1 Introduction

The next few decades are predicted to see a peak in urbanisation due to the rapid growth of the global population [1]. This will demand constant upgrades in the urban infrastructure, especially in the building and transportation sectors. For instance, the growing population will cause increased traffic volumes and vehicle axle loads, requiring the construction of new transport pavements as well as the improvement and maintenance of existing structures [2, 3]. The addition of concrete pavements to the road network is a favourable option due to a faster construction, higher durability, lower maintenance expenditures, and thus, incur a lower cost over their life [4,5,6,7,8,9] when compared to asphalt pavemnts.

A concrete pavement consists of base and subbase layers, typically constructed with Ordinary Portland Cement (OPC) based concrete [10, 11]. Though OPC makes up only about 10% of a typical pavement-grade concrete mix, it causes almost 70% of emissions associated with concrete production [1]. In fact, cement production is responsible for 36% of CO2 emissions in the construction industry and around 8% of the annual total anthropogenic CO2 emission in the world [1, 12]. Therefore, a major reduction of anthropogenic environmental impacts can be achieved by introducing low-carbon alternatives to OPC in pavement construction [13, 14].

The use of Supplementary Cementitious Materials (SCMs) in concrete is a widely accepted strategy to reduce the carbon footprint in the construction industry while achieving the same performance [3, 15]. Finely ground limestone without any heat treatment (i.e., uncalcined) is the commonly used SCM to blend with OPC [1]. However, lime as an SCM is limited to a maximum substitution of around 10% to 15% due to its low reactivity [1]. Fly ash, a by-product from coal power generation, and slag, a by-product from steel manufacturing, are the next most used materials [1, 16,17,18]. These materials are good candidates for a partial or full replacement of OPC binders. However, with the current trends of shifting from coal power generation to newer methods like natural gas; and with steel production shifting over to steel recycling, the availability of these SCMs is likely to decrease in the future [1, 7]. Therefore, research on the identification of new materials suitable for alternative binders development is needed.

After lime, fly ash, and slag, clay has been recognised as the next best potential SCM in the construction industry, considering its wide availability and suitability [1]. Using clay as an SCM is expected to reduce current emission levels associated with the cement industry by 15% to 30% [1].

However, to obtain a reactive form, clay needs to be calcined at high temperatures, (above 650ºC) [19,20,21]. Calcination of clay in large quantities increases the total energy consumption as well as the CO2 emissions during the binder production process. Therefore, incorporating already calcined forms of clay can make binder production significantly more environmentally friendly. In this scenario, waste clay bricks (WCBs) are an ideal candidate, since they have been previously calcined during the brick manufacturing process at temperatures around 850 °C [7, 22,23,24]. Further, large numbers of WCBs are generated worldwide as construction and demolition waste [22, 25, 26]. They are also more likely to be sent to landfills due to the lack of properly established reuse or recycling applications [27]. However, until recently, WCB was investigated only as an aggregate replacement in limited concrete applications.

Recent studies have identified WCBs as a promising alternative or supplementary material to produce low-carbon binders using geopolymer technology [15, 23, 28, 29]. Materials rich in alumina and silica (i.e., aluminosilicate source materials) and alkaline activators are used as the two main ingredients in geopolymer binder preparation. These binders are formed when dissolved silicon (Si) and aluminium (Al) from the surface of the aluminosilicate source particles (or from surface hydration of undissolved particles) undergoes geopolymerisation to form a 3-D aluminosilicate network [30, 31].

The mechanism of geopolymerisation can be explained in terms of three main stages as shown in Fig. 1. First, the dissolution of oxide minerals such as Al2O3 and SiO2 in source materials under highly alkaline conditions initiates the geopolymerisation mechanism [32]. At the next stage, coagulation or gelation of these dissolved mineral oxides occurs, finally leading to polycondensation.

Fig. 1
figure 1

Stages of geopolymerisation

Full size image

The polycondensation process results in a 3-D amorphous aluminosilicate network with a strength similar to or even higher than OPC [33]. The 3-D aluminosilicate network can consist of three different molecular structures based on its chemical bonding; (1) poly(silate) (–Si–O–Al–O–), (2) poly(sialate-siloxo) (–Si–O–Al–O–Si–O–), (3) poly(sialate-disiloxo) (–Si–O–Al–O–Si–O–Si–O–) [34]. The general empirical formula for the 3-D tecto-aluminosilicate framework is denoted by.

Mn[-(SiO2)z–AlO2]n:wH2O,where M is Potassium (K), Sodium (Na), or Calcium (Ca), n is the degree of polycondensation, and z is 1, 2, 3, or >  > 3 [35].

Although there are several studies on developing WCB-based geopolymer binders and geopolymer mortar, research on concrete with WCB as a binder material is very limited. The study by Migunthanna et al. [24] recommends WCB powder as a partial replacement for geopolymer binders to synthesise pavement concrete. Preliminary investigations of using these binders in geopolymer concrete (GPC) production are reported only in the study by Migunthanna et al. [36]. According to the Australian rigid pavement design standards, concrete mixes with 28-day compressive strengths of not less than 5 MPa and 32 MPa can be used as a pavement subbase material and base material, respectively [10, 37]. The GPC mix developed by Migunthanna et al. 2022 [36] was capable of exceeding both these design strength requirements. Therefore, comprehensive studies on these GPC mixes were needed to further verify the suitability of using WCB-based binders in pavement concrete.

When developing concrete trial mixes for rigid pavement applications, it is recommended to have a 28-day target compressive strength of 40 MPa [38]. On the other hand, 28-day concrete flexural strength is a critical parameter in rigid pavement design since it is used as the design strength in determining base thickness [10]. To evaluate the suitability of these kinds of new binder types, more studies related to the fresh concrete properties (workability and setting time), mechanical properties (compressive strength, flexural strength, split-tensile strength) and durability (water absorption, permeability, shrinkage and so on) are necessary [7, 36].

The main purpose of this study is to develop and characterise the mix designs of WCB-based GPC suitable for rigid pavement construction and assess the effect of different production conditions on strength development. The binder mixes consisting of WCB fly ash, and slag as the precursors; and sodium silicate (Na2SiO3) as the activator, are used to develop GPC and a 28-day strength of at least 40 MPa is targeted. The present work restricts itself to non-chemical strategies to optimise the GPC strength, such as the change of curing conditions and WCB particle size. The durability parameters, namely, water absorption, apparent volume of permeable voids (AVPV), resistance to cyclic wetting and drying, and abrasion of GPC are also investigated.

2 Materials and Experimental Methods

2.1 Geopolymer Binder Composition

Four different geopolymer binder compositions including one binary blend of WCB with slag and three ternary blends of WCB with slag and fly ash were used to develop GPC. Detailed information on the development of these binders, including optimisation of binder constituents, mineralogical, microstructural, and chemical characterisation (i.e., x-ray diffraction analysis, scanning electron microscopy, energy-dispersive X-ray spectroscopy, X-ray fluorescence analysis, Fourier-transform infrared spectroscopy), is available in the previous studies by Migunthanna et al. [24] and [36].

All four one-part binders are composed of 10% activator and 90% of aluminosilicate precursor by weight. Anhydrous Na2SiO3 with an activator silica modulus of 1.0 was used as the sole activator. The composition of the one-part binders and the different contents of WCB, slag and fly ash in the precursor is shown in Table 1 and Fig. 2, respectively. According to previous studies, geopolymer paste samples with these selected binders were able to achieve 28-day compressive strengths of up to 92 MPa and high early strengths by recording more than 50% of its 28-day strengths within 72 h [24, 36].

Table 1 Composition of WCB-based binders used to synthesize GPC
Full size table
Fig. 2
figure 2

Precursor composition of different binders. (FA = fly ash, S = slag)

Full size image

WCB, slag, fly ash, and activators were supplied by Eco-Group Melbourne Australia, Independent Cement & Lime Pty Australia, Cement Australia Pty Limited, and Redox Melbourne Pty Limited, respectively. Table 2 summarises the chemical composition of the precursors obtained by x-ray fluorescence (XRF) analysis.

Table 2 Chemical compounds (% weight > 0.1%) present in precursors
Full size table

2.2 GPC Mix Designs

Table 3 summarises the GPC mix designs investigated in this study. GPC was prepared using the four binders given in Table 2. Each concrete mix was named after the fly ash content present in the corresponding binder used. For example, GPC mix ID ‘FA10’ in Table 3 means the concrete made with the ‘FA10-S60’ binder (Table 2, Fig. 1). Fine aggregate content of 35% from the total aggregates, and coarse aggregates of nominal sizes of 14 mm and 7 mm in a weight ratio of 3:2 were used for all GPC mixes.

Table 3 GPC mix designs
Full size table

2.3 Preparation of GPC

One-part geopolymer binders were prepared by dry mixing the precursors and the activator in quantities corresponding to the appropriate binder type as indicated in Table 3. The total mixing time for binder preparation was five minutes. Then, coarse and fine aggregates were added to the binder, and mixing was continued for another five minutes. Subsequently, water was added to the mixture according to the water-to-binder (w/b) ratio corresponding to the selected binder type, and wet mixing continued for about fifteen minutes. This w/b ratio was specifically chosen to achieve a slump of around 50 mm to 65 mm. The slump test was conducted for all the mixes in accordance with AS 1012.3.1:2014 standard [39]. After measuring the slump, the tested fresh concrete was returned to the mixer and remixed with the rest for another minute.

Fresh GPC was then poured into cylindrical moulds with 100 mm diameter and 200 mm height, and prism moulds of size 75 mm × 75 mm × 285 mm to comply with Australian standards (AS) [40, 41]. The moulds were filled in three layers and each layer was subjected to a mechanical vibration for one minute. Fresh GPC samples were covered with polythene sheets and stored in an environmental chamber at 25 °C ± 2 °C. After 24 h, all the samples were demoulded. Then, the samples were subjected to different storage conditions as described in Sect. 3.2. Figure 3 summarises the parameters for the evaluation of the feasibility of producing WCB-based GPC investigated in this study.

Fig. 3
figure 3

Summary of the experimental work on the investigation of the feasibility of producing pavement grade WCB-based GPC

Full size image

2.4 Laboratory Testing

2.4.1 Setting Time

To estimate the setting time of the binders, the Vicat penetrating needle was used for paste samples with normal consistency in accordance with AS/NZS 2350.4:2004 [42]. Fresh paste samples were filled into a conical mould with a base diameter of 82 mm, top diameter of 75 mm, and height of 40 mm. Penetration of the Vicat needle was recorded after 30 s of releasing the set screw. Based on the above standard, the initial setting time was recorded when the sample first refused to penetrate a depth of mould of more than 39 mm. The elapsed time between the addition of water and the first refusal to penetrate more than 0.5 mm is considered the final setting time. However, the Vicat method can not be used to determine the setting behaviour of fresh concrete.

The setting time of fresh concrete was measured by penetration resistance in accordance with AS 1012.18:1996 [43]. Fresh mortar was obtained by removing the coarse aggregate fraction (> 4.75 mm) from the GPC mix design. The extracted mortar was placed in a plastic container with internal dimensions of 150 mm × 150 mm × 150 mm in two layers and compacted by means of a tamping rod applying 25 blows per layer. A circular needle with a diameter of 9 mm and a flat point was used to measure the penetration resistance. First, the needle-bearing surface was brought into contact with the mortar surface and the penetration resistance apparatus was gradually forced vertically downwards until the needle penetrated to a depth of 25 mm within 10 s (Fig. 4). The penetration resistance was measured in approximately hourly intervals and avoiding tests in which subsequent measurements areas were disturbed by the previous tests. The measurements were continued for around 12 h. The penetration resistance was calculated as the ratio of the force required by the bearing face of the needle to penetrate a 25 mm depth to the area of the bearing surface.

Fig. 4
figure 4

Determination of setting time of fresh concrete by penetration resistance

Full size image

According to AS 1012.18:1996 [43], the initial setting time is defined as the elapsed time from the initial contact of the binder and water until a penetration resistance of 3.5 MPa is reached. The elapsed time from the initial contact with water until reaching a penetration resistance of 28 MPa is defined as the final setting time.

2.4.2 Strength

The compressive strength was recorded for all mixes on triplicate concrete cylinders in accordance with AS 1012.9:2014 [40]. Concrete prisms were used to test the flexural strength of GPC mixes in accordance with AS 1012.11: 2014 [41] standards by using a three-point bending setup with a span of 250 mm (Fig. 5). For the optimized GPC, the development of compressive strength, flexural strength, and split tensile strength at 28 days were recorded.

Fig. 5
figure 5

Three-point bending setup to measure flexural strength of GPC

Full size image

2.4.3 Water Absorption and Permeability

The water absorption and the apparent volume of permeable voids (AVPV) of GPC were determined following AS 1012.21:1999 [44] standards. These parameters are significant for assessing the durability of concrete since they indicate the tendency of the concrete to transport aggressive agents through it.

Test specimens were prepared by cutting cylinder specimens into four equal pieces to obtain cylindrical disks with a 50 mm height and 100 mm diameter. To determine the water absorption, the specimens were oven dried at 110ºC to a constant weight and then immersed in water at room temperature for 48 h. To determine the AVPV and the boiled absorption, the test specimens were then boiled in a water bath for 6 h. The mass of the specimens was recorded at each step: the mass of the oven-dried specimen after cooling (Md), the mass of the Surface Saturated Dry (SSD) specimen at the end of the 48-h immersion period (MSSD), and the mass of the SSD specimen after boiling (MSSDb). Subsequently, the boiled specimens were suspended in water and their apparent masses were recorded (Mw). Using the above-desribed data, the following parameters were calculated:

$${\text{Immersed absorption }}\left( {{\text{A}}_{{\text{i}}} } \right) = \frac{{{\text{M}}_{{{\text{SSD}}}} \, - {\text{ M}}_{{\text{d}}} }}{{{\text{M}}_{{\text{d}}} }} \times 100\%$$
(1)
$${\text{Boiled absorption }}\left( {{\text{A}}_{{\text{b}}} } \right) = \frac{{{\text{M}}_{{{\text{SSDb}}}} - {\text{M}}_{{\text{d}}} }}{{{\text{M}}_{{\text{d}}} }} \times 100\%$$
(2)
$${\text{AVPV}} = \frac{{{\text{M}}_{{{\text{SSDb}}}} - {\text{M}}_{{\text{d}}} }}{{{\text{M}}_{{{\text{SSDb}}}} - {\text{M}}_{{\text{w}}} }} \times 100\%$$
(3)

2.4.4 Resistance to Cyclic Wetting and Drying

In practical applications, pavements are often exposed to drying due to sunlight and wetting by rain. Therefore, the ability of concrete to maintain its performance under such field weathering is crucial to pavement’s durability and performance [45].

The resistance of concrete specimens exposed to repeated wetting and drying was evaluated by the cyclic wetting and drying test.The experimental process was inspired by ASTM D559/D559-M-15 [46], the standard test method for wetting and drying of compacted soil–cement mixtures. The GPC samples were subjected to 12 wetting and drying cycles as shown in Fig. 6. At the end of the last exposure (i.e., 12th cycle), the specimens were oven dried at 110 °C for 24 h, and their compressive strengths were tested after the samples cooled down to room temperature. In addition, the weight changes of the concrete cylinders at each stage were recorded.

Fig. 6
figure 6

Wetting and drying cycles

Full size image

2.4.5 Abrasion

To evaluate the wear resistance of concrete pavement under constant exposure to traffic, the abrasion resistance test was carried out on both the top and bottom surfaces of the slabs following BS EN 13,892–4:2002 [47] standards. The test procedure is shown in Fig. 7. Concrete slabs of size 500 mm × 500 mm × 100 mm were cast and tested after 28 days of ageing. Eight points were marked on the concrete surface using a steel plate (Fig. 7a) as a reference for positioning the depth gauge on the surface. The mean depth of wear within the ring pattern was used to assess the wear resistance.

Fig. 7
figure 7

Stages in abrasion testing. a Marking reference points on the slab surface using a steel plate. b Taking the initial depth reading. c Fixing abrasion head to the slab surface and connecting the electric controller. d Taking the depth measurements after testing

Full size image

The abrasion head consisting of three steel wheels at equal radial intervals was connected to the slab surface using four iron screws to keep the setup stable when rotating the wheels. The wheels were rotated at a speed of 180 rpm for 2850 revolutions using an electric motor and controller thus outlining a circular path with a diameter of 225 mm ± 15 mm. The abrasion head was then detached from the concrete slab and the depth change along the wheel path was recorded relative to the initially marked eight reference points (Fig. 7d).

3 Results and Discussion

3.1 Fresh Properties

All fresh GPC mixes were dense and sticky and showed a unique reddish-brown colour (Fig. 8) which was due to the presence of WCB powder in the binder. The high viscosity of the fresh concrete was caused by the Na2SiO3 activator [48]. Although the binder contained only solid activators (i.e., one-part geopolymer), the addition of water and subsequent mixing process facilitated the dissolution of the activators, resulting in a viscous fresh concrete mix similar to two-part GPC.

Fig. 8
figure 8

Fresh GPC with reddish-brown colour

Full size image

3.1.1 Workability

The workability of fresh concrete has a significant influence on the degree of compaction which ultimately affects its strength. In addition, fresh concrete must support efficient transport, placement, and finishing [37]. Even though there are no direct tests available to quantify workability, the slump test is often used to indicate the workability of fresh concrete in pavement construction [37]. The required slump of pavement concrete varies based on the paving technique. According to the Australian pavement design guidelines, concrete with a lower slump (i.e., 20 mm to 40 mm) is used in slip-form construction, since side forms are not used during the paving process. For mechanised and manually operated fixed-form paving, the recommended slump values are in the ranges of 35 mm to 50 mm and 55 mm to 65 mm, respectively [37, 38]. Therefore, in this study, a target slump of 50 mm to 65 mm was selected.

Figure 9 shows the slump test values recorded for all GPC mixes with their corresponding w/b ratios. The amount of water required to achieve a slump within the range of 50 mm to 65 mm was reduced with the increase of fly ash content in the binder. The spherical morphology and low water absorption of fly ash particles were capable of facilitating a homogeneous and workable mixture even at low water contents [15, 49]. All mixes resulted in slump values near the higher end of the required range (i.e., 55 mm to 65 mm). To obtain a lower slump with these GPC mixes, the water content can be slightly reduced. An accompanying improvement in mechanical properties can also be expected.

Fig. 9
figure 9

Slump values and corresponding w/b ratios of GPC mixes

Full size image

3.1.2 Setting Time

The Vicat needle penetration is the most commonly used method to estimate the setting time of OPC and geopolymer binders. The normal consistency pastes used in this method have low water content and are very stiff [50]. On the other hand, concrete has higher water content and aggregates, and exhibits substantially different setting behaviour when compared to the paste samples. Therefore, the setting time studies for paste and GPC samples were carried out separately using the Vicat method and penetration resistance method, respectively.

Figure 10 shows the setting time estimated using the penetration of the Vicat needle for four binder samples used in this study. Here, pastes with normal consistency were used and the w/b ratio was in the range of 0.37 to 0.39. Similar to the w/b ratio in GPC, the amount of water required to achieve normal consistency decreased with the increase of fly ash in the binder. According to AS/NZS 2350.4:2006 standard test setup [42], all binders reached the initial setting time within one hour or less. FA0-S60 binder required the highest water content to achieve normal consistency and took more time to reach the initial set. Water content has a direct influence on the setting time, where higher water contents tend to retard the setting of paste, mortar, or concrete [50,51,52]. FA20-S60 and FA30-S80 binders reached the final setting in respectively 380 min (approximately 6 h) and 410 min (approximately 7 h), and were found to be the fastest setting pastes. Using the Vicat test method, all the binders were estimated to reach the final set in around 8.5 h or less.

Fig. 10
figure 10

Setting time of geopolymer paste with normal consistency according to the Vicat needle penetration

Full size image

Concrete mixes developed using the above-described binders had w/b ratios ranging from 0.45 to 0.49, and they additionally contained aggregates. Therefore, the setting time estimated for normal-consistency binder pastes does not give an accurate estimation of the setting time of the corresponding concrete mix. Furthermore, the Vicat test method does not investigate the behaviour of the mix before reaching the initial setting and does not indicate the hardening rate over time [52, 53].

Figure 11 shows the penetration resistance of the GPC mixes and their initial and final times of set. The setting time obtained for GPC using the penetration resistance was significantly higher than the setting time of paste samples estimated following the Vicat test method. For example, the FA30-S80 binder with normal consistency achieved the initial and final setting time in 40 min and 410 min, respectively. GPC developed with the same binder (i.e., FA30 GPC) reached the initial setting in 290 min and the final setting in 610 min. Similar to the FA0-S60 paste sample, FA0 GPC made with the same binder recorded the highest time to reach the final set. This can be attributed to the higher w/b ratio in both paste and concrete.

Fig. 11
figure 11

Setting time of GPC according to the penetration resistance method

Full size image

Although the Vicat test method is the most commonly followed testing procedure to determine the setting time, the experimental data in this study show that the setting time of a given binder obtained by means of the Vicat method can not give an accurate estimate of the setting time of the resulting concrete made with the same binder (Table 4). The penetration resistance method gave a more reliable estimate of the setting time of fresh GPC while also giving information on the rate of hardening. All the GPC mixes except for mix FA0 achieved initial setting within 4 h to 5 h and final setting time within 10 h to 10.5 h. WCB-based GPC developed in this study showed a high rate of hardening and penetration resistance of 28 MPa in approximately 15 h or less. This hardening behaviour is tremendously advantageous in pavement construction since high early-strength gain is a crucial consideration in pavement concrete.

Table 4 Comparison of initial and final setting times estimated using the Vicat needle penetration and penetration resistance methods
Full size table

3.2 Compressive Strength

Figure 12 shows the 28-day compressive strength of GPC mixes produced with the four types of binders given in Table 1. These samples were demoulded 24 h after preparation and they were stored at ambient temperature (25 °C ± 2 °C) while in contact with air until the test date. The aim of this unsealed ambient curing was to assess the strength development of WCB-based GPC under basic field conditions without any special provisions. Except for the GPC mix FA20, all other samples achieved strengths of more than 32 MPa, which is the minimum design strength requirement for pavement base concrete. However, allowing a safety factor for variations in field conditions and material properties, the target 28-day strength for the trial mixes was actually 40 MPa.

Fig. 12
figure 12

28-day compressive strength of GPC trial 1. Error bars indicate mean ± one standard deviation (SD)

Full size image

Subject to the unsealed production conditions, only the FA30 mix was able to achieve strengths of more than 40 MPa, recording about 41 MPa. This can be attributed to three main reasons. Firstly, the FA30 mix had the lowest w/b ratio of 0.45. Although the FA20 mix has the same water content, the amount of slag present in the FA30 binder was higher than in the FA20 binder. Higher slag facilitates the formation of more Calcium Silicate Hydrates (CSH) in the structure, thereby improving its compressive strength [24, 54, 55]. Finally, the FA30 mix consisted of higher fly ash percentage compared to the other three mixes, which reduced the shrinkage [56]. When all these three factors, namely, the lower water content, higher CSH formation, and lower shrinkage, are combined, a higher strength can be expected.

The results from this first trial suggest the need for modifications to the GPC mix design or production conditions to achieve the target strength for all four mixes. Therefore, in this work, while keeping the mix designs the same, two modifications were applied to the production conditions, namely;

  1. 1.

    Change of WCB particle size

  2. 2.

    Change of curing conditions.

3.2.1 Change of WCB Particle Size

For the initial trial given above, WCB powder of particle size of 150 µm passing with a mean particle size (d50) of 45.8 µm was used. As the first modification, WCB particle size was reduced using an industrial ball mill. This mechanical grinding process reduced the particle size down to d50 of 13.4 µm. Figure 13 shows the particle size distribution corresponding to GPC trials 1 and 2. From GPC trial 2 onwards, WCB with reduced particle size was used for binder preparation.

Fig. 13
figure 13

Particle size distribution of WCB powder used for GPC trial 1 and 2

Full size image

To evaluate the effect of the particle size change, all GPC mixes were repeated by using WCB with reduced particle size and subjected to unsealed ambient curing similar to trial 1. Figure 14 shows the comparison of compressive strengths between the two GPC trials. FA10 and FA20 mixes showed a significant increase in compressive strength after reducing the WCB particle size. Compared to trial 1, FA10 and FA20 mixes showed a 17% and 31% increase in compressive strength, respectively. With 95% confidence levels, FA10 and FA20 mixes were calculated to have p-values of 0.0003 and 0.038, respectively, validating the significant improvement in the strength (under Welch’s t-test, unequal variances). However, the strength change due to the reduction of particle size was insignificant for GPC mixes FA0 (p-value = 0.0849) and FA30 (p-value = 0.0551).

Fig. 14
figure 14

Comparison of change in 28-day compressive strength of GPC after changing the WCB particle size. (i.e., trial 1 vs trial 2). Error bars indicate mean ± one SD

Full size image

Generally, finer particles provide more surface area for reaction and are likely to react completely, making the structure stronger [22, 25, 57]. However, geopolymer formation reaction can be adversely affected when the samples are in contact with air (i.e., unsealed curing). This can be attributed to atmospheric carbonation, which hinders or pauses the geopolymer formation. [58,59,60]. Hence, the reaction process is unable to take complete advantage of the reduced particle size. Therefore, a significant improvement in compressive strengths to exceed the target value could not be obtained.

3.2.2 Change of Curing Conditions

In trial 3, GPC samples were prepared using WCB with reduced particle size, and additionally, the samples were subjected to sealed ambient curing where the demoulded samples were placed inside air-tight polythene bags until the test date. Figure 15 shows the comparison of the 28-day compressive strength of GPC mixes subjected to unsealed and sealed curing. Compared to the unsealed samples, significant strength increase was observed for all the sealed samples (FA0 p-value = 0.008, FA10 p-value = 0.011, FA20 p-value = 0.006, FA30 p-value = 0.006) and all the mixes exceeded the target strength value of 40 MPa. The highest strength of 56 MPa was recorded for FA20 GPC. The lowest strength under sealed ambient curing was recorded as 42 MPa for FA0 GPC which was still higher than the target strength.

Fig. 15
figure 15

Compressive strength of GPC subjected to sealed and unsealed ambient curing. Error bars indicate mean ± one SD

Full size image

According to the study by Lee et al. [60], in fly ash-based geopolymers, sealed samples showed a strength of around 120 MPa to 135 MPa while the unsealed samples from the same mix showed only 60 MPa to 90 MPa. These samples were subjected to elevated curing for 24 h at 70 °C followed by sealing up to the test date. Similarly, Mucsi et al. [51] stated that sealed curing has a direct influence on developing better structural characteristics.

Geopolymer formation occurs in three stages, namely, dissolution, gelation, and polycondensation [32, 33, 61]. Although the final reaction product does not contain water, to facilitate each of these stages, the presence of water within the environment is crucial [60]. Sealing prevents loss of water and atmospheric carbonation, thereby protecting the sample from developing micro-cracks due to dehydration. Zhang et al. [59] verified that sealed curing can result in a denser microstructure with lower crack widths. Sealing also facilitates the reorganisation of the polycondensation process, leading to a better geopolymer formation [60]. Although OPC-based concrete has an established and standardised production process, the geopolymer production still follows locally adapted mixing and curing procedures based on the users’ preferences. Therefore, similar to water curing in OPC concrete, sealed curing can be recommended for GPC.

Sealed curing was further studied by optimising the duration of the sealing period. For this purpose, FA20 mix was selected and the samples were sealed for 3 days and 7 days, and then kept in contact with air until the test date. Figure 16 shows the 28-day compressive strength of the FA20 GPC samples subjected to different sealing periods.

Fig. 16
figure 16

28-day compressive strength of FA20 GPC subjected to different sealing periods. Error bars indicate mean ± one SD

Full size image

Sealing the samples only for the first 3 days still showed a significant improvement in compressive strength compared to the unsealed curing. Further, a corresponding increase in compressive strength was observed with the increase in the seal duration. The mix was able to satisfy the target strength even subject to 3 day sealing period, achieving 28-day compressive strength of 42 MPa. The highest reaction rate of WCB-based geopolymers was observed in the first 15 h and changes in the rates could be observed when varying the curing conditions such as temperature [62, 63]. In addition, sealing in the initial period can facilitate the formation of CSH gels and thus, increase the early strength. In in-situ applications like pavements, it is practical to cover the concrete surface using a polythene sheet or a similar alternative for 3 to 4 days. Longer sealing periods can be beneficial in pre-cast applications or instances which require higher strengths.

3.3 Flexural Strength

According to the Australian pavement design guidelines, the flexural strength of concrete intended to use in rigid pavement base constructions should be in the range of 4 MPa to 5 MPa [37, 38, 63]. Further, concrete with a minimum 28-day flexural strength of 4.5 MPa is recommended for moderate to heavily trafficked roads [37, 63]. As defined in the Australian pavement design guidelines [63], an estimation of the 28-day flexural strength (fcf) in MPa can be obtained in terms of the 28-day compressive strength (fc) in MPa by using the empirical equation given as

$${f}_{cf}=0.75 \times \sqrt{{f}_{c}}$$
(4)

Comparison of flexural strength values of GPC samples obtained experimentally and calculated with Eq. (4) are shown in Fig. 17. The experimental values were obtained by subjecting specimens to sealed ambient curing for about 28 days. The results show that WCB-based GPC can achieve satisfactory flexural strength in 28 days and thus satisfy the basic pavement strength requirements. Compared to the actual value of the flexural strength, the design equation generally gave a lower value.

Fig. 17
figure 17

Flexural strength of WCB-based GPC: Experimental Vs Estimated

Full size image

3.4 Strength Development

From all four GPC mixes developed in this study, the FA20 mix gave the highest 28-day compressive strength of 56 MPa. Therefore, that mix was selected to further investigate strength development characteristics. Here, all the samples were sealed for 3 days and then kept in contact with air until the test date. The suggested method can give better insights into the strength development under proposed in-situ curing.

As shown in Fig. 18a, the FA20 mix had a 3-day compressive strength of 30 MPa, which corresponds to around 71.5% of its 28-day strength. Concrete with high early-strength gain can be considered more suitable for pavement applications. For instance, according to the AASHTO guidelines [64], pavements can be open to traffic only after achieving strength of at least 15 MPa. FA20 GPC in this work was able to gain more than twice the required strength in 3 days. At the end of 7 days, the samples showed compressive strength of 39 MPa.

Fig. 18
figure 18

a Compressive strength and b flexural strength development of FA20 GPC. Error bars indicate mean ± one SD

Full size image

Similarly, within 72 h, WCB-based GPC showed flexural strengths of more than 5 MPa. There was no notable increase in the flexural strength observed for GPC after 3 days. However, the strengths were always larger than the guideline’s requirements. Further, a significant change in flexural strength was not observed after reducing the sealing period. Samples sealed for both 28 day and 3 day periods showed an average flexural strength of 5.1 MPa at 28 days (Figs. 17 and 18b). In addition, FA20 GPC showed an average split tensile strength of 2.6 MPa (SD =  ± 0.44) and 3.4 MPa (SD =  ± 0.59) at 7 and 28 days, respectively.

3.5 Durability Properties

3.5.1 Water Absorption

Water absorption of concrete specimens subjected to sealed ambient curing was determined by immersed and boiled absorption methods. All samples showed a slight increase in water absorption after boiling (Fig. 19). This can be attributed to the removal of air from the specimen due to the boiling, thereby maximising the water flow into the pores inside [44]. All the samples showed water absorption of less than 7.2% and higher than 5%, which agrees well with the study by Negahban et al. [65] according to which the water absorption of OPC can vary in the range of 6.17% to 6.65% while fly ash and slag blended GPC has water absorption in the range of 5.90% to 6.64%. The addition of finer particles to the binder can further reduce water absorption by reducing the pore volume in the concrete matrix [66].

Fig. 19
figure 19

Water absorption of WCB-based one-part GPC after 28 days curing period. Error bars indicate mean ± one SD

Full size image

Generally, concrete with absorption of less than 5% is considered impermeable [38]. Therefore, GPC developed in this study can be categorised as permeable. GPC group FA0 showed the highest water absorption (i.e., 7.2% boiled absorption), and FA20 showed the lowest water absorption (i.e., 6.1% boiled absorption). It can be attributed to the presence of a more porous structure in FA0 compared to FA20 GPC mix. These results validate the obtained compressive strengths, where the FA0 group had the lowest compressive strengths while the FA20 group had the highest.

3.5.2 AVPV

Figure 20 shows the AVPV values obtained for WCB GPC after 28 days of curing. Similar to the water absorption, the lowest AVPV was shown for the FA20 sample which had the highest 28-day compressive strength. These results further verify the less porous and compact structure of the FA20 mix. Further, all GPC mixes showed AVPV of less than 14%, which is the maximum allowable void content for 40 MPa grade concrete [67]. Compared to the AVPV figures reported for OPC concrete and GPC mixes investigated by Nighaban et al.[65], WCB-based geopolymer developed in the current study showed lower AVPV percentages. The current work and the study by Nighaban et al. [65] have used fly ash and slag from the same supplier suggesting that the reduction of permeable voids is mainly due to the addition of finer WCB to the binder.

Fig. 20
figure 20

AVPV of WCB-based one-part GPC after 28 days curing period. Error bars indicate mean ± one SD

Full size image

3.5.3 Cyclic Wetting and Drying

Figure 21a shows the percentage mass change in FA20 GPC specimens exposed to alternate wetting and drying cycles. The positive values indicate the mass gain (i.e., water absorption) by test cylinders at the end of the immersion period. Negative values indicate the mass loss due to oven drying.

Fig. 21
figure 21

a Mass loss of samples exposed to cyclic wetting and drying. b Comparison of compressive strength of FA20 GPC exposed to wetting and drying cycles and reference samples cured at ambient temperature. Error bars indicate mean ± one SD

Full size image

Water absorption during the initial cycles was smaller compared to the later cycles. However, the overall change in mass was not significant, indicating a better geopolymerisation and stable structure. Further, as shown in Fig. 22, no deterioration in the samples was visually observed after exposure to continuous wetting and drying cycles.

Fig. 22
figure 22

FA20 GPC a appearance after exposure to wetting and drying cycles, b reference sample

Full size image

Figure 21b shows the comparison of compressive strength of GPC samples subjected to wetting and drying cycles, and reference samples stored under ambient curing (i.e., 25 °C ± 2 °C, relative humidity 50% ± 2%). A slight reduction in compressive strength (i.e., 3%) was observed for the samples exposed to weathering. However, this reduction was found to be insignificant (p = 0.654). Therefore, more studies investigating the resistance of WCB-based GPC under adverse exposure conditions are suggested.

Based on the previous studies, it is clear that GPC has generally good resistance to wetting and drying [68]. Ganesan et al. [69] investigated the wetting and drying response of fly ash-based GPC by exposing them to 90 weathering cycles. Mass reduction at the end of the exposure period was only 1.005% while OPC concrete exposed to the same conditions showed a mass reduction of 5.45%. Further, compressive strength loss due to exposure to continuous wetting and drying was 14.7% for fly ash-based GPC and 21.62% for OPC concrete [69]. Similarly, Srinivas et al. [70] reported a 9.9% loss of compressive strength for GPC with fly ash and slag after exposure to 28 cycles. OPC under the same exposure conditions showed 10.89% reduction in compressive strength.

3.5.4 Abrasion

Figure 23 shows the FA20 GPC slab top surface before and after the abrasion test. Aggregates were not detached from the surface indicating strong binding between the aggregates and the geopolymer binders. Abrasion reported for slabs subjected to wearing by three steel wheels for about 2850 revolutions is shown in Fig. 24. The average abrasion of the top surface was 1.08 mm and the bottom surface was 0.92 mm.

Fig. 23
figure 23

Top surface of GPC slab a before and b after the abrasion test

Full size image
Fig. 24
figure 24

Wearing depth of GPC slabs due to abrasion. Dash lines indicate the average abrasion for each surface

Full size image

The abrasion resistance of concrete can be influenced by compressive strength, binder type, binder content, aggregate properties, water content, surface finish, and porosity [71]. Abrasion on the bottom surface is generally lesser than on the top surface due to the better compaction in the lower concrete layer. Studies on abrasion resistance of GPC are not often reported in the literature. Therefore, more detailed investigations of the abrasion resistance of GPC are suggested.

4 Conclusions

The use of geopolymer is considered a promising strategy to reduce the carbon footprint in pavement construction. Although there are several studies available on the development of geopolymer binders and mortar using different sources of wastes and industrial or agricultural by-products, there are very limited studies available on developing geopolymer concrete (GPC) using these binders, except with fly ash and slag. The present study investigated the use of binders with waste clay bricks (WCB) as an aluminosilicate source material and solid Na2SiO3 as the sole activator in developing pavement-grade concrete.

From a comprehensive series of experiments on production conditions and optimisation processes, the following conclusions were made:

  • WCB-based GPC was able to achieve the maximum workability requirements for pavement applications with a slump in the range of 50 mm to 65 mm. The addition of fly ash to the binder reduced the water content required to achieve the necessary workability.

  • Estimation of setting time based on the penetration resistance method was found to be more appropriate for GPC than the commonly used Vicat test method. The initial setting of all GPC mixes occurred within 4 to 5 h. All the mixes were able to achieve the final set with a hardening rate of more than 28 MPa in around 10 to 15 h.

  • Sealed ambient curing can be proposed as the best curing method to achieve higher compressive and flexural strengths. The sealing period had a significant influence on strength development, where longer sealing durations were associated with better compressive strengths.

  • Under the sealed curing conditions, GPC made with binder consisting of 20% fly ash in the precursor and, 40% WCB and 60% slag in the remaining precursor (i.e., FA20 GPC with FA20-S60 binder) recorded the highest 28-day compressive strength of 56 MPa. GPC made by using the binder with 30% fly ash in the precursor and, 20% WCB and 80% slag in the remaining precursor (i.e., FA30 GPC with FA30-S80 binder) gave the highest 28-day flexural strength of 6.2 MPa.

  • Water absorption and apparent volume of permeable voids (AVPV) of WCB-based GPC were in a closer range to that of OPC. FA20 mix recorded the lowest water absorption of 5.6% in immersed condition and 6.1% in boiled condition. The same mix gave the lowest AVPV value of around 9.9%.

  • GPC showed better resistance to cyclic wetting and drying reporting around only 3% of strength reduction after weathering. Wearing of the GPC surface due to abrasion was around 1 mm showing better resistance to abrasion.

Therefore, WCB-based GPC mix designs developed in this study can be considered suitable for pavement applications based on the Australian rigid pavement design guidelines and mix design selection criteria. Comprehensive studies investigating durability and field performance are recommended for future studies.