**1. Introduction**

Biomass-derived glucose has received considerable attention as a potential feedstock for the production of useful organic molecules. One attractive target of glucose conversion is the molecule 5-hydroxymethyl furfural (5-HMF): this is regarded as a platform molecule that can be used as a precursor for an extensive range of plastics, polymers, and fuels [1–4]. Acid catalysts are needed to bring about the transformation of glucose, via fructose, into 5-HMF, and both Lewis and Brønsted acids have been implicated in the transformation mechanism [5–8]. Figure 1 shows the most common reaction pathways that have been proposed in the literature [9,10]. It is well established that the isomerisation of glucose to fructose requires a Lewis acid catalyst, and mannose can be produced as an alternative product, while the dehydration of fructose to 5-HMF is catalysed by Brønsted acids that may also result in the further transformation (hydrolysis) to levulinic acid and formic acid [10]. Humins are insoluble carbon-rich materials that are commonly formed as a side-product, and other small molecules may be possible as intermediates and byproducts, depending on the choice of catalyst and solvent [2].

**Figure 1.** Pathway for the conversion of glucose to 5-HMF, showing possible byproducts, as described in the literature [9,10]. Glucose and mannose are shown in their α-pyranose forms and fructose in its α-furanose form.

Heterogeneous catalysts are desirable for the conversion of glucose, and Davis and co-workers pioneered the use of zeolite-based catalysts that contain both Lewis and Brønsted acid sites; in particular, tin-substituted zeolite beta (Sn-beta) has been shown as an effective catalyst for the isomerisation of glucose into fructose and mannose in water [9], or the direct formation of 5-HMF from glucose when a biphasic water/tetrahydrofuran reactor system was used [11]. The stability of the heterogeneous catalyst in the reaction medium is a major challenge in the conversion of glucose [12]. Although Sn-beta is hydrothermally stable at low pH and the solid can be recycled, a significant disadvantage in its use, particularly on an industrial scale, is that the synthesis of the catalyst is a lengthy process requiring a 40 day hydrothermal reaction, and furthermore requires hydrofluoric acid, an extremely toxic and corrosive substance [9]. A challenge is to identify new solution-stable solid–acid catalysts that provide glucose conversion with selectivity towards fructose and 5-HMF; the catalysts must also be easy to prepare, using environmentally benign reagents and processes. In recent years, porous metal–organic frameworks (MOFs) have been studied for wide-ranging applications [13], such as adsorbents [14], gas storage and separation [15], and heterogeneous catalysts [16–20]. MOF materials have many attractive features for catalysis, such as large surface areas and pore sizes, as well as controllable properties, since the wide choice of ligands and metals provides a large variety of structures, porosities, and chemical functionalities. The scope for adding extra functionality, for example, by post-synthesis modification of the organic ligands, offers the possibility of multifunctional heterogeneous catalysts that contain more than one active site for tandem or multi-step conversions [21].

Brønsted acid functionality in MOFs has been introduced in a number of ways [22], but particularly useful is the addition of sulfonic acid (-SO3H) functionalities to the organic ligands, such as by their addition to aromatic polycarboxylate linkers that make up many common MOF structures [23,24]. Lewis acidity may be present in MOF structures in the form of coordinatively unsaturated metal sites, such as metals associated with loosely bound solvent, in addition to the coordinating atoms of the ligands making up the framework [25]. The dehydration of glucose to 5-HMF using MIL-101(Cr) functionalised with sulfonic acid groups was reported by Herbst and Janiak in 2016 [10]. The highest 5-HMF yield (29%) was obtained at 403 K with a 5 hour reaction time in a THF:water (39:1 v:v) solvent. Su et al. studied the same catalyst and found the highest conversions were possible in mixed organic:aqueous solvents, and demonstrated the effectiveness of the solid in a fixed-bed continuous reactor [26]. MIL-101(Cr), UiO-66, and MIL-53(Al) functionalised with sulfonic acid groups have also been studied as catalysts for glucose conversion, and MIL-101(Cr)-SO3H showed high efficiency, with more than 90% yield of 5-HMF from fructose when dimethyl sulfoxide (DMSO) was used as solvent [27]. Previously, MIL-101(Cr) had been used as a host for phosphotungstic acid, thereby adding Brønsted acidity for carbohydrate dehydration [28], while most recently, Guo et al. found that a composite of MIL-101(Cr) and chromium hydroxide gave superior isomerisation of glucose to fructose in ethanol solvent via ethyl fructoside that required hydrolysis in a second step [29]. Other MOFs that have been used for glucose conversion include the zirconium-based NU-1000, with phosphate modification to induce Lewis acidity [30], and the zirconium material UiO-66 with sulfo-modified ligands and inherent Lewis acidity from defects notably allowing conversion of glucose to fructose, along with a significant amount of 5-HMF in water alone [31], and selective conversion of glucose to fructose with iso-propanol solvent [32]. Sulfonated UiO-66 has been further tuned to optimise activity towards glucose conversion [33], compared with other zirconium MOFs [34], and has also been used to effect the esterification of levulinic acid with ethanol [35].

Based on the efficiency of MIL-101(Cr) as a catalyst for glucose conversion, we have now investigated the modification of this material by substitution of the metals making up the structure. MIL-101 is constructed from the cheap and benign ligand benzene-1,4-dicarboxylate (BDC), but can be prepared with various other trivalent cations [36]. The metal Cr is known to be toxic to humans and harmful to the environment, and although these detrimental properties are largely associated with the +6 oxidation state, which is established to be carcinogenic, Cr3+ is considered an irritant, with evidence also for it being an allergen [37]; therefore its release into the environment is clearly undesirable, where it might encounter oxidising conditions. Herein, we report a study of variants of MIL-101 using Fe and Sc as replacements for Cr and their potential as catalysts for glucose conversion. Homogeneous scandium triflate is a powerful Lewis acid catalyst [38], and the synthesis of a range of porous scandium carboxylate MOFs with structures analogous to MIL-88B, MIL-100, and MIL-53 was described previously [39], with MIL-100 (Sc) proven to be an effective Lewis-acid catalyst [40]. Mixed-metal MIL-100 (Sc,Fe) materials have been used as dual acid-redox catalysts [41]. Figure 2 shows the structures of two of the MOFs that are relevant to the current work: MIL-101 and MIL-88B. Both are constructed from the same trimeric building unit, consisting of three octahedrally coordinated metal centres with a common shared oxide, and linked by benzene-1,4-dicarboxylate (BDC), but the connectivity gives different structures [42]. Each octahedral metal centre contains a bound solvent molecule, as indicated by the terminal atom on Figure 2a, which may provide a Lewis acid site.

**Figure 2.** The structures of the MOFs investigated in this work: (**a**) the trimeric building unit showing three {FeO6} octahedral units linked by a common oxygen atom, (**b**) the cubic unit cell of MIL-101, and (**c**) four unit cells of hexagonal MIL-88B. Carbon atoms of the benzene-1,4-dicarboxylate (BDC) linker are shown as grey spheres.
