Tuning the Hydrophobicity and Lewis Acidity of UiO-66-NO 2 with Decanoic Acid as Modulator to Optimise Conversion of Glucose to 5-Hydroxymethylfurfural

: Glucose conversion to 5-hydroxymethylfurfural (HMF) is important to the success of a bioreﬁnery. Herein, metal–organic frameworks (MOFs) with the UiO-66 structure were synthesised with decanoic acid as the modulator and used as the catalyst to optimise HMF yield. PXRD, FTIR, and TGA/DSC techniques were applied to characterise the materials. The analysis results show that the materials assembled from the ligand 2-nitroterephthalic acid and hexameric Zr-oxo clusters contain decanoic acid chemically bound in the framework that inﬂuences porosity, Lewis acidity, and hydrophobicity. The materials exhibit excellent catalytic performance for HMF production from glucose in DMSO as solvent, attributed to their abundant defects and high hydrophobicity due to the addition of the decanoic acid modulator. Inﬂuences of catalyst dosages, reaction duration, and temperature were comprehensively investigated, leading to 98.1% conversion of glucose and 54.5% HMF yield under optimised reaction conditions. The catalytic conversion shows some deterioration after four cycles, yet the reaction selectivity displays no signiﬁcant decline.


Introduction
The efficient conversion of cellulosic biomass, a cheap and abundant resource, can contribute to easing the dependence on fossil-derived hydrocarbons, and alleviate environmental issues and the energy crisis triggered by the excessive consumption of petroleum resources. Among various platform chemicals derived from biomass, 5-hydroxymethylfurfural (HMF) is considered as a promising intermediate compound to produce varieties of commodity chemicals [1][2][3][4]. However, the cost of HMF production is presently prohibitive economically. Therefore, great efforts have been made to reduce the production cost of HMF through innovative technology and catalyst design [5][6][7][8][9][10][11].
Highly selective production of HMF from fructose has been widely studied, and the highest HMF yield can reach up to 98% [12][13][14]. However, synthesising HMF from glucose, which is more abundant and has lower cost compared with fructose, is still challenging [15]. During the conversion, glucose is first isomerised into fructose, followed by the dehydration of fructose into HMF [16]. It is generally accepted that the isomerisation process is catalysed by Lewis acids and dehydration reaction with Brønsted acids. Therefore, designing and tuning catalysts to contain both Lewis and Brønsted acids for glucose conversion is crucial to improve selectivity towards the desired products. From this perspective, metal-organic frameworks (MOFs), in which metal or clusters are connected by ligands to form porous materials, constitute a promising class of catalysts for HMF production due to the ease of combining Lewis and Brønsted acidity within their structures [17]. Moreover, MOFs possess outstanding properties such as high surface area, tunable functional groups, and an atomic-level control over pore structure [18]. It is thus worth exploring the huge

Results and Discussion
The crystallinity and phase purity of the UiO-66-NO 2 -X materials were analysed by using PXRD, as shown in Figure 1. The PXRD patterns of all the samples are characterised by strongest peaks at angles of 7.4 • and 8.5 • , representing the crystal planes (111) and (200) that agree well with that expected for cubic UiO-66 with space group Fm3m [38]. Fitting the profile of the whole PXRD pattern indicates that all the samples possess the same UiO-66 framework topology. This also reveals that all the samples have almost the same unit cell size and that decanoic acid has no significant impact on it. However, differences occur in the patterns around the (111) reflection (the 2θ~7.4 • ), emphasised in Figure 2. As can be seen, there is a broad feature spanning a 2θ range of 4-6 • for the synthesised UiO-66-NO 2 -X. This can be attributed to the reo phase, which is considered as UiO-66 with one-quarter of its clusters missing, according to the study conducted by Shearer et al. [39]. The presence of this broad peak therefore suggests that decanoic acid as modulator causes missing clusters, which leads to the formation of defects in the materials. To test this hypothesis, UiO-66-NO 2 was synthesised with formic acid as modulator under the same conditions. The broad diffraction feature mentioned above is not evident for this sample, indicating that formic acid has not such a significant influence on the orientation of clusters as decanoic acid, which may be due to its small steric effect. To investigate the effect of this defective structure further, the method developed by Feng et al. was applied to quantify the accessible Lewis acidic sites in MOFs, and the results are shown in Table 1 [40]. For the sample synthesised without modulator, the amount of Lewis acidic sites is too low to be detected. A significant enhancement of Lewis acidic sites is observed as amount of decanoic acid is increased from 15 mmol to 25 mmol, and then decreases. In UiO-66-NO 2 -25, the amount of Lewis acidic sites is 0.0061 mmol base per mg, which is about 15 times as high as that of formic acid (0.0004 mmol base per mg), suggesting a profound effect of decanoic acid on the number of Lewis acidic sites. It can be thus concluded that using decanoic acid as modulator is an effective way to enhance the Lewis acidity of UiO-66-NO 2 . steric effect. To investigate the effect of this defective structure further, the method developed by Feng et al. was applied to quantify the accessible Lewis acidic sites in MOFs, and the results are shown in Table 1 [40]. For the sample synthesised without modulator, the amount of Lewis acidic sites is too low to be detected. A significant enhancement of Lewis acidic sites is observed as amount of decanoic acid is increased from 15 mmol to 25 mmol, and then decreases. In UiO-66-NO2-25, the amount of Lewis acidic sites is 0.0061 mmol base per mg, which is about 15 times as high as that of formic acid (0.0004 mmol base per mg), suggesting a profound effect of decanoic acid on the number of Lewis acidic sites. It can be thus concluded that using decanoic acid as modulator is an effective way to enhance the Lewis acidity of UiO-66-NO2. Figure 1. Profile fits to powder X-ray diffraction patterns of UiO-66-NO2-X. The points are measured data, the green line is the final fit, and the red line the difference curve. Refined lattice parameters are provided on each panel. The refinement goodness-of-fit wR = 5.61%, 9.40%, 11.60%, and 14.43 respectively, for X = 15, 20, 25, and 30 in UiO-66-NO2-X. The inset on each plot is an expanded region of the high angle region. Figure 1. Profile fits to powder X-ray diffraction patterns of UiO-66-NO 2 -X. The points are measured data, the green line is the final fit, and the red line the difference curve. Refined lattice parameters are provided on each panel. The refinement goodness-of-fit wR = 5.61%, 9.40%, 11.60%, and 14.43 respectively, for X = 15, 20, 25, and 30 in UiO-66-NO 2 -X. The inset on each plot is an expanded region of the high angle region.
influence on the orientation of clusters as decanoic acid, which may be due to its small steric effect. To investigate the effect of this defective structure further, the method developed by Feng et al. was applied to quantify the accessible Lewis acidic sites in MOFs, and the results are shown in Table 1 [40]. For the sample synthesised without modulator, the amount of Lewis acidic sites is too low to be detected. A significant enhancement of Lewis acidic sites is observed as amount of decanoic acid is increased from 15 mmol to 25 mmol, and then decreases. In UiO-66-NO2-25, the amount of Lewis acidic sites is 0.0061 mmol base per mg, which is about 15 times as high as that of formic acid (0.0004 mmol base per mg), suggesting a profound effect of decanoic acid on the number of Lewis acidic sites. It can be thus concluded that using decanoic acid as modulator is an effective way to enhance the Lewis acidity of UiO-66-NO2. Figure 1. Profile fits to powder X-ray diffraction patterns of UiO-66-NO2-X. The points are measured data, the green line is the final fit, and the red line the difference curve. Refined lattice parameters are provided on each panel. The refinement goodness-of-fit wR = 5.61%, 9.40%, 11.60%, and 14.43 respectively, for X = 15, 20, 25, and 30 in UiO-66-NO2-X. The inset on each plot is an expanded region of the high angle region.  FTIR spectra are shown in Figure 3 and the detailed assignments of absorption bands are listed in Table 2. The peaks in the region of 1400-1600 cm −1 are ascribed to the stretching of C=O, N=O and C=C in the organic ligands, and the anti-phase and in-phase bending of C-H in the benzene ring at 824 cm −1 and 771 cm −1 are also observed. The stretching of Zr-O in Zr-oxo clusters can be seen at 670 cm −1 . A band observed at 1655 cm −1 is attributed to the stretching of carbonyl groups in DMF, revealing some DMF solvent remains in the pores of the framework. The bands at 2930 cm −1 and 2850 cm −1 attributed to methylene groups reveal that some decanoic acid is also present in UiO-66-NO 2 -X. Given the evidence discussed below that shows it is chemically bound, this is likely present as decanoate, coordinated directly at the Zr centres. For other carboxylate modulators in UiO-66, this has proven to be the charge-balancing mechanism [39]. As shown in Figure 3, the intensity of methylene groups exhibits a non-monotonic trend with increasing decanoic acid loading. With the decanoic acid increasing from 15 to 30 mmol, the intensity increases first then decreases, inflecting at 25 mmol. A possible explanation could be that a proportion of channels in UiO-66-NO 2 -X collapse when the amount of the modulator exceeds a certain amount, which inhibits the accumulation of the decanoic acid in MOFs. The result that UiO-66-NO 2 -25 has the highest intensity of decanoic acid agrees with the Lewis acid quantification presented in Table 1. No peak is observed at 1770-1750 cm −1 , which is assigned to the carbonyl group in carboxylic acid, showing that no BDC-NO 2 and decanoic acid in free state exist in the framework, and this result is in agreement with the PXRD analysis results, which show no crystalline free acid.  FTIR spectra are shown in Figure 3 and the detailed assignments of absorption bands are listed in Table 2. The peaks in the region of 1400-1600 cm −1 are ascribed to the stretching of C=O, N=O and C=C in the organic ligands, and the anti-phase and in-phase bending of C-H in the benzene ring at 824 cm −1 and 771 cm −1 are also observed. The stretching of Zr-O in Zr-oxo clusters can be seen at 670 cm −1 . A band observed at 1655 cm −1 is attributed to the stretching of carbonyl groups in DMF, revealing some DMF solvent remains in the pores of the framework. The bands at 2930 cm −1 and 2850 cm −1 attributed to methylene groups reveal that some decanoic acid is also present in UiO-66-NO2-X. Given the evidence discussed below that shows it is chemically bound, this is likely present as decanoate, coordinated directly at the Zr centres. For other carboxylate modulators in UiO-66, this has proven to be the charge-balancing mechanism [39]. As shown in Figure 3, the intensity of methylene groups exhibits a non-monotonic trend with increasing decanoic acid loading. With the decanoic acid increasing from 15 to 30 mmol, the intensity increases first then decreases, inflecting at 25 mmol. A possible explanation could be that a proportion of channels in UiO-66-NO2-X collapse when the amount of the modulator exceeds a certain amount, which inhibits the accumulation of the decanoic acid in MOFs. The result that UiO-66-NO2-25 has the highest intensity of decanoic acid agrees with the Lewis acid quantification presented in Table 1. No peak is observed at 1770-1750 cm −1 , which is assigned to the carbonyl group in carboxylic acid, showing that no BDC-NO2 and decanoic acid in free state exist in the framework, and this result is in agreement with the PXRD analysis results, which show no crystalline free acid.    Nitrogen adsorption-desorption isotherms of all samples were measured at 77 K to elucidate their textural properties. As illustrated in Figure 4, the isotherms of UiO-66-NO 2 -X show Type I patterns. A steep increase of nitrogen adsorption at P/P 0 of 0-0.1 is observed, evidencing a narrow micropore size distribution in UiO-66-NO 2 -X. The textural properties of all samples are shown in Table 3. When no modulator is added, the sample has poor porous properties, with BET surface area and pore volume are 367.52 m 2 /g and 0.136 cm 3 /g, respectively. Introduction of decanoic acid as modulator can improve the porous properties of UiO-66-NO 2 greatly due to the formation of defects, and UiO-66-NO 2 -X exhibit high surface area and pore volume. The effect of modulators has also been reported in literature [41], where the surface area of UiO-66 improved from 439 m 2 /g to 1410 m 2 /g with increasing concentration of modulator. The amount of decanoic acid as modulator has some degree of influence on the textural properties of UiO-66-NO 2 -X but this influence becomes insignificant when the amount of decanoic acid exceeds 25 mmol. The effect of decanoic acid could be explained through two aspects: (1) facilitating the formation of defects to improve porous properties; (2) occupation of decanoic acid to fill the void space and lower porosity. Under the collective influence of these two effects, the UiO-66-NO 2 -20 shows a relatively lower surface area and pore volume compared with other samples. As the amount of decanoic acid is increased from 15 mmol to 20 mmol, the occupation effect becomes significant, and the surface area and pore volume deteriorate. With a further increase in amount of decanoic acid, more defects in MOFs are formed, and the surface area and pore volume improve. The occupation of decanoic acid is suggested by the variation of porous properties of UiO-66-NO 2 -25 after soaking treatment in HCl-DMF mixture for the partial removal of decanoic acid, whereupon the surface area and pore volume increase by nearly 11% and 8%, respectively. Further examination of the existence of decanoic acid molecules and their existing states are investigated and discussed below. Nitrogen adsorption-desorption isotherms of all samples were measured at 77 K to elucidate their textural properties. As illustrated in Figure 4, the isotherms of UiO-66-NO2-X show Type I patterns. A steep increase of nitrogen adsorption at P/P0 of 0-0.1 is observed, evidencing a narrow micropore size distribution in UiO-66-NO2-X. The textural properties of all samples are shown in Table 3. When no modulator is added, the sample has poor porous properties, with BET surface area and pore volume are 367.52 m 2 /g and 0.136 cm 3 /g, respectively. Introduction of decanoic acid as modulator can improve the porous properties of UiO-66-NO2 greatly due to the formation of defects, and UiO-66-NO2-X exhibit high surface area and pore volume. The effect of modulators has also been reported in literature [41], where the surface area of UiO-66 improved from 439 m 2 /g to 1410 m 2 /g with increasing concentration of modulator. The amount of decanoic acid as modulator has some degree of influence on the textural properties of UiO-66-NO2-X but this influence becomes insignificant when the amount of decanoic acid exceeds 25 mmol. The effect of decanoic acid could be explained through two aspects: (1) facilitating the formation of defects to improve porous properties; (2) occupation of decanoic acid to fill the void space and lower porosity. Under the collective influence of these two effects, the UiO-66-NO2-20 shows a relatively lower surface area and pore volume compared with other samples. As the amount of decanoic acid is increased from 15 mmol to 20 mmol, the occupation effect becomes significant, and the surface area and pore volume deteriorate. With a further increase in amount of decanoic acid, more defects in MOFs are formed, and the surface area and pore volume improve. The occupation of decanoic acid is suggested by the variation of porous properties of UiO-66-NO2-25 after soaking treatment in HCl-DMF mixture for the partial removal of decanoic acid, whereupon the surface area and pore volume increase by nearly 11% and 8%, respectively. Further examination of the existence of decanoic acid molecules and their existing states are investigated and discussed below.  Table 3. Analysis of nitrogen adsorption isotherms of UiO-66-NO2-X.

Entry
UiO-66-NO2-X No Modulator X = 15 X = 20 X = 25 X = 30 X = 25 *  Simultaneous TGA-DSC was applied to characterise the thermal stability and chemical composition of UiO-66-NO 2 -X. Figure 5a shows the mass loss curves. The differential scanning calorimetry (DSC) curves are shown in Figure 5b. The differential TG (DTG) curves of UiO-66-NO 2 -X all show a similar trend, and only the curve of UiO-66-NO 2 -25 is given in Figure 5a to highlight the distinctive mass loss events. As temperature is increased from 50 to 150 • C, an approximate 10% mass loss is observed due to the removal of physisorbed water and solvent. A slight loss (about 2%) is attributed to the removal of DMF solvent or the dehydroxylation of the Zr-oxo cluster with the temperature increase from 150 to 300 • C. With temperature increase further up to 600 • C, there exists a continuous mass loss with no clear discontinuities. It is therefore not possible to separate the combustion of linker and combustion of modulator and to assign unambiguously the relative amount of linker, especially bearing in mind variable presence of defects between each sample. Subsequently, ZrO 2 is formed, with a remaining weight of 40% approximately, and no further significant weight loss observed above 600 • C. The final remaining mass of UiO-66-NO 2 -25 is lower than that of other samples which implies a higher decanoic acid percentage. No significant framework decomposition is found below 300 • C, suggesting that UiO-66-NO 2 -X materials show good thermal stability. In Figure 5b, no significant endothermic peak appears at 229 • C for UiO-66-NO 2 -X, at which is the maximum decomposition rate of decanoic acid, showing that no detectable free decanoic acid exists in the MOFs. The maximum decomposition rate of the MOFs prepared with decanoic acid as modulator occurs at 345 • C, and an exothermic peak can be found at the same temperature in Figure 5b, indicating that decanoic acid is combusted with release of heat when it decomposes from the framework. In consideration of the proof from FTIR analysis that methylene groups from decanoic acid do exist in the samples, it can be concluded that the remaining decanoic acid in UiO-66-NO 2 -X is chemically bound in the framework, making its decomposition temperature much higher than that of pure decanoic acid. To test the possibility that decanoic acid is present only in small amounts, TGA-MS analysis was carried out, but this showed no signal related to decanoic acid, validating that none exists in a free state in the UiO-66-NO 2 -X samples. Simultaneous TGA-DSC was applied to characterise the thermal stability and chemical composition of UiO-66-NO2-X. Figure 5a shows the mass loss curves. The differential scanning calorimetry (DSC) curves are shown in Figure 5b. The differential TG (DTG) curves of UiO-66-NO2-X all show a similar trend, and only the curve of UiO-66-NO2-25 is given in Figure 5a to highlight the distinctive mass loss events. As temperature is increased from 50 to 150 °C, an approximate 10% mass loss is observed due to the removal of physisorbed water and solvent. A slight loss (about 2%) is attributed to the removal of DMF solvent or the dehydroxylation of the Zr-oxo cluster with the temperature increase from 150 to 300 °C. With temperature increase further up to 600 °C, there exists a continuous mass loss with no clear discontinuities. It is therefore not possible to separate the combustion of linker and combustion of modulator and to assign unambiguously the relative amount of linker, especially bearing in mind variable presence of defects between each sample. Subsequently, ZrO2 is formed, with a remaining weight of 40% approximately, and no further significant weight loss observed above 600 °C. The final remaining mass of UiO-66-NO2-25 is lower than that of other samples which implies a higher decanoic acid percentage. No significant framework decomposition is found below 300 °C, suggesting that UiO-66-NO2-X materials show good thermal stability. In Figure 5b, no significant endothermic peak appears at 229 °C for UiO-66-NO2-X, at which is the maximum decomposition rate of decanoic acid, showing that no detectable free decanoic acid exists in the MOFs. The maximum decomposition rate of the MOFs prepared with decanoic acid as modulator occurs at 345 °C, and an exothermic peak can be found at the same temperature in Figure 5b, indicating that decanoic acid is combusted with release of heat when it decomposes from the framework. In consideration of the proof from FTIR analysis that methylene groups from decanoic acid do exist in the samples, it can be concluded that the remaining decanoic acid in UiO-66-NO2-X is chemically bound in the framework, making its decomposition temperature much higher than that of pure decanoic acid. To test the possibility that decanoic acid is present only in small amounts, TGA-MS analysis was carried out, but this showed no signal related to decanoic acid, validating that none exists in a free state in the UiO-66-NO2-X samples.   The catalysis results are shown in Figure 6, in which results from oUiO-66-NO 2 synthesised with other modulators (formic acid, butyric acid, benzoic acid, and trifluoroacetic acid) are also given. For fructose dehydration into HMF, conversion of fructose and HMF yield are 80% and 48%, respectively, when no catalyst is added. The reason is that mercaptan is formed for the decomposition of DMSO at high temperature and this can be oxidised into sulfonic acid, an efficient catalyst for fructose dehydration. UiO-66-NO 2 synthesised with different modulators manifests lower catalytic properties, compared with the one prepared without modulator. It is generally accepted that the main source of Brønsted acidity in UiO-66 is µ 3 -OH in the hexameric clusters and this is decisive to the successful conversion of fructose to HMF [37]. As modulators are added during the synthesis of UiO-66, some Brønsted acid sites are lost owing to the missing clusters, leading to the decline of catalytic performance. Surprisingly, this decline is not seen when decanoic acid is used as the modulator, and in fact conversion, yield, and selectivity are increased, revealing that decanoic acid might possess some distinctive effects on conversion of fructose.
For HMF production from glucose, conversion of glucose and HMF yield improve greatly after catalyst addition compared to when no catalyst is used. Among all the samples, UiO-66-NO 2 -25 exhibits the best catalytic properties, with 89.5% conversion of glucose and 44.8% HMF yield, indicating that UiO-66-NO 2 -25 possessed the best combination of Lewis and Brønsted acid sites, facilitating glucose isomerisation and fructose dehydration. According to molecule size calculation, the chain length of decanoic acid is about 13 Å, which is much larger than that of BDC-NO 2 (approximately 6 Å). The coordination of decanoic acid to Zr 4+ might hinder the orientation of Zr-oxo clusters at nearby positions, creating Lewis acidic sites. Moreover, decanoic acid in the framework is likely to increase the hydrophobicity of MOFs, preventing Lewis acidic sites being deactivated by water and increasing the diffusion of water out of the catalyst in favour of glucose conversion into desired product. The hydrophobicity of UiO-66-NO 2 -X is seen qualitatively upon dispersing in water, which is shown in Figure 7. It can be seen that UiO-66-NO 2 -25 floats on the water surface, evidencing a high hydrophobicity. We conclude that it is the synergistic effect of Lewis and Brønsted acidity and hydrophobicity that makes UiO-66-NO 2 -25 an efficient catalyst for HMF production.
The catalysis results are shown in Figure 6, in which results from oUiO-66-NO2 synthesised with other modulators (formic acid, butyric acid, benzoic acid, and trifluoroacetic acid) are also given. For fructose dehydration into HMF, conversion of fructose and HMF yield are 80% and 48%, respectively, when no catalyst is added. The reason is that mercaptan is formed for the decomposition of DMSO at high temperature and this can be oxidised into sulfonic acid, an efficient catalyst for fructose dehydration. UiO-66-NO2 synthesised with different modulators manifests lower catalytic properties, compared with the one prepared without modulator. It is generally accepted that the main source of Brønsted acidity in UiO-66 is μ3-OH in the hexameric clusters and this is decisive to the successful conversion of fructose to HMF [37]. As modulators are added during the synthesis of UiO-66, some Brønsted acid sites are lost owing to the missing clusters, leading to the decline of catalytic performance. Surprisingly, this decline is not seen when decanoic acid is used as the modulator, and in fact conversion, yield, and selectivity are increased, revealing that decanoic acid might possess some distinctive effects on conversion of fructose.
For HMF production from glucose, conversion of glucose and HMF yield improve greatly after catalyst addition compared to when no catalyst is used. Among all the samples, UiO-66-NO2-25 exhibits the best catalytic properties, with 89.5% conversion of glucose and 44.8% HMF yield, indicating that UiO-66-NO2-25 possessed the best combination of Lewis and Brønsted acid sites, facilitating glucose isomerisation and fructose dehydration. According to molecule size calculation, the chain length of decanoic acid is about 13 Å, which is much larger than that of BDC-NO2 (approximately 6 Å). The coordination of decanoic acid to Zr 4+ might hinder the orientation of Zr-oxo clusters at nearby positions, creating Lewis acidic sites. Moreover, decanoic acid in the framework is likely to increase the hydrophobicity of MOFs, preventing Lewis acidic sites being deactivated by water and increasing the diffusion of water out of the catalyst in favour of glucose conversion into desired product. The hydrophobicity of UiO-66-NO2-X is seen qualitatively upon dispersing in water, which is shown in Figure 7. It can be seen that UiO-66-NO2-25 floats on the water surface, evidencing a high hydrophobicity. We conclude that it is the synergistic effect of Lewis and Brønsted acidity and hydrophobicity that makes UiO-66-NO2-25 an efficient catalyst for HMF production.   To further explore the role of decanoic acid remaining in the samples, the c properties of UiO-66-NO2-25 after treatment by soaking in HCl-DMF mixture (0. mL HCl + 20 mL DMF) for glucose conversion were investigated. As shown in F the intensity of methylene groups seen in the FTIR spectrum decreases after s indicating partial removal of decanoic acid from the sample. This can also be va from the enhancement of surface area and pore volume after soaking shown in The decline of catalytic properties after soaking for UiO-66-NO2-25 is shown in F Conversion of glucose and HMF yield drop from 89.5% to 82.8%, 44.5% to respectively, and this decrease is more significant for fructose conversion. It can concluded that decanoic acid remaining in the framework plays an important increasing the catalytic performance for HMF production.  To further explore the role of decanoic acid remaining in the samples, the catalytic properties of UiO-66-NO 2 -25 after treatment by soaking in HCl-DMF mixture (0.6 or 0.8 mL HCl + 20 mL DMF) for glucose conversion were investigated. As shown in Figure 8, the intensity of methylene groups seen in the FTIR spectrum decreases after soaking, indicating partial removal of decanoic acid from the sample. This can also be validated from the enhancement of surface area and pore volume after soaking shown in Table 3. The decline of catalytic properties after soaking for UiO-66-NO 2 -25 is shown in Figure 9. Conversion of glucose and HMF yield drop from 89.5% to 82.8%, 44.5% to 31.3%, respectively, and this decrease is more significant for fructose conversion. It can thus be concluded that decanoic acid remaining in the framework plays an important role in increasing the catalytic performance for HMF production. To further explore the role of decanoic acid remaining in the samples, the catalytic properties of UiO-66-NO2-25 after treatment by soaking in HCl-DMF mixture (0.6 or 0.8 mL HCl + 20 mL DMF) for glucose conversion were investigated. As shown in Figure 8, the intensity of methylene groups seen in the FTIR spectrum decreases after soaking, indicating partial removal of decanoic acid from the sample. This can also be validated from the enhancement of surface area and pore volume after soaking shown in Table 3. The decline of catalytic properties after soaking for UiO-66-NO2-25 is shown in Figure 9. Conversion of glucose and HMF yield drop from 89.5% to 82.8%, 44.5% to 31.3%, respectively, and this decrease is more significant for fructose conversion. It can thus be concluded that decanoic acid remaining in the framework plays an important role in increasing the catalytic performance for HMF production.  UiO-66-NO 2 -X prepared with different amounts of decanoic acid for glucose conversion were also analysed and the reaction results are shown in Figure 10. Conversion of glucose and HMF yield reach the maximum value when the amount of decanoic acid is 20 and 25 mmol, then decreases upon further addition of the modulator. As mentioned above, decanoic acid improves the Lewis acidity and hydrophobicity of MOFs, which favour glucose conversion. Consequently, an enhancement of catalytic performance of UiO-66-NO 2 -X is observed as the amount of modulator increases. However, presence of UiO-66-NO2-X prepared with different amounts of decanoic acid for glucose conversion were also analysed and the reaction results are shown in Figure 10. Conversion of glucose and HMF yield reach the maximum value when the amount of decanoic acid is 20 and 25 mmol, then decreases upon further addition of the modulator. As mentioned above, decanoic acid improves the Lewis acidity and hydrophobicity of MOFs, which favour glucose conversion. Consequently, an enhancement of catalytic performance of UiO-66-NO2-X is observed as the amount of modulator increases. However, presence of further amounts of decanoic acid molecules and imbalance of Lewis and Brønsted acids have a negative impact on glucose conversion.  Figure 11a shows the effect of catalyst dosages on the HMF production from glucose. When 5 mg of catalyst is added, conversion of glucose and HMF yield are 88.8% and 33.9%, respectively, after 4 h, exhibiting the effectiveness of this catalyst. Conversion of glucose and HMF yield improve to 98.1% and 54.5% as the catalyst dosage is increased up to 15 mg. Further increase of catalyst dosage seemingly has no significant promotion  UiO-66-NO2-X prepared with different amounts of decanoic acid conversion were also analysed and the reaction results are shown i Conversion of glucose and HMF yield reach the maximum value when th decanoic acid is 20 and 25 mmol, then decreases upon further addition of th As mentioned above, decanoic acid improves the Lewis acidity and hydr MOFs, which favour glucose conversion. Consequently, an enhancemen performance of UiO-66-NO2-X is observed as the amount of modulat However, presence of further amounts of decanoic acid molecules and Lewis and Brønsted acids have a negative impact on glucose conversion.  Figure 11a shows the effect of catalyst dosages on the HMF prod glucose. When 5 mg of catalyst is added, conversion of glucose and HMF yi and 33.9%, respectively, after 4 h, exhibiting the effectiveness of this catalys of glucose and HMF yield improve to 98.1% and 54.5% as the catalyst dosag up to 15 mg. Further increase of catalyst dosage seemingly has no significa on HMF formation, indicating that higher catalyst dosage might give reactions. Therefore, the catalyst dosage was fixed at 15 mg for subsequen obtain optimum HMF yield.  Figure 11a shows the effect of catalyst dosages on the HMF production from glucose. When 5 mg of catalyst is added, conversion of glucose and HMF yield are 88.8% and 33.9%, respectively, after 4 h, exhibiting the effectiveness of this catalyst. Conversion of glucose and HMF yield improve to 98.1% and 54.5% as the catalyst dosage is increased up to 15 mg. Further increase of catalyst dosage seemingly has no significant promotion on HMF formation, indicating that higher catalyst dosage might give rise to side reactions. Therefore, the catalyst dosage was fixed at 15 mg for subsequent reactions to obtain optimum HMF yield.
As the glucose isomerisation and fructose dehydration are endothermic, temperature has an obvious influence on HMF yield. In order to optimise HMF yield, the process of glucose conversion was monitored as a function of time at three temperatures, which is shown in Figure 11b. A maximum 54.5% HMF yield is achieved at 140 • C after 4 h. Higher temperature (150 • C) can cause HMF decomposition after a longer time, leading to the decline of HMF yield. Moreover, HMF formation proceeds slowly at 130 • C, and HMF yield is 48.0% after 6 h. Furthermore, it can be found that further prolonging of reaction time over optimised conditions decreases HMF yield, which may be due to the decomposition of HMF. The optimal reaction conditions are as follows: 15 mg of catalyst and 27.5 mg glucose at 140 • C for 4 h in 1.25 mL DMSO.
As the glucose isomerisation and fructose dehydration are endothermic, temperature has an obvious influence on HMF yield. In order to optimise HMF yield, the process of glucose conversion was monitored as a function of time at three temperatures, which is shown in Figure 11b. A maximum 54.5% HMF yield is achieved at 140 °C after 4 h. Higher temperature (150 °C) can cause HMF decomposition after a longer time, leading to the decline of HMF yield. Moreover, HMF formation proceeds slowly at 130 °C, and HMF yield is 48.0% after 6 h. Furthermore, it can be found that further prolonging of reaction time over optimised conditions decreases HMF yield, which may be due to the decomposition of HMF. The optimal reaction conditions are as follows: 15 mg of catalyst and 27.5 mg glucose at 140 °C for 4 h in 1.25 mL DMSO. The recyclability of catalyst is important from the perspective of sustainable chemistry and economic cost. Recycling experiments were carried out to investigate the reusability of catalyst with the reaction results shown in Figure 12. After four cycles, glucose conversion declines from about 90% to 50%, and HMF yield from 50% to 20%. The reason for this decline is likely due to the accumulation of humins on the exterior and interior surface of the catalyst, deactivating some active sites and hindering the diffusion of glucose. However, the catalytic selectivity does not show a significant deterioration, suggesting the active sites in the samples maintain their catalytic properties for glucose conversion into HMF instead of other side products.  The recyclability of catalyst is important from the perspective of sustainable chemistry and economic cost. Recycling experiments were carried out to investigate the reusability of catalyst with the reaction results shown in Figure 12. After four cycles, glucose conversion declines from about 90% to 50%, and HMF yield from 50% to 20%. The reason for this decline is likely due to the accumulation of humins on the exterior and interior surface of the catalyst, deactivating some active sites and hindering the diffusion of glucose. However, the catalytic selectivity does not show a significant deterioration, suggesting the active sites in the samples maintain their catalytic properties for glucose conversion into HMF instead of other side products.
which is shown in Figure 11b. A maximum 54.5% HMF yield is achieved at 140 °C afte h. Higher temperature (150 °C) can cause HMF decomposition after a longer tim leading to the decline of HMF yield. Moreover, HMF formation proceeds slowly at °C, and HMF yield is 48.0% after 6 h. Furthermore, it can be found that furt prolonging of reaction time over optimised conditions decreases HMF yield, which m be due to the decomposition of HMF. The optimal reaction conditions are as follows: mg of catalyst and 27.5 mg glucose at 140 °C for 4 h in 1.25 mL DMSO. The recyclability of catalyst is important from the perspective of sustaina chemistry and economic cost. Recycling experiments were carried out to investigate reusability of catalyst with the reaction results shown in Figure 12. After four cyc glucose conversion declines from about 90% to 50%, and HMF yield from 50% to 20 The reason for this decline is likely due to the accumulation of humins on the exterior a interior surface of the catalyst, deactivating some active sites and hindering the diffus of glucose. However, the catalytic selectivity does not show a significant deteriorati suggesting the active sites in the samples maintain their catalytic properties for gluc conversion into HMF instead of other side products.  To investigate the influence of solvent, the glucose conversion was studied in various solvents and mixed solvents, and the results are summarised in Table 4. This shows always a lower HMF yield compared with that in DMSO. As 10% water is added in DMSO, HMF yield declines from 44.8% to 24.3%. HMF yield does not show improvements when MIBK is used to extract HMF from water in a biphasic system. These results suggest the deactivation effect of water on UiO-66-NO 2 -25. HMF yield in DMF is 4.8%, meaning that DMF can deteriorate the catalytic performance significantly, possibly due to its basic characteristics. Water/2-propanol(v/v = 1/9) 4 -3.9 --Not analysed. Table 5 summarises some heterogeneous catalysts previously reported in the literature for glucose conversion to HMF. It can be seen that the UiO-66-NO 2 synthesised with decanoic acid as modulator exhibits desirable catalytic properties with high conversion of glucose and HMF yield in DMSO. Only a few studies have reported higher HMF yield [32,42], but these other high performing catalysts are composite materials with complex synthesis procedures. UiO-66-NO 2 -25 sythesised in our work is a desirable catalyst for glucose conversion, which has a low synthesis cost, high catalytic performance, and outstanding porous properties. For comparsion, the sulfonated MOF MIL-101(Cr) shows very poor performance in DMSO [43]. Some materials manifest much better catalytic performance in aqueous systems compared with the catalyst in our work, yet they are generally different classes of materials, inorganic frameworks or mesoporous silica [44][45][46]. An exception is phosphate-modified NU-1000, where a zirconium-based MOF is modified by post-synthetic treatment with phosphoric acid [22]. This gives an outstanding catalytic performance, but the mass fraction of glucose is as low as 0.018%, which is much lower than that applied in our work (2%). However, for this material HMF yield declines greatly to about 20%, when the glucose concentration increases to 1.8%, and recylability was not invesigated.

Preparation of UiO-66-NO 2
UiO-66-NO 2 was synthesised as follows: 1 mmol ZrCl 4 , 0.5 mmol BDC-NO 2 , and specified moles of decanoic acid were dissolved completely in 18 mL DMF. The mixture was then transferred into a PTFE-lined autoclave and heated at 120 • C for 24 h. The resulting white solid was filtered, washed with DMF twice, and soaked in DMF at 90 • C for 12 h. Finally, the solid was washed with DMF and acetone twice for each solvent, and the obtained UiO-66-NO 2 was dried at 150 • C overnight. Samples of UiO-66-NO 2 with different amounts of decanoic acid are labeled as UiO-66-NO 2 -X, where X represents the number of moles (mmol) of added decanoic acid. Four samples were synthesised with X of 15, 20, 25, and 30. To investigate the effect of decanoic acid, the sample with no modulator added was also prepared for comparison.

Materials Characterisation
The physical and chemical properties of the as-synthesised UiO-66-NO 2 -X were characterised by various techniques. Powder X-ray diffraction (PXRD) analysis was carried out using a Panalytical Empyrean diffractometer (Malvern Panalytical, Malvern, UK) operating with Cu Kα 1/2 radiation to investigate the crystal structure of UiO-66-NO 2 -X. The software GSAS-II was used to perform profile fitting of the diffraction patterns using the Pawley method [47]. Surface area, pore volume, and pore diameter distribution were analysed with N 2 adsorption-desorption experiment at 77 K using a Micromeritics ASAP 2020 apparatus (Micromeritics, Tewkesbury, UK). A Nicolet Nexus 470 Fourier transform infrared (FTIR) spectrometer (Thermo Fisher Scientific, Waltham, MA, USA.) was used to identify the functional groups in the samples. Thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) were conducted using a Mettler Toledo TGA/DSC instrument (Mettler-Toledo, Leicester, UK) under an air flow up to 600 • C with a heating rate of 10 • C/min to investigate thermal stability of the samples.

Catalytic Reactions and Analysis Technique
The catalytic reactions for HMF production from glucose were performed in an oil bath. Glucose (27.5 mg) was added to a 4 mL thick-walled glass vial (Kinesis, Eaton Socon, UK) containing 1.25 mL DMSO, and the mixture was stirred continuously during heating with a magnetic bar to ensure homogeneous conditions. After the chosen reaction time, the mixture was diluted with deionised water and analysed by using high performance liquid chromatography (HPLC).
Glucose and HMF were quantified using an HPLC instrument (Shimadzu, Milton Keynes, UK) equipped with an Aminex HPX-87H 300 mm × 7.8 mm column (Bio-Rad Laboratories, Hercules, CA, USA), UV, and ELSD detectors. The specific parameters are as follows: column temperature of 50 • C, water as the mobile phase at flow rate of 0.4 mL/min, UV detection wavelength at 284 nm, and injection volume of 10 µL. Conversion of glucose and HMF yield were obtained based on standard calibration curves that were fitted from known concentrations of glucose and HMF standard solutions. Conversion of glucose, HMF yield, and reaction selectivity towards HMF are calculated as follows: Coversion glucose (mol%) = moles of glucose reacted moles of starting glucose × 100% (1) Yield HMF (mol%) = moles of HMF produced moles of starting glucose × 100% Seletivity HMF (mol%) = HMF yield glucose conversion × 100% (3)

Quantification of Accessible Lewis Acidic Sites
The accessible Lewis acidic sites were quantified as follows according to literature [40]: about 20 mg of MOFs was treated with 0.5 mL pivalonitrile in toluene for 24 h at room temperature. The resultant solid was centrifuged and washed with toluene thoroughly to remove excess pivalonitrile, then washed with acetone to remove toluene. The final material was dried and digested in NH 4 F/D 2 O. The mixture was analysed by 1 H NMR, in which 1, 4-dioxane was used an internal standard. Then the quantity of pivalonitrile (m pivalonitrile ) in the sample was calculated as follows: where I pivalonitrile is the integral of the pivalonitrile peak, N pivalonitrile is the number of protons corresponding to the pivalonitrile, M pivalonitrile is the molar mass of pivalonitrile, and m standard is the known mass of the standard in the sample.

Recyclability of Catalysts
For recyclability experiments, the catalysts were recovered from the reaction mixture by centrifugation after the completion of the reaction. The recovered catalysts were washed thoroughly with water and acetone, respectively, followed by drying at 150 • C. Subsequently, the dried catalysts were used for four consecutive cycles. To eliminate the influence of a little amount of catalyst loss on recyclability experiments, the first run was duplicated for six times, ensuring that the amount of the recovered catalyst is enough for the subsequent reactions.

Conclusions
UiO-66 MOFs were prepared with 2-nitroterephthalate as linker and decanoic acid as modulator. The decanoic acid present in the MOF structure is chemically bound and influences the porous and catalytic properties, including surface area and pore volume, Lewis acidity, and hydrophobicity. The samples exhibit efficient catalytic performance for conversion of glucose to HMF. Under optimised reaction conditions, conversion of glucose and HMF yield can reach up to 98.1% and 54.5%, respectively. During the process of glucose conversion, humins formed as the side product can deteriorate the catalytic conversion of samples, yet the reaction selectivity towards HMF is almost maintained. Future work must consider further quantification of acidity, both Lewis and Brønsted, to understand more fully the mechanism of catalysis.