Effect of dehydration time and air tightness on pore distribution of potassium and metakaolin-based geopolymer

The Fukushima Daiichi Nuclear Power Plant failed to cool down the nuclear fuels in the reactor core due to a huge earthquake and tsunami. A large amount of spent fuel melted into the cooling water, and radiolysis of the water due to the radioactive contamination produced hydrogen and oxygen. The accumulation of hydrogen may cause hydrogen explosion, and preventing hydrogen explosions has been an issue in the study of waste storage. ¹⁾ The fine radioactive waste particles combine with water to form a slurry, so it is difficult to remove the water. ²⁾

During the decommissioning of Three Mile Island Unit 2 (TMI-2), radioactive waste and water were processed and stored in tanks with Pd-Pt-Al₂O₃ catalyst to promote hydrogen and oxygen recombination. ³⁾ On the other hand, the Fukushima Daiichi Nuclear Power Plant must operate multiple radionuclide removal and other water treatment systems, which will generate a lot more water-carrying radioactive waste than the one of TMI-2. ²⁾ Therefore, it is very important to develop a cost-friendly catalyst and a support composed of metal or inorganic materials and have long-term chemical stability.

Geopolymers are inorganic alumina-silicate materials, prepared by mixing metakaolin powder or another aluminosilicate rich in silica and alumina and alkali activators, such as sodium hydroxide (NaOH) or potassium hydroxide (KOH) and sodium silicate (Na₂SiO₃) or potassium silicate (K₂SiO₃). ⁴⁾ Geopolymers have excellent physical and chemical properties, such as high compressive strength, ^5–10) low shrinkage, ¹¹⁾ low thermal conductivity ^12,13), and alkali or acidic condition resistance. ^14–16) Geopolymer can be hardened at temperatures below 100 ℃ and a large variety of inexpensive raw materials (clay, metakaolin and certain industrial waste) can be used for geopolymer production. ^5,17–20) Consequently, geopolymers have been developed as candidate materials for hydrogen recombination catalyst support in previous research. ^2,21)

The mechanical strength of geopolymers is strongly affected by the content of K₂O and the water coefficient of H₂O/metakaolin, while the compressive strength value of the sample is not significantly different from the SiO₂ to K₂O molar ratio variable. ²²⁾ When a sample is cured at a high temperature, compared to one cured at room temperature, the higher curing temperature will increase the early compressive and flexural strength, and even reach the target value within 1 d. In addition, the effect of temperature depends on the curing time. Curing for only 1 h at elevated temperatures will not cause a significant change in strength development, but a longer treatment time will significantly accelerate the reaction rate and increase the early strength. However, higher curing temperatures above 80 ℃ have a negative impact on the development of hardened geopolymers. ²³⁾ Geopolymers contain pores and penetration of external species occurs through the pore network. ²⁴⁾ The strength of concrete is its most important characteristic and is also affected by these pores. ^25–27) Research ²⁸⁾ has shown that total porosity and pore size distribution are the most important factors affecting the compressive strength of samples, and when the proportion of macropores increases, the compressive strength of the sample decreases. Since concrete is porous, the structure is filled with small pores through which water, air, acid and alkali can pass. ²⁹⁾

The porosity and pore distribution of geopolymers play an important role in the application of catalyst supports. Previous research ²¹⁾ showed that macropores may be formed in the first 4 d. In the present work, potassium and metakaolin-based geopolymer samples were synthesized. The workflow was carried out by evaluating the mechanical strength value and pore size distribution of the geopolymer, which are based on two factors, the varying early-age curing time and air tightness after early-age curing.

Metakaolin was obtained from Sobue Clay Co. Ltd., Japan, with 1.1 wt% of impurities (mostly TiO₂). Amorphous fumed silicon, SiO₂, called EFACO silica, was purchased from Tomoe Engineering Co., Ltd., Japan.

In order to prepare geopolymer composites, an alkaline solution consisting of KOH powder (UNID.Co.Ltd.), potassium silicate solution, K₂SiO₃ (FUJIFILM Wako Pure Chemical Corporation; concentration 50%) and distilled water was prepared. The composition for making the geopolymer is listed in Table I.

The synthesis process of the geopolymer samples is shown in Fig. 1. EFACO silica powder was added to the solution, which was stirred for 1 min. Then, metakaolin was added to the solution, which was stirred for a further 4 min. Finally, the slurry was stirred for a further 5 min in order to make the mixture homogeneous. After the raw materials were mixed, the slurry was cast in φ37 × 60 mm plastic molds, wrapped with plastic film and sealed with lids to ensure good air tightness. The curing process is shown in Table I. The curing process was divided into two periods. The first one, called Curing 1, was dried at 60 ℃ for various times. The Curing 2 process took place at air temperature, and the time was adjusted so that the total curing time was 14 d. There were eight samples, divided into four groups. Each group comprised two samples, which had the same Curing 1 condition but different treatment in Curing 2. The four groups of samples were removed from the oven and stored at room temperature (20 ℃) and relative humidity 45% ± 5% until tested.

Zoom In Zoom Out Reset image size

The mass change after 14 d curing was used to calculate the rate of relative weight loss of the sample. A scanning electron microscope (SEM) was used to measure the pores in this study. Cross-section observation specimens were prepared by cutting a sample perpendicular to the longitudinal axis, 7 mm from the bottom and 7 mm from the top. Approximately 300 pores were observed on the cross-sections to analyze the pore size distribution on the bottom and top of the samples.

To analyze phases in the geopolymer, X-ray diffraction (XRD) with Cu-Kα radiation (0.15418 nm) was used with a scanning angle range of 10°–60° (voltage of 40 kV, current of 40 mA, scan speed of 10° min⁻¹ and step of 0.02°).

We ground the test surface of each sample for about 5 h to ensure that the surface of the sample could be more accurately measured using a microscope to measure the size of the indentation left on the surface of the sample. A Vickers indenter was loaded at 1 kgf for 15 s with 12 measurements for each sample.

The relative weight change of the samples (open lid) in Curing 2 is shown in Fig. 2. After Curing 1 was completed, the weight of all samples decreased faster in 1 or 2 d. However, the relative weight of all samples on the 14th day was about 90% and there was no significant difference. The shrinkage of the samples was obtained by calculating the volume of the 14th day and the initial volume. The results showed that for the samples that were sealed in Curing 2, the shrinkage rate was almost uniform at about 0.5% on the 14th day. For the sample with the lid open in Curing 2, the shrinkage rate of the samples cured at 60 ℃ from 1–4 d in Curing 1 was 2.2%, 2.2%, 2.1% and 2.1%, respectively.

Zoom In Zoom Out Reset image size

SEM images of geopolymer cured at ℃ with different times in Curing 1 are shown in Fig. 3. When the sample was cured at 60 ℃ for 1 d, the SEM image showed that the porosity is slightly larger than that of the other samples. The pore size distributions are shown in Fig. 4. When the samples were cured at 60 ℃ for the same length of time, the average pore size of the samples with a lid and without a lid was not very different. However, for the samples with the lid open, the average pore size ranged from 1.4 × 10² μm to 2.3 × 10² μm, while for the samples with a lid, the average pore size was mainly distributed from 0.9 × 10² μm to 2.0 × 10² μm. The average pore size of the samples with a lid tends to be smaller than those without a lid. The pore size of the top and bottom showed that they were mainly distributed from 1.1 × 10² μm to 2.5 × 10² μm. However, in the range from 1.6 × 10² μm to 2.0 × 10² μm, the frequency of the top pore size distribution was 21% significantly higher than that of the bottom.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The XRD patterns, as shown in Fig. 5, indicated that geopolymer was synthesized, and even some illite and SiO₂·xH₂O crystal peaks appeared in all the samples.

Zoom In Zoom Out Reset image size

The Vickers hardness is shown in Fig. 6. The error bar of each sample shows that the hardness of the sample is unstable. However, the average Vickers hardness did not change significantly with the increase in the number of days in Curing 1. When the sample was cured for 1 d at 60 ℃, its Vickers hardness on the 14th day had reached a high value.

Zoom In Zoom Out Reset image size

As the well-known reaction model assumes, ²⁴⁾ water acts as the medium and participates in several intermediate reactions during geopolymer formation, such as dissolution, polycondensation, etc. A previous report showed that water could exist inside the final products as free or bound water, and thus had an influence on the microstructure. ³⁰⁾ In this study, the relative weight change on the 14th day of metakaolin and potassium-based geopolymer was within 10% difference and was not affected by the length of curing time at 60 ℃ in Curing 1. We infer that the samples may have formed similar microstructures under different curing times in Curing 1. XRD results showed that illite and SiO₂·xH₂O were detected in all samples, confirming the similar microstructure mentioned earlier. The shift of the amorphous halo from 2Ɵ_max = 22° in the metakaolin processor to 2Ɵ_max = 27–30° in the cured geopolymer samples has been commonly explained by structural changes caused in aluminosilicates during geopolymerization. ^5,24)

The SEM image showed a slight increase in porosity in the sample cured for a shorter time in Curing 1. Figures 4(a) and 4(b) showed that regardless of whether the samples had a lid or not, a higher frequency of small pores appeared in the sample with a short curing time in Curing 1. However, the average pore sizes were not very different. Figures 4(c) and 4(d) showed that the distribution of pores at the top tends to be larger. The top pore distribution of all samples showed greater consistency. Figure 4(a) showed that the pores at the bottom tend to be larger as the curing time increases in Curing 1. However, the performance of the average pore size is almost the same, and the Vickers hardness of the geopolymer samples did not change. A report ²³⁾ showed that the produced geopolymer samples can reach high compressive strength after heating at 60 °C for 12 h. This experiment showed that the Vickers hardness of geopolymers can reach a high value of about 90 MPa in 1 d.

Synthesis and dehydration of potassium and metakaolin-based geopolymer were carried out. The pore size distribution and Vickers hardness of geopolymers under different early curing time and air tightness have been studied.

This study shows that when the sample is cured at 60 °C for 1–4 d and the air tightness is subsequently changed, the shrinkage rate remains relatively consistent, at about only 2.1%. We infer that potassium and metakaolin-based geopolymer has low shrinkage performance and is not affected by the length of early curing time.

The results of the Vickers hardness test showed that the sample was cured at 60 ℃ for 1 d, and the Vickers hardness on the 14th day had reached 90 MPa. Longer early curing did not significantly change the Vickers hardness on the 14th day.

The results of pore distribution and average pores size show that the pore size distribution at the top tends to be larger compared to that at the bottom. As the early-age curing time at 60 ℃ increases, the pore distribution at the bottom tends to be large pores, while the pore distribution at the top shows greater consistency. However, for all samples, the average pore size is almost the same at about 2.1 × 10² μm.

This work was supported by ADVAN ENG. Co., Ltd.