Project 2–2: Micro- and Nanofluidic Devices for Sorting and Sensing Biowarfare Agents and Engineering Blood Additives

Principal Investigators: Eric Shaqfeh and Eric Darve (Stanford University)

Microfluidics is the study of suspensions and solutions flowing through channels barely large enough for particles to pass. Micro- fluidic models provide valuable insights on designing portable devices for sensing and identifying biological warfare agents (BWAs), as well as for developing medical treatments for traumatic injuries. Many lab-on-a-chip devices contain fluid channels just wide enough to accommodate a few large molecules such as DNA. Some molecular “bar code” devices contain capillaries that force submicron objects to pass through a channel in single file. Human capillaries can be narrower than a single red blood cell, forcing cells to deform in order to squeeze through.

When the width of a channel is roughly the same size as the particles flowing through it, conventional fluid dynamics principles no longer apply. Essentially all of the particles are in contact with the walls of the channel, leaving no area of free flow in the center. Particle shapes and channel dimensions and geometry take on increasing importance. Irregularly-shaped or elongated particles can form snags and obstructions, like logjams in a stream. Particles adhering to the channel walls (blood clots, for instance) affect the characteristics of the flow. If the particles interact electrostatically with the channel walls or with each other, this adds additional complexity.

Eric Shaqfeh (professor of chemical and mechanical engineering) and Eric Darve (assistant professor of mechanical engineering) are members of Stanford University’s Institute for Computational and Mathematical Engineering. Their work for AHPCRC expands existing modeling and simulation capabilities to construct methods that realistically reproduce and predict the behavior of particles in a microfluidic stream under a wide variety of novel conditions. Their models take into account factors such as electrical forces, flexible or rigid particles of various shapes, Brownian (random) motion, and sedimentation effects.

At present, no computational simulation techniques exist that include all of these factors in the same package. The simulation codes under development by Shaqfeh and Darve handle orientable objects in a flow with hydrodynamic interactions. They are adding and integrating capabilities for complex microfluidic environments and particle shapes, deformable particle surfaces, complex interactions between the solid and liquid phases, and adhesion effects. Other factors in the model include electrical charge effects, mean flow, channel wall interactions, and sedimentation.

Sensing Devices

Electrostatic interactions, channel geometries, and wall adhesion effects can be put to use in sorting and sensing various types of particles, including proteins and viruses. Indeed, several compact field testing devices on the market now use microfluidic channels to detect and identify substances of interest, using length, shape, electrical charge, and chemical affinities as identifying characteristics.

Numerical simulations of particles in microchannels imitate their real-life counterparts—the virtual particles start at a simulated inlet well and are driven electrokinetically through a simulated channel, where they are separated by size or electrical charge, adhere to the walls according to predetermined interaction potentials, or separate into different channels. In the laboratory, real particles would be sent to a sensing area where various analytical devices could identify them and determine their concentrations. In a simulation environment, numerical methods are used to perform tracking and measuring functions. The success of the simulation method is determined by how well it emulates laboratory observations for known systems, and how useful it is for predicting the behavior of new systems.

All relevant interactions among the particulates and with the micro-device walls must be included in a simulation in order to produce realistic predictions of particle behavior. Particles can separate or form aggregates with each other. Layers of particles can build up on the channel walls. In addition to electrokinetic forces, shear, drag, and Brownian motion influence the overall flow.

Capabilities already exist for simulating sedimentation and Brownian motion for rod-like particles and semi- flexible molecules such as DNA. Under the AHPCRC program, Shaqfeh and Darve have completed their computer code for planar channel flow simulations, and this code is being benchmarked and adapted to parallel computing environments. The existing models account for hydrodynamic interactions and the volume excluded by the particles, and the models can use electro-osmosis or pressure as the driving force. The AHPCRC group has used their models to predict nonuniform concentration profiles across small channels as a result of variations in osmotic pressure. They are determining whether this effect can directly separate particles based on length by comparing their simulations with microfluidic experiments conducted by Dr. Samir Mitragotri, an associate professor of chemical engineering at the University of California at Santa Barbara who is an active member of the Institute for Collaborative Biotechnologies (ICB).

The simulation capabilities are being expanded to include complex particle shapes and channel geometries. Dynamic simulations are in progress for nanochannel separations of short double-stranded DNA (up to 100 base pairs). The simulations include centerof- mass and rotational Brownian motion, molecular interaction with the electronic double layer potential (a type of wall interaction effect), and electro-osmotic convection. Scalable algorithm development for the calculations of the long-range interactions (electric and hydrodynamic) is essential in these simulations. The AHPCRC group is focusing on comparison of these simulations with the experimental measurements and modifying them for flow through more complex microfluidic geometries, including multiple channels, T-junctions, and bifurcations (forks in the channel).

Blood circulation

Blood vessels may be thought of as microfluidic channels, and studies of particles in small channels may be applied to the movements of particles in the bloodstream. Such particles include cells, blood additives (oxygen carriers and reconstituted or artificial platelets), particulate (non-dissolved) drugs, and drug delivery agents. Particles may be spherical or rod-shaped, rigid or flexible, and electrically charged or neutral. Particles may be driven through the channels using pressure (e.g., heartbeats). An electric field or electrostatic charges may drive the motion of the particles (electrophoresis) or the surrounding fluid (electro-osmosis), a set of processes known collectively as electrokinesis.

Army medical researchers are especially interested in developing methods for processing human blood platelets so that they can be freeze-dried for storing and shipping, then reconstituted in the field for the treatment of serious wounds (see “The Push for Platelets”). Current techniques for lyophilization (freeze-drying) require a multi-step procedure involving chemical fixers to keep the platelets from agglomerating, and reconstituted platelets are not as effective as platelets in their native state. Despite these factors, lyophilization offers significant advantages in remote field clinics. Recently, the Stanford group completed its first computer simulations of platelet and red blood cell microcirculation behaviors. These simulations show that, at the high shear rates typical for small vessels, red blood cells gradually move toward the center of the blood vessel, forcing the platelets toward the vessel walls, in good agreement with experimental observations.

Source: AHPCRC Bulletin, Vol. 1 No. 4 (2009)