We report experiments characterizing the stratified and filamentary structures formed in the dense core of nanosecond electrical explosion of aluminum wires to understand the physical scenario of electrothermal instability. Direct experimental observations for stratification and filamentation instabilities, as well as the coexistence state of azimuthal strata and vertical filament in the dense plasma column, are presented. The wire core exhibits remarkable different patterns of instability with the decreasing wire length. The shadowgram of shorter wires demonstrates that the instability is transformed from stratified structures to filamentary structures. According to a radial magnetohydrodynamic computation, the wire enters a phase state of negative temperature dependence of resistivity before voltage breakdown. However, filamentary structures are only observed in exploding wires of 1 cm and 0.5 cm in length. The analyses based on experimental and computational results indicate that the increase in internal energy determines the manifestation of instability in the dense core. Filamentation instability occurs when the total energy input is no less than 1.5 times the vaporization energy at the moment of voltage breakdown. The lower limit of energy deposition ensures that the increase in internal energy covers vaporization energy.
The densities of dimethyl carbonate, cyclohexane, and their mixtures were measured for nine compositions at five different temperatures varying from (293.15 to 313.15) K and over the pressure range of (0.1 to 40) MPa. The densities of pure substances and their mixtures at atmospheric pressure were measured with a vibrating-tube densimeter. The densities at elevated pressures were measured with a high-pressure apparatus and a precise analytical balance. The molar volumes Vm, excess molar volumes VmE, isothermal compressibilities κ, and isobaric expansivities α were derived from the experimental densities.
Vapor−liquid equilibrium data for the carbon monoxide + dimethyl carbonate, oxygen + dimethyl carbonate, and dimethyl ether + dimethyl carbonate systems were measured at 293 K, 313 K, 353 K, and 373 K and at elevated pressures up to 12.00 MPa. The measurements were carried out in a cylindrical autoclave with a moveable piston and an observation window. The experimental data were correlated using the Peng−Robinson−Stryjek−Vera equation of state (EOS) with the two-parameter van der Waals II mixing rule.
The wide application of KNN-based lead-free piezoelectric ceramics is constrained by the contradictory relationship between its mechanical quality factor and piezoelectric constant. From an application point of view, searching for chemical composition with enhanced piezoelectric constant (d33) and mechanical quality factor (Qm) is one of the key points of KNN-based ceramics. In this work, KNN-based ceramics with enhanced d33 and high Qm values were obtained by the solid solution method via a donor-acceptor codoping strategy. The donor dopant Ho3+ enhanced d33 values by refining the domain size, while the acceptor dopant (Cu1/3Nb2/3)4+ improved Qm by the formation of defect dipoles. The composition (KNN-5Ho-4CN) exhibits optimal integrated performances, of which d33, Qm, and TC values are 120 pC/N, 850, and 392 °C, respectively. Moreover, the temperature coefficient of resonant frequency (TCF = -429 ppm/K) indicates that KNN-5Ho-4CN ceramic has good temperature stability. This work provides a new insight for developing KNN-based ceramics with enhanced d33 and high Qm.
Recent advances in Large Language Models (LLMs) have achieved remarkable breakthroughs in understanding and responding to user intents. However, their performance lag behind general use cases in some expertise domains, such as Chinese medicine. Existing efforts to incorporate Chinese medicine into LLMs rely on Supervised Fine-Tuning (SFT) with single-turn and distilled dialogue data. These models lack the ability for doctor-like proactive inquiry and multi-turn comprehension and cannot align responses with experts' intentions. In this work, we introduce Zhongjing, the first Chinese medical LLaMA-based LLM that implements an entire training pipeline from continuous pre-training, SFT, to Reinforcement Learning from Human Feedback (RLHF). Additionally, we construct a Chinese multi-turn medical dialogue dataset of 70,000 authentic doctor-patient dialogues, CMtMedQA, which significantly enhances the model's capability for complex dialogue and proactive inquiry initiation. We also define a refined annotation rule and evaluation criteria given the unique characteristics of the biomedical domain. Extensive experimental results show that Zhongjing outperforms baselines in various capacities and matches the performance of ChatGPT in some abilities, despite the 100x parameters. Ablation studies also demonstrate the contributions of each component: pre-training enhances medical knowledge, and RLHF further improves instruction-following ability and safety. Our code, datasets, and models are available at https://github.com/SupritYoung/Zhongjing.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)