Photocatalytic Reduction of CO2 to HCOOH and CO by a Phosphine-Bipyridine-Phosphine Ir(III) Catalyst: Photophysics, Nonadiabatic Effects, Mechanism, and Selectivity.

Photocatalytic CO2 reduction is one of the best solutions to solve the global energy crisis and to realize carbon neutralization. The tetradentate phosphine-bipyridine (bpy)-phosphine (PNNP)-type Ir(III) photocatalyst, Mes-IrPCY2, was reported with a high HCOOH selectivity but the photocatalytic mechanism remains elusive. Herein, we employ electronic structure methods in combination with radiative, nonradiative, and electron transfer rate calculations, to explore the entire photocatalytic cycle to either HCOOH or CO, based on which a new mechanistic scenario is proposed. The catalytic reduction reaction starts from the generation of the precursor metal-to-ligand charge transfer (3 MLCT) state. Subsequently, the divergence happens from the 3 MLCT state, the single electron transfer (SET) and deprotonation process lead to the formation of one-electron-reduced species and Ir(I) species, which initiate the reduction reaction to HCOOH and CO, respectively. Interestingly, the efficient occurrence of proton or electron transfer reduces barriers of critical steps. In addition, nonadiabatic transitions play a nonnegligible role in the cycle. We suggest a lower free-energy barrier in the reaction-limiting step and the very efficient SET in 3 MLCT are cooperatively responsible for a high HCOOH selectivity. The gained mechanistic insights could help chemists to understand, regulate, and design photocatalytic CO2 reduction reaction of similar function-integrated molecular photocatalyst.

Engineering Crystal Growth and Surface Modification of Na3 V2 (PO4 )2 F3 Cathode for High-Energy-Density Sodium-Ion Batteries.

Na3 V2 (PO4 )2 F3 (NVPF) is a suitable cathode for sodium-ion batteries owing to its stable structure. However, the large radius of Na+ restricts diffusion kinetics during charging and discharging. Thus, in this study, a phosphomolybdic acid (PMA)-assisted hydrothermal method is proposed. In the hydrothermal process, the NVPF morphologies vary from bulk to cuboid with varying PMA contents. The optimal channel for accelerated Na+ transmission is obtained by cuboid NVPF. With nitrogen-doping of carbon, the conductivity of NVPF is further enhanced. Combined with crystal growth engineering and surface modification, the optimal nitrogen-doped carbon-covered NVPF cuboid (c-NVPF@NC) exhibits a high initial discharge capacity of 121 mAh g-1 at 0.2 C. Coupled with a commercial hard carbon (CHC) anode, the c-NVPF@NC||CHC full battery delivers 118 mAh g-1 at 0.2 C, thereby achieving a high energy density of 450 Wh kg-1 . Therefore, this work provides a novel strategy for boosting electrochemical performance by crystal growth engineering and surface modification.

Chalcogen Bond Catalysis with Telluronium Cations for Bromination Reaction: Importance of Electrostatic and Polarization Effects.

Recently, chalcogen bond catalysts with telluronium cations have garnered considerable attention in organic reactions. In this work, chalcogen bond catalysis on the bromination reaction of anisole with N-bromosuccinimide (NBS) with the telluronium cationic catalysts has been explored with density functional theory (DFT). The catalytic reaction is divided into two stages: the bromine transfer step and the proton transfer step. Based on the computational results, one can find the rate-determining step is the bromine transfer step. Moreover, the present study elucidates that a stronger chalcogen bond between catalysts and NBS will give better catalytic performance. Additionally, this work also clarified the importance of the electrostatic and polarization effects in the chalcogen bond between the oxygen atom of NBS and the Te atom of the catalyst in this bromination reaction. The electrostatic and polarization effects are significantly influenced by the electron-withdrawing ability of the substitution groups on the catalysts. Moreover, the structure-property relationship between the strength of chalcogen bond, electrostatic effect, polarization effect and catalytic performance are established for the design of more efficient chalcogen bond catalysts.

Thermally activated delayed fluorescence of a Ir(III) complex: absorption and emission properties, nonradiative rates, and mechanism.

One recent experimental study reported a Ir(III) complex with thermally activated delayed fluorescence (TADF) phenomenon in solution, but its luminescent mechanism is elusive. In this work, we combined density functional theory (DFT), time-dependent DFT (TDDFT) and multi-state complete active space second-order perturbation theory (MS-CASPT2) methods to investigate excited-state properties, photophysics, and emission mechanism of this Ir(III) complex. Two main absorption bands observed in experiments can be attributed to the electronic transition from the S0 state to the S1 and S2 states; while, the fluorescence and phosphorescence are generated from the S1 and T1 states, respectively. Both the S1 and T1 states have clear metal-to-ligand charge transfer (MLCT) character. The present computational results reveal a three-state model including the S0, S1 and T1 states to rationalize the TADF behavior. The small energy gap between the S1 and T1 states benefits the forward and reverse intersystem crossing (ISC and rISC) processes. At 300 K, the rISC rate is five orders of magnitude larger than the phosphorescence rate therefore enabling TADF. At 77 K, the rISC rate is sharply decreased but remains close to the phosphorescence rate; therefore, in addition to the phosphorescence, the delayed fluorescence could also contribute to the experimental emission. The estimated TADF lifetime agrees well with experiments, 9.80 vs. 6.67 µs, which further verifies this three-state model.

Theoretical studies on thermally activated delayed fluorescence of "carbene-metal-amide" Cu and Au complexes: geometric structures, excitation characters, and mechanisms.

"Carbene-metal(I)-amide" (CMA) complexes have garnered significant attention due to their remarkable properties and potential TADF applications in organic electronics. However, the atomistic working mechanism is still elusive. Herein, we chose two CMA complexes, i.e., cyclic (alkyl)(amino) carbene-copper[gold](I)-carbazole (CAAC-Cu[Au]-Cz), and employed both DFT and TD-DFT methods, in combination with radiative and nonradiative rate calculations, to investigate geometric and electronic structures of these two complexes in the ground and excited states, including orbital compositions, electronic transitions, absorption and emission spectra, and the luminescence mechanism. It is found that the coplanar or perpendicular conformations are coexistent in the ground state (S0), the lowest excited singlet state (S1), and the triplet state (T1). Both the coplanar and perpendicular S1 and T1 states have similar ligand-to-ligand charge transfer (LLCT) character between CAAC and Cz, and some charge-transfer character between metal atoms and ligands, which is beneficial to minimize the singlet-triplet energy gaps (ΔEST) and increase the spin-orbit coupling (SOC). An interesting three-state (S0, S1, T1) model involving two regions (coplanar and perpendicular) is proposed to rationalize the experimental TADF phenomena in the CMA complexes. In addition to the coplanar ones, the perpendicular S1 and T1 states also play a role in promoting the repopulation of the coplanar S1 exciton, which is a primary source for the delayed fluorescence.

Extending multi-layer energy-based fragment method for excited-state calculations of large covalently bonded fragment systems.

Recently, we developed a low-scaling Multi-Layer Energy-Based Fragment (MLEBF) method for accurate excited-state calculations and nonadiabatic dynamics simulations of nonbonded fragment systems. In this work, we extend the MLEBF method to treat covalently bonded fragment ones. The main idea is cutting a target system into many fragments according to chemical properties. Fragments with dangling bonds are first saturated by chemical groups; then, saturated fragments, together with the original fragments without dangling bonds, are grouped into different layers. The accurate total energy expression is formulated with the many-body energy expansion theory, in combination with the inclusion-exclusion principle that is used to delete the contribution of chemical groups introduced to saturate dangling bonds. Specifically, in a two-layer MLEBF model, the photochemically active and inert layers are calculated with high-level and efficient electronic structure methods, respectively. Intralayer and interlayer energies can be truncated at the two- or three-body interaction level. Subsequently, through several systems, including neutral and charged covalently bonded fragment systems, we demonstrate that MLEBF can provide accurate ground- and excited-state energies and gradients. Finally, we realize the structure, conical intersection, and path optimizations by combining our MLEBF program with commercial and free packages, e.g., ASE and SciPy. These developments make MLEBF a practical and reliable tool for studying complex photochemical and photophysical processes of large nonbonded and bonded fragment systems.

Bone-targeting exosome nanoparticles activate Keap1 / Nrf2 / GPX4 signaling pathway to induce ferroptosis in osteosarcoma cells.

BACKGROUND: In recent years, the development of BMSCs-derived exosomes (EXO) for the treatment of osteosarcoma (OS) is a safe and promising modality for OS treatment, which can effectively deliver drugs to tumor cells in vivo. However, the differences in the drugs carried, and the binding of EXOs to other organs limit their therapeutic efficacy. Therefore, improving the OS-targeting ability of BMSCs EXOs and developing new drugs is crucial for the clinical application of targeted therapy for OS. RESULTS: In this study, we constructed a potential therapeutic nano platform by modifying BMSCs EXOs using the bone-targeting peptide SDSSD and encapsulated capreomycin (CAP) within a shell. These constructed nanoparticles (NPs) showed the ability of homologous targeting and bone-targeting exosomes (BT-EXO) significantly promotes cellular endocytosis in vitro and tumor accumulation in vivo. Furthermore, our results revealed that the constructed NPs induced ferroptosis in OS cells by prompting excessive accumulation of reactive oxygen species (ROS), Fe2+ aggregation, and lipid peroxidation and further identified the potential anticancer molecular mechanism of ferroptosis as transduced by the Keap1/Nrf2/GPX4 signaling pathway. Also, these constructed NP-directed ferroptosis showed significant inhibition of tumor growth in vivo with no significant side effects. CONCLUSION: These results suggest that these constructed NPs have superior anticancer activity in mouse models of OS in vitro and in vivo, providing a new and promising strategy for combining ferroptosis-based chemotherapy with targeted therapy for OS.

Multi-omics perspective on studying reproductive biology in Daphnia sinensis.

Daphnia sinensis is a widespread freshwater microcrustacean. The assembled D. sinensis genome totaled 131.58 Mb with 92.23% of the assembly anchored onto 10 chromosomes. Based on the whole genome information, we further compared the transcriptomic and epigenomic characterization among parthenogenetic females, sexual females and males in D. sinensis. Transcriptomic analysis showed that the up-regulated genes in males were mainly grouped into the cuticle, sex differentiation and methyl farnesoate synthesis, which might play a pivotal role in steering development and reproduction processes. By comparison, the highly expressed genes in parthenogenetic females were mainly grouped into energy metabolism, mitosis, and DNA replication, which might contribute to maintaining rapid production of parthenogenetic females, and nutrient uptake for the growth of neonates. The whole-genome DNA methylation analysis showed that the methylation rate in parthenogenetic females was higher than that in sexual females and males, which might contribute to its rapid response to environment stress.

Nonadiabatic Derivative Couplings Calculated Using Information of Potential Energy Surfaces without Wavefunctions: Ab Initio and Machine Learning Implementations.

In this work, we implemented an approximate algorithm for calculating nonadiabatic coupling matrix elements (NACMEs) of a polyatomic system with ab initio methods and machine learning (ML) models. Utilizing this algorithm, one can calculate NACMEs using only the information of potential energy surfaces (PESs), i.e., energies, and gradients as well as Hessian matrix elements. We used a realistic system, namely CH2NH, to compare NACMEs calculated by this approximate PES-based algorithm and the accurate wavefunction-based algorithm. Our results show that this approximate PES-based algorithm can give very accurate results comparable to the wavefunction-based algorithm except at energetically degenerate points, i.e., conical intersections. We also tested a machine learning (ML)-trained model with this approximate PES-based algorithm, which also supplied similarly accurate NACMEs but more efficiently. The advantage of this PES-based algorithm is its significant potential to combine with electronic structure methods that do not implement wavefunction-based algorithms, low-scaling energy-based fragment methods, etc., and in particular efficient ML models, to compute NACMEs. The present work could encourage further research on nonadiabatic processes of large systems simulated by ab initio nonadiabatic dynamics simulation methods in which NACMEs are always required.

A Deep Learning-Based Approach for Glomeruli Instance Segmentation from Multistained Renal Biopsy Pathologic Images.

Glomeruli instance segmentation from pathologic images is a fundamental step in the automatic analysis of renal biopsies. Glomerular histologic manifestations vary widely among diseases and cases, and several special staining methods are necessary for pathologic diagnosis. A robust model is needed to segment and classify glomeruli with different staining methods and apply in cases with various glomerular pathologic changes. Herein, pathologic images from renal biopsy slides stained with three basic special staining methods were used to build the data sets. The snapshot group included 1970 glomeruli from 516 patients, and the whole-slide image group included 8665 glomeruli from 148 patients. Cascade Mask region-based convolutional neural net architecture was trained to detect, classify, and segment glomeruli into three categories: i) GN, structural normal; ii) global sclerosis; and iii) glomerular with other lesions. In the snapshot group, total glomeruli, GN, global sclerosis, and glomerular with other lesions achieved an F1 score of 0.914, 0.896, 0.681, and 0.756, respectively, which were comparable with those in the whole-slide image group (0.940, 0.839, 0.806, and 0.753, respectively). Among the three categories, GN achieved the best instance segmentation effect in both groups, as determined by average precision, average recall, F1 score, and Mask mean Intersection over Union. The present model segments and classifies multistained glomeruli with efficiency and robustness. It can be applied as the first step for more detailed glomerular histologic analysis.

Thermally Activated Delayed Fluorescence of a Dinuclear Platinum(II) Compound: Mechanism and Roles of an Upper Triplet State.

A dinuclear Pt(II) compound was reported to exhibit thermally activated delayed fluorescence (TADF); however, the luminescence mechanism remains elusive. To reveal relevant excited-state properties and luminescence mechanism of this Pt(II) compound, both density function theory (DFT) and time-dependent DFT (TD-DFT) calculations were carried out in this work. In terms of the results, the S1 and T2 states show mixed intraligand charge transfer (ILCT)/metal-to-ligand CT (MLCT) characters while the T1 state exhibits mixed ILCT/ligand-to-metal CT (LMCT) characters. Mechanistically, a four-state (S0 , S1 , T1 , and T2 ) model is proposed to rationalize the TADF behavior. The reverse intersystem crossing (rISC) process from the initial T1 to final S1 states involves two up-conversion channels (direct T1 →S1 and T2 -mediated T1 →T2 →S1 pathways) and both play crucial roles in TADF. At 300 K, these two channels are much faster than the T1 phosphorescence emission enabling TADF. However, at 80 K, these rISC rates are reduced by several orders of magnitude and become very small, which blocks the TADF emission; instead, only the phosphorescence is observed. These findings rationalize the experimental observation and could provide useful guidance to rational design of organometallic materials with superior TADF performances.

Thermally Activated Delayed Fluorescence Mechanism of a Bicyclic "Carbene-Metal-Amide" Copper Compound: DFT/MRCI Studies and Roles of Excited-State Structure Relaxation.

Herein we investigated the luminescence mechanism of one "carbene-metal-amide" copper compound with thermally activated delayed fluorescence (TADF) using density functional theory (DFT)/multireference configuration interaction, DFT, and time-dependent DFT methods with the polarizable continuum model. The experimentally observed low-energy absorption and emission peaks are assigned to the S1 state, which exhibits clear interligand and partial ligand-to-metal charge-transfer character. Moreover, it was found that a three-state (S0, S1, and T1) model is sufficient to describe the TADF mechanism, and the T2 state should play a negligible role. The calculated S1-T1 energy gap of 0.10 eV and proper spin-orbit couplings facilitate the reverse intersystem crossing (rISC) from T1 to S1. At 298 K, the rISC rate of T1 → S1 (∼106 s-1) is more than 3 orders of magnitude larger than the T1 phosphorescence rate (∼103 s-1), thereby enabling TADF. However, it disappears at 77 K because of a very slow rISC rate (∼101 s-1). The calculated TADF rate, lifetime, and quantum yield agree very well with the experimental data. Methodologically, the present work shows that only considering excited-state information at the Franck-Condon point is insufficient for certain emitting systems and including excited-state structure relaxation is important.

Generalized Ab Initio Nonadiabatic Dynamics Simulation Methods from Molecular to Extended Systems.

Nonadiabatic dynamics simulation has become a powerful tool to describe nonadiabatic effects involved in photophysical processes and photochemical reactions. In the past decade, our group has developed generalized trajectory-based ab initio surface-hopping (GTSH) dynamics simulation methods, which can be used to describe a series of nonadiabatic processes, such as internal conversion, intersystem crossing, excitation energy transfer and charge transfer of molecular systems, and photoinduced nonadiabatic carrier dynamics of extended systems with and without spin-orbit couplings. In this contribution, we will first give a brief introduction to our recently developed methods and related numerical implementations at different computational levels. Later, we will present some of our latest applications in realistic systems, which cover organic molecules, biological proteins, organometallic compounds, periodic organic and inorganic materials, etc. Final discussion is given to challenges and outlooks of ab initio nonadiabatic dynamics simulations.

Effects of doping high-valence transition metal (V, Nb and Zr) ions on the structure and electrochemical performance of LIB cathode material LiNi0.8Co0.1Mn0.1O2.

Ni-rich layered oxides, like LiNi0.8Co0.1Mn0.1O2 (NCM811), have been widely investigated as cathodes due to their high energy density. However, gradual structural transformation during cycling can lead to capacity degradation and potential decay of cathode materials. Herein, we doped high-valence transition metal (TM) ions (V5+, Nb5+, and Zr4+) at the Ni site of NCM811 by first principles simulations and explored the mechanism of doping TMs in NCMs for enhancing the electrochemical performance. Analysis of the calculations shows that doping V, Nb and Zr has an efficient influence on alleviating the Ni oxidation, reducing the loss of oxygen, and facilitating Li+ migration. Moreover, V doping can further suppress the lattice distortion due to the radius of V5+ being close to the radius of Mn4+. In particular, compared with the barrier of the pristine NCM in Li divacancy, the barrier of V-doped NCM reaches the lowest. In conclusion, V is the most favorable dopant for NCM811 to improve the electrochemical properties and achieve both high capacity and cycling stability.

Blue-AsP monolayer as a promising anode material for lithium- and sodium-ion batteries: a DFT study.

Based on first-principle calculations, we proposed a one two-dimensional (2D) blue AsP (b-AsP) monolayer as an ideal anode material for lithium/sodium-ion (Li/Na-ion) batteries for the first time. The b-AsP monolayer possesses thermal and dynamic stabilities. The system undergoes the transition from semiconductor to metal after Li/Na atoms are embedded, which ensures good electric transportation. Most remarkably, our results indicate that the b-AsP monolayer exhibits high theoretical capacities of 1011.2 mA h g-1 (for Li) and 1769.6 mA h g-1 (for Na), low average open circuit voltages of 0.17 eV for Li4AsP and 0.14 eV for Na7AsP systems and ultrafast diffusivity with the low energy barriers of 0.17/0.15 eV and 0.08/0.07 eV of the P/As sides for Li and Na, respectively. Given these exceptional properties, the synthesis of a buckled b-AsP monolayer is desired to achieve a promising electrode material for Li- and Na-ion batteries.

Theoretical insights into the thermal reduction of N2 to NH3 over a single metal atom incorporated nitrogen-doped graphene.

Density functional theory calculations have been performed to study the reaction mechanism of N2 thermal reduction (N2TR) over a single metal atom incorporated nitrogen-doped graphene. Our results reveal that the type of metal atoms and their coordination environments have a significant effect on the catalytic activity of N2TR. Regarding CoN4- and FeN4-embedded graphene sheets that the metal atom is fourfold coordinated, they are inactive for N2TR owing to the poor stability of the adsorbed H2 and N2 molecules. In contrast, if the monodisperse metal atom is surrounded by three N atoms, namely, CoN3/G and FeN3/G show activity toward N2TR, and catalytic conversion of N2 into ammonia is achieved through the associative mechanism rather than the dissociative mechanism. Further investigations show that the synthesis of NH3 over the two surfaces is mainly through the formation of an NHNH* intermediate; however, the detailed reaction mechanisms are sensitive to the type of metal atom introduced into N-doped graphene. Based on the calculated kinetic barriers, FeN3/G exhibits a better catalytic activity for N2TR. The superior performance of FeN3/G can be attributed to the fact that this surface prefers a high spin-polarized state during the whole process of N2TR, while the non-spin polarized state is predicted as the ground state for most of the elementary steps of N2-fixation over CoN3/G. The present study provides theoretical insights into developing graphene-based single atom catalysts with a high activity toward ammonia synthesis through N2TR.

BODIPY-Based Photodynamic Agents for Exclusively Generating Superoxide Radical over Singlet Oxygen.

Developing Type-I photosensitizers is considered as an efficient approach to overcome the deficiency of traditional photodynamic therapy (PDT) for hypoxic tumors. However, it remains a challenge to design photosensitizers for generating reactive oxygen species by the Type-I process. Herein, we report a series of α,ß-linked BODIPY dimers and a trimer that exclusively generate superoxide radical (O2-. ) by the Type-I process upon light irradiation. The triplet formation originates from an effective excited-state relaxation from the initially populated singlet (S1 ) to triplet (T1 ) states via an intermediate triplet (T2 ) state. The low reduction potential and ultralong lifetime of the T1 state facilitate the efficient generation of O2-. by inter-molecular charge transfer to molecular oxygen. The energy gap of T1 -S0 is smaller than that between 3 O2 and 1 O2 thereby precluding the generation of singlet oxygen by the Type-II process. The trimer exhibits superior PDT performance under the hypoxic environment.

Inhibition of leucine-rich repeats and calponin homology domain containing 1 accelerates microglia-mediated neuroinflammation in a rat traumatic spinal cord injury model.

BACKGROUND: Spinal cord injury (SCI) triggers the primary mechanical injury and secondary inflammation-mediated injury. Neuroinflammation-mediated insult causes secondary and extensive neurological damage after SCI. Microglia play a pivotal role in the initiation and progression of post-SCI neuroinflammation. METHODS: To elucidate the significance of LRCH1 to microglial functions, we applied lentivirus-induced LRCH1 knockdown in primary microglia culture and tested the role of LRCH1 in microglia-mediated inflammatory reaction both in vitro and in a rat SCI model. RESULTS: We found that LRCH1 was downregulated in microglia after traumatic SCI. LRCH1 knockdown increased the production of pro-inflammatory cytokines such as IL-1ß, TNF-α, and IL-6 after in vitro priming with lipopolysaccharide and adenosine triphosphate. Furthermore, LRCH1 knockdown promoted the priming-induced microglial polarization towards the pro-inflammatory inducible nitric oxide synthase (iNOS)-expressing microglia. LRCH1 knockdown also enhanced microglia-mediated N27 neuron death after priming. Further analysis revealed that LRCH1 knockdown increased priming-induced activation of p38 mitogen-activated protein kinase (MAPK) and Erk1/2 signaling, which are crucial to the inflammatory response of microglia. When LRCH1-knockdown microglia were adoptively injected into rat spinal cords, they enhanced post-SCI production of pro-inflammatory cytokines, increased SCI-induced recruitment of leukocytes, aggravated SCI-induced tissue damage and neuronal death, and worsened the locomotor function. CONCLUSION: Our study reveals for the first time that LRCH1 serves as a negative regulator of microglia-mediated neuroinflammation after SCI and provides clues for developing novel therapeutic approaches against SCI.

A boron-decorated melon-based carbon nitride as a metal-free photocatalyst for N2 fixation: a DFT study.

On the basis of the electron "acceptance-donation" concept, a boron decorated melon-based carbon nitride (CN) is studied as a metal-free photocatalyst to efficiently reduce N2 to NH3 under visible light irradiation. The results revealed that a boron-interstitial (Bint)-decorated melon-based CN has an outstanding N2 reduction capacity through the enzymatic mechanism with a rather low overpotential (0.32 V). The excellent efficiency and selectivity of Bint-decorated melon-based CN in N2 reduction reaction (NRR) are attributed to the concentrated spin polarization on the B atom, the significant enhancement of visible and infrared light absorption, and the effective inhibition of the competitive hydrogen evolution reaction (HER). Importantly, B-doped melon-based CN has been successfully synthesized in the experiments, so obtaining Bint-decorated melon is promising, while proton transfer from the -NH2 group in CN to the B atom surely will affect the functionality of the catalyst through deactivation of the N2 adsorption site. Our study provides a novel single atom metal-free photocatalyst with high efficiency for NRR, which is conducive to the sustainable synthesis of ammonia.

Efficient Construction of Excited-State Hessian Matrices with Machine Learning Accelerated Multilayer Energy-Based Fragment Method.

Recently, we have developed a multilayer energy-based fragment (MLEBF) method to describe excited states of large systems in which photochemically active and inert regions are separately treated with multiconfigurational and single-reference electronic structure method and their mutual polarization effects are naturally described within the many-body expansion framework. This MLEBF method has been demonstrated to provide highly accurate energies and gradients. In this work, we have further derived the MLEBF method with which highly accurate excited-state Hessian matrices of large systems are efficiently constructed. Moreover, in combination with recently proposed embedded atom neural network (EANN) model we have developed a machine learning (ML) accelerated MLEBF method (i.e., ML-MLEBF) in which photochemically inert region is entirely replaced with trained ML models. ML-MLEBF is found to improve computational efficiency of Hessian matrices in particular for large systems. Furthermore, both MLEBF and ML-MLEBF methods are highly parallel and exhibit low-scaling computational cost with multiple CPUs. The present developments could motivate combining various ML techniques with fragment-based electronic structure methods to explore Hessian-matrix-based excited-state properties of large systems.