1. Continuous-flow bioseparation for sample preparation

There are four highly-desirable benefits associated with continuous-flow preparative separation (spatial separation) when compared with one-dimensional analytical separation (temporal separation):

1) Increased sample throughput ideal for sample preparation based on micro/ nanosystems. Most micro/ nanofluidic systems can only process low quantities of samples due to their small handle volumes. The continuous-flow operation can remove the limitation of the amount of sample the device can analyze, and the fractionated biomolecules can be continuously collected and therefore accumulate over time.     

2) Fractionated biomolecule streams can be easily collected for downstream analysis or subsequent manipulation. By virtue of the continuous flow operation, the fractionated biomolecule streams can be easily recovered or routed to different downstream reaction chambers or detection channels for further analysis. Therefore, the continuous-flow separation scheme is ideal for a highly integrated microanalysis system that includes multiple analytical steps and separation channels and reaction chambers.    

3) Continuous-flow separation permits continuous harvesting of the subset of biomolecules of interest to enhance the specificity and sensitivity for downstream biosensing and detection, whish is highly desirable for integrated bioanalysis microsystems. Operation in continuous flow can enable the integration of the signal over time or to collect the sample over time, thus decreasing the downstream detection limit. This advantage may prove useful for preparative separation of complex biological samples (such as human serum), which has promising implications for proteomic research and biomarker discovery.

4) In one-dimensional analytical separation systems, separation speed and resolution are normally controlled and mediated by a single force field applied along the direction of the separation column (for example, electrostatic force field for capillary electrophoresis systems and hydrodynamic force field for high performance liquid chromatography systems (HPLC)). Therefore, the separation speed and resolution in one-dimensional separation systems are coupled with each other and can often complicate the optimization process of the separation system. The one-dimensional nanofilter arrays we developed is a clear example for such a complication. In the one-dimensional nanofilter arrays, the separation speed and resolution cannot both be enhanced without compromising one another. While in continuous-flow preparative separation systems, the separation speed and resolution are decoupled in two orthogonal directions, and are modulated, respectively, by two independent orthogonal force fields. Therefore, careful regulation of both the force fields in the two orthogonal directions can help achieve rapid separation with concurrent high resolution. 

Figure 1 Continuous-flow separation of different-sized biomolecules (large biomolecule in Green, small biomolecule in Red) in two-dimensional isotropic (A) and anisotropic (B) sieving media. For continuous-flow separation, the vectorial direction of the mobility m determines the trajectory (and separation), and the absolute magnitude of m determines the migration speed. (A) No separation is achieved in an isotropic sieving medium. Different-sized biomolecules follow the same trajectory that is solely determined by the two independent orthogonal electric fields Ex and Ey. (B) In an anisotropic sieving medium, different-sized biomolecules follow different trajectories that are determined by both the electric fields as well as by certain molecular properties (for example, size). The stream deflection angle ¦È is defined with respect to the y-axis.

2. Anisotropic sieving structure: a new paradigm for continuous-flow bioseparation

Continuous-flow separation is highly desirable as we discussed in the previous section. However, in conventional biomolecule separation systems that use a random isotropic sieving medium in their separation channels and chambers (such as gel, liquid gel or ampholytes), continuous-flow separation is not readily possible. Figure 1A shows an example of an isotropic sieving medium that contains a two-dimensional random gel. Upon application of two uniform orthogonal electric fields Ex and Ey in the gel, two different-sized biomolecules can be continuously injected into the gel and form straight molecular steams. Due to the sieving property of the gel, the mobility of two different-sized biomolecules can be different (here we simply assume the smaller red molecule and larger green molecule have mobility of m 1 and m 2, respectively, and m 1 > m 2). The stream deflection angle θ in a two-dimensional sieving medium can be calculated using a simple expression

tanθ = Vx/Vy = (m x/m y)(Ex/Ey)                                                                                             (1)

where Vx and Vy are the migration velocities along the x- and y-axis, respectively, and m x and m y are the two orthogonal mobility along the x- and y-axis, respectively. In an isotropic sieving medium, the mobility is isotropic in nature, thus, in Eq. (1), m x and m y simply cancel out, and the expression of the stream deflection angle θ becomes

tanθ = Ex/Ey                                                                                                                       (2)

Therefore, in an isotropic sieving medium, the stream defection angle θ is solely determined by the two electric fields Ex and Ey, and different-sized biomolecules will follow the same trajectory without separation (Fig. 1A). Here we need to point out that the mobility in the two-dimensional gel is still size-dependent, therefore, the migration speeds of the two different-sized biomolecules will be different.

The structural anisotropy in an anisotropic sieving medium can cause molecules of different properties (for example, size, charge, and hydrophobicity) to follow different migration trajectories, leading to efficient separation. In an anisotropic medium (Fig. 1B), the function of m x/m y in Eq. (1) becomes more complex, and could be a function of both the structural anisotropy of the sieve and molecular properties. Therefore, the expression of the stream deflection angle θ can be modified based on Eq. (1) as

tanθ = Vx/Vy = (m x/m y)(Ex/Ey) = func(size, charge, etc.)(Ex/Ey)                                              (3)

Thus, molecules of different properties will follow different trajectories in the anisotropic sieving medium, if such properties cause the values of the function func(size, charge, etc.) to be dependent on molecular properties. In Fig. 1B, we assume the mobility along the y-axis m y is size-independent, and the mobility along the x-axis m x is size-dependent. Therefore, in Fig. 1B, smaller Red molecule will display a larger deflection angle θ than the larger Green one.  The unknown function f(size, charge, etc.) can be difficult to determine, especially when this function is associated with both molecular properties and the structural anisotropy of the sieve.   

3. A patterned anisotropic nanofilter array (ANA): device design and fabrication

In this section, we will introduce a unique molecular sieving structure design, called the anisotropic nanofilter array (anisotropic nanofilter array, ANA), and its implementation for continuous-flow separation of DNA and proteins based on size. The designed structural anisotropy of the ANA is critical to continuous-flow separation, which is not readily possible with a random isotropic sieving medium. Moreover, the ANA allows for various sieving mechanisms (e.g., Ogston sieving and entropic trapping) to take effect in the separation of biomolecules in very broad biological size ranges.   

The design of the ANA consists of a two-dimensional periodic nanofilter array (Fig. 2). The separation mechanism of the ANA relies on different sieving characteristics along two orthogonal directions within the ANA, which are set perpendicular and parallel to the nanofilter rows (indicated as x- and y-axis, respectively, in Fig. 2). Upon application of an electric field Ey along the positive y-axis, uniformly negative-charged biomolecules injected into the array assume a drift motion in deep channels with a negative velocity Vy that is size-independent. An orthogonal electric field Ex is superimposed along the negative x-axis across the nanofilters, and this field selectively drives the drifting molecules in the deep channel to jump across the nanofilter in the positive x-direction to the adjacent deep channel. Molecular crossings of the nanofilter under the influence of the electric field Ex can be described as biased, thermally activated jumps across free energy barriers at the nanofilter threshold, and these free energy barriers depend on both steric and electrostatic interactions between charged macromolecules and charged nanofilter walls. At high ionic strength where the Debye length is negligible compared to the nanofilter shallow region depth ds, electrostatic interactions between charged macromolecules and charged nanofilter walls are largely screened. The free energy barriers are therefore solely determined by the configurational or conformational entropy loss within the constriction due to the steric constraints or exclusion of the nanofilter wall (a purely steric limit). For biomolecules with diameters smaller than the nanofilter constriction (i.e., Ogston sieving) (Fig. 2A), the entropic energy barrier favors DNA and proteins with a smaller size for passage, resulting in a greater jump passage rate Px for smaller molecules. Therefore, in Ogston sieving, smaller molecules exercise a shorter mean characteristic drift distance L in the deep channels between two consecutive nanofilter crossings, leading to a larger stream deflection angle θ.

For molecules with diameters greater than the nanofilter constriction size, passage requires the molecules to deform and form hernias at the cost of their internal conformational entropy (i.e., entropic trapping). A previous study on long DNA molecules trapped at a similar nanofluidic constriction showed that the activation free energy barrier for DNA escape depends solely on the inverse of the electric field strength (~1/Ex). Furthermore, longer molecules have a larger surface area contacting the constriction and thus have a greater probability to form hernias that initiate the escape process (in other words, they have a higher escape attempt frequency) (Fig. 2B). Therefore, in the entropic trapping regime, longer molecules assume a greater jump passage rate Px, resulting in a larger deflection angle θ.

Here it is worthy to emphasis that in the ANA, the mean characteristic drift distance L between two consecutive nanofilter crossings plays a determinant role for the migration trajectory, and the stream deflection angle θ is directly related to L with the expression of tanθ = (ld+ls)/L, where ld and ls are the deep channel width and nanofilters length, respectively. The mean characteristic drift distance L is determined by both the complex structural geometry of the ANA and the two independent orthogonal fields Ex and Ey. We will discuss more on the theoretical modeling of the mean characteristic drift distance L and the stream deflection angle ¦È in the rest sections of this chapter.

Figure 2 Schematic showing negatively charged macromolecules assuming bidirectional motion in the ANA under the influence of two orthogonal electric fields Ex and Ey as indicated. Dashed lines and arrows indicate migration trajectories projected onto the x-y plane. Nanofilters (with width of ws, length of ls and depth of ds) arranged in rows are separated by deep channels (with width of ld and depth of dd). Rectangular pillars (with width of wp and length of ls) between nanofilters serve as supporting structures to prevent collapse of the top ceiling. (A) Ogston sieving. Shorter molecules (red) are preferred for passage through the nanofilter due to their greater retained configuration freedom, resulting in a greater nanofilter jump passage rate Px (the inset). The mean drift distance L between two consecutive nanofilter crossings plays a determinant role for the migration trajectory, with a shorter L leading to a larger stream deflection angle ¦È that is defined with respect to the negative y-axis. (B) Entropic trapping. Longer molecules (green) have larger surface area contacting the nanofilter threshold (the inset), resulting in a greater probability for hernia formation and thus a greater nanofilter passage rate Px.

Figure 3 Structure of the microfabricated device incorporating the ANA. Scanning electron microscopy images show details of different device regions (clockwise from top right: sample injection channels, sample collection channels, and ANA). The inset shows a photograph of the thumbnail-sized device. The rectangular ANA is 5 mm × 5 mm, and nanofilters (ws = 1 µm, ls = 1 µm and ds = 55 nm) are spaced by 1 µm × 1 µm square-shaped silicon pillars. Deep channels are 1 µm wide (wd) and 300 nm deep (dd). Injection channels connecting sample reservoir (1 mm from the ANA top left corner) inject biomolecule samples as a 30 µm wide stream.

References

1. Fu, J. & Han, J. "A nanofilter array chip for fast gel-free biomolecule separation," Proceedings of the MicroTAS 2005 Symposium, Boston, MA, vol. 2, pp. 1531-1533. (pdf)
2. Fu, J. & Han, J. “Continuous biomolecule separation in a nanofilter,” United States Provisional Patent, MIT, Oct. 2005.
3. Fu, J. & Han. J. “Patterned periodic nanofilter array for continuous-flow biomolecule separation,” American Physical Society National March Meeting, Baltimore, Maryland USA, March 2006. (pdf)
4. Fu, J. & Han, J. "Continuous-flow biomolecule separation through patterned anisotropic nanofluidic sieving structure," Proceedings of the MicroTAS 2006 Symposium, Tokyo, Japan, vol. 1, pp. 519-521. (pdf)