In this section, we introduce digital microfluidic biochips, which is the target architecture of this dissertation. The droplet-based microfluidic biochips discussed in this dissertation are based on the electrowetting on dielectric (EWOD) tech-nique [16], which is one of the most commonly used electrical method for droplet manipulation [49]. EWOD is the actuation method by applying electric field to the fluid-solid interface and changes the tension of this interface based on dc (or low-frequency ac) voltages [11]. One of the advantages of EWOD is high droplet movement speed [16]. Figure 1.4 shows a general digital microfluidic biochip
de-Droplet
Bottom plate Top plate Ground electrode
Control electrodes Hydrophobic
Insulation
Filter fluid
(a)
Droplet Spacing
(b)
Figure 1.5: Cross-section (a) and top (b) views of a general digital microfluidic biochip.
Control wire
Figure 1.6: Illustration of the control wire connection of a digital microfluidic biochip.
veloped by the Duke University. Figure 1.5 (a) shows the cross-section view and Figure 1.5 (b) shows the top view of a biochip based on the principle of EWOD.
There are two plates in a biochip. The top plate contains one ground electrode and the bottom plate contains a set of control electrodes. Droplets containing biological samples and the filter fluid, such as the silicon oil, are sandwiched between the two plates. Note that there is a small spacing between two adjacent cells, as shown in Figure 1.5 (b). With this small spacing, a droplet can move smoothly with a low driving potential. This small spacing also provides routing paths for control wires to connect from external circuits to each electrode.
To move a droplet, a control voltage is applied to one adjacent electrode of this droplet and the electrode under the current position of this droplet is deactived at the same time. A cell is said to be activated if a high voltage is applied to the corresponding control electrode on the bottom plate. An electrical field is generated on an activated cell, and it results an interfacial tension gradient across the spacing between the adjacent electrodes, which causes the transportation of droplets. Based on this method, each droplet can be independently controlled by the electrohydro-dynamic forces generated by an electric field. A droplet can move along a linear array by sequentially activating the electrodes in this linear array. The velocity of a droplet can be up to 20 cm/s [16]. With this method, droplets can move, under user’s control and a system clock, to anywhere of a biochip without micropumps used by the continuous-flow biochips.
The general digital microfluidic biochips allow the individual activation of each cell, and therefore offers a high flexibility of droplet movement. To achieve this, each cell is connected by a control wire from control systems outside a biochip,
Droplet Droplet
Bottom electrodes X direction
Y direction Top electrodes
(a)
Top electrode V
Bottom electrode
Droplet
X Y
(b)
Figure 1.7: Cross-section (a) and top (b) views of a cross-referencing biochip.
as illustrated in Figure 1.6. However, one disadvantage of this type of architecture is that the number of control wires is rapidly increased as system size increases. The large number of control wires affects production cost and increases the complexity of wire routing problem. Therefore, this type of architecture is only suitable for small-scale biochips [58].
Recently, a more scalable biochip architecture is proposed [18]. This ar-chitecture uses a row/column addressing scheme, where a set of electrodes in one row/column is connected to a control pin. We refer to this type of biochip architec-ture as cross-referencing biochips. Figure 1.7 shows the cross-section and top views of a cross-referencing biochip while Figure 1.8 illustrates control wire connection.
A droplet is sandwiched between two plates. A set of electrodes spans a full row (column) in the X-dimension (Y -dimension), and is assigned either a driving or reference voltage. Two sets of electrodes are orthogonally placed, one set each on
Control wire
Figure 1.8: Illustration of the control wire connection of a cross-referencing microflu-idic biochip.
the top and bottom plates as shown in Figure 1.7, and each electrode can be set to either a high or a low voltage. A grid point is “addressed” if there is a potential difference between the upper and lower electrode, i.e., one is high while the other is low. This causes a droplet at a neighboring grid point to move into this location.
The advantage of cross-referencing biochips is that now the number of control wires is proportional to the perimeter of a biochip, instead of its area. Therefore, the production and fabrication costs are reduced.
Due to the flexible, scalable, and dynamically reconfigurable architecture, digital microfluidic biochips have applied to many important applications. For ex-ample, on-chip assays to determine the concentration of target analytes are a natural application of digital microfluidic biochips, such as the in-vitro measurement of glu-cose, lactate, glutamate, and pyruvate in human fluids is very important in clinical diagnosis. Recently, the feasibility of performing a colorimetric enzyme-kinetic glu-cose assay on a digital microfluidic biochip has been demonstrated [45]. Another emerging application area is the clinical diagnostics, especially the point-to-case di-agnostics of diseases [40, 44] Digital microfluidic biochips can also be applied to the environmental monitoring. A scanning-based method to measure airborne particles based on the EWOD method was introduced in [67]. A droplet is used to sweep airborne particles on the filter membrane. Besides, on-chip sequencing by synthe-sis is also an important application. The number of bases in Genbank is increased exponentially. Therefore, a scalable method is needed to handle the exponential growth of the sequence information. Many technologies based on digital microfluidic biochips have been proposed to reduce reagent cost, such as mass spectrometry [26]
and sequencing-by-hybridization [7].
As the number of applications of digital microfluidic biochips increases, the design complexity of digital microfluidic biochips is expected to be significant due to the need of multiple and concurrent assays on a biochip. For example, a prototype has been developed for pyrosequencing with simultaneous execution of 106 fluidic operations [1]. Another example is a lab-on-a-chip system which is designed for protein crystallization with the requirement of concurrent execution of hundreds of operations [59]. A commercial lab-on-a-chip with 600,000 20 µm by 20 µm electrodes is available [3]. For those large systems, traditional full-custom design methodol-ogy is no longer suitable. Therefore, a pressing need to incorporate the support of computer-aided design (CAD) techniques to deliver high-quality design has been raised. Moreover, as pointed out by the International Technology Roadmap for Semiconductors (ITRS) in 2003 [25], the next-generation system-on-a-chip (SoC) is expected to integrate microfluidic based biochips with microelectronic components.
Therefore, in this dissertation, we focus on the synthesis of biochips, including mod-eling, placement, and routing.