Introduction

Several thorough studies have concluded that TiO2-based materials are the most efficient photocatalysts to promote oxidation and reduction reactions, regarding the removal of a wide range of organic and inorganic compounds (Choi et al. 2001; Dong et al. 2015; Eskandarloo et al. 2015; Marinho et al. 2017b; Yang and Lee 2006). However, these materials usually exhibit unsatisfactory photocatalytic efficiency due to their restricted visible-light absorption capacity (Wang et al. 2018b), as well as low surface area and easy recombination of charge carriers (Guayaquil et al. 2017). Therefore, a number of modification approaches are adopted to enhance the materials’ photocatalytic performance, such as manufacturing a nano/mesoporous structure (Adán et al. 2007; Avansi et al. 2014; Guayaquil et al. 2017; Pongwan et al. 2016), combining them with other semiconductors (El Mragui et al. 2019; Qin et al. 2019), and doping and hybridizing them with non-metal (Khedr et al. 2017; Patil et al. 2019) or metal atoms (Adán et al. 2007; Lei et al. 2014b; Pongwan et al. 2016). This element doping strategy plays an essential role in tuning the bandgap structure, extending their light absorption, increasing their charge transfer mobility, and creating more active surface sites (Yan et al. 2013).

Other photocatalysts have also been investigated, including materials based on copper (Huang et al. 2017b), zinc (Gholami et al. 2016), and iron (Roonasi and Mazinani 2017) oxides, as well as cadmium (Wang et al. 2017) and zinc (Wang et al. 2017) sulfides. Although these photocatalysts may present high efficiency in the photo-oxidation/photoreduction of several organic and inorganic pollutants, they also present high preparation costs, making the process expensive and infeasible from a practical point of view. Thus, photocatalytic enhancement using natural, abundant, and cheap raw materials has become attractive in the area of photocatalysis.

Pozzolan is a natural and abundant low-cost resource, consisting mainly of SiO2, Al2O3, and Fe2O3 oxides (Wamba et al. 2017). Italy, Chile, Greece, and Cameroon (Wamba et al. 2017) have deposits of this material and process 3000, 915, 900, and 600 kt/year, respectively. Natural pozzolan is used as a raw material in cement factories and in water treatment as an adsorbent (Wamba et al. 2017). Although studies have focused on natural pozzolans used for improving the performance and durability of self-compacting concrete (Moon et al. 2014; Omrane et al. 2017), only a few studies have concentrated on these materials under enhanced visible-light photocatalysis (Lackhoff et al. 2003; Wang et al. 2018a).

Even though or despite the fact natural pozzolan exhibits poor photocatalytic performance, the mixing of a natural pozzolan with titanium(IV) isopropoxide was an attempt to create an innovative photocatalyst, through a facile and affordable way.

Unlike most organic pollutants, heavy metal ions are not biodegradable into harmless end products as CO2 and H2O (Carolin et al. 2017). Heavy metals are found in wastewater from many types of industries such as chemical, petrochemical, mining, and metallurgical (Carolin et al. 2017; Enuneku et al. 2017). Photocatalysis has been applied successfully over the last few years in treating a wide range of wastewaters polluted with inorganic compounds, including Cr(VI) (Byrne et al. 2017; Soares et al. 2016). Cr(VI) is known as potentially carcinogenic and mutagenic (Trejo-Valdez et al. 2019). The mechanism of the direct photocatalytic reduction of Cr(VI) occurs when semiconductors are excited by UV radiation, leading to the generation of an electron/hole pair (Kumar and Chowdhury 2018). These photogenerated electrons are transferred from the semiconductor conduction band to the chromium species; in parallel, the positive holes can react with H2O/HO− producing HO• radicals (Litter 2015). Positive holes and HO• radicals can result in the reoxidation of Cr(III) to Cr(VI) (Kumar and Chowdhury 2018). In addition, the electron-hole recombination can also suppress the reaction. Several organic agents (oxalic, citric, tartaric acid, etc.) are often added as sacrificial agents, reacting with the HO• radicals or holes and hindering the electron-hole recombination, boosting the Cr(VI) reduction efficiency (Litter 2015). The organic acid oxidation can result in the formation of strong reducing agents (HCOO•−), able to promote an indirect Cr(VI) reduction in the solution bulk (Marinho et al. 2018; Meichtry et al. 2011).

The main goal of this work is the development of a novel hybrid material synthetized from pozzolan and titanium(IV) isopropoxide, as a potential catalyst for the Cr(VI) reduction in aqueous solution. The reduction rate was evaluated as a function of Cr(VI) initial concentration, solution pH, and type of scavenger used to avoid the reoxidation of Cr(III) to Cr(VI).

Materials and methods

Chemicals, reagents, and solutions

All solutions were prepared using deionized water. Chemicals of an analytical grade were used: ethanol (Merck, 99.99%), titanium(IV) isopropoxide (Aldrich, 97%), potassium dichromate (Aldrich, 99.9%), nitric acid (Aldrich, 65%), and sodium hydroxide (Aldrich, 98%). The Cr(VI) solution was prepared by dissolving potassium dichromate (K2Cr2O7) in deionized water.

Synthesis of titanium(IV) isopropoxide nanosheets

Titanium(IV) isopropoxide nanosheets (TnS) were prepared by mixing a solution of ethanol (125 mL) and titanium(IV) isopropoxide (1.25 mmol of Ti(IV)). The suspension was agitated for 1 h. Then, 1.3 mL of water mixed in 50 mL of ethanol was added drop by drop in the suspension. In sequence, the mixture was agitated for 1 h again. The dispersion was centrifuged at 3500 ppm (5 min) and dried in an oven at 80 °C for 4 h. The obtained material was hydrothermally treated in a Teflon reactor with sodium hydroxide solution (10 M) at 140 °C for 36 h. The resultant material was neutralized with nitric acid (3 M) and washed with distilled water until a near-neutral pH (6.6) was obtained. The material was then dried at 100 °C for 24 h.

Natural pozzolan

Natural pozzolan (Pz) was collected on the Cameroon Volcanic Line in Adamawa, Cameroon. The coarse natural pozzolan collected was crushed, sieved, washed, and dried at 105 °C for 24 h in an oven. The dried pozzolan was then pulverized in an agate mortar. The obtained powder was sieved in order to achieve a fraction lower than 53 μm for further chemical modification.

Synthesis of hybrid titanium(IV) isopropoxide/pozzolan nanosheets

Hybrid/pozzolan nanosheets (TnS-Pz) were prepared by mixing natural pozzolan in a solution of ethanol (125 mL) and titanium(IV) isopropoxide (1.25 mmol of Ti(IV) per mass of material). The dispersion was agitated at room temperature for 1 h. Then, 1.3 mL of water mixed in 50 mL of ethanol was added drop by drop in the suspension. In sequence, the mixture was agitated for another 1 h. The dispersion was centrifuged at 3500 rpm (5 min) and dried in an oven at 80 °C for 4 h. The material obtained was hydrothermally treated in a Teflon reactor with a sodium hydroxide solution (10 M) at 140 °C for the 36 h. The resultant material was neutralized using nitric acid (3 M) and washed with distilled water until a near-neutral pH (6.6) was achieved.

The material obtained was then dried at 100 °C for 24 h and kept in plastic tubes for further characterization and applications. An overview of the preparation of the TnS-Pz (Fig. S1A) as well as the scheme of the chemical bond between the titanium and pozzolan (Fig. S1B) is presented in the supplementary material.

Nanosheet material characterization

Nitrogen adsorption isotherms were recorded with a commercial system at − 196 °C after drying for 3 h at 120 °C under low pressure (< 2 mbar). The surface area was determined using the Brunauer, Emmett, and Teller (BET) method. A scanning electron microscope (SEM) equipped with an energy-dispersive X-ray spectroscope (EDX) (TESCAN 3, Sweden) and transmission electronic microscope (TEM) (JEOL JEM-1200 Exll) were both used to observe the surface morphologies and elemental composition.

A Bruker spectrometer of the alpha model Fourier transform infrared spectroscopy (FTIR) with an ATR (attenuated total reflectance) accessory was used to determine the functional groups of the hybrid materials. Sixty-four cumulative scans over the 4000–400 cm−1 range at a resolution of 4 cm−1 recorded the spectrum.

A Shimadzu UV-2450 spectrophotometer, using an ISR-2200 Integrating Sphere Attachment at a spectral range of 200–900 nm, provided the ultraviolet-visible diffuse reflectance (DRUV) and diffuses reflectance. In this analysis, the samples were treated as powder. BaSO4 (Wako Pure Chemical Industries, Ltd.) was used to obtain the baseline. All measurements were carried out at room temperature (25 °C ± 2).

The TnS-Pz samples were analyzed in X-ray diffraction instruments (BRUKER-Diffrac 5000 and PanAnylical-Xpert MPD) operating at 45 kV and 40 mA using Cu-Kα monochromatic radiation (λ = 1.5418 Å), 2θ angle interval from 2 to 100° with step sizes of 0.016° and counting time of 60 s by step, slit widths of 200 μm (D5000), and Xcelerator linear detector with active window of 2.122o (Xpert MPD).

The morphological structure was tested using the X-ray diffraction (XRD) method, performing an ENRAF NONIUS FR590 XRD (Bruker AXS, MA, USA) diffractometer at Cu Kα 1.2 radiation and a step/time scan mode of 0.05 s−1. The XRD pattern of TnS-Pz sample was compared with the diffraction powder file (PDF2) for the following components: diopside aluminian (reference pattern 01-086-0004), anatase, syn (TiO2) (01-089-4221), calcium aluminum silicat (Ca(Al2Si2O8) (01-087-1616), hematite, syn (Fe2O3) (01-089-8103), and titanomagnetite (Fe2.5Ti0.50O4), syn (01-075-1377).

The 57Fe-Mössbauer spectra (Wissel spectrometer) were collected at room temperature under standard transmission geometry in constant acceleration mode where a 57Co source was embedded in the rhodium matrix. The spectrometer was calibrated according to the spectrum of natural iron. The Mössbauer spectra were fitted by the WinNormos-DIST program (Brand 2008). The isomer shift values (δ) are quoted relative to natural iron (δ = 0).

Photocatalysis experimental procedure

The Cr(VI) reduction reactions were performed in a lab-scale annular photoreactor (Fig. S2), as reported elsewhere (Marinho et al. 2017a; Salim et al. 2016). The Supplementary Material presents a detailed description of the photoreactor.

The recirculation glass vessel of the lab-scale prototype was filled with 1.5 L of Cr(VI) solution (0.02–0.20 mM), which was pumped into the annular photoreactor and homogenized by recirculation in the closed system for 15 min in complete darkroom conditions (Q = 1.6 L min−1). The temperature set-point of the refrigerated thermostatic bath was adjusted to maintain the intended solution temperature (25 °C). After which, an organic acid scavenger, citric acid or tartaric acid, was added, tested at [scavenger]0/[Cr(VI)]0 M ratio of 3:1 (based on previous results as reported in the work of Vilar and co-workers (Marinho et al. 2016)), and the pH was adjusted from 3.0 to 6.0, using sulfuric acid/sodium hydroxide. After that, the photocatalyst was added at a concentration of 15 g L−1 and the UVA-Vis lamp was turned on. Samples were taken after each addition before turning on the UVA-Vis lamp (Philips Actinic BL TL 6W/10 1FM/10X25CC), and during the photoreduction period at predetermined times. Fig. S3 shows the spectral power distribution of the UVA-Vis lamp. Other assays were also performed under the following conditions in order to verify the influence of each component: (i) the absence of a catalyst; (ii) the absence of a scavenger; (iii) the absence of a catalyst and a scavenger; (iv) the absence of a scavenger and light.

Ferrioxalate actinometry (6 mM of Fe(III) and 30 mM of oxalic acid) determined a photonic flux reaching the reaction medium of 1.67 × 10−6 einstein s−1, corresponding to a radiant power of 0.56 W (Φ = 1.25 (Kuhn et al. 2004)). Total chromium concentration was determined by atomic absorption spectrometry (AAS, Shimadzu model AA 6300) using an air-acetylene flame. Cr(VI) concentration was evaluated by molecular absorption spectrophotometry. The procedure followed was based on the formation of a pink complex of Cr(VI) with 1.5-diphenylcarbazide in acid solution, with maximum absorbance at 540 nm.

Results and discussion

Characterization of the TnS-Pz sample

The nitrogen adsorption-desorption isotherm was performed to measure the specific surface area and understand the pore nature of the TnS-Pz material. As shown in Fig. 1, the TnS-Pz isotherm curve exhibited a curve type IV isotherm with a hysteresis loop, representing capillary condensation, as is typical for mesoporous materials. The mesoporosity of the TnS-Pz is highlighted by the high external surface area equal to 160.2 m2 g−1 when compared with the micropore material surface area (61.5 m2 g−1). The nitrogen adsorption isotherms confirm the presence of micropores where the low-pressure region (P/P0 < 0.1) exhibits a rapid increase in the uptake of N2 as the relative pressure is increased. Regarding specific surface areas (SBET), the precursor (natural pozzolan) exhibited a very low SBET equal to 13.2 m2 g−1 (Celik et al. 2014); however, the final hybrid material presented well-developed porosity with higher SBET, 271.7 m2 g−1.

Fig. 1
figure 1

N2 adsorption-desorption isotherms of pure pozzolan and TnS-Pz. () adsorption, () desorption

Full size image

As the reflectance spectra (DRUV) help to understand the photocatalytic properties of the materials, the UV-Vis spectra of the TnS-Pz hybrid material were compared with the pure pozzolan (Pz) and titanium(IV) isopropoxide (Fig. 2a). The TnS-Pz material spectra showed a higher absorption region, in the UV spectrum, as well as that the titanium(IV) isopropoxide displays a higher absorption intensity than the other materials. However, from 400 up to 650 nm (visible region), the hybrid material presents a small increase in the absorption intensity compared with the precursors, suggesting the TnS-Pz to be electronically unstable through the interactions between them. Since the pozzolan structure is mainly based on SiO2 and Al2O3, the Si-O-Ti and Al-O-Ti covalent bonds between the pozzolan and titanium(IV) isopropoxide must occur over both structures during the sol-gel process (Wang et al. 2017), changing their characteristics.

Fig. 2
figure 2

a Diffuse reflectance of UV-Vis (DRUV) absorption spectra of natural pozzolan, hybrid material (TnS-Pz), and titanium(IV) isopropoxide, ― titanium(IV) isopropoxide; ― TnS-Pz; ― Pz. b The indirect interband transition energies of the TnS-Pz samples. “A” in the label on the y-axis stands for absorbance, which is proportional to the absorption coefficient α. --- titanium(IV) isopropoxide; --- TnS-Pz; --- Pz

Full size image

The DRUV spectra yielded the estimated bandgap energies as presented in Fig. 2a. The relation between the absorption coefficient (α) and incident photon energy (hν) follows Eq. (1).

$$ \alpha =\frac{{\mathrm{B}}_{\mathrm{i}}\ {\left(\mathrm{h}\nu -{\mathrm{E}}_{\mathrm{g}}\right)}^2}{\lambda } $$
(1)

where Bi is the absorption constant for indirect transitions (Lei et al. 2014a). The plots of (Ahν)1/2 versus hν from the spectral data in Fig. 2a are presented in Fig. 2b. By extrapolating the linear section of the curve (the traced line in Fig. 2b), the material’s bandgap may be indirectly estimated, where a bandgap of around 2.36 eV for hybrid material was calculated (TnS-Pz). Moreover, wherein bandgap energies lower than 3.00 eV correspond to visible-light absorption (Lei et al. 2014a), the bandgap energies of titanium(IV) isopropoxide and natural pozzolan were 2.81 and 2.60 eV, respectively.

FTIR was used to analyze the functional groups present in TnS-Pz and natural pozzolan, both of whose band assignments of both materials are displayed in Table 1. There is a higher number of vibrational bands recorded in the TnS-Pz material than in natural pozzolan, chiefly in 1826 to 2531 cm−1 spectrum. The intense bands located at 3433 and 3331 cm−1 are assignable to hydroxyl stretching and the outer surface hydroxyl vibration of the pozzolan and TnS-Pz (Huang et al. 2017a; Wamba et al. 2017; Zeitler and Brown 1957). Similarly, the broad bands located around 1013 and 1058 cm−1 may be assigned to the Si–O–Si stretching vibration present in the pozzolan and TnS-Pz, respectively. Nevertheless, while the bands at 2923 and 2977 cm–1 correspond to the stretching vibration of –CH2 itself (Wamba et al. 2017), they are observed only in the natural pozzolan band at 958 cm−1, as assigned to the Si–O–Ti vibration (Huang et al. 2017a; Zeitler and Brown 1957). This vibration is only found in the TnS-Pz sample. The observations also confirm the hybridization process.

Table 1 The main FTIR vibrational bands in natural pozzolan and TnS-Pz. All assignments are based according to the literature [24, 28, 29]
Full size table

A scanning electron microscope (SEM) (Fig. 3) was used to examine the morphologies of the natural pozzolan and TnS-Pz. Figure 3a shows the natural pozzolan structure, composed of irregular particle sizes and regular smooth surfaces lacking evident porosity, thereby accounting for the low SBET (13.2 m2 g−1). However, the SEM image from the TnS-Pz material (Fig. 3b) indicates a large number of uneven thin nanosheet shapes, explained by the significant increased porosity (SBET = 271.7 m2 g−1). This modification might correlate to the chemical bond between the natural pozzolan and titanium(IV) isopropoxide (Si–O–Ti and Al–O–Ti) in the hybridization process, as well as by the accumulation of Ti nanoparticles over the surface TnS-Pz in the formed nanosheets.

Fig. 3
figure 3

a, b SEM images of pure pozzolan and TnS-Pz. c TEM images of pure pozzolan and TnS-Pz. d Chemical map for Ti-K energy

Full size image

A TEM analysis was carried out for further morphologic investigation of the synthesized TnS-Pz (Fig. 3c), showing the morphology of the TnS-Pz where the agglomerated structures and nanoplatelets/nanosheets were a minimum of one direction lower than 100 nm in width size.

An energy-dispersive X-ray (EDX) was also used to further analyze the composition of TnS-Pz. The elemental map (Fig. 3d) confirms the presence of Ti particles covering the natural pozzolan and concurs with the proposed structure of the synthetized TnS-Pz.

In the XRD patterns of the TnS-Pz sample (Fig. S4), the main contribution originates from the diopside/diopside aluminian phase (88.7%) and its monoclinic structure containing a C2/c space group. The diopside is a member of the larger pyroxene group (M2M1Si2O6), where the cations may occupy the two non-equivalent sites of both M2 and M1. These sites’ geometry depends on the composition and the space group of the pyroxene member (Biedermann et al. 2015; De Grave and Eeckhout Sigrid 2003; Dyar et al. 2013). The smaller M1 site is sixfold coordinated site—the octahedron is nearly regular and may be occupied by smaller cations (such as Mg2+, Fe2+, Fe3+, Al3+, or Ti4+). The strongly irregular M2 site ranges from six (spodumene) to eightfold (diopside, augite, aergine) larger (Biedermann et al. 2015). The pyroxenes that crystallize in the monoclinic crystal system are clinopyroxenes; the Ca2+-saturated C2/c end-members are diopside (CaMgSi2O6) and hedenbergite (CaFeSi2O6); as well, the intermediate C2/c solid solution is an augite (nominally (Ca, Na) (Mg, Fe, Al, Ti) (Si,Al)2O6). The Ca2+ has a strong preference for the larger M2 site in the synthetic diopside. In natural samples, the substitution of Ca by Na and Mg by Fe is common. When a larger M2 site is vacant, the Fe2+ ions prefer the M2 site; Fe2+ and Fe3+ are otherwise located in smaller M1 sites (Biedermann et al. 2015; De Grave and Eeckhout Sigrid 2003; Dyar et al. 2013).

The minor contribution in the XRD profiles of TnS-Pz sample corresponds to the Ca(Al2Si2O8) compound comprised of a hexagonal structure of titanium oxides (anatase, rutile), and iron (titanium) oxides (hematite and titanomagnetite). The titanomagnetites (Fe3-xTixO4, 0 ≤ x ≤ 1) are solid solutions of magnetite and ulvöspinel (Banerjee et al. 1967; Bowles et al. 2013; Hamdeh et al. 1999; Liu et al. 2014; M. Hassan and Dekan 2013; Sorescu et al. 2012; Tanaka and Kono 1987). The magnetite (x = 0, stoichiometric) has an inverse spinel crystal structure according to the formula (Fe3+)A(Fe3+ Fe2+)B O2−4. The A-sites (tetrahedral sites) are occupied by Fe3+ ions and the B-sites (octahedral sites) by the Fe3+ and Fe2+ ions. At temperatures above the Verwey transition temperature (~ 119 K), a rapid electron exchange occurs between the Fe2+ and Fe3+ ions at the B-sites where a mixed 2.5 + iron valence state is detected (Gunnlaugsson et al. 2008; Kündig and Steven Hargrove 1969; Sorescu et al. 2012). In intermediate compositions that have a more complex spinel structure, Ti4+ ions replace some of the Fe3+ ions on the B-sites; therefore, in order to maintain charge neutrality, another Fe3+ ion at A- or B-site must convert to a Fe2+ ion. The other end member, ulvöspinel (x = 1), is also an inverse spinel, as presented by the formula (Fe2+)A (Fe2+ Ti4+)B O2−4 (Sorescu et al. 2012).

Due to the presence of these disparate phases in the sample where overlapping Mössbauer parameters occur, there is a consequent complex spectra fitted with several Mössbauer subspectra: doublets (two lines of the absorption feature) and sextets (six lines of the absorption feature). The 57Fe-Mössbauer spectrum of the TnS-Pz sample is shown in Fig. 4, together with the total fit and a resolved subspectrum.

Fig. 4
figure 4

The room temperature 57Fe-Mössbauer spectra of the TnS-Pz sample. The solid circles represent the experimental data and the red solid line represents the corresponding fit. The fitted lines of the Mössbauer subspectra are plotted above the main spectrum fit—sextets D (blue), E (black), F (dark cyan), G (orange), H (red), J (magenta), and K (black); doublets A (dark cyan), B (pink), and C (blue). The vertical arrow denotes the relative position of the lowest peak with respect to the basal line. The red solid line in the lower area of the plot represents the error calculated as the difference (Fit-Exp). The largest value of the absolute difference is less than 0.08%

Full size image

The fitted Mössbauer parameters of the subspectrum are presented in Table 2. Numerous initial attempts to fit the spectrum included several sextets, ferric and ferrous doublets, constrained and non-constrained. The initial number of components and fitting constraints was based on prior Mössbauer tests on the Fe-bearing phases present in the samples. As reported in the literature, for the central analyzed spectrum, one or two ferric doublets found for the anatase (Cordischi et al. 1985; Patel et al. 2012; Rodríguez-Torres et al. 2008; Zhu et al. 2005) and rutile (Balcells et al. 2006; Cabrera et al. 2006; Cordischi et al. 1985; Nozik 1972; Sandin et al. 1976; Stampel et al. 1973), two to five ferric and/or ferrous doublets for diopside (Biedermann et al. 2015; De Grave and Eeckhout Sigrid 2003; Dyar et al. 2013; J. G. Stevens 2005; Van Alboom et al. 1991) and augite (Biedermann et al. 2015; J. G. Stevens 2005; Whipple 1974) as well as for other members of the clinopyroxene group (Akasaka 1983; Biedermann et al. 2015; Eeckhout Sigrid and Grave 2003; J. G. Stevens 2005; Redhammer Günther et al. 2006) were considered to be fits. For the magnetic area, one sextet for the hematite (Dyar et al. 2006; J. G. Stevens 2005), two sextets for the magnetite (Dyar et al. 2006; J. G. Stevens 2005; Kündig and Steven Hargrove 1969; M. da Costa et al. 1995), and numerous other studies have detected multiple doublets and/or sextets for titanomagnetite (Banerjee et al. 1967; Bowles et al. 2013; Gattacceca et al. 2013; Hamdeh et al. 1999; J. Sitek 2013; Liu et al. 2014; M. Hassan and Dekan 2013; Morris et al. 1993; Sorescu et al. 2012; Tanaka and Kono 1987); as such, they were taken into account during the fitting procedure. To the best of the authors’ knowledge, there are no reported Mössbauer parameters for the Ca(Al2Si2O8) compound containing a hexagonal structure as found in the TnS-Pz sample. Due to the origin of this phase as an intermediate phase in the thermal transformation of the Ca-LTA zeolite structure into anorthite (Dimitrijević et al. 1996), the measured spectrum may contain ferric and/or ferrous doublets related to this phase resembling those found for other aluminosilicate phases (J. G. Stevens 2005).

Table 2 Room temperature 57Fe-Mössbauer hyperfine parameters for the TnS-Pz sample. A, the relative area of the Mössbauer subspectrum; Γ, line width (FWHM); δ, isomer shift; Δ, quadrupole splitting in a pure electric interaction; 2ε, the quadrupole shift in a combined strong magnetic and weak electric interaction; Bhf, the hyperfine magnetic field. The fitting errors are given in parenthesis
Full size table

Based on the features seen in the fits made incorporating and not incorporating constraints, two key decisions were made after the preliminary fits. The first distinctive attribute was observed for the central section of TnS-Pz spectrum. Due to the overlapping of the Mössbauer lower velocity absorption lines for the ferric and ferrous doublets corresponding to the anatase, rutile, high titanomagnetite, diopside, and augite as well as to the hexagonal Ca(Al2Si2O8) phase, the initial larger number of doublets decreased to only five constrained ones that could be fairly resolved. The constraints imposed were the same line width (FWHM) for the ferric and the same FWHM for the ferrous doublets. The larger number of doublets of imposed constraints neither pointed to any specific pairs of doublets for the abovementioned phases nor provided any reasonable assignments to the observed phases. Therefore, all constraints for the central section of the final fits were lifted. A limited number of the initial doublets merged after lifting; hence, only three doublets in total could be retained in the final fits. Important features in the measured spectrum were also found for the magnetic section at the ambient temperature. There were highly characteristic of spectral features, two sextets with Bhf ~ 49 T; δ ~ 0.3 mms−1 and Bhf ~ 46 T; δ ~ 0.63 mms−1, which generally correspond to the magnetite phase for the A- and B-sites, respectively. These are yet also of the ratio of relative subspectrum areas 2:1 instead of 1:2 (Kündig and Steven Hargrove 1969; M. da Costa et al. 1995). The sizeable line width of the A-site sextet was recorded as well. One interpretation of these lines is that there may be another phase or more phases that contribute to the A-site sextet of the Bhf ~ 49 T in addition to the magnetite phase (Gunnlaugsson et al. 2006; Gunnlaugsson et al. 2003). As the oxidation of titanomagnetite to titanomaghemite (i.e., the “maghemitization” process) is commonly found in nature (de Oliveira et al. 2002; Tanaka and Kono 1987; Xu et al. 1997), the titanomaghemite phase was also taken into account for the evaluation of TnS-Pz spectrum. Other research into titanomagnetite (M. Hassan and Dekan 2013; Sorescu et al. 2012; Tanaka and Kono 1987) has noted the appearance of one or two new sextets accompanied by the simultaneously lowering of hyperfine magnetic fields. Moreover, the broadening of A- and B-site sextets was also been noted with an increase of Ti content in the Ti-doped magnetite phase. More sextets were added to the area of the Bhf ~ 49 T sextet, whereby the range of Bhf ~ 48–52 T described the final fit of three sextets. Moreover, for the TnS-Pz sample, following the preliminary fits, magnetic components were observed below the Bhf ~ 48 T that reached nearly zero quadrupole shifts. These shifts were set to zero in the final fit (Table 2). As regards interpreting the TnS-Pz spectrum in the Bhf ~ 47–52 T range, the fully oxidized magnetite—i.e., the maghemite phase (Fe3+)A (Fe3+5/3□1/3)B O2−4 that is also an inverse spinel (□ denote vacancy)—lacks Fe2+ ions in the structure. This fact was also taken into consideration. Since the A- and B-site sextets had approximate values of the hyperfine magnetic fields, only one sextet of Bhf ~ 50 T (δ ~ 0.32 mms−1) was found in the maghemite spectrum at room temperature (Dyar et al. 2006; J. G. Stevens 2005). Partially oxidized magnetite or the mixture of magnetite and maghemite is able to contribute to the spectra within the same range of a hyperfine parameter (Fock et al. 2017; Joos et al. 2016). The non-stoichiometric Fe3-xO4 contains vacancies presumably on the B-sites in relation to the amount of the Fe2+ in the magnetite structure (A. Gorski and Scherer 2010; Kalska-Szostko et al. 2015) and may have been formed in the sample. From an experimental point of view, since stoichiometric/non-stoichiometric magnetite and maghemite have quite similar inverse spinel structures, it is often difficult to distinguish between the two crystallographic structures or their mixtures since X-ray diffraction is less sensitive than the Mössbauer technique (Fock et al. 2017). In addition to the Ti4+ ions, impurity ions such as Al3+, Ca2+, Mg2+, and Si4+ can be substituted for Fe3+ ions in hematite and maghemite. These impurities contribute to chemical and structural disorder in the crystal structure, leading to additional sextets or distributions of hyperfine magnetic fields in the Mössbauer spectra. Similar distributions can be found for natural (titano)magnetite samples, where the same impurity ions commonly substitute Fe3+ and/or Fe2+ (Bowles et al. 2013). Therefore, for the evaluation of TnS-Pz spectrum, the Mössbauer parameters for the doped hematite, magnetite, and maghemite phases were also considered.

The assignment of the Mössbauer subspectrum presented in Table 2 was done as follows: the Fe3+ D-sextets for the TnS-Pz sample belong to the hematite/titanohematite phase (Dyar et al. 2006; J. G. Stevens 2005; McEnroe et al. 2016). The titanohematites are solid solutions between the two end-members of hematite (α-Fe2O3) and ilmenite (FeTiO3) (McEnroe et al. 2001). The formation of a titanohematite phase must be taken into consideration due to the presence of Ti4+ ions embedded in the titanium(IV) isopropoxide starting material used to produce the TnS-Pz sample. While the Fe3+ E-sextets dominantly originate from the titanohematite phase (McEnroe et al. 2016), minor contributions from the maghemite (J. G. Stevens 2005) and titanomagnetite cannot be entirely excluded (McCammon et al. 1999; Sorescu et al. 2012). The titanomagnetites contributed to the F–J-sextets, where a substantial range of hyperfine magnetic fields (Bhf ~ 40–52 T) and considerable line widths were all recorded. As such, this does indicate that either more than one titanomagnetite phase (Sorescu et al. 2012; Tanaka and Kono 1987) coupled by a non-stoichiometric magnetite may contribute to these sextets or that additional doped hematite (McEnroe et al. 2016) and pure/doped maghemite (Allan et al. 1988; J. G. Stevens 2005; Xu et al. 1997) do so as well. For the TnS-Pz sample, the titanomaghemite may contribute substantially to the sextet due to the fact that the F-sextet exhibited a pure Fe3+ character (Chevrier et al. 2006). The very broad K-sextet may be ascribed only to the titanomagnetites (Sorescu et al. 2012), but the pyrrhotite phase was unable to be excluded from the results of the Mössbauer investigation (Gattacceca et al. 2013). The G–K-sextets with the Fe2.5+ characters indicate Fe-Ti spinel-type phases similar to the magnetite structure of the electron hopping between the octahedral sites.

The three doublets located in the central section of the spectrum partially originate from the dominant diopside/diopside aluminum phase. Similar Mössbauer parameters for the Fe3+ A-doublet were found for the diopside (De Grave and Eeckhout Sigrid 2003; McCanta and Dyar 2017) where the ferric ions usually prefer the M1 site in the pyroxenes. Since it is not commonly found for diopsides (Dyar et al. 2006; J. G. Stevens 2005), a similar doublet was described in a more general manner as Fe3+ in the pyroxenes (Chevrier et al. 2006; Gunnlaugsson et al. 2006; Gunnlaugsson et al. 2003) or as Fe3+ at the M1 site in the clinopyroxene (Akasaka 1983; Dyar et al. 1989). One possible reason for the A-doublet’s major broadening is that it is unable to be assigned solely to one specific phase. Ferric ions embedded in titan oxides also contribute to the same absorption area in the TnS-Pz spectrum (Balcells et al. 2006; Cabrera et al. 2006; Nozik 1972; Patel et al. 2012; Sandin et al. 1976; Stampel et al. 1973; Zhu et al. 2005). The A-doublet may be partially associated with the various superparamagnetic iron oxide/iron hydroxide nanoparticles (smaller than 10–15 nm), which are not easily distinguishable in the Mössbauer spectrum at ambient temperatures (Joos et al. 2016). The Fe2+ B-doublet and Fe2+ C-doublet dominantly belong to the diopside/diopside aluminum phase, where ferrous ions are most likely positioned at the M2 and M1 sites, respectively. Verifying the assignment, similar doublets for the Fe2+ at the M2 and M1 sites were found for the titanaugite (Whipple 1974) and clinopyroxene (Dyar et al. 1989; Dyar et al. 2013). At a minimum, the minor contributions of the high timagnetites (Sorescu et al. 2012) to the Fe2+ paramagnetic section of the TnS-Pz spectrum would be covered by the clinopyroxene’s doublets.

Only an estimate may be given that iron ions, such as (titano)magnetites and (titano)hematite, are embedded more in the magnetic phases present in the Fe-bearing phases present in the sample.

Cr(VI) photoreduction studies

The photocatalytic activity of the TnS-Pz was evaluated through the photoreduction of 0.2 mM of Cr(VI) under UVA-Vis radiation. The experiments were performed in the presence of citric acid as a sacrificial agent in a 3:1 [scavenger]0/[Cr(VI)]0 M ratio, with a catalyst load of 15 mg L−1, pH 3.0, and 25 °C. The preliminary tests showed that the Cr(VI) reduction was negligible after 180 min under UVA-Vis radiation in the absence of catalyst where there was no addition of sacrificial agents (Fig. S5). Likewise, the same behavior was observed in the system with Cr(VI) in presence of only citric acid (Fig. S5). The slight (10%) Cr(VI) photocatalytic reduction observed in the absence of scavenging agents after 180 min of irradiation (Fig. S5) may be chiefly attributed to a short circuiting caused by the continuous reduction and reoxidation of chromium species by holes or hydroxyl radicals when the scavenger agent is not present in the system (Marinho et al. 2017b). The reaction in the presence of citric acid and UVA-Vis (absent of catalysts) leads to a Cr(VI) reduction of 30% in 180 min, whose behavior is explainable by the citric acid oxidation through the photo-induced electron transfer (PET) that acts as electron donor (Mytych and Stasicka 2004).

In order to compare the effectiveness of the titanium(IV) isopropoxide/pozzolan nanosheets (TnS-Pz) in the Cr(VI) photoreduction, the reaction with titanium(IV) isopropoxide nanosheets (TnS) and pure TiO2-P25 was assessed (Fig. 5). The TnS-Pz presented a higher efficiency in the Cr(VI) reduction than the pure TiO2-P25. Moreover, in the TnS-Pz/citric acid/UVA-Vis system, the initial Cr(VI) concentration almost reached complete diminishment after 150 min. However, in its reaction with pure TiO2 (TiO2-P25/citric acid/UVA-Vis system), only 21% of Cr(VI) reduced during the same reaction time. In addition, a tenfold increase on the photocatalytic reduction rate (pseudo first-order kinetic) was found when using TnS-Pz (k = 11 × 10−3 min−1) as a photocatalyst when compared with the TiO2-P25 (k = 1.1 × 10−3 min−1) (Table S1).

Fig. 5
figure 5

Cr(VI) photoreduction over different photocatalysts. () TiO2-P25/citric acid/UVA-Vis, () TnS/citric acid/UVA-Vis, () TnS-PZ/citric acid/UVA-Vis. [Cr(VI)]0 = 0.2 mM; catalyst loading = 15 mg L−1; [citric acid]0/[Cr(VI)]0 M ratio = 3:1; pH = 3.0; and 25 °C

Full size image

Owing to the fact that the system’s radiation source emits up to 440 nm (Fig. S3), the better efficiency of the TnS-Pz may be attributed to the DRUV analysis of the catalyst as well as that the indirect bandgap energy determined was the lowest (EgindirectTnS-Pz = 2.36 eV). The TiO2-P25 demonstrated no satisfactory results under the UVA-Vis, presenting an indirect bandgap value of 3.10 eV (Guayaquil et al. 2017) that is higher than the synthesized hybrid material.

It is important to note that the Cr(total) concentration remained constant (data not shown) in all experiments, thereby indicating no chromium(III) adsorption on the catalyst’s surface.

The effect of pH on the photocatalytic reduction of the Cr(VI) by TnS-Pz/UVA-Vis/citric acid system

The solution pH is crucial to the Cr(VI) reduction in the photocatalytic system with TiO2 (Marinho et al. 2017a; Marinho et al. 2017b). Nevertheless, in order to verify the influence of pH on the TnS-Pz/UVA-Vis/citric acid system, a set of reduction experiments were carried out with 0.2 mM of the Cr(VI), [citric acid]0/[Cr(VI)]0 M ratio = 3:1, 25 °C, varying the pH from 3.0 to 6.0 (Fig. 6). The results shown in Fig. 6 and Table S1 present that, as the pH increases, the reaction rate also substantially decreases. This respective reaction demonstrates that the pH plays a significant role in Cr(VI) reduction. In the pH 3.0 system, a complete reduction (below the detection limit) was achieved in 180 min, while for the pH 4.0 and 5.0, the reduction achieved was roughly 75% over the same time, and the pH 6.0 reaction reduced less than 20% of the Cr(VI). The kinetic constant presented the same behavior, decreasing from 11 × 10−3 to 1.3 × 10−3 min−1 when the pH was increased from 3.0 to 6.0, respectively. These results may stem from the global reduction reaction, which consumes protons and is also favored at lower pH values (Marinho et al. 2016).

Fig. 6
figure 6

The influence of pH on the photocatalytic reduction of Cr(VI) in the TnS-Pz/UVA-Vis/citric acid system at 25 °C. [Cr(VI)]0 = 0.2 mM; catalyst loading = 15 mg L−1; [citric acid]0/[Cr(VI)]0 M ratio = 3:1; pH: () 3.0, () 4.0, () 5.0, () 6.0

Full size image

The effect of the initial concentration of Cr(VI) on the photocatalytic reduction of Cr(VI) by the TnS-Pz/UVA-Vis/citric acid system

The effect of the initial Cr(VI) concentration on its photoreduction was evaluated by varying the corresponding concentration from 0.02 to 0.20 mM at pH 3.0 and 25 °C where the citric acid was proportional at a 3:1 [citric acid]0:[Cr(VI)]0 M ratio. As depicted in Fig. 7, the Cr(VI) reduction’s efficiency gradually decreases while matching the increase of the respective initial concentration. For the lowest concentrations studied (0.02 mM), a total reduction of Cr(VI) was achieved after a reaction time of 45 min. However, for the initial concentrations of the Cr(VI) (0.20 mM), approximately 95% of reduction was achieved in 180 min.

Fig. 7
figure 7

The influence of the initial Cr (VI) concentration on the photocatalytic reduction in the TnS-PZ/UVA-Vis/citric acid system at pH 3.0 and 25 °C. Catalyst loading = 15 mg L−1; [citric acid]0/[Cr(VI)]0 M ratio = 3:1; [Cr (VI)]0: () 0.20 mM, () 0.15 mM, () 0.10 mM (), 0.04 mM, () 0.02 mM

Full size image

Accompanied by the tenfold increase of the initial concentration of Cr(VI) from 0.02 to 0.20 mM, Table S1 shows that the reaction rate decreased approximately 5-fold (k = 52 × 10−3 and k = 11 × 10−3 min−1). Such behavior has been recorded in other studies on TiO2-P25 in suspension (Ku and Jung 2001) and verified (Marinho et al. 2017a), where the initial concentration of Cr(VI) is tripled, causing a 3 to 13-fold reduction in the reaction rate times. The negative effect of higher concentrations of Cr(VI) can be attributed to the surface’s saturation of the catalyst by the Cr(VI) species, thereby preventing the binding of the citric acid species, as similar effects have been reported previously by Chakrabarti et al. (2009) and Valari et al. (2015). In addition, the competition between the Cr(VI) and the citric acid species may occur for the binding sites on the surface of the catalyst, generating a screen effect, where the Cr(VI) is able to impede light entry onto the surface of the catalyst (TnS-Pz), which, in response, decreases the photocatalytic reduction efficiency of Cr(VI).

The effect of the scavenger (by type) on the photocatalytic reduction of Cr(VI) by the TnS-Pz/UVA-Vis system

The photocatalytic Cr(VI) reduction by TiO2 was reported to be enhanced through the addition of tartaric acid as sacrificial agents (Marinho et al. 2017b). Therefore, this scavenging agent was tested on a Cr(VI) (0.2 mM) reduction using the TnS-Pz/UVA-Vis system (15 mg L−1 of TnS-Pz, at pH 3.0 and 25 °C).

As presented in Fig. 8, the results using both scavenger agents were similar as they attained a total Cr(VI) reduction (below the DL of the method) after 180 min of reaction time.

Fig. 8
figure 8

The influence of scavenger type on the photocatalytic reduction of Cr (VI) in the TnS-PZ/UVA-Vis system at pH 3.0 and 25 °C. [Cr (VI)]0 = 0.20 mM, catalyst loading = 15 mg L−1; [scavenger]0/[Cr(VI)]0 M ratio = 3:1; scavenger: () citric acid, () tartaric acid

Full size image

These similar results likely stem from the organic acids being able to oxidize in the catalytic system, resulting in the formation of highly reduced radicals (such as HCOO•−). The ability also allows for an indirect radical-mediated reaction in the solution bulk, thereby intensifying the Cr(VI) reduction.

In the heterogeneous photocatalysis using TiO2, the better efficiency of the tartaric acid systems may be explained by the higher energy of the highest occupied molecular orbital (EHOMO) and the lower adiabatic ionization potential (AIP) of the tartaric acid as it favors the electron transfer within the TiO2 organic acid complex (Wang et al. 2010). However, these thermodynamic driven forces do not appear to be essential in the TnS-Pz/UVA-Vis system.

Conclusion

This work presents the synthetization of a new hybrid material by the sol-gel process using natural pozzolan and titanium(IV) isopropoxide as precursors. Ti–O–Si and Ti–O–Al covalent bonds between the titanium(IV) isopropoxide and pozzolan are shown to be formed during the sol-gel process yielding the hybrid material. The surface area was found to be 271.7 m2 g−1 and of a mesoporous structure. The morphology studies (SEM and TEM) showed that the material presents a hybrid nanosheet structure. The ultraviolet-visible diffuse reflectance (DRUV) absorption detected a varied electronic structure of the material whose material bandgap was estimated as 2.36 eV, which confirms the hybridization process. Comprehensive structural analysis (XRD) has shown the main mineral constituents of hybrid material (Al/Si/Fe/Ti oxides) and confirmed that structural properties—the chemical states on the activated surface sites (MB)—determined their affinity of forming hydroxyl agents.

The Cr(VI) reduction was successfully achieved with the TnS-Pz hybrid material, in the presence of hole/hydroxyl radical scavenger and UVA-Vis radiation. When optimized, a complete 0.2 mM Cr(VI) reduction was achieved after 180 min of reaction time in the presence of a 15 mg L−1 catalyst load, [tartaric acid]0/[Cr(VI)]0 M ratio = 3:1, pH 3, and 25 °C under UVA-Vis radiation.