Flame atomic absorption spectrometer (AA-6880, Shimadzu); Fourier transform infrared spectrometer (type EQUINOX55, BRUKER, Germany); Scanning electron microscope (LEO-1430VP, Zeiss AG, Germany); X-ray powder diffractometer (D8 Advance, AXS GmbH, Bruker, Germany); Electric constant temperature drying oven (DHG-9140A, Shanghai Jinghong Experimental Equipment Co, Ltd.); Vacuum drying oven (DZF-6020, Shanghai Anting Scientific Instrument Factory); Electronic balance (JY302, Shanghai METTLER TOLEDO Instrument Co, Ltd.); Peristaltic pump (BTOO-300M type; Beijing Baoding Town Lange Constant Flow Pump Co, Ltd.); cadmium hollow cathode lamp; grains of sand (grains of sand collected at different points in the Taklamakan Desert); Cadmium carbonate (Tianjin Beilian Fine Chemical Development Co, Ltd.); Ethanol, acetic acid (Tianjin Damao Chemical Reagent Factory); hydrochloric acid, sodium hydroxide (Urumqi Dicheng Chemical Co, Ltd.); Deionized water (deionized water supply room, School of Chemical Engineering, Xinjiang University).

An appropriate amount of sand grains were weighed, deionized water was washed repeatedly and repeatedly to remove dust and water-soluble components on the surface of the sand particles, and then dried at a constant temperature of 80℃ in an oven for 24 h. A certain mass of sand particles as a solid-phase adsorption material was added to a 4×50 mm glass microcolumn, and the separation microcolumn was used in conjunction with a peristaltic pump for dynamic adsorption and desorption experiments.

After the Cd(II) ion standard solution passed through the adsorption separation microcolumn at a certain flow rate, 0.010 mol· L^− 1 hydrochloric acid was eluted, and the Cd(II) ion solution and desorption solution flowing from the microcolumn were quantitatively collected, respectively. The concentration of Cd(II) ions in the original solution, Cd(II) ion solution and desorption solution was determined by FAAS method. Compared with the original solution, the decrease of Cd(II) ion concentration in the column solution indicates that the sand particles have adsorption effect on the Cd(II) ion solution, and the larger the concentration decrease, the better the adsorption effect. The determination of Cd(II) ions in the desorption solution can further prove the adsorption performance of Cd(II) ions by sand particles.

At room temperature, a Cd(II) ionic solution or sample with pH = 6 was passed through the adsorption separation microcolumn at an appropriate flow rate, and then 0.010 mol·L^− 1 hydrochloric acid was eluted through the microcolumn at a natural flow rate, and the concentration of Cd(II) ions in solution was determined by FAAS method. The relevant adsorption parameters are calculated according to equations (1), (2) and (3) (Huang et al. 2025).

C₀ is the concentration of Cd(II) ions in the original solution before adsorption, C_e is the concentration of Cd(II) ions in the column solution, C_d is the concentration of Cd(II) ions in the desorption solution (mg/L), V is the volume of the solution passing through the microcolumn, V_d is the volume of the desorption solution (ml), and W is the dry weight (g) of the adsorbent added.

Surface morphology of natural sand particles before and after adsorption was analyzed using SEM. As shown in Fig. 2(a,b), the pre-adsorption images reveal that sand particles, as granular natural products, exhibit diverse and irregular shapes with a dense, porous, and rough surface morphology. This characteristic enhances the specific surface area of the sand particles, thereby increasing the number of adsorption sites available on their surfaces during the adsorption process. Comparing with Figs. 2(c,d), it is evident that the overall particle contours remain largely unchanged after cadmium ion loading. However, high-magnification images reveal partial blockage of numerous surface pores and smoothing of localized protrusions. These subtle yet discernible morphological changes provide direct qualitative evidence for the successful immobilization of cadmium ions on the surface of natural sand particles.

The composition and functional groups of the sand particles were characterized by Fourier transform infrared (FTIR) spectroscopy. Analysis of the peak assignments in Fig. 3 indicates that the natural sand particles are primarily composed of silica, with silanol (Si-OH) and siloxane (Si-O-Si) groups being the predominant functional groups.

The broad absorption peaks at 3618 cm^-1 and 3424.8 cm^-1 are attributed to the O-H stretching vibrations of water molecules (H-O-H) adsorbed on the sand particle surface. The absorption peaks observed at 2982.8 cm^-1 and 2876.2 cm^-1 correspond to the C-H stretching vibrations from trace organic matter present in the natural sand. The peak at 1796.9 cm^-1 is assigned to a combination band of the Si-O bond. The strong absorption at 1006.3 cm^-1 is attributed to the asymmetric stretching vibration of the siloxane (Si-O-Si) covalent bond. The peak at 875.3 cm^-1 is associated with the symmetric stretching vibration of the Si-O-Si bond, while the absorption at 710.9 cm^-1 corresponds to the bending vibration of the silanol (Si-OH) group. All these characteristic peaks are indicative of the quartz phase of SiO₂ in the natural sand.The coordinating atoms present in the functional groups of sand particles contribute to their adsorption capacity through coordination interactions.

The composition and functional groups of the natural sand particles were further characterized using X-ray diffraction (XRD), and the results are presented in Fig. 4. The strong diffraction peaks observed at 2θ = 20.76°,26.60°,36.51°,50.09°,59.92° are attributed to SiO₂. The diffraction peaks appearing at 2θ = 19.84°,20.79°,21.95°,26.68°,27.96°,36.50°,39.49°,59.98°,67.64° correspond to CaAl₂Si₂O₈·4H₂O. Additionally, the peaks at 2θ = 19.84°,20.79°,21.95°,26.68°,27.96°,36.50°,39.49°,59.98°,67.64° are identified as CaCO₃. The results indicate that the main components of the natural sand particles include SiO₂, Si(OH)_n, Mn(SiO₄)₂, among others. Combined characterization by XRD and FTIR demonstrates that the sand particles, serving as a carrier abundant in silanol and siloxane functional groups, possess certain adsorption capabilities.

The pore structure characteristics of natural sand particles were characterized by nitrogen adsorption-desorption measurements. The specific surface area was determined to be 3.1731 m²/g, with an average pore width of 8.3451 nm. Mesopores close to the average pore size were found to dominate the porosity contribution (0.00589 cm³/g). As shown in Fig. 5, the adsorption-desorption isotherm can be classified as Type IV according to the IUPAC classification. At low relative pressures (p/p₀ < 0.8), the adsorption quantity increased gradually, corresponding to monolayer-to-multilayer physical adsorption of nitrogen on the particle surfaces and within the pore channels. As the relative pressure approached 1.0, a sharp increase in adsorption quantity was observed, which is characteristic of capillary condensation occurring in mesopores (2–50 nm). The adsorption and desorption branches form an H₃-type hysteresis loop in the medium-to-high relative pressure range (p/p₀ ≈ 0.8-1.0). This type of hysteresis is associated with slit-shaped pores, which is consistent with the physical nature of natural sand particles—primarily composed of dense mineral grains such as quartz, where the porosity originates not from intrinsic intraparticle pores but from interparticle voids formed during accumulation, thus exhibiting typical slit-like pore features of a packed structure.

A 0.01 g portion of acid-washed natural sand grains was packed into a 4 mm i.d. micro-column and challenged with Cd(II) standard solutions (20–200 mg·L^-1) at 25 ℃ and a constant loading flow rate of 3 mg·min^-1. Effluent fractions were collected and analysed by on-line FAAS to construct the equilibrium isotherm shown in Fig. 6(a). Uptake increases monotonically with inlet Cd(II) concentration as the enhanced concentration gradient drives mass transfer. A plateau is reached at ≥ 120 mg·L^-1, signalling saturation of the surface binding sites; the corresponding maximum capacity is 10.2 mg·g^-1, in good agreement with the value predicted by the Freundlich model.

The influence of pH on dynamic Cd(II) uptake was examined by injecting 100 mg·L^-1 solutions of varying acidity through the micro-column at 3 mg·min^-1 (Fig. 6b). From pH 1 to 6, removal efficiency increases monotonically, attaining ~ 90% in the narrow window pH 5.5–6.5. At pH < 3, excess H⁺ ions compete directly with Cd(II) for the deprotonated Si-O^- sites, lowering the surface charge and suppressing adsorption. As the pH rises, progressive deprotonation of Si-OH generates negatively charged Si-O^- ligands that electrostatically attract Cd(II), maximising capacity (Nebagha et al. 2015). Above pH 6.5, the slight decline in removal is attributed to the onset of Cd(OH)₂ precipitation, which decreases the activity of free Cd(II) in the liquid phase. Thus, the optimum operational pH is fixed at 5.5–6.5 to ensure maximum surface complexation while avoiding bulk precipitation.

In fixed-bed operation, the loading flow-rate directly controls the residence time of Cd(II) ions within the sand bed and therefore dictates the extent of adsorption. To identify the optimum flow-rate, a 100 mg·L^-1 Cd(II) solution (pH 6.0) was percolated through the micro-column at 0.5-5.0 mL/minwhile effluent concentrations were monitored on-line (Fig. 6c). Between 0.5 and 3 mL/min, removal efficiency plateaus at ~ 90%, indicating that equilibrium is attained within the available contact window. Beyond 3.5 mL·min^-1, efficiency declines steeply because the reduced residence time (< 15 s) is insufficient for Cd(II) to diffuse to the internal binding sites, resulting in premature breakthrough. Consequently, a loading flow-rate of 3 mL·min^-1 was adopted for all subsequent dynamic experiments as the best compromise between capacity and throughput.

The influence of temperature on adsorption is fundamentally governed by thermodynamic principles. During the adsorption process, temperature affects both the chemical reaction rate and the diffusion rate of heavy metal ions on the adsorbent surface, thereby influencing the overall adsorption rate. This study investigated the effect of temperature (ranging from 298 K to 313 K) on the adsorption performance of natural sand particles toward Cd(II). As shown in Fig. 6(d), the adsorption capacity of natural sand for Cd(II) at 298 K was 7.58 mg/g. With increasing temperature, no significant change in adsorption capacity was observed. The relevant thermodynamic parameters for the adsorption process were calculated using Equations (4) and (5)(Li et al. 2024).

The adsorption of Cd(II) onto natural sand particles was evaluated through thermodynamic analysis, where K_d represents the equilibrium constant, R the ideal gas constant, ΔG⁰ (kJ·mol^-1) the Gibbs free energy change, ΔH⁰ (kJ·mol^-1) the enthalpy change, ΔS⁰ (J·mol^-1·K^-1) the entropy change, and T (K) the absolute temperature.

As summarized in Table 1, the calculated thermodynamic parameters reveal that ΔH⁰ > 0, indicating an endothermic adsorption process. The positive value of ΔS⁰ suggests an increase in randomness at the solid-liquid interface during adsorption. Furthermore, the negative ΔG⁰ confirms the spontaneity of the process.

Adsorption isotherms were fitted using adsorption isotherm equations to investigate the adsorption behavior of cadmium ions on solid phases packed with sand particles. Commonly used equations include the Langmuir and Freundlich adsorption isotherms. Adsorption isotherm fitting curves were established based on the Langmuir and Freundlich adsorption isotherm equations (6) ,(7)and (8)(Dagang et al. 2023):

Where C_e is the cadmium ion concentration at adsorption equilibrium, Q_e is the adsorption capacity at equilibrium, Q_m is the maximum adsorption capacity, K represents the Langmuir constant, Kf represents the Freundlich constant, and a is the adsorption exponent. Results are shown in Fig. 7(a) and Table 2. Fitting the experimental data using the Langmuir and Freundlich models yielded correlation coefficients R² of 0.84 and 0.94, respectively. The Freundlich isotherm fit demonstrated significantly better correlation than the Langmuir fit. Therefore, the adsorption behavior of sand particles toward cadmium ions better conforms to the Freundlich adsorption model. The adsorption energies at different sites on the sand particle surface are non-uniform ^[14], and the adsorption of cadmium ions is dominated by a process involving multiple layers of physical adsorption based on loose monolayer chemical adsorption.

The adsorption mechanism of Cd(II) on natural sand particles was analyzed using pseudo-first-order and pseudo-second-order kinetic models, elucidating the patterns of its adsorption behavior. The equations for the pseudo-first-order and pseudo-second-order kinetic models are shown in (9) and (10)(Yang et al. 2025):

The pseudo-first-order and pseudo-second-order kinetic models were applied to fit the adsorption behavior of Cd(II) onto natural sand particles.

The adsorption behavior of Cd(II) onto natural sand particles was analyzed using pseudo-first-order and pseudo-second-order kinetic models. The corresponding parameters for the adsorption process were calculated, and the results are summarized in Table 3. As shown in Fig. 7(b), the correlation coefficient (R²) for the pseudo-first-order model was 0.848, while that for the pseudo-second-order model reached 0.966. The higher R² value of the pseudo-second-order model indicates that it provides a better fit for the experimental data, suggesting that the adsorption process is predominantly governed by chemical interaction mechanisms.