2 Mercosur Congress on Chemical Engineering 4 Mercosur Congress on Process Systems Engineering

th

nd

STUDY OF THE ACTIVE PHASE OF SILVER CATALYSTS FOR ETHYLENE EPOXIDATION

1 NUCAT/PEQ/COPPE - Universidade Federal do Rio de Janeiro Departamento de Engenharia Qumica - Universidade Federal Fluminense

Abstract. Ethylene oxide is a valuable compound due to its versatility as chemical intermediate in many
industrial processes. Its production is a great proportion business, where millions of tons are produced every year. Although ethylene oxide has been produced for more than fifty years by oxidation of ethylene over Ag/Al2O3, the effect of promoter addition is not yet completely understood. The objective of this work is to study the performance of cesium promoted silver catalyst during the epoxidation reaction, in order to understand the interaction of the reactants with the active sites and the transition state. Silver catalysts were characterized by temperature programmed desorption of pre adsorbed oxygen and ethylene, temperature programmed surface reaction and DRIFTS in situ (20% ethylene, 5% oxygen / He). The addition of Cs favored the dispersion with formation of small silver particles over a thin silver film. The decrease of the particle size leads to the increase of silver surface defects where the epoxidizing electrophilic oxygen species are located. In fact, O2 TPD results showed that Cs favored the formation of strong adsorbed electrophilic oxygen. Epoxidation of ethylene did not occurr during C2H4 TPD, indicating that the reactant mixture (C2H4 and O2) modified the structure of silver favoring formation of oxygen subsurface species, which are essential for the epoxidation. TPSR analysis showed that cesium addition allowed the occurrence of the epoxidation reaction at lower temperatures. DRIFTS results showed that there was only carbonate and acetaldehyde formation during ethylene oxidation over -alumina. Ethylene oxide formation occurred with the introduction of the silver, with still certain influence of the support. The Cs promoted silver catalyst presented higher epoxidation activity, and its spectrum presented other intermediate species. The small influence of the support agrees with previous results of better silver covering in the presence of Cs.

Keywords: Silver, Ethylene and Epoxidation.

1. Introduction
Since the pioneering work of Lefort (1931), ethylene oxide has been produced by heterogeneous catalysis. Silver catalysts over -alumina have been used commercially for ethylene epoxidation for more than 50 years, with great technological advances. The production of ethylene oxide is one of the world greatest industrial capacities among the manufactured chemical products where millions of tons are produced every year. Most of ethylene oxide is converted the ethylene glycol, which is used as antifreeze agent or in the polyester fiber production. The main objective of the recent studies is to increase the selectivity. During the 30s the selectivity turned around 40%, and with the gradual addition of moderators (generally chloride hydrocarbons in the reactant feed), the selectivity increased arriving to about 68% in the 60s. In the last years, significant changes in the preparation

To whom all correspondence should be addressed. Address: NUCAT/PEQ/COPPE, UFRJ - Centro de Tecnologia, Bl.G-128, 21945-970, Rio de Janeiro Brazil E-mail: schmal@peq.coppe.ufrj.br

2 Mercosur Congress on Chemical Engineering 4 Mercosur Congress on Process Systems Engineering

nd

method and the introduction of alkalin,e promoters had further increased the selectivity. Nowadays, the currently commercially catalysts operate with selectivity for ethylene oxide around 80%. The effect of promoter addition to silver catalyst is not yet completely understood. Considering only the electronic effects, Cs should decrease the selectivity to ethylene oxide by electronic charge donation to the atomic oxygen (Serafin et al. 1998). Actually, Grant and Lambert (1985) observed that Cs addition on Ag (111) decreased the CO2 production associated to the ethylene oxide formation. They attributed these facts to an increase in the oxygen electronic density, which inhibits the opening of the epoxide ring and thus its decomposition. This conclusion was also sustained by other work (Tan et al. 1987). Campbell (1986) studied the same system on crystalline phase and suggested that Cs structure is more complex, observing the formation of CsO3 species, which can generate some electronic or geometric effects. A recent report showed that Ag (110) suffers a (1x2) reconstruction with Cs adsorption (Serafin et al. 1998), and that Cs increases the CO2 release by silver in the presence of oxygen. This latter effect decreases the ethylene oxide combustion, inhibiting the additional CO2 formation. Authors explain that from the two opposite effects on the selectivity, the suppression of ethylene oxide oxidation prevails compared to the increase surface oxygen electronic density in the presence of Cs, which enhances the ethylene oxide formation. It turns out that Cs acts either by diminishing and stabilizing the Ag particles size, increasing the crystalline lattice defects, where probably the electrophilic oxygen is located (Grant and Lambert 1985), and on the other hand, may decrease the nucleophilic oxygen concentration (surface Ag2O), which is responsible for the ethylene deep oxidation (Bukhtiyarov et al. 1994b). Once the epoxidation rate remains stable (Goncharova et al. 1995) it can be suggested that the ratio of the amounts of regular and defective surface regions, claimed for ethylene epoxidation, remains constant or slightly increases. Therefore, Cs can generate substantial changes in the structure, such as decrease of Ag2O islands or increase (stabilization) of defective regions. However, excess of Cs may block the active surface sites and, thus, decrease the reaction rate in both directions. Processes of ethylene oxide production had been developed conjugating top technology in different knowledge areas. The use of different supports, methods of preparation and introduction of reaction promoters or inhibitors are some of these examples. In the last years a great progress in the study of the ethylene epoxidation over silver catalysts was reached. The biggest advances were due to development of modern surface science techniques, which had been considered as powerful tools and to be used as ally in the search for total understanding of the stages of reaction. In situ surfaces spectroscopy was used to improve catalytic system in order to attain greater yields of ethylene oxide at lower reaction temperatures, minimizing, thus, the operational costs. Many old questions about epoxidation reaction have been elucidated, however there are still many questions concerned to the transition state. Despite the intense research, these new technologies still do not possess necessary sensitivity to perceive the small changes that generate differences in selectivity, of which depend the industrial development. The challenge for the new research is the correlation between surface science electronic and structural data and the small disturbances, which generate significant economic variations in real catalysts. The objective of this work is to study the morphological and structural changes of silver catalysts with the addition of cesium as promoter agent, as well as the interactions between molecules with the active sites, aiming 2

2 Mercosur Congress on Chemical Engineering 4 Mercosur Congress on Process Systems Engineering

nd

at to understand the transition state of surface and surface mechanism reaction. Promoted and non-promoted silver supported -alumina catalysts were prepared and characterized by chemical and spectroscopic techniques, watching the surface transitory phases and molecule dynamics.

2. Experimental
The catalysts were prepared through silver impregnation on -alumina support (NORTON) by a silver lactate solution (Liu and Shen 1995). The sample containing Cs was made by co-impregnation, adding cesium

hydroxide to the silver solution. For silver content determination, quantitative volumetric analysis by Volhard method was used, after previous silver extraction by HNO3 solution (Kobal Jr. and Sartorio, 1981). The determination of Cs content was determined by atomic absorption and the extraction of Cs was made by reflux using deionized water. The reactant feed was composed of 20% ethylene and 5% oxygen diluted in helium. Temperature programmed analyses was carried out in a multipurpose unit equipped with a quartz microreactor where the flow of gases was controlled by a Matheson 8274 flow Metter and with a Balzers mass spectrometer with Prisma QMS 200 quadrupole. The samples were pretreated with a 30 mL/min flow of 20% O2 in He at 350oC for 4 h, cooled at room temperature and being more 18 h under this reactant flow. For ethylene desorption, the samples were submitted the flow of 30 mL/min of pure ethylene 0oC (ice bath) for 1 hour. O2 or C2H4 desorption was carried from 20 to 550oC under 30 mL/min of He flow. Temperature programmed surface reactions was carried under 30 mL/min of reactant feed and the temperature varied from 100 to 800oC, with a rate of 5oC/min. Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS) analyses were carried through model Thermo Nicolet Nexus 470 equipment with MCT-A detector and Spectra-Tech accessory. The samples were dried with 20mL/min of He flow at 150oC, when a spectrum was obtained. After, they were treated with 23 mL/min of 5% O2/He for 2 h at the same temperature, and cooled at room temperature, and another spectra was obtained. Then the gas flow was changed by 19 mL/min of reactant mixture at different reaction temperatures 25, 260, 300 and 400oC, and a third spectra was get. The spectrum of the sample after O2/He treatment was used for subtraction background.

3. Results and Discussion

2 Mercosur Congress on Chemical Engineering 4 Mercosur Congress on Process Systems Engineering

nd

Table 1. Description of the prepared catalysts.

Catalyst 15Ag 15AgCs

Silver content (% Weight) 15.7 15.2

Cesium content (ppm) 211

Ag/Al Ratio (bulk) 0.090 0.085

3.2. Oxygen Temperature Programmed Desorption (O2TPD) The results of O2TPD (m/e = 32) are represented in Figure 1 and show that the -alumina support did not present oxygen adsorption and that the silver containing catalysts presented a capacity of oxygen adsorption. The influence of the cesium addition was evidenced through the total amount of adsorbed oxygen for the cesium promoted sample because while the non promoted catalyst presented a single oxygen desorption peak at 430oC, the cesium promoted catalysts besides this one, presented a greater (5 times) and more strongly adsorbed oxygen species at 500oC. The different oxygen species that can be adsorbed by silver are widely argued in literature, mainly for monocrystals under low pressures conditions. According to Trukhan et al. (1997), the molecular oxygen species are more easily desorbed from silver surface than the nucleophilic oxygen species, which are desorbed at higher temperatures. This nucleophilic oxygen can be transformed into more strong adsorbed species as: reversible subsurface oxygen and irreversible electrophilic oxygen, which is the epoxidizing oxygen.

500 C
m/e = 32

430 C

Intensity (A.U.)

15AgCs 15Ag support

Temperature ( C)

Fig. 1. O2 TPD profiles of silver catalysts and the support (m/e = 32).

The results indicate that the action of the Cs in the epoxidation of ethylene is its capacity to promote greater formation of adsorbed electrophilic oxygen species over silver surface. These comments are in accordance with Bukhtiyarov et al. (1994b), that reported the Cs addition increases the silver crystalline network defects, where

2 Mercosur Congress on Chemical Engineering 4 Mercosur Congress on Process Systems Engineering

nd

epoxidizing oxygen species are located the and, on the other hand, diminishes the concentration of nucleophilic oxygen species (surface Ag2O), which are the responsible for deep ethylene oxidation.

3.3. Ethylene Temperature Programmed Desorption (Et TPD) Silver can chemisorb dissociatively the accumulated molecular oxygen in the neighborhoods of structure defects, to form a surface silver oxide layer. It is necessary to have silver cations (Ag+) to fix the ethylene at catalyst surface and these cations are located near the border between the metal and the silver oxide particles (Bal'zhinimaev, 1999). The results of ethylene TPD analyses are presented in Figure 2, where ethylene desorption (m/e = 28) was not observed, but only the appearance of signal referred to mass 44, which is the main signal, (intensity 100%), of both, partial (C2H4O) and total (CO2) ethylene oxidation reaction products. Signal of mass 29, which is characteristic the secondary oxide ethylene signal, was not detected. This fact is an indicative that ethylene was converted preferentially to CO2. There was not any CO2 desorption from the support, indicating that the ethylene was not adsorbed by -alumina. 15Ag and 15AgCs presented similar desorption profiles, indicating that the Cs presence did not influence total oxidation reaction.

m/e = 44

Intensity (A.U.)

15AgCs

15Ag

support
200 300 400 500 isothermic
o

Temperature ( C)

Fig. 2. Ethylene TPD profiles of silver catalysts and the support (m/e = 44).

Bukhtiyarov et al. (1994b), related that ethylene epoxidation did not occur if silver is treated previously with oxygen and later with ethylene, and they observed the presence of O1s at 530,5 eV by XPS experiments. However when the silver was exposed at reactant feed the epoxidation occurred and the XPS detected one more species (O1s, 528,4 eV). According to the authors, this fact occurs because the surface is activated by reactant mixture, increasing the ionicity of the adsorbed oxygen and favoring, thus, its participation in the ethylene epoxidation. Some authors (Backx et al., 1981, Campbell and Paffett, 1984) had identified the species at 528,4 eV as being subsurface oxygen, which are dissolved in the silver matrix. The presence of this species is told as one of the requisite for epoxidation activity (Borman and Westerterp, 1995). 5

2 Mercosur Congress on Chemical Engineering 4 Mercosur Congress on Process Systems Engineering

nd

3.4. Temperature Programmed Surface Reaction (TPSR) TPSR results are described in Figure 3. The quantification of the reaction products was not possible because the coincidence among the values of mass of the monitored compounds fragments: 28 (ethylene and CO), 44 (CO2 and ethylene oxide). The formation of ethylene oxide can be followed through the secondary mass 29 (COH+) (Bukhtiyarov et al., 1999, 1994a, 1994b, Podgornov et, 2000). The reaction products were monitored by the signals of masses m/e = 44 and 29. For the tested catalysts, the profiles of desorption of the mass fragment 29 presented coincident maxima with mass fragment 44 at the same temperature, indicating that these correspond to the ethylene oxide formation. The results showed that formation of ethylene oxide was favored with the presence of cesium, because the promoted catalysts presented more activity with formation of ethylene oxide at 100oC temperature lower than for non-promoted catalyst.

15AgCs

m/e = 44 m/e = 29x5

Intensity (A.U.)

200

400

600

800

1000

15Ag

200

400

600

800

1000

support

200

400

600
o

800

1000

Temperature C

Fig. 3. Profiles of TPSR of the reaction of oxidation of ethylene of silver catalysts and the support.

3.5. Diffuse Reflectance Fourier Transform Infrared Spectroscopy (DRIFTS) The results of in situ DRIFTS of the ethylene oxidation reaction over -alumina, 15Ag and 15AgCs catalysts are presented in Figure 4. Two different backgrounds were used for subtraction: the catalyst after oxygen treatment at 150oC (Figures A, B and C), and the reactant gas (D). The formation of the different 6

2 Mercosur Congress on Chemical Engineering 4 Mercosur Congress on Process Systems Engineering

nd

products was determined by the presence of adsorption bands, which were ascribed to different compounds according to the literature (Force and Bell, 1975a, 1975b; Kilty and Sachtler, 1974, Goncharova et al., 1995, Tsybula et al., 1995). The gaseous ethylene presents spectrum with bands at four different regions: 3200-2900 cm-1 (region I), 1500-1400 cm-1 (region II), 1220 cm-1 (region III) and 1000-890 cm-1 (region IV). When ethylene is adsorbed in presence of oxygen, another band (region V) raise at 1900-1800 cm-1. Carbon dioxide presents characteristics bands (region VI) at 2370 and 2330 cm-1 and ethylene oxide, besides some bands that rise at the same ethylene regions (I, II and IV) also presents a band at 1245 cm-1.

1043

15Ag 2350 400 C

1195 1245 1580

support
1580-1265

Absorbance (A.U.)

Absorbance (A.U.)

300 C 260 C
o

2370 2330 400 C 300 C 260 C 25 C

25 C

4000

3500

3000

2500

2000
-1

1500

1000

4000

3000

2000
-1

1000

Wavenumber (cm )

Wavenumber (cm )

A
15AgCs 1245 1195

Absorbance (A.U.)

400 C
o

300 C 260 C
o

Absorbance (A.U.)

1580 3750 support 15Ag 15AgCs

25 C

4000

3000

2000
-1

1000

4000

3500

3000

2500

2000
-1

1500

Wavenumber (cm )

Wavenumber (cm )

C D Fig4. Profiles of in situ DRIFTS after reaction of oxidation of ethylene using the sample as background: (A) support,
(B) 15Ag, (C) 15AgCs; and using the reactant feed as background: (D) after reaction at 400oC.

Figure 4 (A) shows the DRIFT spectra of -alumina at different reaction temperatures. The spectrum at 25oC presents the ethylene characteristic bands (I-V). With the increase of the reaction temperature the bands referred to ethylene decreased and at 260oC characteristic bands of carboxylate species (1580 1260 cm-1) and acetaldehyde (1043 cm-1) raised. At higher temperatures the bands referred to the formed CO2 decreased and the 7

1300-900

1000

2 Mercosur Congress on Chemical Engineering 4 Mercosur Congress on Process Systems Engineering

nd

bands referred to carboxylate species and aldehyde became higher, indicating the production of stable undesirable species. The silver catalysts, 15Ag and 15AgCs (Fig. 4B and C, respectively), presented similar profiles, with formation of CO2 higher than that for the support, but giving rise to other band (2350 cm-1) in the same region of gaseous CO2, which is attributed to carbon dioxide adsorbed over silver atom, as showed in Scheme I. At room temperature gaseous ethylene bands were observed, and there was also the presence of a sharp band at 1195 cm-1 related to ethylene adsorbed over silver oxide (Scheme II). These bands decreased with the increase of the temperature with the formation of CO2 and give raise a new band at 1245 cm-1, ascribed to the formation of ethylene oxide. These observations are in accordance with Bal'zhinimaev (1999) who reported that for the formation of the ethylene oxide previous ethylene adsorption over oxidized silver is necessary. Comparing the spectra of the silver catalysts, despite the great decrease of the support influence over the silver impregnated samples, the non-promoted catalyst (15Ag) presented a band referred to an acetate at 1580 cm , indicating a residual influence of the support. The cesium promoted silver catalyst presented only bands ascribed to ethylene, carbon dioxide and ethylene oxide.
O C O Ag O
-1

CH2 CH2 O Ag

Scheme I* Scheme II* * Assign of Infrared bands to surface structures (Force and Bell, 1975a, 1975b).

DRIFT spectra using the reactant feed (20% ethylene, 5% oxygen in He) as analytical background after reaction at 400oC, are presented in Figure 4D. The profiles present a great decrease of support influence at the silver catalysts evidenced by bands at 1300-900 cm-1, but the non-promoted catalyst presented bands referred to acetate (1580 cm-1) and band at 3750 cm-1 ascribed to terminal hydroxyl groups linked with surface support Al atoms. The results indicate that the cesium addition favors the epoxidation reaction by decreasing the formation of undesirable compounds, which are stable at higher temperatures and are formed by support influence, increasing, thus, the selectivity to ethylene oxide. These data are in accordance with previous results showing that silver catalysts present small particle aggregates on the support, while on Ag/Cs promoted catalyst, a thin silver film covered most of the -Al2O3 surface, decreasing, thus the support influence over the silver. The silver film formation was indicated by phase contrast absence at atomic force microscopy (AFM) images of silver promoted catalyst (Carvalho et al., 2003) and also observed through ion scattering spectroscopy (ISS) of the first atomic layer (Epling et al. 1997; Minahan et al. 1997). Serafin et al., 1998 and Grant and Lambert, 1985, also reported that the importance of addition of cesium was its capacity of decrease the influence of alumina, which favors the opening of epoxide ring.

2 Mercosur Congress on Chemical Engineering 4 Mercosur Congress on Process Systems Engineering

nd

4. Conclusions
The oxygen adsorption over silver was influenced by the presence of Cs, which favored the appearance of peak referred to strong adsorbed oxygen electrophilic species, which are considered the epoxidizing species. The reactant feed (C2H4 and O2) modifies the silver structure provoking the appearance of oxygen subsurface species, which are active for the epoxidation of ethylene. Cesium addition promoted the reaction of epoxidation, making that the temperature necessary for ethylene oxide formation decreased at about 100oC. In situ DRIFTS results confirmed the lower influence of the support at Cs promoted silver catalyst, once it did not present any formation of carbonate, aldehyde or hydroxyl species, with a formation of a silver film recovering the support surface, preventing the opening of the epoxide ring.

References
Backx, C., de Groot, C.P.M. and Biloen, P. (1981). Adsorption of Oxygen on Ag(110) Studied by High Resolution ELS and TPD. Surf. Sci., 104, 300. Balzhinimaev, B.S. (1999). Ethylene Epoxidation over Silver Catalysts. Kinet. Catal,. 40, 795. Borman, P.C. and Westerterp, K.R. (1995), An Experimntal Stydy of Kinetics of the Selective Oxidation of Ethene over a Silver on -Alumina Catalysts. Ind.Eng.Chem. Res., 34, 49. Bukhtiyarov, V.I. Prosvirin,I.P. and Kvon, R.I. (1994a), Study of Reactivity of Oxygen States Adsorbed at a Silver Surface Towards Ethylene by XPS, TPD and TPR. Surf. Sci., 320, L47. Bukhtiyarov, V.I., Kaichev, V.V., Podgornov, E.A. and Prosvirin, I.P. (1999). Size Effect in Ethylene Oxidation on Silver Catalysts. Influence of Support and Cs Promoter. Catal. Lett., 57, 233. Bukhtiyarov, V.I. Boronin, A.I. and Savchenko, V.I. (1994b). Stages in the Modification of a Silver Surface for Catalysis of the Partial Oxidation of Ethylene: I. Action of Oxygen. J. Catal., 150, 262. Campbell C.T. (1986). Chlorine Promoters in Selective Ethylene Epoxidation over Ag(111) A Comparison with Ag(110). J. Catal., 99, 28. Campbell, C.T. and Paffett, M.T. (1984). The Interactions of O2, CO and CO2 with Ag (110). Surf. Sci., 143, 517. Carvalho, M.C.N.A., Perez, C.A., Simo, R.A., Passos, F.B. and Schmal, M. (2004). The Promoting Effect of Cesium on Structure and Morphology of Silver Catalysts. An. Acad. Bras. Cinc., 76, 19. Epling W.S., Hoflund G.B. and Minahan D.M. (1997). Study of Cs-Promoted -Alumina-Supported Silver EthyleneEpoxidation Catalysts. IV. Model Study. J. Catal., 171, 490. Force, E.L. and Bell, A.T. (1975)a. Infrared Spectra of Adsorbed Species Present during Oxidation of Ethylene over Silver. J. Catal., 38,440. Force, E.L. and Bell, A.T. (1975)b. The Relationship of Adsorbed Species Observed by Infrared Spectroscopy to the Mechanism of Ethylene Oxidation over Silver. J. Catal., 40, 356. Goncharova, S.N., Paukshtis, E.A. and Balzhinimaev, B.S. (1995). Size Effect in Ethylene Oxidation on Silver Catalysts. Influence of Support and Cs Promoter. App. Catal., 126, 67.

2 Mercosur Congress on Chemical Engineering 4 Mercosur Congress on Process Systems Engineering

nd

Grant, R.B. and Lambert, R.M. (1985). A Single-Crystal Study of Silver-Catalyzed Selective Oxidation and Total Oxidation of Ethylene. J. Catal., 92, 364. Kilty, P.A. and Sachtler, W.M.H. (1974). The Mechanism of the Selective Oxidation of Ethylene to Ethylene Oxide. Cat. Rev. Sci. Eng., 10, 1. Kobal Jr. J. and Sartorio L. (1981). Manual de Anlise Qumica Quantitativa Convencional. Ed. Moderna, So Paulo, SP. Lefort, T.E. (1931). Societ Franaise de Catalyse Generalise Fr. Pat.729, 952, 27. Liu Y. and Shen, S. (1995). Direct Epoxidation of Ethylene in a Dilute Reaction Stream in Oxidative Coupling of Methane. Appl. Catal., A: Gen., 121, 57. Minahan D.M., Hoflund G.B., Epling W.S. and Schoenfeld D. (1997). Study of Cs-Promoted -Alumina-Supported Silver Ethylene-Epoxidation Catalysts. III. Characterization of Cs-Promoted and Nonpromoted Catalysts. J. Catal., 168, 393. Podgornov, E.A., Prosvirin, I.P. and Bukhtiyarov, V.I. (2000). XPS, TPD and TPR Studies of Cs-O Complexes on Silver: their Role in Ethylene Epoxidation. J. Mol. Catal. A, 158, 337. Serafin J.G, Liu A.C. and Seyedmonir S.R. (1998). Surface Science and the Silver-Catalyzed Epoxidation of Ethylene: an Industrial Perspective. J. Catal., 131, 157. Tan S.A., Grant R.B. and Lambert R.M. (1987). The Silver Catalyzed Decomposition of N2O over Ag(111) and Ag/-Al2O3. J. Catal., 104, 156. Trukhan, S.N., Ivanov, V.P. and Balzhinimaev, B.S. (1997). Study of Oxygen Adsorption on Ag/-Al2O3 Catalysts under Ultrahigh/High Vacuum Conditions. Kinet. Catal., 38 565. Tsybula, S.V., Kryukova, G.N., Goncharova, S.N., Shmakov. A.N. and Balzhinimaev, B.S. (1995). Study of the Real Structure of Silver Supported Catalysts of Different Dispersity. J. Catal., 154, 194.

10