Scientific Reports最新文献

This work addresses enhancing the performance of Perovskite solar cells by incorporating ellipsoid plasmonic nanoparticles of Titanium Nitride (TiN) into the pure methylammonium lead iodide (CH₃NH₃PbI₃) active layer. The device structure consists of ITO/TiO₂/CH₃NH₃PbI₃/PEDOT: PSS/Au. The study is conducted using finite difference time domain method for the optical studies and SCAPS-1D software for the electrical parameters, the proposed structure managed to boost the absorption of the perovskite solar cell to levels more than 90% in the visible and near infrared range (NIR), and a broad absorption band from 400 nm to 1200 nm is obtained, which is reflected as an electrical performance with short circuit current density of 46.84 mA/ cm², and open circuit voltage of 0.8924 V, a fill factor of 76.22%, and an impressive power conversion efficiency of 31.8%. These results showcase the plasmonic potential of TiN refractory metal nanoparticles in enhancing the performance of perovskite solar cells.

Controlling the invasive alien Himalayan balsam Impatiens glandulifera is both expensive and time-consuming. The most promising control methods include manual mowing and hand-pulling. However, long-term effectiveness of these methods can be unsatisfactory. We hypothesized that such partial success might result from eradication efforts occurring too late in the season, allowing improperly treated cuttings to flower. Moreover, we experimentally tested whether cut flowering individuals left on the ground could survive long enough to produce viable seeds. The experiment involved plant cuttings and reference individuals. We recorded the number of flowers on each plant, monitored floral visitor activity, and simultaneously measured abiotic factors (air temperature, solar radiation, wind speed). The relationship between insect visitation rates and both flower abundance and solar radiation was nonlinear, highlighting the importance of considering complex environmental effects in pollination dynamics. The seeds were collected, and seed viability was compared between the two treatment groups. We demonstrated that the plant cuttings survived for about three weeks, developed flowers and produced viable seeds. Although the reference individuals had more flowers and were more frequently visited by insects (mainly Bombus pascuorum), they did not develop significantly more viable seeds than the cuttings. This confirms the remarkable survival capacity of I. glandulifera plant cuttings and supports our hypothesis that this feature may hinder the full eradication of the species through mowing and hand-pulling under current guidelines. Based on our results, we recommend changing the control timing.

Pseudomonas aeruginosa is a common pathogen of a wide range of nosocomial infections in humans, as well as foodborne illnesses. The current study focuses on the molecular genotyping of P. aeruginosa recovered from human, water, and food specimens. A total of 350 samples, fifty samples each from human, tap water, fish swamp, chicken meat, minced meat, raw milk, and hospital surface from the Menoufiya governorate, Egypt. P. aeruginosa was detected in 14.28%, including human (26%), tap water (18%), fish swamp (18%), chicken meat (12%), minced meat (10%), raw milk (16%), and hospital surface (0%). The results of testing of 50 P. aeruginosa isolates against sixteen antibiotics revealed a relatively high antibiotic resistance for Amoxicillin (100%), Erythromycin (98%), Cephradine (90%), Colistin (82%), Oxytetracyclin, (79%), Chloramphenicol (70%), Doxycycline (70%), and Kanamycin (62%) and high susceptibility for Imipenem (96%), Apramycin (94%), Amikacin (90%), Norfloxacin (78%), Sulphamethoxazol (86%), Enrofloxacin (64%), and Ofloxacin (60%). Furthermore, PCR was successfully amplified for the toxA, exoS, and oprL virulence genes at 396, 118, and 504 bp, respectively, as well as amplifying the ermB, pelA, blaTEM, and tetA resistance genes at 639, 786, 516, and 570 bp, respectively. The dendrogram investigation by ERIC-PCR of 10 clinical P. aeruginosa isolates revealed two main clusters and 10 different ERIC-PCR patterns. The presence of P. aeruginosa isolates in food may represent a potential public health concern, with the need for further epidemiological studies, as well as whole-genome sequencing and correlations of P. aeruginosa in water, food samples, and human infections.

With the rapid development of large language models (LLMs), an increasing number of applications leverage cloud-based LLM APIs to reduce usage costs. However, since cloud-based models' parameters and gradients are agnostic, users have to manually or use heuristic algorithms to adjust prompts for intervening LLM outputs, which requiring costly optimization procedures. In-context learning (ICL) has recently emerged as a promising paradigm that enables LLMs to adapt to new tasks using examples provided within the input, eliminating the need for parameter updates. Nevertheless, the advancement of ICL is often hindered by the lack of high-quality data, which is often sensitive and different to share. Federated learning (FL) offers a potential solution by enabling collaborative training of distributed LLMs while preserving data privacy. Despite this issues, previous FL approaches that incorporate ICL have struggled with severe straggler problems and challenges associated with heterogeneous non-identically data. To address these problems, we propose an asynchronous distributed bilevel tuning (AsynDBT) algorithm that optimizes both in-context learning samples and prompt fragments based on the feedback from the LLM, thereby enhancing downstream task performance. Benefiting from its distributed architecture, AsynDBT provides privacy protection and adaptability to heterogeneous computing environments. Furthermore, we present a theoretical analysis establishing the convergence guarantees of the proposed algorithm. Extensive experiments conducted on multiple benchmark datasets demonstrate the effectiveness and efficiency of AsynDBT.

Unmanned Aerial Vehicles (UAVs) are increasingly deployed in complex, uncertain, and adversarial environments, yet they remain vulnerable to actuator faults, environmental disturbances, and deliberate interferences such as jamming or spoofing. Conventional control approaches typically address these challenges independently, focusing on either fault-tolerant control (FTC), disturbance rejection, or counter-UAV defense. This paper presents an integrated anti-interference and fault-tolerant control framework that unifies three complementary modules: (i) residual-based fault detection and isolation, (ii) adaptive extended state observer (AESO)-based disturbance estimation and compensation, and (iii) counter-UAV evasive maneuver strategies. The framework is formulated using a nonlinear UAV dynamic model and rigorously analyzed under Lyapunov stability theory. Simulation results demonstrate significant performance improvements compared with baseline controllers. The proposed method reduced position and attitude deviations to below 0.05 m and 0.03 deg, respectively, under simultaneous actuator faults and wind disturbances. Fault and disturbance estimation errors remained below 0.05 and 0.03 units, respectively, ensuring timely control reconfiguration. Furthermore, the UAV achieved a 100% success rate in counter-drone evasive maneuvers while maintaining trajectory stability. These results confirm that the integrated design provides high resilience, rapid recovery, and reliable performance in fault-disturbed and adversarial conditions. By bridging FTC, adaptive disturbance rejection, and counter-UAV defense, the proposed framework advances the state of the art in resilient UAV control for civilian and defense applications.