XU Ning, JIANG Linhua, ZHOU Huaiming, CHU Hongqiang, JIANG Peng

(1. College of Mechanics and Materials, Hohai University, Nanjing 210098, China; 2. Materials & structural Engineering Department, Nanjing Hydraulic Research Institute, Nanjing 210024, China; 3. CCCC Investment Nanjing Co. Ltd., Nanjing 211800, China)

Abstract: The electrical conductivity of graphene/cement composite before and after carbonation was tested by a four-electrode method. The General Effective Media equation was adopted to fit the experiment results. EIS (electrochemical impedance spectroscopy) was employed to study the effect of carbonation on conductivity of graphene/cement composite. The mechanism was analyzed by SEM (scanning electron microscopy). It is revealed that electrical conductivity increases with increasing carbonation depth when the graphene content is less than 2.0%. In this case, the electrical conductivity of composite depends on cement matrix which can be enhanced by carbonation product through filling pores. When the graphene content exceeds 3.3%, the electrical conductivity decreases with increasing carbonation depth. The conductive path is mainly formed by graphene chains which can be broke by carbonation product. The GEM (General Effective Media) equation fits the experimental results well and can be used to calculate the electrical conductivity of graphene/cement composite after carbonation.

Key words: graphene; cement; carbonation; electrical conductivity

1 Introduction

Cement based composites (CBCs) are the most widely used construction materials around the world. Usually, CBCs have very high resistance especially in dry condition. Incorporating conductive materials into CBCs can help improve it’s electrical conductivity to realize self sensing of structure[1-6].

In recent years, one of the most popular used conductive material is graphene. As a novel two-dimensional carbon-based nanomaterial, graphene has incredible electronic mobility (200 000 cm2V-1S-1)[7,8]. CBCs infused with graphene has shown excellent electrical conductivity that exhibits piezoresistivity-based strain or damage sensing ability[9,10]. Baiet al[2]reported that 1.5wt% graphene added into cement could increase three order of magnitude of electrical conductivity. There was a percolation threshold of graphene reinforced cement. When the content of graphene exceeded percolation threshold, a continuous electrical path network can be developed. Sunet al[11]and Taoet al[12]found CBCs infused with graphene present sensitive piezoresistivity that can be utilized for strain or damage sensing of civil infrastructure. Another applications based on the electrical conductivity of graphene CBCs includes monitoring chloride ion penetration in concrete structure and static/dynamic wireless charging for the electric vehicle[13].

The CBCs are a highly-buffered alkaline matrix that consist of C-S-H gel, Ca(OH)2, and ettringite. When exposed to air, the alkaline matrix will react with CO2, commonly known as carbonation. During carbonation process, the reaction product will change the microstructure including pores and compositions. Considerable effort has been placed on characterizing the carbonation mechanism by SEM, EIS, XRD,et al, methods[14-18]. However, the effect of carbonation on the electrical conductivity is rarely reported.

In this paper, the electrical conductivity of graphene/cement composite was tested at different carbonation depth. The GEM equation was used to fit the experiment results. Three electrical equivalent circuit were established considering the graphene content to investigate the effect of carbonation on electrical conductivity by EIS method. In addition, the microstructure of the composite before and after carbonation was studied by means of SEM.

2 Experimental

2.1 Materials

Graphene was manufactured by physical stripping and its properties are summarized in Table 1. Cement was PⅡ 42.5 ordinary portland cement in accordance with GB175-2007 (China standard). The chemical composition of cement is shown in Table 2. Silica fume contained more than 85% SiO2and its average particle size was in the range of 0.1-0.3 μm. The chemical composition of silica fume is presented in Table 3. Sodium dodecyl benzene sulfonate (SDBS) and a polycarboxylate-type water reducer were used to make specimens.

Table 1 Properties of graphene

Table 2 Composition of cement/wt%

Table 3 Physical properties of cement

Table 4 Composition of silica fume/wt%

2.2 Specimens preparation

The binder materials were cement a nd silica fume and the amount of silica fume was 10% of the total binder materials by mass. Graphene was added in concentrations of 0, 0.7%, 1.3%, 1.7%, 2%, 2.3%, 2.6%, 3%, 3.3%, 3.5%, and 3.9% by volume of the total binder materials. The water to binder ration was 0.5. SDBS was added in concentrations of 3% by mass of graphene. Water reducer was added moderately to ensure sufficient workability of the specimens. To get graphene dispersion, graphene, SDBS and water reducer were dispersed firstly in water by magnetic stirring for 20 min and ultrasonic treating for 40 min. Cement and silica fume were blended in advance used a motor-stirrer. The graphene dispersion and binder materials were mixed and stirred for 10 min and then cast in a mould (20 mm×20 mm×60 mm). Four stainless steel gauze electrodes were inserted in the fresh paste as shown in Fig.1. After 24 h, the specimens were demoulded and cured in water under 20 ℃ for 3 months.

Fig.1 The schematic diagram of specimen

2.3 Experimental methods

After cured, the specimens were dried at 60 ℃ for 2 days. Five surfaces were sealed by paraffin then put in a box with the condition of 20% CO2, 20 ℃ and 70% relative humidity to conduct carbonation test. During the carbonation process, a parallel specimen was split and sprayed phenolphthalein solution to examine the carbonation depth every day. When accomplishing carbonation, the specimens were dried at 60 ℃ for 2 days again to test their electrical conductivity by four-electrode method and can be calculated by equation (1):

where, σ is the electrical conductivity, R is the resistance, S is the contact area between stainless steel gauze electrode and paste, L is the length between stainless steel gauze electrode 2 and 3, U is the voltage, and I is the current.

Fig.2 The schematic diagram of electrical conductivity measurement based on four-electrode method

The electrochemical impedance spectroscopy (EIS) test was run over the frequency range from 0.1 Hz to 1 MHz. The microscopy was observed using a Zeiss Sigma HD scanning electron microscopy.

3 Results and discussion

3.1 Electrical conductivity of specimens

The relationship of electrical conductivity with graphene content under different carbonation depths is shown in Fig.3. It can be seen that there was a percolation phenomenon in graphene/cement composite system that similar to Sunet al[19]and Baiet al[2]. Before carbonation, the addition of graphene into cement enhanced the electrical conductivity of system. When the content of graphene was in the range of 2% to 3.3%, the electrical conductivity increased dramatically with increasing graphene content, which could be called the threshold region. After carbonation, the new chemical reaction product and change in pore structure may cause variation of the electrical conductivity, but not have much effect on the threshold region.

Fig.3 Relationship of electrical conductivity with different graphene contents under different carbonation depths

The relative changes of electrical conductivity under different carbonation depths are shown in Fig.4. When the graphene content was less than 2.0%, the carbonation depth had a remarkable effect on the electrical conductivity of the specimens and the conductivity increased uniformly with the carbonation depth increasing. When the graphene content was between 2.0% to 2.6%, the relative changes of electrical conductivity reduced sharply to negative. And the larger carbonation depth, the more variation of relative change. When the graphene content exceeded 2.6%, the relative changes of electrical conductivity show opposite results compared to that less than 2.0%.

Fig.4 Relative changes of electrical conductivity under different carbonation depths

3.2 Fitting of experiment results

According to the Effective Medium Theory, the conductive behavior of materials is not only related to the conductive filling material but also related to the matrix. The effective conductivity of composite depends on the content of each phase, the conductivity, shape, distribution and other factors. On this basis, Mc-Lachlanet alput forward the General Effective Media (GEM) equation[20]:

where,σhis the conductivity of high conductive phase;σlis the conductivity of high conductive phase;σmis the conductivity of composite;tis the key index;φis the content of conductive phase;φcis the percolation threshold.

Fig.5 Curves of GEM equation with a constant value: (a) t; (b) φc

For GEM equation,σh,σl,σm, andφcan be confirmed by experiment. In graphene/cement composite system, the curves of GEM equation with a constant value oftorφcare shown in Fig 5. Whentis constant, the percolation threshold has little effect on the conductivity of the composite, and the width of the percolation range is basically the same. The conductivity tends to accordance with the increase of graphene. When the value ofφcremains the same, the change ofthas a significant effect on the conductivity of the composite. The increase oftwill reduce the percolation range and increase the conductivity finally. The influence ofton the conductivity of the composite is much greater thanφc.

The fitting curves of electrical conductivity with different graphene contents under different carbonation depths by GEM equation are shown in Fig.6. As shown in Fig.6, GEM equation fits the experiment results very well. From Fig.6(a), it can be seen that the percolation threshold was 0.0260 which is among the threshold region. Thus, the electrical conductivity of graphene/cement composite can be calculated by substitutingtandφcinto equation 2. After carbonation, the component and micro-structure will change, causing a changingtandφc. From Figs.6(b-e), the percolation threshold obtained through fitting results tested under 5, 10, 15, and 20 mm carbonation depth were 0.0265, 0.0268, 0.0271, and 0.0272, respectively. The relationship oftandφcwith different carbonation depths is shown in Fig.7. Thetandφcwere linear to the variation of carbonation depth and cab be expressed as:

Fig.6 Fitting curves of electrical conductivity with different graphene contents under different carbonation depths by GEM equation: (a) 0 mm; (b) 5 mm; (c) 10 mm; (d) 15 mm; (e) 20 mm

Fig.7 Relationship of (a) t and (b) φc with carbonation depth

Here,dis the carbonation depth, mm.

In Fig.7(a),tdecreases with the increase of carbonation depth. As key index of graphene/cement composite,tis related to the space dimensionality of cement matrix and the size and shape of graphene. During carbonation process, graphene did not produce chemical reaction thus obtained same size and shape. The cement matrix became more dense because of carbonation therefor ledtto decrease. In Fig.7(b),φcincreased with the increase of carbonation depth. To a certain extent, the percolation threshold can represent the spatial dispersion of graphene in cement matrix. The carbonation product wrapped around graphene and obstructed the conductive pathway, resulting a decrease of electrical conductivity. Consequently, the threshold region shifted to right, which reflected the the nature of reduced dispersion of graphene.

3.3 EIS analysis

The microstructure of graphene/cement composite can be simplified into Fig 8, where the black elliptical slices represent graphene layers. There were three main path formed by graphene. According to Fig.3, when the graphene content was below 2.0% (region A), the graphene layers were far away from each other,which formed insulator path. When the graphene content was between 2.0% to 3.3% (region B), some of the graphene layers overlapped that formed discontinuous path. When the graphene content was more than 3.3% (region C), the graphene layers linked into continuous path. Based on Song’s analytic method, the insulator path and continuous path can be equivalent to capacitance and resistance respectively, while the discontinuous path is series and parallel circuit comprised both of capacitance and resistance.

According to Fig.8, the equivalent circuit of region A, B, and C can be simplified as shown in Fig.9, correspondingly, their electrical model would beQ1(Rct1W1))(Q2(Rct2W2)),Q1(Rct1W1)(QNRN)(Q2(Rct2W2)), andQ1(Rct1W1)(QNRN)(Q2(Rct2W2)), respectively. Here, every last elementQ2(Rct2W2) is related to the chemical reaction between cement matrix and stainless steel gauze electrode, which has no relation to the electrical conductivity of matrix. Thus, the total impedance (Z) of graphene/cement composite would be written as:

Fig.8 Analyzed circuit diagrams of graphene/cement composite

Fig.9 The equivalent circuit of regions A, B, and C

where,Q1is the capacitance existing between graphene and cement;Rct1is the resistance of charge transfer in cement;W1is the Warburg resistance of charge diffusion in cement;QNis the capacitance existing between graphene that close or contact to each other;RNis the resistance of contiguous graphene.

Fig.10 shows the Nyquist plots of the impedance spectrums for the composites with different graphene contents under different carbonation depths. It can be intuitively seen from Fig.10 that the shape of Nyquist plot was similar when the graphene content was consistent. In Fig.10(a), the intersection of the two bulk features decreased with the increasing carbonation depth. In Figs.10(b) and 10(c), the high-frequency arc diameter and position for the composites under different carbonation depths was generally same. This phenomenon is mainly due to the increasing depth of carbonation and solid phase addition, which will lead to the strengthen of the microstructure as well as the increase of the conductivity of cement matrix. In order to quantitatively study how does the carbonation influence on the electrochemical parameters, the detailed values ofRct1andRNcalculated from theQ1(Rct1W1)) andQ1(Rct1W1)(QNRN) model are listed in Table 4. It is found that the value ofRct1decreases with the increasing carbonation depth, while the value ofRNincreases with the depth. The first rule is generally due to the fill of pores by carbonation product, and the latter one is mainly attributed to intercept of the continuous graphene. In addition, the relative changes of electrical conductivity transformed from the values ofRct1andRNare shown in Fig.11. It can be seen that the results based on four-electrode and EIS method are consistent and both linear to carbonation depth.

Fig.10 Nyquist plots measured at different carbonation depths for the composites with different graphene contents: (a) 1.3%; (b) 2.6%; (c) 3.9%

Table 5 Fitting results of Rct1 and RN based on Q1(Rct1W1)) and Q1(Rct1W1)(QNRN) model

Fig.11 Relationship of relative changes of electrical conductivity under different carbonation depths based on four-electrode and EIS method

3.4 Microscopic analysis

Fig.12 shows the microstructure of graphene/cement composite before and after carbonation. The flat flakes were graphene which tiled or embedded in cement matrix. It is found that there is unconspicuous morphological change of graphene flakes after carbonation. In Fig.12(a), graphene close contacted with gel. After carbonation, there were irregularly shaped prominences or sporadic crystals presenting around graphene. Must of them were CaCO3produced by chemical reaction of H2CO2with Ca(OH)2, AFt, and gel. These product of carbonation filled a part of pores and obstructed the connection between graphene flakes. Based on above results, the schematic plot of conductive path of graphene/cement composite after carbonation is exhibited in Fig.13. Before carbonation, the main conductive paths were gel matrix and continuous or discontinuous path formed by graphene. After carbonation, the carbonation product filled in pores, adhered to the surface of graphene and embedded between graphene flakes. Correspondingly, the present of carbonation product enhanced the conductivity of gel but impaired the conductive path of graphene. In region A, the system conducted mainly through gel, thus the electrical conductivity increased by carbonation. In region B, the conductive path of graphene began to form that contributed more conductivity of system. Therefore, the effect of carbonation on conductivity translated from strengthening to weakening. In region C, the system conducted mainly through impeccable graphene network. Carbonation would break the connection causing lower conductivity.

Fig.12 SEM images of graphene/cement composite: (a) Before carbonation; (b) and (c) After carbonation

Fig.13 Schematic plots of conductive path of graphene/cement composite after carbonation

4 Conclusions

a) Carbonation enhanced the electrical conductivity when graphene/cement composite had low content of graphene while reduced when the content of graphene exceeded percolation threshold.

b) The experiment results were fitted well by General Effective Media equation. The key index and percolation threshold were linear to carbonation depth. On the basis of fitted results, the electrical conductivity of carbonated graphene/cement composite could be calculated.

c) The effect of carbonation on the electrical conductivity of graphene/cement composite can be characterized by the equivalent circuit models proposed corresponding to the content of graphene. The parameters extracted from the models can be applied to characterize the variation of electrical conductivity caused by carbonation.The carbonation depth had a linear relationship with the parameters.

d) The SEM observation deciphered and supported the reasonability of microstructure changes proposed for the variation of electrical conductivity caused by carbonation.