Etching a substrate: processes – Gas phase etching of substrate – With measuring – testing – or inspecting
Reexamination Certificate
1999-04-12
2002-05-07
Gulakowski, Randy (Department: 1746)
Etching a substrate: processes
Gas phase etching of substrate
With measuring, testing, or inspecting
C216S059000, C438S014000, C438S016000
Reexamination Certificate
active
06383402
ABSTRACT:
FIELD OF THE INVENTION
The present invention generally relates to the field of plasma processes and, more particularly, to monitoring/evaluating such plasma processes.
BACKGROUND OF THE INVENTION
Plasma is used in various types of industrial-type processes in the semiconductor and printed wiring board industries, as well as in various other industries such as in the medical equipment and automotive industries. One common use of plasma is for etching away materials in an isolated or controlled environment. Various types of materials may be etched by one or more plasma compositions, including glasses, silicon or other substrate materials, organics such as photoresist, waxes, plastics, rubbers, biological agents, and vegetable matter, and metals such as copper, aluminum, titanium, tungsten, and gold. Plasma is also utilized for depositing materials such as organics and metals onto an appropriate surface by various techniques, such as via chemical vapor deposition. Sputtering operations may also utilize plasmas to generate ions which sputter away material from a source (e.g., metals, organics) and deposit these materials onto a target such as a substrate. Surface modification operations also use plasmas, including operations such as surface cleaning, surface activation, surface passivation, surface roughening, surface smoothing, micromachining, hardening, and patterning.
Plasma processing operations can have a significant effect on a company's profit margin. This is particularly true in the semiconductor and printed wiring board industries. Consider that a single semiconductor fabrication facility may have up to 200-300 processing chambers and that each processing chamber in commercial production may process at least about 15-20 wafers per hour. Further consider that an eight inch wafer which is processed in one of these chambers in some cases may be used to produce up to 1,500 semiconductor chips which are each worth at least about $125, and that each of these semiconductor chips are in effect “pre-sold.” Therefore, a single wafer which has undergone an abnormal plasma process and which is scrapped will result in lost revenues of at least about $187,500.
The particular plasma process which acts on the wafer such that a semiconductor device may be formed therefrom is commonly referred to as a plasma recipe. Some skilled in the art refer to a plasma recipe as being a combination of one or more plasma steps, each of which is executed for a fixed period of time. However, “plasma recipe” as used in relation to the present invention means a plasma processing protocol which includes one or more different and distinct plasma steps (e.g., a certain combination of certain steps). “Different and distinct” means that each plasma step produces a different, predetermined result on the product being processed (e.g., a wafer). Differences between plasma steps may be realized by changing one or more process conditions, including without limitation the composition of the plasma, the temperature and pressure in the processing chamber, DC bias, pumping speeds, and power settings. The sequence of the plasma steps, as well as the result of each plasma step, also produces a desired overall or cumulative end result for the plasma recipe.
Plasma processes may be run on wafers in a commercial production facility in the following manner. A cassette or boat which stores a plurality of wafers (e.g., 24) is provided to a location which may be accessed by a wafer handling system associated with one or more processing chambers. One wafer at a time is processed in the chamber, although some chambers may accommodate more than one wafer at a time for simultaneous plasma processing. One or more qualification wafers may be included in each cassette, and the rest are commonly referred to as production wafers. Both the qualification and production wafers are exposed to the same plasma process in the chamber. However, no semi-conductor s devices are formed from a qualification wafer as qualification wafers are processed and retained solely for testing/evaluating the plasma process, whereas semiconductor devices are formed from the production wafers. Further processing operations of these now plasma processed production wafers may be required before semiconductor devices are actually formed from such production wafers.
Monitoring is employed by many plasma processes to evaluate one or more aspects of the process. One common monitoring technique associated with plasma recipes run on wafers is endpoint detection. Current endpoint detection systems attempt to identify when a single plasma step of a given plasma recipe is complete, or more specifically that point in time when the predetermined result associated with the plasma step has been produced on the product. A representative “predetermined” result is when a layer of a multi-layered wafer has been completely removed in a manner defined by a mask or the like. Although prior art systems exist for attempting to identify the endpoint of a single step of a multiple step plasma recipe, no known system is able to identify the endpoint of each step of a multiple step plasma recipe, or even any two steps of a multiple step recipe for that matter.
Having the ability to terminate a given plasma step at its endpoint or just after endpoint is reached would reduce costs in a number of ways. Obviously, the amount of gases which are used to generate the plasma may be reduced by terminating a given plasma step when it has achieved its desired result. More importantly, terminating a given plasma step at or very shortly after its endpoint has been reached prevents the wafer from being over-etched to an undesired degree. Over-etching a wafer removes more material from the wafer than desired, such as by etching away portions of the layer immediately following that which was to be etched, and may also result in the undesirable sputtering of materials onto other portions of the wafer. The resulting effect on the semiconductor device(s) formed from this wafer may reduce the quality of the semiconductor device(s), may go undetected until the semiconductor device(s) has been delivered to the customer which would not be desirable if the device(s) was defective or deficient in any way, or both. Finally, a certain degree of over-etching of a wafer may result in the wafer simply being scrapped.
Endpoint detection is desirable in theory for plasma processes. Certain deficiencies became evident as attempts were made to implement endpoint detection techniques in commercial fabrication facilities. Initially, all known endpoint detection techniques were developed by first chemically analyzing the subject plasma. operation to identify a wavelength to key in on as being indicative of endpoint. Fabrication facilities typically run a multiplicity of plasma recipes. As such, these known endpoint detection techniques increase costs due to the required retention of an experienced chemist. Moreover, these techniques often do not produce the intended result - that is the wavelength which is selected by the chemist may in fact not be at all indicative of endpoint when the plasma step is actually run since it is only “theory” based. A given endpoint detection technique may also be dependent upon the processing chamber on which the technique was developed. Accurate results may not be realized when the endpoint detection technique is used on other processing chambers. Therefore, it would be desirable to have a plasma monitoring system in which the amount of chemical “pre-analysis” is reduced and which would allow the plasma monitoring system to work to an acceptable degree on multiple processing chambers (i.e., a generic plasma monitoring system which was able to identify the relevant endpoint).
Commonly used endpoint detection techniques provide no information on how the plasma process has actually proceeded or the “health” of the plasma process—only if and when an endpoint of the subject plasma step has been reached. Other monitoring techniques which are commonly used in plasma processes suffer from this sa
Smith, Jr. Michael Lane
Stevenson Joel O'Don
Ward Pamela Denise Peardon
Ahmed Shamim
Gulakowski Randy
Marsh & Fischmann & Breyfogle LLP
Sandia Corporation
LandOfFree
Method and apparatus for monitoring plasma processing... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for monitoring plasma processing..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for monitoring plasma processing... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2910634