Modulation of the Bifunctional CrVI to CrIII Photoreduction and Adsorption Capacity in ZrIV and TiIV Benchmark Metal-Organic Frameworks

The presence of hexavalent chromium water pollution is a growing global concern. Among the currently applied technologies to remove CrVI, its adsorption and photocatalytic reduction to CrIII less mobile and toxic forms are the most appealing because of their simplicity, reusability, and low energy consumption. However, little attention has been paid to bifunctional catalysts, that is, materials that can reduce CrVI to CrIII and retain both hexavalent and trivalent chromium species at the same time. In this work, the dual CrVI adsorption–reduction capacity of two iconic photoactive water-stable zirconium and titanium-based metal–organic frameworks (MOFs) has been investigated: UiO-66-NH2 and MIL-125. The bifunctionality of photoactive MOFs depends on different parameters, such as the particle size in MIL-125 or organic linker functionalization/defective positions in UiO-66 type sorbents. For instance, the presence of organic linker defects in UiO-66 has shown to be detrimental for the chromium photoreduction but beneficial for the retention of the CrIII phototransformed species. Both compounds are able to retain from 90 to 98% of the initial chromium present at acidic solutions as well as immobilize the reduced CrIII species, demonstrating the suitability of the materials for CrVI environmental remediation. In addition, it has been demonstrated that adsorption can be carried out also in a continuous flux mode through a diluted photoactive MOF/sand chromatographic column. The obtained results open the perspective to assess the bifunctional sorption and photoreduction ability of a plethora of MOF materials that have been applied for chromium capture and photoreduction purposes. In parallel, this work opens the perspective to develop specific chemical encoding strategies within MOFs to transfer this bifunctionality to other related water remediation applications.


Introduction
Photocatalytic reduction of hexavalent chromium is becoming an environmentally friendly and easily applicable technology to remove highly toxic and mobile chromate oxyanions from polluted water sources [1][2][3][4][5]. Large research efforts have been devoted to demonstrate the efficiency, low cost, and easy implementation of different photocatalysts to reduce hexavalent chromium to trivalent less mobile and toxic species [6][7][8]. Nevertheless, the scientific community has paid less attention to assess whether the photocatalysts are able to immobilize or adsorb the trivalent chromium cations [9] at the time that hexavalent species are photoreduced [10,11].
When approaching water remediation from a circular economy point of view, the conjunction of photocatalysis and adsorption functionalities in the same material is of paramount importance. For instance, at acidic conditions, trivalent chromium ions are still soluble in solution, and henceforth, they do not precipitate as chromium hydroxides/oxides after the photocatalytic reduction of Cr VI [12]. At these acidic solutions, an additional immobilization step is necessary to extract Cr III ions from the aqueous media. Here, the importance of bifunctional photocatalysts that are able to transform and retain chromium species is revealed, since they open the possibility to couple environmental remediation (photocatalysis) to chromium recovery (adsorption) [13]. It is important to note that chromium is currently located at the borderline of the European raw materials map because of its economic importance [14,15]. Indeed, the circular water remediation concept gains even more relevance if it is applied in highly concentrated and acidic water effluents from industrial (e.g., plating, alloying, tanning, textile dyes, pigments . . . ) and mining activities (e.g., acidic water streams) [16], since chromium recovery from these media becomes economically appealing [17]. The problem with acid rising waters derived from industrial platting or tanning activities is even more challenging than the revalorization of exhausted baths. In this case, chromium is highly diluted in large volumes of water, and its recovery by the usual technologies applied to concentrated exhausted electrolytic solutions (e.g., as liquid-liquid extraction, ion exchange and/or electrochemical separation [18] or precipitation, acid leaching, and subsequent precipitation [19,20]) is neither efficient nor economically appealing to revalorize the chromium ions. Diluted hexavalent chromium solutions require high-cost chemicals for its reduction to Cr III until the legal maximum concentration limits for Cr VI (ground water = 0.1 ppm, drinking water = 0.05 ppms) are accomplished. This scenario is where bifunctional photocatalyst sorbents could play a key role to face at the same time the water treatment coupled with the chromium recovery.
At the materials design level, a bifunctional photocatalyst applied for chromium recovery needs to be specifically tailored to (i) capture light and transform it to hole-electron pairs able to generate reactive oxygen radicals, and at the same time, (ii) to adsorb the chemical species transformed during photocatalysis. Therefore, efficient light harvesting [21], semiconducting nature [22], porosity [23][24][25], and chemical affinity to capture the transformed species are some of the key characteristics that need to be specifically encoded within the desired bifunctional catalyst [25]. Classic metal-oxide photocatalysts exhibit excellent photodegradation activities over a variety of pollutants, but their capacity to adsorb or retain the photodegradation intermediates and products is limited by its reduced porosity and surface areas [26][27][28][29]. In this context, metal-organic frameworks (MOFs) stand out as highly porous materials that are able to meet photoactivity and porosity. MOFs can be formed from semiconductor metal-oxide clusters and efficient light-harvesting organic linkers assembled in extended and ordered porous structures [27,30]. The combination of these two characteristics, together with the possibility to further decorate or engineer their structures, both pre and post synthetically, make MOFs the ideal candidates to combine photoreduction and adsorption properties [21,[31][32][33].
Indeed, MOFs have been widely applied for chromium adsorption, photoreduction, and detection purposes, but their combined capacity to photoreduce and at the same time adsorb the phototransformed trivalent chromium ions has been rarely explored [34]. For instance, the photocatalytic activity of UiO-66-NH 2 and MIL-125 to transform Cr VI to Cr III has been proved under varied experimental conditions (e.g., chromate concentration, photocatalyst loading, source of illumination, pH of the media, presence of competitor species...). In addition, our previous study points out that UiO-66-NH 2 is able to sorb quite efficiently both hexavalent and trivalent species from solution, exhibiting also a moderate capacity to chemically reduce the Cr VI to Cr III under dark conditions [35,36]. For instance, the chromium speciation within the Zr-MOF sorbents and photocatalysts has been successfully determined in our previous research by applying a multiple UV-Vis and electron paramagnetic resonance spectroscopy technique. Our findings confirmed that Cr VI to Cr III chemical and photoreduction processes are driven by the intermediate generation of pentavalent chromium transient species inside UiO-66-based materials [37] Nevertheless, this research does not unravel the fate of trivalent chromium phototransformed species in solution. This experimental gap-that is, the monitoring of both hexavalent and trivalent chromium in solution during working conditions-is common for most of the investigations carried out in Cr VI to Cr III photoreduction experiments. Most typically, during the MOF photocatalysis process, the concentration of the hexavalent chromium is experimentally determined by UV-Vis, overriding the monitoring of the trivalent chromium ions that could be still present after the photoreduction. In this work, we have fulfilled this gap, researching whether the photoactive MOFs can also act as efficient sorbents of the photoreduced species. The photoreduction and adsorptive capacities of two iconic photoactive water-stable zirconium and titanium-based MOFs, UiO-66 and MIL-125, have been addressed. For both materials, the experimental monitoring of Cr VI and Cr III in solution during and after the operation has been determined. Going a step forward, photoactive chromatographic columns based on UiO-66-NH 2 have been assembled to perform Cr VI adsorption in a continuous flux mode. Our conclusions clearly point out that the studied bifunctional MOFs are able to completely photoreduce Cr VI to Cr III and at the same time immobilize reduce Cr III species.

Metal-Organic Frameworks Selection, Synthesis, and Characterization
Among the varied MOFs that could potentially exhibit bifunctional photocatalytic and adsorption activities, zirconium amino-terephthalate (UiO-66-NH 2 ) and titanium terephthalate (MIL-125) compounds have been selected. Both MOFs exhibit a cubic structure with an "fcu" topology arising from the connectivity between the inorganic hexanuclear clusters through twelve terephthalate type organic linkers [38,39].
Further, the effect on the photocatalytic and adsorptive properties of (i) the particle size of MIL-125 material and (ii) the organic linker defect density of UiO-66-NH 2 have been studied. MIL-125 has been synthesized as micron-size particles applying hydrothermal synthesis (here and after denoted as MIL-125-H) and as nanoparticles (100-400 nm) under reflux conditions (here and after identified as MIL-125-R). Despite the bulk to surface proton coupled transfer in MIL-125 having been confirmed as highly independent of the particle size [30], it has been observed that the Cr VI to Cr III photoreduction kinetics are greatly affected by this parameter. In parallel, the density of organic linker defects in UiO-66-NH 2 materials has been modulated by the addition of hydrochloric acid to the reaction media. The addition of acid modulators to the synthesis media of Zr-based UiO-66-NH 2 generates a random displacement of amino-terephthalate organic linkers along its "fcu" crystal structure [31,40,41]. Two are the side effects of increasing the density of linker defects on the UiO-66-NH 2 properties: first, an increase of the overall surface area and pore volume, and second, the generation of preferential chemisorption positions for inorganic oxyanions [42][43][44]. Therefore, "defect-free" UiO-66-NH 2 and linker defective UiO-66-NH 2 -def materials were synthesized at the same hydrothermal conditions but adding a controlled volume of HCl in the specific case of the defective sample.
Once synthesized, the quality, particle size, porosity, and optical properties of the materials were fully characterized by X-ray diffraction, UV-Vis spectroscopy, thermogravimetric analysis, transmission electron microscopy, and surface area measurements.
The X-ray diffraction (XRD) patterns of the synthesized samples are fully consistent with the simulated patterns obtained from UiO-66 and MIL-125 [38] structural models, as can be observed in Figure 1a. Profile pattern analyses ( Figure S1) confirm that there are no meaningful displacements of the cell parameters due to the generation of linker defects in UiO-66-NH 2 -def, neither of which are associated to the particle size reduction from MIL-125-H to MIL-125-R (Table S1). XRD patterns show a significant increase of the peak width in the specific case of the samples with smaller particle size, which is ascribed to the reduction of the crystalline domains as a consequence of the synthesis conditions. This experimental evidence is confirmed by the micrographs obtained by means of transmission electron microscopy, as shown in Figure 1d. Indeed, for zirconium amino-terephthalate compounds, the addition of the HCl inorganic modulator induces an increase of the crystal size from 10-20 nm in non-defective UiO-66-NH 2 to 50-100 nm in UiO-66-NH 2 defective samples. A similar particle size dependence on the reaction temperature is found for titanium amino-terephthalate compounds. MIL-125-H, which was synthesized at 150 • C under hydrothermal conditions, exhibits a micron-size particle, whilst MIL-125-R, which was obtained from reflux conditions, crystallizes as spherical particles with diameters ranging from 100 to 400 nm. The light-harvesting capacity of the materials was studied by means of solid UV-Vis spectroscopy ( Figure 1b). The results show that UiO-66 and MIL-125 materials present a very different behavior: while UiO-66 adsorbs in the visible range, the MIL-125 materials adsorb in the UV. UV-Vis spectra of UiO-66-NH 2 samples exhibit two main absorption bands ascribed to the visible light absorption by the amino-terephthalic organic linkers and UV-light harvesting driven by the zirconium inorganic clusters, respectively. As expected, the presence of linker defects in UiO-66-NH 2 -def induces a blue shift displacement of the absorption band associated with the amino-terephthalate building units (UiO-66-NH 2 = 2.76 eV, UiO-66-NH 2 -def = 2.82 eV). The different synthesis conditions (from MIL-125-H to MIL-125-R) induce a slight red shift to the charge-transfer absorption band associated to the titanium oxo-nuclear clusters. The surface area and pore size distribution were determined from CO 2 adsorption isotherms at 273 K ( Figure 1c). The Brunauer-Emmett-Teller (BET) surface area was calculated from the fitting of the linearized form of the BET equation ( Figure S4). The obtained surface areas and microporous distribution (UiO-66 = 906 m 2 /g, UiO-66-NH 2 = 894 m 2 /g, UiO-66-NH 2 -def = 600 m 2 /g, MIL-125-H = 872 m 2 /g, MIL-125-R = 1346 m 2 /g) are in good agreement with previous reports [45,46]. The thermogravimetric analyses shown in Figure S2 allow quantifying the linker defect degree of UiO-66 materials applying the protocol previously described by Shearer et al. [40]. TGA curves of UiO-66-NH 2 compounds show three main weight step losses ascribed to the release of the solvent entrapped within the pores (RT-100 • C), hydroxyl and water species coordinated to the inorganic clusters (150-350 • C), and calcination of the amino-terephthalate organic linkers (350-450 • C). The linker defect degree of UiO-66 samples can be estimated based on the weight loss percentage associated with the organic linker (see SI, Figure S3 and Table S2, for more detailed information). In the specific case of the UiO-66-NH 2 and UiO-66-NH 2def synthesized in this work, 0.6 and 2.2 positions over the twelve carboxyl groups per inorganic clusters present in the ideal structure are vacant. The thermogravimetric analysis curves of MIL-125-H and MIL-125-R compounds show two main weight loss processes associated to (i) the release of water molecules entrapped within the pores (RT-100 • C) and (ii) the calcination of the amino-terephthalate linkers (350-480 • C).
According to that initial characterization, there would be an expected improved photoreduction efficiency for the samples with lower particle size as a result of the shorter migration path to the surface of the electron-hole pairs, which will lead to a more efficient generation of the reactive species that trigger the photoreduction. However, other parameters need to be considered in parallel. The light-harvesting capacity is also a key indicator of the photocatalytic efficiency dependence on the radiation wavelength (e.g., UVA, visible . . . ) of our materials. Finally, linker functionalities of MOFs, such as amine groups, also favor the electron transfer along the framework while improving the chromium adsorption capacity.

Adsorption Kinetics of Cr VI and Cr III Ions
First, the adsorption affinity of UiO-66-NH 2 and MIL-125 samples over Cr VI and Cr III species was assessed by the evaluation of their adsorption kinetics over Cr VI and Cr III solutions at acidic conditions (pH = 3.5) ( Figure 2). As expected from previous works, Zr-based UiO-66-NH 2 materials show fast kinetics and efficiencies over 90% to sorb both hexavalent and trivalent chromium species. For hexavalent chromium, the equilibrium is reached within the first 15 min of the experiment. In agreement with previous works [37], linker defective UiO-66-NH 2 -def shows higher chromate adsorption capacity than the non-defective UiO-66-NH 2 (Figure 2a). It is well established that linker defective positions in Zr-based MOFs act as preferential adsorption points for anionic inorganic and organic species. For instance, hexavalent chromium is stabilized in a wide range of pH values as dichromate or hydrogen dichromate oxyanions [8,47]. On the contrary, MIL-125 materials show a negligible adsorption affinity over negative chromate anions, as expected because of the lack of defective linker positions of titanium terephthalate framework (Figure 2a). Zr-UiO-66-NH 2 and MIL-125 compounds also exhibit adsorption over trivalent chromium, with adsorption capacities ranging from 8 to 15 mg/g for a 5 ppm solution. It is also worthwhile to notice the great impact that the particle size reduction has on the Cr III adsorption capacity of MIL-125 type materials, since nano-sized MIL-125-R overperform the adsorbing capacity of its micron-size MIL-125-H homologue. Opposite, linker defective UiO-66-NH 2 -def materials have larger particles size than that of the non-defective homologue, but their Cr VI absorption capacity overperforms that of the UiO-66-NH 2 materials. In this case, the Cr VI adsorption process is driven by the inner linker defects density instead by surface defects induced by the particle size reduction. In comparison to hexavalent chromate, Cr III adsorption kinetics are slower, with equilibrium times above one hour. The time dependence of Cr VI and Cr III adsorption is well fitted by Bangham and pseudo-second order models (Figure 2b and Figure S5), while the pseudo-first-order model fails to reproduce the tendency above 60 min of adsorption, indicating that Cr VI and Cr III immobilization is not solely governed by diffusion processes, but also by chemisorption and/or chemical Cr VI to Cr III reduction processes into the MOF structure. Indeed, both the obtained adsorption capacity (1/q e for a pseudo-secondorder model and m for Bangham's model) and kinetic rates (1/V 0 for pseudo-second-order model and K T for Bangham's model) obtained from both models agree with the previous discussion regarding the adsorption capacity and affinity over Cr VI and Cr III species of the studied materials. The parameters obtained from the experimental fittings are summarized in Table 1 and allow quantifying the adsorption capacity and affinity of the studied materials over the main chromium species that will be involved in the photocatalysis experiments. As could be observed, some of the results are better fit by the Bangham model and others are better fit by the pseudo-second order model as a result of their different behavior toward chromium adsorption. Table 1. Parameters obtained from the fitting of the Cr VI and Cr III adsorption kinetics shown in Figure 2 by applying the Bangham and pseudo-second-order models.

UiO-66-NH 2 Chromatographic Columns: Continuous Flux Adsorption
When designing continuous flux adsorption experiment, it is necessary to integrate the MOF in a highly permeable chromatographic column that at the same time maximizes the contact between the aqueous solution and the active MOF sorbent. First, the protocol described by Rapti et al. [48] was followed to assemble a UiO-66-NH 2 /sand (1% w/w) based chromatographic column. Its adsorptive capacity was over Cr VI . Results are shown in Figure 3.
The breakthrough adsorption experiment was carried out in dark conditions, with an initial Cr VI concentration of 2.5 ppm and a 1 mL/min flux (Figure 3a). Experimental data for the adsorption stage in the open system was fitted to Thomas and Yoon-Nelson models, but results were not correctly fitted by these models [49]. It is important to point out that the kinetics of adsorption differ significantly in static and continuous flux experiments. Indeed, in static mode, most of the Cr VI adsorption occurs within the five initial minutes, while in continuous adsorption experiments, the sorbent saturation point is reached after 10 h.
Afterwards, we decided to increase the MOF content on the chromatographic column and evaluate its performance under a continuous flux process in a closed loop (see Figure S6 for a detailed information). By applying this strategy, the cumulative effect of the adsorption process in the long-term can be experimentally observed. Henceforth, 10% w/w UiO-66-NH 2 /sand chromatographic columns were prepared, and 100 mL Cr VI solutions of 2.5, 30, and 50 ppm concentrations were pumped at 1 mL/min flux until the adsorption equilibrium was reached. For the experiment performed with a 2.5 ppm concentration, the column was able to adsorb all the Cr VI content in the 100 mL pumped across the column in 400 min. In the specific case of highly concentrated 50 ppm solution, the UiO-66-NH 2 /sand column was able to retain almost 40% of the initial chromium in the adsorption stage. Indeed, a slight increase of the chromium content in the solution was observed after the saturation stage of the material. It is important to point that the color of the column changes appreciably during the experiments, suggesting that the material possess certain degree to induce a chemical reduction of Cr VI to Cr III . Finally, for the experiment performed at an intermediate concentration (Cr VI = 30 ppms), the system was able to capture near 50% under dark conditions. It is clear that the time to reach the adsorption saturation stage is four-fold longer in the continuous flux conditions than in static adsorption experiments. It is interesting to note that the maximum adsorption capacities for the UiO-66-NH 2 material obtained from static sand adsorption experiments are very close. Indeed, 27 mg/g for static experiments, and 27 and 29 mg/g for continuous flux experiments were developed with 30 and 50 ppm solutions. Therefore, just by applying a continuous flux adsorption experiment with column containing 1 g of UiO-66-NH 2 , 12 L of water with 2.5 Cr VI ppm concentration could be depolluted in a closed system.

Bifunctional Photocatalytic Reduction of Cr VI to Cr III and Coupled Adsorption of Cr III Ions
Once the capacity of the different MOFs to adsorb Cr III and Cr VI species has been assessed, their Cr VI to Cr III photocatalytic reduction capacity was investigated. Thus, photoreduction curves were first measured under UVA light in a 5 ppm Cr VI solution and at a sorbent loading of 0.25 g/L (Figure 4). A first adsorption stage in dark reveals that UiO-66-NH 2 type materials are able to capture more efficiently Cr VI ions than the MIL-125 family, which is in good agreement with the Cr VI adsorption kinetics presented above. Nonetheless, a different trend is observed upon illumination, where MIL-125-R exhibits the most efficient photoreduction response and a kinetic rate compared to UiO-66 materials (Figure 4c,d), which can be related with its adsorption in the UVA range. On the contrary, the photoreduction capacity of MIL-125-H is appreciably slower. Even if the lightharvesting capacity is quite similar for both MIL125-R and MIL-125-H samples (Figure 1b), the shorter migration path to the surface of the electron-hole pairs in MIL-125-R nanoparticles with respect to the bigger ones for the MIL-125-H decreases their recombination, and consequently, enhances the generation of the reactive species that trigger the photoreduction process (Figure 4).
In addition to the good photoreduction kinetics of MIL-125 materials, they also exhibit Cr III adsorption capacity. The Cr III concentration after the photoreduction process with MIL-125-R and MIL-125-H is quite low (<0.5 ppm), but it is still higher than the observed for UiO66-NH 2 materials (<0.1 ppm). Thus, in terms of photoreduction capacity plus Cr III recovery, UiO-66 materials show the best combined performance. In addition to a high Cr VI initial adsorption capacity, amino functionalized UiO-66 materials exhibit interesting kinetics for Cr VI to Cr III photoreduction, combined with a negligible loss of the generated Cr III , due to their high adsorption capacity toward this cationic species.
The conclusions drawn by the photocatalytic experiments open relevant questions to properly understand the underlying processes that governs the photocatalytic and adsorptive dual function of the studied materials. In particular, it is important to address how the materials work under visible light illumination with the aim of evaluating their possible performance under solar light, as well as determining how the initial Cr VI concentration just before the illumination stage affects the photoreduction kinetics.
In order to gain further insights regarding the above-mentioned issues, UiO-66-type materials have been selected because of their similarity in terms of crystal structure characteristics, allowing extracting meaningful conclusions from the experimental results. First, their photocatalytic capacity under visible light has been investigated. In order to enhance the photoreduction kinetics under visible light, the charge of the photocatalyst has been increased from the 0.25 g/L used in the previous experiments to 0.35 g/L. All the results are shown in Figure 5. The increase on the MOF's photocatalysts loading generates a higher Cr VI retention in the dark, being the most efficient adsorbent of the UiO-66-NH 2 -def material, followed by UiO-66 and UiO-66-NH 2 . Regarding the photocatalytic process in visible light (Figure 5a), it is observed that amine functionalities play again an important role in the chromium photoreduction compared with UiO-66 non-functionalized samples as a result of their improved light harvesting (Figure 1b). Amino decorated UiO-66-NH 2 materials exhibit higher efficiencies and faster kinetics (Figure 5c) than UiO-66, achieving 100% (UiO-66-NH 2 -def) and 60% (UiO-66-NH 2 ) photocatalytic efficiencies in about 1 h. This indicates the feasibility of applying these photocatalysts under solar irradiation. Regarding Cr III , a small increase in the concentration of this species is observed in the solution during the photoreduction process, but it is still very low (≈0.1 ppm) when compared to the starting Cr VI concentration.
Once it was confirmed that the photoactivity and adsorption-coupled capacity of UiO-66-NH 2 materials are still functional under visible light illumination, chromium photoreduction experiments were performed starting from a similar Cr VI concentration before illumination (Figure 5b). To this end, the initial Cr VI concentration was increased until a value close to 3 ppms was obtained after the initial adsorption stage. Thus, the photocatalytic kinetics were compared starting from similar conditions. As it could be observed in Figure 5b, when the photoreduction process starts from a similar initial concentration, the UiO-66-NH 2 photocatalyst presents improved photoreduction kinetics, which is followed by UiO-66-NH 2 -def and UiO-66 materials. It is important to note that the calculated photocatalytic rates (Figure 5d) are quite different from those obtained from the previous experiments (Figure 5c), where the Cr VI concentration after adsorption differs significantly for the three studied compounds. Indeed, generically speaking, the higher the Cr VI concentration before the illumination stage, the lower the photocatalytic rate.
The photoreduction kinetic rates obtained from Figure 4b,d reveal that this parameter is affected by the particles size and by the light-harvesting capacity in the visible, as well as by the experimental conditions themselves. Independently of the experimental conditions, overall, the Cr III adsorption efficiencies are above 98%, although a slight increment of the Cr III concentration in solution has been observed under certain experimental conditions. Further investigations are needed to elucidate the activation and reusability of the studied materials, as well as the Cr III saturation point under photocatalysis experiments.

Synthesis of Metal-Organic Framework Photocatalysts
UiO-66, UiO-66-NH 2 , and UiO-66-NH 2 -def were prepared through a slightly modified solvothermal synthesis previously reported [41,43]. First, zirconium chloride (Aldrich, San Luis, MO, USA) (0.5418 g) was dissolved in 60 mL of dimethylformamide (DMF, Aldrich, San Luis, MO, USA)) under stirring in a Pyrex ® (Fisher Scientific, Waltham, MA, USA) autoclave. Afterwards, terephthalic (0.38012 g, BDC, Aldrich, San Luis, MO, USA) or amino terephthalic acid (0.4185 g, BDC-NH 2 , Aldrich, San Luis, MO, USA) and distilled water (1.5 mL) were added to the zirconium chloride solution under continuous stirring in order to obtain the UiO-66 and UiO-66-NH 2 samples. UiO-66-NH 2 -def sample synthesis requires the addition of 1 mL of HCl (37% v/v, Aldrich, San Luis, MO, USA) solution to the ZrCl 4 and amino terephthalic DMF mixture. Once a clear solution is obtained, the Pyrex ® reactor was closed and placed in a preheated oven (80 • C for 24 h). After that, the sample was recovered by centrifugation and washed three times overnight with methanol. Finally, the compound was dried at 80 • C for 12 h. An MIL-125-H sample was prepared following the hydrothermal protocol previously established by Dan-Hardi et al. [38]. Reflux synthesis of MIL-125-R was developed applying a protocol adapted from ref [50]. First, 7.70 g of BDC was dissolved in 100 mL of DMF under reflux and continuous stirring at 100 • C for 1 h. Then, 28 mL of methanol (99.9 %, Aldrich, San Luis, MO, USA) was added to the solution and further stirred for 1 h. Afterwards, 8.4 mL of Ti(iPrO) 4 (97 %, Aldrich, San Luis, MO, USA) was added drop by drop to prevent the formation of agglomerates.
Samples were refluxed for three days, recovered, and thoroughly washed with DMF and ethanol to remove the non-reacted reagents.

Characterization Protocols
The MOF photocatalysts were fully characterized by means of XRD, CO 2 adsorption isotherms, electron transmission microscopy, and UV-Vis spectroscopy. The results show that the structural, textural, spectroscopic, optical, and thermal properties are in good agreement with previously reported studies. More detailed information about the characterization protocols can be found in the Supplementary Information.

Adsorption Kinetics
Cr VI and Cr III adsorption kinetics were conducted in the same conditions as those of the photoreduction experiments (5 ppm Cr VI or Cr III solution, 0.25 g/L of adsorbent) but in the absence of light. Then, 5 mL aliquots were taken from the solution at different time intervals until equilibrium. The quantification of Cr VI was performed based on the previously reported diphenyl carbazide colorimetric methodology [51]. Cr III estimation was achieved after oxidizing it to Cr VI and applying the same colorimetric protocol. For that, 5 mL of the chromium solution, 100 µL of a 3.39M H 2 SO 4 (98%, Aldrich, San Luis, MO, USA) solution, and 100 µL of a 22.6 mM solution were added. The mixture was heated at 100 • C for 20 min and then cooled down to room temperature in a water bath. Afterwards, 50 µL of a NaN 3 solution were added to the mixture until the purple color of the potassium permanganate disappears from the solution. Finally, diphenyl carbazide colorimetric methodology was used to quantify the chromium content. Results were fitted to Bangham and pseudo-first and second-order kinetic models [52]. A concentration of 5 ppm was chosen as a model of diluted raising water derived from the electrolytic bath cleaning. This concentration is slightly higher than that usually found in rives and underground waters polluted by the tanning or planting industry (1 to 0.5 ppms).

Photocatalytic Reduction of Cr VI to Cr III
Photocatalysis experiments were conducted in a 5 ppm Cr VI solution using 0.25 g/L and under UVA and Vis light. First, 50 mg of the photocatalyst was immersed in 200 mL solution of Cr VI 5 ppm solution under stirring in the dark. Once an adsorption equilibrium condition was achieved, 5 mL aliquots were taken at different times under illuminated conditions. Cr VI and Cr III quantification was carried out as previously described.

Continuous Flux Adsorption Experiments
Open chromatographic columns of 1% weight of UiO-66-NH 2 (0.0125 g) in sand (1.25 g) were assembled as illustrated in Figure S6. A continuous flux of 1 mL/min of a chromium solution of 2.5 ppm concentration was passed through the column, and the Cr VI concentration was monitored during the process. Closed columns were assembled from a UiO-66-NH 2 (10% w/w, 0.125 g) diluted in 1.25 g of sand. Then, 100 mL of a 1 mL/min flux of 2.5, 30, and 50 ppm chromium solution was pumped through the column until the chromium concentration reaches a stable value, that is, the MOF has reached its saturation point.

Conclusions
The dual photocatalytic and adsorptive functionality of archetypal Zr and Ti-based metal-organic frameworks over the simultaneous Cr VI to Cr III photoreduction and Cr III immobilization has been demonstrated. The particle size reduction of Ti-based MIL-125 has been revealed as a key parameter to enhance the photoactivity of the material, whilst the linker defective chemistry of Zr-based UiO-66 compounds is detrimental for the chromium photoreduction but beneficial for the retention of the Cr III phototransformed species. Considered as a whole, the bifunctional Zr and Ti-based MOFs are able to retain from 90 to 98% of the initial chromium present at acidic solutions. A photodegradation and adsorption coupled process can be performed efficiently on a continuous flux closed loop circuit when the MOF active material is assembled as a chromatographic column. The experimental procedure described in this work opens the future perspective to test the bifunctional catalytic and adsorptive capacity of photoactive MOFs synthesized so far, not just applied for the specific case of chromium, but also for homologous studies such as the As III to As V arsenic photooxidation or emerging pollutant photodegradation coupled to the adsorption of the photogenerated intermediate products. However, it is important to note that further research is needed to consider in a quantitative manner the economic and technical feasibility of MOFs for chromium recovery in real scenarios far from ideal laboratory models. Many questions are raised from the current work, such as (1) to which extend the MOF chemistry is applicable far from diluted rising industrial waters, that is, in high concentration industrial solutions as exhausted metal acidic baths, (2) or at intermediate slightly acidic solutions obtained after the neutralization of the initial bath electrolytes. (3) Can the chromium recovery from the MOF system leading to Cr III solution of 20 g/L concentration be reused in plating electrolyte media? Furthermore, could the process be performed in a repetitive way without damaging the framework and properties of our hybrid inorganic-organic systems?
Supplementary Materials: The following are available online at https://www.mdpi.com/2073-4 344/11/1/51/s1, Description of Materials and processing methods, Characterization techniques and procedures, Performance assessment, DRX, TGA, and BET fittings (Figures S1-S5). Table S1: Cell parameters determined by XRD and Table S2: Calculated weight loss related to ligand calcination step of TGA analysis.  Data Availability Statement: Data available on request due to restrictions. The data presented in this study are available on request from the corresponding author. The data are not publicly available due to possible future exploitation actions of the results presented in this study.