Adsorption of uranium onto goethite is an important partitioning process that controls uranium mobility in subsurface environments, for which many different surface complexation models (SCMs) have been developed. While individual models can fit the data for which they are parameterized, many perform poorly when compared with experimental data covering a broader range of conditions. There is an imperative need to quantitatively evaluate the variations in the models and to develop a more robust model that can be used with more confidence across the wide range of conditions. We conducted an intercomparison and refinement of the SCMs based on a metadata analysis. By seeking the globally best fit to a composite dataset with wide ranges of pH, solid/sorbate ratios, and carbonate concentrations, we developed a series of models with different levels of complexity following a systematic roadmap. The goethite–uranyl–carbonate ternary surface complexes were required in every model. For the spectroscopically informed models, a triple-plane model was found to provide the best fit, but the performance of the double-layer model with bidentate goethite–uranyl and goethite–uranyl–carbonate complexes was also comparable. Nevertheless, the models that ignore the bidentate feature of uranyl surface complexation consistently performed poorly. The goodness of fitting for the models that ignore adsorption of carbonate and the charge distributions was not significantly compromised compared with that of their counterparts that considered those. This approach of model development for a large and varied dataset improved our understanding of U(VI)–goethite surface reactions and can lead to a path for generating a single set of reactions and equilibrium constants for including U(VI) adsorption onto goethite in reactive transport models.

中文翻译：

---

基于元数据分析的 U(VI) 吸附在针铁矿上的表面络合模型的比对和细化

铀在针铁矿上的吸附是控制铀在地下环境中迁移的重要分配过程，为此开发了许多不同的表面络合模型 (SCM)。虽然单个模型可以拟合参数化的数据，但与涵盖更广泛条件的实验数据相比，许多模型表现不佳。迫切需要对模型中的变化进行定量评估，并开发一个更可靠的模型，该模型可以在广泛的条件下更有信心地使用。我们基于元数据分析对 SCM 进行了比对和改进。通过寻找具有广泛 pH、固体/山梨酸酯比率和碳酸盐浓度范围的复合数据集的全局最佳拟合，我们按照系统路线图开发了一系列具有不同复杂程度的模型。每个模型都需要针铁矿-铀酰-碳酸盐三元表面复合物。对于光谱信息模型，发现三平面模型提供最佳拟合，但具有双齿针铁矿 - 铀酰和针铁矿 - 铀酰 - 碳酸盐复合物的双层模型的性能也相当。然而，忽略铀酰表面络合双齿特征的模型始终表现不佳。与考虑这些的对应模型相比，忽略碳酸盐吸附和电荷分布的模型的拟合优度没有显着降低。