Detailed plots of diffusion coefficients, Dm, versus % methanol in CO2 and pressure, P, are presented. Using sub-2 µm particle columns with small ID connector tubing make for large pressure drops, ∆P, across both the system and the column. This creates low average Dm, and large gradients in Dm, slowing the chromatography. The ∆P with 75 µm ID tubing varies non-linearly with flow creating U shaped plots of reduced plate height versus flow rate. Plumbing does but should never change h versus F plots. Using larger ID tubes to estimate the extra-column variance $$\sigma_{e - c}^{2}$$ of smaller tubes, by keeping the aspect ratio (L/r) constant, assumes the same constant Dm in both tubes, which is not true in SFC. Contrary to prediction, calculated tube plate count, N, indicates connector tubing is more dispersive in SFC compared to high performance liquid chromatography (HPLC). Current supercritical fluid chromatographs (SFC) have up to 20 × higher variance than appropriate for use with 2.1 and 3 mm columns ≤ 100 mm long, packed with sub-2 µm particles. Lower volume UV detector flow cells are needed for use with sub-2 µm particles. The modular nature of instruments makes it difficult to further minimize tube lengths while retaining column temperature control. Very small ID tubes, like 75 µm should probably be avoided. It remains difficult to characterize and minimize $$\sigma_{e - c}^{2}$$ .
Read full abstract