Highlights on the properties of the soda-lime-silicate glass residue that enable its use as filler in ultra-high performance concrete

The exponential advancement of cutting-edge technologies in the scope of civil construction, seeks to give cement-based materials the eco-efficient potential linked to mechanical performance that enables different applications. This work aims to evaluate the glass residue regarding the pozzolanic potential through ABNT NBR 5752:2014, as well as to verify whether through the characterization tests of x-ray fluorescence, x-ray diffraction and laser diffraction granulometry, if it is viable of application as supplementary cementitious material (filler), in ultra-high performance concrete. The glass residue submitted to the tests proposed in this study, was crushed in a jaw crusher, milled in a bench ball mill at 47 rpm, and was sieved in a 75 µm opening mesh (ABNT nº 200 mesh). For the test of pozzolanic activity, CP II F-40 class cement, normal sand, water from the public supply network, and superplasticizer additive were used for the mix with 25% of the residue replacing cement, while for the other characterization techniques, the glass residue was applied in its processed form (after sieving), dry or wet. The evaluated glass residue did not reach the minimum rate of 75% established by ABNT NBR 5752:2014, achieving only 45.72%, being classified as non-pozzolanic, which indicates its inert behavior in the presence of calcium hydroxide. The characterization tests confirmed, based on the specialized literature on ultra-high performance concrete, its viability as a filler when adopted as an alternative raw material for presenting chemical and mineralogical composition, in addition to granulometric distribution, very close to those used in studies that demonstrated satisfactory results when using the glass residue as an input.


Introduction
Currently, the design of studies that seek to incorporate the eco-efficient potential into conventional materials is becoming increasingly present due to the considerable increase in the consumption of non-renewable raw materials. Studies with solid residues, for effective use in the civil construction sector, are favorably configured by the industry being one of the most consumed inputs of non-renewable origin, such as natural aggregates for mortar and concrete, as well as with high load environmental, such as Portland cement.
Regarding solid waste, it must be emphasized that the management of industrial by-products needs to consider a work perspective that is supported by a program to mitigate emissions directly at the generating source, as well as through the recycling mechanism itself (Evangelista, Tenório and Oliveira, 2012). Associated with the search for this reduction, the cement industry seeks, through the use of mineral admixtures, to reduce costs with the production of cement, mainly of an energy order, and that, in line with this, it was possible to obtain benefits in the quality of the final product (Garcia, Cabral Junior, Quarcioni and Chotoli, 2015).
Verifying the presence of pozzolanic behavior in industrial by-products makes it possible to classify whether supplementary cementitious materials (SCM) are inert (filler) or reactive (pozzolan). The pozzolanic potential of a material is configured to detonate cementing properties when it reacts in the presence of calcium hydroxide (CH), which originates from the hydrated phases of Portland cement, producing low-density calcium silicate hydrated (C-S-H) (Cunha Oliveira, Chagas, Meira, Carneiro & Melo Neto, 2020;Cunha Oliveira, Meira & Lucena, 2021), similar to that produced by C3S and C2S (tricalcium and dicalcium silicates), and which increases the durability of the cement matrix in the hardened state.
According to Ramanathan, Croly and Suraneni (2020), SCM are increasingly being used to partially replace Portland cement, both to improve the properties of concrete in the long term and to reduce carbon emission rates. Jiang, Ling, Mo and Shi (2019) state that SCM as fly ash, silica fume, ground granulated blast furnace slag, and metakaolin are among the alternatives used to reduce the consumption of Portland cement. In addition, the glass powder, being amorphous and containing a relatively high amount of silica, was also recognized as an SCM with the ability to trigger the pozzolanic reaction (Jiang, Ling, Mo & Shi, 2019).
However, it is noted that in most cases in the literature, for the UHPC mixing project, the quantities of mineral admixtures are provided directly without any detailed explanations or theoretical support, and furthermore, due to the complex cement system of this type of composite, the influence of different mineral admixtures on the hydration kinetics and properties of UHPC still needs further clarification (Yu, Spiesz & Brouwers, 2015). This makes it possible to say, even, that the behavior of the glass powder, and other additions, in presenting reactive potential as a pozzolanic SCM requires complementary evaluations. Vaitkevičius, Šerelis and Hilbig (2014), observed that, when applying powdered soda-lime-silicate glass residue in an attempt to certify it as pozzolanic, it did not have pozzolanic properties as good as microsilica, which consumes almost 5 times more CH than glass powder during hydration of Portland cement. It is worth mentioning that the addition of the glass powder occurred together with the silica fume, resulting in the greater resistance of the study (221 MPa), being an indication that the authors already expected that the residue would act more as a filler than as a pozzolan Using x-ray diffraction, Vaitkevičius, Šerelis and Hilbig (2014) showed that the peak referring to the mineralogical phase of CH still exists in the composition that there is only glass powder, being the 2nd largest among the mixtures, that is, its consumption is not it was completely accomplished. Thus, it is understood that the powdered glass residue tends to have a greater potential to be a SCM with a character of a filler, and not a pozzolan, although Vaitkevičius, Šerelis and Hilbig (2014) also affirm that, the glass powder finely ground with particles up to 100 µm can eventually act as a pozzolanic material. Other factors such as the surface area and its diameter at 50% (d50), can help in understanding the reactivity of the glass, as well as its effect as a chemical activator.
The production of glass waste has a record of generation of powdered residues on a scale of at least 48 thousand tons of powder each year in Brazil (Rodier & Savastano Jr, 2018), and the adoption of powdered residues to improve the properties of the microstructure of cementitious matrix materials is one of the most developed aspects of study in the world due to the interesting characteristics it has: amorphous structure and a large amount of silicon present (Bouchikhi, Benzerzour, Abriak, Maherzi & Mamindy-Pajany, 2019).
Several studies indicate that the glass residue, when finely ground (in granulometric ranges below 75 µm), can present pozzolanic potential, and when used in cement substitution ranges from 10% to 25%, gives satisfactory results (Patel, Tiwari, Shrivastava and Yadav, 2019). In addition to providing better long-term performance, the addition positively reduces the permeability of concrete, resulting in increased durability due to the densification of the material's microstructure through the micro-filler effect (Lee, Hanif, Usman, Sim & Oh, 2018). The application of this residue occurs in various ways in mixtures of concrete and mortar, involving the sphere of aggregates, composing the class of aggregates and acting as a filling material for the larger size of the grains in the mixture, and the sphere of non-aggregates, which due to their high fineness and reactivity make up the group of mineral admixtures (Lee, Hanif, Usman, Sim & Oh, 2018).
Bearing in mind that UHPC systems use high percentages of mineral admixtures to increase their performance in terms of axial compression and durability (Cunha Oliveira, 2020), due to the pore refinement on a microstructural scale resulting in high compactness (Cunha Oliveira, Meira & Lucena, 2021), the glass residue incorporated into the composition as SCM (filler), due to reduced particle size, recommends its use for particle sizes below the fine aggregates, because the specific mass is similar and has less absorption (Soliman & Tagnit-Hamou, 2017a).
The objective of this work is to evaluate the pozzolanic activity index of the glass waste using ABNT NBR 5752:2014, as well as to characterize the residue after beneficiation by means of x-ray fluorescence, x-ray diffraction, and particle size by laser diffraction, to test whether its properties are equivalent to those of the glass waste used as an alternative raw material in UHPC systems. Research, Society and Development, v. 10, n. 3, e59310313801, 2021 (CC BY 4.

Processing of Glass Residue
The residues of silica-lime glass, originating from the disposal of bottles, packaging, and household appliances, of different colors, were collected and sanitized to remove most of the impurities, and then the processing took place, which took place at the Laboratory of Comminution of IFPB -Campus Campina Grande. The first step consisted of crushing the waste by means of a jaw crusher (Figure 1), to reduce the vitreous artifacts to particle sizes suitable for grinding.  Finally, the crushed residue was sieved through a 75 µm opening sieve (ABNT n° 200) with the aid of a mechanical stirrer in 30-minute cycles at a frequency of 8 Hz (Figure 3). All the material passed through the mesh was stored in plastic bags, and the retained material was taken to the ball mill for grinding. Source: Authors.

Pozzolanic Activity Index
Based on ABNT NBR 5752:2014, which designates the method to assess the pozzolanic potential of materials, in powder form, in the presence of Portland cement, CP II F-40 cement from the manufacturer Elizabeth Cimentos, sand was used normal based on ABNT NBR 7214:1982, water from the public supply network in the city of Campina Grande-PB, and superplasticizer additive of the type MasterGlenium® 51 from the manufacturer BASF, for the mixture with 25% of the residue to replace cement.
In order to measure the quantity of materials for the production of the specimens, the quantitative suggested by ABNT NBR 5752:2014 was taken as a parameter, however with the water/cement factor adaptation suggested by it from 0.48 to 0.60, an since the amount of water to be used must recommend an abatement of 225 mm (Cunha Oliveira, Chagas, Meira, Carneiro Source: Authors. Melo Neto, 2018).
The 5x10 cm cylindrical molds were made, with a 1:3 mix (cement and sand), composing two mortars: the reference mortar (A), and the one with the replacement of cement by 25% of the residue (B), with rupture performed at 28 days of age following ABNT NBR 5752:2014. Table 1 shows the quantities for the two mortars manufactured, and Table 2 shows the granulometric composition of normal sand according to ABNT NBR 7214:1982.

X-Ray Fluorescence
The X-Ray Fluorescence (XRF) test was performed at the Materials Characterization Laboratory (CCT/UAEMa), at UFCG -Campus Campina Grande. A dispersive energy spectrometer SHIMADZU Model EDX-720 (Energy Dispersive X-Ray Spectrometer) was used, with generation of x-rays through a Rh target tube, in a vacuum atmosphere and collimator with a 10 mm opening.

X-Ray Diffraction
The X-Ray Diffraction (XRD) test was carried out at the Materials Characterization Laboratory (CCT/UAEMa), at UFCG -Campus Campina Grande, using the powder method. The SHIMADZU Model XRD-6000 equipment (X-Ray Diffractometer) was used, which has a fixed x-ray tube in high vacuum with tungsten filament, with the sample rotation in θ and the arm in 2θ. The radiation used was Kα (monochromatic) for the target metal Copper (Cu), with λ = 1.5406 Å, under a voltage of 40 kV and a current of 30 mA, with slits opening at 1.0°, 1.0° and 0.3 mm. The scanning speed was set at 2°.min -1 , at a sampling step of 0.02° and a stop time of 0.60 s, in the range of 5° to 60°. The Jade 5.0 software was used to identify the clay minerals present, using the ICDD (International Center for Diffraction Data) crystallographic charts.

Laser Diffraction Granulometry
The Laser Diffraction Granulometry test was performed at the Materials Technology Laboratory (CCT/UAEMa), at UFCG -Campus Campina Grande. The particle size distribution was studied using the liquid phase particle dispersion method, associated with an optical measurement process through laser diffraction, combining the proportional relationship between the diffracted laser and the concentration and particle size. The sample was dispersed by sodium hexametaphosphate (HMFNa) + sodium carbonate (Na2CO3), in 470 ml of distilled water, with the aid of a Hamilton Beach® Model N-5000 shaker, at 17,000 rpm for 20 minutes. The CILAS Model 1064 equipment was used, operating in a wet way (DL), which works under data processing using the Fraunhofer algorithm, with results denoted in granulometric curves (histogram and accumulated) in the range of 0.04 to 500 µm.

Pozzolanic Activity Index
Seeking to interpret the result obtained from the pozzolanic potential of the glass residue in producing C-S-H in the presence of Portlandite, with a particle size of less than 75 µm, Table 3 presents the arithmetic mean for the 4 tested specimens of each composition made (mortars A and B). It is clearly noted that for these study conditions, from the processing of the residue for use as a binder to the manufacture of mortars, it did not have any potential to act as a  Table 4 shows the chemical composition of the glass residue, obtained through the XRF test. As expected, high percentages were observed for SiO2 (67.01%), CaO (13.76%), and Na2O (9.58%), confirming that the glass residue used is soda-lime-silicate, in addition to being close to the percentages stipulated by Vogel (1994), Shelby  Figure 4 shows the diffractometric pattern of the glass residue, obtained through the XRD test.
It was possible to observe that, the powdered glass residue presented 90% (d90) of the particles with diameter less than or equal to 77.99 µm, 50% (d50) less than or equal to 36.65 µm, and 10% (d10) less than or equal to 5.35 µm, as shown in Table   5 below. In view of the grinding time of 3 hours at 47 rpm used to obtain the powdered glass residue, the d50 of 36.65 µm obtained is within the expected range, since the input applied to the UHPC was passed in ABNT n° 200 mesh. However, when comparing with the studies that used the powdered glass residue as SCM in UHPC systems, it was noticed that the d50 were lower, to quote: Vaitkevičius, Šerelis and Hilbig (2014) (d50 = 16 µm, using a vibrating disc mill with speed between 750 and 940 rpm); Soliman and Tagnit-Hamou (2016) (d50 = 12 µm), Soliman and Tagnit-Hamou (2017b) (d50 = 12 µm), e Wilson, Soliman, Sorelli and Tagnit-Hamou (2019) (d50 = 12 µm), where both three jobs used air classifier and jet mill with speed between 2,000 and 22,000 rpm. Despite the difference of at least 20 µm as compared to the d50, it is possible to highlight the energy savings in the production of SCM, since the focus on the use of glass waste is to combine sustainable practices with Portland cement matrix materials.

Conclusion
The evaluated glass residue did not reach the minimum rate of 75% established by the standard, achieving only 45.72%, being classified as non-pozzolanic, that is, it remains inert during the hydration process because it does not react with the calcium hydroxide released by the Portland cement. The glass residue characterized by XRF, XRD and laser granulometry had technical conformity in terms of chemical, mineralogical and granulometric composition, being an indication that enables its application as an alternative raw material in UHPC systems, as it is equivalent to the results observed in the literature on the development of UHPC that adopted glass residue as an ecological input.