Optics: measuring and testing – By dispersed light spectroscopy – Utilizing a spectrometer
Reexamination Certificate
2002-04-15
2004-03-02
Evans, F. L. (Department: 2877)
Optics: measuring and testing
By dispersed light spectroscopy
Utilizing a spectrometer
C702S085000, C250S252100
Reexamination Certificate
active
06700661
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to the fields of spectroscopy and of calibrating a spectroscopic device, such as a spectrometer for predicting analyte levels. More particularly, the present invention relates to a method of optimizing wavelength calibration to facilitate the transfer of a calibration model from a primary device to a secondary device.
BACKGROUND OF THE INVENTION
Spectroscopy is based on the analysis of how incident radiation interacts with the vibrational and rotational states of molecules, often within a matrix such as blood or living tissue, which are of analytical interest. Spectrometers and other similar devices that use spectroscopic measurement techniques have gained increased popularity because of the ability to provide fast and non-invasive measurements of concentrations of different chemicals or analytes. In particular, spectrophotometry is a type of spectroscopy commonly used to quantitatively measure properties such as analyte concentration(s) based on spectral energy distribution in the absorption spectrum of a sample solution or medium. In spectrophotometry, the energy distribution is typically analyzed within a range of the visible, ultraviolet, infrared, or near-infrared spectra. For example, near-infrared radiation (NIR) is electromagnetic radiation having a wavelength between about 750 and 2500 nanometers (nm). Near-infrared spectrophotometry generally uses instruments with quartz prisms in monochromators and with lead sulfide photoconductor cells or photodiodes as detectors to observe absorption bands. Near-infrared spectrophotometry is, for example, increasingly being used to measure in vivo analytes such as glucose, fructose, glycerol, and ethanol.
Spectroscopic devices are well known in the art and are described in detail, for instance, in U.S. Pat. Nos. 5,361,758 and 5,771,094, the contents of which are incorporated herein by virtue of this reference. In general, a typical spectrometer system includes a light source which is projected through the sample to be examined, a sample interface mechanism, a spectrometer to separate the light into its component wavelengths, a detector, amplification electronics, and a microprocessor or computer system. By measuring the loss (absorption), between the source and the detector and applying appropriate chemometric or mathematical techniques, it is possible to determine the chemical analytes being examined since different chemicals absorb different amounts of light. The detector or photodetector generally includes a photodiode array of pixels enabling the detector to simultaneously detect the intensities of a number of different spectral components at distinct wavelengths. The intensities at these distinct wavelengths can be used to predict, in turn, the quantities or concentrations of the analyte(s) of interest.
Calibration of spectrometers and of analytical instruments in general is necessary to ensure the accuracy of measurements performed by such devices. In essence, calibration is the development of a model or algorithm that predicts the properties (e.g. analyte concentrations) of a sample from the spectrometer's response. To calibrate a spectrometer, the spectral response of several calibration samples or standards having known concentrations of an analyte of interest is measured. By combining the known concentration data with the measured spectral response data, a calibration model (i.e. a mathematical relationship) can be developed using a “best fit” regression technique (e.g. partial least squares or PLS) between the spectral measurements and analyte or property of interest. The calibration model or algorithm is then stored in a non-volatile memory, such as for example, in a microprocessor system of the device. In most cases, the spectrometer's response is a measure of a number of variables, e.g. a number of different chemical species present in a sample, and so calibration is based on a multivariate calibration model. The spectrometer and its calibration model can then be used to estimate the property or properties (e.g. analyte concentrations) of an unknown sample. By detecting the pixel position of a spectral component measured by the spectrometer, the wavelength of that spectral component is known, and consequently a prediction can be calculated by the microprocessor using the calibration algorithm.
Calibration, i.e. the development and calculation of a calibration algorithm, is generally performed on a primary instrument or device when it is initialized or installed or when any of its components are replaced. This primary instrument is often a member of a group of similar instruments produced by the same manufacturer, and having the same component types, model number, and so on. The other members of this instrument group are hereinafter referred to as “secondary” or “target” instruments. Because calibration is a lengthy and involved process, it is often not practical to individually recalibrate each secondary instrument in a set since this may require, among other difficulties, a large number of calibration samples at the site of each secondary instrument. Instead, for many spectroscopic applications, the calibration model developed for the primary instrument is transferred to each secondary instrument. For example, the transfer of calibration algorithms between primary and secondary instruments is often desirable with NIR spectrophotometers.
However, when a calibration model determined using measurements on a primary instrument is transferred to another instrument of the same type, a loss of accuracy generally occurs. This loss of accuracy is due to the innate differences that exist between any two physical devices, resulting in a variation in spectral responses (i.e. the correlation between pixel locations and wavelength) and affecting the reproducibility of measurements with the secondary device.
Wavelength calibration is performed to reduce the inaccuracy inherent in calibration algorithm transfer between instruments. For this purpose, calibration light source(s) having spectral lines at two or more known wavelengths are typically used to provide wavelength calibration parameters. A calibration light source may be a laser or a mercury lamp, for example. The pixel locations of the spectral lines of the calibration light source are more accurately fitted to the known wavelengths to improve the calibration reference. The calibration model or algorithm can then be adjusted to assign a more accurate wavelength value to each pixel of the multi-pixel detector array.
However, when using a calibration light source such as a laser to locate a selected wavelength on the pixel array of a spectrometer, wavelength calibration to within ±0.2 nm is typically very difficult to achieve, and more accurate wavelength calibration is generally not possible. For many spectroscopic applications, this level of wavelength inaccuracy remains unacceptable, since it can introduce significant errors in analyte concentration predictions or estimations, particularly for measurements performed within complex media such as living issue or blood.
Similarly, it may also be desirable to transfer a calibration developed for a primary instrument to a target instrument, where the target instrument is the primary instrument after a period of time has elapsed. During that time, the primary instrument's response may have changed due to detector instability, temperature variations, drift in the electronics of the primary instrument, or other causes. Consequently, the primary instrument may require subsequent wavelength recalibration to avoid significant inaccuracies.
Thus, wavelength calibration of a pixel (e.g. photodiode) array spectrometer and adjustment of the calibration model to minimize differences between instruments is highly desirable, and there is a need for an improved method of wavelength calibration to facilitate and improve the transfer of a calibration algorithm or model from a primary instrument to a secondary instrument (which may include the primary instrume
Cadell Theodore E
Lu Geng
CME Telemetrix Inc.
Evans F. L.
LandOfFree
Method of optimizing wavelength calibration 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 of optimizing wavelength calibration, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method of optimizing wavelength calibration will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3210859