AgCl ... - J-Stage

Electrochemistry The Electrochemical Society of Japan Article

Received: December 27, 2015 Accepted: February 27, 2016 Published: June 5, 2016 http://dx.doi.org/10.5796/electrochemistry.84.383

Electrochemistry, 84(6), 383–389 (2016)

Electrochemical Characterization of Solid Ag/AgCl Reference Electrode with Different Electrolytes for Corrosion Monitoring of Steel in Concrete Ming JIN,a Linhua JIANG,a,b,c,* Jinxia XU,a Hongqiang CHU,a Debiao TAO,a Shuya BAI,a and Yantao JIAa a b

c

College of Mechanics and Materials, Hohai University, Nanjing 210098, PR China National Engineering Research Center of Water Resource Efficient Utilization and Engineering Safety, 1 Xikang Rd., Nanjing 210098, PR China National Engineering Research Center of Water Resources Efficient Utilization and Engineering Safety

* Corresponding author: [email protected] ABSTRACT Solid Ag/AgCl reference electrodes (RE) with different electrolytes were assembled and characterized in concrete. The performances have been studied with respect to known RE. And the corrosion states of steel have been investigated with respect to embedded and external REs by different electrochemical techniques. Results revealed that half-cell potential of 0.1, 1 and 4.18 M AgCl RE was 20 mV, −30.5 mV and −54.6 mV, respectively. In addition, 4.18 M AgCl RE exhibited the minimum potential drift, the smallest temperature coefficient value and the best reversibility. These corrosion parameters such as rebar potential, corrosion current density (Icorr) and charge transfer resistance (Rct) obtained for steel with respect to these embedded AgCl REs could clearly differentiate the active condition of steel from the passive condition. Besides, Icorr and Rct obtained for steel with respect to external SCE were usually 2 or 3 times larger than these embedded AgCl REs. In such cases embedded AgCl REs could accurately investigate the corrosion rate of steel in concrete. Moreover, the 4.18 M AgCl RE was recommended to monitor the corrosion condition of steel in concrete structure. © The Electrochemical Society of Japan, All rights reserved.

Keywords : Embedded Sensor, AgCl, Reference Electrode, Corrosion Monitoring 1. Introduction Reinforced concrete is concrete where reinforcing bars have been integrated to improve the performance of concrete. However, reinforced concrete may undergo physical deterioration (frost, cracking, fire, etc.), chemical deterioration (acid attack, sea water attack, alkali-aggregate reaction, etc.) and steel corrosion.1 Among them, reinforcement steel corrosion is viewed as a major problem in the maintenance of the structural integrity of structures. Since the corrosion of reinforcement rebar in concrete is electrochemical in nature, therefore, various electrochemical techniques are being adopted to monitor reinforcement steel corrosion, such as half-cell potential,2 linear polarization3 and electrochemical impedance spectroscopy.4 All these techniques call for a reliable reference electrode (RE) which can be embedded in concrete for the reason that the passive/active conditions of steel investigated by the external RE are not accurate.5 The liquid based saturated calomel electrode (SCE) is conventionally used for most laboratory studies in alkaline environments due to its compatibility and easy use, however, this electrode is prepared from mercury and hence any leakage may lead to pollution hazard. Although ASTM C876 specifies the use of Cu/CuSO4 as reference electrode for measurement of steel potential in concrete,6 the leakage of copper sulfate solution can cause contamination of concrete and the IR drop within the concrete cover may lead to erroneous results. It can be clearly seen from the foregoing that there is an utmost need to develop and evaluate the performance of reliable and maintenance-free RE for use in concrete. Such embeddable REs should have the following characteristics:7,8 it should be rugged and maintenance-free, it should have low-temperature coefficient, it also should have stable stability and reliability when compared to a standard RE such as SCE, it should not lead to undesirable contamination of surrounding concrete.

Currently, several research groups are working on developing embeddable REs in concrete. Castro9 and Duffo2 have adopted the metal-metal oxide activated titanium rod as the RE used in concrete, but the potential of this electrode was dependent on pH value. MnO2 RE has been introduced as the embeddable RE in concrete,8,10,11 however, it didn’t belong to the traditional categories of REs because the potential of MnO2 RE could not be described by a theoretical formula or equation. The improved silver/silver chloride (Ag/AgCl) RE has been recommended as the candidate embeddable RE in concrete.12 The drawback of leakage of internal electrolyte can be solved by using polymer gel as electrolyte and cement paste as the bottom layer. This paper deals with the electrochemical studies on the performance characteristics of Ag/AgCl gel reference electrodes containing different electrolytes in concrete. In addition, we compare the embeddable Ag/AgCl RE with the external RE for the corrosion monitoring of reinforcement steel in concrete. 2. Experiment 2.1 Materials used 2.1.1 Fabrication of Ag/AgCl reference electrode Ag/AgCl gel reference electrode consists of three compartments namely a porous hydrated cement paste as bottom layer, KCl solution (0.1, 1 and 4.18 M KCl, respectively) mixed with water retention polymer as the electrolyte and a silver rod with dense AgCl coating inserted in the electrolyte. The abbreviations of Ag/AgCl REs with different kinds of electrolytes are 0.1, 1 and 4.18 M AgCl RE, respectively. 2.1.2 Casting of concrete specimens Concrete specimens were cast with the size of 10 © 10 © 10 cm. A design mix of 1:0.6:1.6:3.2 (cement: water: fine aggregate: coarse aggregate) was used. And the chloride ions were introduced by 383

Electrochemistry, 84(6), 383–389 (2016) adding sodium chloride into the mixing water. The added content of chloride ions was 3% by mass of cement. Moreover, the specimens without chloride ions were cast as the blank specimens. These three kinds of Ag/AgCl REs were embedded centrally into the concrete along with a carbon steel and a Ti electrode. After 24 hours, the concrete specimens were demoulded and cured for three weeks. M0 and M3 represent concrete with 0 and 3% chloride ion, respectively. 2.2 Characterization of Ag/AgCl REs All the electrochemical tests were performed with an Princeton Applied Research (PAR) STAT 2273 Potentiostat. If not specified, the test solution used for contact between concrete and external electrodes was saturated calcium hydroxide solution (SCS), meanwhile, the test was carried out at room temperature (25°C). 2.2.1 Nernst response test Stability test — The potentials of these Ag/AgCl REs embedded in M0 and M3 were monitored with respect to SCE over a period of two months. Effect of temperature- The potential of these Ag/AgCl REs were measured in different temperatures (0, 15, 25, 35, 45 and 55°C) in order to investigate the influence of temperature on the electrode potential. Furthermore, the temperature coefficients of these AgCl REs can be measured. 2.2.2 Polarization test All the polarization tests of Ag/AgCl RE were performed by a three-electrode system assembly consisted of Ag/AgCl RE as working electrode (WE), cylindrical platinum foil as counter electrode (CE) and SCE as RE. Cyclic polarization measurement- The potentiodynamic condition corresponded to the potential sweep rate of 0.2 mV/s and potential ranges of ¹80 to +80 mV from the open circuit potential (OCP). The experiment was run for two cycles. The difference in the shifting of potential from the first cycle and the second cycle was recorded. Potentiodynamic polarization measurement- The potentiodynamic condition corresponded to a potential sweep rate of 1 mV/s and potential ranges of ¹200 to +200 mV from the OCP. Electrochemical impedance spectroscopy (EIS) measurement— The EIS measurement in the frequency range from 10 mHz to 1 MHz was performed. Nyquist plots were recorded for all the embedded REs studied. Galvanostatic polarization measurement- A cathodic direct current density of 0.2 µA/cm2 was applied to all these embedded Ag/AgCl REs for several days. After this, current was interrupted, and Ag/AgCl REs were allowed to recover to their initial OCP. 2.3 Characterization of corrosion condition of reinforcement steel The corrosion condition of reinforcement steel was investigated with respect to these embedded Ag/AgCl RE and external SCE, respectively. The reinforcement steel embedded in concrete specimen was used as the WE. It should be pointed out that when the embeddable AgCl electrode was worked as the RE, the CE was the embeddable Ti electrode. Similarly, if the external SCE worked as the RE, the external cylindrical platinum foil played the role of CE. Half-cell potential (HCP) measurement- The potential of reinforcement steel in M0 and M3 was investigated with respect to Ag/AgCl RE and SCE for one month. Linear polarization (LP) measurement- LP measurement of reinforcement steel in M0 and M3 was carried out with respect to Ag/AgCl RE and SCE, respectively. And LP scan was carried out between ¹10 mV and 10 mV with respect to OCP at a sweep rate of 0.166 mV/s. EIS measurement- Similar to LP measurement, EIS test of reinforcement steel in concrete was carried out with respect to Ag/ AgCl RE and SCE in the frequency range from 10 mHz to 1 MHz. 384

Figure 1. (Color online) Half-cell potential of embedded Ag/ AgCl REs in concrete under passive and active condition.

3. Results and Discussion 3.1 Nernst response test 3.1.1 Stability of Ag/AgCl RE Figure 1 shows the potential evolution with time of these three kinds of Ag/AgCl REs embedded in M0 and M3, respectively. The average half-cell potential of 0.1, 1 and 4.18 M AgCl RE embedded in M0 with respect to SCE is 20 « 5 mV, ¹30.5 « 4.5 mV and ¹54.6 « 2.4 mV, respectively. The maximum variation observed is only «5 mV vs. SCE and this difference in potential is almost negligible, indicating that AgCl REs exhibit perfect stability. The half-cell potential of Ag/AgCl RE in concrete is based on the Ag/ AgCl/KCl equilibrium potential, i.e. Nernst’s law:12–14 o EAg=AgCl ¼ EAg=AgCl 

RT ln aCl nF

ð1Þ

Here, F is the Faraday constant, R is the gas constant, T is the absolute temperature (K), and aCl is the activity of the Cl¹ in solution. When it is at room temperature, and n = 1, the Eq. (1) becomes: EAg=AgCl ¼ Eo  0:059  log aCl

ð2Þ

Thus, according to Eq. (2), the potential of AgCl RE containing 0.1 M KCl electrolyte should be 59 mV higher than that containing 1 M KCl. Moreover, AgCl RE with 1 M KCl should be 36 mV higher than that with 4.18 M KCl. Actually, the potentials of these improved Ag/AgCl REs embedded in concrete agree well with the Nernst’s law. In addition, the difference between the potentials of AgCl REs embedded in M0 and M3 can be observed from Fig. 1. And the biggest difference for 0.1, 1 and 4.18 M AgCl RE during the exposure period is 11 mV, 10 mV and 3.7 mV, respectively. The presence of chloride ions affects little on the potential of embedded AgCl RE. This is quite suited for our interest to use AgCl RE as a true embedded RE for concrete structures. What’s more, for the three kinds of AgCl REs, the 4.18 M AgCl RE shows the minimum drift during the whole testing time and the chloride ions has slightest impact on the electrode potential. 3.1.2 Effect of temperature Reference electrode potential changes with temperature for the reason that both electrochemical reactions and chemical solubilities are affected by temperature. Accordingly, the temperature coefficient (TC), that is dE/dT (mV/°C), varies from one type of RE to another. It is necessary to know the TC in order to minimize errors in potential reading, furthermore, it will be helpful to reduce the error

Electrochemistry, 84(6), 383–389 (2016)

Figure 2. (Color online) Effect of temperature on the potential of Ag/AgCl REs in concrete.

Figure 3. (Color online) Cyclic polarization curve for AgCl REs embedded in concrete.

if the TC is low. Figure 2 shows the potential change of 0.1, 1 and 4.18 M AgCl RE embedded in M0 with temperature. The TC of each electrode can be calculated from these curves. Therefore, calculated TC values of 0.1, 1 and 4.18 M AgCl RE with respect to SCE are 0.02, 0.08 and ¹0.13 mV/°C, respectively. It seems that the 0.1 M AgCl RE has the smallest TC value. But the actual TC value of AgCl RE should be obtained by adding the calculated TC value to the value of SCE. The practical TC value of SCE is 0.22 mV/°C.15 Afterwards, actual TC values of 0.1, 1 and 4.18 M AgCl RE are 0.24, 0.30 and 0.09 mV/°C, respectively. Thus, the 4.18 M AgCl RE has the smallest TC value. When embedding the AgCl RE in the concrete structure to investigate the reinforcement steel potential, the potential error caused by temperature change must be eliminated in order to acquire more accurate potential value of steel. Besides, temperature change will lead to the change of chemical solubility of AgCl coating. For example, raising temperature will increase the solubility of AgCl, which lead to the slope variation of curves of 0.1 and 1 M AgCl RE at 45 and 55°C. It can be clearly found out that there is little variation in the curve of 4.18 M AgCl RE, which can be attributed to that the internal electrolyte in 4.18 M AgCl RE has reached saturated state.

Table 1. Reversibility characteristics of AgCl RE in concrete.

3.2 Polarization tests 3.2.1 Cyclic polarization test of AgCl RE The reversibility characteristic of the AgCl RE in concrete is carried out by cyclic polarization method. The cyclic polarization curves of every kind of AgCl RE in M0 are given in Fig. 3. The reversibility parameters of AgCl RE in concrete derived from the cyclic polarization curves are given in Table 1. Here, it is observed that, the difference in potential between first cycle and second cycle for 0.1, 1 and 4.18 M AgCl RE is 17.6, 6.1 and 6 mV, respectively. It is interesting to found that the difference of 1 M AgCl RE is near to that of 4.18 M AgCl RE. Besides, these differences of 1 and 4.18 M AgCl RE are almost negligible for using AgCl RE in concrete structure. 3.2.2 Potentiodynamic polarization test of AgCl RE The potentiodynamic polarization curves for AgCl REs embedded in M0 are shown in Fig. 4. And the polarization parameters for AgCl RE calculated from the polarization curves are given in Table 2.16 The differences between the half-cell potential (E1/2) and Ecorr values measured for 0.1, 1 and 4.18 M AgCl RE are 1.4, 11.6 and 1.7 mV, respectively. It should be pointed out that the difference for the 0.1 and 4.18 M AgCl RE is less than 2 mV, which proves that no significant change in corrosion potential value is observed. Even a small current is induced during polarization, the corrosion potential of the AgCl RE not changed.

RE

1st (mV)

2nd (mV)

difference (mV)

0.10 M AgCl 1.00 M AgCl 4.18 M AgCl

27.5 ¹24.3 ¹50.9

9.9 ¹30.4 ¹56.9

17.6 6.1 6

Figure 4. (Color online) Potentiodynamic polarization curve for AgCl RE embedded in concrete. This behavior is quite suitable to choose the AgCl RE as an embeddable sensor for concrete structures.8 Besides, the acceptable Icorr value for RE was 0.03 mA cm¹2.17 It can be clearly seen from the Table 2 that the Icorr values of all AgCl REs exceed the acceptable value. The 4.18 M AgCl RE gets the highest Icorr value which indicates that this RE exhibits the best performance of resisting polarization.12 3.2.3 EIS test of AgCl RE The Nyquist diagrams for AgCl REs embedded in M0 are shown in Fig. 5. The Nyquist diagram is usually analyzed by using an equivalent circuit (Fig. 6) consisting of a constant phase-angle element (CPE) in parallel with the charge transfer resistance (Rct, i.e. the polarization resistance.) in addition to the electrolyte resistance (Re).2 In this model, the CPE is used instead of a capacitance to account for the non-ideal capacitive response and characterized by an admittance coefficient (Yo) and a frequency exponent n. Besides, conversion of Yo data into capacitance is very important when 385

Electrochemistry, 84(6), 383–389 (2016) Table 2.

Polarization parameters for AgCl RE in concrete.

RE

E1/2 (mV)

Ecorr (mV)

Difference

bc (mV dec¹1)

ba (mV dec¹1)

Icorr (mA cm¹2)

0.1 M AgCl 1 M AgCl 4.18 M AgCl

20 ¹30.5 ¹54.6

21.4 ¹18.9 ¹56.3

1.4 11.6 1.7

418 1472 1114

436 1583 1231

0.033 0.083 0.095

Table 3. RE 0.1 M AgCl 1 M AgCl 4.18 M AgCl

Impedance parameters for AgCl RE in concrete. Error of Rct Cdl Error of Cdl Rct (©103 ³ cm2) (%) (©10¹4 F cm2) (%) 3.28 0.22 1.58

9 10 8

2.17 4.50 2.57

8.7 11.5 9.5

Figure 5. (Color online) Nyquist plots for AgCl RE embedded in concrete.

CPE

Re Rct Figure 6. Equivalent-circuit model applied to analyze the EIS results for the Ag/AgCl reference electrodes in concrete.

experimental capacitance data are to be used to determine quantitatively system parameters. The conversion procedure is described in detail in literature.18 The charge transfer resistance (Rct) and double layer capacitance (Cdl) values analyzed from the Nyquist diagrams for every kind of AgCl RE are listed in Table 3. The Cdl values for 0.1, 1 and 4.18 M AgCl RE are 2.17 © 10¹4 F cm¹2, 4.50 © 10¹4 F cm¹2 and 2.57 © 10¹4 F cm¹2, respectively. The Cdl value of 1 M AgCl RE is two times as large as those of other AgCl REs. It should be pointed out that the Cdl value of MnO2 RE embedded in concrete is in the order of magnitude of 10¹5 F cm¹2.11 Actually the Cdl value of AgCl RE is more than one order of magnitude larger than the value of MnO2 RE, which means that the AgCl RE shows better performance of capacitive reactance than the MnO2 RE. Besides, the Rct values for 0.1, 1 and 4.18 M AgCl RE are 3.28 © 103 ³ cm2, 0.22 © 103 ³ cm2 and 1.58 © 103 ³ cm2, respectively. It should be pointed out that the Rct value of 1 M AgCl RE is one order of magnitude lower less than the values of other AgCl REs. From EIS tests it was found that 1 M AgCl RE shows the best ability of capacitive reactance, which implies low electrode impedance.2 This capacitive behavior provides tolerance to brief current incursions. Thus, this electrode has the capability to pass brief small currents with a minimum of polarization which is 386

Figure 7. (Color online) Galvanostatic test curve for AgCl RE embedded in concrete.

advantageous when conducting electrochemical measurements on reinforcement steels. 3.2.4 Galvanostatic test of AgCl RE After a quasi-stabilization period that ranged from 1 to 24 hours, a cathodic current density of 0.2 µA/cm2 was applied for a period of 5–6 days. Then the potential was measured for another period of 1 day. Results of the galvanostatic tests for three kinds of AgCl REs are shown in Fig. 7. The potentials of 0.1, 1 and 4.18 M AgCl RE shift 35, 15 and 28 mV, respectively. And the potential recovers quickly after the interruption of the current. Among these three AgCl REs, 1 M AgCl RE potential shifted the minimum value, which can be attributed to the reason that 1 M AgCl RE had the best performance of capacitive reactance. Besides, a polarization resistance is computed from the applied current density during the galvanostatic pulse (0.2 µA/cm2) and the measured potential change. And the resulting values for 0.1, 1 and 4.18 M AgCl RE are 223 k³, 96 k³ and 178.4 k³, respectively. Thus, for a typical input bias current of 1 nA (actually, typical input bias currents of operational amplifiers are 40 nA, but for potentiostat circuits, input bias currents of less than 5 nA are recommended), potential shift ranging from 0.096 to 0.223 mV are obtained. Based on this, the potential shift value is really negligible when embedding the AgCl RE in concrete structures to measure the half-cell potential of reinforcement steel. Then it is concluded that these AgCl REs resist the long-term effect of the application of small bias currents usually found in electrochemical instruments.

Electrochemistry, 84(6), 383–389 (2016) 3.3 Characterization of corrosion condition of steel 3.3.1 Measurement of steel potential in concrete Rebar potential (ER) measured for steel in M0 and M3 with respect to internal AgCl REs and external SCE during the testing period is given in Fig. 8. M0 and M3 represent the passive and active condition of reinforcement steel, respectively. At the end of investigation, the measured half-cell potential of steel with respect to embedded 0.1, 1 and 4.18 M AgCl RE and external SCE under passive condition is ¹240, ¹192, ¹170 and ¹221 mV, respectively. On the other hand, half-cell potential of reinforcement steel with respect to embedded 0.1, 1 and 4.18 M AgCl RE and external SCE under active condition is ¹435, ¹380, ¹353 and ¹410 mV, respectively. Figure 8 indicates that all of these embedded AgCl REs could clearly characterize the active and passive state of reinforcement steel in concrete. Moreover, the ER measured by these embedded AgCl REs shows more stable during the testing period than the potential measured by the external SCE. Which can be attributed to the reason that the IR drop produced by the concrete cover.19 However, the IR drop can be eliminated through placing the solid RE near to the reinforcement steel in the concrete structure. Thus, it is recommended to investigate the potential of reinforcement steel embedded in concrete structure by the internal/embeddable reference electrode.

3.3.2 Linear polarization of steel in concrete The linear polarization curves for steels in M0 and M3 with respect to embedded 0.1, 1 and 4.18 M AgCl RE and external SCE are given in Fig. 9. The ER, Ecorr and Icorr values obtained for different REs in M0 and M3 are reported in Table 4. The polarization resistance (Rp) was introduced to Stern-Geary formula to calculate the Icorr value. The Stern-Geary equation is Icorr ¼

B Rp

ð3Þ

where, B is the Stern-Geary constant. A value of 26 mV has been adopted for active steel and 52 mV for passive steel.20 The ER value measured for steel in M0 with respect to 0.1, 1, 4.18 M AgCl RE and external SCE is ¹240, ¹192, ¹170 and ¹221 mV, respectively, whereas in M3 is ¹410, ¹435, ¹380 and ¹353 mV respectively. What is more, the Ecorr values measured for steel in M0 with respect to 0.1, 1 and 4.18 M AgCl RE and external SCE is ¹250, ¹200, ¹176 and ¹230 mV, respectively, whereas in M3 is ¹455, ¹400, ¹374 and ¹430 mV respectively. Similar to the external SCE, these data measured by the embedded AgCl REs (i.e., 0.1, 1 and 4.18 M AgCl RE) clearly distinguish the behavior of steel in active condition from passive condition. Thus, it is reliable to monitor the change of steel potential in concrete structures by the embedded reference electrodes.

Table 4. Polarization parameters for steel in concrete under passive and active conditions with respect to embedded and external reference electrodes. Concrete specimen

Figure 8. (Color online) Half-potential of steel in concrete under active and passive condition with respect to embedded AgCl REs and external SCE.

(a). passive condition

ER Icorr Ecorr (mV vs. RE) (µA/cm2) (mV vs. RE)

RE

M0

SCE 0.1 M AgCl 1.0 M AgCl 4.18 M AgCl

¹221 ¹240 ¹192 ¹170

0.254 0.155 0.162 0.182

¹230 ¹250 ¹200 ¹176

M3

SCE 0.1 M AgCl 1.0 M AgCl 4.18 M AgCl

¹410 ¹435 ¹380 ¹353

6.017 2.272 3.062 3.653

¹430 ¹455 ¹400 ¹374

(b). active condition

Figure 9. (Color online) Linear polarization curves for steel under active and passive condition in concrete with respect to embedded AgCl REs and external SCE. 387

Ω

Ω

Electrochemistry, 84(6), 383–389 (2016)

Ω

Ω

Figure 10. (Color online) Nyquist plots for steel in steel under active and passive condition in concrete with respect to embedded AgCl REs and external SCE.

In addition, the Icorr value measured for steel in M0 with respect to 0.1, 1, 4.18 M AgCl RE and external SCE is 0.155, 0.162, 0.182 and 0.254 µA/cm2 respectively, whereas in M3 is 2.272, 3.062, 3.653 and 6.017 µA/cm2 respectively. According to literature,20 when Icorr value of steel was lower than 0.1 µA/cm2, the rebar was in passive condition. However, when Icorr value exceeded 1 µA/cm2, the steel got a high corrosion rate. The Icorr value of steel measured by the embedded AgCl REs also clearly distinguish the behavior of steel in active condition from passive condition. It should be noted that the Icorr value measured by the external RE (i.e., SCE) is 2 times higher corrosion rate values than embedded AgCl REs. The variation in the difference in the measured Icorr values between embedded RE and external RE was due to the liquid junction and IR drop problem in concrete caused by the high internal resistance and the heterogeneity of the structure.21 Here it was concluded that more accurate corrosion rate can be obtained through embedded reference electrode than the external reference electrode. Last but not least, Icorr value of steel measured by 0.1 M AgCl is smaller than other AgCl REs. The results reveal that the Icorr value obtained by 0.1 M AgCl is the most accurate in all these embedded REs. 3.3.3 EIS measurement of steel in concrete The impedance curves for steels in M0 and M3 with respect to embedded 0.1, 1 and 4.18 M AgCl RE and external SCE are given in Fig. 10. Generally, Nyquist plots depict a capacitive loop of depressed semicircle. Electrochemically, the EIS curve in the high frequency suggests the resistance between electrolyte solution and the working electrode while that in the low frequency is closely related to the charge transfer resistance of the corrosion process. In order to obtain ideal charge transfer resistance (i.e. the polarization resistance Rp), the EIS data from 10 kHz to 10 mHz is analyzed with the equivalent-circuit shown in Fig. 11, which has been used by other authors.4,22,23 The Rs in high frequency is the solution resistance and the time constant in high frequency (Cf and Rf ) reflects the information of the passive film. Furthermore, the time constant in low frequency (Cdl and Rct) is related to the information of the corrosion reaction on the steel surface. Then the Rct, i.e. Rp, is introduced into the Stern–Geary equation [Eq. (3)] to calculate the corrosion current density (Icorr). Based on this, the corresponding impedance parameters are shown in Table 5. It can be observed from Fig. 10 that the diameter of semicircle measured by embedded RE is smaller than that measured by external SCE. Furthermore, the Rct values obtained is in the order of 104 ³ cm2. Results from Table 5 indicate that higher Rct values are noticed in the passive condition of steel than the active condition of steel. Which means that the Icorr values of steel in passive condition are smaller than the ones of steel in active condition. It should be pointed out that the external SCE shows the higher Icorr values of 388

Cf

Cdl Rs Rf Rct Figure 11. Equivalent-circuit model applied to analyze the EIS results for the steel in concrete. Table 5. Impedance parameters for steel in concrete under passive and active condition with respect to embedded and external RE. Concrete specimen

RE

Rct Icorr Cdl (©104 ³ cm2) (µA/cm2) (©10¹4 F cm2)

M0

SCE 0.1 M AgCl 1.0 M AgCl 4.18 M AgCl

3.5 4.5 4.9 5.0

0.158 0.123 0.113 0.110

5.292 1.607 1.223 1.253

M3

SCE 0.1 M AgCl 1.0 M AgCl 4.18 M AgCl

0.127 0.369 0.381 0.341

8.693 2.992 2.898 3.237

6.611 2.060 1.983 1.296

steel than embedded AgCl REs. For example Icorr value for steel in active condition with respect to these embedded AgCl REs is ca. 3 µA/cm2. However, this value measured by the external SCE is 8.693 µA/cm2. These data clearly prove that the external SCE exhibits almost 3 times than the embedded REs. What is more, double layer capacitance (Cdl) values are also noticed for steel with respect to these embedded REs and external one. Cdl values obtained are in the order of 10¹4 F cm2. Moreover, according to literature,20 Cdl value is inversely proportional to the diameter of the semicircle in the Nyquist plots (Fig. 10). Based on this, higher Cdl values are noticed for active condition when compared to passive condition. For example Cdl values for steel with respect to 0.1 M AgCl RE under active and passive condition are 2.060 © 10¹4 F cm2 and 1.607 © 10¹4 F cm2, respectively. Meanwhile, the Cdl values for

Electrochemistry, 84(6), 383–389 (2016) steel in active condition measured by other REs are also higher than the ones in passive condition. These results confirm the fact that all the embedded AgCl REs perform very well in concrete and differentiate the passive and active condition of reinforcement steel in concrete. However, there is a big difference between the Cdl values measured with respect to embedded REs and external SCE. For instance, Cdl value for steel in active condition with respect to these embedded AgCl REs is approximately 2 © 10¹4 F cm2. But the Cdl value obtained with respect to the external SCE is up to 6.610¹4 F cm2, respectively. These data reveal that the external SCE shows more than 3 times than these embedded REs. And the Icorr value of steel under passive condition measured by the 4.18 M AgCl RE is the smallest. Nevertheless, the Icorr value of steel under active condition measured by 4.18 M AgCl RE is the highest. From above observation it is clearly understood that the corrosion rate of reinforcement steel embedded in concrete is usually more accurate with respect to embedded REs when compared to external RE. Form the viewpoint of service life prediction of field structures, accurate measurements of corrosion rate of steel is a pre-requisite. In such cases these embedded reference electrodes accurately predict the corrosion of reinforcement steel in concrete. 4. Conclusions In the present work it was found that half-cell potential of 0.1, 1 and 4.18 M AgCl RE embedded in concrete with respect to SCE was 20 mV, ¹30.5 mV and ¹54.6 mV, respectively. What’s more, the 4.18 M AgCl RE showed the minimum potential drift and the smallest temperature coefficient value in these three REs. Cyclic polarization test revealed that the difference in potential between first cycle and second cycle for 1 and 4.18 M AgCl RE was only 6 mV, which certified the good reversibility of the embedded RE in concrete. Potentiodynamic polarization test indicated that 4.18 M AgCl RE had the highest Icorr value, which proved 4.18 M AgCl RE exhibited the best performance of resisting polarization. The impedance parameters further confirmed the perfect stability of AgCl RE in concrete. It should be pointed out that 1 M AgCl RE showed the best ability of capacitive reactance and this capacitive behavior provided tolerance to brief current incursions. Last but not least, galvanostatic application of 0.2 µA/cm2 in several days for 0.1, 1 and 4.18 M AgCl RE only caused 35, 15 and 28 mV respectively, which indicated the presence of a finite polarization resistance. The steel potential with respect to the embedded 0.1, 1 and 4.18 M AgCl RE were ¹240, ¹192, ¹170 mV for passive condition and ¹435, ¹380, ¹353 mV for active condition. The results showed that all of these embedded AgCl REs could clearly characterize the active and passive state of steel in concrete. The corrosion current from linear polarization and Rct from impedance technique clearly differentiated the behavior of steel whether in passive or active condition with respect to these embedded AgCl REs. Besides, corrosion parameters such as Icorr and Rct obtained for steel with

respect to external SCE were usually 2 or 3 times larger than these embedded AgCl REs. Here it was concluded that reliable corrosion rate could be obtained through embedded REs. And it was interesting to point out that for service life prediction of field structure, accurate measurement of corrosion parameters of steel was a pre-requisite. In such cases embedded AgCl REs could accurately investigate the corrosion rate of steel in concrete. Acknowledgments The authors gratefully acknowledge the support provided by the National Key Technology Research and Development Program of the Ministry of Science and Technology of China (2015BAB07B04), the Fundamental Research Funds for the Central Universities (No. 2016B04514), the Natural Science Foundation of China (Nos. 51278167, 51278168, 51478164 and 51479051) and the Natural Science Foundation of Jiangsu province under Project (No. BK20131374). References 1. P. K. Mehta, P. J. Monteiro, and M. Education, Concrete: Microstructure, Properties, and Materials, McGraw-Hill New York (2006). 2. G. S. Duffó, S. B. Farina, and C. M. Giordano, Electrochim. Acta, 54, 1010 (2009). 3. J. Xu, L. Jiang, and J. Wang, Constr. Build. Mater., 23, 1902 (2009). 4. R. Liu, L. Jiang, J. Xu, C. Xiong, and Z. Song, Constr. Build. Mater., 56, 16 (2014). 5. S. P. Karthick, S. Muralidharan, V. Saraswathy, and K. Thangavel, Sens. Actuat. B, 192, 303 (2014). 6. C. Astm, Standard Test Method for Half-Cell Potentials of Uncoated Reinforcing Steel in Concrete (1999). 7. W. J. McCarter and Ø. Vennesland, Constr. Build. Mater., 18, 351 (2004). 8. S. Muralidharan, V. Saraswathy, K. Thangavel, and N. Palaniswamy, Sens. Actuat. B, 130, 864 (2008). 9. P. Castro, A. A. Sagues, E. I. Moreno, L. Maldonado, and J. Genesca, Corrosion, 52, 609 (1996). 10. S. Muralidharan, T. Ha, J. Bae, Y. Ha, H. Lee, K. Park, and D. Kim, Mater. Lett., 60, 651 (2006). 11. S. Muralidharan, V. Saraswathy, A. Madhavamayandi, K. Thangavel, and N. Palaniswamy, Electrochim. Acta, 53, 7248 (2008). 12. M. Jin, J. Xu, L. Jiang, G. Gao, H. Chu, C. Xiong, H. Gao, and P. Jiang, Electrochemistry, 82, 1040 (2014). 13. U. Angst, B. Elsener, C. K. Larsen, and Ø. Vennesland, J. Appl. Electrochem., 40, 561 (2010). 14. M. Jin, J. Xu, L. Jiang, Y. Xu, and H. Chu, Ionics, 21, 2981 (2015). 15. R. Myrdal, The Electrochemistry and Characteristics of Embeddable Reference Electrodes for Concrete, Woodhead Publishing, Cambridge, England, p. 3 (2007). 16. L. Yang, Techniques for Corrosion Monitoring, Elsevier (2008). 17. C. S. Brossia and R. G. Kelly, Electrochim. Acta, 41, 2579 (1996). 18. C. H. Hsu and F. Mansfeld, Corrosion, 57, 747 (2001). 19. J. Saxena, K. M. Butler, V. B. Jayaram, S. Kundu, N. V. Arvind, P. Sreeprakash, and M. Hachinger, in 2013 IEEE International Test Conference (ITC), p. 1098 (2003). 20. V. S. Ha-Won Song, Int. J. Electrochem Soc., 2, 1 (2007). 21. S. P. Karthick, S. Muralidharan, V. Saraswathy, and K. Thangavel, Sens. Actuat. B, 192, 303 (2014). 22. J. M. Blengino, M. Keddam, J. P. Labbe, and L. Robbiola, Corros. Sci., 37, 621 (1995). 23. M. B. Valcarce and M. Vázquez, Mater. Chem. Phys., 115, 313 (2009).

389