- Email: [email protected]
Characterization of carbon deposits on coked lithium phosphate catalysts for the rearrangement of propylene oxide
Catalysis Communications 64 (2015) 22–26
Contents lists available at ScienceDirect
Catalysis Communications journal homepage: www.elsevier.com/locate/catcom
Short communication
Characterization of carbon deposits on coked lithium phosphate catalysts for the rearrangement of propylene oxide Li Jiang, Hao Li, Yanan Wang, Weihua Ma ⁎, Qin Zhong Department of Chemical Engineering, Nanjing University of Science and Technology, Nanjing 210094, PR China
a r t i c l e
i n f o
Article history: Received 6 November 2014 Received in revised form 25 January 2015 Accepted 27 January 2015 Available online 28 January 2015 Keywords: Rearrangement Propylene oxide Regeneration Carbon deposits
a b s t r a c t Basic lithium phosphate catalysts for the rearrangement of propylene oxide can be easily deactivated due to carbon deposits. In-situ DRIFTS, TG–DTG, NMR and XPS were used to study carbon deposits. Results showed that carbon deposits were mainly in the form of CxHy and CxHyO, which could be decomposed in a mixture of N2 and O2 atmosphere at higher than 300 °C or in N2 atmosphere near 500 °C. A shorter period of time was needed in the atmosphere containing O2. © 2015 Published by Elsevier B.V.
1. Introduction
2. Experimental
The rearrangement of epoxides to allylic alcohols or carbonyl compounds is a fundamentally important transformation with significant synthetic and industrial utility. A wide range of catalysts are associated with this reaction, and this transformation has been attracting great attentions [1–3]. The constitution of the rearrangement products was influenced by the acidity or basicity of the catalysts [4–6]. Lithium compounds have been attractive candidates as catalysts for the rearrangement of epoxides. For liquid phase catalysis or asymmetric catalysis [7–10], lithium compound bases [6,11,12] or Schlosser's Li/K mixed superbases [13] are often applied in the arrangement of functionalized, aryl or chiral epoxides into allylic alcohols. For epoxides with small molecule, especially propylene oxide (in abbreviation, PO), basic lithium phosphate is often used to catalyze the rearrangement to allyl alcohols [14,15]. There are two processes for the rearrangement of PO to allyl alcohol, gas–solid phase and slurry phase. Gas–solid phase process is much simpler and cleaner than the other, but carbon deposition has always been a major problem. Therefore, it is important to analyze the carbon deposits on the deactivated catalysts. In our previous work, Li3PO4 catalysts gradually got deactivated as reaction time increased and its BET surface area decreased obviously due to carbon (or coke) deposition [16]. This investigation is mainly focused on carbon deposits concurrent with the decomposition or removal of carbon deposits.
2.1. Catalyst preparation
⁎ Corresponding author. E-mail address: [email protected] (W. Ma).
http://dx.doi.org/10.1016/j.catcom.2015.01.026 1566-7367/© 2015 Published by Elsevier B.V.
Lithium hydroxide monohydrate (20.14 g) was dissolved in 125 mL of de-ionized water at 65 °C in a 500 mL three-necked flask with a stirrer. A solution prepared from sodium phosphate dodecahydrate (45.61 g) and water (125 mL) was dropped into the flask with stirring for 1 h. After the mixture was ripened for 3 h, the resultant white precipitate was collected by filtration, dried at 120 °C for 12 h and then calcined at 320 °C for 8 h. Catalytic isomerization reactions were conducted in a φ10 mm fixed-bed reactor at 290 °C using a mixture of N2 and vaporized propylene oxide stream, with weight hourly space velocity (WHSV) value of about 9 h−1 [17]. A series of used or coked lithium phosphate catalysts were obtained. 2.2. Catalyst characterization Used or coked catalysts were characterized by thermo gravimetric analysis (TG, Mettler-Toledo TGA/SDTA851e) at the rate of 20 °C/min in both nitrogen (99.999%, 30 mL/min) and air atmosphere. XPS experiments were carried out on a RBD upgraded PHI-5000C ESCA system (Perkin Elmer). All binding energies were referenced to the C1s neutral carbon peak at 284.6 eV. 13C CP-MAS-NMR spectra were performed on a Bruker DSX400 spectrometer. Tetramethylsilane was used as external reference for the chemical shifts (δ). In situ DRIFTS was used to investigate the decomposition of the carbon deposits at different temperatures. Pure KBr was used as background and the mixture of KBr and the used catalysts (3:1) was used
L. Jiang et al. / Catalysis Communications 64 (2015) 22–26
23
Fig. 1. 13C MAS-NMR (a) and DRIFTS spectra (b) of used catalysts (24 h).
as samples. The spectra were recorded on a Nicolet iS10 spectrometer equipped with an iZ10 Auxiliary Module, which consists of a liquidN2-cooled high-sensitivity MCT-A detector and a DRIFTS cell (SpectraTech) with KBr windows. All spectra were recorded at an accumulation of 20 scans with a resolution of 4 cm−1. The spectra began to be recorded as soon as the sample was heated in pure N2 gas flow. When the temperature reached the set value, O2 was introduced into the cell. The IR data are reported as log 1 / R, where R is the sample reflectance. The function log 1 / R (=“absorbance”) gives a better linear representation of the band intensity against sample surface coverage than that given by the Kubelka–Munk function for strongly absorbing media such as those based on oxides [18]. 3. Results and discussions 3.1. DRIFTS and NMR analysis In order to analyze the types of carbon deposits, DRIFTS and 13C CPMAS-NMR were used to characterize carbon deposits on the used catalyst for 24 h on-stream [19], as shown in Fig. 1. The assignment of the peaks shown in Fig. 1a is given in Table 1. Seen from Fig. 1a and Table 1, there are mainly four groups in carbon deposits, − CH3, − CH2, C–O–C or C–OH, and − C6H5. In the DRIFTS spectrum, the broad peaks at 2828–3024 cm−1 are mainly attributed to symmetrical and asymmetrical C–H stretching vibrations. They are characteristics of alkyl groups of the carbon deposits on catalyst surface. Judged by the peak of δCH at 1449 cm− 1, alkyl species with three or more carbon atoms in the alkyl chains may also exist. Also, olefinic species may exist in the used catalyst because their characteristic stretching vibration frequencies appear at 3000–3024 cm−1. DRIFTS analysis shows that the composition of carbon deposits is very complicated. Carbon deposits mainly consist of both aliphatic and aromatic hydrocarbons containing oxygen element. Other types of
carbon, such as olefinic species, graphite-like carbon, are also possible. Their chemical shifts or peaks may be merged with those of the main four groups into the broad peaks at 34.5, 74.9, and 127.5 ppm shown in Fig. 1a. The − CHO group, with the characteristic peak at 1705 cm−1 in Fig. 1b, can be seen at 210 ppm in Fig. 1a.
3.2. XPS analysis Fig. 2 shows the C1s XPS curves of the fresh and used catalysts. For the fresh catalyst, there are three distinct peaks. The peak at 289.4 eV on the surface of basic lithium can be attributed to C_O bonds of CO2− 3 phosphates [20]. The peak at 284.7 eV indicates the presence of carbon contamination. In the curve of the deactivated catalyst, there are three main distinct peaks, the peak at 283.8 eV indicates the presence of graphitic carbon species in the form of CxHy [21–24]. The peak at 284.6 eV indicates the presence of carbon contamination. And the peak at 285.8 or 285.4 and 285.6 eV can be attributed to the C–O bonds. Moreover, there is a small peak at 288.9 or 288.7 and 288.8 eV, which is due to C_O bonds in species like CxHyO [25,26], a kind of carbon deposit containing O element. The binding energy of C_O bonds in CxHyO species is lower than that in CO23 − groups, since CO23 − disappeared after the catalyst was used for 24 h on-stream at 300 °C.
Table 1 13 C NMR chemical shifts and their possible assignments [18,19]. δ (ppm)
Possible carbon types
18.2 23.0 34.4 39.1 75.0 127.5 134.2 141.7 210.0
CH3 end group to a long paraffinic chain CH2 penultimate carbon in a long paraffinic chain CH and CH2 α to an aromatic ring CH α to the paraffinic chain in alicyclic group CH2 bound to an alkoxyl Aromatic carbons bound to hydrogen Aromatic carbon bound to CH3 Aromatic carbon bound to substituted groups excluding CH3 −CHO
Fig. 2. XPS curves of fresh and used catalysts.
24
L. Jiang et al. / Catalysis Communications 64 (2015) 22–26
Fig. 3. TG–DTG curves of fresh and used catalysts in N2 atmosphere.
3.3. TG–DTG analysis TG–DTG was used to study the decomposition of carbon deposits over different reaction times. The curves in N2 are shown in Fig. 3. For the fresh catalyst, there is a small change in mass (weight loss ratio was about 1.2%) as shown in Fig. 3a. It is due to desorption of water and carbon dioxide adsorbed on catalyst surface. When it turns to the used catalysts, weight loss ratios of the catalysts used for 2 h, 4 h, 6 h and 24 h are 12.5%, 15.4%, 20.5%, and 21.9%, respectively. There is a distinct step down in mass during the temperature range of 300–500 °C, and the weight loss ratios of the catalysts used for 2 h, 4 h, 6 h and 24 h are 11.6%, 10.5%, 15.2%, and 17.6%, respectively. This step is probably caused by the decomposition of hydrocarbon, i.e., carbon deposits. For the catalyst used for 2 h, except the decomposition peak with Tmax = 420 °C there is another shoulder peak at 300–400 °C, which corresponds to a kind of young coke of low molecular weight or with high H/C. However, for the 4 h, there are a main peak at 390 °C and a small broad peak at 500 °C, which correspond to older coke with low H/C. For the 6 h and 24 h, it can be seen that there are other weight changes in mass at higher than 700 °C in TG curves, which correspond to the peaks with Tmax = 728 °C in DTG curves. This may be the decomposition of the aged coke with much lower H/C or the slow oxidation of the graphite-like carbon. Therefore, it can be concluded that carbon deposits formed over different reaction times, especially less than 6 h,
are different. The carbon deposits formed within 2 h are of lower molecular weight and can be easily decomposed. But the carbon deposits formed within 4 h are of higher molecular weight and a bit more difficult to be decomposed than that of 2 h. For the samples used for 6 h and 24 h, part of the carbon deposits can only be removed at about 728 °C. However, the decomposition peaks of the carbon deposits under air atmosphere are all below 450 °C, one is at 435 °C, the other is at 300–400 °C, as shown in supporting Fig. 1. This demonstrates that O2 makes the decomposition of carbon deposits easier, especially for the aged. 3.4. In-situ DRIFTS analysis In-situ DRIFTS was used to characterize the decomposition of carbon deposits. At different temperatures and in different atmospheres, the decomposition velocities of the carbon deposits on the used catalysts were different. The decomposition velocities under different conditions
Table 2 Decomposition rate of carbon deposits under different conditions. Reaction conditions
Decomposition velocities (peak area vs time)
300 °C pure O2 300 °C N2:O2 = 5:3 300 °C N2:O2 = 3:5 325 °C N2:O2 = 3:5 350 °C N2:O2 = 3:5 350 °C pure N2 500 °C pure N2
4.18 3.81 3.13 3.15 14.48 0 32.84
Fig. 4. Evolution of DRIFTS spectra over used catalyst under 300 °C in O2 atmosphere.
L. Jiang et al. / Catalysis Communications 64 (2015) 22–26
25
Fig. 5. Evolution of DRIFTS spectrum over used catalysts under different temperatures and atmosphere: a—300 °C, N2:O2 = 5:3; b—300 °C, N2:O2 = 3:5; c—325 °C, N2:O2 = 3:5; d—350 °C, N2:O2 = 3:5; e—350 °C, N2; f—500 °C, N2.
are shown in Table 2. As shown in Fig. 4, we can only see the distinct changes of the peaks at 2828–3024 cm−1. So in Fig. 5, we only showed the curves of this region (2828–3024 cm−1) to illustrate the decomposition of carbon deposits. When reaction temperature was kept at 300 °C, the coked catalysts were heated in different atmospheres and the spectra are shown in Figs. 4 and 5a–b. As we can see in Fig. 4, the carbon deposits start to be decomposed around 40 min and fade away around 80 min. When the N2:O2 ratio is changed, the situation of carbon decomposition has no obvious change. In Fig. 5e, the catalyst is heated in pure N2 atmosphere at 350 °C, but the peak near 3000 cm−1 of the carbon deposits has no change, indicating that carbon deposits can be hardly decomposed in pure N2 atmosphere at 350 °C or lower than 350 °C. Keeping the atmosphere (a mixture of N2 and O2 at the ratio of 3:5) unchanged, we changed the temperature from 300 °C to 325 °C and 350 °C. The resultant spectra are shown in Fig. 5b–d. At 300 °C, the carbon deposits start to be decomposed from 35 min and fade away around 80 min. When temperature is changed to 325 °C, the decomposition rate is almost the same as that at 300 °C. However, at 350 °C, the decomposition rate increases significantly. Within 10 min, the peak of the carbon deposits has already disappeared. When the coked catalyst is heated to 500 °C in pure N2 atmosphere, the carbon deposits start to be decomposed at about 45 min and soon disappear. In these experiments, it is found that temperature has a certain influence on the decomposition rate of carbon deposits in atmosphere containing O2. When the reaction temperature is increased from 300 °C to 350 °C, decomposition rate changes a bit. But if the reaction temperature is increased to 500 °C, decomposition rate increases rapidly. It is also found that the regenerated catalysts have the same catalytic activity and BET surface area as fresh catalyst, as shown in Table 3.
Table 3 Catalyst activity and surface area of fresh and deactivated Li3PO4 catalysts. Catalyst
SBET (m2/g)
Conversion/%
Fresh-Li3PO4 Deactivated-Li3PO4 Regenerated-Li3PO4
27.8 2.5 24.3
87.0 33.6 86.3
4. Conclusion Carbon deposits formed on the surface of basic lithium phosphate are mainly in the form of CxHy and CxHyO. With the presence of O2, most of the carbon deposits could be removed at about 300 °C. And when the temperature is improved, the decomposition rate changes a bit. But the decomposition of carbon deposits in N2 atmosphere needs a higher temperature, nearly 500 °C. Hence, temperature and atmosphere are important factors influencing the decomposition of carbon deposits. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.catcom.2015.01.026. Acknowledgments The authors are grateful to the Natural Science Foundation (Nos. 21276127 and 51106076). References [1] M.W.C. Robinson, K.S. Pillinger, I. Mabbett, D.A. Timms, A.E. Graham, Copper(II) tetrafluoroborate-promoted Meinwald rearrangement reactions of epoxides, Tetrahedron 66 (2010) 8377–8382. [2] E. Erturk, M. Gollu, A.S. Demir, Efficient rearrangement of epoxides catalyzed by a mixed-valent iron trifluoroacetate [Fe3O(O2CCF3)6(H2O)3], Tetrahedron 66 (2010) 2373–2377. [3] K.A. Bhatia, K.J. Eash, N.M. Leonard, M.C. Oswald, R.S. Mohan, A facile and efficient method for the rearrangement of aryl-substituted epoxides to aldehydes and ketones using bismuth triflate, Tetrahedron Lett. 42 (2001) 8129–8132. [4] G. Neri, G. Rizzo, C. Crisafulli, L. De Luca, A. Donato, M.G. Musolino, R. Pietropaolo, Isomerization of a-pinene oxide to campholenic aldehyde over Lewis acids supported on silica and titania nanoparticles, Appl. Catal. A Gen. 295 (2005) 116–125. [5] G.K. Surya Prakash, T. Mathew, S. Krishnaraj, E.R. Marinez, G.A. Olah, Nafion-H catalysed isomerization of epoxides to aldehydes and ketones, Appl. Catal. A Gen. 181 (1999) 283–288. [6] P.C. Brookes, D.J. Milne, P.J. Murphy, B. Spolaore, Epoxides rearrangements using dilithiated aminoalcohols as chiral bases, Tetrahedron 58 (2002) 4675–4680. [7] A. Seki, M. Asami, Catalytic enantioselective rearrangement of meso-epoxides mediated by chiral lithium amides in the presence of excess cross-linked polymer-bound lithium amides, Tetrahedron 58 (2002) 4655–4663. [8] P. Dinér, Catalytic asymmetric chiral lithium amide-promoted epoxide rearrangement: a NMR spectroscopic and kinetic investigation, Tetrahedron Asymmetry 21 (2010) 2733–2739.
26
L. Jiang et al. / Catalysis Communications 64 (2015) 22–26
[9] S.K. Bertilsson, P.G. Andersson, Asymmetric base-promoted epoxide rearrangement: achiral lithium amides revisited, Tetrahedron 58 (2002) 4665–4668. [10] S.E. De Sousa, P. O'Brien, H.C. Steffens, A new norephedrine-derived chiral base for epoxide rearrangement reactions, Tetrahedron Lett. 40 (1999) 8423–8425. [11] A. Magnus, S.K. Bertilsson, P.G. Andersson, Asymmetric base-mediated epoxide isomerisation, Chem. Soc. Rev. 31 (2002) 223–229. [12] P. O'Brien, P. Poumellec, Chiral base-mediated rearrangement of meso-cyclohexene oxides to allylic alcohols, Tetrahedron Lett. 37 (1996) 8057–8058. [13] P.B. Tivola, A. Deagostino, C. Prandi, P. Venturello, Alkoxydienylstannanes via metalation of α, β-unsaturated and α-phenyl acetals: preparation and synthetic uses in the Stille cross-coupling reaction, J. Chem. Soc., Perkin Trans. 1 4 (2001) 437–441. [14] US Pat., 6 803 491, 2004. [15] R.M. Rojas, J.L. Martin De Vidales, A. Delgado, J.V. Sinisterra, Microstructural and thermal characterization of basic and stoichiometric lithium phosphates, in relation with their catalytic activity, J. Solid State Chem. 106 (1993) 237–252. [16] W.H. Ma, W. Si, W. Wu, Q. Zhong, Structures and catalytic properties of lithium phosphates, Catal. Lett. 141 (2011) 1032–1036. [17] A. Fonseca, P. Zeuthen, J.B. Nagy, Assignment of an average chemical structure to catalyst carbon deposits on the basis of quantitative 13C n.m.r. Spectra, Fuel 75 (1996) 1413–1423. [18] J. Sirita, S. Phanichphant, F.C. Meunier, Quantitative analysis of adsorbate concentrations by diffuse reflectance FT-IR, Anal. Chem. 79 (2007) 3912.
[19] S.Y. Foo, C.K. Cheng, T.H. Nguyen, et al., Carbon deposition and gasification kinetics of used lanthanide-promoted Co–Ni/Al2O3 catalysts from CH4 dry reforming, Catal. Commun. 26 (2012) 183–188. [20] D.A. Netzel, F.P. Miknis, J.M. Mitzel, T.J. Zhang, P.D. Jacobs, H.W. Haynes Jr., Carbon13 solid-state n.m.r. investigation of coke deposits on spent catalysts used in coal liquefaction, Fuel 15 (1996) 1397–1405. [21] L.W. Jia, J. Zhang, M.Q. Shen, et al., The influence of fuel type on carbon deposition over precious metal containing motorcycle catalysts, Catal. Commun. 6 (2005) 757–761. [22] R. Hughes, Deactivation of Catalysts, Science Press, Beijing, 1990. [23] Z. Paál, R. Schlögl, G. Ertl, Photoelectron spectroscopy of polycrystalline platinum catalysts, J. Chem. Soc. Faraday Trans. 88 (1992) 1179–1189. [24] U. Wild, N. Pfänder, R. Schlögl, J. Freseniu, Species analysis of automotive carbon particles: application of XPS for integral analysis of filter samples, J. Anal. Chem. 357 (1997) 420–428. [25] N.M. Rodriguez, P.E. Anderson, A. Wootsch, U. Wild, R. Schlögl, Z. Paál, XPS, EM, and catalytic studies of the accumulation of carbon on Pt black, J. Catal. 197 (2001) 365–377. [26] G. Ertl, H. Knözinger, J. Weitkamp, Handbook of Heterogeous Catalysis, Wiley, New York, 1997.
Contents lists available at ScienceDirect
Catalysis Communications journal homepage: www.elsevier.com/locate/catcom
Short communication
Characterization of carbon deposits on coked lithium phosphate catalysts for the rearrangement of propylene oxide Li Jiang, Hao Li, Yanan Wang, Weihua Ma ⁎, Qin Zhong Department of Chemical Engineering, Nanjing University of Science and Technology, Nanjing 210094, PR China
a r t i c l e
i n f o
Article history: Received 6 November 2014 Received in revised form 25 January 2015 Accepted 27 January 2015 Available online 28 January 2015 Keywords: Rearrangement Propylene oxide Regeneration Carbon deposits
a b s t r a c t Basic lithium phosphate catalysts for the rearrangement of propylene oxide can be easily deactivated due to carbon deposits. In-situ DRIFTS, TG–DTG, NMR and XPS were used to study carbon deposits. Results showed that carbon deposits were mainly in the form of CxHy and CxHyO, which could be decomposed in a mixture of N2 and O2 atmosphere at higher than 300 °C or in N2 atmosphere near 500 °C. A shorter period of time was needed in the atmosphere containing O2. © 2015 Published by Elsevier B.V.
1. Introduction
2. Experimental
The rearrangement of epoxides to allylic alcohols or carbonyl compounds is a fundamentally important transformation with significant synthetic and industrial utility. A wide range of catalysts are associated with this reaction, and this transformation has been attracting great attentions [1–3]. The constitution of the rearrangement products was influenced by the acidity or basicity of the catalysts [4–6]. Lithium compounds have been attractive candidates as catalysts for the rearrangement of epoxides. For liquid phase catalysis or asymmetric catalysis [7–10], lithium compound bases [6,11,12] or Schlosser's Li/K mixed superbases [13] are often applied in the arrangement of functionalized, aryl or chiral epoxides into allylic alcohols. For epoxides with small molecule, especially propylene oxide (in abbreviation, PO), basic lithium phosphate is often used to catalyze the rearrangement to allyl alcohols [14,15]. There are two processes for the rearrangement of PO to allyl alcohol, gas–solid phase and slurry phase. Gas–solid phase process is much simpler and cleaner than the other, but carbon deposition has always been a major problem. Therefore, it is important to analyze the carbon deposits on the deactivated catalysts. In our previous work, Li3PO4 catalysts gradually got deactivated as reaction time increased and its BET surface area decreased obviously due to carbon (or coke) deposition [16]. This investigation is mainly focused on carbon deposits concurrent with the decomposition or removal of carbon deposits.
2.1. Catalyst preparation
⁎ Corresponding author. E-mail address: [email protected] (W. Ma).
http://dx.doi.org/10.1016/j.catcom.2015.01.026 1566-7367/© 2015 Published by Elsevier B.V.
Lithium hydroxide monohydrate (20.14 g) was dissolved in 125 mL of de-ionized water at 65 °C in a 500 mL three-necked flask with a stirrer. A solution prepared from sodium phosphate dodecahydrate (45.61 g) and water (125 mL) was dropped into the flask with stirring for 1 h. After the mixture was ripened for 3 h, the resultant white precipitate was collected by filtration, dried at 120 °C for 12 h and then calcined at 320 °C for 8 h. Catalytic isomerization reactions were conducted in a φ10 mm fixed-bed reactor at 290 °C using a mixture of N2 and vaporized propylene oxide stream, with weight hourly space velocity (WHSV) value of about 9 h−1 [17]. A series of used or coked lithium phosphate catalysts were obtained. 2.2. Catalyst characterization Used or coked catalysts were characterized by thermo gravimetric analysis (TG, Mettler-Toledo TGA/SDTA851e) at the rate of 20 °C/min in both nitrogen (99.999%, 30 mL/min) and air atmosphere. XPS experiments were carried out on a RBD upgraded PHI-5000C ESCA system (Perkin Elmer). All binding energies were referenced to the C1s neutral carbon peak at 284.6 eV. 13C CP-MAS-NMR spectra were performed on a Bruker DSX400 spectrometer. Tetramethylsilane was used as external reference for the chemical shifts (δ). In situ DRIFTS was used to investigate the decomposition of the carbon deposits at different temperatures. Pure KBr was used as background and the mixture of KBr and the used catalysts (3:1) was used
L. Jiang et al. / Catalysis Communications 64 (2015) 22–26
23
Fig. 1. 13C MAS-NMR (a) and DRIFTS spectra (b) of used catalysts (24 h).
as samples. The spectra were recorded on a Nicolet iS10 spectrometer equipped with an iZ10 Auxiliary Module, which consists of a liquidN2-cooled high-sensitivity MCT-A detector and a DRIFTS cell (SpectraTech) with KBr windows. All spectra were recorded at an accumulation of 20 scans with a resolution of 4 cm−1. The spectra began to be recorded as soon as the sample was heated in pure N2 gas flow. When the temperature reached the set value, O2 was introduced into the cell. The IR data are reported as log 1 / R, where R is the sample reflectance. The function log 1 / R (=“absorbance”) gives a better linear representation of the band intensity against sample surface coverage than that given by the Kubelka–Munk function for strongly absorbing media such as those based on oxides [18]. 3. Results and discussions 3.1. DRIFTS and NMR analysis In order to analyze the types of carbon deposits, DRIFTS and 13C CPMAS-NMR were used to characterize carbon deposits on the used catalyst for 24 h on-stream [19], as shown in Fig. 1. The assignment of the peaks shown in Fig. 1a is given in Table 1. Seen from Fig. 1a and Table 1, there are mainly four groups in carbon deposits, − CH3, − CH2, C–O–C or C–OH, and − C6H5. In the DRIFTS spectrum, the broad peaks at 2828–3024 cm−1 are mainly attributed to symmetrical and asymmetrical C–H stretching vibrations. They are characteristics of alkyl groups of the carbon deposits on catalyst surface. Judged by the peak of δCH at 1449 cm− 1, alkyl species with three or more carbon atoms in the alkyl chains may also exist. Also, olefinic species may exist in the used catalyst because their characteristic stretching vibration frequencies appear at 3000–3024 cm−1. DRIFTS analysis shows that the composition of carbon deposits is very complicated. Carbon deposits mainly consist of both aliphatic and aromatic hydrocarbons containing oxygen element. Other types of
carbon, such as olefinic species, graphite-like carbon, are also possible. Their chemical shifts or peaks may be merged with those of the main four groups into the broad peaks at 34.5, 74.9, and 127.5 ppm shown in Fig. 1a. The − CHO group, with the characteristic peak at 1705 cm−1 in Fig. 1b, can be seen at 210 ppm in Fig. 1a.
3.2. XPS analysis Fig. 2 shows the C1s XPS curves of the fresh and used catalysts. For the fresh catalyst, there are three distinct peaks. The peak at 289.4 eV on the surface of basic lithium can be attributed to C_O bonds of CO2− 3 phosphates [20]. The peak at 284.7 eV indicates the presence of carbon contamination. In the curve of the deactivated catalyst, there are three main distinct peaks, the peak at 283.8 eV indicates the presence of graphitic carbon species in the form of CxHy [21–24]. The peak at 284.6 eV indicates the presence of carbon contamination. And the peak at 285.8 or 285.4 and 285.6 eV can be attributed to the C–O bonds. Moreover, there is a small peak at 288.9 or 288.7 and 288.8 eV, which is due to C_O bonds in species like CxHyO [25,26], a kind of carbon deposit containing O element. The binding energy of C_O bonds in CxHyO species is lower than that in CO23 − groups, since CO23 − disappeared after the catalyst was used for 24 h on-stream at 300 °C.
Table 1 13 C NMR chemical shifts and their possible assignments [18,19]. δ (ppm)
Possible carbon types
18.2 23.0 34.4 39.1 75.0 127.5 134.2 141.7 210.0
CH3 end group to a long paraffinic chain CH2 penultimate carbon in a long paraffinic chain CH and CH2 α to an aromatic ring CH α to the paraffinic chain in alicyclic group CH2 bound to an alkoxyl Aromatic carbons bound to hydrogen Aromatic carbon bound to CH3 Aromatic carbon bound to substituted groups excluding CH3 −CHO
Fig. 2. XPS curves of fresh and used catalysts.
24
L. Jiang et al. / Catalysis Communications 64 (2015) 22–26
Fig. 3. TG–DTG curves of fresh and used catalysts in N2 atmosphere.
3.3. TG–DTG analysis TG–DTG was used to study the decomposition of carbon deposits over different reaction times. The curves in N2 are shown in Fig. 3. For the fresh catalyst, there is a small change in mass (weight loss ratio was about 1.2%) as shown in Fig. 3a. It is due to desorption of water and carbon dioxide adsorbed on catalyst surface. When it turns to the used catalysts, weight loss ratios of the catalysts used for 2 h, 4 h, 6 h and 24 h are 12.5%, 15.4%, 20.5%, and 21.9%, respectively. There is a distinct step down in mass during the temperature range of 300–500 °C, and the weight loss ratios of the catalysts used for 2 h, 4 h, 6 h and 24 h are 11.6%, 10.5%, 15.2%, and 17.6%, respectively. This step is probably caused by the decomposition of hydrocarbon, i.e., carbon deposits. For the catalyst used for 2 h, except the decomposition peak with Tmax = 420 °C there is another shoulder peak at 300–400 °C, which corresponds to a kind of young coke of low molecular weight or with high H/C. However, for the 4 h, there are a main peak at 390 °C and a small broad peak at 500 °C, which correspond to older coke with low H/C. For the 6 h and 24 h, it can be seen that there are other weight changes in mass at higher than 700 °C in TG curves, which correspond to the peaks with Tmax = 728 °C in DTG curves. This may be the decomposition of the aged coke with much lower H/C or the slow oxidation of the graphite-like carbon. Therefore, it can be concluded that carbon deposits formed over different reaction times, especially less than 6 h,
are different. The carbon deposits formed within 2 h are of lower molecular weight and can be easily decomposed. But the carbon deposits formed within 4 h are of higher molecular weight and a bit more difficult to be decomposed than that of 2 h. For the samples used for 6 h and 24 h, part of the carbon deposits can only be removed at about 728 °C. However, the decomposition peaks of the carbon deposits under air atmosphere are all below 450 °C, one is at 435 °C, the other is at 300–400 °C, as shown in supporting Fig. 1. This demonstrates that O2 makes the decomposition of carbon deposits easier, especially for the aged. 3.4. In-situ DRIFTS analysis In-situ DRIFTS was used to characterize the decomposition of carbon deposits. At different temperatures and in different atmospheres, the decomposition velocities of the carbon deposits on the used catalysts were different. The decomposition velocities under different conditions
Table 2 Decomposition rate of carbon deposits under different conditions. Reaction conditions
Decomposition velocities (peak area vs time)
300 °C pure O2 300 °C N2:O2 = 5:3 300 °C N2:O2 = 3:5 325 °C N2:O2 = 3:5 350 °C N2:O2 = 3:5 350 °C pure N2 500 °C pure N2
4.18 3.81 3.13 3.15 14.48 0 32.84
Fig. 4. Evolution of DRIFTS spectra over used catalyst under 300 °C in O2 atmosphere.
L. Jiang et al. / Catalysis Communications 64 (2015) 22–26
25
Fig. 5. Evolution of DRIFTS spectrum over used catalysts under different temperatures and atmosphere: a—300 °C, N2:O2 = 5:3; b—300 °C, N2:O2 = 3:5; c—325 °C, N2:O2 = 3:5; d—350 °C, N2:O2 = 3:5; e—350 °C, N2; f—500 °C, N2.
are shown in Table 2. As shown in Fig. 4, we can only see the distinct changes of the peaks at 2828–3024 cm−1. So in Fig. 5, we only showed the curves of this region (2828–3024 cm−1) to illustrate the decomposition of carbon deposits. When reaction temperature was kept at 300 °C, the coked catalysts were heated in different atmospheres and the spectra are shown in Figs. 4 and 5a–b. As we can see in Fig. 4, the carbon deposits start to be decomposed around 40 min and fade away around 80 min. When the N2:O2 ratio is changed, the situation of carbon decomposition has no obvious change. In Fig. 5e, the catalyst is heated in pure N2 atmosphere at 350 °C, but the peak near 3000 cm−1 of the carbon deposits has no change, indicating that carbon deposits can be hardly decomposed in pure N2 atmosphere at 350 °C or lower than 350 °C. Keeping the atmosphere (a mixture of N2 and O2 at the ratio of 3:5) unchanged, we changed the temperature from 300 °C to 325 °C and 350 °C. The resultant spectra are shown in Fig. 5b–d. At 300 °C, the carbon deposits start to be decomposed from 35 min and fade away around 80 min. When temperature is changed to 325 °C, the decomposition rate is almost the same as that at 300 °C. However, at 350 °C, the decomposition rate increases significantly. Within 10 min, the peak of the carbon deposits has already disappeared. When the coked catalyst is heated to 500 °C in pure N2 atmosphere, the carbon deposits start to be decomposed at about 45 min and soon disappear. In these experiments, it is found that temperature has a certain influence on the decomposition rate of carbon deposits in atmosphere containing O2. When the reaction temperature is increased from 300 °C to 350 °C, decomposition rate changes a bit. But if the reaction temperature is increased to 500 °C, decomposition rate increases rapidly. It is also found that the regenerated catalysts have the same catalytic activity and BET surface area as fresh catalyst, as shown in Table 3.
Table 3 Catalyst activity and surface area of fresh and deactivated Li3PO4 catalysts. Catalyst
SBET (m2/g)
Conversion/%
Fresh-Li3PO4 Deactivated-Li3PO4 Regenerated-Li3PO4
27.8 2.5 24.3
87.0 33.6 86.3
4. Conclusion Carbon deposits formed on the surface of basic lithium phosphate are mainly in the form of CxHy and CxHyO. With the presence of O2, most of the carbon deposits could be removed at about 300 °C. And when the temperature is improved, the decomposition rate changes a bit. But the decomposition of carbon deposits in N2 atmosphere needs a higher temperature, nearly 500 °C. Hence, temperature and atmosphere are important factors influencing the decomposition of carbon deposits. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.catcom.2015.01.026. Acknowledgments The authors are grateful to the Natural Science Foundation (Nos. 21276127 and 51106076). References [1] M.W.C. Robinson, K.S. Pillinger, I. Mabbett, D.A. Timms, A.E. Graham, Copper(II) tetrafluoroborate-promoted Meinwald rearrangement reactions of epoxides, Tetrahedron 66 (2010) 8377–8382. [2] E. Erturk, M. Gollu, A.S. Demir, Efficient rearrangement of epoxides catalyzed by a mixed-valent iron trifluoroacetate [Fe3O(O2CCF3)6(H2O)3], Tetrahedron 66 (2010) 2373–2377. [3] K.A. Bhatia, K.J. Eash, N.M. Leonard, M.C. Oswald, R.S. Mohan, A facile and efficient method for the rearrangement of aryl-substituted epoxides to aldehydes and ketones using bismuth triflate, Tetrahedron Lett. 42 (2001) 8129–8132. [4] G. Neri, G. Rizzo, C. Crisafulli, L. De Luca, A. Donato, M.G. Musolino, R. Pietropaolo, Isomerization of a-pinene oxide to campholenic aldehyde over Lewis acids supported on silica and titania nanoparticles, Appl. Catal. A Gen. 295 (2005) 116–125. [5] G.K. Surya Prakash, T. Mathew, S. Krishnaraj, E.R. Marinez, G.A. Olah, Nafion-H catalysed isomerization of epoxides to aldehydes and ketones, Appl. Catal. A Gen. 181 (1999) 283–288. [6] P.C. Brookes, D.J. Milne, P.J. Murphy, B. Spolaore, Epoxides rearrangements using dilithiated aminoalcohols as chiral bases, Tetrahedron 58 (2002) 4675–4680. [7] A. Seki, M. Asami, Catalytic enantioselective rearrangement of meso-epoxides mediated by chiral lithium amides in the presence of excess cross-linked polymer-bound lithium amides, Tetrahedron 58 (2002) 4655–4663. [8] P. Dinér, Catalytic asymmetric chiral lithium amide-promoted epoxide rearrangement: a NMR spectroscopic and kinetic investigation, Tetrahedron Asymmetry 21 (2010) 2733–2739.
26
L. Jiang et al. / Catalysis Communications 64 (2015) 22–26
[9] S.K. Bertilsson, P.G. Andersson, Asymmetric base-promoted epoxide rearrangement: achiral lithium amides revisited, Tetrahedron 58 (2002) 4665–4668. [10] S.E. De Sousa, P. O'Brien, H.C. Steffens, A new norephedrine-derived chiral base for epoxide rearrangement reactions, Tetrahedron Lett. 40 (1999) 8423–8425. [11] A. Magnus, S.K. Bertilsson, P.G. Andersson, Asymmetric base-mediated epoxide isomerisation, Chem. Soc. Rev. 31 (2002) 223–229. [12] P. O'Brien, P. Poumellec, Chiral base-mediated rearrangement of meso-cyclohexene oxides to allylic alcohols, Tetrahedron Lett. 37 (1996) 8057–8058. [13] P.B. Tivola, A. Deagostino, C. Prandi, P. Venturello, Alkoxydienylstannanes via metalation of α, β-unsaturated and α-phenyl acetals: preparation and synthetic uses in the Stille cross-coupling reaction, J. Chem. Soc., Perkin Trans. 1 4 (2001) 437–441. [14] US Pat., 6 803 491, 2004. [15] R.M. Rojas, J.L. Martin De Vidales, A. Delgado, J.V. Sinisterra, Microstructural and thermal characterization of basic and stoichiometric lithium phosphates, in relation with their catalytic activity, J. Solid State Chem. 106 (1993) 237–252. [16] W.H. Ma, W. Si, W. Wu, Q. Zhong, Structures and catalytic properties of lithium phosphates, Catal. Lett. 141 (2011) 1032–1036. [17] A. Fonseca, P. Zeuthen, J.B. Nagy, Assignment of an average chemical structure to catalyst carbon deposits on the basis of quantitative 13C n.m.r. Spectra, Fuel 75 (1996) 1413–1423. [18] J. Sirita, S. Phanichphant, F.C. Meunier, Quantitative analysis of adsorbate concentrations by diffuse reflectance FT-IR, Anal. Chem. 79 (2007) 3912.
[19] S.Y. Foo, C.K. Cheng, T.H. Nguyen, et al., Carbon deposition and gasification kinetics of used lanthanide-promoted Co–Ni/Al2O3 catalysts from CH4 dry reforming, Catal. Commun. 26 (2012) 183–188. [20] D.A. Netzel, F.P. Miknis, J.M. Mitzel, T.J. Zhang, P.D. Jacobs, H.W. Haynes Jr., Carbon13 solid-state n.m.r. investigation of coke deposits on spent catalysts used in coal liquefaction, Fuel 15 (1996) 1397–1405. [21] L.W. Jia, J. Zhang, M.Q. Shen, et al., The influence of fuel type on carbon deposition over precious metal containing motorcycle catalysts, Catal. Commun. 6 (2005) 757–761. [22] R. Hughes, Deactivation of Catalysts, Science Press, Beijing, 1990. [23] Z. Paál, R. Schlögl, G. Ertl, Photoelectron spectroscopy of polycrystalline platinum catalysts, J. Chem. Soc. Faraday Trans. 88 (1992) 1179–1189. [24] U. Wild, N. Pfänder, R. Schlögl, J. Freseniu, Species analysis of automotive carbon particles: application of XPS for integral analysis of filter samples, J. Anal. Chem. 357 (1997) 420–428. [25] N.M. Rodriguez, P.E. Anderson, A. Wootsch, U. Wild, R. Schlögl, Z. Paál, XPS, EM, and catalytic studies of the accumulation of carbon on Pt black, J. Catal. 197 (2001) 365–377. [26] G. Ertl, H. Knözinger, J. Weitkamp, Handbook of Heterogeous Catalysis, Wiley, New York, 1997.