71 |
Investigation of a super performance dew point air cooler and its application in buildingsXu, Peng January 2017 (has links)
Based on extensive literature reviews, technical opportunities were identified to improve the energy efficiency of a dew point air cooler. This applied research aimed to develop a super-performance dew point air cooler to replace or partly replace the conventional energy-intensive air conditioners applicable to buildings. This research followed the methodology of combined theoretical and experimental investigation and a procedure of concept formation, validating and updating. A simulation software was developed and validated to investigate the impacts of the geometric configuration and operational conditions on the unit’s cooling performance and assist the prototype design. As a result, a novel dew point air cooler prototype, featuring innovative structure of the heat and mass exchanger, application of new materials and new processes, unique water distribution and control scheme and exclusive self-developed simulation software, was constructed and tested under controlled laboratory environment. Two patents were generated, one of which has been authorized by the China State Intellectual Property Office and the other has been filed in the Intellectual Property Office of the United Kingdom. Under standard testing conditions, i.e. dry-bulb temperature of 37.8oC and the coincident wet-bulb temperature of 21.1oC, the prototype cooler achieved a wet-bulb cooling effectiveness of 114% and dew-point cooling effectiveness of 75%, yielding a significantly high Coefficient of Performance (COP) of 52.5 at the optimal working air ratio of 0.364. The performance testing was also carried out under various simulated conditions representing the climates of hot & dry, warm & dry, moderate, warm & humid and the wet-bulb effectiveness of the prototype kept in the range 112% to 128% and dew-point effectiveness of 67%-76%, giving a COP of 37.4-52.5. Compared to the conventional vapour compression air conditioners which have a COP of around 3, the prototype cooler had 11-17 times higher COP, leading to a reduction in electrical power consumption by around 92% to 94%. A dedicated case study of the proposed dew point cooler based on conditions in Beijing, a representative city in warm and humid climate, was carried out to predict the annual operational performance, economic rewards, and associated environmental benefits. Compared to the conventional packaged air conditioners, 91.4% of annual power demand could be saved. The annual water consumption is less than 0.3 tonnes to provide the cooling of 2428.1 kWh. And the payback period of static investments would be less than 4 years to replace an equivalent packaged air conditioner. A significant leap forward has been achieved with this study and this is expected to open enormous global business in the very near future, thus bringing about great economic, environmental and sustainability benefits worldwide.
|
72 |
Active Control of Salient Flow Features in the Wake of a Ground VehicleUnknown Date (has links)
Aerodynamics of road vehicles have continued to be a topic of interest due the relationship between fuel efficiency and the environmental impact of passenger vehicles. With the streamlining of ground vehicles combined with years of geometric and shape optimization, other techniques are required to continue to improve upon fuel consumption. One such technique leverages aerodynamics to minimize drag through the implementation of flow control techniques. The current study focuses on the application of active flow control in ground vehicle applications, employing linear arrays of discrete microjets on the rear of a 25 Ahmed model. The locations of the arrays are selected to test the effectiveness of microjet control at directly manipulating the various features found in typical flow fields generated by ground vehicles. Parametric sweeps are conducted to investigate the flow response as a function of jet velocity, momentum, and vehicle scaling. The effect and effciency of the control are quantified through aerodynamic force measurements, while local modifications are investigated via particle image velocimetry and static pressure measurements on the rear surfaces of the model. Microjets proved most effective when utilized for separation control producing a maximum change to the coefficients of drag and lift of -14.0% and -42% of the baseline values, respectively. Control techniques targeting other flow structures such as the C-pillar vortices and trailing wake proved less effective, producing a maximum reduction in drag and lift of -1.2% and -7%. The change in the surface pressure distribution reveals the impact of each flow control strategy on a targeted flow structure, and highlights the complex interaction between the salient flow features found in the wake of the Ahmed model. Areas of pressure recovery on the surface of the model observed for each control technique support the observed changes to the aerodynamic forces. The time averaged, volumetric wake is also reconstructed to characterize the baseline flow field and highlight the effect of control on the three dimensional structure of the near wake region. The results show that separation control has a measurable effect on the flow field including modifications of the locations, size, magnitude, and trajectory of the various structures which comprise the near wake. The observations give insight into desirable modifications and flow topology which lead to an optimal drag configuration for a particular vehicle geometry. / A Dissertation submitted to the Department of Mechanical Engineering in partial fulfillment of the requirements for the degree of Doctor of Philosophy. / Summer Semester 2018. / July 13, 2018. / active flow control, aerodynamics, ahmed model, drag reduction / Includes bibliographical references. / Farrukh Alvi, Professor Directing Dissertation; Sungmoon Jung, University Representative; Rajan Kumar, Committee Member; Kunihiko Taira, Committee Member; Seungyong Hahn, Committee Member.
|
73 |
Understanding the mechanism & effects of stent fracture : a combined experimental & finite element analysisIzadian, Mohammad Hossein January 2018 (has links)
Atherosclerosis is a common heart disease, categorised by a build-up of fatty substances (plaque) in the inner surface of the coronary arteries and causing obstruction to the blood flow to vital organs and other parts of body. Over time, the arteries become narrowed which can lead to serious complications such as angina, heart attack, and stroke. There are several treatments to slow down the progress and reduce the risk, including medication and medical procedures. Percutaneous coronary intervention (PCI) is a non-surgical procedure which reopens blocked arteries and restores the blood flow. In some cases the PCI involves a tiny mesh tube known as a stent, which is placed in the narrowed artery to widen the lumen, support the vessel wall and prevent restenosis. Whilst this is generally successful procedure, stents might cause further problems such as stent fracture, in-stent restenosis, and stent thrombosis. Stent fracture is known to be associated with a number of factors; stent length, stent overlap, vessel tortuosity, degree of calcification of lesions, stent design, and the conditions under which the stent operates. The first part of this thesis presents a design-independent finite element analysis evaluation of the relative stresses induced in a coronary stent when placed in an angulated vessel geometry. This was achieved by idealising the stent to a thin tube, with the structural modulus of the tube representing that of a stent-like structure (this could be adapted for different types of stent structure). The artery and stent were then subjected to a displacement representing a bending movement of 20˚. Furthermore, various artery angles were modelled from 30˚ to 90˚ and each time the angle was transformed in 10˚. This series of finite element analyses computed the stress distribution associated with the displacement, hence quantifying the relationship between the vessel angle and the stress when considering the "hinge-type" movement that the vessel will undergo with each heartbeat. This constant repetitive loading constitutes the most severe mechanical loading that the stent will undergo, which far exceeds the radial expansion/contraction systolic/diastolic of the vessel or any torsional effects. It was observed that changes in stresses within the stent model are directly proportional to the vessel angulation, which stresses increased when the vessel angles became more severe. Furthermore, the bending region where was associated with the hinge-type movement experienced higher amounts of stress in the idealised stent model, and severe vessel angle caused a larger area undergo higher stress. The values increase at a greater rate once an angle of 75 degree has been exceeded, which agrees with clinical observation. Also areas of high stress corresponded to areas where fractures are seen clinically. The second part involved the mechanical testing of 9 samples of four different stent designs; Muilti-Link Vision®, PRO-Kinetic Energy, BioMatrix NeoFlexTM and Promus PREMIER. Stents deployed at nominal pressure into physiological mock artery at initial angle of 90˚, were then subjected to a 20˚ continuous repetitive hinge-type movement, at a rate of approximately 1100rpm (cycles per minute). By 300 million cycles fractures were identified in 7 stents, and are limited to only the Biomatrix design (34.67±28.78 million cycles), exhibiting between one and four strut fractures. Fractures were first seen to occur at 13.5 million cycles, where fractures were observed in 2 stents. All fractures were seen to occur at the ring linker parts of the stent and in the areas which would undergo the most severe tensile and compressive loading. This study shows that artery angulation has a significant impact on the stent stress, and more tortuous vessel increases the risk of stent fracture. Also in vitro experimental work illustrates that stent material and structure play an important role in stent flexibility.
|
74 |
Developments in the theory and applications of the variability response function conceptTeferra, Kirubel January 2011 (has links)
Uncertainty quantification in Civil Engineering applications is crucial to the decision making process in the analysis, design, and retrofitting of infrastructure. The consensus amongst researchers is that deterministic approaches to problem solving can lead to very misleading results, and the assessment of infrastructure performance needs to be addressed within a probabilistic framework. As a result, there is great demand to identify and acquire probabilistic information about uncertain system parameters which affect the performance of a structure. Unfortunately, it is difficult to obtain a full probabilistic description of uncertain system parameters, specifically their spatial correlation structures. In response to this limitation, researchers have sought a means to circumventing the need for a full probabilistic description of system uncertainties in determining structural response statistics. One approach is the Variability Response Function(VRF) concept, introduced by Shinozuka, which decomposes the variability of a response quantity into a deterministic function that is solely dependent on the deterministic components of the structure and the Spectral Density Function (SDF) of the uncertain system parameters modeled as a homogeneous random field. The deterministic function is called the VRF and is analogous to the Green's function of a differential equation. This dissertation explores the limits of the applicability of the VRF concept in Structural Mechanics problems. The VRF concept is applied to nonlinear statically determinate and indeterminate beams as well as plane stress structures where the flexibility is considered to be a random field. In the latter part of the dissertation the VRF concept is applied to the problem of stochastic characterization of homogenized effective properties through an equivalent energy based homogenization technique. The final chapter of this dissertation presents a novel methodology to rapidly generate sample microstructures for random two phase materials.
|
75 |
A novel discrete damage zone model and enhancement of the extended finite element method for fracture mechanics problemsLiu, Xia January 2012 (has links)
This research develops two novel numerical methods for applications in fracture mechanics: (I) A new crack tip enrichment function in the extended finite element method (XFEM), and (II) a discrete damage zone model for quasi-static and fatigue delamination in composites. The first method improves XFEM when applied to general nonlinear materials when crack tip analytical solutions are not available. For linear elastic materials, Branch functions are commonly used as crack tip enrichments. Typically, these are four functions derived from linear elasticity theory and added as additional degrees of freedom. However, for general inelastic material behavior, where the analytical solution and the order of singularity are unknown, Branch functions are typically not used, and only the Heaviside function is employed. This however may introduce numerical error, such as inconsistency in the position of the crack tip. Hence, a special construction of Ramp function is proposed as tip enrichment, which may alleviate some of the problems associated with the Heaviside function when applied to general nonlinear materials, especially ones with no analytical solutions available. The idea is to linearly ramp down the displacement jump on the opposite sides of the crack to the actual crack tip, which may stop the crack at any point within an element, employing only one enrichment function. Moreover, a material length scale that controls the slope of the ramping is introduced to allow for better flexibility in modeling general nonlinear materials. Numerical examples for ideal and hardening elasto-plastic and elasto-viscoplastic materials are given, and the convergence studies show that a better performance is obtained by the proposed Ramp function in comparison with the Heaviside function. Nevertheless, when analytical functions, such as the Hutchinson-Rice-Rosengren (HRR) fields, do exist (for very limited material models), they indeed perform better than the proposed Ramp function. However, they also employ more degrees of freedom per node and hence are more expensive. The second method developed in this thesis is a discrete damage zone model (DDZM) to simulate delamination in composite laminates. The method is aimed at simulating fracture initiation and propagation within the framework of the finite element method. In this approach, rather than employing specific cohesive laws, we employ damage laws to prescribe both interface spring softening and bulk material stiffness degradation to study crack propagation. For a homogeneous isotropic material the same damage law is assumed to hold in both the continuum and the interface elements. The irreversibility of damage naturally accounts for the reduction in material strength and stiffness if the material was previously loaded beyond the elastic limit. The model parameters for interface element are calculated from the principles of linear elastic fracture mechanics. The model is implemented in Abaqus and numerical results for single-mode as well as mixed-mode delamination are presented. The results are in good agreement with those obtained from the virtual crack closure technique (VCCT) and available analytical solutions, thus, illustrating the validity of this approach. The suitability of the method for studying fracture in fiber-matrix composites involving fiber debonding and matrix cracking is demonstrated. Finally, the DDZM method is extended to account for temperature dependent fatigue delamination in composites. The interface element softening is described by a combination of static and fatigue damage growth laws so as to model delamination under high-cycle fatigue. The dependence of fatigue delamination on the ambient temperature is incorporated by introducing an Arrhenius type relation into the damage evolution law. Numerical results for mode I, mode II and mixed mode delamination growth under cyclic loading are presented and the model parameters are calibrated using previously published experimental data. Then, predictions are made under varying mode mix conditions and are compared with numerical results in the literature.
|
76 |
Fabrication, Characterization and Modeling of Functionally Graded MaterialsLee, Po-Hua January 2013 (has links)
In the past few decades, a number of theoretical and experimental studies for design, fabrication and performance analysis of solar panel systems (photovoltaic/thermal systems) have been documented. The existing literature shows that the use of solar energy provides a promising solution to alleviate the shortage of natural resources and the environmental pollution associated with electricity generation. A hybrid solar panel has been invented to integrate photovoltaic (PV) cells onto a substrate through a functionally graded material (FGM) with water tubes cast inside, through which water flow serves as both a heat sink and a solar heat collector. Due to the unique and graded material properties of FGMs, this novel design not only supplies efficient thermal harvest and electrical production, but also provides benefits such as structural integrity and material efficiency.
In this work, a sedimentation method has been used to fabricate aluminum (Al) and high-density polyethylene (HDPE) FGMs. The size effect of aluminum powder on the material gradation along the depth direction is investigated. Aluminum powder or the mixture of Al and HDPE powder is thoroughly mixed and uniformly dispersed in ethanol and then subjected to sedimentation. During the sedimentation process, the concentration of Al and HDPE particles temporally and spatially changes in the depth direction due to the non-uniform motion of particles; this change further affects the effective viscosity of the suspension and thus changes the drag force of particles. A Stokes' law based model is developed to simulate the sedimentation process, demonstrate the effect of manufacturing parameters on sedimentation, and predict the graded microstructure of deposition in the depth direction.
In order to improve the modeling for sedimentation behavior of particles, the Eshelby's equivalent inclusion method (EIM) is presented to determine the interaction between particles, which is not considered in a Stokes' law based model. This method is initially applied to study the case of one drop moving in a viscous fluid; the solution recovers the closed form classic solution when the drop is spherical. Moreover, this method is general and can be applied to the cases of different drop shapes and the interaction between multiple drops. The translation velocities of the drops depend on the relative position, the center-to-center distance of drops, the viscosity and size of drops. For the case of a pair of identical spherical drops, the present method using a linear approximation of the eigenstrain rate has provided a very close solution to the classic explicit solution. If a higher order of the polynomial form of the eigenstrain rate is used, one can expect a more accurate result.
To meet the final goal of mass production of the aforementioned Al-HDPE FGM, a faster and more economical material manufacturing method is proposed through a vibration method. The particle segregation of larger aluminum particles embedded in the concentrated suspension of smaller high-density polyethylene is investigated under vibration with different frequencies and magnitudes. Altering experimental parameters including time and amplitude of vibration, the suspension exhibits different particle segregation patterns: uniform-like, graded and bi-layered. For material characterization, small cylinder films of Al-HDPE system FGM are obtained after the stages of dry, melt and solidification.
Solar panel prototypes are fabricated and tested at different water flow rates and solar irradiation intensities. The temperature distribution in the solar panel is measured and simulated to evaluate the performance of the solar panel. Finite element simulation results are very consistent with the experimental data. The understanding of heat transfer in the hybrid solar panel prototypes gained through this study will provide a foundation for future solar panel design and optimization.
|
77 |
Numerical Modeling of Shear Banding and Dynamic Fracture in MetalsMcAuliffe, Colin J. January 2014 (has links)
Understanding the failure of metals at high strain rate is of utmost importance in the design of a broad range of engineering systems. Numerical methods offer the ability to analyze such complex physics and aid the design of structural systems. The objective of this research will be to develop reliable finite element models for high strain rate failure modelling, incorporating shear bands and fracture. Shear band modelling is explored first, and the subsequent developments are extended to incorporate fracture. Mesh sensitivity, the spurious dependence of failure on the discretization, is a well known hurdle in achieving reliable numerical results for shear bands and fracture, or any other strain softening model. Mesh sensitivity is overcome by regularization, and while details of regularization techniques may differ, all are similar in that a length scale is introduced which serves as a localization limiter. This dissertation contains two main contributions, the first of which presents sev- eral developments in shear band modeling. The importance of using a monolithic nonlinear solver in combination with a PDE model accounting for thermal diffusion is demonstrated. In contrast, excluding one or both of these components leads to un- reliable numerical results. The Pian-Sumihara stress interpolants are also employed in small and finite deformation and shown to significantly improve the computational cost of shear band modelling. This is partly due to the fact that fewer unknowns than an irreducible discretization result from the same mesh, and more significantly, the fact that convergence of numerical results upon mesh refinement is improved drastically. This means coarser meshes are adequate to resolve shear bands, alleviating some of the computational cost of numerical modelling, which are notoriously significant. Since extremely large deformations are present during shear banding, a mesh to mesh transfer algorithm is presented for the Pian Sumihara element and used as part of a remeshing strategy. A practical application of the numerical formulation developed is modelling the shear band failure of a friction stir welded aluminum joint under high rate loading. The energy absorption capacity of these joints are subse- quently analyzed and found to be significantly weaker than untreated aluminum due to the nonhomogeneous material properties of the joint. The second contribution is extending the shear band model described previously to account for fracture by way of the phase field method. The phase field method is modified to account for the contribution of inelastic deformation to the creation of fracture surfaces, which results in a rate and temperature dependent theory for fracture, due to the rate and temperature dependence of plasticity. The combined fracture and shear band model is shown to be capable of representing a wider spectrum of strain rates than either the phase field model or the shear band model alone.
|
78 |
Particle Dynamics Simulation of Microstructure Evolution towards Functionally Graded Material ManufacturingYang, Lingqi January 2015 (has links)
Functionally graded materials (FGMs) have attracted significant attention in academia and industry, because of its unique material properties, such as thermal, electrical, or mechanical properties, which are characterized by the continuous spatial variation in macroscale due to the graded distribution of material phases in the microstructure. In our recent invention, a hybrid solar panel has been fabricated to integrate a FGM layer with water tubes cast inside between the photovoltaic (PV) cell and the substrate, so that the thermal harvest efficiency can be substantially increased, as well as the structural integrity and material efficiency. To scale up the technology from the laboratory to mass production, sedimentation-based methods (static sedimentation and vibration-sedimentation) have been explored as an efficient and economic manufacture method under different solid load, in which aluminum (Al) and high-density polyethylene (HDPE) powders are mixed in a liquid and a graded mixture can be obtained at the end of the sedimentation process. To optimize the manufacturing process, a better understanding of the physics lying behind the experimental observations is required. However, the conventional continuum methods may not be applicable due to the complex fluid effect and particle-particle interactions in this huge many-particle system. My Ph.D. studies aim to develop particle based approaches for fundamental understanding of particle interactions and optimization of the manufacturing process of FGMs. Due to its natural advantage to capture physics at a fine scale and capability to address complex fluid effect and boundary problems, particle based methods are particularly suitable to simulate the fabrication process with a large number of particles. Dissipative particle dynamics (DPD) was used to study the interactions between liquid particles. In this work, we proposed a reduced rough sphere model to use the DPD to understand the interactions between liquid and solid particles and thus simulate the hydrodynamic behavior of solid particle moving in the DPD liquid. And discrete element method (DEM) is used to describe the interaction between solid particles. The DPD/DEM hybrid model has successfully simulated the static sedimentation process and revealed the underlying physicist behind the sedimentation approach for the solid load less than 10 vol%. For a higher solid load, particle jamming is commonly observed if the solid-liquid particle size ratio is not large enough, which however makes the computation cost formidably expensive. Therefore, a modified DEM model is developed, in which liquid particles are disregarded but a drag force as a function of porosity is introduced to simulate the particle-fluid interaction. Therefore, the vibration-sedimentation of a high solid load suspension can be successfully simulated. The particle-based methods are general and can be applied in particle mixing as well, which also involves the microstructure evolution and has a wide application in the pharmaceutical industry, food processing, energetic materials, as well as many other industries. To accelerate the mixing rate, normally the liquid binder is used to wet the granular composite, in which the liquid content is close to zero. The cohesion effect is reported to play an important role in the mixing process. As the volume fraction of the liquid binder increases, the cohesion accelerates the mixing process and enhances the homogeneity of the mixture, however, if it continuously increases, extremely large cohesion force may prevent the mixture from being mixed and segregation occurs. To investigate the high shear mixing process, DEM model is developed to simulate the interaction between solid particles, while the liquid bridge model, which implements the capillary force in the particulate method, is used to describe the cohesion effect. Quantitative and comprehensive studies are performed to investigate the effect of the liquid binder's volume fraction on the mixing rate and homogeneity of the final mixture. The results have shown that there exists a critical volume fraction of the liquid binder to achieve a good homogeneity in the mixture when the filling height is comparable to the blades height. When the filling level is high, cohesion slows down the mixing process and decreases the mixing quality.
|
79 |
A Study of Catalytic Carbon Dioxide Methanation Leading to the Development of Dual Function Materials for Carbon Capture and UtilizationDuyar, Melis Seher January 2015 (has links)
The accumulation of CO₂ emissions in the atmosphere due to industrialization is being held responsible for climate change with increasing certainty by the scientific community. In order to prevent its further accumulation, CO2 must be captured for storage or conversion to useful products. Current materials and processes for CO₂ capture rely on the toxic and corrosive methylethanolamine (MEA) absorbents and are energy intensive due to the large amount of heat that needs to be supplied to release CO₂ from these absorbents. CO₂ storage technologies suffer from a lack of infrastructure for transporting CO₂ from many point sources to the storage sites as well as the need to monitor CO₂ against the risk of leakage in most cases. Conversion of CO₂ to useful products can offer a way of recycling carbon within the industries that produce it, thus creating processes approaching carbon neutrality. This is particularly useful for mitigation of emissions if CO₂ is converted to fuels, which are the major sources of emissions through combustion. This thesis aims to address the issues related to carbon capture and storage (CCS) by coupling a CO₂ conversion process with a CO₂ capture process to design a system that has a more favorable energy balance than existing technologies.
This thesis presents a feasibility study of dual function materials (DFM), which capture CO₂ from an emission source and at the same temperature (320°C) in the same reactor convert it to synthetic natural gas (SNG), requiring no additional heat input. The conversion of CO₂ to SNG is accomplished by supplying hydrogen, which in a real application will be supplied from excess renewable energy (solar and/or wind). The DFM consists of Ru as methanation catalyst and nano dispersed CaO as CO₂ adsorbent, both supported on a porous γ-Al₂O₃ carrier. A spillover process drives CO₂ from the sorbent to the Ru sites where methanation occurs using stored H₂ from excess renewable power. This approach utilizes flue gas sensible heat and eliminates the current energy intensive and corrosive capture (amine solutions) and storage processes without having to transport captured CO₂ or add external heat.
The catalytic component (Ru/γ-Al₂O₃) has been investigated in terms of its suitability for a DFM process. Process conditions for methanation have been optimized. It has been observed that the equilibrium product distribution for CO₂ methanation with a H₂:CO₂ ratio of 4:1 can be attained at a temperature of 280°C with a space velocity of 4720 h⁻¹. TGA-DSC has been employed to observe the sequential adsorption and reaction of CO₂ and H₂ over Ru/γ-Al₂O₃. It was shown that H₂ only reacts with a CO₂-saturated Ru/γ-Al₂O₃ surface but does not adsorb on the bare Ru surface at 260°C, consistent with an Eley-Rideal type reaction. In this rate model CO2 adsorbs strongly on the catalyst surface and reacts with gas phase H₂. Kinetic tests were employed to confirm this observation and demonstrated that the rate dependence on CO₂ and H₂ was also consistent with an Eley-Rideal mechanism. A rate expression according to the Eley-Rideal model at 230°C was developed.
Activation energy, pre-exponential factor and reaction orders with respect to CO₂, H₂, and products CH₄, and H₂O were determined in order to develop an empirical rate equation in a range of commercial significance. Methane was the only hydrocarbon product observed during CO₂ hydrogenation. The activation energy was found to be 66.084 kJ/g-mole CH₄. The empirical reaction order for H₂ was 0.88 and for CO₂ 0.34. Product reaction orders were essentially zero.
|
80 |
Excimer Laser Crystallization of Silicon Thin-Films for Monolithic 3D IntegrationCarta, Fabio January 2015 (has links)
In 1964 the first metal oxide semiconductor (MOS) integrated circuit (IC) became available. Shortly after in 1965 Gordon Moore predicted the pace of the device density increase in ICs. His prediction became a self-fulfilling prophecy, which taking advantage of the formal device scaling rules introduced by Robert Dennard in 1974, drove the evolution of the integrated electronic industry.
In conventional two dimensional ICs, devices are integrated into a single layer of silicon in what is called the front end of line (FEOL) fabrication. Additional layers on top of the devices serve as inter-dielectric isolating layer and metal interconnects and are fabricated in the back end of line (BEOL) process. Scaling the dimension of devices allows for an increase in device density, improvement on device switching speed and reduction of the cost per device. The conjunction of these benefits drove the industry thus far. Over the past decade further scaling the devices while achieving also an increase in performance and cost benefits became extremely difficult. As the dimensional scaling of complementary MOS (CMOS) devices reaches its limits, three dimensional ICs (3DICs) are increasingly being considered as a path to achieve higher device densities. 3DICs offer a way to increase density by using multiple device layers on the same die, reducing the interconnect distance and allowing for a decrease in signal delay. Among different fabrication techniques, monolithic 3D integration is potentially more cost effective but requires high performance devices, a process compatible with transistor integration in the BEOL stack and needs to deliver a high device density and uniformity in order to be adopted by the very large scale integration (VLSI) industry.
This work focuses on a particular laser crystallization technique to achieve monolithic device integration. The technique, called Excimer Laser Crystallization (ELC), makes use of an excimer laser to achieve a large grain polycrystalline thin-film starting from an amorphous layer, allowing integration of high quality thin-film transistors (TFTs). Thus far, the ELC technique has been studied on thin-films typically deposited on top of quartz substrate or Si/SiO₂ wafers. On the other hand state of the art VLSI integration uses more advance BEOL stacks with low-κ material as interlayer dielectrics (ILDs) to passivate the copper (Cu) interconnect lines. This thesis focuses on three different key aspect to enable laser crystallization in the BEOL for device integration: 1. Excimer laser crystallization of amorphous silicon on low-κ dielectric; 2. Excimer laser crystallization of amorphous silicon on BEOL processed wafer; 3. VLSI of TFTs on excimer laser crystallized silicon.
The ELC of amorphous silicon on low-κ dielectric is first explored through one dimension (1D) finite element method (FEM) simulation of the temperature evolution during the laser exposure in two different systems: 1. amorphous silicon deposited on top of SiO₂ dielectric; 2. amorphous silicon deposited on top of low-κ dielectric. Simulations predict that is necessary a lower laser energy for crystallizing the silicon on the low-κ material. Experimental observations confirm the predicted behavior yielding a 35% lower energy for crystallization of thin-film silicon on top of a low-κ dielectric. Material characterization through defect enhanced SEM micrograph, Raman spectroscopy and XRD analysis shows an equivalent material morphology for the two system with a preferential (111) crystal orientation for the SiO₂ system.
Silicon crystallization on BEOL processed wafer is studied through a combination of 1D FEM simulation and experimental observation on a silicon layer deposited on top of a SiO₂dielectric protecting the underlying damascene Cu structure. 1D FEM show that during the silicon laser exposure, because of the short pulse width of the laser (30 ns), the heat is retained in the amorphous silicon layer allowing its melting while keeping the temperature of the Cu lines below 320 °C which is a favorable condition for monolithic integration in the BEOL. Further experimental evidences show the ability of crystallizing a-Si on such structure while preserving the physical and electrical properties of the Cu lines.
The feasibility of monolithic VLSI 3D integration is demonstrated through integration of TFTs devices on 200 mm silicon wafers. The integration process and performance of the TFTs device are modeled through technology computer aided design (TCAD) simulations which are used to define the process flow and the fabrication parameters. Characterization of the TFTs over multiple die yield good device performance and uniformity. TFTs characterized with 1.5 V of supply voltage have a sub-threshold slope down to 79 mV/decade, current density up to 26.3 μA/μm, a threshold voltage of 0.23 V, current On/Off ratio above 10⁵ and device field effect mobility up to 19.8 cm²/(V s) for LPCVD-sourced silicon. Furthermore, the Levinson method allows characterization of the trap density in the thin-film polysilicon devices yielding a mean value 8.13×10¹² cm².
This work present an integration scheme which proves to be compatible with VLSI in the BEOL of wafers. It paves the way to further development which could lead to an high performance, cost effective, monolithic 3D integration approach useful in application such as reconfigurable logic, display, heterogeneous integration and on chip optical communications.
|
Page generated in 0.1037 seconds