Dipole effect on ethylene epoxidation: Influence of alkali metals and chlorine

doi:10.1016/j.jcat.2018.04.008

Journal of Catalysis

Volume 363, July 2018, Pages 18-25

https://doi.org/10.1016/j.jcat.2018.04.008 Get rights and content

Highlights

•
Alkali metals induce electric dipoles on the surface and increase the selectivity.
•
Alkali metals greatly increase the activity by facilitating the oxygen dissociation.
•
Chlorine, surface and subsurface, decreases the activity but markedly increases the selectivity.
•
Interplay of chlorine and cesium is important for the optimal performance.

Abstract

Ethylene epoxidation is one of the most important selective oxidation reactions in industry and is almost invariably carried out on silver catalysts. Empirically it is known that moderate selectivity of silver can be greatly enhanced with cesium and chlorine dopants, whereas a mechanistic understanding of the underlying reason remains elusive. In this work, we present a first-principles explanation of the dipole effect on reaction selectivity, which could, in principle, be effected by alkali metals. We show that, on a silver surface, alkali metals work in a similar fashion thanks to their low electronegative character. The selectivity is impacted by inducing electric dipoles on the catalytic surface, whereas the electronic properties of the catalyst remain largely unperturbed. Cesium, as the least electronegative metal, shows the most pronounced effect. This effect is shown to persist when alkali metals are treated as “naked” atoms or, more realistically, when oxidized. Additionally, alkali metals increase activity of the catalyst by favoring the rate-determining step of oxygen dissociation. Chlorine, which is strongly electronegative, acts differently. Subsurface chlorine would increase selectivity through the same mechanism, while surface chlorine stabilizes alkali metals and moderately increases selectivity. Interplay of both dopants is needed to precisely tailor the catalyst for the best performance.

Graphical abstract

Introduction

Controlled oxidation of ethylene to ethylene epoxide (EO) is an important and experimentally challenging technological process. EO is a precursor to a plethora of chemicals. It is the basis for production of ethylene glycol, detergents, ethanolamines, and so forth. The principal challenge is to prevent the total oxidation of ethylene, which is thermodynamically the most favorable pathway, and instead steer the reaction towards epoxide.

Industrially, the reaction is carried out on silver catalysts, which are mostly doped with cesium and chlorine. By doing so, selectivity of pure silver catalyst, which ranges from 40% to 50%, is improved to over 90%. Despite having been thoroughly studied both experimentally [1], [2], [3], [4], [5], [6], [7], [8] and theoretically [9], [10], [11], [12], [13], [14], the promoter effect of these dopants is still not completely understood, and competing explanations (such as blocking oxygen vacant sites, influence on oxametallocycle formation, electric field, effects on surface reconstruction) have been put forth [12], [15], [16].

Silver is a commonly used catalyst. It is sufficiently mild not to break C–H bonds in the olefin but active enough to dissociate oxygen [17]. There exist other competing catalysts; for instance, pure copper is also predicted to have high selectivity [18]. Interestingly, under reaction conditions, copper-silver alloy transforms into a copper oxide-like layer on top of silver and favorably influences the reaction [19], [20]. However, under industrial reaction conditions pure metal co-exists in equilibrium with the oxide phase, where Ag₂O vastly outperforms Cu₂O and Au₂O [18].

Since the pioneering work of Linić and Barteau [9], [10], it has been well known that epoxide and total combustion pathways proceed through the common intermediate oxametallocycle (OME) [21], [22], which then converts either into EO or, when undergoing full combustion, initially into acetaldehyde (AA) and ultimately to CO₂. It is therefore understood that the difference in activation energies between their respective transition states determines the selectivity of the reaction. What is not known is how Cs and Cl influence the activation energies. Interestingly, the activity of the catalyst is determined earlier in the oxidation process. On pure silver, the rate-determining step is in fact oxygen dissociation [11]. When the catalyst gets doped and/or partially oxidized, this simple picture breaks down. As found by Greeley and Mavrikakis, oxygen might also be subsurface, where it adversely affects the selectivity [23]. On oxidized surfaces, oxide regeneration ceases to be the rate-controlling step [24]. Recently, ethylene epoxidation on silver has also been studied as a model reaction for enhancing activity with plasmonic excitations [25], [26], [27]. It has also been found that catalysts are more active near phase transitions [28].

Linić and Barteau [12] studied the effects of cesium on ethylene epoxidation on Ag(1 1 1) and tentatively argued that its promoter effect stems from the change in the difference between the EO and AA transition states owing to electric field effect. They argue that chlorine, which is strongly electronegative, might diffuse into Ag. Detailed theoretical calculations by Gava et al. cast some doubt on the possibility of interstitial subsurface chlorine [15], while a great deal of experimental data support the notion of subsurface chlorine even at lower temperatures and moderate coverages [5], [29], [30], [31]. It has been shown, however, that chlorine might mitigate the diffusion of oxygen into the subsurface [32].

Ren et al. [33] showed computationally that cesium lowers the oxygen adsorption energy on pure Ag(1 1 1), effectively making oxygen bind less strongly to the silver surface. They do not consider oxygen binding directly to the cesium. At the same time, they showed that cesium doping lowers the activation barrier of epoxide formation in the selectivity-determining step and thus increases selectivity of the catalyst.

For Ag₂O(0 0 1), Özbek et al. argue [16] that chlorine blocks oxygen vacant sites and prevents OME formation and surface reconstructions. Instead, it forces ethylene to react with surface oxygen in a one-step epoxidation [34]. Cesium acts as a co-promoter and stabilizes the non-vacant Ag₂O(0 0 1) surface. However, under industrial conditions, Ag₂O is less stable than pure Ag with a surface oxide [35]. This one-step mechanism, however, has been ruled out time and again for pure silver [36].

In this work, we employ density functional theory (DFT) calculations to shed light on the effects of dipoles on the selectivity of the reaction on Ag(1 1 1). We show how alkali metals on Ag(1 1 1) can bring about surface dipoles and increase selectivity. We show that this effect persists even in oxidizing conditions when Cs is oxidized. Activity is also increased through more favorable oxygen dissociation. The effect of chlorine is less pronounced on Ag(1 1 1), because it mostly serves to stabilize alkali metals and speed up the conversion of OME, but is more important on oxides [16].

Section snippets

Computational details

For electronic structure computations with periodic boundary conditions, the vasp package was used [37], [38], [39], [40]. The Perdew–Burke–Ernzerhof formula [41], [42] was used for evaluating the semilocal exchange and correlation contribution, while the projector augmented wave method [43], [44] was used to describe the interaction between valence electrons and the core. Criteria for relaxation were 0.01 eV/Å for stable structures and 0.03 eV/Å for transition states. The latter were

Thermodynamics

Partial oxidation of ethylene to epoxide is exothermic with an ab initio [CCSD(T)] computed reaction energy of $Δ E_{EO}^{298 K} = - 101$  kJ mol⁻¹ and a standard Gibbs free energy of $Δ G_{EO}^{298 K} = - 77$ kJ mol⁻¹ at 298 K and 1 atm. At the PBE level, which was used for slab calculations throughout this paper, this value is $Δ E_{EO}^{298 K} = - 105$ kJ mol⁻¹, which justifies its use. For further justification, see the comparison with previous work in Table 1.

Thus, successfully producing epoxide is a matter of kinetics. However,

Conclusions

We have shown how alkali metals and chlorine influence the selectivity of ethylene epoxidation on a silver catalyst, modeled as a Ag(1 1 1) surface. Through charge–electric-field interaction alkali metals stabilize the transition state that leads towards ethylene epoxidation relative to the transition state for acetaldehyde formation, which would ultimately yield CO₂. Thus, alkali metals induce a dipole on the surface and increase selectivity. Cesium is most effective because of its lowest

Acknowledgement

We acknowledge financial support from the Knut and Alice Wallenberg Foundation Project 2015.0057, and the Swedish Research Council Project 2015-03773 (A.H.). The computations were performed on resources provided by the Swedish National Infrastructure for Computing (SNIC) at C3SE (Gothenburg) and NSC (Linköping).

References (73)

C.T. Campbell et al.
Appl. Surf. Sci.
(1984)
C.T. Campbell et al.
J. Catal.
(1985)
C.T. Campbell
J. Catal.
(1986)
J.T. Jankowiak et al.
J. Catal.
(2005)
J.T. Jankowiak et al.
J. Catal.
(2005)
M. Atkins et al.
J. Catal.
(2005)
K. Waugh et al.
Catal. Today
(2010)
T.C. Rocha et al.
J. Catal.
(2014)
S. Linić et al.
J. Catal.
(2003)
A. Kokalj et al.
J. Catal.
(2008)

M. Özbek et al.

J. Catal.

(2011)

S. Linić et al.

J. Catal.

(2004)

M. Bowker et al.

Surf. Sci.

(1983)

E.R. Frank et al.

J. Catal.

(1997)

D. Ren et al.

Mol. Catal.

(2017)

G. Kresse et al.

Comput. Mater. Sci.

(1996)

G. Mills et al.

Surf. Sci.

(1995)

A. Michaelides et al.

Chem. Phys. Lett.

(2003)

A. Roldán et al.

Surf. Sci.

(2008)

S. Linić et al.

J. Am. Chem. Soc.

(2002)

S. Linić et al.

J. Am. Chem. Soc.

(2003)

S. Linić et al.

J. Am. Chem. Soc.

(2004)

P. Christopher et al.

J. Am. Chem. Soc.

(2008)

A.C. Lukaski et al.

Catal. Lett.

(2009)

P. Gava et al.

Phys. Rev. B

(2008)

M.O. Özbek et al.

ChemCatChem

(2013)

S. Piccinin et al.

J. Mater. Chem.

(2010)

M. Mavrikakis et al.

J. Phys. Chem. B

(1998)

J.W. Medlin et al.

J. Phys. Chem. B

(1999)

J. Greeley et al.

J. Phys. Chem. C

(2007)

M.-L. Bocquet et al.

J. Am. Chem. Soc.

(2005)

P. Christopher et al.

Nat. Chem.

(2011)

P. Christopher et al.

Nat. Mater.

(2012)

S. Linic et al.

Acc. Chem. Res.

(2013)

M.T. Greiner et al.

J. Am. Chem. Soc.

(2017)

A.G. Shard et al.

J. Phys. Chem. B

(2000)

Cited by (43)

Dispersed Cu (Ni, Co) in MN<inf>3</inf> moiety on graphene as active site via electrolytic water towards electro-epoxidation of ethylene
2024, Applied Surface Science
Ethylene oxide (EO) is an important petrochemical product, widely used in printing and dyeing, automotive, medicine, and other fields. Direct epoxidation of ethylene by electrocatalysis is a simple and sustainable route to produce EO, which can avoid the disadvantage of traditional industrial production, such as high temperature and CO₂ emissions. An efficient electrocatalyst could generate active oxygen intermediate during the electrolytic water process, which escape from the tremendous activation barrier of breaking molecular oxygen. In addition, ethylene and active oxygen are both required to be adsorbed or produce interaction on the same active site, which is more effective to directly transfer oxygen. Motivated by this strategy, we used first-principle to screen the possibility of 27 single transition metal atoms in MN₃ moiety as active site starting from the Gibbs free energy change (ΔG) of *O + C₂H₄ → *+ C₂H₄O. Dispersed Cu, Ni, Co can absorb both active oxygen and ethylene, and exhibits the excellent catalytic performance through stability, selectivity, thermodynamics and kinetics evaluation. Electronic structure analysis indicates that the catalytic activity roots from the high electron transfer from the absorbed ethylene to the empty M–O anti-bonding, which weaken the M–O bonding and promote the oxygen transfer. Notably, for CuN₃@graphene, the applied potential to produce active oxygen intermediate is 1.53 V vs RHE, and the kinetic energy barrier for subsequent oxygen transfer to form EO is 1.00 eV. This study verified the feasibility of single atomic electrocatalysts for EO production and offers valuable clues to design/discover active center to explore experimentally.
A systematic theoretical study the active sites of potassium promoter on the activity of water-gas shift reaction over Pt(1 1 1)
2023, Applied Surface Science
Citation Excerpt :
Since K ions is rather large compared to that of Pt (1.588 vs. 1.455 Å), its electron density caused a Pauli repulsion with the nearby Pt atoms, so it is expected that the Pt atom directly bonded to the K additive did not take part in the catalytic reaction. This is indicating that the effect of alkali metals on the adsorbate adsorption on Pt(1 1 1) is a short-ranged direct interaction through space rather than the indirect interaction through the substrate [15,49–55]. How K atoms affect water adsorption and dissociation through direct steric interactions will be given in the following section.
Alkali metals play an important role as promoters in heterogeneous catalytic reactions, but the systematic study of the nature (or active sites) of promoter is far away from enough. The possible active sites of potassium promoter such as metallic potassium (K), hydroxyl species stabilized potassium (KOH), and partially oxidized potassium (K₂O) species on the Pt-catalyzed water–gas shift reaction (WGSR) was systematically theoretical studied in this work. The calculation results indicated that the adsorption strength follows the trend K₂O/Pt(1 1 1) > KOH/Pt(1 1 1) ≈ K/Pt(1 1 1) > Pt(1 1 1), and the K₂O additive enhances the adsorption properties largest because of more K ions can be availed to bind with adsorbed species. On the other hand, the promotion effect for WGSR with the order: KOH/Pt(1 1 1〉 K₂O/Pt(1 1 1) > K/Pt(1 1 1) ≈ Pt(1 1 1), and the KOH species displays the promising promotion effect due to the facile of O–H/C–H bond cleavage as well as the C-O bond formation. Interestingly, it was shown that the high promotion effect of KOH (or K₂O) on WGSR is related to the strong activation properties of water dissociation assisted by the OH (or O) group involved in KOH (or K₂O). It might be pointed out that the metallic K additive atoms almost have no promotion effect on the Pt-catalyzed WGSR owing to the more difficult in the formation of C-O bond formation compared to that of pure Pt(1 1 1), and the active site of potassium promoter is the KOH-like species. The promotion properties of potassium on the adsorption strength of oxygen-containing species such H₂O could be ascribed to the strong short-ranged direct interaction between the K-species through space. Microkinetic modeling analysis indicated that the carboxyl mechanism is the dominant mechanism of the whole WGSR, followed by the redox mechanism and formate mechanism. It is desirable for that the present work may extend to the alkali promotion character on WGSR to other situation like metal oxide-supported metal clusters.
Optimization of ethylene dichloride (EDC) and ethane concentrations to maximize catalytic ethylene oxide production rate and yield: Experimental study and modeling
2022, Chemical Engineering Science
Citation Excerpt :
They also stated that rhenium performs in coexistence with cesium in the catalyst composition and draws electron density from the Ag site which leads to producing more electrophilic adsorbed oxygen and as a result more EO rather than CO2 will be generated, promoting SEO. Huš and Hellman (2018) using a computational method showed that chlorine which is strongly electronegative acts as an EO selectivity promoter, but it adversely impacts the activity of catalyst, while cesium which is the least electronegative element facilitates oxygen dissociation leading to the increase of both EO selectivity and catalyst activity. So, these two selectivity promoters act through different mechanisms, one by reducing catalyst activity and the other by raising it.
Simultaneous optimization of EDC and ethane concentrations to maximize ethylene oxide production rate (r_EO) over a commercial catalyst at 17.4 bar, 213–240 °C and in the GHSV range of 3000–5000 h⁻¹ was done in this study. The feed composition and operating conditions were chosen similar to those in the industry, so that the findings are practical. EDC and ethane concentrations in the feed are merged into a practical parameter referred to as equivalent EDC (EDC_eq) in this study which is indicative of the chlorinating power of the gas mixture. It was noticed that the higher the temperature, the higher the optimum EDC_eq to maximize r_EO. Although r_EO had a decreasing trend versus EDC_eq for 213 °C, it showed a maximum with EDC_eq for 225 and 240 °C. While CO₂ production rate (r_CO2) monotonically decreased with EDC_eq, EO selectivity experienced an upward trend with EDC_eq, approaching a horizontal asymptote.
Computational and experimental insights into reactive forms of oxygen species on dynamic Ag surfaces under ethylene epoxidation conditions
2022, Journal of Catalysis
The partial oxidation of ethylene to ethylene oxide (EO) proceeds over supported Ag particles. Adsorption of molecular oxygen upon Ag forms the reactive oxygen species responsible for both epoxidation and combustion reactions, however, the catalyst surface structure and the molecular structure and identity of the reactive oxygen species present remain debated despite past use of appropriate forms of in situ spectroscopy. Specifically, there is no consensus for the molecular origin of widely reported Raman features that appear on contact with only O₂ or in reactant mixtures of C₂H₄ and O₂ at pressures and temperatures relevant for the industrial process, in part, due to the lack of concerted ab initio studies that compute vibrational frequencies for oxygen-containing Ag surfaces. Here, we elucidate the molecular structure of the catalyst surface and reactive oxygen species by coordinating spectral deconvolution of transient and steady-state surface-enhanced Raman spectroscopy of Ag catalysts exposed to oxygen (2 – 101 kPa O₂, 523 – 673 K) and mixtures of oxygen and ethylene (2 – 101 kPa O₂, 0.5 – 9.8 kPa C₂H₄, 523 K) with ab initio thermodynamic modeling and vibrational frequency calculations. These comparisons suggest that during EO catalysis reconstructed surface oxides form and partially or completely encase metallic Ag particles. Raman features near 600 cm⁻¹ that persist in O₂ or cofed C₂H₄ represent O–Ag–O structural motifs that form only on surface oxides and high oxygen-coverage reconstructed Ag_xO_y surfaces. We assign features centered near 810 – 840 cm⁻¹ and at higher frequencies, which are ubiquitous throughout the literature, to dioxygen complexes partially embedded within an oxide-like overlayer that forms during exposure either to O₂ or during steady-state epoxidations. Taken together our results implicate the presence of a combination of monatomic and diatomic surface oxygen species, which emerge with appreciable quantities of subsurface oxygen at temperatures and O₂ pressures representative of EO catalysis.
Towards the understanding of promoting effects of Re, Cs and Cl promoters for silver catalysts of ethylene epoxidation: A computational study
2021, Catalysis Today
Citation Excerpt :
Many promoters have been proposed for Ag-based catalysts of ethylene epoxidation, with Cs, Cl, and Re being considered the most important ones. Cesium, the most studied promoter [30–34], was assumed to tune the electronic states of the adsorbed surface oxygen species due to blocking of the combustion sites, and rather affects the transition states to close the oxametallacycle into ethylene oxide [35,36] than its isomerization into acetaldehyde [37]. Chlorine effects also bring about a vivid research interest [38–47].
Promoting effects of Cs, Re oxospecies, and Cl for silver catalysts of ethylene epoxidation are studied computationally using the DFT approach. The promoters are shown to enhance the extent of activation of initial substances, reaction intermediates, and products, while different combinations of promoters are favorable for different stages of the process. The synergistic effect of Cs, Re, and Cl is connected with the adjustment of the substrate location with respect to silver, thus, setting the favorable conditions for rapid transformation of ethylene into ethylene oxide. Re and Cl complement each other to balance and mitigate the effect of the Ag-bound Cs to alter the distribution of electronic density, tuning the electronic states of surface oxygen species by making it more electrophilic, and providing the substrate binding strength that ensures “optimal” substrate residence time and improves the catalyst performance in ethylene epoxidation.
Ethylene oxidation on unpromoted silver catalysts: Reaction pathway and selectivity analysis using DFT calculations
2021, Surface Science
Chemical conversions in catalytic partial oxidation processes of light hydrocarbons are responsible for the production of numerous industrial chemicals, plastics, and intermediates. These processes are relatively expensive to perform, and are typically operated at high thermodynamic inefficiency, so the development of novel, highly efficient catalysts would prove to be very cost effective. Herein, our study focused on surface catalytic mechanisms of the ethylene oxide (EO) formation process. Periodic plane-wave Density Functional Theory (DFT) methods were used to analyze related reaction mechanisms on the Ag(111) surface facet with low coverage. Energetic changes of related species and pathways were calculated. Key surface species are identified to suggest the factors for the observed selectivity during EO formation. Our results are consistent with previous kinetic modeling efforts in the literature which did not employ DFT analysis. Lastly, our study demonstrates how fundamental theoretical investigations and multi-scale modeling techniques are currently impacting the advancement of rational catalyst design and microkinetic modeling techniques in the light hydrocarbon processing industry.

View all citing articles on Scopus

View full text

Dipole effect on ethylene epoxidation: Influence of alkali metals and chlorine

Highlights

Abstract

Graphical abstract

Introduction

Section snippets

Computational details

Thermodynamics

Conclusions

Acknowledgement

Appl. Surf. Sci.

J. Catal.

J. Catal.

J. Catal.

J. Catal.

J. Catal.

Catal. Today

J. Catal.

J. Catal.

J. Catal.

J. Catal.

J. Catal.

Surf. Sci.

J. Catal.

Mol. Catal.

Comput. Mater. Sci.

Surf. Sci.

Chem. Phys. Lett.

Surf. Sci.

J. Am. Chem. Soc.

J. Am. Chem. Soc.

J. Am. Chem. Soc.

J. Am. Chem. Soc.

Catal. Lett.

Phys. Rev. B

ChemCatChem

J. Mater. Chem.

J. Phys. Chem. B

J. Phys. Chem. B

J. Phys. Chem. C

J. Am. Chem. Soc.

Nat. Chem.

Nat. Mater.

Acc. Chem. Res.

J. Am. Chem. Soc.

J. Phys. Chem. B