RILEM Recommended Test Method: AAR-2—Detection of Potential Alkali-Reactivity—Accelerated Mortar-Bar Test Method for Aggregates

Abstract

This draft method was originally prepared by RILEM TC 106-AAR (Alkali-Aggregate Reaction—Accelerated Tests) as TC 106-2, and has been revised by RILEM TC 191-ARP

Annex

(Comments relate to clauses as numbered in method)

1.1 A2 Scope

With some reactive aggregates it has been found that there is a proportion of reactive constituents in the aggregate that leads to a maximum expansion. This proportion is called the “pessimum” content and the relationship between expansion and reactive constituents content is called the “pessimum behaviour” of the reactive aggregate.

1.2 A3 Significance and Use

It is recommended to start all screening of aggregates for their potential alkali-reactivity with a petrographic examination of the aggregate. On the basis of the results of the petrographic examination, a decision can be made with respect to further testing. If aggregates contain more than 2 % (by mass) of porous chert and flint, further testing by means of the accelerated mortar-bar test is not recommended. It is reported (see [16, 17]) that application of the accelerated mortar-bar test to reactive aggregates containing porous chert and flint sometimes gives rise to misleading results and inappropriate approval of such aggregates.

1.3 A7.1 Preparation of the Aggregate Sample

125 μm is the lowest particle size considered in the aggregate sample, both for crushed or natural fines. This is in the light of avoiding discrepancies between the reactivity diagnoses of certain aggregates obtained under different test temperatures. Indeed, at higher temperatures (80 °C), finer particles (e.g. 63 μm or less) may develop a pozzolanic activity that may balance the alkali-reactivity of aggregates, in contrast with what happens at 38 °C (see [18]).

1.3.1 A7.1.2 Crushed Fine Aggregate

Alternatively the equivalent sieves of series B and C of ISO 6274 [2] could be used.

1.3.2 A7.1.6 Combining the Aggregates

If the aggregate sample is composed of:

(a) particles of a single rock, sand, gravel or mineral type that has previously been shown not to have an associated pessimum behaviour and it is intended to use this aggregate unblended; or,

(b) the reactive constituent in a single rock, sand, gravel or mineral type is found at concentrations outside the pessimum limits previously established for the reactive constituent type and it is intended to use this aggregate unblended,

then only a mortar with an aggregate component composed wholly of the test aggregate needs to be tested.

Alternatively, if an aggregate sample is:

(a) composed wholly or in part of a rock, sand, gravel or mineral type which has particles or mineral constituents that have an associated pessimum proportion and the reactive constituents are found in concentrations that are deemed to be potentially reactive, or it is either unknown or not prescribed where the “safe” limits lie; or,

(b) if it has not been established whether the rock or mineral constituent type has or does not have an associated pessimum proportion; or,

(c) if the sample is a blend of aggregate types where the behaviour of the aggregates when combined is unknown; or,

(d) if a petrographic analysis is not available,

then several combinations of test aggregate and non-reactive fine material will need to be tested, as indicated in the following paragraph, in order to identify any pessimum behaviour.

The aggregate to be tested shall be mixed with a reference non-reactive fine material (natural sand or crushed rock) with the same grading as the sample under test. The material used for this purpose shall give an expansion of less than 0.05 % by this test method (See Note A9) and shall not exhibit a pessimum behaviour. It should be preferably a non-siliceous material and previously evaluated. Some suggested proportions of test aggregate and reference non-reactive fine material are shown in Table A.1. As preliminary tests, only three combinations should be used (e.g. I, II and IV in Table A.1). Special aggregates may require some other levels to make sure that any possible pessimum behaviour is identified.

Table A.1 Suggested aggregate proportions

Note A9: This expansion limit is deliberately more restrictive than is generally applied to identify non-reactive aggregates, to ensure that the non-reactive reference aggregate has minimal effect on the results of this test.

1.4 A8 Calculations and Reporting of Results

Criteria for interpreting the results of this test are given in AAR-0, including different maximum percentage expansion criteria for the AAR-2.1 and AAR-2.2 alternative methods.

Recently, an interpretation of the accelerated mortar-bar test results with a kinetic-based method has been proposed. This provides a criterion for defining potentially alkali-reactive aggregates that is an alternative to the maximum percentage expansion criterion.

Correlation of expansion data with a kinetic equation provides a kinetic parameter, ln k, assumed as a reactivity criterion, that is capable of differentiating between reactive and innocuous aggregates regardless of the expansion values obtained at 14 days of immersion of mortar bars in 1 M NaOH solution at 80 °C. This kinetic equation is

$$ \text{E}_{\text{t}} \left( {\% } \right)\text{ = 1 + E}_{{\text{t0}}} \left( {\% } \right) - \text{exp } \left[ {{ - {\rm k}}\left( {{{\rm t} - {\rm t}}_{\text{0}} } \right)^{\text{M}} } \right] $$

(A.1)

where E_t is the percent expansion at the time t (days) of immersion of mortar bars in NaOH solution, E_t0 is the percent expansion at the time t₀ (days) of immersion in NaOH solution, t₀ is the induction period (days) before growth-based kinetics begin, k is a rate constant which combines the effect of nucleation, multidimensional growth, geometry of reaction products and diffusion, and M is an exponential term related to the form and growth of the reaction products. Further details on the development of Eq. (A.1) are given in Johnson and Fournier [13] and Johnson et al. [14].

Solving the logarithmic form of this equation by a least-squares fit yields the reactivity parameter, ln k, as the intercept of the regression line.

A tentative value of three days for t₀ has been proposed in the above references [13, 14], to correlate the accelerated mortar-bar results with Eq. (A.1). Alternatively t₀ may be determined as the value that will give that maximum expansion rate constant, k [15].