Beyond Fossil Fuels: The Role of V-Doped Hydrotalcites in n-Butane Oxidative Dehydrogenation for a Circular Economy

Abstract

This study explores the catalytic performance of V³⁺-modified Mg/Al hydrotalcite-derived materials in the oxidative dehydrogenation (ODH) of n-butane, compared with catalysts derived from pyrovanadate and decavanadate precursors. Different methods for preparing hydrotalcite-like materials were applied to obtain vanadium-containing Mg-Al mixed oxide catalysts for n-butane ODH. The hydrotalcite-like precursors were doped with vanadates (V⁵⁺) via ion exchange or co-precipitation or with V³⁺ cations incorporated into brucite-like layers. During calcination in air or argon flow, different vanadium-containing phases were obtained. Our findings demonstrate that V³⁺-doped hydrotalcites exhibit superior activity and selectivity toward the total C₄H₈ products, with enhanced selectivity for 1,3-butadiene. The highest n-butane conversion was observed for catalysts with an MgO structure and vanadium dispersed in the oxide matrix. A similar conversion level (~44%) was obtained for a spinel-like Mg₂VO₄ catalyst, but only a 15% level was found for the highly crystalline α-Mg₂V₂O₇ catalyst. In contrast, the highest selectivities toward dehydrogenated products were observed for V³⁺-containing and α-Mg₂V₂O₇ catalysts. NH₃- and CO₂-temperature programmed desorption (TPD) analyses showed that high basicity combined with low acidity favors the formation of butene isomers and 1,3-butadiene. This work highlights the strategic potential of tailoring vanadium speciation and hydrotalcite-based catalyst design for low-carbon chemical manufacturing, supporting the transition toward a circular economy.

1. Introduction

1.1. Vanadium Catalysts

Vanadium catalysts have played a significant role in the development of industrial technologies for processing chemical compounds for decades. They are used in sulfuric acid production [1,2], biomass transformation [3], biodiesel production [4], and the oxidation, dehydrogenation, or oxidative dehydrogenation of hydrocarbons [5,6,7,8], as well as in the removal of NOx from effluent gases [9,10], among other applications. However, extensive studies and environmental monitoring have determined that vanadium catalysts can no longer be used in certain applications. Due to their high toxicity and susceptibility to sublimation [11,12,13,14,15,16], vanadium-containing catalysts, especially those with higher oxidation states, have been banned from use in mobile pollution sources.

Vanadium-containing catalysts, particularly Mg-V-O mixed oxides, have been studied for many years as efficient catalysts for the conversion of alkanes to the corresponding alkenes [17]. These catalysts are usually obtained by impregnation of Mg-containing supports with a solution of vanadium salts or by a solid-state reaction between magnesia and vanadia [18,19,20,21,22]. Another possible route is the calcination of V-containing hydrotalcite-like precursors [23,24,25,26,27,28].

1.2. Beyond Fossil Fuels: Transition into Circular Economy

Despite the limitations arising from harmful emissions, the use of vanadium catalysts in closed installations still offers significant advantages. Such catalysts can support the transition to a circular economy not only through their role in pollutant removal [29,30] or the potential for catalyst recycling [31,32,33] but also through their application in processes where fossil fuels can be replaced by renewable raw materials [34,35,36,37]. Consequently, increasing research attention is being devoted to the production of alternative fuels (e.g., biodiesel) and the processing of industrial, agricultural, forestry, food, or plastic waste [38,39,40,41,42].

A circular economy aims to close material loops by transforming waste streams into value-added products [43,44,45]. Producing butane, a key chemical feedstock and fuel, from renewable resources offers both environmental and economic benefits, as it reduces reliance on fossil fuels while valorizing organic waste. Traditionally, n-butane is obtained from petroleum via distillation, but its green analog, biobutane (renewable butane), can be generated through the conversion of biomass or organic residues. Although n-butane production is not as extensively developed as n-butanol fermentation, several studies demonstrated the feasibility of this process. Therefore, butane, selected here as a model molecule, can be considered part of a growing group of renewable materials. The bioeconomy, including biorefineries, is increasingly promoted as a fundamental element of the circular economy, with waste from various origins, particularly food processing residues, serving as the preferred raw material [46].

Processes for producing non-fossil fuel-derived gaseous alkanes or liquefied petroleum gas in form of propane, butane and isobutane blends (so called bio-LPG) or alternative fuels are highly valued and actively investigated. These typically require engineered microbial strains capable of operating efficiently under relevant conditions. For example, butyric acid from waste biomass has been used as a carbon source for microbial fuel production [47]. Halomonas, a robust extremophilic microbial chassis, can operate in non-sterile environments to convert organic compounds into a gaseous alkane mixture. Ideally, microbial fermentation-based bio-LPG production systems could be tuned to control the propane-to-butane ratio according to market needs. Alternative biosynthetic pathways have also been proposed, such as producing propane, isobutane, and n-butane from the amino acids valine, leucine, and isoleucine, respectively, using engineered Escherichia coli (E. coli).

The strong demand for LPG and the pressure to decarbonize energy systems have incentivized the development of both biological and chemical production routes. In addition to microbial conversion, chemical processes play a significant role in maximizing production and waste valorization [48]. Bio-LPG can be produced as either a main product or a by-product in processes such as hydrotreating or dehydrogenating bio-oils, glycerol hydrolysis and fermentation, and gas-phase conversion and synthesis from cellulose or organic waste, as well as in more energy-intensive methods like gasification and pyrolysis.

Compared with bioethanol and biodiesel, microbial alkane and alkene production still suffers from low yields, limiting industrial scalability. Nevertheless, microorganisms such as Saccharomyces cerevisiae and E. coli have been tested for alkane production while using glucose, fatty acids, and aldehydes as carbon sources [49]. Additionally, certain newly isolated strains from natural habitats show exceptional performance. For instance, Aureobasidium melanogenum was reported to be 30% more efficient in alkane and alkene production when grown on inulin.

Selective production of n-butane from renewable carbon sources has also been achieved in model-assisted, two-stage fermentation processes [50]. Here, engineered enzymes converted acetate, produced via electrocatalytic CO reduction, into n-butane, with a modified E. coli strain increasing n-butane titers 168-fold (from 0.04 to 6.74 mg/L). In another example, levulinic acid (LA), a bio-based platform chemical, was transformed over a base-treated Pt/C catalyst in a one-step reaction, yielding 95.5% n-butane via hydrodeoxygenation of LA to valeric acid (VA), followed by VA decarboxylation [51].

The increasing number of studies on biofuel production highlights the growing interest in efficient processes for n-butane oxidative dehydrogenation (ODH). Literature trends (Figure 1) indicate that dehydrogenation of propane and butane was most intensively studied in the late 20th and early 21st centuries. Over the past two decades, however, research interest has shifted toward bio-based production of propane and butane as renewable feedstocks for further transformations, alongside continuing investigation into vanadium-based catalysts.

Thus, when butane is produced via biotechnological routes from small organic molecules or through biomass transformation, it is possible to address several principles of the circular economy according to the 7R scheme: Rethink–Reduce–Reuse–Repair–Refurbish–Recycle–Recover. The replacement of fossil-based feedstocks reduces raw material consumption, prevents (or decreases) waste from food processing, and enables the upcycling of renewable resources (reuse and recycle). Moreover, the adaptation of hydrotalcite precursors for the synthesis of vanadium-doped catalysts within a redefined, renewable-based technology (rethink raw material demand and processing parameters) may play a crucial role in the development of an improved system for platform chemical supply. The last element, recovery of valuable compounds from spent catalysts, may be considered an interesting topic for further studies.

1.3. The Role of Hydrotalcite Precursors

Layered double hydroxides, known as hydrotalcites, consist of brucite-like layers in which part of the divalent cations (M²⁺) is substituted by trivalent ones (M³⁺). The resulting excess positive charge is balanced by hydrated anions (Aⁿ⁻) located between the layers. Thus, the general formula of hydrotalcites (HTs) can be expressed as follows:

[M^II_1−xM^III_x(OH)₂]A_x/nⁿ⁻ · zH₂O

The chemical composition of these materials can be varied over a wide range by substituting metal cations (e.g., Cu²⁺, Ni²⁺, Zn²⁺, Fe³⁺, or Cr³⁺) or interlayer anions (e.g., Cr₂O₇²⁻, Mo₇O₂₄⁶⁻, or V₂O₇⁴⁻) [52,53,54,55]. In the present work, doping within the brucite-like sheets with V³⁺, as well as interlayer modification by introducing vanadates (V⁵⁺), was applied to obtain V-containing Mg-Al hydrotalcite-like precursors. These were subsequently converted into V-containing catalysts with varying vanadium contents and oxidation states. Maintaining a lower oxidation state was intended to achieve higher catalytic activity while reducing harmful emissions. Furthermore, the use of a hydrotalcite precursor was aimed at increasing vanadium dispersion, thereby preventing excessive sintering, phase segregation, and ultimately the formation of vanadium oxide clusters. Hydrotalcite precursors are considered cheap, easy to prepare in a laboratory, and extremely flexible in terms of composition modification and enabling the properties of the catalysts obtained from them to be adjusted [52]. Likewise, the variability of the vanadium oxidation state in hydrotalcite precursors may be preserved in catalysts by applying inert synthesis and calcination conditions, thereby regulating activity in accordance with the principle “the higher the oxidation state of the catalyst, the higher the activity, but the lower the selectivity” [56]. While the catalytic activity of hydrotalcite-derived vanadium catalysts increases with the metal content, excessive loading can lead to the segregation of α-Mg₂V₂O₇, which is associated with non-selective centers. One of the main challenges in catalyst design is to obtain phases that limit oxygen species mobility while ensuring a suitable density for the active sites. Therefore, an optimal balance between metal loading, phase composition, and reducibility is required to avoid overly rapid oxidation that results in total combustion products. Since O²⁻ ions themselves do not possess oxidizing properties, oxidation-reduction cycles must occur to prevent excessive combustion. For this reason, Mg₃(VO₄)₂ is considered a better choice for the active phase than α-Mg₂V₂O₇, as the former corresponds to isolated VO₄ tetrahedra, which can limit the number of oxygen ions available for each adsorbed hydrocarbon molecule. In contrast, the latter consists of V₂O₇ units, which can be viewed as pairs of corner-sharing VO₄ tetrahedra [57].

Another advantageous feature of hydrotalcites is their high basicity. In the first stage of the reaction, basic sites facilitate hydrogen abstraction from alkanes with surface oxygen atoms and subsequently promote butane desorption [18]. Nonetheless, both types of sites, acidic and basic, contribute to the reaction, as oxygen-containing species adsorb on acid sites and are transformed into carboxylates and carbonates, ultimately leading to CO₂ formation [26].

The following properties of hydrotalcites, highlighted in previous studies [21], can be advantageously utilized in the ODH of n-butane, five of which are discussed in this paper:

• Low-cost and relatively simple synthesis;
• The possibility of introducing vanadium in various forms and oxidation states;
• Control over phase composition and active phase dispersion;
• Tunable acidity and basicity (calcined hydrotalcites act as solid bases and basic catalysts);
• The large surface area of hydrotalcite-derived catalysts;
• The potential for recovering components from spent catalysts.

2. Results and Discussion

2.1. Characterization of the Hydrotalcite-like Materials

The following hydrotalcite precursors were used in current work for material characterization and evaluation of catalytic activity (

Table 1. Synthetic route of vanadium incorporation, sample codes, and chemical composition descriptions of hydrotalcite materials.

Synthesis or Route of Vanadium Incorporation	Sample Code *	Chemical Composition of Hydrotalcite (Formula)
starting material for ion exchange	CV0	Mg0.654 Al0.346 (OH)2 (NO3−)0.002 (CO32−)0.172·0.645 H2O
starting material for ion exchange	NV0	Mg0.657 Al0.343 (OH)2 (NO3−)0.302 (CO32−)0.020·0.889 H2O
ion exchange, pH = 9.5	CV3	Mg0.654 Al0.346 (OH)2 (CO32−)0.133 (V2O74−)0.002·0.392 H2O EIE ** = 14.1%
ion exchange, pH = 9.5	CNV7	Mg0.600 Al0.400 (OH)2 (CO32−)0.100 (V2O74−)0.050·0.474 H2O EIE ** = 51.4%
ion exchange, pH = 9.5	NV12	Mg0.704 Al0.296 (OH)2 (CO32−)0.019 (V2O74−)0.064·0.354 H2O EIE ** = 86.9%
ion exchange, pH = 4.5	CV25	Mg0.627 Al0.376 (OH)2 (CO32−)0.009 (V10O286−)0.040·0.450 H2O EIE ** = 52.2%
direct precipitation (D)	DV13	Mg0.696 Al0.304 (OH)2 (CO32−)0.031 (V2O74−)0.075·0.337 H2O EIE ** = 98.5%
precipitation followed by hydrothermal aging in autoclave (A)	AV8	Mg0.690 Al0.239 V0.071 (OH)2 (CO32−)0.295 (NO3−)0.002·0.378 H2O
precipitation followed by hydrothermal aging in autoclave (A)	AV18	Mg0.831 V0.169 (OH)2 (CO32−)0.269 (NO3−)0.004·0.624 H2O

* C = anion in starting material for ion exchange (carbonate); N = anion in starting material for ion exchange (nitrate); V = vanadium content in resulting catalyst. ** EIE = efficiency of ion exchange, calculated with respect to actual layer charge. For calcined samples shown in next chapter, small letter “c” was added to sample names.

).

The X-ray powder diffraction patterns shown in Figure 2 are characteristic of layered double hydroxides (LDHs) with rhombohedral symmetry 3R1 [56]. However, due to different preparation procedures, discrepancies were observed in the crystallinity of the obtained samples, the positions of the basal reflections (00l), and the occurrence of side phases.

Two samples with pyro- and metavanadates intercalated via ion exchange (NV12) or co-precipitation (DV13) exhibited diffuse, poorly developed structures. No shift in the basal reflections, indicative of anion exchange, was observed for the carbonate-containing hydrotalcite (CV3). In contrast, the decavanadate-pillared hydrotalcite (CV25) showed a significant shift for the basal reflection, along with the presence of a side phase, as a result of Mg²⁺ leaching under the acidic conditions used for ion exchange. As shown previously [57], in the case of some cations, such as V³⁺, an additional procedure is necessary to obtain a well-developed structure. Hydrothermal treatment at an elevated temperature and pressure resulted in well-resolved X-ray diffraction (XRD) patterns with sharp peaks for samples AV8 and AV18.

The cell parameters c and a were calculated according to the following equations:

c = (3·d003 + 6·d006)/2 or c = (6·d006 + 9·d009)/2 (for sample CV25)

a = 2·d110

The crystallite sizes k_a and k_c were estimated from line broadening using the Debye–Scherrer equation:

D = k_a = k_c = K λ/(β·cos θ),

where D is the crystalline size, K is the Scherrer constant or the shape factor (equal to 0.89 here), λ is wavelength of CuKα radiation (equal to 1.54184 Å), β is the full width at half maximum (FWHM) of the peaks and θ is the peak position. The results are presented in

Table 2. Cell parameters for V-doped and starting carbonate- or nitrate-containing hydrotalcites.

Sample	Intercalated Anions	Cell Parameter [nm]		Crystallite Size [nm] *
		c	a	kc	ka
CV0 NV0	CO32− NO3−	2.293 2.689	0.3045 0.3045	16 17	31 21
CV25	V10O286−	3.508	0.3035	22	22
CV3 NV12	CO32− ** CO32−/V2O74−	2.283 2.554–2.854	0.3041 0.3046	15 (~6)	31 23
DV13	CO32−/V2O74−	2.358–2.738	0.3034	-	12
AV8 AV18	CO32− CO32−	2.300 2.349	0.3066 0.3111	36 43	49 59

* Calculated using Debye–Sherrer equation. ** Vanadates adsorbed on the external surface.

.

As mentioned above, the sample obtained under acidic conditions (CV25) was successfully intercalated with decavanadates. The c parameter value was only slightly lower than the values reported for V₁₀O₂₈⁶⁻ with its C₂ axis oriented parallel to the brucite-like sheets [53,58,59]. A decrease in the a parameter compared with the parent material (CV0) was also observed, confirming Mg²⁺ extraction from the hydroxide layers. No such effect was observed for the samples contacted with basic vanadate solutions (CV3 and NV12). Based on the c parameter values, it can be concluded that carbonates can be exchanged by higher charge-density anions such as V₁₀O₂₈⁶⁻ but not by pyro- or metavanadates. The small amount of vanadium in sample CV3 is most likely due to adsorption of vanadates on the external crystallite surfaces.

The diffuse patterns of NV12 and DV13 can be attributed to turbostratic disorder, undetectable impurities, or the presence of anion mixtures [59,60]. Deconvolution of the basal reflection (Figure 3) yielded c parameter values close to those for both pyrovanadate- and carbonate-intercalated hydrotalcites [24,61,62]. Therefore, the crystallite size could not be reliably determined from the broadening of the basal reflections. A similar phenomenon was reported for hydrotalcites intercalated with a CO₃²⁻/NO₃⁻ mixture [63].

The results do not provide a clear answer as to which vanadate species was incorporated into the interlayers, as similar c values have been reported for both metavanadate- and pyrovanadate-intercalated hydrotalcites [59,60,64]. For the V³⁺-doped hydrotalcites, the interlayer distances corresponded to those of carbonate-intercalated hydrotalcites (AV8) or were slightly higher (AV18). The latter may be explained by the lower positive charge of the brucite-like layer in AV18 [52]. Complete substitution of Al³⁺ by V³⁺ led to an increase in the a parameter. Hydrothermal treatment also increased the crystallite size; the k_a and k_c values for the AV8 and AV18 samples were 3–4 times higher than those of the other materials. At the expense of smaller crystallites that are dissolved, larger crystals may be formed via Ostwald ripening [65].

The Fourier Transform Infrared (FTIR) spectra (Figure 4) confirmed the presence of V₁₀O₂₈⁶⁻ in the interlayers of CV25. Several infrared bands characteristic of decavanadates were identified: 964 cm⁻¹ (symmetric stretching vibrations of VO₂ or terminal V=O), 813 and 517 cm⁻¹ (antisymmetric and symmetric stretching vibrations in V-O-V chains), 739 cm⁻¹ (V-O stretching), and 673 cm⁻¹ (assigned to H_xV₁₀O₂₈^6−x, x = 1–3) [66,67,68].

All spectra also showed bands associated with OH⁻ stretching vibrations (~3450 cm⁻¹) and bending modes of interlayer water molecules (1626–1650 cm⁻¹). Interlayer carbonates were present in all samples prepared under basic conditions, which were either co-precipitated with carbonates (AV8/AV18) or contaminated with CO₃²⁻ (CV3, NV12, DV13), as indicated by modes at 3060 cm⁻¹ (bridging CO₃²⁻–H₂O), 1360 and 1405 cm⁻¹ (ν₃ stretching CO₃²⁻), and 850 cm⁻¹ (ν₂ stretching CO₃²⁻) [52,69]. One band not observed in other samples appeared in AV18 at 711 cm⁻¹, which was assigned to V-O vibrations in brucite-like layers [70].

It was not possible to distinguish between pyro- and metavanadates in NV12 and DV13 with FTIR. According to various authors, bands around 961–940 cm⁻¹ may arise from vibrations in metavanadate chains or cyclic V₄O₁₂⁴⁻, bands at 925–927 cm⁻¹ correspond to symmetric stretching modes (VO₂) in HV₂O₇³⁻ or (VO₃)_nⁿ⁻, the 785 cm⁻¹ band is assigned to symmetric stretching modes (VO₂) in (VO₃)_nⁿ⁻, and bands at 616–627 cm⁻¹ correspond to V-O-V stretching modes in (VO₃)_nⁿ⁻ [71,72,73].

2.2. Characterization of the Catalysts

The diversity of the preparation procedures yielded a series of hydrotalcite-like materials that, upon heating, formed mixed metal oxides with different chemical compositions, oxidation states for the transition metal, and, consequently, distinct physicochemical properties. A summary of the bulk structural and chemical analyses is given in

Table 3. Phase composition and metal content for V-doped and starting carbonate- and nitrate-containing hydrotalcites.

Sample	Phase Composition	x * [MIII/(MIII + MII)]	Mg [wt%]	Al [wt%]	V * [wt%]
CV0c NV0c	MgO MgO	0.346 (0.33) 0.343 (0.33)	36.1 36.7	21.2 20.7	- -
CV25c	MgO + α-Mg2V2O7	0.376 (0.33)	18.9	12.5	25.5 (30)
CV3c NV12c	MgO MgO + vanadate	0.346 (0.33) 0.296 (0.33)	34.4 31.0	20.2 14.5	2.7 (-) 11.9 (13)
DV13c	MgO + vanadate	0.304 (0.33)	29.6	14.4	13.4 (13)
AV8c AV18c	MgO + spinel Mg2VO4	0.310 (0.33) 0.169 (0.25)	36.1 41.3	13.9 -	7.8 (12) 17.6 (26)

* Brackets show expected values, dash was used to indicate which metal should not be detected according to synthesis mechanism.

.

The highest V loading (25 wt%) was obtained for the CV25c sample, whose LDH precursor was pillared with V₁₀O₂₈⁶⁻. However, this value was lower than expected. This may be due to the high trivalent cation ratio and the large size of the decavanadate species, which caused strong anion-anion repulsion, resulting in less than the required amount of anions being introduced into the interlayers. A different explanation must be considered for the discrepancies V³⁺ in V content observed in both V³⁺-doped samples (AV8c and AV18c): the susceptibility of V³⁺ to oxidation and its large ionic radius hinder quantitative precipitation and incorporation into the brucite-like layer [57,70].

In the other samples, the expected and obtained V loadings were similar, although it cannot be excluded that some vanadates used during synthesis were adsorbed on the surface (CV3c and NV12c) or formed amorphous side phases (NV12c and DV13c).

As shown in Figure 5 and the detailed analysis of phase identification in Figures S1–S5 (Supplementary Materials), the V loading determines the structure of the catalysts. In the CV25c sample, a mixture of highly crystalline α-Mg₂V₂O₇ and MgO was detected, consistent with previous results for similar hydrotalcite-derived catalysts [22,24,25,74] and with phase diagrams of the MgO-V₂O₅ oxide system [75,76,77]. Another V-rich catalyst, AV18c, was obtained from V³⁺-doped hydrotalcite calcined under an inert atmosphere. In this case, a crystalline phase was also detected, but surprisingly, such a structure has not previously been reported from a hydrotalcite-like precursor. A few sharp peaks indicated the presence of spinel-like Mg₂VO₄, with possible variations in its stoichiometry and oxidation state (compare Figure S5, Mg₂V_1.333O₄) instead of orthovanadate Mg₃V₂O₈ [23,25,26,57,70,78,79]. Weak reflections assigned to Mg₂VO₄ were also visible in the XRD pattern of another reduced-V sample, AV8c, although the main component was nearly amorphous MgO (periclase), with broad reflections at 2θ ≈ 36°, 43°, and 63°. Similar features were observed for other low-V catalysts (DV13c and NV12c), but in these cases, an additional vanadate phase and a disturbed pattern in the range of 2θ ≈ 20°, 30°, 32° could be due to the presence of a dispersed orthovanadate phase. The CV3c sample consisted mostly of MgO. However, as was shown recently, upon calcination, gradual segregation of periclase in calcined Mg-containing hydrotalcites is expected [80].

Additional insights into the catalyst structures were obtained from infrared spectra (Figure 6). The most intense bands in CV25c were assigned to α-Mg₂V₂O₇ (970, 919, 848, and 822 cm⁻¹), although some could also arise from Mg₃V₂O₈ vibrations (919 and 822 cm⁻¹) [22,28,57,81]. Similarly, the IR spectra of DV13c and NV12c showed modes at 925 and 918 cm⁻¹ (ν_s VO₄), 835 and 823 cm⁻¹, and 469 and 492 cm⁻¹ (ν_s V-O-V) derived from Mg₃V₂O₈ vibrations, although magnesium pyrovanadate has characteristic modes in similar positions. Analysis of the 1400–1660 cm⁻¹ region provided information on the CO₂ adsorption capacity. In addition to bands of readsorbed water, modes of bidentate CO₃²⁻ (1622–1630 cm⁻¹, ν_as) and unidentate CO₃²⁻ (1521–1523 cm⁻¹, ν_as, 1418–1428 cm⁻¹, ν_s) were observed, surprisingly even for low-surface-area AV8c and AV18c samples [82]. Only CV25c showed no visible CO₂ adsorption.

The catalysts also differed in textural parameters such as the surface area, total pore volume, pore shape, and pore size distribution (

Table 4. Textural parameters and acido-basic properties of the studied catalysts and hydrotalcite-derived reference materials (Mg/Al mixed oxides).

Sample	SBET [m2/g]	VTOTAL [cm3/g]	Pore Size [nm]	Amount of Adsorbed Molecules [μmol/g]
				CO2	NH3
CV0c NV0c	216 164	0.757 0.270	1.6, 6.4 3.2, 5.8	678.8 n.d.	97.5 n.d.
CV25c	14	0.044	-	117.4	20.5
CV3c NV12c	196 79	0.593 0.272	3.2, 8.4 1.4, 5.6	n.d. n.d.	n.d. n.d.
DV13c	104	0.305	1.8, 6.8, 12	*	142.7
AV8c AV18c	28 34	0.099 0.109	- -	* *	117.5 41.4

* Formation of bulk carbonates or strong multilayer adsorption; n.d.—not determined.

). The reduced-V samples (AV8c/AV18c) and the highly loaded CV25c were nearly non-porous solids with surface areas of ~30 m²/g and ~15 m²/g, respectively. The quasi-amorphous structure of the other catalysts resulted in greater accessibility to gas molecules. Surface adsorption of vanadates followed by calcination of CV3 caused only a small decrease in surface area (from 216 to 196 m²/g) and total pore volume (from 0.757 to 0.593 cm³/g) compared with the undoped calcined parent material (CV0c). However, further increasing the V loading promoted sintering, reducing the surface area to 104 m²/g (DV13c) and 79 m²/g (NV12c). The hysteresis loop shapes (Figure 7, left panel) were characteristic of mesoporous materials with bottle-shaped pores [83,84]. The pore size distribution was broad, with the most frequent values being between 5 and 10 nm (Figure 7, right panel). However, it would be worthwhile to consider depositing the precursor on a support with a high specific surface area, such as vermiculite [85,86,87,88,89,90,91], which is resistant to high processing temperatures and has good adsorption capacity. This would improve the dispersion of the active phase and reduce the amount of metal required for the reaction.

Temperature-programmed desorption of probe molecules was used to determine the surface acid-base properties. The amount of NH₃ desorbed varied between 20 and 143 μmol/g, but the acid site density was similar for most samples (~1.4 μmol/m²). Only AV8c, partially doped with reduced vanadium, showed a threefold higher acid site density. All catalysts exhibited a similar NH₃ desorption profile (Figure 8, left panel); desorption began around 100–120 °C with a maximum rate below 200 °C. The broad, asymmetric peaks likely indicate uniformly distributed sites of similar strength, along with diffusion effects and molecule readsorption.

In contrast, the CO₂ desorption profiles (Figure 8, right panel) displayed several peaks, being particularly well resolved for the reduced-V samples (AV8c AV18c). The first desorption stage occurred at 120–145 °C, although for the pyrovanadate sample (CV25c), another maximum appeared at 180 °C, overlapping the first peak. For the spinel-like Mg₂VO₄ catalyst (AV18c), an additional step was observed at 235 °C. This suggests that complete substitution of Al by V and formation of a distinct structure rearranged the basic site strength distribution, favoring strong basic sites. The final significant CO₂ desorption occurred between 320 and 405 °C, with CO₂ remaining on Mg₂VO₄ surfaces up to 465 °C. However, prolonged CO₂ adsorption led to bulk carbonate formation and multilayer adsorption, meaning the results should be considered qualitative.

The reduction profiles of all vanadium catalysts (Figure 9) consist of at least two poorly resolved peaks above 600 °C, likely corresponding to transitions through successive oxidation states [21]. For the CV25c sample, additional peaks occurring between 400 and 600 °C can be attributed to the reduction of small clusters or surface-dispersed VO₄ tetrahedra. Consequently, determining the exact onset of reduction is subject to significant uncertainty. Nevertheless, it can be concluded that the catalyst with a fine crystalline MgO structure containing a dispersed vanadium phase (CNV7c, reference sample with similar V content to AV8c) undergoes bulk reduction at lower temperatures compared with the sample containing crystalline α-Mg₂V₂O₇ (CV25c). Changes in the oxidation state, calculated with the hydrogen consumption (

Table 5. Hydrogen consumption, sample reduction degree, and peak and onset temperatures of reduction for vanadium catalysts.

Sample	Consumption H2 [mmol/g]	Reduction Degree		Temperature [°C]
		[H2]/[V] a	Change of Oxidation State	Maximum	Onset
AV8c AV8c R	0.675 1.467	0.443 0.963	0.9 1.9	(650) (654, 684)	(550) (500)
CNV7c CV22c	1.549 4.222	1.074 0.981	2.1 2.0	(610) (647, 692)	(450) (510)

^a H₂/V molar ratio in catalysts.

), indicate a transition from V⁵⁺ to V³⁺ for both CNV7c and CV25c, as well as for AV8c R, which was examined after the catalytic test in a reaction mixture with a high oxygen content. The fresh AV8c catalyst was reduced to a lesser extent, corresponding approximately to a transition by one oxidation state. It can thus be concluded that the sample obtained after calcination (AV8c) indeed contains V⁴⁺ vanadium in the form of Mg₂VO₄ spinel, while it becomes almost completely oxidized during the catalytic test [79].

X-ray photoelectron (XPS) spectra (Figure 10, Figures S6 and S7 in Supplementary Materials) representing vanadium (V 2p_3/2) species in the catalysts confirmed that the dominant forms in the pyrovanadate- and decavanadate-exchanged materials were V⁵⁺, with peaks identified at 517.0 eV. However, surface species in the reduced vanadium-containing materials were also oxidized, as indicated by peaks recorded at 516.6–517.0 eV. The presence of lower oxidation states in the AV8c and AV18c materials was confirmed by fitting peaks at lower binding energies (BE) (514.8 eV and 515.7 eV) that are characteristic for V³⁺ and V⁴⁺, respectively [92,93,94,95,96]. This observation is consistent with XRD analysis of the catalysts’ structure. The surface chemical composition was derived from the atomic ratios presented in

Table 6. Surface vs. bulk chemical composition of V-doped hydrotalcite-derived catalysts.

	Surface Atomic %
	Mg	Al	V	C	O
CV0c	19.0	13.7	0.0	4.0	63.3
CV3c	16.1	18.8	1.1	2.1	61.9
CV25c	10.3	19.2	7.3	1.7	61.4
CNV7c	14.6	19.2	2.4	1.8	62.1
DV13c	16.4	16.7	3.6	2.6	60.7
AV8c	32.8	20.9	1.4	5.1	39.7
AV8c R	14.8	8.3	0.5	30.9	45.5
AV18c	26.4	0.0	6.3	6.4	61.0
	Surface Molar Ratio			Bulk Molar Ratio
	Mg	Al	V	Mg	Al	V
CV0c	0.581	0.419	0.000	0.654	0.346	0.000
CV3c	0.447	0.523	0.030	0.638	0.338	0.024
CV25c	0.280	0.521	0.199	0.447	0.266	0.287
CNV7c	0.404	0.531	0.066	0.559	0.372	0.069
DV13c	0.448	0.455	0.097	0.605	0.265	0.131
AV8c	0.595	0.379	0.026	0.690	0.239	0.071
AV8c R	0.628	0.352	0.019	n.d.	n.d.	n.d.
AV18c	0.808	0.000	0.192	0.831	0.000	0.169

n.d.—not determined.

. As shown, the catalyst surfaces were enriched mostly in aluminum compared with the bulk composition, which may limit accessibility to the active centers.

2.3. Catalytic Study

Hydrotalcite-derived vanadium-containing catalysts were studied in the ODH of n-butane, using O₂ as an oxidant (Figure 11). The temperature (550 °C) selected for experimentation (Figure S8, Supplementary Materials) allowed reaching moderate-to-high conversion levels while selectivity of the combustion products (CO_x) was limited. Moreover, the oxygen content in the reaction mixture (10 vol.%) was optimum (Figure S9, Supplementary Materials) to obtain the highest yield of the most desired products, such as butenes and butadiene. Although numerous variables must be considered a source of superior performance, some of them stand out, namely phases with reduced vanadium in the AV8c and AV18c catalysts responsible for higher-yield ODH products [26,97] and high acidity for the overall activity, leading to desired products as well as desired ones [21]. The highest n-butane conversion (up to 51%) was obtained for the samples derived from pyro- and metavanadate-pillared LDH precursors (NV12c and DV13c) as well as catalysts containing reduced vanadium (AV8c and AV18c), for which the conversion reached 40–44%. Furthermore, these two latter samples (AV8c and AV18c) exhibited the highest specific activity per surface area (SX) in the range of 6.3–8.3 mmol n-C₄H₁₀/h·m² (Figure 12). Although it was demonstrated that both the AV8c and AV18c samples were characterized with lower reducibility (higher onset temperature), and AV18c presents significant acidity [21], both catalysts showed superior performance in terms of the ODH products’ yields and, to lesser extent, n-butane conversion. It should be considered then that phases formed in the presence of V³⁺ play a crucial role in the dehydrogenation process as well as tuned reduction-oxidation cycles, preventing excessive oxidation [18].

Interestingly, another highly loaded sample, CV25c, also showed high specific activity (5.8 mmol n-C₄H₁₀/h·m²) but only low conversion (15%). Nevertheless, high total selectivity to dehydrogenated species (CV0c, NV0c) showed an extremely high conversion of n-C₄H₁₀ (around 44%), although their main products were CO_x. This was probably due to the fact that on high-surface-area catalysts such as NV12c or CV0c, total oxidation to carbon oxides occurs via acidic centers associated with dispersed Al³⁺ cations [28]. Furthermore, it was shown that the undesirable processes do not require Al³⁺ centers as such, because acid centers can also be formed on the Al-free catalyst (AV18c).

Moreover, the presence of specific structures such as α-Mg₂V₂O₇ (CV25c) or dispersed, easily reducible isolated VO₄ tetrahedra, suggested to be present on the surface of amorphous catalysts like DV13c, NV12c, or even CV3c, may enhance activity toward CO_x formation [98,99]. It was reported previously that the presence of α-Mg₂V₂O₇ on MgO decreases selectivity in the ODH of n-butane [21]. Nevertheless, in the current study, the following highest values for total selectivity of dehydrogenated species were recorded for the samples: 50% (CV25c), 43% (DV13c), 36% (CV3), and 33% (NV12c).

Surprisingly, hydrotalcite doped with reduced vanadium cations V³⁺ provided catalysts that were relatively stable. Over the course of a 2 h experiment (3 h including outgassing and cooling down), neither the conversion nor selectivities to n-butane dehydrogenation or oxidation products changed significantly (Figure 13). Seemingly, the concentration of oxidants also had no influence on the AV8c catalysts’ stability (Figure S10, Supporting Materials). On the other hand, highly loaded catalysts such as CV25c or AV18c lost their initial activity in the first stage of the reaction. Interestingly, catalysts obtained via direct precipitation had preserved activity, while another sample with a much lower amount of vanadium, CV3c, slowly deteriorated over the course of the reaction.

Another possible explanation for the high selectivity in n-butane ODH is the presence of magnesium in V-O-Mg units, which increases the basicity of the mixed oxides compared with V-O-V units. Additionally, the presence of reduced V cations in the structure or their in situ reduction by alkanes can further enhance basicity, leading to faster desorption of butenes [18,19,79,99,100,101].

Due to the synergistic effect between the specific structure, basic properties, and high transition metal loading, the highest specific yields (SYs) of alkenes per surface area (2.8, 3.5, and 3.7 mmol/h·m²) were achieved for catalysts CV25c, AV8c, and AV18c, respectively. The formation of butene isomers was favored on CV25c (36%), whereas 1,3-butadiene formation reached 22% and 34% for AV8c and AV18c, respectively. The latter one, AV18c, also showed lower selectivity to CO_x (34%).

It can be concluded that oxide systems obtained from partially (AV8c) or fully (AV18c) V³⁺-substituted hydrotalcites are the best catalysts for n-butane ODH among all the samples tested. Moreover, AV8c is more stable than its greater V-loaded counterpart. Partial reduction of the catalyst, and therefore increased basicity [79], may facilitate the activation of alkane molecules, but on the other hand, it may favor the desorption of basic dehydrogenation reaction products: alkenes. Unfortunately, progressive deactivation and operational instability of such catalysts should be expected. As shown in the work of Chang et al. [79], for the Mg₂VO₄ catalyst, reversible oxidation to Mg₃(VO₄)₂ occurs, with the formation of excess MgO. This is confirmed by reducibility studies conducted for fresh and used AV8c catalysts as well as XPS studies.

Selection of a suitable V-containing LDH precursor and careful control of the preparation procedure can be highly beneficial for catalytic activity in n-butane ODH and for selectivity toward desired products. Within the studied group, the best catalytic results were obtained when V³⁺-doped hydrotalcites were calcined in an inert atmosphere. However, special care must be taken during the catalytic reaction to avoid excessive oxidation and subsequent deactivation of the catalysts [79]. On the other hand, optimum conditions must also be identified to prevent coke formation.

3. Materials and Methods

The parent Mg-Al hydrotalcites containing either carbonates or nitrates as compensating anions were prepared using a standard co-precipitation method at a constant pH. A solution of Mg(NO₃)₂·6H₂O (51.2 g, 0.2 mol) and Al(NO₃)₃·9H₂O (37.5 g, 0.1 mol) in 200 cm³ of distilled water was added dropwise to 100 cm³ of H₂O or an aqueous solution of Na₂CO₃ (6.6 g, 0.06 mol) to obtain nitrate (NV0) or carbonate (CV0) hydrotalcites, respectively. Co-precipitation was carried out at 60 °C under vigorous stirring at pH = 10.0 ± 0.2, controlled by the simultaneous addition of a 10% aqueous NaOH solution. The suspensions were stirred for 1 h at 60 °C.

The same conditions were applied for the direct co-precipitation of vanadate-intercalated Mg-Al hydrotalcite, except that Mg and Al salts were added to an aqueous solution containing 0.40 mol of V₂O₇⁴⁻ (sample DV13).

The parent hydrotalcites were subjected to ion exchange at pH = 4.5 (CV0) or 9.5 (CV0 and NV0). The pH of an approximately 5% hydrotalcite suspension was adjusted using HNO₃ (1:10) or a 10% NaOH solution, followed by the addition of an appropriate vanadate solution in 25% excess. The vanadate solution was prepared by dissolving 11.7 g of NaVO₃·H₂O in 280 cm³ of H₂O and adjusting the pH to 4.5 for decavanadate or 9.5 for pyrovanadate. The total time for addition and aging at 60 °C was 1 h at pH = 4.5 and 2 h at pH = 9.5. The decavanadate-exchanged sample was denoted CV25, while the pyrovanadate-exchanged sample was denoted NV12. Another sample with pyrovanadates adsorbed on the surface at pH = 9.5 was denoted CV3.

At a constant pH = 9.0 ± 0.2, two samples doped with V³⁺ were obtained. For sample AV18, a mixture of MgCl₂·6H₂O (13.2 g, 0.06 mol) and VCl₃ (5.1 g, 0.03 mol) was used, and for sample AV8, a mixture of MgCl₂·6H₂O (13.8 g, 0.07 mol), AlCl₃·6H₂O (5.5 g, 0.01 mol), and VCl₃ (1.8 g, 0.02 mol) was used. In each case, the salts were dissolved in 65 cm³ of deionized H₂O and added dropwise at room temperature to 100 cm³ of Na₂CO₃ solution (3.6 g, 0.05 mol). A 10% aqueous NaOH solution was used to maintain the pH level. Both samples were subjected to hydrothermal treatment at 150 °C for 6 days under autogenous pressure in the mother solution.

After short aging or hydrothermal treatment, the suspensions were filtered, washed with distilled water, and dried in air at 60 °C for 18 h or in argon at 40 °C for 24 h for the V³⁺-doped samples. An inert atmosphere (Ar) was also applied during calcination of the samples containing reduced vanadium cations, while other samples were thermally treated in static air. Calcination was carried out at 700 °C for 13 h. To distinguish hydrotalcite precursors from mixed oxides, the letter “c” was added to the names of the calcined samples. The number in the sample name indicates the vanadium content (wt%) in the mixed oxide.

Powder XRD patterns of the hydrotalcites (precursors) and the hydrotalcite-derived oxides (catalysts) were obtained using a Philips X’pert PW 3710 diffractometer (CuKα radiation, λ = 1.54184 Å). The XRD patterns were analyzed using Match!—Phase Analysis using Powder Diffraction from Crystal Impact software (Match! 4) [102]. FTIR spectra were recorded on a Bruker IFS 48 spectrometer (Ettlingen, Germany) using the KBr pellet technique. Structural analysis was supported by chemical bulk analysis performed using an elemental analyzer (EuroEA 3000, EuroVector S.p.A., Pavia, Italy), flame atomic absorption spectrometry (F-AAS, ContrAA 800 D, Jena Analytik, Jena, Germany), and ion chromatography (Shimadzu IC-CDD equipped with a Shim-pack IC-C4 cation column and conductivity detector, Kyoto, Japan) after dissolution of the catalysts in nitric acid. XPS spectra were recorded after evacuation in a VG ESCALAB 220i-XL spectrometer (VG Scientific, Waltham, MA, USA).

Prior to further characterization and catalytic tests, the catalysts were pelletized and crushed, and a fraction of the particle was separated (355–500 µm). The textural properties of the calcined materials were determined via low-temperature N₂ adsorption using an ASAP 2010 (Micromeritics, Norcross, GA, USA) sorptometer. The surface acid-base properties were assessed through temperature-programmed desorption (TPD) of NH₃ and CO₂. After outgassing in an N₂ flow and cooling to ambient temperature, a flow of NH₃ (1% NH₃/He, 20 cm³/min) or CO₂ (85 cm³/min) was applied. For temperature-programmed reduction, similar conditions for outgassing were used. Then, after cooling down, the samples were heated in a flow of hydrogen (6% H₂/Ar = 16 mL/min). Desorption and reduction were monitored using a VG SX-200 quadrupole mass spectrometer (VG Instruments, Stamford, CT, USA). The detailed conditions of the TPD and TPR processes are summarized in

Table 7. Experimental conditions for determination of acido-basic properties and reducibility of the V-doped catalysts.

TPR	NH3-TPD	CO2-TPD
outgassing: RT-600 °C, β~30 °C/min He = 20 mL/min	outgassing: RT-600 °C, β~30 °C/min He = 20 mL/min	outgassing: RT-600 °C, β~30 °C/min N2 = 85 mL/min
-	adsorption NH3: 70 °C, 45 min 1% NH3/He = 20 mL/min	adsorption CO2: 70 °C, 4 h CO2 = 85 mL/min
TPR: 200–800 °C, β = 15 °C/min 6% H2/Ar = 16 mL/min mass = 30 mg particle size: 355–500 μm detector: QMS	TPD: 70–600 °C β = 10 °C/min He = 20 mL/min mass = 50 mg particle size: 355–500 μm detector: QMS	TPD: RT-600 °C β = 10 °C/min He = 20 mL/min mass = 50 mg particle size: 355–500 μm detector: QMS

QMS—quadrupole mass spectrometer.

.

The catalytic activity in n-butane ODH was studied in a plug-flow microreactor loaded with 50 mg of the catalyst. After outgassing in N₂ at 550 °C, the catalyst bed was fed a reaction mixture (10% n-C₄H₁₀, 10% O₂, 80% N₂, total flow = 100 cm³/min). The reaction temperature (400–700 °C) as well as the oxygen content (0–18 vol.%) were preselected based on performance (conversion of substrate and selectivity of dehydrogenation and oxidation products). Substrates and products were analyzed online for period of 2 h using a Varian 3400 gas chromatograph (Palo Alto, CA, USA) equipped with TCD and FID detectors. Catalytic performance was expressed as the conversion of a substrate (X, expressed as a percentage) and selectivity of products (S, expressed as a percentage), as well as the specific activity (SX) and specific yield of ODH products (SY) per catalyst surface area (mmol/h·m²).

4. Conclusions

Hydrotalcites doped with vanadium via four different procedures exhibited a variety of XRD patterns within the range typical for these materials.

Samples obtained via pyrovanadate exchange displayed low-crystalline diffuse structures pillared with vanadates, with basal reflections shifted to lower 2θ angles, (NV12) or no significant changes (CV3) when nitrate- or carbonate-containing parent materials were used. The sample precipitated directly in the presence of a vanadate solution (DV13) also showed a diffuse structure, and the position of its basal reflections indicated the presence of pyro- and metavanadates in the interlayer space. In all of the above samples, carbonate contamination was detected.

In the sample pillared with decavanadates (CV25), the hydrotalcite structure was identified along with an additional Mg/Al vanadate phase, formed as a result of acidic leaching during the ion exchange process. The samples doped with V³⁺ (AV8c and AV18c) showed a well-defined hydrotalcite structure pillared with carbonates. The high crystallinity in these materials was attributed to the hydrothermal treatment applied during synthesis.

The vanadium content and oxidation state strongly influenced the phase composition of catalysts obtained via calcination of the corresponding hydrotalcites. All samples with low V contents (up to 13 wt% in the mixed oxide (CV3c, NV12c, DV13c, and AV8c)), regardless of vanadium oxidation state, contained a periclase (MgO) phase and traces of vanadates, likely Mg₃V₂O₈ or Mg₂VO₄ for the samples doped with oxidized or reduced vanadium, respectively. The decavanadate-pillared hydrotalcite transformed into a mixture of MgO and α-Mg₂V₂O₇ (CV25c), whereas the V³⁺-doped sample, after calcination in an inert atmosphere, formed a spinel-like phase of Mg₂VO₄.

Textural properties, such as the surface area and total pore volume, were determined by both V loading and the initial vanadium oxidation state. The hydrotalcites containing reduced V cations (AV8c and AV18c) and the highly V⁵⁺-loaded sample (CV25c) were nearly non-porous, with surface areas of ~30 and ~15 m²/g, respectively. Other catalysts (CV3c, NV12c, and DV13c) exhibited higher surface areas (196, 79, and 104 m²/g, respectively) and bottle-shaped mesopores in the 2 nm and 6–10 nm ranges.

The basicity measurements for the selected samples indicated significantly higher basicity than acidity. However, quantitative CO₂ adsorption analysis for the high surface area DV13c sample and the reduced V-doped samples (AV8c/AV18c) was affected by bulk carbonate formation and multilayer CO₂ adsorption.

In the catalytic study, the highest n-butane conversion (up to 51%) was recorded for the catalysts containing reduced vanadium cations (AV8c and AV18c) and for the amorphous, medium V-loaded, high surface area samples (DV13c and NV12c). The highest specific activity per surface area was observed for AV8c and AV18c (6.3–8.6 mmol n-C₄H₁₀/h·m²) as well as CV25c (5.8 mmol n-C₄H₁₀/h·m²). These same samples also showed the highest specific yield of dehydrogenated products per surface area, reaching up to 3.7 mmol/h·m².

The formation of butene isomers (up to 36%) was favored on highly basic, reduced-vanadium-doped catalysts (AV8c and AV18c) and the highly V⁵⁺-loaded catalyst (CV25c), whereas the highest selectivity for 1,3-butadiene (S = 34%) was observed mainly for the AV8c and AV18c group. Overall, the most effective catalysts in the ODH of n-butane, combining high activity with low total oxidation, were the highly basic catalysts containing reduced vanadium.

To fully utilize the potential of the designed vanadium catalysts, it is necessary to consider depositing the precursor on a support with a high specific surface area. High dispersion of the active phase will contribute to better utilization of the active sites and reduce metal loading. Furthermore, modifiers that further stabilize the active phase should be considered.