The Multi EDGE device – the importance of chip design for high-rate monodisperse droplet formation

This research summary written by Sten ten Klooster is based on the article “Upscaling microfluidic emulsification: the importance of sub-structure design in EDGE devices”.

The project developed a Multi-EDGE device that can produce monodisperse 10-μm droplets at ∼ 300 L⋅m−2⋅h−1 over long periods. They also discuss strategies for upscaling microfluidic-based droplet formation and the importance of overcoming chip manufacturing limitations for the industrial production of emulsions.  

Their work was recently featured in Chemical Engineering Science and showed promising results towards industrial applications for microfluidic-based droplet formation.

Abstract

Microfluidic emulsification devices require low energy usage for highly monodisperse droplet formation. However, current throughputs need improvement for larger-scale microfluidic emulsion production. This study used the Multi EDGE device to produce 10-µm hexadecane droplets in 0.5 wt% SDS solution at 300 L⋅h-1⋅m-2 of chip area. We showed that a higher flow resistance of the dispersed phase supply channels of Multi EDGE could lead to a higher maximum pressure that can be applied before monodispersity is lost, which leads to faster pore refilling. In turn, this could increase the droplet formation frequency per droplet formation unit from 58 s-1 to > 1800 droplets s-1 per droplet formation unit, as shown for a small-scale microfluidic emulsification device that had a higher flow resistance of the dispersed phase supply channels. The fluxes and small droplets of Multi EDGE show these devices’ potential for upscaling, whereas further improvements are still possible by the design of the sub-structure.

Introduction | Upscaling droplet formation

Emulsions are mixtures of two immiscible fluids (e.g. water and oil) and form the basis of many products frequently used in daily life, such as paint, sunscreen, and mayonnaise. Generally, large-scale homogenisers are used to produce emulsions with the oil phase dispersed into the continuous water phase for droplet formation. These processes have some disadvantages: they are energy intensive, and they have limited control over the droplet size (distribution) [1,2], whereas the latter is an important parameter for the product characteristics and stability [3].

Microfluidic emulsification is much less energy intensive (typically 1-5% of that of a high-pressure homogeniser) [1], and it shows excellent control over the droplet size (distribution) [4]. Especially spontaneous microfluidic emulsification is a promising alternative to conventional emulsification techniques because only the dispersed phase flow needs to be controlled to get monodisperse droplet formation [5]. For small droplets (< 10 µm), productivity is in the order of mL⋅h-1, which, in the best case, only allows for small-scale characterisation experiments. 

To further explore its potential, upscaling is required [6]. To improve the productivity of microfluidic emulsification devices, many aspects of microfluidic emulsification devices have been investigated previously [7], but the influence of the sub-structure of the dispersed phase channels has never been considered.

Aims

• To create an upscaled microfluidic emulsification device – the Multi EDGE – and show its potential for large-scale emulsion production. 
• To investigate how the sub-structure of the dispersed phase supply channels can be further optimized to further increase productivity.

Experiment setup | Droplet formation and monitoring

The materials, microfluidic chips, and connectors used in this study are listed below:

• Hexadecane was used as the to-be-dispersed phase and 0.5 wt% sodium dodecyl sulphate as the emulsifier, dissolved in ultrapure water. 
• Ethanol was used to clean the chips.
• The productivity of a small-scale chip (Partitioned EDGE) was compared with an upscaled chip (Multi EDGE). For both chips, the droplet formation units (DFUs) had dimensions of 10 × 2 µm (width × height) (Figure 1, Table 1). 

The main difference between the chips was in the amount of droplet formation units, which was 25 for Partitioned EDGE and 75,000 for Multi EDGE, and in the flow resistance of the sub-structure of the dispersed phase channels (expressed per DFU), which was 3.3-times higher for Partitioned EDGE compared to Multi EDGE. Further details about both chips are given below.

• Fabrication of the Partitioned EDGE chips (Figure 1ab, Table 1) was done by Micronit Microfluidics, Enschede, The Netherlands) [8]. The chip was placed in a chip holder from Micronit (Fluidic Connect PRO Chip Holder with 4515 Inserts, Micronit Microfluidics, Enschede, The Netherlands) and connected with standard tubing.
• For Multi EDGE silicon substrates of 400-µm thickness were used. Deep reactive ion etching was used to fabricate the channels in the substrate (Figure 1) (Cytocentrics B.V., The Netherlands). 
• The Multi EDGE chip was placed in a custom-made module. The Multi EDGE chip was 10 × 10 mm, and the part containing the droplet formation units was 5 × 6 mm, of which the effective pore area was 5 % (Figure 1cd).

The flow of the dispersed and continuous phases was finely tuned with a pressure-driven flow controller (Elveflow OB1, Mk3), and droplet formation was monitored by using an inverted microscope (Axiovert 200 MAT, Carl Zeiss B.V., The Netherlands), which was connected to a high-speed camera (FASTCAM SA-Z, Photron Limited, Japan). 

Key findings | Upscaling monodisperse droplet generation

With the Multi EDGE microfluidic emulsification device (10 × 2 µm droplet formation units (DFUs)), monodisperse hexadecane droplets were produced over 8 h (longest run because of lab closing hours). The lowest applied pressure at which droplet formation was observed was 95 mbar. Up to 130 mbar, the droplets were monodisperse, and the size remained constant at 11.0 µm with a coefficient of variation of < 10 % (Figure 2a). 

Upon increasing the pressure above 130 mbar, a few DFUs produced larger polydisperse droplets (vertical solid line Figure 2a). Increasing the pressure increases both droplet formation frequency (Figure 2b) and the number of active DFUs (Figure 2c) till blow-up is reached, which results in the formation of larger and polydisperse droplets. Thus, increasing the pressure (up to 130 mbar) increased the flux per area increased quadratically (Figure 2d). At 130 mbar, the DFU activity was 93%, which produced, on average, 58 droplets s-1 per DFU, resulting in an oil flux of 313 L⋅m-2⋅h-1.

When comparing Multi EDGE with the small-scale chip Partitioned EDGE, we found that in both chips, droplet formation starts at the same pressure (± 95 mbar). Also, the droplets that they produced – 11 µm for Multi EDGE and 10 µm for Partitioned EDGE – were similar. Upon increasing the pressure, the droplet formation frequency per DFU initially increased faster for Multi-EDGE compared to Partitioned-EDGE (initial slopes in Figure 2b), which is caused by the lower flow resistance of the sub-structure of Multi-EDGE. Just before the blow-up, the maximum droplet formation frequency per droplet formation unit was ∼ 1800 s-1 for Partitioned EDGE (blow-up pressure of 900 mbar) compared with ∼ 58 s-1 for Multi EDGE (blow-up pressure of 130 mbar). 

The productivity of Partitioned EDGE thus shows that the productivity of Multi EDGE can be improved further. To do so, we first have to find the cause of this remarkable result. For that, it is important to understand droplet formation, which we divide into two stages: (1) necking and (2) downtime [9].

During necking, the actual droplet is formed (Figure 3abc), and during the downtime, the meniscus of the oil-water interface moves forward if the applied pressure is higher than its Laplace pressure, after which the oil can leap into the deeper continuous phase channel again and start forming the next droplet (Figure 3cde). The low flow resistance of Multi EDGE caused that at relatively low pressures, the dispersed phase flow becomes too high for the small monodisperse droplet to be formed, resulting in low blow-up pressure. The low blow-up pressure causes that the Laplace pressure of the meniscus just after droplet formation is higher than the applied pressure (Figure 3d), and therefore emulsifier adsorption is required to lower the Laplace pressure before the meniscus can move forward (thus, increasing downtime, and reducing frequency). 

In contrast, the high blow-up pressure of Partitioned EDGE causes that the meniscus also moves forward without surfactants adsorbed at the interface (thus, reducing the downtime, and increasing frequency). The productivity per droplet formation unit can thus be increased through the flow resistance of the dispersed phase sub-structure. This will lead to a blow-up pressure that is substantially higher than the Laplace pressure of the empty interface.

Conclusions | Improved microfluidic emulsion production

Here, we showed that the Multi EDGE microfluidic emulsification device can produce monodisperse hexadecane 10-μm droplets at ∼ 300 L⋅m−2⋅h−1 over long time periods, which brings the device close to industrially relevant values. 

Productivity can be improved further by wisely designing upscaled microfluidic emulsification devices. This can be done by increasing the flow resistance of the dispersed phase sub-structure leading to a larger driving force for the refilling of the droplet formation unit. These insights are helpful, but to get to the industrial production of emulsions produced by microfluidics, we need to overcome chip construction limitations. We would therefore like to collaborate with construction companies to take up the challenge aiming for the next breakthrough with Multi EDGE.

In this study, droplet-based microfluidics was employed to generate monodisperse emulsions. The use of pressure-driven flow-controlled microfluidics allowed the authors to produce very monodisperse droplets at high throughput and with high accuracy.

References