Experimental and numerical study of electrochemical chloride removal from brick and concrete specimens

https://doi.org/10.1016/j.cemconres.2006.09.012Get rights and content

Abstract

The electrochemical technique for chloride extraction (desalination) was applied in galvanostatic mode to cylindrical brick and concrete specimens with a steel bar as reinforcement placed in the centre. The specimens were initially contaminated by immersion in a solution of 35 g/l NaCl. Based on the Nernst–Planck equations, a numerical model was developed considering the interactions between the various ionic species in the pore solution. The model makes it possible to predict the evolution of the chloride profile with time. The numerical and experimental results are compared and the model parameters discussed.

Introduction

Corrosion of steel reinforcement in contact with chloride ions is one of the major causes of degradation of reinforced concrete structures throughout the world. The deterioration induced is characterized by a reduction of the cross-section of the embedded steel and loss of the bond between steel and concrete. For new structures, protective methods such as galvanizing of the reinforcements [1], [2], [3], epoxy coating of steels [4] or addition of inhibitors [5] are used where there is a risk of rebar corrosion. In the case of old structures, repair techniques remain essential for rehabilitating corroded parts. Among the methods available is chloride extraction, also called desalination, by application of an electric field [6], [7], [8], [9], [10]. This method involves the application of a high current density for a short period, typically a few weeks. The steel reinforcement acts as the cathode, and an extended anode is placed in a suitable electrolyte at the concrete surface. The steel cathode repels anions and attracts cations in the concrete pore solution. Therefore, sodium and potassium accumulate at the cathode. This accumulation has been shown to increase the risk of an alkali–silica reaction in concrete containing potentially reactive aggregates [9].

The present work aims at analyzing the transport of ionic species in a saturated medium in order to predict the accumulation of alkaline species near the embedded steel and to define a time criterion for the efficiency of the technique. A model based on a multi-species approach is presented. Moreover, in order to validate the numerical predictions, a desalination programme was conducted on both cylindrical brick and concrete specimens for various periods. The inert material (brick) was chosen for two reasons: (i) to avoid binding between chloride and the matrix, which is dominant in the case of concrete specimens, and (ii) to limit the number of chemical species to be taken into account in the computation. Thus, the model was tested in a simple case before performing simulations with concrete.

Section snippets

Basic equations

According to irreversible thermodynamics and assuming that the activity coefficient is equal to the concentration, the diffusion process in saturated porous media can be completely described through the following set of extended Nernst–Planck equations [11], [12], [13], [14], [15], [16], [17], [18], [19], [20]:Ji=[DiCi+ziDiCiFRTϕ]ϕ=RTF(Ic/F)+iziDiCiizi2DiCi where Ji is the flux of the ion i (mol m 2 s 1), Di the diffusion coefficient (m2 s 1), Ci the concentration (mol m 3), zi the

Experimental programme

The experimental study comprised two parts. The first one dealt with brick specimens (inert material) which provided a test of the model in simplified conditions. The simplified conditions consisted of:

  • (i) the number of species computed was limited (Na+ and Cl), and

  • (ii) chloride binding was negligible.

In the second part, the concrete specimens were tested, corresponding to the real conditions in which the technique is used.

Porosity and density

The porosity accessible to water and the density were measured on specimens of the same dimensions as those described in Section 3.1. Vacuum saturated water specimens were first weighed in water and in air. Then the samples were put in an oven at 105 °C to determine the dry weight. The values p = 0.196 and ρ = 2440 kg/m3 were obtained for the brick specimens and p = 0.11 and ρ = 2627 kg/m3 for the concrete specimens.

Concrete pore solution composition

The knowledge of the initial concentrations of the various ionic species in the

Brick specimens

Fig. 6 shows the measured and computed concentration profiles of chloride ions within the material at four different times. For the current density applied (1 A/m2), the ionic transport was dominated by diffusion. This was due to a strong porosity of the brick (19.6%) and the absence of binding. Concerning the efficiency of the ECE technique, it was found that the chloride content near the steel had been halved after 20 days of treatment.

Concrete specimens

Fig. 7 shows the measured and computed total chloride

Conclusions

Electrochemical chloride removal investigation was conducted on both brick and concrete materials by associating experimental and numerical approaches. The following conclusions can be drawn from the present study:

  • 1.

    The chloride removal and associated ionic movement in the pore solution can be reasonably predicted using the Nernst–Planck equation;

  • 2.

    The ECE technique can significantly reduce the amount of chloride near the steel. However, the use of high current density leads to a rapid increase of

Acknowledgement

The authors would like to thank C. Andrade from ‘Instituto Eduardo Torroja’ of Madrid, for her collaboration in the determination of the concrete pore solution composition by the pore pressing technique.

References (22)

  • C. Andrade et al.

    Mathematical modelling of electrochemical chloride extraction from concrete

    Cement and Concrete Research

    (1995)
  • Cited by (63)

    View all citing articles on Scopus
    View full text