Accelerated Expansion Test Sample Report: Switzerland

Concrete Slab

General Information

A concrete slab containing reactive aggregates was stored at 27°C and 65% RH after production for a few month. Afterwards, the temperature of the storing room varied between 18 and 24°C at approximately 40% RH. The slab was three years old and its average expansion was 0.24% at the time of coring.

The gravity concrete dam is given as a second example for the application of the Swiss method. No details on the several decades old structure are given as it should not be identifiable. The investigation of the dam was triggered, because deformations that indicated concrete expansion were registered during monitoring.

There are only few typical AAR-cracks on the surface of the dam. However, a crack spans the entire length of the gallery located in the interior of the dam.

Only part of the analysis is shown as the example given in paragraph 2 already demonstrates, how the residual expansion measurements can be complemented with microscopy. The purpose of this example is to show the approach to assess the spatial distribution of damage and residual expansion potential in a structure. The horizontal cores taken from the dam (see the following paragraph) are used to show this approach.

Materials and Methods

Four concrete cores (diameter: 100 mm, length: 200 mm) were used for the investigation. Three cores were tested according to the Swiss method.

In order to get additional information on the state of ASR in the concrete at time of delivery, one core was cut and five subsamples were prepared (~50 × 90 mm 2 ). They were dried in an oven at 50°C for three days, epoxy impregnated and polished. The crack-index was determined using a Zeiss Axioplan microscope at a magnification of 100×. The samples were further documented under fluorescent light using a 24 MP camera. As an additional step, the impregnated subsamples were used to produce four samples (discs with a diameter of 50 mm) for scanning electron microscopy (SEM). They were investigated with a Nova NanoSEM 230 FEI at a pressure of 2.0-4.0 × 10 −6 Torr at an acceleration voltage of 12 kV and a beam current of 80-88 µA. An Qxford SSD detector and INCA Energy software with ZAF correction were used for energy-dispersive X-ray spectroscopy (EDS). The concrete after the test was studied in an identical way. Four subsamples were prepared (~50 × 90 mm 2 ) and investigated with the optical microscope. Two samples were produced for studying the concrete with SEM.

The water-soluble alkali content was determined before and after the residual expansion test according to the procedure described in [1,2].

Horizontal and vertical cores with a diameter of 100 mm were taken in three parts of the dam. The vertical and horizontal cores had a length of about 11 m and 5 m, respectively. The samples were sealed immediately after coring to avoid changes in their moisture state. Samples for microscopy were prepared from all cores at different concrete depth always using an 11 cm long section of the concrete core. This allowed to produce thin Sects. (50 × 90 mm 2 ) parallel and perpendicular to length axis of the cores (total of 54 thin sections). Impregnation followed the same protocol described in paragraph 2.2. Crack indices and percentages of aggregates exhibiting cracks typical for ASR were determined.

Data from the vertical cores, results on microstructure acquired by SEM and on water soluble alkali are not shown.

Residual Expansion

The cores show a relatively fast expansion ( Fig. 14.1). The longitudinal expansion (value used for assessment) during conditioning (20°C, 100% RH) is above 0.03%. The irreversible longitudinal expansion at the end of the test of 0.11% indicates that about 0.02% of the swelling during the conditioning is irreversible and must be attributed to ASR. The end of the fast expansion (phase 2) has been reached very quickly after moving the samples into the reactor (temperature of 38°C and 100%

Alkali Content

The water soluble alkalis decrease during the test (Table 14.2). Alkalis can either be leached or bound in ASR products.

Optical Microscopy

General Impression

The concrete shows already pronounced cracking in its delivery state ( Fig. 14.2). The cracks are initiated in aggregates of the sand fraction. Cracks in aggregates >4 mm are practically absent. The cracking is more pronounced after the residual expansion test. In particular, aggregates >4 mm display cracks as well, even when the cracking is much less pronounced than in the sand fraction. The majority of cracks, especially in the cement paste, are filled with ASR products after the test.

cm

Crack-Index

The crack-index was determined both before and after the residual expansion test. The values are presented in Table 14.3. The difference between the crack-indices is 0.17%.

Scanning Electron Microscopy

The samples prepared for SEM ( Fig. 14.3) show cracking restricted to aggregates of the sand fraction (Figs. 14.4 and 14.5) in the state of delivery, as already observed in the optical microscope. Some of the aggregates in the sand fraction contain alkalisilica-glass (determined by EDS) providing a highly reactive phase and a source of alkalis. The majority of the cracks both in aggregates and in the cement paste (width up to 30 µm) are empty or only partly filled with ASR products. There are aggregates of sedimentary, volcanic and magmatic origin, but they will not be described in detail in the following.

Cracking of the concrete is more pronounced with wider cracks up to a width of 60 µm after the residual expansion test. The majority of the cracks in the cement paste are now filled with ASR products (Fig. 14.6). The porosity of the reactive sand grains is increased by further dissolution. Aggregates >4 mm show occasional cracks containing ASR products as well (Fig. 14.7).

Summary and Conclusions

The current experiences with the determination of the residual expansion measurements indicate that the expansion during phase 2 is connected with expansion of the concrete since occurrence of ASR until the time of core extraction [2,3]. Consequently, the expansion during phase 2 indicates that the expansion in the concrete BL A horizontal BL B horizontal BL C horizontal close to the surface has been larger compared to the interior of the dam. This can have two different reasons. First, the concrete in dams often contains more cement and was produced with a lower w/c resulting close to the surface resulting in a higher alkalinity of the concrete pore solution and a higher degree of ASR. Secondly, the exposure of the dam to the south (air side and location of the coring sites) leads to increased temperature during the day increasing SiO 2 solubility and ASR. However, there are no indications detected in microscopy for a significant difference in concrete composition. In any case, the higher expansion rate close to the surface goes together with a higher degree of ASR in this part of the dam, as clearly indicated by the relatively high number of reacted aggregates and the relatively high crack-indices. The expansion during phase 3 represents the kinetics of expansion of the tested concrete [2,3]. Because there is no clear difference in expansion as a function of concrete depth, the degree of ASR-induced damage and the higher expansion rate during phase 2 is likely attributable to the high daily temperature caused by sun radiation. A higher cement content and lower w/c should have resulted in higher expansion during phase 3 of the concrete close to the surface.

Compared to the expansion rates of concrete from other structures, the values for this specific dam are low (Fig. 14.9). This correlates well to the age of the dam and the relative low degree of cracking observed on the concrete surface. Although cracking of the concrete is only small to moderate, many of the cracks as observed in microscopy are not filled with ASR products. This is another sign for a very slow reaction. But it additionally shows that ASR has not reached a "ripe" stage (indicated by reaction products filling cracks and voids) but it still progressing.

In order to model the behavior of the dam in the future, the expansion rates during phase 3 should be used, but the likely temperature-driven higher expansion on the south side should be taken into account. Even if the expansion rates as determined during phase 3 have to be higher than in reality due to the elevated temperature of 38°C, modelling could identify the areas of the dam where problems will likely occur. As the dam is regularly monitored, these in situ measurements could provide a valuable benchmark for the modelling.

Test Results

The concrete expands considerably in the residual expansion test. As typical, the rate of longitudinal expansion in phase 3 is lower compared to phase 2. The irreversible expansion indicates a total longitudinal expansion during the test of 0.11%. Based on the longitudinal expansion during phase 3, an expansion rate of 0.20%/year can be calculated. This is the value to be used as a starting for assessing the expansion of the investigated concrete component. If the studied cores originated from a concrete structure, the boundary conditions like temperature, humidity and stress state would have to be taken into account in an overall assessment.

The more or less constant expansion during phase 3 indicates an internal source of alkalis, as a decrease of expansion with time is usually observed due to leaching. This seems to be confirmed by the presence of aggregate particles of volcanic origin containing alkali-silica-glass.

Microscopy indicates that the very reactive aggregates are present mainly in the sand fraction. The concrete shows already obvious cracking caused by ASR in its state of delivery. The cracking is more pronounced after the test. The difference in the crack index before and after the test is 0.17%. This value should approximately correspond with concrete expansion during the test. Obviously, there is a certain overestimation of expansion by the crack-index in the current case (measured values of 0.11 and 0.14% in longitudinal and diametral direction, respectively), even if the expansion is in the same order of magnitude.

Comparison with Data Base

In comparison with expansion of cores extracted from Swiss concrete structures, the tested laboratory concrete expands fast and reaches very high expansion values (Figs. 14.8 and 14.9). • concrete with a residual expansion potential zero or very low.

• concrete with a moderate residual expansion potential.

• concrete with a high residual expansion potential.

The tested concrete can be assessed as a concrete with fast kinetics and obviously a high residual expansion potential. The expansion rates during phases 2 and 3 are both higher than the ones reported from Swiss concrete structures (Fig. 14.9). Here, it has to be mentioned, that Swiss aggregates are relatively slow reacting. Volcanic rocks containing glass that are present in the studied concrete (see Sec. 14.2.6) are absent.

Gravity Concrete Dam

Results

Close to the south-ward directed surface of the dam the relative number of reacted aggregates in blocks B and C is significantly higher compared to the interior of the dam (Fig. 14.10). There is no such trend visible in block C. The same applies to the crack-indices that are clearly higher close to the surface of the dam in blocks B and C than in the interior (Fig. 14.11) confirming a more pronounced AAR in this area.

The expansion determined in the residual expansion test during phases 2 and 3 was used to calculate an expansion rate for the concrete. The highest expansion rates in all three horizontal cores during phase 2 occur close to the surface (Fig. 14.12). The values during phase 3 are about 10 times lower compared to phase 2 without showing a systematic trend as function of concrete depth (Fig. 14.13). The values in block C during phase 3 indicate somewhat faster kinetics compared to the other block.

Table 14 .

Table 14