Long-Term Effect of Different Particle Size Distributions of Waste Glass Powder on the Mechanical Properties of Concrete

In this research, the employed cement was ordinary Portland cement (OPC) with a strength grade of (42.5 R) produced locally following the Iraqi standard (IQS/5/1984) [54]. The GP used was obtained from locally collected waste-broken glass windows, commonly known as soda-lime glass. To meet the fundamental demands of the pozzolanic behavior of GP and to prevent adverse effects of alkali-silica reactions, the waste glass was ground and then sieved to less than 135 µm of particle size. In order to examine the impact of particle sizes of GP of the same type, two PSDs were used: labeled GP-A, particle sizes ranging from 55 µm to 135 µm (55 µm< GP-A < 135 µm) and labeled GP-B, particle sizes less than 55 µm (55 µm> GP-B). The chemical and mineralogical analysis of the cement and GP used are given in Table 1. Compared to the chemical composition of natural pozzolans of ASTM C618-2015a [32], the combination of the three oxides (SiO₂ + Al₂O₃ + Fe₂O₃) in GP is about 75.16%, which surpasses the 70% requirement for Class N raw and calcined natural pozzolans requirement. GP is, therefore, expected to be capable of producing high-quality pozzolans. The X-ray diffraction (XRD) for GP was performed as shown in Fig 1. From this figure, the amorphous nature of the GP can be seen as there were no clear crystalline peaks to be found. A Laser Particle Analyzer was used to determine the grain size distribution curves for cement and GP, and the results are shown in Fig. 2. The specific gravity of cement, GP-A, and GP-B were determined to be 3.15, 2.50, and 2.505, respectively. The coarse aggregate used was crushed gravel with a nominal maximum size of 12.5 mm. Washed river sand with a maximum size of 4.75 mm was used for fine aggregates. Physical tests of aggregates were carried out following the ASTM designations, and the results are shown in Table 2. The gradation test was carried out on aggregates following ASTM C136 [33]. The aggregate used was within the limits of the specification of ASTM C33 [34], as shown in Fig. 3.

Figure 1.

Figure 2.

Figure 3.

To fully understand the long-term performance of using GP content with different particle sizes on the mechanical behavior of concrete produced with different cement amounts, two sets of concrete mixes were made. The control mix of the first set (LCCM) had a cement content and a water-cement ratio of 331 kg/m³ and 0.54, while the cement content and water-cement ratio of the control mix of the second set (HCCM) were 490 kg/m³ and 0.41, respectively. In comparison to the first, the cement content of the second set is high. The ratios of GP for each of GP-A and GP-B used as cement replacement in the first and second sets were (0%, 5%, and 10%) and (0%, 5%, 10%, and 15%), respectively. Slump test and different mechanical behavior for the mixes mentioned above have been achieved and compared to each other. Further details on the experimental parameters and the tests carried out in this study are provided in Fig. 4.

Figure 4.

To meet the research objectives, twelve mixtures were prepared and cast. The composition and labeling of the concrete mixtures used in this study are presented in Table 3. The mixtures were made using a 0.08 m³ electric tilting mixer following the procedures in ASTM C192 [35] standard.

The cylinders and prisms were prepared and cured under laboratory conditions in the structural laboratory of the Department of Civil Engineering, College of Engineering, University of Sulaimani-Iraqi Kurdistan. After 24 hours of casting, the specimens were demolded and submerged in a water-curing tank at 23 ± 2°C, as recommended by ASTM C192 [35] until the time of the test.

In general, four sets of specimens were prepared for each mixture, and each set of specimens consisted of three individual specimens. The first set was used for the 28-day compressive strength test. The other three sets were used to test the 180-day Compressive strength and elastic modulus, splitting tensile, and flexural strength. The test result for each set was obtained by averaging the specimens’ values of that set. Additional details on the tests, the total number, and the size of the specimens used in this study are provided in Fig. 4. Figure 5 shows the experimental work and tests performed on hardened concrete specimens with and without GP.

Figure 5.

The samples were tested using a universal testing machine (CONTROLS type) with a maximum capacity of 4000 kN. The loading rate used for testing compressive and elastic modulus, flexural strength, and splitting tensile strength was 0.275 MPa/s, 0.02 MPa/s, and 0.023 MPa/s, respectively. The flexural tensile strength was conducted using a third-point loading test.

Two electrical resistance strain gauges of type (BX 120-80 AA) were used to obtain strain measurements for each cylindrical specimen used to test the elastic modulus. The strain gauges were symmetrically fixed at the center of the concrete cylinder’s height using a high-performance adhesive. Stress and the corresponding strain values were measured using the data acquisition system (Microlink 851). Graphic displays were obtained using the data acquisition software “Windmill Software”. Elastic modulus was then determined following ASTM C469 specification [36].

The slump test results for all mixtures containing GP-A and GP-B and compared with the control mixture are shown in Fig. 6. The test results showed that, regardless of the GP particles’ size, there is a systematic increase in workability with an increase in GP replacement. Both GP particle sizes have the same increasing trend. The non-absorbent nature and glassy surface of GP can be considered significant factors in increasing concrete workability [28, 43, 44]. It was also found that, at the same percentage of substitution, the finer-GP (GP-B) yields a relatively higher slump value than the coarse-GP (GP-A).

Figure 6.

Figures 7 through 10 illustrate the short-term (28-day) and long-term (180-day) effects of the use of GP-A and GP-B for cement substitution on the compressive strength of concrete with relatively low and high cement content. From Fig. 7, at 28 days of testing and no matter the size of the GP particle, the use of GP in the relatively low cement content mixtures (LCCM) results in a reduction in compressive strength, and the reduction is proportional to the replacement percentage. However, for the relatively high cement content mixtures (HCCM), the compressive strength of mixtures containing 5% of GP-B and GP-A was found to increase by 5.5% and 1.2% compared to the control mix, respectively, as shown in Fig. 8. It was also found that the use of more than 5% GP, independent of the particle size of GP, reduces the compressive strength at 28 days.

Figure 7.

Figure 8.

Figure 9.

Figure 10.

From the long-term test results (180-days) (Fig. 9), as the curing ages increased, the compressive strength development of LCCM modified with GP-B increases more evidently compared to the control mix. The compressive strength of the mixture modified with 5% of GP-B is almost equal to that of the control mix. However, the compressive strength of mixtures containing 10% of GP-B exceeds the control mixture by 3.04%. In contrast to GP-B, the 180-day compressive strength values of the mixtures modified with 5% and 10% GP-A remain below the control mix value by approximately 5%.

Referring to Fig. 10, For HCCM, using GP-A and GP-B as cement replacement improves the 180-day compressive strength other than a mixture containing 5% of GP-A, where the compressive strength decreased slightly by 1.95% compared to the control mixture.

From the results (Table 4), Both GP-A and GP-B were found to have approximately similar strengths and performance at 28 and 180 days. Considering HCCM and 180-days, the acceptable dosage of GP to be used as a cement replacement independent of the GP particle size is 15%. In terms of LCCM and 180-day testing age, the optimal dosage of GP to replace cement regardless of GP particle size is 10%. Although GP-A and GP-B had approximately similar strengths at 28 days, the improved performance of the finer GP (GP-B) compared to GP-A can be seen at 180 days of testing.

In general, the enhanced compressive strength of GP-modified concrete after 28 days is mainly due to the secondary pozzolanic reaction and micro filler effect of GP particles, whereby they contribute to the densification of the matrix and interface zone, which reduces porosity and improves concrete performance [23]. However, GP’s pozzolanic response at 28 days is low, showing only the inert filler effect of GP particles before the progression of the pozzolanic reaction of GP occurs at a later age.

The cracking and deflection behavior of the concrete members can be indicated by understanding their tensile strength. Therefore, the splitting and flexural tensile strength of the concrete are essential parameters for the flexural and shear members’ design. Figures 11 and 12 show the results of the effect of the use of GP-A and GP-B with different levels of replacement on the 180-day splitting tensile strength of concrete containing low and high content of cement, respectively. The results generally showed that using GP-A and GP-B for LCCM and HCCM enhanced the splitting tensile strength of concrete. It was observed that the particle size distribution has no significant contribution to the enhancement of the splitting tensile strength compared to the amount of GP replacement. The maximum difference between the splitting tensile strength value of GP-A modified concrete and the corresponding GP-B is 0.14 MPa and 0.07 MPa for LCCM and HCCM.

Figure 11.

Figure 12.

Considering the maximum splitting tensile strength obtained, the optimum GP-A and GP-B replacement level is 10% for LCCM and HCCM, respectively. However, the optimal GP content for cement replacement was found to be 10% and 15%, respectively, for LCCM and HCCM compared to the reference mix.

Long-term splitting tensile strength results indicate that the use of GP up to 10% for LCCM and 15% for HCCM, regardless of its particle size, tends to improve concrete porosity through its pozzolanic reactivity.

The results of the 180-day flexural tensile strength of GP-modified concrete with relatively low and high cement contents are shown in Figures 13 and 14. For LCCM, the increase in GP-B levels from 5% to 10% slightly improves the flexural tensile strength by 1.88% and 2.25% compared to the control mixture. In comparison, for GP-A, only a slight improvement was observed at 5% by 1%, and the flexural strength of the concrete mixture containing 10% of GP-A was slightly less than that of the control mixture by 2.5% (Fig. 13). The positive long-term effect of fine GP particles (GP-B) on the flexural strength of concrete can be seen when the use of GP is increased to more than 5%.

Figure 13.

Figure 14.

For HCCM (Fig. 14), using up to 10% GP-B and GP-A, the flexural tensile strength of concrete slightly improved, and beyond which, the flexural strength values are almost equal to that of the reference specimen. The size impact of GP on the flexural strength was observed to be insignificant. For instance, the maximum difference between the flexural strength of the concrete modified with GP-A and GP-B for the same GP amount is 0.05 MPa.

Figures 15 and 16 represent the size effect of GP replacing cement on the 180-day elastic modulus of concrete containing a relatively low and high cement content, respectively. From Fig. 15, it can generally be found that the replacement of cement by GP-A and GP-B in LCCM slightly increases the elastic modulus except for a mixture containing 5% GP-A, where the elastic modulus is almost the same as the control mixture. The optimal content of GP was determined to be 10%, for which the elastic modulus for GP-A and GP-B was 6.6% and 5.1% higher, respectively than the reference mixture.

Figure 15.

Figure 16.

In the case of HCCM, and whatever the size of the GP particles, using GP up to 15% improves elastic modulus except for mixture containing 15% GP-B, where the elastic modulus is almost the same as the control mixture (Fig. 16). Given the maximum elastic modulus obtained, the optimum GP level is 5% for which the elastic modulus for GP-B and GP-A was 12.49% and 9.80% higher, respectively than the reference mixture. Meanwhile, the acceptable GP dosage for cement replacement was found to be 15% for which the elastic modulus for GP-B was equal to the control mixture and 5.02% for GP-A was higher than the reference mixture..

The improvement in the elastic modulus of concrete modified with GP over the long term age was attributed to the pozzolanic activity of GP and the microstructure improvement of concrete [4, 30].

Figures 17 and 18 represent the results of the 180-day compressive stress-strain behavior of low and high cement content mixtures modified with varying percentages of GP-A and GP-B.

Figure 17.

Figure 18.

Similar trends in stress-strain curves can generally be seen in all mixtures (with or without GP) of both HCCM and LCCM. Due to the brittle characteristics of all specimens, the descending part of the curve could not be defined in any of these specimens.

In terms of LCCM, Fig. 17, a relative improvement can be observed in the stress values for the same strain levels of the mixtures modified with 5% and 10% GP-B compared to that of the control mixture. The same improvement can be seen for mixtures modified with 10% GP-A up to approximately 85% of the peak stress, while the stress and the corresponding strain values for the mixture containing 5% GP-A are almost similar to those of the control mixture. The results generally imply that the increase in deformation of concrete specimens with GP was slower than that of the control specimen. The concrete specimens with fine GP (GP-B) exhibited less deformation in comparison with that of GP-A.

Regarding HCCM, Fig. 18, the mixture modified with 5% GP-B shows a steeper stress-strain curve (less deformation) than any other mixtures, while that with 15% GP-B displays more deformation. For the given compressive stress and except a mixture containing 15% GP-B, the GP-containing mixture shows a relatively small strain compared to the control mixture, indicating that the use of GP serves to make the concrete relatively stiffer under the compression load.

Generally, compression toughness reveals the property of energy absorption and damage resistance when the material is damaged by compression. In this study, the compression toughness was obtained by calculating the area under the stress-strain curve to the ultimate stress value [40]. The specific toughness in (%) was also determined by calculating the ratio of toughness to compression strength [41, 42]. Table 5 shows the results of compression toughness and specific toughness in (%) attained from different concrete mixtures with a maximum corresponding strain.

The 180-day results of compression toughness of the concrete mixtures with a relatively low cement content (LCCM) indicate that the use of GP reduces the compression toughness relative to the control mixture, and the maximum reduction was observed when the percentage replacement of GP-A and GP-B is 5% and 10%, respectively. However, for HCCM, using GP up to 5% reduces the compression toughness of concrete, whereas the increase in GP-A and GP-B levels from 5% to 15% enhances the compression toughness. The maximum compression toughness was achieved when the GP-B and GP-A contents were 15% and 10%, respectively.

The relative reduction in the toughness of the concrete specimens with GP compared to the plain concrete (without GP) in LCCM could be due to a slower dissipation of the absorbed energy in the plain concrete than the specimens with GP during failure. This is due to the lower compressive stiffness of plain concrete than specimens containing GP, which can be seen in the compressive stress-strain relationships (Fig. 17). However, in HCCM, the concrete strength of all mixtures is high (exceeding 55 MPa). The absorbed energy dissipation of high strength concrete during a failure is relatively higher than that of normal strength concrete. The same observations were made on GP-concrete specimens. However, the progression of GP’s long-term reaction leads to an improvement in the internal microstructure of concrete (i.e., a relatively higher compressive strength and stiffness), thereby the compression toughness of concrete specimens containing GP above 5% was relatively higher than that of free-GP concrete.