**2.2 From micro- to millifluidics**

The use of femto-picoliter droplets is crucial in a number of experiments, such as biomolecular assays or molecular evolution. However, the upscale of the technology is necessary, when investigating the cells metabolism or their response to one or several stress factors. In this case, the behavior of populations over several generations of cells has to be monitored. This format of the experiments is possible, when the aqueous reactors (droplets) contain sufficient amount of *e.g.* nutrients and other supporting species. In contrast to more common microfluidic approaches, millifluidics appears as an approach for quick screening of larger volume aqueous samples for any quantitative analysis of its biological and chemical content. For comparison: while microfluidics work with picolitre reservoirs, droplets in millifluidic systems reach volumes in the range of 50–200 nL [29]. Therefore, the technology was adapted respectively, by switching from microchannels used in polydimethylsiloxane (PDMS) chips to commercially available off-shelve components, *e.g.* transparent fluorinated ethylene propylene (FEP) tubes with an inner diameter (ID) of about 0.5 mm. Thus, millifluidics emerge as a tool, crucial for a long term monitoring of a large number (up to 103 ) of biochemical reactors containing cell or bacterial populations. The millifluidic technique [13] in this particular case would outperform the microfluidic approach in terms of its better suitability.

In the following we demonstrate several examples of the millifluidic systems that are used to incubate the microbes, investigate their antibiotic susceptibility and even to study their coexistence processes. In these realizations authors demonstrate the fluorescent detection principle, based on the multiple readout of the droplets one-byone to build up the time dependent curves of the process kinetics [29–31].

**Figure 3** demonstrates the typical optical and fluidic setup assembly that is used to fulfill these goals as well as its capabilities [30]. Millifluidic setups typically offer the automatic droplets readout to perform large-scale high resolution assays and calibration. Panels a–d in **Figure 3** depict a schematic diagram of the system, which is divided into two main areas: droplet generation and detection. Sets of optomechanical elements and polymeric capillary tubings are used to assemble such system, being controlled by the custom LabView software. This program enables controlled droplets generation, counting, detection, back-forth motion for repeatable scan of every droplet in the chain. In the detection area, a double fluorescent detector is designed to be responsive to the blue and yellow fluorescent proteins (BFP and YFP), simultaneously (see panels d–g in **Figure 3**). A fluidic pump controls the droplet sequence flow forward and backward to the detectors by infusing and refilling the fluidic circuit. LabView software can automatically measure the growth curves of two strains with high precision during hours and days (**Figure 3e**–**g**). In the following, we applied the setup for the quantitative recording of the growth curves from the multiple droplets, including the co-culture of two strains of *Escherichia coli*, revealing fluorescence according to the BFP and YFP emission spectra. After determining the relationship between fluorescent signal and cell density for each strain, calculation of the limits of detection (LODs) of the droplets analyzer was performed, with LOD for *E. coli* BFP around *ca.* 5000 cells per droplet, and for *E. coli* YFP around 6000 cells per droplet. Growth kinetics of bacteria both, in monoculture (**Figure 3e**–**i**) and co-culture (**Figure 4**) were accurately measured, with various initial cell densities (inoculum). Interestingly, the setup opens possibility to study the organisms' cooperation and coexistence at small scale, by varying the fraction of each bacterial strain inside of the droplet, that affects the final growth curves for these strains. Furthermore, such system represents the elegant way to form a gradient

#### **Figure 3.**

*Millifluidic setup for fluorescent bacteria detection. (a) Schematic of the detection mechanism. (b) Photography of the generated droplets and (c) their fluorescence measurement. (d) Schematics of the double detector for the analysis of two wavelengths, each belonging to a different cell strain: E. coli YFP and E. coli BFP. (e) Growth curve for E. coli YFP monoculture, with example droplet peaks in (f). (g) Growth curve for E. coli BFP monoculture. Adapted from ref. [30].*

**Figure 4.** *Growth kinetics of cocultures from different initial cell density ratios. Adapted from ref. [30].*

of the chemicals, *e.g.* nutrients, drugs, etc. along the droplets chain and challenge the microbes to survive in this stressful environment [29]. This opens the door towards measurements of bacterial resistance to *e.g.* antibiotics, which are of interest for microbiological and clinical applications.

## **2.3 Outlook**

Overall, optical tools played very important role in the establishing and development of the droplets based microfluidics, and the lab-on-a-chip field overall. However, although being very efficient optical approaches still stay bulky. Therefore, one of the main challenges today is the need for the development and integration of novel miniaturized optics-less detection principles [32, 33], outperforming or being complementary to the conventional approaches. In contrast, new devices make the measurement processes independent of the limitations of optical microscopy, *i.e.* dynamic range or use of molecular labels.
