Recent developments in the biphasic
aqueous conversion of biomass-derived
sugars to 5-HMF.
Alexandra Wolters
Literature Thesis (12 EC)
Master Chemistry (joint degree UvA/VU)
Daily supervisor: Dr. Robert-Jan van Putten
Examiner: Prof. dr. Gert-Jan Gruter
Second examiner: Dr. Chris Slootweg
Research performed at UvA / HIMS / Industrial Sustainable Chemistry group
Date of publication Thesis: 13-10-2020
2
Abstract
The conversion of biomass into platform chemicals is gaining a lot of attention. One such platform molecule is 5-hydroxymethylfurfural (HMF), which can be produced by the acid-catalyzed dehydration of sugars. In aqueous medium HMF yield has a maximum around 50% due to side reactions such as the acid-catalyzed hydration of HMF to levulinic acid (LA) and formic acid (FA). One way to minimize these side reactions is to use a biphasic system, consisting of an aqueous phase with the reactants and acid catalyst and an organic phase for in-situ extraction of the formed HMF. This thesis discusses recent mechanistic insights as well as developments in the process chemistry of fructose and glucose dehydration to HMF. Glucose conversions of >90 percent are not uncommon, and HMF yields of >80% are reported. None of the articles provide an isolated yield for HMF and all research is done on small (~5 mL) scale.
3
Contents
1. Introduction 4
2. Mechanism and kinetics 6
3. HMF formation in biphasic systems 16
4. Conclusions 25
4
1. Introduction
Oil is an important resource for fuel and the chemical industry, and the world is highly dependent on it. But the consistent use of crude oil has caused a rapid depletion of the reserves, as well as had a devastating influence on the earth’s ecosystem and environment.1 Biomass is regarded as a
promising alternative to fossil fuels both for the production of fuels and chemicals. Biomass is cheap, widely available and a large natural source of carbohydrates. There is a lot of attention for the conversion of biomass into platform molecules, which then have the potential to be converted into a wide array of end-products.2,3 Specifically, the conversion of biomass to platform chemical
5-hydroxymethylfurfural (HMF) is receiving a lot of interest.4,5 HMF can be derived from cellulose
without breaking a single carbon-carbon bond (see figure 1), and can be transformed into a wide variety of useful compounds (see figure 2).5,6,7
Figure 1 HMF and cellulose, corresponding C-C bonds are coloured red
Carbohydrate feedstocks contain a large amount of oxygen, which needs to be removed in order to create higher value products. One approach for the removal of oxygen is the dehydration of sugars, which can yield HMF. However, there are two important bottlenecks in the dehydration of sugars: low reactivity and selectivity. For the dehydration of fructose to HMF relatively high yields have been reported: up to 50% yield for single phase aqueous systems, up to 70% yield in aqueous biphasic systems and even higher yields (>80%) for ionic liquids and polar aprotic solvents.5 However, fructose
has to be produced from biomass via the isomerization of Glucose first. Therefore, research has shifted to start the production of HMF directly from biomass-derived cellulose and glucose. Unfortunately, yields and selectivities in these processes are generally much lower.
The acid-catalysed dehydration of sugars has been attempted in a wide range of different solvents and with many different catalysts.5 Water is regarded to be a sustainable and
environmentally friendly solvent. It is abundant, cheap and non-toxic. However, in aqueous medium HMF selectivity is low due to the formation of undesirable byproducts.8 HMF can undergo
rehydration to form levulinic acid (LA) and formic acid (FA). Furthermore, sugars, HMF and other reaction intermediates react to form humins, insoluble polymers.9 With regards to yield and
selectivity, ionic liquids and aprotic polar solvents such as DMSO perform by far the best.5,6,10 These
solvents stabilize HMF and thus prevent its rehydration and other side reactions of HMF. However, the isolation of HMF from these solvents proves difficult. 8,11 An important drawback of the
stabilization effect is that it is very difficult to extract HMF from these solvents. Furthermore, both DMSO and ionic liquids are difficult and energy-intensive to evaporate.
Many of the selectivity issues in aqueous mediums can be addressed by employing a biphasic system, which combines organic solvents with water for the in-situ extraction of the formed HMF.5
The extraction of HMF into the organic phase is another tactic to decrease the rehydration of HMF. HMF hydration to LA and FA is acid-catalyzed in water, as is de dehydration of sugars to HMF. This means that when the rate of HMF production is increased, the rate of HMF degradation is also
5 increased. The biphasic system can prevent this to a certain extent. Furthermore, HMF is thought to react with sugars and dehydration intermediates to form humins. This process can also be decreased by a biphasic system, since sugars are not soluble in the applied organic solvents.
Figure 2 potential end products from platform molecule HMF
Due to the importance of scaling-up, easy separation and sustainability, this review will focus on developments over the past decade in aqueous biphasic systems. First, the mechanistic aspects of and some kinetic insights into the conversion of glucose and fructose into HMF are discussed. Then, a summary of interesting new developments in reaction systems for the dehydration reaction is provided. We will review these developments with the possibilities for scaling up and sustainability in mind.
6
2. Mechanism and Kinetics
The conversion of biomass-derived molecules to HMF encounters several bottle-necks, in both selectivity and conversion. Therefore, it is important to know the mechanism of the reaction for optimization purposes, and minimization of by-products. Different mechanisms were found for reactions in DMSO or ionic liquids than for reactions in water, indicating a solvent-dependence.5 In
this paper we will only discuss mechanisms for the reaction in water.
HMF from fructose
HMF can be formed from Fructose by a triple dehydration reaction. As discussed by Van Putten et al., up until 2012 the postulated mechanisms for fructose dehydration could be divided between pathways with cyclic and acyclic intermediates.5 Both mechanisms are shown in figure 3. Currently,
the consensus in literature is that the formation of HMF from fructose proceeds via cyclic intermediates, even though the exact steps that take place have not yet been definitively determined. Some interesting research into the mechanism is reported below.
Figure 3 (a) cyclic and (b) acyclic mechanisms for fructose dehydration to HMF.5
Van Putten and co-workers studied the aqueous dehydration of different ketoses with sulfuric acid as catalyst to investigate the influence of the orientation of the hydroxyl groups on HMF formation.12
The experimental results from high-throughput screening, combined with DFT calculations, showed that different ketoses have different reactivities in the dehydration to HMF, which are most likely due to the orientation of the hydroxyl groups on C3 and C4. The influence of this orientation is best explained by a mechanism via cyclic intermediates. Furthermore, they explain that the tautomeric distribution of ketohexoses in water13 does not explain the difference in reaction rate measured for
7 the different ketoses and that the rate of tautomerization14 under acidic conditions at the reaction
temperature (100, 120 and 140°C) will be much higher than that of the dehydration. Therefore, the tautomeric distribution is not affecting the rate of dehydration.
Conclusions from prior computational research15,16, which suggested hydride transfers from
C1 to C2 and C5 to C4 to be key steps in the dehydration of fructose, were the basis of a series of isotopic labelling experiments by Swift et al.17 They compared kinetics experiments for the
dehydration of C1 deuterated fructose with those for non-labeled fructose, in aqueous medium with added HCl and KCl. The rate for fructose conversion was shown to be significantly lower for deuterium-labelled fructose then regular fructose. This indicates, according to the kinetic isotope effect (KIE), that the rate limiting step is the hydrogen transfer from C1 to C2. HMF yields show a similar trend, reinforcing this mechanism. The proposed mechanism, show in figure 4, comprises of two steps. First, a reversible protonation/dehydration step takes place, followed by the intramolecular C1-C2 hydride shift. The following steps are assumed to be fast and irreversible. They explicitly mention the formation of the intermediate, because the rate of the dehydration was found to be first-order dependent on proton concentration and the hydride shift is not acid-catalysed. Proposed is that the dehydration rate depends on the concentration of the intermediate, and thus on the proton concentration. Interestingly, they report that the tautomeric distribution of fructose has to be included in the kinetic model, otherwise the apparent rate constant of fructose dehydration would be underestimated. This conflicts with the results of Van Putten et al12, described
above. However, they mention that the effect of including the tautomeric distribution decreases with temperature, as the fraction of fructofuranose increases. The difference in fitted rate constant decreases from 2.25 at 90°C to 1.8 at 130°C. They do not mention the factor for temperatures of 160°C and up, that are more generally used for fructose dehydration.5
Figure 4 The first step shows the reversible protonation/dehydration step takes place. The second step includes the proposed rate determining intramolecular C1-C2 hydride shift followed by other fast and irreversible steps.17
HMF from Glucose
For the conversion of glucose to HMF, it is generally assumed that glucose isomerizes to fructose, which is consequently dehydrated to form HMF. 18 However, there is also evidence that points to a
different reaction path involving the direct dehydration of glucose to form HMF. The glucose-fructose isomerization proceeds though a ring-opening and an acyclic intermediate, which is why this mechanism is, somewhat confusingly, referred to as acyclic. The successive fructose dehydration is still assumed to proceed though cyclic intermediates. The direct dehydration of glucose proceeds exclusively through cyclic intermediates and is therefore referred to as the cyclic mechanism.
Acyclic mechanism: via glucose-fructose isomerization
The glucose-fructose isomerization is an important bottleneck in this reaction, due to equilibrium limitations. A one-pot synthesis from glucose to HMF is a tactic to overcome these equilibrium issues.18 However, the glucose-fructose isomerization has long been regarded as enzymatically or
8 that Lewis Acids are also able to catalyse the glucose-fructose isomerization19, opening up more
possibilities for glucose to HMF one-pot tandem reactions.
Zeolites (Sn-beta or Ti-beta) have been reported as efficient Lewis acid catalysts for the isomerization of glucose.19–21 Therefore, they have also been applied in the transformation of glucose
to HMF. In 2010, the group of Davis showed that the zeolite Sn-Beta catalyses the isomerization of glucose to fructose via a 1,2-hydride shift, as opposed to a proton transfer.22 They reacted a 10 wt%
aqueous solution of glucose with Sn-Beta zeolites in a 1:100 Sn:glucose molar ratio. After 45 min at 110°C yields were achieved of 46 wt% glucose, 31 wt% fructose and 9 wt% mannose. Glucose was labelled with deuterium as shown in figure 5. 13C-NMR showed that there was no D/H-exchange taking place under reaction conditions and that the C2-deuterium ended up at C1. In a proton transfer mechanism (see figure 6) the hydrogen is transferred via a base, which would lead to proton exchange with the aqueous solvent. Because the deuterium ends up on fructose C1, and not on water molecules, this supports that the glucose-fructose isomerization takes place via an intramolecular hydride shift.
Figure 5 Conversion of D-labelled glucose to fructose, including proposed transition state showing the 1,2-hydride shift.22
Figure 6 Base-catalyzed proton transfer vs. hydride shift mechanism.22
In 2013, Choudhary et al showed that CrCl3 combined with HCl in solution functions as a
Lewis acid-Brønsted base bifunctional catalyst.23 They achieved 59% HMF yield and 96% glucose
conversion in a biphasic system consisting of NaCl-H2O and THF (1:2). They combined kinetics and a
CrCl3 speciation model to determine that the hydrolysed Cr(III) complex [Cr(H2O)5OH]2+ was the most
active in the isomerization of glucose to fructose. Furthermore, they show a strong interaction between the metal ion and glucose, using X-ray absorption fine structure spectroscopy. Furthermore, they discuss the presence of complex interactions between CrCl3 and HCl. Firstly, the Brønsted acid
9 glucose-fructose isomerization. Secondly, the presence of the Lewis acid in addition to the Brønsted promotes side reactions from fructose, thus increasing fructose consumption while decreasing selectivity to HMF. Interestingly, the catalysis proceeded even in the absence of HCl. The authors note that the dehydration of CrCl3 leading to the active complex, acidifies the solution and thus the
Lewis acid contains an intrinsic Brønsted acid functionality in water. Even though the hydrolysed Cr(III) complex promotes the reaction without the addition of HCl, the optimum results were reached with a combination of both catalysts. This makes perfect sense, since a Lewis acid in water will increase the amount of free protons, and thus the acidity. In the formation of the mentioned active species [Cr(H2O)5OH]2+ a proton has to be formed, according to the following reaction equation:
CrCl3 + 6 H2O → [Cr(H2O)5OH] + 3 Cl- + H+
The functional species being [Cr(H2O)5OH]+2 is supported by Car–Parrinello molecular dynamics
simulations carried out by Mushrif and co-workers.24 Their proposed mechanism is shown in figure 7.
On the left, the mechanism for the glucose ring-opening is shown. First, glucose replaces two water molecules and coordinates to the Cr(III) centre. Then, the hydroxyl coordinated to the metal centre is deprotonated, which enhances the interaction between Cr(III) and glucose. Simultaneously, the hydroxyl group coordinated to the catalyst complex is protonated, which lowers the energy barrier for this step significantly. The authors describe this as the proton being transferred through an intermediate water molecule. The ring opening is completed by transfer of the proton, mediated by water molecules, to the ring oxygen. This mechanism seems logical. However, even though glucose in water is 99.9% in cyclic form, it still actively interchanges between different anomeric forms. This process is also called mutarotation, can be acid- base- or even water-catalysed, and includes a ring-opening. SO the catalyst is probably not essential for this step. The right half of Figure 7igure 7 shows the deprotonation of the C2 hydroxyl group, again through an intermediate water molecule. Subsequently, the hydride shift takes place, which is the rate limiting step in the reaction. Finally, the formed fructose is protonated again (ring-closing) and dissociates from the metal complex, leaving [Cr(H2O)5OH]+2.
Li et al. investigated a wide array of silica-alumina composite (AlSiO) bifunctional catalysts in the transformation of glucose to HMF in a water/THF biphasic system.25 They achieved 58% HMF
yield at 81% glucose conversion. They found that weak Lewis acid sites in the catalyst promote the formation of HMF, while medium to strong lewis acid sites tend to prefer the formation of byproducts such as levulinic acid and humins. Furthermore, they report a volcano-like correlation between HMF selectivity and the amount of Brønsted acid sites relative to Lewis acid sites. This is to be expected because both the rehydration of HA to FA and LA and formation of humins are acid-catalyzed. So an increase in acidic sites leads to a higher rate of HMF production, but also to a faster rehydration of HMF.
Nguyen et al. studied the metal salt catalysed reactions of glucose in water using isotopic labelling experiments combined with 1H and 13C NMR analysis26, in a similar way as the group of Davis
did for zeolites as discussed above. They show that MCl3 salts catalyse the glucose-fructose
isomerization via a 1,2 hydride transfer mechanism. Additionally, they showed that the homogeneous metal salts catalyse the reversible epimerization of glucose to mannose. The biggest portion of mannose is formed through the formation of fructose, which is seen as a C1 mannose peak in the 13C NMR spectrum. However, in the 13C NMR spectrum there is also a peak visible for C2
base-10 catalysed (with NaOH) glucose-fructose isomerization, while this reaction does show C1 mannose in the NMR analysis. The proposed reaction scheme is show in figure 8. With GaCl3 as catalyst and
starting from 10 wt% glucose, they achieved 77.6% glucose, 17.6% fructose, and 2,4% mannose, split up in 2% 13C-1 mannose and 0.4% 13C-2 mannose yields. For other ML
3 catalysts similar results were
achieved.
Figure 7 Proposed mechanism for glucose ring opening (left) and isomerization to fructose (right).24
Figure 8 Summary of different reaction pathways from Glucose in water, catalysed by zeolites or metal salts. All reactions were shown to take place, however the largest yield was glucose, followed by fructose. Other products were found in small quantities (<5%). 26
11 The analysis of experimental data from the aqueous conversion of glucose to HMF with a CrCl3-HCl
system showed that the HMF yield is maximized for different CrCl3 concentrations at a moderate HCl concentration.27 The maximum HMF yield achieved was 15%, with 80% glucose conversion. In a
biphasic reactor (NaCl-H2O/THF ratio 2:1) a HMF yield was achieved of 60% with a glucose conversion
of 90%. The type of optimum yield curve as is shown in figure 9, is referred to as volcano-like behaviour, and points to interactions between HCl and CrCl3 in solution, or the speciation of the active complex, that are relevant to the rate and specificity of the reaction. With the previously described results in mind (the active catalyst complex is [Cr(H2O)5OH]+2), the latter seems more likely.
They did not test other acids in place of HCl. Some interesting points can be concluded from figure 9. A higher concentration of HCl leads to a higher rate of HMF formation, but not a higher yield, which can be attributed to the fact that a higher acid concentration also increases the rate of HMF hydration. Without HCl the reaction also proceeds, which can be explained by the in-situ formation of protons by the Lewis acid in water. Furthermore, for a higher concentration of CrCl3, the optimum
concentration of HCl is also higher then for a lower concentration of HCl. The combination of Lewis and Brønsted acids is effective to overcome equilibrium-limitations in the isomerisation of glucose and leads to high glucose conversions. In figure 10, more volcano curves are shown, which indicate that glucose is first transformed into fructose, which is the dehydrated to HMF. The HMF yield in a single phase aqueous system remains low (<20%) but a high potential for an increased yield is found in biphasic systems.
Extensive investigations into the reaction network of the transformation from Glucose to HMF in a biphasic aqueous NaCl/THF system with AlCl3 as catalyst were performed by Tang and
coworkers.28 They achieved 28% HMF yield for 80% glucose conversion. They designed a kinetic
model based on systematic kinetic experiments, verified by a parity plot and a statistical significance analysis of the kinetic parameters. The proposed reaction network is shown in figure 11. It was figured out which reactions were promoted by which catalyst in the bifunctional catalyst system. The bifunctional catalyst system consisted of species [Al(OH)2] and HCl, which are formed in-situ by the
reaction between AlCl3 and H2O. It was found that the Brønsted acid HCl catalyses the
polymerization of HMF to humins and the rehydration of HMF to LA and FA. The Lewis acid active species [Al(OH)2(aq)] catalyses the glucose-fructose isomerization and the rehydration of HMF to FA
and LA. The dehydration of fructose to HMF is catalysed by both the Lewis and Brønsted acids, as is the direct polymerization of glucose and fructose to humins. This not unexpected, since a Lewis acid in water will also generate protons, as does a Brønsted acid.
12
Figure 9 (a) maximum HMF yield and (b) time to reach the maximum HMF yield for varying CrCl3 concentrations as a
function of HCl concentration.27
Figure 10 Different HCl concentrations in 5mM CrCl3 at 130°C. (a) glucose conversion, (b) fructose yield, and (c) HMF yield as a function of time.29
Figure 11 Proposed Reaction Network for the AlCl3-Catalyzed Conversion of Glucose in NaCl–H2O/THF.28
Cyclic mechanism: direct dehydration of Glucose
Qian and coworkers published a series of papers contradicting the acyclic mechanism an proposed a cyclic mechanism (shown in Figure 12 12) for the acid-catalysed glucose conversion based on Car-Parrinello molecular dynamics simulations (CPMD) coupled with metadynamics (MTD) simulations.30– 32 However, these models only include a Brønsted acid catalyst, where research described above
includes Lewis acids, which generally yield much better results in catalysing the formation of HMF from glucose. The basis of this mechanism is that the hydroxyl groups of glucose form hydrogen bonds with water molecules, which facilitates proton transfer.31 The change of the relative stability of
13 and open paths to the cyclic mechanism. They propose a rate-limiting step consisting of protonation of the C2 hydroxyl group, breaking of the C2-O2 (a) and C1-O5 (b) bonds, and the formation of the C2-O5 bond (Figure 12 12). Fructose is not an intermediate in this reaction path. As mentioned before, glucose in aqueous solutions rapidly interchanges between its different anomers, via a ring-opening reaction. This ring ring-opening takes place at the C2-O2 bond, that is also breaking in this proposed mechanism. However, this bond then closes, after some possible rotations, resulting in circular glucose again.
Qian and Wei have also computed the Brønsted acid catalysed mechanism for glucose-fructose isomerization.32 They agree that this reaction takes place via a 1,2-hydride shift mechanism,
but their proposed reaction path differs from what we have seen before (see figure 13). They propose that first the mechanism follows that of HMF formation, and after formation of the C2-O5 bond, a 1,2-hydride shift takes place, followed by rehydration of C2 to form fructose. The 1,2-hydride shift was calculated to be the rate-limiting step. The ’HMF intermediate’ is widely regarded to be an intermediate in the Brønsted acid catalysed conversion of fructose to HMF.
Figure 12 Proposed cyclic mechanism for conversion of glucose to HMF. 31
Figure 13 Proposed mechanism for the Brønsted acid-catalysed isomerization of glucose to fructose.32
Supporting evidence for the cyclic reaction path was provided by Yang et al. 33 They
combined DFT calculations with isotopic tracing studies using NMR to create a reaction network of the Brønsted acid catalysed reactions of glucose, which is shown in figure 14. They agree with Qian
14 that the first step in the formation of HMF is protonation of the C2 hydroxyl group of glucose. Then, they propose a first dehydration of glucose as the rate limiting step, followed by a reversible reaction to form the HMF intermediate, from which HMF could be formed (see figure 15).
Figure 14 Reaction path analysis for the Brønsted-acid catalysed dehydration of glucose to HMF and byproducts.33
Figure 15 Proposed mechanism of fructose dehydration.33
Conclusion
Currently, the main consensus in literature is that the dehydration of fructose to HMF proceeds via cyclic intermediates, including a 1,2-hydride shift as rate-limiting step. The hydride shift mechanism had been confirmed by isotopic labelling experiments including KIE effect experiments and 1H and 13C
15 The dehydration of glucose is believed to proceed through acyclic isomerization to fructose followed by the cyclic HMF-formation. This is called the acyclic mechanism. The isomerisation of glucose to fructose is Lewis-acid catalysed, while the dehydration of fructose is Brønsted acid catalysed. This distinction is not always clear, since a Lewis acid in water will also form protons and thus exhibit Brønsted acid activity. The reaction is found to proceed best with a moderate amount of acid, since acid not only catalyses the fructose dehydration but also the HMF rehydration resulting in the unwanted by-products levulinic acid and formic acid. The optimal amount of Brønsted acid forms a volcano-like curve with respect to the amount of Lewis acid. The HMF yield seems to be optimal for a wide range of conditions. For example: with a lower amount of catalyst the same yield might still be reached after a longer reaction time, and a shorter reaction time for the same yield could be achieved by increasing the temperature.
A few reports conclude that the dehydration of glucose follows the cyclic, or direct glucose dehydration mechanism, involving direct conversion of the pyranose ring to a 5-membered ring intermediate. However, this research was done for Brønsted acid catalysts, where the other mechanistic studies used Lewis acid catalysts, or a combination of the two, because Brønsted acids alone are not active for the glucose dehydration. Furthermore, glucose is know to spontaneously go through ring-opening reaction in solution. It is also know from practical experiments, such as the one in figure 10, that fructose is formed from glucose before the formation of HMF. Therefore, the acyclic mechanism is more likely to be the most prevalent. Possibly a small amount of glucose also reacts via the cyclic pathway.
16
3. HMF formation in biphasic systems
It has been shown that conversion has become less of a problem, but selectivity and yield of HMF still remain a bottleneck. After enough time or at a high enough concentration with the addition of a simple catalyst, fructose and glucose can be almost fully conversed. In recent publications, conversions of fructose as well as glucose of over 95% are not uncommon. However, due to the formation of byproducts, mainly formic acid, levulinic acid and humins, HMF yield and selectivity remain lower. Tang et al comprehensively showed the added value of the biphasic system.28 They
compared the AlCl3 catalysed conversion of glucose to HMF in water, water/THF and water-NaCl/THF
systems. The results are shown in figure 16. The biphasic systems have a slightly better conversion of glucose, but the biggest difference can be seen in the HMF yield and selectivity. The biphasic system is already yielding significantly better results than the reaction in water (27% HMF yield instead of 12%), the addition of NaCl to the aqueous phase increases the yield and selectivity even more (to 45% HMF yield). There are several reasons for this effect, which we will discuss below.
Figure 16 Comparison of AlCl3-catalyzed glucose conversion in the three solvent systems: (a) glucose conversion, as a
function of time; (b),(c),(d) fructose yield, HMF yield and HMF selectivity, as a function of glucose conversion. Reaction conditions: 1.0 mL H2O, 0.25 M Glc, 25.0 mM AlCl3, 413 K. In the NaCl–H2O/THF biphasic system, NaCl is at saturated
concentration. V(THF):V(H2O) = 3. Figure adapted from reference.28
Multiple factors have been identified in the optimization of these biphasic systems, among others the choice of organic solvent and the addition of certain alkali-metal salts. These factors will be shortly discussed below, after which we will describe some examples of recent advances in the field, which are also summarized in tables 1, 2 and 3.
Choice of organic solvent
The choice of extraction solvent has several arguments. First, the miscibility gap with water is important. At reaction temperature (around 160°C - 200°C) and work-up temperature (usually 0°C or RT) the phases should not mix. Furthermore, the distribution of HMF between the aqueous phase and organic phase, or partition coefficient, should be optimized. A high amount of HMF should be in the organic phase to limit the hydration of HMF and formation of humins, which decrease HMF yield. The yield of HMF is generally seen to increase with an increasing amount of organic solvent. However, a larger amount of organic solvent will cost more time and energy to fully evaporate, which is necessary to isolate HMF. A low boiling point and high vapour pressure is also beneficial for a lower energy work-up. Finally, the choice of organic solvent should consider health and safety and environmental issues.
Esteban and co-workers used COSMO-RS and selection guides to create a top-15 of organic solvents that are promising to use in biphasic systems for the production of HMF, for they are expected to a have a high distribution coefficient of HMF with respect to water.8 Within these 15
17 are methyl propionate, ethyl acetate, n-propyl acetate, isopropyl acetate and MIBK. MIBK is ketone, but the other four recommended solvents are esters. This could be problematic in the biphasic HMF synthesis, because esters readily hydrolyse in acid aqueous media, which is the other phase in the biphasic system. MIBK is already widely applied in HMF synthesis.
Tang et al included extraction solvent THF in their kinetic experiments for the AlCl3 catalysed
conversion of glucose to HMF.28 They found that THF not only functions as extraction solvent, but
also influences the kinetics of the formation of HMF. THF suppresses side-reactions from HMF such as rehydration and polymerization, and obstructs the transformation of fructose to HMF and humins. It is of course imaginable that the reaction products are influenced because of the immediate extraction of HMF into the THF phase. Side reactions such as the formation of humins and the hydration of HMF to FA and LA are hindered because they are acid-catalyzed and require water. The authors also find that THF promotes the formation of FA from glucose and fructose over the formation of LA (see figure 17). They conclude this shows that the effect of the extraction solvent goes further than just miscibility with water and HMF extraction coefficients. However, it is important to note that the authors specifically mention that the HMF yield is measured in both the aqueous and organic phase and they do not mention this for the LA yield. HMF is a moderately polar solvent and Levulinic Acid will dissolve in polar organic solvents. Thus, it could be possible that part of the LA yield is missed in the analysis because it is in the organic layer. This is supported by the fact that the kinetic model points to THF as a reason for the production of LA over FA.
Figure 17 Comparison of the ratio of FA yield : LA yield, as a function of glucose conversion, for the AlCl3-catalyzed glucose
conversion in the three solvent systems. Reaction conditions: 1.0 mL H2O, 0.25 M Glc, 25.0 mM AlCl3, 413 K. In the NaCl–
H2O/THF biphasic system, NaCl is at saturated concentration. V(THF):V(H2O) = 3. Figure adapted from reference.28
Addition of alkali-metal salts
In biphasic systems often a significant amount of alkali-metal salts is added to the aqueous phase, often at saturation concentration.18 This has often been shown to significantly increase sugar
conversion and HMF yield. In general, this effect is attributed to two different effects. Firstly, it decreases the solubility of HMF in the aqueous phase. This is caused by interactions between HMF and the salt, and water and the salt. It promotes HMF to solvate more in the organic phase. Secondly, the salt increase the miscibility gap of the reaction and extraction solvents. More specifically, it increases the upper critical solubility temperature, increasing the reaction temperature at which the two phases are still separated.
Often the added salt is NaCl or KCl. However, Wrigstedt and coworkers reported that NaBr or KBr enhance the HMF yields for the Brønsted-acid catalyzed dehydration of fructose.34 In their
reaction (fructose dehydration with HCl, using different extraction solvents) the bromide salts give a 6-8% higher yield than the sodium salts. However, it is also mentioned that for different catalysts and biphasic systems, different salts give the best results.
18
Substrate Substr. conc.
(wt%) Reaction solvent Extraction solvent Org/aq phase ratio (v/v) Temp (°C) Reaction time
Yield (%) Conversion (%) Selectivity (%) ref
Fructose 45 Water MIBK 4 160 2 h 74 97 76 35
glucose 45 water MIBK 4 160 2 h 6 36 17 35
19 Substrate Substr. conc. (wt%) Reaction solvent Extraction solvent Modifiers Org/aq phase ratio (v/v) catalyst Catalyst loading Temp (°C) Reaction time Yield (%) Conversion (%) Selectivity (%) ref
Glucose 4 Water THF NaCl 10 ZnCl2/HCl 27 mol% 180 60 min 62.7 - - 36
Glucose 4 Water THF NaCl 10 AlCl3/HCl 27 mol% 180 55 min 66.9 - - 36
fructose 10 water MeCN KBr 2 HCl 0.05 M 160** 1 min 90 99 91 34
50 water MeCN KBr 2 HCl 0.05 M 160** 1 min 77 98 79 34
fructose 10 water MeCN KBr 2 HCl 0.05 M 160 40 min 84 95 88 34
Fructose 10 Water MIBK 19 HCl 0.046 M 177 5 min 70 100 70 37
Fructose 10 Water MIBK 19 HCl 0.046 M 177** 5 min 68 100 68 37
Glucose 10 Water MIBK 19 HCl 0.046 M 177 240 min 35 100 35 37
Glucose 10 Water MIBK 19 HCl 0.046 M 177** 120 min 51 100 51 37
Cellulose 10 Water MIBK 19 HCl 0.046 M 177 240 min 34 100 34 37
Cellulose 10 Water MIBK 19 HCl 0.046 M 177** 100 min 43 100 43 37
Glucose 4 water THF NaCl (26
wt%)
3 AlCl3 0.025 M 140 3 h 28 79 35 28
Glucose 2 water MIBK NaCl (20
wt%)
4 AlCl3 & HCl 40 mM
(both)
160 16 min 66 83 80 38
Table 2 Recent advantages in the conversion of fructose, glucose and cellulose to HMF in aqueous biphasic systems using homogeneous catalysts. Substrate concentration is calculated in the reaction solvent phase. *data between brackets is for regenerated catalyst. **microwave assisted heating
20 Substrate Substr. conc. (wt%) Reaction solvent Extractio n solvent Modifiers Org/aq phase ratio (v/v) catalyst Catalyst loading Temp (°C) Reaction time Yield (%) Conversio n (%) Selectivit y (%) ref Fructose 6 water 2-butanol 1.5 Phosphoric carbon 0.8 wt% 160 3 h 81 96 84 39
fructose 17 water DMC 3 (Ce(PO4
)1.5(H2O)(H3O)0.5(H2O)0.5)
Cerium phosphate
7 wt% 150 6 h 68 73 93 40
Glucose 9 water THF NaCl 3 Nb0.15Al0.85Si25O 9 wt% 160 1.5 h 58 81 72 25
Glucose 9 water MIBK 2.3 mesoporous
10Al-MCM-41 catalyst
3.2 wt% 195 150 min 36 87 41 41
Glucose 9 water MIBK NaCl (20
wt%)
2.3 mesoporous
10Al-MCM-41 catalyst
3.2 wt% 195 30 min 62 (47)* 96 (98)* 65 (48)* 41
Glucose 9 water MIBK NaCl (20
wt%)
2.3 H-beta Zeolite 3.2 wt% 195 30 min 56 (21)* 95 (53)* 59 (40)* 41
glucose 10 water GVL KBr 2 Amberlyst-38 (wet) +
CrCl3.6H2O 21 mg in 1.5 ml aq. KBr + 10 mol% 160** 3 min 74 98 76 34 glucose 3 Water/NMP (1:4)
MIBK NaCl 2.3 RCP160M:Al2O3 (1:2)
(B/L)
10 wt% 120 480 min 85 94 90 42
Glucose 9 water THF NaCl 4 TiO2-ZrO2/
amberlyst-70 (1:1)
4 wt% 175 3 h 86 100 86 43
cellulose 9 water THF NaCl 4 TiO2-ZrO2/
amberlyst-70 (1:1)
4 wt% 180 3 h 26 42 61 43
Table 3 Recent advantages in the conversion of fructose, glucose and cellulose to HMF in aqueous biphasic systems using heterogeneous catalysts. Substrate concentration is calculated in the reaction solvent phase. *data between brackets is for regenerated catalyst. **microwave assisted heating
21
Recent advances in the conversion of fructose, glucose and cellulose to HMF using aqueous biphasic systems
Ma and coworkers reported surprisingly good results for the autocatalytic production of HMF in a water-MIBK (1:4) biphasic system35. Aqueous solutions of 45 wt% fructose or glucose combined
with MIBK were heated to 160°C for 2 hours to yield 74% and 6% HMF yield with 76% and 17% selectivity respectively, without the addition of any catalyst. The reactions were performed under nitrogen atmosphere. The analysis was performed by separating the organic and aqueous layers and processing both separately. The aqueous phase was immediately diluted with water, while the organic solvent was evaporated first before water was added. Analysis was done using HPLC methods with an external standard. The extraction ratio (amount of HMF in org phase/in aqueous phase) was 85%. HMF could be isolated by adding NaOH to neutralize the system and remove LA and FA from the MIKB layer. Then, the MIKB layer could be separated, and three additional extractions with MIKB ensured a maximum amount of HMF was removed from the aqueous layer. The organic solvent could be evaporated, yielding HMF with a purity of >98% according to the 1H-NMR spectrum. The isolated
HMF yield is not mentioned in the article, so there is no information about how effective this work-up is quantitatively. The results are definitely notable. However, they were not the first to investigate the autocatalytic HMF production. This mechanism for the dehydration of fructose has been reported before by Ranoux et al44 for a single-phase aqueous system. They propose that the acid
hydration products of HMF, especially FA, are responsible for catalysing the reaction. They support this by a measured drop in pH during the reaction, and the prevention of HMF formation in several concentrated buffers with a pH around 7. They obtained a yield of 43% with 61% selectivity, for 30 wt% fructose in water at 190°C for 40 min. Similar results have been published before, though not often.5
Yang et al employed phosphoric carbon solid acids as catalysts for the dehydration of 6 wt% fructose in a H2O : 2-butanol (1 : 1.5) biphasic system.39 Multiple catalysts were screened with an
increasing amount of phosphoric acid. The authors found that even though the highest HMF yield (75% yield, 78% selectivity) was achieved with the catalyst with the highest acid density (3,36 mmol/g), the best selectivity (74% HMF yield, 83% selectivity) was achieved with a much lower acid density of 0.81 mmol/g. The highest HMF yield also corresponded to the highest measured LA yield. This is in agreement with the other described research where a moderate amount of acid is preferred, because an increased amount of acid speeds up the dehydration of fructose as well as the hydration of HMF, which results in FA and LA. Optimization of reaction conditions lead to the achievement of a maximum HMF yield of 81% with a fructose conversion of 96% after 3 hours at 160°C. The catalysts could be recovered and reused after washing with water and drying at 60°C for 12 hours . After 4 cycles the conversion of fructose decreased with 15% and the HMF yield with 30%. The catalyst was shown to ‘leak’ acid sites. The results with this catalyst are extraordinarily high, and out of line with the other research encountered. They do use a relatively low fructose concentration. The yields were determined using chromatography, the authors do not mention whether they used an internal or external standard. Because the phosphoric carbon catalyst contains Lewis acid sites, it would be interesting to see this catalyst applied to glucose dehydration.
Dibenedetto and coworkers were the first to use organic carbonates (OCs) as HMF-extracting phase.40 More specifically, they used a water:DMC (dimethyl carbonate) system with a 1:3 ratio. The
maximum HMF yield of 68% was achieved at 73% fructose conversion with 93% selectivity, reacting 17 wt% fructose with 7 wt% of a cerium phosphate catalyst ((Ce(PO4)1.5(H2O)(H3O)0.5(H2O)0.5)) for 6
22 and drying under nitrogen air flow. After 5 cycles the HMF yield decreased with 18%. This is due to adsorption of polyol to the catalyst surface, and this effect can be completely reversed by calcination of the catalyst at 550°C.45
Rodrigues Gomes et al compared the homogeneous catalysts ZnCl2 and AlCl3 combined with
HCl in a biphasic system consisting of H2O and THF.36 The aqueous phase was almost saturated with
NaCl, and the organic phase was 10 times the volume of the aqueous phase. 4 wt% glucose could be converted by 27 mol% ZnCl2 or AlCl3 in an hour at 180°C with HMF yields of 63% and 67%
respectively. AlCl3 gave the highest HMF yield from glucose, while ZnCl2 yielded better results when
sucrose or sugarcane molasses were used as feedstock. Conversion and selectivity data are not reported, though the formation of humins was confirmed. The thermal conversion of glucose without added catalyst yielded 20% HMF.
A heterogeneous aluminium-based catalyst for the conversion of glucose to HMF was reported by Moreno-Recio et al.41 They compared a mesoporous MCM-31-type aluminosilicate with
the well-known H-beta zeolite for HMF formation in a water(NaCl)/MIBK (1:2.3) biphasic system for the conversion of 9 wt% glucose at 195°C for 30 minutes. The aluminosilicate catalyst yielded 62% HMF yield at high conversion and 65% selectivity. The zeolite performed slightly worse with 56% HMF yield at 59% selectivity. After regeneration the yields dropped to 47% and 21% respectively. Without the addition of NaCl to the aqueous phase, the aluminosilicate catalyst only yielded 36% HMF and 41% selectivity after 150 minutes.
Atanda et al used a TiO2-ZrO2 (1/1) binary oxide for the conversion of glucose to HMF.43
With a water(NaCl)/THF (1/4) solvent system with 9 wt% glucose and 4 wt% catalyst for 3 hours at 175°C a the HMF yield was 71% with a selectivity of 75%. The replacement of half of the catalyst (in wt) with Amberlyst 70 increased the HMF yield to 86% with a selectivity of 86%. This system, at 180°C for 3 hours under 30 bar of Ar pressure, could convert cellulose to HMF with 42% conversion and 26% yield. Generally a higher temperature is necessary to overcome the intermolecular hydrogen bonds of cellulose, but Amberlyst 70 is not stable at temperatures over 190°C. With 20 wt% Glucose a HMF yield of 74% (complete conversion) could be achieved. Changing the organic phase to dioxane could further increase this yield to 84% (complete conversion). However, dioxane is carcinogenic, and is therefore less preferred as a solvent. All catalytic reactions were carried out for a duration of 3 hours. There is no data on the development of conversion and yield over this time, which makes this duration seem arbitrary.
Wrigstedt, Keskiväli and Repo reported the microwave-enhanced dehydration of both fructose and glucose to HMF.34 For the dehydration of 10 wt% fructose, they applied HCL (0.05 M) as
Brønsted acid catalyst in a water(KBr)/MeCN (1:2) biphasic system to achieve 90% HMF yield (91% selectivity) after 1 minute at 160°C under microwave irradiation. For a higher fructose concentration (50 wt% ) with 0.05 M HCl in the same solvent system, 77% HMF was formed with 79% selectivity. For the dehydration of 10 wt% glucose, Amberlyst-38 (wet) (1.4 wt%) was applied in combination with CrCl3.6H2O (10 mol%) in a water(KBr)/GVL (1:2) biphasic system. After 3 min at 160°C under
microwave irradiation the HMF yield was 74% with 76% selectivity. As a comparison, an experiment was carried out using conventional heating. In 40 minutes, 10 wt% fructose in a water(KBr)/MeCN (1:2) biphasic system with 0.05 M HCl yielded 84% HMF with 88% selectivity. In all reactions the yield of levulinic acid was less than 1%. These results are incredibly high. For the analysis the reaction mixture (4.5 mL total) was diluted with water to 50 mL, after which samples were taken for HPLC analysis. The addition of water likely leads to combination of the phases (the KBr concentration is greatly decreased), this is not mentioned in the article though. Thus, there is no information about
23 the distribution of HMF over the phases. The authors conclude that the microwave irradiation reduced the reaction time, and therefore also the formation of byproducts, LA in particular. The fructose dehydration did indeed proceed faster under microwave irradiation (1 min) than using an oil bath (40 min), but the yield and selectivity are only 6 and 3 percent point higher. Furthermore in all reactions the yield of LA was below 1%. It is therefore not convincing that the microwave irradiation influenced the production (rate) of byproducts. Unfortunately, no control experiment was carried out for the dehydration of glucose using concentional heating.
Li and coworkers investigated the use of silica–alumina composite (AlSiO) catalysts for the dehydration of glucose.25 They employed a water(NaCl)/THF (1:3) biphasic system in which 9 wt%
glucose and 9 wt% catalyst were reacted at 160°C for 1.5 hours. They found that the niobium-doped mesoporous AlSiO catalyst Nb0.15Al0.85Si25O performed best, with a HMF yield of 58% and 72%
selectivity. The best performing catalyst exhibited a high weak Lewis acid/total Lewis acid ratio and corresponding volcano-realtionship for the Brønsted acid/Lewis Acid ratio. This is consistent with the other results discussed in this thesis: weak Lewis acid sites promote HMF selectivity while strong Lewis acid sites promote the formation of LA, FA and humins. The optimal amount of Brønsted acid is moderate and follows a volcano-like optimum curve. The catalyst could be recycled by separating, washing and drying it and after 5 consecutive cycles the HMF yield was 41% with 67% selectivity.
Sweygers et al. investigated the influence of microwaves on the production of HMF from fructose, glucose and cellulose in a biphasic system with a Brønsted acid catalyst.37 They used a
system with water/MIBK in a 1:19 ratio, 10 wt% fructose, glucose or cellulose and 0.046 M HCl as catalyst. The reaction temperature was 177°C, for conventionally heated experiments as well as experiments with microwave irradiation. Their results are summarized in table 4. It is clear that for the transformation of fructose to HMF the microwave irradiation has no notable effect. This is contradicting the results of Wrigstedt et al discussed above34, who found that the use of the
microwave significantly increased the reaction rate. The main differences are the type and ratio of organic solvent, and the type of heating. Wrigstedt et al used an oil bath for conventional heating, while Sweygers et al used a silicon carbide reactor, with the same shape of the borosilicate glass microwave reactor, inside the microwave. The silicon carbide absorbs the microwave energy and transfers it to the sample as conventional heating, in contrast to the borosilicate glass reactor that lets the microwaves penetrate and heat the sample directly. The conversions of glucose and cellulose both show a decrease in reaction time and an increase in the HMF yield and selectivity when microwave heating is applied. Especially notable is the full conversion of all substrates, including cellulose. These results are interesting, but it would be beneficial to repeat this type of experiment with an added glucose-fructose isomerisation catalyst.
A very high yield for the aqueous conversion of Glucose to HMF was reported by Pumrod et al.42 They used a combination of ion exchange resin and aluminium oxide as catalyst, in the ratio
RCP160M:Al2O3 1:2 (B/L). As aqueous phase a combination of water and NMP in the ratio 1:4 was
used, with added NaCl. The aqueous to organic phase ratio was 7:3, with MIBK as the organic phase. The glucose concentration was 3 wt%. At 120°C for 8 hours the achieved HMF yield was 85% with 97% glucose conversion, corresponding to a HMF-selectivity of 90%. If the water:NMP ratio was increased from 1:4 to 1:0, a negative effect on the glucose conversion and HMF yield an selectivity was found, decreasing from 84% to 73%, 54% to 37% and 67% to 37% respectively. This is explained to be a consequence of the NMP and MIBK phases mixing, and the hydrophilic nature of the sulphonic acid group on the catalyst. Notably, the amount of HMF is measured in both layers, the authors do not mention whether the HMF is mostly in the MIBK phase or in the NMP/H2O phase.
24 This is important to know, since HMF is difficult to isolate from aprotic organic solvents such as NMP. Whether a water/NMP ratio of 1/4 is really an aqueous phase might also be up for discussion. The catalyst was shown to be recyclable. After 5 consecutive cycles the HMF yield reduced to 71%, with 79% selectivity.
Guo Heeres and Yue reported another glucose to HMF transformation using the well-known Lewis and Brønsted acid combination of AlCl3 and HCl.38 Interesting about their approach is the use of
a biphasic slug flow capillary reactor to achieve a continuous process (see figure 18). The optimal conditions were a 3 wt% solution of glucose with 20 wt% NaCl, MIBK as the extraction solvent with a aqueous/organic phase ratio of 1:4, 40 mM of both AlCl3 and HCL, at a temperature of 160°C and a
residence time of 16 minutes. The HMF yield was 66%, with 83% glucose conversion and 80% selectivity. The catalysts were reused to yield similar results, without noticeable loss of yield or selectivity after 3 cycles. In between cycles, the aqueous phase was collected, HMF and soluble by-products were extracted and insoluble humins were removed through filtration.
Substrate Heating Reaction time
(min) HMF Yield (%) Conversion (%) Selectivity (%) Fructose Conventional 5 70 100 70 Fructose Microwave 5 68 100 68 Glucose Conventional 240 35 100 35 Glucose Microwave 120 51 100 51 Cellulose Conventional 240 34 100 34 Cellulose Microwave 100 43 100 43
Table 4 Influence of microwaves on dehydration of fructose, glucose and cellulose. Summary of data by Sweygers et al.37
25
4. Conclusions
HMF can be produced from fructose and glucose using a biphasic system with different homogeneous and heterogeneous Brønsted & Lewis acid catalysts. Zeolites and metal salts are very common catalysts, often combined with HCl. The reaction phase is usually aqueous, sometimes mixed with a different solvent, often with added alkali salts. The organic phase can be many different solvents but THF and MIBK are most commonly used. The organic phase enables the in situ extraction of HMF and prevents the formation of byproducts such as LA, FA and humins in the aqueous phase. The aqueous-to-organic phase ratio usually lies between 1:2 and 1:10. The reaction temperatures range from 120 to 195 °C. Reaction times start at a few minutes and go up to 8 hours. For the dehydration of glucose and cellulose there are indications that microwave heating might decrease the reaction time.
Mechanistically, the main consensus in literature is that the dehydration of fructose to HMF proceeds via cyclic intermediates, and the dehydration of glucose proceeds through isomerization to fructose followed by the cyclic HMF-formation. The rate-limiting step in HMF formation is believed to be the 1,2-hydride shift taking place after protonation and the first of three dehydrations. For the dehydration of glucose there is some evidence pointing to the direct glucose dehydration mechanism, involving direct conversion of the pyranose ring to a 5-membered ring intermediate. However, this mechanism appears less likely.
Several factors have been identified that influence the HMF yield and selectivity. The addition of alkali metal salts, often at saturation concentration, to the aqueous phase improves sugar conversion and HMF yield. This is attributed to the facts that the salt increases the miscibility gap between water and the organic solvent, and decreases the solubility of HMF in water, promoting the transfer of HMF into the organic phase. Similarly, the choice of organic solvent is important. A good partition coefficient for HMF between the aqueous and organic phases promotes HMF yield and selectivity. Furthermore, the yield of HMF is generally seen to increase with an increasing amount of organic solvent.
In Van Putten’s comprehensive review from 20135, research was shown that yielded 60-90%
HMF with glucose conversions of 60-90%. In the current review, glucose conversions of >90 percent are not uncommon, and HMF yields of >80% are reported, even though HMF yields around 50% are also still common. So there appears to have been a slight increase in reported yields and selectivities over the years, but is also seems as though the maximum achievable yield might have been reached. At least for the current research setup, which consists of trying new types of Lewis acid catalysts or optimizing reaction conditions.
It is important to note that almost none of the discussed papers describe the isolation and purification process of HMF. Several methods do not separately analyse the aqueous and organic phases, failing to provide data on the extraction coefficient of HMF. None of the articles provide an isolated yield for HMF. Many of the research is done on a small scale of around 5 mL. Sensible next steps would be to focus on the scaling up of the processes and isolating and purifying the HMF yield.
26
5. References
(1) Field, C. B.; Barros, V.; Stocker, T. F.; Dahe, Q.; Jon Dokken, D.; Ebi, K. L.; Mastrandrea, M. D.; Mach, K. J.; Plattner, G. K.; Allen, S. K.; Tignor, M.; Midgley, P. M. Managing the Risks of Extreme Events and
Disasters to Advance Climate Change Adaptation: Special Report of the Intergovernmental Panel on Climate Change; Cambridge University Press, 2012; Vol. 9781107025066.
https://doi.org/10.1017/CBO9781139177245.
(2) Gallezot, P. Conversion of Biomass to Selected Chemical Products. Chem. Soc. Rev. 2012, 41 (4), 1538– 1558. https://doi.org/10.1039/c1cs15147a.
(3) Besson, M.; Gallezot, P.; Pinel, C. Conversion of Biomass into Chemicals over Metal Catalysts. Chemical
Reviews. American Chemical Society February 12, 2014, pp 1827–1870.
https://doi.org/10.1021/cr4002269.
(4) Caes, B. R.; Teixeira, R. E.; Knapp, K. G.; Raines, R. T. Biomass to Furanics: Renewable Routes to Chemicals and Fuels. ACS Sustain. Chem. Eng. 2015, 3 (11), 2591–2605.
https://doi.org/10.1021/acssuschemeng.5b00473.
(5) Van Putten, R. J.; Van Der Waal, J. C.; De Jong, E.; Rasrendra, C. B.; Heeres, H. J.; De Vries, J. G. Hydroxymethylfurfural, a Versatile Platform Chemical Made from Renewable Resources. Chemical
Reviews. American Chemical Society March 13, 2013, pp 1499–1597.
https://doi.org/10.1021/cr300182k.
(6) Teong, S. P.; Yi, G.; Zhang, Y. Hydroxymethylfurfural Production from Bioresources: Past, Present and Future. Green Chemistry. Royal Society of Chemistry March 24, 2014, pp 2015–2026.
https://doi.org/10.1039/c3gc42018c.
(7) Zou, X.; Zhu, C.; Wang, Q.; Yang, G. Catalytic Dehydration of Hexose Sugars to 5-Hydroxymethylfural.
Biofuels, Bioprod. Biorefining 2019, 13 (1), 153–173. https://doi.org/10.1002/bbb.1932.
(8) Esteban, J.; Vorholt, A. J.; Leitner, W. An Overview of the Biphasic Dehydration of Sugars to 5-Hydroxymethylfurfural and Furfural: A Rational Selection of Solvents Using COSMO-RS and Selection Guides. Green Chem. 2020, 22 (7), 2097–2128. https://doi.org/10.1039/c9gc04208c.
(9) Van Zandvoort, I.; Koers, E. J.; Weingarth, M.; Bruijnincx, P. C. A.; Baldus, M.; Weckhuysen, B. M. Structural Characterization of 13C-Enriched Humins and Alkali-Treated 13C Humins by 2D Solid-State NMR. Green Chem. 2015, 17 (8), 4383–4392. https://doi.org/10.1039/c5gc00327j.
(10) Rogers, R. D.; Seddon, K. R. Ionic Liquids--Solvents of the Future? Science (80-. ). 2003, 302 (5646), 792– 793.
(11) Lucas, N.; Nagpure, A. S.; Gurrala, L.; Gogoi, P.; Chilukuri, S. Efficacy of Clay Catalysts for the
Dehydration of Fructose to 5-Hydroxymethyl Furfural in Biphasic Medium. J. Porous Mater. 2020, 1, 3. https://doi.org/10.1007/s10934-020-00943-8.
(12) van Putten, R.-J.; Soetedjo, J. N. M.; Pidko, E. A.; van der Waal, J. C.; Hensen, E. J. M.; de Jong, E.; Heeres, H. J. Dehydration of Different Ketoses and Aldoses to 5-Hydroxymethylfurfural. ChemSusChem
2013, 6 (9), 1681–1687. https://doi.org/10.1002/cssc.201300345.
(13) Que, L.; Gray, G. R. 13C Nuclear Magnetic Resonance Spectra and the Tautomeric Equilibriums of Ketohexoses in Solution. Biochemistry 1974, 13 (1), 146–153. https://doi.org/10.1021/bi00698a023. (14) Wertz, P. W.; Carver, J. C.; Anderson, L. Anatomy of a Complex Mutarotation. Kinetics of
Tautomerization of α-D-Galactopyranose and β-D-Galactopyranose in Water. J. Am. Chem. Soc. 1981,
103 (13), 3916–3922. https://doi.org/10.1021/ja00403a051.
(15) Caratzoulas, S.; Vlachos, D. G. Converting Fructose to 5-Hydroxymethylfurfural: A Quantum
Mechanics/Molecular Mechanics Study of the Mechanism and Energetics. Carbohydr. Res. 2011, 346 (5), 664–672. https://doi.org/10.1016/j.carres.2011.01.029.
(16) Nikbin, N.; Caratzoulas, S.; Vlachos, D. G. A First Principles-Based Microkinetic Model for the Conversion of Fructose to 5-Hydroxymethylfurfural. ChemCatChem 2012, 4 (4), 504–511.
https://doi.org/10.1002/cctc.201100444.
(17) Dallas Swift, T.; Bagia, C.; Choudhary, V.; Peklaris, G.; Nikolakis, V.; Vlachos, D. G. Kinetics of Homogeneous Brønsted Acid Catalyzed Fructose Dehydration and 5-Hydroxymethyl Furfural Rehydration: A Combined Experimental and Computational Study. 2013.
https://doi.org/10.1021/cs4009495.
(18) Rodriguez Quiroz, N.; Norton, A. M.; Nguyen, H.; Vasileiadou, E.; Vlachos, D. G. Homogeneous Metal Salt Solutions for Biomass Upgrading and Other Select Organic Reactions. ACS Catal. 2019, 9 (11), 9923–9952. https://doi.org/10.1021/acscatal.9b01853.
27
(19) Delidovich, I.; Palkovits, R. Catalytic Isomerization of Biomass-Derived Aldoses: A Review.
ChemSusChem. Wiley-VCH Verlag March 21, 2016, pp 547–561.
https://doi.org/10.1002/cssc.201501577.
(20) Moliner, M.; Román-Leshkov, Y.; Davis, M. E. Tin-Containing Zeolites Are Highly Active Catalysts for the Isomerization of Glucose in Water. Proc. Natl. Acad. Sci. U. S. A. 2010, 107 (14), 6164–6168.
https://doi.org/10.1073/pnas.1002358107.
(21) Bermejo-Deval, R.; Assary, R. S.; Nikolla, E.; Moliner, M.; Román-Leshkov, Y.; Hwang, S. J.; Palsdottir, A.; Silverman, D.; Lobo, R. F.; Curtiss, L. A.; Davis, M. E. Metalloenzyme-like Catalyzed Isomerizations of Sugars by Lewis Acid Zeolites. Proc. Natl. Acad. Sci. U. S. A. 2012, 109 (25), 9727–9732.
https://doi.org/10.1073/pnas.1206708109.
(22) Román-Leshkov, Y.; Moliner, M.; Labinger, J. A.; Davis, M. E. Mechanism of Glucose Isomerization Using a Solid Lewis Acid Catalyst in Water. Angew. Chemie - Int. Ed. 2010, 49 (47), 8954–8957.
https://doi.org/10.1002/anie.201004689.
(23) Choudhary, V.; Mushrif, S. H.; Ho, C.; Anderko, A.; Nikolakis, V.; Marinkovic, N. S.; Frenkel, A. I.; Sandler, S. I.; Vlachos, D. G. Insights into the Interplay of Lewis and Brønsted Acid Catalysts in Glucose and Fructose Conversion to 5-(Hydroxymethyl)Furfural and Levulinic Acid in Aqueous Media. J. Am. Chem.
Soc. 2013, 135 (10), 3997–4006. https://doi.org/10.1021/ja3122763.
(24) Mushrif, S. H.; Varghese, J. J.; Vlachos, D. G. Insights into the Cr(Iii) Catalyzed Isomerization Mechanism of Glucose to Fructose in the Presence of Water Using Ab Initio Molecular Dynamics. Phys. Chem.
Chem. Phys. 2014, 16 (36), 19564–19572. https://doi.org/10.1039/c4cp02095b.
(25) Li, X.; Peng, K.; Liu, X.; Xia, Q.; Wang, Y. Comprehensive Understanding of the Role of Brønsted and Lewis Acid Sites in Glucose Conversion into 5-Hydromethylfurfural. ChemCatChem 2017, 9 (14), 2739– 2746. https://doi.org/10.1002/cctc.201601203.
(26) Nguyen, H.; Nikolakis, V.; Vlachos, D. G. Mechanistic Insights into Lewis Acid Metal Salt-Catalyzed Glucose Chemistry in Aqueous Solution. ACS Catal. 2016, 6 (3), 1497–1504.
https://doi.org/10.1021/acscatal.5b02698.
(27) Dallas Swift, T.; Nguyen, H.; Anderko, A.; Nikolakis, V.; Vlachos, D. G. Tandem Lewis/Bronsted Homogeneous Acid Catalysis: Conversion of Glucose to 5-Hydoxymethylfurfural in an Aqueous Chromium(III) Chloride and Hydrochloric Acid Solution. Green Chem. 2015, 17 (10), 4725–4735. https://doi.org/10.1039/C5GC01257K.
(28) Tang, J.; Zhu, L.; Fu, X.; Dai, J.; Guo, X.; Hu, C. Insights into the Kinetics and Reaction Network of Aluminum Chloride-Catalyzed Conversion of Glucose in NaCl-H2O/THF Biphasic System. ACS Catal.
2017, 7 (1), 256–266. https://doi.org/10.1021/acscatal.6b02515.
(29) Swift, T. D.; Nguyen, H.; Erdman, Z.; Kruger, J. S.; Nikolakis, V.; Vlachos, D. G. Tandem Lewis
Acid/Brønsted Acid-Catalyzed Conversion of Carbohydrates to 5-Hydroxymethylfurfural Using Zeolite Beta. J. Catal. 2016, 333, 149–161. https://doi.org/10.1016/j.jcat.2015.10.009.
(30) Qian, X. Mechanisms and Energetics for Brønsted Acid-Catalyzed Glucose Condensation, Dehydration and Isomerization Reactions. In Topics in Catalysis; Springer, 2012; Vol. 55, pp 218–226.
https://doi.org/10.1007/s11244-012-9790-6.
(31) Qian, X. Mechanisms and Energetics for Acid Catalyzed β-D-Glucose Conversion to 5-Hydroxymethylfurfurl. J. Phys. Chem. A 2011, 115 (42), 11740–11748.
https://doi.org/10.1021/jp2041982.
(32) Qian, X.; Wei, X. Glucose Isomerization to Fructose from Ab Initio Molecular Dynamics Simulations. J.
Phys. Chem. B 2012, 116 (35), 10898–10904. https://doi.org/10.1021/jp303842g.
(33) Yang, L.; Tsilomelekis, G.; Caratzoulas, S.; Vlachos, D. G. Mechanism of Brønsted Acid-Catalyzed Glucose Dehydration. ChemSusChem 2015, 8 (8), 1334–1341. https://doi.org/10.1002/cssc.201403264.
(34) Wrigstedt, P.; Keskivali, J.; Repo, T. Microwave-Enhanced Aqueous Biphasic Dehydration of Carbohydrates to 5-Hydroxymethylfurfural. RSC Adv. 2016, 6 (23), 18973–18979.
https://doi.org/10.1039/C5RA25564C.
(35) Ma, H.; Wang, F.; Yu, Y.; Wang, L.; Li, X. Autocatalytic Production of 5-Hydroxymethylfurfural from Fructose-Based Carbohydrates in a Biphasic System and Its Purification. Ind. Eng. Chem. Res. 2015, 54 (10), 2657–2666. https://doi.org/10.1021/ie504791x.
(36) Gomes, G. R.; Rampon, D. S.; Ramos, L. P. Production of 5-(Hydroxymethyl)-Furfural from Water-Soluble Carbohydrates and Sugarcane Molasses. Appl. Catal. A Gen. 2017, 545, 127–133. https://doi.org/10.1016/j.apcata.2017.07.049.
(37) Sweygers, N.; Alewaters, N.; Dewil, R.; Appels, L. Microwave Effects in the Dilute Acid Hydrolysis of Cellulose to 5-Hydroxymethylfurfural. Sci. Rep. 2018, 8 (1), 1–11.
https://doi.org/10.1038/s41598-018-28
26107-y.
(38) Guo, W.; Heeres, H. J.; Yue, J. Continuous Synthesis of 5-Hydroxymethylfurfural from Glucose Using a Combination of AlCl3 and HCl as Catalyst in a Biphasic Slug Flow Capillary Microreactor. Chem. Eng. J.
(Amsterdam, Netherlands) 2020, 381, 122754. https://doi.org/10.1016/j.cej.2019.122754.
(39) Yang, F.; Tong, X.; Xia, F.; Zheng, C.; Qin, L.; Jiang, X. Efficient Hydroxymethylfurfural Production over Phosphoric Carbon Solid Acids. Catal. Letters 2018, 148 (7), 1848–1855.
https://doi.org/10.1007/s10562-018-2396-2.
(40) Dibenedetto, A.; Aresta, M.; di Bitonto, L.; Pastore, C. Organic Carbonates: Efficient Extraction Solvents for the Synthesis of HMF in Aqueous Media with Cerium Phosphates as Catalysts. ChemSusChem 2016,
9 (1), 118–125. https://doi.org/10.1002/cssc.201501181.
(41) Moreno-Recio, M.; Jimenez-Morales, I.; Arias, P. L.; Santamaria-Gonzalez, J.; Maireles-Torres, P. The Key Role of Textural Properties of Aluminosilicates in the Acid-Catalysed Dehydration of Glucose into 5-Hydroxymethylfurfural. ChemistrySelect 2017, 2 (8), 2444–2451.
https://doi.org/10.1002/slct.201700097.
(42) Pumrod, S.; Kaewchada, A.; Roddecha, S.; Jaree, A. 5-HMF Production from Glucose Using Ion Exchange Resin and Alumina as a Dual Catalyst in a Biphasic System. RSC Adv. 2020, 10 (16), 9492–9498.
https://doi.org/10.1039/c9ra09997b.
(43) Atanda, L.; Silahua, A.; Mukundan, S.; Shrotri, A.; Torres-Torres, G.; Beltramini, J. Catalytic Behaviour of TiO2-ZrO2 Binary Oxide Synthesized by Sol-Gel Process for Glucose Conversion to
5-Hydroxymethylfurfural. RSC Adv. 2015, 5 (98), 80346–80352. https://doi.org/10.1039/C5RA15739K. (44) Ranoux, A.; Djanashvili, K.; Arends, I. W. C. E.; Hanefeld, U. 5-Hydroxymethylfurfural Synthesis from
Hexoses Is Autocatalytic. ACS Catal. 2013, 3 (4), 760–763. https://doi.org/10.1021/cs400099a.
(45) Dibenedetto, A.; Aresta, M.; Pastore, C.; Di Bitonto, L.; Angelini, A.; Quaranta, E. Conversion of Fructose into 5-HMF: A Study on the Behaviour of Heterogeneous Cerium-Based Catalysts and Their Stability in Aqueous Media under Mild Conditions. RSC Adv. 2015, 5 (34), 26941–26948.
