Recent Development

CHAPTER 1 OVERVIEW

1.2 Recent Development

Recently, CO2 has been considered as a storage of renewable carbon sources for materials or hydrocarbon synthesis; however, the required energy for the CO2 activation is large due to the thermodynamic stability of CO2. It limits the possibility for using CO2 as one of the carbon sources. For the reason, high energy substances, such as hydrogen, are commonly used for the transformation of CO2. Wei Wang et al. ¹had a schematic illustration for the products of CO2 hydrogenation. The products could be divided into chemicals and fuels. Especially like methane, methanol, dimethyl ether (DMF) and some hydrocarbons are good fuels, easily to storage. Until now, CO2

hydrogenation has been extensively carried out in both homogeneous and heterogeneous catalysis. Homogeneous catalysis gives good selectivity, but get problems in recovery, regeneration, and the separation of catalysts from precursors or products. In contrast, heterogeneous catalysis is right a solution for that even though the relatively poor performance.

Figure 1. Possible products from CO2 hydrogenation.

The typical format of the heterogeneous catalysts is metallic nanoparticles on an oxide support. Hence, shape and size control on metallic nanoparticles and the composition of the oxide support play the important roles because they are usually correlated with active site geometries, which directs the change of the catalytic activity.

John C. Matsubu ² found an isolated metal atom on the oxide support also plays an key role in improving the activity of heterogeneous catalysis with Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS). Based on their findings, John C.

Matsubu et al. focus on two major reactions for CO2 hydrogenation, reverse-water gas shift (r-WGS, CO2 + H2  CO + H2O) and methanation (CO2 + 4H2  CH4 + 2H2O), to study how the isolated active site affects the reaction direction. It turned out that Rhodium nanoparticle sites (RhNP sites) can selectively enhance methanation while Rhodium isolated sites (Rhiso sites) improve r-WGS.

1.3 Research Motivation

Platinum group metal, including Platinum (Pt), Palladium (Pd) and Rhodium (Rh) is generally used in CO2 hydrogenation due to their good catalytic activity. In recent year, more and more researchers engaged in morphosynthesis of Rhodium nanoparticles;

however, most strategies contain the use of organic solvents and strong capping agents such as poly(vinylpyrrolidinone) (PVP) which seriously passivate the surfaces of nanocatalysts.³ Herein, we developed a facile one-pot strategy for the aqueous synthesis of well-defined Rhodium twinned, tetrahedral, and concaved tetrahedral nanocatalysts.

In the synthetic condition, the ionic surfactant hexadecyltrimethyl ammonium bromide (CTAB) was used instead of PVP, which left the catalyst surfaces being easily cleaned for getting better catalytic performance in CO2 hydrogenation.

CHAPTER 2 INTRODUCTION

2.1 Catalysis

Catalysis meaning “something is added to facilitate a reaction, but without been consumed or produced during the whole reaction.” ⁴ was first described by Berzelius in 1836. In chemistry, catalysis can be considered as changing the energy profiles of the reaction pathway to present a lower activation energy (Figure 2.1 blue solid line), which speeds up the reaction rate as comparison to the non-catalyzed reaction pathway (Fig.

2.1, purple dotted line). It sometimes lead to a different product if the intermediate is modulated (green solid line, formation of product B).

Figure 2.1. Energy diagram of a generic reaction and the effects of a catalyst on the reaction profile. Dotted line indicates an uncatalyzed reaction and solid lines means catalyzed reactions which results in different products.  

2.1.1 Homogeneous and Heterogeneous Catalysis

The difference between homogeneous and heterogeneous catalysis is whether the reaction and catalysts in the same or different phases, including two insoluble liquid phases (Figure 2.2). Their advantages and disadvantages are compared in Table 2.1. As summaried, homogeneous catalysts have every single catalytic entity act as the active sites so as to perform very high activity and selectivity.⁴ Nevertheless, the homogeneous catalysts cost much higher than heterogeneous catalysts for separation from the reaction system because of the soluble phase issue. In heterogeneous catalysis, the catalysts are usually loaded on a support for dispersion and stabilization. In this way, they can be easily separated and retrieved from the unreacted reactants and products with physical methods which should be cheap and practical in mass production.

Unfortunately, the active sites of heterogeneous catalysts are located at the surfaces where are easily contaminated by various factors. As a result, the activity (TOF) is mostly poor unless a pretreatment is done before catalytic reactions. In addition, unlike homogeneous catalysts whose selectivity is adjustable with the coordinating ligands, the active sites of heterogeneous catalysts have to be adjusted through the modulation of surface structures. It usually requires the introduction of strong capping agents or other kinds of impurities that probably passivate the active sites, too. Therefore, these issue leaves plenty of space for improvement scientifically.

Table 2.1. Comparison of homogeneous and heterogeneous catalysts.

Figure 2.2. Schematic illustration of heterogeneous phases of (a) liquid-liquid, (b) solid-liquid, and (c) homogeneous liquid phase

2.1.2 Heterogeneous Catalysis

In a typical process of heterogeneous catalysis, it includes adsorption of reactants from fluid phase onto the catalyst surfaces, dissociation of reactants and formation of intermediates over catalyst surfaces, and desorption of products back to the fluid phase.⁵ The whole procedure was firstly demonstrated by a metal catalyzed ethylene hydrogenation^6,7 when a mixing gas of ethylene and hydrogen was used to flow through nickel surfaces.

Adsorption

In the first step of heterogeneous catalysis, “adsorption” involves the activation of reactants by forming strong chemical bonds when reactants attach to the catalyst surfaces. This type of adsorption is called chemical adsorption, and the surface sites binding with reactants are called active sites. As show in Figure 2.3a, ethylene adsorbs on the nickel surface in which dissociation of the carbon C=C double bond and

hydrogen molecules to bind onto the nickel surface. The adsorption of reactants must be monolayered since all the active sites occupied. Van der Waals forces also provides another way for every reactant molecule to adsorb on catalyst surface, this type of weaker adsorption is called physical adsorption.⁵ Even though physical adsorption cannot active the adsorbed reactant, it still can serve as a precursor for chemical adsorption, besides, the surface area of catalyst can be measure by the physical adsorption of monolayer inert gas.

Surface Reaction

The simplest definition of surface reaction is the irreversible conversion of adsorbed reactant into product molecule, like the binding between ethylene and nickel is replaced by the bond between carbon and hydrogen, then leaves from nickel surface (Figure.

2.3b).

Desorption

As shown on Figure 2.3c, the resulting ethane are released from the Ni surface after both hydrogen atoms bind to the ethylene. , leaving the space (active sites) for other reactants.

Figure 2.3. Scheme of ethylene hydrogenation with Ni catalysts. (a) Ethylene and hydrogen molecule adsorbing on the Ni catalyst surface. (b) One side of ethylene reacts with hydrogen and detach from the catalyst. (c) Double sides of ethylene reacts with hydrogen and thus fully desorbs from the Ni catalyst surface.

2.2 Examples of Heterogeneous Catalysis

Nowadays, metallic nanoparticles loading on oxide supports has been widely studied in heterogeneous catalysis, such as ethylene hydrogenation,^8-11 oxidation,^12,13, epoxidation,¹⁴ CO oxidation,¹⁵ and CO2 hydrogenation etc.^16-20 As we know, catalytic performance can be affected by various factors, including catalyst size, shape, chemical composition particle, and the interaction between metal and support.^21-23 (Figure 2.4).

Generally speaking, the catalytic performance increases as the metallic nanocatalysts size decreases due to the increment of the total surface area and thus activity ²¹. Besides, nanocatalyst facets/shapes and compositions lead to different products.^24,25

Figure 2.4. Scheme of metallic catalysts located on oxide support.²²

2.2.1 Support Effect : An example of CO Oxidation

S. Arrii et al. ²⁶ demonstrated a study to show the metallic catalyst activity effect on different support with CO oxidation reaction,. Those results represent that supports can be classified into two categories—“inert” and “active”, showing the same trend with what Schubert proposed in 2001.²⁷ For “inert” materials, such as Al2O3, due to the nonreducibility and low ability to adsorb or store oxygen, lead to the worst activity in CO oxidation, even though Au plays the major role in whole procedure. TiO2 and ZrO2

are composed by reducible oxide, belonging to “active” materials.

Figure 2.5. Use of different support materials can get different CO oxidation activity (Au).

2.2.2 Shape Effect: CO Oxidation as an example

For better comparison, CO oxidation is also used in this section as one of the examples to demonstrate shape-dependent catalytic activity. Rui Wang et al.²⁸ synthesized three kinds of palladium (Pd) nanoparticles following previous methods, including {100} facet dominant nanocubes (20.5 nm), {111} facet dominant octahedra (22.4 nm), and non-controlled spherical nanocrystals (3.9 nm) as shown in Figure 2.6a, b and c.^29-31 After loading onto the support, the characterized TEM images and XRD patterns were also shown in Figure 2.7 and 2.8. According to the TEM images, they proved that all the three shapes didn’t change after loading procedure. The XRD patterns further verified the compositions of catalysts and the major facets of each shape.

Figure 2.6. TEM and HRTEM images of Pd nanocrystals in the shape of (a,b) cube, (c,d) octahedron, and (e,f) sphere. The insets show the corresponding representative modes of the typical shapes.

Figure 2.7. TEM images of as-prepared (a) Pd (cube)/SiO₂, (b) Pd (octahedron)/SiO₂, and (c) Pd (sphere)/SiO₂ catalysts.

Figure 2.8. XRD patterns of (a) Pd (cube)/SiO2, (b) Pd (octahedron)/SiO2, and (c) Pd (sphere)/SiO2 catalysts.

Catalytic performance of CO conversion with increasing temperature shows that Pd activity increases in the order of cube < octahedron < sphere (Figure 2.9). Because Pd spheres have the lowest activation energy (42.6 kJ/mol) while Pd cubes have the highest activation energy (76.5 kJ/mol). It’s worth noting that, the CO oxidation activities of 3.9, 7.5 and 9.6 nm Pd spheres are almost the same, meaning the size effect can be minor here. Thermal programing deposition (TPD) was also carried out at 70°C to figure out the relationship between shape and activation (Figure 2.10). It turned out that

octahedra and spheres with {111} faces get stronger chemical adsorption of CO molecules than the cubes with {100} faces.

Figure 2.9. CO oxidation activity of (a) Pd (cube)/SiO₂, (b) Pd (octahedron)/ SiO₂, and (c) Pd (sphere)/SiO₂ catalysts with 1.0% CO and 1.0% O2 in N2 at a space velocity of 32.4 mL s-1 g-1.

Figure 2.10. CO-TPD spectra of the (a) SiO₂, (b) Pd (cube)/SiO₂, (c) Pd (octahedron)/SiO₂, and (d) Pd (sphere)/SiO₂ catalysts. The inset shows the corresponding CO₂ desorption profiles on the surface of three different catalysts and SiO₂.

2.2.3 Sites Effect: CO

Hydrogenation as an example

The active site geometry of shape-controlled nanocatalysts are regarded correlative with catalytic activity. However, not only the nanoscale particles but also an isolated metal atoms on oxide supports can play a critical role in improving the performance of heterogeneous catalysis. To study how isolated active site affects the reaction, John C.

Matsubu et al.² focused on the two major reactions of CO2 hydrogenation, reverse-water gas shift (r-WGS, CO2 + H2  CO + H2O) and methanation (CO2 + 4H2  CH4

+ 2H2O), finding Rhodium nanoparticle sites (RhNP sites) can selectively enhance methanation. In contrast, Rhodium isolated sites (Rhiso sites) improve r-WGS (Figure 2.11).

Figure 2.11. Scheme of the selectivity between Rhiso sites— rWGS and RhNP sites—

methanation.

Diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) is an infrared spectroscopy technique used on powder samples with no preparation. It was typically used to monitor the bonding states of molecules on catalysts. In the whole procedure of the experiment, the first step is to identify the peaks in the DRIFT spectra (Figure 2.12) In Figure 2.12a, there are three sharp peaks. The peak in the middle denotes that a CO molecule linearly binds on the RhNP sites. The left and right are symmetric and

asymmetric stretching of Rh(CO)2 gem-dicarbonyl species at Rhiso sites. The broad peak at 1850 cm^-1 is of the bridge binding of CO molecules on RhNP sites (Figure 2.12a).

32-35 The results in Figure 2.12b were obtained at different loading weight percent (w.t.%) (Figure 2.12 b), and finally converted in to plot of site fraction vs loading weight percent (Figure 2.12c) according to eq 2.2.

Figure 2.12. (a) DRIFT spectrum obtained from a saturated layer of CO adsorbed at 300 K on 4% Rh/TiO₂. Insets show ball-and-stick models of assigned vibrational modes.

(b) DRIFT spectra of CO on all five weight loadings of Rh/TiO₂catalysts. The spectra are displayed in Kubelka−Munk (KM) units and normalized by the symmetric gem- dicarbonyl peak (2097 cm⁻¹) height to allow for comparison. (c) Site fraction (%) of isolated (Rh_iso) and nanoparticle-based Rh sites (Rh_NP), calculated based on eq. 2.2 and the spectra in (b), as a function of w.t.% Rh.

𝑋 iso = ^𝐼

^iso

^/(𝜀

^iso

^×(

𝐶𝑂 𝑅ℎ )

iso

)

∑ [𝐼

/(𝜀

×( ^𝐶𝑂

𝑅ℎ )

)]

3 𝑖=1

X NP = 1 − X iso

eq.(2.2)

After reaction, site fractions in a similar trend are found within the relational plots of Rhiso sties vs w.t.% (Figure 2.13), which suggests the obvious reaction selectivity on the Rhiso sites in CO2 hydrogenation. To further investigate, 100 mg of Rh nanocatalysts loaded TiO2 was treated with a 10 ml of HCl/H2O2 solution to leach RhNP sites (Figure 2.14a). After treatment, the TOF of methanation showed a drastic decrease (Figure 2.14b, c), again confirming RhNP sites had the specific selectivity in CO2 methanation while Rhiso site had the same trend in r-WGS.

Figure 2.13. Rhiso site fraction and r-WGS TOF plotted as a function of w.t.% Rh at (a) 1 CO₂:4H₂, (b) 3CO₂ : H2, and (c) 10CO₂ : H2 feed ratios. The left axes are Rhiso site fractions, which are displayed in the plots as a black line connecting the measured values for graphical clarity. The green, blue, and red data points correspond to measured r-WGS TOF and are quantified in the right axis of each plot.

Figure 2.14. (a) DRIFT spectra for CO adsorbed on the fresh 2% Rh/TiO2 and HCl/H2O2 leached samples, where spectra are displayed in KM units and normalized by the symmetric gem-dicarbonyl peak (2097 cm−1) height. (b) TOF for both CO2

reduction pathways measured on the fresh 2% Rh/TiO2 and leached samples at 200°C and a feed ratio of 3CO2 : H2. The quantified number of Rh sites for the fresh catalyst was also used for the TOF calculations in the leached sample; see main text for discussion on this. (c) CH4 selectivity plotted for the fresh and leached samples at 200°C and a feed ratio of 3CO2:H2.

2.3 Synthesis of Nanoparticles and Platinum Group Metal

According to discussion before, nanocatalyst morphology undoubtedly plays an important role in catalyst performance. Hence, how to precisely modulate nanocatalysts morphology and correctly set up the connection between the shape of nanocatalysts and catalytic performance of CO2 hydrogenation becomes a challenging but extremely critical task.

2.3.1 Introduction to Nanoparticle Synthesis

When it comes to nanoparticle synthesis, shape control is a topic can never be ignored. Yujie Xiong and Younan Xia ³⁶ categorized most pathways of nanoparticle

growth and summarize as a flow chart in Figure 2.15. Accordingly, the flow chart is roughly divided into two sections including nucleation and formation of seeds, and seed-mediated growth.

Figure 2.15. A schematic illustration of the reaction pathways that lead to Pd nanostructures with different shapes. As the essence of a synthesis, a palladium precursor is reduced to produce Pd atoms, which subsequently aggregate to form nuclei.

Once the nuclei have grown past a certain size, they be- come seeds with a single-crystal, single-twinned, or multiple-twinned structure. If stacking faults are involved, the seeds will grow into plate-like nano- structures. The green, orange, and purple

colors represent the {100}, {111}, and {110} facets, respectively. Twin planes are delineated in the figure with red lines. The parameter R is defined as the ratio between the growth rates along the <100> and <111> axes.

Nucleation and formation of seeds

Generally speaking, nucleation of nanoparticles in a solution phase can be broadly defined as the procedure that atoms in solution form a small cluster, in other word, a seed which has a stable, well-defined crystal structure. This process usually happens in either homogeneous or heterogeneous nucleation that were clarified by Andrea R. Tao et al. ³⁷

In homogeneous nucleation, seed formation and growth occurs simultaneously.

According to La Mer model, reduction of metal ions exhibits a critical reactant concentration in solution. Nucleation will occur when the concentration is over the critical limit. However, further nucleation can be hindered with the gradual consumption of reactants. In this way, the following growth generally take place on the pre-existed nuclei. It says that the formation of nuclei must be quite rapid in order to reach highly shape-monodispersive nanoparticles.

Figure 2.16. Nucleation and growth kinetics of metal nanoparticles. The La Mer model of nucleation, where a critical reactant concentration is required for particle nucleation.

Homogenous nanoparticle dispersions are favored by a single, rapid nucleation event.

Heterogeneous nucleation is typically carried out by introducing pre-formed seed nanoparticles to the reaction solution, indicating the independent processes of seed formation and growth. Therefore, the activation energy of growth, which metallic precursors are reduced on the seed nanocrystals, is definitely lower than homogeneous.

Accordingly, heterogeneous nucleation usually gives higher possibility in the shape control with milder reducing agent or lower temperatures.

Seed-mediated Growth

After nucleation from metal ions, morphology control of nanocrystals can be achieved by adding molecular additives or capping agents, selectively adsorbing on specific crystal planes to induce formation of corresponding morphologies.

2.3.2 Shape Control of Nanoparticle

Taking gold as an example, Mark R. Langille et al. demonstrated a way in controlling nanoparticle morphology using different concentrations of reducing agents to affect the reaction kinetics during synthesis.³⁸ As a result, kinetically favorable shapes often form if increasing the amount of reducing agents in the growth solutions. As Figure 2.17 shows, 0.5 mM and 2 mM of ascorbic acid gives well-defined shape of [111] faceted octahedra and [100] faceted cubes, respectively. However, the trisoctahedra are obtained when the ascorbic acid concentration comes to 10 mM. It is arisen form the reduction rate of gold ions become faster than that of octahedra and cubes and therefore leads to such result.

Figure 2.17. SEM images of reaction products from growth solutions containing 10 mM CTA-Br and (A) 0.5, (B) 2.0, and (C) 10.0 mM ascorbic acid, resulting in the formation of {111}-faceted octahedra, {100}-faceted cubes, and high-index faceted trisoctahedra, respectively. Scale bars in all images are 200 nm.

Halide ions existing in growth solutions affect particle shape, too. In typical, halide ions in the solution form the CTA-X-AuX2– complexes with gold precursors accompanied by the changes in solubility, reduction potential, and reduction rate. The reduction potentials of [AuX2─] complexes and solubility decrease in the order of

[AuCl2─] > [AuBr2─] > [AuI2─] (1.154 > 0.960 > 0.578 V). It results in a more challenging condition in the reduction of gold precursors once if bromide or iodide ions exist in the growth solution, gold ions are much challenging for same amount of ascorbic acid to reduce, thus slowing down gold ion reduction rate. The solubilities of [AuX2─] complexes in fact decrease in the same order, and the decreasing solubilities relative to the high solubility of CTA-X-AuX2– will also slow the rate of gold reduction.

In addition, halide ions can bind to the surface of the gold nanoparticles which inhibit the subsequent growth of on the surfaces. The binding strength between halide ions and particle surfaces is in the order of Cl^– < Br^– < I^–. Hence, the addition of bromide or iodide ions will slow the formation rate of gold particles compared to that of chloride ions. .

To verify how halide ions function on the morphology control of nanoparticles, control experiments using pure CTA-Cl and a CTA-Cl/CTA-Br mixture by adding NaBr to CTA-Cl was carried out for comparison. Figure 2.18A shows that the condition of 50 mM CTA-Cl without adding NaBr results in the formation of trisoctahedra. It indicates that gold precursors are reduced much faster in CTA-Cl system than those in CTA-Br system. In contrast, the condition of 50 mM CTA-Cl, 5 mM NaBr and 0.5 mM HAuCl4 in the growth solution, gives rise to the products of [100]-faced nanocubes (Figure 2.18b). Accordingly, the introduction of Br^– ions to the CTA-Cl system is capable of altering the reduction rate of gold precursors. This estimation was confirmed by a kinetic measurement of the formation of Au⁰ atoms with ICP-AES (Figure 2.18c) in which Au⁰ amount increases with time. As shown, the formation rate of Au⁰ atoms

在文檔中多型態銠金屬奈米晶體之合成與其在二氧化碳氫化上之應用 (頁 16-0)