Room-temperature fabrication of magnetite-boehmite sol-gel composites for heavy metal ions removal

2.1 Chemicals

Aluminum isopropoxide $⩾$ 99%, iron(II) chloride tetrahydrate $⩾$ 98.5%, iron(III) chloride hexahydrate $⩾$ 99%, aqueous solution of ammonia $⩾$ 27.5%, potassium bichromate, hydrochloric acid, sodium hydroxide were obtained from Sigma-Aldrich.

2.2 Ferria hydrosol

Stable ferria hydrosol was prepared in a manner described in our previous work (Drozdov et al., 2016; Shabanova et al., 2016) and illustrated in Scheme 1a. 2.5 g FeCl₂∗4H₂O and 5 g FeCl₃∗6H₂O were dissolved in 100 mL of deionized water under constant stirring (500 rpm). Then, 12 mL of aqueous ammonia solution was added under constant stirring (500 rpm) at room temperature. Formed magnetite precipitate was magnetically separated and washed with deionized water until neutral pH. The washed black precipitate was mixed with 100 mL of deionized water and subjected to ultrasonic treatment (37 kHz, 110 W) under constant stirring (300 rpm). The resulting magnetite sol was afterwards cooled to room temperature. The resulting mass concentration of the ferria sol was 2%.

Close

Scheme 1

Preparation routes of ferria (a) and alumina (b) hydrosols.

Export to PPT

2.3 Alumina hydrosol

Stable alumina hydrosol was prepared using a processing route described in our previous work (Chapurina et al., 2016) and illustrated in Scheme 1b. 3.3 g of aluminum isopropoxide was added to 100 ml of deionized water pre-heated to 80  $°$ C under constant stirring (300 rpm) for 10 min. Then, the resulting suspension was subjected to ultrasonic treatment (37 kHz, 110 W) under constant stirring (300 rpm). The resulting boehmite sol was afterwards cooled to room temperature. The resulting mass concentration of the alumina sol was 2%.

2.4 Ferria-alumina composites

For preparation of composite materials, the freshly prepared ferria hydrosol (0.75–0.25 mL) was mixed with the alumina hydrosol, then the solution was homogenized by vigorous mixing. The prepared solution was dried in an evaporator at 60 °C at normal pressure and the resulting xerogel was ground in a mortar.

2.5 Wastewater

Wastewater sample was taken from rinsing wastewater of degreasing and metal plating process in electroplating laboratory of ITMO University.

2.6 Cr(VI) adsorption and sorbent regeneration

The adsorption isotherms of Cr(VI) ions were studied using 20 mg of the sample dispersed in 10.0 mL bichromate solutions with the initial concentrations of Cr(VI) ions ranging from 10 to 150 mg/L under constant stirring for 60 min. The temperature of adsorption was 25 °C and the pH values of the initial Cr(VI) solutions were 2.0 ± 0.1. After the experiments, the suspensions were filtered through membrane filters with a pore size of 0.2 m and Cr(VI) concentrations in the filtrate was determined by the 1,5-diphenylcarbazide method (Farag et al., 1981). For regeneration, adsorbents were stirred in 20 mL of NaOH (0.005 M) at 25  $°$ C for 30 min. The nanoparticles were magnetically separated, washed with deionized water and reused in the experiments.

2.7 Adsorbent stability

In order to evaluate the stability of the materials, aluminium and iron leaching were measured by atomic absorption spectroscopy. For this purpose the materials were incubated at pH 2 or pH 12 for 10 h at 25 °C under constant stirring, then the supernatant was analyzed using standard procedures (Centi et al., 2000; Strohmeier and Hercules, 1984).

2.8 Characterization methods

Surface area, pore volume, and pore size distributions were investigated by the nitrogen adsorption technique at 77 K using Quantachrome Nova 1200e and subsequent analysis by BET and BJH methods. Prior to analysis, all samples were degassed at 110 °C for 4 h. The particle size distribution and zeta potential were measured using a Photocor Compact-Z analyzer. The crystalline phase and crystallinity of the samples were characterized by X-ray diffraction (Bruker D8 Advance) using Cu K $\alpha$ radiation ( $\lambda$ = 1.5418 Å). The particle morphology and size were investigated by scanning electron microscopy (SEM) using a VEGA3 TESCAN scanning electron microscope equipped with an X-Act EDX detector. The Raman spectra were recorded using the 633 nm He-Ne laser line on a Horiba Jobin-Yvon Micro Raman 300 spectrometer. The laser power on the samples employed was 0.030 mW and 0.344 mW, with 300 s and 120 s exposition per diffraction window, respectively. In all the measurements, 50× Olympus lens, hole of 500 mm, slit of 100 mm and a diffraction grid with 1800 grooves/mm were employed. The samples for transmission electron microscopy (TEM) were obtained by dispersing a small probe in ethanol to form a homogeneous suspension. Then, the suspension was drop-coated on a copper mesh covered with carbon for a TEM analysis (FEI TECNAI G2 F20, at an operating voltage of 200 kV). Spectrophotometry was carried out using an Agilent Cary HP 8454 Diode Array spectrophotometer with TEC. Atomic adsorption spectroscopy was performed using Perkin-Elmer Model 2380 AAS.

3.1 Ferria and alumina hydrosols

The synthesis of ferria-alumina composites was performed in three stages: synthesis of magnetite hydrosol, synthesis of alumina hydrosol, and room temperature sol-gel transition. The synthesis of magnetite hydrosol was performed by the modified coprecipitation route. While the mechanism of the reaction is complex and may involve multiple intermediary products, the overall reaction can be described by Eq. (1) resulting in the formation of truncated octahedral nanoparticles of magnetite (Ahn et al., 2012). The formation of boehmite in the present investigation can be envisaged by the hydrolysis and alcohol condensation reaction as demonstrated in Eq. (2):

The results of material characterization for the both hydrosols are illustrated in Fig. 1(a-h). The hydrosols were obtained in form of jelly-like fluids, remaining stable without precipitation for a long period (Fig. 1a, e). Analysis of their hydrodynamic parameters showed narrow particle size distributions with an average hydrodynamic diameter of 74 and 32 nm for boehmite and magnetite, correspondingly. The high stability of the hydrosols was evidenced by highly positive zeta-potential values of the nanoparticles (NPs): +42 mV at pH 7.0 and +36 mV at pH = 7.4 in the alumina and ferria hydrosols, correspondingly. The enhanced hydrosols stability was assisted by application of ultrasonic (US) irradiation in a course of synthesis, thus improving the hydrophilic properties of the synthesized nanoparticles. Acoustic cavitation caused by US-irradiation leads to formation, growth, and implosive collapse of bubbles at the liquid-solid interface (Prosperetti and Plesset, 1978). The implosive collapse of bubbles generates local heating of the contact area through the adiabatic compression or formation of a shock wave in the gas phase of a bubble, whose speed can reach 100 m/s. This results in instantaneous heating of the particle surface and bordering water layer. Under such circumstances, free radicals such as H^• and OH^• can form and interact with the surface of colloidal particles, resulting in the surface hydroxylation and shifting the isoelectric point of the materials to higher values. The experimentally observed shifts ranged from pH = 8.3 to pH = 10 for the boehmite NPs and from pH = 6.8 to pH = 8.0 for the magnetite NPs (ESI Fig. 1S). Thus the highly positive values of the NPs zeta-potential were achieved even at neutral pH levels and led to the formation of stable hydrosols of alumina and ferria (Fig. 1a, e).

Close

Fig. 1

Alumina (a-d) and ferria (e-h) hydrosols used for preparation of the composites. Visual appearance of the as-synthesized alumina (a) and ferria (e); XRD spectra of the alumina, (reference JCPDS 21-1307 shown as the blue lines) (b), and ferria (reference JCPDS No. 19-0629 shown as red lines) (f); HR-TEM image of the alumina (c) and ferria (g) samples, interplanar spaces are shown in the inserts; HR-SEM image of alumina (d) and ferria (h).

Export to PPT

The alumina hydrosol consisted of nanoparticles with an average crystallite size of 5 × 5 × 2 nm according to TEM and SEM (Fig. 1c, d). Clear visibility of lattice fringes (width = 0.271 nm) in Fig. 1c confirms the crystalline nanoparticle structure. XRD patterns (Fig. 1 of the as-synthesized nanoparticles demonstrated broad peaks, indicating ultra-fine and small crystallite sizes and matched with the XRD pattern for boehmite crystalline phase (reference JCPDS 21-1307). The main peak at 14.45° is attributed to the crystalline plane with Miller indices of (020). Other distinctive peaks at 28.10° (120), 38.38° (031), 49.09° (051), 54.98° (151), 64.62° (022), 71.99° (251) (Fig. 1b) are matched those of the standard boehmite diffraction pattern (Xia et al., 2013). Cell parameters of 2.824 (a), 12.481 (b) and 3.689 (c) agree with the reference values for boehmite crystal phase (Xia et al., 2013). Analysis of the hydrosol hydrodynamic parameters by the dynamic light scattering technique showed narrow size distribution of particles with an average hydrodynamic diameter of 74 nm. Having the zeta-potential value of +42 mV at pH 7.0, the synthesized boehmite nanoparticles formed a jelly-like hydrosol that remained stable without precipitation for a long period (Fig. 1a).

In its turn, the ferria hydrosol consisted of nanoparticles with an average mean particle diameter of 10 ± 2 nm and truncated octahedron morphology according to TEM (Fig. 1g) and SEM (Fig. 1h) images. XRD pattern of the nanoparticles corresponded to magnetite crystalline phase (JCPDS No. 19-0629) and demonstrated the main peak at 35.45° attributed to the crystalline plane with Miller indices of (311). Other distinctive peaks at 18.52° (0 2 0), 30.10° (2 2 0), 43.08° (4 0 0), 53.45° (4 2 2), 56.98° (5 1 1), 62.57° (4 4 0), 70.99° (5 3 3), and 74.02° (4 4 4) (Fig. 1f) matching the standard magnetite diffraction pattern (Cornell and Schwertmann, 2003). Measured lattice spacing d values of 0.297, 0.253, 0.209, 0.161, and 0.148 nm was agreed with the reference values of 0.297, 0.253, 0.209, 0.162, and 0.149 nm for the (2 2 0), (3 1 1), (4 0 0), (5 1 1), and (4 4 0) crystallographic planes of Fe₃O₄, respectively (Cornell and Schwertmann, 2003). This confirmed the inverse spinal structure of the synthesized ferria nanoparticles (space group Fd3m) with the lattice parameter a = 0.8370 nm that correlates with the one provided in JCPDS file No. 19-0629. Fig. 2 illustrates the Raman spectra of the magnetite sample recorded in batch experiments at different laser beam powers. It demonstrates an additional evidence of the magnetite crystal phase in the synthesized materials. The line (a) in Fig. 2 was recorded at a 633 nm He-Ne laser power of 0.030 mW with acquisition time of 300 s per diffraction window. It shows the characteristic broad band, typical for magnetite at 667 cm¹ (Li et al., 2012). Further observations have been in agreement with the known series of phase transformations of magnetite under the Raman laser measurement conditions: as the power was increased to 0.344 and 1.35 mW (acquisition times of 90 and 30 s, respectively), the characteristic maghemite bands were observed at 711, 492, and 372 cm⁻¹ (Fig. 2 line b), and finally, the full oxidation state was characterized by the hematite characteristic bands at 591, 385, 273, and 213 cm⁻¹ (Fig. 2 line c) (Li et al., 2012; Andreeva et al., 2017).

Close

Fig. 2

Raman spectra of the magnetite sample recorded in batch experiments at different laser beam powers.

Export to PPT

Fig. 3 illustrates the magnetic attraction of the mixed hydrosols with various alumina to ferria ratios. As the ferria mass fraction decreases, the attraction of the hydrosol to the magnet is weakening. Due to high zeta potential (+36 mV at pH = 7.4), the magnetite nanoparticles formed stable hydrosol with magnetic fluid-like behavior (Fig. 1e). As shown in Fig. 3, they can be attracted by an external magnetic field together with the dispersion phase (Fig. 3). Since the alumina particles in the mixed hydrosols have similar charge as ferria, these two oxides can be mixed at any ratio without mutual destabilization, thus ensuring a high degree of the colloidal homogeneity of the resulting solution (Fig. 3). The colloidal stability of the blends appeared to be high enough to prevent any notable precipitation during one week of storage on the shelf. The stability of the ferria-alumina blends are close to one observed for pure ferria and alumina hydrosols described earlier (Drozdov et al., 2016; Chapurina et al., 2016) and is thus a distinctive feature of the materials prepared by US-assisted synthesis (Vinogradov and Vinogradov, 2014; Kwan et al., 2015; Gedanken, 2007).

Close

Fig. 3

Visual appearance of ferria and alumina hydrosols and their mixtures.

Export to PPT

3.2 Ferria-alumina sol-gel derived matrices

One of the main features of the synthesized hydrosols is a pristine surface of NPs in the colloidal solution. Their surfaces shall be free of organic molecules or peptizing agents in order to take part in sol-gel process. Upon solvent removal, the synthesized hydrosols transform into mesoporous xerogel matrices with the controllable alumina-to-ferria ratio. All the observed sol-gel transitions were irreversible and resulted in stable ceramic matrices not inclined to spontaneous re-dispersion in water media. The TEM and SEM analyzes showed a uniform distribution of the components in the resulting materials (Fig. 4a-c). The homogeneous coloring of the energy-dispersive X-ray mapping (Fig. 4b) indicates the absence of aggregation that might have occurred at some critical processing stages such as mixing or solvent removal. Such uniform distribution of components in the sol-gel derived materials is a distinctive feature of sol-gel processes performed with highly homogeneous systems prior to gelation (Tabesh et al., 2018; Drozdov et al., 2016; Nassar et al., 2017). The XRD spectra of the composites include the series of peaks associated with their both components (reference JCPDS files 19-0629 and 21-1307), thus proving their presence in crystal form (Fig. 4d-f).

Close

Fig. 4

Characterization of composite matrices. HR-SEM image of the ferria-alumina 1:1 composite (a); EDX image of the ferria-alumina 1:1 composites (b); TEM image of the ferria-alumina 1:1 composite (c); XRD spectra of the ferria-alumina composite with mass fractions ferria-alumina varied from 3:1 to 1:3, respectively (d-f).

Export to PPT

In order to investigate the nature of interactions between magnetite and boehmite nanoparticles in the ferria-alumina composites, a sample of (1:1) was investigated with Raman spectroscopy. All the observed reflections can be attributed to the characteristic peaks of pure magnetite (667 cm⁻¹) and boehmite (684, 497, 462, 362, 338 cm⁻¹) while no any additional vibrational modes corresponding to potential Fe-O-Al bonds were observed. (see ESI Fig. 2S) This suggests no chemical interaction between boehmite and magnetite occurred during the sol-gel processing. Thus the derived materials can be described as two interlaced sol-gel matrices rather than monolithic ceramic material. The textural properties of the formed xerogels are shown in Fig. 5(a-c). The surface areas of pure magnetite and boehmite xerogels were determined as 122 m²/g and 391 m²/g, respectively, as well as their linearly depending ratios in the intermediate composites (Fig. 5a). The total pore volume in the materials also demonstrated a linear increment ranging from 0.258 cm³/g in pure ferria to 0.298 cm³/g in pure alumina xerogels (Fig. 5b). The average pore diameter demonstrated an exponential decay with an increase of alumina mass fraction (Fig. 5c). This can be a result of the two factors. First, the smaller average size of the boehmite nanoparticles favors smaller average pore diameter. Second, the difference in morphology of the nanoparticles (flakes of alumina, Fig. 1c, and truncated tetrahedrons of ferria, Fig. 1g) results in more dense packing of the materials with higher alumina content. The zeta potential of composite materials depends on alumina mass fraction and linearly increases from +18 to +24 mV while moving from pure ferria to pure alumina (see ESI Fig. 3S).

Close

Fig. 5

Textural properties of the composite materials. Surface area as a function of the alumina mass fraction (a); total pore volume as a function of the alumina mass fraction (b); average pore diameter as a function of the alumina mass fraction (c).

Export to PPT

The magnetic properties of the synthesized materials have been characterized, and the corresponding magnetization curves measured for several iron containing composites are shown in Fig. 6. All the curves demonstrate superparamagnetic behavior with high saturation magnetization. They do not demonstrate any hysteresis, neither any remnant magnetization. Variation of the magnetite mass fraction is shown to control the amplitude of the magnetization response, up to the level of 88 emu/g in the pure ferria, which is close to the value of bulk magnetite ( $\approx$ 92 emu/g) (Mikhaylov et al., 2017). As aluminium and its oxides have very weak paramagnetic properties (with permeability at the level of 10⁻⁵ H/m), its presence in the composite material do not affect the magnetic behavior of the material, thus the net magnetic properties of the ferria-alumina composites are determined only by the mass fraction of magnetite nanoparticles. All the measured magnetization curves have been analyzed using the Langevin function (Eq. (3)):

Close

Fig. 6

Magnetization curves of the composite materials. Inset: the low-field.

Export to PPT

The high magnetization properties of the ferria-alumina composites can be used for their magnetically controlled separation in various solutions. An optimal balance between the magnetization and textural properties of the functional ferria-alumina composite can thus be adjusted by the ratio between the constituent components to meet the specific requirements of various applications.

3.3 Chromate absorption mechanism

The combination of high surface area and high magnetization makes the ferria-alumina composites promising adsorbent for water purification. In aqueous systems, the surfaces of the both constituent oxides are covered with hydroxyl groups, which could be protonated or deprotonated to generate charged M- ${\text{OH}}_2^{+}$ or M-O⁻ functional, respectively, groups at pH values below or above the zero point of charge (PZC). In the materials under study the PZC was observed at pH = 8 (for ferria) and pH = 10 (for alumina), thus implying protonation and positive charging of their surfaces at neutral or acidic media. Adsorption of chromate ions to the surface of magnetite or boehmite is a result of their electrostatic interaction and greatly depends on the pH level of the solution. Thus, the electrostatic adsorption of the chromate anions requires the positive charge of the NPs surface. At low pH, a variety of different Cr-anionic species (Cr₂ ${\text{O}}_7^{2-}\text{,}{\text{HCrO}}_4^{-}$ , Cr₃ ${\text{O}}_{10}^{2-}$ , Cr₄ ${\text{O}}_{13}^{2-}$ ) may coexist in water (Yuan et al., 2009). In order to ensure the maximal chromium adsorption capacities the corresponding pH level is typically maintained at 2–3, at which the main form of adsorbed chromate species is represented by ${\text{HCrO}}_4^{-}$ with relatively low adsorption free energy (2.5–0.6 kcal/mol) (Yuan et al., 2009).

Let us now discuss the possible mechanisms of chromium adsorption to ferria and alumina surfaces. Fig. 7 illustrates the different scenarios for adsorption of a chromate ion on the surfaces of ferria (Fig. 7a) and alumina (Fig. 7b). The ferria lattice region contains both Fe²⁺ and Fe³⁺ sites. The surface Fe³⁺ sites are depicted as adsorbing the ${\text{HCrO}}_4^{-}$ ions (Fig. 7a), although both Fe²⁺ and Fe³⁺ surface sites might be involved. In this case, the chromate ion adsorption typically involves two-site (chelating) bonding to the surface, though the single Fe-O-Cr attachment may also occur. Likewise, ${\text{HCrO}}_4^{-}$ and ${\text{CrO}}_4^{2-}$ will be in equilibrium in a solution thus both competing to react at the surface site (Rajput et al., 2016). The ratio of Fe²⁺/Fe³⁺ sites represented in the schemes is obviously arbitrary for these idealized pathways, only suggesting a clue to the mechanism involved. In case of alumina, the adsorption mechanism is more dependent on the pH level of the solution. For pH above 6, the inner-sphere monodentate and bidentate complexes are the dominant pathway for chromate adsorption (Fig. 7b), while for pH ranging between 2 and 3 the outer-sphere complexes are observed. The chromate adsorption is suppressed at pH < 2 and high ionic strength values, because similar reaction mechanisms are responsible for these processes. It is supposed that this suppression is related to limited availability of exchangeable surface groups and those inner-sphere complexes should also be considered (Johnston and Chrysochoou, 2015).

Close

Fig. 7

Adsorption mechanisms for a chromate ion on the ferria (a) and alumina (b) surface.

Export to PPT

3.4 Chromate absorption kinetic

The adsorption kinetics is one of the key features of the adsorption process, important for both modeling and application purposes. The kinetic curves obtained for several composites with various ferria-to-alumina ratios are shown in Fig. 8. These curves clearly demonstrate that the sorption equilibrium depends on the ferria-to-alumina ratio of the composites. The adsorption equilibrium depends on the total surface area of the sorbent material. For the system under study it increases gradually from 23 mg/g to 62 mg/g with an increase of the alumina fraction in the composite xerogel matrices (Fig. 8a). Plotting the t/q_e ratio versus time (Fig. 8b) yields a set of straight linear dependencies for materials with various ferria-to-alumina ratios, Thus the sorption mechanism can be described by the second-order equation (Rodrigues and da Silva, 2009) Eq. (4):

Close

Fig. 8

Adsorption curves of Cr(VI) ions on composite materials. Kinetic curve of the chromate adsorption (a); Pseudo-second-order kinetics curves for the adsorption of Cr(VI) (b).

Export to PPT

The adsorption isotherms have been analyzed using the Langmuir and Freundlich equations (Langmuir, 1918; Freundlich, 1906). The Langmuir equation is associated with an isotherm model assuming that (1) the adsorbent surface is homogeneous; (2) adsorption is limited by monomolecular layer coverage; (3) there are no interactions between adsorbate molecules (Langmuir, 1918). The Langmuir equation can be expressed as follows Eq. (5):

Analysis of the R-squared standard deviations listed in Table 1 suggests that the Langmuir model fits better the experimental data (with R² (>0.941). This is an indication of rather homogeneous distribution of the active sites on the adsorbent surfaces. The 1/n values are in the range of 0–1 for all samples, suggesting that the adsorption is thermodynamically favorable. The adsorption capacity of materials gradually increases with an increase of the alumina mass fraction, in correlation with a change in the specific surface area determined by physical nitrogen adsorption. Compared to magnetite-based adsorbents reported in earlier studies, the composite materials demonstrate significantly enhanced sorption capacities (see ESI Table 1S). This can be an effect of larger surface areas and higher saturation of the surface by hydroxyl groups, as a result of the compositional variation and the applied synthesis conditions.

3.5 Absorbents recycling

As mentioned earlier, the magnetosensitive properties of the composite adsorbents allow performing their magnetic separation with a subsequent regeneration and recycling. In the laboratory experiments this was realized using a permanent magnet, while special flow magnetic separators can be used in real applications, with a performance up to several tons per hour (Yavuz et al., 2009; De Latour, 1973; Rossier et al., 2012). In order to investigate the reuse potential of the composites, the materials were separated magnetically (or by centrifugation for pure alumina) and regenerated in sodium hydroxide solution and subjected to another round of adsorption with totally five batches of adsorption – desorption cycles (Fig. 9). A gradual degradation down to 89–92% of absorption capacity was observed for materials after five cycles, depending on material composition. The degradation rate was slightly higher in composites with higher alumina fraction. This difference in materials regeneration can be attributed to a higher energy of chromate ion adsorption to boehmite crystal structure ( $\mathrm{\Delta }{\text{G}}_{\mathit{ads}}$ −140 kJ/mol) in comparison with adsorption on magnetite ( $\mathrm{\Delta }$ ${\text{G}}_{\mathit{ads}}$ −85 kJ/mol) (Johnston and Chrysochoou, 2015; Suh et al., 2015). Compared to the earlier published results, the recyclability of magnetic ferria-alumina composites are found to possess better regeneration properties than the pristine magnetite- based adsorbents (up to 25% degradation after 2 cycles) (Rajput et al., 2016; Wang et al., 2010) and close to one observed for magnetite-silica based materials (90–95% of adsorptive capacity regeneration) (Topel et al., 2014; Wang et al., 2010).

Close

Fig. 9

Recycling of the composite materials.

Export to PPT

In order to further evaluate the prospects of the composites as recyclable adsorbent, their chemical stability was addressed through evaluation of iron and aluminium ions leaching at low (pH 2.0) and high (pH 12.0) conditions for a sufficiently long time period. The composites were found to be stable to bases (leaching of iron and aluminium ions did not exceed 0.1 wt.% after 10 h of incubation for all the tested materials), while notable leaching was observed in acidic media, reaching 12% and 0.5 wt.% for iron and aluminium respectively (for the 3:1 ferria-alumina composite). While this observation implies gradual degradation of the materials, they still can be regenerated and reused for several times until the complete loss of magnetic responsiveness.

3.6 Purification of real wastewater samples

Finally, we address a set of experiments comparing the purification abilities of the designed composites in real conditions. In these experiments we examined them using wastewater obtained after electroplating process. The real wastewater included the following contaminations prior to the experiment: 100 ppm Cr(VI), 21 ppm Al, 25 ppm Mg, 40 ppm Ca, and 10 ppm B). A sample of artificially contaminated water (100 ppm Cr(VI)) was prepared in the laboratory, and used for simulation in similar experiments. The percentage of chromate uptaking by different materials in these experiments is listed in Table 2. In all the cases the purification performance with the simulation liquid was almost 20% higher than with the industrial effluent. Such difference is in agreement with previously published results for other systems and can be attributed to the competitive adsorption of interfering ions in the effluent (Dakiky et al., 2002; Martín-Lara et al., 2014).