Preparation of geopolymer cementitious materials by combining industrial waste and municipal dewatering sludge: Stabilization, microscopic analysis and water seepage

2.1 Materials

The MS originated from a sewage treatment plant in Wuhan City, Hubei Province, characterized by its black-brown hue, pungent odor, and water content ranging from 75 to 85%. CW was sourced from a CW plant in Wuhan City, Hubei Province, and underwent screening to isolate CW with particle sizes of less than 0.15 mm (FCW), 4.75–10 mm (MCW), and 20–40 mm (CCW) for further processing. FGD, a secondary FGD gypsum, is presented as a yellowish powder. FA, graded as V1500-I, manifested as a gray powder with a density of 2.39 g·cm⁻³ and a strength activity index of 72. GGBS, graded as V500-I, appeared as an off-white powder, boasting a density of 2.90 g·cm⁻³, a specific surface area of 428 m²·kg⁻¹, and a 28-day activity index of 96%. For the alkaline activator, quicklime was selected as the activator, in which the content of CaO was 72.56%. Prior to utilization, FCW, FGD, FA, and GGBS underwent drying in a 105°C constant temperature-drying oven (DHG-9071A, Shanghai Yiheng Science and Technology Instrument Co., Ltd, China) for 24 h. The chemical compositions and contents of the test materials are depicted in Table 1.

The MS was introduced into a cement mortar mixer (JJ-5, Hebei High-Speed Railway Construction Experimental Instrument Co., Ltd, China) and stirred at a low speed of 60 rpm for 10 min until a viscous consistency was achieved. Subsequently, an alkaline activator was added, followed by an additional low-speed stirring for 5 min. Then, GGBS, FA, FGD, and FCW were sequentially added, initially stirred at low speed for 10 min, followed by a high-speed stirring at 120 rpm for 5 min. After thorough mixing of industrial solid waste and MS, the mixture was uniformly pressed into customized steel cylindrical molds with a diameter of 50 mm and a height of 100 mm to form solidified sludge samples SMS1, SMS2, SMS3, and SMS4. The material ratios of the samples are depicted in Figure 1. The dosage of the alkaline activator was 2% of the mixture materials. The alkaline activator dosage (2% CaO by dry mass) was selected based on the optimization results of Zhang et al. [32] and Changjutturas et al. [35] for similar precursor blends, which identified 2% CaO as the level that maximized strength and minimized permeability without exceeding environmental limits.

Figure 1

Proportioning of SMS samples.

2.2 Test methods

After 28 days of maintenance, the SMS underwent a series of F-T cycles. Initially, they were placed in a cryogenic test chamber (Model: Zhejiang Jiesheng Cryogenic Equipment Co., Ltd, China) and subjected to freezing at a controlled temperature of −20°C for 24 h. Subsequently, the frozen samples were transferred to a drying oven and dried for 24 h at a controlled temperature of 50°C. Temperature was ramped at 1°C·min⁻¹ between −20°C and +50°C; the chamber reached the target temperatures in ≈70 min, followed by the specified 24 h dwell periods. This freezing and thawing procedure was repeated for a total of 20 cycles.

A pressure tester (TAW-2000, Jinan Zhonglu Chang Testing Machine Manufacturing Co., Ltd, China) was used to test the unconfined compressive strength of the samples with an axial strain of 3% per minute. Referring to ASTM D5084 and ASTM D7100 test methods, a flexible wall permeameter (PN3230M, Geoequip, USA) was used to test the hydraulic conductivity of the samples. The permeation medium was deionized water, and the perimeter pressure and permeation counterpressure were set to 300 and 30 kPa.

An XRD (Smart Lab SE, Rigaku, Japan), an FTIR spectrometer (Nicolet 6700, Thermo Scientific, America), an SEM (Gemini SEM 300, Zeiss, Germany), an NMR analyzer (PQ-001, Suzhou Newmax Electronic Technology Co., Ltd, China) were employed to analyze the mineral phase composition, surface microforms, elemental composition, and pore structures of SMS subjected to F-T cycles.

The model test of the outdoor landfill cover system is illustrated in Figure 2(a). The model box dimensions were 1.0 m × 0.5 m × 1.0 m (length × width × height) with a tilt angle of 5°. The cover system comprised four layers: a 0.1 m thick vegetative layer (VL) composed of engineering waste soil (EWS) at the surface; followed by a 0.5 m thick capillary barrier layer (CBL) consisting of MS, MCW, and calcium bentonite (CB); a 0.1 m thick drainage layer (DL) comprising CCW in the middle; and finally, a 0.3 m thick IL comprising SMS4 at the bottom. The soil parameters for each layer are shown in Table 2. To monitor the environmental conditions, three rows (noted from left to right as the first, second, and third columns) of soil temperature and humidity sensors (CR-SL01TH, Shandong Weihai Qitong Technology Development Co., Ltd, China) were installed at intervals of 0.1, 0.3, 0.5, 0.7, and 0.9 m from the top of the model box, totaling 15 sensors. A data collector (485-USB, Shandong Jiandarenke Electronic Technology Co., Ltd, China) was utilized to record the volumetric water content at various spatial locations from November 2022 to October 2023. Additionally, rainfall data were collected during this period.

Figure 2

Material testing model: (a) longitudinal section of landfill cover system and (b) sample preparation flowchart.

3.1 Volume shrinkage and unconfined compressive strength

Figure 3(a) illustrates the volumetric shrinkage of SMS1–SMS4 samples. Under the influence of F-T cycles, the SMS exhibited an increasing trend in volumetric shrinkage. Initially, the volumetric shrinkage of SMS1, SMS2, SMS3, and SMS4 samples rose from 1.22, 0.86, 0.90, and 0.35% to 1.59, 1.24, 1.94, and 1.74%, respectively, after five cycles. However, this upward trend notably weakened during cycles 10–20, eventually stabilizing at a constant value. Following 20 F-T cycles, the volumetric shrinkage of SMS1–SMS4 samples reached 1.80, 1.87, 2.04, and 2.06%, respectively, without any observed cracks in the samples. These results indicate the solidified sludge’s robust structural stability. Utilizing this material as a landfill cover can substantially enhance the cover system’s longevity and address the engineering shortcomings associated with the poor durability of traditional compacted clay cover materials, such as bulk shrinkage and cracking.

Figure 3

(a) Volume shrinkage and (b) unconfined compression strength of SMS.

Figure 3(b) depicts the unconfined compressive strengths of SMS1–SMS4 samples. Following the initial two F-T cycles, the unconfined compressive strength of SMS1–SMS4 samples exhibited an increasing trend, rising from 7.63, 6.95, 7.00, and 8.42 MPa to 8.08, 7.45, 8.10, and 11.14 MPa, respectively. Notably, the SMS4 sample demonstrated the most significant strength increase, reaching 32%. Subsequently, the compressive strength of SMS1–SMS4 samples gradually declined to 6.34, 6.00, 8.04, and 10.83 MPa during cycles 3–20. After 20 F-T cycles, the compressive strength of SMS4 slid slightly from 11.14 to 10.83 MPa (Figure 3(b)). This minor decline is mirrored at the microscale: NMR T2 distributions shift rightward as pores ≥0.1 µm proliferate, while SEM images disclose a looser fabric enriched in amorphous phases. Together, these structural defects eclipse any late-stage hydration benefit, yielding the observed 2.8% loss in strength. These findings affirm that SMS exhibits robust resistance to damage from F-T cycles. In regions akin to Northwest and Northeast China, the SMS cover material can maintain high compressive strength over an extended period under repeated high and low temperatures, thereby enhancing the overall stability of the cover system.

At low freezing temperatures, the liquid water within the SMS undergoes a phase change to ice crystals, leading to volume expansion that disrupts the association and arrangement between the material particles. Upon melting at higher temperatures, the ice transitions back to water, causing the pores to shrink. Simultaneously, unreacted FA and GGBS undergo a secondary hydration reaction, generating hydrated calcium silicate, which enhances structural linkage within the sample, consuming water in the process. In the initial F-T cycles, the structural damage inflicted by freezing and thawing is outweighed by the cementing effect of hydration products, increasing the unconfined compressive strength of the SMS. However, in the later stages of the F-T cycle, with reduced water content in the sample, the effect of ice crystal freezing and swelling diminishes, and the hydration reaction nearly ceases. Consequently, some portions of the sample may experience fracturing, leading to a decrease in unconfined compressive strength, while the volumetric shrinkage remains relatively stable.

3.2 Hydraulic conductivity

Figure 4(a) presents the hydraulic conductivity coefficients of SMS1–SMS4 samples. The hydraulic conductivity of SMS demonstrates a positive correlation with the number of F-T cycles. After 20 F-T cycles, the hydraulic conductivity of SMS1 and SMS2 samples increased from 4.51 × 10⁻⁸ and 6.53 × 10⁻⁸ to 1.90 × 10⁻⁷ and 4.13 × 10⁻⁷ cm·s⁻¹, respectively. These values exceeded the engineering requirement stipulating that the hydraulic conductivity of the landfill closure capping layer should be less than 1.0 × 10⁻⁷ cm·s⁻¹. Nonetheless, the hydraulic conductivity coefficients of SMS3 and SMS4 samples after F-T cycles were 9.67 × 10⁻⁸ and 8.87 × 10⁻⁸ cm·s⁻¹, respectively, which still met the impermeability performance criteria of the cover material.

Figure 4

(a) Hydraulic conductivity and (b) T2 distribution curve of SMS.

The hydraulic conductivity is primarily influenced by the degree of pore development within the sample. The NMR T2 distribution curve serves as a reflection of the pore characteristics of the sample. Examining the T2 distribution curve of the SMS4 sample (Figure 4(b)), it is evident that the peaks of the T2 spectra gradually enlarged during the initial ten F-T cycles, while the distribution intervals remained relatively constant. This observation suggests that the pore volume of the SMS increased, yet the pore diameter remained largely unchanged. However, after 20 F-T cycles, a significant rightward skewing of the distribution area of the T2 spectral peaks was observed. This phenomenon indicates that the pore size of the internal pores within the SMS expanded, with small pores evolving into medium or large pores. Throughout the F-T cycle, the pore water within the sample undergoes repeated solid–liquid transformations. Frost expansion or contraction occurs at the interface between pore water and sludge particles or cementation products, leading to decreased particle contact interface, particle dispersion, and ultimately an increase in pore size or enlargement. Consequently, this leads to an elevation in the hydraulic conductivity of the sample.

3.3 Water-resistant performance

The volumetric water content in this solid waste-based landfill closure cover system for the period November 2022 through October 2023 is shown in Figure 5. Volumetric water content in CBL and IL responds to rainfall and evapotranspiration cycle recharge processes. There was a delay in the response of volumetric water content as the depth of cover extended. The volumetric water content response curve in CBL fluctuates at a fast frequency. However, the fluctuation amplitude changes are small and basically consistent with the initial water content. The frequency of volumetric water content fluctuations in IL relative to CBL is significantly slower and the amplitude of fluctuations is larger. The volumetric water content showed an overall increasing trend. However, this layer is still in a non-saturated state, and no water leakage occurs at the bottom. The main reason for the difference in volumetric water content fluctuations in CBL and IL is the pore characteristics of the materials. CBL with large pores has weak capillary action and poor water-blocking capacity, and rainfall rapidly infiltrates into DL through this layer. However, IL is composed of SMS with small pores and low permeability. The strong capillary action of IL caused the IL–DL contact interface region to be in a water-saturated state. The increase in volumetric water content in the IL layer was caused by long-term percolation. The overburden system was set at a 5° slope, which caused water in the overburden layer to seep toward the foot of the slope. So much so that the volumetric water content increases near the foot of the slope relative to the top of the slope in the CBL and IL, which is more pronounced in the IL layer. At the end of monitoring, the water content of CR10, CR11, and CR12 near the upper part of IL was 28.19, 48.00, and 44.52%, respectively, and that of CR13, CR14, and CR15 in the lower part of IL was 23.05, 38.73, and 43.88%, respectively, as shown in Figure 5(a)–(c). The 1-year volumetric water content monitoring data of the cover layer showed that the 0.3 m thick SMS IL still exerted good water-blocking service performance under the complex environmental conditions of low temperature, high temperature, rainfall, and evaporation. The results of the outdoor modeling test can provide data support for the promotion of this solid waste-based cover system.

Figure 5

Volumetric water content of landfill cover systems: (a) first column, (b) second column, and (c) third column.

3.4 Microscopic mechanism

SMS4 samples still have relatively good mechanical properties after F-T cycling. Therefore, the changes in mineral composition, microstructure and elemental composition of SMS4 after different F-T cycles were explored (Figure 6). As can be seen from Figure 6(a), there are obvious diffraction peaks in the XRD patterns of SMS4 samples with different numbers of F-T cycles. The mineral structures contained in the SMS4 samples without F-T cycling are mainly silicon oxide (SiO₂), calcite (CaCO₃), tamarugite, gypsum, ettringite, C–S–H, and analcime. The front of CaCO₃ increased for 1, 5, and 10 F-T cycles compared to the samples that did not experience F-T cycles. The reason for this may be that the SMS itself contains enough calcium-containing compounds within the sludge to allow Ca²⁺ to react with the CO₂ in the air to produce a certain amount of CaCO₃. In addition, the other peaks in the graph do not change much, which also indicates that the F-T cycle has less effect on the hydration products generated inside the SMS [32].

Figure 6

Mineral, microstructure, and F-T failure mechanism analysis of SMS4 sample: (a) XRD atlas, (b) FTIR atlas, and (c) SEM micromorphology and EDS distribution map total spectrum. (d) Model of F-T failure mechanism.

From Figure 6(b), it can be seen that the wave number of the absorption peak of –OH stretching and bending vibration is shifted from 3420.87 cm⁻¹ to the absorption peak with a larger wave number of 3441.02 cm⁻¹, and the wave number of the absorption peak of C–O is shifted from 1423.54 cm⁻¹ to the absorption peak with a larger wave number of 1477.21 cm⁻¹, as compared to the FTIR of the sample which did not undergo the effect of F-T cycling for 20 F-T cycles in the SMS4 sample. This suggests that after 20 F-T cycles of the SMS4 sample, substances containing CO 3 2 − were generated in the SMS [36], and in combination with the XRD pattern the mineral may be CaCO₃. In addition, it can also be seen from the figure that the wave numbers of other functional groups such as S–O and Si–O–Si/Al are basically not shifted.

There are a large number of rods, columns, and agglomerates in the samples during the first five F-T cycles. These materials are stacked on top of each other, forming a spatial skeleton with a compact and regular structure and small and uniform pores. After ten F-T cycles, these rods, columns, and agglomerates were intricately arranged with slightly enlarged pores, but with better integrity. After 20 F-T cycles, the amorphous material increased and the internal structure was loosened (Figure 6(c)). The above phenomenon also further illustrates the pattern of change in the permeability coefficient of SMS. In addition, Al and Na changed less with the increase of F-T cycles. This indicates that the F-T cycles do not damage the C–A–S–H and N–A–S–H gelation products much. However, the changes in the elements Ca, Si, O, C, and S are more prominent. After five F-T cycles, the content of Ca and Si elements increased. This suggests that some of the substances have once again undergone hydration reactions to produce amorphous polymers such as silicates. This may be the reason for the increase in mechanical strength of the samples at the early stage of the F-T cycle. With the increase of the number of F-T cycles, the content of Ca and Si elements decreased slightly [37]. This caused a decrease in the mechanical strength of the samples at the later stage of the F-T cycle.

The F-T failure mechanism of the specimen is shown in Figure 6(d). When the number of F-T cycles is small, the formation of ice crystals in the pore water enlarges some of the pores, resulting in a larger contact area between the melted pore water and the material, which promotes the hydration reaction. As the number of F-T cycles increases, the excess water in the system forms ice crystals after the materials are fully hydrated, leading to volume expansion and damage to the internal structure of the specimen. Due to the repeated phase transition of water, the small pores inside the specimen begin to increase and gradually develop into large pores, thereby releasing the internal pore water pressure. The CaCO₃ and amorphous C–S–H generated by the original hydration reaction also suffered varying degrees of damage during this process [38], resulting in a decrease in the mechanical strength of the samples in the later stage of F-T cycles.