Method for correcting amplitude variations in spectrometers

Abstract

A method of correcting amplitude variations in the output of a spectrometric instrument (300) due to variations in the optical path length (L) through a sample holder (600), the method comprising: exposing a sample in a sample holder (600) to electromagnetic radiation at a plurality of wavenumbers; detecting electromagnetic absorption intensities in the sample at the plurality of wavenumbers; providing the detected absorption intensities indexed for the wavenumbers to a computer device (510) as spectral data; and applying a mathematical transformation (Icorr) to the spectral data in the computer device (510) to correct for amplitude variations in the output of the spectrometric instrument (300), and correcting for amplitude variations by determining a difference (Δ(SB)) between absorbance values at a first derivative of spectral data from a sample of zero material (SBZl) at two different wavenumber ranges. Z ) / ) to calculate the mathematical transformation; and calculating as the difference (Δ(SB Z ) / ) is the mathematical transformation of the inverse correlation function (Icorr).

Description

[0001] The present invention relates to a method of compensating for amplitude variations in the output of a spectrometer of the type used to generate spectral data from an unknown sample held in a sample holder, and in particular to compensating for amplitude variations due to variations in the optical path length through the sample holder.

[0002] In a typical spectrometer for generating spectral data from an unknown sample, a light emitter and a light detector are configured to define a light path in which the sample in question is positioned so that the sample interacts with light. Typically, a sample holder, such as a sample cuvette for a liquid or particulate sample, is used to hold the sample within the light path in a repeatable manner. The sample holder has an internal sample receiving volume and is provided with surfaces, typically opposing surfaces, at least portions of which are transparent to light interacting with the sample. The spacing between these transparent portions defines the optical path length through the sample holder and, therefore, through the sample held in the sample holder.

[0003] A common way to obtain the necessary spectral data in any spectrometer is by generating a transparency (or absorbance) spectrum of the sample. For this purpose, a so-called single beam spectrum (SB S ), which includes spectral data related to both the sample and the spectrometer. In order to separate the spectral data related to the sample, a similar single beam spectrum (SB) is usually measured on a so-called null material such as water or a water-based substance (for example, if the sample to be measured is a liquid) or air (for example, if the sample to be measured is a solid). Z ). This type of single beam spectrum SB Z Contains spectrometer-related and sample spectrum SB S The same effects as those produced by λ are produced, but the effects due to the sample are absent. A zero material spectrum is then employed to provide a wavelength-dependent zero potential across the spectral region within which the spectral data are collected.

[0004] The single beam spectrum (SB) of the sample was then S ) divided by zero single beam spectrum of the material (SB Z ), the wavelengths of the two single beam spectra are the same in all corresponding spectra to obtain the so-called dual beam spectrum (DB S ), the dual beam spectrum is the transparency spectrum of the sample relative to the null material and is essentially only related to the transmission properties of the sample. As is known to all, the negative logarithm of the transparency spectrum is taken 10 The absorbance spectrum of the sample is obtained. These operations are performed in an arithmetic unit of a computing device associated with the spectrometer and arranged integrally with the spectrometer or arranged separately but operatively connected to the spectrometer, for example in the form of a suitably programmed personal computer.

[0005] Over time, the output of a spectrometer tends to change. One aspect of this change can be described as amplitude variation due to different amplitudes being measured at the same wavelength for the same sample in two otherwise similar spectrometers at different times, or in two runs of the same spectrometer. This is typically due to wear of the sample holder causing the spacing between relatively transparent parts to change and therefore causing the optical path length through the sample holder to change. As is known, the Beer-Lambert law states that the absorbance of light by a sample at a given wavenumber (wavelength) is proportional to the optical path length through the sample. Therefore, as the sample holder wears and the optical path length changes, the amplitude of the spectrometer output will also change and need to be compensated for at regular intervals.

[0006] In order to compensate for amplitude variations of the spectrometer, the spectrometer is usually calibrated regularly. Such a standardization is known from US 9,874,515, which discloses a method for determining the path length deviation through a sample in a cuvette. The method comprises: exposing the sample to electromagnetic radiation at a plurality of wave numbers; determining the electromagnetic absorption in the sample at the plurality of wave numbers; determining the zero-beam absorption spectrum (SB) of an absorption band, in particular a zero liquid. Z ), wherein the first wavenumber associated with a first absorption level of a water absorption band in a water absorption band and the second wavenumber associated with a second absorption level of the absorption band are different from the first wavenumber; determining a difference between the first wavenumber and the second wavenumber; and determining a path length deviation based on the difference. Based on this path length difference, an intensity change can be calculated using the Beer-Lambert law and can be compensated for in subsequent spectral measurements.

[0007] Unfortunately, the zero-beam absorption spectrum (SB) thus recorded Z ) contains not only information about the zero material in the cuvette (e.g., zero liquid), but also background information about elements within the light path between the light emitter and the light detector, including those in the atmospheric air, which are independent of the zero material but which affect the light intensity.

[0008] As disclosed in US 9,874,515, such background information can be determined using a ventilation measurement, i.e., a measurement in which the cuvette contains only air. In this case, no sample is present during the spectral analysis, and the single-beam spectrum includes information only about the sample cuvette, the air inside the cuvette, the reflection from the mirror, the emission spectrum of the electromagnetic source, the sensitivity of the detector, etc. However, the spacing between the opposing transparent windows of a typical cuvette is around 50 μm, which makes it difficult to ensure that all sample is removed and only air is present in the cuvette during such a background measurement. It is impractical to disassemble and thoroughly dry the cuvette for each compensation measurement, just as it is impractical to replace the sample cuvette with a dry cuvette for each compensation measurement.

[0009] In order to avoid this, it is also known from US 9,874,515 to make a mathematical estimate of the background spectrum. Such an estimate has been shown to be insufficiently accurate in certain circumstances and for certain applications.

[0010] According to a first aspect of the present invention, a method for correcting amplitude variations in the output of a spectrometric instrument is provided, the method comprising: exposing an unknown sample in a sample holder to electromagnetic radiation at a plurality of wavenumbers; detecting, by the spectrometric instrument, electromagnetic absorption intensities in the unknown sample at the plurality of wavenumbers; making the detected absorption intensities indexed by the wavenumbers accessible to a computer device associated with the spectrometric instrument as spectral data; and applying a mathematical transformation to the spectral data by means of the computer device to correct for amplitude variations in the output of the spectrometric instrument; wherein the method further comprises: calculating the mathematical transformation in the computer device by determining the difference between absorbance values at first derivatives of spectral data from a null material sample at two different wavenumber ranges, the spectral data being the detected absorption intensities indexed by the wavenumbers; and calculating the mathematical transformation as an inverse function of the determined difference. Thus, by calculating the first derivative and subsequently subtracting two signals close to each other, dependence on background spectra is significantly reduced.

[0011] In some embodiments, the spectral data (SB) from zero material Z ) includes wavenumber-normalized spectral data.

[0012] In some embodiments, calculating the mathematical transform further comprises calculating a humidity correction factor as a slope / intercept of a selected humidity region in the absorbance spectrum.

[0013] This humidity correction factor effectively compensates the background information for the effects of humidity and has the advantage that this compensation can be performed independently of any actual knowledge of the background information.

[0014] In these embodiments, calculating the mathematical transformation includes: determining a difference between absorbance values at a first derivative of the spectral data from the zero material sample at two different wavenumber ranges; and calculating the mathematical transformation as a function inversely related to the sum of the determined difference and the humidity correction factor.

[0015] Usefully, the zero material sample is water-based, in particular nominally pure water (water in which any impurities or additives do not make a measurable difference to the measured spectral data of the water). This has the advantage that human error in preparing the zero material sample can be minimised.

[0016] These and other advantages associated with the present invention may become apparent by considering the following description of various aspects of non-limiting exemplary embodiments of the present invention with reference to the accompanying drawings, in which:

[0017] Figure 1 An embodiment of the inventive device according to the invention is schematically illustrated;

[0018] Figure 2 Shown Figure 1 A schematic cross-sectional top view of the sample holder shown in FIG.

[0019] Figure 3 is a block diagram illustrating a method for determining a correction factor according to one embodiment of the present invention;

[0020] Figure 4 is a graphical representation of the first derivative single-beam spectrum of water;

[0021] Figure 5 yes Figure 7 A graphical representation of the region of the single beam spectrum of water shown in , which shows the wavenumber positions corrected for humidity;

[0022] Figure 6 is a graphical representation of the uncorrected absorbance spectrum of milk;

[0023] Figure 7 is a graphical representation of the single-beam spectrum of water; and

[0024] Figure 8 is a graphical representation of the absorbance spectrum of milk corrected according to the method of the present invention.

[0025] In the following, reference will be made to the absorption spectrometer in the context of Figure 1 and 2An embodiment of the inventive apparatus 100 is described. The apparatus 100 comprises a radiation device 200, an interferometric arrangement 300, a detector 400 and a measuring device 500. Furthermore, a sample holder 600 for holding a sample to be analyzed is arranged to be placed in the apparatus 100.

[0026] The radiation device 200 includes a device arranged to Figure 1 and 2 A radiation source 210 that emits polychromatic infrared radiation in the direction indicated by the letter R in FIG.

[0027] The interferometry arrangement 300 comprises the necessary equipment for carrying out Fourier transform spectroscopy, which is well known to those skilled in the art. For example, the interferometry arrangement 300 comprises a collimator for collimating the infrared radiation and further equipment comprised in an interferometer, such as optical components like mirrors and lenses.

[0028] The detector 400 is arranged to detect incoming infrared radiation which is transmitted through the sample holder 600, see further below.

[0029] The measuring device 500 comprises a computer 510 which is connected to the detector 400 to collect unprocessed data on the detected infrared radiation. By means of this connection, the measuring device 500 is arranged to determine the transparency in a discrete plurality of channels positioned equidistantly along the wave number axis. The computer 510 comprises a processor for processing the collected data, suitable computing software and further equipment known to those skilled in the art. In addition, the computer 510 is arranged to store the collected data and the processed data in a memory. According to an embodiment of the invention, a routine using a Fourier transform algorithm is used in order to transform the unprocessed data from the detector 400 into data on the intensity as a function of the wave number. In addition, the computer 510 is arranged to present the data graphically using a two-dimensional plot, see the following reference to Figure 4-8 .

[0030] In the following, the radiation device 200, the interferometric arrangement 300, the detector 400 and the measuring device 500 will be referred to as an FTIR spectrometer or simply a spectrometer. A method of correcting intensity deviations (or amplitude variations) of such an FTIR spectrometer will be described further below.

[0031] A sample holder 600 is placed between the interferometry arrangement 300 and the detector 400. Furthermore, the sample holder 600 is arranged to hold a liquid sample to be spectrally analysed by passing infrared radiation through it. For example, the liquid sample may be milk or wine. In an embodiment of the present invention, the liquid sample mainly comprises water 610, which acts as a reference or so-called "zero" liquid and is used in order to correct for deviations in the cuvette path length. The water sample 610 is placed in a cuvette 620, which is partially made of calcium fluoride. The outer surface of the cuvette 620 is shaped as a rectangular parallelepiped. The cuvette 620 comprises an inner wall 630, a window element 640, a spacer 650, a cavity 660 and a sample space 622 for holding the sample 610, see Figure 2 6. It is clear that the inner wall 630 and the window element 640 are transparent to the infrared radiation transmitted through the sample 610. Note that the spacer 650 does not need to be transparent. For example, the spacer 650 can be composed of plastic. The volume of the sample space 622 can change as the extension of the spacer 650 changes. In effect, the spacer 650 creates a path length for the cuvette 620. Furthermore, there is an inlet 670 for introducing the sample 610 into the sample space 622 and an outlet 680 for removing the sample 610 from the space 622. According to some embodiments, the sample 610 is kept in motion during the measurement so that Figure 2 Flow through sample space 622 from inlet 670 to outlet 680, as indicated by the arrows in FIG. However, in other embodiments, sample 610 is held stationary in sample space 622 during measurement, in which embodiments inlet 670 and outlet 680 may be omitted.

[0032] The distance covered by the infrared radiation in the sample space 622 is called the path length. Figure 1 and Figure 2 The sample 610 is transported through the sample 610 in a direction R at right angles to the sides of the cuvette 620, so the path length L extends along the interior length of the cuvette 620 between the window elements 640. If the cuvette 620 wears, the path length L will change (increase).

[0033] In fact, since the window element 640 in contact with the water sample 610 is made of calcium fluoride, the window element will dissolve over time. During the life of the cuvette 620, the cuvette may also be corroded by other chemicals. For example, the thickness T of the window element 640 (see Figure 2) will become smaller over time. Therefore, the path length L will increase over time, resulting in path length deviation. In addition, note that cuvettes placed in different devices 100 of the same type have different path lengths by default. For example, different path lengths may be caused by different degrees of cuvette dissolution, even if the cuvettes are essentially similar at a certain point in time. In addition, the extension of the spacer 650 may vary between different cuvettes 620, thereby resulting in different path lengths. Therefore, in order to make the characteristics of different devices 100 of the same type more similar and the characteristics of the same device 100 more stable over time, it is necessary to compensate for changes in path length.

[0034] References below Figure 3 The block diagram of FIG. 1 illustrates an exemplary embodiment of a method according to the present invention for correcting intensity deviations in an apparatus 100 (here, for example, an FTIR spectrometer). As will be further understood below, correcting for path length deviations means correcting for intensity deviations. According to an exemplary embodiment of the present invention, the method utilizes a single beam spectrum SB of a zero liquid sample. Z , the zero liquid sample here being a nominally pure water sample (which may contain small amounts of other components, such as about 0.01% by volume of detergent) to detect deviations in the cuvette path length. After the spectrometer 100 has been calibrated using measurements on water samples, the spectrometer can be used to measure other liquid samples, such as milk or wine, to quantitatively determine the components of interest in these samples in a manner well known in the art.

[0035] According to the method of the present invention, the intensity correction I can be determined corr , the intensity correction converts the absorbance value of the sample measured in the sample holder of path length L to the zero material (here water) Corrected to those measured in a sample holder of standard path length L0 This can be described as follows:

[0036]

[0037] It follows that the intensity correction is the ratio between the two path lengths, as follows:

[0038] I corr = L0 / L (2)

[0039] The object of the present invention is to provide a method by which I can be determined without using a reference material with known absorbance. corr method.

[0040] Single beam spectrum SB of zero material (here water) ZThe logarithmic transformation converts the intensity value (y-axis value) into absorbance unit, and then the single beam spectrum SB can be Z Decomposed into its different components, as follows:

[0041]

[0042] Among them SB 空气 is the single-beam spectrum of air, and is the absorbance of water relative to air over a path length of L. This is obtained from equation (3):

[0043]

[0044] in is the absorbance of water relative to air over a path length of L0.

[0045] According to equation (2), equation (4) can be rewritten as:

[0046]

[0047] In order to reduce the influence of the background, usually to a negligible level, the first-order derivative is calculated according to the method of the present invention. Background SB 空气 The derivative of changes slowly, while water absorbs SB Z The derivative of will change much faster, especially in the single beam spectrum of water SB Z At the rising and falling edges of the water absorption band in , therefore according to equation (5):

[0048]

[0049] By taking the derivative at two different wavenumbers (or ranges) x1 and x2 that are close to each other on the x-axis (wavenumber axis), it is also possible to ignore any slope in the background. Preferably, the two different wavenumbers (or ranges) x1 and x2 are chosen so that the slope in the water spectrum relative to air at these points (or ranges) is very different. This can be expressed mathematically as:

[0050]

[0051] It can be rewritten more simply as:

[0052] Δ(SB Z ) / =Δ(SB 空气 ) / -(1 / I corr )·ΔA L0 / (8)

[0053] in:

[0054] Δ(SB Z ) / =log 10 (SB Z (x1)) / -log 10 (SB Z (x2)) / is the difference in slope in the measured single-beam null spectrum of water

[0055] Δ(SB 空气 ) / =log 10 (SB 空气 (x1)) / -log 10 (SB 空气 (x2)) / is the difference in slope of the measured single-beam null spectrum of water to the background

[0056]

[0057] Next, the background can be decomposed into a component SB due to the dry air in the light path 空气,干燥 and optionally a component SB due to the humidity in the air in the light path hum ,therefore:

[0058]

[0059] From equation (8), we can get:

[0060] Δ(SB Z ) / = Δ(SB 空气,干燥 ) / + Δ(SB hum ) / - (1 / I corr ) ·Δ A L0 / (10)

[0061] Then, the intensity correction I corr It can be expressed as

[0062] (1 / I corr ) = (Δ(SB Z ) / / Δ A L0 / ) + (Δ(SB 空气,干燥 ) / / Δ A L0 / ) + (Δ(SB hum ) / / ΔA L0 / ) (11)

[0063] or:

[0064]

[0065] Values c1, c2 and humidity correction (corr hum ) formula needs to be determined experimentally when used:

[0066] The constant c1 is only relevant to the absorption of water relative to air at the nominal path length L0;

[0067] The constant c2 is the contribution to the background due to dry air, specifically but not essentially this may not be the average contribution of the population of devices 100; and

[0068] Correction corr hum It is a single beam spectrum SB based on zero material (here water) Z The correction is for the amplitude of the humidity characteristic in φ, and in some cases where the effect of the amplitude is negligible, the correction can be ignored.

[0069] According to the present invention, corr hum Calculated as the slope / intercept for the selected humidity region as follows:

[0070]

[0071] The constant c3 describes the humidity correction to the intensity I corr the impact of

[0072] The constant c4 is the offset between the intensities at the two peak positions in the absence of humidity. This offset approaches zero as Peak 1 and Peak 2 approach each other.

[0073] Humidity (from water in gaseous form, i.e., water vapor) produces an infrared spectrum with a fringe pattern. The ratio between the valley (peak 2) and the adjacent peak (peak 1) in this fringe pattern can be used as a measure of the amount of water vapor in the light path. By measuring the single beam null spectrum SB of water on one or more devices 100 of the same type using cuvettes with path lengths covering the expected variation, such as 50 μm to 60 μm or 37 μm to 44 μm, the single beam null spectrum SB of water can be measured. z The constants c1, c2, c3, and c4 are determined empirically. The effect on the spectrum of a known sample (e.g., milk or wine or glycerol or other chemical solution) is collected and the constants are adjusted until all spectra are identical. This task only needs to be performed once for a given device type and is used for all devices of that type in the future.

[0074] In some embodiments, the x-axis (wavenumber scale) of the single-beam spectrum is normalized before performing the y-axis (amplitude) correction. This can be achieved by applying a mathematical transformation to the spectrum in a manner well known in the art, whereby the measured data is normalized along the x-axis. In an exemplary embodiment of the present invention, the x-axis normalization is based on the CO2 peak in the infrared range.

[0075] As is known, such x-axis normalization comprises: normalizing the wavenumber scale of the optical spectrum recorded by the device 100 by providing the optical spectrum recorded by the device 100 and comprising a spectral pattern of a component of the atmospheric air in the light path in the device 100; selecting a spectral pattern of a component of the atmospheric air in the device 100 (here CO2 in the air); determining one or more wavenumber-related position values associated with the selected spectral pattern; constructing a mathematical transformation based on the difference between the determined one or more values and the corresponding one or more reference values of the selected spectral pattern, and applying the mathematical transformation to the optical spectrum subsequently recorded by the device 100 to normalize the wavenumber scale.

[0076] Additionally, the humidity correction corr hum It can be used in other methods to determine intensity corrections, such as the method disclosed in US 9,874,515, to compensate for background effects.

[0077] Now refer to Figure 3 The block diagram shown in FIG further describes an exemplary embodiment of the method according to the present invention. The method (block 700) includes exposing a zero liquid having a known amount of at least one component (here a water sample 610) to polychromatic infrared radiation from the radiation device 200 (block 710). The radiation is generated by Figure 1 and Figure 2 The detector 400 detects the incoming infrared radiation that has been transmitted through the interferometric arrangement 300, the water sample 610 and the cuvette 620, thereby determining (block 720) using the measurement device 500 that the range is between 900 cm -1 About 3500cm -1 More specifically, the intensity levels of a discrete set of wavenumbers, usually equally spaced, within this range are determined. The intensity data indexed for the wavenumber data is stored in the memory of the computer 510 as a single beam spectrum SB of the zero material. Z The computer 510 processes the stored data using appropriate mathematical transformations to obtain (block 730) log normalized along the x-axis (wave number scale). 10 Change the intensity (or absorbance) level.

[0078] At least 1650cm -1The first derivative of the absorbance spectrum, log , is calculated (block 740) in the region of the spectrum characterized by water absorption. 10 (SB Z ) / This can be implemented in the computer 510 using the known Savitzky-Golay algorithm. The difference Δ(SB) between the two ranges under the first derivative is calculated (block 740). Z ) / , two of which are close to 1650cm -1 The absorption of the water band is preferably characterized by the higher wave number shoulder of the water band. In the embodiment of the present invention, the range of 1740-1746 cm is used. -1 and 1844-1850cm -1 This is Figure 4 The figure shows a plurality of instruments 100 of the same type (here four instruments: Figure 4 Representative first derivative single beam spectra of water obtained (labeled as Instrument 1, Instrument 2, Instrument 3 and Instrument 4 in the figure).

[0079] The humidity correction corr in use is calculated (block 750) according to equation (13) hum To compensate for the ambient humidity, I corr In the embodiment of the present invention, the range of 1832-1840cm is used. -1 (peak 1) and 1814-1822cm -1 (Peak 2). Figure 5 The figure shows the structure of the dotted line. Figure 4 Representative single beam spectra of water obtained by the same plurality (here four) of instruments 100 mentioned.

[0080] Then calculate (block 760) as the inverse of the inverse intensity correction 1 / I calculated according to equation (12) corr Intensity correction I corr This intensity correction I can then be corr Applied (block 770) to optical spectra subsequently recorded by the apparatus 100 to normalize the absorbance intensity scale (y-axis).

[0081] The following will now be discussed with respect to a plurality of instruments 100 of the same type (herein referred to above) measuring on the same milk sample. Figure 4 and Figure 5An example of the application of this method is described by standardizing the outputs of the four instruments mentioned. It should be understood that in the following figures, which are presented as showing measurements taken by each of the four instruments, the variations between the instruments may be so small that they are visually represented as overlaps. The differences between the four instruments can be better discerned from the numerical values presented in the accompanying tables (including Tables 1 to 3).

[0082] exist Figure 6 The resulting absorbance of milk measured on four different instruments 100 of the same type is plotted against wavenumber. Figure 6 Each of the spectra in is presented as an interpolated curve in a two-dimensional plot with the absorbance amplitude on the vertical axis and the corresponding wavenumber on the horizontal spectral axis. According to an alternative graphical presentation, the plot can be a scatter plot. As can be seen, the absorbance values at the same wavenumber vary slightly between the different instruments. This will result in differences in the amounts of the components of the same milk sample determined chemometrically in a known manner from these absorbance spectra in each of the instruments. When the prediction models for fat, protein, lactose, total solids (TS) and solids-free-fat (SNF) were applied to the uncorrected spectra of a set (here fifteen) of standard milk samples, the following root mean square error (RMSEP) between the instruments and their common mean was obtained - the results are in g / 100 mL:

[0083]

[0084] *The mean error is relative to the mean concentration of the parameter.

[0085] Table 1: Predicted milk components using the uncorrected spectra of each instrument

[0086] This level of performance is unacceptable to users who require nearly identical results from different instruments.

[0087] According to the method according to the invention, a single beam spectrum (SB) of water was obtained on each of the four instruments. Z )(See Figure 7 ), and the first derivative of the absorbance spectrum normalized on the wavenumber axis of each of these single-beam spectra is determined in the computer 510. Figure 4 These first-order derivatives are shown in FIG, which also depicts the position at which the difference Δ(SB Z ) / From the two regions (dashed lines). Figure 7 and Figure 4As can be seen in FIG, the y-axis values at each wave number are slightly different for each of the four instruments 100. The recorded single beam spectra are processed by the computer 500 of the instrument 100, which generates the corresponding spectra to be calculated based on the values given by Figure 5 The intensity values at the two regions shown by the dashed lines in the figure determine the humidity correction corr for each instrument. hum In each instrument, the associated computer 500 then calculates an intensity correction factor I specific to each of the four instruments. corr .

[0088] instrument <![CDATA[Strength correction (I corr )]]> 1 0.9979 2 0.9953 3 0.9740 4 0.8823

[0089] Table 2: Intensity correction factors for each instrument

[0090] The correction factor I for each instrument can then be corr Applied to absorbance data collected by each instrument (e.g. Figure 6 ) and brings the calibrated (or y-axis normalized) absorbance data for each instrument into closer agreement. Figure 8 This is demonstrated in FIG, which shows Figure 6 The absorbance spectrum of milk is shown in Figure 1, but with an appropriate intensity correction factor I for the instrument that recorded the associated milk spectrum. corr Each spectrum was corrected (see Table 2).

[0091] When this correction was applied to a set (here fifteen) of standardized milk samples, the following root mean square error (RMSEP) between the instrument and its common mean was found in the predictions of fat, protein, lactose, total solids (TS), and solids-free-fat (SNF) - the results are in g / 100mL:

[0092]

[0093] *The mean error is relative to the mean concentration of the parameter.

[0094] Table 3: Predicted milk components using the uncorrected spectra of each instrument

[0095] These errors are very low compared to the typical prediction errors of 0.8% relative to the values of these components determined with known wet chemical methods.

[0096] Those skilled in the art will appreciate that application of the method according to the present invention allows for greater normalization of output between different instruments 100 of the same type as well as normalization of output of the same instrument 100 when the path length through the cuvette 620 varies.

[0097] As will be appreciated by those skilled in the art, instead of expressing spectral information about electromagnetic radiation in terms of wave numbers, wavelength or frequency may be used without departing from the invention as claimed.

Claims

1. A method of correcting amplitude variations in the output of a spectrometric instrument due to changes in optical path length (L) through a sample holder (600), the method comprising: exposing an unknown sample in the sample holder (600) to electromagnetic radiation at a plurality of wave numbers; detecting, by the spectrometer, the electromagnetic absorption intensity of the unknown sample at the plurality of wave numbers; providing the detected absorption intensities indexed for wavenumbers as spectral data to a computer device associated with the spectrometric instrument; measuring the absorbance value of the sample relative to a null material in a sample holder; and correcting the intensity I by means of the computer device. corr applied to the generated absorbance values to correct for the amplitude variations in the output of the spectrometric instrument; wherein the method further comprises: determining in the computer device a first derivative log of the logarithmic transformed spectral data from the null material at a first wavenumber range x1 10 (SB Z (x1)) / and the first derivative of the logarithmically transformed spectral data from the zero material at the second wavenumber range x2 log 10 (SB Z (x2)) / The difference between Δ(SB Z ) / To calculate the intensity correction I corr ; and calculate as the difference Δ(SB Z ) / The intensity correction I of the inverse correlation function corr as follows: (1 / I corr )=c1·Δ(SB Z ) / +c2+corr hum Among them, the values c1, c2 and humidity correction corr hum The formula for is determined experimentally when used; the constant c1 is related only to the absorption of water relative to air at the nominal path length L0; the constant c2 is the contribution to the background due to dry air; and the correction corr hum It is a single beam spectroscopy (SB) based on zero material Z Correction of the amplitude of the humidity characteristics in; Among them, the zero material is water.

2. The method according to claim 1 , wherein the intensity correction I is calculated corr Also includes: By determining the spectral data SB from the zero material Z Humidity correction factor corr for the slope / intercept of the selected humidity region hum to correct the spectrometric instrument for the humidity present in the air in the optical path between the radiation device (200) and the detector (400); and to calculate as a difference Δ(SB Z ) / and the humidity correction factor corr hum The intensity correction I corr .

3. The method of claim 1, wherein the zero material comprises a nominally pure water sample. The method according to claim 1 , wherein the absorption intensity is detected by Fourier transform spectroscopy.

5. The method according to claim 1, wherein the spectral data SB from zero material Z Includes wavenumber-normalized spectral data.

6. A device (100), comprising: a spectrum a spectrometric instrument configured to perform Fourier transform spectroscopy on a sample in a sample holder (600); and an associated computing device configured to receive spectral data from the spectrometric instrument and apply an intensity correction stored in a memory accessible by the associated computing device to the spectral data, the transformation correcting the spectral data for amplitude variations in absorption intensity detected by the spectrometric instrument; wherein the associated computing device is programmed to operate to cause the apparatus (100) to perform a method according to any one of the preceding claims.

Images