Electrochemically-Mediated Synthesis of Titanium(III)-Based Metal-Organic Frameworks

Abstract

Although metal-organic frameworks featuring coordinatively-unsaturated transition metal sites are relatively common, examples with redox-active cations are rare. In this report, we describe the electrochemically-mediated synthesis of Ti^III-MIL-101 from the inexpensive Ti⁴⁺ precursor TiCl₄. The framework obtained via electrosynthesis is identical to that prepared from the significantly more expensive and air-sensitive starting material TiCl₃. The above electro-synthetic strategy was also extended to prepare Ti^III-MIL-100 for the first time. This material represents only the second example of a Ti³⁺ based MOF. Gas adsorption, powder X-ray diffraction, SEM, XPS, and TGA experiments demonstrate that this material is superior in quality to samples of the analogous MOF prepared via conventional routes from starting TiCl₃. Given the ease with which the electrosyntheses may be carried out and their compatibility with a broad range of bridging ligands, we expect that this new methodology will find utility for the synthesis of a number of novel materials containing coordinatively-unsaturated, redox-active metal cations.

Graphical Abstract

Metal-organic frameworks (MOFs) have garnered considerable attention, as they display high surface areas,^1–3 tunable pore chemistry,^4–6 and can be prepared from virtually any transition metal cation. These factors distinguish MOFs as highly tunable for numerous applications.⁷ Of the broad material space that is occupied by MOFs, those featuring accessible metal cations on their pore surface, are particularly interesting as such systems can facilitate selective gas binding,⁸ high-density gas storage,⁹ sensing,¹⁰ and serve as sites for small-molecule activation and catalysis.¹¹ However, particularly for the latter which require redox-active metal cation sites, the number of reported materials is limited.^12–16 For example, although numerous Cr³⁺ MOFs exist, there are only two examples containing this metal in the 2+ oxidation state.¹⁷ Similarly, many examples of Ti⁴⁺ metal-organic frameworks have been reported,^18–20 but to date there is a single Ti³⁺-based MOF.²¹ This material, Ti^III-MIL-101 (Figure 1), is based on trimeric clusters connected to terephthalic acid (H₂bdc) and features a high BET surface area, exposed Ti³⁺ sites, and strong Ti–O₂ interactions via the formation of peroxide adducts. Further, its exposed Ti³⁺ sites are substrate accessible and can strongly bind gas molecules such as CO₂ and H₂.

Figure 1.

Portions of the MIL-100 and MIL-101 structure types that depict the large pores in the materials. Use of H₂bdc generates MIL-101; use of H₃btc generates MIL-100.

The preparation of redox-active MOFs generally relies on the use of air-sensitive starting materials, necessitating the use of rigorously air-free reaction conditions and limiting materials discovery throughput. Moreover, for many metal ions in reduced oxidation states, the number of available starting materials are limited, as the synthesis of a metal-organic framework often relies on screening a number of parameters, including counter anion identity, placing additional bottlenecks on discovery of new structures.²² In addressing these limitations, electrochemical methods present an intriguing alternative strategy for synthesis of reduced MOF materials.

Redox reactivity has been coupled to synthesis of MOFs. For example, redox chemistry has been used to drive ligand deprotonation reactions coupled to MOF syntheses.^23–25 Additionally, electrochemical dissolution of metallic anodes to generate oxidized metal ions (i.e., M⁰ → Mⁿ⁺), has been utilized to synthesize Cu²⁺ and Zn²⁺ based MOFs. Despite these examples, we are aware of no published examples, in which electrochemistry has been leveraged to alter the oxidization state of a dissolved metal ion in solution prior to MOF synthesis. Such a strategy is attractive as it can allow for the oxidation state of dissolved ions to be rapidly tuned via controlled potential electrolysis (CPE) for the targeted synthesis of MOFs containing metal centers in non-traditional, reduced-oxidation states. Additionally, this strategy permits the large number of high-oxidation state metal starting materials that are inexpensive and readily accessible to be converted to higher value reduced forms and facilitate the discovery of new materials containing metal ions in reactive oxidation states. This method also has the advantage that a larger number of reaction conditions can be screened, as they can be set up in the presence of air, then deoxygenated prior to electrochemical reduction. To this end, we targeted the synthesis of novel Ti³⁺ frameworks given their scarcity and potential for small molecule storage and activation.

The Ti³⁺ based MOF Ti^III-MIL-101 can be prepared by reaction of exogenous TiCl₃ with terephthalic acid (H₂bdc) at 120 °C in 10:1 N,N’-dimethylformamide (DMF):ethanol for 18 hours under rigorously air-free conditions. In order to prepare this material from Ti³⁺ generated electrochemically from Ti⁴⁺ sources, we utilized cyclic voltammetry (CV) to determine potentials (Figures S1 – S3) at which Ti³⁺ can be generated from a series of Ti⁴⁺ based sources (e.g., TiCl₄, Ti(OⁱPr)₄, Cp₂TiCl₂). In particular, CVs recorded for solutions of TiCl₄ (Figure 2a) showed a cathodic wave corresponding to Ti⁴⁺ reduction at E_cat = −1.20 V (all potentials reported versus Ag/AgNO₃). Preparative scale reduction of TiCl₄ by CPE at E_applied = −1.20 V using a nickel electrode could be monitored by UV-vis spectroscopy (Figure 2b,c). The complete reduction of starting Ti⁴⁺ to Ti³⁺ was accomplished in as little as 60 minutes. A four-hour CPE (E_applied = −1.20 V) of TiCl₄ in the presence H₂bdc and 0.1 M TBAPF₆ as supporting electrolyte in a 10:1 DMF/EtOH mixture using a nickel foam working electrode, followed by heating at 120 °C for 18 hours afforded a crystalline dark purple powder in high yield.

Figure 2.

(a) Cyclic voltammogram recorded of TiCl₄ (1.0 mM) in 10:1 DMF:EtOH containing 0.1 M TBAPF₆ at a scan rate of 100 mV/s. (b) UV-Vis absorption spectra of a TiCl₄/H₂bdc DMF:EtOH solution containing 0.1 M TBAPF₆ during CPE at E_applied = −1.20 V. (c) Change in absorbance at = 650 nm during CPE described in (b). (d) Total current density and overall Faradaic efficiency for TiCl₄ and H⁺ reduction from TiCl₄/H₂bdc DMF:EtOH solution (E_applied = −1.20 V).

Powder X-ray diffraction experiments indicated the resulting material is isostructural to the previously reported Ti^III-MIL-101 phase (Figure S6). The structure is comprised of trimeric Ti³⁺ clusters connected to form tetrahedral building units with bdc^2– anions on their vertices. The tetrahedral units are edge linked to afford a structure containing two large mesoporous cages, one with both pentagonal and hexagonal windows and the other comprised solely of pentagonal windows. Ti^III-MIL-101 obtained via the previously reported solvothermal method from starting Ti³⁺ displays a BET (Langmuir) surface area of 2970 (4440) m²/g.²¹ The corresponding material synthesized via electrochemical generation of Ti³⁺ from TiCl₄ has a BET (Langmuir) surface area of 3285 (4360) m²/g. Although PXRD, SEM, XPS, and TGA analyses indicate the solvothermally and electrochemically generated materials are indistinguishable, synthesis of the latter is accomplished using inexpensive TiCl₄, as opposed to relying on exogenously prepared and highly air-sensitive TiCl₃. Of further note is the fact that commercial TiCl₄ is routinely available in 99.9% (and higher purity), while TiCl₃ is typically prepared via aluminum reduction that result in Ti³⁺ starting materials containing appreciable levels of Al³⁺.

A particular advantage of electrochemically-synthesized reduced metal frameworks, is the management of proton balance throughout the reaction. Significant concentrations of H⁺ are generated during the course of solvothermal MOF syntheses as ligands are deprotonated to form M–L bonds. As many MOF syntheses are dependent on pH, the continually increasing proton concentration can complicate synthesis of phase-pure materials. The electrosynthetic strategy described above proves a convenient means to manage H⁺ build up associated with formation of the bdc^2– anions needed for Ti^III-MIL-101 synthesis. The nickel foam working electrodes used to generate the Ti³⁺ from TiCl₄ show excellent HER activity at significantly lower potentials than those used for the reduction of Ti⁴⁺.²⁶ Accordingly, CPE of the TiCl₄/H₂bdc solutions with the nickel cathode not only generates Ti³⁺, but also reduces the carboxylic acid protons to generate the bdc^2– ligands and H₂, which is evolved from the cathode compartment.

To follow the production of H₂ over the course of the CPE, we monitored the reaction headspace during Ti^III-MIL-101 electrosynthesis via gas chromatography.^27,28 Initial formation of H₂ was detected within as little as 5 min and plateaued after ~3–5 hours. Initial Faradaic efficiencies for HER were <20% and slowly increased to ~70% as concentrations of Ti⁴⁺ were depleted to generate Ti³⁺. In accounting for both the reduction of TiCl₄ to Ti³⁺ and H₂bdc (or H₃btc, vide infra) protons to H₂, we consistently observed net current efficiencies in excess of 75 % over the course of the entire CPE.

Given the potentially broad applicability of MOF electro-synthesis from soluble metal precursors, we sought to employ this method to expand the library of Ti³⁺ MOFs by targeting other multitopic carboxylate-based ligands. Since the μ₃-O centered trimeric cluster is ubiquitous to a large number of MIL-type MOFs, we targeted H₃btc-based MIL-100. This structure has previously been reported for trivalent Al,²⁹ Sc,³⁰ V,³¹ Cr,³² Fe,³³ and mixed-metal systems and features connectivity that is analogous to the MIL-101 structure type, but where ligands occupy sites on the faces rather than vertices of the tetrahedral building units. The electrochemical reduction of a TiCl₄/H₃btc DMF:EtOH (10:1) solution containing 0.1 M TBAPF₆ followed by reaction at 120 °C for 18 hours afforded the target material, as confirmed by powder X-ray diffraction (Figure 3). Interestingly, identical solvothermal reaction conditions utilizing commercial TiCl₃ rather than electrochemically reduced TiCl₄ afforded MOF powders with significantly diminished crystallinity (Figure 3). After thorough solvent exchanges and activation at 100 °C the solvothermally and electrochemically prepared materials showed BET (Langmuir) surface areas of 1304 (1568) and 1736 (2708) m²/g, respectively (Figure 3a). The latter value is in good agreement with those reported for analogous MIL-100 materials.^30,32

Figure 3.

(a) N₂ adsorption in Ti^III-MIL-100 at 77 K. (b) PXRD patterns of electrochemically- (blue) and solvothermally-synthesized (red) Ti^III-MIL-100 as compared to the predicted powder pattern (black). SEM images for (c) solvothermally- and (d) electrochemically-synthesized Ti^III-MIL-100.

Despite several attempts, solvothermal syntheses utilizing exogenous TiCl₃ as starting material consistently afforded frameworks with lower surface areas compared to those obtained using Ti³⁺ that had been electrochemically generated from TiCl₄. The lower surface areas obtained for the solvothermal materials may coincide with a Ti⁴⁺-based framework with MIL-100 topology exhibiting a BET surface area of 1321 m²/g that was recently reported.³⁴ We note that XPS analysis of solvothermally-prepared samples of both Ti^III-MIL-101 and Ti^III-MIL-100 show that each retain chloride as a charge-balancing anion in the frameworks. Electrochemically-prepared samples of these two Ti³⁺-based frameworks both contain fluoride anions (in place of chloride), which is presumed to originate from the TBAPF₆ supporting electrolyte.

To determine whether the presence of supporting electrolyte improved the surface area of the electrochemically prepared Ti^III-MIL-100, TBAPF₆ was added to the solvothermal reaction mixture (exogenous TiCl₃/H₃btc). The thusly obtained dark brown product was both amorphous and nonporous. Further, solvothermal syntheses with exogenous TiCl₃/H₃btc consistently afforded materials of smaller particle size (Figure 3c) as compared to syntheses utilizing electrochemically reduced TiCl₄ (Figure 3d).

Given the near complete dearth of Ti³⁺-based metal-organic frameworks, we sought to further expand the scope of this method to extended pore structures. Although materials of this type based on the M₃O building block have been reported, they typically feature surface areas that are significantly lower than calculated values. MIL-101(Fe)_BPDC, which is a biphenyldicarboxylate-based material that adopts the MIL-101 structure type, displays a BET surface area of just 210 m²/g as compared to the theoretical value of 4500 m²/g. A MIL-100 structure based on an expanded tritopic linker displays a BET surface area of just 26 m²/g vs. the expected value of 3990 m²/g.³⁵ Analogous to the synthesis of Ti^III-MIL-101 and Ti^III-MIL-100, the electrochemical reduction of a TiCl₄ solution containing either 4,4’-biphenyldicarboxylic acid (H₂bpdc) or 4,4’,4”-[1,3,5]triazine-2,4,6-triyl-tris-benzoic acid (H₃tatb) followed by solvothermal reaction affords Ti^III-MIL-101-bpdc and Ti^III-MIL-100-tatb, respectively. Ti^III-MIL-101-bpdc displays a BET surface area of 3263 m²/g and a pore volume of 2.20 cm³/g which are greatly diminished for a solvothermally-prepared sample (2139 m²/g and 1.32 cm³/g). Similarly, electrochemically-synthesized Ti^III-MIL-100-tatb has a BET surface area and pore volume of 5842 m²/g and 4.04 cm³/g as compared to just 740 m²/g and 0.49 cm³/g for solvothermally-synthesized MOF. The pore volume of the electro-chemically-synthesized material is in excellent agreement with the value reported for the analogous Al-based MOF, PCN-333 (3.81 cm³/g).

In conclusion, we have leveraged electrochemical methods to synthesize reduced metal-organic frameworks. Electro-chemical synthesis of Ti^III-MIL-101, Ti^III-MIL-100, Ti^III-MIL-101-bpdc, and Ti^III-MIL-100-tatb using the inexpensive, accessible and highly pure Ti⁴⁺-based starting material, TiCl₄ generated frameworks with crystallinities and adsorption properties that are as good or superior to those of materials synthesized using exogenous/commercial TiCl₃. Furthermore, the electrochemically mediated synthesis of Ti^III-MIL-100 has provided to first reported route to this MOF.

The electrosynthetic methods disclosed in this report not only generate the high-purity Ti³⁺ ions from which Ti^III-MIL-101 and Ti^III-MIL-100 structure types are built, but also provides a means to manage solution pH, as the H⁺ ions generated upon formation of the frameworks’ Ti³⁺–O₂C–Ar bonds are readily reduced and evolved as H₂ gas. Accordingly, we believe the methods outlined herein provide a new general strategy for the electrochemically mediated synthesis of new redox-active metal-organic frameworks and for the isolation of reduced MOF materials.

Figure 4.

SEM images for (a) electrochemically- and (b) solvothermally-synthesized Ti^III-MIL-101-bpdc (c) N₂ adsorption in MIL-101-bpdc at 77 K. SEM images for (d) electrochemically- and (e) solvothermally-synthesized Ti^III-MIL-100-tatb (f) N₂ adsorption in MIL-100 tatb at 77 K.