CVD反应器中成膜与成核竞争机制的实验研究

利用管式热化学气相淀积反应器,研究了化学气相淀积过程中成核与成膜的竞争机制。结果表明,增加反应器操作压力、提高反应温度和反应物浓度,有利于粒子的生成而不利于薄膜淀积。在反应器中,淀积薄膜的表面形态各不相同。     

圆盘反应器中的PET缩聚过程

利用薄膜实验研究了PET缩聚过程反应和传质规律 ,建立了链增长、链降解反应动力学模型和泡沫脱挥时的传质速率方程 ;通过冷模实验研究了圆盘反应器中PET的混合、流动、成膜、膜表面更新以及传质规律 ;综合了反应和反应器研究结果 ,建立了圆盘反应器中的PET缩聚过程模型 ,对圆盘反应器中的PET缩聚过程进行了仿真分析 ,考察了温度、压力、停留时间、转速、催化剂浓度、负荷大小等各种因素的影响     

圆盘反应器成膜性和持液量研究

本文以工业聚酯终缩聚圆盘反应器开发应用为研究背景 ,对不同直径圆盘反应器的成膜性和持液量作了较为详细的研究。首先考察了物料粘弹性、盘径和转速对成膜性的影响 ,以对成膜性观察为基础 ,通过成膜机理分析提出了粘附膜和下垂膜的概念 ,并由此推论圆盘反应器应存在一最佳成膜转速范围 ,然后给出了持液量关联式。不但为该反应器表面更新及脱挥速率的深入研究拓宽思路和视野 ,而且为工业圆盘反应器转速确定及搅拌轴设计提供理论依据。     

圆盘反应器成膜性的持液量研究

本文以工业聚酯终缩聚圆盘反应器开发应用为研究背景，对不同直径圆盘反应器的成膜性和持液量作了较为详细的研究，首先考察了物料粘弹性、盘径和转速对成膜性的影响，以对成膜性观察为基础，通过成膜机理分析提出了粘附膜和下垂膜的概念，并由此推论圆盘反应器应存在一最佳成转速范围，然后给出了持液量关联式，不但为该反应器表面更新及脱挥速率的深入研究拓宽思路和视野，而且为工业圆盘反应器转速确定及搅拌轴设计提供理论依据。     

传统法建立动力学模型的不确定性和沿程法(RAPRA)建模

本文通过数值分析、参数灵敏性分析和实验研究,考察了动力学模型及参数、反应器操作方式和操作条件等因素对模拟结果的影响,认为传统研究方法建立动力学模型具有不确定性.根据反应器设计、模拟要求,提出了动力学建模方法——沿程法,该法在模拟反应器行为时可靠性优于传统法.     

一元或多元成膜助剂在水性涂料成膜中的挥发

综述了水性涂料成膜过程中影响溶剂挥发的因素。主要从成膜助剂的性能参数出发进行讨论。并对成膜第三阶段即扩散阶段中成膜助剂的挥发进行了综述。     

甲醇制烯烃反应动力学及反应器模型研究进展

甲醇制烯烃（MTO）工艺是现代煤化工领域的研究热点,MTO反应动力学及其反应器模型研究是高效反应器开发和工业装置操作优化的基础。本文综述了甲醇制烯烃反应动力学研究进展,详细论述了机理型动力学模型、八集总动力学模型、五集总动力学模型,指出集总动力学模型适用于描述MTO反应过程,如何考虑水、积炭等因素的影响是MTO动力学研究的难点和关键。结合现有动力学模型,评述了MTO反应过程在工业规模的固定床反应器、提升管反应器、循环流化床反应器、湍动流化床反应器中产物分布和转化率模拟情况,结果表明：循环流化床反应器和湍动流化床反应器适合MTO工业过程。最后指出,甲醇制烯烃反应动力学下一步研究方应集中于工业规模流化床反应器气固两相流动模拟,以及与动力学模型结合获得准确预测工业反应结果的MTO反应器模型。     

渗透汽化-酯化反应耦联膜过程动力学模型

建立了渗透汽化 -酯化反应耦联复合膜反应器过程动力学模型及测量复合膜渗透率的方法 .该动力学模型较系统地考虑了复合膜反应器中可能影响酯化反应化学平衡移动的各种因素 .研究结果表明 ,模型的模拟结果能很好地与实验结果相吻合 .     

含较高立构缺陷聚乳酸的非等温结晶动力学

用非等温结晶动力学研究了聚乳酸(PLA)的结晶行为,利用Hoffman-Weeks外推法以及Baur等提出的方程反推出PLA中的右旋组分摩尔分数为7.2%.较高的右旋组分摩尔分数是PLA结晶过程中成核速率低的主要因素.添加质量分数为2%滑石粉等成核剂后,PLA结晶速率没有明显提高;含10%滑石粉的PLA结晶速率略上升,...     

连续快速合成核壳型纳米复合粒子

一种基于二次旋转的高频撞击流反应器实现了连续快速制备核壳型纳米复合粒子的工艺过程。该反应器将均相成核与异相成核过程耦合在一起, 并显著强化了液液多尺度混合过程。通过制备Fe₃O₄/MnOOH纳米复合粒子, 初步探究了包覆率、主流量、支流总量和撞击点位置对包覆过程宏观与本征动力学过程的影响。发现了反应器存在的一些不足之处及改进方法。反应器经不断改进, 有望实现大规模、低成本、高质量生产各种纳米复合粒子的工艺过程。     

Chemical Vapor Deposition of CuOx Films by Cul and O2: Role of Cluster Formation on Film Morphology

CuO _x films were deposited on silica substrates by the chemical vapor deposition (CVD) method, using CuI and O₂ as source gases at low pressure in a tubular reactor. The growth mechanism to obtain a dense and uniformly distributed (in the axial direction in a tubular reactor) film was investigated. It was found that the occurrence of homogenous nucleation caused an abrupt increase of deposition rate and made the film porous. Homogeneous nucleation can be prevented by properly selecting reactant concentration, reactor temperature, and reactor diameter. Based on an aerosol diffusion theory from laminar pipe flow, a method of predicting cluster size in this CVD reaction system was proposed. The result showed that the clusters formed by homogeneous nucleation had an average size of about 1 nm in diameter.     

Modeling CVD synthesis of carbon nanotubes: Nanoparticle formation from ferrocene

Catalyst nanoparticles play an important role in the synthesis of carbon nanotubes. In this paper, we present a two-equation model that can predict the formation process of iron nanoparticles from ferrocene fed into a CVD reactor. The model, combined with an axisymmetric two-dimensional computational fluid dynamics (CFD) simulation, includes the mechanism of nucleation and surface growth of an iron particle and bi-particle collision. The model predicts that the diameter of a particle will increase with an increase in the reaction temperature or the radial distance from the center of the reactor. Iron particles may deposit on the reactor wall; our model predicts that the thickness of the layer consisting of deposited iron particles will decrease with an increase in the axial distance from the entrance. The first prediction was validated by experimental observations reported by other researchers. In addition to the CFD simulation, a dimensional analysis was conducted to find pi-numbers that govern the process of particle formation; three pi-numbers were identified. Furthermore, one-dimensional governing equations were obtained under the assumptions of constant diffusion coefficient and collision frequency function, and solutions for particle diameter were obtained in qualitative agreement with the earlier CFD simulations.     

Growth kinetics of MWCNTs synthesized by a continuous-feed CVD method

Unlike two-step chemical vapor deposition (CVD) methods using pre-deposited catalyst particles, in a continuous-feed CVD process, the liquid feed (consisting of catalytic precursor and hydrocarbon source) is continuously supplied into the reactor causing catalyst particle formation, nucleation of carbon nanotubes (CNTs) and CNT growth to occur simultaneously throughout the reaction period. In order to observe these processes, CVD experiments were conducted for different durations (30 s to 3 h) and the product multiwalled carbon nanotubes (MWCNTs) were characterized using scanning electron microscopy. It was found that the nanotubes did not grow in the vapor phase and that substrates played an important role in the growth by providing a place for them to anchor before growth took place. Based on transmission electron microscopy images, it has been suggested that MWCNTs grew by root-growth mechanism from the catalyst particles that were deposited on the substrate during the early stages. At long process times, continuously supplied feed gas produced additional catalyst particles which were deposited mostly on the growing nanotube mat. Due to weak catalyst-mat interaction, the additional nanotubes grew by tip growth. A comprehensive MWCNT growth model has been presented for the continuous-feed CVD.     

Growth of detector grade CVD diamond films and microscopic interpretation of their efficiency and charge collection distance in the normal and pumped states

Microwave CVD diamond films with very high particle detection efficiency have been obtained using a NIRIM-type reactor. The efficiency and charge collection distance (CCD) of nuclear particle detectors based on these films have been systematically studied as a function both of the methane contents in the growth gas mixture and of the film thickness. In particular, the effects of pre-irradiation with β-particles (pumping) have been thoroughly studied. The detector response to 5.5 MeV α particles was studied using a ²⁴¹Am source. Both efficiency and CCD behave in a markedly different way in the as-grown and pumped states, the effect of pumping being strongly dependent on film thickness. These behaviors are analyzed, taking into account the polycrystalline nature of CVD diamond films, in the framework of a recently proposed model [Appl. Phys. Lett. 75 (1999) 3216] discussing the role of in-grain defects and grain boundaries in determining the charge collection spectra of diamond films in the normal and pumped states. Experimental data are found to agree with what is expected from the model. The influence of pumping on the detector resolution is also discussed.     

Simulation of aerosol dynamics and transport in chemically reacting particulate matter laden flows. Part II: Application to CVD reactors

A highly computationally efficient and accurate semi-implicit numerical technique based on the concept of operator splitting (described in detail in Part I of this paper) has been used to solve the General Dynamic Equation in complex, non-isothermal reacting flows to predict aerosol dynamics and particle deposition rates. The numerically efficient algorithm has made it possible to solve the GDE in complex two-dimensional and three-dimensional simulations, hitherto not done due to the computational intensiveness. Simulations have been performed to elucidate the role of different process parameters on aerosol dynamics and particle deposition rates in idealized and commercial horizontal single wafer CVD reactors. Specifically, it has been demonstrated that the gas phase kinetics, particle formation, growth and deposition rates result in very complex aerosol size distributions in the reactor that cannot be captured with simplistic models that do not couple the GDE to the detailed flow field simulations. Guidelines for minimizing particle contamination in CVD reactors on the basis of the simulation results are presented.     

CFD modeling and optimal design of SiC deposition on the fuel combustion nozzle in a commercial CVD reactor

This work presents a new CFD model developed to investigate the SiC film growth over a fuel combustion nozzle in the commercial hot-wall CVD reactor. A detailed 3D model consisting of the compressive transport models and the reaction mechanism to account for the chemistry of the gas-phase and surface reactions are developed. The velocity, temperature, and concentration profiles inside the reactor are predicted numerically. The multispecies transport and its interplay with the gas and surface reactions are understood to operate the reactor effectively. The natural convection and hydrodynamics stability of the flow is investigated using various dimensionless groups (Re, Pr. Pe, and Gr/Re²). It has been observed that the buoyancy-driven flow is dominant at a large Reynolds number (Re) and Gr/Re² ratio. This results in flow recirculation, which ultimately deteriorates the film uniformity. A sensitivity analysis is performed to identify how critical parameters affect the deposition process. Besides, the optimisation of the CVD reactor is also served by combining the support vector machine, a versatile supervised machine learning method, and the Nelder-Mead algorithm to improve the film quality. Our results demonstrate that the present methodology effectively obtains optimal process conditions, significantly enhancing the film performance.     

Boundary layer chemical vapor synthesis of self-organized radial filled-carbon-nanotube structures

We report a new chemical vapor synthesis method that exploits random fluctuations in the viscous boundary layer between a laminar vapor flow and a surface to yield a not previously observed product: radial ferromagnetically filled-carbon-nanotube structures departing from a central particle. The filling of the nanotube capillary is continuous over a scale much greater than that which can be achieved by conventional CVD. This is a simple method which does not require ultra-fine control of process parameters or highly-engineered reactor components in which a single, self-organized, ordered product is formed in randomly fluctuating vapor in the boundary layer by vapor-, liquid-, solid-phase self-organization. These fluctuations create the thermodynamic conditions for formation of the central particle in the vapor which in turn defines the spherically symmetric diffusion gradient that initiates the radial growth. The subsequent radial growth is driven by the supply of vapor feedstock by local diffusion gradients created by endothermic graphitic carbon formation at the vapor-facing tips of the individual nanotubes and is halted by contact with the surface. The radial structures are the dominant product and the reaction conditions are self-sustaining. We argue that the method has potential for scalable production of metal-carbon nanostructures with other unusual morphologies.     

Cluster Size Determination in the Chemical Vapor Deposition of Aluminum Nitride

Aluminum nitride was prepared in a laminar-flow, tubular reactor using an atmospheric pressure chemical vapor deposition (APCVD) method at a reaction temperature ranging from 700° to 1100°C and AlCl₃ and NH₃ concentrations of 0.4 and 8 mol%, respectively. Films grew on the reactor wall and particles formed in the gas phase. The production rates of films and particles were independently determined. A comprehensive model was constructed to estimate the molecular size of the growth species in the APCVD process to simultaneously form films and particles of AIN from AlCl₃, and NH₃. Transport equations of a dominant growth species used in growing films on the reactor wall and particles in the gas phase in a laminar-flow, tubular reactor were formulated and solved. An assumption made in the model was to use the surface reaction rate constant measured for the film surface for the particle surface. Comparing the film and particle growth data measured experimentally with those obtained from model prediction allows us to conclude that the growth species are clusters ranging in size from 0.8 nm at 700°C, equivalent to 8 units of AIN, to 0.5 nm at 1100°C, equivalent to 1 unit of AIN.     

A new CVD reactor for semiconductor film deposition

The concept and design of a new chemical vapor deposition (CVD) reactor is presented for both epitaxial and nonepitaxial film deposition in semiconductor processing. The reactor is designed in such a way that a stagnant semiconductor source fluid of uniform concentration is provided for the film deposition without causing free or forced convection. The supply of the source gas for the deposition is by diffusion through a porous material such as quartz or graphite. Compared to the low pressure CVD (LPCVD) reactor with mounted wafer configuration, the new reactor should give a better film thickness uniformity and about an order of magnitude reduction in the amount of the source gas required. Further, at least for polycrystalline silicon deposition, the deposition rate can be much higher than is currently practiced with the LPCVD reactor. Design equations for the reactor are given. Details on the design for the polycrystalline silicon deposition are also given.     

Microwave analysis of PACVD diamond deposition reactor based on electromagnetic modelling

The production of high quality diamond films by microwave plasma assisted CVD, with rapid growth rates and good uniformity over large surfaces, requires perfectly optimized reactors from the microwave design point of view. Most MW plasma assisted CVD reactors used for diamond film deposition work on the resonant cavity principle. The design of such reactors relies on 3 choices: i) choice of a suitable resonant mode (i.e. with an electric field structure conducive to plasma ignition), ii) choice of a MW coupling system to excite the cavity, and iii) choice of a quartz window to delimit a reduced pressure zone inside the cavity, so as to obtain the plasma in front of the substrate [1].In this paper, we present an analysis method for MW plasma reactors relying on EM modelling, which allows for the identification of a resonant mode responsible for plasma ignition, applied to an existing reactor exhibiting plasma instabilities and requiring constant supervision.This analysis method, which can be generalised to any resonant cavity reactor, can describe the device behaviour (shape and location of the plasma, occurrence of instabilities) as a function of the various cavity geometrical configurations and to get a first estimate of the process performance.On the basis of such an analysis, it was possible to propose modifications to the reactor considered in order to improve process stability, and obtain higher growth rates. First growth tests done on diamond mono- and nano-crystalline films show excellent material quality and an increase in growth rate by more than an order of magnitude.