A Critical Review of the Physical Properties and Geotechnical Behaviors of Tailing Materials

Abstract

The stability of tailings dams is governed predominantly by the physical properties and geotechnical behavior of their primary construction material—tailings. Consequently, a systematic understanding of these characteristics is of great significance for the rational design and long-term stable operation of tailings dams. This review focuses on the physical properties and geotechnical behavior observed in different types of tailings. In terms of physical properties, the particle size distribution exhibits a pronounced hydraulic classification characteristic within the impoundment, consisting predominantly of silt-sized particles and displaying an overall trend toward finer gradation. The mineralogical and chemical composition is dominated by quartz, hematite, and silicates. However, significant spatial variability exists both between different tailings types and across distinct zones within the same tailings pond. Regarding geotechnical behavior, the permeability of tailings is governed by a fines content threshold: below this threshold, permeability decreases with increasing fines content, while beyond it, the permeability stabilizes. When studying consolidation and compression behavior using slurry specimens, the compression curves exhibit nonlinear characteristics, primarily described by the modified Gibson theory. The shear behavior of tailings is significantly influenced by confining pressure, drainage conditions, anisotropy and stress paths. The presence of transitional behavior leads to the critical state line determined based on a single sampling method erroneously assessing the dilation/cosntraction characteristics of in situ tailings, thereby affecting the assessment of liquefaction risk. Future research should focus on the seepage, consolidation and shear properties of clayey fine-grained tailings and unsaturated tailings, and aim to elucidate the key controlling factors of transitional behavior to enhance the reliability of tailings dam stability assessments.

1. Introduction

Tailings refer to the residual solid waste produced after target metals have been extracted from metal ores using mechanical and chemical methods [1]. With the continued global growth in demand for mineral and metal resources, the volume of tailings generated has increased rapidly. Currently, global annual tailings production stands at approximately 13 billion tones [2,3]. These tailings are typically rich in heavy metals and other potentially harmful substances. If managed improperly, they may pose a serious threat to the ecological environment and human health [4], including groundwater contamination [5], airborne dust [6], soil degradation [7], the spread of heavy metals [8,9], and disasters such as tailings dam failures [10], as illustrated in Figure 1. Consequently, the safe storage and disposal of tailings have become key concerns in the fields of mining engineering and environmental protection. Currently, various methods have been proposed for the treatment and disposal of tailings, such as backfilling underground mines, discharging into lakes or the sea, and dry stacking following dewatering [11,12,13,14]. Among these, the most common method involves conveying tailings in slurry form via pipelines to a tailings pond, where they gradually settle and consolidate under their own weight, ultimately forming a tailings dam. However, tailings dam accidents continue to occur from time to time worldwide. Following a systematic study of 93 types of accidents and public hazards worldwide, researchers at Clark University in the United States found that the severity of tailings dam accidents ranked 18th, behind only nuclear radiation and nuclear explosions [15]. From an engineering perspective, such accidents are primarily linked to the stability of tailings dams during operation and after closure [16], and dam stability is governed predominantly by the physical properties and geotechnical behavior of the primary construction material—tailings. Recent catastrophic tailings dam failures, such as the Mariana (2015) and Brumadinho (2019) disasters in Brazil, have further demonstrated the critical importance of understanding the geotechnical behavior of tailings materials. Post-failure investigations indicated that inadequate drainage conditions, excess pore water pressure buildup, and static liquefaction susceptibility were among the major technical factors contributing to dam instability and failure [17]. These failure mechanisms are intrinsically associated with the permeability behavior, consolidation characteristics, and shear response of tailings under different stress conditions. Therefore, a systematic understanding of the physical properties of tailings and their geotechnical behavior is of great significance for the rational design and long-term stable operation of tailings dams.

In recent years, numerous researchers have conducted extensive studies on the geomechanics of various tailings, achieving significant results. Qiu and Sego [18] systematically investigated the geotechnical properties of copper, gold, coal and oil sands tailings through laboratory testing, highlighting the close relationship between the particle size distribution, permeability and compressibility of tailings, and emphasizing the potential threat to tailings dam stability posed by rising pore water pressure. Wijewickreme et al. [19,20] conducted an in-depth investigation into the mechanical response of fine-grained tailings and mixtures of tailings and waste rock under cyclic loading. They elucidated the strain mechanism of ‘cyclic fluidity’ and demonstrated that liquefaction resistance is influenced by mineralogical composition, void ratio (e) and confining pressure, whilst highlighting the limitations of traditional empirical guidelines in assessing the liquefaction susceptibility of tailings. Wong et al. [21] proposed a one-dimensional consolidation mechanism model for non-separated oil sands tailings (NST) based on mixture theory, providing a theoretical basis for the optimized design of tailings. Schnaid et al. [22] established a highly non-linear critical state line (CSL) through a series of triaxial tests on gold tailings and found that the coupling of state parameters with the small strain height reflected by shear wave velocity can effectively evaluate the static liquefaction trigger conditions of tailings dams. Islam et al. [23] investigated the volume changes and strength evolution during the transition of coal tailings slurry from a slurry state to a soil state, identified the ‘solidification point’ as a key indicator for the transition between slurry and soil behavior, and established a power-law relationship between moisture content and shear strength. Delgado et al. [24] conducted systematic compression and shear path tests on iron tailings based on a critical state soil mechanics (CSSM) framework. They found that CSL is unique under both compression and extension paths, but that the failure paths exhibit significant differences, emphasizing the significant influence of stress paths on the liquefaction trigger behavior of tailings. Dias Neto et al. [25], combining unsaturated soil mechanics theory with numerical simulation methods, systematically assessed the effects of changes in e, material heterogeneity and insufficient compaction during the rainy season on the hydro-mechanical response of dry-stacked iron tailings. They revealed the critical role of compaction quality in maintaining matrix suction, limiting the rise in pore water pressure and ensuring slope stability.

The aforementioned studies indicate that research into the geotechnical behavior of tailings has gradually shifted from early descriptive understanding towards mechanistic explanations. In addition, practical tailings management not only depends on the effectiveness of stabilization and storage technologies, but also on their economic feasibility and engineering applicability. Yıldız et al. [26] pointed out that increasingly stringent requirements for waste characterization, classification, and storage may significantly increase the economic burden of tailings management and disposal. However, most existing studies focus on specific types of tailings or single mechanical properties, with research subjects encompassing a variety of tailings such as gold, copper, iron, coal and oil sands. Due to the complex and variable nature of factors such as particle size distribution [27,28,29], mineralogical composition and chemical composition of tailings [30,31,32], the conclusions drawn from different studies are often highly context-specific, and a unified theoretical framework to systematically explain their behavioral patterns has yet to be established. Therefore, establishing a scientifically reliable yet practically feasible understanding of tailings geotechnical behavior is essential for the development of safe, economical, and sustainable tailings storage strategies.

In view of this, this paper aims to systematically review and evaluate research progress in the field of geotechnical behavior of tailings. Through the synthesis and integration of existing research findings, it focuses on exploring the following core issues: (1) the physicochemical properties of tailings; (2) the geotechnical behavior of tailings, including permeability, consolidation, and shear behavior. On this basis, the paper summarizes the main areas of consensus and existing controversies in current research, identifies research gaps and future directions, and further attempts to establish a more unified understanding of the relationships among physicochemical properties, permeability behavior, consolidation characteristics, and shear response of tailings materials. In particular, the applicability of the CSSM framework to the interpretation of state-dependent tailings behavior is discussed. The findings of this review are expected to provide a systematic theoretical reference for the design, operation, and safety assessment of tailings dams, as well as useful guidance for future geotechnical research and engineering practice involving tailings materials.

2. Review Methodology

2.1. Review Type and Research Question

To ensure objectivity and comprehensiveness, this study employs a critical review methodology based on systematic reviews. This approach combines the structured and transparent literature search process of systematic reviews with the analytical depth of critical reviews to evaluate the geotechnical behavior of tailings materials. This review primarily addresses the following three interrelated research questions (RQs):

RQ1: What are the spatial heterogeneity characteristics and commonalities regarding particle size distribution, mineral composition and chemical composition among different types of tailings?

RQ2: What are the common mechanisms underlying the macroscopic geotechnical behavior (seepage, consolidation, shear) of different types of tailings? What are the applicability limits and theoretical shortcomings of classical geotechnical frameworks when applied to tailings?

RQ3: What key knowledge gaps remain in existing geotechnical studies of tailings? How should future research directions be planned?

2.2. Literature Search Strategy

A comprehensive and systematic literature search was conducted within the Web of Science (WoS) Core Collection database (including SCI-Expanded). The database search was completed and finalized on 20 March 2026. The search syntax employed Boolean operators: TS = (“tailings”) AND TS = (“grain size distribution” OR “mineral composition” OR “chemical composition” OR “permeability” OR “compression-consolidation behavior” OR “shear behavior”).

2.3. Screening Process

To ensure transparency in the screening process, this study adhered to a rigorous multi-stage screening protocol, with the specific workflow illustrated in Figure 2. Based on the aforementioned search terms, a total of 1898 initial records were retrieved. Subsequently, the research field was restricted to “engineering OR mining mineral processing OR construction building technology OR mineralogy”, leaving 1229 papers. The authors independently screened the titles and abstracts of these 1229 articles, excluding those focusing on non-core physical properties and geotechnical topics such as numerical simulation of tailings or environmental impact. On this basis, a full-text quality assessment was conducted on the remaining literature; studies were excluded if they met any of the following criteria: (1) failure to specify the type of tailings under investigation; (2) lack of core data required for this review. Ultimately, over 140 high-quality peer-reviewed papers were included as the core sample for systematic data extraction and comprehensive analysis.

2.4. Scope and Limitations

The tailings materials covered in this review mainly include three types, iron ore tailings, gold tailings, and copper tailings, along with a small number of studies on coal tailings and oil sands tailings. Although this review strives to provide a reliable comprehensive analysis, the following limitations remain: (1) the search scope was restricted to peer-reviewed English-language literature indexed in Web of Science, which may have omitted relevant regional industry standards or patents; (2) this review focuses strictly on the physical properties of tailings and their geotechnical behavior based on laboratory testing—topics such as numerical modelling and environmental impact assessments were not included within the scope of the study.

3. Physical and Chemical Properties of Tailings

3.1. Particle Size Distribution

The particle size distribution of tailings is a key factor influencing their geotechnical behavior. Figure 3 presents the particle size distribution curves for 31 tailings samples, comprising: (a) 14 iron tailings curves; (b) 9 gold tailings curves and (c) 8 copper tailings curves [16,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52]. Overall, the particle size composition of the tailings is dominated by silt and fine sand. As can be seen from Figure 3, even for tailings of the same type, there are certain variations in their particle size distribution. This is primarily related to the processing methods and grinding techniques employed at the mine, as well as the hardness of the constituent minerals themselves [27]. Furthermore, it can be observed from Figure 3b that even samples collected from different locations within the same tailings pond (such as the upper beach (UB), middle beach (MB) and pond (PO)) exhibit significant differences in particle size composition. This phenomenon can be attributed to the hydraulic classification that occurs after the tailings slurry is discharged into the tailings pond, coarse particles settle more rapidly and are deposited preferentially in the immediate vicinity, whilst fine particles settle more slowly and are transported by the water flow to settle at more distant locations. Furthermore, with continuous advances in mineral processing technology, the particle size distribution of tailings materials is showing a trend towards finer grains. For example, in the study by Pi et al. [42], the average particle size of iron tailings was as low as 5 μm. This high content of fine particles will significantly affect the geotechnical behavior of the tailings, such as permeability, compressibility and shear strength.

Table 1. Detailed parameters of the particle size distribution of tailings.

Tailings Type	D50 (mm)	Cu	Cc	Reference
Iron	0.2161	2.25	1.52	Zhang et al. [48]
Iron	0.0109	13.32	1.37	Zhang et al. [48]
Iron	0.12	3.11	1.05	Hu et al. [16]
Iron	0.03	8.82	0.59	Hu et al. [16]
Iron	0.04	19.31	0.68	Wei et al. [46]
Iron	0.05	15.24	1.67	Ke et al. [40]
Iron	0.007	-	-	Mmbando et al. [38]
Iron	0.037	9	2.15	Consoli et al. [37]
Iron	0.097	5.75	1.57	Consoli et al. [37]
Iron	0.075	8.8	3.25	Wagner et al. [45]
Iron	0.104	1.82	0.94	Wagner et al. [45]
Iron	0.103	11.87	3.06	Carvalho et al. [33]
Iron	0.081	10	1.98	Consoli et al. [35]
Iron	0.005	2.57	0.68	Pi et al. [42]
Gold	0.095	7.3	1.4	Chang et al. [34]
Gold	0.053	24.1	2.2	Chang et al. [34]
Gold	0.006	10.5	0.8	Chang et al. [34]
Gold	0.032	-	-	AI-Taehouni et al. [36]
Gold	0.011	7.3	1.4	Li et al. [41]
Gold	0.133	10.4	5.76	Fotovvat et al. [39]
Gold	0.05	17.9	1	Reid et al. [44]
Gold	0.076	32.33	2.89	Zhang et al. [47]
Gold	0.033	44.67	1.6	Nayanthara et al. [43]
Copper	0.097	-	-	Onuagnluchi et al. [52]
Copper	0.123	2.29	1.14	Hu et al. [16]
Copper	0.06	14.84	2.03	Hu et al. [16]
Copper	0.011	7.53	0.82	Xu et al. [49]
Copper	0.26	9.06	2.49	Velten et al. [50]
Copper	0.11	5.23	2.75	Velten et al. [50]
Copper	0.086	23.93	1.54	Velten et al. [51]
Copper	0.084	14.49	4.34	Velten et al. [51]

summarizes the detailed parameters of the particle size distribution for the aforementioned 31 tailings samples, including the mean particle size (D₅₀), the coefficient of uniformity (C_u) and the coefficient of curvature (C_c). The D₅₀ values range from 0.005 mm to 0.26 mm, all falling within the range of silt to fine sand. The range of C_u is from 2.25 to 44.67, whilst the range of C_c is from 0.59 to 5.76, covering a variety of scenarios from poor to good grading. Tailings with good grading are more likely to achieve higher compaction during the compaction process, thereby improving their shear strength. For tailings materials composed of silt and fine sand, the silt content significantly influences their packing behavior. When the silt content is low, silt particles can fill the framework voids formed by fine sand particles, enhancing interlocking and intergranular bonding; at this stage, e decreases as the silt content increases. When the silt content exceeds a certain threshold, fine sand particles gradually ‘float’ within the silt matrix, and e consequently increases. Thus, e exhibits a pattern of first decreasing and then increasing [53]. This variation in e is directly related to the geotechnical behavior of the tailings and influences the stability of the tailings dam. Poor gradation or excessive fine particle content can lead to increased compaction and reduced permeability during the stockpiling process, which, under static or dynamic loads, can easily trigger an increase in pore water pressure, thereby increasing the risk of liquefaction and deformation failure. Therefore, clarifying the particle size distribution characteristics of the tailings and their relationship with the fine particle content is crucial for assessing the long-term safety of tailings dams.

3.2. Mineralogical Composition

Figure 4 summarizes the mineralogical composition of 34 tailings [33,34,35,36,37,39,40,41,43,45,47,51,54,55,56,57,58,59,60,61,62,63,64,65], revealing significant diversity and heterogeneity in their mineralogical composition. The mineral types are predominantly silicates, accompanied by a certain proportion of carbonates and sulfates. Overall, the mineralogical composition of the tailings is jointly governed by the type of ore, the beneficiation process and the storage environment, resulting in marked variations. The statistical results indicate that quartz is the predominant mineral component, with 33 of the 34 tailings samples containing quartz, reflecting that tailings typically form on the basis of a rigid particle skeleton. This is followed by minerals such as hematite and kaolinite, indicating that iron oxides and clay minerals also occur with high frequency in the tailings. Furthermore, minerals such as calcite and chlorite are also relatively common in some tailings, whilst mica minerals, plagioclase and pyroxene minerals occur with moderate frequency. These minerals exhibit significant morphological differences; for example, quartz is predominantly granular in structure, whereas mica and some clay minerals are lamellar or flocculent [66]. Such differences lead to variations in the macroscopic geotechnical behavior of the tailings. Although a small number of sulfate minerals and evaporite minerals occur infrequently, they may have a significant impact on the chemical stability and environmental risks of tailings under specific environmental conditions.

Mineralogical composition does not directly control macroscopic mechanical behavior simply through compositional ratios, but rather exerts its influence indirectly by affecting particle properties, microstructure and structural evolution processes. Under conditions where particle size distribution and morphology are essentially consistent, differences in mineralogical composition alone can significantly alter the mechanical properties of tailings. For example, studies on compacted filtered copper tailings have shown that differences in mineralogical composition may influence the undrained shear response, critical state parameters, and CSL characteristics of tailings materials [51]. However, these responses are also closely associated with factors such as particle breakage, grading characteristics, fines content, and particle morphology, indicating that the mechanical behavior and critical state response of tailings are governed by multiple interacting factors rather than mineralogical composition alone. Concurrently, variations in the combination of different mineral components (such as quartz and iron oxides) lead to changes in particle breakage patterns and morphological evolution processes, thereby governing the stress–strain relationship and hardening or softening behavior of the tailings during loading [33]. On the other hand, the content of clay minerals is considered one of the key factors influencing the mechanical behavior of tailings. As the clay mineral content increases, the cohesion of the tailings gradually rises, the pore structure shifts from macrospores to micropores, and the pore distribution and specific surface area undergo significant changes, thereby affecting their compressibility and shear properties [67]. Furthermore, differences in mineralogical composition are often coupled with particle size distribution and density characteristics. Tailings formed by different mineral processing methods exhibit a synergistic relationship in terms of mineralogical composition and particle size distribution. This coupling effect significantly influences the compressive and shear behavior of the material at low stress levels, whilst under high-stress conditions, the response gradually converges towards a unified critical state [37]. These studies once again highlight the importance of analyzing the mineralogical composition of each tailings deposit, as it is not possible to apply standardized, generalized mineralogical data.

3.3. Chemical Composition

The chemical composition of tailings is typically determined using X-ray fluorescence spectroscopy (XRF), chemical analysis and inductively coupled plasma emission spectrometry (ICP). The main chemical compositions of different types of tailings are summarized in

Table 2. The main chemical composition of different types of tailings.

Tailings Type	Main Composition (wt %)								Reference
	SiO2	Al2O3	CaO	Fe2O3	Na2O	MgO	K2O	SO3
Iron	56.18	6.53	2.44	10.45	-	3.43	-	-	Wei et al. [46]
Iron	44.00	4.90	-	47.1	-	1.4	-	-	Medina et al. [59]
Iron	82.67	4.35	-	12.91	-	-	-	-	Carneiro et al. [55]
Iron	50.80	-	0.60	39.69	-	-	1.56	-	Schatzmayr et al. [62]
Iron	63.47	12.55	3.59	9.79	-	2.87	3.22	0.37	Cao et al. [69]
Iron	60.60	12.26	4.23	14.16	-	3.05	2.32	0.14	Cao et al. [69]
Iron	42.06	11.51	10.50	15.50	-	2.54	5.37	-	Li et al. [70]
Iron	48.10	6.97	8.36	7.08	0.517	26.53	1.03	0.262	Xia et al. [71]
Iron	75.23	2.64	1.47	11.31	0.49	2.10	0.40	0.08	Cheng et al. [72]
Iron	51.40	13.60	6.58	10.20	2.21	5.07	3.06	3.84	Jia et al. [73]
Gold	71.16	13.40	3.70	2.04	1.94	0.55	5.54	0.26	Zhang et al. [65]
Gold	65.21	19.13	3.15	2.42	2.75	0.83	5.32	0.41	Wang et al. [74]
Gold	70.37	13.83	5.85	0.99	3.62	0.18	1.74	0.20	Chen et al. [75]
Gold	66.43	15.62	7.54	0.77	2.49	0.14	4.38	0.13	Chen et al. [76]
Gold	59.60	8.57	10.94	11.59	0.09	1.83	1.53	-	Pyo et al. [77]
Gold	70.29	16.12	1.79	1.31	3.35	0.38	5.38	0.16	Liu et al. [78]
Gold	65.70	14.30	1.88	3.05	-	0.49	-	0.13	Cao et al. [79]
Gold	42.93	9.25	17.56	3.92	3.99	2.58	0.28	-	Wang et al. [80]
Gold	64.97	18.29	3.17	2.47	2.84	-	5.64	-	Li et al. [81]
Gold	71.25	14.28	3.34	3.77	-	0.49	-	0.21	Yang et al. [82]
Copper	56.26	9.38	2.89	8.35	-	2.88	0.98	8.6	Wang et al. [64]
Copper	47.92	14.44	14.14	6.00	-	1.44	3.38	1.07	Li et al. [70]
Copper	50.82	15.82	5.97	14.06	3.39	2.17	4.09	-	Onuaguluchi et al. [52]
Copper	28.15	5.49	35.76	8.39	-	0.13	1.32	11.21	Liu et al. [83]
Copper	40.04	7.40	20.59	5.59	-	9.85	-	2.32	Zhang et al. [84]
Copper	75.00	12.16	0.16	3.60	4.30	0.49	1.85	-	Thomas et al. [85]
Copper	60.90	17.03	2.90	3.86	0.60	1.63	2.63	4.50	Barzegar et al. [86]
Copper	31.85	7.22	20.82	20.17	-	5.92	-	9.33	Chen et al. [87]
Copper	62.73	18.92	2.82	4.93	-	2.47	-	1.82	Chen et al. [87]
Copper	59.70	17.10	5.60	4.70	0.20	0.90	7.50	2.70	Xu et al. [49]

and Figure 5. The analytical results indicate that the primary chemical constituents of the tailings are SiO₂, Al₂O₃, CaO and Fe₂O₃. The presence of these compounds is closely related to their mineralogical composition, particularly corresponding to major phases such as quartz, hematite and various silicate minerals. Figure 5a,b show that SiO₂ is the predominant component in the tailings; furthermore, the CaO content is significantly lower than that of the other three oxides. Additionally, Table 2 lists trace amounts of other oxide components present in the tailings, such as Na₂O, MgO, K₂O and SO₂. It is worth noting that the chemical composition of tailings is not static but is influenced by a combination of factors, including variations in beneficiation processes, differences in ore types, changes in mining faces, and heterogeneity in storage locations within the tailings dam. These factors collectively result in significant spatial variability in the chemical composition of the tailings [68].

In certain ore mining operations, heavy metal elements are often present in the tailings, constituting a major source of potential environmental risk. Heavy metal concentrations are typically determined using ICP analysis.

Table 3. The main heavy metal components in tailings.

Main Composition (ppm)										Reference
Cu	Pb	Zn	Sb	Cr	Mn	As	V	Hg	Cd
20.80	22.30	11.50	41.70	14.50	52.10	37.60	16.90	-	-	Chen et al. [75]
18.60	20.30	11.30	39.70	14.50	50.10	35.60	15.20	-	-	Chen et al. [76]
35.10	32.60	85.80	-	44.90	425.30	-	280.40	-	-	Wang et al. [80]
32.00	62.00	287.00	-	7.00	-	-	-	-	-	Ince et al. [88]
120.00	87.10	1990.00	-	-	-	0.03	-	34.61	20.80	Opiso et al. [89]
120	4.3	71	32	74	1500	1520	59	-	0.3	Kiventerä et al. [90]
310	4.3	71	32	74	-	1520	85	-	85	Kiventerä et al. [91]
195.05	590.79	84.03	-	103.65	-	1788	-	146.8	2.78	Musiige et al. [92]

lists the major heavy metals found in tailings. These include Cu, Cr, Pb, Zn, Se, As, Cd and Hg. Under natural conditions, these heavy metals can be released into the environment through various pathways, primarily including leaching, surface water erosion, physical weathering, microbial metabolic activity, and chemical desorption caused by pH fluctuations. Once released, these heavy metals cause persistent contamination of surrounding soil, surface water and groundwater, thereby threatening the safety of ecosystems and human health. Consequently, the solidification and stabilization of heavy metals in tailings have become key research topics in the field of environmental geotechnical engineering.

3.4. Specific Gravity and Atterberg Limits

The specific gravity (G_s) and Atterberg limits of different types of tailings are summarized in Table 4. It can be seen that there are certain differences in G_s and Atterberg limits among the various types of tailings. This is primarily related to the particle size mentioned in Section 3.1 and the mineralogical composition discussed in Section 3.2. Overall, the G_s of tailings is mainly concentrated within the range of 2.7–3.3, although some variation exists among different mineral types. For example, the G_s of gold tailings is generally stable, mostly ranging between 2.75 and 2.9, whereas iron tailings, due to their higher content of high-density minerals (such as iron oxides), typically have a higher G_s, reaching 3.0 or above and even approaching 3.7. In contrast, the G_s of copper tailings falls within a relatively narrow range, generally 2.7–2.8. These variations are closely related to mineralogical composition; in particular, changes in the content of heavy minerals have a significant impact on G_s.

With regard to plasticity indices, the liquid limit (LL), plastic limit (PL) and plasticity index (PI) of different tailings exhibit considerable variability. Overall, the LL of most tailings is concentrated between 18% and 39%, the PL is approximately 13–24%, and the PI is mainly distributed within the range of 2–18%. Gold tailings typically exhibit low to moderate plasticity, with a PI mostly below 10, although higher values may be attained when the fine-grain content is high. The plasticity of iron tailings varies considerably; some fine-grained iron tailings exhibit high plasticity, whilst coarse-grained tailings generally show no plasticity characteristics. Copper tailings generally exhibit moderate plasticity, particularly fine-grained tailings, where the PI can reach around 10–15. It is worth noting that even within the same type of tailings, plasticity behavior may vary significantly due to differences in mineralogical composition between samples. The Atterberg limit parameters reflect, to a certain extent, the influence of fine-grained content and clay minerals in the tailings. Tailings with higher LL and PI typically possess stronger water absorption capacity and plasticity, generally exhibiting greater compressibility and deformability. Conversely, low-plasticity or non-plastic tailings are generally dominated by coarse grains, exhibiting lower compressibility and higher permeability. Furthermore, variations in the PI may also influence the critical state characteristics and liquefaction potential of tailings [93,94], which is of significant importance for the stability analysis of tailings dams.

Table 4. The G_s and Atterberg limits of different types of tailings.

Tailings Type	Specific Gravity Gs	Liquid Limit LL (%)	Plastic Limit PL (%)	Plasticity Index PI (%)	Reference
Gold	2.9	22.5	20	2.5	Al-Tarhouni et al. [36]
Gold	2.89	-	-	-	Li et al. [41]
Gold	2.75	-	-	-	Fotovvat et al. [39]
Gold	2.78	18	16	2	Reid et al. [44]
Gold	2.75	22.8	13.04	9.76	Zhang et al. [47]
Gold	2.77	24	18.7	5.3	Nayanthara et al. [43]
Gold	2.78	20	16	4	Urbina et al. [95]
Iron (Coarse)	3.23	-	-	-	Hu et al. [16]
Iron (Fine)	3.08	28	19	9	Hu et al. [16]
Iron	2.91	21.3	15.8	5.5	Wei et al. [46]
Iron	2.9	25.2	15.6	9.6	Ke et al. [40]
Iron	3.71	39	21	18	Mmbando et al. [38]
Iron	3.05	-	-	-	Wagner et al. [45]
Iron	2.97	-	-	-	Wagner et al. [45]
Iron (Flotation)	2.76				Wagner et al. [63]
Iron (Slime)	3.26	31	24	7	Wagner et al. [63]
Copper	2.79	26–39	22–27	4–12	Shamsai et al. [96]
Copper (Coarse)	2.77	-	-	-	Hu et al. [16]
Copper (Fine)	2.76	28	13	15	Hu et al. [16]

Table 4. The G_s and Atterberg limits of different types of tailings.

Tailings Type	Specific Gravity Gs	Liquid Limit LL (%)	Plastic Limit PL (%)	Plasticity Index PI (%)	Reference
Gold	2.9	22.5	20	2.5	Al-Tarhouni et al. [36]
Gold	2.89	-	-	-	Li et al. [41]
Gold	2.75	-	-	-	Fotovvat et al. [39]
Gold	2.78	18	16	2	Reid et al. [44]
Gold	2.75	22.8	13.04	9.76	Zhang et al. [47]
Gold	2.77	24	18.7	5.3	Nayanthara et al. [43]
Gold	2.78	20	16	4	Urbina et al. [95]
Iron (Coarse)	3.23	-	-	-	Hu et al. [16]
Iron (Fine)	3.08	28	19	9	Hu et al. [16]
Iron	2.91	21.3	15.8	5.5	Wei et al. [46]
Iron	2.9	25.2	15.6	9.6	Ke et al. [40]
Iron	3.71	39	21	18	Mmbando et al. [38]
Iron	3.05	-	-	-	Wagner et al. [45]
Iron	2.97	-	-	-	Wagner et al. [45]
Iron (Flotation)	2.76				Wagner et al. [63]
Iron (Slime)	3.26	31	24	7	Wagner et al. [63]
Copper	2.79	26–39	22–27	4–12	Shamsai et al. [96]
Copper (Coarse)	2.77	-	-	-	Hu et al. [16]
Copper (Fine)	2.76	28	13	15	Hu et al. [16]

4. Permeability Behavior

The permeability characteristics of tailings are key factors determining the location of the seepage line in tailings dams, the efficiency of drainage and consolidation, and the overall stability of the dam. Due to the complex and variable nature of the particle size distribution, mineralogical composition, chemical composition and environmental conditions of tailings, as mentioned above, their permeability behavior exhibits marked non-linearity and unique characteristics, which traditional empirical formulas in soil mechanics often fail to describe accurately. This section provides a systematic review of the primary factors influencing tailings permeability, the underlying micro-mechanisms, and recent research developments regarding permeability coefficient (k) prediction models.

4.1. Macroscopic Factors Affecting the Permeability of Tailings

4.1.1. Threshold for Fine Particle Content

Fine particle content (typically referring to particles smaller than 0.075 mm) is widely recognized as an important factor influencing tailings permeability behavior; Figure 6 summarizes the relationship between k and fine particle content. Studies by Gan et al. [97] and Fan et al. [98] both indicate that for the tailings materials investigated in their studies, k decreased sharply as the fine particle content increased when the fines content was below approximately 40%, whereas permeability tended to stabilize once the fines content exceeded this range. Gan et al. [97] attribute this phenomenon to a change in the packing state of fine grains within the pores of coarse grains. When fine-grained tailings are scarce, they primarily fill the skeletal voids formed by coarse particles, significantly reducing the effective seepage pathways; when the fine particle content exceeds the threshold, the coarse particles are completely enveloped by the fine-grained matrix, forming a structure where ‘coarse particles are suspended in the fine-grained matrix’. At this stage, seepage is primarily controlled by the fine-grained matrix, and permeability tends to stabilize. Wang et al. [99] further confirmed through laboratory model tests and discrete element simulations that an increase in the fine particle content significantly alters the gradation parameters, porosity and stress–strain distribution of the specimen, thereby influencing the mechanisms of seepage occurrence and development.

It should be noted, however, that the transition range reported in the literature is strongly dependent on particle shape, gradation characteristics, mineralogical composition, depositional fabric, density state, and packing conditions. Consequently, permeability behavior cannot be interpreted solely based on fines content, and caution should be exercised when transferring such transition values across different tailings materials. Taken together, the available evidence suggests that fines content acts as one of several interacting factors governing permeability behavior rather than a universally applicable controlling parameter.

4.1.2. Void Ratio and Stress Levels

The relationship between k and e is a key area of study in classical soil mechanics. Most studies indicate that there is a linear relationship between log k and e for tailings [16,18,100,101,102,103,104]. Figure 7 summarizes the relationship between k and e of tailings reported in various literature sources. Experimental results by Hu et al. [16] on iron and copper tailings show that as pressure increases, e decreases and k decreases significantly. Qiu and Sego [18] conducted experiments on various types of tailings, including coal, gold and copper tailings, and observed the same trend. However, the total e alone is often insufficient to accurately characterize the seepage capacity of tailings, as not all pores contribute to seepage. Zheng et al. [104] utilized nuclear magnetic resonance (NMR) technology to investigate and found that, although the total e of fine-grained tailings may be high, over 97% of the pores are ultra-micro and micro-pores (<0.1 μm), and these pores make a very limited contribution to seepage. By contrast, pores larger than 0.1 μm (including small, medium and large pores) constitute the primary seepage pathways. Therefore, the introduction of the concept of effective e is of great significance for the accurate evaluation of tailings permeability.

As the height of a tailings dam increases, the stress experienced by the tailings at the base can reach several megapascals, significantly altering their permeability characteristics. Ma et al. [102] utilized a proprietary high-pressure consolidation-permeameter to systematically investigate the behavior of tailings of different grain sizes (CLT: clayey tailings; SLT: silty-clayey tailings; STT: silty tailings; SDT: sandy tailings) under high-pressure conditions. The results indicated that when the consolidation pressure exceeded 2 MPa, both the compressibility and permeability of the tailings exhibited a ‘kink’. For coarse-grained tailings, particle breakage is the primary cause of increased compressibility under high pressure, whereas for fine-grained tailings, the role of the diffuse double layer and the oriented arrangement of particles lead to a decrease in compressibility. Gan et al. [105] also pointed out that when the pressure exceeds 200 kPa, seepage channels within the tailings become blocked; at this point, k calculated solely based on the e are overestimated, and corrections must be introduced. The above studies indicate that the influence of stress levels on tailings permeability exhibits non-linear characteristics and is closely related to particle breakage and microstructural evolution.

4.1.3. Chemical and Biological Effects

As discussed in Section 3.2 and Section 3.3, tailings often contain reactive minerals and metal ions, which can alter their permeability through chemical reactions and biological processes. Yin et al. [106] found through laboratory soil column tests that ferrous ions (Fe²⁺) are oxidized during seepage into less soluble iron oxides/hydroxides (such as Fe(OH)₃). These colloidal substances deposit on seepage pathways and geotextiles, causing chemical clogging and significantly reducing the permeability of the tailings (Figure 8a). He et al. [107] investigated the effect of Cu²⁺ on compacted tailings and found that Cu²⁺ compresses the diffusion double layer on the surface of clay particles, altering interparticle forces and thereby increasing the compressibility and permeability of the tailings(Figure 8b). Furthermore, research by de Lucas Pardo et al. [108] revealed that a polychaetae worm (such as Tubifex tubifex) can survive in tailings and, through its biological disturbance, significantly improve the permeability of the tailings, thereby accelerating their dewatering and strength development. These contrasting observations indicate that chemical and biological processes influence permeability primarily through their effects on pore structure and flow pathways. Processes that promote pore clogging or mineral precipitation tend to reduce permeability, whereas processes that increase pore connectivity or modify particle arrangements may enhance seepage capacity. These studies highlight the significant influence of chemical and biological processes on the permeability of tailings. It is therefore necessary to take the impact of these factors into account when designing the impermeabilization of tailings dams.

4.2. Microstructure and Seepage Mechanisms

With advances in testing technology, NMR and CT techniques have become powerful tools for studying the microstructure of tailings. Zheng et al. [104] utilized NMR technology to classify the pores in tailings into ultra-micro, micro, small, meso and macro pores, and clearly identified a pore size of 0.1 μm as the threshold for distinguishing between adsorption pores and flow pores. This classification provides direct evidence for understanding the contradictory phenomenon of ‘high e, low permeability’ in tailings. Li et al. [101] utilized CT scanning technology to reconstruct a three-dimensional pore-channel model of tailings, discovering that permeability is influenced not only by e but also, to a greater extent, by microstructural parameters such as channel radius, pore-to-channel ratio and coordination number. Correlation analysis indicates that the influence of channel radius on permeability is even greater than that of pore radius.

Tailings particles are typically angular in shape, unlike naturally rounded sandy soils, which affects their packing density and pore structure. Li et al. [101] found through CT scanning and image analysis that the particle shape of clayey tailings is closer to circular, whereas sandy tailings exhibit more pronounced angularity. This morphological difference influences the manner of particle contact and the distribution of force chains. Discrete element simulations by Wang et al. [99] indicated that an increase in the fine-grain content alters the force chain network within the tailings specimen, resulting in a reduction in strong force chains and an increase in total force chains. Under the action of seepage forces, the migration of fine grains leads to the restructuring of force chains, thereby affecting the stability and permeability of the tailings structure.

4.3. Permeability Coefficient Prediction Model

Traditional k prediction models, such as the Hazen, Terzaghi, Kozeny and Chapuis models, exhibit significant errors when applied to tailings. Gan et al. [97,105] systematically compared the applicability of these traditional models and found that the results calculated by the Kozeny model differed from measured values by two to three orders of magnitude, whilst the Hazen and Terzaghi equations were only relatively accurate within specific ranges of fine-grain content. Consequently, tailings were classified into three categories based on their fine-grain content, and permeability coefficient prediction models based on d₁₀ and e_f were proposed respectively, significantly improving accuracy. Concurrently, various researchers have conducted studies on k prediction models for tailings; the relevant findings and traditional prediction models are summarized in

Table 5. Summary of permeability coefficient prediction models.

Reference	Formula	Key Parameter	Model Description
Hazen et al. [110]	$k=c{d}_{10}^2$	C, d10	The classical empirical model, which posits that permeability is determined entirely by the finest particles in the soil, is applicable to clean sandy soils with a uniform grain size distribution.
Terzaghi et al. [111]	$k=2d_{10}^2{e}^2$	d10, e	The empirical model has been refined by incorporating the e into the Hazen formula to account for the influence of soil compaction on permeability.
Kozeny et al. [112]	$k=780{\displaystyle \frac{e^3}{{(1-e)}^2}}{d}_9^2$	e, d9	The capillary theory model treats porous media as a bundle of parallel capillaries and establishes the relationship between the permeability coefficient and the porosity and characteristic particle size.
Carman et al. [113]	$k={\displaystyle \frac{Cg}{u_f{\rho }_s^2{S}^2}}·{\displaystyle \frac{e^3}{(1+e)}}$	e, S, C	The Kozeny–Carman model is the most widely used semi-theoretical model; it incorporates the specific surface area(S), taking into account the effects of fluid properties and particle shape.
Chapuis et al. [114]	$k=2.4622{(d_{10}^2{\displaystyle \frac{e^3}{1+e}})}^{0.7825}$	d10, e	The empirical regression model, derived from a large volume of experimental data, establishes a non-linear relationship between k and d10 and e, and is applicable to specific types of soil.
Gan et al. [97]	$k=R{\mathrm{e}}_{\mathrm{f}}{d}_{10}$	R, ef, d10	A new model for tailings introduces the silty void ratio ef = e/FC (where FC represents the fine-grain content), which comprehensively accounts for the effects of effective particle size and the packing state of fine grains on permeability.
Zheng et al. [104]	$k=a{(A_{T_2,>0.1\mathrm{\mu }\mathrm{m}})}^b$	$A_{T_2,>0.1\mathrm{\mu }\mathrm{m}}$	Based on a microstructural model and utilizing NMR technology, a power-law relationship was established using the spectral area corresponding to pores with diameters greater than 0.1 μm T2 as an indicator of effective e.
Ma et al. [102]	$k=C_s{\displaystyle \frac{{(e-a_0{G}_s{w}_l)}^n}{1+e-a_0{G}_s{w}_l}}(1+B_r)$	Gs, wl, Br	The high-pressure correction model incorporates the effective e (excluding the pores occupied by bound water) and Br, and is suitable for predicting the permeability of tailings under high-stress conditions.
Fan et al. [98]	$k={\displaystyle \frac{Cg}{u_f{\rho }_s^2{S}^2}}·{\left({\displaystyle \frac{e^3}{(1+e)}}\right)}^A$	A, e, S, C	The model introduces a coefficient A that accounts for the influence of particle angularity, thereby providing a better description of the permeability of angular particles.
Babaoglu et al. [109]	$k=k_0·{\displaystyle \frac{e^5}{e_0^5}},k=k_0·{\displaystyle \frac{e^5}{e_0^5}}·({\displaystyle \frac{1+e_0}{1+e}})$	k0, e0, e	Single-point calibration of a power-law model using a measured permeability coefficient k0 to calibrate the relationship k∝e5 or k∝e5/(1 + e).
Wong et al. [21]	$k_{NST}=k_f(1-{\displaystyle \frac{1}{1+e_g}})$	kf, eg	The NST permeability model, based on mixture theory, assumes that permeability is controlled by the fine-grained tailings matrix, whilst coarse particles act to block the flow pathways.

. Babaoglu and Simms [109], through the analysis of extensive data, proposed calibrating power-law functions using single-point measured permeability coefficients at high e, and found that k exhibits a high correlation with e⁵/(1 + e) or e⁵, with prediction accuracy within one order of magnitude. On the other hand, Fan et al. [98] modified the Kozeny–Carman equation by introducing a coefficient A that accounts for the influence of particle angularity, thereby enabling a better description of the permeability of tailings containing angular particles. Furthermore, Ma et al. [102] proposed a modified k prediction model under high-pressure conditions that takes into account the effective e and particle breakage rate (B_r). Collectively, these developments indicate that the prediction of tailings permeability has evolved from simple gradation-based empirical equations toward models that explicitly incorporate microstructural characteristics and state-dependent variables. The improved performance of recent models suggests that permeability is governed not only by particle size distribution, but also by pore structure, particle morphology, effective void ratio, and particle breakage processes.

It is worth noting that when selecting a predictive model, the physical properties of the tailings should be taken into account. Rios et al. [103], when comparing iron tailings and natural silt, found that although the two had similar gradations, their permeability and stiffness differed significantly due to variations in particle density and shape; this indicates that general-purpose models require calibration for specific tailings.

5. Compression and Consolidation Behavior

The consolidation and compression of tailings are critical processes in tailings dam engineering that determine storage capacity, drainage efficiency and dam stability. Unlike natural soil, tailings are typically discharged into tailings dams in the form of slurry with a high moisture content; the consolidation process often involves large strains, non-linear compression and permeability characteristics, and is influenced by multiple factors such as particle size distribution, chemical composition and treatment methods. This section provides an overview of the consolidation theory of tailings, as well as the consolidation and compression parameters and influencing factors.

5.1. Theoretical Model of Consolidation and Compression

The traditional Terzaghi small-strain consolidation theory is based on the small-strain assumption and assumes that k and compression index (C_I) are constant; it is therefore not applicable to the large-strain consolidation process of tailings slurries with high moisture content [115,116]. To accurately describe the consolidation behavior of tailings, the finite-strain consolidation theory proposed by Gibson et al. [117,118] has been widely adopted; this theory accounts for large deformations, self-weight effects, and the non-linear compression and permeability relationships of the material.

Jeeravipoolvarn et al. [115] found through 10-m-high standpipe tests that the classical Gibson model overestimates the consolidation rate of fine-grained oil sands tailings, due to its failure to account for time-dependent behaviors such as thixotropy and creep. Based on Gibson’s theory, Bonin et al. [119] employed the piecewise linear model CS2 to simulate the self-consolidation of concentrated tailings; this model accurately reproduces the processes of pore pressure dissipation and settlement. Qin et al. [18], building upon the Gibson model, proposed incorporating the deposition stage into the analytical solution, using e at the onset of air ingress in the consolidation test as the critical point, thereby achieving a unified estimation of both deposition and consolidation settlement.

Independently of Gibson’s theory, Wong et al. [21] established a one-dimensional consolidation mechanism model for NST based on mixture theory. By introducing inter-particle e and interfacial e, this model characterized the three-stage evolution of the NST consolidation process, ‘fine grain controlled, transition, and coarse grain controlled’, providing a new theoretical perspective for understanding the synergistic effects in mixed coarse and fine-grain tailings.

5.2. Compression Characteristics and Key Parameters

It should be noted that the compression and consolidation behavior of tailings is governed by a variety of interacting factors and cannot therefore be described solely in terms of the material’s intrinsic properties. Parameters such as the C_I and consolidation coefficient (C_v) are significantly influenced by the initial structure of the specimen, particle size distribution, mineral composition, stress range, specimen preparation method, and particle fragmentation behavior. Consequently, even for tailings of the same type, different test conditions may result in markedly different compression responses.

5.2.1. Influence of Initial Structure and Specimen Preparation on Compression Curve Shape

The compressibility characteristics of tailings are typically characterized by the relationship between e and effective vertical stress (${\sigma }^{\prime }$). Figure 9 summarizes the compressibility curves of tailings obtained from the literature review. When tests are conducted using slurry-like tailings, their e-log ${\sigma }^{\prime }$ compressibility curves often exhibit distinct piecewise linear or non-linear characteristics. Wong et al. [21] found in their tests on NST slurry that the compression curve could be divided into three linear segments, corresponding to the fine-grain-controlled, transitional and coarse-grain-controlled stages (Figure 9a). Bonin et al. [119] also observed non-linear compression behavior in their self-consolidation tests on concentrated gold tailings slurry and fitted the data using a power-law function (Figure 9b). Experiments by Islam et al. [23] on coal slurry also indicated that, during the transition from a slurry-like state to a soil-like state, the compression curve exhibits a distinct inflection point (Figure 9c). Such non-linear characteristics are primarily attributed to the extremely high initial e of slurry-state tailings and the absence of effective particle-particle contact, resulting in significant irreversible volume shrinkage during the initial drainage phase.

The studies summarized in Figure 9 collectively indicate that the shape of the compression curve is governed not only by the stress level but also by the initial structural state of the tailings. Slurry-like tailings generally possess extremely high e and weak interparticle contacts, resulting in large particle rearrangement and rapid volume reduction during the early stages of consolidation. Consequently, their compression behavior is often characterized by non-linear or segmented responses. In contrast, remolded or compacted tailings already possess a more stable particle skeleton before loading, leading to a more uniform compression mechanism and approximately linear e-log ${\sigma }^{\prime }$ relationships within conventional stress ranges. The transition from non-linear to linear compression behavior therefore reflects the progressive development of particle contacts and load-bearing structures during consolidation. Consistent with this interpretation, when tests are conducted using remolded samples (prepared after drying and compaction), the e-log ${\sigma }^{\prime }$ compression curve of the tailings is typically approximately linear within the stress range under investigation. Hu et al. [16] conducted consolidation tests on remolded samples of iron and copper tailings and found that their compression curves could be described linearly using the conventional C_I (Figure 9d). Tests by Qiu and Sego [18] on remolded samples of copper, gold and coal tailings similarly showed a linear compression trend (Figure 9e). Ma et al. [102] conducted high-pressure consolidation tests on remolded samples of tailings with different particle sizes, revealing that the compression curve is linear in the low-pressure range (<2 MPa), whilst a new linear segment appears in the high-pressure range (≥2 MPa), meaning the curve exhibits segmented linearity rather than a single non-linear behavior (Figure 9f). This change in compression response is primarily associated with the onset of particle breakage under high stress levels, which alters the particle size distribution and packing structure, resulting in additional compressibility.

These findings suggest that the shape of the compression curve depends largely on the initial structure and preparation method of the tailings sample, rather than on its intrinsic properties. Consequently, in engineering practice, the appropriate sample preparation method should be selected based on the actual storage conditions of the tailings in order to determine their compression and consolidation characteristics.

5.2.2. Compression Index

C_I is commonly used to characterize the compressibility of tailings under specific stress conditions and specimen states. However, C_I is not an intrinsic material constant, but a state-dependent parameter influenced by factors such as particle size distribution, mineralogical composition, stress range, and specimen preparation method.

Table 6. The compression index C_I of different types of tailings.

Reference	Tailings Type	Compression Index, CI	Reference	Tailings Type	Compression Index, CI
Wong et al. [21]	Fine oil sands tailings	0.447	Wong et al. [21]	Coarse oil sands tailings	0.036
Wong et al. [21]	NST (Fine content: 21%) stage 1	0.419	Wong et al. [21]	NST (Fine content: 21%) stage 2	0.092
Wong et al. [21]	NST (Fine content: 21%) stage 3	0.041	Bonin et al. [119]	Gold tailings	0.052–0.070
Islam et al. [123]	Coal tailings	0.223	Hu et al. [16]	Fine iron tailings	0.260
Hu et al. [16]	Coarse iron tailings	0.046	Hu et al. [16]	Fine copper tailings	0.085
Hu et al. [16]	Coarse copper tailings	0.025	Qiu and Sego [18]	Copper tailings	0.056–0.094
Qiu and Sego [18]	Gold tailings	0.083–0.156	Qiu and Sego [18]	Coal tailings	0.370–0.396
Carrera et al. [124]	Fine Stava tailings	0.19	Carrera et al. [124]	Coarse Stava tailings	0.11
Li et al. [120]	Fine iron tailings	0.38	Li et al. [120]	Coarse iron tailings	0.20
Ma et al. [102]	SDT (low pressure)	0.069	Ma et al. [102]	SDT (high pressure)	0.355
Ma et al. [102]	STT (low pressure)	0.162	Ma et al. [102]	STT (high pressure)	0.265
Ma et al. [102]	SCT (low pressure)	0.086	Ma et al. [102]	SCT (high pressure)	0.311
Ma et al. [102]	CLT (low pressure)	0.077	Ma et al. [102]	CLT (high pressure)	0.324

summarizes the C_I values for different types of tailings using various test methods. It can be seen that there are significant differences in C_I values between different tailings, ranging from 0.025 (coarse-grained copper tailings) to 0.447 (fine-grained oil sands tailings). Overall, oil sands tailings and coal tailings exhibit significantly higher compressibility than metal tailings: the C_I value for fine-grained oil sands tailings reaches as high as 0.447 [21], whilst that for coal tailings ranges from 0.223 to 0.396 [18,23]; in contrast, the C_I values for metal tailings such as iron, copper and gold are generally below 0.26 [16,18,120]. This difference is primarily attributed to mineralogical composition and plasticity. Oil sands tailings are rich in expansive clays such as montmorillonite and illite, as well as residual bitumen; coal tailings contain large amounts of carbonaceous and clay minerals; whereas metal tailings consist predominantly of low-plasticity, hard particles such as quartz and hematite. Furthermore, for the same type of tailings, the C_I of fine-grained tailings is typically higher than that of coarse-grained tailings, and may even be up to 12 times higher. For example, the C_I value of fine-grained oil sand tailings is 0.036, whereas that of coarse-grained oil sand tailings is 0.446. This reflects the high compressibility and low skeletal strength of fine-grained tailings. The influence of stress level on C_I exhibits non-linear characteristics; Ma et al. [102] found in high-pressure consolidation tests on remolded tailings samples that when the consolidation pressure exceeded 2 MPa, the C_I values of tailings of all particle sizes increased significantly (e.g., SDT increased from 0.069 to 0.355), and the C_I values of different particle sizes tended to converge. This indicates that particle breakage becomes the dominant factor under high stress, weakening the control of the original gradation on compressibility. Furthermore, methods such as freeze–thaw treatment [121], MICP biological treatment [122], and the addition of coarse particles to form NST [21] can all effectively alter the C_I of tailings, providing a theoretical basis for accelerating tailings dewatering in engineering applications.

5.2.3. The Coefficient of Consolidation

The C_v governs pore pressure dissipation and the rate of consolidation, and is a key parameter in engineering design for assessing the rate of settlement in tailings dams. Research indicates that the C_v of tailings is significantly influenced by consolidation pressure and material composition. Wong et al. [21] found that for NST, when the fine-grain content is below 20%, the C_v value can be 10 to 100 times higher than that of pure fine-grain tailings; this is attributed to the coarse-grain skeleton providing a more permeable pathway. Islam et al. [23] conducted constant rate loading tests and found that the C_v of coal tailings decreased by approximately two orders of magnitude as the effective stress increased from low levels to 70–80 kPa, consistent with a sharp decrease in permeability. Jeeravipoolvarn et al. [115] calculated from a 10 m high standpipe test that the C_v of oil sands fine tailings is approximately 15–26 m²/year, whereas after freeze–thaw treatment, C_v can increase by a factor of 10–100 [121]. Ma et al. [102] found in high-pressure consolidation tests that C_v exhibits a piecewise variation with increasing effective stress, with an inflection point occurring near the critical pressure of 2 MPa. Overall, increasing the coarse-grain content, improving drainage efficiency, and implementing pre-treatment measures such as freeze–thaw cycles can all effectively enhance the C_v of tailings and accelerate the consolidation process.

Overall, the compression and consolidation behavior of tailings exhibits strong state dependency and is jointly controlled by material properties, initial structure, and stress conditions. Therefore, caution should be exercised when directly comparing compressibility parameters reported in different studies without considering the corresponding testing conditions and specimen preparation methods.

6. Shear Behavior

The shear behavior of tailings is a key factor in the stability analysis of tailings dams, particularly in the assessment of liquefaction risk. The shear behavior of tailings is typically investigated through laboratory triaxial tests. Numerous researchers have conducted extensive triaxial tests on various types of tailings. This section systematically reviews the strength and deformation characteristics of tailings under monotonic loading conditions, their anisotropic behavior, critical states and static liquefaction mechanisms, as well as their shear behavior under special conditions.

6.1. The Strength and Deformation Characteristics of Tailings Under Monotonic Loading Conditions

6.1.1. Stress–Strain Relationship

The stress–strain response of tailings can be more fundamentally interpreted within the framework of CSSM. In this framework, the contractive or dilative tendency of tailings during shearing is closely related to the position of the initial state relative to the CSL. Tailings with initial states located above the CSL generally exhibit contractive behavior during shearing, whereas specimens located below the CSL tend to exhibit dilation and strain softening. Therefore, factors such as density, confining pressure, drainage condition, particle characteristics, and specimen preparation method influence the stress–strain response primarily through their effects on the state of the material relative to the CSL.

The stress–strain behavior of tailings under triaxial shear conditions exhibits a strong dependence on confining pressure, while drainage conditions significantly influence the effective stress path and mechanical response characteristics [35,37,39,59,124,125,126]. Under drained shear (CD) conditions, at low confining pressures (typically below 0.8–1.2 MPa), dense or initially high density tailings specimens generally exhibit strain softening behavior, characterized by a distinct peak strength on the stress–strain curve, which gradually decreases with increasing axial strain after reaching the peak [35,37,38,126]. For example, high-pressure CD tests on copper tailings conducted by Zhang et al. [126] indicated that, under confining pressures of 0.2–0.8 MPa, the specimens exhibited typical strain-softening behavior, with the axial strain corresponding to the peak strength being approximately 3%. At the same time, specimens under low confining pressure generally exhibit dilation, with volumetric strain first decreasing and then increasing, displaying a characteristic of initial contraction followed by expansion [50,59,126,127]. From the perspective of CSSM, these specimens are generally located below the CSL and therefore exhibit dilative tendencies during shearing. This is attributed to the tumbling and interlocking effects of particles under low confinement; the rearrangement of particles during the shearing process leads to volumetric expansion. When confining pressure rises above 1.2 MPa, even with a high initial density, the tailings tend to exhibit strain hardening; the stress–strain curve shows no distinct peak, and the specimen continues to contract throughout the entire shearing process [37,38,125,126]. As the confining pressure increases, the state of the specimen gradually approaches or moves above the CSL, resulting in more contractive behavior and a transition from strain softening to strain hardening. CD tests on compacted iron tailings conducted by Silva et al. [125] indicate that, under confining pressures of 400–8000 kPa, compacted specimens exhibit a hardening stress–strain curve with no significant strain softening. This transition is generally associated with the suppression of dilation under high confinement, enhanced particle rearrangement, and possible particle breakage; the resulting debris fills the voids, thereby inhibiting dilation.

Under undrained shear (CU) conditions, the stress–strain behavior of tailings exhibits trends generally similar to those observed under drained conditions; however, the evolution of pore pressure significantly influences the effective stress path and mechanical response. From the perspective of CSSM, the contractive or dilative behavior of tailings under undrained loading is closely related to the initial state of the specimen relative to the CSL. Under low confining pressure, compacted specimens may still exhibit strain softening after reaching peak strength, and such behavior is associated with the dilative tendency of specimens initially located below the CSL. Under undrained conditions, pore pressure evolution modifies the effective stress path, often resulting in more pronounced softening behavior than under drained conditions [58,120,124]. Under high confining pressure, undrained shear specimens typically exhibit strain hardening, as the state of the specimen gradually approaches or moves above the CSL, leading to more contractive behavior. Under high stress levels, pore pressure evolution is also influenced by particle rearrangement and particle breakage [124]. For loose tailings, undrained shearing generally exhibits contractive behavior and tends to generate positive pore pressure. Depending on the initial state and confining pressure, the stress–strain response may exhibit strain hardening, ideal plasticity, or strain softening accompanied by significant effective stress reduction [29]. Fotovvat and Sadrekarimi [39] found in their CU tests on gold tailings that loose specimens exhibited pronounced strain softening under low confining pressure, with the degree of softening diminishing as the confining pressure increased. Furthermore, the sample preparation method also influences the stress–strain relationship. Mmbando et al. [38] found that iron tailings samples prepared using the slurry deposition (SD) method exhibited a stronger tendency towards strain hardening than those prepared using the moist tamping (MT) method.

It is worth noting that the tailings involved in the aforementioned studies were primarily sandy tailings or low-plasticity silty tailings. Such tailings have relatively coarse grains and low plasticity, and their mechanical behavior has been systematically characterized within the classical soil mechanics framework. However, with continuous advances in mineral processing techniques and declining ore grades, there is a marked trend towards finer grain sizes in tailings, and the proportion of highly plastic clay tailings is increasing. Compared with sandy tailings, the mechanical response mechanisms of clay tailings are more complex. Currently, research into the triaxial shear behavior of highly plastic clay-rich tailings remains very limited, with a lack of systematic experimental data. This research gap restricts the universality of tailings dam stability assessment methods and also hampers the safe design and risk management of tailings dams with high fine-grain content. Therefore, conducting research into the mechanical properties of highly plastic clay-rich tailings and elucidating their response patterns under complex stress conditions holds significant theoretical and engineering value.

6.1.2. Strength Parameters

The shear strength of tailings is commonly described using the Mohr–Coulomb criterion, with the angle of friction ($\varphi$) and cohesion (c) being governed by various factors such as mineralogical composition, particle shape, particle size distribution and density. As tailings particles are often angular in shape (resulting from crushing and grinding processes), their $\varphi$ is typically high. As shown in

Table 7. The peak strength parameters of different types of tailings.

Reference	Tailings Type	c (kPa)	$\mathit{\varphi }$ (°)
Qiu and Sego [18]	Copper	0	34
	Gold	0	33
Islam [128]	Coal	38.9	22.2
	Red mud	26.3	34.4
	Gold	10.7	36.6
Hu et al. [16]	Iron coarse	8.8	41
	Iron fine	7.4	32
	Copper coarse	32	40
	Copper fine	0	38

Notes: The table shows peak $\varphi$; for the ${\varphi }_{cs}^{\prime }$, please refer to the values given in the text.

, there are significant differences in $\varphi$ among different tailings. Qiu and Sego [18] reported $\varphi$ of 34° and 33° for copper and gold tailings, respectively, with a c of 0 in both cases. Tests on coal tailings, red mud, and gold tailings conducted by Islam [128] indicated that coal tailings exhibit a lower $\varphi$ (22.2°) but higher c (38.9 kPa) due to their elevated clay mineral content, whereas the $\varphi$ of gold tailings can reach 36.6°. Experiments by Hu et al. [16] showed that coarse-grained iron tailings had a $\varphi$ as high as 41°, fine-grained iron tailings 32°, and coarse and fine copper tailings 40° and 38°, respectively. These data indicate that the $\varphi$ of tailings varies widely (approximately 22–41°), with coarse-grained, angular particles typically contributing to a higher $\varphi$, whilst an increase in fine-grained content may reduce it. It should be noted that all values in Table 7 correspond to peak strength parameters. Peak $\varphi$ is influenced by density and dilatancy, and therefore should not be directly compared across different studies without considering testing conditions.

It should be noted that the Mohr–Coulomb criterion typically describes the failure state of soil at peak strength, and its $\varphi$ varies with initial density. In contrast, the critical state friction angle (${\varphi }_{cs}^{\prime }$) is an intrinsic friction parameter attained by tailings under conditions of large deformation and constant shear stress, excluding the effects of density and dilation; consequently, the values are more stable. Carrera et al. [124] found in their study of Stava tailings that the ${\varphi }_{cs}^{\prime }$ for different sand–silt mixtures ranged between 34° and 36°; Tests by Li and Coop [120] on Panzhihua iron tailings indicated that the ${\varphi }_{cs}^{\prime }$ values for the UB, MB and PO tailings were 34.8°, 33.7° and 34.6° respectively, suggesting that when mineral composition is similar, the influence of gradation on the ${\varphi }_{cs}^{\prime }$ is limited. A statistical analysis by Torres-Cruz and Santamarina [66] of various non-plastic tailings showed that the ${\varphi }_{cs}^{\prime }$ ranged around 33° ± 2°, with angular particles exhibiting a significantly higher $\varphi$ than rounded particles. These ${\varphi }_{cs}^{\prime }$ values are generally lower than the aforementioned Mohr–Coulomb peak $\varphi$, reflecting the contribution of dilation to peak strength. For intrinsic comparison of tailings shear behavior (independent of density and stress history), the ${\varphi }_{cs}^{\prime }$ is recommended, as summarized in the text above. At high stress levels, the strength envelope of tailings exhibits distinct non-linear characteristics, reducing the applicability of the linear Mohr–Coulomb criterion. Zhang et al. [126] proposed using a power-law strength criterion, $\tau =0.813{\sigma }^{0.8088}$, to describe the strength behavior of copper tailings under confining pressures ranging from 0 to 5 MPa, with a better fit than the linear Mohr–Coulomb criterion.

6.2. Anisotropy and the Effects of Stress Paths

6.2.1. Structural Anisotropy

Due to processes such as hydraulic deposition or stratigraphic compaction, tailings exhibit significant structural anisotropy, and their mechanical response is closely related to the angle between the principal stress direction and the plane of deposition. Chen et al. [129] conducted shear tests on sandy tailings with different deposition orientations using a modified direct shear tester and found that the variation in the peak friction angle with respect to the angle between the shear plane and the deposition surface could reach 10°, being maximum when the shear plane was perpendicular to the deposition surface and minimum at an angle of 45°. This anisotropy primarily stems from the oriented arrangement of the long axes of the particles and the non-uniform spatial distribution of the particle contact normal. However, the degree of anisotropy in residual strength is relatively low, indicating that some of the initial texture has been re-oriented under large deformations. Chen et al. [130] systematically investigated the progressive failure process of tailings containing fine-grained interlayers with varying inclinations. They found that when the inclination was 60°, the specimen failed due to shear slip along the interlayer, whereas at smaller inclinations, buckling failure occurred. Hou et al. [131] observed the same phenomenon and, using a digital imaging system, demonstrated that strain localization first originated at a specific point within the interlayer before gradually propagating throughout the entire interlayer region.

6.2.2. The Effects of Stress Paths

The stress path has a significant influence on the liquefaction triggering and shear strength of tailings. Fotovvat and Sadrekarimi [39] conducted triaxial compression and extension tests on gold tailings and found that specimens sheared in the same direction as the K₀ consolidation (CKCU) exhibited the highest strength, whereas those consolidated in compression and sheared in extension (CKEU) showed the lowest strength, with the strength ratio differing by several times. Under the constant shear stress drainage (CSD) path, even under drained conditions, tailings may undergo instability failure when the effective stress decreases to a certain level, and the trigger point for instability can be identified using the second-order method (${\mathrm{d}}^2\mathrm{W}=\mathrm{d}{\sigma }^{\prime }\mathrm{d}\epsilon <0$) [39,132]. Similar instability phenomena were observed by Riveros and Sadrekarimi [133] who noted that the friction angle at instability was consistent with the yield friction angle under undrained conditions.

Hollow-cylinder torsional shear tests further confirm that continuous rotation of the principal stress direction angle exacerbates the instability of the tailings. Fanni et al. [58] demonstrated that when the major principal stress direction angle is 45° and the Lode angle is 0°, the undrained yield strength of tailings is significantly lower than that under triaxial compression conditions. The initial shear stress positively contributes to the yield strength, whereas stress reversal (e.g., compression consolidation followed by extension shearing) considerably reduces it. These findings are of great significance for understanding the local instability mechanisms of tailings dams under complex stress states.

6.3. Critical States and Transitional Behavior

6.3.1. Classical Critical State Theory

CSSM provides a fundamental theoretical framework for understanding the state-dependent mechanical behavior and static liquefaction susceptibility of tailings materials. Recent studies have increasingly adopted the CSSM framework to interpret the stress–strain response, contractive/dilative behavior, pore pressure evolution, and liquefaction susceptibility of tailings materials under different density and stress conditions [29,35,39,120,134]. Representative studies on iron tailings [120], gold tailings [39], and Stava tailings [124] consistently indicate that the position of the initial state relative to the CSL governs the contractive or dilative response of tailings during shearing. Within the CSSM framework, the mechanical response of tailings is primarily controlled by the relationship between the initial state and the CSL. Within the CSSM framework, tailings typically exhibit a unique CSL in both the p-q and e-log p’ planes. Li and Coop [120] found that the UB, MB, and PO iron tailings from Panzhihua each exhibit a unique CSL(Figure 10a). The shape and position of the CSL on the e-lop p’ have significant implications for evaluating the state-dependent behavior and liquefaction susceptibility of tailings materials. As shown in Figure 10a, the CSL of tailings on the e-log p’ plane is typically not a straight line but exhibits a distinct curvature. It is relatively gentle in the low-stress region, gradually steepens with increasing stress, and tends to become parallel to the normal compression line (NCL) in the high-stress region. Triaxial tests conducted by Consoli et al. [35] on silty iron tailings within a wide confining pressure range of 0.075–120 MPa revealed that the CSL exhibited a double-curved characteristic (Figure 10b). The curved CSL shape indicates that the state-dependent behavior of tailings varies significantly with stress level. In the low-stress region, even small changes in void ratio may result in substantial variations in the state parameter ($\psi$), thereby strongly influencing the contractive tendency and liquefaction susceptibility of the material. In contrast, under high stress levels, the CSL becomes steeper and the material behavior tends to become less sensitive to void ratio variations, resulting in relatively lower liquefaction susceptibility. Several studies have further suggested that particle breakage under high confining pressure may influence the evolution of the CSL and the state parameter characteristics of tailings materials [35,120,134].

The position and slope of the CSL are governed by the tailings gradation, PI and minimum e. A statistical analysis of 53 tailings materials by Macedo and Vergaray [29] indicated that the intercept $\mathrm{\Gamma }$ of the CSL shows a good linear correlation with LL × G_s (R² = 0.6), whilst the slope $\lambda$ correlates with PI. Torres-Cruz and Santamarina [134] found that the intercept ${\mathrm{\Gamma }}_{100}$ of the CSL on the e-log p’ plane is positively correlated with the minimum e, and that tailings with a mixed gradation (fine-grain content of approximately 30%) exhibit the lowest ${\mathrm{\Gamma }}_{100}$ and the minimum e values, indicating that they are the most compact.

6.3.2. Transitional Behavior

Although the classical CSSM assumes a unique CSL, recent studies have found that some tailings exhibit transitional behavior, meaning that the positions of their NCL and CSL depend on the initial density of the specimen or the sample preparation method [38,43,124]. Specimens with different initial specific volumes exhibit non-convergent behavior during compression and shearing; even when high stresses are applied, their compression paths and critical state points fail to converge onto a single line, making it impossible to define a unique CSL. Cartwright et al. [135] noted that transitional behavior is widespread in tailings; however, as researchers typically do not conduct a sufficient number of tests on specimens with different initial densities, this phenomenon is rarely identified. The causes of transitional behavior remain unclear at present.

Mmbando et al. [38] conducted triaxial tests on viscous iron tailings and found that specimens prepared using the MT and SD methods exhibited two distinctly different CSLs on the e-log p’ plane. Tests by Nayanthara et al. [43] on Australian gold tailings also confirmed the existence of transitional behavior, with the mechanical response of SD specimens being closer to that of in situ specimens, whilst MT specimens exhibited more pronounced strain softening. This suggests that the fabric differences introduced by the specimen preparation methods were not fully eliminated during the shearing process, resulting in non-unique CSL. However, as shown in Figure 10a, despite the use of three different specimen preparation methods for the three tailings, Li and Coop [120] observed a unique CSL for each material. This indicates that the sample preparation method is not the direct cause of the transitional behavior. The occurrence of transitional behavior is also generally considered to be related to the gradation and fine-grain content of the tailings. Carrera et al. [124] found that the compression curves of a 50% sand–50% silt mixture in Stava tailings exhibited the greatest dispersion and the strongest tendency towards transition behavior. The study on Panzhihua iron tailings by Li and Coop [120] revealed that, although none of the three tailings exhibited pronounced transitional behavior, the compression curves of the MB and UB tailings converged more slowly, requiring confining pressures in excess of 20 MPa to converge to a unique NCL.

The microstructural mechanism of transition behavior is attributed to the presence of stable fabric within the tailings. SEM analysis by Nayanthara et al. [43] indicated that the agglomerated structures produced by the MT method and ‘edge-to-face’ particle contacts are the primary causes of transition behavior, whereas the ‘face-to-face’ particle arrangement produced by the SD method is closer to the in situ microstructure. Okewale and Grobler [60] also found in their study of gold tailings that the non-convergent compression behavior of the specimens was related to heterogeneous microstructures (cluster-like structures formed by particle agglomeration) rather than particle breakage. This explains why the influence of initial density on tailings behavior persists even when high stresses are applied. However, SEM results by Li and Coop [120] showed that the microstructural differences produced by different preparation methods were minimal. The authors suggest that this discrepancy may stem from differences in particle shape and mineralogical composition.

The presence of transitional behavior will affect the assessment of the engineering properties of tailings. As shown in Figure 11, when transitional behavior is present in tailings, the CSL determined using a single sample preparation method may incorrectly predict the in situ contractive or dilative behavior of the tailings. For example, a given in situ condition might be classified as contractive based on the MT-CSL, whereas it would be classified as dilative based on the SD-CSL. Such uncertainty has significant implications for the stability assessment of tailings dams and decisions regarding reinforcement. Therefore, an appropriate sample preparation method should be selected based on the actual deposition pattern of the tailings, or at least multiple methods should be employed for comparative validation. In addition, research should be conducted to elucidate the specific causes of the transitional behavior observed in tailings.

6.4. Shear Behavior Under Special Conditions

6.4.1. Cyclic Loading and Liquefaction

Under cyclic loading conditions such as those caused by earthquakes, the cumulative pore pressure and liquefaction behavior of tailings are key factors in seismic response analysis. Cyclic triaxial tests conducted by Hu et al. [16] on iron and copper tailings indicated that the cyclic stress ratio (CSR), consolidation pressure and relative density are the primary factors governing cyclic strength. The pore pressure development pattern of fine-grained tailings differs from the Seed model and is better described by the inverse tangent function model modified by Zhang et al. [136]. Huang et al. [137] found in cyclic triaxial tests on tailings sand from the center-line method that the maximum dynamic shear modulus of underflow tailings sand increased with rising fine-grain content, whereas the opposite was true for overflow tailings sand; the normalized dynamic shear modulus–dynamic strain curve was higher than the average values for sandy soils suggested by Seed, indicating that underflow tailings sand from the center-line method possesses better resistance to deformation. Subsequently, Huang et al. [138] further proposed a modified three-parameter pore pressure stress model, which better describes the ‘rising–steady–rising’ trend of the pore pressure ratio in sandy tailings as the vibration frequency ratio changes. Regarding clay tailings, Liu et al. [139] identified four distinct stages in pore pressure development: ‘slow growth–rapid growth–structural failure–complete liquefaction’, and proposed an improved BiDoseResp pore pressure development model, the results of which demonstrated excellent description of this four-stage trend.

It is worth noting that the method of sample preparation also influences the dynamic response of tailings. In cyclic tests on hydraulically deposited fine-grained iron tailings, Medina et al. [59] found that samples prepared using the MT method and the SD method exhibited markedly different responses under cyclic loading; the cyclic resistance of SD samples was significantly lower than that of MT samples, and the pore pressure rose more rapidly. This result indicates that the appropriate sample preparation method should be selected based on the actual deposition mode of the tailings to accurately assess their dynamic behavior. Regarding the transitional behavior exhibited by the tailings described in Section 6.3.2, Cartwright et al. [135] conducted a comparative study on the cyclic liquefaction characteristics of transitional and non-transitional tailings. The results indicate that the cyclic response of transitional tailings is primarily controlled by the stress history, with initial density having a minor influence; whereas the cyclic resistance of conventional tailings is significantly dependent on initial density. During cyclic loading, even when subjected to substantial strain, the influence of the stress history on the cyclic resistance of transitional tailings specimens remains difficult to eliminate.

6.4.2. The Effects of Freeze–Thaw Cycles

In cold regions, tailings dams are subjected to the severe challenges of long-term freeze–thaw cycles, resulting in a deterioration of their mechanical properties and posing a threat to the long-term stability of the tailings ponds. Cyclic triaxial tests on zinc tailings conducted by Liu and Liu [140] showed that after 1, 5 and 10 freeze–thaw cycles, the strength, elastic modulus and cyclic resistance of the tailings all decreased, whilst the pore pressure at failure increased; after approximately 5 to 10 freeze–thaw cycles, the degradation effect tended to stabilize. Further research revealed that the c of the tailings decreased significantly after freeze–thaw cycles, whilst the $\varphi$ remained largely unchanged, indicating that freeze–thaw action primarily weakens strength by disrupting the bonding between particles [141]. Sun et al. [142] observed via scanning electron microscopy that freeze–thaw cycles disrupted the internal bonding structure of the tailings, increasing the number of voids and microcracks, thereby leading to a decline in macroscopic mechanical properties. Based on the theory of binary media and the linear comparative composite method, Sun et al. [142] established a micro-mechanical strength criterion that accounts for the effects of freeze–thaw cycles, incorporating freeze–thaw parameters into the evolution equation for the volume fraction of friction elements, thereby providing a theoretical tool for the stability assessment of tailings dams in cold regions.

7. Future Work

This paper presents a review of recent progress in the physical properties and geotechnical behavior of tailings. As an artificial soil, the geotechnical properties of tailings have already been the subject of considerable research. However, several key challenges and pressing scientific issues remain to be addressed. Future work could be conducted in the following areas:

• With the continuous advancement of mineral processing technologies and the steady decline in the grade of raw ore, tailings are exhibiting a marked trend towards finer grain sizes. Compared to the silt tailings that have been the primary focus of existing research, clayey tailings resulting from this refinement display significant differences in their geotechnical behavior. However, research on such clayey fine-grained tailings remains insufficient. It is recommended that future studies investigate their permeability, consolidation and compression, and shear behavior.
• As discussed in Section 6.3.2, transitional behavior is widespread in tailings, which severely impacts the assessment of their liquefaction behavior. At present, the specific causes of this transitional behavior in tailings remain unclear. Subsequent research should aim to elucidate the key controlling factors underlying the formation of transitional behavior, thereby enabling a correct assessment of tailings liquefaction behavior and providing a reliable theoretical basis for the stability analysis of tailings dams.
• Tailings constitute a three-phase medium comprising solid, liquid and gas phases. However, most existing studies on the shear behavior of tailings are based on saturated specimens, overlooking the actual influence of the gas phase on soil mechanical behavior. Future research should prioritize the study of the shear characteristics of tailings under unsaturated conditions, systematically investigating the mechanisms by which changes in matrix suction influence the strength, deformation and pore pressure response of tailings, with a view to establishing a mechanical analysis model for tailings that more closely reflects engineering practice.

8. Conclusions

This paper presents a review of the physical properties and geotechnical behavior of tailings. Overall, the reviewed studies indicate that the geotechnical behavior of tailings is controlled by the combined effects of particle characteristics, depositional state, stress level, and microstructural evolution, which jointly influence their permeability, compressibility, and shear behavior. The principal conclusions drawn from this review are as follows:

• The particle size distribution of tailings exhibits distinct characteristics of hydraulic grading within the tailings pond, consisting primarily of silt-sized particles; their mineral and chemical composition is dominated by quartz, hematite and silicates, but there is significant spatial variability both between different tailings and at different locations within the same tailings pond. Consequently, each tailings dam project requires a separate and appropriate analysis. At the same time, there is currently a marked trend towards finer-grained tailings; in the future, research should be conducted into the seepage, consolidation, compression and shear behavior of clayey fine-grained tailings.
• The permeability behavior of tailings is influenced by numerous factors. With regard to particle size distribution, there is a threshold for the content of fine particles; prior to this threshold, permeability deteriorates as the fine particle content increases, whilst beyond the threshold, permeability gradually stabilizes. Furthermore, high pressure significantly alters the permeability characteristics of tailings. In addition, chemical and biological processes must be taken into account during the design of tailings dam impermeabilization measures to prevent the formation of blockages and rapid seepage pathways.
• With regard to the consolidation and compression behavior of tailings, when tests are conducted using slurry specimens, the compression curves exhibit non-linear or piecewise linear characteristics, which are primarily described using the modified Gibson model. As the slurry method more accurately simulates the initial deposition state of tailings in tailings dam engineering, it more closely approximates actual engineering conditions. When tests are conducted using remolded specimens, although the compression curve can be linearly described using traditional compression indices, it similarly exhibits piecewise linear characteristics under high pressure. These observations demonstrate that the compressibility of tailings is strongly dependent on depositional state and stress level, both of which should be considered when extrapolating laboratory results to field conditions.
• The shear behavior of tailings is significantly influenced by confining pressure, drainage conditions, anisotropy and stress paths. The presence of transitional behavior means that CSL determined using a single sampling method may lead to an incorrect assessment of the shear dilation/contraction characteristics of in situ tailings, thereby affecting the assessment of liquefaction risk. Future research should focus on elucidating the key controlling factors underlying the formation of transitional behavior, in order to enhance the reliability of tailings dam stability assessments. These findings emphasize the importance of considering state-dependent behavior when evaluating tailings strength, deformation characteristics, and liquefaction susceptibility.

The findings of this review may provide useful references for the geotechnical characterization, stability assessment, liquefaction evaluation, and engineering design of tailings storage facilities.