我們藉著直接監測粒子的位置變化，來研究似二維微粒電漿液體之局部結構規則度的記憶行為。在微觀尺度，我們可視液體為受隨機擾動之多體強耦合系統。由於觀測工具的缺乏及其廣大的自由度與複雜性，液體的微觀動力行為雖然極為多變、豐富，但顯為人知。近年來，微粒電漿液體的發展提供了一個便於觀測的實驗平台來探討這個問題。它是由帶大量電荷的微米粒子懸浮於電漿中所組成。粒子間強大的庫倫排斥力遠大於一般物質的分子作用力，使粒子的平均間距小於一毫米，因此可由一般的光學顯微 鏡連接CCD 直接觀測紀錄結構及運動狀態。再者，近年來理論物理發展「持續性」的概念，提供了研究記憶或穩定性的方法。我們利用此方法及微粒電漿系統，以實驗証實：在冷液體中，規則和不規則結構狀態的持續時間皆具有普適行為。意指，局部結構變動有non-Makrov 特性。接連的結構狀態(例如：規則-不規則-規則或不規則-規則-不規則)之持續時間傾向以長-短-長或短-長-短之組合出現。此關聯性可傳達至少兩個規則-不規則之週期。粒子間交互作用力傾向於同化周圍結構狀態，因此時間記憶存在於周圍粒子所組成的網路中。當熱擾動增加，時間記憶因而被洗去，局部結構持續時間變為以較無關聯性的stretched exponential 分布。 We investigate the temporal memory behavior of micro-structural order fluctuations in quasi-2D dusty plasma Coulomb liquids through directly monitoring particle positions at the microscopic level. Liquid at the discrete level can be treated as a strongly coupled many body system driven by stochastic noise. Very little is known about its rich fluctuating dynamical behaviors, because of the lack of proper direct observation tools and the complexity under the large number of degrees of freedom. Recently, the development of the dusty plasma liquid formed by the charged micro-meter particles suspended in low pressure discharges provides an inspiring experimental platform with proper scale for addressing this issue through direct optical visualization of the spatiotemporal evolutions of particle positions. Using the idea of persistence, a characterization of memory or stability, we experimentally demonstrate that the temporal persistent length of the ordered and the disordered micro-states both follow power law distribution for the cold liquid. It signi-fies the non-Markovian nature of the fluctuating structure. The correlation probability of the temporal persistent lengths for events in different cycle i and j shows: The order-disorder-order or a disorder-order-disorder sequence is more probable to have the alternating long-short-long or short-long-short combination. The correlation can be propagated to more than two cycles. The memory is carried by the surrounding network through mutual coupling which tends to make neighboring sites alike. Increasing thermal noise level deteriorates the memory and leads to the less correlated stretched exponential distribution of the persistent length.
