瞬态阶跃涡流致热效应:深纳米集成电路动态功耗的电磁本源
Description
本文基于麦克斯韦方程组与非平衡态热力学,提出瞬态阶跃涡流理论(SECT),揭示5纳米及以下工艺节点中动态功耗的电磁熵增本源。逻辑跳变沿激发的电磁场阶跃瞬变在互连中感生阶跃涡流,构成以涡旋电场为广义力、阶跃涡流为广义流的不可逆熵产生过程。本文独立导出电磁熵产生率的微观表达式,证明其与焦耳热宏观自洽,并推导出动态功耗的通用标度律 P ∝ f·L·I²·Γ。模拟芯片的千倍能效源于连续信号规避了阶跃瞬变,使系统维持在低熵轨道。该理论将传统开关电容模型重新解释为准静态极限下的投影,为后摩尔时代低功耗设计提供了第一性原理基础。\n\nThis preprint proposes the Step Eddy Current Theory (SECT) based on Maxwell's equations and non-equilibrium thermodynamics, revealing the electromagnetic entropy production origin of dynamic power consumption in 5nm and beyond. The step transient of electromagnetic fields induced by logic switching generates step eddy currents, forming an irreversible entropy production process with vortex electric field as generalized force and step eddy current as generalized flow. The microscopic expression of electromagnetic entropy production rate is independently derived and shown to be macroscopically self-consistent with Joule heat. A universal scaling law P ∝ f·L·I²·Γ is obtained. The thousand-fold higher energy efficiency of analog computing arises from its continuous signals that avoid step transients, keeping the system on a low-entropy trajectory. The traditional switch-capacitor model is reinterpreted as a quasi-static projection of SECT, providing a first-principles foundation for post-Moore low-power design.
Files
Transient_Step_Eddy_Current_Preprint_v3.0.pdf
Files
(1.0 MB)
| Name | Size | |
|---|---|---|
|
md5:941f922042b3c3dcca845b234703a381
|
435 Bytes | Download |
|
md5:f16ecd71c58d53998e5c880bf93ccb57
|
1.2 kB | Preview Download |
|
md5:3f241752cf97c689b555a82ec6445100
|
897 Bytes | Preview Download |
|
md5:0b229a60aa04b3d380a39345376c19bf
|
3.7 kB | Preview Download |
|
md5:a7022c44f446bafb3d9f210cf30faede
|
2.2 kB | Preview Download |
|
md5:02d74cd7d9f3998c6eb54de5d58ffb2b
|
1.0 MB | Preview Download |
Additional details
Additional titles
- Translated title (Jinyu Chinese)
- Transient Step Eddy Current Thermal Effect: The Electromagnetic Origin of Dynamic Power Consumption in Deep-Nanometer Integrated Circuits
Dates
- Updated
-
2026-05-12Preprint publication date
Software
- Repository URL
- https://doi.org/10.5281/zenodo.17564365