Non-destructive tracing on hydration feature of slag blended cement with electrochemical method

Construction and Building Materials 149 (2017) 467–473

Contents lists available at ScienceDirect

Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat

Non-destructive tracing on hydration feature of slag blended cement with electrochemical method Biqin Dong, Gui Li, Jianchao Zhang, Yuqing Liu, Feng Xing, Shuxian Hong ⇑ School of Civil Engineering, Guangdong Province Key Laboratory of Durability for Marine Civil Engineering, Shenzhen University, Shenzhen 518060, PR China

h i g h l i g h t s
Hydration behavior of slag blended cement is non-destructive traced by electrochemical impedance method.
A novel electrochemical equivalent circuit model is proposed and used to investigate the whole hydration procedure.
Relationship between electrochemical impedance parameter (Rct1) and compressive strength of slag blended cement is built.

a r t i c l e

i n f o

Article history: Received 10 February 2017 Received in revised form 20 April 2017 Accepted 5 May 2017

Keywords: Slag blended cement materials Equivalent circuit model Electrochemical impedance spectroscopy Hydration Compressive strength

a b s t r a c t This paper aims to use electrochemical impedance spectroscopy (EIS) with a novel equivalent circuit (EC) model to examine the hydration behavior of cement materials that incorporate ground granulated blast furnace slag. The experimental results suggest that the electrochemical impedance behavior of blended cement materials vary depending on the slag content. Also, the resistance associated with the ion transport process increases gradually along with the hydration process, and decreases as the slag is incorporated into the cement. In addition, we describe the linear relationship between the Rct1 value and the compressive strength for different slag contents and curing ages. The proposed method of EIS for slag blended cement materials, and the reliability of the new equivalent circuit model, is investigated across the entire hydration process of slag blended cement materials. Finally, the correlation of the compressive strength and the Rct1 value of blended cement materials with different slag contents are analyzed. Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction Supplementary materials (such as slag, fly ash, etc.), which are widely used in cement materials, have the advantage of saving construction costs and reducing environment impact [1–4]. Also, blended cement typically outperforms plain Portland cement. However, ground granulated blast furnace slag, which is the residue and solid waste produced during the steel manufacturing process, often seriously contaminates soil, ground water and oceans, and slag is one of the substitutes that is widely used in cement paste [3–6]. The application of slag dates back to the last century, and a great deal of research has been conducted on it over the past few decades. These studies have shown that granulated blast furnace slag improves the quality of blended cement materials in several ways: (1) improving its long-term strength, (2) reducing the risk of early-age thermal cracking by decreasing heat development and peak temperatures, (3) controlling alkali-aggregate reactions, ⇑ Corresponding author. E-mail address: [email protected] (S. Hong). http://dx.doi.org/10.1016/j.conbuildmat.2017.05.042 0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.

and (4) improving durability [1,7–11]. It is also known that slag is a latent material in hydraulic cement, it is not only rich in SiO2, CaO and Al2O3, but also composed mainly of vitreous body [1,8,9,12–15]. The reaction of slag is slower than the reaction of the clinker, because it is the alkalinity of the pore solution used that drives the dissolution of slag, note that the choice of pore solution determines the concentration of OH- released during the hydration of Portland cement [9,15]. Therefore, activators such as sodium hydroxide, sodium carbonates and sodium silicates have been most widely studied [9,16]. The reactivity of the slag increases at higher temperatures and decreases at higher levels of replacement. And the fineness and composition of granulated blast furnace slag can affect the hydration properties of slag blended cement materials [8–11,17–19]. Hydration, which is known to be a complex physicochemical process, is central to cement materials because it determines the microstructure and macro performance of these materials [20– 26]. Many studies have been done on the hydration of Portland cement using various test methods. Scanning electron microscopy (SEM) is often used to investigate the pore structure development

468

B. Dong et al. / Construction and Building Materials 149 (2017) 467–473

(XRD) [27] and Fourier Transform Infrared Spectrometry (FTIR) [21]. Some researchers monitor the heat released during the hydration of cement materials by calorimetric experiments [28,29]. However, none of these methods be capable of providing continuous testing, because the samples are damaged by the research methods used. Some non-destructive teat methods are used to study the hydration of cement-based paste, such as active acoustic method [30,31] and electrical resistivity method [32–34]. And in our research, a non-destructive steady-state technique method called electrochemical impedance spectroscopy (EIS) that is lowcost, highly-sensitive and convenient is used. It can be used to study the microstructure development that occurred during the hydration of cement materials; because cement materials can be considered to be an electrochemical system, the hydration of which is a complicated electrochemical procedure that involves ion transfer, rearrangement and distribution, so EIS can be applied for investigating the hydration process of blended cement materials [35–39]. The proposed approach also has the advantages of investigating the physicochemical changes of the blended cement materials under different conditions and in various service environments [40–43]. So the proposed method is a promising method to use to explore the microstructural properties of blended cement materials. The objective of this study is to investigate the electrochemical behavior of slag blended cement materials to investigate the hydration behavior of these materials using a new equivalent circuit (EC) model of electrochemical impedance spectroscopy. The influence of slag on the hydration of blended cement materials is analyzed using the electrochemical parameter Rct1 on samples with different slag contents and hydration periods. Then, the correlation of electrochemical parameters (Rct1) to compressive strength, fitted by the new EC (equivalent circuit) model is established, and the effect of different slag content on this correlation is analyzed.

Table 1 Chemical composition and Physical Properties of cement and slag. Chemical composition (Mass %)

Cement

Slag

Calcium oxide (CaO) Silica (SiO2) Alumina (Al2O3) Iron Oxide (Fe2O3) Magnesium oxide (MgO) Sulfur trioxide (SO3) Potassium oxide (K2O) Loss on ignition (LOI)

64.67 18.59 4.62 4.17 2.35 3.32 0.92 1.03

37.73 34.62 11.81 2.73 9.43 1.42 0.65 1.2

of cement-based paste and to collect morphological information about the hydration products [20–22]. Mercury intrusion porosimetry (MIP) is used for measuring pore structure feature of cement-based paste as well [23,24]. The hydration products are analyzed using quantitative methods, including X-ray diffraction

Fig. 1. The system for impedance spectroscopy measurement.

Q

Rs

Rct

W ZF

(a) Q1

Q2

Rct1

Rct2

Rs

(b) Fig. 2. The equivalent circuit model: (a) Randles model; (b) Gu et al.’ model. Rs: the resistance of the electrolyte solution; Q: the double layer capacitance between the electrodes and the electrolyte; Rct: the resistance caused by ion transfer; W: Warburg resistance caused by charge diffusion; ZF: the impedance of Faraday’s procedure that occurs on the surface of the electrodes; Q1: the double layer capacitance between the solid/liquid phases; Rct1: the resistance caused by ion transfer inside the cement sample; Q2: the double layer capacitance between the cement material and the electrodes; Rct2: the resistance caused by the charge transfer on the surface of the electrodes.

469

B. Dong et al. / Construction and Building Materials 149 (2017) 467–473 2. Materials and experiments

Z ¼ Rs þ

2.1. Experimental materials

Z F1 Z F2 þ 1 þ j-Z F1 C d1 1 þ j-Z F2 C d2 Rct1 þ r1 -2 ð1  jÞ 1

For all specimens, ordinary Portland cement (P.O 42.5), was obtained from the Onoda Cement Limited Company in Shenzhen, China, was used. Class I slag, obtained from the Mawan power plant in Shenzhen, was used as a supplementary cement material. The chemical compositions and physical properties of Portland cement and slag are presented in Table 1. Normal tap water was used, and no chemical admixtures were incorporated.

¼ Rs þ

1 þ j-Rct1 C d1 þ j-C d1 ðr1 -2  jr1 -2 Þ 1

1

Rct2 þ r2 -2 ð1  jÞ 1

þ

1 þ j-Rct2 C d2 þ j-C d2 ðr2 -2  jr2 -2 Þ 1

1

ð1Þ

where r1 is the electrical conductivity of cement-based materialsr2 is the electrical conductivity of electrode plates

2.2. Test procedure In this study, blend paste matrices were prepared that had ground granulated blast furnace slag contents of 0%, 10%, 20% and 30%. The water to binder ratio of the specimens was 0.3; they were cast in molds with the dimensions of 40 mm  40 mm  160 mm, and they were then stored in a curing chamber (95 ± 5% RH, 20 ± 2 °C) for 24 h. Finally, the specimens were extracted from the molds and then returned to the chamber for additional curing under the same environmental conditions until the test time. As shown in the Fig. 1, the system of electrochemical impedance measurement was used to measure the electrochemical impedance spectra for the blended cements with different hydration periods (1 h, 8 h, 1d, 3d, 5d, 7d and 28d). The tests were performed using a Potentiostat/Galvanostat 283 instrument (Input Signal: Single power sine; Electrode plates: Model 283 at address 14; Test Frequency: 0.01 Hz-1 MHz). According to GBT 17671-1999 (ISO method), the compressive strength tests were performed using a compression-testing instrument (YAW-300B) at various curing ages (1d, 3d, 5d, 7d and 28d).

- ¼ 2p f W 1 ¼ r1 -2 ð1  jÞ 1

W 2 ¼ r2 -2 ð1  jÞ 1

Cd1 is the double-layer capacitance of the model (Q1)Cd2 is the double-layer capacitance of the model (Q2) When ð- ! 1Þ; that is- >> ðRrct Þ2 ;

C d1 ¼ 1=ðW c Rct1 Þ; C d2 ¼ 1=ðW c Rct2 Þ; Equation can be rewritten as:

3. Result and analysis Electrochemical impedance spectroscopy is employed to study the hydration behavior of slag blended cement materials. There are two existing equivalent circuit (EC) models - the typical EC model (shown in Fig. 2(a)), which is described as Rs(Q(RctW)) [35], and Gu et al.’s EC model [36] (presented in Fig. 2(b)), which is described as Rs(Q1Rct1)(Q2Rct2). The typical equivalent circuit can be used to model the early stages of cement paste hydration, while another electrical EC model, Rs(Q1Rct1)(Q2Rct2) mainly monitors the later hydration of cement mortars that have less water. However, both methods fail to analyze the entire hydration process of blended cement materials [35,36]. Therefore, we propose a novel EC model, Rs(Q1(Rct1W1))(Q2(Rct2W2),) which is demonstrated for the investigation of the entire process of the hydration of blended cement(presented in Fig. 3). Deducing the impedance of this electrochemical equivalent circuit model has been studied [40–43], and is expressed by Eq. (1):

Electrode

Fig. 4. Nyquist curve of the proposed model for hydration procedure of cement materials.

Electrode

Cement Paste

Q1

Q2

Rs Rct1

W1 ZF1

Rct2

W2 ZF2

Fig. 3. The proposed equivalent circuit for hydration procedure of cement materials. W1: Warburg resistance caused by ion diffusion inside the cement sample; W2: Warburg resistance caused by the ion diffusion on the surface of the electrodes; ZF1: the Faraday impedance caused by the Faraday’s procedure inside the blended cement materials; ZF2: the Faraday impedance caused by the Faraday’s procedure between the electrodes and cement materials.

470

B. Dong et al. / Construction and Building Materials 149 (2017) 467–473

 2  2 Rct1 þ Rct2 Rct1 þ Rct2 Z 0  Rs  þ Z 002 ¼ 2 2

ð2Þ

This is an equation of a semi-circle (in the first quadrant) When ð- ! 0Þ; the following equation can be obtained:

Z 0 ¼ Z 00  2r21 C 1  2r22 C 2 þ Rs þ Rct1 þ Rct2

ð3Þ

This is a unit-slope straight-line equation. Based on Eqs. (2) and (3), the classical Nyquist curve in a complex-plane plot can be drawn as shown in Fig. 4. The new EC model is verified by fitting the Rs(Q(RctW)) and Rs(Q1Rct1)(Q2Rct2) models with Nyquist plots, Bode plots and phase angle plots of the slag blended cement materials (w/c = 0.3, slag content = 10%) at a hydration period of 5 days hydration, as shown

20000 Experimental points Rs(Q(RctW)) model curve fitting

Experimental points Rs(Q(RctW)) model curve fitting

4000

Rs(Q1Rct1)(Q2Rct2) model curve fitting

15000

Rs(Q1Rct1)(Q2Rct2) model curve fitting

Rs(Q1(Rct1W1))(Q2(Rct2W2)) model curve fitting

Rs(Q1(Rct1W1))(Q2(Rct2W2)) model curve fitting

10000

Z"(Ohm)

Z"(ohm)

3000

5000

2000

1000 0

0

5000

10000

15000

20000

0

25000

500

Z'(ohm)

1000

1500

2000

2500

Z'(Ohm)

(a)

(a)

30000 Experimental points Rs(Q(RctW)) model curve fitting Rs(Q1Rct1)(Q2Rct2) model curve fitting

Experimental points Rs(Q(RctW)) model curve fitting

12000

Rs(Q1Rct1)(Q2Rct2) model curve fitting

Rs(Q1(Rct1W1))(Q2(Rct2W2)) model curve fitting

Rs(Q1(Rct1W1))(Q2(Rct2W2)) model curve fitting

Z" (ohm)

|Z|(ohm)

20000

10000

8000

4000

0

0 0.1

1

10

100

1000

10000

100000 1000000

0

Frequency(Hz)

5000

10000

15000

20000

Z' (ohm)

(b)

(b)

50 Experimental points Rs(Q(RctW)) model curve fitting

40

Rs(Q1Rct1)(Q2Rct2) model curve fitting

12000

Rs(Q1(Rct1W1))(Q2(Rct2W2)) model curve fitting

Rs(Q1Rct1)(Q2Rct2) model curve fitting Rs(Q1(Rct1W1))(Q2(Rct2W2)) model curve fitting

30

Z" (ohm)

Phase angle(deg)

Experimental points Rs(Q(RctW)) model curve fitting

20

8000

4000

10

0

0 0.1

1

10

100

1000

10000

100000 1000000

0

4000

8000

12000

Frequency (Hz)

Z' (ohm)

(c)

(c)

Fig. 5. Nyquist plots (a), Bode plots (b) and Phase angle plots (c) of electrochemical impedance measurement for slag blended cement materials (w/c = 0.3, slag content = 10%) at 5 days hydration period.

16000

20000

Fig. 6. Nyquist curve of cement pastes (w/c = 0.3, slag content = 10%) at different hydration period: (a) Hydration time: 1 h; (b) Hydration time: 8 h; (c) Hydration time: 28 days.

471

B. Dong et al. / Construction and Building Materials 149 (2017) 467–473

addition, the reaction of the slag could suppress the reaction of the clinker phases due to competition for space. Thus, as the slag replacement increases, the clinker content decreases for samples at the same water to binder ratio. Also, the amount of calcium hydroxide decreases, and the decomposing reaction of the vitreous body of slag slows, because the pH of the pore solution cannot rise adequately. In this case, the degree of density of the microstructure of the slag blended cement materials is lower, making the ion transport process easier [8–9,15–18]. The electrical conduction of slag blended cement materials is conducted by the ions in the free water, and it takes place through the liquid phase as an indirect indicator of pore structure evolution [44,45]. Therefore, this article aims to examine the relationship between electrochemical parameters (Rct1) and compressive strength that is closely related to the pore structure (Fig. 8). Generally, the higher the porosity, the lower the compressive strength. The results show that the compressive strength and Rct1 exhibit similar growth tendencies. At the beginning of the tests, both the Rct1 value and the compressive strength increase sharply, because the hydration continues rapidly for several weeks once the cement materials come into contact with the water. The spaces in the paste fill with the hydration products, and the free water content decreases, eventually resulting in a decrease in porosity and total conductive ions. Then, the increase in the Rct1 value and the compressive strength slow, because there are fewer spaces in which the hydration products could form. Therefore, the volume of the solid phase increases as the pore volume decreased as the hydration progress, resulting in a denser microstructure. As the slag content increases, both the Rct1 value and compressive strength decrease, because the porosity increases and the degree of density of the microstructure is low [23–27]. The correlation between the compressive strength and Rct1 values of the cement-slag paste samples with different amounts of slag incorporation is analyzed (Fig. 9). All of the correlation coefficients, R2 are bigger than 0.98, indicating that compressive strength and Rct1 are significantly correlated. The equation

Experimental points

60

Rs(Q1(Rct1W1))(Q2(Rct2W2)) model curve fitting

55

1 2

Z"(ohm)

10000

3 4 5000

1: slag content=0% 2: slag content=10% 3: slag content=20% 4: slag content=30%

Compressive strength (MPa)

15000

0

3500 3000

50 45

2500

40 compressive strength curve (slag content=0%) compressive strength curve (slag content=10%) compressive strength curve (slag content=20%) compressive strength curve (slag content=30%) Rct1 curve (slag content=0%)

35 30 25

Rct1 curve (slag content=10%) Rct1 curve (slag content=30%)

500

15

0

5000

10000

1500 1000

Rct1 curve (slag content=20%)

20

2000

Rct1 (ohm)

in Fig. 5. It can be seen that the new EC model performs better than the Rs(Q(RctW)) and Rs(Q1Rct1)(Q2Rct2) models. The fitting results of the different models for slag blended cement at different hydration times(1 h, 8 h and 28d) are shown in Fig. 6, and they indicate that the Rs(Q1(Rct1W1))(Q2(Rct2W2)) model analyzes the entire process of mortar hydration more effectively. Thus, the new model Rs(Q1(Rct1W1))(Q2(Rct2W2)) is shown to be more suitable to investigate the entire process of the hydration of blended cement materials. The Rct1 values increase gradually during the hydration process, because the total conductive ions decrease. The volume of the solid phase also increases during hydration. Both of these results cause the microstructure to become denser, which in turn makes the ion transfer more difficult [18,23–26]. From the Nyquist curve and model fitting results for the cement-slag paste samples with different slag content at a hydration period of 3 days (Fig. 7), and the Rct1 values of the blended cement materials with different slag contents and hydration periods (1d, 3d, 5d, 7d, and 28d) (Table 2), it can be seen that the radius of the semicircle and the Rct1 value both decrease as the slag content increases across the entire process of hydration for samples that have the same hydration period. The reasons for these are that slag that is incorporated into the cement-based materials has a filler effect and can proceed to secondary hydration by means of the calcium hydroxide produced during the hydration of Portland cement. On the first day, the hydration process is dominated by the filler effect. There is more space available for the hydration products of the clinker phases for samples with the same water to binder ratio, and the slag particles fill the interstices of the cement particles and act as nucleation sites for the hydration products, accelerating the hydration of the cement. However, the clinker content decreases as the slag content increases, leading to the reduction of hydration. When the slag reaction begins at approximately 2–3 days, the reaction of the slag is slower than that of the clinker, and the dissolution of the slag vitreous body is activated by the increase of OH- produced by the clinker hydration. In

15000

0

200

400

600

800

Time (h)

Z'(Ohm) Fig. 7. Nyquist curve and model fitting for cement-slag paste with different slag content at 3 days hydration period.

Fig. 8. Comparison of compressive strength and Rct1 value for cement-slag paste with different slag content.

Table 2 Rct1 value of slag blended cement materials with different slag content and hydration period. Rct1

1d

3d

5d

7d

28d

0% 10% 20% 30%

1568 1192 819 538

2199 1707 1450 1172

2993 2492 2146 1865

3219 2818 2504 2298

3551 3258 2791 2487

472

B. Dong et al. / Construction and Building Materials 149 (2017) 467–473

Center for Advanced Civil Engineering Materials, Nanjing, P. R. China.

4000 slag content = 0% slag content = 10% slag content = 20% slag content = 30% y = 73.376x - 845.41,R2 = 0.9869 y = 74.084x - 775.15,R2 = 0.9891 y = 67.765x - 482.64,R2 = 0.9979 y = 65.137x - 407.19,R2 = 0.9923

Rct1 (Kohm)

3000

References

2000

1000

0 15

20

25

30

35

40

45

50

55

60

Compressive strength (MPa) Fig. 9. The correlation between the compressive strength and Rct1 values for cement-slag paste with different slag content.

‘‘y = ax + b” is used to express the relationship between compressive strength and Rct1 value, and the slope increases as the slag content increases, while the intercept decreases. Therefore, it is feasible to predict the 28th compressive strength of cement-slag blended materials in accordance with the fitting parameter Rct1 of the electrochemical impedance spectroscopy based on the new EC model. 4. Conclusions In this paper, the influence of slag incorporation on the hydration of cement materials was determined using a new EC model of EIS, and the curves of the slag blended cement materials’ impedance spectroscopy associated with the hydration process were analyzed. The following main conclusions can be drawn: 1) The radius of the semicircle in the classical Nyquist curve and the Rct1 value of the slag blended cement materials decreased as the slag content increased during the entire process of hydration. Because the clinker content decreased as the slag content increased, the amount of hydrates was also reduced. In addition, the reaction of the slag was slower than the reaction of the clinker, and the dissolution of slag vitreous body was prompted by the concentration of OHproduced by the clinker hydration. In the meantime, the reaction of the slag suppressed the reaction of the clinker phases, likely due to competition for space. 2) The electrochemical impedance spectroscopy parameter Rct1 showed a growth trend that was similar to that demonstrated by the compressive strength. The relationship between compressive strength and Rct1 is highly correlated, as shown by the correlation coefficient R2 in each case was greater than 0.98. Also, the slope of the equation increased as the slag content increased, while the intercept showed the opposite trend. Thus, the parameter Rct1 can be used to predict the 28 day compressive strength of the slag blended cement materials.

Acknowledgements The authors would like to acknowledge financial support provided by the National Natural Science Foundation of China (No. 51538007/U1301241/51478270) and the Collaborative Innovation

[1] B. Lothenbach, K. Scrivener, R.D. Hooton, Supplementary cementitious materials, Cem. Concr. Res. 41 (3) (2011) 217–229. [2] S. Kandasamy, M.H. Shehata, Durability of ternary blends containing high calcium fly ash and slag against sodium sulphate attack, Constr. Build. Mater. 53 (3) (2014) 267–272. [3] H. Ma, Multi-scale modeling of the microstructure and transport properties of contemporary concrete, Concr. Constr. (2013). [4] R. Taylor, I.G. Richardson, R. Brydson, Composition and microstructure of 20year-old ordinary Portland cement–ground granulated blast-furnace slag blends containing 0 to 100% slag, Cem. Concr. Res. 40 (40) (2010) 971–983. [5] E. Crossin, The greenhouse gas implications of using ground granulated blast furnace slag as a cement substitute, J. Clean. Prod. 95 (2015) 101–108. [6] R.P. Khatri, V. Sirivivatnanon, W. Gross, Effect of different supplementary cementitious materials on mechanical properties of high performance concrete, Cem. Concr. Res. 25 (1) (1995) 209–220. [7] T. Bakharev, J.G. Sanjayan, Y.B. Cheng, Resistance of alkali-activated slag concrete to alkali–aggregate reaction, Cem. Concr. Res. 31 (2001) 331–334. [8] X. Fu, W. Hou, C. Yang, D. Li, X. Wu, Studies on Portland cement with large amount of slag, Cem. Concr. Res. 30 (4) (2000) 645–649. [9] B.S. Gebregziabiher, R. Thomas, S. Peethamparan, Very early-age reaction kinetics and microstructural development in alkali-activated slag, Cem. Concr. Compos. 55 (2015) 91–102. [10] G.J. Osborne, Durability of Portland blast-furnace slag cement concrete, Cem. Concr. Compos. 21 (1) (1999) 11–21. [11] H. Toutanji, N. Delatte, S. Aggoun, R. Duval, A. Danson, Effect of supplementary cementitious materials on the compressive strength and durability of shortterm cured concrete, Cem. Concr. Res. 34 (2) (2004) 311–319. [12] J.I. Escalante, L.Y. Gómez, K.K. Johal, G. Mendozaa, H. Manchaa, J. Méndeza, Reactivity of blast-furnace slag in Portland cement blends hydrated under different conditions, Cem. Concr. Res. 31 (10) (2001) 1403–1409. [13] S. Song, H.M. Jennings, Pore solution chemistry of alkali-activated ground granulated blast-furnace slag 1, Cem. Concr. Res. 29 (2) (1999) 159–170. [14] M.B. Haha, B. Lothenbach, G.L. Saout, F. Winnefeld, Influence of slag chemistry on the hydration of alkali-activated blast-furnace slag — Part II: Effect of Al2O3, Cem. Concr. Res. 42 (1) (2012) 74–83. [15] B. Kolani, L. Buffo-Lacarrière, A. Sellier, G. Escadeillasa, L. Boutillonb, L. Lingerb, Hydration of slag-blended cements, Cem. Concr. Compos. 34 (9) (2012) 1009– 1018. [16] C. Shi, R.L. Day, Some factors affecting early hydration of alkali-slag cements, Cem. Concr. Res. 26 (3) (1996) 439–447. [17] C. Çetin, S.T. Erdog˘an, M. Tokyay, Effect of particle size and slag content on the early hydration of interground blended cements, Cem. Concr. Compos. 67 (2015) 39–49. [18] E. Berodier, K. Scrivener, Evolution of pore structure in blended systems, Cem. Concr. Res. 73 (2015) 25–35. [19] C.C. Castellano, V.L. Bonavetti, E.F. Irassar, Influence of curing temperature: hydration and strength of cement paste with granulated blast furnace, Matéria 15 (4) (2009) 516–526. [20] X. Feng, E.J. Garboczi, D.P. Bentz, P.E. Stutzman, T.O. Mason, Estimation of the degree of hydration of blended cement pastes by a scanning electron microscope point-counting procedure, Cem. Concr. Res. 34 (10) (2004) 1787–1793. [21] R. Ylmén, U. Jäglid, B.M. Steenari, I. Panas, Early hydration and setting of Portland cement monitored by IR, SEM and Vicat techniques, Cem. Concr. Res. 39 (5) (2009) 433–439. [22] B. Gaël, T. Christelle, E. Gilles, G. Sandrine, S.F. Tristan, Determination of the proportion of anhydrous cement using SEM image analysis, Constr. Build. Mater. 126 (2016) 157–164. [23] H. Ma, Mercury intrusion porosimetry in concrete technology: tips in measurement, pore structure parameter acquisition and application, J. Porous Mater. 21 (2) (2014) 207–215. [24] H. Ma, Z. Li, Realistic pore structure of portland cement paste: experimental study and numerical simulation, Comput. Concr. 11 (4) (2013) 317–336. [25] K. Svinning, A. Høskuldsson, H. Justnes, Prediction of compressive strength up to 28 days from microstructure of Portland cement, Cem. Concr. Compos. 30 (2) (2012) 138–151. [26] K.L. Scrivener, A. Nonat, Hydration of cementitious materials, present and future, Cem. Concr. Res. 41 (7) (2011) 651–665. [27] J.W. Bullard, H.M. Jennings, R.A. Livingston, A. Nonatd, G.W. Scherere, J.S. Schweitzerf, K.L. Scrivenerg, J.J. Thomash, Mechanisms of cement hydration, Cem. Concr. Res. 41 (12) (2010) 1208–1223. [28] V. Tydlitát, T. Matas, R. Cˇerny´, Effect of w / c, and temperature on the earlystage hydration heat development in Portland-limestone cement, Constr. Build. Mater. 50 (2) (2014) 140–147. [29] X. Pang, D.P. Bentz, C. Meyer, G.P. Funkhouser, R. Darbe, A comparison study of Portland cement hydration kinetics as measured by chemical shrinkage and isothermal calorimetry, Cem. Concr. Compos. 39 (5) (2013) 23–32. [30] Y. Lu, H. Ma, Z. Li, Ultrasonic monitoring of the early age hydration of mineral admixtures incorporated concrete using cement-based piezoelectric composite sensors, J. Intell. Mater. Syst. Struct. 26 (3) (2015) 280–291.

B. Dong et al. / Construction and Building Materials 149 (2017) 467–473 [31] J. Zhang, T. Fan, H. Ma, Z. Li, Monitoring setting and hardening of concrete by active acoustic method: Effects of water-to-cement ratio and pozzolanic materials, Constr. Build. Mater. 88 (2015) 118–125. [32] H. Shao, J. Zhang, T. Fan, Z. Li, Electrical method to evaluate elastic modulus of early age concrete, Constr. Build. Mater. 101 (2015) 661–666. [33] B. Dong, J. Zhang, Y. Wang, G. Fang, Y. Liu, F. Xing, Evolutionary trace for early hydration of cement paste using electrical resistivity method, Constr. Build. Mater. 119 (2016) 16–20. [34] X. Wei, L. Xiao, Z. Li, Prediction of standard compressive strength of cement by the electrical resistivity measurement, Constr. Build. Mater. 31 (6) (2012) 341– 346. [35] M. Shi, Z. Chen, J. Sun, Determination of chloride diffusivity in concrete by AC impedance spectroscopy, Cem. Concr. Res. 29 (7) (1999) 1111–1115. [36] P. Gu, X. Ping, J.J. Beaudoinb, Microstructural characterization of the transition zone in cement systems by means of A.C. impedance spectroscopy, Cem. Concr. Res. 23 (3) (1993) 581–591. [37] B. Dong, J. Zhang, Y. Liu, G. Fang, Z. Ding, F. Xing, Tracing hydration feature of aluminophosphate cementitious materials by means of electrochemical impedance method, Constr. Build. Mater. 113 (2016) 997–1006. [38] M. Cabeza, P. Merino, A. Miranda, X.R. Nóvoa, I. Sanchez, Impedance spectroscopy study of hardened portland cement paste, Cem. Concr. Res. 32 (2002) 881–891.

473

[39] K. Brantervik, G.A. Niklasson, Circuit models for cement based materials obtained from impedance spectroscopy, Cem. Concr. Res. 21 (4) (1991) 496– 508. [40] B. Dong, Q. Qiu, Z. Gu, J. Xiang, C. Huang, Y. Fang, Characterization of carbonation behavior of fly ash blended cement materials by the electrochemical impedance spectroscopy method, Cem. Concr. Compos. 65 (2016) 118–127. [41] B. Dong, Q. Qiu, J. Xiang, C. Huang, H. Sun, F. Xing, W. Liu, Electrochemical impedance interpretation of the carbonation behavior for fly ash–slag–cement materials, Constr. Build. Mater. 93 (23) (2015) 933–942. [42] B. Dong, Q. Qiu, J. Xiang, C. Huang, F. Xing, N. Han, Y. Lu, Electrochemical impedance measurement and modeling analysis of the carbonation behavior for cementititous materials, Constr. Build. Mater. 54 (11) (2014) 558–565. [43] B. Dong, Q. Qiu, J. Xiang, C. Huang, F. Xing, N. Han, Study on the carbonation behavior of cement mortar by electrochemical impedance spectroscopy, Materials 7 (1) (2014) 218–231. [44] Z. Liu, Y. Zhang, Q. Jiang, Continuous tracking of the relationship between resistivity and pore structure of cement pastes, Constr. Build. Mater. 53 (53) (2014) 26–31. [45] M. Cabeza, M. Keddam, X.R. Nóvoa, I. Sánchez, H. Takenouti, Impedance spectroscopy to characterize the pore structure during the hardening process of Portland cement paste, Electrochim. Acta 51 (8–9) (2006) 1831–1841.