Data-driven machine learning models for quick prediction of thermal stability properties of OLED materials

Organic light-emitting diode (OLED) materials have exhibited a wide range of applications. However, the further development and commercialization of OLEDs requires higher quality OLED materials, including materials with a high thermal stability. Thermal stability is associated with the glass transition temperature (T_g) and decomposition temperature (T_d), but experimental determinations of these two important properties generally involve a time-consuming and laborious process. Thus, the development of a quick and accurate prediction tool is highly desirable. Motivated by the challenge, we explored machine learning (ML) by constructing a new dataset with more than 1,000 samples collected from a wide range of literature, through which ensemble learning models were explored. Models trained with the LightGBM algorithm exhibited the best prediction performance, where the values of mean absolute error, root mean squared error, and R² were 17.15 K, 24.63 K, and 0.77 for T_g prediction and 24.91 K, 33.88 K, and 0.78 for T_d prediction. The prediction performance and the generalization of the ML models were further tested by two applications, which also exhibited satisfactory results. Experimental validation further demonstrated the reliability and the practical potential of the ML-based models. In order to extend the practical application of the ML-based models, an online prediction platform was constructed. This platform includes the optimal prediction models and all the thermal stability data under study, and it is freely available at http://www.oledtppxmpugroup.com. We expect that this platform will become a useful tool for experimental investigation of T_g and T_d, accelerating the design of OLED materials with desired properties.     

不对称刚性扭曲菲并咪唑分子的设计合成及其深蓝电致发光性能研究

本研究通过不对称、刚性扭曲的分子设计理念,合成了高效深蓝有机电致发光材料MBTPI。该化合物具有很高的分解温度(496℃)与玻璃转化温度(190℃),有利于提高器件的稳定性;不对称刚性扭曲的分子构型有效控制了分子的整体共轭程度,使发光波长在深蓝光区,固体发光量子产率高达74%。理论计算验证了分子不对称扭曲的构型,并且发现甲基的引入对前线轨道分布影响不大,分子保留了较好的双极性质。基于MBTPI的非掺杂器件发射出非常高效的深蓝光。色纯度为(0.15,0.07),非常接近NTSC的蓝光标准(0.14,0.08)。最大外量子效率为4.91%,并且效率滚降很小,为性能最好的非掺杂深蓝光器件之一。     

Calorimetric study and modeling of molecular mobility in amorphous organic pharmaceutical compounds using a modified Adam-Gibbs approach

The purpose of this study is to provide a quantitative characterization of the thermal behavior of amorphous organic pharmaceutical compounds across their glass transition temperature, and to assess their molecular mobility as a function of temperature and time by combining theoretical simulations with experimental measurements using differential scanning calorimetry. A computational approach built on the Boltzmann superposition principle of nonexponential decay and the Adam-Gibbs theory of entropic-dependent structural relaxation is presented. The heat capacities of the crystalline and amorphous forms are incorporated into the simulation in order to accurately assess the entropic fictive temperature as functions of temperature and time under any arbitrary set of experimental conditions. Using this method, we evaluated properties of the glass former, D and T0, and the nonexponentiality index beta, for amorphous salicin, felodipine, and nifedipine, by fitting the simulated glass transition profile with the experimentally determined heat capacity across the glass transition region. From this fit, the evolution of the relaxation time of the model compounds following any thermal cycle, including heating, cooling, and isothermal holds can then be estimated a priori. This study reveals the profound and inextricable effect of thermal history on the molecular mobility of the amorphous materials, and the ability of the glass to undergo fast changes in its molecular motions over an aging process even at low temperatures.     

Anisotropic pseudorotation of the photoexcited triplet state of fullerene C60 in molecular glasses studied by pulse EPR

Spin-polarized echo-detected electron paramagnetic resonance (EPR) spectra and the transversal relaxation rate T2(-1) of the photoexcited triplet state of fullerene C60 molecules were studied in o-terphenyl, 1-methylnaphthalene, and decalin glassy matrices. The model is composed of a fast (correlation time approximately 10(-12) s) pseudorotation of (3)C60 in a local anisotropic potential created by interaction of the fullerene molecule with the surrounding matrix molecules. In simulations, this potential is assumed to be axially symmetric around some axis of a preferable orientation in a matrix cage. The fitted value of the potential was found to depend on the type of glass and to decrease monotonically with a temperature increase. A sharp increase of the T2(-1) temperature dependence was found near 240 K in glassy o-terphenyl and near 100 K in glassy 1-methylnaphthalene and decalin. This increase probably is related to the influence on the pseudorotation of the onset of large-amplitude vibrational molecular motions (dynamical transition in glass) that are known for glasses from neutron scattering and molecular dynamics studies. The obtained results suggest that molecular and spin dynamics of the triplet fullerene are extremely sensitive to molecular motions in glassy materials.     

A NEURAL NETWORK STUDY ON GLASS TRANSITION TEMPERATURE OF POLYMERS

In this paper, an atificial neural network model is adopted to study the glass transition temperature of polymers. Inour artificial neural networks, the input nodes are the characteristic ratio C_∞, the average molecular weigh M_e betweenentanglement points and the molecular weigh M_(mon) of repeating unit. The output node is the glass transition temperature T_g,and the number of the hidden layer is 6. We found that the artificial neural network simulations are accurate in predicting theoutcome for polymers for which it is not trained. The maximum relative error for predicting of the glass transitiontemperature is 3.47%, and the overall average error is only 2.27%. Artificial neural networks may provide some new ideas toinvestigate other properties of the polymers.     

Synthesis,Structure, Properties,and Application of a Carbazole‐Based Diaza[7]helicene in a Deep‐Blue‐Emitting OLED

A carbazole‐based diaza[7]helicene, 2,12‐dihexyl‐2,12‐diaza[7]helicene ( 1 ), was synthesized by a photochemical synthesis and its use as a deep‐blue dopant emitter in an organic light‐emitting diode (OLED) was examined. Compound 1 exhibited good solubility and excellent thermal stability with a high decomposition temperature (T_d=372.1 °C) and a high glass‐transition temperature (T_g, up to 203.0 °C). Single‐crystal structural analysis of the crystalline clathrate ( 1 )₂ ? cyclohexane along with a theoretical investigation revealed a non‐planar‐fused structure of compound 1 , which prevented the close‐packing of molecules in the solid state and kept the molecule in a good amorphous state, which allowed the optimization of the properties of the OLED. A device with a structure of ITO/NPB (50 nm)/CBP:5 % 1 (30 nm)/BCP (20 nm)/Mg:Ag (100 nm)/Ag (50 nm) showed saturated blue light with Commission Internationale de L’Eclairage (CIE) coordinates of (0.15, 0.10); the maximum luminance efficiency and brightness were 0.22 cd A^?1 (0.09 Lm W^?1) and 2365 cd m^?2, respectively. This new class of helicenes, based on carbazole frameworks, not only opens new possibilities for utilizing helicene derivatives in deep‐blue‐emitting OLEDs but may also have potential applications in many other fields, such as molecular recognition and organic nonlinear optical materials.     

分子距边矢量研究链烃与醛酮的定量构效关系

按碳原子及键合特性分类定义并计算了链烃包括烷、烯,炔,双烯,烯炔烃的分子距离－边数矢量（ＭＤＥ）,将１５３个链烃的ＭＤＥ矢量与相应的沸点相关联,得到良好的线性模型,复相关系数Ｒ＝０．９９７６,均方根误差０．９９７５、ＲＭＳ＝４．７２Ｋ和Ｒ＝０．９９７２、ＲＭＳ＝５．１３Ｋ。结果表明模型具有良好的稳定性和预测能力。进一步对杂原用染色因子进行标识,提出了一种适用于含杂原子体系分子结构描述的ＭＤＥ矢量,     

Local segmental relaxation in bidisperse polystyrenes

Dynamic mechanical results are reported for segmental relaxation of monodisperse polystyrenes (PSs) with molecular weights of 0.7, 3, 18, and 104 kg/mol and bidisperse PSs created from blending pairs of these materials. The data for the monodisperse polymers confirm previous findings; namely, there is an increase in the glass‐transition temperature normalized temperature dependence of the segmental relaxation times (fragility) with increasing molecular weight, along with a breakdown of the correlation between the fragility and the breadth of the relaxation function. For both the monodisperse and bidisperse PSs, the glass‐transition temperature is a single function of the average number of chain ends, independent of the nature of the molecular weight distribution. It is also found that these materials exhibit fragilities that uniquely depend on the number‐average molecular weight, that is, on the concentration of chain ends. In blends with linear PS, cyclic PS with a low molecular weight behaves as a high polymer, similar to its neat behavior, reflecting the overriding importance of chain ends. © 2004 Wiley Periodicals, Inc. J Polym Sci Part B: Polym Phys 42: 2604–2611, 2004     

On the use of computational neural networks for the prediction of polymer properties

A general purpose computational paradigm using neural networks is shown to be capable of efficiently predicting properties of polymeric compounds based on the structure and composition of the monomeric repeat unit. Results are discussed for the prediction of the heat capacity, glass transition temperature, melting temperature, change in the heat capacity at the glass transition temperature, degradation temperature, tensile strength and modulus, ultimate elongation, and compressive strength for 11 different families of polymers. The accuracies of the predictions range from 1–13% average absolute error. The worst results were obtained for the mechanical properties (tensile strength and modulus: 13%, 7% elongation: 12%, and compressive strength: 8%) and the best results for the thermal properties (heat capacity, glass transition temperature, and melting point: <4%). A simple modification to the overall method is devised to better take into account the fact that the mechanical properties are experimentally determined with a fairly large range (due to variability in measurement procedures and especially the sample). This modification treats the bounds on the range for the mechanical properties as complex numbers (complex, modular neural networks) and leads to more rapid optimization with a smaller average error (reduced by 3%).Dedicated to Professor Bernhard Wunderlich on the occasion of his 65th birthdayThis research was sponsored by the Division of Materials Sciences, Office of Basic Energy Sciences, U.S. Department of Energy, under Contract No. DE-AC05-84R21400 with Lockheed Martin Energy Systems, Inc. We would like to express our gratitude for the continued collaboration, support, and interest of Prof. Wunderlich in our research. We would also like to thank participants of the 1st DOE Workshop on Applications of Neural Networks in Materials Sciences for useful discussion on materials properties and neural networks.     

A novel approach toward the prediction of the glass transition temperature: Application of the EVM model,a designer QSPR equation for the prediction of acrylate and methacrylate polymers

We describe an original QSPR model called the EVM model (Energy, Volume, Mass) to calculate the glass transition temperature (T_g) of aliphatic acrylate and methacrylate homopolymers using classical molecular mechanics and dynamics. The latter was used to calculate an energy density function related to the cylindrical volume of a 20 monomer unit polymer segment (TSSV, Total Space around a Standard deviation Volume). We then calculated the T_g as a function of this density function and the repeat unit molecular weight, although no interchain interactions were taken into account. For linear and branched aliphatic acrylate and methacrylate polymers, the standard deviation from linear regression was 12 K, and the r² was 0.96. The model allows calculation of the T_g with an average absolute error of error of 10% for linear and branched derivatives not included in the original linear regression analysis. The results obtained with the EVM model are compared with those obtained with Bicerano's model. © 1997 John Wiley & Sons, Inc. J Polym Sci A: Polym Chem 35: 2579–2590, 1997