Titanium coordination compounds: from discrete metal complexes to metal–organic frameworks

Abstract

Owing to their promise in photocatalysis and optoelectronics, titanium based metal–organic frameworks (MOFs) are one of the most appealing classes of MOFs reported to date. Nevertheless, Ti-MOFs are still very scarce because of their challenging synthesis associated with a poor degree of control of their chemistry and crystallization. This review aims at giving an overview of the recent progress in this field focusing on the most relevant existing titanium coordination compounds as well as their promising photoredox properties. Not only Ti-MOFs but also Ti-oxo-clusters will be discussed and particular interest will be dedicated to highlight the different successful synthetic strategies allowing to overcome the still “unpredictable” reactivity of titanium ions, particularly to afford crystalline porous coordination polymers.

---

Hala Assi

Hala Assi, 26 years old, originates from Lebanon. She achieved her Master’s in Materials Chemistry in 2013 at Lebanese University of Beirut. She received her PhD in 2016 from the University of Versailles St Quentin-en-Yvelines/Université Paris Saclay, France, under the supervision of Dr Christian Serre and Dr Thomas Devic. In 2017, she joined the group of Dr Myrtil Kahn at Laboratoire de Chimie de Coordination CNRS, Toulouse, France, as postdoctoral researcher. Her current research focuses on the synthesis of nano-catalyst materials for water splitting.

Georges Mouchaham

Georges Mouchaham was born in Lebanon. He received his MSc (2009) and PhD (2012) degrees in Macro- and Supramolecular Chemistry from the University Paul Sabatier of Toulouse, France, under the supervision of Dr Jean-Pascal Sutter and Dr Nans Roques. In 2013, he joined the “Porous Solids” group at the Institut Lavoisier de Versailles, France, where he worked as a Postdoctoral Fellow under the supervision of Dr Christian Serre and Dr Thomas Devic. Since 2016, he has been a Postdoctoral Fellow at King Abdullah University of Science and Technology (KAUST), Saudi Arabia, working under the supervision of Prof. Mohamed Eddaoudi. His research interests mainly focus on the design, elaboration and study of robust MOFs.

Nathalie Steunou

Nathalie Steunou is professor at the Institut Lavoisier de Versailles (ILV) from the University of Versailles St Quentin-en-Yvelines/Université Paris Saclay. She completed her PhD on the synthesis of Ti-oxo clusters in 1997 in the Laboratory Chimie de la Matière Condensée de Paris at the University Pierre & Marie Curie (UPMC-Paris VI). After a postdoctoral position at the MPI for Colloid and Interfaces (Berlin, Germany), she rejoined the LCMCP as a lecturer for about 11 years working on the development of hybrid materials by assembling metal oxides and biomolecules (biopolymers, enzymes, etc.). In 2010, she moved to ILV reaching her present position. Her current research interests focus on the synthesis of nanomaterials and composites based on metal–organic frameworks for different applications (energy, biosensing, etc.).

Thomas Devic

Thomas Devic studied Chemistry at the Ecole Normale Superieure de Lyon and received his PhD in 2003 from the University of Angers, France, in the field of conducting molecular materials. After a postdoctoral stay at the Cornell University, USA, in the group of Prof. Stephen Lee (2003–2004), he was appointed as a CNRS researcher in the research group “Porous Solids” headed by Gérard Férey at the Institut Lavoisier de Versailles, France. In 2016, he moved to the Institut des Matériaux Jean Rouxel, Nantes, France. His research mainly deals with the synthesis, structural characterization and study of the properties of new hybrid solids, with a special emphasis on the use of non-conventional ligands, in situ studies and structure–property relationships, and more recently their applications in the field of electrochemical energy storage.

Christian Serre

Christian Serre obtained his PhD in 1999 from the University of Versailles St Quentin. After a post-doctoral fellowship in the USA, he became a CNRS researcher in the ‘Porous Solids’ group at Institut Lavoisier (Versailles), taking its lead in 2009 as research director. He moved recently to the Ecole Normale Supérieure and ESPCI (Paris). His research topics deal with the synthesis, structure and applications of porous solids. Christian has received the CNRS bronze medal, a ERC starting grant, the Solid State Chemistry award of the French chemical society and recently the Fondé de l’Etat award from the French academy of sciences.

---

Introduction

Almost two decades after their popularization and uninterrupted exploration, metal–organic frameworks (MOFs) are still fascinating the research community. Due to their tunable crystalline structure, in terms of chemical nature, molecular arrangement as well as size and shape of cavities, MOFs have shown very promising potential applications in very diverse fields such as gas storage and separation, biomedicine, sensing, heterogeneous catalysis, etc.^1,2 Among the numerous MOFs described in the literature, most systems are built up from divalent transition metal cations (Zn²⁺, Cu²⁺, Co²⁺, Ni²⁺, etc.), leading to thousands of open framework architectures with sometimes unprecedented porosity, but often unstable in the presence of water. Increasing the charge of the metal cation (together with its polarizing power: charge over ionic radius) is a powerful strategy to enhance the strength of the cation–ligand bond and thus the chemical stability of the coordination polymer.^3,4 While the use of trivalent cations (Cr³⁺, Al³⁺, Fe³⁺, etc.) has been successful in obtaining porous water stable MOFs, tetravalent cation based MOFs are still scarce. This is mainly due to the high reactivity of such cations leading to the precipitation of their corresponding oxide or, in the best cases, to amorphous or poorly crystallized coordination polymers. One noticeable exception is the Zr(IV). The discovery of the first Zr-carboxylate MOF, UiO-66,⁵ built up from the very robust hexa-nuclear Zr-oxo-cluster, and the easy development of isoreticular or related solids, together with the possibility of using modulators⁶ (allowing tuning of the reactivity of the cations and, hence, the crystallization process), has paved the way to the extremely fast progress in the field of Zr(IV)-MOFs.^7–10

Developing renewable, sustainable and ecological resources has become one of the actual main concerns in order to overcome the problem of the controversial traditional energy sources. Mimicking the natural photosynthesis (in terms of photochemical conversion of light) seems to be a very appealing alternative. Titanium dioxide, TiO₂, is considered nowadays as one of the most successful photocatalysts due to its high efficiency, high chemical stability, high abundance and rather low toxicity. Besides, MOFs, among other materials, have been proven to be very promising for heterogeneous catalysis¹¹ because of their ability to ensure, concomitantly, high accessible surface area, high concentration of actives sites, confined spaces and very good diffusion towards the active sites.

In this context, titanium as a tetravalent Ti⁴⁺ cation appears to be a very attractive candidate to produce MOFs showing not only good chemical stability but also redox activity and photocatalytic properties (Ti⁴⁺ cation may be reduced to Ti³⁺ by photoreduction). While many research groups have focused their attention on the incorporation of this cation, for example in the form of titanium oxo clusters,^12,13 into organic (block copolymers, polymers, dendrimers), inorganic and hybrid organic–inorganic matrices, porous or not,^14–16 such materials do not generally exhibit any long-range order and are almost amorphous.

This review aims at giving an overview of the coordination chemistry of titanium cations and oxygenated ligands, with a specific focus on crystalline compounds. While extended coordination polymers, particularly porous solids (MOFs), will be exhaustively described, the most relevant molecular compounds will also be presented. Besides, attention will be paid to highlight synthetic routes allowing overriding the difficulty encountered with Ti-cations in order to obtain crystalline Ti-MOFs. Relevant photoredox properties of Ti-MOFs and selected Ti-oxo-clusters will also be discussed.

The chemistry of titanium cations

Even though Ti⁴⁺ is a highly attractive metal cation, its use to elaborate photoactive materials (other than TiO₂) remains still rather challenging. The +4 oxidation state is the stable one in ambient (air, water) conditions (see Fig. 1), but due to the low electronegativity of Ti and the high polarizing power of Ti⁴⁺, the latter is prone to fast and spontaneous hydrolysis even in very acidic medium leading to uncontrolled precipitation of TiO₂. This is demonstrated in the titanium Pourbaix diagram (Fig. 1) of Ti in water, where the area of titanium(IV) oxide is strongly dominating, the soluble molecular species occurring only in highly acidic media (pH ≤ 0).¹⁷

Fig. 1 Pourbaix diagrams of titanium calculated for [Ti⁴⁺] = 10⁻³ mol L⁻¹ at 25 °C using the Hydra and Medusa softwares.¹⁷ Green dashed lines: O₂/H₂O and H₂O/H₂ redox couples.

Consequently, the choice of the titanium precursor and the power of the complexing ligand together with the pH of the media are key factors to stabilize this ion in solution. Titanium(IV) coordination chemistry is primarily based on the use of two types of precursors: titanium(IV) chloride and titanium(IV) alkoxides. It is noteworthy that for both precursors, the coordination number of Ti is lower than that of TiO₂ (i.e. 6), which explains the high reactivity of these precursors towards water. Due to the violent hydrolysis and fast release of hydrochloric acid under ambient atmosphere, the use of TiCl₄ is quite scarce, although it was used in the production of monodisperse and pure anatase, brookite or rutile nanoparticles.¹⁸

Because of the risky experimental manipulation of TiCl₄, titanium alkoxides, Ti(OR)₄, are often preferred since they are less readily hydrolyzed than TiCl₄ due to the hydrophobic character of alkoxy groups and/or their steric hindrance. Moreover, Ti cations can also be complexed by a wide range of additional ligands (β-diketonate, carboxylates, phosphonates, catecholates, etc.) which, due to their bidentate characters, lead to an increase of the coordination number, thereby enhancing the stability of these species towards water.

Another way to control the hydrolysis-condensation reaction of Ti lies in the use of Ti-oxo-clusters of well-defined structure. These species of the general formula Ti_nO_m(OR/Cl)_x(L)_y are polynuclear titanium complexes that combine a metal oxo-core and organic complexing ligands (L, again alkoxide, β-diketonate, carboxylate and phenolate). They present a high diversity in terms of size, geometry, degree of O/Ti condensation, coordination number of titanium cations, and organic complexing ligands. These oxo-clusters are generally more stable than titanium alkoxides towards water since they present a higher nuclearity and condensation degree (O/Ti ratio) and possibly ligands (β-diketonates, carboxylates and/or phenolate) that are less hydrolysable than alkoxides. They are usually prepared either at the early stages of the corresponding titanium alkoxide polycondensation or from the controlled hydrolysis of the titanium alkoxide Ti(OR)_x(L)_y in solution.¹⁹ For numerous titanium-oxo clusters, the source of oxolation was also water generated in situ in the reaction medium through either esterification reactions^20,21 or condensation–dehydration reactions of ketones²² while non-hydrolytic routes were also reported.²³ As discussed later, such species can advantageously be used as precursors to build up extended coordination compounds.

Early structures: silicates, phosphates and phosphonates

Before analyzing in depth the case of hybrid Ti-based compounds, some of the first extended inorganic Ti-based structures need to be addressed. The reactivity of TiCl₄ with inorganic anionic ligands such as fluorides, chlorides, sulfates and phosphates has indeed been thoroughly studied.^24–28

In the field of inorganic molecular sieves,²⁹ the first reported microporous materials were zeolites, natural aluminosilicates, known since the 18th century. Since then the attention was turned to the development of new synthetic molecular sieve materials, composed from interconnected octahedral and tetrahedral oxide polyhedra, to obtain new topologies, novel compositions or larger pores.³⁰ Particular attention has been devoted to the introduction of active heteroelements into the frameworks mainly for catalytic or separation purposes. Microporous titanium silicates have particularly attracted interest since the discovery of the first titanium silicate TS-1, an active catalyst in oxidation reactions.³¹ In 1989, a new family of microporous three dimensional titanosilicate frameworks was discovered and named ETS (ETS stands for Engelhard Corporation Titanosilicate). The structure of ETS-10 or A₂TiSi₅O₁₃·4H₂O (A = Na⁺ or K⁺) was reported in 1994 by Anderson et al.,³² and consists of a mixture of TiO₆ octahedra and SiO₄ tetrahedra connected together to form a microporous three dimensional structure. This material is an active base catalyst in the dehydration of tertbutyl alcohol,^33–35 or in the dehydrogenation of alcohols.^35,36 This is due to the presence of TiO₆ bearing two negative charges balanced by K⁺ and Na⁺ cations. The effect of ionic-exchange has also been studied showing the fragility of this material upon treatments under acidic conditions.³⁷ In addition, ETS-4, known also as a precursor for the synthesis of microporous titanium–niobium–silicate materials,³⁸ has been investigated as a size-selective adsorbent.³⁹ By adjusting the pore size of this material, it can be used for the separation of gas mixtures (N₂/CH₄, Ar/O₂ and N₂/O₂).

Focusing on crystalline titanium phosphates, these compounds have been investigated for ion exchange, heterogeneous catalysis, proton conduction, etc.⁴⁰ They are generally obtained by refluxing insoluble salts of amorphous titanium phosphates, already obtained from TiCl₄ under acidic conditions in the presence of an excess of phosphoric acid. Alberti et al.⁴⁰ reported almost 50 years ago the synthesis of monohydrated titanium phosphates of the formula Ti(HPO₄)₂·H₂O while the di-hydrated form Ti(HPO₄)₂·2H₂O was disclosed in 1977 by Alluli et al.⁴¹ The degree of hydration strongly affects the nature of the resulting structure and the arrangement of layers (αTiP and γTiP) – the interlayer distance is ca. 11.6 Å for the di-hydrated form and ca. 7.56 Å for the monohydrate – and the arrangement of the phosphate groups in the lattice. Similar observations have been reported later for zirconium phosphates by Clearfield et al.⁴² Noteworthily, these changes in the crystal structure lead to large differences in the resulting ionic exchange properties. The dihydrated titanium phosphates have shown high stability towards hydrolysis as well as a high ionic exchange capacity in a strongly acidic medium for sodium and strontium ions. At that time, the crystalline structure of these two arrangements of layers was not known. Though, in 1990, the structure of αTiP was solved from PXRD data,⁴³ these two compounds were found to be very similar. Their structure consists of layers of TiO₆ octahedra connected together by phosphate groups to form a two-dimensional structure. In the case of αTiP, HPO₄ groups are distributed alternately above and below the Ti octahedral plane while in the case of γTiP these groups are either PO₄ or H₂PO₄. Later, in 1997, Bortun et al.⁴⁴ studied in detail the TiCl₄/H₃PO₄/H₂O system in order to discover new structures. Two new crystalline fibrous titanium(IV) oxophosphates were obtained, in addition to αTiP, denoted πTiP and ρTiP.

This work has been later extended through the in depth investigation of these systems in the presence of fluorine and/or amines or surfactants leading to a series of new structures denoted MIL-n (MIL stands for Materials of Institut Lavoisier). Hydrothermal conditions were selected to study the formation mechanisms of micro- or mesoporous titanium phosphates, oxyfluorinated or not. The TiO₂/HF/H₃PO₄/amine/water system was the starting point of this work, and two layered oxyfluorinated titanium phosphates were isolated, Ti₂(PO₄)₂F₄·(N₂C₂H₁₀) (MIL-6₂) and Ti₂(PO₄)₂F₄·(N₂C₃H₁₂)·H₂O (MIL-6₃) using either ethylenediamine or 1,3-diaminopropane, respectively.⁴⁵ Their structures were solved from PXRD data and consist of TiO₄F₂ octahedra connected together by PO₄ tetrahedra. The diammoniums are located in the interlayer space and strongly hydrogen bonded to the terminal fluorine atoms, further stabilizing these structures. A mixed valence metal solid, MIL-15 or Ti^IIITi^IVF(PO₄)₂·2H₂O, was also isolated; this three-dimensional framework is built up from TiO₆ or TiO₄(H₂O)₂ octahedra and PO₄ groups delimiting small channels. Noteworthily, a microporous solid labeled MIL-18 or Ti^IV₆O₃(H₂O)₃–(PO₄)₇·(X)₃·H₂O (X = H₃O or NH₄) based on dimers of Ti octahedra was also reported however as an impurity preventing the study of its properties.⁴⁶ Finally, MIL-28n or Ti₃O₂X₂H_y(HPO₄)_x(PO₄)_y·(N₂C_nH_2n+6)_z·2H₂O (n = 2, 3; x = 0, 2; y = 4; z = 3, 2) were obtained; these layered solids, built up from corner sharing chains of TiO₅X octahedra and HPO₄ groups, possess perforated layers where the protonated diamines interact with the terminal Ti–X groups (X = F and OH).⁴⁷ However, in general, the use of charged templates raises several drawbacks, mainly the impossibility to access porosity due to the destruction of the structure upon removal of the organic moieties.

Thus, the use of diphosphonic acid instead of phosphoric acid, as initially proposed by Alberti, Clearfield et al.,^40,42 has been shown to be an alternative route to develop open framework solids without the use of templates. This expansion to hybrid (coordination) compounds led to 3-D open framework structures such as MIL-22 or Ti^IV₃O₂(H₂O)₂(L)₂·2H₂O with L = (O₃P-(CH₂)-PO₃)₂ or methylenediphosphonate. Noteworthily, this solid built up from trimers of TiO₆ or TiO₅(H₂O) octahedra and PO₄ groups possesses a 2-D array of microporous channels filled with free and coordinated water molecules, although their removal does not lead to any accessible porosity towards nitrogen (Fig. 2a).⁴⁸ Another series of 3-D pillared-titanium bisphosphonate structures, MIL-25n (n = 2, 3) formulated Ti^IV(L) with L = O₃P–(CH₂)_n–PO₃, although nonporous, has been reported by Serre et al. (MIL-25₂ in Fig. 2b and MIL-25₃ in Fig. 2c).⁴⁶

Fig. 2 Crystalline structures based on diphosphonate linkers: (a) MIL-22,⁴⁸ (b) MIL-25₂,⁴⁶ (c) MIL-25₃⁴⁶ and (d) MIL-91(Ti).⁴⁹ Color code: Ti, blue; O, red; C, grey; P, orange; N, dark blue. H-atoms of C-atoms were omitted for the sake of clarity.

In 2006, Serre et al.⁴⁹ reported the first porous titanium-based metal–organic framework, denoted MIL-91 or Ti^IVO(H₂-L)·nH₂O (n ≈ 4.5) with L = O₃P–CH₂–NC₄H₈N–CH₂–PO₃ or N,N′-piperazinebismethylenephosphonate. Its structure is reminiscent of the tancoite architecture with corner-sharing chains of TiO₆ octahedra surrounded by HPO₃C groups (Fig. 2d). This delimits ≈4 Å diameter channels, associated with a noticeable BET surface area close to 500 m² g⁻¹. A recent study⁵⁰ revealed that MIL-91 is a very promising candidate for the capture of CO₂. Not only its hydrothermal stability was confirmed but a very high CO₂/N₂ selectivity was measured under post-combustion conditions due to the high adsorption enthalpy of CO₂ (∼40 kJ mol⁻¹) related to the strong confinement effect and the presence of polar P–OH groups. This first example has paved the way for the development of new crystalline porous titanium coordination polymers.

Ti-Based coordination compounds built up from C–O ligands

A considerable number of hybrid (organic–inorganic) compounds based on Ti-cation have been reported so far. As briefly mentioned in the introduction, a large family of Ti-oxo clusters including Ti-oxo-alkoxide, Ti-oxo-alkoxo-carboxylate and Ti-oxo-carboxylate was reported, mainly by the groups of Schubert,⁵¹ Klemperer^52,53 and Sanchez.^22,54 The reader interested in the field could refer to exhaustive recent reviews by Rozes et al.^19,54 and Coppens et al.⁵⁵ Unsurprisingly, Ti-oxo-clusters are, by far, more common compared to coordination polymers.

Indeed, only a few examples of crystalline titanium polymeric coordination compounds, notably MOFs (based on multi-topic organic ligands), have been reported to date, although being desired since almost two decades. Their preparation being very challenging, different methodologies have been followed and, some of them appeared to be promising to overcome the high reactivity of Ti-cations.

The remainder of this review will encompass all Ti-MOFs reported to date classified according to the nature of organic ligands, namely, carboxylate, hydroxycarboxylate, phenolate and catecholate derivatives. As an introductory of each part, prominent examples of discrete Ti-oxo-clusters will be also described, especially those employed for the formation of Ti-MOFs.

Carboxylate ligands

Ti-carboxylate molecular clusters

The richness and the versatility of titanium cation chemistry (highly oxophilic) is clearly reflected by the large number Ti-oxo-clusters of diverse nuclearity reported to date. Indeed, their nuclearity spans from two^56,57 to fifty-two.⁵⁸

Generally, Ti-oxo-carboxylate clusters (general formula Ti_nO_m(OR)_x(OOCR′)_y) are synthesized in solution from a metal cation precursor (very often Ti(OR)₄) in the presence of the carboxylate ligand and/or a controlled amount of water. The main parameters to be controlled are the nature and the amount of the metal cation precursor and/or the ligand (R′COOH), the solvent (nature and amount), as well as the reaction conditions (time, pressure, presence of water or not, inert conditions or not, etc.).^51,54

The titanium oxo core of these Ti_nO_m(OR)_x(OOCR′)_y complexes is formed from water either purposely introduced or generated in situ through esterification. Indeed, the substitution of alkoxide by carboxylate on the Ti ion competes with an esterification reaction that takes place either between the free carboxylic acid in excess and the released alcohol for a carboxylic acid/titanium ratio R′COOH/Ti (R) ≥ 1, or in an intramolecular fashion for R = 1. Moreover, the esterification–polycondensation reactions are kinetically and thermodynamically favoured by increasing the temperature; higher temperature consequently leads to a higher hydrolysis ratio, hence to Ti-oxo-clusters of higher degree of condensation.⁵⁹ The R value plays a key role in the formation of titanium-oxo-carboxylate clusters since it determines the nuclearity of the cluster, its composition and thus its stability in solution. While 1 < R < 3 often results in oxo-carboxylate clusters with condensation rate (O/Ti) between 0.33 and 1, a higher degree of condensation (O/Ti > 1) has also been reported for clusters of higher nuclearity (greater than 12).⁵⁵ Crystal structures of selected examples of oxo-carboxylate clusters are shown in Fig. 4.

Ti⁴⁺ adopts, generally, a six-fold octahedral environment where TiO₆ units are connected through μ₂-oxo and μ₃-oxo bridges. Nonetheless, Ti-oxo-clusters showing under- (tetra- and penta-) and over- (hepta-) coordination of Ti⁴⁺,⁶⁰ or even μ₄-oxo^61–63 have also been reported. Ti_nO_m(Cl)_x(OOCR′)_y are much less common than Ti_nO_m(OR)_x(OOCR′)_y and present only low nuclearities (i.e. 2,^66–71 3^57,64 and 4⁶⁵). In some cases, using TiCl₄ allowed the number of carboxylate ligands (OOCR′/Ti) in the oxo-cluster to be increased. This is the case of the tri-nuclear oxo-cluster [Ti₃(μ₂-O)(μ₃-O)Cl₃(OOCR′)₅] (Fig. 4a; where R′ = ethyl⁵⁷ and p-methyl, p-ethyl, p-fluoro- or p-iodo-phenyl⁶⁴ groups), prepared by reaction of TiCl₄ and carboxylic acid under inert conditions. These clusters present five bridging carboxylate ligands whereas Ti₃-oxo-alkoxo clusters show only two or three.

‘Pure’ Ti-oxo-carboxylate clusters (without alkoxide or chloride groups) are however still very scarce with only two structural types based on mono-topic carboxylate ligands reported to date (Fig. 4g and h). Both are synthesized in the presence of a large excess of carboxylic acids (ca. 10 equivalents) and built up of hexa-coordinated Ti-cations. In the first, Ti₈O₈(OOCR′)₁₆,⁶⁶ Ti-cations are connected through μ₂-oxo bridges and arranged in a wheel-shape (Fig. 4g). The second, Ti₈O₁₀(OOCR′)₁₂,⁶⁷ shows a different packing where μ₂-oxo and μ₃-oxo are bridging Ti-cations (Fig. 4h). As may be expected, due to their quite high condensation degree and the absence of alkoxide or chloride groups, both Ti₈O₈(OOCR′)₁₆ and Ti₈O₁₀(OOCR′)₁₂ present relatively interesting stability towards moisture and nucleophilic species.⁶⁷

These two oxo-clusters also clearly illustrate the lability of the Ti-O iono-covalent bond. For example, Ti₈O₁₀(OOCR′)₁₂ can be prepared from Ti₈O₈(OOCR′)₁₆ under mild conditions, indicating the ability to reconstruct the Ti_nO_m core. More interestingly, different derivatives of Ti₈O₈(OOCR′)₁₆ and Ti₈O₁₀(OOCR′)₁₂ can be prepared by simple ligand exchange without altering the Ti₈-core,^66–68 as already exemplified with other Ti oxo clusters of high nuclearity and condensation degree.¹⁴

Comparable work of carboxylate ligand exchange has been reported recently on the Ti-oxo-alkoxo-carboxylate-phosphonate-cluster.^69,70 Liu et al.⁷¹ have also succeeded in preparing [Ti₆O₆(OⁱPr)₁₀(OOCR′)₂(O₃P-Ph)₂] with 14 different carboxylate functional ligands by exploiting labile coordination sites occupied originally by acetate groups (Fig. 3a). This has allowed tuning rationally the optical bandgap (Fig. 3b) to drive the absorption of light into the visible range with the final aim of optimizing the photocatalytic activity.

Fig. 3 (a) Molecular structure of the [Ti₆O₆(OⁱPr)₁₀(OAc)₂(O₃PR)₂] cluster. Arrows point labile sites. (b) Bandgap summary of the different Ti-oxo-alkoxo-carboxylate-phosphonate-clusters obtained after exchanging acetate ligands by the corresponding electron donor ligands. Color code: Ti, blue; O, red; C, grey; P, orange.⁷¹ Adapted with permission from ref. 71. Copyright 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

Overall, carboxylate based Ti-oxo-clusters show a large panel of nuclearities and geometries. Fig. 4 illustrates the molecular structures of selected carboxylate based Ti-oxo-clusters (including Cl-based one) of various nuclearity (3 to 10) and with the carboxylate/alkoxide and carboxylate/Cl ratio ≥ 1. Among them, Ti₆O₆(OR)₆(OOCR′)₆ (Fig. 4d) has been used in the synthesis of MOFs (see below). It can be described as two staggered Ti-trimers connected through μ₃-oxo bridges and μ₃-OOCR′.

Fig. 4 Crystalline structures of Ti-oxo-carboxylate clusters (including Cl-based ones) of nuclearity ranging between 3 and 10, showing ratio carboxylate/alkoxide (or Cl) ≥ 1. (a) Ti₃O₂Cl₃(OOCR)₅;⁵⁶ (b) Ti₄O₂(OR)₆(OOCR)₆;²¹ (c) Ti₄O₂Cl₆(OOCR)₆;⁶⁵ (d) Ti₆O₆(OR)₆(OOCR)₆;^{56,62,72–78} (e) Ti₆O₄(OR)₈(OOCR)₈;^{21,73,79–84} (f) Ti₆O₆(OR)₂(OOCR)₁₀;⁸⁵ (g) Ti₈O₈(OOCR)₁₆;⁶⁶ (h) Ti₈O₁₀(OOCR)₁₂;⁶⁷ (i) Ti₉O₈(OR)₄(OOCR)₁₆.⁸⁶ Color code: Ti, blue; O, red; C, grey; Cl: green. H-atoms of C-atoms are omitted for the sake of clarity.

Recently, Chun and collaborators^87,88 have shown that supramolecular frameworks built up from the molecular entities Ti₆O₆(OR)₆(OOCR′)₆ (R′ = p-tert-butyl-phenyl or p-amino-phenyl) selectively adsorb CO₂ towards CH₄ and N₂. Similar results have also been obtained for Ti₆O₆(OR)₂(OOCR′)₁₀ (R = p-tert-butyl-phenyl) and Ti₈O₁₀(OOCR′)₁₂ (R′ = p-amino-phenyl).

The high diversity (nuclearity and geometry) and the possible tunable character of Ti-oxo-carboxylate clusters (ligand exchange) make them very promising candidates for pre-designed and controlled synthesis of porous coordination polymers. The following part summarizes the state of the art of existing Ti-carboxylate MOFs. Light will be particularly shed on the difficulties encountered in the elaboration of Ti-MOFs and how they can be overcome.

Ti-carboxylate MOFs

Direct synthesis. Although few amorphous but porous Ti-oxo-carboxylates were described,^89,90 one of the major achievements in Ti carboxylate chemistry came from the preparation in 2009 of the first crystalline and porous titanium carboxylate MOF formulated Ti^IV₈O₈(OH)₄(L)₆ with L = OOC–C₆H₄–COO or terephthalate (BDC) and denoted MIL-125.⁹¹ Noteworthily, only the use of an adequate DMF/MeOH solvent mixture in the presence of titanium isopropoxide [Ti(OⁱPr)₄] and terephthalic acid at 150 °C yielded MIL-125 with very good crystallinity. MIL-125 could also be obtained later with higher yields starting from a precondensed precursor [Ti₈O₈(OOCR)₁₆] (Fig. 4g) instead of the highly reactive titanium isopropoxide.¹⁹ MIL-125 crystallizes in a quasi-cubic tetragonal lattice and its framework is based on cyclic octamers of edge- and corner-sharing TiO₅(OH) octahedra connected to each other through terephthalate linkers defining a distorted fcc network. This leads to both tetrahedral and octahedral cages with respective free diameters of 6 and 12.5 Å (Fig. 5a). MIL-125 shows a significant permanent microporosity (S_BET = 1550 m² g⁻¹; V_p = 0.65 cm³ g⁻¹), high enough for the adsorption of various gases and organic molecules. Interestingly, when alcohols are adsorbed in the cavities under nitrogen, a fast photochromic effect is observed under UV irradiation (Fig. 5b). A color change from white to dark blue takes place without any loss of crystallinity. This behavior is attributed to the partial reduction of Ti⁴⁺ to Ti³⁺ ions,⁹² allowing the photo-oxidation of alcohols to aldehydes. A simple exposure to oxygen allows returning to the initial state indicating the total reversibility of the process. This experiment clearly shows the photoredox properties of the material as well as the promising potential of Ti-based MOFs as photoactive porous materials.

Fig. 5 Crystalline structure of MIL-125:⁹¹ (a) SBU, Ti₈-oxo-cluster, and (b) overall view showing one octahedral cage (yellow sphere) and one tetrahedral cage (light green sphere). (c) Powder of MIL-125 in the presence of benzyl alcohol before and after UV-light irradiation and the corresponding mechanism proposed for the reduction of MIL-125 under irradiation. Color code: Ti, blue; O, red; C, grey.

Indeed, the discovery of MIL-125 was the trigger for various studies not only to obtain new porous Ti-MOFs but also to develop new applications based on their photoactivity.⁹³ Since it was not obvious to synthesize new crystalline Ti-MOFs, efforts have been first focused on tailoring the photocatalytic properties of MIL-125(Ti) by organic functionalization. MIL-125, reported initially with the benzene dicarboxylate ligand, possesses photocatalytic activity only under UV irradiation with an optical bandgap of ca. 3.6 eV (λ = 345 nm).⁹⁴ In order to drive the photoactivity to the visible domain, lowering the highest occupied molecular orbital–lowest unoccupied molecular orbital (HOMO–LUMO) gap of the carboxylated linker was a first approach. Fu et al.⁹⁵ prepared the monoaminated analogue of MIL-125 reported initially by Zlotea et al.,⁹⁶ NH₂-MIL-125, which is a light yellow colored solid with an extra absorption band in the visible-light region, extending the absorption band to ca. 550 nm. As expected, this colored analogue has shown photocatalytic activity under visible-light irradiation for the photoreduction of CO₂ into formate ions in the presence of acetonitrile with triethanolamine (TEOA) as the sacrificial agent.

Similarly, NH₂-MIL-125 showed reversible photochromism during the photocatalytic process, highlighting the reduction of Ti⁴⁺ to Ti³⁺ induced by charge transfer from the organic linker to the inorganic cluster TiO₅(OH) (Ligand-to-Metal Charge Transfer (LMCT)) at the origin of the photocatalytic activity (Fig. 6). As already observed in the literature for NH₂-UiO-66(Zr),⁹⁷ this behavior in the visible range is driven by the NH₂ group of the organic linker and is associated with a reduction of the optical bandgap, experimentally deduced, down to 2.6 eV.

Fig. 6 Proposed mechanism for the photocatalytic reduction of CO₂ with NH₂-MIL-125 under visible light irradiation.⁹⁵ Reprinted with permission from ref. 95. Copyright 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim.

This lowering of the gap induced by –NH₂ groups was further exploited through a combined experimental and computational study by Hendon et al.⁹⁴ The authors confirmed that grafting a –NH₂ group on the phenyl ring resulted in lowering of the optical gap, while the hypothetical diaminated (NH₂)₂-MIL-125 would exhibit an even lower gap, theoretically estimated, of ca. 1.3 eV.

Besides, in order to improve the photocatalytic activity under visible-light irradiation, Nasalevich et al.⁹⁸ have succeeded in post-modifying NH₂-MIL-125 by converting up to 30% of the linkers to dye-like molecular fragments. The obtained MR-MIL-125(Ti) [MR stands for the dye methyl red or N₂-C₆H₄-N(CH₃)₂] shows a significant improvement of its absorption in the visible domain (reaching almost 700 nm) due to the extended conjugation of the aromatic system (Fig. 7).

Fig. 7 Diffuse reflectance spectra of MIL-125 (grey), NH₂-MIL-125 (orange) and MR-MIL-125 (red).⁹⁸ Reprinted with permission from ref. 98. Copyright 2013 The Royal Society of Chemistry.

This solid was later evaluated in the oxidation reaction of benzyl alcohol. An improvement in the photocatalytic performance of 40% in the visible range was observed in comparison to the pristine NH₂-MIL-125. However, this is still lower than the 100% improvement observed in photon absorption for MR-MIL-125. This difference in performance is attributed to other factors affecting the photocatalytic activity such as the molecular adsorption/diffusion properties of the MOF or the absorption of photons of energy too low to drive the photocatalytic transformation.

Horiuchi and collaborators followed another approach to improve photocatalytic performance by encapsulating active species, i.e. platinum nanoparticles, on the surface of NH₂-MIL-125 to obtain Pt/NH₂-MIL-125 (Fig. 8). The composite material showed significant photocatalytic activity in the visible (ca. 500 nm) towards H₂ production.⁹⁹

Fig. 8 Diagram of the photocatalytic H₂ production in the presence of Pt/NH₂-MIL-125.⁹⁹ Reprinted with permission from ref. 99. Copyright 2012 American Chemical Society.

Owing to the very promising photocatalytic properties of MIL-125 and its derivatives, discovering new titanium MOFs became even more attractive. However, new porous Ti-MOFs started to appear only a few years after the discovery of MIL-125s. This is certainly due to the rather high reactivity and fast polycondensation of titanium cations in solution, controlling this step being a key prerequisite for obtaining highly crystalline Ti-MOFs. This poor control of the polymerization of titanium is illustrated by the very different synthetic procedures followed for the few Ti-MOF reported to date. Two main strategies can however be highlighted:

(1) The control of the kinetics of the coordination reaction and/or polycondensation of Ti, through the use of either very complexing solvents¹²⁹ or less reactive titanium precursors such as organometallic complexes¹⁰⁰ and preformed polycondensed titanium oxo-clusters.^19,101,102

(2) The post-synthetic metathesis (PSM) of metal ions and post-synthetic metathesis and oxidation (PSMO).^103–108

Very recently, Bueken et al.¹⁰⁰ reported a new Ti-MOF based on a trinuclear titanium-oxocarboxylate [Ti^IV₃(μ₃-O)(O)₂(COO)₆] cluster. This MOF, denoted COK-69 and formulated Ti^IV₃O₃(L)₃·DMF with L = OOC–C₆H₁₀–COO or 1,4-cyclohexanedicarboxylate (CBC), was prepared by reacting dicyclopentadienyl titanium(IV) dichloride (Cp₂Ti^IVCl₂) and 1,4-cyclohexanedicarboxylic acid (H₂-CBC) in DMF and acetic acid under an inert atmosphere and solvothermal conditions. The reduced form was first obtained [COK-69(Ti^III)] and later oxidized under air exposure [COK-69(Ti^IV)]. The crystalline structure of COK-69 showed that it resembles closely the iron or chromium dicarboxylates MIL-88 and presents also a flexible 3-D framework, although such flexibility arises from a different mechanism (Fig. 9). Note that because of the close pore phase obtained after activation, COK-69 shows a very low porosity towards N₂ (S_BET = 29 m² g⁻¹). The authors postulate that the synthesis of this MOF is induced first by the preformation of the cyclopentadienyl-capped Ti₃(μ₃-O) oxo-cluster preventing the spontaneous hydrolysis of titanium and favouring the direct formation of the [Ti^IV₃(μ₃-O)(μ₁-O)₂(COO)₆] units.

Fig. 9 Crystalline structure of COK-69s:¹⁰⁰ (a) SBU, Ti₃-oxo-cluster (corresponding to closed-pore phase); (b, c) view along the channels of COK-69: (b) closed-pore and (c) open-pore. Color code: Ti, blue; O, red; C, grey.

COK-69 (theoretically estimated optical gap 3.77 eV) showed, under UV irradiation and without any loss of crystallinity, a reversible photoreduction and reoxidation activity in the presence of ethanol and upon air exposure, respectively. It was demonstrated that this process involves simultaneous transfer of both electrons and protons (proton-coupled electron transfer (PCET)).

Using a more reactive titanium precursors, Long et al.¹⁰⁹ have succeeded in preparing the Ti(III) analogue of MIL-101, a MOF based on the trinuclear oxo-carboxylate M(III) oxo-cluster. As for COK-69, Ti(III)-MIL-101 or Ti^III₃O(OCH₂CH₃)(L)₃·2DMF with L = OOC–C₆H₄–COO or terephthalate (BDC) was synthesized under an inert atmosphere in DMF under solvothermal conditions, but by reacting that time H₂-BDC with TiCl₃. Illustration of the proposed composition of the Ti₃O clusters is shown in Fig. 10. Each titanium is coordinated to five oxygen atoms in addition to one labile ligand (solvent molecule). The unsaturated five-coordinated Ti^III (obtained after activation) has shown its ability to react with O₂ molecules generating Ti^IV superoxo and peroxo species. Note that this allows irreversible steep adsorption of O₂.

Fig. 10 Illustration of the proposed composition of the Ti₃O clusters in MIL-101 after (a) heating at 150 °C in vacuum, (b) reacting with O₂ to form titanium(IV) superoxide and (c) reacting with O₂ to form titanium(IV) peroxide. Note that the cluster likely distorts.¹⁰⁹ Reprinted with permission from ref. 109. Copyright 2015 American Chemical Society.

If the direct synthesis (one pot reaction) using titanium salts or organometallic complexes (generally, isolated Ti-cation) as precursors under inert conditions has led to two additional new Ti-carboxylated MOFs initially with Ti ions under a trivalent state, attempts using preformed titanium(IV) oxo-clusters have also been a fruitful alternative. As for metallocene complexes, using Ti-oxo-clusters as a source of titanium should allow a slower reaction, as both ligands (the monotopic one initially on the oxo-cluster and the polytopic one used to build the extended network) show similar coordination ability towards titanium. Two possibilities have been considered. In the best case, a slow exchange between the monotopic and polytopic ligands might take place without altering the nuclearity and the geometry of the preformed oxo-cluster.^{19,102,110–113} Contrarily, the oxo-cluster could act as a reservoir of matter and undergo complete or partial reorganization but still permitting to slow down the reaction between cations and ligands and lead to the crystallization of a coordination polymer.¹⁰¹

This last strategy, initially proposed for MIL-125 (see above), was used by Yuan et al.¹⁰¹ to prepare a new porphyrinic carboxylated Ti-MOF. PCN-22 (PCN stands for porous coordination network), formulated Ti^IV₇O₆(L)₁₂·2DEF with L = tetrakis(4-carboxyphenyl)porphyrin, was obtained by mixing, under solvothermal conditions, a preformed titanium oxo-carboxylate cluster Ti₆O₆(OⁱPr)₆(abz)₆ (with abz = H₂N-C₆H₄-COO or 4-aminobenzoate) (see Fig. 4d) with a porphyrin-based linker H₄-TCPP (tetrakis(4-carboxyphenyl)porphyrin) in the presence of benzoic acid. The structure of PCN-22 revealed that the Ti₆ oxo-cluster was completely transformed during the reaction and the obtained 3-D framework is based on unprecedented Ti₇O₆ oxo-clusters connected to each other by TCPP linkers (Fig. 11). Ti₇O₆ could be described as a pair of Ti₃O₃ units bridged by a Ti-cation. It is a 12-connected cluster composed of seven Ti⁴⁺ ions, two μ₃-O²⁻ ions, two terminal O²⁻ ions, two terminal OH⁻ ions and two bridging solvent molecules (DEF).

Fig. 11 (a) Crystalline structure of PCN-22:¹⁰¹ (a) SBU, Ti₇-oxo-cluster (only O-atoms of DEF molecules coordinated to the central Ti⁴⁺ are represented); (b) view of the framework along channels. Color code: Ti, blue; O, red; C, grey; N, blue; H, white. H-atoms of C-atoms are omitted for the sake of clarity.

PCN-22 presents a high permanent porosity towards N₂ (S_BET = 1284 m² g⁻¹). Besides, thanks to the TCPP moieties, this Ti-MOF shows a broad range of light absorption from 200 nm to 640 nm with an experimentally deduced optical gap of 1.93 eV, the smallest reported to date for pure carboxylate based Ti-MOFs. In addition, the photocurrent measurements indicated that PCN-22 is photoactive under visible-light illumination (>450 nm). Finally, studies revealed that PCN-22 acts as a heterogeneous photocatalyst capable of promoting, under visible-light, alcohol oxidation reaction converting benzyl alcohol to the corresponding aldehyde (conversion rate of 28% after two hours and TON over 100). The proposed mechanism is depicted in Fig. 12.

Fig. 12 Proposed mechanism for the photocatalytic oxidation reaction of benzyl alcohol for the PCN-22/TEMPO system (TEMPO: 2,2,6,6-tetramethylpiperidinyloxyl).¹⁰¹ Reprinted with permission from ref. 101. Copyright 2015 The Royal Society of Chemistry.

The same hexameric Ti-oxo-cluster, Ti₆O₆(OⁱPr)₆(abz)₆, was also used very recently to elaborate an original Ti-MOF. Reported by Nguyen et al., MOF-901, formulated Ti^IV₆O₆(OCH₃)₆(L)₃ with L = OOC-N₂C₂₀H₁₄-COO or 1,4-phenylenebis(methanylylidene)bis(azanylylidene)dibenzoate, is the first MOF obtained by combining the preparation strategies of metal- and covalent-organic frameworks.¹⁰² Its synthesis consists of connecting preformed Ti₆ oxo-carboxylate clusters through imine functions resulting from condensation reactions between terminal –NH₂ groups decorating the aforementioned clusters and carbonyl groups of benzene-1,4-dialdehyde spacer. The overall structure of MOF-901 can be described as perforated 2-D layers, showing triangular apertures, stacked in an AB fashion (Fig. 13). MOF-901 exhibits a permanent porosity towards N₂ with an experimentally deduced S_BET of 550 m² g⁻¹. It is noteworthy that the best crystallinity was nevertheless achieved by mixing all reactants in one-pot (see Fig. 13a) while using the preformed hexameric cluster led to a less crystalline solid.

Fig. 13 MOF-901¹⁰² and MOF-902.¹¹⁴ Scheme summarizing the synthetic procedure. (b) Space filling views of crystalline structures of isoreticular MOFs. The discrete hexameric cluster is shown in the inset of (b). Color code: Ti, blue; C, black; O, red; N, green; H, pink; and second layer, orange. Capping methoxide moieties have been removed for clarity. Reprinted with permission from ref. 114. Copyright 2016 American Chemical Society.

The yellow colored MOF-901 possesses a large absorption spectrum ranging from 340 to 550 nm with a maximum located at 360 nm (experimentally deduced optical gap = 2.65 eV), slightly higher than that measured for PCN-22. Moreover, MOF-901 was successfully used, without loss of crystallinity, as a heterogeneous photocatalyst under visible-light irradiation for the radical polymerization of methyl methacrylate (MMA) with ethyl α-bromophenylacetate as co-initiator. It was found to outperform P-25 TiO₂ and Ti₆O₆(OⁱPr)₆(abz)₆, VNU-1 (VNU stands for Vietnam National University), NH₂-MIL-125 and NH₂-UiO-66.

More recently, MOF-902, isoreticular to MOF-901 with a longer organic linker (bearing two benzene rings instead of one), was reported.¹¹⁴ It was synthesized using the same strategy as that employed for MOF-901 with a slightly modified procedure. It has a similar layered structure with larger triangular apertures of ca. 16 Å but a lower experimental surface area (S_BET of 400 m² g⁻¹). No explanation was provided concerning the latter point.

MOF-902 showed a red-shifted optical absorption band, ranging from 340 nm to 640 nm with a maximum located at 390 nm, probably induced by the higher conjugation system of the organic backbone. Its experimentally deduced band-gap energy was found to be ca. 2.50 eV.

The photoactivity under visible light of MOF-902 was first demonstrated using the same catalytic reaction of MMA polymerization tested for MOF-901. It resulted in the production of polyMMA with a higher molecular weight (31 465 g mol⁻¹vs. 26 850 g mol⁻¹) and a lower polydispersity index (1.11 vs. 1.6) compared to MOF-901. Additionally, MOF-902 showed a significantly better performance than MOF-901 in the photocatalytic polymerization of benzyl methacrylate and styrene monomers without significant loss of performance over consecutive cycles.

Indirect synthesis. Before the burgeoning of new Ti carboxylate based MOFs, post-synthetic metathesis (PSM) was also considered to introduce heteroelements into established porous solids. This approach was carried out first with inorganic nanocrystals,¹⁰⁷ and later on with other bulk materials.^115–117 It was then reported with MOFs in 2009¹¹⁸ allowing to complete exchange between the ion M¹(II) from the framework and either M²(II) or M³(III). This approach appeared to be a very good alternative to avoid the spontaneous hydrolysis of titanium precursors and the high reactivity of the titanium ion. Additionally, it allows targeting a specified topology and cavity size and/or shape for a pre-selected application. Ultimately, the replacement of low valency cations with high valency ones [i.e. M(II) → M(III) → M(IV)] should also lead to more chemically stable MOF analogues.^4,119 Note that the case of the post-synthetic insertion of Ti ions within the core of organic ligands will not be discussed here.

Following the PSM strategy, Cohen and collaborators first succeeded in partially transforming existing MOFs into their titanium analogues, among others. They prepared Zr/Ti-UiO-66¹⁰³ (Zr/Ti < 0.1) and two aminated derivatives: NH₂-Zr/Ti-UiO-66¹⁰⁴ (Zr/Ti = 3.03) and (NH₂)₂-NH₂-Zr/Ti-UiO-66 (Zr/Ti = 2.52). These Ti-analogues were obtained by reacting crystalline Zr-MOFs with a solution of TiCl₄(THF)₂ (THF = tetrahydrofuran) at 85 °C for 5 days (Fig. 14). Various techniques (PXRD, ICP, N₂ sorption measurements) were combined to prove the insertion of Ti, although the exact nature of the coordination sphere of the Ti ions within this solid was not discussed (e.g. substitution of Zr by Ti would lead to unexpected 8-fold coordinated Ti, except if such transformation is accompanied by the formation of defects of the Zr/Ti oxo-cluster).

Fig. 14 Synthesis of aminated Zr- and Zr/Ti-MOF 1(Zr): NH₂-Zr/Ti-UiO-66 (Zr_4.5Ti_1.5O₄(OH)₄(C₈H₇O₄N)₆), and (NH₂)₂-NH₂-Zr/Ti-UiO-66 (Zr_4.3Ti_1.7O₄(OH)₄ (C₈H₇O₄N)_5.17(C₈H₈O₄N₂)_0.83).¹⁰⁴ Reprinted with permission from ref. 104. Copyright 2015 The Royal Society of Chemistry.

If no photocatalytic study was reported for Ti-UiO-66, both aminated Ti-analogues showed photocatalytic CO₂ reduction under visible-light irradiation in the presence of triethanolamine (TEOA) and 1-benzyl-1,4-dihydronicotinamide (BNAH) as sacrificial donor reagents. (NH₂)₂-NH₂-Zr/Ti-UiO-66 showed a higher activity than NH₂-Zr/Ti-UiO-66 (turnover numbers after 6 hours of 6.27 and 4.66, respectively). Such enhancement was attributed to the existence of two absorption paths suitable for the reduction of CO₂ in (NH₂)₂-NH₂-Zr/Ti-UiO-66, while only one in NH₂-Zr/Ti-UiO-66.

A few other studies on post-synthetic exchange (PSE) with the Ti-ion have also been reported.^{101,106,120,121} Among them, Zhou et al.¹⁰⁵ performed a systematic study where two metal(III) carboxylate MOFs (based on M^III₃O₃ trimers: MIL-100(Sc), PCN-333(Sc)) and two metal(II) hydroxycarboxylate solids (based on M^IIO chains: MOF-74(Zn) and MOF-74(Mg)) were converted partially into their corresponding Ti-analogues, and finally tested as photocatalysts. The synthetic strategy, termed high valence metathesis and oxidation (HVMO), consists of post-exchanging cations with Ti³⁺ to obtain an intermediate Ti(III)-MOF analogue followed by mild oxidation under air to yield Ti(IV)-MOF (Fig. 15).

Fig. 15 Schematic illustration of the stepwise HVMO procedure for the design of Ti-MOFs from the template MOFs: (a) MIL-100(Sc) and PCN-333(Sc) metal metathesis with Ti³⁺ followed by metal node oxidation in the air (PCN-333(Sc) and MIL-100(Sc) have the same topology but different structures) and (b) similar process for MOF-74(Zn) and MOF-74(Mg).¹⁰⁵ Reprinted with permission from ref. 105. Copyright 2015 The Royal Society of Chemistry.

The crystalline structure integrity of the obtained Ti-MOFs was confirmed by PXRD and N₂ sorption measurements. The metal exchange ratio indicates variable exchange rate, from 37.9% to 94.7%. The resulting solids exhibit a slightly lower N₂ sorption capacity than the pristine materials, attributed to the presence of extra anions balancing the charge. Diffuse reflectance UV-Vis absorption spectroscopy showed absorption edge at 380 nm for PCN-333(Sc)-Ti and MIL-100(Sc)-Ti and extended absorption up to 660 nm for MOF-74(Ti)s in agreement with the LMCT occurring between 2,5-dihydroxyterephthalic acid (H₂-DOBDC) and Ti centers (see the next section). Therefore, the photocatalytic potential of the Ti-MOFs was examined based on the photo-degradation, under visible-light irradiation, of methylene blue (MB) as a probe reaction. While PCN-333(Sc)-Ti and MIL-100(Sc)-Ti showed only moderate photo-degradation activity after 9 minutes (35% and 64%, respectively), both MOF(74)-Ti revealed excellent conversions up to 98% only after 3 minutes of irradiation (Fig. 16a). The excellent photocatalytic performance of Ti-MOF-74 was attributed to its capability of absorbing more broadly the visible light and a longer excitation lifetime, most probably due to the chain-like connectivity of the Ti-ion. The proposed mechanism is depicted in Fig. 16b.

Fig. 16 (a) Photodegradation of MB using no catalyst (blank), TiO₂, PCN-333(Sc)-Ti, MIL-100(Sc)-Ti, MOF-74(Zn)-Ti and MOF-74(Mg)-Ti with 300 W xenon light irradiation and (b) proposed mechanism of MB degradation of Ti-MOF-74 in the presence of air.¹⁰⁵ Reprinted with permission from ref. 105. Copyright 2015 The Royal Society of Chemistry.

Phenolate based ligands

Although molecular Ti oxo-alkoxo are well studied (see above), crystalline extended networks are far more limited. To the best of our knowledge, a single 3-D Ti alkoxide (i.e. built up from alkyl-derived alcoholate) built up from 1,4-butandiol was reported, albeit soluble (and hence unstable) in most common organic solvents.¹²² As detailed below, aryloxide (phenolate) ligands have been more studied.

As demonstrated by the very good photoactivity of Ti-MOF-74 under visible-light illumination, such ligands succeeded in driving the photoactivity of Ti complexes to the visible-light range. This can be accounted for by the Ti^IV–O_phenolate bonds, known to lead to ligand-to-metal charge transfer (LMCT; between the lone pairs of oxygen atoms and the vacant 3d-orbitals of Ti atoms), giving rise to strong absorption bands in the visible range.^123–127 In addition to providing red-shift in the light absorption spectrum, the high basicity of phenolates (much higher than that of carboxylates) combined with the high oxophilic character of Ti-cations might lead to a strengthening of the interaction between Ti-centers and linkers yielding a better chemical stability. Studies have indeed already proven that, generally, combining a metal cation of high charge density (Zr, Ti) and a ligand showing high pKa values is an efficient strategy towards obtaining highly chemically stable MOFs.^7,128,129

It is worth reminding that the combination between highly reactive species (highly basic phenolates and highly oxophilic Ti-cations) might turn the crystal growth of extended frameworks, i.e. MOFs, into a more complicated task, particularly when considering one-pot synthesis. However, a few crystalline coordination polymers and MOFs based on either carboxyphenolate or pure phenolate derivatives have been reported so far, and will be described below.

Hydroxycarboxylate derivatives

Crystalline compounds based on titanium and hydroxycarboxylate ligands remain very scarce. To our knowledge, only one type of complex prepared with purely α-hydroxybenzoic acid (salicylic acid, H₂-Sal or (OH)-C₆H₄-COOH) has been reported so far. In any case (neutral [Ti(Sal)(H-Sal)₂](BuⁿOH)₂¹²³ or salts^130–132), tris(salicylato)titanate(IV) is constructed from the isolated Ti-center with an octahedral geometry. The Ti-cation is hexacoordinated by three chelating Sal ligands through one O_phenolate and one O_carboxylate atoms (Fig. 17a).

Fig. 17 (a) Crystalline structure of tris(salicylato)titanate(IV);¹²³ (b) crystalline structure of Ti₁₃O₁₀(o-BDC)₄(Sal)₄(OⁱPr)₁₆.¹³³ Color code: Ti, blue; O, red; C, grey. H-atoms are omitted for the sake of clarity.

Very recently Hou et al.¹³³ reported two isostructural Ti-oxo-clusters with a higher nuclearity, Ti₁₃O₁₀(o-O₂C-C₆H₄-CO₂)₄(Sal)₄(OⁱPr)₁₆ and Ti₁₃O₁₀(o-O₂C-C₆H₄-CO₂)₄(Sal-Cl)₄(OⁱPr)₁₆, prepared with a mixture of two ligands (Sal and o-BDC) under solvothermal conditions. Their crystal structures revealed that all the Ti⁴⁺ ions are octahedrally coordinated and assembled together through μ₂-O and μ₃-O bridges, oxygen atoms of chelating o-BDC, and Sal ligands with OⁱPr acting as terminal ligands (Fig. 17b). Despite the fact that these Ti₁₃-clusters are not bearing only Sal ligands, they showed under a full-wavelength light source, similar photocurrent densities of ca. 0.30 μA cm⁻².

Due to their potential benefits for photocatalysis in the visible-light range, the use of carboxyphenolate ligands has expanded to extended coordination frameworks. The first cases are heterometallic Ti/Zn MOFs, labeled ZTOF-1¹³⁴ and ZTOF-2,¹³⁵ obtained using asymmetrical linkers, 2-hydroxyterephthalic acid and 3-hydroxy-2,7-naphthalenedicarboxylic acid, respectively. Though their crystalline structures totally differ, the connectivity of their Zn₆Ti₂ oxo-clusters to the linker remains similar where Ti-cations are coordinated by one O_phenolate and one O_carboxylate atoms (Fig. 18). Both MOFs showed air-stable frameworks and permanent porosity towards different gases (N₂, Ar, CO₂ and H₂) where calculated surface areas are S_BET ZTOF-1 = 1045 m² g⁻¹ and S_BET ZTOF-2 = 1878 m² g⁻¹.

Fig. 18 Crystalline structures of ZTOF-1¹³⁴ (left) and ZTOF-2¹³⁵ (right): (a and c) coordination mode between Zn- and Ti-cations and organic linkers (only one complete linker is represented); (b and d) general views of ZTOF-1 and ZTOF-2, respectively. Color code: Ti, blue; Zn, purple; O, red; C, grey. H-atoms are omitted for the sake of clarity.

NTU-9¹²⁹ (NTU stands for Nanyang Technological University) is one of the scarce MOFs constructed from purely Ti⁴⁺ cations and hydroxycarboxylate derivatives (2,5-dihydroxyterephthalic acid). It is obtained as deep-red single crystals after reacting, under solvothermal conditions, Ti(OⁱPr)₄ with H₄-DOBDC or HOOC-C₆H₂(OH)₂-COOH in acetic acid. The structure of NTU-9 or Ti^IV₂(H-L)₂(H₂-L)_n with L = OOC-C₆H₂(O)₂-COO or 2,5-dihydroxyterephthalate (DOBDC) is built up from Ti-cations octahedra, the six oxygen atoms being issued from phenolate and carboxylate groups, connected together via the organic ligands to form two dimensional perforated honeycomb-like layers. The coordination geometry of titanium ions is very similar to that of tris(salicylato)titanate described above. The stacking of these layers generates 1-D channels showing 11 Å × 11 Å hexagonal apertures (Fig. 19). Despite its stability under air, in water and common organic solvents and after guest removal at 120 °C, the permanent porosity of NTU-9 has not yet been assessed. Nevertheless, owing to the LMCT (confirmed by the density of states (DOS) calculation) at the origin of its deep-color, NTU-9 shows a broad range of light absorption in the visible region with a maximum at ca. 520 nm associated with an experimentally deduced optical gap of 1.72 eV which correlated with the theoretical calculations. Photoelectrochemical studies revealed that the aforesaid MOF shows a rather very weak cathodic photocurrent of ca. 60 nA cm⁻² (measured on the NTU-9/FTO electrode). Finally, NTU-9 showed, under visible-light, a good photocatalytic activity towards the degradation of organic dyes (rhodamine B and methylene blue; degradation completed after 80 min and 20 min, respectively).

Fig. 19 Crystalline structure of NTU-9:¹²⁹ (a) coordination mode between Ti-cation and organic linker; (b) view along the channels of NTU-9. Color code: Ti, blue; O, red; C, grey. H-atoms are omitted for the sake of clarity.

Very recently, the reactivity of 2,5-dihydroxyterephthalic acid (DOBDC or L in the following for MIL-167, MIL-168 and MIL-169) with titanium has been thoroughly investigated by some of us.¹²⁷ Under solvo/hydro-thermal conditions and upon systematic variation of different parameters (solvent, titanium precursors, stoichiometry, modulator, etc.), four crystalline phases have been identified (Fig. 20).

Fig. 20 Crystalline structures of MIL-167, MIL-168 and MIL-169:¹²⁷ (a, c and e) coordination mode between Ti-cation and organic linkers, respectively. Global views showing (b) two interpenetrated 3-D nets in MIL-167, (d) packed 1-D chains in MIL-168 and (f) layers connected through piperazinium cations via H-bonds (dashed blue lines) in MIL-169. Color code: Ti, blue; O, red; C, grey; N, dark blue; H, white. H-atoms of C-atoms are omitted for the sake of clarity.

First, when using Ti(OⁱPr)₄ as a precursor and pure N,N-diethylformamide (DEF) as a solvent, two phases were isolated: a first one very similar to NTU-9¹²⁹ and a second one labelled MIL-168, formulated as Ti^IV(H_xL)_1.5(DEAH)_2−1.5x·nsolv and Ti^IV(L)(cat)·2DEAH, respectively.

The addition of methanol to the reaction medium ultimately leads to the formation of MIL-167, formulated as Ti^IV(L)_1.5(Et₂MeNH)₂·nH₂O. This structure consists of isolated TiO₆ octahedra, with the oxygen atoms arising solely from DOBDC ligands, connected to each other by the ditopic linkers through chelating six-membered rings involving both phenolic and carboxylic oxygen atoms to afford a 3-D chiral network exhibiting 10,3-a topology. Two enantiomeric networks are interpenetrated to afford the final solid, which holds cavities of ca. 6–7 Å free diameter. Interestingly, MIL-167 and NTU-9 are built up from the same coordination motif and hence, without taking into account the counter-ions, are simply polymorphs. The difference between both solids arises from a different alternation of the absolute configuration of the neighboring octahedral metal centers.

Adding catechol (H₂-cat) in the reaction mixture allowed the crystallization of MIL-168. Its structure is built up from isolated TiO₆ octahedra, with the coordinating oxygen atoms arising this time from two DOBDC ligands and one catechol moiety. This defines 1-D zigzag chains running parallel to each other and separated by N–H⋯O hydrogen-bonded diethylammonium (DEAH) cations, arising from the degradation of DEF.

In order to favor the formation of more condensed inorganic units, a tetrameric preformed titanium oxalate cluster, Ti₄O₄(OOC-COO)₇(N₂C₄H₁₀)₃,¹³⁶ was used as the precursor. Under hydrothermal conditions, MIL-169 was obtained. In this structure, each titanium ion is chelated by two DOBDC ligands through the six-membered ring involving phenolate and carboxylate groups already observed in MIL-167, MIL-168 and NTU-9. It is worth noting that the titanium cations are associated in dimers (Ti₂O₁₁) of octahedra bridged by a μ₂-oxo group and bearing terminal water molecules. Dimers are connected through four DOBDC ligands to afford anionic layers separated by hydrogen bonded piperazinium cations and free water molecules.

For all these solids, no noticeable adsorption of N₂ was however detected. However, MIL-167 shows a modest microporous character after adsorption of CO₂ under near-ambient conditions (15 °C) with a maximal adsorption capacity of ca. 2.1 mmol·g⁻¹ at 20 bar. This is due to both the interpenetration and the presence of entrapped bulky Et₂MeNH cations inside the pores, arising again from the degradation of DEF into DEA and then the reaction of DEA with methanol.

All solids present strong absorption in the visible range, much broader than that of the titanium carboxylate systems,^94,95,98,137 associated again with LMCT.^123–127 Their optical spectra have shown a band centered at 250 nm, together with a broad band covering the 300–600 nm window. Noteworthily, under UV irradiation (λ = 280 nm), a noticeable production of hydrogen was detected for MIL-167 with a reaction rate initially ca. 2.6 higher than for MIL-125-NH₂.^99,138 Nevertheless, no activity was detected under visible light irradiation.

Pure phenolate derivatives

Several crystalline coordination complexes based on titanium and monotopic phenolate derivatives¹³⁹ have been reported. They range from mono-,^140–142 di-, tri-nuclear¹⁴⁰ oxo-clusters up to polymeric arrangement (showing μ₂-Cl and μ₂-O bridges)¹⁴³ (Fig. 21). The most common is the di-nuclear one of which different analogues have been reported: hexa-coordinated to only phenolates (Fig. 21a)¹⁴⁴ or phenolate and terminal ligands,^140,145,146 and penta-coordinated¹⁴⁷ (phenolates and chlorine; Fig. 21b).

Fig. 21 Crystalline structures of di-nuclear (a) 6-coordinated¹⁴⁴ and (b) 5-coordinated¹⁴⁷; (c) tri-nuclear 5-coordinated¹⁴⁰ phenolate oxo-clusters; and (d) polymeric¹⁴³ Ti-phenolate chain. Color code: Ti, blue; O, red; C, grey; N, dark blue; Cl, green; F, yellow. H-atoms are not represented for the sake of clarity.

Ditopic phenolate ligands were first used to build up amorphous Ti based hybrid materials as early as in 1990.^148,149 A few years later, Wolczanski and collaborators reported a series of crystalline Ti-oxo-phenolate coordination polymers. Prepared under inert conditions and quoted ‘covalent networks’, they present two- or three-dimensional networks based on isolated Ti-centers or dinuclear oxo-clusters (with or without terminal ligands) connected through various phenolate derivatives (see details in Table 1).^150–154 The nature of the organic solvent seemed to play a role in the coordination mode and, thus, in the resulting coordination network. Among these coordination polymers, the one obtained by reacting Ti(OⁱPr)₄ and 4,4′-dihydroxybiphenyl in THF under an inert atmosphere showed a 3-D open framework based on Ti-dimers (Fig. 22).¹⁵³ If a large pore opening of the 3-D networks is of interest, the potential porosity is once again strongly reduced due to the presence of an interpenetrated network.

Table 1 Summary of the reported titanium coordination polymers (obtained through direct synthesis) discussed in this review

Name	Molecular formula	Organic linker (L)	Inorganic unit nuclearity	Network dimensionality	Porosity	Photoredox property	Ref.
MIL-22	Ti^IV3O2(H2O)2(L)2·2H2O	Methylenediphosphonate	3	3-D	—	—	48
MIL-252	Ti^IV(L)	Ethylenediphosphonate	1	3-D	—	—	46
MIL-253	Ti^IV(L)	Propylenediphosphonate	1	3-D	—	—	46
MIL-91	Ti^IVO(H2-L)·nH2O (n ∼ 4.5)	N,N′-piperazinebismethylenephosphonate	Infinite chain	3-D	S BET = 500 m^2 g^−1 (N2 77 K)	—	49
MIL-125	Ti^IV8O8(OH)4(L)6	Terephthalate	8	3-D	S BET = 1550 m^2 g^−1 (N2 77 K)	Alcohol oxidation (UV light)	91
COK-69op&c	Ti^IV3O3(L)3·DMF	1,4-Cyclohexanedicarboxylate	3	3-D	S BET = 29 m^2 g^−1 (N2 77 K)	Alcohol oxidation (UV light), Eg = 3.77 eV	100
MIL-101(Ti)	Ti^III3O(OCH2CH3)(L)3·2DMF	1,4-Benzenedicarboxylate	3	3-D	S BET = 2970 m^2 g^−1 (N2 77 K)	—	109
PCN-22	Ti^IV7O6(L)12·2DEF	Tetrakis(4-carboxyphenyl)porphyrin	7	3-D	S BET = 1284 m^2 g^−1 (N2 77 K)	Alcohol oxidation (visible light), Eg = 1.93 eV	101
MOF-901	Ti^IV6O6(OCH3)6(L)3	1,4-Phenylenebis(methanylylidene) bis(azanylylidene)dibenzoate	6	2-D	S BET = 550 m^2 g^−1 (N2 77 K)	Mediated radical polymerization of methyl methacrylate (visible light), Eg = 2.65 eV	102
MOF-902	Ti^IV6O6(OCH3)6(L)3	Biphenyl-4,4-diylbis-(methanylylidene) bis(azanylylidene)dibenzoate	6	2-D	S BET = 400 m^2 g^−1 (N2 77 K)	Mediated radical polymerization of methyl methacrylate (visible light), Eg = 2.50 eV	114
NTU-9	Ti^IV2(H-L)2(H2-L)n	2,5-Dihydroxyterephthalate	1	2-D	—	Degradation of organic dyes (visible light), Eg = 1.72 eV	129
MIL-167	Ti^IV(L)1.5(Et2MeNH)2·nH2O	2,5-Dihydroxyterephthalate	1	3-D	Maximal adsorption capacity ∼ 2.1 mmol g^−1 (CO2 20 bar)	Water splitting (UV light)	127
MIL-168	Ti^IV(L)(cat)·2DEAH	2,5-Dihydroxyterephthalate	1	1-D	—	—	127
MIL-169	Ti^IVO0.5(L)(H2O)(H2-pip)0.5·nH2O	2,5-Dihydroxyterephthalate	2	2-D	—	—	127
Ti-phenol	[Ti^IV(L)(H-L)]2(μ-H-L)2	Hydroquinone	2	3-D	—	—	150
Ti-phenol	Ti^IV(L)2·py2	Hydroquinone	1	2-D	—	—	150
Ti-phenol	Ti^IV2(μ1,4-L)2(μ1,4-H-L)2(μ-H-L)2	Hydroquinone	2	3-D	—	—	151
Ti-phenol	Ti^IV2(μ1,4-L)2(μ1,4:η^2,η^1-L)2(OH2)2(H2O)2(H2-L)·x(CH3CN) (x = 1 or 2)	Hydroquinone	2	3-D	—	—	151
Ti-phenol	Ti^IV2(μ1,7-L)2(μ1,7:η^2,η^1-H-L)2(O^iPr)2	2,7-Dihydroxy-naphthalene	2	3-D	—	—	151
Ti-phenol	Ti^IV(μ1,4-L)2py2	Hydroquinone	1	2-D	—	—	152
Ti-phenol	Ti^IV(μ1,4-L)2·py2·py	Hydroquinone	1	2-D	—	—	152
Ti-phenol	Ti^IV(μ1,4-L)2·(4-Ph-py)2	Hydroquinone	1	2-D	—	—	152
Ti-phenol	Ti^IV(μ1,3-L)2·py2	Resorcinol	1	2-D	—	—	152
Ti-phenol	Ti^IV(μ1,3-L)2·(4-Ph-py)2	Resorcinol	1	2-D	—	—	152
Ti-phenol	Ti^IV(μ1,6:η^2,η^1-4,4′-L)0.5(μ1,6:η^2,η^1-4,4′-L)(O^iPr)(HO^iPr)2·THF	4,4′-Dihydroxy-biphenyl	2	3-D	—	—	153
Ti-phenol	Ti^IV(μ1,3-1,4-L)(μ1,3-1,3-L)(1,3-H-L)(HO^iPr)2	Resorcinol	2	2-D	—	—	153
Ti-phenol	Ti^IV(μ2,7-L)2·py2	2,7-Dihydroxy-naphthalene	1	1-D	—	—	154
Ti-phenol	Ti^IV(μ2,7-L)2(4-picoline)2.5	2,7-Dihydroxy-naphthalene	1	2-D	—	—	154
Ti-CAT-5	Ti^IV(L)·2DMA	2,3,6,7,9,11-Hexahydroxytriphenylene	1	3-D	S BET = 450 m^2 g^−1 (N2 77 K)	—	176

---

Fig. 22 Crystal structure of Ti-oxo-biphenolate:¹⁵³ (a) view of the dinuclear oxo-cluster and (b) two interpenetrated nets. Color code: Ti, blue; O, red; C, grey; H, white. H-atoms of C-atoms and pore filling molecules are omitted for the sake of clarity.

Catecholate derivatives

Among the different possible phenolate derivatives, catechol (H₂-cat or C₆H₄(OH)₂) is an interesting candidate due to its interest in biology, electrochemistry and catalysis. In addition, being rather cheap and easy to handle, H₂-cat is ideal for industrial applications. In biology,¹⁵⁵ one can cite siderophores, a class of functionalized catechols. In catalysis, titanium aryloxides are used as homogeneous catalysts for instance for olefine polymerization,¹⁵⁶ oxidation,¹⁵⁷ epoxidation¹⁵⁸ or the enantioselective formation of carbon–carbon bonds.¹⁵⁹ In molecular coordination chemistry, catechol ligands can strongly stabilize various cations in solution,¹²³ and this allows tuning their electrochemical properties. Catechols have also been used to graft Ru(II) photosensitizers on the surface of TiO₂ photoelectrodes in photovoltaic cells.¹⁶⁰ The use of polycatechol derivatives to build up new Ti-MOFs and benefit from the optical/electrochemical properties already demonstrated at the molecular scale is of strong interest. In this part, we will highlight some of the most relevant reported titanium–catecholate complexes that could be considered as secondary building units for the preparation of new hybrid porous solids (MOFs) based on titanium. Metallo-supramolecular self-assembled Ti-polycatecholate complexes^{161,162–165,166,167} (i.e. cages, helicates) are not discussed in this review.

According to the literature, catechol can coordinate metal ions in different fashions (chelation or μ₂-coordination mode), which are summarized in Fig. 23.¹²³

Fig. 23 Coordination modes of the catecholato ligand characterized through single-crystal X-ray diffraction. The number after the letter is the occurrence found in the 5.21 release (January 2001) of the Cambridge Crystallographic Database. Metal and oxygen atoms are colored blue and red, respectively.¹²³ Reprinted with permission from ref. 123. Copyright 2001 American Chemical Society.

In 1920, Rosenheim et al.¹⁶⁸ observed that catechol has a very high affinity for Ti⁴⁺, considered as a strong Lewis acid. They have reported the synthesis of [NH₄]₂[Ti(cat)₃]·H₂O (Fig. 24a), prepared from TiCl₄ in HCl 1 M solution of catechol in a deoxygenated solution of NH₄OH. They pointed out that a catechol solution is able to dissolve hydrated titanium oxides (relatively insoluble) illustrating the high affinity of catechol to strong Lewis acids. According to the authors, this system appeared to be limited to acidic medium, probably because of subsequent hydrolysis and polymerization of titanium(IV).

Fig. 24 Molecular structures of the core of Ti–catechol/ate complexes and clusters cited in this review. (a) [Ti(cat)₃];^2–168,169 (b) [Ti₂O₂(cat)₄];^4–169 (c) [Ti₂(DTBC)₄(HDTBC)₂];^2–169 (d) Ti₂(OⁱPr)₄(HOⁱPr)₂(cat)₂;^60,170 (e) Ti₂(HOⁱPr)₂(cat)₄;¹⁷⁰ (f) Ti₂(cat)₂(cat-H)₄;⁶⁰ (g) Ti₃O(cat)₄(cat-H)₂;¹⁷² (h) Ti₄O(OⁱPr)₆(DTBC)₄:¹⁷³ (i) Ti₆O(cat)₆(OⁱPr)₁₀;⁶⁰ (j) Ti₁₀O₁₂(cat)₈(py)₈;¹⁷⁴ (k) Ti₁₀O₆(OBuⁿ)₁₂(HOBuⁿ)₂(cat)₈¹²³ and (l) Ti₁₇O₂₄(OⁱPr)₁₆(cat)₄.⁶⁰ Color code: Ti, blue; O, red; C, grey; N, dark blue; C(CH₃)₃, lavender. Counter-ions (ammonium in (a) and (c), and potassium in (b)), C-atoms in alcohol/-ate groups, and all H-atoms of C-atoms are not represented for the sake of clarity. N-atoms represent pyridine groups in (j).

In 1984, Borgias et al.¹⁶⁹ reported two new Ti–catecholate complexes K₄[Ti₂(μ₂-O)(cat)₄]·9H₂O (Fig. 24b) and [NH(CH₃CH₂)₃]₂[Ti₂(DTBC)₄(HDTBC)₂]·2CHCl₃ (DTBC = 3,5-Di-tert-butylcatecholate or [(CH₃)₃C]₂C₆H₂(OH)₂) (Fig. 24c), in addition to the complex [NH(CH₃CH₂)₃]₂[Ti(cat)₃], similar to the one reported earlier by Rosenheim et al.¹⁶⁸ All compounds were here prepared in basic conditions, illustrating the effect of the complexing group on the synthetic conditions. If both dimeric complexes are anionic, only in the first Ti⁴⁺ are connected through μ₂-oxo bridges while in the second two O-atoms of two catechol are bridging.

UV-vis spectroscopic studies have shown that Ti⁴⁺ ions are able to stabilize the cat²⁻ ions more strongly than the Fe³⁺ under acidic conditions.¹⁶⁹ Moreover, the absorption spectrum of [NH₄]₂[Ti(cat)₃] does not change from pH 6 to pH 12 even in the absence of an excess of catechol, suggesting that the complex is preserved. Only strong basic conditions lead to the [TiO(cat)₂]₂⁴⁻ dimer.

Dinuclear Ti–catecholate clusters have also been reported by Davidson et al.¹⁷⁰ Two different molecular structures, Ti₂(μ-OⁱPr)₄(μ-HOⁱPr)₂(cat)₂ (Fig. 24d) and Ti₂(μ-HOⁱPr)₂(cat)₄ (Fig. 24e), have been isolated by reacting under inert conditions Ti(OⁱPr)₄ with one and two equivalents of catechol, respectively. A series of Ti₂(μ-OⁱPr)₄(μ-HOⁱPr)₂(R)₂ has also been obtained by performing similar reactions using substituted catechol derivatives (H₂R) where R = 4-methyl-, 4-tertbutyl-, 3,5-tertbutyl-, 3-methoxy-, 4-nitro-, or 3-methyl-catechol. In all these Ti₂–catecholate dimers, catechol bridges between two titanium centers. In contrast to the anionic dimers reported by Borgias et al.,¹⁶⁹ these complexes are neutral. Due to the potential active sites after removal of coordinated alcohol groups, the catalytic activity of both series (activated in vacuum) for the ring-opening-polymerization of caprolactone to afford polycaprolactone (PCL) was evaluated.¹⁷¹ Ti₂(μ-OⁱPr)₄(μ-HOⁱPr)₂(cat)₂ has shown the highest activity.

In 2005, Boyle et al.¹⁷² reported a trinuclear Ti-oxo-catecholate cluster. Single crystals of Ti₃(μ₃-O)(μ-cat)₃(cat)(cat-H)₂ have been isolated by recrystallizing, from pyridine, the red powder of titanium catecholate obtained after the reaction of catechol with [Ti(μ-ONep)(ONep)₃]₂ (ONep = OCH₂CMe₃) in anhydrous THF. Its structure (Fig. 24g) consists of a three hexa-coordinated Ti⁴⁺ connected through a centered μ₃-O bridge, one chelating catechol and two bridging-chelating catechols, and is, to a certain extent, reminiscent of the trimeric unit found in carboxylate MOFs.

Very recently, a tetra-nuclear Ti-oxo-catecholate based cluster has been reported. Ti₄(μ₄-O)(OⁱPr)₆(DTBC)₄ (DTBC = 3,5-di-tert-butylcatecholate) was obtained under solvothermal conditions by reacting Ti(OⁱPr) and H₂-DTBC in isopropanol.¹⁷³ The four Ti-centers adopt a square-shaped arrangement connected through a central μ₄-oxo bridge and bridging bidentate chelate DTBC (Fig. 24h). XPS measurements revealed that it is a mixed valence oxo-cluster with a Ti⁴⁺/Ti³⁺ ratio of ca. 1 : 1.8. Thanks to the bidentate chelate coordination mode of catecholate derivatives, the oxo-cluster shows strong visible absorption promoted by LMCT, and resulting in an optical gap of ca. 2.46 eV (lower than that calculated from DFT, 2.82 eV).

In the quest of elaborating new Ti–catechol based clusters, Benedict and collaborators⁶⁰ succeeded in isolating new titanium oxo-clusters of different nuclearities. Two di-nuclear, Ti₂(OⁱPr)₄(HOⁱPr)₂(cat)₂ and Ti₂(cat)₂(cat-H)₄ (Fig. 24f), and one hexa-nuclear Ti₆(μ₄-O)(cat)₆(OⁱPr)₁₀ (Fig. 24i) were obtained by direct reaction between Ti(OⁱPr)₄ and catechol under inert conditions. If the structure of Ti₂(cat)₂(OⁱPr)₄(HOⁱPr)₂ is a novel polymorph of a previously reported Ti–catechol dimer,¹⁷⁰ Ti₂(cat)₂(cat-H)₄ (Fig. 24f) is the first titanium–catecholate polynuclear complex bearing purely catechol/ate ligands (showing three coordination modes: monodentate, bidentate and μ₂-(O, O′, O′)). In the Ti₆(μ₄-O)(cat)₆(OⁱPr)₁₀ oxo-cluster, the six catechol ligands are connecting Ti⁴⁺ through two coordination modes, μ₂-(O, O′, O′) for two ligands and μ₂-(O, O′, O, O′) for the others, as well as one μ₄-oxo-bridge.

Besides, an oxo-cluster, Ti₁₇(μ₄-O)₄(μ₃-O)₁₆(μ₂-O)₄(OⁱPr)₁₆(cat)₄⁶⁰ (Fig. 24l), of high nuclearity has also been isolated by reacting the pre-formed [Ti₁₇O₂₄(OⁱPr)₂₀] cluster.⁵⁹ The latter consists of one central tetra-coordinated Ti-center surrounded by sixteen Ti centers, among which a few are five-coordinated and bear isopropoxide groups. Upon reaction with catechol, isopropoxide groups are partially exchanged by bidentate catechols, leading to hexa-coordinated Ti.

Theoretical optical gap calculation limited to the aforementioned Ti-oxo-clusters showed no correlation between the size of the cluster and the gap values (Ti₂cat₂: 3.78 eV, Ti₂cat₆: 3.51 eV, Ti₆cat₆: 3.41 eV and Ti₁₇cat₆: 4.71 eV).

Another example of Ti-oxo-catecholate cluster of high nuclearity has been reported by Gigant et al.¹²³ In agreement with previous studies,¹⁷⁵ it was not possible to obtain single crystals from a molar ratio R of catechol to Ti of greater than 1. Depending on the amount of water added (h = H₂O/Ti) to the resulting clear solution, a good yield of microcrystalline precipitate was observed for h = 0.3 and R = 0.75 with Ti(OBuⁿ)₄ or h = 0.4 and R = 0.8 for Ti(OⁱPr)₄. The crystalline structure of the complex obtained with Ti(OBuⁿ)₄ is [Ti₁₀O₆(OBuⁿ)₁₂(HOBuⁿ)₂(cat)₈].¹²³ It consists of two tetrahedral (μ₄-O)Ti₄ building blocks sharing corners with one central edge-sharing dioctahedral unit and leading to a bowtie-shaped molecule. TiO₆ octahedra are linked together through three catechol and two butoxy ligands.

Recently, Chaumont et al.^125,174 prepared several complexes incorporating Ti–O bonds (phenolate) which absorb again between 300 nm and 500 nm (red-orange). They first studied the reactivity of catechol with a titanium oxo-carboxylate complex of the formula Ti₈O₈(OOCCH₂C(CH₃)₃)₁₆ (see Fig. 4g) having at its periphery carboxylato ligands substituted with tert-butyl groups, such groups being used to significantly increase the solubility of the complex in solution.¹²⁵ At a low catechol ratio (1 to 7 equivalent of catechol per Ti₈O₈ cluster), it was possible to form titanium complexes incorporating both carboxylate and catecholate ligands. With an excess of catechol (32 equivalents), the complex was completely converted in the presence of DMAc (dimethylacetamide) and DMF to produce other complexes exclusively containing catechol as ligand, and formulated Ti₂(cat)₄(DMAc)₂ or Ti₂(cat)₄(DMF)₂. Noteworthily, from the recrystallization of these complexes, they were also able to isolate a new [Ti₁₀O₁₂(cat)₈(py)₈] oxo-cluster.¹⁷⁴ NMR DOSY studies further revealed the lability of the pyridine ligands present at the periphery of the cluster, which could be exchanged by other nitrogen ligands (picoline, 4-phenylpyridine or 2-methoxy-5-(pyridin-4-yl)-benzaldehyde). This exchange is however associated with a reorganization of the original cluster.

These results have highlighted the richness of the titanium–catecholate molecular chemistry. At the sight of these works, the use of phenolate linkers allows obtaining various Ti-oxo-clusters with geometries different from those obtained with carboxylate ligands that can be potentially used as SBUs for the design of MOFs.

However, only one example of titanium-based polycatecholate MOF has been reported to date. Published very recently by Nguyen et al., MOF Ti-CAT-5 was obtained as a crystalline powder after reacting, under solvothermal conditions, the tritopic catechol derivative ligand (2,3,6,7,9,11-hexahydroxytriphenylene, H₆-HHTP or C₁₈H₆(OH)₆) and Ti(OⁱPr)₄ under basic conditions in DMF.¹⁷⁶ The structure of this MOF (Fig. 25) consists of an anionic 3-D framework based on isolated Ti-centers octahedrally coordinated to 3 chelating HHTP ligands. The large aperture of this network is occupied by an identical interpenetrated network and organic cations (DMAH) leading to a strong reduction of the accessible surface area (S_BET = 450 m² g⁻¹). Noteworthily, iron and vanadium analogues of Ti-CAT-5 could be isolated. Proton conductivity measurements revealed lower conductivity values for Ti-CAT-5 in comparison with the iron-based structure (σ = 8.2.10⁻⁴ and 5.0.10⁻² S cm⁻¹, respectively, at 25 °C and 98% RH). The authors attributed this to the differences of pore content between the two forms (DMAH vs. sulfates in addition to water).

Fig. 25 Crystalline structure of Ti-CAT-5:¹⁷⁶ (a) coordination mode between Ti-cation and organic linker; (b) view of the two interpenetrated networks. Ti, blue; O, red; C, grey. H-atoms of C-atoms and pore filling molecules are omitted for the sake of clarity.

Conclusion

If the chemistry of titanium based coordination compounds (molecular and polymers) could appear as not “new” in a sense that it has been initiated more than a few decades ago, there has been since then a continuous effort to develop new architectures with a high degree of structural and chemical diversity.

This review proposes a summary of the solid state coordination chemistry of Ti with oxygenated ligands. On the opposite, nitrogen based ligands are far less explored. To our knowledge only one example of coordination polymer based on isolated Ti⁴⁺ ions connected through N-atoms has been reported to date.¹⁷⁷ Its structure consists of 2-D networks based on Ti^IVCl₄ units linked by tetrahedral ligands bearing four nitrile groups. The scarceness of such compounds could be mainly attributed to the high oxophilic character of Ti-cations rendering their crystallization under ambient conditions and/or in the presence of any oxygenated source (alkoxide, water, etc.) very challenging.

The structural and chemical diversity of Ti-based compounds has been significantly expanded so far through a careful control of the synthetic parameters (precursor, hydrolysis rate, etc.); the incorporation of many types of organic oxygenated linkers (i.e. phosphonate, carboxylate, hydroxycarboxylate, phenolate and catecholate) in polymeric hybrid crystalline materials (coordination polymers, MOFs, etc.) appears now feasible. One could raise several intermediate conclusions from this study: (i) on a qualitative basis, the design of Ti coordination compounds is still hardly achievable and/or predictable due to a lower level of control compared to the other systems based on di-, tri-valent and Zr cation based MOFs, (ii) from the structural point of view, depending mainly on the nature of the ligand and the hydrolysis ratio, the nuclearity and condensation rates strongly vary. More specifically, while increasing the complexing ability of the ligand allows preventing the competitive formation of TiO₂ during the synthesis, it also disfavors the formation of large oxo-clusters. For instance, when Ti alkoxides are considered, the nuclearity of the oxo-cluster ranges from Ti₂⁵⁶ to Ti₃₄⁵⁵ while Ti oxo-carboxylate clusters span from tri-, hexa-, hepta-, and nona-nuclear clusters (ratio carboxylate/alkoxide ≥ 1); when turning to phenolate based systems, most structures exhibit a lower degree of nuclearity, typically from single Ti octahedra to dimers or trimers of octahedra, and only very specific strategies can lead to larger clusters.

In terms of optical properties, which are one of the key targets here, the consequences of their chemical or structural characteristics, particularly in the visible-light region, have been discussed. If one shall mainly follow post-synthetic functionalisation strategies (ligand, inclusion of metal complexes or nanoparticles) to introduce visible light absorption in most cases (phosphonates, carboxylates), the LMCT occurring in Ti–O_phenolate bonds is strongly beneficial to the Ti phenolate based systems with a significant enhancement in light absorption in the visible-light range. Although these solids have still been less explored and suffer from a lower thermal stability, this makes such systems very promising in photocatalysis or opto-electronics.

Finally, if one considers more specifically Ti-based MOFs, the recent progress made within the past decade with a series of new microporous solids, from phosphonates, carboxylates to phenolates, through specific direct synthesis or post-synthetic strategies, confirmed the potential benefits of these systems. Although still one of the most challenging domains of MOF chemistry, there is certainly still strong interest to pursue efforts to rationalize and further expand the scope of chemical and structural diversity of Ti based MOFs or molecular solids in the view of new potential applications.

As highlighted in this review, controlling the kinetics of crystallization remains one of the most prominent strategies allowing crystalline and porous Ti-MOFs to be obtained. Nevertheless, the highly unpredictable polycondensation of titanium cations still renders the design of MOFs a quasi-impossible task so far. In this context, in order to exert more control on the elaboration of new MOFs in the future, two possible strategies can be envisioned. The first one would be to better take advantage of the tools of organic chemistry in combination with the plethora of existing Ti-oxo-clusters to form new porous Ti-MOFs;^102,114 in such a case one should however consider that most Ti-oxo-clusters are not very chemically stable (e.g. alkoxide or carboxylate based complexes) and thus one challenge would be to rely on the most stable ones (e.g. catecholates). Another “dream” of the MOF chemists would be, as it was done during the past decade with trivalent based MOFs (e.g. V, Fe or Al) or more recently with zirconium(IV), to find new chemical routes enabling a certain degree of control over the inorganic polycondensation in the presence of the complexing ligands. It was indeed proposed previously with bis-phosphonate ligands leading however to a very limited number of open framework materials, probably due to the intrinsic character of metal–phosphonate connectivity. Very recently a few interesting results were also reported with phenolate based ligands including the synthesis of new Ti-MOFs as single crystals; although once again these solids exhibit a very limited porous character, mainly due to interpenetration issues, the latter approach shall be pursued in order to reach a higher degree of rationale in the synthesis of porous Ti-MOFs.

Acknowledgements

The authors acknowledge past and present coworkers involved in the synthesis and characterization of new MOFs, especially those involved in the chemistry of Ti-MOFs (H. Dan-Hardi, V. Guillerm, N. Guillou, P. Horcajada, T. L. A. Nguyen, F. Nouar, L. C. Pardo-Perez, F. Ragon and G. Férey).

Notes and references

1. H.-C. Zhou, J. R. Long and O. M. Yaghi, Chem. Rev., 2012, 112, 673–674.
2. H.-C. “Joe” Zhou and S. Kitagawa, Chem. Soc. Rev., 2014, 43, 5415–5418.
3. J. J. Low, A. I. Benin, P. Jakubczak, J. F. Abrahamian, S. A. Faheem and R. R. Willis, J. Am. Chem. Soc., 2009, 131, 15834–15842.
4. T. Devic and C. Serre, Chem. Soc. Rev., 2014, 43, 6097–6115.
5. J. H. Cavka, S. Jakobsen, U. Olsbye, N. Guillou, C. Lamberti, S. Bordiga and K. P. Lillerud, J. Am. Chem. Soc., 2008, 130, 13850–13851.
6. A. Schaate, P. Roy, A. Godt, J. Lippke, F. Waltz, M. Wiebcke and P. Behrens, Chem. – Eur. J., 2011, 17, 6643–6651.
7. G. Mouchaham, L. Cooper, N. Guillou, C. Martineau, E. Elkaïm, S. Bourrelly, P. L. Llewellyn, C. Allain, G. Clavier, C. Serre and T. Devic, Angew. Chem., Int. Ed., 2015, 54, 13297–13301.
8. Y. Bai, Y. Dou, L.-H. Xie, W. Rutledge, J.-R. Li and H.-C. Zhou, Chem. Soc. Rev., 2016, 45, 2327–2367.
9. V. Guillerm, F. Ragon, M. Dan-Hardi, T. Devic, M. Vishnuvarthan, B. Campo, A. Vimont, G. Clet, Q. Yang, G. Maurin, G. Férey, A. Vittadini, S. Gross and C. Serre, Angew. Chem., Int. Ed., 2012, 51, 9267–9271.
10. M. Kim and S. M. Cohen, CrystEngComm, 2012, 14, 4096–4104.
11. A. H. Chughtai, N. Ahmad, H. A. Younus, A. Laypkov and F. Verpoort, Chem. Soc. Rev., 2015, 44, 6804–6849.
12. N. Steunou, S. Förster, P. Florian, C. Sanchez and M. Antonietti, J. Mater. Chem., 2002, 12, 3426–3430.
13. A. Strachota, K. Rodzeń, V. Raus, F. Ribot, M. Janata and E. Pavlova, Eur. Polym. J., 2015, 68, 366–378.
14. G. Fornasieri, L. Rozes, S. Le Calvé, B. Alonso, D. Massiot, M. N. Rager, M. Evain, K. Boubekeur and C. Sanchez, J. Am. Chem. Soc., 2005, 127, 4869–4878.
15. L. Nicole, C. Boissière, D. Grosso, A. Quach and C. Sanchez, J. Mater. Chem., 2005, 15, 3598.
16. C. Sanchez, B. Julián, P. Belleville and M. Popall, J. Mater. Chem., 2005, 15, 3559.
17. Hydra/Medusa Chemical Equilibrium, Database and Plotting Software, KTH Royal Institute of Technology, 6th edn, 2004.
18. S. Cassaignon, M. Koelsch and J.-P. Jolivet, J. Mater. Sci., 2007, 42, 6689–6695.
19. L. Rozes and C. Sanchez, Chem. Soc. Rev., 2011, 40, 1006–1030.
20. N. Steunou, F. Robert, K. Boubekeur, F. Ribot and C. Sanchez, Inorg. Chim. Acta, 1998, 279, 144–151.
21. B. Moraru, N. Hüsing, G. Kickelbick, U. Schubert, P. Fratzl and H. Peterlik, Chem. Mater., 2002, 14, 2732–2740.
22. N. Steunou, F. Ribot, K. Boubekeur, J. Maquet and C. Sanchez, New J. Chem., 1999, 23, 1079–1086.
23. R. Schmid, A. Mosset and J. Galy, J. Chem. Soc., Dalton Trans., 1991, 1999–2005.
24. N. N. Greenwood and A. Earnshaw, Chemistry of the elements, Butterworth-Heinemann, Oxford, 2nd edn, 1997.
25. P. Pascal, Nouveau traité de chimie minérale, Masson, Paris, 2nd edn, 1962.
26. F. A. Cotton and G. Wilkinson, Advanced inorganic chemistry, Wiley, New York, 5th edn, 1988.
27. J.-P. Jolivet, De la solution à l’oxide, InterEditions, CNRS, Paris, 1994.
28. H. Auterhoff, Arch. Pharm., 1965, 298, 400.
29. D. W. Breck, J. Chromatogr. Sci., 1975, 13, 18A.
30. D. M. Chapman and A. L. Roe, Zeolites, 1990, 10, 730–737.
31. M. Taramasso, G. Perego and B. Notari, US Pat., US 4410501A, 1983.
32. M. W. Anderson, O. Terasaki, T. Ohsuna, A. Philippou, S. P. MacKay, A. Ferreira, J. Rocha and S. Lidin, Nature, 1994, 367, 347–351.
33. A. Philippou, M. Naderi, J. Rocha and M. W. Anderson, Catal. Lett., 1998, 53, 221–224.
34. J. Rocha, P. Brandão, J. D. Pedrosa de Jesus, A. Philippou and M. W. Anderson, Chem. Commun., 1999, 471–472.
35. P. Brandão, A. Philippou, A. Valente, J. Rocha and M. Anderson, Phys. Chem. Chem. Phys., 2001, 3, 1773–1777.
36. S. Lima, A. S. Dias, Z. Lin, P. Brandão, P. Ferreira, M. Pillinger, J. Rocha, V. Calvino-Casilda and A. A. Valente, Appl. Catal., A, 2008, 339, 21–27.
37. Y. K. Krisnandi, E. E. Lachowski and R. F. Howe, Chem. Mater., 2006, 18, 928–933.
38. J. Rocha, P. Brandão, Z. Lin, A. P. Esculcas, A. Ferreira and M. W. Anderson, J. Phys. Chem., 1996, 100, 14978–14983.
39. S. M. Kuznicki, V. A. Bell, S. Nair, H. W. Hillhouse, R. M. Jacubinas, C. M. Braunbarth, B. H. Toby and M. Tsapatsis, Nature, 2001, 412, 720–724.
40. G. Alberti, P. Cardini-Galli, U. Costantino and E. Torracca, J. Inorg. Nucl. Chem., 1967, 29, 571–578.
41. S. Allulli, C. Ferragina, A. La Ginestra, M. A. Massucci and N. Tomassini, J. Inorg. Nucl. Chem., 1977, 39, 1043–1048.
42. A. Clearfield and J. A. Stynes, J. Inorg. Nucl. Chem., 1964, 26, 117–129.
43. A. N. Christensen, E. K. Andersen, I. G. K. Andersen, G. Alberti, M. Nielsen, M. S. Lehmann and T. Tokii, Acta Chem. Scand., 1990, 44, 865–872.
44. A. I. Bortun, S. A. Khainakov, L. N. Bortun, D. M. Poojary, J. Rodriguez, J. R. Garcia and A. Clearfield, Chem. Mater., 1997, 9, 1805–1811.
45. C. Serre and G. Férey, J. Mater. Chem., 1999, 9, 579–584.
46. C. Serre and G. Férey, Inorg. Chem., 2001, 40, 5350–5353.
47. C. Serre, F. Taulelle and G. Férey, Chem. Mater., 2002, 14, 998–1003.
48. C. Serre and G. Férey, Inorg. Chem., 1999, 38, 5370–5373.
49. C. Serre, J. A. Groves, P. Lightfoot, A. M. Z. Slawin, P. A. Wright, N. Stock, T. Bein, M. Haouas, F. Taulelle and G. Férey, Chem. Mater., 2006, 18, 1451–1457.
50. V. Benoit, R. S. Pillai, A. Orsi, P. Normand, H. Jobic, F. Nouar, P. Billemont, E. Bloch, S. Bourrelly, T. Devic, P. A. Wright, G. de Weireld, C. Serre, G. Maurin and P. L. Llewellyn, J. Mater. Chem. A, 2016, 4, 1383–1389.
51. U. Schubert, J. Mater. Chem., 2005, 15, 3701–3715.
52. V. W. Day, T. A. Eberspacher, W. G. Klemperer and C. W. Park, J. Am. Chem. Soc., 1993, 115, 8469–8470.
53. V. W. Day, T. A. Eberspacher, W. G. Klemperer, C. W. Park and F. S. Rosenberg, J. Am. Chem. Soc., 1991, 113, 8190–8192.
54. L. Rozes, N. Steunou, G. Fornasieri and C. Sanchez, Monatsh. Chem., 2006, 137, 501–528.
55. P. Coppens, Y. Chen and E. Trzop, Chem. Rev., 2014, 114, 9645–9661.
56. A. Pandey, V. D. Gupta and H. Nöth, Eur. J. Inorg. Chem., 2000, 1351–1357.
57. H. Barrow, D. A. Brown, N. W. Alcock, W. Errington and M. G. Wallbridge, J. Chem. Soc., Dalton Trans., 1994, 3533–3538.
58. W.-H. Fang, L. Zhang and J. Zhang, J. Am. Chem. Soc., 2016, 138, 7480–7483.
59. N. Steunou, G. Kickelbick, K. Boubekeur and C. Sanchez, J. Chem. Soc., Dalton Trans., 1999, 3653–3655.
60. J. B. Benedict and P. Coppens, J. Am. Chem. Soc., 2010, 132, 2938–2944.
61. Z. Liu, J. Lei, M. Frasconi, X. Li, D. Cao, Z. Zhu, S. T. Schneebeli, G. C. Schatz and J. F. Stoddart, Angew. Chem., Int. Ed., 2014, 53, 9193–9197.
62. T. J. Boyle, T. M. Alam, C. J. Tafoya and B. L. Scott, Inorg. Chem., 1998, 37, 5588–5594.
63. R. Ghosh, M. Nethaji and A. G. Samuelson, Chem. Commun., 2003, 2556–2557.
64. A. Bashall, D. A. Brown, M. McPartlin and M. G. H. Wallbridge, J. Chem. Soc., Dalton Trans., 1992, 2529–2530.
65. N. W. Alcock, D. A. Brown, S. M. Roe and M. G. H. Wallbridge, J. Chem. Soc., Chem. Commun., 1992, 846–848.
66. T. Frot, S. Cochet, G. Laurent, C. Sassoye, M. Popall, C. Sanchez and L. Rozes, Eur. J. Inorg. Chem., 2010, 5650–5659.
67. T. Frot, J. Marrot, C. Sanchez, L. Rozes and C. Sassoye, Z. Anorg. Allg. Chem., 2013, 639, 2181–2185.
68. T. Frot, PhD thesis, University of Pierre et Marie Curie, 2009.
69. M. Czakler, C. Artner and U. Schubert, Eur. J. Inorg. Chem., 2014, 2038–2045.
70. W.-H. Fang, L. Zhang and J. Zhang, Dalton Trans., 2017, 46, 803–807.
71. J.-X. Liu, M.-Y. Gao, W.-H. Fang, L. Zhang and J. Zhang, Angew. Chem., Int. Ed., 2016, 55, 5160–5165.
72. K. Hong, W. Bak and H. Chun, Inorg. Chem., 2014, 53, 7288–7293.
73. K. Hong and H. Chun, Inorg. Chem., 2013, 52, 9705–9707.
74. T. J. Boyle, R. P. Tyner, T. M. Alam, B. L. Scott, J. W. Ziller and B. G. Potter, J. Am. Chem. Soc., 1999, 121, 12104–12112.
75. P. S. Ammala, S. R. Batten, C. M. Kepert, L. Spiccia, A. M. van den Bergen and B. O. West, Inorg. Chim. Acta, 2003, 353, 75–81.
76. P. Piszczek, M. Richert and A. Wojtczak, Polyhedron, 2008, 27, 602–608.
77. P. Piszczek, A. Grodzicki, M. Richert and A. Wojtczak, Inorg. Chim. Acta, 2004, 357, 2769–2775.
78. R. Papiernik, L. G. Hubert-Pfalzgraf, J. Vaissermann and M. C. H. B. Goncalves, J. Chem. Soc., Dalton Trans., 1998, 2285–2288.
79. G. A. Seisenbaeva, E. Ilina, S. Håkansson and V. G. Kessler, J. Sol–Gel Sci. Technol., 2010, 55, 1–8.
80. A. Rammal, F. Brisach and M. Henry, C. R. Chim., 2002, 5, 59–66.
81. I. Gautier-Luneau, A. Mosset and J. Galy, Z. Kristallogr., 1987, 180, 83–95.
82. S. Cochet, L. Rozes, M. Popall and C. Sanchez, Mater. Sci. Eng., C, 2007, 27, 1401–1405.
83. P. Heinz, M. Puchberger, M. Bendova, S. O. Baumann and U. Schubert, Dalton Trans., 2010, 39, 7640–7644.
84. Y. Gao, F. R. Kogler, H. Peterlik and U. Schubert, J. Mater. Chem., 2006, 16, 3268–3276.
85. T. Frot, C. Sassoye, C. Sanchez and L. Rozes, private communication, 2010.
86. G. Kickelbick and U. Schubert, Eur. J. Inorg. Chem., 1998, 159–161.
87. K. Hong, W. Bak and H. Chun, Inorg. Chem., 2014, 53, 7288–7293.
88. K. Hong and H. Chun, Inorg. Chem., 2013, 52, 9705–9707.
89. M. Sabo, W. Böhlmann and S. Kaskel, J. Mater. Chem., 2006, 16, 2354–2357.
90. C. Maurer, B. Baumgartner, S. Pabisch, J. Akbarzadeh, H. Peterlik and U. Schubert, Dalton Trans., 2014, 43, 950–957.
91. M. Dan-Hardi, C. Serre, T. Frot, L. Rozes, G. Maurin, C. Sanchez and G. Férey, J. Am. Chem. Soc., 2009, 131, 10857–10859.
92. A. Walsh and C. R. A. Catlow, ChemPhysChem, 2010, 11, 2341–2344.
93. Note that the terminology of the field of semiconductors (gap, valence band, conduction band) is often used to describe the optoelectronic properties of MOFs. Nevertheless, most MOFs are indeed not semiconductors, and can be better described from an electronic standpoint as an array of electronically isolated, molecular entities whose HOMO and LUMO levels define absorption bands rather than from the band theory.
94. C. H. Hendon, D. Tiana, M. Fontecave, C. Sanchez, L. D’arras, C. Sassoye, L. Rozes, C. Mellot-Draznieks and A. Walsh, J. Am. Chem. Soc., 2013, 135, 10942–10945.
95. Y. Fu, D. Sun, Y. Chen, R. Huang, Z. Ding, X. Fu and Z. Li, Angew. Chem., Int. Ed., 2012, 51, 3364–3367.
96. C. Zlotea, D. Phanon, M. Mazaj, D. Heurtaux, V. Guillerm, C. Serre, P. Horcajada, T. Devic, E. Magnier, F. Cuevas, G. Férey, P. L. Llewellyn and M. Latroche, Dalton Trans., 2011, 40, 4879–4881.
97. C. Gomes Silva, I. Luz, F. X. Llabrés i Xamena, A. Corma and H. García, Chem. – Eur. J., 2010, 16, 11133–11138.
98. M. A. Nasalevich, M. G. Goesten, T. J. Savenije, F. Kapteijn and J. Gascon, Chem. Commun., 2013, 49, 10575–10577.
99. Y. Horiuchi, T. Toyao, M. Saito, K. Mochizuki, M. Iwata, H. Higashimura, M. Anpo and M. Matsuoka, J. Phys. Chem. C, 2012, 116, 20848–20853.
100. B. Bueken, F. Vermoortele, D. E. P. Vanpoucke, H. Reinsch, C.-C. Tsou, P. Valvekens, T. De Baerdemaeker, R. Ameloot, C. E. A. Kirschhock, V. Van Speybroeck, J. M. Mayer and D. De Vos, Angew. Chem., Int. Ed., 2015, 54, 13912–13917.
101. S. Yuan, T.-F. Liu, D. Feng, J. Tian, K. Wang, J. Qin, Q. Zhang, Y.-P. Chen, M. Bosch, L. Zou, S. J. Teat, S. J. Dalgarno and H.-C. Zhou, Chem. Sci., 2015, 6, 3926–3930.
102. H. L. Nguyen, F. Gándara, H. Furukawa, T. L. H. Doan, K. E. Cordova and O. M. Yaghi, J. Am. Chem. Soc., 2016, 138, 4330–4333.
103. M. Kim, J. F. Cahill, H. Fei, K. A. Prather and S. M. Cohen, J. Am. Chem. Soc., 2012, 134, 18082–18088.
104. Y. Lee, S. Kim, J. K. Kang and S. M. Cohen, Chem. Commun., 2015, 51, 5735–5738.
105. L. Zou, D. Feng, T.-F. Liu, Y.-P. Chen, S. Yuan, K. Wang, X. Wang, S. Fordham and H.-C. Zhou, Chem. Sci., 2016, 7, 1063–1069.
106. C. K. Brozek and M. Dincă, J. Am. Chem. Soc., 2013, 135, 12886–12891.
107. D. H. Son, Science, 2004, 306, 1009–1012.
108. A. S. Yasin, J. Li, N. Wu and T. Musho, Phys. Chem. Chem. Phys., 2016, 18, 12748–12754.
109. J. A. Mason, L. E. Darago, W. W. Lukens and J. R. Long, Inorg. Chem., 2015, 54, 10096–10104.
110. V. Guillerm, S. Gross, C. Serre, T. Devic, M. Bauer and G. Férey, Chem. Commun., 2010, 46, 767–769.
111. S. Surblé, C. Serre, C. Mellot-Draznieks, F. Millange and G. Férey, Chem. Commun., 2006, 284–286.
112. S. Hausdorf, F. Baitalow, T. Böhle, D. Rafaja and F. O. R. L. Mertens, J. Am. Chem. Soc., 2010, 132, 10978–10981.
113. C. Serre, F. Millange, S. Surblé and G. Férey, Angew. Chem., Int. Ed., 2004, 43, 6285–6289.
114. H. L. Nguyen, T. T. Vu, D. Le, T. L. H. Doan, V. Q. Nguyen and N. T. S. Phan, ACS Catal., 2017, 7, 338–342.
115. J. Zhao, L. Mi, J. Hu, H. Hou and Y. Fan, J. Am. Chem. Soc., 2008, 130, 15222–15223.
116. H. Fei, M. R. Bresler and S. R. J. Oliver, J. Am. Chem. Soc., 2011, 133, 11110–11113.
117. H. Fei, C. H. Pham and S. R. J. Oliver, J. Am. Chem. Soc., 2012, 134, 10729–10732.
118. S. Das, H. Kim and K. Kim, J. Am. Chem. Soc., 2009, 131, 3814–3815.
119. T.-F. Liu, L. Zou, D. Feng, Y.-P. Chen, S. Fordham, X. Wang, Y. Liu and H.-C. Zhou, J. Am. Chem. Soc., 2014, 136, 7813–7816.
120. R. J. Comito, K. J. Fritzsching, B. J. Sundell, K. Schmidt-Rohr and M. Dinca, J. Am. Chem. Soc., 2016, 138, 10232–10237.
121. A. Wang, Y. Zhou, Z. Wang, M. Chen, L. Sun and X. Liu, RSC Adv., 2016, 6, 3671–3679.
122. C. J. Chuck, M. G. Davidson, M. D. Jones, G. Kociok-Köhn, M. D. Lunn and S. Wu, Inorg. Chem., 2006, 45, 6595–6597.
123. K. Gigant, A. Rammal and M. Henry, J. Am. Chem. Soc., 2001, 123, 11632–11637.
124. P. Persson, R. Bergström and S. Lunell, J. Phys. Chem. B, 2000, 104, 10348–10351.
125. C. Chaumont, E. Huen, C. Huguenard, P. Mobian and M. Henry, Polyhedron, 2013, 57, 70–76.
126. D. M. Weekes, C. Diebold, P. Mobian, C. Huguenard, L. Allouche and M. Henry, Chem. – Eur. J., 2014, 20, 5092–5101.
127. H. Assi, L. C. Pardo Pérez, G. Mouchaham, F. Ragon, M. Nasalevich, N. Guillou, C. Martineau, H. Chevreau, F. Kapteijn, J. Gascon, P. Fertey, E. Elkaim, C. Serre and T. Devic, Inorg. Chem., 2016, 55, 7192–7199.
128. L. Cooper, N. Guillou, C. Martineau, E. Elkaim, F. Taulelle, C. Serre and T. Devic, Eur. J. Inorg. Chem., 2014, 6281–6289.
129. J. Gao, J. Miao, P.-Z. Li, W. Y. Teng, L. Yang, Y. Zhao, B. Liu and Q. Zhang, Chem. Commun., 2014, 50, 3786–3788.
130. Y.-L. Fu, Z.-W. Xu, J.-L. Ren and S. W. Ng, Acta Crystallogr., Sect. E: Struct. Rep. Online, 2005, 61, m1733–m1734.
131. Y.-L. Fu, Z.-W. Xu, J.-L. Ren and S. W. Ng, Acta Crystallogr., Sect. E: Struct. Rep. Online, 2005, 61, m1730–m1732.
132. A. K. Babko, O. I. Volkova and T. E. Get’man, Zh. Neorg. Khim., 1961, 6, 354–359.
133. J.-L. Hou, W. Luo, Y.-Y. Wu, H.-C. Su, G.-L. Zhang, Q.-Y. Zhu and J. Dai, Dalton Trans., 2015, 44, 19829–19835.
134. K. Hong, W. Bak and H. Chun, Inorg. Chem., 2013, 52, 5645–5647.
135. K. Hong and H. Chun, Chem. Commun., 2013, 49, 10953–10955.
136. Y. Fu, Y. Liu, Z. Shi, B. Li and W. Pang, J. Solid State Chem., 2002, 163, 427–435.
137. M. A. Nasalevich, C. H. Hendon, J. G. Santaclara, K. Svane, B. van der Linden, S. L. Veber, M. V. Fedin, A. J. Houtepen, M. A. van der Veen, F. Kapteijn, A. Walsh and J. Gascon, Sci. Rep., 2016, 6, 23676.
138. M. A. Nasalevich, R. Becker, E. V. Ramos-Fernandez, S. Castellanos, S. L. Veber, M. V. Fedin, F. Kapteijn, J. N. H. Reek, J. I. van der Vlugt and J. Gascon, Energy Environ. Sci., 2015, 8, 364–375.
139. Note that our research on the CSD data base (version 5.37) was limited to phenol/-ate derivatives where o-positions were strictly considered as H-atoms.
140. R. Minhas, R. Duchateau, S. Gambarotta and C. Bensimon, Inorg. Chem., 1992, 31, 4933–4938.
141. A. J. Nielson, J. A. Harrison, C. Shen and J. M. Waters, Polyhedron, 2006, 25, 1729–1736.
142. A. N. Desnoyer, B. Fartel, K. C. MacLeod, B. O. Patrick and K. M. Smith, Organometallics, 2012, 31, 7625–7628.
143. J. A. Suttil, D. S. McGuinness, M. Pichler, M. G. Gardiner, D. H. Morgan and S. J. Evans, Dalton Trans., 2012, 41, 6625–6633.
144. G. W. Svetich and A. A. Voge, Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem., 1972, 28, 1760–1767.
145. R. R. Gowda, D. Chakraborty and V. Ramkumar, Eur. J. Inorg. Chem., 2009, 2981–2993.
146. P. J. Fischer, P. Yuen, V. G. Young and J. E. Ellis, J. Am. Chem. Soc., 1997, 119, 5980–5981.
147. K. Watenpaugh and C. N. Caughlan, Inorg. Chem., 1966, 5, 1782–1786.
148. R. R. Burch, Chem. Mater., 1990, 2, 633–635.
149. Y. Zhao, S.-Y. Lee, N. Becknell, O. M. Yaghi and C. A. Angell, J. Am. Chem. Soc., 2016, 138, 10818–10821.
150. T. P. Vaid, E. B. Lobkovsky and P. T. Wolczanski, J. Am. Chem. Soc., 1997, 119, 8742–8743.
151. T. P. Vaid, J. M. Tanski, J. M. Pette, E. B. Lobkovsky and P. T. Wolczanski, Inorg. Chem., 1999, 38, 3394–3405.
152. J. M. Tanski, T. P. Vaid, E. B. Lobkovsky and P. T. Wolczanski, Inorg. Chem., 2000, 39, 4756–4765.
153. J. M. Tanski, E. B. Lobkovsky and P. Wolczanski, J. Solid State Chem., 2000, 152, 130–140.
154. J. M. Tanski and P. T. Wolczanski, Inorg. Chem., 2001, 40, 2026–2033.
155. K. N. Raymond and C. J. Carrano, Acc. Chem. Res., 1979, 12, 183–190.
156. D. C. H. Oakes, B. S. Kimberley, V. C. Gibson, D. J. Jones, A. J. P. White and D. J. Williams, Chem. Commun., 2004, 2174–2175.
157. M. Fujiwara, Q. Xu, Y. Souma and T. Kobayashi, J. Mol. Catal. A: Chem., 1999, 142, 77–84.
158. Y. Pérez, I. del Hierro, M. Fajardo and A. Otero, J. Organomet. Chem., 2003, 679, 220–228.
159. Y. Chen, S. Yekta and A. K. Yudin, Chem. Rev., 2003, 103, 3155–3212.
160. W. R. Duncan and O. V. Prezhdo, J. Phys. Chem. B, 2005, 109, 365–373.
161. C. Diebold, P. Mobian, C. Huguenard, L. Allouche and M. Henry, Inorg. Chem., 2010, 49, 6369–6371.
162. M. Albrecht, H. Röttele and P. Burger, Chem. – Eur. J., 1996, 2, 1264–1268.
163. M. Albrecht, M. Schneider and R. Fröhlich, New J. Chem., 1998, 22, 753–754.
164. M. Albrecht, M. Schneider and H. Röttele, Angew. Chem., Int. Ed., 1999, 38, 557–559.
165. M. Albrecht, I. Janser, J. Runsink, G. Raabe, P. Weis and R. Fröhlich, Angew. Chem., Int. Ed., 2004, 43, 6662–6666.
166. H. Senouci, B. Millet, C. Volkringer, C. Huguenard, F. Taulelle and M. Henry, C. R. Chim., 2010, 13, 69–96.
167. D. Van Craen, M. Albrecht, G. Raabe, F. Pan and K. Rissanen, Chem. – Eur. J., 2016, 22, 3255–3258.
168. A. Rosenheim and O. Sorge, Ber. Dtsch. Chem. Ges. B, 1920, 53, 932–939.
169. B. A. Borgias, S. R. Cooper, Y. B. Koh and K. N. Raymond, Inorg. Chem., 1984, 23, 1009–1016.
170. M. G. Davidson, M. D. Jones, M. D. Lunn and M. F. Mahon, Inorg. Chem., 2006, 45, 2282–2287.
171. A.-C. Albertsson and I. K. Varma, Biomacromolecules, 2003, 4, 1466–1486.
172. T. J. Boyle, L. J. Tribby, T. M. Alam, S. D. Bunge and G. P. Holland, Polyhedron, 2005, 24, 1143–1152.
173. J. Hou, Q. Zhang, Y. Wu, Y. Liu, L. Du, C.-H. Tung and Y. Wang, Inorg. Chim. Acta, 2016, 443, 279–283.
174. C. Chaumont, P. Mobian and M. Henry, Dalton Trans., 2014, 43, 3416–3419.
175. X. Lei, M. Shang and T. P. Fehlner, Organometallics, 1996, 15, 3779–3781.
176. N. T. T. Nguyen, H. Furukawa, F. Gándara, C. A. Trickett, H. M. Jeong, K. E. Cordova and O. M. Yaghi, J. Am. Chem. Soc., 2015, 137, 15394–15397.
177. F.-Q. Liu, F.-Q. Liu and T. Don Tilley, Chem. Commun., 1998, 103–104.

---

Footnote

† Authors contributed equally to this work.

---