Method for predicting interlayer spacing size mode of layered bimetal oxide by non-chemical experimental method
阅读说明：本技术 非化学实验方法预测层状双金属氧化物层间距大小模式的方法 (Method for predicting interlayer spacing size mode of layered bimetal oxide by non-chemical experimental method ) 是由 刘太行 刘振昌 刘太昂 周晶晶 周央 吴治富 朱峰 刘婷婷 朱鲁阳 刘远 于 2021-09-17 设计创作，主要内容包括：本发明的目的就是为了克服化学方法检测层状双金属氧化物层间距大小模式存在的缺陷,而提供一种低成本、无污染、测试简单、简便快捷的非化学实验方法预测层状双金属氧化物层间距大小模式的方法。整个过程如下：以原子参数为自变量,以层间距大小模式为因变量,构成基础数据；对基础数据中的原子参数数据进行组合转换,得到的综合变量,综合变量和层间距大小模式构成了中间数据；基于中间数据,利用人工神经网络算法建立层状双金属氧化物层间距大小模式的快速识别模型,该模型可以预测新的层状双金属氧化物层间距大小模式。(The invention aims to overcome the defects of the mode of detecting the interlayer spacing of the layered bimetal oxide by a chemical method, and provides a method for predicting the mode of the interlayer spacing of the layered bimetal oxide by a non-chemical experimental method, which has the advantages of low cost, no pollution, simple test, convenience and quickness. The whole process is as follows: forming basic data by taking the atom parameters as independent variables and taking the interlayer spacing size mode as dependent variables; atom parameter data in the basic data are subjected to combined conversion to obtain comprehensive variables, and the comprehensive variables and the interlayer spacing size mode form intermediate data; based on the intermediate data, a rapid identification model of the size mode of the interlayer distance of the layered bimetal oxide is established by utilizing an artificial neural network algorithm, and the model can predict a new size mode of the interlayer distance of the layered bimetal oxide.)
1. A method for predicting the size mode of the interlayer spacing of the layered bimetal oxide by a non-chemical experimental method comprises the following steps:
1) collecting a plurality of layered double hydroxides and interlayer spacing data thereof, defining the interlayer spacing larger than 7.7 as a large interlayer spacing mode and the interlayer spacing smaller than 7.7 as a small interlayer spacing mode according to the numerical values of the layered double hydroxides, collecting atomic parameters of metal atoms in the layered double hydroxides by combining a periodic table of elements, and forming basic data by taking the atomic parameters as independent variables and the interlayer spacing size mode as dependent variables; 2) carrying out combination conversion on atom parameter data in the basic data to obtain comprehensive variables, wherein the comprehensive variables are linear combinations of the atom data, and the comprehensive variables and the interlayer spacing size mode form intermediate data; 3) establishing a rapid identification model of the interlayer spacing size mode of the layered bimetal oxide by utilizing an artificial neural network algorithm based on the intermediate data; 4) collecting new layered bimetal oxide, searching to obtain atomic parameters of the layered bimetal oxide, substituting the parameters into a combined conversion equation, calculating intermediate data of the layered bimetal oxide, substituting the intermediate data into the established rapid identification model of the interlayer spacing size mode of the layered bimetal oxide, and predicting whether the interlayer spacing of the new layered bimetal oxide is a large mode or a small mode.
2. The method for predicting the interlayer spacing size mode of the layered double-metal oxide according to claim 1, wherein the step 2) combines the following transformation equations:
Y1= +3.034[na]-3.328[nb]-3.159[zz/rz]+5.329[ra/rb]-4.945
Y2= -3.839[na]-4.321[nb]-2.331[zz/rz]-6.884[ra/rb]+14.891
Y3= -6.894[na]-11.039[nb]-0.217[zz/rz]-0.215[ra/rb]+8.605
Y4= -18.703[na]-13.466[nb]-1.508[zz/rz]+3.892[ra/rb]+13.625。
Technical Field
The invention relates to the technical field of inorganic material testing, in particular to a method for predicting a size mode of a layered double-metal oxide interlayer spacing by a non-chemical experimental method.
Background
Layered Double Hydroxides (LDHs) are typical anionic clays, are also called hydrotalcite-like compounds, and refer to Hydroxides with hydrotalcite Layered crystal structures, which are composed of two or more metal elements. As a class of host-guest compounds with special structures and functions, LDHs have been the focus of attention in the field of inorganic functional materials in recent years.
LDHs are compounds of the general formula M2+ 1-xM3+ x(OH)2(An-)n/x∙ mH2O, a layered novel functional inorganic material. Wherein:
（1） M2+、M3+respectively representing divalent and trivalent metal cations located on the laminate
(2) X = M in the structural formula3+/(M2++M3+) The x value directly influences the composition and structure of the product
（3） An-Is an interlayer anion
(4) m represents the mole number of free water between layers
The schematic structure of the layered double hydroxide is shown in FIG. 1. The general formula of the LDHs interlamellar spacing is:
dspacing = dlayer + dinter
wherein d islayerRepresents the ply spacing, dinterRepresenting the inter-layer channel height. In general, LDHs with interlayer spacings greater than 7.7 are large interlayer spacing modes and less than 7.7 are small interlayer spacing modes. At present, the size of the interlayer spacing of the LDHs compound can be obtained by measuring through an X-ray diffractometer (XRD), but the measurement of the interlayer spacing of the layered double-metal oxide through the X-ray diffractometer needs overlong analysis period and the analysis process is complex. Because of these limitations, more and more technologists are studying and developing rapid, chemical-intensive methods for predicting the size patterns of the interlayer spacing of layered double-metal oxides.
The artificial neural network is an information processing system for simulating a biological neural network information processing mechanism, is particularly suitable for processing nonlinear data with complex input and output relations, and is an effective means for summarizing rules from chemical production and experimental data. The patent does not carry out chemical experiments, and utilizes microscopic atomic parameters to combine with an artificial neural network algorithm to determine the size mode of the interlayer spacing of the layered bimetal oxide. The method has the advantages of no need of chemical experiments, rapidness, convenience and low cost.
Disclosure of Invention
The invention aims to overcome the defects of the mode of detecting the interlayer spacing of the layered bimetal oxide by a chemical method, and provides a method for predicting the mode of the interlayer spacing of the layered bimetal oxide by a non-chemical experimental method, which has the advantages of low cost, no pollution, simple test, convenience and quickness.
The purpose of the invention can be realized by the following technical scheme:
a method for predicting the size mode of the interlayer spacing of the layered bimetal oxide by a non-chemical experimental method comprises the following steps:
1) collecting a plurality of layered double hydroxides and interlayer spacing data thereof, defining the interlayer spacing larger than 7.7 as a large interlayer spacing mode and the interlayer spacing smaller than 7.7 as a small interlayer spacing mode according to the numerical values of the layered double hydroxides, collecting atomic parameters of metal atoms in the layered double hydroxides by combining a periodic table of elements, and forming basic data by taking the atomic parameters as independent variables and the interlayer spacing size mode as dependent variables;
2) carrying out combination conversion on atom parameter data in the basic data to obtain comprehensive variables, wherein the comprehensive variables are linear combinations of the atom data, and the comprehensive variables and the interlayer spacing size mode form intermediate data;
3) establishing a rapid identification model of the interlayer spacing size mode of the layered bimetal oxide by utilizing an artificial neural network algorithm based on the intermediate data;
4) collecting new layered bimetal oxide, searching to obtain atomic parameters of the layered bimetal oxide, substituting the parameters into a combined conversion equation, calculating intermediate data of the layered bimetal oxide, substituting the intermediate data into the established rapid identification model of the interlayer spacing size mode of the layered bimetal oxide, and predicting whether the interlayer spacing of the new layered bimetal oxide is a large mode or a small mode.
Compared with the prior art, the invention has the following advantages:
the method for the mode of the interlayer spacing size of the layered double-metal oxide is simple and quick: the atomic parameters obtained by utilizing the element periodic table need several minutes, the obtained atomic parameters are imported into the established model, the result can be calculated without several seconds, and the method is convenient and quick and can be completed by only one person.
Secondly, low cost: the method utilizes the atomic parameter-artificial neural network to predict the interlayer spacing size mode of the layered bimetal oxide, and compared with the traditional X-ray diffractometer for measurement, the method does not need to purchase instruments, and is simple to operate and low in cost.
Thirdly, the environment is not polluted: the invention does not use chemical in the whole process and has no pollution to the environment.
Drawings
Fig. 1 is a structural view of a layered double oxide.
FIG. 2 is a schematic diagram of an artificial neural network.
Detailed Description
The invention is described in detail below with reference to the figures and specific embodiments.
The invention relates to a method for predicting the size mode of the interlayer spacing of a layered double-metal oxide by a non-chemical experimental method, which comprises the following steps:
(1) collecting 33 layered double hydroxides and interlayer spacing data thereof, defining the interlayer spacing of more than 7.7 as a large interlayer spacing mode and the interlayer spacing of less than 7.7 as a small interlayer spacing mode according to the values, collecting atomic parameters of metal atoms in the layered double hydroxides by combining a periodic table of elements, and forming basic data by taking the atomic parameters as independent variables and taking interlayer spacing dependent variables. Some example data are shown in table 1;
TABLE 1 interlayer spacing and atomic parameters of layered bimetal oxides
Serial number
Molecular formula
Size mode
n
a
n
b
z
z
/r
z
r
a
/r
b
1
Mg0.63Al0.37(OH)2(CO3)0.19 ∙ 0.62H2O
Small
0.63
0.37
1.08
1.33
2
Mg0.67Al0.33(OH)2(CO3)0.17 ∙ 0.52H2O
Small
0.67
0.33
1.08
1.33
3
Mg0.75Al0.25(OH)2(CO3)0.13 ∙ 0.56H2O
Big (a)
0.75
0.25
1.08
1.33
4
Mg0.80Al0.20(OH)2(CO3)0.10 ∙ 0.48H2O
Big (a)
0.80
0.20
1.08
1.33
5
Co0.63Al0.37(OH)2(CO3)0.19 ∙ 0.80H2O
Small
0.63
0.37
1.08
1.20
6
Co0.67Al0.33(OH)2(CO3)0.17 ∙ 0.68H2O
Small
0.67
0.33
1.08
1.20
7
Co0.75Al0.25(OH)2(CO3)0.13 ∙ 0.71H2O
Small
0.75
0.25
1.08
1.20
8
Co0.80Al0.20(OH)2(CO3)0.10 ∙ 0.60H2O
Small
0.80
0.20
1.08
1.20
9
Mg0.75A10.25(OH)2(C03)0.13 ∙ 0.5H20
Small
0.75
0.25
1.08
1.33
10
Mg0.75Al0.25(OH)2(NO3)0.25 ∙ 0.64H2O
Big (a)
0.75
0.25
0.53
1.33
11
Ni0.75Al0.25(OH)2(CO3)0.13 ∙ 0.49H2O
Big (a)
0.75
0.25
1.08
1.28
Atomic references include: the atomic number of the divalent metal element, the atomic number of the trivalent metal element, the ratio of the valence electron number of the intercalation anion to the thermochemical radius, the ratio of the ionic radius of the divalent metal element to the trivalent metal element, and the like.n a： The number of atoms of the divalent metal element; n isb: the atomic number of the trivalent metal element; z is a radical ofz/rz: the ratio of the number of valence electrons to the thermochemical radius of the intercalated anion; r isa/rb: the ratio of the ionic radii of the divalent metal element to the trivalent metal element.
(2) And carrying out combined transformation on the atomic data to obtain comprehensive variables, wherein the comprehensive variables are linear combinations of the atomic data, and the comprehensive variables and the interlayer spacing form intermediate data. The combined conversion equation is as follows:
Y1= +3.034[na]-3.328[nb]-3.159[zz/rz]+5.329[ra/rb]-4.945
Y2= -3.839[na]-4.321[nb]-2.331[zz/rz]-6.884[ra/rb]+14.891
Y3= -6.894[na]-11.039[nb]-0.217[zz/rz]-0.215[ra/rb]+8.605
Y4= -18.703[na]-13.466[nb]-1.508[zz/rz]+3.892[ra/rb]+13.625
the combined transformed data are shown in table 2.
TABLE 2 Combined transformed data
Y1
Y2
Y3
Y4
-0.59
-0.80
-0.34
0.41
-0.33
-0.78
-0.18
0.20
0.17
-0.74
0.15
-0.22
0.49
-0.72
0.36
-0.48
-1.28
0.10
-0.32
-0.10
-1.03
0.11
-0.15
-0.31
-0.52
0.15
0.18
-0.73
-0.20
0.18
0.39
-0.99
0.17
-0.74
0.15
-0.22
1.91
0.54
0.27
0.61
-0.09
-0.40
0.17
-0.42
(3) Based on the intermediate data, an artificial neural network algorithm is utilized to establish a rapid identification model of the interlayer spacing size mode of the layered bimetal oxide, and the artificial neural network inputs the number of layer nodes 4, the number of hidden layer nodes 3 and outputs the number of layer nodes 1.
(4) Collecting 4 new layered double-metal oxides, searching to obtain atomic parameters of the layered double-metal oxides, substituting the parameters into a combined conversion equation, calculating intermediate data of the layered double-metal oxides, substituting the intermediate data into a rapid identification model for establishing a layer spacing size mode of the layered double-metal oxides, and predicting whether the layer spacing of the new layered double-metal oxides is a large mode or a small mode. Four new layered double oxides with atomic parameters as shown in table 3.
TABLE 3 New predicted data atom parameters
na
nb
zz/rz
ra/rb
0.63
0.37
1.08
1.33
0.67
0.33
0.55
1.37
0.67
0.33
0.55
1.18
0.75
0.25
1.08
1.26
Example 1: the identification models of 33 layered double-metal oxide interlayer spacing size modes are established by using the artificial neural network by taking the intermediate data as an independent variable and the interlayer spacing size mode as a target variable, and the model accuracy is shown in table 4.
TABLE 4 modeling results
Example 2: the leave-one-out results of the model for recognition of the 33 layered double-metal oxide interlayer spacing size patterns are shown in table 5. The leave-one-out cross-validation assumes that there are N samples, each sample being a test sample and the other N-1 samples being training samples. This results in N classifiers, N test results. The average of these N results is used to measure the performance of the model.
TABLE 5 leave-one-out results
Example 3: and (4) forecasting results of the size mode of the interlayer spacing of the 4 new layered bimetal oxides. Substituting 4 layered double-metal oxide atomic parameters into a combined transformation equation to obtain new intermediate data of the layered double-metal oxide, substituting the intermediate data into an artificial neural network model, and forecasting the interlayer distance size mode of the 4 new layered double-metal oxides. The intermediate data and the forecast results are shown in table 6.
TABLE 6 intermediate data of forecast samples
Serial number
na
nb
zz/rz
ra/rb
Forecast results
1
-0.59
-0.80
-0.34
0.41
Small model
2
1.55
0.18
-0.07
1.15
Big mode
3
0.54
1.49
-0.03
0.41
Big mode
4
-0.20
-0.26
0.17
-0.49
Small model