Optimized optical system design for endpoint detection

Details Optimized optical system design for endpoint detection Optimized optical system design for endpoint detection

1999-11-24

2003-04-15

Rutledge, Della J (Department: 2851)

Photography

Fluid-treating apparatus

Fluid application to one side only of photographic medium

C156S345250, C156S250000, C156S559000, C156S356000, C156S365000, C156S118000, C156S712000

Reexamination Certificate

active

06547458

ABSTRACT:

FIELD OF THE INVENTION
The present invention concerns the field of optical monitoring of integrated circuit fabrication.
BACKGROUND OF THE INVENTION
During the fabrication of integrated circuits, a desired circuit pattern for a given layer of the integrated circuit is etched into a dielectric film. To accomplish this, a photoresist material is disposed on the area of the dielectric film where etching is not desired. An etching gas, which is chemically reactive with the dielectric material and less chemically reactive with the photoresist, is generated in a plasma. The plasma is then supplied to the surface of the dielectric being etched, causing the etching gas to diffuse into the surface of the dielectric film. The etching gas chemically reacts with the dielectric film to form a volatile by-product. The volatile by-product is then desorbed from the surface of the dielectric film and diffuses into the bulk of the etching gas.
After the pattern is etched into the dielectric layer, the photoresist that was used to define the metal circuit pattern on the dielectric layer is removed. Any post-etch residues including sidewall polymer deposition also must be thoroughly removed or stripped from the underlying layer. One dry process used to strip photoresist and photoresist residues from the dielectric layer is known as ashing. The process of ashing is similar to the etching process. The gas used for ashing, however, is more chemically reactive with the photoresist than with-the dielectric. The ashing gas chemically reacts with the photoresist to form a volatile by-product. The volatile by-product diffuses into the bulk of the ashing gas. After the ashing process is complete, the etched pattern is filled with copper or other conductive material.
Optical emission spectroscopy has previously been used to determine the end point of the etching process by providing information about the etching gas and the by-product of the etching gas and dielectric material. The technique relies on the change in the emission intensity of characteristic optical radiation from the dielectric by-product in the plasma. Excited atoms or molecules emit light when electrons relax from a higher energy state to, a lower energy state. Atoms and molecules of different chemical compounds emit a series of unique spectral lines. The emission intensity for each chemical compound within the plasma depends on the relative concentration of the chemical compound in the plasma. A typical optical emission spectroscopy apparatus operates by measuring the emission intensities of the reactive etching gas and the by-product of the etching gas and the dielectric. The emission of the by-product decreases and finally stops when an endpoint is reached. The optical emission spectroscopy apparatus senses the declining emission intensity of the by-product to determine this endpoint.
It is very important to accurately determine the endpoint of stripping, etching, or residue-removal processing of wafers for semiconductor devices. Accurate endpoint detection improves throughput and minimizes damage to other wafer layers. Over-ashing and under-ashing produce undesirable patterns in the integrated circuit wafer. It is difficult to accurately determine endpoint, because process chambers for semiconductor wafer processing offer very little diagnostic access. Optical emission spectroscopy (OES) has been used to determine endpoint measurements, but is inaccurate because of poor optical access to the region of the wafer. If radiant wafer heating lamps are used to promote the ashing process, the light used to heat the wafer interferes with the light emitted at the wafer surface, used to determine endpoint.
In a radiantly heated wafer process chamber the lamps used to heat the wafer emit broadband, blackbody light. The intensity of the broadband, black-body lamp emission may be several orders of magnitude greater than the emission by the reactant by-product of the coating and the plasma being monitored to detect end-point. The broadband, black-body light is most intense during a ramp phase when the lamps are on at full power. The stray light from the lamps becomes a critical issue during ramp phase. Both stray light outside the spectrometer but associated with the optics leading into the spectrometer and stray light which makes its way into the spectrometer make monitoring of the reactant by-product to determine endpoint more difficult. Stray light enters the optics monitoring system, because dirt and coating films, deposited on the optics during the ashing process, and imperfections in the optics scatter stray light into the spectrometer. Additionally, the far wall reflects and emits stray light into the field of view of the optics. The light emitted by the volatile by-product is in part reflected and diffused out of the field of view of the spectrometer by the coating on the optics and optics imperfections. Unwanted lamp light is in part diverted into the field of view of the spectrometer by the coating and dirt on the optics and imperfections in the optics. Coating and dirt on the optics reduces the by-product light, which reaches the spectrometer input and scatters lamp light into the spectrometer input. The increase in lamp light and decrease in volatile by-product light which enter the spectrometer result in a reduced signal-to-noise or signal-to-background ratio, which degrades the performance of the optical detection system.
Spectrometers are designed to direct light into a particular path for each wavelength for measurement. To accomplish this, spectrometers include numerous internal surfaces, off which the light either reflects, refracts, or is transmitted. These internal surfaces have surface imperfections and become dirty causing a small percentage of the lamp light, which enters the spectrometer, to be diffusely scattered. When lamps are used to radiantly heat the wafer a small percent of the high intensity lamp light leads to a moderately high level of randomly scattered light inside the spectrometer. Some lamp light, although at wavelengths other than the wavelength of the light emitted by the by-product, enters wavelength channels of the spectrometer designed to measure light emitted by the volatile by-product. Stray light inside the spectrometer is background or noise, scattered off the imperfect interior surface of the spectrometer. The stray light in the spectrometer results in a reduced signal-to-noise or signal-to-background ratio.
There are special problems with downstream process chambers with regard to the number of available OES signals and accessing the volume from which maximum signal is produced by the ashing and etching reactions. Often, there is both broadband light from the radiant heating lamps and the plasma itself which varies during the ashing and the etching processes, which make it more difficult to measure the light emitted by the wafer. Special cases of low ash rate processes, resulting in weak signals, also make endpoint detection more difficult.
The design of a standard etcher is distinct from the design of a downstream asher in two respects. First, the pressures associated with etchers are much lower. The lower pressure of etchers allows molecular flow, where given molecules of the reactant by-product move freely around the process chamber and bounce off numerous chamber walls many times, creating a substantially uniform distribution of by-product within the reaction chamber. Secondly, the plasma fills the process chamber and remains in the process chamber. For these two reasons, there is often nearly uniform signal strength from the reactant by-product in etchers.
The pressure in a downstream plasma asher is increased. Applicants have observed a transitional region of laminar viscous flow slightly above the wafer surface due to the increase in pressure. The reactant by-product of the coating being removed and the plasma is contained near the wafer surface by the laminar flow. The by-product being monitored is formed on or near the surface of the wafer and is constrained in this region by the laminar flo

Affiliated with

Also associated with

Rating

Optimized optical system design for endpoint detection does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Optimized optical system design for endpoint detection, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Optimized optical system design for endpoint detection will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3102851