Introduction

The northwest region of China is rich in coal resources, and the phenomenon of spontaneous combustion is quite common due to factors such as the degree of coal metamorphism, moisture content, oxygen enrichment, coal heat accumulation, and coal composition1,2,3,4. Overlying strata in coal mines often consist of rock layers and loess layers. The stability of the strata, lithology of the rock layers, physicochemical properties of the loess, and the porosity structure of the loess will change with the variation in temperatures caused by coal spontaneous combustion5. Although previous studies mainly focused on the impact of combustion temperature on the overlying rock layers, there are still deficiencies in terms of the physical and chemical properties of the overlying loess layer and the engineering application transformation of the results.

In terms of physicochemical properties, factors such as the strength, resistivity, wave velocity, and porosity structure of the soil are particularly important in experiments. In heating experiments of loess, the color of the loess surface changes to varying degrees, which is why colorimetric measurements are introduced in the experiments to quantify the color information of the loess surface6,7,8,9. The strength indicators of soil include tensile strength, compressive strength, and shear strength, with tensile strength being relatively low. Brazilian splitting tests, known for their simplicity and low equipment requirements, are widely used to determine tensile strength10,11,12. During mechanical loading, acoustic emission monitoring can be employed to observe crack development in real time during the loading process13,14,15,16. Resistivity is mainly influenced by the moisture content in the loess, the conductivity of its constituent materials, and the test frequency. A general trend is observed where higher frequencies, stronger conductivity of the constituents, and higher moisture content lead to lower resistivity17,18,19. Electrical methods are widely used in fields such as mineral exploration, hydrogeology, and engineering geology for this reason. Wave velocity is primarily related to the composition of the material. Differences in wave velocity conduction paths, such as those caused by cracks in the material, can result in abrupt changes in wave velocity. This principle is also widely applied in engineering geology and mineral development20,21. The strength, resistivity, and wave velocity of the soil are closely linked to its porosity structure. Current mainstream methods for determining the porosity structure of materials include mercury intrusion, nuclear magnetic resonance, and nitrogen adsorption. In nitrogen adsorption experiments, pores are classified into macropores (r > 50 nm), mesopores (50 nm > r > 2 nm), and micropores (r < 2 nm) based on their size22,23. The volume distribution of different pore types also contributes to variations in the physicochemical properties of loess. The research content of the aforementioned researchers is shown in Table 1.

Table 1 Study of high-temperature loess.
Full size table

In this study, loess samples were heated to 200 °C, 400 °C, 600 °C, 800 °C, and 1000 °C, and subsequent tests were conducted to measure their mechanical properties, wave velocity, resistivity, colorimetric properties, and porosity. These findings will contribute to improving the understanding of the physicochemical properties and the mechanisms of property changes in loess within the temperature range of 25–1000 °C. And the research results have been applied in engineering in a way that is of guiding significance.

Experiment overview

Sample preparation

The loess samples were collected from the Q2 layer in Xi’an City, Shaanxi Province, which is 12 m below the surface. The moisture content was 11.5%, the dry density was 1.39 g/cm3, the cohesion was 8.6 kPa, and the internal friction angle was 26.10°. First, the loess was dried at 105 °C and then cooled to room temperature. It was then sieved and ground into powder. Water was added to the mixture, and the initial moisture content of the loess was adjusted to 20%. The mixture was sealed in a bag and left to stand for 24 h. Using a hydraulic jack, the powdered loess was compressed in the mold to a fixed scale of 20 mm, forming a cylindrical loess sample with a diameter × height of 61.8 × 20 mm (ASTM2009). The dry density of the sample was measured to be approximately 1.40 g/cm3. Finally, the loess sample was placed in the SDTGA200 box furnace and heated at a heating rate of 5 °C per minute to the following temperatures: 200 °C, 400 °C, 600 °C, 800 °C, and 1000 °C.

Test equipment

The equipment and instruments used in this experiment include the JW-BK112W specific surface area and pore size analyzer, the DS5-8A full-information acoustic emission instrument, the RSM-SY6(C) non-metallic ultrasonic testing device, the NR10QC colorimeter, the TH2816A LCR digital bridge meter, and the SHT4605 microcomputer-controlled electro-hydraulic servo universal testing machine.

Test procedure

After treating these samples in high-temperature environments of 200 °C, 400 °C, 600 °C, 800 °C and 1000 °C, their resistivity, wave velocity, chromaticity and porosity (porosity volume) were measured in a test room with a temperature of 20 °C and an environmental humidity of 55%. After completing these tests, the samples were placed on a Brazilian splitting testing machine and loaded at a rate of 0.05 mm/min24. During the loading process, mechanical data such as force–displacement-time were collected. The acoustic emission probe was adhered to the loading base in contact with the sample using hot melt adhesive. These probes were used to capture acoustic emission signals during the sample loading process. The experimental procedure is shown in Fig. 1.

Fig. 1
Fig. 1
Full size image

Experimental flow chart.

Results

Effects of different high-temperature treatments on the mechanical and acoustic emission characteristics of loess

In the initial stage of loading (the first 20 s for temperatures between 25–400 °C and the first 200 s for 800–1000 °C), as stress was applied, the compaction of the loess matrix reduced the pore spaces, triggering the first acoustic emissions in Fig. 2. During the later stage of loading, as the stress reached its peak, the acoustic signals surged sharply, reaching their maximum values. Significant cracks appeared at the top of the sample and nearly propagated along its diameter. At 25 °C and 200 °C, the ringing counts remained below 200; however, once the temperature exceeded 200 °C, a dramatic increase in ringing counts occurred. The peak values at 400 °C, 800 °C, and 1000 °C reached 68,690, 582,427, and 756,363, respectively. In general, high ringing counts were mainly observed during the failure stage of the sample, while little or no activity was recorded during the compaction phase.

Fig. 2
Fig. 2
Full size image

Mechanical characteristics and acoustic emission signals of loess after high-temperature treatment.

After the loess samples were subjected to high-temperature treatments, their tensile strength was significantly enhanced in Fig. 3. Notably, the tensile strength increased dramatically from 0.0084 MPa at 25 °C to 0.195 MPa at 200 °C, representing a 2221% increase. At 400 °C and 800 °C, the increase in tensile strength was relatively slower, with values of 0.345 MPa, 0.442 MPa, and 0.487 MPa, respectively. The time required to reach the peak load also increased with the rise in heating temperature, showing a positive correlation with tensile strength. The elastic modulus of the sample also significantly increased with the rise in heating temperature. Especially at 1000℃, the elastic modulus rose from 66.2 MPa at 800℃ to 234.5 MPa.

Fig. 3
Fig. 3
Full size image

Stress and axial displacement curves of loess after high-temperature treatment.

Effect of different high-temperature treatments on the wave velocity of loess

The wave velocity of loess initially decreases and then increases with the rise in heating temperature. Between 200 °C and 600 °C, the wave velocity decreases, while from 600 °C to 1000 °C, the wave velocity increases. The relationship between wave velocity and temperature variation is shown in Fig. 4.

Fig. 4
Fig. 4
Full size image

Variation of wave velocity with temperature after high-temperature treatment of loess.

Effect of different high-temperature treatments on the resistivity of loess

Figure 5 illustrates the resistivity of loess after different high-temperature treatments. A clear positive correlation exists between loess resistivity and frequency. Under the same temperature conditions, an increase in test frequency results in a decrease in loess resistivity. When the frequency exceeds 1 kHz, the resistivity decreases significantly. At test frequencies of 50 Hz and 1 kHz, resistivity decreases with increasing temperature, showing reductions of − 82.9% and − 77.9%, respectively. However, at test frequencies of 10 kHz, 100 kHz, and 200 kHz, resistivity generally increases with rising temperature, with increases of 1698.49%, 3739.77%, and 1132.80%, respectively.

Fig. 5
Fig. 5
Full size image

Loess resistivity at different temperatures and frequencies.

Effect of different high-temperature treatments on the porosity of loess

According to the International Union of Pure and Applied Chemistry (IUPAC) classification, pores can be divided into three types based on pore size: micropores (r < 2 nm), mesopores (2 nm < r < 50 nm), and macropores (r > 50 nm)23. As shown in the pie chart in Fig. 6, at 25 °C, the loess pores are primarily micropores. As the temperature increases, the proportion of micropores gradually decreases, from 46% at 400 °C to 34.6% at 600 °C, 10.4% at 800 °C, and 2.4% at 1000 °C. Conversely, the volume fraction of mesopores increases, rising from 42.4% at 25 °C to 50.9% at 400 °C, 65.4% at 600 °C, and 89.6% at 800 °C. At 1000 °C, the mesopore volume fraction slightly decreases, showing a 4.5% reduction compared to 800 °C. Macropores are somewhat developed at 400 °C and 1000 °C.

Fig. 6
Fig. 6
Full size image

Proportions of different pore types in loess at different temperatures and the corresponding pore volume for each pore size.

Effect of different high-temperature treatments on the colorimetric properties of loess

The color distribution of loess at different temperatures is shown in Fig. 7. Here, L* represents brightness (ΔL* > 0, indicating a whitish tone; L* < 0, indicating a blackish tone), a* represents the red/green value (a* > 0, indicating a reddish tone; a* < 0, indicating a greenish tone), and b* represents the yellow/blue value (b* > 0, indicating a yellowish tone; b* < 0, indicating a bluish tone)25,26,27. The brightness of loess increases gradually with temperature up to 600 °C, after which it decreases. The loess turns progressively red with increasing temperature up to 800 °C, with a particularly noticeable increase in redness between 600 °C and 800 °C. At 800–1000 °C, the redness (a*) decreases. At 25 °C, prior to any high-temperature treatment, loess exhibits a b* value < 0, indicating a bluish tone. As the temperature increases, the b* value gradually rises; after 600 °C, b* decreases and stabilizes. ΔE* represents the total color difference of loess, which can be expressed using formula (1). The total color difference increases with temperature, peaking at 600 °C before decreasing.

$$\Delta E^{*} = \sqrt {\left( {L^{*} - L_{{0}} } \right)^{2} + \left( {{\text{a}}^{*} - {\text{a}}_{{0}} } \right)^{2} + \left( {b^{*} - {\text{b}}_{{0}} } \right)^{2} }$$
(1)
Fig. 7
Fig. 7
Full size image

Color characteristics of loess at different temperatures.

Here, L0 represents the brightness of the sample at 25 °C, a0 represents the initial red/green value at 25 °C, and b0 represents the initial yellow/blue value at 25 °C.

Discussion

The mineral composition of loess mainly consists of quartz (over 50% of the total composition), feldspar (20–30%), along with trace amounts of mica, carbonate minerals, iron-bearing minerals, and clay minerals. With the increase in heating temperature, both the mineral composition and the structural arrangement of the loess undergo changes.

In the early stage of heating, specifically between 25 °C and 100 °C, the main processes involve the evaporation of free water, adsorbed water, and some bound water within the loess pores (Eq. 2). As the temperature increases further, the crystallization water in the minerals begins to dehydrate. Before 570 °C, the main component of loess, α-quartz, gradually expands with heat. At 570 °C, it undergoes a phase transformation to β-quartz (Eq. 3)28. During this process, between 470 °C and 540 °C, kaolinite transforms into metakaolinite. The dehydroxylation of kaolinite produces Al2O3 and SiO2, which intensifies the vitrification and cementation of the loess’s pore structure (Eq. 4). This is one of the primary reasons why the proportion of micropores gradually decreases with increasing temperature. The gradual closure of the pore structure also contributes to the enhancement of loess strength29. Between 400 and 600 °C, thermal cracks appear on the surface of the loess, which is consistent with previous experimental results. The formation of these cracks allows more air to enter the sample, thereby affecting the decrease in wave velocity. When the crack development rate exceeds the dehydroxylation rate of kaolinite, it leads to a reduction in wave velocity. Subsequently, the agglutination of clay minerals and the formation of iron-bearing mineral oxidation products result in lower porosity in the loess. These factors contribute to the increase in wave velocity after 600 °C30. Between 600 and 900 °C, CaCO3 in the loess decomposes into CaO and CO2. During the resistivity measurements, the primary factors influencing resistivity are pore water and the polarization effects. As the temperature rises and the pore water content decreases, the influence on resistivity shifts to the polarization effect. Polarization is closely related to frequency; at lower frequencies, more types of polarization occur. This follows the trend that resistivity increases as frequency decreases. At frequencies of 10 kHz, 100 kHz, and 200 kHz, the polarization effect disappears significantly, allowing resistivity to better reflect the presence of pore water and mineral crystallization water in the loess samples. The decrease in pore water and crystallization water content leads to an increase in resistivity31. In addition, CaO reacts with water vapor in the air to form Ca(OH)2. When low-frequency alternating current passes through the loess sample, it dissociates into Ca2⁺ and OH−, and under the influence of the polarization effect, a decrease in resistivity is observed.

$${\text{H}}_{2} {\text{O}}\mathop = \limits^{{25-100\;^{ \circ } {\text{C}}}} {\text{H}}_{2} {\text{O}} \uparrow$$
(2)
$$a{\text{-}}quartz\to ^{{570\;^{ \circ } {\text{C}}}} b{\text{-}}quartz$$
(3)
$${\text{Al}}_{{4}} \left[ {{\text{Si}}_{{4}} {\text{O}}_{{{10}}} } \right]\left( {{\text{OH}}} \right)_{8} \to ^{{470-540\;^{ \circ } {\text{C}}}} {\text{Al}}_{2} {\text{O}}_{{3}} + 4{\text{SiO}}_{2} + 4{\text{H}}_{2} {\text{O}}$$
(4)

Among the many iron-bearing minerals, the primary minerals influencing the colorimetric properties are ferrous oxide (FeO), goethite (FeOOH), pyrite (FeS2), marcasite (Fe1−xS), limonite (Fe2O3·nH2O), hematite (Fe2O3), and magnetite (Fe3O4). These minerals exhibit colors such as black, yellow–brown, gray-yellow, copper-yellow with a reddish tint, brown, reddish, and iron-black, respectively. The most noticeable effect is reflected in the a* value, which represents the red/green component. As the heating temperature increases, the a* value of loess first increases and then decreases. The primary factor influencing the a* value is the variation in the proportion of iron-bearing minerals. Below 200–300 °C, the main iron-bearing minerals in loess are FeO and small amounts of FeOOH. As the temperature rises, FeO and FeOOH gradually react to form Fe2O3 (Eqs. 5–6), leading to an increase in the red component of the a* value. Since the sample does not experience uniform heating during the experiment, the reactions described by Eqs. 5 and 6 continue as the temperature increases. According to the a* results, this reaction likely persists until around 800 °C, where the Fe2O3 content reaches its peak. Between 800 °C and 1000 °C, Fe2O3 further reacts with oxygen (O2) in the air to form Fe3O4, which gives rise to a dark color. At 950–1000 °C, Fe1−xS reacts with O2 and hematite to form magnetite. This is reflected in the a* and L* values, showing a decrease in red, brightness, and yellow values (Eqs. 7–9)28,32. As a major component of the loess skeleton, the process from FeO to Fe2O3 and then to the dense Fe3O4 strengthens the loess, further improving its mechanical properties.

Generally speaking, an increase in porosity leads to a decrease in wave velocity. However, during the high-temperature stage (470—540 °C), a fundamental change occurred on the surface of the sample. On one hand, the glassification of clay minerals and the oxidation/re-crystallization of iron oxides formed rigid cementing bridges between the particles, significantly enhancing the elastic modulus of the solid matrix. At 600 °C, due to heat conduction within the sample, a phase transition also occurred internally, starting the glassification process33,34,35. The pores were filled with new cementing substances and iron oxides between the pores, resulting in a decrease in porosity. This is consistent with the result of the reduced pore volume shown in Fig. 5. On the other hand, the pores measured by nitrogen adsorption are mainly mesopores and macropores. Under the influence of high-temperature cementation, the shapes of these pores may tend to be more spherical and their connectivity deteriorates, thereby reducing the scattering effect on the propagation of elastic waves. The extreme enhancement of the solid skeleton stiffness becomes the dominant factor controlling the wave velocity, leading to the observed recovery of the wave velocity. We have supplemented this mechanism in the “Discussion” section and cited relevant literature to strengthen the argument.

$$4{\text{FeO}} + {\text{O}}_{2} \to ^{\Delta } 2{\text{Fe}}_{2} {\text{O}}_{3}$$
(5)
$$2{\text{FeOOH}}\to ^{\Delta } {\text{Fe}}_{{2}} {\text{O}}_{{3}} + {\text{H}}_{2} {\text{O}}$$
(6)
$$6{\text{Fe}}_{2} {\text{O}}_{3} + 2{\text{O}}_{{2}} \to ^{{950-1000\;^{ \circ } {\text{C}}}} 4{\text{Fe}}_{{3}} {\text{O}}_{{4}} + 3{\text{O}}_{{2}}$$
(7)
$$3{\text{Fe}}\left( {1 - x} \right){\text{S}} + (5 - 2x){\text{O}}_{{2}} \to ^{{950-1000\;^{ \circ } {\text{C}}}} (1 - x){\text{Fe}}_{{3}} {\text{O}}_{{4}} + 3{\text{SO}}_{{2}}$$
(8)
$$7{\text{Fe}}\left( {1 - x} \right){\text{S}} + (10 - 4x){\text{Fe}}_{{2}} {\text{O}}_{{3}} + (10 - 4x){\text{O}}_{{2}} \to ^{{950-1000\;^{ \circ } {\text{C}}}} (9 - 5x){\text{Fe}}_{{3}} {\text{O}}_{4} + 7{\text{SO}}_{{2}}$$
(9)

Engineering applications and prospects

Based on the research results, we can establish a temperature threshold monitoring and early warning system. Distributed temperature sensors are buried in the overlying loess in the burned area, with a focus on the critical temperature threshold of 600°, to achieve the stage identification and early warning of thermal damage. Additionally, a portable colorimeter can be used to measure the color of the surface or drill core samples. A significant increase in the a* value can be used to determine whether the temperature has entered the high-temperature stage (> 600 °C), assisting in the assessment of the degree of spontaneous combustion. High-frequency electrical methods (> 10 kHz) can also be used for ground or borehole detection. If the resistivity significantly increases locally, it may indicate that the area has entered a high-temperature and low-water content state, suggesting an active area of spontaneous combustion. For the loess layers that have entered the high-temperature stage (> 800 °C), although the strength is high, the brittleness increases. It is recommended to use low-temperature grouting (such as silicate-based cementitious materials) for crack filling and cementation reinforcement to prevent sudden brittle cracking. Thermal isolation engineering can be implemented at the boundary of the spontaneous combustion area, such as injecting gel foam or laying insulation boards, to inhibit the transfer of heat to the surrounding soil layers and control the spread of spontaneous combustion. We can also integrate multi-source data such as wave velocity, resistivity, and color to establish a ‘temperature-property’ corresponding database for loess in the burned area, and develop a dynamic evaluation model for the stability of strata based on property changes, providing decision support for surface subsidence and ecological restoration after mine closure.

The heating process in this study was conducted under laboratory conditions of uniform temperature, absence of external loads and no seepage water, which are ideal for uniform temperature distribution and stable conditions. However, spontaneous combustion of coal seams in the field is often accompanied by complex factors such as uneven temperature distribution, large thermal gradients, fluctuating oxygen supply, participation of groundwater, and the influence of overlying stratum stress. These differences may lead to variations in the degree and uniformity of mineral phase changes. Inhomogeneous heating in the field may result in incomplete mineral transformation in certain areas, forming mixed mineral zones, which can affect the overall mechanical behavior. Additionally, the coupling of thermal stress and stratum stress in the field may cause fractures to expand along the dominant direction, thereby altering the anisotropy of permeability and wave velocity. The presence of groundwater may accelerate or inhibit certain high-temperature reactions (such as the hydration of CaO), and affect the evolution path of resistivity and pore structure. Therefore, although this study has revealed the basic laws of the evolution of yellow soil properties under high temperatures, its quantitative results need to be comprehensively corrected and scaled down when applied to actual coal seam spontaneous combustion areas. Future research can use large-scale physical model experiments or field-in-situ coupled simulations to further approach the coupled process of heat, force, and chemical reactions in the real environment.We will consider adding in-situ monitoring of the high-temperature phase transition process and quantitative analysis of minerals during the experiment to further optimize the experiment.

Conclusion

In this study, loess samples were heated at 200 °C, 400 °C, 600 °C, 800 °C, and 1000 °C, and their wave velocity, resistivity, colorimetric properties, nitrogen adsorption, and Brazilian splitting tests were conducted. Acoustic emission monitoring was also performed during mechanical loading. The main conclusions are as follows:

  1. 1.

    The primary iron-bearing minerals in loess are affected by heating temperature. Below 300 °C, FeO gradually transforms into Fe2O3 with increasing temperature. After 800 °C, Fe2O3 further transforms into Fe3O4, thereby influencing the color characteristics.

  2. 2.

    After heating to 470 °C, the boundaries between clay minerals in loess become blurred, gradually converting into a glassy cementing material, which leads to pore filling, especially the reduction in micropore volume.

  3. 3.

    As the heating temperature increases, the loess strength is significantly enhanced due to changes in pore structure and the formation of new products, particularly between 800 and 1000 °C.

  4. 4.

    The resistivity of loess decreases with increasing temperature at low frequencies, such as 50 Hz and 1 kHz. However, at higher frequencies (10 kHz, 100 kHz, and 200 kHz), the resistivity increases as the temperature rises.