In this work, we study hafnium zirconate (HZO) based ferroelectric (FE) capacitors (FeCAP) intended for random access memory and compute-in-memory (CiM) applications. We show that there exists a discrepancy between the coercive voltages (V_C) extracted from a small-signal capacitance–voltage (C-V) measurement ({V_C}_CV) and dynamic polarization-voltage (P-V) measurement ({V_C}_PV): |{V_C}_CV| < |{V_C}_PV|. Utilizing the Nucleation Limited Switching (NLS) framework, we attribute the origin of this discrepancy to the difference in voltage–time dynamics experienced by the FE during the C-V and P-V measurements, resulting in a higher switching probability at the same applied voltage for a longer time during the C-V measurement: |{V_C}_CV| < |{V_C}_PV|. Therefore, experimentally obtained V_C values do not solely reflect an intrinsic material property of the FE as they depend on the voltage–time dynamics of the measurement.

在这项工作中，我们研究了用于随机存取存储器和内存计算 (CiM) 应用的基于锆酸铪 (HZO) 的铁电 (FE) 电容器 (FeCAP)。我们表明，从小信号电容电压 (C-V) 测量中提取的矫顽电压 (V _C ) ({V _C } _CV } _PV ):|{V _C } _CV |<|{V _C } _PV |。利用成核限制开关 (NLS) 框架，我们将这种差异的根源归因于 FE 在 C-V 和 P-V 测量期间经历的电压-时间动态差异，从而导致在相同的施加电压下，在更长的时间内具有更高的开关概率。 C-V 测量期间的时间：|{V _C } _CV | <|{V _C } _PV |。因此，实验获得的 V _C 值并不仅仅反映了 FE 的固有材料属性，因为它们取决于测量的电压-时间动态。