Optics: measuring and testing – By dispersed light spectroscopy – With sample excitation
Reexamination Certificate
2001-07-06
2003-07-15
Font, Frank G. (Department: 2877)
Optics: measuring and testing
By dispersed light spectroscopy
With sample excitation
C356S316000
Reexamination Certificate
active
06594010
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to an analyzer having a charge coupled device based emission spectrometer for ultra-high purity gas analysis.
BACKGROUND OF THE INVENTION
For many years, gaseous emission spectroscopy has been used for the analysis of nitrogen in argon (see e.g., U.S. Pat. No. 3,032,654). A commonly used emission source for this technique is a low-energy argon plasma, also known as a silent electric discharge (SED). This technology has improved over the years to lower the limit-of-detection (LOD) to single digit parts-per-billion (ppb) levels; for example, through the use of electro-optical modulation (see, e.g., U.S. Pat. No. 5,412,467). Further improvements in sample cell design, electronics, and the microprocessor platform have led to the current generation of spectroscopic analyzers. The current practice of using multiple detectors and optical filters allows for the simultaneous analysis of multiple impurities if suitable emission wavelengths can be found.
The block diagram in
FIG. 1A
shows the emission and detection systems utilized in connection with early analyzers that perform conventional emission spectroscopy. Similarly,
FIG. 1B
is a block diagram for a state-of-the-art analyzer design using electro-optic modulation, as described in U.S. Pat. No. 5,412,467. In both types of systems, a high voltage transformer
1
powers a light source
2
containing a gaseous sample to be analyzed. The gases are excited by the voltage to produce optical emission lines (an emission spectrum) characteristic of each gas (impurity) in the sample. Narrow bandpass optical filters
3
isolate the strongest emission line corresponding to each impurity. Photomultipliers (PMTs)
5
convert the light output from each impurity to a current which is amplified by a frequency selective amplifier, either a fixed amplifier
6
a
as in
FIG. 1A
or a tuned amplifier
6
b
as in
FIG. 1B
, and readout
7
. The conventional system uses a chopper wheel
4
to interrupt (or modulate) the light to the PMT. Whereas, the electro-optic modulation system uses a frequency doubler
8
and variable frequency oscillator
9
to modulate the light to the PMT.
To date, each generation of emission spectrometer has shared a common detection scheme. The emission line of the impurity of interest is isolated by a narrow bandpass optical filter and converted to an electrical signal through the use of a photomultiplier tube. The PMT has been the detector of choice for numerous applications in low light level spectroscopy due to the inherent high electronic gain possible through the use of the PMT. In addition to sensitivity, the PMT is also rugged, reliable, low cost, and stable over long periods of time. These are important attributes when used in a continuous-use application, such as emission spectroscopy. However, PMTs do pose several problems when used as detectors for emission spectroscopy. PMTs are comparatively large devices by today's standards, particularly when several PMTs must be used in a single analyzer. Although PMTs are low cost, the high-quality narrow bandpass filters are not, especially when several filters are needed. Moreover, the narrow bandpass filters, which isolate the emission line of interest for a given impurity, also prevent evaluation of the background light level at the wavelength chosen for analysis.
The background light level at the impurity emission wavelength of interest can change for a variety of reasons, such as changes in temperature, sample gas pressure, excitation conditions, or other impurities entering the discharge. It is extremely difficult to distinguish background light level shifts from a changing concentration of the impurity of interest when only the emission intensity at the wavelength of interest is known. Shifts in background light level can result in problems with long term baseline drift, nonlinear calibration curves, and cross sensitivity to other impurities. These are all serious problems when attempting to perform impurity analysis on impurities with measurements in the parts per billion.
FIG. 2
illustrates the problem inherent in using the PMT and optical filter approach.
FIG. 2
shows six emission spectra labeled A-F; that respectively correspond to 86, 56, 38, 25, 9 and 0 ppb concentrations of moisture (water vapor) in an argon sample gas. Each spectrum shows the region of the ultraviolet (UV) spectrum where both moisture and nitrogen have characteristic emission lines. Note that the addition of moisture causes a rise in the baseline light level, particularly in the region of the spectrum (333-360 nm) where nitrogen characteristically emits. If a PMT and optical filter are used, this increased light level could be interpreted as coming from a nitrogen impurity, resulting in an erroneously high concentration of nitrogen being reported. However, if the baseline light level shift is evaluated properly, the fact that no nitrogen emission peak is present can be correctly determined, and hence the nitrogen concentration is actually zero. The same argument applies to baseline shifts due to other factors, as mentioned above, which show up as noise and drift in the analytical results if not taken into account. Two approaches have been proposed to address the problem of changing baseline light level.
First, a separate PMT detector can be dedicated to determining the baseline emission light level rather than analyzing for an impurity. This is done by choosing a narrow bandpass filter that isolates a wavelength region of the sample gas emission spectrum close to, but not including, the impurity emission lines of interest. The analyzer then uses the ratio of the signal from the PMT measuring the impurity emission and the signal from the PMT monitoring the baseline. This approach eliminates many of the problems of the baseline emission light level. However, this technique is more complicated and requires either an additional PMT and optical filter or a reduction in the number of impurities which can be detected.
In the second approach, the baseline drift and some of the nonlinearity in the calibration curve of the analyzer are compensated for mathematically. The application of such a correction to each impurity analysis is implemented as part of the operating program of the analyzer. However, this approach is only possible if the nonlinearity is well characterized from previous experimental work.
The ability of a charge coupled device (CCD) array to easily evaluate the entire region of the spectrum of interest makes them an attractive detector choice for a number of spectroscopic methods. CCD arrays have been used in place of PMTs and narrow bandpass filters for spectroscopic applications for a number of years and small, low-cost, commercial units are available. The best known units are used for inductively coupled plasma (ICP) emission spectroscopy. These applications are well understood, but involve the use of very intense emission sources, typically ICP or microwave sources. These emission sources are far more intense and more energetic than the low-level emission sources in gas emission analyzers.
CCD arrays consist of an array of detector elements (pixels), each of which is a photodiode. However, CCDs lack the inherent high gain capability of a PMT. In this respect, the pixels act like photographic film. Low light images can be captured using longer integration times, much like a long exposure time is used with a conventional camera. However, long integration times worsen a problem inherent to CCD arrays; the so-called dark or thermal noise. If an array is left in complete darkness, it will generate a unique noise signature that is a function primarily of integration time and temperature. Managing this changing noise signature is key to using this technology when low intensity sources are to be detected.
Because of this dark noise problem, a brighter emission is needed from the impurity of interest in order to generate a useable signal from the CCD array detector. Heretofore, applications that normally use low light
Malczewski Mark Leonard
Martin Wayne Donald
Wegrzyn Joseph
Font Frank G.
Lauchman Layla
Praxair Technology Inc.
Schwartz Iurie A.
LandOfFree
Emission spectrometer having a charge coupled device detector does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Emission spectrometer having a charge coupled device detector, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Emission spectrometer having a charge coupled device detector will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3093387