To assess the risk of corrosion of reinforcement steel in concrete structures expose to marine or deicing salt environments is required to determine the chloride concentration profile in concrete from the exposed surface inwards. Due to the cost and limitations of methods of sampling for obtaining chloride concentration profiles in concrete, continuous improvement of existing methods and development novel approaches are required.
At the National Laboratory of Civil Engineering (LNEC), Portugal, an ongoing study aims to assess the influence of sampling and of the methodology of collecting concrete powder in determining the chloride content. Slabs of conventional concrete were produced, with and without addition of chloride, to extract cores. From these cores, concrete dust was collected by two methods: cutting and crushing, and dry grinding, with the proportion of aggregate in the test area of each specimen of concrete being evaluated. Knowledge of the chloride concentration profile in concrete from different methods for obtaining dust samples was assessed for establishment of future experimental campaigns to quantify the mass of the concrete dust sample for the considered depth.
Test and main results
For the cutting and crushing method, two slices of each core were cut with 10 mm thickness: one near the concrete surface, at a depth of 2 to 12 mm, and the other at half of the depth, 45 to 55 mm, as shown in Figure 1. These slices were crushed with a jaw crusher, providing a sample of about 100 g of crushed concrete for each slice. The samples were prepared for chemical determination of chloride content, according to an LNEC´s internal methodology. First the size of the sample was reduced through quartering, thus obtaining a smaller sample of about 20 to 35 g. This sample was passed through the 125-μm sieve, and the fraction retained was crushed into a fine powder in a ball mill. This fine powder was passed through the sieve, and the small fraction retained was crushed with the mortar and pestle, until the entire sample passed through the 125-μm sieve.
For the grinding method, the cores were placed in the grinding device, with the commercial name of Profile Grinder Kit – PF1100, of German Instruments A/S (GI), to collect the dust from a 10 mm layer, at depth of 2 to 12 mm, as shown in Figure 2. After grinding this layer, the cores were cut, to allow grinding of another 10 mm layer, at the depth of 45 to 55 mm. This grinding equipment performs a circular grind of 73 mm diameter, thus giving results comparable to the 75 mm diameter of the cutting and crushing method. After the concrete dust was collected from all grinded layers, about 90 g per sample, these samples were passed through the 125 μm sieve, according to according to an LNEC´s internal methodology. Because all the samples from the grinding method consisted in a fine powder, there was no need for quartering. The fraction retained on the sieve was crushed in a ball mill and passed again through the sieve. The small fraction retained was crushed with the mortar and pestle, until the entire sample passed through the 125-μm sieve.
Figure 3 illustrates the variation of the individual results of the chloride contents at 2-12 mm and 45-55 mm depths for both methods used and Figure 4 illustrates the variation of the result of chloride content (mean value of n individual results) as a function of the number of individual results considered (n) for the two methods.
From the analysis of Figure 3 to Figure 4, the following aspects can be emphasized:
The chloride contents obtained in the two depths differ clearly, the higher values having been obtained in the depth 2-12 mm, closer to the surface (Figure 3). This is due to the segregation of the cement paste close to the finishing surface, accumulating and retaining a greater amount of paste and/or mortar.
For the 2-12 mm depth (Figure 4), the difference in chloride contents between the sum of the chloride constituents contribution (0.13%) and those obtained by the cutting and crushing method (0.19%) was 0.06%, and for the grinding method (0.21%) was 0.08%. For the 45-55 mm depth the difference was null, allowing the following conclusions: i) both methodologies (sampling + collecting method) did not interfere on the chloride contents at depths ≥ 45 mm, allowing the correct estimation of the chloride contents; ii) the cement paste segregation inherent to the finishing face loses its effect at depths ≥ 45 mm.
The results at the 45-55 mm depth confirmed that the consideration of the average of all the individual values for each methodology (n=5) led to same value as obtained by the sum of the chloride constituents contribution. This leads to conclude that, for depths where the segregation effects do not occur, the average of at least 5 determinations will lead to a good estimate of the chloride contents.