*2.3. Pd-Based Catalysts*

Pd-based catalysts also showed good catalytic performance for the aerobic oxidation of HMF into FDCA. According to the reported Pd-based catalysts for the aerobic oxidation of HMF, results along with their reaction conditions are summarized in Table 3.


**Table 3.** Summary of reported results for the aerobic oxidation of HMF to FDCA over Pd catalysts.

a PVP = Polyvinylpyrrolidone b CC = Carbonaceous Catalyst c HAP = Hydroxyapatite d HT= Hydrotalcite.

The performance and particle size of di fferent Pd-based catalysts are found to be dependent on the choice of the support. Siyo et al. studied the e ffect of catalytic support by depositing PVP stabilized Pd-nanoparticles on di fferent metal oxide (Al2O3, TiO2, KF/Al2O3, and ZrO2/La2O3) supports in ethylene glycol with average particle size of 1.8 nm [67]. Pd/ZrO2/La2O3 catalyst showed to be a promising catalyst for HMF oxidation reaction, with a FDCA yield of 90% at 90 ◦C and 1 bar O2 pressure after 8 h of reaction time (Table 3). TEM indicated that there was no obvious sintering of Pd nanoparticles in the spent Pd/ZrO2/La2O3 catalyst, whereas other catalysts showed serious aggregation. XPS confirmed that the majority of Pd remained as the metallic state and that the electronic structure of the Pd nanoparticles was unchanged in the spent Pd/ZrO2/La2O3 catalyst. The strong interaction between Pd nanoparticles and the metal oxide supports makes the catalysts more stable, to which its efficient catalytic performance is attributed. Siyo et al. further studied PVP stabilized Pd NP catalyst using ethylene glycol at three di fferent Pd/NaOH ratios [68]. Di fferently sized Pd nanoparticles are obtained, with mean diameters between 1.8 nm and 4.4 nm, by changing the Pd/NaOH ratios. It was concluded that smaller Pd-nanoparticles showed higher FDCA yield. The highest FDCA yield obtained is 90% with greater than 99% HMF conversion over 1.8 nm Pd-nanoparticle catalyst synthesized with 1:4 molar ratio of Pd/NaOH [68]. These newly designed Pd-catalysts were able to store for long time (up to one month) in alkaline medium under ambient conditions without any change in their activity.

Owing to the ease in catalyst separation after its use, Zhang et al. prepared a Pd catalyst with magnetic properties [70]. In this catalyst hydroxyapatite (HAP) was layered on Fe2O3 using coating technique and Pd2+ ions are exchanged with HAP's Ca2+ ions with further reduction to produce Pd<sup>0</sup> nanoparticles. This novel catalyst showed excellent catalytic activity with almost 93% of FDCA yield and 97% HMF conversion under optimal reaction conditions (T = 100 ◦C, PO2 = 1 bar, t = 6 h). This catalyst, due to its magnetic properties, is very easy to recycle as it can be easily separated using an external magne<sup>t</sup> without any change in catalytic activity. In another study, similar super paramagnetic catalyst was prepared by Mei and coworkers, in which graphene-based C-Fe2O3-Pd catalysts were synthesized using reliable graphene oxide [71]. Pd nanoparticles and Fe2O3 particles are simultaneously deposited on surface of graphene oxide by one pot solvothermal technique. The reaction temperature and the base concentration notably a ffected the HMF conversion and FDCA selectivity. This carbon catalyst demonstrated high catalytic activity for oxidation reactions of HMF with FDCA with a yield of 91.8% and HMF conversion of 98.2% under mild reaction parameters (80 ◦C, O2, 1 bar, 4 h) with a base/substrate ratio of 0.5. The same research group further studied the Pd-based magnetic catalysts and e ffectively immobilized the Pd-nanoparticles on core-shell structure of C@Fe3O4 (C acted as shell and Fe3O4 as core) with further reduction to generate Pd/C@Fe3O4 catalysts [72]. In this technique, no excess reducing agents and capping reagents were used which makes it a clean and environmentally benign technique. Under optimal conditions, Pd/C@Fe3O4 catalyst showed high catalytic activity in the aerobic oxidation of HMF to FDCA with 98.4% HMF conversion and 87.8% FDCA yield at 80 ◦C for 6 h reaction.

Bimetallic catalysts has also been studied and found to be an e fficient approach for the aerobic oxidation of HMF to FDCA. Lolli et al. studied bimetallic Pd-catalysts for HMF oxidation to FDCA by preparing Pd-Au/TiO2 nanoparticle catalysts along with their monometallic counterpart Pd/TiO2 and revealed that minor alloying of Pd with gold (Pd:Au = 1:6) can enhance the FDCA yield from 9% to 85% with complete HMF conversion at 70 ◦C and 10 bar O2 pressure for4hreaction [73]. Major reaction routes involved in FDCA production on pyrolized Pd-Au/TiO2 in the presence of base are summarized in Scheme 6. According to this study, once HMFCA is formed, the bimetallic catalyst showed a reaction route with no further oxidation due to the substantial inability of Pd and its alloy to oxidize the hydroxymethyl group of HMFCA [73]. FDCA molecule is produced via oxidation of alcohol part of the HMF molecule to produce DFF, which upon further oxidation gives FFCA. This produced FFCA has the options of producing FDCA or being oxidized to HMFCA which will again show the same behavior of no further oxidation. Recently, similar Pd-Au bimetallic catalysts were also synthesized on hydrotalcite (HT) support by Xia et al and their catalytic performance compared

with that of their monometallic counterparts Pd/HT and Au/HT catalysts [74]. High FDCA yield (up to 90%) was achieved over bimetallic Pd-Au/Ht catalyst with Au:Pd = 4:1 at 60 ◦C and 1 bar O2 pressure after 6 h of reaction. Similar to the previous report, the enhanced catalytic performance over a bimetallic catalyst is attributed to the synergistic effect between two metals and the formation of smaller particle size of the catalyst that facilities the HMF oxidation reaction to produce FDCA. In addition, Gupta et al. further studied the bimetallic M0.9-Pd0.1 (M = Ni, Co or Cu) alloy nanoparticles supported on in situ prepared Mg(OH)2 for the aerobic oxidation of HMF to FDCA without using an external base additive [75]. Ni0.9-Pd0.1/Mg(OH)2 was found to be the most efficient, and afforded superior catalytic activity with a FDCA yield of 89% compared to Co- or Cu-based bimetallic nanoparticles at 100 ◦C. The basicity of support could facilitate the activation of the hydroxyl group of HMF, leading to the enhanced FDCA production. Furthermore, Wang et al. reported another HT supported Pd nanoparticle catalyst in base-free environment, in which Mg-Al-CO3 HT supported Pd catalysts are synthesized with different Mg/Al ratios [76]. The Mg/Al molar ratio 5:1 with 2% Pd metal loading (2%Pd/HT-5) appeared to be best in terms of catalytic activity due to formation of weak basic sites (OH- groups). FDCA yields greater than 99% are achieved with >99% HMF conversion for 8 h under mild reaction conditions (T = 100 ◦C and P = 1 bar).

**Scheme 6.** Reaction pathways for HMF oxidation over Pd-Au/TiO2 catalyst (reproduced from [73] with permission from Elsevier, copyright 2015).

#### *2.4. Other Noble Metal Catalysts*

The aerobic oxidation of HMF to FDCA can also be normally catalyzed with different Ru-based catalysts at moderate reaction temperatures ranging from 100–150 ◦C, with O2/air pressure of 1 to 40 bar and base additives [77–83]. Majorly, Ru-based catalysts were reported for the oxidation of HMF to DFF in organic solvents [84]. While only a few examples were extended to the oxidation of HMF to FDCA in water [85]. The performance of different Ru-based catalysts, for catalytic aerobic oxidation of HMF to FDCA is summarized, and details are also given in Table 4.

Yi et al. compared the effect of weak bases (K2CO3, Na2CO3, HT, and CaCO3) with strong base (NaOH) over commercial Ru/C catalyst and concluded that the stronger base lead towards lower FDCA yield (69%) (Table 1) due to degradation of HMF at higher pH [78]. The weaker the base, the higher the FDCA yield. The maximum FDCA yield (95%) was attained by the use of CaCO3 and FDCA obtained is in the form of its calcium salt because of its lower solubility.


**Table 4.** Summary of reported literature on HMF oxidation to FDCA over Ru catalysts.

a ACNaOCl= Activated Carbon oxidized with sodium hypochlorite b HT= Hydrotalcite c HAP = Hydroxyapatite.

In this reaction system, oxidation of HMF occurred through hydroxyl part rather than aldehyde part. Ru-catalyst favors the production of FDCA through DFF route instead of HMFCA route due to oxidation of –OH group of HMF to –CHO group (Scheme 7). It is concluded that the conversion of –OH to –CHO is very fast step as compared to subsequent conversion of –CHO to –COOH, which is the rate determining step [78]. Over Ru/C catalyst, the oxidation of –CHO to –COOH is in less time (5 h) due to presence of base, which is a challenge in base-free conditions. This is why base-free reaction was carried out for relatively longer reaction time (10 h) to ge<sup>t</sup> 88% FDCA yield at similar reaction conditions (120 ◦C, O2, 2 bar).

**Scheme 7.** Reaction route for oxidation of HMF to produce FDCA over Ru/C catalysts (reproduced from [78] with permission from the Royal Society of Chemistry, copyright 2016).

Kerdi and coworkers studied the modification of catalyst support using various doped carbons as supporting materials for Ru-based catalysts to investigate the consequence of surface properties as well as pore structure of carbon on oxidation rates [79]. Activated carbon was oxidized using sodium hypochlorite (NaOCl) and the Ru metal was impregnated on this oxidized ACNaOCl support to prepare Ru/ACNaOCl catalyst with particle size of 2 nm. Moderate results with FDCA yield of 55% with complete HMF conversion are obtained after4hreaction at 100 ◦C over a modified supported Ru catalysts [79].

Normally, the HMF oxidation reaction takes place in organic solvents or water, but this reaction is also investigated in ionic liquids (ILs) over Ru(OH)x catalysts in base-free environment due to redox stability, negligible vapor pressure, non-flammability and unique dissolving abilities of ILs [80]. In this study, several different supports (TiO2, Fe2O3, ZrO2, CeO2, HAP, HT, MgO, and La2O3) on Ru(OH)x were tested along with various ILs. As shown in Table 4, Ru(OH)x/La2O3 gives reasonable catalytic

activity, with a FDCA yield of 48% and high HMF conversion (98%) at 100 ◦C and elevated pressure (30 bar O2). Ru(OH)x/HT also appeared to be active in ILs with 99% HMF conversion but FDCA yield obtained is very low (19%) at high temperature (140 ◦C) and ambient pressure after 24 h of reaction. High temperature (140 ◦C) is selected to decrease the influence of viscosity of the mixture formed in ILs and catalyst [80]. As the idea for using ionic liquids is flopped because of their high cost, low FDCA yield and unfeasible large scale production, researchers continue to search for base-free phase oxidation over modified support Ru catalysts. In another work, MnCo2O4 spinel supported Ru-catalyst (Ru/MnCo2O4) with 4% metal loading shows an exceptionally high FDCA yield of 99.1% in comparison with other modified support catalysts (Ru/CoMn2O4 and Ru/MnCo2CO3) under moderate reaction conditions (T = 120 ◦C, Pair = 24 bar and t = 10 h) with minor FFCA impurities [81]. In this study, it is concluded that catalyst supports structure plays vital role in catalytic activity. The high FDCA yield is due to Lewis and BrØnsted active acid sites on catalytic surface which is confirmed by NH3-TPD results (Figure 4). Variation in supports to CoMn2O4 and MnCo2CO3 adversely affects the catalytic activity and FDCA yield is decreased to 82.2% and 69.9%, respectively (Table 4).

**Figure 4.** (**a**) NH3-TPD and (**b**) summarized data obtained from NH3-TPD for CoMn2O4, MnCo2O4 and Ru/MnCo2O4 catalyst (reproduced from [81] with permission from the Royal Society of Chemistry, copyright 2017).

Ru/HAP also shows good catalytic activity with 100% of HMF conversion and more than 99% of FDCA yield in base-free conditions [82]. This reaction is carried out in severe reaction conditions (T = 140 ◦C, P = 10 bar and t = 24 h) which is very challenging for practical applications. High catalytic performance of Ru/HAP catalyst is credited to formation of acidic-basic sites due to well dispersed Ru nanoparticles on HAP support. Christian and coworkers studied FDCA production via HMF oxidation over Ru catalyst supported on high surface area zirconia (ZrO2) in a base-free environment [83]. High FDCA yield (97%) is achieved with 100% HMF conversion after 16 h reaction at similar severe reaction conditions (T = 120 ◦C and P = 10 bar). The catalytic tests in this investigation revealed that the small size of Ru particles due to utilization of high surface area ZrO2 is crucial reason for better catalytic performance.

Rhodium (Rh) metal, in comparison with other noble metals (Ru, Pd, Au, and Pt), also has similar potential to act as catalytic active site in heterogeneous catalysis. However, this metal is not broadly explored by the researchers for catalytic oxidation of HMF to FDCA. This could be because of its high cost and less availability. Vuyyuru et al. have taken first step to use Rh to prepare a Rh/C catalyst for catalytic oxidation of HMF to FDCA [86]. The catalytic activity is compared for HMF oxidation reaction using different noble metals on carbon support at mild temperature (50 ◦C) and oxygen pressure (10 bar). The FDCA yield obtained in this work is 12.62%, which is comparatively low, with HMF conversion of 82%, after 4 h of reaction time over Rh/C catalysts. Ag-based catalysts were also studied for the aerobic oxidation of HMF to FDCA [87]. Inferior activities and selectivity to FDCA were shown with Ag-based catalysts, with HMFCA as the primary product. In addition, leaching of Ag was also demonstrated to be another issue during catalysis. Therefore, rooms are available for the

improvement of catalytic performance of those noble-metal based catalysts in the aerobic oxidation of HMF into FDCA.

To sum up the overall results for the aerobic oxidation of HMF to FDCA over noble metal catalysts, a lot of progress has been achieved recently. Au catalysts are more e ffective catalysts in terms of stability and selectivity in comparison with Pt, Pd, Ru, and Rh based catalysts owing to a better ability in resistance to water and oxygen. However, deactivation of Au catalysts, and intermediates depositing on catalytic active sites are still observed for Au catalysts. Most of the studied noble metal catalysts use base additives to achieve high FDCA yield, to facilitate the oxidation of aldehyde part of HMF and keep the formed FDCA to dissolve into the solutions, in which the strong adsorption of products on the catalysts can be prevented. The aerial oxidation of HMF in base-free environment is a more green process, and more appropriate for sustainable chemistry. This entails a more research focus on base-free catalytic systems for the oxidation of HMF to FDCA. Applying an appropriate support and using an alloy strategy might make the catalysts show high activity and stability for the base-free oxidation of HMF into FDCA.

#### **3. Non-Noble Metal Catalysts for FDCA Production**

Noble metal catalysts are generally considered as active and stable in the field of catalysis, but due to their higher costs and less availability, it is of interest to design non-noble metal catalysts with high efficiency and excellent stability. Therefore, research shifted towards non-precious metal catalysts for aerobic oxidation of HMF to FDCA with prominent catalytic performance, and a reasonable progress has been achieved until now to ge<sup>t</sup> active and stable catalysts [88] (Table 5).


**Table 5.** Results of the oxidation of HMF to FDCA over non-noble metal catalysts.

a POP = Porous organic polymer; b MR-Co-Py = Merrifield Resin supported Co(II)-*meso*-tetra(4-pyridylporhyrin).

Saha et al. prepared a thermally stable, robust structured iron catalyst (Fe/POP) by the integration of Fe3<sup>+</sup> on the center of porphyrin ring supported on porous organic polymer (POP) to study the catalytic performance of aerobic oxidation of HMF to FDCA in aqueous medium [89]. This inexpensive metal catalyst can be reused without any significant loss of activity because the oxidation state of Fe remains intact after the reaction over this catalyst. As a result, complete HMF conversion was achieved with high FDCA yield (85%) at 100 ◦C and 10 bar O2 pressure for 10 h reaction. The metal active site of Fe3<sup>+</sup>-POP catalyst plays a significant role and a plausible radical chain mechanism for HMF oxidation would involve thermal autoxidation of organic substrate (R–H) to peroxides (ROOH) which further lead to FDCA product through Fenton-type cleavage of RO–OH bond over Fe. Later on, a stable cobalt (II)-*meso*-tetra(4-pyridyl) porphyrin supported on Merrifield resin catalyst was developed (abbreviated as Merrified Resin-Co-Py) by Gao et al. and studied the e ffect of various oxidants [90]. This catalyst showed excellent catalytic activity (FDCA yield = 90.4%, and HMF conversion = 95.6%) at 100 ◦C in the presence of *tert*-butylhydroperoxide (*t*-BuOOH) as oxidant after 24 h reaction (Table 5). On the

other hand, in the presence of O2 as oxidant, no FDCA was detected after 24 h of reaction. Methyl nitrile (CH3CN) was found to be the best solvent in this reaction system. Furthermore, Jain et al. introduced a low-cost Li2CoMn3O8 (spinel-mixed metal oxide) catalyst prepared by gel pyrolysis method to study the catalytic oxidation of HMF to FDCA in the presence of sodium bromide and acetic acid [91]. Although a reasonable FDCA yield (80%) is obtained but use of high temperature (150 ◦C) along with acetic acid and sodium bromide additives are the main drawbacks for this route.

Similar to noble metal magnetic catalysts, Wang et al. prepared non-noble metal (nano-Fe3O4-CoOx) catalysts with magnetic properties [92]. As demonstrated earlier, this catalyst can also be easily recovered using external magne<sup>t</sup> because of its magnetic properties. This magnetic catalyst showed 97.2% of HMF conversion with FDCA yield (68.6%) and reasonable reusability of catalyst with minor mass loss at 80 ◦C for 12 h reaction (Table 5). Experimental results in this study demonstrated that the first step of HMF oxidation to produce FFCA is initiated by BrØnsted base, even without the presence of catalyst, whereas, the second step to produce FDCA from FFCA, requires the presence of catalysts. Similar conclusions were reached in another study of the catalytic oxidation of HMF to produce FFCA over a low cost Mn0.75/Fe0.25 heterogeneous metal catalyst [100].

Ionic liquids or ionic fluids have also been studied as solvents for this reaction system together with low cost transition metal oxide catalysts [93,94]. Several combinations of iron oxide (Fe2O3) and zirconia (ZrO2) have been used with di fferent Fe to Zr ratios to develop a highly e fficient catalyst. For this system, even though the attained FDCA yield is low, excellent conversion of HMF attracts researchers to dig more about its mechanism. The results illustrated that change in the reaction parameters and using di fferent Fe to Zr proportions can hardly improve the catalytic activity (Table 5).

Manganese (Mn) based catalysts were also reported for the aerobic oxidation of HMF with base additives. MnO2/NaHCO3 system was reported by Hayashi et al. [95]. A FDCA yield of 91% with a complete HMF conversion could be obtained at 100 ◦C and 10 bar O2 after 24 h of reaction. However, catalyst deactivation was observed in the third cycle of reuse runs, mainly owing to adsorbed humin species covering the active sites. Catalyst reactivation could be achieved by calcination at 300 ◦C in air. Hayashi et al. further studied the impact of the structure of MnO2 crystal on the performance in the areobic oxidation of HMF to FDCA through combined computational and experimental studies [101]. They demonstrated that reaction rates per surface area for the slowest step, FFCA oxidation to FDCA step, decrease in the order of β-MnO2 > λ-MnO2 > γ-MnO2 ≈ α-MnO2 > δ-MnO2 > ε-MnO2 on the basis of good agreements achieved between experimental results with the DFT calculations. β-MnO2 exceeds that of the previously reported activated MnO2. The successful synthesis of high-surface-area β-MnO2 could significantly improve the catalytic activity for the aerobic oxidation of HMF to FDCA. Notably, a porous 2D Mn2O3 nanoflakes was prepared by a facile thermal treatment of a Mn-based metal-organic framework (MOF) precursor and applied for oxidation of HMF at 100 ◦C and 14 bar O2 with NaHCO3 [96]. A FDCA yield of 99.5% at complete conversion of HMF could be achieved after a reaction time of 24 h. However, a slight catalyst deactivation was observed during the recycle experiments.

Mixed/binary oxides have also been applied for the aerobic oxidation of HMF and showed enhanced catalytic performances as compared with their mono-oxide counterparts. Han et al prepared a mixed oxide MnOx-CeO2 (Mn/Ce = 6) catalyst by co-precipitation method, which a fforded a high FDCA yield of 91% with a HMF conversion of 98% at 110 ◦C with KHCO3 after 15 h reaction [97]. Structural analysis of mixed oxide catalyst revealed that the Mn4<sup>+</sup> and Ce3+ on catalytic surface played the pivotal role as the active sites for HMF oxidation to FDCA. A mechanism involving the Mn4<sup>+</sup> active center, the lattice oxygen transfer from CeO2 to Mn oxide and the activation of O2 on CeO2 was proposed for the enhanced the performance. Mn-Co binary oxides catalysts with di fferent Mn/Co molar ratios were also studied by Zhang et al. for catalytic oxidation of HMF to FDCA [98]. The MnCo2O4 catalyst with a Mn/Co molar ratios of 1/2, showed a HMF conversion of 99.5% and a FDCA yield of 70.9% at 100 ◦C and 10 bar O2 with KHCO3 for 24 h, which was significantly better than Mn3O4, Co3O4 and Mn-Co binary oxides with other Mn/Co molar ratios. The enhanced catalytic activity was attributed to the presence of Mn3<sup>+</sup> ions and the high oxygen mobility and reducibility.

The aerobic base-free oxidation of HMF to FDCA over non-precious metal catalysts has been limited reported. The use of organic solvent, organic peroxide and base additives may promote the product yield but undoubtedly hamper the green footprint of renewable FEDA production. It is still rather di fficult to use non-noble metal based catalysts for the aerobic oxidation of HMF to FDCA in water and without base additives. Recently, Gao et al. prepared a Mn-Co mixed oxide catalyst (Co3O4/MnxCo) with Co3O4 nanoparticles well-dispersed on amorphous Mn-Co-O solid solutions by co-precipitation method [99]. They claimed that a FDCA yield of >99% could be obtained with Co3O4/Mn0.2Co where the Mn/Co ratio was of 0.2, at 1 bar O2 and 140 ◦C after 24 h without base additives. The high content of both Lewis (Mn<sup>4</sup>+) and Brønsted (Co–O–H+) acid sites on the surface, leading to an excellent ability of HMF adsorption and COOH group formation, as well as the enhanced oxygen mobility. This catalyst was shown stable after a minor deactivation (≤8%) during six recycling uses and its activity could be entirely regenerated by calcination in air.

According to the results of the aerobic oxidation of HMF to FDCA over non-noble metal based catalysts (Fe-, Co-, and Mn-based), an inferior catalytic selectivity to FDCA is normally shown as compared to noble metal catalysts. The activation of O2 is not very e fficient (especially Fe- and Co-based) and a strong oxidant (e.g., *t*-BuOOH) is more normally required to obtain a good selectivity of FDCA. Especially, base additives and organic solvent are often required to improve the FDCA yield. The stability of non-noble metal based catalyst su ffers from more issues, i.e., metal leaching (especially Ni- and Cu-based), change of the active phase, and coverage of active species. Still, owing to the advance in the cost, it is promising to further explore non-precious transition-metal catalysts considering the lower catalyst cost for the upcoming practical production of FDCA. Therefore, further effort should be continuously devoted to develop new approaches for designing e fficient non-precious transition-metal catalysts for the aerobic oxidation of HMF to FDCA.

#### **4. Conclusions and Perspectives**

The catalytic aerobic oxidation of biomass-derived HMF to FDCA is currently a hot topic, especially since FDCA exhibits the potential to replace petrochemical-derived terephthalic acid, one of the most widely used monomers in polymers, for the production of a series of biopolymers. According to the recent results of the aerobic oxidation of HMF to FDCA over supported metal catalysts, noble-metal catalysts are the most studied. Much progress and numerous breakthroughs have already been made in catalyst design and understanding the reaction mechanism. Although the application of inexpensive transition-metal catalysts might o ffer promising prospects in the practical synthesis of FDCA, the main issue of the non-noble metal-based catalyst is the inferior selectivity for FDCA compared to the noble-metal analogues, based on currently reported methods. Additional research e fforts should be devoted to develop new methods based on non-noble transition-metal catalysts that can improve the selectivity of FDCA, especially with O2 as the oxidant and without using additional base. The performance of the catalyst and the reaction pathway are highly dependent on the properties of the catalyst (i.e., the active phase, support, particle size) and reaction conditions (i.e., oxygen pressure, oxygen flow rate, pH, and temperature). Among the noble metal-based catalysts, Au-based catalysts appear to show a better performance in catalyst selectivity and stability for the aerobic oxidation of HMF into FDCA in water, compared to the Pt-, Pd-, Ru-, and Rh-based catalysts, owing to the better resistance to water and O2. Nevertheless, deactivation of the Au-based catalysts by the deposition of the byproducts or intermediates on its active sites is also observed, as mentioned in some examples. For further improvement of the Au-based catalyst, a bimetallic alloy approach, achieved by alloying of a second metal with Au, has been applied and shown to be e ffective, with a higher catalytic activity and improved stability as compared to the monometallic counterparts. Although many bi-functional combinations have shown promising outcomes with rate enhancement or a synergistic e ffect. More details into how such e ffect of di fferent function sites comes from metals or other sites with an optimal

site-balance, have been limitedly studied. Better insight of the reaction mechanisms involved needs to be provided on atomic levels for the oxidation of HMF into FDCA.

Normally, excess base is used to promote the oxidation of HMF, which can not only facilitate the reaction, but also transform the formed FDCA into a salt form dissolved in aqueous solution. Otherwise, the strong adsorption of the carboxylic acids on the catalyst can hinder the further process of this reaction. The use of excess base can however also lead to a more expensive process, which is also less green. Therefore, it is necessary to develop a base-free oxidation system, which is more cost-e ffective and environmentally benign approach, appreciated in sustainable chemistry. Selecting an appropriate support with basic sites for the catalysts has shown advantages in catalytic activity and stability for the base-free oxidation of HMF into FDCA in recent examples. However, catalyst deactivation owing to a loss in basic sites has been observed with the support during reactions, and the stability of the catalyst is found to be a challenge for the catalyst. More e ffort needs to be put into finding out how to stabilize the required functional sites on the catalyst, and advanced strategies of catalyst design for preparing specific structures, i.e., single-atom metal, core-shell, sub-cluster segregated, multi-shell and random homogeneous alloys, etc., deserve more thoroughly exploration in research.

Most reported studies with high FDCA yields in the aerobic oxidation of HMF, have been conducted in dilute HMF solutions (0.5–2.1 wt%), which is unreal for the practical production of FDCA on an industrial scale. The limitation is attributed to the highly reactive functional groups in HMF, which can lead to the formation of undesired solid byproducts, namely humins, via complex side reactions (i.e., condensation and polymerization). Many fewer examples of concentrated HMF substrates are studied in the oxidation of HMF to FDCA. Recently, an approach for stabilizing the active formyl group of HMF by the acetalization with 1,3-propanediol was reported, which enables production of a high yield of FDCA and low humin formation, even in solutions of up to 20 wt% HMF acetal, by aerobic oxidation in the presence of a base additive [38]. In addition, extremely low concentration of HMF, together with their short-term reaction time often applied in reported cases, may underestimate other issues, particularly catalyst stability. Catalyst deactivation might not be well recognized in the liquid phase with a higher concentration of HMF (polar, aqueous and even corrosive). Thus, research e fforts are still needed to conduct the reactions under a more practical concentration of HMF, and further understand the deeper fundamentals of catalyst stability challenges, in order to develop innovative and creative approaches.

The reaction mechanisms involved in the aerobic oxidation of HMF were revealed with some metal-based catalysts, mainly by applying isotope labeling and mass spectrometric techniques. Still, more e fforts need to be devoted to the development of modern in-situ characterization technologies, to provide deep insights into the intrinsic kinetics and mechanisms. In addition, the state-of-the-art *operando* characterization methods combined with various spectroscopy techniques are also necessary for understanding the deep fundamentals of the nature of the intrinsic active sites for each elemental step in HMF oxidation, which might dynamically evaluate during catalysis [102]. The adequate understanding of the reaction mechanism will elucidate a more detailed understanding of catalytic chemistry. The deep insights on the active site can greatly benefit the rational design of catalysts, even with the use of the non-noble metals to prepare more e fficient and stable catalysts for the oxidation of HMF to FDCA.

Although many significant achievements have been made for the aerobic oxidation of HMF to FDCA over metal-based heterogeneous catalysts, further improvements are still required for scaling up to an industrially large-scale production of FDCA. The mass balances for the aerobic oxidation of HMF to FDCA based on laboratory data need to be accurate and correct. The mass balance for the oxidation of HMF was not always mentioned or even not fully closed in many cases. Those unknown parameters might result in a significant amount of economic loss during scaling up processes if not properly done. Current approaches for the pioneering processes for the production of FDCA from HMF are technically feasible but not economically viable, mainly owing to the high price of HMF and its limited availability. Development of energetically and economically viable processes is a long-term task which requires

extensive time, e fforts and normally involves interdisciplinary knowledge of process engineering, chemistry, material science, etc. Only with a full grasp of the knowledge and reorganization of the fundamental details and catalytic challenges, we may develop an economically feasible approach for realizing industrially large-scale production of FDCA in the near future, which will alleviate our society's dependence on the traditional fossil resources.

**Author Contributions:** Conceptualization, S.H. and W.L.; validation, W.L.; formal analysis, L.L. and W.L.; investigation, S.H., L.L. and W.L.; resources A.W. and W.L.; writing—original draft preparation, S.H.; writing—review and editing, L.L.; supervision, A.W. and W.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Key Projects for Fundamental Research and Development of China (2018YFB1501602) and the National Natural Science Foundation of China (21703238, 21690084).

**Acknowledgments:** The authors acknowledge the National Key Projects for Fundamental Research and Development of China (2018YFB1501602), the National Natural Science Foundation of China (21703238, 21690084) and the CAS-TWAS President's Fellowship Program between Chinese Academy of Sciences (CAS) and The World Academy of Sciences (TWAS) for financial support.

**Conflicts of Interest:** The authors declare no conflict of interest.
