Cement-Stabilized Waste Sand as Sustainable Construction Materials for Foundations and Highway Roads

In this study, cement-treated waste sand as a by-product material produced from Al-Ahsa quarries (Saudi Arabia) was experimentally tested and investigated as a base course material for the foundation of structures and roads. The study aimed to use the waste sand as a construction material by improving its strength, bearing capacity, and stiffness. The waste sand was mixed with different percentages of Portland cement content (0, 2, 4, 6, and 8%) at the maximum dry density and optimum water content of the standard Proctor compaction conditions of a non-treated sample. Unconfined compressive strength and California Bearing Ratio (CBR) tests for different curing times were conducted. X-ray diffraction (XRD), laser-scanning microscopy (LSM), and X-ray spectroscopy (XPS) were used to explore the microstructure and composition of the treated sand. The results showed that the compressive strength, initial tangent modulus, and CBR of the treated sand increase with the increase in cement content and curing time. Furthermore, good correlations were established among the strength, initial tangent modulus, and CBR. Based on the obtained results, cement-stabilized waste sand is a potential material for use in construction. This is expected to save the environment and reduce the cost of road construction.


Introduction
The conservation of the Earth's natural resources has become an important and critical issue for the continued existence and prosperity of human environments. Huge amounts of natural resources have been utilized for the last two centuries due to rapid industrialization, enormous growth in our population, and the continuous trend of urbanization. The construction industry consumes huge amounts of these natural materials, exhausts natural resources, and causes associated ecological issues. For the last few decades there has been enormous pressure on the construction industry to incorporate sustainability. Searching for new materials, especially the utilization of waste materials from different industries, will help preserve the Earth's natural materials and sustain the natural environment. The use of waste sand in the construction industry has manifold advantages. It can help in the protection of our natural resources, decrease environmental and health hazard issues, reduce a burden on landfills, and contribute to the economy. relationships between strength, initial tangent modulus, and the California bearing ratio (CBR) of the treated waste sand materials for practical use in the construction industry.

Classification of Waste Sand
Two major quarry areas are producing limestone aggregates for construction purposes near the Al-Ahsa area. The first quarry is located east of Al-Ahsa City in the Abu Hadriyah area (Dammam road) and the second is located west of the city on Ryadh-Khurais road. The waste sand used in this work was collocated from the second quarry. Figure 1 shows the location of the used waste sand and Figure 2 shows the stockpiled sand produced at this location. According to the Unified Classification System [33,34] the waste sand is classified as poorly graded sand, while according to the American Association of State Highway and Transportation Officials (AASHTO) system [35] the sand is classified as A3. Figure 3 shows the grain-size distribution of the waste sand and Table 1 summarizes its physical properties and classification.
Materials 2019, 12, x FOR PEER REVIEW 3 of 18 relationships between strength, initial tangent modulus, and the California bearing ratio (CBR) of the treated waste sand materials for practical use in the construction industry.

Classification of Waste Sand
Two major quarry areas are producing limestone aggregates for construction purposes near the Al-Ahsa area. The first quarry is located east of Al-Ahsa City in the Abu Hadriyah area (Dammam road) and the second is located west of the city on Ryadh-Khurais road. The waste sand used in this work was collocated from the second quarry. Figure 1 shows the location of the used waste sand and Figure 2 shows the stockpiled sand produced at this location. According to the Unified Classification System [33,34] the waste sand is classified as poorly graded sand, while according to the American Association of State Highway and Transportation Officials (AASHTO) system [35] the sand is classified as A3. Figure 3 shows the grain-size distribution of the waste sand and Table 1 summarizes its physical properties and classification.

Mineralogical Analysis
Mineralogical analysis using x-ray diffraction (XRD) was carried out using a Rigaku Mini Flex II X-Ray Diffractometer at 10 • to 80 • . The step size maintained throughout the test was 0.01. The results showed that most of the waste sand particles consist of calcite mineral and small amount of quartz, as shown in Figure 4.

Morphologic Analysis
A morphologic analysis was performed using a laser-scanning microscope (LSM). Sand samples were separated into two portions (a fine portion with particle size smaller than 0.425 mm, and a coarse portion with particle size greater than 0.425 mm). Images of the coarser portion were first obtained in the tile-scan (combination image) mode using the 5× objective lens. A tiles-scan image was collected on a large sampling area~12.5 × 12.5 mm 2 with each tile showing~2.5 mm × 2.5 mm. A high-magnification laser microscopic Z-scan image was obtained for the fine portion of sand keeping the fixed X scale bar at 200 µm. A confocal hole of 70 µm was used for the 402 nm diode laser reflection.

Mineralogical Analysis
Mineralogical analysis using x-ray diffraction (XRD) was carried out using a Rigaku Mini Flex II X-Ray Diffractometer at 10° to 80°. The step size maintained throughout the test was 0.01. The results showed that most of the waste sand particles consist of calcite mineral and small amount of quartz, as shown in Figure 4.

Morphologic Analysis
A morphologic analysis was performed using a laser-scanning microscope (LSM). Sand samples were separated into two portions (a fine portion with particle size smaller than 0.425 mm, and a coarse portion with particle size greater than 0.425 mm). Images of the coarser portion were first obtained in the tile-scan (combination image) mode using the 5× objective lens. A tiles-scan image was collected on a large sampling area ~12.5 × 12.5 mm 2 with each tile showing ~2.5 mm × 2.5 mm. A highmagnification laser microscopic Z-scan image was obtained for the fine portion of sand keeping the fixed X scale bar at 200 μm. A confocal hole of 70 μm was used for the 402 nm diode laser reflection.

Elemental Composition Analysis
Elemental composition analysis was performed using x-ray photoelectron spectroscopy (XPS). The data was recorded in an Omicron (ESCA) spectrometer using a Mg Kα X-ray line (1254 eV photon energy) and all the spectra were calibrated by adventitious carbon peak position (284.8 eV). The survey spectrum of the XPS was recorded with 0.5 eV energy resolution while high-resolution XPS for the signature peaks of elements was recorded with 0.02 eV energy resolution.

Cement
High sulfate resistance (type V) Portland cement with low tricalcium aluminum (C3A less than 5%) was used in this work. Type V cement is available in the market and it is suitable for use in roads and foundation construction, since those structures are expected to be exposed to high levels of

Elemental Composition Analysis
Elemental composition analysis was performed using x-ray photoelectron spectroscopy (XPS). The data was recorded in an Omicron (ESCA) spectrometer using a Mg K α X-ray line (1254 eV photon energy) and all the spectra were calibrated by adventitious carbon peak position (284.8 eV). The survey spectrum of the XPS was recorded with 0.5 eV energy resolution while high-resolution XPS for the signature peaks of elements was recorded with 0.02 eV energy resolution.

Cement
High sulfate resistance (type V) Portland cement with low tricalcium aluminum (C 3 A less than 5%) was used in this work. Type V cement is available in the market and it is suitable for use in roads and foundation construction, since those structures are expected to be exposed to high levels of sulfate ions. Tables 2-4 show the chemical, physical, and mechanical properties of the used cement produced in Saudi Arabia [36].

Water
The water used in the test was tap water and, according to AASHTO T-26, it has less than 1000 ppm of chloride (CL −2 ) and less than 3000 ppm sulfates (SO 4 +2 ).

Experimental Program
The testing program was designed to achieve the objectives of this work. The program focused on the investigation of the behavior of the treated waste sand at different percentages (0, 2, 4, 6, and 8%) of Portland cement added to the waste at the maximum dry density and optimum water content of the non-treated sand, using the standard Proctor compaction ASTM D698-07 method A [37]. The major tests included in this work were: the unconfined compressive strength (q u ) test according to ASTM D2166 [38] and the California bearing ratio (CBR) test according to ASTM D1883-07 [39]. The two types of tests were performed for different curing times (7, 14, and 28 days). After molding, the treated sand samples were tightly wrapped and sealed in plastic sheets to maintain the optimum water content. Table 5 shows the tests performed at different percentages of Portland cement and different curing times. Figure 5 shows the standard Proctor compaction curve of the waste sand.

Physical and Chemical Characteristics of the Waste Sand
Results of laser-scanning microscope (LSM) analysis showed that the coarse part of the sand (particle size greater than 0.425 mm) are non-homogenous and angular in shape with sharp edges. In contrast, the fine part (particle size less than 0.425 mm) is almost homogenously rounded in shape, as shown in Figure 6. The composition details of the waste sand from x-ray photoelectron

Physical and Chemical Characteristics of the Waste Sand
Results of laser-scanning microscope (LSM) analysis showed that the coarse part of the sand (particle size greater than 0.425 mm) are non-homogenous and angular in shape with sharp edges. In contrast, the fine part (particle size less than 0.425 mm) is almost homogenously rounded in shape, as shown in Figure 6. The composition details of the waste sand from x-ray photoelectron spectroscopy (XPS) are presented in Figure 7. Figure 7a demonstrates the presence of calcium, silicon, oxygen, and some atmospheric carbon in the waste sand samples. The photoelectron count per second (cps) for the three-high-resolution XPS peaks, Si2p, Ca2p and O1s, are shown in Figure 7b-d. XPS peak areas calculated for the Si2p and Ca2p peaks are 221 and 329 sq.-cps, respectively. Using the standard empirical atomic scattering peak factors, which are 0.27 (Si2p 3/2 ) and 1.58 (Ca2p 3/2 ), we found actual atomic percentages by multiplying them with the observed sq.-cps values [41]. The amount of lime and silica in the sample were found to be 89.7% and 10.3%, respectively. Moreover, the native oxide, also known as crystalline oxide, was present in the waste sand sample. A clear deconvolution in the XPS O1s peak of the native oxide was observed at 530 eV energy as shown in Figure 7d. Similarly, the sub-micron sized nature of the waste sand particles was evident from the presence of the significant surface oxide O1s peak at 532.3 eV. Lee and Oh (2004) [42] reported similar higher-energy O1s peak positions in the surface oxides placed at 532.5 eV in the XPS spectrum. A high ratio of surface oxides compared to native oxides means higher surface reactivity (due to the smaller micro crystallize structure). This structure is expected to work effectively with the cement and develop a strong bond between the sand grains. deconvolution in the XPS O1s peak of the native oxide was observed at 530 eV energy as shown in Figure 7d. Similarly, the sub-micron sized nature of the waste sand particles was evident from the presence of the significant surface oxide O1s peak at 532.3 eV. Lee and Oh (2004) [42] reported similar higher-energy O1s peak positions in the surface oxides placed at 532.5 eV in the XPS spectrum. A high ratio of surface oxides compared to native oxides means higher surface reactivity (due to the smaller micro crystallize structure). This structure is expected to work effectively with the cement and develop a strong bond between the sand grains.

Unconfined Compressive Strength (q u )
Unconfined compression tests were performed at the maximum dry density and optimum water content of the waste sand prepared at standard Proctor compaction (Method A). The sample dimensions were 102 mm in diameter and 116 mm. in height. Different percentages of cement were added to the sand (0, 2, 4, 6, and 8%). The samples were sealed with plastic sheets and subjected to curing times of 7, 14, and 28 days at room temperature (22-23 • C). A universal testing machine was used to perform the tests as shown in Figure 8a. The results of the tests showed that the unconfined compressive strength increased tremendously with the increase in cement content, as shown in Figure 9. In addition, the results showed that the strength of the treated sand increased with the increase in curing time, especially for the 28 days curing time. This behavior was expected since the cement particles coat and bind the sand grains, which in turn increase the resistant forces at the contact points of the sand grains. The obtained compressive strength results were supported by XPS spectrum analysis that showed a high ratio of surface oxide and surface reactivity of the waste sand particles, which in turn contributed to developing strong bonds at the contact areas of the sand grains. As the curing time increased, more hydration (cement reaction with water) took place, and consequently the strength increased. Figure 10 shows the percentage increase in q u with cement content. In this figure, it can be seen that with adding just 2% cement, the increase in q u of the treated waste sand was 500% of the non-treated sample, and with an 8% increase in cement, the q u was 2500%.

Unconfined Compressive Strength (qu)
Unconfined compression tests were performed at the maximum dry density and optimum water content of the waste sand prepared at standard Proctor compaction (Method A). The sample dimensions were 102 mm in diameter and 116 mm. in height. Different percentages of cement were added to the sand (0, 2, 4, 6, and 8%). The samples were sealed with plastic sheets and subjected to curing times of 7, 14, and 28 days at room temperature (22-23 °C). A universal testing machine was used to perform the tests as shown in Figure 8a. The results of the tests showed that the unconfined compressive strength increased tremendously with the increase in cement content, as shown in Figure 9. In addition, the results showed that the strength of the treated sand increased with the increase in curing time, especially for the 28 days curing time. This behavior was expected since the cement particles coat and bind the sand grains, which in turn increase the resistant forces at the contact points of the sand grains. The obtained compressive strength results were supported by XPS spectrum analysis that showed a high ratio of surface oxide and surface reactivity of the waste sand particles, which in turn contributed to developing strong bonds at the contact areas of the sand grains. As the curing time increased, more hydration (cement reaction with water) took place, and consequently the strength increased. Figure 10 shows the percentage increase in qu with cement content. In this figure, it can be seen that with adding just 2% cement, the increase in qu of the treated waste sand was 500% of the non-treated sample, and with an 8% increase in cement, the qu was 2500%.

Initial Tangent Modulus (E i )
The initial tangent modulus of the treated waste sand was also evaluated at different percentages of cement and curing times. The results in Figure 11 show that the initial tangent modulus increased with the increase in both the percentage of cement and curing time. The increase in E i with cement content and curing time, as discussed before, was due to: (i) The bonding effect of cement at the contact points of the sand grains, and (ii) the hydration process that is a function of the curing time. Figure 12 shows the percentage increase in E i with cement content. In this figure, it can be seen that, by using a small percentage of cement (between 2 and 4%), the E i increased by 200-700%. The substantial improvement in the initial tangent modulus of the treated sand is expected to reduce the deformation in the base course layer, which in turn will reduce the damage in the road surface. As seen in the same figure, with 8% added cement, which is a relatively high amount, the increase in E i of the treated sand is 2000%.

Initial Tangent Modulus (Ei)
The initial tangent modulus of the treated waste sand was also evaluated at different percentages of cement and curing times. The results in Figure 11 show that the initial tangent modulus increased with the increase in both the percentage of cement and curing time. The increase in Ei with cement content and curing time, as discussed before, was due to: i) The bonding effect of cement at the contact points of the sand grains, and ii) the hydration process that is a function of the curing time. Figure 12 shows the percentage increase in Ei with cement content. In this figure, it can be seen that, by using a

California Bearing Ratio (CBR)
Un-soaked CBR tests (Figure 8b) were performed on the treated waste sand at the maximum dry density and optimum water content of standard Proctor and at the same percentages of cement content and curing times used for the unconfined compression tests. The results in Figure 13 shows that the CBR value increased with the increase in the cement content and the curing time. Figure 14 shows the percentage increase in the CBR value with cement content. The results indicated that by

California Bearing Ratio (CBR)
Un-soaked CBR tests (Figure 8b) were performed on the treated waste sand at the maximum dry density and optimum water content of standard Proctor and at the same percentages of cement content and curing times used for the unconfined compression tests. The results in Figure 13 shows that the CBR value increased with the increase in the cement content and the curing time. Figure 14 shows the percentage increase in the CBR value with cement content. The results indicated that by

California Bearing Ratio (CBR)
Un-soaked CBR tests (Figure 8b) were performed on the treated waste sand at the maximum dry density and optimum water content of standard Proctor and at the same percentages of cement content and curing times used for the unconfined compression tests. The results in Figure 13 shows that the CBR value increased with the increase in the cement content and the curing time. Figure 14 shows the percentage increase in the CBR value with cement content. The results indicated that by just using a small amount of cement (1.0-2.0%), the CBR value increased by 4000-8000%. For cement content of 8%, the CBR value increased by 12,000%, which is extremely high. just using a small amount of cement (1.0-2.0%), the CBR value increased by 4000-8000%. For cement content of 8%, the CBR value increased by 12,000%, which is extremely high.

Relationships among qu, Ei and CBR Value of the Treated Waste Sand
Based on the results obtained from the unconfined compression and CBR tests of the treated waste sand, useful and practical relationships can be drawn among the unconfined compressive strength (qu), initial tangent modulus (Ei), and CBR values, as shown in Figures 15-17. Figure 15 shows that as the qu increased, the Ei increased. Furthermore, the data shows a strong linear just using a small amount of cement (1.0-2.0%), the CBR value increased by 4000-8000%. For cement content of 8%, the CBR value increased by 12,000%, which is extremely high.

Relationships among qu, Ei and CBR Value of the Treated Waste Sand
Based on the results obtained from the unconfined compression and CBR tests of the treated waste sand, useful and practical relationships can be drawn among the unconfined compressive strength (qu), initial tangent modulus (Ei), and CBR values, as shown in Figures 15-17. Figure 15 shows that as the qu increased, the Ei increased. Furthermore, the data shows a strong linear

Relationships among q u , E i and CBR Value of the Treated Waste Sand
Based on the results obtained from the unconfined compression and CBR tests of the treated waste sand, useful and practical relationships can be drawn among the unconfined compressive strength (q u ), initial tangent modulus (E i ), and CBR values, as shown in Figures 15-17. Figure 15 shows that as the q u increased, the E i increased. Furthermore, the data shows a strong linear relationship between q u and E i . The relationship is simple and can be expressed as, E i = 55q u with a correlation factor, R 2 = 0.98. Figure 16 shows the relationship between the E i and CBR values in %. In this figure, it can be seen that as the CBR value increased, the E i increased. The data can be also correlated with a simple linear regression as, E i = 0.4 CBR (%), with regression factor, R 2 = 0.82. Figure 17 shows the relation between the q u and CBR values. In this figure, it can be seen that as the q u increased, the CBR value increased. The data can also be presented in a simple relation such as CBR (%) = 120 q u , with a correlation factor, R 2 = 0.80. relationship between qu and Ei. The relationship is simple and can be expressed as, Ei = 55qu with a correlation factor, R 2 = 0.98. Figure 16 shows the relationship between the Ei and CBR values in %. In this figure, it can be seen that as the CBR value increased, the Ei increased. The data can be also correlated with a simple linear regression as, Ei = 0.4 CBR (%), with regression factor, R 2 = 0.82. Figure  17 shows the relation between the qu and CBR values. In this figure, it can be seen that as the qu increased, the CBR value increased. The data can also be presented in a simple relation such as CBR (%) = 120 qu, with a correlation factor, R 2 = 0.80.  relationship between qu and Ei. The relationship is simple and can be expressed as, Ei = 55qu with a correlation factor, R 2 = 0.98. Figure 16 shows the relationship between the Ei and CBR values in %. In this figure, it can be seen that as the CBR value increased, the Ei increased. The data can be also correlated with a simple linear regression as, Ei = 0.4 CBR (%), with regression factor, R 2 = 0.82. Figure  17 shows the relation between the qu and CBR values. In this figure, it can be seen that as the qu increased, the CBR value increased. The data can also be presented in a simple relation such as CBR (%) = 120 qu, with a correlation factor, R 2 = 0.80.

Conclusions
An experimental program was performed to investigate the effect of adding different percentages of Portland cement as a stabilizer in the engineering properties of waste sand used as base course materials for the foundations of roads and buildings. In general, and based on the results of the unconfined compressive strength and the California bearing ratio, cement-stabilized waste