Natural and Chemically Modified Post-Mining Clays—Structural and Surface Properties and Preliminary Tests on Copper Sorption

Abstract

The structural and surface properties of natural and modified Pliocene clays from lignite mining are investigated in the paper. Chemical modifications are made using hydrofluoric acid (HF), sulfuric acid (H₂SO₄), hydrochloric acid (HCl), nitric acid (HNO₃), sodium hydroxide (NaOH), and hydrogen peroxide (H₂O₂), at a concentration of 1 mol/dm³. Scanning electron microscopy is used to detect the morphology of the samples. Nitrogen adsorption isotherms were recorded to determine the specific surface area (SSA), mesoporosity, microporosity, and fractal dimensions. The raw clay has an SSA of 66 m²/g. The most promising changes in the structural properties are caused by modifications with HF or H₂SO₄ (e.g., the SSA increased by about 60%). In addition, the raw and modified clays are used in preliminary tests with Cu(II) sorption, which were performed in batch static method at initial Cu(II) concentrations of 25, 50, 80, 100, 200, 300, and 500 mg/dm³ in 1% aqueous suspensions of the clayey material. The maximum sorption of Cu(II) on the raw material was 15 mg/g. The structural changes after the modifications roughly reflect the capabilities of the adsorbents for Cu(II) adsorption. The modifications with HF and H₂SO₄ bring a similar improvement in Cu(II) adsorption, which is around 20–25% greater than for the raw material. The structural properties of investigated clays and their adsorptive capabilities indicate they could be used as low-cost adsorbents (e.g., for industrial water pretreatment).

1. Introduction

Lignite extraction via surface mining results in the removal of accompanying minerals, which create an overburden of geological deposits above lignite. The overburden consists of loose sedimentary material (gravel and sand), plastic rocks (claystones and mudstone), and more compact rocks (sandstones, lake chalk, and others) [1]. There are no exact data for the overburden that arises each year due to surface mining, in particular due to lignite mining. In Poland, lignite mining requires the removal of 110–130 million m³ per year of overburden [1]. To roughly estimate the overburden removal in the world, the above number can be multiplied by the “lignite extraction in the world” (1342 × 10⁶ t) to “lignite extraction in Poland” (66 × 10⁶ t) ratio, to obtain around 2500 million m³ per year (2013). There are no specific data from individual mines on the amount of clays exploited during the extraction of lignite. It is estimated, however, that the quantities of mined overburden masses are many times higher than the mass of lignite extracted. The clayey minerals covering the lignite beds create deposits with thickness ranging from a few meters to several dozen meters. For example, the average ratio of overburden to bed thickness ranges from 2.2:1 in Turów to 6.6:1 in Adamów [2].

Abundant resources of clays around the world and their low price contribute to their popularity. They are non-toxic, readily accessible, and easy to process and prepare. Depending on the quality parameters, clays from lignite mining are used mainly in the refractory industry in building ceramics, such as bricks, ceramic tiles, sanitary ware, and stoneware products. They are also used for the construction of waterproofing layers in landfills, for rock-sealing, and as insulation materials in water reservoirs [3,4]. Clays, including beidellite ones, have also been applied as natural sorbents in adsorption processes for water and wastewater treatment [5,6,7,8,9,10,11]. They often combine ion exchange and molecular sieve adsorptive properties, and are susceptible to a wide range of modifications to these properties [12]. The presence of active surface centers in clay minerals also makes them good catalysts for various reactions occurring in nature. They are used, for example, as catalysts for the removal of organic pollutants from water and in industrial processes [13,14]. They are active in cracking reactions [15], synthesis [16], polymerization [17], isomerization [18], and oxidation [19].

Sustainable and rational management of mineral resources requires searching for new methods for management of mining wastes. Therefore, although the focus of the research was directed to the structural and surface properties of post-mining waste rocks, the motivation was the possible use of beidellite clays from surface mining as inexpensive adsorbents of metal ions. The use of clays as low-cost adsorbents is in accordance with the directions for mining waste management indicated in directive 2006/21/EC on the management of waste from extractive industries [20]. There is also a noticeable trend of using natural materials in adsorption, including both waste rocks [21] as well as agricultural waste materials [22]. The clays as well as other materials are often modified thermally [23,24] or chemically [25,26,27] to enhance their adsorptive properties. As reported in [28], clay minerals are good adsorbents for metals, such as Cd(II), Cr(III), Cr(VI), Hg(II), Co(II), Cu(II), Zn(II), Pb(II), Ni(II), and Mn(II). Polluting substances containing metal ions are released into the environment with insufficiently treated industrial and municipal sewage, as well as with waters originating from runoff from agricultural areas [29] or metalliferous rocks [30]. Wastewater containing metal ions originates mainly in the processing of non-ferrous metals, in the textile and chemical industries, in electroplating, and in production of alloys, cables, and others [31]. Metal ions show strong toxic properties and pose a threat to living organisms. For example, an excess of copper in water is highly toxic to biological activity, which limits the self-cleaning processes [32]. Therefore, initial tests against Cu(II) sorption on the tested material were also performed. According to [28], the adsorption capacity for Cu(II) sorption on various clay minerals ranges from 2.35 mg/g on palygorskite to 54.07 mg/g on immobilized bentonite.

In the research, the naturally occurring as well as chemically modified beidellite clays from lignite mining are considered. The aim was to determine and enhance the structural and surface properties of Pliocene clays, as well as perform preliminary tests for Cu(II) sorption. Acidic, alkaline, and hydrogen peroxide modifications were tested. Although they produce adsorbents without highly ordered porous structures, they are simple and inexpensive, and can significantly increase the specific surface and porosity, leading to improvement in the adsorption of various contaminants. Acid activation consists of treating minerals with acid solutions. This results in the introduction of H₃O+ ions into exchangeable positions, increased surface acidity and partial violation of the crystal structure, and changes in surface area and porosity. The aim is to obtain a partially dissolved material with increased surface area, porosity, and surface acidity [33]. During acid activation, the number of weakly acidic surface functional groups increases, while the number of groups with a stronger acidic character decreases. The dissolving of Al predominates over Si in acidic treatments. The materials produced are widely available, relatively inexpensive sources of protons, which are effective in many important reactions and processes for the industry [34]. Materials obtained in this way can be used to purify water, as well as wastewater from anthropogenic pollutants. Aluminates treated with acids are used as catalysts (e.g., in oil cracking). The main difference between acidic and alkaline activation of clay minerals is that in the first case, the process leads to degradation and amorphization of the original silicate structure, while in the second case, through the partial degradation stage, skeletal silicates with favorable sorption, ion-exchange, and catalytic properties can be formed. Bases were also used to obtain synthetic zeolites formed from kaolin and bentonite or from fly ash from coal combustion [35]. However, the course of these reactions is strongly dependent on factors such as the type and degree of ordering of the clayey matter, alkali concentration in the solution, time and temperature of activation, mixing intensity of the suspension, and the ratio between solid and liquid phases in the alkaline suspension.

2. Materials and Methods

2.1. Materials

The samples of clay rocks were taken from the Pliocene clayey sandy layers from an opencast lignite mine located in the central part of Poland. The lignite deposits, as well as accompanying sediments (clays, mud, and peat deposits), are associated with the tertiary formations of the Polish lowlands. The samples were acquired from a dump site where the beidellite clays are selectively stored as a waste mineral, which is used by the mine directly or after processing, but mostly for sale to external customers.

2.2. Sample Preparation

The samples were ground below a particle size of 100 μm and air dried. The analyses were done for the unmodified material as well as modified chemically by acidic, alkaline, and hydrogen peroxide activation to oxidize the organic components of the clays. The modifications were performed using the standard procedures described in [36,37,38]. Acidification of the samples was carried out with hydrofluoric, sulfuric, nitric, and hydrochloric acids. Alkaline activation was performed with sodium hydroxide. In the tests, 1 molar solutions of hydrofluoric acid (HF), H₂SO₄, H₂NO₃, HCl, NaOH, and H₂O₂ were used. Here, 30 g of raw clay was mixed with a 1 molar solution of the suitable chemical in a shaking water bath (357 type, Elpin-Plus s.c., Lubawa, Poland) at 70 °C for 6 h, after which the samples were allowed to stand for 14 h in solution. Finally, the precipitate was filtered off and washed several times with deionized water until reaching natural pH. The samples of the clays used in our investigations are marked as follows:

• S1—unmodified clay;
• S2—clay activated with HF;
• S3—clay activated with H₂O₂;
• S4—clay activated with H₂SO₄;
• S5—clay activated with HCl;
• S6—clay activated with HNO₃;
• S7—clay activated with NaOH.

2.3. Characterization of Material

The ash content (non-flammable constituents) was determined at 815 °C according to Polish standard PN-ISO 1171:2002P [39]. The chemical composition was determined via XRF (X-ray fluorescence) on a sequence spectrometer PW 1404 (Philips, Amsterdam, The Netherlands). Scanning electron microscopy (SEM) images were obtained on a Zeiss Supra 55 VP with 7.5 μm aperture and 2 nm resolution. The images were made for raw and modified samples in the form of powder of grain size below 100 μm. Nitrogen adsorption/desorption isotherms (T = 77 K) were measured, employing a commercial volumetric adsorption setup provided by Autosorb iQ (Quantachrome Instruments, relative uncertainty below 0.15%). The volume ($v$) of adsorbed nitrogen was measured as a function of the relative equilibrium pressure ($p/p_{\mathrm{s}}=x$), where $p$ is the equilibrium pressure, and $p_{\mathrm{s}}$ is the saturated vapor pressure over the flat surface of the liquid. The relative pressure range was from around 10⁻⁵ to above 0.99. The nitrogen adsorption/desorption values were expressed in cm³ @STP (at standard temperature and pressure) per gram of adsorbent. Prior to nitrogen adsorption, the clay samples were degassed at 105 °C for several hours at a pressure of 10⁻⁴ mmHg.

The specific surface area was determined using standard BET (Brunauer-Emmett-Teller) technique (Equations (A1)–(A3) in Appendix A) with several linear forms of the BET Equation (A3). However, they all resulted in negative (i.e., non-physical) $C$ values or poor matching with experimental data for some samples. Therefore, the modified BET equation (A4), which takes into account the finite number of adsorbate layers, $n$, was finally used. The value of $n$ determined the best possible fit (i.e., to approach 1 for the determination coefficient ($R^2$)). The search interval was 1 ≤ $n$ ≤ 5.

The pore size distribution was calculated using the standard BJH (Barrett-Joyner-Halenda) method with the Kelvin equation. Calculations were performed using the standard Harkins–Jura statistical thickness of the adsorbate layer, as well as for the experimentally determined statistical thickness, whose values were tabulated in [40]. The details are given in Appendix A (Equations (A5)–(A7) and Figure A1).

The microporosity analysis was performed using the Dubinin–Radushkevich (DR) equation. Relative pressures up to 0.1 were considered. Calculations were also made using the Dubinin–Astakhov equation.

The microporosity was also evaluated via the t-plot method, with the statistical thickness of the adsorbate layer as mentioned above. The α_S analysis was performed using the Dunino given in [41] as the reference adsorbent. Its adsorption isotherm has a shape similar to the adsorption isotherms of the tested samples, and is characterized by $S_{\mathrm{BET}}$ = 9.1 m²/g and $v\left(0.4\right)$ = 3.3087 cm³/g (STP).

The fractal dimension ($D$) is a measure of self-similarity of a geometrical object at various scales. For a surface, $2\le D<3$, and the higher $D$ the more folded and crimpled the surface. Therefore, the fractal dimension is often used to characterize the surface of porous materials [42,43]. In this paper, it was calculated by applying the Kiselev equation and the method described in [44]. In accordance with this approach, the following equation is valid:

$$\ln \left[-\frac{{\int }_{v\left(x\right)}^{v_{\max }}\ln x\mathrm{d}v\left(x\right)}{r_{\mathrm{c}}^2\left(x\right)}\right]=\mathrm{const}+D\ln \frac{{\left(v_{\max }-v\left(x\right)\right)}^{\frac{1}{3}}}{r_{\mathrm{c}}\left(x\right)},$$

(1)

where $r_{\mathrm{c}}\left(x\right)$ is the Kelvin radius corresponding to the relative pressure $x$. The value of $D$ was calculated by fitting the experimental data with a straight line in the range of 0.4 ≤ $x$ ≤ 0.99. Equation (1) is similar to the equation resulting from the thermodynamic method [45], but the arrangement of the experimental points is almost rectilinear in the pressure range broader than in the thermodynamic method or the Frenkel–Halsey–Hill method.

2.4. Cu(II) Sorption

Sorption capacity for Cu(II) was determined for raw and chemically modified samples. Sorption tests were carried out using a synthetic solution of copper nitrate (Cu(NO₃)₂⋅3H₂O) in batch static method. Water solutions of copper ions at initial concentrations of 25, 50, 80, 100, 200, 300, and 500 mg/dm³ were added to 1% aqueous suspensions of clay. The pH of the solutions was 4 and was determined using a solution of 0.1 mol/dm³ HCl or 0.1 mol/dm³ NaOH. The flasks with prepared samples were shaken on a rotary shaker for 2 h at 20 °C, after which they were placed in a dark room for 22 h. Then, the water from the conical flasks was decanted and centrifuged (MPW-2 type, MPW Med. Instruments, Warsaw, Poland) at 2500 rpm to remove the settling material from the solution. In the obtained eluates, the concentration of metal ions was analyzed using an inductively coupled plasma atomic emission spectrometer (ICP-AES). Three series of measurements were carried out for each sample, and in the following considerations the average values are considered. The amount of metal absorbed ($q$) is calculated as follows:

$$q=\frac{C_{\mathrm{i}}-C_{\mathrm{e}}}{m}V\left[\frac{\mathrm{mg}}{\mathrm{g}}\right],$$

(2)

where $C_{\mathrm{i}}$ and $C_{\mathrm{e}}$ are the initial and equilibrium concentration of Cu(II), respectively, $V$ is the volume of the solution, and $m$ is the mass of the sediment. In mathematical description of Cu(II) adsorption, the Freundlich and Langmuir isotherms were used, along with Freundlich–Langmuir, two-site Langmuir, and several other more complicated isotherms [46]. However, apart from the Langmuir and Freundlich isotherms the estimates for errors of parameter values were usually larger than the parameter values. Therefore, only results for the Freundlich and Langmuir isotherms will be presented. The Freundlich isotherm is purely empirical and can be written as follows:

$$q=K_{\mathrm{F}}{C}_{\mathrm{e}}^n\mathrm{or}\ln q=\ln K_{\mathrm{F}}+n\ln C_{\mathrm{e}}$$

(3)

where $K_{\mathrm{F}}$ and $n$ are Freundlich isotherm parameters. The Langmuir isotherm is derived assuming a kinetic model of adsorption, and has the following form

$$q=\frac{Q{K}_{\mathrm{L}}{C}_{\mathrm{e}}}{1+K_{\mathrm{L}}{C}_{\mathrm{e}}}\mathrm{or}\frac{C_{\mathrm{e}}}{q}=\frac{1}{Q{K}_{\mathrm{L}}}+\frac{1}{Q}{C}_{\mathrm{e}}$$

(4)

where $Q$ (sorption capacity) and $K_{\mathrm{L}}$ (Langmuir constant) are Langmuir isotherm parameters. The values of parameters $K_{\mathrm{F}}$, $n$, $K_{\mathrm{L}}$, and $Q$ were determined from the linear forms of the above equations.

3. Results and Discussion

3.1. Mineralogical and Chemical Composition

The rocks in the natural state have a light gray tint. In all samples, a high content of clay minerals was detected, ranging from 60% to 95%. The main clay minerals found in the mineral composition are smectites (beidelite), kaolinite, and calcite. In addition, quartz was found in small amounts in all samples, as well as trace amounts of siderite. The mentioned results for mineral composition are the average of three samples. The ash content is 91%, which corresponds to a loss on ignition (LOI) of about 9%. The chemical composition is presented in

Table 1. Chemical composition of the tested rocks.

Component	SiO2	Al2O3	Fe2O3	CaO	MgO	Na2O	K2O	MnO	P2O5	TiO2	SO3	LOI
wt %	56.1	23.2	4.05	2.7	0.97	1.4	0.47	0.97	0.17	0.03	0.08	9.2

. It reflects the mineral composition. Large contents of SiO₂ and Al₂O₃ are related to the presence of beidellite, kaolinite, and quartz. A relatively low content of Na₂O results from lack of feldspars and micas in the examined clays.

3.2. Nitrogen Adsorption and Desorption Isotherms

The nitrogen adsorption and desorption isotherms of the raw and chemically modified samples are presented in Figure 1. Hysteresis loops are visible on all of the isotherms. The adsorption and desorption branches join at a relative pressure of around 0.45. The hysteresis loops can be classified as H3/H4 (International Union of Pure and Applied Chemistry, IUPAC) or B/D (de Boer). H3 type is characteristic for loose aggregates of plate-like particles typical for some clays, and also in the case of pore networks containing macropores not completely filled with condensate. According to [47], H3-type loops indicate the presence of plaque-like mesopores with spaces between parallel plates. This seems to be confirmed in the subsequent section.

Classification of the isotherms themselves presents some difficulties. On the one hand, the presence of a hysteresis loop indicates type IV according to the IUPAC classification, but there is either no plateau characteristic for this type or it is difficult to observe (reduced to almost one point). On the other hand, the shape of the adsorption isotherms is best suited to type II, which in turn is characterized by a lack of hysteresis and no saturation at relative pressure near to 1; this is characteristic for nonporous and macroporous adsorbents [48]. Type II isotherms correspond to single- and multilayer physical adsorption on porous adsorbents. According to [49], the isotherms can be classified as pseudo-type II, describing delayed capillary condensation due to the small rigidity of the aggregate structure of the adsorbent.

Table 2. Pore volume of the tested samples.

Parameter	S1 (Raw)	S2 (HF)	S3 (H2O2)	S4 (H2SO4)	S5 (HCl)	S6 (HNO3)	S7 (NaOH)
$x_{\max }$	0.997	0.994	0.995	0.994	0.994	0.995	0.996
$d_{\max }$, nm	295	319	375	300	317	377	463
$V_{\max }$, mm3/g a	270	305	65	167	99	102	72
$V$($x$ = 0.99), mm3/g a,b	165	227	43	146	88	87	66

^a Evaluated as $V=kv$, where $k$ = 0.00154678, $v$ is volume at STP. ^b Volume of pores with diameter up to 190 nm (i.e., filled at a relative pressure of 0.99). Note: HF = hydrofluoric acid.

summarizes the pore volumes at the maximum pressure and the corresponding maximum pore sizes. These volumes cannot be compared directly because of the lack of a plateau on the isotherms, so Table 2 also shows volumes of pores filled at a relative pressure of 0.99, with a diameter of up to 190 nm (according to the Kelvin equation). These volumes range from 43 mm³/g for sample S3 to 227 mm³/g for S2. The pore volume in the raw sample is relatively large (165 mm³/g). The modification with HF increases this volume by approximately 37%, whereas the remaining modifications decrease it. The volume is slightly smaller in the sample treated with H₂SO₄, but in the samples modified with HCl and HNO₃ the volume decreased about twice, and in the samples modified with NaOH and H₂O₂ to about 40% and 26% of the initial value respectively.

3.3. Morphology

SEM images of samples S1–S3 are presented in Figure 2. In all images there are visible plates, which confirm the above mentioned remark about H3-type loops. The SEM image for sample S2 is shown in Figure 2c. In comparison to the raw material (Figure 2a,b), the plates seem cleaner and more emphasized. As shown in the subsequent sections, sample S2 had one of the largest specific surface areas (SSA) out of the tested samples. Figure 2d shows the SEM image for sample S3, which turned out to have the lowest SSA. Indeed, the plates seem to be “dusty” and the spaces between them are filled and unavailable.

3.4. Specific Surface Area

The parameters related to the SSA are shown in

Table 3. Specific surface area (SSA) of the tested samples determined via modified BET method.

Parameter	S1 (Raw)	S2 (HF)	S3 (H2O2)	S4 (H2SO4)	S5 (HCl)	S6 (HNO3)	S7 (NaOH)
$S_{\mathrm{BET}}$, m2/g	66.1(1)	105.8(1)	32.0(1)	106.5(1)	67.0(1)	67.9(1)	42.8(1)
$v_{\mathrm{m}}$, cm3/g (STP)	15.18(1)	24.30(1)	7.36(1)	24.48(1)	15.40(1)	15.61(1)	9.83(1)
$n$	3.1(1)	4.4(1)	3.0(1)	3.6(1)	3.1(1)	3.2(1)	3.1(1)
$C$	226(5)	122(3)	97(2)	91(2)	165(3)	173(4)	183(3)

The digit in the parentheses indicates the uncertainty of the preceding digit.

. The values of $n$ fall in the range of 3–4.4 and the values of $C$ range from 91 to 226, which means that point B (the knee) is well localized, and the interpretation of $v_{\mathrm{m}}$ as a monolayer volume is justified [48]. The determination coefficient ($R^2$) in all cases exceeds the value of 0.9999.

The values of the SSA range from 32 m²/g for sample S3 to 106.5 m²/g for sample S4. They can be described by relations S4 ≈ S2 > S6 ≈ S5 ≈ S1 > S7 > S3. The raw rock is characterized by an SSA of ca. 66 m²/g. The activation with hydrofluoric or sulfuric acid increases the SSA up to approximately 106 m²/g (i.e., over 60%). The modification with nitric acid or hydrochloric acid practically has no influence on the SSA value. The sodium base modification, on the other hand, reduces the surface to about 43 m²/g, and the modification with hydrogen peroxide decreases the SSA more than twice (down to 32 m²/g). A similar SSA has been revealed, for example, in bentonite modified with HCl (84.6 m²/g) [50], as well as in a bentonite-based magnetic composite (94.81 m²/g) [51]. Bhattacharyya and Gupta [52] obtained an SSA of 19.8 m²/g for montmorillonite, which after modification with 0.25 mol/dm³ H₂SO₄ increased to 52.3 m²/g.

Acids may affect the structure and chemical composition of layered silicates in various ways. Changes in the structure of minerals depend on the type of layered silicate and acid treatment conditions (type of acid used, its concentration and temperature). The use of 1 mol/dm³ HF or H₂SO₄ was sufficient to affect the octahedral and partially tetrahedral layers, thus partially replacing Al²⁺, Mg²⁺, Fe³⁺, and other ions with hydronium ones. The activation with HF of low concentrations, in addition to partial degradation of the clay structure, causes substitution of OH⁻ and network oxygen with F⁻ ions [53]. This leads to an increase in the amount and power of active surface centers, in particular the acid centers of the Brønsted and Lewis type. Structural changes caused by fluorination increase the surface and porosity, mainly in the range of mesopores (Table 3 and

Table 4. Summary of BJH analysis results using statistical adsorbate thickness $t_{\mathrm{ChJ}}$; for comparison, the pore volume obtained directly from the adsorption isotherm and BET surface area are also shown.

Parameter	S1 (Raw)	S2 (HF)	S3 (H2O2)	S4 (H2SO4)	S5 (HCl)	S6 (HNO3)	S7 (NaOH)
Pore size
Dominant size, nm	4.0	4.0	4.2	4.0	4.2	4.2	4.2
Average size a, nm	15.2	9.9	7.1	5.5	5.5	5.5	6.2
Average size b, nm	16.3	11.5	8.1	6.3	5.9	6.0	6.7
Pore volume
Total, mm3/g	269.5	305.1	64.5	166.9	99.3	101.8	71.6
Cumulative, mm3/g	269.6	320.1	68.3	180.6	105.0	108.2	78.9
macropores	177.9	113.1	28.3	31.2	18.6	21.5	8.6
mesopores	85.3	201.3	37.8	136.4	79.1	80.0	70.1
micropores c	6.4	5.7	2.2	13.1	7.3	6.7	0.2
Pore surface area
BET, m2/g	66.1	105.8	32.0	106.5	67.0	67.9	42.8
Cumulative, m2/g	70.7	129.2	38.5	131.0	76.8	79.4	50.5
macropores	4.7	2.1	0.6	0.7	0.5	0.5	0.2
mesopores	46.8	107.8	31.2	92.8	54.6	58.7	49.5
micropores c	19.2	19.2	6.7	37.4	21.7	20.1	0.8

Note: ^a $4V_{\mathrm{cum}}/S_{\mathrm{cum}}$, ^b $4V_{\max }/S_{\mathrm{BET}}$ , where $V_{\max }$ is the total pore volume. ^c Cumulative value minus meso- and macropore values.

). This has been documented by studies on acidic properties of montmorillonite [54], as well as sepiolite [55].

Dissolution of silica tetrahedrons in the clay mineral structure is unlikely to occur at low acid concentrations (below ca. 0.1 mol/dm³), as they are relatively stable and their destruction occurs only at higher acid concentrations [56]. The use of more concentrated mineral acids can cause leaching of cations located in octahedral sublayers. Usually, these are Mg²⁺, Al²⁺, and Fe³⁺ cations. A part of them may form oxide aggregates in the solution under appropriate conditions. Octahedral layers, on the other hand, are much more susceptible to dissolving at low acid concentrations [57]. The structure of the tested clay minerals was significantly more resistant to 1 mol/dm³ HCl and HNO₃, which only revealed a slight increase in SSA.

Bases affect the structure of the silicates in various ways, too. In this case, the SSA of the sample modified with NaOH decreases. Hence, the base causes a passivation process, involving partial blockage of the pores. A similar effect was reported in [58]. Jozefaciuk and Bowanko [37] found strong elution of silica by NaOH from bentonite rocks. It is, therefore, likely that in the present case, NaOH elutes silica from the octahedral layers, which then completely or partially blocks the micropores and smaller mesopores. The absence of small mesopores is evident in Figure 3, depicting the pore size distribution.

The use of H₂O₂ to improve structural properties did not succeed, which was somewhat surprising, since activation with H₂O₂ usually increases the SSA [59]. However, it may also happen that the SSA was reduced [38]. This may be caused by several reasons affecting the effectiveness of H₂O₂ treatment, such as concentration, time of treatment, and presence of various catalysts, SiO₂, and elements that build up in the mineral structure (e.g., iron) [60].

3.5. Mesoporosity

Although the BJH analysis is based on idealized pore geometries, it can deliver some information on mesoporosity. The analysis was performed for the statistical adsorbate thicknesses given in Appendix A. The most reliable results were assumed to be those for which the cumulative pore volumes were closest to relevant maximum values obtained from the adsorption isotherms. In each case, the best result was achieved for the thickness $t_{\mathrm{ChJ}}(x)$ (see Figure A1 in Appendix A). The size distribution of the pores is shown in Figure 3, whereas a comparison of the experimental results with the BJH analysis is presented in Table 4. The analysis with $t_{\mathrm{HJ}}(x)$ gives similar results, but with clearly higher cumulative volume and area values.

The dominant pore size is about 4.0 nm for samples S1, S2, and S4, and 4.2 nm for the remaining samples. However, the changes are very small and in the range of the uncertainty of the measurements and the analysis method. It is evident that neither HF nor H₂SO₄ increase the dominant pore size. The increase in SSA in this case results from the appearance of mesopores of size 2–15 nm (see Figure 3). In turn, HCl or HNO₃ increase the number of mesopores up to 10 nm in size. However, this does not cause a significant increase in the SSA. A small number of mesopores with a size up to 10 nm also appear in sample S7, but the mesopores of size below 4 nm seem to be almost eliminated, thus resulting in a decrease in the SSA. A similar effect is apparent for sample S3, in which there are virtually no smaller or larger pores, so that only a portion of those dominant in the raw sample remains accessible for adsorption.

The average pore size (assuming a cylindrical shape) was about 15 nm in the raw sample, whereas it decreased to 5.5–9.9 nm in the modified ones. The modifications, thus, led to either forming a number of pores with a diameter below 15 nm, or eliminating pores larger than 15 nm. This is consistent with earlier remarks.

In general, the cumulative volumes are usually larger than the measured ones. Likewise, the cumulative surface area is much larger than the one following from the BET analysis. This is due to the assumptions made in the BJH method concerning the shape of the pores as well as the assumed adsorbate thickness as a function of pressure. In spite of this, it is worth noting that the fraction of macropores evidently dominates in the raw sample (65% of the total pore volume), and the volume fraction of mesopores is about 31% of the total pore volume. After chemical treatment it considerably rises to above 50% of the total pore volume.

3.6. Microporosity

The results of microporosity analysis using the DR theory are shown in

Table 5. Microporosity analysis.

Quantity	S1 (Raw)	S2 (HF)	S3 (H2O2)	S4 (H2SO4)	S5 (HCl)	S6 (HNO3)	S7 (NaOH)
DR theory
$V_0$, mm3/g	24.7(4)	36.9(8)	12.1(2)	36.5(7)	23.9(4)	25.0(4)	15.8(2)
t-plot
$V_{\mathrm{micro}}$, mm3/g	6.6(2)	3.9(2)	1.3(1)	2.3(1)	4.7(3)	4.8(3)	3.7(2)
$S_{\mathrm{ext}}$, m2/g	39.9(5)	75.8(5)	22.3(2)	78.1(3)	44.0(6)	44.6(6)	26.8(3)
$S_{\mathrm{micro}}$, m2/g	21.5(6)	27.4(6)	7.2(3)	23.6(4)	18.4(7)	19.3(7)	13.0(4)

The digits in the parentheses indicate the uncertainty of the preceding digits.

. The matching of the linear form of the DR equation to the experimental data was very good ($R^2$ ≥ 0.99). The volume of micropores determined by this method ranges from about 12 mm³/g for sample S3 to about 37 mm³/g for sample S2. The volume of the micropores of the raw sample is about 25 mm³/g. The modifications with HF or H₂SO₄ result in an increase in micropore volume by about 50%. The other two acids (HCl and HNO₃) practically do not change the volume of the micropores. The use of NaOH significantly reduces this volume. The largest drop (by roughly 50%) was obtained with the use of H₂O₂. These results are consistent with the BET SSA values. The values of the exponent $m$ according to the Dubinin–Astakhov theory were also calculated to get the best fit. The interval 2 ≤ $m$ ≤ 5 was considered. In each case, the best results were obtained for $m$ at close to 2.

Table 5 also shows the results of t-plot analysis using the statistical layer thickness $t_{\mathrm{ChJ}}$ (the results using $t_{\mathrm{HJ}}$ were very similar). Although a linear fit of the experimental data in the range 1–1.5 monolayer thickness is very good ($R^2$ is close to 1), the volumes of micropores calculated using this method are much smaller than using the DR equation and range from 1.3 to 6.7 mm³/g. The relationships between samples are slightly different too—the largest volume of micropores was obtained in the raw sample. Modification with HCl or HNO₃ slightly reduces the content of micropores. Even smaller volumes of micropores are obtained for the samples modified with NaOH, as well as HF and H₂SO₄. The smallest micropore volume was for the sample treated with H₂O₂. The rather small volume of micropores for samples S2 and S4, deviating from the expected values greater than the raw sample, can be explained by the presence of relatively large micropores in these samples. They are not yet fully filled at the monolayer formation stage, and the assumptions of the t-plot method are not true in such a case.

The external surface area fulfills the relations analogous to the total SSA determined by the BET method. The surface area of micropores, understood as the difference between the BET surface and the external surface, is 21.5 m²/g in the raw sample. The modification with HF or H₂SO₄ increases this surface, whereas other modifications reduce it. The largest (almost a three-fold) drop is observed in a sample modified with H₂O₂.

The results of the α_S analysis of the examined samples are presented in Figure 4. Two α_S ranges were considered: 0.5–0.9 and 2–5. The tables included in the drawings show the available surface area and the volume of already filled pores at particular stages. The slope of the straight line in the first interval, approximately corresponding to the formation of a monolayer, is proportional to the surface area. These values agree with the results obtained using the BET method. The intersection of this line with the vertical axis indicates the volume of pores occupied in the previous condensation stage. For S2 and S4, the lines intersect slightly below zero, so it is assumed that the intersection is at zero in these cases.

The presented α_S graphs do not contain the characteristic c-swing associated with capillary condensation (except for sample S2, in which a slight c-swing is noticeable). This is because capillary condensation takes place at high relative pressures in the samples under consideration, which is characteristic for pseudo-second-type isotherms. However, all samples are characterized by a distinct f-swing, indicating the presence of micropores.

3.7. Fractal Dimension

The results of the fractal dimension are shown in

Table 6. Fractal dimension of tested samples.

Quantity	S1 (raw)	S2 (HF)	S3 (H2O2)	S4 (H2SO4)	S5 (HCl)	S6 (HNO3)	S7 (NaOH)
$D_{\mathrm{ads}}$	2.49(1)	2.72(1)	2.61(2)	2.75(1)	2.74(1)	2.76(1)	2.75(1)
$D_{\mathrm{des}}$	2.55(1)	2.84(3)	2.58(6)	2.89(3)	2.92(4)	2.93(4)	2.98(3)

The digits in the parentheses indicate the uncertainty of the preceding digits.

. The relative pressure range was 0.4–0.99. Matching the data to a straight line is very good ($R^2$ > 0.99). The smallest fractal dimension is observed in the raw sample (2.49–2.55). Acid modifications cause an increase in the dimension to about 2.72–2.93. The treatment with NaOH results in a similar increase in the fractal dimension. The smallest increase in the fractal dimension is recorded for the H₂O₂ modified sample (2.58–2.61). It is worth noting that the fractal dimension was also calculated using the thermodynamic and Frenkel–Halsey–Hill methods, but in both cases the range of data linearity was much smaller.

3.8. Cu(II) Sorption

Changes in the Cu(II) concentration in the solution as a result of sorption processes are presented in Figure 5. The amount of sorbed Cu(II) ions increased in the series H₂O₂ < NaOH < raw ≈ HCl ≈ HNO₃ < H₂SO₄ < HF. The raw sample was characterized by a maximum sorption (obtained at the initial concentration of 500 mg/dm³) equal to about 15 mg/g. Similar values were obtained for samples modified with HCl and HNO₃. The samples modified with HF and H₂SO₄ proved to be the best sorbents of copper ions (maximum sorption capacity of about 19 mg/g and 18 mg/g, respectively). The adsorbents showing the weakest Cu(II) sorption are samples modified with NaOH and H₂O₂. They adsorbed 13 mg/g and 12 mg/g of Cu(II), respectively, at the maximum initial concentration of metal in the solution. The cation exchange capacity (CEC) of the studied clays is between 31.68 and 57.41 cmol₍₊₎/kg [61]. Comparing the maximum amount of copper ions retained with their CEC, it was found that Cu(II) ions in all samples tested were bound below the CEC value. All values of sorption capacity obtained in the experiment are lower than for activated carbons and also some biosorbents, such as pomegranate peels, for which the maximum sorption capacity values were 43.47 mg/g and 30.12 mg/g, respectively [62,63]. On the other hand, they are similar to those obtained for kaolinite and montmorillonite and their acid-activated forms, for which the Langmuir monolayer capacity was 4.3–28.0 mg/g [64] and much higher than for unmodified bentonite (8.33 mg/g) [65].

Based on the obtained results, the parameters of Freundlich and Langmuir isotherms were determined (

Table 7. Parameters of Freundlich and Langmuir isotherms for Cu(II) sorption on the tested clays.

Parameter	S1 (Raw)	S2 (HF)	S3 (H2O2)	S4 (H2SO4)	S5 (HCl)	S6 (HNO3)	S7 (NaOH)
Freundlich isotherm
$K_{\mathrm{F}},\mathrm{mg}/\mathrm{g}{\mathrm{dm}}^{3n}/{\mathrm{mg}}^n$	1.4(4)	1.4(3)	1.4(3)	1.9(3)	1.8(3)	1.9(3)	1.3(2)
$n$	0.43(7)	0.46(5)	0.38(5)	0.40(4)	0.38(5)	0.36(4)	0.41(4)
$R^2$	0.871	0.943	0.926	0.943	0.933	0.948	0.945
Langmuir isotherm
$Q,\mathrm{mg}/\mathrm{g}$	16.4(7)	21(1)	13.0(5)	19(1)	16.7(7)	16.2(8)	14.4(9)
$K_{\mathrm{L}},{\mathrm{dm}}^3/\mathrm{mg}$	0.022(4)	0.017(4)	0.025(4)	0.025(5)	0.027(5)	0.027(6)	0.019(4)
$R^2$	0.992	0.976	0.993	0.987	0.990	0.987	0.983

Note: $R^2$—determination coefficient; the digits in the parentheses indicate the uncertainty of the preceding digits.

), obtaining in all cases a better fit for the Langmuir model ($R^2$ from 0.976 to 0.993), although in the case of Freundlich isotherms it was also acceptable ($R^2$ from 0.871 to 0.948). For this reason, only the Langmuir isotherms are shown in Figure 5. The highest ability for Cu(II) retention by samples S2 and S4 is confirmed by the Langmuir maximum adsorptive capacity, $Q$, which is 21 ± 1 mg/g and 19 ± 1 mg/g, respectively. This means an improvement in the adsorption capacity by 29% and 18%, respectively. The higher $K_{\mathrm{L}}$ value for the S4 sample means that the adsorbent saturation occurred at lower values of copper ion concentration than in the case of S2.

Good estimation of the experimental data by the Langmuir isotherm confirms that the binding of copper ions by the clay is limited to the monomolecular layer covering the entire surface of the adsorbent. Cu(II) ions may, thus, be bound by chemisorption (specific sorption) to form stable ionic or covalent bonds between metal ions and the sorbent surface [66]. Strong binding of copper ions is very beneficial due to their immobilization, low susceptibility to release, and reduction of reactivity.

The sorption of copper ions on the tested materials decreases with an increase in the initial concentration of Cu(II) (Figure 6). The percentage of sorption is considerable and ranges from 19% to 86% for the initial concentrations tested. The percentage of sorption for raw material is from 56% to 80% at lower initial Cu(II) concentrations. At an initial concentration of 500 mg/dm³, it is 23% for the natural sample and about 30% for the acid-activated HF and H₂SO₄ (S2 and S4); however, for S3 and S7 samples activated with base and hydrogen peroxide, the percentage of sorption is about 20%. Samples S3 and S7 adsorbed the smallest amounts of Cu(II), most likely as a result of the passivation process, which is reflected in SSA and porosity changes (Table 3, Table 4 and Table 5).

3.9. Industrial Wastewater Pretreatment

The results presented in the above sections indicate that the considered post-mining rocks could be used as low cost adsorbents for industrial wastewater pretreatment. The idea of the process is presented in Figure 7. First, the rocks are prepared for metal ion adsorption. After grinding them to a suitable grain size, the material is placed in the adsorption column. Alternatively, it can be modified by a 1 mol/dm³ H₂SO₄ or 1 mol/dm³ HF before placement in the column, but this stage is optional and may only be required to enhance the adsorption process. The raw industrial wastewater is directed into the settling tank for preliminary sedimentation and pH correction. The decanted wastewater is then transferred to the column for metal ion adsorption. The pretreated wastewater can then be directed for another treatment stage, depending on the degree of contamination and the required degree of purification. To enhance the adsorption process and avoid unnecessary downtime, two or more columns can be used in a cycle, so that when one column is working the other is being prepared for use by cleaning and filling with another portion of adsorbent. The used adsorbent can be directed for utilization, or alternatively it can be washed and reused in another cycle of the adsorption process.

4. Conclusions

Based on the above results, the following conclusions can be formulated:

• Raw clays under testing have an SSA of approximately 66 m²/g; modifications with 1 mol/dm³ HF or H₂SO₄ increase the SSA over 60%; modifications with 1 mol/dm³ HCl or HNO₃ practically do not change the SSA; treatment with 1 mol/dm³ NaOH or H₂O₂ results in a significant decrease in SSA. The SSA of the tested samples can be described by relations: H₂SO₄ ≈ HF > HNO₃ ≈ HCl ≈ raw > NaO > H₂O₂.
• Adsorption isotherms (pseudo-second type with H3 hysteresis loop) indicate the presence of non-rigid aggregates, which is typical for some clays.
• Chemical treatment changes the pore size distribution—acids produce small and medium mesopores, NaOH removes small mesopores, and H₂O₂ comprises the majority of mesopores.
• The volume of micropores according to DR theory corresponds to relations analogous to the SSA. The presence of micropores is also confirmed by α_S analysis.
• The fractal dimension of the raw sample is 2.49–2.55. After chemical treatment, it increases to 2.61–2.98.
• Cu(II) sorption results reflect the structural and surface changes in the samples. The highest sorption capacity is obtained for modifications with HF or H₂SO₄.
• The investigated clays may be considered as prospective inexpensive adsorbents. Two aspects of cleaner production are obtained, in lignite extraction as well as in ion metal removal from specific industrial wastewater. For such a purpose, HF and H₂SO₄ produce the most favorable changes—they result in a significant increase in SSA, the formation of small and medium mesopores, as well as micropores (according to the DR theory). This shows a reflection in tests of Cu(II) sorption.