Method for deriving metallocene catalyst through genetic algorithm and device for modeling metallocene catalyst using same
By employing density functional theory and genetic algorithms to optimize metallocene catalysts based on structural indices, the method addresses the inefficiencies in existing catalyst design processes, achieving efficient and cost-effective catalyst development for polyolefin polymerization.
Patent Information
- Application Number
- PCT/KR2024/002067
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-15
- Filing Date
- 2024-02-14
- Publication Date
- 2024-11-21
AI Technical Summary
The existing methods for designing and synthesizing metallocene catalysts for polyolefin polymerization are inefficient, requiring extensive experimental analysis and resource-intensive processes, and lack effective virtual screening technologies to predict catalyst performance.
A method using density functional theory (DFT) to calculate electronic and steric structural indices, which are integrated into a fitness function to select and optimize metallocene catalysts through a genetic algorithm, allowing for the prediction of catalyst performance and the derivation of new catalyst structures without extensive experimentation.
This approach significantly reduces the time and resources required for catalyst design, enabling the efficient derivation of metallocene catalysts with optimized performance for polyolefin polymerization, improving economic feasibility and efficiency.
Smart Images

Figure KR2024002067_21112024_PF_FP_ABST
Abstract
Description
A method for deriving a metallocene catalyst using a genetic algorithm and a metallocene catalyst modeling device using the same.
[0001] The present invention relates to a method for deriving a metallocene catalyst using density functional theory (DFT), a fitness function defined by electronic structural indices and stereostructural indices, and a genetic algorithm, a method for predicting the performance of a metallocene catalyst, and a metallocene catalyst modeling device.
[0002]
[0003] To manufacture polyolefin (PO) polymers with superior properties, the molecular weight distribution of the polymer and the polymerization distribution of comonomers can be controlled depending on the metallocene catalyst used. Therefore, research on metallocene catalysts for the excellent synthesis of polyolefins is crucial.
[0004] However, there are limitations to rationally designing a new catalyst based solely on experimental interpretation of polyolefin polymerization results, and synthesizing the designed catalyst and evaluating its performance has the disadvantage of being time-consuming and wasting resources.
[0005] Meanwhile, in recent years, various research fields have been studying virtual screening technology that utilizes computer simulations to design materials and verify their performance in a virtual environment without direct experiments. However, research on screening methods for deriving metallocene catalysts for excellent polyolefin synthesis has been insufficient.
[0006] (Patent Document 1) KR 2022-0129783 A
[0007]
[0008] The purpose of this invention is to provide a method for deriving a metallocene catalyst, which defines a fitness function related to catalyst properties as electronic structural indices and stereostructural indices, calculates electronic structural indices and stereostructural indices from a metallocene catalyst population divided into upper, bridge, and lower groups according to the catalyst structure through density functional theory (DFT), introduces these into a fitness function to obtain a fitness function value, and then selects a metallocene catalyst having a fitness function value greater than a preset value to form a group, and performs a genetic algorithm on the formed group.
[0009]
[0010] One aspect of the present invention relates to a method for deriving a metallocene catalyst.
[0011] 1. A method for deriving a metallocene catalyst according to an embodiment of the present invention comprises: (a) defining a fitness function related to catalyst properties using electronic structural indices and stereostructural indices; (b) forming a population of metallocene catalysts divided into upper, middle, and lower groups according to catalyst structures; (c) calculating electronic structural indices and stereostructural indices of the metallocene catalyst corresponding to the electronic structural indices and stereostructural indices of the fitness function using density functional theory (DFT); (d) obtaining a fitness function value from the electronic structural indices and stereostructural indices; (e) forming a group by selecting metallocene catalysts having a value equal to or greater than a preset value from the fitness function values; (f) performing crossover and mutation on metallocene catalysts within the group to generate offspring entities; and (g) repeating steps (c) to (f) multiple times on the generated offspring entities, and terminating the process.
[0012] 2. In the above 1 specific example, the step of defining a fitness function related to the catalyst properties using electronic structural indices and stereostructural indices may include: a step of analyzing a catalytic reaction mechanism from any metallocene catalyst; a step of deriving electronic structural indices and stereostructural indices based on the activation energy of each reaction step, the free energy difference between reaction intermediates, and the change in the catalyst structure from the analyzed catalytic reaction mechanism; and a step of defining a fitness function based on a multivariate linear regression-based statistical model that predicts experimental result values from the derived electronic structural indices and stereostructural indices.
[0013] 3. In the above 1-2 specific examples, the electronic structural indicator may be at least one indicator selected from the group consisting of the charge of the metallocene central metal ion, orbital energy, and electron occupancy of a specific orbital.
[0014] 4. In the above 1-3. specific example, the stereostructural indicator may be at least one indicator selected from the group consisting of the dihedral angle between the upper ligand ring and the lower ligand ring, the average bond length between the metallocene central metal ion and the upper ligand ring or the lower ligand ring, and the bond angle of the upper ligand ring-metallocene central metal-lower ligand ring.
[0015] 5. In the above 1-4 specific examples, the fitness function may be one function selected from the group consisting of a catalytic activity function, a copolymerization ratio function, a polymer molecular weight function, a polymer hydrogen reactivity function, a polymer molecular weight distribution function, a polymer density function, a polymer transparency function, a polymer MI function, a polymer MFR function, a polymer Tg function, a polymer melting point function, and a polymer mechanical property function.
[0016] 6. In the above 1-5 specific examples, the copolymerization ratio function may be a function defined by [Mathematical Formula 3], which is derived from [Mathematical Formula 1] and [Mathematical Formula 2].
[0017] [Mathematical Formula 1]
[0018] ΔΔG= -3.206×M charge + 0.016×CpMCp - 0.038×CpCp - 12.344×MCp - 0.351×MX + 33.933
[0019] (In the above mathematical formula 1,
[0020] M charge is the charge of the metallocene central metal ion, CpMCp is the bond angle of the upper ligand ring-metallocene central metal-lower ligand ring, CpCp is the dihedral angle between the upper ligand ring and the lower ligand ring, MCp is the average bond length between the metallocene central metal ion and the upper ligand ring or the lower ligand ring, and MX is the average bond length between the metallocene central metal ion and the halogen ion.
[0021] [Equation 2]
[0022] v = e ΔΔG / RT
[0023] (In Equation 2, R is the gas constant and T is the absolute temperature)
[0024]
[0025] [Equation 3]
[0026] Copolymerization ratio function =
[0027] 7. In the above 1-6 specific examples, in the step of forming a group by selecting a metallocene catalyst having a value higher than a preset value from the fitness function value (e), the preset value of the copolymerization ratio function among the fitness functions can be selected as the top 50% value.
[0028] 8. In the above specific examples 1-7, the catalytic activity function may be a conversion frequency function defined by [Mathematical Formula 7], derived from [Mathematical Formula 4] to [Mathematical Formula 6] below.
[0029] [Equation 4]
[0030] ΔG D = 86.868n4d -169.6 (n 4d < 2.17)
[0031] -48.979n 4d + 125.32 (n 4d > 2.17)
[0032] (In the above mathematical formula 4,
[0033] n 4d is the electron occupancy of the d orbital of the metallocene central metal ion)
[0034] [Equation 5]
[0035] ΔG Monomer = -0.1653 θ + 32.903
[0036] (In the above mathematical expression 5,
[0037] θ is the open angle between the upper and lower ligand rings measured in the pre-catalyst structure)
[0038] [Equation 6]
[0039] ΔG RXN = ΔG D + ΔG Monomer
[0040] [Equation 7]
[0041] Conversion frequency (10 6 h -1 ) = -0.6207 ΔG RXN + 20.3
[0042] 9. In the above 1-8 specific examples, in the step of forming a group by selecting a metallocene catalyst having a value higher than a preset value from the fitness function value (e), the preset value of the catalyst activity function among the fitness functions can be selected from the top 50% values.
[0043] 10. In the above 1-9 specific examples, the polymer molecular weight function may be a function defined by [Mathematical Formula 9], derived from [Mathematical Formula 8] below.
[0044] [Equation 8]
[0045] M wrZrCp= -27141000 ×MCp + 60339000 (MCp < 2.256)
[0046] 4137300 ×MCp - 9131700 (MCp > 2.256)
[0047] (In the above mathematical expression 8,
[0048] MCp is the average bond length between the metallocene central metal ion and the upper or lower ligand ring.
[0049] [Equation 9]
[0050] Polymer molecular weight = 0.8075×M wrZrCp - 9419.03×CpCp+485525.7×d Eth-H + n 4d + 5455121
[0051] (In the above mathematical formula 9,
[0052] CpCp is the dihedral angle between the upper and lower ligand rings, and n 4d is the electron occupancy of the d orbital of the metallocene central metal ion, and d Eth-H is the distance between the hydrogens forming the β-agostic bond of the olefin monomer ethylene and the zirconium ion.
[0053] 11. In the above 1-10 specific examples, in the step of forming a group by selecting a metallocene catalyst having a value higher than a preset value from the fitness function value (e), the preset value of the polymer molecular weight function among the fitness functions can select the top 50% value.
[0054] 12. In the above 1-11 specific examples, in the step of generating offspring individuals by performing crossing and mutation on the metallocene catalysts within the group (f), the step of (f-1) crossing the metallocene catalysts within the group according to a crossing rate; and (f-2) generating offspring individuals by mutating the metallocene catalysts within the crossed group according to a mutation rate.
[0055] 13. In the above specific examples 1-12, the crossover rate may be 0.3 to 0.5.
[0056] 14. In the above specific examples 1-13, the mutation rate may be 0.3 to 0.5.
[0057] 15. Another aspect of the present invention relates to a method for predicting the performance of a metallocene catalyst. The method comprises the steps of calculating electronic structural indices and stereostructural indices from a target metallocene catalyst using density functional theory (DFT); obtaining a fitness function value from the electronic structural indices and stereostructural indices; and predicting the performance of the metallocene catalyst using the fitness function.
[0058] 16. Another aspect of the present invention relates to a metallocene catalyst modeling device. The device includes a processor that defines a fitness function related to catalyst properties using electronic structural indices and stereostructural indices, calculates electronic structural indices and stereostructural indices using density functional theory (DFT), and obtains a fitness function value from the electronic structural indices and stereostructural indices; and a memory necessary for the processor to perform calculations.
[0059] 17. In the above 16 specific examples, the processor can select metallocenes based on the fitness function value, and perform a genetic algorithm on the selected metallocene catalysts to derive a new catalyst structure optimized for each performance.
[0060] 18. In the above 16-17 specific examples, the genetic algorithm may perform crossover and mutation on the selected metallocene catalysts to generate offspring individuals, obtain a fitness function value for the generated offspring individuals, and select again to generate offspring individuals of the next generation, repeating the process multiple times.
[0061] 19. In the above 16 specific examples, the processor can predict the metallocene catalyst performance by the fitness function value.
[0062] 20. In the above 16-19 specific examples, the electronic structural indicator is at least one indicator selected from the group consisting of the charge of the metallocene central metal ion, orbital energy, and electron occupancy of a specific orbital, the stereostructural indicator is at least one indicator selected from the group consisting of the dihedral angle between the upper ligand ring and the lower ligand ring, the average bond length between the metallocene central metal ion and the upper ligand ring or the lower ligand ring, and the bond angle of the upper ligand ring-metallocene central metal-lower ligand ring, and the fitness function may be at least one function selected from the group consisting of a catalytic activity function, a copolymerizability ratio function, a polymer molecular weight function, a polymer hydrogen reactivity function, a polymer molecular weight distribution function, a polymer density function, a polymer transparency function, a polymer MI function, a polymer MFR function, a polymer Tg function, a polymer melting point function, and a polymer mechanical property function.
[0063]
[0064] The metallocene catalyst derivation method not only predicts metallocene catalyst performance such as catalytic activity, copolymerization ratio, and polymer molecular weight through a fitness function defined by density functional theory (DFT), electronic structural indices, and stereostructural indices, but also effectively derives a new catalyst structure optimized for each performance through a genetic algorithm, so it has the advantage of improving economic feasibility and efficiency when designing metallocene catalysts.
[0065]
[0066] Figure 1 is a schematic diagram showing the pre-catalyst structure.
[0067] Figure 2 is a diagram schematically showing the definitions of electronic structural indicators and stereostructural indicators.
[0068] Figure 3 is a diagram showing the types of ligands corresponding to the top or bottom of the metallocene catalyst in the population.
[0069] Figure 4 is a diagram showing the types of bridges corresponding to the interruption of the metallocene catalyst in the population.
[0070] Figure 5 is a diagram showing a metallocene catalyst obtained as a final offspring entity by a method for deriving a copolymerization ratio function.
[0071] Figure 6 is a diagram showing the progress of an automated quantum computing program calculation and index value extraction code for obtaining information on metallocene catalyst progeny individuals and a genetic algorithm code for optimizing characteristic values (fitness function values).
[0072] Figure 7 is a schematic diagram of a metallocene catalyst modeling device according to one specific example of the present invention.
[0073]
[0074] The above objectives, other objectives, features, and advantages will be readily understood through the following preferred embodiments, illustrated in the accompanying drawings. However, the embodiments described herein are not limited to the embodiments described herein and may be embodied in other forms. Rather, the embodiments introduced herein are provided to ensure that the disclosed content is thorough and complete, and to sufficiently convey the technical concepts to those skilled in the art.
[0075] In this specification, terms such as “include” or “have” are intended to specify the presence of a feature, number, step, operation, component, part, or combination thereof described in the specification, but should be understood not to exclude in advance the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.
[0076] Unless otherwise specified, all numbers, values, and / or expressions expressing quantities of ingredients, reaction conditions, polymer compositions, and blends used herein are approximations that inherently reflect, among other things, the various uncertainties of measurement that arise in obtaining such values, and therefore should be understood as being modified in all instances by the term "about." Furthermore, whenever a numerical range is disclosed herein, such range is continuous and includes every value from the minimum value to the maximum value inclusive, unless otherwise indicated. Furthermore, whenever such a range refers to an integer, every integer from the minimum value to the maximum value inclusive, unless otherwise indicated, is included.
[0077] In this specification, when a range is described for a variable, the variable will be understood to include all values within the described range including the described endpoints of the range. For example, the range "5 to 10" will be understood to include the values 5, 6, 7, 8, 9, and 10, as well as any subranges such as 6 to 10, 7 to 10, 6 to 9, 7 to 9, etc., and any value between integers that fall within the described range, such as 5.5, 6.5, 7.5, 5.5 to 8.5, and 6.5 to 9. Also, for example, a range of "10% to 30%" would be understood to include all integers up to and including 30%, as well as any subranges such as 10% to 15%, 12% to 18%, 20% to 30%, etc., and any value between reasonable integers within the stated range, such as 10.5%, 15.5%, 25.5%, etc.
[0078] Previously, research on screening methods for deriving metallocene catalysts for excellent synthesis of polyolefins using computer simulations was insufficient.
[0079] Accordingly, the inventors of the present invention conducted extensive research to solve the above-mentioned problem, and as a result, calculated electronic structural index values and stereostructural index values from metallocene catalysts within a metallocene catalyst population divided into upper, bridge, and lower groups according to the catalyst structure through the density functional theory (DFT) theory, introduced these into a fitness function to obtain a fitness function value, and then selected metallocene catalysts having a fitness function value higher than a preset value to form a group, and performed a genetic algorithm on the formed group to derive a metallocene catalyst. In this case, not only can a specific catalyst performance be predicted through the density functional theory (DFT) and the fitness function, but a new catalyst structure optimized for each performance can be effectively derived through the genetic algorithm, and this has been completed.
[0080] A method for deriving a metallocene catalyst according to one aspect comprises: (a) a step of defining a fitness function related to catalyst properties using electronic structural indices and stereostructural indices (S10); (b) a step of forming a population of metallocene catalysts divided into upper, middle, and lower groups according to catalyst structures (S20); (c) a step of calculating electronic structural indices and stereostructural indices of the metallocene catalyst corresponding to the electronic structural indices and stereostructural indices of the fitness function using density functional theory (DFT) (S30); (d) a step of obtaining a fitness function value from the electronic structural indices and stereostructural indices (S40); (e) a step of forming a group by selecting metallocene catalysts having a value equal to or greater than a preset value from the fitness function values (S50); (f) a step of generating offspring entities by performing crossover and mutation on the metallocene catalysts within the group (S60); And (g) a step (S70) of repeating steps (c) to (f) multiple times for the generated offspring individuals, and terminating the step.
[0081] Depending on the implementation example, each step may be performed on a workstation / cluster.
[0082] Step (a) (S10) according to an embodiment is a step of defining a fitness function related to catalyst properties using electronic structural indices and stereostructural indices.
[0083] According to an embodiment, the fitness function may be of various types depending on the specific characteristics or performance of the catalyst, and may be, for example, one function selected from the group consisting of a catalytic activity function, a copolymerizability ratio function, a polymer molecular weight function, a polymer hydrogen reactivity function, a polymer molecular weight distribution function, a polymer density function, a polymer transparency function, a polymer MI function, a polymer MFR function, a polymer Tg function, a polymer melting point function, and a polymer mechanical property function.
[0084] According to one embodiment, the step of defining a fitness function using electronic structural indices and stereostructural indices may include: analyzing a catalytic reaction mechanism from an arbitrary metallocene catalyst; deriving electronic structural indices and stereostructural indices based on the activation energy of each reaction step, the free energy difference between reaction intermediates, and the structural change of the catalyst from the analyzed catalytic reaction mechanism; and defining a fitness function based on a multivariate linear regression-based statistical model that predicts experimental result values from the derived electronic structural indices and stereostructural indices.
[0085] According to one embodiment, the step of deriving electronic structural indices and stereostructural indices from the catalytic reaction mechanism includes designing a pre-catalyst structure before an active state, and then, when the catalytic reaction mechanism progresses by introducing an olefin monomer, identifying the activation energy of each reaction step, the free energy difference between reaction intermediates, and the structural change of the catalyst, and then deriving electronic structural indices and stereostructural indices based on this. At this time, Fig. 1 is a schematic diagram showing the pre-catalyst structure. Referring to this, in the catalytic reaction of the pre-catalyst structure copolymerization, the structure of the catalyst before showing activity with the olefin monomer may be a metallocene catalyst structure in which a halogen (X) is bonded to a metallocene central metal (M), and preferably, a zirconium metallocene structure in which chlorine is bonded.
[0086] In one embodiment, a multivariate linear regression based statistical model predicts the dependent variable (outcome) Y=β1X1+ β2X2+… + β p X p It can be an equation with two or more independent variables (characteristics) expressed as +ε.
[0087] That is, the fitness function can be defined by analyzing each mechanism according to the catalyst characteristics, deriving the optimal electronic structural indicators and stereostructural indicators suitable for each characteristic, and then defining the fitness function accordingly.
[0088] According to one implementation example, the electronic structural indicator of the fitness function may be one or more indicators selected from the group consisting of charge of the metallocene central metal ion, orbital energy, electron occupancy of a specific orbital, Highest-Occupied Molecular Orbital (HOMO), Lowest-Unoccupied Molecular Orbital (LUMO), Mulliken charge, Atomic polar tensor (APT) charge, and Natural Population Analysis (NPA) charge.
[0089] In one embodiment, the metallocene central metal may be zirconium (Zr), hafnium (Hf), titanium (Ti), or the like, and in one embodiment, the metallocene central metal may be zirconium (Zr), and the electron occupancy of a specific orbital may be the electron occupancy of a zirconium orbital.
[0090] In one implementation example, the conformational indices of the fitness function include the dihedral angle between the upper ligand ring and the lower ligand ring, the average bond length between the metallocene central metal ion and the upper ligand ring or the lower ligand ring, the bond angle of the upper ligand ring-metallocene central metal-lower ligand ring, and the percent buried volume (%V bur ), quantitative value of dispersing power (P int ), Solvent-Accessible Surface Area (SASA), and Cone angle.
[0091] According to one implementation example, the copolymerization ratio function among the fitness functions may be a function defined to secure the characteristics of a metallocene catalyst capable of producing a polymer having a high comonomer ratio. Specifically, based on the degree of steric hindrance between each olefin monomer (Ethylene, 1-hexene) and the upper ligand ring or the lower ligand ring from the catalytic reaction mechanism, the degree of interaction between the metallocene central metal ion and each olefin monomer, and the difference in binding free energy of each reaction intermediate, electronic structural indices (charge value of the metallocene central metal ion) and stereostructural indices (dihedral angle of the upper ligand ring and the lower ligand ring, bond length and bond angle between the metallocene central metal ion and the upper ligand ring or the lower ligand ring) may be derived, and a function, which is derived from [Mathematical Formula 1] and [Mathematical Formula 2] below, may be defined as [Mathematical Formula 3] according to a multivariate linear regression statistical model.
[0092] According to one embodiment, the copolymerization ratio function may be a function defined to secure the characteristics of a metallocene catalyst capable of producing a polymer having a high proportion of 1-hexene as a comonomer. Specifically, the following mathematical expression 1 can be derived using multivariate linear regression to derive the insertion reaction activation energy difference (ΔΔG) between ethylene and 1-hexene as olefin monomers. By deriving the activation energy difference, the copolymerization ratio of ethylene and 1-hexene of the produced polymer can be predicted and confirmed.
[0093] [Mathematical Formula 1]
[0094] ΔΔG= -3.206×M charge + 0.016×CpMCp - 0.038×CpCp - 12.344×MCp - 0.351×MX + 33.933
[0095] In the above mathematical formula 1,
[0096] M chargeis the charge of the metallocene central metal ion, CpMCp is the bond angle of the upper ligand ring-metallocene central metal-lower ligand ring, CpCp is the dihedral angle between the upper ligand ring and the lower ligand ring, MCp is the average bond length between the metallocene central metal ion and the upper ligand ring or the lower ligand ring, and MX may be the average bond length between the metallocene central metal ion and the halogen ion.
[0097] Figure 2 is a schematic diagram showing the definition of electronic structural indicators and stereostructural indicators. Referring to Figure 2, M charge may be the charge of the metallocene central metal ion. In addition, CpMCp may be the bond angle of the upper ligand ring-metallocene central metal-lower ligand ring, and may be the angle (β) formed by the center of the upper ligand ring (Cp1), the center of the metallocene central metal (M), and the center of the lower ligand ring (Cp2). CpCp may be the dihedral angle (α) formed by the plane of the upper ligand ring (Cp1) and the plane of the lower ligand ring (Cp2). In addition, MCp may be the average of the distance (L1) from the center of the metallocene central metal (M) to the center of the plane of the upper ligand ring (Cp1) and the distance (L2) from the center of the metallocene central metal (M) to the center of the plane of the lower ligand ring (Cp2). Additionally, MX may be the average of the distance (L3) from the metallocene central metal (M) center to the halogen (X, preferably, Cl1), and the distance (L4) from the metallocene central metal (M) center to the halogen (X, preferably, Cl2).
[0098] In one embodiment, M charge can be the charge of the metallocene central metal, and Cp1 and Cp2 can independently be a substituted or unsubstituted cyclopentadienyl group, indenyl group, or fluorenyl group. At this time, in the substituted or unsubstituted, the substituents are independently a halogen group, a cyano group, a nitro group, a C1-C 20 Alkyl group (C1-C 10Silane group substituted with alkyl group can be substituted), C1-C 20 Aryl group (C1-C 10 Silane group substituted with alkyl group can be substituted), C1-C 20 Allyl group (C1-C 10 It may be at least one selected from the group consisting of a silane group substituted with an alkyl group), and a naphthalenyl group. The metallocene central metal (M) may be zirconium. And the halogen X may be chlorine (Cl).
[0099] According to one embodiment, an equation such as Equation 2 can be derived to calculate the relative rate ratio at which olefin monomers, preferably, ethylene and 1-hexene, are inserted into a polymer using the difference in activation energy calculated from Equation 1.
[0100] [Equation 2]
[0101] v = e ΔΔG / RT
[0102] (In the above mathematical equation 2, R is the gas constant and T is the absolute temperature)
[0103] Using the relative rate ratio calculated from the above mathematical equation 2, a copolymerization rate function such as mathematical equation 3 can be derived to predict the copolymerization rate from the rate at which ethylene and 1-hexene are inserted into the polymer.
[0104] [Equation 3]
[0105]
[0106]
[0107] In one embodiment, the catalytic activity function among the fitness functions may be a function defined to secure a characteristic having a high conversion frequency of the polymerization reaction. At this time, based on the activation energy of each reaction step from the catalytic reaction mechanism and the free energy difference between the reactants, intermediates, and products along the entire polymerization reaction path, an electronic structural index (metallocene central metal ion charge value) and a stereostructural index (open angle, which is the angle formed by the metallocene central metal ion and the carbon atoms of the upper ligand ring and the lower ligand ring) are derived, and a conversion frequency function, which is [Mathematical Formula 7], derived from [Mathematical Formula 4], [Mathematical Formula 5], and [Mathematical Formula 6], may be defined based on the correlation between the activation energy and free energy difference values and the indices of the two groups.
[0108] [Equation 4]
[0109] ΔG D = 86.868n 4d -169.6 (n 4d < 2.17)
[0110] -48.979n 4d + 125.32 (n 4d > 2.17)
[0111] In the above mathematical expression 4, n 4d is the electron occupancy of the d orbital of the metallocene central metal ion. In a specific example, it may be the electron occupancy of the 4d orbital of the zirconium ion.
[0112] [Equation 5]
[0113] ΔG Monomer = -0.1653 θ + 32.903
[0114] In the above mathematical expression 5, θ may be the open angle of the upper ligand ring and the lower ligand ring measured in the pre-catalyst structure. At this time, the pre-catalyst structure is the structure of the catalyst before showing activity with the olefin monomer in the catalytic reaction of copolymerization, and may be a metallocene catalyst structure in which a halogen (X) is bonded to a metallocene central metal (M), and preferably, may be a zirconium metallocene structure in the form of a chlorine bond.
[0115] [Equation 6]
[0116] ΔG RXN = ΔG D + ΔG Monomer
[0117]
[0118] [Equation 7]
[0119] Conversion frequency (10 6 h -1 ) = -0.6207 ΔG RXN + 20.3
[0120]
[0121] According to one embodiment, the polymer molecular weight function among the fitness functions may be a function defined to secure the characteristic of producing a polymer with a large molecular weight. At this time, based on the fact that the termination reaction that determines the molecular weight of the polymer from the catalytic reaction mechanism is affected by changes in the structure and electronic structure of the catalyst, an electronic structural index (electron occupancy of the d orbital of the metallocene central metal ion) and a stereostructural index (average bond length of the metallocene central metal ion and the upper ligand ring or the lower ligand ring, dihedral angle between the upper ligand ring and the lower ligand ring, distance between hydrogens forming a β-agostic bond with an olefin monomer and the metallocene central metal ion) may be derived, and a function, which is [Mathematical Formula 9] derived from [Mathematical Formula 8] below, may be defined according to a multivariate linear regression-based statistical model.
[0122] [Equation 8]
[0123] M wrZrCp= -27141000 ×MCp + 60339000 (MCp < 2.256)
[0124] 4137300 ×MCp - 9131700 (MCp > 2.256)
[0125]
[0126] In the above mathematical formula 8, MCp may be the average bond length between the metallocene central metal ion and the upper ligand ring or the lower ligand ring. In this case, the metallocene central metal ion may be a zirconium ion.
[0127] [Equation 9]
[0128] Polymer molecular weight = 0.8075×M wrZrCp - 9419.03×CpCp + 485525.7×d Eth-H + n 4d + 5455121
[0129]
[0130] In the above mathematical formula 9, CpCp is the dihedral angle between the upper ligand ring and the lower ligand ring, and n 4d is the electron occupancy of the d orbital of the metallocene central metal ion, and d Eth-H may be the distance between the hydrogens forming the β-agostic bond of the olefin monomer and the metallocene central metal ion. In this case, the metallocene central metal may be zirconium, and the olefin monomer may be ethylene.
[0131] Step (b) (S20) according to an embodiment is a step of forming a population of metallocene catalysts divided into upper, middle, and lower parts according to the catalyst structure.
[0132] According to one embodiment, the upper end according to the catalyst structure is an upper ligand ring located at the upper end based on the metallocene central metal, which may be a substituted or unsubstituted cyclopentadienyl group, indenyl group, or fluorenyl group. In this case, in the substituted or unsubstituted case, the substituents are independently a halogen group, a cyano group, a nitro group, or a C1-C 20 Alkyl group (C1-C 10Silane group substituted with alkyl group can be substituted), C1-C 20 Aryl group (C1-C 10 Silane group substituted with alkyl group can be substituted), C1-C 20 Allyl group (C1-C 10 It may be at least one selected from the group consisting of a silane group substituted with an alkyl group, a silane group substituted with an alkyl group, and a naphthalenyl group. According to one embodiment, the upper end according to the catalyst structure may be a ligand ring represented by Lig_0 to Lig_34 (see FIG. 3).
[0133] In one embodiment, the interruption according to the catalyst structure may be a bridge portion connecting the top and bottom within the metallocene catalyst. In one embodiment, the interruption according to the catalyst structure may be a bridge represented by Br_0 to Br_13 (see FIG. 4).
[0134] According to one embodiment, the lower ligand ring located at the lower end based on the metallocene central metal according to the catalyst structure may be a substituted or unsubstituted cyclopentadienyl group, indenyl group, or fluorenyl group. In this case, in the substituted or unsubstituted case, the substituents are independently a halogen group, a cyano group, a nitro group, or a C1-C 20 Alkyl group (C1-C 10 Silane group substituted with alkyl group can be substituted), C1-C 20 Aryl group (C1-C 10 Silane group substituted with alkyl group can be substituted), C1-C 20 Allyl group (C1-C 10 It may be at least one selected from the group consisting of a silane group substituted with an alkyl group), and a naphthalenyl group. In one embodiment, the lower part according to the catalyst structure may be a ligand ring indicated as Lig_0 to Lig_34 (see FIG. 3).
[0135] In one embodiment, the top, middle, and bottom metallocene catalysts can be populated from code (gen_3d_structure.py) that combines three partial structures to form one complete metallocene catalyst structure.
[0136] According to one implementation example, the electronic structural indicator value or stereostructural indicator value of the subsequent step can be obtained from each of the metallocene catalysts included in the population obtained from step (b).
[0137] Step (c) (S30) according to an embodiment is a step of obtaining electronic structural index values and stereostructural index values corresponding to the electronic structural index and stereostructural index of the fitness function defined in step (a) through calculation using density functional theory (DFT) from each metallocene catalyst in the population formed in step (b).
[0138] According to one embodiment, the electronic structural index value of the metallocene catalyst may be at least one index value selected from the group consisting of charge, orbital energy, electron occupancy of a specific orbital, Highest-Occupied Molecular Orbital (HOMO), Lowest-Unoccupied Molecular Orbital (LUMO), Mulliken charge, Atomic polar tensor (APT) charge, and Natural Population Analysis (NPA) charge of each metallocene central metal ion in a population of the metallocene catalyst. Preferably, the electronic structural index value may be a specific fitness function according to a specific catalyst property among fitness functions, for example, a catalytic activity function, a copolymerizability ratio function, or a polymer molecular weight function.
[0139] According to one implementation example, the electronic structural index calculation for obtaining the electronic structural index value can be calculated through a code (utils.py) that extracts the above type of index value from the output value of the calculation through density functional theory.
[0140] According to one embodiment, the stereostructural index values of the metallocene catalyst are the dihedral angle between the upper ligand ring and the lower ligand ring of each metallocene catalyst in the metallocene catalyst population, the average bond length between the metallocene central metal ion and the upper ligand ring or the lower ligand ring, the bond angle of the upper ligand ring-metallocene central metal-lower ligand ring, and the percent buried volume (%V bur ), quantitative value of dispersing power (P int ), Solvent-Accessible Surface Area (SASA), and Cone angle. Preferably, it may be a stereostructural index value corresponding to a specific fitness function according to a specific catalyst property among the fitness functions, for example, a catalytic activity function, a copolymerization ratio function, or a polymer molecular weight function.
[0141] According to one implementation example, the calculation of the stereostructural index to obtain the stereostructural index value can be calculated through a code (metallocene.py) that extracts the above type of index value from the output value of the calculation through density functional theory.
[0142] According to an embodiment, the electronic structural indicator values and stereostructural indicator values obtained from step (c) can be used to obtain the fitness function values of the subsequent step.
[0143] Step (d) (S40) according to an embodiment is a step of obtaining a fitness function value of a fitness function defined by electronic structural indices and stereostructural indices from the electronic structural indices and stereostructural indices obtained in step (c).
[0144] According to the implementation example, the fitness function calculation to obtain the fitness function value can be calculated through a multivariate linear regression-based statistical model included in the utils.py code based on the indicator values extracted from the output of the calculation through the density functional theory.
[0145] According to an embodiment, the fitness function value obtained from step (d) can be used to select a metallocene catalyst that exceeds a preset value by checking whether the fitness function value exceeds a preset value set for each fitness function in a subsequent step.
[0146] That is, the metallocene catalyst derivation method has the advantage of being able to effectively predict metallocene catalyst performance, such as catalytic activity, copolymerization ratio, and polymer molecular weight, through a fitness function defined by density functional theory (DFT), electronic structural indicators, and stereostructural indicators.
[0147] Step (e) (S50) according to an embodiment is a step of forming a group by selecting metallocene catalysts that are at least in the top 50% (preset value) for each specific fitness function from the fitness function values obtained from step (d).
[0148] Step (f) (S60) according to one embodiment is a step of generating offspring individuals by performing crossover and mutation from a group in which only the metallocene catalysts with a top 50% or higher fitness function value are selected. Specifically, the step may include: (f-1) a step of crossing the metallocene catalysts within the group according to a crossover rate; and (f-2) a step of generating offspring individuals by mutating the metallocene catalysts within the crossed group according to a mutation rate.
[0149] According to one embodiment, the crossing performed in step (f) means crossing between ligands or bridges between the metallocene catalysts within the group formed in step (e), specifically, the ligands or bridges corresponding to the top, middle, or bottom of each catalyst between the metallocene catalysts within the group formed in step (d) may cross depending on the crossing rate.
[0150] According to one implementation, the crossover may occur according to the crossover rate. In this case, the crossover rate represents the probability that one metallocene catalyst in the group will crossover with another metallocene catalyst, and may be 0.3 to 0.5. For example, if the crossover rate is 0.1, there is a 10% probability that the ligand or bridge of one metallocene catalyst in the group will crossover with the ligand or bridge of another metallocene catalyst with a 10% probability. If the crossover rate is too low, the range of metallocene catalyst structures to be searched in one step will be narrowed, which may lead to a disadvantage in that the search time may be very long or only a narrow range may be repeatedly searched when finding the optimum point. In addition, if the crossover rate is too high, there is a disadvantage in that the metallocene structure being searched in one step may differ greatly from the metallocene structure of the optimum point due to excessive structural change during the optimization process, which may increase the possibility that the fitness function values may diverge.
[0151] In one embodiment, the mutation performed in step (f) means that the ligand or bridge of the metallocene catalyst, whether or not the crossing is performed, is mutated to another ligand or bridge.
[0152] In one embodiment, the mutation may be a mutation rate. The mutation rate represents the probability that the ligand or bridge corresponding to the top, middle, or bottom of each metallocene catalyst in the group will be mutated to a different ligand or bridge, and may be 0.3 to 0.5. For example, if the mutation rate is 0.001, there is a 0.1% probability that the ligand or bridge of the metallocene catalyst will be mutated to a different ligand or bridge, or that the different ligand or bridge will be mutated to a ligand or bridge of the metallocene catalyst.
[0153] In step (g) (S70) according to an embodiment, the object (metallocene catalyst) performed up to step (f) is treated as a descendant object, the fitness function value of the descendant object is obtained, and the group formation, crossover, and mutation processes can be performed repeatedly again.
[0154] Specifically, the offspring individuals generated according to the step (f) above can be generated by obtaining the electronic structural index values and stereostructural index values of the offspring individuals in the step (c), obtaining the fitness function value in the step (d), grouping the metallocene catalysts having a fitness function value greater than or equal to a preset value in the step (e), and performing crossover and mutation in the step (f). At this time, the preset values in the offspring individuals can be adjusted to suit the purpose, such as the top 50% or more, the top 60% or more, the top 70% or more, and the top 80% or more.
[0155] In one embodiment, the above-described iterative process may be repeated until a target number of generations is reached, and the number of generations may be set according to circumstances. Preferably, the above-described offspring generation process may be repeated 2 to 50 times. If the number of generations is too low, the range of metallocene catalyst structures to be explored may be narrowed, and if the number of generations is too high, the optimization of a specific fitness function value based on a genetic algorithm may take a long time.
[0156] In one embodiment, the preset value of the copolymerization ratio function for the final offspring individual among the fitness functions may be selected from the top 20% of values or may be 0.11 or higher. When the preset value is satisfied, there is an advantage in that the content ratio of the comonomer (1-hexene) of the polymer produced through metallocene polymerization is high.
[0157] In one embodiment, the preset value of the catalytic activity function for the final offspring individual among the fitness functions may be selected from the top 20% of values or may be 4,500,000 / hour or more. Satisfying the preset value has the advantage of enabling rapid polymerization reaction and rapid production of products.
[0158] In one embodiment, the preset value of the polymer molecular weight function for the final offspring entity among the fitness functions may be selected from the top 20% of values or may be 350,000 or greater. This has the advantage that, when the preset value is satisfied, a polymer having a molecular weight greater than the preset value can be generated, thereby producing a product having the target properties.
[0159] In other words, since the metallocene catalyst derivation method can effectively derive a new catalyst structure optimized for each performance with a simplified computational amount through a genetic algorithm without considering all cases of possible structures of the metallocene catalyst and without calculating all fitness function values of all cases, it has the advantage of improving economic feasibility and efficiency when designing a metallocene catalyst.
[0160] Another aspect of the present invention relates to a metallocene catalyst modeling device. Fig. 7 is a schematic diagram of a metallocene catalyst modeling device according to one embodiment of the present invention.
[0161] As illustrated in FIG. 7, the modeling device includes a processor (10) and a memory (20). The processor (10) defines a fitness function related to catalyst characteristics as an electronic structural index and a stereostructural index, calculates electronic structural index values and stereostructural index values through density functional theory (DFT), and obtains a fitness function value from the electronic structural index values and stereostructural index values.
[0162] In one specific example, the processor (10) can select metallocenes based on the fitness function value and perform a genetic algorithm on the selected metallocene catalysts to derive a new catalyst structure optimized for each performance.
[0163] The above genetic algorithm can repeat the process of performing crossover and mutation on the selected metallocene catalysts to generate offspring individuals, obtaining a fitness function value for the generated offspring individuals, and selecting again to generate offspring individuals of the next generation multiple times.
[0164] In another specific example, the processor (10) can predict the metallocene catalyst performance by the fitness function value.
[0165] The above electronic structural indicator is at least one indicator selected from the group consisting of the charge of the metallocene central metal ion, orbital energy, and electron occupancy of a specific orbital, the above stereostructural indicator is at least one indicator selected from the group consisting of the dihedral angle between the upper ligand ring and the lower ligand ring, the average bond length between the metallocene central metal ion and the upper ligand ring or the lower ligand ring, and the bond angle of the upper ligand ring-metallocene central metal-lower ligand ring, and the above fitness function may be at least one function selected from the group consisting of a catalytic activity function, a copolymerizability ratio function, a polymer molecular weight function, a polymer hydrogen reactivity function, a polymer molecular weight distribution function, a polymer density function, a polymer transparency function, a polymer MI function, a polymer MFR function, a polymer Tg function, a polymer melting point function, and a polymer mechanical property function.
[0166] The above memory (20) can provide storage space required for the processor (10) to perform operations.
[0167]
[0168] The present invention will be described in more detail below through examples. However, these examples are intended to exemplify the present invention and the scope of the present invention is not limited to these examples.
[0169]
[0170]
[0171] Example 1: Method for deriving a metallocene catalyst (copolymerization ratio function)
[0172] <Definition of the fitness function>
[0173] The catalytic reaction mechanism was analyzed using a quantum computational program based on density functional theory for a random metallocene catalyst. Then, based on the steric hindrance effect between the monomer / comonomer and the upper or lower ligand ring, the difference in interaction with the zirconium ion, and the free energy difference of each reaction intermediate, electronic structural parameters (zirconium ion charge) and stereostructural parameters (dihedral angle between the upper and lower ligand rings, bond angles between the zirconium ion and the upper or lower ligand rings, and distance between the zirconium ion and the center of the upper or lower ligand ring) were derived from the catalytic reaction mechanism. Then, a goodness-of-fit function was defined based on a multivariate linear regression-based statistical model from the derived electronic and stereostructural parameters. Specifically, the function [Equation 3], which is derived from [Equation 1] and [Equation 2] below, was defined.
[0174] Specifically, the following mathematical equation 1 can be derived using multivariate linear regression to derive the insertion reaction activation energy difference (ΔΔG) between ethylene and 1-hexene. At this time, ethylene and 1-hexene were selected because they are the basic units of a general PO polymerization reaction, and by deriving the insertion reaction activation energy difference (ΔΔG), the copolymerization ratio can be predicted from the insertion ratio of ethylene and 1-hexene into the polymer.
[0175] [Mathematical Formula 1]
[0176]
[0177] At this time, in the above mathematical expression 1, M charge is the charge of the zirconium ion, which is the metallocene central metal ion, CpMCp is the bond angle of the upper ligand ring-zirconium ion-lower ligand ring, CpCp is the dihedral angle between the upper ligand ring and the lower ligand ring, MCp is the average bond length between the zirconium ion and the upper ligand ring or the lower ligand ring, and MX is the average bond length between the zirconium ion and the halogen chloride ion.
[0178] Using the difference in activation energy from the above mathematical equation 1, an equation such as mathematical equation 2 was derived to calculate the relative rate ratio at which ethylene and 1-hexene are inserted into the polymer.
[0179] [Equation 2]
[0180] v = e ΔΔG / RT
[0181] In order to calculate the ratio of ethylene and 1-hexene inserted into the polymer from the above mathematical formula 2, a copolymerization ratio function such as mathematical formula 3 was derived.
[0182] [Equation 3]
[0183] Copolymerization ratio function =
[0184]
[0185] <Population formation of metallocene catalysts>
[0186] Figure 3 is a diagram showing the ligand types corresponding to the top or bottom of the metallocene catalyst in the population. Figure 4 is a diagram showing the bridge types corresponding to the middle of the metallocene catalyst in the population.
[0187] Referring to Figures 3 and 4, the upper and middle ligand rings indicated as Lig_0 to Lig_34 were applied, respectively, and the middle used a bridge indicated as Br_0 to Br_13. At this time, Cp in the bridge can be a substituted or unsubstituted cyclopentadienyl group, indenyl group, or fluoroenyl group. At this time, in the substituted or unsubstituted, the substituents are independently a halogen group, a cyano group, a nitro group, a C1-C 20 Alkyl group (C1-C 10 Silane group substituted with alkyl group can be substituted), C1-C 20 Aryl group (C1-C 10 Silane group substituted with alkyl group can be substituted), C1-C 20 Allyl group (C1-C 10 It may be at least one selected from the group consisting of a silane group substituted with an alkyl group), and a naphthalenyl group.
[0188] At this time, the code used to form the population was a code (gen_3d_structure.py) that combines three partial structures of the upper, middle, and lower metallocene catalysts to form one complete metallocene catalyst structure.
[0189] Table 1 shows the metallocene catalysts in the population and their copolymerization function values (fitness function values).
[0190]
[0191] Catalytic CopolymerizationCatalytic CopolymerizationCatalytic Copolymerization33_4_330.177330_1_180.148533_7_330.137033_4_330.177331_2_330.148333_6_180.136633_8_330.168633_1_180.147933_13_180.136533_1_330.168332_12_320.147331_8_320.136430_4_330.168333_4_18 0.147133_14_330.135933_4_300.167632_11_320.147133_8_320.135832_1_330.166031_12_330.146633_5_180.135232_1_310.165932_3_330.146632_14_300.133833_0_330.165731_2_320.145830_8_180.132533_2_330.165533_7_ 300.145518_4_330.131233_1_300.165532_7_320.145533_15_330.129830_1_330.165531_4_180.145318_1_300.128833_4_310.163431_4_180.145332_1_150.126931_4_330.163131_10_330.145332_1_170.126932_4_330.162832_1 2_330.145131_15_320.126632_1_320.162631_13_330.144827_1_330.125833_4_320.162431_9_330.144417_4_330.125431_4_320.161832_8_330.144416_4_330.124830_1_300.161531_5_330.144215_4_330.124633_10_330.160032 5_320.144132_1_140.124033_16_330.159032_16_320.143932_1_140.124033_12_330.158331_7_330.143514_4_320.123233_9_330.158331_11_320.143427_4_330.123133_13_330.157631_5_320.143315_1_300.122033_6_330.157631_5_320.143314_4_330.122033_11_330.157533_3_300.142833_1_160.121931_1_320.156332_3_320.142017_1_330.121932_5_300.151133_0_300.141333_1_170.121933_12_300.150931_3_330.141116_1_330.121830_2_310.150 932_8_320.141015_1_330.121232_16_330.150833_8_300.140433_4_160.121033_3_330.150731_8_330.140433_4_160.121033_11_300.150230_8_310.140333_1_190.120733_2_180.148533_12_180.137333_4_170.1206.
[0192]
[0193] Referring to Table 1, the metallocene catalysts within the population are represented as “top_middle_bottom” respectively. For example, in the case of catalyst (33_4_33), it means that the metallocene catalyst is composed of ‘Ligand(Lig)_33’ in Fig. 3 as the top, ‘Bridge(Br)_4’ in Fig. 4 as the middle, and ‘Ligand(Lig)_33’ as the bottom.
[0194] Accordingly, 102 metallocene catalysts were formed as a population, from catalyst (33_4_33) to catalyst (33_4_17).
[0195] <Calculation of each indicator value and fitness function value using density functional theory (DFT)>
[0196] The electronic structural indices and stereostructural indices of the fitness function and their corresponding respective index values were obtained from each metallocene catalyst in the population through calculations using density functional theory (DFT).
[0197] Specifically, the three-dimensional structural indices, such as the bond length and bond angle between the zirconium ion, the metallocene central metal, and the ligand, and the bond length between the zirconium ion and the halogen chloride ion, were obtained through a quantum computational program based on the density functional theory. The methodology used here is to optimize the metallocene structure by utilizing 1) TPSSTPSS or 2) PBE0 as the functional and def2-svp as the basis function. Each indices value was measured from the output of the calculation using the metallocene.py code.
[0198] The charge value of the zirconium ion, which is an electronic structural index value, was extracted using the utils.py code from the output value of quantum calculation using 1) M06-2X or 2) PBE0 functional and def2-tzvp basis function in the metallocene structure that completed structural optimization using the above method.
[0199] Each of the electronic structural index values and stereostructural index values obtained through the above DFT were substituted into the copolymerization ratio function (Mathematical Formula 1 to Mathematical Formula 3) to obtain the copolymerization ratio value for each metallocene catalyst, and the results are shown in Table 1 above.
[0200] <Selection of metallocene catalysts exceeding a preset value from the population>
[0201] Referring to Table 1, it was confirmed that the metallocene catalyst containing both Lig_30 to Lig_33 at the top and bottom had relatively high copolymerizability. On the other hand, it was confirmed that the metallocene catalyst containing Lig_14 to Lig_18 at either the top or bottom had next-order copolymerizability. In addition, it was confirmed that the metallocene catalyst having a C-bridge as a bridge had better copolymerizability than the metallocene catalyst having a Si-bridge, and it was confirmed that the difference in copolymerizability depending on the presence or absence of an aliphatic group or an aryl group in the bridge was minimal.
[0202] Accordingly, the top 30 catalysts were selected (excluding duplicates) from the copolymerization function values of all metallocene catalysts generated through the selection, crossover, and mutation processes to form a group, which is shown in Table 2.
[0203]
[0204] Catalytic copolymerizationCatalytic copolymerization33_4_330.177330_1_300.161533_8_330.168633_10_330.160033_1_330.168333_16_330.159033_4_300.167633_12_330.158332_1_330.166033_9_330.158332_1_310.165933_13_330.157633_0_330.165733_6_330.157633_2_33 0.165533_11_330.157533_1_300.165531_1_320.156333_4_310.163432_5_300.151131_4_330.163133_12_300.150932_4_330.162830_2_310.150932_1_320.162632_16_330.150833_4_320.162433_3_330.150731_4_320.161833_11_300.1502
[0205]
[0206] Genetic Algorithm
[0207] According to Fig. 6, the genetic algorithm was performed using a code (gen_3d_structure.py) that generates input values for quantum computing program calculations with a metallocene catalyst within the group, a code (GA.py) that obtains the fitness function values of the metallocene catalyst (metallocene.py, utils.py) and evaluates, crosses over, and mutates based on these values, and a code (database.py) that stores each index value and fitness function value of the generated metallocene catalyst structure in a database.
[0208] Specifically, the metallocene catalysts within the group according to Fig. 6 were crossed at a crossing rate of 0.3 using the GA.py code. Then, they were mutated at a mutation rate of 0.5 using the GA.py code.
[0209] Next, offspring individuals are obtained, and the copolymerization ratio function value for the offspring individuals is obtained again, and the metallocene catalysts with values exceeding the top 50% are grouped, and then crossover and mutation are performed again to generate the next generation of offspring individuals. At this time, the number of generations was repeated from 2 to 50, and this was performed using utils.py and main.py (code that inputs setting values for operating the genetic algorithm).
[0210] Accordingly, the metallocene catalysts with the top 20% of the upper copolymerization ratio obtained as the final offspring are shown in Figure 5.
[0211]
[0212] Example 2: Method for deriving a metallocene catalyst (catalytic activity function)
[0213] <Definition of the fitness function>
[0214] The catalytic reaction mechanism was analyzed using a quantum computational program based on density functional theory for a random metallocene catalyst. Then, based on the activation energy of each reaction step and the free energy difference between the reactants, intermediates, and products along the entire polymerization reaction path from the catalytic reaction mechanism, electronic structural parameters (electron occupancy of the 4d orbital of the zirconium ion, the metallocene central metal) and stereostructural parameters (open angles of the upper and lower ligand rings) were derived. Then, a fitness function was defined based on the activation energy and free energy difference values and the correlation between the two groups of parameters from the derived electronic and stereostructural parameters. Specifically, the function [Equation 7], which is derived from [Equations 4], [Equations 5], and [Equations 6] below, was defined.
[0215] Specifically, the activation energy (ΔG RXN ) was divided into two stages depending on the participation of cocatalyst and monomer in the reaction, and the free energy (ΔG) of each reaction stage D ,ΔG Monomer ) can be derived from the following mathematical equations 4 to 6, which can be derived from electronic structural indices and stereostructural indices. Finally, the activation energy (ΔG RXN ) can be used to predict the transition frequency.
[0216] [Equation 4]
[0217]
[0218] In the above mathematical expression 4, n 4d is the electron occupancy of the d orbital of the metallocene central metal ion.
[0219] [Equation 5]
[0220] ΔG Monomer = -0.1653 θ + 32.903
[0221] In the above mathematical expression 5, θ is the open angle of the upper ligand ring and the lower ligand ring measured in the pre-catalyst structure.
[0222] [Equation 6]
[0223] ΔG RXN = ΔG D + ΔG Monomer
[0224]
[0225] [Equation 7]
[0226] Conversion frequency (10 6 h -1 ) = -0.6207 ΔG RXN + 20.3
[0227]
[0228] <Population formation of metallocene catalysts>
[0229] Figure 3 is a diagram showing the ligand types corresponding to the top or bottom of the metallocene catalyst in the population. Figure 4 is a diagram showing the bridge types corresponding to the middle of the metallocene catalyst in the population.
[0230] Referring to Figures 3 and 4, the upper and middle ligand rings indicated as Lig_0 to Lig_34 were applied, respectively, and the middle used a bridge indicated as Br_0 to Br_13. At this time, Cp in the bridge can be a substituted or unsubstituted cyclopentadienyl group, indenyl group, or fluoroenyl group. At this time, in the substituted or unsubstituted, the substituents are independently a halogen group, a cyano group, a nitro group, a C1-C 20 Alkyl group (C1-C 10 Silane group substituted with alkyl group can be substituted), C1-C 20 Aryl group (C1-C 10 Silane group substituted with alkyl group can be substituted), C1-C 20 Allyl group (C1-C 10 It may be at least one selected from the group consisting of a silane group substituted with an alkyl group), and a naphthalenyl group.
[0231] At this time, the code used to form the population was a code (gen_3d_structure.py) that combines three partial structures of the upper, middle, and lower metallocene catalysts to form one complete metallocene catalyst structure.
[0232] Table 3 shows the catalytic activity function values (fitness function values) of the top 30 catalysts in terms of catalytic activity within the metallocene catalyst population.
[0233]
[0234]
[0235]
[0236] <Calculation of each indicator value and fitness function value using density functional theory (DFT)>
[0237] The electronic structural indices and stereostructural indices of the fitness function and their corresponding respective index values were obtained from each metallocene catalyst in the population through calculations using density functional theory (DFT).
[0238] Specifically, the three-dimensional structural indices, such as the bond length and bond angle between the zirconium ion, the metallocene central metal, and the ligand, and the bond length between the zirconium ion and the halogen chloride ion, were obtained through a quantum computational program based on the density functional theory. The methodology used here is to optimize the metallocene structure by utilizing 1) TPSSTPSS or 2) PBE0 as the functional and def2-svp as the basis function. Each indices value was measured from the output of the calculation using the metallocene.py code.
[0239] The charge value of the zirconium ion, which is an electronic structural index value, was extracted using the utils.py code from the output value of quantum calculation using 1) M06-2X or 2) PBE0 functional and def2-tzvp basis function in the metallocene structure that completed structural optimization using the above method.
[0240] Each of the electronic structural index values and stereostructural index values obtained through the above DFT were substituted into the catalytic activity function (Equations 4 to 7) to obtain the catalytic activity function value for each metallocene catalyst, and the results for the metallocene catalyst having a catalytic activity function value of 4,500,000 / hour or more are shown in Table 3 above.
[0241]
[0242] <Selection of metallocene catalysts exceeding a preset value from the population>
[0243] Referring to Table 3, it was confirmed that the metallocene catalyst containing Lig_10 or Lig_30 to Lig_34 at the top and bottom had relatively high catalytic activity. On the other hand, it was confirmed that the metallocene catalyst containing Br_4 to Br_7 or Br_9 to Br_13 containing at least one aryl group as a bridge had relatively high catalytic activity.
[0244] Accordingly, the top 30 catalysts were selected (excluding duplicates) from the catalytic activity function values of all metallocene catalysts generated through the selection, crossover, and mutation processes to form a group, which is shown in Table 3.
[0245] Genetic Algorithm
[0246] According to Fig. 6, the code for generating inputs for performing a genetic algorithm quantum computing program calculation with a metallocene catalyst within the group (gen_3d_structure.py), the code for obtaining the fitness function value of the metallocene catalyst (metallocene.py, utils.py) and evaluating, performing crossover and mutation based on this (GA.py), and the code for storing each index value and fitness function value of the generated metallocene catalyst structure in a database (database.py) were performed.
[0247] Specifically, the metallocene catalysts within the group according to Fig. 6 were crossed at a crossing rate of 0.3 using the GA.py code. Then, they were mutated at a mutation rate of 0.5 using the GA.py code.
[0248] Next, offspring individuals are obtained, the catalytic activity function values are obtained again for the offspring individuals, the metallocene catalysts with values exceeding the top 50% are grouped, and then crossover and mutation are performed again to generate the next generation of offspring individuals. At this time, the number of generations was repeated from 2 to 50, and this was performed using utils.py and main.py (code that inputs the setting values for operating the genetic algorithm).
[0249] Accordingly, the top 30 catalysts in terms of catalytic activity among the metallocene catalysts with a catalytic activity function value of 4,500,000 / hour or more obtained as the final offspring are shown in Table 3.
[0250]
[0251] Example 3: Method for deriving a metallocene catalyst (polymer molecular weight function)
[0252] <Definition of the fitness function>
[0253] According to one embodiment, the polymer molecular weight function among the fitness functions may be a function defined to secure the characteristics of a metallocene catalyst capable of producing a polymer having a high comonomer (1-hexene) ratio. At this time, based on the analysis of the catalytic reaction mechanism that the termination reaction determining the molecular weight of the polymer is affected by changes in the structure and electronic structure of the catalyst, an electronic structural index (electron occupancy of the 4d orbital of the zirconium ion, which is the central metal of the metallocene) and a stereostructural index (average bond length of the zirconium ion and the upper ligand ring or the lower ligand ring, the dihedral angle between the upper ligand ring and the lower ligand ring, the distance between the zirconium ion and the β-hydrogen of ethylene) may be derived, and a function, which is [Mathematical Formula 9] derived from [Mathematical Formula 8] below, may be defined according to a multivariate linear regression-based statistical model.
[0254] Specifically, the activation energy (ΔG) for several possible termination reaction pathways can be calculated using density functional-based quantum calculations. RXN) was confirmed to be an important factor in determining the termination reaction by comparing the structure of the metallocene complex in which the monomer is coordinated to the zirconium ion. Therefore, the change in the three-dimensional structure of the catalyst in the termination reaction was analyzed to establish electronic structural indices and three-dimensional structural indices that affect the termination reaction. The following mathematical equations 8 and 9, which can derive the molecular weight of the produced polymer through multivariate linear regression analysis, a statistical analysis, were defined. The difference in activation energy (ΔG) was determined by the above method. RXN ) can be used to predict the molecular weight of a polymer.
[0255] [Equation 8]
[0256]
[0257] In the above mathematical expression 8, MCp is the average bond length between the zirconium ion, which is the metallocene central metal ion, and the upper ligand ring or the lower ligand ring.
[0258] [Equation 9]
[0259]
[0260] In the above mathematical formula 9, CpCp is the dihedral angle between the upper ligand ring and the lower ligand ring, and n 4d is the electron occupancy of the d orbital of the metallocene central metal ion, and d Eth-H is the distance between the hydrogens forming the β-agostic bond of the olefin monomer ethylene and the zirconium ion.
[0261]
[0262] <Population formation of metallocene catalysts>
[0263] Figure 3 is a diagram showing the ligand types corresponding to the top or bottom of the metallocene catalyst in the population. Figure 4 is a diagram showing the bridge types corresponding to the middle of the metallocene catalyst in the population.
[0264] Referring to Figures 3 and 4, the upper and middle ligand rings indicated as Lig_0 to Lig_34 were applied, respectively, and the middle used a bridge indicated as Br_0 to Br_13. At this time, Cp in the bridge can be a substituted or unsubstituted cyclopentadienyl group, indenyl group, or fluoroenyl group. At this time, in the substituted or unsubstituted, the substituents are independently a halogen group, a cyano group, a nitro group, a C1-C 20 Alkyl group (C1-C 10 Silane group substituted with alkyl group can be substituted), C1-C 20 Aryl group (C1-C 10 Silane group substituted with alkyl group can be substituted), C1-C 20 Allyl group (C1-C 10 It may be at least one selected from the group consisting of a silane group substituted with an alkyl group), and a naphthalenyl group.
[0265] At this time, the code used to form the population was a code (gen_3d_structure.py) that combines three partial structures of the upper, middle, and lower metallocene catalysts to form one complete metallocene catalyst structure.
[0266] Table 4 shows metallocene catalysts having a polymer molecular weight function value (fitness function value) of 350,000 or more within the metallocene catalyst population.
[0267]
[0268] Catalyst molecular weightCatalyst molecular weight15_9_30405,19617_10_31365,61217_6_31402,40517_5_31361,22318_10_30397,40516_13_30359,83515_14_31392,93817_13_30359,49015_7_30385,74316_1_31357,61918_5_31383,66218_6_31350,51315_10_31375,638
[0269]
[0270] <Calculation of each indicator value and fitness function value using density functional theory (DFT)>
[0271] The electronic structural indices and stereostructural indices of the fitness function and their corresponding respective index values were obtained from each metallocene catalyst in the population through calculations using density functional theory (DFT).
[0272] Specifically, the three-dimensional structural indices, such as the bond length and bond angle between the zirconium ion, the metallocene central metal, and the ligand, and the bond length between the zirconium ion and the halogen chloride ion, were obtained through a quantum computational program based on the density functional theory. The methodology used here is to optimize the metallocene structure by utilizing 1) TPSSTPSS or 2) PBE0 as the functional and def2-svp as the basis function. Each indices value was measured from the output of the calculation using the metallocene.py code.
[0273] The charge value of the zirconium ion, which is an electronic structural index value, was extracted using the utils.py code from the output value of quantum calculation using 1) M06-2X or 2) PBE0 functional and def2-tzvp basis function in the metallocene structure that completed structural optimization using the above method.
[0274] Each of the electronic structural index values and stereostructural index values obtained through the above DFT were substituted into the polymer molecular weight function (Mathematical Equations 8 to 9) to obtain the polymer molecular weight function value for each metallocene catalyst, and the results are shown in Table 4 above.
[0275]
[0276] <Selection of metallocene catalysts exceeding a preset value from the population>
[0277] Referring to Table 4, it was confirmed that the metallocene catalyst containing both Lig_30 and Lig_31 at the top and bottom had a relatively high polymer molecular weight. On the other hand, it was confirmed that the metallocene catalyst containing both Br_5 and Br_7 or Br_9 and Br_14 containing at least one aryl group as a bridge had a relatively high polymer molecular weight.
[0278] Accordingly, 13 catalysts with a molecular weight function value of 350,000 or more among all metallocene catalysts produced through the selection, crossover, and mutation processes were selected (excluding duplicates) to form a group, which is shown in Table 4.
[0279] Genetic Algorithm
[0280] According to Fig. 6, the genetic algorithm was performed using a code (gen_3d_structure.py) that generates input values for quantum computing program calculations with a metallocene catalyst within the group, a code (GA.py) that obtains the fitness function values of the metallocene catalyst (metallocene.py, utils.py) and evaluates, crosses over, and mutates based on these values, and a code (database.py) that stores each index value and fitness function value of the generated metallocene catalyst structure in a database.
[0281] Specifically, the metallocene catalysts within the group according to Fig. 6 were crossed at a crossing rate of 0.3 using the GA.py code. Then, they were mutated at a mutation rate of 0.5 using the GA.py code.
[0282] Next, offspring individuals are obtained, and the polymer molecular weight function values are obtained again for the offspring individuals, and the metallocene catalysts with values exceeding the top 50% are grouped, and then crossover and mutation are performed again to generate the next generation of offspring individuals. At this time, the number of generations was repeated from 2 to 50, and this was performed using utils.py and main.py (code that inputs the setting values for operating the genetic algorithm).
[0283] Accordingly, the metallocene catalysts obtained as the final offspring individuals having a molecular weight function value of 350,000 or more are shown in Table 4.
[0284]
[0285] That is, the method for deriving a metallocene catalyst according to the present invention not only effectively predicts metallocene catalyst performance such as catalytic activity, copolymerization ratio, and polymer molecular weight through a fitness function defined by density functional theory (DFT), electronic structural indices, and stereostructural indices, but also effectively derives a novel catalyst structure optimized for each performance with a simplified amount of computation through a genetic algorithm, so it has the advantage of improving economic feasibility and efficiency when designing a metallocene catalyst.
[0286]
[0287] Simple modifications or changes of the present invention can be easily implemented by a person having ordinary skill in the art, and all such modifications or changes can be considered to be included in the scope of the present invention.
Claims
1. (a) A step of defining a fitness function related to catalyst properties using electronic structural indices and stereostructural indices; (b) forming a population of metallocene catalysts divided into upper, middle, and lower groups according to the catalyst structure; (c) a step of calculating the electronic structural index values and stereostructural index values of the metallocene catalyst corresponding to the electronic structural index and stereostructural index of the fitness function from density functional theory (DFT); (d) a step of obtaining a fitness function value from the electronic structural index values and stereostructural index values; (e) a step of forming a group by selecting a metallocene catalyst having a value higher than a preset value from the above fitness function value; (f) performing crossover and mutation on the metallocene catalysts within the group to generate offspring individuals; and (g) A method for deriving a metallocene catalyst, characterized in that it comprises a step of repeating steps (c) to (f) multiple times on the generated offspring individuals, and terminating the step.
2. In paragraph 1, The step of defining the fitness function related to the above (a) catalyst properties as electronic structural indicators and stereostructural indicators A step of analyzing the catalytic reaction mechanism from any metallocene catalyst; A step of deriving electronic structural indicators and stereostructural indicators based on the activation energy of each reaction step, the free energy difference of reaction intermediates, and the structural change of the catalyst from the catalytic reaction mechanism analyzed above; and A method for deriving a metallocene catalyst, comprising: defining a fitness function according to a multivariate linear regression-based statistical model that predicts experimental result values from the electronic structural indicators and stereostructural indicators derived above; 3. In paragraph 1, A method for deriving a metallocene catalyst, wherein the above electronic structural indicator is at least one indicator selected from the group consisting of the charge of the metallocene central metal ion, orbital energy, and electron occupancy of a specific orbital.
4. In paragraph 1, A method for deriving a metallocene catalyst, wherein the above three-dimensional structural indicator is at least one indicator selected from the group consisting of a dihedral angle between an upper ligand ring and a lower ligand ring, an average bond length between a metallocene central metal ion and an upper ligand ring or a lower ligand ring, and a bond angle of an upper ligand ring-metallocene central metal-lower ligand ring.
5. In paragraph 1, A method for deriving a metallocene catalyst, wherein the above fitness function is one function selected from the group consisting of a catalytic activity function, a copolymerization ratio function, a polymer molecular weight function, a polymer hydrogen reactivity function, a polymer molecular weight distribution function, a polymer density function, a polymer transparency function, a polymer MI function, a polymer MFR function, a polymer Tg function, a polymer melting point function, and a polymer mechanical property function.
6. In paragraph 5, The above copolymerization ratio function is a function defined by [Mathematical Formula 3] derived from [Mathematical Formula 1] and [Mathematical Formula 2] below, and a method for deriving a metallocene catalyst. [Mathematical Formula 1] (In the above mathematical formula 1, M charge is the charge of the metallocene central metal ion, CpMCp is the bond angle of the upper ligand ring-metallocene central metal-lower ligand ring, CpCp is the dihedral angle between the upper ligand ring and the lower ligand ring, MCp is the average bond length between the metallocene central metal ion and the upper ligand ring or the lower ligand ring, and MX is the average bond length between the metallocene central metal ion and the halogen ion. [Equation 2] [Equation 3] 7. In paragraph 1, In the step of forming a group by selecting a metallocene catalyst having a value higher than a preset value from the fitness function value (e) above, A method for deriving a metallocene catalyst by selecting the top 50% values as the preset value of the copolymerization ratio function among the fitness functions.
8. In paragraph 5, A method for deriving a metallocene catalyst, wherein the above catalytic activity function is a conversion frequency function defined by [Mathematical Formula 7] derived from [Mathematical Formula 4] to [Mathematical Formula 6] below. [Equation 4] (In the above mathematical formula 4, n 4d is the electron occupancy of the d orbital of the metallocene central metal ion) [Equation 5] ΔG Monomer = -0.1653 θ + 32.903 (In the above mathematical expression 5, θ is the open angle between the upper and lower ligand rings measured in the pre-catalyst structure) [Equation 6] ΔG RXN = ΔG D + ΔG Monomer [Equation 7] Conversion frequency (10 6 h -1 ) = -0.6207 ΔG RXN + 20.3 9. In paragraph 1, In the step of forming a group by selecting a metallocene catalyst having a value higher than a preset value from the fitness function value (e) above, A method for deriving a metallocene catalyst by selecting the top 50% values as the preset value of the catalytic activity function among the fitness functions.
10. In paragraph 5, A method for deriving a metallocene catalyst, wherein the above polymer molecular weight function is a function defined by [Mathematical Formula 9] derived from [Mathematical Formula 8] below. [Equation 8] M wrZrCp = -27141000 ×MCp + 60339000 (MCp < 2.256) 4137300 ×MCp - 9131700 (MCp > 2.256) (In the above mathematical expression 8, MCp is the average bond length between the metallocene central metal ion and the upper or lower ligand ring. [Equation 9] Polymer molecular weight = 0.8075×M wrZrCp - 9419.03×CpCp+485525.7×d Eth-H +n 4d +5455121 (In the above mathematical formula 9, CpCp is the dihedral angle between the upper and lower ligand rings, and n 4d is the electron occupancy of the d orbital of the metallocene central metal ion, and d Eth-H is the distance between the hydrogen bonding between the olefin monomer and the β-agostic bond of the metallocene central metal ion.
11. In paragraph 1, In the step of forming a group by selecting a metallocene catalyst having a value higher than a preset value from the fitness function value (e) above, A method for deriving a metallocene catalyst by selecting the top 50% values as the preset value of the polymer molecular weight function among the fitness functions.
12. In paragraph 1, In the step of generating offspring individuals by performing crossover and mutation on the metallocene catalysts within the above (f) group, (f-1) a step of crossing the metallocene catalysts within the above group according to the crossing rate; and (f-2) A method for deriving a metallocene catalyst, comprising: a step of generating offspring individuals by mutating the metallocene catalyst within the above-mentioned crossed group according to a mutation rate.
13. In paragraph 12, A method for deriving a metallocene catalyst, wherein the above crossover ratio is 0.3 to 0.
5.
14. In paragraph 12, A method for deriving a metallocene catalyst, wherein the mutation rate is 0.3 to 0.
5.
15. For the target metallocene catalyst, electronic structural index values and stereostructural index values are calculated from the metallocene catalyst using density functional theory (DFT); Obtaining a fitness function value from the above electronic structural index values and stereostructural index values; and Predicting the performance of a metallocene catalyst through the above fitness function; A method for predicting the performance of a metallocene catalyst including steps.
16. A processor that defines a fitness function related to catalyst properties as electronic structural indices and stereostructural indices, calculates electronic structural indices and stereostructural indices using density functional theory (DFT), and obtains a fitness function value from the electronic structural indices and stereostructural indices; and Memory required for the above processor to perform operations; A metallocene catalyst modeling device comprising:
17. In paragraph 16, A metallocene catalyst modeling device, wherein the processor selects metallocenes based on the fitness function value and performs a genetic algorithm on the selected metallocene catalysts to derive a new catalyst structure optimized for each performance.
18. In paragraph 17, A metallocene catalyst modeling device, wherein the genetic algorithm performs crossover and mutation on selected metallocene catalysts to generate offspring individuals, obtains a fitness function value for the generated offspring individuals, and selects them again to generate offspring individuals of the next generation, repeating this process multiple times.
19. In paragraph 16, The above processor is a metallocene catalyst modeling device that predicts the metallocene catalyst performance by the above fitness function value.
20. In paragraph 16, The above electronic structural indicator is at least one indicator selected from the group consisting of the charge of the metallocene central metal ion, orbital energy, and electron occupancy of a specific orbital, The above three-dimensional structural indicator is at least one indicator selected from the group consisting of the dihedral angle between the upper ligand ring and the lower ligand ring, the average bond length between the metallocene central metal ion and the upper ligand ring or the lower ligand ring, and the bond angle of the upper ligand ring-metallocene central metal-lower ligand ring. A metallocene catalyst modeling device, wherein the above fitness function is one function selected from the group consisting of a catalytic activity function, a copolymerization ratio function, a polymer molecular weight function, a polymer hydrogen reactivity function, a polymer molecular weight distribution function, a polymer density function, a polymer transparency function, a polymer MI function, a polymer MFR function, a polymer Tg function, a polymer melting point function, and a polymer mechanical property function.
Citation Information
Patent Citations
Stair for working
KR1020210128745A
Method for predicting property of polymers
KR102095523B1
Method for predicting properties of polyethylene resin, and preparation method of polyethylene resin using the same
KR102524480B1
Method of selecting polyolefins based on rheological properties
WO2006107374A1
