Microchip Flow Cytometry

   We have developed microfluidic devices that incorporate electrokinetic focusing [1,2] to spatially confine particles and monitor their spectral characteristics. Microchip operating conditions, flow rates, counting efficiencies, and particle throughputs have been investigated using mixtures of fluorescently labeled and unlabeled particles. Figure 1 shows a 5 second integrated CCD exposure of particle transport through the microfabricated focusing chamber. The streaks observed in the microchip channels are due to the fluorescence from individual particles as they travel from the sample channel, through the focusing chamber, and down the analysis channel. The fluorescently labeled particles were 1.88 µm in diameter and at a concentration of ~ 1.5 x 107 particles/mL. The length of the arrows correspond to the observed fluid velocities and are proportional to the field strengths in each of the channels. The stream width, measured at a point 50 µm beyond the focusing chamber, was 8.0 µm at maximum extent. This is ~ 15% of the channel width and corresponds to the sample to analysis field strength ratio which was 0.15 (100 V/cm in the sample channel and 700 V/cm in the analysis channel).

   Figure 2 shows typical scatter/fluorescence coincidence data collected from a 50/50 mixture of 0.97 µm fluorescently labeled and 1.94 µm unlabeled latex particles. Data from such runs were generally acquired for 60 seconds to gather enough information for statistical treatment, however, only a 10 second segment of the run is shown so that the individual peaks can be seen. The inset in Figure 2 is an expanded view of a 0.4 second section of the run. Each peak in the scatter channel (A) represents a single particle passing through the interrogation region, with the labeled particles also appearing in the fluorescence channel (B). The sample to focusing field strength ratio (Esample/Efocus) was 0.18 and yielded a sample stream width of ~ 7-8 µm. The average particle throughput was 13.5 particles/s which is within experimental error of the expected 15 + 1 particles/s. In addition, the corresponding peak intensity in the scatter channel for the 0.97 µm diameter labeled particles is ~ 0.25 of the signal from the 1.94 µm unlabeled particles. These observations are consistent with that expected for forward angle Mie scattering where the scattering intensity should be proportional to the particle surface area and, therefore, to the square of the particle diameter.

   Figure 3A shows a scatter plot of the same set of data displayed in Figure 2. The boxed regions in Figure 3A contain 95% of the total number of particles that were detected in this experiment. Figure 3B shows a histogram of the peak intensities from the scatter data in Figure 3A. The histogram has a bimodal distribution, which is consistent with the two sizes of particles analyzed. The location of the arrow drawn in the plot was determined using the fluorescence channel data where 368 fluorescently labeled 0.972 µm particles were observed. This number was subtracted from the total number of peaks counted in the scatter channel (808) to obtain the number of peaks due to the unlabeled 1.94 µm particles (440). The two types of particles were also run individually to confirm that the two peaks seen in the signal intensity distribution in Figure 3B were actually due to the two different sized particles.

   In this work, microchip cytometry with electrokinetic focusing was performed to spatially confine samples of particles varying in type, size, and ratio, and to count these particles using coincident light scattering and fluorescence detection. Particles of varying sizes can be distinguished by their scattering intensities. In addition, sample throughput and detection efficiency were shown to be affected by the sample to focusing field strength ratio applied to the chip. Microchip cytometry has the potential to be less expensive, consume less solvent, and require less bulky instrumentation than conventional flow cytometry. Rare event cell sorting can also be carried out on a microchip, without the dilution which occurs with conventional flow cytometers. Methods of increasing sample throughput, such as applying higher operating field strengths and using microchips with different and/or multiple channel geometries, are being explored to achieve 0.1 to 1 kHz counting rates.

(1) Jacobson, S. C., Ramsey, J. M. "Electrokinetic Focusing in Microfabricated Channel Structures," Analytical Chemistry, 1997, 69, 3212-3217.
(2) Schrum, D. P., Culbertson, C. T., Jacobson, S. C., Ramsey, J. M. "Microchip Flow Cytometry Using Electrokinetic Focusing," Analytical Chemistry, 1999, (submitted).

back to the Homepage

back to Laboratory on a chip subdirectory

back to Projects subdirectory