We present a study of the electrophoresis of long DNA in a strong electric field through a hexagonal array of cylindrical microscale posts spaced such that the pore size is commensurate with equilibrium coil size of the DNA. Experimental mobility, dispersivity, and videomicroscopy data indicate that the DNA frequently collide with the posts, contradicting previous Brownian dynamics studies using a uniform electric field. We demonstrate via simulations that the frequent collisions, which are essential to separations in these devices, are due to the nonuniform electric field, highlighting the importance of accounting for electric-field gradients when modeling DNA transport in microfluidic devices.

Several continuous-time random walk (CTRW) models exist to predict the dynamics of DNA in micropost arrays, but none of them quantitatively describes the separation seen in experiments or simulations. In Part I of this series, we examine the assumptions underlying these models by observing single molecules of λ DNA during electrophoresis in a regular, hexagonal array of oxidized silicon posts. Our analysis takes advantage of a combination of single-molecule videomicroscopy and previous Brownian dynamics simulations. Using a custom-tracking program, we automatically identify DNA-post collisions and thus study a large ensemble of events. Our results show that the hold-up time and the distance between collisions for consecutive collisions are uncorrelated. The distance between collisions is a random variable, but it can be smaller than the minimum value predicted by existing models of DNA transport in post arrays. The current CTRW models correctly predict the exponential decay in the probability density of the collision hold-up times, but they fail to account for the influence of finite-sized posts on short hold-up times. The shortcomings of the existing models identified here motivate the development of a new CTRW approach, which is presented in Part II of this series.

The "channeling hypothesis" of DNA electrophoresis in sparse, ordered arrays of posts predicts that the DNA will move through the array relatively unhindered if ͑i͒ the spacing between the posts is larger than the DNA coil and ͑ii͒ the electric field lines are straight. We tested this hypothesis by studying the electrophoretic separation of a small plasmid DNA ͑pUC19, 2686 base pairs͒ and a large, linear DNA ͑-DNA, 48 500 base pairs͒ in a hexagonal array of 1 m diameter posts with a pitch of 7 m. At low electric field strengths, these DNAs are separated due to the long-lived, rope-over-pulley collisions of -DNA with the posts. The resolution is lost as the electric field increases due to the onset of channeling by the -DNA. Using a diffusive model, we show that channeling arises at low electric fields due to the finite size of the array. This channeling is not intrinsic to the system and is attenuated by increasing the size of the array. Higher electric fields lead to intrinsic channeling, which is attributed to the disparate time scales for a rope-over-pulley collision and transverse diffusion between collisions. The onset of channeling is a gradual process, in agreement with extant Brownian dynamics simulation data. Even at weak electric fields, the electrophoretic mobility of -DNA in the array is considerably higher than would be expected if the DNA frequently collided with the posts.

Separating long DNA in a microfabricated post array requires (tens of) thousands of posts in the separation channel. Moving from microposts to nanoposts is thus a fabrication challenge owing to the large area that needs to be nanopatterned. The authors implemented an oxygen plasma etching method in conjunction with conventional optical photolithography and deep trench etching that led to centimeter-long microchannels containing either 360 or 460 nm diameter posts in a hexagonal array with a 3 μm spacing. Separations of the XhoI λ-DNA digest in the device indicate that these sparse nanopost arrays are an improvement over the equivalent micropost array with only a marginal increase in fabrication cost. The fabrication method described here is broadly applicable to biological microfluidic and nanofluidic platforms that require nanoscale features with micrometer-scale spacing.

We present experimental results for the collision of a single λ-DNA molecule with isolated cylindrical nanoposts whose radii range from 350 nm to 1.37 μm. The experimental apparatus was fabricated in oxidized silicon using a plasma-thinning method. The probability density for the dimensionless holdup time is described by a gamma distribution, which accurately represents the behavior of the probability density for X collisions and U/J collisions for short-lived and long-lived collisions, respectively. The shape parameter of the gamma distribution reflects the nature of the short-lived collisions, whereas the scaling parameter captures the role of the finite size of the post on the U/J collisions. When rendered in an appropriate dimensionless form, the existing models for the role of electric field gradients extend to all post sizes, indicating that 350 nm is still well above the point-sized limit. Our experiments provide insights into a regime that is very challenging for numerical simulations but highly relevant for separation processes.