Introduction

Engineered cementitious composite (ECC) is a special type of fibre reinforced cementitious composite, micromechanically designed to exhibit high tensile ductility and enhanced durability [1]. Despite its improved properties, ECC is susceptible to fire hazard due to its dense and compact microstructure. To address this issue, use of various of types of polymer, metallic or natural fibres have been assessed and low melting point polymer fibres such as polypropylene (PP), linear low-density polythene (LLDPE), and polyvinyl alcohol (PVA) have proven to be particularly efficient in preventing spalling during fire exposure [2]. These fibres melt and vaporize under high temperatures, creating an interconnected network of vacant channels. This network improves the dissipation of pore pressure, thereby enhancing resistance to spalling and overall mechanical performance.

However, it should be noted that the formation of the interconnected network and risk of spalling is greatly dependent on the specimen size. It is easier for the moisture to escape in a smaller sized specimen due to shorter path. Therefore, smaller sized specimens may have better spalling resistance and residual performance when compared to the larger sized specimens [3]. Additionally, this phenomenon might be intensified by the ‘size effect’ observed during uniaxial compression testing. It is well established that the likelihood of encountering a critical-sized crack increases with the number of cracks i.e., volume of the specimen [4]. Therefore, the combined influence of formation of vacant channels and critical-sized cracks may have a significant impact on the post-exposure compressive performance.

Existing studies on ECC utilizes a wide range of specimen sizes to assess the elevated temperature performance and spalling resistance, making it challenging to make justified comparisons among the existing data and predict the actual fire resistance. The magnitude of this variability is highlighted by examining the specimen sizes employed within the last five years, as listed in Table 1. It can be observed that not only there is a significant variation in the size of the specimens, the shape and aspect ratio of the specimens also differs which makes it hard to provide a valid comparison of the existing data. Therefore. it is important to define a coherent link between specimen size and its impact before establishing fire test guidelines for fibre reinforced cementitious composites (FRCC).

Table 1 Specimen size used in existing studies for residual compressive performance analysis
Full size table

In order to address this issue, various researchers have studied the effect of specimen size on the elevated temperature compressive performance of FRCC. Zheng et al. [28, 29] reported that the residual compressive strength tested on 70.7 mm cubic steel fibre reinforced specimens was higher than that tested on the prism (70.7 × 70.7 × 220 mm). For instance, the residual compressive strength at 400 °C retained by approximately 76% for cubic specimens with 3% (by volume) steel fibres, whereas the retention was just 16% for the prism specimens. Bamonte and Gambarova [30] analysed the size effect in PP-FRCC and also observed relatively higher strength in 40 mm cubic specimens as compared to 36 mm diameter × 110 mm height cylinders. In both studies, larger sized specimens suffered higher loss which may have been due to the delay in formation of interconnected network. On the other hand, Akca and Zihnioglu [31] found better compressive performance in 100 mm cubic specimens as compared to 100 × 100 × 50 mm specimens. 100 mm cubic specimens retained around 60% compressive strength at 600 °C, whereas the retention for the specimens with 50 mm height was 54%. They concluded that the specimens with smaller size could result in higher loss due to rapid heating. Contrary to these findings, Erdem [3] reported that the residual compressive strength was unaffected by the change in the size of specimens. Cylinder specimens with diameter 50 mm, 100 mm, 150 mm, and height to diameter ratio 2:1 showed similar crack pattern and the influence of elevated temperature on residual compressive strength was also similar with roughly equivalent retention observed until 800 °C. The conflicting findings from these studies highlight a lack of clear correlation between specimen size and performance under elevated temperatures. Consequently, a comprehensive analysis is required to better understand the effect of temperature on different sized specimens, enabling future studies to provide meaningful comparisons with existing data.

Therefore, the main objective of this paper is to extensively investigate how specimen size and shape influence the residual compressive strength of high strength ECC (HSECC). In the authors’ previous work, a binary blend of ground granulated blast furnace slag (GGBFS) and cement with GGBFS: cement = 1:1 with hybrid polyethylene (PE)-steel fibre was found to have a superior spalling resistance and residual compressive performance [6, 20]. Consequently, the same mix has been employed in the present study to analyse the residual compressive strength and provide a thorough explanation of the effect of temperature on different sized specimens. Microstructure characterization techniques including scanning electron microscopy (SEM), and X-ray diffraction (XRD) are further utilized to analyse the influence of size at different temperature range. Following a careful and thorough comparison, the variation in the performance of different sized specimen is correlated and a clear effect of specimen size and shape on residual strength has been presented.

Experimental program

Raw materials

ECC mix with cement to GGBFS ratio of 1:1 was used in this study. Details of the mix design are given in Table 2. The main binder materials in the matrix were general purpose cement (specific gravity 3.15 and fineness 350 m2/kg) and GGBFS (specific gravity 2.91 and fineness 445 m2/kg). Their chemical compositions are further outlined in Table 3. The aggregate used was oven-dried sand, passing through 300 microns with average particle diameter of 200 microns. In addition to this, water and high range water reducer (HRWR) were used as a solution to ensure sufficient flowability of the mix. HRWR was poly-carboxylic ether type with brand name ADVA650 and it was procured from GCP Applied Technologies, Australia. Additionally, hybrid combination of PE and steel fibre was also used to impart better spalling resistance and residual performance as recommended in authors’ previous work [6, 20, 32]. Further details on the properties of fibres are presented in Table 4.

Table 2 Mix proportions used in this study
Full size table
Table 3 Chemical composition of the binder material
Full size table
Table 4 Properties of the fibre used
Full size table

Specimen preparation and curing

The mixing was done in a 100-litre mixer with rotating shaft as per the procedure illustrated in Fig. 1. Firstly, the sand was poured in the mixer along with other binder material and mixed for around 1 min. Approximately 10% of the water + HRWR solution was added initially to reduce the plume. After the dry mixing, the remaining water + HRWR solution was added, and the mixing continued for 2–3 min. Additional HRWR was adjusted based on flowability requirements to achieve a homogeneous mix. Subsequently, fibres were gradually added, and mixing continued for 8–10 min until uniform distribution was achieved. Once the mix was prepared, it was transferred directly into each of the greased moulds in a single pour to avoid any layering effect and proper compaction was ensured through the use of a vibration table.

Fig. 1
figure 1

Mixing procedure of HSECC

Full size image

The types of specimens considered are (Fig. 2): cubes of size 50 mm, 75 mm, and 100 mm, 75 mm × 75 mm × 150 mm prism, cylinders of diameter 40 mm, 75 mm, 100 mm, and 150 mm with height to diameter ratio of 2:1. All the different sized specimens were cast in a single batch except the 150 mm diameter × 300 mm height cylinders to eliminate any variation due to mixing. Cylinders of 150 mm diameter were cast in three batches due to the limited number of moulds available. Nevertheless, for each of the casted batch, specimens were tested at ambient conditions in order to confirm that their mechanical properties are consistent.

Fig. 2
figure 2

Types of specimens used in the present study

Full size image

The specimens were demoulded 18–24 h after casting and were placed in fog room for curing for 28 days at 23 ± 1 °C and 95 ± 5% relative humidity. Thereafter, all the specimens were transferred to 20 ± 1 °C and 55 ± 5% relative humidity and were kept there for another 28 days. The casting procedure is further illustrated through Fig. 3.

Fig. 3
figure 3

Different stages of HSECC casting

Full size image

Testing program

After the completion of curing period, all the specimens were exposed to 200, 400, 600, and 800 °C following the heating profile depicted in Fig. 4a. The heating was done at a constant rate of 1 °C/min to avoid unnecessary thermal gradients that could affect the resulting strength. The dwell duration was confirmed by measuring the temperature at the centre and surface of a 150 mm cube when the temperature increased from 20 to 1000 °C. It was observed that approximately 1.3 h were required for the core of the specimens to reach the target temperature (Fig. 4b). Given that the shortest distance from the centroid of the 150 mm cube to any of its sides (75 mm) was equal to or greater than all other specimens tested in this study, a 2-h hold duration was considered adequate to ensure thorough heating of all specimen cores. This duration has also been employed consistently in previous studies [2, 33, 34].

Fig. 4
figure 4

Temperature-time profile used for elevated temperature testing

Full size image

After the completion of the dwelling period, the specimens were allowed to cool down naturally by turning off the furnace. Thereafter, specimens were tested under compression at a displacement rate of 0.05 mm/min. An average was calculated for each temperature range and specimen size by conducting tests on three specimens in each group. This resulted in a total of 120 specimens tested, comprising five temperature ranges, eight different specimen sizes, and three specimens per set. Cube and prism specimens were tested in the direction perpendicular to the casting, whereas cylinder specimens underwent testing parallel to the casting direction.

Following the compression testing, small samples were taken from the core of each HSECC specimen set. These samples were immersed in acetone for a minimum of 48 h to halt further hydration and then dried at 60 °C for an additional 24 h. Subsequently, a portion of the sample, both in solid and powdered form, was utilized for characterisation study to perform SEM and XRD analysis, respectively. For microstructural analysis, solid samples were initially carbon-coated and examined using a Phenom XL SEM operating at 15 kV and 1 Pa chamber pressure. Elemental composition of the samples was further analysed using the Phenom energy dispersive spectroscopy (EDS) system. XRD patterns were obtained using a Bruker D8 Advance XRD system with an accelerating voltage and current of 40 kV and 40 mA, respectively. Powdered samples were mounted in a Bruker back-loading sample holder and scanned from 5° to 80° 2θ, with a step size of 0.02° 2θ and a counting time of 2.5 s per step.

Results and discussions

Qualitative characterisation

Figures 5 and 6 show the colour change in the specimens of square and circular cross-section after exposure to elevated temperatures. It is evident that the colour change is quite similar across all types of specimens. As the temperature increases from 20 to 200 °C, the specimens turn to light-yellow shade, which gradually transitions to white at 800 °C. More importantly, none of the specimens experienced spalling. The general crack pattern was also similar with no observable cracks until reaching 600 °C. However, at 800 °C, more noticeable cracks appeared in the 50 mm cube and 40 mm cylinder specimens. It was likely due to their smaller size within their corresponding series. Since the dwelling duration was kept constant for the heating, it is expected that the smaller sized specimens may have achieved isothermal stage before the large specimens, leading to a longer dwell time at the reference temperature. This extended exposure likely contributed to the increased formation of cracks in these smaller sized specimens.

Fig. 5
figure 5

Surface of specimens with circular cross-section at elevated temperatures

Full size image
Fig. 6
figure 6

Surface of specimens with square cross-section at elevated temperatures

Full size image

Effect of specimen size on mass loss

Mass loss at elevated temperature is mainly caused by the loss of free water, absorbed interlayer water and chemically bound water. In addition, the melting of fibre and decomposition of matrix also slightly contributes to this loss [2]. Therefore, it is a relatively direct process primarily influenced by the matrix composition and level of heat exposure. Furthermore, the most substantial loss of mass, attributed mainly to water evaporation, occurs before 400 °C. As a result, the trend in mass loss significantly differs up to this temperature, depending on the extent of exposure.

Specimens with square cross-section

Figure 7 compares the average mass loss of specimens with square cross-section at elevated temperatures. It can be observed that the range of mass loss change is very similar for the size varying from 75 to 100 mm with very small deviation. The average mass loss increased to 6.2–6.9% at 200 °C, showing a steep rise till 400 °C where the mass loss becomes 13.6–14.3%. With further increase in temperature, rate of rise in mass loss decreases, and it only climbs to 15–16% at 800 °C. The 50 mm cube showed an obviously higher mass loss up to 200 °C, after which the rate of mass loss relatively reduced. It is well established that the mass loss rise is very steep until 400 °C due to the evaporation of free water, physically absorbed water, and the interlayer water [2]. The smaller 50 mm cube specimens reached the isothermal state more quickly, potentially staying at that temperature longer, which might have led to more evaporation and consequently greater loss of mass. This likely explains why the 50 mm cubes experienced the highest mass loss, while the 100 mm cubes had the lowest. However, as the temperature increased to 400 °C, where the evaporation of both free and interlayer water nears completion, the rate of mass loss stabilized. After this temperature, the relative difference between the specimens with similar cross-sectional shape but different size was found to be very small. For instance, mass loss observed in 50 mm cube was 13.1% at 400 °C, whereas it was 13.6% and 13.9% in 75 mm and 100 mm cube respectively.

Fig. 7
figure 7

Average mass loss in HSECC specimens with square cross section

Full size image

For the 75 × 75 × 150 mm prims specimens with same cross-section area but higher aspect ratio, the mass loss was similar as cubes until 200 °C as shown in Fig. 8 and the relative rate of mass loss only slightly increased at higher temperature ranges. For instance, mass loss at 800 °C was 15.7% in 75 mm cubic specimens, whereas it was 16% in prism. This slightly higher mass loss could be attributed to factors such as relative distribution of fibres and presence of better interconnected channel network in specimen with higher surface area.

Fig. 8
figure 8

Comparison of average mass loss between 75 mm cube and 75 mm prism

Full size image

Specimens with circular cross-section

The mass loss in specimens with circular cross-section also followed a similar trend as square cross-section specimen (Fig. 9). The mass loss among cylinders of diameters ranging from 75 to 150 mm was in a comparable range, with values between 6.1–7.5% at 200 °C, increasing to 15.9–16.9% at 800 °C. An exception was noted in the smaller 40 mm cylinder specimens at 200 °C, which exhibited higher mass loss, consistent with observations for 50 mm cubic specimens. The smaller 40 mm cylinder specimens reached the isothermal state more quickly, potentially prolonging their exposure and resulting in greater evaporation and subsequent mass loss. This further explains why the 40 mm cylinders experienced the highest mass loss, while the 100 mm cylinders showed the lowest. Moreover, 150 mm cylinder specimen suffered lower mass loss than 100 mm cylinder specimens at 400 °C, the reason for which is not particularly clear. It may have been due to variation in fibre distribution or different evaporation rate. Further research is required to corroborate this. After 400 °C, the trend of mass loss became consistent, and it marginally increased with increase in specimen size.

Fig. 9
figure 9

Average mass loss in HSECC specimens with circular cross section

Full size image

Additionally, prism specimens also experienced a slightly higher mass loss than 75 mm cylinder specimens with same aspect ratio as shown in Fig. 10. This may have again been due to the simultaneous exposure of higher surface area in case of prisms (prism surface area = 56,250 mm2 vs. 75 mm cylinder surface area = 44178.64 mm2) leading to higher dehydration, and consequently, higher mass loss.

Fig. 10
figure 10

Comparison of average mass loss between 75 mm cylinder and 75 mm prism

Full size image

In general, it is evident that change in specimen size has minimal effect on mass loss. The variation is likely to occur at temperatures below 400 °C. This is because of the fact that mass loss is mainly contributed by the evaporation of free and chemical bound water which predominantly occurs prior to reaching this temperature. Additionally, the formation of interconnected channels typically does not complete before this point due to the higher vaporisation point of PE fibres [32]. Therefore, the degree of evaporation may differ based on the duration of exposure, potentially affecting smaller specimens differently. However, beyond 400 °C, the impact of exposure duration becomes less distinct [35] and hence, the time to reach the isothermal state has very little influence on the total mass loss as evident from the obtained results.

Effect of specimen size on compressive strength

Ambient compressive strength

Figure 11 illustrates the compressive strength of different sized specimens at room temperature. It can be observed that for the square cross-section size varying from 50 to 100 mm, compressive strength varied from 97 to 112 MPa with 75 mm prism showing the least compressive strength. The cylinders also had comparable strength with the exception of 40 mm diameter cylinder which showed 130 MPa strength due to its smallest size in the set. This strength variation in HSECC with specimen size is not in line with behaviour observed for normal concrete where the cubes tend to have greater strength than cylinders or prisms in compression. For normal concrete, the stress becomes concentrated along specimen’s edge in case of cubes with lateral expansion being restraint by loading plate. Whereas the slender specimens also experience ‘length effect’ along with the stress concentration leading to greater decrease in the strength [36, 37]. However, this phenomenon is less pronounced in the context of HSECC due to the presence of a strong interfacial transition zone and a homogenous microstructure resulting from the incorporation of very fine aggregates (< 300 microns). It should further be noted that the effect of lateral expansion is more noticeable due to the difference in loading and placement direction [37]. This may be the reason of lesser strength in case of prism specimens. In general, it was observed that as the specimen size increased, the compressive strength decreased. This was due to the presence of smaller fracture process zone in small sized specimens which resulted in larger absorption capacity [38]. However, the difference was not as significant as typically observed in case of normal concrete.

Fig. 11
figure 11

Average compressive strength of different sized specimens at ambient temperature

Full size image

Residual compressive strength

Specimens with square cross-section

Figure 12 compares the normalised residual compressive strength of specimens with square cross section. The normalised compressive strength has been defined as the ratio of the compressive strength at a specific temperature to the average room temperature compressive strength. It can be observed that the residual strength varied with the size until 600 °C and became fairly similar at 800 °C. However, the exact trend of variation is not distinctly apparent due to a high deviation. Initial strength retention was in the range 86.4–98.8% at 200 °C with 100 mm cube showing the highest strength. On further increase in the temperature, the strength decreased as expected and this decrease was highest in 50 mm cube. This could be attributed to the 50 mm cube reaching the isothermal state faster than the other specimens, leading to higher initial decomposition in this temperature range.

In general, the effect of elevated temperature was less severe for larger sized specimens. This is likely because specimens with a higher volume-to-surface area ratio suffer less damage during initial heating stages due to the longer time required to reach the same temperature level. Notably, at 800 °C, the strength difference among various sized cubic specimens was minimal. This similarity suggests that at higher temperatures, the decomposition process reaches a near-completion stage, resulting in comparable phases of decomposition and consequently similar strength. CSH is transformed to C2S and C3S at this temperature potentially contributing to a degraded matrix [2] and hence, the strength significantly decreased irrespective of the specimen size. Nonetheless, the specimen with higher volume and surface area may show different trend as further explained in the following section.

Fig. 12
figure 12

Average residual compressive strength of specimens with square cross section

Full size image

Interestingly, the overall trend is not in line with the rise in mass loss with specimen size (Fig. 7). This may have mainly been due to the dependency of the mass loss and compressive strength on different factors. Compressive strength at elevated temperatures is influenced by not only dehydration or evaporation of water, but decomposition and extent of microcracks may also significantly alter the strength. Therefore, unlike mass loss which mainly accounts for evaporation of water, the influence of degree and extent of temperature exposure will be more apparent for compressive strength. This was also confirmed by Yu et al. [35] who noted that the mass loss is affected with increase in duration of exposure only until 400 °C and thereafter, the effect of duration of exposure is minimal. However, the effect of duration was clearly noticeable for compressive strength at all temperature ranges. This observation is particularly crucial in the present case, as the time required to attain the isothermal state in different sized specimens will be different, affecting the degree to which specimens are heated.

The decrease in case of prism was also higher when compared with 75 mm cube as shown in Fig. 13. At a temperature of 600 °C, 75 mm cube exhibited around 75% strength retention, whereas the strength retention in case of prism was approximately 67%. The higher decrease may have been mainly due to the larger height or surface area of the prism specimen directly exposed to the temperature, making it more susceptible to damage. This has also been previously highlighted by Zheng et al. [28] who inferred that the specimen with larger height suffer more damage. Moreover, a difference in the placement and loading direction as well as the lateral expansion due to size effect may have further amplified this effect. Nevertheless, as the temperature increased to 800 °C, the influence of size on the decomposition of hydration products diminished, resulting in approximately 60% strength retention. This may have again been due to the complete decomposition of hydration phases at this temperature as also evidenced in existing literature [2].

Fig. 13
figure 13

Comparison of average residual compressive strength between 75 mm cube and 75 mm prism

Full size image
Specimens with circular cross-section

Similar trend was observed for specimens with circular cross section as shown in Fig. 14 with distinctly evident size effect. With increase in the temperature, the specimens with lower size suffered higher increase; however, the difference tends to diminish at higher temperature range. This may have also been due to the lower volume (V) to surface area (SA) ratio of smaller sized specimens. At 200 °C, the lowest strength retention was seen in 40 mm cylinders (V/SA = 8) which retained around 84.6% strength, whereas the percentage retention was 91.6%, 93% and 95.17% for 75 mm diameter (V/SA = 15), 100 mm diameter (V/SA = 20) and 150 mm diameter cylinders (V/SA = 30) respectively. The same trend was observed with increase in temperature, the reason for the higher strength retention in 75 mm diameter cylinders at 600 °C is not clearly understood and warrants further investigation. At 800 °C, the effect of size of cylinders subsided and around 39% strength retention was observed for all specimen sizes. This suggests that at higher temperature range where the hydration products approach near complete decomposition stage, the influence of specimen size on strength may become insignificant.

Fig. 14
figure 14

Average residual compressive strength of specimens with circular cross section

Full size image

The same reason could also be attributed to the greater strength reduction in case of 75 mm diameter cylinders compared to 75 mm prisms despite having the same aspect ratio as shown in Fig. 15. The reduction in strength was approximately 20% for 75 mm prisms, whereas 75 mm cylinder showed around 26% reduction at 400 °C. Since both 75 mm cylinder and prism have same volume to surface area ratio (V/SA = 15), a lesser decrease in case of prism could be related to its higher volume. Specimens with larger volumes are less susceptible to decomposition due to reduced exposure of their core at specific temperatures. Therefore, prisms experienced less strength reduction at elevated temperatures compared to cylinders, except at 200 °C. The strength loss below 400 °C is influenced by multitude of factors including fibre melting, pore-pressure dissipation, and the loss of chemically bound water. Additionally, there is a change in cohesion forces between the CSH layers which could have both positive or negative effect depending on the pozzolanic reactivity of the matrix constituents [39, 40]. Therefore, no distinct trend associated with size effect was observed in this temperature range and a detailed analysis of these observations may be required for establishing a clear trend at this temperature. Beyond 600 °C, the influence of these factors tends to diminish, especially since the fibres are nearly completely decomposed and the water is almost entirely evaporated and hence, a relatively consistent trend was observed.

Fig. 15
figure 15

Comparison of average residual compressive strength between 75 mm cylinder and 75 mm prism

Full size image

If the strength of specimens with different aspect and shape ratio is compared, the percentage decrease was observed to be higher in the specimens with higher aspect ratio. As shown in Fig. 16, cubic specimens suffered less decrease as compared to the cylinder specimens and the trend remained consistent from 200 to 800 °C. Considering the observed trend, surface area and volume to surface area ratio emerges as the primary parameter to consider. For the specimens with greater aspect ratio, the decrease will be generally higher due to their higher surface area directly exposed to heat load. For the specimens with roughly similar surface area, volume to surface area ratio could be used to define the variation trend with rise in temperature. It should further be noted the influence of specimen size may become less pronounced at higher temperatures, particularly at 800 °C, likely due to the decomposition that occurs. However, this effect also is contingent on the specimen type (cylinder vs. cube vs. prism) and how quickly the core reaches the specified temperature and the duration it remains at that temperature.

Fig. 16
figure 16

Residual compressive strength of different sized HSECC specimens

Full size image

Microstructural characterisation

SEM analysis

SEM analysis was performed to further compare the behaviour of different sized specimens. The analysis focused on the temperature range of 400–600 °C, where a consistent decomposition takes place, and the respective strength trend was also consistent. A comparison was made to distinctly assess the effect of specimen size (75 mm vs. 150 mm diameter cylinders) and specimen cross-section shape (75 mm cylinder and 75 mm prism).

Figure 17 shows the SEM micrographs of cylinder specimen with 75 mm and 150 mm diameter after exposure to 400 °C. It can be observed that the decomposition stage is almost similar in both specimens. Both micrographs exhibit compact microstructure with no observable degradation or microcracks. Furthermore, energy dispersive spectroscopy analysis confirmed the presence of the same types of elements in the analysed region, indicating that the dehydration or decomposition stage did not significantly differ between cylinders of different diameters. However, the difference in the residual strength might have occurred due to the varying relative quantity of different hydration compounds. As a result, phase analysis would be necessary to further understand this trend, which has been included in the subsequent section.

Fig. 17
figure 17

SEM micrographs of a 75 mm b 150 mm diameter cylinder specimens after 400 °C exposure

Full size image

Figure 18 further shows the SEM micrographs of the specimen of different cross-sectional shape (75 mm prism and cylinder) after 600 °C exposure at two levels of magnification (1000× and 2500×). In both types of specimens, there were no observable microcracks near the molten fibre, and the decomposition stages of the material appeared similar. However, at 2500× magnification, microcracks were more evident in the cylinder samples compared to the prism samples. These findings align with the results from compressive strength tests, which indicated that cylindrical samples exhibited lower strength retention. This provides further evidence that the presence of these microcracks likely contributed significantly to the observed differences between these specimens.

Fig. 18
figure 18

SEM micrographs of specimens with 75 mm cross-section a cylinder b prism after 600 °C exposure (×1000, ×2500 magnification)

Full size image

XRD analysis

Previous section examined how elevated temperatures affect microcrack formation in specimens of various sizes using SEM analysis. However, the influence of the size difference on decomposition processes could be better understood with XRD analysis. Figure 19 shows the X-ray diffraction curves of cylinder HSECC specimens after exposure to 400 and 600 °C, where major peaks of portlandite (18.04°, 34.11°, 47.13°, 64.28°), calcite (29.45°), ettringite (9.07°), albite (between 26.7–28°) and alite (31–33°) are highlighted. In addition, the peaks of quartz are also shown to establish a baseline for the relative comparison of different sized specimens.

Fig. 19
figure 19

X-ray diffractograms of cylinder specimens at temperature exposure of a 400 °C b 600 °C [A—Alite, A′—Albite, C—Calcite, E—Ettringite, P—Portlandite, Q—Quartz]

Full size image

The XRD results indicate that amount of ettringite is minimal in both specimens exposed to 400 and 600 °C, indicating that the decomposition had already initiated. Furthermore, as the temperature increases from 400 to 600 °C, the intensity of the portlandite peak decreases, also suggesting an ongoing decomposition process. Subsequently, at temperatures beyond 600 °C, the peaks corresponding to calcite and other decomposed compounds become more pronounced. Additionally, a comparison can be further made among different cross-sectional area cylinders by comparing the major peak of portlandite at 18.04°. In specimens exposed to 400 °C (Fig. 19a), there is a noticeable increase in the intensity of this peak as the cylinder diameter increases from 40 to 150 mm, indicating a higher presence of hydration compounds. A similar trend is observed in specimens exposed to 600 °C (Fig. 19b). These findings are consistent with mechanical performance tests, which also indicate that larger specimens are less affected by elevated temperatures, possibly due to differences in the duration they remain at a specific temperature after reaching isothermal state.

If compared among the different shapes (75 mm prism, 75 mm cube vs. 75 mm cylinder), the peak phases and respective decomposition patterns was found to be almost identical (Fig. 20). There were no observable differences in the phases, indicating that the decomposition progresses at a similar rate for different specimen types. This finding suggests that altering the shape of the specimen might not significantly impact the decomposition pattern when exposed to high temperatures. However, the extent of microcracks may vary as observed through SEM analysis and this may have resulted into the slight variation in the strength.

Fig. 20
figure 20

X-ray diffractograms of specimens with 75 mm cross-section after 600 °C exposure [A—Alite, A′—Albite, C—Calcite, E—Ettringite, P—Portlandite, Q—Quartz]

Full size image

Conclusions

This study investigated the influence of size of HSECC specimens on the compressive performance after exposure to temperatures from 200 to 800 °C. Cylinder specimens with 40–150 mm diameter (height/diameter = 2:1), cubes with size 50–100 mm side and prism of size 75 × 75 × 150 mm were selected to study the effect of specimen size and shape on the residual compressive strength. Based on the observations, the following conclusions can be drawn:

  • Specimen size did not have significant influence on mass loss especially for temperature above 400 °C. A slight variation was observed at 200 °C which could be attributed to the higher mass loss caused by prolonged exposure duration. However, residual compressive strength was found to be considerably influenced by both specimen size and shape. This was in clear contradiction to the existing study [8] that had indicated no significant impact of size on strength under high temperature exposure. Nevertheless, this difference diminishes at higher temperature ranges, particularly beyond 600 °C.

  • The obtained results suggested a strong correlation between the residual strength and aspect ratio or the volume to surface area of specimens. In general, the cubic specimens (aspect ratio 1) suffered less strength loss for the same cross-section sized prisms (aspect ratio 2). It was also noted that specimens with lower volume-to-surface-area (at constant aspect ratio) deteriorated more rapidly due to faster heating at a specific temperature.

  • Based on the consistency of results and optimal utilization of material, the use of 75–100 mm diameter cylinder specimens (with 2-hour exposure duration) is recommended for elevated temperature studies on HSECC. Adoption of this size would ensure coherent results with minimal influence of external factors such as fibre melting, placement direction or heightened exposure which are critical below 400 °C.

  • While specimen type clearly influenced compressive strength variations, establishing a definitive relationship between specimen size or shape and resulting residual compressive strength may require further experimental trials, considering time to reach the isothermal stage and varying heating rates. Future studies should explore precise dwell durations based on isothermal stages for specimens of different sizes to establish clearer trends.