Gross Bnaya, Volotsenko Irina, Sallem Yuval, Yadid Nahala, Bonamassa Ivan, Havlin Shlomo, Frydman Aviad
Network Science Institute, Northeastern University, Boston, MA, USA.
Department of Physics, Northeastern University, Boston, MA, USA.
Nat Commun. 2025 Jul 1;16(1):5869. doi: 10.1038/s41467-025-61127-z.
Phase transitions are fundamental features of statistical physics. While the well-studied continuous phase transitions are known to be controlled by external global changes affecting the order parameter, the origin of abrupt transitions is not fully clear. Here we show that abrupt phase transitions may occur due to a unique internal random spatial cascading mechanism, arising from dependency interactions. We experimentally unveil the underlying mechanism of the abrupt transition in interdependent superconducting networks to be governed by a unique metastable state of a long-living resistance cascading plateau. This plateau is characterized by spontaneous cascading events that occur at random locations and last for thousands of seconds, followed by a sudden global phase shift of the system. The plateau time length changes with the system size and distance from criticality, obeying scaling laws with critical exponents. Furthermore, like epidemic spreading, these changes are characterized by a branching factor which equals exactly one at the critical point and deviates from one off criticality. Importantly, the branching factor provides an early warning for the closeness of critical catastrophic cascades yielding system collapse.
相变是统计物理学的基本特征。虽然人们熟知的连续相变是由影响序参量的外部全局变化所控制，但突变相变的起源尚不完全清楚。在此我们表明，突变相变可能源于一种独特的内部随机空间级联机制，该机制由依赖相互作用产生。我们通过实验揭示了相互依赖的超导网络中突变相变的潜在机制，其受一个长寿命电阻级联平台的独特亚稳态支配。这个平台的特征是在随机位置发生并持续数千秒的自发级联事件，随后系统会突然发生全局相移。平台时间长度随系统大小和与临界状态的距离而变化，遵循具有临界指数的标度律。此外，与流行病传播一样，这些变化的特征是一个分支因子，它在临界点恰好等于1，在临界状态之外偏离1。重要的是，分支因子为导致系统崩溃的临界灾难性级联的临近提供了早期预警。
