Ferrihydrite is one of the most important iron-containing minerals on Earth. Yet determination of its atomic-scale structure has been frustrated by its intrinsically poor crystallinity. The key difficulty is that physically-different models can appear consistent with the same experimental data. Using X-ray total scattering and a nancomposite reverse Monte Carlo approach, we evaluate the two principal contending models—one a multi-phase system without tetrahedral iron(III), and the other a single phase with tetrahedral iron(III). Our methodology is unique in considering explicitly the complex nanocomposite structure the material adopts: namely, crystalline domains embedded in a poorly-ordered matrix. The multi-phase model requires unphysical structural rearrangements to fit the data, whereas the single-phase model accounts for the data straightforwardly. Hence the latter provides the more accurate description of the short- and intermediate-range order of ferrihydrite. We discuss how this approach might allow experiment-driven (in)validation of complex models for important nanostructured phases beyond ferrihydrite.

水铁矿是地球上最重要的含铁矿物之一。然而，其原子级结构的确定因其本质上较差的结晶度而受挫。关键困难在于物理上不同的模型可能看起来与相同的实验数据一致。使用 X 射线全散射和纳米复合材料反向蒙特卡罗方法，我们评估了两个主要的竞争模型——一个是没有四面体铁 (III) 的多相系统，另一个是具有四面体铁 (III) 的单相系统。我们的方法在明确考虑材料采用的复杂纳米复合结构方面是独一无二的：即嵌入在有序矩阵中的晶体域。多相模型需要非物理结构重排来拟合数据，而单相模型直接解释了数据。因此，后者更准确地描述了水铁矿的短程和中程有序。我们讨论了这种方法如何允许对水铁矿以外的重要纳米结构相的复杂模型进行实验驱动（在）验证。