Microfluidic Free Flow Electrophoresis (μFFE) – What would you do with a separation if you didn't need to wait?

Michael T. Bowser
Department of Chemistry, University of Minnesota, Minneapolis, MN
Email: bowser@umn.edu
Web: http://www.chem.umn.edu/groups/bowser/

Time is never far from a separation scientist's thoughts. Time is a currency that can be sacrificed for resolution, or vice versa. Entire workflows (and coffee break schedules) revolve around the time to produce an analysis and the number of analyses that can be performed in a day. Considering the number of separations performed in a high throughput lab, even shaving a minute off a routine analysis could save thousands of dollars. So after a career of waiting (albeit across a wide range of time scales) it is almost difficult for the separation scientist in me to consider the question: What if time didn't matter?

Free Flow Electrophoresis.

Free Flow Electrophoresis is one of a limited number of separation techniques capable of separating a continuous stream of analyte. The separation mechanism is shown schematically in Figure 1. Analyte enters a planar separation channel in a continuous flow stream. An electric field is applied perpendicularly to the direction of flow, deflecting charged analytes as they travel through the flow chamber. Note that the separation occurs in space, not time. Transit time through the device is the same for all analytes as determined by the linear velocity of the buffer flow. FFE is not a new technique. Conventional macroscale FFE was first demonstrated over 50 years ago.1 The dimensions and flow volumes of these instruments make FFE more amenable to preparative separations of proteins, organelles and cells than quantitative applications. The relatively large dimensions also make precise control of flow paths and detector integration challenging. As such, conventional FFE has remained a niche technique.

Microfluidic Free Flow Electrophoresis (μFFE).

FFE was first demonstrated on the microscale in 1994.2,3 Early devices fabricated in silicon or PDMS were limited by restrictions imposed by the substrate material. We first fabricated an all glass device in 2005.4 These early designs were severely limited by electrolysis bubbles generated at the electrodes. Since then designs5 and buffer additives6 have been introduced to mediate the effect of bubble generation, making long term operation of μFFE devices feasible for the first time. The separation theory of μFFE separations has been developed, clearly demonstrating the relationships between electric field, buffer flow rate, band position and resolution.7 Free zone, MEKC, isoelectric focusing and isotachophoresis separations have all been demonstrated on μFFE devices.5 μFFE has finally emerged as a viable technique. The question is: how best to take advantage of the continuous nature of μFFE separations? Our efforts have focused on two areas where μFFE outperforms existing separation strategies: high speed monitoring and microscale purification.

High Speed Monitoring.

μFFE offers the selectivity of a separation with the time response of a sensor. This suggests a re-evaluation of what applications are best suited for separations vs. those that are better addressed using sensors. Note that while the transit time through the separation chamber is on the order of 10-20 seconds, μFFE is capable of measuring much faster changes. If the concentration of analyte entering the device changes, this change is maintained in the flow stream as it travels through the device. The temporal response of the device is therefore limited by how fast the detector can record images. We routinely record fluorescence images using a CCD camera every 100 ms, but much faster detector frequencies are possible. The high sampling rate of μFFE opens the door to some interesting applications. Numerous measurements can be taken in a relatively short time period, allowing the signal averaging strategies more often used in spectroscopy to be applied. We have demonstrated a 24-fold enhancement in S/N by recording 500 images over a period as short as 2 minutes.8 μFFE is capable of monitoring samples with changing compositions. For example, we introduced a sample where we titrated a fluorescently labeled aptamer with a continuous gradient of increasing target protein concentration.9 This allowed the equilibrium mixture to be assessed at 300 different concentrations in 5 minutes. Similarly we have introduced gradients in the separation buffer. Using this strategy we were able to measure the μFFE separation of a mixture of fluorescently labeled amino acids at 60 different cyclodextrin concentrations, again in 5 minutes.10

Microscale Purification.

Preparative microscale separations present their own unique challenges. Volumes are often too small to recover a reasonable amount of material. Traditional separations require precise timing and valving for effective fraction collection. μFFE presents a unique solution to both of these issues. The continuous flow of analyte into the device introduces enough material to make microscale purifications feasible. Analyte typically flows into the device at ~100 nL/min, allowing a μL of material to be purified every 10 minutes. To recover more material all you need to do is wait longer. Fraction collection is as simple as directing analyte streams to different exit channels. Once it is running no timing or valving is required, greatly simplifying the fabrication and operation of the device. We have recently demonstrated the microscale purification potential of μFFE by using this device to isolate high affinity aptamers for IgE.11 Notably, we were able to double the number of sequences used in the selection while decreasing the concentration of the library 10-fold in comparison to previous CE based selections.

Figure 1. A) Schematic of the μFFE separation mechanism. B) Image of a μFFE device. C) μFFE separation of fluorescein (1), an impurity (2), rhodamine 110 (3), and rhodamine 123 (4). The electric field was 259 V/cm with the cathode at the right.

Literature References

  1. Roman, M. C.; Brown, P. R. Anal. Chem. 1994, 66, 86A-94A.
  2. Raymond, D. E.; Manz, A.; Widmer, H. M. Anal. Chem. 1994, 66, 2858-2865.
  3. Raymond, D. E.; Manz, A.; Widmer, H. M. Anal. Chem. 1996, 68, 2515-2522.
  4. Fonslow, B. R.; Bowser, M. Anal. Chem. 2005, 77, 5706-10.
  5. Turgeon, R. T.; Bowser, M. T. Anal. Bioanal. Chem. 2009, 394, 187-198.
  6. Frost, N. W.; Bowser, M. Lab on a Chip 2010, 10, 1231-36.
  7. Fonslow, B. R.; Bowser, M. T. Anal. Chem. 2006, 78, 8236-8244.
  8. Turgeon, R. T.; Bowser, M. T. Electrophoresis 2009, 30, 1342-1348.
  9. Turgeon, R. T.; Fonslow, B. R.; Bowser, M. T. Anal. Chem. 2010, 82, 3636-3641.
  10. Fonslow, B. R.; Bowser, M. T. Anal. Chem. 2008, 80, 3182 -3189.
  11. Jing, M.; Bowser, M. T. Lab on a Chip, 2011, 11, 3703-3709.

Michael T. Bowser
University of Minnesota
Chemistry Department