A Fast Dielectrophoretic Cell Sorter

1
A Fast Dielectrophoretic Cell Sorter
Shih-Chi Chen,
Tom Hunt,
Ankur Mehta,
Ashish Shah,
Kate Thompson,
Zhiyu Zhang
2
Flow Cytometry and Cell Sorters

• Flow Cytometry characterization of cells as they
  flow past sensors measuring optical, electrical,
  or other properties
• Cell Sorters partition a mixture of cells based
  on characteristics detected through flow cytometry

Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
3
Applications of Cell Sorters

• Isolation of human chromosomes for construction
  of genetic libraries
• Diagnose leukemia and other cancers
• Potential use in treatment of chronic diseases
  like HIV and diabetes

Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
4
Current Technology

• Commercial sorters
• Process 10,000 100,000 cells / sec
• Not portable
• Prohibitively expensive
• MEMS sorters
• Research devices sort 100 cells / sec
• Potential for greater portability
• Potential for lower cost
• Potential for disposable chips

Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
5
Project Goal

• Design a micro-scale cell sorter
• Sort 1,000 cells/second
• Use dielectrophoresis for sorting
• Cause no permanent damage to the cells
• lt 1 Pa Shear Stress on cells
• lt 70 mV across cell membrane
• Maximum Temperature lt 39 C

Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
6
Design Overview

• Three Main Design Components
• Flow Focuser
• Optical Detection System
• Cell Sorter

Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
7
Active Region
100 um
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
8
Focuser
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
9
Analytical Model
Device Goal 1,000 cells / sec 1 Error
100 ?m spacing Cell Velocity 0.1 m/s

• Fully Developed Flow
• Parabolic Velocity Profile
• u umax (1 r2/R2)

• Shear Stress
• ? du/dy

? umax - umax (1 rcell2/Rchannel2) rcell
Minimum Channel Diameter 32 ?m
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
10
Liquid / Liquid Sheath Flow

• Reynolds Number
• Re ? D u / ?

• Conservation of Mass
• ? dmin / dt ? dmout / dt

sheath flow
sample flow
sheath flow

• 2D1v1 D2v2 Davcore/2
• D2 v2 d vcore

Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
11
Horizontal Focusing
2D ANSYS Finite Element Model Predicts 35 ?m
sample diameter focused to 10 ?m maximum
velocity of 0.06 m/s
Sheath
Velocity Vector Plot
Cells
Sheath
Particle Trace
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
12
Horizontal Focusing
Cell Solution
Sheath Flow
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
13
Vertical Focusing
2D ANSYS Finite Element Model Predicts 35 ?m
sample diameter focused to 14 ?m maximum
velocity of 0.096 m/s
Velocity Vector Plot
Particle Trace
Sheath
Cells
Sheath
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
14
Vertical Focusing
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
15
3D Fluid Modeling
35 um
35 um
35 um
70 um
35 um
35 um
35 um
200 um
250 um
150 um
35 um
100 um
100 um
100 um
100 um
100 um
35 um
100 um
Inlet Horizontal Sheath Velocity 0.0083
m/s Inlet Horizontal Core Velocity 0.03
m/s Inlet Vertical Sheath Velocity 0.015
m/s Goal Maximum Velocity 0.1 m/s
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
16
3D Fluid Modeling
Pressure
3D Simulation Confirms Max Velocity 0.100459
m/s Pressure Drop 7.6 kPa Focusing Successful
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
17
3D Fluid Modeling
Isometric Vector Plot Around Electrodes
Vector Plot
Particles closest to electrodes move 2.5 to 3.5
?m Particles furthest from electrodes move up to
2 ?m
Top View Particle Trace Around Electrodes
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
18
DEP Sorting
Design challenge at least an order of magnitude
more DEP force than typical experiments, with a
high conductivity medium.
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
19
DEP Cell Spacer

• Random cell spacing in the flow direction from
  focuser introduces possibility of sorting error
• Uniform spacing allows up to fivefold increase in
  throughput
• Idea Impose spacing with DEP quadrupole traps

DEP forces are too weak! We must endure random
spacing
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
20
DEP Sorting paddle electrodes
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
21
DEP Sorting paddle electrodes
How far can the paddle electrodes push the cells
in 0.15msec?
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
22
DEP Sorting splitter electrodes
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
23
Heating issues

• Joule heating (sE2) causes 106 times more heat
  generation than viscous dissipation
• The temperature rise is excessive, even with
  refrigerated input fluid, so extra cooling
  channels are necessary.

Temperature profile after all of the electrodes
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
24
Cooling
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
25
DEP Sorting drive electronics
Configurable delay, pulse width
Electronic switches
Thresholding
Optical Signal
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
26
Device Exploded View
Gold Electrodes
Top Cover Slip
PDMS Flow Channel
Bottom Pyrex Wafer
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
27
Fabrication Steps
Glass and Electrodes
Deposit Au in SU-8 Wells
Drill, Laser Cutting
Flow Channel
SU-8 Photoresist Mold
Injection Molded PDMS
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
28
Assembly
PDMS on carrier wafer
Fully processed bottom glass
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
29
Assembly
Fully processed top glass
PDMS and bottom glass with carrier wafer removed
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
30
Packaging

• Need to access device from outside world
• Syringe pumps for fluid supply
• O-rings for fluid connections
• Electrical connections with
• Wire bonds, zebra stripe

www.syringepump.com/
1 cm
1.5 cm
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
31
Packaging
Ceramic Carrier
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
32
Packaging
Device Unit
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
33
Packaging
ZIF Socket
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
34
Final Assembly
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
35
Close Up
Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
36
System View
Microscope/ Optics
Circuitry
Device Sandwich
Top Cover Slip
PDMS Layer
Bottom Pyrex
Ceramic Carrier
Ceramic Carrier
ZIF Socket

ZIF Socket

Syringe Pump
ZIF Socket

Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
37
Second Order Effects

• First iteration modeled only main forces
• Other effects to be added in the future include
• Fluid flow around paddles
• Pushes cell away from electrodes, decreasing DEP
  force
• Electrohydrodynamic flow
• Within an order of magnitude of DEP force.
  Fortunately it acts in the same direction
• More accurate E-field modeling
• Including higher multipole moments, non-uniform
  forces across the cell volume, and 3D fringing
  fields

Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
38
Device Limitations

• Up to 2 orders of magnitude slower than existing
  macroscale flow cytometers
• Not quite portable requires cooling unit to
  chill cells and solution to 5o C
• Cheaper than current systems, but many process
  steps mean it might not be disposable
• PDMS not commercially viable - swells with
  prolonged exposure to liquids
• But its getting there!

Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
39
Future Analysis

• Spacer
• Ensuring regular cell spacing allows up to
  fivefold increase in throughput with the same
  flow parameters
• Alternative actuation
• DEP is very weak, and it generates a lot of heat
• Optical tweezers or magnetic sorting may be more
  effective

Background ? Goal ? Design/Modeling ?
Fabrication ? Packaging ? Discussion
40

Questions?