Diphenyl carbonate (DPC) is an important chemical intermediate for many organic compounds and polymers such as polycarbonates, p-hydroxybenzoic acid polyesters, and aliphatic monoisocyanates ^{[1, 2]}. In recent years, the synthesis of DPC has received more attention due to the extensive use of polycarbonates ^[3]. The industrial synthesis of DPC is based on the highly toxic phosgene ^[1]. There are also two non-phosgene methods, transesterification ^{[4, 5]} and oxidative carbonylation of phenol ^[6], which are under study. Among these, the oxidative carbonylation of phenol with CO and O₂, shown in Eq. (1), is an attractive method as it is a one-step process with H₂O being the sole byproduct in principle.

The oxidative carbonylation of phenol is a multi-step electron transfer reaction ^[7]. It is normally catalyzed by Pd compounds with various co-catalysts ^{[8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20]}. Vavasori et al. ^[7] studied the reaction using Pd(OAc)₂ or PdBr₂ as catalyst, along with benzoquinone and Co, Mn, and Cu salts as cocatalysts. In particular, the influence of the cocatalysts was discussed. Xue et al. ^[10] prepared an embedded Pd-Cu-O/SiO₂ catalyst using a water-in-oil microemulsion as nano-reactors. Its catalytic performance with Cu(OAc)₂, hydroquinone, and tetrabutylammonium bromide (TBAB) acting as cocatalysts was evaluated. The DPC yield was 35.4% with reduced Pd leaching and improved stability. Ce compounds are effective cocatalysts for the homogenerous Pd-catalyzed oxidative carbonylation of phenol ^[8]. In addition, Ce can also be used as the second metal component or support, in the form of CeO₂, as the heterougeneous catalyst for this reaction. Wu et al. ^[21] prepared a Pd catalyst supported on Mn-doped CeO₂ for the oxidative carbonylation of phenol and obtained a DPC yield of 9.3%. Wang ^[22] reported a Pd-Ce-O/SiO₂ catalyst prepared by microemulsion that gave 53.7% DPC yield under optimized conditions.

CeO₂ with the fluorite structure is a significant rare earth oxide in industrial catalysis. It can act as an oxygen buffer by absorbing and releasing oxygen through a fast Ce³⁺/Ce⁴⁺ cycle involving the participation of the lattice oxygen species. It has been extensively studied and applied in heterogeneous catalysis ^{[23, 24, 25, 26, 27]}. Usually, its unique redox property is dependent on the size of the CeO₂ particles. The morphology also influences the redox feature of CeO₂ significantly through the exposed crystalline planes. A typical example is CeO₂ nanorods with a one-dimensional structure, which showed a higher CO oxidation activity than other conventional nanoparticles due to the exposed (110) crystal surface of CeO₂ ^{[28, 29]}. In addition, CeO₂ nanotubes also show excellent activity for CO oxidation. This was explained by the combination of the large surface-to-volume area with the inner side of the nanotubes providing more active sites ^{[28, 30]}. Therefore, a good catalytic performance for DPC synthesis utilizing CeO₂ with the one-dimensional structure as the support for a heterogeneous Pd catalyst is expected.

In this work, CeO₂ nanotubes (CeO₂-NT) and a Pd-O/CeO₂-NT catalyst were synthesized by the liquid phase deposition-hydrothermal method using carbon nanotubes (CNTs) as the template. The catalytic performance for the oxidative carbonylation of phenol to DPC was evaluated and compared with that using zero-dimensional CeO₂ nanoparticles supporting a Pd catalyst. The effects of the reaction conditions were also studied for the Pd-O/CeO₂-NT catalyst.

A 4 g sample of CNTs (length 0.5-2 μm, outer diameter 30-50 nm, purity > 95%, Chengdu Organic Chemistry Co., Ltd., Chinese Academy of Sciences) was heated in 30% HNO₃ (600 mL) under reflux at 120 °C for 24 h. Then the CNTs were cooled to room temperature. After that, they were separated by filtration and washed with deionized water and anhydrous ethanol repeatedly until pH = 7. The CNTs were dried at 80 °C.

CeO₂-NT was synthesized by the liquid phase deposition ^[31] hydrothermal method using the pretreated CNTs as the template. First, the CNTs (0.54 g) were dispersed in a solution composed of Ce(NO₃)₃ aqueous solution (0.15 mol/L, 20 mL) and anhydrous ethanol (215 mL) with ultrasonic treatment for 2 h. Then, with vigorous stirring, aqueous NaOH solution (0.125 mol/L) was added dropwise into the mixture until the pH value was 10. After stirring for 30 min, the suspension was transferred to a stainless steel autoclave for the hydrothermal process at 110 °C for 24 h. Third, the solid was collected by filtration, repeatedly washed with deionized water and anhydrous ethanol until the pH was 7, and then dried at 80 °C. Finally, the solid was calcined at 500 °C for 4 h in air using a heating rate of 5 °C/min. The sample was denoted as CeO₂-NT.

The preparation of Pd-O/CeO₂-NT was similar to the preparation of CeO₂-NT above. The only difference was that an amount of PdCl₂ in aqueous ammonia, along with the suspension, was also added into the autoclave for the hydrothermal process. The other operation processes were the same as those for the preparation of CeO₂-NT. The obtained sample was denoted as Pd-O/CeO₂-NT with a nominal Pd loading of 1.0 wt%.

For comparison with Pd-O/CeO₂-NT, a Pd-O/CeO₂-P catalyst in which CeO₂ was zero-dimensional particles was prepared using a water-in-oil microemulsion as a nanoreactor. The microemusion consisted of cyclohexane (100 mL), Triton X-100 (10.8 mL), PdCl₂ in aqueous ammonia (0.25 mmol PdCl₂, 10 mL aqueous ammonia), and n-hexanol (2.8 mL). Ce(NO₃)₃ aqueous solution (3 mL, 5.2 mol/L) was added to the microemulsion and hydrolyzed at 30 °C for 5 h. Then, the solution was centrifuged to separate the precipitate, which was thoroughly washed with ethanol and deionized water, dried at 80 °C overnight, and calcined in the air at 500 °C for 4 h. The sample was denoted as Pd-O/CeO₂-P with a nominal Pd loading of 1.0 wt%.

Transmission electron microscope (TEM) images and selected area electron diffraction (SAED) were obtained with a PHILIPS TECNOL 20 at an acceleration voltage of 200 kV. The specific surface areas of the samples were calculated by the BET method by N₂ adsorption-desorption with a Micromeritics ASAP 2020 M+C porosity analyzer. X-ray diffraction (XRD) patterns were recorded on a Rigaku D/Max-2500 X-ray diffractometer with Cu K_α (40 kV, 100 mA) radiation and a secondary beam graphite monochromator (SS/DS = 1°, RS 0.15 mm, Counter SC) at the scanning 2θ range of 5°-90°. Raman spectra were measured by a Renishaw inVia Reflex microspectrometer. The Pd content in the catalyst was determined by a Thermo Scientific iCAP 7400 ICP-OES.

Temperature-programmed reduction by hydrogen (H₂-TPR) was carried out on a Micromeritics Auto Chem II-2920 apparatus. All samples (0.1 g) were pretreated in the flow of Ar (50 mL/min) at room temperature for 5 min. Then the flowing gas was switched to 10% H₂/Ar mixture (50 mL/min) and the sample was heated to 1000 °C at a ramping rate of 10 °C/min. The H₂ consumption was monitored by a thermal conductivity detector (TCD).

Oxidative carbonylation of phenol was performed in a 50-mL Teflon-lined stainless steel autoclave with a magnetic stirrer. In a typical experiment, the catalyst and co-catalyst (Cu(OAc)₂, tetrabutylammonium bromide, and hydroguinone) along with phenol and the solvent (dichloromethane) and 4A molecular sieve were introduced into the autoclave. Then, the autoclave was pressurized with a mixture of CO and O₂. At the end of the reaction, the autoclave was cooled to room temperature and vented. The catalyst and 4A molecular sieve were removed by filtration.

The product mixture was identified qualitatively using a Thermo Fisher TRACE DSQ GC-MS with a BPX5 (30 m × 0.25 mm × 0.25 μm) capillary column. The injection port temperature was 250 °C. The column temperature was programed and the initial column temperature was fixed at 40 °C for 3 min and then increased to 250 °C at a rate of 10 °C/min. The final temperature was maintained for 5 min. The carrier gas was He at a constant flow of 1 mL/min. A mass spectrometer was used in the electron impact (EI) mode at 200 °C.

A Knauer K2600 HPLC was used to analyze the reaction liquid quantitatively. The analytical conditions were as follows: a Venusil XBP C18 (4.6 mm × 150 mm, 5 μm) chromatographic column was used at a flow rate of 0.6 mL/min. CH₃OH/H₂O (65/35, V/V) was used as the mobile phase and the detection wavelength was 254 nm. The column temperature was kept at 30 °C and the injection was 20 μL. Quantitative analysis of diphenyl carbonate and phenol was carried out using the external standard method.

TEM images of CeO₂ and the supported Pd catalysts were recorded and shown in Fig. 1. Figures 1(a) and 1(b) are images of CeO₂-NT, which displayed the outer diameter of 25 nm and lengths of more than 300 nm with open ends. The tube wall was composed of CeO₂ particles with the size of 4-9 nm. The selected area electron diffraction (SAED) pattern of CeO₂-NT is also shown in Fig. 1(b). Its ring diffraction indicated the polycrystalline structure of CeO₂-NT. Moreover, the rings were not smooth with clear spots, which was due to the small amount of CeO₂ crystals in the small selected area diffraction electron beam. The TEM image of Pd-O/CeO₂-NT is shown in Fig. 1(c). The tubular structure of CeO₂ was retained. However, it was difficult to distinguish Pd particles from the CeO₂ support. This may be attributed to the strong interaction between Pd and CeO₂, which inhibited the growth of the Pd particles. On the other hand, the high electron density of CeO₂ also adversely affected finding Pd particles in the TEM image ^[32]. Boronin et al. ^[33] indicated that most of the Pd particles on a CeO₂ support could not be observed, even when HR-TEM was utilized. Figure 1(d) is the TEM image of Pd-O/CeO₂-P prepared by the microemulsion method. The sizes of the zero-dimensional CeO₂ particles were 7 to 15 nm. Similarly, Pd particles could not be observed.

Fig. 1. TEM images of CeO2 and supported Pd catalyst. (a, b) CeO2-NT; (c) Pd-O/CeO2-NT; (d) Pd-O/CeO2-P; Inset of (b): SAED pattern of CeO2-NT.

The XRD patterns of Pd-O/CeO₂-NT and Pd-O/CeO₂-P are shown in Fig. 2. There was no significant difference between the two patterns. The peaks at 2θ = 28.48°, 33.12°, 47.59°, 56.46°, 59.07°, 69.44°, 76.75°, and 79.16° were attributed, respectively, to the diffraction of the (111), (200), (220), (311), (222), (400), (331), and (420) planes of CeO₂ with the cubic structure (JCPDS 65-5923). Diffraction peaks for Pd compounds were not found in XRD patterns, possibly because of the high dispersion of the Pd particles of low weight content on the CeO₂ support ^[34].

Fig. 2. XRD patterns of the Pd-O/CeO2 catalysts.

The lattice parameter of CeO₂ in Pd-O/CeO₂-NT (0.5402 nm) was smaller than that in Pd-O/CeO₂-P (0.5406 nm), which is shown in Table 1. This indicated that more Ce⁴⁺ ions were substituted by Pd²⁺ or Pd⁴⁺ ions in Pd-O/CeO₂-NT. Because the radius of the Pd²⁺ ion (0.84 Å) or Pd⁴⁺ ion (0.62 Å) is smaller than that of Ce⁴⁺ (0.99 Å), Pd-O/CeO₂-NT has a small lattice parameter.

Table 1 Textural properties and catalytic performance of Pd-O/CeO2 for oxidative carbonylation of phenol. Catalyst ABET (m2/g) Pore size (nm) Pore volume (cm3/g) Lattice parameter (nm) Phenol conversion (%) DPC selectivity (%) Pd-O/CeO2-NT 108.8 14.7 0.4 0.5402 53.9 70.9 Pd-O/CeO2-P 93.8 5.9 0.2 0.5406 25.0 30.5 Reaction conditions: 100 °C, 8 h, O2 0.6 MPa, CO 6.6 MPa, phenol 25.5 mmol, 20 mL CH2Cl2, 0.4 g catalyst (Pd 0.04 mmol), 4.0 g 4A molecular sieve, Pd/Cu(OAc)2/TBAB/H2BQ = 1/5/40/40 in molar ratio.

Table 1 Textural properties and catalytic performance of Pd-O/CeO2 for oxidative carbonylation of phenol.

The surface area and porosity of the Pd-O/CeO₂-NT and Pd-O/CeO₂-P catalysts were investigated by N₂ adsorption-desorption at -196 °C. As displayed in Fig. 3, both samples exhibit type IV adsorption isotherms with hysteresis loops, which are due to capillary condensation in mesoporous pores ^[35]. The hysteresis loops of Pd-O/CeO₂-NT and Pd-O/CeO₂-P were H3 type and H2 type, respectively, which suggested different pore structures. Figure 3 also shows that the pore size distribution of Pd-O/CeO₂-NT was in a relatively large scale and the pores of Pd-O/CeO₂-P were small and uniform in the mesoporous range. Moreover, the specific surface area of Pd-O/CeO₂-NT (108.8 m²/g) was larger than that of Pd-O/CeO₂-P (93.8 m²/g), which was due to its unique tubular structure and small CeO₂ particles.

Fig. 3. N2 adsorption-desorption isotherms (a) and pore size distribution (b) of the Pd-O/CeO2 catalyst.

The two CeO₂ supported Pd catalysts were evaluated for their activity in the oxidative carbonylation of phenol to DPC. The results are shown in Table 1. The Pd-O/CeO₂-NT catalyst was superior to Pd-O/CeO₂-P in both phenol conversion and DPC selectivity.

To explain the difference between the catalytic performance of the two catalysts, H₂-TPR of the catalysts and CeO₂ was carried out. The results are shown in Fig. 4 and Table 2. Two reduction peaks were observed at 478 and 810 °C over CeO₂-P. For CeO₂, the low temperature peak at 400-600 °C was due to the reduction of surface oxygen species attached to surface Ce⁴⁺ ions in an octahedral coordination and the high temperature peak over 700 °C was due to the reduction of oxygen anion bonded to two Ce⁴⁺ ions in the bulk phase ^{[36, 37]}. In the TPR curve of CeO₂-NT, there was an extra strong peak at 307 °C in addition to the surface oxygen peak at 430 °C and the bulk phase oxygen peak at 815 °C. The H₂ consumption corresponding to the reduction peaks (Table 2) suggested that the bulk oxygen contents in CeO₂-P and CeO₂-NT were almost the same. However, the surface oxygen content in CeO₂-NT was much lower than that in CeO₂-P. However, if the H₂ consumption of the reduction peak at 307 °C was taken into account, the surface oxygen content of CeO₂-NT (503.1 mmol/g) approximately equals to that of CeO₂-P (515.4 mmol/g). Thus, the peak at 307 °C was ascribed to a kind of surface oxygen species that can be reduced more easily. The new surface oxygen species was denoted as surface O-II. Its generation is due to the special tubular structure of CeO₂-NT. Furthermore, this implies that the surface oxygen on CeO₂-NT can participate in the catalytic reaction more readily than CeO₂-P. Zhou et al. ^[30] prepared a kind of highly reducible CeO₂ nanotubes and found that its surface oxygen reduction started at 200 °C. Shan et al. ^[38] also found surface oxygen species that could be reduced at 180 °C in their 3D CeO₂. They attributed this to the increase of surface oxygen vacancies on CeO₂, which can activate molecular oxygen easily, which then resulted in more adsorbed oxygen species that can be easily reduced. Rao ^[39] believed that the reduction temperature of the CeO₂ surface oxygen was related to the coordination number of O^2-. The less coordinated oxygen species are reduced at a lower temperature than those with a higher coordination.

Fig. 4. TPR profiles of CeO2 and Pd-O/CeO2. (1) CeO2-P; (2) CeO2-NT; (3) Pd-O/CeO2-P; (4) Pd-O/CeO2-NT.

Table 2 H2-TPR results of CeO2 and Pd-O/CeO2. Sample Peak temperature (°C) H2 uptake (mmol/g) Mixed O* Surface O-II Surface O Bulk O Mixed O Surface O-II Surface O Bulk O CeO2-P — — 478 810 — — 515.4 663.7 CeO2-NT — 307 430 815 — 226.8 276.3 663.5 Pd-O/CeO2-P 200 — 432 778 520.2 — 75.8 448.9 Pd-O/CeO2-NT — 96 299 785 — 603.2 747.5 570.9 * Sum of redution of PdO and surface O.

Table 2 H2-TPR results of CeO2 and Pd-O/CeO2.

The TPR profiles of the CeO₂ supported Pd catalyst changed significantly compared with those of CeO₂. The surface oxygen peak became very weak with a H₂ consumption of 75.8 mmol/g in the TPR curve of Pd-O/CeO₂-P. Moreover, there were two other reduction peaks at 200 and 778 °C. The latter was ascribed to the reduction of the bulk oxygen of CeO₂. The former was considered as the sum of the reduction of PdO and surface oxygen ^[40]. Since the reduction temperature of PdO is 50 °C ^[41], the interaction between PdO and CeO₂-P hindered the reduction of PdO ^[40]. This would also account for the decrease of the reduction temperature of the CeO₂ surface oxygen species from 478 to 200 °C.

In the Pd-O/CeO₂-NT the bulk oxygen reduction temperature was 785 °C, and the surface oxygen reduction temperature decreased to 299 °C. There was another strong reduction peak at 96 °C. Furthermore, a reverse peak at 65 °C indicated that the sample released H₂. This can be explained by the fact that metal Pd can adsorb H₂ to form a hydride (PdH_x) at room temperature under 0.013 atm H₂ pressure ^[42]. PdH_x decomposed to release H₂ at an elevated temperature. This also revealed that the Pd species on the Pd-O/CeO₂-NT surface could be reduced at an even lower temperature. Cargnello et al. ^[43] found that PdO/CeO₂ prepared by impregnation could be reduced to Pd/CeO₂ at room temperature. This was also the case of TiO₂ supported Pd catalyst ^[40]. This was attributed to the highly dispersed Pd species on the support. This also suggested that the peak at 96 °C corr esponded to the reduction of surface O-II only and did not include the reduction of PdO. It can be suggested that the decrease of the surface oxygen reduction temperature in Pd-O/CeO₂-NT was the result of hydrogen spillover ^[41]. From the data in Table 2, it can be found that the hydrogen consumption by the surface oxygen species on Pd-O/CeO₂-NT increased (1350.7 mmol/g) compared with those of CeO₂-NT (503.1 mmol/g) and Pd-O/CeO₂-P (596.0 mmol/g). This can be assigned to the incr e ase of oxygen vacancie s on the surface of Pd-O/CeO₂-NT. Figure 5 shows the Raman spectra of the catalysts with a 325-nm excitation laser line. It can be seen that all the catalysts have two distinct bands at 449 and 581 cm^-1. The former is ascribed to the Raman active F_2g mode vibration of CeO₂, and the latter is attributed to oxygen vacancies ^[44]. The peak areas of the two bands were calculated and the results shown in Fig. 5, which are denoted as A449 and A581, respectively. The ratio of A581/A449 revealed the concentration of oxygen vacancies in the catalyst. It can be seen that the A581/A449 value for Pd-O/CeO₂-NT (0.58) was higher than that for Pd-O/CeO₂-P (0.43), reflecting that there was a higher concentration of oxygen vacancies in the Pd-O/CeO₂-NT catalyst. According to the XRD result, there were more Ce⁴⁺ substituted by Pd²⁺/Pd⁴⁺ in Pd-O/CeO₂-NT, which led to more oxygen vacancies ^[45]. O₂ can be adsorbed on oxygen vacancies to generate the oxygen species that can be reduced easily ^[38]. Furthermore, the H₂ consumption may also be generated by the spillover of hydrogen atoms, which originated from H₂ dissociative adsorption on the highly dispersed metallic Pd onto the s u rface of CeO₂.

Fig. 5.Raman spectra of Pd-O/CeO2-NT (a) and Pd-O/CeO2-P (b).

According to the above TPR results, the Pd-O/CeO₂-NT catalyst showed a high phenol conversion because of increased surface oxygen species. Meanwhile, there was a strong interaction between the highly dispersed Pd and CeO₂. This favored electron transfer from Pd(0) to Ce(IV); therefore the active sites can be regenerated to Pd(II), which promoted the synthesis of DPC and increased its selectivity.

The influence of reaction time on the oxidative carbonylation of phenol over the Pd-O/CeO₂-NT catalyst was studied. The results are shown in Fig. 6. Phenol conversion increased quickly from 23.3% to 58.2% as the reaction time was extended from 5 to 8 h. However, the DPC selectivity changed in a different way from the phenol conversion. The DPC selectivity was 96.3% when the reaction time was 5 h. With the prolonging of time, the DPC selectivity gradually decreased. In particular, there was an obvious decrease from 93.6% to 86.8% when the time was changed from 7 to 8 h. During the reaction, phenol can be oxidized by O₂ to benzoquinone and dimeric and trimeric compounds and polymers, which was proved by the brown reaction solution. The polymers can accumulate on the catalyst surface and cover the active sites. Therefore, DPC synthesis was decreased when the reaction time was increased. However, the side reactions were less influenced and phenol conversion still increased. As a result, the DPC selectivity decreased gradually with reaction time. Moreover, DPC can react with H₂O, another product of the oxidative carbonylation reaction, to give the formation of phenol and CO₂. This also decreased the DPC selectivity.

Fig. 6.Effect of reaction time on oxidative carbonylation of phenol. Reaction conditions: 100 °C, O2 0.6 MPa, CO 6.6 MPa, phenol 25.5 mmol, 20 mL CH2Cl2, 0.4 g catalyst (Pd 0.04 mmol), 4.0 g 4A molecular sieve, Pd/Cu(OAc)2/TBAB/H2BQ = 1/5/40/40 in molar ratio.

The effect of reaction temperature on the oxidative carbonylation of phenol over Pd-O/CeO₂-NT catalyst is shown in Fig. 7. Phenol conversion gradually increased from 17.7% to 60.7% as the reaction temperature increased from 80 to 110 °C. Meanwhile, DPC selectivity increased from 88.1% to 95.3%. With further temperature elevation to 120 °C, the phenol conversion changed little and DPC selectivity decreased to 85.7%. The oxidative carbonylation of phenol is an exothermic reaction. A high temperature would retard the reaction by decreasing the equilibrium constant. However, from the viewpoint of reaction kinetics, a high temperature accelerates the reaction rate. Therefore, there is an optimum reaction temperature for the oxidative carbonylation of phenol. Furthermore, phenol is easily oxidized by O₂ at a high temperature to generate polymer sideproducts, which would cover the active sites of the catalyst and prevent the formation of DPC. Then the DPC selectivity decreased.

Fig. 7.Effect of reaction temperature on oxidative carbonylation of phenol. Reaction conditions: 7 h, O2 0.6 MPa, CO 6.6 MPa, phenol 25.5 mmol, 20 mL CH2Cl2, 0.4 g catalyst (Pd 0.04 mmol), 4.0 g 4A molecular sieve, Pd/Cu(OAc)2/TBAB/H2BQ = 1/5/40/40 in molar ratio.

Figure 8 shows the effect of reaction pressure on the oxidative carbonylation of phenol. The partial pressure ratio of CO/O₂ was kept at 11/1 to ensure that the CO concentration was beyond the explosive limit. Phenol conversion increased from 41.3% to 60.7% as the CO pressure changed from 4.4 to 6.6 MPa. After that, it remained unchanged with the further increase of CO pressure. The DPC selectivity decreased from 97.2% to 87.7% when the CO pressure increased from 4.4 to 7.7 MPa. In a volume-reduced reaction with gaseous reactants, such as the oxidative carbonylation of phenol, a higher pressure can accelerate the reaction. CO adsorption on the catalyst is also promoted at a higher pressure with the result of increasing phenol conversion. When the CO adsorption reaches saturation on the catalyst surface as the pressure increases, the phenol conversion then remains constant. In addition, the side reactions of phenol oxidation can also be accelerated with increasing pressure. The side reactions occur much more easily than DPC synthesis. Thus the DPC selectivity decreased with increasing pressure.

Fig. 8.Effect of reaction pressure on oxidative carbonylation of phenol. Reaction conditions: 110 °C, 7 h, phenol 25.5 mmol, 20 mL CH2Cl2, 0.4 g catalyst (Pd 0.04 mmol), 4.0 g 4A molecular sieve, Pd/Cu(OAc)2/TBAB/H2BQ = 1/5/40/40 in molar ratio.

The effect of Pd-O/CeO₂-NT catalyst loading on the oxidative carbonylation of phenol was investigated. The results are shown in Fig. 9. Phenol conversion increasesd from 46.4% to 68.8% with increasing catalyst loading due to the increase of active sites. The DPC selectivity, however, increased at first and then decreased gradually. This indicated that some side reactions were accelerated also. An overall consideration of phenol conversion and DPC selectivity showed that the optimal catalyst loading was 0.6 g (Pd/phenol = 1/425, molar ratio) under the given conditions. The optimized results were as follows. Phenol conversion was 67.7% with 93.3% DPC selectivity, and the turnover frequency (TOF) was 38.3 mol-DPC/(mol-Pd·h).

Fig. 9.TEM image of once used Pd-O/CeO2-NT.

After the reaction, the Pd-O/CeO₂-NT catalyst was washed with ethanol throughly, dried at 80 °C, and calcined at 500 °C for 1 h to eliminate residual organic compounds. Then it was re-evaluated for the oxidative carbonylation of phenol to DPC. The results showed that only 29.3% phenol was converted with 62.3% DPC selectivity. That is, Pd-O/CeO₂-NT has partly lost its catalytic activity. The TEM image of the once-used Pd-O/CeO₂-NT was recorded to reveal its change. This is shown in Fig. 10. Compared with Fig. 1(c), it can be concluded that most of the CeO₂ nanotubes had collapsed into small particles during the reaction. The TPR curve of the once-used Pd-O/CeO₂-NT is shown in Fig. 11. The peak corresponding to the reduction of surface O-II at 96 °C in the TPR curve of the fresh Pd-O/CeO₂-NT has disappeared, while the peaks assigned to the decomposition of PdH_x, surface oxygen, and bulk oxygen still existed. This is evidence that surface O-II was generated by the special structure of Pd-O/CeO₂-NT. Due to the destruction of the nanotube during the reaction, Pd-O/CeO₂-NT has lost its activity. In addition, ICP analysis was conducted to show Pd leaching during the reaction. The results showed there were 0.56% and 0.36% Pd content in the fresh and once-used Pd-O/CeO₂-NT, respectively. Therefore, this was also a cause of the deactivation.

Fig. 10.Effect of catalyst loading on oxidative carbonylation of phenol. Reaction conditions: 110 °C, 7 h, O2 0.6 MPa, CO 6.6 MPa, phenol 25.5 mmol, 20 mL CH2Cl2, 4.0 g 4A molecular sieve, Pd/Cu(OAc)2/TBAB/H2BQ = 1/5/40/40 in molar ratio.

Fig. 11.TPR profile of once used Pd-O/CeO2-NT.

CeO₂ nanotubes (CeO₂-NT) were prepared by a liquid phase deposition-hydrothermal method using CNTs as the template. The CeO₂-NT has outer diameters of about 25 nm and lengths of more than 300 nm. A Pd-O/CeO₂-NT catalyst was prepared using the CeO₂-NT as support, which has surface oxygen species that can be reduced at a lower temperature. This led to a high activity in the catalytic reaction. Pd-O/CeO₂-NT was used as the catalyst for the oxidative carbonylation of phenol and it gave better activity and DPC selectivity than Pd-O/CeO₂-P. When the CO pressure was 6.6 MPa and catalyst loading was Pd/phenol = 1/425 (molar ratio), the phenol conversion was 67.7% with 93.3% DPC selectivity at 110 °C after 7 h. After the reaction, the tubular structure of Pd-O/CeO₂-NT was destroyed and the surface oxygen species that was reduced at a lower temperature disappeared. Moreover, Pd leaching also occurred during the reaction. All these resulted in a decrease of the catalytic performance of Pd-O/CeO₂-NT.

碳酸二苯酯(DPC)是一种重要的化工中间体,可用于合成许多有机化合物及高分子材料,如聚碳酸酯、对羟基苯甲酸聚酯和脂肪族单异氰酸酯等^{[1, 2]}.近年来,由于聚碳酸酯的广泛应用,导致其原料DPC的需求大增,因而DPC的合成研究成为人们关注的热点^[3].DPC的合成方法主要有三种:光气法^[1]、酯交换法^{[4, 5]}和氧化羰基化法^[6].氧化羰基化法以苯酚、CO和O₂为原料一步直接合成DPC,具有工艺简单、原料便宜等特点^[1],其反应如下:

苯酚氧化羰基化反应是一个多步电子转移催化体系^[7],通常使用Pd及其化合物作为催化剂,同时加入各种助剂来提高催化性能^{[8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20]}.Vavasori等^[7]使用Pd(OAc)₂或PdBr₂为催化剂,对苯醌和Co,Mn,Cu盐为助剂,催化苯酚氧化羰基化反应;并分析了助剂在反应中所起的作用.Xue等^[10]采用微乳液为纳米反应器制备了包覆型催化剂Pd-Cu-O/SiO₂,同时加入Cu(OAc)₂、氢醌和四丁基溴化胺(TBAB)为助剂,DPC产率可达35.4%;同时,还减轻了Pd的流失,提高了催化剂重复使用性能.对于Pd催化的苯酚氧化羰基化反应,Ce化合物也是一种有效的助剂^[8].此外,Ce还可作为第二金属组分,或者作为载体(CeO₂)促进Pd催化的苯酚氧化羰基化反应.邬茂等^[21]制备了Mn掺杂CeO₂负载Pd催化剂,用于催化苯酚氧化羰基化反应,DPC收率为9.3%.王志苗^[22]采用微乳液法制备了Pd-Ce-O/SiO₂催化剂,在优化条件下,DPC收率可达53.7%.

CeO₂是工业催化中最具意义的稀土氧化物.CeO₂具有独特的萤石结构,拥有氧化还原离子对(Ce³⁺/Ce⁴⁺),Ce⁴⁺和Ce³⁺之间可以实现快速可逆的氧化还原循环.将这种快速的储放氧能力应用到非均相反应中,可以使O₂快速地从气相主体扩散到CeO₂固体表面形成表面氧,在多相催化中有广泛的应用前景^{[23, 24, 25, 26, 27]}.一般来说,小尺寸、大比表面积的CeO₂纳米颗粒具有较高的催化活性和选择性.然而,研究发现,CeO₂的催化性能不仅与粒子大小有关,而且与形貌密切相关.典型的例子是一维单晶结构的CeO₂纳米棒,因其选择性显露出高度活性的(110)晶面,更容易产生氧空位,因而表现出较常规纳米粒子更高的CO氧化活性^{[28, 29]}.此外,CeO₂纳米管因其显露管外和管内两个可利用的表面,与常规粒子相比具有更大的表面-体积比率,因而显示出更好的还原和氧存储能力^{[28, 30]}.因此,以一维结构CeO₂作为载体制备负载型Pd催化剂,可望获得较好的催化合成DPC性能.

本文以碳纳米管为模板,通过液相沉积-水热法制备出CeO₂纳米管(CeO₂-NT),以其为载体制备出Pd-O/CeO₂-NT催化剂,用于催化苯酚氧化羰基化合成DPC反应.考察了反应条件对其催化性能的影响,并与0维CeO₂颗粒负载Pd催化剂进行了对比.

称取4.0gCNTs(长度0.5-2μm,外径30-50nm,纯度>95%,购自中国科学院成都有机化学有限公司),放入1000mL三口烧瓶中;加入600mL硝酸溶液(30%),在120°C回流24h.静置冷却至室温后,用乙醇和去离子水洗涤至中性,80°C烘干备用.

以处理的CNTs为模板,通过液相沉积^[31]-水热法制备管状纳米CeO₂(CeO₂-NT).称取0.54g预处理CNTs放入500mL烧杯中,加入215mL无水乙醇和20mLCe(NO₃)₃溶液(0.15mol/L),将其放入超声水浴中分散2h.然后,在剧烈搅拌的同时,将NaOH溶液(0.125mol/L)缓慢加入到上述混合物中,调节pH值到10,继续搅拌30min.将所得悬浮液转移至晶化釜中,在110°C水热处理24h.过滤分离所得固体,用乙醇和去离子水洗涤滤饼至滤液中性,然后在80°C烘干至恒重.将所得固体在500°C焙烧4h,升温速率为5°C/min.所得样品记为CeO₂-NT.

与CeO₂-NT制备过程相似,只是将悬浮液转移至晶化釜的同时,向釜中加入一定量的PdCl₂氨水溶液,所得样品记为Pd-O/CeO₂-NT,其中Pd理论负载量为1.0wt%.

为了对比,采用W/O型微乳液为纳米反应器制备Pd-O/CeO₂-P催化剂,该催化剂中载体CeO₂形貌为0维颗粒.操作步骤如下:将100mL环己烷、10.8mLTritonX-100和一定量的PdCl₂氨水溶液(0.25mmolPdCl₂,10mL氨水)置于三口烧瓶中,搅拌得到粘稠乳浊液,然后加入2.8mL正己醇,混合体系变为澄清透明的W/O型微乳液.将此微乳液放入30°C恒温水浴中,再加入一定量的Ce(NO₃)₃水溶液(3mL,5.2mol/L),使其在氨水催化下进行水解反应5h.反应结束后,用无水乙醇反复离心洗涤多次,于80°C下干燥至恒重.将所得固体在500 °C焙烧4h,得到Pd-O/CeO₂-P催化剂,其中Pd理论负载量为1.0wt%.

采用PHILIPS公司TECNOL20型透射电子显微镜(TEM)观察样品形貌和尺寸,并进行选区电子衍射(SAED)分析,仪器工作电压为200kV.采用美国Micromeritics公司生产的ASAP2020M+C型比表面和孔隙度分析仪,对催化剂样品进行分析,样品的比表面积采用BET法计算.采用Rigaku D/Max-2500型X射线衍射仪(XRD,40kV,100mA)测定样品的晶体结构,Cu靶,石墨单色滤光片,狭缝SS/DS=1°,RS0.15mm,计数器SC,扫描范围2θ=5°-90°.采用Renishaw公司inViaReflex激光显微光谱仪测定CeO₂和催化剂的Raman光谱.采用Thermo Scientific公司iCAP 7400型等离子体发射光谱仪(ICP-OES)测定催化剂中的Pd含量.

采用美国Micromeritics公司Auto Chem II-2920型化学吸附分析仪进行H₂-程序升温还原(H₂-TPR)分析.称取0.1g样品,首先用Ar(50mL/min)于室温下吹扫5min,然后通入H₂-Ar(10%/90%)混合气(50mL/min);待基线平稳后,开始升温至1000°C,升温速率10 °C/min,利用TCD检测耗氢量.

催化剂活性评价在内衬聚四氟乙烯的不锈钢反应釜中进行,反应釜有效容积50mL.将催化剂、助催化剂Cu(OAc)₂、四丁基溴化胺(TBAB)、氢醌、4A分子筛、CH₂Cl₂和苯酚加入反应釜中,密封后用N₂置换其中的空气;然后分别充入O₂和CO,升温至预定温度,并反应一定时间.反应结束后,冷却至室温,取出反应物.过滤去除催化剂和4A分子筛,对液相产物进行色谱分析.

产物定性分析采用美国Thermo公司TRACEDSQ气相色谱-质谱联用仪.色谱条件:BPX5色谱柱(30mx0.25mmx0.25μm),进样口温度250°C;色谱柱初始温度40°C,以10°C/min升至250°C(5min);载气为He,流速1mL/min;分流进样,分流比50:1,进样量0.02μL.质谱条件:电子轰击(EI)离子源,轰击能量70eV;离子源温度200°C,传输线温度250°C;扫描范围40-500amu.

产物定量分析使用德国Knauer公司的K2600型高效液相色谱仪,VenusilXBPC18色谱柱,5μm,4.6mm×150mm,流动相V(甲醇):V(水)=65:35,检测波长为254nm,流速0.6mL/min,进样量20μL,柱温30°C.采用外标法对碳酸二苯酯和苯酚进行定量.

CeO₂及负载Pd催化剂的TEM照片见图1.由图可知,CeO₂-NT为管状结构,外径~25nm,长度>300nm,其管壁由粒径4-9nm的CeO₂晶粒组成.图1(b)中插图为选区电子衍射(SAED),由其衍射环可知CeO₂-NT为多晶结构;并且,该衍射环 并不平滑,环中还存在明显的衍射斑点,这是由于选区较小,产生衍射的CeO₂单晶颗粒较少所致.另外,Pd-O/CeO₂-NT中CeO₂的管状结构仍然存在,但很难分辨出Pd的存在.这是由于Pd与CeO₂之间存在较强的相互作用,抑制了Pd的生长,因而Pd颗粒尺寸较小;另一方面,由于CeO₂电子密度较高,采用TEM很难区分其中的Pd颗粒^[32].Boronin等^[33]也发现,即使采用高分辨率TEM,CeO₂载体上的大部分Pd颗粒也是观察不到的.图1(d)为采用微乳液法制备的Pd-O/CeO₂-P的TEM照片,纳米CeO₂颗粒的直径在7-15nm.同样,也无法观察到Pd颗粒的存在.

图2为Pd-O/CeO₂-NT和Pd-O/CeO₂-P催化剂的XRD谱.由图可知,二者没有明显的区别,均在28.48,33.12,47.59,56.46,59.07,69.44,76.75和79.16处出现衍射峰,分别对应于立方晶系CeO₂的(111),(200),(220),(311),(222),(400),(331)和(420)晶面(JCPDS65-5923).同时未检测到Pd化合物的特征衍射峰.这是由于CeO₂表面存在着大量氧空位,与Pd发生相互作用时,Pd离子可占据这些氧空位,从而抑制了Pd颗粒的聚积,使Pd化合物处于高度分散状态^[34],无法形成较强的衍射峰.

由XRD表征结果计算了两种催化剂中CeO₂的晶格参数(见表1).Pd-O/CeO₂-NT晶格参数(0.5402nm)小于Pd-O/CeO₂-P(0.5406nm),这是由于在前者中有较多的Pd²⁺/Pd⁴⁺进入到CeO₂晶格所致.由于Pd²⁺/Pd⁴⁺半径(0.84Å/0.62Å)小于Ce⁴⁺(0.99Å),因此掺杂较多Pd²⁺/Pd⁴⁺的Pd-O/CeO₂-NT晶格参数较小.

图3为Pd-O/CeO₂-NT和Pd-O/CeO₂-P样品的N₂吸附-脱附曲线和孔径分布图.可以看出,两种催化剂的吸附-脱附曲线均为IV型,伴有明显的滞后环,这是由于中孔毛细凝聚所引起的^[35];其中,Pd-O/CeO₂-NT的吸附-脱附滞后环为H3型,而Pd-O/CeO₂-P的为典型的H2型,说明二者具有不同的孔结构.还可看出,Pd-O/CeO₂-NT表面的孔径分布较为分散,而Pd-O/CeO₂-P则具有小而均一的孔,主要分布在介孔范围内.此外,Pd-O/CeO₂-NT的比表面积为108.8m²/g,高于Pd-O/CeO₂-P(93.8m²/g),这应是前者的管状结构及较小的晶粒共同作用的结果.

将前述两种不同形貌CeO₂负载的Pd催化剂用于苯酚氧化羰基化反应,结果见表1.可以看出,无论是苯酚转化率,还是目标产物DPC的选择性,Pd-O/CeO₂-NT均优于Pd-O/CeO₂-P.

为分析催化性能存在差异的原因,对各催化剂及其载体进行了H₂-TPR表征,结果见图4和表2.由图可知,CeO₂-P样品存在2个还原峰,分别位于478和810 °C.通常认为,在CeO₂的TPR曲线中,400-600°C的低温峰对应于与表面Ce⁴⁺离子以八面体配位形式相结合的表面氧负离子的还原;700°C以上的高温峰则对应于与体相两个Ce⁴⁺ 离子相结合的体相氧负离子的还原^{[36, 37]}.可以看出,CeO₂-NT上除了表面氧还原峰(430°C)和体相氧还原峰(815°C)外,在307°C还出现了一个强还原峰.由表2可知,CeO₂-P与CeO₂-NT的体相氧含量基本相同,而CeO₂-NT的表面氧含量却大大降低;但是,如果将其与307°C还原峰对应的氧含 量 相加,则其和(503.1mmol/g)与CeO₂-P表面氧含量(515.4mmol/g)基本相同.因此,307°C还原峰对应的是一种更容易被还原的表面氧,可用SurfaceO-II来表示,该氧物种的产生与CeO₂-NT的管状结构有关.同时,这也说明CeO₂-NT比CeO₂-P更容易被还原,其表面氧更容易参与到催化反应中.Zhou等^[30]制备了高还原性的CeO₂纳米管,发现在200°C即开始发生表面氧还原.Shan等^[38]也发现,在制备的3DCeO₂的TPR曲线中出现还原温度为180°C的氧物种.他们认为,这是由于CeO₂表面氧空位增加,更容易活化分子氧,产生了较多的易还原吸附氧物种所致.Rao^[39]则认为CeO₂表面氧还原温度的高低与O^2-的配位数有关;配位数低的还原温度低,反之亦然.

将Pd负载到CeO₂载体上后,H₂-TPR曲线发生了明显变化.Pd-O/CeO₂-P样品中,表面氧的还原变得十分微弱,其H₂消耗量仅为75.8mmol/g.此外,还存在200和778°C两个还原峰,后者归属于体相氧的还原;而前者则可认为是由PdO的还原与CeO₂-P表面氧的还原共同作用所致^[40].纯PdO的还原温度为50°C^[41],负载后其还原温度的升高以及CeO₂-P表面氧还原温度的降低都可归因于PdO与CeO₂-P之间的相互作用^[40].

在Pd-O/CeO₂-NT的TPR曲线中,体相氧的还原峰位于785°C,表面氧的则降至299°C.而在低于200°C的区域出现了一个强还原峰,峰顶温度为96°C.此外,在65°C还出现了一个倒峰,这说明样品中释放出了H₂.众所周知,在室温下,当H2压力超过0.013atm时,Pd即可吸附H₂形成氢化物(PdH_x)^[42];升高温度时,PdH_x分解即释放出H₂.这也说明Pd-O/CeO₂-NT表面的Pd物种可在更低的温度下被还原.Cargnello等^[43]发现,采用浸渍法制备的PdO/CeO₂在室温下即可被还原为Pd;在TiO₂载体上也有类似现象^[40].这是Pd被高度分散到CeO₂-NT表面所产生 的结 果 .此外,这也说明96°C还原峰对应的全部为SurfaceO-II的还原,不包括PdO的还原.Pd-O/CeO₂-NT表面氧还原温度的降低则是溢流氢作用的结果^[41].由表2可知,与CeO₂-NT(503.1mmol/g)和Pd-O/CeO₂-P(596.0mmol/g)相比,Pd-O/CeO₂-NT的表面氧 总耗氢量大大增加(1350.7mmol/g).这应该是由于Pd-O/CeO₂-NT表面氧空位密度增加所致.图5是Pd-O/CeO₂-NT和Pd-O/CeO₂-P的Raman谱.由图可知,449cm^-1附近的Raman峰对应于萤石结构CeO₂的F_2g特征振动;而581cm^-1处Raman峰则对应于CeO₂表面氧空位^[44]如果以氧空位Raman峰与F_2g特征Raman峰的峰面积之比表示氧空位的相对密度,则Pd-O/CeO₂-NT表面氧空位密度(0.58)高于Pd-O/CeO₂-P(0.43).由前述XRD结果可知,在Pd-O/CeO₂-NT中有较多的Pd进入到CeO₂晶格中,导致其表面形成了较多的氧空位^[45].分子氧O₂可以在氧空位上发生强化学吸附,生成易还原的表面氧物种^[38].另一方面,Pd-O/CeO₂-NT表面部分氢消耗也可能是H₂在高度分散的金属Pd表面发生解离吸附,产生的H原子溢流到CeO₂载体上而产生的.

综上可知,在Pd-O/CeO₂-NT催化剂中,CeO₂表面氧活性和密度均有所增加,因而以其为催化剂,苯酚转化率较高.同时,由于Pd被高度分散到载体表面,并部分进入CeO₂晶格,形成固溶体,Pd与CeO₂之间存在着强相互作用,更加有利于Pd和Ce之间的电子转移.反应中被还原的Pd(0)可以容易地向C e(IV)转移电子,重新被氧化为Pd(II),恢复生成DPC的能力,从而具有较高的选择性.

图6为反应时间对苯酚氧化羰基化反应的影响.由图可知,当反应5h时,苯酚转化率为23.3%,DPC选择性为96.3%.随着反应时间的增加,苯酚转化率迅速增加,至8h时,苯酚转化率为58.2%.DPC选择性则随反应时间的延长而逐渐降低;特别是反应时间由7h延长至8h 时,DPC选择性由93.6%迅速降为86.8%.在反应体系中,苯酚会发生氧化副反应,生成苯醌,进一步氧化会生成一些多聚物,反应溶液呈现棕黑色也证明了这一点.这些聚合物随着反应时间延长而不断积累,覆盖在催化剂表面,阻碍了Pd活性中心与反应物的接触,从而降低了DPC生成速率;但苯酚氧化副反应受影响较小,因此转化率持续增加,DPC选择性会逐渐降低.此外,反应时间延长时,DPC还会与体系中生成的微量水发生水解副反应,生成苯酚和CO₂,也会降低其选择性.

图7考察了反应温度对苯酚氧化羰基化反应的影响.由图可知,当反应在80°C进行时,苯酚转化率为17.7%,DPC选择性为88.1%.随着温度的升高,苯酚转化率逐渐增加,至110°C时达到60.7%;继续升高温度,苯酚转化率基本不变.DPC选择性则随着反应温度的升高而缓慢增加,至110°C时达到95.3 %;继续升高温度至120°C,DPC选择性降为85.7%.苯酚氧化羰基化合成DPC为放热反应,温度升高会导致平衡常数下降,不利于反应的进行;但从动力学角度考虑,升高温度会加快反应速率.因此,反应温度应存在一个最佳值.此外,当反应温度过高时,苯酚容易发生聚合反应,降低了DPC选择性;而苯酚多聚体也可覆盖在催化活性中心上,阻碍反应的进行,抑制了DPC的生成.

图8为反应压力对苯酚氧化羰基化反应的影响.为保证反应气氛处于CO爆炸极限之外,将CO/O₂(分压比)限定为11/1.由图可知,当CO压力由4.4MPa增加到6.6MPa时,苯酚转化率由41.3%上升到60.7%;继续提高反应压力,苯酚转化率基本不变.DPC选择性则随CO压力从4.4MPa增加到7.7MPa从97.2%逐渐降到87.7%.苯酚氧化羰基化反应是一个体积缩小的反应,提高压力有利于反应进行;另外,提高压力也有利于CO在催化剂上的吸附,因此苯酚转化率随反应压力的增加而增加.但是,由于催化剂用量的限制,当CO在催化剂表面的吸附随着压力增加而基本趋于饱和时,转化率即不再发生变化.在主反应得到加速的同时,增大压力也有利于苯酚氧化副反应,并且该类氧化反应更容易进行,因此DPC选择性随反应压力的增加而降低.

图9考察了Pd-O/CeO₂-NT用量对苯酚氧化羰基化反应的影响.由图可知,随着催化剂用量的增加,苯酚转化率由46.4%逐渐增加至68.8%.增加催化剂用量提供了更多的活性中心,使反应速率加快,因而苯酚转化率提高;而DPC选择性则随催化剂用量的增加而先增加后降低.综合考虑,当Pd-O/CeO₂-NT用量为0.6g(Pd/苯酚=1/425,摩尔比)时效果最佳,苯酚转化率为67.7%,DPC选择性为93.3%,催化剂转换频率为38.3mol-DPC/(mol-Pd·h).

将反应后的Pd-O/CeO₂-NT催化剂用乙醇充分洗涤,80°C干燥,并在500°C焙烧1h,除去残留的有机物.将其再次用于催化苯酚氧化羰基化反应,苯酚转化率为29.3%,DPC选择性为62.3%.图10给出了使用一次后Pd-O/CeO₂-NT的TEM照片;与图1(c)对比可知,Pd-O/CeO₂-NT催化剂中大部分纳米管都已坍塌为纳米颗粒.并且,对反应后Pd-O/CeO₂-NT进行了TPR表征(图11),发现其中并不存在对应于氧物种surfaceO-II的还原峰,说明经过一次反应后,该氧物种已经消失.而其他表面氧和体相氧的还原以及PdH_x分解倒峰仍然存在.结合TEM表征结果,可认为易还原氧物种surface O-II的产生与Pd-O/CeO₂-NT的管状结构有关.经过一次反应后,Pd-O/CeO₂-NT的管状结构被破坏,氧物种surface O-II消失,因而其催化苯酚氧化羰基化反应性能明显降低.此外,ICP分析发现,反应前后催化剂中Pd含量分别为0.56%和0.36%.活性组分在反应中的流失也是催化剂失活的重要原因.

以CNTs为模板,采用液相沉积-水热法制备了管状纳米CeO₂-NT,其外径~25nm,长度＞300nm.以其为载体制备了Pd-O/CeO₂-NT催化剂,该催化剂的表面氧在低温下即可被还原,具有更好的反应活性.对于催化苯酚氧化羰基化反应,Pd-O/CeO₂-NT催化剂的活性和碳酸二苯酯(DPC)选择性均优于Pd-O/CeO₂-P催化剂.当CO压力为6.6MPa,催化剂量为Pd/苯酚=1/425(摩尔比)时,在110°C反应7h,苯酚转化率为67.7%,DPC选择性为93.3%.使用一次后Pd-O/CeO₂-NT的管状结构被破坏,易还原氧物种消失,并且活性组分Pd流失明显,因此其催化性能明显降低.