The usage of fossil fuels causes serious problems like the energy crisis and global warming. In order to solve these problems, so far much attention has been paid to the development of renewable energies such as solar or wind energy. Biofuel produced from biomass is one of the potential alternatives. First-generation biofuels (i.e. biodiesel) produced from corn and soybean oil have proved that biomass-to-biofuel conversion is possible; however, the use of edible agriculture as a source will cause other problems such as food deficiency.¹  Therefore, second-generation biofuels generated from non-edible lignocellulosic biomass have attracted more attention recently.

Lignocellulosic (or so-called ‘wood-based’) biomass consists of three major components: cellulose (41%), hemicellulose (28%), and lignin (27%).²  Generally, cellulose and hemicellulose can be used to produce bioethanol, and lignin offers a broad spectrum of conversion (thermal cracking, fast pyrolysis, and complete gasification) to achieve valuable chemicals and transportation fuels.³  So far, a great deal of effort has been put toward the degradation of cellulose with enzymes,⁴  mineral acids,⁵  bases,⁶  and supercritical water.⁷  The enzymatic hydrolysis of cellulose is effective, but the system is sensitive to contaminants originating from other biomass components. Furthermore, pre-treatment of cellulose (e.g., ammonia or steam treatments in a high-pressure process or mechanical milling) is typically required to increase the accessible area of cellulose for a reasonable rate of enzymatic hydrolysis.⁸  Mineral acids have been extensively investigated to catalyze hydrolysis at a variety of acid concentrations and temperatures. A rather high temperature (180–230 °C) has been used in order to obtain an acceptable rate of cellulose hydrolysis. Furthermore, degradation of the resulting glucose becomes an issue at such high temperatures.⁹ 

Recently, ionic liquids (ILs) have attracted a lot of attention and have been utilized as solvents for the degradation of the lignocellulosic biomass.^10–14  The importance of ionic liquids in cellulose dissolution has been emphasized in several reviews.^15–17  ILs are a kind of novel green solvent. They are organic salts with relatively low melting points. In other words, ILs usually appear as crystals under normal conditions; however, they can be melted and dissociated into two ionic parts at relatively high temperatures (usually less than 100 °C). In contrast to other crystalline salts (e.g. NaCl), the attractive characteristic of ILs is that they can transform into a liquid phase.

The utilization of ILs for the dissolution of lignocellulose started in early 2000. Numerous papers have been published on controlling the viscosity and polarity of ionic liquids by varying their ionic structures.^16,17  The main focus of these papers was the solubility of the synthesized ILs toward different carbohydrates such as glucose, sucrose, amylose, cellulose, and so on. In 2002, Rogers et al. reported that cellulose could be dissolved in ILs at 100 °C.¹⁰  The solubility of cellulose in ionic liquids results from its anions. It can disrupt the hydrogen bonds between polysaccharide chains of cellulose and then dissolve it.¹⁸  This discovery started a new pathway to deal with cellulose at low temperatures and ambient pressure.

In 2007, Zhang and his co-workers discovered that CrCl₂ in [EMIM]Cl (1-ethyl-3-methylimidazolium chloride, an imidazolium-type ionic liquid) can efficiently catalyze the glucose-to-HMF conversion.¹⁹  HMF is a promising platform chemical because it can further transform to a widely used biofuel called 2,5-dimethylfuran (DMF)²⁰  and other useful materials.²¹  Since then, many have worked on the production of HMF from cellulose or glucose in ionic liquid systems. Binder and Raines combined HCl, CrCl₂ or CrCl₃, DMA/LiCl and [EMIM]Cl to convert cellulose to HMF;¹³  Zhang and his co-workers used CrCl₂/CuCl₂ as catalysts in [EMIM]Cl;¹⁴  Han and his co-workers also discovered SnCl₄ in [EMIM]BF₄ can convert glucose to HMF with a high yield;²²  Riisager and his co-workers discussed HMF produced from lanthanide-containing ionic liquid systems;²³  Bell and Chidambaram discovered that 12-molybdophosphoric acid in [EMIM]Cl/acetonitrile or [BMIM]Cl/acetonitrile can selectively convert glucose to HMF.²⁴  Although there has been much research focused on the addition of various kinds of catalysts in ionic liquid systems, very few papers discussed the effects of reaction conditions (such as dissolution temperatures and times of ILs, reaction temperatures and times, and the amounts of water) on the conversion efficiency in ionic liquids without additional catalysts.²⁵  In fact, in the above-mentioned papers, HMF could still be produced when using ILs only (no other additives), although the yields were very low. This indicates that ILs in these systems serve not only as solvents but also as catalysts. We suggest that the low HMF yield was because the reaction conditions for HMF production in these cases were not optimized. For example, Zhao et al. has shown that the yield of HMF converted from fructose was greatly affected by the reaction temperature in an [EMIM]Cl only system.¹⁴  Very recently, Binder and Raines discussed the sequence and timing of the addition of water into the cellulosic conversion and showed that an optimal sequence and timing strongly affected the conversion efficiency.²⁶ 

In recent decades, cellulase was broadly studied for the hydrolysis of cellulose.^27–31  Cellulase is a mixture of enzymes containing three main components: (1) endo-1,4-beta-d-glucanase (EG) which randomly cleaves the cellulose chain to lower the crystallinity; (2) cellobiohydrolase (CBH) which degrades cellulose by releasing cellobiose units; (3) beta-glucosidase which hydrolyzes cellobiose and other oligomers to get glucose units. To date, the reaction conditions and the hydrolytic processes of hydrolyzing cellulose by using free cellulase have been optimized with a glucose yield as high as 70%.²⁹  However, one critical problem when using cellulase as a catalyst is the easy deactivation of cellulase by environmental factors (e.g., temperature), which greatly hinder its practical use in industry.^32,33  In order to overcome such difficulties, the immobilization of cellulase onto solid materials is a feasible way to enhance its stability.

Many research papers have demonstrated that immobilizing cellulase onto organic and inorganic materials could improve the stability and reusability of cellulase without reducing its catalytic ability.^34–41  Among the host materials, mesoporous silica materials have gained much attention because of their large specific surface area, high mechanical strength, and tunable surface functionality.^34,35  Recently, Sakaguchi et al. studied the encapsulation of cellulase by using mesoporous silica SBA-15 with various pore sizes as hosts.³⁶  They found that the enzymatic activity of cellulase strongly depended on the pore size of SBA-15. The best performance of cellulase could be obtained when using SBA-15 with pore diameter around 8.9 nm. However, the structure of SBA-15 is 2D hexagonal with length of several μm, which would inhibit the adsorption of cellulase into the inner surface of the SBA-15 and result in a low adsorption amount. Lu et al. studied the effect of surface functionalities of a mesoporous silica FDU-12 (pore size is around 25.4 nm) on the immobilization of cellulase.³⁵  They functionalized FDU-12 with phenyl, thiol, amino and vinyl groups. The results showed that the electrostatic and hydrophobic interactions between cellulase and functionalized FDU-12 play significant roles on the activity and stability of immobilized cellulase. Amine-functionalized FDU-12 adsorbed the largest amount of cellulase but exhibited the lowest activity. They explained this was due to the interaction between amine groups of FDU-12 and the carboxyl groups of catalytic site of cellulase which thereby inhibited the activity of cellulase. In contrast, vinyl-functionalized FDU-12 not only maintained the activity of cellulase up to 80% but also temporal enzyme stability owing to the existence of hydrophobic groups. Despite these pioneering studies, none of them has studied different immobilization methods (i.e., physical adsorption and chemical binding) on the efficiency of cellulase, and cellulosic hydrolysis by immobilized cellulase has never been reported yet.

5-Hydroxymethylfurfural (HMF), converted from lignocellulosic biomass, is considered one of the “top value-added chemicals”; this results from its utilization as a building-block platform between biomass and promising chemical intermediates, such as 2,5-furandicarboxylic acid (FDCA),⁴²  2,5-dimethylfuran (DMF),⁴³  5-ethoxymethylfurfural (EMF),⁴⁴  and ethyl levulinate (EL),⁴⁵  which have been studied extensively in recent years and demonstrate the significance of HMF.

HMF has been successfully generated from fructose, glucose, and cellulose using various kinds of reaction systems with homogeneous or heterogeneous catalysts.^46–48  The mechanism of cellulose-to-HMF conversion is still unclear, but the conversion can be divided into several reactions. First, cellulose is usually pre-treated by alkaline, acid, or certain ionic solutions to destroy its rigid framework. The pre-treated cellulose then goes through the depolymerization process in an acidic system in order to break the 1,4-β-glycosidic bonds of cellulose and produce glucose. Subsequently, glucose converts to fructose via isomerization, which is a so-called Lobry de Bruyn–Alberda van Ekenstein transformation.⁴⁹  Finally, the dehydration of fructose generates HMF. The mechanism of fructose-to-HMF conversion has been discussed in numerous studies.^50–52  The micro-kinetic model for this three-stage water-removed process⁵³  has been constructed to determine an apparent activation energy.⁵⁴  In addition, according to computational results, both the estimated equilibrium constant and activation energy can be greatly influenced by reaction conditions, including temperature, solvent, and catalysts.⁵⁵ 

Different solvents have been used in the fructose-to-HMF conversion because of the contrast between water-soluble reactants (e.g., fructose and glucose) and organic-solvent-soluble products (e.g., HMF). The careful selection of solvents can promote the preferential reaction and enhance product yield. The Dumesic group studied the effects of solvents on the dehydration of fructose in biphasic systems, and demonstrated the catalytic ability of dimethylsulfoxide (DMSO), which is able to suppress the undesired side reactions effectively.⁵⁵  Recently, ILs have been widely used as both catalysts and solvents for producing HMF from lignocelluloses because of their comparatively higher catalytic activity and adjustable composition.⁵⁶  However, despite the excellent activity and recyclability of ILs, their potential is restricted to laboratory-scale experiments due to high costs. Therefore, a low-price solvent with the desired properties (e.g., high boiling point and low viscosity) such as DMSO can have more potential in industrial applications.

In recent years, several groups have reported the production of HMF from fructose in DMSO-based reaction systems via homogeneous and heterogeneous catalysts, including acids, salts, and metal ions.⁵⁷  The Dumesic group has investigated the catalytic capabilities of various homogeneous mineral acids.⁵⁸  Recently, Wang et al. used carbon-based p-toluenesulfonic acid (TsOH) at 130 °C for 1.5 h resulting in a 91.2% yield of HMF.⁵⁹  Although these pioneering studies showed high yields of HMF, harsher reaction conditions are always needed in such homogeneous catalytic systems. From economic and sustainable viewpoints, scientists have turned to heterogeneous solid catalysts and mild reaction conditions. For example, the Sidhpuria group immobilized ILs onto silica particles as an efficient heterogeneous catalyst for fructose-to-HMF conversion with a yield of 63% in a DMSO system.⁶⁰ 

Mesoporous silica nanoparticles (MSNs) have attracted a great deal of attention in the field of catalysis because of their high surface area and controllable pore size. In addition, abundant SiOH groups on the surface of MSNs provide the possibility of further functionalization with other organic groups.⁶¹  For example, the Lin group has used a co-condensation method to functionalize MSNs with a general acid (i.e., a ureidopropyl (UDP) group) and a base (i.e., a 3-[2-(2-aminoethylamino)ethylamino]-propyl (AEP) group) as a cooperative acid–base catalyst for aldol, Henry, and cyanosilylation reactions.⁶²  We also have used a grafting method to functionalize MSNs with several metal–histidine complexes for H₂O₂-assisted tooth bleaching. However, the conventional MSNs synthesized from the cationic surfactant cetyltrimethylammonium bromide (CTAB) exhibit a pore size of around 2 nm. For several catalytic reactions involving large molecules (e.g., proteins or cellulose), this pore size is too small to allow the reactants to diffuse into the mesopores, thus losing the advantage of high surface area inside the MSNs. Therefore, the synthesis of MSNs with pore sizes large than 10 nm is highly desirable.

The MSNs with large pore sizes can be synthesized through two approaches: (1) using high-molecular-weight surfactants as templates; and (2) adding hydrophobic additives as swelling agents. For example, the Zhao group succeeded in synthesising mesoporous silica with an ultra-large pore size of approximately 37.0 nm by using a high-molecular-weight surfactant (poly(ethylene oxide)-b-poly(methyl methacrylate); PEO-b-PMMA).⁶³  The same group has also reported the addition of 1,3,5-trimethylbenzene (TMB) as a swelling agent to synthesize mesoporous silica with a large pore of 25.4 nm (as denoted as FDU-12).^63,64  The Lu group has functionalized FDU-12 materials with phenyl, thiol, amino, and vinyl groups and studied the effect of these functional groups on the immobilization efficacy of an enzyme (i.e., cellulase).⁶⁵  Despite these pioneering studies, researchers have not yet utilized large-pored MSNs with various functional groups for cellulosic conversion in ionic liquid systems. In general, cellulosic conversion involves three main reactions: (1) cellobiose-to-glucose depolymerization, (2) glucose-to-fructose isomerization, (3) fructose-to-HMF dehydration. These reactions need acid, base, and acid catalysts, respectively, as illustrated in Scheme 1.1. Consequently, to synthesize large-pored MSNs with both acid and base functionalities as a new cooperative solid catalyst would be helpful for one-pot cellulose-to-HMF conversion.

Optimal reaction conditions with respect to temperature, the amount of catalyst and the reaction time are crucial to maximising the final yield of cellulosic hydrolysis when using cellulase as a catalyst. Therefore, we first optimized the reaction temperatures, the amount of free cellulase and the reaction time. As shown in Figure 1.1a, 15 mg of cellulose was hydrolyzed using free cellulase (50 Unit, 1 Unit indicates the amount of enzyme that can catalyze 1×10⁻⁶ mole substrate in one minute) as the catalyst at different temperatures for 24 h; the maximum yield of glucose was obtained at 50 °C. By repeating the experiment, we found that 50 °C is also the most stable operating condition for cellulose. Thus, we chose 50 °C as the suitable temperature for cellulase-assisted cellulose hydrolysis. From the economic point of view, an optimal amount of cellulase means the minimum amount of cellulase while keeping the maximum yield of glucose. Various amounts of cellulase, ranging from 1 mg to 23 mg, were used for the hydrolysis of cellulose (15 mg) at 50 °C for 24 h. Based on the results in Figure 1.1b, we found that the optimal amount of celluase was 25 Unit (i.e., 4.5 mg). More cellulase than 25 Unit did not increase the yield of glucose at the current operation coniditions. After obtaining the optimal reaction temperature and the amount of cellulase, the optimal reaction time was also examined. 15 mg of cellulose was hydrolyzed using free cellulase (25 Unit) at 50 °C for various reaction times (i.e., ranging from 3 to 48 hours). According to the results shown in Figure 1.1c, in order to reach 90% glucose yield, cellulose has to be hydrolyzed for at least 24 h although 80% glucose yield could be obtained in 12 h. For consistency, here we chose 24 hours as the optimal reaction time.

The morphology and porous properties of the synthesized MSNs with two different pore sizes (NB these are referred to as large-pore MSNs (LPMSNs) and small-pore MSNs (SPMSNs) hereafter) are characterized with scanning electron microscopy (SEM) and nitrogen adsorption–desorption isotherms. The SEM images in Figure 1.2a and b show the uniform and spherical morphology for both LPMSNs and SPMSNs with diameters around 600 and 150 nm, respectively. In Figure 1.2c and d, LPMSNs and SPMSNs exhibit type III and type IV nitrogen adsorption–desorption isotherms, respectively. The type III isotherm of LPMSNs exhibits prominent adsorption at high relative pressures (P/P₀), which is indicative of adsorption in macropores. In contrast, the type IV isotherm of SPMSNs has been widely shown to occur in a typical mesoporous material with a two-dimensional hexagonal structure. The Brunauer–Emmett–Teller (BET) specific surface areas for LPMSNs and SPMSNs are 262.6 and 820.1 m² g⁻¹, respectively. In addition, the pore-size distribution calculated from the Barrett–Joyner–Halenda (BJH) method clearly shows that LPMSNs exhibit a broad pore size around 20–40 nm while SPMSNs exhibit a narrow pore size around 2–5 nm, as depicted in Figure 1.2e and f, respectively. The structural properties of LPMSNs and SPMSNs are summarized in Table 1.1.

In addition to pore size and surface area, the surface functionality of the MSN also affects the immobilization of cellulase. We qualitatively and quantitatively study the functional groups on the SPMSNs and LPMSNs by ²⁹Si and ¹³C solid-state nuclear magnetic resonance (NMR). Because the synthetic methods for SPMSNs and LPMSNs are different, the organic functional groups and their amounts on the surface of the prepared materials are different. As shown in the ²⁹Si NMR spectra (Figure 1.3a), LPMSNs exhibited Q³, Q⁴, T³ and T² bonds, indicating that there are Si–O–H and Si–O–C bonds in the material. Because we added 3-aminopropyltrimethoxysilane (APTMS) during the synthesis of LPMSN, the Si–O–C bonds should be due to the presence of APTMS. On the other hand, SPMSN exhibited only Q³ and Q⁴ peaks that represent the presence of Si–O–Si and Si–O–H bonds, as shown in Figure 1.3b. These two materials were used as hosts for physical adsorption of cellulase. For chemically binding cellulase, we further functionalize LPMSN with carboxyl groups by reacting LPMSN with an organosilane 3-triethoxysilylpropyl succinic acid anhydride (TESP-SA). Based on the ¹³C NMR spectra of LPMSN and TESP-SA-functionalized LPMSN in Figure 1.3c and d, respectively, we could conclude that the TESP-SA was successfully grafted onto the surface of LPMSN. The TESP-SA-functionalized LPMSN should exhibit two functional groups, i.e., the amino group from APTMS and the carboxyl group from TESP-SA, on its surfaces. The amounts of the functional groups on the SPMSN, LPMSN, and TESP-SA-functionalized LPMSN were calculated and summarized in Table 1.1.

Cellulase was immobilized into SPMSN and LPMSN by physical adsorption and into TESP-SA-functionalized LPMSN by chemical binding. The amounts of cellulase immobilized into these materials were quantitatively measured by UV-Vis spectroscopy. As shown in Table 1.2, on the basis of the same amount of host materials (i.e., 50 mg), the amounts of the immobilized cellulase are 14.6 mg, 23.4 mg, and 19.2 mg for SPMSN, LPMSN, and TESP-SA-functionalized LPMSN, respectively. The highest amount of immobilized cellulase was found in the case of cellulase-adsorbed LPMSN, which was 1.6 times the amount in cellulase-adsorbed SPMSN. Although the SPMSN exhibits a higher surface area than LPMSN, the small pore size of SPMSN (i.e., 2–5 nm) made the diffusion of cellulase into the pore difficult due to the large size of cellulase (around 8 nm). Therefore, a pore size larger than 8 nm in MSNs is essential for the immobilization of cellulase. In addition, although the amount of cellulase chemically bonded with TESP-SA-functionalized LPMSN was less than that of cellulase-adsorbed LPMSN, the amount was still larger than 4.5 mg which had been considered to be the minimum amount for maximum glucose production under optimal reaction conditions.

Another factor affecting the immobilization of cellulase is surface charge. The surface charges of cellulase and SPMSN at pH=4.8 are both negative (−6.7 and −14.8 mV, respectively). Therefore, the electrostatic interaction between SPMSN and cellulase was negligible. The surface charge of LPMSN is around zero (+1.0 mV) because of the existence of both Si–OH and Si–NH₂ groups. Therefore, in addition to the larger pore size of LPMSN, the increased adsorption amount in the case of cellulase-adsorbed LPMSN also resulted from the electrostatic interaction between cellulase and Si–NH₂ groups of LPMSN. It is worth noting that the surface charge of TESP-SA-functionalized LPMSN was negative (−40.5 mV) owing to the existence of carboxylic acid groups in TESP-SA. Therefore the immobilized cellulase amount in this case was less than that of cellulase-adsorbed LPMSN. However, we have confirmed that the immobilized cellulases here were covalently linked with TESP-SA-functionalized LPMSN by ¹³C NMR, and such chemically linked cellulases could avoid the problem of cellulase detachment, resulting in excellent stability (see discussion below).

As we have previously examined, 4.5 mg of free cellulase was enough to convert 15 mg of cellulose to glucose with a high yield of around 85% at 50 °C for 24 hours. However, although the amounts of cellulase immobilized in the MSN materials are all larger than 4.5 mg, their glucose yields were all less than 85% at the same optimal reaction conditions. As shown in Figure 1.4 and Table 1.2, the glucose yields for cellulase-adsorbed SPMSN, cellulase-adsorbed LPMSN, and cellulase-linked TESP-SA LPMSN were 33.30%, 77.89%, and 83.79%, respectively. It is clearly seen that the cellulase chemically bonded with TESP-SA functionalized LPMSN exhibited almost the same activity with free cellulase. For the case of cellulase-adsorbed SPMSN, the glucose yield was the smallest, which might be due to the formation of a byproduct (i.e., cellobiose with yield of 46.14%). Because the cellulase-adsorbed LPMSN catalyst still exhibited a respectable glucose yield of 77.89%, we suggest that the low glucose yield for cellulase-adsorbed SPMSN was due to the small pore size. When the pore size of MSN was small, not only could cellulase not be adsorbed in the pores but also the pre-treated cellulose could not diffuse into the pores. This result again proved the significance of suitable pore size when using MSN materials as hosts.

One may be concerned that the chemically linked cellulase will lose its activity due to the change of its conformation via covalent binding. The Lu group has utilized amino-group-functionalized mesoporous silica materials to immobilize cellulase and found that the activity of the immobilized cellulase decreased although the amount of immobilized material was large, as compared to free cellulase.³⁵  They suggested that the amino group of the materials would bind to the catalytic domains of the cellulase, thereby reducing its catalytic ability. In contrast to their result, our data indicated that the activity of chemically linked cellulase was similar with that of free cellulase. We conclude that the carboxylic-group-functionalized LPMSN used in this study would bind with the cellulose binding domains (not catalytic domains) of the cellulase, thereby retaining the activity of cellulase. In fact, several papers have also reported the preservation/enhancement of cellulolytic activity by connecting the cellulose binding domains (CBD) of cellulase with scaffolds.⁶⁶ 

One of the advantages of immobilizing cellulase within porous materials is to increase the stability of cellulase. To examine the stability of the immobilized cellulase, this study tests different cellulase-immobilized materials, including free cellulase, cellulase-adsorbed SPMSN, cellulase-adsorbed LPMSN and cellulase-linked TESP-SA-functionalized LPMSN by storing these catalysts at room temperature (usually cellulase should be stored at 4 °C). After 23 days, these materials were used to hydrolyze cellulose, and the results are refered as “After_storage” in contrast to “Before_storage”, which involved catalysts before stability experiments. Figure 1.4 and Table 1.2 show that after 23 days the glucose yields for free cellulase, cellulase-adsorbed SPMSN, cellulase-adsorbed LPMSN, and cellulase-linked TESP-SA-functionalized LPMSN all decreased, yielding 53.39%, 4.59%, 10.64%, and 82.15%, respectively. The corresponding percentage loss of glucose yields for free cellulase and cellulase-immobilized catalysts were 37.78%, 83.30%, 86.56% and 4.61%, respectively.

The stability experiments in this study reveal three important findings. (i) The SPMSN and LPMSN could not protect cellulase when the cellulase was merely immobilized by physical adsorption. The large percentage loss (over 80%) of glucose yield in these two cases indicated that cellulase easily detached from MSN at room temperature, resulting in deactivation. (ii) In addition to glucose, cellobiose was formed in the cases of free cellulase and cellulase-adsorbed SPMSN and LPMSN. This result indicates that the activity of cellobiohydrolase, an enzyme in cellulase that hydrolyzes disaccharides into individual monosaccharides, decreased after storage at room temperature for 23 days. (iii) The cellulase chemically linked to TESP-SA-functionalized LPMSN exhibited the best stability. This indicates that the chemical bonding between cellulase and TESP-SA-functionalized LPMSN decreased its hydrolytic activity while preserving the catalytic specificity toward cellulose-to-glucose conversion, which is the best way to immobilize cellulase with a stable efficiency.

The synthetic process of the bi-functionalized MSN is shown in Scheme 1.2 and is described as follows: Brij-97 was used as the template and was first dissolved in 180 g of deionized water. Then, APTMS and DOP were added to the Brij-97 solution with stirring at room temperature. After stirring for 30 min, organosilanes (i.e., MPTMS and CPTES) were added to the reaction system along with TEOS, and the whole system, with a composition (in molar ratios) of water: Brij-97 : TEOS : MPTMS : CPTES=433 : 0.293 : 1 : 0.009 : 0.009, was prepared and kept stirred for 24 h at room temperature. The mixture was subsequently heated at 100 °C overnight. Finally, the precipitated solid was collected by filtration and washed sequentially with water and methanol. It is worth noting that the template can be extracted by this washing step. To further convert the thiol group of (MP+CP)-MSN to a sulfonic group, the (MP+CP)-MSN was oxidized in an H₂O₂ solution with a composition of (MP+CP)-MSN : H₂O : MeOH : H₂O₂=0.5 g : 10 ml : 10 ml : 10 ml. After reaction at room temperature overnight, the obtained sample (i.e., (HSO₃+CP)-MSN) was washed and dried in vacuum. To further functionalize the (HSO₃+CP)-MSN with ionic liquid, the (HSO₃+CP)-MSN sample and solid imidazole were degassed for 3 h before the addition of anhydrous benzene and chlorobutane with the molar ratio of imidazole : chlorobutane=1 : 2. After one day of reflux, the product (i.e., (HSO₃+ILs)-MSN) was collected through filtration and washed with anhydrous benzene. Finally, the (HSO₃+ILs)-MSN sample was immersed in CrCl₂ solution, and the whole mixture was kept stirred overnight. CrCl₂ can be physically absorbed on the surface of (HSO₃+IL)-MSN. The final product ([HSO₃+(ILs/CrCl₂)]-MSN) was collected by centrifugation.

The morphology and porous properties of the series of synthesized bi-functionalized MSN were characterized with SEM and nitrogen adsorption–desorption isotherms. The SEM image shows an uniform and spherical morphology for bi-functionalized MSN with particle size of sub-microns. In addition, the bi-functionalized MSNs exhibit a type IV nitrogen adsorption–desorption isotherm, indicating multilayer adsorption by capillary condensation. Moreover, the results of BET specific surface area and pore-size distribution calculated from the BJH method are listed in Table 1.3 together with other structural properties of bi-functionalized MSNs.

Next, we qualitatively and quantitatively investigated the functional groups on the bi-functionalized MSN using ²⁹Si and ¹³C solid-state NMR. The ¹³C NMR spectrum contains ten identified signals, and these results evidenced the successful grafting of the organosilane MPTMS and ILs on the MSN. Additionally, we further quantified the amounts of each functional group on the MSN by ²⁹Si NMR. There were Qⁿ and Tⁿ peaks, which represents the relative amount of silica framework and their covalent bonding with organosilanes. The amounts of functional groups are summarized in Table 1.3. In addition, results of element analysis indicated the presence of the sulfonic acid (S, around 2.3%) and ionic liquid (N, around 1.7%).

In order to demonstrate the effect of bi-functionalized MSN on the fructose-to-HMF conversion, the reaction was executed without catalysts (blank sample), with MSN and bi-functionalized MSN. The result including the efficacy of fructose conversion, HMF yield, and selectivity is depicted in Figure 1.5a. The cases labeled ‘blank’ (no catalyst) and ‘MSN’ (with non-functionalized MSN as the catalyst) showed low conversion (around 25%) and almost no HMF yield, indicating that the dehydration of fructose to generate HMF was difficult to achieve in such reaction conditions (i.e., DMSO solvent, 90 °C, 3 h). In contrast, with the same reaction conditions, the reaction with the presence of the bi-functionalized MSN surprisingly exhibited enhanced fructose conversion (almost 100%) and HMF yields as high as 72.5%, as shown in Figure 1.5a. These results clearly demonstrate the effectiveness of the bi-functionalized MSN on the catalytic production of HMF, which was due to the contributions of the functional groups of R–HSO₃ acid and [EMIM]Cl/CrCl₂ ionic liquid. It has been reported that the dehydration of fructose can be promoted with the assistance of different homogeneous acids and metal chlorides.⁶¹  Here we report a successful functionalization of both sulfonic acid and ionic liquid/metal chloride onto the surface of MSN materials as a new method to provide efficient heterogeneous catalysts. In addition to the sulfonic acid groups, chloride ions can act as an effective catalyst due to their nucleophilicity, and the acidic C-2 proton of the imidazolium part of the ionic liquid also promotes the dehydration of fructose.⁵²  Furthermore, in contrast to other solid nanoparticles, MSN exhibits a high surface area and large mesopores, which should enhance the efficiency greatly, owing to the increased number of reaction sites.

Reusability of the [HSO₃+(ILs/CrCl₂)]-MSN was further studied over five cycles. As shown in Figure 1.5b, it is seen that the conversions of fructose were maintained at almost 98% until the fifth run. Additionally, no significant loss of HMF yield was observed. It means that the grafted functional groups (i.e. sulfonic acid (HSO₃) and ILs) did not leach during the complicated reiterating process and further hold their original activity without obvious decay. Therefore, the recyclability of the synthesized bi-functionalized material has been exactly confirmed.

We further study the kinetics of the fructose-to-HMF conversion and compare the rate constants, reaction orders, and activation energies for the systems with and without bi-functionalized MSN catalysts. First, we conducted two reaction systems (i.e., with and without catalyst) at the same reaction temperature (i.e., 90 °C) but for different reaction time periods. We observed the variation of HMF yields with reaction time and the data is shown as Figure 1.6a. The reaction rates of both cases are initially higher but tend to be almost constant after a time-span, indicating that the reactants have already run out, resulting in no more HMF being produced. However, adding our synthesized catalysts, [HSO₃+(ILs/CrCl₂)]-MSN, obviously accelerates the generation of HMF to some extent. Three hours later, there is about 73.4% of HMF, which is nearly three times larger than the percentage in blank (25.3%). That is to say, our material has an effect on the catalytic fructose conversion.

For the purpose of identifying the relationships between kinetic parameters (such as rate constants, reaction orders, and activation energies) and the addition of the bi-functionalized catalyst, we further comprehensively constructed the kinetics profiles at different temperatures for systems with and without [HSO₃+(ILs/CrCl₂)]-MSN (Figure 1.6a). We supposed that the ultimate product of fructose dehydration is HMF only, i.e. we did not take other by-products into account. Besides, we assumed the degradation of HMF would not occur under our designed condition (90 °C, 0–6 h). This hypothesis can be supported by the test results of HMF decomposition, as shown in Figure 1.6b. In this experiment, HMF is considered as the reactant, placed in an identical environment as before. From the data shown, we can see that the apparent decay of HMF (15.18%) could be noticed only under sufficient reaction temperature and time (140 °C, 6 h), which is very far from our practical operating conditions. That is to say, in our milder conditions (90 °C, 3 h), the decomposition of HMF is negligible and this outcome bolsters the previous model we have set up.

In Figure 1.7, we analyzed the kinetics profiles in order to systematically understand the shifts of each kinetics parameter caused by the catalysts. Referring to the previous published research,⁵⁴  we assumed that the transformation of fructose is a 1st order process and that the reaction rate could be expressed as follows:

where t is the reaction time and C is an arbitrary constant. Therefore, we plotted a figure with −ln(1−X) as the y-axis and t as the x-axis, fitting the data linearly, and evaluated reaction constants from the slopes. As shown in Figure 1.7, there is an obvious increase of k in the presence of bi-functionalized MSNs, confirming their ability to promote this reaction. Next, we calculated the activation energy (E_a) of each system from rate constants we obtained by the Arrhenius equation. The E_a values of systems with and without catalysts are 67.5 and 80.05 kJ mol⁻¹, respectively. This fact indicated that our addition of [HSO₃+(ILs/CrCl₂)]-MSN has altered the reaction route to a certain degree, consequently lowering the activation energy and leading to a higher reaction rate.

To functionalize LPMSN with acid and base groups, the organosilane, i.e., MPTMS and APTMS, was grafted onto the surface of LPMSN. Typically, 1 g of LPMSN in a two-necked round-bottom flask was degassed in vacuum at 110 °C for 3 h. After that, dried toluene (40 mL) was injected into the flask under nitrogen atmosphere, followed by injecting organosilanes. The amount of organosilane used was 1.3 times the amount of silanol group on the LPMSN that was previously calculated by solid-state NMR (i.e., 6 mmol g⁻¹). Then, the mixture was heated and refluxed at 110 °C for 24 h. Finally, the acid- and/or base-functionalized LPMSN was collected by filtration, washed with toluene several times in order to remove the residual reactant, and dried in vacuum. The resulting samples were called LPMSN-NH₂ and LPMSN-SH. Then, the LPMSN-SH was oxided to become LPMSN-SO₃H by modifying a published procedure.⁶⁷  Typically, 0.5 g of LPMSN-SH was added to the mixture of hydrogen peroxide (10 mL), deionized water (10 mL), and methanol (10 mL). The mixture was stirred at room temperature for 12 h. After that, the resulting precipitate was collected by filtration, washed with deionized water several times and dried in vacuum. The resulting sample was named LPMSN-SO₃H. LPMSN-SO₃H and LPMSN-NH₂ are used as acid and base solid catalysts, respectively, in this study. For preparation of bi-functionalized LPMSN exhibiting both acid and base groups (denoted as LPMSN-Both), APTMS was grafted onto the pre-synthesized LPMSN-SO₃H using the same grafting process described above.

The cellulosic conversion includes pre-treatment and reaction. For pre-treatment, cellulose (15 mg) was added into [EMIM]Cl (150 μL), and the whole mixture was heated at 120 °C for 0.5 h with stirring for dissolution of cellulose. For reaction, LPMSN-based catalysts (4 mg) and water (16.67 μL) were added to the cellulose/[EMIM]Cl solution while keeping heating at 120 °C for another 3 h. All reactions were repeated three times and the average yields of products were obtained. After the optimization of the reaction conditions beforehand, the amount of catalyst was determined as 4 mg. The conversion process for cellobiose, glucose and fructose was the same as that for cellulose except for exclusion of the pre-treatment step. When using glucose and fructose as reactants, water did not be added into the reaction systems.

Before functionalizing LPMSN with other functional groups, we qualitatively and quantitatively investigated the amounts of hydroxyl group in the LPMSN using ²⁹Si and ¹³C solid-state NMR. As shown in the ²⁹Si NMR spectra, LPMSN exhibited Q³, Q⁴, T³ and T² bonds, indicating that there are Si–O–H and Si–O–C bonds in the material. Since 3-aminopropyltrimethoxysilane (APTMS) was added during the synthesis of LPMSN, the Si–O–C bonds should be due to the presence of APTMS. For functionalization of LPMSN with acid (SO₃H) or base (NH₂) groups, the as-synthesized LPMSN was further reacted with an organosilane 3-(mercaptopropyl)trimethoxysilane (MPTMS) or APTMS, respectively. As shown in the ¹³C NMR spectra of LPMSN-SO₃H (Figure 1.8a), three peaks at 11, 18 and 54 ppm correspond to the carbons of the Si–CH₂–CH₂–CH₂–SO₃H from left to right, respectively, indicating the appearance of the acid functionality. On the other hand, there are three distinct peaks at approximately 11, 22 and 42 ppm representing the carbons of the Si–CH₂–CH₂–CH₂–NH₂ from left to right, respectively,⁶⁸  as depicted in the ¹³C NMR spectra of LPMSN-NH₂, proving the existence of the base functionality (Figure 1.8b). The peak of around 71 ppm indicates the existence of Brij97 residue.

Since the as-synthesized LPMSNs also exhibit base functionality, we quantified the amounts of functional groups for all four samples (i.e., LPMSN, LPMSN-SO₃H, LPMSN-NH₂ and LPMSN-Both) by ²⁹Si NMR in order to distinguish the degree of different functionality. After deconvolution of ²⁹Si NMR peaks, we calculated the amounts of hydroxyl group and functional group on the surface of each sample. As summarized in Table 1.4, on the basis of the functionality of LPMSN (i.e., 1.06 mmol g⁻¹), LPMSN-Both exhibited the highest amount of functional groups (i.e., 2.32 mmol g⁻¹), indicating the successful addition of both acid and base groups. Although LPMSN contains base groups (from APTMS during synthesis), the amount of its functional group is less than LPMSN-NH₂ that was further grafted with APTMS. In addition, the surface areas and pore sizes of LPMSN-SO₃H, LPMSN-NH₂ and LPMSN-Both decreased as compared to those of LPMSN, indicating a pore filling effect upon functionalization (Table 1.4).⁶⁹ 

The acidity of each LPMSN catalyst was estimated using following process. Samples were added into properly chosen indicator solution, and the color change of the solution was observed. As summarized in Table 1.4, the sequence of acidity (i.e., lower pK_a value) is LPMSN-SO₃H>LPMSN>LPMSN-Both>LPMSN-NH₂, indicating that LPMSN-SO₃H is the strongest acid catalyst and LPMSN-NH₂ is the strongest base catalyst. Although LPMSN also contains NH₂ groups, the silanol groups on the surface of LPMSN would provide acidity, giving a weak acid property in total.

Fructose (15 mg) and LPMSN-based catalysts (4 mg) were added into [EMIM]Cl (150 μL), and the whole mixture was heated at 120 °C for 3 h. As shown in Figure 1.9a, the yields of HMF converted from fructose with the presence of four LPMSN-based catalysts were similar with each other (in the range around 66–70%). It has been considered that an acid catalyst is necessary for the dehydration of fructose to produce HMF. However, we suggested that fructose could be easily converted into 5-HMF in the ILs system with a high temperature (120 °C) because such conditions (ILs and high temperature) favor dehydration. Therefore, there is no difference between all LPMSN-based catalysts. Other groups have also reported similar findings.

The yields of HMF converted from glucose with the presence of four LPMSN-based catalysts are shown in Figure 1.9b. It can be clearly seen that the cases with LPMSN-Both and LPMSN-NH₂ catalysts exhibited the highest yield around 13%, in contrast to that of LPMSN-SO₃H (ca. 10%) and LPMSN (ca. 7%). The conversion of glucose to HMF involves two steps: isomerization of glucose to fructose and dehydration of fructose to HMF. In general, base catalysts are helpful for isomerization of glucose to fructose.⁵¹  From the result of fructose-to-HMF conversion, we have confirmed that four different LPMSN catalysts had a similar effect on the dehydration of fructose to HMF; in other words, HMF can be easily converted from fructose in our reaction system. Therefore, the high HMF yields for the cases using LPMSN-Both and LPMSN-NH₂ indicate that base catalysts indeed promote the production of fructose converted from glucose, and the glucose-to-fructose conversion can be regarded as the rate-determining step in the glucose-to-HMF conversion.

To stimulate the structure of pre-hydrolyzed cellulose, we used cellobiose as the reactant and studied its conversion with the presence of LPMSN-based catalysts. The yields of glucose and HMF converted from cellobiose are shown in Figure 1.9c. It can be seen that the LPMSN-SO₃H exhibited the highest yields of both glucose and HMF (25.6 and 18.9%, respectively). The cellobiose-to-HMF conversion contains three steps: hydrolysis of cellobiose-to-glucose, isomerization of glucose-to-fructose and dehydration of fructose to HMF. Because acid catalysts can facilitate the first step, we suggest that it is the reason why the highest yield of glucose appeared in the case of LPMSN-SO₃H. In addition, it can also be proposed that the cellobiose-to-glucose is the rate-determining step of the three reactions. Therefore, catalysts with stronger acidity (i.e., LPMSN-SO₃H and LPMSN) would favor the production of glucose that was then converted to HMF, resulting high yields of glucose and HMF.

We further directly used pre-hydrolyzed cellulose as the reactant and performed the cellulose-to-HMF conversion in ionic liquid system with the presentence of four LPMSN-based catalysts. As shown in Figure 1.9d, the high yields of glucose and HMF were found in the cases of LPMSN-SO₃H (35.8 and 19.2%, respectively) and LPMSN-Both (36.3 and 14.7%, respectively). It is predictable that LPMSN-SO₃H showed the highest efficiency toward cellulosic conversion because it exhibits the strongest acidity that can facilitate the hydrolysis of cellulose. In fact, several groups have also synthesized SO₃H-functionalized ILs for effective cellulosic conversion.⁷⁰  However, it is surprising to us that the LPMSN-Both also provided high yields of glucose and HMF, even its acidity is less than LPMSN-SO₃H and LPMSN.

Since the cellulosic conversion involves a series of complicated reactions that need different acid and base catalysts in each step, we propose that the enhanced efficacy of LPMSN-Both could be attributed to the cooperative catalysis of both acid and base functional groups in the LPMSN-Both. In order to prove our hypothesis, we used a mixture of LPMSN-SO₃H and LPMSN-NH₂ (1 : 1 in weight ratio) as the catalyst for the cellulose-to-HMF conversion at the same reaction conditions. The reaction efficacy of the mixed catalyst was compared with that of LPMSN-Both. As shown in Figure 1.10, the yields of cellobiose, glucose, and HMF for the case of mixed catalyst were all similar to those for the LPMSN-Both case, indicating that the acid-and-base mixed catalyst exhibited the same efficacy as acid- and base-containing catalysts. The results above indeed prove that the unusual catalytic enhancement is a strong indication of the existence of cooperation between the acid (SO₃H) and base (NH₂) groups in our LPMSN-Both system.

In this chapter, we have demonstrated the successful synthesis of multi-functionalized MSNs as effective, reliable, and re-usable solid catalysts for cellulosic biomass conversion. In the enzyme-assisted catalytic system, we optimized the reaction conditions for cellulase-immobilized solid catalysts in cellulosic hydrolysis. For the first time, carboxyl-group-functionalized MSNs with large pore size of 40 nm were synthesized and used to chemically link cellulase. The proposed cellulase-assisted biocatalyst exhibits a high cellulose-to-glucose conversion efficiency (over 80%) with outstanding stability. In the chemical-assisted catalytic system, we demonstrated the synthesis of MSNs with both acid and ionic liquid groups. Such bi-functionalized MSN solid catalysts have enhanced the production of HMF from fructose dehydration in mild conditions using DMSO as a solvent. The kinetics study has indicated that our bi-functionalized MSN could accelerate fructose dehydration by reducing the activation energy required. In addition, we also demonstrated the synthesis of large-pored mesoporous silica nanoparticles (LPMSN) and functionalization of LPMSN with acid, base, and both acid and base groups. The functionalized LPMSN-based catalysts have showed enhanced catalytic efficacy toward cellulosic conversion including fructose-to-HMF dehydration, glucose-to-fructose isomerization, and cellobiose-to-glucose hydrolysis. The bi-functionalized LPMSN enhanced the yields of glucose and HMF directed converted from cellulose, indicating the cooperative catalytic ability. We envisage that the multi-functionalized LPMSN materials could serve as new selective catalysts for other important reactions.