Transforming Waste Red-Bed Mudstone into Sustainable Controlled Low-Strength Materials: Mix Design for Enhanced Engineering Performance

Abstract

Red-bed mudstone from civil excavation is often treated as waste due to its poor water stability and tendency to disintegrate. This study proposes a sustainable approach for its utilization in controlled low-strength material (CLSM) by blending it with cement and water. Laboratory tests evaluated the fresh properties (i.e., flowability, bleeding rate, setting time, and subsidence rate) and hardened properties (i.e., compressive strength, drying shrinkage, and wet–dry durability) of the CLSM. The analysis focused on two main parameters: cement-to-soil ratio (C/S) and water-to-solid ratio (W/S). The results show that increasing W/S significantly improves flowability, while increasing C/S also contributes positively. Flowability decreased exponentially over time, with an approximately 30% loss recorded after 3 h. Bleeding and subsidence rates rose sharply with higher W/S but were only marginally affected by C/S. To meet performance requirements, W/S should be kept below 52%. In addition, the setting times remained within 24 h for all mixtures tested. Compressive strength showed a negative correlation with W/S and a positive correlation with C/S. When C/S ranged from 8% to 16% and W/S from 44% to 56%, the compressive strengths ranged from 0.3 MPa to 1.22 MPa, meeting typical backfilling needs. Drying shrinkage was correlated positively with water loss, and it decreased with greater C/S. Notably, cement’s addition significantly enhanced water stability. At a C/S of 12%, the specimens remained intact after 13 wet–dry cycles, retaining over 80% of their initial strength. Based on these findings, predictive models for strength and flowability were developed, and a mix design procedure was proposed. This resulted in two optimized proportions suitable for confined backfilling. This study provides a scientific basis for the resource-oriented reuse of red-bed mudstone in civil engineering projects.

1. Introduction

Red-bed mudstone presents significant challenges in civil engineering due to its tendency to disintegrate and soften when exposed to water, leading to a considerable reduction in strength and compromised engineering performance. Construction projects in regions abundant in red-bed mudstone often encounter issues such as slope instability, subgrade deformation, and pavement cracking [1]. In China, red-bed mudstone is widely distributed across the southeastern, northwestern, and southwestern regions with the Sichuan Basin, covering an area of 165,000 km², and serving as a representative region in the southwest [2]. The humid and wet climate in this area exacerbates engineering failures, as demonstrated by the Bazhong-to-Guang’an Expressway project, which experienced 192 landslides attributed to red-bed mudstone, with remediation costs exceeding 55 million USD [2]. Internationally, areas such as western Kyushu (Japan), northwestern India, the western United States, northern Chile, central and southern South Africa, and Western Australia also face similar challenges associated with red-bed mudstone [1,3].

To mitigate the disaster risks associated with red-bed mudstone, excavated soil is often discarded during infrastructure projects, with sand and gravel commonly used as alternative backfill materials [4,5]. However, this practice significantly increases transportation costs and conflicts with modern environmental standards. Disposing of red-bed mudstone waste requires additional land, while excessive extraction of natural sand and gravel accelerates vegetation’s degradation and may trigger landslides [6,7,8]. Another issue arises in backfilling projects with restricted working space—such as pipe galleries, bridge abutments, culvert backs, and retaining wall zones—where compacting traditional sand and gravel becomes difficult [9]. In response to these challenges, converting red-bed mudstone waste into controlled low-strength material (CLSM) presents a promising solution. CLSM typically consists of fine aggregates, cement, and water, with admixtures added as needed. It offers high flowability and can self-level under its own weight, requiring little to no vibration for adequate densification [9]. The American Concrete Institute (ACI) defines CLSM as a self-compacting, flowable material that requires no mechanical compaction. As early as the 1960s, the United States Bureau of Reclamation (USBR) employed CLSM for backfilling pipeline trenches, demonstrating its effective backfilling performance. Compared with conventional sand and gravel, CLSM is more suitable for backfilling in confined areas and provides notable economic advantages. First, it reduces material costs by decreasing reliance on natural aggregates. Second, it lowers transportation costs by reducing both the removal of waste soil and the delivery of new materials. Finally, due to its self-compacting nature, CLSM eliminates the need for layer-by-layer compaction, thereby reducing the labor input and reliance on equipment [10,11,12,13]. In recent years, CLSM has played an increasingly significant role in the resource-efficient utilization of solid waste, driven by infrastructure development [14]. Numerous studies have explored the incorporation of industrial byproducts—such as dewatered sludge, waste glass, incineration bottom ash, and rubber particles—as aggregates in CLSM for infrastructure backfilling. These practices not only promote resource efficiency but also offer substantial ecological and environmental benefits, enhancing the sustainability of construction projects [15,16,17,18]. In addition, several studies have investigated the use of waste soil. Findings show that employing waste soil as an aggregate for on-site CLSM backfilling reduces dependency on natural aggregates and alleviates land occupation issues caused by soil disposal. This approach supports environmentally sustainable construction and provides a practical solution for backfilling in space-constrained areas [19,20]. For instance, Qian et al. [10] implemented in situ backfilling with low-plasticity clay from construction sites in Shanghai, while Zhao et al. [21] successfully produced CLSM with stable performance using silty clay generated during subway construction.

However, most existing studies remain focused on soils with favorable engineering properties, while investigations into lower-quality materials—particularly red-bed mudstone—are relatively scarce [22]. This is primarily due to the tendency of red-bed mudstone to disintegrate and soften upon exposure to water, resulting in significant strength reduction and poor durability, which pose multiple challenges for mix design [20]. In particular, a pronounced trade-off exists between the flowability of red-bed mudstone and other engineering performance parameters. As a key property for ensuring self-compaction, flowability is especially critical in the design of CLSMs. Hu et al. [4] investigated the effects of water content and polycarboxylate superplasticizer content on the flowability and bleeding behavior of a CLSM prepared with red-bed mudstone from engineering excavation. Their findings indicated that achieving adequate flowability requires a higher water content, which often leads to bleeding rates that exceed design thresholds. The American Concrete Institute (ACI) has also highlighted the considerable difficulties associated with mix design when using engineering waste soils. Such soils typically possess small particles and large specific surface areas, demanding significantly more water to produce a fluid mixture that meets the flowability requirements. However, excessive water content can result in elevated bleeding, prolonged setting times, and increased void ratios, all of which hinder cement hydration and reduce the overall strength of the hardened material. In particular, the poor water stability of red-bed mudstone severely compromises its strength and durability. Due to these limitations, studies involving red-bed mudstone for CLSM production remain scarce. Therefore, comprehensive research is needed to evaluate the bleeding rate, setting time, subsidence characteristics, and the strength and durability of hardened CLSM made from red-bed mudstone, in order to fully assess its feasibility for engineering use [4,5,9].

This study produced a CLSM using waste red-bed mudstone combined with cement and water. Tests on the fresh mixtures included flowability, bleeding rate, subsidence rate, and setting time, to investigate the effects of the cement-to-soil ratio and water-to-solid ratio on the fresh properties of red-bed mudstone-based CLSM. Compressive strength tests were subsequently conducted to evaluate the mechanical performance of the hardened material. Based on the experimental data and normalization techniques, predictive models were developed for flowability and compressive strength. Considering the poor water stability of red-bed mudstone, additional tests were carried out to assess residual strength under wet–dry cycles and the drying shrinkage rate, thereby evaluating the material’s durability. A mix design process was established for CLSMs incorporating red-bed mudstone, through which suitable mix proportions were proposed for backfilling behind abutments, culverts, retaining walls, and subgrades. The findings provide a scientific basis for the feasibility and mix design of CLSMs prepared from engineering waste soils, facilitating the conversion of environmental liabilities into valuable construction resources.

2. Materials and Methods

2.1. Materials

The red-bed mudstone was sourced from a road construction site in suburban Chengdu, Sichuan Province. The X-ray diffraction (XRD) pattern (Figure 1) indicated that the primary mineral components of the red-bed mudstone were quartz, muscovite, calcite, and chlorite. In accordance with JTG D3430-2020 [23], basic physical property tests were conducted. The results are presented in

Table 1. Index properties of red-bed mudstone.

Index Property	Value	Unit
Appearance	Maroon colored	-
Specific gravity	2.696	-
Maximum dry density	1.95	g/cm3
Optimal moisture content (OMC)	10.65	%
Liquid limit (LL)	31.5	%
Plasticity index (PI)	14.4	-
Particles < 75 μm	78.9	%
Free swelling ratio	14.92	%
Standard moisture absorption	0.86	%

and Figure 2 (particle size distribution curve). With a liquid limit below 50% and a plasticity index of 14.4, the red-bed mudstone was classified as low-liquid-limit clay according to the standard (ASTM D2487-17) [24].

The cement used in the experiment was PO 42.5 ordinary Portland cement, supplied by Yangchun Cement Co., Ltd. (Zhucheng, China), which is widely used domestically. It had an initial setting time of 182 min and a final setting time of 249 min. The main chemical composition is listed in

Table 2. Chemical composition of cement.

Composition	SiO2	Al2O3	Fe2O3	CaO	MgO	SO3	Cl−
Content (%)	24.99	8.26	4.03	51.42	3.71	2.51	0.043

. The particle size distribution curve, measured using a Malvern Mastersizer 2000 laser particle size analyzer, is presented in Figure 2. Additionally, tap water was used throughout the experiment.

2.2. Mix Proportions

In the experiment, the following ratios were defined:

(1)

The ratio of the mass of cement to the mass of soil particles, termed the cement-to-soil ratio (C/S).

(2)

The ratio of the mass of water to the total mass of solid particles, referred to as the water-to-solid ratio (W/S).

The C/S is a key parameter governing the compressive strength of the CLSM, and its minimum value should be determined based on strength design requirements. According to DBJ51/T 188-2022 [25], C/S is generally limited to 25% to balance strength and cost. The W/S primarily affects the flowability of the CLSM. As soil is the main constituent, the lower limit of W/S should not fall below the liquid limit. For the red-bed mudstone used in this study, the W/S should not be less than 32%. Additionally, to mitigate bleeding, the upper limit of W/S must be controlled. Prior to formal testing, preliminary trials were conducted to define the appropriate mix parameter ranges. The final selected ranges were as follows: C/S from 4% to 16%, in 4% intervals; and W/S from 44% to 56%, in 4% intervals. The mix proportions of the specimens are listed as IDs 1-16 in

Table 3. Mix proportions for soil-based CLSM mixes.

ID	C/S (%)	W/S (%)	Dry Red-Bed Mudstone (kg/m3)	Cement (kg/m3)	Water (kg/m3)
1	4	44	1189	48	576
2	4	48	1133	45	598
3	4	52	1082	43	619
4	4	56	1036	41	638
5	8	44	1148	92	555
6	8	48	1093	87	577
7	8	52	1044	84	597
8	8	56	999	80	615
9	12	44	1109	133	537
10	12	48	1056	127	558
11	12	52	1009	121	577
12	12	56	965	116	594
13	16	44	1073	172	519
14	16	48	1022	163	539
15	16	52	975	156	558
16	16	56	933	149	575

and were used to evaluate the feasibility of producing CLSM using red-bed mudstone, cement, and water.

2.3. Test Methods

2.3.1. Types of Tests

The experimental procedure is outlined in Figure 3. Cement and water were first combined in a mortar mixer for 60 s. The red-bed mudstone was then added, and the mixture was further stirred for an additional 180 s. After mixing, the CLSM was tested for flowability, bleeding rate, setting time, and subsidence rate. A portion of the mixture was poured into molds, labeled, and cured in a standard curing room at 20 ± 2 °C with approximately 95% relative humidity. After the designated curing period, the samples were tested for compressive strength, drying shrinkage rate, and water stability.

2.3.2. Testing of Fresh Mix Properties

(1)

Flowability test

Flowability is a critical design parameter for CLSMs, as higher flowability enhances self-compaction during construction. In accordance with the Japanese standard JHS A313-1992 [26], a cylindrical container (80 mm in diameter and height) was filled with the fresh CLSM mixture. The container was then slowly lifted over a 5 s period, allowing the material to spread freely. Once the flow ceased, the diameters in two perpendicular directions were measured, and the average value was recorded as the flowability (f) of the CLSM. This testing method is consistent with the procedure proposed in ASTM D6103 [27], with the main difference being that the cylindrical mold used in ASTM D6103 has a larger volume.

(2)

Bleeding rate test

After pouring, the solid particles in the CLSM settle under the influence of gravity, resulting in surface bleeding. To assess this, the bleeding rate (B_r) test was conducted, following JTG 3420-2020 and ASTM C940-16 [23,28]. A 1000 mL graduated cylinder, filled with 1000 mL of fresh CLSM, was covered with plastic parafilm to prevent evaporation, as shown in Figure 3c. During the first hour, bleed water on the specimen’s surface was extracted every 10 min using a pipette and transferred to a beaker. After 1 h, water was collected every 30 min until bleeding ceased. The bleeding rate was then calculated as the ratio of the total mass of extracted water to the initial water content in the CLSM mixture. It is important to note that engineering and testing standards may have slight variations in their specific requirements for the bleeding rate. To maintain scientific rigor, this study kept the bleeding rate below 5% [5,9,29,30,31,32,33].

(3)

Setting time test

The setting time (T_s) is a critical parameter for planning CLSM construction [5,34,35]. According to the Chinese standard JGJ/T70-2009 [36], T_s is determined using a ZKS-100 setting time tester. The CLSM mixture is placed in a container and allowed to stand for 2 h before testing begins, with measurements taken every 30 min. During each test, a needle with a cross-sectional area (A_p) of 30 mm² is inserted 25 mm into the CLSM, and the static pressure (N_p) is recorded. The penetration resistance (f_p) is then calculated using Equation (1). When f_p reaches 0.3 MPa, measurements are taken every 15 min until f_p reaches 0.7 MPa. The time corresponding to f_p = 0.5 MPa, derived from the plotted curve of f_p versus time, is defined as T_s.

$$f_p={\displaystyle \frac{N_p}{A_p}}$$

(1)

(4)

Subsidence rate test

The sedimentation of solid particles and air discharge following CLSM mixing can contribute to increased subsidence. To assess this, the subsidence rate (h_sr) of the CLSM was measured [31,32,37,38]. The CLSM mixture was placed in a cylindrical container with a capacity of 1000 mL and marked sidewalls. The container was sealed with plastic film and a lid. The height of the mixture was periodically recorded, and h_sr was calculated using Equation (2):

$$h_{sr}={\displaystyle \frac{h_0-h_s}{h_0}}$$

(2)

where h_sr (%) is the subsidence rate, h₀ (mm) is the initial height of the mixture, and h_s (mm) is the stable subsidence height of the mixture.

2.3.3. Testing of Hardened Mix Properties

(1)

Compressive strength test

The CLSM was mixed and cast into cubic molds with a side length of 70.7 mm, and then demolded after 28 d of curing in a standard curing room. In accordance with the Chinese standard JGJ/T 233-2011 [39], the 28 d unconfined compressive strength (q_u0) test was performed using a testing machine specifically designed for lime-stabilized soil.

(2)

Drying shrinkage rate test

The high water content in CLSMs can lead to drying shrinkage and cracking [9,10,37]. To investigate this, the drying shrinkage rate was tested, following the Chinese standard JGJ/T70-2009 [36]. Freshly mixed CLSM was poured into prismatic molds (40 mm × 40 mm × 160 mm). After 7 d of curing, the initial length (L₀) of the specimens was measured in a standard curing room. The specimens were then transferred to a controlled environment maintained at (20 ± 2) °C and (60 ± 5) % relative humidity. Length measurements were taken on days 14, 21, 28, and 56, and the drying shrinkage rate was calculated using Equation (3). The specimens’ mass was also recorded during each measurement to determine the water loss rate, providing supplementary data for analyzing drying shrinkage.

$${\epsilon }_{at}={\displaystyle \frac{L_0-L_t}{L-L_d}}$$

(3)

where ε_at (%) is the drying shrinkage rate, L₀ (mm) is the length of the specimen after curing for 7 d, L (160 mm) is the length of the molds, L_d (20 ± 2 mm) is the length of the shrink screw inserted into the mixture, and L_t (mm) is the length of the specimen after curing for t days.

(3)

Water stability test

The poor water stability of red-bed mudstone may lead to insufficient water stability in the prepared CLSM. To assess this, a dry–wet cycling test was conducted based on the ASTM D559 testing methods [40,41,42]. Specimens cured for 28 d were dried in a 45 °C electric oven for 48 h and then immersed in water at 20 °C for 24 h to complete one dry–wet cycle. For each mixture, the specimens were divided into two groups. One group was cured for 27 d, soaked in water for 24 h, and then tested for compressive strength (q_u0). The other group underwent N (1, 3, 5, 7, 9, 11, and 13) dry–wet cycles before testing, with the compressive strength recorded as q_un. The residual strength (r_qn) was calculated using Equation (4) to assess the water stability of the CLSM:

$$r_{qn}={\displaystyle \frac{q_{un}}{q_{u0}}}$$

(4)

3. Results and Discussion

In this section, the effects of varying cement-to-soil ratios (C/S) and water-to-solid ratios (W/S) on the fresh and hardened properties of the CLSM mixtures, respectively, are analyzed. Notably, the CLSM is required to exhibit adequate flowability to enable self-compaction and attain the expected strength upon setting. Accordingly, both flowability and compressive strength must strictly comply with design specifications. Moreover, the experiments revealed well-defined normalized relationships between flowability and compressive strength and the C/S and W/S. Accordingly, specimens ID1–16 were selected to systematically investigate the effects of C/S and W/S on flowability and compressive strength. In contrast, other performance parameters are not strictly standardized. For example, the setting time is generally controlled within a reasonable range, depending on specific construction scheduling requirements. During testing, it was observed that, within the studied mix design range, the effect of W/S on the setting time followed consistent trends at fixed C/S values, while the influence of C/S remained steady at fixed W/S values. Based on these observations, seven representative specimens were selected to assess the setting time response: one group with a constant C/S of 12% and W/S varying from 44% to 56%, and another group with a constant W/S of 52% and C/S ranging from 4% to 16%. These seven groups were also used to evaluate the bleeding rate, subsidence rate, drying shrinkage rate, and water stability, using a consistent analytical framework.

3.1. Properties of the Fresh Mix

3.1.1. Flowability

(1)

Initial flowability

The test results for initial flowability (f₀) are shown in Figure 4. Except for the sample with a water-to-solid ratio (W/S) of 44%, all other samples achieved flowability exceeding 180 ± 20 mm, meeting the requirement of the technical specification (JHS A 313-1992) [26]: that the self-leveling flowability should not be less than 180 mm. For different backfilling applications, flowability can be adjusted based on the working area size and the availability of vibration compaction. Additionally, it can be observed from Figure 4 that flowability increases with higher W/S and cement-to-soil ratio (C/S). Water in the sample exists as bound water on particle surfaces and free water between particles, with flowability primarily determined by the free water content. When C/S is constant and W increases, the free water content rises, leading to increased particle spacing and a reduction in cohesive and frictional forces, thereby improving the flowability. Moreover, during cement hydration, free Ca²⁺ ions react with Na⁺ and K⁺ ions on soil particle surfaces, reducing the thickness of the electric double layer. This reaction thins the bound water layer on the particle surfaces, increases the free water content, and further enhances the flowability. Therefore, increasing C/S also improves the flowability of the sample [5,43,44].

To quantitatively analyze the effect of W/S on flowability, a normalization factor X_f(W/S) for flowability was introduced [5], as defined in Equation (5). The calculation results are shown in Figure 5a, where the error bars indicate the difference between the predicted and actual values. As illustrated in Figure 5a, the fitted functional relationship between X_f(W/S) and W/S is expressed in Equation (7). Similarly, a normalization factor X_f(C/S) for f and C/S was introduced based on Figure 5b, as defined in Equation (6). The fitted relationship between X_f(C/S) and C/S is provided in Equation (8).

$$X_f(W/S)={\displaystyle \frac{f_0(C/S,W/S)-f_0(C/S,44)}{f_0(C/S,56)-f_0(C/S,44)}}$$

(5)

$$X_f(C/S)={\displaystyle \frac{f_0(C/S,W/S)-f_0(4,W/S)}{f_0(16,W/S)-f_0(4,W/S)}}$$

(6)

where f₀(16, W/S) and f₀(4, W/S) represent flowability at C/S values of 16% and 4%, respectively, while f₀(C/S, 56) and f₀(C/S, 44) represent flowability at W/S values of 56% and 44%, respectively.

By simultaneously solving Equations (5)–(8) and substituting the key data points f₀(16, 56), f₀(16, 44), f₀(4, 56), and f₀(4, 44), the expression for flowability as a function of C/S and W/S can be derived as shown in Equation (9). Figure 6 demonstrates that the regression formula aligns well with the experimental data within the ranges of W/S from 44% to 56% and C/S from 4% to 16%.

$$X_f(W/S)=8.45(W/S)-3.72$$

(7)

$$X_f(C/S)=8.43(C/S)-0.32$$

(8)

$$f_0(C/S,W/S)=2493(C/S)(W/S)+1004(W/S)-845(C/S)-307$$

(9)

(2)

Time-dependent variation in flowability

After mixing, the flowability of the CLSM at a given time t is denoted as f(t), as shown in Figure 7a. The results indicate that f(t) decreases over time, with the rate of decrease gradually diminishing. To quantitatively describe the relationship between f(t) and time t, this study introduces the flowability loss rate r_f(t), as defined in Equation (10). The r_f(t) exhibits a logarithmic relationship with time, as illustrated in Figure 7b. At 180 min, the average value of r_f(t) is approximately 30%. This aspect warrants particular attention, as the rate of flowability loss directly affects the workable time between mixing and final placement. If transportation and casting are not completed within this effective window, the material progressively loses its flowability, thereby impairing its self-compacting performance and posing challenges to both construction progress and quality. For instance, although the specimen with (W/S)48(C/S)12 satisfied the self-compacting requirement in terms of initial flowability, it fell below the required specification after 60 min. In addition, the evolution of r_f(t) over time is influenced by the water-to-solid ratio (W/S) and cement-to-soil ratio (C/S). When W/S is held constant, r_f(t) increases with higher values of C/S. For instance, at 180 min, the r_f(t) for the (W/S)52(C/S)4 mixture was 26%, while it rose to 35% for (W/S)52(C/S)16. This increase was likely due to the accelerated setting process caused by the higher cement content. Conversely, when C/S was fixed at 12%, increasing W/S from 44% to 56% resulted in a reduction in r_f(t) from 34% to 27%. This can be attributed to the delayed formation of the cementitious skeleton due to the higher water content, which slowed down the rate of flowability loss.

$$r_f(t)=[f_0-f(t)]/f_0$$

(10)

where r_f(t) is the flowability loss rate at time t, f₀ is the initial flowability, and f(t) is the flowability at time t.

3.1.2. Bleeding Rate

The bleeding rate (B_r) of the specimens is shown in Figure 8. B_r is primarily influenced by the water-to-solid ratio (W/S), with the cement-to-soil ratio (C/S) having a relatively minor effect. When C/S is 12% and W/S increases from 44% to 56%, B_r rises from 2.1% to 7.2%, exceeding acceptable engineering limits (5%) [5,9,29,30,31,32,33]. Following mixing, solid particles in the CLSM settle under gravity, and free water replaces surface solids, leading to bleeding. As W/S increases, the amount of free water in the mixture grows significantly, resulting in a pronounced increase in B_r. Excessive bleeding can form an undesirable weak layer on the surface, which may compromise the long-term mechanical performance of the CLSM. Higher flowability is often associated with increased free water content, which may elevate the bleeding rate and compromise the early-stage uniformity and structural stability of the specimen. To satisfy the design requirements for the B_r of a CLSM incorporating red-bed mudstone, the W/S should be limited to below 52%.

3.1.3. Setting Time

The setting time (T_s) test results of the specimens are summarized in Figure 9. When the cement-to-soil ratio (C/S) was 12%, increasing the water-to-solid ratio (W/S) from 44% to 56% resulted in a 7 h extension of T_s. This was due to the higher water content, which significantly delayed the formation of a cohesive skeleton through cement hydration [5,45]. Conversely, when W/S = 52% and C/S increased from 4% to 16%, Ts decreased from 19 h to 9 h. This can be attributed to the increased cement content, which intensifies the hydration reaction and accelerates the formation of the cementitious skeleton, thereby promoting the setting and hardening of the specimen. In addition, no standard specification is defined for setting time, and it should be determined based on specific construction requirements. For typical engineering applications—such as subgrade backfilling and backfilling behind abutments—a setting time within 24 h is generally considered to be appropriate, and it must not exceed 36 h. The test results indicated that, within the mix design range of W/S = 44–56% and C/S = 4–16%, the setting time of the prepared CLSM remained within 24 h [5,34,35].

3.1.4. Subsidence Rate

The subsidence rate (h_sr) test results of the specimens are shown in Figure 10. The h_sr rose rapidly during the first 12 h, with no further subsidence observed from 30 h to 28 d. This suggests that subsidence predominantly occurred before the sample had hardened. Additionally, the h_sr was less influenced by the cement-to-soil ratio (C/S) but strongly affected by the water-to-solid ratio (W/S). When C/S was 12%, an increase in W/S from 44% to 56% caused the 28 d h_sr to rise from 1% to 5%, approaching the permissible limit for typical engineering applications [38]. This was primarily due to the increase in W/S, which exacerbated the bleeding. Notably, the subsidence behavior is illustrated in Figure 11, where the subsidence of the mixture (h₀ − h_s) exceeds the bleeding height (h_W − h_s). This indicates that h_sr in CLSM is influenced not only by bleeding but also by other factors, such as air expulsion [31].

3.2. Properties of the Hardened Mix

3.2.1. Compressive Strength

The compressive strength (q_u0) test results are presented in Figure 12. The experimental results show that, except for samples with a cement-to-soil ratio (C/S) of 4%, CLSM with C/S ranging from 8% to 16% and a water-to-solid ratio (W/S) between 44% and 56% achieved a q_u0 of 0.3 MPa to 1.3 MPa. These values meet the requirements for backfilling in engineering applications such as bridge abutment backs, culvert backs, retaining wall backs, and general subgrades [25].

In addition, at a constant C/S, q_u0 decreases linearly as W/S increases. Conversely, at a constant W/S, q_u0 increases linearly with higher C/S. Cement chemically improves red-bed mudstone, which is a critical factor in determining its suitability for CLSM production. An increase in C/S enhances q_u0, primarily due to hydration reactions between cement minerals (e.g., tetracalcium aluminoferrite) and water, resulting in the formation of calcium silicate hydrate (C-S-H), calcium hydroxide (CH), and needle-shaped ettringite (Aft), as shown in the SEM images in Figure 13a [14]. Subsequently, CH reacts with SiO₂ and other active components in the red-bed mudstone via pozzolanic reactions, generating additional C-S-H. These hydration products, particularly C-S-H and Aft, interweave to form a dense network structure that effectively encapsulates soil particles, as observed in Figure 14a [5,14,31,44]. This microstructural skeleton plays a crucial role in improving the unconfined compressive strength (q_u0) by enhancing particle bonding, reducing porosity, and restricting deformation. To further validate the composition of the hydration products contributing to strength gain, Energy-Dispersive X-ray Spectroscopy (EDS) was conducted on the needle-like and network structures shown in Figure 13a and Figure 14a. The EDS mapping in Figure 13b indicates that the needle-shaped structure is likely Aft [5,14,46], while Figure 14b shows that the network contains Ca, O, Si, Al, and Fe. This elemental composition corresponds to the expected formulae of C-S-H and Aft, confirming their presence. Since both C-S-H and Aft possess cementitious and space-filling characteristics, their formation and spatial distribution directly contribute to the development of a stronger, more cohesive soil matrix, thereby explaining the observed increase in q_u0 with higher cement content [5,14,47].

The variation in q_u0 demonstrates a clear relationship with C/S and W/S. In line with the derivation method for flowability (f₀), a normalization method was employed by introducing the normalization factors X_q(C/S) and X_q(W/S) to establish an expression for q_u0 as a function of C/S and W/S. The estimation accuracy was validated within the ranges of C/S = 4–16% and W/S = 44–56%, and the final form is presented in Equation (11):

$$q_{u0}(C/S,W/S)=-17.54(C/S)(W/S)+15.20(C/S)-1.04(W/S)+0.46 {(\mathrm{R}}^2=0.978)$$

(11)

3.2.2. Drying Shrinkage Rate

The test results of the drying shrinkage rate (ε_at) of the specimens are summarized in Figure 15a, and the water loss rate (w_L) results are presented in Figure 15b. As shown in Figure 15a, all samples, except for (C/S)4(W/S)52, maintained ε_at below 1.25%, with no cracking observed.

As shown in Figure 15a, at a constant water-to-solid ratio (W/S), the shrinkage caused by the hydration reaction of cement results in an increase in the early ε_at [48]. However, between 17 and 56 d, ε_at decreases as the cement-to-soil ratio (C/S) increases. This can be attributed to reduced hydration activity after 14 d, which weakens the drying shrinkage. Furthermore, hydration products, including C-S-H and ettringite, occupy the voids between soil particles, enhancing the sample’s ability to withstand shrinkage stress [32,49,50]. Therefore, when W/S is 52%, the sample with C/S at 16% shows the highest early-stage shrinkage but the lowest shrinkage at 56 d. As shown in Figure 15a, when C/S = 12%, drying shrinkage becomes increasingly pronounced with higher W/S at all ages [51]. Meanwhile, Figure 15b indicates that ε_at and w_L exhibit similar trends over time. In addition, the inset in Figure 15b demonstrates a positive correlation between ε_at and w_L. This suggests that water loss is the primary cause of drying shrinkage in the sample, highlighting the importance of strictly controlling the upper limit of water content in engineering applications.

3.2.3. Water Stability

The residual strength (r_qn) of the specimens after N wetting–drying cycles is presented in Figure 16. Except for the (C/S)4(W/S)52 specimen, all others remained intact after 13 wet–dry cycles. Notably, when the cement-to-soil ratio (C/S) was 12% and the water-to-solid ratio (W/S) ranged from 44% to 56%, the residual strength after 13 cycles exceeded 80% of the initial strength. This result differs significantly from the typical disintegration of untreated red-bed mudstone under wet–dry cycles, indicating that the incorporation of cement markedly enhances its water stability [2]. Moreover, most of the specimens exhibited an initial increase in r_qn, followed by a decline, and then stabilized after nine cycles. During the early cycles, elevated temperatures and sufficient moisture promoted cement hydration, thereby increasing the strength of the samples, particularly those with a higher cement content [10,14,52]. As the cycles progressed, repeated drying shrinkage and wetting expansion induced deformation stresses that exceeded the structural strength, causing stress concentrations at weak points between soil particles, and leading to the formation of microcracks [5,10,52], as illustrated in Figure 17. These microcracks facilitated water ingress, further degrading the internal structure and reducing the compressive strength. However, the decline in r_qn slowed after nine cycles. This stabilization can be attributed to the widening of internal cracks, which, after several cycles, provide space for deformation and mitigate the damaging effects of shrinkage and expansion on the structure [10,52]. This trend is consistent with previous findings regarding the water stability of CLSMs [10,14,52].

Figure 16 demonstrates that, after 13 wet–dry cycles, the residual strength (r_qn) decreased by 26% when the cement-to-soil ratio (C/S) was 12% and the water-to-solid ratio (W/S) increased from 44% to 56%, indicating significant deterioration. To ensure water stability, it is crucial to impose strict limits on W/S [31]. At W/S = 52%, increasing C/S significantly improved the water stability. For example, while (C/S)4(W/S)52 disintegrated, (C/S)8(W/S)52 remained intact, as illustrated in Figure 17. This improvement can be attributed to the increased cement hydration products, which reduce interconnected pores and block permeability channels, thereby mitigating damage from wet–dry cycles [5,53,54]. These findings highlight the critical role of cement in enhancing the water stability of red-bed mudstone, supporting its application in CLSM production.

3.3. Mix Design

3.3.1. Reasonable Range of Mix Design

Based on the experimental data, a mix design was developed for CLSMs used in subgrade, culvert backfill, retaining wall backfill, and bridge abutment applications. The first step involved identifying the optimal ranges of the C/S and W/S. Figure 18a presents the performance of CLSM with a fixed C/S of 12% and W/S values of 44%, 48%, 52%, and 56%, while Figure 18b shows results for a fixed W/S of 52% and C/S values of 4%, 8%, 12%, and 16%. In both panels, the bleeding rate, subsidence rate, and drying shrinkage rate are expressed as reciprocals; higher reciprocal values indicate lower actual values and, thus, more favorable mix proportions. The setting time is generally constrained within 24 h based on project timelines. Overall, higher performance values in the figures indicate a more optimal mix design. The performance indicators and corresponding evaluation criteria are provided in

Table 4. Design standards for CLSMs in various engineering applications.

Engineering Performance	Engineering Type	Design Standard	Standard/Reference
Flowability	I	≥180 mm	JHS A 313-1992 [26]
	II	≥200 mm	JHS A 313-1992 [26], DBJ51/T 188-2022 [25], and References [5,25,46]
Bleeding rate	I	<5%	References [5,9,29,30,31,32,33]
	II	<5%
Setting time	I	<24 h	References [5,34,35]
	II	<24 h
Subsidence rate	I	<5%	Reference [38]
	II	<5%
Compressive strength	I	≥0.8 MPa	JHS A 313-1992 [26] and DBJ51/T 188-2022 [25]
	II	≥1.0 MPa
Drying shrinkage rate	I	<1.25%	DBJ51/T 188-2022 [25]
	II	<1.25%
Residual strength	I	≥0.8 MPa	DBJ51/T 188-2022 [25] and Reference [5]
	II	≥1.0 MPa

Note: Type I refers to backfilling behind abutments, culverts, and retaining walls; type II refers to subgrade backfilling.

. It is important to note that no specific standard currently exists for drying shrinkage. This study recommends a design limit of 1.25%, based on the observation that specimens with shrinkage below this threshold showed no visible cracking at 56 d [25]. CLSM is commonly used for underground backfilling, where exposure to wet–dry cycles is minimal. However, in non-underground applications, durability under such conditions must be considered. This research focused on evaluating red-bed mudstone-based CLSM in wet–dry cycling environments. According to the design criteria, specimens must retain structural integrity after 13 cycles and maintain a residual strength above the target value ($q_{u_{\mathrm{a}}}$) [25,54]. As shown in Table 4 and Figure 18a, while a W/S of 56% yields high flowability, it fails to meet other performance requirements, particularly for the bleeding rate (reciprocal > 20) and compressive strength. With W/S reduced to 52%, bleeding remains excessive. Prior analysis suggests that increasing C/S has a limited effect on bleeding, indicating that W/S should be kept below 52%. However, when W/S is 44%, the flowability drops below 180 mm, falling short of self-compacting criteria. Therefore, the suitable W/S range is 44–52%. Further examination of Figure 18b and Table 4 reveals that increasing C/S generally improves performance metrics, except for a slight increase in subsidence rate. While a higher C/S enhances overall performance, it should be limited to 25% to balance cost and environmental impact. As shown in Figure 18b, specimens with W/S = 52% and C/S = 4% exhibited disintegration and shrinkage cracking (1/ε_at is 0), which disappeared when C/S was increased to 8%. Thus, to ensure stability, C/S should not fall below 8% within the W/S range of 44–52%. In summary, the recommended mix design range for CLSMs incorporating red-bed mudstone in subgrades and related infrastructure applications is W/S = 44–52% and C/S = 8–25%.

3.3.2. Optimal Mix Proportion

Based on the above analysis, an optimal mix design was developed within the identified optimal range. A suitable mix should minimize water and cement consumption while fulfilling all engineering performance requirements, thereby reducing both economic costs and the environmental burden associated with cement. The cement content is typically dependent on the amount of water used. To achieve sufficient flowability, a higher water content is often required, which, in turn, increases the cement dosage needed to maintain strength. Therefore, this study adopts the minimum flowability requirement as the design baseline to minimize the usage of both water and cement. Flowability and compressive strength were calculated using Equations (9) and (11), respectively. Specimens were then prepared according to the corresponding C/S and W/S values, and additional tests were conducted to verify compliance with practical engineering requirements. The finalized mix design procedure is illustrated in Figure 19. According to the design method shown in Figure 19, two mixture schemes were developed for abutment, culvert, retaining wall, and subgrade backfilling, without considering long-distance transportation. The corresponding physical and mechanical properties of these mixtures are summarized in

Table 5. The mixture proportions and engineering performance of the designed mixtures.

Sample	f0/mm	Br/%	TS/h	hsr/%	qu0/MPa	εat/%	qu13/MPa
(C/S)12(W/S)47	203	3.57	10	2.5	0.821	1.01	0.805
(C/S)14.6(W/S)46.5	180	3.11	8	3.0	1.033	0.97	1.078

. As shown in Table 5, the engineering performance indicators of CLSM prepared with the two mix designs meet the requirements for backfilling applications. Therefore, with a properly designed mix proportion, red-bed mudstone can be used for CLSM production. Additionally, when CLSM with other engineering performance requirements is needed, the mix can be adjusted based on the design criteria using Figure 19.

4. Conclusions

To facilitate the efficient utilization of waste red-bed mudstone, this study proposes blending the material with cement and water to produce a CLSM. The investigation examined the effects of cement-to-soil ratio (C/S) and water-to-solid ratio (W/S) on the fresh and hardened properties of CLSM. The primary findings can be summarized as follows:

Firstly, the initial flowability of the specimens significantly improves with increased W/S, and also with higher C/S. For C/S values ranging from 4% to 16%, W/S should not fall below 44%, to ensure self-compacting properties. A predictive model for flowability, based on a normalization method, accurately describes the relationship between flowability, W/S, and C/S. Notably, flowability diminishes exponentially over time, with the loss rate reaching approximately 30% within 3 h.

Secondly, bleeding becomes more pronounced at higher W/S and is minimally influenced by C/S. To maintain the bleeding rate below 5%, W/S must not exceed 52%. Increased W/S prolongs the setting time, whereas higher C/S notably shortens it. Within the studied ranges of W/S (44–56%) and C/S (4–16%), the setting times remain within 24 h. The subsidence rate rises with increased W/S and shows minimal sensitivity to C/S; hence, controlling the upper limit of W/S (56%) is necessary to satisfy the design criteria.

Thirdly, compressive strength displays a strong linear negative correlation with W/S and significantly increases with higher C/S. With C/S ranging from 8% to 16% and W/S between 44% and 56%, compressive strength values from 0.3 MPa to 1.22 MPa are achievable, meeting the backfilling requirements of most projects. A normalization-based estimation model relating compressive strength to variations in C/S and W/S was also developed.

Fourthly, drying shrinkage correlates positively with water loss, increasing rapidly within the initial 17 d and stabilizing after 56 d. At 56 d, shrinkage exhibits a positive correlation with W/S and a negative correlation with C/S. The addition of cement significantly enhances the water stability of red-bed mudstone. At a cement content of 12% and W/S ranging from 44% to 56%, specimens remain intact after 13 wet–dry cycles, maintaining residual strengths above 80% of the initial values. Residual strength initially increases during early cycles, subsequently declines, and finally stabilizes. In the stabilized phase, residual strength markedly decreases at higher W/S values and increases at higher C/S values. Thus, engineering applications require strict control of the upper limit of W/S and the lower limit of C/S.

Lastly, through an optimized mixture design, red-bed mudstone can be used to produce CLSMs with satisfactory engineering performance. Based on experimental outcomes, a mixture design methodology was established. According to this procedure, the mixture proportion (C/S)12(W/S)47 is suitable for backfilling applications behind abutments, culverts, and retaining walls, whereas the mixture proportion (C/S)14.6(W/S)46.5 meets the requirements for subgrade backfill. These findings offer practical guidance for resource-oriented reuse of excavated soil in engineering projects.