柴达木盆地西部地区的南翼山油田卤水锂资源丰富, 共存元素复杂, 传统的火焰原子吸收光谱法在测定该体系中锂含量时, 基体干扰严重。 而且油田卤水在各蒸发浓缩阶段的离子浓度变化幅度较大, 一般的基体匹配法繁琐不便。 研究采用Design of Experiments(DOE) 实验设计, 通过干扰元素的显著性分析、 消电离剂的选择及干扰模型的建立, 对传统的火焰原子吸收测定锂的分析方法进行了优化。 运用部分因子实验设计研究了南翼山油田卤水中钙、 锶、 钾、 钠、 镁、 铵、 硼等主要共存离子及离子间交互作用对锂分析的影响规律, 考察了各干扰元素的显著性程度。 研究表明, 钙、 锶、 镁、 钠以及钙*硼在锂测定分析过程中存在显著干扰, 其显著性从大到小排序为钙>锶>镁>钙*硼>钠。 针对钙、 锶、 镁、 钙*硼干扰, 可加入消电离剂进行沉淀消除, 通过比较分析, 草酸钾作为消电离剂加入的除干扰效果最佳, 锂测定相对误差从-20.75%降低至-12.15%; 对于样品中的钠干扰, 运用响应曲面实验设计, 拟合方程建立干扰模型, 通过方差分析及拟合度分析, 回归方程各项p值均为0.000, 方程的R-sq, R-sq（调整）与R-sq（预测）分别为99.96%, 99.96%以及99.95%, 表明所建立的模型及方程各项显著, 且回归方程拟合度较好。 实验以各蒸发浓缩阶段的南翼山实际卤水与西藏龙木错实际卤水为样品, 对消电离剂和干扰模型进行实验验证。 结果显示, 加入草酸钾消电离剂后, 锂加标回收率在89.30%~98.60%之间; 使用钠干扰模型校正后, 锂加标回收率可提升至98.88%~101.40%, 表明锂测定的准确度得到大幅提高。 该方法不仅适用于南翼山油田卤水分离的整个过程, 也同样适用于其他盐湖卤水, 可以为盐湖企业锂元素的准确测定提供技术支持。

The oil-field water in Nanyi Mountain, located in the west of Qaidam Basin of Qinghai-Tibet Plateau, is rich in lithium, calcium and other components. Thus, it was difficult to determine the exact content of lithium due to the matrix interference problems using traditional flame atomic absorption spectrometry, especially in the process of different evaporation stages. Though the general matrix matching method can resolve this issue, it was tedious and time-consuming. In this work, DOE was applied to study the influences of main coexisting ions, and the flame atomic absorption spectrometry was optimized based on the analysis of main coexisting elements, releasing agent selection as well as the establishment of the interference model. The results indicated that Ca2+, Sr2+, Mg2+, Na+ and Ca2+*B had the significant interferences on lithium determination, and the sequencing was Ca2+＞Sr2+＞Mg2+＞Ca2+*B＞Na+. The comparative study suggested that potassium oxalate showed excellent performances as the releasing agent to eliminate the influences of Ca2+, Sr2+, Mg2+, Na+ and Ca2+*B. And The relative error of lithium determination was reduced from -20.75% to -12.15% Meanwhile, results showed the response surface experimental design was beneficial to eliminate the influences of Na+, which was proved by the variance analysis and fitting degree analysis. In order to verify the conclusion further, different kinds of oil-field water were used. The results manifested the standard recovery of lithium was 89.30%~98.60% after adding potassium oxalate as the releasing agent, and the standard recovery increased to 98.88%~101.40% after the correction using the sodium interference model. All of the data indicated that the accuracy of lithium determination was improved greatly. The optimized method was not only applicable to the whole separation process of the oil-field water in Nanyi Mountain, but also applicable to other brines. And it proposed a guideline for the accurate determination of lithium for enterprises.