**3. Discussion**

The MIL-88B material provides the Lewis acidity that catalyses isomerisation of glucose to fructose, and a cooperative effect of the Fe and Sc in the mixed-metal MIL-88B MOF provides optimum catalytic activity to maximise the 5-HMF yields, while the DMSO solvent catalyses the fructose dehydration to 5-HMF. It is notable that the mixed-metal material shows enhanced reactivity compared to the iron or scandium catalysts, and the presence of a locally disordered structure, such as uncoordinated ligands, may be partly responsible for this. Further work is needed to elucidate the mechanism of reaction, such as NMR studies with isotopic substitution, especially since it is likely that the surface of the MOF crystallites provides the active sites for catalysis and that solvent effects may be at play in this chemistry, but our work shows how one-step synthesis of MOF catalysts under moderate conditions provides a convenient route to acid catalysts for biomass transformations. The long-term stability of the catalysts and their implementation in flow reactors is the next step towards realistic use.

#### **4. Materials and Methods**

#### *4.1. Materials and Chemicals*

All reagents were purchased and used without further purification. FeCl3·6H2O (99.9%) was obtained from Sigma Aldrich (Gillingham, United Kingdom). ScCl3·6H2O (99.9%) and benzene-1,4-dicarboxylic acid (H2BDC, 98%) were obtained from Alfa Aesar (Heysham, United Kingdom). Monosodium 2-sulfobenzene-1,4-dicarboxylate (NaH2MSBDC) was obtained from Tokyo Chemical Industry (TCI, >98%, Tokyo, Japan). DMF was purchased from Fischer Scientific (Loughborough, United Kingdom) and absolute ethanol was purchased from VWR Chemicals.

#### *4.2. Synthesis of MIL-101(Fe,Sc)*

MIL-101(Fe) was prepared following the method of Zhu et al. [43], FeCl3·6H2O (0.894 g, 3.3 mmol), H2BDC (0.280 g, 1.65 mmol), and DMF (20 ml) were mixed with stirring at room temperature for 2 h. The mixture was transferred to a Teflon-lined autoclave (45 mL) and heated at 383 K for either 3 or 40 hours (see below). After cooling to room temperature, the mixture was filtered and the solid product was washed with ethanol (20 mL) at 60 ◦C for 3 h. The product was dried at 60 ◦C overnight. Variation of metal ratio (Fe:Sc) and ligand using MSBDC in place of BDC was investigated by following the same synthesis method, but replacing the FeCl3·6H2O with ScCl3·6H2O and/or the H2BDC with NaH2MSBDC.

#### *4.3. Synthesis of MIL-88B(Sc)*

For comparison, MIL-88B(Sc) was prepared using a method based on that of Mowat et al. [39], where 0.22 g (1.45 mmol), scandium (III) chloride, 0.14 g (8.43 mmol) H2BDC, and 9 mL DMF (Fisher Scientific) were stirred for 5 minutes prior to transfer into a 45 mL Teflon -lined autoclave and heated to 140 ◦C for 48 h before cooling to room temperature. The solid product was isolated by suction filtration and washed with 25 mL ethanol by stirring at room temperature for 24 h. The recovered solid was finally dried at 70 ◦C in the air for 24 h.

### *4.4. Materials Characterisation*

For sample identification, PXRD patterns were recorded on a D5000 Siemens di ffractometer with Cu K <sup>α</sup>1/2 radiation (λ = 1.54184 Å) and 2θ= 2−30◦ (step size 0.02◦ in 2θ), operated at 40 kV and 40 mA. High resolution PXRD patterns were recorded using a Panalytical X'Pert Pro MPD (Malvern Panalytical, Malvern, United Kingdom), equipped with monochromatic Cu Kα<sup>1</sup> radiation (λ = 1.54056 Å) and a PIXcel solid state detector. Profile fitting of the powder patterns was performed using the GSAS software (revision 1188, Los Alamos, USA) [69] implemented using the EXPGUI interface [70]. FT-IR spectra were measured at room temperature in the range 400−4000 cm<sup>−</sup><sup>1</sup> using a Platinum-ATR Bruker Alfa instrument (Bruker Optics GmbH, Ettlingen, Germany). The stability of the catalysts was investigated by TGA using a Mettler Toledo TGA/DSC1 instrument (Leicester, United Kingdom) under ambient air pressure and a heating rate of 10 ◦C min−1. Samples were heated in air from 25 ◦C to 1000 ◦C. Selected elemental compositions of materials were determined by energy dispersive X-ray (EDX) analysis using a FEI scanning electron microscope (SEM, Fei UK Limited, Altrincham, United Kingdom). Elemental analysis for metals was performed by MEDAC Ltd, Surrey, United Kingdom, using the ICP-OES method after acid digestion. XPS was measured using a Kratos AXIS Ultra DLD (Manchester, United Kingdom). XPS measurements were carried out in a UHV system with a base pressure of 5 × 10−<sup>11</sup> mbar. The sample was excited with X-rays from a monochromated Al K α source (1486.7 eV), with the photo-electrons being detected at a 90◦ take-o ff angle with respect to the sample surface. Curve fitting was performed using the CasaXPS package, incorporating Voigt (mixed Gaussian–Lorentzian) line shapes and Shirley backgrounds for all regions except the Sc 2*p* region, whereaU2Tougaard background was found to be more appropriate.

The acidity of materials was characterised using ammonia TPD. An excess of 0.02 vol% ammonia in helium was dosed onto 50–70 mg of a catalyst contained in a quartz tube at 100 ◦C (to minimise physisorption). The ammonia was then desorbed from the catalysts by heating the material to 350 ◦C at a ramp rate of 2 ◦C min−1. To ensure the complete desorption of ammonia from the material, the temperature was then maintained at 350 ◦C for 6 hours. The amount of ammonia desorbed from the catalyst was measured using a mass spectrometer at *m*/*z* = 17 with the interference with water vapour taken into account. A blank experiment (used as a baseline) was performed with a fresh catalyst, but without ammonia pre-adsorption. Surface area measurements were performed using nitrogen adsorption via the Brunauer–Emmett–Teller (BET) method using a QUADRASORB (gas sorption surface area analyzer) (Quantachrome UK, Hook, United Kingdom) after degassing samples under vacuum at 100 ◦C.
