivity.

ity coefcient. It is noted that the relationship between the DF and

producing concrete with a high FT resistivity.

[30] Cwirzen A, Habermehl-Cwirzen K. The effect of carbon nano- and microberson strength and residual cumulative strain of mortars subjected to freeze

reteAcknowledgements

This research was nancially supported by the National Naturalelectrical resistivity for NC3 is very close to that of NC5.

4. Conclusions

In order to discover the effects of NKC on the behavior of con-crete exposed to freezethaw actions, experimental investigationswere performed in the laboratory. Concrete specimens containing0%, 1%, 3%, and 5% NKC were prepared for the rapid FT tests.We recorded the physical, mechanical, chloride diffusion, and elec-trical resistivity properties of the concrete specimens during the FT cycles. The effects of NKC on the property of concrete sustainingFT actions were analyzed frommacro-, meso-, and micro-levels indetail. Based on our experimental results, the following conclu-sions are presented:

(1) The XRD and SEM data indicates that the lling and activat-ing effects of NKC improved the durability of the concreteduring FT actions. Although the control specimens wereseverely damaged after 125 FT cycles, the samples contain-ing 3% and 5% NKC were still in good condition.

(2) It is found that the distribution of the porosity inside of theconcrete specimens obeys the normal distribution beforeand after FT test. The mean value of the porosity in the con-crete samples decreases linearly, in relation to the increasedamount of NKC additives during the FT tests.

(3) The concrete specimens with 3% and 5% NKC showed a lowerdecrease in compressive strength throughout the FT tests.After 125 FT cycles, a 5% NKC additive in concretesimproved the compressive strength by up to 34% controlsamples.

(4) Electrical resistivity of the concretes with NKC was higherthan the normal concrete before and after the repeated FT cycles. It is noted that the electrical resistivity in the con-crete samples with 5% NKC achieved a 64% increase, as com-pared to the control samples, after 75 FT cycles.

(5) Chloride diffusion coefcient of the concretes with NKC waslower than the normal concrete before and after therepeated FT cycles. It is noted that the chloride diffusioncoefcient in the concrete samples with 5% NKC achieved a59% reduction, as compared to the control samples, after75 FT cycles.

(6) A potential relationship exists between the E/E0 and thenumber of FT cycles for all of the concrete specimensexposed to FT actions. Results indicate that sustaining upto 125 FT cycles leads to about a 23% difference in thedynamic modulus of elasticity in the 3% NKC-modied con-crete and the control concrete.

(7) Our results provide the relationship between theanti-freezing durability coefcient in regards to the numberof FT cycles, compressive strength, chloride diffusion coef-cient, and the electrical resistivity of the concrete. It is con-cluded that adding 3% and 5% NKC may be benecial inbetween the DF and the electrical resistivity of the concrete duringthe FT action. In other words, the electrical resistivity of the con-crete exponentially decreases along with the anti-freezing durabil-

Y. Fan et al. / Cement & ConcScience Foundation of China (Grant No. 51178069), the NationalNatural Science Foundation of China (Grant No. 50708010), theProgram for New Century Excellent Talents in University (Grantthaw cycles. J Adv Concr Technol 2013;11:808.[31] Chang TP, Shih JY, Yang KM, Hsiao TC. Material properties of Portland cement

paste with nano-montmorillonite. J Mater Sci 2007;42:747887.[32] Garboczi EJ. Concrete nanoscience and nanotechnology: denitions andNo. NCET-11-0860), and the Liaoning Provincial Fund forDistinguished Young Scholars.

References

[1] Pigeon M, Pleau R. Durability of concrete in cold climate. London, UK: E & FNSpon; 1995.

[2] Rostasy FS, Weib R, Wiedemann G. Changes of pore structure of cement mortardue to temperature. Cem Concr Res 1980;10:15764.

[3] Kumara R, Bhattacharjee B. Porosity, pore size distribution and in situ strengthof concrete. Cem Concr Res 2003;33:15564.

[4] Powers TC, Helmuth RA. Theory of volume changes in hardened Portlandcement paste during freezing. Highway Res Board Bull 1953;32:28597.

[5] Nagrockiene D, Skripknas G, Girskas G. Predicting frost resistance of concretewith different coarse aggregate concentration by porosity parameters. MaterSci 2011;17:2037.

[6] Cai H, Liu X. Freezethaw durability of concrete: ice formation process inpores. Cem Concr Res 1998;28:12817.

[7] Marchand J, Pigeon M, Setzer M. Freezethaw durability of concrete. London,UK: E & FN Spon; 1997.

[8] Kuo WY, Huang JS, Lin CH. Effects of organo-modied montmorillonite onstrengths and permeability of cement mortars. Cem Concr Res2006;36:88695.

[9] Mardani-Aghabaglou A, Andi-akir , Ramyar K. Freezethaw resistance andtransport properties of high-volume y ash roller compacted concretedesigned by maximum density method. Cem Concr Compos 2013;37:25966.

[10] Ayub T, Shaq N, Khan SU. Durability of concrete with different mineraladmixtures: a review. Int J Civil, Archit Sci Eng 2013;7:1728.

[11] Chung C-W, Shon C-S, Kim Y-S. Chloride ion diffusivity of y ash and silicafume concretes exposed to freezethaw cycles. Constr Build Mater2010;24:173945.

[12] Deja J. Freezing and de-icing salt resistance of blast furnace slag concretes.Cem Concr Compos 2003;25:35761.

[13] Sobolev K. Mechano-chemical modication of cement with high volumes ofblast furnace slag. Cem Concr Compos 2005;27:84853.

[14] Olgun M. The effects and optimization of additives for expansive clays underfreezethaw conditions. Cold Reg Sci Technol 2013;93:3646.

[15] Richardson AE, Coventry KA, Wilkinson S. Freeze/thaw durability of concretewith synthetic bre additions. Cold Reg Sci Technol 2012;8384:4956.

[16] Bleszynski R, Hooton RD, Thomas MDA, Rogers CA. Durability of ternary blendconcrete with silica fume and blast-furnace slag: laboratory and outdoorexposure site studies. ACI Mater J 2002;99:499508.

[17] Chidiac SE, Panesar DK. Evolution of mechanical properties of concretecontaining ground granulated blast furnace slag and effects on the scalingresistance test at 28 days. Cement Concr Compos 2008;30:6371.

[18] Van den Heede P, Furniere J, De Belie N. Inuence of air entraining agents ondeicing salt scaling resistance and transport properties of high-volume y ashconcrete. Cem Concr Compos 2013;37:293303.

[19] Ji T. Preliminary study on the water permeability and microstructure ofconcrete incorporating nano-SiO2. Cem Concr Res 2005;35:19437.

[20] Cardenas HE, Struble LJ. Electrokinetic nanoparticle treatment of hardenedcement paste for reduction of permeability. J Mater Civ Eng 2006;18:55460.

[21] Gaitero JJ, Campillo I, Guerrero A. Reduction of the calcium leaching rate ofcement paste by addition of silica nanoparticles. Cem Concr Res2008;38:11128.

[22] Gaitero JJ, Campillo I, Mondal P, Shah SP. Small changes can make a greatdifference. Transport Res Rec: J Transport Res Board 2010;2141:15.

[23] Porro A, Dolado JS, Gaitero JJ, Manzano H. Nanotechnology and concreteconcepts and approach. Transport Res Rec: J Transport Res Board2010;2142:1279.

[24] Karagl F, Demirboga R, Kaygusuz MA, Yadollahi MM, Polat R. The inuence ofcalcium nitrate as antifreeze admixture on the compressive strength concreteexposed to low temperatures. Cold Reg Sci Technol 2013;89:305.

[25] Kucharczykova B, Kersner Z, Pospichal O, et al. Inuence of freezethaw cycleson fracture parameters values of lightweight concrete. Proc Eng2010;2:95966.

[26] Zhang M, Malhotra V. Characteristics of a thermally activated alumino-silicatepozzolanic material and its use in concrete. Cem Concr Res 1995;25:171325.

[27] Veras-Agullo J, Chozas-Ligero V, Portillo-Rico D, Garcia-Casas M, Gutierrez-Martinez A, Mieres-Royo J. Mortar and concrete reinforced withnanomaterials. Nanotechnol Construct 2009;3:3838.

[28] Sanchez F, Sobole K. Nanotechnology in concrete a review. Constr BuildMater 2010;24:206071.

[29] Behfarnia K, Salem N. The effects of nano-silica and nano-alumina on frostresistance of normal concrete. Constr Build Mater 2013;48:5804.

Composites 62 (2015) 112 11applications. Nanotechnol Construct 2009;3:818.[33] Gartner E, Macphee D. A physico-chemical basis for novel cementitious

binders. Cem Concr Res 2011;41:73649.

[34] Jennings HM, Bullard JW, Thomas JJ, Andrade JE, Chen JJ, Scherer GW.Characterization and modeling of pores and surfaces in cement paste:correlations to processing and properties. J Adv Construct Technol2008;1:529.

[35] Khan MI. Nanostructure and microstructure of cement concrete incorporatingmulticementitious composites. Transport Res Rec: J Transport Res Board2010;2141:217.

[36] Farzadnia N, Ali AAA, Demirboga R, Anwar MP. Effect of halloysite nanoclay onmechanical properties, thermal behavior and microstructure of cementmortars. Cem Concr Res 2013;48:97104.

[37] Fan Y, Zhang S, Kawashima S, Shah SP. Inuence of kaolinite clay on thechloride diffusion property of cement-based materials. Cem Concr Compos2014;45:11724.

[38] He X, Shi X. Chloride permeability and microstructure of Portland cementmortars incorporating nanomaterials. Transport Res Rec: J Transport Res Board2008;2070:1321.

[39] Morsy MS, Alsayed SH, Aqel M. Effect of nano-clay on mechanical propertiesand microstructure of ordinary Portland cement mortar. Int J Civil Environ Eng2010;10:237.

[40] Tregger N, Pakula M, Shah SP. Inuence of micro- and nanoclays on fresh stateof concrete. Transport Res Rec 2010;2141:6874.

[41] Varga G. The structure of kaolinite and metakaolinite. ptoanyag2007;59:69.

[42] Test methods of cement and concrete for highway engineering (JTGE30-2005).Beijing: China Communications Press; 2005.

[43] Test methods of concrete for long-term and durability (GBJ82-85). Beijing:China Building Industry Press; 1997.

[44] Tang L, Nilsson LO. Rapid determination of chloride diffusivity of concrete byapplying an electric eld. ACI Mater J 1992;49:4953.

[45] Fan Y, Luan H. Pore structure in concrete exposed to acid deposit. Constr BuildMater 2013;49:40716.

[46] Rajabipour F, Weiss WJ, Abraham DM. In situ electric conductivitymeasurements to assess moisture and ionic transport in concrete. In:Proceedings of the International RILEM Symposium (CD-ROM), NorthwesternUniversity, Evanston, IL, USA; March 2126 2004.

[47] Koleva DA, Copuroglu O, van Breugel K, Ye G, de Wit JHW. Electrical resistivityand microstructural properties of concrete materials in conditions of currentow. Cem Concr Compos 2008;38:73144.

[48] Sengul O. Use of electrical resistivity as an indicator for durability. Constr BuildMater 2014;73:43441.

12 Y. Fan et al. / Cement & Concrete Composites 62 (2015) 112

Effects of nNano-kaolinite cClay on the fFreezethaw rResistance of cConcrete1 Introduction2 Experimental study2.1 Material properties2.2 Specimen preparation2.3 Methods2.3.1 SEM/EDS2.3.2 X-ray power diffraction2.3.3 Pore characteristic2.3.4 Mass loss2.3.5 Chloride diffusion coefficient (Dcl)2.3.6 Electrical resistivity2.3.7 Compressive strength2.3.8 Dynamic modulus of elasticity (E)

3 Results and discussions3.1 Microstructure of the mortars before FT actions3.1.1 SEM3.1.2 XRD

3.2 Macro-, meso- and microscopic appearance3.2.1 Macro-appearance3.2.2 Pore characteristic3.2.2.1 CT digital image3.2.2.2 Porosity evolution of concrete

3.2.3 Micro-appearance

3.3 Mass loss3.4 Chloride diffusion property3.5 Effects of NKC on electrical resistivity in concrete3.6 Compressive strength3.7 Dynamic modulus of elasticity3.8 The relative dynamic modulus of elasticity E/E03.9 Anti-freezing durability of concrete

4 ConclusionsAcknowledgementsReferences