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

全文

(1)國立臺灣師範大學化學系碩士論文. 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. Morphology-Dependent CO2 Hydrogenation with Rhodium nanocatalysts. 研究生 : 陳韋潔. 指導教授 : 郭俊宏博士. 中華民國一百零五年七月.

(2) 摘要. 銠金屬奈米晶體因其良好的活性使其在催化反應的研究上備受矚目。在本論文中，我們描述如何利用一步一鍋化的水相合成法，藉由調控反應劑的使用量，成功合成了具有不同表面晶面的銠金屬奈米晶體，並將其應用至二氧化碳氫化反應上，探討生成甲烷的選擇性與晶體結構之關係。我們探討的銠金屬奈米結構包含凹面、鑿面與雙晶相奈米晶體，其中凹面晶體為具有高均一性的四面體結構。此三種形態可藉由在反應條件中，調整界面活性劑—溴化十六烷基三甲铵的濃度高低而互相轉變。以凹面四面體的條件為核心，溴化十六烷基三甲铵濃度較低時可形成鑿面奈米晶體，然而較高時可得到具雙晶相的奈米晶體。此三種結構經由電子顯微鏡、X 光電子能譜儀等分析可被證實為面心立方的純銠金屬，但在表面有因暴露空氣下而產生的氧化銠層。在二氧化碳氫化的催化反應中，此三種結構皆被分散在氧化鋯上製備成非勻相觸媒。在條件測試中，三者皆在溫度高於攝氏 400 度才有明顯二氧化碳轉化率，且轉化率隨溫度上升。然而，甲烷的生成只有在凹面四面體的結構上甚為明顯，其餘皆產生高比率的一氧化碳。經參考文獻結果與推測探討，此凹面四面體銠奈米觸媒因具有 110 的晶面存在於凹面中，而改變了催化反應的途徑，提高了甲烷生成的選擇性，而存在於表面的 111 與 100 晶面則趨向於一氧化碳反應途徑的生成。此項發現不僅證實了觸媒表面結構與產物選擇性的關係，更提供一個極重要的方法，將二氧化碳高選擇地轉化成高經濟價值的甲烷。. I.

(3) ABSTRACT Rhodium (Rh) is a widely used metal by virtue of its superior performance in hydrogenation. In this thesis, the one-pot aqueous synthesis was developed to synthesize shaped Rh nanocrystals as catalysts and applied to CO2 hydrogenation for investigating the relation the between crystal morphology and the selectivity in methane formation. The three kinds of Rh catalysts we investigated were concaved, excavated and twinned nanocrystals, among which the concaved nanocrystals were in highly monodispersive size and tetrahedral shapes. The three nanocrystals could be obtained in the same reaction condition except for the concentration of capping agent, CTAB. Compared to the use amount of CTAB in the synthesis of concaved tetrahedra, the lower CTAB amount induced the formation of excavated nanocrystals while the higher resulted in the twinned ones. Characterized by SEM, TEM, and XPS, all shaped nanocrystals were confirmed the pure Rh with the Rh2O3 layers over their surfaces because of the exposure in the air. In CO2 hydrogenation, they exhibited significant activity in CO2 conversion when the temperature was over 400˚C. Notably, a relatively high yield ratio of methane was observed taking place when the concaved Rh tetrahedra were used as catalysts. In contrast, the other two gave high yield ratio of CO. Referring to former literature and the results of structural analysis. The high selectivity of methane formation possibly came from the existence of 110 crystal faces over the concaved surfaces of a tetrahedron; however, the 111 and 100 crystal faces typically caused the favored formation of CO. This is undoubtedly an exciting discovery which not only validates the relation between crystal structure and product selectivity in CO2 hydrogenation, but also provides a promising road leading to efficient conversion of CO2 to the value-added chemicals, such as methane.. II.

(4) 中文關鍵字奈米晶體銠金屬一鍋法合成二氧化碳氫化甲烷一氧化碳. English Keyword Nanocrystals Rhodium One-Pot Synthesis CO2 Hydrogenation Methane Carbon monoxide. III.

(5) ACKNOWLEDGEMENT. 昔日剛進入實驗室的情景依然記憶猶新，而今轉眼間已過了兩年。感謝郭俊宏老師的指導，才能有今日得以完成一個有階段性成果的研究，也謝謝美瑩姐在操作 SEM 技巧的指導，和測試樣品時的鼎力相助。在中研院化學所的兩年時光讓我遇到了許多才華洋溢的人，得到了許多無法從紙本上學到的知識。在全英文的實驗室會議報告中依然可以侃侃而談的柏蘅學長，為實驗室的報告標準展現了最佳典範；不管是學校課業上的困難或是實驗原理上的疑惑，在猶如人形百科全書般博學多聞的智文學長面前都可以獲得解答；從做事嚴謹目的明確的家俊學長身上，學到了做任何事都需要掌握節奏切中要點，面對困難時依然要勇於面對堅持下去；在實驗觸礁時，煉明學長也用豐富的學理知識，從不一樣的觀點切入使我豁然開朗。兩年來，謝謝 Sonya、Tika、Reddy、苡娸、世承的陪伴與鼓舞，讓我即使在實驗上遇到挫折碰壁也不致灰心放棄。謝謝口試委員李位仁老師、黃暄益老師還有王迪彥老師們，在口試時從不同的觀點給我更多的指導與建議，讓我得到更多收穫。最後謝謝佳慧和玳寧的幫忙，祝你們未來的實驗順利。. IV.

(6) TABLE OF CONTENENT 摘要................................................................................................................................. I ABSTRACT......................................................................................................... II ACKNOWLEDGEMENT .................................................................................. IV TABLE OF CONTENENT .................................................................................. V LIST OF FIGURES ........................................................................................... VII LIST OF TALBES ....................................................................................................XIII CHAPTER 1 OVERVIEW ......................................................................................... 1 1.1. Research Background ................................................................................1. 1.2. Recent Development................................................................................... 2. CHAPTER 2 INTRODUCTION ................................................................................4 2.1 Catalysis...........................................................................................................4 2.1.1 Homogeneous and Heterogeneous Catalysis........................................5 2.1.2 Heterogeneous Catalysis ......................................................................6 2.2 Examples of Heterogeneous Catalysis ............................................................. 8 2.2.1 Support Effect : An example of CO Oxidation ......................................9 2.2.2 Shape Effect: CO Oxidation as an example ........................................10 2.2.3 Sites Effect: CO2 Hydrogenation as an example ................................ 14 2.3 Synthesis of Nanoparticles and Platinum Group Metal ................................ 17 2.3.1 Introduction to Nanoparticle Synthesis ..............................................17 2.3.2 Shape Control of Nanoparticle ........................................................... 21 CHAPTER 3 EXPERIMENTAL SECTION ....................................................... 29 3.1 Chemicals .......................................................................................................29 V.

(7) 3.1.1 Synthesis of Rhodium Nanoparticles .................................................. 29 3.2 Instrumentation .............................................................................................. 29 3.3 Procedure .......................................................................................................30 3.3.1 Synthesis of Excavated Tetrahedron, Concaved Tetrahedron, and Twinned Nanoparticles ................................................................................ 30 3.3.2 Preparation of Rhodium Catalyst ....................................................... 30 3.3.3 Catalytic Performance Experiment for CO2 Hydrogenation ..............31 3.4 Characterization ............................................................................................ 32 3.4.1 Scanning Electron Microscopy (SEM)................................................ 32 3.4.2 Transmission Electron Microscope (TEM) .........................................32 3.4.4 X-ray Photoelectron Spectroscopy (XPS) ...........................................32 3.4.5 Inductively Coupled Plasma with Atomic Emission Spectroscopy (ICP-AES) ....................................................................................................33 CHAPTER 4 RESULTS AND DISCUSSION ......................................................... 36 4.1 Shape-Controlled Rhodium (Rh) Nanoparticles ............................................36 4.1.1 The Role of Reducing Agent ................................................................ 39 4.1.2 The Influence of Surfactant .................................................................43 4.1.3 Temperature effect ..............................................................................46 4.2 Test of Catalytic Performance .......................................................................47 4.2.1 Spectrum Identification and Stability Test ..........................................47 4.2.2 Shape-dependent Catalytic Performance ...........................................51 CONCLUSION ..........................................................................................................55 REFERENCES ...........................................................................................................56. VI.

(8) LIST OF FIGURE. Figure 1. Possible products from CO2 hydrogenation. ................................................. 2 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. ................................................ 4 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. ...................................... 8 Figure 2.4. Scheme of metallic catalysts located on oxide support.22 .......................... 9 Figure 2.5. Use of different support materials can get different CO oxidation activity (Au). ............................................................................................................................. 10 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 ......................................................................................... 11 Figure 2.7. TEM images of as-prepared (a) Pd (cube)/SiO2, (b) Pd (octahedron)/SiO2, and (c) Pd (sphere)/SiO2 catalysts................................................................................ 12 Figure 2.8. XRD patterns of (a) Pd (cube)/SiO2, (b) Pd (octahedron)/SiO2, and (c) Pd (sphere)/SiO2 catalysts ................................................................................................ 12 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. ........................................................................................................... 13. VII.

(9) 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. ............................................................................................................................. 13 Figure 2.11. Scheme of the selectivity between Rhiso sites—rWGS and RhNP sites— methanation. ................................................................................................................. 14 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/TiO2 catalysts. 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 (Rhiso) and nanoparticle-based Rh sites (RhNP), calculated based on eq. 2.2 and the spectra in (b), as a function of w.t.% Rh. ........................................................ 15 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. ..................................... 16 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 VIII.

(10) discussion on this. (c) CH4 selectivity plotted for the fresh and leached samples at 200°C and a feed ratio of 3CO2:H2. ............................................................................. 17 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 become seeds with a singlecrystal, single-twinned, or multiple-twinned structure. If stacking faults are involved, the seeds will grow into plate-like nanostructures. 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. ..................................................... 18 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. ...................................................................................................................................... 20 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. ................................ 21 Figure 2.18. (a,b) SEM images of reaction products from growth solutions containing 50 mM CTA-Cl and (a) 0.0 and (b) 5.0 mM NaBr, resulting in the formation of highindex faceted trisoctahedra and {100}-faceted cubes, respectively. Scale bars are 200 nm. (c) ICP- AES kinetics data of the reactions containing 0.0 mM (black squares) and 5.0 mM (red triangles) NaBr. ....................................................................................... 23. IX.

(11) Figure 2.19. (a−c) SEM images of reaction products from growth solutions containing 50 mM CTA-Cl and (a) 0.0, (b) 10.0, and (c) 75.0 μM NaI, resulting in the formation of high-index faceted trisoctahedra, {111}-faceted octahedra, and {111}-faceted truncated bitetrahedra, respectively. Scale bars: 200 nm. The truncated bitetrahedra are the planar twinned analogue of the single crystalline octahedra. (d) ICP-AES kinetics data of the reactions containing 0.0 μM (black squares) and 75.0 μM (red triangles) NaI. ...................................................................................................................................... 24 Figure 2.22. Structural characterizations of Rh concave tetrahedrons synthesized using the standard procedure (TriEG and 145 °C): (a) low- magnification TEM image, (b) tilted TEM images of two Rh concave tetrahedrons recorded at different tilting angles, (c, d) high-resolution HAADF STEM images of a concave tetrahedron recorded along [111] and [211] zone axes, respectively, together with the corresponding atomic models in the insets. The scale bars in b are 10 nm. ................................................................ 26 Figure 2.23. (a−d) TEM images of Rh nanocrystals obtained using the standard procedure at different reaction times after the injection of Rh(III) precursor: (a) 6, (b) 15, (c) 30, and (d) 60 min. The scale bars in all the insets represent 5 nm. (e) A schematic illustration showing the major steps involved in the formation of Rh concave tetrahedrons: (1) corner-selected deposition, (2) diffusion from corners to edges at a rate of V2, and (3) diffusion from corners/edges to side faces at a rate of V3. ........... 27 Figure 2.24. (a−c) TEM images of Rh nanocrystals synthesized using the standard procedure except for the use of different polyols as the solvents: (a) EG, (b) DiEG, and (c) TetraEG. (d) A plot of the average edge lengths of the Rh nanocrystals obtained from different polyols. ................................................................................................. 28. X.

(12) Figure 3.1 Schematic illustration of (a) wall-coated open tubular column (WCOT), (b) porous-layer open tubular column (PLOT), (c) support-coated open tubular column (SCOT)......................................................................................................................... 34 Figure 3.2 Schematic interpretation of the TCD working principle. .......................... 35 Figure 3.2. The scheme of the heterogeneous catalysis system. (a) sampling system, (b) reaction section, (c) analysis section. ........................................................................... 35 Figure 4.1. (a) Low-, (b) middle- and (c) high-magnification TEM images of Rh tetrahedral nanoparticles. ............................................................................................. 37 Figure 4.2. (a) TEM image and (b) the corresponding SAED pattern of a concaved Rh. (c) Size-distribution histogram concaved tetrahedral with 100 particles counted. ...... 38 Figure 4.3. 3d XPS fitting result before loading on oxide support. ............................ 38 Figure 4.4. (a) low-magnification and (b) middle-magnification TEM image using methanol as reducing agent. (c) low-magnification (d) middle-magnification and (e) HR-TEM image of a cubic Rh nanoparticle prepared with formaldehyde to replace formic acid. The lattice parameter (1.9 Å and 1.87 Å ) is referred to rhodium 200 faces. ...................................................................................................................................... 41 Figure 4.5. TEM images of Rh nanocrystals prepared with (a) 0.05 M, (b) 0.1 M, (c) 0.2 M, (d) 0.3 M, (e) 0.4 M concentrations of formic acid. The small picture beside (e) is the scheme for hierarchical structure. ................................................................. 42 Figure 4.6. TEM images of Rh nanoparticles prepared with (a) 0.005 M, (b) 0.01 M, (c) 0.025 M, (d) 0.05 M (e) 0.075 M CTAB. ............................................................... 44 Figure 4.7. Magnified TEM images of Rh nanoparticles prepared with (a) 0.005M, (b) 0.01M, (c) 0.025M, (d) 0.05M. (e) 0.075M CTAB. .................................................... 45 Figure 4.8. (a) low-, (b) middle- and (c) high-magnification TEM images of Rh nanoparticles prepared with 0.01M CTAC instead of CTAB...................................... 45. XI.

(13) Figure 4.9. TEM images of Rh nanoparticles synthesized at (a) 75 °C, (b) 80 °C, (c) 90 °C, (d) 100 °C and (e) 110 °C. ..................................................................................... 47 Figure 4.10. Intensity vs retention time plot for peak identification of CO2 hydrogenation.. Different species are labeled by numbers. ...................................... 49 Figure 4.11. (a) Peak deconvolution of the H2 and CO signals. (b) Localized peak fitting results in the range of 3.15 to 3.65 min.. ........................................................... 49 Figure 4.12. Calibration lines for signal area vs gas flow mass of (a) CO2 (b) CO and (c) CH4. The R-square values of all reach 0.99. .......................................................... 50 Figure 4.13. CO2 conversion as a function of temperature (°C) of twinned nanoparticles (black bar), concaved tetrahedra (red bar) and excavated nanoparticles (blue bar). ... 52 Figure 4.14. CO2 Conversion (blue bar), CH4 and CO formation (red and blue bar) as a function of temperature (°C) with the catalysts of (a) concaved tetrahedra, (b) excavated nanocrystals, and (c) twinned nanocrystals. ............................................... 54. XII.

(14) LIST OF TALBES Table 2.1. Comparison of homogeneous and heterogeneous catalysts.. ....................... 6 Table 4.1. Standard Formation Gibbs Free Energy (ΔG) of the selected chemicals. . 40 Table 4.2. CO2 Conversion value for each temperature. ............................................. 51 Table 4.3. CO2 conversion, CH4 and CO selectivity value for each temperature. ...... 53. XIII.

(15) CHAPTER 1 OVERVIEW 1.1 Research Background The global warming is the most urgent issue requiring solution in the coming century. Behind this issue, the greenhouse gases (GHC) are the major factor in which CO2 occupies the highest ratio apart from the water vapor. So far, numerous strategies have been asserted to ease the CO2 pollution including CO2 conversion, capture and storage. Whatever the strategy, the ultimate target is to convert CO2 to value-added chemicals. Nevertheless, activation of CO2 to trigger a synthesis reaction requires harsh conditions, i.e. high temperature and pressure, because it is a very stable molecule. High temperature and pressure are always the killers to catalysts. Accordingly, how to prepare the catalysts sustainable in activity and thermostability interests chemists very much. In spite of the high efficiency of homogeneous catalysts, heterogeneous catalysts are still favorable in the industry due to their advantages of easy separation and collection during mass production. However, unlike the homogeneous catalysts with tunable activity and selectivity of the active centers by changed the coordinated ligands, the active sites over the surfaces of the heterogeneous catalysts have to be modulated with distinct strategies. In typical, three ways such as atomic modulation of surface structures, compositions, and strain were widely taken for years. In these three ways, the d-band structures of the catalyst surfaces are altered by which the coordination between reactant molecules and catalyst surfaces could be probably benefited.. 1.

(16) Wrapping catalysts with porous materials such as silica, MOFs, and zeolites is also a popular strategy which could not only raise the catalyst thermostability but also provide them additional selectivity to reactant molecules.. 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. 2.

(17) 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. 3.

(18) 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. . 4.

(19) 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. 5.

(20) 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 hydrogenation6,7 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 6.

(21) 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.. 7.

(22) 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).. 8.

(23) 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.. 9.

(24) 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 including {100} facet dominant. methods,. 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.. 10.

(25) 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.. 11.

(26) 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 12.

(27) 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. 13.

(28) 2.2.3 Sites Effect: CO2 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 14.

(29) 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/TiO2 catalysts. The spectra are displayed in Kubelka−Munk (KM) units and normalized by the symmetric gemdicarbonyl peak (2097 cm−1) height to allow for comparison. (c) Site fraction (%) of isolated (Rhiso) and nanoparticle-based Rh sites (RhNP), calculated based on eq. 2.2 and the spectra in (b), as a function of w.t.% Rh.. 15.

(30) 𝐶𝑂. 𝑋iso =. 𝐼iso/(𝜀iso×(𝑅ℎ)iso). , 𝐶𝑂 3 ∑𝑖=1[𝐼i/(𝜀i×( )i)] 𝑅ℎ. XNP = 1 − Xiso 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.. 16.

(31) 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 17.

(32) 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 become seeds with a singlecrystal, single-twinned, or multiple-twinned structure. If stacking faults are involved, the seeds will grow into plate-like nanostructures. The green, orange, and purple. 18.

(33) 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.. 19.

(34) 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.. 20.

(35) 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. 21.

(36) [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 is higher in the condition without bromide ions than that with 5 mM bromide ions.. 22.

(37) Figure 2.18. (a,b) SEM images of reaction products from growth solutions containing 50 mM CTA-Cl and (a) 0.0 and (b) 5.0 mM NaBr, resulting in the formation of highindex faceted trisoctahedra and {100}-faceted cubes, respectively. Scale bars are 200 nm. (c) ICP- AES kinetics data of the reactions containing 0.0 mM (black squares) and 5.0 mM (red triangles) NaBr.. A similar result happened in the case of iodide-added CTA-Cl reaction solution. In Figure 2.19 a, b and c, the results are obtained from the conditions containing 0, 10, 75 µM NaI. As shown trisocthedra, octahedra and [111]-truncated bitetrahedra, respectively. The corresponding kinetic measurements of the formation of trisocthedra and bitetrahedra with ICP-AES are shown in Figure 2.19d. It turns out that the case of iodide ions behave as bromide ions which slow the formation rate of Au0 atoms.. 23.

(38) Figure 2.19. (a−c) SEM images of reaction products from growth solutions containing 50 mM CTA-Cl and (a) 0.0, (b) 10.0, and (c) 75.0 μM NaI, resulting in the formation of high-index faceted trisoctahedra, {111}-faceted octahedra, and {111}-faceted truncated bitetrahedra, respectively. Scale bars: 200 nm. The truncated bitetrahedra are the planar twinned analogue of the single crystalline octahedra. (d) ICP-AES kinetics data of the reactions containing 0.0 μM (black squares) and 75.0 μM (red triangles) NaI.. 2.3.4 Platinum Group Metals ─ Rhodium (Rh) Typical shaped nanostructures of f.c.c. metals such as nanocubes, octahedrons, nanoprisms have been discussed in previous section. Herein, some unique nanostructures with modulated surfaces will be discussed with Rhodium as the example. Shuifen Xie et al. have synthesized the concaved Rh tetrahedra, as shown in Figure 2.22, by tuning the reduction kinetics.39 According to the results of TEM tilting (Figure 2.22b) and high-resolution imaging (Figure 2.22c,d), the 111 facets are indeed concaved arisen form either passivated growth or etching. To investigate the mechanism, ex-situ time-dependent study during reduction of Rh precursors was. 24.

(39) carried out (Figure 2.23). As a result, newly formed rhodium particles were mixtures in various shapes at a short time stage (Figure 2.23a). Next, a preferable growth of Rh nanoparticles at the corners of small tetrahedra were observed at the reaction time of 15 minutes, resulting in the multipod nanostructures (Fig. 2.23b). When the reaction came to the 30th minute, completed concaved tetrahedra began to show up (Figure 2.23c). Higher yield would be achieved with the reaction time extended to an hour (Figure 2.23d). Notably, the role of citric acid is actually a kind of capping agents favoring in 111 facet capping.. Due to the reason, plenty of Rh atoms reduced in the. initial nucleation stage could not subsequently deposit on the pre-formed Rh small particles but preferentially at the corners of tetrahedral seeds. The speculated mechanism is schematically interpreted in Figure 2.23e. Apart from citric acid, solvent is also a factor having significant influence on the reduction kinetics. The reducing ability of polyol solvents comes from the hydroxyl groups which decreases with the elongation of hydrocarbon chain. Accordingly, the reduction rates of Rh precursors with different polyol solvents change in the order: ethylene glycol (EG) > diethylene glycol (DiEG) > triethylene glycol (TriEG) > tetraethylene glycol (TetraEG). Which can refer to the resulting particle size. The biggest particle were obtained if TetraEG was used as the solvent (Figure 2.24c).. 25.

(40) Figure 2.22. Structural characterizations of Rh concave tetrahedrons synthesized using the standard procedure (TriEG and 145 °C): (a) low- magnification TEM image, (b) tilted TEM images of two Rh concave tetrahedrons recorded at different tilting angles, (c, d) high-resolution HAADF STEM images of a concave tetrahedron recorded along [111] and [211] zone axes, respectively, together with the corresponding atomic models in the insets. The scale bars in b are 10 nm.. 26.

(41) Figure 2.23. (a−d) TEM images of Rh nanocrystals obtained using the standard procedure at different reaction times after the injection of Rh(III) precursor: (a) 6, (b) 15, (c) 30, and (d) 60 min. The scale bars in all the insets represent 5 nm. (e) A schematic illustration showing the major steps involved in the formation of Rh concave tetrahedrons: (1) corner-selected deposition, (2) diffusion from corners to edges at a rate of V2, and (3) diffusion from corners/edges to side faces at a rate of V3.. 27.

(42) Figure 2.24. (a−c) TEM images of Rh nanocrystals synthesized using the standard procedure except for the use of different polyols as the solvents: (a) EG, (b) DiEG, and (c) TetraEG. (d) A plot of the average edge lengths of the Rh nanocrystals obtained from different polyols.. 28.

(43) CHAPTER 3. EXPERIMENTAL SECTION. 3.1 Chemicals 3.1.1 Synthesis of Rhodium Nanoparticles Rhodium(III) bromide hydrate (RhBr3·xH2O, Alfa Aesar), formic acid (HCOOH, 98%, Sigma-Aldrich), hexadecyltrimethylammonium chloride (CH3(CH2)15N(Cl) (CH3)3 CTAC, 95%, TCI), hexadecyltrimethylammonium bromide (CH3(CH2)15N(Br) (CH3)3 CTAB, 95%, TCI), hydroxylamine hydrochloride (NH2OH·HCl, 99%, SigmaAldrich), formalin (HCHO, 24%, 63 pure Chemicals), methanol (CH3OH, 99.9%, Fluka), zirconium oxide (ZrO2, <100 nm, Aldrich), carbon dioxide gas (CO2, 99.9995%, Fung Ming Industrial Co.), hydrogen gas (H2, 99.9995%, Shen-Yi Co.), helium gas (He, 99.9995%, C. C. Gaseous CO.) were used without further purification. Deionized water (18.2 MΩ·cm, Sartorius arium pro) was used as solvent in all experiment.. 3.2 Instrumentation Temperature-controllable hotplate (Fisher Scientific), borosilicate glass vial (22 mL) with plastic caps, centrifuge (Eppendorf, centrifuge 5804), micro-centrifuge (Thermo Scientific, Heraeus Pico 17), micropipettes (Eppendorf, range: 1-10 mL, 200-1000 µL, 2-200 µL) with appropriate disposable tips, poly(propylene) centrifuge tubes with capacity of 50 mL (Corning CentriStar), poly(propylene) micro-centrifuge tubes with capacity of 1.5 mL (Fisher brand), and ultrasonicator (Elma E60H Elmaonic). 29.

(44) Mass flow controller (range: 20 sccm, 200 sccm, 100 sccm, ALICAT), gas chromatography with TCD detector (7890B and G3432, AGILENT) and six-port rotary valve, GC separation column (30 m x 0.32 mm, GS-GASPRO, AGILENT) costumed reaction oven (Carbolite, R.T. to 1150 °C) with temperature controller (Kindleuro)... 3.3 Procedure 3.3.1 Synthesis of Excavated Nanoparticles, Concaved Tetrahedron, and Twinned Nanoparticles An oil bath was preheated to 90 °C and set under stirring at a rate of 350 rpm for a while to ensure the ultimate temperature is stable. On the other hand, the precursor solution was prepared mixing 10 ml of de-ionized water, CTAB, 0.02 M RhBr3 600µL, and formic acid in a glass vial. CTAB was used as the ionic capping agent and the formic acid served as the reducing agent. However, the CTAB concentration varied from 0.005 to 0.075 M, and the formic acid concentration was adjusted from 0.05 M to 0.4 M, the best amount for synthesis excavated, tetrahedral, and twinned nanoparticles is 0.005M, 0.01M, and 0.075 CTAB with 0.2M formic acid. The vial was then heated in oil bath without stirring for 18 hours, followed by centrifuging at 11,000 rpm for 20 minutes, rhodium nanoparticles could be obtained.. 3.3.2 Preparation of Rhodium Catalyst First, appropriate amount of commercial ZrO2 powder was dispersed in 2-3 mL of deionized water and trace mount of CTAB in a glass vial, followed by the addition of Rh nanoparticle solution under stirring at 700 rpm to make 1 w.t.% heterogeneous. 30.

(45) catalysts. The stirring was run for 3 hours and centrifuged at 5,000 rpm for 10 minutes. The solution typically exhibit two separate layers, including the upper supernatant and the precipitation of catalysts. After carefully removing the upper supernatant, the precipitated catalysts were washed with DI water and collected again by centrifuging. At last, the catalyst powder was placed in an oven for drying at 40 °C. Before loading into the U-shaped stainless steel reaction tube (1/8 inch), the dried catalysts were diluted with the sea sand for more efficient reaction.. 3.3.3 Catalytic Performance Experiment for CO2 Hydrogenation The catalytic experiment has been done in a continuous flow reactor. The catalyst was placed in a 1/8 stainless steel reaction tube. Reduction of Catalysts The catalysts were reduced in 10% H2/He mixing gas at a constant flow of 100 sccm under atmospheric pressure. The reduction temperature was raised from room temperature up to 400 °C, held for an hour. When the reduction procedure was done, the reactor was cooled with a continuous pure He flow of 100 sccm for 30 minutes.. Catalytic Activity Measurement The standard catalytic performance was tested by using a U-shaped, 1/8 inch stainless steel reaction tube at atmospheric pressure. 100 mg of catalysts was loaded in a U-shaped reactor of the stainless tube followed by sealing both open ends with quartz wool. All gas flows were controlled by digital mass flow controllers (ALICAT). After the reduction of catalyst, the reaction gas (CO2 and H2, ¼ , v/v) was mixed with He (37.5%, v/v), 40 standard cubic centimeters per minute (sccm) total flow rate of gas 31.

(46) was introduce to reactor tube. Measurements were carried out at various temperatures from 300 to 500 °C, and 1 h for each temperature. The product gas stream was quantified using a gas chromatograph (GC) (Angilet, 7890B) with TCD detector.. 3.4 Characterization 3.4.1 Scanning Electron Microscopy (SEM) The low-mag images were analyzed using a field-emission scanning electron microscope (FESEM, ZESS ULTRA plus). All SEM samples were prepared by dropping the NP solutions on silicon wafers in the planar size of 0.2 x 0.2 cm2.. 3.4.2 Transmission Electron Microscope (TEM) The morphologies of rhodium NPs and their corresponding high-resolution lattice images were characterized by a field-emission transmission electron microscopy (JEOL-JEM 2100F) operated at 200 KeV. All samples were prepared by dropping NP solutions on the carbon-coated copper grids (Formvar).. 3.4.4 X-ray Photoelectron Spectroscopy (XPS) The surface chemical state of rhodium NPs was measured by PHI Quantera SXM (Ulvac-Phi, Inc.) using single optical scanning device (scanning monochromated) Al anode as X-ray source. All samples were prepared following the same way to that of making SEM samples.. 32.

(47) 3.4.5 Inductively Coupled Plasma with Atomic Emission Spectroscopy (ICP-AES) The accurate content of Rhodium atoms in the nanoparticle solutions were always verified by the ICP-AES (Agilent, Varian 720-ES). For the measurement, samples have to be prepared by dissolving trace amount of Rh nanoparticle solution in aqua regia along with dilution.. 3.4.6 GC-TCD Gas Chromatography (GC) A normal gas chromatography, includes at least four parts requiring constant optimization, which are samples, carrier gas (mobile phase), column (stationary phase) and detector. After injection and gasification of samples, ample elution of the tubing and loop of the catalytic system has to be done with inert carrier gas, such as N2, Ar, He inactive to the analytes, for bringing the gas samples through the capillary column. The capillary column works based on the different strength of the interaction between sample molecules and inside stuffing, which results in the different retention times of the analytical signals. According to the retention time, the content of samples can be confirmed easily.. Capillary Column According to the coating manner on the inner walls, three kinds of capillary columns are commonly used, PLOT, WCOT and SCOT (Figure 3.1). PLOT column packs a layer of porous solid support on the inner wall. Although it gives good performance in sieving gas molecules, plenty amount of polar molecules will retain inside the solid support and causes the issues in experiments. In contrast, WCOT is coated with a thin. 33.

(48) layer of liquid stationary phase on the inner wall, giving a better separation and resolution. In 1979, a new kind of WCOT, support-coated open tubular column (SCOT) was released. SCOT is the one merging the advantages of POLT and WCOT. It has not only the special treated silica porous solid support but also the polyimide coated on the outer wall to enhance the physical strength, resulting in excellent resolution.. (a). (b). Capillary Column Liquid Stationary Phase Porous Solid Support Porous Solid Support Coated with Liquid Stationary Phase. (c). Figure 3.1 Schematic illustration of (a) wall-coated open tubular column (WCOT), (b) porous-layer open tubular column (PLOT), (c) support-coated open tubular column (SCOT). Detector – Thermal Conductivity Detector (TCD) The working principle of TCD is related to the composition of sampling gases. When reference flow, typically inert carrier gas like N2 or He, goes through the measurement channel and meanwhile alters the resistance (R4 in Figure 3.2). A difference in the resistance is kept constantly between R2 and R4 as the reference to that between R3 and R1 which is mainly induced by the analyte gases. In this principle, all samples can exhibit distinct thermal conductivity to that of the reference flow (N2 or He) and therefore become distinguishable in TCD detection which is non-destructive.. 34.

(49) Figure 3.2 Schematic interpretation of the TCD working principle.. Heterogeneous Catalysis System The solid-gas heterogeneous catalysis system are divided into three parts, that is (a) sampling section where four mass flow controllers, a mixing chamber, and a checking valve is installed; (b) reaction section where a U-shaped stainless tube, a heating furnace, and a pressure controller are set; (c) analysis section where a 6-port loop injection valve, the GC-TCD system, and a PC are connected up for ultimate signal collection and data processing.. (a). (b). (c). Figure 3.2. The scheme of the heterogeneous catalysis system. (a) sampling system, (b) reaction section, (c) analysis section.. 35.

(50) CHAPTER 4 RESULTS AND DISCUSSION. 4.1 Shape-Controlled Rhodium (Rh) Nanoparticles As mentioned in the experimental section, the synthetic procedure of Rh nanocatalysts is a one-pot method in which RhBr3, cetyltrimethylammonium bromide (CTAB) and formic acid are mixed in a. glass sample vial sealed with a cap followed. by heating at 90 °C in an oil bath for 18 hours. Formic acid is generally oxidized as CO2 to release electrons or decomposes into CO and H2, playing the dual roles of reducing and shaping agents. In Figure 4.1, Rh tetrahedral (TD) with all surfaces concaved were obtained with moderate amounts of formic acid and CTAB. During the process of nanoparticle formation, partial formic acid would decompose into carbon monoxide and water (HCOOH  CO + H2O) when heating in the oil bath. CO molecules have been known a very strong shaping agent inhibiting the growth of 111 crystal faces of f.c.c. metals. 40,41. In addition, a tetrahedron is known a structure with four 111 facets and therefore the concaved tetrahedra are possibly the major products as a significant amount of CO molecules exist. Figure 4.2 is the TEM image of a single concaved Rh tetrahedron and its corresponding selected-area electron diffraction (SAED) pattern. Apparently, the concaved TD is a symmetric single crystalline. Figure 4.3 shows the 3d XPS spectrum and the fitting curves of the concaved Rh TDs. The peaks at 306 and 311 eV denote the 3d5/2 and 3d3/2 of Rh. After fitting, the concaved TDs are known to have the Rh2O3. 36.

(51) composition, possibly due to oxide layers on the surfaces. It is a general phenomenon observed in the products made with aqueous synthesis. Besides, the peaks of RhBrx are also found because of the unreacted precursors. 42. Figure 4.1. (a) Low-, (b) middle- and (c) high-magnification TEM images of Rh tetrahedral nanoparticles.. 37.

(52) (a). (b) 022 220 202. [111]. 50 nm 45. (c). 17.1±1.3 nm. 40. Percent Frequency (%). 35 30 25 20 15 10 5 0 10. 12. 14. 16. 18. 20. 22. 24. Diameter (nm). Figure 4.2. (a) TEM image and (b) the corresponding SAED pattern of a concaved Rh. (c) Size-distribution histogram concaved tetrahedral with 100 particles counted.. Figure 4.3. 3d XPS fitting result before loading on oxide support.. 38.

(53) 4.1.1 The Role of Reducing Agent Reducing Agent The control experiments that using methanol and formaldehyde instead of formic acid were done to understand the influence of reducing agents. In Figure 4.6a and b, the TEM images show that the Rh nanoparticles obtained with methanol are very tiny, around 2 nm with ill-defined shapes. In Figure 4.6c-e, the Rh products reduced by formaldehyde are the mixtures of cubic and irregular nanoparticles in which the nanocubes are dominant in yield. The difference in the results of Rh nanoparticles implies a critical factor controlling the morphology, which is the reducing ability of reducing agents. Table 4.1 collects the standard Gibbs free energy of formation of selected chemicals. 43-45. Based on the information, the free energy of the oxidation of methanol,. formaldehyde and formic acid could be calculated. According to eq. 1 to 4, the formation of CO2 from the oxidation of methanol should be the most favored pathway during synthesis (∆G = −738.39 kJ/mol) while that of CO is the least (∆G = −6.94 kJ/mol). It is similar in the case of formaldehyde as shown in eq. 5 to 7. The free energy of formaldehyde oxidation to become CO2 is also the most favored (∆G = −541.89 kJ/mol). However, the free energy of formic acid oxidation to become CO2 (∆G = −12.95 kJ/mol) is harder than that of becoming CO (∆G = −33.01 kJ/mol). It means that a significant amount of CO would be generated during synthesis. As mentioned, CO is known a shaping agent which inhibits the growth of 111 faces of f.c.c. metal. When methanol and formaldehyde are used, least CO gas is released to the reaction system and therefore the particle morphologies are not 111 face dominant. The truth is confirmed by Figure 4.6 in which the particles obtained with methanol and formaldehyde are irregular and cubic, respectively. The methanol has strong reducing 39.

(54) ability; thus the rhodium precursor is quickly reduced to form tiny particles. The formaldehyde has relatively weak reducing ability but still release least CO during synthesis. Without the interference of CO chemical adsorption, the existing Br─ ions dominate in the particle shaping which leads to 100 faceted nanocrystals.46-49. Table 4.1. Standard Formation Gibbs Free Energy (ΔG) of the selected chemicals.. Standard Gibbs Free Energy (ΔG, kJ/mol) of Formation H2. O2. 0. 0. H2O. CO. -237.13 -137.17. CH3OH + 130.23. CO2. HCOH. -394.36. -89.6. 1 O  HCOH 2 2 -89.6. HCOOH CH3OH -361.35. -130.23. + H2O ΔG = -196.5 (kJ/mol) -237.13. Eq.(4.1). CH3OH + O2  HCOOH + H2O ΔG = -468.25 (kJ/mol) Eq.(4.2) 130.23. . CH3OH 130.23. CH3OH + 130.23. -237.13. -361.35. CO -137.17. + 2H2. 3 O  CO2 + 2H2O 2 2 -394.36 -474.26. HCOH + 1 O2  HCOOH 89.6. 2. -361.35.  CO2 +. H2O. 89.6. -394.36. -237.13. HCOH.  CO +. 89.6. -137.17. HCOH + O2. H2. 40. ΔG = -6.94 (kJ/mol). Eq.(4.3). ΔG = -738.39 (kJ/mol) Eq.(4.4). ΔG = -271.35 (kJ/mol). Eq.(4.5). ΔG = -541.89 (kJ/mol). Eq.(4.6). ΔG = -47.57 (kJ/mol). Eq.(4.7).

(55) (a). (b). 50 nm. 20 nm. (c). (d). (e). 1.9 Å. 1.87 Å 5 nm. 10 nm. 50 nm. Figure 4.4. (a) low-magnification and (b) middle-magnification TEM image using methanol as reducing agent. (c) low-magnification (d) middle-magnification and (e) HR-TEM image of a cubic Rh nanoparticle prepared with formaldehyde to replace formic acid. The lattice parameter (1.9 Å and 1.87 Å ) is referred to rhodium 200 faces.. HCOOH  361.35. HCOOH  361.35. CO. +. -137.17. CO2. -394.36. H2O. ΔG = -12.95 (kJ/mol). Eq.(4.8). H2. ΔG = -33.01 (kJ/mol). Eq.(4.9). -237.13. +. 41.

(56) Formic Acid Concentration In previous section, we realized formic acid is the major factor to obtain concaved tetrahedra due to the release of CO. Accordingly, the concentration of formic acid is worth discussion. In Figure 4.7, very small tetrahedra sized 12.43 nm were obtained when the concentration is very low (0.005 M). Nevertheless, the reduction took relatively long time for completing synthesis (2 days). When the concentration is raised to 0.1 M, the tetrahedral skeletons with incomplete 111 faces are obtained. It clearly tells that the growth takes place on the edges and vertices of a small tetrahedral seed as the priority followed by the growth of the 111 crystal faces with increasing concentration of formic acid. The results of selective growth can be observed again when the concentration goes up to 0.4 M, in which the hierarchical nanostructures are obtained.39.. Figure 4.7. TEM images of Rh nanocrystals prepared with (a) 0.05 M, (b) 0.1 M, (c) 0.2 M, (d) 0.3 M, (e) 0.4 M concentrations of formic acid. The small picture beside (e) is the scheme for hierarchical structure. 42.

(57) 4.1.2 The Influence of Surfactant CTAB Concentration In the synthesis, both of the from CTAB and CO released from formic acid have shaping ability on the particle morphology. Since the influence of CO has been discussed, that of Br─ ions requires further investigation as well. For this purpose, the added CTAB concentration was varied from 0.005 M, 0.01 M, 0.025 M, 0.05 M to 0.075 M and their corresponding TEM results are shown in Figure 4.8. According to the TEM images, two phenomena, the shrinking size and the structure evolution of Rh nanoparticles, were observed. It has been known the variation of particle size mainly depended on the amount of CTA+ cations.. 50. In contrast, the structure evolution was. closely related to the concentration of Br─ ions. Figure 4.9 shows the zoom-in TEM images of those in Figure 4.8. In the extremely low CTAB concentration, the Rh nanoparticles formed as the excavated Rh nanostructures with mixed decahedra and tetrahedra. The result possibly arose from either [CTA+] or [Br ─]. To verify which factor dominated, the CTAC (0.01 M) was used as the capping agent instead of CTAB. As shown in Figure 4.10, similar excavated nanostructures like those in Figure 4.8a and 4.9a were obtained. It is apparently different to the condition of 0.01 M CTAB in which the typical concaved Rh tetrahedral were got. This denotes that Br─ ions have stronger binding energy than Cl─ ions and it in fact has been validated and discussed in previous literatures.38 Br─ ions were also regarded the species inducing the formation of twin structures. 51,52. The fact can be easily confirmed by Figure 4.8b-e and 4.9b-e. The Rh. nanostructures evolved from concaved TDs to twin particles through the mixtures of concaved TDs, bipyramids and twin particles with increasing [CTAB]. It turns out that the influence of Br─ ions gradually got significant in the competition with CO 43.

(58) molecules when the [CTAB] is over 0.01 M.. Overall, the two parts of CTAB, CTA +. and Br─, have different influences on the particle growth. The CTA+ ions mostly provide size control. However, the Br─ ions competing with the CO from formic acid would induce twin structures formation.. (a). (b). (c). 50 nm. 50 nm. 50 nm. (d). (e). 50 nm. 50 nm. Figure 4.8. TEM images of Rh nanoparticles prepared with (a) 0.005 M, (b) 0.01 M, (c) 0.025 M, (d) 0.05 M (e) 0.075 M CTAB.. 44.

(59) (a). (b). (c). 20 nm. 20 nm. 20 nm. (d). (e). 20 nm. 20 nm. Figure 4.9. Magnified TEM images of Rh nanoparticles prepared with (a) 0.005M, (b) 0.01M, (c) 0.025M, (d) 0.05M. (e) 0.075M CTAB.. (a). (b). 50 nm. (c). 200 nm. 20 nm. Figure 4.10. (a) low-, (b) middle- and (c) high-magnification TEM images of Rh nanoparticles prepared with 0.01M CTAC instead of CTAB. 45.

(60) 4.1.3 Temperature effect Ultimately, temperature is the last factor requires discussion. Heating a reaction means applying energy to trigger nucleation accompanied by growth. Thus, different temperatures usually lead to different results attributed to varied reduction rates of precursors. Figure 4.12 shows the TEM results of Rh nanocrystals prepared at different temperatures. In the condition of 75˚C, the products were irregular agglomerates formed in a very dilute concentration (Figure 4.12a). It tells that higher energy should be input. When the temperature was risen up to 80˚C, it was surprising that the octahedral nanoparticles formed instead of tetrahedra (Figure 4.12b). In contrast to the tetrahedral shape, an octahedron of f.c.c. metal is a more thermodynamically stable morphology which is commonly obtained. It comes from a very slow reduction rate of precursors and therefore least nucleation and slow growth. However, the concaved tetrahedral structures were always obtained when the temperature went over 80˚C (Figure 4.12c-e). They formed in similar size whatever the temperature was if being over the critical temperature. It indicates that the tetrahedra should be the kinetic products resulting from both the CO shaping and rapid growth. In conclusion, the influence of temperature is mainly on the growth kinetics. 75˚C is the lowest temperature limit to trigger nucleation and 80˚C is the critical temperature to the structure evolution.. 46.

(61) (a). (b). (c). 50 nm. 50 nm. 50 nm. (d). (e). 50 nm. 50 nm. Figure 4.12. TEM images of Rh nanoparticles synthesized at (a) 75 °C, (b) 80 °C, (c) 90 °C, (d) 100 °C and (e) 110 °C.. 4.2 Test of Catalytic Performance As mentioned in Chapter 3, the working principle of TCD is to distinguish the difference of resistance between the flows of analytes and reference. Nevertheless, TCD cannot tell the exact chemical composition of an analytes, while the number of compounds. Accordingly, peak identification must be carried out in advance for the following test on the catalytic performance of CO2 hydrogenation.. 4.2.1 Spectrum Identification and Stability Test Spectrum Identification In general, there are multiple products generated after a catalytic reaction. To accurately identify each product, two ways are typically utilized. The first is to search 47.

(62) for references from the database. However, if references are not fitting to the actual experiment conditions, building up reliable references using corresponding pure compounds become necessary. In addition, calibration lines of concentration vs signal abundance should be always carried out to quantify specific products. Regarding to CO2 hydrogenation, the products can be miscellaneous. They typically include unreacted CO2, H2, methane, methanol, formic acid, and some further products generated based on the mentioned low-carbon compounds. In our experiment, the major product obtained with Rh nanocatalysts was methane. Helium (He) was used as the carrier gas for the whole catalytic system. To get correct signals with the TCD, He was also used as the reference gas, the same to the carrier gas, so that the He gas would not contribute any signal (black line in Figure 4.13).. .. Figure 4.13 is the plot of signal abundance vs retention time of the reactants and products in the CO2 hydrogenation. The signal of H2 (red) is a set of positive and negative peaks, locating at the retention time of 3.258 and 3.345 min, because of the difference in conductivity to the carrier gas. 53 The other references such as CO2 (blue), CO (pink) and CH4 (green) show up at the retention time of 5.33, 3.383 and 3.5 min, respectively. The dark blue curve shows the result after the CO2 hydrogenation. The inset graph shows more localized range of the intensity vs retention time plot from 3.1 to 3.7 nm. The curve can be deconvoluted as 1 to 4 area and well-fitted with H2 (1), CO (2), CH4 (3) and CO2 (4). Notably, the broad peak located in-between the range of 3.30 to 3.44 min is due to the overlap of the H2 and CO signals (Figure 4.14).. 48.

(63) He H2. He H2 CO2 CO CH4. CO2. CO2+H2. CO CH4 CO2+H2. Intensity (a.u.). 2. 4. 3. 1 3.1. 3.2. 3.3. 3.4. 3.5. 3.6. 3.7. Rentention time (min). 2. 3. 1. 3.5. 4.0. 4.5. 5.0. 5.5. 6.0. Rentention time (min). Figure 4.13. Intensity vs retention time plot for peak identification of CO2 hydrogenation. Different species are labeled by numbers.. (a). (b). Intensity Intensity(a.u.) (a.u.). 2. CO2+H2. He H2. H2. CO2. CO Sum. CO CH4 CO2+H2. 1 Program. 2. 3. 1 3.30. 3.32. 3.34. 3.36. 3.38. 3.40. 3.42. 3.44. 3.15. Retention time (min). 3.20. 3.25. 3.30. 3.35. 3.40. 3.45. 3.50. 3.55. 3.60. 3.65. Rentention time (min). Figure 4.14. (a) Peak deconvolution of the H2 and CO signals. (b) Localized peak fitting results in the range of 3.15 to 3.65 min.. 49.