Introduction

As a miniaturized experimental platform, microfluidic chips enable extremely sensitive, efficient, low-consumption, safe and environment-friendly biochemical reactions, revealing great potential in a diverse array of applications ranging from chemical synthesis to biological assays [1,2,3,4,5]. However, with fixed microfluidic channel networks, a given chip is dedicated to a specific reaction, which prevents its practical usage in different assays. More importantly, external auxiliary devices (e.g. pressure pumps for sampling, the microscope for observation, and power sources) significantly limit the portability of such lab-on-a-chip (LoC) systems, making microfluidics a technology for laboratory use only. Recently, in addition to conventional microfluidic chips, some new paradigms, for instance, digital microfluidic (DMF) chips [6,7,8,9] and surface-directed microfluidic (SDMF) systems [10,11,12], have emerged as alternatives. The former chips are generally reconfigurable for desired combination of droplet operations, since they realize fluidic operations, such as dispensing, transporting, splitting and merging droplets, by electrowetting effect using electrode arrays [13, 14]. The latter chips drive fluids through capillary action of hydrophilic patterns on an open chip, thus they do not require affiliated driving equipment [15, 16]. Although fluid handling can be realized in a controlled fashion, influence from the electric field with respect to DMF chips and the cross-contamination issues of SDMF chips significantly limit their practical applications [17,18,19,20]. Therefore, neo microfluidic systems that feature self-controlled fluids handling, reconfigurable channel networks, real portability and extensible functionalities are highly desirable.

In fact, nature creatures have already demonstrated sophisticated capabilities to control small volume liquid, which provides the inspiration for developing novel microfluidic devices [21,22,23]. For instance, rice leaves control the directional sliding of droplets using anisotropic superhydrophobicity [24, 25]; Nepenthes pitcher plants surface enables continuous directional water transport along the peristome surface [26]. These natural phenomena provide a hint that self-controlled liquid transport at microliter-scale can be realized by engineering functional surfaces with unique wettability. However, towards microfluidic chip applications, there are several very strict requirements with respect to droplet manipulation, including energies for driving droplets, directional droplet transport, quantitative droplet splitting and merging, as well as avoiding cross contamination. Consequently, the innovation of nature inspired microfluidic devices that are capable sophisticated droplet manipulation remains a big challenge.

In this work, we report the laser fabrication of modular droplet chips (MDC) that can serve as a reconfigurable, miniature and portable experimental platform for self-propelled droplets manipulations and even on-chip reactions. Gravity is used as the driving force for droplets motion. Flexible manipulation of liquid droplets has been achieved by embossing a series of well-designed 3D superhydrophobic structures on PDMS jigsaws. To endow artificial solid surfaces with desired dewetting properties that are comparable to natural creatures, we embossed superhydrophobic structures including sunken and protuberant morphologies on PDMS slices. To manipulate droplets freely as a chip, we fabricated a series of superhydrophobic embossments on PDMS slices that are designed as modules interlocking with each other, and thus enabling chip functions such as droplets transporting, splitting and merging without mass loss. Flexible assembly of a set of modules achieves different chip functions, for instance, solution dilution, SERS detection, cell labeling and chemical synthesis, revealing great potential for developing next-generation LoC systems.

Methods

CO2 laser with a central wavelength of 10.6 μm is employed for the fabrication. Laser process programs are designed using CorelDRAW. Firstly, embossed structures are engraved on PMMA through subtractive laser manufacturing. Pre-designed structures can be directly created through a layer-by-layer scanning manner. The scanning step length is fixed at 100 μm, and the scanning speed is 100 mm/s. Laser power density is measured to be ~ 6300 W/cm2. After the subtractive laser manufacturing, embossed structures form on the PMMA surface. The structured PMMA substrate is washed in water with the help of ultrasonic cleaning, and dried in air for further usage. Secondly, PDMS (contained 10% cross-linking agent, Sylgard 184, Dow Corning) is adopted to duplicate the embossed structures on the PMMA template through a soft lithography process, and cured on a hotplate for 2 h at 80 °C for solidification. In this way, an inverted embossed structure formed on PDMS. Then, another laser treatment is performed to nanotexture the PDMS. In the laser nanotexturing process, the scanning step length is fixed at 100 μm, and the scanning speed is 100 mm/s, the laser power density is 5700 W/cm2. After laser nanotexturing, the embossed structures become superhydrophobic since the laser treatment introduce micro/nanostructures on PDMS. Finally, a superhydrophobic PDMS chip with embossed functional structures is obtained. To achieve a modular design, the squire chips is a fixed size of 3 × 3 cm, in which a tooth (5 × 5 mm) or a gap (5 × 5 mm) is designed on each side for flexible linkage.

Results and discussion

The design of such MDC aims at bridging the gap between macroscopic apparatus and microfluidic chips (Fig. 1a). Laboratory glassware can be easily assembled into experimental instruments that enable various routine chemical synthesis and analysis. However, their consumption of reagents is between several to hundreds of milliliters, which is a waste of chemicals and a threat to the environment. For microfluidic chips, very small dose of reactants (e.g., microliter or even nanoliter scales) is enough, however, microfluidic chips are not reconfigurable, and thus are only specific to a certain experiment. In the case of MDC, only several drops of reagents (one droplet, 5 ~ 150 μL) are needed for the entire experiments. The MDC not only possesses the advantages of ease of fluid handling without the need of any coupled instruments, but also features the merits of microfluidics, including low reagents consumption, safety, environmental friendliness, high sensitivity and efficiency. To meet the requirements of different experiments, we designed a whole set of chip modules on which different superhydrophobic structures have been embossed for different droplet operations (e.g., moving, replacement, merging, splitting, mixing, and collection of droplets, Fig. 1b). We fabricated the embossed superhydrophobic structures via direct laser writing (DLW). The method enables facile, mask-free and high-efficiency fabrication of large-area superhydrophobic chips with three-dimensional (3D) embossments, thus providing an approach to emboss desired superhydrophobic structures for droplet operation. Figure 1c shows the photographs of different chip modules (3 × 3 cm in size), which have been tailored into jigsaw puzzle shapes. The inlet and the outlet of an individual module is fixed at the middle of the sides, thus the assembly of individual modules into a functional chip is very flexible and easy (Fig. 1d). More importantly, the chip is reconfigurable, thus it is suitable for different experiments. Gravity is used as a driving force for droplets operation; thus, no additional instrument is necessary. To convert the gravitational potential energy to kinetic energy, the chips can be tilted at a certain degree. Owning to the surface superhydrophobicity, high energy conversion efficiency can be achieved, which can guarantee a sufficient driving force for droplet operation. The conversion efficiency further depends on the tilted angle and the morphologies of the embossed structures (Supplementary Note 1.1, Fig. S1).

Fig. 1
figure 1

Design principle of MDCs. a An MDC is a gravity driven droplet-handling chip without using external auxiliary power. It contains compatible functional modules, just like macroscopic apparatus, that can be flexibly assembled as an integrated experiment platform. In addition, it is several centimeters in size, and enables systematically manipulating microliter-sized droplets, just like a portable microfluidic chip. b Sketches of basic functional modules adopted in MDCs. c Typical presentations of modules with embossed structures constructed by sunken grooves and protuberant ridges. d An MDC assembled by 12 modules. The scale bar is 1 cm

Full size image

Figure 2a shows the schematic illustration of our fabrication procedures, including laser structuring, soft lithography and laser nanotexturing (see Supplementary materials for methods) [27]. We first embossed an inverse structure as a template on PMMA through subtractive laser manufacturing (Fig. 2b). After that, PDMS is employed to duplicate the embossed structures through a soft lithography process (Fig. 2c). To alter the surface wettability, laser treatment is performed to nanotexture the PDMS, forming the superhydrophobic surface due to the combined effects of carbonization and micronanostructuring (Fig. 2d, e) [28].

Fig. 2
figure 2

Fabrication and manipulations of droplets on MDCs. a Schematic illustration of laser fabrication process of DMC. b A typical PMMA template fabricated via DLW; c A PDMS overpass structure fabricated via soft lithography; d A superhydrophobic overpass emboss fabricated after laser modification; e A SEM image of DMC. Plenty of nanostructures make DMC superhydrophobic. The contact angle of water droplet is 152°. Bar is 500 nm. f A moving-path assignment chip that made of a “L” shape channel jointed with three branches. Droplets (10.0 μL) released from Height-1, Height-2, Height-3, are assigned to Channel-a, Channel-b, Channel-c, respectively. g A moving-path assignment chip made of an embossed tunnel and a “L” shape branch. Drop-1 (20.0 μL), whose diameter (3.4 mm) is much larger than the width (W, 2 mm) of the tunnel, is pushed to Channel-a; Drop-2 (10.0 μL), whose diameter (2.7 mm) is slightly larger than W, is squashed and assigned to Channel-b. Cross-sections of the tunnel with a 20.0 μL or 10.0 μL droplet are inserted at the bottom corners to further illustrate the size comparison. h A droplet replacement chip made of a sunken groove with two protuberant pillars. Moving Drop-2 collides and replaces stationary Drop-1 without merging. A model and a physical image are inserted on the top. i Cross-sections of an overpass chip made of opposite slopes. Drop-2 jumps over Drop-1 along the white dotted line, from Slope-1 to Slope-2. Inserted images are top view of the jump process. The scale bar is 3 mm. Detail calculations are shown in Supplementary

Full size image

To operate a water droplet freely, controlling its moving track is a basic requirement. We designed and fabricated sunken grooves and protuberant ridges, as well as the combination of them to guide droplets transport (Supplementary Movie 1, Supplementary Note 1.2, Fig. S2). Generally, the grooved channel can control the swerving movement of a droplet, providing its moving speed is below 189 mm/s, since the maximum trapping force is calculated to be ~ 23 nN. To guarantee a stable control of high-speed droplet, protuberant ridges structures that can rebound droplets have been embossed. By combining the grooves and the protuberant ridges structures together, more stable droplets delivery can be achieved (Fig. S2d). Besides, as a droplet chip, controllable assignment of the droplet moving path is also very important for droplet delivery. In this work, by embossing special superhydrophobic structures (Supplementary Movie 2, Supplementary Note 1.3), different droplets can be assigned into different channels according to their physical parameters, such as the speed, size, and kinetic energy. We first demonstrated the assignment droplets with different instant speeds by releasing them at different heights on a titled (10°) chip (Fig. 2f). Droplet released from Height-1, Height-2 and Height-3 can gain an initial speed of ~ 118 mm/s, ~ 201 mm/s and ~ 328 mm/s for the horizontal projectile motions, and thus be assigned to accepting channels of Channel-a, Channel-b and Channel-c, respectively. Droplets moving-path can be also controlled according to their geometry sizes (Fig. 2g). By tuning the width of the embossed tunnel, droplets with different volume can be assigned into different channels. Besides, the moving-path assignment can be also achieved according to the kinetic energy of a droplet (Fig. S3). When a shallow Channel-b jointed with a deep branch Channel-a, a droplet released from different heights at Channel-b will either be trapped in Channel-a or skip over it, which depends on its kinetic energy. The kinetic energy of Drop-3 (trapped in Channel-a) and Drop-4 (skipped over it) is calculated to be 65 and 83 nJ, respectively (at the joint). The above-mentioned structures demonstrate the precise control over the droplet moving-path.

In addition to the droplet movements, splitting, replacing and merging of droplets are also basic functions for droplet chips (Supplementary Movie 3). Splitting droplets accurately enables high-through assays through parallel reactions [29]. By embossing some sharp structures, quantitative droplets splitting can be realized provided the kinetic energy of a droplet is large enough (over 5980 nJ for a 40 μL droplet) to overcome the surface energy (Supplementary Note 1.4, Fig. S4) [30]. Moreover, by tuning the colliding position, it is possible to control the splitting ratio of a droplet, quantitatively (Fig. S4b ~ d, Supplementary Table 1). On the other hand, to operate droplets reactions on a chip, it is necessary to get droplets merged controllably. To reach this end, a droplet can be trapped first, then another moving droplet can be guided to collide and mix with it, and take it away subsequently. A common structure used for droplets merging is the “Y” shape channels including a shallow channel and a deep branch channel (Supplementary Note 1.5, Fig. S5). First, a droplet with relatively small kinetic energy can be trapped at the joint; then another droplet with relatively large kinetic energy will be guided to merge with the trapped droplet and bring it away along the channel. In addition to merging, one droplet would bounce and replace the other one when two droplets meet (Fig. 2h), if the air between them did not have enough time to escape (see Supplementary Note 1.6 for quantitative calculation) [31, 32]. Droplets replacing can be used for sequence reactions.

For 2D microfluidic chips, there exist a limitation that restricts the miniaturization and integration of complex channel networks: two kinds of liquids cannot synchronously pass across a channel junction without mixing. Thus, the channels cannot cross each other. To solve this problem, we designed and fabricated an overpass structure (Fig. 2i, Supplementary Movie 4). By integrating a pair of wedge structures to the channel junction, a droplet can jump over the crossed channel, thus two droplets can pass across the channel without mixing. Such a “overpass” structure would consume ~ 310 nJ of kinetic energy of a droplet (Supplementary Note 1.7, Fig. S6).

By embossing a series of well-designed superhydrophobic structures on one PDMS slice enables on-chip droplet reactions. We demonstrated here the chromogenic reactions (Supplementary Movie 5, Fig. S7). A droplet of pH indicator (Bromothymol Blue) was split into three droplets that fell into different cells, respectively. Then, droplets of acid and alkali released from different channels were guided to mix with each other or directly go to the cells that contained pH indicator. In this way, three chromogenic reactions based on droplets have been realized within several seconds. However, the droplet chip is not reconfigurable. To overcome this limitation, we embossed various superhydrophobic structures as different modules. Flexible assembly of these functional modules can form different chip networks, thus enabling various reactions (Supplementary Movie 6). Figure 3a illustrates a droplet splitting chip assembled through five modules. Two parallel sharp ridges are designed against a groove (module-2), which can quantitatively split Drop-M into three droplets. After the gravity induced accelerating process, the Droplet-M can gain an instantaneous velocity of about 180 m/s in 1 s (blue line, Fig. 3b). When it knocks against the sharp protuberant ridges on module No.2, the droplet was split to three droplets (Drop-M1, 2, 3) that were collected by modules No. 3, 4,5, respectively. Figure 3b records the moving tracks of the as-generated droplets. When these droplets passed the junction between modules, a slight deceleration can be detected. Figure 3c, d shows droplets assigning modules. Smaller droplets (Drop-1, < 35 μL) can be assigned to module-2; whereas relative larger droplets (Drop-2, > 40 μL) can be assigned to module-3. The moving track analysis clearly demonstrate the moving-path assignment of the two droplets of different size (Fig. 3e). For chemical reactions, the droplet chip requires merging & oscillatory functions; this can be achieved by assembling a series of groove, merging and oscillatory modules (Fig. 3f). A droplet of hydrochloric acid (Drop-2) was introduced to merge with a droplet of alkalized bromothymol blue (Drop-1) first. Then the merged droplet (Drop-1 + 2) moved along the “Z-shape” oscillatory modules for efficient mixing. Continuous color change from blue to green can be observed due to neutralization reaction. Figure 3g shows the moving track of the droplets, which also reveals their instantaneous velocity at different position.

Fig. 3
figure 3

Typical representations of MDCs. a A trisected splitting MDC assembled by 5 modules. Drop-M released from module No.1 is split into three parts on module No.2, which are then introduced onto module No.3, No.4, No.5, respectively. b Moving track analysis of Drop-M and Drop-M1, 2, 3 shown in (a). c, d) A moving-path assignment MDC assembled by 3 modules. Smaller Drop-1 (35.0 μL) is assigned to module No.2, while larger Drop-2 (40.0 μL) is assigned to module No.3. e Moving track analysis of Drop-1 and 2 shown in (c) and (d). f A mixing & oscillatory MDC assembled by 7 modules. Hydrochloric acid (Drop-2, 30.0 μL, 0.2 mol/L) released from module No.1 will merge with alkalized bromothymol blue (Drop-1, 30.0 μL 0.1 mol/L), then move along oscillatory modules. During oscillatory, the color of Drop-1 + 2 changed from blue to green, finally to yellow. The scale bar is 1 cm. g Moving track analysis of Droplet-1, 2 and 1 + 2 shown in (f)

Full size image

Using these chip modules, more functional chips can be flexibly assembled for different experiments. As a typical example, a 10-fold droplets dilution chip is demonstrated using splitting, moving-path control, merging, oscillatory shaker, and collecting modules (totally 15 modules, Fig. 4a and b, Supplementary Movie 7, Supplementary Note 2.1). A general dropper is employed to inject droplets with an average volume of 44.0 μL on the MDC. The dilution process only requires input one drop of initial solution (R6G aqueous solution) and two drops of water. By repeating this process, 102-fold, and 103-fold dilution can be achieved. To evaluate the dilution accuracy, the concentrations of the diluted droplets were tested (Fig. 4c). The maximum deviation of droplets diluted from 10− 1 mol/L R6G droplets by 10-fold is 4% among ten droplets, and the average concentration is 1.01 × 10− 2 mol/L, revealing satisfactory accuracy. In the case of 102-fold, and 103-fold dilution, the deviations became lager slightly. Nevertheless, the average concentrations are still close to theoretical value. Besides, the concentrations of 10− 4 mol/L droplets that diluted from 10− 3, 10− 2, 10− 1 mol/L R6G droplets, respectively, are also tested (Fig. S8), in which similar dilution accuracy has been achieved (Fig. 4c). The deviations might come from the difference in the volume of individual droplets, the splitting ratio and the evaporation of water, which are inevitable. Even so, such a quantitative droplet dilution chip is promising for real out-lab assays.

Fig. 4
figure 4

An MDC for droplet dilution and SERS detection. a A 10-fold droplet dilution MDC assembled by 15 modules, including 1:4 droplets splitting modules (No.1 and No.4), 1:1 droplets splitting modules (No.9 and No.12), moving-track control modules (No.6 and No.14), droplets merging modules (No.3 and No.11), oscillatory shaking modules (No.7 and No.8), droplets collector (No.2, No.5, No.10, No.13 and No.15). The red dotted lines are the tracks of R6G droplet. The white dotted lines are the tracks of water droplets. b The video screenshots of 10-fold dilution process of 44.0 L 10− 1 mol/L R6G droplet, which is gradually fading, indicating it is diluted. Detail explanations are shown in Supplementary. c Accuracy analysis of 10-, 102-, 103-fold dilution of 10− 1 mol/L R6G droplets. The inserted error bars represent the concentration deviations among ten droplets at each dilution fold. d SERS spectra of R6G droplets at four different concentrations that are diluted from 10− 5 mol/L R6G droplets. Before SERS detections, R6G molecular are enriched on MDC by hydrophobic enrichment effect. The inserted images are 44.0 L R6G droplets before and after enrichment

Full size image

Using this MDC, surface enhanced Raman scattering (SERS) experiments can be carried out easily. Typically, R6G droplets of different concentrations diluted above can be easily mixed with a droplet containing silver nanoparticles (AgNPs, Supplementary Note 2.2, Fig. S9). After the evaporation of water, both R6G molecules and AgNPs are enriched on superhydrophobic surfaces within a small region. The aggregation of AgNPs can form abundant “hot spot” under laser irradiation, thus sensitive SERS detection can be realized (Fig. 4d) [33]. The superhydrophobicity is quite helpful for the enrichment of analytes. By using superhydrophobic surface to control concentration, much high sensitivities are achievable [34]. On the other hand, to avoid the evaporation of solvents during long time experiments, ionic liquids can be used instead of water. As a typical example, the polymerization reaction of aniline was carried out based on this MDC (Supplementary Note 2.3, Fig. S10) [35].

In addition, MDC could be used in biology research (Supplementary Note 2.4, Fig. S11). Biomolecular, especially proteins, are easily adsorbed by the solid surfaces, which is a serious issue that limits the practical applications of DMF [36]. Superhydrophobic surfaces could prevent protein adsorption effectively due to the presence of a thin air layer between the droplets and the solid substrate [20]. To confirm this hypothesis, droplets that contain high concentration of anti-human IgG (whole molecule)-FITC were released to move on a superhydrophobic MDC (Supplementary Note 2.5, Fig. 5), then the labeled and unlabeled areas (A1 and A2) of the MDC were identified from fluorescence images to calculate protein coverage rate (CR, \( \frac{{\mathbf{A}}_{\mathbf{1}}}{{\mathbf{A}}_{\mathbf{1}}+{\mathbf{A}}_{\mathbf{2}}}\times \mathbf{100}\% \)). Notably, no fluorescence signals can be detected on the MDC after transporting protein droplets for 10 times, demonstrating good antipollution capability. Even after transporting the droplets for 1000 times, the protein CR is only 0.0225%. On the contrary, for general PDMS chip without superhydrophobic modifications, considerable proteins residual can be observed after only once transport of a protein droplet. According to these results, MDC that features antipollution property could effectively reduce the biomolecular residual, emerging as a reusable experimental platform for biochemical reactions [37, 38].

Fig. 5
figure 5

Proteins residual test on MDC. a Fluorescence image of the surface of MDC after transporting protein droplets several times. Inset numbers are CP. b The dependence of CP and related times of usage. c Fluorescence image of PDMS chip (control experiment) after transporting protein droplets 1 time. Instant test shows the presence of proteins on such a PDMS chip. Even after water rinsing, 96% of protein residuals can be detected. Ultrasonic cleaning can remove most of the protein residuals, while the CP is still much larger than that of MDC used for 1000 times. The scale bar is 200 μm

Full size image

From a practical perspective, currently available microfluidic devices suffer from the limitations with respect to suitability for different reactions, cross-contamination in reuse and the necessary external auxiliary equipment. For conventional microfluidic chips with fixed microfluidic channel network, a chip is generally dedicated to a specific reaction, which lacks general applicability in different analysis. For emerging DMF and SDMF chips, the issue of cross-contamination is ineluctable, which also restricts practical usage. More importantly, most of the microfluidic chips themselves are miniature devices with small dimension and portability. Nevertheless, the use of microfluidic chips usually involves complex external auxiliary equipment, for instance, pumps and microscope. All of the above-mentioned limitations make microfluidics a technology for laboratory use only. And therefore, it is highly desirable to develop neo microfluidic systems that can address these issues.

Conclusions

Herein, inspired by the droplets transport on natural micro/nanostructures, we designed and fabricated MDC that permit flexible manipulation of various droplets by embossing superhydrophobic structures including sunken and protuberant morphologies on PDMS slices. Gravity has been used as a driving force, accurate droplets operations including droplets transporting, splitting, merging, replacing, jumping, oscillating, and collecting have been realized without the use of any couple instruments. The resultant superhydrophobic PDMS modules can be assembled into functional droplet chips for various experiments. Bearing a modular design, the chip is reconfigurable. Using MDCs, common experiments, such as biological label, optical detection, solution dilution, and chemical synthesis, can be realized. By taking advantage of the superhydrophobicity, reagents residual can be avoided, thus the chip modules are reusable. The MDC features controllable fluids handling, reconfigurable chips structures, portability and extensible functionalities, revealing great potential for the next generation of lab-on-chip systems.