用ECR-CVD法合成小管徑奈米碳管

Abstract

本研究利用電子迴旋共振化學氣相沉積法(ECR-CVD)，以甲烷(CH4)與氫氣(H2)為反應氣體，CoCrPtOx為觸媒於矽晶片上成功的開發小管徑(平均直徑小於8 nm)之奈米碳管。首先，以CoCrPt為靶材，在Ar/O2的氣氛下，利用物理氣相沉積法(PVD)於矽晶片上鍍一層金屬觸媒薄膜，再者，有些試片在鍍上CoCrPtOx前，先在矽晶片上鍍上Al2O3薄膜作為緩衝層，接著把試片至於ECR-CVD系統中進行氫電漿前處理，使觸媒薄膜變成均勻分佈的奈米觸媒粒子最後在ECR-CVD系統中合成奈米結構。本實驗也研究在前處理與合成奈米結構的過程中於試片上覆蓋一片矽晶片對其結果有何影響。試片在每個步驟的形貌，將藉由掃描電子顯微技術(SEM)加以分析，沉積後的各種奈米結構及其性質將藉由SEM、穿透電子顯微技術(TEM)、能量散佈光譜儀(EDS)拉曼光譜技術(Raman spectroscopy)、光電子能譜儀(XPS)以及場發射J-E測量法加以分析探討。從本研究結果中可獲得下列結論。 當試片有覆蓋一片矽晶片時，在較高的載台偏壓(如: -100 V) ，CH4/H2比例大於0.5/50 sccm/sccm，較高的微波功率(~1000 W) ，觸媒膜厚小於2 nm 與工作壓力為 4 Torr時，合成出基底成長的主要為4壁的奈米碳管，其平均管徑為~7.9 nm，而在其中也可發現少量管徑約為3.6 nm的雙壁奈米碳管。覆蓋Si晶片在此的作用是阻止離子直接轟擊於試片表面與改變氣流成為水平流向，因此需要45分鐘的成長時間在試片的中央才可以得到均勻的奈米結構。可得到此小尺寸的觸媒粒子與小管徑的多壁奈米碳管的原因為，PtO2的爆炸式化學還原反應使觸媒前趨物薄膜爆散成奈米觸媒顆粒，與Cr2O3在進行氫電漿前處理與初期合成奈米結構時，可有效的阻止奈米觸媒顆粒的聚集。 根據改變實驗參數所得到之結果顯示，前處理後所得到的觸媒粒子大小與觸媒薄膜的厚度和前處理時的條件有關，但是當觸媒薄膜的厚度減少到2 nm或者2 nm以下時，較小的觸媒粒子大小其大小會到達一定值並變為主要的粒徑尺寸，所以當觸媒薄膜的厚度從10 nm減少到2 nm或2 nm以下時，成長的奈米碳管平均管徑會從~27.3 nm減少至一定值，此值為~7.9 nm，實驗結果也顯示當觸媒薄膜的厚度為1 nm並且有10 nm Al2O3當做緩衝層，所得到的奈米碳管的管徑大小與此定值並無明顯改變，推論原因是細小的觸媒粒子融點較大觸媒粒子為低，所以奈米碳管似乎較易在小觸媒粒子處合成。關於CH4/H2流量比例的影響，當CH4/H2流量比例增加到足夠值(如:15/50 sccm/sccm)時，奈米碳管的形貌會從義大利麵狀轉變為準直性的奈米碳管，這是因為較高的流量比例與較高的微波功率可提供較多碳源以增加奈米碳管的成核密度，於是鄰近的奈米碳管互相推擠生長，因而簇擁成準直性的奈米碳管。除此之外，在較低的壓力下雖然氣體的解離率會稍微增加，但是相對的提供奈米碳管生長的碳源也較少，因此壓力在4 x 10-3 Torr以下不易合成奈米碳管。 當試片沒有覆蓋一片矽晶片並且壓力在4 x 10-3 Torr時，基材偏壓由-50 V增加到-100 V，所得到的奈米結構會從奈米碳片(CNSs)轉變為矽奈米錐(SNCs)。SNCs的成長機制是因為基材負偏壓的大小會影響正離子轟擊試片表面的強度，而最終所得到的奈米結構是沉積速率與蝕刻速率達到平衡的結果。當在高基材偏壓時，沉積速率遠小於蝕刻速率，在觸媒粒子之間沒有被保護的矽基材會被蝕刻，然而觸媒粒子也會逐漸被轟擊變小，因此被蝕刻的矽基材與逐漸被轟擊變小的觸媒粒子就形成了SNCs。而在低基材偏壓時產生的CNSs，似乎支持文獻中的成長機制，其機制是在低基材偏壓時，觸媒粒子所產生的側向電場大於基材的垂直電場，因此碳就隨著電場的方向側向沉積。 關於成長出的小管徑奈米碳管的性質，拉曼分析結果顯示其拉曼IG/ID比分佈為0.52~0.88，而場發射分析指出其起始電壓分佈為5.2~8.1V/□m (在電流密度0.01 mA/cm2時) ，其中0.88與5.2 V/□m為準直性的奈米碳管所量測到之值。結果也顯示奈米碳管的管徑大小不是影響拉曼IG/ID比與起始電壓的主要因素。

The process to synthesize the small sized (< 8 nm in average diameter) carbon nanotubes (CNTs) on Si wafer was successfully developed by electron cyclotron resonance chemical vapor deposition (ECR-CVD) method with CH4 and H2 as source gases, CoCrPtOx as catalyst. The catalyst was first sputtered on Si wafer by physical vapor deposition (PVD) method with CoCrPt (Co 57.08 %, Cr 10.97 %, Pt 31.95 %) alloy as target under Ar/O2 (=10/30 sccm/sccm) atmosphere, and then followed by H-plasma pretreatment in ECR-CVD system to obtain the well-distributed catalyst particles. Some of the specimens were coated with Al2O3 film before catalyst coating to act as buffer layer. The CNTs were then deposited on the pretreated specimens in ECR-CVD system. Effects of covering the Si wafer on the specimen during catalyst pretreatment and CNTs deposition were examined. The morphologies of the specimen at each process step were examined by SEM. The as-deposited nanostructures and their properties were characterized by SEM, TEM, Raman spectroscopy, XPS and field emission J-E measurements. From the experimental results, the following conclusions can be drawn. For specimens covered by Si wafer, the as-deposited nanostructures are mainly base-growth CNTs and consist of mainly four-walled CNTs with an average diameter about 7.9 nm and few 3.6 nm double-walled CNTs (DWNTs) under the following conditions: higher negative substrate bias (e.g. -100 V), CH4/H2 > 0.5/50 sccm/sccm, higher microwave power (~1000 W), < 2 nm catalyst thickness and under ~ 4 Torr pressure. Effect of specimen protection by Si wafer is essentially to avoid direct ion bombardment and to guide the flow in horizontal direction. This is why 45 min deposition time is required to obtain uniform nanostructures at the center of the specimen. The main reason to achieve the small sized catalysts and CNTs is basically due to the explosive reduction reaction of PtOx and anti-agglomeration effect of the nano-sized pores on Cr2O3 surface during H-plasma pretreatment and the initial stage of nanostructure deposition. Regarding effects of process parameters, the results show that the pretreated catalyst particle size distribution is a function of catalyst thickness and the pretreatment conditions; but the smaller particle sizes reach a constant value and become the dominated sizes by decreasing the catalyst thickness to or below 2 nm,. Therefore, it is found that the average CNTs size for specimens changing catalyst thickness from 10 nm to 2 nm or below will vary from 27.3 nm to a constant value of 7.9 nm. The results also indicate that application of 10 nm Al2O3 buffer layer for 1 nm catalyst layer has no obvious change to this constant value. This is due to the fact that the small dominated particle size seems to be the preferred sites to grow the CNTs due to a lower melting temperature by comparing with the bigger sizes. As to effect of CH4/H2 flow ratios, morphologies of the as-deposited CNTs can be varied from spaghetti-like to become well-aligned by increasing the ratio to a higher enough value, e.g. 15/50 sccm/sccm. A higher ratio and microwave power are essentially to supply more carbon species and to enhance higher nucleation density for CNTs formation; therefore the crowding effect of the neighbor CNTs becomes the driving force to produce the well-aligned CNTs. In addition, a lower pressure may slightly increase the dissociation efficiency of the gases, but may cause a lower carbon species for CNTs growth. This is why the pressure below 4 x 10-3 Torr can result no significant CNTs formation. For specimens without covering with Si wafer and under 4 x 10-3 Torr pressure, the results show that morphologies of the as-deposited nanostructures vary from carbon nanosheets (CNSs) to silicon nanocones (SNCs) by increasing the substrate bias from -50 V to -100 V. Effect of the negative substrate bias is basically to enhance the positive ion bombardment. The balance between the deposition rate and the etching rate of bombardment determines the final deposited nanostructures. At higher substrate bias, the deposition rate is much lower then the etching rate. The mechanism to form SNCs at higher negative bias may result from progressive etching of Si wafer between catalyst particles due to bombardment protection of catalyst particles and from a progressive decrease in catalyst sizes due to bombardment. The mechanism to form CNSs at lower substrate bias seems to support the proposed mechanism in the literature, i.e. the lateral electric field created by the catalyst particles is stronger than the vertical electric field due to substrate bias; therefore, carbon deposition can follow the electric field to grow horizontally in sidewise fashion. Regarding properties of the small sized CNTs, the results appeal that the Raman IG/ID ratio are ranging from 0.52 ~ 0.88, and field emission turn-on voltages from 5.2 to 8.1 V/□m (for current density 0.01 mA/cm2), where 0.88 and 5.2 V/□m are corresponding to the values for the well-aligned CNTs. The results also show that the size of CNTs is not the main factor to determine Raman IG/ID ratio and field emission turn-on voltage of CNTs.