Introduction

Glass ceramic is a composite material composed of a mini-crystal phase and residual glass phase, which is made by controlling the crystallization of basic glass with a specific composition during the heating process [1,2,3,4,5]. Because its performance combines the advantages of glass, ceramic, and stone, it can be used as a building material, a functional material, and a structural material.

A large amount of metallurgical solid waste is accumulated and cannot be efficient utilization in ironmaking industry, which resulted in environment pollutions and resource waste [6,7,8,9]. Ti-bearing blast furnace slag is one of the metallurgical solid wastes produced in the blast furnace ironmaking process using vanadium-titanium-bearing magnetite. The main components of Ti-bearing blast furnace slag are SiO2, CaO, MgO, and Al2O3, which are also important components of silicate-based glass ceramics [10,11,12]. In addition, a small amount of TiO2 can be a good crystal-nucleating agent, which can promote nucleation and crystallization [13, 14]. In this work, the low Ti-bearing blast furnace slag was applied to prepare glass ceramics, which includes approximately 6% TiO2. This kind of low content TiO2 can be used as nucleating agent for the preparation of glass ceramics, and no additional nucleating agent is required. In addition, compared to steel slag, low Ti-bearing blast furnace slag has a lower alkalinity, which is closer to the alkalinity of silicate-based glass ceramics. Therefore, the low Ti-bearing blast furnace slag has great potentials to prepare slag-based glass ceramic.

The performance of glass ceramics largely depends on crystallization behaviors of basic glass [6, 15]. The crystallization behaviors of glass are mainly affected by nucleation and crystal growth, and the main influencing factors include nucleation agents, components, and heat treatment process [16,17,18,19]. According to previous reports, SiO2 was the main oxide in silicate glass, and its structure in glass had a decisive influence on the properties of silicate glass [20,21,22]. CaO existed in silicate network as a network modifier, which can polarize bridging oxygen and weaken Si–O bond [23, 24]. SiO2 and CaO were as the main components of silicate-based glass ceramics, and played important roles in the crystallization process of glass ceramics [24,25,26,27]. Li et al. [28] reported that the microstructure, mechanical, thermal, and electrical properties of BaO-Al2O3-B2O3-SiO2 glass ceramics were improved, and the flexural strength and the Young’s modulus of glass ceramics remarkably increased from 67 to 150 MPa and 45 to 66 GPa, respectively, with the increasing SiO2 content. Zhou et al. [29] investigated the effect of CaO/SiO2 ratio on the microstructures, electrical properties, and mechanical characteristics of the CaO-B2O3-SiO2 glass ceramics which synthesized by sol–gel method, and indicated that the increase of CaO content promoted the crystallization and increased the flexural strength. Hou et al. [30] studied the effect of different CaO/SiO2 ratio and heat treatment parameters on the main crystalline phase of SiO2-CaO-Al2O3-Na2O glass ceramics, and showed that the existence of nepheline had enhanced the mechanical properties of glass ceramics. Fu et al. [31] optimized CaO/SiO2 ratio and B2O3 content of CaO-B2O3-SiO2 glass ceramics, and reported that the crystallization ability of glasses increased with the increase of B2O3 and CaO/SiO2 ratio, which resulted in an increased permittivity of glass ceramics. Besides, the influence of CaO/SiO2 ratio on the preparation of glass ceramics from traditional blast furnace slag has been investigated, including crystallization ability, crystal phase, and mechanical properties [10, 32, 33]. However, the effect of CaO/SiO2 ratio on the properties of Ti-bearing blast furnace slag-based glass ceramics has been less reported.

In our previous work [34], the influence of TiO2 on the crystallization behavior of Ti-bearing blast furnace slag-based glass ceramics had been investigated and showed that the good performance and crystallization behaviors of glass–ceramic were obtained when the content of TiO2 was about 4%. To further explore the influence of CaO/SiO2 mass ratio on the Ti-bearing blast furnace slag-based glass ceramics, a series of Ti-bearing blast furnace slag-based glass ceramics with varying CaO/SiO2 mass ratio were prepared. Additionally, the variation range of CaO/SiO2 ratio were also considered by that the content of TiO2 can be maintained at approximately 4%. Moreover, the crystallization behaviors were analyzed combining with differential scanning calorimetry (DSC), X-ray diffraction (XRD), and scanning electron microscopy (SEM). The mechanical properties and acid and alkali resistance of the prepared glass ceramics were analyzed. This work can not only provide a theoretical reference for controlling crystallization and preparing slag-based glass–ceramic, but also provide experimental technical support for the effective resource utilization of Ti-bearing blast furnace slag.

Experimental procedures

Sample preparation

Ti-bearing blast furnace slag was obtained from an iron and steel plant in China, and the chemical compositions were characterized by X-ray fluorescence spectroscopy (XRF), as showed in Table 1. The CaO/SiO2 ratio of the original Ti-bearing blast furnace slag was approximately 1.1, and the chemical reagent SiO2 was added to adjust the R (CaO/SiO2) of specimens to be 0.3, 0.4, 0.5, and 0.6, respectively. The chemical compositions of the samples are showed in Table 2.

Table 1 Chemical composition of Ti-bearing blast furnace slag (wt%)
Full size table
Table 2 Chemical composition of basic glass (wt%)
Full size table

The prepared specimens were ground in an agate mortar for 0.5 h after accurately weighing to mix the various components uniformly. Then, the powder specimens were placed into a corundum crucible, and the temperature was raised from room temperature to 1500 °C in air holding 4 h. After that, the specimens were removed and quenched with water to obtain the basic glass specimens, which were identified by XRD, as showed in Fig. 1.

Fig. 1
figure 1

XRD patterns of basic glasses

Full size image

After grounding the basic glasses into powder with 200 mesh, 1% polyvinyl alcohol and 5% zinc stearate as binders were added. The FYD-30 electric powder compactor was used to press glass powders into the cylinder with a diameter of 8 mm for Vickers hardness detection. The cuboids with 50 × 6 × 6 mm were also pressed for flexural strength detection. The pressed specimens were placed on a platinum sheet and heated to 600 °C for 1 h in a high-temperature tube furnace to reduce the glass stress and remove the binders. According to the results of DSC, the transition temperature (Tg) of the prepared glasses was determined at the range of 740–760 ℃, and then the nucleation temperature was estimated by 800 ℃, because the optimum nucleation temperature usually lie at 40–80 ℃ above Tg of glasses [35]. Then, the specimens were hold in its crystallization temperature for 1.5 h to obtain the glass ceramics. The specific flowchart is showed in Fig. 2 [36].

Fig. 2
figure 2

The specific preparation process of Ti-bearing blast furnace slag-based glass ceramics [36]

Full size image

Analysis methods

X-ray fluorescence spectroscopy (XRF) (XRF-1800, Shimadzu, Japan) was used to obtain the chemical composition of Ti-bearing blast furnace slag. The basic glass samples were examined by DSC (449F3, NETZSCH, Germany) in air from room temperature to 1300 °C at heating rate of 5 °C/min, to obtain the transition temperature and crystallization temperature. The Fourier transform infrared spectroscopy (FT-IR) (iS50, Thermo Fisher, USA), with a spectral resolution of 4 cm−1, was recorded in the wavenumber range of 400–4000 cm−1 to analyze the microstructure of basic glasses. The phase of glass ceramics was identified by X-ray diffraction (XRD) (PANalytical X’Pert Powder, Spectris Pte, Netherlands). The scanning electron microscopy with an energy dispersive spectrometer (SEM–EDS) (JSM-7800F, JEOL, Japan) was used to observe the crystal phase and morphology of glass ceramics.

Regarding the physical performances of the prepared glass ceramics, the density was determined using Eq. (1), as follows:

$$\uprho =\frac{{m}_{0}}{h*\pi {*r}^{2}}$$
(1)

where ρ is the density of glass ceramics, \({m}_{0}\) is the mass of glass ceramics, \(h\) is the height of cylindrical sample, and \(r\) is the radius of cylindrical sample. The vickers hardness was measured by a microhardness-tester (MH-5L, Shanghai, China) with a measuring force of 300 N and a load time of 15 s. Besides, the electronic universal material testing machine (CMT4303, MTS, American) was used to detect the flexural strength of specimens with a span of 40 mm and the descent speed of 0.5 mm/min, and the flexural strength values were calculated by the equation according to GB/T 6569–2006, as shown as Eq. (2):

$${\sigma }_{f}=\frac{3FL}{2b{d}^{2}}$$
(2)

where \({\sigma }_{f}\) is the flexural strength (MPa), F is the maximum load (N), and L is the lower span of fixture(mm). Besides, b and d are the width and height of the sample (mm), respectively. Moreover, the surface and weight changes were observed after immersing the glass ceramics in 5% H2SO4 and 5% NaOH solutions for 24 h to evaluate the acid and alkali corrosion resistance of glass ceramics. The chemical stability was calculated by weighing the samples before corrosion (m0) and after corrosion (m1):

$$\mathrm{K}\left(\mathrm{\%}\right)=\frac{{m}_{1}}{{m}_{0}}\times 100$$
(3)

Results and discussion

Crystallization behaviors of Ti-bearing blast furnace slag-based glass ceramics

The DSC curves of Ti-bearing blast furnace slag-based glass samples with different CaO/SiO2 ratio at heating rate of 5℃/min are showed in Fig. 3(a). The transition temperature (Tg) and crystallization temperature (Tp) can be determined as the endothermic and exothermic peak temperature, respectively. The variations in transition and crystallization temperature of the prepared basic glass samples are showed in Fig. 3(b), demonstrating that the crystallization temperature obviously decreased with the increasing CaO/SiO2 ratio, and only negligible changes of Tg were detected. Besides, as the increase of CaO/SiO2 ratio, the crystallization peak intensity increased, indicating that the crystallization ability enhanced. Because Ca2+ ions has an effect on enriching oxygen, it accelerated to destroy the bridging oxygen bond of Si–O-Si units in the Si–O tetrahedron [32, 37], thereby creating more non-bridging oxygen bonds, reducing the degree of network polymerization, and decreased the crystallization activation energy, that resulted in stronger crystallization. This can be further confirmed from the FT-IR spectra of basic glasses, as showed in Fig. 4. According to previous studies [38,39,40], the FT-IR spectra curves can be divided into two domains: the 600–800, and 800–1200 cm−1 bands. The weaker band at 600–800 cm−1 was attributed to the bending vibrations of [AlO4]. The highest intensity bands in 800–1200 cm−1 range corresponded to the stretching vibrations of [SiO4] tetrahedral network, and the main peak shifted to low wave number with CaO/SiO2 ratios increasing. Moreover, it was observed that the intensity of the band centered at about 845 cm–1 gradually strengthen with the increase of CaO/SiO2 ratio, which was assigned to Q0 (Q is the tetrahedron with Si atom as the center and n is the amount of bridging oxygen in it) [38]. The specific schematic diagram of [SiO4] tetrahedral, [AlO4] tetrahedral, and Si–O-Si bridging oxygen is showed in Fig. 5. Si–O-Si means that the two [SiO4] structure units are connected by sharing one oxygen, which is here called a bridging oxygen. The oxygen with only one [SiO4] connected is called non-bridging oxygen. When four oxygens in one [SiO4] unit are non-bridging oxygen, this is called Q0 which is isolated island structure. The increase of Q0 indicated the decrease of polymerization degree. Therefore, the degree of polymerization of the investigated glasses decreased with increase of CaO/SiO2 ratio, so the atoms’ transfer for crystallization became easier, which resulted in a decreased crystallization temperature and an increased crystallization ability.

Fig. 3
figure 3

The result of DSC with different CaO/SiO2 mass ratios. (a) DSC curve of basic glass; (b) the changes of transition and crystallization temperature

Full size image
Fig. 4
figure 4

FTIR spectra of prepared glasses with different CaO/SiO2 ratios between 500 and 1400 cm−1

Full size image
Fig. 5
figure 5

Schematic diagram of [SiO4] tetrahedral, [AlO4] tetrahedral, and Si–O-Si bridging oxygen

Full size image

The crystal phases of Ti-bearing blast furnace slag-based glass ceramics were analyzed by combination with XRD and SEM–EDS. XRD patterns of the glass–ceramic samples with different CaO/SiO2 ratios are showed in Fig. 6. It can be observed that the main crystal phases transformed from CaAl2Si2O8 and CaMgSi2O6 to CaAl2Si2O8, CaMgSi2O6, and Ca2MgSi2O7 as the CaO/SiO2 ratio increased from 0.3 to 0.6. When CaO/SiO2 ratio was less than 0.5, the ratio of Ca and Si in the precipitated crystals was 1/2. Due to the increase of CaO/SiO2 ratio, the crystal phase of Ca2MgSi2O7 containing larger ratio of Ca and Si was precipitated at CaO/SiO2 = 0.5, and it increased as CaO/SiO2 ratio continued to increase. As seen from the XRD patterns, the crystals were still dominated by CaAl2Si2O8 as an increased CaO/SiO2 ratio.

Fig. 6
figure 6

XRD patterns of glass ceramics with different CaO/SiO2 mass ratios

Full size image

Moreover, to further explore the shape and distribution of crystals, the glass ceramics were polished and analyzed by SEM–EDS, which are exhibited in Fig. 7. It can be observed that when CaO/SiO2 ratio was 0.3, the massive crystals and fine needle-like crystals were observed in the SEM images, the needle-like crystals grown around the glass phase, and there were still a lot of glass phases, and the crystals distribution in the glass was not uniform. With the increase of CaO/SiO2 ratio, the glass phase decreased, the crystal phase and the glass phase were uniformly distributed, and the crystal shape changed to dendritic, the size decreased. When CaO/SiO2 ratio reached 0.6, the size of dendritic crystals in glass ceramics increased. Combining the DSC curve and FT-IR analysis results showed that when CaO/SiO2 ratio increased, the glass was easier to crystallize. Glasses have proper crystallization ability that can make the crystal size small and uniformly distributed in the glass ceramics, but when the crystallization ability is too strong, the crystals will grow excessively, which will increase the crystal size. In addition, it can be observed that the results of EDS point detections showed that the crystal phase in the glass ceramics was approximately to CaAl2Si2O8. Other phases were not identified which may be due to their low content and/or small size. The primary crystal phase is consistent with the XRD results.

Fig. 7
figure 7

SEM images of the polished slag-based glass ceramics. (a) S1, (b) S2, (c) S3, and (d) S4

Full size image

The crystallite phase in glass ceramics is as the framework, and the glass phase is as a continuous phase to bond the crystallite phases together. For the case of fine crystal grains, dense, and uniform distribution, it can further improve the overall performance of the material. Therefore, it can be regarded as a good crystal distribution state when the CaO/SiO2 ratio is 0.5.

Properties of Ti-bearing blast furnace slag-based glass ceramics

Density and vickers hardness analysis

The physical properties of the prepared glass ceramics are presented in Table 3, including density, vickers hardness, and flexural strength. The density of all specimens were in the range of 2.68–2.85 g/cm3. Compared with the original Ti-bearing blast furnace slag-based glass–ceramic (the density is 2.16 g/cm3), the density increased significantly with a decreased CaO/SiO2 ratio. As showed in Table 3, as increasing CaO/SiO2 ratio, the vickers hardness gradually decreased, since the glass ceramics with the CaO/SiO2 ratio of 0.3 had relatively few crystals, which showed the highest vickers hardness of 1045.10 MPa. According to the results of DSC, the crystallization ability increased with the increase of CaO/SiO2 ratio, but excessive crystallization ability may reduce the hardness of glass ceramics[25], and the grain growth may be another reason. According to the Hall–Petch formula, the relationship between hardness of a bulk material and grain size can be described by Eq. (4) [30, 41, 42]:

Table 3 The physical properties of Ti-bearing blast furnace slag-based glass ceramics
Full size table
$${H}_{g}={H}_{0}+k/\sqrt{D}$$
(4)

where \({H}_{g}\) is the hardness of grain interior;\(D\) is the grain size; \({H}_{0}\) is the intrinsic hardness of the grain and \(k\) is a constant. Equation (4) indicates that the increase of grain size will worsen the hardness of sample. Thus, with increasing CaO/SiO2 ratio, the crystallization ability increased and the crystal gradually grown up, so the vickers hardness decreased.

Flexural strength performance analysis

The effect of CaO/SiO2 ratio on flexural strength of the glass ceramics is showed in Fig. 8. The flexural strength first increased and then decreased as the CaO/SiO2 ratio increased from 0.3 to 1.1, and the maximum flexural strength value of 109.58 MPa was obtained at CaO/SiO2 = 0.5. Besides, compared with the glass–ceramic prepared by the original Ti-bearing blast furnace slag, decreasing the CaO/SiO2 ratio significantly improved the flexural strength of the glass–ceramic. According to the results of SEM and DSC, the crystallization ability increased with the increase of CaO/SiO2 ratio, and when the CaO/SiO2 ratio was 0.5, the largest number of small-sized crystals was formed in glass ceramics and was uniformly distributed, so the glass ceramics at this condition had excellent flexural performance. As the CaO/SiO2 ratio continued to increase, the size of crystals in the glass ceramics increased and the flexural strength decreased. It can be showed that within a certain range of crystallization ability, the increase of crystallization ability was helpful for improving the flexural strength, while too high crystallization ability was not conducive. This is consistent with the results of Hou et al. [30].

Fig. 8
figure 8

The change of the flexural strength of glass–ceramic with different CaO/SiO2 mass ratios

Full size image

Acid and alkali resistance performance analysis

The prepared glass ceramics were immersed in 5% H2SO4 and 5% NaOH solutions for 24 h. Figure 9 shows the change of the sample surface before and after acid and alkali corrosion test. It can be observed that the glass ceramics showed a certain amount of fading after soaking in 5% H2SO4 solution for 24 h, especially at high alkalinity, indicating that the acid resistance performance was slightly weakened as an increased CaO/SiO2 ratio. However, no obvious changes were observed on the surface of samples after soaking in 5% NaOH solution for 24 h. The weight changes of the glass–ceramic samples were calculated to evaluate the acid and alkali resistance performance, listed in Table 4. Compared with the glass ceramics prepared from the original slag, reducing the ratio of CaO/SiO2 had a certain positive effect on the acid and alkali resistance of glass ceramics. Therefore, it can be concluded that all of the prepared glass ceramics showed good acid and alkali resistance, especially the alkali resistance, all over 99.9%.

Fig. 9
figure 9

The surface condition of the glass ceramics before and after the acid and alkali resistance test

Full size image
Table 4 The acid and alkali resistance performance of the glass ceramics with different ratios of CaO/SiO2
Full size table

The results of flexural strength, hardness, and corrosion have been compared to other glass ceramics prepared by blast furnace slag [43,44,45,46]. The prepared glass ceramics showed excellent properties, such as flexural strength, acid resistance, and alkali resistance for the samples S2 and S3 (Table 5). The flexural strength of the studied titanium-containing blast furnace slag glass ceramics can reach up to 109.58 MPa at CaO/SiO2 = 0.5. The decrease of CaO/SiO2 from 1.1 to 0.5, and the flexural strength significantly increased from 26.09 to 109.58 MPa. On the whole, the flexural strength of the prepared glass ceramics is slightly higher than that of most other blast furnace slags. Although the flexural strength was slightly lower than that of the glass ceramics prepared by Deng et al. [45], the current glass ceramics exhibited better resistance to acid and alkali corrosion.

Table 5 Comparison of main properties of glass ceramics prepared with better properties with those of others’ works
Full size table

Conclusions

In this work, a series glass ceramics were successfully prepared from Ti-bearing blast furnace slag. The effect of CaO/SiO2 ratio on the crystallization behaviors, mechanical properties, and acid and alkali resistance properties of the prepared Ti-bearing blast furnace slag-based glass ceramics were investigated. The following conclusions can be drawn:

  1. (1)

    The crystallization temperature obviously decreased with the increase of CaO/SiO2 ratio, and the higher CaO/SiO2 ratios leaded to stronger crystallization ability.

  2. (2)

    The main crystal phases transformed from CaAl2Si2O8 and CaMgSi2O6 to CaAl2Si2O8, CaMgSi2O6, and Ca2MgSi2O7 as the CaO/SiO2 ratio increased from 0.3 to 0.6. When the CaO/SiO2 increased to 0.5, a small amount of akermanite precipitated.

  3. (3)

    The vickers hardness gradually decreased, and the flexural strength first increased and then decreased with an increased CaO/SiO2. The glass ceramic with CaO/SiO2 = 0.5 exhibited the highest flexural strength of 109.58 MPa. The prepared glass ceramics showed good acid and alkali resistance (> 98.30%), especially alkali resistance. Therefore, the best candidate for CaO/SiO2 ratio in the investigated Ti-bearing blast furnace slag-based glass ceramics was selected as 0.5.