Abstract: A novel method for analyzing the pulse train resulting from a scanned beam particle monitor is described. The method enhances signal-to-noise ratio and significantly reduces particle false alarm rate. Performed in the time domain, the method filters noise pulses that do not occur at the scanner frequency. The analysis further identifies particle-pulse-envelopes (PPE's) by performing a forward-looking and backward-looking autocorrelation. Gaussian fits are subsequently applied to identified particle-pulse envelopes, to determine particle characteristics, such as size and speed.

Abstract: A novel method for analyzing the pulse train resulting from a scanned beam particle monitor is described. The method enhances signal-to-noise ratio and significantly reduces particle false alarm rate. Performed in the time domain, the method filters noise pulses that do not occur at the scanner frequency. The analysis further identifies particle-pulse-envelopes (PPE's) by performing a forward-looking and backward-looking autocorrelation. Gaussian fits are subsequently applied to identified particle-pulse envelopes, to determine particle characteristics, such as size and speed.

Abstract: A particle sensor for a semiconductor device fabrication tool scans a laser beam through a measurement volume immediately adjacent a wafer during processing and detects light scattered by particles adjacent the wafer. Scanning provides a real-time count of particles without interfering with processing. Detected light can be forward-scattered, side-scattered, or back-scattered depending on available optical access for a detector portion of the sensor. A pulse in the intensity of scattered light results each time a particle is scanned. Because the scanning velocity is high relative to the particle velocity, each particle may be scanned several times while the particle is in the measurement volume. Analysis fits a series of pulses observed for a single particle to a Gaussian distribution to determine a size, position, and velocity for each particle and a time-resolved particle count of the particles.

Abstract: A method applicable to an ensemble laser diffraction (ELD) instrument computes a particle size distribution in real time after correction for the multiple scattering phenomena. In one embodiment, a numerical method, similar to the Newton's method, is provided to iteratively calculate the single scattering mode. The present method is hence suitable for use, with high accuracy, in real time controlling and monitoring applications.

Abstract: An ensemble scattering particle sizing system employing optical means and unique methodology which generates reliable data relative to particle concentration, size distribution and spatial distribution, including axial spatial distribution, for particles disposed in light transmitting medium gas with a given sample volume, means including a beam transmitter, a transform lens, an apertured image plane and a relay lens are strategically associated with a sample volume and detector means to produce the desired results.