Addressing the Mystery of Sample Loss

By September 24, 2020

Many years ago, I was hired by a biopharma company to develop analytical methods for therapeutic oligonucleotides. When a new colleague warned me about sample loss on LC columns I looked at him with skepticism. “Yes,” he insisted, “part of your DNA sample may be irreversibly bound to the column and you will never see it again!” I admit, I thought that was a quirky idea. Afterall, what goes up, must come down. What is injected on a column must elute. Right?


As I gained more experience with oligonucleotides I learned that the first several injections of nucleic acids on a new column showed an appreciable sample loss and characteristic peak tailing. It always took several injections for this problem to go away and achieve a “passivation” of the column.

Years later I experienced a similar problem with phosphopeptides. This time my task was LC-MS analysis of small amounts of protein digests. I quickly realized that singly phosphorylated peptides would elute from the column just fine, while doubly phosphorylated peptides had nasty tailing, and triply phosphorylated species eluted as a hump on my baseline, that I can identify only thanks to my mass spectrometer. Phosphopeptides with four or more phosphorylated species did not give me recognizable peaks and were impossible to quantify, unless a few creative tricks were applied (e.g. basic mobile phase pH or EDTA in the sample). I realized that non-specific sample adsorption to LC hardware is much more common than I expected.

Many users have never experienced the problem of sample loss due to the metallic surfaces of the LC columns. Like me, they mostly practice analysis of neutral small molecules, which behave chromatographically well. Even experienced analytical chemists tend to underestimate the problem of compounds “sticking” to metal oxide surfaces. They may even think it is a quirky idea.

Well, those who analyze sialylated glycans, tricarboxylic organic acids, phosphopeptides or basically any acidic analytes, know better. I encourage all analytical scientists to follow this new blog series. Here, my colleagues and I will discuss the solution to this common analytical challenge. There are several tricks and exciting improvements in LC technology that you may want to learn about. So please keep reading, it will save you time and trouble in your future analytical projects!