Hybrid biopolymer/metal–organic framework 3D-sponges towards the capture of the 'big five' heavy metals

Hybrid biopolymer/metal–organic framework 3D-sponges towards the capture of the ‘big five’ heavy metals María Calles García a , Hugo Salazar a , Sylvia Britto b , Oleksandr Tomchuk b , Pedro M. Martins c,d , Arunava Pradhan c,d , Fernanda C´ assio c,d , Senentxu Lanceros Mendez a,e , Koro de la Caba a,f , Pedro Guerrero a,f,g,* , Viktor Petrenko a,e,* , Roberto Fern´ andez de Luis a,* a BCMaterials, Basque Center for Materials, Applications and Nanostructures, UPV/EHU Science Park, 48940, Leioa, Spain b ISIS Neutron and Muon Source, Science and Technology Facilities Council, Rutherford Appleton Laboratory, Didcot, OX11 0QX, United Kingdom c Centre of Molecular and Environmental Biology, University of Minho, 4710-057, Braga, Portugal d IB-S – Institute for Research and Innovation on Bio-Sustainability, University of Minho, 4710-057, Braga, Portugal e IKERBASQUE, Basque Foundation for Science, 48009, Bilbao, Spain f BIOMAT Research Group, University of the Basque Country (UPV/EHU), Escuela de Ingeniería de Gipuzkoa, Plaza de Europa 1, 20018 Donostia-San Sebasti´ an, Spain g Proteinmat Materials SL, Avenida de Tolosa 72, 20018, Donostia-San Sebasti´ an, Spain ARTICLE INFO Keywords: Adsorption Biopolymers Filters Heavy metals Metal-organic frameworks ABSTRACT Access to clean water in isolated regions remains a major challenge, particularly due to contamination by the five most prevalent heavy metals: Hg(II), Pb(II), Cd(II), As(III/V), and Cr(VI). Traditional sorbents are limited in their ability to capture all the “big five” heavy metals, since they occur as cationic, neutral, or anionic species under standard conditions. To address this challenge, we have integrated a thiol rich Zr(IV)- Metal-Organic Framework (MOF), namely BCM-1, into a soy protein (SPI) and chitin (CHI) sponge in order to engineer a 3D-hybrid water filter. The components and the composite systems were thoroughly characterised by conventional means. Additionally, neutron imaging was used to reveal the 3D-interconnected microto macroporous structure of the filters, while Small-Angle X-ray Scattering (SAXS) confirmed the presence of BCM-1 as monodisperse nanoparticles. The 3D-sponge combines mechanical stability, high permeability, and broad chemical affinity, allowing the efficient removal of all five heavy metals through simple adjustments of its activation conditions. Adsorption experiments demonstrated over 90 % removal for most target metals depending if the hybrid-sponge is employed as synthesised, or after activating at pH =1. When tested with 1 ppm solutions, they exhibit adsorption efficiencies for Hg(II), Pb(II), Cd(II), As(V), and Cr(VI) of 60.8/100 %, 94.4/74.8 %, 15.7/69.1 %, 100/38.2 %, 5.7/ 100 %, and 13.5/97.4 %, before and after the activation of the 3D-sponge, respectively. The metrics are consistently maintained over three adsorption/desorption cycles in surface water samples. On the whole, this work provides a scalable and sustainable approach to combine biopolymers and MOFs for real-world water remediation applications and highlights the key role of their protonation state on their absorptive properties. 1. Introduction Heavy metal pollution remains a real threat to the local environment and populations on these isolated areas that lack water treatment plants. Here, the development of portable adsorption systems emerges as one of the most promising technologies to access potable/drinking water. The rational design of an adsorbent is intrinsically governed by the physicochemical properties of the target metal species to be removed from the environment. In environmental science, the term “heavy metal” is commonly associated with the acute toxicity of these ions. In light of * Corresponding authors at: BCMaterials, Basque Center for Materials, Applications and Nanostructures, UPV/EHU Science Park, 48940, Leioa, Spain E-mail addresses: [email protected] (M. Calles García), [email protected] (H. Salazar), [email protected] (S. Britto), oleksandr. [email protected] (O. Tomchuk), [email protected] (P.M. Martins), [email protected] (A. Pradhan), [email protected] (F. C´ assio), [email protected] (S. Lanceros Mendez), [email protected] (K. de la Caba), [email protected] (P. Guerrero), viktor. [email protected] (V. Petrenko), [email protected] (R. Fern´ andez de Luis). Contents lists available at ScienceDirect Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej https://doi.org/10.1016/j.cej.2025.169442 Received 3 July 2025; Received in revised form 25 September 2025; Accepted 8 October 2025 Chemical Engineering Journal 524 (2025) 169442 Available online 13 October 2025 1385-8947/© 2025 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC license ( http://creativecommons.org/licenses/bync/4.0/ ). this, the ‘big five’ members of this family—chromium, arsenic, cadmium, mercury, and lead—are readily identifiable due to their impact in human health and environmental harm. In fact, there are several scenarios where some of the “big five” heavy metals are concurrently present in polluted waters, as in the case with industrial effluents, acid mine drainages or textile and tannery wastewater. From a technological standpoint, designing an adsorption system able to concurrently capture the ‘big five’ heavy metals is a challenge owing to the diverse speciation that the “big five” members show in water media. [1]. In the usual pH/ eH conditions found in polluted waters (i.e. pH 4–9 and eH –200 to +600 mV) [2], Pb(II) and Cd(II) are usually stabilized as isolated cations, Hg(II) and As(III) tend to form non-charged HgCl 2 and As(OH) 3 species, and As(V) and Cr(VI) are complexed as H 2 AsO 4 − /HAsO 4 2− and HCrO 4 − /Cr 2 O 7 −2 oxyanions, respectively. All in all, the concurrent capture of the “big five” requires the development of a multifaceted sorbent able to trap cationic, neutral, and anionic species through a combination of adsorption mechanisms. In this context, MOFs have emerged as highly promising microporous sorbents, exhibiting exceptional capacities and kinetics for the sequestration of heavy metals [3]. Zr-MOFs stand out among the MOF -family because of their hydrolytic and chemical stability, at the same time that show an outstanding framework tunability. These features together enable the selective capture of anionic species at defect sites within the structure, while cationic or neutral species may be effectively absorbed through chemisorption or electrostatic interactions at the functional groups installed at the linker sites [4]. However, despite their chemical versatility, no single Zr-MOF has yet demonstrated the capacity to capture of all the ‘big five’ members. Moreover, the inherent powdered nature of MOFs continues to pose a significant barrier to their application in practical water treatment [5]. When MOF powders are employed in a bed-configuration, they usually need to be pelletized or shaped to prevent a backpressure increase during the water permeation. At the same time, shaping MOFs is usually linked to a partial decrease of their surface area, which is in turn associated to crystallinity-loss or surface inactivation arising from the processing conditions (i.e. pressurization, temperature, solvents…). [6] The inclusion of MOFs into classic polymer matrices (i.e. PVDF for its chemical resistance [7], Polyurethanes for flexibility [8], polysulfone (PSU) and polyethersulfone (PES) for thermal stability [9], and polyamide (PA) for selective permeability [10]) is one of the potential strategies explored to overcome this limitation. Most usually, synthetic polymers provide mechanical support for MOF particles adding negligible or secondary absorption functions to the composite system. By using conventional polymer processing techniques—such as molding, casting, or electrospinning, a wide range of composites has been assembled and applied for water remediation, including heavy metals adsorption. However, the design of MOF/polymer composites still faces limitations that restrict their practical application, including structural heterogeneities from non-uniform MOF distribution within the polymer—such as partial encapsulation [11]—and limited permeability for their application in continuous-flow systems. This is precisely where the integration of MOFs into biopolymeric composites comes into play. Biopolymers usually are easier and greener to process than synthetic polymers. In addition, depending on their surface chemistry, they show an active role to chelate metal ions. Their easy processing has paved the way for the design of three-dimensional composites featuring interconnected microto macroporous structures as aerogels, monoliths, and 3D-printed structures. In addition to exhibit mechanical robustness, MOF/biopolymer composites posses at the same time synergic absorption functions arising from the combination of biopolymer and MOF components (Scheme 1). [7,12] While previous studies have reported the integration of MOFs with natural biopolymers for heavy metal adsorption, such as chitosan–Zr-MOF cryogels or dualfunctionalised polymer/MOF sponges for targeted ion removal, [8,9] these often focus on a limited number of metal species and lack studied on their simultaneous capture. Notable examples of successful composites include chitosan and iron metal-organic framework sponges, which demonstrated exceptional adsorption capacities of 2857–2326 mg/g for organic pollutants, or 3D printed porous chitosan/UIO-66 filters based on carboxyl acids functionalized alginate and chitosan, which have shown outstanding heavy metal removal capabilities. [13] Another promising system involves Fe/MOF-5@chitosan composite films synthesised by doping 5 wt% of Fe/MOF-5 into chitosan films, which combine the mechanical properties of the biopolymer matrix with the high surface area and selectivity of the MOF component. [14] In contrast to synthetic polymers, the chemical versatility and green processability of MOF/biopolymer composites are counterbalanced by a partial limitation in their chemical stability window, which restricts the range of conditions under which these composites can be applied. Nevertheless, they remain highly suitable for the typical environments of surface and groundwater sources contaminated with heavy metals. In this study, we have leveraged the rich chemical functionality of soy protein isolate (SPI) and chitin (CHI) biopolymers to construct a metal-chelator MOF-functionalised polymeric 3D-sponge capable of simultaneously capturing all the ‘big five’ heavy metals [15]. This performance is attributed to the presence of amino, hydroxyl, and carbonyl functional groups inherent to the SPI and CHI matrices, in conjunction with the defective and thiol-rich architecture of the zirconium mercaptosuccinate MOF—hereafter referred to as BCM-1 [16]. The thiol and amino rich structures of the BCM-1 and SPICHI components assures the efficient capture of soft to intermediate cationic and neutral heavy metal species; whilst the linker-defective positions in BCM-1 endows the filter with the capacity to capture oxyanions after its activation under acidic conditions. Multiple characterisation techniques have been employed to fully understand the atomic to macro-metric scale structure of our system, including X-ray diffraction (XRD), small-angle X-ray scattering and Scheme 1. Strategy employed in this research. M. Calles García et al. Chemical Engineering Journal 524 (2025) 169442 2 neutron imaging. Thus, our research has uncovered the beneficial and detrimental adsorption effects arising from the combination of these three components in a highly permeable sponge technology, identifying the potential routes to fully unblock its potential for its application as heavy metal sorbents in real scenarios. 2. Materials and methods 2.1. BCM-1 synthesis BCM-1 was prepared by scaling up tenfold the hydrothermal synthesis reported in the previous study of our research team [17]. To this end, zirconium chloride (2.33 g, 10 mmol) was dissolved in 10 mL of distilled water under magnetic stirring in a 50 mL Pyrex ® reactor. Afterwards, 570 μ L of formic acid (FA) was added under continuous stirring at room temperature. Lastly, thiomalic acid was added to the metal solution while stirring vigorously until a clear solution was obtained. The reactor was sealed and placed in a preheated oven at 120 ◦C for 24 h. After the crystallisation of BCM-1 was completed, the reactor was allowed to cool down to room temperature and the MOF powder was recovered by centrifugation (Hettich EBA21, 6500 rpm, 20 min) and washed with water (50 mL, three times) overnight. Finally, the sample was dried at 80 ◦C for 12 h. Another three samples, two of BCM-1 cubic variant, and one of the BCM-2 hexagonal variant, were synthesised by varying the addition of FA into the reaction media to modulate the particle size and the final structure of the zirconium mercaptosuccinate MOFs. [18,19] The two BCM-1 samples were exclusively employed to assess their ecotoxicity in case of leaching during adsorption. BCM-2 sample was employed to elucidate whether the crystal structure of the MOF affects its eco-toxicity. 2.2. Preparation of SPICHI@BCM-1 composite 3D-sponges It is important to note that the content of BCM-1 in the membranes was varied by controlling its addition in the composite filters’ formulation. For the preparation of SPICHI@BCM-1 filter with a 6 wt% of BCM-1 (hereafter denoted as SPICHI@BCM-1(6 %)), firstly, 5 g of SPI, 0.5 g of BCM-1 (6 wt%) and 1.5 g of CHI (30 wt% based on the SPI dry basis) were mixed in 125 mL of distilled water and stirred for 30 min. Then 30 wt% glycerol (based on the SPI dry basis) was added to the solution. The pH was adjusted to a value of 7 with a NaOH (1 M) solution to later heat it at 80 ◦C for 30 min under magnetic stirring. Finally, the blend was poured into moulds, kept in a freezer at −22 ◦C for 48 h, and freeze-dried for 72 h to obtain the composite filters. The same protocol was followed for the preparation of SPICHI@BCM-1(20 %), but 2 g BCM1 were added during the formulation of the blend. 2.3. Structural, textural and functional characterisation The experimental details of the physic-chemical characterisation protocols (i.e. XRD, Scanning Electron Microscopy (SEM), Fourier Transform Infrared Spectroscopy (FTIR), Thermogravimetric Analysis (TGA), Differential Scanning Calorimetry (DSC)) employed for the characterisation of BCM-1, SPICHI, and SPICHI@BCM-1 are detailed in the Supporting Information. Below, we describe the experimental conditions for SAXS, neutron imaging and ecotoxicity experiments. Finally, we outline the protocols for evaluating the metal adsorption of the microporous BCM-1, the SPICHI biopolymer, and the SPICHI@BCM-1 composite filters. 2.3.1. Small-angle X-ray scattering SAXS measurements were conducted on I22 beamline at Diamond Light Source (Didcot, U.K.) [20]. Data were collected with a Pilatus P3–2 M detector. SAXS data were reduced and azimuthally averaged to obtain the isotropic scattering intensity as a function of scattering vector modulus q =4 π /λ sin(θ/2), where θ is the scattering angle and λ =1 Å is the X-ray wavelength, using the DAWN software package [21,22]. All scattering curves were fitted according to functions provided by SasView 6.0.1, including Indirect Fourier Transformation analysis routine. 2.3.2. Neutron imaging Neutron tomography measurements were carried out at IMAT, the neutron imaging instrument at the ISIS Neutron and Muon Source (Didcot, U.K.). Samples were wrapped in aluminium foil and placed inside an aluminium tube, which was mounted on a rotation stage. Tomographic projections were acquired by rotating the sample over 360 o , with angular steps chosen to satisfy the Nyquist criterion. The instrument geometry was configured to a collimation ratio (L/D) of 260 where L is the distance from the pinhole to the detector and D is the pinhole aperture. Images were taken with the Andor Ikon-L 936 CCD camera coupled with a 105 mm lens and an 80 μ m thick ZnS/6LiF scintillator providing a pixel size of 48 μ m. An exposure time of 60 s was used for each projection. Flat-field images (recorded without the sample in the beam) and dark-field images (recorded with the beam off) were acquired for image normalisation. Tomographic reconstructions of the projection data were carried out using the Mantid Imaging [18] software package. [23] Porosity analysis was performed using Porespy. [24] For the porosity profile analysis, a ROI that falls within the sample area was first selected for the analysis. Image segmentation was then done using Otsu thresholding before calculation of percentage porosity. 2.3.3. Assessment of heavy metals adsorption kinetics and capacity The pH values of the stock heavy metals solutions, having a concentration of 100 ppm, were measured and found to be within the range of 5.5 to 6.6 (As(III) =6.6, As(V) =5.9, Cr(VI) =4.8, Hg(II) =6.4, Cd (II) =6.1, Pb(II) =5.5). The pH was approximately maintained within the same value range when the solutions were diluted to perform the adsorption experiments described in the following. The metal speciation dependence on the pH, eV and metal concentration was calculated with Hydra-Medusa software. 2.3.3.1. Preliminary adsorption. Experiments were conducted at room temperature with an adsorbent dosage of 1 mg mL −1 . For a conventional batch experiment, around 10 mg of the MOF, SPICHI or SPICHI@MOF were immersed in 10 mL of metal solution. After a given time (4 h), the suspension was centrifuged, the supernatant was collected and then filtered using a hydrophilic 0.22 μ m filter. The metal concentration in the preand post-adsorption conditions was analysed by Inductively Coupled Plasma Atomic Emission Spectroscopy (ICP −AES). 2.3.3.2. Adsorption isotherms. 10 mg of BCM-1, SPICHI, or SPICHI@BCM-1(6 and 20 %) were immersed in 10 mL of heavy metal solutions (i.e. NaAsO 2 , H 3 AsO 4 , K 2 Cr 2 O 7 , HgCl 2 , Cd(NO 3 ) 2 ⋅4H 2 O and Pb (NO 3 ) 2 ) with concentrations ranging from 0.5 to 2000 ppm. The isotherms were fitted to Langmuir and Freundlich models. 2.3.3.3. Kinetic experiments. 50 mg of BCM-1, SPICHI, or SPICHI@MOF were added to 50 mL of the heavy metal solutions at a concentration of 10 ppm. The solutions were stirred for 4 h to ensure that the adsorption equilibrium was reached. Aliquots of the samples were taken at different times of the experiments to follow up on the heavy metal concentration decay during adsorption. The kinetic curves were fitted according to pseudo-first, pseudo-second order, and intraparticle diffusion non-linear forms described in Eqs. (1), (2), and (3), respectively [25]: Qt=Qe(1−e(− K1t))(1) Qt=Q2 cK2t 1+QcK2t(2) Qt=kidt 1 / 2 +C(3) M. Calles García et al. Chemical Engineering Journal 524 (2025) 169442 3 where Q e and Q t (mg/g) are the adsorption capacities of heavy metals at equilibrium and at a specific time, respectively. K 1 (min −1 ) is the pseudo-first order adsorption rate constant, K 2 (g/mg min)) is the pseudo-second order adsorption rate constant, K id is the intraparticle diffusion rate constant (mg/g min -1/2 ), t is the time (min), and C is a parameter related to the boundary layer effect. The adsorption capacity (q e ) of BCM-1, SPICHI, and SPICHI@BCM-1 was determined based on the Eq. (4): qe=(Co−Ceq)×VT (m)TOTAL ≡mg g(4) where V T denotes the volume of the solution expressed in millilitres (mL), m indicates the mass of the adsorbent in grams (g), and C 0 and C e represent the initial and equilibrium concentrations of the metals within the solution, respectively. 2.3.3.4. Reusability. The adsorption and desorption efficiency of SPICHI@BCM-1(20 %) was studied over 3 cycles. A solution of all the five heavy metals – having a concentration of 1 ppm of each of the individual ions - was prepared (pH =6.0). For the desorption and regeneration, SPICHI@BCM-1(20 %) was immersed into a 0.1 M HCl solution and stirred for 2 h. Afterwards, an aliquot was withdrawn to quantify the concentration of the heavy metals in the elution solution; and in turn, to calculate the efficiency of the desorption. SPICHI@BCM-1(20 %) was then immersed into ultrapure water and stirred for another 2 h. After this washing step, employed to neutralize the pH of the membrane, the filter was employed in a new absorption cycle. Fig. 1. Illustration of the crystal structures of (a) BCM-1, (b) chitin and (c) soy protein; (d) XRD patterns of BCM-1 (1), SPICHI (2), SPICHI@BCM-1(6 %) (3) and SPICHI@BCM-1(20 %) (4); (e) profile fitting of SPICHI@BCM-1(6 %); (f) TGA of BCM-1 (1), SPICHI (2), SPICHI@BCM-1(6 %) (3) and SPICHI@BCM-1(20 %) (4); and (g-h) Acute lethal effects (mortality in % of unexposed control) of three MOFs of different size and/or structure, designated as BCM-2 (hexagonal), BCM-1 m (cubic micrometric) and BCM-1n (cubic nanometric) on (g) freshwater juvenile rotifer Brachionus calyciflorus (24 h exposure), (h) freshwater macroinvertebrate shredders Allogamus sp. (96 h exposure) and (i) Limnephilus sp. (96 h exposure). Mean ±SEM, n =3. M. Calles García et al. Chemical Engineering Journal 524 (2025) 169442 4 3. Results and discussion 3.1. Characterisation of MOF and SPI/CHI@MOF composite 3D-sponges 3.1.1. Multiscale characterisation of composite 3D-sponges BCM-1, SPICHI, and SPICHI@BCM-1 were first characterised by XRD after their synthesis, washing and activation at 80 ◦C at atmospheric pressure. Notably, the synthesis conditions for BCM-1 were scaled up tenfold compared to our previous study, yielding nearly 1 g of material per batch using just 10 mL of water as a solvent. The XRD pattern of BCM-1 matched the simulated data calculated from the cubic structural model with a F23 symmetry. [18] The fullprofile matching analysis confirms that the crystallographic cell and space group symmetry of the scaled-up BCM-1 closely align with those reported in our previous study with only a minor difference (Fig. S1). As illustrated in Fig. 1a, the crystal structure of BCM-1 consists of the archetypal hexameric [Zr 6 ( μ 3 O) 4 ( μ 3 OH) 4 ] 12+ clusters connected by mercaptosuccinic linkers. The 12-connected framework adopts a typical cubic “fcu” topology, where the disorder of the linkers results in a pore space decorated with -SH functions (Fig. 1a). In contrast to the nonscaled homologue, the diffraction pattern of the scaled-up BCM-1 shows a low-intensity pre-diffraction maximum at approximately 6◦. This feature is linked to the presence of long-range correlated nanodomains of cluster defects within its crystal structure. [26]. These defective nanoregions induce a reduction of the local symmetry to a P23 space group, as illustrated by the full profile matching analysis shown in Fig. S1. As expected, the XRD pattern of SPICHI exhibits low crystallinity, with broad diffraction maxima that arise from the disordered packing of CHI and SPI structures. As shown in Fig. 1b, the crystal structure of CHI consists of carbohydrate chains oriented along the c-axis and packed along the [100] and [010] directions. SPI adopts a globular-like structure due to the folding of amino acid chains sequence [27] (Fig. 1c), which further stacks into a three-dimensional framework stabilized by weak interactions. When CHI and SPI are combined, the XRD pattern revealed three key features: a low-intensity maximum near 9◦, and three broad signals at approximately 18◦, 29◦, and 42◦. Although the low crystallinity of SPICHI limits the extraction of precise structural information, it is evident that the packing of CHI and SPI units leads to a distinct arrangement in comparison to previously studied chitosan or chitosan/pectin composites [28], whose diffraction patterns exhibit slightly different features. No significant changes are observed in the diffraction data when BCM-1 is integrated into SPICHI. The patterns of SPICHI@MOF sponges show the three broad signals at 2θ =19.3, 27.5 and 40.2◦attributed to the SPICHI scaffold. In addition, they exhibit the two characteristic diffraction maxima of BCM-1 at low 2θ values of 8.5◦and 9.8◦. The intensity of the maxima increases as the MOF concentration in the composite raises (Fig. 1d). Last but not least, a profile fitting of the XRDdata was performed to semi quantify the presence of BCM-1 in the 3Dsponge (Fig. 1e). As expected, the intensity ratio between the most intense diffraction features of MOF and SPICHI components increases as the MOF content does. 3.1.2. Physicochemical characterisation FTIR analysis was carried out to elucidate whether any chemical interactions between SPICHI and BCM-1 occur when integrating both components in the 3D-sponge. The FTIR spectra of SPICHI@BCM-1 mainly show the characteristic fingerprints of SPICHI with subtle contribution from BCM-1 (Fig. S15). SPICHI exhibited three characteristic bands common to all proteins [15]: amide I (1630 cm −1 ), attributed to C – – O stretching; amide II (1540 cm −1 ), associated with N – H bending; and amide III (1235 cm −1 ), corresponding to C – N stretching and N – H bending, indicative of β-sheet structures [29]. By analysing hydrogen bonding within the polymer’s FTIR spectrum, the amide A region (3310–3270 cm −1 ) is also evaluated to follow up variations of the N – H stretching vibration. This region is part of the Fermi resonance doublet, and its second component, amide B, which exhibits weak absorption in the 3100–3030 cm −1 range [30]. Additionally, stretching vibrations of C – H bonds can be observed between 2960 and 2865 cm −1 . Further, the distinctive absorption bands of C – O bonds from CHI component are present in the 1180–953 cm −1 range [16]. BCM-1 shows a broad absorption signal ranging from 2600 to 3680 cm −1 corresponding to O–H stretching vibrations from water adsorbed from room conditions. Two absorption maxima located at approximately 1587 and 1415 cm −1 [31] dominate the contribution in the MOF spectrum and are assigned to ʋ asym and ʋ sym of (COO − ) groups belonging to the organic linkers. In addition, a weak band at 2560 cm −1 is assigned to the stretching vibration of S – H [32]. It is interesting to note that the two weak signals observed at approximately 1720 cm −1 and 1322 cm −1 are attributed to the C – – O and C – O stretching vibrations, and in turn, confirm the presence of uncoordinated linkers-structural defects in BCM-1. Finally, a prominent band appears near 650 cm −1 , which is related to Zr – O stretching vibrations [33]. The signals from SPICHI largely dominated the FTIR spectra of SPICHI@MOF composites. However, subtracting the SPICHI contribution from the SPICHI@BCM-1 data allows identifying the absorption band associated with ʋ asym (COO − ) of mercaptosuccinate linkers. Similarly, the signal at 650 cm −1 is attributed to Zr – O vibrational modes, which further confirms the presence of BCM-1 in the composite. As expected, the intensity of these bands increases with the BCM-1 content in the filters. Since none of the reported bands exhibit a significant shift in position when compared to the ones of SPICHI and BCM-1, strong chemical interactions are not expected between the surface functional groups of both materials integrated in the 3D-sponges. This tendency is also observed in the thermal properties, where the profile of SPICHI@BCM-1 is the weighted average of the TGA-curves for SPICHI and BCM-1. The thermogravimetry of BCM-1 shows three thermal processes (Fig. 1f). Initially, BCM-1 exhibits a weight loss attributed to the evaporation of adsorbed moisture (25–130 ◦C). This is followed by the release of the formic acid modulator, coordinated water and hydroxyl molecules (150–250 ◦C). The last degradation stage, occurring between 250 and 650 ◦C, corresponds to the calcination of the organic linker. The weight loss associated to the last two thermal processes was employed to quantify the average linker defects per formula in BCM-1. As detailed in the Table S1, 1.5 linker defects per formula unit were estimated based on the weight loss of the thermogravimetric curve between 220 and 550 ◦C. The thermal degradation of SPICHI starts above 150 ◦C, continue with several multistep weight losses, and ends after the overall calcination of the polymer. A minor inorganic residue was observed as a known impurity of SPICHI. The thermal degradation of composite 3D-sponges resembles that of SPICHI with a subtle difference. The remaining weight loss in the studied composites increases with the content of BCM-1, linked with the generation of ZrO 2 as residue. By analysing the weight percentage of the final residue of the TGA profiles, the final MOF wt% in the SPICHI@BCM-1 sponges was estimated. As detailed in the Table S2, BCM-1 content is slightly lower than the expected but still shows a qualitative direct correlation with the initial weight used for the sponges´assembly. 3.1.3. Ecotoxicological assessment Ecotoxicity assays were employed to estimate whether the release of BCM-1 particles from the biopolymeric composites pose a risk to the environment. The freshwater rotifer species Brachionus calyciflorus is a small metazoan invertebrate, ecologically relevant in lentic ecosystems and is generally sensitive to a wide range of contaminants, including metal nanoparticles [34]. On the other hand, Allogamus sp. and Limnephilus sp. are macroinvertebrates widely distributed in Iberian streams, ecologically relevant in lotic ecosystems, and, they are also sensitive to water quality [35–37], [38]. M. Calles García et al. Chemical Engineering Journal 524 (2025) 169442 5 Two BCM-1 materials with different particle sizes were studied for the ecotoxicological assays: BCM-1 m (micron-sized MOF particles) and BCM-1n (nanometric MOF particles). In addition, the zirconium mercaptosuccinate hexagonal variant, namely BCM-2, was synthesised for sake of comparison. The 24 h acute exposure of the freshwater rotifers B. calyciflorus to BCM-1 and BCM-2 samples did not show significant mortality up to 10 mg L −1 MOF concentration (Fig. 1g). Two-way ANOVA revealed that only the concentration of the MOFs had a significant effect (P <0.05), whereas the effects of particle size and MOF structure type were statistically insignificant (P >0.05; Fig. 1g–i). Tukey’s multiple comparisons tests showed that exposure only to 50 mg L −1 of BCM-1n induced significant mortality (P <0.05) on rotifers, although it was only about the 13 % (Fig. 1g). BCM-1 m and BCM-2 did not show any significant mortality effects (P >0.05) (Fig. 1h-i) on this freshwater micro-invertebrate. The 96 h acute exposure of the two different stream macroinvertebrate species, Allogamus sp. and Limnephilus sp., showed slightly varied results at the highest exposure concentration (Fig. 1h-i). None of the species exhibited mortality in unexposed controls after 96 h. Twoway ANOVAs confirmed the negligible effects of type and concentration of MOFs on Limnephilus sp. (P >0.05). However, exposure of Allogamus sp. showed similar response patterns upon exposure to three MOFs as observed for rotifers. Exposure only to 50 mg L −1 of BCM-1n induced significant mortality (P <0.05) of rotifers, however, the mortality was about 7 % (Fig. 1h). In contrast, no mortality was observed for Limnephilus sp. even after exposure to 50 mg L −1 of each MOF (Fig. 1i). Acute exposure tests clearly indicate that at environmentally relevant exposure concentrations, none of the three MOFs exhibited lethal toxicity to the microor macroinvertebrates. A low mortality (<15 %) for B. calyciflorus and Allogamus sp. was observed only at the highest exposure concentration of BCM-1n, probably due to its smaller size compared to BCM-1 m and BCM-2. The outcomes of our acute exposure tests also emphasised the importance of the ecotoxicological studies of MOFs considering their application to environmental remediation. 3.1.4. Nano to micrometric scale characterisation of hybrid 3D-sponges In the following, the combined information obtained from SEM, SAXS and Neutron Imaging were employed to reconstruct the nano to micrometric scale characteristics of SPICHI and SPICHI@BCM-1 3Dsponges. Microscopy pictures and neutron imaging experiments were conducted to acquire a deeper information of the internal porosity of a representative reconstructed neutron tomography slice of the composite sponge (Fig. 2a-b). The tomographic slice exhibits regions with varying grey-scale intensities (arbitrary units), which reflect the normalised neutron attenuation (Fig. 2b). Higher grey values correspond to regions of greater neutron attenuation (lower transmission), whereas lower grey values indicate reduced attenuation (higher transmission). Quantitative analysis of the macroporosity distribution along the filter depth, performed through pore segmentation, revealed a minimum pore diameter exceeding 200 μ m (equivalent to more than four pixels) (Fig. 2c), and an interconnected porous system where the percentage of porosity variates with the depth of the sponge. This information is showing a picture of the macroscopic structuration of the micrometric interconnected pore structure in the 3D-sponge, observed both by conventional microscopy and neutron imaging. SEM presents challenges in directly visualising the BCM-1 nanoparticles stabilized within the polymeric matrix. Even so, it enables the extraction of information about their impact on the nano to micrometric structure of the porous composite. First, as derived from the SEM images analysis by ImageJ software, the MOF shows a relatively homogeneous/ monodisperse particle size distribution with an average particle diameter of 42.5 nm (Fig. 2d). Second, a slight difference can be observed between the surface morphology of SPICHI (Fig. 2e) and SPICHI@BCM1 (Fig. 2f). For both samples, a floppy porous structure is revealed, which in turn is formed by the interconnection of thin SPICHI lamellar walls. When the MOF nanoparticles are integrated into the polymer matrix, the average size of the interconnected cavities within the membrane is reduced (Fig. 2f). This experimental evidence suggests that BCM-1 can modulate the pore structure of the polymeric scaffold, most probably altering the interaction between the composite and the solvent during the lyophilisation stage. This phenomenon has been reported for chitosan/MOF-808 beads. [39] SAXS experiments were conducted to gain further insights into the nanostructure of the sponges. According to the SAXS spectra (Fig. 2g) SPICHI presents a smooth curve with a power law behaviour, characteristic of large and extended polydisperse structures displaying fractallike organisation [40]. The sum of two power laws (I(q) =scale 1 ⋅q -p1 + scale 2 ⋅q -p2 were used to fit experimental data. The power law exponents, p 1 =4.1(1) and p 2 =2.6(1), were obtained from the fitting and are tentatively attributed to pores of micrometre size or larger in the sample. Upon integration of BCM-1 to SPICHI, in addition to the power-law scattering observed in both the lowand high q-regions, an additional SAXS signal emerges in the intermediate q-range. This signal is characterised by two oscillations, indicative of fairly monodisperse nanoparticles in the system. It is important to note that the power-law exponents are nearly identical across all samples, indicating a similar macro pore-structure in both the polymeric sponge and the MOF-hybrid homologues. The Indirect Fourier Transform (IFT) [41] for the SAXS data of SPICHI@BCM-1 was done at intermediate q-values [42]. IFT enables representing the scattering data in terms of the pair distance distribution function (PDD), p(r) (Fig. 2h). Despite the SAXS curves represented as PDD are quite similar, a slight difference associated to the size of the BCM-1 component, was observed. Namely R g =24.2 nm and Rg =19.6 nm values were obtained from the fitting of the SAXS spectra of SPICHI@BCM-1(6 %) and (20 %), correspondingly (Fig. 2h). Taking into account the spherical shape of BCM-1 nanoparticles, we can estimate the size of the spheres from the well-known relation between radius of gyration, Rg and radius of the sphere, R, Rg 2 =3 5R2 [43].Thus, it is concluded that SPICHI@BCM-1(20 %) corresponds well to the size of initial MOF particles, whilst some degree of agglomeration can be observed in SPICHI@BCM-1(6 %). The bell-like form of the PDD functions indicates the symmetrical nature of the particle size distribution, which significantly matches the one observed by the particle size distribution obtained by SEM (Fig. 2i). This is confirmed by the analysis of the polydispersity indices based on the invariants of the scattering curves, which indicates that the size distribution is closest to gaussian [44]. 3.2. Functional characterisation 3.2.1. Heavy metals adsorption screening Before presenting the experimental results in this section, it is important to note that the heavy-metal capture mechanism of the SPICHI@BCM-1 composite relies on a combination of absorption within the MOF pore structure and adsorption at both the surface of the biopolymer scaffold and the MOF/biopolymer interface. In the following discussion, and for the sake of simplicity, the term adsorption will be used. However, it should be noted that, depending on whether the MOF, SPICHI, or SPICHI@MOF components are employed, absorption, adsorption, or a combination of both processes may be responsible for the immobilization of the metal species. The stability of the membranes was first evaluated in 100 ppm solutions of the five studies heavy metals and in aqueous solutions of different pH values (Table S5). No significant pH changes were observed after immersing 10 mg of SPICHI@BCM-1 (20 %) sponge in 10 mL of the aforementioned solutions, except for the solution with an initial pH of 8.7, which decreased to 7.2 upon contact with the membrane. UV–Vis spectra of acidic solutions (pH ≤5) following contact with the sponge exhibited an increase of the absorbance within 200 and 250 nm, consistent with the release of minor components from SPICHI@BCM-1 M. Calles García et al. Chemical Engineering Journal 524 (2025) 169442 6 Fig. 2. (a) Optical microscope image of the SPICHI@BCM-1; (b) Neutron tomography; (c) Macroporosity distribution obtained from the processing of neutron tomography data (d-f) SEM images of (d) BCM-1, (e) SPICHI and (f) SPICHI@BCM-1; (g) SAXS data for SPICHI and SPICHI@BCM-1 filters. Experimental data: dots. Fitting: solid lines; (h) The obtained p(r) functions for SPICHI and SPICHI@BCM-1, together with Rg and Ø are presented in the inset; (i) Particle size distribution obtained from the analysis of the SEM images. M. Calles García et al. Chemical Engineering Journal 524 (2025) 169442 7 (Fig. S41). This release appears negligible, as similar absorbance values are observed when only very low concentrations of thiomalic acid or chitosan are dissolved in water. Accordingly, for subsequent experiments, the pH of the metal solutions was maintained at slightly acidic conditions to minimize the hybrid 3D-sponges´destabilization. Preliminary adsorption experiments were conducted for BCM-1, SPICHI, SPICHI@BCM-1(6 %) and SPICHI@BCM-1 (20 %) in 5 ppm solutions of the “big five” heavy metals—Hg(II), Pb(II), Cd(II), Cr(VI), As (III), and As(V) (Fig. 3). The results point that SPICHI has itself adsorption efficiencies over the 90 % for Hg(II), Pb(II), and Cd(II). In contrast, it showed negligible capacity to adsorb Cr(VI), As(V), and As (III), which remain stabilized in solution as anionic and neutral species, respectively. In contrast, BCM-1 shows excellent efficiency for the capture of Cr(VI), As(V), and As(III). When SPICHI and BCM-1 are combined into a 3D sponge-like configuration, their adsorption performance is modulated depending on the metal ion. The adsorption capacities of SPICHI/BCM-1 composites for Hg(II), Pb(II), and As(III) closely match the weighted average performance of the individual SPICHI and BCM-1 components. Interestingly, a synergistic effect was observed for Cd(II) adsorption, with SPICHI@BCM-1 composites outperforming both individual materials. In contrast, the SPICHI@BCM-1 hybrids significantly underperformed the weight-averaged figures obtained from BCM-1 and SPICHI for As(III), for Cr(VI) and As(V). Since BCM-1 is the active component for the capture of the latter species, its immobilization within the composite appears to partially block the access of anionic metalloid species to the MOF particles. This fact can be easily explained considering the isoelectric point of SPI, CHI and BCM-1. At the intermediate pH values where the adsorption experiments have been carried out (Table S5), SPI [45] and CHI [46] exhibit a slightly negatively charged surface, which may repulse arsenate and chromate anions to access the structure of the filter, and hence, the BCM-1 particles. Although BCM-1 exhibits an isoelectric point of approximately 3.9 (Fig. S32), its chemisorption affinity over oxyanions enables breaking the electrostatic repulsion forces at surface scale. Thus, increasing the BCM-1 content enhances the composite’s adsorption capacity for heavy metals that are stabilized as oxyanions. However, the experimentally obtained capacity is significantly lower than that predicted from the weighted-average calculated capacity (Fig. 3b), confirming that the blocking effect of the biopolymer scaffold on anions outweighs the chemisorption affinity of BCM-1. 3.2.2. Adsorption kinetics and isotherms It is important to note that the equilibrium capacities obtained from the kinetic experiments are significantly higher than those reported in the initial screening (Fig. 3), because 10 ppm initial metal concentration was employed instead of the 5ppms of the previous assays. Even though the adsorption trends obtained in preliminary experiments are mirrored in the kinetics study. The kinetic curves allow gaining significant insights into the time dependence of adsorption. To this end, the concentration of the “big five” heavy metals was monitored over an adsorption period of up to 240 min using BCM-1, SPICHI, and SPICHI@BCM-1 (20 %) as sorbents (Fig. 4). Pseudo-first and pseudo-second order models were employed to fit the kinetic data and to extract the parameters describing the sorbate–sorbent affinity and equilibrium adsorption capacity (Continuous line in Fig. 4). The fittings were employed to calculate the kinetic profile of SPICHI@BCM-1(20 %) based on the weight-averaged contributions from the BCM-1 and SPICHI kinetic curves (dashed line in Fig. 4). For detailed information about the calculations, the reader may consult the experimental section. As shown in the Fig. 4, BCM-1 exhibits a fast uptake with equilibrium times below 30 min for Cr(VI), As(V), As(III) and Hg(II). These metals are usually stabilized as (Cr 2 O 7 ) 2− /(HCrO 4 ) − , (H 2 AsO 4 ) − oxyanions; neutral arsenite (H 3 AsO 4 ) and HgCl 2 species in water (Fig. S33-S36), respectively. Consistent with previous studies [47–51], the capture of these species by Zr-MOFs with thiol-rich linkers is based on chemisorption at the defective positions of BCM-1 structure, and/or through the generation of metal/sulphide (i.e. S – As and S – Hg) bridges. In contrast, even if SPICHI shows negligible adsorption for Cr(VI) and As(V), the thiol-rich structure of SPI welcomes the capture of Hg(II) and As(III), likely through the formation of S – As and S – Hg covalent linkages as well. In comparison, the kinetics of their capture are usually slower than those shown by BCM-1. The microporous ordered nature of BCM-1 plays a beneficial role in accelerating the uptake of these heavy metals, compared to a SPICHI composite, which shows an overall lower surface area. It is worth noting that, although the adsorption kinetics of SPICHI@BCM-1 follows in general the weighted average curve derived from the individual kinetic profiles of SPICHI and BCM-1, notable positive and negative deviations are observed in terms of adsorption rate, equilibrium, and capacity. In detail, Hg(II) uptake by SPICHI@BCM-1 follows a similar and rapid adsorption kinetics to that calculated for the initial stage of the experiment; however, the maximum uptake is lower than expected. Similarly, the kinetics and capacity for As(III) uptake in the composite sponge are above the calculated value. This conclusion suggests a beneficial synergistic adsorption, likely from the interphase between the BCM-1 and SPICHI components combined in the filter. In contrast, SPICHI@BCM-1 exhibits a strong inhibition of adsorption kinetics for arsenate and chromate anions, indicating that SPICHI plays a dominant repealing role for oxyanions. The adsorption kinetics for Cd(II) and Pb(II) ionsusually stabilized as cationic species in water (Fig. S37-S38)- are slower than those Fig. 3. (a) Adsorption capacity (mg/g) of BCM-1, SPICHI and SPICHI@BCM-1 over Hg(II), Pb(II), Cd(II), Cr(VI), As(III) and As(V). Experiments were performed with an initial metal ion concentration of 5 ppm, using 1 mg adsorbent/mL dosage in 10 mL of solution. Measurements were done after 4 h. (b) Adsorption difference obtained from the difference between the experimental value and the one calculated from the weighted average of the contributions arising from the MOF and polymer components of the composite filters. M. Calles García et al. Chemical Engineering Journal 524 (2025) 169442 8 observed for anionic or neutral heavy metals. The equilibrium time expands above 60 min for Pb(II) and above 90 min for Cd(II). The time dependence of the metal uptake varies significantly from BCM-1 to SPICHI. Surprisingly, the response of SPICHI@BCM-1 fairly escapes from the expected calculated intermediate tendency. Pb(II) capture by SPICHI@BCM-1 shows a S-type profile that suggest a double-site-based capture. Once the first site is occupied by the metal ions, the adsorption is fastened until saturation of the second site. For Cd(II), the composite sponge surpasses the performance of the individual components in terms of adsorption capacity and rate. Once the individual and composite systems’ general adsorption trends and kinetics were identified, adsorption isotherms for the “big five” members were investigated. Isotherms provide two key parameters: (a) the saturation capacity, which reflects the density of available adsorption sites, and (b) the adsorption affinity, which indicates the material’s efficiency in capturing specific heavy metal ions at low concentrations. These two parameters were quantified by fitting the experimental data to the Freundlich isotherm model, as indicated by the continuous lines in Fig. 5. Although alternative models—such as Langmuir and Temkin—were also considered, they yielded significantly poorer fits. The Freundlich-fitting of the isotherms indicates that the adsorption in BCM-1, SPICHI, and especially SPICHI@BCM-1 are heterogeneous, suggesting the presence of multiple adsorption sites with slightly different natures. [52] A similar trend is observed for BCM-1. Although its crystal structure suggests the presence of homogeneous distribution and nature of the adsorption sites, the experimental data indicate a certain degree of heterogeneity. This deviation may be attributed to surface adsorption, which can lead to a behaviour that diverges from the ideal single-site adsorption occurring at the internal pore space of BCM-1. Such surface effects likely contribute to the nonideal isotherms observed in our study. [18] The isotherms of BCM-1, SPICHI, and SPICHI@BCM-1 (20 %) show two distinct trends linked to the adsorption of: (I) metals stabilized in cationic forms and (II) oxyanions or neutral species. The first group of Fig. 4. Adsorption kinetics of BCM-1, SPICHI and SPICHI@BCM-1 over Hg(II), Pb(II), Cd(II), Cr(VI), As(III) and As(V). Triplicate experiments were performed with an initial metal ion concentration of 10 ppm, using 1 mg adsorbent/mL dosage in a 50 mL of metal solution. M. Calles García et al. Chemical Engineering Journal 524 (2025) 169442 9 Aquat. Toxicol. 180 (2016) 227–235, https://doi.org/10.1016/j. aquatox.2016.09.017. [37] L.-L. Dong, H.-X. Wang, T. Ding, W. Li, G. Zhang, Effects of TiO2 nanoparticles on the life-table parameters, antioxidant indices, and swimming speed of the freshwater rotifer Brachionus calyciflorus, J Exp Zool A Ecol Integr Physiol 333 (2020) 230–239, https://doi.org/10.1002/jez.2343. [38] G.I. Tovar Jimenez, A. Valverde, C. Mendes-Felipe, S. Wuttke, A. Fidalgo-Marijuan, E.S. Larrea, L. Lezama, F. Zheng, J. Reguera, S. Lanceros-M´ endez, M.I. Arriortua, G. Copello, R.F. de Luis, Chitin/metal-organic framework composites as widerange adsorbent, ChemSusChem n/a (2021), https://doi.org/10.1002/ cssc.202100675. [39] P.W. Schmidt, Small-Angle Scattering Studies of Disordered, Porous and Fraetal Systems, 1991. [40] O. Glatter, Determination of particle size distribution functions from small angle scattering data by means of the indirect transformation Metho, J. Appl. Cryst. 10 (1977) 415–421. [41] SasView, (n.d.). [42] L. A., S.D.I. Feigin, Structure Analysis by Small-angle X-ray and Neutron Scattering, 1987. [43] O.V. Tomchuk, O.P. Kryshtal, E. Juszy´ nska-Gałązka, W.M. Zając, Particle-size polydispersity analysis based on the unified exponential/power-law approach to small-angle scattering, J. Appl. Cryst. 56 (2023) 1099–1107, https://doi.org/ 10.1107/S1600576723004752. [44] Z. Wang, L. Zhang, X. Zhang, M. Zeng, Z. He, J. Chen, Interfacial rheology and foaming properties of soy protein and hydrolysates under acid condition, Food Biophys. 16 (2021) 484–491, https://doi.org/10.1007/s11483-021-09685-9. [45] R. Hartmann, M. Beaumont, E. Pasquie, T. Rosenau, R. Serna-Guerrero, N-alkylated chitin nanocrystals as a collector in malachite flotation, ACS Sustain. Chem. Eng. 10 (2022) 10570–10578, https://doi.org/10.1021/acssuschemeng.2c01978. [46] C.O. Audu, H.G.T. Nguyen, C.-Y. Chang, M.J. Katz, L. Mao, O.K. Farha, J.T. Hupp, S.T. Nguyen, The dual capture of AsV and AsIII by UiO-66 and analogues, Chem. Sci. 7 (2016) 6492–6498, https://doi.org/10.1039/C6SC00490C. [47] H. Hu, J. Chen, F. Zhou, M. Nie, D. Hou, H. Liu, M. Delgado-Baquerizo, H. Ni, W. Huang, J. Zhou, X. Song, X. Cao, B. Sun, J. Zhang, T.W. Crowther, Y. Liang, Relative increases in CH4 and CO2 emissions from wetlands under global warming dependent on soil carbon substrates, Nat. Geosci. 17 (2024) 26–31, https://doi. org/10.1038/s41561-023-01345-6. [48] X. Wang, R. Lu, B. Pan, C. Yang, M. Zhuansun, J. Li, Y. Xu, S.F. Hung, G. Zheng, Y. Li, Z. Wang, Y. Wang, Enhanced carbon-carbon coupling at interfaces with abrupt coordination number changes, ChemSusChem 17 (2024), https://doi.org/ 10.1002/cssc.202400150. [49] H. Salazar, M. Rosales, I. Zarandona, J. Serra, B.F. Gonçalves, A. Valverde, L. P. Cavalcanti, S. Lanceros-Mendez, A. García, K. de la Caba, P. Guerrero, P. M. Martins, R. Fern´ andez de Luis, Metal-organic framework functionalized chitosan/pectin membranes for solar-driven photo-oxidation and adsorption of arsenic, Chem. Eng. J. 497 (2024) 154417, https://doi.org/10.1016/j. cej.2024.154417. [50] R. Brito-Pereira, J.M. Queir´ os, L. Celaya-Azcoaga, R.F. de Luiz, P. Martins, S. Lanceros-Mendez, Hexavalent chromium dual water remediation and sensing based on hybrid polymer/metal-organic framework composites, J. Environ. Chem. Eng. 12 (2024) 113839, https://doi.org/10.1016/j.jece.2024.113839. [51] H. Salazar, M. Rosales, I. Zarandona, J. Serra, B.F. Gonçalves, A. Valverde, L. P. Cavalcanti, S. Lanceros-Mendez, A. García, K. de la Caba, P. Guerrero, P. M. Martins, R. Fern´ andez de Luis, Metal-organic framework functionalized chitosan/pectin membranes for solar-driven photo-oxidation and adsorption of arsenic, Chem. Eng. J. 497 (2024) 154417, https://doi.org/10.1016/j. cej.2024.154417. [52] G.P. Pincetti-Zúniga, L.A. Richards, L. Daniele, A.J. Boyce, D.A. Polya, Hydrochemical characterization, spatial distribution, and geochemical controls on arsenic and boron in waters from arid Arica and Parinacota, northern Chile, Sci. Total Environ. 806 (2022) 150206, https://doi.org/10.1016/j. scitotenv.2021.150206. [53] M. Rosales, R. Gauvin, R. Espinoza-Gonz´ alez, S. Lanceros-M´ endez, R.F. de Luis, J. Vicente, N. Brodusch, S. Bessette, B. Rodríguez, H. Estay, A. García, Unveiling autarchic photo-thermal capabilities of morphology-tailored SnO2 nanostructures for boosting arsenic photo-oxidation, J. Environ. Chem. Eng. 13 (2025) 115736, https://doi.org/10.1016/j.jece.2025.115736. [54] S. Santoro, J. Occhiuzzi, M. Aquino, A. Politano, S. Straface, G. D’Andrea, C. Carrillo, R. Mallada, A. Garcia, H. Estay, D. Xevgenos, P. Argurio, E. Curcio, Green photocatalytic mixed matrix membranes for simultaneous arsenic photooxidation and water recovery via membrane distillation, Sep. Purif. Technol. 342 (2024) 127042, https://doi.org/10.1016/j.seppur.2024.127042. [55] S.M. Abramov, J. Tejada, L. Grimm, F. Sch¨ adler, A. Bulaev, E.J. Tomaszewski, J. M. Byrne, D. Straub, H. Thorwarth, R. Amils, S. Kleindienst, A. Kappler, Role of biogenic Fe(III) minerals as a sink and carrier of heavy metals in the Rio Tinto, Spain, Sci. Total Environ. 718 (2020) 137294, https://doi.org/10.1016/j. scitotenv.2020.137294. [56] W. Li, J. Wang, Z. Zhan, J. Luo, D. Zhou, Y. Lei, Extraction of valuable metals from acid mine drainage by an electrochemically activated limestone system, Nat. Commun. 16 (2025) 6719, https://doi.org/10.1038/s41467-025-62045-w. [57] M.J. Valenzuela-Diaz, A. Navarrete-Calvo, M.A. Caraballo, J. McPhee, A. Garcia, J. P. Correa-Burrows, L. Navarro-Valdivia, Hydrogeochemical and environmental water quality standards in the overlap between high mountainous natural protected areas and copper mining activities (Mapocho river upper basin, Santiago, Chile), J Hydrol (Amst) 588 (2020) 125063, https://doi.org/10.1016/j. jhydrol.2020.125063. [58] A.T. Luís, J.C. Fortes, M. Santisteban, J.M. D´ avila, M.A. Caraballo, J.M. TerronesSaeta, J. Diaz-Curiel, J.A. Grande, Relationships between hydrogeochemistry and diatoms in acid mine drainage affected media: the case of Iberian pyrite belt; functioning models for an all metallogenetic province, J. Geochem. Explor. 264 (2024) 107537, https://doi.org/10.1016/j.gexplo.2024.107537. [59] J. G´ omez, H. Giraldo, G. Piaggio, L. Barros, M. Quilaqueo, Y.M. Quintero, A. García, S. Santoro, E. Curcio, H. Estay, Recovery of copper sulfate from acidic mine waters by membrane crystallization, J Memb Sci 700 (2024) 122707, https:// doi.org/10.1016/j.memsci.2024.122707. [60] R.N. Lieberman, M. Izquierdo, P. C´ ordoba, N. Moreno Palmerola, X. Querol, A. M. S´ anchez de la Campa, O. Font, H. Cohen, Y. Knop, R. Torres-S´ anchez, D. S´ anchez-Rodas, C. Mu˜ noz-Quiros, J.D. de la Rosa, The geochemical evolution of brines from phosphogypsum deposits in Huelva (SW Spain) and its environmental implications, Sci. Total Environ. 700 (2020) 134444, https://doi.org/10.1016/j. scitotenv.2019.134444. [61] M.L. Peralta Ramos, J.A. Gonz´ alez, S.G. Albornoz, C.J. P´ erez, M.E. Villanueva, S. A. Giorgieri, G.J. Copello, Chitin hydrogel reinforced with TiO2 nanoparticles as an arsenic sorbent, Chem. Eng. J. 285 (2016) 581–587, https://doi.org/10.1016/j. cej.2015.10.035. [62] J. Li, Z. Yang, T. Ding, Y.-J. Song, H.-C. Li, D. Li, S. Chen, F. Xu, The role of surface functional groups of pectin and pectin-based materials on the adsorption of heavy metal ions and dyes, Carbohydr. Polym. 276 (2022) 118789, https://doi.org/ 10.1016/j.carbpol.2021.118789. [63] S. Rehman, J. Wu, M. Hanif, T. Wang, M. Huang, T. Hongmei, S. Zhongbo, P. Wu, F. Raziq, Enhancing cadmium immobilization on halloysite through varied dissolved organic matter: exploring the role of halloysite aggregation, J. Environ. Chem. Eng. 12 (2024) 111898, https://doi.org/10.1016/j.jece.2024.111898. [64] L. Celaya-Azcoaga, A. Crespi, W. Shepard, R. Quezada, M.L. Peralta Ramos, M. Cavallo, I. Irastorza Epelde, H. Garcia Garcia, G. Barandika, F. Bonino, B. Baz´ an, A. García, G.J. Copello, O. Castillo, U. Silv´ an, R. Fern´ andez de Luis, Metal-organic chelator frameworks for arsenic-based cancer treatment, J. Colloid Interface Sci. 691 (2025) 137335, https://doi.org/10.1016/j.jcis.2025.137335. [65] M. Mon, F. Lloret, J. Ferrando-Soria, C. Mart˜ A-Gastaldo, D. Armentano, E. Pardo, Selective and efficient removal of mercury from aqueous media with the highly flexible arms of a BioMOF, Angewandte Chemie - International Edition 55 (2016) 11167–11172, https://doi.org/10.1002/anie.201606015. [66] A. García, M. Rosales, M. Thomas, G. Golemme, Arsenic photocatalytic oxidation over TiO2-loaded SBA-15, J. Environ. Chem. Eng. 9 (2021) 106443, https://doi. org/10.1016/j.jece.2021.106443. [67] M. Rosales, J. Orive, R. Espinoza-Gonz´ alez, R. Fern´ andez de Luis, R. Gauvin, N. Brodusch, B. Rodríguez, F. Gracia, A. García, Evaluating the bi-functional capacity for arsenic photo-oxidation and adsorption on anatase TiO2 nanostructures with tunable morphology, Chem. Eng. J. 415 (2021) 128906, https://doi.org/10.1016/j.cej.2021.128906. [68] C.O. Audu, H.G.T. Nguyen, C.Y. Chang, M.J. Katz, L. Mao, O.K. Farha, J.T. Hupp, S. T. Nguyen, The dual capture of AsV and AsIII by UiO-66 and analogues, Chem. Sci. 7 (2016) 6492–6498, https://doi.org/10.1039/c6sc00490c. [69] M. Rosales, J. Orive, R. Espinoza-Gonz´ alez, R. Fern´ andez de Luis, R. Gauvin, N. Brodusch, B. Rodríguez, F. Gracia, A. García, Evaluating the bi-functional capacity for arsenic photo-oxidation and adsorption on anatase TiO2 nanostructures with tunable morphology, Chem. Eng. J. 415 (2021) 128906, https://doi.org/10.1016/j.cej.2021.128906. [70] P.G. Saiz, B. Gonzalez Navarrete, S. Dutta, E. Vidal Martín, A. Reizabal, I. Oyarzabal, S. Wuttke, S. Lanceros-M´ endez, M. Rosales, A. García, R. Fern´ andez de Luis, Metal-organic frameworks for dual photo-oxidation and capture of arsenic from water, ChemSusChem 17 (2024), https://doi.org/10.1002/ cssc.202400592. [71] P.G. Saiz, N. Iglesias, B. Gonz´ alez Navarrete, M. Rosales, Y.M. Quintero, A. Reizabal, J. Orive, A. Fidalgo Marijuan, E.S. Larrea, A.C. Lopes, L. Lezama, A. García, S. Lanceros-Mendez, M.I. Arriortua, R. Fern´ andez de Luis, Chromium speciation in zirconium-based metal–organic frameworks for environmental remediation, Chemistry 26 (2020) 13861–13872, https://doi.org/10.1002/ chem.202001435. [72] P.G. Saiz, A. Valverde, B. Gonzalez-Navarrete, M. Rosales, Y.M. Quintero, A. Fidalgo-Marijuan, J. Orive, A. Reizabal, E.S. Larrea, M.I. Arriortua, S. LancerosM´ endez, A. García, R. Fern´ andez de Luis, Modulation of the bifunctional CrVI to CrIII photoreduction and adsorption capacity in ZrIV and TiIV benchmark metalorganic frameworks, Catalysts 11 (2021) 51, https://doi.org/10.3390/ catal11010051. [73] S. Tatay, S. Martínez-Gim´ enez, A. Rubio-Gaspar, E. G´ omez-Oliveira, J. Castells-Gil, Z. Dong, ´ A. Mayoral, N. Almora-Barrios, N.M. Padial, C. Martí-Gastaldo, Synthetic control of correlated disorder in UiO-66 frameworks, Nat. Commun. 14 (2023) 6962, https://doi.org/10.1038/s41467-023-41936-w. [74] K.P. Øien, S. Wragg, D. Reinsch, H. Svelle, S. Bordiga, S. Lamberti, C. Lillerud, Detailed structure analysis of atomic positions and defects in zirconium metal − organic frameworks, Cryst. Growth Des. 14 (2014) 5370–5372. [75] K.P. Shearer, G.C. Chavan, S. Ethiraj, J. Vitillo, J.G. Svelle, S. Olsbye, U. Lamberti, C. Bordiga, S. Lillerud, Tuned to perfection: ironing out the defects in metalorganic framework UiO-66, Chem. Mater. 26 (2014) 4068–4071, https://doi.org/ 10.1021/cm501859p. [76] H. Furukawa, F. G´ andara, Y.B. Zhang, J. Jiang, W.L. Queen, M.R. Hudson, O. M. Yaghi, Water adsorption in porous metal-organic frameworks and related materials, J. Am. Chem. Soc. 136 (2014) 4369–4381, https://doi.org/10.1021/ ja500330a. [77] A. Valverde, E. Alkain, N.A. Rio-L´ opez, L. Lezama, A. Fidalgo-Marijuan, J.M. Laza, S. Wuttke, J.M. Porro, I. Oyarzabal, M. Jim´ enez-Ruiz, V. García Sakai, P.L. Arias, M. Calles García et al. Chemical Engineering Journal 524 (2025) 169442 16 I. Agirrezabal-Telleria, R. Fern´ andez de Luis, Enzyme-mimicking of copper-sites in metal–organic frameworks for oxidative degradation of phenolic compounds, J Mater Chem A Mater 12 (2024) 4555–4571, https://doi.org/10.1039/ D3TA06198A. [78] A. Valverde, G.I. Tovar, N.A. Rio-L´ opez, D. Torres, M. Rosales, S. Wuttke, A. Fidalgo-Marijuan, J.M. Porro, M. Jim´ enez-Ruiz, V. García Sakai, A. García, J. M. Laza, J.L. Vilas-Vilela, L. Lezama, M.I. Arriortua, G.J. Copello, R. Fern´ andez de Luis, Designing metal-chelator-like traps by encoding amino acids in zirconiumbased metal–organic frameworks, Chem. Mater. 34 (2022) 9666–9684, https://doi. org/10.1021/acs.chemmater.2c02431. [79] J. Jiang, F. G´ andara, Y.B. Zhang, K. Na, O.M. Yaghi, W.G. Klemperer, Superacidity in sulfated metal-organic framework-808, J. Am. Chem. Soc. 136 (2014) 12844–12847, https://doi.org/10.1021/ja507119n. [80] A.J. Howarth, M.J. Katz, T.C. Wang, A.E. Platero-Prats, K.W. Chapman, J.T. Hupp, O.K. Farha, High efficiency adsorption and removal of selenate and selenite from water using metal-organic frameworks, J. Am. Chem. Soc. 137 (2015) 7488–7494, https://doi.org/10.1021/jacs.5b03904. [81] A. Barth, Infrared spectroscopy of proteins, Biochimica et Biophysica Acta (BBA) 1767 (2007) 1073–1101, https://doi.org/10.1016/j.bbabio.2007.06.004. [82] G. Socrates, G., Infrared and Raman Characteristic Group Frequencies, 3rd ed, 2004. Chichester. [83] P. Liang, S. Liu, M. Li, W. Xiong, X. Yao, T. Xing, K. Tian, Effective adsorption and removal of Cr(VI) from wastewater using magnetic composites prepared by synergistic effect of polypyrrole and covalent organic frameworks, Sep. Purif. Technol. 336 (2024) 126222, https://doi.org/10.1016/j.seppur.2023.126222. [84] A.G. Al Lafi, A. Khuder, Removal of Cr(VI) from aqueous solutions by activated carbon and its composite with P2W17O61: a spectroscopic study to reveal adsorption mechanism, Heliyon 11 (2025) e41862, https://doi.org/10.1016/j. heliyon.2025.e41862. [85] A.G.B. Pereira, E.C. Muniz, Y.-L. Hsieh, 1H NMR and 1H–13C HSQC surface characterization of chitosan–chitin sheath-core nanowhiskers, Carbohydr. Polym. 123 (2015) 46–52, https://doi.org/10.1016/j.carbpol.2015.01.017. [86] G. Sun, X. Liu, D.J. McClements, S. Liu, B. Li, Y. Li, Chitin nanofibers improve the stability and functional performance of Pickering emulsions formed from colloidal zein, J. Colloid Interface Sci. 589 (2021) 388–400, https://doi.org/10.1016/j. jcis.2021.01.017. [87] J. Li, G. Huang, H. Qian, F. Pi, Fabrication of soy protein isolate - high methoxyl pectin composite emulsions for improving the stability and bioavailability of carotenoids, Food Biosci. 53 (2023) 102738, https://doi.org/10.1016/j. fbio.2023.102738. M. Calles García et al. Chemical Engineering Journal 524 (2025) 169442 17