JP2011076975A - Excitation energy predicting device, and excitation energy predicting method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device which reduces time and cost-benefit load and predicts the exciting energy for a pigment, with full high accuracy. <P>SOLUTION: The exciting energy predicting device includes a holding means which holds correction information indicative of the relation between an exciting energy, for a reference pigment, obtained by an experiment and an excitation energy, for the reference pigment, that is calculated based on the chemical structure formula of the reference pigment using a molecular orbital method; an input-accepting means which accepts input of the chemical structure formula of a subject pigment for which an exciting energy is to be predicted; a calculating means which calculates the excitation energy for the subject pigment based on the chemical structure formula, by using the molecular orbital method; and a correction means which corrects the excitation energy, for the subject pigment, calculated by the calculating means, using the correction information held by the holding means. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an excitation energy prediction apparatus, an excitation energy prediction method, and a program, and more particularly to a technique for screening dyes having appropriate excitation energy at low cost and at high speed in development of dyes used for dye-sensitized solar cells and the like.

  In designing the dye, it is necessary to optimize the excitation energy in order to obtain the desired absorption and emission. In optimizing the excitation energy, the excitation energy can be known by referring to literature values for existing dyes. However, in the case of a new dye, it is necessary to obtain the excitation energy by experiments or the like.

  However, when the excitation energy is obtained by experiments, the time and cost burden for synthesizing the dye is large. In order to develop new dyes with the desired excitation energy, it is desirable to try many new dyes and narrow down to the desired dyes. However, it is time-consuming and economical to screen many new dyes by organic synthesis. There is a lot of waste.

J. Ridley et al. Theoret. Chim. Acta 32, 111, 1973 K. Hara et al. J. Phys. Chem. B 107, 597, 2003 K. Hara et al. J. Phys. Chem. B 109, 15476, 2005 Y. Kurashige et al. J. Phys. Chem. A 111, 5544, 2007 Tatsuya Tachikawa and three others, "Electron spectrum calculation of condensed polycyclic aromatic hydrocarbons by semi-empirical molecular orbital method using New-γ", [online], [October 1, 2009 search], Internet <Http://polaris.hoshi.ac.jp/cicsj26/jfiles/JP05.pdf> Other reference information “Merocyanine dye”, 4-C organic dye system Table 1, [online], [October 1, 2009 search], Internet <http://www.jpo.go.jp/shiryou/s_sonota/ hyoujun_gijutsu / solar_cell / 4_c_1_b.htm>

  Therefore, a means for calculating the excitation energy based on the electronic state theory by assuming the molecular structure of the dye using the molecular orbital method is conceivable. However, in the case of a non-empirical method such as a time-dependent density functional method, for example, in spite of a large calculation time burden, the calculation accuracy is often not sufficient for dye design applications.

  In the case of a semi-empirical molecular orbital method such as the ZINDO method (Non-Patent Document 1), the excitation energy of a large number of dyes can be calculated in about one thousandth of the time of the non-empirical molecular orbital method. Is big. However, there is a problem that the calculation accuracy is lower than that of the non-empirical method, the result is not reliable, and is not suitable for dye screening.

  For example, in designing a dye used in a dye-sensitized solar cell, not only the peak position of excitation energy but also information on the energy at the absorption edge may be required (Non-Patent Document 2 and Non-Patent Document 3). Since the dye exists in a condensed system such as in a solvent, the absorption / emission spectrum has a large non-uniform width due to the interaction with the solvent molecule. For this reason, the absorption edge greatly deviates from the peak position. Recently, the absorption peak position of excitation energy for coumarin-based dyes and the like has been investigated by a non-empirical method such as a time-dependent density functional method (Non-Patent Document 4). There is currently no suitable method for empirical calculations.

  In view of the above circumstances, an object of the present invention is to provide an apparatus that reduces the time and cost burden and predicts the excitation energy of a dye with sufficient accuracy.

  According to the present invention, the correction indicating the relationship between the excitation energy of the reference dye obtained by experiment and the excitation energy of the reference dye calculated using the molecular orbital method based on the chemical structural formula of the reference dye. A holding means for holding information; an input receiving means for receiving an input of a chemical structural formula of a target dye whose excitation energy is to be predicted; and an excitation energy of the target dye using the molecular orbital method based on the chemical structural formula There is provided an excitation energy predicting device comprising: a calculating unit that calculates the correction energy; and a correcting unit that corrects the excitation energy of the target dye calculated by the calculating unit using the correction information held by the holding unit. .

  In the present invention, the molecular orbital method may be a semi-empirical molecular orbital method.

  In the present invention, the molecular orbital method may be a ZINDO method.

  In the present invention, the holding means includes the plurality of references calculated using the molecular orbital method based on excitation energies of a plurality of reference dyes obtained by experiments and chemical structural formulas of the plurality of reference dyes. A correction formula calculated using the excitation energy of the dye can be held as the correction information.

  In the present invention, the correction formula may be calculated by regression analysis.

  In the present invention, the regression analysis may be a regression analysis represented by a quadratic expression of excitation energy with a constant term as zero.

  In the present invention, the regression analysis can be a regression analysis represented by a quadratic expression of excitation energy where a constant term is zero and a secondary coefficient is 1.

  In the present invention, the holding means classifies the reference dyes into a plurality of groups according to a chemical structural formula, holds the correction information for each of the plurality of groups, and the correction means is input by the input receiving means. Determination means for determining to which group the chemical structural formula that has been received belongs, acquisition means for acquiring the correction information of the group determined by the determination means from the holding means, and the acquisition means acquired by the acquisition means The first correction means for correcting the excitation energy of the target dye calculated by the calculation means by using correction information can be used.

  According to the present invention, based on the chemical structural formula of the target dye whose excitation energy is to be predicted, the excitation energy of the target dye is calculated using the molecular orbital method, and the excitation energy of the reference dye obtained by experiment, Using the correction information indicating the relationship with the excitation energy of the reference dye calculated using the molecular orbital method based on the chemical structural formula of the reference dye, the calculated excitation energy of the target dye is corrected. An excitation energy prediction method is provided.

  In the present invention, the molecular orbital method is a ZINDO method, and the correction information is obtained based on an excitation energy of a plurality of reference dyes obtained by experiments and a chemical structural formula of each of the plurality of reference dyes. Using the excitation energies of the plurality of reference dyes calculated using the correction formula, a correction formula calculated by regression analysis can be obtained.

  According to the present invention, there is provided a program for predicting excitation energy, an input receiving step for receiving an input of a chemical structural formula of a target dye whose excitation energy is to be predicted, and a molecular orbital method based on the chemical structural formula. The calculation step of calculating the excitation energy of the target dye using the above, the excitation energy of the reference dye obtained by experiment, and the reference calculated using the molecular orbital method based on the chemical structural formula of the reference dye A correction program for correcting the excitation energy of the target dye calculated in the calculation step using correction information indicating a relationship between the excitation energy of the dye and a program for causing the computer to execute the correction step is provided.

  In the present invention, the molecular orbital method is a ZINDO method, and the correction information is obtained based on an excitation energy of a plurality of reference dyes obtained by experiments and a chemical structural formula of each of the plurality of reference dyes. Using the excitation energies of the plurality of reference dyes calculated using the correction formula, a correction formula calculated by regression analysis can be obtained.

  In the present invention, after calculating the excitation energy (calculated value) of the target dye whose excitation energy should be predicted using the molecular orbital method, the excitation energy (actually measured value) of the reference dye obtained by experiment and the molecular orbital method The calculated excitation energy (calculated value) of the target dye is corrected using correction information indicating the relationship between the excitation energy (calculated value) of the reference dye calculated using the same molecular orbital method. According to such a technique, when a new dye having a desired excitation energy is developed, excitation energy with sufficient accuracy can be obtained by means of calculating the excitation energy based on the electronic state theory without obtaining an actual measurement value by experiment. Can be obtained. As a result, it is possible to reduce the time and economic burden.

  In the present invention, the excitation energy (measured value) of the reference dye obtained by experiment can have not only the value of the peak position of the excitation energy but also the value of the absorption edge. As a result, it is possible to reduce the time and economic burden, calculate the value of the peak position of the excitation energy of the new dye with high accuracy, and also calculate the value of the absorption edge.

  According to the present invention, it is possible to reduce the time and economical burden and predict the excitation energy of the dye with sufficient accuracy.

It is an example of the functional block diagram of the excitation energy prediction apparatus of Embodiment 1. It is an example of the functional block diagram of the excitation energy prediction apparatus of Embodiment 2. It is a figure which shows the chemical structural formula of a coumarin-type pigment | dye. It is a figure which shows the chemical structural formula of a merocyanine pigment | dye. It is a figure which shows the chemical structural formula of a merocyanine pigment | dye.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In all the drawings, the same components are denoted by the same reference numerals, and description thereof will be omitted as appropriate.

  Note that each unit constituting the device of each embodiment includes a CPU, a memory, and a program loaded in the memory of an arbitrary computer (a program stored in the memory from the stage of shipping the device in advance, or a removable medium such as a CD). (Including programs downloaded from media, servers on the Internet, etc.), storage units such as hard disks for storing the programs, network connection interfaces, displays, etc., and any other combination of hardware and software. . It will be understood by those skilled in the art that there are various modifications to the implementation method and apparatus.

  Also, the functional block diagrams of FIGS. 1 and 2 used in the following description show functional unit blocks instead of hardware unit configurations. In these drawings, the excitation energy predicting device of each embodiment is described as being realized by one physically separated device, but the means for realizing it is not limited to this. That is, two or more physically separated devices may be connected by wire or wirelessly, and the excitation energy predicting device of each embodiment may be realized by the plurality of devices.

  First, an outline of the present invention will be described.

  In the present invention, for example, a calculated value of excitation energy calculated using a semi-empirical molecular orbital method such as the ZINDO method is input to a correction formula calculated by learning from training data, for example, and the calculated value of excitation energy is calculated. To improve the accuracy of the predicted value of excitation energy.

  Here, the “training data” means a set of data in which the chemical structural formula of the reference dye is associated with the excitation energy (actual measurement value) of the reference dye obtained by the ultraviolet / visible spectrum measurement experiment. This “excitation energy (experimental value)” may be the value of the peak position of the excitation energy, the value of the absorption edge, or both. The “reference dye” corresponds to any dye whose chemical structural formula is clear and whose measured value of excitation energy is obtained by past ultraviolet / visible spectrum measurement experiments. For example, it may be a dye whose chemical structural formula and actual measurement value of excitation energy are described in the literature, or may be a dye that has been experimentally clarified to reveal the chemical structural formula and excitation energy. .

  Further, “learning” refers to an excitation energy (calculated value) calculated using a semi-empirical molecular orbital method (for example, “ZINDO method”) based on a chemical structural formula of a reference dye included in training data, This means a process of calculating a correction equation by regression analysis or the like using the excitation energy (actual value) of the reference dye held as training data.

Next, an embodiment of the present invention will be described.
<Embodiment 1>

  FIG. 1 shows an example of a functional block diagram and a processing flow of the excitation energy prediction apparatus of the present embodiment. As shown in the figure, this excitation energy prediction device is roughly divided into two parts, a data training unit 10 and an excitation energy prediction unit 20.

  The data training unit 10 includes a known pigment data unit 11, a first excitation energy calculation unit 12, and a learning unit 13.

The known dye data section 11 has a chemical structural formula that is clear, and an actual measured value of excitation energy obtained from past ultraviolet / visible spectrum measurement experiments and the like. : Z _{E (} hereinafter referred to as “actually measured excitation energy Z _E ”).

  The means for acquiring the data to be held is not particularly limited. For example, it may be stored in the memory in advance from the shipment stage of the excitation energy prediction device, or accept data input from the user of the excitation energy prediction device. It may be retained or a combination thereof.

The means for receiving data input from the user is not particularly limited. For example, the excitation energy prediction apparatus of the present embodiment includes a user interface realized by a device such as a keyboard, a mouse, and a touch panel display. Via the user, input of numerical data, character data, and graphic data may be received from the user. Alternatively, the excitation energy prediction device of the present embodiment is configured to be able to communicate with other external devices, portable storage devices, etc., and data stored in other external devices, portable storage devices, etc. Data input may be accepted by reading. In addition, the excitation energy prediction apparatus of this embodiment is configured to be connectable to the Internet, and accesses a predetermined server on the network periodically or intermittently (in advance, the excitation energy prediction apparatus of this embodiment Holds the URL of the server to be accessed), and new data may be acquired. This server, such as sales manufacturer manage the excitation energy prediction apparatus, the data associated with the chemical structure and measured excitation energy Z _E regularly or intermittently, or may be arranged in such a way as to update.

The first excitation energy calculation unit 12 receives the chemical structural formula held by the known dye data unit 11 as an input, optimizes the structure using, for example, the PM3 method or the AM1 method, and then performs a molecular orbital method such as a semi-empirical molecular orbital method, As a more detailed example, the ZINDO method is used to calculate the excitation energy (calculated value: Z _C1 , hereinafter referred to as “calculated excitation energy Z _C1 ”) using the optimized chemical structural formula as an input. Specifically, the first excitation energy calculation unit 12 includes a PM3 method and AM1 method for structure optimization, a molecular orbital method for calculating excitation energy, such as a semi-empirical molecular orbital method, As an example, a calculation program for realizing a calculation process by the ZINDO method is held, and the calculation excitation energy Z _C1 is calculated by executing the calculation program in combination. The calculated calculated excitation energy Z _C1 is transferred to the learning unit 13 described below in association with the inputted chemical structural formula.

Learning unit 13 acquires the measured excitation energy Z _E from known dyes data unit 11, also, to obtain the calculated excitation energy Z _C1 from the first excitation energy calculation unit 12. Then, the chemical structural formula as a key and associates the measured excitation energy Z _E of the same dye and compute the excitation energy Z _C1. Then, by using the association data, linking the calculated excitation energy Z _C1 on the measured excitation energies Z _E calculates a correction equation f. For example, the correction formula f may be calculated by using regression analysis or a neural network. Specifically, a calculation program for realizing a regression analysis or a calculation process using a neural network is held, and the calculation of the correction formula f is realized by executing the calculation program. The calculated correction formula f is stored in the internal memory.

  Here, in dye design, a derivative is often produced by changing the atoms or atomic groups in the peripheral portion while maintaining the basic structure. Therefore, the learning unit 13 may classify the pigments into a plurality of groups according to the chemical structural formula, and calculate the correction formula for each group using only the reference pigment belonging to each group. In this way, the reliability of the predicted excitation energy value obtained in this embodiment is improved.

  This group includes, for example, (1) metal complex, (2) porphyrin, (3) phthalocyanine, (4) nonmetal complex, (5) polymethine, (6) polyene, (7) styryl. (8) azo, (9) carbonium, (10) xanthene, (11) coumarin, (12) diketopyrrolopyrrole, (13) squarylium, (14) anthracene, (15) It may be classified as perylene, (16), etc. This classification is merely an example, and the classification may be performed more finely, or may be classified into groups smaller than the above example.

  The learning unit 13 is not particularly limited as a specific means for classifying the reference pigments that have received the input into the group. For example, the learning unit 13 stores data indicating a characteristic basic structure for each group. It is possible to determine which group each dye belongs to by determining which group's basic structure exists in the chemical structural formula of each reference dye. Good. There are no particular restrictions on the means to determine which basic structure exists. For example, using a method such as character recognition or image recognition, it is determined which element is connected to which element by how many lines. By doing so, it may be realized. Alternatively, as a means for the known dye data unit 11 to accept the input of the chemical structure of the reference dye, an interface that allows selection of items such as elements to be used, single bonds, and double bonds is held, and these items are stored. Use it to accept chemical structure inputs. Then, it may be realized by determining which item is used and which item is connected to the chemical structural formula that has received the input. In addition, when the known dye data unit 11 receives the input of the chemical structural formula of the reference dye, it also receives an input of which group the reference dye belongs to, and classifies the reference dye into groups according to the input group. Such means may be used.

  When the learning unit 13 calculates the correction formula for each group, the known pigment data unit 11 classifies and holds the acquired reference pigments for each basic structure by the same means as described above. Also good.

  Next, a method preferable when the learning unit 13 of the present embodiment calculates the correction formula f by regression analysis will be described in detail.

First, from a plurality of test calculations actually performed by the present inventor, assuming that the predicted value (or reproduced value) of excitation energy by regression analysis is Z _P , the form of the single regression equation (1) is the original ZINDO calculation. It was found that the improvement from the value (Z _C ) was insufficient.

Z _P = α ₁ Z _C + α ₀ (1)

Therefore, the present inventor considered increasing the explanatory variables included in the regression equation. In this case, the introduction of the first explanatory variable Z _C unrelated explanatory variables, may latter effect is very large, there is a possibility of outputting the predicted values very bad with respect to the dye that is not included in the training data . That is, robustness with respect to prediction is lost. Therefore, the present inventor considered adding a power of the first explanatory variable. At this time, except for the special case where the first explanatory variable can be a set consisting of only 0 and 1, the second and subsequent explanatory variables are not linearly dependent on the previous explanatory variables. Therefore, there is no problem in regression calculation. Therefore, the present inventor considered the following Nth order regression equation (2).

^{_{Z P = α N Z C N}} + α N-1 Z C N-1 + ···· + α 1 Z C + α 0 (2)

In this equation (2), if the constant term (α ₀ ) is not zero, the predicted value (Z _P ) is strongly influenced by the constant term when Z _C is small. For example, if the constant term is negative, a non-physical situation occurs in which negative excitation energy is predicted when Z _C is sufficiently small. For this reason, it is appropriate to set α ₀ = 0. On the other hand, when the higher-order terms of Z _C are present, since the predicted value with high excitation energy side (Z _P) would change abruptly, which is not preferable from the viewpoint of robustness. Further, seemingly Generally the explanatory variables increases, but approaches 1 coefficient of determination indicating the goodness of true regression calculation, the degree of freedom adjusted coefficient of determination increased variable Given (hereinafter, corrected R ₂ hereinafter) effect May not be observed, but rather the correction R ₂ may be reduced. From the above, it has been found that it is preferable to use the form of the quadratic expression (3) without an intercept in the present invention.

Z _P = α ₁ Z _C + α ₂ Z _C ² (3)

Further, when α ₁ = 1 is set in the equation (3), the quadratic equation (4) is obtained.

Z _P = Z _C + α ₂ Z _C ² (4)

This equation (4) can be interpreted as (predicted value: Z _P ) = (excitation energy calculated using molecular orbital method (eg, ZINDO method)) + (second-order correction term of the excitation energy), Since the explanatory variables decrease, the reproduction accuracy decreases, but the robustness increases when the number of learning data is not large. That is, in the present invention, it has been found preferable to use the formula (4).

  The excitation energy prediction unit 20 includes a new dye data unit 21, a second excitation energy calculation unit 22, and an output unit 23.

  The new dye data unit 21 holds data on the chemical structural formula of the target dye for which the excitation energy received from the user is desired. The means for receiving data input from the user is not particularly limited. For example, a user interface realized by a device such as a keyboard, a mouse, or a touch panel is provided, and data input from the user is received via this user interface. Also good. Or, it is configured to be communicable with other external devices and portable storage devices, and by accepting data input by reading data stored in other external devices and portable storage devices Also good. The data of the chemical structural formula of the target dye held by the new dye data section 21 is passed to the second excitation energy calculation section 22 described below. The new dye data unit 21 may continue to hold the data of the chemical structural formula of the target dye after passing the data of the chemical structural formula of the target dye to the second excitation energy calculating unit 22 or discard it. May be.

The second excitation energy calculation unit 22 receives the chemical structural formula of the target dye held by the new dye data unit 21 as an input, and is the same method as the first excitation energy calculation unit 12 (preferably, the same calculation program). Is used to optimize the structure and calculate the excitation energy (calculated value: Z _C2 , hereinafter referred to as “calculated excitation energy Z _C2 ”). Thereafter, the calculated calculated excitation energy Z _C2 is input to the correction formula f calculated by the learning unit 13, and the calculation result (excitation energy (predicted value: Z _p , hereinafter referred to as “predicted excitation energy Z _p ”)). Ask. The calculation result (predicted excitation energy Z _p )) is passed to the output unit 23 described below.

Note that when the learning unit 13 classifies the dyes into a plurality of groups according to the chemical structural formula and calculates and holds the correction formula for each group, the second excitation energy calculation unit 22 includes the new dye data unit 21. It is determined to which group the chemical structural formula that has received the input from belongs, and thereafter, the correction formula f of the determined group is acquired from the learning unit 13, and the calculated calculation excitation is added to the acquired correction formula f. The energy Z _C2 is input, and the calculation result (predicted excitation energy Z _p ) is obtained. As a means for determining to which group the chemical structural formula that has accepted the input belongs, the same means as the means described in the learning unit 13 can be used.

The output unit 23, a prediction excitation energy Z _p to the second excitation energy calculation unit 22 calculates and outputs by using a display or a speaker. Other, the calculated prediction excitation energy Z _p is printed on a paper medium, it may be output.

  In the present embodiment, the molecular orbital calculation is performed at a semi-empirical level, so the processing speed is high. Moreover, the excitation energy prediction apparatus of this embodiment can also be realized by a combination of existing software, and can be realized at an extremely low price. This makes it possible to calculate excitation energy for each dye in parallel by using multiple excitation energy prediction devices, even when screening a desirable dye from a very large number of new dyes, and further processing. Speed can be increased.

  In the excitation energy calculation, the excited state of interest in the dye design is a low excited state corresponding to electronic excitation from the highest occupied orbital (HOMO) to the lowest unoccupied orbital (LUMO), that is, a HOMO-LUMO transition. Therefore, in this embodiment, it is also possible to shorten the excitation energy calculation time by including only a low excited state in the calculation target.

  As described above, the excitation energy prediction apparatus according to the present embodiment can predict the excitation energy of the dye with an accuracy that allows the experiment to be considered as a substitute while maintaining the low calculation load of the semi-empirical calculation method. That is, the excitation energy prediction apparatus of the present embodiment finds a correction equation that reproduces the experimental value from the calculated value assuming an appropriate regression equation, and calculates the excitation energy of a desired new dye using the equation. Thus, it is possible to design an apparatus that gives a calculation accuracy higher than that of an ab initio calculation method without increasing the calculation cost. This makes it possible to perform dye screening without actually synthesizing a large number of dyes, and as a result, it is possible to reduce the time and cost of developing new dyes.

Here, Non-Patent Document 5 discloses a technique for calculating excitation energy using the ZINDO method and regression analysis. However, this technique uses regression analysis to determine the parameter values used in the ZINDO method. According to such a technique, an existing program for ZINDO calculation cannot be used, and it is necessary to make an appropriate improvement in order to use this technique. In addition, reference data for regression analysis needs to be newly collected, and the time and cost burden is large. On the other hand, in the case of the present embodiment, it is possible to use an existing program for ZINDO calculation without improving it, and the time and cost burden is small. In addition, it is possible to use existing literature values as reference data for regression analysis, and in this respect, the time and cost burden is small.
<Embodiment 2>

  FIG. 2 shows an example of a functional block diagram of the excitation energy prediction apparatus of the present embodiment. As shown in the figure, the excitation energy prediction apparatus includes a holding unit 30, an input receiving unit 40, a calculating unit 50, and a correcting unit 60. In addition, you may have the output part 23. FIG.

  The holding unit 30 is correction information indicating the relationship between the excitation energy of the reference dye obtained by experiment and the excitation energy of the reference dye calculated by using the molecular orbital method based on the chemical structural formula of the reference dye. Configured to hold. The holding unit 30 may classify the reference dyes into a plurality of groups according to the chemical structure, and hold correction information for each of the plurality of groups.

  The molecular orbital method here may be, for example, a semi-empirical molecular orbital method, and more specifically, may be a ZINDO method.

  The correction information here is used by the molecular orbital method (calculation unit 50 described below) based on the excitation energy of a plurality of reference dyes obtained through experiments and the chemical structural formulas of the plurality of reference dyes. It may be a correction formula calculated using the excitation energies of the plurality of reference dyes calculated using the same molecular orbital method. This correction formula may be calculated by regression analysis. For example, it may be a correction formula calculated by regression analysis represented by a quadratic expression (3) of excitation energy with a constant term set to zero. Alternatively, it may be a correction formula calculated by regression analysis represented by a quadratic expression (4) of excitation energy with a constant term of zero and a quadratic coefficient of 1.

Z _P = α ₁ Z _C + α ₂ Z _C ² (3)

Z _P = Z _C + α ₂ Z _C ² (4)

  The concepts such as the plurality of groups are the same as those described in the data training unit 10 of the first embodiment.

  The holding unit 30 may have the same configuration as the data training unit 10 described in the first embodiment. Or the data training part 10 is implement | achieved by the apparatus different from the excitation energy prediction apparatus of this embodiment, and the holding | maintenance part 30 hold | maintains the correction formula acquired by wired / wireless communication from the apparatus which has the data training part 10. You may comprise.

  The input receiving unit 40 is configured to receive an input of a chemical structural formula of a target dye whose excitation energy is to be predicted. The input receiving unit 40 can be realized by the same configuration as the new pigment data unit 21 described in the first embodiment.

  The calculation unit 50 uses the molecular orbital method, for example, the semi-empirical molecular orbital method, and more specifically, the ZINDO method, based on the chemical structural formula received by the input receiving unit 40 to excite the target dye. It is configured to calculate energy.

  The correction unit 60 is configured to correct the excitation energy of the target dye calculated by the calculation unit 50 using the correction information held by the holding unit 30. The correction here corresponds to, for example, a process of inputting the excitation energy calculated by the calculation unit 50 to the correction formula held by the holding unit 30 and calculating the calculation result as corrected excitation energy.

  The correction unit 60 may include a determination unit 61, an acquisition unit 62, and a first correction unit 63. The determination unit 61 is configured to determine to which group the chemical structural formula received by the input receiving unit 40 belongs. This group is a group in which the reference dyes are classified according to the chemical structure so that the holding unit 30 holds a plurality of correction information for each group. The acquisition unit 62 is configured to acquire the correction information of the group determined by the determination unit 61 from the holding unit 30. The first correction unit 63 is configured to correct the excitation energy of the target dye calculated by the calculation unit 50 using the correction information acquired by the acquisition unit 62.

  The calculation unit 50 and the correction unit 60 can be realized by the same configuration as the second excitation energy calculation unit 22 described in the first embodiment.

  Also in the excitation energy prediction apparatus of this embodiment, the same effect as the excitation energy prediction apparatus of Embodiment 1 is realizable.

  The following inventions are also disclosed by the description of the first and second embodiments.

  Based on the chemical structural formula of the target dye whose excitation energy is to be predicted, the excitation energy of the target dye is calculated using the molecular orbital method, and the excitation energy of the reference dye obtained by experiment and the chemical structure of the reference dye An excitation energy prediction method for correcting the calculated excitation energy of the target dye using correction information indicating the relationship between the excitation energy of the reference dye calculated using the molecular orbital method based on the formula .

  In the excitation energy prediction method, the molecular orbital method is a ZINDO method, and the correction information is based on excitation energy of a plurality of reference dyes obtained by experiments and chemical structural formulas of the plurality of reference dyes. A method for predicting excitation energy, which is a correction formula calculated by regression analysis using excitation energy of the plurality of reference dyes calculated using the ZINDO method.

  A program for predicting excitation energy, an input receiving step for receiving an input of a chemical structural formula of a target dye whose excitation energy is to be predicted, and the target dye using a molecular orbital method based on the chemical structural formula A calculation step for calculating the excitation energy of the reference dye, the excitation energy of the reference dye obtained by an experiment, the excitation energy of the reference dye calculated using the molecular orbital method based on the chemical structural formula of the reference dye, And a correction step of correcting the excitation energy of the target dye calculated in the calculation step by using correction information indicating the relationship of the above.

In the program, the molecular orbital method is a ZINDO method, and the correction information is calculated based on the excitation energy of a plurality of reference dyes obtained by experiments and the chemical structural formulas of the plurality of reference dyes. A program which is a correction formula calculated by regression analysis using the excitation energies of the plurality of reference dyes calculated by using.
<Example>

<Example 1: Reproduction of excitation energy of coumarin dye (absorption peak position)>

  Table 1 shows the excitation energy experiments of nine coumarin dyes (C343, NKX-2195, NKX-2311, NKX-2384, NKX-2388, NKX-2393, NKX-2398, NKX-2510, NKX-2586). Value (absorption peak), calculated excitation energy by ZINDO method, and reproduced value of excitation energy by equations (3) and (4) (energy unit is eV, experimental value is Non-patent document 2). FIG. 3 shows chemical structural formulas of nine coumarin dyes. Equations (3) and (4) were calculated by regression analysis using the data for the nine coumarin dyes.

The original calculated value by the ZINDO method overestimates the experimental value by about 12.4 to 31.3% (20.2% on average). (3) The reproduced value according to the formula (Z _P = 1.106104Z _C −0.10667Z _C ² ) remains at a deviation of −5.1 to + 10.2% (on average + 0.02%) with respect to the experimental value. ing. The correction _{R 2} was obtained a value close to 0.855 and 1, regression analysis is good. In addition, the reproduction value according to the equation (4) (Z _P = Z _C −0.05477Z _C ² ) remains within a deviation (−0.01% on average) of −5.4 to + 9.6% with respect to the experimental value. ing. Correction _{R 2} is a 0.829, again regression analysis is good.

<Example 2: Reproduction of excitation energy of coumarin dye (energy value at absorption edge (threshold value))>

  Table 2 shows the experimental values (energy values at the absorption edge) of the excitation energy of the nine coumarin dyes as in Example 1, calculated excitation energy by the ZINDO method, and absorptions by the equations (3) and (4). The energy value at the end and the reproduced value are shown (the unit of energy is eV, and the experimental value is according to Non-Patent Document 2). Equations (3) and (4) were calculated by regression analysis using the data for the nine coumarin dyes.

(3) The reproduction value according to the formula (Z _P = 0.629647Z _C + 0.014883Z _C ² ) remains at a deviation of −5.8 to + 13.6% with respect to the experimental value. The correction _{R 2} was obtained a value close to 0.852 and 1, regression analysis is good. The reproduced value by equation (4) _{(Z P = Z C -0.10446Z C} 2) is remained -6.0 + 11.5% deviation with respect to the experimental values. Correction _{R 2} is a regression analysis in 0.850 is good.

<Example 3: Prediction of excitation energy of coumarin dye>

Table 3 shows that NKX-2777, which is not included in the training data of Example 1, is an unknown dye, the formula (3) of Example 1 (Z _P = 1.1061054Z _C −0.10667Z _C ² ), and (4) A predicted value of excitation energy (peak position) is calculated using the formula (Z _P = Z _C −0.05477Z _C ² ), and this predicted value and an experimental value (the experimental value is according to Non-Patent Document 3). And a value calculated by the time-dependent density functional (TDDFT) method (the unit of energy is eV). The TDDFT calculation used the SCRF method to incorporate the effect in methanol solvent at the B3LYP / 6-31 + G (d, p) level.

Similarly, excitation energy using the formula (3 _P ) of Example 2 (Z _P = 0.629647Z _C + 0.014883Z _C ² ) and the formula (4) (Z _P = Z _C −0.10446 Z _C ² ). A predicted value of (energy value at the absorption edge) is calculated and Table 3 shows a comparison between this predicted value and an experimental value (the experimental value is according to Non-Patent Document 3). Since the energy value of the absorption edge cannot be calculated by the TDDFT method, there is no comparison with the calculated value by this method.

  The predicted values of the excitation energy (peak position) calculated using the equations (3) and (4) are both closer to the experimental values than the TDDFT method. Further, the excitation energy (energy value at the absorption edge) calculated using the equations (3) and (4) is close to the experimental value.

<Example 4: Reproduction of excitation energy of merocyanine dye (absorption peak position)>

  Table 4 shows experimental values (absorption peaks) of excitation energies of four merocyanine dyes (Ma (2) -N, Mb (2) -N, Mb (2) ′, Ma (2) -NM), The excitation energy calculation value by the ZINDO method and the reproduction value of the excitation energy by the equations (3) and (4) are shown (the unit of energy is eV). 4 and 5 show chemical structural formulas of the four merocyanine dyes. Equations (3) and (4) were calculated by regression analysis using the data of the four merocyanine dyes.

The calculated value by the ZINDO method overestimates the experimental value by about 33.7 to 41.1% (average of 38.9%). (3) The reproduction value according to the equation (Z _P = 0.916564Z _C −0.0579Z _C ² ) remains at a deviation of −3.6 to + 1.9% with respect to the experimental value. In addition, the reproduced value according to the equation (4) (Z _P = Z _C −0.08252Z _C ² ) remains at a deviation of −3.5 to + 2.0% with respect to the experimental value. Correction _{R 2} is (3) and (4) respectively when the expression, worse than Example 1 become 0.499 and 0.663. This is presumably because the deviation between the original calculated value by the ZINDO method and the experimental value is larger than that in Example 1 and the number of pigments serving as learning data is small. However, it is a sufficient value in the calculation of excitation energy (absorption peak position). Further, if necessary, the accuracy can be further improved by taking measures such as increasing the number of pigments as learning data.

<Example 5: Reproduction of excitation energy of merocyanine dye (energy value of absorption edge (threshold value))>

  Table 5 shows the experimental values (energy values at the absorption edge) of the excitation energy of the same four merocyanine dyes as in Example 4, the calculated excitation energy by the ZINDO method, and the absorption by the equations (3) and (4). The reproducible value of the energy value at the end is shown (the unit of energy is eV). Equations (3) and (4) were calculated by regression analysis using the data of the four merocyanine dyes.

(3) The reproduction value according to the equation (Z _P = 1.563423Z _C −0.27134Z _C ² ) remains at a deviation of −2.9 to + 1.3% with respect to the experimental value. In addition, the reproduction value according to the equation (4) (Z _P = Z _C −0.1052Z _C ² ) remains at a deviation of −3.2 to + 3.6% with respect to the experimental value. For the same reason as in the fourth embodiment, the correction R ₂ is 0.499 and 0.665, respectively, slightly worse in the cases of the expressions (3) and (4). However, this is a sufficient value for calculating the excitation energy (energy value at the absorption edge). Further, if necessary, the accuracy can be further improved by taking measures such as increasing the number of pigments as learning data.

<Example 6: Prediction of excitation energy of merocyanine dye>

Table 6 shows that Mb (5) -N and Mb (2) -M, which are not included in the training data of Example 4, are unknown dyes, and the expression (3) of Example 4 (Z _P = 0.916564Z _C −0.0579Z _C ² ) and (4) using formula (Z _P = Z _C −0.08252Z _C ² ), a predicted value of excitation energy (peak position) is calculated, and this predicted value and an experimental value are calculated. (Experimental values are based on Non-Patent Document 6) and values calculated by the TDDFT method are compared (unit of energy is eV). The TDDFT calculation used the SCRF method to incorporate the effect in methanol solvent at the B3LYP / 6-31 + G (d, p) level.

Similarly, excitation is performed using (3) formula (Z _P = 1.563423Z _C −0.27134Z _C ² ) and (4) formula (Z _P = Z _C −0.1052Z _C ² ) of Example 5. Table 6 shows a predicted value of energy (absorption edge energy value (threshold value)) and a comparison between the predicted value and the experimental value. Since the energy value of the absorption edge cannot be calculated by the TDDFT method, there is no comparison with the calculated value by this method.

The predicted values of the excitation energy (peak position) calculated using the equations (3) and (4) are both closer to the experimental values than the TDDFT method. Further, the excitation energy (energy value at the absorption edge) calculated using the equations (3) and (4) is close to the experimental value.

DESCRIPTION OF SYMBOLS 10 Data training part 11 Known pigment | dye data part 12 1st excitation energy calculation part 13 Learning part 20 Excitation energy prediction part 21 New pigment | dye data part 22 2nd excitation energy calculation part 23 Output part 30 Holding | maintenance part 40 Input reception part 50 Calculation part 60 Correction unit 61 Judgment unit 62 Acquisition unit 63 First correction unit

Claims (12)

Holding means for holding correction information indicating the relationship between the excitation energy of the reference dye obtained by experiment and the excitation energy of the reference dye calculated using the molecular orbital method based on the chemical structural formula of the reference dye When,
An input receiving means for receiving an input of a chemical structural formula of a target dye whose excitation energy is to be predicted;
Based on the chemical structural formula, calculation means for calculating the excitation energy of the target dye using the molecular orbital method,
Correction means for correcting the excitation energy of the target dye calculated by the calculation means using the correction information held by the holding means;
An excitation energy prediction apparatus having: The excitation energy prediction apparatus according to claim 1,
The molecular orbital method is an excitation energy prediction apparatus that is a semi-empirical molecular orbital method. The excitation energy prediction apparatus according to claim 2,
The molecular orbital method is an excitation energy prediction apparatus that is a ZINDO method. In the excitation energy prediction apparatus as described in any one of Claim 1 to 3,
The holding means is
The excitation energy of a plurality of reference dyes obtained by experiments and the excitation energy of the plurality of reference dyes calculated using the molecular orbital method based on the chemical structural formula of each of the plurality of reference dyes are used. An excitation energy prediction apparatus that holds the correction formula calculated in the above as the correction information. The excitation energy prediction apparatus according to claim 4,
The correction equation is an excitation energy prediction device calculated by regression analysis. The excitation energy prediction apparatus according to claim 5,
The regression analysis is an excitation energy prediction device that is a regression analysis expressed by a quadratic expression of excitation energy with a constant term being zero. The excitation energy prediction apparatus according to claim 5,
The regression analysis is an excitation energy prediction device that is a regression analysis represented by a quadratic expression of excitation energy with a constant term of zero and a secondary coefficient of 1. In the excitation energy prediction device according to any one of claims 1 to 7,
The holding means is
Classifying the reference dyes into a plurality of groups according to the chemical structural formula, holding the correction information for each of the plurality of groups,
The correction means includes
A determination unit that determines which group the chemical structural formula received by the input reception unit belongs;
Acquisition means for acquiring the correction information of the group determined by the determination means from the holding means;
An excitation energy prediction apparatus comprising: first correction means that corrects the excitation energy of the target dye calculated by the calculation means using the correction information acquired by the acquisition means. Based on the chemical structural formula of the target dye whose excitation energy is to be predicted, the excitation energy of the target dye is calculated using the molecular orbital method,
Using correction information indicating the relationship between the excitation energy of the reference dye obtained by experiment and the excitation energy of the reference dye calculated using the molecular orbital method based on the chemical structural formula of the reference dye An excitation energy prediction method for correcting the calculated excitation energy of the target dye. The method of predicting excitation energy according to claim 9,
The molecular orbital method is a ZINDO method,
The correction information includes the excitation energy of a plurality of reference dyes obtained by experiments, the excitation energy of the plurality of reference dyes calculated using the ZINDO method based on the chemical structural formula of each of the plurality of reference dyes, An excitation energy prediction method, which is a correction formula calculated by regression analysis using. A program for predicting excitation energy,
An input receiving step for receiving an input of a chemical structural formula of a target dye whose excitation energy is to be predicted;
Based on the chemical structural formula, a calculation step of calculating the excitation energy of the target dye using a molecular orbital method,
Using correction information indicating the relationship between the excitation energy of the reference dye obtained by experiment and the excitation energy of the reference dye calculated using the molecular orbital method based on the chemical structural formula of the reference dye, A correction step of correcting the excitation energy of the target dye calculated in the calculation step;
A program that causes a computer to execute. The program according to claim 11,
The molecular orbital method is a ZINDO method,
The correction information includes the excitation energy of a plurality of reference dyes obtained by experiments, the excitation energy of the plurality of reference dyes calculated using the ZINDO method based on the chemical structural formula of each of the plurality of reference dyes, A program that is a correction formula calculated by regression analysis using.

Images