Response of Reinforced Concrete Columns Embedded with PET Bottles Under Axial Compression

Abstract

This study explores the potential use of Polyethylene Terephthalate (PET) plastic bottles as void makers in short reinforced concrete columns under pure axial compression. Such a scheme promotes sustainability by decreasing the consumption of concrete and reducing the pollution that comes with the disposal of PET bottles. The experimental component of this study consisted of testing 16 reinforced concrete columns divided into two groups, based on the cross-section dimensions. One group contained eight columns of a length of 900 mm with a net cross-sectional area of about 40,000 mm², while the second group contained eight columns of a length of 1100 mm with a net cross-sectional area of about 62,500 mm². The diameter of the void within the small cross-section group was 100 mm and within the large cross-section group was 265 mm. The experimental program includes pairs of solid and corresponding void specimens with consideration of the size of the longitudinal steel reinforcement, lateral tie spacing, and concrete compressive strength. The tests are conducted using a universal test machine under displacement-controlled loading conditions with the help of strain gauges and Linear Variable differential transformers (LVDTs). The analysis of the test results showed that the columns that were embedded with a small void that occupied about 30% of the core area exhibited reductions of 9% in the ultimate capacity, 14% in initial stiffness, 20% in ductility, and 1% in residual strength. On the other hand, the columns that contained a large void occupying about 60% of the core area demonstrated reductions of 24% in the ultimate capacity, 34% in initial stiffness, and 26% in ductility, although the residual strength was slightly increased by 5%. The reason for the deficiency in the structural response in the latter case is because the void occupied a significant fraction of the concrete core. The theoretical part of this study showed that the ACI 318 code provisions can reasonably predict the uniaxial compressive strength of columns embedded with PET bottles if the void does not occupy more than 30% of the concrete core. This study confirmed that short columns embedded with relatively small voids made from PET bottles and subjected to pure axial compression create a balance between sustainability benefits and a structural performance tradeoff.

1. Introduction

Concrete is the most widely used construction material in the world due to its ease of use and the availability of constituent materials [1]. However, its extensive use results in massive amounts of pollution, particularly through the production of cement. This includes gases that increase global warming such as carbon dioxide and toxic gases that negatively impact health such as nitrous oxide and sulfur [2,3]. It is important to note that the worldwide population is increasing at an unprecedented rate, and with that, the demand for infrastructure—and consequently, concrete production—also rises, leading to an inevitable increase in harmful pollution to the environment [4,5].

Another major environmental concern is the excessive use and disposal of plastic. Plastic production has drastically increased over recent years, reaching 460 million tons [6]. Non-biodegradable materials, such as plastic bottles, pose significant harm to the environment in multiple ways. Firstly, since plastic does not biodegrade, disposing of it in water bodies can severely impact marine life through ingestion and entanglement [7]. Moreover, the large polymer chains that make up plastic can be toxic to certain species.

In the last two decades, efforts have been made to recycle concrete by crushing demolished concrete rubbles into a coarse and fine aggregate, and then use it in new concrete mixes. Such an approach serves as a sustainable alternative to virgin aggregates, reducing landfill waste and the environmental impact of mining. However, more work is urgently needed in the field of sustainable construction to address climate change, pollution, and resource depletion.

In building structures, columns support vertical loads from the roof and floors primarily through compression, with or without bending. They are crucial members in a structural system because their failure often leads to patrial or total collapse. Reinforcing concrete columns with steel is necessary for improving their stiffness, strength, ductility, and structural integrity. Concrete columns are reinforced both longitudinally by rebars and transversely by ties. The longitudinal rebars help in increasing the compressive and flexural strengths, whereas the transverse ties provide confinement to the concrete core, offer extra shear strength, and prohibit the rebars from buckling from the steel cage after concrete cover spalling.

Extensive past research has explored the incorporation of plastic waste in concrete mixtures, primarily as a partial replacement for fine/coarse aggregates or as fibers to enhance the sustainability and environmental performance of concrete. For example, in 2012 Kitsutaka and Uchida [8] investigated the integration of PET bottles into concrete slabs as embedded voids to develop lightweight, multifunctional concrete. Their findings indicated a weight reduction of approximately 16% compared to solid concrete slabs. However, the compressive strength varied significantly with the bottle spacing; the most closely spaced configuration led to a 55% reduction in strength, highlighting the structural trade-offs involved. Later on, Foti [9] explored the use of recycled PET fibers, derived from waste bottles, as a discrete reinforcement in concrete beams. The results demonstrated improved ductility and good bonding between the PET fibers and the cement matrix, suggesting potential applications in pavements, slabs, and masonry layers.

Ahmad et al. [10] evaluated the use of PET bottle waste as a fine aggregate in concrete paver blocks. Concrete mixes with a PET content above 5% exhibited reduced compressive strength, lower workability, and higher porosity, indicating a limit to acceptable substitution levels without compromising the quality. Waroonkun et al. [11] developed non-load-bearing concrete blocks using plastic flakes as a fine aggregate replacement. Their optimized mix—comprising 80% plastic flakes and 20% natural aggregate with a 0.5 water–cement ratio—yielded the highest compressive strength among the tested specimens. Babafemi et al. [12] provided a comprehensive review of the influence of recycled plastic waste on both the fresh and hardened properties of concrete. They concluded that while plastic inclusion may reduce the mechanical performance, concrete containing such materials can still meet structural requirements in certain applications. Ongpeng et al. [13] used PET bottle strips to confine rectangular concrete columns, reporting strength improvements between 19% and 70% and notable increases in the axial strain capacity. Similarly, Robleh [14] examined the arrangement of PET bottles within self-compacting concrete blocks. The blocks met ASTM standards for non-load-bearing applications and exhibited a 56% reduction in thermal conductivity compared to conventional blocks.

Almohana et al. [15] reviewed the broader benefits of incorporating plastic waste in concrete, including improved thermal and acoustic insulation and reduced reliance on natural resources. However, they cautioned that the decreased density could impact the structural integrity and long-term durability. Related studies have also explored the use of other lightweight inclusions. For example, Ling et al. [16] embedded polystyrene spheres in reinforced concrete beams to reduce the dead load. Replacing 8.7% of the concrete volume with polystyrene increased the beam strength by 8%, although larger spheres reduced the compressive strength, highlighting the importance of optimal geometry and spacing. Al-Hadithi et al. [17] assessed the effect of waste plastic fibers on concrete’s mechanical and thermal properties. Increasing the fiber volume fraction and aspect ratio improved the impact resistance and ductility, but reduced the density and thermal conductivity, underscoring the trade-offs in using such materials.

Lee et al. [18] evaluated the performance of fastening systems in concrete panels containing plastic voids using the Concrete Capacity method. They observed that truncated cone failure was the most common mode and recommended a reduction in the effective embedment depth for design purposes. Mashaan [19] reviewed the mechanical behavior of concrete incorporating various types of plastic waste, including PET, HDPE, LDPE, PVC, and PP. The study noted high variability in the strength performance due to the differing properties of plastic types and highlighted a significant research gap in assessing the durability. Attia et al. [20] investigated the use of plastic tube fibers in structural concrete. Their optimal mix contained 1.5% by volume of 20 mm long fibers, which enhanced the ductility and toughness but increased the microporosity and reduced the ultrasonic pulse velocity. Finally, Mohamedsalih et al. [21] studied the substitution of a coarse aggregate with plastic waste. While the concrete density remained unaffected, the mechanical properties began to degrade significantly beyond a 2.5% replacement level. Mixes with up to 2.5% plastic maintained comparable strength to conventional concrete, suggesting feasibility for structural applications under certain conditions.

Focusing more on concrete columns, Rasoul et al. [22] conducted a study on the implementation of plastic fibers in concrete columns. The paper studied columns containing different volumetric percentages of fibers ranging between 0% and 1%. The authors found out that while the addition of fibers decreased the axial compressive strength of the columns, it improved the concrete splitting resistance, as demonstrated by an increase of 46% in the cracking load. Fayed and Mansour [23] discussed the behavior of reinforced concrete columns incorporating up to 3% recycled plastic fibers under the action of eccentric loading. The results of the research demonstrated that the addition of fibers prevented concrete cover delamination, which improved the ductility by 71% for the case of a fiber content of 1%.

It is apparent that most past studies on the implementation of PET in concrete structures concentrated on either using the plastic as an aggregate in concrete mixes or shredding it as fibers to enhance the tensile strength of the material. Limited studies have addressed the novel potential of single-use plastic bottles as void formers within structural concrete elements, which is the subject of the current study. The proposed method promotes green building agendas and is easy to implement in structural applications involving concrete work.

2. Research Significance, Scope, and Objectives

The study at hand focuses on developing a sustainable construction approach in which non-biodegradable PET bottles are firmly stacked together and then embedded within structural concrete elements to provide a void and serve as a permanent disposal site for plastic. This novel concept offers multiple advantages, including reducing cement consumption and lowering carbon dioxide emissions. The proposed method is innovative, easy to implement, and particularly suitable for underdeveloped regions due to its cost-effectiveness, practical implementation, and positive sustainability traits. It should be noted that the voids that are made from the PET bottles in this study are strictly used as forms since they do not possess enough stiffness to influence the mechanical properties of the concrete material. Hence, the impact of the void on the structural capacity and integrity of the members will be minimal.

This study involved fabrication, instrumentation, and testing 16 half-scale reinforced concrete square columns, of which eight have a small cross-section and eight have a large cross-section. Each group consisted of four solid and four voided columns. The parameters that were considered in this study include the amount of longitudinal reinforcements, lateral tie spacing, and concrete compressive strength. Each of the columns was subjected to an axial load applied at the center of the cross-section, resulting in concentric axial compression. Due to their small slenderness ratios, the columns are classified as short columns, meaning buckling was not a factor in this study. The experimental load–deflection relationships were analyzed for various performance metrics, including the strength, ductility, stiffness, and residual strength. The results were compared with the ACI 318-25 design code provisions [24] to assess their validity and accuracy.

The objectives of this research were to explore the feasibility and effectiveness of using PET bottles as void makers in reinforced concrete columns under the action of concentric loading. Specifically, this study seeks to:

• Develop a practical method of construction for introducing voids of different sizes into reinforced concrete columns through the use of PET bottles.
• Investigate through experimental testing the impact of adding voids in the form of PET bottles into reinforced concrete columns on the structural response in terms of the stiffness, ultimate strength, ductility, and residual strength.
• Check whether or not the current reinforced concrete code provisions for axially loaded columns are applicable for voided columns created by embedding PET bottles within the concrete core.

3. Experimental Program

The experimental program consisted of a total of 16 columns of reasonable size, of which 8 had small cross-section and 8 had large cross-section. Each cross-sectional category included 4 solid and 4 corresponding voided specimens. Each pair of solid and voided columns focused on a specific parameter in order to study its impact on the structural response. In all cases, the clear concrete cover on the transverse reinforcement was equal to 25 mm. To prohibit early failure due to stress concentration within the loaded or supported end regions of the column specimens during testing, such regions were externally jacketed with structural steel channels, made solid and reinforced with closely spaced ties at spacing of 50 mm.

The small-cross-section columns were 900 mm long with square cross-section of 200 mm side for the solid specimens and square cross-section of 200 mm side for the voided ones. The void was created from 100 mm diameter flexible PET bottles that were firmly stacked inside each other for enhanced stiffness and secured within the central region of the specimen with the help of steel wires attached to the steel cage. Such a method of void making proved to be effective as the voids did not bend, become damaged, or were displaced during the construction of the concrete specimens. Within the cross-section, the void occupied about 30% of the core area. The first pair (denoted by 1-S and 1-V) served as control, utilizing relatively low nominal concrete strength of 20 MPa, 4 ϕ 12 mm longitudinal bars, and ϕ 10 mm transverse ties at 150 mm spacing. The second pair (2-S and 2-V) included larger amount of longitudinal steel in the form of 4 ϕ 16 mm bars instead of the 4 ϕ 12 mm. The third pair (3-S and 3-V) contained lateral steel ties that were closely spaced at 75 mm instead of the 150 mm employed in the control. The fourth pair (4-S and 4-V) employed higher nominal concrete strength of 35 MPa instead of the 20 MPa used in control. Figure 1 illustrates the dimensions and characteristics of column specimens that are part of group 1.

On the other hand, the large columns were 1100 mm long with square cross-section of side 250 mm for the solid specimens and square cross-section of side 350 mm for the voided ones. The void was created from a single 265 mm diameter rigid PET water bottle of 19-L capacity that was placed within the central region of the specimen. In this case, the void constituted about 60% of the core area, with most of the remaining core concrete located in the vicinity of the corners where the rebars were positioned. The first pair (denoted by 1’-S and 1’-V) served as control, utilizing relatively low nominal concrete strength of 20 MPa, 4 ϕ 12 mm longitudinal bars, and ϕ 10 mm transverse ties at 200 mm spacing. The second pair (2’-S and 2’-V) included larger amount of longitudinal steel in the form of 4 ϕ 16 mm bars instead of 4 ϕ 12 mm. The third pair (3’-S and 3’-V) contained lateral steel ties that were closely spaced at 100 mm instead of the 200 mm employed in the control. The fourth pair (4’-S and 4’-V) employed higher nominal concrete strength of 35 MPa instead of the 20 MPa used in control. Figure 2 shows the dimensions and characteristics of column specimens included in group 2.

Prior to casting, two 10 mm long strain gauges were placed on two diagonally opposite longitudinal rebars after grinding the surface to determine the strain in the steel rebars during testing. In addition, one 10 mm strain gauge was mounted on the surface of a tie located at mid-depth within the central region of the columns. All preliminary tasks, including formwork preparation, steel cage assembly, concrete mixing and placement, vibration, and curing were carried out at a local company that specializes in concrete work in the emirates of Sharjah, UAE.

Table 1. Details of the specimens used in the experimental program.

Column Label	Cross-Section	L (mm)	b (mm)	f’c (MPa)	s (mm)	As (mm2)	Dv (mm)
1-S	Solid	900	200	20	150	452	-
1-V	Void	900	220	20	150	452	100
2-S	Solid	900	200	20	150	804	-
2-V	Void	900	220	20	150	804	100
3-S	Solid	900	200	20	75	452	-
3-V	Void	900	220	20	75	452	100
4-S	Solid	900	200	35	150	452	-
4-V	Void	900	220	35	150	452	100
1’-S	Solid	1100	250	20	200	452	-
1’-V	Void	1100	350	20	200	452	265
2’-S	Solid	1100	250	20	200	804	-
2’-V	Void	1100	350	20	200	804	265
3’-S	Solid	1100	250	20	100	452	-
3’-V	Void	1100	350	20	100	452	265
4’-S	Solid	1100	250	35	200	452	-
4’-V	Void	1100	350	35	200	452	265

includes the details of the specimens. In the table, L denotes the length of the specimen, b is the side dimension of the square cross-section, actual f’_c is the target or nominal concrete compressive strength, s is the lateral tie spacing, A_s is the total cross-sectional area of the longitudinal steel reinforcement, and D_v is the diameter of the void made from the PET bottle(s).

3.1. Specimen Fabrication

Before the casting of the concrete, the voids that were made from PET bottles needed to be prepared. For the small voids, each column included four 1.5-L soda bottles (diameter of 100 mm and wall thickness 0.5 mm) that were cut and tightly stacked together to create a perfect cylindrical shape of diameter 100 mm and length 300 mm that was placed within the central part of the column. As for the larger columns, a single 19-L water dispenser plastic cylindrical bottle (diameter of 265 mm, height of 300 mm, and wall thickness 2.5 mm) was used as a void, which was positioned within the middle region of the column specimens. The percentage of plastic within the net concrete cross-sectional area was about 1.5% for the small void specimens and 3% for the large void specimens. Compared to a solid cross-section having the same outside dimensions, the area of the void occupied about 16% of the gross cross-sectional area in the small specimens and 45% of the gross cross-sectional area in the large specimens. Figure 3 shows the process of preparing the small voids from the PET bottles used in the small specimens. In practice, it is anticipated that the method of making voids from PET bottles would be automated in a factory.

The column specimens used in this study were fabricated at a local company in the emirate of Sharjah that specializes in precast concrete work. Care was exercised in the making of the specimens in order to reduce the presence of anomalies. The manufacturing process started by making the steel cages without securing the middle ties in place to create space for the installation of the void. Once the void was installed and secured at its specified location within the central part of the specimens by steel wires, the ties were moved into their place and attached to the longitudinal bars. Two strain gauges were then installed on diagonally opposite longitudinal rebars after surface smoothing by grinding and covered with water-proofing material. The assembled steel cage was then placed inside a wood formwork oriented in a horizontal position and containing proper spacers placed on the sides and bottom to ensure proper concrete cover on reinforcement. The reason for the horizontal concrete casting was to achieve better-quality concrete work by conveniently placing the freshly cast concrete in the formwork on a vibrating bed. The formwork was then placed on a large shaking bed, filled with concrete, and vibrated to remove air pockets and voids from the concrete mix and ensure a strong bond between the rebars and surrounding concrete. Figure 4 shows the fabrication of the specimens.

3.2. Material Properties

This section focuses on the stress–strain graphs from tests on concrete and steel samples. These relationships will aid in understanding the experimental test results and predicting the structural response of the columns in the theoretical component of this study.

3.2.1. Concrete

For the concrete, 12 small-scale samples from each mix were collected on both casting days. One-half of the samples were 150 mm cubes and the other half were 150 mm by 300 mm cylinders, with each shape further divided into two groups: one for a target (i.e., nominal) 20 MPa strength and the other for a target 35 MPa strength. The mix design proportions for the 20 MPa concrete consisted of 1009 kg of coarse aggregate, 868 kg of fine aggregate, 338 kg of cement, and 186 kg of water. On the other hand, the mix design proportions for the 35 MPa concrete consisted of 959 kg of coarse aggregate, 764 kg of fine aggregate, 449 kg of cement, and 225 kg of water. All samples underwent compressive strength testing using a compression testing machine at a displacement-controlled loading condition on the date of the column test following the relevant ASTM standard [25]. Figure 5 and Figure 6 show the stress–strain relationships for the cube and cylinder samples, respectively. The actual average cube and cylinder concrete compressive strengths at the time of testing for the 20 MPa and 35 MPa concrete nominal strengths are provided in

Table 2. Summary of actual concrete compressive strength results.

Column Type	Average Low-Strength Concrete Compressive Strength (MPa)		Average High-Strength Concrete Compressive Strength (MPa)
	Cube	Cylinder	Cube	Cylinder
Small-Section Columns	32.36	28.15	42.04	38.35
Large-Section Columns	24.62	21.83	39.06	35.02

. The mass density of the considered low- and high-strength concrete averaged about 2400 kg/m³.

3.2.2. Steel

For the steel bars and ties used in this study, three 210 mm samples were taken from each of the considered steel reinforcements with diameters of 10 mm, 12 mm, and 16 mm, resulting in a total of nine samples. These samples were tested in a tensile test under a displacement-controlled environment using a UTM following the relevant ASTM standard [26]. After testing, the data were collected and plotted in a stress–strain curve to assess the material properties. Figure 7 below shows the stress–strain curves obtained from the tests of the steel rebar samples. Note that the 10 mm rebars had a yield strength of 600 MPa while the 12 mm and 16 mm rebars had a yield strength of 550 MPa.

3.3. Instrumentation and Test Setup

Following the fabrication of the column specimens and transportation from the casting place to the laboratory testing facility, further preparation of the samples for testing was carried out. First, the samples were painted white and 50 mm gridlines were drawn on the columns to enhance crack visibility and tracking during the test. Subsequently, holes were drilled into the columns to accommodate the screws and bolts required for securing the LVDTs at their supports. It is to be noted that the LVDTs were placed at the center of the specimens between two points within the central part of the columns that were 280 mm apart. This spacing was selected to monitor the region of interest where the 300 mm long void is located. Furthermore, external steel channels were attached and tightened at the top and bottom ends of the columns to confine such regions and discourage crack formation at the extremities of the specimens due to possibility of stress concentration. This mechanism made the central region weaker than the ends and ensured that the LVDTs and strain gauges could provide accurate readings. Finally, each column was placed inside a 2400 kN capacity UTM in the structural laboratory at the American University of Sharjah and subjected to a purely axial compression in a displacement-controlled test environment. Figure 8 shows a schematic diagram of the test setup.

4. Results and Discussion

A displacement-controlled loading protocol was implemented to capture the pure axial compressive behavior of the columns, including beyond their ultimate strength. At a loading rate of 0.3 mm/minute, the duration of each test averaged approximately 30 min. The failure criterion for the test was defined as a 50% drop in the ultimate load after the attainment of the peak capacity. Following the conclusion of the tests, additional destructive loading was applied manually through the test machine to the columns in order to emphasize the failure mode. Figure 9 and Figure 10 show, respectively, the eight column specimens with small cross-sections and eight large cross-sections following the conclusion of the tests.

Four LVDTs were positioned on all side faces between two points that were 280 mm apart and located in the central region of the columns within the mid-height part. These devises were needed to acquire axial displacements within the critical region of the column during the test since deflection readings obtained from the UTM do not address the nonuniformity of the displacement along the column length. Additionally, three strain gauges were attached to the steel cages—two placed on diagonally opposite longitudinal rebars and one on a tie located within the central region of the column. The two strain gauges on the rebars were needed to record the axial strain in the rebars during testing, while the strain gauge on the tie was required to capture the lateral expansion of the concrete core due to the Poisson effect.

Figure 11 shows the load versus strain relationships for the eight small-cross-section columns and eight large-cross-section columns, in which positive strain corresponds to tension in the lateral ties while negative strain corresponds to compression in the longitudinal bars. Note that the strain gauges that were not properly working during the tests are excluded from the plot. Also, only one of the two strain gauges that were mounted on the longitudinal bars was considered in the figure in order to preserve the clarity. Overall, the results indicate that out of 12 functioning strain gauges that were installed on the transverse reinforcements, only 2 of them reached the yielding strain of 0.0025 and beyond. With regard to the maximum strain in the longitudinal reinforcement, 11 out of the 16 tested columns had at least one strain gauge showing a strain level above the yield value of 0.0025.

Data from the tests were used to generate uniaxial load–displacement relationships for all the 16 considered columns in this study. Generally, all specimens exhibited a linear increase in load versus displacement up to approximately 50% of the ultimate compressive capacity, followed by a nonlinear slow increase with a progressively decreasing slope until reaching the ultimate load. After reaching the peak capacity, the load either decreased suddenly in a brittle fashion or nonlinearly with a negatively decreasing slope until the total collapse of the specimen, often resulting in an incline shear failure plane.

Crack propagation on the specimens was documented through photographs that were taken during the tests. In general, the cracking patterns revealed a nonuniform crack formation on almost all sides of the columns, indicating no eccentricity from the actuator platen. Furthermore, the ultimate compressive strength values of the specimens displayed a consistent trend that was closely aligned with the expected theoretical capacity across all the specimens. As expected, columns with a higher concrete strength and larger-diameter rebars exhibited a higher ultimate strength than corresponding ones with a lower concrete strength and smaller-diameter rebars.

4.1. Load–Strain Relationship

Figure 12 shows the overall load–strain relationships for the eight small- and eight large-cross-section columns. As seen in both figures, all graphs followed the typical load–displacement behavior of a concentrically loaded reinforced concrete column. The curves begin with an initial close-to-linear ascending region for both small and large columns. This is followed by a gradual decline in stiffness until reaching the ultimate strength, after which the load-carrying capacity decreases until failure occurs, typically when the final load drops to approximately half of the ultimate load. Most of the tested column specimens exhibited some form of residual compressive strength after the peak capacity. In general, voided columns demonstrated load–displacement behavior similar to their solid counterparts, with the smaller-cross-section set of columns showing a closer correlation than the larger-cross-section column set.

For the small-cross-section columns, the load–deflection graphs initially overlap, showing a steep semi-linear ascending and nearly identical initial slope for all specimens. As the load approaches the ultimate value, the curves become nonlinear, reaching a maximum capacity at strain levels between 0.002 and 0.004. After the peak, they start to diverge, leading to different post-peak behaviors that had an impact on the ductility and residual strength of the columns. As expected, the highest recorded load was 1250 kN for specimens 2-S and 4-S, attributed to their larger rebars (16 mm instead of 12 mm) and stronger concrete (35 MPa nominal instead of 20 MPa). The lowest recorded load was 950 kN for specimen 1-V, the voided column with the weakest attributes (low concrete strength, a small reinforcement ratio, and/or large lateral tie spacing). When comparing the overall behavior of solid and voided columns, a slight reduction in performance was observed for the voided columns. This reduction is attributed to their unconfined central concrete core due to the presence of the void. With regard to the residual strength, both solid and voided columns demonstrated a similar ability to retain a significant portion of their ultimate strength after the peak load. Specifically, specimens 3-S and 3-V that contained closely placed ties exhibited extended post-peak load retention before experiencing a decline. This behavior is attributed to the confining effect of the lateral reinforcement, which can resist crack penetration into the core, enhance the structural ductility, and delay strength degradation under increased axial deformation.

When examining the large-cross-section columns, a similar trend is observed in their overall behavior, as all specimens exhibited nearly identical initial stiffness. As with the smaller columns, divergence in the results occurred after the stiffness began to decline under increased loading, continuing until the column reached its ultimate load, followed by a decrease in the load capacity until it dropped to approximately 50% of the ultimate load. However, compared to the smaller columns, the variation between the results in the large columns was more pronounced, with a wider envelope encompassing the data. This discrepancy is likely due to the proportionally larger voids in the large-cross-section columns relative to their overall size. To elaborate, the small columns contained a 100 mm diameter void within a 200 mm × 200 mm cross-section, whereas the large columns had a 265 mm diameter void within a 350 mm × 350 mm cross-section. Relative to the gross cross-sectional area, the created void occupied about 20% of the small-cross specimens and 80% of the large-section specimens. With regard to the core area that is bound by the tie, the void occupied about 30% of the core in the small specimens and 60% of the core in the large specimens. Compared to the smaller columns, the greater void percentage in the larger columns have significantly impacted the structural response and contributed to increased variability in their behavior. Also, the large void within the core of the columns that have large cross-sections resulted in most of the concrete being located within the unconfined cover, which made it susceptible to a deficiency in strength and ductility.

In general, columns with higher concrete strength performed the best, with specimens 4’-S and 4’-V achieving ultimate loads of 1800 kN and 1550 kN, respectively. The lowest recorded ultimate load was 930 kN for specimen 1’-V. Comparing the overall behavior of solid and voided large columns revealed several key observations. A significant reduction in the load-carrying capacity was observed in specimens 1’-V and 3’-V compared to their solid counterparts, 1’-S and 3’-S, with ultimate loads dropping from approximately 1350 kN to around 1000 kN. Conversely, columns 2 and 4 showed only minor differences in their load-carrying capacity between the solid and voided specimens. The load capacity decreased from 1650 kN to 1300 kN for 2’-S and 2’-V, and from 1800 kN to 1550 kN for 4’-S and 4’-V. As previously mentioned, all modified parameters resulted in a better performance compared to the base samples, 1’-S and 1’-V. The specific differences and impacts of the individual parameter will be further analyzed in the following sections.

4.2. Effect of Longitudinal Reinforcement Ratio

Clearly, an increase in the size and number of longitudinal bars in a column will result in higher strength. In the experimental investigation, 14 column samples contained four 12 mm diameter longitudinal steel rebars and 4 column samples contained four 16 mm diameter rebars (2-S, 2-V, 2’-S, 2’-V). The increase in the rebar size from ϕ 12 mm to ϕ 16 mm amplified the total cross-sectional area of the steel within the cross-section by 78%. Figure 13 presents a comparison between both solid and voided columns numbered 1 and 2 for both small- and large-section columns. Obviously, the primary observation is that increasing the rebar size significantly enhanced the load-carrying capacity of the voided columns. For the small columns, increasing the rebar size from ϕ 12 mm to ϕ 16 mm resulted in a noticeable increase in the maximum load-carrying capacity. The solid columns exhibited an approximately 24% increase in capacity, rising from 1000 kN in 1-S to 1245 kN in 2-S. Similarly, the voided columns showed an about 16% increase, with the capacity rising from 950 kN in 1-V to 1100 kN in 2-V. This improvement is expected due to the increased reinforcement area. However, the smaller relative increase in voided columns can be attributed to the absence of a core, leading to reduced confinement and increased stress on the rebars. A similar trend was observed in the columns that have a large cross-section. The solid columns exhibited a substantial increase in the load capacity, rising from approximately 1400 kN in 1’-S to 1650 kN in 2’-S, marking a 20% increase. Meanwhile, the voided columns experienced a remarkable 40% increase, with the capacity rising from 930 kN in 1’-V to 1305 kN in 2’-V. The greater increase in the voided large-section columns can be attributed to the significant size of the voids, which left little confined concrete that possessed a low capacity. Consequently, the rebars played a much larger role in load resistance, making the increase in the rebar size particularly effective. When comparing the solid columns to their voided counterparts, a clear trend was noticeable. For the small columns, the transition from solid to voided specimens resulted in moderate reductions in the load-carrying capacity, with 1-S and 2-S experiencing decreases of 5.5% and 11.5%, respectively. However, the large columns exhibited more substantial reductions, with 1’-S and 1’-V showing a 32% drop and 2’-S and 2’-V experiencing a 21.1% decrease. These larger reductions in the columns with a large cross-section are primarily due to the significantly larger voids present, which led to a further loss of confinement and structural capacity.

4.3. Effect of Transverse Reinforcement Ratio

Lateral reinforcement in columns contributes towards the confinement of concrete within the core, which in turn impacts the strength and ductility of the member. In this study, all small and large specimens included a 150 mm and 200 mm tie spacing, respectively, within the central region where the void is located, except for those labeled 3 (3-S, 3-V, 3’-S, and 3’-V), which had a reduced tie spacing of 75 mm for the small columns and 100 mm for the large columns. Figure 14 presents a comparison between the solid and voided columns labeled 1 and 3 for both small and large specimens. The primary observation from these results is that decreasing the tie spacing had a minimal impact on the uniaxial compressive capacity of the columns. For the solid columns, reducing the tie spacing from 150 mm to 75 mm resulted in a slight increase of 6.5% in the load capacity, rising from 1000 kN in 1-S to 1070 kN in 3-S. A similar trend was observed in the large solid columns, with load capacities of 1380 kN in 1’-S and 1370 kN in 3’-S, indicating negligible variation. For the voided columns, a comparable trend was observed, although the effect of closer ties was slightly more pronounced. The load capacity increased by 5.9% when comparing 1-V (950 kN) to 3-V (1005 kN), while the large voided columns showed a 7.7% increase, rising from 930 kN in 1’-V to 1000 kN in 3’-V. This limited effect aligns with expectations that are based on prior tests, as reducing tie spacing primarily enhances the ductility and deformation capability of columns rather than significantly influencing their load-carrying capacity. When comparing solid with voided columns, some key observations emerge. The small columns exhibited a minimal decrease in the load-carrying capacity, approximately 6%, when transitioning from solid to voided specimens. However, the large columns experienced a more substantial reduction, with the load capacity decreasing by approximately 30%. This substantial drop in strength can be credited to the presence of exceptionally large voids, which resulted in the minimal confinement of the core.

4.4. Effect of Concrete Compressive Strength

In lightly reinforced concrete columns subjected to concentric loading, the concrete strength contributes the most towards the compressive strength of the member. In this study, 12 of the 16 columns were made with lower-strength concrete with a target compressive strength of 20 MPa, while the remaining four columns labeled 4 (4-S, 4-V, 4’-S, and 4’-V) were made with higher-strength concrete with a target strength of 35 MPa. As reported earlier, the actual lower and higher concrete cylinder strengths for the columns with a small cross-section at the time of testing were 28.15 MPa and 38.35 MPa, respectively. The corresponding actual strength for the columns with a large cross-section were 21.83 MPa and 35.02 MPa, respectively. Figure 15 presents the load–displacement graphs for the column specimens numbered 1 and 4 for both small and large column specimens. As expected, the results clearly indicate that increasing the concrete strength enhanced the load-carrying capacity of all specimens, although not by the same proportion. For the small-section columns, when the concrete compressive strength increased from 28.15 MPa to 38.35 MPa, the load-carrying capacity increased by 20% and 11% for the solid and voided specimens, respectively, with values rising from 1005 kN to 1210 kN for solid columns and from 950 kN to 1050 kN for voided columns. The large columns exhibited an even greater increase, with load capacities rising from 1380 kN to 1830 kN for solid columns and from 930 kN to 1550 kN for voided columns, reflecting increases of 32% and 67%, respectively. The substantial improvement observed in the columns with a large cross-section due to a change in the concrete strength can be attributed to the fact that the major source of strength in such columns came from the concrete, with little contribution from the longitudinal steel that amounted to 0.72% of the gross area (compared to 1.1% of the gross area in the small-section columns). When comparing the solid and voided specimens within group 4, a slight reduction in strength was observed due to confinement issues. The load capacity decreased by 13% when transitioning from the solid sample 4-S to the voided sample 4-V and by 15% when moving from the solid sample 4’-S to the voided sample 4’-V.

4.5. Ductility Analysis

Ductility is a very important property when it comes to structural behavior because ductile members have the ability to deform to some degree without rupturing, thus providing warning to users prior to collapse. Several methods have been proposed in the literature to quantify displacement ductility using a ductility index. One of the most common and straightforward approaches for measuring ductility involves the ratio of column shortening corresponding to 85% of the peak capacity on the descending part of the load–deflection relationship, denoted by (Δ_0.85)_post, to the column shortening corresponding to 85% of the peak capacity on the ascending part of the load–deflection relationship, denoted by (Δ_0.85)_pre, as shown below [27]:

$$\mu ={\displaystyle \frac{{{(∆}_{0.85})}_{post}}{{{(∆}_{0.85})}_{pre}}}$$

(1)

As expected, this ductility measure implies that the wider the load–deflection curve, the higher the corresponding ductility of the structural member. Figure 16 below shows the displacement ductility indices for all 16 column samples.

As a general observation, all displacement values ranged between 2 and 3.5, with an average of approximately 2.6. A key trend observed was that voided columns with at least one enhanced parameter (columns numbered 2, 3, and 4) exhibited ductility behavior closely resembling their solid counterparts. For example, pairs 2-S and 2-V, 4-S and 4-V, 2’-S and 2’-V, and 3’-S and 3’-V demonstrated comparable ductility performance. Their respective ductility values were 2.40 and 2.19, 2.44 and 2.09, 3.09 and 3.19, and 2.25 and 2.29. The highest recorded ductility value was 3.5, observed in 3-S. This result aligns with expectations, as 3-S is a solid column with closely spaced ties, leading to enhanced confinement and increased ductility. This trend was also evident in voided columns, with 3-V achieving the highest ductility among voided specimens. The largest reductions in the ductility index from a solid column to the corresponding void column were observed in the pair of specimens labeled 1, 3, 1’, and 4’, leading to reductions of 28%, 24.5%, 27%, and 34.3%, respectively. The largest reduction in the ductility of column 4’-V from its 4’-S counterpart could be due to the combined effect of a small amount of core concrete, weak confinement within the core, and the brittle nature of the higher-strength concrete used in the specimens. For columns with more of a longitudinal reinforcement ratio, the voided columns performed well compared to their solid counterparts. Sample 2-V exhibited only a minor drop in ductility relative to 2-S, while 2’-V unexpectedly showed a slight increase in ductility compared to 2’-S. Lastly, for columns with stronger concrete (group 4), the small specimens 4-S and 4-V displayed very similar results, yielding ductility values of 2.43 and 2.09, respectively. This outcome suggests that stronger concrete compensated for the reduced confinement in 4-V, leading to a comparable performance. However, for large specimens, a significant reduction in ductility was observed when transitioning from 4’-S to 4’-V, decreasing from 3.09 to 2.03. This drop is likely due to the large void size, which resulted in a lack of confined concrete and, consequently, lower ductility.

4.6. Analysis of Stiffness at Service Load Level

The initial stiffness of columns along the ascending part of the load–displacement relationship is an important property that is often related to the serviceability limit state, particularly deflection under service loads and prior to cracking. It is determined as the slope of the straight line that joins the origin to the point on the load–deflection curve corresponding to 50% of the peak load. In this study, the 50% value was selected as an approximate representation of the service load level. Note that the displacement in the considered graph is the average axial compressive net displacement between two points located 280 mm apart within the central region of the column, obtained from the four LVDTs that are mounted on the four surfaces. Figure 17 shows a summary of the stiffness at a service load level for all the tested columns.

From the results, multiple different observations can be made. Firstly, the voids had an overall slight negative impact on the stiffness of the columns due to the presence of relatively thin regions within the cross-section. The average stiffness for the eight small-section samples was 2656 kN/mm, whereas the average stiffness for the eight large-section samples was found to be 3859 kN/mm. The largest stiffness value found was 5854 kN/mm, belonging to the solid 4’-S column. This was an expected outcome because it is the largest solid specimen that was made with the strongest concrete. The lowest stiffness value was 2105 kN/mm, belonging to the voided 3-V column. The largest drop in stiffness between a solid column and its voided counterpart was 34%, observed in the 1’-pair and 2’-pair. These drops were expected because they occurred in the voided columns with large-cross-section samples which included a large void that greatly reduced the confinement and, hence, the resulting stiffness of these samples. On the other hand, the smallest drop in stiffness between a solid column and its voided counterpart belonged to the two-pair, in which the stiffness was virtually identical in the solid and its equivalent voided columns. This can be attributed to the fact that bigger rebars were able to offset any negative effects that came with the presence of voids.

4.7. Residual Strength Analysis

Residual strength is a measure of the ability of a column to support extreme loads without failing, which is critical for maintaining the overall stability of the structural system. It can be a useful structural attribute in determining the cause of a structural fault, assessing potential damage to a structure and evaluating the feasibility of repairing the structure. In this study, the residual strength for the tested columns is quantified by considering the ratio of the load on the descending part of the load–deflection relationship corresponding to twice the displacement at peak load to the deflection at peak load. Figure 18 shows a summary of the results of the residual strength, presented as a percentage of the ultimate capacity, for all column specimens considered in this study.

Overall, the results show a beneficial residual strength for all specimens, amounting to an average of 65%. When comparing the overall behavior of voided columns to their solid counterparts, it was found that for the most part, the voided columns were similar and, in some cases, better than their corresponding solid columns, with an average difference of just 5%. The largest recorded residual strength was 84% for column 3-S, which had a solid cross-section and contained closer tie spacing, which helped with the overall confinement of the course following clear cover spalling. The lowest observed residual strength values were for column specimens 4-V, 1’-V, and 3’-S. The low values for 4-V and 1’-V are due to the lack of confinement as a result of the embedded void, which affected the ability of these columns to retain loads at a large displacement level. With regard to column 3’-S, although the LVDTs that were mounted on the surfaces of this column became detached shortly after attaining the peak capacity due to extensive cracking within the LVDT region, the load from the UTM was found to be constant at 686 kN, which was used to find the residual strength value.

4.8. Impact of Void on Structural Behavior

To fully evaluate the impact of a void on the structural behavior of a column, the areas of the concrete core and cover within the cross-section need to be identified and calculated. This is because the strength and ductility of confined concrete within the core are more enhanced than the corresponding ones of the unconfined concrete within the cover. Figure 19 identifies the concrete core and cover areas within the cross-section of the solid, small-void and large-void specimens considered in this study.

Table 3. Information on the concrete and void areas for the solid and void specimens.

Area Component	Type of Section
	Small Solid	Small Void	Large Solid	Large Void
Core Concrete (mm2)	22,500	21,046	40,000	34,845
Cover Concrete (mm2)	17,500	19,500	22,500	32,500
Void Area (mm2)		7854		55,155
% Void within Core		27.2%		61.3%

shows the areas of concrete within the cover and the core (including the area of the longitudinal rebars), as well as the areas of the void (when applicable), for the small and large solid and void specimens. Relative to the corresponding solid specimens, the results show that the amount of concrete within the core in the void specimens are much smaller in the large column samples than in the small ones. For the small columns, the percentage of the void area relative to the total core area is 27.2%. Conversely, for the large columns, the percentage of the void area relative to the total core area is 61.3%. Hence, there is a much higher fraction of confined concrete in the voided small specimens than in the larger ones. Such an outcome explains the relatively better structural performance of the specimens that contain a small void compared with the specimens containing a large void.

4.9. Comparison with ACI 318-25 Code Equation

An experimental study would not be complete without the consideration of a theoretical component. In this section, the ultimate compressive strength results are compared with those obtained from the provisions of the ACI 318-25 [24] code for the case of pure axial loading,

$$P_n=0.80\{0.85{f^{’}}_c\left(A_g-A_{st}\right)+f_y{A}_{st}\}$$

(2)

where f’_c is the actual concrete compressive strength, A_g is the area of concrete within the cross-section, A_st is the total area of the steel rebars, and f_y is the actual yield strength of the steel. Although it is not explicitly stated in the ACI code, Equation (2) is assumed to be more suitable for regular columns with a reasonable size of a core that is free of regions of a stress concentration. The ratios of the experimental-to-theoretical ultimate strength for all 16 tested column specimens are shown in Figure 20.

As expected, the results showed that the majority of the columns had an experimental-to-predicted strength ratio larger than 1.0. Overall, the calculated ratios ranged between 0.80 and 1.34, with an average ratio equal to 1.08. All but one of the eight solid columns had a strength ratio above 1.0. On the other hand, four of the eight voided columns had a ratio below 1.0. Furthermore, three of the four voided columns whose strength was over-predicted by the ACI 318 code equation were associated with the large void group. Based on the findings of this theoretical study, it can be concluded that the uniaxial compressive strength equation in the ACI 318 code may not be applicable to columns embedded with voids that occupy a large fraction of the core. Due to the limitations of the current study and relatively small number of tested specimens, the authors believe it is inappropriate at this stage to propose modifications to structural code equations.

5. Conclusions

This study investigated the potential implementation of Polyethylene Terephthalate (PET) bottles as void formers in non-slender reinforced concrete columns subjected to pure axial compression. Such an approach reduces the volume of concrete and offers a sustainable disposal method for plastic bottles made from a non-biodegradable material. The findings of this study lead to the following conclusions:

• This study successfully introduced a novel method for the permanent disposal of PET plastic bottles that can be embedded inside reinforced concrete columns with the help of wires that are attached to the steel cage. This approach contributes to the sustainable construction mandate by reducing plastic waste accumulation in landfills.
• Columns containing small voids that occupied about 30% of the core exhibited a structural performance somewhat comparable to their solid counterparts. The load–displacement curves showed a maximum of a 9% reduction in the axial load-carrying capacity relative to their equivalent solid complements. Reductions in the stiffness, ductility, and residual strength were also reasonably small at 14%, 20%, and 1%, respectively, indicating that the introduction of small PET bottle voids does not significantly compromise the columns’ performance.
• Columns containing large voids that occupied about 60% of the core demonstrated a significantly weaker performance than the corresponding solid ones. The inclusion of large PET voids led to decreased confinement, which adversely affected the axial load–displacement behavior. On average, the reductions in the stiffness, ductility, and load-carrying capacity were 24%, 34%, and 26%, respectively, with a 5% increase in residual strength, indicating poor structural viability for large voids.
• As expected, increasing the longitudinal reinforcement size from ϕ 12 mm to ϕ 16 mm improved the load-carrying capacity of both solid and voided columns. However, the benefit was more pronounced in the solid columns, highlighting the diminished effectiveness of the reinforcement in poorly confined (voided) members.
• Reducing the ϕ 10 mm tie spacing from 150 mm to 75 mm in the small columns and from 200 mm to 100 mm in the large columns enhanced the confinement, leading to higher stiffness and ductility. This was particularly noticeable in the voided specimens, thus indicating that closely spaced ties can partially mitigate the adverse effects of internal voids.
• As expected, columns made with relatively higher-strength concrete demonstrated an improved overall performance in terms of the load-carrying capacity and stiffness. This enhancement was consistent across both solid and voided columns, although the effect was more pronounced in the solid than in the void columns.
• The ACI 318 building code provisions concerning the compressive strength of concentrically loaded reinforced concrete columns can be reasonably used to predict the ultimate capacity of columns embedded with voids made from PET bottles if the size of the void does not exceed 30% of the core area.

In summary, this study showed that the use of PET drinking water bottles as void formers is practical and can be effective in reinforced concrete columns, as long as the area of the void within the cross-section does not exceed about 30% of the core area that is bound by the ties. It should be noted that the above conclusions are based on the testing of one sample of each investigated parameter, which is common practice in research involving large-scale testing. They are also applicable to concentrically loaded, lightly reinforced concrete columns having square cross-sections embedded with cylindrical voids made from PET bottles. While the manual process of making voids from PET bottles in the laboratory was easy, practical, and fast, the authors believe that automation in a factory may be needed for large scale applications. Future studies on the subject may address eccentrically loaded members, different cross-section and void shapes, and a wide range of material properties and reinforcement ratios. Additionally, durability concerns and long-term effects due to creep and shrinkage may also be considered by researchers interested in pursuing this topic further in the future.