ABSTRACT

The measurement of on site corrosion rate is a technique that has been extended for the increasing necessity to determine the service life of concrete structures. The most reliable method is the Polarization Resistance determination by means of the modulated confinement of the current. In present paper, a calibration of corrosion rate measurements with the modulated confinement of the current method (MCC) is given, comparing the electrochemical results with the gravimetric losses of the rebars. This gravimetric comparison is the reference for the correct calibration of the corrosion rate measurement, following Rilem Recommendation of TC-154.

Some concrete slabs with different chloride contents (0%, 0.15%, 0.30% y 0.60% by weight of cement) were made. In some cases, a hole was drilled over some steels for an additional chloride introduction during the test in order to develop very localized attack. Corrosion rate measurements were made using a portable corrosion rate meter with confinement of the current, and with a standard portable potentiostat. Measurements were taken starting on the fabrication day and during four moths. After this period of time, slabs were broken and the bars were cleaned and weighed.

The comparison of electrochemical and gravimetric metal losses confirms the reliability of the modulated confinement of the current method, except when the concrete is very young due to the very low resistivity. Confinement is then not always reliable when the concrete is very wet, and the attenuation of potential method has to be used instead.

1 INTRODUCTION

It is known that reinforcement corrosion is the main distress behind the present concern regarding concrete durability. However in spite of the very numerous papers published on the subject, relatively few are devoted to the development of on-site techniques in general, and even less to the measurement of corrosion. However, it is recognised the importance of an accurate (non destructive on-site) identification of the zones suffering corrosion, in order to make calculations of the residual life based in the loss of reinforcement cross section.

The only electrochemical technique with quantitative ability regarding the corrosion rate measurement in reinforcement is the so called Polarization Resistance, R_p (1). This technique has been extensively used in the laboratory. It is based on the application of a small electrical perturbation to the metal by means of a counter and a reference electrode. Providing the electrical signal is uniformly distributed throughout the reinforcement, the D E/D I ratio defines R_p. The corrosion current, I_corr, is inversely proportional to R_p, I_corr= B/R_p where B is a constant. R_p can be measured by means of D.C. or A.C. techniques (2), both of which have specific features in order to obtain a reliable corrosion current value in agreement with gravimetric losses.

On site measurements: Modulated confinement of the current (guard ring) method.

Direct estimation of true R_p values from DE/D I measurements is usually unfeasible in large real concrete structures. This is because the applied electric signal tends to vanish with distance from the counter electrode, CE rather than spread uniformly across the working electrode, WE. Therefore, the polarization by the electric signal is not uniform, and it reaches a certain distance that is named the critical length, Lcrit.

Hence, DE/DI measurements on large structures using a small counter electrode provides an Apparent polarization resistance (R_p^app) that differs from the True R_p value depending on the experimental conditions (3). Thus, if the metal is actively corroding, the current applied from a small CE located on the concrete surface is 'drained' very efficiently by the rebars and it tends to confine itself on a small surface. Conversely, if the metal is passive and R_p is high, the current applied tends to spread far away (e.g., around 50 cm) from the application point. Therefore, the Apparent R_p approaches the True R_p for actively corroding reinforcement. But when the steel is passive, the large distance reached by the current necessitates a quantitative treatment.

There are several ways of accounting for a True R_p value, among which the most extended one is the use of a guard ring (3), in order to confine the current in a particular rebar area, as Figure 1 depicts. The measurement is made by applying a galvanostatic step, lasting 30-100 seconds, from the central counter. Then, another counter current is applied from the external ring, and this external current is modulated by means of the two reference electrodes called "ring controllers" in order to equilibrate internal and external currents, which enables a correct confinement, and therefore, calculation of R_p. By means of this electrical delimitation to a small zone of the polarized area, any localised spot or pit can be first, localised, and second its measurement can be made by minimising the inherent error of R_p.

However, not all guarded techniques are efficient. Only that using a "Modulated Confinement" controlled by two small sensors for the guard ring control placed between the central auxiliary electrode and the ring, shown in Figure 1, is able to efficiently confine the current within a predetermined area. The use of guard rings without this control leads into too high values of the I_corr for moderate and low values, and the error introduced in the case of very localized pits, is very high.

Fig 1: Corrosion rate measurement with the modulated confinement of the current method.

There are on the market several devices reporting to measure in large structures the corrosion rate, I_corr (4-5), however those not using a modulated confinement only can give more or less accurate I_corr values when the reinforcement is actively corroding (6). Without modulated confinement the I_corr values result too high, > 0.1 mA/cm² in passive state.

Rilem TC-154 has prepared a Recommendation (6) for the measurement of R_p on-site. There it is suggested to use slabs of a certain size for checking the reliability of available devices. In present paper results are given on such suggested calibration for the device Gecor 06 using modulated confinement.

2 EXPERIMENTAL

Fig 3: Corrosion rate measurement on one tested stab using the Gecor 06.

Three slabs with different chloride contents added on the mixing water were fabricated. Sodium chloride was used in concentrations of 0.15, 0.30 and 0.60% of chloride ion by weight of concrete. On this way localised corrosion mechanism was activated since fabrication time.

Bars were previously cleaned and weighed. The exposed area was limited using an insulating tape. Bars 1 to 6 where embedded in the slab with 0.15% of chloride, 7 to 12 in the slab with 0.30% of chloride, and 13 to 18 in the slab with 0.60% of chloride.

Dimensions of these slabs were 40x35x6 cm, and there were six corrugated carbon steels (500 Mpa) embedded on each one (6mm diameter and 45 cm length), as Figure 2 and Figure 3 show. Bars were placed in parallel separated 5 cm, and externally connected by a wire.

Fig 2: Holes disposition over steels 4 and 2 on slabs 0.15 y 0.60 % de Cl-.

Apart from this, on slabs with 0.15% and 0.6% two holes were drilled over bars number 2 and 4 for an extra addition of chloride every 15 days (CaCl₂) (Figure 2). On this way, it is possible to generate a very localized corrosion in these zones in order to evaluate the effect and how the corrosion rate meter is able to detect the generated pits.

Another slab of a larger size was prepared without chlorides. The larger size is needed due to the polarized area is much more extended when the steel remains passive. Then, a slab of 100x85x8 cm in size with 9 bars embedded was as well prepared.

Corrosion rate, I_corr, and simultaneous Corrosion potential, E_corr, and electrical resistance, R_e, were measured on each slab during more than 100 days.

The measurements were made using two different equipments:

• A standard potentiostat used for R_p measurements in the laboratory. This potentiostat was used with a three electrode system, connecting as a working electrode, the bar embedded, as a counter electrode another parallel bar and using as reference electrode the saturated calomel one (SCE).
• For the measurements of the modulated confinement of the current technique, the portable corrosion rate meter Gecor 06 was used (Figure 3). It uses a Cu/CuSO₄ electrode as reference ones. Therefore the E_corr values differ of about 80 mV between the results of the normal potentiostat and those of Gecor 06.

2.1 Sequence of measurements
All the bars in the four slabs were connected by a wire from the casting of the concrete. When the standard potentiostat was used, the bars are disconnected and one bar is taken as working electrode while the parallel one as counter electrode. The reference electrode is placed over and in the middle of the bar acting as working electrode. When finished the measurement, the operation is repeated in the rest of the bars. They are finally again connected until next measurement. In the case of Gecor 06, the bars are mantained connected all the time. The locations were Gecor 06 was placed are shown in Figure 4. In each location the amount of rebar introduced in the calculations of Icorr may be different, because sometimes the circle bellow the device encounters one or two bars. The mean I_corr of each bar is calculated by averaging the four measurements made on each bar. This averaged I_corr of each bar made by means of Gecor 06 is compared with the value obtained throught the standard potentiostat.

Finally, the comparison between electrochemical and real weight loss was made making a gravimetrical study of the bars after finishing the test. At this moment, bars were cleaned and weighed, and the weight loss was compared with the electrochemical weight loss calculated by the integration with time of the I_corr values, using Faraday law. The electrochemical loss from the standard potentiostat is the averaged value of the whole bar, while that obtained from Gecor 06, results of averaging the four measurements on each bar (Figure 4).

Fig 4: Points identification for the measurements with Gecor 06.

3 RESULTS AND DISCUSSION

Just after casting the concrete, Corrosion rate values measured in all the slabs containing chlorides with both methods are over 10 mA/cm². This is due to the low concrete resistivity when the concrete is fresh. Then, after a few days of curing, concrete resistivity increases, and corrosion rate values decrease, starting to be more or less stable. In general, as higher is the chloride content, higher results the corrosion rate measured. When an additional concentration of chloride is added on the holes drilled on the concrete surface, an increase of the corrosion rate value can be detected

At the initial times, differences were found between both electrochemical methods (the standard potentiostat and the corrosion rate meter with modulated confinement of the current), as Figure 5 shows. They were attributed to the fact that, the standard potentiostat obtained an averaged value for the whole bar and the measurement is made in the bars electrically isolated while with Gecor 06, all the bars remain connected, and is able to pick better the localized attack.

Fig 5: Corrosion rate vs. time measured with the standard potentiostat (bars isolated) and the Gecor 06 (bars connected) on the slabs with 0.30 % of chloride content.

An example of I_corr measurements taken on slabs with 0.30% of chloride is shown on Figure 5. On the left side, standard potentiostat results are shown, while on the right side results corresponds to the Gecor 06.

In Figure 6, weight losses obtained on the three tested slabs are shown. With Gecor 06 the agreement found was quite good while the results of electrochemical weight losses obtained from the potentiostat were too low, indicating that the low I_corr values recorded at the initial times after casting are not representing the behaviour of the bars when connected. In addition the relatively large size of the bars does not allows the potentiostat to pick the high corrosion rate at the pits growing on the holes made.

Fig 6: Electrochemical and gravimetrical weight losses. Correction applying a pitting factor.

In order to try to improve the comparison, corrosion rate values measured in the areas closed to the extra chloride addition were recalculated. On these zones, pits detected during the visual inspection were deeper than in the rest of the bars. In these cases, I_corr values in these points were underestimated in a factor around 10. Error in Gecor 06 is smaller than in potentiostat because the first one confines de current in a small area, and the corrosion rate value for one bar results from the average of the values measured on different points, while the potentiostat polarize all the bar, averaging the corrosion rate of the steel, without localizing the pits.

Also, on Figure 6 are shown results after recalculating the I_corr values in points affected by the pits. This correction was made multiplying by 10 the corrosion rate values of the bars with the pits. In Gecor 06 results, only were corrected the measurement points over the pits, but in the case of the potentiostat all the affected bar was corrected (multiplying by ten).

After the correction, electrochemical losses are compared again with the gravimetrical results (Figure 6 right side). In spite the potentiostat results improve, the Gecor 06 still provides better agreement with gravimetrics.

This confirms that the modulated confinement of the current method for corrosion rate measurements on site is reliable, and only it is necessary to make correction by multiplying by ten when very localized pits are generated.

As was mentioned on the experimental part, another slab of a larger size was prepared without chlorides. In this case, was observed that during the first days after the fabrication, the modulated confinement of the current method provided very high values, while the standard potentiostat provided coherent low I_corr values. This high values measured with Gecor 06 were due to the very low concrete resistivity at this time, so, in fresh concrete modulated confinement method does not provide an accurately confinement, and the I_corr value is higher than was expected.

Therefore, another corrosion rate measurement method was tested. This is called Potential Attenuation Technique, and is included on the Gecor 08 device. As is shown on Figure 7, potential attenuation technique provides similar results than the standard potentiostat, so, could be a good alternative on site in very low concrete resistivity structures.

Fig 7: Corrosion rate measured on the slab without chlorides.

When the concrete is very wet, its resistivity may be so low that the confinement by the guard ring of the current cannot be well achieved because the area polarized is very large. For these conditions, another measurement method has been developed, the so called measurement of the potential attenuation with the distance, which is based in the direct measurement of the critical length. The sensor consists in a small disc acting as the only counter electrode, which has in its centre the reference electrode for the recording of E_corr. Other three reference electrodes are placed aligned with the E_corr one at fixed distances.

For the measurement, a potentiostatic step, lasting between 30-100 seconds, is applied to the bar. This applied potential step attenuates with the distance. From the distance (L_crit) reached by the signal and certain geometrical considerations of the bars diameter, it is possible to calculate the true Rp (referred to a particular steel area).

This method is not applied for normal non-wet concretes due to it cannot, in these cases, localize well the isolated corroding areas (pits).

4 CONCLUSIONS

1. Modulated confinement of the current method provides similar values than the gravimetrical technique, even without making any correction in pitting areas. Maximum error detected for this technique without applying the pitting factor is 53%. A standard potentiostat gives less similar results due to the large size of the exposed area.
2. After the correction of corrosion rate measurements (pitting factor = 10), confinement measurements are closer to gravimetrical results than before the correction. The error on this case is around 7%. The corrected values for potentiostat measurements without current confinement are better but still too low, being the error in this case around 26%.
3. In passive bars, the standard potentiostat and the Potential Attenuation method are more reliable at the initial times after casting. This is due to the so low concrete resistivity on this moment. Modulated confinement gives correct results when the resistivity increases due to the hardening.

5 REFERENCES

1. ANDRADE, C. AND GÓNZALEZ, J.A., "Quantitative measurements of corrosion rate of reinforcing steels embedded in concrete using polarization resistance measurements", Werkst. Korros., 29, 515 (1978).
2. ANDRADE, C., CASTELO, V., ALONSO, C. AND GONZÁLEZ, J.A., "The determination of the corrosion rate of steel embedded in concrete by the Rp on A.C. Impedance methods," ASTM-STP 906, pp. 43-64. (1986).
3. FELIÚ, S., GONZÁLEZ, J.A., FELIÚ, S.JR., AND ANDRADE, C., "Confinement of the electrical signal or in-situ measurement of Polarization Resistance in Reinforced concrete," ACI Mater. J., 87, pp 457. (1990).
4. ELSENER B, KLINGHOFFER O, FROLUND T, RISLUND E, SCHIEGG Y, BÖHNI H. "Assessment of reinforcement corrosion by means of galvanostatic pulse technique" Proc. International Conference Repair of Concrete Structures, ed. A. Blankvoll, Svolvaer Norway p.391. 1997.
5. KLINGHOFFER O, KOFOED B, FROLUND T. "Monitoring of reinforcement corrosion by means of embedded sensors and portable polarization technique". International conference: Corrosion and Rehabilitation of Reinforced Concrete Structures. Orlando, December 1998.
6. ANDRADE C, ALONSO C "Test methods for on-site corrosion rate measurement of concrete reinforcements by means of the polarization resistance method". Rilem TC-154-ECM: "Electrochemical techniques for measuring corrosion in concrete". 2000.