Two of the simulations discussed in a prior article (Hammond et al 2019 Nucl. Fusion 59 066035) were affected by a simulation glitch. We repeated the affected calculations and discuss them here. The overall conclusions are essentially unchanged, though the details are different. In particular, observations that we referred to as ‘concerted bursting’ were caused primarily by non-physical heating and cooling applied by the thermostat after most atoms’ velocities were deleted (for reasons that are not known for certain). The phenomenon of one bubble bursting and causing another nearby bubble to burst does exist, though its effects are much less spectacular in the absence of non-physical driving forces. The observation of an interconnected network of sub-surface cavities formed by burst bubbles is real, and the observation of holes on the surface 1–2 nm in diameter is also confirmed.

中文翻译：

---

更正和附录：大规模分子动力学模拟（2019 Nucl。Fusion 59 066035 [https://iopscience.iop.org/article/10.1088/1741-4326 / ab12f6]）

先前文章（Hammond et al 2019 Nucl.Fusion 59 066035）中讨论的两个仿真受到仿真故障的影响。我们重复了受影响的计算，并在此处进行了讨论。尽管细节不同，但总体结论基本上没有变化。特别是，我们称为“连续爆发”的观察主要是由于删除了大多数原子的速度后，恒温器进行了非物理加热和冷却（出于某些未知的原因）。确实存在一个气泡破裂并导致附近另一个气泡破裂的现象，尽管在没有非物理驱动力的情况下其影响远不那么明显。观察到由破裂的气泡形成的地下空腔相互连接的网络是真实的，