Hardware Reference
In-Depth Information
Tabl e 4. 1
An example of
module library for
dilution/mixing operations [
6
]
Size of mixing module
Completion time (s)
2
4-array
2
2
3-array
4
2
2-array
6
1
4-linear array
3
In order to overcome the drawbacks associated with characterization, biochips
integrated with sensing systems, i.e., cyberphysical microfluidic biochips, are being
developed [
1
,
4
,
9
,
10
]. The feedback provided by the sensing system enables
real-time concentration checking, error detection, and error correction for fluidic
operations [
1
]. Therefore, essential operations such as droplet dispensing and mix-
ing can be precisely implemented on cyberphysical microfluidic biochips without
the need for specifying a module library or characterizing a bioassay [
1
,
9
]. However,
today's synthesis algorithms for mapping biochemistry protocols to the chip still
rely on characterization procedures for bioassays [
11
-
13
]. Hence the advantages of
cyberphysical integration are not fully exploited, and precious samples/reagents and
time are consumed and wasted during characterization.
In order to map the protocol of a bioassay to the detailed implementations of
fluidic operations on the biochip, bioassay synthesis algorithms need to provide all
the information required for the execution of the bioassay, including the scheduling
of operations, transportation paths of droplets, module placement of operations, the
pin-assignment configuration of the biochip, and the wire-routing solution for the
biochip. However, current synthesis algorithms cannot provide all the details listed
above due to the following limitations:
1. Prior work is oblivious to variability and uncertainty in biochemical processes.
The competition-time of fluidic operations in practical applications may be
different from the time defined in a module library, thus the accuracy of the
scheduling results for fluidic operations cannot be guaranteed.
2. On-line computation in prior work does not provide information about the paths
of droplet transportation.
To overcome the above drawbacks, we propose a new design flow for cyberphys-
ical microfluidic biochips. The key contributions and benefits of this chapter are as
follows:
1. We propose the design of microfluidic biochips that uses multiple clock fre-
quencies. The execution time of the bioassay can be reduced without additional
degradation of electrodes or hardware cost.
2. In order to handle the uncertainties in completion-time of fluidic operations,
we propose an “operation-interdependence-aware” synthesis algorithm—the first
on-chip biochemistry synthesis procedure that does not use the module library
as a design guideline. Using this algorithm, the characterization process can be
removed. The algorithm results in a design approach that considers completion-
time uncertainties for fluidic operations, hence it leads to the improved accuracy
of fluidic operations.
