Enhancing single-cell encapsulation in droplet microfluidics with fine-tunable on-chip sample enrichment

Tao Tang,Hao Zhao,Shaofei Shen,Like Yang,Chwee Teck Lim

DOI: https://doi.org/10.1038/s41378-023-00631-y

2024-01-02

Microsystems & Nanoengineering

Abstract:Abstract Single-cell encapsulation in droplet microfluidics is commonly hindered by the tradeoff between cell suspension density and on-chip focusing performance. In this study, we introduce a novel droplet microfluidic chip to overcome this challenge. The chip comprises a double spiral focusing unit, a flow resistance-based sample enrichment module with fine-tunable outlets, and a crossflow droplet generation unit. Utilizing a low-density cell/bead suspension (2 × 10 6 objects/mL), cells/beads are focused into a near-equidistant linear arrangement within the double spiral microchannel. The excess water phase is diverted while cells/beads remain focused and sequentially encapsulated in individual droplets. Focusing performance was assessed through numerical simulations and experiments at three flow rates (40, 60, 80 μL/min), demonstrating successful focusing at 40 and 80 μL/min for beads and cells, respectively. In addition, both simulation and experimental results revealed that the flow resistance at the sample enrichment module is adjustable by punching different outlets, allowing over 50% of the aqueous phase to be removed. YOLOv8n-based droplet detection algorithms realized the counting of cells/beads in droplets, statistically demonstrating single-cell and bead encapsulation rates of 72.2% and 79.2%, respectively. All the results indicate that this on-chip sample enrichment approach can be further developed and employed as a critical component in single-cell encapsulation in water-in-oil droplets.

nanoscience & nanotechnology,instruments & instrumentation

What problem does this paper attempt to address?

### What problem does this paper attempt to solve?

This paper aims to solve the problem of low single - cell encapsulation efficiency in droplet microfluidic technology. Specifically, traditional droplet microfluidic devices face the following challenges during single - cell encapsulation:

1. **Trade - off between initial cell density and chip focusing performance**: Although a higher cell suspension density can theoretically increase the encapsulation rate, it will lead to increased interactions between cells, which in turn reduces the focusing and encapsulation efficiency. On the other hand, a lower cell density may result in too many empty droplets.

2. **Inefficiency of random encapsulation**: According to the Poisson distribution, in the case of random encapsulation, the theoretically maximum single - cell encapsulation rate is 37%, which leads to a large amount of sample and material waste.

To solve these problems, the author designed a new type of droplet microfluidic chip, which includes the following key components:

- **Double - helix focusing unit**: Used to focus a low - density cell / bead suspension (2 × 10^6 objects/mL) into a nearly equidistant linear arrangement.
- **Flow - resistance - based sample enrichment module**: Optimizes cell concentration by adjusting the number of outlets to remove the excess aqueous phase.
- **Cross - flow droplet generation unit**: Ensures that the focused cells or beads can be encapsulated into droplets one by one.

Through these improvements, this study achieved a higher single - cell encapsulation rate and reduced the waste of samples and materials. The experimental results show that under optimized conditions, the encapsulation rates of single - cells and beads reached 72.2% and 79.2% respectively.

In addition, this study also developed a droplet detection algorithm based on YOLOv8n to count the number of cells or beads in droplets, further verifying the effect of single - cell encapsulation.

### Formula

The formula involved in the paper is the Poisson distribution, which is used to calculate the probability that a droplet contains a specific number of objects:

\[ P(\lambda; k)=\frac{\lambda^{k}e^{-\lambda}}{k!} \]

where:
- \( \lambda \) is the average number of objects (cells or beads) contained in each droplet.
- \( k \) is the actual number of objects contained in the droplet.

By adjusting droplet generation parameters, such as flow rate, sample concentration, etc., the value of \( \lambda \) can be controlled, thereby optimizing the single - cell encapsulation effect.