1 Introduction
Tunnel lining construction may take place in challenging environmental conditions, including high-geothermal areas characterized by temperatures exceeding 70 °C in extreme cases [
1,
2]. Although the influence of high temperature on the stability and durability of structures has been investigated [
3,
4], the health and safety of workers and related labor productivity have been relatively ignored [
5]. Shotcrete technology has emerged as a popular choice, for constructing tunnel linings due to its ability to rapidly set, to provide support to the surrounding rock, and to offer high flexibility without the need for formwork, particularly in inaccessible locations [
6,
7]. However, the rebound of shotcrete remains an issue, and the quality of shotcrete greatly depends on the nozzle angle and distance to the substrate, highlighting the necessity of automated applications [
8].
Replacing shotcrete technology with extrusion-based 3-dimensional concrete printing (3DCP) technology in high-geothermal environments is expected to reduce labor. Among these newly developed technologies with a high degree of automation, 3DCP is more suitable for large-scale elements than others such as particle bed methods [
9,
10]. Furthermore, extrusion-based 3DCP offers advantages over automated spraying, as it ensures no rebound and provides high surface quality. During the extrusion process, fresh cementitious materials are deposited in a layer-by-layer manner. The extrusion-based printing technology was primarily developed in the Contour crafting project with the aim of building structures in outer space, where the environment is even more hostile than the high-geothermal situation [
11,
12].
More recently, extrusion-based 3DCP has been further applied in hundreds of research projects and industrial applications [
13,
14]. The authors have conducted several trials of 3D printing tunnel linings, as depicted in Fig.1 [
15–
17], using Portland cement-based mixtures with cellulose ethers and re-dispersible polymer powders, to achieve robust adhesion between printed layers and the excavated rock wall. Other influencing factors such as the surface inclination and the substrate properties (roughness level and moisture content) have also been studied [
17].
As a novel technology, 3D printing with concrete still faces several challenges, including the need for higher thixotropy, often achieved by increasing binder content, to meet diverse pumping and deposition requirements. In general, the binder content of printable mixtures is more than twice that of conventional mold-cast concrete [
18]. In addition, the production of Portland cement is a major contributor to global CO
2 emissions, accounting for approximately 8%. This has raised significant concerns due to the reduced sustainability [
19–
21].
To promote the sustainability of 3D printable materials, reducing the binder quantity and augmenting the aggregate content is a potential strategy. However, the usage of a peristaltic pump for continuous extrusion in 3D printing poses a challenge due to the limited capacity of the associated rotational worm chamber. This limitation may result in blockages when increasing the aggregate content. An alternative approach is to incorporate waste by-products, including ground granulated blast furnace slag (GGBFS), fly ash, and silica fume, as partial substitutes for Portland cement. These by-products exhibit high reactivity, cost-effectiveness, and widespread availability, making them advantageous [
22,
23]. On the other hand, using alkali-activated material (AAM) could be an option, with less carbon emission due to the complete replacement of Portland cement with silica and alumina rich waste materials.
Within the realm of 3DCP, extensive research has been conducted on AAMs due to their noteworthy sustainability benefits when compared to Portland cement-based mixtures. In summary, the current studies on 3D printable AAMs are mainly restricted to the pumping, extrusion, and deposition at early ages for general 3D concrete printing [
24,
25]. By using a trial-and-error approach, different kinds of precursors and activators have been combined and tested to formulate mixtures that are suitable for printing [
26,
27]. In addition to that, one study focused on examining the capability of alkali-activated slag mixtures to retain their shape, specifically by incorporating nano clay and nucleation seeds [
28]. Apart from the chemical intervention, a small microwave heating zone right at the position of the printing head allows the alkali-activated material to gain its properties in a short period due to the accelerated polycondensation reaction [
29].
The investigation of alkali-activated materials, particularly in high-temperature environments, with the aim of improving the adhesion performance of 3D-printed tunnel linings in fresh and hardened stages remains unexplored. In this study, three different alkali-activated slag mixtures with different Ms ratios (SiO2/Na2O) of 0, 1, and 2 were formulated. Two different temperatures, 20 and 40 °C, were considered, representing normal and high-temperature environments. The adhesion of the alkali-activated slag mixtures was assessed in both the fresh and hardened states through a tack test and a pull-off test, respectively. Additionally, the structural build-up rate of the mixtures at varying temperatures was evaluated using a small-amplitude oscillation shear (SAOS) test. To further understand the mechanical behavior, flexural strength, and compressive strength were measured using prismatic and cubic specimens, respectively, and these were cured at different temperatures for 7 d. Furthermore, microstructure characterization was performed, including the quantification of pore structure using mercury intrusion porosimetry (MIP).
2 Experimental program
2.1 Materials and mixture proportions
The precursor utilized in the study was GGBFS, which was sourced from ORCEM, a company based in the Netherlands. The chemical composition of the GGBFS is provided in Tab.1. The mean particle size (
) of the GGBFS is approximately 9 μm. Additional information regarding the X-ray diffraction pattern, particle size distribution, and morphology of the GGBFS particles can be found in the authors’ previous work [
30]. Three alkaline solutions with Ms ratios of 0, 1, and 2 were used in this study. Pure sodium hydroxide solution was prepared by adding caustic soda with a purity of 97% to deionized water. The increase in the Ms ratio was achieved by combining liquid sodium silicate solution (SiO
2 27.5%, Na
2O 8.25%, water 64.25%) with the sodium hydroxide solution.
Silica sand, with particle sizes ranging from 0 to 2 mm, was employed in the study. The specific particle size distribution details can be found in Ref. [
17]. To enhance the shape stability of the materials utilized in 3D printing, attapulgite nano clay obtained from Acti-gel, a company based in the Netherlands, was incorporated into all the mixtures.
Three mixtures were formulated in this study, as shown in Tab.2. The printability of the mixtures was evaluated in the authors’ previous work [
31]. All mixtures maintained a water-to-solid binder (
w/
sb) ratio of 0.37 and a sand-to-solid binder (
s/
sb) ratio of 1. The Na
2O content in the mixtures was 5% based on the weight of the precursors. Additionally, the SiO
2 content varied for the different mixtures: 0% for Ms0, 5% for Ms1, and 10% for Ms2, calculated based on the weight of the precursors. It is important to note that the combined weight of GGBFS, nano clay, and the dry portion of the activator solutions constituted the solid binder (
sb). For all mixtures, a consistent dosage of 3% nano clay was used, determined by the weight of the precursors.
Paste samples were used for the SAOS tests and scanning electron microscopy, while mortar samples with the addition of silica sand were used for the other tests. To examine how different temperatures affect the adhesion performance of fresh alkali-activated slag mixtures, the raw materials, including sand, GGBFS, nano clay, and alkali solutions were placed at different temperatures (20 and 40 °C) for 24 h before the mixing process. All mixtures were prepared using a Hobart planetary mixer and the same mixing procedure was adopted as in the previous study [
32].
2.2 Testing procedures
2.2.1 Tack tests
Typically, the adhesion performance of cementitious materials is assessed once they have hardened, while the adhesion at the fresh stage is rarely evaluated [
33]. The tack test offers a valuable method for evaluating the adhesive performance of 3D printable materials, as has already been proven in previous studies by the authors [
15,
16]. For the tack test in this study, an Anton Paar rheometer (MCR 102) was utilized, as schematically shown in Fig.2. The diameters of circular parallel plates for the tack test were 50 mm. To simulate the rough excavated rock wall, circular sandpapers with a consistent diameter of 50 mm and a root mean square deviation of 0.18 mm were affixed to both plates [
17].
With a brief mixing and testing period (at 20 °C) lasting only a few minutes; it was assumed that the decrease in temperature did not significantly affect the results of the tack test. The tack test involved pouring the fresh material onto a plastic cylindrical mold with a 50 mm inner diameter and a 20 mm height, placed on a bottom plate. Subsequently, the mold was removed, and the top plate was lowered with a constant speed of 50 μm/s to compress the fresh sample until the distance between the plates reached 10 mm. As the samples were cast with the same size and were squeezed with the same speed, it can be assumed that the initial state of the samples remained the same for all series before testing. Following an 8-min waiting period after the binder and activator came into contact, the tack test was conducted in a linearly increasing force mode. The loading rate was maintained at a constant value of 0.04 N/s, determined from the stepwise increasing force mode in the 3D printing process [
17]. The evolution of tensile force and separation displacement over time was recorded during the tack test. Two tests were conducted for each mixture, using freshly prepared mixtures.
2.2.2 Flow table tests
Flow diameters were measured using flow table tests, following the standard EN 1015-3 [
34]. The flow table’s supporting plate was centrally positioned, and the mold was placed atop it. Two layers of fresh material were carefully introduced into the mold, after which the flow table underwent 15 drops to ensure even dispersion of the material. This step was performed 8 min post-binder/activator contact, with no consideration given to any temperature reduction effects for the measurement of flow diameter. The material’s diameters were measured perpendicularly, with two testing processes for each mixture.
2.2.3 Small-amplitude oscillation shear tests
For the SAOS tests, the Anton Paar MCR 102 rheometer was utilized. After mixing, the paste samples were transferred into cup containers with a diameter of 28 mm and depth of 70 mm. Subsequently, a six-blade vane (blade height 16 mm, blade diameter 22 mm) was inserted into the fresh samples. A water bath was used to keep the temperature constant (either 20 or 40 °C) during the measurement period. Before conducting the oscillatory time sweep tests, oscillatory strain sweep tests were performed to identify the linear viscoelastic region (LVER). The strain was logarithmically increased from 0.001% to 20% at a constant frequency of 1 Hz. For the time sweep tests, a constant strain amplitude of 0.005% and a constant frequency of 1 Hz were selected, which aligns well with the values chosen by other researchers [
35–
37]. Pastes were pre-sheared at 200 s
−1 for 30 s and rested for another 30 s. After that, the oscillatory time sweep tests were carried out to record three parameters, including storage modulus (
), loss modulus (
), and loss factor, starting at the age of 8 min (i.e., the time after contact between binder and activator) [
38]. The measurements were repeated twice to ensure accuracy and reliability.
2.2.4 Mechanical tests
At the age of 7 d, flexural tests, compression tests, and pull-off tests were performed using mortar specimens. Fresh mortar mixtures were cast in prismatic molds (40 mm × 40 mm × 160 mm) and covered with foil to avoid water evaporation. Specimens were cured at different temperatures of 20 or 40 °C until the measurement.
The flexural strength was determined using the three-point bending test following the standard EN 12390-5. The test was repeated three times to ensure the accuracy and reliability of the results [
39]. The distance between two supporting rolls was 120 mm. In addition, the compressive strength was measured following the standard EN 12390-3 with three measurements for each series [
40].
The bond strength in the hardened stage was measured using pull-off tests, following the standard EN 1542 [
41]. Square sandblasted concrete blocks measuring 100 mm in length, 100 mm in width, and 50 mm in height were used as a replacement for the excavated rock wall for tunnel linings. The blocks were kept in environments with specific temperatures of either 20 or 40 °C for 24 h before casting. Because the macro surface textures of rough substrate cannot be filled by the printed material [
15], the fresh material was placed manually on the substrate concrete block and compacted using a trowel to ensure full contact with the substrate. Molds were used for casting and the thickness of the overlay was around 10 mm. Subsequently, the specimens were covered with plastic foil and subjected to curing at the specific temperatures as above.
To perform the pull-off test, cores measuring 25 mm in diameter were drilled and steel dollies were attached to the top surface of the core using epoxy resin, as schematically shown in Fig.3. An automatic pull-off machine (Proceq DY-2) was used and three tests were done for each series.
2.2.5 Shrinkage measurements
Drying shrinkage was evaluated using mold-cast specimens (40 mm × 40 mm × 160 mm), following the standard ASTM C596-01 [
42]. The specimens were cured within an environment maintained at specific temperatures as prescribed by the experimental conditions. At the age of 1 d, the specimens were removed from the molds, and their surfaces were left exposed to the air. Subsequently, demountable mechanical strain gauge (DEMEC) pins were affixed at a gauge length of 100 mm on two sides of the beams. The specimens were then maintained in a vertical position and subjected to constant temperatures of 20 or 40 °C throughout the entire measurement duration. The change in length was measured using a digital DEMEC instrument at different ages (1, 3, 7, 14, and 28 d, after demolding). Three specimens were tested for each series.
2.2.6 Mercury intrusion porosimetry
MIP was used to evaluate the pore structure of 7-d-old mortar samples. Thermo Scientific Pascal 140 and 440 series instruments were employed for MIP analysis. The samples were cured at 20 or 40 °C, and the hydration process was halted after 7 d using a solution exchange technique. Specifically, the samples were first placed in a solution of 2-propanol for 24 h, then in an oven (40 °C) for 24 h. After that, the samples were vacuum dried at 20 °C for 7 d to reduce the microstructural damage [
43]. The samples were then oven-dried for 1 d at 40 °C and vacuum-dried for 14 d at 20 °C before further analysis [
43]. The test for each mixture was performed twice to ensure the accuracy and reliability of the results.
3 Results and discussion
3.1 Adhesion performance in the fresh state
The tensile force and separation displacement evolution observed during the tack test indicated the adhesion performance of the printed material in its fresh state. The tensile force versus separation displacement curves at different temperatures of 20 or 40 °C are shown in Fig.4 (one representative curve for each series). During the observation, it was noted that upon reaching the peak tensile force at the failure point, the tensile force rapidly dropped to zero as fracture occurred either at the interface or within the bulk zone of the fresh material [
17,
32].
The results of the peak tensile force values are further listed in Tab.3. At 20 °C, the mixture Ms0 showed the peak tensile force value of 6.7 N, while the peak tensile forces were 5.3 and 6.3 N for the mixtures Ms1 and Ms2, respectively. Prior research has suggested that the shear resistance of the fresh material plays a crucial role in governing the adhesive performance [
17,
44].
In this study, the flow diameter of the fresh material measured in flow table tests was used to characterize the shear resistance, as shown in Fig.5. The peak tensile force and the flow diameter exhibited a clear inverse relationship. Specifically, the mixture Ms0, which obtained the lowest flow diameter, showed the highest peak tensile force, while the mixture Ms1 gained the highest flow diameter but the lowest peak tensile force.
At 40 °C, the top plate failed to reach the initial measuring position for the mixture Ms0 due to its high structural build-up rate. This was also reflected by the flow diameter of the mixture Ms0, where almost no deformation was observed (diameter 102 mm). Thus the tensile force evolution curve was not recorded. In contrast, the peak tensile force of the mixture Ms1 was higher than that of the mixture Ms2, which is opposite trend to that observed at 20 °C. This difference is attributed to the higher shear resistance arising from the increased structural build-up rate of the Ms1 mixture at higher temperatures. Further discussion on this point will be presented in the next section.
As shown in Fig.4, the fresh material displayed minimal deformation in the early stage as the applied tensile force increased. Subsequently, the fresh material underwent shear flow after reaching the yield point. The fracture propagated at either the interface or within the bulk material, resulting in a rapid upward separation displacement of the top plate. The critical separation displacement, defined as the separation displacement value when the tensile force reached the peak value [
15], is listed in Tab.3. A larger critical separation displacement can be observed for the mixture with a lower peak tensile force and lower shear resistance. For example, the mixture Ms1 showed the lowest peak tensile force but the highest critical separation displacement at 20 °C. The critical strain of cementitious materials reduces as age increases [
45]. The mixture Ms2, characterized by a higher Ms value, displayed a decrease in the critical strain linked with the peak tensile force. This finding suggests that the stiff material, featuring enhanced shear resistance, impeded inward flow toward the central point, leading to restricted separation displacement in the vertical direction during the tack test [
44,
46].
In addition, the dissipated energy, which was calculated based on the integration of the tensile force over the separation displacement, is tabulated in Tab.3. Generally, the dissipated energy was higher for the mixture with a higher critical separation displacement, which required more energy consumption for the inward flow of the material.
3.2 Structural build-up
Oscillatory strain sweep test results are shown in Fig.6(a) and Fig.6(b). The storage moduli showed plateau behavior until the strain amplitude reached 0.1%. Subsequently, the storage moduli demonstrated a tendency to decrease after surpassing a specific strain value, known as the critical strain [
47,
48]. The critical strain value for all mixtures was approximately 0.1%. To confine the material within the LVER and achieve precise characterization, a fixed strain amplitude of 0.005% was employed for time-sweep tests across all mixtures. This selection guaranteed that the material remained within the LVER during testing.
The evolution of the storage moduli of mixtures at different temperatures is shown in Fig.7(a) and Fig.7(b). In the case of the mixture Ms0 activated with pure sodium hydroxide solution, the higher pH of the solution facilitated the dissolution of silica and alumina, leading to the rapid formation of reaction products. This was evident from the higher initial rate of increase of the storage modulus [
49,
50]. However, the formation of a reaction ring around the GGBFS particles could impede the subsequent dissolution of anions into the solution. This phenomenon can help to explain the lower rate of structural build-up observed during the later stages of the process [
51]. By contrast, the mixtures Ms1 and Ms2 activated by sodium silicate showed a lower storage moduli in the early stage due to the absorption of silicate anions onto the GGBFS particle surface, resulting in a dispersed suspension [
52–
54]. Despite this, a significant number of oligomers were generated as a result of the dissolution of the source materials and the interactions between various ions [
55]. Over time, a notable and rapid rise in the storage moduli was observed for the mixtures Ms1 and Ms2, surpassing the value obtained for the Ms0 mixture. This phenomenon can be attributed to the increased availability of nucleation sites provided by silicates present in the activator solution, resulting in higher precipitation of reaction products. Additionally, condensation reactions occurred between the formed oligomers, further contributing to the observed increase in the storage moduli. As a result, the solidification process was accelerated with more formation of products [
56].
In addition, the structural build-up rate of the mixtures at 40 °C was remarkably higher than that at 20 °C. For example, the mixture Ms0 at 20 °C reached the measuring limitation of around 500 kPa after around 76 min, while the period for the mixture Ms0 at 40 °C was merely around 12 min. Similarly, a rapid early-age strength gain of alkali-activated slag concrete was reported when cured at a higher temperature [
57,
58]. This is due to the elevated temperature breaking the reaction ring covering GGBFS particles and promoting the dissolution of ions into the solution, thus resulting in improved structural build-up [
59]. Interestingly, it could be found that the mixture Ms1 seemed more sensitive to temperature in the first 100 s as compared to the other two mixtures. The storage moduli of Ms1 started to increase rapidly from the beginning when the temperature increased to 40 °C. As shown in Fig.7, the storage moduli of Ms1 reached around 300 kPa at 40 °C after 100 s, while the value was merely around 100 kPa at 20 °C. This indicated a simultaneous formation of products without the induction period. However, the fast formation of products also resulted in the insufficient dissolution of GGBFS particles and a lower reaction rate at the later stage at 40 °C. In practical situations, the high geothermal environment would be beneficial for increasing the structural build-up rate. Therefore, the promoted strength gain can be used to avoid punching failure caused by loose rock [
16,
60].
The evolution of the loss modulus of mixtures at various temperatures is shown in Fig.8(a) and Fig.8(b). The loss moduli reflects the irreversibly dissipative response of the material. Compared with the storage moduli shown in Fig.7, the values of the loss modulus were much lower, especially for the mixture Ms0 with a high storage moduli in the early beginning. This indicated the formation of a solid-like inner structure of the material immediately after mixing, due to rapid coagulation. Furthermore, the decrease in the loss modulus can be attributed to the enhanced solidification rate of the material [
54]. In this study, the mixtures tested at 40 °C showed a much faster decrease in loss modulus, which can be associated with the fast change in viscous dissipation.
Fig.9(a) and Fig.9(b) depict the temporal evolution of the loss factor, providing insight into the relative ratio of viscous to elastic characteristics exhibited by the mixtures at temperatures of 20 and 40 °C. Each series is represented by a single curve. A decrease in the loss factor from 1 to 0 signifies the transition from an ideal viscous state to an ideal solid state [
61]. The percolation time, defined as the point at which the loss factor reaches 0, denotes the duration required for the formation of a percolated elastic inner structure. It also serves as an indicator of the transition from a viscous state to an elastic state [
53,
57]. Notably, the mixture Ms0 exhibited a lower initial loss factor, suggesting a more rigid and elastic behavior. Similarly, the mixtures at 40 °C experienced a faster drop in the loss factor to 0 compared to those at 20 °C, indicating an accelerated transition to the elastic state.
3.3 Mechanical behavior
The mechanical strength of mixtures at different temperatures is presented in Fig.10(a)–Fig.10(c), respectively. At 20 °C, the mixture Ms1 obtained the highest mechanical performance, while the mixture Ms0 had the lowest strength. For example, as compared to the mixture Ms1 with a bond strength of 1.1 MPa, its counterpart mixture Ms0 had a bond strength value of merely 0.7 MPa. Similar results, that the increase in the ratio of silicate to hydroxide is beneficial for better mechanical performance, have been reported previously [
61]. The highest strength of the mixture Ms1 was attributed to the presence of silicate that reacted with calcium ions that dissolved from the GGBFS grains and formed calcium silicate hydrate (C-S-H) [
62]. It is important to note that the flexural and bond strengths of the mixture Ms2 were lower than those of the mixture Ms1. This difference can be primarily attributed to the higher shrinkage resulting from the increased silicate moduli in the Ms2 mixture (see drying shrinkage values in Subsection 3.4) [
63].
The compressive strength increased when the Ms value increased from 0 to 1 because of the enhancement of interaction between the introduced silicates (mainly monomer and dimmer silicon groups) and dissolved calcium and aluminum, forming more reaction products, further resulting in higher compressive strength [
64,
65]. However, with an increase in the Ms value in the activator solution from 1 to 2, the pH value in the solution decreased due to the introduction of excessive silicates, primarily in the form of trimer and octamer silicon groups. This reduction in pH hindered further dissolution and led to a lower reaction degree in the mixture.
Another interesting point is that the increased temperature promoted the strength gain for the mixtures with sodium silicate (especially Ms2). This is because the solubility of essential ions increases at a higher temperature, and therefore essential ingredients appear for the precipitation of reaction products [
66]. During the SAOS test, a corresponding pattern was identified where heightened temperatures expedited the process of structural formation, leading to increased production of products. In the context of tunnel lining printing, the robust binding properties of Ms1 and Ms2 mixtures in their hardened state ensure the creation of a structurally sound configuration.
3.4 Shrinkage behavior
The drying shrinkage strains are illustrated in Fig.11. Higher Ms values resulted in significantly higher drying shrinkage values, which is consistent with previous studies [
67]. For sodium silicate-activated mixtures, a denser microstructure formed (see the measured pore structure in Subsection 3.5). As a result, the tensile stress of capillary pores during drying was increased, leading to higher drying shrinkage values [
63]. Another plausible reason is that the formation of silica or silica-rich gel during the reaction led to high drying shrinkage values [
68].
In addition, drying shrinkage values of the mixtures with higher Ms values (such as the mixture Ms2) were lower at 40 °C than at 20 °C. As seen in Fig.11, the drying shrinkage of the mixture Ms2 reached 11426 με at 20 °C, while the value of the same mixture was 8310 με at 40 °C. Reduced shrinkage at 40 °C could be explained by the mechanism that C-S-H formed during high-temperature treatment has less water content [
57,
69]. For the 3D printing of tunnel linings in a high-geothermal environment, high temperature is beneficial in decreasing the drying shrinkage and in avoiding structure deterioration.
3.5 Pore structure
The pore structural characterizations of the mixtures at different temperatures, 20 and 40 °C, are shown in Fig.12(a) and Fig.12(b), respectively. The critical pore sizes. which correspond to the peaks of the curves, are further listed in Tab.4. It can be observed that the mixture with a higher Ms value showed lower critical pore size, at both 20 and 40 °C. As stated, a pure sodium hydroxide activator with a high pH value can promote the dissolution of GGBFS particles in the early stage, leading to the formation of non-uniformly dispersive hydration products, and thereby a loose pore structure [
70]. In the case of the alkali activator with a higher Ms value, the presence of sodium silicate allows for a greater concentration of silicate ions in the pore solution. This, in turn, facilitates the formation of a larger quantity of products during the precipitation process. As a result, a compacted pore structure can be obtained for the mixture with more content of sodium silicate [
69,
71]. Concerning the influence of temperature, however, the results showed a limited influence of raising the temperature on the critical pore size.
The cumulative pore volume of the mixtures at different temperatures, 20 and 40 °C, is shown in Fig.13(a) and Fig.13(b), respectively. Furthermore, the porosity was derived from the cumulative pore volume, as listed in Tab.4. With increase in the Ms value, a lower porosity was obtained. For example, at 20 °C, the porosity was 12.6% for the mixture Ms0, while the value was merely 5.9% for the mixture Ms2. As explained, the mixture incorporating more silicate would form more products and a denser pore structure, thus a lower total porosity.
Furthermore, the increase in temperature reduced the porosity for all mixtures with different Ms values. These findings align with previous studies that have demonstrated how alkali-activated materials tend to develop a denser matrix with reduced porosity and finer pore size as the curing temperature is elevated [
72,
73]. This phenomenon can be attributed to the fact that higher temperatures facilitate the dissolution of GGBFS particles and the crystallization of hydration products. Consequently, the pores within the matrix are better filled, resulting in an improved pore structure [
74–
76]. The improved mechanical properties observed in mixtures cured at higher temperatures are attributed to their lower porosity and finer pore size distribution. This reduction in porosity and refinement of pore sizes leads to enhanced interparticle bonding and increased material density, ultimately resulting in superior mechanical performance [
76].
In the context of tunnel lining construction, elevated temperatures are advantageous for both structural performance and durability. However, it is important to consider that excessive temperatures may lead to the loss of pore water, which can cause degradation of the pore structure and an increase in permeability [
70].
4 Conclusions
In the present study, an alkali-activated slag mixture is used for the 3D printing of tunnel linings in a high-geothermal environment. The effect of different temperatures, 20 and 40 °C, on the performance of the mixtures with different Ms values is investigated. In light of the experimental findings and the subsequent analysis, the following conclusions can be deduced.
1) The adhesive performance at 20 °C primarily depends on the shear resistance of the material, whereas at 40 °C, the structural build-up rate significantly increases, resulting in improved shear resistance and adhesive performance.
2) The mixtures activated by sodium silicate (Ms1 and Ms2) display a slower initial structural build-up rate compared to the sodium hydroxide-activated mixture, due to the latter’s higher pH value. Nevertheless, the storage modulus of sodium silicate-activated mixtures increase significantly with time. This can be explained by the greater number of nucleation sites available, resulting in more effective precipitation of reaction products.
3) The mixture Ms1 has the highest bond strength at 20 °C in the hardened state, while the mixture Ms2 with the highest Ms ratio obtains the highest bond performance at 40 °C, which can be attributed to the promotion of the reaction by the high temperature.
4) High temperature is beneficial in decreasing the drying shrinkage of alkali-activated mixtures with high Ms values, which could be due to the lower water content of formed C-S-H.
5) The mixtures activated by sodium silicate have finer pore structures as compared to those activated by sodium hydroxide. As the temperature is elevated, it results in the formation of a denser matrix characterized by lower porosity and finer pore size distribution.
PDF
(4659KB)
