Gustavus Adolphus College, Department of Chemistry, Saint Peter, MN, USA.
Department of Chemical and Biomolecular Engineering, University of Delaware, Newark, DE 19716, USA.
J Chromatogr A. 2018 Feb 9;1536:185-194. doi: 10.1016/j.chroma.2016.12.084. Epub 2016 Dec 31.
The ongoing movement in HPLC toward the use of small columns packed with small particles for high speed separations results in eluted peaks with very small volumetric variances. Avoiding degradation of separation performance under these conditions requires careful consideration of all sources of extra-column peak dispersion. Recent trends towards decreased diameters of connecting capillaries and increased flow rates for analytical-scale separations can result in Reynolds numbers that exceed 2000. This raises the possibility of a transition from laminar to turbulent flow, thereby resulting in a higher than expected pressure drop across the capillary at a given flow rate. In this study we collected pressure drop data as a function of flow rate under many conditions relevant to modern HPLC. The variables studied included capillary diameter (50-120μm) and length (100-550mm), acetonitrile/water composition (0-100%), and temperature (20-80°C). Most of the work involved stainless steel capillaries, but a subset of experiments involved fused silica. We then used the experimental data to train a model that enables prediction of pressure drops for all of the conditions studied. We find that a single global friction factor profile is sufficient to predict pressure drops as a function of flow rate that are in qualitative agreement with the experimental results. The quantitative accuracy of these predictions is generally quite good, with a mean prediction error of about 2% over the entire range of conditions studied. Predictions for some outlying capillaries are not as good, with errors as high as -40%. This variability is probably due mainly to capillary-to-capillary variability, especially in the wall roughness, which is difficult to characterize definitively. We believe the model described here will be very useful to practicing chromatographers for predicting the conditions under which turbulent flow might develop in their connecting capillaries, and the magnitude of the pressure drop increase over that expected if flow were exclusively laminar.
高效液相色谱法(HPLC)目前的发展趋势是使用小颗粒填充的小柱进行高速分离,这导致洗脱峰的体积变化非常小。在这些条件下避免分离性能下降需要仔细考虑所有柱外峰展宽的来源。最近,分析规模分离中连接毛细管直径的减小和流速的增加趋势可能导致雷诺数超过 2000。这增加了从层流向湍流过渡的可能性,从而导致在给定流速下通过毛细管的压降高于预期。在这项研究中,我们在许多与现代 HPLC 相关的条件下收集了作为流速函数的压降数据。研究的变量包括毛细管直径(50-120μm)和长度(100-550mm)、乙腈/水组成(0-100%)和温度(20-80°C)。大部分工作涉及不锈钢毛细管,但一部分实验涉及熔融石英。然后,我们使用实验数据训练了一个模型,该模型能够预测所有研究条件下的压降。我们发现,单个全局摩擦系数分布足以预测流速与压降的关系,与实验结果定性一致。这些预测的定量精度通常相当好,在整个研究条件范围内平均预测误差约为 2%。对于一些离群的毛细管,预测结果不是很好,误差高达-40%。这种可变性可能主要归因于毛细管之间的差异,特别是在壁粗糙度方面,这很难明确地进行表征。我们相信,这里描述的模型对于实践色谱学家来说将非常有用,可以预测在其连接毛细管中可能发生湍流的条件,以及与完全层流时相比,压降增加的幅度。