Chapter 8
Generating Droplet Sequences
The passive droplet routing mechanism allows to route a payload droplet (containing the biological sample) through different paths of a microfluidic network and,
by this, allows for executing different experiments on a single device. Recall, if
the payload is supposed to take a non-default successor at any bifurcation in the
network, it has to be made sure that another droplet (called header droplet) arrives
before and flows through the default successor (i.e., the header temporarily “blocks”
the default successor for closely following droplets).
In order to establish a routing for the payload, payload and header droplets have
to be injected into the network so that headers block the default successors which
should not be taken by the payload. Hence, headers have to arrive right before the
payload at corresponding bifurcations.
For ring and bus architectures as, e.g., proposed in [11, 24, 27, 60, 80], the
injection time of the header and payload droplets can be calculated by a formula
because of their regular structure. However, this is not the case for more complex architectures. Especially for application-specific architectures as proposed in
Chap. 7 and [48], it is not obvious how many headers are needed and when to inject
these headers relatively to the payload. In other words, what droplet sequence should
be injected.
More precisely, the injected droplet sequence has to ensure that each bifurcation
where a droplet is supposed to take the non-default successor is temporarily blocked
by a header droplet. Furthermore, the time a header requires in order to flow into
the channel to be blocked depends on the flow rates in the microfluidic network.
These flow rates however constantly change due to the resistances caused by the
flow of droplets and make it hard to estimate whether droplets unintentionally
influence their respective ways. These nontrivial interdependencies motivate an
automatic method for the generation of droplet sequences since, in particular for
larger networks, it is infeasible to conduct all considerations manually.
Exactly these flow rate changes and interdependencies are described by the 1D
analysis model reviewed in Sect. 3.2. This model allows to determine the speeds and,
© Springer Nature Switzerland AG 2020
A. Grimmer, R. Wille, Designing Droplet Microfluidic Networks,
https://doi.org/10.1007/978-3-030-20713-7_8
99
Generating Droplet Sequences
The passive droplet routing mechanism allows to route a payload droplet (containing the biological sample) through different paths of a microfluidic network and,
by this, allows for executing different experiments on a single device. Recall, if
the payload is supposed to take a non-default successor at any bifurcation in the
network, it has to be made sure that another droplet (called header droplet) arrives
before and flows through the default successor (i.e., the header temporarily “blocks”
the default successor for closely following droplets).
In order to establish a routing for the payload, payload and header droplets have
to be injected into the network so that headers block the default successors which
should not be taken by the payload. Hence, headers have to arrive right before the
payload at corresponding bifurcations.
For ring and bus architectures as, e.g., proposed in [11, 24, 27, 60, 80], the
injection time of the header and payload droplets can be calculated by a formula
because of their regular structure. However, this is not the case for more complex architectures. Especially for application-specific architectures as proposed in
Chap. 7 and [48], it is not obvious how many headers are needed and when to inject
these headers relatively to the payload. In other words, what droplet sequence should
be injected.
More precisely, the injected droplet sequence has to ensure that each bifurcation
where a droplet is supposed to take the non-default successor is temporarily blocked
by a header droplet. Furthermore, the time a header requires in order to flow into
the channel to be blocked depends on the flow rates in the microfluidic network.
These flow rates however constantly change due to the resistances caused by the
flow of droplets and make it hard to estimate whether droplets unintentionally
influence their respective ways. These nontrivial interdependencies motivate an
automatic method for the generation of droplet sequences since, in particular for
larger networks, it is infeasible to conduct all considerations manually.
Exactly these flow rate changes and interdependencies are described by the 1D
analysis model reviewed in Sect. 3.2. This model allows to determine the speeds and,
© Springer Nature Switzerland AG 2020
A. Grimmer, R. Wille, Designing Droplet Microfluidic Networks,
https://doi.org/10.1007/978-3-030-20713-7_8
99