JP2741376B2 - Multi-component analysis method in spectroscopic analysis - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of irradiating a sample with light from a light source and analyzing the concentration of multiple components contained in the sample based on an absorption spectrum obtained at that time.

[0002]

2. Description of the Related Art At present, a method for analyzing the concentration of multiple components contained in a sample by performing a matrix operation on the basis of an absorption spectrum obtained using, for example, FTIR (Fourier transform infrared spectrometer) is known. Although several types are known, conventionally, any one of the methods is generally employed.

[0003]

By the way, in the exhaust gas of an automobile (hereinafter referred to as own exhaust gas), CO, CO ₂ ,
NO, H ₂ O, HC (hydrocarbon) and the like are contained as main components. Of these, CO, CO ₂ , NO, H ₂ O
Has a sharp and characteristic absorption, as is clear from the absorption spectra of FIGS. 8 to 11 because the molecular structure is simple. These components have a relatively high concentration in their own exhaust gas and therefore have a strong absorption. Also, 1800 ~
Focusing on the wave number range of 2400 cm ⁻¹ , there is almost no interference component for these components.

On the other hand, HC includes CH ₄ (methane), C ₂
It contains a very large number of compounds such as H ₆ (ethane) and C ₃ H ₈ (propane). 16 to 19 respectively show C
_{2 H 6, C 3 H 8} , n-C 4 H 10 ( butane),
are shown the absorption spectrum of the iso -C ₄ H ₁₀ (isobutane), as is clear from these figures, is that they are common to these coupling a large number of hydrocarbons, it has an absorption at approximately the same area In addition, the absorption itself is broad and similar to each other (however, CH ₄ and C ₂ H ₂ (acetylene) having relatively simple molecular structures have sharp peaks).

[0005] As described above, the main components contained in the self-exhaust gas have different absorption characteristics between CO, CO ₂ , NO, H ₂ O, and HC. _{2 H 6, C 3 H 8} , n-C 4 H 10, iso -C
₄ In such H ₁₀ and the simple CH _4, C ₂ H ₂ molecular structure, different absorption characteristics. Therefore, when one type of method is employed as in the related art, it is difficult to accurately quantify many components contained in the own exhaust gas.

SUMMARY OF THE INVENTION The present invention has been made in consideration of the above-mentioned problems, and has as its object to provide a multi-component analysis in a spectroscopic analysis capable of simultaneously and accurately and efficiently quantitatively analyzing multi-components. It is to provide a method.

[0007]

In order to achieve the above object, in the present invention, a sample is irradiated with light from a light source, and arithmetic processing is performed on the basis of an absorption spectrum obtained at that time. a method of simultaneously quantitatively analyzing the concentration of multi-component contained, the multi-component based on the wave number range used shape your Yopi quantification of the peak of the absorption grouped into a plurality of component groups, the information on each component group Set
The concentration of each component is determined using the most appropriate quantitative calculation method according to the absorption peak over the wave number range used for the amount .

[0008]

[Action] For example, when each quantitative analysis of components of the self exhaust gas, CO, CO 2, _NO, for components such as _H 2 _O, from where the absorption peak is sharp, the present applicant or October 1990 13 Patent application dated [Japanese Patent Application 2-
274204 (Japanese Patent Application Laid-Open No. 4-148830)] , a quantitative analysis method based on the concept of relative absorbance can be suitably applied. By using this method, the amount of memory used in a computer and the calculation can be reduced. The time required can be reduced as much as possible, and the components can be quantitatively determined with high accuracy.

[0009] Among the HCs, CH ₄ , C
Regarding HC except for ₂ H ₂ , since absorption itself is broad, PLS which has been generally used in the past has been used.
(Partial Least Square; Partial Least Square) method is used. According to this, compared to the relative absorbance method, the amount of memory used and the time required for calculation are increased, but each component can be quantified with high accuracy.

As described above, by adopting an appropriate quantification method for each component to be quantified, the accuracy of the entire analyzer can be improved, and multiple components can be simultaneously quantified with high accuracy. Then, it is possible to minimize the memory usage and the calculation speed on the computer.

[0011]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 shows an example of an apparatus for implementing a multi-component analysis method in spectroscopic analysis according to the present invention.
3 shows a schematic configuration of R1. This FTIR
Reference numeral 1 denotes an analyzer 2 which processes an interferogram which is an output of the analyzer 2 and a movable mirror 7 which will be described later.
And a computer 3 for issuing a control command to the drive mechanism.

The analyzing section 2 includes a light source 4 configured to emit parallel infrared light, a beam splitter 5, a fixed mirror 6, and a movable mirror 7 that moves in the X and Y directions by a driving mechanism (not shown). It comprises an interference mechanism 8, a cell 9 for accommodating a measurement sample and the like, and irradiated with infrared light from the light source 4 via the interference mechanism 8, and a detector 10 such as a semiconductor detector.

In the computer 3, an averaging processing unit 11 for averaging the interferogram output from the detector 10, and a high-speed Fourier transform processing for performing a high-speed Fourier transform on the output data from the averaging processing unit 11 A unit 12, a calculation unit 13 for performing a spectrum calculation and a quantitative calculation based on output data from the fast Fourier transform processing unit 12, and the like, and a control unit (not shown) are provided.

In the FTIR 1 configured as described above, the comparison sample or the measurement sample is stored in the cell 9 and the interferograms of the comparison sample and the measurement sample are measured. Each of these interferograms is Fourier transformed to obtain a power spectrum, that is, a spectrum of light transmitted through the cell 9. Next, the ratio of the power spectrum of the measurement sample to the power spectrum of the comparison sample is determined, and this value is converted to an absorbance scale to obtain an absorption spectrum.

Next, the details of the multi-component analysis method of the present invention will be described. Now, as an example, the resolution is 0.25 cm ^-1.
Consider the case of using FTIR1 to quantify multicomponents in own exhaust gas discharged from an LPG (liquefied petroleum gas) vehicle.
The main exhaust gas contains, as main components, CO, CO ₂ , NO, H ₂ O, and HC having 4 or less carbon atoms.
Then, CO, CO ₂ , NO, H
₂ O, NH ₃ , CH ₄ , C ₂ H ₄ (ethylene), C ₂ H
_{6, C} 3 _{H 6 (propylene), C 3 H 8, n} -C 4 H
_10, given the 12 components of the iso-C ₄ H _10, as already described, these components are based on the shape and the absorption region of the absorption peak, for example, four components as shown in the following Table 1 Can be grouped into groups.

[0016]

[Table 1]

The first component group in Table 1 is a component group having absorption in the wave number range of 1800 to 2400 cm ^-1 as shown in FIGS. 8 to 11. The components belonging to this component group have a characteristic and sharp absorption peak due to their simple molecular structure, and have a strong absorption due to the relatively high concentration of their own exhaust gas. Then, the second component group is, as shown in FIGS.
A component group also having a sharp absorption peak in the wave number range of cm ^-1 . Further, the third component group includes, as shown in FIGS.
The second in the wave number range of 800 to 1100 cm ^-1 which overlaps with the second component group
The component group has a broader absorption than the component group. In the wave number ranges of the second and third component groups, CO ₂ and H ₂ O in addition to these components also have absorption. The fourth component group is
As shown in FIGS. 16 to 19, the components are broad and have substantially similar absorption in the same wave number range of 2800 to 3100 cm ⁻¹ . In the wave number range of the fourth component group, H ₂ O,
CH ₄ , C ₂ H ₄ and C ₃ H ₆ also have absorption.

Since the first and second component groups have sharp absorption, the above-mentioned Japanese Patent Application No. 2-274204 is disclosed.
No. (JP-A-4-148830) .

First, the concept of the sum of relative absorbances introduced in this method will be briefly described.
The spectrum output from the spectrometer is generally
Noise from various sources,
The expression (1) can be expressed as follows: A (v _i ) = Cα (v _i ) + ε _i (2) Here, ε _i is noise included in the spectrum.

Then, focusing on the difference between any two points in the spectrum, the equation (2) can be expressed as follows: A (ν _p ) −A (ν _b ) = C [α (ν _p ) −α (ν _b )] + Ε _p −ε _b (3), indicating that the relative absorbance between such two points also obeys Lambert-Beer's law. here,
p and q are suffixes indicating a wave number point corresponding to a spectrum peak and a base, respectively.

In general, the spectrum of a substance, particularly the spectrum of a gas, contains a large number of peaks, and a large number of pairs of the wave point ν _p and the base wave point ν _b of a peak specific to a substance can be selected.

FIG. 2 shows an example of a typical gas absorption spectrum and the corresponding relative absorbances L ₁ , L ₂ , and L ₃ . Each relative absorbance L _1, L _2, L
₃ corresponds to the value of the above equation (3).

By taking the sum of such values, Σ _k [A (ν _p ) −A (ν _b )] _k = CΣ _k [α (ν _p ) −α (ν _b )] + Σ _k (ε _p −ε _b ) _k (4) This equation (4) also maintains Lambert-Beer's law. From equation (4), it can be seen that the second term on the right-hand side averages out the random noise lurking in the spectrum and cancels disturbances such as drift of the baseline.

Here, the value calculated by the above equation (4) is
Defined as MAS (Multiple Absorption Sum) value. That is, MAS≡Σ _k [A (ν _p ) −A (ν _b )] _k (5) In particular, the MAS value obtained for a reference spectrum of a single component and a known concentration is calculated using the M
AS _i .

FIG. 3 shows an example of specifying wavenumber points for three gas types included in the generalized component group. In the figure, curves I, II, and III represent reference spectra of gases X, Y, and Z, respectively. Then, a set of wave number points for gas _{X Φ X: (X p,} X b) k, the set of wave number points for gas _{Y Φ Y: (Y p,} Y b) k, the set of wave number points for gas Z [Phi _Z: (Z _p , Z _b ) _k are designated respectively. In specifying these peak-base pairs, it is preferable to obtain as large a relative absorbance as possible and not to be affected by mutual absorption.

Here, the story is generalized, and the above-mentioned set of wave point designation values for m components is represented by Φ _i (i = 1 to m).
And the peak of component i for one absorption spectrum
The operation of calculating the set of base pairs as in the above equation (4) is defined as the following equation. Σ _k [A (ν _p ) −A (ν _b )] ≡A ◎ Φ _i (6) where ◎ is a new operator, and equation (6) gives the MAS value.

Therefore, MAS _i = A ◎ Φ _i (7), and m MAS _i are calculated for m Φ _i . If MAS _i , which is a set of m numbers, is denoted by Ψ, on the other hand, this can be considered as a spectrum having m elements. Ψ = (MAS _i : i = 1 to m) (8)

The new spectrum region (component spectrum region) has a horizontal axis representing gas type i and a vertical axis representing the sum M of relative absorbances.
AS _i can be graphically displayed. Therefore, it can be considered that the equation (7) is obtained by converting the spectrum A into the component spectrum Ψ using the set Φ.
This will be described below by applying the above generalized component group.

FIG. 4 shows the three types of gases X, Y and Z shown in FIG.
Represents the absorption spectrum of those mixed at arbitrary concentrations, that is, the absorption spectrum of the generalized component group as a whole. By applying the wave point specification values of the peak-base pairs of the three gases specified in FIG. 3 to the absorption spectrum of FIG. 4, X,
The sum of the relative absorbances represented by Y and Z is determined.
Then, as shown in FIG. 5, the sum of these relative absorbances is plotted in a region where the horizontal axis represents the gas type and the vertical axis represents the sum of the relative absorbances.

As described above, the influence of noise is largely eliminated from the component spectrum obtained in this way, and the influence of interference between spectra between gas species is suppressed. As shown in FIGS. 4 and 5, if conversion without interference between components can be performed, the sum of the relative absorbances in the component spectrum region is proportional to the respective component concentrations. Obtainable.

However, in the actual gas absorption spectrum, it is rare that the wave point point designation without interference can be rarely performed as shown in FIGS. 4 and 5, and some interference remains even in the converted component spectrum region. Is common.

Therefore, in the quantitative analysis using the component spectrum region, it is necessary to correct the above-mentioned slightly remaining interference, and the following processing is performed in a calibration step of creating a matrix (calibration matrix) used for interference correction, that is, It is divided into a step of preparing data for interference correction and an estimation step of calculating unknown density using the data.

In the calibration stage, a calibration matrix is obtained as follows. Now, when generalizing to the case where there are reference spectra α _i (i = 1 to m) of unit concentration for m components, the above-mentioned set of wave number point designation values Φ _i is applied to these reference spectra, and by applying a transformation represented by 7), the component spectrum corresponding to α _i Ψ _{i (i}
= 1 to m). More specifically, it can be obtained by calculating the sum of relative absorbances when a plurality of gases are each used as a sample alone.

FIGS. 6A, 6B, and 6C show reference component spectra of three gas types X, Y, and Z included in the generalized component group. In this case, FIG. The calibration matrix is

[0035]

(Equation 1)

## EQU1 ##

On the other hand, the component spectrum of the component group can be obtained as shown in FIG. 7 from the absorption spectrum obtained by spectrally analyzing the sample to be measured in the wave number range assigned to the component group. The component spectrum Ψ _{u in} this case is

[0038]

(Equation 2)

## EQU1 ## In general, m gas components (i = 1
Assuming that the unknown concentration of each gas component is C ₁ (i = ₁ to m) in the mixed gas of で_u ), _{ｕ u} is represented by a linear combination of _{１ 1} . That is, _{ｕ u} = Ψ ₁ · C ₁ + Ψ ₂ · C ₂ +... + Ψ _m · C _m (9) always holds, and when this is rewritten using a matrix, Ψ _u = ΩC (10) It can be expressed as. Here, C is a vector of unknown concentration, Omega in matrix with rows [psi _1, which is referred to as a calibration matrix in the component spectral region.

Therefore, assuming that the unknown concentrations of the three gas species X, Y, and Z of the above-described component group are C _X , C _Y , and C _Z ,

[0041]

(Equation 3)

Is as follows. In the calibration step, the Ω needs to be determined accurately. Also, as already mentioned,
_{Since i} is a vector with high linear independence, that is, a vector with little interference, Ω is a matrix from which a stable inverse matrix can be obtained.

Next, in the estimation stage, the above equation (10) is expressed as C
By solving for, i.e., as C = Ω ^-1 Ψ _u, it is possible to estimate the unknown concentration.

Then, also in the above equation (11),

[0045]

(Equation 4)

Thus, C _X , C _Y , and C _Z can be obtained.

Since the calculated inverse matrix has a high linear independence of Ω, a stable solution can be obtained. Therefore, the error in the numerical calculation of the estimated concentration is extremely small.

As described above, the relative absorbance method is particularly suitable for the quantification of a component capable of obtaining a large relative absorbance by reducing the influence of absorption of other components, that is, a component having a sharp characteristic absorption peak. . Further, the quantitative calculation processing is relatively simple, the calculation speed is high, and the memory usage is small. While the relative absorbance method having such characteristics is to apply to the first and second group of components, in this case, the wave number range to be used for quantitation need not be the whole area of the infrared, for example, the first
Regarding the component group, 1800 to 2400 cm ⁻¹ , the second
The component group may have a limited range of 800 to 1400 cm ^-1 .

Further, as for the first component group, since there are no other interference components, the four components (CO, CO ₂ , NO, H
_In consideration of only ₂ O), a quantitative calculation may be performed in the wave number range. In addition, for the second component group, a quantitative calculation is performed in the wave number range in consideration of other components (H ₂ O, CO ₂ , C ₂ H ₄ , and C ₃ H ₆ ) that cause interference. in this way,
The components quantified by the same relative absorbance method are also the first component group and the second component group.
By dividing into component groups, the amount of memory used on the computer can be reduced, and quantitative calculation can be performed at higher speed. However, since the quantitative calculation method for the first component group and the second component group is the same, it is a matter of course that the quantitative determination may be performed for one component group.

Next, regarding the third and fourth component groups,
Since the absorption is broad and the wave number range and the shape of the absorption are similar to each other, the accuracy tends to be insufficient when the relative absorbance method is applied. Therefore, a technique called PLS method, which has been generally used, is used for these component groups. In this method, a matrix for concentration calculation is created in advance using an absorption spectrum having a known concentration of a compound to be quantified as a reference spectrum. Note that this PLS method is, for example, Prac
tical Fourier Transform Infrared Spectroscopy, Aca
demic Press, Lindberg, W., et al., Anal.Chem., 19
83, 55, 643, Haaland, DM, Thomas, EV, Anal.Che
m., 1988, 60, 1193; Geladi, P., Kowalsky, BR,
Anal.Chem. Acta., 1986, 185, Wold, S., et al., Mat.
It is described in detail in rix Pencil, Springer Verlag, etc.

The PLS method requires a larger amount of memory and a lower calculation speed than the relative absorbance method, but can quantitatively determine a broad absorption with relatively high accuracy. And even when using this PLS method, the wave number range used for quantification does not need to be the whole infrared region. For example, for the third component group, 800 to 1100 cm ⁻¹ , and for the fourth component group, A limited range of 2800-3100 cm ^-1 may be used. Also,
As for the third component group, CO ₂ ,
In consideration of H ₂ O and NH ₃ , a quantitative calculation is performed in the wave number range. Similarly, for the fourth component group, H ₂ O, C
In consideration of H ₄ , C ₂ H ₄ , and C ₃ H ₆ , a quantitative calculation is performed in the wave number range. In this manner, the components quantified by the same PLS method are also divided into the third component group and the fourth component group, so that the memory usage on the computer can be reduced and the quantitative calculation speed can be increased. However, since the third component group and the fourth component group have the same quantitative calculation method, it goes without saying that the quantification may be performed collectively in one component group.

As will be understood from the above description, when simultaneously quantifying multiple components, the components to be measured are divided into a plurality of groups, and groups which are easy to quantify (for example, the absorption peak is sharp or characteristic, For a group consisting of components where a certain region does not overlap with other components or the absorption intensity is strong, of the quantitative methods that provide sufficient accuracy, simple methods of arithmetic processing and memory Apply less method. Conversely, groups that are difficult to quantify (groups composed of components such as broad absorption with few features and overlapping of other components with the absorption region)
For, the most appropriate method is applied in terms of accuracy, even if some conditions (such as calculation speed and memory usage) are sacrificed.

In these cases, the wave numbers used in the calculation may overlap as in the above-described example. In addition, the type and number of the quantification method need not be limited. For example, CLS (Classical Least Squa
The re) method may be used. Furthermore, the present invention is not limited to gas analyzers, nor is it limited to analysis using FTIR.

[0054]

As described above, according to the present invention,
By employing an appropriate quantification method for each component to be quantified, the accuracy of the entire analyzer can be improved, and multiple components can be simultaneously quantified with high accuracy. Then, the memory and the calculation speed on the computer can be minimized.

[Brief description of the drawings]

FIG. 1 is a diagram schematically showing an example of an apparatus for carrying out the method of the present invention.

FIG. 2 is a diagram of a schematic absorption spectrum of a gas and a relative absorbance thereof for explaining a calculation method used in an embodiment of the method of the present invention.

FIG. 3 is a diagram showing an example of specifying a wave number point for a gas type by generalizing the component group in the embodiment.

FIG. 4 is a diagram showing an absorption spectrum of a generalized component group in the example.

FIG. 5 is a diagram showing an example of a component spectrum calculated from an absorption spectrum of a generalized component group in the embodiment.

FIG. 6 is a diagram showing a reference component spectrum corresponding to a generalized component group in the embodiment.

FIG. 7 is a diagram showing a component spectrum calculated from an absorption spectrum of a measurement sample in the example.

FIG. 8 is a diagram showing an absorption spectrum of CO.

FIG. 9 is a diagram showing an absorption spectrum of CO ₂ .

FIG. 10 is a diagram showing an absorption spectrum of NO.

FIG. 11 shows an absorption spectrum of H ₂ O.

FIG. 12 shows an absorption spectrum of NH ₃ .

FIG. 13 shows an absorption spectrum of CH ₄ .

FIG. 14 shows an absorption spectrum of C ₂ H ₄ .

FIG. 15 shows an absorption spectrum of C ₃ H ₆ .

FIG. 16 shows an absorption spectrum of C ₂ H ₆ .

FIG. 17 shows an absorption spectrum of C ₃ H ₈ .

FIG. 18 is a diagram showing an absorption spectrum of nC ₄ H ₁₀ .

FIG. 19 shows an absorption spectrum of iso-C ₄ H ₁₀ .

[Explanation of symbols]

 1 FTIR, 2 analysis part, 4 light source.

 ──────────────────────────────────────────────────続 き Continuation of the front page (56) References JP-A-54-84781 (JP, A) JP-A-4-60442 (JP, A) JP-A-64-73239 (JP, A) JP-A-3- 114441 (JP, A)

Claims (1)

(57) [Claims] 1. A method for irradiating a sample with light from a light source and performing an arithmetic process based on an absorption spectrum obtained at that time to simultaneously quantitatively analyze the concentrations of multiple components contained in the sample. based on the wave number range to use components in the shape and quantification of the peak of the absorption grouped into a plurality of component groups, used in the quantification for each component group
A multi-component analysis method in spectral analysis, wherein the concentration of each component is determined using the most appropriate quantitative calculation method according to the absorption peak over a range of wave numbers .