545931 Combined Solid- and Gas Phase Kinetic Modelling: Ethylene Oxychlorinationover K-Promoted Cu/Al2O3 Catalyst

Wednesday, June 5, 2019: 3:27 PM
Texas Ballroom A (Grand Hyatt San Antonio)
De Chen1, Terje Fuglerud2, Endre Fenes3, Hongfei Ma3, Magnus Rønning4, Samuel Regli5, Yanying Qi5 and Kumar Rout6, (1)Chemical Engineering, Norwegian University of Science and Technology, Trondheim, Norway, (2)INOVYN, Porsgrunn, Norway, (3)Norwegian University of Science and Technology, Trondheim, Norway, (4)Department of Chemical Engineering, NTNU, Trondheim, Norway, (5)NTNU, Trondheim, Norway, (6)SINTEF, Trondheim, Norway

Combined Solid- and Gas Phase Kinetic Modelling: Ethylene Oxychlorinationover K-Promoted Cu/Al2O3 Catalyst

Endre Fenes1, Hongfei Ma1, Yanying Qi1, Samuel Regli1, Kumar R. Rout2*, Terje Fuglerud3, Magnus Rønning1, De Chen1*

1 – Department of Chemical Engineering, NTNU, Trondheim, 7491, Norway

2 – SINTEF, Trondheim, 7491, Norway

3 – INOVYN Norway, Porsgrunn, 3936, Norway 

*Corresponding author: kumarranjan.rout@sintef.no, de.chen@ntnu.no

<>Introduction

Ethylene oxychlorination is an industrially important process where the CuCl2/γ-Al2O3 based catalyst undergoes a reduction-oxidation process (Eq.1).[1] The catalyst is disposed to by product formation over alumina acid sites,  volatilization of Cu+ species and hotspot formation due to the exothermicity of the reaction, all of which is industrially abated with K addition. While K simply titrates alumina acid sites, [1] the nature of K interaction with copper species and the subsequent effect on the reduction - and oxidation reaction kinetics is still an open question. Building on our kinetic model of the neat CuCl2/γ-Al2O3 catalyst, [2] wherein we demonstrated that the nature of the active site depends on the degree of reduction, we extend this approach to K doped catalysts in combination with DFT modelling and operando XAS spectroscopy, the two latter are currently underway.   

                                                        (1) <>Experimental

The catalyst were prepared through incipient wetness impregnation of CuCl2•2H2O and KCl and contains 5wt% Cu and 1.2 wt% K. The catalyst was characterized with ex situ XRD, BET and XPS. Step transient reduction (PC2H4=0.065) and oxidation (PO2=0.3) experiments were performed at differ temperatures to obtain the kinetic parameters. During co-feeding experiments, C2H4, O2 and HCl were introduced simultaneously (T=503K, Ptot=1Bar, PC2H4=0.009, PO2=0.0045, PHCl=0.018). The effluent gases were monitored my mass spectrometry and Cu oxidation state quantified thorough operando UV-Vis-NIR spectroscopy.[3]  <>Results

During the step transient reduction, ethylene gradually removes Cl from the catalyst, reducing Cu2+ to Cu+ and the Cu coordination from four to two. However, when plotting the reaction rate with respect to the amount of Cl remaining in the catalyst, it is immediately clear that two active sites are necessary in order to properly describe the overall reaction (r1). The presence of at least two active sites is also supported by temperature programmed reduction and preliminary DFT results. (not shown) Initially only four-coordinated Cu (r1,I) is present on the catalyst as CuCl2. Then, as Cl is removed and the Cl concentration of the catalysts drops below a certain concentration (Cl**) defined by Gaussian deconvolution, three-coordinated Cu appear (r1,II). As the catalyst is almost completely reduced (Cl*), no more four-coordinated Cu can be sustained and only the three coordinated site remains active until all reducible Cu is converted into the inert, two-coordinated CuCl. The kinetic model describing the reduction reaction is summarized below. (Eq. 2-4)

                                                                                                                                     (2)

                                                                                                                   (3)

                                                                                                            (4)

Similar for the oxidation step, there is initially only one active site, two coordinated CuCl. At some point (CuCl**), three and four coordinated Cu coexist and as the catalyst is almost completely oxidized (CuCl*), only three-coordinated Cu sites are active. Thus, a kinetic model similar to the reduction step is proposed (Eq. 5-7) with the addition of a desoprtion term to account for O2 dissociation. It is noted that the third step (Eq. 1) is generally considered to be much faster than the redox steps and thus assumed kinetically irrelevant.[3]

                                                                                                                                                  (5)

                                                                                                               (6)

                                                                                                         (7)

The above model was then fitted towards the step transient experiments of the neat and K doped catalyst at different temperatures in order to obtain the activation energy and preexponential factor which are presented in Table 1. From the table, it is evident that K addition lowers the reduction activation energy although the largest difference is in the preexponential factor of the four coordinated CuCl2 which decreased almost three orders of magnitude, resulting in a decreased reduction turnover frequency (TOF) compared to the neat catalyst. This indicates that entropy changes is more significant in K doped catalysts than in the neat catalyst, which could be explained by formation of supported catalyst particles in the K doped catalyst with larger degree of crystallinity compared to  a more disorded, partial monolayer coverage of CuCl2 on the neat catalyst as indicated by TEM images. (not shown) Results from combined in situ XAS and UV-Vis-NIR experiments at the Swiss-Norwegian

Table 1. Kinetic parameters determined from step transient experiments.

Reduction

Oxidation

Catalyst

Parameter

Part I

Part II

Part I

Part II

5Cu

A [gCu / mol atm s]

1.69E+08

4.94E+07

1.7E+05

2.91E+06

Ea [kJ/mol]

60.25

65.40

41.22

52.27

1.2K5Cu

A [gCu / mol atm s]

2.76E+05

1.58E+06

3.63E+05

2.04E+08

Ea [kJ/mol]

45.22

51.89

40.34

70.17

beamline might elucidate the exact nature of the catalyst particles further and especially the tendency of KCuClx mixed salt formation. To our knowledge, we are the first to have brought synthesized samples of KCuCl3 as a reference material.

With respect to the oxidation step, the oxidation of the two coordinated site is fairly similar at the investigated K loading while K addition increases both the preexponential factor and activation energy for the oxidation of the three-coordinated site. The overall effect is a slightly increased oxidation rate compared to the neat catalyst, and in combination with the decreased reduction activity, this should result in a lower concentration of volatile CuCl, minimizing Cu volatilization.

The kinetic parameters in Table 1 was applied in a 1D reactor model to simulate conditions were all reactants were introduced to the reactor simultaneously and compared to the actual performance of the laboratory reactor. No further fitting of the kinetic parameters were carried out and the estimated reaction rate (2.23E-04 [mol/molCu s]) corresponded well with the experimentally observed one (2.28E-04 [mol/molCu s]). The experimental Cu2+ concentration was quantified through operando UV-Vis-NIR spectroscopy, and it is noted that the slight discrepancy between the theoretical modelling and experimental data in Figure 2 might be due to carbon formation increasing the abosorption of the visible spectrum. It is though clear that the proposed kinetic model is capable of predicting both the gas phase kinetics as well as the evolution of the active site. While both redox steps were modelled with two active sites in order to account for the changing electronic properties of the active site, resulting from altered Cu coordination as the reacton proceeds, the actual transformation might be continnuous in nature. Although this is a topic we are currently persuing from several angles, the two site approach is a computationally inexpensive approch with physically meningful active sites that are more easily envisioned.

<>Conclusions

A kinetic model, capable of accurately predicting both gas phase and solid state kinetics while taking into account the dynamics of an changing active site is developed from step transient experiments. The kinetic model is capable of accurately predicting the behavior of a laboratory scale reactor during co-feeding conditions. The proposed kinetic model is expected to be applicable in other systems where the catalyst partakes in redox reactions, and highlights the common assumption that the reaction kinetics is a function of temperature and partial pressures alone while ignoring the impact of altering the active site coordination sphere. Both DFT and analysis of combined, operando UV-Vis-NIR and XAS is currently underway and is expected to help elucidate further details about the catalytic system. References

[1] N. B. Muddada, U. Olsbye, L. Caccialupi, F. Cavani, G. Leofanti, D. Gianolio, S. Bordiga, C. Lamberti, Phys. Chem. Chem. Phys. 12 (2010) 5605.

[2] E. Fenes, Y. Qi, M. F. Baidoo, H. Ma, K.R. Rout, T. Fuglerud and D. Chen, In preparation, (2019)

[3] K.R. Rout, M. F. Baidoo, E. Fenes, T. Fuglerud and D. Chen, J. Catal. 352, (2017), 218-228

Figure 1. a) Reduction reaction rate with respect to reducible Cl. b) Oxidation reaction rate with respect to available CuCl.

Figure 2. Comparison of the evolution of Cu2+ species during co-feeding experiments on the 1.2K5Cu catalyst. (T=503K, Ptot=1Bar, PC2H4=0.009, PO2=0.0045, PHCl=0.018).


Extended Abstract: File Not Uploaded
See more of this Session: Oxidative Transformations of Light Hydrocarbons Rapid Talks
See more of this Group/Topical: General Submissions