Extensive Screening of Green Solvents for Safe and Sustainable UiO-66 Synthesis

. Zirconium based Metal-Organic Framework UiO-66 is to date considered one of the benchmark compound among stable MOFs and it has attracted a huge attention for its employment in many strategic applications. Large scale production of UiO-66 for industrial purposes requires the use of safe and green solvents, fulfilling the green chemistry principles and able to replace the use of N,N -Dimethyl-Formamide (DMF), which, despite its toxicity, is still considered the most efficient solvent for obtaining UiO-66 of high quality. Herein we report on a survey of about 40 different solvents with different polarity, boiling point and acidity, used for the laboratory scale synthesis of high quality UiO-66 crystals. The solvents were chosen according the European REACH Regulation 1907/2006 among those having low cost, low toxicity and fully biodegradable. Concerning MOF synthesis, the relevant parameters chosen for establishing the quality of the results obtained are the degree are the crystallinity, microporosity and specific surface area, yield and solvent recyclability. Taking into account also the chemical physical properties of all the solvents, a color code was assigned in order to give a final green assessment for the UiO-66 synthesis. Defectivity of the obtained products, the use of acidic modulators and the use of alternative Zr-salts have been also taken into consideration. Preliminary results lead to conclude that GVL (γ-valerolactone) is among the most promising solvents for replacing DMF in UiO-66 MOF synthesis.

source and DMF as solvent able to well dissolve H2-BDC and Zr salt. The formation of zirconium oxo-clusters required the presence of oxygen sources which can arise from water present in DMF, in the hydrated salts or from direct addition of water during synthesis. The use of acidic modulators and of proton scavengers have been also extensively investigated. 19,20 During the last years, several synthetic strategies have been tried for the synthesis of DMF-free UiO-66, involving the use of alternative methods such as mechanochemical approaches 21 , accelerated aging 22 and microwaves 23 or the use of nontoxic solvents such as water, green solvents from biomasses and from industrial production waste. 24 Water based syntheses were found to be effective for UiO-66 MOFs based on substituted X-BDC ligands (with X = F, Cl, Br, NO2) or nitrogen-containing BDC-like linkers such as 2,5-pirydine or 2,5-pirazine dicarboxylic acids (H2-PDC, H2-PyDC) with higher acidity than simple H2-BDC acid, which increased the solubility in water. 25,26 However, the use of water has proven to be not suitable for the preparation of H2-BDC UiO-66 due to its low solubility in water, while the zirconium salt is highly soluble. This difference in the "solution-state" make the crystals growth very difficult. 27 In fact, solvents play a dramatic role in influencing the chemical and environmental efficiency of a process, and also the preparation of MOF materials is tremendously affected by the medium used both on the properties of the material obtained and on its applicability. In almost every chemical production, it is urgent the need for developing new, safer, and cleaner synthetics routes in agreement to the regulatory requirements in the European Union aiming at reducing or banning the use of chemicals that may be harmful to human health or the environment. Representative is the case of the European REACH Regulation 1907/2006 that envisages a mechanism which forces companies to apply for an authorization if they want to use or distribute chemicals (including solvents) identified as Substances of Very High Concern (SVHC). 28 Equivalent rules apply to the manufacturing, the commercialization and use SVHC, and especially to largely used solvents. Health and safety regulations have nowadays a major impact on solvent selection and in particular, chlorinated hydrocarbons, aromatics, and dipolar aprotic solvents such as N,Ndimethylformamide (DMF), are identified as hazardous due to their well-known chronic toxicity effects. 28 In this context, it is fundamental the identification of safer media to replace those highly toxic and most of the fundamental and industrial research is looking for novel safe solvent candidates with a special attention to those deriving from waste of large manufacturing process or from biomasses. [29][30][31][32][33][34][35][36][37][38][39] The ideal solvent for the synthesis of H2-BDC UiO-66 is responsible to perform in many different directions. An effective solvent should be able to completely dissolve the reagents whilst playing a fundamental role in the formation of the crystalline structure. As an example, the commonly used DMF, has an active role in H2-BDC UiO-66 synthesis by entering the coordination sphere of the zirconium steering the formation of the crystal lattice formation. 40 In this contribution, we report an extensive study on the screening of solvents that could be eventually be selected as alternative for the synthesis of BDC-UiO-66. Based on our previous studies, 24,41-46 we have considered several different classes of solvents and grouped in terms of their chemical-physical features, e.g. dielectric constant (Ɛ) and boiling point (Bp). We have compared their efficiency to produce BDC-UiO-66 using a standard protocol and classified their performance based the ability to actually form the desired UiO-66 MOF, the full width at half maximum (FWHM) of (111) diffraction peak, the crystal size, the BET surface area achieved and the yield. The screening allowed to identify a small number of selected solvents which afforded high quality UiO-66 crystals as pure phase without the formation of co-products whereas those yielding co-products or MOFs with low crystallinity degree will be object of future dedicated studies.

Chemicals
All chemicals are commercially available and used without further purification. First, DMF synthesis without modulator of UiO-66 was tried according to literature 11 : ZrCl4 (0.5 mmol), one equivalent of H2-BDC and three equivalents of water were dissolved in 100 mL of DMF in a Teflon reactor and the mixture was sonicated until complete dissolution of reagents occurred. After this, the reactor was put in thermostated oven at 120°C. After 16 hours MOF was soaked in DMF, water and acetone to remove unreacted starting materials and the reaction media.
At the end the UiO-66 was put in oven at 80°C to remove the solvent used to soak. Figure 1 shows the general reaction scheme. In the second step the reaction was scaled down to verify the viability of the process using a lower amount of solvent and therefore a higher concentration of reagents to improve the efficiency of the process. The procedure was the same as reported above but the volume of the solvent amount was 2 mL and the ZrCl4 0.5 mmol. The final concentration of zirconium salt was 0.25M.
Finally, the procedure was optimized by adding 30 equivalents of acetic acid (AcOH).

Synthesis of UiO-66 using green solvents at the optimized conditions
The synthetic procedures were standardized in the following way: In a screw capped 4 mL vial, H2-BDC (83 mg, 0.5 mmol) was dissolved in the chosen solvent* (4 mL), followed by the addition of water (27 μL, 3 eq), acetic acid (860 μL, 30 eq) and ZrCl4 (116 mg, 0.5 mmol). The mixture was sonicated until dissolution. The vial was then put inside a Teflon reactor and placed in a thermostated oven at 120°C for 16 h. After completion of the reaction, the solid was centrifuged and washed with MeOH (two time one-hour soaking), water (one times one-hour soaking) and   thin-film transistor processing, the solvent color code is inspired by the GlaxoSmithKline, GSK, and CHEM 21 solvent selection guides for the pharmaceutical industry. [48][49][50][51] These guides account for the physical and (eco)toxicity properties of solvents by transforming them into a scale to determine their greenness. The color codes for bps, viscosity and crystal size of MOFs are reported in the table 3. Other than the parameters already used in previous work for the green assessment we added here a color code also for the crystallinity degree of the obtained MOF. DMF is also included as reference entry. These codes are also combined in a composite color incorporating all these requirements, to give a ranking by default and "ranking after discussion" of each solvent.
Particularly, the preferred solvents, i.e. solvents presenting a few issues, are displayed with the green color code; the yellow color code has been used for problematic solvents, i.e. solvents that can be used but their implementation may present issues or uncertainties; not recommended solvents are identified with red color code, i.e. the constraints on the solvent use are very high.
Boiling points are already considered in the overall greenness assessment but, a separate column was included in Table 2 listing the solvent b.ps, to take specifically into account their suitability for UIO-66 synthesis. The associated color codes are based on the ranges defined in Table 3a.
Plurafac LF 120 Consequently, the ranking "by default" color code is dominated by the crystal size (Table 3b).
Finally, a ranking "after discussion" was assessed ( Table 2, last column), as the result of an overall evaluation of the solvent greenness, MOFs crystal size and surface area, and so on further improvement and possible applications based on the obtained morphological features. First, in the ranking "after discussion" DMF, despite the quality of the MOF obtained, is listed as "not recommended" due to its critical toxicological profile. We did not modify the ranking by default for "preferred" solvents ("green" code) leading to UIO-66 MOFs with crystal size higher than 450 Å, an arbitrary value we chosen in order to establish a good crystallinity degree. Moreover, we assigned the color "green" to those solvents leading to a "green" crystal size but that featured present an intermediate BET surface area (yellow code). The reason for this choice is the fact that further enhancement in the MOF synthesis could lead to higher values.
Furthermore, we assigned "yellow code" to the combination "green" crystal size/ "red" surface area and "red" code to the combination "yellow" crystal size/ "red" surface area and vice versa.
Crystal size values lower than 300 Å regardless of the BET surface area value achieved a red score.   Despite the many experiments and the tentative optimization of specific conditions, most of the solvents screened were not able to successfully allow the MOF formation. In these cases, UiO-66 of AcOH yielded compounds with high crystallinity degree and porosity comparable to those found for UiO-66 of good quality. 55 The exception is the experiment carried out in Steposol (ST) (entry 39bis) in which the best crystallinity was obtained without modulators. Syntheses at the optimized conditions which gave the best results are those made in the following solvents: γ-Valero-lactone (entry 11) Propylene carbonate (PC) (entry 23) 1-(2-hydroxyehtyl)-1-pyrrolidone (1-(2-HE)-2-P) (entry 2), t-amylmethylether (TAME) (entry 37) and 1,3-propandiol (1,3-PD) (entry 3). In two cases, those done in dimethyl carbonate (entry 22) and in dimethyl adipate (entry 6), the syntheses were successful but, due to the presence of other phases, the optimization was not done. Table 4 reports on the main features of obtained materials.  Figure 2 shows the XRPD patterns and the N2 adsorption and desorption analysis of the MOFs obtained (entries 1-6).
The materials are pure phases with good degree of crystallinity and BET surface area values comparable to conventional UiO-66 syntheses. 18 The MOF obtained in 1-(2-HE)-2-P presents the XRPD pattern having characteristic peaks of a defective structure, with broad signals in the low angle regions, as previously reported by Bordiga et al. 56 This is also confirmed by the high specific surface area and micropore volume which is typical of defective UiO-66 with missing clusters defects. 18 The experiment carried out in GVL is quite relevant since the MOF possesses a good crystallinity degree and it was obtained at a high yield (84%). Moreover, the GVL is a solvent which is rapidly replacing DMF, especially in the field of organic synthesis. 45 Another interesting result has been obtained by using Steposol (ST). Figure 3 shows the XRPD patterns and the BET analysis of the MOFs obtained with ST as solvent using the same reaction condition above reported except the use of acetic acid as modulator. As in the case of 1-(2-HE)-2-P, in this solvent the MOF obtained has an XRPD pattern with characteristic broad low-angle peaks of a defective structure, as also confirmed by its higher specific surface area than the MOF obtained using DMF. (see table 3). The only case in which the solvent could be included in the compound formula is the UiO-66 synthesized in PC. As a matter of fact this solvent was not removed before starting decomposition, as observed from the TGA curve ( figure S8). The presence of the solvent could be also responsible of the low specific surface area if compared to the other syntheses. PC 1 H-NMR peaks (figure S12) on the hydrolyzed MOFs confirmed this hypothesis. Figure    In fact, they are biodegradable and with no bioaccumulation issues. All the solvents here proposed, show a better and less critical toxicological profile then DMF itself. Focusing on the dermal LD50 (rabbit) data listed in the table above, we can observe an important increase of this value in all the green solvents tested. Moreover GVL, PC and Steposol, are not considered to be potential carcinogenic by IARC, while DMF is classified as "possibly carcinogenic to humans" (IARC Group 2B), and for this reason is labelled as SVHC by the European REACH regulation.
The use of new solvents deriving from biomass and/or industrial waste is a definitely cleaner and safer opportunity, although currently it may result to be more expensive, especially in the case of biomass-derived chemicals. This is certainly related to the fact that their preparation is based on different transformations but also that this area of chemical production is still at its infancy and as the technologies related to this industrial sector will be optimized and the proper raw materials made available, the cost associated to processes is expected to decrease with the increase of the market requests.