Comparison the Adsorption Capacity of Ukrainian Tuff and Basalt with Zeolite–Manganese Removal from Water Solution

Manganese is an undesirable element in tap water but is common in the groundwater. Several methods can be used for manganese removal, including adsorption. Mined rocks are commonly evaluated as adsorbents and it was the objective of this paper – to investigate the Ukrainian volcanic tuff and basaltic rock from the Ivanodolinsky quarry and compare it with Ukrainian zeolite as well as with literature data. The research was based on equilibrated batch tests at a temperature of 10°C and slightly acidic pH. The data were treated using Langmuir and Freundlich models in the linear form. The results indicated the spontaneous and favourable adsorption of manganese. The volcanic tuff was characterized by the highest adsorption capacity, twice higher than basalt and zeolite. The heterogeneity of the active adsorption sites on the tuff was also greater and resulted from the diversity of the mineral composition. Considering the literature data, the properties of tuff are worth further research.


INTRODUCTION
The presence of manganese in the water environment can be caused by the anthropogenic pressure; however, manganese can also be a natural component. Manganese is especially common in groundwater and when water is collected for the supply purposes, manganese is removed for the operational and sanitary reasons [Latterman 1999; Kowal and Świderska-Bróż 2007]. Usually, manganese is removed as a result of homogeneous or heterogeneous oxidation to insoluble oxyhydroxides being mechanically separated on filtering bed [Vries et al. 2017]. Various methods are used for the oxidation: adding chemical oxidants such as potassium permanganate or sodium hypochlorite; aerating and adding alkali; aerating and using beds containing manganese oxides as catalysts, sometimes being of biogenic origin [Tekerlekopoulou et al. 2013].
Another solution for the removal of heavy or transition metals from water is the use of sorption materials. The metals dissociated in water can be adsorbed by various materials, such as activated carbon, resins, mineral sorbents or biosorbents [Sen Gupta and Bhattacharyya 2011;Jiang et al. 2015; Renu et al. 2017]. Activated carbon and resins are highly effective sorbents, but their production costs are also high [Cegłowski and Schroeder 2015]. Low cost materials for heavy and transition metals removal are biosorbents or natural minerals. Mined rocks are characterized by a varied mineral composition, which significantly determines their adsorption properties. For this reason, their suitability is determined in technological research.
The adsorption properties of materials are assessed using various factors, which can be determined under equilibrium conditions. In order to predict adsorption, two-parameter isotherms are used, as well as more-parameter models. The constants of isotherm models give information about the nature of adsorption process and interactions between adsorbent and removing particles. The primary parameter is the adsorption capacity, which is used by many authors to compare materials in terms of the effectiveness of their use in the adsorption process [ The aim was to determine the adsorption properties of the Ukrainian volcanic tuff and basaltic rock in manganese solution and to compare them with the results obtained for Ukrainian zeolite as well as with the literature data. These studies are a part of the research on the use of natural mineral adsorbents to remove metals from water, in this case manganese.

MATERIALS AND METHODS
The investigated rock materials were the volcanic tuff (T) and basalt (B) from the Ivanodolinsky quarry (Rivne region, Ukraine) as well as the zeolite (Z) from the Sokirnitskoe deposit (Zakarpattia region, Ukraine). The tuff mainly consists of saponite (58% w/w) and a smaller amount of quartz (22%), hematite (17%) and analcime (3%), whereas the composition of basaltic rock is dominated by andesine (93 %) with an admixture of saponite (7 %) [Melnychuk et al. 2018]. The main component of zeolite is clinoptilolite (70-80%) [Rogov 2009]. The rocks were used in the natural form without thermal treatment, prepared only by grinding and sieving (< 0.1 mm).
The adsorption equilibrium experiment was carried out in conical flasks by shaking the rock materials in Mn-solutions at 10°C during 24 h. The Mn-solutions were obtained by dissolution of manganese(II) chloride tetrahydrate (pure reagent by Chempur) in distilled water. The solutions were characterized by the Mn concentration in the range 2-17 mg/L and pH 5.96 ± 0.08. Two types of experimental series differed in the ratio of rock mass to the solution volume: 1g/1L and 2g/1L. After shaking, the solutions were separated from the mixtures by centrifugation in the time 7 minutes at 1500 rpm. The Mn concentration in solutions before and after adsorption was measured by means of the AAS technique using a PG Instruments spectrometer. In the solutions, pH was also measured using a gel electrode and Hach HQ40D-Multimeter.
Equilibrium adsorption capacity q e (mg/g) was calculated according to the equation: where: C 0 (mg/L) and C e (mg/L) are the Mn concentrations in solution before adsorption and at equilibrium, respectively, V (L) is the volume of solution and m (g) is the mass of rock sample.
The data obtained from the adsorption experiment were treated using the Langmuir and Freundlich models. Linear regression was used; the applied isotherm equations are presented in the Table 1. The fit of the models to the experimental data was validated using coefficient of determination (R 2 ), chi-square statistic (χ 2 ) and mean absolute percentage error (MAPE). In chi-square test 2-tailed significance level (p) was calculated for 7 degrees of freedom.

RESULTS AND DISCUSSION
The results of fitting isotherm models of Mn adsorption at equilibrium on the tested materials are shown in Table 2 and in Figures 1-3. On the basis of R 2 , χ 2 and MAPE values, it has been determined that the adsorption of Mn onto volcanic tuff and basalt is very well described by both the Among the tested materials, the highest maximum adsorption capacity (q m ) has the tuff contained saponite and hematite, minerals characterized by significant adsorption properties relative to heavy and transition metals and radionuclides ]. It was twice higher than for zeolite rich in clinoptilolite and basalt consisting mainly of andesine. The ratio of adsorbent mass to the solution volume is significant in the case of tuff and zeolite, higher maximum capacities were achieved using ratio 1g/1L, not 2g/1L. This is important when designing adsorption systems.
Comparing the obtained results with the results of other studies on natural mineral adsorbents (Table 3), it can be concluded that volcanic tuff is characterized by high adsorption capacity. This comparison required an expression of the adsorption capacity in mmoles·g -1 due to the study of various elements. The adsorption capacity of basaltic rock and zeolite can be described as an intermediate. It should be noted that in the presented study, the temperature was much lower than in other studies, since the groundwater temperature was simulated. Our previous study indicated that the Mn adsorption rate and removal efficiency do not change at temperature 10-25°C [Reczek et al. 2020]. However, differences in experimental conditions and material preparation may be the reason for obtaining lower adsorption capacity of the Sokirnitskoe zeolite (clinoptilolite) than in paper of Korablew et al. [2017]. The maximal adsorption capacity of manganese dioxide modified siliceous rock used for groundwater treatment is 1.25 mg·g -1 [Michel and Kiedryńska 2012] several times lower than the capacities of volcanic tuff and basaltic rock. In this regard, the properties of volcanic tuff seem to be promising.
Another important factor affecting the adsorption results is pH of solution. In the present study, the MnCl 2 solution was slightly acidic (pH 5.96 ± 0.08) which corresponds well with slightly acidic characteristics of ground water. During Mn adsorption, the pH grew and at equilibrium the pH increased on average by 0.54 ± 0.09 in tuff and   Basalt has the highest affinity for Mn represented by the lowest value of the Langmuir's b parameter. Importantly, increasing the concentration of tuff and zeolite from 1g/1L to 2g/1L decreased its affinity for Mn, represented by higher b values. If it is assumed that good adsorbent is characterized by high q m and low b values of constants in Langmuir model, then it is preferable to use tuff and zeolites in the ratio 1g/1L. On  the basis of the values of the n parameter in the Freundlich model, it can be concluded that Mn adsorption is more spontaneous on the tuff than on the basaltic rock. However, the differences are rather small and n is close to 2, the limit value between moderately difficult and good adsorption properties of material [Treybal 1981].
On the basis of the b parameter values from Langmuir model expressed in L mmole -1 the Gibb's free energy change (ΔG°) was calculated as follows: where: R -universal gas constant (J·mol -1 ·K -1 ), T -absolute temperature (K).

CONCLUSIONS
The volcanic tuff and basalt from the Ivanodolinsky quarry in Ukraine effectively adsorbed manganese from the water solution. Adsorption occurred spontaneously at 10°C and in slightly acidic pH. There are beneficial conditions when the material is planned to be used in groundwater treatment line.
The adsorption of manganese on the volcanic tuff was slightly better described by the Freundlich model (F≥L), whereas the Langmuir model slightly better predicted the adsorption on basaltic rock (L≥F) and definitely on the zeolite (L>>F). This may be related to the greater heterogeneity of the active adsorption sites on the tuff showing a greater variety in mineral composition than basalt and zeolite.
The tuff was characterized by the highest adsorption capacity -twice higher than basalt from the same quarry and zeolite, the reference material. The comparison with other natural mineral adsorbents classifies tuff by high adsorption capacity and basalt by intermediate. The results obtained are promising for further research.