2.3 Hardware and Software Setup
17
(i.e., one-by-one upon request). In the following, these two categories for the droplet
generation are outlined:
• Droplet trains: An often applied device for incessantly producing droplet trains
are T-junctions. In a T-junction, the main channel containing the continuous
phase perpendicularly intersects with the channel containing the dispersed phase.
As can be seen in Fig. 1.1 (cf. page 4), the dispersed phase enters the main
channel and the main stream breaks the dispersed phase into droplets. The size
of droplets depends on the volumetric flow rates, geometries, and viscosities.
Further approaches and details for incessantly producing droplet trains are
reviewed in [55].
• Droplet on Demand: When droplets need to be injected one-by-one upon request
(i.e., at dedicated times),the so-called droplet-on-demand is required. To this
end, the droplets can be generated with internal valves [33, 119, 130], external
valves [17], electric fields [83, 84], or with pressure pulses [57, 118, 122].
Example 2.4 Let’s again consider the microfluidic network shown in Fig. 2.1a.
Here, a T-junction is used to generate droplet trains. Therefore, both the continuous
and the dispersed phase are driven by a pump. Overall, the droplet generation
should lead to well-controlled droplet sizes.
In order to fabricate a microfluidic network in a device, a large variety of
materials and techniques can be used. Especially the selection of the materials
used for the device constitutes an important step [117], i.e. the material needs
to allow the formation of droplets and needs to be adequate for the conducted
(bio)-chemical experiment. Prototypes are most commonly produced in polydimethylsiloxane (PDMS) using a soft-lithography process because PDMS allows
for a rapid and low-cost fabrication, is biocompatible for many experiments, and is
transparent [74]. But also other materials like stiff polymers (polycarbonate (PC),
cyclic olefin copolymer (COC), polymethylmethacrylate (PMMA), polytetrafluoroethylene (PTFE), fluorinated ethylene propylene (FEP)), glass, or silicone [68]
can be used in combination with chemical modifications to get desired channel
surfaces. Furthermore, the fabrication technique can also be 3D-printing, injection
molding, or milling.
2.3 Hardware and Software Setup
All design methods proposed in the following have been implemented and experimentally evaluated. For the evaluation, a machine with an Intel Core i7-7700K (at
4.2 GHz) having 32 GB of main memory and running 64-bit Ubuntu 16.04.4 LTS
has been used. All proposed methods, except the method presented in Chap. 5,
have been implemented and evaluated in Java (using OpenJDK in version 1.8.0_181
and a 64-Bit Server VM) and Matlab (using version R2016b). For the online tool
presented in Chap. 5, JavaScript in combination with the Bootstrap library has been
used. Furthermore, the SMT solver Z3 [25] in version 4.6.0 has been applied for the
methods presented in Chaps. 7 and 8.
17
(i.e., one-by-one upon request). In the following, these two categories for the droplet
generation are outlined:
• Droplet trains: An often applied device for incessantly producing droplet trains
are T-junctions. In a T-junction, the main channel containing the continuous
phase perpendicularly intersects with the channel containing the dispersed phase.
As can be seen in Fig. 1.1 (cf. page 4), the dispersed phase enters the main
channel and the main stream breaks the dispersed phase into droplets. The size
of droplets depends on the volumetric flow rates, geometries, and viscosities.
Further approaches and details for incessantly producing droplet trains are
reviewed in [55].
• Droplet on Demand: When droplets need to be injected one-by-one upon request
(i.e., at dedicated times),the so-called droplet-on-demand is required. To this
end, the droplets can be generated with internal valves [33, 119, 130], external
valves [17], electric fields [83, 84], or with pressure pulses [57, 118, 122].
Example 2.4 Let’s again consider the microfluidic network shown in Fig. 2.1a.
Here, a T-junction is used to generate droplet trains. Therefore, both the continuous
and the dispersed phase are driven by a pump. Overall, the droplet generation
should lead to well-controlled droplet sizes.
In order to fabricate a microfluidic network in a device, a large variety of
materials and techniques can be used. Especially the selection of the materials
used for the device constitutes an important step [117], i.e. the material needs
to allow the formation of droplets and needs to be adequate for the conducted
(bio)-chemical experiment. Prototypes are most commonly produced in polydimethylsiloxane (PDMS) using a soft-lithography process because PDMS allows
for a rapid and low-cost fabrication, is biocompatible for many experiments, and is
transparent [74]. But also other materials like stiff polymers (polycarbonate (PC),
cyclic olefin copolymer (COC), polymethylmethacrylate (PMMA), polytetrafluoroethylene (PTFE), fluorinated ethylene propylene (FEP)), glass, or silicone [68]
can be used in combination with chemical modifications to get desired channel
surfaces. Furthermore, the fabrication technique can also be 3D-printing, injection
molding, or milling.
2.3 Hardware and Software Setup
All design methods proposed in the following have been implemented and experimentally evaluated. For the evaluation, a machine with an Intel Core i7-7700K (at
4.2 GHz) having 32 GB of main memory and running 64-bit Ubuntu 16.04.4 LTS
has been used. All proposed methods, except the method presented in Chap. 5,
have been implemented and evaluated in Java (using OpenJDK in version 1.8.0_181
and a 64-Bit Server VM) and Matlab (using version R2016b). For the online tool
presented in Chap. 5, JavaScript in combination with the Bootstrap library has been
used. Furthermore, the SMT solver Z3 [25] in version 4.6.0 has been applied for the
methods presented in Chaps. 7 and 8.