A new strategy to eliminate sample mixing during in-tube solid phase microextraction

• Sample mixing may complicate separation or detection for some sample types.
• The degree of sample mixing during IT-SPME was evaluated.
• A switching valve and an additional LC pump were added to eliminate the problem.
• Three groups of compounds were studied to prove the effectiveness of the developed method for washing mixing.

During in-tube solid phase microextraction, sample mixing with mobile phase contained in the autosampler tubing during extraction may result in some amount of sample becoming entrained in the mobile phase rather than returning to the sample vial or being directed to waste after extraction. In cases where target analytes have relatively low affinity for the sorbent on the wall of the capillary, mixing can impact data quality. Where the sample contains components that may interfere with either the separation (e.g. proteins) or detection (e.g. ions with MS detection), additional difficulties can arise. In the current research, the magnitude of the sample mixing effect was illustrated by analyzing ranitidine and a series of polycyclic aromatic hydrocarbons (PAH). The sample volume equivalent of mixing was calculated as 37 μL for ranitidine and 20 μL for PAHs using the same inner diameter of capillary. To address this issue, a novel approach involving adding a switching valve located between the metering pump and the capillary was developed. Capillary flush conditions, draw/eject speed and extraction time were optimized for ranitidine with the result that in the final method, no mixing of sample with mobile phase was apparent in the detected amounts. To provide information on a compound class with intermediate polarity, two β-blockers were also extracted using the optimized washing conditions respectively. The results indicated that the issue of sample mixing had been resolved for these as well. Finally, in-tube SPME calibration of these three analyte classes was shown to be highly linear, providing further indication that sample mixing was not impacting data quality. Available literature on the subject was surveyed, and a discussion on the rational selection of conditions to guide method development was also provided.