We performed molecular dynamics simulations to study how well some of the water models used in simulations describe shocked states. Water in our simulations was described using three different models. One was an often-used all-atom TIP4P/2005 model, while the other two were coarse-grained models used with the MARTINI force field: non-polarizable and polarizable MARTINI water. The all-atom model provided results in good agreement with Hugoniot curves (for data on pressure versus specific volume or, equivalently, on shock wave velocity versus “piston” velocity) describing shocked states in the whole range of pressures (up to 11 GPa) under study. If simulations of shocked states of water using coarse-grained models were performed for short time periods, we observed that data obtained for shocked states at low pressure were fairly accurate compared to experimental Hugoniot curves. Polarizable MARTINI water still provided a good description of Hugoniot curves for pressures up to 11 GPa, while the results for the non-polarizable MARTINI water substantially deviated from the Hugoniot curves. We also calculated the temperature of the Hugoniot states and observed that for TIP4P/2005 water, they were consistent with those from theoretical calculations, while both coarse-grained models predicted much higher temperatures. These high temperatures for MARTINI water can be explained by the loss of degrees of freedom due to coarse-graining procedure.

中文翻译：

---

Hugoniot激波状态下的粗粒水模型和全原子水模型的比较计算研究

我们进行了分子动力学模拟，以研究模拟中使用的某些水模型对冲击状态的描述程度。我们在模拟中使用三种不同的模型描述了水。一个是经常使用的全原子TIP4P / 2005模型，另外两个是用于MARTINI力场的粗粒度模型：不可极化和可极化的MARTINI水。全原子模型提供的结果与描述整个压力范围（最高11 GPa）中的冲击状态的Hugoniot曲线（用于压力与比体积的数据，或者等效地，关于冲击波速度与“活塞”速度的数据）非常吻合。在研究中。如果在短时间内使用粗粒度模型对水的冲击状态进行了模拟，我们观察到，与实验Hugoniot曲线相比，在低压冲击状态下获得的数据相当准确。可极化的MARTINI水在高达11 GPa的压力下仍能很好地描述Hugoniot曲线，而不可极化的MARTINI水的结果则大大偏离了Hugoniot曲线。我们还计算了Hugoniot态的温度，并观察到对于TIP4P / 2005水，它们与理论计算的结果一致，而两个粗粒度模型都预测温度要高得多。MARTINI水的这些高温可以通过粗粒度过程导致的自由度损失来解释。而不可极化的MARTINI水的结果则大大偏离了Hugoniot曲线。我们还计算了Hugoniot态的温度，并观察到对于TIP4P / 2005水，它们与理论计算的结果一致，而两个粗粒度模型都预测温度要高得多。MARTINI水的这些高温可以通过粗粒度过程导致的自由度损失来解释。而不可极化的MARTINI水的结果则大大偏离了Hugoniot曲线。我们还计算了Hugoniot态的温度，并观察到对于TIP4P / 2005水，它们与理论计算的结果一致，而两个粗粒度模型都预测温度要高得多。MARTINI水的这些高温可以通过粗粒度过程导致的自由度损失来解释。