Introduction

Water contamination due to organic dyes and radionuclides is a challenging issue for the modern world. Different industries such as pharmaceutical, food, paper and leather industries use large quantities of dyes. Particularly, textile industries have largely increased dyes in their effluents in terms of both types and volume (Robinson et al. 2001). According to an estimation, about 10,000 tonnes of dyes of different varieties are produced every year in the world (Forgacs et al. 2004). About 15% of these dyes flow into the industrial effluents, which cause turbidity and reduce the penetration of sun rays deep into the water (Houas et al. 2001). This affects aquatic plants by decreasing the rate of photosynthesis in them. Consequently, the deoxygenation of water disturbs life activities in fish and other aquatic animals. Furthermore, organic dyes can also enter the food chain and distress human health as well as local flora and fauna (Khurana et al. 2018).

Apart from dyes, radionuclides such as U(VI) is a serious cause of environmental pollution. Nuclear technology plays an important role in agriculture, power and defence system of a country. However, nuclear power plants, nuclear weapon test sites, uranium ore processing and weathering of rocks are important sources of uranium in the environment (Child and Hotchkis 2013). 235U and 238U are the natural radioactive forms of uranium with a half-life of 7.0 × 108 and 4.4 × 109 years, respectively. Therefore, they exist for a long period of time in the environment and eventually accumulate in the plants and animals and cause toxic health problems in them. In animals and humans, they are absorbed in the gastrointestinal tract via drinking water or food and cause kidney failure in them (Bayramoglu et al. 2015; Konietzka 2015). They enter the plants via the root system and accumulate in their different parts (Dushenkov et al. 1997). Health problems due to organic dyes and uranium emphasize their removal from contaminated water and have become the focus of worldwide research topics.

Among the many methods used for the removal of unwanted species from wastewater, the adsorption technique is easy, cost-effective, simple to operate and eco-friendly (Ayub et al. 2017; Khan et al. 2018). A variety of adsorbents such as activated carbons (de Luna et al. 2013) and double-layer hydroxide composites (Blaisi et al. 2018) have been applied for the removal of EBT from aqueous solutions. Similarly, coal fly ashes (Chen et al. 2017) and layered titanate nanowires (Yin et al. 2017) have been utilized for the removal of U(VI) from water solutions. However, these adsorbents had no prominent bonding sites for EBT and U(VI) ions. On the other hand, EBT and U(VI) could form strong complexes with magnesium and calcium ions. Moreover, the carbon-loaded adsorbents have remarkable adsorption capabilities for U(VI) (Liao et al. 2019). Therefore, we proposed to use rubbish tissue paper–derived carbon-doped MgO and CaCO3 for the elimination of EBT and U(VI) from aqueous environment.

The morphologies and physical properties of MgO and CaCO3 are dependent on their way of preparation. Calcination of Mg(OH)2 in air at 450°C results in a blur mass of MgO powders (Moussavi and Mahmoudi 2009). The sol-gel combustion of Mg(NO3)2 with urea, citric acid and oxalic acid at 350°C results in irregular cubes of MgO (Nassar et al. 2017). Similarly, the morphology of CaCO3 depends on the reaction temperature (Seo et al. 2005). CaCO3 obtained by the simultaneous mixing of CaCl2 with NH4HCO3 (1:1 ratio) at 30°C had mushroom-like lamellar structures with a diameter of 20–35 μm (Chen and Xiang 2009). The diameter of the particles decreased to 10–20 μm when the temperature was increased to 50°C. The morphology changed to rod shape at 60°C. Similarly, rod-shaped crystals with a constant diameter of 1.0–5.0 μm and a length of 15–40 μm were formed at 80°C. Recently, we synthesized them by using a different approach. Firstly, we achieved amorphous carbon by the wet oxidation of rubbish tissue papers. Then it was used as reducing as well as doping agent in the in situ hydrothermal treatment of Mg(OH)2 and Ca(OH)2 to obtain C-MgO and C-CaCO3. The C-MgO and C-CaCO3 achieved in this way had unique morphologies. Here carbon played the following significant roles: (I) At high temperature and pressure, carbon reacts with water and in a series of reactions might produce certain gases such as water gas and carbon dioxide gas, which could react with Mg(OH)2 and Ca(OH)2 to form MgO and CaCO3; (II) Doping with carbon reduces the sticky and gelatinous nature of MgO and CaCO3 in aqueous solutions and gives them unique morphologies; (III) Doping with carbon changes the pHpzc values of CaCO3 and MgO and consequently changes their adsorption properties. Both naked CaCO3 and MgO are suitable for the adsorption of anions only (Hu et al. 2010; Smiciklas et al. 2008), whereas C-CaCO3 and C-MgO are found to be suitable for the adsorption of both anions and cations.

Although numerous studies on U(VI) and EBT sorption are reported in the literature, the interaction mechanisms of EBT and U(VI) with C-MgO or C-CaCO3 have not been investigated theoretically. This study performs density functional theory (DFT) calculations to elucidate molecular-level interactions between C-MgO/C-CaCO3 and EBT/U(VI). Such interactions are difficult to ascertain experimentally. However, computational chemistry provides a great help in the study of molecular-level interactions and provides an easy, quick and cheap way to explain the interaction mechanism between pollutants and adsorbents.

The purpose of this study is to introduce a new and economic way for the synthesis of C-MgO and C-CaCO3 and their comparative adsorption capacities for the removal of EBT and U(VI) from wastewater. It was also intended to reuse the rubbish tissue papers for more valuable purposes in the synthesis and doping of MgO and CaCO3. The adsorption capabilities of C-MgO and C-CaCO3 for EBT and U(VI) have been optimized under different experimental conditions. The structures, energies, electronic properties and adsorption behaviour of EBT and U(VI) onto C-MgO and C-CaCO3 are also elucidated by DFT simulations.

Materials and methods

Synthesis of carbon

The reagents utilized in the batch sorption study of EBT/U(VI) onto C-MgO/C-CaCO3 were obtained from Sinopharm Chemical Regent Co., Ltd. (Beijing, China). Carbon was synthesized by the wet oxidation of rubbish tissue papers. In a typical reaction, 5.60 mL of conc. H2SO4 was slowly added to 5.0 g of the used tissue papers in a 200-mL beaker. The ingredients were continuously mixed with a wooden stick, and finally, the black powder was washed to the pH close to Milli-Q water. After drying at 110°C for 5.0 h, the product was grinded to a fine powder and stored in a plastic bag for further use.

Synthesis of C-MgO and C-CaCO3

For this purpose, MgCl2.6H2O (1.051 g, 5.17 mmol) and NaOH (0.414 g, 10.34 mmol) were separately added to 30 mL Milli-Q water. Then the MgCl2.6H2O solution was added dropwise to NaOH solution under vigorous stirring. After 30 min stirring at 40°C, the reaction mixture along with 0.5 g carbon was transferred to a Teflon-lined stainless-steel autoclave. The reaction mixture was subjected to 140°C for 6.0 h. The obtained C-MgO was washed three times with Milli-Q water and dried up at 110°C for 5.0 h. The same process, using CaCl2 instead of MgCl2, was repeated for the synthesis of C-CaCO3. All the possible reactions involved in the conversion of Mg(OH)2 and Ca(OH)2 to C-MgO and C-CaCO3 are presented in SI.

Characterization

The morphologies of the synthesized materials were characterized by scanning electron microscopy (S-2500, Japan) and transmission electron microscopy (TEM, Hitachi-7650). The BET surface area was measured by QUADRASORBSI, Quantachrome USA. The powder X-ray diffraction (XRD) pattern was performed with Rigaku-D/max2200 diffractometer in the 2θ range of 5–70° and a scanning rate of 0.05°/s. Fourier transform infrared (FT-IR) spectra were obtained by Nicolet Magana-IR 750 spectrometer. The thermogravimetric analysis (TGA) was performed with a Shimadzu TGA Q5000 thermal analyser in the range of 50–900°C and a heating rate of 10°C min−1. The ZETASIZER 3000 HSA system was used for measuring the zeta potential values of the adsorbents. The XPS analysis was performed by Thermo Escalab 250 electron spectrometer operating at 150 W Al Kα radiation.

Batch sorption experiments

Batch sorption experiments for EBT and U(VI) were performed at different experimental conditions. Suitable amounts of EBT (2.17 × 10−5 to 5.85 × 10−4 mol/L) and U(VI) (2.54 × 10−5 to 2.79 × 10−4 mol/L) along with adsorbent (0.05 to 0.20 g/L) were transferred to a series of 10-mL polyethylene test tubes and diluted to the required concentrations. The adsorption was optimized by performing pH experiments in the range of 2.0–10.0. Small volumes of 0.1 mol/L HCl or NaOH were used to acquire the desired pH. The effect of adsorbent amount was accomplished in the range of 0.05–0.80 g/L. The kinetic experiments were conducted in the time range of 10 min to 13.0 h. The adsorption equilibria were achieved at 25, 40 and 55°C. The samples were gently shaken at 150 rpm in a thermostatic shaker in the dark to avoid photo-degradation of the pollutant species. After the equilibration time, the samples were filtered through Nylon-66 syringe filters (0.22 μm) to separate the adsorbent from the suspension. The concentration of U(VI) in the supernatant was determined with arsenazo-III spectrophotometric technique at 650 nm while that of EBT was recorded at 520 nm. The adsorption percentage was determined as adsorption (%) = (Co− Ce)/Co × 100 while the equilibrium concentration was calculated as, qe = (Co− Ce) × V/m, where Co and Ce (mol/L), qe (mol/g), V (L) and m (g) are the initial concentration, equilibrium concentration, adsorption amount at equilibrium, volume of the suspension and weight of the adsorbent, respectively.

Results and discussion

Characterization

The microstructures and surface characteristics of carbon, C-MgO and C-CaCO3 were accomplished by SEM and TEM images. The SEM image (Fig. 1a) exposes amorphous morphology of granular carbon. The C-MgO had rod-shaped morphology (Fig. 1b), whereas C-CaCO3 appeared as plate-like structures with well-defined edges (Fig. 1c). The TEM image of carbon (Fig. 1e) discloses the homogenously distributed granules. The rod-shaped C-MgO nanoparticles with lattice fringes of 1.70 Å were 63.31 nm long (Fig. S1a) and 7.71 nm wide (Fig. S1b) and were covered with carbon particles (Fig. 1f). The interplanar spacing of C-CaCO3 (calcite) nanoplatelets was 2.77 Å (Fig. 1g) with an average diameter and a particle area of 53.23 nm and 1.89 × 10−6 m2 (Fig. S1c and d), respectively. Moreover, the high-resolution transmission electron microscopy (HRTEM) images of the boxed area in Fig. 1 f and g confirmed the single-crystalline nature of C-MgO and C-CaCO3 materials. The energy dispersive X-ray microanalysis (EDXMA, Fig. 1d, h) demonstrates the presence of carbon, magnesium and oxygen in C-MgO and carbon, calcium and oxygen in C-CaCO3. Calculated from the EDXMA analysis, the concentrations of the surface carbon in C-MgO and C-CaCO3 were 10.86 and 31.38 at%, respectively. The powder XRD patterns of the synthesized materials are shown in Fig. 2a. In the XRD pattern of carbon, the broad peak at 2θ = 21.67o indicates its amorphous nature (Rajan et al. 2014). The face-centred cubic-structured MgO nanorods with lattice parameter, a = 4.21 Å, had distinctive planes (111), (200), (220), (311) and (222) at 2θ = 37.12°, 43.04°, 62.26°, 74.68° and 79.08°, respectively (JCPDS PDF No. 00-001-1235) (Devi et al. 2012). The crystallite size of C-MgO calculated from the most intense peak (200) was 4.87 nm, whereas it was 5.0 nm for monocrystalline MgO achieved by sol-gel method (Goryczka et al. 2010). The major reflections for calcite, (104) and (110) appeared at 2θ = 29.28° and 36.04°, respectively (JCPDS PDF No. 01-086-2340) (Peng et al. 2013). The lattice parameter (a) of C-CaCO3 (calcite) was 9.76 Å with crystallite size of 18.38 nm. The sharp peaks in C-MgO and C-CaCO3 XRD patterns indicated their ultra-fine nature. Furthermore, the XRD pattern evidenced that doping with carbon had no effect on the crystal structures and brought no structural deformations to MgO and CaCO3.

Fig. 1
figure 1

SEM images of a carbon, b C-MgO and c C-CaCO3. d Elemental mapping and EDX spectrum of C-MgO. TEM images of e carbon, f C-MgO and g C-CaCO3. h Elemental mapping and EDX spectrum of C-CaCO3

Full size image
Fig. 2
figure 2

a XRD patterns, b FT-IR spectra, c N2 adsorption-desorption isotherms and d TGA curves of the synthesized materials

Full size image

In the FT-IR analysis of all the materials (Fig. 2b), the strong peak at 3434 cm−1 was due to the physically attached water molecules. In the FT-IR spectrum of C-MgO, a small peak at 1155 cm−1 corresponded to H− ions bonded to Mg2+ at different coordination cites while the characteristic Mg-O stretching vibrations appeared at 430 cm−1 (Mageshwari et al. 2013). In the case of CaCO3, a strong peak at 1404 cm−1 along with a shoulder peak (1510 cm−1) was due to CO32− ions (Wang et al. 2017). The other FT-IR characteristic peaks for CaCO3 were found at 3441, 1461, 1037, 871 and 770 cm−1, which were in good agreement with the reported result (Saraya and Rokbaa 2016).

The Brunauer-Emmett-Teller (BET) surface areas of pure carbon, C-MgO and C-CaCO3 were found to be 24, 54 and 29 m2/g (Fig. 2c) with the average pore diameters (Barrett-Joyner-Halenda (BJH)) of 3.7, 5.1 and 18.6 nm and total pore volumes of 0.091, 0.134 and 0.727 cm3/g, respectively (Fig. S2). Larger pore diameters represented the mesoporous nature of the materials and were useful for the strong uptake of U(VI) and EBT. The pore diameters of both C-MgO and C-CaCO3 were very large compared with the molecular diameter of EBT (1.55 nm) (Loera-Serna et al. 2017) and the ionic radius of U(VI) (0.086 nm) (Ohmichi et al. 1981). Therefore, it could be easy for EBT and U(VI) to diffuse inside the adsorbent materials.

The TGA curve (Fig. 2d) shows that the oxidation of pure carbon completed at 550°C and entirely converted to CO2 gas (Oliveira et al. 2002). The physically adsorbed water molecules were detached (5% weight loss) from C-MgO at 200°C followed by the evolution of CO2 gas (55% weight loss) at 500°C, leaving behind a constant mass of MgO (40% weight) as solid residue. Till 500°C, the TGA thermograph of C-CaCO3 was similar to that of C-MgO. The moisture content was lost in the first jump. Carbon was removed in the second step to form CO2 gas (60% weight loss). In the third step, at 770°C, the remaining CaCO3 started to decompose into CaO (20% constant mass left behind) and CO2 gas (10% weight loss). The possible reactions that carbon, C-MgO and C-CaCO3 might have undergone during thermogravimetric analysis, are presented in SI.

pHpzc is an important characteristic property of an adsorbent and determines its adsorption mode towards a pollutant. The pHpzc of pure carbon is neutral (7.0) or slightly basic (7.9) (Hao et al. 2004; Kodama and Sekiguchi 2006). The pHpzc value of naked CaCO3 lies in the range of 8.0–9.5 (Smiciklas et al. 2008) while that of naked MgO is 12.4 and is favourable for the adsorption of anions only (Hu et al. 2010). Interestingly, the pHpzc of C-CaCO3 and C-MgO are 2.50 and 3.34, respectively, and allow the adsorption of both anions (such as EBT) and cations (such as UO22+).

XPS analysis was performed to characterize the surface states of the as-prepared materials. The survey spectra of C-MgO and C-CaCO3 (Fig. 3a) show distinct peaks at 49.86 eV (Mg 2p) and 347.42 eV (Ca 2p). The Mg 2p peak in C-MgO spectrum at 49.4 eV was assigned to magnesium-bonded hydroxyl groups (Mg-OH), while the peak at 50.5 eV was assigned to magnesium-bonded oxygen atoms (Mg-O, Fig. 3b) (Li et al. 2014). The C 1s peak of each material was deconvoluted into several Gaussian peaks. The carbon peaks were found in binding energy range from 281.0–292.0 eV (Fig. 3c). In the C 1s spectrum of MgO (111) surface, four different peaks at 284.6, 286.2, 288.0 and 290.3 eV, were assigned to C-C, C-O, C=O and O-C=O, respectively (Zou et al. 2017). The Mg 1s peak at 1303 eV was supplemented by Auger peaks between 300 and 390 eV (encircled area in Fig. 3a which is enlarged in Fig. 3e). The Auger peaks have large chemical shifts and are useful for chemical state analysis. The O 1s XPS spectrum of C-MgO (Fig. 3d) was deconvoluted into two components at 530.1 eV (surface oxide O2−) and 532.1 eV (surface OH−) (Cui et al. 2018; Yao et al. 2000). The peak at higher binding energy was due to chemisorbed water. As shown in the survey spectrum, the Ca 2p and Mg (KLL) peaks can be easily misunderstood, because Ca 2p appears in the range same to that of Mg Auger peaks. However, the characteristic spin-orbit splitting of the suspected Ca 2p peak (Ca 2p3/2 and Ca 2p1/2 components) splits by 3.4 eV with an intensity ratio of 2:1. The presence of Ca 2p3/2 at 347.42 eV and Ca 2p1/2 at 350.03 eV (Chen et al. 2018), in the high-resolution XPS spectrum of Ca 2p (Fig. 3f), confirmed the existence of CaCO3 in the sample. C 1s spectrum of C-CaCO3 (Fig. 3g) was deconvoluted into three Gaussian peaks at 284.6 eV (C-C), 285.6 eV (C-O) and 289.2 eV (CO32−), with full width at half maximum (FWHM) values of 1.39, 2.45 and 2.43, respectively (Chu et al. 2013). The C-CaCO3 O 1s XPS spectrum (Fig. 3h) was deconvoluted into Ca-O (531.1 eV) and C-O (533.3 eV), with FWHM values of 1.76 and 2.07, respectively (Ni and Ratner 2008). This suggested that some Ca atoms on the surface of CaCO3 existed in the oxide form. The percentage of different functional groups present in C-CaCO3 and C-MgO, calculated from C 1s and O 1s peaks are shown in Fig. S3. The percent concentrations of C 1s deconvoluted peak areas of each functional group present in C-CaCO3 were 40.70% (C-C), 20.94% (CO32−) and 38.36% (C-O), and those of O 1s were 73.87% (Ca-O) and 26.13% (C-O) (Fig. S3a). The percent concentrations of C 1s deconvoluted peak areas assigned to each functional group in C-MgO were 67.71% (C-C), 9.93% (C-O), 11.55% (C=O), and 10.82% (O-C=O), and those of O 1s were 36.16% (O2−) and 63.34% (OH−) (Fig. S3b). High concentration of the surface bonded OH− groups can prominently enhance the adsorption ability of C-MgO towards EBT and U(VI). The relative percent distribution of elements in both adsorbents is given in Table S1. The carbon content of C-MgO (10.36%) was less than that of C-CaCO3 (31.38%). However, it was homogeneously distributed in C-MgO (Fig. 1d), while aggregated in C-CaCO3 (Fig. 1h). The homogeneous distribution of carbon could be beneficial for the high surface area and adsorption capacity of C-MgO. The atomic ratios and binding energies of all elements in C-MgO and C-CaCO3 are given in Table 1. Moreover, various metal oxides and their peak positions are tabulated in Table S2.

Fig. 3
figure 3

a XPS survey scan spectra of adsorbent materials, b XPS spectra of Mg 2p, c C 1s, d O 1s and e Mg (KLL) and Auger peaks of C-MgO. XPS spectra of f Ca 2p, g C 1s and h O 1s of C-CaCO3. e is the enlarged photo of the encircled area in a

Full size image
Table 1 XPS surface elemental analysis of C-MgO and C-CaCO3
Full size table

Effect of solid content

The effect of solid content was studied to achieve maximum adsorption of pollutants with minimum possible amount of adsorbent. Figure 4 a and b elucidate the effect of C-CaCO3 and C-MgO dose on U(VI) sorption while their effect on EBT sorption is given in Fig. S4a and b. The experiments showed that the adsorption of EBT and U(VI) increased significantly with increasing the adsorbent amount from 0.05 to 0.20 g/L. Figure 4 a and b and Fig. S4a and b further illustrate that the partition ratio (Kd) increased proportionally with the increase in the amount of adsorbent till it reached 0.20 g/L and then decreased with further increase of the adsorbent amount. The maximum adsorption of both EBT and U(VI) was observed at 0.20 g/L of C-CaCO3 and C-MgO.

Fig. 4
figure 4

Effect of the adsorbent dose on the adsorption and partition ratio of U(VI) onto a C-CaCO3 and b C-MgO, at Co[U(VI)] = 7.61 × 10−5 mol/L, pH = 6.0, C[NaCl] = 0.01 mol/L and T = 25 °C. Effect of pH on the speciation and adsorption of c EBT and d U(VI) at Co[EBT] = 8.67 × 10−5 mol/L, Co[U(VI)] = 7.61 × 10−5 mol/L, m/V = 0.20 g/L, T = 25°C (C[NaCl] = 0.01 mol/L for U(VI) only). e Determination of zeta potential as a function of pH in 0.01 mol/L NaCl solution. f Effect of ionic strength on U(VI) sorption at Co[U(VI)] = 7.61 × 10−5 mol/L, m/V = 0.20 g/L and T = 25°C. The experimental conditions are not applicable to the speciation of EBT

Full size image

Effect of solution pH

The pH of the solution has great impact on the adsorption properties and chemical state of an adsorbent. Similarly, it influences the distribution of organic and inorganic species in aqueous solutions. The effect of pH on the adsorption of EBT and U(VI), and the relative distribution of their species as a function of pH are shown in Fig. 4c and d. A rapid decrease in the adsorption of EBT onto C-CaCO3 was observed when the pH was increased from 3.0 to 5.0 (Fig. 4c). This was because the surface of C-CaCO3 was negatively charged at pH higher than its pHpzc (2.50, Fig. 4e) which might lead to strong repulsions with EBT anions. Beyond pH = 3.0, the increase in negative charge on both adsorbents as well as EBT intensified the electrostatic repulsions between them. Consequently, the adsorption of EBT decreased continuously throughout the whole pH range. On the other hand, the adsorption of U(VI) followed a totally different route. Unlike EBT, the adsorption of U(VI) was low at low pH. However, it increased rapidly with increasing pH and reached to the peak value at pH = 6.0, followed by a gradual decrease. The relative distribution of U(VI) species (Fig. 4d) shows that at pH = 6.0, UO22+, (UO2)2(OH)22+ and (UO2)3(OH)5+ ions were dominant species in the solution. At low pH, the competition between H+, Na+ and UO22+ ions suppressed the sorption of UO22+ ions (Wang et al. 2016). The high adsorption efficiency at pH = 6.0 was due to the strong surface complexation and electrostatic interactions between the positively charged U(VI) species and the negatively charged C-CaCO3 and C-MgO surfaces. At pH = 9.0–10.0, the electrostatic repulsion between U(VI) anions, i.e., (UO2)3(OH)7−, UO2(OH)3− and (UO2)(OH)42− and the negatively charged adsorbents resulted in low adsorption (Ding et al. 2015). Figure 4f shows that the adsorption of U(VI) decreased only by 0.6% when the NaCl concentration increased from 0.001 to 0.03 mol/L. A small effect of ionic strength suggested the formation of inner-sphere surface complexes of U(VI) with both adsorbents (Hu et al. 2015). However, the gradual increase in the adsorption after 0.03 mol/L NaCl was due to the salting-out effect of NaCl (Chen et al. 2017).

Adsorption kinetics

Adsorption kinetics was helpful in explaining the basic steps such as mass transport and chemical reaction involved in the adsorption process. The pseudo-first-order model (PFO, Eq. (1)) (Vilela et al. 2018) and pseudo-second-order model (PSO, Eq. (2)) (Balasubramanian et al. 2019) were applied to investigate the kinetic data and are expressed as:

$$ {q}_{\mathrm{t}}={q}_{\mathrm{e}}\left(1-{e}^{-{k}_1t}\right) $$
(1)
$$ {q}_{\mathrm{t}}=\frac{q_{\mathrm{e}}^2{k}_2t}{1+{q}_{\mathrm{e}}{k}_2t} $$
(2)

where qe and qt (mol/g) referred to the equilibrium and instantaneous adsorption of pollutants and k1 (1/h) and k2 (g/mol·h) are the PFO and PSO rate constants. Due to the availability of more binding sites on the adsorbent surface, the initial adsorption rate of EBT onto C-MgO and C-CaCO3 was rapid and reached equilibrium in 2.5 and 3.3 h, respectively (Fig. 5a). Similarly, the equilibrium time for U(VI) onto C-MgO and C-CaCO3 was 1.6 and 3.0 h, respectively (Fig. 5b). The respective values of qe, k1 and k2 are listed in Table S3. It shows that the EBT and U(VI) sorption on C-MgO was faster than C-CaCO3, i.e., kC-MgO > kC-CaCO3, which further confirmed that EBT and U(VI) sorption onto C-MgO was much efficient than C-CaCO3. The kinetic data agreed fairly well with the PSO simulation. The good matching of kinetic data to PSO model was further confirmed by the percent relative error factor (RE (%)) calculated from Eq. (3).

$$ \mathrm{RE}\left(\%\right)=\frac{\left|{q}_{\mathrm{e},\mathrm{calc}.}-{q}_{\mathrm{e},\exp .}\right|}{q_{\mathrm{e},\exp .}}\times 100 $$
(3)

where qe,calc. and qe,exp. are the calculated and experimental adsorption capacities, respectively. The values of RE (%) for PSO were lower than that of PFO (Table S3), confirming a better fit of the PSO model to the experimental data. The agreement of kinetic data with PSO model suggested that EBT and U(VI) sorption onto C-CaCO3 and C-MgO was a chemical process (Zhao et al. 2014; Zou et al. 2015) that might involve the exchange of electrons between the pollutant species and adsorbent (Khan et al. 2018). To further investigate the adsorption kinetics of EBT and U(VI) sorption, the initial adsorption rate (h (mol/g·h)) and half-equilibrium time (t1/2 (h)) were calculated from Eqs. (4) and (5), respectively, and the results are recorded in Table S3.

$$ h={k}_2{q}_{\mathrm{e}}^2 $$
(4)
$$ {t}_{1/2}=\frac{1}{k_2{q}_{\mathrm{e}}} $$
(5)
Fig. 5
figure 5

Time-dependent sorption curves of a EBT and b U(VI). Liquid film diffusion model fitting of c EBT and d U(VI). Intraparticle diffusion model for e EBT and f U(VI). Initial sorption model for g EBT and h U(VI), at Co[EBT] = 8.67 × 10−5 mol/L, Co[U(VI)] = 7.61 × 10−5 mol/L, m/V = 0.20 g/L and T = 25°C. (pH = 2.0 for EBT sorption while pH = 6.0 and C[NaCl] = 0.01 mol/L for U(VI) sorption)

Full size image

By changing the adsorbent from C-CaCO3 to C-MgO, the h values increased from 1.61 × 10−15 to 2.01 × 10−15 mol/g·h for EBT and from 5.59 × 10−15 to 11.3 × 10−15 mol/g h for U(VI) ions, while t1/2 values reduced from 9.0 × 1010 to 7.6 × 1010 h for EBT while this decrease for U(VI) was 3.0 × 1010 h to 2.0 × 1010 h.

Diffusion is a key phenomenon involved in the process of adsorption. The migration of the pollutant species from the bulk of the solution to the liquid film boundary of the adsorbent is referred to as the external surface adsorption. Besides this, the other processes involved in the adsorption are intraparticle diffusion and film diffusion.

The liquid diffusion model (Eq. (6)) was applied to study the film diffusion process.

$$ \ln \left(1-\frac{q_{\mathrm{t}}}{q_{\mathrm{e}}}\right)=-{k}_{\mathrm{fd}}t $$
(6)

Here, the qt/qe is the equilibrium sorption ratio and kfd (h−1) is the rate constant for sorption. The zero intercept of the straight line obtained by plotting -ln(1-qt/qe) vs. t represents that sorption kinetics is governed by the film diffusion process (Chen et al. 2017). However, for both EBT and U(VI) sorption these intercepts were greater than zero (Fig. 5c, d), which shows that the film diffusion was not the only process involved in the adsorption of U(VI) and EBT.

The kinetic data was also simulated by intraparticle diffusion model (IPD, Eq. (7)) (Weber and Morris 1963), to understand the role of diffusion in the adsorption process.

$$ {q}_t={k}_{\mathrm{p}i}{t}^{0.5}+{C}_{\mathrm{i}} $$
(7)

The values of kpi (mol/g·h1/2) and Ci (mol/g) along with R2, obtained from Eq. (7) are tabulated in Table S3. The IPD plots obtained by using piecewise regression of qt vs. t0.5 are shown in Fig. 5e, f. They indicate that the IPD plot of EBT was comprised of three regions while that of U(VI) consisted of two regions. The values of IPD model parameters (Table S3) reveal that the adsorption rate constants were in the order of kp1 > kp2 > kp3. High adsorption rate in the first region was due to the availability of a large number of free adsorption sites on the adsorbent surfaces. In the second region, the rate-determining step, the adsorption rate slowed down and became almost insignificant in the third stage, where approximately all the adsorption points were covered by the pollutant species. The adsorption process is controlled by IPD when the plot of qt vs. t0.5 passes through the origin. However, Fig. 5 e and f indicate that the IPD plots did not pass through the origin, which reflected that some interactions other than IPD were also involved in EBT and U(VI) sorption (Ahmed et al. 2019). The intercept (Ci) is related to the boundary layer effect (McKay et al. 1980) which increases directly with the increase in Ci. The Ci values (Table S3) were in the order C1 < C2 < C3, indicating that the boundary layer effect continued to increase from region-1 to region-3. The initial adsorption factor (Ri) was computed from the IPD model, which was helpful to further understand the importance of boundary layer diffusion (BLD) in the adsorption process (Wu et al. 2009). For the determination of Ri, Eq. (7) can be written as:

$$ {q}_{\mathrm{e}}={k}_{\mathrm{p}i}{t}_{\mathrm{e}}^{0.5}+{C}_i $$
(8)

where te (h) is the equilibrium time and qe (mol/g) is the amount of EBT/U(VI) adsorbed at t = te. Subtracting Eqs. (8) and (7) and on rearrangement gives Eq. (9).

$$ \left(\frac{q_{\mathrm{t}}}{q_{\mathrm{e}}}\right)=1-{R}_{\mathrm{i}}\left[1-{\left(\frac{t}{t_{\mathrm{e}}}\right)}^{0.5}\right] $$
(9)

where\( {R}_{\mathrm{i}}=\frac{k_{\mathrm{p}i}{t}_{\mathrm{e}}^{0.5}}{q_{\mathrm{e}}} \) is defined as the initial adsorption factor of the IPD model. Ri can be obtained from Eq. (8) as:

$$ {R}_{\mathrm{i}}=1-\frac{C_{\mathrm{i}}}{q_{\mathrm{e}}} $$
(10)

Ri = 1 for Ci = 0 means that there is no initial adsorption in the system and adsorption is governed merely by IPD. Ri = 0 for Ci = qe refers to no real adsorption and some other processes such as coagulation and aggregation may involve in the system. However, as the ratio of Ci/qe decreases the Ri increases and IPD becomes the prominent adsorption mechanism. The Ri values, classified into four zones, are given in Table S4. In zone 0 and 1, the adsorption is controlled by IPD throughout the process. The adsorption reaches to about 80% in zones 2 and 3. In zone 4, the initial adsorption occupies more than 90% of the adsorption. The initial adsorption characteristic curves based on the IPD model (Eq. (9)) are presented in Fig. 5 g and h. The Ri values for EBT adsorption onto C-CaCO3 and C-MgO were 0.67 and 0.84, respectively. According to Table S4, the adsorption of EBT comes into zone 2 and exhibits intermediate initial adsorption. Similarly, the Ri values for U(VI) sorption onto C-CaCO3 and C-MgO were 0.35 and 0.49, respectively, which fall into zone 3 and showed strong initial adsorption. One reason for low adsorption of EBT was its larger molecular size than U(VI) (Wu et al. 2009).

Sorption isotherms and thermodynamic studies

Isotherm study of U(VI) and EBT sorption was performed at different temperatures because it affects not only the state of pollutant anions and cations but also the sorption efficiency of adsorbents. EBT and U(VI) sorption data were simulated by Langmuir (Eq. (11)) (Yang and Fox 2018) and Freundlich (Eq. (12)) (Choi et al. 2019) isotherm models.

$$ {q}_{\mathrm{e}}=\frac{q_{\mathrm{m}}b{C}_{\mathrm{e}}}{1+{bC}_{\mathrm{e}}} $$
(11)
$$ {q}_{\mathrm{e}}={K}_{\mathrm{F}}{C_{\mathrm{e}}}^n $$
(12)

where qm (mol/g) is the monolayer maximum adsorption, Ce (mol/L) is the equilibrium concentration of pollutants in solution, b (L/mol), KF (mol1–n·Ln/g) and n are the constants of relative models.

Langmuir isotherm is useful to describe homogeneous adsorption. It was applied to calculate the monolayer maximum adsorption capacity of C-CaCO3 and C-MgO for EBT and U(VI) ions. Freundlich isotherm was applied to depict the heterogeneity of EBT and U(VI) sorption onto C-CaCO3 and C-MgO surfaces. Table 2 encloses the values of the respective parameters of the Langmuir and Freundlich models calculated at 25°C, while those calculated at 40 and 55°C are listed in Table S5. The values of n indicated that EBT and U(VI) sorption onto C-CaCO3 and C-MgO was a favourable process (Dada et al. 2012). However, the data did not converge well with the Freundlich model and the regression coefficient was lower as compared to the Langmuir isotherm model. The better fit of the Langmuir model was further confirmed by the relative error factor, calculated from Eq. (3). As shown in Table 2, the RE (%) values for Langmuir model were lower than those of the Freundlich model, indicating a better fit of the sorption data to the Langmuir model. Moreover, it was proved with an important characteristic of the Langmuir model namely the dimensionless constant separation factor (RL), that the adsorption of EBT and U(VI) onto C-CaCO3 and C-MgO was a favourable process (detail is given in SI).

Table 2 Langmuir and Freundlich model constants and correlation coefficients for the adsorption of EBT and U(VI) onto C-CaCO3 and C-MgO at 25°C
Full size table

The monolayer adsorption capacities of C-CaCO3 and C-MgO calculated at 25°C were 2.53 × 10−4 and 3.62 × 10−4 mol/g for EBT and 5.92 × 10−4 and 8.10 × 10−4 mol/g for U(VI) ions, respectively. Table 3 shows that C-CaCO3 and C-MgO has comparatively good adsorption abilities for EBT and U(VI) ions. The EBT and U(VI) sorption isotherms at 25, 40 and 55°C are presented in Fig. 6a–d. The rapid increase in the adsorption capacity with increasing temperature suggested that EBT and U(VI) sorption was more favourable at high temperature.

Table 3 Comparison of the adsorption capacities of different adsorbents for EBT and U(VI) calculated at 25°C
Full size table
Fig. 6
figure 6

Sorption curves of a EBT and b U(VI) onto C-CaCO3, sorption curves of c EBT and d U(VI) onto C-MgO at m/V = 0.20 g/L and contact time = 24 h, pH = 2.0 for EBT sorption while pH = 6.0 and C[NaCl] = 0.01 mol/L for U(VI) sorption, =25°C, =40°C, and =55°C, solid line: Langmuir model, dash line: Freundlich model. van’t Hoff plots for e EBT and f U(VI) sorption

Full size image

Thermodynamic study was helpful in explaining the feasibility of sorption reaction. The Van’t Hoff Eq. (13) was applied to calculate the important thermodynamic parameters, the standard entropy change (∆S°, J/mol·K) and standard enthalpy change (∆H°, kJ/mol).

$$ \ln \frac{C_{\mathrm{a}}}{C_{\mathrm{e}}}={lnK}_{\mathrm{c}}=\frac{{\Delta S}^{{}^{\circ}}}{R}-\frac{{\Delta H}^{{}^{\circ}}}{R}\times \frac{1}{T} $$
(13)

where Ca (mol/L), Kc, T (K) and R (8.314 J/mol·K) are the solid phase equilibrium concentration, equilibrium constant, thermodynamic temperature and universal gas constant, respectively. The plots of lnKc vs. 1/T (Fig. 6e, f) gave a straight line with intercept and slope equal to ∆S°/R and ∆H°/R, respectively. The change in standard Gibbs free energy of adsorption (∆G°, kJ/mol) was measured from Eq. (14):

$$ \Delta {G}^{\circ }=\Delta {H}^{\circ }-T\Delta {S}^{\circ } $$
(14)

Table 4 shows the values of thermodynamic parameters obtained from Eqs. (13) and (14). The negative values of ∆G° and positive values of ∆H° and ∆S° revealed that sorption of EBT and U(VI) onto C-CaCO3 and C-MgO was spontaneous, endothermic and random at solid-solution interface. The decrease in ∆G° values with increasing temperature proved that sorption became more favourable at high temperature due to the dehydration of the pollutant ions (Liu et al. 2015). The results obtained at 25, 40 and 55°C displayed that the adsorption of EBT and U(VI) on both adsorbents increased with the increase in temperature, which might be attributed to the tendency of the pollutant species to escape from the bulk phase to the solid phase.

Table 4 Thermodynamic parameters for EBT and U(VI) sorption on C-CaCO3 and C-MgO
Full size table

DFT calculations

The DFT theoretical calculations (Gonze et al. 2009) were important to explore the interaction mechanism of EBT and U(VI) ions with MgO and CaCO3 at the molecular level. The geometric optimization, sorption energy calculation and charge density analysis were performed by Vienna ab initio Simulation Package (VASP) (Kresse and Furthmüller 1996). The exchange-correction function was described by the Perdew-Burke-Ernzerhof (PBE) formulation of the generalized gradient approximation (GGA) (Perdew et al. 1996). Crystal structures and other DFT conditions are given in the SI.

Figure 7 represents the most stable geometries for EBT and U(VI) sorption onto MgO and CaCO3. As shown in Fig. 7 a and b, EBT molecule orients parallel to the adsorbent’s surface. The distance between EBT and CaCO3 (1.778 Å) was much shorter than that of EBT and MgO (4.354 Å). However, Mg-O(EBT) bond length (2.178 Å) was shorter than Ca-O(EBT) bonds, i.e., Ca1-O1 (2.713 Å), Ca1-O2 (2.457 Å), Ca2-O3 (2.474 Å) and Ca2-O4 (3.522 Å). Similarly, the U1-O1 distance in MgO system (2.099 Å) was shorter than that of CaCO3 system (2.271 Å, Fig. 7c, d). Therefore, from a geometric perspective, the electrostatic interactions of EBT and U(VI) with MgO were much stronger than CaCO3. The sorption energy (ΔE) was calculated from ΔE = Esub + EEBT/U − Ecomp, where Ecomp represents the total energy of the target complex system, Esub and EEBT/U correspond to the energies of the substrate (CaCO3 or MgO) and the isolated EBT or U(VI) molecules, respectively. High sorption energy of EBT-MgO (20.62 kcal/mol) than EBT-CaCO3 (10.21 kcal/mol), and that of U-MgO (63.41 kcal/mol) than U-CaCO3 (34.29 kcal/mol) indicated that the sorption ability of MgO for EBT and U(VI) was better than that of CaCO3.

Fig. 7
figure 7

Geometric structures of the complexes of a EBT-CaCO3, b EBT-MgO, c U-CaCO3 and d U-MgO. The charge density differences of e EBT-CaCO3 system, f EBT-MgO system, g U-CaCO3 system and h U-MgO system. Purple and green colours represent positive and negative charges, respectively, and all bond lengths and distances are given in Å

Full size image

In order to further clarify the adsorption mechanism, the differential charge density distributions of both adsorption systems were calculated as depicted in Fig. 7e–h. The three-dimensional charge density difference plot with an isosurface value of 1.0 × 10−5 e·Å3 was obtained by subtracting the calculated electronic charge of the individual parts. The green isosurface represented the regions where the electron density was increased while the purple one represented a reduction in the charge density. In case of EBT, the adsorption process induced the intra- and inter-molecular charge redistribution not only in the EBT molecules but also in the basal plane of CaCO3. While in the case of U(VI), the uranyl ions (UO22+) were positively charged and the charge transfer process took place only in the type of “atom to atom”. The results of charge density distribution showed that the electrostatic interactions in EBT-MgO and U-MgO were stronger than EBT-CaCO3 and U-CaCO3, which were consistent with the experimental results. Short bond distances, high sorption energies and increased charge density all support high sorption ability of MgO for both EBT and U(VI) as compared to CaCO3.

Reusability

Reusability is an important property of an adsorbent which reduces the cost of the adsorption process. Both C-CaCO3 and C-MgO were recycled five times to check their reusability for EBT and U(VI) sorption. After the adsorption of EBT, 1.0 g of C-CaCO3 and C-MgO was stirred in 15 mL NaOH (1.0 mol/L) and shaken at 25°C. After 24 h, the adsorbents were separated from the solution and washed with 15 mL Milli-Q water followed by 10 mL ethanol and finally dried at 110°C for 5 h. The dried mass was reused for the adsorption of EBT and the results are presented in Fig. 8. It indicates that in five cycles, the adsorption capacity of C-CaCO3 and C-MgO for EBT decreased by 15.4 and 11.2% respectively. The big decrease in adsorption capacity of C-CaCO3 was due to its strong complexation with EBT. Like EBT, the same process was repeated for the desorption of U(VI), however, instead of NaOH, 1.0 mol/L HCl was used as eluent. Figure 8 shows that in the first three cycles, the sorption capacities of C-CaCO3 and C-MgO for U(VI) decreased by 8.5 and 6.1% respectively, while in the fourth and fifth cycle, these decreased by 13.7 and 10.3% respectively. The reason for the big decrease in the adsorption ability in the fourth and fifth cycle would be due to the corrosion of the adsorbents in acidic medium and incomplete separation of the pollutant species. These results indicate that both adsorbents can be reused five times for the effective elimination of EBT and U(VI) from wastewater.

Fig. 8
figure 8

Recyclability of C-CaCO3 and C-MgO for the removal of EBT and U(VI) ions from water solutions at Co[EBT] = 8.67 × 10−5 mol/L, Co[U(VI)] = 7.61 × 10−5 mol/L, m/V = 0.20 g/L, contact time = 24 h and T = 25°C. (pH = 2.0 for EBT sorption, while pH = 6.0 and C[NaCl] = 0.01 mol/L for U(VI) sorption)

Full size image

Conclusion

In this study, EBT and U(VI) sorption onto C-CaCO3 and C-MgO was explored by batch technique and DFT theoretical calculations. The adsorption of EBT and U(VI) ions was pH dependent; however, ionic strength had little effect (0.6%) on the adsorption process. The adsorption amount of EBT and U(VI) both raised with the increasing amount of adsorbents and reached its maximum value at 0.20 g/L of adsorbent. This showed that satisfying adsorption results could be achieved with small amounts of C-CaCO3 and C-MgO. The adsorption data of EBT and U(VI) well agreed with PSO kinetic and Langmuir isotherm models. Thermodynamically, the adsorption reaction of EBT and U(VI) on the surfaces of C-CaCO3 and C-MgO was favourable. Results of the sorption data of both EBT and U(VI) were concordant with DFT calculations. It can be concluded from the easy preparation, strong interactions with EBT and U(VI) and reusability that both C-CaCO3 and C-MgO can be useful adsorbents in the possible environmental pollution cleanup.