Unlocking the potential of 3D printed microfluidics for mass spectrometry analysis using liquid infused surfaces

Abstract

Combining microfluidics with mass spectrometry (MS) analysis has great potential for enabling new analytical applications and simplifying existing MS workflows. The rapid development of 3D printing technology has enabled direct fabrication of microfluidic channels using consumer grade 3D printers, which holds great promise to facilitate the adoption of microfluidic devices by the MS community. However, photo polymerization-based 3D printed devices have an issue with chemical leeching, which can introduce contaminant molecules that may present as isobaric ions and/or severely suppress the ionization of target analytes when combined with MS analysis. Although extra cure and washing steps have alleviated the leeching issue, many such contaminant peaks can still show up in mass spectra. In this work, we report a simple surface modification strategy to isolate the chemical leachates from the channel solution thereby eliminating the contaminant peaks for MS analysis. The channel was prepared by fabricating a layer of polydimethylsiloxane graft followed by wetting the graft using silicone oil. The resulting liquid infused surface (LIS) showed significant reduction in contaminant peaks and improvement in the signal intensity of target analytes. The coating showed good stability after long-term usage (7 days) and long-term storage (∼6 months). Finally, the utility of the coating strategy was demonstrated by printing herringbone microfluidic mixers for studying fast reaction kinetics, which obtained comparable reaction rates to literature values. The effectiveness, simplicity, and stability of the present method will promote the adoption of 3D printed microdevices by the MS community.

Figure 1.

(a) Schematic of the preparation of a liquid infused surface on the surface of 3D printed microchannels. (b) Pictures for 3D printed straight microchannel. Scanning electron microscope (SEM) images of (c) uncoated channel surface, (10 KV and 1.00K) (d) channel surface coated with poly dimethyldimethoxysilane, (10 KV and 1.00K) and (e) channel surface with poly dimethyldimethoxysilane graft and silicone oil. (10 KV and 1.00K) The scale bars in the images represent 50 μm.

Figure 2.

Mass spectra of 1 μM clarithromycin in aqueous solution with different surface treatment conditions. (a) untreated 3D printed channel, (b) LIS coated 3D printed channel, (c) PTFE tubing, and (d) 3D printed channel washing at 50 μL/min with methanol for 30 min and then photocured for 3 hours.

Figure 3.

Mass spectra of 1 μM ubiquitin in aqueous solution containing 50 mM ammonium acetate with different surface treatment conditions. (a) untreated 3D printed channel, (b) LIS coated 3D printed channel. Mass spectra of 1 μM single strand DNA in aqueous solution with different surface treatment conditions in negative ion mode. (c) untreated 3D printed channel (the peaks shown in the spectrum are not from the DNA molecule), (d) LIS coated 3D printed channel.

Figure 4.

Stability study of the LIS surface using 1 μM clarithromycin in aqueous solution under different conditions. Mass spectrum of 1 μM clarithromycin upon passing through (a) LIS coated 3D printed channel after one day of 100 min continuous flow, (b) LIS coated 3D printed channel after 5 days of 100 min continuous flow, (c) LIS coated 3D printed channel after 7 days of 100 min continuous flow, (d) LIS coated 3D printed channel after storing at room temperature for ~6 months.

Figure 5.

(a) Design of 3D serpentine channel with two inlets and one outlet with a picture of the 3D device. (b) Schematic of the enzyme kinetics measurement experiment. (c) Progress curves for HRP catalysis reactions at different H₂O₂ concentrations. (d) Lineweaver-Burk plot of the reciprocal initial reaction rates obtained from (c) vs the reciprocal of the H₂O₂ concentration.

Figure 6.

Reaction kinetics between DCIP and L-AA. Mixing of water and fluorescein solution (a) without and (b) with herringbone structures. The scale bars in the images represent 200 μm. Mass spectrum for the reaction under a flow rate of 50 μL/min (c) without and (d) with herringbone structure. (e) Plot of ln(c/c₀) vs time (s) of the DCIP and L-AA reaction. (f) Zoomed-in range for the reaction time from 0.05 s to 0.22 s for deriving the pseudo first-order reaction rate constant. The error bars are the standard deviations of triplicate measurements.