Catalytic Oxydehydrogenation of Propane and Butane Under Free-of-Oxygen Atmospheres

Abstract

This review examines the state-of-the-art in catalytic oxidative dehydrogenation of propane and butane to olefins under oxygen-free gas-phase conditions (FOx–PCODH and FOx–BCODH). First, it is compared with traditional methods for C₃–C₄ olefin production, using reaction enthalpies as an indicator of the process energy required. The role of catalyst oxygen species, including lattice and chemisorbed oxygen, is discussed in relation to olefin formation under oxygen-free gas-phase operation. This is followed by a review of V₂O₅-based catalysts, with particular emphasis on co-catalysts and catalyst supports. The performance of V₂O₅-based catalysts for FOx–PCODH is assessed using overall olefin formation rates. Special attention is given to catalyst performance in the mini-fluidized CREC Riser Simulator, which exhibit the highest overall olefin formation rates. These studies are also valuable to shed light into the FOx–PCODH reaction mechanism and to determine paraffin conversion, olefin selectivity, oxygen availability, and coke formation. This data can be used to determine FOx–PCODH kinetic parameters by assessing the effects of temperature, contact time, paraffin partial pressure, and catalyst-to-feedstock ratio. Finally, this review analyzes Barracuda-based numerical simulations of a large-scale downer reactor using statistically derived kinetic parameters. These simulations are employed to confirm the viability of the FOx–PCODH process at an industrial scale.

1. Introduction

Light olefins, also known as lower alkenes with the formula C_nH_2n, such as ethylene, propylene, and butenes, are key components in the petrochemical industry. The production of light olefins is motivated by the need to manufacture a wide range of products, including polymers, plastics, and chemicals, such as acrylic acid, isopropanol, and epoxides [1,2,3,4,5]. These chemical intermediates can be transformed into household appliances, plastic films, packaging and labels, textiles, automotive components, and laboratory equipment [1,6]. Consequently, due to increased consumption of their derived products, the demand for light olefins has been increasing steadily. In 2020, the worldwide demand for propylene reached 115.2 million metric tons (MMT). It is expected to continue growing to 158.4 MMT in 2027 [7].

Light paraffins such as ethane, propane, and butane are the primary components of natural gas. These compounds are alkanes with fewer than five carbon atoms and are generally represented by the formula C_nH_2n+2 [8]. Ethane can be steam-cracked via homogeneous, endothermic reactions to produce mainly ethylene [9,10,11,12]. Although this is a well-established large-scale process, industrial companies remain reluctant to introduce process modifications given already affected large scale investments, despite its significant negative environmental impact.

Likewise, light olefins such as propylene, 1-butene, and iso-butene can also be produced via energy demanding steam cracking and fluid catalytic cracking [13,14,15,16,17]. These technologies often use feedstocks such as vacuum refinery hydrocarbon residues, gas oil, and de-asphalted oil, with olefins being a byproduct [1,3,13,18]. These processes require high temperatures, large amounts of energy, and have low selectivity to light olefins [6,19,20]. In addition, large quantities of coke are formed, leading to catalyst deactivation, sintering, and frequent plant shutdowns [3,6,21,22,23].

Therefore, there is both opportunity and interest in the development of environmentally acceptable and economically viable technologies to meet the market light-olefin demand. A promising alternative for producing propylene and butenes from light paraffins by using more environmentally friendly technologies is catalytic oxidative dehydrogenation (CODH), in which oxygen is co-fed with the paraffins in gas–phase mixtures. CODH renders the reaction mildly exothermic while limiting coke formation [24,25,26]. Nevertheless, CODH selectivity to olefins can be negatively affected by the exothermic combustion of light paraffins, which generates CO_x, water, and other byproducts, such as methane. CODH with oxygen in the gas phase also raises concerns regarding the formation of flammable and potentially explosive hydrocarbon–oxygen mixtures in the gas phase [9,27,28,29].

To address these various issues, CODH under oxygen-free conditions in the gas phase (FOx–CODH) has been proposed. This process uses the catalyst lattice to provide the needed oxygen, offering the intrinsic advantage of converting light paraffins into olefins, with high selectivity [1,9,30,31,32].

The present review focuses on FOx–CODH, with special consideration given to vanadium doped on different catalyst supports. Efficient methodologies for catalyst performance evaluations that use the CREC Riser Simulator invented and developed at the Chemical Reactor Engineering Center (CREC) laboratories at the University of Western Ontario, Canada [33] are also reported. Furthermore, the present review highlights the value of this mini-fluidized bed CREC Riser Simulator unit for the evaluation of light paraffin conversion and the determination of olefin selectivity, olefin/H₂ ratio, and coke formation. Through this approach, one can develop a mechanistic and kinetic analysis of FOx–CODH together with the establishment of key reaction parameters. This review also examines a FOx–CODH integrated process in a large-scale downflow reactor by using Barracuda^TM 17.3.0 software simulations.

2. Stoichiometry and Reaction Enthalpies

Table 1 reports standard reaction enthalpies (∆Hr) at 298 K for the conversion of propane into olefins via (a) dehydrogenation, (b) oxydehydrogenation with oxygen being provided in the gas phase, and (c) the FOx–CODH process.

In Section A of Table 1, non-oxidative propane dehydrogenation stoichiometries are shown (Equations (2)–(6)). These reactions are thermodynamically limited and strongly endothermic [14,34,35,36,37]. As a result, these processes need to be conducted at high temperatures (>650 °C) while using low partial pressures of light paraffins to yield limited amounts of undesired products (e.g., CO, CO₂, CH₄, coke) [37,38,39,40,41].

On the other hand, the ODH of light paraffins, formulated using Equation (1), is an exothermic reaction:

$${\mathrm{C}}_{\mathrm{n}}{\mathrm{H}}_{2\mathrm{n}+2}\left(\mathrm{g}\right) + 0.5{\mathrm{O}}_2\left(\mathrm{g}\right) \to {\mathrm{C}}_{\mathrm{n}}{\mathrm{H}}_{2\mathrm{n}}\left(\mathrm{g}\right) + {\mathrm{H}}_2\mathrm{O}\left(\mathrm{g}\right)$$

(1)

Thus, there is no need for thermal energy to be supplied to either the ODH or CODH process (See Table 1, sections B to D). This is in contrast with the endothermic nature of light olefin dehydrogenation, where significant amounts of heat need to be provided [8,42].

As examples of the CODH process, one can consider both propane oxydehydrogenation (PODH) with oxygen being supplied in the gas phase (Equations (7)–(16)) and FOx–PCODH (17)–(21)) in which gas-phase oxygen is absent. As described in Table 1, PODH may be influenced by successive secondary oxidation reactions that may lead to the formation of undesired byproducts (Equations (18)–(21)) [27,43]. Therefore, for a better management of these undesirable reaction steps, it is critical to implement FOx–PCODH catalysts such as V₂O₅ under oxygen-free gas phase conditions with high olefin selectivity.

Table 1. Reaction enthalpies at 298 K and 1 atm for propane (${\mathrm{C}}_3{\mathrm{H}}_8$) dehydrogenation, propane Oxydehydrogenation, and propane catalytic oxydehydrogenation under oxygen-free conditions.

Method	Possible Chemical Reactions	Eq	∆Hr a
(A) Propane Catalytic Dehydrogenation [1]	${\mathrm{C}}_3{\mathrm{H}}_8{ \leftrightarrow \mathrm{C}}_3{\mathrm{H}}_6+{ \mathrm{H}}_2$	(2)	124.3
	${\mathrm{C}}_3{\mathrm{H}}_8 \to \mathrm{C}{\mathrm{H}}_4+{\mathrm{C}}_2{\mathrm{H}}_4$	(3)	81.4
	${\mathrm{C}}_2{\mathrm{H}}_4+{\mathrm{H}}_2{ \to \mathrm{C}}_2{\mathrm{H}}_6$	(4)	−137.2
	${\mathrm{C}}_3{\mathrm{H}}_8+{\mathrm{H}}_2 \to \mathrm{C}{\mathrm{H}}_4+{\mathrm{C}}_2{\mathrm{H}}_6$	(5)	−55.7
	${\mathrm{C}}_3{\mathrm{H}}_8 \to 3\mathrm{C}+4{\mathrm{H}}_2$	(6)	103.9
(B) Propane Catalytic Oxydehydrogenation (PODH) [1,31]	${\mathrm{C}}_3{\mathrm{H}}_8+0.5{\mathrm{O}}_2 \to {\mathrm{C}}_3{\mathrm{H}}_{6 }+{\mathrm{H}}_2\mathrm{O}$	(7)	−117.6
	${\mathrm{C}}_3{\mathrm{H}}_8+5{\mathrm{O}}_2 \to 3\mathrm{C}{\mathrm{O}}_{2 }+4{\mathrm{H}}_2\mathrm{O}$	(8)	−2044.0
	${\mathrm{C}}_3{\mathrm{H}}_8+3.5{\mathrm{O}}_2 \to 3\mathrm{CO}+4{\mathrm{H}}_2\mathrm{O}$	(9)	−1195.1
	${\mathrm{C}}_3{\mathrm{H}}_6+4.5{\mathrm{O}}_2 \to 3\mathrm{C}{\mathrm{O}}_{2 }+3{\mathrm{H}}_2\mathrm{O}$	(10)	−1926.4
	${\mathrm{C}}_3{\mathrm{H}}_6+3{\mathrm{O}}_2 \to 3\mathrm{CO}+3{\mathrm{H}}_2\mathrm{O}$	(11)	−1077.5
	$\mathrm{CO}+0.5{\mathrm{O}}_2 \to \mathrm{C}{\mathrm{O}}_2$	(12)	−283.0
	${\mathrm{C}}_3{\mathrm{H}}_8+1.5{\mathrm{O}}_2 \to {\mathrm{C}}_2{\mathrm{H}}_4+\mathrm{CO}+2{\mathrm{H}}_2\mathrm{O}$	(13)	−437.9
	${\mathrm{C}}_3{\mathrm{H}}_8+2{\mathrm{O}}_2 \to {\mathrm{C}}_2{\mathrm{H}}_4+\mathrm{C}{\mathrm{O}}_2+2{\mathrm{H}}_2\mathrm{O}$	(14)	−720.8
	${\mathrm{C}}_3{\mathrm{H}}_6+{\mathrm{O}}_2 \to {\mathrm{C}}_2{\mathrm{H}}_4+\mathrm{CO}+{\mathrm{H}}_2\mathrm{O}$	(15)	−320.3
	${\mathrm{C}}_3{\mathrm{H}}_6+1.5{\mathrm{O}}_2 \to {\mathrm{C}}_2{\mathrm{H}}_4+\mathrm{C}{\mathrm{O}}_2+{\mathrm{H}}_2\mathrm{O}$	(16)	−603.3
(C) Propane Catalytic Oxydehydrogenation under Oxygen-Free Conditions (FOx–PCODH) [1]	${\mathrm{C}}_3{\mathrm{H}}_8+{\mathrm{V}}_2{\mathrm{O}}_5={{\mathrm{V}}_2{\mathrm{O}}_4+\mathrm{C}}_3{\mathrm{H}}_6+{\mathrm{H}}_2\mathrm{O}$	(17)	5.9
	${\mathrm{C}}_3{\mathrm{H}}_8+10{\mathrm{V}}_2{\mathrm{O}}_5=10{\mathrm{V}}_2{\mathrm{O}}_4+3\mathrm{C}{\mathrm{O}}_2+4{\mathrm{H}}_2\mathrm{O}$	(18)	−809.7
	${\mathrm{C}}_3{\mathrm{H}}_8+7{\mathrm{V}}_2{\mathrm{O}}_5=7{\mathrm{V}}_2{\mathrm{O}}_4+3\mathrm{CO}+4{\mathrm{H}}_2\mathrm{O}$	(19)	−331.1
	${\mathrm{C}}_3{\mathrm{H}}_6+9{\mathrm{V}}_2{\mathrm{O}}_5=9{\mathrm{V}}_2{\mathrm{O}}_4+3\mathrm{C}{\mathrm{O}}_2+3{\mathrm{H}}_2\mathrm{O}$	(20)	−815.6
	${\mathrm{C}}_3{\mathrm{H}}_6+6{\mathrm{V}}_2{\mathrm{O}}_5=6{\mathrm{V}}_2{\mathrm{O}}_4+3\mathrm{CO}+3{\mathrm{H}}_2\mathrm{O}$	(21)	−337.0
(D) FOx–PCODH Chemisorbed Oxygen Parallel Reactions [44]	${\mathrm{C}}_3{\mathrm{H}}_8+\mathrm{O-S} \stackrel{ \mathrm{cat}}{\to } {\mathrm{C}}_3{\mathrm{H}}_6+{\mathrm{H}}_2\mathrm{O}+ \mathrm{-S}$	(22)	NA b
	${\mathrm{C}}_3{\mathrm{H}}_8+10\mathrm{O-S} \stackrel{ \mathrm{cat}}{\to } 3\mathrm{C}{\mathrm{O}}_2+4{\mathrm{H}}_2\mathrm{O}+10\mathrm{-S}$	(23)	NA
	${\mathrm{C}}_3{\mathrm{H}}_8+7\mathrm{O-S} \stackrel{\mathrm{cat}}{\to } +3\mathrm{CO}+4{\mathrm{H}}_2\mathrm{O}+7\mathrm{-S}$	(24)	NA
	${\mathrm{C}}_3{\mathrm{H}}_6+9\mathrm{O-S} \stackrel{ \mathrm{cat}}{\to } +3\mathrm{C}{\mathrm{O}}_2+3{\mathrm{H}}_2\mathrm{O}+9\mathrm{-S}$	(25)	NA
	${\mathrm{C}}_3{\mathrm{H}}_6+6\mathrm{O-S} \stackrel{ \mathrm{cat}}{\to } +3\mathrm{CO}+3{\mathrm{H}}_2\mathrm{O}+6\mathrm{-S}$	(26)	NA
(E) FOx–PCODH Catalyst Regeneration [1]	$2{\mathrm{V}}_2{\mathrm{O}}_4+{\mathrm{O}}_2 \to 2{\mathrm{V}}_2{\mathrm{O}}_5$	(27)	−246.9
(F) FOx–PCODH Catalyst Site Readsorption	$\mathrm{-S}+0.5{\mathrm{O}}_2 \to \mathrm{O-S}$	(28)	NA

Notes: ^a ΔHr in kJ·mol⁻¹, ^b NA = Non-Available.

Table 1. Reaction enthalpies at 298 K and 1 atm for propane (${\mathrm{C}}_3{\mathrm{H}}_8$) dehydrogenation, propane Oxydehydrogenation, and propane catalytic oxydehydrogenation under oxygen-free conditions.

Method	Possible Chemical Reactions	Eq	∆Hr a
(A) Propane Catalytic Dehydrogenation [1]	${\mathrm{C}}_3{\mathrm{H}}_8{ \leftrightarrow \mathrm{C}}_3{\mathrm{H}}_6+{ \mathrm{H}}_2$	(2)	124.3
	${\mathrm{C}}_3{\mathrm{H}}_8 \to \mathrm{C}{\mathrm{H}}_4+{\mathrm{C}}_2{\mathrm{H}}_4$	(3)	81.4
	${\mathrm{C}}_2{\mathrm{H}}_4+{\mathrm{H}}_2{ \to \mathrm{C}}_2{\mathrm{H}}_6$	(4)	−137.2
	${\mathrm{C}}_3{\mathrm{H}}_8+{\mathrm{H}}_2 \to \mathrm{C}{\mathrm{H}}_4+{\mathrm{C}}_2{\mathrm{H}}_6$	(5)	−55.7
	${\mathrm{C}}_3{\mathrm{H}}_8 \to 3\mathrm{C}+4{\mathrm{H}}_2$	(6)	103.9
(B) Propane Catalytic Oxydehydrogenation (PODH) [1,31]	${\mathrm{C}}_3{\mathrm{H}}_8+0.5{\mathrm{O}}_2 \to {\mathrm{C}}_3{\mathrm{H}}_{6 }+{\mathrm{H}}_2\mathrm{O}$	(7)	−117.6
	${\mathrm{C}}_3{\mathrm{H}}_8+5{\mathrm{O}}_2 \to 3\mathrm{C}{\mathrm{O}}_{2 }+4{\mathrm{H}}_2\mathrm{O}$	(8)	−2044.0
	${\mathrm{C}}_3{\mathrm{H}}_8+3.5{\mathrm{O}}_2 \to 3\mathrm{CO}+4{\mathrm{H}}_2\mathrm{O}$	(9)	−1195.1
	${\mathrm{C}}_3{\mathrm{H}}_6+4.5{\mathrm{O}}_2 \to 3\mathrm{C}{\mathrm{O}}_{2 }+3{\mathrm{H}}_2\mathrm{O}$	(10)	−1926.4
	${\mathrm{C}}_3{\mathrm{H}}_6+3{\mathrm{O}}_2 \to 3\mathrm{CO}+3{\mathrm{H}}_2\mathrm{O}$	(11)	−1077.5
	$\mathrm{CO}+0.5{\mathrm{O}}_2 \to \mathrm{C}{\mathrm{O}}_2$	(12)	−283.0
	${\mathrm{C}}_3{\mathrm{H}}_8+1.5{\mathrm{O}}_2 \to {\mathrm{C}}_2{\mathrm{H}}_4+\mathrm{CO}+2{\mathrm{H}}_2\mathrm{O}$	(13)	−437.9
	${\mathrm{C}}_3{\mathrm{H}}_8+2{\mathrm{O}}_2 \to {\mathrm{C}}_2{\mathrm{H}}_4+\mathrm{C}{\mathrm{O}}_2+2{\mathrm{H}}_2\mathrm{O}$	(14)	−720.8
	${\mathrm{C}}_3{\mathrm{H}}_6+{\mathrm{O}}_2 \to {\mathrm{C}}_2{\mathrm{H}}_4+\mathrm{CO}+{\mathrm{H}}_2\mathrm{O}$	(15)	−320.3
	${\mathrm{C}}_3{\mathrm{H}}_6+1.5{\mathrm{O}}_2 \to {\mathrm{C}}_2{\mathrm{H}}_4+\mathrm{C}{\mathrm{O}}_2+{\mathrm{H}}_2\mathrm{O}$	(16)	−603.3
(C) Propane Catalytic Oxydehydrogenation under Oxygen-Free Conditions (FOx–PCODH) [1]	${\mathrm{C}}_3{\mathrm{H}}_8+{\mathrm{V}}_2{\mathrm{O}}_5={{\mathrm{V}}_2{\mathrm{O}}_4+\mathrm{C}}_3{\mathrm{H}}_6+{\mathrm{H}}_2\mathrm{O}$	(17)	5.9
	${\mathrm{C}}_3{\mathrm{H}}_8+10{\mathrm{V}}_2{\mathrm{O}}_5=10{\mathrm{V}}_2{\mathrm{O}}_4+3\mathrm{C}{\mathrm{O}}_2+4{\mathrm{H}}_2\mathrm{O}$	(18)	−809.7
	${\mathrm{C}}_3{\mathrm{H}}_8+7{\mathrm{V}}_2{\mathrm{O}}_5=7{\mathrm{V}}_2{\mathrm{O}}_4+3\mathrm{CO}+4{\mathrm{H}}_2\mathrm{O}$	(19)	−331.1
	${\mathrm{C}}_3{\mathrm{H}}_6+9{\mathrm{V}}_2{\mathrm{O}}_5=9{\mathrm{V}}_2{\mathrm{O}}_4+3\mathrm{C}{\mathrm{O}}_2+3{\mathrm{H}}_2\mathrm{O}$	(20)	−815.6
	${\mathrm{C}}_3{\mathrm{H}}_6+6{\mathrm{V}}_2{\mathrm{O}}_5=6{\mathrm{V}}_2{\mathrm{O}}_4+3\mathrm{CO}+3{\mathrm{H}}_2\mathrm{O}$	(21)	−337.0
(D) FOx–PCODH Chemisorbed Oxygen Parallel Reactions [44]	${\mathrm{C}}_3{\mathrm{H}}_8+\mathrm{O-S} \stackrel{ \mathrm{cat}}{\to } {\mathrm{C}}_3{\mathrm{H}}_6+{\mathrm{H}}_2\mathrm{O}+ \mathrm{-S}$	(22)	NA b
	${\mathrm{C}}_3{\mathrm{H}}_8+10\mathrm{O-S} \stackrel{ \mathrm{cat}}{\to } 3\mathrm{C}{\mathrm{O}}_2+4{\mathrm{H}}_2\mathrm{O}+10\mathrm{-S}$	(23)	NA
	${\mathrm{C}}_3{\mathrm{H}}_8+7\mathrm{O-S} \stackrel{\mathrm{cat}}{\to } +3\mathrm{CO}+4{\mathrm{H}}_2\mathrm{O}+7\mathrm{-S}$	(24)	NA
	${\mathrm{C}}_3{\mathrm{H}}_6+9\mathrm{O-S} \stackrel{ \mathrm{cat}}{\to } +3\mathrm{C}{\mathrm{O}}_2+3{\mathrm{H}}_2\mathrm{O}+9\mathrm{-S}$	(25)	NA
	${\mathrm{C}}_3{\mathrm{H}}_6+6\mathrm{O-S} \stackrel{ \mathrm{cat}}{\to } +3\mathrm{CO}+3{\mathrm{H}}_2\mathrm{O}+6\mathrm{-S}$	(26)	NA
(E) FOx–PCODH Catalyst Regeneration [1]	$2{\mathrm{V}}_2{\mathrm{O}}_4+{\mathrm{O}}_2 \to 2{\mathrm{V}}_2{\mathrm{O}}_5$	(27)	−246.9
(F) FOx–PCODH Catalyst Site Readsorption	$\mathrm{-S}+0.5{\mathrm{O}}_2 \to \mathrm{O-S}$	(28)	NA

Notes: ^a ΔHr in kJ·mol⁻¹, ^b NA = Non-Available.

While the above postulated reactions for FOx–PCODH are all thermodynamically allowed, there is, however, still debate on whether oxydehydrogenation truly involves totally or partially oxidized vanadium chemical species, as described in Section C of Table 1 with Equations (17)–(21), or alternatively, if it involves chemisorbed oxygen obtained from other surface chemical species, as described in Section D of Table 1, and as follows:

$${\mathrm{C}}_{\mathrm{n}}{\mathrm{H}}_{2\mathrm{n}+2}\left(\mathrm{g}\right) + \mathrm{-O-S} \left(\mathrm{s}\right) \to {\mathrm{C}}_{\mathrm{n}}{\mathrm{H}}_{2\mathrm{n}}\left(\mathrm{g}\right) +{ \mathrm{H}}_2\mathrm{O}\left(\mathrm{g}\right) + \mathrm{-S}$$

(29)

with O-S representing the chemisorbed oxygen in the catalyst lattice.

Thus, the FOx–CODH process takes advantage of the catalyst’s structural lattice oxygen, also referred to as nucleophilic oxygen, O²⁻ or O_latt [45,46], with the amount of oxygen available being a function of the state of reduction of the metal oxide catalyst (e.g., V⁵⁺ to V³⁺) [6]. However, when the catalyst is fresh or has been regenerated, there can be other forms of oxygen available, such as weakly adsorbed or chemisorbed oxygen (also referred to as electrophilic oxygen species like O_elec, O₂⁻, or O⁻) [45,47].

As a result, by assuming a Mars-van Krevelen mechanism [48], in which one mole of propane reacts with one mole of lattice oxygen (C₃H₈ + -O- → C₃H₆ + H₂O), the amount of lattice oxygen accessible in the catalyst can be estimated. This mechanism is based on the metal oxide content and the stoichiometric oxygen available, as described in Figure 1. It is worth mentioning that the work of Rostom and de Lasa [49] showed that the amount of lattice oxygen available from a 7.5V/ZrO₂-γAl₂O₃ catalyst employed for FOx–PCODH exceeded the propane converted that was fed via successive propane injections by a factor of three, as shown in Appendix A.

Therefore, if the lattice oxygen were the only active species, one would expect it to be depleted after the third successive injection. Results show, however, that the propane conversion with stable high olefin yield continues over 10 consecutive injections, indicating that FOx–PCODH cannot be entirely attributable to the lattice oxygen [49].

Although a greater reduction of the catalyst (e.g., V₂O₅→V₂O₃) could, in principle, increase the availability of lattice oxygen, this is unlikely to occur under the FOx–PCODH operating conditions studied by Rostom and de Lasa [49]. By using an XPS analysis, these authors reported only the presence of V⁵⁺ and V⁴⁺ species on the used catalyst over the 10 consecutive injections. This sustained catalyst activity over multiple injections, therefore, suggests the involvement of extra oxygen species beyond those in the crystal lattice. In our view, it is important for future research to consider alternative reaction pathways that fully account for the sources of oxygen that contribute to the FOx–PCODH process.

One should mention that the V₂O₅ catalyst can be activated with air at 550–600 °C, as described by Equation (27) of Table 1. This V₂O₅ catalyst activation facilitates the integration of catalytic oxydehydrogenation and catalyst reactivation (also known as regeneration), in a single process, which includes an FOx–PCODH fluidized reactor and an air fluidized regeneration unit, both operating within a close temperature range [49].

It is interesting to note that one can study FOx–PCODH together with catalyst regeneration using either Route 1 (Equations (17)–(21) and (27)) or Route 2 (Equations (22) and (28)). While Route 1 is a function of the chemical state of vanadium only, Route 2 depends on the chemisorbed oxygen and the catalyst (O-S) bond strength. Thus, by establishing the oxygen chemisorption enthalpy, the reaction enthalpies of the reactions described by Equations in Sections D and F of Table 1 can be calculated. However, even if the specific values of the FOx–PCODH reaction enthalpies differ for Route 1 and Route 2, the overall heat of reaction (which is the algebraic addition of the enthalpy of reaction for the FOx–PCODH, and the heat of catalyst regeneration (Sections E and F, Table 1)) for both routes yield the same values. This observation is critical, as it removes the overall reaction enthalpy uncertainty in an integrated process, configured with a FOx–PCODH reactor and a catalyst regenerator.

It is noteworthy that the concepts that apply to the free-of-oxygen oxidative dehydrogenation of propane discussed in this review also apply to butane, as recently reported by Bin Sulayman and de Lasa [44].

Thus, it is our view that FOx–CODH offers a new approach for light paraffin conversion into olefins, with significant savings of thermal energy and no required natural gas consumption, providing a process free of CO₂ emissions.

3. FOx–PCODH Vanadium-Based Catalysts

The FOx–CODH of light paraffins has been conducted by employing metal oxide catalysts, including vanadium oxide-based catalysts, vanadium–magnesium–oxygen systems, and molybdate-based catalysts, among others. The general stoichiometric equation for the oxidative dehydrogenation of propane to propylene over metal oxide catalysts is given by Equation (30). Once the catalyst becomes deactivated through partial reduction, it can be regenerated using oxygen from air, as shown in Equation (31) [1].

$${\mathrm{C}}_3{\mathrm{H}}_8+\mathrm{M}{\mathrm{O}}_{\mathrm{x} }\to { \mathrm{C}}_3{\mathrm{H}}_6+{\mathrm{H}}_2\mathrm{O}+\mathrm{M}{\mathrm{O}}_{\mathrm{x}-1}$$

(30)

$$\mathrm{M}{\mathrm{O}}_{\mathrm{x}-1}+1/2{\mathrm{O}}_2\left(\mathrm{air}\right) \to \mathrm{M}{\mathrm{O}}_{\mathrm{x}}$$

(31)

One should note that the redox properties of vanadium-based catalysts play a crucial role in reactions that proceed via the Mars–van Krevelen mechanism. In this mechanism, the lattice oxygen from vanadium oxides reacts with hydrocarbon molecules to form alkenes and water. The resulting reduced vanadium species are subsequently re-oxidized by gaseous molecular oxygen in the regenerator, thereby replenishing the active catalyst sites [50].

FOx–PCODH involves three simultaneous and consecutive processes: the oxidative dehydrogenation of propane, the combustion of propane to form CO_x, and the combustion of propylene to form CO_x, as shown in Figure 2 and previously summarized in Table 1. The combustion of both propane and propylene limits both propylene yields and selectivities. Moreover, it is widely recognized that the propane CODH process can be both thermally and kinetically constrained, as selectivity toward the desired propylene decreases with increasing propane conversion [1]. As a result, a portion of both the propane feed and the propylene product can be further oxidized to carbon oxides.

A critical aspect in designing effective catalysts for alkane FOx–CODH is the stoichiometric control of oxygen supply to active sites. One must ensure that it meets the requirements for the selective oxidation of hydrocarbons to desired products or intermediates. This approach avoids the supply of excess oxygen, which promotes complete oxidation to undesired products. Strategies to achieve this controlled oxygen supply include the use of metals with intermediate oxidation states and the design of catalysts in which active oxygen species are dispersed among a few neighboring active sites. This can be further enhanced through the partial modification of oxide surfaces via sub-stoichiometric treatments that deactivate specific regions of the catalyst [8,51].

FOx–PCODH has been extensively studied using vanadium-based catalysts. Due to their low cost, high catalytic activity, thermal stability, and large specific surface area, supported vanadium oxides (VO_x) are among the most promising catalysts for dehydrogenation reactions. VO_x-based catalysts for alkane oxidation contain a variety of vanadium species in different oxidation states (V⁵⁺, V⁴⁺, V³⁺), including oligomeric Vⁿ⁺ species, isolated vanadium sites, and crystalline V₂O₅ phases. The distribution and reactivity of these species are strongly influenced by the nature of the support, the vanadium loading, and the preparation method [37,52].

Various synthesis methods, such as wet impregnation, incipient wetness impregnation, flame spray pyrolysis, and sol–gel, have been employed to optimize the dispersion and redox properties of vanadium-based catalysts [18,30,52,53].

Supported catalysts offer several advantages over unsupported systems, including improved control over metal loading and oxide active species dispersion, as well as greater flexibility in tuning physicochemical properties [1,25]. In the oxidative dehydrogenation of light alkanes, a wide range of supports has been investigated, including SiO₂, Al₂O₃, TiO₂, ZrO₂, CeO₂, Nb₂O₅, MgO, and zeolites [18,54,55]. Typically, a soluble vanadium precursor, such as ammonium metavanadate (NH₄VO₃), is deposited onto the support in a quantity equal to or close to the support pore volume via incipient wetness impregnation. This method generally yields a distribution of vanadium species on the support surface, including both V₂O₅ nanoparticles and isolated vanadyl species [18].

Table A1. Oxidative dehydrogenation of propane in different catalytic systems over the last decade.

Study	Reactor	Catalyst Composition	T [°C]	P [atm]	C%	S%	Y%	Wcat [g]	Time [min]	${{\overline{\mathbf{r}}}_{{\mathbf{C}}_3{\mathbf{H}}_6}}^{\mathrm{a}}$
[19]	Fluidized bed	7.5V/ZrO2-γ-Al2O3 (1:1)	550	1	25	94	23.5	0.4	0.33	21.6
[19]	Fluidized bed	7.5V/γ-Al2O3	550	1	26	90	23.4	0.4	0.33	21.5
[39]	Fixed bed	6V/30CeAl	600	1.4	46.6	93.5	43.6	0.5	30.00	9.9
[20]	Fluidized bed	VOx/CaO-γ-Al2O3	640	1	25.5	94.2	24.0	0.5	0.17	8.5
[56]	Fixed bed	Mo on VOx/Al2O3 V/Mo = 6	500	1.4	36	89	32.0	0.5	-	6.9
[19]	Fluidized bed	7.5V/ZrO2-γ-Al2O3 (1:1)	500	1	8	91	7.3	0.40	0.33	6.7
[20]	Fluidized bed	VOx/CaO	640	1	24.4	77.1	18.8	0.5	0.17	6.6
[6]	Fluidized bed	VOx/CaO-γ-Al2O3 (1:1)	640	1	65.1	85.2	55.5	0.5	0.52	6.3
[20]	Fluidized bed	VOx/CaO-γ-Al2O3	640	1	65	85	55.3	0.5	0.52	6.3
[58]	Fixed bed	VOx/SiO2	500	1	14.5	68	9.9	0.10	30.00	6.2
[57]	Fixed bed	Re on VOx/Al2O3 V/Re = 6	525	1.2	31.7	83.1	26.3	0.5	3.00	6.2
[59]	Fluidized bed	VOx/θ-Al2O3/BaO	625	1	97	50	49.0	0.5	0.50	5.7
[57]	Fixed bed	W on VOx/Al2O3 V/W = 6	525	1.2	30.4	79	24.0	0.5	3.00	5.7
[20]	Fluidized bed	VOx/CaO	640	1	64.1	70.3	45.1	0.5	0.52	5.1
[46]	Fixed bed	10V-3Ce/Al	600	1	24.7	83.4	20.6	0.5	3–15	5.1
[59]	Fluidized bed	VOx/θ-Al2O3/CaO	625	1	86	45	39.0	0.5	0.50	4.6
[20]	Fluidized bed	VOx/CaO-γ-Al2O3	640	1	14	91.6	12.8	0.5	0.17	4.5
[59]	Fluidized bed	V/θ-Al2O3	650	1	83	45	38.0	0.5	0.50	4.4
[30]	Fixed bed	Co3O4/SiO2	600	1	9.3	75.9	7.1	0.2	8	4.3
[6]	Fluidized bed	VOx/CaO–γ-Al2O3 (1:1)	640	1	13	92	12.0	0.5	0.17	4.2
[57]	Fixed bed	VOx/Al2O3	525	1.2	24.4	68	16.6	0.5	3.00	3.9
[20]	Fluidized bed	VOx/CaO	640	1	15	73.3	11.0	0.5	0.17	3.9
[59]	Fluidized bed	VOx/MgO-θ-Al2O3	650	1	97	34	32.0	0.5	0.50	3.9
[47]	Fixed bed	0.1 V/TiO2	500	1.4	9	95	8.6	0.25	-	3.7
[59]	Fluidized bed	VOx/MgO-θ-Al2O3	600	1	94	33	31.0	0.5	0.50	3.7
[59]	Fluidized bed	VOx/CaO-θ-Al2O3	650	1	63	39	24.6	0.5	0.50	2.9
[20]	Fluidized bed	VOx/CaO-γ-Al2O3	550	1	10.3	78.3	8.1	0.5	0.17	2.8
[30]	Fixed bed	VOx/SiO2	450	1	4	88.3	3.5	0.2	8	2.2
[20]	Fluidized bed	VOx/CaO	550	1	9.2	62.2	5.7	0.5	0.17	2.0
[59]	Fluidized bed	VOx/BaO-θ-Al2O3	600	1	65	24	15.6	0.5	0.50	1.8
[59]	Fluidized bed	V/θ-Al2O3	600	1	57	9	5.1	0.5	0.50	0.6
[30]	Fixed bed	NiO/SiO2	425	1	2	14.8	0.3	0.2	8	0.2

^a Overall rate of propylene formation [mmolC₃H₆·(gcat h)⁻¹]. Note: C% = propane conversion percentage, S% = C₄H₆ selectivity percentage, Y% = yield percentage.

in the Appendix C summarizes the various catalysts employed in FOx–PCODH, as reported in the technical literature. Vanadium remains the most widely researched active component. Recent studies focus on its modification either through atomic-scale doping with metals such as molybdenum (Mo), tungsten (W), and rhenium (Re) [56,57], or by combining it with oxygen carrier materials, such as ceria (CeO₂) nanodomains and Ca-doped SrFeO₃ perovskites [39,58]. Common supports used in these systems include silica (SiO₂), titanium dioxide (TiO₂), and alumina phases (γ-Al₂O₃ or θ-Al₂O₃) [6,30,47,59].

The acidity of the support strongly influences catalytic performance and product selectivity. Because propylene interacts more strongly with acidic sites than the propane feedstock, high surface acidity promotes the adsorption of propylene, increasing its residence time on the catalyst surface and augmenting the chances for further transformation into undesired products (e.g., overoxidation to CO_x and coke) [6,19]. To mitigate this effect, researchers employ basic or less acidic oxides to modify catalyst supports, forming mixed oxides such as CaO–γ-Al₂O₃, BaO–θ-Al₂O₃, MgO–θ-Al₂O₃, CeO₂–γ-Al₂O₃, and ZrO₂–γ-Al₂O₃. These modifications neutralize acid sites and moderate the binding strength between propylene and the catalyst surface [6,20,46,49,59]. Alternatively, and to address this issue, a UiO-66 support based on confinement agents has recently been explored to prevent the access of hydrocarbons to sites responsible for overoxidation during PCODH [60].

Desirable isolated VO₄ species are also formed at low vanadium loadings, before reaching monolayer surface coverage. At higher vanadium loadings, these isolated sites tend to promote the formation of undesirable agglomerates into polyvanadates, eventually leading to the creation of V₂O₅ crystallites, which cause the complete combustion of propane [1]. Raman spectroscopy can help to identify these undesirable V₂O₅ crystallites by a sharp band around 1040 cm⁻¹. This band corresponds to a V=O stretching vibration [30,47].

For the catalyst to be used in fluidized-bed reactors the particle size must promote good fluidization. The reported optimal average size is approximately 87 to 90 μm [49,53]. Additionally, typical prepared fluidizable catalysts displayed the following properties: (a) specific surface: 14–200 m²/g, (b) vanadium loading: 5–10%, (c) average crystallite size: <4 nm, (d) ammonia acidity: 8.26 cm³ STP/gcat free of vanadium species and 13 cm³ STP/gcat with added vanadium species [6,49,53,59].

FOx–PCODH Vanadium-Based Catalysts Performance

The FOx–PCODH catalyst performance depends on several factors, such as active metal dispersion, support effects, and redox properties. Although alumina is widely used in oxidative dehydrogenation and in modified vanadium systems, silica with well-dispersed vanadium is often cited as a stronger candidate for achieving high selectivity in CODH reactions, given its anticipated lower acidity.

In this context, the catalysts used in FOx–CODH processes must be prepared on supports with a suitable particle size and density (e.g., 40–90 μm and 1.21–1.73 g/cm³, respectively [53,61]) to ensure good fluidization under the selected operating conditions. In addition, the supports should possess an adequate specific surface area and pore structure (e.g., approximately 200 m²/g, and 72–109 Å mean pore size [49,53]) in order to facilitate the incipient wetness impregnation of vanadium precursor species.

Figure 3a reports literature data illustrating the relationship between propane conversion and propylene selectivity in FOx–PCODH systems. The observed trend is that there is a trade-off between activity and selectivity, demonstrating the difficulty of achieving both simultaneously. It is shown, however, that FOx–PCODH catalysts with high conversions (e.g., 65.1% reported in [6]) can still maintain a competitive selectivity (85.2%). However, surpassing both thresholds simultaneously remains challenging.

While the majority of studies report propane conversion and propylene selectivity as evaluation metrics, these parameters provide a limited view of the catalytic performance. One could instead consider the turnover frequency (TOF), which requires knowing the number of active catalyst sites for its calculation [47]. However, given the lack of such information, reaction rates based on the unit weight of the catalyst can be used instead as a basis for comparison [39,56]. Thus, for the FOx–PCODH case, an overall olefin formation rate can be established as a key parameter by using Equation (32) as follows:

$${\displaystyle \genfrac{}{}{0pt}{}{\mathrm{Overall} \mathrm{Rate} \mathrm{of}}{\mathrm{Propylene} \mathrm{Formation}}}{\overline{\mathrm{r}}}_{{\mathrm{C}}_3{\mathrm{H}}_6} = {\displaystyle \frac{\mathrm{Reactant} \mathrm{input} \times { \mathrm{C}}_3{\mathrm{H}}_8\mathrm{Conversion} \times {\mathrm{C}}_3{\mathrm{H}}_8 \mathrm{Selectivity}}{\mathrm{Reaction} \mathrm{Time} \times \mathrm{Weight} \mathrm{of} \mathrm{Catalyst} }}$$

(32)

It should be noted that the overall rates of propylene formation derived from Equation (32) and reported in Figure 3b were calculated based on catalyst weight and reaction time, using the highest values reported in the literature for each study considered. Interestingly, the application of Equation (32) yields performance results that differ significantly from those observed in Figure 3a. One might expect that catalysts with the highest conversion and selectivity would also exhibit the highest olefin productivity; however, this is not necessarily the case. The overall rate of propylene formation established at a set temperature also depends on operating conditions such as contact time and catalyst loading. Consequently, catalysts with moderate conversion and selectivity can outperform others in terms of olefin productivity when operating under conditions that favor higher reaction rates (e.g., low amounts of catalysts required, or short reaction times). Thus, these results suggest that catalytic performance should be evaluated based on olefin productivity (overall formation rate), rather than relying solely on conversion–selectivity relationships.

It is important to note that only a limited number of the best catalysts evaluated for FOx–PCODH are of the fluidizable type, as described by the green bars in Figure 3b [6,19,20,59]. One should mention that fluidizable catalysts were evaluated in the CREC Riser Simulator by Hugo de Lasa and co-workers at the University of Western Ontario [1,19], as well as by Mohammad Mozahar Hossain and collaborators at the King Fahd University of Petroleum and Minerals [6,20,59]. These fluidizable catalysts consistently demonstrated both suitable fluidization behavior and relatively high overall propylene formation rates. As such, they can be considered as promising candidates for further investigation and for potential application in the scale-up of circulating fluidized bed FOx–PCODH processes. However, additional studies are still recommended to confirm their stability, attrition resistance, and performance under extended time-on-stream conditions.

These results underscore the importance of not only catalyst development but also their evaluation in representative laboratory-scale reactor systems. In this context, the CREC Riser Simulator enables catalyst testing under conditions that closely approximate those of industrial fluidized riser and downer reactors. Such a system allows for the controlled assessment of key operating variables, including temperature, reactant partial pressure, reaction time, and catalyst-to-propane ratio, thereby providing a more relevant basis for catalyst screening, kinetic analysis, process scale-up, and future process development.

4. FOx–BCODH Vanadium-Based Catalysts

The objective of butane oxidative dehydrogenation (BODH) is to produce valuable C4 olefins (such as 1-butene, isobutene, and 1,3-butadiene) while overcoming the limitations of traditional processes like steam cracking and fluid catalytic cracking. As previously noted, these conventional routes typically suffer from high energy consumption due to elevated operating temperatures, thermodynamic constraints, and rapid catalyst deactivation [62,63,64].

Although ODH is an exothermic alternative, a major challenge is, as seen in the case of PODH, the undesired deep oxidation of butanes to carbon oxides (CO_x) [65]. The catalytic oxidative dehydrogenation of butane under a gas-phase oxygen-free atmosphere (FOx–BCODH) addresses this issue by enabling dehydrogenation through the lattice oxygen of solid metal oxide catalysts (e.g., vanadium oxides), rather than relying on co-fed molecular oxygen.

FOx–BCODH can involve five parallel processes as described in Figure 4: (i) the target reaction, in which n-butane reacts with lattice oxygen to form the desired C₄ olefins and water; (ii) the cracking of n-butane into lighter C₁–C₃ hydrocarbons (e.g., methane, ethane, ethylene, propane, and propylene); (iii) the combustion of n-butane to CO_x; (iv) the secondary combustion of the formed butenes to unwanted CO_x; and (v) the secondary combustion of the lighter hydrocarbons [66].

Regarding supported catalysts used in FOx–BCODH, researchers frequently employ γ-alumina (γ-Al₂O₃). As observed in these studies, the acidic nature of alumina supports can cause unwanted cracking, carbon deposition (coking), and deep oxidation. To address these issues, various strategies focus on modifying the acid–base properties of the support through doping. One reported approach involves adding MgO to vanadium-based alumina catalysts. MgO neutralizes strong acidic sites and reduces the metal–support interactions that contribute to over-oxidation [63]. Similarly, cerium doping decreases support acidity, enhances the thermal stability of the catalyst, and mitigates deactivation [49].

Unsupported catalysts such as zinc–molybdenum–vanadium mixed oxide (ZnMo_0.45V_0.55O_x) have also been investigated for iso-butane FOx–CODH to form isobutene. Researchers have modified this material by developing a solvent-free grinding synthesis that uses Vitamin C as a pore-forming agent, lowering the valence states of the metals and generating oxygen vacancies that enhance the reactivity of the lattice oxygen [17].

FOx–BCODH Vanadium-Based Catalysts Performance

Table 2. FOx–BCODH using different catalysts over the last decade.

Study	Reactor	Catalyst	T [°C]	C%	S%	Y%	Time [min]	${\overline{\mathbf{r}}}_{{\mathbf{C}}_4}$a
[44]	Fluidized bed	5 V/MgO-γAl2O3 (1:1)	500	27	87	23.5	0.17	19.8
[63]	Fluidized bed	5 V/MgO-γAl2O3 (1:1)	500	27	86	23.2	0.17	14.7
[62]	Fluidized bed	5.0VOx/0.2Ce-γ-Al2O3	450	10.7	62.4	6.7	0.08	2.1
[17]	Fixed bed	ZnMo0.45V0.55Ox-0.3 vitamin C	500	17.6	82.6	14.5	10.0	1.3
[17]	Fixed bed	ZnMo0.45V0.55Ox-0.3 Sorbitol	500	11.6	81.5	9.5	10.0	0.9
[17]	Fixed bed	ZnMo0.45V0.55Ox-0.3 citric acid	500	10.3	43.2	4.4	10.0	0.4
[17]	Fixed bed	ZnMo0.45V0.55Ox	500	8	22.6	1.8	10.0	0.2

^a Overall rate of butene formation [mmol C₄(g_cat·h)⁻¹]. Notes: C% = butane conversion percentage, S% = C₄ selectivity percentage, Y% = yield percentage.

summarizes the performance of various BCODH catalytic systems. The limited number of studies available, while promising, indicate that FOx–BCODH is still in its early stages. Throughout the reported data, reactor configuration plays an important role, with fluidized-bed systems significantly outperforming fixed-bed reactors in terms of conversion, yield, and overall olefin productivity rates [44,62,63]. This trend suggests that efficient gas–solid contact and continuous catalyst redox cycling are essential under oxygen-free conditions. It is also worth noting that, although high selectivities can be achieved, overall performance remains constrained by the limited conversion [62].

5. Experimental FOx–CODH Reactors with Fluidized Catalysts

As discussed in Section 3 and Section 4, experimental investigations of FOx–PCODH have primarily employed two bench-scale reactor configurations: fluidized-bed and fixed-bed reactors. Among fluidized-bed systems, the CREC Riser Simulator has been one of the valuable laboratory scale reactors used to study FOx–BCODH under short contact-time conditions, typically ranging from 5 to 31 seconds [6,19,59]. Maintaining such brief exposure times is critical, as it minimizes deep oxidation pathways and thereby enhances olefin selectivity.

Figure 5 illustrates the CREC Riser Simulator unit, a 53 cm³ mini fluidized-bed batch reactor that can be used for evaluating fluidizable catalysts and developing kinetic models for paraffin CODH [1,33]. The reactor consists of an upper shell and a lower shell, a catalyst basket, and a high-speed impeller operating between 4500 and 7000 rpm. These components ensure stable fluidization and minimize transfer limitations. Typical catalyst loadings range from 0.40 to 0.76 g per experiment. Light paraffins are injected in pulses of 1.2–10 mL per run, while reaction temperatures are maintained between 475 and 650 °C, depending on catalyst formulation and on the targeted conversion and selectivity levels.

The operation of the CREC Riser Simulator involves achieving the desired emulated reactivity found in a continuous riser or downer. The steps involved are as follows (see Figure 5b):

(a)

Run preparation—Step 1: The V₁, 4PV, and V₂ valves are set to the “open” position and the 6PV is set to the “bypass sample loop” position with lines 1, 2, 3, 4, 5, 7 being interconnected forming a single continuous flow path, while the V₄ valve is set to the “vent” position.

(b)

Run Preparation—Step 2: The 4PV is turned to the “closed” position and the V₄ valve is connected to the vacuum box, with the sample loop of the 6PV being set to the “load position”.

(c)

Run preparation—Step 3: The 6PV is turned back to the “bypass sample loop position”.

(d)

End of Reaction Time Step: The 4PV is turned to the “open” position to allow reactant products to be quickly removed from the reactor and sent to the vacuum box.

(e)

Post Reaction Step: The vacuum box contents are mixed via a mixing device. Following this, the 6PV is turned to the “load” position and a product gas sample with a gas carrier is sent to the GC via lines six and eight.

Typically, following a series of 10 propane injections performed according to the protocols described above, a continuous flow of air is introduced for 15 min at 575 °C [49]. This procedure facilitates the combustion of the coke deposited during the reaction and ensures complete regeneration (reoxygenation) of the FOx–CODH catalyst. Catalyst samples collected before the coke combustion step are analyzed by Total Organic Carbon (TOC) analysis to quantify the coke formed during the reaction.

Typical CODH pressure profiles in the CREC Riser Simulator during and after the reaction are presented in Figure 6. Immediately after the paraffin sample is injected into the reactor using a syringe, the pressure rises sharply from atmospheric conditions. As the FOx–CODH reaction proceeds, the pressure continues to increase due to product formation. Once the designated reaction time has elapsed, the unreacted feed and reaction products are transferred almost instantaneously to a vacuum box, producing the characteristic sudden pressure drop observed in the profile.

Researchers both at the University of Western Ontario, Canada and King Fahd University of Petroleum and Minerals, Sausi Arabia successfully evaluated FOx–CODH by studying ethane, propane, and butane conversions and selectivities, using the Riser Simulator reactor [6,9,19,20,32,44,49,53,59,62,63,67,68]. Their work has involved various VO_x/γ-Al₂O₃ catalyst configurations and different paraffin reactants (ethane, propane, and butane) under oxygen-free conditions.

Table 3. FOx–CODH catalysts evaluated in the CREC Riser Simulator.

Paraffin	Catalyst	Paraffin Conversion	Olefin Selectivity	Olefin Formation Rate a	Reference
Propane	7.5V/ZrO2-γ-Al2O3 (1:1)	25%	94%	21.6	[19,49]
	(10–5%) VOx/γ-Al2O3	11.73–15.11%	67.65–85.89%	7.7–12.6	[53]
	VOx/CaO-γ-Al2O3	25%	94%	8.5	[20]
	VOx/CaO–γ-Al2O3 (1:1)	65.1%	85.2%	6.3	[6]
	VOx/θ-Al2O3/BaO	97%	50%	5.7	[59]
Butane	VOx/MgO-γAl2O3	27%	87%	19.8	[44,63]
	5.0VOx/0.2Ce-γ-Al2O3	10.7%	62.4%	2.1	[62]

^a Overall rate of olefin formation [mmol C₄(g_cat·h)⁻¹].

reports various vanadium-based catalysts for propane and butane FOx–PCODH, assessed in recent years in the CREC Riser Simulator. Several catalyst formulations, particularly ZrO₂-γ-Al₂O₃ and CaO-γ-Al₂O₃ support for propane and MgO-γ-Al₂O₃ for butane, achieved selectivities of around 85–94% while maintaining a moderate paraffin conversion. It is worth noting that in these experiments, at the evaluated catalyst loadings, the CREC Riser Simulator was able to reproduce the anticipated hydrodynamics and reaction conditions characteristic of continuous industrial riser or downer fluidized-bed units, despite operating in batch mode. Moreover, the high overall olefin formation rates achieved under short-contact-time conditions further demonstrate the suitability of fluidized-bed reactors for oxygen-free CODH applications.

Additionally, it was observed that after the first two injections propylene selectivity increased while CO_x formation decreased. This trend was attributed to the reduction of the available labile lattice oxygen [53]. This demonstrates that achieving the highest propylene selectivities in FOx–CODH requires the consumption of such labile oxygen in the catalyst, leading to a partially reduced catalyst state.

Fluidized bed reactors surpass fixed bed reactors as technologies of choice for exothermic FOx–PCODH and FOx–BCODH. Exothermic fixed-bed reactors have issues with reaction temperature control, both along the reactor unit and inside the catalyst pellets. Axial and radial temperature gradients in these units may also promote undesired total oxidation reactions and accelerate catalyst deactivation. Consequently, fixed beds require special design approaches to manage heat accumulation (catalyst dilution, multitubular reactors with heat exchange), which increase the operational complexity and costs [1].

In contrast, circulating fluidized beds, as the ones envisioned for FOx–CODH, provide better heat and mass transfer properties, thus permitting quasi-isothermal operation with minimal diffusional issues in the fluidized particles. Fluidized reactors allow catalyst circulation between reaction and regeneration units, which is helpful for an integrated FOx-CODH process with continuous lattice oxygen regeneration.

6. FOx-CODH Reaction Network Validation

Establishing the validity of the FOx–PCODH and FOx–BCODH reaction networks requires integrating the reaction stoichiometry, the catalyst characterization, the experimental product analysis, the controlled parameter variation, and the mathematical modeling.

An initial step in developing and optimizing FOx–PCODH and FOx–BCODH reaction networks is to conduct experimental catalytic runs in reactor units loaded with fluidizable inert support particles (e.g., Al₂O₃) under the chosen temperatures, partial pressures, and reaction times required to minimize light paraffin thermal conversion [6,45,49,53]. These “blank” experiments typically yield mostly unconverted propane with only trace amounts of methane, ethane, and CO₂. This confirms that paraffin thermal conversion is negligible under these investigated conditions.

FOx–CODH reactions are monitored during experimental runs. This is accomplished through the analysis of reactor effluents by using a gas chromatograph (GC) equipped with: (a) a thermal conductivity detector (TCD), (b) a flame ionization detector (FID), and (c) a mass spectrometry detector (MS) [20,39,53]. These analytical techniques enable the identification and quantification of products such as (a) unconverted paraffins, (b) olefins, (c) CO, CO₂, H₂O, H₂, CH₄, and (d) other light hydrocarbons. This comprehensive analysis allows for carbon balance calculations and the monitoring of species consumption and formation as the lattice oxygen is depleted and the catalyst progressively deactivates.

The proposed triangular reaction network reported in Figure 2 can also be validated through selectivity–conversion analysis. The propylene and carbon oxide selectivities at near-zero propane conversion show that these species are formed from propane via a parallel reaction network. One can notice, however, that at a high propane conversion, olefin selectivity typically decreases while CO_x formation increases. This indicates that propylene, as a primary FOx–PCODH product, is depleted later via an undesirable secondary oxidation into carbon oxides [20,53].

This triangular reaction network was validated by using in situ Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS) [39,46,47]. This technique enables the observation of chemical double bond formation. DRIFTS reveal C–H bond stretching and V–OH bond formation, indicating that propane activation proceeds via hydrogen abstraction at vanadium–oxygen surface sites. Additionally, C=C stretching vibrations were detected, confirming the formation of propylene intermediates. Furthermore, C=O stretching bands observed in the 1660–1680 cm⁻¹ region also suggest the formation of oxygenated species, such as acetone, which is considered an intermediate product obtained prior to the generation of CO_x [39,47].

In addition to evaluating gas-phase products, catalyst deactivation caused by coke formation can be assessed [53,63]. In these cases, the amount of carbon deposited on the spent catalyst is quantified using a Total Organic Carbon (TOC) Analyzer. By using this method, the deposited coke is determined via its combustion to CO₂. It is subsequently measured by using an infrared gas analyzer.

Hydrogen formation in the reactor effluent can also serve as a useful indicator to evaluate catalytic oxidative dehydrogenation (CODH) mechanisms [39,63]. In this context, by studying FOx–BCODH, our team has proposed that the C₄-olefin/H₂ ratio can be used as a diagnostic parameter for mechanism confirmation [63,69]. In non-oxidative dehydrogenation, this ratio is expected to fall between 0.5 and 1, in accordance with the reaction stoichiometry. However, significantly higher values observed (ranging from 3.09 to 3.99, and up to 7.34) indicate the suppression of H₂ formation. This is attributed to the oxidative dehydrogenation reaction, where hydrogen is removed primarily as H₂O through the contribution of lattice oxygen [63]. This validation approach can also be extended to FOx–PCODH.

7. Kinetics Models

7.1. FOx–PCODH

The kinetic models proposed to describe FOx–PCODH in recent years can generally be classified as either Langmuir–Hinshelwood (L–H) [19,20] or Mars–van Krevelen (MvK) models [31]. The L–H approach assumes that reactant species are adsorbed onto catalyst active sites prior to the reaction. In contrast, the MvK model proposes that the lattice oxygen from metal oxides within the catalyst plays a key role in limiting the reaction of hydrocarbon molecules [18]. On this basis, kinetics models have been developed based on reactions frameworks for FOx–PCODH by using the CREC Riser Simulator and a series-parallel propane reaction network, as described in Figure 2 of Section 3.

These models rely on multiple interrelated equations. In L–H models, the surface coverage of adsorbed species (θᵢ) is described by using adsorption equilibrium constants (Kᵢ) in combination with reactant partial pressures (pᵢ) [19] or concentrations (Cᵢ) [20]. To account for the decrease in available oxygen, given that the reaction utilizes lattice oxygen rather than gaseous oxygen, these models incorporate an exponential decay term, $\exp (-\mathrm{\lambda }{\mathrm{X}}_{\mathrm{C}3\mathrm{H}8})$, where λ is the decay constant, and X_C3H8 represents the propane conversion. The resulting r_i reaction rates obtained by combining these expressions within the L–H framework [19], are presented in Equations (33)–(35) as follows:

$${\mathrm{r}}_1 = {\displaystyle \frac{{\mathrm{k}}_1{\mathrm{K}}_{{\mathrm{C}}_3{\mathrm{H}}_8}{\mathrm{P}}_{{\mathrm{C}}_3{\mathrm{H}}_8}}{1 + {\mathrm{K}}_{{\mathrm{C}}_3{\mathrm{H}}_8}{\mathrm{P}}_{{\mathrm{C}}_3{\mathrm{H}}_8} + {\mathrm{K}}_{{\mathrm{C}}_3{\mathrm{H}}_6}{\mathrm{P}}_{{\mathrm{C}}_3{\mathrm{H}}_6} + {\mathrm{K}}_{{\mathrm{CO}}_{\mathrm{x}}}{\mathrm{P}}_{{\mathrm{CO}}_{\mathrm{x}}}}} \times \exp \left(-\mathrm{\lambda }{\mathrm{X}}_{{\mathrm{C}}_3{\mathrm{H}}_8}\right)$$

(33)

$${\mathrm{r}}_2={\displaystyle \frac{{\mathrm{k}}_2{\mathrm{K}}_{{\mathrm{C}}_3{\mathrm{H}}_8}{\mathrm{P}}_{{\mathrm{C}}_3{\mathrm{H}}_8}}{1+{ \mathrm{K}}_{{\mathrm{C}}_3{\mathrm{H}}_8}{\mathrm{P}}_{{\mathrm{C}}_3{\mathrm{H}}_8}+{\mathrm{K}}_{{\mathrm{C}}_3{\mathrm{H}}_6}{\mathrm{P}}_{{\mathrm{C}}_3{\mathrm{H}}_6}+{\mathrm{K}}_{{\mathrm{CO}}_{\mathrm{x}}}{\mathrm{P}}_{{\mathrm{CO}}_{\mathrm{x}}}}} \times \exp \left(-\mathrm{\lambda }{\mathrm{X}}_{{\mathrm{C}}_3{\mathrm{H}}_8}\right)$$

(34)

$${\mathrm{r}}_3={\displaystyle \frac{{\mathrm{k}}_3{\mathrm{K}}_{{\mathrm{C}}_3{\mathrm{H}}_6}{\mathrm{P}}_{{\mathrm{C}}_3{\mathrm{H}}_6}}{1+{\mathrm{K}}_{{\mathrm{C}}_3{\mathrm{H}}_8}{\mathrm{P}}_{{\mathrm{C}}_3{\mathrm{H}}_8}+{\mathrm{K}}_{{\mathrm{C}}_3{\mathrm{H}}_6}{\mathrm{P}}_{{\mathrm{C}}_3{\mathrm{H}}_6}+{\mathrm{K}}_{{\mathrm{CO}}_{\mathrm{x}}}{\mathrm{P}}_{{\mathrm{CO}}_{\mathrm{x}}}}} \times \exp \left(-\mathrm{\lambda }{\mathrm{X}}_{{\mathrm{C}}_3{\mathrm{H}}_8}\right)$$

(35)

Furthermore, both the kinetic and adsorption parameters (rate constants k_i and adsorption constants K_i) together with the activation energies are obtained by using Arrhenius equations, as shown in Equations (36) and (37). In these equations, ${\mathrm{k}}_{\mathrm{i}}^0$ and ${\mathrm{K}}_{\mathrm{i}}^0$ are the reaction and adsorption constant pre-exponential factors respectively, E_i is the activation energy, ${\Delta \mathrm{H}}_{\mathrm{i}}$ is the heat of adsorption, R is the universal gas constant, and T_m is the median temperature. These mathematical expressions are particularly advantageous because they are centered around a reference or median temperature T_m used to reduce the cross-correlation between activation energies and pre-exponential factors [19,31]. In this way, the overall reactor behavior is represented by a set of first-order differential mass–balance equations with the parameters ${\mathrm{k}}_{\mathrm{i}}^0$, ${\mathrm{K}}_{\mathrm{i}}^0$, E_i, and ${\Delta \mathrm{H}}_{\mathrm{i}}$ having to be estimated.

$${\mathrm{k}}_{\mathrm{i}}={ \mathrm{k}}_{\mathrm{i}}^0\exp \left[{\displaystyle \frac{-{\mathrm{E}}_{\mathrm{i}}}{\mathrm{R}}}\left({\displaystyle \frac{1}{\mathrm{T}}}-{\displaystyle \frac{1}{{\mathrm{T}}_{\mathrm{m}}}}\right)\right]$$

(36)

$${\mathrm{K}}_{\mathrm{i}}={\mathrm{K}}_{\mathrm{i}}^0\exp \left[{\displaystyle \frac{-{\Delta \mathrm{H}}_{\mathrm{i}}}{\mathrm{R}}}\left({\displaystyle \frac{1}{\mathrm{T}}}-{\displaystyle \frac{1}{{\mathrm{T}}_{\mathrm{m}}}}\right)\right]$$

(37)

Hossain [20] and Al-Ghamdi et al. [31] estimated adsorption constants and kinetic parameters simultaneously within a single regression framework. In contrast, Rostom et al. [19] highlighted that the non-linear nature of the rate equations renders the parameter estimation procedure susceptible to over-parameterization. To address this issue, Rostom et al. [19] reduced the number of independent fitting parameters in the ODH model by determining the adsorption constants prior to kinetic model fitting, and by using separate adsorption isotherm experiments conducted with the CREC Riser Simulator. Furthermore, Rostom et al. [19] reported negligible coke formation and, consequently, set the decay constant λ to zero.

Once the system of kinetic model equations was established, the kinetic model was solved numerically by using appropriate differential equations solvers, data-fitting techniques, and statistical and physical criteria to validate the various chemical species partial pressures. This was accomplished when the sum of squared residuals (SSQ or SSR) between the predicted and experimental species partial pressures was minimized. Additional indicators of good fit included a normal distribution of residuals, high correlation coefficients, low cross-correlation among parameters, and reduced confidence intervals [19,20,44,66].

By following this procedure, Hossain [20] demonstrated that replacing a pure CaO support with a mixed CaO–γ-Al₂O₃ support reduced the activation energy required for propylene formation from 126.7 to 120.3 kJ·mol⁻¹, while increasing the activation energy associated with undesired CO₂ formation from 32.8 to 55.1 kJ·mol⁻¹. The activation energy required for the secondary combustion of propylene remained approximately constant at 53.7 kJ·mol⁻¹. Similarly, Al-Ghamdi et al. [31] reported activation energies of up to 124.9 kJ·mol⁻¹ for propylene formation, along with values of 52.8 kJ·mol⁻¹ and 53.8 kJ·mol⁻¹ for the combustion of propane and propylene, respectively, when using a VO_x/γ-Al₂O₃ catalyst.

In contrast, Rostom et al. [19] reported lower activation energies when employing a VO_x/ZrO₂–γ-Al₂O₃ catalyst, with values of 55.7 kJ·mol⁻¹ required for propylene formation and 33.3 kJ·mol⁻¹ being needed for propane combustion. Despite the relatively low activation energy required for combustion, the catalyst maintained a high propylene selectivity, which was attributed to a larger pre-exponential factor k⁰. The activation energy for secondary propylene combustion was 98.5 kJ·mol⁻¹, thereby limiting consecutive oxidation reactions. Additionally, it was confirmed that the cross-correlation coefficients among the estimated parameters were well below unity, indicating that the assigned parameter values were only weakly affected by one another (see

Table 4. Reported kinetic parameter for FOx–PCODH adapted from Rostom and de Lasa [19].

Reaction			${\mathbf{k}}_{\mathbf{i}}^0$		Ei
Propylene formation i=1			2.82 × 10−5 ± 1.15 × 10−6		55.7 ± 7.58
Propane combustion i=2			1.65 × 10−6 ± 1.02 × 10−7		33.3 ± 7.58
Secondary propylene combustion i=3			4.80 × 10−6 ± 2.29 × 10−6		98.5 ± 15.56
Correlation Matrix	${\mathrm{k}}_1^0$	${\mathrm{k}}_2^0$	${\mathrm{k}}_3^0$	E1	E2	E3
${\mathrm{k}}_1^0$	1
${\mathrm{k}}_2^0$	−0.84	1
${\mathrm{k}}_3^0$	0.83	−0.94	1
E1	−0.21	0.04	−0.20	1
E2	−0.03	0.07	0.13	−0.68	1
E3	0.52	−0.55	0.75	−0.59	0.70	1

).

Beyond ODH reaction kinetics, kinetic modeling has also benefited from the evaluation of catalyst acidity. Ayandiran et al. [6] investigated ammonia temperature-programmed desorption (NH₃-TPD) using the Cvetanovic and Amenomiya model. This model showed good agreement with experimental data, with correlation coefficients being close to unity and residuals being normally distributed. Their results indicated that increasing the CaO loading led to an increase in the activation energy of desorption, from 39.2 kJ/mol to 96.3 kJ/mol, suggesting stronger interactions between the active sites and the support.

7.2. FOx–BCODH

Early kinetic descriptions of BODH were primarily based on redox mechanisms, such as those of the MvK model. More recent studies, however, have adopted the L–H model to describe reactions under oxygen-free gas-phase conditions. Similar to the FOx–PCODH kinetic models, recent L–H formulations for FOx–BCODH incorporate kinetic and adsorption parameters (k_i and K_i), with Arrhenius-type expressions having temperature-centering approximations. These estimations are used to describe temperature dependence, and exponential decay functions to account for catalyst deactivation due to lattice oxygen depletion [44,66].

Both the KFUPM-Saudi Arabia research group and the CREC-UWO group have been leading kinetic modeling research using the CREC Riser Simulator. Lucky et al. [66] formulated two alternative Langmuir–Hinshelwood (L–H) models for a VO_x/CeO₂–γ-Al₂O₃ catalyst, based on the five-step reaction network, previously shown in Figure 4. In Model I, a single type of adsorption site (“X”) was assumed, with all sites having equivalent activity and the product selectivity was limited by the accessibility of the reactants to the lattice oxygen. In Model II, two types of adsorption sites (“X” and “Z”) were considered, where the mildly acidic X sites promoted ODH to C₄-olefins, while the more strongly acidic Z sites favored cracking into lighter hydrocarbons. These results indicated that the single-site adsorption model provided the best fit to the experimental data. Accordingly, authors conducted a simultaneous regression of both the intrinsic kinetic parameters and the adsorption constants, resulting in a total of 17 parameters being calculated.

In contrast, a recent approach proposed by Bin Sulayman and de Lasa [44] was developed, in which a L–H model using a VO_x/MgO–γ-Al₂O₃ catalyst was based on a simplified triangular parallel–series reaction network comprised of only three primary reactions, as illustrated in Figure 7. Light cracking products (C₁–C₃ hydrocarbons) were excluded from the kinetic model due to their minimal yields (<2%). Additionally, Bin Sulayman and de Lasa set the decay parameter λ to zero, as coke formation was negligible (<0.02 wt%). To further simplify parameter estimation, the adsorption constants were determined through separate adsorption experiments conducted in the CREC Riser Simulator reactor, following the approach previously used by Rostom et al. [19] for FOx–PCODH. As a result, the number of parameters to be estimated was reduced to six, thereby significantly reducing model complexity.

When comparing the CREC Riser Simulator results, it is worth mentioning that both CREC-UWO and KFUPM groups reported very similar activation energies for the primary formation of C₄-olefins (90.2 kJ/mol [66] and 89 kJ/mol, respectively [44]). In contrast, for the primary combustion of n-butane and for the secondary combustion of C₄-olefins, Lucky et al. [66] estimated activation energies of 121.6 kJ/mol and 81.0 kJ/mol, respectively, whereas Bin Sulayman and de Lasa [44] showed significantly lower values of 42 kJ/mol and 44 kJ/mol, respectively. These lower values suggest that the corresponding reactions are controlled by relatively small pre-exponential factors rather than by high activation energy barriers. Overall, the triangular L–H model proposed by Bin Sulayman and de Lasa demonstrated good agreement with experimental data. The intrinsic parameters were estimated under a high number of degrees of freedom (>420), with narrow 95% confidence intervals and cross-correlation coefficients below 0.89.

8. Coke Formation as a Challenge for CODH Large Scale Implementation

Coke formation, or carbon deposition on catalysts, is presently a major limitation in conventional olefin production processes. The adoption of the FOx–CODH technology offers a promising pathway to mitigate this issue. As discussed in earlier sections, FOx–CODH utilizes lattice oxygen to facilitate hydrogen removal and enable water formation via an exothermic reaction that is thermodynamically favored at relatively low temperatures (approximately 450–600 °C). Operating within this lower temperature range significantly suppresses thermal cracking pathways, which are the primary contributors to coke formation, thereby enhancing catalyst stability and extending process longevity [1,6,39,46].

Recent studies have consistently demonstrated that under optimal conditions, coke formation can have a minimal effect on FOx–CODH activity. Coke can therefore be neglected as a deactivation factor, as discussed in Section 7 [19,44,63]. Similarly, experiments using a VO_x/γ-Al₂O₃ catalyst at 550 °C, with ten 10 ml consecutive propane injections, led to 1–1.5 wt% carbon deposition only [53]. Likewise, studies evaluating VO_x/TiO₂ and W/Re-doped VO_x catalysts at 500 °C reported carbon mass balances between 96% and 98%, indicating minimal carbon loss in the form of coke [47,57]. On this basis, it is speculated that in the FOx–CODH reaction, when there is an excess of lattice oxygen, coke forms primarily on the support matrix (e.g., γ-alumina) rather than on the active sites.

Furthermore, coke formation can also occur when the lattice oxygen is completely depleted. Under these conditions, non-oxidative dehydrogenation and cracking reactions can occur [39,45,56]. As a result, coke formation can lead to losses of valuable light paraffin feedstock as coke and subsequently to both the unwanted consumption of light paraffins and the severe deactivation of the catalyst. However, if the FOx–CODH reactor is operated under slight excess oxygen conditions, while considering the oxygen stochiometric requirements, coke formation can be mitigated by modifying or replacing acidic supports. For instance, fluidizable γ-Al₂O₃ with a strong acidic character can be substituted with lower-acidity catalyst support materials, such as fluidizable silica [6,19,59]. With this approach, one can expect minimal coke formation, negligible catalyst deactivation, and improved overall FOx–CODH catalytic performance with the entire light paraffin feedstock being employed for CODH, as discussed in Section 3 and Section 4 of this review.

9. Future Perspectives for FOx–CODH

To overcome the challenges associated with maintaining sufficient lattice oxygen for the reaction in a continuous unit at an industrial scale, researchers have proposed the use of a twin circulating fluidized bed reactor system for FOx–CODH processes [1,6]. In this configuration, as illustrated in Figure 8, catalyst particles continuously circulate between a downer reactor unit, where the FOx–CODH reaction takes place, and a separate dense fluidized bed regenerator used for catalyst reoxidation. Within the regenerator, the partially reduced catalyst is reoxidized through exposure to air, replenishing the lattice oxygen, and combusting any coke that may have been formed. Additionally, multiple cyclone separators are required at the outlets of both the downer and the regenerator to separate product gases from catalyst particles. Given that the reaction and regeneration stages are spatially separated, each can be optimized independently, allowing for improved process control and overall system performance. Furthermore, the twin fluidized bed configuration is also equipped with a catalyst cooler, allowing the CODH catalyst to be thermally reconditioned (temperature reduced) prior to being fed back to the CODH downer unit.

By using the Computational Particle Fluid Dynamics (CPFD) Barracuda software, Rostom et al. [61] proposed an industrial downer reactor, with a length of 20 m, with particle cluster velocities between 2.8 and 3.5 m/s, and with gas–catalyst contact times between 5 and 7 s. It was demonstrated that this type of simulation plays an essential role in the scale-up and optimization of industrial reactor configurations. This is the case given that CPFD provides valuable information regarding hydrodynamics, gas–solid flow behavior, and reaction kinetics.

Thermal management is another critical consideration, as both the FOx–CODH reaction and catalyst regeneration are exothermic. CPFD simulations predict temperature increases of 11–13 °C along the downer reactor [61]. To maintain stable continuous operation, a fraction of the hot catalyst flow exiting the ODH reactor, as described in Figure 8, is partially cooled during its transport back to the reactor inlet, thereby helping to regulate the system temperature [31].

Furthermore, as observed in the studies discussed, CODH has typically been investigated using pure ethane, propane, or butane as feedstocks. However, in industrial practice, these paraffins are commonly present as mixtures, such as in natural gas, liquefied natural gas (LNG), and shale gas streams rich in light alkanes (methane, ethane, propane, and butane) [18,28,70]. Consequently, the disparity between experimental conditions and practical industrial feed compositions highlights the need for further research on the direct synthesis of olefins from mixed paraffin feeds.

Investigating CODH using mixed-component feeds (e.g., ethane–propane mixtures) could also provide valuable insights into catalyst behavior, including competitive adsorption, selectivity patterns, and potential inhibitory interactions between compounds. Therefore, using advanced experimental tools, such as the CREC Riser Simulator to evaluate the performance of fluidizable vanadium-based catalysts, while employing multi-component feeds that more closely resemble real-world conditions, particularly oxygen-free environments, represents a promising direction for future research.

10. Conclusions

• Vanadium-based catalysts supported on γ-Al₂O₃, ZrO₂, SiO₂, or mixed oxides have demonstrated to be valuable for FOx–CODH performance. A combination of vanadium with promoters such as Mo, Re, W, or CaO improves catalyst stability, redox behavior, and dispersion.
• Sustained paraffin conversion suggests that oxygen sources other than lattice oxygen participate in FOx–CODH. Further research is needed to identify the oxygen species involved in the reaction.
• Conversion and selectivity alone are not sufficient indicators to evaluate catalyst performance. A more reliable metric is the overall olefin formation rate, since catalysts with moderate conversion and selectivity may still be more efficient if they operate at shorter reaction times or are present in the reactor in smaller amounts.
• The experimental evaluation of CODH vanadium-based fluidizable catalysts under free oxygen conditions (FOx–CODH), particularly when employing the CREC Riser Simulator, shows excellent 14–21 mmol C₄/(gcat h) overall olefin formation rates. These experimental assessments also lead to trustworthy kinetic models with accurate parameters with reduced spans and cross-correlation coefficients.
• The numerical simulation of fluidizable FOx–CODH catalysts in a continuous downer unit obtained by using the CPFD Barracuda software allows one to establish the expected light paraffin conversion, the olefin selectivity and the overall process productivity together with the removal of the heat of reaction. This minimizes undesirable reactions such as cracking and coke formation.
• The numerical simulation of the entire CODH process, including catalyst circulation and regeneration, obtained by using the CPFD Barracuda software allows one to predict the conditions required for FOx–CODH process scale-up and for the stable operation of an industrial unit. This prevents undesirable coke accumulation as well as promoting adequate CODH catalyst re-oxygenation.
• The suggested process integration, including a downer reactor and a dense phase fluidized bed, provides insights on how the FOx–CODH technology could be implemented at the industrial scale, which still nowadays remains a significant challenge.