TiO2-Doped CeO2 Nanorod Catalyst for Direct Conversion of CO2 and CH3OH to Dimethyl Carbonate: Catalytic Performance and Kinetic Study.

A new class of TiO2-doped CeO2 nanorods was synthesized via a modified hydrothermal method, and these nanorods were first used as catalysts for the direct synthesis of dimethyl carbonate (DMC) from CO2 and CH3OH in a fixed-bed reactor. The micromorphologies and physical-chemical properties of nanorods were characterized by transmission electron microscopy, X-ray diffraction, N2 adsorption, inductively coupled plasma atomic emission spectrometry, X-ray photoelectron spectroscopy, and temperature-programmed desorption of ammonia and carbon dioxide (NH3-TPD and CO2-TPD). The effects of the TiO2 doping ratio on the catalytic performances were fully investigated. By doping TiO2, the surface acid-base sites of CeO2 nanorods can be obviously promoted and the catalytic activity can be raised evidently. Ti0.04Ce0.96O2 nanorod catalysts exhibited remarkably high activity with a methanol conversion of 5.38% with DMC selectivity of 83.1%. Furthermore, kinetic and mechanistic investigations based on the initial rate method were conducted. Over the Ti0.04Ce0.96O2 nanorod catalyst, the apparent activation energy of the reaction was 46.3 kJ/mol. The reaction rate law was determined to be of positive first-order to the CO2 concentration and the catalyst loading amount. These results were practically identical with the prediction of the Langmuir-Hinshelwood mechanism in which the steps of CO2 adsorption and activation are considered as rate-determining steps.

Introduction

Carbon dioxide (CO2) is a major greenhouse gas that contributes to Earth’s global warming.[1] The emission of CO2 has been significantly increased within the past years and is still increasing each year. Over the last decade, the utilization of CO2 has attracted scientists’ attention because it is known as a naturally plentiful and recyclable carbon source for the production of chemical feedstocks.[2,3]

Dimethyl carbonate (DMC) is an important chemical material[4−6] because it serves as an environmentally friendly intermediate and a nontoxic substitute for poisonous and corrosive phosgene and dimethyl sulfate in the reaction of carbonylation and methylation,[7−10] as well as a promising fuel additive.[11,12] Because of the high demand, a lot of efforts have been made to find a sustainable way to mass-produce DMC[13] and several routes have been developed: carbonylation of methanol,[14] transesterification of carbonates,[15] methanolysis of urea,[16] and direct synthesis of DMC from CO2 and methanol[17−20] technologies.[17−20]

Among all of these routes, the first three synthesis methods are still far from perfect with rigorous conduct conditions, highly toxic and severely corrosive.[13,21] Therefore, the direct synthesis of DMC from CO2 and methanol has gradually become the most attractive approach as such a method is environmentally friendly not only for developing new carbon resources but also for reducing greenhouse gas emission.[22] However, there also exist significant challenges. DMC productivity is relatively low due to the fact that CO2 is highly thermodynamically stable and kinetically inert as well as because of the deactivation of catalysts induced by water formation in the reaction process.[9,23]

A number of methods to improve the DMC yield have been developed,[24] including adding coreagents[25] and removing water by dehydrants[26] in the reaction systems. Furthermore, some new technologies, such as the photoassistant,[27] electroassistant,[28] membrane separation,[29] and supercritical CO2[10,30] technologies have been introduced. However, the DMC yield is far from satisfaction. Although all of these approaches are pursued today, developing efficient catalysts, especially heterogeneous catalysts, is still considered to be the most effective route.[31,32] Compared with homogeneous catalysts, heterogeneous catalysts are difficult to deactivate, convenient to separate, easy to recycle, and environmentally friendly.[13,33] Hence, the exploitation of heterogeneous catalysts is regarded as the first choice for the direct synthesis of DMC, especially metal oxide catalysts,[34,35] heteropoly acid catalysts,[36] and bimetallic supported catalysts.[23]

As an excellent metal oxide catalyst with both Lewis acid and base properties, CeO2 particularly shows high catalytic activities in the direct synthesis of DMC.[37,38] There are a number of references in which it is also employed as a catalyst in the reaction involving dehydration.[39] In addition, on the basis of previous studies, various morphologies of dispersible CeO2 nanocrystals have been synthesized including nanoparticles, polyhedra, nanocubes, and nanorods.[40−42] Also, through theoretical simulation, it was found that three low-index crystal planes exposed on CeO2 nanocrystals were strongly affected by their morphology, exhibiting different properties and further influencing the catalytic activity for the direct synthesis of DMC.[43,44] Previous studies of Wang et al. have revealed that the CeO2 nanorod catalyst with the most (1 1 0) facet exhibited the most favorable catalytic activities for the direct synthesis of DMC when compared to those of CeO2 nanocubes with the (1 0 0) facet and CeO2 nano-octahedra with the (111) facet.[45]

In addition to neat CeO2, TiO2-modified CeO2 has attracted significant attention for catalytic applications due to the enhanced properties. Various studies on TiO2-doped CeO2 catalysts have shown enhancements in redox reactions.[46,47] It was proved that the oxygen deficiency and acidity of the surface can be significantly enhanced by doping TiO2 into CeO2.[48] Thus, we carried out further research on TiO2-doped CeO2 nanorod catalysts.

In this work, we first report a new class of TiO2-promoted CeO2 nanorod catalysts for the direct synthesis of DMC from CO2 and methanol. The recyclability of the catalysts was also explored. We also presented a detailed kinetic model for the direct synthesis of DMC in a continuous fixed bed reactor over CeO2-based catalysts. These will be certainly helpful for a comprehensive understanding of the reaction.

Results and Discussion

Morphology of the Prepared Catalysts

The morphology of CeO2 and TiCe1–O2 nanorod catalysts was observed by transmission electron microscopy (TEM), and the images are shown in Figure S1. From TEM images of pure CeO2 nanorods (Figure S1a,b), it can be seen that CeO2 exhibits intact nanorods. The morphological parameters measured from TEM observation show that CeO2 nanorods have an average length of around 50–100 nm with an external diameter of about 10 nm. After being modified by TiO2, the nanorod structure was not destroyed and the size of nanorods remained almost unchanged, suggesting the well dispersion of TiO2 among the nanorods.

Microstructure of the Prepared Catalysts

The crystal structures of CeO2 and TiCe1–O2 nanorod catalysts were investigated by X-ray diffraction (XRD), and the spectra are shown in Figure .

Figure 1

XRD patterns of CeO2 and TiCe1–O2 nanorods: (a) CeO2, (b) Ti0.02Ce0.98O2, (c) Ti0.04Ce0.96O2, (d) Ti0.06Ce0.94O2, (e) Ti0.08Ce0.92O2, and (f) Ti0.1Ce0.9O2.

For neat CeO2 nanorods (Figure a), the diffraction peaks (at 2θ of 28.6° (111), 33.1° (200), 47.6° (220), 56.4° (311), 59.2° (222), and 69.5° (400)) can be assigned to the cubic phase of CeO2 (JCPDS 65-5923). After doping, the spectra of Ti0.02Ce0.98O2 and Ti0.04Ce0.96O2 nanorods (Figure b,c) remained unchanged, indicating that TiO2 disperses well among the nanorods. However, for Ti0.06Ce0.94O2, Ti0.08Ce0.92O2, and Ti0.1Ce0.9O2, diffraction peaks of TiO2 (2θ of 27.6° (110) and 41.3° (111)) can be observed, suggesting that TiO2 particles are bigger and appear individually in nanorods. Furthermore, the average grain size (Table ) in the series of nanorods decreases from 11.04 to 8.19 nm in the fluorite structure along with the doping of TiO2.

Table 1

Textural Data of CeO2 and TiCe1–O2 Nanorods According to N2 Adsorption, XRD, and Inductively Coupled Plasma Atomic Emission Spectroscopy (ICP-AES) Investigations

parameters of nanorodsa	CeO2	Ti0.02Ce0.98O2	Ti0.04Ce0.96O2	Ti0.06Ce0.94O2	Ti0.08Ce0.92O2	Ti0.1Ce0.9O2
BET surface area (m2/g)	75.3	71.9	71.5	50.2	51.3	40.7
pore volume (m3/g)	0.299	0.380	0.308	0.234	0.218	0.191
average pore size (nm)	15.9	21.2	17.2	18.6	17.0	18.7
average grain size (nm)b	11.0	9.89	9.32	8.72	8.72	8.19
content of Ti in bulk phase (%)c		2.08%	3.97%	6.10%	8.12%	9.93%
content of Ti on surface (%)d		2.15%	4.06%	6.13%	8.04%	10.1%

The nanorods were calcined in air under 600 °C for 4 h.

Average grain size was calculated by the Scherrer formula according to the XRD result.

Content of Ti in the bulk phase was calculated with O elimination according to the ICP-AES result.

Content of Ti on the surface was calculated with O elimination according to the X-ray photoelectron spectroscopy (XPS) result.

Textures of the catalysts are listed in Table . The Brunauer–Emmett–Teller (BET) specific surface area of the nanorods decreases slightly from 75.3 to 71.5 m2/g upon increasing the TiO2 doping ratio to 0.04, indicating that the introduction of TiO2 almost has no effect on the surface area within a suitable ratio. However, with a higher TiO2 content (Ti0.06Ce0.94O2, Ti0.08Ce0.92O2, and Ti0.1Ce0.9O2), the specific surface area decreases significantly, which may cause the exposure of less active sites and thus lead to a lower activity. It also shows that the pore volume of nanorods increases slightly and then drops evidently as TiO2 doping ratio increases. To sum up, the samples of Ti0.02Ce0.98O2 and Ti0.04Ce0.69O2 nanorods have almost the same specific surface area, larger pore volume, and smaller average grain sizes when compared to those of neat CeO2 nanorods. The bulk phase and surface compositions of prepared nanorods were investigated by ICP-AES and XPS, respectively. It can be seen that the content of Ti matches well with the added proportion of raw materials both on the surface and in bulk phase. This illustrates that the composition of nanorods is homogeneous.

XPS Investigations of the Prepared Catalysts

XPS analysis was performed to lucubrate the surface chemical state of Ce, Ti, and O in prepared nanorods. Ce 3d, Ti 2p, and O 1s spectra of CeO2 and TiCe1–O2 nanorods are shown in Figures –4, respectively.

Figure 2

XPS Ce 3d spectra of CeO2 and TiCe1–O2 nanorods: (a) CeO2, (b) Ti0.02Ce0.98O2, (c) Ti0.04Ce0.96O2, (d) Ti0.06Ce0.94O2, (e) Ti0.08Ce0.92O2, and (f) Ti0.1Ce0.9O2.

Figure 4

XPS O 1s spectra of TiCe1–O2 nanorods: (a) CeO2, (b) Ti0.02Ce0.98O2, (c) Ti0.04Ce0.96O2, (d) Ti0.06Ce0.94O2, (e) Ti0.08Ce0.92O2, and (f) Ti0.1Ce0.9O2.

The labels used for identifying Ce 3d XPS peaks were established by Burroughs et al.,[49] where V and U indicate the spin–orbit couplings of 3d5/2 and 3d3/2, respectively. The peaks referred to as v, v″, and v‴ are contributed by CeO2 and assigned to a mixture of Ce IV (3d94f2) O (2p4), Ce IV (3d94f1) O (2p5), and Ce IV (3d94f0) O (2p6), respectively.[48,50] The same peak assignment is applied to u structures. Peaks v0 and v′ are assigned to a mixture of Ce III (3d94f2) O (2p5) and Ce III (3d94f1) O (2p6), respectively. The Ce 3d spectrum of CeO2 (Figure a) shows six peaks at 917.9, 908.3, 902.1, 899.9, 888.9, and 883.4 eV. These peaks indicate the existence of Ce4+ and progressively decrease with the increased amount of Ti doping. In contrast, the intensity of peaks at 903.7 and 887.8 eV increases as the Ti doping ratio increases, which reveals the increase of Ce3+ concentration. On the basis of the curve-fitting method established by Watanabe et al.,[48] the XPS spectra forecast that the surface concentration of Ce3+ in nanorods varies with the amount of Ti doping from 5.1% in CeO2 nanorods to 22.4% in Ti0.1Ce0.9O2 nanorods. Ce3+ concentration shows an apparently linear relationship with Ti doping, as shown in Figure S3. As a consequence, introducing TiO2 into CeO2 leads to the partial reduction of Ce4+.

Concomitant with the change of Ce in CeO2 nanorods upon Ti doping, the chemical state of Ti also changes. Figure shows Ti 2p1/2 and Ti 2p3/2 spectra of TiCe1–O2 nanorods. The Ti 2p3/2 peak shifts from 458.0 eV in Ti0.02Ce0.98O2 to 458.6 eV in Ti0.1Ce0.9O2, representing intermediate TiO2 incorporation into nanorods. Thus, the XPS result reveals the significant reduction and higher degree of oxygen deficiency of Ce and Ti due to the doping.

Figure 3

XPS Ti 2d spectra of TiCe1–O2 nanorods: (a) Ti0.02Ce0.98O2, (b) Ti0.04Ce0.96O2, (c) Ti0.06Ce0.94O2, (d) Ti0.08Ce0.92O2, and (e) Ti0.1Ce0.9O2.

Figure shows the O 1s spectra of prepared nanorods. The binding energy of the surface oxygen is pretty susceptible to the concentrations of Ce and Ti in TiCe1–O2 nanorods, shifting from 528.8 to 529.4 eV upon increasing the Ti concentration. Furthermore, O 1s peaks corresponding to TiO2 and CeO2 do not appear in the spectra individually, suggesting that Ti and Ce chemically interact with each other in nanorods. Thus, TiCe1–O2 nanorods are not simple mixtures of two different oxides but rather a homogeneous solid solution with chemical interactions.

Temperature-Programmed Desorption (TPD) Investigations of the Prepared Catalysts

To further explore the surface acid–base properties, NH3-TPD and CO2-TPD techniques were employed and the results are shown in Figures S4 and S5, respectively. All profiles can be deconvoluted into three Gaussian peaks.

Figure S4 displays the NH3-TPD profiles of CeO2 and TiCe1–O2 nanorods. The amount of acidic sites and peak positions are summarized in Table . For all nanorods, three small peaks assigned to weakly (100–200 °C), moderately (200–400 °C), and strongly (400–600 °C) acidic sites can be found. For Ti0.02Ce0.98O2 and Ti0.04Ce0.96O2 nanorods, more moderately acidic sites are formed and increase with the increasing TiO2 content, which has already been proved to be beneficial to the formation of DMC.[51] However, for Ti0.06Ce0.94O2, Ti0.08Ce0.92O2, and Ti0.1Ce0.9O2 nanorods, the amount of moderately acidic sites decreases, which can be ascribed to the following reasons: the specific surface area decreases rapidly upon increasing the TiO2 content, which can then result in less exposed active sites.

Table 2

Quantification of the NH3-TPD Profiles of CeO2 and TiCe1–O2 Nanorods

	weakly		moderately		strongly
nanorod samplesa	T (°C)	amount (μmol/g)	T (°C)	amount (μmol/g)	T (°C)	amount (μmol/g)	total
CeO2	175.0	63	352.3	135	507.2	54	252
Ti0.02Ce0.98O2	174.3	68	348.0	174	510.2	63	305
Ti0.04Ce0.96O2	172.8	69	343.7	201	512.6	73	342
Ti0.06Ce0.94O2	170.8	61	339.4	197	513.9	75	334
Ti0.08Ce0.92O2	168.5	52	335.7	187	514.9	78	317
Ti0.1Ce0.9O2	166.0	46	332.2	179	514.8	82	307

The nanorods were calcined in air under 600 °C for 4 h.

Figure S5 displays CO2-TPD profiles of CeO2 and TiCe1–O2 nanorods. The amount of basic sites and peak positions are summarized in Table . For all nanorods, three small peaks assigned to weakly (100–200 °C), moderately (200–400 °C), and strongly (400–600 °C) basic sites can be found. Similar to acidic sites, moderately basic sites increase first for CeO2, Ti0.02Ce0.98O2, and Ti0.04Ce0.96O2 nanorods and then decrease in the following order: Ti0.06Ce0.94O2, Ti0.08Ce0.92O2, and Ti0.1Ce0.9O2 nanorods. It can be observed that the Ti0.04Ce0.96O2 nanorod possesses the largest amount of both acidity and basicity. This is mainly due to the increase of Ce–O, Ti–O pairs on the surface, which can provide more active sites, as well as keeping undiminished specific surface area when compared to that of neat CeO2 nanorods.

Table 3

Quantification of the CO2-TPD Profiles of CeO2 and TiCe1–O2 Nanorods

	weakly		moderately		strongly
nanorod samplesa	T (°C)	amount (μmol/g)	T (°C)	amount (μmol/g)	T (°C)	amount (μmol/g)	total
CeO2	164.8	23	333.7	43	512.6	17	83
Ti0.02Ce0.98O2	164.0	26	328.9	58	503.0	20	104
Ti0.04Ce0.96O2	163.1	28	325.5	73	496.8	24	125
Ti0.06Ce0.94O2	162.1	24	322.9	72	493.0	22	118
Ti0.08Ce0.92O2	161.1	20	321.0	67	490.5	19	107
Ti0.1Ce0.9O2	160.0	17	319.5	63	488.9	18	98

The nanorods were calcined in air under 600 °C for 4 h.

Catalytic Performance

The effects of TiO2 content on the catalytic performance were investigated, and the catalytic process was carried out in a 50 mL autoclave microreactor with high-velocity mechanical stirring. The results are shown in Table and serve as a basis for original selection of nanorod catalysts.

Table 4

Influence of TiO2 Content on DMC Productivity

nanorod catalysta	DMC yield (mmol g/cat)b
none	0
CeO2	3.20
Ti0.02Ce0.98O2	4.02
Ti0.04Ce0.96O2	4.70
Ti0.06Ce0.94O2	4.56
Ti0.08Ce0.92O2	4.33
Ti0.1Ce0.9O2	4.20

The nanorods were calcined in air under 600 °C for 4 h.

The reaction conditions are as follow: 0.20 g of catalysts; 15 mL of MeOH; initial pressure, 5 MPa; reaction temperature, 120 °C, reaction time, 5 h.

The neat CeO2 nanorod catalyst shows much lower DMC yield than that of TiCe1–O2 nanorods. The DMC yield increases with the increasing TiO2 content, reaching a maximum, and then decreases with a further increase in the TiO2 amount. Among the catalysts examined, the Ti0.04Ce0.96O2 nanorod catalyst shows the highest DMC yield of 4.70 mmol g/cat, much higher than that of the neat CeO2 nanorod catalyst, which shows the activity of 3.20 mmol g/cat.

In association to the result of N2 adsorption, Ti0.06Ce0.94O2, Ti0.08Ce0.92O2, and Ti0.1Ce0.9O2 nanorod catalysts with a smaller surface area exhibit lower catalytic performance, suggesting that the catalytic activity is correlated to the surface area of the catalysts. Small surface area generally causes the exposure of less active sites and leads to a lower catalytic activity.[52,53]

In addition, according to the mechanism of the synthesis of DMC from CH3OH and CO2, methanol and CO2 are absorbed and activated on the acidic and basic sites of the catalyst, respectively.[54] Therefore, in this study, the low catalytic activity of the neat CeO2 nanorod catalyst may result from the smaller amount of acidic and basic sites on the surface, which leads to less activated methanol and CO2. TiCe1–O2 nanorod catalysts exhibit higher catalytic activities than those of neat CeO2 nanorods. It can be attributed to the increased amount of moderately acidic and basic sites with the addition of TiO2.

Figures S6 and S7 show the relationships between the catalytic activity and the amount of moderately acidic and basic sites, respectively. It can be seen that the DMC yield increases linearly upon increasing moderately acidic and basic sites. Combining TPD, XRD, and N2 adsorption results, it is revealed that the coexistence of nanorod structures of CeO2 and TiO2 favors the formation of moderately acidic and basic sites. The lower catalytic activity of Ti0.06Ce0.94O2, Ti0.08Ce0.92O2, and Ti0.1Ce0.9O2 nanorods is presumably due to the decrease of specific surface area, which reduces the surface moderately acidic and basic sites.

After the original selection, further research of DMC formation from methanol and CO2 with Ti0.04Ce0.96O2 nanorods was carried out in a fixed-bed microreactor (Scheme S1). The effects of different catalytic conditions on the activities of Ti0.04Ce0.96O2 nanorod catalysts for DMC synthesis were fully investigated, and the results are summarized in Table . The catalytic performance gives a crest value at 140 °C while the reaction temperature varies from 110 to 160 °C. Higher pressures are favorable for methanol conversion but adverse to DMC selectivity, leading to the highest yield of DMC at 1.0 MPa. The catalytic activities decline slightly with the increasing space velocity. Therefore, reaction conditions were selected for further exploration at 140 °C and 1.0 MPa and space velocity of 360 h–1.

Table 5

Effects of Reaction Conditions on the Catalytic Performances of DMC Formation from CO2 and CH3OH over Ti0.04Ce0.96O2 Nanorod Catalystsa

Different Reaction Temperatures (°C): Pressure: 0.8 MPa; Space Velocity: 360 h–1
temperature (°C)	methanol conversion (%)	DMC selectivity (%)	yield (%)
110	2.77	92.5	2.56
120	3.61	91.5	3.48
130	4.16	89.9	3.74
140	5.20	83.8	4.36
150	5.41	77.9	4.21
160	6.29	66.5	4.18

Ti0.04Ce0.96O2 nanorods were calcined in air under 600 °C for 4 h. Catalyst weight: 2.0 g. Time on stream: 7 h. The results were the average data between 2 and 4 h with time on stream.

Further investigation of the recyclability of Ti0.04Ce0.96O2 nanorods was carried out, and the used nanorods were recovered by recalcination before reusing them for DMC formation under the same catalytic conditions. SBET and catalytic activities of the recycled catalysts are presented in Table . Both the catalytic activities and SBET were found dropping as the number of recycles increases, which is mainly due to the redispersion of active sites during the regeneration of catalyst. However, Ti0.04Ce0.96O2 nanorods exhibit pretty good stability for the direct synthesis of DMC from CO2 and methanol.

Table 6

Recyclability Study of the Ti0.04Ce0.96O2 Nanorod Catalyst for the Direct DMC Formation from CO2 and Methanola

entry	recycle number	methanol conversion (%)	DMC selectivity (%)	yield (%)	SBET (m2/g)
1	fresh	5.38	83.1	4.47	71.5
2	first	5.01	79.5	3.98	66.8
3	second	4.67	76.2	3.40	61.9
4	third	4.22	72.4	3.06	56.9

Ti0.04Ce0.96O2 nanorods were calcined in air under 600 °C for 4 h. Catalyst weight: 2.0 g. Time on stream: 7 h. The results were the average data between 2 and 4 h with time on stream.

Effect of Dehydrant on DMC Formation

To explore the effect of dehydrant, further DMC synthesis reactions were carried out with various dehydrants, including 2,2-dimethoxypropane, cyclohexene oxide, and 2-cyanopyridine dissolved in reactant methanol (dehydrants/methanol = 5.0 wt %). As shown in Table , 2-cyanopyridine turned out to be the best, enhancing DMC yield from 4.47 to 7.83%. In addition, 2,2-dimethoxypropane exhibited excellently improved effect with DMC yield of 6.76%. Surprisingly, cyclohexene oxide also showed pretty good promotion of DMC yield of 6.30%. These results indicate that the dehydrant played important roles in direct DMC synthesis from CO2 and methanol with a positive effect on the DMC yield.

Table 7

Effects of Dehydrants on the Catalytic Performances of DMC Formation from CO2 and CH3OH over Ti0.04Ce0.96O2 Nanorod Catalystsa

entry	dehydrant
1
2	2,2-dimethoxypropane
3	cyclohexene oxide
4	2-cyanopyridine

Ti0.04Ce0.96O2 nanorods were calcined in air under 600 °C for 4 h. Catalyst weight: 2.0 g. Time on stream: 7 h. The results were the average data between 2 and 4 h with time on stream.

Kinetic Insights

Further kinetic exploration for the direct synthesis of DMC with the Ti0.04Ce0.96O2 nanorod catalyst was conducted in an optimized fixed-bed reactor (Schedule S2). On the basis of former literature reports, two different mechanisms are involved and the apparent rate law can be obtained based on the controlling step of each mechanism further (Table ).[55,56] Regarding the Eley–Rideal (ER) mechanism, it assumes that CO2 reacts directly with the activated MeOH group adsorbed on the catalyst surface because the adsorption of MeOH seems to be much stronger with the MeOH* and CO2 combination (S2) as the rate-determining step. While considering the adsorption of all species on the catalyst, the Langmuir–Hinshelwood (LH) mechanism declares that CO2 and methanol should interact with the catalyst surface in two separate steps. However, there are controversial opinions on the controlling step for the LH mechanism, of which Tomishige et al.[55] and Marin et al.[57] regard the formation of DMC* (S3) and the activation of CO2 (S1) as rate-determining steps, respectively. In this research, we take both viewpoints into consideration. For convenience, the apparent reaction rate expressions can be deduced from the mechanisms considering each step as an elementary reaction and defining the controlling step (full details are given in Supporting Information).

Table 8

Langmuir–Hinshelwood and Eley–Rideal mechanisms based on refs (54, 55)a

elementary reaction	Eley–Rideal mechanism	Langmuir–Hinshelwood mechanism
S1	MeOH + * ↔ MeOH*	CO2+ * ↔ CO2*
S2	MeOH* + CO2 ↔ MC*	MeOH + * ↔ MeOH*
S3	MC* + MeOH* ↔ DMC + H2O + *	2MeOH* + CO2* ↔ DMC* + H2O* + *
S4		DMC* ↔ DMC + *
S5		H2O* ↔ H2O + *
controlling step:	S2	S3 or S1
apparent rate law	R = k[CO2][MeOH][*]	R = k[CO2][MeOH]2[*] or R = k[CO2][*]3

* = Active sites; MC = methyl carbonate.

It can be inferred that the initial rate should be a function contrast the reactants concentration and catalysts loading amount. In addition, the rate constant should be strictly controlled by the reaction temperature and activation energy of the catalysts. Therefore, a series of initial rate measurement experiments were carried out in which the amount of CO2, methanol, and Ti0.04Ce0.96O2 nanorod catalyst was varied while retaining constant pressure of 1.0 MPa and temperature of 140 °C. The reaction was terminated at 120 min because the initial rate remains linear within 2 h, and the conversion of methanol and productivity of DMC were quantified by gas chromatography. The detailed reaction conditions and results of these experiments are shown in Table S1.

As predicted, there is an apparent direct positive correlation between the consumption of catalyst versus the initial rate as the catalyst load increases from 1.0 to 1.6 g (Figure ). The ln–ln plot of the initial rate versus active site concentration (measured in total surface area and expressed in m2) reveals the existence of an approximately first-order reaction, revealing that the Ti0.04Ce0.96O2 nanorods catalyzed heterogeneously without any mass transport limitations. The result indicates that the amount of catalyst loading strongly impacts the initial rate of the reaction, suggesting that the rate-limiting step of the reaction occurs on the surface of the catalysts.

Figure 5

Kinetic study of the initial rate of DMC production vs catalyst loading. Active sites are measured in total surface area and expressed in m2.

Table shows the analysis results of DMC formation with different CO2 and MeOH feedings. For the ER mechanism, we consider the combination of MeOH* and CO2 as the rate-determining step; thus, there should be a positive linear relationship between PCO·PMeOH and the initial rate. The ln–ln plots of these two parameters are given in Figure S8, and the negative slope of the fitting line suggests that the ER mechanism is inadequate to expound the kinetic process of DMC formation with the Ti0.04Ce0.96O2 nanorod catalyst in a fixed-bed reactor. These mainly account for wrongly ignoring the direct adsorption and activation of CO2 on the catalyst surface though it is much weaker than that of methanol.

Table 9

Analysis Results of the Initial Rate with Different CO2 and Methanol Feedings According to Table S1a

run	PCO2 (MPa)	PMeOH (MPa)	PCO2·PMeOH2	PCO2·PMeOH	initial rate of DMC (mmol/min)
1	0.764	0.236	0.043	0.181	0.910 × 10–2
2	0.795	0.205	0.033	0.163	0.944 × 10–2
3	0.819	0.181	0.027	0.148	0.976 × 10–2
4	0.838	0.162	0.022	0.136	1.021 × 10–2
5	0.801	0.199	0.032	0.159	0.952 × 10–2
6	0.729	0.271	0.054	0.198	0.853 × 10–2
7	0.698	0.302	0.064	0.211	0.842 × 10–2
8	0.669	0.331	0.073	0.222	0.784 × 10–2

Pressure 1.0 MPa; temperature 140 °C; catalyst weight 1.0 g; time on stream 2 h.

For the LH mechanism, with the MeOH* and CO2* combination as the rate-determining step, there should be a positive linear relationship between PCO·PMeOH2 and the initial rate, which is drawn as ln–ln plots in Figure S9 and which shows a negative liner dependence actually. Moreover, regarding the adsorption and activation of CO2 as the rate-controlling step, there is a direct relationship between PCO and the initial rate as the feeding of reactants varied (Figure ), giving a slope in close proximity to +1 (1.080). This indicates that the observed experimental rate equation was roughly as follows: initial rate = k [CO2]−1 [*]−1, which is consistent with the prediction of the LH mechanism with CO2 activation as the rate-determining step.

Figure 6

Kinetic study of the initial rate of DMC production with different CO2 and methanol feedings based on the LH mechanism with the activation of CO2 as the rate-controlling step.

Further investigation of the initial rate was carried out at 125, 130, 135, and 140 °C while maintaining constant 1.0 g load of the Ti0.04Ce0.96O2 nanorod catalyst, constant 1.0 MPa reaction pressure, and consistent feeding flow of methanol and CO2. The results are summarized in the Arrhenius plot (Figure ). The liner fitting slope of −5.565 indicates an apparent activation energy of 46.3 ± 5.0 kJ/mol for the Ti0.04Ce0.96O2 nanorod catalyst, which is lower than that of the neat CeO2 nanorod catalyst (65 kJ/mol).[57] This forcefully suggests that the incorporation of TiO2 in nanorods has improved the catalytic activity of the Ti0.04Ce0.96O2 nanorod catalyst through the improvements in surface active sites.

Figure 7

Arrhenius plot composed of initial rate data for the direct synthesis of DMC with Ti0.04Ce0.96O2 and neat CeO2 nanorod catalyst. The apparent activation energy (Ea) of this reaction was found to be 46.3 ± 5.0 kJ/mol for the Ti0.04Ce0.96O2 nanorod catalyst and 67.9 ± 7.5 kJ/mol for the neat CeO2 nanorod catalyst.

Conclusions

In summary, a new class of TiCe1–O2 nanorods catalysts for the direct DMC formation in a fixed-bed reactor is first synthesized and reported. The influences of the TiO2 doping ratio in the nanorods and various reaction conditions on the catalytic performances were investigated. It is found that the catalytic performance of ceria nanorods can be greatly enhanced with the introduction of TiO2 because the nanorod catalysts with more surface acidic–basic sites presented better catalytic activities. The Ti0.04Ce0.96O2 nanorod catalyst exhibits superior catalytic performances than those of neat CeO2 and TiCe1–O2 nanorods with other TiO2 doping ratios. Under optimal reaction conditions, the Ti0.04Ce0.96O2 nanorod catalyst exhibits the highest catalytic performance with a methanol conversion of 5.38% and DMC selectivity of 83.1% in a fixed-bed reactor. On the basis of the kinetic experiments utilizing an approach of the initial rate method, the reaction process and kinetics were studied for the direct synthesis of DMC over the Ti0.04Ce0.96O2 nanorod catalyst. An apparent rate equation is given: initial rate = k · [CO2]−1· [*]−1. These results are consistent with the prediction of the Langmuir–Hinshelwood mechanism where CO2 adsorbs directly on the catalyst surface with the CO2 activation as the rate-controlling step. Moreover, the activation energy barrier of the Ti0.04Ce0.96O2 nanorod catalyst is determined to be 46.3 ± 5.0 kJ/mol, which is lower than 67.9 ± 7.5 kJ/mol for neat CeO2 nanorods. However, the activation barrier of 46.3 kJ/mol remains unsatisfying. As a consequence, we forecast that the yield of DMC can be further improved by decreasing the activation energy barrier required for CO2 activation, maximizing the surface area of catalysts, and operating the catalytic process with a proper ratio of CO2 to methanol in the feed gas.

Experimental Section

Materials

Cerium(III) nitrate hydrate (Ce(NO3)3·6H2O; 99.5% metal basis) was purchased from Aladdin Co., Ltd. (Shanghai, China). Titanium(IV) oxysulfate-sulfuric acid hydrate (TiOSO4·xH2SO4·xH2O; synthesis grade, titanyl sulfate content 93.85%) was purchased from Macklin Biochemical Co., Ltd. (Shanghai China). Sodium hydroxide (99.5%, AR) was obtained from Guangdong Chemical Reagent Factory (Guangzhou, China). All of the chemicals were used without further purification.

Catalyst Preparation

Titania-doped Ceria nanorod catalysts with different TiO2 contents (x = 0.02, 0.04, 0.06, 0.08, and 0.10) were prepared using a modified hydrothermal method[42,57] that incorporated lyophilization to ensure the dryness of the catalysts. Briefly, stoichiometric Ce(NO3)3·6H2O and TiOSO4·xH2SO4·xH2O were mixed with 40 mL of a 6 M sodium hydroxide aqueous solution in a 100 mL Teflon liner. The liner was sealed in a stainless steel autoclave and retained in a convection oven for 24 h at 100 °C. The solid products were centrifuged and washed with ethanol and water several times until the supernatant became neutral. Afterward, the presynthesized nanorods were frozen in liquid nitrogen (−196 °C) and then freeze-dried in a bulk tray dryer (Four-ring Science Instrument Plant Beijing Corporation, Beijing, China) at a sublimating temperature of −40 °C and a pressure of 3 mbar. Finally, the catalysts were calcined at 600 °C under an air atmosphere by placing on a quartz boat in the center of a Muffle furnace and then slowly cool to room temperature. For contrast, ceria nanorod catalysts were prepared by the same method without the addition of TiOSO4·xH2SO4·xH2O.

Catalyst Characterization

Transmission electron microscopy (TEM) measurements were carried out on JEOL JSM-2010HR at 200 kV. The samples were dispersed into ethanol by ultrasonication and then dropped onto the copper grid and dried in air at room temperature.

X-ray powder diffraction was conducted on a Rigaku Dmax 2200 diffractometer. Graphite monochromatized Cu Kα radiation (λ = 0.154178 nm) at 40 kV/30 mA and a scan rate of 5°/min were applied in this characterization. The grain sizes of the particles were calculated based on the Scherrer equation.

N2 adsorption measurement was performed on a Micrometrics ASAP-2020 nitrogen adsorption apparatus. The sample was pretreated in a nitrogen flow at 200 °C for 2 h. When cooling to room temperature, the sample adsorbed N2 in liquid nitrogen under a flow rate of 110 mL/min. The parameters of the particles were calculated from the adsorption results using the BET methods.

Inductively coupled plasma atomic emission spectrometry (ICP-AES) measurements were performed on an Atomscan Advantage spectrometer, Thermo Ash Jarrell Corporation. The samples were digested by strong nitric acid in a Teflon autoclave and then diluted by 5% nitric acid solution. The content of TiO2 in nanorods was calculated according to eq .The X-ray photoelectron spectrum (XPS) of the sample was recorded on a ESCALAB250 (Thermo-VG Scientific) spectrometer with a monochromatized Al Kα source at 1486.6 eV, 15 kV, and 150 w. Survey scan spectra in the 1100–0 eV binding energy range were recorded with pass energy of 20 eV, and detailed spectra were recorded for the sample. An estimation of the amount of Ce(III) can be obtained from the intensity of the v0 (u0) and v′ (u′) lines according to eq .[58,59]Temperature-programmed desorption (TPD) was obtained on a Quantachrom Chem-BET 3000 apparatus according to the method established in our former research.[37] Samples were pretreated in nitrogen flow at 200 °C for 1 h. After cooling to room temperature, the samples were saturated with 10%CO2/90%N2 or 10%NH3/90%N2 for 30 min under a flow rate of 60 mL/min. Then, the physically adsorbed CO2 or NH3 was removed by streaming with 30%N2/70% He for 2 h under a flow rate of 50 mL/min. The samples were then heated up to 600 °C at a heating rate of 8 °C/min in N2/He. In addition, the amount of desorbed NH3/CO2 (in mmol/g) was obtained by back-titration, in which HCl/NaOH (0.01 mol/L) was used as the adsorbent of NH3/CO2 and NaOH/HCl (0.01 mol/L) was used as the titrant with a mixed indicator (1% bromocresol green ethanol solution and 2% methyl red ethanol solution with a volume ratio of 3:1).[60]

Catalytic Performance Measurement

The original selection of catalysts was conducted in a 50 mL autoclave with high-velocity mechanical stirring. Fresh catalyst (0.2 g) and absolute methanol (20 mL) were put into the autoclave. CO2 was purged into the reactor and replaced the air inside several times to assure oxygen-free reaction conditions. The initial pressure of CO2 was set as 5 MPa, and then it was reacted at 120 °C for 5 h. The resultants were filtrated by a PES membrane (pore size, 0.45 μm) and then measured by GC-7900II equipped with a flame ionization detector (FID).

Direct synthesis of DMC from methanol and CO2 was carried out in a continuous tubular fixed-bed microreactor. The diagram of the apparatus is shown as Schedule S1. Nitrogen was purged into the tubular reactor to assure oxygen-free reaction conditions before the catalyst bed was heated up to a set temperature. CO2 was purged into the methanol container to get a CO2/methanol mixed gas. The mixed gas was then charged into the reactor. The resulting productions were measured by GC-7900II equipped with a flame ionization detector (FID) through a six-way valve.

Further kinetics exploring experiments were carried out in an optimized fixed-bed reactor equipped with a high-performance liquid chromatography (HPLC) pump. The diagram of the apparatus is shown as Schedule S2. Nitrogen was purged into the tubular reactor to replace the air inside. CO2 was purged into the reactor directly. Liquid methanol was pumped into the reactor through the HPLC pump. The resulting productions were gathered offline and then measured by GC-7900II equipped with a flame ionization detector (FID).