Selection of EOCs
Among the commonly used 100 pharmaceuticals,11 anti-hypertensive and antibiotic drugs are two of the most consumed groups. Arterial hypertension is one of the main chronic health problems, being extremely prevalent (20%) among the older adult population.12 β-blockers are among the most consumed drugs used in arterial hypertension treatment.13 During 2020, the consumption of β-blockers ranged around 24,495 (defined daily doses per 1,000 inhabitants per day).14 In particular At, a β-blocker drug primarily used to treat patients suffering from various heart disorders such as high blood pressure, chest pain (angina), migraines, and irregular heartbeats,15 is ranked the 31st most prescribed drug in United States of America in 2014.16 In addition, At cannot be fully metabolized by the human body, undergoing incomplete absorption (ca. 90%) and being largely excreted unchanged in the urine,17–19 exponentially increasing its presence in water (from 0.35 to 2.21 mg·L-1).20,21 Further, the presence of At in water may cause severe toxic effects (e.g., dizziness, feeling tired, depression, chances of heart failure, shortness of breath, and cause bronchospasm), being also accumulated in breast milk, which is related to the immaturity of the renal function of neonate breast-fed infants.15,22
On the other hand, because of their importance for human and veterinary medicine, but also because of their persistence, sulfonamides, quinolones and trimethoprim are the most widely detected antibiotics in water.23 Only in Spain, in 2020 the daily dose of prescribed antibiotics by the public health system were 19 per 1000 inhabitants per day according to the Spanish Agency for Medicines and Health Products. Among the most widely prescribed groups of antibiotics, we can find β-lactams (53.4%), macrolides and lincosamide (11.3%), quinolones (9.8%), tetracycline (8.0%), and sulfonamides and trimethoprim (2.5%).24 Regarding the veterinary use of antibiotics, in the European Union, the bestseller antibiotics were tetracyclines (30.4%), β-lactams (26.9%) and sulfonamides (9.2%).25 Particularly, the sulfonamide sulfamethazine (SMT), frequently used in pigs and cattle, has been detected in natural environments such as soil or water, ranging from a few ng to tens of mg per liter or kg of soil.26,27 Large proportions of this compound is excreted unchanged in feces and urine given its incomplete metabolism.28 Furthermore, the presence of SMT in water is associated with significant risks for humans due to the development of microbial resistances, even at low doses.29
Screening of the SMT and At photodegradation and MOF stability
For the efficient removal of EOCs from water, we have selected 4 Ti-MOFs based on: i) Ti-oxoclusters with redox and photocatalytic activity that could be exploited in the photodegradation of EOCs,8 ii) their a priori remarkable hydrolytic stability (mostly tested in pure MilliQ water), iii) their exceptional porosity (see SI, Table S1), in some cases compatible with the SMT and At dimensions (11 x 5 x 5 Å3, and 15 x 7 x 5 Å3, respectively; estimated by Vesta considering van der Waals radii), and iv) in terms of cost, they could be considered affordable for large scale production in the future.30 The series comprises: 1) the MIL-100(Ti) or [Ti3(µ3-O)O(OH)2(BTC)2] (MIL: Materials Institute Lavoisier) based on the trimesate ligand (H3BTC), which combines a high chemical stability and mesoporosity (∼25 & 29 Å, accessible via ∼5 & 8.6 Å windows; SBET ∼ 1300 m2·g-1) with photoactive Ti3(µ3-O) metal-oxoclusters;31 2) the MIL-125-NH2 or [Ti8O8(OH)4(BDC-NH2)6] built from Ti-oxoclusters coordinated to the 2-aminoterephthalate ligand (BDC-NH2),32 exhibiting a high microporosity (∼13 & 6 Å, accessible via ∼6 Å windows; SBET ∼ 1400 m2·g-1) and robustness; 3) the MUV-10(Ca) or [Ti3Ca3(µ3-O)3(BTC)4] porous solid (MUV: Materials of the University of Valencia) built from the interlinking of fully deprotonated trimesate anions and tetranuclear TiIV2CaII2(µ3-O)2(H2O)4(CO2)8 clusters,33 with an important porosity (∼10 Å, accessible via ∼5 Å windows; SBET ∼ 1000 m2·g-1); and 4) the recently reported IEF-11 or [Ti2O3(SQ)] (IEF: IMDEA Energy Framework),34 based on photo/redox 2D titanium layers and the squarate (SQ, 3,4-dihydroxycyclobut-3-ene-1,2-dionate) as organic linker (see detailed properties in Table S1), with small porosity (∼4.5 Å; SBET ∼ 120 m2·g-1) and exceptional hydrolytic stability.
First, the SMT and At photodegradation capacity and the matrix chemical stability of the selected Ti-based MOFs were studied by using the At and SMT concentration normally reported in the environment (5 ppm,35 and 35 ppm,20,21 respectively). The At and SMT photodegradation capacity strongly depends on the used MOF, decreasing in the following order: (At) MIL-100(Ti) > MIL-125-NH2 > > IEF-11 > MUV-10(Ca); and (SMT) MIL-100(Ti) > MIL-125-NH2 > > MUV-10(Ca) > IEF-11 (Fig. 2).
Interestingly, these results highlight the remarkable ability of the MIL-100(Ti) to remove both EOCs, eliminating 100% of At and SMT in only 2 and 4 h, respectively. The plateau of degradation is reached only after 1 h in all studied materials, except for the continuous At degradation of IEF-11, and SMT and At degradation of MIL-100(Ti) where it took only 2 h. Factors like MOF structure/porosity (pore accessibility, surface, volume, tortuosity, connectivity, particle size, etc.) and nature (ligand, cluster structure, band gap, external surface, etc.) might influence the pharmaceuticals degradation capacity. For instance, compared with the rest of studied Ti-MOFs, the lower accessibility of IEF-11, with a pore and window size of 4.5 Å, might hamper the accessibility of At and SMT and, therefore, their photodegradation. In fact, there is no SMT degradation when using IEF-11 and the kinetics of degradation of At is ca. 11-fold lower than MIL-100(Ti) (Table 1). Regarding the chemical nature of MOFs, there is not a direct relationship between the calculated band gap values (Figure S1) and the photocatalytic capacity of these materials in the studied reaction.
Figure 2. Comparative photodegradation evolution of At (a) and SMT (b) using different Ti-MOFs. For clarity, degradation of MOFs is omitted here (although included in the SI, Figure S2&S3).
Table 1
Total At and SMT photodegradation (after 5 h, %), MOF degradation (%), and kinetic constant (M− 1·h− 1) for all studied materials.
MOF | Photodegraded At (%) MOF (%) | Kinetic constant (M-1·h-1) R2 | Photodegraded SMT (%) MOF (%) | Kinetic constant (M-1·h-1) |
MIL-100(Ti) | 100 ± 0 3.8 ± 0.1 | 12199 0.985 | 100 ± 0 3.2 ± 0.3 | 112013 0.964 |
MIL-100(Fe) | 65.0 ± 4.4 9.7 ± 1.5 | 247 0.957 | 66.4 ± 10.4 1.0 ± 0.1 | 24937 0.996 |
MIL-125-NH2 | 66.0 ± 1.1 38.8 ± 1.2 | 122555 0.985 | 69.5 ± 5.2 39.3 ± 0.4 | 70213 0.958 |
MUV-10(Ca) | 43.9 ± 14.7 41.0 ± 1.9 | 64034 0.995 | 34.4 ± 5.6 36.1 ± 6.4 | 4946.2 0.896 |
IEF-11 | 56.0 ± 12.6 0.78 ± 0.01a | 1124.4 0.982 | 0.00 ± 0.01 0.35 ± 0.01a | - |
a Value corresponding to 24 h. |
For comparison, the At and SMT adsorption capacity of the studied Ti-MOFs was performed in absence of light. As in the photocatalytic studies, the adsorption of At and SMT strongly depends on the MOF, decreasing in the following order: (At) MIL-125-NH2 > MIL-100(Ti) > MUV-10(Ca); and (SMT) MIL-125-NH2 > MIL-100(Ti) ~ MUV-10(Ca) (Figure S2-S4). Unlike its remarkable pharmaceuticals photodegradation capacity, MIL-100(Ti) can eliminate by adsorption only 13.5 ± 3.9% of At and 0 ± 3.9% of SMT after 5 h. Again, in a view of providing a comparison, the same experiment was performed using the Fe-based MIL-100(Fe) analogue, obtaining similar adsorption capacities for At (8.2 ± 5.5%) and SMT (14.2 ± 8.0%) than the MIL-100(Ti). The formation of interactions between the At and the MIL-100(Ti) is evidenced by Fourier transform infrared spectroscopy (FTIR) by comparing the spectra of free At to that of the empty and the At loaded material (At@MIL-100(Ti)). Specifically, there is a shift in the wavelengths of pure At (from 2964 and 2922 cm-1 to 2955 and 2925 cm-1 for At and At@MIL-100(Ti), respectively; see SI, Figure S5), characteristic of the ν(C-H) group of the C9 (near the ether group C9-O-benzene).36 In contrast, the MIL-125-NH2 was able to adsorb up to 66.6 ± 1.1% of At and 50.2 ± 3.4% of SMT after 5 h. However, the release of the previously adsorbed EOCs was evidenced after 1 h, probably related with the degradation of the framework during the drug adsorption process (SI, Figure S2&S3). This is supported by the significant leaching of the linker (ca. 18% in both SMT and At adsorption after 5 h), confirmed by the similar FTIR spectra of the drug-loaded MIL-125-NH2 and the free linker H2BDC-NH2 (SI, Figure S6), and the amorphization of the MIL-125-NH2 framework after the adsorption process (SI, Figure S7&S8).
On the other hand, it is imperative to get fast degradation kinetics in order to achieve more efficient removal processes. The comparison of the At and SMT degradation kinetics has been performed through the fitting of the data to a second order kinetics according to Eq. (1) (see experimental section and SI, Figure S9). Although the data have also been fitted to a zero and first order kinetics, a better correlation was found using the second order. This is in agreement with previously reported photoactive MOFs and composites (i.e., MIL-125-NH2 and AgNC@MIL-125-NH2 in the degradation of SMT or methylene blue).10 The best degradation rate is obtained when using MIL-125-NH2 in the At degradation, with a k value of 122,555 M-1·h-1, which is ca. 2, 10 and 109-fold higher than MUV-10(Ca), MIL-100(Ti), and IEF-11, respectively. In the case of SMT, the best degradation rate is obtained for MIL-100(Ti), with a k value of 112,013 M-1·h-1, which is ca. 2, 5, and 23-fold higher than MIL-125-NH2, MIL-100 (Fe), and MUV-10(Ca), respectively. Finally, when evaluating the kinetics of the photocatalytic process using MIL-100(Ti) and MIL-125-NH2, the particle size and the external surface may also play important roles. It is generally accepted that smaller particles might favor the catalytic reactions. Tentatively, we can argue that the smaller MIL-125-NH2 particles (ca. 240 nm with an estimated external surface of 225 mg2·g-1 by t-plot method (p/p0 from 0.3 to 0.6)) than MIL-100(Ti) (> 1 µm with an external surface of 195 mg2·g-1) will favor the transport of EOCs and degradation products.
Aside from the EOCs elimination capacity, the chemical stability of the framework is a limiting parameter in water decontamination processes. The point is to avoid an extra contamination of water due to the leaching of the MOFs constituents (e.g., ligands, metals). Therefore, and in contrast with the vast majority of reports on this topic, we have here considered the possible release of the MOF’s building blocks. The chemical stability of a coordination network mainly depends on the strength of the metal to linker bonds, which can be estimated according to the hard and soft acids and bases (HSAB) principle,37 and the competition with reactive species found in the solution. Thus, the release of the constitutive ligand was monitored by HPLC and the crystalline structure was checked by XRPD. At this point, it should be noted that the degradation of the frameworks was not evaluated via the quantification of the released metallic species in solution as, under the working conditions (tap water, pH = 6) and according to the Pourbaix diagram,38 the vast majority of the degraded metallic species would be precipitated (i.e., TiO2), underestimating the MOF degradation. Through the linker release, MOFs degradation could be also overestimated, as it can be adsorbed in the porosity of the framework (e.g., poorly activated/purified solids). However, we should rule out the release of any species that could be toxic, which in imperative for water decontamination. During the At and SMT photodegradation, the chemical stability of the framework in At or SMT-contaminated tap water decreased in the following order: (At) IEF-11 > MIL-100(Ti) > > MIL-125-NH2 ~ MUV-10(Ca); and (SMT) IEF-11 > MIL-100 (Ti) > > MUV-10(Ca) ~ MIL-125-NH2 (Table 1). Note here that the stability ranking is very similar for At and SMT, ruling out an important effect of the drug nature on the MOF degradation. In addition, the XRPD patterns of the At- and SMT-containing MOFs evidenced that the At or SMT loading process does not alter the crystalline structure of MIL-100(Ti) and IEF-11, but there is an important peak broadening in the MUV-10(Ca) and MIL-125-NH2 materials, consistent with a crystallinity loss (SI, Figure S7&S8). Remarkably, MIL-100(Ti) and IEF-11 show a high chemical stability under the working conditions with only ca. 3 and 0.7% degradation, respectively, after 5 h in contact with the contaminants’ solutions.
Considering the above results, the most promising photocatalyst for the At and SMT removal in water is MIL-100(Ti), demonstrating not only an exceptionally high and fast EOCs degradation (100% in 2 and 4 h, respectively), but also a moderate matrix degradation (ca. 9%). This is particularly important considering that the median oral lethal dose (LD50) for H3BTC is ca. 4 times lower than the one for SMT and At (in rats, LD50 is 8.4, 2 and > 2 g·kg-1 for H3BTC, SMT and At). Therefore, the proposed photodegradation method is efficiently improving the water quality within a short time. The results on the photodegradation of At and SMT using MIL-100(Ti) are on the range or even overpass the results obtained with other MOFs, MOF composites or other materials (SI, Table S2).
At this point, in an attempt to rationalize the complex process involved in the At and SMT photodegradation using MIL-100(Ti), we have explored the photocatalytic performance of the isostructural iron analogue MIL-100(Fe) under equal conditions. It should be pointed that even the different band gap values of the Ti and Fe based MIL-100 materials (3.48 and 2.73 eV, respectively, SI, Figure S1), both suspensions were irradiated in all the UV-vis range. As expected, the Ti-based MIL-100 is more efficient than its Fe counterpart, with a 100% and ca. 65% of At and SMT degradation after 5 h-irradiation, evidencing the crucial role of the titanium trimeric clusters. When comparing the kinetic of the process (Table 1), the degradation rates of At and SMT by MIL-100(Ti) is ca. 49 and 4.5-fold higher than the Fe isostructural MOF. Both materials are chemically stable under the studied conditions, with only ca. 10 and 1% of MIL-100(Fe) degradation in At and SMT solutions, respectively. However, in the case of MIL-100(Fe), the degradation of the framework seems to be affected by the nature of the contaminant, being favored in the presence of At. The amide of At may preferentially interact with the Fe metal sites than with the Ti ones (more oxophilic). Finally, the accessible porosity of the framework is not a determining factor when comparing the Fe and Ti-based MIL-100 materials. Although the accessible porosity of MIL-100(Fe) is double than the one of MIL-100(Ti) (SI, Table S1), this last material is the most effective in At and SMT degradation.
Photodegradation products using MIL-100(Ti)
The hazard of these EOCs does not only relay on their concentration or toxicity, but also their metabolites or degradation products, which can sometimes be more harmful than the parent compounds.39 The degradation products of At and SMT using MIL-100(Ti) were thus independently analyzed. Previous studies have described the photochemical behavior of At and SMT, determining that the degradation occurs through the cleavage at various positions (Figure S10).40 The identification of the photodegradation products of At and SMT formed after 5 h-irradiation of the At-MIL-100(Ti) and SMT-MIL-100(Ti) systems in aqueous medium was carried out using UHPLC/MS, supported with fragmentation patterns obtained from MS/MS experiments. Although the determination of the EOCs degradation pathway is out of the scope of this work, we have successfully identified some intermediates deduced from their estimated molecular weight, allowing to assess the potential toxicity of the resulting products.
Particularly for At, cleavage of the side chain and the addition of the hydroxyl group to the parent compound were found to be the two main degradation pathways (Figure S10&S11). Thus, the fragment ion m/z 134 (3-(isopropyl)propane-1,2-diol) was attributed to the product from the chain cleavage, and the ions m/z 167 2-(2,4-dihydroxyphenyl)acetamide and 2-(3,4-dihydroxyphenyl)acetamide to the chain cleavage and oxidation. These photodegradation pathways and intermediates have been frequently reported in the photodegradation of At.41–43
On the other hand, in the SMT photodegradation study, the fragment ion m/z 216.7 (N-(4,6-dimethylpyrimidin-2-yl)benzene-1,4-diamine) was attributed to the product from SO2 extrusion, a phenomenon frequently shown in sulfonamides (Figure S10&S12).44 The attack of a hydroxyl radical at the C-N bond of the benzene ring might result in the fragment derived from the pyrimidinyl portion (m/z 124), leading to the formation of the 2-amino-4,6-dimethoxypyrimidine product. The cracking of the N-containing benzene ring is attributed to the formation of the fragment ions m/z 197 (4-amino-N-(iminomethylene)benzenesulfonamide) and m/z 213 (4-(2-imino-4,6-dimethylpyrimidin-1(2H)-yl)cyclohexamine), as previously reported.45
Importantly, regarding the potential toxicity of the degradation products, neither of the intermediated formed through At and SMT photodegradation exhibit acute toxicity,41 supporting then the significant improvement of the resulting water quality.