CN104502495A - Method for extracting pure spectrum of gas chromatography-mass spectroscopy - Google Patents

Abstract

The invention provides a method for extracting a pure spectrum of gas chromatography-mass spectroscopy. The method comprises the following steps: expressing original GC-MS data into a matrix form; subtracting adjacent scanning mass spectrums in the matrix, and respectively saving obtained absolute values of the positive values and the negative values in individual data sets; respectively shifting the two data sets obtained by subtracting, and adding for combining the two data sets to obtain a matrix again; obtaining a matrix with relatively pure spectrum data by utilizing the newly synthesized matrix; repeating the previous steps for several times, thereby obtaining a matrix of the pure spectrum data; finally, performing data reduction on the matrix obtained by performing difference operation twice, and merging to form another novel matrix, wherein the line number of the another novel matrix is two times that of the line number of the previous matrix; and performing matrix reconstruction on the merged matrix in the previous step, solving the matrix representing the pure composition spectrum data, and finishing the extraction of the pure spectrum. According to the method disclosed by the invention, the resolution ratio of the spectrum of the gas chromatography-mass spectroscopy can be improved.

Description

The pure spectrogram extracting method of a kind of gaschromatographic mass spectrometry

Technical field

The present invention relates to a kind of gas chromatograph-mass spectrometer (GCMS).

Background technology

Gas chromatograph-mass spectrometer (GCMS) is widely used in every field as a kind of analytical instrument, its data handling system as its important component part, for extracting pure mass spectrogram and qualitative analysis.

Blend sample is analyzed through gas chromatography-mass spectrography system (GC-MS), the material that wherein retention time is close can not be completely segregated after chromatogram, cause chromatographic peak overlapping, the mass spectrogram obtained can comprise the ion of common effluent, also comprise some ions produced from pollution of ion source thing, column bleed thing, matrix interference thing simultaneously, cause the accuracy rate of qualitative analysis and reliability to reduce the most at last.The compound that nature is known simultaneously has hundreds thousand of kinds, and each analysis design mothod all can produce a large amount of data, if only adopt artificial method to identify spectrogram and qualitative analysis more difficult, be necessary adopt computer assisted form to extract pure mass spectrogram and carry out qualitative analysis.GC-MS data handling system is the indispensable pith of GC-MS.

Summary of the invention

The object of the invention is to propose one and can put forward high-resolution gas chromatograph-mass spectrometer (GCMS) spectrogram disposal route, to improve the ability of its qualitative analysis.Technical scheme of the present invention is as follows:

The pure spectrogram extracting method of a kind of gaschromatographic mass spectrometry, comprises the following steps:

(1) original GC-MS data are expressed as matrix form;

(2) scanning mass spectrogram adjacent in matrix is subtracted each other, be kept at independent data set by what obtain respectively on the occasion of the absolute value with negative value;

(3) carry out shifting function new addition of laying equal stress on be combined into a matrix to subtracting each other two data sets obtained respectively;

(4) matrix of new synthesis is utilized to obtain the matrix of comparatively pure mass spectrometric data;

(5) repeat above-mentioned step (2) to (5) several times, obtain the matrix of more pure mass spectrometric data;

(6) data-reduction is carried out to the matrix that last twice calculus of differences obtains, and then the new matrix of composition after merging, the line number of new matrix be before the twice of line number of matrix;

(7) matrix reconstruction is carried out to the matrix after the merging of (6) step, try to achieve the matrix representing pure composition mass spectrometric data, complete the extraction of pure spectrogram.

Accompanying drawing explanation

The pure mass spectrogram of Fig. 1 composition A and composition B, (a) is composition A, and (b) is composition B.

Fig. 2 (a) is original total ion current figure, and (b) is mass spectrogram corresponding to maximal value place.

Fig. 3 (a) is the total ion current figure with the signal increased sweep time, and Fig. 3 (b) and (c) are respectively composition A and the composition B mass spectrogram corresponding at chromatographic peak maximal value place.

Fig. 4 (a) is the total ion current figure with the signal reduced sweep time, and Fig. 4 (b) and (c) are respectively composition A and the composition B mass spectrogram corresponding at chromatographic peak maximal value place.

The total ion current figure synthesized after the difference GC-MS method that Fig. 5 (a) adopts the present invention to propose, Fig. 4 (b) and (c) are respectively composition A and the composition B mass spectrogram corresponding at chromatographic peak maximal value place.

The mass spectrogram at retention time 12.6039 place in Fig. 6 AMDIS result.

The mass spectrogram at retention time 12.606 place in the self-editing arithmetic result of Fig. 7.

The mass spectrogram at retention time 15.6339 place in Fig. 8 AMDIS result.

The mass spectrogram at retention time 15.6355 place in the self-editing arithmetic result of Fig. 9.

Embodiment

In desirable chromatogram, the chromatographic peak of often kind of composition should be the form of bar graph, but moves in chromatographic column due to often kind of composition and occupy certain space, and the chromatographic peak of the actual often kind of composition obtained all has certain width.If containing the composition that retention time is close in tested blend sample, then compose summit in total ion current figure and overlap, a kind of mass spectrogram of composition can introduce the interference of other component ion, finally affect the result of qualitative analysis.In order to reduce the peak width of chromatographic peak and improve the resolution of chromatogram, the present invention proposes a kind of method based on difference GC-MS.

First the matrix S of the raw data of GC-MS with a R × N represent by the present invention, the mass spectrometric data obtained of each scanning of every a line representative of matrix S, its each row representative be mass chromatography data corresponding to some mass numbers.

The present invention adopts a kind of difference GC-MS to carry out the process of chromatogram, mainly comprise two steps, first the mass spectrometric data of adjacent for GC-MS scanning is subtracted each other, namely the adjacent rows respective value of matrix S is subtracted each other, the difference obtained is saved in two data set U and V by positive and negative respectively, and what wherein matrix V was preserved is the absolute value of negative difference.

$\begin{array}{cc}U=\left[\begin{array}{cccc}{u}_{11}& u_{12}& ...& u_{1N}\\ u_{21}& u_{22}& ...& u_{2N}\\ ...& ...& ...& ...\\ u_{R1}& u_{R2}& ...& u_{\mathrm{RN}}\end{array}\right]& V=\left[\begin{array}{cccc}{v}_{11}& v_{12}& ...& v_{1N}\\ v_{21}& v_{22}& ...& v_{2N}\\ ...& ...& ...& ...\\ v_{R1}& v_{R2}& ...& v_{\mathrm{RN}}\end{array}\right]\end{array}---\left(1\right)$

In formula, u _{(i-1) j}=s _ij-s _{(i-1) j}, i=2,3, R, j=1,2, N is (if s _ij-s _{(i-1) j}> 0)).V _{(i-1) j}=| s _ij-s _{(i-1) j}|, i=2,3, R, j=1,2, N is (if s _ij-s _{(i-1) j}< 0)), s _ijit is the value of matrix S i-th row jth row.Other elements in U and V are zero.Positive differential data shows that signal increases in time, and negative differential data then shows that signal reduces in time.

The second step of this algorithm is that matrix U and V are reassembled into a matrix.Concrete operations are, U and V is shifted with behavior unit, and object is that to make the chromatographic peak belonging to same composition reach sweep time of maximal value identical.Matrix U is shifted downwards, V upward displacement, and the size of shift value depends on the peak width of chromatographic peak.Then, by the matrix U after displacement and V summation, a new data set S is obtained ₁.New data matrix comprises the information the same with raw data, and chromatographic resolution rate is significantly improved.

Because of the new data matrix S obtained by this algorithm ₁also can regard common raw data as, therefore repeatedly can apply the new data of difference algorithm process, comprise difference and displacement summation, such chromatographic resolution rate has further raising.

Be described in detail below in conjunction with drawings and Examples.

In emulation experiment, the GC-MS packet of simulation containing two kinds of composition A and B, the mass spectrometric data of these two kinds of compositions from NIST standard mass spectrometric data storehouse, as shown in Figure 1.

Suppose retention time difference Δ t=8s (scanning times) of two kinds of compositions.The concentration be modeled to point by Gauss's peak shape is with the change curve of sweep time, the concentration of two kinds of compositions reaches time phase difference 8 scanning of maximal value, as can be seen from Fig. 2 (a), in total ion current figure, the chromatogram peak complete-superposing of two kinds of compositions, mass spectrogram corresponding to chromatographic peak maximal value place is the superposition of composition A and composition B mass spectrogram, as Fig. 2 (b).

Adopt difference GC-MS algorithm for above-mentioned simulated data, the total ion current figure that the first step of this algorithm obtains and mass spectrogram corresponding to peak value maximum obtain result as shown in Figure 3.

Total ion current figure in Fig. 3 (a) represents the signal increased with sweep time, namely corresponds to and is added corresponding for all row of matrix U (as Suo Shi equation (1)) data obtained.As can be seen from total ion current figure, the chromatographic peak of composition A and composition B separates significantly, and mass spectrogram corresponding to chromatographic peak maximal value place is respectively the mass spectrogram of composition A and composition B, as shown in Fig. 3 (b) He 3 (c).

Total ion current figure in Fig. 4 (a) represents the signal reduced with sweep time, namely corresponds to and is added corresponding for all row of matrix V (as Suo Shi equation (1)) data obtained.As can be seen from total ion current figure, the chromatographic peak of composition A and composition B separates significantly, and mass spectrogram corresponding to chromatographic peak maximal value place is respectively the mass spectrogram of composition A and composition B, as shown in Fig. 4 (b) He 4 (c).

The second step of difference algorithm aligns value difference divided data matrix U and negative value differential data matrix V is shifted and reconfigures, and move to right by Fig. 3 (a), Fig. 4 (a) moves to left, then by the chart adding after displacement.The size of shift value equals the half of the difference of substance A number of scans and the difference mean value of substance B number of scans in Fig. 3 (a) and Fig. 4 (a), and value is 4, and object makes the chromatographic peak of same material reach maximal value at one time.The result that this step operation obtains as shown in Figure 5.

As can be seen from positive and negative difference total ion current figure and the total ion current figure reconfigured, the chromatographic peak of A and B two kinds of compositions separates significantly, obtains cleaner mass spectrogram.The GC-MS data obtained by difference algorithm regard original GC-MS data as, can repeat the step of above-mentioned difference GC-MS algorithm, obtain the new spectrogram after second order difference GC-MS algorithm.

Mass spectrogram above-mentioned steps obtained compares with standard spectrogram in NIST storehouse, and result for retrieval is as shown in table 1.

Table 1 result for retrieval

As can be seen from Table 1, difference GC-MS algorithm can not only by the chromatographic peak of composition A and B in original total ion current figure separately, and the matching degree of the mass spectrogram obtained and standard spectrogram also increases, and improves the reliability of retrieval.

Below, based on the new data obtained for above-mentioned difference GC-MS algorithm again through a series of matrix operation, the matrix representing pure mass spectrometric data is tried to achieve.

Original GC-MS tables of data is shown as matrix form as follows:

S _R×N＝C _R×KF _K×KΔ′ _K×N(2)

In formula, matrix S _{r × N}represent original GC-MS data, Matrix C _{r × K}comprise the chromatogram of normalized pure composition, its each row are normalized to unit area, matrix F _{k × K}be a diagonal matrix, be used for quantizing often kind of composition, matrix Δ ' _{k × N}comprise the mass spectrum of normalized pure composition, its every a line is normalized to unit strength.

For matrix S _{r × N}apply above-mentioned difference GC-MS algorithm, namely adjacent mass spectrogram subtracts each other, and then is shifted to differential data and sues for peace, and obtains matrix for matrix repeat above-mentioned steps, obtain matrix by that analogy, after applying n difference GC-MS algorithm, matrix is obtained after n difference GC-MS algorithm, obtain:

$S_{R\&times;N}^n=C_{R\&times;K}^n{F}_{K\&times;K}^n{\mathrm{\&Delta;}}_{K\&times;N}^{\&prime;}---\left(4\right)$

Matrix and matrix respectively to Matrix C _{r × K}and matrix F _{k × K}the matrix obtained after carrying out n difference GC-MS algorithm.The mass spectrometric data matrix Δ of pure composition ' _{k × N}by the impact of difference GC-MS algorithm, after n difference algorithm be still matrix Δ ' _{k × N}.

As n=1, formula (4) can be write as following form.

$S_{R\&times;N}^1=C_{R\&times;K}^1{F}_{K\&times;K}^1{\mathrm{\&Delta;}}_{K\&times;N}^{\&prime;}---\left(5\right)$

Matrix with matrix Δ ' _{k × N}product can be written as following form.

$F_{K\&times;K}^1{\mathrm{\&Delta;}}_{K\&times;N}^{\&prime;}={\left(C_{R\&times;K}^1\right)}^{-1}{S}_{R\&times;N}^1---\left(6\right)$

In formula, matrix it is matrix inverse matrix.The approximate matrix A rebuild with one _{r × K}replace matrix so equation (4-6) can be write as following form.

$F_{K\&times;K}^1{\mathrm{\&Delta;}}_{K\&times;N}^{\&prime;}={\left(A_{R\&times;K}^T{A}_{R\&times;K}\right)}^{-1}{A}_{R\&times;K}^T{S}_{R\&times;N}^1---\left(7\right)$

The mass spectrometric data matrix Δ of pure composition ' _{k × N}by to matrix product Δ ' _{k × N}normalization and obtain.Equation (4) is better than equation (2), main because at data matrix in, eliminate the interference of ground unrest, and chromatographic resolution rate is improved significantly.In equation (5), get n=1, mainly because after application first difference GC-MS algorithm, chromatographic peak type still can be approximated to be Gaussian peak type.

Before solve equation (7), need reconstruction matrix A _{r × K}.Matrix A _{r × K}process of reconstruction be divided into two steps, the first step is composition detection, and second step is the chromatogram rebuilding it for each composition of detecting.

1, composition detection

What the algorithm of composition detection was mainly applied is matrix character during convergence, the From Spectral Signal that the first often plants composition mainly concentrates on matrix two row in and matrix two row in, and the process of difference GC-MS algorithm replaces, namely $S_{R\&times;N}^{n-1}=S_{R\&times;N}^{n-3}.$

Matrix the another one character of convergence is at matrix or matrix in, in the spectrogram of adjacent scanning, intensity ratio is greater than assuming that at matrix in the intensity level of adjacent scanning corresponding to same mass number be x and y and x > y, so exist the intensity level of middle correspondence will be x+y and x-y.When the ratio of x and y is equal with the ratio of x+y and x-y, equation is shown below.

$\frac{x}{y}=\frac{x+y}{x-y}---\left(8\right)$

Can be released by formula (8)

$x=(1+\sqrt{2})y,x=(1-\sqrt{2})y---\left(9\right)$

Because intensity level can not be negative value, therefore can obtain namely mean at matrix in or matrix in, the intensity level of adjacent scanning is greater than

Matrix and matrix in the information that comprises be redundancy, therefore, under the prerequisite of not loss of information, need matrix and matrix carry out data-reduction.First search in the nonzero value element of two matrixes, if their element s _i,jand s _{(i+1), j}ratio exceed so only retain larger element, other element is all set to zero.After data-reduction, matrix and matrix comprise complementary information, then will with be combined into a matrix R _{2R × N}, matrix R _{2R × N}line number be 2 × R, shown in (10).

$R_{2R\&times;N}=\left[\begin{array}{cccc}{r}_{11}& r_{12}& ...& r_{1N}\\ r_{21}& r_{22}& ...& r_{2N}\\ ...& ...& ...& ...\\ r_{2R,1}& r_{2R,2}& ...& r_{2R,N}\end{array}\right]---\left(10\right)$

If matrix R _{2R × N}the all elements sum of the i-th row and the number of all nonzero elements of this line is all greater than the minimum value of setting, so think to there is a kind of composition at place's sweep time that this line is corresponding.

2, the reconstruction of composition chromatogram

Original GC-MS data matrix S _{r × N}the matrix obtained after first difference GC-MS algorithm be expressed as form.

$S_{R\&times;N}^1=\left[\begin{array}{cccc}{s}_{11}^1& s_{12}^1& ...& s_{1N}^1\\ s_{21}^1& s_{22}^1& ...& s_{2N}^1\\ ...& ...& ...& ...\\ s_{R1}^1& s_{R2}^1& ...& s_{\mathrm{RN}}^1\end{array}\right]---\left(11\right)$

The data matrix of corresponding total ion current figure is expressed as form.

$S^{\mathrm{TIC}\left(1\right)}={(a_1^1,a_2^1,a_3^1,...,a_R^1)}^T---\left(12\right)$

In formula, $a_i^1={\mathrm{\&Sigma;}}_{j=1}^N{s}_{\mathrm{ij}}^1,i=\mathrm{1,2},\&CenterDot;\&CenterDot;\&CenterDot;R.$

Matrix A _{r × K}reconstruction be by S ^{tIC (1)}in data carry out that matching completes.First the quantity supposing the pure composition that first step composition detection detects is k, and the scanning position residing for often kind of composition is t ₁, t ₂... t _k.Total ionic strength adjustment buffer angle value corresponding to scanning place of composition is planted to i-th (i=1 ~ k) and its adjacent scanning intensity carry out Gauss curve fitting, namely the data obtained form matrix A _{r × K}i-th row.Successively aforesaid operations is carried out to the composition detected in k, finally can obtain matrix A _{r × K}k column data, complete matrix A _{r × K}reconstruction.

At reconstruction matrix A _{r × K}afterwards, by equation (7) can try to achieve the matrix Δ that represents pure composition mass spectrometric data ' _{k × N}, complete the extraction of pure spectrogram.

Experimental verification is carried out to the method that the present invention proposes below, the key instrument that experiment adopts: gas chromatograph-mass spectrometer (GCMS), U.S. Agilent Products, model is 7890A/5975C.Test specimen: DDV STD.

After test specimen passes through the pure spectrogram extraction algorithm process of business software AMDIS deconvolution process and the present invention's proposition, the retention time of the compound of acquisition is to such as shown in table 2 and table 3.Potpourri comprises nine kinds of compositions altogether, and table 2 lists the pure composition that the raw data that collects extracts after business software AMDIS process, and table 3 item lists the result of self-editing pure spectrogram extraction algorithm.

Table 2AMDIS result

The self-editing algorithm process result of table 3

As can be seen from table 2 and table 3, self-editing pure spectrogram extraction algorithm can well by pure constituents extraction out, the result extracted with business software AMDIS is basically identical, although both have fine distinction in retention time, but the retention time of total ion current figure only has relative meaning, do not affect the extraction of pure spectrogram and the qualitative analysis of composition.

Fig. 6 and Fig. 7 compared for the Pyridinium that AMDIS and self-editing algorithm extract, 1-(carboxymethyl)-, hydroxide, the mass spectrogram of inner salt, and Fig. 8 and Fig. 9 compares the Carbonic acid that AMDIS and self-editing algorithm extract, the mass spectrogram of 1-methylethyl phenyl ester.By the comparison to spectrogram, the pure mass spectrogram that AMDIS and self-editing algorithm extract is basically identical.

The result extracted by contrast AMDIS and self-editing algorithm and spectrogram, both results have good consistance, and self-editing spectrogram extraction algorithm can reach the level of AMDIS substantially, can obtain cleaner spectrogram to realize reliable qualitative analysis.

Claims (1)

1. the pure spectrogram extracting method of gaschromatographic mass spectrometry, comprises the following steps:

(1) original GC-MS data are expressed as matrix form;

(2) scanning mass spectrogram adjacent in matrix is subtracted each other, be kept at independent data set by what obtain respectively on the occasion of the absolute value with negative value;

(3) carry out shifting function new addition of laying equal stress on be combined into a matrix to subtracting each other two data sets obtained respectively;

(4) matrix of new synthesis is utilized to obtain the matrix of comparatively pure mass spectrometric data;

(5) repeat above-mentioned step (2) to (5) several times, obtain the matrix of more pure mass spectrometric data;

(6) data-reduction is carried out to the matrix that last twice calculus of differences obtains, and then the new matrix of composition after merging, the line number of new matrix be before the twice of line number of matrix;

(7) matrix reconstruction is carried out to the matrix after the merging of (6) step, try to achieve the matrix representing pure composition mass spectrometric data, complete the extraction of pure spectrogram.