三维光子输运在分相光催化气凝胶中的研究：双连续形态如何控制速率常数
Three-dimensional photon transport in spinodal photocatalytic aerogels: how bicontinuous morphology controls kinetic rate constants
AI总结：
研究通过结合3D分相结构与GPU MC方法，揭示双连续形态对光催化速率常数的影响，发现光子通道效应导致固相接收到的光子比体积平均多50%-70%，并指出有效介质理论无法完全解释分相结构的内在差异。
AI中文摘要：
基于硅气凝胶中的金红石TiO2的多孔整体光催化剂在空气净化中具有潜力。其双连续分相结构提供高比表面积和强光散射。然而，提取内在动力学速率需要准确的光学模型。当前方法将复杂的3D孔隙网络简化为均质1D板，这种近似在分相几何中误差未知。我们结合Cahn-Hilliard模拟的3D分相掩膜与GPU MC光子输运方法，量化了这一误差。我们引入了一种固相荧光估计器，考虑催化位点分布，将其与体积平均和扩散近似进行比较。在孔隙率0.70时，固相接收到的光子比体积平均多50%，在0.90时增加到70%。这种优先照明源于通过孔隙通道的准弹道路径，称为光子通道效应。提取的速率描述符在3D MC和扩散模型之间差异达34%。均质控制显示，总73%差异中约50%是双连续结构固有的，无法通过有效介质理论修正。这些结果为此类光催化剂的动能提取提供了首次定量修正，并建立了连接合成粗化、孔隙尺寸和光效率的设计规则。
英文摘要：
Anatase TiO2 dispersed in a silica aerogel matrix forms porous monolithic photocatalysts promising for indoor air purification and VOC photo-degradation. Their bicontinuous spinodal architecture provides high surface area, structural integrity, and intense UV multiple scattering. Extracting kinetic descriptors from apparent rate constants requires an optical model for the photon fluence, and current practice replaces the 3D pore network by a homogeneous 1D slab with effective optical coefficients -- an approximation never quantified for spinodal geometries. We combine realistic 3D spinodal pore masks from Cahn-Hilliard spectral integration with 3D GPU-accelerated Monte Carlo (MC) photon-migration simulations, introducing the solid-phase-weighted estimator $Φ_{3D}$ and comparing it with the conventional volume average $Φ_{1D}$ and the diffusion-approximation (DA) prediction $Φ_{DA}$. The ratio $Φ_{3D}/Φ_{1D}$ equals 1.515 at porosity $ϕ=0.70$ and 1.774 at $ϕ=0.90$: the solid phase is preferentially illuminated, traced to optical phase contrast. Since pores neither absorb nor scatter, photons cross quasi-ballistically, and the void phase sustains only $r=0.514$ of the solid-phase fluence, diluting the volume average with dark voxels. This ratio is porosity-independent, so the concentration ratio obeys the exact relation $1/[ϕr+(1-ϕ)]$, reproducing both simulated values. The kinetic descriptor $k_{surf}\cdot K_{ads}$ is underestimated by 20% via the DA route and overestimated by 51.5% via the volume-averaged route. Homogeneous-slab MC simulations at matched optical thickness show that 84% of the fluence-weighted RMS discrepancy between MC and DA is intrinsic to the bicontinuous architecture and beyond any effective-medium DA. These results give a quantitative correction for kinetic extraction and design rules linking porosity, pore size and photon utilisation.
