Fluid characterization by resonant nanomechanical sensing

Microfluidic technologies allow handling and characterizing liquid samples on the micro- to picoliter scale. Thereby, the viscosity and mass density are key properties of such samples, because they characterize their flow behavior. The viscosity of a liquid indicates its resistance to flow, whereas the density quantifies the mass per volume. Molecular transformations, such as chemical polymerization, protein folding and aggregation, or nucleic acid hybridization, influence both properties. Therefore, measuring them is fundamental for basic research, quality control, and process monitoring. Since many, especially biological, samples are only available in small quantities and/or expensive, reducing sample consumption is essential. Furthermore, the acquisition time of viscosity measurements nowadays is on the order of minutes, limiting the characterization of large numbers of samples. Hence, increasing the time resolution and the throughput is another significant requirement.
It was early noticed that the dynamics of nanomechanical resonators are strongly influenced by the surrounding fluid. This effect can be utilized to measure the fluid properties, specifically the viscosity and mass density. In this thesis, resonant nanomechanical cantilevers were, therefore, employed with focus on the application of higher modes of vibration.
First, a suitable method to excite and detect the strongly damped cantilever resonances encountered in liquid was realized: Photothermal excitation uses an intensity-modulated laser to induce cantilever vibration. Its direct and local energy transfer avoids distortions arising in prevalent excitation methods, such as piezo-acoustic excitation, and results in spurious-free resonance spectra. To detect the nanometer vibrations of the cantilevers, a second laser was used in an optical beam deflection configuration. Such optical excitation/detection method is accurate and robust, however, it is only suitable for transparent liquids. Technical details about the developed setup are provided in the appendix of this thesis. Due to the small dimensions of the microfluidic channel containing the cantilever sensors, the influence of proximate surfaces was investigated. Placing a vibrating cantilever below a critical distance to a surface induces squeeze-film damping. The magnitude and range of this undesirable effect on higher mode vibrations was characterized and incorporated in the fluid channel design. The above findings are generally applicable to atomic force microscopy and nanomechanical sensing in liquid.
Next, the ability of the sensor to measure viscosity and mass density of liquids was assessed. Dynamic properties of the cantilever resonator were derived from resonance spectra and converted into the surrounding liquid properties, using adapted hydrodynamic models. Multiple modes of vibration covered a broad frequency range in the order of kHz to MHz. A stringent temperature control was implemented, due to the high temperature dependency of the measured parameters. To investigate time-resolved processes, free-radical polymerization reactions were tracked and characterized. The shear-thinning behavior of the polymer solutions, i.e., the non-Newtonian effect of decreasing viscosity with increasing frequency, was resolved by the instrument. The time to characterize a 5 µL sample was on the order of 1 min.
Finally, the setup was optimized for automated high-throughput screening of microliter sample droplets. The droplets were generated by an automated sampler and separated by fluorinated oil. To achieve the required time resolution, a higher vibrational mode was tracked using two phase-locked loop demodulators. This allowed to derive the viscosity and mass density of the liquid surrounding the resonator with a temporal resolution of about 1 ms. The instrument was able to detect ~1 µL droplets at a rate on the order of 1 s per droplet.
The developed viscosity and mass density sensor opens several possibilities. We recently initiated the study of stimulus-responsive polymers for glucose sensing and the unfolding behavior of proteins. This, by solely measuring changes in viscosity after introducing the analyte or inducing denaturation. Future work could involve monitoring of RNA hybridization and protein aggregation into fibrils.