1 

 

Optical Biosensors First Semester Report

Fall Semester 2008

by Joel Kindt

Lauren Netherton

Prepared to partially fulfill the requirements for ECE401

Department of Electrical and Computer Engineering Colorado State University

Fort Collins, Colorado 80523

Report Approved: __________________________________ Project Advisor

__________________________________

Senior Design Coordinator

2 

 

ABSTRACT

Currently, cancer detection is a difficult, long and invasive process. Many times the symptoms are unclear or tumors are detected far into the stages of cancer. Clinical diagnostics aim to recognize abnormal characteristics as efficiently and quickly as possible. The optical biosensor is a faster, cheaper alternative for cancer cell detection. With this new machine for cancer screening integrated into the clinic, a more comprehensive healthcare tool would be more widely available to health care professionals. This project addresses optical biosensors. A cell is trapped using dielectrophoresis, and spectral data is used to determine if the cell is cancerous.

In order to make further progress on automating the optical biosensor several specifications of the current set up needed to be quantified. The velocity and voltage requirements for successful trapping were measured and plotted. These obey the concepts behind dielectrophoretic (DEP) force strength and appear reasonable. In order to further understand the DEP force electrostatic modeling was simulated and collected. To better understand the direction of the DEP force, positive DEP (pDEP) and negative DEP (nDEP) dependence was explained based on both the permittivity of the medium and of the cell. The final characteristic examined was the impedance of channel for air, DI (deionized) water, and phosphate buffered saline (PBS).

There were several characteristics of the channel that were examined: trapping requirements, electrostatic modeling, and channel impedance. At higher voltages we can trap cells flowing at a faster rate than when applying a lower voltage. The maximum trapping velocity depends on the voltage applied, not a set velocity of 40 µm/s as decided previously. Maxwell 3D application was used to create the electric field contours for our trap design. We measured the impedance of the channel for air, DI water, and PBS versus frequency. Our main goals are to produce a working simulation of the current microfluidic channel and to compile background information on spectra collection and analysis.

3 

 

TABLE OF CONTENTS

Title………………………………………………………………………………………………. 1

Abstract…………………………………………………………………………………………... 2

Table of Contents………………………………………………………………………………… 3

List of Figures and Tables………………………………………………………………………... 5

I: Introduction……………………………………………………………………………………. 6

II: Prior Research………………………………………………………………………………… 8

A. Fabrication…………………………………………………………………………… 8

B. Flow Control…………………………………………………………………………. 8

C. Optical Detection Circuit…………………………………………………………….. 8

D. Cell Trapping………………………………………………………………………… 8

III: Trapping Requirements……………………………………………………………………… 9

IV: Dielectrophoretic Force…………………………………………………………………….. 14

V: Electrostatic Modeling………………………………………………………………………. 15

A. Electrostatic Field Simulation in Maxwell 3D……………………………………… 15

B. Electric Field and DEP Force using MATLAB…………………………………….. 17

VI: Positive and Negative DEP…………………………………………………………………. 18

VII: Experimental Setup………………………………………………………………………... 19

VIII: Impedance of Channel……………………………………………………………………. 20

IX: Conclusions and Future Work……………………………………………………………… 22

A. Conclusions…………………………………………………………………………. 22

B. Future Work………………………………………………………………………… 22

4 

 

References………………………………………………………………………………………. 24

Appendix A – Abbreviations…………………………………………………………………… 25

Appendix B – Budget…………………………………………………………………………… 25

Appendix C – Semester Goals………………………………………………………………….. 26

Appendix D – MATLAB Code…………………………………………………………………. 28

Acknowledgements……………………………………………………………………………... 28

5 

 

LIST OF FIGURES

Figure 1…….. Diagram of optical biosensor containing microfluidic channel …………………..7

Figure 2…….. Example spectra from healthy cell ………………………………………………. 7

Figure 3…….. Example spectra from cancerous cell …………………………………………… 7

Figure 4…….. Velocity vs. voltage for prostate cells at 52MHz ………………………………. 10

Figure 5…….. Velocity vs. voltage for yeast cells at 30MHz …………………………………. 11

Figure 6…….. Velocity vs. position along channel for yeast cells at 50MHz cells …………… 12

Figure 7…….. Velocity vs. position along channel for yeast cells at 25MHz …………………. 13

Figure 8…….. Normalized velocity vs. position along channel for yeast cells ………………... 13

Figure 9…….. DEP trap………………………………………………………………………… 14

Figure 10…… Actual electrodes on the chip…………………………………………………… 16

Figure 11…… Model of the electrodes in Maxwell 3D ……………………………………….. 16

Figure 12…… Electric field magnitude ………………………………………………………... 17

Figure 13…… DEP force magnitude…………………………………………………………… 17

Figure 14…… Diagram of positive and negative DEP ………………………………………… 18

Figure 15…… K factor for two different conductivities ………………………………………. 18

Figure 16…… Diagram of experimental setup ………………………………………………… 19

Figure 17…… Voltage divider …………………………………………………………………..21

Figure 18…… Voltage divider with channel and known resistor………………………………. 21

Figure 19…… Impedance of the channel for various mediums………………………………... 21

LIST OF TABLES

Table 1………Velocity measurements at 52MHz using prostate cells …………………………..9

Table 2……... Velocity measurements at 30MHz using yeast cells …………………………….11

 

6 

 

Chapter I: Introduction

According to the Morris Animal Foundation, “one in four dogs will die of cancer” and

“60% of golden retrievers die of cancer" [1]. Will your dog die of cancer? What would you think

if you could find all the cancerous cells and remove them? How could you tell which ones are

cancerous? If it works for dogs, can we apply it for other species like humans? The optical

biosensor is based on this idea of cancer cell detection. A quick, inexpensive cancer diagnostic

device would be an efficient and vital addition to any animal clinic or doctor’s office. With this

new machine for cancer screening integrated into the clinic or office visit standard blood test, a

more comprehensive healthcare tool would be more widely available to health care

professionals. How can you tell which cells are cancerous? It is know that a cancer cell has an

enlarged nucleus. This is possibly caused by the uncontrolled synthesis of Deoxyribonucleic acid

(DNA) within the nucleus. The optical biosensor uses a technique to differentiate between

cancerous and non-cancerous cells by looking at the optical spectrum collected when a cell is in

a resonator. Let’s look at an example.

Imagine taking a piece of fruit and placing it in an organ pipe. You have to guess the fruit

from the resulting sound. Intuitively, the change in pitch would vary based on the size and shape

of the fruit. For example, an orange is much smaller, rounder, and softer than a watermelon so

the resulting pitches would be quite different. This is similar to identifying a cancer cell, but in

this case an optical spectrum from a resonator would be used.

Below is a diagram of an optical biosensor containing a microfluidic channel. The

rectangular glass sheet is the bottom of the channel. The yellow shows the gold electrodes on the

glass. The bar electrode connects to ground. The trap is the square part of the electrode on the

left and connects to an AC voltage source. Below the trap is a high powered infrared light

emitting diode (LED) light source. Around the trap area is an arrangement of mirrors to form a

resonator. A spectrometer shown in the upper right corner analyzes the light waves emitted from

the trap. Analysis of the spectra determines the state of the cell in the trap. Example spectra for

an empty trap, healthy cell, and cancerous cell are shown below. Particular wavelengths go

through the resonator when a cell is in the trap just like certain frequencies go through the organ

7 

 

pipe dependent on the fruit inside.

Figure 1. Diagram of optical biosensor containing microfluidic channel

Figure 2. Example spectra from healthy cell Figure 3. Example spectra from cancerous cell

Chapter 2 outlines some prior research on this project. Chapter 3 discusses requirements

for trapping cells and leads into the concept of Dielectrophoretic force in Chapter 4. Chapters 5

and 6 cover computer modeling of this force and the electric field. From this information Chapter

7 discusses the two kinds of dielectrophoresis (DEP): positive and negative. The experimental set

up is explained in Chapter 8 followed by the channel impedance characteristics in Chapter 9.

Chapter 10 summarizes our findings for the semester.

8 

 

Chapter II: Prior Research

Last year, the project focused on a few areas: fabrication, flow control, optical detection,

and automated trapping.

A. Fabrication

The previous team fabricated and tested DEP chips. The chip contains both the

microfluidic channel and the DEP trapping circuit. More specifically, they compared the

fabrication and overall performance between using polydimethylsiloxane (PDMS) versus glass to

create the DEP channels. The use of these chips in experiments also tested the quality of the

fabrication.

B. Flow Control

The team wrote a Labview program to regulate the syringe pump made using a motorized

micrometer. This computer interface is easily customizable to the experiment being conducted.

From the control, the pump duty cycle can be entered manually. This is extremely useful for

slowing the fluid flow and for consistently recreating experimental procedures.

C. Optical Detection Circuit

The optical trigger circuit was created last year. The goal was to identify the presence of

a cell in the trap. The team was able to design and test two different circuit designs. The team

determined the main issues to be time response, noise, and defined ideas for future circuit

designs. The detection circuit is the first part of automating the trapping procedure.

D. Cell Trapping

An RF switch was investigated to energize the DEP trap when a cell was detected by the

optical detector. This utilized a Data Acquisition Unit (DAQ) to read the output voltages from

the detection circuit to determine when the switch should turn on. Last year’s team eliminated a

few approaches that did not work due to high frequency interference.

9 

 

Chapter III: Trapping Requirements

Previous research documented acceptable standards of fluid flow to be successful at

trapping cells. Trapping cells allows the spectrometer to collect spectra from the trap. The

spectra can be interpreted and analyzed to determine if the cell is cancerous. In order to ensure

trapping, the flow rate must be slow enough. Last year’s documentation states that flow rate must

be less than or equal to 40 µm/s based on the strength of DEP force from the trap. This velocity

trapping requirement was further investigated. The velocity and voltage relationship was

investigated because the DEP force to oppose fluid flow depends on the electric field created by

the applied voltage. The concept behind dielectrophoretic (DEP) force is described in a later

section. DEP force is also dependent on the frequency of the applied voltage so two different

frequencies were examined.

The first experiment used prostate cancer cells in a medium of phosphate buffered saline

(PBS). The voltage was an AC signal with magnitude 16V at the frequency 52 Mhz. Only three

cells flowed through the trap in the duration of the experiment. None of the cells were trapped.

The velocities were between 54 and 135 µm/s. The data is recorded in the table and graph below.

This supports the theory that fluid flow must be approximately 40 µm/s or less in order to trap.

The next experiment was conducted with yeast cells.

52MHz voltage (V)  trapping velocity (um/s) 

 not trapped velocity  (um/s) 

16     54 

16     96.71 

16     135.87  Table 1. Velocity measurements at 52MHz using prostate cells

10 

 

Figure 4. Velocity vs. voltage for prostate cells at 52MHz

The second experiment used yeast cells in a medium of phosphate buffered saline (PBS).

The voltage was an AC signal with varying magnitudes at 30 Mhz. Fourteen cells flowed

through the trap in the duration of this experiment. Nine cells were successfully trapped and four

were not. The cells appeared to be quite large, suggesting that they were possibly a clump of

cells. The velocity range was between 15 and 143 µm/s. The data is recorded in the table and the

graph below. This flow was controlled using the Labview fluid pump duty cycle program.

Higher velocities occur right after the pump has pulsed during the duty cycle to accelerate flow

through the channel. We can see that at low voltages we need a slower fluid flow in order to trap.

If we use higher voltages, we can trap faster flowing cells. Four of the nine trapped cells were

flowing faster than the standard 40 µm/s trapping velocity set previously. This disproves the

theory that cells must flow at or slower than 40 µm/s in order to be successfully trapped. The

data is shown in the table and graph below. In conclusion, successful trapping velocities are

dependent on the applied voltage and frequency. This is reasonable because the DEP force is

stronger with larger electric fields, created by higher voltages, and varies with frequency.

11 

 

30MHz voltage (V)  trapping velocity (um/s)   not trapped velocity(um/s) 

10  14.93    

9  41.7    

9  31.25    

8  100    

7  55.56    

6  33.33    

5  33.33    

4  16.67  142.86 

4     100 

3  14.93  25 

3     25 

2     16.67  Table 2. Velocity measurements at 30MHz using yeast cells

Figure 5. Velocity vs. voltage for yeast cells at 30MHz

There are a few factors that could have affected the results previously shown. Following

experiments should be conducted to eliminate multivariable issues. Possible result differences

can be due to the following: cell type, frequency, viability of the cells, and medium differences.

These large variations in velocities were the basis for another experiment. We looked at

the effect of trap location on the fluid velocity of yeast cells. The plots show data at three trap

12 

 

positions along the channel. Position one is closest to the beginning of the channel where the

fluid is connected and position three is towards the outlet end of the channel. At 50MHz, we

observed fluid velocity at 16V, shown in red, and at 0V, shown in blue. The fluid velocity tended

to increase in the center position in relation to the other positions.

Figure 6. Velocity vs. position along channel for yeast cells at 50MHz

Next the velocity profile was examined at 25MHz at 16V and 0V. The resulting data at

both voltages suggests that fluid velocity decreases in the center position then increases a little

toward the end of the channel. This appears to be the opposite effect at 50MHz. The last plot in

this section is the normalized velocity versus position. This was created for each frequency. The

velocities at 16V were divided by the corresponding velocities at the 0V control case. This

confirmed the previously mentioned observations of the fluid velocity changes along the trap

positions of the channel.

13 

 

 

Figure 7. Velocity vs. position along channel for yeast cells at 25MHz

Figure 8. Normalized velocity vs. position along channel for yeast cells

14 

 

Chapter IV: Dielectrophoretic Force

The ability to trap cells is based on a concept from electromagnetics called

dielectrophoresis. A non-uniform electric field will polarize the molecules within a dielectric, or

insulator, inducing a dieletrophoretic (DEP) force on the dielectric. This force is used to hold the

cell in the trap on the chip. We create a non-uniform electric field by applying an AC voltage to

the electrode with the square trap and ground to the bar electrode. The figure below would

demonstrate flow from right to left with the trap to the right of the grounded electrode bar. The

DEP force will oppose the fluid flow to hold the cell in the trap. Now we need to research what

determines the DEP force.

The DEP force equation is 2 ε Re K E. We can see the

force depends on several variables within the experiment. The r represents

the radius of the cell. εm is the permittivity of the medium or the ability of a

material to polarize due to application of an electric field. E is the gradient

of the square of the electric field. Re[k] is the real part of the polarization

factor. The value of the K factor is determined by the following relationship

between the complex permittivity of both the cell and the medium:

2

The complex permittivity is also dependent on conductivity and angular frequency .

The conductivity, , is a measure of the ability to conduct an electric current. The angular

frequency is in radians and calculated by 2 . parameters: cell radius, cell and medium

permittivity, cell and medium conductivity, frequency, and the electric field. This is shown

below in the DEP force equations.

2 ε Re2

E 2 ε Re2 2 E

The chapters on electrostatic modeling and positive and negative DEP will expand on the

detailed effects of the electric field and K factor on the DEP force.

Figure 9. DEP trap. Reproduced from [4]. 

15 

 

Chapter V: Electrostatic Modeling

A. Electrostatic Simulation in Maxwell 3D A picture of the actual electrodes on the chip is seen in Figure 10. In this figure, one can

also see the channel in which the fluid is pumped into. By knowing the dimensions of the

electrodes and the channel, a simulation of the electric field can be produced using Maxwell 3D,

illustrated in Figure 11. This software is convenient since one can select the materials to be

used. For example, the electrodes are chosen to be gold, while the substrate is selected as glass.

To simulate the electric field, an excitation force needs to be assigned to the electrodes.

A voltage excitation of 10V is chosen for the electrode with the square trap, whereas an

excitation of 0V is assigned to the bar electrode to represent ground. One should note this is a

DC voltage, whereas an AC voltage is actually applied to the electrodes. Using a DC voltage for

electrostatic modeling is reasonable since it shows the magnitude of the electric field.

Furthermore, a DC voltage is practical because the AC frequency that is used corresponds

to a wavelength that is longer than the distance between the trap electrode and the ground

electrode. The wavelength is found with the equation:

where λ is the wavelength, v is the phase velocity, and f is the frequency. Assuming the medium

is water, the phase velocity is:

3 10  

1.33334

3 10  

where c is the speed of light in a vacuum, and n is the index of refraction for water. Therefore,

the wavelength can now be found assuming the AC frequency is 50 MHz:

34 3 10

50 

34 3 10

50 10 1 4.5 4.5

Consequently, the wavelength is much greater than the distance between the trap electrode and

ground electrode, so using a DC voltage is reasonable. The AC frequency should still be

considered with its effects on the medium, particularly with electrolysis.

 

16 

 

 

In Maxwell 3D, the electric field is solved for the entire 3D area of the channel.

However, the main interest of the electric field is a 2D plane in the center of the trap. This plane

can be found by exporting the electric field data in the center of the trap to MATLAB. Next, a

program called plotc.m is used to plot the electric field contours. The contour plot indicates a

higher electric field magnitude with red, and a lower magnitude with blue.

Figure 10. Actual electrodes on the chip.

 

 

Figure 11. Model of the electrodes in Maxwell 3D.

17 

 

B. Electric Field and DEP Force using MATLAB

The DEP force is of interest since it is the force that traps the cell. From the electric field

data, the DEP force can be found with the following equation:

2 ε Re K E

To quickly model the magnitude of the DEP force as a function of the electric field, the equation

can be normalized to be:

| E |

From this approximation, a program in MATLAB was written that takes the electric field as an

input and solves for the normalized DEP force. The program plotc.m is then used to plot the

contours of the normalized DEP force. The contour plots of the electric field and the DEP force

are illustrated in Figures 12 and 13, respectively.

One should expect the magnitude of the DEP force is greatest where the electric field

contours are closest together. This is because the del operator, which takes the derivative in the

Y and Z directions, is greatest for the largest change in the electric field. Even though the

magnitude of the DEP force is shown, the direction of the DEP force is of interest as well. The

direction of the DEP force will be discussed in the next chapter.

Figure 12. Electric field magnitude. Figure 13. DEP force magnitude.

18 

 

Chapter VI: Positive and Negative DEP

The direction of the DEP force will determine if the force is positive or negative DEP.

As seen in Figure 14, positive DEP is indicated if the arrows point towards each other, while

negative DEP is indicated if the arrows point away from each other. The arrows are drawn on

the magnitude of electric field plot, so that one can see both the electric field and the direction of

the DEP force.

The direction of the DEP force is found by looking at the K factor of the DEP force

equation:

2

where the K factor is a function of the complex permittivity between the medium and the cell,

and the complex permittivity is a function of frequency as seen:

Thus, the K factor is a function of frequency. The real part of the K factor versus frequency is

illustrated in Figure 15 for a typical mammalian white blood cell [a]. In the figure, both low (σ =

0.1 S/m) and high (σ = 1.5 S/m) conductivities are shown. The DEP force is positive when the K

factor is greater than zero, and the DEP force is negative when the K factor is less than zero.

Figure 14. Diagram of positive and negative DEP.

Figure 15. K factor for two different conductivities. Reproduced from [5].

19 

 

Chapter VII: Experimental Setup

A majority of the experimental setup was completed by the senior design team last year,

as described in Chapter II. Figure 16 illustrates some of the prior work, including the amplifier

and control electronics, and the OFIS sensor chip. The experiment works by the syringe

pumping cells into the channel. To achieve a velocity that is slow enough to trap the cells, a

LabVIEW program regulates a duty cycle for the custom syringe pump. The duty cycle is

usually set to have a long “off” time and a short “on” time. Even though the duty cycle creates

pulses, the flow is relatively steady because magnitude and time of the pulse are very small.

When the cell approaches the DEP trap, the person observing the cell turns on the

function generator to apply the AC voltage. If the velocity of the cell is slow enough and the

applied AC voltage has the required magnitude and frequency, then the cell is likely to be

trapped. A previous discussion on trapping requirements is found in Chapter III. The high

power infrared LED then reflects off the cell and is measured by the spectrometer. If the spectral

data matches that of a cancerous cell, then the cell is cancerous. After the spectral data is taken,

the cell is released by turning off the function generator.

Figure 16. Diagram of experimental setup.

20 

 

Chapter VIII: Impedance of Channel

One of the experiments of interest was to measure the impedance of the channel. The

impedance of the channel was determined by using a voltage divider, as seen in Figure 17. The

voltage divider was implemented such that a known resistor was used for impedance Z2, and the

channel impedance was Z1. Figure 18 illustrates the connection between the channel and the

known resistor. From the voltage divider equation, the channel impedance Z1 can be found.

·

1

The AC voltages Vin and Vout were then measured with an oscilloscope. However, only the

magnitudes of the AC voltages were measured, such that only the magnitude of Z1 could be

found. Further experiments should be done to measure the phase of Vin and Vout so that the

phase of Z1 is known.

The channel impedance Z1 was then measured for various mediums and frequencies.

Figure 19 includes the impedance of the channel for air, deionized (DI) water, and phosphate

buffered saline (PBS). As seen in the figure, the mediums of air and DI water exhibit capacitive

impedance. This is because a capacitor is defined with the impedance Z = 1/(jωC), so increasing

the frequency decreases the impedance.

In addition, the real part of the impedance is important to observe because it dissipates

power as given by:

where V is the constant voltage source and R1 is the resistance. The equation shows that the

power dissipation will increase as the resistance decreases. Consequently, PBS will dissipate the

most power at lower frequencies, because it has lower impedance. The power dissipation will

warm up the water surrounding the electrodes, and will affect the spectra.

21 

 

Figure 17. Voltage divider. Figure 18. Voltage divider with channel and known resistor.

Figure 19. Impedance of the channel for various mediums.

22 

 

Chapter IX: Conclusions and Future Work

A. Conclusions

The accomplishments from this semester were defining velocity and voltage trapping

requirements, verifying trapping requirements against the theory behind DEP trapping,

simulating the electric field for our current trap, and simulating an example of the Voldman trap.

Through experiments varying voltage and fluid velocities we determined some basic

trends in the ability to trap cells. We tested the theory from last year that cells must flow slower

than 40 µm/s in order to be trapped. At higher voltages we can trap cells flowing at a faster rate

than when applying a lower voltage. The maximum trapping velocity depends on the voltage

applied, not a set velocity of 40 µm/s as decided previously.

Trapping success is based on the combination of velocity and voltage. Trapping also

changes based on the AC voltage frequency in the experiment. The trend that as voltage

increases the velocity can be faster to trap is the same, but the exact combination of velocity and

voltage will shift based on the frequency value. These theories make sense because of the DEP

force proportionality to the gradient of the electric field.

Maxwell 3D application was used to create the electric field contours for our trap design.

These were explained in the Electrostatic Modeling section above. Another program is being

investigated and tested to be used in coordination with the DEP force simulation.

DEP force simulations were conducted using a MATLAB code published by Dr. Joel

Voldman. We found that the public program was not fully functional. There is a test to confirm

accurate functionality. We were able to trace the outputs and correct the input parameters to get a

closer result to the expected test output. Once we have the correct format of the electric field, the

test output should match the expected one for the program.

B. Future Work

The main areas of concern for next semester are computer simulation and spectral data.

Our main goals are to produce a working simulation of the current microfluidic channel and to

compile background information on spectra collection and analysis. Some accomplishments

23 

 

from this semester focused on computer modeling of the fluid flow, electric field and DEP force.

Future work needs to finalize and verify the progress completed thus far on the models. Since we

can now record spectra, we also need to better understand the spectral data.

The following outlines the process plan for computer modeling of the channel. If we are

able to access a specific electrostatic modeling program, we plan to simulate a working DEP

force model of our system. We plan to model this and the fluid flow of the channel by the

beginning of March. Next, we will verify this using experimental data to compare holding

voltage, and flow rates at specific conditions. Once successfully completed, we can design and

test another trap until the beginning of April. Otherwise, the focus will shift to correcting the

model. From there we will have to decide next steps toward fabrication and testing of the new

design.

In addition to modeling, we plan to collect spectra, research spectral shifts, and define

interpreting spectral shifts. We will conduct weekly experiments collecting spectra. The

frequency of our experiments depends on our access to cells. Using prior research and

publications, we will create a database on interpreting spectral data as cancerous versus non-

cancerous cells. We will include our corresponding experimental data, spectra, and analysis.

Hopefully, we will attain different types of cancer cells to test. We can include a section

comparing spectra shifts from different types of cancer cells.

The deliverables will be documentation on specific accomplishments. By the end of the

semester we should have instructional documentation on: electrostatic simulation, DEP force

simulation and comparison to actual data, experimental setup and protocol for running

experiments with spectra collection, spectral shift expectations and how to analyze spectral

shifts.

The goals for next semester aim to improve the accuracy of the Optical Biosensor project

through computer modeling, spectral theory and experimental proof. A huge accomplishment

would be to have accurate computer modeling of the DEP force and verification of the

theoretical calculations through experimental data. We also should have instructions on

interpreting spectra and some data to confirm our conclusions.

24 

 

REFERENCES

[1] J. Voldman, R. Braff, M. Toner, M. Gray, and M. Schmidt, “Holding Forces of Single-Particle Dielectrophoretic Traps,” Biophysical Journal, Vol. 80, pp 531-541, January 2001. [2] W. Wang,H. Shao, K. Lear, “Lab-on-a-Chip Single Particle Dielectrophoretic (DEP) Traps” PowerPoint presentation given March 2006. [3] Optofluidic Intracavity Spectroscopy of Canine Lymphoma and Lymphocytes Lear, Kevin L.; Shao, Hua; Wang, Weina; Lana, Susan E.; LEOS Summer Topical Meetings, 2007 Digest of the IEEE, 23-25 July 2007 Page(s):121 – 122 [4] http://www.engr.colostate.edu/ece-sr-design/AY07/biosensors/ECE402_Report_1.09.pdf

[5] J. Voldman, “Electrical Forces for Microscale Cell Manipulation” Annu. Rev. Biomed. Engr. 2006. Pages 425-454.

25 

 

APPENDICES

Appendix A: Abbreviations DEP-Dieletrophoresis/dielectrophoretic OFIS-Optofluidic Intracavity Spectroscopy PBS-Phosphate Buffered Saline

Appendix B: Budget

We were able to save our money for this semester because we used materials purchased by the previous senior design team. We have a budget plan for next semester shown above.

26 

 

Appendix C: Semester Goals

Overall Goals for Senior Design Team:

1. Quantitative modeling of micro fluidic channel flow a. Fluent/Gambit

i. Time: 40 hours ii. Resources: Professors, graduate students that have used the software, user

groups iii. Risk:

1. Time: This could take a lot of time so need to time manage the different duties

2. Resources: Not able to find the right people to help iv. Milestones:

1. Design basic channel with reservoirs 2. Include nanotube inlet 3. Include trap dimensions in channel 4. Design multiple chips 5. Collect data and compare data for different chips

v. Definition of Success: Success gauged by completion of milestones b. Electrostatic modeling with Maxwell

i. Time: 10 hrs ii. Resources: Voldman, online tutorials and user groups

iii. Risk: 1. May not be able to export data to model DEP forces with Matlab 2. Time: Could take away from other duties if time is not managed

carefully iv. Milestones:

1. Display cross-section of electric field data at middle of DEP trap 2. Explore similarities between Maxwell and FEMLAB, which is the

actual program used to export the electric field data 3. Export electric field data from Maxwell to model DEP forces with

Matlab 4. If it is deemed not possible to export data from Maxwell, then look

into purchasing license for FEMLAB (now COMOSOL Multiphysics)

v. Definition of Success: Success gauged by completion of milestones c. Modeling DEP forces with MatLab

i. Time: 30hrs if no issues arise (excludes experiments time to calculate flow rate)

ii. Resources: Voldman and graduate students, proficient MatLab users iii. Risk:

1. Time: could take longer to precisely model our setup than expected and reduce the focus from other duties

2. Resources: MatLab groups not focused on streamforce modeling, Voldman and students no longer supporting the software

27 

 

iv. Milestones: 1. Working test example from Rosenthal et al 2005 publication 2. Simulate DEP model with our chip parameters 3. Simulate our chip with electric field data from Maxwell 4. Flow rate analysis between experiments and modeling

v. Definition of Success: Success gauged by completion of milestones

2. Collection and analysis of cell data i. Time: Unlimited

ii. Resources: Need viable cells, information about viability tests, what is solution made of and how to make diluted solution, money

iii. Risk: 1. Time- Devoting too much time and neglecting other duties 2. Money- Lack of data may lose grant opportunities 3. Inconclusive data 4. Availability of cells and running experiments while cells are viable

iv. Milestones: 1. Trapping cells and data collection (velocity, frequencies, voltages,

bubble formation, time between flowing cells, position of cells, etc)

2. Spectra 3. Theoretical analysis 4. Experimental analysis

v. Definition of Success: Success will be gauged by completion of milestones

3. Design (fabrication-Weina) new DEP mask i. Time: Coordination with design and fabrication will be time dependent

(don’t have a gauge of how long this could take) ii. Resources: Simulations in Maxwell for electrostatic design specifications,

MatLab DEP modeling, Fluent simulations to help design for fluid flow iii. Risk:

1. Time- Takes a lot of time from other duties 2. Money- Need money for materials 3. Lack of materials- Not able to get materials or too much money

iv. Milestones: 1. Come up with functional designs 2. Actually fabricating chips from the design

v. Definition of Success: Success will be gauged by completion of milestones

28 

 

4. Analyze how bubble size affects channel flow i. Time: 10 hours

ii. Resources: Taking all the data collected and breaking it into comparable data, program to graph data

iii. Risk: Taking too much time and not focusing on other duties iv. Milestones:

1. Collecting relevant data 2. Organizing data 3. Plotting data

v. Definition of Success: Success gauged by completion of milestones

Appendix D: MATLAB Code

DEP force modeling: Rosenthal, A., Taff, B. M., & Voldman, J. Quantitative modeling of dielectrophoretic traps. Lab on a Chip 6 (2006), 508-515.

Link: http://www.rle.mit.edu/biomicro/publications.htm

Acknowledgements

A heartfelt thank you to Dr. Lear, Dr. Kisker, Weina, and Hailey for guidance, advice, time, and brainpower working on this project!