close

Вход

Забыли?

вход по аккаунту

?

Non-thermal effects of pulsed microwave fields on catecholamine release from chromaffin cells: Exposure system design and characterization, and experimental data

код для вставкиСкачать
University of Nevada, Reno
Non-Thermal Effects of Pulsed Microwave Fields on
Catecholamine Release from Chromaffin Cells: Exposure System
Design and Characterization, and Experimental Data
A dissertation submitted in partial fulfillment of the requirements for the degree of
Doctor of Philosophy, in Electrical Engineering
by
Jihwan Yoon
Dr. Indira Chatterjee/Dissertation Advisor
Dr. Gale Craviso/Dissertation Co-Advisor
December, 2008
3339154
Copyright 2009 by
Yoon, Jihwan
All rights reserved
3339154
2009
Co-Advisor
i
Dissertation Abstract
Over the last several decades, innumerable investigations have been conducted to
define the bioeffects of microwave (MW) fields on cells, tissue and whole animals, not
only stemming from concerns of public safety from exposure to wireless communication
systems, but also for the purpose of developing beneficial applications in therapeutic and
diagnostic procedures. Studies with cells have indicated that certain exposure parameters
related to MW fields may influence particular functions of cells by the interactions
between specific cellular constituents in cells and the applied MW fields.
The goal of this research is to identify specific MW parameters in the frequency
range 1 to 6 GHz that can induce possible nonthermal effects on catecholamine release
from chromaffin cells. For the study, a free-space broadband in vitro MW exposure
system consisting of a cell perfusion apparatus (CPA), which allows for on-line
monitoring of catecholamine release, placed in the far field of a horn antenna, has been
designed, constructed and tested. The system was characterized by measurements
performed in an anechoic chamber as well as numerically using the Finite-Difference
Time-Domain (FDTD) method, which computes detailed distributions of the electric field
and specific absorption rate (SAR). All aspects of the experiments were computer
ii
controlled hence minimizing the possibility of human error during experiments.
Computations and experiments performed using an early design of the free-space
exposure system led us to conclude that the system had certain limitations, which were
(1) unsatisfactory homogeneity of the electric field, SAR and temperature distribution
and (2) insufficiently high magnitude of the electric field at the location of the cells. In
order to overcome these limitations, modifications were made to the free-space exposure
system including placing the CPA in the near field, rather than in the far field of the horn
antenna and replacing the original CPA with one having smaller dimensions. These
modifications significantly improved the homogeneity of the electric field, SAR and
temperature distribution in the region containing the cells, and also increased the
magnitude of the electric field to the maximum possible level achievable with the
existing MW equipment. A series of carefully controlled experiments were carried out
with this modified free space exposure system to quantify changes in catecholamine
release under various parameters of MW field exposure. These experiments indicated that
MW induced effects on catecholamine release occurs most often when a pulsed frequency
sweep (PFS) with a 100 ms pulse width in the frequency range 5-6 GHz is used. However,
not all experiments carried out using these MW exposure parameters showed an effect,
iii
indicating a lack of reproductivity. This could be because the maximum electric field
magnitude capable of being produced by our exposure system is close to a threshold
value for causing the effect on catecholamine release.
In an effect to increase the probability of inducing reproducible robust changes in
catecholamine release, a second exposure system was designed that can produce much
larger electric fields. The new system is based on a planar exposure chamber embedded
in a low loss, high dielectric constant substrate material. Initial computations using the
FDTD method have shown that the newly designed exposure system will be capable of
producing an electric field that is 20 - 40 fold greater in magnitude than our original
exposure system, while at the same time maintaining acceptable homogeneity of the
electric field over the region containing the cells for the entire frequency range of interest.
iv
Acknowledgements
First of all, I would like to express my appreciation to my advisor, Dr. Indira
Chatterjee, for her support, patience, and encouragement throughout my graduate studies.
Her technical and editorial advice was essential to the completion of this dissertation and
she has taught me numerous lessons and insights on the workings of academic research.
Another person who deserves credit is my co-advisor Dr. Gale Craviso in the
Pharmacology Department who has provided tremendous support in the biological
aspects of the research, and the editorial help.
My thanks also go to the members of my committee: Drs. Bruce Johnson, James
Henson, and Suk-Wah Tam-Chang, who have been extremely helpful in their academic
advisement. I would also like to take this opportunity to thank Mr. Dana McPherson who
was always there for me to help in finding instruments, to provide technical support and
simply to listen to my crazy ideas.
I would like to appreciate lab members, Horace Goff, Todd Hagan, and Michael
Lambrecht for their occasional help and their friendship toward me.
I like to thank my friends in Korea and the U.S. for their continuous support.
Last, but not least, I would like to thank my family for supporting everything during the
v
past several years and my wife, Minkyung, who has love, kindness, patience, and
goodness, receive my deepest love for her dedication, support and encouragement during
my graduate studies.
This dissertation is lovingly dedicated to my wife and my son Joseph, and most
of all, my Lord, Jesus who always makes the best way for me.
I would also like to acknowledge support of this work by the Air Force Office of
Scientific Research under Grants F49620-03-1-0267, F49620-03-1-0262, FA9550-04-10194, FA9550-05-1-0308 and FA9550-06-1-0377.
vi
Table of Contents
Dissertation Abstract
i
Acknowledgements
iv
Table of Contents
vi
List of Figures
xiv
List of Tables
xxiv
Chapter 1: Introduction: General Introduction
1
1.1 In Vivo Studies Addressing MW-Induced Effects on the Nervous System.
3
1.2 In Vivo Studies Addressing MW-Induced Effects on Neurotransmitters
4
1.3 Cell Model Capable of Showing Changes in Exocytosis during MW
Exposure
5
1.4 Exposure System
6
1.5 Dissertation Outline
6
1.6 References
8
Chapter 2: Possible Mechanisms for the Interaction between MW Fields
and Biological Systems
13
2.1 Introduction
13
2.2 Exocytosis
14
2.3 Possible Non-thermal Mechanisms Underlying MW-Induced Bioeffects
14
2.3.1 Force Effects on Proteins
15
vii
2.3.1.a Changes in Binding Ligands
16
2.3.1.b Changes in Protein Conformation
17
2.3.1.c Effects Inside of Cells
18
2.3.2 Electrical Current Induced in the Extracellular Medium
19
2.4 Conclusion
20
2.5 References
21
Brief Introduction of Chapter 3
23
Chapter 3: Design, Characterization and Optimization of a Broadband Mini
Exposure Chamber for Studying On Line Monitoring of Catecholamine
Release from Chromaffin Cells Exposed to Microwave Radiation:
Finite-Difference Time-Domain Technique
24
Abstract
25
3.1 Introduction
26
3.2 Methodology
28
3.2.1 Cell Perfusion
28
3.2.2 Instrument Shielding
30
3.2.3 MEC
30
3.2.4 MW Signal Generation
31
3.3 FDTD Modeling
36
3.3.1 Dielectric properties used in the FDTD modeling
36
3.3.2 Horn Antenna
38
3.3.3 Exposure System
39
viii
3.4 Results and Discussion
43
3.4.1 Antenna Gain and Far-Field Radiation Pattern
43
3.4.2 Validation of the effectiveness of the MEC
46
3.4.3 Field and SAR Distribution across the GFF
53
3.4.4 Improving SAR Homogeneity for the Entire 1-6 GHz Frequency
Range
57
3.5 Conclusions
60
3.6 Acknowledgements
62
3.7 References
63
Chapter 4: Improvements to the Perfusion/On-Line Detection System and
Additional Details
66
4.1 Introduction
66
4.2 Pressure Stabilization
66
4.2.1 Compression of the GFF by the CPA Screw
67
4.2.2 Number of Cells Loaded onto the GFF
72
4.2.3 Use of Microtubing
73
4.2.4 Particulates in the BSS
75
4.2.5 Use of a 5 µm Mesh
76
4.3 Improvements to the Injection System
76
4.4 Improvements to the CPA Temperature Controller
77
4.5 Electrochemical Detector (ECD)
79
4.5.1 Principle of Electrochemical Detection
80
ix
4.5.2 Sensitivity and Calibration
82
4.5.3 Linearity
84
4.5.4 Detection Limit
87
4.5.5 Hysteresis
87
4.5.6 Repeatability
88
4.5.7 Specificity and Selectivity
88
4.5.8 Factors that Can Generate Artifacts
89
4.5.8.a Fluctuations in BSS Flow Rate
90
4.5.8.b Addition of an ECD Temperature Controller
93
4.5.9 Area under the Peak
98
4.6 Automation
103
4.7 References
105
Chapter 5: Characterization and Optimization of the Exposure System Based on
Measurements and Numerical Computations
106
5.1 Introduction
106
5.2 Detailed Antenna Characterization
106
5.2.1 S Parameters of the Broadband Horn Antenna
106
5.2.1.a Measurement of the S11 Parameter of the Broadband Horn Antenna
107
5.2.1.b Computation of the S11 Parameter of the Broadband Horn Antenna
108
5.2.1.c Design of the Antenna Feed Port in the XFDTD Model
110
5.2.1.d Validation of the Broadband Horn Antenna
112
5.2.1.e Conclusion
116
x
5.2.2 Three Dimensional Broadband Horn Antenna Radiation Pattern
117
5.2.3 E Field Distribution Inside the MEC
119
5.3 The CPA in the Near-Field Range
5.3.1 E Field Distribution in the Near Field in the Absence of the CPA
122
123
5.3.2 E Field Distributions on the GFF in the Near Field of the Horn Antenna 125
5.3.3 Magnitude of the E Field on the GFF placed in the Near Field
5.4 Use of the Smaller CPA
127
128
5.4.1 Temperature Gradient
128
5.4.2 Improvements to Decrease the Temperature Gradient
130
5.4.3 Distributions and Magnitudes of the E Field on the GFF
130
5.5 Actual Power
5.5.1 Power Cable and Coupler
5.6 References
Chapter 6: Results and Discussion
133
133
137
138
6.1. Introduction
138
6.2 Statistical Analysis Overview
138
6.3 Analysis of Data from Control Experiments
139
6.3.1 Method
140
6.3.2 Typical Experimental Profiles
140
6.3.3 Mathematical Model for the Decline in the Area under the Peak
143
6.3.4 Assessment of the Range of Variation in a Control Experiment
147
6.4 Analysis of Data from Microwave Exposure Experiments
149
xi
6.4.1 Statistical Analysis of Results
149
6.4.2 Summarizing the Results of the MW Exposure Experiments
154
6.4.3 Assessing the Number of Bioeffects
160
6.4.4 Assessing the Magnitude of the Bioeffects
161
6.4.5 Patterns and Magnitudes of the Bioeffects
162
6.4.6 Identifying MW Exposure Parameters that Induce the Most Robust
Bioeffects
6.5 Discussion
163
169
6.5.1 Description of the 5-6 GHz PFS Signal
170
6.5.2 Waveform of the PFS MW Fields
171
6.5.3 Effect of Frequency and Power Window
173
6.5.4 Threshold Value of the E field
174
6.6 Conclusion
175
6.7 References
176
Chapter 7: Future Works
177
7.1. Introduction
177
7.2 Cell Holder and Antenna
177
7.3 Limitations of the Free Space Exposure System and Solutions
177
7.3.1 Large Aperture Size of the Horn Antenna
178
7.3.2 Reflections from the Edges of the GFF
179
7.4 The Newly Designed Exposure System
181
7.5 Temperature Control
184
xii
7.6 References
Appendixes
187
188
Appendix A: Cell Preparation Protocol
188
Appendix B: Cell Loading Protocol
200
Appendix C: Control / MW Experiment Protocol
209
Appendix D: System Cleaning Process
214
Appendix E: E Field and SAR Distribution Calculator (written in Mathcad 2001) 217
Appendix F: Procedure of Statistical Analysis for Identification of Bioeffect
226
Appendix G: ECD Range Setting
234
Appendix G.1. Output Range Setting
235
Appendix G.2. Background Current of the ECD
236
Appendix G.3. Selection of the Flowcell
239
Appendix G.4. Selection of the Flow Rate
239
Appendix G.5. A Fixed Output Range
240
Appendix G.6. Responses in the Non-Linear Range of the ECD
242
Appendix G.7. Conclusion
244
Appendix G.8. References
244
Appendix H: LabVIEW Programs
245
Appendix H.1. Pump Controller Program
246
Appendix H.2. Temperature Controller Program
246
Appendix H.3. Injector Controller Program
247
Appendix H.4. ECD Monitoring Program
247
xiii
Appendix H.5. MW Controller Program
248
Appendix H.6. Amplifier Controller Program
248
Appendix H.7. Power Monitoring Program
249
Appendix H.8. ECD Inlet Temperature Controller
249
Appendix H.9. Data Logging Program
250
Appendix I: Unreliable Experiments
251
xiv
List of Figures
2.1. Schematic diagram of the process of exocytosis. After acetylcholine (Ach, green
dots) binds to nicotinic cholinergic receptors ① and the chromaffin cell becomes
depolarized, influx of calcium (blue dot) via voltage gated calcium channels occurs ②
and leads to fusion of the catecholamine containing vesicles with the plasma membrane
③ and subsequent release of catecholamines (red dots) to the extracellular space ④.
2.2. Simplified electrical model of a cell membrane. The plasma membrane can be
expressed as a resistor and a capacitance in parallel.
3.1. Photograph (a) of the CPA showing the location of the glass inlet tubing wound with
nichrome wire, and where temperature probes are placed and (b) SolidWorks model of
the CPA, showing an exploded view of the filter holder and its components.
3. 2. Photograph showing a close-up view of the MEC.
3.3. Photograph (a) and SolidWorks diagram (b) of the spiral antenna (all dimensions are
in mm).
3.4. Schematic (a) of the entire free space MW exposure system showing the
experimental arrangement inside the anechoic chamber and adjacent screen room;
Photograph of the exposure system inside the anechoic chamber.
(b)
3.5. Measured complex relative permittivity (dielectric constant ε΄, dielectric loss ε˝) and
conductivity of GFF soaked with BSS over the frequency range 1 to 6 GHz.
3.6. Photograph (a) and SolidWorks model (b) of the broadband horn antenna.
3.7. XFDTD model (numerical mesh) of the entire exposure system. (a) 3D perspective
view and (b) cross sectional view through the central horizontal plane.
3.8. SolidWorks model (a) and (b) XFDTD model (3D perspective view) of the CPA;
(c) XFDTD cross sectional view of the gasket and the GFF that is composed of 3.5 mm
base cell size and 200 µm adaptive mesh cell size. The GFF has an overall radius of 12
xv
mm, with an inner radius of 11 mm available to the cells due to the 2 mm wide gasket. (d)
Close up view of the center cross section of the XFDTD mesh showing the BSS layer.
3.9. Comparison (a) of XFDTD computed and measured gains of the broadband horn
antenna; (b) Comparison of the XFDTD computed and measured H plane radiation
pattern and (c) Comparison of the XFDTD computed and measured E plane radiation
pattern.
ur
3.10. E field in the central horizontal plane that contains the GFF. (a) XFDTD
contour plot; (b) XFDTD surface plot of the region shown in (a) by the rectangular box.
ur
(c) Histogram and descriptive statistics of the XFDTD computed E field distribution
for the region in (9) within the MEC in the plane containing the GFF. (d) Comparison of
XFDTD (solid line) and EIRP results (dashed line). Input power to the horn antenna =
250 W.
3.11. Comparison of measured (solid line) and EIRP results (dashed line) at 3.3 GHz
ur
where the inhomogeneity in the E field distribution was the greatest. Input power to
the horn antenna = 10 dBm.
ur
3.12. Measured E field in the region containing the GFF (shown by the shaded region
in Figure 2. 9d). Frequency = 3.5GHz and input power to the horn antenna = 10 dBm.
ur
3.13. XFDTD computed E field contour plot for 1 GHz (a), 3.5 GHz (c) and 6 GHz (e)
ur
and histogram and descriptive statistics of the E field distribution for 1 GHz (b), 3.5
GHz (d) and 6 GHz (f). All results are normalized and are for the plane of the GFF.
Input power to the horn antenna = 250 W.
3.14. XFDTD computed SAR contour plot for 1 GHz (a), 3.5 GHz (c) and 6 GHz (e) and
histogram and descriptive statistics of the SAR distribution for 1 GHz (b), 3.5 GHz (d)
and 6 GHz (f). All results are normalized and are for the plane of the GFF. Input power
to the horn antenna = 250 W. Total absorbed power at 1, 3.5 and 6 GHz is 0.00631, 9.478
and 2.518 mW, respectively.
ur
3.15. XFDTD computed contour plot of the E field (a) and SAR (c), and histogram and
ur
descriptive statistics of the E field (b) and SAR (d) for a GFF having a region of radius
xvi
7 mm available to cells at 6 GHz. Total absorbed power = 1.838 mW.
4.1. Schematic of the current improved free space MW exposure system showing the
experimental arrangements inside the anechoic chamber and adjacent screen room.
4.2. Drawing of (a) the upper and lower halves of the CPA, (b) cross-sectional view of the
assembled CPA and (c) blown-up view of the components circled in blue in (b).
4.3. The CPA assembled with the rubber gaskets, 350 µm meshes, 5 µm mesh and GFF
when (a) the GFF is maintained at the original thickness and (b) the GFF is compressed
by the nylon meshes due to over tightening.
4.4. System pressure monitored by the pressure meter location shown in (Figure 4.1). The
graph shows an increase in the system pressure over time due to over tightening of the
CPA.
4.5. Illustration of the CPA with 0.6 mm gaskets (a). The E field distribution over the
GFF with (b), and without (c) the additional BSS layer.
4.6. Illustration of the CPA using a 0.4 mm and a 0.6 mm gasket.
4.7. Illustration of the GFF (a) with evenly distributed cells and (b) with too many cells
present.
4.8. Diagram of perfusion system showing the locations of the microtubing. Microtubing
① is the DMPP injection tubing, microtubing ② is the inlet tubing for the ECD and
microtubing ③ is the outlet tubing of the system to maintain constant pressure.
4.9. Diagram of the injection system (a) before and (b) after modification.
4.10. Detailed schematics of the CPA and surrounding components (a) before and (b)
after modifications. In the modified design, the temperature probes are located much
closer to the GFF and hence the cells.
xvii
4.11. Cross-flow electrochemical transducer. (a) Exploded view [6] and (b) illustration of
a top cross-sectional view of the transducer showing the path of the BSS (blue line), the
low-volume thin layer cell (circled in red) and the glassy carbon electrode.
4.12. Calibration curve for the ECD for an output range setting of 0 – 50 nA, showing
that the response is linear with the amount of injected CA (dotted line).
4.13. Illustration of the non-linear curve (red curve) and the compensated linear curve
(dotted line).
4.14. Calibration curve for the ECD including the non-linear range (50-500 nA) when the
output range is set to 50 nA. The response is non-linear between 470 to 500 nA.
4.15. Calibration curve for the ECD in the non-linear range (50 to 500 nA) and the
detection limit when the output range is set to 50 nA. Circled in red is an exploded view
of the top of a peak that is outside the detection limit.
4.16. Response of the ECD to repeated injections of a constant volume of CA standard.
Injections were delivered every 10 minutes for up to four hours.
4.17. ECD output is proportional to the flow rate of the BSS. Flow rate was gradually
increased and then decreased by 0.4 ml/min.
4.18. Schematic diagram of how the ECD output is affected by stopping the flow of the
BSS for 10 seconds.
4.19. Effect of temperature on the ECD response. (a) Diagram showing the flowcell [6]
with a resistor based thermometer mounted on the wall of the auxiliary electrode block
and (b) the CA standard response while the temperature of the BSS was increased.
4.20. Response of the ECD during exposure of the CPA to a MW field at 3.5 GHz using
maximum MW power. Heating of the BSS by the MW field caused the baseline response
of the ECD to change.
4.21. Diagram of the computer controlled temperature controller that maintains the
xviii
temperature at the inlet of the flowcell to within ± 0.45 oC of the computer controller set
point.
4.22. Experimental profile showing the temperature at the CPA inlet and CPA outlet (label
1) and the ECD inlet (label 2) with the temperature controller either turned off (left panel)
or turned on (right panel). Pressure and flow rate (label 3) and power (label 4) are also
shown. When maximum MW power was applied at 3.5 GHz, the temperature at the ECD
inlet increased by 2 oC when the ECD inlet temperature controller was off and was
maintained constant to within 35 ± 0.45 oC when the ECD inlet temperature controller
was on.
4.23. Height of and area under the CA peak.
4.24. Simplified diagram of Figure 4.1 showing the points at which concentration of the
DMPP and the released CA, where location (1) is the outlet of the drug injection, (2) is
the inlet of the CPA and (3) is the inlet of the ECD.
4.25. Illustration of (a) a bolus of injected DMPP flowing in a tubing and the trailing of
DMPP that occurs along the walls due to viscosity of the BSS. The simulated speed of the
flow of DMPP in a cross-sectional view through (b) the vertical plane and (c) the center
horizontal plane of a tubing, obtained using COSMOSFloWorks, are also shown. The
inner diameter of the tubing used for the simulation was the same as that comparing the
actual setup.
4.26. The cross-sectional view of the CPA showing multiple flow paths.
4.27. Illustration of the concentration profile of the CA at the inlet of the ECD.
5.1. The measured S11 parameter of the broadband horn antenna over the frequency range
1 to 6 GHz.
5.2. Comparison of the measured and calculated S11 parameter of the broadband horn
antenna.
5.3. Cross-sectional view of the center horizontal plane of (a) the entire broadband horn
xix
antenna and (b) an exploded view of the antenna’s feeding components (yellow circle in
(b))
5.4. SolidWorks model of the four step matching port that is connected to the input port
of the horn antenna to reduce the capacitance problem.
5.5. The SolidWorks model (a) and (b) XFDTD model (cross-sectional view) of two four
step matching ports connected back to back.
5.6. The calculated S parameters and impedance of two four step matching ports
connected back to back.
5.7. (a) Final XFDTD mesh of the horn antenna with the matching feed port, (b) crosssectional view of the horn antenna, and (c) close-up view of the feed port.
5.8. Measured and computed S11 parameter of the broadband horn antenna.
5.9. The experimental setup that measures the three dimensional radiation pattern.
5.10. Measured (a) E plane patterns and (b) H plane patterns for 3.5 GHz.
5.11. Illustration of the system that measures the E field at the center horizontal plane that
contains the GFF.
5.12. E field distributions for the frequency range 1–6 GHz in steps of 0.1 GHz (total of
50 frequencies) measured using a computer controlled electromechanical actuator and a
spiral antenna.
5.13. Illustration of size comparison of the spiral antenna and the GFF.
5.14. Computed near field distribution at the plane of the GFF. Frequency = 3.5 GHz.
5.15. Photograph of the horn antenna (a) with a red line indicating where the E field is
calculated; (b) magnitude of the E field along the red line shown in (a).
5.16. Exploded view of Figure 3.1 that shows the region where the E field magnitude is
xx
high and homogeneous regardless of the operating frequency.
5.17. Comparison of the E field distributions in the near and far fields at 1, 3.5 and 6 GHz.
5.18. The cause of the improvement to the homogeneity of the E field. The red, green and
blue rays represent the direct, diffracted and reflected waves, respectively.
5.19. Photograph of the cells on the GFF showing that the cells are uniformly distributed
over the entire GFF. For visualization, the cells were stained with the dye neutral red.
5.20. Measured (a) and calculated (b) temperature distribution on the 24 mm diameter
GFF
5.21. Measured (a) and calculated (b) temperature distribution on the GFF of diameter 10
mm.
5.22. Comparison of the E field distributions when the original larger CPA is placed in
the far and near fields with the E field distribution when the smaller CPA is placed in the
near field at 1, 3.5 and 6 GHz.
5.23. Measured attenuation due to the high power cable with the coupler attached to one
end.
6.1. A typical CA release profile for a control experiment.
6.2. Bar graph showing the area under the CA peaks for the experiment shown in Figure
6.1.
6.3. Normalized results of all control experiments performed on cells from different cell
preparations and at different lengths of time in culture.
6.4. Average of peaks ± S.D. The numbers on the graph indicate the number of peaks
used to calculate the average.
xxi
6.5. Areas under the CA peaks for a control experiment, superimposed with a trend line,
as calculated by SigmaPlot.
6.6. Control experiments performed on different dates showing the trend lines calculated
using SigmaPlot.
6.7. Data from an actual experiment depicting the CA peaks, the calculated trend line and
the 95 % prediction band.
6.8. Experimental profile of a representative MW exposure experiment. Red arrows
indicate a decrease and blue arrows indicate an increase relative to neighboring peaks.
6.9. The areas under the CA peaks and the trend line for the MW exposure experiment
shown in Figure 6.8. The orange arrows indicate peaks lower than the trend line and the
green arrows indicate peaks higher than the trend line.
6.10. Normalized trend lines and CA peaks of the MW exposure experiment performed
on 04/02/2008 and the normalized trend line of the corresponding control experiment.
The 95 % prediction band for the control experiment is also shown.
6.11. The final result of the Mathcad analysis program showing the normalized control
trend line, 95 % prediction band, modified MW exposure trend line, and modified CA
peaks for the MW exposure experiment performed on 04/02/2008.
6.12. Analysis procedure for the MW exposure experiment performed on 02/12/2008
showing (a) the area under the CA peaks with the trend line, (b) the statistical result
generated by the Mathcad program (Appendix F) and (c) a table summarizing the results
shown in (b).
6.13. Cellular responses to DMPP and the corresponding trend lines for 32 MW exposure
experiments. The dates on which the experiments were performed are also shown.
6.14. Histogram of the percent difference for the peaks that indicate a bioeffect.
6.15. Illustration of the three main patterns of MW bioeffects. The bioeffect occurs (a)
xxii
during, (b) after, and (c) both during and after MW exposure.
6.16. Profile obtained for cells exposed to a 5-6 GHz PFS (sweep time of 0.5 s) field 5 to
6 GHz showing a robust bioeffect.
6.17. Profile obtained for a second experiment performed immediately after the
experiment shown in Figure 6.16.
6.18. Illustration of the 5-6 GHz frequency sweep in steps of 0.2 GHz with a sweep time
of 500 ms.
6.19. Illustration of the 5-6 GHz PFS with a sweep time of 500 ms, a pulse width of 100
ms and a repetition rate of 1 Hz.
6.20. Illustration of the 5-6 GHz PFS with a sweep time of 500 ms, a pulse width of 100
ms and a repetition rate of 1 Hz at the location of the cells. The magnitudes of the wave at
different frequencies vary since the gain of the amplifier, the gain of the broadband horn
antenna and the loss of the power cable vary at different frequencies. The magnitudes
were found using results obtained by XFDTD and the measured cable loss.
7.1. Horn antenna. (a) Photograph showing the plane where the magnitude of the E field
was measured at 3.5 GHz and (b) the measured E field distribution over the plane shown
in (a).
7.2. Photograph of the horn antenna (a) together with a depiction of the relative size of
the GFF. (b) Illustration of the flare structure in the horn antenna in free space showing
that only a small amount of the MW field (dotted line in red) propagates toward the GFF.
(c) Illustration of the flare structure embedded into a high dielectric substrate showing
that a larger amount of the MW field (dotted line in red) propagates toward the GFF.
7.3. Illustration of a cross-sectional view of the CPA with the incident and reflected MW
fields at the various dielectric interfaces.
7.4. SolidWorks model of (a) the Vivaldi antenna and dielectric substrate, (b) the Vivaldi
antenna embedded in the dielectric substrate and (c) a GFF located in front of the Vivaldi
xxiii
antenna.
7.5. SolidWorks drawing of the entire exposure system.
7.6. A calculated snapshot of the E field propagating in the plane of the GFF computed
using XFDTD (Frequency = 3.5 GHz).
7.7. The E field distributions in the region of the GFF embedded in the newly designed
exposure system at 1, 3.5 and 6 GHz. % refers to the area of the GFF homogenous to
within 30 % (% refers to the area of the GFF homogeneous to within 30 %).
7.8. Exploded view of the upper block of the exposure system where a cooling system is
implemented. The coolant will flow through the surface of the substrate material to
control the temperature at the location where the cells are immobilized during MW
exposure.
xxiv
List of Tables
3.1. Measured reflectivities of the absorber material as a function of frequency and angle
of incidence.
ur
3.2 Homogeneity and mean values of E field and SAR, as well as total absorbed power
in the GFF computed using XFDTD at various frequencies in the range 1 - 6 GHz.
3.3 Variation in homogeneity as a function of radius of the GFF versus frequency.
5.1. The calculated E field magnitudes on the GFF for the near and far field.
5.2. Comparisons of the mean, maximum and minimum values of the E field magnitude
on the GFF for the original larger CPA placed in the far and near fields with the values for
the smaller CPA placed in the near field of the horn antenna.
5.3. Maximum power output of the amplifier at different frequencies for CW.
5.4. Adjusted E field magnitudes (rectangle in red) based on the maximum power and the
cable loss.
6.1. Summary of results for each MW exposure experiment.
6.2. Summary of the number of bioeffects (data from Table 6.1).
6.3. Summary of the patterns and magnitude of the 91 peaks showing a bioeffect.
6.4. Analysis of the results for all MW experiments using CW, AM, pulse keying and
Gaussian pulses.
6.5. Analysis of the results for PFS MW experiments.
6.6. Summary of the results shown in Table 6.4 and Table 6.5.
xxv
6.7. Comparison of the results for 5-6 GHz PFS and those for other frequency ranges.
6.8. Comparison of the results for exposures with 10 ns, 100 ms and 100 µs pulse widths
for 5-6 GHz PFS.
6.9. Magnitude of the E field at different frequencies. The magnitude is increased as
frequency is increased except at 2 GHz due to the large reflections at this frequency
(Figure 5.8).
7.1. Comparison of the magnitude of the E field on the GFF for the free space exposure
system and the newly designed system.
1
Chapter 1
General Introduction
It has been reported that microwave (MW) fields can affect biological systems
that range from microbes to animals, including humans [1-5]. The interaction between a
MW field and a biological system can take place through either thermal or non-thermal
mechanisms [1, 4, 6-9]. Thermal mechanisms are likely due to changes in temperaturedependent biochemical reactions in tissues, whereas non-thermal mechanisms are those
that are not directly associated with a temperature change but instead are due to an
interaction of the MW electric field with cellular constituents. These latter interactions
due to MW fields have been proposed to occur in several ways and include the following:
(1) an alteration in the transport or orientation of ions that can interrupt electrical and
chemical signals that go from the outside of a cell to the inside of a cell [10, 11], (2)
changes in the binding of ligands to cell surface receptors that can alter the function of
the receptors [8, 12, 13] and (3) changes in protein conformation that can affect the
conductance of ion channels, and the activity of enzymes and receptors [14, 15].
Moreover, when bioeffects due to MW fields have been reported, it was observed that
these effects occurred only in certain bands of frequencies, and at certain field intensities
and modulations [1, 16-18, 19]. Thus, specific parameters of MW fields (i.e., specific
frequency, power, and modulation schemes) appear to be required to induce non-thermal
bioeffects.
To date, the majority of research on MW-induced bioeffects has been carried out
to address the growing concern for widespread use of wireless communication systems.
2
Thus, most MW parameters found to induce bioeffects are those commonly used in
telecommunications, for example, continuous wave (CW) and amplitude or pulse
modulated waves at 0.9, 1.71, 1.8, and 2.45 GHz [19-24]. Hence, many bioeffects and
possible interactions between MW fields and biological components have been reported
at these specific frequencies [25, 26]. Other researchers have searched for potential
therapeutic applications of MW fields at these and other frequencies and these efforts
have led to the use of MW fields in a variety of clinical settings, such as MW-induced
stimulation of tissue repair and regeneration [1]. However, the mechanisms underlying
these effects are not clear. Hence, more research is needed to provide reproducible and
predictive results [4].
The research described in this dissertation is directed at the potential development
of novel clinical applications of MW fields. To follow up on studies that have reported
nonthermal effects of MW fields on biological systems, the frequency range investigated
in this research was chosen to be 1–6 GHz, which includes frequencies used in most of
the modern telecommunication systems. Also, the depth of penetration in tissue in this
frequency range is several inches [7], which is important since the goal is to use MW
fields for therapeutic applications.
The ultimate biological target in this study is the nervous system, since numerous
studies have reported that MW fields can elicit non-thermal effects on nervous system
functions. A review of the studies that have provided insights into how the nervous
system is affected by MW fields is presented in the following sections.
3
1.1 In Vivo Studies Addressing MW-Induced Effects on the Nervous
System.
There have been occasional reports of headache, lassitude, sleeplessness and
irritability among workers in the vicinity of MW generating equipment that radiates low
levels of MW fields (<10 W/kg for occupational exposure), indicating that humans
exposed to low levels of MW fields experience nervous system effects [1]. More direct
evidence for the existence of possible non-thermal effects on the nervous system have
been derived from controlled investigations on the activity of brain such as,
electroencephalograms (EGG) or assessing cognitive functions in animals and human,
during exposure to CW and pulse modulated MW fields within the frequency range 0.9 to
6 GHz [1-5, 27-30]. As examples, MW fields at 2.45 and 4 GHz increased the total power
of EEG spectra in rats [31] and in rabbits 2.45 GHz MW fields increased the number of
spindle-shaped firings and slow waves in the EEG [32] accompanied same study also
reported a change in the discharge frequency of neurons in the visual cortex [32]. An
assessment of MW effects on humans performed at 0.9 GHz has shown an increase of the
REM EEG spectral power density [33], a suppressive effect on REM sleep [33] and, in
some cases, a sleep-inducing effect [34]. At 2.9 GHz, MW exposure evoked changes in
the waking and sleep EEG pattern [35].
Investigations of MW fields on cognitive performance showed a significant
increase in response time and an increase in memory load in humans during an n-back
task by 0.9 GHz MW fields [36], a deficit in the spatial working memory function of rats
by 2.45 GHz MW fields [37], and a field-dependent improvement in immediate memory
in human subjects after exposure to a 1.8 GHz field [38].
4
Since the synaptic release of neurotransmitters is a primary mechanism for
intercellular communication in the central nervous system that could underlie the effects
just described, MW-induced effects on neurotransmitter systems are reviewed in the next
section.
1.2 In Vivo Studies Addressing MW-Induced Effects on
Neurotransmitters
Earlier studies reported changes in various neurotransmitters (catecholamines,
serotonin, and acetylcholine) in the brain of animals after exposure to high intensity
(<8 W/kg) MW fields [39-41]. However, the effects reported in those studies were
thought to be thermal in nature.
Other studies performed at lower intensities (< 10 mW/cm2) have shown a
decrease of cholinergic function in rats in a time-dependent fashion due to exposure to
pulsed or CW 2.45 GHz MW fields [30], an increase in epinephrine levels in the brain of
rats after 2.375 GHz MW exposure [42], a decrease in acetylcholineesterase (AChE)
activity in guinea pig brain by pulsed 3 GHz MW fields [43], a decrease in activity of
succinate dehydrogenase in the hypothalamus, and of hypothalamic catecholamine (CA)
and monoamine oxidase in rats by MW exposure at 3 GHz [44], and a significant
decrease (40 %) in mean AChE release from rat hippocampus in response to 2.45 GHz
fields [45].
While in vivo studies have provided evidence of non-thermal MW-induced effects
on central neurotransmitters, in vitro investigation of non-thermal MW effects on
neurotransmitters at the frequency range of interest (1-6 GHz) have, to the best of our
knowledge, not yet emerged. The goal of this research is to identify MW exposure
5
parameters that can elicit effects on exocytosis, the process by which neurotransmitters
are released from nerve terminals. The biological system that was used for these studies
is described in the next section.
1.3 Cell Model Capable of Showing Changes in Exocytosis during MW
Exposure
In vitro research that would enable us to investigate the effect of MW fields on
neurotransmitter release utilized a well-established model of neural-type cells, isolated
bovine adrenal chromaffin cells that synthesize, store, and release CA. The release of CA
from these cells occurs via a process called exocytosis. As stated above, this is the same
process by which neurotransmitters are released from nerve terminals and hence
underlies the primary mechanism for intercellular communication in the nervous system.
Chromaffin cells release the CA epinephrine and norepinephrine either by spontaneous
depolarization or in response to activation of nicotinic acetylcholine receptors by agonists
such as 1,1-dimethy-l-4-phenylpiperazinium (DMPP) [7, 8, 46-51]. Hence, if MW
parameters that can affect exocytosis either in the presence or absence of a nicotinic
receptor stimulus can be identified, this could potentially lead to clinical applications,
such as the ability to modulate nervous system function in a non-invasive and controlled
manner in patients with neurological diseases. Chromaffin cells provide an excellent cell
model for this research because CA release can be monitored on-line by electrochemical
detection [52-55] during MW exposure. A similar approach has been used by this
laboratory for the frequency range 750-1000 MHz [56]. A brief explanation of the MW
exposure system used for the 1-6 GHz frequency range is described in the next section.
6
1.4 Exposure System
As discussed in the previous section, the goal of this research is to identify MW
parameters that affect the process of exocytosis in chromaffin cells, using applied MW
fields in the 1-6 GHz range. In order to identify specific MW parameters that induce
effects over this broad frequency range, the approach taken was to actively change MW
parameters, such as frequency, modulation and power level during a single exposure in
order to generate complex E field patterns. This approach contrasts with previously
reported studies that used a fixed frequency, such as one commonly used by
telecommunication systems. Hence, the MW radiation devices used in the other studies,
such as a waveguide [22, 23, 57-59], horn antenna [21], and a coplanar waveguide [45,
60], were not adaptable for our application due to their narrow frequency band. Instead,
our study employed a broadband horn antenna based free-space type of MW exposure
system that was capable of monitoring released CA on-line. The design and
characterization were carefully performed since many previously reported results were
subject to technical criticism or the effects reported simply were not present when followup studies were conducted under better controlled conditions [13, 61, 62].
1.5 Dissertation Outline
In this dissertation, Chapter 2 discusses in more detail the possible mechanisms of
interaction between MW fields and biological systems. Chapters 3-5 provide a detailed
description of the design, characterization and optimization of the in vitro 1-6 GHz free
7
space exposure system. The results of series of carefully controlled exposure experiments
using amplitude modulation (AM), frequency modulation (FM), pulse modulation, phase
modulation (PM), Gaussian pulse and pulsed frequency sweep (PFS) are described in
Chapter 6 where the MW parameters that induce the most robust bioeffects are identified.
Chapter 7 describes a novel exposure system that overcomes some of the identified
limitations of the current exposure system and is proposed for future studies.
1.6 References
8
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
S. Banik, S. Bandyopadhyay, and S. Ganguly, "Bioeffects of microwave–a brief
review," Bioresource Technology, vol. 87, pp. 155–159, 2003.
J. A. D’Andrea, C. K. Chou, S. A. Johnson, and E. R. Adair, "Microwave effects
on the nervous system," Bioelectromagnetics Supplement, vol. 24, pp. 107-147,
2003.
M. H. Repacholi, "Low-Level Exposure to Radiofrequency Electromagnetic
Fields: Health Effects and Research Needs," Bioelectromagnetics, vol. 19, pp. 1-9,
1998.
M. Swicord and Q. Balzano, "Has Electromagnetic Energy in the Band 0.1–100
GHz Useful Medical Applications? A Review of Mechanisms and Biological
Database Offers Dim Prospects," IEEE Transactions on Plasma Science, vol. 36,
pp. 1638-1649, 2008.
J. R. Jauchem, "Effects of low-level radio-frequency (3 KHz to 300 GHz) energy
on human cardiovascular, reproductive, immune, and other systems: A review of
the recent literature," International Journal of Hygiene and Environmental Health,
vol. 211, pp. 1-29, 2008.
K. R. Foster, "Thermal and Nonthermal Mechanisms of Interaction of RadioFrequency Energy with Biological Systems," IEEE Transactions on Plasma
Science, vol. 28, pp. 15-23, 2000.
H. P. Schwan, "Interaction of Microwave and Radio Frequency Radiation with
Biological Systems," IEEE Transactions on Microwave Theory and Techniques,
vol. 19, pp. 146-152, 1971.
L. J. Challis, "Mechanisms for Interaction Between RF Fields and Biological
Tissue," Bioelectromagnetics Supplement, pp. 98-106, 2005.
M. S. Markov, "Pulsed electromagnetic field therapy history, state of the art and
future," Environmentalist, vol. 27, pp. 465-475.
R. Adey, "Effects of microwaves on cells and molecules," Nature, vol. 333, pp.
401, 1988.
F. Barnes and Y. Kwon, "A Theoretical Study of the Effects of RF Fields in
theVicinity of Membranes," Bioelectromagnetics, vol. 26, pp. 118-124, 2005.
A. Chiabrera, B. Bianco, E. Moggia, and J. J. Kaufman, "Zeeman-Stark Modeling
of the RF EMF InteractionWith Ligand Binding," Bioelectromagnetics, vol. 21,
pp. 312-324, 2000.
C. J. Thompson, Y. S. Yang, V. Anderson, and A. W. Wood, "A cooperative
model for Ca++ efflux windowing from cell membranes exposed to
electromagnetic radiation," Bioelectromagnetics, vol. 21, pp. 455-464, 2000.
B. Wang, C. Liu, Hong.Zhang, and R. Fan, "Review on the Mechanisms of NonThermal Bioeffects of Electromagnetic Waves and the Experimental Study on
Bioeffects of Transient Electromagnetic Fields on Cells and Its Applications,"
presented at Infrared and Millimeter Waves, 2000. Conference Digest. 2000 25th,
Beijing, 2000.
J. A. Laurence, P. W. French, R. A. Lindner, and D. R. Mckenzie, "Biological
Effects of Electromagnetic Fields-Mechanisms for the E4ects of Pulsed
Microwave Radiation on Protein Conformation," Journal of Theoretical Biology,
vol. 206, 2000.
9
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
E. Postow and M. L. Swicord, Modulated fields and “window” effects.
S. Lin-Liu and W. R. Adey, "Low frequency amplitude modulated microwave
fields change calcium efflux rates from synaptosomes," Bioelectromagnetics, vol.
3, pp. 309-322, 1982.
K. R. Foster and M. H. Repacholi, "Biological effects of radiofrequency fields:
Does modulation matter?" Radiation Research, vol. 162, pp. 219-225, 2004.
H. Hinrikus, J. Lass, and V. Tuulik, "Low-level Microwave Effects on EEG,"
presented at 22nd Annual EMBS International Conference, Chicago IL, July 23-28,
2000.
S. Inomata-Terada, S. Okabe, N. Arai, R. Hanajima, YasuoTerao, T. Frubayashi,
and Y. Ugawa, "Effects of High Frequency Electromagnetic Field (EMF) Emitted
by Mobile Phones on the Human Motor Cortex," Bioelectromagnetics, vol. 28, pp.
553-561, 2007.
G. Kubinyi, G. Thuroczy, J. Bakos, E. Boloni, H. Sinay, and L. D. Szabo, "Effect
of Continuous-Wave and Amplitude-Modulated 2.45 GHz Microwave Radiation
on the Liver and Brain Aminoacyl-Transfer RNA Synthetases of in Utero
Exposed Mice," Bioelectromagnetics, vol. 17, pp. 497-503, 1996.
S. Lee, D. Johnson, K. Dunbar, H. Dong, X. Ge, Y. C. Kim, C. Wing, N.
Jayathilaka, N. Emmanuel, C. Q. Zhou, H. L. Gerber, C. C. Tseng, and S. M.
Wang, "2.45 GHz radiofrequency fields alter gene expression in cultured human
cells," FEBS Letters, vol. 579, pp. 4829–4836, 2005.
A. G. Pakhomov, P. Gasek, L. Allen, B. E. Stuck, and M. R. Murphy,
"Comparison of Dose Dependences for Bioeffects of Continuous-Wave and HighPeak Power Microwave Emissions Using Gel-Suspended Cell Cultures,"
Bioelectromagnetics, vol. 23, pp. 158167, 2002.
J. L. Kiel, J. E. Parker, P. J. Morales, J. L. Alls, Patrick A. Mason, R. L. Seaman,
S. P. Mathur, and E. A. Holwitt, "Pulsed Microwave Induced Bioeffects," IEEE
TRANSACTIONS ON PLASMA SCIENCE, vol. 28, pp. 161-167, 2000.
E. Maby, R. L. B. Jeannes, G. Faucon, C. Liegeois-Chauvel, and R. D. Seze,
"Effects of GSM Signals on Auditory Evoked Responses," Bioelectromagnetics,
vol. 26, pp. 341-350, 2005.
D. Hamblin, A. Wood, R. Croft, and C. Stough, "Examining the effects of
electromagnetic fields emitted by GSM mobile phones on human event-related
potentials and performance during an auditory task," Clinical Neurophysiology,
vol. 115, pp. 171-178, 2004.
K.-A. Hossmann and D. M. Hermann, "Effects of Electromagnetic Radiation of
Mobile Phones on the Central Nervous System," Bioelectromagnetics, vol. 24, pp.
49-62, 2003.
H. Lai, J. Zabawska, and H. A, "Sodium dependent high choline uptake in
hippocampus and frontal cortex of the rat affected by actual restraint stress,"
Brain Research, vol. 66, pp. 537–540, 1986.
H. Lai, A. Horita, C. K. Chou, and A. W. Guy, "A Review of microwave
irradiation and actions of psychoactive drugs," IEEE Engineering in Medicine and
Biology, vol. 6, pp. 31–36, 1987.
10
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
H. Lai, "Research on the neurological effects of nonionizing radiation at the
University of Washington," Bioelectromagnetics, vol. 13, pp. 513–526, 1992.
G. Thuroczy, G. Kubinyi, M. Bodo, J. Bakos, and L. D. Szabo, "Simultaneous
Response of Brain Electrical Activity (EEG) and Cerebral Circulation (REG) to
Microwave Exposure in Rats," Reviews on Environmental Health, vol. 10, pp.
135-148, 1994.
R. A. Chizhenkova, "Slow Potentials and Spike Unit Activity of the Cerebral
Cortex of Rabbits Exposed to Microwaves," Bioelectromagnetics, vol. 9, pp. 337345, 1988.
K. Mann and J. Röschke, "Effects of Pulsed High-Frequency Electromagnetic
Fields on Human Sleep," Neuropsychobiology, vol. 33, pp. 41-47, 1996.
P. Wagner, J. Roschke, K. Mann, W. Hiller, and C. Frank, "Human Sleep under
the Influence of Pulsed Radiofrequency Electromagnetic Fields: A
Polysomnographic Study using Standardized Conditions," vol. 19, pp. 199-202,
1998.
R. Huber, T. Graf, K. A. Cote, L. Wittmann, E. Gallmann, D. Matter, J. Schuderer,
N. Kuster, A. A. Borbely, and P. Achermann, "Exposure to Pulsed HighFrequency Electromagnetic Field during Waking Affects Human Sleep EEG,"
NeuroReport, vol. 11, pp. 3321-3325, 2000.
M. Koivisto, C. M. Krause, A. Revonsuo, M. Laine, and H. Hamalainen, "The
Effects of Electromagnetic Field Emitted by GSM Phones on Working Memory,"
NeuroReport, vol. 11, pp. 1641-1643, 2000.
H. Lai, A. Horita, and A. W. Guy, "Microwave Irradiation Affects Radial Arm
Maze Performance in the Rat.," Bioelectromagnetics, vol. 15, pp. 95-104, 1994.
J. W. Smythe and B. Costall, "Mobile Phone Use Facilitates Memory In Male, but
not Female, Subjects," Neuroreport, vol. 14, pp. 243-246, 2003.
C. N. Catravas, J. B. Katz, J. Takenaga, and J. R. Abbott, "Biochemical changes
in the brain of rats exposed to microwaves of low power density (symposium
summary)," J Microwave Power, vol. 11, pp. 147-148, 1976.
A. T. Modak, W. B. Stavinoha, and U. P. Dean, "Effect of short electromagnetic
pulses on brain acetylcholine content and spontaneous motor activity in mice,"
Bioelectromagnetics, vol. 2, pp. 89-92, 1981.
S. H. Snyder, "The effect of microwave irradiation on the turnover rate of
serotonin and norepinephrine and the effect of microwave metabolizing
enzymes," U.S. Army Medical Research and Development Command, 1971.
A. N. Grin, "Effects of microwaves on catecholamine metabolism in brain," US
Joint Pub Research Device Rep JPRS 72606, 1974.
S. Baranski, "Histological and histochemical effects of microwave irradiation on
the central nervous system of rabbits and guinea pigs," American Journal of
Physical Medicine, vol. 51, pp. 182-190, 1972.
H. Chiang and G. D. Yao, "Effects of Pulsed Microwave Radiation Pre -and PostNatally on the Developing Brain in Mice," Journal of Bioelectricity, vol. 6, pp.
197-204, 1987.
11
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
[53]
[54]
[55]
[56]
[57]
[58]
G. Testylier, L. Tonduli, R. Malabiau, and J. C. Debouzy, "Effectsof Exposureto
LowLevel Radiofrequency Fields on Acetylcholine Release in Hippocampus of
Freely Moving Rats," Bioelectromagnetics, vol. 23, pp. 249-255, 2002.
J. Teng, D. C. d. Toural, F. Duhamel, and A. V. Vorst, "No nonthermal effect
observed under microwave irradiation of spinal cord," IEEE Transactions on
Microwave Theory and Techniques, vol. 44, pp. 1942–1948, 1996.
G.D’Inzeo, P. Bernardi, S. Pisa, and L. Tarricone, "Ionic channels gating under
EM exposure: A stochastic model," Bioelectrochemistry and Bioenergetics, vol.
29, pp. 290-304, 1993.
L. Tarricone, "Estimation of the physiological parameters of a membrane channel
exposed to electromagnetic fields using the GPOF approach," Electro- &
Magnetobiology, vol. 19, pp. 259270, 2000.
A. W. Chiabrera, M. Grattarola, and R. Viviani, "Interaction between
electromagnetic fields and cells: microelectrophoretic effect on ligands and
surface receptors," Bioelectromagnetics, vol. 5, pp. 173-191, 1984.
G. D'Inzeo, P. Bernardi, F. Eusebi, F. Grassi, C. Tamburello, and B. M. Zani,
"Microwave effects on acetylcholine-induced channels in cultured chick
myotubes," Bioelectromagnetics, vol. 9, pp. 363-372, 1988.
S. K. Dutta, B. Ghosh, and C. F. Blackman, "Radiofrequency radiation-induced
calcium ion flux enhancement from human and other neuroblastoma cells in
culture," Bioelectromagnetics, vol. 10, pp. 197-202, 1989.
R. Borges, F. Sala, and A. García, "Continuous monitoring of catecholamine
release from perfused cat adrenals," Journal of Neuroscience Methods, vol. 16, pp.
289-300, 1986.
M. G. López, C. Montiel, C. J. Herrero, E. García-Palomero, I. Mayorgas, J. M.
Hernández-Guijo, M. Villarroya, R. Olivares, L. Gandía, J. M. McIntosh, B. M.
Olivera, and A. G. García, "Unmasking the functions of the chromaffin cell α7
nicotinic receptor by using short pulses of acetylcholine and selective blockers,"
Proc. Natl. Acad. Sci. USA, vol. 95, pp. 14184-14189, 1998.
M. Herrera, L.-S. Kao, D. J. Curran, and E. W. Westhead, "Flow-injection
analysis of catecholamine secretion from bovine adrenal medulla cells on
microbeads," Analytical Biochemistry, vol. 144, pp. 218-227, 1985.
D. J. Green and R. L. Perlman, "On-Line Measurement of Catecholamine
Secretion," Analytical Biochemistry, pp. 270-276, 1981.
T. Hagan, I. Chatterjee, D. McPherson, and G. L. Craviso, "A novel waveguidebased radio frequency/microwave exposure system for studying nonthermal
effects on neurotransmitter release-Finite-Difference Time-Domain modeling,"
IEEE Trans. Plasma Science, vol. 32, 2004.
T. A. Litovitz, L. M. Penafiel, J. M. Farrel, D. Krause, R. Meister, and J. M.
Mullins, "Bioeffects Induced by Exposure to Microwaves Are Mitigated by
Superposition of ELF Noise," Bioelectromagnetics, vol. 18, pp. 422-430, 1997.
C.-K. Chou and A. W. Guy, "Effects of Electromagnetic Fields on Isolated Neme
and Muscle Preparations," IEEE Transactions on Microwave Theory and
Techniques, vol. 26, pp. 141-147, 1978.
12
[59]
[60]
[61]
[62]
J. E. H. Tattersall, I. R. Scott, S. J. Wood, J. J. Nettell, M. K. Bevir, Z. Wang, N.
P. Somasiri, and X. Chen, "Effects of Low Intensity Radiofrequency
Electromagnetic Fields on Electrical Activity in Rat Hippocampal Slices," Brane
Research, vol. 904, pp. 43-53, 2001.
D. Platano, P. Mesirca, A. Paffi, M. Pellegrino, M. Liberti, F. Apollonio, F.
Bersani, and G. Aicardi1, "Acute Exposure to Low-Level CW and GSMModulated 900 MHz Radiofrequency Does Not Affect Ba2+ Currents Through
Voltage-Gated Calcium Channels in Rat Cortical Neurons," Bioelectromagnetics,
vol. 28, pp. 599-607, 2007.
K. R. Foster and W. F. Pickard, "Microwaves: The risks of risk research," Nature,
vol. 330, pp. 531-532, 1987.
C. Eulitz, P. Ullsperger, G. Freude, and T. Elbert, "Mobile Phones Modulate
Response Patterns of Human Brain Activity," NeuroReport, vol. 9, pp. 3229-3232,
1998.
Chapter 2
13
Possible Mechanisms for the Interaction between MW Fields
and Biological Systems
2.1 Introduction
The goal of this research is to identify MW exposure parameters that can affect
the release of CA from chromaffin cells. The types of MW exposure parameters that are
available using current equipment in our laboratory are continuous wave (CW),
amplitude modulation (AM), pulse modulation, phase modulation (PM), frequency
modulation (FM), frequency sweep and power sweep. In addition, the signal generator
allows a combination of the modulations listed above. For example, the signal generator
can enable functions for AM and frequency sweep simultaneously to generate a signal
whose amplitude is modulated and whose carrier frequency is swept at the same time.
Thus, there are available an infinite number of combinations of modulations that could be
tested. In order to accomplish the goal of the research within a reasonable time frame, it
was necessary to narrow down the modulation parameters to those that would have the
greatest likelihood of producing a bioeffect. This was accomplished by a thorough
investigation of the literature to identify those MW exposure parameters in the 1 to 6
GHz frequency range that were reported to cause effects on biological systems. Many of
the studies were cited in Chapter one. The following section focuses on possible nonthermal mechanisms of interaction between MW fields and biological systems.
2.2 Exocytosis
14
The release of CA from adrenal chromaffin cells occurs via a process called
exocytosis and it is this process that we envision can be altered by MW fields in a nonthermal manner. Within chromaffin cells, CA are stored in membrane-bound vesicles. In
vivo, the interaction of the neurotransmitter acetylcholine with nicotinic cholinergic
receptors on the plasma membrane results in cell depolarization that in turn results in
calcium entry into the cells via voltage gated calcium channels. The rise in intercellular
calcium leads to exocytosis, which is the fusion of the CA storing vesicles with the
plasma membrane. CA are then released to the outside of the cell. This process is shown
diagrammatically in Figure 2.1.
For experiments, CA release is stimulated by application of the drug 1,1dimethyl-4-phenylpiperazinium (DMPP), which is a stable nicotinic receptor agonist.
Thus we are monitoring a physiologic process that occurs in vivo and looking for effects
of MW fields on this process.
2.3 Possible Non-thermal Mechanisms Underlying MW-Induced
Bioeffects
Non-thermal effects on a biological system are caused by the interactions between
MW fields and cellular constituents. Numerous interactions have been theoretically
explained and experimentally investigated, and possible mechanisms by which the
release of CA from chromaffin cells could be affected are summarized in the following
sections.
15
Figure 2.1. Schematic diagram of the process of exocytosis. After acetylcholine (Ach,
green dots) binds to nicotinic cholinergic receptors ① and the chromaffin cell becomes
depolarized, influx of calcium (blue dot) via voltage gated calcium channels occurs ②
and leads to fusion of the CA containing vesicles with the plasma membrane ③ and
subsequent release of CA (red dots) to the extracellular space ④.
2.3.1 Force Effects on Proteins
MW fields can induce a torque on molecules, which can result in displacement of
ions, vibrations in bound charges, and rotation and reorientation of dipolar molecules
such as water [1]. This type of effect cannot be observed in low-level electromagnetic
fields, due to the random thermal agitation existing in biological systems [2]. Thus, for
signals from external fields to be discernible by cells, the signal strength should be above
this endogenous background level [2-5]. This suggests that there is a minimum intensity
16
of applied field, i.e., a threshold field strength [2], for eliciting effects on molecules, such
as plasma membrane proteins. Thus, the exposure systems should be able to achieve
power levels much higher than the threshold field strength to produce a robust and
repeatable bioeffect [5, 6]. The mechanism also suggests a cutoff frequency (above which
no response is observed) due to viscous drag or other mechanical forces on the particle
[2].
With respect to plasma membrane proteins, the interaction between MW fields
and molecules can affect different functions, such as binding ligands and protein
conformation. Each of these interactions due to force effects are described in the
following three sections.
2.3.1.a Changes in Binding Ligands
It is well known that a ligand binds to a receptor and changes the conformation of
the receptor to control its function [7] and many studies report that the binding of ligands
to receptors may be altered by MW fields [3, 8-11]. An empty crevice of a receptor
attracts a free ligand in its endogenous potential energy well. The probability that a ligand
binds with a receptor is represented as P, called the ion binding probability. Applied MW
fields possibly cause energy losses of the ligand ion due to collisions inside the receptor
crevice and thus can interrupt the attracting endogenous force due to changes in the
potential energy of the ion in the binding site, and hence change the P value [8]. The
results presented in [3] and [8] show that the P value calculated by a quantum biophysical
model can be significantly changed by the presence of a MW field. Moreover, the change
17
in P value is reported to be both frequency dependent and power dependent [2].
Theoretical studies have shown that changes in the binding of ligands can be
induced by high E field exposure (>10 kV/m) that are sufficient to counteract thermal
noise [2, 12]. However, experimental studies have found effects at E fields lower than
guideline values (10 W/kg for occupational or 2 W/kg for public exposure) and recent
research suggests that there are some amplication mechanisms of bioeffects at work in
cellular process [6, 10, 13-15]. This indicates that there may exist a phenomenon called a
“power window”, that is, a certain power density or range of power densities that lead to
prominent bioeffects.
As described in the previous section, the release of CA from chromaffin cells
involves the binding of acetylcholine (ACh) to nicotinic cholinergic receptors. If the P
value for the binding of ACh to its receptor is decreased, i.e., the number of bindings
between ACh molecules and nicotinic cholinergic receptors is reduced due to the
presence of MW fields, the extent to which the cell is depolarized would vary, which
would lead to a change in the amount of released CA.
2.3.1.b Changes in Protein Conformation
Proteins consist of a sequence or chain of amino acids connected by peptide
bonds. The chain often contains loops and or helices due to Van der Waals and other
types of non-covalent attractive forces. The way in which the chain is arranged in space is
called the conformation and the environment where a protein is located is important for
maintaining the physiologically relevant conformation and hence proper functioning of
the protein. When MW fields are applied to proteins, the environment can possibly be
18
changed due to the force between the MW E field and the polar chains of the amino acids
[16, 17]. Small changes in protein conformation can result in significant effects on
biological function.
During the process of exocytosis, influx of calcium occurs via voltage gated
calcium channels. If MW fields change, for example, the conformation of voltage gated
calcium channels, the function of the channels can be altered [17], which may affect the
fusion of the CA containing vesicles with the plasma membrane, and thus change the
amount of released CA.
2.3.1.c Effects Inside of Cells
The plasma membrane is considered as a lossy (nonzero conductivity) dielectric
material [18], and hence can be modeled as shown in Figure 2.1. When a low frequency
E field is applied to the cells, most of the voltage is dropped across the membrane due to
the resistance of the membrane (Figure 2.2). The situation changes when fields at higher
frequencies are applied. The capacitance of the membrane (Figure 2.2) causes the E field
within the membrane to be equal to or possibly less than the average field at
frequencies higher than 1 GHz [7].
19
Figure 2.2. Simplified electrical model of a cell membrane. The plasma membrane can be
expressed as a resistor and a capacitance in parallel.
As the E field passes through a cell, it may induce non-thermal effects, such as
changing gene expression and affecting enzyme activities inside the cell. In chromaffin
cells, the mechanisms responsible for fusion of vesicles with the plasma membrane may
be affected, thus changing the amount of CA released.
2.3.2 Electrical Current Induced in the Extracellular Medium
The previously explained interactions are due to the forces produced by electric
fields that directly affect proteins in biological systems. There can also be other
interactions that occur in an indirect manner. For example, MW field-induced current in
the conductive extracellular medium can affect ions, voltage dependent ion channels or
other charged molecules on the plasma membrane [19, 20]. For ion channels, , such
20
interactions have been reported to include decreased frequency of single channel
openings and increased rates of rapid, burst-like firing [5]. If such effects are exerted on
voltage dependent ion channels on chromaffin cell membranes, influx of calcium can be
altered, hence changing the amount of released CA.
2.4 Conclusion
Based on the previous discussions, there are many possible ways that MW E
fields could interact with chromaffin cells and affect the process of CA release. In order
to be able to understand the non-thermal mechanisms underlying any effects on the
exocytotic process, robust and reproducible bioeffects need to be induced and the specific
MW exposure parameters that elicit the effects carefully delineated. Therefore, carrying
out this investigation requires a well characterized exposure system. The design,
characterization, and optimization of the exposure system are described in Chapters 3, 4
and 5.
21
2.5 References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
H. P. Schwan, "Biological effects of non-ionizing radiations: Cellular properties
and interactions," Annals of Biomedical Engineering, vol. 16, pp. 245-263, 1988.
K. R. Foster, "Thermal and Nonthermal Mechanisms of Interaction of RadioFrequency Energy with Biological Systems," IEEE Transactions on Plasma
Science, vol. 28, pp. 15-23, 2000.
C. J. Thompson, Y. S. Yang, V. Anderson, and A. W. Wood, "A cooperative
model for Ca++ efflux windowing from cell membranes exposed to
electromagnetic radiation," Bioelectromagnetics, vol. 21, pp. 455-464, 2000.
R. Adair, "Constraints on biological effects of weak extremely-low-frequency
electromagnetic fields," Physical Review A, vol. 15, pp. 1039-1048, 1991.
M. H. Repacholi, "Low-Level Exposure to Radiofrequency Electromagnetic
Fields: Health Effects and Research Needs," Bioelectromagnetics, vol. 19, pp. 1-9,
1998.
A. G. Pakhomov, P. Gasek, L. Allen, B. E. Stuck, and M. R. Murphy,
"Comparison of Dose Dependences for Bioeffects of Continuous-Wave and HighPeak Power Microwave Emissions Using Gel-Suspended Cell Cultures,"
Bioelectromagnetics, vol. 23, pp. 158167, 2002.
L. J. Challis, "Mechanisms for Interaction Between RF Fields and Biological
Tissue," Bioelectromagnetics Supplement, pp. 98-106, 2005.
A. Chiabrera, B. Bianco, E. Moggia, and J. J. Kaufman, "Zeeman-Stark Modeling
of the RF EMF InteractionWith Ligand Binding," Bioelectromagnetics, vol. 21,
pp. 312-324, 2000.
S. Lin-Liu and W. R. Adey, "Low frequency amplitude modulated microwave
fields change calcium efflux rates from synaptosomes," Bioelectromagnetics, vol.
3, pp. 309-322, 1982.
C. Blackman, S. Benane, J. Elder, D. House, J. Lampe, and J. Faulk, "Induction of
calcium-ion efflux from brain tissue by radiofrequency radiation: effect of sample
number and modulation frequency on the power-density window,"
Bioelectromagnetics, vol. 1, pp. 35-43, 1980.
K. R. Foster and M. H. Repacholi, "Biological effects of radiofrequency fields:
Does modulation matter?" Radiation Research, vol. 162, pp. 219-225, 2004.
H. P. Schwan, "Interaction of Microwave and Radio Frequency Radiation with
Biological Systems," IEEE Transactions on Microwave Theory and Techniques,
vol. 19, pp. 146-152, 1971.
M. Zhadin and F. Barnes, "Frequency and Amplitude Windows in the Combined
Action of DC and Low Frequency AC Magnetic Fields on IonThermal Motion in
a Macromolecule:Theoretical Analysis," Bioelectromagnetics, vol. 26, pp. 323330, 2005.
W. Adey, "Frequency and power windowing in tissue interactions with weak
electromagnetic fields," Proceedings of the IEEE, vol. 68, pp. 119-125, 1980.
B. Bianco, A. Chiabrera, E. Moggia, and T. Tommasi, "Enhancement of the
interaction between low-intensity R.F. e.m. fields and ligand binding due to cell
basal metabolism," Wireless Networks, vol. 3, pp. 477-487, 1997.
22
[16]
[17]
[18]
[19]
[20]
J. A. Laurence, P. W. French, R. A. Lindner, and D. R. Mckenzie, "Biological
Effects of Electromagnetic Fields-Mechanisms for the E4ects of Pulsed
Microwave Radiation on Protein Conformation," Journal of Theoretical Biology,
vol. 206, 2000.
M. H. Repacholi and B. Greenebaum, "Interaction of Static and Extremely Low
Frequency Electric and Magnetic Fields with Living Systems: Health Effects and
Research Needs," Bioelectromagnetics, vol. 20, pp. 133-160, 1999.
M. D. Kotnik T, "Second-order model of membrane electric field induced by
alternating external electric fields.," IEEE Transactions on Biomedical
Engineering, vol. 47, pp. 1074-1081, 2000.
W. F. Pickard and E. G. Moros, "Energy Deposition Processes in Biological
Tissue:Nonthermal Biohazards Seem Unlikely inthe Ultra-High
FrequencyRange," Bioelectromagnetics, vol. 22, pp. 97-105, 2001.
F. Barnes and Y. Kwon, "A Theoretical Study of the Effects of RF Fields in
theVicinity of Membranes," Bioelectromagnetics, vol. 26, pp. 118-124, 2005.
23
Brief Introduction of Chapter 3
This research focuses on searching for potently useful bioeffects of MW fields on a
well-established model of neural-type cells, isolated bovine adrenal chromaffin cells in
the frequency range 1 to 6 GHz. In order to investigate a diverse range of MW exposure
parameters that have the possibility of eliciting non-thermal bioeffects on chromaffin
cells, a well characterized exposure system is essential for obtaining reliable data since
many positive experimental results reported by previous researchers have been criticized
for poorly designed and characterized exposure systems leading to unreliable results.
Hence, the exposure system for this research was carefully designed, characterized and
optimized using the finite-difference time-domain (FDTD) technique and a finite element
based flow and temperature based simulation. Chapter 3, which is a reproduction of a
paper that appeared in the IEEE Transactions on Plasma Science, describes part of this
effort, and the following Chapters 4 and 5 explain in more detail the optimization process
for the electrical and biological aspects of the exposure system.
24
Chapter 3
Design, Characterization and Optimization of a Broadband
Mini Exposure Chamber for Studying On Line Monitoring of
Catecholamine Release from Chromaffin Cells Exposed to
Microwave Radiation: Finite-Difference Time-Domain
Technique
The complete paper is reprinted from: Jihwan Yoon, Indira Chatterjee, Dana McPherson, and
Gale L. Craviso, “Design, Characterization and Optimization of a Broadband Mini
Exposure Chamber for Studying On Line Monitoring of Catecholamine Release from
Chromaffin Cells Exposed to Microwave Radiation: Finite-Difference Time-Domain
Technique”, IEEE Transactions on Plasma Science, volume 34, Issue 4, August 2006,
page 1455-1469. © 2006 IEEE
25
Abstract
A free space in vitro exposure system for identifying specific microwave
parameters in the frequency range 1 to 6 GHz that can induce nonthermal effects on
exocytosis, the process by which neurotransmitter release occurs, has been designed,
constructed, characterized and optimized. The exposure system is placed within an
anechoic chamber and incorporates continuous on-line monitoring of basal and
stimulated catecholamine release from cultured bovine adrenal medullary chromaffin
cells, a well-established model of neural-type cells. The cells are immobilized inside a
cell perfusion apparatus and continuously superfused with temperature controlled
balanced salt solution, with the entire cell perfusion apparatus placed within a mini
exposure chamber constructed out of a microwave absorbing material. All relevant
equipment for carrying out experiments is shielded from the microwave field by being
housed in an aluminum conductor box located behind the mini exposure chamber.
Detailed distributions of the electric field and specific absorption rate at the location of
the cells within the mini exposure chamber were computed using the Finite-Difference
Time-Domain method. Finite-Difference Time-Domain computations were also used for
optimizing the exposure system so that the highest intensity of electric field could be
delivered under dynamic temperature control and with an acceptable degree of field
homogeneity (to within 30%) over the entire frequency range of 1-6 GHz. A major
finding is that maintaining an acceptable level of homogeneity of the electric field and
specific absorption rate for exposing cells to 1- 6 GHz microwave fields requires a
different distribution of the cells within the cell perfusion apparatus for exposures carried
out at the lower versus the higher end of the frequency range of interest.
26
3.1 Introduction
Studies carried out both in vitro and in vivo show that microwave (MW) fields can
produce effects on the nervous system [1]. In studies where effects due to heating appear
to have been ruled out, several mechanisms were proposed as the basis for the
nonthermal bioresponses [2, 3]. One hypothesis is that MW fields interact with specific
cellular constituents, such as neurotransmitter receptors and membrane ion channels [4-6],
and that important determinants for eliciting the effects pertain to the parameters of the
MW fields that are applied, such as frequency, power, modulation schemes, and whether
the fields are pulsed or continuous wave (CW) [7].
Because the synaptic release of neurotransmitters is a primary mechanism for
intercellular communication in the nervous system, of particular interest to us is the
question of whether MW fields can induce nonthermal effects on neurotransmitter release.
If specific MW parameters can be identified that elicit selective, reproducible and
interpretable effects on this important physiological process, this could set the stage for
future studies directed at developing the use of MW fields to alter nervous system
function in a controlled, non-deleterious manner, which would have clinical applications.
To explore the potential use of MW fields for altering neurotransmitter release, we are
undertaking an in vitro study that will use a well-established model of neural-type cells,
isolated bovine adrenal chromaffin cells that synthesize, store and release catecholamines,
and an exposure system that will allow identification of MW parameters over the 1 to 6
GHz frequency range that produce nonthermal effects on basal and stimulated
catecholamine release. The basic experimental approach is similar to that previously
described for identifying nonthermal effects on catecholamine release from chromaffin
27
cells exposed to radiofrequency/MW fields in the frequency range 0.75 to 1.12 GHz,
which uses a waveguide-based exposure system. Essentially, chromaffin cells are
immobilized within a cell perfusion apparatus (CPA), which provides continuous
superfusion of the cells with a balanced salt solution (BSS) and continuous monitoring of
catecholamine release by electrochemical detection, using an in-line electrochemical
detector (ECD) during exposures [8]. Because the physical size of a standard waveguide
for the higher frequency range of 1 to 6 GHz is too small to accommodate the CPA, a
different exposure system was required. The approach taken was to conduct the
exposures in free space inside an anechoic chamber.
This paper describes the design, construction, characterization and optimization of a
free-space exposure system that similarly incorporates a CPA for on-line monitoring of
catecholamine release during exposures to MW fields over the 1 to 6 GHz range.
Important features of the exposure system are the placement of the CPA within a mini
exposure chamber (MEC) to reduce reflections from the horn antenna signal source and
the shielding of all relevant equipment such as the ECD by a conductor box located
behind the MEC [9]. Numerical Finite-Difference Time-Domain (FDTD) modeling, a
tool for computing electromagnetic fields in biological systems [10 - 13], was used to
characterize and optimize the design of the exposure system so that it provides the
ur
maximum possible nonthermal level of electric ( E ) field as well as acceptable
ur
homogeneity (to within 30%; [14, 15]) of the E field and specific absorption rate
(SAR) in the region where the cells are immobilized within the CPA. To achieve these
objectives, the FDTD model took into account the geometries and the dielectric
properties over the 1 to 6 GHz frequency range of all components of the MW exposure
28
system. Based on the results of the numerical computations, the exposure system
designed should enable us to perform experiments under well-defined and controlled
conditions and thus allow us to obtain interpretable and reproducible results.
3.2 Methodology
3.2.1 Cell Perfusion
Cultured bovine adrenal medullary chromaffin cells are prepared and cultured
according to the method described previously [8]. For experiments, the cells are
immobilized inside a plastic filter holder (the CPA; Figure 3.1.a) to allow for continuous
superfusion of the cells with a BSS. All components of the CPA have been previously
described [8]. Briefly, inside the CPA are three nylon filters sandwiching a single glass
fiber filter (GFF) on which the chromaffin cells are immobilized (Figure 3.1.b). Dead
space within the filter holder is minimized with dental putty. The CPA is entirely
composed of non-metallic, low dielectric permittivity materials in order to cause minimal
distortion of the MW field in which it is placed [8].
The temperature of the BSS entering and exiting the CPA is continuously monitored
by fluoroptic temperature probes (Luxtron Model 790 Fluoroptic thermometer with
model SFW-5 non-perturbing temperature probes) inserted into the BSS inlet and outlet
tubing of the CPA. The temperature of the BSS entering the CPA is maintained at 36.5 ±
0.2°C by means of a glass inlet tube to the CPA around which is wound a 1 foot long
piece of nichrome wire of radius 0.5 mm that serves as the heat source.
29
Inlet Glass Tubing with Nichrome Wire
CPA
Upper
Filter
Holder
Dental
Putty
Gasket
Glass
Fiber
Filter
Temperature
Probes
350 µm
Nylon
Mesh
5 µm
Nylon
Mesh
Lower
Filter
Holder
(a)
(b)
Figure 3.1. Photograph (a) of the CPA showing the location of the glass inlet tubing
wound with nichrome wire, and where temperature probes are placed and (b) SolidWorks
model of the CPA, showing an exploded view of the filter holder and its components.
Current through the nichrome wire is supplied by a constant voltage source (Astron RS35A) that works in conjunction with a custom-built current controller utilizing a series of
FETs (field effects transistors). The temperature of the BSS in the inlet tubing measured
by the Luxtron thermometer is used in a feedback loop to the computer. The set point
temperature of 36.5 oC is compared to the inlet temperature and is used to provide the
control to the current controller. Because our system incorporates this dynamic
30
temperature control, the temperature of the BSS entering the CPA is maintained at 36.5 ±
0.2 oC both in the absence and presence of MW fields.
3.2.2 Instrument Shielding
The experimental setup requires that several different instruments (e.g., ECD,
thermometer) be placed in close proximity to the CPA. Because the presence of metal in
the instruments could cause significant perturbation in the MW fields as well as possible
damage to the instruments themselves due to the MW fields, a rectangular aluminum
enclosure (RAE) of dimensions 1 x 1 x 0.5 m was constructed to house all relevant
instruments and shield them from the MW fields. The RAE was placed behind a MEC,
which is described in the next section.
3.2.3 MEC
Since the presence of the RAE in the anechoic chamber would cause large reflections
ur
of the MW field, a MEC was designed to house the CPA and minimize E field
reflections. The MEC, which was constructed out of MDF (medium density fiberboard;
Sierrapine, Roseville, CA), consisted of a back wall (dimensions 1 x 1 m) and four side
walls (dimensions 1 x 0.34 m) each making an angle of 45o with the back wall (Figure
3.2). The angle of 45o was chosen based on measured oblique reflectivities of a highperformance vented microwave pyramidal absorber material (AEP-8-V, Advanced
ElectroMagnetics Inc., Santee, CA) that was glued to the inside of the five walls of the
MEC [16]. These measurements indicated that the reflectivity was less than or equal to
31
14 dB for angles of incidence between 40º and 50º in the frequency range 1.5 – 6 GHz
(Table 3.1).
Table 3.1. Measured reflectivities of the absorber material as a function of frequency and
angle of incidence.
Frequency
(GHz)
Incident Angle (degrees)
20
30
40
50
60
70
-14
-13
-10
1.5
-12
-15
dB
-16
2.0
-7
-16
-15
-16
-12
-13
2.5
-7
-11
-14
-17
-17
-17
3.0
-11
-10
-20
-20
-22
-22
3.5
-8
-7
-18
-31
-29
-33
4.0
-8
-15
-26
-32
-34
-27
4.5
-15
-14
-21
-33
-37
-31
5.0
-18
-23
-44
-29
-21
-25
5.5
-18
-18
-30
-30
-31
-40
6.0
-18
-21
-35
-29
-25
-44
3.2.4 MW Signal Generation
The MW exposure system utilizes a high-power broadband horn antenna (Model AH-81,
Instruments for Industry, Ronkonkoma, NY) to provide pulsed or CW fields. The horn
antenna is mounted inside the anechoic chamber on a computer-controlled tower and is
energized by a signal generator (Model 8665B, Agilent Technologies, Palo Alto, CA) and
a pulsed/CW high-power broadband amplifier (Model TD 81-250, Instruments for
Industry), which are located in the adjacent screen room.
32
Pyramidal Absorber
Material
MDF
Stand
CPA
Figure 3.2. Photograph showing a close-up view of the MEC.
The power supplied by the amplifier is monitored by a power meter and sensor (Models
E4416A and E9325A respectively, Agilent Technologies) connected via a directional
coupler (Model 069Y7-73047-701 for 1 to 4 GHz and Model 069Y7-73047-702 for 4 to
6 GHz, Mega Industries, Gorham, ME). The MEC is placed in the far field of the horn
antenna (1.5 m at 1 GHz), which is calculated based on the measured half power beam
ur
widths in the H and E planes [17]. During experiments, the E field is monitored
33
continuously using one of three identical broadband (frequency range 0.8 - 7.5 GHz)
spiral antennas designed and fabricated in our laboratory (Figure 3.3) using FR4 substrate
(Kepro, Fenton, MO) and a printed circuit board plotter (Model 91S/VS, LPKF, Garbsen,
Germany). The gains of the spiral antennas were measured using the well known three
antenna method [18 - 20]. A possible error associated with this method is polarization
mismatch, an error which can be as large as the axial ratio of the spiral antenna. We have
corrected for polarization mismatch by rotating the spiral antennas with respect to each
other until a peak response was obtained. Since the outer dimension of the spiral antenna
is greater than the diameter of the GFF, it cannot be placed at the location of the GFF or
in its vicinity inside the MEC during experiments. Hence, it is placed at a fixed location
just outside the MEC as shown in Figures 3.4.a and 3.4.b that represent, respectively, a
schematic of the overall MW exposure system and a photograph of the exposure system
inside the anechoic chamber. Due to the unavailability in our laboratory of a non-
ur
ur
perturbing miniature E field probe, it was not possible to measure the detailed E field
distribution in the region where the chromaffin cells will be located in experiments.
ur
Instead, an average value of the E field in the region containing the GFF was obtained
ur
by performing a calibration that related the measured E field at the location of the
ur
spiral antenna outside the MEC (Figures 3.4.a and b), the average E field in the region
containing the GFF (also measured using the spiral antenna) and the input power to the
horn antenna.
34
(a)
(b)
Figure 3.3. Photograph (a) and SolidWorks diagram (b) of the spiral antenna (all
dimensions are in mm).
35
During experiments, this calibration serves also to continuously monitor and adjust the
ur
average E field at the location of the GFF via a feedback program written in LabVIEW
(National Instruments version 7.0), where the signal received by the spiral antenna is
measured by a spectrum analyzer (model 70205A, Hewlett Packard) located in the screen
room. All aspects of the exposure parameters are computer-controlled using locally
written LabVIEW programs.
(a)
36
Location of the
Spiral Antenna
MEC
Horn Antenna
CPA
RAE
Wood Stand
Tower
(b)
Figure 3.4. Schematic (a) of the entire free space MW exposure system showing the
experimental arrangement inside the anechoic chamber and adjacent screen room; (b)
Photograph of the exposure system inside the anechoic chamber.
3.3 FDTD Modeling
3.3.1. Dielectric properties used in the FDTD modeling
The dielectric properties of all the materials comprising the CPA were the same as
those used previously [8]. The complex dielectric permittivity of the GFF saturated with
BSS was measured in our laboratory using an open-ended dielectric probe kit (Model
85070A, Hewlett Packard, Palo Alto, CA) and twenty stacked GFFs [8]. As shown in
Figure 3.5, there was a three-fold increase in conductivity from 1 to 6 GHz. This
37
variation in dielectric properties with frequency was directly incorporated into the FDTD
model. The horn antenna and aluminum backing of the MEC were modeled as perfect
conductors and the dielectric properties of the absorber material within the MEC were
measured over the frequency range 0.13 – 20 GHz using the open-ended dielectric probe
kit. The measured complex permittivity obtained by averaging measurements [21] at 20
different locations on a sample of the absorber material was then used to obtain the
parameters for the Debye equation that was used in the FDTD method for calculating the
dielectric properties of the absorber material [12].
60
10
8
40
6
30
4
20
2
10
0
1
2
3
4
5
6
Frequency (GHz)
Figure 3.5. Measured complex relative permittivity (dielectric constant ε΄, dielectric loss
ε˝) and conductivity of GFF soaked with BSS over the frequency range 1 to 6 GHz.
Conductivity (S/m)
Relative Permittivity
50
38
3.3.2. Horn Antenna
To characterize the gain and far-field radiation pattern of the broadband horn antenna
(Figure 3.6.a), a detailed 3D model of the antenna (dimensions 21 x 26.5 x 15.9 cm) was
constructed (Figure 3.6.b) using the commercially available CAD package SolidWorks
2005 (Solid Works, Concord, MA). This model was imported into a commercially
available FDTD software package XFDTD (BioPro version 6.2, Remcom Corporation,
State College, PA), in which the horn antenna geometry was broken up into a fine mesh
consisting of cubical Yee cells [12] having a side length of 0.5 mm. Because the horn
antenna determines the MW field incident on the CPA and hence the cells, the geometry
of the antenna was rendered in the XFDTD model. The horn antenna as well as its
coaxial feed were modeled as perfect conductors. A broadband excitation source was
selected to excite the antenna using a 50 Ω coaxial port as suggested in XFDTD
(Reference Manual). The source was a modulated Gaussian pulse centered at 3.5 GHz,
having a pulse width of 946.903 ps. The default time step was set to 0.9623 ps, and the
simulation ran for 5000 time steps (2.7797 ns).
Coax Connector
Flared Structure
Antenna
Holder
(a)
(b)
Figure 3.6. Photograph (a) and SolidWorks model (b) of the broadband horn antenna.
39
3.3.3 Exposure System
A detailed geometric model (Figure 3.7) of the entire MW exposure system (horn
antenna, MEC and the CPA; overall dimensions 114.4 x 113.1 x 218.4 cm) was
constructed using SolidWorks. The model also included a sheet of aluminum to represent
the side of the RAE in direct contact with the back wall of the MEC. The CPA (Figures
3.8.a and b) is located 5 cm from the tip of the absorber material directly behind it, and
the horn antenna is placed at a distance of 1.5 m from the CPA in order to remain in the
far-field over the entire frequency range 1 – 6 GHz.
The model was imported into XFDTD and broken up into a mesh comprised of a
main grid of base cubical Yee cells having a side length of 3.5 mm, which was the
smallest Yee cell size that could accommodate the dimensions of the model with the
ur
available computer resources. To achieve better accuracy in computing the E field and
the SAR distribution at the location of the cells on the GFF, a more accurate XFDTD
model of the GFF was created. The GFF is circular in geometry with a radius of 12 mm
(the radius actually available for cell distribution during experiments is 11 mm because of
the gasket) and a thickness of 0.26 mm and consists of a very fine glass mesh. For the
XFDTD model, the GFF was approximated as a solid dielectric material broken up into a
mesh comprised of cubical Yee cells of dimension 0.2 mm using the adaptive meshing
capability of XFDTD.
40
(a)
Conducting
Plane
MEC
CPA
Horn
Antenna
MDF
Frame
(b)
Figure 3.7. XFDTD model (numerical mesh) of the entire exposure system. (a) 3D
perspective view and (b) cross sectional view through the central horizontal plane.
41
The much smaller XFDTD mesh better represents the glass mesh size of the actual GFF
(Figure 3.8.c). The solid dielectric material representing the GFF had the dielectric
properties of GFF saturated with BSS (shown in Figure 3.5). In addition, there was a 0.4
mm thick BSS layer directly above the GFF to represent as closely as possible the actual
situation in the CPA where the space between the two sections of the filter holder created
by the silicone gaskets creates a 0.4 mm gap so that the chromaffin cells do not get
crushed. Figure 3.8.d is a close up view of the center cross section of the XFDTD mesh
that shows this BSS layer. A Liao absorbing boundary [12] surrounds the entire solution
space of the model with a thickness of eight base cells.
As in the actual exposure system, the launch in the FDTD model was excited by a
discrete sinusoidal 50-Ω MW source that is connected to the N-type coaxial connector of
the horn antenna. The CW input power to the horn antenna was set to 250 W which is the
maximum power that the power amplifier can deliver. The simulation at different
frequencies in the range 1 – 6 GHz each ran for 30,000 time steps and the detailed spatial
ur
distributions of E field and SAR in the region containing the chromaffin cells on the
GFF computed.
42
(a)
(b)
GFF
Filter Holder
Base Mesh
Silicone
Rubber
Gasket
Adaptive Mesh
(c)
43
(d)
Figure 3.8. SolidWorks model (a) and (b) XFDTD model (3D perspective view) of the
CPA; (c) XFDTD cross sectional view of the gasket and the GFF that is composed of 3.5
mm base cell size and 200 µm adaptive mesh cell size. The GFF has an overall radius
of 12 mm, with an inner radius of 11 mm available to the cells due to the 2 mm wide
gasket. (d) Close up view of the center cross section of the XFDTD mesh showing the
BSS layer.
3.4 Results and Discussion
3.4.1 Antenna Gain and Far-Field Radiation Pattern
In order to validate the accuracy of the XFDTD horn antenna model, the computed
gain and far-field radiation pattern of the antenna in the frequency range 1 – 6 GHz were
compared with those measured in the anechoic chamber using the three antenna method
[18 - 20]. The other two antennas used were two of the identical spiral antennas whose
gains had been previously measured (section II.D). The gains of the two identical spiral
antennas measured using the three antenna method with the horn antenna as the third
antenna were then compared with the gains obtained using the three identical spiral
antennas. In the three antenna method, if the gain of one of the antennas is in error, the
44
gains of the other two antennas used will also be incorrect. By comparing the spiral
antenna gains with those previously measured, it was ensured that the gain of the horn is
accurate. We found that the gains of the spiral antennas measured with the horn antenna
as the third antenna were very close to the gains measured using the three spiral antennas.
In addition we corrected for polarization mismatch, and applied a gain correction as
suggested in the literature [19]. As shown in Figure 3.9.a for the frequency range 1 – 6
GHz, there was a maximum difference of 2.45 dB between the XFDTD computed results
(mean = 10.693 ± 2.202 dB) and the measured results (mean = 10.436 ± 2.498 dB) and a
statistical analysis using a Student’s paired t test showed further that the differences were
not statistically significant (P = 0.065; here and elsewhere, probability of difference
values are considered statistically significant at the P ≤ 0.05 level). When the computed
and measured far field radiation patterns in the H and E planes at 3.5 GHz (center of the
band of interest) were compared, the agreement overall was very good (Figures 3.9.b and
c), except at angles between 30o and 60o and between 300o and 330o. The larger
difference between the computed and measured results at these angles is mainly due to
the maximum usable distance inside the anechoic chamber as well as due to the
difference in space surrounding the horn antenna in the simulation (Liao absorbing
boundary) versus the finite anechoic chamber volume [22, 23]. Since the antenna gain
and far-field radiation pattern are strongly dependent on the geometry of an antenna, the
close agreement that was obtained between the measured and XFDTD results ensures that
the geometry of the antenna used in the XFDTD model is accurate. This in turn
contributes to the accuracy of the results obtained with the complete model of the MW
exposure system.
45
16
14
Gain (dB)
12
10
8
6
4
Calculated gain
Measured gain (smoothed)
2
0
1
2
3
4
Frequency (GHz)
(a)
(b)
5
6
46
(c)
Figure 3.9. Comparison (a) of XFDTD computed and measured gains of the broadband
horn antenna; (b) Comparison of the XFDTD computed and measured H plane radiation
pattern and (c) Comparison of the XFDTD computed and measured E plane radiation
pattern.
3.4.2 Validation of the Effectiveness of the MEC
For monitoring catecholamine release on-line during continuous cell superfusion, the
effluent leaving the CPA goes directly to an ECD used in the amperometric mode. To
obtain well-defined amperometric spikes that reliably reflect when catecholamine release
is stimulated, an important design consideration is that the tubing that connects the CPA
to the ECD be of the shortest length possible. Consequently, as stated previously, the
ECD has to be located inside the anechoic chamber rather than in the adjacent screen
47
room where all instrumentation is typically placed to avoid being exposed to the MW
fields. In addition to the ECD, the equipment required to superfuse the cells with
temperature-controlled BSS, which include a peristaltic pump and the thermometer for
monitoring temperature of the perfused BSS, also must be located in the immediate
vicinity of the CPA and thus inside the anechoic chamber. To address this issue, all
instrumentation is placed inside a RAE located close to the CPA. The effectiveness of the
RAE in ensuring that the output of the ECD in particular was not affected by the MW
field in the frequency range 1 to 6 GHz was confirmed by showing that the response of
the ECD to catecholamine standards was the same in the presence and absence of the
maximum field intensity that would be used in experiments.
The CPA is placed inside a MEC, the design of which took into account the need to
ur
minimize the E field reflected from the RAE and to provide an acceptable level of
ur
homogeneity of the E field within the MEC where the CPA is located. To confirm that
ur
the MEC met these objectives, the detailed E field distribution within the MEC was
ur
computed using XFDTD. Figure 3.10.a shows that the E field within the absorber on
the back wall of the MEC, which is in the region immediately adjacent to the wall of the
RAE, is rapidly attenuated and is negligible (46 dB below the value at the input to the
horn antenna). Hence, reflection from the RAE, which would be the main source of
ur
reflection, is negligible. Also, the E field within the MEC (mean = 295.18 ± 12.32
V/m) is relatively uniform (Figure 3.10.b and c) and has an acceptable homogeneity (to
within 17.7 %), suggesting that reflections from the RAE as well as any other reflections,
such as scattering from the absorber, are negligible.
48
ur
When the E field within the MEC in the plane where the GFF would be located
(rectangular region in Figure 3.10.a) is calculated as a function of distance from the horn
antenna using the effective isotropic radiated power (EIRP) [24] and the corresponding
ur
E field values computed using XFDTD with an input power of 250W to the horn
antenna (Figure 3.10.d), the difference between the two values was to within 6.2 % and
ur
was not statistically significant (P = 0.40; mean = 312.8 ± 7.94 V/m for XFDTD E field
ur
and mean = 313.72 ± 11.82 V/m for EIRP E field). This is further evidence that the
MEC successfully met the design requirements discussed above (minimizing reflections
to gain maximum homogeneity), and also proves that the XFDTD calculation is reliable.
An additional way that was used to validate that the MEC met the design requirements
ur
was to measure the E field within the MEC in the region where the GFF and hence the
chromaffin cells would be located in actual experiments (rectangular region in Figure
ur
3.9.a). The E field was measured using the broadband spiral antenna described
previously, with the spiral antenna accurately positioned at varying distances from the tip
of the back absorber within the MEC by means of a computer-controlled electromechanical actuator built in our laboratory. A locally written LabVIEW program allowed
ur
the spiral antenna to be moved in steps of 1 mm and the E field intensity up to 300 mm
away from the absorber tips in the frequency range 1 – 6 GHz in steps of 0.1 GHz (total
ur
of 50 frequencies) to be automatically logged. The measured E field values for a
10 dBm input power supplied by the signal generator to the spiral antenna showed that
ur
the E field within the MEC in the region where the GFF will be placed in experiments
49
did not vary appreciably with distance from the back absorber over the frequency range
1 to 6 GHz.
Approximate Location of GFF
Y
X
(a)
(b)
50
(c)
Approximate GFF Location
360
Electric Field Intensity (V/m)
320
280
240
200
160
120
1.60
1.65
1.70
1.75
1.80
Distance From the Antenna (m)
(d)
ur
Figure10. E field in the central horizontal plane that contains the GFF. (a) XFDTD
contour plot; (b) XFDTD surface plot of the region shown in (a) byurthe rectangular box.
(c) Histogram and descriptive statistics of the XFDTD computed E field distribution
for the region in (9) within the MEC in the plane containing the GFF. (d) Comparison of
XFDTD (solid line) and EIRP results (dashed line). Input power to the horn antenna =
250 W.
51
ur
At 3.3 GHz where the measured E field showed the most variability, the difference
ur
between the E field calculated using the EIRP and that computed using XFDTD
(Figure 3.11) was not statistically significant (P = 0.574; mean = 1.258 ± 0.102 V/m for
the measured field and mean = 1.261 ± 0.096 V/m for the EIRP field). Thus, both the
ur
FDTD computations and the measured results indicate that the E field has an
acceptable level of homogeneity in the region where the GFF containing the chromaffin
cells will be located in actual experiments within the MEC.
2.0
Electric Field (V/m)
1.5
1.0
0.5
0.0
1.60
1.65
1.70
1.75
1.80
Distance From the Antenna (m)
Figure 3.11. Comparison of measured (solid
line) and EIRP results (dashed line) at 3.3
ur
GHz where the inhomogeneity in the E field distribution was the greatest. Input
power to the horn antenna = 10 dBm.
52
Both the measured and computed results were also used to pinpoint the location
within the MEC (that is, within the rectangular region in Figure 3.10.a) where the
ur
homogeneity of the E field is the greatest and thus where the GFF loaded with
chromaffin cells should be positioned. Over the frequency range of 1 – 6 GHz, the
ur
homogeneity of the E field was found to be greatest at a location 52 mm from the tips
of the back absorber, 150 mm from the absorber tips of the side walls and 150 mm from
ur
the tips of the top and bottom absorber. In addition, the E field measured in the region
that will contain the 12 mm radius GFF (52 – 76 mm from the back absorber tips, as
shown in Figure12), was homogeneous to within 1.9 % (mean = 1.077 ± 0.006 V/m with
10 dBm input power to the horn antenna), which agrees well with the FDTD computed
results that indicate a homogeneity to within 2.2 % (mean = 313.3 ± 2.72 V/m with
250 W input power to the horn antenna).
1.14
Electric Field (V/m)
1.12
1.10
1.08
1.06
1.04
1.02
1.00
52
54
56
58
60
62
64
66
68
70
72
74
76
Distance from Absorber Tip (mm)
ur
Figure 3.12. Measured E field in the region containing the GFF (shown by the shaded
region in Figure 3.9d). Frequency = 3.5GHz and input power to the horn antenna = 10
dBm.
53
3.4.3 E Field and SAR Distribution across the GFF
ur
When the incident E field from the horn antenna is parallel to the plane of the GFF
(parallel orientation), the coupling, which is determined mainly by the tangential
ur
boundary condition, is most efficient, causing the E field in the region of the GFF to be
ur
higher than if the incident E field plane is perpendicular to the plane of the GFF
(perpendicular orientation). The parallel orientation was therefore modeled, and the
ur
distributions of the E field and SAR over the GFF for the lowest frequency (1 GHz),
center frequency (3.5 GHz) and highest frequency (6 GHz) in the frequency range of
interest with an input power 250 Watt, were calculated. At 1 GHz, the distributions of
ur
both E field (Figure 3.13.a) and SAR (Figure 3.14.a) were somewhat asymmetrical.
ur
The E field value at the edge of the GFF closest to the horn antenna was the highest
and decreased gradually towards the edge of the GFF farthest from the antenna. This
ur
variation in E field is expected due to attenuation caused by the GFF that is saturated
ur
with BSS which is a lossy material. At 3.5 GHz, the E field (Figure 3.13.c) and SAR
distribution (Figure 3.14.c) results showed that due to the rapid attenuation, there was a
region of almost zero SAR, or in other words a “cold spot”, on the side of the GFF
ur
farthest from the horn antenna. A comparison of the E field distributions at 3.5 GHz
(Figure 3.13.c) and 6 GHz (Figure 3.13.e) showed that as frequency continued to increase,
ur
the region of high E field shifted from the edge of the GFF closest to the horn antenna
towards the center of the GFF, with more “cold” regions formed in the GFF. The SAR
ur
distribution (Figures 3.14.c and e) showed a similar pattern as for the E field at each
54
ur
frequency. More detailed information regarding the E field and the SAR distribution at
each frequency are given in Figure 3.13.b,d and f and Figure 3.14.b,d and f, respectively.
ur
When the percentage area of the GFF that is homogeneous to within 30% for the E
field and SAR at each frequency was examined (Table 3.2), it can be seen that there was
only a slight decrease in homogeneity as frequency increased from 1 to 3.5 GHz.
However, at 6 GHz, the percentage area of the GFF that was homogeneous to within
ur
30 % for the E field and SAR decreased markedly. This dramatic decrease in
homogeneity was first evident between 4 and 5 GHz (Table 3.2). Values of the total
absorbed power in the GFF are also shown and indicate that the highest absorbed power
is around 3.5 GHz.
ur
Table 3.2. Homogeneity and mean values of E field and SAR, as well as total absorbed
power in the GFF computed using XFDTD at various frequencies in the range 1 - 6 GHz.
ur
SAR
E Field
Total Absorbed Power
Frequency
(mW)
(GHz)
A
(%)
Mean
(V/m)
S.D.
(%)
S.D.
(%)
A
(%)
Mean
(W/Kg)
S.D.
(%)
S.D.
(%)
1
78
10.3
±2.0
±19.4
47
0.05
±0.02
±34.1
0.00631
3.5
73
300.1
±61.2
±20.4
38
67.25
±23.9
±35.5
9.478
4
67.1
234.6
±41.8
±17.8
36.8
46.72
±15.8
±34.2
6.45
5
40.7
120.3
±30.5
±25.3
21
17.05
±8.44
±50.4
2.378
6
38
110.3
±41.2
±37.4
18.7
18.35
±12.2
±66.5
2.518
A = area (%) of the GFF homogeneous to within 30%.
XFDTD calculations also performed for the perpendicular orientation of the incident
ur
E field indicated that this orientation greatly improves the homogeneity in the region
55
ur
1 GHz (a), 3.5 GHz (c) and 6
Figure 3.13. XFDTD computed E field contour plot for
ur
GHz (e) and histogram and descriptive statistics of the E field distribution for 1 GHz
(b), 3.5 GHz (d) and 6 GHz (f). All results are normalized and are for the plane of the
GFF. Input power to the horn antenna = 250 W.
56
Figure 3.14. XFDTD computed SAR contour plot for 1 GHz (a), 3.5 GHz (c) and 6 GHz
(e) and histogram and descriptive statistics of the SAR distribution for 1 GHz (b), 3.5
GHz (d) and 6 GHz (f). All results are normalized and are for the plane of the GFF. Input
power to the horn antenna = 250 W. Total absorbed power at 1, 3.5 and 6 GHz is 0.00631,
9.478 and 2.518 mW, respectively.
57
containing the cells even at the higher frequencies in the band of interest. However,
ur
with the same input power to the horn antenna, the E field and SAR values were less
than for the parallel orientation by about two orders of magnitude, due to the
dominance of the perpendicular boundary condition. Thus, this orientation was not
ur
considered suitable for experiments where the goal is to maximize the E field in the
region of the GFF, at the same time maintaining nonthermal conditions.
3.4.4 Improving SAR Homogeneity for the Entire 1-6 GHz Frequency
Range.
Because our experimental approach will monitor catecholamine release from the
entire population of chromaffin cells present on the GFF, reliable interpretation of results
ur
as well as reproducibility of MW effects requires that the E field and SAR distribution
over the GFF be as homogenous as possible. Therefore, improving the homogeneity of
ur
the E field and SAR distribution on the GFF at the higher frequencies in the 1 to 6 GHz
range means determining possible causes for the inhomogeneities, and two were
identified. One was that the diameter of the GFF becomes comparable to the wavelength
in the lossy BSS soaked GFF or small enough to result in the formation of a standing
wave due to reflections from the edge of the GFF. The other was that the increase in
conductivity of the BSS as frequency increases (Figure 3.5) could cause inhomogeneities.
In order to determine the relative contribution of frequency and GFF diameter and
frequency to reducing the homogeneity of the SAR distribution, FDTD simulations were
performed that modeled three frequencies (1, 3.5 and 6 GHz) and three radii of the GFF
58
that could be available to the cells for immobilization. The conditions for the latter were:
1) our basic CPA setup (depicted in Figure 3.8.c) where the GFF has a region of radius
11 mm available to the cells by using a 2 mm wide gasket; 2) a GFF that has only a
region of radius 7 mm available to the cells by using a wider (6 mm) gasket, and 3) a
smaller CPA that accommodates a GFF filter with a region of radius 5 mm available to
the cells. To focus more directly on the effect of the filter radius and frequency on
homogeneity, the FDTD calculations were performed in the absence of the MEC and the
antenna was replaced with a plane wave source. The results of a two-way analysis of
variance (ANOVA) showed that both the radius of the GFF available for cell distribution
and frequency have a significant effect on the SAR homogeneity (P = 0.041 and P =
0.015 for filter radius and frequency, respectively). As shown in Table 3.3, homogeneity
(expressed in terms of the percentage of area that was homogeneous to within 30%)
improved as the radius of the GFF decreased, either by increasing the size of the gasket,
as shown in Figure 3.15, or using a smaller CPA that in turn accommodates a smaller
GFF filter. The improvement in SAR homogeneity was observed at each frequency. On
the other hand, as frequency increased, homogeneity decreased regardless of the radius of
the GFF. Because experiments are planned for the entire 1 to 6 GHz frequency range,
ur
options available to us for obtaining the best possible homogeneity of E field and SAR
while exposing chromaffin cells to 1 – 6 GHz MW fields rely on tailoring the distribution
of chromaffin cells on the GFF for each frequency. At the lower frequencies this may
simply mean determining the optimal gasket size that should be used with the CPA to
ur
obtain maximal homogeneity for the E field and SAR. At the higher frequencies,
59
Table 3.3. Variation in homogeneity as a function of radius of the GFF versus frequency.
% refers to the area of the GFF homogeneous to within 30%.
however, a more challenging approach would be needed that would involve strategically
ur
placing the cells in regions of the GFF that would have homogeneous E field
distributions. For our waveguide-based exposure system [8] we had already determined
the optimal experimental conditions for on-line monitoring of catecholamine release from
perfused chromaffin cells using the CPA described here (for example, the number of cells
needed to give a good amperometric signal, the conditions for loading the cells into the
CPA, the maximal pressure that the cells can withstand once immobilized within the CPA,
the flow rate for superfusion, etc.). Thus, changing the distribution of the cells across the
GFF by concentrating the cells in certain areas may alter, for example, flow rate and
pressure, that may in turn also alter basal catecholamine release and release stimulated in
response to a drug added to the BSS perfusing the cells. Such changes in experimental
conditions could make comparison of MW effects at different frequencies difficult.
Consequently, ongoing studies are working out the conditions for minimizing such
60
changes using this approach, with the desired location of the cells on the GFF verified by
neutral red staining of the cells.
ur
Figure 3.15. XFDTD computed contour plotur of the E field (a) and SAR (c), and
histogram and descriptive statistics of the E field (b) and SAR (d) for a GFF having a
region of radius 7 mm available to cells at 6 GHz. Total absorbed power = 1.838 mW.
3.5 Conclusions
The details for the design, characterization and optimization of a free-space MW
exposure system that will be used for identifying MW parameters over the frequency
range 1 to 6 GHz that elicit nonthermal effects on catecholamine release from chromaffin
61
cells have been presented. The rationale for the design of the exposure system was
governed by several factors. These include our use of an experimental approach in which
catecholamine release is to be monitored on-line during the exposures, the desire to have
all parameters associated with the MW source signal monitored and controlled, and the
ability to maintain temperature to within ± 0.2°C of a setpoint by means of a dynamic
control system that operates under both pulsed and CW field conditions. Our use of the
FDTD method enabled the characterization and optimization of the entire exposure
ur
ur
system so that the maximum possible E field as well as an acceptable level of E field
homogeneity in the region where the chromaffin cells will be located during experiments
could be obtained. To validate the FDTD model, the results of the FDTD computations
were compared, whenever possible, to results obtained by other methods, including direct
measurements, and agreement was good. Thus, FDTD modeling will continue to be used
ur
for further optimization of the exposure system to ensure acceptable levels of E field
ur
homogeneity while obtaining maximal possible E field intensity over the entire 1 to 6
ur
GHz frequency range. As we have shown, E field and SAR homogeneity decrease as
frequency increases from 1 to 6 GHz under our standard cell perfusion conditions. FDTD
modeling has given us insight into how the size of the region on the GFF that is available
to the cells during experiments can be manipulated to counteract this frequency effect.
Finally, since the location of the cells on the GFF can be determined at the end of the
experiment by staining the cells with the dye neutral red, FDTD modeling can also be
ur
used to correlate any changes in catecholamine release with the distribution of the E
field and SAR.
62
3.6 Acknowledgements
The authors would like to thank Todd Hagan for assistance with the FDTD simulations,
David Brouse for assistance with the cell perfusion setup, and Dave Thomas, Antenna
Consulting Engineer, for his advice with respect to the antenna gain measurements.
63
3.7 References
[1] J. A. D’Andrea, C.K. Chou, S. A. Johnson, and E. R. Adair, “Microwave effects on
the nervous system,” Bioelectromagnetics Supplement, vol. 24, pp.S107-S147, 2003.
[2] K.R. Foster and H.P. Schwan, “Dielectric properties of tissues”, In C. Polk and E.
Postow (Eds.), CRC Handbook of Biological Effects of Electromagnetic Fields, Boca
Raton, FL: CRC Press, 1995, pp. 25-102.
[3] A.W. Chiabrera, B. Bianco, E. Moggia, T. Tommasi and J.J. Kaufman, “Recent
advances in the biophysical modeling of radiofrequency electromagnetic field
interactions with living systems”. In G.L. Carlo (Ed.), “Wireless Phones and Health”,
Norwell, MA: Kluwer Academic Publishers,1998, pp. 135-164.
[4] A.W. Chiabrera, M. Grattarola and R. Viviani, “Interaction between electromagnetic
fields and cells: microelectrophoretic effect on ligands and surface receptors,”
Bioelectromagnetics, vol. 5, pp. 173-191, 1984.
[5] G. D'Inzeo, P. Bernardi, F. Eusebi, F. Grassi, C. Tamburello and B.M. Zani,
“Microwave effects on acetylcholine-induced channels in cultured chick myotubes,”
Bioelectromagnetics, vol. 9, pp. 363-372, 1988.
[6] S.K. Dutta, B. Ghosh and C.F. Blackman, "Radiofrequency radiation-induced calcium
ion flux enhancement from human and other neuroblastoma cells in culture"
Bioelectromagnetics, vol. 10, pp. 197-202, 1989.
[7] E. Postow and M.L. Swicord, Modulated fields and “window” effects, In C. Polk and
E. Postow (Eds.), CRC Handbook of Biological Effects of Electromagnetic Fields, Boca
Raton, FL: CRC Press, 1986, pp. 425-461.
[8] T. Hagan, I. Chatterjee, D. McPherson and G. L. Craviso, “A novel waveguide-based
radio frequency/microwave exposure system for studying nonthermal effects on
neurotransmitter release-Finite-Difference Time-Domain modeling” IEEE Trans. Plasma
Science, vol. 32, No.4, Aug. 2004.
[9] J. Yoon, I.Chatterjee, D. McPherson and G.L. Craviso, “Design and characterization
of a broadband mini exposure chamber for studying catecholamine release from
chromaffin cells due to nonthermal levels of 1 – 6 GHz continuous and pulsed microwave
radiation – Finite-Difference Time-Domain computations”, in Symposium Record
Abstracts, ElectroMed2005 Fourth International Symposium on Nonthermal
Medical/Biological Treatments using Electromagnetic Fields and Ionized Gases, Portland,
OR, May 16-18, 2005, pp. 95 – 98.
[10] A. W. Guy, C.K. Chou and J.A. McDougall, “A quarter century of in vitro research:
A new look at exposure methods”, Bioelectromagnetics, vol. 20, pp. 21 – 39, 1999.
64
[11] O.P. Gandhi, G. Lazzi and C.M. Furse, “Electromagnetic absorption in the human
head and neck for mobile telephones at 935 and 1900 MHz”, IEEE Trans. Microwave
Theory Tech., vol. 44, pp. 1884 – 1897, 1996.
[12] K.S. Kunz and R.J. Luebbers, The Finite Difference Time Domain Method for
Electromagnetics, Boca Raton, FL: CRC Press, 1993.
[13] M. Popovic, S.C. Hagness and A. Taflove, “Finite-Difference Time-Domain analysis
of a complete transverse electromagnetic cell loaded with liquid biological media in
culture dishes”, IEEE Trans. Biomed. Eng., vol. 45, pp. 1067 – 1076, 1998.
[14] N. Kuster and F. Schönborn, “Recommended minimal requirements and
development guidelines for exposure setups of bio-experiments addressing the health risk
concern of wireless communications,” Bioelectromagnetics, vol. 21, pp. 508-514, 2000
[15] F. Schönborn, K. Poković, A. M.Wobus, and Niels Kuster, “Design, optimization,
realization, and analysis of an in vitro system for the exposure of embryonic stem cells at
1.71GHz,” Bioelectromagnetics, vol. 21, pp. 372-384, 2000.
[16] K. Shimada, T. Hayashi, M. Tokuda, “Fully compact anechoic chamber using the
pyramidal ferrite absorber for immunity test”, 2000 IEEE International Symposium on
EMC, vol. 1, pp. 21 – 25, Aug. 2000.
[17] Qin Yu, “Measurement of 1-40 GHz radiated electric field spurious emissions” 2001
IEEE International Symposium on EMC. vol. 1, pp. 13-17, Aug. 2001.
[18] Y.T. Lo and S.W. Lee, “Antenna handbook theory, applications, and design” New
York: Van Nostrand Reinhold Company, c1988.
[19] P. J. Sroka, "Nomograph saves time in converting antenna gain," Microwaves and
RF, pp. 54 – 55, Mar. 1974.
[20] D. Thomas, Antenna Consulting Engineer, Gardnerville, Nevada, personal
communication.
[21] B.-K. Chung and H.-T. Chuah, “Modeling of RF absorber for application in the
design of anechoic chamber”, Progress in Electromagnetics Research, PIER 43, pp. 273285, 2003.
[22] C. Bruns, P. Leuchtmann and R. Vahldieck, “Analysis and simulation of a 1 - 18
GHz broadband double-ridged horn antenna”, IEEE Transactions on Electromagnetic
Compatibility, Vol. 45, pp. 55 - 60, 2003.
65
[23] C. Bruns, P. Leuchtmann and R. Vahldieck, “Comprehensive analysis and
simulation of a 1 - 18 GHz broadband parabolic reflector horn antenna system”, IEEE
Transactions on Antennas and Propagation, vol. 51, pp. 1418 - 1422, 2003.
[24] T. S. Rappaport, “Wireless Communications: Principles and Practice”, Upper Saddle
River, NJ: Prentice Hall PRT, 1996, chapter 3.
66
Chapter 4
Improvements to the Perfusion/On-Line Detection System and
Additional Details
4.1. Introduction
The design, characterization and optimization of the broadband mini exposure
chamber are described fully in Chapter 3. This section gives a detailed description of the
various components of the perfusion/on-line detection system: pressure stabilization
(Section 4.2), injection system (Section 4.3), CPA temperature controller (Section 4.4),
ECD (Section 4.5), and automation (Section 4.6). Although these components were
initially described in [1], they have since been modified to achieve better performance
and characterized fully since each one plays an important role in conducting wellcontrolled biological experiments. Hence, the schematic of the original system (Figure
3.4.a) is now replaced by Figure 4.1, which is the schematic of the current improved free
space microwave (MW) exposure system.
4.2 Pressure Stabilization
Initial trials of the perfusion system showed that the pressure of the BSS flowing
within the system can vary gradually or fluctuate rapidly during an experiment. Such
changes in pressure affect the amount of CA released from the chromaffin cells, and
67
hence the output of the ECD. In the following section, the causes of such pressure
changes are identified, and the methods used to eliminate them are described.
Figure 4.1. Schematic of the current improved free space MW exposure system showing
the experimental arrangements inside the anechoic chamber and adjacent screen room.
4.2.1 Compression of the GFF by the CPA Screw
As discussed in Section 3.2.1, the two halves of the CPA screw into each other
(Figures 4.2.a, b and c).
68
(a)
(b)
(c)
Figure 4.2. Drawing of (a) the upper and lower halves of the CPA, (b) cross-sectional
view of the assembled CPA and (c) blown-up view of the components circled in blue in
(b).
69
The resulting compression of the rubber gaskets, as illustrated in Figure 4.3.a, serves to
prevent leakage of the BSS. Ideally, the thickness of the two compressed rubber gaskets
should be equal to the total thickness of the stacked components within the CPA (two 350
µm nylon meshes, the GFF, and the 5 µm mesh), so that the GFF, which is where the
cells are located, does not get compressed. Over tightening the two halves of the CPA
will compress the rubber gaskets to the point where the GFF is also compressed, as
illustrated in Figure 4.3.b.
(a)
(b)
Figure 4.3. The CPA assembled with the rubber gaskets, 350 µm meshes, 5 µm mesh and
GFF when (a) the GFF is maintained at the original thickness and (b) the GFF is
compressed by the nylon meshes due to over tightening.
After several experiments were conducted, it became apparent that the gaskets
being used in the smaller CPA were not thick enough to prevent compression of the GFF
when the two halves of the CPA were screwed together. As indicated previously,
complete tightening of the assembly is required in order to avoid leakage of the BSS.
While the compression of the GFF did not stop the flow of BSS, it increased the
70
resistance of the BSS to pass through the GFF, thus causing a gradual pressure increase
over time as shown in Figure 4.4. The pressure increase will affect the amount of drug
injected (Section 4.3) and hence the ECD output.
Figure 4.4. System pressure monitored by the pressure meter (location shown in Figure
4.1). The graph shows an increase in the system pressure over time due to over tightening
of the CPA.
The simplest way of removing a pressure increase due to compression of the
GFF is to slightly unscrew the GFF when an increase in pressure is detected. The
feasibility of doing this during an actual experiment when a pressure increase was
detected was tried several times. The conclusion reached was that it was best not to
unscrew the CPA. First, the moment the CPA was unscrewed, the pressure of the BSS
inside the CPA fluctuated suddenly, causing a stimulation of the cells that resulted in an
unwanted release of CA. Second, even though unscrewing the CPA reduced the
71
compression of the GFF, the GFF did not recover to its original thickness. Thus, the
pressure still increased. Third, unscrewing the CPA caused the GFF to tear, creating
particulates that caused even more fluctuations of system pressure.
The best solution therefore, was not to compress the GFF from the start. This can
be done using thicker gaskets (0.6 mm thick rather than 0.4 mm gaskets) such that even
after being compressed, an empty space will form above the upper 350 µm mesh and be
filled with BSS during perfusion (Figure 4.5.a). In this way, the GFF will obviously not
get compressed. The results of several experiments performed with 0.6 mm gaskets
showed that there was no increase in pressure. However, as determined by XFDTD
simulations performed at 3.5 GHz with 0.6 mm gaskets and a 0.2 mm thick layer of BSS
above the 350 µm mesh, the homogeneity of the E field and SAR on the GFF (Figure
4.5.b) calculated (using a program shown in Appendix E) was 21.7 % less than without
the BSS layer (Figure 4.5.c). Thus, the presence of the layer of BSS above the upper 350
µm mesh significantly affected the homogeneity of the E field and SAR on the GFF. To
improve the E field distribution in the region containing the cells while maintaining a
constant pressure, the solution was to use a combination of a 0.4 mm and a 0.6 mm
gasket (Figure 4.6). By placing the thinner gasket above the upper GFF, the gap filled
with the BSS was eliminated, hence improving the homogeneity. The thicker gasket that
was placed under the GFF now provided more space for all the components so that the
GFF did not get compressed.
72
(a)
(b)
(c)
Figure 4.5. Illustration of the CPA with 0.6 mm gaskets (a). The E field distribution over
the GFF with (b), and without (c) the additional BSS layer.
Figure 4.6. Illustration of the CPA using a 0.4 mm and a 0.6 mm gasket.
4.2.2 Number of Cells Loaded onto the GFF
When the chromaffin cells are loaded onto the GFF (Appendix B), they may not
be distributed consistently each time; for example, they can be concentrated in the central
area of the GFF, or be spread throughout the surface. Moreover, the cells can either
remain on the GFF surface or penetrate into the GFF, as illustrated in Figure 4.7.a. If too
73
many cells are distributed on the GFF, as illustrated in Figure 4.7.b, resistance to the flow
of the BSS in the system increases, thus causing the pressure of the BSS flowing within
the system to increase (Figure 4.4). On the other hand, if too few cells are loaded, the
amount of CA released may be too small to detect, causing a low signal to noise ratio in
the ECD output.
(a)
(b)
Figure 4.7. Illustration of the GFF (a) with evenly distributed cells and (b) with too many
cells present.
Experiments carried out with different numbers of cells loaded onto the GFF
showed that using a total of around one million cells does not increase the system
pressure, while giving a good response to DMPP over the baseline, as measured by the
ECD. This indicates that accurate cell counting is important. Cell number is recorded for
each cell preparation and the method used to count the cells is described in Appendix A.
4.2.3 Use of Microtubing
There are three separate pieces of microtubing that are inserted at different
locations in the system and these are shown in Figure 4.8. One is the microtubing used
74
for the delivery of DMPP (labeled 1) to a point right above the location of the cells. It is
flexible (Scientific Commodities INC. BB311-32) and with a small inner (0.25 mm) and
outer (0.51 mm) diameter, in order to fit inside a W connector (4 mm in diameter),
along with the inlet temperature probes (1.5 mm in diameter). The second piece of
microtubing is used for the ECD (labeled 2). It is made of hard plastic, with an inner
diameter of 0.25 mm and an outer diameter of 1.59 mm. This microtubing is supplied by
the manufacturer of the ECD and is designed specifically for the ECD to ensure
accuracy of the ECD output. The last piece of microtubing (labeled 3) is the same
microtubing used for delivery of DMPP to the cells (labeled 1) and is inserted at the end
of the system, allowing the system pressure to increase by 1 - 2 psi. This overall
increase in pressure of the system helps to maintain the system pressure at a constant
level during the course of an experiment. Tubings with an inner diameters 0.8 or 1 mm
are used for the rest of the system so that the BSS flows at a sufficient rate throughout
the system to avoid dilution of the injected DMPP and the CA that are released.
Since the smallest inner diameter of the microtubings is only 0.25 mm, any
particulates in the BSS, such as pieces of torn GFF or debris from protein build-up at any
point in the system, may clog the microtubing, causing a sudden pressure fluctuation. In
order to keep the system flowing well, it is cleaned after each experiment by flushing the
system sequentially with 0.1 N NaOH, DI water, 40 % ethanol, and finally copious
amounts of DI water (Appendix D). Because of the implementation of this rigorous
cleaning process after each experiment, pressure fluctuations due to particulates in the
system from prior use rarely occur.
75
Figure 4.8. Diagram of perfusion system showing the locations of the microtubing.
Microtubing ① is the DMPP injection tubing, microtubing ② is the inlet tubing for the
ECD and microtubing ③ is the outlet tubing of the system to maintain constant pressure.
4.2.4 Particulates in the BSS
The system pressure can also be increased by particulates present in the BSS
solution. For example, the salts used for preparing the BSS contain very fine particulate
matter. Also, because the BSS solution contains glucose, it can promote the growth of the
bacteria that can clog the microtubing, and/or the 5 µm mesh. To prevent any particles
from clogging the system, the BSS is filtered prior to use using a 0.22 µm filter
(Millipore GSWP 04700) and used within 24 hours (Appendix B).
76
4.2.5 Use of a 5 µm Mesh
Within the CPA, a 5 µm mesh is placed immediately under the GFF in order to
trap any cells, torn pieces of the GFF, or particulates in the BSS and thereby prevent the
passage of these materials into the ECD. However, if the 5 µm mesh is used several times,
it eventually gets clogged, especially by pieces of the GFF that are easily torn off when
the GFF is being removed from the CPA. To prevent this from happening, a new 5 µm
mesh is cut and used for each experiment (Appendix B).
4.3 Improvements to the Injection System
Figure 4.9.a is an exploded view of the injection system that was originally
implemented in the MW exposure system (Chapter 3). In this setup, the drug stimulus
was injected directly to the location of the cells via a 50 cm in length piece of tubing
(Figure 4.9.a, circled in red). To fill this tubing at the beginning of each experiment with
the drug stimulus, six to seven injections were needed. This was because the volume
needed to fill the tubing is large (392.3 mm3) and the pressure that builds up in the main
BSS inlet tubing (about 3-4 psi) pushes against the BSS containing the drug. However, by
adding an additional perfusion component in which BSS is pumped to the drug injector
tubing, as shown in Figure 4.9.b, the volume needed to fill the drug tubing is reduced to
7.9 mm3 and the pressure against the injector is reduced by half. Therefore, only three
injections are now needed to fill the tubing. A drawback of this method is that the
injected drug stimulus may get diluted while being delivered to the cells since the drug is
injected directly into the continuously flowing stream of BSS. Multiple experiments
77
using CA standards in place of DMPP showed that the dilution of the drug stimulus
would be minimal.
(a)
(b)
Figure 4.9. Diagram of the injection system (a) before and (b) after modification.
4.4 Improvements to the CPA Temperature Controller
The temperature controller that was first built was based on one used for the
78
waveguide exposure system described in [2] that also incorporated a CPA. Figure 4.10
shows detailed schematics of the CPA before and after modification.
In the original setup, the dynamic temperature controller was designed to control
the temperature of the BSS in the CPA at the location of the tip of the inlet temperature
probe [1] (the region circled in blue in Figure 4.10.a), which was 2.5 cm from the
location of the cells. Hence the temperature of the region where the cells are actually
located was neither monitored nor directly controlled by the temperature controller. The
outlet temperature probe, which monitors the temperature at the outlet of the CPA to
assess increases in temperature of the BSS during MW exposure of the GFF, was also
located at a distance (1.5 mm) from the GFF (circled in red in Figure 4.10.a). Hence, any
temperature changes that occurred during MW exposure of the GFF were not monitored
properly.
Figure 4.10.b illustrates the modified arrangement of the temperature control
system. A four way W-connector (shown circled in red) was connected to the inlet port
and a three way T-connector (shown circled in blue) was connected to the outlet port of
the CPA. These connectors allowed insertion of the temperature probes immediately
above and below the GFF where the cells are located. Hence, measurement of
temperature is performed as close as practically possible to the location of the cells, thus
providing better temperature control and monitoring in the immediate vicinity of the cells.
79
(a)
(b)
Figure 4.10. Detailed schematics of the CPA and surrounding components (a) before and
(b) after modifications. In the modified design, the temperature probes are located much
closer to the GFF and hence the cells.
Experiments performed under conditions where maximal power available from the MW
equipment was used to expose the CPA to MW fields showed that the temperature
control system was able to maintain the temperature to 36.5±0.15 oC at the location of the
cells during MW exposure.
4.5 Electrochemical Detector (ECD)
The ECD is the primary sensing device used to monitor the release of CA from
chromaffin cells. Among numerous analytes (where the analyte refers to the
80
biological/chemical species to be detected) that the ECD can detect, the target of
detection in this research are the CA norepinephrine and epinephrine that are synthesized
and released by chromaffin cells. Thus the ECD in this project is specifically serving as a
biosensor. For a biosensor to be useful, it must fulfill a number of requirements that
indicate its ability to measure a physical quantity. These requirements are sensitivity,
accurate calibration, linearity, detection limit, hysteresis, repeatability, specificity, and
selectivity [3, 4].
Understanding the principle behind electrochemical detection is necessary to
operate the device properly, to analyze the experimental data correctly, and to prevent
artifacts caused by inadvertent careless operation of the instrument. Therefore, the basic
operating principle behind the ECD is first explained. The characterization of the ECD as
a biosensor, as well as the sources of possible artifacts, are then described in detail in
following sections.
4.5.1 Principle of Electrochemical Detection [3-5]
In our experiments, the ECD is used in the amperometric mode, which allows
measurement of the concentration of the analyte on-line during experiments. In this mode,
the solution (BSS) containing the analyte (CA) flows past an electrode and a small
fraction of the analyte is oxidized, generating an electrical current that is proportional to
the concentration of the analytes at the electrode surface.
Amperometric detection is accomplished using a cross-flow electrochemical
transducer (Figure 4.11.a), which is composed of a silver/silver chloride reference
electrode (BASi MF-2078), working electrode block (BASi MF-1000), auxiliary
81
electrode block (BASi MF-1093), and a Teflon gasket (BASi MF-1046). The Teflon
gasket (0.05 mm thickness) is sandwiched between the working electrode block and the
auxiliary electrode block, forming a low-volume thin layer cell (circled in red in Figure
4.11.b) where the solution under test passes through. On the surface of the working
electrode block, a glassy carbon electrode is placed.
(a)
(b)
Figure 4.11. Cross-flow electrochemical transducer. (a) Exploded view [6] and (b)
illustration of a top cross-sectional view of the transducer showing the path of the BSS
(blue line), the low-volume thin layer cell (circled in red) and the glassy carbon electrode.
The electrode is held at a fixed potential relative to the reference electrode to oxidize the
analyte, and the electrons released during the oxidation are subsequently converted to a
current, which is recorded on a computer.
The response of the detector is described by the equation (1)
i = n ⋅ F ⋅ A ⋅ KT ⋅ c ⋅ u1/ 3
(1)
82
where n is the number of electrons per molecule involved in the reaction, F is the Faraday
constant (9.65 x 10 4 C / mole), A is the area of the working electrode, KT is the limiting
mass transfer coefficient, c is the analyte concentration, and u is the linear velocity of the
mobile phase over the surface of the electrode. In order for i to depend only on the
concentration (c), all the other values in equation (1) should be maintained constant.
Since n, F, A, and KT are constant, the value of i depends on c and u. Since u is
maintained constant by the peristaltic pump during experiments, hence i is dependent
only on c.
4.5.2 Sensitivity and Calibration
The sensitivity of a biosensor is a measure of its ability to detect small changes in
analyte concentration. As described in the previous section, the ECD electrode is held at a
fixed potential to oxidize the CA in the BSS. The factor that contributes to the sensitivity
of an ECD is the fixed potential. The higher the potential that is applied to the working
electrode the higher is the amount of oxidization, thus increasing the sensitivity of the
ECD. However, too high a potential may cause a relatively high background current as
well as a high level of noise that reduces the linear region of the output (see Appendix
G.5. for details).
In order to find an optimal potential for the electrode, a literature search was
conducted for papers that describe measurements of CA release from chromaffin cells
using an ECD [7-11] and based on these, several potentials were experimentally tested.
Consecutive experiments performed with each of the potentials reported in the literature
(0.6 V [8] and 0.65 [10]) showed that a potential of 0.65 V gave a high response from the
83
cells when stimulated with DMPP as well as a low baseline. The noise level at 0.65 V
was low compared to a high signal-to-noise ratio (SNR). For these reasons, a potential of
0.65 V has been used in all experiments described in this dissertation.
A biosensor is calibrated by varying the concentration of the target analyte and
measuring the response at each concentration. Hence, the ECD was calibrated using a
commercially available preparation of a CA standard that is a mixture of epinephrine,
norepinephrine and dopamine. For the experiment, different amounts of the CA standard
were injected into the ECD and the response measured, as shown in Figure 4.12. The
sensitivity (nA/ng) defined as the slope of the linear region, was found to be 1.17 (nA/ng).
Figure 4.12. Calibration curve for the ECD for an output range setting of 0 – 50 nA,
showing that the response is linear with the amount of injected CA (dotted line).
84
Figure 4.12 only shows the linear region, where the ECD output changes linearly
with the concentration of CA. However, in practice, there is a range of concentration
where the graph is not perfectly linear and where the ECD is still useful over this nonlinear range. Linearity is discussed in the next section.
4.5.3 Linearity
It is desirable that the response of a biosensor be linear over a broad range of
analyte concentrations, i.e. a biosensor should have a constant sensitivity over the
concentration range that may be encountered in an application. This linear range depends
on the output range setting (Appendix G.1), background current (Appendix G.2),
selection of the flowcell (Appendix G.3), selection of the flow rate (Appendix G.4), and
offset functions in the ECD (Appendix G.6). Hence it is necessary to understand these
factors in order to maximize the linear range. Each of these factors is described in
Appendix G, where the discussion leads to the conclusion that the fixed output range
setting of 50 nA provides a broad detection range (0 – 500 nA) and high sensitivity
(Appendix G.5).
Figure 4.12 shows that the response of the range under 50 nA is linear, and hence
the ECD meets the criteria for linearity that is a requirement of a biosensor. However, the
range of injected CA that shows a linear response is very narrow (20 - 37.5 ng), which is
problem because the response of the cells, i.e. the amount of CA released from the cells
when stimulated with DMPP, varies widely from experiment to experiment (Appendix
G.5). It should be noted that the ECD can measure a ten times greater current (500 nA)
than its output range setting (50 nA) ([6] and Appendix G.1). Hence, to obtain a broader
85
linear range, a range that is larger than the output range setting (50 - 500 nA) was
investigated. If the response within this range is non-linear, then the heights of the peaks
can be obtained in an indirect manner. Once the non-linear region is carefully
characterized i.e., a non-linear curve is obtained, as illustrated in Figure 4.13, the heights
of peaks outside the linear region can then be compensated for and used as a linear value
[3].
Figure 4.13. Illustration of the non-linear curve (red curve) and the compensated linear
curve (dotted line).
In order to ensure the feasibility of using the non-linear region, a detailed
characterization was performed as shown in Figure 4.14. It was observed that most of the
range was linear except the range inbetween 470 and 500 nA, indicating that the only
range required to be compensated for and used as a linear value (dotted line) is 470 to
500 nA. Consecutive experiments performed at a higher output range setting (100 nA)
86
showed that the response between 470 to 500 nA is linear, indicating that the nonlinearity shown in this range using an output range setting of 50 nA is not because of
non-linearity in oxidization but instead because of a non-linear response of the LC-4C
that measures, amplifies, and outputs the current in this region.
Figure 4.14. Calibration curve for the ECD including the non-linear range (50-500 nA)
when the output range is set to 50 nA. The response is non-linear between 470 to 500 nA.
It should be noted that the heights of CA peaks obtained from all experiments performed
(Chapter 6) were less than 300 nA. Thus, the results presented in this dissertation were
derived from CA peaks that did not require a compensation process.
87
4.5.4 Detection Limit
The detection limit of a biosensor is the upper limit of the region where the
measured concentration of the analyte becomes inaccurate. As explained in Section
4.5.3.a, the ECD can measure a ten times greater current (500 nA) than its output range
setting (50 nA). Peaks greater than 500 nA are displayed as a horizontal flat line (Figure
4.15, circled in red). Hence, the detection limit of the ECD in our application is 500 nA.
Figure 4.15. Calibration curve for the ECD in the non-linear range (50 to 500 nA) and the
detection limit when the output range is set to 50 nA. Circled in red is an exploded view
of the top of a peak that is outside the detection limit.
4.5.5 Hysteresis
The flowcell is affected by previous measurements because a film builds up on
the surface of the carbon electrode due to heavy oxidization. For this reason, the ECD
may not always produce the same reading when measuring the same concentration of an
88
analyte. The maximum range of the expected error caused by the previous experiment is
defined as hysteresis. This can cause lowering of sensitivity and an increase in nonlinearity of the ECD response during successive experiments. For this reason the
electrode is properly cleaned before each experiment using the protocol described in
Appendix D.
4.5.6 Repeatability
It is important that the response of the ECD to the same amount of CA be
constant throughout an experiment. Although a film builds up on the surface of the
carbon electrode due to heavy oxidization, it does not build up fast enough to affect the
response during the course of a four hour experiment. In order to verify this, a constant
volume (amount) of the CA standard was repeatedly injected into the ECD using the
same injection protocol as that used for DMPP during a typical experiment (Appendix C).
This consisted of a 10 second pulse of the CA standard injected every 10 minutes, for a
total of 20 injections. Figure 4.16 shows that the response of the ECD to each injection of
the CA standard is constant, indicating that the ECD response does not change during the
course of a four hour experiment.
4.5.7 Specificity and Selectivity
Specificity is defined as the ability of a biosensor to recognize a single analyte
when several different analytes are present in the same sample. Since detecting only one
analyte is often impossible, the term selectivity is used instead of specificity.
89
Figure 4.16. Response of the ECD to repeated injections of a constant volume of CA
standard. Injections were delivered every 10 minutes for up to four hours.
It should be noted that the ECD does not have good specificity. This is because
the ECD will oxidize all substances capable of being oxidized at the set potential.
However, as described in Appendix G.2, a series of experiments established the
selectivity of the ECD for recognizing CA released from the chromaffin cells.
4.5.8 Factors that Can Generate Artifacts
Two specific factors have been identified that can affect the response of the ECD.
These are a change in the speed of the flow of the BSS and a change in the temperature of
the BSS.
90
4.5.8.a Fluctuations in BSS Flow Rate
As described in equation 1 (Section 4.5.1), the ECD response is proportional to
the speed (u) of the solution passing through the flowcell. That is, if the speed of the flow
is increased, then the output current (i) is also increased, and vice versa. To quantify how
much the response of the ECD changes due to variations in flow rate of the BSS, two
types of experiments were carried out. In the first experiment, the response of the ECD
was monitored while the flow rate was gradually increased from 0.8 ml/min to 1.2
ml/min and then decreased to 0.8 ml/min, i.e., a 0.4 ml/min overall change in flow rate
that is much larger than that which would occur in an actual experiment. As shown in
Figure 4.17, the background current was, as expected, proportional to the flow rate. The
0.4 ml/min increase and decrease in flow rate caused an increase and decrease,
respectively, of approximately 4.0 nA in the ECD output, which is about 10 % of steadystate background current. Since the maximal change in flow rate detected during the
course of a four hour experiment never exceeds 0.05 ml/min, the maximal change in the
ECD response for the background current as well as for the height of the peaks
corresponding to stimulated CA release is only 1.25 %. Thus, the magnitude of the
fluctuations in BSS flow rate that occur during a typical experiment is not enough to
significantly affect the ECD response.
91
Figure 4.17. ECD output is proportional to the flow rate of the BSS. Flow rate was
gradually increased and then decreased by 0.4 ml/min.
The second experiment assessed the effect of a sudden fluctuation in flow rate on
the ECD response. The strategy used was to stop the flow (flow rate = 1 ml/min) by
simultaneously pinching the ECD inlet tubing (labeled 2 in Figure 4.8) using a plastic
pinch valve and stopping the peristaltic pump for 10 seconds; the pinch valve was then
removed and the pump restarted. As expected, this had a greater effect on the ECD
response than a gradual change in flow rate. As shown in Figure 4.18, the background
current increased suddenly and then gradually decreased. However, the current stabilized
at a higher level than the initial background. Thus, a short-term fluctuation of only 10
seconds can cause a long-lived change (greater than 5 minutes) in the background current
92
(circled in red), underscoring the importance of preventing sudden fluctuations in the
flow rate of the BSS during experiments.
Figure 4.18. Schematic diagram of how the ECD output is affected by stopping the flow
of the BSS for 10 seconds.
The main reasons why sudden fluctuations in the flow rate of the BSS can occur
as well as the methods used to prevent them have been described previously (Section 4.2).
However, the pump controller whose purpose is to maintain the flow of the BSS at a
constant rate can also cause sudden alterations in the flow rate of the BSS. Such
alterations can occur, for example, when a large volume of DMPP is injected. While the
controller compensates for the additional volume injected into the flow of the BSS, an
oscillation in the flow can be introduced by the pump controller, which eventually
disappears. This fluctuation by the pump controller can be eliminated by locking the
speed of the pump, once a pump speed of 1 ml/minute is achieved by the pump controller
after cell loading. That is, the automatic flow controller is turned off during an
experiment (Appendix B). Also, in order to reduce flow rate fluctuations due to injections
93
of DMPP, a minimum volume of DMPP that does not affect the flow rate, yet produces
measurable CA peaks, was determined by a trial and error method. The optimal volume
was found to be 10 µl with a DMPP concentration of 75 µM, and these values were used
for all experiments.
4.5.8.b Addition of an ECD Temperature Controller
The ECD user manual indicates that changes in temperature may affect the
response of the ECD [6]. This is not surprising since the electrochemical reaction
(oxidization) that takes place is temperature dependent. To verify how changes in
temperature affect the response of the ECD, the response to repeated injections (every 2
minutes) of the CA standard was measured as the temperature of the BSS at the CPA
inlet was increased from 30 to 35 oC at a rate of 1 oC/minute using the CPA inlet
temperature controller (described in Section 4.4). The change in temperature at the ECD
flowcell was monitored by a resistor-based thermometer mounted on the wall of the
auxiliary electrode block (Figure 4.19.a). Figure 4.19.b shows that as the BSS
temperature at the inlet was increased, the temperature of the auxiliary electrode block
also gradually increased due to the BSS heating the metal body of the auxiliary electrode.
With respect to the ECD profile, both the background current and the area under the peak
increased (15.2 % and 23.2 % respectively) as the temperature of the BSS solution
increased. This result demonstrates that the response of the ECD is affected by an
increase in the temperature of the BSS.
The primary way that the temperature of the BSS could increase during an
experiment is when the cells are exposed to MW fields. To assess this, the CPA without
94
(a)
(b)
Figure 4.19. Effect of temperature on the ECD response. (a) Diagram showing the
flowcell [6] with a resistor based thermometer mounted on the wall of the auxiliary
electrode block and (b) the CA standard response while the temperature of the BSS was
increased.
95
cells was exposed to a MW field at a frequency of 3.5 GHz, the center frequency of the
range of interest (1-6 GHz), and the ECD response was measured as shown in Figure
4.20.
Figure 4.20. Response of the ECD during exposure of the CPA to a MW field at 3.5 GHz
using maximum MW power. Heating of the BSS by the MW field caused the baseline
response of the ECD to change.
When the maximum MW power was applied with different ramping schemes, not only
does the temperature of the BSS at both the inlet and the outlet of the CPA increase, but
the baseline response of the ECD also changes, showing that heating of the BSS due to
MW fields can cause an artifact in the ECD response.
This MW heating induced artifact can be prevented by maintaining the
temperature of the BSS flowing into the flowcell constant to within ± 0.45 oC of
96
computer controller set point during MW field exposure. To control temperature this
precisely, a computer controlled temperature controller comprised of nichrome wire
wound around the ECD inlet tubing and a needle type thermocouple to monitor the
temperature and provide feedback to a LabVIEW program was constructed (Figure 4.21).
The effectiveness of this temperature controller system is shown in Figure 4.22.
Figure 4.21. Diagram of the computer controlled temperature controller that maintains
the temperature at the inlet of the flowcell to within ± 0.45 oC of the computer controller
set point.
When the ECD inlet temperature controller was turned off during MW exposure, the
temperature at the ECD inlet increased by 2 oC (labeled 2). However, such a temperature
change was not detected (labeled 2) during MW exposure when the temperature
controller was operating. Instead the inlet temperature of the ECD was maintained at 35
o
C.
97
Figure 4.22. Experimental profile showing the temperature at the CPA inlet and CPA
outlet (label 1) and the ECD inlet (label 2) with the temperature controller either turned
off (left panel) or turned on (right panel). Pressure and flow rate (label 3) and power
(label 4) are also shown. When maximum MW power was applied at 3.5 GHz, the
temperature at the ECD inlet increased by 2 oC when the ECD inlet temperature
controller was off and was maintained constant to within 35 ± 0.45 oC when the ECD
inlet temperature controller was on.
For CW exposures, which produced the greatest likelihood of heating the BSS,
the temperature of the ECD inlet was maintained at 35 oC. For pulsed MW experiments,
it was maintained at 30 oC since the maximum temperature increase that would be
produced would be only 0.2 oC and thus not enough to increase the temperature of the
BSS at the ECD inlet.
98
4.5.9 Area under the Peak
During both control and MW exposure experiments, the cells were stimulated by
DMPP every 10 minutes to elect the release of CA. The amount of released CA was
displayed by the ECD as a CA peak. It should be noted that there exist two possible ways
of assessing the amount of released CA: (1) the height of the peak and (2) the area under
the peak (Figure 4.23).
Figure 4.23. Height of and area under the CA peak.
The question is which of these parametesr provides the most reliable assessment of the
amount of released CA during an experiment. In order to answer this question, it is
necessary to understand what the height and area of the peak represent. This can best be
accomplished by flow analysis. The concentration of DMPP at the outlet of the drug
injector (location 1 in Figure 4.24) is expected to be homogeneous. However, as
99
Figure 4.24. Simplified diagram of Figure 4.1 showing the points at which concentration
of the DMPP and the released CA, where location (1) is the outlet of the drug injection,
(2) is the inlet of the CPA and (3) is the inlet of the ECD.
DMPP, which is injected as a 10 second pulse, flows from location (1) to location (2),
trailing occurs (Figure 4.25a). This is because the speed of the liquid flowing in the
middle of a tubing is faster compared to that along the tubing wall due to the viscosity of
the liquid [12]. Flow modeling (Figures 4.25.b and c) performed using the commercially
available software package COSMOSFloWorks (SolidWorks Corp.) confirmed the
existence of different speeds of liquid flowing in a tubing.
100
(a)
(b)
(c)
Figure 4.25. Illustration of (a) a bolus of injected DMPP flowing in a tubing and the
trailing of DMPP that occurs along the walls due to viscosity of the BSS. The simulated
speed of the flow of DMPP in a cross-sectional view through (b) the vertical plane and
(c) the center horizontal plane of a tubing, obtained using COSMOSFloWorks, are also
shown. The inner diameter of the tubing used for the simulation was the same as that
comparing the actual setup.
101
This means that the DMPP that will eventually reach the cells is not uniform in its
distribution of drugs molecules.
Once DMPP enters the CPA and reaches the cells, CA are instantly released. The
next consideration pertains to the flow of the released CA. It should be noted that there
are several paths that the flow of BSS can take in passing through the CPA (Figure 4.26).
These paths are through the center of the CPA (black lines), over the 350 µm meshes
(gray lines) or inside the GFF (pink lines). It is expected that the flow of BSS in the
Figure 4.26. The cross-sectional view of the CPA showing multiple flow paths.
102
center of the CPA (black) takes the shortest time to pass through the CPA, whereas the
flow paths further away from the center of the CPA take longer. These time differences
caused by the multiple flow paths of the BSS in the CPA increase the extent of trailing of
the released CA, and hence increase the inhomogeneity of the concentration distribution.
Moreover, the CA are further diluted and the extent of the trailing is increased while
passing through the tubing between the CPA outlet and the ECD inlet (location 3). The
expected concentration profile of the CA based on the ECD output is illustrated in Figure
4.27.
Figure 4.27. Illustration of the concentration profile of the CA at the inlet of the ECD.
103
Figure 4.27 indicates that the height of a CA peak only represents a small portion
of released CA (box in yellow), possibly the CA delivered by the flow at the center of the
CPA (black line in Figure 4.26), while ignoring the part of the flow that still contains
released CA (labeled 1). Therefore, in order to more accurately reflect all the released CA,
the area under the peak was used for all the analyses presented in this dissertation. It
should be noted that the measured height of the peaks and the area under peaks showed
very similar patterns for most of the experiments, with the exception of experiments
where CA peaks were small (Section 6.8). Thus, the height of the peaks was also
representative of the concentration of released CA.
4.6 Automation
As shown in Figure 4.1, instruments that are used in the exposure system are
either computer-controlled or computer-monitored via programs written in-house using
LabVIEW (National Instruments version 7.0), where the red and orange lines connected
to the computer represent control cables that are used to communicate with each
instrument. After the cells are loaded onto the GFF, the drug injection amount and times
of application are typed in, and MW exposure parameters are set. The experiment can
then be carried out automatically, hence greatly reducing human error.
There are seven programs that constitute the automation process. A brief
description of the automation process, describing which programs perform what tasks, is
as follows.
Once the cells are loaded onto the GFF, the pump controller program gradually
increases the speed of the peristaltic pump from zero to the point where the flow rate
104
becomes 1 ml/min in 15 minutes (Appendix H.1). Over this same time period, the
temperature controller program increases the temperature at the location of the cells
from room temperature to 36.5 oC (Appendix H.2). The flow and temperature controller
programs are programmed in such a way as to minimize stimulation of the cells by
sudden pressure or temperature changes. Once the temperature of the BSS inlet reaches
36.5 oC, it is maintained by the temperature controller program during the entire
experimental period. The injector controller program then injects the drug stimulus as
instructed, typically every 10 minutes, to elicit the release of CA from the cells
(Appendix H.3), and the responses can be reviewed by the operator through the ECD
monitoring program, which monitors and records both basal release and stimulated
responses throughout an experiment (Appendix H.4). The signal generator generates the
MW field with specific parameters entered by the operator using the MW controller
program (Appendix H.5). The amplifier, which can be enabled or disabled by the
amplifier controller program, then delivers the MW field to the horn antenna (Appendix
H.6). During MW exposure, the power monitoring program monitors and records the
power delivered to the antenna (Appendix H.7). The temperature of the CPA inlet is
controlled by, and the temperature of the auxiliary electrode block in the ECD and the
temperature inside RAE (Section 4.5.7.b), are monitored by the ECD inlet temperature
controller (Appendix H.8). All the monitored data obtained from the LabVIEW programs
listed above (Appendixes H.1 – H.8) are recorded into a text file by the data logging
program (Appendix H.9).
105
4.7 References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
J. Yoon, I. Chatterjee, D. McPherson, and G. L. Craviso, "Design,
Characterization and Optimization of a Broadband Mini Exposure Chamber for
Studying Catecholamine Release from Chromaffin Cells Exposed to Microwave
Radiation: Finite-Difference Time-Domain Technique," IEEE Transactions On
Plasma Science, vol. 34, pp. 1455-1469, 2006.
T. Hagan, I. Chatterjee, D. McPherson, and G. L. Craviso, "A novel waveguidebased radio frequency/microwave exposure system for studying nonthermal
effects on neurotransmitter release-Finite-Difference Time-Domain modeling,"
IEEE Trans. Plasma Science, vol. 32, 2004.
D. G. Buerk, Biosensors Theory and Applications. Lancaster, PA: Technomic
Publishing Co., Inc, 1993.
J. Yoon, "A Microstrip-Based Radio-Frequency Biosensor for the Detection of
Bacteria in Water," in Electrical Engineering, vol. M.S. Reno: University of
Nevada, Reno, 2003.
J. Yoon, I. Chatterjee, and G. L. Craviso, "Feasibility study of using two types of
microstripline bandpass filters for the detection of bacteria in water," presented at
Proceedings of the 37th European Microwave Conference, Munich, Germany,
2007.
Instruction Manual: Principles of EC Detection and Troubleshooting Guide. West
Lafayette Indiana: Bioanalytical Systems, Inc, 1994.
B. Lara, M. G. Lopez, M. Villarroya, L. Gandia, L. Cleeman, M. Morad, and A. G.
Garcia, "A Caffein-Sensitive Ca2+ store modulates K+-evoked secretion in
chromaffin cells," American journal of physiology. Cell physiology, vol. 41, pp.
1211-1221, 1997.
D. J. Green and R. L. Perlman, "On-Line Measurement of Catecholamine
Secretion," Analytical Biochemistry, pp. 270-276, 1981.
M. Herrera, L.-S. Kao, D. J. Curran, and E. W. Westhead, "Flow-injection
analysis of catecholamine secretion from bovine adrenal medulla cells on
microbeads," Analytical Biochemistry, vol. 144, pp. 218-227, 1985.
M. G. López, C. Montiel, C. J. Herrero, E. García-Palomero, I. Mayorgas, J. M.
Hernández-Guijo, M. Villarroya, R. Olivares, L. Gandía, J. M. McIntosh, B. M.
Olivera, and A. G. García, "Unmasking the functions of the chromaffin cell α7
nicotinic receptor by using short pulses of acetylcholine and selective blockers,"
Proc. Natl. Acad. Sci. USA, vol. 95, pp. 14184-14189, 1998.
R. Borges, F. Sala, and A. García, "Continuous monitoring of catecholamine
release from perfused cat adrenals," Journal of Neuroscience Methods, vol. 16, pp.
289-300, 1986.
R. W. Fox and A. T.McDonald, Introduction to Flow Mechanics, 3rd ed. New
York: John Willey & Sons, 1985.
106
Chapter 5
Characterization and Optimization of the Exposure System
Based on Measurements and Numerical Computations
5.1 Introduction
The design, characterization and optimization of the broadband MEC using a
CPA are described in Chapter 3. This chapter describes more detailed characterizations
that were omitted in Chapter 3 as well as improvements made to the system since the
paper [1] was published: detailed broadband horn antenna characterization, detailed E
field distribution on the GFF, use of the near field to obtain greater E field magnitude,
temperature distribution on the GFF, and calculation of the measured power delivered to
the horn antenna and temperature increase due to the MW exposure.
5.2 Detailed Antenna Characterization
The antenna gain and radiation patterns of the broadband horn antenna were
discussed in Section 3.4.1. In this section, additional experimental and numerical
characterizations performed for the broadband horn antenna are described.
5.2.1 S Parameters of the Broadband Horn Antenna
The performance of an antenna can be characterized by a set of parameters called
scattering (S) parameters. For an antenna, the S11 parameter, which quantifies how much
107
of the input signal is reflected, is especially important, since the S11 parameter provides
the acceptable operating frequency range and input impedance of the antenna under test.
The measured S11 parameter for the broadband horn antenna was compared to that
computed using XFDTD in order to validate the accuracy of the SolidWorks model of the
antenna.
5.2.1.a Measurement of the S11 Parameter of the Broadband Horn
Antenna
The S11 parameter of the broadband horn antenna was measured inside the
anechoic chamber using a network analyzer (Hewlett-Packard 8720C) and is plotted in
Figure 5.1. It is observed that the S11 at frequency range 1 to 6 GHz is lower than -3 dB
for the entire frequency range, indicating less than half of the input power is reflected at
the port of the antenna, i.e., the broadband antenna functions satisfactory over the entire
frequency range of interest (1 to 6 GHz)
108
Figure 5.1. The measured S11 parameter of the broadband horn antenna over the
frequency range 1 to 6 GHz.
5.2.1.b Computation of the S11 Parameter of the Broadband Horn
Antenna
In order to build an accurate SolidWorks model of the horn antenna, the horn
antenna was completely disassembled and the dimensions of all components were
carefully measured using calipers. The model was then imported into XFDTD and the S11
parameter was computed and compared to the measured value (Figure 5.2). It is observed
that the S11 values at the lower frequencies were comparable to the measured values,
however, the S11 values at frequencies higher than 3 GHz were greater than the measured
109
Figure 5.2. Comparison of the measured and calculated S11 parameter of the broadband
horn antenna.
values. The network analyzer used to measure the S11 parameter were carefully calibrated
and the performance was tested using standard attenuators whose values are known, and
no errors were found. This indicates that there may exist an error or errors in the
SolidWorks model. To find the sources of error in the modeling, first, the geometry of the
SolidWorks model was carefully compared to that of the actual antenna, and no major
differences that could cause the difference between the measured and simulated S11
parameter were found.
The error in the computation of the S11 was identified after intensive research as
110
being due to an unexpected capacitance induced at the feed port of the XFDTD model. It
was recognized that the accuracy of any XFDTD computation depends heavily on the
proper design of the feed port. Thus, the detailed design of the feed port is described in
the following section.
5.2.1.c Design of the Antenna Feed Port in the XFDTD Model
Figure 5.3.a shows the cross-sectional view of the center horizontal plane of the
broadband horn antenna and Figure 5.3.b is an exploded view of the feed components
(circle in Figure 5.3.a). The green line shown within the circle in Figure 5.3.b is called
“port element” in the XFDTD model, which excites the antenna with MW. The length of
the port element is the dimension of one mesh and thus the distance between the center
conductor of the horn antenna and the ground plane is only 1 mm in this specific
arrangement. Since the ground plane and center conductor at different potentials are
placed close to each other (0.5 mm apart), they behave like a capacitor, i.e., when the port
element is excited with MW, the two conductors (circled in yellow in Figure 5.3) couple
to each other, and hence the calculated S11 parameter becomes inaccurate at higher
frequencies (above 3 GHz) as shown in Figure 5.2.
In order to reduce the capacitance between the ground plane and the center
conductor, a four step matching port (Figure 5.4) was designed and inserted inbetween
the ground plane and the horn antenna port. The matching port was designed to reduce
the capacitance between the ground plane and the horn antenna port while maintaining an
input impedance of 50 Ω. The matching port is composed of four 50 Ω coaxial cables,
each of which has a different center conductor diameter and dielectric constant. The
111
(a)
(b)
Figure 5.3. Cross-sectional view of the center horizontal plane of (a) the entire broadband
horn antenna and (b) an exploded view of the antenna’s feeding components (yellow
circle in (b))
critical part of the design for the matching port is step 1, shown in Figure 5.4. The key
technique for step 1 is to reduce the capacitance between the two conductors shown in the
yellow circle (Figure 5.3.b) by minimizing the diameter of the center conductor. This idea
is realized by using an “electric wire” in the XFDTD design, the thinnest possible
diameter of a conductor in the software, as the center conductor. The relative permittivity
of the dielectric material that makes this section 50 Ω was chosen to be 19 as determined
by a trial and error method obtained from a separate XFDTD computation. This 50 Ω
coaxial cable with a very thin center conductor removes the unnecessary capacitance
between the center conductor and the ground plane. Step 4 is an air filled coaxial cable
that has the same diameter (8 mm) as the horn antenna input port. The diameters of the
center conductor for steps 2 and 3 increase to compensate the discontinuity problem that
112
occurs when the dimension of a transmission line changes, introducing parasitic
reactances that can lead to phase and amplitude errors, input and output mismatch, and
possibly spurious coupling [2].
Figure 5.4. SolidWorks model of the four step matching port that is connected to the
input port of the horn antenna to reduce the capacitance problem.
5.2.1.d Validation of the Broadband Horn Antenna
In order to validate the performance of the four step matching port, two four step
matching ports were connected back to back as shown in Figure 5.5, and the scattering
parameters (S11 and S21) and the input impedance were computed using XFDTD. If the
port is designed properly, the impedance of the entire coaxial cable should be maintained
at 50 Ω, and there is very little reflection of the input signal at the feeding port (i.e. a
113
(a)
(b)
Figure 5.5. The SolidWorks model (a) and (b) XFDTD model (cross-sectional view) of
two four step matching ports connected back to back.
small value of S11 is expected) with most of the input signal transmitted to the output port
(0 dB value for S21 is expected). The calculated S parameters (Figure 5.6.a) indicate that
the S11 value is close to -24 dB (mean = -24.008 ± 6.854 dB) and the S21 value is close to
0 dB (mean = -0.03192 ± 0.02502 dB), meaning that most of the input signal is not
reflected back to the source but rather passes through the feed port into the antenna over
114
the entire frequency range of interest (1 to 6 GHz). The calculated impedance (Figure
5.6) is maintained at 50 Ω over the entire frequency range (mean = 50.752 ± 5.356 Ω).
The results clearly show that the use of two back to back four step matching ports
removed the capacitance problem discussed in the previous section, since, unlike Figure
5.2, there is no sign of a short circuit due to the capacitance at the higher frequencies.
Figure 5.6. The calculated S parameters and impedance of two four step matching ports
connected back to back.
The complete XFDTD model of the antenna with the four step matching port
connected to the input port of the horn antenna is shown in Figure 5.7. Figure 5.7.a is a
three dimensional view of the entire antenna mesh with the designed four step matching
port connected (yellow circle) at the input port of the horn antenna. The cross-sectional
view of the horn antenna (Figure 5.7.b) and its exploded view (Figure 5.7.c) shows how
115
the matching circuit is connected to the horn antenna.
(a)
(b)
(c)
Figure 5.7. (a) Final XFDTD mesh of the horn antenna with the matching feed port, (b)
cross-sectional view of the horn antenna, and (c) close-up view of the feed port.
The computed S11 parameter (Figure 5.8) was then compared with the measured
value. The mean ± S.D. of difference between the measured and computed S11 parameter
was 1.52 ± 1.58 dB. This result leads to the conclusion that the broadband horn antenna
has been accurately modeled, and hence can be used for all computations in XFDTD.
116
Figure 5.8. Measured and computed S11 parameter of the broadband horn antenna.
5.2.1.e Conclusion
A comparison of the S parameters of the broadband horn antenna numerically
computed using the FDTD method and measured with a network analyzer confirmed that
there was an error in the XFDTD modeling. In order to remove the error in the
computation, a four step matching port was designed and implemented in the model and
enabled the error to be removed in the computation. The S11 parameter of the broadband
horn antenna with the four step matching port was computed and compared with the
measured value over the frequency range of interest. The result confirmed that the
117
computed and measured parameters were comparable, confirming that the broadband
horn antenna was now correctly designed.
5.2.2 Three Dimensional Broadband Horn Antenna Radiation Pattern
The two dimensional horn antenna radiation patterns at 3.5 GHz was
presented in Figure 3.9. A more descriptive and complete radiation pattern of an antenna
can be provided by a three dimensional radiation pattern, which provides the relative
magnitude of the E-field in various directions. In order to plot the three dimensional
radiation pattern, the E field was measured using a computer controlled tower and a turn
table placed in the anechoic chamber. A program written in Visual Basic moves the
receiving antenna to different height and turns the transmitting antenna (antenna to be
characterized) to measure the E field.
The experimental setup that measures the three dimensional radiation pattern is
shown in Figure 5.9. The program initially moved the receiving antenna to 1 m higher
than the transmitting broadband horn antenna placed at 1.5 m from the receiving antenna,
and then turn the receiving antenna through 360o while measuring and recording the E
field every 2 oC. Once the radiation pattern of the top plane was completed, the program
lowered the receiving antenna in steps of 10 cm, and measured the radiation pattern at
each plane. The program was terminated when the receiving antenna completed the
measurement for the plane 1 meter below the broadband horn antenna (shown dashed).
The receiving antenna frequency range did not cover the range from 1 to 2 GHz.
Hence, the patterns for this frequency range were not measured. A total number of 64800
readings for the nine frequencies (2 GHz to 6 GHz with 0.5 GHz step) were successfully
118
obtained without the presence of the operator in the anechoic chamber and the measured
radiation patterns at 3.5 GHz for both the E and H planes are presented in Figure 5.10.
Figure 5.9. The experimental setup that measures the three dimensional radiation pattern.
(a)
(b)
Figure 5.10. Measured (a) E plane patterns and (b) H plane patterns for 3.5 GHz.
119
As expected, the intensity of E field in the middle of the tower, where the two
antennas are aligned, is the strongest for both E and H plane, and the E field intensity is
decreased as the height of the tower is moved away from the middle of the tower for all
the frequencies measured (2 GHz to 6 GHz). The radiation patterns show a decreased
pattern width as the frequency increased and hence higher directivity.
5.2.3 E Field Distribution Inside the MEC
As described in Section 3.4.2, the purpose of the MEC is to achieve a
homogeneous E field distribution at the location of the cells on the GFF by reducing the
reflections from the RAE. The system used to measure the E field distribution inside the
MEC was described in Section 3.4.2 and is illustrated in Figure 5.11. As described in
Section 3.4.2, the E field up to 300 mm away from the absorber tips (label 1) in the
frequency range 1–6 GHz in steps of 0.1 GHz (total of 50 frequencies) was successfully
Figure 5.11. Illustration of the system that measures the E field at the center horizontal
plane that contains the GFF.
120
measured. The E field distributions for the entire frequency range (50 distributions) are
plotted in one graph (Figure 5.12) to observe any field inhomogeneities. Since E-field
distributions at 50 different frequencies are plotted on a single graph, it is hard to observe
a single distribution, yet it appears that there exist no significant E field distribution
Figure 5.12. E field distributions for the frequency range 1–6 GHz in steps of 0.1 GHz
(total of 50 frequencies) measured using a computer controlled electromechanical
actuator and a spiral antenna.
121
inhomogeneities, proving that the MEC properly reduces reflections from the RAE and
that E field distributions within the MEC are fairly uniform for the entire frequency range.
Although the distributions in Figure 5.12 provide a good overview of the E field
distribution in a wide region inside the MEC, it is not possible to achieve detailed E field
distributions in the small region containing the cells. The limitation on this measurement
was that the measured E fields were averaged over the entire surface of the spiral antenna
which was much larger than the GFF surface as illustrated in Figure 5.13. Since the E
field on the GFF is of particular interest, another method is necessary to quantify the E
field at the location of the cells.
Figure 5.13. Illustration of size comparison of the spiral antenna and the GFF.
Due to the unavailability in our laboratory of a nonperturbing miniature E field probe, it
was not possible to measure the E field distribution at the location of the cells on the GFF,
122
and hence the XFDTD method was used to compute all E field distributions.
5.3 The CPA in the Near-Field Range
A possible way of inducing repeatable and robust non-thermal bioeffects by MW
fields is to expose the cells to high magnitude E fields [3]. As explained in Section 3.2.4,
the current free space exposure system is arranged to expose the cells to a plane wave,
and therefore the CPA is placed in the far field (1.5 meter at 1 GHz) of the horn antenna.
However, since the maximum E field magnitude possible in the far field of the current
exposure system does not appear to produce any non-thermal bioeffects, it is reasonable
to take advantage of the higher magnitude E fields in the near field region of the horn
antenna. In this case, the homogeneity of the fields is an important issue since the field
distribution in the near field of a horn antenna is fairly non-uniform and varies with
frequency [4]. Thus it is important to locate the region where the E field is both high and
homogeneous regardless of the operating frequency and to place the GFF in this region.
Preliminary experiments performed to date where the chromaffin cells were
exposed to continuous wave (CW), amplitude modulated and pulse modulated fields in
the frequency 1 – 6 GHz range with an input power of 250 Watts to the horn antenna
have shown no changes in CA release. A possible reason for the lack of an effect may be
because the E field magnitude is not high enough to induce an effect. Moving the CPA to
a location in the near field enabled us to expose the chromaffin cells to E fields that were
3-10 times higher than that in the far field.
The E field and SAR distributions at different frequencies in the frequency range
of interest were computed using XFDTD, and an optimum location in the near field
123
region where the E field magnitude was 3-10 times higher than in the far field and where
the homogeneity was acceptable in the region containing the GFF was successfully
identified.
5.3.1 E Field Distribution in the Near Field in the Absence of the CPA
Due to reflections and scattering from the metallic edges of the horn antenna
which can cause phase cancellations and reinforcement of the fields, the near field
distribution tends to be more non-uniform than the far field distribution (Figure 5.14).
Figure 5.14. Computed near field distribution at the plane of the GFF. Frequency = 3.5
GHz.
The E field along the line indicated by the red arrow in Figure 5.15.a in the absence of the
GFF at different frequencies in the range 1 – 6 GHz was computed using XFDTD and the
results are shown in Figure 5.15.b. It is observed that, as expected, the E field varies in a
complex manner at all frequencies close to the horn antenna, i.e. in the near field, but
beyond a certain distance the variations disappear. The graph also shows that the E field
distribution depends on the operating frequency and the distance from the antenna in the
near field region. Even though a large region near the horn antenna did not have an
acceptable level of E field homogeneity, the E field distributions computed for the entire
124
frequency range showed that there does exist a region of dimensions of the order of the
diameter of the GFF where the level of homogeneity is acceptable regardless of
frequency (Figure 5.16, circled in white).
(a)
(b)
Figure 5.15. Photograph of the horn antenna (a) with a red line indicating where the E
field is calculated; (b) magnitude of the E field along the red line shown in (a).
125
Figure 5.16. Exploded view of Figure 3.1 that shows the region where the E field
magnitude is high and homogeneous regardless of the operating frequency.
5.3.2 E Field Distributions on the GFF in the Near Field of the Horn
Antenna
Detailed E field distributions in the near and far fields in the region of the GFF for
the lowest frequency (1 GHz), center frequency (3.5 GHz) and highest frequency (6
GHz) in the range of interest with an input power of 250 W to the horn antenna were
computed using XFDTD and are shown in Figure 5.17. Here, the goal is to obtain the
variations in E field over the GFF to within 30%, which has been suggested as a
reasonable homogeneity to aim for [5, 6]. Homogeneity is shown as the percentage of the
area of the GFF homogeneous to within 30 % (Figure 5.17). A comparison of the E field
distributions in the near and far fields showed that the homogeneities of the E field were
significantly increased by placing the CPA in the near field region.
126
Figure 5.17. Comparison of the E field distributions in the near and far fields at 1, 3.5 and
6 GHz. The distributions were calculated using a program shown in Appendix E.
The cause of these improvements to the homogeneity of the E field distributions
in the near field is shown in Figure 5.18. In the near field region, the diffracted (green)
and reflected (blue) waves are not negligible, hence they reach the GFF region from
various directions, as illustrated in Figure 5.18, to the locations where the E field would
have been low due to the attenuation by the BSS solution and the reflection at the edge of
the GFF forming a standing wave (Section 3.4.3). This significantly improves the
homogeneity in the region containing the cells.
127
.
Figure 5.18. The cause of the improvement to the homogeneity of the E field. The red,
green and blue rays represent the direct, diffracted and reflected waves, respectively.
5.3.3 Magnitude of the E Field on the GFF placed in the Near Field
The main goal of placing the CPA in the near field region of the horn antenna is
to obtain a greater magnitude of the E field at the location of the cells than was possible
in the far field. Using the numerical calculations of the E field distributions shown in
Figure 5.17, the minimum, mean and maximum magnitude of the E field for the lowest
frequency (1 GHz), center frequency (3.5 GHz) and highest frequency (6 GHz) of interest
were found and are presented in Table 5.1 together with the corresponding values for the
far field. It was observed that a maximum 10 fold increase was obtained in the mean
value of the E field magnitude hence confirming that the magnitude of the E field
increased significantly in the near field.
128
Table 5.1. The calculated E field magnitudes on the GFF for the near and far field.
5.4 Use of the Smaller CPA
Several measurements and computations suggested that a smaller CPA with a
GFF diameter of 10 mm performs better than the original CPA with a GFF diameter of 24
mm, in terms of homogeneity of temperature and E field distributions. The CPA was
replaced with the smaller CPA and used for all experiments. In this section, the reasons
for switching to the smaller CPA are described.
5.4.1 Temperature Gradient
Temperature control profile (Chapter 6) shows that the inlet temperature is
maintained fairly constant to within 36.5 ± 0.15 oC. The measuring point, however, is
restricted to the center of the GFF as shown in Figure 4.10. Figure 5.19 is a photograph of
the cells on the GFF with the dye neutral red and shows that the cells are uniformly
distributed over the entire GFF, hence it is necessary to ensure that the cells are
129
maintained at physiological temperature not only at the center but over the entire region
of the GFF.
Figure 5.19. Photograph of the cells on the GFF showing that the cells are uniformly
distributed over the entire GFF. For visualization, the cells were stained with the dye
neutral red.
Since it is practically impossible with our current set-up to insert another non-perturbing
temperature probe inside the CPA, it is thus not possible to measure the temperature
distribution in detail over the entire region of GFF during experiments. Hence an
experiment was performed using a modified CPA that allows insertion of a needle type
temperature probe (Omega Model HYPO-33-1-T-60-SMPW-M) so that it makes contact
with the GFF during perfusion of the BSS. As shown in Figure 5.20, a temperature
gradient exists in the region of the GFF where the cells are immobilized. Actual
measurement of the temperature at different locations on the GFF (Figure 5.20 (a))
showed a 5.8 ℃ difference from the center to the edge of the GFF. Thermal modeling
(Figure 5.20 (b)) performed using the commercially available software package
COSMOSFloWorks (SolidWorks Corp.) confirmed the existence of this temperature
130
gradient [7]. Subsequent thermal modeling showed that the temperature gradient is due to
the heat transferred from the BSS to the CPA via conduction and the tendency of the BSS
to flow faster through the central region of the GFF.
(a)
(b)
Figure 5.20. Measured (a) and calculated (b) temperature distribution on the 24 mm
diameter GFF
5.4.2 Improvements to Decrease the Temperature Gradient
To decrease the temperature gradient, the CPA was replaced with one having
smaller dimensions that incorporates a GFF of diameter 10 mm. Both the measured and
calculated temperature distributions on the GFF (Figures. 5.21 (a) and (b)) show that the
temperature difference between the center and the edge of the GFF is decreased from
5.8 ℃ to 2.5 ℃ and that between the inlet and outlet BSS is decreased from 4 ℃ to 1 ℃.
5.4.3 Distributions and Magnitudes of the E Field on the GFF
The distribution of the E field on the GFF in the smaller CPA was computed
131
using XFDTD at 1, 3.5 and 6 GHz and compared with the E field distribution for the
original CPA (Figure 5.22).
Figure 5.21. Measured (a) and calculated (b) temperature distribution on the GFF of
diameter 10 mm.
Unlike in the case of the E field distributions for the original larger CPA, there is no
region of low E field on the GFF within the smaller CPA because now the diameter of the
GFF is small compared to the wavelength and hence does not support the formation of a
standing wave over the entire frequency range of 1 – 6 GHz. As a result, the homogeneity
of the E field over the region containing the cells is to within 30 % for most of the
frequency range of interest.
It is also observed that, unlike in the case of the original larger CPA, the E field
magnitude for the smaller CPA (Table 5.2) increases with frequency. This is because the
132
Figure 5.22. Comparison of the E field distributions when the original larger CPA is
placed in the far and near fields with the E field distribution when the smaller CPA is
placed in the near field at 1, 3.5 and 6 GHz. The distributions were calculated using a
program shown in Appendix E.
133
diameter of the GFF in the smaller CPA is comparable to the wavelength at the highest
frequency in the band of interest (6 GHz). Thus the magnitude of the E field is maximum
at the highest frequency and decreases as frequency decreases.
Table 5.2. Comparisons of the mean, maximum and minimum values of the E field
magnitude on the GFF for the original larger CPA placed in the far and near fields with
the values for the smaller CPA placed in the near field of the horn antenna.
5.5 Actual Power
It should be noted that the E field distributions calculated by XFDTD so far are
for an input power of 250 W to the horn antenna which is the maximum power rating of
the MW amplifier. The actual power delivered to the horn antenna is not 250 W, since
there exists cable losses and also the fact that the maximum power output of the MW
amplifier varies with operating frequency. Hence, the power computed using XFDTD in
Table 5.2 should be corrected based on the actual maximum power output of the MW
amplifier and the cable loss as discussed below.
5.5.1 Power Cable and Coupler
134
As shown in Figure 4.1, a high power cable (Model 800108-36 Instruments for
Industry) connected between the directional coupler and the power amplifier delivers the
power generated by the MW amplifier to the horn antenna. During this transmission, the
power from the amplifier is attenuated by intrinsic losses in the cable. In order to
minimize cable loss, a minimum possible length of the cable (2 feet) is selected. The
cable loss measured with the couplers (1-4 GHz and 4-6 GHz) attached to one end of the
cable is plotted in Figure 5.23, and the maximum measured loss was approximately 0.5
dB at 5.5 GHz.
Figure 5.23. Measured attenuation due to the high power cable with the coupler attached
to one end.
The maximum power of the amplifier is defined as the power level slightly
smaller than the level where the amplifier reaches saturation and switches to the standby
135
mode while the input power is maintained lower than the maximum input power of the
amplifier (0 dBm). The maximum power of the amplifier in CW mode was measured at
various frequencies and shown in the Table 5.3. E field magnitudes on the GFF (shown in
Table 5.2) are adjusted based on the cable loss shown in Figure 5.23 and maximum
power shown in Table 5.3 and are tabulated in Table 5.4 (rectangle in red).
Table 5.3. Maximum power output of the amplifier at different frequencies for CW.
136
Table 5.4. Adjusted E field magnitudes (rectangle in red) based on the maximum power
and the cable loss.
137
5.6 References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
J. Yoon, I. Chatterjee, D. McPherson, and G. L. Craviso, "Design,
Characterization and Optimization of a Broadband Mini Exposure Chamber for
Studying Catecholamine Release from Chromaffin Cells Exposed to Microwave
Radiation: Finite-Difference Time-Domain Technique," IEEE Transactions On
Plasma Science, vol. 34, pp. 1455-1469, 2006.
Y. Xu and R. G. Bosisio, "An Effective Approach for Study of Multiple
Discontinuities of Transmission Lines," IEEE Transactions on Microwave Theory
and Techniques, vol. 43, pp. 2585-2589, 1995.
J. F. Kolb, S. Kono, and K. H. Schoenbach, "Nanoscend Pulsed Electric Field
Generators for the Study of Subcellular Effects," Bioelectromagnetics, vol. 27, pp.
172-187, 2006.
W. L. Stutzman and G. A. Thiele, Antenna Theory and Design, 2nd ed. New
York: John Wiley & Sons, Inc., 1998.
N. Kuster and F. Schönborn, "Recommended minimal requirements and
development guidelines for exposure setups of bio-experiments addressing the
health risk concern of wireless communications," Bioelectromagnetics, vol. 21, pp.
508-514, 2000.
F. Schönborn, K. Poković, A. M.Wobus, and N. Kuster, "Design, optimization,
realization, and analysis of an in vitro system for the exposure of embryonic stem
cells at 1.71GHz," Bioelectromagnetics, vol. 21, 2000.
R. Misra, I. Chatterjee, J. Yoon, D. McPherson, and G. L. Craviso, "Thermal
Modeling of a Free Space Exposure System for On-Line Monitoring of
Neurotransmitter Release from Cells Exposed to Microwave Fields,"
Bioelectromagnetics 29th annual meeting, Kanazawa, Japan, 2007.
138
Chapter 6
Results and Discussion
6.1 Introduction
With the design, characterization, and optimization of the broadband MEC
completed and the exposure system tested, a series of experiments were performed to
determine the MW exposure parameters that could possibly induce a bioeffect (i.e., an
alteration in CA release) in chromaffin cells stimulated with the nicotinic receptor agonist
DMPP.
The first step in the experimental part of this research was to perform several
control experiments. Experiments with different MW exposure parameters were then
performed, and the results were analyzed to assess whether CA release was affected by
MW fields. Each of these steps are described in this chapter.
6.2 Statistical Analysis Overview
The goal of this research was to identify MW exposure parameters that could
affect DMPP stimulated release of CA from chromaffin cells. Thus, changes in CA
release observed during the application of MW fields constitute the “bioeffect” being
assessed, and were reflected by changes in the area of CA peaks as measured by the ECD.
Experimental results showed that some CA peaks were significantly changed in the
139
presence of MW fields and others were only slightly or not changed. Hence it was critical
to employ a statistical technique that would allow identification of a bioeffect.
The technique used was to calculate a trend line based on the well known method
of non-linear regression curve fitting. Control experimental data (obtained in the absence
of MW exposure) provided the characteristics of the trend line and the variation of the
CA peaks. Thus, if a specific CA peak in a MW exposure experiment was “significantly
different” from the trend line (Section 6.4.1), that difference was considered to represent
a bioeffect. Here, “significantly different” was defined as the CA peak being outside the
statistically estimated value, called a prediction band.
The next section describes the analysis of control experimental results used to
generate trend lines.
6.3 Analysis of Data from Control Experiments
Control experiments utilized the same methodology as MW exposure
experiments, only without the presence of a MW field (Appendix C). In general, repeated
applications of DMPP resulted in a progressive decrease in CA release from bovine
adrenal chromaffin cells, even though the same amount of DMPP was injected at regular
intervals. At some point in time, the areas of the CA peaks eventually stabilized. The
trends for this type of CA release profile were represented by mathematical expressions
that were used in further analysis, as described in the following sections.
140
6.3.1 Method
The protocol used for the preparation of chromaffin cells is described in
Appendix A and the protocol for conducting a control experiment is described in
Appendixes B and C. Briefly, approximately 1 million cells in medium are transferred
from a 60 mm Petri dish to a glass tube, centrifuged for 10 minutes at room temperature,
and then resuspended in BSS and immobilized on the GFF. The immobilized cells are
perfused with BSS preheated to 36.5 oC, and basal CA release and that stimulated by
DMPP are monitored by the ECD. DMPP (10 µl of a 75 µM solution) is injected every
10 minutes; at the end of the experiment, 20 µl of DMPP is injected to ensure that the
cells are not desensitized and thus able to release twice the amount of CA in response to
the larger stimulus.
6.3.2 Typical Experimental Profiles
Figure 6.1 shows a typical profile for a control experiment performed with cells
that had been maintained in culture for 7 days. The first peak represents CA release
stimulated in response to the sudden fluctuation in pressure and flow rate while the CPA
with the cells is being connected to the flow-through system [1] (Sections 4.5.3.b). The
second peak represents the first response to DMPP. It is smaller than successive peaks
because the injection syringe requires one or more injections to overcome the pressure
built up in the main tubing (described in Section 4.3). As DMPP continues to be injected,
the overall areas under the peaks gradually decrease, which agrees with results in the
141
literature [1]. The last peak, which is larger, represents the response of the cells to twice
the amount of DMPP.
Figure 6.1. A typical CA release profile for a control experiment.
Figure 6.2 is a bar graph that shows the area under the CA peaks for the experiment
shown in Figure 6.1.
The results of all the control experiments carried out using cells from different
cell preparations and at different times in culture were normalized with respect to the
142
Figure 6.2. Bar graph showing the area under the CA peaks for the experiment shown in
Figure 6.1.
maximum peak (typically the second or third peak) and plotted in a single graph for
comparison (Figure 6.3). In all experiments, the peaks gradually decreased in magnitude
but with slightly different decay rates. The time at which the peaks became constant also
varied.
143
Figure 6.3. Normalized results of all control experiments performed on cells from
different cell preparations and at different lengths of time in culture.
In Figure 6.4, the average of the peaks from the control experiments were plotted in the
form of a bar graph (Figure 6.4) to show the overall variation in the decay pattern.
6.3.3 Mathematical Model for the Decline in the Area under the Peak
Among many possible models, the pattern of decline shown in Figure 6.4 fits the
exponential decay model, expressed as
y = a ⋅ e −bx + c
(6.1)
144
Figure 6.4. Average of peaks ± S.D. The numbers on the graph indicate the number of
peaks used to calculate the average.
where a, b, and c are constants to be determined. To prove that the pattern follows
equation 6.1, a non-linear regression method [2] using SigmaPlot was applied and an
example is shown in Figure 6.5. The areas under the peaks are plotted (bars) with the
fitted curve or trend line (curve in red) calculated using SigmaPlot. The constants shown
in equation 6.1 that were calculated by SigmaPlot were a = 2214.4, b = 0.0982 and c =
485.9 for this specific experiment and hence the equation for the trend line is given by
y = 2214.4 ⋅ e −0.0982 x + 485.9
(6.2)
145
Figure 6.5. Areas under the CA peaks for a control experiment, superimposed with a
trend line, as calculated by SigmaPlot.
In order to prove that the decay pattern of a control experiment always follows equation
6.1, all of the experimental data points from 8 separate control experiments were fitted to
the equation by the non-linear regression method in SigmaPlot and superimposed with
their corresponding trend lines (Figure 6.6). The non-linear regression curve fitting
method in SigmaPlot reports whether the data used are successfully fitted to the given
equation or not. The results show that for each control experiment, the peaks were
successfully fitted to equation 6.1, indicating that the pattern of decline of the peaks in all
of the control experiments indeed follows the exponential decay model. It should be
noted that intrinsic variations in the response of the cells exist in all control experiments
146
since no trend line perfectly matches all the peaks in a particular experiment. This
variation is natural and thus expected in any biological system. During MW experiment,
peaks greater or less than the variation of the corresponding control experiment would
therefore be considered indicative of a bioeffect. Hence, proper assessment of the
intrinsic variation that is present within a given control experiment is extremely important
to pinpoint the existence of a bioeffect. The procedure used to identify this variability is
described in the next section.
(a)
(b)
(c)
(d)
147
(e)
(f)
(g)
(h)
Figure 6.6. Control experiments performed on different dates showing the trend lines
calculated using SigmaPlot.
6.3.4 Assessment of the Range of Variation in a Control Experiment
As discussed in the previous section, variations in the response of the cells to
successive applications of DMPP needs to be mathematically represented to make proper
comparisons with the results of MW exposure experiments and hence allow identification
of a bioeffect. In statistics, such a variation is called the variability of samples, and the
range of the variability can be estimated by a statistical method called “prediction
interval”. This method provides an interval or range within which the next single
148
observation from the given population will be contained [3, 4]. When it is applied for the
analysis of MW exposed cells, it provides the interval within which CA peaks are
expected to fall based on the corresponding control experiment, if there is no bioeffect.
Thus, if any of the peaks fall outside this interval, this would be indicative of a bioeffect.
The prediction interval is mathematically expressed as
y ± t (α / 2; n − 1) ⋅ s ⋅ 1 +
1
n
(6.3)
where y is mean of the data, α is the critical value (0.95), s is standard deviation of the
data, n is number of data points (CA peaks), and t (α / 2; n − 1) is a value found in the
well known t distribution table [3, 4]. The band formed by the prediction interval for all x
values is called a prediction band and an example of a prediction band generated for a
control experiment using SigmaPlot is shown in Figure 6.7.
Figure 6.7. Data from an actual experiment depicting the CA peaks, the calculated trend
line and the 95 % prediction band.
149
Hence, if CA peaks obtained during exposure to MW fields are outside the 95 %
prediction band of the corresponding control experiment (from the same cell preparation),
these peaks would be either larger or smaller than the intrinsic variation, and thus indicate
a bioeffect.
6.4 Analysis of Data from Microwave Exposure Experiments
Several experiments have been performed with chromaffin cells exposed to MW
fields using the protocol presented in Appendixes B and C. In this section, the statistical
analysis used to identify CA peaks representing bioeffects due to MW exposure of the
cells is explained, the results of the statistical analysis are summarized and the MW
exposure parameters that induce the most robust bioeffects are identified.
6.4.1 Statistical Analysis of Results
As explained in the previous section, CA peaks naturally decay over the course
of an experiment and the trend line that fits the peaks can be calculated using curve fitting
(non-linear regression). The same method was adapted for analyzing all MW exposure
experiments.
The experimental profile of a representative MW exposure experiment is shown
in Figure 6.8. It is seen that most of the peaks decrease exponentially as in the case of the
control experiments (Section 6.3.3); exceptions are the 13th, 14th, 16th, 17th 19th 20th and
21st peaks. These peaks are either higher or lower than their neighboring peaks. Since any
changes in the CPA inlet temperature, CPA outlet temperature, pressure and flow rate,
150
which may affect the DMPP response (Sections 4.2 and 4.5), are within acceptable limits
during the entire experiment, the peaks indicated by the red and blue arrows could
represent a bioeffect due to the MW field.
Figure 6.8. Experimental profile of a representative MW exposure experiment. Red
arrows indicate a decrease and blue arrows indicate an increase relative to neighboring
peaks.
151
The remainder of this section describes the trend line analysis used as a first step
to identify bioeffects.
The trend line for the exponential decay of peaks 4 to 21 as shown in Figure 6.9
was calculated using the curve fitting (non-linear regression) method in SigmaPlot. As a
result, this trend line can be considered to represent expected values of CA peaks when
no MW field is present.
Figure 6.9. The areas under the CA peaks and the trend line for the MW exposure
experiment shown in Figure 6.8. The orange arrows indicate peaks lower than the trend
line and the green arrows indicate peaks higher than the trend line.
As discussed in Section 6.3.3, CA peaks obtained in the MW exposure
experiments that are greater or less than the natural variation of the corresponding control
experiment are considered to be indicative of a bioeffect. Hence, in order to determine the
152
existence of a bioeffect in a MW exposure experiment, the 95 % prediction band of the
corresponding control experiment (Figure 6.7) can be superimposed on the CA peaks of a
MW exposure experiment (such as Figure 6.9) and hence any CA peaks that are outside
the prediction band can be identified. The problem encountered in this method is that the
two graphs do not completely overlap because the decay ratio of the DMPP response is
different from experiment to experiment (Section 6.3.2). For example, the normalized
trend line and the 95 % prediction band of the control experiment (Figure 6.7)
superimposed on the normalized MW exposure experiment performed on 04/02/2008
(Figure 6.10) shows that the decay ratios of the two trend lines are different, and hence
one cannot directly apply the 95 % prediction band of the control experiment to the MW
exposure experiment. Thus, in order to identify peaks that indicate a bioeffect, it is
usually necessary to first modify the MW exposure trend line to fit the control trend line
so that they overlap (Figure 6.10, dotted gray lines). The entire process is programmed in
Mathcad, and the algorithm is described in Appendix F. Briefly, the percent difference
between the control trend line and the CA peaks for the MW exposure experiment are
calculated using equation 6.4.
% difference( x ) = Normalized Control Trend Line( x )
− Normalized MW CA Peak ( x )
(6.4)
153
Figure 6.10. Normalized trend lines and CA peaks of the MW exposure experiment
performed on 04/02/2008 and the normalized trend line of the corresponding control
experiment. The 95 % prediction band for the control experiment is also shown.
The parameters of the trend line for the MW exposure experiment are multiplied
by the unknown parameters u1, u2 and u3.
Modified Trend Line = a ⋅ u1 ⋅ e −b⋅u2 x + c ⋅ u 3
(6.5)
This modified trend line for the MW exposure experiment (equation 6.5) will overlap
with the corresponding control trend line, and hence the three unknown values in
equation 6.5 can be found by substituting points on the control trend line in equation 6.5
or using the built in function in Mathcad, such as the “solving equation” function. The
new CA peaks for the modified trend line are calculated using equation 6.6 which is
equation 6.4 solved for the normalized CA peak:
154
Normalized MW CA Peak ( x ) = Normalized MW Trend Line( x )
− % difference( x )
(6.6)
The final Mathcad result is shown in Figure 6.11. It is seen that the two trend lines
overlap and there is only one CA peak (circle in red) that is outside the 95 % prediction
band (-6.97 % in Table 1 in Appendix F).
Figure 6.11. The final result of the Mathcad analysis program showing the normalized
control trend line, 95 % prediction band, modified MW exposure trend line, and modified
CA peaks for the MW exposure experiment performed on 04/02/2008.
6.4.2 Summarizing the Results of the MW Exposure Experiments
Figure 6.12 provides an example of how the results of each exposure experiment
were analyzed and summarized. First, the trend line for the CA peaks was obtained
(Figure 6.12.a). As observed for this experiment, many peaks fell far from the trend line.
155
Figure 6.12.b shows the statistical result generated by the Mathcad program (procedure
described in Appendix F). The CA peaks (‘x’ in Figure 6.12.b) outside the 95 %
prediction band (purple lines) are indicative of a bioeffect. Figure 6.12.c summarizes
Figure 6.12.b. The first line of the table (gray box) gives the description of the MW
parameters used in the exposure part of the experiment. In the second line, CA peaks
indicating a bioeffect are marked either ▲ or ▼, where ▲ indicates CA peaks greater
than the upper range of the 95 % prediction band (enhanced response to DMPP) and ▼
indicates CA peaks lower than the lower range of the 95 % prediction band (suppressed
response to DMPP). If there is no symbol, it simply means that the peak is within the
95 % prediction band, and hence no bioeffect is indicated. The numbers in the third line
are the percent differences between the trend line and the CA peak, which is referred to in
this dissertation as the “magnitude of the bioeffect”. A negative number indicates a CA
peak lower than the corresponding trend line (suppressed response to DMPP) and a
(a)
156
(b)
(c)
Figure 6.12. Analysis procedure for the MW exposure experiment performed on
02/12/2008 showing (a) the area under the CA peaks with the trend line, (b) the statistical
result generated by the Mathcad program (Appendix F) and (c) a table summarizing the
results shown in (b).
A summary of the results of 32 MW experiments are given in Figure 6.13 and
Table 6.1. Excluded from this summary are experiments where the data may be
unreliable due to either small CA peaks, suboptimal experimental conditions or sudden
baseline shifts (Appendix I). A detailed analysis of the results is given in the following
sections.
157
158
Figure 6.13. Cellular responses to DMPP and the corresponding trend lines for 32 MW
exposure experiments. The dates on which the experiments were performed are also
shown.
159
Table 6.1. Summary of results for each MW exposure experiment.
160
6.4.3 Assessing the Number of Bioeffects
The total number of bioeffects that were observed are presented in three different
ways: (1) the number of experiments that have at least one bioeffect, (2) the number of
MW exposures that have at least one bioeffect and (3) the number of peaks that are
indicative of a bioeffect. The results are summarized in Table 6.2. For (1), out of a total
number of 32 experiments, 21 experiments (65.6 %) contained at least one peak that is
indicative of a bioeffect. This means that more than half of the experiments performed
using different MW exposure parameters showed a bioeffect. For (2), out of a total
161
number of 81 exposures, 33 exposures (40.7 %) contained at least one peak indicative of
a bioeffect. By analyzing the peaks within the 33 exposures showing a bioeffect, several
patterns of bioeffects (Section 6.4.5) as well as the MW exposure parameters that induce
the most robust bioeffects were identified (Section 6.4.6). Finally, for (3), out of a total
number of 412 CA peaks, 91 peaks were indicative of a bioeffect (22.1 %). From the 91
peaks showing a bioeffect, the magnitudes of the bioeffects were obtained (Section 6.4.4).
Each of these results is discussed in the following sections.
Table 6.2. Summary of the number of bioeffects (data from Table 6.1).
6.4.4 Assessing the Magnitude of the Bioeffects
The magnitude of the bioeffects (for the peaks indicative of a bioeffect in Table
6.2) was assessed as shown in the histogram in Figure 6.14. Negative values represent a
peak that is lower than the corresponding trend line, i.e., a decrease in the response to
DMPP, while positive values represent a peak that is higher than the trend line, i.e., an
enhancement of the response to DMPP. It was found that there were more negative
162
values (72 peaks, 76 %) than positive values (23 peaks, 24 %), which indicates that MW
exposure causes a suppression of the response to DMPP more often than an enhancement.
Figure 6.14. Histogram of the percent difference for the peaks that indicate a bioeffect.
In addition, the overall magnitude of the percent difference for negative
bioeffects (17.82 %) was greater than that for positive bioeffects (10.09 %), meaning that
suppression of the effect of DMPP on CA release from the cells by MW fields was more
pronounced than enhancement of CA release.
6.4.5 Patterns and Magnitudes of the Bioeffects
From Figure 6.6 and Table 6.1, one can observe the range of the patterns and
magnitudes, respectively, of the induced bioeffects. Three main patterns that have
emerged are depicted in Figure 6.15 and are also as follows: the bioeffect occurs (a) only
163
during MW exposure, (b) only after MW exposure, and (c) both during and after MW
exposure. The number of peaks showing these patterns, as well as their magnitude
(percent differences) are tabulated in Table 6.3. Of the 91 peaks to which a bioeffect is
attributed, 14 peaks show a bioeffect during MW exposure (Figure 6.15.a), 4 peaks show
a bioeffect after MW exposure (Figure 6.15.b), and 71 peaks show a bioeffect both
during and after MW exposure (Figure 6.15.c). Hence it can be concluded that, if a
bioeffect is induced by MW exposure, the effect is induced during exposure and in some
cases tends to last even after the MW field is turned off.
(a)
(b)
(c)
Figure 6.15. Illustration of the three main patterns of MW bioeffects. The bioeffect
occurs (a) during, (b) after, and (c) both during and after MW exposure.
The magnitude was greatest (17.13 %) for bioeffects occurring during and after exposure
(Figure 6.15.c) and slightly less (12.58 %) for bioeffects that occurred during exposure
only (Figure 6.15.a).
6.4.6 Identifying MW Exposure Parameters that Induce the Most
Robust Bioeffects
To identify the MW exposure parameters in the 1 to 6 GHz frequency range that
164
Table 6.3. Summary of the patterns and magnitude of the 91 peaks showing a bioeffect.
induce the most robust bioeffects, experiments were performed at fixed frequencies using
a variety of parameters, such as continuous wave (CW), pulse keying, amplitude
modulation (AM), pulsed AM, phase modulation (PM) and Gaussian pulses. In most of
the initial experiments performed, bioeffects were not observed. Exceptions included the
experiment performed on 01/07/2008 that used a pulsed (75 ms pulse width) MW field at
3.5 GHz and the experiment performed on 12/05/2007 that used a PM MW field in which
the operating frequency during exposure changed. In both of these experiments, CA
peaks were depressed during and after MW exposure.
As follow up to these observations, experiments were subsequently performed
using not only pulsed MW fields with different pulse widths and with a broader
frequency band (instead of a single fixed frequency) but also a more complex MW
scheme that actively changed the operating frequency and pulse width during MW
exposure. This latter exposure scheme is called a pulsed frequency sweep (PFS).
165
Table 6.4. Analysis of the results for all MW experiments using CW, AM, pulse keying
and Gaussian pulses.
For experiments that used this latter MW exposure scheme, the PFS with the
largest span width (1 GHz) was chosen so that the entire frequency range of interest
could be covered in five experiments (1 to 2, 2 to 3, 3 to 4, 4 to 5 and 5 to 6 GHz). In all
experiments, the power was set to the maximum value available from the amplifier
(Section 5.5.1) and the sweep rate was set to the minimum possible value (500 ms) in
order to introduce a maximum number of frequencies over a short period of time.
Three different pulse durations, 10 ns, 100 µs and 100 ms, were used for each frequency
sweep and the results are presented in Table 6.5.
166
Table 6.5. Analysis of the results for PFS MW experiments.
The results of these experiments indicate that bioeffects were induced most often
by sweeping the frequency during exposure. Table 6.6 shows that 55 % of exposures with
167
PFS exhibited a bioeffect while only 26 % of exposures with all other MW parameters
showed a bioeffect.
Table 6.6. Summary of the results shown in Table 6.4 and Table 6.5.
When MW fields with PFS in the range 5 to 6 GHz and 100 ms pulse width were applied
to the cells, the areas of the CA peaks significantly increased during exposure, and then
decreased when the field was turned off (Figure 6.16). Thereafter, the CA peaks
gradually increased toward the trend line and leveled off to a value close to the trend line
during the last DMPP injection. There was no significant temperature, pressure, and flow
rate change during the experiment, and hence the change in CA peaks was attributed to
MW field exposure.
A second experiment performed immediately after the experiment whose results
are shown in Figure 6.16 shows a very similar response, with the area of the CA peaks
increasing during MW exposure and then decreasing after MW exposure (Figure 6.17).
Table 6.7 shows that 72 % of the exposures using a 5-6 GHz PFS field elicit a bioeffect,
while only 43 % of the exposures for the other four frequency sweep ranges combined
show bioeffects. These results indicate that bioeffects occur most often for the frequency
range of 5-6 GHz.
168
Figure 6.16. Profile obtained for cells exposed to a 5-6 GHz PFS (sweep time of 0.5 s)
field 5 to 6 GHz showing a robust bioeffect.
Figure 6.17. Profile obtained for a second experiment performed immediately after the
experiment shown in Figure 6.16.
169
In addition, the magnitude of the bioeffect is also greater, i.e., the change in the
area of the CA peaks for a 5-6 GHz PFS fields, is also larger than for PFS exposures in
the other frequency ranges. Among the 5-6 GHz PFS exposures, the one using a 100 ms
pulse shows the greatest change in the area of the CA peaks and hence the largest
bioeffect. (Table 6.8).
Table 6.7. Comparison of the results for 5-6 GHz PFS and those for other frequency
ranges.
6.5 Discussion
As presented in the previous section, several experiments with different MW
exposure parameters were performed and the results were analyzed and summarized. It
was observed that the greatest number of bioeffects were induced when PFS was utilized
with maximum power as the MW exposure method (Table 6.6). Of the PFS exposure
experiments that were carried out, those using the frequency range of 5-6 GHz (Table
6.7) and a pulse width of 100 ms (Table 6.8) showed the most significant bioeffects. A
170
Table 6.8. Comparison of the results for exposures with 10 ns, 100 ms and 100 µs pulse
widths for 5-6 GHz PFS.
detailed explanation of the PFS signal and possible reasons as to why PFS induces the
strongest bioeffects are described in the following sections.
6.5.1 Description of the 5-6 GHz PFS Signal
A complete description of the “5-6 GHz PFS” that was found to induce the
strongest bioeffect is as follows: 5-6 GHz PFS with maximum power (maximum power is
defined in Section 5.5.1), pulse width of 100 ms, repetition rate of 1 to 2 Hz and sweep
time of 500 ms. Although PFS is a pulsed MW signal, the maximum peak power of the
MW field can increase the temperature at the location of the cells if the pulse width of the
171
PFS MW field is too long and applied repeatedly to the cells. Experimentally it was
determined that 100 ms was the maximum pulse width that could be used to prevent an
increase in the temperature of the BSS at the outlet of the CPA of more than 0.2 oC
during PFS exposures. A repetition rate of 1 to 2 Hz was similarly determined by trial
and error to maintain the temperature of the BSS to within 0.2 oC of 36.5 oC during an
experiment. A repetition rate of 1 to 2 Hz was found to meet this need. Because the goal
of our exposure schemes was to introduce maximum changes in frequency and magnitude
of the E field in as short a time as possible, a short sweep time was desirable. The sweep
time used, 500 ms, was the shortest sweep time that the signal generator was capable of
delivering.
6.5.2 Waveform of the PFS MW Fields
Understanding possible mechanisms of the observed bioeffects produced by the
5-6 GHz PFS fields (Section 6.5.5) requires a detailed waveform analysis of the PFS.
The 5-6 GHz PFS exposure scheme is a combination of a frequency sweep and
pulse keying that is controlled by a LabVIEW program (Section 4.6). There exists a
limitation on the frequency shift time, i.e., the minimum time that it takes the signal
generator to change its operating frequency, of 100 ms. The frequency spectrum
measured using a spectrum analyzer during a sweep between 5 and 6 GHz showed that
the signal generator was capable of generating discrete frequencies of 5, 5.2, 5.4, 5.6, 5.8,
and 6 GHz. Thus, if the signal is not pulsed, the waveform of a continuous single sweep
will look like the illustration shown in Figure 6.18. When this frequency swept signal is
172
pulsed with a pulse width of 100 ms and a repetition rate of 1 Hz, the resulting output is
as illustrated in Figure 6.19.
Figure 6.18. Illustration of the 5-6 GHz frequency sweep in steps of 0.2 GHz with a
sweep time of 500 ms.
It is also known that the gain of the amplifier, the gain of the broadband horn
antenna (section 3.4.1) and the loss of the power cable (section 5.5.1) connecting the
amplifier to the horn antenna vary with frequency, hence resulting in a variation in the
magnitude of the E field at the location of the cells over the frequency band being
delivered. The E field intensities at the location of the cells based on XFDTD calculations
and measured power loss at different frequencies is shown in Figure 6.20.
173
Figure 6.19. Illustration of the 5-6 GHz PFS with a sweep time of 500 ms, a pulse width
of 100 ms and a repetition rate of 1 Hz.
Figure 6.20. Illustration of the 5-6 GHz PFS with a sweep time of 500 ms, a pulse width
of 100 ms and a repetition rate of 1 Hz at the location of the cells. The magnitudes of the
wave at different frequencies vary since the gain of the amplifier, the gain of the
broadband horn antenna and the loss of the power cable vary at different frequencies. The
magnitudes were found using results obtained by XFDTD and the measured cable loss.
6.5.3 Effect of Frequency and Power Windows
Researches have reported the existence of “frequency and power windows”, that
is, a specific range of frequency and power intensity that produce the greatest effect [5-
174
11]. The fact that the bioeffects observed in this study occurred most often and more
robust in a certain frequency range (5-6 GHz) indicates that one or more parameters of
the applied 5-6 GHz PFS (Figure 6.20) may be within a frequency and power window for
producing effects on the particular biological system used for these exposures.
6.5.4 Threshold Value of the E field
Another possible reason why bioeffects have been found to occur most often in
the at 5 to 6 GHz range rather than at 1 to 5 GHz is that the wavelength at the higher
frequencies (7.07-8.49 mm at 5-6 GHz) becomes comparable to the diameter of the GFF
(10 mm) inside the CPA, increasing the magnitude of the E field at the location of the
cells (Table 6.9). This in turn could mean that the magnitude of the E field now exceeds a
threshold value required to induce bioeffects reproducibly (Section 2.3.4). The fact that
PFS MW exposure experiments at the other frequencies (1 – 5 GHz) also sometimes
induced a bioeffect (Table 6.7) may be because the magnitude of the E fields at the
location of the cells was close to the threshold value.
175
Table 6.9. Magnitude of the E field at different frequencies. The magnitude is increased
as frequency is increased except at 2 GHz due to the large reflections at this frequency
(Figure 5.8).
6.6 Conclusion
The goal of this research was to identify MW exposure parameters that affect the
release of CA from chromaffin cells using MW fields within the 1-6 GHz frequency
range. In order to increase the likelihood of achieving this goal, MW exposure parameters
were actively changed during a single exposure period to generate complex E field
patterns. Hence, a PFS scheme was implemented and this exposure paradigm enabled us
to successfully identify MW parameters that induce robust bioeffects (Section 6.4.6). The
mechanism that elicits the bioeffects is not clear. However, it is possibly that the bioeffect
observed are related to frequency and power window effects (Section 6.5.4) or a
threshold value of the E field (Section 6.5.3). Taken as whole, these findings suggest that
MW exposure parameters that induce bioeffects can be identified using novel approaches,
such as delivering MW fields that produce complex E field patterns.
176
6.7 References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
D. J. Green and R. L. Perlman, "On-Line Measurement of Catecholamine
Secretion," Analytical Biochemistry, pp. 270-276, 1981.
W. W. Hines, D. C. Montgomery, D. M. Goldsman, and C. M. Borror,
Probability and Statistics in Engineering, 4th ed. Hoboken, NJ: John Wiley &
Sons, Inc., 2002.
P. R. Nelson, M. Coffin, and K. A. F. Copeland, Introductory Statistics for
Engineering Experimentation. San Diego, California: Academic Press, 2003.
H. M. Wadsworth, Handbook of Statistical Methods for Engineers and Scientists.
New York, NY: McGraw-Hill, 1990.
W. Adey, "Frequency and power windowing in tissue interactions with weak
electromagnetic fields," Proceedings of the IEEE, vol. 68, pp. 119-125, 1980.
M. Zhadin and F. Barnes, "Frequency and Amplitude Windows in the Combined
Action of DC and Low Frequency AC Magnetic Fields on IonThermal Motion in
a Macromolecule:Theoretical Analysis," Bioelectromagnetics, vol. 26, pp. 323330, 2005.
S. Banik, S. Bandyopadhyay, and S. Ganguly, "Bioeffects of microwave–a brief
review," Bioresource Technology, vol. 87, pp. 155–159, 2003.
E. Postow and M. L. Swicord, Modulated fields and “window” effects.
S. Lin-Liu and W. R. Adey, "Low frequency amplitude modulated microwave
fields change calcium efflux rates from synaptosomes," Bioelectromagnetics, vol.
3, pp. 309-322, 1982.
K. R. Foster and M. H. Repacholi, "Biological effects of radiofrequency fields:
Does modulation matter?" Radiation Research, vol. 162, pp. 219-225, 2004.
H. Hinrikus, J. Lass, and V. Tuulik, "Low-level Microwave Effects on EEG,"
presented at 22nd Annual EMBS International Conference, Chicago IL, July 23-28,
2000.
F. S. Barnes, "Cell membrane temperature rate sensitivity predicted from the
Nernst equation," Bioelectromagnetics, vol. 5, pp. 113-115, 1984.
K. R. Foster, "Thermal and Nonthermal Mechanisms of Interaction of RadioFrequency Energy with Biological Systems," IEEE Transactions on Plasma
Science, vol. 28, pp. 15-23, 2000.
M. H. Repacholi, "Low-Level Exposure to Radiofrequency Electromagnetic
Fields: Health Effects and Research Needs," Bioelectromagnetics, vol. 19, pp. 1-9,
1998.
J. L. Kiel, J. E. Parker, P. J. Morales, J. L. Alls, Patrick A. Mason, R. L. Seaman,
S. P. Mathur, and E. A. Holwitt, "Pulsed Microwave Induced Bioeffects," IEEE
TRANSACTIONS ON PLASMA SCIENCE, vol. 28, pp. 161-167, 2000.
F. Bouchet and Y. Boulard, "Ultrastructural changes following treatment with a
microwave pulse in the oocyst of Eimeria magna Perard, 1925," Parasitology
research, vol. 77, pp. 585-589, 1991.
M. S. Markov, "Pulsed electromagnetic field therapy history, state of the art and
future," Environmentalist, vol. 27, pp. 465-475.
177
Chapter 7
Future Work
7.1 Introduction
Designing, building, optimizing and testing the free space exposure system was
completed and the system used in systematic experiments to identify the specific MW
parameters (PFS, 5-6 GHz) that appear to induce non-thermal effects on CA release from
chromaffin cells. As a follow-up to these studies, the design of another exposure system
that can produce much larger E fields at the location of the cells has been initiated. This
new exposure system could possibly enable us to identify an E field threshold value that
causes changes in CA release.
7.2 Cell Holder and Antenna
As described in Section 5.4, the smaller CPA that incorporates a GFF of diameter
10 mm provides excellent homogeneity in both temperature and E field distributions in
the region containing the cells. Hence, it was decided to use the smaller GFF as the cell
chamber for the new exposure system. The design also incorporates a Vivaldi antenna to
produce the MW field since this antenna is broadband and easy to fabricate via
commonly used techniques for printed circuit boards (PCB).
7.3 Limitations of the Free Space Exposure System and Solutions
178
In the free space exposure system, all the power radiated from the horn antenna is
not efficiently delivered to the location of the cells due to the small size of the GFF and
spreading of the fields. Thus, E fields are only of the order of 1-2 kV/m. This limitation is
overcome in the design of the new exposure system.
7.3.1 Large Aperture Size of the Horn Antenna
The computed and measured radiation patterns of the horn antenna at
3.5
GHz shown in Figures 3.9 (a) and (b) indicate that the horn antenna has a fairly broad
beamwidth both in the H and E planes. This was confirmed by the E field distribution
(Figure 7.1.b) measured at 3 m from the horn antenna (yellow plane shown in Figure 7.1.
a) at 3.5 GHz using a spiral antenna (Figure 3.3) and a computer controlled turntable.
Figure 7.1. Horn antenna. (a) Photograph showing the plane where the magnitude of the
E field was measured at 3.5 GHz and (b) the measured E field distribution over the plane
shown in (a).
179
The E field was uniformly distributed over a large area instead of being focused in the
central region in front of the horn antenna where the cells are placed.
The dimensions of the aperture of the horn antenna used in the free space
exposure system are 24 x 14 cm while the dimension of the GFF exposed to the MW
fields is only 1 cm (Figure 7.2.a). Hence, only a small part of the MW field actually
propagates to the location of the GFF (Figure 7.2.a. line in blue), the rest being radiated
to regions not containing the GFF (Figure 7.2.a, lines in red). This means that this
particular antenna is not desirable for use in an exposure system that requires a large E
field over a small target region. Hence, a method was sought to “focus” the E field onto
the region containing the cells.
Embedding a broadband radiating element (that has a flare structure similar to
that characteristic of the horn antenna; Figure 7.2.b) in a high dielectric constant substrate
(Figures 7.2.c) solves this problem. In this way, the dimensions of the flare structure can
be much smaller than in free space and hence radiate more of the field to the location of
the GFF (dotted line in Figure 7.2.c). A co-planar type of antenna based on this flare
structure, called a Vivaldi antenna, is the one utilized in the new exposure system.
7.3.2 Reflections from the Edges of the GFF
As the BSS flows through the CPA for perfusing the chromaffin cells, the
components inserted into the CPA, such as the nylon meshes and the GFF are soaked
with the BSS (Figure 3.1). Hence, during exposure, the MW fields are first incident on
180
(a)
(b)
(c)
Figure 7.2. Photograph of the horn antenna (a) together with a depiction of the relative
size of the GFF. (b) Illustration of the flare structure in the horn antenna in free space
showing that only a small amount of the MW field (dotted line in red) propagates toward
the GFF. (c) Illustration of the flare structure embedded into a high dielectric substrate
showing that a larger amount of the MW field (dotted line in red) propagates toward the
GFF.
the CPA, and then pass through the GFF soaked with the BSS (Figure 7.3, line in red).
The CPA is made of polypropylene whose relative dielectric constant is 2.3. Because of
the low relative dielectric constant difference, reflections at the dielectric interface
between the air and the GFF are negligible. However, reflections at the dielectric
interface between the CPA and the BSS (Figure 7.3, lines in blue) are significant due to
the high relative dielectric constant of the GFF soaked with the BSS, which was
measured to be ~50 over the entire1 to 6 GHz frequency range [1]. Because of the large
difference in relative dielectric permittivity of the two materials, about 80 % of the E
fields propagating perpendicular to the GFF are reflected.
The reflections at the interface between the GFF and the CPA can only be
reduced by matching the relative dielectric permittivity of the two materials. In the new
exposure system, dielectric matching is accomplished by embedding the GFF and Vivaldi
181
antenna described in Section 7.3.1 in a high dielectric constant substrate material.
Figure 7.3. Illustration of a cross-sectional view of the CPA with the incident and
reflected MW fields at the various dielectric interfaces.
7.4 The Newly Designed Exposure System
Based on the discussion in Section 7.2, a broadband Vivaldi antenna has been
designed for the frequency range 1 – 6 GHz (Figure 7.4.a) [2-7], and embedded in a 2
mm thick dielectric substrate material with a relative dielectric constant of 20 (Model CSTOCK AK, Cuming Microwave Corp.) (Figure 7.4.b). The GFF is placed at the open
end of the Vivaldi antenna where a part of the substrate material is removed to form a
holder (1 mm in depth and 10 mm in diameter) for the GFF (Figure 7.4.c).
182
(a)
(b)
(c)
Figure 7.4. SolidWorks model of (a) the Vivaldi antenna and dielectric substrate, (b) the
Vivaldi antenna embedded in the dielectric substrate and (c) a GFF located in front of the
Vivaldi antenna.
A detailed geometric model (Figure 7.5) of the entire system was constructed
using SolidWorks. The antenna is embedded in the dielectric substrate and the bottom
portion of the CPA with the GFF is screwed into a threaded block glued at the bottom of
the dielectric material. Inlet and outlet ports with temperature probes are inserted into the
upper and bottom block, respectively. The geometry of the entire exposure system was
imported into XFDTD and a snapshot of the propagating E field at the plane where the
cells would be immobilized was calculated (Figure 7.6). It was observed that there exists
a region where the MW fields are high as well as fairly homogeneous (circled in blue).
This region was chosen for the location of the GFF.
183
Figure 7.5. SolidWorks drawing of the entire exposure system.
Figure 7. 6. A calculated snapshot of the E field propagating in the plane of the GFF
computed using XFDTD (Frequency = 3.5 GHz).
The magnitude of the E field over the GFF for the lowest frequency (1 GHz),
center frequency (3.5 GHz) and highest frequency (6 GHz) in the frequency range of
184
interest, with an input power 250 W, was computed using XFDTD and is presented in
Table 7.1. The calculations show that at the location of the GFF there is about a 20 to 40
fold increase over the maximum E field obtained in the free space exposure system.
Table 7.1. Comparison of the magnitude of the E field on the GFF for the free space
exposure system and the newly designed system.
Detailed E field distributions over the GFF in the newly designed exposure
system for the lowest frequency (1 GHz), center frequency (3.5 GHz) and highest
frequency (6 GHz) in the frequency range of interest computed using XFDTD are shown
in Figure 7. 7. The results show that the homogeneity of the E field distributions in the
new exposure system are acceptable.
7.5 Temperature Control
Since the magnitude of the MW field is very high at the location of the cells, it is
possible that heating by the MW field can exceed the capability of the dynamic
185
Figure 7.7. The E field distributions in the region of the GFF embedded in the newly
designed exposure system at 1, 3.5 and 6 GHz. % refers to the area of the GFF
homogenous to within 30 % (% refers to the area of the GFF homogeneous to within
30 %).
temperature controller to maintain the temperature of the cells to within 0.2 oC of the set
point of the BSS. Hence, an additional system for controlling temperature has to be
implemented at the upper block of the exposure system. As shown in Figure 7.8, the
upper block is compressed against the dielectric substrate via a silicone gasket and this
provides an empty space for coolant to flow on the surface of the substrate. If inlet and
outlet ports are drilled in the upper block, this will allow a fast stream of coolant to cool
the surface of the substrate material immediately above the GFF during MW field
exposure and thus help to maintain the temperature to within ±0.2 oC of 36.5 oC at the
location of the cells.
186
Figure 7.8. Exploded view of the upper block of the exposure system where a cooling
system is implemented. The coolant will flow through the surface of the substrate
material to control the temperature at the location where the cells are immobilized during
MW exposure.
In conclusion, a newly designed system will be built, characterized and used to
obtain more repeatable and robust bioeffects using the experimentally identified MW
exposure parameters (Section 6.5.5) that were formed to produce changes in CA release.
This system will provide valuable information, such as the existence of a threshold field
strength and also allow better assessment of the presence of bioeffects at the other
frequencies (1-5 GHz) since E fields at all the frequencies would be similar (Table 7.1).
7.6 References
187
[1]
[2]
[3]
[4]
[5]
[6]
[7]
T. Hagan, I. Chatterjee, D. McPherson, and G. L. Craviso, "A novel waveguidebased radio frequency/microwave exposure system for studying nonthermal
effects on neurotransmitter release-Finite-Difference Time-Domain modeling,"
IEEE Trans. Plasma Science, vol. 32, pp. 1668-1676, Aug. 2004.
A. Sutinjo and E. Tung, "The Design of a Dual Polarized Vivaldi Array,"
Microwave Journal, 2004.
M. Chiappe and G. L. Gragnani, "Theoretical and Numerical Analysis of the SelfScaling Properties of the Exponentially Tapered Slot-Line Antenna," Microwave
And Optical Technology Letters, vol. 45, pp. 485-491, 2004.
M. Chiappe and G. L. Gragnani, "Vivaldi Antennas for Microwave Imaging:
Theoretical Analysis and Design Considerations," Transactions on
Instrumentation and Measurement, vol. 55, pp. 1885-1891, 2006.
W. Grammer and K. S. Yngvcsson, "Coplanar Waveguide Transitions to Slotline:
Design and Microprobe Characterization," Transactions on Microwave Theory
and Techniques, vol. 41, pp. 1653-1658, 1993.
P. Knott and A. Bell, "Coaxially-fed tapered slot antenna," ELECTRONICS
LETTERS, vol. 37, pp. 1103-1104, 2001.
M. C. Greenberg, K. L. Virga, and C. L. Hammond, "Performance Characteristics of
the Dual Exponentially Tapered Slot Antenna (DETSA) for Wireless
Communications Applications," IEEE Transactions on Vehicular Technology, vol.
52, pp. 305-312, 2003.
188
Appendix A
Cell Preparation Protocol
This protocol is provided by Dr. Craviso’s laboratory.
PROCEDURE FOR ISOLATING BOVINE ADRENAL CHROMAFFIN CELLS
189
The procedure is designed for isolating cells from four glands; If time permits, collect at least 2
extra glands to ensure that there will be enough intact tissue. Cells are prepared in Dr. Craviso’s
laboratory.
TWO DAYS BEFORE PREP
Abbreviations
HBSS (no BSA) = Ca2+/Mg2+-free Hank’s Balanced Salt Solution
HBSS + BSA = Ca2+/Mg2+-free Hank’s Balanced Salt Solution + Bovine Serum Albumin
Remove a bottle of bovine calf serum and antibiotic/antimycotic from the freezer and place into
the refrigerator to thaw slowly.
Check that the following are prepared: 500 mL HBSS (no BSA),
2 L HBSS+BSA (can add BSA day before), 3L Ham’s F-12 Complete, 20% BSA, sufficient
aliquots of collagenase B, 0.2 M CaCl2, 5 ml of 6 mg/mL Ara-C, trypan blue, neutral red, 70%
EtOH (from 95% bulk stock) and 70% high-quality EtOH (from 100% stock).
DAY BEFORE PREP
Finish preparing any reagents. Discard the oldest prep of cells from the incubator and check the
water levels; make sure that CO2 tanks do not need to be replaced.
Perfusion apparatus: Check integrity of tubing (Cole-Parmer; Masterflex 96300-14) and catheters
(Fisher Scientific # 14-170-12E) by pumping milliQ H2O through them.
Make sure that each
catheter has perforations spanning 3/4” of the terminal ends and that they are not clogged.
To
make new perforations, use a 22-gauge needle and perforate through both walls of tubing. This
will ensure that hole diameters are large enough for optimal perfusion.
Put the following items on the lab cart: those on the top shelf are items to be used and those on
the lower shelf are for backup (note: ITEMS MARKED WITH AN ASTERISK ARE
AUTOCLAVED)
Top Shelf:
Four 150-mL conical tubes*
Two 1-L wide-mouth bottles*
One stainless steel strainer*
190
One 95 µm (octagonal mesh) & one 53 µm (square mesh) filter*
Four 400-ml beakers*
Two glass Petri dishes*
Two 15-mL glass vials
One 10-mL syringe (dedicated for filtering the collagenase)
One 0.45 µm Cameo filter
One 0.22 µm Cameo filter
One package sterile sutures
Three dissecting scissors * which bear the engraving ROBOZ RS-5912, German stainless;
these scissors are to be used for dissection only)
Two forceps*
One saw-toothed forceps*
One saw-toothed hemostat*
Bottom Shelf:
Four 400-ml beakers*
One 1L wide mouth bottle*
Additional scissors and forceps*
Ensure that there are a sufficient number of vented T-175 tissue culture flasks and 100 x 20 mm
Petri dishes stocked in their respective cupboard/drawer for easy access.
Check that the laminar flow hood area is stocked with: filled EtOH squirt-bottles, all sizes of
sterilized glass pipettes; sterilized Pasteur pipettes; Kimwipes, paper towels, gauze, gloves, etc.
Fill a 250-mL wide-mouth bottle with fresh 70% EtOH, cap and place in hood for catheter
sterilization, to be performed on the day of prep.
Prepare the Styrofoam ice cooler (for transporting the glands on ice) and the travel-box (items
needed for removing the fat from the glands): take gloves, 70% EtOH, knife, large curved
scissors, paper towels, lab coat, a sandwich baggie, a clean blue bench diaper, and icepacks to
slaughterhouse.
Call the slaughterhouse and confirm that they will be slaughtering cows the following morning.
DAY OF PREP
191
MORNING IN THE LAB:
Ask someone to place 500 mL HBSS (no BSA), 2 L HBSS + BSA
(four 500-mL bottles), and 3 L Ham’s Complete (three 1-L bottles) into the 37°C water bath at
least 2 hours before dissection begins (usually around 10 am)
OBTAINING ADRENAL GLANDS
AT THE SLAUGHTERHOUSE:
•
Slaughterhouse workers will give you the glands, which will be embedded in fat
•
Trim fat off gland carefully using the large curved scissors
•
If a gland has been cut through by accident, discard
•
Cover adreno-lumbar vein with finger and rinse with 70% EtOH
•
Place gland in the sandwich baggie kept in the cooler with the icepacks.
•
Continue until a minimum of four glands are obtained, additional glands if time permits.
BACK AT THE LAB:
STERILIZE THE PERFUSION TUBING AND CATHETERS:
•
Immerse the perfusion inlet tubes and catheters in the 250-mL wide mouth bottle
containing 70% EtOH.
Pump at 7.0-9.0 for approximately 30-45 minutes.
Ensure that
the entire length of each catheter is submerged.
LAMINAR FLOW HOOD PREP (two people dissecting):
•
Place the ice cooler and the cart with all of the needed supplies within reaching distance
of the laminar flow hood
•
Swab the inside of the laminar flow hood with bulk 70% EtOH using a Kimwipe.
(Bulk
EtOH is used for sterilizing the work area and one’s gloves anytime his/her hands leave
and re-enter the hood.
•
High quality EtOH is used only for rinsing the glands.)
Fill the two blue dissecting trays and the white bowl with crushed ice.
Place a blue tray
and a pair of dissecting scissors where each person will be sitting.
•
Place the white bowl with the ice in the center of the hood and place a 400-mL beaker
with 200 ml HBSS (no BSA) into it
•
Set up each blue dissection tray as follows:
glass Petri dish on the Kimwipe.
place a Kimwipe on the ice, then place a
Push the dish down into the ice.
Open a sterile 4” x
4” gauze and, handling only the corner, place the gauze over the dish.
mL HBSS (no BSA) through the gauze into the dish.
Pour about 10-15
(The gauze will be “floating” on
the Petri dish at this point; this is okay. As dissection takes place, the gauze will get
192
pushed down into the dish and soak up the HBSS.) Save the gauze wrapper as an extra
sterile surface if needed.
DISSECTING THE GLANDS:
Note: Once dissection has begun, one should avoid getting up from his/her chair until the
dissection is finished and the dissected glands are in the HBSS on ice
•
Each person should remove one gland at a time from the ice cooler and rinse with 70%
high quality EtOH.
•
Remove outer cortex as follows: using the dissecting scissors, cut a circle through the
cortex around the adreno-lumbar vein, leaving a collar of tissue of approximately 3 mm
radius from the vein; cut the remaining cortex into two halves by cutting around the
perimeter of the gland; separate the cortex from the medulla one side at a time; cut a
groove under the cortex collar surrounding the vein. This will provide a lip of cortex
under which to secure the suture when catheterizing the glands.
•
Rinse each dissected gland with 70% EtOH and let the excess drip off prior to placing the
glands into the 200 ml beaker containing the ice-cold HBSS (no BSA) in the bowl of ice.
RINSE CATHETERS OF EtOH:
•
Drain all EtOH out of the catheters, and place catheters and intake tubes into a 400-mL
beaker containing about 150 ml HBSS (no BSA). Pump the HBSS through the catheters
for about three minutes.
Drape the tubing through the stabilizing device (an upright
clamp by the pump) to keep the catheters from falling out of the beaker.
Repeat this
rinse procedure using a separate 400-mL beaker containing 200 mL HBSS+BSA.
Turn
pump off before catheterization.
After this point, BSA will be present in the HBSS for all subsequent steps.
CATHETERIZATION:
•
Remove ice trays from under hood and swab the work area with 70% EtOH. Open a
4”x 4” sterile gauze and leave it on its wrapper.
Wet the gauze with HBSS+BSA.
Remove a sterile suture from its package using a forceps.
Double-up the suture and tie a
loose knot, placing the loop on the soaked 4”x 4” gauze.
•
Catheterize the gland (medulla) one at a time.
Secure each gland by the cortex collar
using a saw-toothed forceps and place it on the soaked gauze; position the cortex collar in
193
the loop of the suture.
Using fingers and/or forceps and/or scissors, work the suture into
the groove under the cortex lip, and tie the suture loosely under the lip.
•
Carefully insert a perfusion catheter into the adreno-lumbar vein. Gently work the
catheter into the full length of the vein by palpation (twisting the tube back and forth
helps). Make sure that the holes at the end of the catheter are inside the gland.
If you
can not get the catheter in far enough to cover up the holes, then remove the catheter and
snip off the tip to shorten the length of the perforated end.
Tighten and tie off the suture.
(The suture must be tight enough so that the gland will not slide off the catheter and the
perfusate will not flow out of the entry point of the catheter, but not so tight as to tie the
catheter closed and cut off the flow of perfusate.)
•
Rinse the medulla with 70% EtOH and let the excess EtOH drip off the gland. Place the
catheterized medulla into a 400-mL beaker containing the 200 mL HBSS+BSA at room
temperature.
•
Make sure the gland is submerged.
When all medulla are catheterized, perfuse (speed 2.5; make sure the correct direction of
the pump is used) with HBSS+BSA solution for about 5 minutes to flush out red blood
cells. Repeat this rinse procedure if the solution becomes excessively cloudy. During
this time, prepare the collagenase.
PREPARING
THE COLLAGENASE SOLUTION
•
Pour 50 mL / gland HBSS+BSA into a 400-mL beaker.
•
Using a 10-mL disposable pipette, dissolve by trituration 25 mg / gland of collagenase B
in a 15-mL glass vial containing approximately 10 mL HBSS+BSA.
Place the vial in a
styrofoam holder to prevent it from tipping over.
•
Use the dedicated 10 ml syringe for the following filtration steps.
collagenase first with a 0.45 µm filter into a second 15-mL glass vial.
Filter the
Then filter with
the 0.22 µm filter directly into the 400-ml beaker containing the 50 mL / gland
HBSS+BSA.
•
Add 12.5 µL 0.2 M CaCl2 per gland (i.e. per 50 mL HBSS+BSA) for 50 µM final
concentration.
(The 0.2 M stock solution is stored in the R219 refrigerator in a plastic
centrifuge tube which is labeled sterile and used for the CELL PREP ONLY!)
DIGESTION:
194
•
One at a time, transfer the catheterized medulla and the intake tube to the 400-mL
beaker containing the HBSS+BSA/collagenase/Ca2+ solution that is placed inside a
heating block to maintain the temperature of the HBSS at 30°C (make sure that the
temperature is monitored with a thermometer placed next to beaker inside the
heating block).
Perfuse the glands (speed 4); until they appear puffy and swollen (anywhere from 45
min to 2 hours). Make sure the glands are submerged and that the intake fitting is
not touching the glands.
•
Check to make sure that digested tissue does not enter tubing and clog the
perfusion. Observe glands through the beaker to see the degree of digestion. If
tubing becomes clogged, remove the tubing from the gland, purge with
HBSS+BSA/collagenase/Ca2+ solution to remove clogged tissue, then reinsert the
catheter.
Remove any digested glands and proceed with Cell Isolation to
prevent over-digestion.Clamp off catheter tubes using hemostats; reduce pump
speed/pressure accordingly.
CELL ISOLATION:
•
One at a time, transfer the digested glands to a new 400-mL beaker containing 200 mL
HBSS+BSA. Remove the catheters using a sterile forceps and place them together with
their intake tubes in a 400 ml beaker containing 400 m of MilliQ water.
Pump water
through and change occasionally.
•
To isolate the cells, use the saw-toothed hemostat and grab the cortex collar of one of the
medulla and hit the gland against the walls of the beaker, until the tissue loses its form.
The HBSS will become cloudy with cells.
Repeat for the remaining medullae. Note the
extent of tissue breakdown.
•
Place a 95 µm mesh filter funnel into a 1L bottle, and then place the stainless steel
strainer on top of the funnel. Wet the filter mesh with about 50 mL HBSS+BSA.
•
Slowly pour the dissociated tissue mixture through the filter apparatus. Rinse debris in
strainer with 50 mL HBSS+BSA.
•
Bring volume of filtrate up to 100 ml/gland using HBSS+BSA and gently swirl to mix.
•
Pour 100 ml of filtrate into each of the 150 ml conical tubes, then add 25 mL
HBSS+BSA to each tube.
Swirl gently.
195
•
Perform four centrifugation steps @ room temperature to remove red blood cells and
debris.
Use a separate Pasteur pipette for each aspiration step, and a separate10-mL
wide-bore glass pipette for each trituration step.
Bring volume up by pouring the HBSS
or medium directly into the centrifugation tubes.
a) 600 rpm for 6 minutes; you should see a pellet of whitish cells topped by a band of
red blood cells.
Aspirate so that 10 mL remains; resuspend the pellet by trituration
using a 10-ml wide-bore glass pipet; and bring the volume up to 100 mL with
HBSS+BSA.
b) 600 rpm for 4 minutes. The red band of blood cells should be gone; aspirate to 10
mL; resuspend cells by trituration; bring volume up to 100 mL with HBSS+BSA.
* At this point if there has been excessive cell damage a non-dissociable mass of mucuslike material containing released DNA may appear.
If this happens, remove the mass
with a sterile 1-mL pipette before continuing.
c) 500 rpm for 4 minutes; aspirate to 10 mL; resuspend by trituration; bring volume up
to 100 mL with Ham’s F-12 Complete.
d) 500 rpm for 3 minutes; aspirate to 10 mL; resuspend by trituration; bring volume up
to 100 mL with Ham’s F-12 Complete.
•
Place a 53 µm filter funnel onto a 1L bottle and wet the filter with 50 ml of Ham’s F-12
Complete.
•
Filter the cells through the 53 µm mesh into the 1L bottle. Rinse the funnel with about 50
mL of Ham’s F-12 Complete.
COUNTING THE CELLS:
•
Using a 1-mL disposable pipette, transfer about 1 mL of cell suspension from the 1L
bottle to a 12 x 75 mm test tube. Be sure to suspend the cells by swirling before taking
the aliquot, and take aliquots swiftly, as the cells will settle very quickly when not being
swirled.
•
To another 12 x 75 mm test tube, add 200 µL trypan blue from the sterile stock solution.
Now, using a P-200 with a snipped tip, transfer 200 µL of the cell suspension to the tube
containing trypan blue.
Again, make sure the cells are in suspension by tapping the tube
(do not vortex) before taking the aliquot.
•
Mix tube contents well by tapping with finger (do not vortex), and using a P-20 with a
snipped tip, load both sides (10µL/ side) of the hemacytometer marked “for cell prep
only” (LP2). Count cells, live (clear) and dead (blue). # Cells/ml = average of live cell
196
counts x 2(dilution factor) x 104(hemacytometer factor/ml). % dead = (average of dead
counts/ average of total counts) x (100).
•
Dilute cells with Ham's F-12 Complete Medium to a concentration of 4 x105 cells/ml.
•
Add appropriate amount of Ara-C to diluent volume to get a concentration of 1.5 µl/ml
from a stock of 6 mg/mL
•
Place hemacytometer in hot water with detergent to soak.
Be sure to clean the
hemacytometer before leaving, as it will need to be used the same evening.
DIFFERENTIAL PLATING STEP:
•
Using a 50 ml pipette, add 50 ml of the properly diluted cell solution into T-175 culture
flasks that have vented tops.
•
Suspend the cells uniformly before each transfer.
Place culture flasks in incubator (36.5°C, 5% CO2) for 5 ½ hours. Note the number of
flasks used.
CLEANUP BEFORE LEAVING:
•
Continue pumping fresh milliQ H2O through perfusion tubing and catheters.
•
If time permits, rinse catheters with 0.1 N NaOH.
finally wrap the catheters back up in plastic.
Rinse again with water and
At some later time, have the lab
assistant rinse the catheters out with several rinses of water, checking the pH until
back to neutral.
•
Clean hemacytometer.
•
Make sure all glassware, instruments, etc., are submerged in soapy water.
Dissecting instruments, nylon meshes, vials and other small items should be in the
small wash tub, while beakers and larger items are kept separate in the larger wash
tub.
AT NIGHT:
PLACING CELLS IN CULTURE:
•
Under the horizontal laminar flow hood, remove the 2” sterile stir bar from its storage
bottle of EtOH with sterile forceps and dry by placing on a sterile gauze.
(It has a
raised belt that keeps it from grinding cells against the bottom of the container that will
be used to distribute the cells from. Place the stir bar into a sterile 2-L Erlenmeyer flask
(or a 1-L bottle if a small volume anticipated.)
197
•
Remove the tissue culture flasks from incubator and place inside the hood. View the
contents of one flask under microscope before continuing.
Gently swirl flasks to
dislodge any chromaffin cells which may have lightly adhered to plated down nonchromaffin cells. Note suspended chromaffin cells (whether in clusters or single) and the
presence of attached cells. Then carefully pour the contents of each flask into the
Erlenmeyer flask or 1L bottle and stir to evenly suspend the cells. Observe the plated
down cells in one of the empty flasks.
•
Count the cells stained as described above using both neutral red (NR) and trypan blue.
Do not vortex cells. With respect to neutral red, count the numbers of NR+ and NR- to
get % chromaffin cells (stained red). Using the average number of NR+ cells, the cells
are then diluted or concentrated to exactly 1 x 105 cells/mL in Ham's F-12 Complete
medium.
(Hemacytometer for neutral red is labeled CP1.) Determine the number of
dead cells from the trypan blue count.
•
Compare the totals from each side of the hemacytometer as well as the totals from each
type of stain.
recount.
•
If numbers are approximately the same, continue with dilution: if not,
Again, use the chromaffin cell count (NR+) as final cell count for dilution.
Add Ara-C to bring to a concentration of 1.5 µl/ml (from a stock of 6 mg/mL) and mix
(don’t forget to take into account the amount of Ara-C that was added earlier at the
differential plating step).
•
Distribute the cells as desired, noting how and by whom cells are to be used. For general
lab studies, this is typically 40 mL into 100 x 20 mm Petri dishes. Place dishes
immediately in the incubator (36.5°C, 5% CO2).
For obtaining the CA profile of each
prep, place 2.0 mL into each of four 35-mm Falcon Petri dishes.
•
Before leaving, make sure all medium etc. is placed back in the refrigerator and the water
bath is turned off.
•
Leave hemacytometers to soak in hot detergent overnight.
Collect all waste tissue using the Ziploc bag, and discard it in the animal facility waste
freezer.
NEXT DAY:
•
Rinse and clean hemacytometers, air dry and put away.
•
Observe cells and record appearance.
•
Finish cleaning the perfusion tubing and catheters as necessary.
•
Make sure that the lab assistant is notified
198
REAGENTS AND SOLUTIONS NEEDED FOR PREP
Ara-C, 6 mg/mL (cytosine β-D-arabinofuranoside):
1) Dissolve 30mg Ara-C (Sigma #C-6645, R219) in 5mL dH2O.
2) Using a 5mL syringe, sterile filter (0.22µm) into a sterile 14mL round bottom tube.
3) Can be stored up to two weeks at 4 °C.
20% Bovine Serum Albumin (BSA):
1) Weigh off 5g BSA (Sigma A9418, tissue culture use only) into a 50 mL plastic centrifuge
tube. Add ~20ml dH2O; gently dissolve by both vortexing and letting stand. Then bring
up to 25mL.
2) Sterile filter (0.22 µm) into a sterile 50-mL centrifuge tube. (Use a Steriflip)
3) Store at 4 °C.
Ham’s Medium:
Ensure that you have the following:
Five 10.6 g packets of F-12 Nutrient Mixture with L-glutamine, without sodium bicarbonate
(Gibco BRL 21700-109, R219).
5.88g dry NaHCO3 (from tissue culture room)
5.95g HEPES (from tissue culture room)
Five autoclaved 1-L bottles, and one autoclaved 500-mL bottle.
1) Place about 4.75L of dH2O in a 5L beaker. Add stir bar. While stirring, add F-12 mixture,
HEPES, and NaHCO3. Rinse F-12 jar twice.
2) Dilute to 5L.
3) Adjust pH to 7.2 with 5N NaOH (from tissue culture room).
4) Filter sterilize using a millipore prefilter and a 0.22 µm bell filter. Discard the first 50mL of
filtrate. Fill the 1-L bottles to 900 mL, and put the remaining ~450 mL into the 500-mL bottle.
Label the bottles with numbers in the order that they are filled.
5) Store at 4 °C.
Ham’s Complete Medium:
Add the following components to 900ml Ham’s Medium under the hood.
100 mL Bovine Calf Serum (Hyclone Laboratories, SH30073.03).
10 mL of Antibiotic-Antimycotic (GibcoBRL 15240-062, F219):
1) Make sure that both components are completely thawed, sometimes requiring one
whole day in the refrigerator, followed by a short time in the water bath.
Neutral Red (0.03%):
1) Dissolve 90 mg NaCl and 3 mg neutral red (TC cabinet) in 10 ml dH2O.
2) Filter sterilize through a 0.22 µm filter under the hood.
3) Divide into 500 µl aliquots and store at -20°C (F219).
199
Trypan Blue (0.4%):
1) Dissolve 400 mg of Trypan Blue (Sigma, TC cabinet) and 900 mg of NaCl in 100 ml dH2O.
2) Filter sterilize under the hood with a 0.22 µm Nalgene filter.
Ca2+/Mg2+ Free Hank’s Balanced Salt Solution:
Ensure that you have eight 500-mL bottles (for 4L). Use reagents in box marked
only”
“tissue culture
1) Dissolve all salts in dH2O to 90% volume.
2) Add Phenol Red. Stir on stir plate.
3) Bring up to full volume with dH2O.
4) Titrate with 5N NaOH to pH 7.2.
5) Time yourself for 1 mL per 100 mL = 5 Minutes; Discard the first 50 mL.
6) Filter using a Millipore prefilter and a 0.22 µm Bell Filter into autoclaved 500 ml bottles.
7) Store at 4°C.
0.2M CaCl2:
Dissolve 2.94g CaCl2•2H2O (FW=147.0) in 100 mL milliQ H2O and autoclave for about 20
minutes on liquid setting. It is helpful to aliquot some 0.2M CaCl2 into a 12.0 mL plastic tube.
Both containers should be stored in the refrigerator (R219).
Ca2+/Mg2+ Free Hank’s Balanced Salt Solution + BSA:
Add 500µL of 20% BSA per 500mL Hank’s.
200
Appendix B
Cell Loading Protocol
201
Protocol for On-Line CA Secretion Studies
Preparatory
1. CPA components
Make sure dental putties are inserted into the upper and lower filter holders. Dental
putty are made of a vinyl polysiloxane (Dentsply/Calk: REPROSIL® Impression
Material) which fills the dead space of each side of the filter holder, leaving only a
central flow channel having a diameter of 1.8 mm for the BSS to pass through.
Prepare 9 mm diameter 350 µm mesh (SMALL PARTS, Inc. #CMN-0350-C) disks
by stenciling an image of a disk and cutting with small scissors.
Prepare 11 mm diameter 5 µm nylon mesh (SMALL PARTS, Inc. #CMN-0005-D)
disks by stenciling an image of a disk and cutting with small scissors.
202
Prepare 11 mm diametric GFC glass fiber filters (Whaman # 1822 042) by placing
the filter on a polypropylene cutting board and using a 11 mm punch and hammer to
stamp out the correctly sided disk. Up to 5 filters can be generated from one 42.5
mm filter disc.
2. Solutions - see end of protocol for preparation (section VII)
3. Cleaning the setup – see Day Before Experiment and Appendix D.
4. Flowcell assembly
Power of the ECD should be turned off prior to assembling or disassembling the
flowcell. The ECD flowcell is assembled by placing the Teflon gasket within the
metal assembly, as shown in the following figure.
Day Before Experiment
1. Cell Transport Protocol
Cells should be transferred from the incubator in the Howard Bldg to the incubator
in HREL at least one day using the styrofoam boxes labeled “Biological Materialsnon-hazardous.” Any number of Petri dishes may be transferred, as long as they fit
in transfer container. Dishes are sealed with parafilm to prevent leakage (~6x1/2
inch strip stretched to fit), transferred immediately to HREL. Avoid shaking the cells
as much as possible during transport, and keep box level. When unwrapping, place
the dishes on a flat surface. Wipe any spilled medium with a paper towel.
2. Make sure at least 100 mL of 5X balanced salt solution (BSS) is available.
203
3. Soak silicon gaskets, 5 µm mesh and 350 µm mesh in H2O overnight.
4. Make sure the ECD is off. Make sure the flow cell is assembled properly. Prime the
pump using DI water by placing the inlet tube in a beaker and engaging the pump
with flow rate of 1 ml/min. Remove the clamp from pump. Allow gravity to feed
water through system overnight.
5. Check to ensure sufficient amount 5 mM DMPP stock prepared in Dr. Craviso’s
laboratory.
Day of Experiment
I. EQUIPMENT
1. Turn on the computer and all instruments in the Rectangular Aluminum Enclosure
(RAE).
2. Follow instructions in section II for preparing 1xBSS. Remove the DI water from the
inlet container. Add 300 ml of 1x BSS.
3. The computer programs that run the equipment and record the data are called VIs (for
Virtual Instrument). Open the relevant VIs by running “Data Logging.vi”.
4. Run FlowControl.vi, ECD.vi and ECD_Luxtron.vi. Set flow to 1 ml/min and wait for
the pressure to stabilize. Read the pressure and flow rate of the BSS flow through
system.
¾ Pressure should be around 2 ~ 3 ± 0.07 psi for the flow rate set at 1 ml/min.
¾ Flow rate should be kept at 1 ± 0.2 ml/min.
5. If the pressure or flow rate is not within the above range, some part of the tubing may
be obstructed. Find the obstructed part of tubing and clean the tubing using .1M
NaOH and a 70 ml syringe. System must be flushed for 10 minutes with DI water
before continuing.
6. Run Luxtron.vi. Make sure the temperature is maintained following range.
¾ Inlet Temperature should be (36.5 ± 0.15 oC).
¾ Outlet Temperature should be (≈ 34.5 oC).
II. PREPARING 1 X BSS
1.
2.
3.
4.
For 500 ml: Pour 100 mL of the 5X BSS into a 100 mL graduated cylinder.
Transfer the 5X BSS to a 1 L volumetric flask, rinsing the graduated cylinder twice
with H2O, adding the rinse to the volumetric flask, and complete the volume to 500
ml with milliQ H2O, swirling to mix.
Transfer contents of volumetric flask to a 1 L side-arm flask and add 0.9 g glucose
and 0.5 mL 2M CaCl2 (2 mMf) while stirring on a stir plate. Make sure glucose
dissolves.
Filter the BSS using a Millipore 0.22 um nitrocellulose filter (46 mm in diameter)
and the Millipore glass filter assembly as follows.
204
5.
6.
With the flask on a stir plate, plug the top of the flask with a size 8 rubber stopper
covered in foil and connect the side-arm flask to the vacuum pump. Turn on stir plate
and vacuum pump, degassing for ~30 min.
After degassing, fill the designated BSS container with BSS. Perform all transfers
carefully so as not to re-introduce bubbles into the degassed working solution.
Use same procedure for preparing one liter of BSS, making sure to use correct
amounts of CaCl2 and glucose.
III. SETTING UP THE FILTER HOLDER ASSEMBLY AND PERFUSION SYSTEM
1. Hold the top half of the filter holder upside down. Add in this order (all component
pre-wetted with BSS): a silicone gasket (O.D. 13 mm and I.D. 9 mm ), a 350 µm
nylon screen (9 mm), a GFF disk (11 mm), a 5 µm nylon mesh (9 mm), a second 350
µm screen, and the silicone gasket. Screw in the bottom half of the filter holder and
check to make sure that the gaskets form a watertight seal.
2. Secure the CPA into the test tube clamp, connect a 3 ml, luer type, glass syringe to
the top of the filter holder, connect a silicone tubing with a Polycarbonate luer valve
(Cole-Parmer # EW-30600-06) to the bottom of the holder and attach a 50 ml plastic
syringe to the open end of the tube.
3. Fill the glass syringe with BSS and pull the BSS through the GFF by pulling on the
50 ml syringe plunger. Stop when 1 ml of BSS is left in the glass syringe using the
pinch clamp attached to the outlet tubing. Be sure to remove any bubbles stuck in the
CPA.
4. Attach the BSS inlet line via the luer lock to the top of the filter holder; then attach
the effluent tubing segment to the bottom of the filter holder.
5. Observe the CPA to make sure the CPA does not leak. If the CPA leaks, then
205
disassemble the CPA and assemble it again with a new GFF.
IV. PREPARING THE CELLS
1. 1 x 106 cells obtained from a 100 mm dish containing 3-4 x 106 cells/40 ml is used for
each experiment.
2. If BSS at 4 oC, place the BSS bottle on lab bench for enough time to reach room
temperature.
3. The day before the experiment, take one 100 mm Petri dish from the incubator, and
observe the cells under a microscope. Place the dish under the sterile hood, with the hood
fan on. If the cells are attached on the bottom of the Petri dish, gently scrape them loose
with a rubber policeman.
4. Using a 50 ml sterile disposable pipette, transfer the cells and medium to a sterile 100 ml
glass bottle.
5. Gently swirl, and depending on the total cell number, evenly distribute the cells into 3 or
4 aliquots to obtain approximately 1 million cells for each aliquot. Each aliquot is placed
into a 60 mm Petri dish and returned to the cell culture incubator.
6. Add 200 µl tyrosine and 200 µl ascorbic Acid (see section VII). Incubate overnight.
7. Take out a 60 mm Petri dish from the incubator and transfer the cells and medium to a 40
ml centrifuge tube. Place the centrifuge tube into a centrifuge (Eppendorf 5810; swinging
bucket rotor).
8. Centrifuge cells at 400 rpm for 10 minutes at room temperature.
9. Leave the cells pelleted in the tube in the centrifuge until ready to use (consistently 40
minutes total time from start of centrifugation to cell loading in the CPA).
V. LOADING AND EQUILIBRATING THE CELLS
1. After centrifugation and pelleting of the cells, prepare the injection syringe with DMPP.
2. Remove the CPA from the perfusion system and in its place, insert a polycarbonate
luer valve (shown in Figure below) that allows the BSS continuously to flow the
system.
3. Secure the CPA into the test tube clamp. Connect a 10 ml, luer type glass syringe to
the top of the filter holder, then connect a silicone tubing with a Polycarbonate luer
valve to the bottom of the holder and attach a 50 ml plastic syringe to the open end of
the tube.
206
4. Fill the glass syringe with BSS and pull the BSS through the GFF by pulling on the
50 ml syringe plunger. Stop when 1 ml of BSS is left in the glass syringe using the
pinch clamp attached to the outlet tubing. Be sure to remove any bubbles stuck in the
CPA.
5. Replace the 10 ml syringe with a 1 ml syringe. Check to make sure the syringe is
perfectly horizontal (use the small round leveler).
6. Fill the 1 ml syringe with BSS.
7. Using the microman positive displacement pipetter set at 250 uL, transfer cells (from
bottom of centrifuge tube) to 1 ml glass syringe attached to CPA, after the 40 minute rest
period.
8. Open the pinch clamp to allow the cells to drop into the CPA and distribute onto the GFF.
9. Add BSS as needed to prevent air bubbles from entering the CPA. Leave a small
amount of BSS in the syringe before closing the pinch clamp.
10. Record the pump speed that makes the flow rate 1 ml/min from the pump controller
program. Enter the recorded pump speed to the manual pump speed entry titled
“Pump1 speed” (Appendix H1), then switch the pump controller from automatic to
manual mode. This will stop the active flow rate controller and lock the pump speed
until the end of experiment.
11. Remove the glass syringe and take the CPA with the silicon tubing to the free space
exposure system.
12. Remove the plastic coupler, then place the injection syringe into the injector.
13. Connect the inlet tubing to the top of the CPA. Make sure no bubbles are introduced
into the CPA. Remove the attached silicone tubing from the CPA and connect the
outlet to the bottom of the CPA.
207
14. Start Luxtron.vi, ramp up temperature from 26 ℃ to 36.5 ℃ for 600 seconds.
VI. CHECKING CELL RESPONSIVENESS
1. The cells release a large amount of CA right after cell loading due to the changes in
flow rate and pressure inside the CPA. Monitor and record the maximum peak height
of the CA released by the cell loading.
2. After the pressure and temperature are stabilized, inject the DMPP to stimulate the
cells.
VII. SOLUTIONS
Balanced Salt Solution (BSS)
1) Dissolve all salts in a beaker of milliQ water at 90% volume with a stir bar on a stirrer.
2) Titrate with 5N NaOH to a pH of 7.4.
3) Complete the volume of the mixture by adding milliQ water (use a graduated
cylinder). Mix well. Store at 4°C.
On day of use, make approximate volume of 1xBSS (500 ml per experiment) and add
glucose ( ) and calcium chloride (2mMf).
CaCl2 2 M stock:
Dissolve 29.4 g of CaCl2 * 2H2O in 100 ml milliQ water. Mix well. Ready to use.
Store at 4°C.
DMPP Solution:
75 µM DMPP: To prepare 1 ml DMPP solution, transfer 985 µl of BSS using 1000 µl
pipette and 15 µl of 5 mM DMPP stock (prepared in Dr. Craviso’s laboratory) using 20 µl
pipette into a 50 ml plastic test tube. Fill entire injection syringe with 75 µM DMPP by
pulling plunger of the syringe.
Tyrosine:
208
To prepare a 50 mL 2mM stock solution of tyrosine, weigh off 0.018g of
Tyrosine(mw= 181.2g/mol) and dissolve in 50 mL of Ham’s F12 incomplete (prepared in
Dr. Craviso’s laboratory). Crystals dissolve readily in H2O but take longer in Ham’s F12
so let incubate for a few hours at 37 oC with frequent agitation (shaking solution) every
30 minutes. Once a majority of the crystals have dissolved, use a steri-flip filter to
sterilize the solution. Store at 4 oC.
Ascorbic Acid:
To prepare 5 ml of a 40 mM stock solution, weigh off 0.035g of scorbic acid
(MW=176.1g) and dissolve with 5 ml of milli-Q H2O in a glass vial. Use a 5 ml syringe
and a 0.22 µm sterilization filter to sterilize. Store at 4 oC.
209
Appendix C
Control / MW Experiment Protocol
210
I. CONTROL EXPERIMENT PROTOCOL
1. Perform procedure described in Appendix B.
2. Run “Injector.vi.”, where the front panel of the program will appear (Figure 1).
Figure 1. Front panel of the Injector.vi.
3. Type the following in the entries on the front panel
¾ Addition: 10 (10 minutes injection interval)
¾ Delivered Volume:10000.0 (10 µl injection)
¾ Delivered Rate: 1000.0 (1 µl/sec injection rate)
¾ Start Time: Type desired time for the first injection
¾ Addition: Type desired injection interval, typically 10 minutes
¾ Press “ADD” button 20 times to inject 20 times.
II. MW EXPOSURE EXPERIMENT
Perform the procedure described in Section I for the injection of DMPP and then add
the exposure parameters
.
1. Run “Discrete RF Control.vi” (Figure 2) and press the OPEN button (circle in
red) to run a subvi “Mod Type Seletion.vi”(Figure 3).
211
Figure 2. Front panel of “Discrete RF Control.vi” that controls the MW exposure
parameters.
Figure 3. Front panel of “Mod Type Seletion.vi” program that sets the time and MW
exposure parameters.
2. Set the exposure time (Section III) and select the MW exposure parameters using
the Mod Type Seletion.vi program.
III. EXPOSURE TIME
During MW exposure experiments, the cells were exposed to MW fields for
either (1) a short-term exposure of 10 minutes or (2) a long-term exposure of 30 minutes.
212
Detailed time sequences for both the short-term and the long-term exposures as well as
the protocol for DMPP injection are presented below.
For the short-term exposure and DMPP injection protocol (shown in Figure 4),
the cells are stimulated by DMPP every 10 minutes (labeled 1) with an injection time of
10 seconds (labeled 2). The MW exposure is programmed to start 1 minute (labeled 3)
after injection of DMPP 1 in order to provide sufficient time for the injected DMPP
(injection 1) to pass to the GFF where the cells are immobilized and for the released CA
to be detected by the ECD (peak 1; 40-50 seconds). Hence, the cells are exposed to
MW fields for 9 minutes (labeled 4) prior to the second DMPP injection (injection 2) and
for an additional minute after the second DMPP injection (labeled 5), thus giving a total
MW exposure time of 10 minutes.
The time sequence for the long-term exposure experiments is shown in Figure 5.
Here, the cells are exposed for 9 (labeled 2), 19 (labeled 3), and 29 (labeled 4) minutes
prior to injections 1, 2 and 3, respectively, and the total exposure time to the MW field is
30 minutes (labeled 1).
213
Figure 4. Time sequence for the short-term MW exposure and DMPP injection protocol.
Figure 5. Time sequence for the long-term MW exposure and DMPP injection protocol.
214
Appendix D
System Cleaning Process
215
I. AFTER THE EXPERIMENT
1. Have available at least 100 mL of 0.1 N NaOH, 100 ml of 40 % ethanol, and 500
ml of fresh DI H2O.
2. Remove the CPA, and replace with the plastic connector (Label 1 in Figure 1).
3. Make sure the flowcell is assembled (Figure below) and the ECD is turned off,
otherwise unwanted oxidization by the cleaning solution could occur.
4. Turn off the power supplier of temperature controllers for the CPA and ECD.
5. Place the perfusion system inlet tubing into the container with 0.1 N NaOH; set
the speed on the Pump Controller.vi to 0.5 ml/min and start the program. Allow
at least 10 minutes for the base to flow through the system.
6. Replace the base with DI water. Allow at least 10 minutes for the water to
completely rinse the system.
7. Replace the DI water with 40 % ethanol. Allow at least 10 minutes for the
ethanol to flow through the system.
8. Flush the system with copious amounts of DI water.
9. Stop the pump.
II. OVERNIGHT
1. Place a 3 mm diameter piece of silicone tubing (O.D. 6 mm) filled with DI water
at the end of the system (Label 2 in Figure 2).
2. Prime the system by pumping 1 ml/min of DI water for 10 minutes to make sure
the entire system is filled completely with DI water.
3. Shut off pump.
4. Allow gravity to continue feed water through the system overnight. Make sure
there is at least 500 ml of water in the container.
216
Figure 1. Simplified schematic diagram of the exposure system prepared for cleaning.
217
Appendix E
E Field and SAR Distribution Calculator
(written in Mathcad 2001)
218
219
220
221
222
223
224
225
226
Appendix F
Procedure of Statistical Analysis for Assessment of a
Bioeffect
227
A control experiment and one of its corresponding MW exposure
experiment performed on 04/02/2008 are analyzed, and the protocol as
well as output of the Mathcad program are described below.
Step (1): Import all the control experiment data from SigmaPlot to the Mathcad program.
This data includes the areas under the CA peaks, the trend line and the 95 % prediction
band. Plot all the data to ensure that they have been properly imported into Mathcad, i.e.
that the trend line follows the CA peaks and is located at the center of the 95 %
confidence band as shown in Figure 1. The equation for the specific control experiment is
given by
ycontrol= 2007.35 ⋅ e −0.0982⋅ x + 485.91
(1)
Step (2): Import all the MW exposure data from SigmaPlot to the Mathcad program. This
includes the areas under the CA peaks and the trend line. Plot the trend line and CA peaks
for the MW exposure experiment to ensure that they have been properly imported into
Mathcad, i.e. that the trend line follows the CA peaks as shown in Figure 2. The equation
for the specific MW exposure experiment is given by
ymw= 433.00 ⋅ e -0.1509⋅x + 699.48
(2)
228
Figure 1. Imported control experiment data showing the CA peaks, the trend line, and the
95% prediction band.
229
Figure 2. The trend line generated by Mathcad and the CA peaks for the MW exposure
experiment.
Step (3): Once the two graphs (Figures 1 and 2) show that both sets of experimental data
have been properly imported, normalize the curves to the maximum trend line values for
each graph and multiply by 100 to make the maximum value equal to 100 (%).
Expressing the normalized values as a percent instead of a fraction is necessary for the
accuracy of calculations, since Mathcad can accurately calculate only to a limited number
of decimal places. Plot the graphs to ensure that the normalization is properly performed.
Figures 3 (a) and (b) show the normalized curves, which indicate that both the trend lines
begin at 100 %, the corresponding curves follow the CA peaks and the overall shapes are
the same as in Figures 1 and 2 respectively.
230
(a)
(b)
Figure 3. (a) Normalized control and (b) MW exposure experimental profile obtained
using Mathcad.
The normalized trend line equations for the control and MW exposure experiments are
given respectively by
normalized ycontrol= 87.07 ⋅ e −0.0982⋅x + 21.08
(3)
normalized ymw= 40.40 ⋅ e -0.1509⋅x + 65.26
(4)
Step (4): Calculate the percent difference between the trend line (equation 4) and CA
peaks for the MW exposure experiment using equation 5.
% difference( x ) = Normalized Trend Line( x ) − Normalized CA Peak ( x )
(5)
where x = injection number.
The percent differences for the specific experiment are shown in Table 1. These values
will be used later to find the modified CA peaks.
231
Table 1. The percent difference between the CA peaks and the corresponding trend line
for the MW exposure experiment calculated using Equation 5.
Step (5): As discussed above, the trend line of the MW exposure experiment is modified
to fit the trend line of the control experiment. This is realized by multiplying the
parameters a, b and c (equation 4) by unknown parameters u1, u2 and u3. For example,
parameters of the trend line for the MW exposure experiment (equation 4) are multiplied
by unknown parameters, u1, u2 and u3, as shown in equation 6.
y mw = 40.40 ⋅ u 1 ⋅ e -0.1509⋅x⋅u 2 + 65.26 ⋅ u 3
(6)
Step (6): Equation 6 is the equation for the modified trend line of the MW exposure
experiment, which will overlap the trend line of the control experiment (equation 3).
Hence, if proper u1, u2 and u3 values are obtained, the two equations (equation 3 and 6)
will be almost identical. Hence, subtracting equation 3 from equation 6 and summing
over all injection values will result in zero or a value close to zero:
232
∑ [(8 7.07 ⋅ e
−0.0982⋅ x
) (
)
+ 21.08 − 40.40 ⋅ u1 ⋅ e -0.1509⋅x⋅u2 + 65.26 ⋅ u3 ] = 0
(7)
x
where x = injection number. Mathcad can solve such a problem and find the unknown
values using a function called “solving equation” which uses the well known quasiNewton method. Here, the calculated values are u1 = 2.155, u2 = 0.651 and u3 = 0.323 and
the modified trend line for the “modified MW exposure experiment” is given by:
y = 40.40 ⋅ 2.16 ⋅ e -0.1509⋅x⋅0.651 + 65.26 ⋅ 0.323 = 86.86 ⋅ e -0.098⋅x + 21.08 (8)
Step (7): The last step is to calculate the CA peaks based on the modified trend line and
the percent difference shown in Table 6.1. Solving equation 5 for the CA peaks:
CA Peak ( x ) = Modified Trend Line( x ) − % difference( x )
(9)
where x=injection number. The final Mathcad result is shown in Figure 4. It is seen that
the two trend lines overlap and there is only one CA peak (circle in red) that is outside the
95% prediction band (-6.97% in table 6.1).
233
Figure 4. The final result of the Mathcad analysis program showing the normalized
control trend line, 95% prediction band, modified MW exposure trend line, and modified
CA peaks for the MW exposure experiment.
234
Appendix G
The ECD Setting
235
Appendix G.1: Output Range Setting
As described in Section 4.5.1, the current released by the oxidization process is
proportional to the concentration of the analyte. However, the electronic device within
the ECD that measures, amplifies, and outputs the electric current, LC-4C (BASi), may
not always have an output that is proportional to the concentration of the analyte when
the height of the peak is outside the “output range” setting. It is important to understand
how to properly set the output range since this setting is closely related to the linearity,
sensitivity and noise level of the response.
The output detection range button is located on the front panel (Figure 1, circle in
red) of the LC-4C and can be chosen in the range from 0.1 nA to 50 µA in 18 steps. The
ECD is, however, designed to read and display responses ten times higher than its range
setting. For instance, if the detection range setting is set to 10 nA, the ECD can measure a
response as high as 100 nA (ten times higher than 10 nA). What is important about the
range is that the response within the range is designed to be linear, i.e., the height of the
response peak is proportional to the concentration of analyte. Thus, in order for a
response to be linear, the detection range should be set to higher than the largest peaks to
be detected. For example, if the maximum peak height is 19.5 nA, the closest
predetermined range, 20 nA, should be selected.
Proper setting of the output range is also important to achieve high sensitivity
and low noise responses. When the output range is set to one that is much higher than the
amplitude of the detected peaks (for example, the largest range (50 µA) is selected when
detecting 10 nA peaks), the height of the peaks cannot be compared accurately since the
50 µA range is not sensitive enough to distinguish the heights of such small peaks. The
236
larger output ranges result in larger noise in the output, and hence if small heights of
peaks are measured using a large output range, the noise level may be too large to
distinguish it from the actual signal. In this project, the range is selected to be 50 nA.
Figure 1. The front panel of the LC-4C that measures, amplifies, and outputs the current
caused by the oxidization process. The keys circled in red are used to set the output range.
Appendix G.2: Background Current of the ECD
As described in the previous section, the linear region is restricted by the output
range setting. Hence, obtaining a low background current is important for achieving a
wider range of linearity. There are three major factors that affect the level of the
background current of the ECD: (1) the state of equilibrium, (2) the presence of
electrolyte in the BSS [1], and (3) the basal rate of CA release from the chromaffin cells.
Figure 2 illustrates the baseline change due to each of these factors.
As soon as the flowcell switch is turned on (i.e., 0.65 V is applied to the flowcell),
a large current is measured, as shown in region ① of Figure 2. This is due to additional
237
charge required for the electrode interface to achieve the applied potential difference
(0.65 V). The response then rapidly decreases and stabilizes, which is called the “state of
equilibrium” [2]. The manufacturer (contacted via e-mail) claims that the equilibrium
time should be less than 15 minutes, yet it was found that it varies slightly from this value
depending on the flowcell. The background current in the equilibrium state is close to
zero when DI water flows over the working electrode.
Region ② shows an increase in the baseline when ID water is replaced with BSS.
The increased background is due to the presence of a spontaneous electrical double layer
caused by the presence of electrolytes in the BSS [1]. The constant current caused by the
oxidization of the constituents of the continually perfused BSS is called the “steady-state
background current”. It was found that the intensity of the steady-state background
current varies widely with each flowcell.
After the cells are loaded onto the GFF within the CPA and the CPA connected
to the on-line perfusion system, the cells initially release a large amount of CA (Region
3). This is due to a sudden change in flow of the BSS (0 to 1 ml/min) presented to the
GFF, as well as the pressure fluctuation (0 to 4 psi) that occurs at the instant the CPA is
connected to the on-line perfusion system (Appendix B). After the cells are loaded and
the background current reaches some maximum value, the current gradually decreases to
a level that reflects the amount of CA released spontaneously (basal release) from the
cells (Region 3). This basal release gradually decreases over a period of time ranging
from 30 to 60 minutes and becomes constant. This time depends on (1) the sensitivity of
the cells to the manipulations associated with cell preparation and cell loading, and (2)
the pressure and temperature change existing once the perfusion of the cells is started.
238
Figure 2. Various ECD outputs. Region #1: the liquid passing through the ECD flowcell
is DI water; region #2: the DI water is replaced with BSS; region #3: the cells are loaded
in the system.
Sensitivity of the cells to being manipulated is a natural characteristic of these cells.
Hence, a protocol has been developed by trial and error that minimizes stimulation of the
cells (Appendix B).
In summary, the baseline of the response becomes stable after the state of
equilibrium for the flowcell, the steady-state background current for the BSS, and the
constant basal release of CA from the cells are established. In other words, if one of these
components has not reached steady state, a constant baseline for the ECD output will not
be achieved. The times for equilibrium for the flowcell and steady-state background
239
current were dependent on the flowcell that was used. Of the flow cells available, the
selection of the best one is described in the next section.
Appendix G.3: Selection of the Flowcell
It was found that different flowcells from the same manufacturer exhibit
different levels of performance. For example, one particular flowcell took several hours
to reach a steady-state background current, whereas another flowcell took only thirty
minutes. The sensitivity of detection between the flowcells was also different. These
differences are due to variations in fabrication of the carbon electrode during
manufacturing or after use for example, an irremovable film can build up on the surface
of the carbon electrode due to heavy oxidization). Of the three flowcell assemblies
available in our laboratory, the one that provides the most sensitive detection of CA as
well as the shortest time to reach equilibrium and steady state background current was
selected and used for all experiments.
Appendix G.4: Selection of the Flow Rate
As discussed in Section 4.5.1, in order for the ECD output value (i) to depend
only on the concentration of CA (c), the flow of the BSS (u) should be maintained
constant. The actual flow rate is also important. For example, if the response to DMPP is
observed as very broad peaks, the flow rate is set too low. This is because both the DMPP
that stimulates the cells and the CA that are released from the cells tend to get diluted
when the BSS perfusion rate is too slow. On the other hand, when the BSS flow rate is
240
too high, the pressure in the system is increased. This result is not desirable either, since
the cells should be maintained at a pressure low enough to achieve optimal simulation by
DMPP. Various flow rates that were used by two different research groups that also
performed experiments with perfused chromaffin cells [3, 4] were tested. It was found
that a flow rate of 1 ml/minute generates a pressure of only 3-4 psi in our system as well
as provides narrow CA peaks. Thus a flow rate of 1 ml/minute was used in all
experiments.
Appendix G.5: A Fixed Output Range
Figure 3 illustrates a typical response of the ECD obtained when chromaffin cells
are stimulated with DMPP. The ECD response includes a steady-state background current,
noise and peaks due to oxidization of the stimulated release of CA from the cells. As
shown in Figure 3, the final height of a peak produced by released CA is the summation
of the height of the peak (“signal”) and the background current.
As explained in Appendix G.1, the response is linear only within the output
range determined by the user. Hence, in Figure 3, the heights of the second and third
peaks are proportional to the concentration of CA, whereas the first peak height is not
expected to be proportional to the CA concentration.
The overall amplitude of the signal and background current is a function of (1) the cells
(i.e., how they were handled and number of cells loaded onto the GFF), (2) the pressure
and flow rate of the BSS, and (3) the condition of the glassy carbon electrode. Although
factors listed above (2 and 3) are well controlled, basal CA release can vary, resulting in
different steady-state background currents. The response of the cells to DMPP can also
241
vary. Hence, from experiment to experiment, the overall height of the peaks also varies.
This variation can range from a few nA to a few hundred nA.
Because it is not possible to know how the cells will respond until after the CPA
is loaded with the cells and connected to the system, and the cells are stimulated several
times with DMPP, the best output range of the ECD cannot be chosen at the beginning of
an experiment. Moreover, it is best not to change the output range setting as an
Figure 3. Illustration of a typical response of the ECD, where the chromaffin cells are
stimulated by DMPP. Total peak height is determined by the amplitude of the signal, the
amplitude of the steady-state background current and noise of the ECD at the output
range setting.
experiment is in progress since changing the setting causes fluctuation of the ECD output
and also the background current may be different before and after changing the setting. In
addition, the height of the peaks at different output ranges is slightly different, and hence
the peaks before and after the setting change cannot be compared.
242
To obviate these problems, a fixed output range that covers a variety of peak
heights that provides a good SNR as well as a linear response regardless of varying peak
heights from experiment to experiment was determined by trial and error. It was found
that an output range of 50 nA fulfils these requirements. Problems associated with using a
fixed output range of 50 nA become apparent in cases where the output of the peak is
either too small or too large: if the peak is too small (less than 5 nA), the output of the
ECD is unreliable and hence the experiment is discarded. If the peak is too large (larger
than the output range setting), a method developed for handling peaks outside the output
range setting is used, as described in the next section.
Appendix G.6: Manual and Zero Offset
Lowering the background current to reduce the height of the peaks is one way
to enable analysis of peaks whose heights exceed the 50 nA setting. The EC-4C includes
two functions, the “manual offset” and the “zero offset” (Figure 4) that null out the
background current to establish the operating baseline at low current values. The manual
offset function reduces the background current by the amount set by user (using manual
offset adjuster). The zero offset function reduces the background current value
automatically, functioning similar to the manual offset function, except that the amount
reduced will be the value of the current when the zero offset button is initially pressed.
The background current value will be reset to zero whenever the zero offset button is
pressed.
In our experience, we found that the zero offset and manual offset functions
are only useful when the baseline current is at steady state. As discussed in Appendix G.2,
243
once the CPA is connected to the system, it takes a long time for CA release to reach
steady state. Moreover, both the rate of decrease and the time it takes to reach steadystate can vary substantially from experiment to experiment and in most cases a perfect
steady-state condition is not achieved. Hence, if the background current is arbitrarily
decreased by the user, the background current often becomes lower than zero. Hence in
most of the experiments performed, the offset function is not used. Exceptions are in
cases where the background current reaches steady state in the early stages of the
experiment or the background current is excessively high.
Figure 4. The front panel of the LC-4C that measures, amplifies, and outputs the current
caused by the oxidization process. The switches within the circles are the manual offset
(circle in red) and auto zero offset (circle in blue) adjusters.
Even though the ECD includes functions specifically designed to reduce the height of the
peaks, there functions cannot be used for most of the experiments due to continuously
decreasing background current; hence heights of peaks larger than the detection range
(such as the first peak in Figure 3) may not be directly obtained from the ECD.
244
In order to measure the peaks outside of the output range, the non-linear region
of the ECD is used. The details are presented in Section 4.5.3.
Appendix G.7. Conclusion
The output range setting of the ECD chosen for this research, that uses a flow
rate of 1 ml/min, is a fixed output setting of 50 nA. In order to handle peaks outside the
output range (50 nA), the non-linear region of the ECD is used (Section 4.5.3), instead of
the offset functions included in the ECD.
Appendix G.8. References
[1]
[2]
[3]
[4]
W. Wang and S. A. Soper, Bio-MEMS:Technologies and Applications. New York,
NY: CRC Press, 2006.
Instruction Manual: Principles of EC Detection and Troubleshooting Guide. West
Lafayette Indiana: Bioanalytical Systems, Inc, 1994.
R. Borges, F. Sala, and A. García, "Continuous monitoring of catecholamine
release from perfused cat adrenals," Journal of Neuroscience Methods, vol. 16, pp.
289-300, 1986.
S. Lin-Liu and W. R. Adey, "Low frequency amplitude modulated microwave fields
change calcium efflux rates from synaptosomes," Bioelectromagnetics, vol. 3, pp.
309-322, 1982.
245
Appendix H
LabVIEW Programs
246
Appendix H.1. Pump Controller Program
The pump controller program controls the speed of the peristaltic pump. There are two
operational modes: (1) automatic mode and (2) manual control mode. The automatic
control mode dynamically maintains the flow rate (ml/min) to a set point, whereas the
manual control mode locks the speed of the pump (rpm) to a set point. In a typical
experiment, the program maintains the flow rate at 1 ml/min.
Appendix H.2. Temperature Controller Program
The temperature controller program controls and monitors the temperature at the inlet of
the CPA and at the outlet of the CPA. During a typical experiment, the temperature of the
CPA inlet is maintained at 36.5 oC.
247
Appendix H.3. Injector Controller Program
The injector controller program injects DMPP at the times defined by the user. In a
typical experiment, the injector injects DMPP every 10 minutes.
Appendix H.4. ECD Monitoring Program
The released CA are monitored by the ECD, and the output of the ECD is monitored and
recorded by the ECD monitoring program.
248
Appendix H.5. MW Controller Program
The MW controller program controls the signal generator to generate specific MW
parameters. The MW parameters that the program can enable/disable include the standard
modulations (CW, AM, FM, PM and phase modulation), power level, frequency, and
combinations of the existing modulations (PFS).
Appendix H.6. Amplifier Controller Program
The amplifier controller program can enable and disable the amplifier, display forward
and returned power from the amplifier, and monitor the status of the amplifier.
249
Appendix H.7. Power Monitoring Program
.
During MW exposure, the power monitoring program monitors and records the average
power delivered to the antenna.
Appendix H.8. ECD Inlet Temperature Controller
The temperature of the BSS that flows into the flowcell is controlled by the ECD inlet
temperature controller. Other temperatures that may possibly affect the ECD response,
including room temperature and the temperature on the auxiliary electrode block in the
flowcell assembly, are also measured and recorded by this program.
250
Appendix H.9. Data Logging Program
All the activities including current time, ECD output, CPA inlet and outlet temperatures,
ECD temperatures, pressure, flow rate, injection status, MW parameters, are recorded on
an instrument control computer by the data logging program for analysis.
251
Appendix I
Unreliable Experiments
252
Among a total of 51 experiments performed with MW exposure, 16 (31 %) were
categorized as “unreliable experiments” and are not included for analysis in this
dissertation. Experiments were considered to be unreliable if the ECD output has (1)
small CA peaks or (2) fluctuations in the baseline.
I. Small CA Peaks
The trailing of DMPP and released CA (Section 4.5.9) form shoulders on a CA peak, as
illustrated in Figure 1. Several experiments indicated that the effect of trailing can vary
slightly during the course of lengthy (4-5 hours) experiments due to several factors, such
as changes in room temperature that affect the flow rate of the peristaltic pump or
changes in the smoothness of the internal surface of the tubings due to deposits on the
tubing walls during experiments, etc. In these cases, the shape of the shoulders changes
causing variations in the area of the CA peak. While amount of this variation is negligible
when the height of the CA peak is large (Figure 1.a), the variation is noticeable when the
height of the CA peak is small (Figure 1.b). The heights of the CA peaks in the control
experiment shown in Figure 2.a are close to or less than 1 nA, and the peak heights and
areas under the peaks do not match, as shown in Figure 2.b. Comparisons performed
between the heights of and areas under the peaks indicate that the heights of peaks less
than 5 nA exhibit such a disagreement. Hence, all experiments with peak heights less
than 5 nA are ignored in the analyses.
253
(a)
(b)
Figure 1. The shoulder of a CA peak for (a) higher and (b) lower peak heights.
(a)
(b)
Figure 1. (a) Control experiment with small CA peaks (≤ 1nA). (a) ECD profile and (b)
the area under the peaks superimposed with the height of peaks to show the mismatches.
254
II. Baseline Fluctuation
If a sudden fluctuation of the baseline is observed during an experiment, the
experiment is also considered to be unreliable. Figure 3 is an example of an ECD profile,
with the area under the peaks (pink) superimposed over the peaks. In this experiment, the
baseline suddenly fluctuated in the middle of the experiment causing the heights and
areas of the peaks to also increase.
Figure 3. An experimental profile for the case in which there was baseline fluctuation.
Baseline fluctuations sometimes occurred both in control experiments and MW exposure
experiments. The reason for this type of baseline fluctuation has not been clearly
identified, and those experiments that had baseline fluctuations were not considered in
this dissertation
Документ
Категория
Без категории
Просмотров
0
Размер файла
9 853 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа