What is the difference between dehydrogenation and oxidation

Intermediates in light color and transition states connected with dashed lines are energetically disfavored. Figure 6. Full symbols denote propene production, and empty symbols denote CO 2 production. Inset energies correspond to the activation barriers. Figure 7. Full symbols denote propene production and empty symbols denote CO 2 production. Inset values denote the slopes, which correspond to the reaction order with respect to the oxidant. Figure 8.

Lines are guides for the eye. Figure 9. Inset values denote the slopes, which correspond to the reaction order with respect to propane. Figure Top Selectivity toward propene left and CO 2 right , bottom catalyst activity left and propene yield right.

The oxidant used is O 2 top or N 2 O bottom. Dashed lines represent the cumulative number of molecules produced propene, CO 2 or consumed C 3 H 8 per active site and use the right axis. The contribution of the European Commission by funding a Horizon project BiZeolCat through grant agreement is greatly appreciated. More by David Bajec. Cite this: ACS Catal.

Published by American Chemical Society. Article Views Altmetric -. Abstract High Resolution Image. Short-chained olefins, such as propene propylene and butadiene, are important precursor chemicals in the production of plastics, synthetic rubbers, copolymers, epoxides, various organic acids, acrylonitrile, nylon, and so forth. They are predominantly extracted from higher hydrocarbons during steam cracking and fluid catalytic cracking. Alternatively, dehydrogenation can be performed with oxidants, such as air, oxygen, N 2 O, or CO 2 , releasing hydrogen as water in a strongly exothermic reaction.

To steer the reaction toward propene and avoid extensive cracking non-oxidative or oxidation oxidative , appropriate catalysts must be used. Rather than consisting of semi- noble metals, they are made of inexpensive source materials.

Both processes suffer from persistent coking of the catalyst, which must be regenerated or changed often, negatively impacting the catalyst longevity and the process economics.

The use of chromia dates back to the s, 17 while the first commercial technology was the Pacol process from , using alumina-supported platinum catalysts. Working on chromia-based catalysts, Suzuki and Kaneko already in proposed a macrokinetic model. Nijhuis et al. Purely first-principles theoretical descriptions of chromia-based catalysts for propane dehydrogenation remain scarce. These are used as the two extrema of the realistic catalyst surface.

Using extensive KMC modeling, we investigated the kinetic parameters for propane dehydrogenation on each of the surfaces temperature and pressure dependence , the effect of varying oxidation state modeled as a varying ratio between the surfaces , and the effect of the oxidant used none, O 2 , N 2 O, or CO 2.

We identified the rate-determining steps, the simplified reaction rate law descriptions, and the side products formed. We show that the oxidized and reduced surfaces behave radically different and how the oxidants influence the selectivity and activity. Although both surfaces bind saturated hydrocarbons weakly, double and triple bond-containing hydrocarbons are strongly adsorbed on the oxidized surface.

As a consequence, the latter exhibit a much greater activity but poor selectivity as CO 2 is mostly produced. Similarly, using O 2 as a strong oxidant increases the activity and suppresses the selectivity in comparison to using N 2 O. Without the oxidant, the oxidized surface is eventually reduced. Using multiscale modeling, we demonstrate a Goldilocks effect.

For optimum conversion of propane to propene, the surface should be partially oxidized. Lastly, we show how the catalyst activity decreases due to coking. This is, to the best of our knowledge, the first multiscale study of an industrially relevant catalyst in realistic conditions for propane dehydrogenation.

Moreover, since the transition between the oxidized and reduced sites is included, this is effectively a model where the catalyst changes during the reaction. Computational Details. The simulation parameters chosen were consistent with our previous work 28,29 for comparability.

The plane waves were expanded to the energy cut-off of eV. The Grimme D3 correction was used to describe the dispersion interactions. Geometry relaxations were performed with a force threshold of 0. The surface was modeled with 12 layers, of which the bottom six were immovable in their bulk positions. The dipole correction was used because the slabs were asymmetrical. Among the adsorption reactions, we distinguish simple non-activated adsorptions for instance, hydrocarbons and activated dissociative adsorptions hydrogen, CO 2 , N 2 O, O 2 , etc.

The former is a purely kinetic event eq 3 , while the latter is an ER reaction eq 2. Furthermore, k B denotes the Boltzmann constant and h the Planck constant.

It is decomposed into the electronic interaction, E int negative , and the distortion energies of the adsorbate, E dis , and surface, E surf,dis both positive. All energies are zero-point energy-corrected. We model the catalyst as the surface of Cr 2 O 3 with 12 layers. To prevent lateral interactions of adsorbates across the adjoining cells, a supercell is used. Wang and Smith 50 performed extensive first-principles simulations of this surface and constructed a phase diagram.

They showed that five surface phases can exist between two extrema segregated Cr atoms and condensed oxygen. Essentially, these different terminations are congruent with different oxidation states of the chromium atoms at the surface. Several experimental studies have shown that such descriptions of the surface are appropriate. Maurice et al. However, the situation during the reaction, where oxidants O 2 and N 2 O and reducing species hydrocarbons are present, is more complicated.

First, these conditions are close to a phase transition and a small increase in the temperature or decrease of the oxygen pressure would render the D surface more stable. Second, the accuracy of DFT calculations is limited in terms of chemical accuracy, meaning that surface diagrams can easily be shifted for — K and several factors for pressure. Third, catalysts are most active near phase transitions. Fourth, while oxygen is an oxidant, propane is a moderate reducing agent. Lastly, in this work, we study the reaction in both, oxidative and non-oxidative regimes, where O 2 strong oxidant , N 2 O, and CO 2 soft oxidant of various concentrations or, alternatively, no oxidant are used.

This would necessitate taking into account all five surface structures and their interconversion on-the-fly, which adds too many layers of complexity to a simple model. Noting that the B, C, and D surfaces are essentially the E surface of different surface oxygen in the form of chromyl coverages, we construct a simplified model as an alternative.

In our model, we use the A surface, which was already used in our previous studies, 28,29 and the E surface in varying fractions. The model has three types of active sites: chromium atoms on A Cr red , oxygen atoms on A O red , and oxygen atoms on E O ox. The E surface can lose the surface oxygen as O 2 or, more realistically, H 2 O, being converted to A. As depicted in Figure 1 , the lattice has a quasi-hexagonal symmetry, where each chromium atom is connected to six nearest chromium atoms and three nearest oxygen atoms and vice-versa.

High Resolution Image. The kinetic analysis was performed as KMC simulations. Kinetic and thermodynamics parameters, as obtained from the DFT data, were used in the KMC model to probe the reaction at various temperatures, reactant concentrations effectively pressures , and catalyst compositions. The simulations were performed using Zacros, which is a graph-theoretical implementation of the KMC approach. In this approach, the Hamiltonian is calculated within the energetic model, accounting for the number of adsorbed clusters on the lattice and their interactions.

The simulations were carried out on a quasi-hexagonal lattice, as shown in Figure 1 , which is commensurate with the lattice of the DFT model. There are two types of surface sites: one corresponding to the exposed chromium atoms and the other to oxygen atoms. Oxygen atoms are linked to six oxygen atoms in the hexagonal arrangement and the nearest three chromium atoms.

In total, each surface site has a connectivity of nine Figure 2. To account for the effect of surface oxidation, two different surfaces are investigated: reduced A and oxidized E. The underlying lattice is the same, but the chromium sites are not exposed on the oxidized surface.

Thus, three active site types are considered O ox , Cr red , and O red , while the Cr ox sites in the KMC are considered inert having no physical counterpart, they are only included to construct the lattice more easily. The reaction constants and, consequently, the reaction mechanism are different on the two investigated surfaces and were individually determined by separate DFT calculations. The simulations were run with 19 different seeds and then averaged.

Adsorption and diffusion reactions were treated as fast equilibrated events stiffness-scaled. The model is checked to be thermodynamically consistent. All reactions paths on the catalyst on the oxidized or reduced surface yield the same reaction energy as is the energy difference between the products and reactants in the gaseous phase. In the kinetic model, two sets of adsorption energies and reaction barriers are available, depending on the oxidation state of the active site.

Results and Discussion. Saturated hydrocarbons, of which methane, ethane, and propane were included in the model, interact with the surfaces merely through weak and non-specific van der Waals interactions. This results in high barriers for their activation see the section Reaction Mechanism and Table 4 , low surface coverages during the reaction, and negligible surface perturbation.

As summarized in Table 1 , the adsorption is weaker on the oxidized surface 0. Molecular hydrogen does not bind to the surface.

The attractive interaction is wholly due to electronic effects being weaker on the oxidized surface due to the higher electron density on the surface, repelling the saturated hydrocarbons. Effectively, the oxidized surface exhibits an acidic character with exposed oxygen atoms, which readily take on hydrogen atoms.

Table 1. Hydrocarbons with multiple bonds propene, propyne, ethene, and ethyne react with the two surfaces differently. The latter interaction is approximately 1 order of magnitude stronger, accompanied by the strong electronic interaction and geometric effect. Such strong interactions have a profound effect on the reaction selectivity as the intermediates do not readily desorb but instead undergo further dehydrogenation or cracking, as shown later on.

KMC simulations offer insight into the behavior of the catalyst structure with an atomistic resolution provided sufficient input data are available.

Lateral interactions are key to transcending a mean-field description. This was shown to be sufficient in our previous work, 28 while including all possible lateral interactions is impractical due to the sheer number of combinatorial possibilities. With a few exceptions, these interactions are weaker on the oxidized surface, where they are also generally repulsive.

Table 2. On the reduced surface, hydrogen atoms bound to oxygen atoms can only recombine into H 2 and desorb. Recombination of hydrogen atoms on two adjacent oxygen surface atoms yields chemisorbed water, which can desorb. The ensuing oxygen vacancy can migrate across the surface with a kinetic barrier of 0.

This migration is limited to the oxidized part of the catalyst this is relevant only in the mixed composition. If all surface oxygen atoms are lost, the oxidized surface E is equivalent to fully reduced A. The oxygen vacancy can be replenished by N 2 O in an exothermic reaction, yielding N 2 and the fully oxidized surface.

The use of CO 2 , however, is calculated to be less effective due to a high barrier and strong endothermicity. When two adjacent oxygen vacancies form, they are easily filled by dissociative adsorption of O 2.

This reaction is two-step. First, O 2 strongly adsorbs near the vacancies and then it dissociates. These reactions, effectively enabling a transformation between A and E, are summarized in Table 3 and included in the kinetic model. See Figure 3 for the structures involved. Table 3. The surfaces A and E represent two extrema. The reaction mechanism for propane dehydrogenation, although consisting of 74 individual steps, is composed of only four types of reactions adsorption, ER reactions, diffusion, and Langmuir—Hinshelwood reactions.

The steps with the corresponding barriers and reaction energies are listed in Table 4. Propane, propene, propyne, ethane, ethene, ethyne, and methane can adsorb vide supra , while molecular hydrogen interacts weakly and non-specifically with either surface steps 8— The carbon species either bind too strongly all unstable intermediates and multiple bond-containing species on the oxidized surface , rendering them immobile, or too weakly, making diffusion comparable with desorption.

Table 4. As shown in Figure 4 , we systematically include all possible dehydrogenation steps, where a hydrogen atom is removed from the adsorbate. As it takes two steps two hydrogen atoms must be removed to convert a single bond to a double bond, dehydrogenation reactions generally link a stable hydrocarbon and a monoradical either as a reactant or as a product.

Deep dehydrogenation reactions violate this rule, yielding multiple radicals. While they are generally unlikely on the reduced surface, the oxidized surface is so active that deep dehydrogenations readily proceed, occasionally surpassing normal dehydrogenation routes for instance, CH 3 CH 2 is preferentially deep-dehydrogenated to CH 3 CH instead of forming ethene. Intramolecular hydrogen migrations are omitted because their activation barriers exceed those of the dehydrogenation reactions.

Upon a weak physisorption of propane 0. The potential energy surface of the reaction steps is depicted in Figure 5. In a nutshell, dehydrogenations are exothermic and kinetically more accessible on the oxidized surface and endothermic with higher barriers on the reduced surface.

Although C2 hydrocarbons enter the reaction as partially dehydrogenated species formed in various cracking reactions of the C3 intermediates, we model the entire pathway.

Consequently, very little ethene is formed. The increased reactivity of the oxidized surface is mirrored in much greater cracking activity as well. While on the reduced surface, most cracking reactions have high barriers above 2.

As it will be shown later on, this greater activity of the oxidized surface manifests in both higher TOFs for the production of olefins and increased cracking, causing the formation of C2 and C1 products, and coking. The coke formed is usually burnt away with cycles of excess oxygen as CO 2.

Using no oxidant, the products of propane dehydrogenation are propene and hydrogen and the reaction is endothermic. When an oxidant is used, propene, CO 2 , side products propyne, ethene, etc. The true apparent activation energies are shown in Figure 6.

For the production of propene, this value is 1. The larger value for O 2 does not imply that the reaction proceeds slower, which is clearly shown in Figure 6. Instead, the larger value reflects a great temperature dependence, while the overall TOF is still larger. On the reduced surface Figure 6 , the apparent activation barrier is 1.

As shown in Figure 6 , the behavior of this surface is, as expected, between the both extrema. The apparent activation barrier for propene production is 1. When using O 2 , CO 2 is still the main product, while with N 2 O, propene begins to predominate at lower temperatures. A simple mathematical reasoning shows that the reactions with higher activation energies start to predominate at higher temperatures. However, it is known experimentally that at higher temperatures, hydrocarbons will convert to CO 2 in oxidative conditions.

This apparent inconsistency with our model is reconciled as follows. At higher temperatures, combustion proceeds homogeneously, while we are interested only in the performance of the catalyst, that is, surface reactions. Although the production of propene increases with temperature, this is in real life offset by homogeneous combustion.

Second, real-life scenarios deal with significant catalyst deterioration at higher temperatures due to deactivation, coking, sintering, and so forth. While we include coking vide infra , other transformations of the catalyst have not been included in the model because we are interested in the performance of the catalyst in its pristine reduced and oxidized forms and not in the description of the process per se.

The effect of the oxidant used and its pressure on the reaction is shown in Figure 7. At 1 bar of propane and K, the partial pressure of O 2 and N 2 O was varied from 10 —3 to 40 bar over the oxidized and reduced surfaces. On the former, increasing the partial pressure of the oxidant has little effect on the selectivity. The reaction order with respect to O 2 and N 2 O is 0.

On the mixed surface, the effect is different. Since CO 2 production proceeds only on the oxidized surface, it is not surprising that the reaction order remains virtually unchanged.

For the production of propene, the reaction order drops to 0. In these more realistic conditions, the effect of oxidant pressure on the selectivity is predicted to have an important effect. This has clear implications for selectivity, which is shown in Figure 8.

On the mixed surface, the effect is pronounced. Whereas at lower oxidant pressures, the production of propene predominates, the selectivity precipitously drops at higher pressures.

With oxygen, above 0. On the oxidized surface, the reaction order of propene production with respect to propane is 0. On the reduced surface, the reaction order with respect to propane is close to unity, consistent with our previous work. For the production of CO 2 , the pressure of propane is irrelevant when using O 2 or N 2 O as oxidants without the oxidant, there is a positive trend with very low absolute TOF, which is attributed to the loss of surface oxygen atoms.

For the production of propene, the reaction orders with respect to propane are 0. Higher propane pressures are thus expected to be advantageous for the production of propene. In Figure 10 , we show a stark difference between the performances of the three surfaces at different propane pressures.

On the oxidized surface, the selectivity toward propene remains poor and only marginally improves even when the propane pressure shoots up. On the reduced surface, propane is almost exclusively produced. In all instances, N 2 O as a soft oxidant exhibits better selectivity, which is expected.

The analysis so far has focused on the effects of temperature and reactant pressure, which were tested on three arbitrary surface oxidation levels: fully oxidized, fully reduced, and mixed While the oxidized surface exhibits high activity and low selectivity, the reduced surface has the exact opposite properties. Thus, a closer look into the optimum surface condition is warranted.

We construct a series of mixed surfaces, as shown in Figure 1. As expected, the selectivity toward propene drops as the degree of oxidation increases. The drop is more precipitous when using O 2 as opposed to NO 2. The selectivity toward CO 2 displays a converse pattern because no other products are produced in any meaningful amounts; temperatures in excess of K were found to be required for the production of propyne and C 2 products.

However, the overall catalyst activity increases with the temperature. Thus, a maximum in propene yield was expected at some intermediate degree of oxidation. When using O 2 , the optimum degree of oxidation seems to be around 0. For better insight into the catalyst performance, we study the catalyst surface, whose temporal evolution is shown in Figure The rationale is as follows.

The empty oxidized sites are readily available for the reaction. The surface oxygen vacancies actually correspond to the reduced catalyst surface, which is much less active vide supra and can be used as a surrogate for reversible catalyst inhibition. Different carbon species, however, constitute irreversible deactivation of the catalyst.

While C 3 and most C 2 species adsorb reversibly since they can undergo further dehydrogenation, the buildup of C 1 is indicative of coking. The C 1 concentration steadily increases with time but can be burnt away with oxygen, regenerating the catalyst in our model, we do not account for structural changes of the catalyst such treatment might cause. For comparison, the cumulative production of propene and CO 2 and the consumption of propane is also shown on the secondary axes.

We observe a few clear trends. First, the fraction of empty oxidized sites decreases with temperature and is higher under O 2 than under N 2 O, where it drops to zero. Conversely, the fraction of empty reduced sites increases and approaches unity at higher temperatures.

In most instances, this is the predominant surface motif. These results explain the observed kinetic behavior of the catalyst under different conditions. It is clear that under N 2 O, the surface is more reduced, which is beneficial for the selectivity toward propene.

Moreover, we see that initially only CO 2 is produced red dashed line and only when the catalyst is sufficiently reduced is propene also formed blue dashed line. Catalyst deactivation is a common problem with oxidation reactions. As coke forms on the catalyst surface, the activity drops. Deactivation is a complex topic, which a first-principles kinetic model cannot fully reproduce. We make no attempt at describing the structural deterioration of the surface, such as sintering, phase transitions, or decomposition.

We define deactivation as the buildup of dead-end carbon species, which cannot further dehydrogenate even upon C—C cracking.

This might very well be an artifact of the model as both species describe the coked surface. Thus, we treat both species cumulatively in our analysis. We show that the deactivation of the catalyst can be described with the Arrhenius kinetics.

The apparent activation barrier and pre-exponential factor for the deactivation are dependent on the oxidant used: 1. A much stronger temperature dependence when using a stronger oxidant is an expected result.

This shows that the oxidized surface deactivates faster than the reduced surface when O 2 is used as the oxidant. Soft oxidants, such as N 2 O, cause a much slower catalyst deactivation.

In Figure 13 , we show the catalyst surface after 9. When the catalyst is fully coked Figure 13 , it can be recovered by burning away the coke as CO 2.

N 2 O and CO 2 are too weak oxidants for surface recovery. In real operation, cycles of catalyst coking and regeneration repeat with an hourly cadence because the catalyst is not fully oxidized.

As shown previously, the reduced surface gets coked much slower. It is difficult and often not very informative to try applying idealized models aimed at understanding the fundamentals in simplified conditions to real-life scenarios.

Assumptions pristine crystal lattice devoid of defects and phase boundaries, constant pressure and temperature, suppression of unwanted side reactions, reactor design, etc. Reactions which do involve gain or loss of one or more oxygen atoms are usually referred to as 'oxygenase' and 'reductase' reactions, and are the subject of section For the most part, when talking about redox reactions in organic chemistry we are dealing with a small set of very recognizable functional group transformations.

It is therefore very worthwhile to become familiar with the idea of 'oxidation states' as applied to organic functional groups. By comparing the relative number of bonds to hydrogen atoms, we can order the familiar functional groups according to oxidation state. We'll take a series of single carbon compounds as an example. Methane, with four carbon-hydrogen bonds, is highly reduced. Next in the series is methanol one less carbon-hydrogen bond, one more carbon-oxygen bond , followed by formaldehyde, formate, and finally carbon dioxide at the highly oxidized end of the group.

This pattern holds true for the relevant functional groups on organic molecules with two or more carbon atoms:. Notice that in the series of two-carbon compounds above, ethanol and ethene are considered to be in the same oxidation state. You know already that alcohols and alkenes are interconverted by way of addition or elimination of water section When an alcohol is dehydrated to form an alkene, one of the two carbons loses a C-H bond and gains a C-C bond, and thus is oxidized.

However, the other carbon loses a C-O bond and gains a C-C bond, and thus is considered to be reduced. Nothing but sales either way sale out and Counter sale. The main difference is FBI agents have federal jurisdiction within certain limits but police generally have jurisdiction only within their home city or state. Federal agents normally investigate federal crimes such as bank robbery, but do not normally investigate local jurisdiction crimes.

Disinfectants are antibacterial agents that are applied to inorganic surfaces. They should generally be distinguished from antiseptics that destroy pathogens on living tissue. Oh, in europe, unions do not pretend not to be political parties. In the USA, unions pretend not to be agents of a political party. Ketones resist oxidation by most oxidising agents, including potassium dichromate and molecular oxygen.

Log in. Elements and Compounds. Study now. See Answer. Best Answer. Study guides. Chemistry 20 cards. How does a buffer work. What happens in a neutralization reaction. What is a conjugate acid-base pair. Why is water considered to be neutral. Physics 20 cards. Which term explains whether an object's velocity has increased or decreased over time.

Which of these is a characteristic of nonmetals. What is the only factor needed to calculate change in velocity due to acceleration of gravity 9. What term is used to describe splitting a large atomic nucleus into two smaller ones. Vaping Study Guide 3 cards. Propylene Glycol.

Q: What is the difference between oxidation and dehydrogenation agents? Write your answer Related questions. How do you tell the difference between oxidation and reduction reactions?

How do Oxidation agents affect the natural and artificial color pigment? What is the difference between corrosive and caustic?