Ability of the R3 test to evaluate differences in early age reactivity of 16 industrial ground granulated blast furnace slags (GGBS)

Abstract

Ground granulated blast furnace slag (GGBS) is a glassy by-product of pig iron production and is commonly used in concrete industry to replace cement and thereby lower the carbon footprint of the material. Large variations in reactivity exist depending on the GGBS physical and chemical features. Here we investigate the ability of three rapid calorimetric methods to evaluate the reactivity of GGBS. On a set of 16 industrial GGBS, we show that 24 h heat release, using the R3-protocol, correlates well with 2d compressive strength of standard mortars using 75 wt% GGBS. The correlation of R3-test results (R² = 0.87) is better than for traditional reactivity indices calculated from chemical composition. Furthermore, we present data on the repeatability of the test protocol and show that the R3-protocol is very sensitive to sample fineness. Finally, XRD patterns show that slight differences in phase assemblage exist between the most and least reactive GGBS.

Introduction

Ground granulated blast furnace slag (GGBS) is a glassy by-product of pig iron production and has been used in blended cements for over 140 years [1]. Besides economic aspects the use of GGBS at high substitution levels has the advantage of lowering CO₂ footprint of concrete, decreasing heat development during setting and increasing physical and chemical resistance of the material [[1], [2], [3], [4], [5], [6]]. However, short term compressive strength development of GGBS is below that of original Portland cement (OPC), but reactivity under same conditions, can differ widely between different GGBS [1,4,[7], [8], [9], [10]].

The intrinsic reactivity of GGBS depends on different parameters as chemical composition, glass content and structure or quenching parameters of the slag [1,9,[11], [12], [13]]. This makes modification of production parameters of GGBS – so called upstream modification - a promising route to increase its reactivity [[13], [14], [15]]. Another way to increase the reactivity of GGBS is the modification of the activation system (downstream) by increasing fineness, increasing reaction temperature, adapting solution chemistry or by adding chemical activators as CaCl₂ [[16], [17], [18]].

The present work originates in the European research project Actislag, aiming to develop a second generation GGBS reaching cement substitution rates of 80 wt% while keeping the specifications of CEM II. Therefore, both activation routes are considered: upstream modifications of GGBS chemistry/microstructure and downstream modifications of the activation system. In order to reach the research goals, it is essential to be able to quickly assess the reactivity of a large number of GGBS. Furthermore, as upstream modifications will be tested in lab scale trials, minimal sample mass should be used for the reactivity tests.

A multitude of test methods are available to predict the reactivity of supplementary cementitious materials (SCM). Usually chemical tests are used to predict short term compressive strength, see Snellings and Scrivener (2016) and Li et al. (2018) for a review [19,20]. Standardized test methods use Ca consumption during hydration of the SCM (e.g. Chapelle test, Frattini test). These were designed for pozzolanic material and are difficult to apply to slags, as GGBS contain a substantial amount of reactive Ca [19]. Further methods aim to determine the degree of reaction of SCMs: most notably selective dissolution, SEM-IA, XRD, and NMR (²⁷Al or ²⁹Si) are presented in scientific literature but are often time consuming and with moderate accuracy at early age [21].

In recent publications heat development during hydration of a simplified system showed good correlation with compressive strength of different SCM containing mortars [19,20,22]. In these studies, a sulphate containing alkaline solution is added to a mix of portlandite, carbonates and the respective SCM. The heat development during the following reaction is measured by isothermal calorimetry. The reaction can be accelerated by increasing the temperature conditions. This setup avoids the use of other hydraulic material as OPC, and thus the risk of overprinting the signal of intrinsic SCM reactivity [[23], [24], [25]]. By this protocol a wide range of SCMs including GGBS, fly ash, natural pozzolans and calcined clays were tested and classified according to their reactivity. In Avet et al. (2016) the simplified test was used to predict the pozzolanic activity of 7 calcined clays, naming it R3 test [22]. In that study a good prediction of compressive strength of mortars containing 30 wt% calcined clays was obtained after 1d of calorimetric measurement at 40 °C for all setting ages. Li et al. 2018 reached a good prediction of 28d compressive strength of mortars at a replacement rate of 30 wt% of cement by different SCMs. They used the same calorimetric test, but the best correlation between 28d compressive strength and heat development was reach for 3 and 7d of calorimetric measures. Other studies used a similar test design to determine hydraulic or pozzolanic properties of different materials, including GGBS [[26], [27], [28]]. However, it was never attempted to use the R3test to rank different GGBS according to their early age reactivity.

Another promising method to judge GGBS reactivity in literature is proposed by Kashani et al. (2014) [29]. In this study pure 1 M NaOH is used as activation system which gives exploitable results in <12 h in ambient temperature isothermal calorimetry, nevertheless no comparison between different GGBS was attempted.

In the present study, 16 industrial GGBS from different sources were examined in terms of composition and mechanical performance in binder pastes. Subsequently different calorimetric protocols existing in literature were tested in order to develop a rapid and reliable method using minimal sample mass to evaluate GGBS short term performance in cementitious systems. The three main objectives of this study are:

• Evaluation of existing calorimetric tests for the assessment of young age compressive strength of GGBS. Main focus is on protocols based on the R3test and the optimization of grinding parameters.

• Application of an optimized calorimetric test on a set of 16 industrial slags and comparison of test results with actual compressive strength values in order to evaluate the capacity of the test to evaluate the reactivity of different GGBS.

• Discussion of sources of differences in reactivity of the analyzed GGBS, especially chemical composition of the slags and mineral phases formed during the calorimetric test.