The Impact of Temperature on the Removal of Inorganic Contaminants Typical of Urban Stormwater

Appropriate management of urban stormwater requires consideration of both water quantity, resulting from flood control requirements, and water quality, being a consequence of contaminant distribution via runoff water. This article focuses on the impact of temperature on the efficiency of stormwater treatment processes in permeable infiltration systems. Studies of the removal capacity of activated carbon, diatomite, halloysite, limestone sand and zeolite for select heavy metals (Cu and Zn) and biogenes (NH4-N and PO4-P) were performed in batch conditions at 3, 6, 10, 15, 22, 30 and 40 ◦C at low initial concentrations, and maximum sorption capacities determined at 3, 10, 22 and 40 ◦C. A decrease in temperature to 3 ◦C reduced the maximum sorption capacities (Qmax) of the applied materials in the range of 10% for diatomite uptake of PO4-P, to 46% for halloysite uptake of Cu. Only the value of Qmax for halloysite, limestone sand and diatomite for NH4-N uptake decreased slightly with temperature increase. A positive correlation was also observed for the equilibrium sorption (Qe) of Cu and Zn for analyses performed at low concentrations (with the exception of Zn sorption on limestone sand). In turn, for biogenes a rising trend was observed only in the range of 3 ◦C to 22 ◦C, whereas further temperature increase caused a decrease of Qe. Temperature had the largest influence on the removal of copper and the smallest on the removal of phosphates. It was also observed that the impact of temperature on the process of phosphate removal on all materials and ammonium ions on all materials, with the exception of zeolite, was negligible.


Introduction
Controlling urban stormwater, particularly in areas with impermeable covers such as road pavements, parking lots, airports and roofs, is an important issue if local flooding or larger floods after rainfall is to be prevented. Many methods have been proposed to mitigate this issue [1][2][3][4]. The amount of runoff from urbanized areas may be reduced by increasing the retention of runoff water close to its source by centralized and non-centralized sustainable urban drainage systems [2,3]. One of the solutions recommended as best management practice is the use of systems that increase stormwater infiltration. The advantages of this solution are the attenuation of peak runoff, reduction of direct removal of runoff water to the reservoir, and groundwater supply [1,5]. In the last decade, aside from infiltration systems (filter strips, soakaways, infiltration trenches, infiltration basins) and significantly influence the daily, monthly and annual variability of the removal efficiency of transported contaminants. Most reports are focused on the analysis of the intensity of contaminant removal at ambient temperatures [30][31][32][33] or on assessing the impact of temperature increases above room temperature on the intensification of the process [34][35][36][37]; fewer studies indicate how temperature decrease, even to 0 °C, influences the ability of reactive materials to remove contaminants [38,39].
The main aim of the studies here was to determine the qualitative and quantitative impact of temperature on the sorption of selected heavy metals (Cu and Zn) and dissolved forms of biogenic substances (NH4-N and PO4-P) on five different reactive materials. Contaminant removal was determined both at low concentrations and at concentrations sufficient to attain a maximum loading capacity, significantly exceeding those that are characteristic of stormwater.

Materials and Methods
Activated carbon and unamended mineral materials, e.g., zeolite, limestone sand, diatomite and halloysite, were used in the analyses (Figure 1). The mineral materials were not subject to pretreatment to ensure use of the most economically favorable material, as these relatively cheap materials are widely used for water treatment processes [40][41][42][43], and also in the phytostabilisation of soils contaminated by heavy metals [44,45]. The only exception is activated carbon, which was selected as a reference material due to its wide application in the treatment of contaminants in the soil-water environment using permeable reactive barrier technology [46][47][48][49]. Surface morphology and oxide composition of the applied materials were determined using scanning electron microscopy (FEG Quanta 250, USA) and X-ray spectroscopy with energy dispersion X-ray microscopy -EDS by EDAX (USA). The specific surface area (SBET) of the reactive materials was calculated using the Brunauer-Emmett-Teller (BET) method based on the nitrogen adsorption-desorption isotherm at 77 K using ASAP 2020M by Micromeritics (Norcross, GA, USA).
Electrical conductivity (EC) and pH of the reactive materials were measured in an aqueous suspension of 1:2.5 (w/v) and 1:5 (w/v) using a conductivity and pH multimeter pH/LF 12 (Schott, Munich, Germany). The cation exchangeable capacity (CEC) of Na + , K + , Mg 2+ , and Ca 2+ was determined at pH 7 with ammonium acetate [50] and their concentrations measured by an Atomic Absorption Spectrophotometer iCE-3000 (Thermo Scientific, Waltham, MA, USA).
Analysis of the effect of temperature on the intensity of contaminant removal was performed in two series: (1) for low initial concentrations of contaminants (Zn -7.5 mg/L, Cu -30 mg/L, PO4-P -10 mg/L, NH4-N -10 mg/L), insufficient for the exhaustion of the maximum adsorption capacity of the materials, at 3 °C, 6 °C, 10 °C, 15 °C, 22 °C, 30 °C and 40 °C; and (2) for high initial concentrations of the contaminants, in the range from 0.5 to 100 mg/L for heavy metals, phosphates from 0.25 to 50 mg/L, and ammonium ions from 0.5 to 100 mg/L at 3 °C, 10 °C, 22 °C and 40 °C. The concentration of heavy metals and biogenic components in stormwater is varied and can range from 0.007 to 0.963 mg Surface morphology and oxide composition of the applied materials were determined using scanning electron microscopy (FEG Quanta 250, USA) and X-ray spectroscopy with energy dispersion X-ray microscopy-EDS by EDAX (USA). The specific surface area (S BET ) of the reactive materials was calculated using the Brunauer-Emmett-Teller (BET) method based on the nitrogen adsorption-desorption isotherm at 77 K using ASAP 2020M by Micromeritics (Norcross, GA, USA).
Electrical conductivity (EC) and pH of the reactive materials were measured in an aqueous suspension of 1:2.5 (w/v) and 1:5 (w/v) using a conductivity and pH multimeter pH/LF 12 (Schott, Munich, Germany). The cation exchangeable capacity (CEC) of Na + , K + , Mg 2+ , and Ca 2+ was determined at pH 7 with ammonium acetate [50] and their concentrations measured by an Atomic Absorption Spectrophotometer iCE-3000 (Thermo Scientific, Waltham, MA, USA).
Analysis of the effect of temperature on the intensity of contaminant removal was performed in two series: (1) for low initial concentrations of contaminants (Zn-7.5 mg/L, Cu-30 mg/L, PO 4 -P-10 mg/L, NH 4 -N-10 mg/L), insufficient for the exhaustion of the maximum adsorption capacity of the materials, at 3 • C, 6 • C, 10 • C, 15 • C, 22 • C, 30 • C and 40 • C; and (2) for high initial concentrations of the contaminants, in the range from 0.5 to 100 mg/L for heavy metals, phosphates from 0.25 to 50 mg/L, and ammonium ions from 0.5 to 100 mg/L at 3 • C, 10 • C, 22 • C and 40 • C. The concentration of heavy metals and biogenic components in stormwater is varied and can range from 0.007 to 0.963 mg Cu/L, 0.015 to 5.070 mg Zn/L, 0.73 to 17.00 mg NH 4 -N/L and from 0.02 to 14.30 mgP/L (from 0.10 to 0.40 mg PO 4 -P/L) [8][9][10]15,17]. The concentrations of tested components, assumed here, exceeded the concentrations observed in-situ, but they ensured the unification of the test scheme and, in most cases, the exhaustion of materials capacity.
In both series, 0.500 ± 0.001 g of reactive material was shaken with 50 ± 0.5 mL of the solution for 24 h on a DOS orbital shaker (Elmi, Bergstrasse, Germany) at 160 rpm within a thermostatic cabinet ST 1 (Pol-Eko, WodzisławŚląski, Poland) to ensure a stable temperature (±0.1 • C) for the duration of the experiment. All analyses were performed in triplicate, for which arithmetic means and standard deviations were calculated and presented on charts as error bars.
Chemical reagents (CuCl 2 ·2H 2 O, ZnCl 2 , NH 4 Cl, K 2 HPO 4 ) and demineralized water were used in all analyses. Following the experiment, samples were collected and filtered through a 0.45 µm filter for quantitative analyses (Whatman, Maidstone, UK). Samples containing heavy metals were acidified by adding 0.5 mL concentrated nitric acid and stored at 3 • C for later analysis. The concentrations of dissolved Cu and Zn ions were determined using atomic adsorption spectroscopy iCE-3000 (Thermo Scientific, Waltham, MA, USA) in an air-acetylene flame using wavelengths of 324.8 nm and 213.9 nm respectively. Solutions containing ammonium and phosphate ions were immediately analysed using a DR-6000 UV-vis spectrophotometer (Hach Lange, Loveland, CO, USA). Ammonium ions were determined using the silver nitrate titration with a chromate indicator (Mohr's method) [51], whereas phosphates were determined using the application of ammonium molybdate [52].
The results are presented in form of relationship between the equilibrium sorption capacity Q e (mg/g) (Q e = (C 0 − C e ) V/m) and temperature (series I) and in form of the relationship between the equilibrium concentration in the solution (C e ) and the equilibrium sorption capacity Q e (mg/g) (series II). In the calculations, V (L) is the solution volume, m (g) is the sample mass, and C 0 (mg/L) is the initial concentration of the contaminant. The maximum adsorption capacity of reactive materials Q max and the Langmuir equilibrium constant K L (L/g) were determined by matching the Hanes-Woolf's linear form of Langmuir isotherm [53] to the obtained results with application of Solver in Microsoft Excel: Additionally, the results were used to determine parameters describing the thermodynamics of the processes. Enthalpy change ∆H 0 (J/mol) and entropy change ∆S 0 (J/mol·K) were calculated in accordance with van't Hoff equation: the dimensionless equilibrium constant K C was based on parameters of the Langmuir isotherm [53]: and the Gibbs energy change ∆G 0 (J/mol) was calculated from the relationship: where R (J/(mol·K) is the universal gas constant (equal to 8.3144), T (K) the absolute temperature, M w (g/mol) is the molecular weight of the adsorbate, and factor 55.5 is the number of moles of pure water per liter. Pearson's correlation was used to identify the statistical significance of the relationship between the parameters of contaminant sorption capacity (Q e ) and temperature. Statistical significance was at the level of p < 0.05. Moreover, equations describing the relationship between the equilibrium sorption capacity and temperature were estimated by using the multinomial regression model and statistical significance was tested using Statistica (StatSoft, Tulsa, OK, USA).

Characteristics of the Materials
The surface characteristics determined on the basis of nitrogen adsorption/desorption isotherms are presented in Table 1. Activated carbon is a microporous material, as evidenced by the micropore surface and volume values (respectively 366.41 m 2 /g and 0.162 cm 3 /g) and relatively low value of mesopore surface (177.10 m 2 /g) in relation to the total specific surface of the material (782.67 m 2 /g). Diatomite, zeolite and halloysite are mesoporous materials and therefore are characterized by average specific surface areas with mesopores in the shape of narrow fissures with irregular dimensions and shapes. The largest specific surface and volume of mesopores among these minerals were observed in halloysite (respectively 48.33 m 2 /g and 0.193 cm 3 /g), and the smallest in diatomite (respectively 29.32 m 2 /g and 0.05 cm 3 /g). Limestone sand can be classified as a non-porous material. Scanning Electron Microscope (SEM) images and chemical micro-range analyses (EDS) of reactive materials are presented in Figure 2. The observations are in accordance with the results presented in Table 1. It is worth noting that the activated carbon sample has the most expanded porous surface, with spherical pores attaining dimensions from over ten to several hundreds of micrometers. SEM analysis of limestone sand confirmed the carbonate character of this material, whereas micro area chemical analysis of calcium and oxygen pointed to calcite (CaCO 3 ). SEM-EDS analyses indicated the presence of minerals from the silica group in the diatomite sample and allowed for the petrographic classification of this material as diatomaceous earth. SEM micro images confirmed the moderately developed surface of zeolite and halloysite, and in the case of the latter material, the presence of characteristic nano tubes, which, however, were not very distinct at this magnification. Additionally, EDS spectra pointed to the presence of: silicon (Si), oxygen (O), aluminum (Al), potassium (K), calcium (Ca) and iron (Fe) in the zeolite sample, and of: iron (Fe), aluminum (Al), silicon (Si), magnesium (Mg), phosphorus (P), calcium (Ca) and titanium (Ti) in the halloysite sample.
The pH, electrical conductivity (EC) and cation exchange capacity (CEC) of the applied reactive materials are shown in Table 2. Studies of EC have allowed to arrange the analyzed materials as follows in increasing order: limestone sand ≈ diatomite < zeolite < halloysite << activated carbon. The analyzed materials can be subdivided with regard to pH values into 3 groups: (1) reaction below neutral (pH < 7)-diatomite and halloysite, (2) reaction between 7 and 8-zeolite, and (3) reaction > 8-limestone sand and activated carbon. The pH of the materials plays a large role in contamination removal, particularly of heavy metals, but also of ammonium ions [57]. In the case of water pH exceeding 7, beside ammonium ions (NH 4 -N), water contains gaseous ammonia (NH 3 ), and at pH = 11.1 free ammonia amounts to 99%. In solutions with larger values of pH heavy metals precipitate as hydroxides. In turn, at pH < 7 for Pb, pH < 7.5 for Cu, pH < 8.7 for Zn and pH < 9.8 for Ni, the dominant form of metals in solution is their ionic form, which may be removed via adsorption processes. Therefore, it may be assumed that in the case of water containing Zn and Cu with activated carbon and limestone sand, precipitation will play a significant role in the removal of these contaminants due to their high pH values.  Cation exchange capacity (CEC) indicates the number of negatively charged sites on the material surface able to exchange cations (Na + , K + , Ca 2+ and Mg 2+ ). Zeolite was found to have the largest CEC, followed by diatomite, halloysite, activated carbon and limestone sand ( Table 2). Khorshidi and Lu [58] demonstrate that this characteristic not only presents an evaluation of the ion exchange processes responsible for contaminant removal, but also indirectly properties such as water adsorption, ability of water transfer and contaminant migration.

Effect of Temperature at Low Initial Concentrations
The influence of temperature on the removal of heavy metals at initial concentrations of 30 mg Cu/L and 7.5 mg Zn/L in addition to empirically fitted relationships linking these two parameters, are presented in Figure 3, whilst the Pearson's correlation coefficients are presented in Table 3. The  Cation exchange capacity (CEC) indicates the number of negatively charged sites on the material surface able to exchange cations (Na + , K + , Ca 2+ and Mg 2+ ). Zeolite was found to have the largest CEC, followed by diatomite, halloysite, activated carbon and limestone sand ( Table 2). Khorshidi and Lu [58] demonstrate that this characteristic not only presents an evaluation of the ion exchange processes responsible for contaminant removal, but also indirectly properties such as water adsorption, ability of water transfer and contaminant migration.

Effect of Temperature at Low Initial Concentrations
The influence of temperature on the removal of heavy metals at initial concentrations of 30 mg Cu/L and 7.5 mg Zn/L in addition to empirically fitted relationships linking these two parameters, are presented in Figure 3, whilst the Pearson's correlation coefficients are presented in Table 3. The relationships have been presented as polynomials only in the case of statistical significance of all function terms, otherwise a linear relationship is assumed. The intensity of removal of copper and zinc ions changed with temperature and was dependent on the reactive material and metal; however, correlation analysis showed a positive relationship between these variables. Only in the case of Zn removal using limestone sand were the values of Pearson's correlation coefficient (R = 0.224) and statistical significance (p = 0.3288) such that the relationship was statistically not significant.
Appl. Sci. 2019, 9, x FOR PEER REVIEW 7 of 22 relationships have been presented as polynomials only in the case of statistical significance of all function terms, otherwise a linear relationship is assumed. The intensity of removal of copper and zinc ions changed with temperature and was dependent on the reactive material and metal; however, correlation analysis showed a positive relationship between these variables. Only in the case of Zn removal using limestone sand were the values of Pearson's correlation coefficient (R = 0.224) and statistical significance (p = 0.3288) such that the relationship was statistically not significant.     relationships have been presented as polynomials only in the case of statistical significance of all function terms, otherwise a linear relationship is assumed. The intensity of removal of copper and zinc ions changed with temperature and was dependent on the reactive material and metal; however, correlation analysis showed a positive relationship between these variables. Only in the case of Zn removal using limestone sand were the values of Pearson's correlation coefficient (R = 0.224) and statistical significance (p = 0.3288) such that the relationship was statistically not significant.  Removal of Cu was least influenced by the changes of temperature for activated carbon (∆Q e = 0.075 mg/g) and most influenced for limestone sand (∆Q e = 0.933 mg/g). The reactive materials used in the investigations can be arranged with regard to the change of the equilibrium sorption capacity (∆Q e for Cu ions) as follows: activated carbon < diatomite (∆Q e = 0.230 mg/g) < zeolite (∆Q e = 0.341 mg/g) < halloysite (∆Q e = 0.631 mg/g) < limestone sand. An analogous series obtained for zinc ions is as follows: zeolite < halloysite < activated carbon < diatomite, and for this contaminant, the changes of the equilibrium sorption capacity did not exceed 0.051 mg/g. The series does not include limestone sand, because a continuous increase of Q e with temperature increase was not noted.
The probable reason for the shape of the curves describing the obtained results ( Figure 3) was the restricted mass of Cu and Zn (below the maximum sorption capacity), which was insufficient to fill the available negatively charges sites. At initial Zn concentrations equal to 7.5 mg/L (lower than for Cu), temperature change in the range of 3 • C to 40 • C had no practical significance, whereas at initial Cu concentrations equal to 30 mg/L, decrease of temperature to 3 • C resulted in restricting the degree of ion removal particularly on limestone sand and halloysite.

Equilibrium Tests
Equilibrium tests demonstrated that for both metals temperature influences contaminant removal Figures 4 and 5. However, the shape of curves showing the dependence for Zn and limestone sand at 22 • C and 40 • C ( Figure 5), potentially pointing to a precipitation process, did not allow for the determination of the maximum adsorption capacity of this material with regard to Zn ions. The concentration of heavy metals may be regulated by processes depending both on the metal type, but also geochemical conditions influencing metal speciation. Dissolution of Ca 2+ and release of HCO 3 linked with temperature increase may lead to precipitation of Zn in form of ZnCO 3 . As shown by previous studies, the formation of smithsonite is the main mechanism of Zn 2+ ion removal by limestone materials [41]. Additionally, the intensity of heavy metal removal on limestone materials may be linked with chemisorption of metal ions resulting from the rough surfaces of the material, precipitation of metals in the form of metal hydroxides and carbonate, and ion exchange with calcium [59]. In the studies performed herein, precipitation of heavy metals could take place at low initial concentrations, because the pH of equilibrium solutions was below 8.74 for Zn and 8.04 for Cu, whereas in a higher range of initial concentrations, the pH of equilibrium solutions was below 6.88 for Zn and 5.91 for Cu. In the case of the remaining materials, precipitation of Zn and Cu was less significant, because pH was within 5.26 to 7.55 for activated carbon, 5.07 to 7.52 for halloysite, 5.08 to 6.56 for zeolite, and 5.05 to 6.95 for diatomite. Removal of heavy metals on zeolite is generally via ion exchange [36], but also takes place through electrostatic adsorption to negatively charged sites on the zeolite surface [31]. Moreover, heavy metals are removed on halloysite and diatomite via ion exchange and surface complexation respectively on permanently charged sites and variably charged sites [40,60].
An analysis of heavy metal (Cu and Zn) removal equilibrium studies showed a greater influence of temperature on the removal of copper ions compared to zinc ions. Comparison of maximum adsorption capacities of the analyzed reactive materials Q max for Cu and Zn and the constants of adsorption equilibrium K L determined by describing the obtained results with the Langmuir isotherm equation (Equation (1)) is presented in Figure 6 and Table 4, respectively. Removal of heavy metals (Cu and Zn) on all reactive materials was well described by the Langmuir isotherm with the determination coefficient of R 2 ≥ 0.932 (Table 4). The only exception was removal of Cu on activated carbon and on zeolite at 40 • C, described by the Langmuir isotherm with determination coefficient at 0.415 and 0.799, respectively. The material with the best ability to remove Cu ions was activated carbon, followed by limestone sand, halloysite and diatomite, whereas Zn ions were best removed by limestone sand, zeolite, halloysite, diatomite and activated carbon. It cannot be stated unequivocally that the copper ions were removed with larger intensity than zinc ions for all materials. Such a relationship was observed only for activated carbon. Halloysite and limestone sand were more useful for zinc removal. However, it should be emphasized that the maximum adsorption capacity of limestone sand with regard to zinc at 22 • C and 40 • C was not determined in the range of the analyzed concentrations A decrease of temperature from 22 • C (standard temperature for batch analyses) to 3 • C, reduced the maximum adsorption capacity of Cu by approximately 41% for halloysite (by 2.585 mg/g). The next reactive material in the series was zeolite (by 2.407 mg/g), next-limestone sand (by 2.062 mg/g), activated carbon (by 1.462 mg/g) and diatomite (by 1.431 mg/g). For Zn ions the efficiency of the reactive materials attained the following order: zeolite (decrease by 2.148 mg/g), halloysite (by 1.669 mg/g), activated carbon (by 1.198 mg/g) and diatomite (by 0.749 mg/g). During their studies on the application of sodium zeolite in groundwater treatment in the Antarctic using the method of permeable reactive barriers, Woinarski (Parkville, VIC, Australia) in [38] observed a decrease of Q max by about 32% in consequence of temperature fall from 22 • C to 2 • C. In the analysis presented herein, for raw zeolite this value was 31% at temperature change from 22 • C to 3 • C and 44% at change from 40 • C to 3 • C.        Based on parameter R L ( Figure 6) calculated from the Langmuir equilibrium constant K L (R L = 1/(1 + K L ·C 0 )), the processes of heavy metal removal on the analyzed materials may be characterized as favorable, because the R L value was within 0 to 1. Moreover, at higher concentrations of metals in the solution, R L attained values close zero (R L < 0.05), therefore sorption was close to irreversible. It was also observed that with temperature rise, the R L value decreased for all materials except Cu removal on activated carbon and Zn removal on limestone sand. In the first case, the most probable cause is the excess capacity of activated carbon at 40 • C, whereas in the second, it is the large contribution of precipitation processes in the removal of zinc on limestone sand. Based on parameter RL ( Figure 6) calculated from the Langmuir equilibrium constant KL (RL = 1/(1 + KL·C0)), the processes of heavy metal removal on the analyzed materials may be characterized as favorable, because the RL value was within 0 to 1. Moreover, at higher concentrations of metals in the solution, RL attained values close zero (RL < 0.05), therefore sorption was close to irreversible. It was also observed that with temperature rise, the RL value decreased for all materials except Cu removal on activated carbon and Zn removal on limestone sand. In the first case, the most probable cause is the excess capacity of activated carbon at 40 °C, whereas in the second, it is the large contribution of precipitation processes in the removal of zinc on limestone sand.   Statistical analysis of the relationships between maximum adsorption capacities with regard to Cu determined at temperatures from 3 • C to 40 • C and the measured properties of reactive materials showed the lack of statistical significance for the specific surface area, average pore size and volume, pH of the materials and content of the fraction below 0.5 mm on the value of Q max . A positive correlation was observed only with the content of Na 2 O in the composition of the materials (Pearson's coefficient R larger than 0.936), which was 1.23% for activated carbon, 0.96% for zeolite and 0.37% for limestone sand, whereas Na 2 O was not observed among the oxides in diatomite and halloysite. Additionally, a distinct positive correlation was observed between the difference of Q max determined for the most extreme temperatures (∆Q max ) and the content of K 2 O (R = 0.879), and a negative correlation with the content of Fe 2 O 3 (R = −0.926). The studied zeolite is characterized by the presence of K 2 O at the level of 3.51% and Fe 2 O 3 at the level of 1.58, diatomite at the level of 0% and 37.04% respectively, limestone sand at the level of 1.18% and 4.30% respectively, and halloysite at the level of 0.42% and 3.95% respectively. In equilibrium tests for zinc, there was a positive correlation between Q max and the average pore size (R > 0.886), and also between ∆Q max and the CaO content (R = 0.965). The percentage contribution of CaO was 86.93% for limestone sand, 3.17% for zeolite, 1.37% for activated carbon, and 0.67% for halloysite, whereas CaO was not noted in the diatomite sample. It should be emphasized that in this case correlation with the reaction of equilibrium solutions and the pH of the reactive materials was also not observed. The possible explanation was that in batch studies, the ratio between the solution volume to the mass of the reactive material was large (0.500 g material per 50 mL solution), which influenced smaller changes of the reactions of equilibrium solutions (pH e in the range from 5.05 to 8.04 for Cu and from 5.87 to 8.74 for Zn) in relation to the initial solutions (pH 0 in the range from 5.83 for Cu and 6.51 for Zn to 7.00 for distilled water).

Effect of Temperature at Low Initial Concentrations
Change of the equilibrium sorption capacity of reactive materials with regard to ammonium ions and phosphates in the temperature range from 3 • C to 30 • C with empirical fits are presented in Figure 7. Results of tests conducted at 40 • C were not presented for biogenic components because the differences in the measured equilibrium concentrations for the three replicates exceeded 10%. At initial concentrations of contaminants at 10 mg/L NH 4 -N and PO 4 -P, no significant influence of temperature on the intensity of biogenic components removal was observed.
The sorption capacity of the analyzed materials in relation to ammonium ions increased with temperature change from 3 • C to 22 • C. However, the difference between total amounts of ammonium ions removed by the materials at 3 • C and 22 • C was not large, and varied from 0.038 mg/g for halloysite to 0.174 mg/g for diatomite in the studied temperature range. Further temperature increase caused higher intensity of contaminant removal on limestone sand, whereas a decrease of NH 4 -N removal was observed in the case of the remaining materials. However, taking into account practical significance, such variability will not have substantial impact on the degree of stormwater treatment. All terms of the polynomial functions turned out to be statistically significant for all materials with the exception of functions describing removal of ammonium ions on limestone sand. For this material, a description of the observations by a polynomial equation did not cause an increase of the determination coefficient value; therefore, Figure 7 shows a linear relationship. Additionally, the efficiencies of diatomite, halloysite, limestone sand and activated carbon to remove NH 4 -N ions were similar; only zeolite was characterized by a higher sorption capacity in comparison to the remaining materials.
In the case of phosphate ion removal on all analyzed materials, there was a statistically significant (at p < 0.05) positive correlation between temperature and the dependent variable, i.e., sorption capacity (Table 3). However, similar to the case of ammonium ions, the changes of the equilibrium sorption capacity were insignificant, in the range from 0.019 mg/g for limestone sand to 0.299 mg/g for halloysite. It should be emphasized that for these materials temperature increase above room levels was linked with a decrease of Q e . These relationships were described with a polynomial function and the analysis of polynomial regression showed statistical significance of all three terms of the equation for all materials. The material that best removed phosphates from water solutions was activated carbon, followed by halloysite and diatomite, whereas phosphates were least removed by limestone sand and zeolite.
Appl. Sci. 2019, 9, x FOR PEER REVIEW 13 of 22 function and the analysis of polynomial regression showed statistical significance of all three terms of the equation for all materials. The material that best removed phosphates from water solutions was activated carbon, followed by halloysite and diatomite, whereas phosphates were least removed by limestone sand and zeolite. However, negative correlations between the sorption capacity of the applied materials and temperature increase have been documented in the literature, e.g., for some natural zeolites [61,62] and halloysite [63]. A relationship similar to that presented herein was observed by Jing et al. [64] for halloysite from a Chinese reservoir; in those studies, the sorption capacity increased with rising temperature. In turn, Zheng et al. [34] observed that the removal efficiency of the ammonium ion increased with rising temperature at temperatures below 25 °C, and further temperature increase caused decrease in the efficiency of contamination removal. These authors have explained this phenomenon by the low rate of chemical adsorption, which increased with temperature rise to room temperature. However, in the case of exothermic processes, further temperature increase causes larger desorption of contaminants. A similar trend of temperature impact on the removal processes was also observed for biogenic components at initial concentrations of 10 mg/L, but the thermodynamic parameters of the process in the temperature range of 3 °C to 22 °C (ΔH 0 > 0) pointed to the endothermic character of the process (Table 4). However, negative correlations between the sorption capacity of the applied materials and temperature increase have been documented in the literature, e.g., for some natural zeolites [61,62] and halloysite [63]. A relationship similar to that presented herein was observed by Jing et al. [64] for halloysite from a Chinese reservoir; in those studies, the sorption capacity increased with rising temperature. In turn, Zheng et al. [34] observed that the removal efficiency of the ammonium ion increased with rising temperature at temperatures below 25 • C, and further temperature increase caused decrease in the efficiency of contamination removal. These authors have explained this phenomenon by the low rate of chemical adsorption, which increased with temperature rise to room temperature. However, in the case of exothermic processes, further temperature increase causes larger desorption of contaminants. A similar trend of temperature impact on the removal processes was also observed for biogenic components at initial concentrations of 10 mg/L, but the thermodynamic parameters of the process in the temperature range of 3 • C to 22 • C (∆H 0 > 0) pointed to the endothermic character of the process (Table 4).

Equilibrium Tests
The dependence of the ammonium ion concentration in the solution and the sorption capacity in equilibrium conditions is presented in Figure 8. The assumed initial concentration range was too small to exhaust the maximum sorption capacity of the zeolite. Therefore, the parameters of the Langmuir equation were determined at R 2 < 0.833 and fitted Langmuir isotherms where not presented. Ammonium ions represent only part of the nitrogen compounds present in stormwater and their content increases in agricultural areas [8,9,12]. According to Toor et al. [65], the contribution of nitrates in total nitrogen in stormwater is about 58%, whereas all other forms, including ammonium ions, contribute to 42%. The maximum permissible concentration of ammonium ions according to the World Health Organization (WHO, Geneva, Switzerland) is 10 mg/L. Emission control of this compound is particularly crucial with regard to the oxidation of ammonium ions, resulting in the production of nitrites and nitrates harmful for human health and at the same time extremely difficult to remove in in-situ conditions [7,20]. The usefulness of zeolites in the removal of ammonium ions during ion exchange processes has been widely documented [65,66]. An important property influencing the intensity of these processes is the cation exchangeable capacity (CEC) [66][67][68][69], which was at 118.02 meq/100 g for the zeolite studied. CEC of the remaining material was in the range from 1.12 meq/100 g for limestone sand to 13.88 meq/100 g for diatomite (Table 2), and the intensity of NH 4 -N removal on these materials was lower than for zeolite. Statistical analysis showed a positive correlation of CEC both with the maximum adsorption capacity Q max (R = 0.996), as well as the difference of the Q max values calculated for extreme temperatures ∆Q max (R = 0.971). A similar degree of reduction of the ammonium ion concentration due to contact with halloysite, limestone sand, activated carbon and diatomite was not the result of gaseous ammonia release caused by solution pH increase. The average value of the pH of equilibrium solutions was from 6.8 for diatomite to 7.9 for activated carbon, and in such conditions the percentage content of ammonium was below 1.5%. There was a positive relationship between K 2 O and MnO, and the maximum adsorption capacity and ∆Q max (R respectively larger than 0.938 and 0.984), and the CEC value was positively correlated with K 2 O (R = 0.938) and MnO (R = 0.984) contents.

Equilibrium Tests
The dependence of the ammonium ion concentration in the solution and the sorption capacity in equilibrium conditions is presented in Figure 8. The assumed initial concentration range was too small to exhaust the maximum sorption capacity of the zeolite. Therefore, the parameters of the Langmuir equation were determined at R 2 < 0.833 and fitted Langmuir isotherms where not presented. Ammonium ions represent only part of the nitrogen compounds present in stormwater and their content increases in agricultural areas [8,9,12]. According to Toor et al. [65], the contribution of nitrates in total nitrogen in stormwater is about 58%, whereas all other forms, including ammonium ions, contribute to 42%. The maximum permissible concentration of ammonium ions according to the World Health Organization (WHO, Geneva, Switzerland) is 10 mg/L. Emission control of this compound is particularly crucial with regard to the oxidation of ammonium ions, resulting in the production of nitrites and nitrates harmful for human health and at the same time extremely difficult to remove in in-situ conditions [7,20]. The usefulness of zeolites in the removal of ammonium ions during ion exchange processes has been widely documented [65,66]. An important property influencing the intensity of these processes is the cation exchangeable capacity (CEC) [66][67][68][69], which was at 118.02 meq/100 g for the zeolite studied. CEC of the remaining material was in the range from 1.12 meq/100 g for limestone sand to 13.88 meq/100 g for diatomite (Table 2), and the intensity of NH4-N removal on these materials was lower than for zeolite. Statistical analysis showed a positive correlation of CEC both with the maximum adsorption capacity Qmax (R = 0.996), as well as the difference of the Qmax values calculated for extreme temperatures ΔQmax (R = 0.971). A similar degree of reduction of the ammonium ion concentration due to contact with halloysite, limestone sand, activated carbon and diatomite was not the result of gaseous ammonia release caused by solution pH increase. The average value of the pH of equilibrium solutions was from 6.8 for diatomite to 7.9 for activated carbon, and in such conditions the percentage content of ammonium was below 1.5%. There was a positive relationship between K2O and MnO, and the maximum adsorption capacity and ΔQmax (R respectively larger than 0.938 and 0.984), and the CEC value was positively correlated with K2O (R = 0.938) and MnO (R = 0.984) contents.  The curves showing the relationship of equilibrium concentration and sorption capacity for diatomite attained a specific shape (Figure 8). A significant influence of temperature on the equilibrium concentration of ammonium ions was observed for this material at low initial concentrations of this contaminant, whereas at initial concentrations exceeding 20 mg NH 4 -N/L the influence of temperature was insignificant. Tan and Hameed [70] indicated difficulties in the interpretation of the results resulting from the influence of temperature and pH on the shapes of equilibrium isotherms. As shown by Jing et al. [71], ammonium ions are removed on diatomite due to physical adsorption. They also showed that diatomite is negatively charged in an environment with pH exceeding 3.0, whereas according to Li et al. [72] this takes place at pH above 6. It was observed herein that the pH of equilibrium solutions increased with increase of contamination load, attaining a maximum of 7.53. Therefore, it may be concluded that at lower pH values (corresponding to lower initial concentrations of NH 4 -N), the surface of diatomite was less negatively charged, and temperature increase could influence the slight increase of the removal processes intensity. Moreover, Jing et al. [64] have observed that removal of NH 4 -N on halloysite is not the result of electrostatic interaction. They have evidenced that halloysite had a higher negatively charged surface, but the maximum sorption capacity was relatively small, similarly as in the analyses presented herein.
The intensity of phosphate removal from aqueous solutions and the influence of temperature on these processes were more variable than in the case of ammonium ions (Figures 9 and 10). In samples with a zero value of the phosphate initial concentration, the phosphate concentration in solutions being in contact with zeolite and halloysite attained values above zero after reaching equilibrium.
These observations were concordant with the results of studies on the oxide content in selected reactive materials. The presence of P 2 O 5 was observed only in these two materials: 0.39% in zeolite and 0.54% in halloysite. Average removal of phosphates was calculated at 0.14 mg/g of zeolite and 0.19 mg/g for halloysite, respectively. The curves showing the relationship of equilibrium concentration and sorption capacity for diatomite attained a specific shape (Figure 8). A significant influence of temperature on the equilibrium concentration of ammonium ions was observed for this material at low initial concentrations of this contaminant, whereas at initial concentrations exceeding 20 mg NH4-N/L the influence of temperature was insignificant. Tan and Hameed [70] indicated difficulties in the interpretation of the results resulting from the influence of temperature and pH on the shapes of equilibrium isotherms. As shown by Jing et al. [71], ammonium ions are removed on diatomite due to physical adsorption. They also showed that diatomite is negatively charged in an environment with pH exceeding 3.0, whereas according to Li et al. [72] this takes place at pH above 6. It was observed herein that the pH of equilibrium solutions increased with increase of contamination load, attaining a maximum of 7.53. Therefore, it may be concluded that at lower pH values (corresponding to lower initial concentrations of NH4-N), the surface of diatomite was less negatively charged, and temperature increase could influence the slight increase of the removal processes intensity. Moreover, Jing et al. [64] have observed that removal of NH4-N on halloysite is not the result of electrostatic interaction. They have evidenced that halloysite had a higher negatively charged surface, but the maximum sorption capacity was relatively small, similarly as in the analyses presented herein.
The intensity of phosphate removal from aqueous solutions and the influence of temperature on these processes were more variable than in the case of ammonium ions (Figures 9 and 10). In samples with a zero value of the phosphate initial concentration, the phosphate concentration in solutions being in contact with zeolite and halloysite attained values above zero after reaching equilibrium.
These observations were concordant with the results of studies on the oxide content in selected reactive materials. The presence of P2O5 was observed only in these two materials: 0.39% in zeolite and 0.54% in halloysite. Average removal of phosphates was calculated at 0.14 mg/g of zeolite and 0.19 mg/g for halloysite, respectively. The mechanisms of binding or removal of phosphorus compounds may be very variable and depend on pH, temperature, presence of competing compounds, salinity, occurrence of organic matter or grain-size distribution [73]. For example, high values of environmental pH may stimulate phosphorus adsorption through ligand exchange of HPO4 2− or H2PO4 − ions present in the solution with OHions present on the surfaces of solid particles [74]. Reddy and DeLaune [75] stated that temperature increase causes increase of the energy at which phosphorus sorption is bound in soils and sediments and shortening of the time needed for the formation of crystalline forms. Moreover, The mechanisms of binding or removal of phosphorus compounds may be very variable and depend on pH, temperature, presence of competing compounds, salinity, occurrence of organic matter or grain-size distribution [73]. For example, high values of environmental pH may stimulate phosphorus adsorption through ligand exchange of HPO 4 2− or H 2 PO 4 − ions present in the solution with OHions present on the surfaces of solid particles [74]. Reddy and DeLaune [75] stated that temperature increase causes increase of the energy at which phosphorus sorption is bound in soils and sediments and shortening of the time needed for the formation of crystalline forms. Moreover, as pointed out by Reddy and DeLaune [75], in the presence of limestone rocks, orthophosphates are adsorbed on the surface of the material forming CaHPO 4 or replacing CO 3 2− by HPO 4 2− . However, taking into account the measured maximum sorption capacities of limestone sand (from 0.205 to 0.300 mg/g), it can be concluded that such processes did not take place. This is most probably related with the pH of equilibrium solutions below 7.5, because most mentioned processes take place in an alkaline environment. Moreover, according to Yang et al. [76], the zeta-potential of halloysite is at pH range of 6 to 7, whereas Veerabadran et al. [77] suggested that the external surface of halloysite exhibits a negative zeta-potential over a pH range of 2.5-8.5. The pH of equilibrium solutions for this material is in the range of 6.07 to 6.93 and for zeolites, i.e., in the range of 6.74 to 7.74. In the case of the second material, phosphorus removal may be caused by electrostatic adsorption, because the point of zero charge of this material can range from 8.5 to 9.0 [30]. as pointed out by Reddy and DeLaune [75], in the presence of limestone rocks, orthophosphates are adsorbed on the surface of the material forming CaHPO4 or replacing CO3 2− by HPO4 2− . However, taking into account the measured maximum sorption capacities of limestone sand (from 0.205 to 0.300 mg/g), it can be concluded that such processes did not take place. This is most probably related with the pH of equilibrium solutions below 7.5, because most mentioned processes take place in an alkaline environment. Moreover, according to Yang et al. [76], the zeta-potential of halloysite is at pH range of 6 to 7, whereas Veerabadran et al. [77] suggested that the external surface of halloysite exhibits a negative zeta-potential over a pH range of 2.5-8.5. The pH of equilibrium solutions for this material is in the range of 6.07 to 6.93 and for zeolites, i.e., in the range of 6.74 to 7.74. In the case of the second material, phosphorus removal may be caused by electrostatic adsorption, because the point of zero charge of this material can range from 8.5 to 9.0 [30]. Equilibrium tests for ammonium and phosphate ions were performed only in the temperature range from 3 °C to 22 °C; therefore, the relationship of the maximum adsorption capacity and temperature has not been described by a polynomial function (Figure 10). Similarly, as in analyses of series I, ammonium ions were removed most intensely by zeolite (Qmax in the range from 8.450 to 9.723 mg/g); the maximum sorption capacities of the remaining materials attained similar values from 0.960 (activated carbon) to 1.233 mg/g (diatomite). In the case of phosphates, the material with the highest removal capacity was activated carbon (Qmax in the range from 2.114 to 2.634 mg/g), followed by halloysite (from 1.224 to 1.617 mg/g), diatomite (from 1.329 to 1.477 mg/g), zeolite (from 0.668 to 0.797 mg/g) and limestone sand (from 0.204 to 0.300 mg/g). It should be emphasized that a slight decrease of Qmax with regard to NH4-N was noted for diatomite, limestone sand and halloysite, but these changes were below 0.146 mg/g.
Langmuir parameters for the removal of ammonium ions were determined at R 2 from 0.982 to 0.999 (Table 4). The only exception are results obtained for zeolite (R 2 from 0.771 to 0.833). Lower values of R 2 could be caused by the non-exhaustion of the maximum sorption capacity of the materials in the established range of contaminants. The results of phosphate removal (R 2 from 0.548 to 0.986) were described least precisely, despite the fact that numerous authors have shown the Equilibrium tests for ammonium and phosphate ions were performed only in the temperature range from 3 • C to 22 • C; therefore, the relationship of the maximum adsorption capacity and temperature has not been described by a polynomial function (Figure 10). Similarly, as in analyses of series I, ammonium ions were removed most intensely by zeolite (Q max in the range from 8.450 to 9.723 mg/g); the maximum sorption capacities of the remaining materials attained similar values from 0.960 (activated carbon) to 1.233 mg/g (diatomite). In the case of phosphates, the material with the highest removal capacity was activated carbon (Q max in the range from 2.114 to 2.634 mg/g), followed by halloysite (from 1.224 to 1.617 mg/g), diatomite (from 1.329 to 1.477 mg/g), zeolite (from 0.668 to 0.797 mg/g) and limestone sand (from 0.204 to 0.300 mg/g). It should be emphasized that a slight decrease of Q max with regard to NH 4 -N was noted for diatomite, limestone sand and halloysite, but these changes were below 0.146 mg/g.
Langmuir parameters for the removal of ammonium ions were determined at R 2 from 0.982 to 0.999 (Table 4). The only exception are results obtained for zeolite (R 2 from 0.771 to 0.833). Lower values of R 2 could be caused by the non-exhaustion of the maximum sorption capacity of the materials in the established range of contaminants. The results of phosphate removal (R 2 from 0.548 to 0.986) were described least precisely, despite the fact that numerous authors have shown the usefulness of the Langmuir isotherm for the description of phosphate sorption on minerals and organic materials [78,79]. This could be caused by the low abilities of the applied materials for the removal of such contaminants. According to Reddy and DeLaune [75], the increase of removal capacity with increasing temperature may be related to the increase of the bonding energy of phosphorus sorption and precipitation rates.
Evaluations of parameter R L calculated on the basis of the Langmuir adsorption constant indicated that the sorption of ammonium and phosphate ions in the analyzed range of concentrations and temperatures is favorable (0 < R L < 1), and removal of phosphates at low concentrations of these contaminants in the solution is similar to linear and less favorable. This is confirmed by values of R L close to one. Moreover, the most favorable adsorption process was determined for the highest temperatures for all materials and contaminants, as evidenced by lower R L values. However, these differences were insignificant in the case of NH 4 -N removal on zeolite and phosphates on all materials. Jing et al. [64] also observed an increase of the maximum sorption capacities of halloysite with regard to ammonium ions with temperature increase (from 0.514 mgNH 4 -N/g at 15 • C to 0.737 mgNH 4 -N/g at 40 • C), but in these investigations the values of Langmuir equilibrium constant K L decreased (therefore R L increased) with temperature change from 15 • C to 40 • C as opposed to the analysis in this paper. Table 4 shows thermodynamic parameters calculated at various temperatures. ∆H 0 and ∆S 0 were calculated from the slope and intercept of van't Hoff plots of lnK C against 1/T and the ∆G 0 from Equation (4). The results showed that the process of removal of all heavy metals and biogenic substances on zeolite, halloysite, diatomite, limestone sand and activated carbon was endothermic, as evidenced by the increase of contaminant removal degree with temperature and positive values of ∆H 0 [80]. Further, negative values of ∆G 0 indicated that the removal processes of the analyzed contaminants on 5 reactive materials had a spontaneous nature and was thermodynamically favorable in the given experiment conditions [81]. In turn, positive values of ∆S 0 indicated that contaminant removal was spontaneous at high temperatures, reflected the affinity of reactive materials to the analyzed contaminants, and confirmed the increased randomness at the solid-liquid interface during adsorption [82]. Based on the obtained calculation results, it may be assumed that the processes of Cu removal on diatomite and zeolite, NH 4 -N removal on activated carbon, and PO 4 -P removal on diatomite represented physisorption. Removal of zinc on all five materials, copper on limestone sand and halloysite, ammonium ions on zeolite, limestone sand and halloysite and phosphates on all tested materials except diatomite resulted from electrostatic interaction. The mechanism of chemical sorption played a role only in copper and ammonium ion removal with the application of activated carbon and diatomite. These conclusions were drawn based on an overview by Anastopoulos et al. [83], according to which the value of ∆H 0 < 20 kJ/mol is characteristic for physisorption, ∆H 0 in the range of 20 to 80 kJ/mol is typical for electrostatic interaction, a ∆H 0 in the range of 80 to 450 kJ/mol is characteristic for chemical adsorption.

Potential Application of the Study Results
The studies conducted have demonstrated that ambient temperatures may influence the removal of contaminants transported with stormwater from urban areas through permeable treatment zones. Analyzing the results of conducted tests, the contaminants load CL (milligrams of added contaminants per grams of reactive material) treated to a certain level of contaminants concentration was determined ( Table 5). The permissible concentrations of contaminants transported with sewage to water or to soil were assumed, 0.5 mg Cu/L, 2 mg Zn/L, 10 mg NH 4 -N/L and 5 mg PO 4 -P/L, respectively. Values of the CL may have a practical significance. Although at higher loads the maximum sorption capacity of the materials was not exceeded, the equilibrium concentration exceeded the permissible level. The conclusions drawn from the analysis of CL values are consistent with conclusions based on changes of Q max values. The highest impact of the temperature decrease was observed for Cu removal on activated carbon and Zn removal on limestone sand and the lowest for NH 4 -N removal on diatomite, limestone sand, halloysite and activated carbon. The values of CL and Q max determined at low and high temperatures may be helpful during design stage of the dimensions and maintenance frequency of treatment zone operated in cold region.
Moreover, the studies confirmed that contaminants present in stormwater from urbanized areas cannot be removed with one universal reactive material. Only the use of a mixture of several reactive materials can ensure the removal of a larger spectrum of contaminants. The conclusion is consistent with those presented by other researchers [17,18,20,29,30]. The obtained test results indicate that the highest removal efficiency of Cu and PO 4 -P has activated carbon, of Zn-limestone sand, and of NH 4 -N-zeolite.

Conclusions
This study highlights the impact of temperature on the removal efficiency of selected inorganic contaminants typical for stormwater on activated carbon and low-cost mineral adsorbents (diatomite, halloysite, limestone sand and zeolite). The investigation revealed that ambient temperature determining the conditions in the sub-surface soil layer, and therefore, in the sub-surface permeable treatment zone, may influence the removal of contaminants transported with stormwater from urban areas. The sorption efficiency of the studied materials (activated carbon, diatomite, halloysite, limestone sand and zeolite) became weaker with decreasing temperature, particularly for heavy metals (from 28% to 46%) compared to biogenic substances (from −7% to 32%). However, the basic restriction of practical application of the obtained results is the use of mono-component model solutions instead of real stormwater samples. The presence of other dissolved components, competing for active locations on the surfaces of reactive materials, as well as organic matter and suspensions may influence the actual sorption abilities of the materials. In turn, the achievement of this research was to gain a better understanding of how the changes of temperature may affect the removal of heavy metals (Cu and Zn) and nutrients (NH 4 -N and PO 4 -P) from solutions being in contact with selected reactive materials.
It has been revealed that the intensity of contaminant removal variably changed with temperature depending on the reactive material and the contaminant; however, analysis of the correlation has shown a positive dependence between these two variables. The test results demonstrated that in conditions of low initial concentrations this correlation is true for heavy metals in the tested range of temperatures (from 3 • C to 40 • C) and for biogenes in the range of 3 • C to 22 • C. Further increase of temperature caused the decrease of equilibrium sorption capacity of all materials tested. The only exception was the removal of ammonium ions on limestone sand. The investigations provided a conclusion that the decrease of temperature had the greatest impact on the maximum adsorption capacities for copper, followed by zinc and ammonium (on zeolite). Moreover, the values of the analyzed parameter (Q max ) were slightly influenced by temperature changes for phosphate removal on all analyzed reactive materials, as well as ammonium ion removal on halloysite, limestone sand, activated carbon and diatomite. Furthermore, it is worth noting that temperature decrease more significantly reduced the sorption capacities of reactive materials in high concentration solutions. Unfortunately, analysis of the results of the performed investigations precludes defining a universal concentration range for all reactive materials, for which the impact of the analyzed factor may be avoided.
The performed studies indicate that the material showing the highest efficiency of Cu and PO 4 -P removal is activated carbon, of Zn is limestone sand, and of NH 4 -N is zeolite. The remaining materials were characterized by a medium efficiency of heavy metal removal and a low efficiency of biogene removal from mono-component solutions. Based on the values of maximum sorption capacity, treatment processes were most restricted by temperature fall in the case of Cu removal on activated carbon and Zn removal on limestone sand. The performed calculations indicate that the removal processes were favorable in nature, being close to irreversible for higher concentrations of heavy metals, and less favorable for lower concentrations of phosphates. Moreover, all observed processes can be characterized as endothermic and spontaneous at high temperatures.