Transport of Cu²⁺ in Unsaturated Porous Medium with Humic Acid/Iron Oxide Nanoparticle (Fe₃O₄) Amendment

Abstract

Humic acid (HA) and iron oxide (such as Fe₃O₄) nanoparticles are widely distributed in soil, and their complex embedded in soil might affect the transport and fate of Cu²⁺ in the vadose zone, while Cu²⁺ is a serious threat to the underlying groundwater. In this study, we synthesized a composite of Fe₃O₄ nanoparticles coated with HA (HA@NPs) using as an amendment a packed sand matrix in the transport column experiments. The impacts of HA content and ion strength (IS) on Cu²⁺ transport in the unsaturated columns were investigated. The results showed that HA exhibited a stronger inhibition effect on Cu²⁺ transport, and a higher IS enhanced the mobility of Cu²⁺ in an unsaturated porous medium in the presence of HA@NPs. The recovery ratio (R_r) of Cu²⁺ breakthrough in the column decreased from 66.56% to 3.94% while the mass concentration ratio C_HA/C_NPs increased from 0 to 50 in the HA@NPs complex. The R_r increased by 1.64 times while the IS increased from 0 to 100 mM. Batch adsorption experiments, kinetics and isotherm models, and Fourier transform infrared (FTIR) spectra analysis were implemented to elucidate the underlying mechanism. It was found that HA embedded in the sand matrix could bind Cu²⁺ by forming stable chelate, while the IS-dependent Cu²⁺ transport could be attributed to the competitive adsorption between Na⁺ and Cu²⁺. Our study demonstrates that the physicochemical environment, as well as the presence of iron oxide nanoparticles and natural organic matter, can significantly impact Cu²⁺ transport in unsaturated porous medium.

1. Introduction

Copper (Cu) pollution poses a dangerous threat to the environment. In particular, China has the most widespread copper soil pollution problem [1,2]. According to the China Soil Pollution Survey in 2014, 2.1% of the contaminated sites in tests had soil polluted by copper [3]. Ionic copper (Cu²⁺), namely the soluble state of copper, is one of the key soil pollutants that must be removed because of its biotoxicity and mobility [4]. Copper-bearing wastewater from smelting plants, chemical mechanical planarization, and mine drainage could break through the vadose zone and infiltrate into groundwater. This can pose serious risk to the environment and human health, as groundwater supply accounts for 17.5% of the total water supply in China [5]. Understanding the factors affecting transport and fate of Cu²⁺ in unsaturated porous media is thus essential. In particular, while humic acid (HA) and iron oxide (such as Fe₃O₄) particles are widely distributed in the soil, there have been few studies that examine how the embedded Fe₃O₄ particles and HA affect Cu²⁺ transport in unsaturated porous medium [6,7].

HA is the main organic component in the soil [8], and represents approximately 60% of soil organics [9]. A large number of branched hydrophilic and hydrophobic moieties, as well as functional groups (i.e., carboxy, carbonyl, and hydroxyl), are linked to each other by the backbone of the aliphatic or aromatic units in HA [10]. This complex chemical structure determines the sensitivity and the physicochemical properties of HA to any given environmental factors [11]. HA becomes a repository for many pollutants due to its strong surface adsorption caused by the complexation and chelation between the functional groups of HA and the environmental pollutants [12]. The interactions between HA and heavy metals are mainly chemical complexation, cationic proton exchange, electrostatic adsorption, and flocculation [8,13]. The number of oxygen functional groups on the surface of the HA particles is a crucial factor that determines the nature of its interaction with heavy metals [14,15]. Hydrophilic and hydrophobic organics usually enhanced the mobility of heavy metals in clean sand, since the adsorption of heavy metals on the sand surface was inhibited after they were coated by these organics. The mediated transport depended on the inherent mobility of the organics in sand, and usually followed the order of hydrophilic matters > hydrophobic matters [16]. However, if the dissolved organic matter (DOM) was injected continuously, then ternay complexes (metals–DOM–sand medium) might form and substantially inhibit the mobility of heavy metals [17]. Ionic strength (IS) is recognized to compress the double electric layer and consequently influence the electro kinetic characteristics of macromolecular substance [18]. Thus, the reaction between HA and Cu²⁺ may be significantly affected by IS. The impacts of IS on the transport of mobile HA and colloid/nanoparticles (NPs) through porous medium have been reported [19]. Because of increased charge screening and charge neutralization effects, the increase of IS can weaken electrostatic repulsion between suspended NPs and between NPs and the medium particles, typically resulting in NPs aggregation and strong deposition at porous medium surfaces [18,19]. However, the role of IS in the transport of Cu²⁺ in porous medium with embedded HA/ Fe₃O₄ NPs is not well understood.

Moreover, it is important to understand the transportation and interaction between NPs and heavy metals in unsaturated porous medium because the vadose zone represents both the medium that is polluted with Cu²⁺ and the natural barrier that prevents Cu²⁺ from being transported into groundwater [20]. According to survey work done by Corapciogl et al. (1999) [21], there is an average of 10⁸–10¹⁷ colloidal particles (biocolloids, inorganic colloids, and organic colloids) per liter in the groundwater. This indicates that colloids/NPs can carry pollutants such as heavy metals into the groundwater [22]. Most studies focus on the transportation of colloids in a saturated porous medium, where the colloids are adsorbed across the solid-water interface (SWI) and deposited in the medium through multi-particle bridging and the physical interception of steric hindrance [23,24]. Bradford et al. showed that the colloid adsorption across an air-water interfaces (AWI) in an unsaturated porous medium is both strong and irreversible because of the negative charge of the gas–liquid interface [25,26]. Processes such as the deposition and transportation of colloids and heavy metals are controlled by the chemical and hydrodynamic conditions of the porous medium, such as ionic strength, pH, moisture content, and flow rate [26,27]. Few studies have examined the heavy metal transport in an unsaturated porous medium embedded with NPs coated with organic material. In particular, Fe₃O₄ NPs usually appear in soil accompanied by Cu²⁺ in mining areas [28,29], or magnetic Fe₃O₄ NPs are often added artificially for soil remediation because they can be easily separated from the soil matrix through an external magnet to avoid secondary pollution [30,31].

The objective of this work was thus to examine the influence of the presence of HA-coated Fe₃O₄ NPs (HA@NPs) embedded in porous medium and solution IS on Cu²⁺ transport through the vadose zone. The approach focused on a systematic identification of the mechanism that mediated Cu²⁺ transport in unsaturated porous medium with HA@NPs amendment. The Cu²⁺ transport column experiment results were interpreted through monitoring of breakthrough curves (BTCs), a batch adsorption experiment, and surface property measurements of targeted compounds.

2. Materials and Methods

2.1. Porous Media

For the porous medium in our experiments, we collected natural river sand (sifted through 1 mm sieve) from the Yitong River in Changchun, China. Sands with an average diameter of 0.53 mm represent the medium in the vadose zone overlying the local aquifer. The sand was air dried and washed with deionized (DI) water five times to remove impurities. The sand was then immersed in 4% HCl for 24 h, rinsed with deionized water, soaked in a 1 M NaOH solution for another 24 h, and finally washed repeatedly with deionized water until the filtrate had a pH of 5.5 [32,33]. The pH 5.5 used herein aimed to avoid precipitation of Cu. The sand was then dried at 105 °C overnight.

2.2. Preparation of Humic Acid (HA)-Coated Iron Oxide (Fe₃O₄) Nanoparticles (NPs)

The Fe₃O₄ NPs were purchased from the Shanghai Macklin Biochemical Technology Co., Ltd., China with a diameter of 20 nm (purity ≥ 99.5%). We added 3.4 g of the NPs to 1.0 L of ultra-pure water, mixed the solution in a sonicator bath for 6 h, and finally deflocculated the solution by heating it at 70 °C with sufficient stirring until the slurry changed to be a stable suspension. In order to avoid the particle aggregation, the NPs were ultrasonicated for 3 h (Fuyang F-060S, 450 W, 40 KHz) again prior to being coated with HA, the average size of NPs was 584 nm measured by ZetaSizer Nano ZS (Malvern Instruments, Worcestershire, UK) (Figure S1 in Supplementary Material). The stock suspension was stored at 4 °C.

We dissolved 1.7 g of HA (acquired from the Tianjin Guangfu Fine Chemical Research Institute, Tianjin, China) in 100 mL of 1 M NaOH under constant stirring at 25°C until the HA had dissolved into the solution. We then introduced ultrapure water to create a final solution with mass concentration ratio C_HA/C_NPs values of 10, 20, 30, and 50. Every mixture of HA and Fe₃O₄ NPs was stirred for 50 h; we obtained HA@NPs by injecting that mixture with 0.01 M HCl and 0.01 M NaOH until the pH reached a value of 5.5. The suspension of HA@NPs was then mixed with the cleaned sand at 25 °C until the HA@NPs fully coated the surface of the sand particles.

2.3. Characterization of HA, Fe₃O₄ NPs, and HA-Coated Fe₃O₄ NPs

Fourier transform infrared (FTIR) spectra of sand, HA, Fe₃O₄ NPs, and HA@NPs were recorded in the range of 600–4000 cm⁻¹ using a 330 FTIR microscope (Thermo Fisher Scientific, Waltham, MA USA) [33]. The zeta potential (ZP) and hydrodynamic diameter of the aforementioned four mediums was measured using ZetaSizer Nano ZS (Malvern Instruments, Worcestershire, UK) [34]. We measured the magnetic properties of the Fe₃O₄ NPs and HA@NPs using a vibrating sample magnetometer (Magnetic Property Measurement System XL-7, Quantum Design, San Diego, CA USA) following the process described by Ghosh et al. (2011) [35].

2.4. Column Transport Experiments

Transport experiments were conducted by pumping Cu²⁺ solution (50 mg L⁻¹) through an acrylic column packed with a sand matrix (Figure 1). The acrylic column was 8 cm in diameter and 50 cm in height; porous circular plates were installed at the top and bottom of the column to prevent particles from escaping. The column was packed with 3.4 kg of granular media with a density of 1.36 g/cm³. The porosity of the packed medium was 0.378 on average, and the hydraulic dispersion coefficient was estimated to be 0.178 cm² min⁻¹. Two peristaltic pumps at the inlet and outlet of the column were used to regulate the moisture content and to control the downward flow rate. The target moisture content (80% water saturation) was obtained according to the method developed by Esfandyari Bayat et al. 2015 [36]. The tensiometers (Meter T5x, München, Germany) at both ends read the same values, indicating uniform flow conditions.

There are two groups of column experiment in this study. (1) Group one was performed to examine the impact of the presence of HA coated Fe₃O₄ NPs (HA@NPs) embedded in porous medium on Cu²⁺ transport. Six types of packed medium in the columns were sand only (sand), sand containing Fe₃O₄ NPs (NPs + sand), 1HA@NPs + sand, 2HA@NPs + sand, 3HA@NPs + sand, and 5HA@NPs + sand, wherein the coefficients of HA, I = 1, 2, 3, and 5, mean i g HA embedded in 1 kg sand. The content of Fe₃O₄ NPs in the sand was 0.1 g kg⁻¹. Accordingly, the C_HA/C_NPs was raised from 0 to 10, 20, 30, and 50, respectively. The solution IS was 5 mM and pH = 5.5. (2) In group two, three types of packed medium were used: sand, NPs + sand, and 1HA@NPs + sand. The Cu²⁺ breakthrough curves under IS 0, 5, 10, 50, 100 mM were investigated in each medium to indicate the role of IS. The solution pH was 5.5.

For each column experiment, 10 pore volumes (PVs) of the background electrolyte solution of interest were first passed through the column to ensure that the column was fully equilibrated with this solution. This process also aimed to remove all possible re-mobile particles in the fixed-bed matrix in the column, and the spatial distribution of embedded particles in the column was determined immediately after this elution process. The mechanism for the stability of the packed medium was elucidated by Derjaguin–Landau–Verwey–Overbeek (DLVO) interaction energy calculations [33].

Then, suspension of Cu²⁺ within the same background electrolyte composition was injected into the packed column for 16.5 PVs (phase 1), followed by elution with Cu²⁺-free solution again (phase 2). The outflow from the columns was connected by an auto fraction sampler (Huxi CBS-A 100, Shanghai, China), and the effluent Cu²⁺ concentration was monitored using an inductively coupled plasma optical emission spectrometer (ICP-OES, Avio 200, PerkinElmer, Waltham, MA, USA). The effluent mass recovery rate (R_r) was calculated by integrating the BTCs.

2.5. Batch Adsorption–Desorption Experiments

Both batch adsorption and desorption experiments were carried out to explain the Cu²⁺ transport behaviors in different packed mediums and under different IS. In the adsorption experiments, three types of medium, sand, NPs + sand, and 1HA@NPs + sand were used as adsorbents. We measured the Cu²⁺ sorption kinetics by mixing 20 g of targeted adsorbent with 100 mL of Cu²⁺ solution (50 mg/L) with a pH of 5.5 and IS of 0, 5, 10, 50, and 100 mM in Teflon centrifuge tubes. The time required for equilibrium was set as 5 h following a previous kinetic study [37], and the differences of the concentration values of Cu²⁺ were actually less than 1% after 3 h indicating the feasibility of 5 h. The adsorption dynamics of Cu²⁺ in each adsorption system was obtained by measuring Cu²⁺ concentration at time interval of 5–60 min. The tubes at specific sampling time were centrifuged for 30 min at 12,000 rpm, and the supernatant was filtered by a 0.45-μm membrane, followed by Cu²⁺ measurement by ICP-OES. The Cu²⁺ desorption kinetics were then measured by mixing the Cu²⁺-rich porous medium with 100 mL of a 0.01 M NaCl solution and shaking the solution at room temperature (23–25 °C) for 5 h. The Cu²⁺ concentrations of the solutions were determined at the same time interval as the aforementioned adsorption experiment. The mechanisms underlying the adsorption process were further investigated using the pseudo-1st order and pseudo-2nd order rate equations to fit the adsorption experimental data as described by Chen et al. [38]. The adsorption equilibrium data were fitted using the Langmuir and Freundlich isotherms to provide quantitative information for adsorption isotherms, respectively [39].

3. Results and Discussion

3.1. Surface Properties of HA, Fe₃O₄ NPs, HA@NPs, and Sand

FTIR spectra were used to identify the surface functional groups for the HA, Fe₃O₄ NPs, 1HA@NPs, and sand (Figure 2a). For sand, the main peaks around 998 and 773 cm⁻¹ are attributed to symmetric Si-O-Si vibrations [40]. The FTIR spectrum of HA reveals a broad band at 3255 cm⁻¹, attributed to O–H and trace N–H stretching [41]. The peak at 1565 cm⁻¹ and 1378 cm⁻¹ are identified as N–H bending vibration and C–N stretching vibrations while peaks around wavelengths 1007 cm⁻¹ and 912 cm⁻¹ correspond to C–H stretching [42]. The band located at 1039 cm⁻¹ in the FTIR spectrum of Fe₃O₄ NPs indicates the symmetric stretching vibration of the Fe–O stretching. While the NPs are coated by HA, the main peak is assigned to C–H stretching. Both the surface hydroxyls (Si–OH) on sand and (3623 cm⁻¹) abundant –OH on HA can provide adsorption sites for Cu²⁺.

Figure 2b,c shows the zeta potentials and hydrodynamic diameters of different samples when IS varied from 0 to 100 mM. The electrokinetic measurements showed that the HA, 1HA@NPs and sand were negatively (approximately –25 mV) charged over the range of solution conditions investigated, and the Fe₃O₄ NPs is positively charged (approximately 25 mV). This indicates considerable electrostatic attractive forces between Cu²⁺ and these compounds. The particles size increased obviously with IS, while the zeta potentials became less negative. Increasing the IS results in compression of diffuse double layers around charged surfaces, thus reducing the repulsive forces between the particles and decreasing their stability to form aggregates [32]. By contrast, the HA coating increased the stability of Fe₃O₄ NPs (Figure 2c). As shown in Figure 2d, the hysteresis loops of Fe₃O₄ NPs and 1HA@NPs are standard S-shaped curves without remanence, indicating that the coated HA did not destroy the intrinsic superparamagnetism. Adding the HA coating to the NPs decreased the saturation magnetization value because HA has diamagnetic resistance, and a magnetically dead layer was formed with oxygen and other elements on the surface of the NPs [35]. Some studies demonstrated that the magnetic properties of NPs were attenuated through interaction with the HA surface groups [43,44,45], our magnetic coercivity values for NPs and 1HA@NPs are similar to those of previously published values [46]. Thus, HA reduced the magnetic force between the HA@NPs particles and improved their stability. Moreover, the decreased magnetization still could be observed after rinsing of the HA@NPs, it thus also proved the stability of the HA–NP complex.

3.2. Stability of HA-Fe₃O₄ NP-Sand Fixed-Bed in the Column

One important premise of the transport experiments is the stability of the fixed-bed matrix in the column. The breakthrough curves of Cu²⁺ should not be disrupted by the possible presence of mobile particles. Herein, we discuss the possible stabilization mechanisms for the HA-Fe₃O₄ NP-sand system. Calculated DLVO interaction energies (Φ) between sand and the NPs or 1HA@NPs are plotted as functions of separation distance in Figure 3, and the data are depicted in terms of kT, where k is the Boltzmann constant and T is Kelvin temperature. For NPs, as shown in Section 3.1, the positively charged surface made this scenario as a “favorable condition” for attachment on negative sand surface. NPs were readily deposited into the primary energy minimum (Φ_primary-min) (Figure 3a). For 1HA@NPs, while the presence of the high Φ_max would prevent the deposition of particles into the primary energy minimum (Figure 3b), they could be immobilized within the secondary energy minimum (Φ_second-min) [33]. Indeed, the calculated interaction energy profiles showed the presence of the Φ_second-min and the depth of the Φ_second-min increased with IS, which indicated that it became more likely for HA@NPs to be retained with sand in the column under high IS. The release of NPs and 1HA@NPs from the column during sequential elution experiment and retention profiles of the two particles after 10 PVs elution finished at IS = 0 mM are shown in Figure 3c,d. The release of immobilized particles could be negligible after 8 PVs elution (near 0 mg L⁻¹), and the distributions of particles along the column were uniform.

3.3. Effect of the HA-Coated Fe₃O₄ NPs Embedded in Sand Matrix on Cu²⁺ Transport

As shown in Figure 4, the normalized effluent Cu²⁺ concentration C/C₀ was plotted as a function of PV. The deposition of Cu²⁺ increased (C/C₀ decreased) with increasing HA content in all packed media (Figure 4a and

Table 1. Quantifying the breakthrough curves of Cu²⁺ under different experimental conditions.

Medium	IS (mM)	pH	(C/C0)max	Rr
Sand	5	5.5	0.859	66.96%
NPs + sand	5	5.5	0.857	66.56%
HA@NPs + sand	5	5.5	0.810	47.40%
2HA@NPs + sand	5	5.5	0.647	33.72%
3HA@NPs + sand	5	5.5	0.560	19.21%
5HA@NPs + sand	5	5.5	0.170	3.94%
Sand	0	5.5	0.811	62.98%
	5	5.5	0.859	66.96%
	10	5.5	0.878	66.71%
	50	5.5	0.932	78.85%
	100	5.5	0.967	87.39%
NPs + sand	0	5.5	0.832	65.71%
	5	5.5	0.857	66.56%
	10	5.5	0.863	70.21%
	50	5.5	0.918	79.75%
	100	5.5	0.926	85.92%
HA@NPs + sand	0	5.5	0.756	44.83%
	5	5.5	0.810	47.40%
	10	5.5	0.830	50.70%
	50	5.5	0.889	65.12%
	100	5.5	0.910	73.53%

). The maximum breakthrough concentrations, (C/C₀)_max, were decreased from 0.857 to 0.810, 0.647, 0.560, and 0.170 finally when C_HA/C_NPs was raised from 0 (NPs + sand) to 10 (1HA@NPs + sand), 20 (2HA@NPs + sand), 30 (3HA@NPs + sand), and 50 (5HA@NPs + sand), respectively. The corresponding R_r values were 66.56%, 47.40%, 33.72%, 19.21%, and 3.94%, respectively. As a control, the (C/C₀)_max and R_r in pure sand were 0.859and 66.96%. These results indicate that HA impedes the transportation of Cu²⁺.

As shown in Equations (1) and (2), Cu²⁺ was first adsorbed on the surface of HA@NPs and sand by electrostatic attraction. Then, as shown in Equation (3), Cu²⁺ complexed with the functional groups (–COOH, –C=O) on the surface of HA@NPs and sand, which was controlled by proton exchange between Cu²⁺ and H⁺ [34,47]. This HA@NPs that occurred in the potential layer of HA, is irreversible for Cu²⁺. Additionally, some Cu²⁺ attached on HA@NPs to produce the stable chelate, and thus Cu²⁺ was immobilized by the HA@NPs.

Sur(H⁺/Na⁺)_m + Cu²⁺ → Sur(H⁺/Na⁺)_m-r…Cu²⁺ + r(H⁺/Na⁺),

(1)

Sur-O⁻ + Cu²⁺ → Sur-O⁻…Cu²⁺,

(2)

Sur-(O_nH_m)^-n+m + Cu²⁺ → Sur-(O_nH_m-Cu)^-n+m-r+2 + rH⁺,

(3)

where “Sur” is the particle surface, “...” is an electrostatic action, and n, m, and r are constants. The Cu²⁺ ratio is stable in Equations (1)–(3) [48].

HA@NPs either adhered to the sand surface or accumulated in the porous medium AWI. According to Akbour et al. (2013) [10], the adsorption of HA on the surface of positively charged sand is primarily dependent on electrostatic interaction, and secondarily dependent on the interaction between the non-polar sites of the sand and the hydrophobic groups of the HA. Even if HA carried the same charge as the sand, HA could also be adsorbed onto the sand surface, and the amount of adsorption would surpass the amount of desorption because the HA function groups undergo ligand exchange with the sand [12,49]. The HA@NPs adhering to the sand might be caused by the similarity in the surface chemistries as HA. For the SWI and the AWI in the unsaturated porous medium, some studies indicate that the increased hydrodynamic diameter of HA@NPs could result in physical interception such as multi-particle bridging, steric hindrance, or surface deposition [23]. When the hydrodynamic diameters of the particles exceeded the AWI thickness, HA@NPs were more easily captured by the nearby AWI [26]. Additionally, the oscillating ring produced by water absorption between the particles hindered the migration of HA@NPs [50].

3.4. Effect of Ion Strength (IS) on Cu²⁺ Transport in Diverse Unsaturated Porous Media

Figure 4b–d shows the transport of Cu²⁺ in three types of porous medium with different IS values. The (C/C₀)_max values of Cu²⁺ in each BTC and R_r are shown in Table 1. Taking the medium HA@NPs + sand as an example, when the value of IS increased from 0 to 5, 10, 50, and 100 mM, the (C/C₀)_max values of Cu²⁺ increased from 0.756 to 0.810, 0.830, 0.889, and 0.910, respectively, while the corresponding R_r = 44.83%, 47.40%, 50.70%, 65.12%, and 73.53%, respectively. This IS-dependent Cu²⁺ transport can be attributed to the competitive adsorption between Na⁺ and Cu²⁺. As the IS of the solution increased, a large amount of Na⁺ ions occupied the negatively charged adsorption sites on the surface of the sand and HA [51]. Those Na⁺ ions neutralized the negative charges on the surface of the sand and HA, which reduced both the electrostatic force between the Cu²⁺ and the negative ions adsorbed at the electrical adsorption sites and the total number of Cu²⁺ ions on the surface of sand and HA. Based on the limit equation and the Bader–Shock theory, IS is negatively correlated with the activity coefficient of ions; the higher IS value corresponds to a lower initial activity of Cu²⁺ in the solution [52]. As a result, more Cu²⁺ were transported through the column.

3.5. Adsorption Characteristics of Cu²⁺ on Different Types of Media

Figure 5 depicts the time-dependent Cu²⁺ adsorption and desorption capacity of three media across varying IS (0, 5, 10, 50, and 100 mM), pure sand, NPs + sand, and 1HA@NPs + sand. In all adsorption media, the Cu²⁺ concentration in the liquid phase exhibited a rapid decrease within the initial 1–1.5 h which contributed to > 90% of the total adsorption, followed by a gradual decrease. Compared with the adsorption processes, the desorption of Cu²⁺ was relatively weak (<6 mg/L after 300 min). The aqueous Cu²⁺ concentration in the desorption system was an order of magnitude lower than that in the adsorption system. This is in accord with the observation in transport experiments. The C/C₀ of Cu²⁺ quickly reached 0 after elution by 4–5 PVs Cu²⁺-free solution (phase 2), suggesting that the release of pre-deposited Cu²⁺ could be negligible under the same IS.

Table 2. Fitted parameters of the adsorption kinetics models for the adsorption of Cu²⁺ in different media, with the IS of 0–100 mM.

Porous Medium	Parameters	IS (mM)
		0	5	10	50	100
Sand	qe,exp (mg kg−1)	0.0698	0.0538	0.0458	0.0308	0.0243
	qe,cal (mg kg−1)	0.0706	0.0558	0.0464	0.0312	0.0255
	k2 (min−1)	0.0018	0.0017	0.0016	0.0015	0.0011
	R2	0.996	0.995	0.991	0.984	0.979
NPs + sand	qe,exp (mg kg−1)	0.0673	0.0643	0.0625	0.0323	0.0215
	qe,cal (mg kg−1)	0.0691	0.0666	0.0643	0.0394	0.0230
	k2 (min−1)	0.0022	0.0021	0.0021	0.0020	0.0018
	R2	0.997	0.998	0.996	0.986	0.955
1HA@NPs + sand	qe,exp (mg kg−1)	0.1380	0.1100	0.1030	0.0990	0.0895
	qe,cal (mg kg−1)	0.1390	0.1110	0.1000	0.1050	0.0895
	k2 (min−1)	0.0028	0.0022	0.0021	0.0021	0.0018
	R2	0.999	0.999	0.999	0.999	0.999

and Figure 6a–c present the data fitting results of the pseudo-2nd-order kinetic model, which delineates that the entire adsorption process is generally controlled by chemisorption [39]. The results demonstrate the superior performance of the 2nd-order model with correlation coefficients R² > 0.98 and modeled q_e values closer to the measured results. The R² values of the pseudo-1st-order model were all less than 0.89 (data not shown here). This indicates the greater suitability of the pseudo-2nd-order model in describing the adsorption mechanism and consequently illustrates the predominance of chemical interactions thereof. The fitting results of the Cu²⁺ adsorption isotherm data indicated that the Langmuir model exhibited a stronger correlation (R² = 0.987–0.997) than the Freundlich model (

Table 3. Fitted parameters of the isotherm models for the adsorption of Cu²⁺ in different media, with an IS of 5 mM.

Models	Parameters	Porous Medium
		Sand	NPs + Sand	1HA@NPs + Sand
Langmuir	qm (mg kg−1)	79.365	90.909	130.208
	KL (L mg−1)	0.075	0.121	0.133
	R2	0.987	0.997	0.997
Freundlich	1/n	0.490	0.364	0.337
	KF (L mg−1)	10.132	20.696	31.414
	R2	0.924	0.922	0.985

and Figure 6d). This suggests that the binding of Cu²⁺ on the porous medium in testing is a monolayer sorption.

The influences of IS and HA coating on Cu²⁺ adsorption thus supports the results of Cu²⁺ transport in the columns. The increasing HA concentration was linked with an enhancement in Cu²⁺ adsorption on the media, which was confirmed by the increase in k₂ and K_L (Table 3). Electrolytes inhibited Cu²⁺ adsorption and accelerated Cu²⁺ desorption on the surface of mediums. The scenario with higher IS displayed lower k₂, which thus led to more Cu²⁺ breakthrough in the column. This phenomenon could still be explained by either the competition for adsorption sites [53] or the competition for complexation sites on the outer-sphere surface between Na⁺ and Cu²⁺ [54,55]. An X-ray energy dispersive spectrometer mapping graph was provided in Figure S2 to confirm the Cu adsorption.

4. Conclusions

In this study, we conducted laboratory experiments in order to investigate the transport of Cu²⁺ in unsaturated columns with humic acid/iron oxide NP (Fe₃O₄) amendment under different IS conditions. We found that the presence of HA stabilized Fe₃O₄ NPs in a sand matrix and was effective in restraining the transport of Cu²⁺. The mobility of Cu²⁺ in the packed sand columns was enhanced at higher IS values. Complexation, electrostatic attractive force, hydration, oxygen-containing functional groups, surface attachment, and AWI attachment are all mechanisms or features that may be responsible for the retention and transport of Cu²⁺ in the unsaturated porous media embedded with HA@NPs. These conclusions were supported by the adsorption tests and modeling analyses. The presence of HA and NPs in the medium, and the IS change altered the fate, transport and diffusion risk of Cu²⁺ contamination during unsaturated seepage. Note that only monovalent electrolyte solution (NaCl) was used for IS modulation in the current study. We suggested that studies of the influences of cation composition in divalent solutions (Ca²⁺ or Mg²⁺) or in mixed systems (Na-Ca-Mg) are necessary, because Na⁺, Ca²⁺ and Mg²⁺ generally coexist in natural groundwater and their relative composition varies with hydro-geochemical process. In addition to Fe₃O₄ NPs, many natural organic/inorganic colloids and engineered nanomaterials are present in the soil environment, which also tend to associate with dissolved organic materials and heavy metals. Our findings thus provide a reference for delineating the ternary interactions and the consequences of these interactions, which will be an improved comprehension of the environmental risk of heavy metals in complex systems containing organic materials and colloids/NPs.