奈米結構碳材之製備、成長機制及其甲醇氧化反應觸媒之電觸媒效能

Abstract

本研究主要是開發奈米製程以合成具有可調控孔洞結構之碳奈米結構、探索它們對甲醇氧化反應之催化效能，以及討論碳奈米管(CNTs)之成長機制。孔洞性碳材料(CPMs)以及碳奈米管(CNTs)乃利用沸石或二氧化矽為模板，以C2H2和H2為反應氣體，分別在無或有鐵觸媒的情況下藉模板輔助化學氣相沉積(CVD)法合成之。這些二氧化矽模板材料，包括了Y型沸石、MCM-41、MCM-48、SBA-15和二氧化矽光子晶體(PC)等等，它均由傳統方法製備之。至於直接甲醇燃料電池陽極之製備，乃是先將孔洞性碳材料或碳奈米管與H2PtCl6混合，以CVD製程在H2和523 K溫度環境下以將Pt前趨物還原，使Pt分散在具有孔洞的CPMs或CNTs陽極擔體上，反應時間為30分鐘。採用下列方法來分析在每一研究階段之奈米材料之結構與性質：穿透式電子顯微技術、X-光繞射技術、循環伏安法、化學吸附、熱重、拉曼以及氮氣吸脫附等分析。從實驗結果，可得出以下結論： 關於CPM之製備，模板輔助化學氣相沉積製程可以成功製備CPM，所得之CPM孔徑介於1到400 nm 之間，本製程使用之氣態反應於1073 K下可以大幅縮短製程時間至1小時內，而須經反覆乾燥及除水步驟的傳統液態反應法，則需費時一天。本製程的另一個優點，在於可以複製各種尺度的微孔洞、中孔洞以及大孔洞碳材料，相對的微孔碳材微孔洞碳材料則難以藉由傳統方法製備，達到複製各種尺度的。在移除二氧化矽或沸石模板後之CPMs的IG/ID比値大約介於0.7-0.8之間，此比値與CPMs之孔徑無關。 關於碳奈米管之製備，中孔洞矽膜板成功的被用來製備CNTs。其製程是在CVD沉積之前將鐵觸媒顆粒填入膜板之孔洞中。藉由穿透式電子顯微技術之分析發現這裡的CNTs主要是多壁碳奈米管(MWCNTs)，且它們的尺寸可以成功的被合成和控制，直徑為3到17 nm的CNTs分別是使用孔洞直徑為3到18 nm的二氧化矽模板。在移除二氧化矽模板後的MWCNTs之IG/ID比値與商業化單壁碳奈米管、商業化XC-72之IG/ID之比值大約相同。 關於分散有Pt觸媒的CPMs所製之DMFC陽極之甲醇氧化效能而言，循環伏安曲線中之結果顯示正向電流密度峰值，If,(122到655A/g Pt)，隨著CPMs之孔徑增加而增加(從1到400 nm)，其中If是代表陽極甲醇氧化反應(MOR)活性之指標。其原因可能是在反應時甲醇之質量傳輸受到CPMs之孔洞的侷限。正向與逆向電流密度峰值之比值代表陽極的CO-容忍度之指標；相對於孔洞尺寸為1到400 nm之CPMs，其值分別為4.9到1.0。其中CO-容忍度是代表Pt顆粒表面對抗CO毒化之能力，其中佔據Pt表面的CO是MOR操作過程中的中間產物，它使得陽極失活。因此一般來講，CO-容忍度是與Pt顆粒大小無關的。然而此研究結果顯示，小於1 nm的Pt顆粒數量增加將使得CO容忍力上升，這指出小於1 nm之Pt顆粒的表面性質由於奈米效應而有明顯改變。然而其背後的原因有待進一步的研究。 關於分散有Pt處媒的CNTs所製陽極之甲醇氧化效能而言。結果顯示其陽極反應活性(354-414 A/g Pt)隨者管徑增加而增加(3到17 nm)。穿透式電子顯微技術分析亦指出具有較大之管徑使得它較不易於聚集成束，Pt顆粒被包埋在管束中的或然率就比較小，也就有較好的Pt分散。這可能是MOR活性隨者管徑增加而增加的原因。結果也指出再這些情形下CO-容忍度約介在1.2到1.3之間,且CNTs尺寸以及CNTs中之Fe觸媒對於容忍度沒有明顯的效應。此外，在這裡Fe-Pt 合金對CO-容忍度並沒有顯著的影響，這是由於碳管的隔離效應使Fe免於和Pt變成合金。此外，藉由比較CPM-與CNT-輔助陽極之效能，結果指出較大的孔徑或管徑可提升MOR活性。而CO-容忍度與碳奈米結構材料之型態無關，且大量< 1 nm 的Pt顆粒會導致較高的CO-容忍度。換句話說，在本研究中，孔徑400-nm或管徑17-nm之CPMs或CNTs輔助陽極具有可導致較好的MOR活性。此外，小尺寸Pt的數量在孔徑為1 nm的CPMs輔助陽極上之數量多於其它的碳材料輔助陽極，因此具有最佳之CO-容忍度。 關於跨越觸媒顆粒之溫度差對於碳奈米管成長機制影響之研究，碳奈米管乃藉由熱以及電漿化學輔助化學氣相沉積法沉積於不同基材，例如中孔洞SBA-15以及矽晶圓等，以C2H2及H2為反應氣體，Co為觸媒。結果顯示，ΔT其定義為觸媒顆粒頂端(接近氣氛處)以及底端(接近基材處)之局部溫度差，是決定CNT頂端或底端成長機制的重要製程參數。也顯示成長模式在ΔT > 0, ~ 0, < 0時，分別傾向於頂端-、洋蔥狀的-、底端-成長模式。利用ΔT可以成功的用來解釋為何底端-和頂端-成長模式的碳奈米管通常分別有較大的趨勢發生在熱-以及電漿輔助-化學氣相沉積製程中，還可進一步地藉由改變ΔT的設計成功的改變在熱-以及電漿輔助-化學氣相沉積製程中的成長模式。

In this work, nanofabrication processes to synthesize various carbon nanostructures with tunable pore structure, to explore their DMFC (direct methanol fuel cell) anode applications and to examine growth mechanisms of CNTs (carbon nanotubes) were developed. The CPMs (carbon porous materials) and CNTs were synthesized without and with Fe as catalyst, respectively, using template-assisted CVD process with C2H2 and H2 as the reaction gases. The templates are made of zeolite or silica, and consist of commercial Zeolite-Y, MCM-41, MCM-48, SBA-15, and PC (photonic crystal) porous silicas. The templates were prepared by conventional methods. For DMFC anode preparation, anodes were fabricated by mixing CPMs or CNTs with H2PtCl6, followed by a CVD process to reduce Pt-precursor and to disperse Pt element on porous CPMs or CNTs anode in H2 atmosphere at 523 K for 30 min. The structure and properties of the nanostructured materials after each step were characterized by transmission electron microscopy (TEM), X-ray diffratometry (XRD), cyclic-voltammetry (C-V), chemisorption, thermogravimetric (TG), Raman, and N2 adsorption/desorption analyses. From the experimental results, the following conclusions can be drawn. For CPMs fabrication, the developed silica template-assisted CVD process could be successfully used to control CPMs with pore diameters ranging from 1 nm – 400 nm, using gas reactants at 1073 K to reduce the process time to less than 1 h, in contrast to about 1 day by the conventional processes using liquid reactants and requiring multiple dehydration and drying steps. The additional advantage of the new process is that all range of microporous, mesoporous and macroporous carbons can be replicated. In contrast, microporous carbons are difficult to be fabricated by conventional methods. The IG/ID ratios of CPMs after silica or zeolite removal are about 0.7 - 0.8, independent of pore size. For CNTs fabrication, mesoporous silica templates were also successfully used to fabricate CNTs, by filling Fe catalyst particles into the pores of template before CNTs deposition. From TEM examination, it is found that the CNTs are mainly MWCNTs (multiwalled CNTs), and their sizes can be successfully synthesized and controlled, ranging from 3 nm to 17 nm in diameter by using the pore diameters of silica templates ranging from 3 nm to 18 nm, respectively. The IG/ID ratios of MWCNTs after silica removal are about the same as CPMs, commercial SWCNTs and XC-72 activated carbon material. Regarding methanol oxidation reaction (MOR) performance of DMFC anode made of CPMs with dispersed Pt, the results show that peak forward mass activity If (from 122 to 655 A/g of Pt) in the C-V curves, which is an index representing the MOR activity of the anode, increases with increasing pore size (from 1 to 400 nm in diameter) of CPMs. The reason may be due to limitation of pore of CPMs on mass transfer of methanol during reaction. From the C-V curves of the anode, the peak forward to reverse mass activity ratio, If/Ir, is an index representing CO-tolerance of the anode; the results indicate that the values ranging from 4.9 to 1.0 are corresponding to the pore sizes of CPMs ranging from 1 to 400 nm, respectively. The CO-tolerance represents the ability of surface condition of Pt-particles to stand poisoning by the occupied CO, where CO is an intermediate product of MOR during operation, making the anode inactive. Therefore CO-tolerance, in general, is not sensitive to Pt-particle size. However, the results indicate that a higher tolerance value is closely related to a greater amount of smaller Pt particles with sizes less than 1 nm in diameter, signifying surface conditions may change significantly for particle size less than 1 nm due to nano effect. Further study is required to find the reasons behind that. Regarding MOR performance of DMFC anode made of CNTs with dispersed Pt on outside of the tubes, the results show that the MOR activity (from 354 to 414 A/g of Pt) of the anode increases with increasing tube diameter (from 3 to 17 nm). The TEM examination also indicates that a larger tube size give rise to a less tube bundling, less probability for Pt particles to be embedded within the bundles and so more uniform Pt particle distribution. This may be the reasons for a greater MOR activity for a greater tube size. The results also indicate that the CO-tolerances in these cases are around 1.2 – 1.3, where effects of tube size of CNTs and amount of Fe catalyst in CNTs on tolerance are not significant. It signifies there are no significant differences in surface conditions for Pt-particles on different tube sizes. In addition, there is no significant amount of Pt-particles with sizes less than 1 nm, as discussed in the previous paragraph. Furthermore, the reported effect of Fe-Pt alloys on CO-tolerance is not obvious in theses cases due to isolation effect of the tubes to the Fe-particles from alloying with Pt. Furthermore, by comparing the performance between the CPMs- and CNT-assisted anodes, the results indicate that a greater pore size or tube diameter gives rise to a greater MOR activity, and a greater amount of Pt particles of < 1 nm results in a higher CO-tolerance, independent of the type of carbon nanostructured materials. In other words, the CPMs- or CNT-assisted anodes with 400-nm pore size or 17-nm tube diameter can result in a better MOR activity amoung the present working cases. In addition, the amount of smaller Pt particles (< 1 nm) in the CPM-assisted anode with 1-nm pore size is much greater than other carbon-assisted anode cases, therefore it results in the best CO-tolerance. About effect of temperature difference across a catalyst particle on growth mechanism of CNTs, growth of CNTs by thermal and plasma-enhanced CVD on various substrates, such as mesoporous SBA-15 and Si wafer, with C2H2 and H2 as reaction gases and Co as catalyst were conducted. The results show thatΔT, defined as local temperature difference between the top (close to gas atmosphere) and bottom (in contact with the substrate) sides of a catalyst particle, is an important parameter to determine whether growth mode being tip- or base-growth mechanisms. When ΔT > 0 , ~ 0 and < 0, the results show that growth mode of CNTs is more favor to be tip-, onion-like- and base-growth modes, respectively. TheΔT parameter proposed to determine growth modes can be successfully adopted to explain why the base- and tip-growth CNTs are common in thermal CVD and plasma-enhanced CVD processes, respectively. Furthermore, few experiments designed to changeΔT to vary the growth modes in both thermal and plasma-enhanced CVD processes were successfully conducted.