Selective ketonization of propionic acid on Fe-MFI zeolites: crucial roles of acid strength and density

Abstract

The ketonization reaction offers a convenient approach to remove oxygen and increase the carbon chain length of carboxylic acids without consuming H₂. Conventional zeolites with a strong Brønsted acid site (BAS) are very active for ketonization; however, they suffer from low ketone selectivity and fast deactivation owing to the facile secondary and tertiary reactions occurring on the strong BAS. Herein, a series of Fe-MFI zeolites (Si/Fe = 80–180) with a weaker BAS, i.e., Fe–OH–Si, than Al-MFI were prepared, characterized and tested for ketonization of propionic acid at 350 °C and atmospheric pressure. Compared with Al-MFI-180, although Fe-MFI-180 with its weaker BAS strength moderately reduces the activity for propionic acid ketonization (turnover frequency (TOF) of 6.97 and 3.80 min⁻¹, respectively), it significantly reduces the activity of secondary (aldol condensation) and tertiary (aromatics formation) reactions (3-pentanone conversion TOF of 1.85 and 0.33 min⁻¹, respectively), resulting in 3-pentanone as the dominant primary product as well as improved stability. The Fe-MFI zeolites not only showed high TOF (2.00–3.80 min⁻¹) but also improved the selectivity for 3-pentanone and enhanced stability compared to Al-MFI with strong BAS. These results demonstrate a strategy for weakening the strength of BAS of zeolites to reduce the activity of the secondary and tertiary reactions and thereby improve the selectivity and stability of ketonization of carboxylic acids.

---

1. Introduction

Abundant and renewable biomass is considered a promising resource that can serve as an alternative or supplement to fossil fuels and has attracted extensive interest and research.^1–3 Carboxylic acids with varying carbon chain lengths can be produced via depolymerization of cellulose and hemicellulose of biomass,^4,5 fermentation of wet waste,⁶ and hydrolysis of triglycerides in fats and oils,⁷ and therefore represent an important type of biomass derived oxygenate. The acidic and corrosive properties of carboxylic acids create difficulties in their storage, transportation and utilization,^8–11 making conversion of carboxylic acids necessary. Ketonization reaction converts two carboxylic acids into a longer chain ketone and eliminates CO₂ and H₂O in the meantime, which offers a feasible approach to partially remove oxygen and increase the carbon chain length without consuming H₂. Moreover, the produced ketones are important chemicals and can be further converted.^12,13

Various types of metal oxide-based catalysts have been explored for ketonization of carboxylic acids. Among these metal oxides, the amphoteric TiO₂, CeO₂ and ZrO₂ are the most intensively investigated owing to their high activity.^14–17 Both acid and base sites of metal oxides are involved in catalyzing different elementary steps of ketonization.^18,19 Thus, acid–base properties of metal oxides play a crucial role in ketonization.^20–22 The earth abundant FeO_x catalysts also show notable activity in the ketonization reaction.^23–27 It has been reported that propionic acid conversions of 30% and >90% were achieved on FeO_x prepared via hydrothermal and nanocasting syntheses, respectively, at 335 °C and space time of 2 h.²⁶ The core–shell Fe@Fe₃O₄ catalyst showed a 70% yield of acetone in ketonization of acetic acid at 400 °C and space time of 0.06 h.²³ The complex and inhomogeneous surface structure of the FeO_x catalyst may create controversy and uncertainty on active site structures.

Zeolite catalysts have also been widely explored for the conversion of carboxylic acids.^28–34 The Brønsted acid sites (BAS) of HZSM-5 act as active sites to mediate efficient conversion of carboxylic acids to ketones.²⁹ However, the BAS are also very active for the secondary and tertiary reactions, resulting in significantly lowered selectivity for ketones due to further conversion. Recently, a heteroatom zeolite with a Lewis acid site (LAS) obtained via substitution of Al with 4+ valence metals has been reported to improve the selectivity for ketone by retarding the secondary and tertiary reactions.^28,35,36 Among these zeolites, the Zr-Beta and TS-1 show excellent selectivity and stability for ketonization, which achieve 3-pentanone selectivity >95% at propionic acid conversions of 57.5% and 70% (350 °C and space time of 2 h), respectively.^35,36 Although Zr-Beta and TS-1 show excellent performance for ketonization, the preparation of these zeolites is complex and costly, which may hinder their practical applications. In addition, although the selectivity toward ketone is significantly improved, the activity of heteroatom Lewis acid zeolites is much lower than that of the conventional zeolites with BAS. Hence, there is still an urgent need to develop inexpensive zeolites with high activity and selectivity for ketonization of carboxylic acids.

Al-MFI zeolites, i.e., HZSM-5, have been widely applied in many industrially important reactions, due to their medium pore size and strong acid sites. However, their strong acidity also limits the application of Al-MFI zeolites for reactions that only require weak acidity.^37,38 Due to the similarity of Fe³⁺ and Al³⁺, Fe³⁺ is highly accessible to the framework of zeolites, which makes the synthesis of zeolites with elevated Fe content in the framework a relatively straightforward process. Fe-MFI zeolites with relatively weaker acidity have been reported to show excellent performance for a number of reactions, such as the conversion of methanol or dimethyl ether to olefins.^38–41 However, there is no open research reported focusing on the application of Fe heteroatom zeolites for ketonization of carboxylic acids.

In this contribution, a series of Fe-MFI zeolites were synthesized hydrothermally, fully characterized, and applied for ketonization of propionic acid. The results demonstrated that the Fe-MFI contains mainly BAS (framework bridging Fe–OH–Si) with weaker acid strength relative to Al-MFI, which significantly reduces the secondary and tertiary reactions for further conversion of the primary product of 3-pentanone, therefore resulting in 3-pentanone as the dominant product as well as improved stability. Moreover, though the Fe-MFI is less active than Al-MFI, it is much more active than heteroatom zeolites with LAS.

2. Experimental

2.1. Preparation of Fe-MFI and Al-MFI zeolites

All Fe-MFI and Al-MFI zeolites were synthesized via a hydrothermal process.⁴² Initially, 10.72 g tetraethyl orthosilicate (TEOS, 98%, ACS reagent grade, Kermel) was added dropwise to a mixture of 12.24 g tetrapropylammonium hydroxide (TPAOH, 40 wt% in water, Shanghai DiBo) and 26.8 g deionized H₂O under vigorous stirring for 2 h. Subsequently, a calculated amount of Fe(NO₃)₃·9H₂O (98%, Heowns) dissolved in 10 g of H₂O was added to the clear solution under stirring, and then stirred for 4 h. The resultant solution was heated at 65 °C for 0.5 h to evaporate ethanol originating from the hydrolysis of TEOS. Subsequently, the mixture was transferred into a 100 mL Teflon-lined stainless-steel autoclave and subjected to hydrothermal synthesis at 170 °C for 24 h. After that, the solid product was centrifuged, washed with H₂O, dried at 80 °C for 12 h, and finally calcined at 550 °C for 6 h in air with a heating rate of 2 °C min⁻¹. The prepared zeolite was designated as Fe-MFI-x, where x represents the Si/Fe molar ratio (x = 180, 160, 140, 120, 100 and 80).

The procedure for the preparation of Al-MFI-180 (Si/Al molar ratio of 180) was similar to that of Fe-MFI, except for the replacement of Fe(NO₃)₃·9H₂O with Al(NO₃)₃·9H₂O (99%, Macklin). A pure silicon zeolite, designated as MFI-0, was prepared by omitting the addition of the aluminum or iron source. A reference sample of Fe₂O₃/MFI-0 (Si/Fe = 180) was prepared via an impregnation method.⁴³ The MFI-0 was impregnated with a Fe(NO₃)₃·9H₂O solution for 12 h, dried at 100 °C for 12 h, and finally calcined in air at 500 °C (which is high enough for decomposition of the Fe precursor) for 4 h with a heating rate of 2 °C min⁻¹. All samples were pressed and crushed to 40–60 mesh for catalytic runs and characterization.

2.2. Characterization

The crystalline structure of zeolites was analyzed by powder X-ray diffraction (XRD) using a Rigaku D/Max-2500 diffractometer with a Cu Kα radiation source (40 kV and 20 mA). The XRD diffractograms were collected in the 2θ range of 5–50° at a rate of 5° min⁻¹. The elemental composition of zeolites was quantified using inductively coupled plasma optical emission spectrometry (ICP-OES, Varian). The samples were dissolved in HF and then complexed with a saturated boric acid solution. The physical property of the zeolites was measured via N₂ adsorption on an APSP 2460 analyzer (Micromeritics). Prior to the test, 150 mg catalyst was degassed at 300 °C for 8 h under vacuum. The specific surface area (S_BET) was calculated using the Brunauer–Emmett–Teller (BET) method. The micropore surface area and pore volume were determined by the t-plot method, and the total pore volume was obtained at P/P₀ = 0.99. The morphology of zeolites was observed using an Apreo S LovAc scanning electron microscope (SEM) at an accelerating voltage of 30 kV. High angle annular dark field scanning transmission electron microscopy (HAADF-STEM) observation and energy dispersive X-ray spectroscopy (EDX) analysis were performed on a JEOL JEM-F200 electron microscope. The ultraviolet-visible (UV-vis) spectra were recorded on a UV-2600 spectrophotometer (Shimadzu) in the wavelength range of 200–700 nm.

The NH₃ temperature-programmed desorption (NH₃-TPD) was carried out on a Chemisorb 2750 (Micrometrics) with a thermal conductive detector (TCD). 200 mg zeolite sample was placed in a U-tube reactor and pretreated with flowing He (30 mL min⁻¹) at 400 °C for 1 h. After the sample was cooled to 95 °C, NH₃ was introduced for 1 h. Then, the sample was purged with He for 1 h to remove physically adsorbed NH₃. Finally, the sample was ramped up to 500 °C at a rate of 10 °C min⁻¹. Fourier transform infrared (FT-IR) spectra were recorded on a PerkinElmer Frontier spectrometer, equipped with a deuterated triglycine sulfate (DTGS) detector and a transmission cell (Harrick). A self-supporting wafer of 30 mg was placed into the cell and treated with flowing Ar (30 mL min⁻¹) at 350 °C for 1 h. Thereafter, the cell was cooled to 150 °C, and a background spectrum was recorded. Pyridine or propionic acid (99%, J&K) vapor was introduced to the sample for adsorption for 2 h. The vapor and physically adsorbed pyridine or propionic acid was removed by flowing Ar for 2 h. The sample was then heated up to 350 °C. All spectra were recorded with a resolution of 4 cm⁻¹ and 32 scans.

2.3. Catalytic performance

Ketonization of propionic acid (99%, J&K) was investigated in a fixed-bed tubular quartz reactor.³⁵ Zeolite sample (40–60 mesh) was loaded in the middle of the reactor tube, and was pretreated at 350 °C for 0.5 h in flowing Ar (30 mL min⁻¹). Subsequently, propionic acid was introduced via a KDS-100 syringe pump (KD Scientific), and was vaporized at a high temperature before entering the reactor. The partial pressure of propionic acid was maintained at 4 kPa for typical runs. All lines were maintained at 220 °C to prevent condensation. The products were analyzed online on a gas chromatograph (GC, Agilent 7890B) equipped with an HP-INNOwax column (60 m × 320 μm × 0.5 μm) and a flame ionization detector (FID). The space time (W/F, h), defined as the ratio of the weight of the catalyst (W, g) to the organic feed mass flow rate (F, g h⁻¹), was varied in the range of 0.04–8 h by varying both the mass of the catalyst (0.0125–0.1500 g) and the feed flow rate (0.02–0.30 mL h⁻¹). For a typical run with a W/F of 2 h, the mass and volume of the catalyst, the propionic acid feed flow rate, and the residence time were 0.0990 g, 0.15 cm³, 0.05 mL h⁻¹ and 1.5 s, respectively. The conversion of propionic acid and selectivity for products were calculated by equations reported in previous work.⁴⁴ The carbon balance was >94% for all runs.

3. Results and discussion

3.1. Characterization of Al-MFI and Fe-MFI zeolites

The XRD patterns of all samples show characteristic MFI diffraction peaks at 2θ of 7.9°, 8.8°, 23.1°, 23.9° and 24.4° (Fig. S1†), confirming the successful formation of the MFI structure in all zeolites. Notably, the diffraction peaks of neither Al₂O₃ nor Fe₂O₃ are discernible, which indicates that the Al or Fe species are highly dispersed in all zeolites. The crystallinities of Al-MFI and Fe-MFI zeolites relative to Fe-MFI-100, with the highest crystallinity, are all >95% (Table 1), proving that all MFI zeolites are well crystallized and varying the content of Al or Fe has little effect on the crystallinity. Moreover, the XRD estimated crystallite sizes of all MFI zeolites are similar (52–64 nm). In addition, the measured Si/M (M = Al or Fe) molar ratios agree well with the target values (Table 1), indicating that Al or Fe was doped into the zeolites.

Table 1 Physicochemical properties of samples

Sample	Si/M^a	S BET (m^2 g^−1)	S micro (m^2 g^−1)	S external (m^2 g^−1)	Relative crystallinity^e (%)	D (nm)
a M represents Al or Fe, which was determined using ICP. b BET surface areas were calculated using the data in the range of 0.01 < P/P0 < 0.1 and fitted with the consistency criteria. c Determined using the t-plot. d S external = SBET − Smicro. e Crystallite size calculated using XRD.
MFI-0	—	399	272	127	91	64.4
Al-MFI-180	170	427	311	116	95	59.4
Fe-MFI-180	168	431	300	131	96	61.6
Fe-MFI-160	150	427	306	121	99	56.0
Fe-MFI-140	137	401	280	121	97	56.5
Fe-MFI-120	113	397	278	119	95	56.4
Fe-MFI-100	98	424	302	122	100	59.3
Fe-MFI-80	74	411	306	105	99	52.4

---

As illustrated in Fig. S2,† all zeolites exhibit type IV adsorption–desorption isotherms with H3 hysteresis loops within the range of 0.4 < P/P₀ < 0.99. The H3 type corresponds to slit-shaped pores formed by the aggregation of plate-like particles, which are indicative of capillary condensation in the mesopores of zeolites originating from agglomerates of crystallites.⁴⁵ As reported in Tables 1 and S1,† all zeolite samples display comparable specific surface area (S_BET, 397–431 m² g⁻¹), micropore area (S_micro, 272–311 m² g⁻¹), external surface area (S_external, 105–131 m² g⁻¹), total pore volume (V_total, 0.204–0.242 cm³ g⁻¹) and micropore volume (V_micro, 0.126–0.141 cm³ g⁻¹), implying that varying the Si/Fe molar ratio has a negligible effect on the pore structure of Fe-MFI zeolites. In addition, the surface areas of mesopores estimated by the BJH method (S_meso, Table S1†) are also similar, 52–74 m² g⁻¹ for all zeolites. Thus, we can conclude that these Fe-MFI and Al-MFI samples have similar textural properties. It should be noted that the mesopore volume of Fe-MFI decreases slightly with decreasing the Si/Fe ratio, which may be attributed to increased clogging of the pores by the agglomeration of Fe species.

SEM images of MFI-0, Al-MFI-180, Fe-MFI-180 and Fe-MFI-120 zeolites all show uniform nanoscale elliptical shapes with similar sizes of 366 ± 36 nm, 337 ± 40 nm, 321 ± 36 nm, and 333 ± 46 nm, respectively (Fig. S3†). These sizes are much larger than the primary crystallite sizes calculated by XRD (Table 1), indicating that the several primary crystallites may be assembled into a large crystallite, which is also confirmed by TEM (Fig. S4†). Both the primary crystallite sizes estimated from XRD and crystallite sizes estimated from SEM only varied slightly for all zeolites, which indicate similar diffusion pathlengths for all zeolites and therefore should have a similar effect on intra-crystallite diffusion of reactants. The HAADF-STEM images of Fe-MFI-180 and Al-MFI-180 (Fig. 1) further confirm the elliptical shape morphology. Elemental mapping images show that the Fe or Al, Si and O are homogeneously dispersed in both zeolites. No aggregation of Fe or Al species can be identified.

Fig. 1 (a) HAADF-STEM image and (b–d) EDX mapping analysis of Al-MFI-180; (e) HAADF-STEM image and (f–h) EDX mapping analysis of Fe-MFI-180.

The fine coordination structure of Al or Fe species in MFI zeolites was probed by UV-vis. As shown in Fig. 2a, the band at wavelength <250 nm can be attributed to tetrahedrally coordinated Fe³⁺ or Al³⁺ species in the framework.^46,47 The bands in the ranges of 250–350, 350–450, and >450 nm can be assigned to oligomeric FeO_x or AlO_x species within the zeolite channel, FeO_x or AlO_x clusters^46,48 and bulk metal oxides,^46,47 respectively. The Fe₂O₃/MFI shows broad adsorption bands in the range of 300–650 nm, indicating that the Fe species are mainly in the form of extra-framework FeO_x clusters and particles.⁴⁹ In contrast, the Fe-MFI and Al-MFI samples show bands dominantly at <250 nm, indicating that the framework tetrahedral Fe or Al is the dominant species. The distribution of Fe or Al species derived from curve fittings (Fig. S5†) is reported in Table 2. For both Al-MFI-180 and Fe-MFI-180, the framework tetrahedral species is ∼90%. As the Fe content increases (or Si/Fe decreases), the tetrahedral Fe decreases gradually from 89.8% for Fe-MFI-180 to 75.1% for Fe-MFI-80, while the oligomeric FeO_x species increases gradually from 9.3% to 21.8%. This trend indicates reduced framework Fe species while increased extra-framework Fe species with increasing Fe content, which is in agreement with the general rule for preparation of heteroatom zeolites.⁵⁰ Note that the content of both FeO_x clusters (<5%) and particles (<2%) is low in all Fe-MFI samples, and therefore they were unable to be detected in XRD and TEM.

Fig. 2 (a) UV-vis spectra of MFI-0, Fe₂O₃/MFI-0, Al-MFI and Fe-MFI zeolites and (b) NH₃-TPD profiles of MFI-0, Fe₂O₃/MFI-0, Al-MFI and Fe-MFI zeolites.

Table 2 Percentage of Al or Fe species in various MFI zeolites from UV-vis spectra

Catalyst	Framework Al or Fe (%)	Oligomeric Al or Fe (%)	Oxide cluster (%)	Bulk oxide (%)
Al-MFI-180	90.2	6.8	2.9	0
Fe-MFI-180	89.8	9.3	0.9	0
Fe-MFI-160	85.7	11.9	2.4	0
Fe-MFI-140	83.0	12.5	4.4	0
Fe-MFI-120	81.5	17.8	0.6	0
Fe-MFI-100	77.6	18.4	2.3	1.7
Fe-MFI-80	75.1	21.8	1.2	1.8

---

The acidic properties of catalysts were probed by NH₃-TPD (Fig. 2b). The MFI-0 only shows a small peak at a low temperature of 140 °C, pointing to the weak acidity. Compared with MFI-0, the Fe₂O₃/MFI-0 exhibits a slightly stronger peak at 158 °C, which likely results from the weak LASs of Fe cations on FeO_x. Different to MFI-0 and Fe₂O₃/MFI-0, the Al-MFI-180 shows two strong peaks at 170 and 400 °C, which are assignable to NH₃ desorption from weak and strong acid sites, respectively.^49,51 Similarly, the Fe-MFI-180 also displays weak and strong acid sites at desorption temperatures of 170 and 340 °C, respectively. The strong acid sites are related to BAS originating from bridging hydroxyls, i.e., Al–OH–Si or Fe–OH–Si.^40,52,53 It should be noted that the desorption temperature from the strong acid sites on Fe-MFI-180 is lower than that of Al-MFI-180 by 60 °C, reflecting the weaker strength of strong acid sites on Fe-MFI-180 relative to Al-MFI-180. Decreasing the Si/Fe ratio from 180 to 120 has little effect on the desorption peak positions, while a further decrease to 80 causes the peaks to shift slightly to a higher temperature of 350 °C, which likely results from the extra-framework Fe species enhancing the acid strength of vicinal framework Fe species.^54,55 Quantification (with the deduction of base line drift at temperatures >400 °C) results in terms of both mass and surface area (Tables 3 and S2†) clearly indicate that the density of acid sites increases with decreasing the content Si/Fe ratio.

Table 3 Total acid density, BAS density, LAS density, BAS/(BAS + LAS), theoretical acid density, and theoretical BAS density of zeolites

Catalyst	Total acid density^a (μmol g^−1)	BAS density^b (μmol g^−1)	LAS density^b (μmol g^−1)	BAS/(BAS + LAS)^c	Theoretical acid density^d (μmol g^−1)	Theoretical BAS density^e (μmol g^−1)
a Calculated by NH3-TPD. b Calculated by FT-IR spectra of pyridine adsorption. c Density ratio of BAS/(BAS + LAS). d Calculated by theoretical Si/M molar ratio. e Calculated by theoretical acid density and percentage of framework Al or Fe.
MFI-0	18.6	—	—	—	—	—
Fe2O3/MFI-0	35.3	—	1.6	—	—	—
Al-MFI-180	107.2	69.2	4.2	0.942	92.2	83.2
Fe-MFI-180	100.5	61.7	3.9	0.940	92.0	82.6
Fe-MFI-160	107.1	76.5	5.3	0.935	103.4	86.4
Fe-MFI-140	110.6	90.7	11.0	0.892	118.1	96.3
Fe-MFI-120	129.2	100.8	26.4	0.792	137.6	107.2
Fe-MFI-100	147.8	112.8	33.5	0.771	164.9	112.3
Fe-MFI-80	170.1	101.6	43.7	0.699	205.6	128.9

---

The acidic properties of the MFI zeolites were further studied by FT-IR of pyridine adsorption. As illustrated in Fig. 3a, the MFI-0 barely shows pyridine adsorption peaks, reflecting the weak acidic property. In contrast, all Al-MFI and Fe-MFI zeolites exhibit characteristic bands of protonated pyridine on BAS at 1636, 1545, and 1490 cm⁻¹,⁵⁶ pyridine coordinated on LAS at 1621, 1611, 1490, and 1450 cm⁻¹,⁵⁷ and hydrogen bonded pyridine on Si–OH at 1597 and 1447 cm⁻¹.^57–59 Notably, the reference sample of Fe₂O₃/MFI shows hydrogen bonded pyridine at 1597 and 1447 cm⁻¹ as well as coordinated pyridine on LAS at 1450 cm⁻¹, without BAS being detected. To further compare the acid strength of Al-MFI-180 and Fe-MFI-180, temperature programmed desorption of pyridine was also followed by FT-IR (Fig. S6†). The integral band intensities at 1545 and 1450 cm⁻¹ were normalized to respective intensities at 150 °C and plotted as a function of desorption temperature. As illustrated in Fig. 3b, the normalized integral band intensity of BAS on Fe-MFI-180 decreases more quickly than that on Al-MFI-180, which indicates the weaker strength of BAS on Fe-MFI and agrees well with NH₃-TPD results. The normalized integral band intensity of LAS on Fe-MFI-180 (Fig. 3b and S7†) decreases much faster than that of BAS, pointing to the much weaker strength of LAS relative to BAS. In addition, the decrease rate is similar and even slightly faster than that on Al-MFI-180 (Fig. S7†), which indicates similar weak strength of LAS in Al-MFI-180 and Fe-MFI-180. Quantification results of the densities of BAS and LAS using the integral band intensities at 1545 and 1450 cm⁻¹ with the integrated molar extinction coefficients of 1.46 and 2.22 cm μmol⁻¹,⁶⁰ respectively, are summarized in Tables 3 and S2.† It is not surprising that the densities of both BAS and LAS increase with decreasing Si/Fe ratio. However, the fraction of BAS, i.e., the density ratio of BAS/(BAS + LAS), decreases with decreasing Si/Fe, reflecting increased fraction of extra-framework Fe at the expense of framework tetrahedral Fe. Note that the BAS density decreases when the Si/Fe ratio decreases from 100 to 80, suggesting a significant reduction in the fraction of Fe entering the framework. Interestingly, the BAS/(BAS + LAS) can be well correlated with the fraction of framework Fe estimated from the UV-vis spectra (Fig. S8†).

Fig. 3 (a) FT-IR spectra of pyridine adsorption on MFI-0, Fe₂O₃/MFI-0, Al-MFI and Fe-MFI zeolites at 150 °C and (b) normalized integral band intensity at 1545 cm⁻¹ (BAS) as a function of desorption temperature on Al-MFI-180 and Fe-MFI-180. The band intensities in figure (b) were normalized to those at 150 °C.

Propionic acid adsorption at 150 °C (Fig. 4) followed by temperature programmed desorption (Fig. 4 and S10†) was tracked by FT-IR to analyze the surface species during reaction. All samples show a distinctive band centering at ∼1740 cm⁻¹ (Fig. 4a), which can be ascribed to the vibration of C=O (ν(C=O)) of hydrogen bonded monodentate propionic acid on Si–OH groups.^35,61 This feature is accompanied by a negative band at 3725 cm⁻¹ and a broad positive band at 3400 cm⁻¹ (Fig. 4b), which are attributed to the perturbation of Si–OH groups upon adsorption of propionic acid. Apparently, these bands on MFI-0 are much weaker than those on Al-MFI-180, Fe-MFI-180 and Fe-MFI-120 zeolites, which is consistent with the much weaker –OH stretching bands of MFI-0 than other zeolites in the region of 3000–4000 cm⁻¹ (Fig. S9†). There are several small bands in 1500–1300 cm⁻¹, which originate from C–H bending vibrations (δ(C–H)) of CH₃ and CH₂ groups, that is, CH₃ bending (δ(CH₃), 1466 cm⁻¹), CH₂ bending (δ(CH₂), 1418 cm⁻¹), and CH bending (δ(CH), 1386 cm⁻¹).⁶¹ In addition, the peaks at 2989, 2950, 2924 and 2891 cm⁻¹ are attributed to the antisymmetric and symmetric stretching vibration of C–H of adsorbed propionic acid (ν_as(CH₃) at 2989 cm⁻¹, ν_as(CH₂) at 2950 cm⁻¹, ν_s(CH₃) at 2924 cm⁻¹, and ν_s(CH₂) at 2891 cm⁻¹).^35,62

Fig. 4 FT-IR spectra of propionic acid adsorption over MFI-0, Fe₂O₃/MFI-0, Al-MFI-180, Fe-MFI-180 and Fe-MFI-120 at 150 °C in the ranges of (a) 1850–1300 cm⁻¹ and (b) 4000–2500 cm⁻¹.

Compared with MFI-0 and Fe₂O₃/MFI-0, all samples (Fe-MFI-180, Fe-MFI-120 and Al-MFI-180) containing BAS show a distinctive band at 1650 cm⁻¹. This band is located at a lower wavenumber than that of ν(C=O) of propionic acid bonded at Si–OH (1740 cm⁻¹) and framework tetrahedral Ti centers of Ti-MFI (1680 cm⁻¹),³⁵ indicating a stronger perturbation of the C=O bond due to interaction with active centers. In addition, Fe-MFI-120 with higher density of BAS shows stronger band intensity than Al-MFI-180 and Fe-MFI-180. Therefore, we assign this band to ν(C=O) of protonated propionic acid at BAS.⁶³ Increasing the temperature from 150 °C to 350 °C resulted in a reduction in intensity for all bands (Fig. S10†), which can be related to either desorption or conversion of adsorbed propionic acid. The normalized band intensities at 1740 and 1650 cm⁻¹ relative to those at 150 °C were plotted as a function of desorption temperature (Fig. 5). The band at 1650 cm⁻¹ decreases more rapidly than that at 1740 cm⁻¹ for all samples, suggesting that the conversion of propionic acid on BAS is more pronounced than the desorption of propionic acid from Si–OH groups. In particular, the decrease rate of the band at 1650 cm⁻¹ follows the order of Al-MFI-180 > Fe-MFI-180 > Fe-MFI-120, which agrees with the order of turnover frequency (TOF) for propionic acid conversion (see below). In addition, MFI-0 that only contains propionic acid on Si–OH at 1740 cm⁻¹ exhibits negligible activity, which excludes the possibility of Si–OH groups as the active sites. Hence, BAS are the likely active sites for ketonization.

Fig. 5 Normalized integral band intensities at 1650 cm⁻¹ (solid lines, BAS) and 1740 cm⁻¹ (dotted lines, Si–OH) of FT-IR spectra during temperature-programmed desorption of propionic acid over Al-MFI-180, Fe-MFI-180 and Fe-MFI-120. The band intensities were normalized to those at 150 °C.

3.2. Effect of acid strength of BAS

To explore the effect of acid strength of BAS, Al-MFI-180 and Fe-MFI-180 with identical heteroatom contents were investigated for conversion of propionic acid at 350 °C as a function of space time (W/F). The Al-MFI-180 is much more active than Fe-MFI-180, since a much shorter W/F (2 h vs. 8 h) is required to achieve 100% conversion of propionic acid (Fig. 6a). The product distributions as a function of conversion, which results from varying W/F, are compared in Fig. 6b and c. For Al-MFI-180 (Fig. 6b), 3-pentanone is the dominant product at low conversions (<30%), indicating that ketonization of propionic acid is the primary reaction (Fig. 7). The yield of 3-pentanone passes a maximum of 44.0% at a conversion of 65.0%, and then decreases to 25.9% at a conversion of 98.1%. In contrast, the yield of aromatics increases significantly to 56.1% at a conversion of 98.1%, becoming the major product. In the meantime, the yields of C_1–2, C₃, and C_4–7 also increase at high conversions. These trends of product evolution as a function of conversion agree with previously proposed major reaction pathways on zeolites with BAS (Fig. 7),^29,44,64i.e., (1) ketonization of propionic acid to 3-pentanone; (2) aldol condensation of 3-pentanone to aldol dimer and/or trimer, which may form aromatics directly through dehydration, ring closure and hydride transfer reactions;^44,65 (3) the aldol dimer and trimer may also undergo cracking to C_4–7 hydrocarbons first, and then form aromatics; (4) dealkylation of aromatics and cracking of C_4–7 produce C_1–3 hydrocarbons. A different product evolution behavior is observed on Fe-MFI-180 (Fig. 6c). The yield of the primary product (3-pentanone) always increases with increasing conversion, and reaches 62.5% at a conversion of 98.6%. Although the yields of aromatics C_1–2, C₃, and C_4–7 increase at conversion >60%, they appear as minor products even at conversions approaching 100%. These trends indicate that the aldol condensation and sequential tertiary reactions are strongly inhibited on Fe-MFI-180 with weaker acidity.

Fig. 6 (a) Conversion of propionic acid as a function of W/F over Al-MFI-180 and Fe-MFI-180; yield of products as a function of propionic acid conversion over (b) Al-MFI-180 and (c) Fe-MFI-180. Reaction conditions: T = 350 °C, P_{propionic acid} = 4 kPa, time on stream (TOS) = 0.5 h.

Fig. 7 Proposed major reaction pathway of propionic acid conversion on Al-MFI-180 and Fe-MFI-180. The thickness of the arrows is indicative of the reactivity of ketonization and aldol condensation over Al-MFI-180 and Fe-MFI-180.

The conversion of propionic acid and selectivity to 3-pentanone for Al-MFI-180 and Fe-MFI-180 were compared at an identical W/F of 2 h (Table 4). Although the Al-MFI-180 shows much higher conversion of propionic acid than Fe-MFI-180 (98.7% vs. 59.0%), the selectivity for 3-pentanone is much lower (26.2% vs. 85.9%). In comparison, the conversion is much lower on the reference samples of Fe₂O₃/MFI-0 (24.8%) and MFI-0 (2.3%) without BAS, pointing to that the BAS is essential to achieve high conversions. The TOF of propionic acid conversion was estimated using the intrinsic reaction rate measured under differential conditions (conversion <15%) free of mass transfer limitations (see details in Fig. S11 and Table S3†) and it was assumed that the Fe species are atomically dispersed in zeolite. As listed in Table 4, the TOF on Fe-MFI-180 (3.80 min⁻¹) is about 1/2 of that on Al-MFI-180 (6.97 min⁻¹), while it is 4.4 times higher than that on Fe₂O₃/MFI-0 (0.85 min⁻¹). This comparison indicates that Fe-MFI-180 with BAS is more efficient for ketonization than Fe₂O₃/MFI-0 without BAS, while it is less active than Al-MFI-180 with stronger BAS.

Table 4 Conversion of propionic acid, selectivity for 3-pentanone, TOF of propionic acid conversion and TOF of 3-pentanone conversion over different catalysts

Catalysts	Conversion^a (%)	Selectivity^a (%)	TOFketonization^b (min^−1)	TOFaldolization^c (min^−1)
a Conversion of propionic acid and selectivity for 3-pentanone. Reaction conditions: T = 350 °C, W/F = 2 h, PPropionic acid = 4 kPa, TOS = 0.5 h. b TOF of propionic acid conversion, reaction conditions: T = 350 °C, Ppropionic acid = 4 kPa, TOS = 0.5 h, conversion of propionic acid <15%. c TOF of 3-pentanone conversion, reaction conditions: T = 350 °C, P3-pentanone = 4 kPa, TOS = 0.5 h, conversion of 3-pentanone <15%.
MFI-0	2.3	86.0	—	—
Fe2O3/MFI-0	24.8	91.9	0.85	—
Fe-MFI-180	59.0	85.9	3.80	0.33
Al-MFI-180	98.7	26.2	6.97	1.85

---

The 12 h stability test of Al-MFI-180 and Fe-MFI-180 was also performed at 350 °C and W/F of 2 h. As illustrated in Fig. 8a, the propionic acid conversion on Al-MFI-180 drops remarkably from 98.7% to 29.4% in the initial 3 h. In the meantime, the yields of aromatics, C_1–2, C₃ and C_4–7 decrease to almost zero in the initial 3 h, while the yield of 3-pentanone increases first and then decreases to 26.9%, which further confirms that 3-pentanone is the primary product and can be converted further. After that, the conversion decreases slowly to 16.7% at the end of 12 h of reaction, with 3-pentanone being the dominant product. In contrast, the conversion of propionic acid on Fe-MFI-180 decreases slowly from 59.0% to 20.7% during a 12 h run, with 3-pentanone always being the dominant product (Fig. 8b). Hence, it can be concluded that the Fe-MFI-180 with weaker BAS is more stable and selective for the production of 3-pentanone than Al-MFI-180 with stronger BAS. In addition, it appears that the formation of secondary and tertiary products (aromatics, C_1–2, C₃ and C_4–7) results in faster deactivation of Al-MFI-180 than Fe-MFI-180.

Fig. 8 Conversion of propionic acid and yield of products on (a) Al-MFI-180 and (b) Fe-MFI-180 as a function of reaction time; conversion of 3-pentanone and yield of products on (c) Al-MFI-180 and (d) Fe-MFI-180 as a function of reaction time. Reaction conditions: T = 350 °C, P_{propionic acid} or P_3-pentanone = 4 kPa, W/F = 2 h.

To evaluate the reactivity of 3-pentanone toward secondary and tertiary products as well as its influence on the stability, 3-pentanone was fed as the reactant at 350 °C with a W/F of 2 h. As shown in Fig. 8c, the conversion of 3-pentanone on Al-MFI-180 approaches 100% in the initial 4 h, which implies that the catalyst has more sites than that is required to achieve 100% conversion. After that, the conversion drops quickly to 61.9% at the end of 12 h, suggesting fast deactivation of the catalyst. Aromatics are the major products (initial yields of 57.2%) over the C_1–2, C₃, and C_4–7 nonaromatic hydrocarbons. The decrease trend of their yields resembles that of conversion, which implies that the tertiary reaction for aromatic formation is strongly influenced by the deactivation. In contrast, the initial conversion on Fe-MFI-180 is rather low (22.4%), and decreases slowly to 13.7% after 12 h of reaction (Fig. 8d). Aromatics are also the major products, while their decreasing trend is much milder. These results indicate that the Fe-MFI-180 is less active for the conversion of 3-pentanone to aromatics, which in turn improves the stability. The TOF of 3-pentanone conversion was also measured under differential conditions and is reported in Table 4. The TOF of Al-MFI-180 (1.85 min⁻¹) is 5.6 times higher than that of Fe-MFI-180 (0.33 min⁻¹), confirming that Al-MFI-180 with stronger BAS is much more active for further conversion of 3-pentanone than Fe-MFI-180 with weaker BAS. Furthermore, the TOF of propionic acid (6.97 min⁻¹) is 3.8 times higher than that of the 3-pentanone (1.85 min⁻¹) on Al-MFI-180, whereas the TOF of propionic acid (3.80 min⁻¹) is 11.5 times higher than that of 3-pentanone (0.33 min⁻¹) on Fe-MFI-180. This significant difference further indicates that the Fe-MFI-180 with weaker BAS is much less active for further conversion of 3-pentanone, resulting in 3-pentanone as the dominant product with improved stability on Fe-MFI-180 in comparison to Al-MFI-180 with stronger BAS.

3.3. Effect of acid density of BAS

To explore the impact of BAS density and screen the optimal catalyst, a series of Fe-MFI zeolites with varying Si/Fe ratio were tested for ketonization of propionic acid at 350 °C with a W/F of 2 h. As the structural properties (specific surface area, pore volume, and crystallite size) of the Fe-MFI zeolites are similar, the catalytic performance should be mainly influenced by the acidic property, which is dependent on the content and distribution of Fe species in the zeolites. As illustrated in Fig. 9a, the conversion of propionic acid increases monotonously from 59.0% to 79.5% as the Si/Fe ratio is decreased from 180 to 100. Further reduction of Si/Fe ratio to 80 leads to a slight decline in the conversion of propionic acid to 77.9%, which likely stems from the decreased density of BAS at a low Si/Fe ratio of 80. This is confirmed by the fact that the conversion can be linearly correlated with the density of BAS in these zeolites (Fig. S12†), indicating that BAS are the active sites. However, the selectivity for 3-pentanone decreases monotonously from 86.5% to 68.5%, which likely results from the increased conversion of propionic acid as well as the varied acidic property. The intrinsic reaction rate was measured under differential conditions at 350 °C, by varying the W/F to ensure that the conversion is <15%. Under these conditions, 3-pentanone is the only product. The measured intrinsic reaction rates show a volcano dependence on the Si/Fe ratio (Fig. 9b). The highest rates of 0.435 and 0.438 mmol g⁻¹ min⁻¹ are achieved at Si/Fe ratios of 120 and 100, respectively, which have the highest density of BAS. The TOF of ketonization was then estimated assuming that Fe species are atomically distributed in the zeolites. As shown in Fig. 9b, the TOF decreases slowly and linearly from 3.80 to 3.16 min⁻¹, as the Si/Fe ratio decreases from 180 to 120. When the Si/Fe further decreases from 120 to 80, the TOF decreases quickly and linearly to 2.00 min⁻¹. This trend agrees with the fraction of framework Fe content or the fraction of BAS in total acid (Fig. S13†) as a function of Si/Fe ratio, confirming that BAS are the active sites for ketonization. The decrease in TOF with decreasing Si/Fe ratio reflects that more Fe forms extra-framework species. In addition, the highest TOF is achieved on Fe-MFI-180 with the longest distance between vicinal Fe centers, which also indicates that a single BAS instead of two neighboring BASs is required to mediate the ketonization reaction. Considering the Fe content, intrinsic reaction rate and TOF, the Fe-MFI-120 appears to be the optimal catalyst. Notably, the TOF of 3.16 min⁻¹ on optimal Fe-MFI-120 is much higher than those on Lewis acidic heteroatom zeolites, such as Zr-Beta (0.19 min⁻¹ for a Si/Zr ratio of 50), Ti-Beta (0.40 min⁻¹ for a Si/Ti ratio of 40), and TS-1 (0.68 min⁻¹ for a Si/Ti ratio of 80).^35,36 This comparison highlights that Fe-MFI with weak BAS is active and selective for the ketonization of carboxylic acids.

Fig. 9 (a) Conversion of propionic acid and selectivity of 3-pentanone over Fe-MFI zeolites as a function of Si/Fe ratio; (b) intrinsic reaction rate and TOF of Fe-MFI as a function of Si/Fe ratio; (c) Arrhenius plots of intrinsic reaction rates of propionic acid conversion on Al-MFI-180 and Fe-MFI zeolites. Reaction conditions: (a) T = 350 °C, W/F = 2 h, P_{propionic acid} = 4 kPa, TOS = 0.5 h; (b) T = 350 °C, P_{propionic acid} = 4 kPa, TOS = 0.5 h, conversion of propionic acid <15%; (c) T = 330–370 °C, P_{propionic acid} = 4 kPa, TOS = 0.5 h, conversion of propionic acid <15%.

The Arrhenius plots of intrinsic reaction rate dependence on reaction temperature (330–370 °C) of Al-MFI-180 and selected Fe-MFI samples are compared in Fig. 9c. The derived apparent activation energies (E_a) of ketonization on Fe-MFI-180, Fe-MFI-120 and Fe-MFI-80 are similar, i.e., 51–53 kJ mol⁻¹, which suggests a similar reaction mechanism and the same structure of active sites. These values are higher than the E_a on Al-MFI-180 (28 kJ mol⁻¹), which agrees with the lower TOF on Fe-MFI zeolites and indicates that BAS with stronger acidity facilitate ketonization reaction. Notably, these values are lower than the E_a (93 kJ mol⁻¹) on TS-1 (Si/Ti = 40), consistent with the higher TOF on Fe-MFI (3.80–2.00 min⁻¹) than on TS-1 (0.57 min⁻¹).³⁵ This comparison highlights that the BAS in Fe-MFI is more favorable than LAS on TS-1 with the same MFI pore architecture.

The samples of Fe-MFI-120 and Fe-MFI-80 were selected for the stability test at 350 °C and W/F of 2 h. As shown in Fig. 10a, the decline in propionic acid conversion is more rapid in the initial 6 h and then relatively slow in the next 6 h for Fe-MFI-120 and Fe-MFI-80. In this period of fast deactivation, the yields of aromatics decrease to zero (Fig. 10b), indicating that the initial fast deactivation is related with deactivation for formation of aromatics from the tertiary reactions. Alternatively, the yield of 3-pentanone was almost constant in the initial 2 h and then decreased on Fe-MFI-120 and Fe-MFI-80, which was likely owing to the reduced converting of 3-pentanone to aromatics in the first 2 h. Overall, the deactivation rate in the 12 h test follows the order of Fe-MFI-80 > Fe-MFI-120 > Fe-MFI-180. The slowest deactivation of Fe-MFI-180 is apparently related to its lowest density of BAS (69.2 μmol g⁻¹). As the densities of BAS on Fe-MFI-120 and Fe-MFI-80 (100.8 and 101.6 μmol g⁻¹, respectively) are similar, the faster deactivation of Fe-MFI-80 is likely related to more extra-framework FeO_x species and the slightly stronger acid strength resulted from extra-framework FeO_x species (see NH₃-TPD in Fig. 2b) which facilitate the formation of more aromatics. As shown in Fig. 10c, although the yields of both 3-pentanone and aromatics are similar on different samples at conversions <40%, the yields vary slightly at conversions >40%. The Fe-MFI-80 produces more aromatics but less 3-pentanone relative to Fe-MFI-120 and Fe-MFI-180 at identical conversions, resulting in faster deactivation. It should be noted that the yields of both aromatics and 3-pentanone are similar on Fe-MFI-120 and Fe-MFI-180 at identical conversion (Fig. 10c) and the yield of 3-pentanone in the 12 h test is the highest on Fe-MFI-120 (Fig. 10b), which indicates that Fe-MFI-120 is the optimal catalyst for ketonization.

Fig. 10 Effect of reaction time on the (a) conversion of propionic acid, (b) yield of 3-pentanone and aromatics, (c) yield of 3-pentanone and aromatics as a function of conversion of propionic acid on Fe-MFI-180, Fe-MFI-120 and Fe-MFI-80. Reaction conditions: T = 350 °C, W/F = 2 h, P_{propionic acid} = 4 kPa.

Finally, the catalytic activity of Fe-MFI-120 was compared with that of widely explored metal oxide catalysts (Table 5). The intrinsic ketonization reaction rate of Fe-MFI-120 in terms of mass is lower than those of MnO₂,⁶⁶ CeO₂-P,¹⁵ and TiO₂(001),²⁰ but is higher than those of ZrO₂-PR, Al₂O₃, and La₂O₃ under similar reaction conditions. Moreover, when the ketonization rate was normalized to mole of metal (Fe for Fe-MFI-120), the rate on Fe-MFI-120 is 39.5 and 34.8 times higher than that on CeO₂-P and MnO₂, respectively. Evidently, the Fe centers (BAS) are dispersed in the framework of zeolites, which make the active centers accessible to the reactant. In contrast, the active centers of metal oxides (acid–base pair) are only accessible to the reactant on the surface. This comparison highlights the advantages of zeolites. It should be noted that the stability of Fe-MFI is significantly improved relative to Al-MFI, and is comparable with that of some metal oxide catalysts.^67–69

Table 5 Catalytic activity of Fe-MFI-120 and metal oxides for the ketonization of carboxylic acid to ketone

Catalyst	Reaction conditions	Rate^a (mmol gcat^−1 min^−1)	Rate^b (mol molM^−1 min^−1)	Ref.
a Intrinsic reaction rates calculated by the mass of the catalyst used. b Intrinsic reaction rates were estimated in terms of mol of metal in the catalyst used.
Fe-MFI-120	T = 350 °C, W/F = 0.06 h	0.44	3.20	This work
m-ZrO2-PR	T = 350 °C, W/F = 0.8 s g mL^−1	0.18	0.022	70
TiO2(001)	T = 350 °C, W/F = 0.05 h	0.67	0.054	20
CeO2-P	T = 350 °C, W/F = 0.05 h	0.47	0.081	15
MnO2	T = 350 °C, W/F = 0.29 s g mL^−1	1.06	0.092	66
Al2O3	T = 400 °C, W/F = 0.29 s g mL^−1	0.019	0.00097	70
La2O3	T = 350 °C, W/F = 0.8 s g mL^−1	0.052	0.0085	70

---

4. Conclusion

Fe-MFI zeolites with varying Si/Fe molar ratio (80–180) were successfully synthesized via a hydrothermal method, fully characterized (by XRD, N₂ adsorption, SEM, HAADF-STEM, UV-vis, NH₃-TPD, FT-IR of pyridine adsorption, and FT-IR of propionic acid TPD), and tested for ketonization of propionic acid at 350 °C. The framework Fe species form BAS, i.e., bridging Fe–OH–Si, which shows weaker acid strength than Al-MFI. Compared with Al-MFI-180, although the Fe-MFI-180 with weaker strength of BAS moderately reduces the activity for ketonization of propionic acid (TOF of 6.97 and 3.80 min⁻¹, respectively), it significantly reduces the activity for secondary reaction (aldol condensation) and tertiary reaction (aromatic formation) (TOF of 3-pentanone conversion of 1.85 and 0.33 min⁻¹, respectively), resulting in the primary product of 3-pentanone as the dominant product and in turn improving the stability. Decreasing Si/Fe ratio leads to reduced fraction of framework tetrahedral Fe species, and the Si/Fe ratio of 120 was identified as optimal considering the Fe content, intrinsic reaction rate, TOF, selectivity and stability. The results of this work demonstrate a strategy of weakening the strength of Brønsted acid sites of zeolites to reduce the activity of the secondary and tertiary reactions and therefore to improve the selectivity and stability of ketonization of carboxylic acids.

Data availability

All data are available in the manuscript and its ESI.†

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (22278299), the Fundamental Research Funds for the Central Universities, and the Haihe Laboratory of Sustainable Chemical Transformations.

References

1. R. S. Varma, ACS Sustainable Chem. Eng., 2019, 7, 6458–6470.
2. Z. Wang, L. Yuan and C. Tang, Acc. Chem. Res., 2017, 50, 1762–1773.
3. B. Boekaerts and B. F. Sels, Appl. Catal., B, 2021, 283, 119607.
4. Y. Liu, S. Wu, H. Zhang and R. Xiao, Bioresour. Technol., 2021, 338, 125519.
5. Z. Yang, Q. Ge and X. Zhu, Green Chem., 2024, 26, 8068–8099.
6. N. A. Huq, G. R. Hafenstine, X. Huo, H. Nguyen, S. M. Tifft, D. R. Conklin, D. Stück, J. Stunkel, Z. Yang, J. S. Heyne, M. R. Wiatrowski, Y. Zhang, L. Tao, J. Zhu, C. S. McEnally, E. D. Christensen, C. Hays, K. M. Van Allsburg, K. A. Unocic, H. M. Meyer, Z. Abdullah and D. R. Vardon, Proc. Natl. Acad. Sci. U. S. A., 2021, 118, e2023008118.
7. E. Menalla, J. G. Serna, D. Cantero and M. J. Cocero, Chem. Eng. J., 2024, 493, 152391.
8. E. Ranzi, P. E. A. Debiagi and A. Frassoldati, ACS Sustainable Chem. Eng., 2017, 5, 2867–2881.
9. G. Wang, Y. Dai, H. Yang, Q. Xiong, K. Wang, J. Zhou, Y. Li and S. Wang, Energy Fuels, 2020, 34, 15557–15578.
10. T. Kan, V. Strezov and T. J. Evans, Renewable Sustainable Energy Rev., 2016, 57, 1126–1140.
11. S. Wang, G. Dai, H. Yang and Z. Luo, Prog. Energy Combust. Sci., 2017, 62, 33–86.
12. M. Wu, M. Liu, P. Guo, T. Lu, Z. Wu and W. Luo, ChemCatChem, 2024, 16, e202400896.
13. B. Boekaerts, M. Vandeputte, K. Navaré, J. Van Aelst, K. Van Acker, J. Cocquyt, C. Van Caneyt, P. Van Puyvelde and B. F. Sels, Green Chem., 2021, 23, 7137–7161.
14. J. Pittayatornkul, T. Maluangnont, S. Jongpatiwut, P. Praserthdam, M. Ogawa and T. Sooknoi, React. Chem. Eng., 2024, 9, 2345–2357.
15. Z. Liu, Y. Guo, H. Wang and X. Zhu, Appl. Catal., A, 2024, 672, 119593.
16. J. D. Maron, D. Cesari, S. Banu Rameesdeen, T. Tabanelli, A. Fasolini, F. Basile and F. Cavani, Green Chem., 2023, 25, 7381–7392.
17. K. Wu, M. Yang, W. Pu, Y. Wu, Y. Shi and H.-s. Hu, ACS Sustainable Chem. Eng., 2017, 5, 3509–3516.
18. S. Wang and E. Iglesia, J. Phys. Chem. C, 2017, 121, 18030–18046.
19. P. Tingelstad, R. L. Peters, S. K. Regli, A. Miró I Rovira, K. Rajendran and D. Chen, Catal. Today, 2025, 443, 114961.
20. J. Huang, L. Li, X. Wu, Y. Guo, Z. Yang, H. Wang, Q. Ge and X. Zhu, Catal. Sci. Technol., 2023, 13, 5924–5937.
21. R. A. L. Baylon, J. Sun, L. Kovarik, M. Engelhard, H. Li, A. D. Winkelman and Y. Wang, Appl. Catal., B, 2018, 234, 337–346.
22. S. Tosoni and G. Pacchioni, J. Catal., 2016, 344, 465–473.
23. S. Yang, B. Qian, Y. Wang, K. Taira, Q. Zhou, K. Wilson, A. F. Lee and L. Zhang, Appl. Catal., B, 2023, 322, 122106.
24. J. A. Bennett, C. M. A. Parlett, M. A. Isaacs, L. J. Durndell, L. Olivi, A. F. Lee and K. Wilson, ChemCatChem, 2017, 9, 1648–1654.
25. H. Nguyen, Y. Wang, D. Moglia, J. Fu, W. Zheng, M. Orazov and D. G. Vlachos, Catal. Sci. Technol., 2021, 11, 2762–2769.
26. E. Heracleous, D. Gu, F. Schüth, J. A. Bennett, M. A. Isaacs, A. F. Lee, K. Wilson and A. A. Lappas, Biomass Convers. Biorefin., 2017, 7, 319–329.
27. S. Wang and E. Iglesia, J. Catal., 2017, 345, 183–206.
28. Z. Yang, Q. Yu, Y. Guo, X. Wu, H. Wang, J. Han, Q. Ge and X. Zhu, Microporous Mesoporous Mater., 2022, 330, 111625.
29. A. Gumidyala, T. Sooknoi and S. Crossley, J. Catal., 2016, 340, 76–84.
30. A. Wang, D. Austin, A. Karmakar, G. M. Bernard, V. K. Michaelis, M. M. Yung, H. Zeng and H. Song, ACS Catal., 2017, 7, 3681–3692.
31. T. Yan, L. Yang, W. Dai, G. Wu, N. Guan, M. Hunger and L. Li, ACS Catal., 2019, 9, 9726–9738.
32. B. Yan, L. Z. Tao, Y. Liang and B. Q. Xu, ChemSusChem, 2014, 7, 1568–1578.
33. E. V. Fufachev, B. M. Weckhuysen and P. C. A. Bruijnincx, Green Chem., 2020, 22, 3229–3238.
34. T. Goculdas, S. Deshpande, W. Zheng, R. J. Gorte, S. Sadula and D. G. Vlachos, Green Chem., 2023, 25, 614–626.
35. Z. Yang, Q. Yu, H. Wang, Q. Ge and X. Zhu, J. Catal., 2024, 429, 115247.
36. Q. Yu, Y. Guo, X. Wu, Z. Yang, H. Wang, Q. Ge and X. Zhu, ACS Sustainable Chem. Eng., 2021, 9, 7982–7992.
37. J. Zhang, X. Tang, H. Yi, Q. Yu, Y. Zhang, J. Wei and Y. Yuan, Appl. Catal., A, 2022, 630, 118467.
38. M. Yabushita, H. Kobayashi, A. Neya, M. Nakaya, S. Maki, M. Matsubara, K. Kanie and A. Muramatsu, CrystEngComm, 2020, 22, 7556–7564.
39. K. Miyake, Y. Hirota, K. Ono, Y. Uchida, M. Miyamoto and N. Nishiyama, New J. Chem., 2017, 41, 2235–2240.
40. K.-Y. Lee, S.-W. Lee and S.-K. Ihm, Ind. Eng. Chem. Res., 2014, 53, 10072–10079.
41. M. Shahami and D. F. Shantz, Catal. Sci. Technol., 2019, 9, 2945–2951.
42. H. Li, Y. Zhai, X. Zhang, G. Lv, Y. Shen, X. Wang, T. Jiang and Y. Wu, Ind. Eng. Chem. Res., 2020, 59, 10289–10297.
43. Y. Zhai, F. Wang, X. Zhang, G. Lv, Y. Wu, T. Jiang, Q. Zhang, M. Li, M. Li and Y. Liu, Nano Res., 2021, 14, 4304–4313.
44. J. Cao, S. Ding, H. Wang, J. Han, Q. Ge and X. Zhu, Ind. Eng. Chem. Res., 2019, 58, 10307–10316.
45. K. S. W. Sing, D. H. Everett, R. A. W. Haul, L. Moscou, R. A. Pierotti, J. Rouquerol and T. Siemieniewska, Pure Appl. Chem., 1985, 57, 603–619.
46. P. Xiao, Y. Wang, J. N. Kondo and T. Yokoi, Appl. Catal., A, 2019, 579, 159–167.
47. K. Chen, X. Wu, J. Zhao, H. Zhao, A. Li, Q. Zhang, T. Xia, P. Liu, B. Meng, W. Song, X. Zhu, H. Liu, X. Gao, C. Xu and B. Shen, J. Catal., 2022, 413, 735–750.
48. E. Hensen, Q. Zhu, R. Janssen, P. Magusin, P. Kooyman and R. Vansanten, J. Catal., 2005, 233, 123–135.
49. Y. S. Kanga, D. K. Leea and P. Stroeve, Thin Solid Films, 1998, 327–329, 541–544.
50. Q. Wu, C. Xu, L. Zhu, X. Meng and F.-S. Xiao, Catal. Today, 2022, 390–391, 2–11.
51. E. Yuan, G. Wu, W. Dai, N. Guan and L. Li, Catal. Sci. Technol., 2017, 7, 3036–3044.
52. S. P. Yuan, J. G. Wang, Y. W. Li and H. Jiao, J. Phys. Chem. A, 2002, 106, 8167–8172.
53. X. Jiang, X. Su, X. Bai, Y. Li, L. Yang, K. Zhang, Y. Zhang, Y. Liu and W. Wu, Microporous Mesoporous Mater., 2018, 263, 243–250.
54. A. Primo and H. Garcia, Chem. Soc. Rev., 2014, 43, 7548–7561.
55. C. Liu, G. Li, E. J. M. Hensen and E. A. Pidko, J. Catal., 2016, 344, 570–577.
56. T. Li, T. Chen, Y. Ye, P. Dong, T. Wang, Q. Cui, C. Wang and Y. Yue, Chin. J. Chem. Eng., 2024, 66, 51–59.
57. A. Sakthivel, S. E. Dapurkar, N. M. Gupta, S. K. Kulshreshtha and P. Selvam, Microporous Mesoporous Mater., 2003, 65, 177–187.
58. R. Mokaya, W. Jones, Z. Luan, M. D. Alba and J. Klinowski, Catal. Lett., 1996, 37, 113.
59. W. Wu and E. Weitz, Appl. Surf. Sci., 2014, 316, 405–415.
60. J. Datka, A. M. Turek, J. M. Jehng and I. E. Wachs, J. Catal., 1992, 135, 186–199.
61. J. D. Kammert, A. Chemburkar, N. Miyake, M. Neurock and R. J. Davis, ACS Catal., 2021, 11, 1435–1455.
62. Z. Zhang, C. E. Berdugo-Díaz, D. T. Bregante, H. Zhang and D. W. Flaherty, ACS Catal., 2022, 12, 1481–1496.
63. L. M. Parker, Stud. Surf. Sci. Catal., 1988, 36, 589–595.
64. X. Wang, S. Ding, H. Wang, X. Liu, J. Han, Q. Ge and X. Zhu, Appl. Catal., A, 2017, 545, 79–89.
65. O. Kikhtyanin, R. Bulánek, K. Frolich, J. Čejka and D. Kubička, J. Mol. Catal. A: Chem., 2016, 424, 358–368.
66. O. Nagashima, S. Sato, R. Takahashi and T. Sodesawa, J. Mol. Catal. A: Chem., 2005, 227, 231–239.
67. S. Ding, H. Wang, J. Han, X. Zhu and Q. Ge, Ind. Eng. Chem. Res., 2018, 57, 17086–17096.
68. S. T. Almutairi, E. F. Kozhevnikova and I. V. Kozhevnikov, Appl. Catal., A, 2018, 565, 135–145.
69. Y. Yamada, M. Segawa, F. Sato, T. Kojima and S. Sato, J. Mol. Catal. A: Chem., 2011, 346, 79–86.
70. J. D. Maron, L. Bellotti, A. Baldelli, A. Fasolini, N. Schiaroli, C. Lucarelli, F. Cavani and T. Tabanelli, Sustainable Chem., 2022, 3, 58–75.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5cy00175g

---