3.4 Case Study
41
Here, some dimensions are given by the application or by fabrication limitations,
e.g.:
• The droplets’ size and spacing determine the trapping well radius r and the length
L 1 , i.e. r is chosen so that a droplet fills most of the trap and L 1 is chosen so that
following droplets do not contact with the trapped droplet.
• The fabrication limits the gap width w gap and gap length L 2 (i.e., the pillars
require a minimum size, otherwise they could be peeled off from the silicon
wafer).
• The properties of the used phases are chosen so that they are suited for forming
droplets and for the application.
However, besides that, all remaining issues such as the length of the bypass
channel or the applied pressures have to be explicitly specified by the designer.
This results in a tedious task because even small microfluidic networks are
composed of several channels which, together with the droplets, yield to plenty
of interdependencies. In fact, the authors of [13] spent a significant portion of their
work discussing the respective issues in detail, which is why readers are referred to
this work for a more detailed treatment.
Overall, deriving the specification requires the definition and consideration
of plenty of variables, which all affect the intended behavior (i.e., whether the
objectives are fulfilled). In this task of deriving the specification, the designer relies
on his/her expert knowledge and often applies simplifications and assumptions.
Especially the time-dependent resistances caused by droplets are often simplified
or completely ignored when deriving the specification as it is impossible to consider
all droplet states and positions by hand. For example, in [13] only a fixed number of
droplets contained in the bypass are assumed, which cause an additional resistance.
Prototyping and Testing
The task reviewed above eventually yields a specification which is supposed
to realize the intended behavior. However, due to applied simplifications and
assumptions, the designer cannot be sure whether an implementation based on the
specification indeed realizes the desired behavior and whether all objectives are
fulfilled under all settings.
Therefore, as a next step, it is tested whether the specification realizes the desired
functionality by using (physical) experiments. To this end, the designer fabricates
the design based on the derived specification. This first requires a physical design
(also called layout or mask) to be drawn from the specification (e.g., as a vector
graphic), which can be used as input for the production process. Afterwards, this is
used to fabricate the device using, e.g., a soft-lithography technique, 3D-printing,
or milling. Then, the resulting device is validated, i.e. experiments are conducted to
check whether the device indeed shows the desired behavior. This is the first time in
the design process in which the designer can observe the effects of his/her choices
41
Here, some dimensions are given by the application or by fabrication limitations,
e.g.:
• The droplets’ size and spacing determine the trapping well radius r and the length
L 1 , i.e. r is chosen so that a droplet fills most of the trap and L 1 is chosen so that
following droplets do not contact with the trapped droplet.
• The fabrication limits the gap width w gap and gap length L 2 (i.e., the pillars
require a minimum size, otherwise they could be peeled off from the silicon
wafer).
• The properties of the used phases are chosen so that they are suited for forming
droplets and for the application.
However, besides that, all remaining issues such as the length of the bypass
channel or the applied pressures have to be explicitly specified by the designer.
This results in a tedious task because even small microfluidic networks are
composed of several channels which, together with the droplets, yield to plenty
of interdependencies. In fact, the authors of [13] spent a significant portion of their
work discussing the respective issues in detail, which is why readers are referred to
this work for a more detailed treatment.
Overall, deriving the specification requires the definition and consideration
of plenty of variables, which all affect the intended behavior (i.e., whether the
objectives are fulfilled). In this task of deriving the specification, the designer relies
on his/her expert knowledge and often applies simplifications and assumptions.
Especially the time-dependent resistances caused by droplets are often simplified
or completely ignored when deriving the specification as it is impossible to consider
all droplet states and positions by hand. For example, in [13] only a fixed number of
droplets contained in the bypass are assumed, which cause an additional resistance.
Prototyping and Testing
The task reviewed above eventually yields a specification which is supposed
to realize the intended behavior. However, due to applied simplifications and
assumptions, the designer cannot be sure whether an implementation based on the
specification indeed realizes the desired behavior and whether all objectives are
fulfilled under all settings.
Therefore, as a next step, it is tested whether the specification realizes the desired
functionality by using (physical) experiments. To this end, the designer fabricates
the design based on the derived specification. This first requires a physical design
(also called layout or mask) to be drawn from the specification (e.g., as a vector
graphic), which can be used as input for the production process. Afterwards, this is
used to fabricate the device using, e.g., a soft-lithography technique, 3D-printing,
or milling. Then, the resulting device is validated, i.e. experiments are conducted to
check whether the device indeed shows the desired behavior. This is the first time in
the design process in which the designer can observe the effects of his/her choices