New insights into the use of glass cullet in cement composites - Long term examinations

Abstract

Finely ground glass cullet is a very attractive pozzolanic additive, as it is an excellent source of active silica. Moreover, crushed glass is sometimes used due to its aesthetic and decorative qualities. Unfortunately, it is also the source of alkalis. The packaging glass cullet contains even more than 16% alkalis, which are slowly but continuously released into the hydrating system. This could be a reason for the occurrence of ASR in hardened concrete. In this work, OPC mortars containing finely ground glass powder (as a part of the binder) and glass sand (as quartz sand replacement) in various proportions were examined for 10 years. The compressive strength, phase composition, and expansion tests were carried out, along with microstructural observations. Many researchers indicated positive effects of the use of glass as a mineral additive, however, the presented long-term results show that the applicability of glass cullet as a mineral additive is disputable. This is the first work that presents the results of 10-year-long examinations of glass-containing mortars. It also shows a new, previously undescribed phenomenon: the glass hydration process has a multistage, not a continuous character.

Introduction

The problem of glass cullet, especially heavily coloured, becomes urgent for two reasons. First, the glassworks are reluctant to use such recycled material in the production process. Second and more important, the global cement market is currently focused on the search for alternative mineral additives due to the limitations in CO₂ emissions, climate changes and the decreasing availability of ashes and slags. Glass cullet seems to be an excellent candidate for such applications. The number of reports on the use of previously unprecedented mineral additives as supplementary cementitious materials in various building materials is continuously growing [[1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13]]. In the previous papers, glass cullet was examined mainly as an aggregate or a pozzolanic additive to OPC, however, it is known that the material can lead to ASR [[14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25]]. Glass cullet is a good source of active silica, calcium and alumina and from this point of view, it shows a great potential as a highly active pozzolanic additive [[26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36]], a component of blended binders [[37], [38], [39], [40], [41], [42], [43]] or alkali activated materials [[44], [45], [46], [47]]. Attempts have also been made to use it as a component of pro-ecological binders [[48], [49], [50], [51]] or glass-modified concretes and mortars [[52], [53], [54], [55], [56], [57], [58], [59], [60], [61], [62], [63], [64], [65]]. However, the material contains a large amount of both alkali and active silica which, in the long term, can become a source of serious problems as the main reason behind ASR. It is also commonly known that the presence of alkalis increases the dissolution rate of SiO₂ - an important factor determining the reactivity of materials during alkali activation [66]. The challenge is to find a way to take advantage of the presence of both active silica and alkalis while mitigating the harmful ASR effects.

ASR was first examined by Stanton [67] in 1940, however, the process is still not fully explained [14]. The chemical activity of glasses strongly depends on their chemical composition and specific surface area, as shown in hundreds of previous papers, for example: [3,56,[68], [69], [70], [71], [72], [73], [74], [75], [76], [77], [78], [79], [80], [81], [82], [83]]. There were many attempts to use ground and coarse glass in the form of powder (GP), sand (GS) or aggregates [17,54,[84], [85], [86], [87], [88], [89]] as a component of binders, however, in the light of the low hydration rate of the material, the conducted research was usually short, for example: 28 days [63,90], 56 days [91], 90 days [33,92], 91 days [93,94], 180 days [72], 210 days [35], 356 days [95,96], 560 days [18], 1000 days [97] and many others. There is also the paper by A. F. Omran et al. who studied GP-containing concrete cores obtained 7 years after casting and compared the results to laboratory-cured samples [98]. The laboratory examinations were, however, continued for only 91 days, while curing time is extremely important in the case of glass, as shown in the previously published paper [99]. Many researchers indicate the role of calcium and mutual calcium-sodium relations in special binders [38] and in ASR [14,24,[100], [101], [102], [103]], as it seems to control the both the ASR reaction and the related expansion process. Most authors indicate a positive influence of GP on the compressive strength of OPC mortars, but in this particular case, the effect was examined over a 10-year period of hydration under natural conditions. Long term examinations of GP-containing systems are important, especially in the context of ASR caused by the presence of glass cullet.

The first hypothesis – in theory, glass is much more chemically active than quartz sand due to its thermodynamic properties. Therefore, in alkaline solutions, finely ground GP will react, in terms of the chemistry of cement, very fast - compared to quartz sand. During the hydration of OPC in the presence of GP, a lot of alkalis will be released into the reacting system, which will accelerate the hydration in the early period, as noted in many papers concerning the use of GP as a component of binders [48,62,[104], [105], [106]]. At the early stage of hydration (in the case of this work, up to 90 days), the released alkalis will be physically trapped, but mostly not chemically bound, within the hydrates (C–S–H or C-A-S-H phases). Therefore, they can be released back into the reacting system, as confirmed in previous examinations [107]. On the other hand, when large grains of glass are used as the component of mortars, their reactivity is significantly lower due to the lower specific surface area. Their hydration will proceed for years, but with the same result – releasing alkalis into the reacting system.

The second hypothesis – in theory, during the long-term hydration of systems that contain GP, GS or both, the amount of alkalis will continuously increase, with all of the related consequences – the reacting system will slowly transform from the conventional “OPC-like” hydration to alkali activation. In the presented case, the hydrating system contained only GP, GS, quartz sand, and CEM I 42.5R. However, the majority of currently produced binders contain also SCMs such as fly ashes, slags, clay minerals, etc. that can be alkali-activated and therefore, the effect may be beneficial. The increasing amount of alkalis could, however, exclude the possible use of most available aggregates due to the risk of ASR and among others, it is one of the reasons for the limitation of alkali content in cement in standard regulations.

It is known that the grain size distribution of glass influences the expansion in Portland cement systems [19,23]. It is also known that GP can mitigate the expansion, which is explained by the size of the grains and the chemical reactions [24]. Small grains dissolve too quickly, while large pieces are not reactive enough to cause stresses capable of breaking the hardened matrix. Therefore, there is a range of grain sizes for which ASR is most likely to occur [20,108,109].

The dominant view in the literature is that the addition of GP reduces expansion due to the pozzolanic reaction [16,24,91,108,[110], [111], [112], [113]], which binds the alkalis in the solid phase and consequently, decreases their amount in the liquid phase. As a result, the structure is tighter, the water circulation is more difficult and thus, also the ASR rate is lower. However, it should be taken into account that most of the tests were performed on the basis of ASTM C 1260 or ASTM C 1293. The problem with the glass-containing systems is that in both procedures, an additional amount of alkali is introduced together with the glass which accelerates its dissolution. Thus, it is not entirely clear whether the mitigation of the expansion is related to the accelerated dissolution of the fine glass grains, the pozzolanic reaction, the dissipation of stresses or some other reason.

However, in this study, it was shown that for both GS and GP, the expansion slowed after 4 years of hydration, and the compressive strength slowly started to rise again. The objective of the work was to determine the influence of GP and two fractions of GS on the properties of hardened mortars after long periods of hydration. The two different fractions of GS were used due to the reported influence of glass grain size on ASR [87,114,115]. Samples were cured and examined for 10 years under laboratory conditions.