Hard Disk Manufacturing Research Articles

An EWMA chart for high dimensional process with multi-class out-of-control information via random forest learning

ABSTRACT Modern manufacturing and quality monitoring involve multi-class out-of-control (OOC) information from the training sample. It is essential to use such information during online monitoring of data streams from complex processes. In this paper, a monitoring framework is designed by combining the random forest technique with the exponentially weighted moving average method for monitoring complex processes with multi-class OOC information. To be specific, a process surveillance technique in the form of a control chart is proposed based on the probability that the online data is classified as an in-control (IC) sample, and the control chart triggers an alarm when the probability is lower than the control limit. Our numerical findings based on the Monte–Carlo simulation show that the proposed control chart performs more effectively than its competitors under various distributions and data types, especially for high-dimensional cases when multi-class OOC information is known in advance. Moreover, the proposed method is illustrated with an application using the data related to the hard disk manufacturing processes.

A Deep Learning System for Recognizing and Recovering Contaminated Slider Serial Numbers in Hard Disk Manufacturing Processes.

This paper outlines a system for detecting printing errors and misidentifications on hard disk drive sliders, which may contribute to shipping tracking problems and incorrect product delivery to end users. A deep-learning-based technique is proposed for determining the printed identity of a slider serial number from images captured by a digital camera. Our approach starts with image preprocessing methods that deal with differences in lighting and printing positions and then progresses to deep learning character detection based on the You-Only-Look-Once (YOLO) v4 algorithm and finally character classification. For character classification, four convolutional neural networks (CNN) were compared for accuracy and effectiveness: DarkNet-19, EfficientNet-B0, ResNet-50, and DenseNet-201. Experimenting on almost 15,000 photographs yielded accuracy greater than 99% on four CNN networks, proving the feasibility of the proposed technique. The EfficientNet-B0 network outperformed highly qualified human readers with the best recovery rate (98.4%) and fastest inference time (256.91 ms).

Signal Analysis of Scratch-Detection on Magnetic Disc by Using Light Reflection Approach

In hard disk manufacturing, a process of quality inspection of magnetic disc is crucially focused on finding traces of scratch that occur on the surface by various sources such as production process, machinery, tribology or recording head. It may affect the efficiency of magnetic disc’s read/write ability. Many approaches have been proposed to detect the scratch by either destructive or non-destructive testing. In this study, it shows an analysis of signals from the detected scratches on the magnetic disc by reflection light method, using a CD pick-up head. It works by producing an incidental laser beam to the surface and then detecting it by a photodiode detector. The results showed that when the laser beam is incident onto the magnetic disc with/ without scratch, reflecting lights to the photodiode are different and the signal changes according to the intensity of the incidental/reflecting lights.

Identification of individual features in areal surface topography data by means of template matching and the ring projection transform

Starting from areal surface topography data as provided by current commercial three-dimensional (3D) profilometers and 3D digital microscopes, this work investigates the problem of automatically identifying and extracting functionally relevant, individual features within the acquisition area. Feature identification is achieved by adopting an original template-matching algorithmic procedure, based on applying the ring projection transform in combination with a parametric template. The proposed algorithmic procedure addresses in particular template-matching scenarios where significant variability may be associated with the features to be compared to the reference template. The algorithm is applied to a test case involving the characterization of the surface texture of a superabrasive polishing tool used in hard-disk manufacturing.

Human Factor in Equipment Safety of Hard Disk Manufacturing

The focus of this paper is on the human factors aspect of equipment safety focusing on process hazard analysis and safety programs. The objectives of this study are to determine the level of awareness, involvement and perceptions on occupational safety and health among workers. These are accomplished by conducting a survey using questionnaire and interviews involving 98 respondents in two hard disk manufacturing companies. Majority of the respondents were executives with high academic qualifications and possessed high level of safety awareness. The results conclude that despite the existence of various legislations, human factors and its interactions with facilities and management system is the main contributor to incidents. It is also found that most of the incidents involved power driven machineries. For the improvement of equipment safety, it is recommended that Process Hazard Analysis techniques be used to play significant role, along with both engineering and administrative control measures and safety programs. Keywords: Equipment safety; industrial incidents; process hazard analysis; hard disk industry

Human Factor in Equipment Safety of Hard Disk Manufacturing

The focus of this paper is on the human factors aspect of equipment safety focusing on process hazard analysis and safety programs. The objectives of this study are to determine the level of awareness, involvement and perceptions on occupational safety and health among workers. These are accomplished by conducting a survey using questionnaire and interviews involving 98 respondents in two hard disk manufacturing companies. Majority of the respondents were executives with high academic qualifications and possessed high level of safety awareness. The results conclude that despite the existence of various legislations, human factors and its interactions with facilities and management system is the main contributor to incidents. It is also found that most of the incidents involved power driven machineries. For the improvement of equipment safety, it is recommended that Process Hazard Analysis techniques be used to play significant role, along with both engineering and administrative control measures and safety programs. Keywords: Equipment safety; industrial incidents; process hazard analysis; hard disk industry

Calculation on Temperature Rise of CMP Process: Roughness Effects Considered

Chemical mechanical polishing (CMP) has been widely accepted in modern integrated circuit (IC) industries and hard disk manufacturing processes, to insure wafer surface with high level of global and local planarity required. In CMP process, temperature rise has two-edged influences: temperature-rise accelerates the chemical activity and the motion of nano-particles contained in the slurry through which material removal ratio (MRR) is enhanced; the other side of the same coin is, however, that it will soften the pad surface and subsequently reduce the MRR. Furthermore, it is found that temperature-rise would cause agglomeration of nano-particles, which would cause the wafer surface defect. The net effects of temperature thus should be under investigation with scrutiny. In an attempt to study the temperature variation and influencing rule, in this paper we firstly establish the flow equation considering pad roughness, coupling the energy equation on the basis of thermodynamics. Then, by taking numerical simulation which is carried out to give out the temperature distribution of conventional CMP process. The results show that temperature-rise in the CMP process is very minor. The research will surely shed some lights on the mechanism of CMP and lay a feasible foundation for possible future utilization.

A Preliminary Model Accounting for the Suction Pressure in CMP Process

Chemical mechanical polishing (CMP) is a widely used technique to achieve high level of global and local planarity required in modern integrate circuit (IC) industries and hard disk manufacturing process, etc., which pleas for concentrate researches. The main purpose of the present research is in an attempt to express the counterintuitive experimental aftermath: the ‘negative’ pressure, i.e., a suction force occurred in conventional commercial CMP process. A preliminary two tiers wafer-scale flow model for CMP is presented considering the roughness as well as the elasticity of the bulk pad substrate. Numerical simulations were conducted to elucidate the contact pressure and flow pressure distributions. The results show that a divergence region appears near the leading edge, which contributes to the suction pressure. A stress-richened area near the edges will give rise to over polishing. The research aftermaths agree well with the experiments, that validate the proposed analysis to some extend. This will shed lights on the mechanism of CMP process, which for a long time is considered as a black art where empirical or semi-empirical data are dependent upon to optimize the CMP parameters.

Nanoparticle scanning and detection on flat and structured surfaces using fluorescence microscopy

A new technique is proposed for the scanning and detection of nanoparticles on flat substrates and three-dimensional structures using fluorescence microscopy. This technique is utilized for particle removal measurements especially in semiconductor and hard disk manufacturing. This fluorescent particle scanning technique enables nanoscale particle detection. The technique shows that single particles down to 63 nm could be detected and counted. The technique is also capable of detecting particles in trenches that are as deep as 500 microm.

Superlattice source and its application in Co/Pd multilayers for perpendicular magnetic recording

This article reports a sputtering system, the Intevac Superlattice Source (SLS), designed to meet the needs of making perpendicular multilayers for hard disk manufacturing. The SLS is a two-sided sputtering station in which each side is capable of depositing multiple sequential bilayers of two different film materials. Each side of the source contains a magnetron (sharing the same centerline as the substrate) and a rotating cathode on which are mounted two semicircular targets. The two rotating cathodes (one on each side of the source) are synchronized to ensure matched elements sputtering simultaneously and to prevent cross contamination between target materials. The design easily integrates onto standard hard disk manufacturing systems and can deposit up to 5 bilayers per second. Using this source, a series of Co/Pd multilayers has been made. We investigated the effect of different seedlayers, Pd, Ti, and Ru, on the coercivity and microstructure of Co/Pd multilayers. We found that although the crystalline growth is very different for each of the three seed layers studied, all three lead to higher coercivity and unique squareness of Co/Pd multilayers as compared to multilayers made without any seedlayer. The coercivity achieved for a 300 Å thick seedlayer [Co 4 Å/Pd 10 Å]×15 multilayer is 6253, 6871, and 6920 Oe for Pd, Ti, and Ru seedlayers, respectively.

Fuzzy rule learning during simulation of manufacturing resources

Knowledge-based simulation is an effective way to deal with a wide variety of uncertainty and complexity issues. In this context, knowledge acquisition plays a central role and addresses what has been a bottleneck in most simulation models. In this paper, a new knowledge auto-acquisition technique for simulation of manufacturing resources using genetic algorithm (GA) to learn fuzzy knowledge is proposed. Given the fuzzy variables and the number of fuzzy sets, the fuzzy rules will be automatically deduced via learning. In the proposed approach, all potential fuzzy relations among fuzzy variables are enumerated into a form of matrix function, manipulated by means of GA operating on the simulation model. This novel matrix equation can be de-numerated into symbolic rules, put into dynamic simulation resources, and simulated interactively. The paper concludes with an application in a final assembly environment of a hard disk manufacturing company.

Removing Spherical Silica Particles from Si, Ge and NiP Substrates by KrF Excimer Laser

Particle adhesion and removal are of great interest in IC process and high-density hard disk manufacturing. Increasing effort is required to remove submicron particles. We investigated the adhesion and removal of spherical silica particles from silicon (Si), germanium (Ge) and NiP substrates using a KrF excimer laser. Atomic force microscope (AFM) point-contact mode measurement showed that the tip adhesion force on Ge surface is greater than that on Si or NiP surface, and the Hamaker constant of NiP is about 2/3 of that of Ge substrate. The cleaning efficiency and cleaning thresholds for standard-size particles with a diameter of 1.0 µm were measured. The cleaning efficiency was sensitive to laser fluence. The threshold laser fluences for Si, Ge and NiP are 100, 30 and 8 mJ/cm2, respectively. The different cleaning curves are not only due to the difference in adhesion on the substrates, but also due to the different optical and thermal properties of the substrates.

An introduction to sputtering of magnetic and magneto-optic thin films for data recording

After a simple introduction to sputtering, the different techniques that have been used for depositing magnetic films are outlined. Magnetron sputtering is the most common but ion plating, magnetic ion beam and argon ion beam source sputtering provide alternatives for specific applications. Ion plating and argon ion beam source sputtering can be used for amorphous data recording films whereas ion beam deposition produces a high degree of crystallinity. Magnetron sputtering in a Winchester hard disk manufacturing unit is described in some detail. All types of optical discs require sputtered films and the requirements are reviewed for CDROM, WORM and magneto-optic WREM media. Finally, some of the novel multi-layer sputtered magneto-optic films are briefly mentioned.