高能量飞秒激光脉冲对硅和玻璃的附带损伤

采用高能量飞秒激光脉冲对硅片和载玻片的未抛光表面进行了烧蚀,对烧蚀后样品的表面形貌和纳米力学性能进行了检测。用显微镜对烧蚀后样品的表面形貌进行检测,硅片烧蚀后正面形貌随位置不同发生渐变,而烧蚀后玻璃正面烧蚀区与未烧蚀区有明显界限。用原子力显微镜对烧蚀后样品的预抛光背面进行了检测,发现硅片烧蚀前后背面形貌未发生显著变化;而玻璃烧蚀前后背面形貌发生了明显改变,且烧蚀区与未烧蚀区的背面形貌有较为明显的界限。利用原位纳米力学测量仪对样品抛光背面的纳米力学性能进行了测量,发现硅片背面纳米力学性能有与正面烧蚀区渐变形貌相对应的规律性变化,而玻璃烧蚀前后背面的纳米力学性能则无显著变化。样品表面微观形貌和纳米力学性能的规律性变化在一定程度上反映了高能量飞秒激光脉冲在样品背面产生附带损伤的分布情况。     

SUBSURFACE DAMAGE IN GRINDING TITANIUM ALUMINIDE

Abstract

This paper reports on an experimental investigation of grinding-induced subsurface damage in gamma titanium aluminide (γ-TiAl). Grinding was carried out with resin-bond diamond wheels having the same concentration but different grit sizes, and with a vitreous-bond silicon carbide wheel using various depths of cut. The extent of the subsurface damage was determined by a bonded interface technique and optical microscopy. The grinding-induced damage beneath the ground surfaces was found to consist of plastic deformation and microcracks. The severity and depth of the subsurface damage zone increased with an increase in the abrasive grit size and the depth of cut. Microcracks were observed when grinding was performed with the silicon carbide wheel at a depth of cut of 100 and 125 μm.     

PHOEBE,a prototype scanning laser-feedback microscope for imaging biological cells in aqueous media

Based on the principle of laser-feedback interferometry (LFI), a laser-feedback microscope (LFM) has been constructed capable of providing an axial (z) resolution of a target surface topography of ～ 1 nm and a lateral (x, y) resolution of ～ 200 nm when used with a high-numerical-aperture oil-immersion microscope objective. LFI is a form of interferometry in which a laser's intensity is modulated by light re-entering the illuminating laser. Interfering with the light circulating in the laser resonant cavity, this back-reflected light gives information about an object's position and reflectivity. Using a 1-mW He–Ne (λ = 632·8 nm) laser, this microscope (PHOEBE) is capable of obtaining 256 × 256-pixel images over fields from (10 μm × 10 μm) to (120 μm × 120 μm) in ～ 30 s. An electromechanical feedback circuit holds the optical pathlength between the laser output mirror and a point on the scanned object constant; this allows two types of images (surface topography and surface reflectivity) to be obtained simultaneously. For biological cells, imaging can be accomplished using back-reflected light originating from small refractive-index changes (> 0·02) at cell membrane/water interfaces; alternatively, the optical pathlength through the cell interior can be measured point-by-point by growing or placing a cell suspension on a higher-reflecting substrate (glass or a silicon wafer). Advantages of the laser-feedback microscope in comparison to other confocal optical microscopes include: the simplicity of the single-axis interferometric design; the confocal property of the laser-feedback microscope (a virtual pinhole), which is achieved by the requirement that only light that re-enters the laser meeting the stringent frequency, spatial (TEM₀₀), and coherence requirements of the laser cavity resonator mode modulate the laser intensity; and the improved axial resolution, which is based on interferometric measurement of optical amplitude and phase rather than by use of a pinhole as in other types of confocal microscopes.     

SUBSURFACE DAMAGE IN GRINDING TITANIUM ALUMINIDE

This paper reports on an experimental investigation of grinding-induced subsurface damage in gamma titanium aluminide (γ-TiAl). Grinding was carried out with resin-bond diamond wheels having the same concentration but different grit sizes, and with a vitreous-bond silicon carbide wheel using various depths of cut. The extent of the subsurface damage was determined by a bonded interface technique and optical microscopy. The grinding-induced damage beneath the ground surfaces was found to consist of plastic deformation and microcracks. The severity and depth of the subsurface damage zone increased with an increase in the abrasive grit size and the depth of cut. Microcracks were observed when grinding was performed with the silicon carbide wheel at a depth of cut of 100 and 125 μm.     

A state-of-the-art review of ductile cutting of silicon wafers for semiconductor and microelectronics industries

Silicon is a typical functional material for semiconductor and optical industry. Many hi-tech products like lenses in thermal imaging, solar cells, and some key products of semiconductor industry are made of single crystal silicon. Silicon wafers are used as substrate to build vast majority of semiconductor and microelectronic devices. To meet high surge in demand for microelectronics based products in recent years, the development of rapid and cost efficient processes is inevitable to produce silicon wafers with high-quality surface finish. The current industry uses a sequence of processes such as slicing, edge grinding, finishing, lapping, polishing, back thinning, and dicing. Most of these processes use grinding grains or abrasives for material removal. The mechanism of material removal in these processes is fracture based which imparts subsurface damage when abrasive particles penetrate into the substrate surface. Most of these traditional processes are extremely slow and inefficient for machining wafers in bulk quantity. Moreover, the depth of subsurface damage caused by these processes can be up to few microns and it is too costly and time consuming to remove this damage by heavy chemical–mechanical polishing process. Therefore, semiconductor industry requires some alternative process that is rapid and cost effective for machining silicon wafers. Ductile cutting of silicon wafer has the potential to replace the tradition wafer machining processes efficiently. If implemented effectively in industry, ductile cutting of silicon wafers should reduce the time and cost of wafer machining and consequently improve the productivity of the process. This paper reviews and discusses machining characteristics associated with ductile cutting of silicon wafers. The limitations of traditional wafer fabrication, the driving factors for switching to ductile cutting technology, basic mechanism of ductile cutting, cutting mechanics, cutting forces, surface topography, thermal aspects, and important factors affecting these machining characteristics have been discussed to give a systematic insight into the technology.     

硅片低损伤磨削砂轮及其磨削性能

针对传统金刚石砂轮磨削硅片存在的表面/亚表面损伤问题,研制了一种用于硅片化学机械磨削加工的新型常温固化结合剂软磨料砂轮。根据化学机械磨削加工原理和单晶硅的材料特性,设计的软磨料砂轮以氧化铈为磨料,二氧化硅为添加剂,氯氧镁为结合剂。研究了软磨料砂轮的制备工艺,分析了软磨料砂轮的微观组织结构和成分。通过测量加工硅片的表面粗糙度、表面微观形貌和表面/亚表面损伤,进一步研究了软磨料砂轮的磨削性能。最后,与同粒度金刚石砂轮磨削和化学机械抛光(CMP)加工的硅片进行了对比分析。结果表明,采用软磨料砂轮磨削的硅片其表面粗糙度Ra1nm,亚表面损伤仅为深度30nm的非晶层,远好于金刚石砂轮磨削硅片,接近于CMP的加工水平,实现了硅片的低损伤磨削加工。     

A THRESHOLD MODEL FOR LASER CLEANING OF LARGER SILICON WAFERS

High cleaning quality for silicon wafers without damage is a challenge in laser cleaning technologies. Laser cleaning of Al₂O₃ micro-particles, which are the main contaminants of silicon wafer lapping and polishing solutions used in industry, from silicon wafers was studied for determining laser energy for high efficient particle removal while not causing damage to the wafers. As the cleaning force is generated from laser-energy absorption and conduction of the wafer, heat-conduction model on silicon wafer was developed during laser irradiation using a finer finite element method, from which cleaning force exerting on the particles greater than the adhesion force between the particle and the substrate, but less than the wafer damage energy of laser input was determined. Calculations of the laser energy threshold values for both particle cleaning and wafer damage were conducted for silicon wafers of 200 mm in diameter and 0.2 mm in thickness, and they were found to be about 60 mJ/cm² and 320 mJ/cm², respectively. The laser energy threshold model was finally verified experimentally using a KrF Excimer laser and found to be in good agreement with the experimental data. With the cleaning parameters from the model, the cleaning efficiency of as high as 98% has been achieved.     

工件旋转法磨削硅片的磨粒切削深度模型

半导体器件制造中,工件旋转法磨削是大尺寸硅片正面平坦化加工和背面薄化加工最广泛应用的加工方法。磨粒切削深度是反映磨削条件综合作用的磨削参量,其大小直接影响磨削工件的表面/亚表面质量,研究工件旋转法磨削的磨粒切削深度模型对于实现硅片高效率高质量磨削加工具有重要的指导意义。通过分析工件旋转法磨削过程中砂轮、磨粒和硅片之间的相对运动,建立磨粒切削深度模型,得到磨粒切削深度与砂轮直径和齿宽、加工参数以及工件表面作用位置间的数学关系。根据推导的磨粒切削深度公式,进一步研究工件旋转法磨削硅片时产生的亚表面损伤沿工件半径方向的变化趋势以及加工条件对磨削硅片亚表面损伤的影响规律,并进行试验验证。结果表明,工件旋转法磨削硅片的亚表面损伤深度沿硅片半径方向从边缘到中心逐渐减小,随着砂轮磨粒粒径、砂轮进给速度、工件转速的增大和砂轮转速的减小,加工硅片的亚表面损伤也随之变大,试验结果与模型分析结果一致。     

湿式机械化学磨削单晶硅的软磨料砂轮及其磨削性能

针对干式机械化学磨削(Mechanical chemical grinding, MCG)单晶硅过程中易产生磨削烧伤、粉尘多、加工环境差等问题,研制一种可用于湿式MCG单晶硅的新型软磨料砂轮,并对砂轮的磨削性能及其磨削单晶硅的材料去除机理进行研究。根据湿式机械化学磨削单晶硅的加工原理和要求,制备出以二氧化硅为磨料、改性耐水树脂为结合剂的新型软磨料砂轮。采用研制的软磨料砂轮对单晶硅进行磨削试验,通过检测加工硅片的表面/亚表面质量对湿式MCG软磨料砂轮的磨削性能进行分析,并与传统金刚石砂轮、干式MCG软磨料砂轮的磨削性能进行对比。采用X射线光电子能谱仪对磨削前后硅片的表面成分进行检测,分析湿式MCG加工硅片过程中发生的化学反应。结果表明,采用湿式MCG软磨料砂轮加工硅片的表面粗糙度Ra值为0.98 nm,亚表面损伤层深度为15 nm,湿式MCG软磨料砂轮磨削硅片的表面/亚表面质量远优于传统金刚石砂轮,达到干式MCG软磨料砂轮的加工效果,可实现湿磨工况下硅片的低损伤磨削加工。在湿式MCG过程中,单晶硅、二氧化硅磨粒与水发生了化学反应,在硅片表面生成易于去除的硅酸化合物,硅酸化合物进一步通过砂...     

飞秒激光参数对硅表面微结构影响的研究

为了制备出表面具有准规则排列的微米量级锥形尖峰结构的黑硅材料,在SF6气体氛围中,用一定能量密度的飞秒脉冲激光照射单晶硅片表面。针对激光通量和激光峰值功率这两个参量分别进行实验,具体分析了15fs和130fs脉冲宽度的飞秒激光脉冲作用下硅表面微结构的形成,不同实验条件下制备出的硅微纳结构也有明显的差异。研究表明,在同一背景气体下,激光的峰值功率对硅表面微结构的形成起着决定性的作用。     

激光清洗硅片表面Al2O3颗粒的试验和理论分析

以KrF准分子激光器为激光源,对目前工业上常用的硅片研磨抛光液的主要成分Al₂O₃颗粒进行激光清洗的试验和理论分析。建立一维热传导模型,利用有限元分析软件MSC.MarC模拟硅片表面的温度随激光作用时间和能量密度的分布。通过理论计算,量化了颗粒所受到的清洗力以及其与硅片表面之间的粘附力,理论预测出1 μm Al₂O₃颗粒的激光清洗阈值为60 mJ/cm²。在理论分析的指导下,利用248 nm、30 ns的KrF准分子激光进行单因素试验,研究激光能量密度、脉冲个数、激光束入射角度对激光干法清洗效率的影响,并且实验验证了清洗模型以及场增强效应对激光清洗结果的影响。     

单晶硅片磨削表面相变研究

为了揭示硅片自旋转磨削加工过程中材料的去除机理,采用显微拉曼光谱仪研究了硅片磨削表面的相变。结果表明：半精磨和精磨硅片表面存在-Si相、Si-III相、Si-IV相和Si-XII相,这表明磨削过程中Si-I相发生了高压金属相变（Si-II相）,Si-II相容易以塑性方式去除。粗磨硅片表面没有明显的多晶硅,只有少量的非晶硅出现,材料以脆性断裂方式去除。从粗磨到精磨,材料去除方式由脆性断裂去除向塑性去除过渡。粗磨向半精磨过渡时,相变强度越大,材料的塑性去除程度越大;半精磨向精磨过渡时,相变强度越小,材料的塑性去除程度越大。     

A new model of grit cutting depth in wafer rotational grinding considering the effect of the grinding wheel,workpiece characteristics,and grinding parameters

Wafer rotational grinding is widely employed for back-thinning and flattening of semiconducting wafers during the manufacturing process of integrated circuits. Grit cutting depth is a comprehensive indicator that characterizes overall grinding conditions, such as the wheel structure, geometry, abrasive grit size, and grinding parameters. Furthermore, grit cutting depth directly affects wafer surface/subsurface quality, grinding force, and wheel performance. The existing grit cutting depth models for wafer rotational grinding cannot provide reasonable results due to the complex grinding process under extremely small grit cutting depth. In this paper, a new grit cutting depth model for wafer rotational grinding is proposed which considers machining parameters, wheel grit shape, wheel surface topography, effective grit number, and elastic deformation of the wheel grit and the workpiece during the grinding process. In addition, based on grit cutting depth and ground surface roughness relationship, a series of grinding experiments under various grit cutting depths are conducted to produce silicon wafers with various surface roughness values and compare the predictive accuracy of the proposed model and the existing models. The results indicate that predictions obtained by the proposed model are in better agreement with the experimental results, while accuracy is improved by 40%–60% compared to the previous models.     

Effect of surface etching on the lubricated sliding wear of an eutectic aluminium–silicon alloy

The paper presents a study of the formation of wear grooves on near-eutectic aluminium–silicon alloy flats, by sliding a steel ball. The formation of the grooves are tracked on etched and unetched flats as functions of normal load and sliding distance. The groove is initially formed by plastic flow, and then expanded by micro-abrasion as the ball continues to slide on the groove. Etching causes surface hardening of the alloy, but, more importantly creates a surface topology that reduces the peak contact pressure, which discourages further plastic flow in the subsurface. This effect is rationalised using an existing contact mechanical model of indentation of rough surfaces.     

Analyzing Elastic-Plastic Real Rough Surface Contact in Running-in

A numerical method is presented for evaluating the elastic-elastic contact of real rough surface contacts during running-in. For the surface contact, an elastic-plastic model based on the variational method is applied to analyze the pressure distribution and contact area of worn surfaces during running-in. In conjunction with the classical statistic model of Greenwood and Williamson, the numerical result showed that the plasticity index Ψ was decreased to one in the elastic range as running-in proceeded. In comparison with the Hertzian solution, the influence of the asperities is very significant on the pressure distribution, thereafter causing a higher peak value of contact pressure. For the subsurface, the interior stress from the von Mises criterion was calculated to evaluate the subsurface stress field subject to both normal and tangential forces. In the calculated of the interior stress, the total stress is decomposed into a fluctuating component and a smooth component. The fluctuating part is solved by using FFT from the concept of the convolution theorem while the smooth part is obtained directly by analytical solution. Calculations of contact area and subsurface stress on experimentally produced surfaces whose topography has been determined using an atomic force microscope and friction coefficient front sliding have been carried out. The results showed that asperities and friction coefficient gave rise to stress increase in the near-surface stress field and produced a high stress zone towards the surface. As a result, transverse asperity cracking was produced. The calculations and supporting experimental evidence clearly confirmed that the reduction of peak pressure during running-in decreased the plastic deformation of contact.     

利用图像处理技术评价硅片表面清洗率

介绍了一种基于Matlab图像处理工具箱技术的评价硅片表面污染颗粒激光清洗率的新方法。借助Matlab图像处理工具箱,对清洗前后硅片表面光学显微镜照片进行处理,编写硅片表面激光干法清洗率的评价程序,统计清洗前后硅片表面评价区域的污染颗粒个数,对清洗效果进行定量评价。研究结果证明,利用此方法统计的颗粒数准确度达97.6%,得到的激光清洗率准确度达99.2%。结果表明,借助图像处理技术评定清洗效果是一种高效、快速、准确的新方法。     

Pad surface and coefficient of friction correlation analysis during friction between felt pad and single-crystal silicon

In wafer polishing pad surface plays a crucial role in the polishing process. With the increase of friction time between pad and wafer, the pad becomes flattened or glazed with particles clogging the pores of the pad and forming a layer of slurry residue and wafer particles, leading to changes of COF, material removal rates and higher defects on the wafer surface. Thus, this study aims to determine the correlation between pad surface deformation, slurry adhesive rate and Coefficient of friction (COF) during friction between felt pad and single -crystal silicon, to analyze the relationship between pad condition and COF. The real-time COF between felt pad and single-crystal silicon wafer are tested which are sorted in groups depending on various loads and oscillation frequencies and surfaces of felt pads measuring by Scanning electron microscope (SEM) are compared. The correlation between pad surface deformation and abrasive adhesion and COF is evaluated through analyzing the experiment results.

     

润湿性梯度集油表面制备及摩擦性能研究*

为了解决润滑油发生迁移，使接触面润滑油中断，导致润滑失效的问题，制备一种具有高集油性能的疏油-亲油-疏油的梯度表面。采用化学气相沉积的方法，在硅片表面沉积一层单分子膜，并采用接触角测量仪、UMT摩擦磨损试验机、共聚焦显微镜等对该样品进行表征，研究该类表面在限量供油条件下的润滑性能。结果表明：将油滴滴在疏油/亲油交界处，油滴能够迅速地从疏油区域向亲油区域运动；点接触往复运动摩擦实验结果表明，梯度表面硅片的摩擦因数明显低于原始的硅片，且梯度表面硅片的表面磨痕深度比原始硅片浅。各种实验结果表明，所制备的疏油-亲油-疏油梯度表面能够起到集油的作用，避免了润滑油中断的问题。     

弹流润滑条件下织构表面亚表层特性研究

为研究在弹流润滑状态下表面形貌对亚表层特性的影响，利用激光加工方法获得2种微凹坑型织构表面形貌，通过将实测的表面形貌坐标输入弹流润滑数值计算程序得到油膜压力和膜厚分布；以对应工况的油膜压力作为表面法向压应力，利用Rabinowicz经验公式算出剪切应力；将表面法向压应力和切向剪应力叠加后对弹流润滑界面亚表层特性进行仿真研究。结果表明：表面织构使亚表层应力分布发生显著改变；微凹坑直径、卷吸速度对亚表层应力的大小与分布有不同的影响；亚表层变形在摩擦过程中呈现随深度增加先缓慢减小后快速下降的规律，研究结果将为通过表面形貌设计改善轴承等零件受力状况提供理论支持。     

On the stiction of MEMS materials

Stiction is a serious problem in microelectromechanical systems (MEMS) due to their large surface area-to-volume ratio. Stiction is closely related to surface forces, which greatly depend on the materials used, surface topography and surface treatment process. In this paper, we investigate surface energies and stiction of commonly used MEMS materials by contact angle measurements and atomic force microscopy (AFM). Dispersive and polar components of surface energies are calculated by Owens–Wendt–Rabel–Kaelble method. Silicon and silicon-related materials have higher polar surface energies than SU-8 and poly-methylmethacrylate (PMMA), thereby have larger surface energies and enhanced tendency for stiction. The nano-scale adhesion forces between Si₃N₄ tip and surfaces obtained by AFM further verified that silicon wafer with native oxide has 3–4 times higher adhesion force than SU-8 and PMMA. It has been shown that the materials with higher surface energy have higher sticton/adhesion forces. The topography of surface influences the contact angle and stiction, and is also discussed in the paper.