As a solution to the environmental and energy problems, the catalytic upgrading of renewable biomass to value-added transportation fuels and chemicals is of key interest [1]. The production of high carbon diesel (C₁₂–C₂₂) and jet-fuel (C₈–C₁₆) range alkanes, which are widely used as transportation fuels, is drawing immense attention [2]. Biomass-derived platform compounds are light molecules with up to six carbon atoms, therefore, a chain growth process is imperative to synthesize high carbon fuel precursors [3]. After total hydrodeoxygenation, thus-prepared oxygen-rich fuel precursors can be converted to alkanes and used as transportation fuels [4].

Following the pioneering work of Dumesic et al. [5], several types of C–C bond coupling reactions, such as aldol condensation [6, 7], hydroxyalkylation/alkylation [8-11], Diels-Alder reaction [12], benzoin coupling [13], ketonization [14], Michael addition and Robinson annulation [15, 16], were developed to produce fuel precursors. Among the C–C bond coupling reactions, aldol condensation is the best known C–C forming synthetic reaction, which involves the transformation of two carbonyl compounds and requires at least one carbonyl compound containing active α-hydrogen atoms [17]. Since biomass-derived carbonyl compounds are numerous, as shown in Scheme 1, employing aldol condensation reaction to increase carbon chain length has attracted considerable interest [18]. Typically, a series of fuel precursors have been obtained by the aldol condensation of furfural with biomass-derived ketones, such as acetone, methyl isobutyl ketone (MIBK), 2-pentanone, 2-heptanone, 3-pentanone, cyclopentanone, acetylacetone, angelica lactone and LA ester [5, 7, 16, 19-25]. In addition, acetone, MIBK and cyclopentanone can be selectively converted to fuel precursors by its self–aldol condensation [26-28]. Specifically, some biomass-derived carbonyl compounds (furfural, acetone, MIBK and cyclohexanone) can be currently produced in industrial scale [18, 29-32], making the promising production of fuel precursors through aldol condensation of biomass-derived carbonyl compounds possible. Besides its application to biomass transformation, the aldol reaction is also important in industry, either in polymers, or in fine chemical and pharmaceutical industry [17, 33, 34]. In short, aldol condensation is a cornerstone of synthetic chemistry and has received immense attention.

Generally, aldol condensation can proceed in the presence of base, acid and acid-base bifunctional catalysts. In most cases, basic catalysts are applied because they can directly abstract α-protons, forming carbanions to attack –C=O bond and then promote C–C bond formation. Homogeneous base catalysts, such as NaOH and Ca(OH)₂ [35-39], are generally employed to catalyze aldol condensation, but have significant disadvantages, including non-recyclability, complex separation, equipment corrosion and waste water production. To overcome these drawbacks, several heterogeneous basic catalysts were used to push aldol condensation. For example, CaO, CaAl-hydrotalcite and MgAl-hydrotalcite were considered as a promising alternative to avoid the above problems [19, 27, 40, 41]. As we know, basic metal oxides have high sensitivity to the ambient CO₂ and face a challenge for stability in the presence of water, while aldol adducts can easily dehydrate to enone groups and produce water. Solid acid catalysts, such as TiO₂, Hf-, Sn-, and Zr-Beta zeolites, also emerged to promote aldol condensation owing to their recyclability and water resistance [42-48]. However, the examples of using acidic materials as catalysts for aldol condensation are still few. In addition, the hydrodeoxygenation of bio-fuel precursors is catalyzed by bifunctional metal/solid acid catalysts, so it would be promising if an excellent solid acid is found for aldol condensation, in this case, the aldol condensation and following hydrodeoxygenation can be integrated over bifunctional metal-acid catalysts and the separation can be omitted. Therefore, it is highly desirable to develop efficient solid acid catalysts to promote aldol condensation.

Nb₂O₅ is an extremely promising catalyst owing to its excellent functionality, such as water resistance, high stability, suitable acidity and C–O bond cleavage ability [49-53]. Therefore, increasing attention has been paid to use Nb₂O₅ as catalyst for the catalytic upgrading of biomass to value-added transportation fuels and chemicals. These excellent properties are fully exploited in the hydrodeoxygenation of biomass and biomass-derived oxygenates. Nb-based catalysts have a unique feature, that is, NbO_x species play a key role in C–O bond cleavage, especially in the cleavage of tetrahydrofuran ring, leading to the excellent performance in the hydrodeoxygenation reaction [52]. Obviously, Nb₂O₅ catalyst exhibited the superior performance in comparison with other solid acid catalysts (ZrO₂, Al₂O₃, TiO₂ and H-ZSM-5) [51-53].

Herein, we report that Nb₂O₅ is taken as a solid acid catalyst to catalyze aldol condensation of biomass-derived carbonyl chemicals and exhibits excellent activity and stability. Investigations on the originate of high activity reveal that Nb₂O₅ catalyst can efficiently activate C=O bond of carbonyl molecules to form a metal enolate intermediate, leading to the excellent performance. We also demonstrate the scope of its application to other biomass-derived carbonyl compounds. Finally, we combine the two functions of Nb₂O₅, namely, cleavage of C–O bond which is reported before [52] and activation of C=O bond, and hydrogenation function of palladium, to develop a one-pot process of the direct production of liquid alkanes from biomass-derived carbonyl molecules over a multifunctional Pd/Nb₂O₅ catalyst.

Furfural, 4-heptanone, MIBK, cyclopentanone, acetone and 2, 4-pentanedione were purchased from Shanghai Titan Technology Co., Ltd. Pd(NO₃)₂ aqueous solution was purchased from Heraeus Materials Technology Shanghai Co., Ltd. The MgO and CaO catalysts were purchased from Aladdin Industrial Inc. Before the tests, the metal oxide catalysts were pretreated in N₂ flow at 500 ℃ for 4 h. All other chemicals were purchased from Sinopharm Chemical Reagent Co., Ltd.

The Nb precursor (niobium oxalate) was prepared according to the literature procedure [54]. Niobium oxalate solutions with a Nb concentration of 0.4 mol L^–1 were prepared for use. The layered Nb₂O₅ was synthesized by changing the substrates and reaction conditions following Murayama's work [55]. Typically, 19.2 g of prepared niobium oxalate and 0.711 g of ammonium oxalate were dissolved in 50 mL deionized water, and then the transparent solution was aged in a Teflon-lined autoclave for 24 h at 180 ℃. After cooling down, the solid was filtered, washed with distilled water and then dried at 50 ℃ overnight. Finally, the Nb₂O₅ sample was obtained by calcination at 400 ℃ for 4 h in air with a linear heating ramp of 10 ℃ min^–1.

2% Pd/Nb₂O₅ catalyst was prepared by incipient wetness impregnation of L-Nb₂O₅ with the aqueous solution of Pd(NO₃)₂. After impregnation, the catalyst was dried at 100 ℃ for 12 h, followed by calcination in air at 400 ℃ for 5 h with a linear heating ramp of 2 ℃ min^–1.

Magnesium-aluminium hydrotalcite (MgAl-HT) with Mg/Al atomic ratio of 2:1 was prepared according to the literature [56]. Typically, an aqueous solution was prepared by adding 0.093 mol Mg(NO₃)₂·6H₂O and 0.0465 mol Al(NO₃)₃·9H₂O to 100 g of H₂O. A second solution containing 0.219 mol NaOH and 0.0565 mol Na₂CO₃ in 100 g of H₂O was slowly added to the Mg/Al aqueous solution under constant stirring at 70 ℃. Subsequently, the precipitate as obtained was aged at 70 ℃ for 20 h, filtered and washed with deionized water until pH = 7. The product was dried in an oven at 80 ℃ overnight and calcined in nitrogen flow at 400 ℃ for 8 h. The mixed oxides were rehydrated in decarbonated water at 80 ℃ for 3 h. Then the activated catalyst was obtained by drying under vacuum at 100 ℃ and denominated as MgAl-HT.

The aldol condensation of furfural and 4-heptanone was performed in a 50 mL Teflon-lined stainless-steel autoclave. Typically, furfural (0.48 g, 5 mmol), 4-heptanone (2.28 g, 20 mmol) and catalyst (0.1 g) were transferred into the autoclave. The reactor was charged to 1 MPa N₂ pressure. The reaction was then performed at 110–140 ℃ under magnetic stirring for 4–7 h. The reactor was quenched in an ice-water bath to stop the reaction immediately. The liquid solution was separated from the solid catalyst by centrifugation and analyzed qualitatively by GC-MS (Agilent 7890A-5975C) and quantitatively by GC equipped with a flame ionization detector (FID, Agilent 7890) using 2-octanone as the internal standard. Conversion of furfural and yield of products were calculated by using the equation: furfural conversion = (moles of furfural reacted) / (moles of starting furfural) × 100%, single aldol adduct yield = (moles of single aldol adduct) / (moles of starting furfural) × 100%, double aldol adduct yield = (moles of double aldol adduct) / (moles of starting furfural / 2) × 100%. The experimental section of aldol condensation involving other molecules is the same as above process.

The one-pot conversion of furfural and 4-heptanone into branched alkanes was conducted in a 50 mL Teflon-lined stainless-steel autoclave. Furfural (0.96 g, 10 mmol), 4-heptanone (0.57 g, 5 mmol), cyclohexane (5.00 g) and 2%Pd/Nb₂O₅ (0.3 g) were transferred into the autoclave. The reactor was charged to 1 MPa N₂ pressure. The reaction was then performed at 130 ℃ under magnetic stirring for 24 h. After the aldol reaction, the reactor was purged with H₂ three times, and charged to 6 MPa H₂ pressure. The reaction was then performed at 220 ℃ and 6 MPa for 24 h. After the HDO reaction, the autoclave was quenched in an ice-water bath to room temperature. The liquid solution was separated from the solid catalyst by centrifugation and analyzed qualitatively by GC-MS (Agilent 7890A-5975C) and quantitatively by GC-FID (Agilent 7890) using C₁₅ chain alkane as the internal standard. Here, we employed pure straight chain alkanes (C₁₂, C₁₇) to replace targeted alkanes as standards, because pure targeted alkanes cannot be purchased or synthesized at present. The product yields were calculated by using the equation: yield = (moles of product) / (moles of starting 4-heptanone) × 100%.

DRIFTS of acetone adsorption were collected with a NICOLET iS50 FT-IR spectrometer equipped with an MCT/A detector. Firstly, the catalysts were pretreated in situ in the cell in N₂ at 400 ℃ for 40 min and background spectra were recorded at 200, 150, 100, and 50 ℃, respectively. Then acetone with N₂ was bubbled into the in situ cell for 40 min. Next, the cell was purged with N₂ at 35 ℃ for 30 min and the adsorption spectra of acetone was recorded. Finally, the temperature was increased to 70 ℃ and 100 ℃, respectively, and the adsorption spectra of acetone were recorded with increasing the temperature.

Powder XRD patterns were recorded with a Rigaku D/max-2550VB/PC diffractometer by using Cu K_α (λ = 0.15406 nm) radiation that was operated at 40 kV and 40 mA.

N₂ adsorption-desorption isotherms were measured at –196 ℃ by using a Micromeritics ASAP 2020 M sorption analyzer. The BET method was used to calculate the specific surface area.

Scanning electron microscopy (SEM) images of samples were recorded on a Nova NanoSEM 450 microscopy at 3 kV. Transmission electron microscopy (TEM) images of samples were obtained on a JEOL Model 2100F electron microscopy at 200 kV.

The actual Pd loading in the sample was detected by inductively coupled plasma-atomic emission spectroscopy (ICP-AES) on a Perkin-Elmer Optima 2100 DV spectrometer.

H₂-TPR was carried out on a Micromeritics AutoChem II 2920 System in 10%H₂/He, and 0.10 g sample was loaded.

Infrared (IR) spectra of pyridine adsorption were recorded on a Nicolet NEXUS 670 FT-IR spectrometer. The samples were pressed into self-supporting disks and placed in an IR cell attached to a closed glass-circulation system. The disk was dehydrated by heating at 400 ℃ for 1 h under vacuum in order to remove the physically adsorbed water. After the cell was cooled to room temperature, the IR spectrum was recorded as background. Pyridine vapor was then introduced into the cell at room temperature until equilibrium was reached, and then a second spectrum was recorded. Subsequent evacuation was performed at 100 ℃ for 10 min followed by spectral acquisitions. The spectra presented were obtained by subtracting the spectra recorded before and after pyridine adsorption.

NH₃-TPD was carried out in an apparatus (PX200, Tianjin Golden Eagle Technology Co. Ltd.) equipped with a thermal conductivity detector (TCD). The samples (100 mg) were loaded into a U-shaped quartz tube. Prior to TPD measurements, the samples were pretreated in flowing N₂ (45 mL min^–1) for 1 h at 500 ℃, and then cooled to 50 ℃. NH₃ was adsorbed onto the samples by exposure to flowing 10% NH₃ in N₂ gas mixture (50 mL min^–1) for 45 min at 50 ℃. Residual and physical adsorbed NH₃ was removed by purging the samples with flowing N₂ (45 mL min^–1) at 90 ℃ for 1 h. Desorption of NH₃ was performed by heating the samples at a rate of 10 ℃ min^–1 under flowing N₂ (45 mL min^–1) from 90 ℃ to 600 ℃.

Temperature-programmed desorption of carbon dioxide (CO₂-TPD) as a probe molecule was carried out on a Micromeritics AutoChem 2920 equipped with mass spectrometry using a quadruple spectrometer (MS OmniStar). 100 mg of the catalyst was placed in a quartz sample tube. Before TPD experiments the catalysts were outgassed at 300 ℃ for 90 min in a flow of helium. Subsequently, the catalysts were cooled down to 50 ℃ and treated under a CO₂ flow for 40 min. Weakly adsorbed CO₂ was removed by flushing with He at 50 ℃ for 30 min. The desorption of CO₂ was measured by heating the catalyst from 50 ℃ to 800 ℃ at a heating rate of 10 ℃ min^–1 in a He flow. The desorbed products were analyzed by mass spectrometry.

A suitable probe reaction is of significant importance for the catalyst screening of aldol condensation. Here, we employed the aldol condensation of furfural with 4-heptanone (Scheme 2) as a probe reaction to screen catalysts. 4-Heptanone is also a biomass-derived molecule and can be synthesized by the esterification-ketonization of biomass-derived n-butyric acid or condensation of ABE fermentation mixture [57-59]. There are three reasons for the selection of this probe reaction. (1) The aldol condensation of furfural with biomass-derived ketones is the most common according to previous reports [3]. (2) Compared with common biomass-derived ketones, the steric hindrance of 4-heptanone is more severe, so the aldol reaction would be more challenging. In this way, the catalyst selected in this reaction can be easily used to other aldol reactions. (3) To our knowledge, there is no report on the production of branched alkanes by using 4-heptanone and furfural as precursors, so we likely provide an alternative strategy for the synthesis of branched alkanes. Generally, aldol condensation is carried out in the presence of basic catalysts. Therefore, common solid-base catalysts, such as MgO, CaO and magnesium-aluminum hydrotalcite (MgAl-HT) (the powder XRD pattern of as-synthesized MgAl-HT see Fig. S1), were tested for the aldol condensation of furfural with 4-heptanone and compared with Nb₂O₅, a solid acid catalyst, the results are shown in Table 1. Surprisingly, Nb₂O₅ has obviously superior activity over the solid-base catalysts; the yield of aldol adducts (FH and FHF) reached 70.8% over Nb₂O₅ catalyst, it was lower than 30% over each solid-base catalyst. Given these results, we first suspected that Nb₂O₅ is an amphoteric metal oxide that may have both acidic and basic sites. It is recognized by all that Nb₂O₅ is an acid catalyst. To check whether Nb₂O₅ has basic sites or not, CO₂-TPD was conducted and the result is presented in Fig. S2. It can be seen that there are no obvious CO₂ desorption peaks, indicating that Nb₂O₅ has no basic sites. As common Lewis acid catalysts, ZrO₂ and Al₂O₃ were also used for the aldol condensation of furfural with 4-heptanone but have relatively poor catalytic performance, and the yields of aldol adducts were only 5.5% and 29.8%, respectively. The above results show that Nb₂O₅ is an excellent catalyst for the aldol condensation of furfural with 4-heptanone. The influence of reaction conditions on the yields of the aldol products was investigated and the results (Fig. S3) show that a reaction at 130 ℃ for 6 h is optimal for the aldol condensation of furfural with 4-heptanone. Theoretically, the prior aldol condensation product should be a β-hydroxyketone. However, the GC result of the aldol condensation of furfural with 4-heptanone (Fig. S4) shows that the obtained products are mainly α, β-unsaturated carbonyl compounds (FH and FHF), rather than β-hydroxyketones. Driven by the conjugated structure of furan ring and C=C bond, the β-hydroxyketone can be easily converted into the stable α, β-unsaturated carbonyl compound. Besides this, it is well known that acid catalysts can promote the dehydration reaction, leading to the conversion of β-hydroxyketones into α, β-unsaturated carbonyl compounds.

To understand the origin of the excellent activity of Nb₂O₅ catalyst, first of all, nitrogen adsorption/desorption was carried out to measure the BET surface areas of Nb₂O₅, ZrO₂ and Al₂O₃. Obviously, the BET surface area of Nb₂O₅ (207 m² g^–1) is close to that of Al₂O₃ (218 m² g^–1) and higher than that of ZrO₂ (66 m² g^–1), but it is inconsistent with the trend of catalytic activity. Therefore, we can conclude that the difference of the BET surface area cannot significantly affect the catalytic activity of the catalysts in the aldol condensation.

First, we guessed that the amount of acidity was responsible for the catalytic performance and therefore NH₃-TPD was carried out to measure the amount of acid and was shown in Fig. 1. It can be seen that Al₂O₃ catalyst has markedly higher acid amount than Nb₂O₅ and ZrO₂, but cannot follow the sequence of catalytic activity. Interestingly, it is noticed that Nb₂O₅ has a certain amount of Br nsted acid sites but ZrO₂ and Al₂O₃ have hardly any according to the result of Py-FTIR spectra (Fig. 2). The clear difference offers a possibility, that is, the excellent performance of Nb₂O₅ catalyst may be attributed to the Br nsted acid sites on the catalyst. In order to verify this, we employed Amberlyst-15 as catalyst, which is a typical Br nsted acid catalyst, to conduct the aldol condensation under the same conditions. The results showed that the conversion of furfural and the yield of aldol adducts were 50.5% and 40.6%, respectively, significantly lower than that (82.2% and 70.8%) over Nb₂O₅ catalyst, which indicated that Br nsted acid sites are not the main origin of the excellent activity of Nb₂O₅ catalyst.

For understanding the origin of high activity, DRIFTS experiments of acetone adsorption were performed (Fig. 3) to test the activation ability of Nb₂O₅, ZrO₂ and Al₂O₃ to C=O bond. Acetone, which is a simple biomass-derived carbonyl molecule, was chosen as the probe molecule. For acetone, the peak appeared at 1736 cm^–1 region is ascribed to the C=O stretching vibration, the band at 1213 and 1366 cm^–1 are attributed to C–CH₃ deformation modes. After adsorption of acetone on these three catalysts for 30 min and purging with N₂ for 30 min at 35 ℃, a red shift in the C=O stretching vibration was observed in all three catalysts. But, the red shift for the Nb₂O₅ catalyst (1695 cm^–1, Fig. 3a) is much larger than that for ZrO₂ (1703 cm^–1, Fig. 3b) and Al₂O₃ catalysts (1703 cm^–1, Fig. 3c). The red shift of the C=O bond can be interpreted as the strong adsorption of C=O bond on catalyst, leading to weakness in the strength of the C=O bond. This result indicates that Nb₂O₅ catalyst has stronger activation ability for C=O bond than ZrO₂ and Al₂O₃, which would be the origin of the excellent activity of Nb₂O₅ catalyst. To explore the activation behavior of the C=O bond by reaction temperature, the cell was purged with N₂ at 70 and 100 ℃, respectively, and the adsorption spectra of acetone were recorded. It can be seen that the red shift is more severe with the increase of temperature and the red shift on Nb₂O₅ catalyst was also more evident (1681 cm^–1) than that of ZrO₂ (1695 cm^–1) and Al₂O₃ catalysts (1700 cm^–1). The striking difference further confirms that Nb₂O₅ has the unique ability for C=O bond activation. It is worth noting that the intensity of a new band at 1557 cm^–1 gradually increased over Nb₂O₅ with the rise of temperature, but it did not occur on ZrO₂ and Al₂O₃. The band at 1557 cm^–1 can be assigned to C=C stretching vibration. There are two possibilities for the origin of C=C stretching vibration: (1) with the activation of the C=O bond, it is gradually converted into an enolate intermediate containing a C=C bond, and thus-generated enolate can undergo nucleophilic addition with another carbonyl molecule to get target product; (2) with the increasing of temperature, the self-aldol condensation of acetone could happen over Nb₂O₅, thus-generated adduct is a typical α, β-unsaturated carbonyl compound containing a C=C bond, so the band at 1557 cm^–1 may be interpreted as the C=C bond of the self-aldol adduct of acetone. The two possibilities are both in good consistent with the observed catalytic activity. Meanwhile, the enolate intermediate certainly exists in every case, it is merely due to poor stability of the enolate intermediate and cannot be seized by the FT-IR spectra in the second case. Overall, the DRIFTS study of acetone adsorption reveals that Nb₂O₅ catalyst has excellent activation ability for C=O bond and thus generates an enolate intermediate for aldol condensation, leading to the high activity.

Based on the DRIFTS study of acetone adsorption, we propose a possible reaction mechanism of Nb₂O₅-catalyzed aldol condensation for two carbonyl compounds, i.e., carbonyl compound 1 containing active α-hydrogen atoms with carbonyl compound 2, as shown in Scheme 3. First, the metal center of Nb₂O₅ polarizes the C=O bond of 1, which forms a metal enolate. Subsequently, the metal enolate undergoes nucleophilic addition to C=O bond in molecule 2 and form a new C–C bond, which affords 3. Finally, 3 can be easily converted to an enone compound 4 by the dehydration reaction in the presence of acidic Nb₂O₅.

We were interested to extend the applicability of Nb₂O₅ to other biomass-derived carbonyl molecules-involved aldol condensation. To explore the scope of the reaction, we investigated two typical aldol condensation, the cross condensation of two different carbonyl molecules and the self-condensation of the same carbonyl molecule (for a structural analysis, see Figs. S5–S11 in the Supporting Information). Firstly, the aldol condensation of furfural with biomass-derived ketones, such as acetone, methyl isobutyl ketone (MIBK) and 2, 4-pentanedione, were carried out over Nb₂O₅ catalyst under the same conditions as the probe reaction system and the results are summarized in Scheme S1. It was found that a high yield of aldol adducts was obtained in the acetone system and 2, 4-pentanedione system. However, in the MIBK system, the yield (60.2%) of aldol adducts is medium, indicating steric hindrance of MIBK could be more obvious than acetone and 2, 4-pentanedione. Clearly, the alpha-H of 2, 4-pentanedione is more active than that of MIBK owing to the presence of two electron withdrawing groups in 2, 4-pentanedione, which is a possible factor that higher yield of aldol adducts was obtained in the 2, 4-pentanedione system. We tried to increase reaction time to improve the yield of aldol adducts and the results (Scheme 4) show that extending the reaction time from 4 h to 12 h significantly increased the yield of the aldol adducts of furfural with MIBK (from 60.2% to 81.7%). In addition, the self-aldol condensation of cyclopentanone over Nb₂O₅ catalyst was carried out and the result showed that a moderate yield of the aldol adducts was obtained. Inspired by the MIBK system, we extended reaction time to 12 h and achieved a high yield (87.0%) of aldol adducts (Scheme 4). The above results fully demonstrate that Nb₂O₅ is an efficient catalyst for biomass-related aldol reactions.

The recyclability of the Nb₂O₅ catalyst for the aldol condensation of 4-heptanone with furfural was investigated by reusing the catalyst in consecutive catalytic runs. After reaction, the color change of the catalyst from original white to canary yellow was observed, and the conversion of furfural and the yield of aldol adducts decreased from 55.0% to 20.3% and from 48.9% to 17.9%, respectively. To discover the reason for this catalyst deactivation, thermogravimetric analysis (TG) was performed to characterize the deactivated catalyst. The result (Fig. S12) showed that ca. 12% carbon was deposited on the deactivated catalyst, which was ascribed to polymerization of the reactants and/or products in the presence of acid catalyst [7, 40]. After calcination at 400 ℃ for 1 h, the catalyst was re-generated and then recharged for the next run. It can be seen from Fig. 4 that no obvious change was found in the catalytic activity (58.6% conversion and 50.4% yield). Re-measurement of XRD pattern (Fig. S13) shows that there was no phase change after reaction, and N₂ sorption shows that the surface area was still high (170 m² g^–1). The above result clearly demonstrates the excellent reusability of Nb₂O₅ catalyst for aldol condensation.

Previously, our group demonstrated that metal-supported Nb-based catalysts, such as Pd/NbOPO₄, Pt/NbOPO₄, Pd/Nb₂O₅ and Pd/Nb₂O₅/SiO₂, are excellent catalysts for the total hydrodeoxygenation of biomass and biomass-derived oxygenates [48-53, 60]. This is because of the abilities of noble metals for H₂ activation and dissociation, NbO_x species for C–O bond cleavage and acid sites on support for dehydration. Given this and the above results, we speculated that a one-pot process for the direct production of liquid alkanes from biomass-derived carbonyl molecules would be possible over a multifunctional metal-supported Nb₂O₅ by combining the aldol condensation with the following total hydrodeoxygenation. Therefore, we draw an experimental scheme to realize one-pot process by using Pd/Nb₂O₅ as catalyst and the detailed process is presented in the experimental section. The details of the Pd/Nb₂O₅ catalyst characterization are provided in the Supporting Information. A key in the one-pot two-step process is to change the reaction atmosphere, as the aldol condensation and the following HDO should be conducted in the atmosphere of nitrogen (as a shielding gas) and hydrogen (as a reactant gas), respectively. Firstly, furfural, 4-heptanone, cyclohexane and 2%Pd/Nb₂O₅ were transferred into the autoclave to conduct aldol condensation. The reactor was charged to 1 MPa N₂ pressure as a shielding gas. After the aldol reaction, the reactor charged to 6 MPa H₂ pressure to perform the following HDO reaction. After the HDO reaction, the liquid solution was separated from the solid catalyst by centrifugation and analyzed qualitatively by GC-MS and quantitatively by GC. After the one-pot two-step process, no oxygen-containing products were observed in the reaction effluent. The yield of alkanes is shown in Fig. 5. As expected, the yield of C_12–17 alkanes reached up to 61.7%, starting from 4-heptanone and furfural. A small amount of C–C cleavage product (C₁₀ alkane) was generated by the hydrocracking of the branched C₂ alkyl groups because the hydrocracking is easier at the branched carbon atom [11]. In addition, C₇ alkanes were produced owing to incomplete reaction of 4-heptanone. The condensation product yield was 66.4% after the first step. Compare for the total yield (61.7%) of one-pot process, the above result indicates that the aldol condensation is the rate-determine step and the following hydrodeoxygenation is easy to implement, which is consistent with our previous results [48, 49]. In short, the total yield of C_10–17 alkanes reached up to 62.5% in the one-pot process. The above results integrating the aldol condensation with the total HDO over Pd/Nb₂O₅ are of interest to save considerable energy consumption and time required for intermediate separation processes, which might be an extremely promising reaction system for the industrial-scale production of liquid alkanes from biomass-derived carbonyl molecules.

Based on the above results and previous reports [51, 52], Pd/Nb₂O₅ as a multifunctional catalyst plays four roles in the conversion of biomass-derived carbonyl molecules into liquid alkanes: (1) Nb₂O₅ helps to activate C=O bond of carbonyl molecules to promote aldol condensation; (2) the metal is used for H₂ activation and dissociation; (3) NbOx species promote C–O bond cleavage; and (4) the niobium-based solid acid catalyzes the dehydration.

We have presented the exceptional activity and reusability of Nb₂O₅ catalyst for the aldol condensation of biomass-derived carbonyl compounds. Among the investigated catalysts (CaO, MgO, MgAl-HT, ZrO₂, Al₂O₃ and Nb₂O₅), Nb₂O₅ showed superior activity for the aldol condensation of furfural with 4-heptanone. The yield of aldol adducts over Nb₂O₅ catalyst reached about 70%, while it was lower than 30% over other catalysts. The superior activity of Nb₂O₅ catalyst for aldol condensation is found to originate from its excellent activation ability for C=O bond, which plays an important role in forming a metal enolate intermediate and leading to the high performance. Meanwhile, other biomass-related aldol reactions can be efficiently conducted to produce a series of fuel precursors over Nb₂O₅ catalyst. Considering that Nb-based catalysts could effectively promote the C–O bond cleavage, we provide a procedure for performing the aldol condensation and the total hydrodeoxygenation in one-pot over multifunctional Pd/Nb₂O₅ catalyst. This work not only provides a highly efficient catalyst for known aldol condensation, but also promotes energy-saving conversion of biomass-derived carbonyl compounds into transportation fuels.

The authors declare no competing financial interest.

This work was financially supported by the National Natural Science Foundation of China (No. 21832002, 21872050, and 21808063), the Fundamental Research Funds for the Central Universities (222201718003), the Science and Technology Commission of Shanghai Municipality (18ZR1408500, 10dz2220500), and the "Zhang Jiangshu" Excellent Ph.D. Project of ECUST.