Elsevier

Cement and Concrete Research

Volume 76, October 2015, Pages 107-120
Cement and Concrete Research

A thermodynamic and kinetic model for paste–aggregate interactions and the alkali–silica reaction

https://doi.org/10.1016/j.cemconres.2015.05.004Get rights and content

Abstract

A new conceptual model is developed for ASR formation based on geochemical principles tied to aqueous speciation, silica solubility, kinetically controlled mineral dissolution, and diffusion. ASR development is driven largely by pH and silica gradients that establish geochemical microenvironments between paste and aggregate, with gradients the strongest within the aggregate adjacent to the paste boundary (i.e., where ASR initially forms). Super-saturation of magadiite and okenite (crystalline ASR surrogates) occurs in the zone defined by gradients in pH, dissolved silica, Na+, and Ca2 +. This model provides a thermodynamic rather than kinetic explanation of why quartz generally behaves differently from amorphous silica: quartz solubility does not produce sufficiently high concentrations of H4SiO4 to super-saturate magadiite, whereas amorphous silica does. The model also explains why pozzolans do not generate ASR: their fine-grained character precludes formation of chemical gradients. Finally, these gradients have interesting implications beyond the development of ASR, creating unique biogeochemical environments.

Introduction

Expansion due to reaction between cement and aggregate has been recognized for over 70 years as a cause of premature degradation in concrete [1]. Although various deleterious reactions have been observed, the alkali–silica reaction (ASR) remains the most infamous, forming an alkali-rich silica gel that swells on hydration and that is associated with cracking tied to premature degradation of the concrete. ASR has a worldwide distribution [2], occurring over time in concretes that contain both available alkali (sodium and/or potassium) and certain types of silica-rich aggregates [3], [4].

For the almost sixty years since its presentation, the conceptual model of Powers and Steinour [5], [6] for ASR has remained the basis for most interpretations of the mechanisms by which the reaction proceeds. Although some aspects of the model have been refined by subsequent studies, the model remains largely intact [7].

A central tenet of the Powers and Steinour (P&S) model is that ASR initiates by an attack of silica-rich phases (such as opal, chert, strained quartz, volcanic glass) by alkalis in the pore fluids from the cement paste: specifically, alkali cations and hydroxyl anions diffuse into the silica-rich phase, disrupting the silica framework to form what is termed “ASR gel”. In this model, the dissolution kinetics of these reactive silica phases generates silica gel whereas unstrained quartz is unreactive. This model successfully explains many observations of ASR, including the following two:

  • Often, gel forms initially within the bounds of the aggregate, producing the darker reaction rim on the aggregate that is one characteristic of ASR as well as forming gel in fractures that transect the aggregate.

  • Gel within the bounds of the aggregate is enriched in alkalis and depleted in calcium, despite the fact that calcium dominates the chemistry of the hydrated paste outside the bounds of the aggregate.

In this study, we offer an alternative conceptual model for ASR based on geochemical principles and tools that have evolved since the Powers and Steinour model was introduced. Our conceptual model is based on dissolution and precipitation driven by an acid–base disequilibrium across the cement–aggregate interface; this disequilibrium establishes a chemical gradient that drives dissolution and transport processes on both sides of the interface. The net result is the formation of a geochemical microenvironment defined by an accumulation of aqueous silica and alkali within the aggregate boundary and by a gradient in pH and aqueous silica species, ultimately resulting in the super-saturation and precipitation of ASR gel.

In developing this conceptual model, we rely on calculations from reactive-transport simulations [8], [9], focusing on simple, well-documented physical processes: dissolution, precipitation, equilibrium aqueous speciation, and diffusion. We introduce the conceptual model by focusing on the simplified cement system CaO–SiO2–H2O. The use of a simplified system allows us to explore quantitatively a few key processes that will occur in both the simple and full systems. Additional components (Na+, Cl, HCO3, and SO42 ) are added to the simplified reference-case to develop the conceptual model with respect to other solid phases and fluid species. In developing the model, we directly evaluate the impact of two key uncertainties: dissolution rate (including reactive surface area) and C–S–H solubility (including both Ca:Si ratio and differences in published solubility products). Although the lack of thermodynamic information on ASR gels precludes addressing the formation of gels directly, conceptual aspects of gel formation are explored using saturation of crystalline surrogates for ASR gel and considering mechanisms of silica gelation. We do not, however, consider other possible crystalline surrogates for ASR gel (such as analcime [10] or erionite [11]), which have also been reported as crystalline products associated with ASR; these would require adding an additional chemical complexity to the system (e.g., Al3 +).

Like the original model of Powers & Steinour [5], [6], the geochemical model presented here is able to explain the two key observations of ASR described above. In addition, it provides a sound physical basis for explaining the roles of silica mineralogy (i.e., amorphous silica vs. unstrained quartz) and of grain size (e.g., problematic siliceous aggregate vs. beneficial, fine-grained siliceous pozzolans), as well as for understanding the impact of lithium on gel formation. Importantly, our model has distinct differences in the driving mechanisms for ASR gel formation, which, we hope, will lead to new insights on predicting and controlling ASR. More specifically, dissolution occurs on both sides of the cement–aggregate interface, stressing the role of reactions in the paste in the development of ASR gel, particularly the role of dissolution of calcium–silicate–hydrate paste. An outcome of this is that pH is buffered in the paste, allowing a gradient to develop inside the aggregate. Although the acid–base disequilibrium between paste and aggregate has been noted widely in ASR research, it's implications for the mechanisms of ASR have not been fully explored. In our model, this gradient has numerous implications, including the establishment of silica gradients that drive diffusion, the shifting of aqueous (alkali-) silica species across the cement–aggregate boundary, and the development of stability fields within the aggregate for ASR reaction products. In general, this pH gradient drives the development of unique geochemical environments at the boundary between paste and aggregate that could have implications beyond the formation of ASR (including, for example, the formation of microbial niches, as discussed briefly below).

Section snippets

Modeling approach

Pan et al. [12] present a comprehensive review of computational modeling that has been applied to ASR. Although several studies have attempted to couple the chemical and mechanical effects in order to explain both the reaction and the associated expansion, most have followed the basic chemical mechanisms of Powers and Steinour [5] and Dent Glasser and Kataoka [3], assuming the diffusion of alkali and hydroxyl to the reacting silica-rich particle.

We follow a different approach in this study,

Dissolution and diffusion for reference case

The reference case scenario examined the evolution of the system when the rates of moles released for each phase (k·A) are equivalent for the two rate-limiting phases (C–S–H and amorphous silica) and for the case of Ca:Si = 1.70 for C–S–H. The stoichiometry for these phases was set such that both had one mole of silica per mole of mineral (i.e., the silica release rates are equivalent); however, the molar volumes for the two phases differ significantly (124.1 cm3 mol 1 for C–S–H (A); 29 cm3 mol 1 for

Discussion

The basic geochemical model presented here entails an acid–base reaction between the aggregate and the paste, resulting in dissolution and precipitation at the boundary between the two regions. In general, the acidic silica phases in the aggregate react with the basic phases in the paste (C–S–H and portlandite) through diffusion in the fluid within fractures and/or pores.

The two regions have equilibrium pH's that are distinct: fluids in equilibrium with the paste have pH > 12, whereas fluids in

Acknowledgments

We would like to thank: R. Grover and J. Barela of the New Mexico State Highway and Transportation Department for providing core samples of ASR-impacted concrete as well as enlightening discussions on ASR; C. Nicholson-Guthrie and G. Guthrie for insights into the use of ninhydrin; Satish Karra for assistance installing and running PFLOTRAN. We would also like to thank two anonymous reviewers for comments and suggestions on an earlier version of the manuscript; in particular the suggestion by

References (48)

  • J.D. Rimstidt et al.

    The kinetics of silica–water reactions

    Geochim. Cosmochim. Acta

    (1980)
  • J.J. Mazer et al.

    Dissolution kinetics of silica glass as a function of pH between 40 and 85 °C

    J. Non-Cryst. Solids

    (1994)
  • A. Seidel et al.

    The kinetics of dissolution of silica ‘Monospher’ into water at different concentrations of background electrolyte

    Solid State Ionics

    (1997)
  • D. Wolff-Boenisch et al.

    The effect of crystallinity on dissolution rates and CO2 consumption capacity of silicates

    Geochim. Cosmochim. Acta

    (2006)
  • D. Wolff-Boenisch et al.

    The dissolution rates of natural glasses as a function of their composition at pH 4 and 10.6, and temperatures from 25 to 74 °C

    Geochim. Cosmochim. Acta

    (2004)
  • A.W. Harris et al.

    Testing of models for the dissolution of cements—leaching of synthetic C–S–H gels

    Cem. Concr. Res.

    (2002)
  • J.J. Chen et al.

    Solubility and structure of calcium silicate hydrate

    Cem. Concr. Res.

    (2004)
  • C.J. Benmore et al.

    The structure of alkali silicate gel by total scattering methods

    Cem. Concr. Res.

    (2010)
  • G.D. Guthrie et al.

    A simple environmentally friendly, and chemically specific method for the identification and evaluation of the alkali–silica reaction

    Cem. Concr. Res.

    (1997)
  • C. Tremblay et al.

    Experimental investigation of the mechanisms by which LiNO3 is effective against ASR

    Cem. Concr. Res.

    (2010)
  • X. Feng et al.

    Summary of research on the effect of LiNO3 on alkali–silica reaction in new concrete

    Cem. Concr. Res.

    (2010)
  • W.L. Marshall et al.

    Amorphous silica solubilities—II. Effect of aqueous salt solutions at 25 °C

    Geochim. Cosmochim. Acta

    (1980)
  • S. Multon et al.

    Effects of aggregate size and alkali content on ASR expansion

    Cem. Concr. Res.

    (2010)
  • M. Gutiérrez-Padilla et al.

    Biogenic sulfuric acid attack on different types of commercially produced concrete sewer pipes

    Cem. Concr. Res.

    (2010)
  • Cited by (20)

    • A 3D reactive transport model for simulation of the chemical reaction process of ASR at microscale

      2022, Cement and Concrete Research
      Citation Excerpt :

      In available ASR thermodynamic models [39,41,87], magadiite, okenite and kanemite have been used as surrogates for low Ca/Si ASR gel based on their similarities in composition. We have recalculated their equilibrium constants based on the thermodynamic data in [39,41] with H2SiO4 or SiO2(aq) and H+ transferred to the used ion H2SiO42− and OH− in this paper. The calculated equilibrium constants are 1 × 10−3, 1 × 10−18 and 1 for magadiite, okenite and kanemite respectively.

    • The role of calcium on the formation of alkali-silica reaction products

      2019, Cement and Concrete Research
      Citation Excerpt :

      First attempts at applying thermodynamic models to ASR have been made by Kim and Olek [26], and by Guthrie and Carey [27]. However, due to the lack of experimental solubility measurements and derived solubility products, Kim and Olek used the estimated solubilities for two hypothetical ASR products (K2Ca4Si6O17·10.5H2O and Na2Ca4Si6O17·10.5H2O) to describe the sequence of ASR [26], while Guthrie and Carey used the thermodynamic data for magadiite (NaSi7O14·4.5H2O) and okenite (CaSi2O5·2H2O) as surrogates for an alkali-silicate and a high calcium-silicate ASR products [27]. These calculations showed that it is in principle possible to predict the conditions for formation of ASR products, but also highlighted the need to use realistic chemical compositions and solubility products to predict the conditions for formation of ASR products.

    View all citing articles on Scopus
    View full text