The impact of changes in pore structure on the compressive strength of sulphoaluminate cement concrete at high temperature
Published Online: Jul 06, 2021
Page range: 75 - 85
Received: Feb 25, 2021
Accepted: Mar 12, 2021
DOI: https://doi.org/10.2478/msp-2021-0006
Keywords
© 2021 J.J.K. Tchekwagep et al., published by Sciendo
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Previous research has highlighted the risks presented by the dehydration and loss of strength experienced by concrete due to exposure to elevated temperatures. Al-Salloum et al. [1] exposed concrete to temperatures of 100 °C and 200 °C for periods of 1, 2 or 3 hours, concluding that significant degradation of bonding strength occurred at 200 °C. Ma et al. [2] described the serious damage to the micro-structure and meso-structure that concrete suffers due to fire, including general mechanical decay and even detrimental effects at the structural level, and called for additional studies of the behaviour of cement composites at high temperatures. We know that under the influence of elevated temperature, the internal pore structure of Portland cement concrete changes. Pore structure is one of the most important characteristics of concrete generally, and a change in pore structure significantly affects the mechanical properties of the material. Therefore, it is important to study the changes that occur in SACC pore structure and mechanical properties after exposure to elevated temperatures.
Tchekwagep et al. [3] found that at elevated temperatures, due to the decomposition of hydration products, the strength and stress-strain behaviour of SACC are affected, therefore, understanding the micro structure pore system as the temperature increased is necessary to further explain the strength decrease. Kumar et al. [4] used the mercury intrusion porosimetry (MIP) method to test the pore structure of Portland cement concrete under drying at 60 °C and 105 °C. Galle found that, compared with freeze-drying, elevated temperature changed the pore structure of Portland cement concrete and increased its pore diameter more extensively; this finding has guided our choice of the MIP method to characterize the pore structure in the present work. Ermiati et al. [5] determined that at temperatures of 150 °C and 200 °C, the permeability of the Portland cement concrete structure was significantly greater. This was due to the increase in the width of pores and micro-cracks in the concrete at elevated temperature. Amadu et al. [6] studied the pore structure characteristics of steel fibre concrete at 20 °C, 400 °C and 800 °C, finding that after exposure to elevated temperature, the pores with diameter greater than 200 nm increased significantly and those with a pore diameter less than 50 nm decreased accordingly. Liu [7] divided the pores into several categories with regard to their impact on concrete performance: harmless (< 20 nm), less harmful (20 to 50 nm), harmful (50 to 200 nm) and more harmful (> 200 nm); this categorization is helpful to the present work, the purpose of which is to identify and characterize the range of pores present in SACC as the temperature increases. Wu Zhongwei proposed increasing the number of pores below 50 nm and reducing the number above 100 nm so as to improve performance. Anovitz [8] believed that only pores with a diameter greater than 132 nm affected the strength and permeability of Portland cement concrete. Kodur [9] tested the compressive strength of Portland cement concrete at 20 °C to 800 °C, and the results showed that with increased temperature, the strength continually decreased. Eurocode [10], ASCE [11] reported similar results. du Plessis [12] showed that when the temperature was higher than 400 °C, the dehydration of chemically bound water in Portland cement concrete had a greater impact on the strength.
There are three types of pores in concrete: gel pores (< 10 nm), capillary pores (10 – 200 nm) and macropores (> 200 nm) [13]. Gel pores are isolated pores distributed in the hydration product. Capillary pores are not filled with hydration product. Large holes, or macropores, occur due to uneven vibration during concrete pouring. Under the influence of elevated temperature, the pore structure of SACC changes mainly due to the evaporation of capillary water, physically bound water and chemically bound water. This change in pore structure could lead to a significant decrease in the concrete’s essential structural properties, such as strength and durability. It is generally believed that at 100 °C, the capillary water and physical bound water evaporate completely as found by Mir [14]. At 20 °C, the capillary water and physically bound water with weaker binding force are relatively stable in the concrete. When the temperature exceeds 100 °C, the C–S–H gel, ettringite and Al(OH)3 undergo further dehydration, gradually decompose and dissipate. At the same time, the pore walls of the gel pores, formed by chemically bound water, will peel off, adjacent pores will gradually communicate, and the pore structure will continue to change.
At present, most work on the pore structure of concrete has relied on the mercury injection method. When this method is used to measure the micropore structure, the mercury can easily destroy the pore structure under external pressure, resulting in inaccurate test results. Therefore, this work uses both the mercury intrusion method and the nitrogen adsorption method, along with compressive strength testing, to analyse the changes in the pore structure of SACC at 100 °C, 200 °C and 300 °C and to gain further understanding of the resulting impact on the concrete’s strength after exposure to these temperatures.
Concrete of grade C40 was adopted in this work. SAC P·O 42.5 grade cement from Wuhan Yadong Yangfang Cement Group Co. Ltd., Shandong province, China was used for the testing; its chemical composition is shown in Table 1. The coarse aggregate was 5–20 mm continuous graded granite crushed stone; the fine aggregate was natural river sand with a fineness modulus of 2.3; the water was ordinary tap water. Table 2 shows the mixture proportions used in this study. Standard 10 cm cube specimens were used. They were demoulded after one day and then cured in a standard curing room. It has been suggested that thermal shrinkage, internal shrinkage and incomplete hydration might affect the results of testing like that proposed in this experiment. Therefore, we elected to extend the standard curing of our specimens to 12 months.
Chemical proportions of cement in concrete
| Oxides |
|||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| CaO | Al2O3 | SO3 | SiO2 | Fe2O3 | MgO | TiO2 | K2O | SrO | Na2O | Cl | P2O5 |
| 519.4 | 86.2 | 86.2 | 40.5 | 80.0 | 13.1 | 5.1 | 6.8 | 59.9 | 0.5 | 1.3 | 0.5 |
| 45.28% | 17.51% | 15.76% | 9.19% | 2.50% | 1.90% | 0.75% | 0.48% | 0.17% | 0.19%0.11 | % | 0.10% |
Mix proportions of SACC
| Cement kg/m3 | Water kg/m3 | Aggregate kg/m3 | Sand kg/m3 | Water reducer % |
|---|---|---|---|---|
| 430 | 160 | 975 | 835 | 1% |
After the curing was completed, the specimens were put into drying ovens at 20 ± 3 °C for drying treatment until the sample quality was no longer changing; then they were subjected to varying temperatures from 100 °C to 300 °C. The specimens were placed inside a temperature-controlled electric furnace. The temperature was increased at a rate of 4 °C/min up to the test temperature, which was then held constant for 4 hours to attain steady state conditions (thermal equilibrium). After the four-hour hold, the temperature was reduced to ambient (20 °C) at a controlled rate of 2 °C/min. After that, the compressive strength of each specimen was measured. To ensure the accuracy of the testing, the average test strength of three specimens per temperature was taken as the compressive strength of the SACC specimen. After the compressive test, the central part of each SACC specimen was taken. We took several small SACC pieces without coarse aggregate as test samples for the mercury intrusion method and nitrogen adsorption method for each temperature. As noted above, the mercury intrusion method can destroy the pore structure, resulting in inaccurate test results, whereas the nitrogen adsorption method is suitable to fully measure small holes. Therefore, the nitrogen adsorption method was used to determine the gel pores smaller than 10 nm, and the mercury intrusion method was used to determine the pores larger than 10 nm.
The test of SACC compressive strength was based on the ‘Test code for hydraulic concrete’ (SL352-2006) [15] procedure, which calls for polishing the surface of the test piece before the start of the test and ensuring that it is flat. After the start of the test, load was applied continuously at a speed of 0.5 MPa/s, and the testing machine automatically recorded the failure load.
The formula used for calculating the compressive strength was:
The sample was tested using a Quantachrome Pore Master 33G mercury porosimeter. The volume of the internal pores of the measured object was obtained by measuring the pressure and the consumption volume of liquid mercury following the specific expression [4]:
In the formula, P is the mercury intrusion pressure (Pa); γ is the surface tension of mercury, generally 480 × 10−3 N/m; θ contact angle being 130 °C in this study and d pore diameter (size). The experimental process was carried out in strict accordance with the oil and gas industry standard SY/T 5346-2005, ‘Determination of Rock Capillary Pressure Curve’ [16]. Compared with the nitrogen adsorption method, the mercury intrusion method has higher measurement accuracy for macropores. Therefore, only data on pores larger than 10 nm were considered for analysis using this method.
This test used the US Quantachrome AutosorbiQ2-C two-station physical adsorption device. The partial pressure (P/P0) can range from 5×10−8 to 0.995, the specific surface area measurement range is greater than 0.29 m2/g, and the measurement repeatability error is less than 1.6%. The degassing treatment was carried out at 20 °C, 100 °C, 200 °C and 300 °C, respectively, until the vacuum pressure reached a range from 39.5 Pa to 53.6 Pa. After the treatment, the test can obtain the isothermal absorption-desorption curves of the samples at different degassing temperatures. Relative to the mercury intrusion method, the nitrogen adsorption method produces larger errors in the measurement of macropores. Therefore, only the data on pores below 10 nm were considered for analysis during the processing of the data obtained by this method. According to the isothermal adsorption-desorption curve standard, [17] the quenched solid density functional theory (QSDFT) method [18] was used to obtain the pore size distribution characteristic curve of the SACC sample.
The compressive strength test results of the SACC samples at different temperatures are shown in Table 3. As noted above, the compressive strengths reported here represent the average values of three specimens per exposure temperature. The compressive strength decreased with increasing temperature. Taking the compressive strength of SACC at 20 °C as the standard state, the relative decrease in compressive strength of concrete at other temperatures, R, can be defined as:
Compressive strength results of the SACC
| Specimens types | Temperature (°C) |
|||
|---|---|---|---|---|
| 20 Ref | 100 | 200 | 300 | |
| Compressive strength | 51.31 | 39.52 | 37.88 | 28 |
| R/% | 0 | 22.97 | 26.17 | 45.42 |
In that formula, σ0 is the compressive strength of SACC at 20 °C (MPa); σ is the compressive strength of SACC at other temperatures (MPa). At 100 °C, the compressive strength of SACC decreased by 23% relative to that at 20 °C. As the temperature continued to rise, the strength continued to decrease, although the rate of change in strength varied, with slower change between 100 °C and 200 °C. The strength reduction value reached 45.42% at 300 °C, which has a large adverse effect on the performance of the SACC.
The mercury intrusion porosimetry (MIP) method (Fig. 1) shows the cumulative mercury ingress of SACC samples that were conditioned at different temperatures. It can be seen from Fig. 1 that the porosity is relatively sensitive to temperature conditioning. For temperature exposures of 20 °C, 100 °C, 200 °C and 300 °C, the average porosity of SACC was 2.98 ml/g, 4.76 ml/g, 5.53 ml/g and 7.71 ml/g, respectively; therefore, the porosity increased with the exposure temperature. When the exposure temperature changed from 20 °C to 100 °C, the porosity change was considerable for all the samples. Moreover, for all the exposure temperatures, the curves of cumulative mercury intake of the SACC samples also exhibited considerable change. The threshold pore size at which the SACC started to uptake a large amount of mercury had a close relationship with temperature. At exposure temperatures of 20 °C, 100 °C, 200 °C and 300 °C, the threshold pore size was 100 nm, 130 nm, 7.5 μm and 20 μm, respectively. In summary, the higher the temperature, the greater the porosity of the SACC and the size of the pores.
Fig. 1
Cumulative intrusion-pore diameter curves of samples (MIP).
The pore size distribution of SACC samples after exposure to different temperatures is shown in Fig. 2 and Table 4. It can be seen that the pore size distribution curve of SACC shifted to the left at higher temperatures and the overall pore size increased. Compared to the 20 °C control samples, the capillary pore volume of specimens exposed to 100 °C rose from 2.03% to 2.46%, and the macropore volume rose from 0.25% to 0.30%. At the same time, the ratio of pores and macropores changed proportionally. This indicates that as the temperature increases, the total pore volume in the SACC increases and the pore structure changes. This may be because, as the physically bound water in the inner layer of SACC evaporates between 20 °C and 100 °C, the evaporation is accompanied by a decomposition of chemically bound water, which causes changes in the pore structure. When the temperature was increased to 200 °C and 300 °C, the trend of the pore size distribution curve of SACC was similar, and the most probable pore size continued to increase (to 94.91 nm and 99.89 nm, respectively); while the overall pore volume gradually decreased, the macropore volume gradually increased. Based on the data shown in Fig. 1, the capillary pore volume increased steadily; on the other hand, the macropore volume jumped between 20 °C and 100 °C and then gradually decreased again with further temperature increase. At the same time, the chemically bound water dissipated, the pore volume of SACC continued to increase, and the pore structure started to change. The transitional pores with diameter of 10–100 nm were found to generally increase with temperature. This expansion of pore size is known to cause the concrete structure to shrink.
Fig. 2
Pore size distribution curves of the sample (MIP).
Pore size distribution parameters of specimens (MIP).
| Temperature (°C) | Average porosity (ml/g) | PHg | Threshold diameter (nm) | Most probable aperture (nm) | |
|---|---|---|---|---|---|
| Capillary pores (10 – 200 nm) | Air voids (> 200 nm) | ||||
| 20 (ref.) | 2.89 | 0.20 | 2.98 | 100 | 27.26 |
| 100 | 3.30 | 0.60 | 4.76 | 130 | 59.03 |
| 200 | 4.38 | 0.90 | 5.53 | 7500 | 98.91 |
| 300 | 4.98 | 1.89 | 7.71 | 20000 | 118.40 |
For specimens exposed to a temperature of 300 °C, two major peaks, at 300 °C and 200 °C, appear in the pore size distribution curve. The largest peak was at 117.3 nm. The log differential intrusion versus the pore size diameter has a local peak. This indicates that there were more large pores distributed around 15 μm in the SACC. In the specimens heated to 300 °C, the macropore volume of the SACC increased significantly. The large amount of decomposition of chemically bound water damaged the initial pore structure.
According to Kelvin’s equation, under very low relative pressure (<0.01), the micropores are filled sequentially. Under low pressure, nitrogen is adsorbed in a single layer on the microporous surface; when the relative pressure gradually increases, the process changes to multilayer adsorption, and the capillary agglomeration starts; then, under relatively high pressure, capillary agglomeration occurs. After the adsorption is completed, the desorption process begins. The relative pressure gradually decreases, and when the relative pressure corresponding to the Kelvin radius of the hole is reached, capillary evaporation occurs. When condensation and evaporative adsorption isotherms branch due to different relative pressures, adsorption loops occur; the shape of the adsorption loop reflects the pore characteristics of the material.
Due to the large amount of data, Fig. 3 shows only the adsorption isotherms of SACC samples at 20 °C to 300 °C. As shown in Fig. 3, the absorption isotherm of the sample has an obvious ‘hysteresis loop.’ At different degassing temperatures, the hysteresis loops of each sample are similar in shape, and they are all at a medium relative pressure (0.3 to 0.5). There are obvious inflection points, indicating the presence of tubular capillaries with open ends. When the relative pressure is low, the adsorption isotherms basically coincide, indicating that most of the pores in the small pore size range are airtight pores closed at one end. According to International Pure Chemistry and applying the type of hysteresis recommended by IUPAC, Fig. 3 shows that the curve is a III-type curve with H3 hysteresis loop–type model.
Fig. 3
Isotherm adsorption.
The high relative pressure area does not show any adsorption limitation; there are parallel plate-shaped cracked pores, the hysteresis loop is wide, and the mesopore distribution is relatively wide. Fig. 3 also shows that the higher the degassing temperature, the lower the adsorption isotherm of the sample. This phenomenon indicates that with exposure to elevated temperature, the macropores in the sample increase, the pore volume increases, and the specific surface area decreases, resulting in a decrease in the adsorption capacity of the sample.
The pore size distribution of the SACC samples based on nitrogen adsorption is shown in Fig. 4 and Table 5. The most probable pore size of SACC changed for specimens exposed to different temperatures, and it is distributed from 5.3 to 5.5 nm. As the temperature increased, the peak pore size distribution of SACC gradually decreased.
Fig. 4
Pore size distribution of SACC (N2 absorption).
Pore size distribution parameters of specimens (N2 adsorption).
| Temperature (°C) | Average Micropore porosity (ml/g) | Average threshold diameter (nm) | Most probable pore size (nm) | Average peak value (ml/g) |
|---|---|---|---|---|
| 20 | 0.066 | 9.92 | 5.20 | 0.051 |
| 100 | 0.055 | 11.96 | 5.50 | 0.042 |
| 200 | 0.050 | 13.57 | 5.50 | 0.037 |
| 300 | 0.043 | 14.49 | 5.20 | 0.032 |
The pore size distribution curve of the SACC samples was notably different for the 100 °C specimens compared to the 20 °C controls, and the gel pore volume was much different. This indicates that the evaporation of physically bound water and capillary water affected the rate of change in the gel pore structure of SACC. For specimens exposed to 200 °C and 300 °C, the gel pore volume also decreased. However, the volume reduction was lower compared to the previous examples. At elevated temperatures, as the chemically bound water decomposed, the gel pores gradually connected with adjacent pores. At 300 °C, the peak pore size distribution and the gel pore volume were greatly reduced, the pore connectivity increased, and the pore structure was damaged.
In concrete, water exists in three forms: chemically bound water, physically bound water and capillary water. Among these, chemically bound water is found in the primary hydration product as a result of the hydration process. Physically bound water is water adsorbed on the surface of solid materials, and its density is generally greater than that of standard water. Capillary water is free moisture contained in the pores. Under the effect of heating, the internal moisture is gradually lost. For the time frame of the tests, when the temperature was held at 20 °C, the capillary water in the concrete and part of the outer physically bound water were essentially stable. When the temperature was held at 100 °C, all the physically bound water and the capillary water evaporated.
Gawin et al. [19] proposed that the mass loss of chemically bound water varies with temperature. The equation for the degree of change is
The total porosity of the gel as measured by the nitrogen adsorption method, corresponding to the relationship between total porosity and mass loss of chemically bound water, is shown in Table 6 and Fig. 5. It can be seen from Fig. 5 that the total porosity increases linearly with the loss of chemically bound water. The linear correlation coefficient is very high at 0.98, although the statistical reliability is relatively low, since the coefficient is based on only three measurements.
The relationship between the total porosity of concrete and the mass loss of dimensionless chemically bound water.
| Temperature (°C) | Total porosity (ml/g) | Dimensionless chemically bound water loss (wd/c) |
|---|---|---|
| 100 | 9.9 | 0.034 |
| 200 | 11.9 | 0.075 |
| 300 | 14.4 | 0.113 |
Fig. 5
Linear fitting of experimental porosity and theoretical models.
The relationship between concrete compressive strength and porosity proposed by Schiller [20] is
Fig. 6
The relationship between porosity and compressive strength of SACC.
In summary, the compressive strength of the SACC decreases at elevated temperatures. This reduction in strength may be caused by the changes in pore structure. The compressive strength and porosity of SACC are strongly correlated. The influence of temperature on the pore structure and compressive strength of SACC can be described as follows. When hardened SACC is exposed for an extended duration of time to a temperature between 20 °C and 100 °C, capillary water and physically bound water begin to evaporate, but the water and hydration products in the gel pores do not evaporate. As Tables 4 and 5 show, when the samples were heated from 20 °C to 100 °C, the volume of gel pores increased somewhat and the volume of capillary pores and macropores increased substantially. The pore size distribution curves for those two cases were very similar to each other, but the pore structure changed and the pore void ratio increased by 0.6%, due to the evaporation of capillary and physically bound water at 100 °C. The compressive strength was reduced by 23%, and there was a strong correlation between the changes in porosity and in compressive strength.
As the temperature increased further (from 100 °C to 200 °C), the SACC became dehydrated, the chemically bound water (e.g., C-S-H gel, ettringite and Al(OH)3) were gradually lost, and the gel pores were destroyed as shown in Table 4. It can be seen from Table 5 that the volume of gel pores in the SACC decreased with each increase in temperature of exposure. As the water distributed in the isolated gel pores evaporates, it generates internal pressure on the pores, which increases the volume of gel pores to the extent that they nearly connect with each other. The compressive strength of SACC exposed to a temperature of 200 °C decreased compared to the sample exposed to 100 °C. At 100 °C and 200 °C, due to the evaporation of chemically bound water, the internal porosity of the concrete increases continuously, and the water in the gel pores generates stress, which causes the gel pores to be destroyed. The gel pores gradually communicate with the surrounding pores, and the pore structure changes. As the temperature increases, the compressive strength of SACC keeps decreasing.
When the specimen is exposed to a temperature of 300 °C, there is a distinctive local peak in the pore distribution curve at 15 nm, as shown in Fig. 2, which does not correspond to any of the other temperature exposures. This shows that with the continuous loss of chemically bound water, and compared with the temperature at 200 °C, the concrete has more large-scale pores distributed around 15 nm in size. At the same time, Tables 4 and 5 show that the pore volume of the gel is greatly reduced. At this point, the volume of the macropores increased significantly. By the time the temperature reaches 300 °C, the pore structure has become seriously damaged, with a correspondingly extensive reduction of 45.42% in compressive strength relative to the strength at 20 °C.
The pore structure of SACC materials is very sensitive to increased temperature. As the ambient temperature rises, the porosity of SACC materials increases continually as well. When the temperature reaches 100 °C, the change in porosity is related to the mass of chemically bound water in SACC. That loss is linear in nature.
Under elevated temperature conditions, the pore structure of SACC also changes, due to the loss of chemically bound water. By 100 °C, because of the loss of capillary water and physically bound water, the pore structure of SACC begins to change. As the temperature increases beyond 100 °C, the chemically bound water begins to decompose and evaporate continuously. Between 100 °C and 200 °C, the pore structure of SACC changes steadily, in a linear relationship to the change in temperature. When the temperature reaches 300 °C, the pore structure is greatly damaged.
The relationship between the compressive strength and porosity of SACC at different temperatures can be quantitatively reflected by the Schiller formula. The results of this experiment indicate a strongly linear correlation with the formula. However, the regression was based on limited data, so further research is warranted to validate this observation.