RU2805385C2 - Method for correcting amplitude changes in spectrometer - Google Patents

Abstract

FIELD: spectroscopic measurements.

SUBSTANCE: invention concerns a method for correcting changes in the amplitude of the output signal of a spectrometric device. The method includes exposing a sample in a sample holder to electromagnetic radiation with a plurality of wave numbers, determining absorption intensities of the electromagnetic radiation in the sample with multiple wave numbers, and providing the determined absorption intensities indexed by wave number as spectral data to a computer device. The amplitude of the output signal is corrected by applying a mathematical transformation (I_corr). Mathematical transformation (I_corr) is calculated by determining the difference (Δ(SB_Z)^/) between the first derivative of log-transformed spectral data (SB_Z) for the sample of zero material in the first range (x₁) of wave numbers and the first derivative of log-transformed spectral data (SB_Z) for the sample of zero material in the second range (x₂) of wave numbers and calculating the mathematical transformation (I_corr) as a function inversely dependent on a certain difference (Δ(SB_Z)^/).

EFFECT: increasing the accuracy of measurements.

6 cl, 8 dwg, 3 tbl

Description

[0001] The present invention relates to a method for compensating for changes in amplitude at the output of a spectrometer, a type of spectrometer for generating spectral data for unknown samples contained in sample holders, and, in particular, for compensating for changes in amplitude due to changes in the optical path length through the sample holder.

[0002] In typical spectrometers configured to generate spectral data for unknown samples, the light emitter and light detector are configured to determine the path of light along which the sample in question is located when the sample interacts with the light. Typically, a sample holder, for example containing a liquid or solid sample cuvette, is used to place samples within the light path in a reproducible manner. The sample holder has an internal volume for receiving the sample and has surfaces, typically opposing surfaces, at least portions of which are transparent to light interacting with the sample. The separation between said transparent areas limits the length of the optical path through the sample holder and thus through the sample contained in the sample holder.

[0003] A common method for obtaining the required spectral data in any spectrometer is to generate a transmittance (or absorption) spectrum of the sample. To do this, a so-called single-beam spectrum (SB _S ) is obtained, which contains spectral data related to both the sample and the spectrometer. To isolate spectral data related to a sample, a similar single-beam spectrum (SB _Z ) is typically measured on a so-called null material such as water or a water-based material if, for example, the sample to be measured is liquid or air , or if, for example, the sample to be measured is a solid. These single-beam SB _Z spectra provide the same spectrometer-related effects as the SB _S spectra of the sample, but there are no sample-related effects. The spectrum of the null material is then used to provide a wavelength-dependent null in the spectral region in which the spectral data is collected.

[0004] The single-beam spectrum of the sample (SB _S ) is then divided by the single-beam spectrum of the null material (SB _Z ) at the same wavelengths in all relevant spectra to obtain the so-called double-beam spectrum of the sample (DB _S ), which is essentially the transmittance spectrum of the sample relative to zero material and relates only to the transmission properties of the sample. As is well known, calculating the negative log ₁₀ for it yields the absorption spectrum for the sample. These operations are performed in the arithmetic unit of the computing device, which is connected to the spectrometer and which is built into the spectrometer or is a separate device, but functionally connected to the spectrometer, for example, in the form of a suitably programmed personal computer.

[0005] Over time, the spectrometer output changes. One aspect of this change can be described as a change in amplitude, which results in different measurements at the same wavelengths for the same sample in two otherwise similar spectrometers, or in two runs using the same spectrometer at different times. amplitude values. Typically this is due to wear on the sample holder, which changes the distance between opposing transparent parts and therefore changes the optical path length through the sample holder. As is known, according to the Beer-Lambert law, the absorption of light by a sample at a given wave number (wavelength) is proportional to the length of the optical path through the sample. Thus, as the sample holder wears and the optical path length changes, the amplitude of the spectrometer output signal changes and requires regular compensation.

[0006] To compensate for changes in the amplitude of the spectrometer, the spectrometer is typically standardized periodically. This standardization is known from US Pat. No. 9,874,515, which discloses a method for determining the deviation of the optical path length through a sample in a cuvette. This method involves: exposing a sample to multiple wavenumber electromagnetic radiation, determining the absorption of electromagnetic radiation in the multiple wavenumber sample, determining the first wavenumber associated with the first absorption level from the absorption band, in particular the absorption band for water in the spectrum (SB _Z ) absorption of the zero ray of the zero liquid, and a second wavenumber associated with a second absorption level from said absorption band, the second wavenumber being different from the first wavenumber, determining the difference between the first wavenumber and the second wavenumber, and determining the deviation of the optical path length based on the specified difference. From this difference in optical path length, the change in intensity can be calculated using the Beer-Lambert law and compensated for in subsequent spectral measurements.

[0007] Unfortunately, the zero beam absorption spectrum (SB _Z ) recorded in this way includes information not only about the zero material (eg, zero liquid) in the cuvette, but also background elemental information, including information about atmospheric air , within the light path between the light emitter and the light detector, which is not associated with the null material but affects the light intensity.

[0008] As disclosed in US Pat. No. 9,874,515, this background information can be determined using the air measurement result, i.e. measurements in which the cuvette contains only air. In this case, the sample is absent during spectral analysis, and the single-beam spectrum contains information only about the sample cuvette, the air inside the cuvette, reflection from mirrors, emission spectrum of the electromagnetic source, detector sensitivity, etc. However, the distance between the opposing clear windows of a typical cuvette is about 50 µm, making it difficult to remove the entire sample and ensure that the cuvette is only air-filled during such background measurements. Disassembling and thoroughly drying the cuvette for each compensation measurement is not practical, nor is replacing the sample cuvette with a dry one for each compensation measurement.

[0009] To avoid this, the mathematical estimation of the background spectrum known from US patent 9,874,515 is used. This assessment has proven to be insufficiently accurate in certain circumstances and for certain applications.

[0010] According to a first aspect of the present invention, there is provided a method for correcting changes in the amplitude of an output signal of a spectrometer instrument, comprising exposing an unknown sample in a sample holder to electromagnetic radiation having a plurality of wave numbers; determination, using a spectrometric device, of the absorption intensity of electromagnetic radiation in an unknown sample with many wave numbers; transmitting to a computer device associated with the spectrometric instrument certain absorption intensities, indexed by wave number, as spectral data; and applying, by means of a computer device, a mathematical transformation to the spectral data to correct for changes in the amplitude of the output signal of the spectrometer; wherein the method further includes calculating on a computer device a mathematical transformation by determining the difference between absorption values in two different wavenumber ranges in the first derivative of the spectral data, determining the wavenumber-indexed absorption intensity for a sample of the zero material; and calculating the mathematical transformation as a function inversely of the specified difference. Thus, by calculating the first derivative and then subtracting two signals close to each other, the dependence on the background spectrum is significantly reduced.

[0011] In some embodiments, calculating the mathematical transformation also includes calculating a slope/intercept moisture correction factor for selected moisture regions in the absorption spectrum.

[0012] This humidity correction factor effectively compensates for the effect of humidity in the background information and has the advantage that this compensation can be performed independently of any actual known background information.

[0013] In these embodiments, calculating the mathematical transformation includes determining the difference between absorption values in two different wavenumber ranges in the first derivative of the spectral data for a null material sample; and calculating a mathematical transformation as a function inversely dependent on the sum of the determined difference and the moisture correction factor.

[0014] Preferably, the null material sample is based on water, particularly nominally pure water (water in which any impurities or additives do not appreciably affect the measured spectral data of the water). This provides the advantage that human errors in preparing a sample of zero material can be minimized.

[0015] These and other advantages associated with the present invention will become apparent upon reading the following description of non-limiting and exemplary embodiments of the present invention with reference to the accompanying drawings, in which:

in fig. 1 schematically shows an embodiment of the proposed apparatus according to the present invention; in fig. 2 is a schematic top cross-sectional view of the sample holder shown in FIG. 1; in fig. 3 is a flowchart illustrating a method for determining a correction factor according to one embodiment of the present invention; in fig. 4 a graphical representation of single-beam spectra of water (first derivative) is presented; in fig. 5 presents a graphical representation of the region of single-beam spectra of water shown in Fig. 7, which shows the positions of the wave numbers for humidity correction; in fig. 6 a graphical representation of the uncorrected absorption spectra for milk is presented; in fig. 7 a graphical representation of single-beam spectra of water is presented; And in fig. 8 is a graphical representation of absorption spectra for milk corrected according to the method of the present invention.

[0016] Next, with reference to FIG. 1 and 2, an embodiment of the apparatus 100 of the present invention will be described in the context of absorption spectroscopy. The apparatus 100 includes an emitting device 200, an interferometric device 300, a detector 400, and a measuring device 500. In addition, a sample holder 600 for housing a sample to be analyzed is configured to be placed in the apparatus 100.

[0017] The emitting device 200 includes a radiation source 210 that is configured to emit polychromatic infrared radiation in the direction indicated by R in FIG. 1 and 2.

[0018] Interferometric device 300 contains the necessary equipment for implementing Fourier transform spectroscopy, which is well known to those skilled in the art. For example, interferometric device 300 includes a collimator that collimates infrared radiation, and additional equipment contained in the interferometer, such as optical components such as mirrors and lenses.

[0019] The detector 400 is configured to detect incoming infrared radiation that passes through the sample holder 600, see below for more details.

[0020] The measurement device 500 includes a computer 510 that is coupled to the detector 400 to collect raw data about the detected infrared radiation. Using this connection, the measuring device 500 is configured to determine the transmittance in a discrete number of channels spaced at regular intervals along the wavenumber axis. Computer 510 includes a processor for processing collected data, associated computer software, and additional equipment well known to those skilled in the art. In addition, the computer 510 is configured to store collected data and processed data in a storage device. According to the present embodiment, a procedure including Fourier transform algorithms is used to convert the raw data from detector 400 into intensity data as a function of wavenumber. In addition, the computer 510 is configured to graphically represent data in the form of two-dimensional graphs, see FIG. 4-8, which are referenced below.

[0021] The emitting device 200, the interferometric device 300, the detector 400 and the measuring device 500 will hereinafter be referred to as an FTIR spectrometer or simply a spectrometer. A method for correcting intensity deviations (or amplitude changes) of said FTIR spectrometer will now be described below.

[0022] The sample holder 600 is placed between the interferometric device 300 and the detector 400. In addition, the sample holder 600 is configured to receive a liquid sample to be performed spectral analysis by passing infrared radiation through it. For example, the liquid sample may be milk or wine. In the present embodiment, the liquid sample mainly contains water 610, which serves as a reference or so-called “null” fluid and is used to perform correction for path length deviations in the cuvette, see more below. A water sample 610 is placed in a cuvette 620, which is partially composed of calcium fluoride. The outer surface of the cuvette 620 has the shape of a rectangular parallelepiped. Cuvette 620 has interior walls 630, window elements 640, dividers 650, cavities 660, and a sample space 622 for housing sample 610, see sectional top view of FIG. 2. It will be understood that the interior walls 630 and window elements 640 are transparent to infrared radiation that is passed through the sample 610. It should be noted that the dividers 650 do not necessarily need to be transparent. For example, the spacers 650 may be made of plastic. The volume of the sample space 622 can be varied by changing the length of the spacers 650. As seen in the drawing, the spacers 650 define the optical path length in the cell 620. In addition, there is an inlet 670 for introducing sample 610 into the sample space 622 and an outlet 680 for removing sample 610 from space 622. In some embodiments, sample 610 is in motion during measurement, flowing from inlet 670 to outlet 680 through sample space 622, as indicated by arrows in FIG. 2. However, in other embodiments, sample 610 remains stationary in sample space 622 during measurement, and in such embodiments, inlet 670 and outlet 680 may be eliminated.

[0023] The distance that infrared radiation travels in sample space 622 is referred to as the optical path length. Since the radiation passes through the sample 610 at right angles to the side edge of the cuvette 620, in the direction R in FIG. 1 and fig. 2, the optical path length L is the same as the internal length of the cuvette 620 between the window elements 640. As the cuvette 620 wears, the optical path length L changes (increases).

[0024] Since the window elements 640 in contact with the water sample 610 are made of calcium fluoride, they will dissolve over time. During the life of the 620 cuvette, it may also be damaged by other chemicals. For example, the thickness T (see FIG. 2) of window elements 640 decreases over time. Consequently, the length L of the path will increase over time, i.e. There will be a deviation in the optical path length. In addition, it should be noted that cuvettes placed in different devices of the same type 100 have different optical path lengths by default. For example, different optical path lengths can result from diluting the cells to different degrees, even if the cells were essentially the same at some point in time. In addition, the length of the spacers 650 may vary between different cells 620, resulting in variations in the optical path length. Thus, in order for the performance of different devices of the same type 100 to be more similar, and the performance of the same device 100 to be more stable over time, it is necessary to compensate for the change in optical path length.

[0025] An exemplary embodiment of a method for correcting intensity deviations in an apparatus 100 (in this case, for example, an FTIR spectrometer) according to the present invention is described below with reference to the flowchart shown in FIG. 3. As will be clear from the following description, correcting the optical path length deviation also means correcting the intensity deviation. In this exemplary embodiment, the method uses a single beam spectrum SB _Z of a zero liquid sample, which in this case is a sample of nominally pure water (the water sample may contain small amounts of other components, for example about 0.01% by volume detergent ), to determine deviations in the optical path length in the cuvette. Once the spectrometer 100 has been adjusted using measurements on a water sample, it can be used to make measurements on other liquid samples, such as milk or wine, to quantify the components of interest in those samples in a manner well known in the art.

[0026] According to the present method, an intensity correction, I _corr , by which the absorption values are adjusted ( ) for a sample relative to a zero material (in this case water), for which measurements are performed in a sample holder with optical path length L, to the specified values ( ), measured in a sample holder with a standard optical path length _L0 . This can be expressed as follows:

It follows that the intensity correction is the following relationship between these two optical path lengths:

[0027] It is an object of the present invention to provide a method by which I _corr can be determined without the use of a reference material having known absorbance.

[0028] A logarithmic transformation of the single-beam spectrum of the null material SB _Z , in this case water, provides a conversion of intensity values (y-axis values) into absorption units, after which the single-beam spectrum of SB _Z can be decomposed into its various components as follows:

where SB _{of air} is the single-beam spectrum of air, and represents the absorption for water relative to air over the optical path length L. From equation (3) it follows that:

Where represents the absorption for water relative to air over the optical path length _L0 .

Equation (2) can be rewritten as equation (4) as follows:

[0029] To reduce background influence to typically negligible levels, the present method calculates the first derivative. The derivative of the background _air SB changes slowly, while the derivative for the water SB _Z absorption will change much more quickly, particularly at the leading and trailing edges of the water absorption bands in the single-beam spectrum of water SB _Z , thus from equation (5) follows:

When taking derivatives for two different wavenumbers (or ranges), x ₁ and x ₂ , located close to each other on the x-axis (wavenumber axis), any background tilt can also be neglected. Preferably, two different wavenumbers (or ranges), x ₁ and x ₂ , are chosen so that the slopes in the spectrum of water relative to the spectrum of air at these points (or ranges) are very different. Mathematically this can be expressed as follows:

And you can simply rewrite it like this:

Where:

Δ(SB _Z ) ^/ = log ₁₀ (SB _Z (x ₁ )) ^/ - log ₁₀ (SB _Z (x ₂ )) ^/ represents the difference in slope for the measured single-beam null spectrum of water Δ(SB_air)^/=log₁₀(SB_air(x₁))^/- log₁₀(SB_air(x₂))^/ represents the difference in slope for the background of the measured single-beam null spectrum of water ^/ = ( ^/ (x ₁ ) - ^/ ( _x2 )) Difference in the slope of the absorption spectrum for water relative to the spectrum of air at the nominal optical path length

[0030] The background can then be decomposed into a component due to dry air in the light path, SB _{air, dry} , and, optionally, a component due to humidity of the air in the light path, Sb _wet. , Thus:

From equation (8) we get:

Then the intensity correction I _corr can be expressed as follows:

or:

In the case of using the values of c ₁ , c ₂ and the formula for correction for humidity (corr _{. humidity} ), they must be determined experimentally:

the constant c ₁ depends solely on the absorption for water relative to air at the nominal optical path length L ₀ ;

the constant c ₂ represents the contribution to the background due to dry air, in particular, but not predominantly, this may be the average contribution for a plurality of devices 100; And

correction, corr _{. humidity} is a correction based on the amplitude of the moisture signature in the single-beam SB _Z spectrum of the zero material (in this case water), and can be ignored in some cases where its effect is negligible.

In accordance with the present invention, _corr. calculated as slope/intercept in selected moisture regions as follows:

constant c ₃ describes the effect of humidity on intensity correction I _corr ; And

the constant c ₄ represents the offset between the intensity values at the two peak positions in the absence of humidity. It approaches zero as peak ₁ and peak ₂ approach each other.

[0031] Humidity (due to water in the gas phase, i.e. water vapor) provides an infrared spectrum with an interference fringe pattern. The relationship 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's path . The constants c ₁ , c ₂ , c ₃ and c ₄ are determined empirically by measuring the single-beam null spectrum SB _{z of} water on one or more apparatus 100 of the same type using cells with optical path lengths covering the expected change, for example, from 50 μm to 60 μm or from 37 µm to 44 µm. Record the effect on the spectrum of a known sample (for example, milk, wine, glycerol or other chemical solution) and adjust the constants until all spectra are identical. This task needs to be performed only once for a given machine type and can be used in the future for all machines of that type.

[0032] In some embodiments, the x-axis (or wavenumber scale) of the single-beam spectrum is standardized before performing the y-axis (amplitude) correction. This can be achieved using a method well known in the art, which consists of applying a mathematical transformation to the spectrum by which the measured data is standardized along the x-axis. In the present exemplary embodiment, the x-axis standardization is based on the infrared CO ₂ peak.

[0033] As is known, said x-axis standardization involves standardizing the wavenumber scale of the optical spectrum recorded by apparatus 100 by providing an optical spectrum recorded by apparatus 100 containing spectral patterns obtained for components of atmospheric air along the path of light in the apparatus 100; selecting the spectral pattern obtained for the components of atmospheric air in the apparatus 100, in this case CO ₂ in the air; determining one or more wavenumber dependent position values associated with the selected spectral pattern; performing a mathematical transformation based on the difference between the determined value or values and a corresponding reference value or values for the selected spectral pattern, and applying the mathematical transformation to the optical spectra sequentially recorded by the apparatus 100 to standardize the wavenumber scale.

[0034] In addition, humidity correction corr _. may be used in other methods for determining intensity correction, such as, for example, the method disclosed in US Pat. No. 9,874,515, to compensate for background effects.

[0035] Next, with reference to the block diagram shown in FIG. 3, an exemplary embodiment of the method according to the present invention will also be described. The method (block 700) includes exposing a null liquid containing a known amount of at least one component (in this case a water sample) 610 to polychromatic infrared radiation (block 710) from an emitter 200. The radiation is depicted by wavy lines in FIG. 1 and 2. Detector 400 detects incoming infrared radiation that has been transmitted through interferometric device 300, water sample 610, and cuvette 620, thereby determining (block 720) intensity levels for wavenumbers ranging from about 900 cm ^-1 to about 3500 cm ^-1 using measuring device 500. More specifically, in this range intensity levels are determined for a discrete set of typically equally spaced wavenumbers. The intensity data, indexed by the wavenumber data, is stored in the computer memory 510 as a single beam spectrum of the null material SB _Z . Computer 510 processes the stored data using a suitable mathematical transformation to produce (block 730) log ₁₀ -transformed intensity (or absorbance) levels standardized on the x-axis (wavenumber scale).

[0036] Calculate the first derivative, log ₁₀ (SB _Z ) ^/ of this absorption spectrum (block 740) at least in the spectral region characterized by absorption for water of approximately 1650 cm ^-1 . This may be performed on computer 510 using the well-known Savitzky-Golay algorithm. The difference between the two bands in the first derivative Δ(SB _Z ) ^/ is calculated (block 740), where the two bands are characterized by proximity to the water absorption band of approximately 1650 cm ^-1 , preferably on the higher wavenumber arm of the water band. In the present embodiment, the ranges 1740-1746 cm ^-1 and 1844-1850 cm ^-1 are used. This is illustrated by the dotted line in FIG. 4, which shows representative single-beam spectra of first derivative water obtained using a plurality of instruments 100 of the same type (in this case, four: these are designated instrument 1, instrument 2, instrument 3, and instrument 4 in FIG. 4).

[0037] Humidity correction, _{humid corr.} , when used, is calculated (block 750) in accordance with equation (13) to compensate for the effect of ambient humidity on the intensity correction I _corr . In the present embodiment, the ranges of 1832-1840 cm ^-1 ( peak ₁ ) and 1814-1822 cm ^-1 ( peak ₂ ) are used. This is illustrated by the dotted line in FIG. 5, which shows representative single-beam spectra of water obtained using the same plurality (in this case four) of instruments 100 mentioned above in relation to FIG. 4.

[0038] The intensity correction, _Icorr, is then calculated (block 760) as the reciprocal of the intensity correction, 1/ _Icorr , calculated in accordance with equation (12). This intensity correction, _Icor , can then be applied (block 770) to the optical spectra sequentially recorded by apparatus 100 to standardize the absorption intensity scale (y-axis).

[0039] Next, an example of applying this method to standardize the output of multiple devices 100 of the same type will be described, in this case the four devices mentioned above in relation to FIG. 4 and 5, with the help of which measurements are performed on the same milk sample. It should be understood that in the following figures, which are provided to illustrate measurements made with each of these four instruments, the differences between the instruments may be so small as to visually appear to overlap. The differences between these four instruments can be more easily distinguished by the numerical values presented in the accompanying tables, from Table 1 to Table 3 inclusive.

[0040] The absorbance result for milk measured on four different instruments 100 of the same type is shown in FIG. 6 depending on wave numbers. Each of the spectra in Fig. 6 is presented as an interpolation curve on a two-dimensional graph with the absorption value on the vertical axis and the corresponding wavenumber on the horizontal axis of the spectrum. In an alternative graphical representation, the graph may be a scatter plot. As can be seen in the drawing, the absorption values at the same wavenumber differ slightly between instruments. This leads to differences in the amount of components of the same milk sample determined chemometrically based on these absorption spectra in a known manner in each of the instruments. By applying prediction models for fat, protein, lactose, total solids (TS), and solids non-fat (SNF) to uncorrected spectra for a set, in this case fifteen standard milk samples, the following Root Mean Square Error (RMSEP) between instruments and their overall average - results in g/100 ml:

Fat Protein Lactose T.S. SNF Device 1 0.147 0.132 0.188 0.479 0.349 Device 2 0.134 0.141 0.195 0.490 0.352 Device 3 0.040 0.041 0.068 0.146 0.117 Device 4 0.320 0.313 0.450 1.115 0.818 Average 0.190 0.185 0.265 0.659 0.482 Relative error* 5.30% 5.08% 5.28% 5.11% 5.15% *Average error relative to the average concentration for the parameter.
Table 1: Prediction for milk components using corrected spectra for each instrument

This level of performance is unacceptable for a user who requires essentially the same results from different instruments.

[0041] Following the method of the present invention, a single-beam spectrum of water (SB _Z ) is obtained on each of the four instruments (see FIG. 7) and the first derivative of the absorption spectrum is determined at computer 510 for each of these single-beam spectra, standardized along the wavenumber axis. . These first derivatives are shown in Fig. 4, which also shows (by dotted lines) two areas for which it is necessary to calculate the difference Δ(SB _Z ) ^/ . As can be seen in FIG. 7 and fig. 4, the Y-axis values for each wavenumber are slightly different for each of the four instruments 100. The recorded single-beam spectrum is processed by the computer 500 of the instrument 100, which generates a corresponding spectrum to determine the humidity correction, _corr. , for each instrument, based on the intensity values in the two regions shown by the dashed lines in FIG. 5. Then, in each device, the corresponding computer 500 calculates the intensity correction factor I _corr specific for each of the four devices.

Device Intensity correction ( I _corr ) 1 0.9979 2 0.9953 3 0.9740 4 0.8823 Table 2: Intensity correction factors for each device

[0042] The correction factor _Icor for each instrument can then be applied to the absorbance data collected by each instrument (e.g., absorbance data for the same milk sample shown in FIG. 6) and corrected (or standardized to y-axis) of the absorbance data for each instrument to provide even better consistency. This is illustrated in FIG. 8, which shows the absorption spectra for milk shown in FIG. 6, but each of them is corrected using the corresponding intensity correction factor, I _corr , for the device with which the corresponding spectrum of milk was recorded (see table 2) .

[0043] By applying this correction to a set of, in this case fifteen, standard milk samples, the following inter-instrument root mean square errors (RMSEP) and their overall average are obtained when performing predictions for fat, protein, lactose, total solids (TS ) and solids free of fat (SNF) - results in g/100 ml:

Fat Protein Lactose T.S. SNF Device 1 0.007 0.011 0.006 0.013 0.010 Device 2 0.007 0.010 0.017 0.034 0.022 Device 3 0.016 0.014 0.008 0.041 0.021 Device 4 0.012 0.017 0.010 0.022 0.012 Average 0.011 0.014 0.011 0.030 0.017 Relative error* 0.32% 0.39% 0.23% 0.24% 0.19% *Average error relative to the average concentration for the parameter.
Table 3: Prediction for milk components using corrected spectra for each instrument

These errors are very small compared to the typical prediction error of 0.8% with respect to the values of the indicated components determined by known wet chemical methods.

[0044] It will be apparent to one skilled in the art that the application of the method in accordance with the present invention allows for greater standardization of output signals between different devices 100 of the same type, as well as replacing the standardization of output signals of the same device 100 as the length of the optical paths through ditch 620.

[0045] As will also be apparent to one skilled in the art, instead of expressing spectral information about electromagnetic radiation using wavenumber, wavelength or frequency can be used without departing from the claimed invention.

Claims (6)

1. A method for correcting changes in the amplitude of the output signal of a spectrometer instrument (300) due to changes in the length (L) of an optical path through a sample holder (600), comprising exposing an unknown sample in the sample holder (600) to electromagnetic radiation of multiple wave numbers; determining, using a spectrometric device (300), the absorption intensities of electromagnetic radiation in an unknown sample with a plurality of wave numbers; providing to a computer device (510) coupled to the spectrometer instrument (300) determined wavenumber-indexed absorption intensities as spectral data; generating, using a computer device (510), absorption values relative to the zero material based on the spectral data; and applying, by a computer device (510), a mathematical transformation ( _Icorr ) to the generated absorption values to correct changes in the amplitude of the output signal of the spectrometer instrument (300); wherein the method further includes calculating on the computer device (510) a mathematical transformation (I _corr ) by determining the difference (Δ(SB _Z ) ^/ ) between the first derivative of the logarithmically transformed spectral data (SB _Z ) for a sample of zero material in the first range (x ₁ ) wavenumbers and the first derivative of logarithmically transformed spectral data (SB _Z ) for a sample of zero material in the second range (x ₂ ) of wavenumbers; and calculating the mathematical transformation (I _corr ) as a function inversely dependent on a certain difference (Δ(SB _Z ) ^/ ). 2. The method according to claim 1, according to which the calculation of the mathematical transformation (I _corr ) also includes correction for air humidity in the light path between the emitting device (200) and the detector (400) of the spectrometric device (300) by determining the humidity correction factor (corr). _moisture ) as slope/intercept in selected moisture regions in spectral data (SB _Z ) for a sample of zero material; and calculating the mathematical transformation ( _Icorr ) as a function inversely dependent on the sum of a certain difference (Δ(SB _Z ) ^/ ) and the humidity correction factor (corr _{. humidity} ). 3. The method of claim 1, wherein the sample of zero material contains a sample of nominally pure water. 4. The method according to claim 1, according to which the absorption intensity values are determined using Fourier transform spectroscopy. 5. The method of claim 1, wherein the spectral data for the null material (SB _Z ) comprises spectral data standardized by wavenumber. 6. An apparatus (100) comprising a spectrometric instrument (300) configured to perform Fourier transform spectroscopy on a sample in a sample holder (600) and an associated computing device (510) configured to receive spectral data from the spectrometric instrument (300) and applying to them a mathematical transformation stored in a storage device accessible to the associated computing device (510), which corrects the transformation of the spectral data to change the amplitude based on the absorption intensity values determined by the spectrometer instrument (300); wherein the associated computing device (510) is programmed to operate such that the apparatus (100) performs the method according to any of the preceding paragraphs.