Logo: to the web site of Uppsala University

In this work, we have studied how the choice of the continuous phase oils affects droplet-internal acoustic manipulation and mapped the acoustic properties of the selected oils to evaluate their compatibility with droplet acoustofluidic methods. The selection of continuous phase included hydrocarbon, fluorinated, and silicone oils. To map the acoustic properties of the oils, we measured their speed of sound and density. We then studied the acoustic performance of each oil for droplet-internal manipulation through experiments and finite-element simulations (COMSOL Multiphysics^®, COMSOL, Stockholm, Sweden). From our results, we conclude that a match between the speed of sound of the continuous and dispersed phases is strongly correlated to the generation of a strong and uniform acoustic field inside the droplet. We demonstrate that conventionally favoured fluorinated oils in droplet microfluidics are no longer the best choice when considering droplet-internal acoustic focusing. Instead, hydrocarbon oils, especially linseed oil, are most suitable for this specific application as they generate stable and monodisperse droplets and bear the most resemblance to water in terms of acoustic properties. We believe this collection of data will serve the acoustofluidics community by providing results that aid in the selection of continuous phase in future droplet acoustofluidic studies and data for performing acoustofluidic simulations.

Droplet-based microfluidics involves the generation of discrete volumes in a continuous and immiscible phase. It is an attractive technique for biological, especially single-cell level, assays due to its ability to encapsulate particles and cells within individual reaction vessels at high throughput. Bulk-wave-acoustophoresis is a contactless and label free method to spatially localize particles. It has emerged as a suitable technique to concentrate and separate droplet contents on-chip. In traditional droplet microfluidics for biological applications fluorinated oils, e.g. Novec HFE-7500, are favored due to insolubility for most organic compounds, which prevents cross-contamination between droplets and their high gas permeability. However, a strong and uniform field cannot be formed inside aqueous droplets generated in HFE-7500 due to a mismatch in the acoustic properties of the two phases. Previously, the individual contributions of the acoustic properties, i.e. density, compressibility, and speed of sound, to the focusing quality were not fully investigated. We demonstrate in this work, that the speed of sound dominates the focusing performance. It is shown experimentally and numerically that matching the speed of sound between continuous and discrete phase is paramount to high quality acoustic focusing. To this end the acoustic focusing of polystyrene particles inside aqueous droplets immersed in different continuous phase fluids (hydrocarbon oil, fluorinated oil and silicone oil) is investigated. We demonstrate that a match in the speed of sound between the two phases results in higher droplet-internal particle focusing quality.

Droplet-based microfluidics involves the generation of discrete volumes of a dispersed phase (often water) in a continuous and immiscible phase (oil). Bulk acoustic wave acoustophoresis can be combined with droplet microfluidics for droplet handling, or handling of droplet content. Fluorinated oils, e.g. Novec HFE-7500, possess properties such as high permeability for gases and insolubility for most organic compounds, which makes them favorable for droplet-based applications. However, it has been observed that a deterioration in the average magnitude of the primary acoustic radiation force inside water droplets generated in HFE-7500 results in poor acoustophoretic focusing quality [2,3]. As a follow up to these studies, we compared the droplet internal acoustophoretic focusing performance inside aqueous droplets generated in a selection of oils that have previously been used in two-phase applications [4]. Based on these experiments, we conclude that a match in the speed of sound between the dispersed and continuous phase is paramount to high quality acoustic focusing. Here, we present simulation results of four selected cases, aqueous droplets immersed in linseed oil, light mineral oil, silicone oil 50 cSt and HFE-7500, and offer a theoretical explanation to why it is speed of sound that dominates acoustophoretic focusing performance.

The ability to manipulate particles or cells inside water-in-oil droplets is essential to various chemical, biological and biomedical assays performed using droplet microfluidics systems. Bulk acoustophoresis permits spatial localization of cells and particles in a label-free and non-contact manner by utilizing acoustic standing wave and is therefore a suitable particle manipulation method for integration with droplet microfluidics. Fluorinated oils, e.g., Novec HFE-7500, are most commonly used as the continuous phase in typical droplet applications as they are permeable to gases and insoluble for most organic compounds. Previous studies, however, have concluded that sufficient acoustic focusing capability cannot be achieved inside aqueous droplets generated in HFE-7500 due to a mismatch in the acoustic properties of the two phases used. The critical acoustic properties that underlie this difference in acoustic focusing capability were however not identified. This work presents a study where the suitability of a selection of oils, from three categories commonly used in droplet microfluidics (hydrocarbon oil, fluorinated oil and silicone oil), was examined for acoustic particle manipulation in water-in-oil droplets. We hypothesized that when performing acoustic focusing inside aqueous droplets confined in a microfluidic channel (i.e., plugs), due to the presence of a thin oil film between the channel wall and the droplet, the acoustic impedance between the two fluid phases must match in order to increase acoustic coupling efficiency between the two liquid phases. Through this experimental work, we demonstrate that hydrocarbon oils are most compatible with acoustic focusing in a two-phase system that generates channel-confined water-in-oil droplets due to matching acoustic properties with water. We also conclude that acoustic impedance matching between the two fluid phases is not critical for optimal droplet-internal particle manipulation. Our results suggest that a match in the speed of sound of the two phases may be more relevant in practice.

Droplet microfluidics is a promising platform for biochemical applications where compartmentalized droplets serve as individual vials. Droplets are formed by using two immiscible phases, the continuous phase and the dispersed phase, making up the droplets. Droplets are interesting because they can provide fast, parallel reactions with low reagent consumption. Microscale particles, such as cells, can be encapsulated in the droplets and chemical reagents can be added via a pico-injector. However, removal of droplet background signal is hard to achieved by conventional methods, especially if you do not want to risk losing the encapsulated cells. In this thesis, I present a droplet microfluidic system that can achieve this, via droplet-internal particle manipulation using acoustophoresis.

This droplet microfluidic system contains pico-injection and droplet split with acoustophoresis. The pico-injection is used to add fresh solution into the droplets and the droplet split with acoustophoresis is used to remove the droplet supernatant. With the combination of the pico-injector and the droplet split, the background signal of the droplets can be reduced and the cell medium in the droplets can be exchanged. This droplet microfluidic system can also be used to control timing of enzyme reactions by initiating the reaction by adding enzyme-coupled beads via the pico-injector and taking a sample from the droplets at specific time points via side channels. In this work, I have also investigated how the design of the droplet split could be optimized to obtain high particle recovery and enrichment. Finally, acoustic properties of a selection of oils that can used as the continuous phase were mapped to optimize the droplet system for acoustophoresis.

This thesis explores the biochemical applications performed by the droplet acoustofluidics, in-droplet time-controlled enzyme reaction and medium exchange for in-droplet cell culture. Furthermore, the droplet acoustofluidics has the potential to study the reaction kinetics by other enzymes and achieve long-term in-droplet cell culture.

Droplet microfluidics is a powerful method used to characterize chemical reactions at high throughput. Often detection is performed via in-line optical readout, which puts high demands on the detection system or makes detection of low concentration substrates challenging. Here, we have developed a droplet acoustofluidic chip for time-controlled reactions that can be combined with off-line optical readout. The principle of the platform is demonstrated by the enzymatic conversion of fluorescein diphosphate to fluorescein by alkaline phosphatase. The novelty of this work is that the time of the enzymatic reaction is controlled by physically removing the enzymes from the droplets instead of using chemical inhibitors. This is advantageous as inhibitors could potentially interact with the readout. Droplets containing substrate were generated on the chip, and enzyme-coupled microbeads were added into the droplets via pico-injection. The reaction starts as soon as the enzyme/bead complexes are added, and the reaction is stopped when the microbeads are removed from the droplets at a channel bifurcation. The encapsulated microbeads were focused in the droplets by acoustophoresis during the split, leaving the product in the side daughter droplet to be collected for the analysis (without beads). The time of the reaction was controlled by using different outlets, positioned at different lengths from the pico-injector. The enzymatic conversion could be measured with fluorescence readout in a separate PDMS based assay chip. We show the ability to perform time-controlled enzymatic assays in droplet microfluidics coupled to an off-line optical readout, without the need of enzyme inhibitors.

Acoustophoresis is a commonly used technique to manipulate microbeads suspended in single-phase fluids [1]. Fornell et al. demonstrated that acoustophoresis can also be used to manipulate microbeads in droplets when the acoustic properties of the continuous phase match with the acoustic properties of the dispersed phase [2]. Water-in-oil droplets can be considered to be ideal vials for liquid mixing and reactions, as they allow for high throughput, low cross-contamination and low reagent consumption. Droplet microfluidics has previously been applied to study enzyme reactions [3,4]. However, these approaches are limited by on-chip detection and analysis since the enzyme is mixed with the product of the reaction inside the droplets. Here, we present an acoustofluidic chip where we can start the enzymatic reaction by injecting enzyme-coupled microbeads into the droplets containing the substrate by pico-injection and then stop the enzymatic reaction by removing the enzyme-coupled microbeads using acoustophoresis [5]. The analysis can then be performed off-chip, which allows for longer exposure times compared to performing the analysis in-line.

Droplet microfluidics is a powerful method to characterize chemical reactions at high throughput. Often detection is performed via in-line optical readout, which puts high demands on the detection system and makes detection of low concentration substrates challenging. Here, we have developed a droplet acoustofluidic chip for time-controlled microbead-based reactions that is combined with off-line optical readout.

We have developed a fast and simple method for fabricating microfluidic channels in silicon using direct laser writing. The laser microfabrication process was optimised to generate microfluidic channels with vertical walls suitable for acoustic particle focusing by bulk acoustic waves. The width of the acoustic resonance channel was designed to be 380 µm, branching into a trifurcation with 127 µm wide side outlet channels. The optimised settings used to make the microfluidic channels were 50% laser radiation power, 10 kHz pulse frequency and 35 passes. With these settings, six chips could be ablated in 5 h. The microfluidic channels were sealed with a glass wafer using adhesive bonding, diced into individual chips, and a piezoelectric transducer was glued to each chip. With acoustic actuation at 2.03 MHz a half wavelength resonance mode was generated in the microfluidic channel, and polystyrene microparticles (10 µm diameter) were focused along the centre-line of the channel. The presented fabrication process is especially interesting for research purposes as it opens up for rapid prototyping of silicon-glass microfluidic chips for acoustofluidic applications.