Next Article in Journal
The Chemical Composition and Surface Texture of Transparent Heavy Minerals from Core LQ24 in the Changjiang Delta
Next Article in Special Issue
Spanish Bentonites: A Review and New Data on Their Geology, Mineralogy, and Crystal Chemistry
Previous Article in Journal
Effect of Comminution Methods on Low-Rank Coal Bubble–Particle Attachment/Detachment: Implications for Flotation
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Minerals
Volume 9
Issue 7
10.3390/min9070453
minerals-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Evaluation of the Sorption Potential of Mineral Materials Using Tetracycline as a Model Pollutant

by
Jaime Cuevas
1,*,
Nisael Dirocie
2,
Felipe Yunta
2,
Carlos García Delgado
1,
Daniel E. González Santamaría
1,
Ana Isabel Ruiz
1,
Raúl Fernández
1 and
Enrique Eymar
2
1
Department of Geology and Geochemistry, Autonomous University of Madrid, 28049 Madrid, Spain
2
Department of Agricultural Chemistry and Food Sciences, Autonomous University of Madrid, 28049 Madrid, Spain
*
Author to whom correspondence should be addressed.
Minerals 2019, 9(7), 453; https://doi.org/10.3390/min9070453
Submission received: 18 June 2019 / Revised: 15 July 2019 / Accepted: 18 July 2019 / Published: 21 July 2019
(This article belongs to the Special Issue Special Clays and Their Applications)
Browse Figures
Versions Notes

Abstract

:
Tetracycline (TC) is among the most used antibiotics in animal feedstock in the EU. Antibiotics’ persistence as emerging pollutants in the environment is evidenced by their long half-life in residual organic-mineral sediments and waters. The risk associated with this persistence favours antibiotic-resistant microbiota, affecting human health and ecosystems. The purpose of the present work is to assess the adsorption of TC into natural clay minerals, synthetic iron hydroxides and calcined sewage sludge. TC adsorption isotherms were performed in three replicated batch tests at three different pH values (4, 6, 8) and TC concentrations (33–1176 mg·L−1). X-Ray diffraction (XRD) mineralogy, cation exchange capacity (CEC), Brunauer, Emmett and Teller specific surface area (BET-SSA) and point of zero charge salt effect (PZSE) were determined for the characterization of materials. Sorption was analysed by means of fitting Langmuir and Freundlich adsorption models, which showed good fitting parameters for the studied materials. Low-charge montmorillonite (LC Mnt) is displays the best sorption capacity for TC at maximum TC concentration (350–300 mgTC·g−1) in the whole range of pH (4–8). Sepiolite and smectites adsorbed 200–250 mgTC·g−1, while illite, calcined sludge or iron hydroxides present the lowest adsorption capacity (<100 mgTC·g−1). Nevertheless, illite, sepiolite and ferrihydrite display high adsorption intensities at low to medium TC concentrations (<300 mg·L−1), even at pH 8, as is expected in wastewater environmental conditions.

1. Introduction

Tetracycline (TC) is the most used antibiotic in animal feedstock in the EU. In 2016, TCs, penicillin and sulphonamides were the most-sold antibiotics, accounting for 32%, 26% and 12%, respectively, of use in food-producing species for those 25 countries delivering data from 2011 to 2016 [1]. They are used in medicine and agriculture [2]. The persistence of these emerging pollutants in the environment is due to their long half-life, on the order of magnitude of 0.5 years, so they become relatively immobilized in residual organic-mineral complexes incorporated into sediments and soils [3]. This behaviour favours their accumulation in sediments and eventual release due to seasonal flooding related to Mediterranean hydrological regimes [4]. In addition, wastewater treatment plants are not efficient at reducing or degrading antibiotics and other pharmaceutical soluble products, so they become adsorbed on biowaste debris, responsible for their fixation [5]. The risk associated with this persistence is the development of antibiotic-resistant microbiota that can affect ecosystems and human health [6].
The removal efficiency (%) of TC and other antibiotics in wastewater treatment plants has been reported to be 80–100% depending on the water retention times, which need to reach up to 60 days to obtain high removal efficiency [7]. Typical effluents show remaining TC concentrations of <1 µg·L−1. Such amounts could be easily retained by natural clay minerals, which can adsorb from 5 mg·g−1 in the case of kaolinite [8] to 20 and >300 mg·g−1 in the case of 2:1 sheet silicates as illite and several montmorillonites, respectively [8,9,10]. This is to say that 1 g of mineral could adsorb the TC contents of >300 m3 of wastewater.
TC sorption onto high-surface-area (50–100 m2·g−1) swelling sheet silicates such as montmorillonite (Mnt) has been reported elsewhere [11,12,13]. The acid–base aqueous chemistry of the TC molecule makes its sorption potentially favourable at acidic pH (<4) by means of cation exchange mechanisms [14]. The protonation or deprotonation of three different functional groups in the TC molecule (Figure 1) causes the TC to be positive (TCH3+ pH < 3.3), neutral zwitterionic (TCH2± 3.3 < pH < 7.7) or negative (TCH− pH > 7.7) [15,16]. Therefore, TC will mainly be in the zwitterionic form under environmental conditions. In fact, the 2:1 structures can also allocate TC at circumneutral pH (7–8) when the molecule becomes neutral or negatively charged by means of interlayer TCH−–cation–polar surface interaction bridges [9]. These studies did not explore the potential differences in organic compound–clay interaction due to the particular crystallochemical nature of smectites, including distinct features of charge density (high charge, low charge), distribution of layer charge (tetrahedral, octahedral), chemical composition [17] or the inherent acid–base surface chemistry of the sorbents used [18].
Mineral sorbents involve a large group of raw materials of natural origin that are commonly used due to their advantages [20] (e.g., relatively low cost and ready availability). In this group, it is worth mentioning the clay minerals; these nanostructure materials cover several types of hydrous aluminium and magnesium phyllosilicates that differ in the arrangement of the tetrahedral and octahedral sheets and the distribution of surface electric charge. They exhibit special properties such as high sorption, ion exchange capability or swelling behaviour, resulting from their nonpareil structure, the presence of surface OH groups and weak electrostatic interactions between layers and/or sheets and the exchangeable cations [21].
A number of high surface available mineral materials have been used in this study for screening their potential capacity to adsorb TC as a model for ionic water-soluble emerging pollutants. The aim of this work is to assess the adsorption capacity of TC onto mineral sorbents such as clay minerals and synthetic iron hydroxides. The sorption capacity of natural clay minerals, synthetic iron hydroxides and calcined sewage sludge on TC has been evaluated by: (1) studying their interaction in batch experiments, determining the sorption potential of each mineral sorbent through adsorption isotherm analysis; and (2) discussing this capacity in terms of TC speciation and solid surface interaction through determination of the point of zero charge salt effect (PZSE) of each material under study. The whole work is intended to contribute to selecting easily available mineral products to retain emerging pollutants coming from pharmaceutical industries in order to study potential innovative processes to improve water and crop soil safety for ecosystems and human health.

2. Materials and Methods

2.1. Tetracycline

The TC used in this work was obtained commercially from Sigma-Aldrich (Steinheim, Germany) for veterinary use as T3383, TC hydrochloride (MW 480.898 g/mol), with purity >95%.

2.2. Mineral Sorbents

The potentially sorbent materials used for this study were selected based on their high surface area properties and on previous knowledge gained during their use in several characterization studies [21,22,23,24,25]. Several clays, iron hydroxides and calcined sewage sludge were used. Figure 2 shows the broad peaks characterizing the XRD powder patterns of the selected materials in which the crystal size coherent diffraction domain has been calculated between 13 Å (synthetic ferrihydrite, Fh) and 280 Å (MX-80 (LC Mnt), parallel to a sheet silicate 2:1 structure dimension). This confirmed the very small grain sizes for the particles forming the selected materials. A more detailed crystal–chemical description of the nature of the clay minerals and materials used is given in Section 3.1 and the Supplementary Materials (S1).

2.2.1. Clay Minerals

Sepiolite, with the idealized structural formula [Mg8·Si12O30 (OH)4·(OH2)4]·8H2O, was supplied by TOLSA S.A. (Madrid, Spain), commercialized under the PANSIL100® (PAN100) brand and extracted from mineral deposits from the Madrid basin, as along with stevensite. PAN100 is a mineral compound of approximately 95% sepiolite (<5% feldspars and quartz; <1% calcite). Sepiolite is a natural sheet silicate of fibrous morphology with no permanent layer charge [26].
Stevensite was supplied commercially by TOLSA, S.A. under the MINCLEAR N100® (MINC100) brand (>90% stevensite, 5% dolomite, <5% quartz and feldspars). Stevensite was supplied commercially by TOLSA, S.A. under the MINCLEAR N100® (MINC100) brand (>90% stevensite, 5% dolomite, <5% quartz and feldspars). Stevensite belongs to the trioctahedral smectites 2:1 sheet silicates group [27]. In stevensite, the permanent layer charge arises from the existence of divalent cation vacancies (M+0.5·[Mg2.5□0.5·Si4O10·(OH)2]·nH2O; □: vacancy; M+: cation counterion for permanent negative layer charge) (i.e., [28,29]).
Four Al-dioctahedral smectites have been used in the study: a beidellite type, one low-charge montmorillonite and two high-charge Mnt. Beidellite (ideally (M+0.5·[(Al,Fe(III))3·Si3.5Al0.5·O10(OH)2]·nH2O)) is the main component of the fine fraction of clay minerals included in arkose sandy sediments [10]. MX-80 bentonite supplied by AmColl (American Colloid Company, Hoffman Estate, IL, USA) (95% smectite, 5% quartz and feldspars) contains Mnt (M+0.35·[(Al1.55, Fe(III)0.20,Mg0.25)·Si3.90Al0.10·O10(OH)2]·nH2O; It is a low-charge montmorillonite (LC Mnt) (M+ < 0.374, [30]), and has Na+ as the dominant interlayer cation. FEBEX (>95% smectite, <5% albite, cristobalite, quartz) and MMt-Chile (>95% smectite, <5% pyrophillite, albite, heulandite, quartz) bentonites are high-charge Mnt (HC Mnt) (M+ > 0.426), (M+0.50·[(Al1.35,Fe(III)0.20,Mg0.45)·Si3.95Al0.95·O10(OH)2]·nH2O) and (M+0.55·[(Al1.35, Fe(III)0.10,Mg0.60)·Si4.00·O10(OH)2]·nH2O), respectively; MMt-Chile is characterized by a low iron content in comparison to the other Mnt (HC Mnt NFe).
A red clay from Carboneros (Bailén, Jaén; [31]) was supplied commercially by Cerámica Comercial Bailén S.A. (Bailén, Spain) This clay is more than 70% illite, 20% quartz and feldspars and 5% dolomite. Illite 2:1 is a di-octahedral clay mineral with non-exchangeable dehydrated K+ cation in the interlayer region.

2.2.2. Iron Hydroxides: Ferrihydrite and Goethite

Two-line ferrihydrite and goethite was synthesized as described elsewhere [7,8]. Either 5 or 1 M NaOH solution was slowly added to 0.4 M Fe(NO3)3 up to a final pH of approximately 7.0. The resulting sample suspension was allowed to age for 22 h at 20 °C in the dark. It was then repeatedly washed with ultrapure water until the conductivity was less than 10 μS·cm−1. The aged suspension was freeze-dried and ground to obtain a reddish-brown powder. A similar procedure was followed for goethite synthesis. In fact, NaOH volumes were slowly added to a 0.4 M Fe(NO3)3 solution to a final pH of 12. The resulting suspension was allowed to age for 72 h at 65 °C in the dark, and after several washes, the resulting dialyzed suspension was freeze-dried and ground to obtain a yellowish powder [32].

2.2.3. Calcined Sewage Sludge

This residue comes from the gasification of a water treatment plant sewage sludge. It is a dried and crushed material (<2 mm size) calcined at 800 °C as the residue of a gasification process [33].

2.3. Characterization of Materials

All the materials were characterized by means of XRD powder patterns (Figure 2) confirming their crystalline properties and estimating their crystal size (Table 1). The mineralogical analysis of the samples was carried out by XRD using a PANalytical B.V. (Almelo, The Neederlands) X’Pert PRO Theta/2Theta diffractometer with Ge (111) as the primary monochromator. This procedure allowed for the selection of CuKα1 radiation analysed with an X’Celerator detector. The samples were registered in the range of 3° < 2θ < 70°, with a step size of 0.0167° and a counting time of 100 s for each step. The samples were analysed by the random powder method.
The Brunauer, Emmett and Teller (BET) equation was used to determine the adsorption capacity of materials by nitrogen gas adsorption at 77 °K in a Gemini V Micromeritics® equipment after degassing the powdered sample under N2 flow for 18 h at 90 °C (UNE 22-164/94).
The cation exchange capacity (CEC) was determined by the Cu-trien complex method, as described in previous papers using these materials [22,23,24].
The PZSE (point of zero charge salt effect) for all study materials was determined by acid and alkaline potentiometric titrations at four ionic strength concentrations, 0.001, 0.01, 0.05 and 0.2 M NaCl, as described by [32]. Thus, the addition of protons and hydroxides was calculated by subtracting those remaining in solution with the solid material from those that remain in solution without the solid material for each ionic strength concentration. Finally, interpolation data were calculated and H+–OH− data were determined at regular pH increments.

2.4. Batch Sorption Experiments

The TC concentrations used in the present study are large and far from the real concentrations expected in the environment, which should be <50 mg/L and are usually <0.2 mg/L, as reviewed by [7,16]. However, they were used to cover the potential ion exchange capacity of the high-surface-area mineral sorbents. Mnt have approximately 1 mmol (negative charge)/g, which corresponds approximately to the maximum TC concentration prepared (1.0 mmol TC/g adsorbent). In addition, the pH used in the batch experiments was designed to cover a slightly acidic pH range, corresponding to the pH of acetogenic organic leachates, to urban wastewaters, which lie commonly in the pH interval 6–8. In these pH conditions, the TC zwitterionic species (TCH2±) is predominant. According to [19], at pH = 4, 20% of TC in solution is in the TCH3+ form, and at pH = 8, the species TCH− reach 60%, while at pH = 6, 100% of TC is as TCH2±.
The TC sorption capacity onto tested materials was studied in batch experiments adopting the usual performance for adsorption isotherms. The estimated required time to equilibrate the adsorption of TC in Mnt and other mineral constituents was found in the literature, with significant variations ranging from 0.5 to 48 h [34,35,36,37]. An experiment was performed at three different adjusted pH values, 4, 6 and 8. The time was selected according to the change of the adjusted pH of clay minerals suspensions after 12 h soaking. A 0.5 pH unit change was accepted for the isotherm experiments. This change was maximum for stevensite adjusted to acid pH after 2 h (Figure 3). The other materials underwent lower changes in the whole pH range. According to [14], >80% of TC becomes adsorbed at 2 h, with the sorption rate very slow from this time to 48 h. Then, 2.5 ± 0.01 g of clay minerals and 4 ± 0.01 g of iron hydroxides were ultrasonically dispersed in 1 L of deionized water for 5 min at 20 kHz and 200 W. The process was repeated five times after a cycle of 30 min of continuous stirring. Alkali or acid volumes for pH adjustment were recorded, and dilution factors were taken into consideration for final calculations to reach the working pH values of 4, 6 and 8. In parallel, a 10 g TC·L−1 standard solution was used to spike the total dissolved TC concentration to obtain 33, 132, 322, 625 and 1176 mg·L−1. The batch experiment was performed in 20-mL polypropylene centrifuge tubes [38]. Each tube contained 15 mL of mineral suspension, 0.087 ± 0.002 g of NaCl (0.1 M of electrolyte) and TC in different concentrations adjusted to the desired pH. A blank solution was prepared containing only material suspension at 0.1 M NaCl and for each study pH value. Polypropylene centrifuge tubes were left to interact in an orbital shaker at 120 rpm. and 25 ± 2 °C for 2 h in the dark to avoid TC photodegradation processes [39]. Three replicates were arranged for each of the batch experiments, i.e., for all pH conditions, the experiments were performed three times in triplicate for each material. The dissolved TC concentration was spectrophotometrically determined on the supernatant fraction at 254 nm in a Spectronic™1200 UV-VIS spectrometer (Spectronic, Spain), after centrifugation at 10,000 rpm. for 5 min.

2.4.1. Adsorption Isotherms

The TC adsorption pattern was assessed using Langmuir and Freundlich isotherm models. Isotherm parameters from both models were determined by non-linear curve fitting by minimizing the sum of squares between the calculated and experimental concentrations of TC [40]. Molar units (TC, MW: 480.898 g/mol) and mass units (sorbents) have been used in isotherm calculations according to [41]. µmol and g are used to give comprehensive entire numbers in fitted parameters and graphical representation of the isotherms. For the Langmuir model, the adsorption isotherm can be mathematically expressed as:
Cs = (K × Cm × Ce)/(1 + K × Ce),
where Cs: TC concentration adsorbed (µmol·g−1) at equilibrium; Ce: TC concentration in solution at equilibrium (µmol·L−1); Cm: maximum adsorption or apparent adsorption (µmol·g−1); K: equilibrium constant (L·µmol−1).
For the Freundlich model, the adsorption isotherm can be adapted to the following equation:
Cs = Kf × Ce(1⁄n),
where Cs: adsorbed amount of TC per unit amount of sorbent (µmol·g−1); Ce: initial concentration of TC in the bulk solution (µmol·L−1). Kf: constant that indicates adsorption capacity of the sorbent ((L/mg)1/n); and n: is a constant that indicates adsorption intensity. It is restricted to >1 values. A range between 2 and 10 indicates good adsorption intensity [42].

2.4.2. Statistical Analysis

Factorial analysis of variance (ANOVA) was initially conducted (STATGRAPHICS 5.1 for Windows®; Statgraphics technologies, The Plains, VA, USA. Thus, sorbent materials, initial TC concentrations (33,2; 131.6; 322.6; 625.0 and 1176.4 mg·L−1) and tested pH values (4, 6 and 8) were entered as factors in order to assess the significant difference level (p < 0.05). One-way ANOVA (SPSS 24.0 for Windows; IBM company, Armonk, NY, USA) was later performed among those significant interactions, and a Tukey post hoc test (p < 0.05) was conducted.
Multivariate statistical methods were used in two ways, as follows: to classify the different materials according to the analytical properties; and to discriminate them from adsorption outcomes. Discriminant analysis was performed to distinguish among the studied materials based on the set of experimental constants from the Freundlich and Langmuir models. Thus, solid phases were grouped (clusters) by comparing the two functions defined by the Langmuir and Freundlich equations corresponding to adsorption and intensity capacity, respectively. Conglomerate dispersion diagrams were determined by the Euclidean quadratic K-average method by using the STATGRAPHICS® v.16.1 statistical package (Version 16.1, Statgraphics technologies, The Plains, VA, USA).

3. Results and Discussion

3.1. Physical and Mineralogical Characterization of Synthesized and Natural Sorbent Materials

The mineralogical composition of the studied materials can be described from the XRD powder patterns. Quartz and dolomite were detected as the main impurities determined in several of the smectite, sepiolite and illite clay mineral materials studied. They are high-grade clay materials with <10% impurities. Synthetic iron hydroxides are virtually monomineral samples of goethite (goet) and ferrihydrite (fh). Fh (two line type) is characterized by the presence of two broad bands at 2.54 Å ((hkl):(200)) and 1.49 Å ((hkl):(300,213)). In the case of the calcined sewage sludge XRD analysis showed that this is an amorphous glass with quartz and feldspar impurities (Figure 2).
The specific surface area of the different materials, complementary to the BET SSA method, can be estimated from the crystal size calculation using the Scherrer equation (angular breadth of an XRD reflection at half maximum peak height; ∆2θ = λ/Lcosθ [43], where L is the crystal thickness in nm (λ CuKα 0.154 nm) in the perpendicular direction to a characteristic (hkl) plane, and θ is the angle for the existing Bragg reflection condition (XRD reflections in Figure 2). A hexagonal cell is assumed for a single crystal of ferrihydrite, characterizing the size of the synthesized material, with an equal number of unit cells in the a direction, as measured by means of the (h00) reflections (2.54–1.49 Å; two-line ferrihydrite) or the c direction (c is the height of the hexagonal cell: a = 5.08 Å, c = 9.40 Å). For goethite and 2:1 clay minerals, a rectangular (a × b) polyhedron of height c is considered to build the unit cell. Goethite crystal size, determined by the (h00) reflections, is used assuming equal development of unit cells in the a, b or c directions (4.59 × 9.96 × 3.02 Å). The number of unit cells is calculated by dividing the crystal size by the cell parameter a (for n cells) and multiplying na × nb × nc to obtain the total number of cells (N) in a single crystal of goethite. If the density is known, the weight of a unit cell (Wc) can be calculated. The weight of a single crystal is N × Wc. The area of the crystal parallelogram is calculated as 2na × nb + 2nb × nc + 2na × nc, and therefore the specific surface area in m2g−1 is calculated by transforming Å units to m. In the case of 2:1 sheet silicates, the b dimension is calculated as d(060) × 6, the c dimension as d(001), and the a dimension is taken from the literature as 5.2 Å for aluminium di-octahedral species (1.494–1.503 Å for the (060) reflection) and 5.3 Å for stevensite (1.523 Å for the (060) reflection (Figure 2)). The sepiolite surface area was calculated using the (011) reflection and crystal size calculation by the procedure given by [44]. The calculated areas were corrected by the semi-quantitative amount of clay minerals in the characterized materials. Table 1 gives data on crystal size and calculated surface areas.
There is a good correlation between the calculated and measured surface areas, excluding Sewage S. (not possible to calculate) and ferrihydrite (Figure 4), where presumably some disorder in the crystalline clusters could affect XRD reflection broadening, overestimating this parameter. The BET surface area obtained for the synthesized ferrihydrite was 253.3 ± 0.9 m2·g−1. This result is in line with those obtained by [32] and within the interval 200–350 m2·g−1 indicated by [45], although [46] listed values up to 720 m2·g−1, obtained by various techniques and adsorbents. The synthesized goethite was at 59.0 ± 0.3 m2·g−1, in agreement with the interval 30–100 m2·g−1 [47,48] and with the XRD calculated surface area. Based on these contrasted surface areas, materials such as ferrihydrite, sepiolite or stevensite should offer the highest available external crystal grain surfaces for sorption processes.

3.2. Point of Zero Salt Effect

The materials tested for the PZSE were intentionally not purified to determine the acid–base surface behaviour in their natural state as far as they are intended to be used as an inexpensive alternative for pollutant retention in wastewater treatment plants. Raw materials may develop an acid-neutralizing capacity if carbonate impurities or other dissolution reactions are taking place. In this case, there will not be a pronounced ion-strength effect in the behaviour of the materials if the proton adsorption processes are not dominant. These processes can also take place in hydroxide phases, intimately attached to clay minerals in the raw materials.
The PZSE of the high-grade Mnt smectitic clay minerals group (HC Mnt NFe, HC Mnt and LC Mnt) approached zero charge (<±0.03 mmol proton charge) at pH values in the range 5–10 (Figure 5), with 8.2–8.4 (Table 2) being the crossing values of the different ion strength acid–base titration media. These data are consistent with those found by [49] for Na-montmorillonites, or those cited in [12] for Ca Mnt. HC Mnt NFe, HC Mnt and beidellite, all high-charge smectites, exhibit a lower surface proton buffering than LC Mnt. Otherwise, all of them have a PZSE very close to the PZC of alumina or aluminol groups in montmorillonite [50], a behaviour that was predicted in modelling studies [51] or observed in other acid–base titration experiments [52]. The high-charge smectites may have a strong electrostatic attraction between layers, leading to very stable face-to-face particle arrangements in aqueous suspensions [53]. Their acid–base behaviour in suspension is restricted to edge surfaces with very low proton interlayer interactions. In the case of the Mnt LC bentonite, the existence of less stable stacking of interlayers with more swelling ability, even at high electrolyte concentrations of up to 0.5 M of NaCl [54], can favour proton adsorption in some exposed free interlayers. In our case, proton consumption increased preferentially at high ion strength. This is not expected for Mnt, which should increase both PZC pH and proton adsorption in diluted media due to proton exchange in the permanent negative charge sites [18]. In summary, titration curves for all the studied bentonites are typical for materials with non-permanent charged external hydroxylated edge surface groups, where electrical double layers may become thin enough to enhance proton sorption as electrolyte concentration increases.
The pH evolution as a function of NaCl concentration for stevensite and sepiolite is shown in Figure 5. The stevensitic material shows PZSE at pH 8.2, although in this case the charge balance is very close to 0 from pH 8–9, possibly indicating that PZSE has a slightly higher pH corresponding to the contribution of Mg–OH groups. The NaCl concentration has no influence, and there is no shifting between the curves. The presence of dolomite can contribute to proton buffering. However, the slow kinetics in dolomite dissolution, depending on crystal size, (2–4 orders of magnitude at pH 2–5, regarding calcite), may limit the reactivity of this mineral [56]. In the case of the illite material (Figure 6), a similar 5% content of dolomite does not have a significant effect on proton buffering compared with the stevensitic material. The ability of stevensitic clay minerals to buffer pH, once it has been adjusted multiple times over several hours to pH < 4 (dolomite was dissolved), is confirmed in Figure 3. The behaviour of stevensite contrasted with sepiolite, with PZSE at pH 9.8, in accordance with the data provided by [57] at pH values from 9.8–10.4. Sepiolite exhibits a relatively high capacity for proton sorption due to its high external specific surface.
PZSE values for ferrihydrite and goethite were 8.0 and 8.3, respectively (Figure 6). This result is in agreement with the results achieved for the synthesized ferrihydrite in a CO2-free atmosphere, found in the range of pH 7–9 [58]. A very similar value was obtained for goethite at pH 8.3, which again is in agreement with the results obtained by [47] in the pH interval 8.2–8.6. The PZSE for illite was 8.2, similar to that obtained for di-octahedral Mnt by [50]. The splitting of the acid–base titration curves with increasing proton adsorption at low pH follows the expected behaviour of a surface potential when a DDL (Diffuse Double Layer) is developed without an internal permanent negative charge. In contrast, no electrical double layer effect is developed in calcined sewage sludge, although a relatively high potential for neutralizing protons was found. Taking into account that there were no free alkali hydroxides in the sample, siliceous matrix dissolution may favour the acid neutralization capacity of these surfaces. PZSE was obtained within the pH interval 5–8.5 for the calcined sewage sludge. This material was not found in previous studies since their physicochemical characteristics are highly heterogeneous and dependent on the sludge source.

3.3. Interaction of Tetracycline with Mineral Sorbents

3.3.1. Adsorption Capacity

The mineral sorbent materials have been compared in terms of their sorption efficiency by means of determining their sorption capacity (mg·g−1) of TC, which is shown in Figure 7.
The pH parameter does not show a significant effect in the adsorption rate of TC on the different sorbents under study (p = 0.179) when factorial ANOVA analysis is done. Some correlation is detected between pH and HC Mnt NFe sorbent concentration, but the pH can be included as a dependent variable. Thus, one-way ANOVA analysis is conducted by entering the sorbent as a single factor for each of the concentrations under study.
The LC Mnt sorbent shows the highest TC sorption capacity for all the tested concentrations. Thus, LC Mnt sorbed a mean of 13.2, 51.2, 125.5, 242 and 381.2 mg TC·g−1 when interacted with initial TC concentrations of 33.2, 131.6, 332.6, 625.0 and 1176.5 mg·L−1, respectively. As the initial concentrations increase, the HC Mnt NFe sorption efficiency increases, reaching the highest adsorption concentration (342.14 mg·g−1) as the LC Mnt (381.24 mg∙g−1) at 1176.5 mg∙L−1 TC concentration. On the other hand, sorbents such as Fh, Goet and Sewage S. showed the less capacity of TC sorption for all the concentrations under study showing significant differences at 322.6; 625.0 and 1176.5 mg TC∙L−1 with respect to the rest of sorbents. In the case of illite, its sorption capacity decreases as the initial TC concentration increases, reaching the lowest sorbed concentrations (32.29 and 41.02 mg TC∙L−1) at TC doses of 625.0 and 1176.5 mg∙L−1, respectively.
The materials with higher sorption capacity, smectites and sepiolite, increase their relative differences compared to the materials with lower adsorption capacity (iron hdroxydes, illite and calcined sewage sludge) when the TC concentration increases from 132 mg·L−1 to 1176 mg·L−1. At the lowest TC concentration (33 mg L−1), the specific surface area and acid–base properties play a significant role in adsorption. High surface area iron hydroxide (ferrihydrite) and illite have adsorption capacities of 8–10 mg TC/g, comparable to smectites, at pH 8, near their PZSE and relatively low adsorption rates at low pH, where TC and the mineral surface both have a positive charge. Conversely, other materials such as the smectite materials with PZSE at pH > 8.2, in general, do not exhibit systematic sorption differences as a function of pH. The exception is HC Mnt NFe, an ideal non-Fe high-charge montmorillonite, which shows a drastic decrease in TC adsorption at pH = 8 for all TC concentrations. This indicates that clay minerals may adsorb TC specifically (permanent interlayer negative charge) when the cationic dimethylammonium functional group remains with positive charge at pH < 8. When the TC concentration is increased to 1176 mg·L−1, the relative adsorption capacity for the materials with variable layer charge and low CEC, Fe-hydroxides and calcined sewage sludge decrease compared to smectites and sepiolite.
Smectites exhibit, in general, a high TC sorption capacity [12,19], which has no significant dependence on pH. The hydrated interlayer region of 2:1 sheet silicates has the potential to shelter TC independently of its net charge: either cationic or anionic poles of TC molecule could host solvating interlayer cations or be tied to hydroxyl groups at the interlayer by hydrogen bonds or other polar interactions [13,59]. This is in agreement with LC Mnt performance, which exhibit a similar adsorption capacity independently of pH, indicating that a complex interaction with TC is taking place, probably due to a delocalized bonding interaction between the TC functional ionic groups and the interlayer cations.
It is worth mentioning that sepiolite maintains a high sorption capacity in the whole experimental range of concentrations, at comparable sorption efficiencies of HC smectites, beidellite or stevensite. At a high concentration, no significant pH dependence is observed, although it tends to have a maximum adsorption at pH 8–6, where the TCH2± zwitterion species predominates compared to the slight predominance of TCH3+ in acidic conditions. PZSE of sepiolite maintain positive proton adsorption until pH 9–10, which can favour the adsorption of the anionic pole of TC, also taking into account the high specific surface area for interaction. In any case, the LC Mnt exhibit the best sorption capacity in the whole pH range and the optimum efficiency compared to the other studied materials.

3.3.2. Adsorption Isotherms

Adsorption isotherms, as a function of pH and TC concentration, are represented for each sorbent material, using the Freundlich and Langmuir equations as fitting models (Figure 8 and Figure 9; Table 3 and Table 4). The quadratic correlation values (R2) in general indicate good fitting to the experimental results. The parameter n of the Freundlich model was ~≥1, except for goethite at pH 4–6, meaning that the adsorption in most of the mineral sorbents is favourable. For the whole pH interval, the highest n value is obtained for the illite and smectite group, with HC Mnt NFe having the lowest value of n for this group.
HC Mnt NFe, goethite and sepiolite at acid pH show linear isotherms (Figure 8 and Figure 9), with Freundlich n values approaching to 1, and Cm in the Langmuir model is very high regarding their CEC. This means that these materials could be modelled with Linear partition Kd models and that they do not show strong interaction with TC [36]. The Cm parameter in the Langmuir model is the maximum adsorption capacity and should be directly related to the CEC of materials, in the case cation exchange interactions play a major role in TC adsorption. With the exception of the HC Mnt NFe, the smectite group and illite present Cm values within the same order of magnitude of their CEC. However, they are not fully correlated because LC Mnt exhibits the maximum sorption capacity (90–120 cmol(+)/kg) compared to high-charge Mnt with higher CEC. In the case of stevensite or illite, with a relatively low CEC, the Cm values are consistent with those shown in Table 2 and Table 4. The case of LC Mnt again shows that TC sorption is not solely a matter of cation solvation or cation exchange. Low-charge montmorillonite can allocate more TC because 2:1 sheets are less tightly stacked than high-charge smectites. Then a more effective internal layer surface can be made available [53].
In the case of sepiolite, TC adsorption is not related to CEC, and hence is better explained by linear or Freundlich models, in agreement with the strong physical sorption of the molecule at the external surface. Sepiolite and palygorskite are fibrous clay minerals of high surface area with a high sorption capacity for polar and non-polar organic compounds. Palygorskite has been proven to adsorb higher amounts of TC at a basic pH (9) than in acidic media (<6) [15], which is in agreement with the studied sepiolite.
As can be observed for both models, the quadratic correlation coefficient R2 of each material is high (R2 > 0.900), with the exception of the high-charge montmorillonite (R2 = 0.83) and sewage sludge (R2 = 0.74), for the Freundlich and Langmuir models, respectively. Because of this, the R2 coefficient cannot be used as a unique discrimination parameter to choose which model has better fitting to the experimental isotherms obtained in this work. Hence, discriminant analysis was performed to classify the different mineral sorbents in accordance with adsorption isotherm parameters.
The classification of different materials according to the model constants after the model isotherm fitting is shown in Figure 10. Graphs have been displayed for pH values, comparing both constants (n and K) for the Freundlich and Langmuir models. A sorbent material is considered to be suitable if the adsorption intensity, defined by Freundlich parameter n is higher than 1. Likewise, when n is between 2 and 10, it is considered to have a high intensity [42]. In addition to n, the material needs to have a high adsorption capacity, defined by the parameter K. Adsorption capacity is better defined in the case of the Langmuir model. The criteria for choosing the material with the best qualities will be based on the highest maximum adsorption (highest Cm) and high adsorption intensity (K > 1) for this model. Values of Cm > 120 and low Langmuir K (<1) correspond to linear isotherms or to unrealistic values for maximum capacity of smectites. Table 5 shows that LC Mnt is a very good material for TC sorption in the entire pH range for high TC concentrations, followed by stevensite at neutral to basic pH, which exhibits the highest intensity of adsorption and a good adsorption capacity. It is interesting to note that, for more realistic low TC concentrations and at neutral to basic pH conditions, calcined sewage sludge, illite, sepiolite, and ferrihydrite, in this order, exhibit sufficiently good adsorption properties.
Conglomerate analysis for Freundlich and Langmuir model parameters grouped the materials in terms of adsorption intensity and adsorption capacity, respectively. Groups 1 and 2 (Figure 10) associate iron hydroxides, sepiolite and HC Mnt NFe, which had anomalous Cm values. This parameter is a factor that separates HC Mnt NFe out of the smectitic materials. Group 3 in both models discriminates illite as a singular material for its high intensity of adsorption interaction. Group 4 includes calcined sewage sludge with smectites with high intensity of adsorption interaction, with the smectites being set apart in Langmuir group 5 for their high adsorption capacity (LC Mnt and beidellite). Other clay minerals, such as sepiolite and illite, show low capacity but strong intensity of adsorption, as well as calcined sludge. Their strong interaction properties may become interesting for the adsorption of pharmaceuticals in low load regimes.
The association of materials into different groups, using isotherms’ model parameters, reveal that these are not simple relationships considering the other different physical–chemical properties determined as SSA (BET or crystal size derived) or PZSE. At a low TC concentration, sepiolite, stevensite or ferrihydrite, with the highest BET, exhibit good sorption capacity independent of pH. At high TC loads the planar structures of smectites hold more TC, independent of SSA. High-charge smectites adsorb less TC, then, as has been stated; their performance is not solely a matter of cation exchange and charge density. The zero charge (salt effect) of these non-purified materials ranges from a nearly neutral pH (6–7) to 8–9. In a near-neutral pH environment zwitterionic species of TC can use ions (cation and anions) to establish bridge bonds with the available charged surface, following the concepts developed by [13]. Ideal montmorillonite with high charge and low iron substitution failed to retain TC at a slightly basic pH, which suggests that the chemical composition can play a significant role. Finally, high-charge smectites, including tetrahedral charged beidellite, do not produce differentiated performance in agreement with the acid–base properties [49], which indicates that the charge density and the tendency to maintain sTable 2:1 layer stacks are critical factors for TC sorption.

4. Conclusions

The study confirms the capacity of smectitic materials to entrap high quantities of tetracycline. Low-charge montmorillonite displays a more favourable sorption potential, both in terms of interaction intensity and concentration range of tetracycline, independent of pH. Therefore, wastewater treatments for TC should be based on this kind of mineral material.
Low-charge smectite or stevensite has the highest tetracycline adsorption intensity/capacity in a broad pH range from 4 to 8, and at low TC concentrations, which includes typical wastewater characteristics. The relative pH unspecific adsorption capacity agrees with the significant role of other non-ionic (solvation) or electrostatic type interactions rather than purely cation exchange processes in these types of minerals. At a basic pH, high-charge montmorillonite decreases its adsorption capacity due to the stronger electrostatic cation interaction between the clay mineral layers.
Those materials that have displayed a low capacity for sorption at high tetracycline concentrations should not be discarded in its use, because of the low concentration levels of these pollutants in the environment. Specifically, materials such as illite, sepiolite and ferrihydrite display high adsorption intensities at low to medium concentrations. Presumably, these materials are suitable to design pre-filters to implement permeable reactive barriers or to be tested with other emergent pollutants. Further efforts are being carried out in order to test these materials in a real wastewater treatment plant in order to promote the use of these easy available, eco-efficient materials.

Supplementary Materials

The following are available online at https://www.mdpi.com/2075-163X/9/7/453/s1, supplementary material S1: Detailed description of crystal-chemical related properties of clays.

Author Contributions

E.E.: conceptualization, supervision. J.C., N.D., F.Y.: methodology, F.Y.: software, validation, formal analysis, A.I.R., R.F.: investigation; resources; data curation, N.D., J.C.: writing—original draft preparation and conceptualization, C.G.D., D.E.G.S.: writing—review and editing.

Funding

This work has been economically supported by the Ministry of Economy and Competitiveness of Spain (CTM2013-47874-C2-2-R and AGL2016-78490-R).

Acknowledgments

In this section you can acknowledge any support given which is not covered by the author contribution or funding sections. This may include administrative and technical support, or donations in kind (e.g., materials used for experiments).

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. European Medicines Agency, E.S.V.A.C.E. Sales of Veterinary Antimicrobial Agents in 30 European Countries in 2016. (EMA/275982/2018); European Medicines Agency: Amsterdam, The Netherlands, 2018; p. 184. [Google Scholar]
  2. Stockwell, V.O.; Duffy, B. Use of antibiotics in plant agriculture. Revue Scientifique et Technique-Office International des Epizooties 2012, 31, 199–210. [Google Scholar] [CrossRef]
  3. Chee-Sanford, J.C.; Mackie, R.I.; Koike, S.; Krapac, I.G.; Lin, Y.F.; Yannarell, A.C.; Maxwell, S.; Aminov, R.I. Fate and transport of antibiotic residues and antibiotic resistance genes following land application of manure waste. J. Environ. Qual. 2009, 38, 1086–1108. [Google Scholar] [CrossRef] [PubMed]
  4. Osorio, V.; Perez, S.; Ginebreda, A.; Barcelo, D. Pharmaceuticals on a sewage impacted section of a Mediterranean River (Llobregat River, NE Spain) and their relationship with hydrological conditions. Environ. Sci. Pollut. Res. Int. 2012, 19, 1013–1025. [Google Scholar] [CrossRef] [PubMed]
  5. Kim, S.; Eichhorn, P.; Jensen, J.N.; Weber, A.S.; Aga, D.S. Removal of antibiotics in wastewater: Effect of hydraulic and solid retention times on the fate of tetracycline in the activated sludge process. Environ. Sci. Technol. 2005, 39, 5816–5823. [Google Scholar] [CrossRef] [PubMed]
  6. Harris, S.; Cormican, M.; Cummins, E. The effect of conventional wastewater treatment on the levels of antimicrobial-resistant bacteria in effluent: A meta-analysis of current studies. Environ. Geochem. Health 2012, 34, 749–762. [Google Scholar] [CrossRef] [PubMed]
  7. Michael, I.; Rizzo, L.; McArdell, C.S.; Manaia, C.M.; Merlin, C.; Schwartz, T.; Dagot, C.; Fatta-Kassinos, D. Urban wastewater treatment plants as hotspots for the release of antibiotics in the environment: A review. Water Res. 2013, 47, 957–995. [Google Scholar] [CrossRef] [PubMed]
  8. Li, Z.; Schulz, L.; Ackley, C.; Fenske, N. Adsorption of tetracycline on kaolinite with pH-dependent surface charges. Journal of colloid and interface science 2010, 351, 254–260. [Google Scholar] [CrossRef]
  9. Chang, P.-H.; Li, Z.; Jean, J.-S.; Jiang, W.-T.; Wang, C.-J.; Lin, K.-H. Adsorption of tetracycline on 2:1 layered non-swelling clay mineral illite. Appl. Clay Sci. 2012, 67–68, 158–163. [Google Scholar] [CrossRef]
  10. Antón-Herrero, R.; García-Delgado, C.; Alonso-Izquierdo, M.; García-Rodríguez, G.; Cuevas, J.; Eymar, E. Comparative adsorption of tetracyclines on biochars and stevensite: Looking for the most effective adsorbent. Appl. Clay Sci. 2018, 160, 162–172. [Google Scholar] [CrossRef]
  11. Sithole, B.B.; Guy, R.D. Models For Tetracycline In Aquatic Environments. l. lnteraction with Bentonite Clay Systems. Water Air Soil Pollut. 1987, 32, 303–314. [Google Scholar] [CrossRef]
  12. Chang, P.-H.; Li, Z.; Jiang, W.-T.; Kuo, C.-Y.; Jean, J.-S. Adsorption of tetracycline on montmorillonite: Influence of solution pH, temperature, and ionic strength. Desalinat. Water Treat. 2014, 1–13. [Google Scholar] [CrossRef]
  13. Aristilde, L.; Lanson, B.; Miehe-Brendle, J.; Marichal, C.; Charlet, L. Enhanced interlayer trapping of a tetracycline antibiotic within montmorillonite layers in the presence of Ca and Mg. J. Colloid Interface Sci. 2016, 464, 153–159. [Google Scholar] [CrossRef] [PubMed]
  14. Wang, Y.-J.; Jia, D.-A.; Sun, R.-J.; Zhu, H.-W.; Zhou, D.M. Adsorption and Cosorption of Tetracycline and Copper(II) on Montmorillonite as Affected by Solution pH. Environ. Sci. Technol. 2008, 42, 3254–3259. [Google Scholar] [CrossRef] [PubMed]
  15. Chang, P.H.; Li, Z.; Yu, T.L.; Munkhbayer, S.; Kuo, T.H.; Hung, Y.C.; Jean, J.S.; Lin, K.H. Sorptive removal of tetracycline from water by palygorskite. J. Hazard. Mater. 2009, 165, 148–155. [Google Scholar] [CrossRef] [PubMed]
  16. Borghi, A.A.; Palma, M.S.A. Tetracycline: Production, waste treatment and environmental impact assessment. Brazilian J. Pharmaceut. Sci. 2014, 50, 25–40. [Google Scholar] [CrossRef]
  17. Ertem, G.; Steudel, A.; Emmerich, K.; Lagaly, G.; Schuhmann, R. Correlation between the extent of catalytic activity and charge density of montmorillonites. Astrobiology 2010, 10, 743–749. [Google Scholar] [CrossRef] [PubMed]
  18. Kraepiel, A.M.L.; Keller, K.; Morel, F.M.M. On the Acid-Base Chemistry of Permanently Charged Minerals. Environ. Sci. Technol. 1998, 32, 2829–2838. [Google Scholar] [CrossRef]
  19. Chang, P.-H.; Li, Z.; Jiang, W.-T.; Jean, J.-S. Adsorption and intercalation of tetracycline by swelling clay minerals. Appl. Clay Sci. 2009, 46, 27–36. [Google Scholar] [CrossRef]
  20. Bandura, L.; Woszuk, A.; Kołodyńska, D.; Franus, W. Application of Mineral Sorbents for Removal of Petroleum Substances: A Review. Minerals 2017, 7, 37. [Google Scholar] [CrossRef]
  21. Fernández, R.; Ruiz, A.I.; Cuevas, J. The role of smectite composition on the hyperalkaline alteration of bentonite. Appl. Clay Sci. 2014, 95, 83–94. [Google Scholar] [CrossRef]
  22. Ruiz, A.I.; Fernandez, R.; Sanchez Jimenez, N.; Rodriguez Rastrero, M.; Regadio, M.; de Soto, I.S.; Cuevas, J. Improvement of attenuation functions of a clayey sandstone for landfill leachate containment by bentonite addition. Sci. Total Environ. 2012, 419, 81–89. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  23. Cuevas, J.; Leguey, S.; Garralon, A.; Rastrero, M.; Procopio, J.; Sevilla, M.; Jimenez, N.; Abad, R.; Garrido, A. Behavior of kaolinite and illite-based clays as landfill barriers. Appl. Clay Sci. 2008. [Google Scholar] [CrossRef]
  24. Cuevas, J.; Fernández, R.; Ortega, A.; Ruiz, A.I. Comparison of Alternative Bentonites for Potential use as Buffer and Sealing Materials in the Swiss Concept for Radioactive Waste Disposal. Part 2: Results; Nagra ed. CH-5430: Wettingen, Switzerland, 2014; p. 90. [Google Scholar]
  25. Cuevas, J.; De La Villa, R.V.; Ramirez, S.; Petit, S.; Meunier, A.; Leguey, S. Chemistry of Mg smectites in lacustrine sediments from the vicalvaro sepiolite deposit, madrid neogene basin (Spain). Clays Clay Miner. 2003, 51, 457–472. [Google Scholar] [CrossRef]
  26. Álvarez, A.; Santarén, J.; Esteban-Cubillo, A.; Aparicio, P. Current industrial applications of palygorskite and sepiolite. Dev. Clay Sci. 2011, 3, 281–298. [Google Scholar] [CrossRef]
  27. Guven, N. Bentonites—Clays for molecular engineering. Elements 2009, 5, 89–92. [Google Scholar] [CrossRef]
  28. Tosca, N.J.; Wright, V.P. Diagenetic pathways linked to labile Mg-clays in lacustrine carbonate reservoirs: A model for the origin of secondary porosity in the Cretaceous pre-salt Barra Velha Formation, offshore Brazil. Geolog. Soc. Lond. Spec. Publicat. 2018, 435, 33–46. [Google Scholar] [CrossRef]
  29. Benhammou, A.; Yaacoubi, A.; Nibou, L.; Tanouti, B. Adsorption of metal ions onto Moroccan stevensite: Kinetic and isotherm studies. J. Colloid Interface Sci. 2005, 282, 320–326. [Google Scholar] [CrossRef] [PubMed]
  30. Emmerich, K.; Wolters, F.; Kahr, G.; Lagaly, G. Clay profiling: The classification of montmorillonites. Clays Clay Miner. 2009, 57, 104–114. [Google Scholar] [CrossRef]
  31. Vázquez, M.; Jiménez-Millán, J. Clay raw materials from the Triassic Red Beds (Northern Jaén, Spain) for making ceramic construction materials. Materiales de Construcción 2004, 54, 5–20. [Google Scholar] [CrossRef]
  32. Yunta, F.; Di Foggia, M.; Bellido-Diaz, V.; Morales-Calderon, M.; Tessarin, P.; Lopez-Rayo, S.; Tinti, A.; Kovacs, K.; Klencsar, Z.; Fodor, F.; et al. Blood meal-based compound. Good choice as iron fertilizer for organic farming. J. Agricult. Food Chem. 2013, 61, 3995–4003. [Google Scholar] [CrossRef]
  33. Sánchez, J.M.; Murillo, J.M.; Otero, J.; Plaza, F.; Garralon, G. Thermal valorisation of dried sewage sludge through gasification. In Proceedings of the 10th International Conference on Environmental Science and Technology 2007, Kos Island, Grrece, 5–7 September 2007; pp. A1277–A1286. [Google Scholar]
  34. Parolo, M.E.; Savini, M.C.; Vallés, J.M.; Baschini, M.T.; Avena, M.J. Tetracycline adsorption on montmorillonite: pH and ionic strength effects. Appl. Clay Sci. 2008, 40, 179–186. [Google Scholar] [CrossRef]
  35. Guler, U.A.; Sarioglu, M. Removal of tetracycline from wastewater using pumice stone: Equilibrium, kinetic and thermodynamic studies. J. Environ. Health Sci. Eng. 2014, 12, 11. [Google Scholar] [CrossRef] [PubMed]
  36. Gong, W.; Liu, X.; He, H.; Wang, L.; Dai, G. Quantitatively modeling soil-water distribution coefficients of three antibiotics using soil physicochemical properties. Chemosphere 2012, 89, 825–831. [Google Scholar] [CrossRef] [PubMed]
  37. Avisar, D.; Primor, O.; Gozlan, I.; Mamane, H. Sorption of sulfonamides and tetracyclines to montmorillonite clay. Water Air Soil Pollut. 2009, 209, 439–450. [Google Scholar] [CrossRef]
  38. Ciarlone, A.E.; Fry, B.W.; Ziemer, D.M. Some observations on the adsorption of tetracyclines to glass and plastic labware. Microchem. J. 1990, 42, 250–255. [Google Scholar] [CrossRef]
  39. Wammer, K.H.; Slattery, M.T.; Stemig, A.M.; Ditty, J.L. Tetracycline photolysis in natural waters: Loss of antibacterial activity. Chemosphere 2011, 85, 1505–1510. [Google Scholar] [CrossRef] [PubMed]
  40. Bolster, C.H. Microsoft Excel Spreadsheets for Fitting Sorption Data; USDA-ARS: Madison, WI, USA, 2010. [Google Scholar]
  41. Sposito, G. The Chemistry of Soils, 2nd ed.; Oxford University Press, Inc.: New York, NY, USA, 2008; p. 342. [Google Scholar]
  42. Masel, R.I. Principles of Adsorption and Reaction on Solid Surfaces; John Wiley & Sons, Inc.: New York, NY, USA, 1996; p. 793. [Google Scholar]
  43. Brindley, G.W.; Brown, G. Crystal Structures of Clay Minerals and Their X-ray Identification; Mineralogical Society: London, UK, 1980; Volume 5. [Google Scholar]
  44. Hibino, T.; Tsunashima, A.; Yamazaki, A.; Otsuka, R. Model calculation of sepiolite surface areas. Clays Clay Miner. 1995, 43, 391–396. [Google Scholar] [CrossRef]
  45. Jambor, J.L.; Dutrizac, J.E. Occurrence and constitution of natural and synthetic ferrihydrite, a widespread iron oxyhydroxide. Chem. Rev. 1998, 98, 2549–2585. [Google Scholar] [CrossRef]
  46. Dzombak, D.A.; Morel, F.M.M. Surface Complexation Modeling: Hydrous Ferric Oxide; John Wiley & Sons: New York, NY, USA, 1990; p. 416. [Google Scholar]
  47. Kosmulski, M. pH-dependent surface charging and points of zero charge. IV. Update and new approach. J. Colloid Interface Sci. 2009, 337, 439–448. [Google Scholar] [CrossRef]
  48. Antelo, J.; Avena, M.; Fiol, S.; Lopez, R.; Arce, F. Effects of pH and ionic strength on the adsorption of phosphate and arsenate at the goethite-water interface. J. Colloid Interface Sci. 2005, 285, 476–486. [Google Scholar] [CrossRef]
  49. Kriaa, A.; Hamdi, N.; Srasra, E. Acid–Base Chemistry of montmorillonitic and beidellitic-montmorillonitic smectite. Russian J. Electrochem. 2007, 43, 167–177. [Google Scholar] [CrossRef]
  50. Tombácz, E.; Szekeres, M. Colloidal behavior of aqueous montmorillonite suspensions: The specific role of pH in the presence of indifferent electrolytes. Appl. Clay Sci. 2004, 27, 75–94. [Google Scholar] [CrossRef]
  51. Tournassat, C.; Davis, J.; Chiaberge, C.; Grangeon, S.; Bourg, U.C. Modeling the Acid-Base Properties of Montmorillonite Edge Surfaces. Environ. Sci. Technol. 2016, 50, 13436–13445. [Google Scholar] [CrossRef] [PubMed]
  52. Arfaoui, S.; Hamdi, N.; Frini-Srasra, N.; Srasra, E. Determination of point of zero charge of PILCS with single and mixed oxide pillars prepared from Tunisian-smectite. Geochem. Int. 2012, 50, 447–454. [Google Scholar] [CrossRef]
  53. Christidis, G.E.; Blum, A.E.; Eberl, D.D. Influence of layer charge and charge distribution of smectites on the flow behaviour and swelling of bentonites. Appl. Clay Sci. 2006, 34, 125–138. [Google Scholar] [CrossRef]
  54. Wilson, M.J.; Wilson, L.; Patey, I. The influence of individual clay minerals on formation damage of reservoir sandstones: A critical review with some new insights. Clay Miner. 2014, 49, 147–164. [Google Scholar] [CrossRef]
  55. González-Santamaría, D.E.; López, E.; Ruiz, A.I.; Fernández, R.; Ortega, A.; Cuevas, J. Adsorption of phenantrene by stevensite and sepiolite. Clay Miner. 2017. In press. [Google Scholar]
  56. Pokrovsky, O.S.; Golubev, S.V.; Schott, J. Dissolution kinetics of calcite, dolomite and magnesite at 25 °C and 0 to 50 atm pCO2. Chem. Geol. 2005, 217, 239–255. [Google Scholar] [CrossRef]
  57. Santos, S.C.R.; Boaventura, R.A.R. Adsorption of cationic and anionic azo dyes on sepiolite clay: Equilibrium and kinetic studies in batch mode. J. Environ. Chem. Eng. 2016, 4, 1473–1483. [Google Scholar] [CrossRef]
  58. Villalobos, M.; Antelo, J. A unified surface structural model for ferrihydrite: Proton charge, electrolyte binding, and arsenate adsorption. Revista Internacional de Contaminación Ambiental 2011, 27, 139–151. [Google Scholar]
  59. MacEwan, D.M.C.; Wilson, M.J. Interlayer and intercalation complexes of clay minerals. In Crystal Structures of Clay minerals and their X-Ray Identification; Brindley, G.W., Brown, G., Eds.; Mineralogical Society: London, UK, 1980; Volume 5, pp. 197–248. [Google Scholar]
Minerals 09 00453 g001 550
Figure 1. Structure of the TC molecule and proton dissociation constants for acid–base active groups, after [16,19]. Image Source: PubChem, URL: https://pubchem.ncbi.nlm.nih.gov. Data deposited in or computed by PubChem.
Figure 1. Structure of the TC molecule and proton dissociation constants for acid–base active groups, after [16,19]. Image Source: PubChem, URL: https://pubchem.ncbi.nlm.nih.gov. Data deposited in or computed by PubChem.
Minerals 09 00453 g001
Minerals 09 00453 g002 550
Figure 2. X-ray diffraction (XRD) powder patterns from different mineral sorbent materials. The main reflections identifying minerals and used in crystal size calculations (csc) are quoted and marked in Å units. (hkl) The Miller index is incorporated for csc peak reflections. Sep: sepiolite, Sme: smectite, Ilt: illite; Gth: Goethite; qtz: quartz, sheet sil: sheet silicates; Dol: dolomite. Materials: Sepiolite: Pansil®; Stevensite: Minclear 100®; HC Mnt: FEBEX high-charge montmorillonite; HC Mnt NFe: Mnt-Chile High-charge montmorillonite low Fe content; Goet: Synthetic Goethite; Fh: Synthetic Ferrihydite; Sewage S.: Calcined Sewage Sludge.
Figure 2. X-ray diffraction (XRD) powder patterns from different mineral sorbent materials. The main reflections identifying minerals and used in crystal size calculations (csc) are quoted and marked in Å units. (hkl) The Miller index is incorporated for csc peak reflections. Sep: sepiolite, Sme: smectite, Ilt: illite; Gth: Goethite; qtz: quartz, sheet sil: sheet silicates; Dol: dolomite. Materials: Sepiolite: Pansil®; Stevensite: Minclear 100®; HC Mnt: FEBEX high-charge montmorillonite; HC Mnt NFe: Mnt-Chile High-charge montmorillonite low Fe content; Goet: Synthetic Goethite; Fh: Synthetic Ferrihydite; Sewage S.: Calcined Sewage Sludge.
Minerals 09 00453 g002
Minerals 09 00453 g003 550
Figure 3. Evolution of pH as a function of time in clay mineral suspensions (1 g/40 cm3) in 0.1 M NaCl, 0.5 g/L TC after adjusting the pH to a stable value of 3.5. The pH data were recorded every 10 s for 12 h, under continuous stirring. HC Mnt NFe: High-charge, low Iron content Mnt.
Figure 3. Evolution of pH as a function of time in clay mineral suspensions (1 g/40 cm3) in 0.1 M NaCl, 0.5 g/L TC after adjusting the pH to a stable value of 3.5. The pH data were recorded every 10 s for 12 h, under continuous stirring. HC Mnt NFe: High-charge, low Iron content Mnt.
Minerals 09 00453 g003
Minerals 09 00453 g004 550
Figure 4. Specific surface area of the different geosorbents comparing Brunauer, Emmett and Teller (BET) method nitrogen adsorption measurements and calculated surface area by crystal size XRD determination based on crystallographic symmetry and XRD reflection breadth measurements. Linear fit excludes ferrihydrite (Fh). Materials: Sepiolite: Pansil®; Stevensite: Minclear 100®; HC Mnt: FEBEX high-charge montmorillonite; HC Mnt NFe: Mnt-Chile High-charge montmorillonite, low Fe content; Goet: Synthetic Goethite; Fh: Synthetic Ferrihydite; Sewage S.: Calcined Sewage Sludge.
Figure 4. Specific surface area of the different geosorbents comparing Brunauer, Emmett and Teller (BET) method nitrogen adsorption measurements and calculated surface area by crystal size XRD determination based on crystallographic symmetry and XRD reflection breadth measurements. Linear fit excludes ferrihydrite (Fh). Materials: Sepiolite: Pansil®; Stevensite: Minclear 100®; HC Mnt: FEBEX high-charge montmorillonite; HC Mnt NFe: Mnt-Chile High-charge montmorillonite, low Fe content; Goet: Synthetic Goethite; Fh: Synthetic Ferrihydite; Sewage S.: Calcined Sewage Sludge.
Minerals 09 00453 g004
Minerals 09 00453 g005 550
Figure 5. Experimental acid–base titration of sepiolite and smectitic materials for the point zero salt effect (PZSE) method obtained at various ionic strengths at 25 °C.
Figure 5. Experimental acid–base titration of sepiolite and smectitic materials for the point zero salt effect (PZSE) method obtained at various ionic strengths at 25 °C.
Minerals 09 00453 g005
Minerals 09 00453 g006 550
Figure 6. Experimental acid–base titration curves for iron oxyhydroxides, calcined sludge and illite obtained by the PZSE method at three ionic strengths and 25 °C.
Figure 6. Experimental acid–base titration curves for iron oxyhydroxides, calcined sludge and illite obtained by the PZSE method at three ionic strengths and 25 °C.
Minerals 09 00453 g006
Minerals 09 00453 g007 550
Figure 7. Sorption capacity of tetracycline (TC) (mg∙g−1) on tested mineral sorbents; Ferrihydrite (Fh), goethite (Goet), illite, calcined sewage sludge (Sewage S.), sepiolite, stevensite, non-iron high-charge montmorillonite (HC Mnt NFe), high-charge montmorillonite (HC Mnt), low-charge montmorillonite (LC Mnt), beidellite at those pH levels of 4 (white), 6 (grey) and 8 (black) and at those tested TC concentrations of (I) 33.2; (II) 131.6; (III) 322.6; (IV) 625.0 and (V) 1176.5 mg TC·L−1.
Figure 7. Sorption capacity of tetracycline (TC) (mg∙g−1) on tested mineral sorbents; Ferrihydrite (Fh), goethite (Goet), illite, calcined sewage sludge (Sewage S.), sepiolite, stevensite, non-iron high-charge montmorillonite (HC Mnt NFe), high-charge montmorillonite (HC Mnt), low-charge montmorillonite (LC Mnt), beidellite at those pH levels of 4 (white), 6 (grey) and 8 (black) and at those tested TC concentrations of (I) 33.2; (II) 131.6; (III) 322.6; (IV) 625.0 and (V) 1176.5 mg TC·L−1.
Minerals 09 00453 g007
Minerals 09 00453 g008 550
Figure 8. pH dependence of TC adsorption isotherms for sepiolite and smectitic materials at 25 °C in 0.1 M of NaCl aqueous media. pH = 4 in black; pH = 6 in blue; pH = 8 in red. Solid lines correspond to Langmuir fit. The Freundlich model fitting is plotted as dots. Freundlich and Langmuir models are inserted for comparison. Fitting parameters are listed in Table 3 (Freundlich) and Table 4 (Langmuir).
Figure 8. pH dependence of TC adsorption isotherms for sepiolite and smectitic materials at 25 °C in 0.1 M of NaCl aqueous media. pH = 4 in black; pH = 6 in blue; pH = 8 in red. Solid lines correspond to Langmuir fit. The Freundlich model fitting is plotted as dots. Freundlich and Langmuir models are inserted for comparison. Fitting parameters are listed in Table 3 (Freundlich) and Table 4 (Langmuir).
Minerals 09 00453 g008
Minerals 09 00453 g009 550
Figure 9. pH dependence of TC adsorption isotherms for iron hydroxides, calcined sewage sludge and illite at 25 °C in 0.1 M of NaCl aqueous media. pH = 4 in black; pH = 6 in blue; pH = 8 in red. Solid lines correspond to Langmuir fit. The Freundlich model fitting is plotted as dots. Freundlich and Langmuir models are inserted for comparison. Fitting parameters are listed in Table 3 (Freundlich) and Table 4 (Langmuir).
Figure 9. pH dependence of TC adsorption isotherms for iron hydroxides, calcined sewage sludge and illite at 25 °C in 0.1 M of NaCl aqueous media. pH = 4 in black; pH = 6 in blue; pH = 8 in red. Solid lines correspond to Langmuir fit. The Freundlich model fitting is plotted as dots. Freundlich and Langmuir models are inserted for comparison. Fitting parameters are listed in Table 3 (Freundlich) and Table 4 (Langmuir).
Minerals 09 00453 g009
Minerals 09 00453 g010 550
Figure 10. Material classification according to the statistical analysis of Freundlich and Langmuir isotherm parameters (K and n) as a function of pH interaction. Conglomerates dispersion diagram determined by the Euclidean quadratic K-average method implemented in STATGRAPHICS® v.16.1 software. (n_4, n_6, n_8 = n value for each pH; Cm_4, Cm_6, Cm_8 = Cm; K_4, K_6, K_8 = K value for each pH value). Shaded areas mark the optimum values defined for adsorption intensity and capacity for the fitted parameters (see text and Table 4).
Figure 10. Material classification according to the statistical analysis of Freundlich and Langmuir isotherm parameters (K and n) as a function of pH interaction. Conglomerates dispersion diagram determined by the Euclidean quadratic K-average method implemented in STATGRAPHICS® v.16.1 software. (n_4, n_6, n_8 = n value for each pH; Cm_4, Cm_6, Cm_8 = Cm; K_4, K_6, K_8 = K value for each pH value). Shaded areas mark the optimum values defined for adsorption intensity and capacity for the fitted parameters (see text and Table 4).
Minerals 09 00453 g010
Table
Table 1. Crystallographic and physical properties of the adsorbents used in this study.
Table 1. Crystallographic and physical properties of the adsorbents used in this study.
MaterialCell Parameters (Å)Cell Volume/DensityCrystal Thickness% Clay MineralBET Surface AreaCalculated Surface Area
(hkl) *abc(Å3)g/cm3c (001) ** 2:1 (Å)b (060) ** 2:1 (Å)wt. %m2/gm2/g
Ferrihydrite (100)5.085.089.42103.9613 100253561
Goethite (100)4.5969.9573.0211384.29157 1005988
Illite5.209.0210.04692.80251197703964
Beidellite5.208.9714.56762.80191217606150
Mnt HC NFe5.209.0213.06102.80112175904793
Mnt LC5.208.9912.55842.80107280903458
Mnt HC5.209.0014.56792.8054249955468
Stevensite5.309.1514.57032.60199490223186
Sepiolite (110)5.2826.9513.419022.40141 95293274
* (hkl) Miller indices of the corresponding XRD reflections used to calculate the crystal thickness. **: Crystal thicknesses of 2:1 sheet silicates were determined in two directions (b and c) for sheet silicates. Materials: Sepiolite: Pansil®; Stevensite: Minclear 100®; HC Mnt: FEBEX high-charge montmorillonite; HC Mnt NFe: Mnt-Chile High-charge montmorillonite, low Fe content; Goet: Synthetic Goethite; Fh: Synthetic Ferrihydite; Calcined Sewage Sludge is not considered for BET surface area calculation as it is not a specific mineral with crystal structure. Its BET surface area is 16 m2/g.
Table
Table 2. Physicochemical characteristics of sorbent materials used in this study.
Table 2. Physicochemical characteristics of sorbent materials used in this study.
MaterialCEC (cmol(+)·kg−1)PZSE (pH)
Illite (Dolomite < 5%)13(1)8.2 **
Beidellite61(2)8,3 **
Stevensite (Dolomite < 5%)60(3)8.2 **
Sepiolite20(3)9.8 **
LC Mnt 83(4)8.3 **
HC Mnt 102(4)8.4 **
HC Mnt NFe96(4)8.3 **
Sewage S.N.D.5–8.5 **
FhN.D.8.0 **
GoetN.D.8.3 **
CEC: cation exchange capacity. (1): [23](2): [22]; (3): [55]; (4): [21]. (**): Determined in this work. Materials: Sepiolite: Pansil®; Stevensite: Minclear 100®; HC Mnt: FEBEX high-charge montmorillonite; HC Mnt NFe: Mnt-Chile High-charge montmorillonite, low Fe content; Goet: Synthetic Goethite; Fh: Synthetic Ferrihydite; Sewage S.: Calcined Sewage Sludge. N.D.: Not Determined.
Table
Table 3. Freundlich isotherm fitting parameters.
Table 3. Freundlich isotherm fitting parameters.
Mineral SorbentKf (L·g−1)nR2
pH 4pH 6pH 8pH 4pH 6pH 8pH 4pH 6pH 8
Stevensite5.7953.531.751.753.672.700.9670.9280.970
Sepiolite0.503.031.121.121.371.390.9970.9950.997
LC Mnt73.365.092.362.362.682.820.9640.9530.971
HC Mnt22.025.082.452.452.442.220.8330.8300.835
HC Mnt N Fe5.902.411.241.241.051.110.9650.9830.889
Beidellite12.110.481.881.881.752.060.9780.9900.963
Goet0.010.0590.720.720.940.290.9980.9980.970
Fh0.070.620.940.941.331.440.9980.9950.992
Sewage S.1.557.561.541.543.464.390.9660.8600.956
Illite5.0413.683.443.444.432.540.9570.9290.987
Quadratic correlation values (R2).
Table
Table 4. Langmuir isotherm fitting parameters.
Table 4. Langmuir isotherm fitting parameters.
Mineral SorbentK (L·μmol−1) × 10−3* Cm (cmol·kg−1)CEC (cmol·kg−1)R2
pH 4pH 6pH 8pH 4pH 6pH 8 pH 4pH 6pH 8
Stevensite2.0013.04.0060.739.142.6600.9710.9790.965
Sepiolite0.100.405.00392160125200.9970.9930.998
LC Mnt1.209.909.2011687.689.8830.9970.9990.998
HC Mnt3.503.602.3052.058.662.11020.9510.9450.901
HC Mnt N Fe1.000.200.20251759156960.9780.9860.905
Beidellite1.901.602.7076.289.774.8610.9960.9960.999
Goet0.000.000.1015367.842.1n.d.0.9700.9980.982
Fh0.100.101.0026195.561.6n.d.0.9940.9980.991
Sewage S.0.605.5012.738.67.202.72n.d.0.9890.9830.742
Illite4.6011.92.304.967.4716.8130.9750.9860.932
* Recalculated to be compared to CEC (cation exchange capacity). Quadratic correlation values (R2). Values in bold are remarked concerning they are very similar in the model prediction compared to determined CEC.
Table
Table 5. Classification of materials for their adsorption intensity and capacity (see text).
Table 5. Classification of materials for their adsorption intensity and capacity (see text).
Intensity/CapacityOptimum Properties (1)High Intensity at Low TC Concentrations
Low pHNeutral pH (6–8)
Freundlich n > 2; K > 10; LC Mnt HC MntStevensite, LC Mnt, HC Mnt Illite, Sewage S. (pH 6–8)
Langmuir Cm(10–120 cmol/kg) (2);
Klangmuir > 1		MX-80, Beidellite, HC Mnt, StevensiteLC Mnt, Beidellite, HC Mnt, StevensiteSepiolite, Fh pH (8)
(1) materials are listed according to “best” criteria; i.e., Freundlich: higher n followed by higher K. (2) >120 cmol/kg is considered unrealistically high values not fitted to Langmuir.

Share and Cite

MDPI and ACS Style

Cuevas, J.; Dirocie, N.; Yunta, F.; García Delgado, C.; González Santamaría, D.E.; Ruiz, A.I.; Fernández, R.; Eymar, E. Evaluation of the Sorption Potential of Mineral Materials Using Tetracycline as a Model Pollutant. Minerals 2019, 9, 453. https://doi.org/10.3390/min9070453

AMA Style

Cuevas J, Dirocie N, Yunta F, García Delgado C, González Santamaría DE, Ruiz AI, Fernández R, Eymar E. Evaluation of the Sorption Potential of Mineral Materials Using Tetracycline as a Model Pollutant. Minerals. 2019; 9(7):453. https://doi.org/10.3390/min9070453

Chicago/Turabian Style

Cuevas, Jaime, Nisael Dirocie, Felipe Yunta, Carlos García Delgado, Daniel E. González Santamaría, Ana Isabel Ruiz, Raúl Fernández, and Enrique Eymar. 2019. "Evaluation of the Sorption Potential of Mineral Materials Using Tetracycline as a Model Pollutant" Minerals 9, no. 7: 453. https://doi.org/10.3390/min9070453

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop