A Modular Microfluidic Device via Multimaterial 3D Printing for Emulsion Generation

3D-printing (3DP) technology has been developing rapidly. However, limited studies on the contribution of 3DP technology, especially multimaterial 3DP technology, to droplet-microfluidics have been reported. In this paper, multimaterial 3D-printed devices for the pneumatic control of emulsion generation have been reported. A 3D coaxial flexible channel with other rigid structures has been designed and printed monolithically. Numerical and experimental studies have demonstrated that this flexible channel can be excited by the air pressure and then deform in a controllable way, which can provide the active control of droplet generation. Furthermore, a novel modular microfluidic device for double emulsion generation has been designed and fabricated, which consists of three modules: function module, T-junction module, and co-flow module. The function module can be replaced by (1) Single-inlet module, (2) Pneumatic Control Unit (PCU) module and (3) Dual-inlet module. Different modules can be easily assembled for different double emulsion production. By using the PCU module, double emulsions with different number of inner droplets have been successfully produced without complicated operation of flow rates of different phases. By using single and dual inlet module, various double emulsions with different number of encapsulated droplets or encapsulated droplets with different compositions have been successfully produced, respectively.


Legends of Supplementary Movies S1 to S3
1.1 Supplementary Movie S1: Active control of double emulsion generation The video demonstrates that double emulsions with 1-4 of inner droplets are generated with excitation frequency: 3 Hz, 6 Hz, 9 Hz and 12 Hz, respectively. The flow condition is QI = 6 μl/min, QM = 14 μl/min, and QO = 180 μl/min. The video is recorded with 300 frames per second (fps) and played with 30 fps.

Supplementary Movie S3:
Double emulsions with different compositions of encapsulated droplets This video demonstrates that double emulsions with variable compositions of inner droplets are generated by changing the flow rate ratio of the inner phase while keeping it a constant of QI = 15 μl/min. The other flow condition is QM = 16 μl/min and QO = 160 μl/min. The video is recorded with 300 fps and played with 30 fps.

Experimental and numerical analysis on the deformation of flexible channels
15 pieces of specimens (38×7×4.9 mm 3 ) were printed in total, and 5 specimens for each printing directions: 0° (in the printline direction), 45° and 90° as shown in Fig. S2a. The strain-stress relation was acquired using a 5940 Series Single Column Table Top Systems (Instron ® ) with the customized grip and fixture. The maximum strain change that can be applied in our device is about 30%. As shown in Fig. S2a, different printing directions have little influence on the Young's modulus. Up to the range of 30% strain change, the TangoPlus material has shown a linear strain-stress relation, and still lies in the elastic deformation regime. Such elastic deformation is adequate in our experiments for emulsion generation. The Young's modulus is therefore calculated as 0.504 MPa. The Poisson's ratio of TangoPlus material is 0.495-0.499, provided by supplier.
Numerical studies on the deformation of flexible channels were conducted using Abaqus CAE (Dassault Systè mes). As shown in Fig. S2b, uniform pressure is radially applied to the 3D coaxial flexible channel, whose ends both connect to the rigid part. The inner and outer diameters of the flexible channel are 550 µm and 4 mm, respectively, and the flexible channel length is 5mm. These dimensions are in accordance with our actual printed one. For this structure, axisymmetric assumption in vertical direction and symmetry about the horizontal middle plane can be applied, and therefore we can choose 1/8 of flexible channel to make simulations. As shown in Fig. S2b, the upper plane has a fixed support as the boundary condition (BC). The two sides are limited by the axisymmetric condition. BC of the bottom plane is plane symmetry towards the horizontal plane. The additional pressure resulting from the liquid flow in the channel can be negligible due to the small flow rates used in microfluidics for emulsion generation. Simulation results have shown that the center part of channels has the largest radial displacement of about 15 µm under the applied pressure of 200 mbar, and the displacement decreases along the vertical channel. Axial and tangential displacements can be negligible. According to the stability analysis, a critical pressure over 2 bar will cause the channel buckled and squashed. More than 50,000 cycles with 300 mbar pressure are applied for 10 PCUs, and no one has been found broken or leaked, which has demonstrated that our design and fabrication is robust.
As shown in simulation results (Fig. S2b), when a static pressure is applied, the flexible channel will deform accordingly. The local displacement of the inner channel surface in the radial direction is determined by the pressure P and the axial position z , expressed as   , I P z . The volume change () VP caused by this deformation can be achieved through an integration of the displacement over the inner channel surface as shown in equation (S1).
where C D and L are the inner diameter and length of the channel respectively. The volume change has also been measured by observing the change of liquid meniscus in the glass capillary at downstream. As shown in Fig. S2c, the simulation results agree well with the experimental results. We have also found that the volume change with respect to the applied pressure shows a linear relation, and therefore, the volume change can be expressed with equation (S2).
()  V P KP (S2) where K = 0.61 nl/mbar is obtained experimentally. It is noteworthy that although the deformation of the flexible channel can be complicated, the volume change resulting from this deformation can be described as a simple linear relation with respect to the applied pressure, which can benefit the engineering applications.
If a periodic pressure () Pt with an excitation frequency F f is applied, fF has influence on the deformation of the channel and corresponding volume change (Fig. S2c inset). According to the experimental results, we can express their relation as follows:

Flow rate estimation
The applied pressure waveform has been found to play a minor role in the droplet generation in our experiments, so we present the pressure: By replacing the term P in equation (S4) with equation (S5), we obtain the volume change (t) V with respect to the applied pressure and time t : The actual flow rate can be described with a sum of the fixed inner flow rate I Q supplied by a syringe pump, and the additional flow rate, which is the derivative of volume change () Vt over time t , expressed as equation (S7). ( ) ( ) cos(2 )