WikiPatents - Community Patent Review
Create Free Account  |  License or Sell Your Patent  |  WikiPatents Marketplace  |  WikiPatents Blog
Username:  Password:  
    
Advanced Search
Plasma constituent analysis by interferometric techniques    
United States Patent5225888   
Link to this pagehttp://www.wikipatents.com/5225888.html
Inventor(s)Selwyn; Gary S. (Hopewell Junction, NY); Walkup; Robert E. (Ossining, NY)
AbstractAn interferometer (18 or 40) is used to identify trace constituents in a plasma during processing semiconductor devices such as transistors. Light emissions collected from the processing chamber (10) are collimated by lens (14) and transmitted to the interferometer (18 or 40) which selectively allows therethrough particular wavelengths of light which are characteristic of the excitation emissions of certain atoms such as sodium and copper. The light intensity at the selected wavelengths is sensed by a photomultiplier tube (30). In one embodiment, the interferometer (18) is a Fabry-Perot type interferometer and the separation of the plates (20 and 22) which form the Fabry-Perot etalon is controlled using a piezoelectric driver (26). A signal processor (34) correlates the sensed light emissions from the photomultiplier tube (30) with the selected wavelength that is determined by the piezoelectric driver (26). In another embodiment, the interferometer (40) is a narrow bandpass interferometric filter which is tiltable with respect to the collimated incident light from the processing chamber (10). Tilting a narrow bandpass interferometric filter (42) with respect to incident light changes the path length through the filter (42) and allows for the selective transmission of certain wavelengths of light. By rapidly tilting the narrow bandpass interferometric filter (42) at a rate between 5-300 Hz with respect to the incident light, a narrow range of wavelengths on the order of 3 nm can be scanned.
   














 Title Information Submit all comments and votes
 
Patent Text Patent PDF Print Page Summary File History
Plain text PDF images Print Summary File History
Drawing from US Patent 5225888
Plasma constituent analysis by interferometric techniques - US Patent 5225888 Drawing
Plasma constituent analysis by interferometric techniques
Inventor     Selwyn; Gary S. (Hopewell Junction, NY); Walkup; Robert E. (Ossining, NY)
Owner/Assignee     International Business Machines Corporation (Armonk, NY)
Patent assignment
All assignments
Publication Date     July 6, 1993
Application Number     07/633,811
PAIR File History     Application Data   Transaction History
Image File Wrapper   Patent Term   Fees
Litigation
Filing Date     December 26, 1990
US Classification     356/454
Int'l Classification     G01B 009/02
Examiner     Turner; Samuel A.
Assistant Examiner    
Attorney/Law Firm     Whitham & Marhoefer
Address
Parent Case    
Priority Data    
USPTO Field of Search     356/346 356/352 356/351 250/339 250/343 250/227.29 359/578 156/626
Patent Tags     plasma constituent analysis interferometric techniques
   
Enter a comma (,) or semicolon (;) between multiple tag words/phrases.
Describe this patent:
 Amusing   
 Clever   
 Complex   
 Efficient   
 Historic   
 Important   
 Innovative   
 Interesting   
 Practical   
 Simple   
[no votes]
Patent WIKI

Share information and news about this patent, including information and news about the technology, inventors, company, ligation and licensing.

 References Submit all comments and votes
 
*references marked with an asterisk below are user-added references
 U.S. References
 
Add a new US reference:  
ReferenceRelevancyCommentsReferenceRelevancyComments
3740144



[0 after 0 votes]
3914055



[0 after 0 votes]
5059784
Northrup
250/237R
Oct,1991

[0 after 0 votes]
4999013
Zoechbauer
356/454
Mar,1991

[0 after 0 votes]
4838694
Betz

Jun,1989

[0 after 0 votes]
4758304
McNeil
216/60
Jul,1988

[0 after 0 votes]
4743114
Crane, Jr.
356/454
May,1988

[0 after 0 votes]
4735507
Crane, Jr.
356/491
Apr,1988

[0 after 0 votes]
4732480
Fortunato
356/453
Mar,1988

[0 after 0 votes]
4717446
Nagy
438/14
Jan,1988

[0 after 0 votes]
4711573
Wijntjes
356/452
Dec,1987

[0 after 0 votes]
4680084
Heimann
216/60
Jul,1987

[0 after 0 votes]
4675072
Bennett
216/60
Jun,1987

[0 after 0 votes]
4602981
Chen
438/17
Jul,1986

[0 after 0 votes]
4525067
Hernandez
356/454
Jun,1985

[0 after 0 votes]
4482248
Papuchon
356/454
Nov,1984

[0 after 0 votes]
4454001
Sternheim
216/60
Jun,1984

[0 after 0 votes]
4448486
Evans
359/559
May,1984

[0 after 0 votes]
4204771
Shull
356/454
May,1980

[0 after 0 votes]
4092070
Smithline
356/454
May,1978

[0 after 0 votes]
4005937
Barrett
356/454
Dec,1969

[0 after 0 votes]
4035643
Barrett
250/339.13
Dec,1969

[0 after 0 votes]
 Foreign References
 Other References
 Market Review Submit all comments and votes
   
Market Size
Estimate the gross annual revenues of the relevant market sector:
> $10B
$5B - $10B
$2B - $5B
$500M - $2B
$100M - $500M
$10M - $100M
$1M - $10M
$500K - $1M
$100K - $500K
< $100K
[No votes]
$0
 
$0   $2.5B   $5B   $7.5B   $10B
Market Share
Estimate the percentage of the relevant market sector this invention will capture:
75% - 100%
50% - 74.99%
25% - 49.99%
10 - 24.99%
5 - 9.99%
2 - 4.99%
1 - 1.99%
< 1%
[No votes]
0.0%
 
0%   25%   50%   75%   100%
Reasonable Royalty
What percentage of gross sales should the inventor or assignee be paid?
75% - 100%
50% - 74.99%
25% - 49.99%
10 - 24.99%
5 - 9.99%
2 - 4.99%
1 - 1.99%
< 1%
[No votes]
0.0%
 
0%   25%   50%   75%   100%
Public's "Guesstimation" of Royalty Value
Market SizeN/A[No votes]
xMarket ShareN/A[No votes]
xReasonable RoyaltyN/A[No votes]

N/A

License Availablity
If you are NOT the owner or assignee, answer here:
Yes, license is available for purchase

No, license is not currently available



[No votes]
License Availablity
If you ARE the owner or assignee, answer here:
Yes, license is available for purchase

No, license is not currently available



[No votes]
Competitive Advantage
Does this invention have a significant competitive advantage over similar technologies?
Yes

No



[No votes]
Most helpful competitive advantage comment
[No comments]

Commercial Alternatives
Are there viable commercial alternatives for this invention?
Yes

No



[No votes]
Most helpful commercial alternative comment
[No comments]

 Technical Review Submit all comments and votes
 Claims Submit all comments and votes
 


Having thus described our invention, what we claim as new and desire to secure by Letters Patent is as follows:

1. A system for monitoring a plasma during solid state device manufacturing, comprising:

means for collecting light from a plasma reactor;

means for collimating said light from said plasma reactor to form a collimated beam of light;

an interferometer positioned in the path of said collimated beam of light, said interferometer including a selection means for selectively permitting only certain wavelengths of light in said collimated beam of light to pass through said interferometer and a tiltable bandpass interferometer filter, said selection means includes a means for driving said tiltable bandpass interferometer filter to first and second angular orientations relative to said collimated beam of light, said first and second angular orientations of said bandpass interferometric filter respectively defining a first and second wavelengths of light which pass through the interferometer, said means for driving modulates said tiltable bandpass interferometric filter between said first and said second angular orientations on a periodic cycle, said means for driving includes an audio speaker which emits acoustic waves, said acoustic waves being selectively transmitted from said audio speaker according to the output from an audio driver;

means for collecting light in said certain wavelengths which pass through said interferometer;

means for sensing an intensity of said collected light in said certain wavelengths; and

means for correlating said means for sensing with said selection means.

2. A system for monitoring a plasma during solid state device manufacturing, comprising:

means for collecting light from a plasma reactor;

means for collimating said light from said plasma reactor to form a collimated beam of light;

an interferometer positioned in the path of said collimated beam of light, said interferometer including a selection means for selectively permitting only certain wavelengths of light in said collimated beam of light to pass through said interferometer and a tiltable bandpass interferometer filter, said selection means includes a means for driving said tiltable bandpass interferometer filter to first and second angular orientations relative to said collimated beam of light, said first and second angular orientations of said bandpass interferometric filter respectively defining a first and second wavelengths of light which pass through the interferometer, said means for driving modulates said tiltable bandpass interferometric filter between said first and said second angular orientations on a periodic cycle, said means for driving including a means for dampening the movement of said tiltable bandpass interferometric filter;

means for collecting light in said certain wavelengths which pass through said interferometer;

means for sensing an intensity of said collected light in said certain wavelengths; and

means for correlating said means for sensing with said selection means.

3. A system for monitoring a plasma during solid state device manufacturing, comprising:

means for collecting light from a plasma reactor;

means for collimating said light from said plasma reactor to form a collimated beam of light;

an interferometer positioned in the path of said collimated beam of light, said interferometer including a selection means for selectively permitting only certain wavelengths of light in said collimated beam of light to pass through said interferometer;

means for collecting light in said certain wavelengths which pass through said interferometer;

means for sensing an intensity of said collected light in said certain wavelengths;

means for correlating said means for sensing with said selection means; and

a means for subtracting background emissions in said plasma reactor sensed by said means for sensing from said intensity of said collected light in said certain wavelengths.

4. A method for monitoring trace constituents during semiconductor processing, comprising the steps of:

collecting light from a plasma reactor;

collimating the light collected from said plasma reactor;

directing the collimated light towards an interferometer;

controlling said interferometer to selectively permit certain wavelengths of light to pass through said interferometer;

collecting the light at the certain wavelengths which passes through said interferometer;

sensing the intensity of light collected at certain wavelengths;

correlating the sensed intensity with said certain wavelengths of light which pass through said interferometer; and

subtracting background emissions from said light collected at said certain wavelengths.

5. A method as recited in claim 4 wherein said step of subtracting includes the step of comparing a peak height of said light collected at said certain wavelengths with a baseline emission.

6. A method for monitoring trace constituents during semiconductor processing, comprising the steps of:

collecting light from a plasma reactor;

collimating the light collected from said plasma reactor;

directing the collimated light towards an interferometer;

controlling said interferometer to selectively permit certain wavelengths of light to pass through said interferometer;

collecting the light at the certain wavelengths which passes through said interferometer;

sensing the intensity of light collected at certain wavelengths, said sensing step includes the steps of sensing a first emission intensity at a first wavelength where a particular atomic emission should occur and sensing a second emission intensity at a second wavelength where said particular atomic emission does not occur, and subtracting said second emission intensity from said first emission intensity;

correlating the sensed intensity with said certain wavelengths of light which pass through said interferometer.
 Description Submit all comments and votes
 


BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention generally relates to solid state device manufacturing and, more particularly, to an interferometric optical emission detection system used for trace constituent contamination monitoring and endpoint detection.

2. Description of the Prior Art

Sodium (Na) and other alkali metals are unwanted impurities which adversely affect the operation of transistors and other solid state devices. Sodium is a mobile ion and will cause the gate of a transistor to turn on and off at different voltages. Variable operation of the gate will, in turn, cause inconsistent operation of other devices such as memory chips and the like. Sodium contamination of transistors can result from human contact with chemicals and tools used in device manufacturing (i.e., sodium ions can be transported from a person's body onto a processing tool simply by the person touching the tool), impurities present in processing reagents such as photoresist, from the manufacture and storage of these processing reagents, by the sodium which is naturally present in the atmosphere corrupting the processing plant environment (i.e., fog, mist, and rainwater seepage can carry sodium ions from the atmosphere into a processing plant), or by other means. Field Effect Transistors (FETs) are four orders of magnitude more sensitive to alkali metals than bipolars; therefore, sodium contamination is a particular concern in FET processing.

Sodium emission, as well as other trace constituent emissions, can be detected optically. Most of today's optical emission detecting systems used in transistor processing can be categorized as either monochromators with photodiode array detectors or as scanning monochromators with a photomultiplier detector. These types of optical emission systems are available from the following companies: Tracor-Northern, Plasma-Therm Analytical, Xenix, and EG&G PAR. One problem with these types of optical emission systems is that they offer low resolution detection; often greater than 1 nm and sometimes as poor as 10-20 nm. Low resolution detection results in insufficient wavelength dispersion and convolution of the desired emission signal with unwanted background and other emissive interferences. Because of this, low resolution detection also provides low sensitivity for trace species detection against the intense background emission of plasma processing tools. This is further compounded by low light throughput which is typical of conventional monochromators.

As a result of the low sensitivity of conventional optical emission systems, the usual method used for detecting sodium impurities in FETs is by electrical measurement at the end of the chip fabrication process. Clearly, it would be advantageous to have some means which is sensitive enough to detect trace levels of an impurity such as sodium ion before a large number of chips have been damaged. Ideally, potential processing problems could be identified when unacceptable levels of sodium are detected during the fabrication of FETs, and the processing could be halted temporarily to clean the tools and/or check the sodium levels in the processing reagents.

Proper etching endpoint detection is also a major concern in transistor manufacturing. One prior art method of determining the endpoint of material removal is simply timing the etch process according to the rate of material removal. To use a timed etch procedure, it is necessary to empirically determine the time at which all of the desired material has been removed, but the underlying layer has not been etched significantly. Timing is not an ideal procedure for determining the proper etching endpoint because it is indirect and the consequences of improper etching are significant. Underetching will cause a degradation of the gain of the transistor, while overetching will result in a degraded contact between the intrinsic and extrinsic base regions.

Optical emission detection has been proposed as a means for detecting the endpoint at which an etching operation should be halted. A monochromator or bandpass filter can be used to select wavelengths of light at which a desired optical emission will occur. High spectral resolution is always preferable to low spectral resolution for monitoring and analysis of plasma species because, with sufficiently high resolution, it is possible to minimize interference from other emitting plasma species. Moreover, if light throughput is held constant, peak height will increase relative to the background as spectral resolution is improved. Accordingly, high spectral resolution can provide considerable advantages in sensitivity for etch endpoint detection and in the identification of weakly emitting plasma species such as in sodium ion contaminant analysis. Unfortunately, the usual methods for obtaining high spectral resolution, e.g., narrowing monochromator slits or utilizing larger monochromators, have practical limits for use in transistor manufacturing applications. When the slits of a monochromator are narrowed, higher resolution is achieved; however, the light throughput is significantly reduced and, thus, the overall sensitivity is reduced. Using larger monochromators is unacceptable because they would require costly clean room space and are generally impractical for manufacturing applications.

Laser-induced fluorescence (LIF) is a newer technique used in transistor manufacturing for identifying trench etch endpoints and can be used to identify trace constituents. For example, U.S. Pat. 4,675,072 to Bennett et al. discloses an LIF system used to detect and control the reactive ion etch (RIE) through of a given layer in a wafer by detecting a large change in the concentration of a selected minor species from the wafer in the etching plasma. Although LIF is generally used for monitoring major species, LIF may be used to detect trace Na ion contaminants during etching; however, Na only fluoresces at the excitation wavelength of 588.996 nm or 589.593 nm and since scattered laser light is a severe limitation on the sensitivity of LIF, the presence of Na may not be easily detected because of unavoidable interference from scattered laser light off the tool walls and windows. Furthermore, methods to reduce scattered laser light, such as the use of Brewster angle windows and light baffles, are impractical for use on commercial etch tools. LIF may also be used to detect trace copper atoms which appear in aluminized lines as a means to detect an etch endpoint. However, copper atoms will exhibit fluorescence at 324.775 nm which is a wavelength that would require frequency doubling of the dye laser; a process generally regarded as infeasible for manufacturing applications because more powerful lasers are required for this task and a skilled laser operator would be required.

Several laser interferometer techniques are now in standard practice for optically detecting line widths and etch endpoints. U.S. Pat. No. 4,454,001 to Sternheim et al. and U.S. Pat. No. 4,680,084 to Heimann et al. are directed to etch monitoring using laser interferometric methods whereby the thickness of the region being etched is simultaneously monitored. U.S. Pat. No. 4,838,694 to Betz et al. discloses a laser interferometry process which uses the reflected laser beam. U.S. Pat. No. 4,717,446 to Nagy et al. discloses a method, using a monitor wafer which is correlated to the endpoints of the working wafer, for detecting the endpoint of the etch of an epitaxially grown silicon whereby a laser is used to measure the etch rate of the monitor wafer by measuring the reflected light off the oxide layer. U.S. Pat. No. 4,602,981 to Chen et al. is directed an impedance monitoring technique for plasma etching wherein endpoints are detected by an impedance change of the plasma; however, the Chen et al. reference does disclose that laser interferometry is well established in the art and points out that the laser measures the thickness of the film removed as the etch process proceeds. U.S. Pat. No. 4,758,304 to McNeil et al. discloses an apparatus for ion etching which utilizes an interferometer for surface monitoring. The methods described in Chen et al. and McNeil et al. are generally insensitive to the precise wavelength of the detected light. Before the invention thereof by the applicants, interferometers have not been used in solid state device manufacturing for detecting trace emitting species in a complex background spectra.

Furthermore, several techniques for controlling interferometers are now in common practice. U.S. Pat. No. 4,482,248 to Papuchon et al. discloses an interferometer used for optical filtering. U.S. Pat. No. 4,711,573 to Wijntjes et al. discloses a dynamic mirror alignment control which utilizes an interferogram for analyzing sample materials wherein a closed loop servo motor is used to maintain position orientation. U.S. Pat. No. 4,448,486 to Evans discloses the use of mirrors in a Fabry-Perot interferometer used to change the bandwidth of the Fabry-Perot interferometer. The applicants invention describes completely new methods for using and controlling interferometers in transistor manufacturing.

SUMMARY OF THE INVENTION

It is therefore an object of the present invention to provide methods of using light wave interferometry for the detection of trace plasma species.

It is another object of this invention to overcome the low light throughput and corresponding low resolution disadvantages of using a monochromator for wavelength selection to monitor optical emissions of trace plasma species.

It is yet another object of this invention to describe a new, low-cost, durable optical emission detection system which operates using interferometric principles.

According to the invention, light wave interferometry is used in solid state device manufacturing processes to identify trace plasma species. By using light wave interferometry, high resolution optical emission detection is achieved without the use of a monochromator. The light wave interferometry methods of the present invention result in a high light throughput efficiency, high spectral resolution, and an associated improvement in detection sensitivity and selectivity. The light wave interferometry methods of the present invention are useful in etch endpoint analysis and in the identification of important trace contaminants such as sodium impurities encountered during FET fabrication.

One method involves the use of a Fabry-Perot interferometer and relies upon the constructive and destructive interferences of entering light between two highly reflective parallel plates separated by a known distance of air. As discussed by R. P. Feynman et al. in The Feynman Lectures on Physics, Vol. III, Addison-Wesley, Reading, Mass., pp. 4-10, 1965, an interferometer operates on the principle of maintaining a standing wave between two reflective plates. The standing wave criterion will be met when, k.sub.f =j.pi./L, where k.sub.f is the wavenumber (k=2.pi./.lambda.) of the jth mode (j=any integer) and L is the separation of plates. By varying the separation between the two plates, which can be accomplished with a piezoelectric drive or similar means, the transmitted light may be tuned with very high resolution (e.g., 0.001 to 0.08 nm) over a narrow wavelength range. Such a system can typically transmit 70% of the light at the selected wavelength which compares very favorably with monochromator systems that have lower resolution and transmit only 5-10% of the light. Since multiple orders are also transmitted by the air-spaced etalon, it is also necessary to use prefilters to block out multiple order transmission.

Air-spaced Fabry-Perot interferometers are commercially available and have been primarily used in transistor and other solid state device manufacturing for linewidth measurements of major emitting species. For example, R. Walkup et al. disclosed the use of air spaced Fabry-Perot interferometers for linewidth measurements in J. Chem. Phys., 84, 2668 (1986). In contrast to prior uses of Fabry-Perot interferometers in solid state device manufacturing, the present invention anticipates the use of a Fabry-Perot etalon for end point detection and for the selective detection of trace emitting species from a complex background spectra. The present invention's use of the Fabry-Perot interferometer takes advantage of the high light throughput which occurs with the Fabry-Perot interferometer which is not present in other means of optical emission detection (e.g., monochromators). High light throughput provides no major advantage in line width analysis because a major plasma species, as opposed to a trace species, is usually being monitored during line width analysis.

Another method by which light wave interferometry can be used to monitor trace species during transistor manufacturing is to use a tiltable narrow bandpass interference filter to scan a small range of wavelengths. Narrow bandpass interference filters are commercially available from Oriel Optics, Andover Corp., Spindler, and Hoya and include a series of thin dielectric layers which are used to produce constructive or destructive interference of the incident light. Narrow bandpass interference filters are similar to the air spaced Fabry-Perot etalon in principle; however, the separation between layers in a narrow bandpass interference filter is fixed and several dielectric layers are used instead of air. While the resolution of a narrow bandpass interference filter is typically somewhat lower than with a Fabry-Perot interferometer, multiple order transmission problems associated with Fabry-Perot interferometers are not present when a narrow bandpass interference filter is used. This is because a Fabry-Perot interferometer operates in a single reflective cavity and, narrow bandpass interference filters operate by multiple reflections between successive dielectric layers. In this way, sensitivity to mode effects is minimized. Narrow bandpass interference filters have been widely used in materials processing plasmas for endpoint detection. In fact, some commercial reactors come equipped with specific bandpass interference filters, such as the Zylin metal etch reactor which is equipped with a 10 nm bandpass filter for aluminum (Al) atom detection. Harshberger et al. have described a reactor equipped with a bandpass interference filter in J. Electronic Materials, 7, 429 (1978). Interference filters are typically used in the same way as low resolution optical spectrometers for etch endpoint detection. The high light throughput of these filters provides an advantage over a fixed wavelength monochromator, and also has a lower capital cost and greater reliability than a monochromator. However, the fixed wavelength selection of the filter makes these light detection systems vulnerable to errors resulting from unwanted detection of interferences within the transmission range of the filter. The absolute emission intensity measured by the bandpass filter also includes intensity from the underlying continuous emission of the plasma. This continuous emission can be significant and can dominate over weak, wavelength-specific emission from trace or impurity plasma species. The continuum can result from electron-ion recombination or changes in the electron energy distribution in the plasma. Fixed wavelength detection using filters means that the desired emission may be indistinguishable from other emissions in the plasma.

It is known that bandpass interferometric filters can be tuned to shorter wavelengths by tilting the filter with respect to the incident light. Tilting the filter effectively changes the path length of the light between the dielectric layers without changing the layer separation or composition (as pointed out above, the distance between layers is fixed). Hence, the transmission peak of a given bandpass interferometric filter can be tuned slightly to shorter wavelengths by tilting the filter to an appropriate orientation. FIGS. 1 and 2, which are taken from the Oriel Optics catalog of 1986, illustrate the effect of varying the angle of incidence of a collimated beam on bandpass interferometric filter. FIG. 1 shows that if the angle of incidence is increased, the peak wavelength is shifted to shorter wavelengths. This shift is due to the fact that the path difference between the direct transmitted beam and the beam formed by the multiple reflection decreases. The angle shift in peak wavelength can be estimated for angles less than thirty degrees using equation 1: ##EQU1## where: a is the angle of incidence,

.lambda..sub.a is the peak wavelength at angle a,

.lambda..sub.o is the peak wavelength at normal incidence, and

N.sub.e is the incidence angle shift factor or the "effective index of refraction" of the filter.

N.sub.e is not the actual refractive index of the spacer layer, rather it is dependent on both the high and low index materials in the filter. Oriel Optics experimentally determined that N.sub.e in a zinc sulfide/cryolite system is 1.45 for a cryolite spacer layer and 2.0 for a zinc sulfide spacer layer. FIG. 2 plots the approximate wavelength shift versus the angle of incidence, as calculated according to equation 1, for two bandpass interferometric filters which each include a sulfide/cryolite system and where one uses a cryolite spacer layer and the other uses a zinc sulfide spacer layer. For small angles (e.g., under 30 degrees), the shape of the band pass does not change appreciably except for a small reduction in transmission. However, at larger angles band pass shape becomes highly distorted. For 5 nm and 10 nm filters, shifting by about 10 nm without drastic change in the band pass shape can be achieved.

Narrow and very narrow bandpass interference filters are a subset of bandpass interferometric filters and include filters having a full-width half maximum (FHWM) ranging between 0.1 nm and 1 nm. The distinctions between bandpass interferometric filters, narrow bandpass filters, and very narrow bandpass interference filters are qualitative. The FHWM refers to the error in wavelength detection measured at one half the peak height. Narrow bandpass interference filters are available from the Andover Corp. of Massachusetts as well as other vendors. Narrow bandpass interference filters show a critical dependence on filter orientation to collimated incident light. For example, if a very narrow bandpass interference filter having an FHWM=0.2 nm varied by 5.degree. from normal incidence, an approximately 1 nm shift in peak wavelength will occur without significant broadening of the transmission curve. Hence, unlike 5 nm and 10 nm bandpass interferometric filters which shift about 10 nm without drastic change in band pass shape, very narrow bandpass interference filters have considerably less range, but have improved spectral resolution over that range.

The present invention uses the "problem" of angular dependent wavelength shifts with narrow bandpass filters as a means to greatly improve optical emission detection sensitivity. Specifically, the invention contemplates mounting a narrow bandpass filter in a holder which can be rapidly tilted with respect to the collimated light collected from a plasma reactor. By tilting the bandpass filter 5.degree., an approximate range of 1 nm can be scanned with a bandpass of better than 0.2 nm at the half height or 0.05 nm at 90% transmission maximum. FIG. 3 shows a plot of the transmission of the filter versus wavelength where the bandpass is 0.2 nm at the half height and 0.05 nm at 90% transmission and shows how the plot shifts as the narrow bandpass filter is tilted. A photomultiplier tube is used to detect the transmitted light through the tilting narrow bandpass filter. The output from the photomultiplier tube is sent to an oscilloscope or lock-in amplifier which is synchronized with filter movement. Therefore, the lock-in amplifier can sensitively measure a change in the photomultiplier signal as a function of filter orientation.

Since atomic emission lines are extremely sharp (e.g., on the order of 0.0001 nm) any atomic emission detected will show a very strong signal dependence on filter orientation. This dependency on filter orientation provides the advantage of effectively eliminating background emission noise (i.e., background emission is considered to be the emission which is relatively constant over the range of wavelengths scanned). Eliminating background emission enables the detection of v