The listing or discussion of a prior-published document in this specification should not necessarily be taken as an acknowledgement that the document is part of the state of the art or is common general knowledge.

Biomass, particularly non-edible lignocellulosic biomass, is one of the most abundant natural materials and it is widely available around the world at a relatively low cost (S.H. Zhu, et al., Catal. Sci. Technol., 2015, 5, 3845-3858). The three main components of biomass are normally: 40-50% cellulose; 20^10% hemicellulose; and 20-30% lignin. As a vital renewable alternative to fossil fuel, the conversion of lignocellulosic biomass into renewable fuels and chemicals is undoubtedly an ideal route for biomass utilisation and has attracted much attention. As will be appreciated, materials obtained may be chemicals that may be useful for their inherent properties, or because they are intermediates or starting materials for valuable chemical products (e.g. pharmaceuticals).

Although there are conventional biological processes that can be used to convert biomass to useful materials (typically the fermentation of biomass-based carbohydrate), a chemocatalytic process provides opportunities to increase the process efficiency because the catalyst and the operation conditions can be readily engineered and optimised.

Given this opportunity, various routes for the catalytic transformation of biomass and/or biomass-derived feedstock have been explored. Amongst which, the transformation of cellulose/hexose into valuable molecules in a one-pot approach is one of the most attractive routes for biomass utilisation (e.g. see S. Biella, et al., J. Catal., 2002, 206, 242-247; Y. Onal, et al., J. Catal., 2004, 223, 122-133; X. S. Tan, et al., Chem. Commun., 2009, 7179-7181 ; C. Baatz and U. PruBe, J. Catal., 2007, 249, 34-40). However, in order to directly transform biomass into valuable downstream chemicals in a one-pot process, multi-functional catalytic systems capable of catalysing each reaction are needed. As will be appreciated, a one-pot process is advantageous as it reduces or eliminates isolation and/or purification steps that may otherwise have been necessary between chemical reactions. To date, a number of catalysts have been developed for one-pot biomass transformation (N. Yan, et al., J. Am. Chem. Soc. 2006, 128, 8714-8715; P. Wang, et al., Biores. Technol. 2011 , 102, 4179-4183). However, most of them are either very expensive and/or are not efficient. Since the one-pot degradation of biomass involves a series of complicated cascade reactions, the careful design and functionalisation of the catalyst is necessary to obtain a high yield and selectivity for the desired downstream products.

In addition, most catalysts developed to date are: not recyclable; easily deactivated after a few reaction cycles; or corrosive to the reaction apparatus. For example, homogeneous catalysts like H2SO4 and NaOH are very corrosive chemicals that can damage the reaction apparatus, so they are not particularly desirable catalysts even though they exhibit excellent catalytic activity in one or more of the reactions in the desired cascade. Further, after the reactions have completed on an industrial scale, tons of water would be required to neutralise these chemicals, which is not environmentally-friendly. Heterogeneous catalysts, on the other hand, are known to be recyclable. However, their catalytic activity is usually far lower than that of homogeneous catalysts and some of them are easily deactivated after recycling.

Given the above, there remains a need for new materials for use as catalysts for converting biomass feedstock to useful downstream chemicals. These materials have to be easy to synthesise, recyclable and environmentally-friendly. More importantly, they have to be able to demonstrate good selectivity and efficiency in the catalytic reaction and can degrade a wide range of feedstocks, ranging from simple sugars to complex carbohydrates.

The discovery and development of graphene quantum dots (GQDs) as nano-sized sp ² - hybridised carbon nanosheets have led to its versatile applications in biology, energy, photovoltaics, and catalysis (S. Bak, et al., Current Applied Physics, 2016, 16, 1 192-1201 ). Importantly, GQDs are cheap, abundantly available and environmentally-friendly.

Summary of Invention The current invention relates to graphene quantum dot materials that have been surprisingly found to be particularly useful in the catalytic conversion of biomass to higher-value chemicals (e.g. 5-hydroxymethylfurfural and 2,5-diformylfuran). The yield and the selectivity of the target product are significantly improved by using functionalised graphene quantum dots (GQDs), an inexpensive carbon-based material. The GQDs exhibit remarkable catalytic efficiency (as high as homogeneous catalysts) and can be recycled and reused without obvious loss of catalytic activity.

Thus, in a first aspect of the invention, there is provided a nanoparticulate material comprising: graphene quantum dots having a first and second surface, each surface comprising functional groups selected from one or more of the group consisting of O-based functional groups, N-based functional groups, S-based functional groups, P-based functional groups, and B-based functional groups; and

metal-containing nanoparticles attached to the first or second surface of one or more graphene quantum dots, wherein:

each graphene quantum dot is a nanosheet having from one to five layers of graphene and a lateral diameter of from 1 nm to 1 ,000 nm; and

the functional groups on the first and second surfaces of the graphene quantum dots account for from 0.1 to 60 wt% of the total weight of the nanoparticulate material.

In embodiments of the first aspect of the invention:

(Aa) the metal-containing nanoparticles may be selected from one or more of the group consisting of metal sulfide nanoparticles, metal phosphide nanoparticles or, more particularly, metal nanoparticles and metal oxide nanoparticles, for example:

(aA) the metal nanoparticles may be selected from one or more of the group consisting of Ru or, more particularly, Ni, Au, Pt, and alloys thereof;

(bA) the metal oxide nanoparticles may be selected from one or more of the group consisting of M0O3 and V ₂ 0 ₅ nanoparticles;

(cA) the metal sulfide nanoparticles may be M0S2;

(dA) the metal phosphide nanoparticles may be selected from the group consisting

(Ab) the metal in the metal-containing nanoparticles may account for from 20 to 60 wt% of the total weight of the nanoparticulate material, and/or the functional groups on the first and second surfaces of the graphene quantum dots may account for from 0.1 to 60 wt%, such as from 0.15 to 50 wt%, of the total weight of the nanoparticulate material (e.g. the metal in the metal-containing nanoparticles may account for from 30 to 50 wt% of the total weight of the nanoparticulate material, and/or the functional groups on the first and second surfaces of the graphene quantum dots may account for from 0.2 to 50 wt%, such as from 0.2 to 2 wt%, such as from 0.2 to 1 wt%, of the total weight of the nanoparticulate material); (Ac) the metal-containing nanoparticles may have an average diameter of from 1 to 1 ,000 nm, such as from 50 to 500 nm, such as 100 to 200 nm.

In particular embodiments of the first aspect of the invention that may be mentioned herein, the metal-containing nanoparticles may be M0O3 nanoparticles. It will be appreciated that such nanoparticles may, in certain embodiments, have the features of (Ab) and (Ac) above.

In a second aspect of the invention, there is also provided a nanoparticulate material comprising:

graphene quantum dots having a first and second surface, each surface comprising functional groups selected from one or more of the group consisting of O-based functional groups, N-based functional groups, S-based functional groups, P-based functional groups, and B-based functional groups, wherein:

each graphene quantum dot is a nanosheet having from one to five layers of graphene and a lateral diameter of from 1 nm to 1 ,000 nm; and

the functional groups on the first and second surfaces of the graphene quantum dots account for from 0.1 to 80 wt% of the total weight of the nanoparticulate material.

In embodiments of the second aspect of the invention, when the nanoparticulate material does not comprise metal-containing nanoparticles, the functional groups on the first and second surfaces of the graphene quantum dots may account for from 0.1 to 80 wt%, such as from 1 to 10 wt%, such as from 3 to 8 wt%, of the total weight of the nanoparticulate material.

In further embodiments of the second aspect of the invention, when the nanoparticulate material does not comprise metal-containing nanoparticles, the nanoparticulate material may comprise about 1 .05 mmol/g of weak acid sites (<270°C), about 2.31 mmol/g of medium acid sites (from 270 to 400°C) and about 0.22 mmol/g of strong acid sites (>400°C), as measured by temperature programmed desorption of ammonia. As will be appreciated, the first and second aspects of the invention are related to one another and so in embodiments of the first and second aspects of the invention:

(i) the O-based functional groups are selected from one or more of the group consisting of hydroxyl groups, carboxylic acid groups, ketone groups, aldehyde groups, ester groups, carbonate ester groups and epoxy groups;

(ii) the N-based functional groups are selected from one or more of the group consisting of amino groups, amide groups, nitro groups, and pyridinyl groups; (iii) the S-based functional groups are selected from one or more of the group consisting of thiol groups, sulphide groups, disulphide groups and sulfonic acid groups;

(iv) the P-based functional groups are selected from one or more of the group consisting of phosphate groups and phosphonic acid groups;

(v) the B-based functional groups are selected from one or more of the group consisting of boronic acid groups, borinic groups and borinate ester groups;

(vi) the functional groups on the first and second surfaces of the graphene quantum dots are selected from the group comprising of hydroxyl, carboxylic acid, sulfonic acid, and combinations thereof;

(vii) each graphene quantum dot has an average lateral diameter less than or equal to 40 nm, such as from 1 to 40 nm, such as from 10 to less than 40 nm;

(viii) the nanoparticulate material has a BET surface area of from 50 to 500 m ² /g, such as 90 to 350 m ² /g, such as from 100 to 300 m ² /g;

(ix) each graphene quantum dot is a nanosheet having from one to two layers of graphene; (x) the nanoparticulate material comprises from 0.1 mmol/g to 10 mmol/g of acidic sites, such as from 2 mmol/g to 8 mmol/g as measured by temperature programmed desorption of ammonia;

(xi) the nanoparticulate material forms agglomerates that have an average diameter of from 1 nm to 1 ,000 nm, such as from 15 nm to 500 nm, such as from 25 to 100 nm, such as from 50 to 75 nm.

Unless otherwise stated, all embodiments of the first or second aspects of the invention may be combined with the features of the first or second aspect of the invention, respectively, whether individually or in any combination (provided said combination is not excluded due to technical incompatibility).

As noted above, the materials described above may be useful in the conversion of biomass into useful chemicals. Thus, in a third aspect of the invention, there is provided a method of converting a biomass feedstock to 2,5-diformylfuran, the method comprising the step of contacting the biomass feedstock with a catalytic amount of a nanoparticulate material as defined in the first or second aspect of the invention, or any technically sensible combination of their respective embodiments.

In embodiments of the third aspect of the invention:

(Ba) the feedstock may be selected from one or more of the group selected from xylose, cellulose, or, more particularly, sucrose, fructose, glucose, cellobiose, and inulin (e.g. the feedstock may be fructose and/or cellulose); (Bb) the catalytic amount of the nanoparticulate material may be from 0.1 to 40 wt% of the weight of the feedstock, such as from 0.5 to 20 wt%, such as from 1 to 10 wt%, such as 2 wt%;

(Be) the nanoparticulate material may be as defined in the first aspect of the invention and any technically sensible combination of its embodiments;

(Bd) when the nanoparticulate material is as defined in the second aspect of the invention and any technically sensible combination of its embodiments, when the nanoparticulate material does not comprise metal-containing nanoparticles, the process may further comprise a step of oxidising 5-hydroxymethylfurfural to 2,5-diformylfuran.

In a fourth aspect of the invention, there is provided a method of converting a biomass feedstock to 5-hydroxymethylfurfural, the method comprising the step of contacting the biomass feedstock with a catalytic amount of a nanoparticulate material as defined in the second aspect of the invention and any technically sensible combination of its embodiments, where the nanoparticulate material does not comprise metal-containing nanoparticles.

In embodiments of the fourth aspect of the invention:

(Ba) the feedstock may be selected from one or more of the group selected from xylose, cellulose, or, more particularly, sucrose, fructose, glucose, cellobiose, and inulin (e.g. the feedstock may be fructose and/or cellulose);

(Bb) the catalytic amount of the nanoparticulate material may be from 0.1 to 40 wt% of the weight of the feedstock, such as from 0.5 to 20 wt%, such as from 1 to 10 wt%, such as 2 wt%.

Brief Description of Drawings

Fig. 1 Depicts a schematic transformation of fructose 10 to 5-HMF 30 and DFF 50, using SGQDs 60 and Mo0 ₃ /SGQDs 70 respectively.

Fig. 2 Depicts the synthesis of SGQDs and M0O3/SGQDS of the current invention.

Fig. 3 Depicts the (a) FTIR spectra of the as-synthesised SGQDs and Mo0 ₃ /SGQDs of the current invention in comparison with GQDs; (b) FTIR spectra (wider range) of SGQDs in comparison with GQDs; and (c) Raman spectra of SGQDs and GQDs. Fig. 4 Depicts the TEM images of (a) SGQDs and (b) Mo0 ₃ /SGQDs of the current invention; (c) TEM images of SGQDs with corresponding EDS mapping for C, O and S elements (top), and for S element only (bottom); (d) TEM-EDS line scanning with EDS spectra for C, O, S elemental distribution along the line; (e) AFM image of SGQDs, with the inset showing the height profile along the indicated line; (f) diameter distribution of SGQDs (n = 110); and (e) height distribution of SGQDs (n = 152). Fig. 5 Depicts the XPS characterisation of SGQDs of the current invention in comparison with GQDs: (a) full XPS spectrum of SGQDs of the current invention in comparison to that of GQDs; (b) deconvoluted C 1 s peaks of SGQDs and GQDs; (c) deconvoluted O 1 s peak of SGQDs and GQDs; and (d) S 2p peak of SGQDs. Fig. 6 Depicts the pyridine-adsorption FTIR spectra (for acidity characterisation) of SGQDs of the current invention in comparison to that of GQDs.

Fig. 7 Depicts the effects of (a) solvent compositions, and (b) initial fructose concentrations on the dehydration of fructose to 5-HMF, using the SGQDs of the current invention. The product yields are referred to the left y-axis, while the fructose conversion rates are referred to the right y-axis.

Fig. 8 Depicts the correlation between (a) SO3H density and 5-HMF yield; and (b) total acidity and fructose conversion for the SGQDs of the current invention in comparison with various heterogeneous catalysts.

Fig. 9 Depicts (a) the recycling process for the SGQDs of the current invention; and (b-c) the performance of the recycled SGQDs in dehydrating fructose to 5-HMF. Fig. 10 Depicts the (a) FTIR spectra; and (b) pyridine-adsorption FTIR spectra of the spent SGQDs of the current invention which had been used for a total of eight times (i.e. re-used seven times after initial use).

Fig. 11 Depicts an (a) TEM image of the spent SGQDs after five times of recycling; and b) TEM-EDS line scanning with EDS spectra for C, O, S elemental distribution along the line.

Description

It has been discovered that graphene quantum dots can provide a catalytic efficiency as high as homogeneous catalysts due to their quasi-homogeneity, which allows them to be homogeneously dispersed in aqueous solution. In addition, it has also been found that such graphene quantum dots can also mimic heterogeneous catalysts, as they can be recycled and reused after reaction. In addition, it has also been advantageously found that graphene quantum dots can be functionalised to endow them with multiple functions for specific purposes, as disclosed in more detail herein. Disclosed herein are both graphene quantum dots that have been functionalised to carry suitable functional groups, and graphene quantum dots that have been combined with metal- containing nanoparticles, which may provide added catalytic functionality.

Thus, there is disclosed a nanoparticulate material comprising:

graphene quantum dots having a first and second surface, each surface comprising functional groups selected from one or more of the group consisting of O-based functional groups, N-based functional groups, S-based functional groups, P-based functional groups, and B-based functional groups, wherein:

each graphene quantum dot is a nanosheet having from one to five layers of graphene and a lateral diameter of from 1 nm to 1 ,000 nm; and

the functional groups on the first and second surfaces of the graphene quantum dots account for from 0.1 to 80 wt% of the total weight of the nanoparticulate material.

In embodiments of the above, the nanoparticulate material may be substantially free of metals, though it will be appreciated that the nanoparticulate material may contain some metals as an impurity. In the context of the current invention, "substantially free of X" is intended to mean that no more than 1 wt% (e.g. less than 0.5 wt%, less than 0.1 wt%, such as less than 0.001 wt%, such as less than 0.0001 wt%) of substance X is present. In other aspects and embodiments of the invention, the nanoparticulate material may actually contain nanoparticles that contain a metal. As such, there is also provided a nanoparticulate material comprising:

graphene quantum dots having a first and second surface, each surface comprising functional groups selected from one or more of the group consisting of O-based functional groups, N-based functional groups, S-based functional groups, P-based functional groups, and B-based functional groups; and

metal-containing nanoparticles attached to the first or second surface of one or more graphene quantum dots, wherein:

each graphene quantum dot is a nanosheet having from one to five layers of graphene and a lateral diameter of from 1 nm to 1 ,000 nm; and

the functional groups on the first and second surfaces of the graphene quantum dots account for from 0.1 to 60 wt% of the total weight of the nanoparticulate material. When used herein, the tem "graphene quantum dot" takes its ordinary meaning in the art, with the proviso that there may only be from one to five of layers of graphene in the quantum dot. As will be appreciated, the actual number of layers in each of the graphene quantum dots in a batch of the nanoparticulate material may vary, such that some dots may only contain one layer, while others may contain five layers. In particular embodiments of the invention that may be mentioned herein, the graphene quantum dots may have only one or two layers of graphene. As such, while the "thickness" as measured through an axis perpendicular to the layers of graphene is therefore limited to a maximum thickness by virtue of the number of layers of graphene that may be present in said quantum dots, the other dimensions of the quantum dot are not so limited. For example, the lateral diameter of each graphene quantum dot may be from 1 nm to 1 ,000 nm. Again, a variety of lateral diameters may be present in a sample of the nanoparticulate materials described above. In particular embodiments of the invention, the lateral diameter may be considered to be the average lateral diameter of the graphene quantum dots in the nanoparticulate material and this may range from 1 nm to 500, such as from 2 nm to 250 nm, such as from 20 nm to 100 nm, such as less than or equal to 40 nm, such as from 1 nm to 40 nm, such as from 10 nm to less than 40 nm.

Unless otherwise stated herein, when there is more than one numerical range provided for a particular feature, any of the point values disclosed may be combined with any of the other point values disclosed to form a new range. For example in the numerical range provided above for lateral diameter, the average lateral diameter may be from 2 to 500 nm, from 1 to 2 nm, from 100 to 250 nm, from 250 to 500 nm, from 1 nm to less than 40 nm, from 2 nm to 10 nm, and any other combination.

When mentioned herein, the first and second surfaces of the graphene quantum dots refers to the exposed top and bottom surfaces of the layer(s) of graphene in the quantum dot.

As noted above, the surfaces of the graphene quantum dots have been functionalised to include O-based functional groups, N-based functional groups, S-based functional groups, P- based functional groups, and B-based functional groups. In embodiments of the invention where the nanoparticulate material does not contain metal- containing nanoparticles, the functional groups on the first and second surfaces of the graphene quantum dots may account for from 0.1 to 80 wt%, such as from 1 to 10 wt%, such as from 3 to 8 wt%, of the total weight of the nanoparticulate material.

In contrast, in embodiments of the invention where the nanoparticulate material also contains metal-containing nanoparticles, the metal in the metal-containing nanoparticles may account for from 20 to 60 wt% of the total weight of the nanoparticulate material, and/or the functional groups on the first and second surfaces of the graphene quantum dots may account for from 0.1 to 60 wt%, such as from 0.15 to 50 wt%, of the total weight of the nanoparticulate material. For example, the metal in the metal-containing nanoparticles may account for from 30 to 50 wt% of the total weight of the nanoparticulate material, and/or the functional groups on the first and second surfaces of the graphene quantum dots may account for from 0.2 to 50 wt%, such as from 0.2 to 2 wt%, such as from 0.2 to 1 wt%, of the total weight of the nanoparticulate material. O-based functional groups may be selected from the list comprising, but not limited to, hydroxyl groups, carboxylic acid groups, ketone groups, aldehyde groups, ester groups, carbonate ester groups, epoxy groups, and combinations thereof. Unless otherwise stated herein, all groups containing a C=0 group may be referred to as a carbonyl group (e.g. in particular embodiments the carbonyl groups may be selected from one or more of the group consisting of carboxylic acid groups, ketone groups, aldehyde groups and ester groups). It is noted that the presence of O-based functional groups may be inherently present in a graphene material. For example, each of the O-based functional groups mentioned above may be inherently present in reduced graphene oxide. In particular, reduced graphene oxide may contain one or more of hydroxyl groups, carboxylic acid groups, ketone groups, aldehyde groups, and epoxy groups.

S-based functional groups may include, but are not limited to thiol groups, sulphide groups, disulphide groups, sulfonic acid groups and combinations thereof. S-based functional groups may be introduced through the use of a suitable sulfonation agent. Suitable sulfonation agents that may be mentioned herein include, but are not limited to, chlorosulfonic acid, sulphuric acid, sulphur trioxide and the like. A particular sulfonation agent that may be mentioned herein is chlorosulfonic acid.

N-based functional groups may include, but are not limited to amino groups, amide groups, nitro groups, pyridinyl groups and combinations thereof. The inclusion of such N-based functional groups may occur through reaction with O-based functional groups already present in the graphene to be used (e.g. amide formation with a pre-existing carboxylic acid functional group) or by any other suitable method. For example, adenosine triphosphate can be used as a precursor to provide pyridinyl and pyrrolyl groups onto the surface of the graphene quantum dots (e.g. Nanoscale, 2015, 7, 8159-8165). The nitro groups can be introduced onto the graphene quantum dots using nitric acid.

P-based functional groups may include, but are not limited to phosphate groups, phosphonic acid groups and combinations thereof. B-based functional groups may include, but are not limited to boronic acid groups, borinic groups, borinate ester groups and combinations thereof. The inclusion of such P-based and B-based functional groups may occur through reaction with O-based functional groups already present in the graphene to be used. For example, adenosine triphosphate can be used as a precursor to provide phosphonic acid groups at the edge of the graphene quantum dots (e.g. Nanoscale, 2015, 7, 8159-8165).

As will be appreciated, any combination of 0-, S-, N-, P and B- based functional groups that is technically achievable is intended to fall within the scope of the disclosed invention. For example, in particular embodiments of the invention the functional groups on the first and second surfaces of the graphene quantum dots may be selected from the group comprising hydroxyl, carboxylic acid, sulfonic acid, and combinations thereof. As will be appreciated, the presence of acidic functional groups on the graphene quantum dots may contribute to the overall acidity of the nanoparticulate material, which overall acidity may be measured and quantified. For example, the nanoparticulate material may have from 0.1 mmol/g to 10 mmol/g of acidic sites, such as from 2 mmol/g to 8 mmol/g of acidic sites, as measured by temperature programmed desorption of ammonia, or other suitable molecules (e.g. pyridine etc). When used herein, "acidic site" may refer to sites that display Lewis acidity or Bronsted acidity. It will be appreciated that the nanoparticulate material may have some sites that display Lewis acidity and others that display Bronsted acidity. A discussion of Lewis acidity and Bronsted acidity is provided in the experimental section below. In particular embodiments of the invention (e.g. embodiments not including metal-containing nanoparticles), the nanoparticulate material may have about 1 .05 mmol/g of weak acid sites (<270°C), about 2.31 mmol/g of medium acid sites (from 270 to 400°C) and about 0.22 mmol/g of strong acid sites (>400°C), as measured by temperature programmed desorption of ammonia, or other suitable molecules (e.g. pyridine etc). The nanoparticulate materials disclosed herein above (whether including a metal-containing nanoparticle or not) may display one or more of the following properties: (Ca) a BET surface area of from 50 to 500 m ² /g, such as 90 to 350 m ² /g, such as from 100 to 300 m ² /g; and/or

(Cb) the nanoparticulate material may forms agglomerates that have an average diameter of from 1 nm to 1 ,000 nm, such as from 15 nm to 500 nm, such as from 25 to 100 nm, such as from 50 to 75 nm.

In embodiments of the invention where metal-containing nanoparticles are present, the metal- containing nanoparticles may be selected from selected from one or more of the group consisting of metal phosphide nanoparticles, metal sulfide nanoparticles or, more particularly, metal nanoparticles and metal oxide nanoparticles. Any suitable metal may be used as the metal nanoparticle. Examples of suitable metals that may be used to form a metal nanoparticle for use in the current invention include, but are not limited to, Ru or, more particularly, Ni, Au, Pt, and alloys thereof. Any suitable metal oxide may be used as the metal oxide nanoparticle. Examples of suitable metal oxides that may be used to form a metal oxide nanoparticle for use in the current invention include, but are not limited to, M0O3 and V ₂ 0 ₅ . Any suitable metal sulfide may be used as the metal sulfide nanoparticle. An example of a suitable metal sulfide that may be used to form a metal sulfide nanoparticle for use in the current invention is M0S2. Any suitable metal phosphide may be used as the metal phosphide nanoparticle. Examples of suitable metal phosphides that may be used to form a metal phosphide nanoparticle for use in the current invention include, but are not limited to, Ni ₂ P and Ni^Ps- In particular embodiments of the invention, the metal-containing nanoparticles may be M0O3 nanoparticles.

The size of the metal-containing nanoparticles is not particularly limited, provided that they can be still considered to be nanoparticles. For example, the metal-containing nanoparticles may have an average diameter of from 1 to 1 ,000 nm, such as from 50 to 500 nm, such as 100 to 200 nm.

In certain embodiments of the invention, a nanoparticulate material containing metal nanoparticles may be one in which the functional groups on the first and second surfaces of the graphene quantum dots may be selected from the group comprising hydroxyl, carboxylic acid, sulfonic acid, and combinations thereof and the metal-containing nanoparticles may be M0O3 nanoparticles.

As will be appreciated, the functionalised graphene quantum dots as the nanoparticulate material may be achieved by the methods described in the examples section below and using the materials described above. As an example, a carbonaceous material may first be dispersed in nitric acid and subjected to sonication and the resulting suspension reacted with a sulfonation agent to obtain S- and O-functionalised graphene materials (following centrifugation, neutralisation and dialysis) to provide a nanoparticulate material according to the invention that does not include a metal-containing nanoparticle. Alternatively, pre-obtained graphene may be functionalised directly by reaction with a suitable functionalisation agent (e.g. a sulfonation agent) in a suitable solvent (e.g. water and/or an organic solvent) to obtain a nanoparticulate material according to the invention that does not include a metal-containing nanoparticle. The introduction of the metal-containing nanoparticles may be achieved by dispersing said material into distilled water (with sonication) to form a homogeneous solution/suspension, then adjusting the pH of the solution to about 1 (e.g. using HCI) and then adding in a suitable metal-containing nanoparticle precursor material (e.g. ammonium molybdate) with a suitable reducing agent (e.g. pyrrole). The resulting mixture may then be autoclaved (e.g. at 180 °C for 10 h), centrifuged, washed and dried before being calcined at a temperature of from 300 to 500 °C for 1 -3 h under air to provide a nanoparticulate material that also includes a metal-containing nanoparticle. As will be appreciated any suitable synthetic method may be used, either based on the methods disclosed herein or based upon the common general knowledge of a practitioner in this field.

It has been surprisingly found that the materials disclosed herein can be used as catalysts for one-pot synthesis of 2,5-diformylfuran and/or 5-hydroxymethylfurfural from fructose, and other feedstocks. The materials disclosed herein are inexpensive carbon materials and their application as catalysts will be more cost effective than many commercial catalysts (e.g. zeolites and MCM-41 ) when applied to industrial settings. When used as catalysts, the nanoparticulate materials unify the merits of heterogeneous and homogeneous catalysts, which results in highly efficient catalytic performance in the biomass conversion to valuable platform molecules (e.g. 2,5-diformylfuran and/or 5-hydroxymethylfurfural). In addition, it is also important to note that the reaction conditions needed are much milder than many of the currently used processes to manufacture these materials.

Thus, there is also disclosed herein a method of converting a biomass feedstock to 2,5- diformylfuran, the method comprising the step of contacting the biomass feedstock with a catalytic amount of a nanoparticulate material as defined hereinbefore. In embodiments of the invention, when the nanoparticulate material also includes a metal-containing nanoparticle, the feedstock may be converted without additional steps being necessary. However, when the nanoparticulate material does not include a metal-containing nanoparticle, the feedstock may be converted initially to 5-hydroxymethylfurfural and a further oxidation step may be needed to convert 5-hydroxymethylfurfural to 2,5-diformylfuran. Said oxidation step, where needed, may be any suitable oxidation step. When the nanoparticulate material contains metal-containing nanoparticles, then the full conversion from the feedstock to 2,5-diformylfuran proceeds without any need for further introduction of reagents. For example, in the presence of such nanoparticulate materials, 2,5- diformylfuran can be obtained in a yield of up to 78% at 160°C for 2 hours in DMSO solvent. This is at least 2 times more effective than the existing catalysts, such as V ₂ 0 ₅ @carbon nanotube.

Commercially, high yields of 2,5-diformylfuran are only currently obtained from the oxidation of 5-hydroxymethylfurfural, which is a very expensive platform molecule. The nanoparticulate materials disclosed herein which also include a metal-containing nanoparticle, therefore provide a facile and practical protocol for synthesising 2,5-diformylfuran directly from fructose (or other abundant biomass-derived sources), which may open up new opportunities for the industrial scale production of 2,5-diformylfuran for use in the chemical and pharmaceutical industries.

There is also disclosed a method of converting a biomass feedstock to 5-hydroxymethylfurfural, the method comprising the step of contacting the biomass feedstock with a catalytic amount of a nanoparticulate material as defined above, where the nanoparticulate material does not comprise metal-containing nanoparticles. When used as a catalyst to convert fructose to 5- hydroxymet ylfurfural, the nanoparticulate materials that do not contain metal nanoparticles can provide the product in a yield of up to 92% at 150 °C after only 2-hour reaction time. Without wishing to be bound by theory, the good dispersion of the disclosed nanoparticles in the reaction mixture in combination with the functional groups attached to said nanoparticles, which act as catalytic sites, may be responsible for the high efficiency in this reaction. In addition, the nanoparticulate materials can be reused for a theoretically unlimited number of times without any obvious loss of catalytic activity, based on the results discussed in the examples section below (e.g. no loss of activity was noticed after 8 uses of the catalytic material). In contrast, many commercial catalysts for this reaction cannot be recycled, or their catalytic activity significantly declines following recycling for a 2 ^nd time or 3 ^rd time.

The conversions described above may make use of any suitable feedstock. Examples of suitable feedstock include, but are not limited to xylose, cellulose, or, more particularly, sucrose, fructose, glucose, cellobiose, inulin and combinations thereof. In particular examples of the methods of conversion discussed herein, the feedstock may be fructose and/or cellulose. The conversions of the feedstock described above may be conducted in any suitable solvent (e.g. such as water, DMSO, isopropanol, butanol, methyl isobutyl ketone or valerolactone), in any suitable reaction vessel (e.g. moving-bed reactor, tube reactor, stirred tank reactor, etc.), at a suitable reaction temperature (e.g. from 120 - 200°C) and for a suitable period of time (e.g. from 1 to 6 hours (or the equivalent in a flow reactor setup)). As will be appreciated, the nanoparticulate material is intended to be used as a catalyst, but any suitable amount may be used. For example, the nanoparticulate material may be present in an amount of from 0.1 to 40 wt% relative to the weight of the feedstock, such as from 0.5 to 20 wt%, such as from 1 to 10 wt%, such as 2 wt%.

In certain embodiments, a biphasic solvent system may be used, said system may comprise an aqueous phase (e.g. water alone or in combination with a water-miscible organic solvent) and an organic phase (e.g. one or more water-immiscible solvents). For example, in certain embodiments, the aqueous phase may contain DMSO and water in a v:v ratio of 7:3 and an organic phase that contains methyl isobutyl ketone (MIBK): butanol in a v:v ratio of 7:3, where the wt:wt ratio of the aqueous phase to organic phase may be 1 :2. Such a solvent system may be particularly useful in combination with the nanoparticulate materials disclosed herein that do not include a metal-containing nanoparticle and so may be particularly suited to the conversion of the feedstocks mentioned herein to 5-hydroxymethylfurfural.

In other embodiments, a homogeneous solvent system comprising DMSO and a suitable ester solvent (e.g. valerolactone) may be used. In such solvent systems, the v:v ratio of DMSO to valerolactone may conveniently be from 2:1 to 10:1 , such as 8:2. While the nanoparticulate materials disclosed herein have been shown to be useful in the conversion of biomass feedstocks to 5-hydroxymethylfurfural and/or 2,5-diformylfuran, the nanoparticulate materials may be functionalised in a number of ways that may suit the needs of other specific reactions. This is because there are a variety of exchangeable oxygenated groups exposed on the layered graphene nanosheets and their inherent electron-donating capacity allows for variable functionalisation to be conducted (e.g. as discussed herein before). This feature may allow the nanoparticulate materials disclosed herein to be tailored for the degradation of different biomass feedstocks or different compositions of feedstock (i.e. a specific type of biomass or a mixture of several specific biomass types), as well as to achieve different valuable platform molecules as desired.

Further aspects and embodiments of the invention will now be described by reference to the following non-limiting examples. Examples

The current invention relates to the use of functionalised graphene quantum dots (GQDs) in converting carbohydrates to useful and valuable products like 5-(hydroxymethyl)furfural (5-HMF) and 2,5-diformylfuran (DFF). This is shown in Fig. 1 in which fructose 10 undergoes dehydration 20 to form 5-HMF 30, in the presence of the mono-functional GQDs 60, which in this application refers to the sulfonated GQDs. The 5-HMF 30 can then undergo oxidation 40 to form DFF 50. The one-pot transformation of 10 to 50 can be achieved with the use of bi- functional GQDs 70, which in this application refers to the molybdenum oxides-grafted sulfonated GQDs.

Materials Fructose (≥ 99%, Alfa Aesar), glucose (99.5%, Sigma-Aldrich), sucrose (99%, Sigma-Aldrich), cellobiose (≥ 98%, Sigma-Aldrich), xylose (≥ 99%, Sigma-Aldrich), ZSM-5 (Si/AI = 30, Alfa Aesar), MCM-41 (Sigma-Aldrich), SBA-15 (Sigma-Aldrich), Nation (Sigma-Aldrich), Amberlyst-15 (hydrogen form, Sigma-Aldrich), inulin from chicory (Sigma-Aldrich), cellulose powder (Sigma-Aldrich), 5-(hydroxymethyl)furfural (5-HMF, ≥ 99%, Sigma-Aldrich), 2,5- diformylfuran (DFF,≥ 99%, Alfa Aesar), ammonium molybdate tetrahydrate (≥ 99%, Alfa Aesar), pyrrole (99%, Sigma-Aldrich), chlorosulfonic acid (98%, Sigma-Aldrich), methoxytrimethylsilane (99%, Sigma-Aldrich), carbon black (Cobat Corporation), carbon nanotube (Sigma-Aldrich), activated carbon (Johnson Matthey, United Kingdom) and other chemicals are commercially available. All the chemicals were used without further purification. Butyl sultone pyridinium-based Bronsted ionic liquid, [PyBS][HS04], was prepared according to the previous report {Phys. Chem. Chem. Phys. 2016, 18, 32723-32734).

General Methods The sulfonated graphene quantum dots (SGQDs) and molybdenum oxide-grafted sulfonated graphene quantum dots (M0O3/SGQDS) catalysts were finely ground to a particle size of less than 1 mm prior to characterisation. Fourier transform infrared (FTIR) spectra of the catalysts were obtained on a Digilab FTS 3100 FTIR spectrometer with a 4 cm ¹ resolution in the range of 400-4000 cm ¹ using a standard KBr disk technique. Alternatively, the FT-IR measurements were performed on a Bruker Tensor spectrometer. The transmission electron microscopy (TEM) images were collected using a JEOL 21 OOF TEM microscope with a ZrO/W Schottky Field Emission Gun with an accelerating voltage of 200 kV. A scanning TEM-energy-dispersive X-ray (STEM-EDX) spectrometer with high-angle annular dark field was used for mapping the elemental distribution of the samples. Atomic force microscopy (AFM) images were collected using Bruker Dimension Icon system. Raman spectra of the samples (Renishaw InVia Reflex Raman) were obtained at an excitation wavelength of 633 nm.

The carbon, oxygen and sulfur contents on the surface of GQDs and SGQDs were determined by X-ray photoelectron spectroscopy (XPS, Escalab MK II spectrometer), using an Al X-ray excitation source (Ka = 1487.6 eV). The amount of functional groups bonded to SGQDs was estimated by elemental analysis (vario EL cube, ELEMENTAL) and cation-exchange analysis (M. Hara, et al., Angew. Chem. Int. Ed., 2004, 43, 2955-2958; S. Suganuma, et al., J. Am. Chem. Soc, 2008, 130, 12787-12793). The density of S0 ₃ H groups were calculated based on the sulfur content obtained from the elemental analysis. Subsequently the contents of SO3H + COOH and SO3H + COOH + OH were determined by the exchange of Na ⁺ in aqueous solutions of NaCI and NaOH, respectively. The proportions of each functional group were estimated based on the analysis results. Methods for catalyst testing and analysis of the products

Preliminary method

The catalytic reaction was performed in a 50 mL flask equipped with a condenser. In a typical run, the carbohydrate feedstock (500 mg), functionalised GQDs (10 mg), and solvent (10 mL) were added into the flask. The mixture was then heated to the desired temperature with stirring. After the completion of the reaction, the reaction mixture was cooled in an ice bath and the furan-derived products were extracted using methyl isobutyl ketone (MIBK). The functionalised GQDs which were well-dispersed in the solvent, were dialysed to remove the products and other impurities prior to reusing for the next run.

The aqueous phase of the reaction mixture was diluted with deionised water and filtered with a 0.45 mm syringe filter prior to analysis. Both the aqueous and organic phases of the reaction mixture were analysed for the products. The yield of 5-HMF or DFF was determined using high-performance liquid chromatography (HPLC, Agilent 1 100 Series) equipped with a Bio- Rad Aminex HPX-87H pre-packed column (300 mm x 7.8 mm) and an auto-sampler (Agilent G1329A). Sulfuric acid (H2SO4, 5 mM) was employed as the mobile phase at a flowing rate of 0.6 mL/min at 60 °C. An external standard was used for quantifying the products and the identities of the compounds were determined by comparing their retention times with those of the pure compounds. The main products produced from the dehydration of fructose are 5- HMF and DFF. The remaining by-products (furfural, levulinic acid and formic acid) were detected by gas chromatography-mass spectrometry (GC-MS) as the concentrations of these by-products were too low to be quantified by HPLC analysis.

Optimised method The catalytic reaction was performed in a 10 mL flask equipped with a low-temperature condenser at 1 °C. In a typical run, the carbohydrate feedstock, catalyst and solvent were charged into the flask and the mixture was heated at 170 °C for 2 h with stirring. After the completion of the reaction, the reaction mixture was quenched in an ice bath. The 5-HMF product stayed predominately in the upper phase (organic phase), while the catalyst and the unreacted hydrophilic reactants remained in the lower phase (aqueous phase). The SGQDs in the aqueous phase were purified by dialysis, using a dialysis bag with a 500 Da molecular weight cut-off (MWCO), for subsequent use in another catalytic reaction.

The 5-HMF product in both the organic and aqueous phases were sampled and analysed by HPLC (Agilent 1 100 Series) equipped with a XDB-C18 column and a UV detector at 320 nm, using methanol: water (2:8, v/v) as the mobile phase. The carbohydrate feedstocks in both phases were analysed by the same HPLC equipped with a Bio-Rad Aminex HPX-87H prepacked column (300 mm x 7.8 mm) and refractive index detectors (RID), using 5 mM of H2SO4 as the mobile phase (J.N. Chheda, ef al., Green Chem., 2007, 9, 342-350). An external standard was used for quantifying the products and the identities of the compounds were determined by comparing their retention times with those of the pure compounds. The carbohydrate conversion, the molar yields of 5-HMF and other products were calculated from the product concentration in both the organic and aqueous phases. Example 1. Synthesis of sulfonated graphene quantum dots (SGQDs) and molybdenum oxides-grafted sulfonated graphene quantum dots (MoCVSGQDs)

Synthesis of sulfonated graphene quantum dots (SGQDs) In the initial attempt as shown in Fig. 2, the carbon black 210 was first dispersed in 6 M of nitric acid 220 under sonication for 1 -2 h to give a black dispersed suspension of the carbon black materials. The sulfonating agent 230, chlorosulfonic acid, was then added dropwise into the black dispersed suspension and the mixture was heated at 130°C for 36 h under vigorous stirring. This involves the process 280 of acidic cutting and sulfonation which breaks down the bulk carbon materials into smaller sizes. After cooling to room temperature, the suspension was centrifuged at 8000-10000 rpm for 20 min to settle the unreacted carbon black particles to the bottom. The supernatant was then diluted with deionised water and neutralised with concentrated NaOH solution before filtering through a syringe membrane with pore size of 0.2 μπι. The mixture was then dialysed in deionised water for one to three days, using a dialysis bag with a MWCO of 1000 Da, to remove salts. The purified mixture was freeze-dried to give the SGQDs 240 as a dark-brown powder.

In a subsequent optimised attempt, the GQDs was first synthesised according to a reported procedure using carbon black as the raw material (J. Mat. Chem., 2012, 22, 8764-8766). To achieve sulfonation, 1 g of GQDs was pre-treated by calcination under N ₂ at 300 °C for 1 h to form a dark solid, which was subsequently dispersed in chloroform containing 5 ml_ of chlorosulfonic acid. The mixture was continuously refluxed under N ₂ with vigorous stirring for 24 h at 70 °C. After cooling to room temperature, the suspension was centrifuged, diluted with deionised water and neutralised with concentrated NaOH solution. The mixture was further filtered with a syringe membrane of pore size 0.2 pm, followed by dialysis for 3 d in deionsied water (using a dialysis bag with MWCO of 500 Da) to remove salts. The SGQDs was finally obtained after freeze-drying. Other sulfonated carbon materials (e.g. sulfonated carbon nanotube, sulfonated activated carbon) were synthesised using the same procedure.

The sulfonated GQDs synthesised in the above methods is denoted as "SGQDs" and this term is used throughout the following examples to refer to the materials synthesised in this example.

Synthesis of MoOn/SGQDs

After obtaining the SGQDs, the molybdenum active sites in the form of molybdenum oxides were subsequently grafted onto the SGQDs (Fig. 2). Typically, 20 mg of SGQDs was first dispersed in deionised water by ultrasonication for 1 -2 h to form a homogeneous solution. Thereafter, the pH of the solution was adjusted to pH 1 with the addition of HCI. The molybdenum precursor 250, ammonium molybdate (100 mg), was added into the reaction mixture, followed by the addition of pyrrole (254 μΙ_) as the reducing agent. After complete dissolution of the reactants, the mixture was next transferred into a Teflon-lined stainless steel autoclave 260 and was heated at 180 °C for 10 h in an oven. This allowed the grafting of molybdenum onto SGQDs as shown in process 290. The reaction mixture was then centrifuged, washed thoroughly with deionised water, and dried at 60 °C for 12-24 h. Finally, the product was calcined at 400 °C for 3 h in air with a heating rate of 5 °C/min to obtain the molybdenum oxides-grafted SGQDs catalyst 270.

The as-synthesised molybdenum ox ides -grafted SGQDs is denoted as "MoCVSGQDs" and this term is used throughout the following examples to refer to the materials synthesised in this example.

Example 2. Characterisation of SGQDs and M0O3/SGQDS by Fourier transform Infrared (FTIR) and Raman spectroscopy (for SGQDs)

The as-synthesised SGQDs and M0O3/SGQDS of Example 1 were characterised by FTIR spectroscopy to determine the chemical functionalities on the respective GQDs. Fig. 3a shows the IR spectra for GQDs, SGQDs, and M0O3/SGQDS. In contrast to the non-functionalised GQDs, the IR spectra for SGQDs and M0O3/SGQDS clearly showed the vibrational band at 1260 cm ¹ due to the S=0 stretches of SO3H, which suggests the presence of sulfonic groups on both the SGQDs and Mo0 ₃ /SGQDs catalysts.

In another comparative study between the as-synthesised GQDs and SGQDs of Example 1 , these two materials were further characterised by FTIR and Raman spectroscopy. The vibrational peaks originated from epoxides (C-O-C at -1280-1330 cm ^"1 ), sp ² -hybridised C=C (in-plane vibrations at -1580 cm ^"1 ), and edge groups such as carbonyls or carboxyls (C=0 or COOH at -1650-1750 cm ^"1 ) were observed in GQDs (Fig. 3b).

In contrast, the intensity of the epoxide peak decreased significantly in SGQDs. The vibrational bands at 1130 cm ^"1 and 1215 cm ^"1 were assigned to S=0 asymmetric and symmetric stretches of -SO3H, which further confirmed the successful grafting of sulfonic groups onto the GQDs. The broadband at 2300-2700 cm ^"1 was assigned to an overtone (Fermi resonance) of the bending mode of -ΟΗ···0 = originated from a strong hydrogen bond, which is often found in strong liquid Bronsted acids such as CF3SO3H (S. Suganuma, et al., J. Am. Chem. Soc, 2008, 130, 12787-12793). This observation implies that the -OH groups on the SGQDs, which were in close proximity to the -SO3H groups, were responsible for the strong hydrogen bonds.

In the Raman spectra as shown in Fig. 3c, the GQDs gave a D band (Ai _g breathing mode) at 1368 cm ^"1 and a G band (E _2g G mode) at 1590 cm ¹ , which were due to sp ³ carbon atoms from the defects and sp ² carbon atoms in the graphitic hexagonal lattice respectively (G. Eda, et al., Adv. Mater., 2010, 22, 505-509). In comparison to SGQDs, the ratio of the intensity of D band to G band was higher for SGQDs than that for GQDs, which suggests the disruption of graphitic structure upon introduction of SO3H groups on the GQDs. In addition, a blue shift of the G band was observed in SGQDs, which was probably due to the formation of single-double bond alternation within the sp ² carbon domains when two hydroxyl groups were grafted adjacently on the basal plane (K.N. Kudin, et al., Nano Lett, 2008, 8, 36-41 ).

Example 3. Characterisation of SGQDs and M0O3/SGQDS by transmission electron microscopy (TEM), Energy-dispersive X-ray spectroscopy (EDS) and Atomic Force Microscopy (AFM)

To understand the morphologies of SGQDs and M0O3/SGQDS, TEM imaging was carried out and the TEM images are as shown in Fig. 4a and b. The SGQDs were found to be composed of randomly oriented particles which were predominantly in the size range of 10-20 nm. The morphology of the SGQDs is consistent with the previous study reporting the TEM images of GQDs, which demonstrates that the SGQDs retained the characteristic morphology of GQDs even after sulfonation treatment (G. Liu, etal., Nanoscale, 2017, 9, 4934-4943). Fig. 4b shows that the M0O3/SGQDS possessed irregular nanoparticles with size ranging from 100-200 nm. These nanoparticles were ascribed to the molybdenum oxides on the GQDs sheets. The GQDs and the molybdenum oxides on the M0O3/SGQDS displayed a lattice spacing of 0.243 nm and 0.320 nm respectively.

In a further characterisation study by TEM, it was observed that SGQDs had a diameter about 10.4 nm (±3.1 nm, n = 110) with a lattice spacing of 0.245 nm corresponding to (1120) plane of graphene (Fig. 4c and d).The SGQDs were further characterised by EDS, in which the EDS elemental mapping revealed that the S element was uniformly distributed on the basal plane (presumably on both sides) of SGQD, instead of residing at the edge sites (Fig. 4c). This was further confirmed by the TEM-EDS line scanning spectrum (Fig. 4d). Such homogeneous conjugation of SO3H groups is desirable for full utilisation of active catalytic sites on the GQDs. Furthermore, characterisation by AFM revealed that the SGQDs had an average thickness of 1.06 nm (± 0.34 nm, n = 152), in which they exist mostly as single or double layers (Fig. 4e and g)-

Example 4. Characterisation of the SGQDs by X-ray Photoelectron Spectroscopy (XPS), elemental analysis and cation-exchange experiments The as-synthesised SGQDs of Example 1 were characterised by XPS to determine the chemical states of the elements as well as the elemental composition. Other than the common C1s peak at 284.5 eV and 01 s peak at 530.8 eV observed from the XPS spectra in both SGQDs and GQDs (Fig. 5a), additional S2s peak at 233.1 eV and S2p peak at 169.8 eV were also observed in SGQDs. This indicates the successful conjugation of SO3H groups to the GQDs. The atom percentages of carbon, oxygen, and sulfur in SGQDs were 60.53%, 34.71 %, 4.76% respectively, as determined from the XPS spectra. The sulfur content was much higher than other sulfonated carbon material, implying that the as-synthesised SGQDs should have stronger Bronsted acidity (J.Y. Ji, etal., Chem. Sci., 2011 , 2, 484-487; Z. Sun, et al., Cellulose, 2015, 22, 675-682; A.V. Nakhate, G.D. Yadav, ACS Sustain. Chem. Eng., 2016, 4, 1963- 1973). In addition, the total O/S atom ratio of -7:1 in SGQDs was much higher than the exact O/S atom ratio in the -SO3H groups (3:1 ), therefore suggesting that many oxygen-containing groups (e.g. -COOH, -OH) remained on the SGQDs after sulfonation. The high-resolution C1 s XPS peak of original GQDs can be deconvoluted into C=C (graphitic carbon) at 284.7 eV, C-C (sp ³ -hybridised carbon) at 285.4 eV, C-O-C/C-OH (epoxy/hydroxyl) at 286.8 eV and C(=0)-OH (carboxylic group) at 289.5 eV (Fig. 5b). In comparison, the peak intensity at 286.8 eV was lower in SGQDs, indicating that many epoxy groups on the basal plane (which were more vulnerable to reduction as compared with hydroxyl groups) were removed after the thermal annealing (R. Maiti, etal., Nanotechnology, 2014, 25, 495704). This is consistent with the FTIR spectrum of SGQDs in Fig. 3b which shows a weaker epoxy peak as compared to the GQDs. Taken together, the oxygen functionalities in SGQDs were mainly phenolic OH, COOH/C=0, and S0 ₃ H. Similar to GQDs, the SGQDs gave a C=C peak at -284.7 eV, suggesting that the graphitic structure was preserved after sulfonation (Fig. 5b). In addition, the GQDs and SGQDs shared similar 01 s spectra (Fig. 5c), whereas the S2p spectrum for SGQDs displayed a new peak at 169.8 eV which was due to the SO ₃ H groups (Fig. 5d).

Further, elemental analysis and cation-exchange experiment were also carried out to confirm the composition of the SGQDs, which was determined to be CH0.465O0 573So.o79- The amount of SO ₃ H, COOH, and phenolic OH groups were calculated to be 3.1 , 1.5, and 3.5 mmol- g ^~1 respectively.

Example 5. Acidity characterisation of SGQDs by pyridine-adsorption FTIR and temperature programmed desorption of ammonia (NH3-TPD) Method

The acidic sites on GQDs and SGQDs were determined by pyridine adsorption characterised by FTIR. In this experiment, FTIR spectra were recorded on a Nicolet-6700 FT-IR spectrophotometer with a resolution of 4.0 cm ¹ . A self-supported wafer of the test sample was placed inside a quartz IR cell equipped with CaF ₂ windows. The samples were pre-treated under vacuum (0.01 Pa) at 350 °C for 2 h and the background spectra were recorded after cooling the sample to room temperature. The wafer was exposed to excess pyridine vapour at room temperature for 0.5 h, followed by the removal of excess pyridine by evacuating the samples at 150 °C and 300 °C, respectively, for 0.5 h. The IR spectra of adsorbed pyridine on the samples were then collected after cooling the samples to room temperature. The spectra of pyridine adsorption were obtained by subtracting the sample spectrum with the background spectrum. The acidic strength of GQDs and SGQDs was measured by temperature programmed desorption of ammonia (NH ₃ -TPD) on a Micromeritics AutoChem 2910 instrument. Before adsorption, the samples were pre-treated under an Ar stream at 300 °C. Subsequently, they were cooled to 100 °C with an Ar flow rate of 20 cm ³ /min before commencing the ammonia adsorption process. The adsorption step was carried out by allowing small pulses of ammonia in Ar at 100 °C until saturation was reached (normally 1 h). The samples were exposed to an Ar flow of 50 cm ³ /min for 2 h at 100 °C. Finally, the desorption was performed from 100 °C to 600 °C at 10 °C/min and maintained at 600 °C for 10 min under an Ar stream of 50 cm ³ /min, in which the adsorbate was completely desorbed under this condition. Results and Discussions

The acidity of GQDs and SGQDs was determined by pyridine-adsorption FTIR as described in the method above. After pyridine desorption at 150 °C, a small peak at 1456 cm ^"1 (indicative of Lewis acidity) and a prominent peak at 1531 cm ^"1 (indicative of Bronsted acidity) were observed in SGQDs (Fig. 6). The Lewis acidity of SGQDs probably originated from the dehydration of Bronsted acid sites or electron inductive effect of the S=0 bonds as demonstrated by Upare et al., while the Bronsted acidity originated from the -SO3H and -COOH groups (P.P. Upare, et al., Green Chem., 2013, 15, 2935-2943). When the desorption temperature was increased to 300 °C, the Lewis acidity peak remained unchanged while the Bronsted acidity peak decreased, obviously due to the loss of weak Bransted acidity. In comparison, the non-sulfonated GQDs showed weaker Bransted acidity and no obvious Lewis acidity. Given these, it was observed that other than the significant enhancement of Bransted acidity by the addition of SO3H groups to SGQDs, the sulfonation process also induced a small amount of Lewis acidity onto SGQDs. The ratio of Bronsted acidity to Lewis acidity for SGQDs was determined to be 3.73 (Fig. 6), based on the absorbance intensity of the respective IR peaks in the FTIR spectrum at 300°C.

The acid strength and amount were determined by temperature programmed desorption of ammonia (NH3-TPD). It was observed that SGQDs contained 1.05 mmol/g of weak acid sites (<270 °C), 2.31 mmol/g of medium acid sites (27C 00 °C) and 0.22 mmol/g of strong acid sites (>400 °C), while 1.23 mmol/g of weak acid sites and 0.35 mmol/g of medium acid sites were observed in GQDs (Table 1 ). Evidently, the medium acid sites were predominantly found in SGQDs and the acidity distribution was in accordance with the co-existence of weak and strong Bronsted acidity as revealed by the pyridine-adsorption FTIR experiments. The observation that GQDs had no strong acidic sites and a much lesser amount of medium acidic sites was also consistent with the observation of the weak Bronsted acidity in GQDs.

Table 1. Amount of acid sites in GQDs and SGQDs, determined by temperature programmed desorption of ammonia

Example 6. Catalytic performance of SGQDs in converting fructose to 5-HMF in different solvent systems and different initial fructose concentrations

Both the original GQDs and SGQDs can disperse homogeneously in aqueous solution to serve as quasi-homogenous acidic catalysts. Therefore, the conversion of fructose into 5-HMF using SGQDs was carried out in different solvent systems, using the optimised method described above in "Methods for catalyst testing and analysis of products". Subsequently, the reaction was carried out with different initial fructose concentrations to study the effect on the yield of 5-HMF.

Effects of different solvent systems As shown in Fig. 7a, the production yield of 5-HMF was very low (17.1 %) in pure water and this was accompanied by a considerable amount of condensation compounds (43.5%). This was probably due to the use of water, which may allow many reaction pathways as it is a nonselective solvent for Bronsted acid-catalysed conversion of carbohydrates (G. Yang, et ai, J. Catal., 2012, 295, 122-132).

The addition of organic solvent can potentially extract the 5-HMF product and consequently avoid the further reactions of 5-HMF in aqueous phase (e.g. condensation and rehydration). With the use of a biphasic system comprising methyl isobutyl ketone (MIBK): butanol (v/v of 7:3) as the organic solvent, and water as the aqueous phase (organic solvent: water at w/w of 2:1 ), the conversion of fructose and the yield of 5-HMF increased to 46.1 % and 22.3% respectively. In addition, the 5-HMF produced was extracted into the organic phase with high purity. Such in-situ extraction not only improved the yield and purity of 5-HMF, but also provided a convenient method for collecting the pure product. However, it was observed that a large amount of undesired condensation compounds (42.1 %) was produced.

The use of a mixture of DMSO and water (DMSO: water with v/v of 7:3) as solvent significantly enhanced the conversion of fructose and the yield of 5-HMF to 89.5% and 41 .7% respectively while the yield of undesired condensation compounds was reduced to 38.1% (Fig. 7a). However, it was difficult to extract 5-HMF from the solvent by distillation given that DMSO has a very high boiling point (189 °C).

Given this, a biphasic system consisting of an aqueous phase (DMSO:water with v/v of 7:3) and an organic phase (MIBK:butanol with v/v of 7:3) was used, such that the w/w ratio of aqueous phase to the organic phase was 1 :2. The conversion of fructose and the yield of 5- HMF were increased to 91 .8% and 51 .7% respectively, while the yield of the condensation product was decreased to 36.5%. Advantageously, this system allowed the high purity 5-HMF produced to be collected easily by extracting the organic phase from the biphasic system on standing and evaporating the organic solvents. On the other hand, the SGQDs remained in the aqueous phase which can be collected and purified for subsequent reuse.

Effects of different initial fructose concentrations

A high fructose concentration of 20wt% was used in the experiments described above. Generally, a high substrate concentration is preferred to improve the absolute production yield in a chemo-catalytic process. Typically, the increase of fructose concentration can lead to a higher conversion of fructose due to the production of more condensation products through polymerisation of fructose (J. Wang, etal., AIChEJ., 2013, 59, 2558-2566). However, a higher initial fructose concentration may not necessarily give a high yield of 5-HMF. In fact, this can lead to a lower yield of 5-HMF instead due to the formation of condensation compounds. Therefore, most of the current methods do not use initial fructose concentrations of more than 5 wt% for the reason mentioned above (Y. Li, et al., Bioresour. Technol., 2013, 133, 347-353; Y. Song, et al., Catalysts, 2016, 6, 49-59).

As expected, when the initial fructose concentration was reduced to 5 wt%, the yield of 5-HMF was further increased to 65.3% (Fig. 7b), using the same biphasic solvent consisting of an aqueous phase (DMSO:water with v/v of 7:3) and an organic phase (MIBK:butanol with v/v of 7:3), with an aqueous to organic phase w/w ratio of 1 :2. Nevertheless, the SGQDs were able to perform well in a high substrate concentration and this was probably due to the abundant medium acid sites which favoured the dehydration of fructose to 5-HMF. Therefore, a substrate concentration of 20 wt% of fructose was chosen to balance the absolute production yield and the selectivity of 5-HMF.

Example 7. Comparative studies of the catalytic performance of SGQDs with other catalysts in dehydrating fructose to 5-HMF Initial studies using the preliminary method

The comparative study involving the dehydration of fructose to 5-HMF was first conducted in 10 mL DMSO in the presence of 0.5 g fructose and 10 mg of catalyst at 150 °C for 2 h (adapted from the preliminary method in "Methods for catalyst testing and analysis of the products"). The commonly used catalysts for dehydration of fructose include homogeneous catalysts like H2SO4 and butyl sultone pyridinium-based ionic liquid [PyBS][HSC>4], and heterogeneous catalysts like zeolite-based catalyst (ZSM-5), AI-MCM-41 and AI-SBA-15. Doping of MCM-41 and SBA-15 with aluminium oxide was carried out in-house to achieve a Si/AI molar ratio of 30 using conventional techniques. That is, AI-MCM-41 was prepared according to Fuel, 2009, 88, 461-468; and AI-SBA-15 was synthesised according to Green Chem., 2014, 16, 4985- 4993.

Table 2 shows the efficiency of the SGQDs of Example 1 for the dehydration of fructose, in comparison with the commonly used Bransted acid catalysts. The SGQDs exhibited high conversion of fructose (as high as 100%) and superior selectivity for the 5-HMF product, as compared to the heterogeneous catalysts. Although the H2SO4 and [PyBS][HS04] ionic liquid displayed a selectivity for 5-HMF comparable to that of SGQDs, they are corrosive, non- environmentally-friendly, or very expensive. These factors therefore directly limit the scaled- up production of 5-HMF. In contrast, the SGQDs are an inexpensive carbon-based material and they displayed highly efficient reactivity similar to homogeneous catalysts, due to their efficient functional groups and excellent dispersion in the solvent.

Table 2. Effects of different catalysts on the dehydration of fructose to 5-HMF in DMSO, using the reaction conditions adapted from the preliminary method ³

aReaction conditions: The reaction was conducted in 10 mL DMSO in the presence of 0.5 g fructose and 10 mg catalyst at 150 °C for 2 h.

Subsequently, comparative studies of the various catalysts using 20 wt% fructose concentration in a biphasic solvent system were carried out (adapted from the preliminary method in "Methods for catalyst testing and analysis of the products"). The reactions were carried out at 170 °C for a reaction time of 2 h in 5 mL of biphasic solvent consisting of an organic phase (MIBK:2-butanol with v/v of 7:3) and an aqueous phase (DMSO:water with v/v of 5:5), with an aqueous to organic phase w/w ratio of 1 :2.

The SGQDs displayed the best catalytic efficiency among the heterogeneous catalysts and even showed activity comparable to that of the homogeneous catalysts (0.25 M HCI). It was noted that SGQDs showed a higher yield of 5-HMF, as well as a higher selectivity than any of the heterogeneous catalysts presented in Table 3. The quasi-homogeneity of SGQDs in the reaction medium and the strong Bronsted acidity were responsible for their excellent catalytic activity. The conventional homogeneous Bronsted acid catalysts such as the [PyBS][HS04] ionic liquid and 0.25 M HCI gave 5-HMF yields of 31.6% and 44.1% respectively. Though 0.25 M HCI may have exhibited slightly higher yield and selectivity of 5-HMF than SGQDs, it is corrosive and non-environmentally friendly which therefore makes it impractical for use in a large-scale production of 5-HMF. Table 3. Effects of different catalysts on the dehydration of fructose with initial fructose concentration of 20 wt % in the biphasic solvent system (using the reaction conditions adapted from the preliminary method) ^b

bReaction conditions: The initial fructose concentration with respect to the aqueous phase was 20 wt% and the amount of catalyst used was 10 wt% with respect to the weight of the feedstock. The reactions were carried out at 170 °C for 2 h in a biphasic solvent (5 ml.) consisting of an aqueous phase (DMSO:water with v/v of 5:5) and an organic phase (MIBK:butanol with v/v of 7:3), with an aqueous to organic phase w/w ratio of 1 :2. Subsequent comparative studies using the optimised method

In another attempt using the optimised method adapted from "Methods for catalyst testing and analysis of the products", various acid catalysts were tested at the same dosage of 40 mg with initial fructose concentration of 20 wt% (with respect to aqueous phase) and in a biphasic solvent consisting of an aqueous phase (DMSO:water with v/v of 7:3) and an organic phase (MIBK:butanol with v/v of 7:3), with an aqueous to organic phase w/w ratio of 1 :2. In these studies, the weight of the aqueous phase was 2.1 g, therefore 420 mg of fructose was used. In addition, the total acidity and SO3H density were determined by titration and elemental analysis respectively.

As shown in Table 4, it was observed that even weakly acidic GQDs gave a 5-HMF yield of 31 .2% and a fructose conversion of 54.6%, outperforming other non-SOsH functionalised graphene oxide sheets with similar total acidity (P.P. Upare, et ai, Green Chem. 2013, 15, 2935-2943). This implies that the quasi-homogeneity of GQDs is desirable as this property allows full exposure of the active sites to the reactants in the same phase, therefore providing unhindered mass transport. The SGQDs of Example 1 were observed to be more superior than the original GQDs because of the higher amount of medium Bronsted acid sites. In addition, the SGQDs also performed better than the common heterogeneous catalysts (Amberlyst-15, ZSM-5 and Nation), sulfonated carbon nanotube (SCNT), and sulfonated activated carbon (SAC), though some of these catalysts had a relatively higher SO3H density and total acidity (Fig. 8a and b; and entry 7-1 1 in Table 4). This suggests that the quasi-homogeneity and other properties of SGQDs play important role in the catalytic reactions which allows the active sites on the SGQDs to be utilised fully. Remarkably, the SGQDs also exhibited better catalytic activity than the commonly used homogeneous catalysts ([PyBS][HS04] ionic liquid and H2SO4) which have much higher total acidity than SGQDs (entry 4 and 6, Table 4). This also implies that the catalytic performance of SGQDs may be determined by its physicochemical properties, in which the chemical moieties on SGQDs may exert synergistic effects with the physical properties. Although HCI gave slightly higher 5-HMF yield than SGQDs, it is corrosive, environmental unfriendly and not recoverable.

Interestingly, in contrast to other heterogeneous or homogeneous catalysts, the SGQDs gave almost 100% carbon mass conversion of fructose to 5-HMF and the condensation compounds (without apparent carbon loss to rehydration reaction and production of CO2). This means that fructose dehydration and condensation are the only two dominating reactions catalysed in tandem by SGQDs, which therefore gave this catalyst its selectivity.

Table 4. Subsequent comparative studies of different catalysts on the dehydration of fructose with initial fructose concentration of 20 wt% of aqueous phase of the biphasic solvent system

(using the reaction conditions adapted from the optimised method) ^c

cReaction condition: The initial fructose concentration with respect to the aqueous phase was 20 wt% and 40 mg of the catalysts was used. In these studies, the weight of the aqueous phase was 2.1 g, therefore 420 mg of fructose was used. The reactions were performed at 170 °C for 2 h in a biphasic solvent consisting of an aqueous phase (DMSO:water with v/v of 7:3) and an organic phase (MIBK:butanol with v/v of 7:3), with an aqueous to organic phase w/w ratio of 1 :2.

dTotal acidity was determined by titration.

eS03H density was determined by elemental analysis. Example 8. Biomass transformation of different carbohydrate feedstocks to 5-HMF by SGQDs

The SGQDs of Example 1 was further evaluated for biomass transformation of various carbohydrate feedstocks to 5-HMF, using the reaction conditions adapted from the optimised method in "Methods for catalyst testing and analysis of the products". The reactions were conducted using: 60 mg of catalysts; an initial fructose concentration of 10 wt% (with respect to aqueous phase); and in a biphasic solvent consisting of an aqueous phase (DMSO:water with v/v of 7:3) and an organic phase (MIBK:butanol with v/v of 7:3), with an aqueous to organic phase w/w ratio of 1 :2. Specifically, the weights of the aqueous phase and biomass were 2.1 g and 210 mg respectively.

As shown in Table 5, the dehydrations of inulin and xylose catalysed by SGQDs gave a 5- HMF yield of 52.3% and 18.3% respectively. Xylose is the major compound in hemi-cellulose, while inulin is a fructose polymer that can be obtained from chicory. Given this, the successful degradation of these compounds suggests the potential use of SGQDs in biomass transformation of these widely available feedstocks to useful products.

Sucrose is the most abundant disaccharide and it contains a glucose and a fructose molecule. The catalytic degradation of sucrose using SGQDs gave a 39.2% yield of 5-HMF, while the same reaction carried out on glucose gave a 5-HMF yield of 19.5% yield and a 45.1% conversion rate (entry 4, Table 5). In contrast to fructose dehydration which proceeds via a cyclic furanose intermediate pathway, glucose dehydration undertakes an acyclic intermediate pathway which is possibly mediated by Bransted acidic sites on SGQDs (X. Qian, et al., Carbohydr. Res. 2005, 340, 2319-2327). Due to the stable acyclic structure of the sugar intermediates, the energy barrier for 5-HMF production is much higher from glucose than from fructose. Interestingly, the yield of 5-HMF produced from the degradation of glucose by SGQDs was considerably higher than that obtained from commonly used SO3H- functionalised catalysts which gave yields typically below 5% (P. Wang, et al., Chem. Eng. Commun., 2016, 203, 1507-1514). Therefore, it is deduced that the Lewis acid sites on SGQDs catalysed the isomerisation of glucose into fructose with a low energy barrier and the Bronsted acidic sites subsequently converted the fructose into 5-HMF.

As compared to glucose, the degradation of cellobiose gave a slight decrease of 5-HMF yield (13.2%) but much higher conversion (91 .2%). Surprisingly, a good 5-HMF yield of 22.2 wt% and a conversion rate of 78.9% were produced from cellulose under atmospheric pressure.

Cellulose is the most abundant biomass resources and its decomposition can involve a cascade of complicated reactions (Y.L. Wang, etal., Nat. Commun., 2013, 4, 2141 ). Therefore, the ability of SGQDs to readily catalyse cellobiose and cellulose into 5-HMF is advantageous and is probably due to the abundant medium Bronsted acid sites on SGQDs (Fig. 6). These results demonstrate the unique potential of SGQDs in transforming non-edible lignocellulosic biomass into highly valuable compounds.

Table 5. Conversion of various carbohydrate feedstocks to 5-HMF by SGQDs, using the reaction conditions adapted from the optimised method'

Molar yield (%)

Conversion

Entry Substrate Condensation Rehydration

(%) 5-HMF

compounds compounds

1 Inulin 90.7 52.3 - 0.87

2 Xylose 41 .9 18.3 14.9 7.85

3 Sucrose 79.5 39.2 - 1.68 Molar yield (%)

Conversion

Entry Substrate Condensation Rehydration

(%) 5-HMF

compounds compounds

4 Glucose 45.1 19.5 4.68 1 .35

5 Cellobiose 91 .2 13.2 19.4 0.1 1

6 Ball-milled cellulose ⁹ 78.9 22.2 40.7 10.2

'Reaction conditions: The initial concentration of substrate with respect to the aqueous phase was 10 wt% and 60 mg of catalyst was used. Specifically, the weights of the aqueous phase and biomass were 2.1 g and 210 mg respectively. The reactions were performed at 170 °C for 2 h in a biphasic solvent consisting of an aqueous phase (DMSO:water with v/v of 7:3) and an organic phase (MIBK:butanol with v/v of 7:3), with an aqueous to organic phase w/w ratio of 1 :2.

⁹ The reaction was conducted in a 15 mL autoclave in the presence of 25 mg cellulose and 100 mg SGQDs under hydrothermal for 2 h at 170 °C. Only the yields of the products for entry 6 are in weight percentage (wt%). The yields for the remaining entries are in molar percentage.

Example 9. Comparative studies of the catalytic performance of MoGVSGQDs with other catalysts in converting fructose to DFF

The conversion of fructose to DFF using the M0O3/SGQDS of Example 1 , in comparison with other catalysts, were carried out in 10 mL of DMSO in the presence of 0.5 g of fructose and 10 mg of catalyst at 160 °C for 2 h (adapted from the preliminary method in "Methods for catalyst testing and analysis of the products"). The efficiencies of several representative catalysts for the one-pot transformation of fructose to DFF were determined and compared with that of the said MoOVSGQDs (Table 6). These catalysts include molybdenum oxides- modified materials based on: SBA-15 (M0O3/SBA-I 5); sulfonated carbon nanotubes (M0O3/SCNT) ; and activated carbon (M0O3/ Activated carbon). Vanadium oxides-modified sulfonated carbon nanotubes (V2O5/SCNT) were also used in this study. These catalysts were prepared according to a reported method with some modifications (Fabien L. Grasset, et al., RSC Advances, 2013, 3, 9942).

Among the catalysts tested, the M0O3/SGQDS which were modified with sulfonic groups and M0O3 active sites, exhibited the highest yield of 78% for DFF. The superiority of the M0O3/SGQDS can be attributed to their quasi-homogeneity, which allowed the reactants to access the catalytic active sites readily. In comparison, M0O3/SBA-I 5 which is a commonly used mesoporous catalyst, resulted in a lower conversion of fructose and poor selectivity for the DFF product. Further, M0O3/SCNT and MoOVActivated carbon also exhibited lower catalytic activities as compared to M0O3/SGQDS. The V ₂ 0 ₅ -modified sulfonated carbon nanotubes (SCNT) which are commonly used in the oxidation of 5-HMF to DFF, showed significantly lower DFF yield as compared to the M0O3/SGQDS (Fabien L. Grasset, et al., RSC Advances, 2013, 3, 9942). This can be attributed to the high oxidising power of the vanadium active sites (V ₂ 0 ₅ ), leading to the over-oxidation of DFF.

Table 6. Catalytic performance of different catalysts on the one-pot transformation of fructose to DFF, using the reaction conditions adapted from the preliminary method ^h

hReaction conditions: The reactions were conducted in 10 mL DMSO in the presence of 0.5g fructose and 10 mg catalyst at 160 °C for 2 h.

Example 10. Biomass transformation of cellulose to DFF by M0O3/SGQDS In addition, the said M0O3/SGQDS of Example 1 were also applied in the degradation of cellulose, which is the most abundant source of biomass in the world. The reactions were carried out in 5 mL of solvent in the presence of 50 mg of cellulose and 10 mg of catalyst at 160 °C for 2 h (adapted from the preliminary method in "Methods for catalyst testing and analysis of the products").

In the presence of M0O3/SGQDS, the cellulose conversion reached as high as 70.5% and a 30.5% yield of DFF was obtained using DMSO-butyrolactone (8:2 v/v) as the solvent (Table 7). The remarkable results of the M0O3/SGQDS suggests the feasibility of the use of such catalysts for industrial production of DFF from low-cost biomass. Table 7. Biomass transformation of cellulose to DFF by MoCVSGQDs, using the reaction conditions adapted from the preliminary method'

'Reaction conditions: The reactions were conducted in 5 mL of the solvent in the presence of 50 mg cellulose and 10 mg catalyst at 160 °C for 2 h.

Example 11. Recyclability of SGQDs for dehydration of fructose to 5-HMF

The recycling process of the SGQDs for subsequent reactions is depicted in Fig. 9a. After the completion of the catalytic reactions, the biphasic reaction mixture was left to stand, which then allowed phase separation 110 to take place. The 5-HMF or other organic compounds as- produced were extracted into the organic phase 140, while the SGQDs were homogeneously dispersed in the aqueous phase 150. The 5-HMF produced was collected by evaporating the organic phase, while the SGQDs were separated from the aqueous phase by dialysis and drying (120) to give the SGQDs powder 160. Subsequently, the resuspension (130) of SGQDs 160 in the solvent gave a homogeneous dispersion of the SGQDs 170, which was used for the next catalytic reaction.

After the initial studies that were conducted in 10 mL of DMSO in the presence of 0.5 g fructose and 10 mg of catalyst at 150 °C for 2 h, the reusability of the SGQDs was investigated (in Example 7, "Initial studies using the preliminary method"). Advantageously, these SGQDs can be reused for at least eight times without any obvious loss of catalytic activity (Fig. 9b), and this was significantly more superior than the commonly used heterogeneous catalysts (e.g. ZSM-5, AI-MCM-41 , AI-SBA-15). The recyclability study was also carried out for the SGQDs used in the optimised method (in Example 7, "Subsequent comparative studies using the optimised method"), in which various acid catalysts were tested at the same dosage of 40 mg with initial fructose concentration of 20 wt% (with respect to aqueous phase) and in a biphasic solvent system of 7:3 (v/v) DMSO:water and 7:3 (v/v) MIBK:butanol, with an aqueous to organic phase w/w ratio of 1 :2.

As shown in Fig. 9c, even after seven times of reuse, 36.8% of 5-HMF yield and 81.3% of fructose conversion were achieved using the spent SGQDs. This is likely due to the minimal loss of -SO3H functional groups (-0.85 mol% S loss) as revealed by element analyses. On the other hand, FTIR characterisation shows that -SO3H groups, Bronsted acidity, and Lewis acidity were still well-preserved in SGQDs after they had been reused for seven times (Fig. 10a and b).

Furthermore, the TEM and TEM-EDS characterisations of spent SGQDs indicated that the morphology, size and element contents of SGQDs were well-preserved (Fig. 1 1 a and b). The spent SGQDs also remained well-dispersed in the solvents even after several times of reuse, which suggests that the catalysts were not deactivated by the insoluble humins which may be produced during the conversion of carbohydrates to 5-HMF. Given the above, it was observed that the SGQDs of the current invention possess both good catalytic efficiency and recyclability, which serve the criteria of a good quasi-homogeneous carbocatalyst

Example 12. Mechanism and the role of -OH groups on SGQDs in dehydrating fructose to 5- HMF It was deduced that the good catalytic efficiency of SGQDs was due to the -OH groups on SGQDs which allowed the strong absorption of fructose via strong hydrogen bonding, while the -COOH and -SO3H groups released protons to activate the fructose.

Method

To identify the effect of -OH groups in SGQDs on the catalytic performance, methoxytrimethylsilane (MTS) was used to block the groups via the condensation of -OH in SGQDs with MTS (P.P. Upare, et al., Green Chem., 2013, 15, 2935-2943; A . Mursito, et al., IOP Conference Series: Earth and Environmental Science, IOP Publishing (2018), p. 012069).

In a typical procedure, 1.0 g of SGQDs was dispersed in 50 mL of toluene to form a suspension. 1.65 mmol MTS was added to the suspension and the resulting mixture was reacted at 70 °C for 12 h and then allowed to cool to room temperature. After filtration, the solid was washed three times with anhydrous ethanol, and finally dried overnight at 100 °C to give the MTS- modified SGQDs (designated as SGQDs-MTS). Results and discussion

The blocking of the -OH groups on SGQDs by MTS resulted in the decrease of the hydrophilicity index (HI) from 2.86 to 2.61 . In addition, the 5-HMF yield and conversion rate obtained from SGQD-MTS were drastically reduced to 29.1 % and 69.8% respectively, as compared to 51 .7% and 91.8% of SGQDs (Table 4). This experiment unambiguously confirmed the critical importance of -OH groups on SGQD. When the -OH groups were not present, the decrease in catalytic performance was attributed to: (i) weakened interactions between the SGQDs and fructose; and (ii) weakened Bronsted acidity from -SO3H groups due to their hydrogen bonding with fructose.