Coal fly ash-based porous geopolymers prepared by a combined foaming process: adsorption mechanism analysis and environmental impacts assessment

Abstract

In this work, coal fly ash, a widely available industrial solid waste in Xinjiang, China, was utilized as the aluminosilicate source to synthesize coal fly ash-based porous geopolymers via a novel foaming strategy combining saponification and hydrogen peroxide. This method enabled the formation of coal fly ash-based porous geopolymers with a maximum porosity of 61.63%. Methylene blue and Pb²⁺ were selected as representative contaminants to evaluate the adsorption behavior of coal fly ash-based porous geopolymers under optimized conditions, with maximum adsorption capacities of 65.85 mg/g for methylene blue and 264.55 mg/g for Pb²⁺. The isotherm and kinetic model fitting results suggested that both adsorption processes follow a combined chemical-physical mechanism. FT-IR and XPS further revealed the relationship between structural features and adsorption behavior, highlighting the combined effect of pore architecture, surface functional groups, and amorphous gel in governing adsorption behavior. Life cycle assessment using openLCA indicated relatively low environmental impacts, largely attributed to the utilization of coal fly ash as a waste-derived raw material. Using decarbonized electricity and low-toxicity reagents is the direction for subsequent research optimization. Also, the cost of coal fly ash-based porous geopolymers is only $0.154 kg^-1. This work provides theoretical support for design of efficient, green, and cost-effective adsorbents.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-48956-8.

Introduction

Persistent organic dyes and inorganic heavy metals such as methylene blue (MB) and Pb²⁺ tend to accumulate in aquatic systems, continuously deteriorating water quality, ultimately exerting adverse effects on human health^1,2. The effective elimination of dyes and heavy metal contaminants from water has become an urgent environmental challenge. Compared to other treatment technologies such as photodegradation, ultrafiltration, and biofilms, adsorption is widely regarded as a prominent technique for wastewater treatment owing to its simplicity, cost-effectiveness, and high removal efficiency³. Currently, adsorbents such as silica⁴, polyacrylonitrile⁵, and graphene oxide⁶ have been employed for the treatment of wastewater containing dyes and heavy metals. However, high production costs and energy consumption, as well as the complexity of preparation, hinder their widespread application. Therefore, the fabrication of innovative adsorption materials that are cost-effective, efficient, and environmental-friendly has become a major research focus.

Geopolymers are three-dimensional aluminosilicate network structures formed from precursors rich in SiO₂ and Al₂O₃ through dissolution, polymerization, and solidification under alkaline activation⁷. The concept was initially proposed by Davidovits in the late 1970s⁸. Its molecular structure consists of a tetrahedral anionic framework composed of [SiO₄]^– and [AlO₄]^– units, with charge-balancing cations (such as K⁺ and Na⁺) capable of undergoing ion exchange with cations in solution⁹. Recently, increasing attention has been devoted to the valorization of industrial solid wastes as functional materials for environmental remediation. Cheng et al. demonstrated the feasibility of utilizing metallurgical waste for gas purification¹⁰. They also have explored the application of reclaimed sintered ore in the adsorption and reduction of CO from flue gas¹¹. These studies highlight the potential of industrial by-products as low-cost functional materials for pollution control. Due to the wide distribution, coal fly ash (CFA), an available and low-cost industrial solid waste, which is rich in reactive SiO₂ and Al₂O₃, favors the preparation of adsorbent. Also, it is more in line with the actual carbon emission control policy and circular economy in China. Previous studies have demonstrated the feasibility of using CFA as a precursor to prepare geopolymers and applied it to the field of adsorption^6,12.

To prepare geopolymers with open porous structures, previous studies have demonstrated that H₂O₂ and aluminum powder could generate porous structures within the material. However, increasing the foaming agent dosage resulted in a gradual decrease in compressive strength^13,14. To enhance the structural stability of porous geopolymers, Jaabay et al.¹⁵ successfully prepared porous geopolymers with a porosity of up to 72.8% by employing a synergistic foaming strategy using a saponification agent (olive oil) combined with H₂O₂. However, this study mainly focused on material synthesis and structural characteristics, while the potential application of such porous geopolymers for pollutant removal has not been systematically investigated. Moreover, most existing studies tend to merely attribute the removal mechanisms to general physical or chemical interactions^16,17. In fact, structural features such as pore architecture, surface functional groups, and gel-phase composition are critical factors governing mass transfer, active site availability, and pollutant binding. Despite their importance, the structure-function relationship in porous geopolymer adsorbents remains insufficiently understood. Therefore, establishing a clear correlation between material structure and adsorption behavior is essential for optimizing the design of geopolymer-based adsorbents.

In addition to adsorption performance, the environmental sustainability of adsorbent materials has attracted increasing attention in recent years. Life cycle assessment (LCA) is widely recognized as an effective tool for quantifying the environmental impacts associated with material production, use, and disposal^18–20. da Silva et al.²¹ conducted a full LCA on the LDH-MgFe production for nitrate removal, which revealed that NaOH is the primary factor contributing to the increased environmental impacts throughout the life cycle. Sharma et al.²² reported an LCA for the first time for yeast lipid-based biofuels and chemicals from the Indian perspective. The results indicated that yeast lipids are a potential renewable energy source, yet it is still necessary to seek alternatives to toxic solvents and reduce energy consumption to mitigate environmental impacts. However, most studies on geopolymer-based adsorbents primarily focus on adsorption capacity and mechanisms, while the environmental impacts related to their preparation and application are rarely evaluated. In particular, the environmental implications of porous geopolymer adsorbents prepared using foaming strategies remain insufficiently explored. Therefore, integrating adsorption performance with life cycle assessment is essential for comprehensively evaluating the sustainability of such waste-derived materials.

In this study, CFA from a coal-fired power station in Xinjiang, China was employed as the raw material by using the saponification/H₂O₂ combined method to synthesize porous geopolymers for the efficient removal of MB and Pb²⁺. To elucidate the underlying adsorption mechanisms, multiple isotherm models, kinetic analyses, and advanced characterization techniques were employed. The analysis considered multiple material attributes, providing new insights into the structure-function relationship between the porous geopolymer structure and its adsorption performance. Meanwhile, the environmental impacts and costs of the adsorbent were evaluated by the LCA method, enabling an integrated assessment of adsorption performance and environmental sustainability.

This work offers an environmentally sustainable, efficient, and low-cost solution for water pollution control, aligning with the principles of circular economy. The deep adsorption mechanisms and LCA provide a more comprehensive theoretical foundation for the design and application of solid waste-based adsorbents in environmental remediation.

Experimental conditions

Materials (including Table S1 and Table S2) section is in the Supporting Information.

Coal fly ash-based porous geopolymer preparation

CFA was selected as the primary raw material, and alkaline activators with different moduli were pre-prepared for subsequent use. Following the direct foaming method²³, 100 g of CFA was mixed rapidly with alkaline activators and water until a homogeneous paste was formed. Subsequently, H₂O₂ and olive oil were added sequentially, and the mixture was stirred slowly until complete foaming status was achieved. The slurry was then molded and cured in an oven at a controlled temperature for a fixed duration. After curing, the samples were ground through a sieve with an opening size of 150 μm, and the powdered CBPG (< 150 μm) was used in all adsorption experiments. Given that increased porosity has been reported to influence adsorption performance actively²⁴, a single-factor experimental design was employed to optimize the synthesis parameters, with porosity as the primary evaluation index. The specific synthesis conditions are listed in Table 1.

Table 1.

CBPG preparation conditions.

Influencing factors	Contact time	H2O2	Olive oil	Alkaline activators modulus	Temperature
	h	wt%	wt%	/	°C
Parameters	5	1.5	1.5	1.2	50
	6	2.0	2.0	1.4	60
	7	2.5	2.5	1.6	70
	8	3.0	3.0	1.8	80
	9	3.5	3.5	2.0	90

Adsorption tests

A specific amount of CBPG synthesized under optimal conditions was subjected to separate adsorption experiments with MB and lead ion (Pb²⁺) solutions at room temperature, using a shaking speed of 200 rpm. After equilibrium, the suspensions were centrifuged and the supernatants analyzed. MB concentration was determined via UV-Vis spectrophotometry, while Pb²⁺ levels were determined using inductively coupled plasma emission spectrometer (ICP-OES). The effects of pH, adsorbent dosage, initial concentration and contact time were systematically examined. Isotherm and kinetic models were fitted, and FT-IR and XPS analyses were performed on CBPG before and after adsorption to elucidate the adsorption mechanisms.

The adsorption capacity (q_e) and removal efficiency (E) were determined according to Eq. (1) and Eq. (2), respectively:

1

2

where q_e (mg/g) represents the amount of pollutant adsorbed per unit mass of porous geopolymer, C₀ and C_e (mg/L) denote the initial and equilibrium concentrations of the pollutant, V (L) denotes the volume of the solution, and m (g) corresponds to the mass of the adsorbent used.

The experimental data were fitted using the Langmuir, Freundlich, Temkin, and Dubinin-Radushkevich (D-R) isotherm models.

The Langmuir model postulates that active sites on the adsorbent surface are distributed homogeneously, with no interactions occurring between adsorbed molecules. It is expressed by Eq. (3)²⁵:

3

where q_m (mg/g) is the maximum adsorption capacity. K_L (L/mg) denotes the Langmuir constant which is related to adsorption equilibrium performance. Once K_L is obtained, the separation factor R_L can be computed via Eq. (4):

4

If R_L = 1, the adsorption exhibits linear behavior; R_L > 1 indicates unfavorable adsorption; R_L = 0 implies irreversible adsorption; 0 < R_L < 1 suggests favorable adsorption.

The Freundlich isotherm describes a non-ideal and reversible adsorption process. It is not limited to monolayer adsorption and can describe multilayer adsorption behavior. In this model, adsorption occurs on a heterogeneous surface with non-uniform distribution of active sites. It is expressed by Eq. (5)¹²:

5

K_F (mg⁽¹⁻ⁿ⁾Lⁿ/g) represents the Freundlich constant, and the value of n serves to indicate the type of isotherm. When n ranges from 1 to 10, a higher value suggests stronger interactions between the adsorbent and adsorbate, which promotes the adsorption process. Specifically, when n = 1, the model is considered linear, indicating that the adsorption energy is uniform across all active sites.

The Temkin isotherm posits a linear decline in the heat of adsorption as surface coverage increases, reflecting either surface heterogeneity or interactions between adsorbed molecules. Equation (6) expresses the model²⁶:

6

where B_T = RT/B, R denotes the universal gas constant (8.314 J/mol·K), T represents the absolute temperature (K), B (J/mol) is associated with adsorption heat, and A_T (L/g) stands for the equilibrium binding constant.

The D-R isotherm model, based on micropore filling theory, is primarily used to describe the adsorption behavior of adsorbents toward solutes in gaseous or aqueous phases. In this model, adsorption occurs by filling micropores through potential energy fields rather than simple surface coverage. The adsorption capacity is governed by both the microporous volume and the adsorption potential energy. The model is presented in Eq. (7)²⁷:

7

in which K denotes D-R constant related to the adsorption free energy (mol²/KJ²). ε represents the Polanyi potential (KJ/mol). The model also allows the calculation of the mean adsorption free energy E (KJ/mol) via formula Eq. (8):

8

if E < 8 KJ/mol, the adsorption process is classified as physical; where as if E > 8 KJ/mol, the process is considered chemical.

The development of adsorption kinetic models enables the elucidation of rate-controlling mechanisms and predict the adsorption behavior over time. The experimental data were fitted using pseudo-first-order, pseudo-second-order²⁸, and intraparticle diffusion models²⁹. Equation (9) expresses the pseudo-first-order model:

9

where k₁ denotes the rate constant (g/mg·min), and q_t represents the quantity of pollutant adsorbed per unit mass of adsorbent at time t (mg/g).

The pseudo-second-order model is described by Eqs. (10),

10

in which k₂ represents the rate constant (g/mg·min).

The intraparticle diffusion model provides insight into the transport behavior of adsorbates within the interior of adsorbent particles. This model postulates that diffusion within the internal pores of the adsorbent, in a homogeneously mixed solution, is the rate-limiting step throughout the overall process of adsorption. The formula is given as Eq. (11):

11

where k_i is the intraparticle diffusion rate constant (mg/(g·min^1/2)), and C_i is a constant related to the thickness of the boundary layer (mg/g). If the plot of q_t versus the square root of time displays multiple linear segments, this suggests that diffusion occurs in multiple sequential stages.

Life cycle assessment

Purpose and scope

To evaluate the life-cycle environmental impacts and energy consumption of CBPG adsorbent prepared from CFA for MB and Pb²⁺ removal, openLCA 2.5 was applied which follows the guidelines formulated in ISO 14044. The system boundary is ‘cradle to grave’, including raw material transport, CBPG production, adsorption, and end-of-life (disposal of spent adsorbent). Fig. 1 presents the input/output processes of the hypothetical scenario. The functional unit (FU) is defined as removal of 1 t MB/Pb²⁺. The life cycle inventory was originally based on laboratory-scale experiments normalized to 1 mg pollutant removed, and results were linearly scaled up to 1 t. This extrapolation assumes proportional scaling of inputs and outputs, without considering potential process intensification or efficiency improvements at industrial scale.

Fig. 1.

System boundaries and inputs/outputs of producing CBPG absorbent to remove 1 t MB/Pb²⁺.

Life cycle inventory

The CML-IA baseline study was used to analyz the impact assessment. Table 2 shows the life cycle inventory of each stage, including all input and output data for each unit which was derived from the experimental results of our laboratory, relevant literature and ecoinvent database. The effective yield is approximately 71–72%, with the remainder lost as evaporated water and gases. There are some assumptions when collecting the data above: (1) CFA was treated as an industrial waste-derived material and modeled using a cut-off approach, meaning that no upstream environmental burdens from the coal combustion process were allocated to CFA. This assumption follows the allocation rules in ISO 14044, and is commonly adopted in LCA studies involving waste-derived materials. (2) The materials and energy used in the construction of facilities and equipment in all processes are outside the system boundaries³⁰. (3) The computation is carried out under optimal conditions for adsorption. Based on the market prices of materials and energy, the cost of CBPG in its life cycle was calculated. Labor costs, equipment maintenance, transport and so on were calculated as 10% of total cost, according to existing studies³¹. The impact assessment was conducted using the CML-IA baseline method, and the main environmental impact categories considered in this study included global warming potential, acidification, eutrophication, human toxicity, and so on.

Table 2.

Life cycle inventory of each stage.

Process		Inputs			Outputs
		Flows	Regional datasets	Amount	Flows	Amount
CBPG preparation		Coal fly ash	RoW	0.8 kg	CBPG, produced	1 kg
		Sodium silicate	RoW	0.23016 kg
		NaOH	RoW	0.09072 kg
		H2O2	RoW	0.02664 kg
		Olive oil	GLO	0.0184 kg
		Deionized water	RoW	0.24 kg
		Electricity	CN, XJ	0.24 kWh
		Transport, lorry	RoW	0.2576 tkm
		Transport, train	CN	1.0302 tkm
Adsorption	Pb2+	CBPG, produced	/	3.78E-6 kg	Pb2+ removed	1 mg
		Electricity	CN, XJ	4.73E-4 kWh	CBPG spent	3.78E-6 kg
	MB	CBPG, produced	/	1.52E-5 kg	MB removed	1 mg
		Electricity	CN, XJ	8.45E-4 kWh	CBPG spent	1.52E-5 kg
CBPG spent landfill		Transport, lorry	RoW	0.025 tkm	CBPG spent	1 kg
		Hazardous waste, for underground deposit	/	1 kg

(Notes: Electricity consumption is based on the low-voltage electricity mix from the openLCA background database. Transportation distances in Xinjiang were estimated according to laboratory-scale procurement conditions with an assumed 80% railway and 20% road transport structure.)

Characterization

The porosity of CBPG was determined using a direct-reading solid density meter based on Archimedes’ principle (Shanghai Yixin Scientific Instrument Co., China). The mineralogical composition of CFA and CBPG was analyzed by X-ray diffraction (XRD, D8 Advance, Bruker, Germany). Phase identification was subsequently performed using MDI Jade 9 software (Version 9.0, Materials Data Inc., Livermore, CA, USA; https://materialsdata.com/). The functional groups and chemical structures of CFA and CBPG before and after adsorption were examined by Fourier-transform infrared spectroscopy (FT-IR, Vertex 70, Bruker, Germany). The surface morphology, elemental distribution, and microstructure of CFA and CBPG were characterized using scanning electron microscopy coupled with energy spectrum analysis (SEM-EDS, Supra 55, Zeiss, Germany). The specific surface area, pore volume, and pore size distribution of CBPG were determined using a Brunauer-Emmett-Teller analyzer (BET, ASAP 2460, Micromeritics, USA) to evaluate its porous structure. The elemental composition and chemical states on the geopolymer surface were further confirmed by X-ray photoelectron spectroscopy (XPS, ESCALAB 250Xi, Thermo Fisher, USA). The concentration of MB in solution was measured using a UV-visible spectrophotometer (L8, Yixin, China), while the concentration of Pb²⁺ was determined using inductively coupled plasma-optical emission spectrometry (ICP-OES, Optima 8000, PerkinElmer, USA).

Results and discussion

Geopolymers characterization

Optimization of synthesis conditions

Through a series of single-factor experiments, the optimal synthesis conditions for CBPG were determined. The porosity exhibited a mean value of 61.63% with a standard deviation of 2.46% (RSD ≈ 4%), indicating stable pore formation during the combined saponification/H₂O₂ process. These conditions included an alkali activator modulus of 1.6, 3 mL of foaming agent, 2.5 mL of foam stabilizer, a curing temperature of 70 ℃, and a curing time of 6 h. The resulting CBPG sample was used for subsequent characterization and adsorption experiments.

XRD analysis

The CBPG sample synthesized under the optimized conditions described in Sect. 3.1.1 was used for structural characterization. Fig. 2 presented the XRD patterns of CFA and CBPG. The CFA was primarily composed of crystalline phases including quartz (SiO₂, PDF#99-000-3084), anhydrite (CaSO₄, PDF#99-000-0120), mullite (Al₆Si₂O₁₃, PDF#99-000-2542), and magnetite (Fe₃O₄, PDF#99-000-0294), showing distinct diffraction peaks. After alkali activation, CBPG exhibited the formation of a calcium silicate phase (C₂S, 2CaO·SiO₂, PDF#98-000-0275), which contributed to enhancing mechanical strength and durability³². During the curing process, Ca²⁺ ions reacted with CO₂ from the air to form CaCO₃ (PDF#00–005-0586), which also exerted a favorable impact on enhancing the compressive strength of the geopolymer³³.

Fig. 2.

XRD patterns of CFA and CBPG.

Compared with CFA, the XRD peak intensities of CBPG were significantly reduced, indicating that alkali activation induced polymerization reactions, resulting in the conversion of crystalline phases into amorphous gel structures³⁴. The broad hump observed between 15–30° 2θ range in CFA, attributed to its amorphous content, shifted to a higher angle range (20–35°) after geopolymerization, which is a typical feature of amorphous aluminosilicate gel formation³⁵. This suggests the generation of N(C)-A-S-H gels during the geopolymerization of CFA. Additionally, a short diffraction band appeared at 29–31° 2θ range in CBPG, corresponding to C-S-H gel formation³⁶. Moreover, silicon sites in the Si-O chains and the calcium sites in the interlayer region of C-S-H can interact with pollutants via coordination or ion-exchange mechanisms, effectively immobilizing them within the material structure^37,38. This dual adsorption capability is particularly significant for the removal of pollutants from water environments.

FT-IR analysis

The FT-IR spectra of CFA and CBPG are shown in Fig. 3, with principal absorption bands categorized in Table S3. The characteristic peaks at 3434 cm^{− 1} and 1654 cm^{− 1} in CFA are assigned to the stretching and bending vibrations of O-H groups. In CBPG, these peaks become narrower and exhibit increased intensity, indicating that the adsorbed water is transformed into chemically bound water during the geopolymerization process. A broad peak is observed at 1034 cm^{− 1} in CFA, which corresponds to the stretching vibration of amorphous aluminosilicates Si-O-M (M = Si or Al). As the geopolymerization reaction proceeds, this peak gradually shifts to lower wavenumbers, becomes narrower, and increases in intensity, indicating the formation of a highly cross-linked aluminosilicate polymeric network³⁹. This also suggests that part of the crystalline phase in the raw materials has been converted into a glassy amorphous N(C)-A-S-H gel, which serves as further evidence of geopolymerization⁴⁰. The absorption bands observed at 464 cm^{− 1} in CFA and 457 cm^{− 1} in CBPG are attributed to the bending vibrations of Si-O-M (M = Si or Al). The band at 710 cm^{− 1} is assigned to tetrahedral aluminum ([AlO₄]^–). Within the 400–800 cm^{− 1} range, the Si-O and Al-O bending vibrations in CFA appear weak and scattered, indicating a disordered mineral framework, while the corresponding peaks in CBPG show significantly enhanced intensity, reflecting a more ordered mineral framework structure.

Fig. 3.

FT-IR spectrum of CFA and CBPG.

pHpzc analysis

The point of zero charge (pHpzc) is defined as the pH value at which the number of positive charges equals the number of negative charges present on the adsorbent surface. As shown in Fig. 4, the pHpzc of CBPG is approximately 9.7. Therefore, the geopolymer surface carries a positive charge when the pH < 9.7, whereas it bears a negative charge at pH > 9.7⁴¹. This surface charge behavior is expected to significantly influence the adsorption of charged pollutants. In particular, when the solution pH exceeds the pHpzc, the negatively charged surface can promote electrostatic attraction toward cationic species.

Fig. 4.

Point of zero charge (pHpzc) of geopolymers.

SEM-EDX analysis

The microstructures of CFA and CBPG are illustrated in the SEM images in Fig. 5, which also include EDS data. Fig. 5(a) shows that spherical particles are present in the raw CFA, with an overall irregular and relatively dense structure, indicating a low specific surface area. In contrast, CBPG exhibits significant changes, with its microstructure characterized by a gel phase, irregularly sized pores, and partially unreacted CFA particles. Among these, the dense gel phase, which forms the geopolymer matrix and contributes the majority of its mechanical strength, is the primary product of the geopolymerization reaction⁴². This phase is formed by the depolymerization and reorganization of Si-O and Al-O bonds in CFA under alkaline conditions. In combination with XRD and FT-IR results, it is further confirmed that geopolymerization occurred under the action of alkali activators, forming N(C)-A-S-H gel. The macropores observed in the morphology images are primarily attributed to the decomposition of added H₂O₂, which released gas and formed bubbles within the material. The formation of micropores may be associated with a saponification reaction triggered by the addition of olive oil. This reaction helps stabilize the pore structure and facilitates the formation of fine pores. The development of porosity significantly enhances the specific surface area of the geopolymer, providing more active sites for adsorption⁴³. Additionally, some inert spherical CFA particles that remain unreacted are encapsulated by the gel phase. This structure retains the morphology of the raw material and provides a potential support for enhanced adsorption performance. EDS analysis reveals an increase in Si, O, and Na contents and a corresponding decrease in Al and Ca contents after geopolymerization, indicating the successful formation of a sodium aluminosilicate hydrate (N-A-S-H) gel network enriched in silica and sodium.

Fig. 5.

SEM micrographs and EDS spectrum of CFA (a and c) and CBPG (b and d).

Adsorption studies

Effect of solution pH

The influence of solution pH on MB and Pb²⁺ adsorption is presented in Fig. 6(a-b). The initial MB concentration was 20 mg/L, with an adsorbent dosage of 0.2 g/L and a contact time of 90 min. For Pb²⁺, the initial concentration was 60 mg/L, the dosage was 0.6 g/L, the contact time was also 90 min, and pH varied from 2 to 6 to avoid alkaline precipitation.

Fig. 6.

Effect of pH on the adsorption of MB (a) and Pb²⁺ (b) on CBPG.

In Fig. 6(a), the removal efficiency and adsorption capacity of CBPG for MB increased as pH rose, peaking at pH 10 (removal rate 69.77%, adsorption capacity 69.77 mg/g). A slight decline was observed at pH 12. MB exists in aqueous solutions as either undissociated molecules (MB°) or cations (MB⁺). At pH = 3, approximately 86% of MB is in the undissociated form. When pH = pKa = 3.8, MB° and MB⁺ coexist in equal proportions, while at pH ≥ 6, MB⁺ dominates⁴⁴.Therefore, under acidic conditions, abundant H⁺ competed with MB⁺ for active adsorption sites on CBPG. As pH increased, the concentration of H⁺ decreased, weakening this competition and leading to improved MB adsorption. When the pH rose to 10, which was greater than pHpzc, OH^– enhanced electrostatic attraction between MB⁺ and the negatively charged surface of CBPG. Additionally, OH^– may form complexes with dye molecules or other ions, promoting the formation of MB precipitates on the geopolymer surface, further promoting adsorption. Thus, the adsorption mechanism involves both electrostatic attraction and precipitation⁴⁵. However, a decline at pH 12 was observed, likely due to pore structure damage or the deactivation of surface sites in highly alkaline environments. Thus, subsequent MB adsorption experiments were performed at pH 10.

In the Pb²⁺ solution, CBPG exhibited poor adsorption performance at pH 2, due to strong H⁺ competition. As pH increased, adsorption improved, reaching a maximum at pH = 5 (removal rate 99.63%, adsorption capacity 103.88 mg/g). However, further increase to pH = 6 may lead to partial hydrolysis of Pb²⁺ and the formation of colloidal complexes⁴⁶, which may precipitate on the porous surface of CBPG and block access to internal pores, thereby reducing adsorption efficiency. Hence, pH 5 was selected for subsequent Pb²⁺ adsorption experiments.

Effect of adsorbent dose

Figure 7(a-b) shows the effects of CBPG dosage on the adsorption efficiency for MB and Pb²⁺ solutions, respectively. In the experiments, the initial concentration of MB was 20 mg/L, the solution pH was set to 10, and the contact time was 90 min. For Pb²⁺, the initial concentration was 60 mg/L, with a pH of 5 and the same contact time.

Fig. 7.

Effect of adsorbent dose on the adsorption of MB (a) and Pb²⁺ (b) on CBPG.

The results demonstrated that as the adsorbent dosage increased, the removal efficiency of MB rose from 75.9% to 96.7%, while that of Pb²⁺ increased from 86.5% to 99.4%. This enhancement can be ascribed to the increased number of available surface active sites, which facilitated greater interaction between the adsorbent and the pollutants⁴⁷. With the initial pollutant concentration held constant, the presence of more adsorption sites allowed for more effective binding of MB and Pb²⁺ ions, thereby significantly improving removal efficiency. However, further increases in dosage resulted in a gradual plateauing of removal efficiency. This can be attributed to the potential overlap of active sites at higher dosages, which may reduce effective contact between the adsorbent and the pollutants and increase the diffusion path length into the internal pores of the material⁴⁸. Considering both cost-efficiency and adsorption performance, the optimal CBPG dosages selected for subsequent MB and Pb²⁺ adsorption experiments were 0.9 g/L and 0.4 g/L, respectively.

Effect of initial dye concentration and isotherms

The corresponding fitting curves and parameters of isotherm models are presented in Fig. 8 and summarized in Table 4. Both systems had a contact time of 90 min. As the initial concentration increased, the removal efficiency of both pollutants initially rose rapidly but subsequently leveled off, due to saturation of active sites on CBPG⁴⁹.

Fig. 8.

Non-linear (Langmuir, Freundlich, Temkin, and Dubinin-Radushkevich) adsorption isotherms for MB (a) and Pb²⁺ (b) adsorption on CBPG.

Table 4.

Non-linear adsorption isotherms (Langmuir, Freundlich, Temkin, and Dubinin-Radushkevich) parameters for MB and Pb²⁺ adsorption on studied samples.

Models	Parameters	Pollutants
		MB	Pb(II)
Langmuir	qm (mg/g)	94.34	249.71
	KL (L/mg)	0.06	5.11
	r2	0.84	0.93
Freundlich	KF (mg(1−n)Ln/g)	13.21	160.54
	n	2.28	6.92
	r2	0.91	0.88
Temkin	AT (L/g)	1.90	392.55
	BT	14.08	28.04
	r2	0.83	0.92
Dubinin-Radushkevich	qm (mg/g)	70.24	244.62
	E (KJ/mol)	0.16	3.93
	r2	0.68	0.92

For MB, the maximum adsorption capacity was 65.85 mg/g, and the Freundlich model provided the best fit (r² = 0.91). The value of Freundlich constant (K_F) was 13.21, and the heterogeneity factor (n) was 2.28, indicating favorable adsorption and multilayer adsorption on a heterogeneous surface⁵⁰. In contrast, the adsorption of Pb²⁺ was better described by the Langmuir model (r² = 0.93), with a maximum adsorption capacity of 264.55 mg/g. This indicates a monolayer adsorption process occurring on a homogeneous surface, with no significant interaction between adsorbed species. The R_L was within the range of 0–1, confirming the favorability of the adsorption process. The Temkin model yielded a poor fit for MB (r² = 0.83), but showed a relatively good fit for Pb²⁺ (r² = 0.92), suggesting that thermodynamic interactions and initial adsorption energy play a more pronounced role in Pb²⁺ adsorption. Based on the D-R model, the correlation coefficient for MB adsorption was low (r² = 0.68), indicating that this model is not suitable for accurately describing the adsorption behavior. Instead, for Pb²⁺ adsorption, the D-R model showed a reasonable correlation coefficient (r² = 0.92), and the calculated E value (3.93 KJ/mol) was below 8 KJ/mol, indicating that physical interactions dominate the overall adsorption energy profile.

In summary, the adsorption of MB exhibited pronounced surface heterogeneity, while Pb²⁺ adsorption was dominated by uniform monolayer coverage. Notably, Pb²⁺ displayed significantly higher adsorption energy and capacity compared to MB, and was mainly governed by physical interactions, highlighting the superior selectivity of CBPG toward heavy metal ions.

Effect of contact time and kinetics

Figure 9 and Table 5 show the fitting curves and parameters of the kinetic models and the intraparticle diffusion model. The adsorption capacities of CBPG for MB and Pb²⁺ increased rapidly at the initial stage, then gradually slowed, and eventually reached equilibrium. Both kinetic models yielded high correlation coefficients (r² ≈ 1.00) for different pollutants. In particular, the pseudo-second-order model showed better agreement between the calculated and experimental adsorption capacities, suggesting that the adsorption rate of both MB and Pb²⁺ by CBPG is governed by surface active sites⁵¹.

Fig. 9.

Fitting plots of adsorption kinetics models (a-b) and intraparticle diffusion model (c-d) for MB and Pb²⁺ adsorption on CBPG.

Table 5.

Parameters of pseudo-first-order model, pseudo-second-order model and intraparticle diffusion model for MB and Pb²⁺ adsorption on CBPG.

Models	Parameters	Pollutants
		MB	Pb(II)
Experiment	qexp (mg/g)	36.55	143.38
Pseudo-first-order	qm (mg/g)	34.69	139.17
	k1 (g/mg·min)	0.33	0.97
	r2	0.97	0.99
Pseudo-second-order	qm (mg/g)	36.09	143.70
	k2 (g/mg·min)	0.02	0.0064
	r2	0.99	1.00
Intraparticle diffusion	ki−film (mg/(g·min1/2))	1.26	5.57
	Ci−film (mg/g)	26.76	108.56
	r2i−film	0.98	0.98
	ki−pore (mg/(g·min1/2))	0.69	0.83
	Ci−pore (mg/g)	29.38	135.42
	r2i−pore	0.77	0.91
	ki−ad (mg/(g·min1/2))	0.20	0.10
	Ci−ad (mg/g)	34.44	142.06
	r2i−ad	0.98	0.82

The intraparticle diffusion model revealed the diffusion stages and rate-limiting steps throughout the entire adsorption process. The fitting results in Fig. 9(c-d) exhibited three distinct linear regions, corresponding to external surface diffusion, intraparticle diffusion, and the equilibrium stage. During the external diffusion phase, pollutants rapidly migrated from the bulk solution to the CBPG surface under a high concentration gradient. The diffusion rate constants (k_i) were the highest at this stage, particularly for Pb²⁺ (k_i−film = 5.57 (mg/(g·min^1/2)) compared to that of MB (k_i−film = 1.26 (mg/(g·min^1/2)), likely due to the smaller hydrated ionic radius of Pb²⁺, which facilitates its penetration through the boundary layer, whereas the larger molecular size of MB imposes steric hindrance during migration. In the intraparticle diffusion phase, the rate constants for both pollutants decreased, indicating that mass transfer became increasingly restricted. At this stage, the outer surface of CBPG was saturated with adsorbates, requiring MB and Pb²⁺ to diffuse into internal pores to interact with functional groups or undergo ion exchange. The fit for MB in this phase was relatively poor (r² = 0.77), possibly due to the heterogeneity of the pore structure. In the equilibrium stage, due to the decrease in pollutant concentration in the solution and the coverage of the CBPG surface by pollutants, external mass transfer is greatly reduced. Therefore, surface or particle diffusion basically does not occur during this stage, and the adsorption process is in a state of equilibrium. However, the fit for Pb²⁺ in this stage was suboptimal, likely due to additional surface complexation or precipitation reactions that are not fully accounted for in the model, thus complicating the interface behavior⁵². For both pollutants, none of the fitted lines passed through the origin, indicating that intraparticle diffusion served as the main rate-limiting step in the adsorption process⁵³.

Collectively, it can be inferred that the adsorption of MB on CBPG follows a combined physical-chemical mechanism, with a strong dependence on pore structure. In contrast, the adsorption of Pb²⁺ appears to be governed by more complex synergistic mechanisms, including ion exchange, surface complexation, and precipitation.

FT-IR analysis

Figure 10 shows the FT-IR spectra of CBPG before and after adsorption of MB and Pb²⁺, denoted as CBPG-MB and CBPG-Pb²⁺, respectively. A new absorption peak appeared at 1351 cm^{− 1} in the CBPG-MB spectrum, corresponding to the bending vibration of methyl groups, confirming the successful adsorption of MB onto the CBPG surface⁵⁴. Compared with CBPG, the O-H stretching vibration peak at 3436 cm^{− 1} became significantly weakened and broadened, and the O-H bending vibration at 1654 cm^{− 1} also shifted, suggesting that both interactions with MB and Pb²⁺ between surface hydroxyl groups and the adsorbates via hydrogen bonding or coordination. For CBPG-Pb²⁺, these changes may also suggest the formation of surface precipitates. The N-O stretching peak at 1384 cm^{− 1} shifted to a higher wavenumber, implying that coordination or complexation with surface functional groups altered the local electronic environment. Minor shifts in the Si-O-M (M = Si or Al) peaks at 1011 and 457 cm^{− 1} suggest ion exchange-induced structural rearrangements⁵⁵. The absorption peak at 710 cm^{− 1} corresponds to the bending vibration of tetrahedrally coordinated aluminum [AlO₄]^–. After adsorption, this peak shifted to lower wavenumbers. For CBPG-MB, this shift may result from hydrogen bonding or van der Waals interactions between MB molecules and the Al-O framework, slightly altering the Al-O vibrational mode. In contrast, Pb²⁺ is more likely to coordinate directly with the oxygen atoms in [AlO₄]^–, weakening the Al-O bond strength.

Fig. 10.

FT-IR spectrum of CBPG before and after MB and Pb²⁺ adsorption.

XPS analysis

To further elucidate the adsorption mechanisms of CBPG toward MB and Pb²⁺, XPS analysis was conducted on the samples before and after adsorption. Fig. 11 presents the corresponding spectra. As shown in the survey spectra Fig. 11(a), a significant increase in the N 1 s peak intensity was observed for CBPG-MB, while distinct Pb 4f peaks appeared in CBPG-Pb²⁺, confirming the successful adsorption of both pollutants by CBPG. Moreover, the disappearance of the Na 1 s peak and the marked attenuation of the Ca 2p peak in the CBPG-Pb²⁺ spectrum suggest that ion exchange occurred between Pb²⁺ and Na⁺/Ca²⁺ during the adsorption process, supporting prior findings.

Fig. 11.

Survey (a), C 1 s (b), O 1 s (c) N1s (d) Si 2p (e) and Al 2p (f) XPS spectra of CBPG before and after MB/Pb²⁺ adsorption.

In the high-resolution O 1 s spectra Fig. 11(c), the Si-OH bending energy shifted to higher values after adsorption of both pollutants. This is attributed to the deprotonation of Si-OH groups to Si-O^– in aqueous solution, facilitating electrostatic or coordinative interactions with MB and Pb²⁺⁵⁶. Moreover, Si-OH groups may also interact with MB through hydrogen bonding or π-π interactions, which reduce the local electron density around the oxygen atoms and result in an increase in binding energy.

In the high-resolution N 1 s spectra Fig. 11(d), for CBPG-MB, the N 1 s peak corresponding to C-NH₂ shifted from 399.96 eV to 399.56 eV, indicating that amino functional groups have an affinity for MB⁵⁷. This shift suggests that MB molecules formed more stable complexes on the CBPG surface or engaged in hydrogen bonding with -NH₂ groups. In CBPG-Pb²⁺, a downward shift in the N 1 s peak, along with the FT-IR shift of the N-O vibration peak to higher wavenumbers, suggests the possible involvement of coordination interactions during Pb²⁺ adsorption.

Combining the Si 2p and Al 2p spectra in the Fig. 11(e-f), after the adsorption of MB and Pb²⁺ respectively, the binding energy of Al increased, while that of Si decreased slightly. This phenomenon can be attributed to the electronic coupling behavior within the Si-O-Al framework. During adsorption, the [AlO₄]^– units likely served as main active sites, undergoing electrostatic attraction or surface complexation with the cationic pollutants. As a result, the electron cloud density around Al atoms decreased, leading to an increase in binding energy. Accordingly, the O atoms in the Si-O-Al structure compensate electrons from the Si side, causing the electron cloud density of the Si atoms to increase relatively, resulting in a decrease in binding energy.

Fig. 12 summarizes the adsorption mechanism of MB and Pb²⁺ Overall, MB adsorption on CBPG is predominantly governed by reversible physical interactions, including electrostatic attraction and hydrogen bonding. In contrast, Pb²⁺ adsorption is mainly driven by strong electrostatic attraction due to its higher charge density, with partial contributions from ion exchange and surface complexation. Therefore, both systems exhibit physical-chemical synergistic behavior, but physical interactions dominate the overall energy profile, while localized chemical interactions enhance adsorption stability.

Fig. 12.

The adsorption mechanism of CBPG for MB and Pb²⁺.

Regeneration and reuse experiments

Figure 13 reflects the adsorption capacity of CBPG for MB (C₀ = 80 mg/L, pH = 10, t = 90 min, dosage of 0.9 g/L) and Pb²⁺ (C₀ = 80 mg/L, pH = 5, t = 90 min, dosage of 0.4 g/L) after four adsorption-desorption regeneration experiments. With an increase in the number of cycles, both the removal efficiency and adsorption capacity for the two pollutants exhibited a decreasing trend. After four cycles, the adsorption capacity of CBPG for MB and Pb²⁺ remained at 41.53 mg/g and 153.48 mg/g, respectively, all above 60% of the original adsorption capacity. The overall trend demonstrates that CBPG still retain excellent structural stability and practical application potential after multiple cycles.

Fig. 13.

The change of adsorption capacity for MB (a) and Pb²⁺ (b) with the number of cycles.

Application prospects

Life cycle assessment

Figure 14 shows the environmental impacts of each stage of CBPG used in MB/Pb²⁺ adsorption. This corresponds to a direct linear extrapolation from laboratory FU (1 mg) to 1 t removal, and this simplification does not fully capture potential changes in energy efficiency and process performance at industrial scale. Therefore, the results should be interpreted as an indicative assessment rather than a precise industrial-scale evaluation. Due to the difference in adsorption capacity, the environmental impacts of MB are generally higher than those of Pb²⁺. Overall, electricity used in adsorption accounts for the largest share of impacts in multiple categories. Chemical inputs contribute disproportionately to toxicity-related categories due to their upstream manufacturing burdens and high characterization factors in toxicity models. Therefore, small mass fractions can produce large toxicity impacts. Accordingly, energy optimization and substitution with low-toxicity reagents can represent the highest-leverage mitigation options. Moreover, negative contributions were observed in the contribution tree, mainly due to the avoided environmental burdens associated with the utilization of industrial solid waste (CFA) as raw material. By substituting conventional raw materials, the use of CFA reduces the environmental impacts that would otherwise be generated during the production of virgin materials. Taking Global Warming Potential (GWP100a), which is considered in numerous literature, as an example³⁰, when impacts are reported per 1 t pollutants removed, CBPG exhibits low life-cycle GWP (34498 and 25039 kg CO₂ eq per ton of MB and Pb²⁺ removed especially) owing to its high adsorption capacity and the valorization of fly ash as a waste feedstock.

Fig. 14.

The environmental impacts (per t pollutant removed) of each stage of CBPG used in MB/Pb²⁺ adsorption.

In the life cycle of CBPG, electricity consumption and chemical usage contribute significantly to the overall environmental impacts, whereas CFA, as an industrial solid waste, was treated without allocating upstream environmental burdens in the baseline scenario. Therefore, a sensitivity analysis was performed by varying these key parameters (± 20%) to evaluate the robustness of the LCA results. Table 6 shows the sensitivity analysis results of various environmental impact categories under different parameter variations. The results indicate that electricity consumption (± 20%) has a noticeable influence on most environmental impact categories, with variations of approximately 20%. Among them, the impact on terrestrial ecotoxicity (TE) is the smallest, with a variation of only ± 3.5%. When the consumption of major chemicals varies by ± 20%, the TE value exhibits the largest change (16.8%), which can be attributed to the relatively high contribution of chemical production processes to this category. In comparison, when the environmental burdens of CFA were considered, except for abiotic depletion (fossil fuels), most environmental impact categories of MB adsorption change by around 10%, while for Pb²⁺ adsorption, the value is 5%. Overall, although variations in key parameters lead to moderate changes in individual environmental impact categories, the relative environmental advantages of CBPG remain unchanged, indicating that the LCA results of this study are reasonably robust.

Table 6.

Sensitivity analysis of main variables to environmental indexes.

MB	Electricity	Major chemicals	CFA
MB
Pb (II)

Comprehensive comparison with reported adsorbents

While life cycle assessment provides insight into the environmental sustainability of CBPG, practical application also requires consideration of economic feasibility. Therefore, a cost analysis was conducted based on raw material consumption, energy input, and auxiliary expenses during the preparation process (as shown in Table S4). The calculated production cost of CBPG is approximately 0.154 $/kg, which is relatively low compared to many reported adsorbents. This cost advantage is mainly attributed to the utilization of coal fly ash as a waste-derived precursor and the avoidance of high-temperature treatment.

To further evaluate the practical competitiveness of CBPG, a comparative analysis was conducted with previously reported geopolymer-based and conventional adsorbents. As summarized in Table 7, representative adsorbents were compared in terms of adsorption capacity, reusability, and cost estimation. In particular, the maximum adsorption capacity of CBPG for Pb²⁺ (264.55 mg/g) significantly exceeds many geopolymers reported. Also, the production cost of CBPG (0.154 $/kg) is considerably lower than that of activated carbon (36.62 $/kg) and also lower than other porous geopolymer systems reported in the literature (0.30 $/kg). This combination of high adsorption capacity and low cost highlights the strong economic competitiveness of CBPG. The overall performance indicates that CBPG is a promising candidate for practical wastewater treatment applications.

Table 7.

Comparison of different adsorbents.

Adsorbents	Pollutants	Adsorption conditions	qm (mg/g)	Reusability	Costs	References
Metakaolin geopolymers	Cu(II)	pH = 5.5, dosage = 0.2 g/L, t = 120 min, C0 = 100 mg/L	97.67 mg/g	5 cycles	Not given	Yu et al.58,
Coal gangue and red mud geopolymers	Pb(II)	pH = 5, dosage = 0.6 g/L, t = 200 min, C0 = 140 mg/L	137.7 mg/g	5 cycles	Not given	Wang59,
CFA/steel slag geopolymers	MB	pH = 7, dosage = 3 g/L, t = 240 min	33.30 mg/g	3 cycles	Not given	Yan60,
	Cu(II)	pH = 5, dosage = 4 g/L, t = 420 min	24.15 mg/g	3 cycles
Coal gangue/fly ash porous geopolymers	MB	pH = 6, dosage = 1 g/L, t = 300 min, C0 = 200 mg/L	35.60 mg/g	5 cycles	0.30 $/kg	Fang et al.16,
	Pb(II)		98.31 mg/g
	Cu(II)		85.67 mg/g
Activated carbon	Pb(II)	Not given	120 mg/g	Not given	36.62 $/kg	Qi61,
Coal fly ash porous geopolymers	MB	pH = 10, dosage = 0.9 g/L, t = 90 min, C0 = 100 mg/L	65.85 mg/g	4 cycles	0.154 $/kg	This work
	Pb(II)	pH = 5, dosage = 0.4 g/L, t = 60 min, C0 = 150 mg/L	264.55 mg/g

Conclusion

1. CBPG was successfully synthesized from CFA using the saponification/H₂O₂ combined method. Under optimal conditions (alkali modulus = 1.6, 3 mL foaming agent, 2.5 mL foam stabilizer, 70 °C for 6 h), the resulting CBPG exhibited a high porosity of 61.63%, which possesses a typical amorphous aluminosilicate gel structure and mesoporous characteristics.

2. The maximum adsorption capacities were 65.85 mg/g for MB and 264.55 mg/g for Pb²⁺. During adsorption, MB molecules diffuse into the mesoporous structure and interact with -OH groups via hydrogen bonding or bind to [AlO₄]^– sites through electrostatic attraction. While Pb²⁺ ions are driven into the internal matrix by initial adsorption potential, where they either coordinate with active sites in the C-S-H or N-A-S-H gels, or undergo ion exchange and precipitation with -OH groups.

3. In the LCA of CBPG used as adsorbents, the overall environmental impacts are low, attributed to the utilization of coal fly ash as a waste-derived raw material. Combined with the low production cost of CBPG (0.154 $/kg), this study provides a promising and sustainable strategy for developing solid-waste-derived adsorbents for wastewater treatment. Subsequent research can enhance environmental benefits from using decarbonized electricity and low-toxicity reagents.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Xinyue Chang: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Project administration, Methodology, Investigation, Formal analysis, Data curation. Yuanrong Yi: Writing – review & editing, Project administration, Investigation, Funding acquisition, Data curation, Conceptualization. Jiale Bai: Visualization, Validation, Formal analysis. Xiaohan Zhou: Writing – review & editing, Visualization, Project administration. Shuo Ma: Writing – review & editing, Supervision, Data curation.

Funding

This work was supported financially by the Xinjiang Uygur Autonomous Region Major Science and Technology Special Project Topic 4 (No. 2024A03002–4).

Data availability

Data will be made available from the corresponding author or the first author on request.

Declarations

Competing interests

The authors declare no competing interests.