Exploring the Potential of TNT and Aniline Coexistence to Enhance Their Transports in Saturated Chinese Loess

Abstract

To determine the interactions between TNT and aniline adsorptions and the potential to enhance their transports in saturated Chinese loess, batch and column tests were conducted. The batch tests show that their adsorptions inhibit each other when they coexist, and their inhibitions depend on their concentrations, implying that their coexistence has the potential to enhance their transports of each other in the saturated loess. The column tests confirm this speculation, while aniline enhances TNT transport more obviously than TNT does. These findings are ascribed to TNT adsorption being primarily through surface adsorption, while aniline adsorption mainly takes place via electrostatic adsorption and inner pore diffusion adsorption, as well as surface adsorption. There is a certain competitive relationship in their adsorptions on the loess because they have same and different adsorption sites; in particular, electrostatic force is greater than surface force. Therefore, these inhibitions on adsorption are conducive to the existences of TNT and aniline in the water rather than being fixed on the loess, thus enhancing their transports in the saturated loess, indicating that their coexistence can increase the risk of soil and even groundwater pollution.

1. Introduction

With the rapid development of industry, agriculture, and urbanization, the combined pollution caused by the simultaneous coexistence of two or more pollutants is a common phenomenon in soil and groundwater that has become a critical problem in the field of environmental remediation [1,2,3]. Exploring the environmental behaviors of combined pollutants is an essential issue for their remediation over decades, and the coexisting pollutants, including heavy metals (e.g., Cu, Pb and Cd) and organic matter (e.g., 2,4,6-trinitrotoluene (TNT), aniline and toluene), have been studied extensively to explain their adsorption characteristics on ambient media [4]. It was found that adsorption had a notable influence on the transport of objective pollutants in soil and groundwater [3,4]. As we known, the surface properties of adsorbents can be changed by adsorbing one pollutant, thereby affecting the adsorption for other pollutants [3,5]. Consequently, it is reasonable to speculate that the interaction between multiple pollutants may inevitably affect their transport in soil and groundwater systems. As expected, several studies conducted on the combined pollution of heavy metals have proved that there are interactive effects between heavy metals’ adsorptions and transports [6,7], but few studies have been conducted on multiple organic pollutants [8]. For example, pollution in military training ranges is typically a combination of two or more organic compounds due to the incomplete detonation of munitions, i.e., TNT, hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX), and octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine (HMX), which has led to the transportation of these compounds into groundwater [9,10]. TNT has been found in groundwater at some munition production sites, including the Nebraska Ordnance Plant [11]; nevertheless, much of the TNT pollution remains in or near the soil surface, and TNT concentration and its plume are smaller than the associated RDX in the studied groundwater [11]. However, Craig and Taylor (2011) [12] reported that the concentration of TNT (4 mg/L) was higher than RDX concentration (2.5 mg/L) in their studied groundwater. Obviously, there is an inconsistency between the two studies and that is a puzzle. Although the inconsistency is likely due to microbial activity, no specific microorganisms were identified to be responsible for the degradation of these organic pollutants [13]. In our view, we speculate that the inconsistency might strongly depend on the coexistence of multiple organic pollutants in the soil–water system, which then result in certain interacting behaviors, including adsorption and transport among coexisting pollutants. And yet, to the best of our knowledge, little attention has been given to exploring the interactions and competitive effects of one organic pollutant on the other, and no specific study conducted to solve the puzzle has been reported.

Loess is a special soil extensively distributed in Central Asia, the Middle East, Russia and North America [13]. The Loess Plateau in north-central China is the most concentrated and largest loess area. It is generally considered that the loess is transported by air and then deposited in the manner of a loosely packed soil skeleton without chemical weathering. The loess particles are mainly silty quartz with dimensions of less than 75 µm, and the small quartz grain is reported to be produced by energetic processes like glacial grinding and tectonic activities [14]. In general, partition is always involved unavoidably in the process of objective pollutant adsorption on the studied medium because soil organic matter (SOM) is an important part of the soil or sediment [15], and thus it is difficult to determine the contribution of adsorption and characterize the interactions of coexisting pollutants. Unlike other soil, loess is very low in SOM, with calcite as the main component [14]. Meanwhile, lots of studies have proved that loess has great potential as a promising adsorbent to adsorb numerous pollutants including organic compounds (such as TNT, aniline, toluene and naphthalene) [16,17] and heavy metals (such as Cd, Pb and Zn) [13,18]. Hence, given these aspects, loess is an ideal adsorbent in natural environmental pollution remediation for studying interactions of multiple organics.

Currently, multiple organic pollutants coexisting in soil and groundwater pose the most challenging issue in the deterioration of the health of soil and groundwater ecosystems. Here, it is particularly important to understand their occurrence, transport and impact for pollution control. Therefore, this present study was carried out with Chinese loess sampled from Xi’an, China, to investigate the characteristics of the interactions between TNT and aniline adsorptions, and the potential of TNT and aniline coexistence to enhance their transports in a loess–water system, and then to clarify the mechanism enhancing their transport in a saturated loess–water system. The expected findings can provide new perspectives for assessing the potential risk of multiple pollutants coexisting in a soil–water system.

2. Materials and Methods

Chinese loess, as a typical Quaternary loess, was sampled from Xi′an, China. Following previous reports [19,20], TNT and aniline chemical reagents were used, and batch and column tests were conducted in the present study. Batch tests were performed by mixing 20 g of the prepared loess with 200 mL TNT solution at concentrations ranging from 10.0 to 100.0 mg/L in the flask. Column tests were conducted with a polymethyl methacrylate tube (8 cm × 45 cm, internal diameter × length). The 25.0 mg/L TNT solution was continuously dispensed to each loess column from the top at a flow rate of 40 mm/h using a peristaltic pump. The same procedures were followed to test aniline and to test for coexistence.

The physicochemical properties of the loess were determined according to the national standard method [13,19]. The morphological characteristics were tested by scanning electron microscopy. The surface area was measured using the N₂ adsorption method by Autosorb 1-MP [21]. Organic matter content was determined via the potassium dichromate oxidation method [21]. The pH, zeta potential, and the cation exchange capacity of the loess was also determined as Hu et al. (2016) reported [21]. The primary chemical constituents were characterized by inductively coupled plasma–mass spectrometry. X-ray powder diffraction (XRD) was employed to characterize the crystal of the loess operated at 40 kV and 30 mA with Cu K_α1 radiation (k = 0.154 nm). The XRD data were collected in the scanning range of 2θ = 10~90°.

Aniline and TNT were analyzed via high performance liquid chromatography [19,20,22]. The experiment data were analyzed via the statistical software SPSS 11.5 and the results were considered statistically significant at a 95% confidence limit, corresponding to a probability level p ≤ 0.05 [22].

3. Results and Discussion

3.1. Chinese Loess Characterization

The characterization of Chinese loess has been reported in other references [13,20]. Most of its physical–chemical properties, including chemical constituents, specific surface area, pH and Zeta potential, can be found in the references [13,20]. It should be noted that the SOM of the loess without human disturbance is rather low and just about 0.56% (5.6 mg/g), which is lower than the SOM of the other land-use loess and the soil closed to that of wasteland, proving that SOM is subjected to human activities while SOM is very low for Chinese loess without human disturbance [14]. Thus, the pollutants were fixed on the loess resulting from adsorption rather than partition, because the partition is controlled by SOM [23] as we all know.

The SEM micrographs of the loess (Figure S1) show that the loess particles have many macro-micropores and micrometric particles with different sizes. Furthermore, abundant disoriented particles and bulges stacked randomly can also be clearly observed. The loess has abundant pores and heterogeneity of surface structure. This special structure guarantees the loess a large specific surface area, up to 24.1 m²/g, suggesting that loess has a high affinity to pollutants in soil–water systems [13,20]. As expected, loess is an ideal medium to use as an adsorbent in the present study.

3.2. Batch Tests

3.2.1. Adsorption Kinetics

The amount of TNT or aniline absorbed on per unit mass of the loess (q_t, mg/g) can be calculated by the formula ${\mathrm{q}}_{\mathrm{t}}=\left({\mathrm{C}}_0-{\mathrm{C}}_{\mathrm{t}}\right)·\mathrm{V}/\mathrm{m}$, where C₀ is the initial concentration of TNT or aniline (mg/L), C_t is the concentration of TNT or aniline at a reaction time (mg/L), V is the solution volume (L), and m is the mass of the loess (g), respectively. The adsorption kinetic test data (Figure S2a) show the TNT amount adsorbed on the loess (q_t) during the reaction time at TNT initial concentration (C₀) from 25.0 to 100.0 mg/L at an interval of 25.0 mg/L under 25 ± 2 °C, and their maximum adsorption amounts were 118.78, 239.15, 299.60 and 359.61 mg/kg, respectively, suggesting that the adsorption capacity increases with an increasing initial concentration. The higher the TNT initial concentration, the greater the driving force between the loess and water, and the loess adsorption ability is enhanced. This phenomenon was also reported by Arthur et al. (2017) [24].

TNT adsorption on the loess has two stages: a fast adsorption stage and a slow adsorption stage (Figure S2). The fast stage occurred within 4 hours, when almost 80% of the adsorption was completed in this stage, and then was followed by slow stage until reaching a steady state. The equilibrium times were 10, 9.5, 8 and 7 hours for TNT initial concentrations of 25.0, 50.0, 75.0 and 100.0 mg/L, respectively. This illustrates that the adsorption equilibrium time was shortened as the TNT initial concentration increased, as a great concentration difference between loess and water enhanced the initial adsorption rate. The equilibrium time was close to that of raw soil with 3.03% SOM reported by Hao et al. (2018) [25], and much shorter than that of soil studied by Dontsova et al. (2006) [26] because SOM was richer than the current loess used. The equilibrium time was approximately 18 h in 2% and 5% fulvic acid (FA) soil, and even reached 20 h in 10% FA soil [25]. Similarly, Eriksson et al. (2004) [8] reported that the equilibrium time of TNT bound to dissolved organic matter (DOM) extended to about 1 day. These results further confirm that TNT is fixed on the loess via adsorption rather than the partition of SOM.

The kinetic data were analyzed using four kinetic equations, including the pseudo-first order kinetic (PFOK) (Equation (S1)), the pseudo-second order kinetic (PSOK) (Equation (S2)), the Weber–Morris intraparticle diffusion model (WMIDM) (Equation (S3)) and the Elovich equation (EE) (Equation (S4)) [27,28,29], and the results are provided (Table S1).

The four kinetic models fit well to describe TNT adsorption (Table S1), and the coefficient values (R²) were in the order PSOK > PFOK > WMIDM > EE in all the tests. The ranges of R² for PSOK are from 0.998 to 0.999, so are close to 1.0, and the theoretical adsorption amounts are 40.98, 115.06, 252.49 and 375.94 mg/kg, calculated based on the equation, and are consistent with the experimental values of 43.12, 127.56, 239.15 and 359.61 mg/kg. The results suggest that TNT adsorption on the studied loess perfectly complies with PSOK. Hao et al. (2018) [25] also found a similar phenomenon, indicating that TNT adsorption on loess particles was a controlled physical and chemical process involving complexation and diffusion to surface active sites within the loess soil matrix [30,31], and the adsorption was considered to be a complex mass transfer process depending on liquid film diffusion, surface adsorption and intraparticle diffusion [32].

The initial adsorption rate (h) (mg/(kg·min)) can be determined by the formula $\mathrm{h}={\mathrm{k}}_2{·\mathrm{q}}_{\mathrm{e}}^2$ for TNT adsorption on the loess [33]. The values of h increased with TNT concentration, increasing from 10.0 to 100.0 mg/L (Table S1). The change was due to the increase in TNT adsorption [34], and the high h values of 2.762 mg/(kg·min) and 3.676 mg/(kg·min) for TNT concentration at 50.0 mg/L and 100.0 mg/L demonstrated that TNT adsorption on loess might be via surface interchange mechanisms until all the surface-active groups were saturated [35].

The R² value of PSOK is more than 0.9, indicating that PSOK is also suitable for describing the adsorption process of TNT on the loess, and the process is largely due to physical adsorption, while film diffusion is a rate-limiting factor [34]. The regression was conducted by WMIDM and, as expected, its R² values were close to 0.9 (Table S1), implying that the TNT adsorption process is attributed to particle diffusion through the solution to the external surface of the loess or the boundary layer of the solute molecule, where intraparticle diffusion is a rate-limited step [35]. This further proves that not only film diffusion but also intraparticle and boundary layer diffusion are the adsorption-controlling factors [31]. However, their parameters, including k₃ and c, are rather small and almost unchanged, while only decreasing a little with increasing TNT concentration, from 10.0 mg/L to 100.0 mg/L, showing that the control potential is rather small and can even be ignored. Thus, the results demonstrate that TNT adsorption on the loess is largely due to surface physical adsorption and then chemical adsorption. Compared with PSOK, the R² values of EE are small, close to 0.8, confirming that TNT adsorption is not on the whole due to chemical adsorption [18].

Like the TNT adsorption on the loess, aniline adsorption has also two stages (fast and slow) and then reaches equilibrium (Figure S2b). Regardless of whether the initial concentration of aniline is 10.0, 25.0 or 75.0 mg/L, the experimental results fitted well with four studied models (Table S1), implying that the adsorption of aniline and TNT on the loess has similar mechanisms. However, compared with TNT adsorption, the fast and slow stages of aniline adsorption were short, 40 min and 80 min, and the equilibrium time was 120 min, which is significantly shorter than in the report by Eriksson et al. (2004) [8]. The adsorption capacities of aniline were 77.92, 159.38 and 286.67 mg/kg at initial concentrations of 10.0, 25.0 and 75.0 mg/L, larger than those of TNT under the same conditions [32].

The parameters of aniline adsorption kinetics are also listed in Table S4. Obviously, the values of R² are in an order of PSOK > PFOK > WMIDM > EE. The results show the four models are also suitable to describe the adsorption process of aniline, suggesting that aniline and TNT share similar adsorption mechanisms. Furthermore, the FTIR of the loess before and after the adsorption of TNT and aniline presented few differences (Figure S3), indicating that the basic physical and chemical functional group structures of the loess had not been altered after the adsorption of TNT or aniline, further proving the similarity of TNT and aniline adsorption mechanisms. Still, there are also obvious differences in their corresponding parameters, such as the rate constants of k₁, k₂, k₃ and h for aniline being much greater than those for TNT, indicating that aniline is more easily adsorbed by the loess than TNT. In particular, the value of k₃ for aniline is more than one hundred times that for TNT, illustrating that aniline has a stronger intra-particle diffusion driving force on the loess than that of TNT.

Moreover, at the same experimental conditions, the adsorption capacity and rate of TNT were higher than those of aniline, but its equilibrium time was longer than that of aniline. The h value of PSOK for TNT adsorption was smaller than that for aniline adsorption (Table S1). The adsorption process for TNT is mainly controlled by surface physical adsorption at the first stage, while aniline is controlled by surface electrostatic adsorption. These differences in adsorption behaviors are attributed to their differences between adsorption processes. Firstly, the molecular size of aniline (7.0 Å) is larger than that of TNT (5.9 Å), while the small molecules can easily access into the pores and surfaces of loess, and secondly, the solubility of aniline and TNT in the water is 34.97 and 0.13 g/L at 25 ℃, respectively; as the solubility decides the hydrophilic and hydrophobic interactions, the adsorption of a hydrophilic compound is less favored than a hydrophobic one, so aniline adheres more easily on loess than TNT.

3.2.2. Adsorption Isotherms

The adsorption isotherm could describe the distribution of TNT and aniline between aqueous and loess, and then the isotherm data were simulated by three mathematical equations—the Langmuir (Equation (S5)), Freundlich (Equation (S6)) and Redlich Peterson models (Equation (S7)) [28,36] and the results were obtained (Table S2).

It was found that (Table S2) the adsorption of TNT was well-fitted to the three adsorption isotherm models, with correlation coefficients R²>0.90, and the order of R² value was Freundlich> Langmuir> Redlich–Peterson model. Obviously, the test data were more consistent with the Freundlich model, indicating that physical interaction is the main mechanism of adsorption. The predicted parameter K_F in the Freundlich model presents an increasing trend, i.e., 11.69, 16.18, 25.16, 37.32 and 48.15 ((mg/kg)·(L/mg)^1/n), with the temperature increasing from 278 to 318 K, implying that the process of TNT being adsorbed by the loess is an exothermic reaction. The parameter n is 1.69, 1.58, 1.59, 1.68 and 1.77 at 278, 288, 298, 308 and 318 K, respectively, and these values are much greater than 1.0. These results suggest that the Freundlich equation can also describe the adsorption behavior of TNT on the loess well. Combining the results observed by Sharma et al. (2013) [15], TNT adsorption on the loess is mainly monolayer adsorption [37] and the loess surface heterogeneity and anisotropy have few influences on the adsorption [32]. These further prove that TNT adsorption on the loess is mainly physical adsorption supplemented with chemical adsorption. Similarly, the R² values of the Langmuir model are more than 0.94, also illustrating that TNT adsorption on the loess is more like monolayer adsorption. The predicted parameter q_m in the Langmuir model shows an increasing trend, i.e., 230.42, 390.63, 529.10, 609.06 and 786.32 mg/kg with temperature increasing from 278 to 318 K, illustrating that TNT adsorption is sensitive to temperature, and this is consistent with the results of the Freundlich isotherm analysis.

The R² values of the Redlich–Peterson model were greater than 0.91, showing that TNT adsorption on the loess involves monolayer and multilayer adsorption simultaneously [32]. The exponent g reflects the heterogeneity of the adsorbent, and when g values are equal to 0 and 1, the Redlich–Peterson model converges to Henry and Langmuir isotherms, respectively. The present study shows that g is in the range of 0.53~0.54, revealing the loess is heterogenous [37], which is in agreement with the SEM results (Figure S1). The values of the predicted parameter B in the Redlich–Peterson model are less than 1.0, further proving that the adsorption has a characteristic of the Langmuir adsorption model, where the adsorption mainly occurs on the surface of the loess and is dominated by monolayer adsorption accompanied by multilayer adsorption [20].

By comparing their adsorption behaviors (Table S2), aniline adsorption has similar characteristics to TNT adsorption regarding the loess, and also has some differences. Firstly, the values of the parameter n in the Freundlich model for aniline adsorption are 4.25, 4.52, 4.67, 5.15 and 5.17, which are much greater than those for TNT adsorption, demonstrating that aniline is subject to nonlinear adsorption rather than linear adsorption. Meanwhile, the values of K_L obtained for aniline (from 4.23 to 5.21 L/mg) are greater than the values for TNT (from 2.09 to 2.60 L/mg), indicating a high energy for aniline adsorption [18]. Secondly, from the results of the Redlich–Peterson model, both g and B values for aniline adsorption (0.68, 0.70, 0.53, 0.62, 0.63 and 143.26, 114.84, 93.21, 71.53, 68.32) are greater than those for TNT adsorption (0.58, 0.53, 0.53, 0.54, 0.55 and 0.37, 0.39, 0.41, 0.42, 0.46), illustrating that the isothermal adsorption tends towards a Freundlich model. All the B values for aniline adsorption were higher than 1.0, which means that the adsorption mainly occurs on the heterogeneous surface of the loess via multilayer adsorption supplemented with monolayer adsorption. In this regard, aniline adsorption on the loess is mainly chemical adsorption followed by physical adsorption, which agrees with the results obtained from the kinetic analysis. Thirdly, the values of q_m from the Langmuir model for aniline adsorption did not increase as TNT adsorption but decreased from 208.77 to 167.32 mg/kg, with the temperature increasing from 278 to 318 K, proving that the aniline adsorption process is exothermic and benefits from low temperatures.

3.2.3. Thermodynamic Analysis

Based on Gibbs free energy theory, the thermodynamic behavior of adsorption processes can be determined using Equations (S8) and (S9) [38]. Thus, TNT adsorption on the loess was investigated at a given TNT concentration with different temperatures (278~318 K), and these data were analyzed and the results were obtained (Table S3).

ΔG⁰ was negative with a range from −8.93 to −3.91 kJ·mol⁻¹ for TNT adsorption on the loess, demonstrating the spontaneity of adsorption. It is also evident that the adsorption process was not only a physical process but also driven by a chemical synergistic effect. ΔG⁰ increases with decreasing temperature, and the increase range doubled, approximately, implying that the potential for spontaneity becomes stronger, and therefore TNT adsorption on the loess is sensitive to temperature and favorable with high temperatures.

The positive values of ΔH⁰ were obtained for TNT at all tested temperatures, i.e., 408.24, 278.02, 200.68, 158.19, 143.18 and 130.18 kJ·mol⁻¹, indicating that the adsorption process is endothermic and driven by the synergistic forces of physics and chemistry, and that the binding between soil particles and TNT has physical and chemical effects [39]. Similar phenomena have also been found in the studies on NH₄⁺ [33] and the adsorption of heavy metals like Pb and Cd on loess [13]. The positive value of ΔS⁰ implies that TNT is less hydrated in the loess layer than in the aqueous solution, and the orderliness is increased between TNT molecules and loess in the studied system [40]. Moreover, Hefne et al. (2008) [33] noted that the positive value of ΔS⁰ appeared due to the redistribution of energy between adsorbate and adsorbent. The distribution of rotational and translational energy among some molecules will enhance adsorption by producing positive ΔS⁰, and simultaneously the randomness will increase at the solid–solution interface during the process of adsorption. Thus, the absolute values of ΔS⁰ are small, at 1.0 and even 0.48, suggesting that there is a weak binding force for TNT adsorption on the loess. All this understanding is consistent with the results obtained in the previous adsorption kinetics and isotherms above.

The analysis of ΔG⁰ showed that aniline adsorption on the loess was also feasible and spontaneous thermodynamically (Table S4). However, the increase in ΔG⁰ values with temperature decrease was not like that for the results from TNT adsorption, implying that the potential for spontaneity becomes weaker, and the adsorption process is exothermic and favored by low temperatures [41]. In particular, the absolute value of ΔG⁰ for aniline is much higher than that for TNT, showing that the loess tends to adsorb aniline rather than TNT, and this is on account of their different chemical properties, as aniline has a stronger polarity and higher ratio of chemical adsorption compared to TNT.

Negative values of ΔH⁰ were obtained for aniline adsorption at all studied temperatures, further confirming that the process is exothermic, unlike that for TNT. The ΔH⁰ values increase with temperature increase, and their absolute values are much smaller than those of TNT, showing that aniline adsorption on the loess is far less sensitive to its concentration than that of TNT, which is consistent with the analysis obtained from isothermal experiments. This is also supported by the negative values of ΔS⁰, which are −47.27, −37.77, −34.52, −29.25, −27.92 and −22.28 kJ/(mol·K), respectively. The negative values of ΔS⁰ increase with temperature increasing and become doubled under the test conditions; moreover, the absolute values of ΔS⁰ for aniline are two orders of magnitude greater than those for TNT, revealing that the driving force for aniline adsorption is relatively larger than that for TNT adsorption.

In general, the TNT adsorption on the loess is endothermic and dominated by a monolayer accompanied by multilayer adsorption, and the surface heterogeneity and anisotropy of the loess have little influence on the adsorption. Aniline adsorption is exothermic, and the adsorption mainly occurs on the heterogeneous surface of the loess via multilayer adsorption supplemented by monolayer adsorption. Obviously, the adsorption performances of aniline and TNT on the loess are different, which would likely induce their interaction when they coexist in a loess–water system.

3.2.4. Interaction between TNT and Aniline Adsorption on the Saturated Loess

The effect of TNT on aniline adsorption on the loess is shown in Figure 1, under initial aniline concentrations of 10.0 and 25.0 mg/L.

Aniline adsorption rates decreased gradually from 80.62 and 66.19 to 73.96% and 62.03%, with TNT concentration increasing from 0 mg/L (control) to 25.0 mg/L, and the decreased percentage was 6.66 and 4.16% at the initial aniline concentrations of 10.0 mg/L and 25.0 mg/L (Figure 1), respectively, showing that TNT and aniline compete in their adsorption on the loess and that TNT has an inhibitory potential for aniline adsorption. The inhibition increases with TNT concentration increase, which is supported by other reports [1,41]. The inhibition of TNT on aniline adsorption is ascribed to its similar adsorption sites and driving forces, including Van der Waals, etc. Moreover, the statistical analysis of the tests shows that there is a linear relationship between initial TNT concentration (C) and aniline adsorption rate (ƞ %), and Equations (1) and (2) were obtained, indicating that

$$ƞ=-0.2637\mathrm{C}+80.68\hspace{1em}\left(\mathrm{aniline}\mathrm{concentration}10.0\mathrm{mg}/\mathrm{L}\right)$$

(1)

$$ƞ=-0.1709\mathrm{C}+66.31\hspace{1em}\left(\mathrm{aniline}\mathrm{concentration}25.0\mathrm{mg}/\mathrm{L}\right)$$

(2)

where the inhibition effect is more obvious with the aniline concentration of 10.0 mg/L compared to that of 25.0 mg/L, as the slope of the former straight line is higher than the latter.

Figure 2 shows the effect of aniline on TNT adsorption on the loess. The adsorption rate of TNT decreases as the aniline concentration increases, which means that aniline inhibits TNT adsorption, as TNT does to aniline adsorption. With an aniline concentration increase from 0 (control) to 25.0 mg/L, TNT adsorption rates are drawn from 48.11 and 38.56% to 6.23 and 28.15%, and the decline ranges are 21.88 and 10.41% with 50.0 and 100.0 mg/L TNT concentration. Obviously, compared with the effect of TNT on aniline (Figure 1), the decline of aniline on TNT adsorption is far greater than that of TNT on aniline, indicating that aniline has a stronger influence on TNT adsorption on the loess than that for its adsorption affected by TNT. However, when aniline concentration is lower than 10.0 mg/L, aniline has hardly any influence on TNT adsorption on the loess under the experimental conditions, and this is a significant difference from the influence of TNT on aniline adsorption. This phenomenon can be ascribed to the molecular polar force being larger than the surface force driving aniline adsorption on the loess preferentially over TNT. More coincidently, TNT adsorption on the loess is mainly via surface physical adsorption, so aniline at low initial concentrations is directly adsorbed by the loess rather than being in competition with TNT [41]. However, when the aniline concentration is high, surplus aniline molecules compete with TNT adsorption on the loess, and the influence becomes obvious and strong. Thus, there is also a linear relationship between aniline and its influence on TNT adsorption, as that for TNT done on aniline, and then the relationship can be described as in Equations (3) and (4), with initial TNT concentrations of 50.0 and 100.0 mg/L, respectively.

$$ƞ=-0.2185\mathrm{C}+46.811(\mathrm{TNT}\mathrm{concentration}50.0\mathrm{mg}/\mathrm{L})$$

(3)

$$ƞ=-0.0677\mathrm{C}+34.545\left(\mathrm{TNT}\mathrm{concentration}100.0\mathrm{mg}/\mathrm{L}\right)$$

(4)

Obviously, the inhibition of aniline on TNT adsorption has a significant relationship to initial TNT concentration, and decreases with increases in its concentration (Equations (3) and (4)). The slopes of these two lines are 0.2185 and 0.0677, which are very close to the 0.2637 and 0.1709 obtained by TNT influence on aniline adsorption on the loess (Equations (1) and (2)). The characteristics are very similar, which further proves the correctness of the mechanism analysis for TNT and aniline adsorption on the loess.

3.3. Column Tests

The results of the batch tests reveal that there is a significant interaction between TNT and aniline adsorption on the loess, and we speculate that the interaction may also exist in their transportation when they coexist. To identify the hypothesis, two groups of column tests were conducted and the results are shown in Figure 3.

In the first group tests (Figure 3a), when aniline and TNT exist alone, the breakthrough points of the start and end are 4.5 d and 3.5 d, 7 d and 7 d with the aniline and TNT concentrations of 10.0 mg/L and 100.0 mg/L, respectively. These show that aniline has a smaller mobility than that of TNT, which agrees with the conclusion obtained from the batch tests: that the adsorption capacity of aniline is larger than that of TNT. As a comparison, when aniline and TNT coexist, their start and end breakthrough points are changed to 3.5 d and 2.5 d, 6.5 d and 7 d at 10.0 mg/L and 100.0 mg/L of aniline and TNT, respectively, which also shows that the mobility of aniline is smaller than that of TNT, and the same phenomenon is reported by Dontsova et al. (2006) [26]. More significantly, compared with the breakthrough points for aniline and TNT existing alone at 10.0 mg/L and 100.0 mg/L, their breakthrough points at the start and end appear ahead, proving that there is indeed interaction between aniline and TNT adsorptions on the loess, and they promote each other to migrate in the saturated loess when they coexist. The ahead-of-breakthrough time is different when aniline is shorter than TNT, and the breakthrough curve of aniline shrinks its opening more obviously than that of TNT, further showing that the influence of aniline on TNT adsorption is more remarkable than that of TNT on aniline adsorption. These results are the same as those obtained from the batch tests, which give a good explanation for our hypothesis.

To further prove the significant influence of aniline, initial concentrations with 25.0 mg/L aniline and 50.0 mg/L TNT were conducted, and their breakthrough curves are shown in Figure 3b. When aniline and TNT exist alone, their breakthrough points at the start and end are about 2.5 and 5.0, 5.5 and 9.5 d, respectively. Nevertheless, when aniline and TNT coexist, the breakthrough points are changed to 2.0 d and 2.0 d, 4.0 d and 5.0 d, and the time appears to be ahead, about 0.5 and 3.0, 1.5 and 4.5 d compared with that of their existence alone. Consequently, it is clear that the ahead-of-time for aniline is shorter than that for TNT; meanwhile, the breakthrough curves shrink their openings to 2.0 and 3.0 d for aniline and TNT when they coexist, and the shrink extent is smaller for aniline than for TNT. All this proves that the influences of aniline on TNT adsorption are stronger than those of TNT on aniline, and this is supported by the results of the batch tests study.

From the above, the organics of aniline and TNT have different polarities, which determine their diverse mechanisms and performances of adsorption on the loess. And moreover, the organics showed different inhibitions to each other when they coexisted. These inhibition effects are conducive to TNT and aniline preferring the water solution rather than being fixed on loess to enhance their migration in the saturated loess–water system. Therefore, there are interactions for multiple pollutants, and their transport behaviors will inevitably be affected in the combined pollution soil environment [1,2,32]. In our present study, aniline and TNT compete to be adsorbed on the loess and then their transport in the saturated loess is significantly enhanced. Similarly, Wang et al. (2009) [13] also reported that the amounts of 2,4-dinitrophenol or benzoic acid as polar organics adsorbed on the loess were reduced when they existed simultaneously, due to their competitive adsorption, but their transport characteristics were not discussed. Contrarily, the results produced by Xie et al. (2017) [37] and Wang et al. (2021) [18] revealed that Pb²⁺ adsorption could create more binding sites for NH₄⁺ adsorption, and NH₄⁺ adsorption increased the surface area of loess when Pb²⁺ and NH₄⁺ coexisted. Eventually, their adsorption became stronger and mobility was inhibited due to the existence of each other [42]. This suggests that their interaction will lead to the environmental behaviors of pollutants varying in soil–water systems. It is well known that there might be TNT, RDX, HMX and their derivatives (i.e., aniline, nitrobenzene and Pb²⁺) in firing/training ranges, and their adsorption performances depend on the chemical structures and properties of the adsorbents and coexisting substances. In this regard, the present results indicate that the transport of TNT and aniline could be underestimated if the interaction is not considered, and the coexistence of pollutants can increase the risk of deep soil and even groundwater pollution [43].

4. Conclusions

The study reveals that the characteristics of TNT and aniline adsorptions when they coexist are significantly different from those when they exist alone, and there are noticeable interactions and inhibitions in the process of TNT and aniline adsorptions on the loess. As a result, their transports were enhanced each other due to their coexistence under the studied conditions, which increased the risk of soil and groundwater pollution. Therefore, under complex pollution conditions, the interaction between multiple pollutants must be considered when evaluating their adsorptions and transports, and more studies need to be conducted to further reveal their interaction behaviors among multiple pollutants in the combined pollution environment.