Diverse Optimal Molecular Libraries for Organic ... - ACS Publications

Subscriber access provided by GAZI UNIV

Article

Diverse Optimal Molecular Libraries for Organic Light Emitting Diodes Chetan Raj Rupakheti, Rachael Al-Saadon, Yuqi Zhang, Aaron M Virshup, Peng Zhang, Weitao Yang, and David N. Beratan J. Chem. Theory Comput., Just Accepted Manuscript • DOI: 10.1021/acs.jctc.5b00829 • Publication Date (Web): 07 Mar 2016 Downloaded from http://pubs.acs.org on March 8, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Chemical Theory and Computation is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Chemical Theory and Computation

Diverse Optimal Molecular Libraries for Organic Light Emitting Diodes Chetan Rupakheti1, Rachael Al-Saadon2, Yuqi Zhang2, Aaron M. Virshup2,+, Peng Zhang2, Weitao Yang2,3*, David N. Beratan2,3,4* 1

Program in Computational Biology and Bioinformatics, Duke University, Durham NC 27708 2 Department of Chemistry, Duke University, Durham NC 27708 3 4

+

Department of Physics, Duke University, Durham NC 27708

Department of Biochemistry, Duke University, Durham NC 27708

current address: Bio/Nano Research Group, Autodesk Inc. Pier 9, San Francisco 94111

* corresponding authors

Abstract Organic light emitting diodes (OLEDs) have wide-ranging applications from lighting to device displays. However, the repertoire of organic molecules with efficient blue emission is limited. To address this limitation, we have developed a strategy to design property-optimized, diversity-oriented libraries of structures with favorable fluorescence properties. This approach identifies novel diverse candidate organic molecules for blue emission with strong oscillator strengths and low singlet-triplet energy gaps that favor thermally activated delayed fluorescence (TADF) emission.

1

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

▪ Introduction Organic light emitting diodes (OLEDs) have applications ranging from lighting to electronic displays.1,2 Advantages of OLEDs over traditional technologies include their energy efficiency and compactness.1,2 OLEDs convert electrical charge into light by the radiative recombination of electron-hole pairs.3 The performance of blue OLEDs is lower than that for emitters of other colors.4 Unlike fluorescence based materials that emit from singlet exciton states with efficiency limited to ~ 25% (excitons exist in 25% singlet and 75% triplet states, assuming electrons and holes are injected with random spin orientation)5, phosphorescent materials can funnel both singlet and triplet excitons into emissive states, thus endowing OLEDs with higher efficiencies.6 Known phosphorescent OLED materials are based on iridium and platinum complexes.6 The aim of this study is to identify prospects for high efficiency organic blue-emitting molecules. Our approach is based on designing property-optimized diverse molecular libraries7 with favorable fluorescence properties. OLED Structures OLEDs can be single or multi-layered,1,2 and property optimization can focus on the individual layers or the interfaces. A typical OLED structure is shown in Figure 12 :

2


Page 2 of 30

Page 3 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


eh+ h+eS1

25%

75%

T1

Intersystem Crossing

Anode

Fluorescence

Phosphorescence

Cathode

Figure 1. Light emission mechanism from an OLED device. Holes (h+) are supplied by the anode and electrons (e-) by the cathode.

The electron and hole pairs in the emissive layer of an OLED form excitons that exist either in singlet (S1) or triplet states (T1). Since the injected charges are not spin correlated, 75% of the excited states produced are triplets. The excitons can decay radiatively, via fluorescence or phosphorescence. Intersystem crossing (from T1 to S1) can be facilitated by heavy metals (such as iridium) and is enhanced if the singlet and triplet states are close in energy.2 First generation blue emitting OLEDs harvest luminescence from singlet-excited states. Second generation OLED materials (PHOLEDs) harvest luminescence from triplet excited states of transition metal complexes that mix singlet and triplet states via spinorbit coupling. The exciton spin state may be, αα, ββ, αβ+βα, or αβ-βα, where α and β represent electron spins. The first three states are triplets and the last is a singlet. In

3



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

OLEDs based purely on fluorescence, the radiative decay of triplet excitons (75%) is spin forbidden, and only the singlet excitons (25%) produce appreciable luminescence.8,9 Organometallic LEDs that emit from triplet states are coupled by spin-orbit interactions to higher lying singlets and to the ground state, enhancing the emission. However, the requisite metals (such as iridium) are costly.8,9 Third generation OLEDs are based on fluorophores that produce thermally activated delayed fluorescence (TADF). TADF emission can be achieved when the energy gap between triplet and singlet states is sufficiently low that excitons in triplet states can transition into singlet states (intersystem crossing, Figure 1) by thermal activation. TADF thus circumvents the emission quantum yield limitations defined by spin statistics. Two properties are required to realize efficient TADF: a small energy gap between the lowest energy singlet and triplet excited states and a large transition dipole moment for fluorescence.8,9 Recent experiments showed impressive results for TADF with internal quantum yields (ratio of radiative electron-hole recombination to the total radiative and nonradiative recombination)10 and external quantum yields (ratio of the number of photons emitted from the LED to the number of electrons passing through the device)2 for singlets and triplets up to 100% and 25%, respectively.6 There are few purely organic compounds with high efficiency TADF, and molecular design is based on the formation of a charge transfer intermediate state (with a donor-acceptor-donor motif).8,9 Developing strategies to expand the TADF mechanism to organic structures beyond the known donor-acceptor-donor framework is appealing and could be facilitated by comprehensive chemical space explorations. Here, we extend the strategy for diverse optimal molecular library design based on the ACSESS strategy7 to

4


Page 4 of 30

Page 5 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


design libraries of 1) blue emitting fluorescent structures that emit dominantly from singlet states and 2) higher efficiency blue emitters utilizing TADF emission. ▪ Methods Using Property-Optimizing ACSESS To Design Fluorescent Emitters. We briefly summarize the approach for diverse optimal molecular library construction using the algorithm for chemical space exploration with stochastic search (ACSESS) approach. The details of the ACSESS algorithm appear in Refs [7] and [11]. The previous implementations of ACSESS aimed to maximize library diversity of chemically stable organic molecules11 or to maximize the diversity of molecules while optimizing a ground state molecular property.7 Here, we extend the ACSESS approach to discover diverse chromophores with targeted fluorescent properties. As in earlier ACSESS calculations, we use a 40 dimensional autocorrelation descriptor to define a molecule’s coordinates in chemical space.11 This descriptor set encodes the structural and physio-chemical properties of a molecule in a vector of fixed length (in this case 40) and describes topological correlations between atomic properties for a molecule. 11 The examples of properties used here are atomic number, GasteigerMarsili partial charge, atomic polarizability, and topological steric index.11 This descriptor set is used to compute distances among sampled molecules and to sample a diverse set of molecules within ACSESS. We coupled the ACSESS code with the Gaussian0912, ZINDO/S13, DFTB+12 programs to compute molecular excitation energies and oscillator strengths. As part of the property assessment, the objective function of ACSESS was extended to include: 1) the lowest singlet excitation energy, 2) the lowest triplet excitation energy, 3) the gap between the singlet (S1) and the triplet (T1) excitation

5



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

energies, and 4) the transition’s oscillator strength. The molecules accepted in the ACSESS library are required to satisfy multiple criteria, in contrast to our earlier single property biased library design. As extended here, ACSESS samples from chemical space by iteratively maximizing the nearest-neighbor distance (diversity) of a subset of compounds in the space of all possible fluorescent emitters. There are four main steps in the extended property-optimizing ACSESS calculation: (1) initialize a library with known blue emitters, (2) breed new compounds, (3) remove compounds that are not stable or do not have electronic properties that satisfy the multiple-property thresholds, and (4) select a maximally-diverse subset of propertyqualified structures. The new molecules within ACSESS are predicted using a set of mutations moves that modifies the atoms, bonds, and rings in the molecules. The crossover moves generate compounds by combining relevant fragments of the parent compounds (segmented at a randomly chosen position). To ensure we generate stable molecules, various filters are applied to the molecules that neglect molecules that are unfavorable.11 To optimize the electronic property values iteratively, we tighten the property value cutoffs linearly with each iteration step, until a desired threshold is reached for each property. Initially, the threshold is set to a less stringent value (based on the electronic property of interest) to ensure that the population does not collapse to zero because of the fitness constraint. The flowchart of the multi-objective property-biased ACSESS scheme used here is shown in the Figure 2.

6


Page 6 of 30

Page 7 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Initialize Library with Known Blue Emitter

Modify Molecular Structures Properties to Optimize

Filter based on Chemical Stability and Electronic Properties

1) Validate excitation energy within blue 2) Validate oscillator strength value

Select a Maximally Diverse Set

No

3) If required, Validate S1T1 gap

Final Iteration? Yes Done

Figure 2. ACSESS extension for fluorescent-molecule design based on emission wavelength, oscillator strength, and singlet-triplet excitation energy gap. Beginning with a few known organic blue emitters,13,14 the goal of the molecular design was to generate a library of diverse structures with singlet excitation energies in the 2.5-4.0 eV range, with an optimal singlet (S1)-triplet (T1) energy gap (∆ES1-T1) at most ~0.3eV (we chose ~0.3eV following previous published design structures9,15 and considering the errors involved in the semiempirical calculations).

7



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Quantum Chemical Computation of Singlet Excitation Energies and Oscillator Strengths. During the ACSESS library generation, the molecular property values are computed on the fly. We use semiempirical quantum chemistry methods to reduce the computational costs. The ground state geometries are optimized using the semiempirical AM1 approach.16 Vertical electronic excitation energies and oscillator strength are computed for the lowest singlet transitions using ZINDO/S calculations.17 The reliability of a semiempirical approach was assessed by comparing the excitation energies for a set of published organic chromophores18,19 with the results of time-dependent density functional theory (TDDFT) calculations.20,21 We randomly selected a set of polycyclic, aromatic hydrocarbons to compare the TDDFT and semiempirical excitation energies.18,19 TDDFT and semiempirical calculations were performed using the Gaussian09 package.22 The geometry optimizations were performed using the CAM-B3LYP23 functional with the 6-31G* basis set.24 Frequency analysis was carried out for all molecules in the validation set to ensure that the optimized structures were local energy minima. The optimized geometries were used in TDDFT excited states calculations with the same functional and basis sets that were used for geometry optimization. Design of Fluorescent Emitters that Utilizes only Singlet Excitation State. Within this extended property-biased ACSESS framework, we used chemical stability and synthetic feasibility criteria, applied previously to generate GDB libraries,25,26 and we computed the electronic excitation energies and oscillator strengths of molecules at each iteration of the ACSESS algorithm. We focused on generating molecules with singlet excitation energies in the 2.5-4.0 eV range (with ≥ 0.2 oscillator

8


Page 8 of 30

Page 9 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


strength). Semiempirical AM1 and ZINDO/S methods were used to compute the excitation energies and oscillator strengths of the candidate molecules. The molecular structures were generated using OpenEye27 and OMEGA libraries,28 and optimized at the AM1 level. As a final step, to assess the reliability of the semiempirical calculations, the property values for the optimized ACSESS library were further analyzed using DFT and TDDFT computations: the molecular structure was re-optimized at the CAM-B3LYP/631G* level of theory and the lowest vertical singlet excitation energies and oscillator strengths were computed at the TD-CAM-B3LYP/6-31G* level based on the DFT optimized geometry. Designing the TADF Library by Optimizing the Singlet - Triplet Gap. The triplet excitons for organic molecules are not useful, per se, in OLED materials because of their long phosphorescence lifetimes. The TADF mechanism funnels T1 states to nearby S1 states, and thus enhances emission intensity.8 In the condensed phase at room temperature, the initially excited molecules typically relax to S1 or T1 excited states on sub-picosecond time scales.29 If S1 is near the potential minimum of T1, the small S1-T1 energy gap will enhance the intersystem crossing from T1 to S1.8 Thus, a key feature of TADF structures is a small S1-T1 energy gap near the T1 equilibrium geometry. The intersystem crossing facilitated at room temperature by a small S1-T1 gap can greatly increase the efficiency of OLED materials. Unlike previously reported OLED computational designs,15 our PO-ACSESS based optimization criteria computes the S1-T1 gap at the triplet-optimized geometry where the intersystem crossing process is most likely to occur. Also, previous TADF designs focused on molecules with S1 and T1 excited states that have charge-transfer character favoring

9



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

degeneracy of the S1 and T1 states.6,8,9,15 Here, we do not constrain our search to species with charge transfer excited state character, and we explore the chemical space in a property-biased manner to favor near degenerate singlet and triplet excited states.8 We used the library optimized by ACSESS in the previous step (fluorescent emitters with 25% quantum yield) and minimized the singlet and triplet excitation energy gap, requiring the energy gap (∆ES1-T1) within a few kBT (kBT = 0.026eV). For the design, we also used chemical stability and synthetic feasibility filters that were used previously to generate the small molecule universe (SMU) of compounds.11 ACSESS was extended to compute the energies of the generated molecules at their minima on the lowest triplet state potential energy surface. In summary, the multi-objective design criteria were: 1) vertical singlet excitation energy in the 2.4 - 4.1 eV range (blue region); 2) oscillator strength ≥ 0.4; 3) ∆ES1-T1 ≤ 0.3 eV at the triplet excited state geometry minimum. Our benchmark calculations indicate that the singlet-triplet gap derived from ZINDO/S methods is offset from those found in the TDDFT analysis (see results section). Since the accuracy of computed S1-T1 gap to identify potential TADF molecules is more sensitive to the error compared to the previous problem of identifying singlet emitters, we instead used DFTB and linear-response DFTB methods12 since the methods are known to compute the geometries and excitation energies of organic molecules reliably. After obtaining the library, we validated the computed energies using TDDFT and RICC2 methods.30 We chose a reasonably low ∆ES1-T1 value of 0.3 eV and achieved this objective by discarding compounds that do not match our criteria 1-3 at each iteration. To ensure that the number of promising candidates in each iteration pool does not collapse to

10


Page 10 of 30

Page 11 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


zero, we gradually tightened our convergence criteria. First, we constrained the singlet excited state energies of these compounds within the 2.4 - 4.1 eV range. Then, we linearly increased the oscillator strength cutoff and, finally, we linearly decreased the singlet-triplet energy gap requirement until we reached the targeted values of all three criteria. More specifically, we used the following steps to optimize the library of candidate TADF structures: ▪

Ground singlet state (S0) optimization using the DFTB Hamiltonian

▪

Linear response DFTB calculation of the lowest singlet excitation energy and oscillator strength based on the S0 optimized geometry

▪

Lowest triplet state (T1) geometry optimization using the DFTB Hamiltonian

▪

Linear response DFTB calculation of the lowest singlet and triplet excitation energies using the optimized T1 geometry

▪

Results and Discussion

Validating the Semiempirical Excitation Energies. A set of 44 published organic chromophores18,19 was used to compare computed singlet and triplet excitation energies from ZINDO/S and TDDFT methods. Figure 4 shows the comparison of the computed vertical electronic excitation energies of the lowest singlet excited state based on the ZINDO/S and TDDFT methods. The average difference between ZINDO/S and TDDFT computed vertical electronic excitation energies is ~0.7 eV (Figure 3A). The average difference between the two methods for computing the lowest triplet excited state energy is ~0.9 eV (Figure 3B). This study suggests that using semiempirical calculations to design the singlet emitters is feasible,

11



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

since our criterion for blue emission ranges from 2.4 to 4.1 eV. However, for the TADF designs involving small singlet-triplet excitation energy gaps (less than 0.3 eV), the ZINDO/S calculations are not sufficiently accurate. We introduced more reliable electronic structure methods based on linear response DFTB12 methods within ACSESS to improve the accuracies and we validated the calculations using the RICC2 method. A

B

12


Page 12 of 30

Page 13 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3. (A) Comparison of vertical electronic excitation energies for the lowest singlet transitions of the validation set (average error ~0.7 eV). (B) Comparison of vertical electronic excitation energies for the lowest triplet transition of the validation set (average error ~0.9 eV). Property-biased ACSESS-based Library of Singlet Emitters. We used property-optimizing ACSESS calculations (Figure 2) to construct a library of diverse structures with property values targeted for blue emission and substantial oscillator strengths with the chemical stability and synthetic feasibility described above25 to generate a diverse library of 195 fluorescent structures that fit the property constraints. The structures of 4 representative molecules of the generated library are shown in Figure 4.

13



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 4. ACSESS generated potential blue-emitting organic structures. Shown are some representative structures of the 195 compound library with average singlet excitation energy and oscillator strength (3.88 eV and 0.8, respectively). The blue ovals indicated the chromophore units of the molecules. The predicted singlet excitation energy of the library, on average, is 3.88 eV (320 nm) with an average oscillator strength of 0.8. To compare our generated structures with known structures, we collected 48 published organic blue emitters.13,14,31–35 All of the property-optimizing ACSESS generated molecules are structurally unique compared to these published blue emitters. None of the molecules published or patented were found to have the same chromophores (blue ovals, Figure 4) as in our generated structures, base on a SciFinder search. Also, the known and registered molecules within SciFinder have only ~65% structural match with the molecules in our generated library. As such, the molecules in our library most likely have not been investigated before for their fluorescence properties. The lowest singlet excited states for structures containing this unique chromophore are dominated by π π*transitions of the noted motifs. The HOMO and LUMO for one of the representative structures are shown in Figures 5A and 5B, respectively. To visualize the difference between the ACSESS generated compounds and the published structures, we projected the published blue emitters onto the selforganizing map trained on the 195 property biased ACSESS generated compounds (Figure 6B). We noticed that ACSESS sampled compounds cover chemical space broadly compared to the known compounds (Figures 6A and 6B). The known compounds are localized to small regions and occupy only a small portion of the ACSESS sampled chemical space. The white areas in Figure 6B indicate regions where no published

14


Page 14 of 30

Page 15 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


compounds were mapped to the SOM trained on ACSESS sampled compounds. Comparing Figures 6A and 6B, we find that the property-biased ACSESS generated compounds sample regions of chemical space not populated by published compounds. A

B

15



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 5. Figures show the HOMO and LUMO of a representative structure sampled using ACSESS. (A) Shows HOMO and (B) shows LUMO of the representative molecule. A

Number of Compounds Per Grid

B

Figure 6. The x- and y-axes represent position in a SOM grid. (A) The ACSESS generated blue emitting compounds are all shown in a self-organizing map (SOM) where

16


Page 16 of 30

Page 17 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


each blue dot represents a compound. (B) Projection of known compounds in the SOM trained on ACSESS generated SOM. Validating the ZINDO/S Excitation Energy Calculations with TDDFT and Linear Response RICC2/Def2-SVP for the ACSESS-designed Library. To test the validity of the computed transition energies of the designed structures, we compared the results of the ZINDO/S computations with those of TDDFT analysis for 45 maximally diverse structures (a representative subset of the pool of ACSESS generated molecules)36 among the 195 generated structures. The structures were optimized; their ground state energies and excitation energies, and their transition dipole moment were computed using the CAM-B3LYP functional with a 6-31G* basis set. TDDFT computed excitation energies agree with the ZINDO/S computed energies for 43 of the 45 molecules to lie in the blue emission region (2.4-4.1 eV). Figure 7 compares ZINDO and TDDFT computed singlet excitation energies. Since the average difference for the computed singlet excitation energy between TDDFT and ZINDO/S is 0.13 eV, ZINDO/S reproduces the DFT excitation energies to the level of accuracy required to remain in the blue region of the spectrum. The oscillator strength computed using the semiempirical approach has a mean error of 0.44 compared to the TDDFT computed values. The diverse library has an average TDDFT computed oscillator strength of ~0.4. We further validated the ZINDO/S and TDDFT computed excitation energies and oscillator strength of some molecules using the linear-response RICC2 method with the Def2-SVP basis set implemented in the Turbomol package.37 For further testing, we randomly choose five molecules to reduce the computational cost, and we found that the average singlet excitation energy was 3.6 eV, with a standard deviation of 0.027 eV. The

17



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

average oscillator strength of the molecules was 0.31, with a standard deviation of 0.143. From this analysis, we learned that the ZINDO/S and TDDFT computed singlet excitation energies are sufficiently reliable for designing compounds with blue fluorescence.

Figure 7. Comparison of vertical electronic excitation energies for the lowest singlet transitions of the maximally diverse set (average error of ~0.13 eV between ZINDO/S and TDDFT). The excitation energies are in eV units. Property-optimized ACSESS Library of TADF Emitters. We used the extended property-biased ACSESS method to design a library of novel TADF emitters (Figure 3) where we initialized the library with property-optimizing ACSESS sampled blue fluorescent emitters (Figure 4). The initial library of 195 compounds had an average ∆ES1-T1 of ~1.8 eV. We ran ACSESS for 50 iterations where ~90 molecules were generated per iteration that satisfied the required ∆ES1-T1 constraint.

18


Page 18 of 30

Page 19 of 30

Each generated molecule was subjected to geometry optimization and singlet and triplet excitation energy calculation. We reached our goal of optimizing the average excitation energy gap of the molecules to below 0.3 eV in 35 iterations (Figure 8). At the same time, we optimized the average oscillator strength to 0.45 (Figure 9) and the vertical singlet energy in the blue region (2.4 - 4.1 eV) of the spectrum (Figure 10). The full run was completed in 7 days on a 16 core CPU. The majority of the computational time (~80%) was spent on the semiempirical electronic structure computations. 2 Average Min Max

1.8 1.6 1.4

S1-T1 Gap (eV)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1.2 1 0.8 0.6 0.4 0.2 0 0

5

10

15

20

25

30

Generations

19


35

40

45

50


Figure 8. Singlet-triplet energy gap optimization derived from the property-biased ACSESS calculation. We reach the goal of shrinking ∆ES1-T1 to ~0.3 eV within 35 iterations. 0.7

0.6

0.5

Oscillator Strength

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 30

0.4

0.3 Average Min Max

0.2

0.1

0 0

5

10

15

20

25

30

35

40

45

50

Generations Figure 9. Oscillator strength optimization using the property-optimizing ACSESS method. We increased the computed oscillator strength to ~ 0.45 in the 35 iterations.

20


Page 21 of 30

4

Average Min Max

3.5

Singlet Energy (eV)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


3

2.5 0

5

10

15

20

25

30

35

40

45

50

Generations Figure 10. Singlet excitation energy at each iteration of the property-optimizing ACSESS calculation. While optimizing the S1-T1 energy gap and the oscillator strength, all of our compounds’ excitation energies remain in the blue region (2.5-3.9 eV is the range for the sampled molecules). Figure 8 indicates that we can minimize the singlet-triplet energy gap iteratively to 0.3 eV on average. The advantage of designing near-degenerate singlet and triplet excited-state energies is that it favors intersystem crossing and improves the overall OLED quantum yield. Figures 9 and 10 indicate that, while optimizing the gap, we can

21



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

also discover compounds that emit in the blue region (2.5-3.9 eV) with strong oscillator strength. The optimization of the key properties involved in the TADF design may expand the diversity and number of synthetic targets. Some of the molecules with favorable computed semiempirical properties are shown in Figure 11.

Figure 11. Some representative molecules generated with the property-optimizing ACSESS method with semiempirical computations of the average S1-T1 gap of ~0.2 eV, average excitation energy of ~3.3 eV, and average oscillator strength of ~0.6. Validation of Semiempirical Computed Gap using TDDFT and Linear Response RICC2/Def2-SVP Method. From the PO-ACSESS designed library, we chose 15 molecules to validate the S1-T1 gaps using TDDFT and RICC2/Def2-SVP methods. The DFTB computed triplet

22


Page 22 of 30

Page 23 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


molecular geometries were subjected to TDDFT calculation using RCAM-B3LYP functional and 6-31(d) basis set. To further validate the DFTB and TDDFT computed S1T1 gaps, we used the optimized geometries using DFTB methods and re-computed the S1T1 gaps using RICC2/Def2-SVP37 and spin-flip TDDFT analysis (with the CAM-B3LYP functional and Def2-SVP basis set).38 Table 1 shows the comparison of the TDDFT and RICC2 computed energy gaps of the molecules found using PO-ACSESS to be within 0.3 eV. The spin-flip TDDFT computed results are shown in supplementary Table 1. Table 1: Validation of S1-T1 gap using TDDFT and linear response RICC2 methods for molecules with energy gaps from semiempirical methods ≤ 0.3 eV TDDFT: RCAM-B3LYP/ Linear Response RICC2/ 6-31(d) Def2-SVP

Smiles CCC(CCCC(C)C(C)C)CC(C)(C)C(=CC1=C(C(=C2C(=N1)CCC2=CC)C)C)N=C(CN)N

0.5632

CCC(C)C(=CC1=C(C(=C2C(=N1)CCC2=C)C)C)N=C(CNNC(=N)CN)N

0.5464

0.02694 0.01537

CCC(C)C(=CC1=C(C(=C2C(=N1)CCC2=CC)C)C)N=C(CNNC(=N)CN)N

0.5613

0.02417

CC=C1CCC=2C1=C(C(=C(N2)C=C(C(C)C)N=C(CNNC(=N)CN)N)C)C

0.564

0.02509

CC=1C(=C2C(=NC1C=C(C(C)CCCC=C(C)C)N=C(CN)N)CCC2=C(C)CC3CC(C(C=C3)NC)(C)C)C

0.5523

0.02203

CC=C1CCC=2C1=C(C(=C(N2)C=C(C(C)CCCCC(C)(C)C(C=C)NC)N=C(CN)N)C)C

0.5678

0.03265

CCC(C)(C(C)CCC(C)C)C(=CC1=C(C(=C2C(=N1)CCC2=CC)C)C)N=C(C(C)N)N

0.5648

0.02203

CCCCC(CCCNC)CC(C)(C)C(CCCC(C)C(=CC1=C(C(=C2C(=N1)CCC2=CC)C)C)N=C(CN)N)C(C)C

0.5629

0.02914

CC=C1CCC=2C1=C(C(=C(N2)C(=CN=C(CN)N)CC=C(C)C(C)C)C)C

2.4804

1.38757

CCC(C)C(C)(C)CC(C)CCCC(=CC1=C(C(=C2C(=N1)CCC2=CC)C)C)N=C(CN)N

1.0272

0.36257

CC=C1CCC=2C1=C(C(=C(N2)C=C(C(C)CCCC(C)CCCC(C)C(C)C)N=C(CN)N)C)C

1.8683

0.90194

CC=C1C2=C(C(=C(N=C2CO1)C=C(CCN)N=C(CCN)N)C)C

1.5972

0.79958

CCC(C)C(C)(C)CC(C)CCCC(=CC1=C(C(=C2C(=N1)CCC2=C)C)C)N=C(CN)N

0.9988

0.34732

CC=C1CCC=2C1=C(C(=C(N2)C=C(C(C)C(C)(C(C=C)NC)N)N=C(CN)N)C)C

1.8032

0.88089

CCCCC(=C(C)CCC=C(C)C(C)C)CCC(=NC(=CC1=C(C(=C2C(=N1)CCC2=CC)C)C)CC)N

1.834

0.88361

The results from linear response RICC2 suggest that several of our designed compounds have a S1-T1 energy gap appropriate for supporting the TADF mechanism. The average error between the computed energy gap from linear response DFTB and TDDFT methods is ~0.8 eV. The TDDFT and DFTB methods computed gaps are offset by ~0.6 eV and ~0.3eV respectively, compared to the results from liner-response RICC2 analysis. We also found that the linear-response DFTB method computed gap is consistent with the S1-T1 gap computed using the spin flip TDDFT method (under prediction by ~0.1 eV, see supporting information Table 1). We recognize the limitations

23



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

of linear-response DFTB analysis and also of TDDFT methods to compute S1-T1 gaps. Nonetheless, 8 of 15 analyzed molecules had singlet-triplet energy gaps within a few kBT of room temperature, based on the RICC2 method. Hence, DFTB analysis can be used to assist in the discovery of promising molecular structures for potential uses as TADF based emitters. ▪

Conclusions We used the extended property-optimizing ACSESS method to construct a library

of structurally diverse organic compounds that are predicted to emit blue light (2.4-4.1 eV). The library contains 195 molecules, most of which (~70%) are novel structures that do not appear in the literature. The average of the computed light emission wavelengths for our designed library is 484 nm. The accuracy of the predicted emission wavelength was validated using TDDFT and RICC2 methods. As a demonstration of our extended multi-objective optimization approach, we optimized the singlet-triplet energy gap by utilizing only semiempirical computed oscillator strengths, singlet-triplet energy gaps, and maintaining the emission in the 2.4-4.1 eV range. We found ~60 structures with a computed ∆ES1-T1 ~0.3 eV that are predicted to emit in the blue and to have reasonably high oscillator strengths. We validated the computed gaps using TDDFT and RICC2 methods and found that the gaps for many designed molecules under 0.3 eV (within a few kBT at room temperature). Previous computational designs for TADF structures relied on the D-A-D motif,15 while our designs are not restricted to a specific motif. We were able to search an even broader chemical space to discover potential TADF emitters. We also showed that linear-response DFTB theory may be used to navigate chemical space to

24


Page 24 of 30

Page 25 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


identify lead TADF structures, with the aim of accelerating the computational design of these novel emitters.

AUTHOR INFORMATION

Corresponding Author *Phone: (919) 660-1526. E-mail: [email protected] *Phone: (919) 660-1562. E-mail: [email protected]

Notes

The authors declare no competing financial interest.

Acknowledgement

Support for this project was provided by the Samsung Advanced Institute of Technology (SAIT) Global Research Outreach (GRO) program. We thank OpenEye Scientific Software (Santa Fe, NM) for the use of Omega, OEChem, MolProp, and other tools.

References

(1)

Thejokalyani, N.; Dhoble, S. J. Novel Approaches for Energy Efficient Solid State Lighting by RGB Organic Light Emitting Diodes – A Review. Renew. Sustain. Energy Rev. 2014, 32, 448–467.

(2)

Thejo Kalyani, N.; Dhoble, S. J. Organic Light Emitting Diodes: Energy Saving Lighting technology—A Review. Renew. Sustain. Energy Rev. 2012, 16, 2696– 2723.

(3)

Friend, R.; Gymer, R.; Holmes, A.; Burroughes, J.; Marks, R.; Taliani, C.; Bradley, D.; Dos Santos, D.; Brédas, J.; Lögdlund, M.; Salaneck, R. Electroluminescence in Conjugated Polymers. Nature 1999, 397, 121–128.

(4)

Shen, J. Y.; Lee, C. Y.; Huang, T.-H.; Lin, J. T.; Tao, Y.-T.; Chien, C.-H.; Tsai, C.

25



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

High Tg Blue Emitting Materials for Electroluminescent Devices. Journal of Materials Chemistry, 2005, 15, 2455. (5)

Rothberg, L. J.; Lovinger, A. J. Status of and Prospects for Organic Electroluminescence. Journal of Materials Research, 1996, 11, 3174–3187.

(6)

Goushi, K.; Yoshida, K.; Sato, K.; Adachi, C. Organic Light-Emitting Diodes Employing Efficient Reverse Intersystem Crossing for Triplet-to-Singlet State Conversion. Nature Photonics, 2012, 6, 253–258.

(7)

Rupakheti, C.; Virshup, A.; Yang, W.; Beratan, D. N. A Strategy to Discover Diverse Optimal Molecules in the Small Molecule Universe. J. Chem. Inf. Model. 2015, 1–31.

(8)

Reineke, S. Organic Light-Emitting Diodes: Phosphorescence Meets Its Match. Nat. Photonics 2014, 8, 269–270.

(9)

Zhang, Q.; Li, B.; Huang, S.; Nomura, H.; Tanaka, H.; Adachi, C. Efficient Blue Organic Light-Emitting Diodes Employing Thermally Activated Delayed Fluorescence. Nat. Photonics 2014, 8, 326–332.

(10)

Zhao, H.; Liu, G.; Zhang, J.; Arif, R. A.; Tansu, N. Analysis of Internal Quantum Efficiency and Current Injection Efficiency in III-Nitride Light-Emitting Diodes. 2013, 9, 212–225.

(11)

Virshup, A. M.; Contreras-García, J.; Wipf, P.; Yang, W.; Beratan, D. N. Stochastic Voyages into Uncharted Chemical Space Produce a Representative Library of All Possible Drug-like Compounds. J. Am. Chem. Soc. 2013, 135, 7296–7303.

(12)

Aradi, B.; Hourahine, B.; Frauenheim, T. DFTB+, a Sparse Matrix-Based Implementation of the DFTB Method. J. Phys. Chem. A 2007, 111, 5678–5684.

(13)

Han, C.; Zhao, F.; Zhang, Z.; Zhu, L.; Xu, H.; Li, J.; Ma, D.; Yan, P. Constructing Low-Triplet-Energy Hosts for Highly Efficient Blue PHOLEDs: Controlling Charge and Exciton Capture in Doping Systems. Chem. Mater. 2013, 25, 4966– 4976.

(14)

Cui, L.-S.; Dong, S.-C.; Liu, Y.; Xu, M.-F.; Li, Q.; Jiang, Z.-Q.; Liao, L.-S. MetaLinked Spirobifluorene/phosphine Oxide Hybrids as Host Materials for Deep Blue Phosphorescent Organic Light-Emitting Diodes. Org. Electron. 2013, 14, 1924– 1930.

(15)

Shu, Y.; Levine, B. G. Simulated Evolution of Fluorophores for Light Emitting Diodes. J. Chem. Phys. 2015, 142, 104104.

(16)

Dewar, M. J. S.; Thiel, W. Ground States of Molecules. 38. The MNDO Method. Approximations and Parameters. J. Am. Chem. Soc. 1977, 99, 4899–4907.

(17)

Ridley, J.; Zerner, M. An Intermediate Neglect of Differential Overlap Technique for Spectroscopy: Pyrrole and the Azines. Theor. Chim. Acta 1973, 32, 111–134.

(18)

Parac, M.; Grimme, S. A TDDFT Study of the Lowest Excitation Energies of

26


Page 26 of 30

Page 27 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Polycyclic Aromatic Hydrocarbons. Chem. Phys. 2003, 292, 11–21. (19)

Schreiber, M.; Silva-Junior, M. R.; Sauer, S. P. A.; Thiel, W. Benchmarks for Electronically Excited States: CASPT2, CC2, CCSD, and CC3. J. Chem. Phys. 2008, 128.

(20)

Runge, E.; Gross, E. K. U. Density-Functional Theory for Time-Dependent Systems. Phys. Rev. Lett. 1984, 52, 997–1000.

(21)

Casida, M. E. Time-Dependent Density-Functional Theory for Molecules and Molecular Solids. Journal of Molecular Structure: THEOCHEM, 2009, 914, 3–18.

(22)

Frisch, M. J.; G. W. Trucks; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; X. Li, H. P. H.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A.; Jr., J. E. P.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Bro, and D. J. F. Gaussian 09, Revision A.02. Gaussian 09, Revision A.02, 2009.

(23)

Yanai, T.; Tew, D. P.; Handy, N. C. A New Hybrid Exchange-Correlation Functional Using the Coulomb-Attenuating Method (CAM-B3LYP). Chem. Phys. Lett. 2004, 393, 51–57.

(24)

Hariharan, P. C.; Pople, J. A. The Influence of Polarization Functions on Molecular Orbital Hydrogenation Energies. Theor. Chim. Acta 1973, 28, 213–222.

(25)

Blum, L. C.; Reymond, J.-L. 970 Million Druglike Small Molecules for Virtual Screening in the Chemical Universe Database GDB-13. J. Am. Chem. Soc. 2009, 131, 8732–8733.

(26)

Ruddigkeit, L.; van Deursen, R.; Blum, L. C.; Reymond, J.-L. Enumeration of 166 Billion Organic Small Molecules in the Chemical Universe Database GDB-17. J. Chem. Inf. Model. 2012, 52, 2864–2875.

(27)

OEChem 1.7.5, OpenEye Scientific Software, Inc., Santa Fe, NM USA, www.eyesopen.com, 2014.

(28)

Hawkins, P. C. D.; Skillman, A. G.; Warren, G. L.; Ellingson, B. A.; Stahl, M. T. Conformer Generation with OMEGA: Algorithm and Validation Using High Quality Structures from the Protein Databank and Cambridge Structural Database.

27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

J. Chem. Inf. Model. 2010, 50, 572–584. (29)

Burghardt, I.; Hughes, K. H.; Martinazzo, R.; Tamura, H.; Gindensperger, E.; Köppel, H.; Cederbaum, L. S. Conical Intersections Coupled to an Environment. In Conical Intersections; 2011; pp. 301–346.

(30)

Christiansen, O.; Koch, H.; Jørgensen, P. The Second-Order Approximate Coupled Cluster Singles and Doubles Model CC2. Chem. Phys. Lett. 1995, 243, 409–418.

(31)

Lee, C. W.; Lee, J. Y. High Quantum Efficiency in Blue Phosphorescent Organic Light Emitting Diodes Using Ortho-Substituted High Triplet Energy Host Materials. Org. Electron. 2013, 14, 1602–1607.

(32)

Wu, S.; Aonuma, M.; Zhang, Q.; Huang, S.; Nakagawa, T.; Kuwabara, K.; Adachi, C. High-Efficiency Deep-Blue Organic Light-Emitting Diodes Based on a Thermally Activated Delayed Fluorescence Emitter. J. Mater. Chem. C 2014, 2, 421–424.

(33)

Kim, B. S.; Lee, J. Y. Engineering of Mixed Host for High External Quantum Efficiency above 25% in Green Thermally Activated Delayed Fluorescence Device. Adv. Funct. Mater. 2014, 24, 3970–3977.

(34)

Teiber, M.; Giebeler, S.; Lessing, T.; Muller, T. J. J. Efficient Pseudo-FiveComponent Coupling-Fiesselmann Synthesis of Luminescent Oligothiophenes and Their Modification. Org. Biomol. Chem. 2013, 11, 3541–3552.

(35)

Lin, N.; Qiao, J.; Duan, L.; Wang, L.; Qiu, Y. Molecular Understanding of the Chemical Stability of Organic Materials for OLEDs: A Comparative Study on Sulfonyl, Phosphine-Oxide, and Carbonyl-Containing Host Materials. J. Phys. Chem. C 2014, 118, 7569–7578.

(36)

Schmuker, M.; Givehchi, A.; Schneider, G. Impact of Different Software Implementations on the Performance of the Maxmin Method for Diverse Subset Selection. Mol. Divers. 2004, 8, 421–425.

(37)

Furche, F.; Ahlrichs, R.; Hättig, C.; Klopper, W.; Sierka, M.; Weigend, F. Turbomole. Wiley Interdiscip. Rev. Comput. Mol. Sci. 2014, 4, 91–100.

(38)

Shao, Y.; Gan, Z.; Epifanovsky, E.; Gilbert, A. T. B.; Wormit, M.; Kussmann, J.; Lange, A. W.; Behn, A.; Deng, J.; Feng, X.; Ghosh, D.; Goldey, M.; Horn, P. R.; Jacobson, L. D.; Kaliman, I.; Khaliullin, R. Z.; Ku, T.; Landau, A.; Liu, J.; Proynov, E. I.; Rhee, Y. M.; Richard, R. M.; Rohrdanz, M. A.; Steele, R. P.; Sundstrom, E. J.; Lee, H.; III, W.; Zimmerman, P. M.; Zuev, D.; Albrecht, B.; Alguire, E.; Austin, B.; Beran, G. J. O.; Bernard, Y. A.; Berquist, E.; Brandhorst, K.; Bravaya, K. B.; Brown, S. T.; Casanova, D.; Chang, C.-M.; Chen, Y.; Chien, S. H.; Closser, K. D.; Crittenden, D. L.; Diedenhofen, M.; Distasio, R. A.; Do, H.; Dutoi, A. D.; Edgar, R. G.; Fatehi, S.; Fusti-Molnar, L.; Ghysels, A.; GolubevaZadorozhnaya, A.; Gomes, J.; Hanson-Heine, M. W. D.; Harbach, P. H. P.; Hauser, A. W.; Hohenstein, E. G.; Holden, Z. C.; Jagau, T.-C.; Ji, H.; Kaduk, B.; Khistyaev, K.; Kim, J.; Kim, J.; King, R. A.; Klunzinger, P.; Kosenkov, D.; Kowalczyk, T.; Krauter, C. M.; Lao, K. U.; Ele, A.; Laurent, D.; Lawler, K. V;

28


Page 28 of 30

Page 29 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Levchenko, S. V; Lin, C. Y.; Liu, F.; Livshits, E.; Lochan, R. C.; Luenser, A.; Manohar, P.; Manzer, S. F.; Mao, S.-P.; Mardirossian, N.; Marenich, A. V; Maurer, S. A.; Mayhall, N. J.; Neuscamman, E.; Oana, C. M.; Olivares-Amaya, R.; O ’neill, D. P.; Parkhill, J. A.; Perrine, T. M.; Peverati, R.; Prociuk, A.; Rehn, D. R.; Rosta, E.; Russ, N. J.; Sharada, S. M.; Sharma, S.; Small, D. W.; Sodt, A.; Stein, T.; Stück, D.; Su, Y.-C.; Thom, A. J. W.; Tsuchimochi, T.; Vanovschi, V.; Vogt, L.; Vydrov, O.; Wang, T.; Watson, M. A.; Wenzel, J.; White, A.; Williams, C. F.; Yang, J.; Yeganeh, S.; Yost, S. R.; You, Z.-Q.; Zhang, I. Y.; Zhang, X.; Zhao, Y.; Brooks, B. R.; Chan, G. K. L.; Chipman, D. M.; Cramer, C. J.; Goddard III Ah, W. A.; Gordon, M. S.; Hehre, W. J.; Klamt, A.; Schaefer III Aj, H. F.; Schmidt, M. W.; Sherrill, C. D.; Truhlar, D. G.; Warshel, A.; Xu, X.; AspuruGuzik, A.; Baer, R.; Bell, A. T.; Besley, N. A.; Chai, J.-D.; Dreuw, A.; Dunietz, B. D.; Furlani, T. R.; Gwaltney, S. R.; Hsu, C.-P.; Jung, Y.; Kong, J.; Lambrecht, D. S.; Liang, W.; Ochsenfeld, C.; Rassolov, V. A.; Slipchenko, L. V; Subotnik, J. E.; Voorhis, T. Van; Herbert, J. M.; Krylov, A. I.; Gill, P. M. W.; Head-Gordon, M. Advances in Molecular Quantum Chemistry Contained in the Q-Chem 4 Program Package. Mol. Phys. 2015, 113, 184–215.

29



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57

Page 30 of 30

e-‐ h⁺

h⁺e-‐ 75%

25% S1

ISC


T1

Diverse Optimal Molecular Libraries for Organic ... - ACS Publications

Recommend Documents