Search

The search for new two-dimensional (2D) carbon nitride (CN) materials with desirable properties is an important and critical area of research. This study presents a series of 2D CN materials with appealing features such as semimetal or flat bands. By utilizing group theory and graph theory, we perform a random search of the configurational space of 2D CN monolayers, leading to the identification of 721 new CN materials. Extensive density functional theory calculations were then conducted to evaluate their properties. We found that 57 materials exhibit semimetal behavior, while 664 display characteristics of semiconductors and normal metals. Notably, materials such as 51-5-16-C 3 N-r568 and 55-5-20-C 4 N-r568 feature topological Dirac semimetals with protected 1D topological boundary states. Moreover, 31 materials possess flat bands near their Fermi level, including two with kagome-like flat bands that induce intrinsic spontaneous magnetization due to partially-occupied flat bands. Our results offer several promising candidates for exploring physical phenomena related to semimetals or flat bands in 2D CN materials. [Display omitted] [ABSTRACT FROM AUTHOR]

• This work constitutes a timely and important advancement in the interdisciplinary area of photocatalytic water splitting and machine learning technology. • The machine learning screening method proposed in this paper can greatly accelerate the progress of photocatalyst discovery. We showed that ten potential photocatalysts can be selected from 860 metal oxides within seconds based on the established screening model. • A prediction model of energy level was established based on regression algorithm. By using regression algorithm, we can train the prediction model based on a relatively small dataset, and the accuracy of the prediction model is very high (R 2 = 0.908). • An effective photocatalyst of CsYO 2 was selected. We used DFT calculations to probe the free energy change during hydrogen emission reaction and oxygen emission reaction, and the effectiveness of CsYO 2 was verified. • Besides, the machine learning screening method proposed in the paper can be extended to screen multi-component metal oxide photocatalysts by establishing its corresponding material library. Photocatalytic water splitting represents a sustainable avenue for hydrogen production, yet the traditional trial-and-error approach for identifying efficient photocatalysts among numerous candidates is cumbersome. In this paper, we introduce a machine learning (ML)-driven approach for rapid screening of promising metal oxide photocatalysts for hydrogen generation. This methodology comprises three key steps: the construction of a comprehensive material library, ML-accelerated material screening, and Density Functional Theory (DFT) calculations for verification. Through this process, we successfully identify ten metal oxides with good energy levels and charge carrier transport properties from an initial library of 860 candidates. Notably, CsYO 2 emerges as a standout candidate, exhibiting excellent photocatalytic water-splitting performance on its (001) crystal face, with Y atoms identified as active sites through DFT calculations. This study demonstrates the efficacy of applying machine learning to the precise and rapid screening of vast material databases, paving the way for discovering novel and efficient photocatalysts. [Display omitted] [ABSTRACT FROM AUTHOR]

Acute promyelocytic leukemia (APL) is often accompanied by potentially fatal coagulopathy, especially in high-risk APL. Bleeding, particularly severe bleeding is the leading cause of early death (ED). Meanwhile, thrombosis, the other major coagulopathic complication, is being increasingly recognized. However, predictors of thrombohemorrhagic disorders are still not well investigated. In this study, we retrospectively studied 83 patients with high-risk APL and categorized them into severe bleeding, thrombosis and no evident events groups. Severe bleeding was observed in 15 patients, nearly half of whom died of hemorrhage, while thrombosis was observed in 12 patients. Risk factor analysis showed that high WBC (>58.76 × 109/L) (p = 0.001) and prolonged PT (>17.7 s) (p = 0.015) could be independent predictors for severe bleeding, while high WBC/D-dimer>5.12 (p = 0.002) and low D-dimer/FIB<5.14 (p = 0.03) could be independent predictors for thrombosis in high-risk APL patients. Moreover, there are significant differences in WBC/D-dimer and D-dimer/FIB between DIC and Non-DIC groups (p < 0.001). Notably, we found that the WBC/D-dimer was dramatically higher in the thrombotic group than in the other two groups at the time of admission or during the first week of induction therapy. High WBC and prolonged PT could predict severe bleeding in high-risk APL patients, while high WBC/D-dimer and low D-dimer/FIB could be independent predictors for thrombosis. For high-risk APL, WBC/D-dimer and D-dimer/FIB are also beneficial in the diagnosis of DIC. WBC/D-dimer might help early identification of thrombosis at the time of admission or during the first week of induction therapy. • High WBC and prolonged PT could predict severe bleeding in high-risk APL patients. • WBC/D-dimer and D-dimer/FIB can be beneficial for early identification of thrombosis and the diagnosis of DIC. • PT, D-dimer, FDP/FIB, D-dimer/FIB and WBC/D-dimer could differentiate severe hemorrhage from thrombosis. • WBC/D-dimer might help early identification of thrombosis at admission or during the first week of induction therapy. [ABSTRACT FROM AUTHOR]