论文标题
部分可观测时空混沌系统的无模型预测
Competing oxygen evolution reaction mechanisms revealed by high-speed compressive Raman imaging
论文作者
论文摘要
储层计算是预测湍流的有力工具,其简单的架构具有处理大型系统的计算效率。然而,其实现通常需要完整的状态向量测量和系统非线性知识。我们使用非线性投影函数将系统测量扩展到高维空间,然后将其输入到储层中以获得预测。我们展示了这种储层计算网络在时空混沌系统上的应用,该系统模拟了湍流的若干特征。我们表明,使用径向基函数作为非线性投影器,即使只有部分观测并且不知道控制方程,也能稳健地捕捉复杂的系统非线性。最后,我们表明,当测量稀疏、不完整且带有噪声,甚至控制方程变得不准确时,我们的网络仍然可以产生相当准确的预测,从而为实际湍流系统的无模型预测铺平了道路。
Transition metal oxides are state-of-the-art materials for catalysing the oxygen evolution reaction (OER), whose slow kinetics currently limit the efficiency of water electrolysis. However, microscale physicochemical heterogeneity between particles, dynamic reactions both in the bulk and at the surface, and an interplay between particle reactivity and electrolyte makes probing the OER challenging. Here, we overcome these limitations by applying state-of-the-art compressive Raman imaging to uncover competing bias-dependent mechanisms for the OER in a solid electrocatalyst, α-Li2IrO3. By spatially and temporally tracking changes in the in- and out-of-plane Ir-O stretching modes - identified by density functional theory calculations - we follow catalytic activation and charge accumulation following ion exchange under a variety of electrolytes, particle compositions and cycling conditions. We extract velocities of phase fronts and demonstrate that at low overpotentials oxygen is evolved by the combination of an electrochemical-chemical mechanism and a classical electrocatalytic adsorbate mechanism, whereas at high overpotentials only the latter occurs. These results provide strategies to promote mechanisms for enhanced OER performances, and highlight the power of compressive Raman imaging for low-cost, chemically specific tracking of microscale reaction dynamics in a broad range of systems where ion and electron exchange can be coupled to structural changes, i.e. catalysts, battery materials, memristors, etc.