International Concrete Abstracts Portal

Alkali-silica reaction is responsible for concrete cracking, but when initial microcracking is present, does it influence the reaction and, if so, how? This was the problem the authors tried to solve through the following experiments. Four sets of 7 x 7 x 28-cm test concrete bars were prepared with a potentially reactive aggregate. One set was kept as a control, while two others were mechanically microcracked by applying stresses corresponding to 75 and 100 percent of the breaking stress. The fourth set was used to determine the minimum stress that could be applied to the bars. The resulting microcracking was estimated by measuring the ultrasonic wave velocity and by scanning electron microscopy. The evolution of the disorders was tracked by measurement of dimensional variations. The bars were cured at 38 C (100 F) with a moisture content of 100 percent in accordance with standard testing procedure. After 2 years of observation, the authors noted the following developments. The original microcracking had significantly increased the speed of the material's response to the alkali reaction; at the same time, the number of disorders that were consequences of the reaction seemed noticeably higher. Also, cyclic behavior was evident, which induced a dormant stage corresponding to the filling of the microcracking by the reaction gel, and also induced an active stage leading to additional microcracking. Such a sequence of dormant and active stages should affect all the bars tested, but was actually totally evident only on the bars that were initially subjected to significant cracking. This study clearly shows the important role played by initial microcracking on the future of concrete and, consequently, the choice and implementation of solutions that could reduce concrete disorders.

Although aggregates in the concrete matrix are regarded primarily as inert, certain aggregates have been identified as deleterious due to their chemical reactivity in an alkaline environment. Despite extensive research on the various aspects of this problem, a rational model that comprehensively explains the rate of the chemical reaction and resulting expansion has not yet been presented. Paper deals primarily with modeling of the chemical reactions and ensuing expansion in the case of alkali-silica reaction. The chemical reaction phase has been assumed to be governed by the rate of diffusion of hydroxide and alkali ions into the aggregate. The model also assumes the existence of a porous zone around the aggregate and that expansion is initiated only after the amount of reaction products exceeds the volume of this porous zone. An attempt has also been made to discuss some experimental results in the light of the proposed model and provide some of the analytical results arrived at using the model. It was found that by carrying out a slightly modified version of the quick chemical test, the apparent diffusion coefficients of the hydroxide ions can be estimated and the results can be used to accurately estimate the expansion ensuing during the mortar bar tests. Analytical results also indicate that certain characteristic features of alkali-aggregate reaction-related expansion, such as the existence of an incubation period before the onset of expansion, varying rates of expansion, and the shapes of the expansion-time curves, can be explained using the model proposed by the authors.

A test method was developed to predict the risks of ASR expansion for actual field concrete compositions. Concrete prisms (7 x 7 x 28 cm) were cast, demolded, and measured for their initial length. They were then immersed in an alkaline solution at 150 C for 3 weeks in individual stainless steel containers and their lengths were monitored weekly. The concentration of the solution was adapted to match as closely as possible the composition of the interstitial pore solution by summing up the contribution of the binder constituents to the effective sodium and potassium contents, respectively. Despite this fact, the alkali balance before and after treatment of the prisms shows that the concrete is enriched in Na 2O and/or K 2O during the cure. The expansions reached at 150 C after 3 weeks were compared to those obtained at 100 percent relative humidity in air at 38 and 60 C after 12 and 4 months, respectively. Good correlations were obtained for the 67 different concrete mixes tested. Consequently, an expansion value of 0.11 percent was proposed as a provisional limit, which is rather conservative. Reaction products were studied by SEM, optical microscopy, and electron microprobe as gels, and semi-organized and crystallized compounds, and were shown to be present in cracks and pores as well as within the paste. The composition of the products formed at 150 C seems to be restricted to a narrower range of Ca, Si, Na, and K concentrations than what has been reported at lower temperatures.

Petrographic examination and the South African mortar bar test have been performed at SINTEF--Structures and Concrete during the last 2 to 3 years to evaluate the reactivity of Norwegian aggregates to be used in concrete structures. Paper presents the relationships between these two test methods. The purpose of the petrographic examination is to identify, quantify, and group different rock types in an aggregate. These groups are: reactive (with known reactive field performance), potentially reactive, and innocuous aggregates. In Norway, further testing by the mortar bar test is recommended when petrographic examination indicates 20 percent of reactive or potentially reactive rock types in the aggregates. The mortar bar expansion after 14 days of exposure is used for the evaluation of potential expansivity of the aggregates. One main conclusion from the investigation is that mortar bar expansion increases to an upper level with increasing content of reactive rocks in the aggregates. Beyond a "marginal" amount of reactive rocks in aggregates, the mortar bar expansion increases no further. A significant difference in mortar bar expansion between different reactive rock types has not been found. The established limit of 20 percent of reactive rocks in aggregates appears, in most cases, sufficient for classifying aggregates as innocuous; however, no verification of the limit has been made.

The transport of ions related to the penetration of chlorides into concrete has been studied in the field by drilling 100-mm concrete cores from a marine bridge column. A 4-year-old concrete column in Sweden was selected. The concrete was of high quality (i.e., frost- and sulfate-resistant, with a low-heat, low-alkali portland cement with a maximum water-cement ratio of 0.40) according to new Swedish recommendations. Concrete cores were drilled from the submerged, splash, and atmospheric zones. Selective rinding from the concrete surface (profile grinding) revealed concentration profiles of acid-soluble chlorides, carbonates, sulfates, and water-soluble alkalies. Selected parts of the concrete surface were examined by SEM and thin-section microscopy for microstructural studies. Laboratory estimates of chloride diffusivities were carried out on 6-month-old laboratory concrete of similar mix proportions, and also on unexposed parts of drilled concrete cores. Chloride diffusivities obtained from laboratory exposure were then compared with the values obtained from the field concentration profiles, from both the bridge column and a field station, using Fick's second law of diffusion. Maximum chloride diffusivities calculated from the field profiles after 4 years of exposure were more than ten times lower than those obtained from the same concrete in the laboratory. Clearly, there are important mechanistic problems associated with laboratory procedures, resulting in serious misjudgments, if such laboratory tests are used for linear extrapolation of the service life for marine concretes.