University of Southern California Dissertations and Theses

Development of implantable Parylene-based MEMS technologies for cortical applications

(USC Thesis Other)

Development of implantable Parylene-based MEMS technologies for cortical applications

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content
DEVELOPMENT OF IMPLANTABLE PARYLENE-BASED MEMS
TECHNOLOGIES FOR CORTICAL APPLICATIONS

by

Brian Jung Kim

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BIOMEDICAL ENGINEERING)

August 2015

Copyright 2015 Brian Jung Kim
I

To my family
II

ACKNOWLEDGEMENTS
It is with a grateful and overall amazed heart that I have finally sat down to write
the acknowledgements portion of my dissertation. It has been a long journey from the
beginning, and there are many individuals that I would like to thank for helping me during
this experience of a lifetime. It truly takes a village to raise a child, and I am grateful for
all of the support that I received during my academic growth over the past 5 years.
First and foremost, I would like to thank my parents, without whom I would not
have had the support and encouragement to pursue my educational aspirations. They have
sacrificed so much so that my sister and I would have opportunities that were unavailable
to them. My mother and father both inspire me in their hard-working characters, love for
each other and for family, and prayerful lives. I am grateful and appreciative of their
sacrifices for our family− thank you so much mom and dad, this is for you. Secondly, my
younger sister Tammy has always been a strong source of encouragement and advice from
trivial topics to important life decisions. I respect and look up to her, and it is with a great
sense of pride and appreciation that I also dedicate this work to her to show how grateful I
am to have her as a sister and a friend. I am very proud to be known as her brother and am
excited to see the woman that she will grow to be. Lastly, I would like to thank the entire
Won family, especially the cousins, from whom I have learned a great deal on the
importance of family and unconditional love and support. Thank you guys for being an
inspiration to me, for your continual encouragement, and for raising me right.
Before moving on, I cannot forget to mention a group of brothers that have become
family to me since college. William, Minq, DC, Sammy: thank you for your friendship,
camaraderie, brotherhood, and support over the past 5 years. Our reunions, inside jokes,
and conversations have brought me much laughter and happiness in the most difficult of
times. I look forward to future adventures with you brothers into our older ages. Legends
for life!
III

My path to USC began with the Viterbi SURE program that Dr. Ellis Meng was
gracious enough to accept me into. Having been on the other side of the scrutiny for
incoming undergraduate researchers, I thank Ellis that she gave me a chance to prove
myself in the challenging field of research. That summer as an undergraduate research
student within her lab sealed my interest within the bioMEMS field and definitely solidified
my goals in obtaining a Ph.D. Having the opportunity to work under Ellis over the past 5
years, I have learned so much and developed considerably under her supervision. As a
researcher, Ellis’s creativity and problem-solving acumen have inspired me to open my
mind and think outside the box to conceive new, more efficient solutions to frustrating
problems. As a mentor, she was always there to offer her advice and provide guidance and
direction whenever I needed it. Thank you so much Ellis for all that you have done for me
over the years!
My more technical lab skills and device development knowledge I can attribute to
my mentorship under Dr. Christian Gutierrez, who I think of much more than a senior
graduate student mentor, but as an older brother. Though I came in as a summer
undergraduate student, Christian took me under his wing and shared his insights and secrets
on conducting research in the development of Parylene-based devices. I am grateful that I
was paired with Christian during the SURE program, and that I had the chance to work
with him during my first year of my Ph.D. I will always remember his two lessons
regarding processing within the cleanroom the first time he took me in: namely on keeping
mindful of the thermal budget of materials and maintaining process latitude during recipe
development. His brotherly personality and insightful mind have greatly inspired me to
become the engineer and person that I am today. I cannot describe how grateful I am to
him as much of my own intellectual development as a researcher within our field has been
a result of his guidance and support. Thank you so much for everything Christian!
The past 5 years could not have been as enjoyable as it was without the
companionship of the members of the Biomedical Microsystems Laboratory. As I always
say, our lab is more than a just a lab, it’s a family, and I am deeply indebted to the
IV

Biomedical Microsystems Lab for their support and friendship over the years. Dr. Jonathan
Kuo, Dr. Seth Hara, and Dr. Curtis Lee: working with you guys truly has been an honor
and has confirmed my approval for the “trial by fire” method in learning and research. The
amount of progress that we achieved in such a limited time speaks to how awesome our
team was! Thank you for being more than lab-mates, but older brothers to me as well. Dr.
Heidi Tu: thank you for all your insightful Parylene advice and encouraging me to join
GSBME. My involvement with lab and BME activities are all because of you! Dr. Roya
Sheybani: without you Roya, our lab would be in shambles. Thank you for your outgoing
personality and passion for creating a truly family environment within lab and amongst its
members. Lawrence Yu: Law, I am grateful for your enthusiasm for research, keen eye for
efficiency, and overall camaraderie throughout the years. I will miss our conversations and
diving deeply into our various hobbies to further our extracurricular education and
development. Angelica Cobo: señora, thank you for your companionship, your similar taste
in music and comedy, and your interest in my fitness and well-being. I am going to miss
our Colombian food outings, early morning music listening sessions, and lively banter.
Alex Baldwin: after working with you for this short amount of time, I can already tell that
you are going to do great things. Don’t stop your eagerness to learn and excitement for
figuring things out. Jessica Ortigoza: Jess, thank you for your friendship and overall
kindheartedness within the lab. I’m going to be carefully watching for your results in
improving Parylene fabrication processes; I’m excited for the results! Ahuva Weltman: our
resident doctor, your ability to learn and pick up skills so quickly is amazing, and your
friendliness is contagious. I enjoyed working with you and am excited for your future
successes as an M.D./Ph.D. And finally Dr. Kee Scholten: I know that it has only been a
little time, but in that short time your advice has been invaluable and overall comprehensive
knowledge on almost everything (trivia!) is awesome. Your excitement for research and
passion for the pursuit of knowledge is contagious and inspirational; I look forward to
tracking your future research and publications and reading about your lab!
And with that, I would like to leave the current and future members of the
Biomedical Microsystems Lab with a quote from Dr. Richard Feynman: “Study hard what
V

interests you the most in the most undisciplined, irreverent and original manner possible.”
Be provocative, challenge notions, and do not limit yourselves; for in the most outrageous
of thoughts and ideas, there lies innovation. #mengsters

VI

TABLE OF CONTENTS
ACKNOWLEDGEMENTS .................................................................................................... II
TABLE OF CONTENTS ..................................................................................................... VI
LIST OF TABLES ........................................................................................................... VIII
LIST OF FIGURES ............................................................................................................. X
ABSTRACT ............................................................................................................. XXII
PARYLENE AND PARYLENE-BASED MEMS DEVICES ................................ 1
1.1 INTRODUCTION TO POLY(P-XYLYLENE) ............................................................ 1
1.2 PARYLENE C FOR MEMS ................................................................................. 6
1.3 PARYLENE C MICROMACHINING METHODS ....................................................... 7
1.4 PARYLENE-BASED IMPLANTABLE DEVICES ..................................................... 17
1.5 PARYLENE IN THE BRAIN ................................................................................ 19
1.6 CHALLENGES OF UTILIZING PARYLENE AS A STRUCTURAL MATERIAL ............ 21
1.7 OBJECTIVES .................................................................................................... 25
1.8 REFERENCES ................................................................................................... 26
PARYLENE-BASED EC-MEMS SENSOR FOR STUDIES OF INTRACORTICAL
PROBE INSERTION MECHANICS ............................................................... 38
2.1 BACKGROUND ................................................................................................ 38
2.2 ELECTROCHEMICAL IMPEDANCE-BASED TRANSDUCTION METHOD ................. 42
2.3 SENSOR DESIGN .............................................................................................. 45
2.4 SENSOR FABRICATION .................................................................................... 47
2.5 DEVICE PACKAGING ....................................................................................... 49
2.6 SENSOR CHARACTERIZATION .......................................................................... 50
2.7 ANNEALING EFFECTS ON SENSOR PERFORMANCE ........................................... 59
2.8 SENSOR INSTRUMENTATION ........................................................................... 64
2.9 BENCHTOP INSERTION EXPERIMENTS .............................................................. 66
2.10 INSERTION MECHANICS CONSIDERATIONS ...................................................... 73
2.11 SUMMARY ...................................................................................................... 76
2.12 REFERENCES ................................................................................................... 77
THREE-DIMENSIONAL PARYLENE SHEATH ELECTRODE ARRAY FOR
CHRONIC APPLICATIONS ......................................................................... 82
3.1 BACKGROUND ................................................................................................ 82
3.2 PARYLENE SHEATH ELECTRODE ..................................................................... 87
3.3 DEVELOPMENT OF THERMOFORMING PROCESS ............................................... 91
3.4 EFFECTS OF THERMOFORMING ........................................................................ 98
3.5 DEVELOPMENT OF PACKAGING OF PARYLENE SHEATH ELECTRODE .............. 111
VII

3.6 IMPLANTATION METHODS OF PARYLENE SHEATH ELECTRODE...................... 122
3.7 SUMMARY .................................................................................................... 137
3.8 REFERENCES ................................................................................................. 138
PARYLENE MEMS PATENCY SENSOR FOR ASSESSMENT OF
HYDROCEPHALUS SHUNT OBSTRUCTION .............................................. 149
4.1 BACKGROUND .............................................................................................. 149
4.2 SENSOR DESIGN AND OPERATION .................................................................. 155
4.3 SENSOR FABRICATION AND PACKAGING ....................................................... 157
4.4 SENSOR CHARACTERIZATION ........................................................................ 160
4.5 BENCHTOP TESTING ...................................................................................... 178
4.6 FUTURE STUDIES .......................................................................................... 183
4.7 SUMMARY .................................................................................................... 186
4.8 REFERENCES ................................................................................................. 187
CONCLUSION ......................................................................................... 193
APPENDIX A: INTRACORTICAL FORCE SENSOR PROCESS FLOW .............................. 196
APPENDIX B: PARYLENE SHEATH ELECTRODE ARRAY PROCESS FLOW .................. 198
APPENDIX C: PARYLENE-BASED PATENCY SENSOR PROCESS FLOW........................ 201
APPENDIX D: PARYLENE DEVICE RELEASE AND PACKAGING FOR ZIF CONNECTOR
.............................................................................................................. 203
VIII

LIST OF TABLES
Table 1-1. Material properties of Parylene N, C, D, and HT (adopted from [5]) ...............4
Table 2-1. Material properties for Parylene C used for FEM deflection analysis of
microchannel structure. ..........................................................................................51
Table 2-2. Table of crosstalk values obtained for sensor calibration experiments.
Crosstalk was calculated by a ratio of the percent impedance change of the site
recorded over the percent impedance change of the site indented. Mean ± SE, n
= 3. *For rows 1 and 4, n = 2. ................................................................................59
Table 2-3. Summary of insertion studies in literature exploring normal insertion forces
of intracortical probe insertion. ..............................................................................68
Table 2-4. Representative table of encountered normal forces for a single trial insertion
experiment into 0.5% agarose for different speeds. ...............................................70
Table 2-5. Table of observed distributions of impedances along the length of the probe
for two different insertion experiments..................................................................72
Table 3-1. Summary of the recent development of polymer-based neural probes. ..........86
Table 3-2. Summary of testing conditions for soak temperature and time variation
effects on chemical and mechanical properties of Parylene thin films following
thermoforming. ......................................................................................................99
Table 3-3. Summary of shrinkage percentages for contact pad and inter-pad regions as
well as total cable shrinkages for 8 and 16 channel Parylene FFCs.
a
These n
values correspond to 5 measurements taken between two individual samples. ..102
Table 3-4. Contact angle measurements obtained from Parylene films following the
thermoforming process with varying Ts (ts = 6 hrs). Results indicated that there
was no change in the surface energy with varying soak temperature. .................106
Table 3-5. Contact angle measurements obtained from Parylene films following the
thermoforming process with varying ts (Ts = 200°C). Results indicated that there
was no change in the surface energy with varying soak time. .............................106
Table 3-6. Young’s modulus measurements obtained via nanoindentation of
thermoformed films for varying Ts (ts = 6 hrs). n values correspond to number
of points on one large area sample. ......................................................................107
IX

Table 3-7. Young’s modulus measurements obtained via nanoindentation of
thermoformed films for varying ts (Ts = 200°C). n values correspond to number
of points on one large area sample. ......................................................................108
Table 4-1. Ionic formulation of artificial CSF used in benchtop experiments.
Compound weights are given for mixture into 1 L of DI water. .........................161
Table 4-2. Obtained optimal impedance measurement frequencies and sensitivities for
electrodes of the Parylene-based EC-MEMS patency sensor. .............................163
Table 4-3. Measured standard deviations (n = 4) of 6-holed catheter impedance
measurements (varying blocked hole positions) to compare efficacy of
electrode design in improving hole specificity. ...................................................171

X

LIST OF FIGURES
Figure 1-1. A schematic overview of the Gorham deposition as demonstrated for
depositing poly(chloro-p-xylylene). (1) The precursor dimer is evaporated. (2)
In the furnace, the dimer is pyrolyzed to cleave the dimer into monomer units.
(3) The reactive monomers adsorb to the substrate within the deposition
chamber and coats the samples. (4) A cold trap is also used within the setup to
trap stray unreacted monomer from entering the vacuum pump. ........................... 2
Figure 1-2. Chemical structure of poly(p-xylylene) variants with both SCS (Parylene)
and Kisco (diX) trade names................................................................................... 5
Figure 1-3. Cartoon illustrating the differences between bulk and surface
micromachining as well as the formation of isotropic and anisotropic sidewall
profiles following bulk silicon etching using a mask. ............................................ 8
Figure 1-4. (a) Cartoon representation of photolithography involving the patterning of
photoresist on Parylene. Note that for negative photoresists, the exposed areas
where the light hits are not developed away in solution (“if it hits, it sits”), and
for positive photoresists, the areas where the light hits are developed away (“if
it shows, it goes”). (b) Examples of e-beam metal evaporation, where PR is used
as a temporary mask for metal patterning and is removed with an acetone soak,
(deposition) and plasma etching, where PR used as a protective mask, (etch)
processes commonly used during Parylene micromachining. .............................. 10
Figure 2-1. An illustration of the three components of insertion mechanics during and
following insertion: (a) normal insertion force, (b) probe-tissue interfacial force,
and (c) micromotion force. ................................................................................... 40
Figure 2-2. Schematic of the Randles equivalent circuit model of an electrode-
electrolyte interface. .............................................................................................. 43
Figure 2-3. Equivalent circuit model of measuring the electrochemical impedance
between a pair of electrodes in an electrolyte. Note the use of the simplified
Randles circuit model at the two electrode-electrolyte interfaces. ....................... 44
Figure 2-4. Schematic of electrochemical impedance-based transduction method of
contact forces utilizing a flexible, mechanically compliant microchamber. ........ 45
Figure 2-5. (a) Optical micrograph of full sensor array and integrated Parylene cable.
Microchannel structure is highlighted in pink. (b) Scanning electron micrograph
(SEM) image showing top view of the sensor array with a fluidic port at the end
of the microchannel. (c) Top-down image of Parylene microchannel (pink)
indicating the electrodes and fluidic ports located at the ends. (d) SEM showing
the cross section of the 20 µm Parylene microchannel sensing structure. (e)
XI

Microelectrode layout and sensor numbering based on adjacent electrode pairs.
............................................................................................................................... 46
Figure 2-6. Operation principle of the Parylene microchannel-based interfacial force
sensor (a) pre and (b) post insertion into tissue. ................................................... 47
Figure 2-7. Overview of fabrication process for the Parylene-based EC-MEMS sensor
array. (a) Platinum electrodes and contact pads were patterned by liftoff onto a
Parylene substrate; an insulation layer was also deposited and etched to reveal
electrode sites. (b) Sacrificial photoresist was spun on and patterned to form the
microchannel structure. (c) Another Parylene layer was deposited to form the
deformable top membrane structure and fluidic ports were etched into the top
using oxygen plasma; the device perimeter was then etched using O2 plasma.
(d) Devices were released off the carrier wafer and soaked in acetone to remove
the sacrificial photoresist. Then the devices were filled with 1× PBS prior to
testing. ................................................................................................................... 47
Figure 2-8. Image showing integrated Parylene flex cable inserted into a ZIF-ZIF
connector for electrical connection to the sensor. ................................................. 49
Figure 2-9. Diagram of custom multiplexing PCB used to obtain measurements from
sensor array. .......................................................................................................... 50
Figure 2-10. Screenshots of FEM results obtained in SolidWorks illustrating (a) the
entire model and (b) zoomed in image of the first sensor..................................... 51
Figure 2-11. Load-displacement results obtained through FEM analysis of the each
sensor along the length of the microchannel. Results indicate a consistent
mechanical response regardless of sensor position along the length of a single
microchannel. The artifact in the results for sensor 7 (orange) arise from the
presence of fluidic ports in the model. .................................................................. 52
Figure 2-12. Load-displacement results obtained from FEM analysis of a single sensor
limited to a 20 µm deflection due to the sensor floor. Results indicated that the
working sensor range was between 0-40 mN. ...................................................... 53
Figure 2-13. (a) Load-displacement setup for mechanical and calibration testing of
Parylene force sensor array. (b) Optical micrograph taken using a camera
underneath testing setup illustrating deflection probe (pink) displacing into
sensor element. ...................................................................................................... 54
Figure 2-14. Impedance (a) magnitude and (b) phase plots of 18 individual sensors.
An fmeas of 1 kHz (dashed line) was selected and corresponds to maximum
resistive response at a phase value near 0°. .......................................................... 55
XII

Figure 2-15. Load-displacement plot for the Parylene microchannel sensing element
of the EC-MEMS sensor array confirming FEM results as operational
uniformity was maintained despite their differing locations along the fluidically
coupled channel. ................................................................................................... 56
Figure 2-16. Loading and unloading cycle results of (a) four different sensors at
maximum deflection and (b) at various deflection depths indicating negligible
hysteresis. .............................................................................................................. 57
Figure 2-17. Obtained calibration curve for the EC-MEMS sensor showing great
linearity within the force range of 0-60 mN. ........................................................ 58
Figure 2-18. SEM images of top-down views of (a) untreated and (b) annealed sensing
structures. (c) Representative profilometry measurement of untreated and
annealed Parylene microchannel sensing structure illustrating ~3% shrinkage
following annealing. ............................................................................................. 61
Figure 2-19. Results of load-deflection tests comparing untreated and annealed sensors
illustrating an increased structural stiffness of ~1.6x after annealing. ................. 61
Figure 2-20. EIS plots of (a) impedance magnitude and (b) phase for electrodes of
untreated (black) and annealed (blue) Parylene sensors. A change in the slope
of the capacitive region of the magnitude plot and narrowing of the phase also
in the capacitive region following annealing corresponds to a decrease in the
surface roughness of the electrode. Dashed line in (b) corresponds to the
measurement frequency chosen as the peak of the phase plot. ............................. 62
Figure 2-21. A marked difference in sensor calibration performance (reduction of
~24%) was observed due to the combination of mechanical and electrochemical
effects of annealing on Parylene sensors. Sensitivity (α) units are in normalized
impedance/mN. ..................................................................................................... 63
Figure 2-22. (a) Image of the ceramic cortical shank instrumented with fully packaged
sensor array. (b) Image of two designs of the cortical probe, sharp and flat, that
were instrumented with two different designs of the sensor array. ...................... 64
Figure 2-23. Illustration depicting the instrumentation of the sensors onto the cortical
probes. Following release of the sensor array off the wafer, the device is aligned
and affixed to the cortical probe using biocompatible adhesive. Following
attachment, the instrumented sensors are filled and ready for testing. ................. 65
Figure 2-24. Image of instrumented probes on an acrylic jig for benchtop insertion
testing. ................................................................................................................... 65
XIII

Figure 2-25. (a) Schematic of benchtop insertion testing setup. (b) Screenshot image
of user GUI during benchtop testing: (Left) a video stream of the insertion
process using a USB microscope. (Right) heat map of GUI to show the
distribution of force magnitudes along the length of the probe. Images illustrate
the progression of the instrumented probe from (i) before insertion, (ii) only tip
inserted, (iii) half of probe inserted....................................................................... 67
Figure 2-26. Representative plot of normal force (purple dash-dot) as well as
impedance responses (solid) for an insertion experiment into 0.5% agarose. ...... 71
Figure 2-27. Results from simulated micromotion experiments. Micromotion was
produced by displacing the agarose substrate ±1 mm normal to shank face.
Highlighted region contains 3 micromotion events. ............................................. 73
Figure 2-28. Illustration of probe insertion forming a triangular profile similar to a
mode 1 opening crack. .......................................................................................... 74
Figure 3-1. Images of the (a) Microwire array (adapted from [12], Copyright 2003
National Academy of Sciences, U.S.A.), (b) Utah electrode array (Reprinted by
permission from Macmillan Publishers Ltd: Nature Neuroscience [13],
copyright 2002), (c) Michigan array (© 2008 IEEE [14]), all developed
intracortical recording technologies. ..................................................................... 84
Figure 3-2. Photograph of 3D Parylene sheath probe highlighting the probe tip and a
portion of the integrated Parylene cable. .............................................................. 88
Figure 3-3. Optical micrographs of the three design iterations of the Parylene sheath
electrode: (a) the initial design with 4 external electrodes on the top surface of
the sheath, (b) revised design with the 4 electrodes moved to the outer periphery
of the probe, (c) a 1x2 PSE array with 2 sheath probes at the tip. [Scale bar of
inset is 200 µm] ..................................................................................................... 89
Figure 3-4. Overview of fabrication steps for 3D Parylene sheath probe. [Note:
fabrication steps as drawn are shown on an already cut-out substrate, however
the devices are not cut to shape during the actual process until step d.] (a) 5 µm
of Parylene was first deposited onto a carrier wafer. (b) Pt electrodes (2000 Å)
were e-beam evaporated and patterned using lift-off. Parylene insulation (1 µm)
was deposited and patterned to expose the electrode sites using oxygen plasma.
(c) A sacrificial layer of photoresist was spun on and patterned to form the
microchannel structure. (d) Parylene (5 µm) was deposited on top of the
sacrificial layer to complete the microchannel. Openings were etched into the
ends of the microchannel and the border of the device was also etched to form
the final sheath structure and shape. (e) The sacrificial photoresist is removed
using a sequential acetone, isopropyl alcohol, and DI water soak........................ 91
XIV

Figure 3-5. Top and cross-sectional drawings illustrating test samples used in
thermoforming experiments: (left) Parylene on Parylene samples and (right)
Parylene-metal-Parylene (PMP) sandwiches. ....................................................... 93
Figure 3-6. Illustration of semicrystalline, thermoplastic Parylene polymer
highlighting crystalline and amorphous regions. .................................................. 94
Figure 3-7. Optical micrographs of molds used to thermoform Parylene 3D structures.
(a) Top and (b) side-view of a microwire inserted into a Parylene microchannel
to form conical structures. (c) A PMP cable is wrapped around a glass rod to
form a macro-coil structure. .................................................................................. 94
Figure 3-8. (a) Parylene sheath probe after release from wafer with flat 2D
microchannel structure. (b) Microwire mold inserted into the microchannel for
thermoforming process. (c) Thermoformed result showing opened 3D sheath
structure................................................................................................................. 96
Figure 3-9. Sequential photographs (side-view of probe) demonstrating the mechanical
robustness of the stiffened cone structure. (a) A deflection probe positioned
above the sheath probe. (b) 100 µm displacement of the top of the sheath with
the deflection probe. (c) Retraction of the deflection probe allowed the sheath
to return to its original shape. Sheath movement is highlighted with a black
outline. .................................................................................................................. 96
Figure 3-10. Displacement (black squares) and load force (red triangles) measurements
taken during deflection studies of thermoformed PSE sheath structure. Inset and
plot indicate the (1) starting position, (2) deflection of 200 µm, and (3) return
to initial position indicating following release of load. Sheath remains flexible
and is robust to mechanical perturbations. Sheath deformation is outlined in
black to facilitate visualization. Scale bar is 200 µm. .......................................... 97
Figure 3-11. Examples of thermoformed Parylene 3D structures. (a) A thermoformed
PP conical structure with a sharp taper. (b) A thermoformed PMP cylindrical
structure with electrode sites on the inside and outside of the cylinder. (c) An
SEM image of the cylindrical PMP structure. (d) A thermoformed PMP device
formed from a Parylene microchannel with electrodes on the top. (e) A
thermoformed PMP device formed from a perforated Parylene microchannel.
(f) An SEM image of a pair of thermoformed conical devices highlighting the
repeatability of the process. [Scale bar = 200 µm] ............................................... 98
Figure 3-12. Examples of thermoformed macro-scale coils and micro-scale cones in
dual thermoformed (a) 4 channel and (b) 8 channel PMP devices. .................... 100
Figure 3-13. (a) Image of PMP device with a 3D conical structure with electrode sites
on the outer face; dashed line indicates the region with high tensile strains
XV

leading to cracked electrodes. (b) An SEM image of an electrode (white box)
illustrating the formation of cracks along the surface following thermoforming
due to high tensile strains. ................................................................................... 100
Figure 3-14. Micrograph images of contact pad regions of (a) 16 channel and (b) 8
channel PMP cables illustrating bulk shrinkage associated with the
thermoforming process. ...................................................................................... 101
Figure 3-15. Comparison between a normal PMP cable and those that underwent
thermal oxidative degradation due to a leak in the vacuum oven. Note the
discoloration of the films as well as brittle failure in the PP sample during
handling............................................................................................................... 103
Figure 3-16. Absorbance measurements from FTIR analysis of thermoformed samples.
Varying the (a) soak temperature and (b) soak time had no effect on the
chemical composition following the thermoforming process. ............................ 104
Figure 3-17. Absorbance results from FTIR measurements of untreated Parylene,
Parylene thermoformed under vacuum, and Parylene thermoformed in an
oxygen rich environment (leak in vacuum oven). Thermoforming within an
oxygen rich environment results in an additional carbonyl peak that is not
present in samples thermoformed under vacuum. .............................................. 105
Figure 3-18. Representative CV curves of untreated (blue dashed) and thermoformed
(black solid) electrodes. ...................................................................................... 109
Figure 3-19. Representative plots impedance magnitude (a) and phase (b) of EIS
measurements of untreated (red triangles) and thermoformed (black squares)
electrode. ............................................................................................................. 110
Figure 3-20. Representative 500 msec trace from a 300-Hz high-pass filtered
electrophysiological record from thermoformed PSE at 21 days post-
implantation into the rat motor cortex. ............................................................... 110
Figure 3-21. (a) As-fabricated Parylene sheath probe highlighting the contact pads for
integration with ZIF connectors. (b) Photograph of fully packaged dual-probe
array with two probes secured into 2 ZIF connectors on the flexPCB. .............. 112
Figure 3-22. (a) Optical micrograph of Parylene flexible cable indicating the contact
pad terminated ends for insertion into a commercial ZIF connector. (b)
Schematic of operation of ZIF connector. .......................................................... 114
Figure 3-23. (a) Image of bare Wing PCB prior to cutting and bending to shape. (b)
Formed Wing PCB following heat treatment to form bent orientation designed
XVI

to allow for connection to 4 PSEs. (c) Wing PCB inserted into the rigid ZIF-
Omnetics board that allowed for integration with Plexon system. ..................... 114
Figure 3-24. Optical micrograph of the California PCB and inserter tool used for initial
in vivo insertion experiments. ............................................................................. 115
Figure 3-25. Image of bare hFlex initially designed to allow for 4 PSEs; the arm piece
could fold back to allow for all 4 PSEs to be implanted at once. ....................... 116
Figure 3-26. Image of halved hFlex and wire soldered to the body for use as the
GND/REF used for 28 day study. ....................................................................... 117
Figure 3-27. Image comparing the older hFlex (left) and the newer, shorter hFlex
(right). ................................................................................................................. 118
Figure 3-28. Image of new flexible ZIF-Omnetics PCB for more robust connections to
animals during neurophysiological measurements. ............................................ 119
Figure 3-29. Image of the FlexPCB used to form the PSEA from two 1x2 arrays. A 16
channel ZIF connector was used to connect each 1x2 array (one on each side)
of the FlexPCB, which has leads to a BM10 receptacle connector (in black). ... 120
Figure 3-30. Image of the PSEA FlexPCB and the external BM10-Omnetics flexible
adaptor board used for the chronic in vivo studies. ............................................. 121
Figure 3-31. (a) Optical micrograph of side-mounted probe to microwire. (b) Optical
micrograph of tip-mounted probe to microwire.................................................. 123
Figure 3-32. Image of first inserter tool design with 4 tapered microwires soldered to
a larger handling cylinder. .................................................................................. 125
Figure 3-33. Silicon-based inserter tool with a PSE-California PCB affixed with PEG
for insertion. ........................................................................................................ 125
Figure 3-34. (a) A custom microwire-based introducer tool used for implantation for
dual-probe array. (b) The dual-probe array was temporarily affixed to the
introducer tool using PEG, which dissolved following insertion. PEG was used
in 2 locations: cable region for strain relief and support and probe-microwire
tip for stiffness during insertion. ......................................................................... 127
Figure 3-35. Image sequence of a representative benchtop agarose insertion test. (a)
The array-tool assembly was attached to stereotaxic apparatus and positioned
over the spot of interest. (b) The array-tool assembly was inserted into the
agarose by hand, and 1× PBS applied to begin the dissolution of PEG. (c) The
dual-probe array was detached from the tool to allow for tool withdrawal; note
XVII

that the Parylene cable was still attached to the tool. (d) After complete
dissolution of PEG (5 minutes), the inserter tool was withdrawn from the
agarose, and the dual-probe array remained implanted. ..................................... 128
Figure 3-36. Optical micrograph of PSEA attached to an acrylic inserter tool modified
to include a fiduciary marker wire for automated motorized insertions. ............ 129
Figure 3-37. Image sequence of benchtop insertion test indicating that the probe
withdrew ~90 m following the tool removal. (a) A photograph taken after
insertion but prior to tool withdrawal where the probe tip depth is indicated by
the black dotted line. (b) A photograph following tool withdrawal where the
new probe tip depth is indicated by the white dotted line................................... 130
Figure 3-38. Image of rat implanted with PSEA. The bone cement head cap and top
portion of the FlexPCB can be seen atop the animal’s head............................... 132
Figure 3-39. Sequence of images illustrating the in vivo implantation process for the
PSEA. (a) The PSEA-tool assembly is lowered to the surface of the cortex
following dura removal. (b) The PSEA-tool is inserted into the cortex to the
required depth. Bone cement is also added to the rear of the PSEA to hold the
array in place following removal of the inserter tool. (c) The inserter tool is
removed following application of saline and gel foam is placed over the
craniotomy. (d) Lastly a head cap is formed using bone cement. ....................... 133
Figure 3-40. Sample microphotographs of the cerebral cortex double-stained with
antibodies for NeuN (brown) and GFAP (dark-blue) through the following
probe tips: (a) sharp, (b) moderate, and (c) blunt. The 10 µm sections were cut
perpendicular to the probe track (the probe was removed during brain
dissection). .......................................................................................................... 135
Figure 3-41. Quantification of (a) neuronal and (b) astrocytic density for different
probe tip shapes (blunt, moderate, and sharp) at 28 days (n = 13 animals),
indicating the best performance out of the sharp design..................................... 135
Figure 3-42. Image of shear-induced failure of the FlexPCB due to repeated mating
cycles. This can be improved by using a thicker FlexPCB material for the
headstage. ............................................................................................................ 136
Figure 4-1. (a) Cartoon of 2D view of ventricular system within the brain comprising
of the lateral ventricles, third ventricle, and fourth ventricle connected via
narrow passages. Choroid plexus is shown in red. (b) Cartoon anatomy of brain
illustrating the pathway of CSF from its production within the ventricles, and
towards the subarachnoid space (SAS) where it is absorbed. ............................. 150
XVIII

Figure 4-2. (a) Implantable Medtronic Delta valve shunt system consisting of ports for
proximal and distal catheters to control CSF flow from the ventricles. (Image
modified from [8]) (b) Comparison of distal catheter placement (pink) VP
shunts (peritoneal cavity) and VA shunts (atrium of the heart). ......................... 152
Figure 4-3. (a) Equivalent circuit model of two electrodes in an electrolyte,
highlighting the solution resistance. (b) Conceptual cartoon of impedance
sensing mechanism of patency sensor. Two electrodes on the internal and
external surfaces of the catheter are fluidically connected via the drainage ports.
Obstruction of these ports impedes the ionic conduction path between the
electrodes, and (b) the electrochemical impedance between the electrodes
increases for measurements above a certain frequency (fmeas). ........................... 156
Figure 4-4. Process flow for fabrication of Parylene patency sensor. (a) Parylene C (12
µm) is deposited on a silicon carrier wafer. (b) Pt electrodes (2000 Å) are
deposited via e-beam evaporation and patterned using liftoff. (c) An insulation
layer of Parylene C (12 µm) is deposited and electrode sites are exposed using
oxygen plasma etching. (d) Devices are released from wafer by gentle peeling
while immersed within DI water. ........................................................................ 157
Figure 4-5. (a) Optical micrograph of v1 Parylene device with 4 electrode designs to
see the effect of electrode size on performance. (b) Electrically packaged v1
Parylene device using a ZIF connector and integrated flat flexible cable (FFC).
(c) Electrically packaged v2 Parylene device with single electrode design and
biocompatible epoxy (yellow) for encapsulation. ............................................... 159
Figure 4-6. Fluidically packaged sensors in (a) cap and (b) inline modules used for
benchtop testing. The inline module was designed for sensor integration with
external ventricular drainage systems for clinical validation studies. In this form
factor, sensors are curled within the lumen to permit uninterrupted flow. ......... 159
Figure 4-7. (a) Mock silicone catheters (ID = 1 mm) with varying number of holes
used for benchtop testing. (b) Magnified image of manually punched hole (Ø =
1 mm) using a 15 gauge coring needle. .............................................................. 160
Figure 4-8. (a) Image of experimental setup demonstrating fluidic connection between
the catheter and sensor module. (b) Experiments were conducted using either
the (i) cap or inline module with a (ii) syringe or peristaltic pump for static and
flow conditions, respectively. Impedance was measured between the sensor
electrode and platinum ground electrode in the beaker. ..................................... 161
Figure 4-9. (a) Electrochemical impedance spectroscopy results of the impedance
magnitude of the E1 electrode indicating variations for measurements
frequencies > 10 kHz between different catheter blockages. (b) Electrochemical
impedance spectroscopy results of the impedance phase of the four electrode
XIX

sizes demonstrating varying optimal measurement frequencies (where phase =
0°). Dotted line indicates 10 kHz, the fmeas observed for the E1 electrode. ........ 163
Figure 4-10. Representative patency curve obtained for the E1 electrode within the cap
module indicating an inverse relationship between measure impedance
magnitude and the number of open holes (percent blockage). ........................... 164
Figure 4-11. Patency curve obtained for v2 sensor within inline module confirming a
similar shape to v1 results with the E1 electrode within the cap module. .......... 165
Figure 4-12. (a) Two different catheter hole orientations (normal and reversed) used to
assess the transduction mechanism of the sensor. (b) Resultant patency curves
of the two catheter orientations indicate that the sensor was not measuring the
number of open holes, but rather the position of the first open hole, following a
“path of least electrochemical resistance” effect. ............................................... 166
Figure 4-13. (a) Two different catheter sets (normal and alternate) used to assess the
dependency of the sensor response on the first hole. Total number of open holes
are given above each catheter drawing. Note that each catheter between the sets
has the same top-most open hole position but a different number of open holes
(e.g. 4 holed-catheter of the Normal set has the same top-most hole position of
the 6-holed catheter in the Alternate set). (b) Resultant patency curves of the
two catheter sets indicate that even with varying number of total open holes, the
measured impedance is dependent only on the top most hole, following a “path
of least electrochemical resistance” effect, and thus a more linear response in
the case for blockages from the top of the catheter downwards. ........................ 167
Figure 4-14. 4 electrode orientations designed to improve the hole specificity of the
sensor. ................................................................................................................. 168
Figure 4-15. (a) Impedances measured using the inline module for (b) a 4 catheter set
with six open holes of varying positions. Note that impedances are constant for
catheters 2-4 while catheter 1 has a large increase in impedance due to the
patency of the top-most hole. .............................................................................. 169
Figure 4-16. Patency curves obtained for the 4 electrode orientations: (a) orientation
A, (b) orientation B, (c) orientation C top pair, (d) orientation C bottom pair,
(e) orientation D top pair, and (f) orientation D bottom pair. These plots
highlight the measured impedances for the 6-holed catheter set given in Figure
4-15b. .................................................................................................................. 170
Figure 4-17. Images of explanted proximal catheters that indicate the tissue obstruction
within the top most holes. (Reproduced with permission from the Journal of
Neurosurgery [26]).............................................................................................. 172
XX

Figure 4-18. (a) Time-lapse images (1 →4) of a dye drop experiment into a 16-holed
catheter within aCSF under flow (0.3 mL/min). Note that though the dye was
placed at the bottom holes, a blue dye stream is not observed until the diffused
dye reaches the top hole (image 3). Even in image 4, though the dye is within
the proximity of the other holes, dye flow is still only observed through the most
proximal hole. (b) Magnified image of catheter illustrating the blue flow stream
into the catheter only at the most proximal hole (white circle). ......................... 172
Figure 4-19. (a) Patency curves obtained for the inline modules for varying
temperatures 32–44°C, indicating a decrease in baseline, but retaining a similar
sensitivity. (b) Within clinically relevant temperatures between 34-38°C, no
large variations were recorded. ........................................................................... 174
Figure 4-20. Patency curves measured for inline modules indicate that there is a
decrease in impedance of ~8% with the addition of flow, but no variation among
flow rates within the range of 0.03-0.6 ml/min. ................................................. 175
Figure 4-21. Measured impedance of patency sensor for a 16-holed catheter in a bottle
brain within flow for 14 days demonstrating low drift of the sensor response. .. 176
Figure 4-22. Hydrogen peroxide plasma sterilization had no effect on the patency curve
of the inline module. ........................................................................................... 177
Figure 4-23. (a) Electrochemical impedance spectroscopy and (b) cyclic voltammetry
measurements of patency electrode packaged in inline module pre and post
H2O2 plasma sterilization. Results indicate to changes to the electrode surface
properties following sterilization. ....................................................................... 178
Figure 4-24. Image of catheter pinching experiment to model obstruction of the lumen
of the catheter using a stylus. .............................................................................. 179
Figure 4-25. (a) Observed sensor response to catheter deflection using a stylus to close
off the lumen. (b) Results indicate that the sensor response varies inversely due
to changes in the cross sectional area between the electrodes due to lumen
obstruction........................................................................................................... 179
Figure 4-26. (a) Transient blockage experiments of sheathing/unsheathing a 16-holed
catheter illustrated (b) real-time measurement capabilities of the sensors.
Obstruction events are labeled with an X. .......................................................... 180
Figure 4-27. Dynamic obstruction study using PEG coated mock catheters in aCSF,
demonstrating tracking of PEG dissolution (“reverse blockage”) using sensor.
(a) In the first experiment, all holes of the catheter except for the bottom-most
hole were covered with PEG and sensor response was measured during PEG
dissolution. (i) PEG coated catheter is placed in aCSF, (ii) PEG dissolves from
XXI

the top-most hole, (iii) top hole is open, (iv) all holes are open. These results
confirm the dependence of the sensor on the first hole. (b) The second
experiment revealed a similar response although only the top most hole was
covered with PEG. .............................................................................................. 181
Figure 4-28. (a) Patency curve for device S12 tested in human CSF solution indicating
preservation of performance. (b) 5 hr drift measurement of sensor response in
configuration for use with EVD systems indicating a slow downward drift of
5% over 5 hours. Additional studies are required for further validation of sensor
performance ........................................................................................................ 183
Figure 4-29. (a) Cartoon schematic and (b) image of inline module packaged sensor
for integration with external ventricular drains (EVDs) within clinics. ............. 184
Figure 4-30. (a) 3D model of implantable packaging envisioned for the sensor for
placement in between the proximal catheter and valve. (b) Cross sectional view
of the implantable packaging. (c) Images of 3D printed implantable module to
scale for initial testing. ........................................................................................ 186

XXII

The development of implantable devices for biomedical applications has been
heavily galvanized by the introduction of microelectromechanical systems (or MEMS)
technologies. Stemming from efforts within the semiconductor industry, MEMS enables
the miniaturization of devices while maintaining, and in some cases, improving their
functionality. MEMS are ideal for implantable applications, as smaller devices can reduce
the foreign body response once implanted and allow for better integration within the body.
However, though these micro-devices are a significant improvement over previous implant
technologies, there is still much room for improvement.
Current implants are marred by rigid substrates such as metals, ceramics, and
silicon that create a mechanical mismatch between the implant and the surrounding tissue
and shortens the implant lifetime. In addition, typical MEMS devices are sensitive to wet,
saline in vivo environments due to the materials used, which can affect chronic
performance within the body. The following work centers on the development of Parylene
C-based technologies as an improved paradigm for implantable devices due to their
superior material properties and integratable capabilities. Parylene C is a USP class VI
polymer (highest possible rating of biocompatibility for plastics) and is compatible with
micromachining processes, which enables the construction of MEMS devices using a
Parylene C thin film substrate and architecture. In addition, Parylene C’s low Young’s
modulus (~2-3 GPa) and thus flexibility compared to silicon or metal allows for improved
tissue integration as well as simple integration with current implant hardware to add
functionality.
The following work details Parylene C-based approaches to filling the needs for
two thrusts within the cortical research field: namely, improving chronically implanted
intracortical electrode technologies for neural prosthetics and improving current
ABSTRACT
XXIII

hydrocephalus treatment. Chapters 2 and 3 delineate two Parylene C-based approaches in
improving the chronic reliability of implanted intracortical electrode technologies. A
Parylene C-based force sensor array is demonstrated in chapter 2 that leverages the
flexibility of Parylene C for use as a deformable membrane to measure tissue forces
enacting on the implanted electrode. The thin-film and low-profile nature of the sensor
array also allows for easy integration of the sensor array on the face of these intracortical
probes to assess implantation and in vivo mechanics once implanted. These sensors will
enable, for the first time, quantitative measurement of interactional forces between the
probe and tissue in situ. It is the hope that studies with these instrumented sensors will help
establish design parameters for probe size, shape, and electrode placement for more
improved solutions for chronically implanted intracortical electrodes.
In chapter 3, a secondary approach in improving the reliability of chronically
implanted cortical electrodes is presented, in the development of an electrode technology
fabricated on a Parylene C substrate. As a solution to the mechanical mismatch between
the state-of-the-art, rigid metal or silicon probes and the surrounding neural tissue, a
flexible, thin-film Parylene-based approach is presented and implemented in rat models to
assess chronic performance. In the development of this device, a thermoforming process is
introduced where three dimensional (3D) structures are constructed from planar Parylene
devices using thermal shaping. The effects of the process on the material and chemical
properties of the polymer are also presented to further characterize this post-process as a
new technique in establishing 3D Parylene structures. Electrical packaging and implant
strategies are also discussed to create a basis for implementation of these flexible devices
for in vivo applications.
Chapter 4 discusses a Parylene C-based approach in improving current
hydrocephalus treatment. Hydrocephalus is a chronic condition of the excess accumulation
of fluid within the brain, usually treated by implanting a catheter system (“shunt”) that is
marred by a failure rate of 40% within the first year due to catheter obstruction. Though
failure rates are high, there are no current practical methods to assess shunt patency;
XXIV

physicians rely on nonspecific symptomatic indicators (e.g. headaches, nausea) to report
shunt failure, which pose severe risks for the patient. Towards an improved technique to
assess the failure of these implanted shunts, a Parylene-based patency sensor designed to
integrate with shunt hardware is presented. The sensor adds continuous monitoring
functionality to current treatment methods to assess shunt dynamics and treatment efficacy.
The present work sets the groundwork for clinical packaging and testing for this Parylene-
based device that can be extended to any device for use within the clinic.
Within this work, is it the author’s aim to not only demonstrate the capability of
Parylene C-based devices to fulfill needs for implantable applications, but also extend the
basis of knowledge on the processing, packaging, and implantation strategies for Parylene-
based approaches within the MEMS field. It is the hope of the author that the following
discussions may aid in the advancement of Parylene-based MEMS technologies and
encourage the use of Parylene as a substrate material in the development of novel
implantable biomedical devices.

1

1.1 Introduction to poly(p-xylylene)
Poly(p-xylylene), the chemical name of Parylene, was first synthesized by Michael
Mojzesz Szwarc in 1947 in a study of the by-products of thermal decomposition of p-
xylyene [1]. This process however demonstrated poor yield as well as secondary gaseous
byproducts non-ideal to mass polymerization. It was not until the development of a stable
dimer precursor and optimization of a chemical vapor deposition (CVD) polymerization
process by William Gorham at Union Carbide [2] that a commercially viable form of the
material was introduced.
The Gorham process (Figure 1-1), which is carried out under vacuum to increase
the mean free path to the substrate to be coated, begins with a granular dimer precursor, di-
p-xylylene (or [2.2] paracyclophane). This precursor is vaporized via sublimation and then
pyrolyzed at a temperature above 550°C to cleave the dimer into its reactive radical
monomer. Within the deposition chamber (at room temperature), the reactive monomer
adsorbs to all exposed surfaces and begins to spontaneously polymerize to form conformal,
poly(p-xylylene) films. A thorough review of the deposition and polymerization process,
as well as a proposed chemical model, is given by Fortin et al. [3]. This process enables
the control of deposition parameters (vaporizer, pyrolysis temperatures and chamber
pressure) and also allows for the full conversion of the dimer into the polymer film without
any byproducts. Another advantage of this process is that the substrate is coated at room
temperature, allowing compatibility with thermally sensitive materials.

PARYLENE AND PARYLENE-BASED MEMS DEVICES
2

Figure 1-1. A schematic overview of the Gorham deposition as demonstrated for
depositing poly(chloro-p-xylylene). (1) The precursor dimer is evaporated. (2) In the
furnace, the dimer is pyrolyzed to cleave the dimer into monomer units. (3) The reactive
monomers adsorb to the substrate within the deposition chamber and coats the samples. (4)
A cold trap is also used within the setup to trap stray unreacted monomer from entering the
vacuum pump.

Currently, the commercial market for poly(p-xylylenes) is dominated by two
industry leaders, Specialty Coating Systems (SCS; “Parylene” trade name) and Kisco
Conformal Coating, LLC (“diX” trade name), with each manufacturing their own dimer
sources. For both companies, the breakthrough of the Gorham process to efficiently
produce poly(p-xylylenes) contributed to the synthesis of other chemical variants of
poly(p-xylylenes) with added functional groups on the monomer unit. Though these
variants utilize the same Gorham polymerization process (but with different starting
dimers), they have different material properties due to the functional groups and are thus
used for different applications.
There are more than 10 commercially available variants of poly(p-xylylenes) to
date (Figure 1-2). The most commonly used within literature are Parylene N (the same
poly(p-xylylene) discovered by Szwarc), Parylene C (poly(chloro-p-xylylene)), Parylene
D (poly(dichloro-p-xylylene)), and Parylene HT, also named AF-4 (poly(tetrafluoro-p-
3
xylylene)), all from SCS. A brief comparison of the material properties of these four
polymers is detailed in Table 1-1. In short, Parylene N is primarily used as a dielectric and
has been shown to have excellent crevice penetration ability, but has the slowest deposition
rate. Parylene C has been the most popular for biological applications because it was the
first variant to attain the ISO-10993, USP class VI rating (the highest biocompatibility
rating for plastics), and has excellent water and gas barrier properties compared to Parylene
N. It is important to note that Parylene N and Parylene HT have also since received the
ISO-10993, USP Class VI rating. Parylene D has greater thermal stability but otherwise
similar properties to Parylene C and more recently has been replaced in use with the
discovery of Parylene HT. Parylene HT is becoming more popular, largely because of its
improved properties: lower dielectric constant, higher ultraviolet stability, better crevice
penetration, higher thermal stability, and lower moisture absorption. However, the material
requires substrate cooling for acceptable deposition rates and yield [4].

4

Table 1-1. Material properties of Parylene N, C, D, and HT (adopted from [5])
Property Parylene N Parylene C Parylene D Parylene HT
Dielectric Strength
(V/mil), 1 mil film
7,000 5,600 5,500 5,400
Dielectric Constant
60 Hz 2.65 3.15 2.84 2.21
1 kHz 2.65 3.10 2.82 2.20
1 MHz 2.65 2.95 2.80 2.17
Young’s Modulus (GPa) 2.413 2.758 2.62 2.551
Index of Refraction 1.661 1.639 1.669 1.559
Yield Strength (MPa) 42.06 55.16 62.05 34.47
Elongation to Break (%) Up to 250 Up to 200 Up to 200 Up to 200
Coefficient of Friction
Static 0.25 0.29 0.33 0.145
Dynamic 0.25 0.29 0.31 0.130
Density (g/cm
3
) 1.10-1.12 1.289 1.418 1.32
Melting Point (°C) 420 290 380 >500
Thermal Conductivity at
25°C (W/[m·K])
0.126 0.084 - 0.096
Gas Permeability at 25°C,
(cc·mm)/(m
2
·day·atm)

N 2 3.0 0.4 1.8 4.8
O 2 15.4 2.8 12.6 23.5
CO 2 84.3 3.0 5.1 95.4
H 2 212.6 43.3 94.5 -
Water Absorption
(% after 24 hours)
<0.1 <0.1 <0.1 <0.01
Specific Heat at 20°C
(cal/g·C)
0.20 0.17 - 1.04

5

Figure 1-2. Chemical structure of poly(p-xylylene) variants with both SCS (Parylene) and
Kisco (diX) trade names.

There are other, more reactive p-xylylenes, with functional groups that make them
ideal for specific applications, such as Parylene A (diX A and diX AM from Kisco) [6, 7]
that contains amine groups for better adhesion performance with biological specimens.
Other work within literature have synthesized polymer films with other active functional
groups such as a carbonyl [8], aldehyde [9], ester [10], anhydride [10], alkyne [11, 12],
alcohol [13], ethyl [14], and photoactivatable phenylacetyl [15] that can provide improved
material properties and under additional chemical reactions, further chemically modify the
polymer surface. Overall, many chemical poly(p-xylylene) variants have been produced
for different applications, but the polymer predominantly used for biomedical applications
6
of microelectromechanical systems (MEMS) is Parylene C because of its properties
discussed in the next session, and thus is the focus of this work.
1.2 Parylene C for MEMS
Regardless of the variant, Parylenes have ideal properties for barrier applications
as their benzene backbones make them chemically inert, and the conformal, uniform
deposition process allows for optimal coverage for encapsulation. Due to these
characteristics, in addition to its high dielectric strength, Parylene C first found use as a
coating for implantable electronics [16, 17], following early experiments that validated the
polymer’s biocompatibility [16]. After the introduction to the MEMS community in 1997
as a coating material for fluidic interconnects [18], Parylene C gained popularity as a
MEMS material due to the advantages of its deposition process and compatibility with
standard micromachining and photolithographic processes. Its gas phase, pinhole-free
polymerization at room temperature and effective gap fill made the coating process
compatible with a variety of MEMS materials and structures for applications as a coating
(e.g. antistiction coating due to Parylene C’s low coefficient of friction) and a structural
material. Parylene C has also been shown to deposit with low to no intrinsic stress, though
stresses in the film can increase following processing methods that apply heat and plasma
[19]. The ability to form smooth films that incur little optical scattering and high
transmittance in the visible spectrum makes Parylene C amenable to applications requiring
optical transparency [20-23]. Currently, free film Parylene C devices are increasing in
popularity compared to other polymer (e.g. polyimide (PI) and poly (di-methyl siloxane
(PDMS)) devices due to its simpler deposition process and combination of high flexibility
and mechanical strength.
Specifically for biomedical MEMS (i.e. bioMEMS), Parylene C has been widely
adopted for its proven biocompatibility (ISO-10993, USP class VI rating) and chemical
and solvent inertness which are imparted by its chemical structure. As the deposition
process does not require any additives (unlike epoxies) and has no harmful by-products,
Parylene C has been the standard for the coating of implantable devices as well as a
7
structural MEMS material for biomedical devices. Numerous published studies have tested
the biocompatibility of Parylene C both in vitro [24] and in vivo [17], and its biostability,
low cytotoxicity, and resistance against hydrolytic degradation have been strong arguments
for its use as a biomedical material [25-27].
1.3 Parylene C micromachining methods
Prior to a discussion on the present work on the development of Parylene C-based
MEMS technologies, it is beneficial to familiarize the reader with background on
fabrication techniques with regards to the processing and micromachining of Parylene C
(hereon referred to as Parylene).
1.3.1 Fundamentals of microfabrication
Research in the development of microfabrication techniques and technologies that
comprise the field of MEMS began with efforts within the semiconductor industry with the
demonstration of the first transistor at Bell Laboratories in 1947. However, it wasn’t until
a seminal talk given by physicist and Nobel laureate Dr. Richard Feynman, entitled
“There’s Plenty of Room at the Bottom” to the American Physical Society in 1959 that the
intellectual endeavor of micro-scale electro-mechanical machines was launched. Within
this talk, Dr. Feynman detailed challenges to the scientific community to develop new
equipment and processes to scale down technologies (and even text!) to micro and nano-
sizes that only existed within the imaginations of researchers. With this colossal paradigm
shift in creating devices at such small scales, the field of MEMS was born and the
exploration of such devices and technologies began.
In practice, microfabrication processes can be broken down into two categories:
bulk and surface micromachining (Figure 1-3). Bulk micromachining involves processing
directly on the substrate material itself, usually silicon, to construct a MEMS device, while
surface micromachining utilizes the substrate as a carrier or a base that the device would
be fabricated on. The latter consists of the deposition and patterning of various layers of
materials to build structures on a layer-by-layer approach to realize devices, which can then
8
be released from the carrier wafer. Micromachining of Parylene-based devices, at least
within this work, largely involve surface micromachining on Parylene layers on a silicon
carrier wafer.

Figure 1-3. Cartoon illustrating the differences between bulk and surface micromachining
as well as the formation of isotropic and anisotropic sidewall profiles following bulk silicon
etching using a mask.

Micromachining techniques largely involve a combination of three techniques to
realize MEMS devices: photolithography, etching processes, and deposition processes
(Figure 1-4).
1. Photolithography: Photolithography is the process by which a photo-sensitive
material is deposited and patterned on a wafer for a variety of steps in
microfabrication, including use as a mask for etching/deposition steps or to
create sacrificial or structural elements of the device. This process involves the
use of a photo-patternable organic polymer, commonly known as photoresist
(PR), which can be spun onto a wafer and patterned with UV light. Depending
on the type of the photoresist, i.e. positive or negative tone, the UV light can
either make exposed regions more soluble (positive) or insoluble (negative)
within a developer solution. After UV exposure, the PR coated wafers are
“developed” within this solution to realize patterns for subsequent fabrication
steps.
9
2. Etching processes: Etching (or removal) processes involve steps to remove
material from a wafer by using physical, chemical, or a combination of both
methods. Etching is divided into two categories depending on the technique:
wet etching involves the use of immersion in a solution to remove materials on
the wafer, while dry etching typically has plasma-based techniques for removal
without the need for wet solutions. The choice of etching method largely
depends on the desired material to be etched. It is important to also choose the
proper masking material on top of the layer to be etched to ensure that the
desired areas to be etched are removed and those that are to remain are protected
from the etch. These processes can either produce isotropic or anisotropic
sidewall profiles (Figure 1-3) also depending on the method chosen.
3. Deposition processes: Deposition (or additive) processes involve steps to
deposit a layer of material on the wafer. Deposition can be accomplished by a
variety of steps including evaporation (for metals), chemical vapor deposition
(CVD; for some polymers including Parylene), and pouring/spinning (used for
PR). The choice of deposition process varies on the compatibility of the process
with the material to be coated (e.g. thermal compatibility). Adhesion of the
deposited material on the underlying layer is also important to consider. Certain
deposition techniques can have great step-coverage or “conformality”
depending on the material and technique used. Such is the case of the deposition
of Parylene, which has outstanding conformality during its CVD deposition.

10

Figure 1-4. (a) Cartoon representation of photolithography involving the patterning of
photoresist on Parylene. Note that for negative photoresists, the exposed areas where the
light hits are not developed away in solution (“if it hits, it sits”), and for positive
photoresists, the areas where the light hits are developed away (“if it shows, it goes”). (b)
Examples of e-beam metal evaporation, where PR is used as a temporary mask for metal
patterning and is removed with an acetone soak, (deposition) and plasma etching, where
PR used as a protective mask, (etch) processes commonly used during Parylene
micromachining.

With a basic understanding with these three processes used for microfabrication, a
brief review of fabrication techniques on the micromachining of Parylene can be presented
in the following sections.

11
1.3.2 Parylene etching processes
Etching techniques for Parylene are limited to physical and dry processes largely
due to the high solvent inertness of the polymer. There have been reports of the wet etching
of Parylene using chloronapthelene or benzoyl benzoate [28], but only at extreme
temperatures (>150°C), which can make this process incompatible with photolithography
steps. Dry etching techniques have been found to be the most effective and practical to etch
Parylene. For a more detailed review of plasma etching of Parylene, the author refers the
reader to Meng et al. [29].
Parylene etching can be accomplished using oxygen-based plasma etching [29-32],
reactive ion beam etching (RIBE) [33], and reactive ion etching (RIE/deep RIE or DRIE)
[29, 34, 35]. For these methods, the etching mechanism involves benzene ring opening
using reactive oxygen radicals [36, 37]. These techniques produce an isotropic etch profile,
as RIE and DRIE methods can create aspect ratios of 2:1, while plasma etching is limited
to a 1:1 ratio. A switched chemistry etch that involves cycling through (1) deposition of
C4F8-based Teflon like polymer as a sidewall passivation layer, (2) etching in SF6 plasma,
and (3) etching in O2 plasma was found to improve the anisotropy to produce fairly vertical
sidewalls [29].
Many materials have been explored as etch masks for Parylene including
photoresists, aluminum, oxides, spin on glass, nitride, and α-silicon [29, 38-40], but
photoresist and sputtered aluminum remain the most popular, due to ease of patterning and
hard mask qualities, respectively. However, the etch selectivity between photoresist and
Parylene is very low (1:1) and may not be optimal when etching thick Parylene layers (>10
µm) [29]. Aluminum etch masks are not compatible with DRIE processes, as they have
been shown to be sputter and redeposit during the etch process [41]. It is important to
consider these factors during process design.
Other methods for Parylene removal include laser ablation (266 nm) [42-44], one
of the first techniques for Parylene etching to expose electrode sites, and manual removal
by peeling. For the latter, pre coating of release agents, such as 2% Micro-90 lab cleaning
12
solution (International Products Corporation, Burlington, NJ), on the substrate as well as
immersion in water [45-47] can aid in the removal of the film without damage.
1.3.3 Parylene deposition techniques
As the CVD of Parylene is a relatively simple and tunable process, many variants
of the routine coating method have been investigated to achieve novel films and structures.
A prominent technique within literature is the deposition of Parylene onto molds to create
structures. Three-dimensional (3D) devices have been constructed by depositing Parylene
onto structural molds (e.g. patterned photoresist, bulk etched silicon) to form hemispherical,
bump electrodes [48], pockets for silicon chips [49], and 3D micro electrode arrays [50,
51]. One work of note utilized a two-photon polymerization (TPP) process to create 3D
nano/micro structures out of photoresist with high resolution (< 100 nm) and coated the
structures with Parylene to form precise 3D Parylene structures [52]. Super-hydrophobic
Parylene films (contact angles of ~155°) were also constructed by depositing Parylene onto
a silicon mold with micro/nanostructures to attain advantages of a super-hydrophobic
surface while maintaining the transparency and flexibility of Parylene [53]. Other works
have looked into the use of more inexpensive PDMS molds to develop bond-less, tape-
capped Parylene-based microfluidics [54] as well as 3D penetrating microelectrodes [55].
Beyond molds, Parylene deposition onto different surfaces has also been used to
fabricate unique films, such as the Parylene on liquid deposition (PoLD) technique [56],
also known as the solid on liquid deposition (SOLID) process [57] developed around the
same time, involving the deposition of Parylene on liquids. By depositing Parylene on a
low vapor pressure liquid between 1 and 7 Pa (e.g. glycerin, silicone), Parylene
encapsulated liquids are formed, which can be used for complex optical devices, including
microliquid lenses [21], liquid prisms [22], and a micro droplet array for displays [20].
Alternatively, the liquid can serve as a sacrificial layer to fabricate microfluidic devices
without the need for molds, polymer sacrificial structures, or channel bonding [58]. A
closer inspection of the polymer interface between the Parylene and liquid reveals a rough,
porous surface, unlike the planar, smooth surface on the other side, the properties of which
13
can vary with the liquid substrate that was coated [59]. These porous surfaces have been
utilized to form gas exchange membranes [59].
A more in-depth look at the CVD and polymerization of Parylene C reveals that the
deposition reaction is under kinetic control. That is, the rate-limiting steps for the reactions
are the (1) the adsorption of the monomer onto the substrate, (2) surface migration and bulk
diffusion of the monomer, and (3) chemical reactions involving the initiation or
propagation of the polymer [3]. By devising deposition processes that interfere with these
different steps, films with different properties can be synthesized including: a porous
Parylene film for use as an ultrafilter that uses evaporating glycerin vapors during the
deposition process to hinder polymer growth [60], the deposition of Parylene nanofibers
using oblique angle polymerization (OAP), and a membrane template technique that
physically hinders the diffusion of the monomer units [61]. In another example, von
Metzen et al. used an aperture to impede the diffusion of monomer units, as a technique
for controlled, tapered deposition of Parylene [62].
Furthermore, processes that inhibit these deposition mechanics enable selective
deposition of Parylene. In most cases, heat can be used to prohibit the deposition of
Parylene using heaters fabricated on the substrate [63, 64], as a localized temperature
increase (>140°C [64]) can reduce the deposition rate in that region [65, 66]. Another
technique utilizes deposited transition metals (e.g. iron, gold, silver, platinum) in the form
of evaporated thin films, salts, and organometallic complexes to delay the initiation and
propagation steps of Parylene polymerization [67]. The mechanism for this inhibition relies
on deactivation of the radical monomers once they are adsorbed to the metal surface,
prohibiting initiation of polymer growth [67]. The efficacy of this method varies with
choice of transition metal (group 8 transition metals were the most effect at inhibiting
polymer growth) and is relevant only for thin film (< 1.7 µm) Parylene deposition. This
technique has been applied to create polycrystalline and nano-hill structured Parylene [68].
For thicker films, Parylene polymerization continues as secondary monomers adsorb on
top of the deactivated monomer layer, and conformality is reasserted [67].
14
A quick note is mentioned here on the adhesion of Parylene on different materials.
Though the deposition of Parylene results in strong adhesion for many substrates, the bond
affinity of Parylene is not universal; adhesion can be improved through the use of a silane-
based adhesion promoter, A-174 ([(gamma-(methacryloyloxy)propyl]trimethoxysilane).
A-174 has been found to greatly increase the adhesion of Parylene for silicon devices [69]
as well as platinum surfaces [70]. However, the typical immersion-based A-174 treatment
process readily dissolves common positive photoresist materials, which must be considered
during process design.
1.3.4 Surface modification of Parylene
Modification of the surface of Parylene C can result in useful functionalization (as
mentioned previously) as well as improvements in the material’s performance for specific
applications. Plasma treatments of Parylene surfaces have been explored to improve the
adhesion of various materials to Parylene. These techniques are advantageous as they
modify only the surface properties of the polymer without affecting the bulk. Oxygen
plasma and ion beam treatment of Parylene prior to gold deposition was found to increase
adhesion properties largely due to the added carbonyl functional groups following the
process as well as improved mechanical interlocking due to increased roughness [71]. Pre-
plasma treatment of various polymers prior to Parylene coating has also been shown to
improve adhesion. In a study observing in situ plasma treatment (i.e. plasma treatment
within the Parylene deposition chamber) of poly(tetrafluoroethylene) (PTFE),
poly(propylene) (PP), poly(methylmethacrylate) (PMMA), and glass substrates during
Parylene deposition, argon, oxygen, and methane plasma treatment all contributed to a
qualitative increase in adhesion, however the mechanisms for improved adhesion were
different for each chemistry [72]. Methane plasma treatment was found to be the most
consistent method to increase adhesion [72], due to the deposition of a hydrophobic surface
layer with additional free radical sites during methane plasma treatment. This is also
supported by a study involving the deposition of another hydrophobic plasma polymer
(trimethylsilane) to increase adhesion of Parylene to metal surfaces [73]. Oxygen plasma
15
treatment of PDMS has also been observed to increase Parylene adhesion up to a four-fold
improvement [74].
Plasma treatment can also be used to change the surface energy of Parylene C,
normally hydrophobic (native contact angle = ~80-90°), to create films that are either more
hydrophilic or hydrophobic, depending on the treatment parameters and the gases involved.
In one case, oxygen plasma was used to create hydrophilic Parylene surfaces by
introducing oxygen-related polar functional groups (e.g. carboxyl, hydroxyl) onto the
surface, but the hydrophobic nature of Parylene was found to revert after a week, but to
only 40-50% of its initial state [75]. Consecutive O2-SF6 (COS) plasma treatment has been
found to achieve contact angles of 169.4° to create super-hydrophobic surfaces, through a
combination of O2 plasma induced surface roughening and fluorine-based chemical
modification of the surface with SF6 [76].
For biomedical applications, plasma treatment of Parylene C surfaces can improve
cell adhesion, typically poor for untreated Parylene surfaces [24, 77]. The oxygen plasma
treated surfaces aid in cellular adhesion due to increased hydrogen bonding between the
surface and water molecules of the cellular material. Another method to improve cellular
adhesion is to adhere proteins onto as-deposited Parylene films (e.g. horse serum, bovine
serum albumin, immunoglobulin G, fibronectin, Matrigel) by soaking the films within the
solution [24, 78]. Pre-treatment of the Parylene film with oxygen plasma in this case has
been found to decrease the adsorption of proteins [24]. In cases where protein adhesion is
required within Parylene lumen structures, a pre-coating of poly(d-lysine) was found to
improve the wettability of the Parylene surface and to wick protein mixtures [79]. Other
methods for surface modification of Parylene C for biomedical applications involve UV-
induced photooxidation (1 or 2 hour treatments) to create a hydrophilic surfaces by
producing carboxyl and aldehyde groups [78]. The addition of other functional groups to
the polymer surface using Friedel-Crafts acylation to add thiol and poly(N-
isopropylacrylamide) (pNIPAM) groups for improved gold film and tissue adhesion,
respectively, [8] has also been demonstrated.
16

1.3.5 Thermal treatment of Parylene
Parylene C is a thermoplastic material, and is susceptible to thermal treatment
during or after fabrication. Post-fabrication thermal annealing of Parylene can increase
crystallinity and stiffness [80] and reduce water vapor permeation [81] in the annealed film.
Application of temperature and pressure facilitates Parylene to polymer bonding for unique
applications, such as forming microchannel structures [82]. By exposing Parylene-polymer
constructs to high temperatures (greater than the glass transition point (Tg) of Parylene, 60-
90°C, but below its melting temperature (Tm), 290°C [82]) while applying a bonding
pressure, mechanical fusing of Parylene into the second polymer can be achieved and aid
in bond formation. Plasma activation of the Parylene C layer to create radical species can
further aid in this process. This mechanism can be used to create Parylene to SiO 2 bonds
at 280°C with O2 plasma treatment [83], as well as Parylene to photoresist (SU-8, AZ 4620)
bonds, for temperatures greater than 90°C with O2 plasma treatment [84]. Bonding strength
was found to increase with increased bonding temperature [84]. Parylene-Parylene bonding
has also been demonstrated at both die [85] and wafer [86-88] levels, to construct devices
or to achieve an intermediate glue layer for wafer-level bonding.
Other techniques leveraging the thermoplastic property of Parylene includes a hot-
embossing process, where a nickel mold was pressed into Parylene films at 150°C to form
an imprint with <2.32% dimensional deviation [89]. Thermal forming of Parylene free
films has also been demonstrated by annealing multi-layer Parylene devices with varying
thickness or differing Parylene variants (Parylene C, N structures) to create residual stress
differences to form self-curling films [90-92]. Metal molds have also been used to
thermally shape Parylene planar films into curved [93, 94] and 3D structures [95].
Additional details on this thermal forming, or “thermoforming,” process of Parylene in
addition to its effects on the material properties of Parylene will be presented in the
subsequent chapters.
17
An exotic technique that leverages the thermal treatment of Parylene is the
pyrolysis of the polymer to form conductive polymer films [96-99]. In this technique,
Parylene is used as a carbon source, and is pyrolyzed at 900-1200°C under nitrogen
atmosphere to form a conductive thin film for use as an electrode surface. This process has
been observed to produce a 15-20% reduction in thickness following pyrolysis [96]. This
process takes advantage of the conformal deposition and patternable properties of Parylene
to form planar and 3D carbon electrodes and structures, by first fabricating the Parylene
structures and then undergoing the pyrolysis process.
1.4 Parylene-based implantable devices
Increasingly, micro-machined Parylene MEMS released entirely from carrier
substrates, so-called free film devices, have become more prevalent for implantable device
applications. In addition to its USP class VI biocompatibility rating for bioinertness,
Parylene’s low Young’s modulus substantiates its obvious choice as an implant material
and substrate. As current implantable devices are normally constructed from rigid materials
(e.g. silicon, ceramic, metal), the flexibility of these free film devices helps minimize
damage during and while implanted, and for some applications, allows for conformal
attachment to surrounding tissue. This flexibility is also advantageous for integrating these
Parylene devices into/onto existing implant architectures (e.g. electrodes, catheters,
implant housing) to provide additional functionality, such as feedback sensing, to current
hardware. Free film Parylene devices can also integrate transducer, electrical components
(e.g. coils [100, 101], discrete electronics [102], and chips [49, 102]), and flexible electrical
connections into a single, encapsulated structure, minimizing unnecessary complexity. For
some applications, the need for bulky packaging often required of implants for protection
from the harsh saline environment of the body can also be removed, as Parylene devices
have been shown to operate well in vivo without hermetic packaging.
The compatibility of Parylene with micromachining processes, as detailed in the
previous section, is also beneficial largely due to the benefits that come with MEMS
processing: namely, in creating precise designs using layout software and batch fabrication.
18
Following fabrication, these free film devices are normally released from a carrier substrate
(e.g. silicon) using sacrificial layers (e.g. silicon, oxides, photoresist, aluminum) that are
removed following the final process steps. Mechanical peeling, assisted by water
immersion, has also facilitated the release of devices without damaging the structure [45-
47]. A brief review on the current state-of-the-art for Parylene implantable devices is given
in the following section.
1.4.1 State-of-the-art Parylene-based implantable devices
There has been considerable work using Parylene as a substrate material for
implantable neural prostheses. Specifically for cortical recording and stimulation
applications, Parylene has been used to develop both penetrating [46, 47] and non-
penetrating microelectrode arrays [23, 103, 104]. In one device, the transparency of
Parylene was leveraged to allow for optical stimulation of cortical cells for optogenetics
while maintaining electrical recording functionality using transparent, indium tin oxide
(ITO) electrodes [23]. In another effort, sandwiching multiple layers of Parylene and metal
traces, high-density neural probes with 256 electrodes on a single shank have been
produced [103]. Parylene-based cuff electrodes for recording from nerve fibers in the
peripheral nervous system have also been developed that utilize thermal curling [92] or
built in ratcheting structures [105] to create conformal wrapping around fibers for
improved signals. Planar Parylene spinal cord stimulators have also been fabricated and
chronic performance demonstrated in rat animal models [106].
In the field of sensory neural prosthetics, Parylene has become a popular material
for retinal implants that involve electrical stimulation of retinal ganglion cells to induce
image percepts. High electrode density, curved retinal implants were demonstrated using
both a multi-Parylene-metal layer process and thermal shaping of the finished device to
match the curvature of the retina [107]. In addition to exploiting thermal stress to curl
Parylene devices, the inclusion of thin, “score” lines enable enabled a “origami like”
manipulation of film shape [108]. An all-inclusive, wireless Parylene retinal implant was
also demonstrated that incorporated stimulation electrodes, discrete electronic components,
19
and a radio frequency (RF) inductive powering coil, on a single Parylene substrate [102].
3D penetrating electrodes for the retina exploited a micromachined silicon mold to form
sharp Parylene needle-like structures [50]. Thermal curling of a planar Parylene electrode
array was also demonstrated to match the spiral structure of the cochlea, while using
Parylene as a thin, flexible substrate to enable a higher density of electrodes than the current
state of the art for cochlear implants [90].
The boundaries for Parylene processing and device development are expanding, as
evidenced by the devices mentioned previously and an increasing number of publications
and conference presentations within the MEMS and biomedical engineering communities.
In the following section, we introduce the basis of the present work: the development of
Parylene-based devices for cortical applications.
1.5 Parylene in the brain
Alongside the success of Parylene-based devices for implantable neural
applications mentioned above, this work centers on the application of Parylene within two
thrusts of research in the cortical field: namely (1) the development of a reliable
intracortical prosthesis in realizing brain computer or brain machine interfaces to restore
functionality for loss-of-limb or paralyzed patients and (2) the improvement of treatment
methods for hydrocephalus, a chronic condition of the excess accumulation of fluid within
the brain that affects 1 in 500 infants born today.
1.5.1 Need for reliable intracortical interface technology
A vast network of neurons within the brain transmits signals to carry commands to
complete a variety of everyday tasks, such as the movement of an arm. As these signals are
largely electrical in nature, a correctly placed interface, i.e. an implanted intracortical
electrode, can read these signals to construct enabling technology for motor control. This
is the basis for brain-machine interfaces (BMI), where arrays of intracortical electrodes are
implanted within the brain to measure the extracellular electrical activity associated with
movement transmitted from neurons within the cortex. The development of these
20
chronically implanted electrodes is a large effort within biomedical research not only to
realize BMI technologies, but also to study the activities and organization of the brain.
However, the development of a reliable interface to chronically record from these
neurons has been a herculean task, largely due to biological and non-biological factors that
lead to premature device failure. Biological factors include the foreign body response to
the implant, in addition to a cascade of inflammatory responses from the trauma associated
with implantation that creates a glial scar around the implant. Over time, a combination of
these two factors isolate the implant from the body, in effect blocking the measurement
electrodes from the neural signals. In addition, common non-biological factors, such as
electrode corrosion or cable breakage stemming from device design further exacerbates
chronic performance. There is a need for improved electrode technology for better
biological integration and device design to advance the current state of the art for more
reliable operation.
1.5.2 Need for improved hydrocephalus treatment technology
Hydrocephalus is a chronic condition characterized by the accumulation of excess
cerebrospinal fluid (CSF) within the ventricles of the brain, resulting in elevated
intracranial pressure and severe neurological defects and death. Currently, hydrocephalus
is treated by implanting a shunt, a multi-holed silicone catheter and valve system, into the
ventricles to drain the excess CSF. Though effective, shunts fail at an alarming rate of 40%
within the first year. There are many causes of shunt failure, including mechanical issues
and infection, but the most common is blood and tissue obstruction of the drainage ports.
Though these failure rates are high, there are currently no reliable and convenient
methods to detect shunt obstruction failure. Obstruction can present itself in only vague
symptoms, such as headaches and nausea, which can be confused with other medical
conditions. Currently, the clinical standard to assess obstruction is to use imaging of the
brain or invasive shunt taps, where a needle is punctured through the skin and to the shunt
to assess flow. However, these methods are subjective and involve secondary risk to the
21
patients. There is a need for a non-invasive method to periodically assess shunt patency so
that treatment can be improved through the timely and accurate diagnoses of shunt failure.
1.5.3 Parylene perspectives
Both thrusts mentioned above highlight a need that an implantable Parylene-based
device is capable of filling, but there is still work to be done concurrently with regards to
the development and maturation of Parylene as a device material. Specifically, issues with
fabrication, adhesion between inter-Parylene layers, and packaging methods deter many
groups from using Parylene for in vivo applications. Details of these three areas are
addressed in the following section.
1.6 Challenges of utilizing Parylene as a structural material
1.6.1 Fabrication considerations
The comparatively narrow range of processing temperatures compatible with
Parylene remains one of the largest obstacles when developing bioMEMS with the material.
Parylene is a semi-crystalline polymer and thus is very sensitive to temperature. As
mentioned previously, any process that nears the Tg of Parylene can induce bulk material
changes that can impede processing [82]. For example, standard soft bake temperatures for
photoresists spun on silicon wafers are ~120°C; for Parylene, this temperature must be
reduced. It follows that certain recipes developed for silicon are not compatible with
Parylene. There are additional small nuances in processing Parylene; for example during
O2 plasma dry etching, cooling or rest steps must be included to prevent Parylene being
exposed to temperatures above its Tg for long periods of time due to heating from the
plasma exposure. These thermal considerations extend to optimal operating conditions of
Parylene C devices; mismatch of coefficient of thermal expansion (CTE) between Parylene
and the coated substrate has been found to create mechanical failure for devices under
relatively high temperature soaking conditions [109]. UV exposure of a large area of
unmasked positive photoresist (e.g. in the fabrication of micro-sacrificial structures) during
photolithography can also be detrimental to Parylene processing. Upon exposure, the UV-
22
initiated chemical reaction of Novolak-based photoresists (AZ) can create an increase in
temperature of 200°C [110] that can affect the underlying Parylene, such as in the creation
of bubbles during this process that diffuse through the Parylene layer and causes points of
delamination [111].
In addition to issues during fabrication, the thin film nature of Parylene creates
difficulties in forming high-aspect ratio three-dimensional structures. Thus Parylene-based
device designs are limited to the process limitations of the polymer, and often require
ingenuity or exotic fabrication steps to realize structures that are relatively more trivial for
the bulk micromachining of silicon. For these reasons, there is a steep learning curve
associated with processing with Parylene that can cause a variety of complications and
poor device yield, even for the most experienced groups working with the polymer.
1.6.2 Wet adhesion issues
It is well established that the adhesion between Parylene-Parylene layers and
Parylene-metal (e.g. thin film gold, titanium, platinum) layers can be compromised during
long term soaking conditions, with device life lasting from days [100, 112-114] to more
than a year [69]. Before analyzing the underlying mechanism, it is helpful to define the
nature of polymer/polymer-metal adhesion. Adhesion between two surfaces can generally
be categorized as one of two interactions: (1) mechanical interlocking of one surface to the
other at size-scales ranging from molecular to polymer chain lengths or (2) ionic or
covalent bonding between the molecules at the interface [115].
For Parylene-Parylene layers formed during CVD, the interactions between the
bottom Parylene surface and the top can comprise a combination of these two mechanisms
[3]. The top layer forms when vaporized monomer units either physically adsorb to the
surface or bond chemically with surface-bound free radicals. Parylene monomer units can
propagate from these sites to finish polymerizing to create a top layer that is adhered to the
bottom layer. Parylene-metal or metal-Parylene layer interactions are predominantly
chemical, primarily resulting from first and second order bonds between the carbon atoms
of Parylene monomer units and the various metal atoms (titanium and oxygen atoms of
23
titanium-oxides, gold trimer units, and electronegative platinum atoms) [116]. For both
Parylene-Parylene and Parylene-metal adhesion, bonds are susceptible to failure under
chronic soak conditions. Proposed mechanisms for soak delamination include water
intrusion physically weakening mechanical adsorption and accentuating slight differences
in surface energies between the two layers [72].
Soaking induced delamination is strongly dependent on the presence of voids or
contaminants at the bond interface [117]. Water vapor transport through the Parylene C
layer condenses at these void sites underneath the coating and leads to eventual
delamination [118]. This motivates additional cleaning processes prior to Parylene
deposition (e.g. dilute HF bath [119]) to prevent void formation. The thickness of the
Parylene has also been shown to affect adhesion; thicker Parylene coatings (>10 µm)
reduce and slow water vapor permeation [120], however water penetration and subsequent
delamination of layers up to 40 µm thick has been observed [121]. Increases in Parylene
layer thickness will improve the lifetime of the device, but will not prevent its eventual
failure.
The inevitable failure of these Parylene-Parylene and Parylene-metal devices have
prompted the development of a variety of adhesion techniques that center on improving
both the mechanical and chemical interactions between the layers. Improving the
mechanical interactions have centered on two fronts: namely, (1) generating additional
polymer chains interweaving the top and bottom layers of Parylene by using high
temperature (above Tg) annealing during deposition using in situ heating [122] as well as
a post-treatment [93, 106, 107], and (2) using macro level surface roughness, substrate
porosity, or mechanical anchor points to provide additional regions for mechanical
interlocking of the top layer [123]. Improving the chemical interactions have largely
focused on using plasma treatments (e.g. argon, methane, oxygen [70, 72, 73]) or chemical
coatings (e.g. propylene carbonate [119]) to functionalize the Parylene and metal surfaces
and add functional or reactive groups to create more hydrophobic surfaces prior to Parylene
deposition [8, 113, 124]. In one study, the use of oxygen plasma treatment on Parylene and
24
platinum surfaces prior to Parylene deposition was found to decrease the adhesion [70].
The integration of plasma sources within the deposition chamber [72, 73] to plasma treat
the substrates prior to Parylene deposition without breaking vacuum have yielded
promising results in improving adhesion.
It is important to note that the use of these treatments for implantable devices
presents another hurdle, as the process should not only be compatible with Parylene, but
also not compromise the overall biocompatibility of the device, especially for these plasma-
based or chemical coatings. Overall, though there have been many efforts, Parylene-
Parylene and Parylene-metal adhesion still remains an issue and a large deterrent for using
Parylene as a device material.
1.6.3 Packaging issues
Lastly, a large issue for implantable Parylene-based devices is the difficulties
associated with making electrical connections to the soft Parylene-metal substrate. Taking
the intracortical recording electrode as an example, it is vital to capture the electrical neural
signals for proper processing to enable the benefits of neural prosthetic technology.
However to do this, electrical connections are necessary from the Parylene device to the
proper neurophysiological recording system. Typically to make electrical connections to
MEMS devices, a variety of techniques are used: such as soldering, wire bonding, and ball
grid array (BGA) bonding. However, these are all tedious processes, forming single pad-
by-pad connections in the case of soldering and wire bonding, and often require adverse
bonding conditions for Parylene (e.g. high heat, ultrasound vibration) that can cause
significant damage to the thin metal films on the Parylene substrate.
There have been methods developed that are more compatible with polymer
substrates, such as microflex interconnect technology, which uses ball or wedge bonds to
form microrivets through holes that are designed within the contact pad of Parylene cables
[125, 126]. However, this technique still requires a wirebonder that utilizes high energy
(e.g. heat, ultrasound) to form mechanical bonds that can potentially damage Parylene and
is tediously conducted on a pad-by-pad basis. Other methods involve conductive pastes,
25
such as flip-chip bonding using anisotropic conductive films (ACFs) [127] or manually
applying conductive epoxies [128], but these techniques also require high temperatures
(relative to the Tg of Parylene) for curing and often utilize materials that can compromise
the biocompatibility of the device. There is much more work to be done in this sense, to
develop reliable packaging methods for Parylene-based MEMS devices that not only utilize
techniques that are compatible with Parylene but also do not compromise device integrity.
1.7 Objectives
With all of this in mind, it is the aim of the present work to accomplish the following
research aim:
To advance the development of implantable Parylene-based devices by providing
new knowledge and process techniques related to the fabrication, packaging, and
integration/implementation of Parylene-based MEMS.
This work focuses on advancing Parylene-based MEMS technology specifically for
cortical applications with regards to the needs mentioned previously. Towards more
reliable cortical electrode technologies, the first two chapters sets to apply two Parylene-
based approaches on improving chronic in vivo performance: (1) the development of
electrode-integrated sensor technology to assess the mechanical interactions at the interface
between the implant and surrounding neural tissue to obtain quantitative data on how to
improve chronic performance and (2) the development of a Parylene-based electrode
technology that leverages the flexibility and biocompatibility of the polymer material in
improving chronic reliability. Towards improving the current treatment for hydrocephalus,
the third chapter introduces an implantable Parylene-based sensor to quantitatively assess
the obstruction of the shunt catheter, designed to integrate with current shunt technologies
used within the clinic.
In the following chapters, a novel Parylene processing technique is introduced,
effects of heat-treatment on Parylene films and devices is analyzed, and packaging, testing,
and implementation methods for clinically relevant implantable Parylene-based devices are
26
presented. It is the hope of the author that the following discussions may aid in the
advancement of Parylene-based MEMS technologies and encourage the use of Parylene as
a substrate material in the development of novel implantable biomedical devices.
1.8 References
[1] M. Szwarc, "Some remarks on the CH2=benzene=CH2 molecule," Discussions of
the Faraday Society, vol. 2, pp. 46-49, 1947.
[2] W. F. Gorham, "A New, General Synthetic Method for the Preparation of Linear
Poly-p-xylylenes," Journal of Polymer Science Part A-1: Polymer Chemistry, vol. 4, pp.
3027-3039, 1966.
[3] J. B. Fortin and T.-M. Lu, "Deposition kinetics for polymerization via the Gorham
route," in Chemical Vapor Deposition Polymerization, ed: Springer, 2004, pp. 41-55.
[4] W. R. Dolbier and W. F. Beach, "Parylene-AF4: a polymer with exceptional
dielectric and thermal properties," Journal of fluorine chemistry, vol. 122, pp. 97-104, 2003.
[5] "SCS Parlyene Properties," ed: Specialty Coating Systems, 2010.
[6] J. Lahann, et al., "Synthesis of Amino[2.2]paracyclophanes—Beneficial
Monomers for Bioactive Coating of Medical Implant Materials," Angewandte Chemie
International Edition, vol. 40, pp. 726-728, 2001.
[7] B.-J. Jeon, et al., "Parylene-A coated microplate for covalent immobilization of
proteins and peptides," Journal of Immunological Methods, vol. 353, pp. 44-48, 2010.
[8] P. N. Wahjudi, et al., "Improvement of metal and tissue adhesion on surface-
modified parylene C," Journal of Biomedical Materials Research Part A, vol. 89, pp. 206-
214, 2009.
[9] H. Nandivada, et al., "Vapor-Based Synthesis of Poly[(4-formyl-p-xylylene)-co-
(p-xylylene)] and Its Use for Biomimetic Surface Modifications," Macromolecular Rapid
Communications, vol. 26, pp. 1794-1799, 2005.
[10] J. Lahann, et al., "Reactive Polymer Coatings: A Platform for Patterning Proteins
and Mammalian Cells onto a Broad Range of Materials," Langmuir, vol. 18, pp. 3632-3638,
2002/04/01 2002.
[11] H. Nandivada, et al., "Reactive Polymer Coatings that “Click”," Angewandte
Chemie International Edition, vol. 45, pp. 3360-3363, 2006.
27
[12] J. J. Senkevich, et al., "Thermomechanical Properties of Parylene X, A Room-
Temperature Chemical Vapor Depositable Crosslinkable Polymer," Chemical Vapor
Deposition, vol. 13, pp. 55-59, 2007.
[13] J. Lahann and R. Langer, "Surface-Initiated Ring-Opening Polymerization of ϵ-
Caprolactone from a Patterned Poly(hydroxymethyl- p-xylylene)," Macromolecular Rapid
Communications, vol. 22, pp. 968-971, 2001.
[14] J. J. Senkevich, et al., "Unique structure/properties of chemical vapor deposited
parylene E," Journal of Vacuum Science & Technology A: Vacuum, Surfaces, and Films,
vol. 20, pp. 1445-1449, 2002.
[15] K. Y. Suh, et al., "A Novel Photodefinable Reactive Polymer Coating and Its Use
for Microfabrication of Hydrogel Elements," Advanced Materials, vol. 16, pp. 1401-1405,
2004.
[16] G. E. Loeb, et al., "Parylene as a Chronically Stable, Reproducible Microelectrode
Insulator," Biomedical Engineering, IEEE Transactions on, vol. BME-24, pp. 121-128,
1977.
[17] E. M. Schmidt, et al., "Long-term implants of Parylene-C coated microelectrodes,"
Medical and Biological Engineering and Computing, vol. 26, pp. 96-101, 1988/01/01 1988.
[18] P. F. Man, et al., "Microfluidic plastic interconnects for multi-bioanalysis chip
modules," Bellingham WA, WA 98227-0010, United States, 1997, pp. 196-200.
[19] T. Zöpfl, et al., "Characterisation of the intrinsic stress in micromachined parylene
membranes," in SPIE Europe Microtechnologies for the New Millennium, 2009, pp.
73621M-73621M-11.
[20] N. Binh-Khiem, et al., "Active Parylene-encapsulated droplets for displays,"
Polymer Preprints, vol. 49, p. 931, 2008.
[21] N. Binh-Khiem, et al., "Polymer thin film deposited on liquid for varifocal
encapsulated liquid lenses," Applied Physics Letters, vol. 93, p. 124101, 2008.
[22] Y. Yoshihata, et al., "Micro liquid prism," in Micro Electro Mechanical Systems,
2009. MEMS 2009. IEEE 22nd International Conference on, 2009, pp. 967-970.
[23] P. Ledochowitsch, et al., "A transparent μECoG array for simultaneous recording
and optogenetic stimulation," in Engineering in Medicine and Biology Society, EMBC,
2011 Annual International Conference of the IEEE, 2011, pp. 2937-2940.
28
[24] T. Y. Chang, et al., "Cell and protein compatibility of parylene-C surfaces,"
Langmuir, vol. 23, pp. 11718-11725, Nov 6 2007.
[25] J. I. Kroschwitz, Kirk-Othmer Encyclopedia of Chemical Technology. New York:
John Wiley & Sons, Inc., 1998.
[26] J. Lahann, "Vapor-based polymer coatings for potential biomedical applications,"
Polymer International, vol. 55, pp. 1361-1370, Dec 2006.
[27] B. A. Weisenberg and D. L. Mooradian, "Hemocompatibility of materials used in
microelectromechanical systems: Platelet adhesion and morphology in vitro," Journal of
Biomedical Materials Research, vol. 60, pp. 283-291, 2002.
[28] Specialty Coating Systems, "Solvent Resistance of Parylene," 2001.
[29] E. Meng, et al., "Plasma removal of parylene c," Journal of Micromechanics and
Microengineering, vol. 18, p. 045004, Apr 2008.
[30] T. E. Nowlin and D. F. Smith, Jr., "Surface Characterization of Plasma-Treated
Poly-p-Xylylene Films," Journal of Applied Polymer Science, vol. 25, pp. 1619-1632, 1980.
[31] X. Q. Wang, et al., "A Parylene Micro Check Valve," in Proceedings of the IEEE
Microelectromechanical Systems Conference, Orlando, FL, 1999, pp. 177-182.
[32] B. P. Levy, et al., "Definition of the Geometric Area of a Microelectrode Tip by
Plasma-Etching of Parylene," Ieee Transactions on Biomedical Engineering, vol. 33, pp.
1046-1049, Nov 1986.
[33] B. Ratier, et al., "Vapor deposition polymerization and reactive ion beam etching
of poly(p-xylylene) films for waveguide applications," Optical Materials, vol. 12, pp. 229-
233, 1999.
[34] N. Majid, et al., "The Parylene-Aluminum Multilayer Interconnection System for
Wafer Scale Integration and Wafer Scale Hybrid Packaging," Journal of Electronic
Materials, vol. 18, pp. 301-311, Mar 1989.
[35] J. T. C. Yeh and K. R. Grebe, "Patterning of poly-para-xylylenes by reactive ion
etching," Journal of Vacuum Science & Technology A, vol. 1, pp. 604-608, 1983.
[36] R. R. A. Callahan, et al., "Downstream oxygen etching characteristics of
polymers," Journal of Vacuum Science & Technolology B, vol. 21, pp. 1496-1500, 2003.
29
[37] R. R. A. Callahan, et al., "Effects of gas pressure and substrate temperature on the
etching of parylene-N using a remote microwave oxygen plasma," Journal of Vacuum
Science & Technology B, vol. 19, pp. 725-731, 2001.
[38] J. Yeh and K. Grebe, "Patterning of poly ‐ para ‐ xylylenes by reactive ion
etching," Journal of Vacuum Science & Technology A, vol. 1, pp. 604-608, 1983.
[39] N. Majid, et al., "The parylene-aluminum multilayer interconnection system for
wafer scale integration and wafer scale hybrid packaging," Journal of Electronic Materials,
vol. 18, pp. 301-311, 1989.
[40] T. Standaert, et al., "High-density plasma patterning of low dielectric constant
polymers: A comparison between polytetrafluoroethylene, parylene-N, and poly (arylene
ether)," Journal of Vacuum Science & Technology A, vol. 19, pp. 435-446, 2001.
[41] S. Selvarasah, et al., "A reusable high aspect ratio parylene-C shadow mask
technology for diverse micropatterning applications," Sensors and Actuators a-Physical,
vol. 145, pp. 306-315, Jul-Aug 2008.
[42] G. E. Loeb, et al., "Toward the ultimate metal microelectrode," Journal of
Neuroscience Methods, vol. 63, pp. 175-183, Dec 1995.
[43] E. M. Schmidt, et al., "Laser exposure of Parylene-C insulated microelectrodes,"
Journal of Neuroscience Methods, vol. 62, pp. 89-92, Nov 1995.
[44] J. D. Weiland, et al., "Recessed electrodes formed by laser ablation of parylene
coated, micromachined silicon probes," in Engineering in Medicine and Biology Society,
1997. Proceedings of the 19th Annual International Conference of the IEEE, 1997, pp.
2273-2276 vol.5.
[45] D. C. Rodger, et al., "Flexible parylene-based multielectrode array technology for
high-density neural stimulation and recording," Sensors and Actuators B-Chemical, vol.
132, pp. 449-460, Jun 2008.
[46] B. J. Kim, et al., "3D Parylene sheath neural probe for chronic recordings," Journal
of Neural Engineering, vol. 10, p. 045002, 2013.
[47] J. T. W. Kuo, et al., "Novel flexible Parylene neural probe with 3D sheath structure
for enhancing tissue integration," Lab on a Chip, vol. 13, pp. 554-561, 2013.
[48] Y. Rui, et al., "Parylene-based implantable Pt-black coated flexible 3-D
hemispherical microelectrode arrays for improved neural interfaces," Microsystem
technologies, vol. 17, pp. 437-442, 2011.
30
[49] R. Huang and Y. Tai, "Parylene-Pocket Chip Integration," in Micro Electro
Mechanical Systems, 2009. MEMS 2009. IEEE 22nd International Conference on, 2009,
pp. 749-752.
[50] R. Wang, et al., "Fabrication and characterization of a parylene-based three-
dimensional microelectrode array for use in retinal prosthesis," Microelectromechanical
Systems, Journal of, vol. 19, pp. 367-374, 2010.
[51] R. Wang, et al., "Fabrication and properties of 3D flexible parylene-based
microelectrode array with silicon tips," in Micro Electro Mechanical Systems (MEMS),
2011 IEEE 24th International Conference on, 2011, pp. 253-256.
[52] M. Kurihara, et al., "3d laser lithography combined with parylene coating for the
rapid fabrication of 3d microstructures," in Micro Electro Mechanical Systems (MEMS),
2012 IEEE 25th International Conference on, 2012, pp. 196-199.
[53] B. Lu, et al., "Highly flexible, transparent and patternable parylene-C
superhydrophobic films with high and low adhesion," in Micro Electro Mechanical
Systems (MEMS), 2011 IEEE 24th International Conference on, 2011, pp. 1143-1146.
[54] B. J. Kim, et al., "A novel method to fabricate parylene-based flexible microfluidic
platforms with commercial adhesive tape," Journal of Micromechanics and
Microengineering, vol. 25, p. 017003, 2015.
[55] Y. Nishinaka, et al., "Fabrication of Polymer Microneedle Electrodes Coated with
Nanoporous Parylene," Japanese Journal of Applied Physics, vol. 52, p. 06GL10, 2013.
[56] N. Binh-Khiem, et al., "Tensile film stress of parylene deposited on liquid,"
Langmuir, vol. 26, pp. 18771-18775, 2010.
[57] J. Charmet, et al., "Solid on liquid deposition," Thin Solid Films, vol. 518, pp. 5061-
5065, 2010.
[58] S. Komatsu, et al., "Peristaltic micropump fabricated by depositing parylene
directly on liquid," in Solid-State Sensors, Actuators and Microsystems Conference, 2009.
TRANSDUCERS 2009. International, 2009, pp. 2007-2010.
[59] N. Binh-Khiem, et al., "Porous Parylene and effects of liquid on Parylene films
deposited on liquid," in Micro Electro Mechanical Systems (MEMS), 2011 IEEE 24th
International Conference on, 2011, pp. 111-114.
31
[60] G. S. Prihandana, et al., "Polyethersulfone membrane coated with nanoporous
Parylene for ultrafiltration," Microelectromechanical Systems, Journal of, vol. 21, pp.
1288-1290, 2012.
[61] G. Demirel, et al., "Template-based and template-free preparation of
nanostructured parylene via oblique angle polymerization," Thin Solid Films, vol. 518, pp.
4252-4255, 2010.
[62] R. P. von Metzen, et al., "Diffusion-limited deposition of Parylene C,"
Microelectromechanical Systems, Journal of, vol. 20, pp. 239-250, 2011.
[63] E. M. Charlson, et al., "Temperature selective deposition of Parylene-C," IEEE
transactions on bio-medical engineering, vol. 39, pp. 202-206, 1992/02// 1992.
[64] Y. Xu and Y.-C. Tai, "Selective deposition of parylene C for underwater shear-
stress sensors," in TRANSDUCERS, Solid-State Sensors, Actuators and Microsystems, 12th
International Conference on, 2003, 2003, pp. 1307-1310.
[65] P. Kramer, et al., "Polymerization of Para-Xylylene Derivatives (Parylene
Polymerization) 1. Deposition Kinetics for Parylene-N and Parylene-C," Journal of
Polymer Science, Polymer Chemistry Edition, vol. 22, pp. 475-491, 1984.
[66] M. Gazicki, et al., "Polymerization of para-xylylene derivatives (parylene
polymerization) 2. heat-effects during deposition of parylene-C at different temperatures,"
Journal of Polymer Science, Part A: Polymer Chemistry, vol. 23, pp. 2255-2277, 1985.
[67] K. M. Vaeth and K. F. Jensen, "Transition metals for selective chemical vapor
deposition of parylene-based polymers," Chemistry of Materials, vol. 12, pp. 1305-1313,
2000.
[68] Y. Zhou, et al., "Bottom-up fabrication of special parylene films based on selective
growth on Au-coated surface," in Nano/Micro Engineered and Molecular Systems, 2009.
NEMS 2009. 4th IEEE International Conference on, 2009, pp. 441-444.
[69] J. M. Hsu, et al., "Encapsulation of an Integrated Neural Interface Device With
Parylene C," Biomedical Engineering, IEEE Transactions on, vol. 56, pp. 23-29, 2009.
[70] C. Hassler, et al., "Characterization of parylene C as an encapsulation material for
implanted neural prostheses," Journal of Biomedical Materials Research Part B: Applied
Biomaterials, vol. 93B, pp. 266-274, 2010.
32
[71] J. H. Lee, et al., "Microstructure and adhesion of Au deposited on parylene-c
substrate with surface modification for potential immunoassay application," Plasma
Science, IEEE Transactions on, vol. 32, pp. 505-509, 2004.
[72] A. K. Sharma and H. Yasuda, "Effect of glow discharge treatment of substrates on
parylene‐substrate adhesion," Journal of Vacuum Science and Technology, vol.
21, pp. 994-998, 1982.
[73] D. Zeniieh, et al., "Parylene-C Thin Film For Biocompatible Encapsulations with
Very Strong Adhesion and Superior Barrier Properties," in 21st International Symposium
on Plasma Chemistry (ISPC 21), Queensland, Australia, 2013.
[74] P. Rezai, et al., "Plasma enhanced bonding of polydimethylsiloxane (PDMS) with
parylene," in Solid-State Sensors, Actuators and Microsystems Conference
(TRANSDUCERS), 2011 16th International, 2011, pp. 1340-1343.
[75] T. Trantidou, et al., "Oxygen plasma induced hydrophilicity of Parylene-C thin
films," Applied Surface Science, vol. 261, pp. 43-51, 2012.
[76] X. Bi, et al., "Super Hydrophobic Parylene-C Produced by Consecutive O2 and
SF6 Plasma Treatment," Microelectromechanical Systems, Journal of, vol. 23, pp. 628-
635, 2014.
[77] T. Hoshino, et al., "Improvement of neuronal cell adhesiveness on parylene with
oxygen plasma treatment," Journal of bioscience and bioengineering, vol. 113, pp. 395-
398, 2012.
[78] E. Delivopoulos, et al., "Guided growth of neurons and glia using microfabricated
patterns of parylene-C on a SiO 2 background," Biomaterials, vol. 30, pp. 2048-2058, 2009.
[79] C. D. Lee, et al., "Matrigel coatings for Parylene sheath neural probes," Journal of
Biomedical Materials Research Part B: Applied Biomaterials, 2015.
[80] J.-M. Hsu, et al., "Effect of Thermal and Deposition Processes on Surface
Morphology, Crystallinity, and Adhesion of Parylene-C," Sensors and Materials, vol. 20,
pp. 071-086, 2008.
[81] P. Menon, et al., "Characterization of water vapor permeation through thin film
Parylene C," in Solid-State Sensors, Actuators and Microsystems Conference, 2009.
TRANSDUCERS 2009. International, 2009, pp. 1892-1895.
33
[82] H. S. Noh, et al., "Parylene micromolding, a rapid and low-cost fabrication method
for parylene microchannel," Sensors and Actuators B-Chemical, vol. 102, pp. 78-85, Sep
1 2004.
[83] A. T. Ciftlik and M. Gijs, "A low-temperature parylene-to-silicon dioxide bonding
technique for high-pressure microfluidics," Journal of Micromechanics and
Microengineering, vol. 21, p. 035011, 2011.
[84] J.-C. Chang, et al., "A low-temperature parylene-C-to-silicon bonding using photo-
patternable adhesives and its applications," in Solid-State Sensors, Actuators and
Microsystems (TRANSDUCERS & EUROSENSORS XXVII), 2013 Transducers &
Eurosensors XXVII: The 17th International Conference on, 2013, pp. 2217-2220.
[85] D. Ziegler, et al., "Fabrication of Flexible Neural Probes With Built-In Microfluidic
Channels by Thermal Bonding of Parylene," Microelectromechanical Systems, Journal of,
vol. 15, pp. 1477-1482, 2006.
[86] H. S. Noh, et al., "Wafer bonding using microwave heating of parylene
intermediate layers," Journal of Micromechanics and Microengineering, vol. 14, p. 625,
2004.
[87] H. Kim and K. Najafi, "Characterization of low-temperature wafer bonding using
thin-film parylene," Microelectromechanical Systems, Journal of, vol. 14, pp. 1347-1355,
2005.
[88] Q. Shu, et al., "Wafer bonding with intermediate parylene layer," in Solid-State and
Integrated-Circuit Technology, 2008. ICSICT 2008. 9th International Conference on, 2008,
pp. 2428-2431.
[89] S. W. Youn, et al., "Fabrication of a micro patterned parylene-C master by hot-
embossing and its application to metallic mold replication," Journal of Micromechanics
and Microengineering, vol. 17, pp. 1402-1413, Jul 2007.
[90] A. C. Johnson and K. D. Wise, "A self-curling monolithically-backed active high-
density cochlear electrode array," in Micro Electro Mechanical Systems (MEMS), 2012
IEEE 25th International Conference on, 2012, pp. 914-917.
[91] S. Yamagiwa, et al., "Self-curling and-sticking flexible substrate for ECoG
electrode array," in Micro Electro Mechanical Systems (MEMS), 2013 IEEE 26th
International Conference on, 2013, pp. 480-483.
[92] X. Kang, et al., "Self-Closed Parylene Cuff Electrode for Peripheral Nerve
Recording," Microelectromechanical Systems, Journal of, vol. PP, pp. 1-1, 2015.
34
[93] W. Li, et al., "Integrated Flexible Ocular Coil for Power and Data Transfer in
Retinal Prostheses," in Engineering in Medicine and Biology Society, 2005. IEEE-EMBS
2005. 27th Annual International Conference of the, 2005, pp. 1028-1031.
[94] R. Huang and Y.-C. Tai, "Flexible parylene-based 3-D coiled cable," in
Nano/Micro Engineered and Molecular Systems (NEMS), 2010 5th IEEE International
Conference on, 2010, pp. 317-320.
[95] B. Kim, et al., "Formation of three-dimensional Parylene C structures via
thermoforming," Journal of Micromechanics and Microengineering, vol. 24, p. 065003,
2014.
[96] S. Konishi, et al., "Parylene-pyrolyzed carbon for mems applications," in
Proceedings of the IEEE Microelectromechanical Systems Conference, Maastricht,
Netherlands, 2004, pp. 161-164.
[97] K. Naka and S. Konishi, "Design and fabrication of pyrolyzed polymer micro and
nano structures," in Micro-NanoMechatronics and Human Science, 2005 IEEE
International Symposium on, 2005, pp. 103-108.
[98] K. C. Morton, et al., "Carbon electrode fabrication from pyrolyzed parylene c,"
Analytical Chemistry, vol. 83, pp. 5447-5452, 2011.
[99] K. C. Morton, et al., "Pyrolyzed Carbon Film Diodes," ACS applied materials &
interfaces, vol. 5, pp. 10673-10681, 2013.
[100] W. Li, et al., "Flexible parylene packaged intraocular coil for retinal prostheses,"
in Proceedings of 2006 International Conference on Microtechnologies in Medicine and
Biology, Okinawa, Japan, 2006, pp. 105-108.
[101] X. Sun, et al., "Design and fabrication of flexible parylene-based inductors with
electroplated NiFe magnetic core for wireless power transmission system," in Nano/Micro
Engineered and Molecular Systems (NEMS), 2012 7th IEEE International Conference on,
2012, pp. 238-242.
[102] W. Li, et al., "Parylene-based integrated wireless single-channel neurostimulator,"
Sensors and Actuators A: Physical, vol. 166, pp. 193-200, 2011.
[103] P. Ledochowitsch, et al., "Fabrication and testing of a large area, high density,
parylene MEMS µECoG array," in Micro Electro Mechanical Systems (MEMS), 2011
IEEE 24th International Conference on, 2011, pp. 1031-1034.
35
[104] C. Lei-Chun, et al., "A parylene-C based 16 channels flexible bio-electrode for
ECoG recording," in SENSORS, 2014 IEEE, 2014, pp. 877-880.
[105] H. Yu, et al., "A Parylene Self-Locking Cuff Electrode for Peripheral Nerve
Stimulation and Recording," 2014.
[106] M. S. Nandra, et al., "A Parylene-based microelectrode array implant for spinal
cord stimulation in rats," in Micro Electro Mechanical Systems (MEMS), 2011 IEEE 24th
International Conference on, 2011, pp. 1007-1010.
[107] D. C. Rodger, et al., "Flexible parylene-based multielectrode array technology for
high-density neural stimulation and recording," Sensors and Actuators B: Chemical, vol.
132, pp. 449-460, 2008.
[108] Y. Liu, et al., "Parylene Origami Structure for Introcular Implantation," in
Proceedings of the 17th International Conference on Soild-State Sensors, Actuators and
Microsystems (TRANSDUCERS 2013), Barcelona, Spain, 2013.
[109] W. Li, et al., "Wafer-level parylene packaging with integrated RF electronics for
wireless retinal prostheses," Microelectromechanical Systems, Journal of, vol. 19, pp. 735-
742, 2010.
[110] A. Reiser, et al., "The molecular mechanism of novolak–diazonaphthoquinone
resists," European Polymer Journal, vol. 38, pp. 619-629, 2002.
[111] J. T. W. Kuo, "Development of Micromachined Technologies for Neural
Interfaces," Ph.D. Dissertation, Biomedical Engineering, University of Southern California,
Los Angeles, CA, 2013.
[112] W. Li, Rodger, D.C., Menon, P.R., Tai, Y.C., "Corrosion behavior of parylene-
metal-parylene thin films in saline," ECS Transactions, vol. 11, pp. 1-6, 2008.
[113] J. P. Seymour, et al., "The insulation performance of reactive parylene films in
implantable electronic devices," Biomaterials, vol. 30, pp. 6158-6167, 2009.
[114] R. P. von Metzen and T. Stieglitz, "The effects of annealing on mechanical,
chemical, and physical properties and structural stability of Parylene C," Biomedical
Microdevices, vol. 15, pp. 727-735, 2013.
[115] J. Fortin and T.-M. Lu, "A model for the chemical vapor deposition of poly (para-
xylylene)(parylene) thin films," Chemistry of Materials, vol. 14, pp. 1945-1949, 2002.
36
[116] A. V. Vasenkov, "Atomistic modeling of parylene-metal interactions for surface
micro-structuring," Journal of Molecular Modeling, vol. 17, pp. 3219-3228, 2011.
[117] H.-I. Kuo, et al., "Development of micropackage technology for biomedical
implantable microdevices using parylene C as water vapor barrier coatings," in Sensors,
2010 IEEE, 2010, pp. 438-441.
[118] A. Vanhoestenberghe and N. Donaldson, "Corrosion of silicon integrated circuits
and lifetime predictions in implantable electronic devices," Journal of Neural Engineering,
vol. 10, p. 031002, 2013.
[119] J. H. Chang, et al., "Adhesion-enhancing surface treatments for parylene
deposition," in Solid-State Sensors, Actuators and Microsystems Conference
(TRANSDUCERS), 2011 16th International, 2011, pp. 390-393.
[120] H.-w. Lo and Y.-C. Tai, "Characterization of parylene as a water barrier via buried-
in pentacene moisture sensors for soaking tests," in Nano/Micro Engineered and Molecular
Systems, 2007. NEMS'07. 2nd IEEE International Conference on, 2007, pp. 872-875.
[121] J.-C. Chang, et al., "Reliable packaging for parylene-based flexible retinal
implant," in Solid-State Sensors, Actuators and Microsystems (TRANSDUCERS &
EUROSENSORS XXVII), 2013 Transducers & Eurosensors XXVII: The 17th International
Conference on, 2013, pp. 2612-2615.
[122] D. Kang, et al., "In situ heating to improve adhesion for parylene-on-parylene
deposition," in Nano/Micro Engineered and Molecular Systems (NEMS), 2012 7th IEEE
International Conference on, 2012, pp. 226-229.
[123] R. Huang and Y. C. Tai, "Parylene to silicon adhesion enhancement," in Solid-State
Sensors, Actuators and Microsystems Conference, 2009. TRANSDUCERS 2009.
International, 2009, pp. 1027-1030.
[124] J. J. Senkevich, et al., "The facile surface modification of poly (p-xylylene)
ultrathin films," Colloids and Surfaces A: Physicochemical and Engineering Aspects, vol.
216, pp. 167-173, 2003.
[125] T. Stieglitz, et al., "A Flexible Retina Implant For People Suffering from Retinitis
Pigmentosa," in International Functional Electrical Stimulation Society, 1999, pp. 61-64.
[126] H. Beutel, et al., "Versatile 'Microflex'-based Interconnection Technique," in SPIE
Conference on Smart Electronics and MEMS, San Diego, CA, 1998, pp. 174-182.
37
[127] W. Fan, et al., "A multi-shank silk-backed parylene neural probe for reliable
chronic recording," in Solid-State Sensors, Actuators and Microsystems (TRANSDUCERS
& EUROSENSORS XXVII), 2013 Transducers & Eurosensors XXVII: The 17th
International Conference on, 2013, pp. 888-891.
[128] R. Huang and Y. C. Tai, "Parylene-Pocket Chip Integration," in Micro Electro
Mechanical Systems, 2009. MEMS 2009. IEEE 22nd International Conference on, 2009,
pp. 749-752.

38

Understanding the interface between intracortical probes and the surrounding
cortical tissue is a subject of many efforts to improve the longevity and reliability of
implantable intracortical electrodes. Exploring the mechanics at this interface however is
faced with difficulties largely because of the wet, corrosive, and soft environment. In this
chapter, we present a Parylene-MEMS sensor that has been developed to meet this need,
with a unique electrochemical-based sensing mechanism to assess the mechanics at this
complex interface.
2.1 Background
2.1.1 Sensors for intracortical electrodes
There has been significant progress in the MEMS field on the development of
sensors used for a variety of applications within the auto industry (pressure sensors),
consumer electronics (accelerometers and gyroscopes), and healthcare (biosensors).
Specifically in the area of implantable devices for the medical field, there is still much
work to be done. Currently, MEMS-based biomedical sensors feature traditional silicon-
based technologies that are not well-suited for operation within the wet, corrosive
environment of the body. Consequently, these sensors require complicated hermetic
packaging to ensure that the sensitive electronics and sensors are protected from the
surrounding environment. However, hermetic enclosures add bulk and increase device size,
potentially creating long-term reliability issues, additional failure modes, and reducing
sensor performance in vivo. These disadvantages can trump the benefits of size scaling and
simplicity that are only achievable using MEMS devices. There is a need for the

PARYLENE-BASED EC-MEMS SENSOR FOR STUDIES OF
INTRACORTICAL PROBE INSERTION MECHANICS
39
development of flexible, low-profile, robust sensors that are capable of working within the
body’s wet environment without the need for complicated packaging.
One area in which the development of this sensor technology can be advantageous
is neural prosthetics, or more specifically, intracortical neuronal recording and stimulating
probe technologies. Currently, the potential of implantable intracortical prosthetics for
sensory or motor enabling technologies is challenged by mechanics-induced biological
issues encountered by the probe in vivo [1]. The initial stab wound injury from implantation
of these rigid probes, the resultant disruption of the blood-brain-barrier (BBB), as well as
the chronic immune response encountered by the probe as a foreign body object within the
cortical tissue have been found to diminish their chronic efficacy [2]. In addition, the
persistent presence of micromotion due to vascular pulsing induces a constant mechanical
agitation between the soft tissue and rigid probes [3]. Exploring these mechanical
interactions at the prosthesis-tissue interface can motivate the development of novel, more
reliable designs that aim to function over long periods in vivo, such that tissue damage and
the related immunological responses are minimized [4].
2.1.2 Studies of probe implantation mechanics
Among these mechanical interactions, there are three components that are of
interest during and after probe insertion: (1) the normal insertion force, (2) probe-tissue
interfacial forces, and (3) micromotion forces (Figure 2-1). The normal insertion force is
the aggregate force experienced by the entire probe during the implantation procedure. This
force is representative of the initial dimpling and compression of the tissue as well as
penetration through the meninges. Following penetration, there are two main forces that
dominate the mechanics during probe insertion: tissue displacement forces imposed by the
probe on tissue and frictional forces caused by the probe tip moving through tissue [4].
Measurements of normal insertion force during this period combine these two phenomena
into a single measurement obscuring the different interactions. Therefore, during this stage
of implantation, we are more interested in the probe-tissue interfacial forces, which are the
normal and shear forces that are encountered on the lateral faces of the probe caused by the
40
surrounding neural tissue. These allow for a more accurate measurement of the tissue
displacement forces and can be correlated with resulting tissue strains that cause further
damage to surrounding vasculature and neuronal macrostructures [1]. Lastly, after the
probe has reached the desired depth to target a specific layer of neurons, micromotion
forces are generated on the probe by small displacements from vascular pulsing. Though
not an exhaustive list, these three forces can provide great insight into the mechanical
interactions during and following probe implantation to aid in more efficacious designs.

Figure 2-1. An illustration of the three components of insertion mechanics during and
following insertion: (a) normal insertion force, (b) probe-tissue interfacial force, and (c)
micromotion force.

In addition to providing information for the development of improved device
designs, obtaining quantitative data on the effects of various intracortical probe
technologies on these different mechanical interactions (i.e. normal insertion force,
interfacial forces, and micromotion forces) can serve as a metric in comparing the
performance of various probe designs [5]. If studies were conducted to correlate specific
magnitudes of these forces to subsequent tissue damage or probe performance, standards
could be set to objectively evaluate device designs.
In light of these reasons, there have been efforts in literature to characterize these
41
mechanical interactions. Previously, external load cells were used to measure macroscale
normal insertion forces [5-7] as well as to compare differing probe tip designs [5, 8]. But
in these studies, out-of-plane stresses acting on the probe surface were not measured, which
provides limited insight into interfacial phenomena. Also, studies that compare normal
insertion force measurements alone have found poor correlation between normal force
magnitudes and chronic probe performance [8]. Other efforts such as modeling [9-11] and
ex vivo probe implantation experiments utilizing cross-sectional viewing planes [12] have
provided great insight into the theoretical distribution of strains and forces during
implantation, at the probe-tissue interface, and during micromotion events. However,
quantitative data obtained from surgical and in vivo conditions to capture these mechanical
interactions during insertion and micromotion events in real time is lacking.
Technologies to integrate sensors on probes have been developed, but many of
these sensors are difficult to apply and provide limited access to interfacial mechanics. For
example, CMOS sensors integrated on a silicon probe were used to measure effects of
insertion-derived probe deflection stresses during implantation [13]. Though resultant
probe deflection due to insertion into tissue does capture some interfacial forces, the
method cannot detect interfacial forces normal to the probe from the surrounding neural
tissue. Also, these sensors are currently limited to silicon probes. Furthermore, commercial
probes readily available to neuroscience researchers do not offer the option of sensor
integration with their electrodes. A more versatile approach is necessary to enable
comparative studies between different electrode designs and materials.
Integrated sensors must be able to: accommodate small shank areas, integrate with
CMOS-incompatible electrode shank materials, and operate in aqueous and saline
environments. By utilizing sensor technology along with histological methods, benchtop
or in vivo studies (such as comparing probe designs, materials, and insertion parameters)
can be conducted to correlate interfacial mechanical interaction data with chronic neural
probe integration and performance. In addition, these integrated devices can be used to
effectively characterize micromotion-related forces and their effects on long-term
42
performance. Understanding these design factors and corresponding mechanical
interactions at the prosthesis-tissue interface in vivo is one large step towards the realization
of reliable probe technologies that minimize tissue damage and the related immunological
responses that hinder the progress of chronically implanted neural prosthetic technologies
[4]. To address this unmet need, we adapted our Parylene C-based electrochemical-MEMS
(EC-MEMS) sensing technology [14-16] to transduce force and pressure distributions
directly on the probe surface.

2.2 Electrochemical impedance-based transduction method
Prior to detailing sensor design, the transduction method of these sensors is briefly
reviewed. EC-MEMS sensors use electrochemical impedance as a sensing modality to
transduce various signals [14]. These sensors usually consist of a pair of electrodes exposed
to an electrolyte (a conductive medium) and enclosed within a Parylene microchamber with
fluidic ports that allow fluid exchange with the local environment. To understand the
impedance-based transduction mechanism, it is useful to consider the well-known
simplified Randles circuit model of the electrode-electrolyte interface [17].
In Randles’ equivalent circuit of the electrode-electrolyte interface, both the
double-layer capacitance (Cdl) and the charge transfer resistance (Rct) are used to
approximate the phenomena of charge being transduced from the electrolyte by the
electrodes (Figure 2-2). At this interface, there exists two layers of charge known as the
double-layer, named by Helmholtz in 1879, formed by corresponding ions and molecules
in the electrolyte responding to excess charges on the surface of the electrode [18]. The
double layer is responsible for capacitive and resistive effects to the signals that traverse
the electrode to the electrolyte or vice versa [19].

43

Figure 2-2. Schematic of the Randles equivalent circuit model of an electrode-electrolyte
interface.

With the model in mind, we can introduce the impedance-based transduction
method. An alternating current (AC) signal is used to measure the electrochemical
impedance between a pair of electrodes in an electrolyte solution. If the frequency of this
AC signal is sufficiently high (fmeas), these double-layer effects are diminished as the
solution resistance (Rs) dominates the impedance response. Thus, in this sensing
configuration, the sensor is effectively a variable resistor, and a number of variables (cross
sectional area (A), distance between the electrodes (l), conductivity of solution (ρ)) can
establish a change in the measured solution resistance, and thus electrochemical impedance,
accordingly (Figure 2-3). The complex impedance for a two-electrode system is expressed
in (1) with the high-frequency approximation provided in (2).

44

Figure 2-3. Equivalent circuit model of measuring the electrochemical impedance between
a pair of electrodes in an electrolyte. Note the use of the simplified Randles circuit model
at the two electrode-electrolyte interfaces.

Z =
1
jωC
dl
R
ct
1
jωC
dl
+ R
ct
+ R
s
+
1
jωC
dl
R
ct
1
jωC
dl
+ R
ct
=
2R
ct
1+R
ct
jωC
dl
+ R
s
(2-1)

Z ≈ R
s
=
ρl
A
; for sufficiently high ω, where ω = 2πf (2-2)

The principle of sensor operation relies on external contact forces deforming a
compliant electrolyte-filled structure in the normal direction and redistributing the fluid
contained within the chamber. This alters the volumetric ionic conduction path of current-
carrying ions in the fluid, in effect modifying the cross sectional area between the two
electrodes, and registering as a change in the magnitude of the solution impedance (Figure
2-4). In this manner, impedance variations can be correlated to mechanical interfacial
contact forces exerted by the tissue in contact with the top of the sensor array. However,
because of the use of an electrochemical sensing scheme, the sensor is particularly sensitive
to changes in electrolyte compared to other membrane-based force sensors. Here, there
exists a trade-off between electrolyte sensitivity and robust performance in wet
environments; current membrane-based force sensors require bulk, hermetic packaging to
resist the high saline environment of the body. By leveraging the electrolytic environment
45
for sensing, the EC-MEMS sensor design eliminates the bulky packaging required by
current sensors and results in a low-profile and flexible device ideal for a chronic implant.

Figure 2-4. Schematic of electrochemical impedance-based transduction method of
contact forces utilizing a flexible, mechanically compliant microchamber.

2.3 Sensor design
The intracortical Parylene-based EC-MEMS sensor array technology is based on
distributed sensors arranged along a Parylene microchannel-sensing element (100 µm x 4.5
mm x 24 µm) to match the footprint of a ceramic-based cortical probe [20], notable for
exceptional chronic recording and electrochemical sensing performance within monkey
cortical work, for initial testing. The microchannel encloses eight fluidically-coupled
platinum (Pt) electrodes (100 µm x 130 µm) arranged at a pitch of 500 µm, with adjacent
pairs forming a sensor unit (Figure 2-5). Each array consists of seven sensor “pixels” along
the length of the probe; sensor 1 begins at the tip of the probe and sensor 7 is at its base
(Figure 2-5e). This sensor configuration was selected to maximize sensor density over the
area of interest defined by the cortical probe layout and to maintain transducer symmetry.
Etched fluidic ports at the channel ends allow the surrounding electrolyte to fill the channel.
Fluid can flow freely between the external environment and the channel via these access
46
ports. A Parylene flat flexible cable (FFC) with eight leads was also integrated into the
device design in order to establish electrical contact via a zero-insertion-force (ZIF)
connector.

Figure 2-5. (a) Optical micrograph of full sensor array and integrated Parylene cable.
Microchannel structure is highlighted in pink. (b) Scanning electron micrograph (SEM)
image showing top view of the sensor array with a fluidic port at the end of the
microchannel. (c) Top-down image of Parylene microchannel (pink) indicating the
electrodes and fluidic ports located at the ends. (d) SEM showing the cross section of the
20 µm Parylene microchannel sensing structure. (e) Microelectrode layout and sensor
numbering based on adjacent electrode pairs.

Because of the flexible and thin Parylene microchannel structure (4 µm thickness),
contact forces (i.e. elastic and spring forces) from surrounding cortical tissue will deform
the top surface along the length of the probe (Figure 2-6). It is important to note here that
the sensors are insensitive to hydrostatic forces because the fluidic ports allow for an open
fluidic connection between the inside and outside of the channel; the sensing element
deforms primarily due to mechanical contact forces on the top channel surface. The seven
sensor pixels along this length quantify the distribution of interfacial forces that the probe
imparts on the surrounding cortical tissue during insertion, extending a similar concept
47
from previous sensor technology developed to quantify interfacial forces produced during
implantation of an epiretinal implant onto the retinal surface [21]. The flexible substrate
(total thickness = 10 µm) also allows for simple, low-profile instrumentation of the ceramic
cortical probes. The microchannel and electrode layout can be modified to fit probe
footprints of a variety of designs, demonstrating its robustness and versatility.

Figure 2-6. Operation principle of the Parylene microchannel-based interfacial force
sensor (a) pre and (b) post insertion into tissue.

2.4 Sensor fabrication
Standard surface micromachining processes were used for sensor array fabrication
(Figure 2-7). All fabrication processes were performed at low temperatures (90 ºC) to
prevent excessive thermal cycling of the Parylene structural material. Parylene was chosen
as the sensor material for its biocompatibility, mechanical strength, electrical insulation
properties, and compatibility with the micromachining processes.

Figure 2-7. Overview of fabrication process for the Parylene-based EC-MEMS sensor
array. (a) Platinum electrodes and contact pads were patterned by liftoff onto a Parylene
substrate; an insulation layer was also deposited and etched to reveal electrode sites. (b)
Sacrificial photoresist was spun on and patterned to form the microchannel structure. (c)
Another Parylene layer was deposited to form the deformable top membrane structure and
48
fluidic ports were etched into the top using oxygen plasma; the device perimeter was then
etched using O2 plasma. (d) Devices were released off the carrier wafer and soaked in
acetone to remove the sacrificial photoresist. Then the devices were filled with 1× PBS
prior to testing.

Platinum electrodes and contact pads (2000 Å) were patterned by e-beam
evaporation and lift-off on a Parylene substrate (5 µm) coated onto a silicon carrier wafer.
Following deposition of a Parylene insulation layer (1.5 µm), openings for electrodes and
contact pads were etched using oxygen plasma. A sacrificial photoresist (PR) process was
used to form the microchannel, but because of the relatively tall height (20 µm), two PR
layers (AZ 4620; AZ Electronic Materials, Branchburg, NJ) were spun on (each 10 µm
thick) to accurately and uniformly establish the desired microchannel height. Following
this, a 4 µm thick layer of Parylene was deposited to form the final device structure. This
thickness was important as it set the sensor mechanics; thicker films would displace less
than thinner microchannels. A thickness of 4 µm was chosen based on previous sensor
work on measuring contact forces between epiretinal implants and retinal tissue [14].
Fluidic access ports at the channel ends were opened by oxygen plasma. Finally, to
facilitate precise shaping of the array for attachment to a fine ceramic shank tip, arrays
were singulated using oxygen plasma.
Devices were released from the wafer by stripping the protective photoresist mask
for the cut-out etch with acetone, and then submerging the wafer in DI water. The
hydrophobicity of the Parylene polymer allowed the arrays to easily separate from the
native oxide of the silicon carrier wafer and lift-off the surface. Sacrificial photoresist
within the channel was then removed by immersion in acetone and isopropyl alcohol
following rinse in deionized water. The individual arrays were then let to air dry overnight.
The mechanical integrity and dimensions of the Parylene microstructure was sufficient to
avoid stiction during the drying process.

49
2.5 Device packaging
Following fabrication, sensors underwent post-processing steps for electrical
packaging to the measurement system. Electrical connections to the integrated Parylene
cable were made via an 8 contact ZIF connector (8 channel, 0.5 mm pitch; Hirose Electric
Co., Simi Valley, CA), a hinge-based connector that allows for a zero-insertion-force
method when the cable is inserted into the bay of the connector. As the connector requires
a cable thickness of 300 µm, a poly(etheretherketone) (PEEK; 8504K13, McMaster-Carr,
Aurora, OH) polymer backing was applied as a stiffener on the back of the contact pad
region. This connection scheme enabled rapid, epoxy-less connections to multiple contact
pads supported on our flexible Parylene substrate [22] (Figure 2-8).

Figure 2-8. Image showing integrated Parylene flex cable inserted into a ZIF-ZIF
connector for electrical connection to the sensor.

The ZIF connector used in this packaging method was a ZIF-ZIF adaptor that
connected the sensor array to a commercial polymer flexible cable that was thicker and
more robust for handling; the ZIF-ZIF adaptor was made by soldering the legs of two
separate ZIF connectors together. The addition of the polymer flexible cable allowed for
more robust, repeatable connections into a custom PCB via another ZIF connector
compared to the thin Parylene cable.

50
The custom PCB was designed to be a multiplexing board that allowed for user-
controlled switching between various pairs of electrodes to measure the impedance using
a precision LCR meter (Agilent e4980a, Agilent Technologies, Santa Clara, CA). Digital
signals sent through a DAQ and LabVIEW graphical user interface (GUI) set multiplexer
lines to connect various electrodes to the working and counter leads of the LCR meter.
Impedance was measured at fmeas, with an amplitude of 1 Vp-p. A diagram of the PCB is
presented below (Figure 2-9).

Figure 2-9. Diagram of custom multiplexing PCB used to obtain measurements from
sensor array.

2.6 Sensor characterization
2.6.1 FEM analysis
Prior to sensor characterization, finite element modeling (FEM) was used to
analyze the mechanical response of the Parylene microchannel. A 3D model of the
Parylene microchannel was constructed in SolidWorks (Dassault Systèmes, SolidWorks
Corp., Waltham, MA) and a nonlinear-static analysis was carried out to determine the
51
resultant deflections from an applied load force (Figure 2-10). The material properties used
for Parylene C in the model are given in the table below:

Table 2-1. Material properties for Parylene C used for FEM deflection analysis of
microchannel structure.
Parameter Value
Material Parylene C
Material Type Linear Elastic Isotropic
Young’s modulus 3.00 GPa
Poisson’s ratio 0.4
Mass density 1289 kg/m
3

Shear modulus 318.90 MPa
Tensile strength 68.9 MPa
Yield strength 55.2 MPa

Figure 2-10. Screenshots of FEM results obtained in SolidWorks illustrating (a) the entire
model and (b) zoomed in image of the first sensor.

In the first model of the complete Parylene microchannel, load forces up to 65 mN
were applied to the top of each sensor unit. Each sensor unit comprised a surface area of
approximately 60,000 µm
2
(600 µm x 100 µm) as measured between adjacent electrode
midlines. Load forces were applied uniformly over the top surface of each sensor unit,
centered between the electrode pairs. Within the model, seven of these regions were created
52
and forces were applied over these sensor units. The simulation results indicated, for an
applied force range between 0–55 mN, that the sensor mechanics were quite similar
regardless of location along the microchannel (Figure 2-11). In this simulation however,
deflections exceeded 20 µm within the model as a no-contact boundary condition was not
applied to the bottom surface of the microchannel, i.e. the top surface was allowed to
deform through the bottom surface at high forces.

Figure 2-11. Load-displacement results obtained through FEM analysis of the each sensor
along the length of the microchannel. Results indicate a consistent mechanical response
regardless of sensor position along the length of a single microchannel. The artifact in the
results for sensor 7 (orange) arise from the presence of fluidic ports in the model.

As the previous modeling study with the full channel demonstrated that the
mechanical deflections were consistent from sensor to senor regardless of position along
the length of the microchannel, a single sensor unit model was created in SolidWorks
resembling a smaller microchannel with open ends (similar to sensors 2-6) and a top
surface area of 60,000 µm
2
(600 µm x 100 µm x 20 µm). These simulation results indicated
53
that the working sensor range was between 0-40 mN, and demonstrated a similar sigmoidal
mechanical response of the sensor unit (Figure 2-12).

Figure 2-12. Load-displacement results obtained from FEM analysis of a single sensor
limited to a 20 µm deflection due to the sensor floor. Results indicated that the working
sensor range was between 0-40 mN.

2.6.2 Sensor characterization setup
Following the simulations, mechanical characterization of the sensors was carried
out on the benchtop. Because the sensing principle (and thus sensor operation) requires the
sensor to be immersed, the sensor array was characterized in a custom acrylic jig designed
to keep the sensing element immersed in an electrolyte (1× PBS; Figure 2-13). A flat,
circular, probe tip (diameter = 630 µm), attached to a motorized z–axis stage, was used to
indent the microchannel in 2 µm increments until reaching the maximum deflection of 20
µm. The probe diameter was chosen to cover the entire surface of a sensor unit (60,000
µm
2
). An in-line load cell (0-50 g, LCFA-50; Omega Engineering, Stamford, CT) was used
to measure the applied force of the indenting probe during displacement. During this
calibration procedure, real-time impedance was measured across all sensor units using
LabVIEW software. A 10-minute relaxation step was used between calibrations of sensors
54
to ensure that the microchannel structure returned to the baseline reading (impedance ≤
0.1% of initial value). Also, to further decouple the sensor units of the array, sensors were
calibrated in an order chosen to prevent the sequential testing of two adjacent sensors.

Figure 2-13. (a) Load-displacement setup for mechanical and calibration testing of
Parylene force sensor array. (b) Optical micrograph taken using a camera underneath
testing setup illustrating deflection probe (pink) displacing into sensor element.

2.6.3 Calibration and mechanical characterization
Prior to calibration, EIS was performed across all sensor electrode pairs to
determine the optimal fmeas at which the solution resistance dominates the complex
impedance response. An excitation frequency of ~2 kHz was found to correspond to the
maximum resistive response at a phase value near 0° (Figure 2-14). However, to ensure
signal fidelity from parasitic capacitances of the measurement system, all subsequent
impedance measurements were performed at 1 kHz, which was sufficient to isolate the
solution resistance.

55

Figure 2-14. Impedance (a) magnitude and (b) phase plots of 18 individual sensors. An
fmeas of 1 kHz (dashed line) was selected and corresponds to maximum resistive response
at a phase value near 0°.

Mechanical characterization of the Parylene microchannel immersed in solution
confirmed modeling results as sensors demonstrated operational uniformity despite their
differing locations along the fluidically coupled channel. The Parylene microchannel was
found to deform uniformly along the length of the channel, as the displacement per applied
force was similar across runs for each sensor (Figure 2-15). However the operational force
range and sensor mechanical response deviated from the expected results within the FEM
experiments. These differences can be attributed to inaccuracies in terms of the modeling
parameters (namely Parylene’s Young’s modulus). Also, assumptions of the model such
as an ideal rectangular cross section, a perfectly flat top surface, and strictly normal
application of the load force, may differ from experimental conditions. Nevertheless, the
models were helpful in providing estimated ranges of operational forces, with which the
measured response of the sensor agreed.

56

Figure 2-15. Load-displacement plot for the Parylene microchannel sensing element of the
EC-MEMS sensor array confirming FEM results as operational uniformity was maintained
despite their differing locations along the fluidically coupled channel.

Plotting cyclic loading/unloading force-displacement data generated curves that
capture the sensor’s dynamic mechanical behavior. The channel structures were deformed
at a rate of 1 µm/s, held at a maximum deflection of 20 µm for 20 s and then returned to
the initial position at 1 µm/s; the next loading cycle was initiated 30 s following the return
to resting position. Five consecutive loading/unloading cycles were applied. Results
indicated negligible hysteresis regardless of the sensor position (Fig. 10). Although higher
frequency dynamic analysis was not performed, the frequency response of the Parylene
microchannel structure in air was calculated to be 570 kHz using FEM software (COMSOL,
Burlington, MA). It is expected that this frequency would be further reduced to account for
fluidic damping to a conservative estimate of 10-100 Hz, as observed in other resonant
frequency studies of Parylene thin film diaphragms [14, 23]. Fortunately, as the interfacial
force dynamics within the cortex are very slow (maximum ~3–5 Hz of micromotion due
to vascular pulsing [6]), the current sensor design is more than sufficient for this application.
Additional dynamic loading tests were also performed to assess the effects of
varying step-sizes (4, 6, 12, 20 µm) on possible deflection-linked dynamic mechanical
responses. Results indicated reproducible and repeatable displacement-load force curves
57
for varying displacements and small hysteresis over five cycles (0.01 Hz) at each depth. In
some instances however, considerable hysteresis was observed, which suggests yield and
process variations that can create non-uniform microchannel structures and thus varying
mechanical responses.

Figure 2-16. Loading and unloading cycle results of (a) four different sensors at maximum
deflection and (b) at various deflection depths indicating negligible hysteresis.

A sensitivity of 0.127 ± 0.001 percentage change in impedance/mN (%/mN; mean
± SE, n = 6) was found for each sensor unit (Fig. 11) from the calibration curve, within a
sensor working range of 0-60 mN. Without utilizing any signal conditioning techniques,
the resolution of the sensor is dependent on the accuracy of the impedance measuring
device. Using the high precision LCR meter (“short” measurement function, ± ~0.2%
impedance magnitude), the resolution of the measurement systems was calculated to be ~
±1 mN.

58

Figure 2-17. Obtained calibration curve for the EC-MEMS sensor showing great linearity
within the force range of 0-60 mN.

Mechanical crosstalk was also assessed by indenting a single sensor and monitoring
the responses of the other sensors in the array concurrently (Table 2-2). Crosstalk was
calculated as the ratio of the percentage change in impedance of the site recorded to that of
the site indented. The colored table facilitates visualization of relative crosstalk magnitudes.
As indicated in the table below, mechanical crosstalk between sensors was generally ~2-
10%, allowing for independent discrimination of forces based on location along the length
of the sensor array.

59
Table 2-2. Table of crosstalk values obtained for sensor calibration experiments.
Crosstalk was calculated by a ratio of the percent impedance change of the site recorded
over the percent impedance change of the site indented. Mean ± SE, n = 3. *For rows 1
and 4, n = 2.

Site Recorded
Site Indented
1 2 3 4 5 6 7
1 1.00±0.00 0.03±0.01 0.01±0.01 -0.02±0.03 -0.01±0.02 -0.01±0.03 -0.01±0.01
2 0.23±0.27 1.00±0.00 -0.03±0.03 -0.04±0.03 -0.02±0.02 -0.13±0.12 -0.19±0.13
3 -0.14±0.10 0.05±0.20 1.00±0.00 -0.06±0.07 -0.05±0.05 -0.07±0.08 -0.05±0.05
4 0.03±0.02 -0.25±0.27 0.08±0.21 1.00±0.00 0.09±0.04 0.07±0.04 0.08±0.03
5 -0.03±0.02 -0.03±0.03 -0.01±0.01 0.05±0.02 1.00±0.00 0.02±0.02 0.01±0.01
6 -0.01±0.01 -0.01±0.02 0.02±0.01 -0.01±0.02 -0.07±0.14 1.00±0.00 0.05±0.02
7 -0.02±0.00 0.01±0.01 -0.01±0.01 0.03±0.03 0.02±0.02 0.12±0.02 1.00±0.00

2.7 Annealing effects on sensor performance
As mentioned previously, multi-layered Parylene devices encounter delamination
failure during prolonged soak conditions that can have detrimental effects on device
performance. One technique used to improve the adhesion between the layers is an
annealing process designed to increase polymer chain entanglement between the two
Parylene layers and provide additional physical anchoring between the chains to resist
delamination. Annealing is achieved by treating the device to temperatures above the glass
transition point (Tg) of Parylene to provide enough energy for movement of the polymer
chains.
However, the heat treatment of Parylene is also known to cause crystallinity
changes to the bulk polymer (hence the importance of Parylene’s thermal budget during
device fabrication) [24], which may or may not influence device performance. To explore
the annealing effects on multi-layered Parylene device performance, we investigated the
post-annealing performance of the Parylene-based EC-MEMS force sensor array. The
devices were annealed to prevent delamination resulting from chronic soaking in saline.
More specifically, sensor arrays were placed within a vacuum oven (10 mTorr), which was
ramped at 1.6°C/min to 200°C. Following a thermal soak time of 48 hours, the devices
were cooled overnight (~15 hours) under vacuum, and then removed for post-process
60
characterization. Untreated sensors, just released off the wafer, were also tested for
comparison.
As the mechanical and electrochemical properties of the sensor drive its
performance, the effects of annealing on these two aspects were analyzed. The bulk
material and mechanical properties of the sensors were assessed before and after annealing
via scanning electron microscopy (SEM), profilometry, and load-deflection tests of the
sensing structures immersed in an electrolyte solution (1x PBS). Two point EIS (1× PBS,
20 Hz–1 MHz), a commonly used electrochemical technique to assess electrode surface
properties, was used between adjacent electrodes to determine changes to the
electrochemical properties of the sensing electrodes.
2.7.1 Mechanical characterization
SEM analysis revealed no significant qualitative differences between the two
samples (Figure 2-18a). However, profilometric measurements indicated a ~3% reduction
in the height (~1 µm) of the microchannel structure following annealing (Figure 2-18b),
analogous to bulk shrinkage effects previously observed in lateral thin film structures [25].
Analysis of the electrode spacing indicated no significant shrinkage of the pitch of the
electrodes.

61

Figure 2-18. SEM images of top-down views of (a) untreated and (b) annealed sensing
structures. (c) Representative profilometry measurement of untreated and annealed
Parylene microchannel sensing structure illustrating ~3% shrinkage following annealing.

Load-deflection tests revealed a ~1.6x increase in the stiffness of the structure
following annealing (Figure 2-19), likely due to a combination of microchannel height
shrinkage and an increase in the elastic modulus of annealed Parylene bulk polymer
attributed to increased polymer crystallinity [26].

Figure 2-19. Results of load-deflection tests comparing untreated and annealed sensors
illustrating an increased structural stiffness of ~1.6x after annealing.
62

2.7.2 Electrochemical characterization
EIS results indicated an increase in the slope of the capacitive region of the
magnitude plot and narrowing of the phase plot, attributed to a decrease in the surface
roughness of the electrodes following annealing, due to the smoothing of the thin-film
metal [27]. The smoothing reduces the capacitance dispersion of the Cdl element of the
Randles circuit model, which in turn diminishes the deviation from the ideal capacitive
impedance within the model in these frequencies, creating the observed response [28].
Despite these slight differences, the measurement frequency (peak of phase plot) remained
consistent following annealing (Figure 2-20). The effect on the electrochemical properties
of the electrodes can be hypothesized to be correlated with electrode size-dependent
phenomena in altering electrode surface properties; annealing effects on electrochemical
properties of Pt on Parylene electrodes have been demonstrated to be magnified in devices
with smaller electrode sizes (Ø = 40 µm) [29].

Figure 2-20. EIS plots of (a) impedance magnitude and (b) phase for electrodes of
untreated (black) and annealed (blue) Parylene sensors. A change in the slope of the
capacitive region of the magnitude plot and narrowing of the phase also in the capacitive
region following annealing corresponds to a decrease in the surface roughness of the
electrode. Dashed line in (b) corresponds to the measurement frequency chosen as the peak
of the phase plot.
63

2.7.3 Impact on sensor performance
A combination of the mechanical and electrochemical changes following annealing
significantly altered sensor performance as observed during calibration (Figure 2-21).
Device sensitivity was decreased from αuntreated = 1.25×10
-3
(normalized impedance/mN) to
αannealed = 9.25×10
-4
(normalized impedance/mN). A reduction in the sensitivity of the
sensor by 24% after the annealing process follows the increase in stiffness of the sensing
structure. For a stiffer channel, applied forces induce smaller channel deflections, thus
reducing sensor sensitivity. Finite element modeling studies utilizing COMSOL to model
the observed changes following annealing elucidated that the increase in stiffness of the
Parylene film contributed the most (rather than microchannel height reduction or decrease
in electroactive surface area) to the differences of sensor performance for these devices.

Figure 2-21. A marked difference in sensor calibration performance (reduction of ~24%)
was observed due to the combination of mechanical and electrochemical effects of
annealing on Parylene sensors. Sensitivity (α) units are in normalized impedance/mN.

64
2.8 Sensor instrumentation
Following complete sensor characterization, instrumentation of the sensors on
cortical probes was achieved (Figure 2-22). The cortical probes used within this study were
provided by a collaborator, Dr. Greg Gerhardt of the University of Kentucky [30]. These
cortical probes are made from a ceramic substrate, chosen to improve signal-to-noise ratios
specifically for electrochemical sensing applications as well as improve biocompatibility,
as ceramics have shown promising inert responses in vivo. To ensure that the recording and
electrochemical sensing capabilities of the ceramic probe was not lost, the sensor arrays
were attached to the back-side (non-electrode-side) of the probes using a thin layer of
biocompatible superglue (Adhesive Systems MG 100 USP Class VI) (Figure 2-23). A
biocompatible adhesive was chosen to minimize immunological responses to the
instrumented shank. The sensor array was manually aligned and adhered to the ceramic
probe. Handling tabs designed into the array allowed for precise manual placement. For
benchtop testing, instrumented ceramic electrode shanks were then interfaced to an acrylic
test jig, which allowed attachment to the insertion setup (Figure 2-24).

Figure 2-22. (a) Image of the ceramic cortical shank instrumented with fully packaged
sensor array. (b) Image of two designs of the cortical probe, sharp and flat, that were
instrumented with two different designs of the sensor array.

65

Figure 2-23. Illustration depicting the instrumentation of the sensors onto the cortical
probes. Following release of the sensor array off the wafer, the device is aligned and affixed
to the cortical probe using biocompatible adhesive. Following attachment, the instrumented
sensors are filled and ready for testing.

Figure 2-24. Image of instrumented probes on an acrylic jig for benchtop insertion testing.

66
2.9 Benchtop insertion experiments
2.9.1 Insertion setup
The cortical sensor array was demonstrated on benchtop by replicating an
implantation via probe insertion into 0.5% agarose (A9539-50G; Sigma-Aldrich, St. Louis,
MO), a commonly used model for cortical tissue [31, 32], immersed in a 1× phosphate
buffered saline (PBS) solution to mimic in vivo conditions. Prior to the experiment, sensor
arrays were first filled via immersion with 1× PBS electrolyte following a short immersion
in isopropyl alcohol (a Parylene wetting step to facilitate diffusion of the polar PBS into
the microchannel). Then, instrumented ceramic probes were inserted to a depth of 3.5 mm
into agarose at three different speeds: 0.01 mm/s (slow), 0.03 mm/s (medium), and 0.1
mm/s (fast), held at the maximum displacement for five minutes, and extracted at the same
speed while sensor array impedances were measured in real time. Impedance was measured
by a LabVIEW-interfaced precision LCR meter (1 Vpp sinusoid, fmeas = 1 kHz) via a
multiplexing PCB for multi-channel impedance measurement across the seven sensors of
the array.
An external 50 g load cell was used for measurement of normal forces generated
during insertion. The probe and load cell were affixed to a motorized micropositioning
stage that controlled insertion speed and depth into the tissue phantom (Figure 2-25a). The
motorized stage, LCR meter, and PCB digital switching all interfaced with LabVIEW to
give the user control of insertion speed and depth as well as a real-time image (graphical)
of the impedance changes (interfacial forces) across the ceramic probe (Figure 2-25b).

67

Figure 2-25. (a) Schematic of benchtop insertion testing setup. (b) Screenshot image of
user GUI during benchtop testing: (Left) a video stream of the insertion process using a
USB microscope. (Right) heat map of GUI to show the distribution of force magnitudes
along the length of the probe. Images illustrate the progression of the instrumented probe
from (i) before insertion, (ii) only tip inserted, (iii) half of probe inserted.

2.9.2 Normal insertion forces
Normal neural probe insertion forces were reported by others [4, 33-38], but are
briefly mentioned here to provide a more complete picture of insertion mechanics.
Observed normal forces during insertion of the probe are attributed to three main forces
encountered during implantation: cutting forces at the probe tip during probe penetration,
and frictional forces between the probe surfaces and surrounding tissue and tissue
deformation forces following penetration [4, 33]. Because of this, normal insertion forces
observed during probe insertion can vary depending on a multitude of factors attributed to
probe implantation: e.g. probe dimensions, probe material and surface (i.e. surface
chemistry, cleanliness), insertion depth, insertion speed, etc. Examples of maximum
normal forces observed for differing probe designs and insertion parameters in literature
are given in Table 2-3 on the following pages.
68
Table 2-3. Summary of insertion studies in literature exploring normal insertion forces of intracortical probe insertion.
Group
Probe
Material
Probe
Shape
Tip
Angle
Width
(µm)
Thickness
(µm)
Insertion
Speed
(mm/s)
Insertion
Depth
(mm)
Model
Dura
Retracted
Normal
Fmax
(mN)
Jensen et al.
2006 [35]
Silicon
Nitride
Sharp 4° 200 25 2 2 Cortex Y 1.03
Jensen et al.
2006 [35]
Tungsten Sharp 4° 50 Ø - 2 2 Cortex Y 0.74
Jensen et al.
2006 [35]
Tungsten Sharp 10° 50 Ø - 2 2 Cortex Y 0.99
Jensen et al.
2006 [35]
Tungsten Sharp 10° 150 Ø - 2 2 Cortex Y 1.65
Sharp et al.
2006 [37]
Steel Flat - 100 Ø - 0.80 1.5
Olfactory
Bulb
- 0.6
Paralikar et al.
2006 [36]
Tungsten Flat - 50 Ø - 0.01 1 Cortex Y 3.65
Hosseini et al.
2007 [34]
Silicon Sharp 17° 120 100 0.017 3
0.6%
Agar +
PE foil
- 43
Hosseini et al.
2007 [34]
Glass Sharp 49.3° 500 175 1.667 6
0.6%
Agar +
PE foil
- 33.6
Hosseini et al.
2007 [34]
Tungsten Sharp 15°-20° 135 Ø - 0.017 3
0.6%
Agar +
PE foil
- 13.8
Hosseini et al.
2007 [34]
Polyimide Sharp 66° 460 10 0.083 4
0.6%
Agar
- 0.33
69
Sharp et al.
2009 [4]
Stainless
Steel
Flat - 100 Ø - 0.822 1.5 Cortex Y 0.833
Sharp et al.
2009 [4]
Stainless
Steel
Flat - 100 Ø - 0.104 1.5 Cortex Y 0.564
Sharp et al.
2009 [4]
Stainless
Steel
Flat - 100 Ø - 0.011 1.5 Cortex Y 0.519
Sharp et al.
2009 [4]
Stainless
Steel
Flat - 200 Ø - 0.822 1.5 Cortex Y 2.481
Sharp et al.
2009 [4]
Stainless
Steel
Flat - 200 Ø - 0.104 1.5 Cortex Y 1.7
Sharp et al.
2009 [4]
Stainless
Steel
Flat - 200 Ø - 0.011 1.5 Cortex Y 1.27
Sharp et al.
2009 [4]
Stainless
Steel
Sharp 20° 200 Ø - 0.822 1.5 Cortex Y 1.358
Welkenhuysen
et al. 2011 [38]
Silicon Sharp 20° 200 100 0.01 9 Cortex Y 1.144
Welkenhuysen
et al. 2011 [38]
Silicon Sharp 20° 200 100 0.1 9 Cortex Y 4.535
Andrei et al.
2012 [33]
Silicon Sharp 30° 300 100 0.01 3 Thalamus Y 1.5
Andrei et al.
2012 [33]
Silicon Sharp 30° 300 100 0.1 3 Thalamus Y 4.5

70

In benchtop insertion experiments for the instrumented ceramic probes in the
present study (sharp shape, tip angle ~33.4°, probe thickness = 127 µm, probe width = 120
µm), normal forces encountered for the three speeds were comparable to those of similar
dimensions (Table 2-4). Slightly higher normal forces for these ceramic probes can be
attributed to the ceramic material and cleanliness of the probe surface; contrary to other
studies in literature, these probes were not cleaned in any way prior to benchtop testing,
which can have an effect on increased frictional forces during the implantation due to
increased surface area [33, 35]. We also observed that with increasing insertion speeds, the
normal insertion forces increase, agreeing with other studies in literature [4, 33, 38], with
forces nearly doubling with a 10× increase in speed (Table 2-4). This increase in normal
force at faster speeds is due to the increased resistance to insertion by the tissue (i.e.
increased tissue deformation and frictional forces) that is encountered with increases in
insertion speed. Slower speeds can allow for the tissue to relax during the insertion and
create a damping-type effect on the normal insertion force [33]. For a more thorough
explanation of normal insertion forces during probe insertion, we point the reader towards
two references that have great modeling and in vivo data, analyzing each attribute of the
normal insertion force in detail [4, 33].

Table 2-4. Representative table of encountered normal forces for a single trial insertion
experiment into 0.5% agarose for different speeds.
Insertion speed
(mm/s)
Maximum normal force
(mN)
0.01 7.99
0.03 8.94
0.1 15.60

71
2.9.3 Shank interfacial forces
Two insertion experiments were performed. In both insertion experiments, some of
the observed normalized impedance values (and thus forces) greatly exceeded the
calibrated linear range of the sensor (Figure 2-26). Possible sources for the observed are:
(1) observing force ranges much higher than expected and (2) differing displacement
profiles (i.e. non-normal to top surface) of the sensor from what was characterized. Slight
changes in electrolyte conductivity from agarose gel filling into the channel through the
fluidic ports during insertion may occur, but may be unlikely as diffusion from a gel into
solution would occur at a longer time scale than mechanical deformation during insertion
at these speeds. Despite this however, the percentage change in impedance data suggests a
trend in interfacial pressures experienced as a function of insertion speed.

Figure 2-26. Representative plot of normal force (purple dash-dot) as well as impedance
responses (solid) for an insertion experiment into 0.5% agarose.

Noting that sensors 1-7 begin at the shank tip and move to its base, data indicated
that a majority of the interfacial forces during insertion are within the first millimeter of
the electrode shank (the first sensor). This is also evidenced by the results across all three
72
insertion speeds; the maximum percentage change in impedance of the first sensor is
considerably higher than that of the remaining sensors higher up on the shank (Table 2-5).
We attribute the interfacial forces experienced at the tip during insertion to the tissue
displacement and propagation of the electrode shank track generated by the probe.
Additional force is imposed by tissue as it is being displaced during insertion.
By varying insertion speed, we observed that interfacial forces at the shank tip (first
millimeter) decrease as the speed increases. Slower insertion speeds allow for a longer
travel time through tissue, increasing shank-tissue adhesion and thus the frictional forces
between the shank and tissue. Results confirm that faster speeds decrease the interfacial
forces experienced at the tip (by tissue displacement and frictional forces) [1] as the
maximum impedance change is lower by nearly 60%.
This effect of speed on interfacial forces is consistent throughout the length of the
shank. At the slowest speed, increased frictional forces resulting in additional array
deformation during insertion give rise to higher average impedance change in other
sensors; Sensors 2-7 recorded an average maximum impedance change of 12% compared
to 2% and 1% at the medium and fast speeds, respectively. Therefore, faster speeds also
reduce adhesion forces between the shank and surrounding tissue along the length of the
tip.
Table 2-5. Table of observed distributions of impedances along the length of the probe
for two different insertion experiments.
Experiment #
Insertion speed
(mm/s)
Maximum ΔZ%: Sensor 1
(shank tip)
Average maximum ΔZ%:
Sensors 2-7 (n=6, Mean SE)
1
0.01 102 % 20 ± 1.9 %
0.03 70 % 7 ± 0.8 %
0.1 26 % 4 ± 0.4 %
2
0.01 87 % 12 ± 3 %
0.03 38 % 2 ± 1 %
0.1 28 % 1 ± 0.6 %

73
2.9.4 Micromotion
Electrode shank instrumentation also provides the benefit of additional sensing
capabilities. More specifically, it has been reported that micromotion of the brain occurs
due to pulsed blood flow and respiration [39]. Micromotion events can affect implanted
probes and result in disrupted recording due to positional changes. We also demonstrate
the ability to detect micromotion events such that sensory feedback can be used to
distinguish neural signals from micromotion events (Figure 2-27).

Figure 2-27. Results from simulated micromotion experiments. Micromotion was
produced by displacing the agarose substrate ±1 mm normal to shank face. Highlighted
region contains 3 micromotion events.

2.10 Insertion mechanics considerations
Benchtop insertion data indicated that interfacial forces were concentrated at the tip.
As these interfacial forces are a result of tissue strain during insertion, we would expect to
find a similar trend in vivo with high tissue strains localized at the tip region. This is
supported by modeling [9, 11] and in vitro [4] efforts that assess tissue strains during the
insertion of stiff probes. The probe track is often compared to a mode 1, or opening mode,
of a crack, which produces a triangular profile that lends to higher interfacial forces at the
74
tip (Figure 2-28). These results are further confirmed by an ex vivo study by Bjornsson et
al. [12], where probe insertions established twice the tissue strain at the tip compared to
the base of the probe. Even within in vivo experiments [38, 40], increased glial response
and reduced neuronal presence was observed at the tip of the probe, most likely due to the
relatively high forces and tissues strains that can cause additional damage. Our work and
those in literature suggest that probe designs should have electrodes concentrated away
from the tip to minimize recording zones within these highly damaged regions.

Figure 2-28. Illustration of probe insertion forming a triangular profile similar to a mode
1 opening crack.

In addition to probe design, another important factor that influences tissue damage
is the speed of insertion. The precise effect of speed on tissue damage is still debated, but
in general, slower speeds can drag and compress the surrounding tissue and create a path
of damage that can extend the kill zone of the implant beyond the probe track as far as 300
µm away [12, 38]. Faster insertion speeds can reduce tissue damage by swiftly transecting
the tissue and vasculature during implantation, which cause significantly less tissue
compression [38, 41]. In comparing insertion speeds, our results indicated that faster speeds
produced smaller interfacial forces, which suggests smaller tissue strains (and thus
damage) during insertion. This result is further supported by ex vivo insertion mechanics
75
studies where faster insertion rates (~2 mm/s) were found to decrease microvasculature
damage and tissue strain compared to slower speeds (125-500 µm/s) [12]. Unfortunately,
tissue responses to insertion speeds of 125 and 500 µm/s [12] and 10 and 100 µm/s [38]
were found to be similar, suggesting that insertion mechanics are comparable within the
range of 10-500 µm/s for probes with cross sectional dimensions of ~100 x 100 µm. Probes
with larger dimensions (e.g. ~1 mm in diameter for deep-brain stimulation electrodes)
might observe greater differences between speeds.
Insertion speeds greater than 2 mm/s have been known to compensate for blunt
probe design, as observed in a speed of 8.3 m/s used to pneumatically implant the three-
dimensional, 100 site Utah electrode array to avoid elastic compression of the cortical
surface [42]. On the other side of the spectrum, some efforts target very slow speeds, on
the range of neuronal or microglial movement (0.42 µm/s), to displace cells and vasculature
without tearing them, and have also demonstrated improved tissue immune response [37].
Overall these results suggest that speeds (whether extremely fast or slow) are chosen based
on the probe structure and design to minimize tissue damage, and should be considered on
an individual probe to probe basis. It is the hope that additional experiments with this
technology, in combination with histological data and in vivo neurophysiological
measurements, can help to clarify or confirm some of these debated topics to realize the
promise of neural prosthetics.
A quick note is to be made regarding the insertion technique. Previous work by
Edell et al. [41] illustrated that any type of lateral movement during probe insertion (e.g.
caused by vibrations of the probe) can extend the kill zone of the probe track and be
detrimental for probe performance. This supports the use of a stereotaxic apparatus and
low vibration motorized drive to limit any extraneous wobble that may be present in the
manual insertion of probes. As seen in a multitude of insertion mechanics studies [12, 35,
37, 38, 40, 41], miniature microdrives and microstages are used to insert probes into the
cortex to prevent unnecessary damage and have even demonstrated longer recording
76
longevity [43]. In some cases a vibration isolation table [38] and suspended animal
respiration [41] is used for further vibration prevention.

2.11 Summary
Within the field of biomedical engineering, EC-MEMS sensors open a new door in
the development of a novel sensor/sensing mechanism that is simple to fabricate and
implement. Here we have demonstrated the design, fabrication, characterization, and
benchtop implementation of a Parylene-based force sensor array designed for mechanics
studies of implanted intracortical probes. We have demonstrated effects of annealing, or
any type of heat-treatment at elevated temperatures, on Parylene-based devices, including
the mechanical and electrochemical changes and resulting effects on the sensor
performance. We have also presented a method to attach these thin, low-profile sensors on
ceramic cortical probes and have begun conversations on quantitative data on the effects
of insertion speed on the interfacial forces and micromotion. Force ranges and sensitivity
for a particular EC-MEMS force sensor can also be tuned by changes in device geometry
and thickness to accommodate interfacial force ranges expected during insertion.
Specifically in looking at insertion mechanics, an analysis limited to normal
insertion forces might suggest that slower insertion speeds benefit insertion by decreasing
the forces encountered during probe implantation. However, before this conclusion can be
made, it is also important to consider interfacial forces and the effects of tissue
displacement and frictional forces; faster insertion speeds not only reduce the forces at the
shank tip created during insertion, but also the frictional forces observed. Measurement of
interfacial forces elucidates the locations and magnitudes of tissue displacement and
frictional forces during insertion.
Although the ideas presented here limit the technology to measuring force, there
are a wide variety of other applications for Parylene substrate, impedance-based devices
that opens an entirely new subfield of MEMS-based medical sensors. Currently, there is
77
great interest within the MEMS research community on flexible polymer-based sensors,
and these Parylene-based EC-MEMS sensors represent a critical advancement, especially
in the area of sensors compatible with surgical or in vivo use. It is the hope of the author
that the development of this Parylene-based novel sensing system will spur research on
additional sensors/sensing modalities outside of traditional silicon-based sensors currently
used in medical technology, to allow for the development of biomimetic or bio-inspired
sensors that are more integratable with the human body.

2.12 References
[1] V. S. Polikov, et al., "Response of brain tissue to chronically implanted neural
electrodes," J. Neurosci. Methods, vol. 148, pp. 1-18, 2005.
[2] A. P. Kelsey, et al., "Stab injury and device implantation within the brain results in
inversely multiphasic neuroinflammatory and neurodegenerative responses," Journal of
Neural Engineering, vol. 9, p. 046020, 2012.
[3] A. Gilletti and J. Muthuswamy, "Brain micromotion around implants in the rodent
somatosensory cortex," Journal of Neural Engineering, vol. 3, pp. 189-195, Sep 2006.
[4] A. A. Sharp, et al., "In Vivo Penetration Mechanics and Mechanical Properties of
Mouse Brain Tissue at Micrometer Scales," Biomedical Engineering, IEEE Transactions
on, vol. 56, pp. 45-53, 2009.
[5] R. Hoffmann, et al., "Comparative Study on the Insertion Behavior of Cerebral
Microprobes," in Engineering in Medicine and Biology Society, 2007. EMBS 2007. 29th
Annual International Conference of the IEEE, 2007, pp. 4711-4714.
[6] R. Das, et al., "A Benchtop System to Assess Cortical Neural Interface
Micromechanics," Biomedical Engineering, IEEE Transactions on, vol. 54, pp. 1089-1096,
2007.
[7] T. ChunXiang and H. Jiping, "Monitoring Insertion Force and Electrode Impedance
during Implantation of Microwire Electrodes," in Engineering in Medicine and Biology
Society, 2005. IEEE-EMBS 2005. 27th Annual International Conference of the, 2005, pp.
7333-7336.
78
[8] D. H. Szarowski, et al., "Brain responses to micro-machined silicon devices," Brain
Research, vol. 983, pp. 23-35, Sep 5 2003.
[9] S. P. DiMaio and S. E. Salcudean, "Needle insertion modelling and simulation," in
Robotics and Automation, 2002. Proceedings. ICRA '02. IEEE International Conference
on, 2002, pp. 2098-2105 vol.2.
[10] L. Hyunjung, et al., "Biomechanical analysis of silicon microelectrode-induced
strain in the brain," Journal of Neural Engineering, vol. 2, p. 81, 2005.
[11] J. Subbaroyan, et al., "A finite-element model of the mechanical effects of
implantable microelectrodes in the cerebral cortex," Journal of Neural Engineering, vol. 2,
p. 103, 2005.
[12] C. S. Bjornsson, et al., "Effects of insertion conditions on tissue strain and vascular
damage during neuroprosthetic device insertion," Journal of Neural Engineering, vol. 3, p.
196, 2006.
[13] K. Seidl, et al., "CMOS-based high-density silicon microprobe for stress mapping
in intracortical applications," in Micro Electro Mechanical Systems (MEMS), 2010 IEEE
23rd International Conference on, 2010, pp. 35-38.
[14] C. A. Gutierrez and E. Meng, "Impedance-Based Force Transduction Within Fluid-
Filled Parylene Microstructures," Journal of Microelectromechanical Systems, vol. 20, pp.
1098-1108, Oct 2011.
[15] C. A. Gutierrez, et al., "An Implantable All-Parylene Liquid-Impedance Based
Mems Force Sensor," MEMS 2010: 23rd IEEE International Conference on Micro Electro
Mechanical Systems, Technical Digest, pp. 600-603, 2010.
[16] C. A. Gutierrez and E. Meng, "A Dual Function Parylene-based Biomimetic Tactile
Sensor and Actuator for Next Generation Mechanically Responsive Microelectrode
Arrays," in Transducers, Denver, CO, USA, 2009, pp. 2194-2197.
[17] J. E. B. Randles, "Kinetics of rapid electrode reactions," Discussions of the Faraday
Society, vol. 1, pp. 11-19, 1947.
[18] H. Helmholtz, "Studien uber electricshe Grenzschichten," Annalen der Physik, vol.
243, pp. 337-382, 1879.
[19] D. A. Robinson, "The electrical properties of metal microelectrodes," Proceedings
of the IEEE, vol. 56, pp. 1065-1071, 1968.
79
[20] B. J. Kim, et al., "Parylene-based electrochemical-MEMS force sensor array for
assessing neural probe insertion mechanics," in Micro Electro Mechanical Systems
(MEMS), 2012 IEEE 25th International Conference on, 2012, pp. 124-127.
[21] C. A. Gutierrez, "Development of Flexible Polymer-Based MEMS Technologies
for Integrated Mechanical Sensing in Neuroprosthetic Systems," Ph.D. Dissertation,
Biomedical Engineering, University of Southern California, Los Angeles, 2011.
[22] C. A. Gutierrez, et al., "Epoxy-less packaging methods for electrical contact to
parylene-based flat flexible cables," in Solid-State Sensors, Actuators and Microsystems
Conference (TRANSDUCERS), 2011 16th International, 2011, pp. 2299-2302.
[23] W. Jiun-Min, et al., "Design and fabrication of a diaphragm type thermo-buckled
microactuators," in Nano/Micro Engineered and Molecular Systems, 2006. NEMS '06. 1st
IEEE International Conference on, 2006, pp. 873-876.
[24] J.-M. Hsu, et al., "Effect of Thermal and Deposition Processes on Surface
Morphology, Crystallinity, and Adhesion of Parylene-C," Sensors and Materials, vol. 20,
pp. 071-086, 2008.
[25] B. J. Kim, et al., "Three dimensional transformation of Parylene thin film structures
via thermoforming," in Micro Electro Mechanical Systems (MEMS), 2013 IEEE 26th
International Conference on, 2013, pp. 339-342.
[26] R. Metzen and T. Stieglitz, "The effects of annealing on mechanical, chemical, and
physical properties and structural stability of Parylene C," Biomedical Microdevices, pp.
1-9, 2013/03/15 2013.
[27] M. Grosser and U. Schmid, "The impact of annealing temperature and time on the
electrical performance of Ti/Pt thin films," Applied Surface Science, vol. 256, pp. 4564-
4569, 2010.
[28] T. Pajkossy, "Impedance spectroscopy at interfaces of metals and aqueous solutions
- Surface roughness, CPE and related issues," Solid State Ionics, vol. 176, pp. 1997-2003,
Aug 2005.
[29] B. J. Kim, et al., "Evaluation of Post-Fabrication Thermoforming Process for
Intracortical Parylene Sheath Electrode," in 6th International IEEE EMBS Neural
Engineering Conference, 2013, pp. 379-382.
[30] J. J. Burmeister, et al., "Ceramic-based multisite microelectrodes for
electrochemical recordings," Analytical Chemistry, vol. 72, pp. 187-192, Jan 1 2000.
80
[31] Z. J. Chen, et al., "A realistic brain tissue phantom for intraparenchymal infusion
studies," Journal of Neurosurgery, vol. 101, pp. 314-322, Aug 2004.
[32] F. Pervin and W. W. Chen, "Mechanically Similar Gel Simulants for Brain
Tissues," in SEM Annual Conference, Indianapolis, Indiana, USA, 2010, pp. 9-13.
[33] A. Andrei, et al., "A response surface model predicting the in vivo insertion
behavior of micromachined neural implants," Journal of Neural Engineering, vol. 9, p.
016005, 2012.
[34] N. Haj Hosseini, et al., "Comparative study on the insertion behavior of cerebral
microprobes," in Engineering in Medicine and Biology Society, 2007. EMBS 2007. 29th
Annual International Conference of the IEEE, 2007, pp. 4711-4714.
[35] W. Jensen, et al., "In-vivo implant mechanics of flexible, silicon-based ACREO
microelectrode arrays in rat cerebral cortex," Biomedical Engineering, IEEE Transactions
on, vol. 53, pp. 934-940, 2006.
[36] K. J. Paralikar, et al., "Collagenase-aided Insertion of Intracortical Microelectrode
Arrays: Evaluation of Insertion Force and Chronic Recording Performance," in
Engineering in Medicine and Biology Society, 2006. EMBS '06. 28th Annual International
Conference of the IEEE, 2006, pp. 2958-2961.
[37] A. A. Sharp, et al., "Toward a self-deploying shape memory polymer neuronal
electrode," Journal of Neural Engineering, vol. 3, p. L23, 2006.
[38] M. Welkenhuysen, et al., "Effect of Insertion Speed on Tissue Response and
Insertion Mechanics of a Chronically Implanted Silicon-Based Neural Probe," Biomedical
Engineering, IEEE Transactions on, vol. 58, pp. 3250-3259, 2011.
[39] L. Xindong, et al., "Stability of the interface between neural tissue and chronically
implanted intracortical microelectrodes," Rehabilitation Engineering, IEEE Transactions
on, vol. 7, pp. 315-326, 1999.
[40] A. C. Hoogerwerf and K. D. Wise, "A three-dimensional microelectrode array for
chronic neural recording," Biomedical Engineering, IEEE Transactions on, vol. 41, pp.
1136-1146, 1994.
[41] D. J. Edell, et al., "Factors influencing the biocompatibility of insertable silicon
microshafts in cerebral cortex," Biomedical Engineering, IEEE Transactions on, vol. 39,
pp. 635-643, 1992.
81
[42] P. Rousche and R. Normann, "A method for pneumatically inserting an array of
penetrating electrodes into cortical tissue," Annals of Biomedical Engineering, vol. 20, pp.
413-422, 1992/07/01 1992.
[43] R. L. Rennaker, et al., "A comparison of chronic multi-channel cortical
implantation techniques: manual versus mechanical insertion," Journal of Neuroscience
Methods, vol. 142, pp. 169-176, 2005.

82

The potential of intracortical prosthetics has been marred by the limited in vivo
lifetime of current technology due to biological and non-biological failure modes. If
intracortical prosthetic devices functioned reliably under chronic conditions (in the range
of multiple years), the promise of intracortical recording electrodes as a neural prosthetic
can be realized. In an effort to improve the reliability of the current state-of-the art, we
present a novel Parylene-based approach, leveraging Parylene’s thin, flexible substrate and
compatibility with batch microfabrication techniques with a unique probe architecture to
enhance implant-tissue integration in vivo.

3.1 Background
Amidst a complex cellular environment of vasculature and support cells such as
astrocytes and microglia, the brain consists of a vast network of neurons that facilitate the
transmission of varying efferent/afferent electrical signals to control and receive
information, respectively, to and from the body. These signals are abundant as they carry
the commands to complete a variety of tasks such as the movement of the arm, which itself
is an elaborate conversation of movement commands as well as feedback signals from
proprioceptive receptors to successfully execute a reach task. As these signals are electrical
in nature, we have the opportunity to intercept these commands (record) or mimic them
(stimulate) using implanted intracortical electrodes to construct enabling technology for
motor control. This is the basis for neural prosthetics in the development of brain-machine
interfaces (BMIs) to restore functionality to loss-of-limb or paralyzed patients [1, 2].

THREE-DIMENSIONAL PARYLENE SHEATH ELECTRODE
ARRAY FOR CHRONIC APPLICATIONS
83

3.1.1 Intracortical electrodes for recording
In the case of recording these command signals for use in BMIs, arrays of
intracortical electrodes implanted within the brain are used to measure the extracellular
electrical activity associated with limb movement transmitted from neurons within the
cortex. These chronically implanted intracortical recording electrode arrays are at the cusp
of biomedical research in an effort to not only realize BMI technologies, but also to study
the activities and organization of the brain. For these applications, there are two main
classes of implantable electrodes that are commonly used for research and clinical
applications: microwires and micromachined probes from bulk silicon.
The microwire type electrodes were one of the first in development, and consist of
a tungsten or iridium microwire that has been electrochemically etched to a very fine taper
for insertion in the cortex [3]. Each microwire is insulated with a passivation layer
(Parylene or polyimide) and laser-ablated at the tips to reveal the recording site, allowing
for specificity and neuronal targeting during recording. Arrays of these microwires can be
constructed by affixing individual microwires into an organized pattern (depending on the
application) onto a base plate (Figure 3-1a). Though extensively used in a variety of animal
and human studies, these devices fail in chronic applications largely due to process
variations and unreliable fabrication techniques, as observed in long-term reliability testing
of these probes [4].
Two silicon-based micromachined chronic probe technologies have enjoyed
widespread adoption: the monolithic, “bed-of-needles” Utah Electrode Array (UEA) [5]
and the planar, multi-site Michigan Array (MA) [6]. The UEA consists of tapered-tip
silicon needles insulated in a polymer, with exposed, platinum electrodes at the tip (Figure
3-1b) and has found success in cats [7], non-human primates [8], as well as human subjects
[9]. The MA is also formed using microfabrication techniques, but entails a sharp silicon
shank with multiple electrode sites along the length of the shank (Figure 3-1c) designed
with integrated CMOS electronics [10] and the capability of expanding to 3D array
84
conformations [11]. These array platforms allow for a dense packing of recording sites and
have demonstrated neural recordings in their respective models over years. However,
reliable implementation of these devices has proven to be difficult as these chronically
implanted neural probes are susceptible to a number of failure modes, detailed in the
following section, reducing their longevity in vivo.

Figure 3-1. Images of the (a) Microwire array (adapted from [12], Copyright 2003 National
Academy of Sciences, U.S.A.), (b) Utah electrode array (Reprinted by permission from
Macmillan Publishers Ltd: Nature Neuroscience [13], copyright 2002), (c) Michigan array
(© 2008 IEEE [14]), all developed intracortical recording technologies.

3.1.2 Issues with implant-tissue interface
In addition to the device-based (non-biological) failure modes (e.g. insulation
delamination, electrode corrosion), which hinge on device construction and packaging
reliability, several biological failure modes have been found to reduce the chronic neural
recording ability due to issues at the probe-tissue interface [15]. As the probe is initially
inserted into tissue, it severely damages vasculature, extracellular matrix, and the processes
of neurons and other glial cells, beginning a wound healing response to the stab wound.
More specifically, the insertion trauma causes the disruption of the blood brain barrier
(BBB) [16], forming a “leaky” BBB, which releases blood-born macrophages (and thus
cytokines) as well as oxidative reactants (such as iron [4]) that cause inflammatory and
neurotoxic conditions surrounding the implant. This situation is further exacerbated by the
immune response of the brain to the chronic presence of a foreign body object [17, 18].
85
Also, there is considerable mechanical mismatch between the stiffness of the metal
microwire or silicon probe and soft cortical tissue, inflicting continuous damage, especially
during micromotion events within the tissue caused by vascular pulsing [19, 20]. These
inflammatory and immune response processes contribute to the formation of a glial cell
encapsulation, or “glial scar”, around the probe and the retraction and death of neurons
near the probe, which prevents high fidelity recordings [21]. Tethering between skull
anchoring of electrical connectors and the implanted probe further contributes to adverse
micromotion or development of interfacial stresses between the stiff neural probes and
cortical tissue [22].
3.1.3 Improving intracortical interfaces
To mediate the previously mentioned issues during chronic implementation of
neural probes, there have been many efforts to increase the reliability of probes in vivo by
a variety of ingenious approaches such as altering probe shape for better probe-tissue
integration (e.g. open probe architectures and movement of recording sites near the edges
of the shank) or adding bioactive coatings (e.g. various anti-inflammatory and neurotrophic
factors on a matrix) to reduce the immune response and encourage neuronal proliferation
[23-25]. One of the major thrusts in the advancement of chronic neural probe technology
is the movement away from stiff, silicon-based probes to more flexible, compliant
polymer-based probes in an effort to create devices that are more conformal and cause less
damage to cortical tissue once implanted [26-30]. Polymer-based MEMS neural probes
allow for flexibility and mechanical material matching that is not capable with microwire
or silicon-based technology, while maintaining the advantages of MEMS-based neural
probes, notably a large number and density of recording sites per probe within a small
footprint. More specifically polyimide (PI), poly(di-methylsiloxane) (PDMS), and more
exotic, customized polymers (e.g. shape memory polymers, functionalized hydrogels) have
all been used in the development of polymer-based intracortical probes and have
demonstrated much success. A summary of the recent development of polymer-based
neural probes is presented below (Table 3-1) to illustrate the wide variety of polymer
86
substrates, electrode materials, design, and insertion strategies used to implant these
compliant probes within the cortex.

Table 3-1. Summary of the recent development of polymer-based neural probes.
Group
Substrate
Material
Electrode
Material
Shape Biofactors
Insertion
Strategy
Takeuchi et al.
2003 [31]
Polyimide Ti
3D shank
array
-
Nickel
backing
plate
Takeuchi et al.
2004 [32]
Parylene Au Shank
Drug
injection via
microchannel
PEG coating
Pellinen et al.
2005 [33]
Parylene Pt/Pt Black Shank
Drug
injection via
microchannel
Not
necessary
Kato et al. 2006
[34]
Parylene Au Shank
Factor
loading in
groove and
coating
PEG coating
Seymour et al.
2006 [24]
SU-8/Parylene - Shank -
Not
necessary
Mercanzini et al.
2007 [35]
Polyimide Pt Shank Coating
Not
necessary
Mercanzini et al.
2007 [36]
Polyimide Pt Shank - -
Kozai et al. 2009
[37]
PDMS/Polyimide - Shank -
SAM coated
insertion
shuttle
Egert et al. 2011
[23]
Parylene Au Shank -
Structural
design
Fan et al. 2011
[25]
Polyimide Ir Fish-bone - Silk coating
Gilgunn et al.
2011[38]
Parylene Pt Wires/shank
Loading in
CMC shuttle
CMC shuttle
Kuo et al. 2012
[39]
Parylene Pt
Conical
sheath
Coating
Insertion
microwire
Kozai et al. 2012
[40]
Carbon fiber/
Parylene N
PEDOT Microwire PEGMA
Not
necessary
Capadona et al.
2012 [41]
Nanocomposite/
Parylene
Au Shank -
Not
necessary
Wu et al. 2013
[42]
Parylene Au Shank - Silk coating
Ware et al. 2014
[43]
Shape memory
polymer (SMP)
Au Shank -
Not
necessary

87
Interestingly, a glass-based probe technology, the neurotrophic cone electrode (NE),
was successfully demonstrated in chronic applications in both animals and humans; the
longevity of recordings is attributed to the unique hollow conical tip coated with growth
factors [44-49]. The NE consists of a glass cone with microwire electrodes manually
affixed within the inner lumen of the cone. The cone is coated with nerve growth factors
to promote the growth of neural processes towards the electrodes as well as to allow for
better integration with the cortical tissue. The release of these factors by the neural probe
in vivo may counteract the adverse physiological responses reported in other chronically
implanted probes. By having the ingrowth of neurons closer to the electrode sites over time,
improved neural signal quality can be achieved as the distance between the neurons and
electrodes is decreased [50]. Also, the effects of micromotion are diminished as the growth
of tissue within and through the cone facilitate integration and anchoring of the probe with
the surrounding tissue.
The NE however, requires a manual and labor intensive process for its fabrication
and possesses a rigid glass structure with a large mechanical mismatch with surrounding
neural tissue. Here we present a Parylene-based sheath electrode that mimics the conical
design and coating strategy of the NE for chronic intracortical recording. By taking
advantage of micromachining processes and polymer construction, the Parylene sheath
electrode seeks to combine the advantages of polymer-MEMS technologies, namely small,
flexible form factor and batch fabrication, with the design of the NE. It is the aim of the
three-dimensional (3D) Parylene-based sheath electrode to employ the sheath design and
neural ingrowth concept of the NE but with a polymer-based substrate to improve tissue
integration and allow more reliable, long-term performance in vivo.
3.2 Parylene sheath electrode
3.2.1 Design
The 3D flexible Parylene sheath electrode (PSE) is constructed from surface-
micromachined Parylene for neuronal recording [51, 52] (Figure 3-2). The PSE consists of
88
a hollow sheath structure with four electrode sites within the lumen and another four
electrode sites on the outer perimeter of the sheath. Dimensions were chosen such that
following implantation, the electrodes would be positioned in layers IV-VI of the rat M1
cortex. The aim of the PSE is to utilize a drug-coated, sheath-based design to minimize the
immune response and encourage neuronal proliferation and growth towards and within the
sheath, which houses the electrode sites and follows the previous success of the NE. The
PSE however, also has added benefits of a flexible, thin (12 µm) substrate material that
would further reduce the mechanical mismatch (and subsequent damage) between the PSE
and surrounding cortical tissue compared to the rigid, glass-based NE. Also because of
Parylene’s compatibility with microfabrication techniques, probes with precise designs can
be batch fabricated to ensure probe-to-probe repeatability and thus increased reliability. An
extended Parylene ribbon cable from the sheath probe tip is also designed in for additional
flexibility and terminates in a linear contact pad array for establishing electrical
connections to the neural probe utilizing a zero-insertion-force (ZIF) connector.

Figure 3-2. Photograph of 3D Parylene sheath probe highlighting the probe tip and a
portion of the integrated Parylene cable.

In this body of work, the PSE had three main design iterations: (1) an initial PSE
with its outer electrodes on the top surface of the sheath, (2) a revised design of the PSE
89
with outer electrodes on the wing edges of the probe used for an acute implantation study,
and lastly (3) a 1 x 2 PSE array, the final design used for a chronic study, that had a tapered
cone and contained 2 PSEs on a single fabricated die (Figure 3-3). 2 of these 1 x 2 PSE
arrays were combined in flexible PCB packaging to form the Parylene sheath electrode
array or PSEA.

Figure 3-3. Optical micrographs of the three design iterations of the Parylene sheath
electrode: (a) the initial design with 4 external electrodes on the top surface of the sheath,
(b) revised design with the 4 electrodes moved to the outer periphery of the probe, (c) a
1x2 PSE array with 2 sheath probes at the tip. [Scale bar of inset is 200 µm]

3.2.2 Fabrication
The 3D Parylene sheath neural probe with integrated Parylene cable was surface
micromachined using a layer-by-layer process, building upon a supported Parylene
substrate to form a Parylene sheath (Figure 3-4). Parylene was chosen as the probe substrate
and structural material because of its USP class VI rating, compatibility with
micromachining processes, and flexibility to improve the mechanical mismatch between
the probe and tissue. Following a dehydration bake (120ºC, 20 minutes), a 5 µm layer of
Parylene was deposited on a bare silicon wafer with its native oxide layer intact. The native
oxide aided release of these sheath probes off of the wafer. Platinum (Pt) metal (2000 Å)
was deposited using e-beam evaporation and patterned using photolithography and a lift-
off process to pattern the inner and outer electrodes (45 µm diameter), leads (10 µm width),
90
and contact pads (350 µm by 3.5 mm). Platinum was chosen as the electrode material for
its biocompatibility, inertness within biological environments [53], and long track record
for neural recordings. A titanium adhesion layer was not used as Pt has been reported to
have good adhesion with Parylene substrates [54], however oxygen plasma was used to
treat and descum the surface prior to metal deposition [55].
The different design iterations of the PSE had varying electrode positions largely
to increase yield with newer designs and overcome process difficulties. In the first design,
the outer electrodes on top of the sheath were fabricated using a dual-layer liftoff technique
on top of the sheath that allowed for continuous electrical wiring across the sidewall of the
Parylene microchannel [51]. However, it was observed that the mechanical opening of the
sheath during post-processing produced tensile stresses sufficient to crack electrode sites
for the probe dimensions investigated here. Therefore, outer electrodes were moved to the
sheath perimeter for the second design iteration. This design simplified the fabrication
process, reduced probe fabrication time, and increased yield. For this reason, the peripheral
outer electrode design remained into the development of the PSEA.
An insulation layer of Parylene (1 µm) was then deposited, patterned, and etched
using oxygen plasma to expose the electrode sites as well as the contact pads. A thick
sacrificial photoresist layer (~10 µm) was spun on and patterned to form the microchannel
structure that would later become the inner lumen of the 3D Parylene sheath. To form the
top surface of the microchannel, a 5 µm layer of Parylene was deposited and openings were
etched at both ends to form the final 2D sheath structure; in the same step the outer
electrodes and contact pads were etched and exposed. The devices were then cut-out using
oxygen plasma, defining the probe shape and cable length. The sacrificial photoresist was
removed using a sequential soak in acetone, isopropyl alcohol, and deionized (DI) water to
reveal the inner lumen of the sheath structure. Sheath probes were released from the wafer
by gently peeling the Parylene devices from the carrier wafer while immersed in DI water.
These microchannel Parylene devices are then thermally formed, or “thermoformed,” to
form the 3D sheath shape. For the PSEA, the fabrication process mentioned prior remained
91
consistent, expect for the addition of an additional perforation etch step that added
perforations to the design for better tissue integration via an open probe architecture.

Figure 3-4. Overview of fabrication steps for 3D Parylene sheath probe. [Note: fabrication
steps as drawn are shown on an already cut-out substrate, however the devices are not cut
to shape during the actual process until step d.] (a) 5 µm of Parylene was first deposited
onto a carrier wafer. (b) Pt electrodes (2000 Å) were e-beam evaporated and patterned
using lift-off. Parylene insulation (1 µm) was deposited and patterned to expose the
electrode sites using oxygen plasma. (c) A sacrificial layer of photoresist was spun on and
patterned to form the microchannel structure. (d) Parylene (5 µm) was deposited on top of
the sacrificial layer to complete the microchannel. Openings were etched into the ends of
the microchannel and the border of the device was also etched to form the final sheath
structure and shape. (e) The sacrificial photoresist is removed using a sequential acetone,
isopropyl alcohol, and DI water soak.

3.3 Development of thermoforming process
3.3.1 Thermoforming of Parylene
Prior to presenting the thermoforming process for forming the 3D sheath structure
for the PSE, it would be helpful to introduce the thermoforming process for Parylene in
creating 3D structures. As mentioned previously, conventional methods to produce 3D
92
Parylene structures predominantly utilize two main strategies: (1) the use of temporary
sacrificial structures such as photoresist [56, 57] or wax [58] to shape Parylene, or (2)
Parylene deposition on molds with topological positive/negative features [59] that then
require film release. These well-established processes, though effective, are characterized
by complex process steps; sacrificial structures require extra steps to ensure their complete
removal, and removal of films from molds without film or other structural damage can be
difficult. Also, these techniques may not be suitable to achieve structures with large aspect
ratios, as the Parylene structure is limited by the photolithographic or mold fabrication
processes employed. To address these limitations, some studies have focused on leveraging
the thermoplastic nature of Parylene [60] to form 3D structures using a hot embossing
process [61], as well as a modified polymer bonding technique [62] for both die level [63,
64] and wafer level bonding [62, 65] to assemble Parylene structures via bonding of a top
open structure and a bottom substrate layer. These processes successfully produced
microchannel structures while eliminating the need for a sacrificial material.
Here, we present an alternative method to form 3D Parylene structures adapted
from previous work in the thermal shaping of Parylene thin films [60, 66, 67]. This strategy
leverages the thermoplastic property of Parylene to thermally form, or “thermoform,” thin
film Parylene into 3D structures in a similar manner to previous efforts involving other
common polymers. Traditionally used as a technique to form large three-dimensional (3D)
structures from planar polymer films in industry-center mass production processes,
thermoforming has been recently extended into the micro-realm to produce micro-scale 3D
structures from polymer films [68]. This forming process involves the reshaping of
thermoplastic polymers, enabled by softening the material by heating above its glass
transition temperature (Tg) while simultaneously confining the material within a mold of
the desired configuration. Upon cooling and removal of the mold, the desired shape is
retained. Specifically within biomedical applications, thermoforming of different polymers
(e.g. polystyrene, polyethylene terephthalate, polylactic acid) has allowed for successful,
mass-scale, repeatable production of lab-on-a-chip as well as other in vitro microfluidic
systems [69-72].
93
Free standing Parylene films (single Parylene thin films or Parylene-Parylene (PP)
dual layers) or devices (Parylene-metal-Parylene (PMP) sandwiches) (Figure 3-5) are
thermoformed by annealing the assemblies (film and mold) in the desired configuration for
a set process time (soak time or tS) at a certain process temperature (soak temperature or
TS). As a semicrystalline polymer, Parylene chains consist of amorphous and crystalline
regions [73] (Figure 3-6). By treating the polymer with soak temperatures between the
glass transition point (~60-90°C) and the melting point (290°C) [63, 74], the amorphous
regions of the polymer gain energy to rearrange, allowing for softening of the polymer and
subsequent molecular relaxation and reorganization of the chains [75, 76]. In
thermoforming experiments, the soak temperature was not increased beyond the melting
point to avoid device or film complications due to large and potentially damaging changes
to the material properties following treatment above that temperature [77].

Figure 3-5. Top and cross-sectional drawings illustrating test samples used in
thermoforming experiments: (left) Parylene on Parylene samples and (right) Parylene-
metal-Parylene (PMP) sandwiches.

94

Figure 3-6. Illustration of semicrystalline, thermoplastic Parylene polymer highlighting
crystalline and amorphous regions.

Parylene free films or devices were assembled into a specific orientation by using
a mold (e.g. microwire, glass rod) to form the desired final shape (Figure 3-7). These
assemblies were then placed within a vacuum oven (TVO-2, Cascade Tek Inc., Longmont,
CO) during the subsequent thermoforming process. It is essential to conduct the process in
a vacuum environment as Parylene undergoes thermal oxidative degradation at
temperatures >125 °C in the presence of oxygen [78]. For the purposes of this study, the
vacuum environment was maintained at 10 mTorr. In previous studies, nitrogen backflow
(10 SCFH) was also used as an extra precaution to remove any oxygen [79], however in
the present study, nitrogen backflow was not present as it was determined that vacuum
alone was sufficient in preventing thermal oxidative degradation.

Figure 3-7. Optical micrographs of molds used to thermoform Parylene 3D structures. (a)
Top and (b) side-view of a microwire inserted into a Parylene microchannel to form conical
structures. (c) A PMP cable is wrapped around a glass rod to form a macro-coil structure.

95
Following placement of the assembled devices into the vacuum oven, the
temperature was increased at a rate of ~1.6°C/minute until the desired soak temperature
was achieved, after which the temperature was maintained for a given soak time. Soak
times were determined by comparable studies of heat treatment of Parylene films and
devices within literature, ranging from 30 mins for Parylene-Parylene bonding [62] to 48
hrs for shaping of Parylene C and improving adhesion between layers [80]. Following a
slow overnight (~15 hrs) cool down of the samples under vacuum by turning off the heating
element, the assemblies were removed from the oven and the molds were easily removed
to reveal the final thermoformed shape.
In addition to the formation of a 3D shape, the final device also attains added
benefits attributed to the annealing of Parylene layers, such as increased adhesion [81],
decreased water permeability and increased dielectric strength [75], and improvement in
thermal stability at higher temperatures, namely >37°C for biological applications [73].
Thermoforming Parylene allows for a simple process to create desired 3D structures with
tunable material properties and expands the potential range of applications of Parylene-
based microelectromechanical systems (MEMS).

3.3.2 Thermoforming the Parylene sheath electrode
To form the 3D sheath shape from the fabricated microchannel of the PSE, a
microwire mold was inserted into the sheath for the thermoforming process (Figure 3-8).
These custom microwire molds were fabricated from tungsten (MicroProbes for Life
Science, Gaithersburg, MD) and stainless steel (Cooner Wire Co., Chatsworth, CA) with
tip ends shaped to match the specific taper designs and were used to thermoform the 3D
sheath structures. Following insertion of the microwire molds into the microchannel, the
opened probes were placed within a vacuum oven and heated to 200ºC. Following thermal
treatment and cooling, the molds were easily removed without the need for release agents,
and probes were packaged for implantation.

96

Figure 3-8. (a) Parylene sheath probe after release from wafer with flat 2D microchannel
structure. (b) Microwire mold inserted into the microchannel for thermoforming process.
(c) Thermoformed result showing opened 3D sheath structure.

Thermoforming with a microwire mold resulted in 3D sheath structures, with the
stiffness of the structure increasing sixty times corresponding to increased crystallinity
resulting from thermal treatment of Parylene [82]. However, the sheath was still flexible
enough to return to its original shape following deformations of 100 µm (Figure 3-9) and
200 µm (Figure 3-10) and did not contribute significantly to the overall stiffness of the
probe.

Figure 3-9. Sequential photographs (side-view of probe) demonstrating the mechanical
robustness of the stiffened cone structure. (a) A deflection probe positioned above the
sheath probe. (b) 100 µm displacement of the top of the sheath with the deflection probe.
(c) Retraction of the deflection probe allowed the sheath to return to its original shape.
Sheath movement is highlighted with a black outline.

97

Figure 3-10. Displacement (black squares) and load force (red triangles) measurements
taken during deflection studies of thermoformed PSE sheath structure. Inset and plot
indicate the (1) starting position, (2) deflection of 200 µm, and (3) return to initial position
indicating following release of load. Sheath remains flexible and is robust to mechanical
perturbations. Sheath deformation is outlined in black to facilitate visualization. Scale bar
is 200 µm.

Conical sheaths of varying tapers ranging from 200 and 300 µm base to 70 µm tip
in diameter and cylindrical sheaths of 300 µm in diameter were produced using the
thermoforming post process. This technique was also successful in forming 3D structures
out of a Parylene microchannel with metal electrode sites (first design iteration) on top of
the structure as well as Parylene microchannels with 15 µm perforations (1x2 PSE array),
exhibiting the robustness of the technique across different initial Parylene structures
(Figure 3-11).

98

Figure 3-11. Examples of thermoformed Parylene 3D structures. (a) A thermoformed PP
conical structure with a sharp taper. (b) A thermoformed PMP cylindrical structure with
electrode sites on the inside and outside of the cylinder. (c) An SEM image of the
cylindrical PMP structure. (d) A thermoformed PMP device formed from a Parylene
microchannel with electrodes on the top. (e) A thermoformed PMP device formed from a
perforated Parylene microchannel. (f) An SEM image of a pair of thermoformed conical
devices highlighting the repeatability of the process. [Scale bar = 200 µm]

3.4 Effects of thermoforming
The chemical, mechanical, and electrochemical effects of thermoforming on the
Parylene bulk material as well as on Parylene-metal devices (i.e. Parylene sheath electrode)
were assessed. Chemical and mechanical characterization of the effects of the
thermoforming process was carried out on Parylene film samples (10 µm thick) that were
deposited on 2.54 x 7.62 cm glass slides that were cut into 10 x 20 mm rectangles (for
chemical characterization) or 10 x 10 mm squares (for mechanical characterization) using
a sharp blade and subsequently released from the glass slides by gentle peeling. The
chemical and mechanical changes following variations to the soak temperature and soak
time during the thermoforming process were investigated to reveal any material
modifications that may have occurred as a result of thermoforming, as well as trends
relevant to process tunability. Experiments in which soak temperature variation was
99
evaluated were performed under a constant soak time of 6 hrs (soak temperatures: 100, 120,
160, 180, and 200°C), whereas soak time variation experiments were conducted under a
constant soak temperature of 200°C (soak times: 0.5, 2, 6, 12, 24, and 48 hrs). A summary
of the testing conditions is given in Table 3-2.

Table 3-2. Summary of testing conditions for soak temperature and time variation effects
on chemical and mechanical properties of Parylene thin films following thermoforming.
Soak temperature variation (°C),
tS = 6 hrs
Soak time variation (hrs),
TS = 200°C
80 0.5
100 2
120 6
160 12
180 24
200 48

3.4.1 Bulk device effects
Thermoforming was performed in sequential processes as well; two-stage
thermoforming was demonstrated in forming 3D micro and macro-structures for device
applications; the hollow sheath cone structure was first formed in one step, and a secondary
macro-coil structure in the incorporated Parylene cable was formed in the second
thermoforming step without affecting the initial 3D Parylene microstructure (Figure 3-12).

100

Figure 3-12. Examples of thermoformed macro-scale coils and micro-scale cones in dual
thermoformed (a) 4 channel and (b) 8 channel PMP devices.

Bulk dimensional changes occurred as a result of thermoforming and may impact
the final desired structure based on the pre-thermoformed design. In thermoformed
structures with a small radius of curvature (such as the outer surface of a micro-cone: r =
30-90 µm), high tensile strains can lead to cracking of electrodes following thermoforming
if located in the curved region (Figure 3-13). It is essential to pre-plan areas of high strains
based on the final desired geometry following the thermoforming process and avoid
placement of thin film metal electrodes in these areas.

Figure 3-13. (a) Image of PMP device with a 3D conical structure with electrode sites on
the outer face; dashed line indicates the region with high tensile strains leading to cracked
electrodes. (b) An SEM image of an electrode (white box) illustrating the formation of
cracks along the surface following thermoforming due to high tensile strains.

101
Additionally, these bulk dimensional changes can impact applications in which the
precise electrode or trace pitch is crucial. As a result of the reorganization of the polymer
chains, that is the reorientation of the amorphous regions of the polymer into more ordered
crystalline portions, a degree of shrinkage exists following the thermoforming process. In
the case of Parylene flat flexible cables (FFCs), direct electrical connections to multiple
contact pads have been demonstrated using commercially available ZIF connectors [83].
However, these ZIF connectors require precise spacing between metal contact pads for
proper insertion and alignment into the connector. Following thermoforming of 8 and 16
channel Parylene FFCs, metal contact pad and inter-pad Parylene spacing shrinkage of
~1.5% was observed (Figure 3-14) and is summarized in Table 3-3. Statistical significance
was measured using a Mann-Whitney nonparametric test in a statistical software package
(Prism, GraphPad Software, Inc., La Jolla, CA).

Figure 3-14. Micrograph images of contact pad regions of (a) 16 channel and (b) 8 channel
PMP cables illustrating bulk shrinkage associated with the thermoforming process.

In comparing the shrinkage between contact pad areas (PMP regions) and inter-pad
areas (PP regions), it was found that both regions were comparable in shrinkage (no
significant difference) for the conditions tested, but larger scale samples may exhibit a
102
different trend as relaxation of built-in intrinsic stress during the deposition of the metal
[84, 85] can cause additional bulk material changes alongside shrinkage of the polymer.
This shrinkage error propagates in contact pad layouts with larger channel numbers (e.g.
32 channels [79]), leading to the prevention of proper pin-pad alignment. In the case of 8
and 16 channels, the dimensional changes were tolerable. However, dimensional changes
must be accounted for during the design process for large channel count cables.

Table 3-3. Summary of shrinkage percentages for contact pad and inter-pad regions as
well as total cable shrinkages for 8 and 16 channel Parylene FFCs.
a
These n values
correspond to 5 measurements taken between two individual samples.
Parylene FFC Region Shrinkage Percentage
(%; Mean ± SD)
Statistical Significance?
8 Channel FFC Contact pad 1.63 ± 0.59 (n = 11) Yes, p = 0.0008
Inter-pad 2.05 ± 0.81 (n = 11) Yes, p < 0.0001
Total pitch 1.21 ± 0.30 (n = 10
a
) Yes, p = 0.0001
16 Channel FFC Contact pad 1.30 ± 0.35 (n = 31) Yes, p < 0.0001
Inter-pad 1.40 ± 0.71 (n = 32) Yes, p < 0.0001
Total pitch 1.33 ± 0.23 (n = 10
a
) Yes, p < 0.0001

Lastly, the need for a vacuum environment for thermoforming was confirmed as
treatment of devices and free films at 200°C for 48 hrs in an oxygen-rich environment
caused by a leak within the vacuum oven resulted in discoloration (yellowing) attributed
to thermal oxidative degradation (Figure 3-15); samples also exhibited increased brittleness
similar to that reported in previous studies [86]. The resulting samples were in fact too
brittle to endure any standard handling. The oxidative degradation of the Parylene samples
following thermoforming in an oxygen-rich environment was also confirmed in chemical
characterization of the bulk polymer by the presence of carbonyl groups on the surface of
Parylene detailed in the following section.

103

Figure 3-15. Comparison between a normal PMP cable and those that underwent thermal
oxidative degradation due to a leak in the vacuum oven. Note the discoloration of the films
as well as brittle failure in the PP sample during handling.

3.4.2 Chemical characterization
Any changes in the chemical properties of thermoformed Parylene may impact
potential functionality of the material in biological or chemical applications. Properties of
untreated Parylene have been thoroughly characterized as an encapsulation material for
medical implants [87, 88] or as coating materials for in vitro experiments [89, 90]. For this
reason, any possible changes to the bulk and surface chemical properties of the
thermoformed films were assessed to observe modifications of the material following the
annealing process.
Changes in chemical functionality within the bulk Parylene following
thermoforming were assessed using Fourier transform infrared spectroscopy (FTIR;
Nicolet iS 10 FT-IR Spectrometer, Thermo Scientific, West Palm Beach, FL), an effective
method to identify the presence of certain functional groups within a material based on
unique resonance absorption peaks of wavelengths in the infrared. Changes to the surface
properties of Parylene films were assessed by measuring contact angles (10 µL DI H 2O
droplet, Model 290-F1, Ramé-Hart Instrument Co., Succasunna, NJ), a method to analyze
surface energy using the wetting angle formed by a liquid droplet on a solid surface.
104
FTIR analysis on the Parylene samples following thermoforming within vacuum
indicated no new chemical functionalities within the bulk polymer (Figure 3-16). Varying
soak temperature and time had no significant effect on FTIR measurements, as the spectra
for Parylene was consistent over all thermoformed samples. This analysis was critical in
the detection of carbonyl groups that have been observed in Parylene spectra following
heat treatment due to thermal oxidation of the polymer [86], which can alter the surface
and bulk chemistries. In our samples, the carbonyl addition was prevented by
thermoforming Parylene samples in a vacuum environment (Figure 3-17).

Figure 3-16. Absorbance measurements from FTIR analysis of thermoformed samples.
Varying the (a) soak temperature and (b) soak time had no effect on the chemical
composition following the thermoforming process.

105

Figure 3-17. Absorbance results from FTIR measurements of untreated Parylene, Parylene
thermoformed under vacuum, and Parylene thermoformed in an oxygen rich environment
(leak in vacuum oven). Thermoforming within an oxygen rich environment results in an
additional carbonyl peak that is not present in samples thermoformed under vacuum.

An analysis of the sample surfaces conducted using contact angle measurements
indicated no significant changes to the surface energy of Parylene as a function of either
soak temperature or time. Samples thermoformed under varying soak temperatures and
times exhibited contact angles that varied only by a few degrees which demonstrated that
thermoforming in a low oxygen environment did not significantly alter the surface energy
of Parylene (Table 3-4, Table 3-5).

106
Table 3-4. Contact angle measurements obtained from Parylene films following the
thermoforming process with varying Ts (ts = 6 hrs). Results indicated that there was no
change in the surface energy with varying soak temperature.
Ts (°C) Contact Angle
(°; Mean ± SD, n = 6)
0 86.7 ± 0.3
80 87.9 ± 0.8
100 88.8 ± 1.5
120 85.2 ± 1.4
160 89.8 ± 1.3
180 88.3 ± 1.0
200 87.9 ± 1.3

Table 3-5. Contact angle measurements obtained from Parylene films following the
thermoforming process with varying ts (Ts = 200°C). Results indicated that there was no
change in the surface energy with varying soak time.
ts (hrs) Contact Angle
(°; Mean ± SD, n = 5)
0 86.7 ± 0.3
0.5 87.1 ± 2.0
2 88.7 ± 2.1
6 87.9 ± 1.3
12 89.7 ± 1.5
24 89.7 ± 1.2
48 88.2 ± 0.3

3.4.3 Mechanical characterization
Annealing of Parylene films has been reported to increase the crystallinity of the
polymer as the amorphous regions of the polymer gain more ordered structure following
thermal treatment [73]; increases in crystallinity were demonstrated to result in increased
tensile strength and overall stiffness [86]. In this study, changes to the mechanical
functionality of the Parylene films following thermoforming were analyzed by measuring
changes to the Young’s modulus of thermoformed films using nanoindentation (Berkovich
tip, 4 um displacement, MTS Nano Indenter XP, Agilent Technologies Inc., Santa Clara,
CA). The changes to the Young’s moduli of thermoformed films of varying conditions
would alter the stiffness and thus mechanical functionality of 3D Parylene structures of
107
differing geometries. Such changes also have implications on the possibility of tuning the
mechanical properties of Parylene via annealing processes. The significance of the changes
to Young’s modulus was determined using a one-way ANOVA test along with a test for
linear trend using a statistical software package (Prism, GraphPad Software, Inc., La Jolla,
CA).
Nanoindentation measurements indicated that the Young’s moduli of
thermoformed Parylene samples increased following the annealing process, agreeing with
changes in crystallinity reported within literature (Table 3-6, Table 3-7). This increase in
crystallinity stems from the development of new crystalline domains within the Parylene’s
amorphous regions [73], as well as the formation of a new second crystalline phase [75].
Variation of soak temperature had a larger effect on mechanical stiffness changes
(increased 37%) largely due to the additional energy in the system at greater temperatures
leading to increased reorganization and crystalline ordering of the polymer (one-way
ANOVA: p < 0.0001; test for linear trend: p < 0.0001) (Table 3-6). The smaller change in
stiffness as a function of soak time compared to soak temperature (no significant
difference) can be explained by observations in literature in which Parylene crystallinity
does not vary greatly following the first couple of minutes of annealing; the kinetics of
crystallization (i.e. crystallinity and crystallite size changes) have been shown to be
relatively quick [73] and reach a plateau in a matter of minutes [91] as suggested by
crystallinity measurements using X-ray diffraction of annealed Parylene films.

Table 3-6. Young’s modulus measurements obtained via nanoindentation of
thermoformed films for varying Ts (ts = 6 hrs). n values correspond to number of points
on one large area sample.
TS (°C) E (GPa; Mean ± SE)
Untreated 2.69 ± 0.03 (n = 29)
80 2.90 ± 0.03 (n = 31)
100 3.10 ± 0.03 (n = 33)
120 3.17 ± 0.03 (n = 37)
160 3.57 ± 0.03 (n = 25)
180 3.77 ± 0.02 (n = 35)
200 3.71 ± 0.03 (n = 26)
108

Table 3-7. Young’s modulus measurements obtained via nanoindentation of
thermoformed films for varying ts (Ts = 200°C). n values correspond to number of points
on one large area sample.
tS (hrs) E (GPa; Mean ± SE)
0.5 3.69 ± 0.03 (n = 31)
2 3.76 ± 0.05 (n = 12)
6 3.71 ± 0.03 (n = 26)
12 3.78 ± 0.03 (n = 28)
24 3.78 ± 0.03 (n = 23)
48 3.71 ± 0.03 (n = 26)

3.4.4 Electrochemical characterization
The recording quality of intracortical electrodes have been largely tied to the
electrochemical properties of the electrode surfaces once implanted [15]. Thermoforming
effects on electrode (and thus recording) properties of the PSE were assessed using cyclic
voltammetry (CV; 1x PBS, -0.6 to 0.8V, scan rate of 50 mV/s for 3 cycles, Ag/AgCl
reference) and electrochemical impedance spectroscopy (EIS; 1x PBS, 1-100,000 Hz,
Ag/AgCl reference), which are widely used to assess electrode properties and investigate
changes to the electrode surfaces [92]. CV and EIS measurements were taken of electrodes
before and after the thermoforming process to observe possible annealing effects on the
electrode material, and give insight into the recording capability of the PSE.

109

Figure 3-18. Representative CV curves of untreated (blue dashed) and thermoformed
(black solid) electrodes.

CV measurements of the platinum electrode surface indicated a noticeable
reduction in the area inside the curve compared to the untreated electrodes, indicating a
decreased charge storage capacity (Figure 3-18). Electrochemical impedance also
supported an altered electrode surface as the impedance magnitude had increased and the
phase plot had shifted to the right; the 1 kHz impedance magnitude increased by 375% to
~125 kΩ following thermoforming (Figure 3-19). Taken together with the CV results, we
hypothesize that one or more phenomena may have contributed to the observed changes:
(1) enhanced sealing of poorly adhered Parylene-Parylene or Parylene-metal regions of the
device due to annealing, (2) contamination of the surface of the platinum electrodes by
chlorine molecules mobilized from Parylene C during annealing that diffuse through
microvoids within platinum grain boundaries formed following platinum thin-film
annealing [93, 94], and/or (3) a combination of the relaxation effects of the intrinsic
residual stress of the deposited thin-film platinum [85] and possible formation of
microcracks during thin-film platinum annealing [95]. Additional information on the
possible causes of the altered electrochemical properties of thin film electrodes supported
on thermoformed Parylene substrates can be found here [96].
110

Figure 3-19. Representative plots impedance magnitude (a) and phase (b) of EIS
measurements of untreated (red triangles) and thermoformed (black squares) electrode.

However, despite the chemical or mechanical disruptions to the electrode surface
caused by thermoforming, the electrodes were still able to detect resolvable neural signals
in vivo [52] (Figure 3-20).

Figure 3-20. Representative 500 msec trace from a 300-Hz high-pass filtered
electrophysiological record from thermoformed PSE at 21 days post-implantation into the
rat motor cortex.

111
3.5 Development of packaging of Parylene sheath electrode
One of the largest non-biological failure modes of chronic probes is packaging
related issues of breakage or corrosion of the electrical connections from the electrodes to
the neurophysiological measurement connectors [4, 17, 97, 98]. Issues due to poor
mechanical strength of implanted wires, improper hermetic passivation to protect the
implant from the wet environment of the body, and connector breakage have spurred the
efforts focused on packaging and connector development for metal or silicon-based neural
probes in research [99-101] and commercial [102-104] settings.
However, the development of packaging methods for Parylene-based devices is not
as trivial as their more rigid counterparts. Because of the soft Parylene substrate, traditional
means of establishing electrical connections to contact pads that require heat and
mechanical abrasion (e.g. soldering, wire bonding) cannot be used. Previous efforts to wire
bond and solder direct leads to Parylene-Pt cables were found to be very difficult and time
consuming. We previously presented a ZIF connection scheme that allowed for rapid,
robust, and repeatable connections to multi-channel Parylene cables [83]. Because of the
commercial availability, low cost of this connector for prototyping and batch production,
and simplicity of use for Parylene cables, design iterations of the packaging of the Parylene
sheath electrode centered on the use of the ZIF connector to make electrical connections
(Figure 3-21).

112

Figure 3-21. (a) As-fabricated Parylene sheath probe highlighting the contact pads for
integration with ZIF connectors. (b) Photograph of fully packaged dual-probe array with
two probes secured into 2 ZIF connectors on the flexPCB.

3.5.1 Zero-insertion-force connector
The zero-insertion-force connector used in our connection scheme (FH19SC-8S-
0.5SH(05); HIROSE Electric U.S.A., Inc., Simi Valley, CA) is an actuator-based connector
that allows a flat cable to be inserted into a bay within the connector and a hinge actuator
closes on top of the cable to seal in the cable and make electrical contact. Depending on
the type of ZIF connector (top or bottom contact), the metal contacts are positioned on the
top or bottom region of the cable bay. These ZIF connectors terminate at the other end with
solder legs that allow for soldering onto a PCB.

3.5.2 Omnetics connector
To allow for proper integration with the Plexon neurophysiological measurement
system used by our collaborators at Huntington Medical Research Institutes (HMRI), the
Parylene sheath electrode arrays had to terminate in a female Omnetics micro-strip
connector (NPD, type VV; Omnetics Connector Corporation, Minneapolis, MN). These
micro connectors are the industry standard for use in acute and chronic intracortical
stimulating and recording electrodes largely due to their small size and robustness in build
to endure animal-related stresses during the lifetime of the implant.
113

3.5.3 Connection paradigm of PSE
The connection strategy of the PSE was a two-part approach that had (1) a ZIF-
based head stage that would connect the PSE to an intermediate connector and remain
implanted and (2) an external connector board that would interface between the
intermediate connector and the Omnetics to fit into the neurophysiological recording
system. This two-part method was intended to reduce the costs and design iteration
turnovers during the prototyping stage, largely due to the high-costs and long lead times of
the Omnetics connectors as well as to allow for modular configurations of arrays of the
PSEs.
For the headstage, we planned the use of a flexible PCB (flexPCB, polyimide) that
would contain the appropriate ZIF connectors for connection to multiple PSEs (to
configure arrays) electrically integrated with an intermediate connector. The ZIF connector
requires a minimum cable thickness of 300 µm to ensure a tight and secure fit inside of the
connector. As the Parylene flexible cable fabricated with the PSE is a total thickness of
~12 µm, a poly(etheretherketone) (PEEK; 8504K13, McMaster-Carr, Aurora, OH)
stiffener backing was affixed to the contact pad region of the Parylene flex cable to allow
for proper insertion into the ZIF connector (Figure 3-22). Although hermeticity is not
required because the ZIF connector portion is not implanted on or under the skull but above,
marine epoxy (Loctite®, Henkel Corp., Westlake, OH) was used to provide passivation of
the ZIF connectors. The intermediate connector for the initial designs began as a similar
large channel count ZIF connector, but was later changed into a board-to-board, header-
socket-based Hirose BM10 connector (HIROSE Electric U.S.A., Inc., Simi Valley, CA) in
the final design iteration of the PSEA for better reliability. This intermediate connector
interfaced with a mating connector on an external adaptor board, which had contact pads
to outfit the 40 pin female Omnetics connector. Initially the external adaptor board was a
rigid PCB, but was later changed to a flexible substrate to facilitate making connections to
the PSE when the rat’s head was in a difficult position.
114

Figure 3-22. (a) Optical micrograph of Parylene flexible cable indicating the contact pad
terminated ends for insertion into a commercial ZIF connector. (b) Schematic of operation
of ZIF connector.

3.5.4 FlexPCB design iterations
3.5.4.1 Wing-California PCB
For the very first connector design, a flexPCB was designed that allowed for 4 PSEs
to attach to two separate wings and a ZIF-based intermediate connector (also known as the
Wing PCB, Figure 3-23):

Figure 3-23. (a) Image of bare Wing PCB prior to cutting and bending to shape. (b) Formed
Wing PCB following heat treatment to form bent orientation designed to allow for
connection to 4 PSEs. (c) Wing PCB inserted into the rigid ZIF-Omnetics board that
allowed for integration with Plexon system.

115
The Wing PCB had four 8 channel ZIF connectors (2 on each side of each wing) to
allow for electrical connections for 4 PSEs. These 8-channel ZIF connectors had traces that
lead to a contact pad arrangement designed for connection with a 33 channel ZIF connector
(32 channels for the 4 probes of 8 electrodes as well as one more channel for an onboard
REF/GND) on the external rigid adaptor board. This rigid PCB connected a 33 channel
0.3 mm pitch ZIF connector with the 36 channel (40 pin) Omnetics connector for
connections to the neurophysiological measurement system.
However, to bring the 4 probes closer to the desired array dimension of a 1 x 1 mm
square (to target the barrel cortex of the rat) the large wingspan of the Wing PCB had to be
reduced. This was achieved by using a large scale thermoforming process utilizing a hot-
air gun. The heat treatment of the Wing PCB while it was in a bent orientation as seen in
Figure 3-23b locked in the final shape. Despite this, for initial implantation experiments,
the Wing PCB was too large and bulky to be handled with the first generation of inserter
tools. Thus the PCBs were halved (California PCB; Figure 3-24), and initially a single ZIF
connector was soldered onto the California PCB to allow for implantation of a single probe
for testing.

Figure 3-24. Optical micrograph of the California PCB and inserter tool used for initial in
vivo insertion experiments.

116
These California PCBs however were still too bulky and long for implantation; the
collaborators desired a shorter and smaller PCB that would be easier to attach to the top of
the head to form a tower with bone cement, and also would not be such a burden on the rat
following surgery.

3.5.4.2 hFlex PCB
In the next design iteration, a more compact and shorter version of the Wing PCB
was designed that had an extended arm piece (thus named the hFlex) that was to be folded
back and thermoformed to allow for 4 probes to be implanted at once (Figure 3-25). This
design also was modular by simply cutting the arm piece to realize a dual-probe array. The
hFlex was similar to the Wing PCB in that it allowed for 4 PSEs to be connected to the
flexPCB body (4 ZIF connectors) and also terminated in contact pads that would plug into
the ZIF connector of the same external rigid adaptor board. Much like the Wing PCB, the
ZIF connectors for the hFlex were positioned on the front and backsides of each at a set
pitch of 1 mm to match the rostro-caudal extent of the M1 motor cortex in the rat.

Figure 3-25. Image of bare hFlex initially designed to allow for 4 PSEs; the arm piece
could fold back to allow for all 4 PSEs to be implanted at once.

117
For the initial acute implantation study (28 day study), folding the arm and
thermoforming the final shape proved difficult and did not produce reliable results. Thus
the arm of the hFlex was cut to leave only 2 probes for implantation, simplifying the
insertion method and creating a basis for future versions (Figure 3-26). A wire was also
soldered to the hFlex to connect the GND/REF to a stainless screw via wire wrap that was
used to fasten a headplate to the skull. This dual probe array version of the FlexPCB was
very successful in providing reliable results for the 28 day study.

Figure 3-26. Image of halved hFlex and wire soldered to the body for use as the GND/REF
used for 28 day study.

3.5.4.3 FlexPCB v1: dual probe arrays
Stemming from the success of the halved hFlex dual probe array for the 28 day
study, slight modifications were made per the request of the surgeons. Because the portion
of the hFlex that protruded from the rat’s skull was a too long, the entire length was
shortened to decrease the length that was susceptible to breakage and also aided in
developing less of a disturbance to the animal (Figure 3-27). The h arm was also removed
from the design to create a dedicated dual PSE array to finish off the 28 day study.
118

Figure 3-27. Image comparing the older hFlex (left) and the newer, shorter hFlex (right).

Also, the rigid PCB that was once used to connect the hFlex to the recording system
was found to not be flexible or robust enough to give quality signals while the animal’s
head was in a skewed position. Therefore a longer, more flexible PCB was design to allow
for the bending of the connector to reach the h flex regardless of the animal’s head position
(Figure 3-28). This aided in ease for the surgeon and animal caretakers for obtaining
impedance and neural signal data.
119

Figure 3-28. Image of new flexible ZIF-Omnetics PCB for more robust connections to
animals during neurophysiological measurements.

Another change that was made to the system during the 28 day study was the
removal of the REF/GND wire that was attached to the hFlex. Loss of recording fidelity
mid study was found due to corrosion of this wire at the interface with the implanted
stainless steel screw. Thus, moving forward, a wire was attached to the external flexible
adaptor board and was clipped to the headplate to be used as the REF/GND during
electrical testing and measurements.

3.5.4.4 FlexPCB v2: PSEA
The PSEA and its packaging was designed concurrently following the lessons
learned of the previous packaging attempts. With a similar concept of the FlexPCB v1, a
single flexPCB substrate is used to connect two 1x2 PSE arrays via a 16 channel ZIF
connector to an intermediate connector (Figure 3-29). In this design however, the
intermediate connector was switched to a header/receptacle, socket-based BM10 connector
(HIROSE Electric U.S.A., Inc., Simi Valley, CA). This design change stemmed from the
unreliable performance of the actuator hinge of the ZIF connector to make repeatable
mating connections for a month. By utilizing a socket-based connector, the durable
120
connection scheme made it possible for longer mating cycles for the planned 6 month study.
Also the change from the ZIF intermediate connector not only allowed for a thicker
FlexPCB (a major stress point and failure mode during the in vivo studies), but also allowed
for performance robustness as the rat’s head does not have to be in a specific orientation
(and maintained in that orientation) to take a measurement. The header portion is attached
to an external BM10-Omnetics adaptor, and snaps into place onto the receptacle on the
array for easy attachment (Figure 3-30).

Figure 3-29. Image of the FlexPCB used to form the PSEA from two 1x2 arrays. A 16
channel ZIF connector was used to connect each 1x2 array (one on each side) of the
FlexPCB, which has leads to a BM10 receptacle connector (in black).

Note that in the figure above, the array FlexPCB has a small tab that remains attached to
the introducer tool during the implantation process. This allows for a stabilization and
overall PCB support during the insertion process.

121

Figure 3-30. Image of the PSEA FlexPCB and the external BM10-Omnetics flexible
adaptor board used for the chronic in vivo studies.

The FlexPCB v2 and external BM10-Omnetics flexible adaptor board was
successful for the 3-6 month study that was used to verify the sheath-based design during
chronic implant conditions. However, a failure mode was encountered in the polyimide
material used in the FlexPCB to not be robust enough to withstand the connect/reconnect
cycles of the BM10 connectors; some arrays faced a failure mode of shearing off after the
connect/reconnect cycle. As mentioned previously, because we use the BM10 connector,
we can make the flexPCB thicker to prevent this, or move to a more rigid material to
produce the headstage portion of the connector. There were also issues with dirt from the
cage environment lodging into the socket of the connector, requiring cleaning before each
recording step. This can be mediated in future studies by using a cap to enclose the
connector for protection.

122
3.6 Implantation methods of Parylene sheath electrode
3.6.1 Insertion techniques for flexible probes
In assessing implantation techniques for the traditional silicon UEA and MA, the
rigid substrates of the probes allow for relatively simple insertions into the cortical tissue
by means of a manual (i.e. with tweezers) or tool-based insertion straight into the tissue.
Surgical implantation of flexible cortical probes, however, is more challenging; because of
the compliance of the sheath probe, additional assistance is required to possess temporary
stiffness during tissue penetration for a straight path into the desired depth into tissue to
minimize tissue damage caused by buckling and movement during insertion [37]. For
example, to insert the probe through the pia mater into rat cortex, probes must withstand a
buckling load of 1 mN; this value is larger if the dura mater is not retracted [105].
Two insertion strategies have often been utilized for flat, thin film probes: (1) a
coating to temporarily stiffen the probe to allow for insertion and subsequent dissolution
to recover flexibility (e.g. poly(ethylene glycol) (PEG) [34], carboxy-methylcellulose
(CMC) [38], silk fibrion [25]) and/or (2) a supportive backing on which the probe is
temporarily attached as a guide, which is removed following implantation [37]. Beyond
the use of additional materials or support tools, there is also current work in the
development of novel probe materials that have tunable rigidity properties as another
insertion strategy; a probe can be stiff when implanted and then change its properties to
become flexible after it has reached its insertion depth [41, 43].

3.6.2 Development of insertion techniques for Parylene sheath electrode
For the development of the insertion strategy for the PSE and the PSEA, the 3D
hollow sheath structures posed additional constraints to the insertion techniques for flexible
probes mentioned prior. Because of the non-planar structure of the sheath, the PSE would
be difficult to attach to a flat supportive backing such as SU-8 and silicon used in literature.
123
Also understanding the effects of high temperature on Parylene devices, coating materials
and processes had to be constrained to temperatures underneath the Tg of Parylene.
With these two points in mind, the inserter tool strategy developed for the PSE
consists of two parts: (1) microwires integrated with a support shuttle that allows for
handling and insertion and (2) a PEG stiffener coating that temporarily attaches the probe
to the support backing as well as to the guide microwires. Tungsten microwires (250 µm
diameter) were chosen as the attachment shuttle for the probe as they were successfully
demonstrated as neural probe shanks and their rounded shape enabled a closer match to the
Parylene sheath probe during attachment to the guide. To adhere the probes to the inserter
tool, PEG, a commonly used and well tested polymer in neural implants [106] that
dissolves in saline and has a low melting point (~60°C for PEG 3000), was chosen. A
droplet of molten PEG was applied to the microwire-probe interface using a clean solder
iron tip (80°C), which allowed for controlled application to form an overcoat and attach
the PSE to the microwire. PEG was similarly applied between the integrated Parylene cable
of the probe and the support shuttle for handling robustness. For adhering probes to the
guide microwires, initial designs were side-mount oriented, where the probe was attached
to the side of a sharp microwire, and later designs took advantage of a tip mount approach,
where the probe was attached at the tip end of a blunt microwire to reduce the insertion
probe track and take advantage of the natural taper of the cone structure for an improved
insertion into the tissue (Figure 3-31).

Figure 3-31. (a) Optical micrograph of side-mounted probe to microwire. (b) Optical
micrograph of tip-mounted probe to microwire.

124
For the implantation procedure, the probe-tool assembly was tension clamped to an
arm of a stereotaxic apparatus. A manual screw drive (and later a motorized drive with
precise displacement and insertion velocity control) controlled the displacement of the
probe-tool assembly into the cortex following craniotomy and retraction of the dura mater.
The insertion depth was tracked by the stereotaxic apparatus as well as depth markers that
were designed into the Parylene flexible cable of the PSE. Following insertion, 1× PBS
was applied to the PEG interfaces to dissolve the polymer and allow for detachment.
Inserter tools were retracted after complete dissolution of the PEG (5 minutes), leaving the
PSE within the cortical tissue.

3.6.2.1 Silicon insertion shuttle
The first iteration of the inserter tool consisted of 4 tungsten microwires in a 2 x 2
orientation at pitch of 1 x 1 mm soldered around the circumference of a stainless steel wire;
precise placement of the tungsten microwires around the stainless steel wire was achieved
using a custom soldering jig (Figure 3-32). Although simple to fabricate, the 4 pronged
inserter tool was much too difficult to implement with the prototype probe designs as the
cables were far too thick and inflexible to allow for PEG attachment for all 4 tools at the
same time. A support structure was found to be necessary that would allow the FlexPCB
to rest on the tool during the transport to surgery as well as during the insertion process.
This created multiple strain points that damaged the probes and caused some to undergo a
tear failure during transport and surgery. These results indicated a need for a support shuttle
that could hold and support the FlexPCB, the Parylene flexible cable of the PSE, and the
probe itself.

125

Figure 3-32. Image of first inserter tool design with 4 tapered microwires soldered to a
larger handling cylinder.

For the first support shuttle design, bulk silicon was explored as the shuttle backing
because of its rigidity and simplicity in scribing to precise dimensions. Si rectangles (6 mm
x 5 cm) were scribed from a 384 µm thick Si wafer and a single microwire was glued to
the backing using a biocompatible adhesive. A single PSE with integrated Parylene cable
and California PCB was affixed to the silicon insertion shuttle using PEG 1000 at the cable-
shuttle interface and PEG 8000 at the probe-shuttle interface (Figure 3-33). The support
backing eased the transport and handling of the PSEs to the surgery site and during the
implantation process compared to the first iteration of the inserter tool.

Figure 3-33. Silicon-based inserter tool with a PSE-California PCB affixed with PEG for
insertion.
126

However, initial benchtop insertion experiments demonstrated that the silicon
backing for the inserter tool was too brittle. The silicon backing was not robust enough to
endure surgeon handling and clamping to the stereotaxic apparatus for insertion, and often
fractured above the implant site, which dropped harmful silicon shards onto the tissue.
Another alternative was necessary that was made of a more robust material.

3.6.2.2 Acrylic insertion shuttle
The search for a more robust material ended with the use of acrylic, an inexpensive
plastic capable of being laser machined to accurate dimensions and thus batch fabricated,
as the insertion shuttle for the halved hFlex PCBs and for the following designs (Figure
3-34). To form the acrylic shuttle, an acrylic piece (1.5 mm thick) was laser cut to a precise
shape to facilitate handling and ease of clamping to the stereotaxic arm and also had
rastered grooves to allow precise placement of microwires, which were glued in place using
a biocompatible adhesive. The flex PCBs were similarly attached to the acrylic insertion
shuttles using PEG 3000 at both the cable-shuttle interface as well as the probe-shuttle
interface. The PEG used was changed to PEG 3000 from the originally used PEG 1000 at
the cable-shuttle interface and PEG 8000 at the probe-shuttle interface to provide a more
robust attachment at the cable-shuttle interface as well as be less difficult to work with at
the probe-shuttle interface because of the high melting point and short work time of the
PEG 8000.

127

Figure 3-34. (a) A custom microwire-based introducer tool used for implantation for dual-
probe array. (b) The dual-probe array was temporarily affixed to the introducer tool using
PEG, which dissolved following insertion. PEG was used in 2 locations: cable region for
strain relief and support and probe-microwire tip for stiffness during insertion.

The acrylic support tool was robust enough to endure rough handling by surgeons
as well as sustain multiple clamp cycles in the stereotaxic apparatus without fracture. The
acrylic-based inserter tool was used in successful insertions of the Parylene sheath probe
into agarose brain models as well as cortical tissue during in vivo experiments, which
validated the robustness of the acrylic shuttle. An acrylic shuttle tool with 4 microwires (2
x 2 orientation) was also used for the insertion of the PSEA.
A quick note is mentioned here in the change from the side-mount to tip-mount
attachment orientation at the probe-shuttle interface. Although both adhering methods were
successfully employed in surgical insertions, the tip mount was chosen because of its low-
profile shape, which reduced the insertion track dimensions and minimized tissue damage.

3.6.3 Benchtop testing
Benchtop insertion tests of the PSE and PSEA were performed using 0.5% agarose
gel (A9539-50G; Sigma-Aldrich, St. Louis, MO), a commonly use cortical tissue gel
phantom that has a similar stiffness properties of the tissue [107, 108]. For benchtop
insertions tests, two setups were used to mimic the two procedures used during surgery:
(1) manual insertion and the later developed (2) motorized insertion. For manual insertion
128
tests, a stereotaxic apparatus, similar to what is used during in vivo insertions, was
employed. The probe-tool assembly was clamped onto an arm of the stereotaxic apparatus
and was manually inserted into the agarose by turning a screw drive. Insertion depth was
verified by using depth markers on the integrated Parylene cable of the PSE as well as on
markings on the stereotaxic apparatus (Figure 3-35).

Figure 3-35. Image sequence of a representative benchtop agarose insertion test. (a) The
array-tool assembly was attached to stereotaxic apparatus and positioned over the spot of
interest. (b) The array-tool assembly was inserted into the agarose by hand, and 1× PBS
applied to begin the dissolution of PEG. (c) The dual-probe array was detached from the
tool to allow for tool withdrawal; note that the Parylene cable was still attached to the tool.
(d) After complete dissolution of PEG (5 minutes), the inserter tool was withdrawn from
the agarose, and the dual-probe array remained implanted.

These manual insertions however were found to not be consistent in terms of
insertion speed and depth and also introduced vibrations from manually turning the screw
drive that could cause additional damage during the implantation process [109]. To
automate the process, a motorized insertion strategy using a velocity and displacement
controlled motorized stage via LabVIEW and a very thin fiduciary marker wire to
129
determine the starting position was used to produce more accurate and repeatable results
(Figure 3-36). The motorized insertion procedure was carried out as follows: initially the
probe-tool assembly was manually lowered using a screw drive until the tip of the fiduciary
marker was touching the agarose surface. Because the fiduciary marker is set at a specific
distance from the bottom of the PSE (1 mm), the motorized stage was programmed to
displace into the agarose a depth of 2.5 mm (1 mm + 1.5 mm required depth) at a set
insertion speed (0.8 mm/s). Following PEG dissolution, the inserter tool was also removed
at a fast velocity, leaving the PSE implanted within the gel.

Figure 3-36. Optical micrograph of PSEA attached to an acrylic inserter tool modified to
include a fiduciary marker wire for automated motorized insertions.

For both insertion methods however, following removal of the inserter tool from
the gel, closer inspection of the cross sectional area of the gel block during the withdrawal
process indicated that the PSE withdrew upwards 86.34 ± 36.02 m (Mean ± SD, n = 6
trials) (Figure 3-37). This is likely attributed to surface tension at the PEG-inserter tool
interface, forming adhesive forces between the PSE and tool. To compensate for this, the
surgery insertion depth for subsequent in vivo experiments was adjusted appropriately.
Despite this, the motorized insertion process was found to facilitate the process and add a
130
degree of confidence to the repeatability and reduced insertion trauma for the implantation
procedure.

Figure 3-37. Image sequence of benchtop insertion test indicating that the probe withdrew
~90 m following the tool removal. (a) A photograph taken after insertion but prior to tool
withdrawal where the probe tip depth is indicated by the black dotted line. (b) A photograph
following tool withdrawal where the new probe tip depth is indicated by the white dotted
line.

A quick note is mentioned here regarding the motorized insertion speeds chosen for
the implantation process, both on benchtop and in vivo. As mentioned in the previous
section on the Parylene-based EC-MEMS sensor, recommended insertion speeds are
debated within the field, and vary based on the geometry, design, and material of the probe.
However for probes with similar size and cross-sectional areas of the PSE, a speed of 2
mm/s demonstrated success and reduced tissue strains [110]. Faster speeds have been found
to reduce the damage for relatively blunt probes [111] (the UEA “bed of needles” design
is inserted at a pneumatic 8 m/s [112]). For this reason, we believed that a faster insertion
speed was better for the PSE and PSEA. Unfortunately, the maximum speed of the
motorized drive used for in vivo studies was limited to 0.8 mm/s, so all motorized tests
were conducted at this velocity.

131
3.6.4 In vivo testing
In vivo implantations for electrophysiological measurements with the PSE were
carried out in two studies: (1) a shorter, “acute,” 28 day study with the FlexPCB v1 dual
PSE arrays to evaluate and down-select three probe designs and three coating compositions,
and (2) a longer, “chronic,” 3-6 month study with the FlexPCB v2 PSEA to validate the
efficacy and long term feasibility of the PSE design.
For in vivo tests, the animal preparations are as follows [all procedures for the
animal experiments were in accordance with the animal protocol approved by the
Huntington Medical Research Institutes Institutional Animal Care and Use Committee
(HMRI IACUC) and in compliance with the Animal Welfare Act.]:
The array-tool assembly was first ethylene-oxide (EtO) sterilized for 24 hours using
a room-temperature sterilizing system (Anprolene AN74i, Andersen Products, Haw River,
NC) and stored in sterile packaging at 4°C until the time of implantation (~1-2 days)
following manufacturer specifications to reduce the degradation rate of the coatings [113].
The chronic implantation surgery was performed in the rat M1 motor cortex of young, male
Sprague Dawley rats (> 320g) as follows. Anesthesia was induced by placing the rat in a
chamber filled with 4% isoflurane in 1 lpm oxygen. The animal was then placed in a
stereotaxic frame (Small animal stereotaxis; Kopf Instruments, Tujunga, CA) and 1-3%
isoflurane in 1 lpm oxygen was administered via a nose cone. Following a 2 cm long
midline incision and retraction of the skin, six holes were drilled into the skull to allow for
attachment of a headplate (used as reference/ground) with six stainless steel screws. A
craniotomy was made centered at +1.5 mm AP and 1.5 mm ML from bregma, exposing
the M1 cortex. The dura layer over the target area was incised and retracted using the tip
of a syringe needle. The array-tool assembly was positioned over the opening region via a
stereotaxic apparatus and for the 28 day study was manually inserted into the cortex with
a micromanipulator until reaching the required depth of 2 mm under visual inspection using
binocular loupes (Kepler 6.0x; Care-Optics Industrial Co., LTD., Shenzhen, China).
Following insertion, Kwik-Sil (World Precision Instruments, Sarasota, FL) was applied by
132
syringe to the dual-probe array to anchor it in the cortex. After allowing five minutes for
Kwik-Sil curing, saline was applied to allow for PEG dissolution and separation of the
dual-probe array from the introducer tool. The tool was then withdrawn. Additional Kwik-
Sil was applied to seal the craniotomy and for cable support. Bone cement was applied to
anchor the array completely within a head cap. The skin was then closed around the
implanted assembly and sutured shut (4-0 nylon monofilament sutures; Keebomed Inc.,
Morton Grove, IL). The animal was kept on a warming pad in a recovery cage until it was
fully awake and sternal, at which point the animal was returned to its housing cage (Figure
3-38).

Figure 3-38. Image of rat implanted with PSEA. The bone cement head cap and top portion
of the FlexPCB can be seen atop the animal’s head.

Motorized implant procedures differed from the previous protocol in that following
the craniotomy and dural retraction, the array-tool assembly was manually lowered using
the stereotaxic apparatus until the fiduciary marker touched the exposed cortical surface.
The set displacement was then inputted into the motor control and the array-tool assembly
was displaced 1.5 mm into the tissue at 0.8 mm/s. Prior to saline application however, bone
133
cement was used to anchor the array and flexPCB to the skull to allow for tool withdrawal.
Once the bone cement was set, saline was applied to dissolve the PEG and the tool was
withdrawn to leave the PSEA implanted (Figure 3-39). In histology preparations from the
previous 28 day studies, it was found that the applied Kwik-Sil made it difficult to keep
the PSE within the tissue to analyze the histology with the probes intact. For the 3-6 month
studies, the Kwik-Sil was replaced with gel foam, a polymer that initially remains solid but
takes on the surrounding cortical fluid to dissolve, with bone cement to form the head cap.

Figure 3-39. Sequence of images illustrating the in vivo implantation process for the PSEA.
(a) The PSEA-tool assembly is lowered to the surface of the cortex following dura removal.
(b) The PSEA-tool is inserted into the cortex to the required depth. Bone cement is also
added to the rear of the PSEA to hold the array in place following removal of the inserter
tool. (c) The inserter tool is removed following application of saline and gel foam is placed
over the craniotomy. (d) Lastly a head cap is formed using bone cement.

3.6.4.1 28 day acute study results
Detailed results for the 28 day study are out of the scope for the current work, but
are briefly mentioned here for completeness. The acrylic inserter shuttles were robust in
repeatably inserting the dual-PSE arrays into the rat cortex, and the implantation procedure
134
was found to be sufficient to complete the study. 13 out of 19 rats were successful in
obtaining data to down-select the coatings and the three different designs; 6 rats were
excluded because of failures with flexPCB stress breakage from the connect/reconnect
cycles as well as the headplate become loose and coming off due to a small number of
screws used in initial implants. For the dual PSE arrays, the typical rise and stabilization
of the impedances and a stabilization of the probe-tissue interface at 28 days was reported
in analyzing the neural noise, signal-to-noise ratio (SNR), and event rate.
Coating analysis indicated that an immunosuppressant coating, dexamethasone,
suspended in a Matrigel hydrogel provided some improvements on the recording event rate.
Histological and neurophysiological results also indicated that in the comparison of three
different probe tip designs (sharp taper, medium taper, and cylindrical) the sharpest tip had
the best performance, largely attributed to the additional damage to the tissue during/upon
implantation caused by a blunt, larger tip (Figure 3-40, Figure 3-41). These results are
further supported by a microscopically-monitored cortical insertion study, demonstrating
that sharp tips do not produce more micro-vessel severance (compared to blunt tips); in
contrast, blunt tips produce more tissue strain, resulting in dragging and subsequent rupture
of microvascular networks [114]. Full details of the 28 day study can be found here [52].

135

Figure 3-40. Sample microphotographs of the cerebral cortex double-stained with
antibodies for NeuN (brown) and GFAP (dark-blue) through the following probe tips: (a)
sharp, (b) moderate, and (c) blunt. The 10 µm sections were cut perpendicular to the probe
track (the probe was removed during brain dissection).

Figure 3-41. Quantification of (a) neuronal and (b) astrocytic density for different probe
tip shapes (blunt, moderate, and sharp) at 28 days (n = 13 animals), indicating the best
performance out of the sharp design.

3.6.4.2 3-6 month chronic study results
Analysis of the 3-6 month chronic study is still underway, but in analyzing the
packaging and insertion strategy, the acrylic shuttle and motorized insertion method were
found to be repeatable and simplified the surgical procedure. However, in looking at some
of the rat subjects that were terminated early, similar failure modes as the previous 28 day
study are apparent. Namely, the flexPCBs still undergo shear failure despite the change in
136
connector and thickening of the flexPCB due to repeated mating cycles of the connector
(Figure 3-42) and the head caps would detach from the skull due to a lack of adhesion
between the bone cement and the skull. In later surgeries, the latter was addressed by using
a surgical glue that greatly improved the adhesion and eliminated the failure.

Figure 3-42. Image of shear-induced failure of the FlexPCB due to repeated mating cycles.
This can be improved by using a thicker FlexPCB material for the headstage.

3.6.4.3 Surgery lessons learned
As a collective lessons learned over the development of the PSE and PSEA and in
dealing with Parylene devices for implantable applications, the most important lessons
learned is that the device needs to be robust enough to endure surgical handling. With
Parylene and other softer polymeric materials, the devices are not as robust as stiffer or
more rigid materials such as silicon or glass, and thus need to be adapted or modified in a
way to ensure that the devices are handled properly during the surgical procedure; the
flexibility of the device comes as a double-edged sword as seen in the development of
additional inserter tool technology for the PSE and PSEA.

137
Also, it is important to be mindful of the material in choosing various post-
processes, treatments, and other compatible technologies. It was mentioned previously that
the thermal budget of Parylene is important during the fabrication process, and the thermal
sensitivity of Parylene remains of importance when choosing different processes. For
example, in looking at sterilization methods for the PSE, rather than typical hot-steam
sterilization (i.e. autoclave), an alternate EtO process was used as the high humidity and
temperature environment of the autoclave can cause unwanted side-effects to the Parylene
devices. Although we are able to leverage this property of Parylene to create 3D structures,
failure to consider these effects can be detrimental in the development of Parylene-based
technology.

3.7 Summary
In summary, the PSE shows much promise as a recording electrode technology,
with hopes to claim similar success of the NE. Here we presented a means to fabricate the
3D sheath structure of the PSE using a thermoforming process, and have further
characterized the process and its effects on the chemical and mechanical properties of the
polymer. A packaging method for long-term implementation of these Parylene devices was
also developed, and though faced some issues due to strains while connecting to the
neurophysiological measurements, worked well for our applications but improvements can
be made. A reliable and repeatable insertion protocol was also presented, that along with a
motorized inserter, provided great results in both the 28 day and 6 month studies. An
interesting note is that the Parylene-Parylene adhesion of these devices in vivo has found
to last considerably longer that in soak testing of these same PSE devices. There are still
additional studies that must be conducted in analyzing the benchtop soak testing of these
Parylene-based devices; for these results the author refers to another work currently being
conducted by another member within our group [115].
138
The development of this thermoforming process in constructing 3D structures from
Parylene widens the opportunities for Parylene to be used as a substrate material for a
variety of applications that may require complex structures that are difficult to fabricate
using standard micromachining processes. Along with the studies shown here discussing
the effects of various thermoforming parameters on the final outcome, this work sets the
basis for further leveraging the thermoplastic nature of Parylene for additional device
designs that were once thought as unachievable with these thin film polymers. The
presented progress on packaging methods for Parylene-based devices can also be applied
to a variety of applications outside of intracortical devices, as the ZIF-BM10 headstage
have provided us with much success.
However, another note is to be made on improving the reliability of these
implantable Parylene-based devices by moving away from wired devices and towards
wireless communication methods. There needs to be additional work done to further
advance Parylene devices by integrating planar coils within the design or integrate devices
with complete systems that contain antennae for wireless transmission or receiving [116-
118]. It is upon achieving this level of Parylene systems development that can expand the
possible applications for Parylene-based devices as implantable systems throughout the
body.
3.8 References
[1] M. A. Lebedev and M. A. L. Nicolelis, "Brain–machine interfaces: past, present
and future," Trends in Neurosciences, vol. 29, pp. 536-546, 2006.
[2] M. A. L. Nicolelis, "Brain-machine interfaces to restore motor function and probe
neural circuits," Nature Reviews Neuroscience, vol. 4, pp. 417-422, May 2003.
[3] D. McCreery, et al., "Arrays for chronic functional microstimulation of the
lumbosacral spinal cord," Neural Systems and Rehabilitation Engineering, IEEE
Transactions on, vol. 12, pp. 195-207, 2004.
[4] A. Prasad, et al., "Comprehensive characterization and failure modes of tungsten
microwire arrays in chronic neural implants," Journal of Neural Engineering, vol. 9, p.
056015, 2012.
139
[5] P. K. Campbell, et al., "A silicon-based, three-dimensional neural interface:
manufacturing processes for an intracortical electrode array," Biomedical Engineering,
IEEE Transactions on, vol. 38, pp. 758-768, 1991.
[6] K. Najafi and K. D. Wise, "An implantable multielectrode array with on-chip signal
processing," Solid-State Circuits, IEEE Journal of, vol. 21, pp. 1035-1044, 1986.
[7] E. M. Maynard, et al., "Neuronal interactions improve cortical population coding
of movement direction," Journal of Neuroscience, vol. 19, pp. 8083-93, Sep 15 1999.
[8] S. Suner, et al., "Reliability of signals from a chronically implanted, silicon-based
electrode array in non-human primate primary motor cortex," IEEE Trans Neural Syst
Rehabil Eng, vol. 13, pp. 524-41, Dec 2005.
[9] L. R. Hochberg, et al., "Neuronal ensemble control of prosthetic devices by a
human with tetraplegia," Nature, vol. 442, pp. 164-71, Jul 13 2006.
[10] A. M. Sodagar, et al., "An Implantable 64-Channel Wireless Microsystem for
Single-Unit Neural Recording," Solid-State Circuits, IEEE Journal of, vol. 44, pp. 2591-
2604, 2009.
[11] G. E. Perlin and K. D. Wise, "A compact architecture for three-dimensional neural
microelectrode arrays," in Engineering in Medicine and Biology Society, 2008. EMBS 2008.
30th Annual International Conference of the IEEE, 2008, pp. 5806-5809.
[12] M. A. L. Nicolelis, et al., "Chronic, multisite, multielectrode recordings in macaque
monkeys," Proceedings of the National Academy of Sciences, vol. 100, p. 11041, 2003.
[13] J. P. Donoghue, "Connecting cortex to machines: Recent advances in brain
interfaces," Nature Neuroscience, vol. 5, pp. 1085-1188, 2002.
[14] K. D. Wise, et al., "Microelectrodes, Microelectronics, and Implantable Neural
Microsystems," Proceedings of the IEEE, vol. 96, pp. 1184-1202, 2008.
[15] P. Abhishek and C. S. Justin, "Quantifying long-term microelectrode array
functionality using chronic in vivo impedance testing," Journal of Neural Engineering, vol.
9, p. 026028, 2012.
[16] T. Saxena, et al., "The impact of chronic blood–brain barrier breach on intracortical
electrode function," Biomaterials, vol. 34, pp. 4703-4713, 2013.
[17] M. P. Ward, et al., "Toward a comparison of microelectrodes for acute and chronic
recordings," Brain Research, vol. 1282, pp. 183-200, Jul 28 2009.
140
[18] G. C. McConnell, et al., "Implanted neural electrodes cause chronic, local
inflammation that is correlated with local neurodegeneration," Journal of Neural
Engineering, vol. 6, Oct 2009.
[19] D. H. Szarowski, et al., "Brain responses to micro-machined silicon devices," Brain
Research, vol. 983, pp. 23-35, Sep 5 2003.
[20] A. Gilletti and J. Muthuswamy, "Brain micromotion around implants in the rodent
somatosensory cortex," Journal of Neural Engineering, vol. 3, pp. 189-195, Sep 2006.
[21] J. N. Turner, et al., "Cerebral astrocyte response to micromachined silicon
implants," Experimental Neurology, vol. 156, pp. 33-49, Mar 1999.
[22] Y.-T. Kim, et al., "Chronic response of adult rat brain tissue to implants anchored
to the skull," Biomaterials, vol. 25, pp. 2229-2237, 2004.
[23] D. Egert, et al., "Parylene microprobes with engineered stiffness and shape for
improved insertion," in Solid-State Sensors, Actuators and Microsystems Conference
(TRANSDUCERS), 2011 16th International, 2011, pp. 198-201.
[24] J. P. Seymour and D. R. Kipke, "Fabrication of Polymer Neural Probes with Sub-
cellular Features for Reduced Tissue Encapsulation," in Engineering in Medicine and
Biology Society, 2006. EMBS '06. 28th Annual International Conference of the IEEE, 2006,
pp. 4606-4609.
[25] W. Fan, et al., "A flexible fish-bone-shaped neural probe strengthened by
biodegradable silk coating for enhanced biocompatibility," in Solid-State Sensors,
Actuators and Microsystems Conference (TRANSDUCERS), 2011 16th International, 2011,
pp. 966-969.
[26] P. J. Rousche, et al., "Flexible polyimide-based intracortical electrode arrays with
bioactive capability," Biomedical Engineering, IEEE Transactions on, vol. 48, pp. 361-
371, 2001.
[27] S. Takeuchi, et al., "Parylene flexible neural probes integrated with microfluidic
channels," Lab Chip, vol. 5, pp. 519-523, 2005.
[28] L. Sung Eun, et al., "A Flexible Depth Probe Using Liquid Crystal Polymer,"
Biomedical Engineering, IEEE Transactions on, vol. 59, pp. 2085-2094, 2012.
[29] A. Andrei, et al., "Fabrication and successful in-vivo implantation of a flexible
neural implant with a hybrid polyimide-silicon design," in Engineering in Medicine and
141
Biology Society (EMBC), 2012 Annual International Conference of the IEEE, 2012, pp.
3890-3893.
[30] A. Fomani and R. R. Mansour, "Fabrication and characterization of the flexible
neural microprobes with improved structural design," Sensors and Actuators A: Physical,
vol. 168, pp. 233-241, 2011.
[31] S. Takeuchi, et al., "3D flexible multichannel probe array," in Micro Electro
Mechanical Systems, 2003. MEMS-03 Kyoto. IEEE The Sixteenth Annual International
Conference on, 2003, pp. 367-370.
[32] S. Takeuchi, et al., "Parylene flexible neural probe with micro fluidic channel," in
Micro Electro Mechanical Systems, 2004. 17th IEEE International Conference on.
(MEMS), 2004, pp. 208-211.
[33] D. S. Pellinen, et al., "Multifunctional Flexible Parylene-Based Intracortical
Microelectrodes," in Engineering in Medicine and Biology Society, 2005. IEEE-EMBS
2005. 27th Annual International Conference of the, 2005, pp. 5272-5275.
[34] Y. Kato, et al., "Preliminary Study of Multichannel Flexible Neural Probes Coated
with Hybrid Biodegradable Polymer," in Engineering in Medicine and Biology Society,
2006. EMBS '06. 28th Annual International Conference of the IEEE, 2006, pp. 660-663.
[35] A. Mercanzini, et al., "Controlled Release Drug Coatings on Flexible Neural
Probes," in Engineering in Medicine and Biology Society, 2007. EMBS 2007. 29th Annual
International Conference of the IEEE, 2007, pp. 6612-6615.
[36] A. Mercanzini, et al., "Demonstration of cortical recording and reduced
inflammatory response using flexible polymer neural probes," in Micro Electro
Mechanical Systems, 2007. MEMS. IEEE 20th International Conference on, 2007, pp. 573-
576.
[37] T. D. Y. Kozai and D. R. Kipke, "Insertion shuttle with carboxyl terminated self-
assembled monolayer coatings for implanting flexible polymer neural probes in the brain,"
Journal of Neuroscience Methods, vol. 184, pp. 199-205, 2009.
[38] P. J. Gilgunn, et al., "An ultra-compliant, scalable neural probe with molded
biodissolvable delivery vehicle," in Micro Electro Mechanical Systems (MEMS), 2012
IEEE 25th International Conference on, 2012, pp. 56-59.
[39] J. T. W. Kuo, et al., "Fabrication of 3D Parylene Sheath Probes For Realiable
Neuroprosthetic Recordings," presented at the Solid-State Sensors, Actuators, and
Microsystems Workshop, Hilton Head Island, South Carolina, 2012.
142
[40] T. D. Kozai, et al., "Ultrasmall implantable composite microelectrodes with
bioactive surfaces for chronic neural interfaces," Nat Mater, vol. 11, pp. 1065-73, Dec 2012.
[41] J. R. Capadona, et al., "Mechanically adaptive nanocomposites for neural
interfacing," MRS Bulletin, vol. 37, pp. 581-589, 2012.
[42] W. Fan, et al., "A multi-shank silk-backed parylene neural probe for reliable
chronic recording," in Solid-State Sensors, Actuators and Microsystems (TRANSDUCERS
& EUROSENSORS XXVII), 2013 Transducers & Eurosensors XXVII: The 17th
International Conference on, 2013, pp. 888-891.
[43] T. Ware, et al., "Thiol-ene/acrylate substrates for softening intracortical
electrodes," Journal of Biomedical Materials Research Part B: Applied Biomaterials, vol.
102, pp. 1-11, 2014.
[44] P. Kennedy, et al., "Using human extra-cortical local field potentials to control a
switch," Journal of neural engineering, vol. 1, p. 72, 2004.
[45] P. R. Kennedy, "The cone electrode: a long-term electrode that records from
neurites grown onto its recording surface," J. Neurosci. Methods, vol. 29, pp. 181-193,
1989.
[46] P. R. Kennedy and R. A. Bakay, "Restoration of neural output from a paralyzed
patient by a direct brain connection," Neuroreport, vol. 9, pp. 1707-11, Jun 1 1998.
[47] P. R. Kennedy, et al., "The cone electrode: ultrastructural studies following long-
term recording in rat and monkey cortex," Neuroscience letters, vol. 142, pp. 89-94, 1992.
[48] P. R. Kennedy, et al., "Behavioral correlates of action potentials recorded
chronically inside the Cone Electrode," Neuroreport, vol. 3, pp. 605-8, Jul 1992.
[49] J. Bartels, et al., "Neurotrophic electrode: Method of assembly and implantation
into human motor speech cortex," Journal of Neuroscience Methods, vol. 174, pp. 168-176,
Sep 30 2008.
[50] K. P. Eaton and C. S. Henriquez, "Confounded spikes generated by synchrony
within neural tissue models," Neurocomputing, vol. 65, pp. 851-857, Jun 2005.
[51] J. T. W. Kuo, et al., "Fabrication of 3D Parylene Sheath Probes for Reliable
Neuroprosthetic Recordings," presented at the Hilton Head 2012: Solid State Sensors,
Actuators and Microsystems Workshop, Hilton Head Island, South Carolina, USA, 2012.
143
[52] B. J. Kim, et al., "3D Parylene sheath neural probe for chronic recordings," Journal
of Neural Engineering, vol. 10, p. 045002, 2013.
[53] E. Meng, Biomedical Microsystems, 1 ed.: CRC Press, 2010.
[54] A. V. Vasenkov, "Atomistic modeling of parylene-metal interactions for surface
micro-structuring," Journal of Molecular Modeling, vol. 17, pp. 3219-3228, 2011.
[55] L. Jeong Hoon, et al., "Microstructure and adhesion of Au deposited on parylene-c
substrate with surface modification for potential immunoassay application," Plasma
Science, IEEE Transactions on, vol. 32, pp. 505-509, 2004.
[56] P. Changlin, et al., "Electrolysis-based Parylene Balloon Actuators for Movable
Neural Probes," in Nano/Micro Engineered and Molecular Systems, 2007. NEMS '07. 2nd
IEEE International Conference on, 2007, pp. 913-916.
[57] K. Walsh, et al., "Photoresist as a sacrificial layer by dissolution in acetone," in
Micro Electro Mechanical Systems, 2001. MEMS 2001. The 14th IEEE International
Conference on, 2001, pp. 114-117.
[58] F. Guo-Hua and K. Eun-Sok, "Universal concept for fabricating micron to
millimeter sized 3-D parylene structures on rigid and flexible substrates," in Micro Electro
Mechanical Systems, 2003. MEMS-03 Kyoto. IEEE The Sixteenth Annual International
Conference on, 2003, pp. 594-597.
[59] W. Renxin, et al., "Fabrication and Characterization of a Parylene-Based Three-
Dimensional Microelectrode Array for Use in Retinal Prosthesis," Microelectromechanical
Systems, Journal of, vol. 19, pp. 367-374, 2010.
[60] R. Huang and Y.-C. Tai, "Flexible parylene-based 3-D coiled cable," in
Nano/Micro Engineered and Molecular Systems (NEMS), 2010 5th IEEE International
Conference on, 2010, pp. 317-320.
[61] S. W. Youn, et al., "Fabrication of a micro patterned parylene-C master by hot-
embossing and its application to metallic mold replication," Journal of Micromechanics
and Microengineering, vol. 17, pp. 1402-1413, Jul 2007.
[62] H. Kim and K. Najafi, "Characterization of low-temperature wafer bonding using
thin-film parylene," Microelectromechanical Systems, Journal of, vol. 14, pp. 1347-1355,
2005.
[63] H.-S. Noh, et al., "Parylene micromolding, a rapid and low-cost fabrication method
for parylene microchannel," Sensors and Actuators B, vol. 102, pp. 78-85, 2004.
144
[64] Y. Yu-Chang, et al., "A pneumatic neural probe structure fabricated by parylene
thermal bonding technique," in Micro Electro Mechanical Systems (MEMS), 2012 IEEE
25th International Conference on, 2012, pp. 381-384.
[65] H. S. Noh, et al., "Wafer bonding using microwave heating of parylene
intermediate layers," Journal of Micromechanics and Microengineering, vol. 14, p. 625,
2004.
[66] C. A. Gutierrez, "Development of Flexible Polymer-Based MEMS Technologies
for Integrated Mechanical Sensing in Neuroprosthetic Systems," Ph.D. Dissertation,
Biomedical Engineering, University of Southern California, Los Angeles, 2011.
[67] W. Li, et al., "Integrated Flexible Ocular Coil for Power and Data Transfer in
Retinal Prostheses," in Engineering in Medicine and Biology Society, 2005. IEEE-EMBS
2005. 27th Annual International Conference of the, 2005, pp. 1028-1031.
[68] R. Truckenmuller, et al., "Thermoforming of Film-Based Biomedical
Microdevices," Advanced Materials, vol. 23, pp. 1311-1329, 2011.
[69] H. Dreuth and C. Heiden, "Thermoplastic structuring of thin polymer films,"
Sensors and Actuators A: Physical, vol. 78, pp. 198-204, 1999.
[70] M. Ikeuchi and K. Ikuta, "The Membrane Micro Emboss (MeME) Process for
Fabricating 3-D Microfluidic Device Formed from Thin Polymer Membrane," in
Proceedings of the 10th International Conference on Miniaturized Systems for Chemistry
and Life Sciences, Micro Total Analysis Systems, µTAS, Tokyo, Japan, 2006, pp. 693-695.
[71] M. Ikeuchi and K. Ikuta, "Membrane micro emboss following excimer laser
ablation (MeME-X) process for pressure-driven micro active catheter," in Micro Electro
Mechanical Systems, 2008. MEMS 2008. IEEE 21st International Conference on, 2008, pp.
62-65.
[72] R. Truckenmüller, et al., "Low-cost thermoforming of micro fluidic analysis chips,"
Journal of Micromechanics and Microengineering, vol. 12, p. 375, 2002.
[73] J.-M. Hsu, et al., "Effect of Thermal and Deposition Processes on Surface
Morphology, Crystallinity, and Adhesion of Parylene-C," Sensors and Materials, vol. 20,
pp. 071-086, 2008.
[74] A. Kahouli, et al., "Ac-conductivity and dielectric relaxations above glass transition
temperature for parylene-C thin films," Applied Physics a-Materials Science & Processing,
vol. 106, pp. 909-913, Mar 2012.
145
[75] E. M. Davis, et al., "The influence of thermal history of structure and water
transport in Parylene C coatings," Polymer, vol. 52, pp. 5378-5386, 2011.
[76] A. Kahouli, et al., "Structural and dielectric study of parylene C thin films," Applied
Physics Letters, vol. 94, Apr 2009.
[77] J. J. Senkevich, et al., "Correlation between bond cleavage in parylene N and the
degradation of its dielectric properties," Electrochemical and Solid State Letters, vol. 7, pp.
G56-G58, 2004.
[78] D. W. Grattan and M. Bilz, "The Thermal Aging of Parylene and the Effect of
Antioxidant," Studies in Conservation, vol. 36, pp. 44-52, 1991.
[79] B. J. Kim, et al., "Three dimensional transformation of Parylene thin film structures
via thermoforming," in Micro Electro Mechanical Systems (MEMS), 2013 IEEE 26th
International Conference on, 2013, pp. 339-342.
[80] D. C. Rodger, et al., "Flexible parylene-based multielectrode array technology for
high-density neural stimulation and recording," Sensors and Actuators B: Chemical, vol.
132, pp. 449-460, 2008.
[81] R. Huang and Y. C. Tai, "Parylene to silicon adhesion enhancement," in Solid-State
Sensors, Actuators and Microsystems Conference, 2009. TRANSDUCERS 2009.
International, 2009, pp. 1027-1030.
[82] B. J. Kim, et al., "Three Dimensional Transformation of Parylene Thin Film
Structures Via Thermoforming," in The 26th IEEE International Conference on Micro
Electro Mechanical Systems (MEMS), Taipei, Taiwan, 2013.
[83] C. A. Gutierrez, et al., "Epoxy-less packaging methods for electrical contact to
parylene-based flat flexible cables," in Solid-State Sensors, Actuators and Microsystems
Conference (TRANSDUCERS), 2011 16th International, 2011, pp. 2299-2302.
[84] T. Pajkossy, "Impedance spectroscopy at interfaces of metals and aqueous solutions
- Surface roughness, CPE and related issues," Solid State Ionics, vol. 176, pp. 1997-2003,
Aug 2005.
[85] M. Grosser and U. Schmid, "The impact of annealing temperature and time on the
electrical performance of Ti/Pt thin films," Applied Surface Science, vol. 256, pp. 4564-
4569, 2010.
146
[86] R. Metzen and T. Stieglitz, "The effects of annealing on mechanical, chemical, and
physical properties and structural stability of Parylene C," Biomedical Microdevices, pp.
1-9, 2013/03/15 2013.
[87] A. Sharma, et al., "Evaluation of the packaging and encapsulation reliability in fully
integrated, fully wireless 100 channel Utah Slant Electrode Array (USEA): Implications
for long term functionality," Sensors and Actuators a-Physical, vol. 188, pp. 167-172, Dec
2012.
[88] B. D. Winslow, et al., "A comparison of the tissue response to chronically
implanted Parylene-C-coated and uncoated planar silicon microelectrode arrays in rat
cortex," Biomaterials, vol. 31, pp. 9163-9172, Dec 2010.
[89] E. Delivopoulos and A. F. Murray, "Controlled Adhesion and Growth of Long
Term Glial and Neuronal Cultures on Parylene-C," Plos One, vol. 6, Sep 2011.
[90] L. Wei and A. Lakhtakia, "Towards bioinspired parylene-C coatings of implant
surfaces," in Bioinspiration, Biomimetics, and Bioreplication 2012, San Diego, CA, 2012,
p. 83390R.
[91] J. C.-H. L. Lin, "MEMS for Glaucoma," Ph.D. Dissertation, California Institute of
Technology, 2012.
[92] S. F. Cogan, "Neural Stimulation and Recording Electrodes," Annual Review of
Biomedical Engineering, vol. 10, pp. 275-309, 2008.
[93] N. Beshchasna, et al., "Influence of artificial body fluids and medical sterilization
procedures on chemical stability of Parylene C," in Electronic Components and
Technology Conference (ECTC), 2010 Proceedings 60th, 2010, pp. 1846-1852.
[94] W. Gruber, et al., "Strain Relaxation and Vacancy Creation in Thin Platinum
Films," Physical Review Letters, vol. 107, p. 265501, 2011.
[95] W. Sripumkhai, et al., "Effect of Annealing Temperature on Platinum Thin Films
Prepared by Electron Beam Evaporation," Journal of the Microscopy Society of Thailand,
vol. 24, pp. 51-54, 2010.
[96] S. A. Hara, et al., "An Electrochemical Investigation of the Impact of
Microfabrication Techniques on Polymer-based Microelectrode Neural Interfaces,"
Journal of Microelectromechanical Systems, Submitted. 2015.
[97] S. Kim, et al., "Integrated wireless neural interface based on the Utah electrode
array," Biomedical Microdevices, vol. 11, pp. 453-466, 2009/04/01 2009.
147
[98] R. J. Vetter, et al., "Chronic neural recording using silicon-substrate microelectrode
arrays implanted in cerebral cortex," Biomedical Engineering, IEEE Transactions on, vol.
51, pp. 896-904, 2004.
[99] E. C. Gill, et al., "High-Density Feedthrough Technology for Hermetic Biomedical
Micropackaging," MRS Online Proceedings Library, vol. 1572, pp. null-null, 2013.
[100] J.-M. Hsu, et al., "Characterization of a-SiC thin films as an encapsulation material
for integrated silicon based neural interface devices," Thin Solid Films, vol. 516, pp. 34-
41, 2007.
[101] X. Xie, et al., "Plasma-assisted atomic layer deposition of Al2O3 and parylene C
bi-layer encapsulation for chronic implantable electronics," Applied Physics Letters, vol.
101, p. 093702, 2012.
[102] A. Neuro Connectors. Available: http://www.omnetics.com/neuro/
[103] A. ZIF-Clip Based Microwire Arrays. Available:
http://www.tdt.com/products/MW16.htm
[104] A. W-Series Systems. Available: http://www.trianglebiosystems.com/w-series-
systems.html
[105] B. A. Wester, et al., "Development and characterization of in vivo flexible
electrodes compatible with large tissue displacements," Journal of Neural Engineering, vol.
6, Apr 2009.
[106] K. B. Bjugstad, et al., "Biocompatibility of poly(ethylene glycol)-based hydrogels
in the brain: An analysis of the glial response across space and time," Journal of Biomedical
Materials Research Part A, vol. 95A, pp. 79-91, Oct 2010.
[107] Z. J. Chen, et al., "A realistic brain tissue phantom for intraparenchymal infusion
studies," Journal of Neurosurgery, vol. 101, pp. 314-322, Aug 2004.
[108] F. Pervin and W. W. Chen, "Mechanically Similar Gel Simulants for Brain
Tissues," in SEM Annual Conference, Indianapolis, Indiana, USA, 2010, pp. 9-13.
[109] D. J. Edell, et al., "Factors influencing the biocompatibility of insertable silicon
microshafts in cerebral cortex," Biomedical Engineering, IEEE Transactions on, vol. 39,
pp. 635-643, 1992.
148
[110] M. Welkenhuysen, et al., "Effect of Insertion Speed on Tissue Response and
Insertion Mechanics of a Chronically Implanted Silicon-Based Neural Probe," Biomedical
Engineering, IEEE Transactions on, vol. 58, pp. 3250-3259, 2011.
[111] W. Jensen, et al., "In-vivo implant mechanics of flexible, silicon-based ACREO
microelectrode arrays in rat cerebral cortex," Biomedical Engineering, IEEE Transactions
on, vol. 53, pp. 934-940, 2006.
[112] P. Rousche and R. Normann, "A method for pneumatically inserting an array of
penetrating electrodes into cortical tissue," Annals of Biomedical Engineering, vol. 20, pp.
413-422, 1992/07/01 1992.
[113] S.-A. Inc. Nerve Growth Factor-2.5S (N6009) [Online].
[114] C. S. Bjornsson, et al., "Effects of insertion conditions on tissue strain and vascular
damage during neuroprosthetic device insertion," Journal of Neural Engineering, vol. 3, p.
196, 2006.
[115] S. A. Hara, "The Electrochemical Evaluation of Parylene-based Electrodes for
Neural Applications," Ph.D. Dissertation, Biomedical Engineering, University of Southern
California, 2015.
[116] P. J. Chen, et al., "Q-enhanced fold-and-bond MEMS inductors," in Nano/Micro
Engineered and Molecular Systems, 2008. NEMS 2008. 3rd IEEE International
Conference on, 2008, pp. 869-872.
[117] W. Li, et al., "Integrated flexible ocular coil for power and data transfer in retinal
prostheses," in Engineering in Medicine and Biology Society, 2005. IEEE-EMBS 2005.
27th Annual International Conference of the, 2006, pp. 1028-1031.
[118] W. Li, et al., "Parylene-based integrated wireless single-channel neurostimulator,"
Sensors and Actuators A: Physical, vol. 166, pp. 193-200, 2011
149

“The treatment of hydrocephalus in infancy remains one of the most unsatisfactory and
discouraging problems the neurological surgeon is called upon to treat. Penfield’s
statement made in 1935 that ‘all surgeons who continue to face the hydrocephalus problem
require the support of fortified optimism’, is just as pertinent today.” – Franc Ingraham,
1948
Though Dr. Ingraham’s statement was made in 1948, much of it still applies today.
Shunt-based treatment for hydrocephalus remains plagued by high failure rates due to
obstructive events of the proximal catheter. Despite this, there are currently no practical
methods to precisely detect shunt failure. Treatment can be improved through an integrated
system that continuously monitors shunt dynamics to assess treatment efficacy. Towards
this aim, we present one element of a multi-sensor system designed to assess the patency
(i.e. obstruction) of the proximal catheter during the lifetime of the shunt. This integrated
patency sensor will help determine timely and accurate diagnoses of shunt failure for
improved treatment for hydrocephalus.

4.1 Background
4.1.1 Introduction to hydrocephalus
Hydrocephalus is a chronic condition characterized by the accumulation of excess
cerebrospinal fluid (CSF) within the ventricles of the brain due to an imbalance in its
production and drainage. The exact cause of hydrocephalus can vary, but there are several

PARYLENE MEMS PATENCY SENSOR FOR ASSESSMENT OF
HYDROCEPHALUS SHUNT OBSTRUCTION
150
factors that lead to its development, such as congenital malformations, tumor, trauma,
meningitis, and hemorrhages; it can be both congenital and acquired [1, 2].
Prior to further discussion of its etiology, it is useful to present a brief overview of
the relevant structures within the brain pertinent to hydrocephalus. The ventricular system
of the brain consist of four structures that are all connected via narrow passages: two lateral
ventricles, the third ventricle, and the fourth ventricle (Figure 4-1a). CSF is primarily
produced by the choroid plexus, a group of cells that line each of the ventricles. It drains
out of the fourth ventricle towards the subarachnoid space (SAS), a CSF-filled cushion-
like layer that covers the surface of the brain and the spinal cord, where it is then absorbed
into the bloodstream (Figure 4-1b). Any impairment of this drainage and absorption
process can lead to the onset of hydrocephalus.

Figure 4-1. (a) Cartoon of 2D view of ventricular system within the brain comprising of
the lateral ventricles, third ventricle, and fourth ventricle connected via narrow passages.
Choroid plexus is shown in red. (b) Cartoon anatomy of brain illustrating the pathway of
CSF from its production within the ventricles, and towards the subarachnoid space (SAS)
where it is absorbed.

151
Hydrocephalus is classified by the nature of disruption to CSF hydrodynamics,
either in its drainage or absorption. Non-communicating (obstructive) hydrocephalus is
derived from an impairment of the fluidic drainage pathway of CSF out of the ventricular
structures, while communicating hydrocephalus results from an impairment in the
absorption of CSF into the bloodstream occurring within the SAS. Regardless of the form,
the prohibition of CSF drainage or absorption leads to an increase in CSF volume within
the ventricles and a subsequent increase in intracranial pressure (ICP). The large
accumulation of CSF expands the ventricles and causes swelling and deformation of the
brain, leading to significant cortical injury and inflammation, manifesting itself
symptomatically through neurological defects (e.g. learning disabilities, epilepsy, the
Hakim triad [3]) and often death. One notable form of communicating hydrocephalus is
normal pressure hydrocephalus, where swelling of the ventricles does not cause a
correlating increase in ICP and is currently the basis for a large degree of clinical research
as not much is known about its cause and treatment [4, 5].

4.1.2 Hydrocephalus treatment methods
There are two common surgical treatments for hydrocephalus: endoscopic third
ventriculostomy (ETV) and ventricular shunt insertion. In the former case, an
intraventricular endoscopic approach is used to create a hole in the floor of the third
ventricle to create a CSF diversion pathway from the ventricular space to the basilar
cisterns of the SAS. Though this is the preferred method for surgeons due to the lack of
implanted hardware, many patients are excluded because of anatomic complications (e.g.
prepontine scarring, large massa intermedia, or slit ventricle syndrome). ETVs also can
have high failure rates (overall success rate of 56%), depending on the etiology of
hydrocephalus, due to either inadequate CSF diversion via the SAS spaces or subsequent
obstruction of the stoma due to scar formation [6].

152
The implantation of a shunt is the most common treatment for both forms of
hydrocephalus and has been successfully performed since the 1950s [7]. An implanted
shunt system consists of a multi-pore silicone proximal catheter, valve system, and a
silicone distal catheter designed to drain excess CSF from the ventricles (Figure 4-2a). The
proximal catheter, placed within the ventricle, is connected to a pressure- or flow-
controlled one-way valve (set pressure or programmable) that directs the fluid out of the
ventricle and through the distal catheter to drain the CSF either within the peritoneal cavity
(ventriculoperitoneal (VP) shunt), pleural cavity (ventriculopleural (VPL) shunt), or within
the atrium of the heart (ventriculo-atrial (VA) shunt) to be reabsorbed (Figure 4-2b). In the
case that a temporary (<2 weeks) treatment is necessary, an external ventricular drain
(EVD) system can be used, where the implanted proximal catheter directly drains to an
external fluidic container at the patient’s bedside.

Figure 4-2. (a) Implantable Medtronic Delta valve shunt system consisting of ports for
proximal and distal catheters to control CSF flow from the ventricles. (Image modified
from [8]) (b) Comparison of distal catheter placement (pink) VP shunts (peritoneal cavity)
and VA shunts (atrium of the heart).

Though effective, these shunts fail at an alarming rate of 40% within the first year
[9] and 80% by ten years, largely due to complications. There are many causes of shunt
failure, including mechanical (i.e. hardware-related) issues [10] and infection [11], but the
most common is obstruction of the drainage ports (70% of patients) [12-15]. Although
153
obstruction of the shunt cannot be attributed to a single cause, there have been many studies
within literature that have found obstructive shunt failure due to ingrowth or attachment of
choroid plexus tissue, inflammatory cells, blood components, and other cellular debris [14,
16-20]. This cellular and tissue attachment has been attributed to the typical foreign body
response for implants in addition to disruption of the blood-brain barrier upon implantation
[19-21], increase in shear stress around high flow regions (i.e. at the holes) [22, 23], as well
as proximity to tissue structures (i.e. improper placement of the shunt) [18, 20, 24-26] .
Shunt failures can impede treatment and efficacy of the shunt, and cause severe symptoms,
which can lead to emergent shunt revision surgery [27].
Despite the severity of this problem, there are currently no reliable and convenient
methods to detect shunt obstruction failure. Obstruction can only present itself in non-
specific symptoms, such as headaches and nausea, which can be misdiagnosed. In addition,
obstruction can also present itself acutely requiring urgent or emergent surgical
intervention. The current clinical standard to assess shunt obstruction is the use of static
imaging of the brain (i.e., magnetic resonance imaging (MRI), computed tomography (CT)
scans) for initial analysis, followed by an invasive shunt tap procedure, where a needle is
punctured through the skin into a reservoir within the implanted valve to assess CSF flow.
These methods are not only unreliable and largely subjective based on the clinician’s
expertise, but may also involve secondary risks to patients such as radiation and surgical
complications. Given the high risk and incidence associated with shunt failure, an
improved paradigm in hydrocephalus treatment would be the development of a “smart
shunt,” capable of periodically assessing ICP, CSF flow, and shunt patency using
integrated sensors so that treatment can be improved through the timely and accurate
diagnoses of shunt failure [28].

4.1.3 “Smart shunt” paradigm
The development of microelectromechanical systems (MEMS) technology has
allowed for the miniaturization of sensors for integration with shunts, and have brought an
154
onset of various micro-scale pressure [29-31] and flow sensors [32-36] to assess shunt
status towards the aim of this “smart shunt.” Several of the large shunt-manufacturing
companies have also developed sensors that are designed to work as add-ons to current
implanted shunt systems for on-demand, single-time-point measurements of either ICP or
CSF flow data; these include sensors from Medtronic, Raumedic, Radionics, Meithke,
Issys, Codman Neuro-DePuy Synthes, ShuntCheck, and Transonic Systems. These devices
however utilize typical MEMS silicon-based sensors that suffer from issues related to long-
term in vivo performance due to their chronic contact with biological fluids. In addition,
there have been reports of clinically relevant electrostatic discharges affecting the baseline
of these sensors [37]. Additionally, some of these devices are not fully integrated with
shunt technologies, creating temporary disruptions during measurement, and are sensitive
to patient movement and activity, confounding readings [3, 38]. Beyond issues related to
the robustness of the technology, these sensors are limited to single time points of ICP and
CSF flow, and do not continuously monitor shunt performance.
Such devices, capable only of monitoring ICP and CSF flow rates, lack accurate
assessments of obstructive failure and fall short of a true “smart shunt.” Patency data must
be inferred from both ICP and CSF flow measurements, which necessitates continuous (i.e.
not a single-time point) measurement as well as real time signal processing and analysis
[38-40]. These measurements are further complicated in the case of normal pressure
hydrocephalus, where ICP measurements do not necessarily correlate with the typical
prognosis for hydrocephalus. This has spurred research beyond the conventional sensing
paradigms of pressure and flow and towards other parameters such as ventricular volume
[41, 42], ventricular size [43], intracranial compliance [44], and a form of shunt patency
[45] to provide more direct measurements of shunt efficacy.
Towards this aim, we have developed a Parylene-based device that employs
electrochemical sensing for patency monitoring, leveraging the wet in vivo environment
and eliminating concerns with sensor integration. Building off of previous work in the
development of Parylene-based electrochemical-MEMS sensors, we have developed a
155
multi-sensor system consisting of Parylene-based pressure [46], flow [47], and patency
[48] sensors designed to be integrated with existing shunt systems and provide reliable
diagnosis of shunt status. The combination of all three sensors enables continuous
measurement of ICP, CSF flow, and shunt patency in one package and provides
measurement redundancy for more robust clinical performance. The choice of Parylene C
as a medium offers substrate flexibility, to facilitate multiple packaging schema (e.g.
implantable module for VP shunt systems or larger non-invasive module for use in EVDs),
batch-scale fabrication, through compatibility with conventional micromachining
processes, and high biocompatibility, by exploiting Parylene’s proven track record as a
USP class VI material.
In this chapter, we focus on the development and characterization of the patency
sensor as capable of providing direct, periodic measurements of shunt obstruction. This
sensor utilizes a very simple transduction scheme to assess the degree of shunt obstruction
by measuring changes in electrochemical impedance in the conductive path through the
drainage ports in the shunt catheter. This patency sensor is the first of its kind for use in
hydrocephalus catheters, and will enable quantitative monitoring of shunt performance and
more importantly, provide accurate and timely diagnosis of impending failure to improve
treatment for hydrocephalus patients.

4.2 Sensor design and operation
The sensing mechanism, similar to previously presented electrochemical-MEMS
sensors [49-51], comprises only the measurement of impedance between a pair of
electrodes immersed in an electrolyte solution (Figure 4-3a). The Parylene-based patency
sensor consists of two electrodes, one positioned on each internal and external surface of
the catheter, such that the catheter ports establish an ionic conductive path between them
(Figure 4-3b). When measuring the electrochemical impedance at a sufficiently high
frequency (fmeas) to isolate the solution resistance, any disturbances in the volumetric
156
conduction path between the two electrodes (i.e. port blockage) will register as impedance
changes (Figure 4-3c).

Figure 4-3. (a) Equivalent circuit model of two electrodes in an electrolyte, highlighting
the solution resistance. (b) Conceptual cartoon of impedance sensing mechanism of
patency sensor. Two electrodes on the internal and external surfaces of the catheter are
fluidically connected via the drainage ports. Obstruction of these ports impedes the ionic
conduction path between the electrodes, and (b) the electrochemical impedance between
the electrodes increases for measurements above a certain frequency (fmeas).

Thus in this configuration, the sensor is effectively a variable resistor, and various
parameters can induce changes in the measured resistance following Equation 4-1: either
(1) the resistivity of the solution (i.e. the ionic concentration of the solution) (ρ), (2) the
cross sectional area between the electrodes (A), or (3) the distance between the electrodes
(l).

Z ≈ R
s
=
ρl
A
; for sufficiently high f (4-1)

A decrease in the number of open holes in the catheter, a consequence of shunt
obstruction, will decrease the available cross sectional area between the pair of electrodes,
closing off the available conductive paths. Thus, we expect that the relationship between
157
measured impedance and obstruction should follow an inverse relationship (i.e. 1/A) per
Equation 1.

4.3 Sensor fabrication and packaging
The patency sensor was constructed using standard surface micromachining
processes for Parylene MEMS devices. Platinum (Pt) electrodes (2000 Å) were e-beam
evaporated and patterned using the liftoff method on a Parylene C substrate (12 µm)
deposited on a silicon carrier wafer. The choice of a 12 µm base layer rather than a 5 µm
layer used in the previous devices stemmed from advice obtained from Specialty Coating
Systems in using film thickness > 10 µm for better soak performance. Platinum was chosen
as it has shown outstanding inertness for in vivo and electrochemical sensing applications
[52]. The electrodes were insulated by depositing another layer of Parylene C (12 µm) and
electrode sites were exposed via oxygen plasma etching. The final device was released by
a complete cut out etch in oxygen plasma.

Figure 4-4. Process flow for fabrication of Parylene patency sensor. (a) Parylene C (12
µm) is deposited on a silicon carrier wafer. (b) Pt electrodes (2000 Å) are deposited via e-
beam evaporation and patterned using liftoff. (c) An insulation layer of Parylene C (12 µm)
is deposited and electrode sites are exposed using oxygen plasma etching. (d) Devices are
released from wafer by gentle peeling while immersed within DI water.

158
Free film sensor dies were released from the carrier wafer by first stripping the
protective photoresist mask used during the final release etch, and then gentle peeling using
tweezers; immersing the wafer in deionized (DI) water facilitated the process. Released
devices were electrically packaged by fitting the contact pad end of the sensor dies into a
zero-insertion-force (ZIF) connector (12 channel, 0.5 mm pitch; Hirose Electric Co., Simi
valley, CA) soldered onto a flat flexible cable (FFC; Molex Inc., Lisle, IL), a commonly
used technique for Parylene devices [53] (Figure 4-5). As the ZIF connector requires an
inserted cable of a certain thickness and stiffness, the contact pad regions of the Parylene
device were affixed to a poly(etheretherketone) (PEEK) stiffener backing (300 µm) prior
to insertion into the connector. The ZIF connector region was encapsulated in EpoTek 353-
NDT biocompatible epoxy (Epoxy Technology, Inc., Billerica, MA), before further
assembly into the module (Figure 4-5c).
The electrically packaged sensor was then inserted into two different luer-lock
modules (cap and inline configurations) designed for early acute validation studies in
humans using an EVD system (Figure 4-6) as a critical translational step towards the
development of an implantable sensor module. In one arrangement, the cap module, the
Parylene device was first affixed within a slit of a rubber stopper on top of a cap housing
using EpoTek 353-NDT biocompatible epoxy (Epoxy Technology, Inc., Billerica, MA),
which was then filled with artificial CSF (aCSF). A 3-way valve system allowed for
attachment of the cap module to the rest of the testing system. In a second arrangement,
the inline module was formed by affixing the sensor platform within a milled slit in a luer-
lock adaptor using EpoTek 353-NDT biocompatible epoxy (Epoxy Technology, Inc.,
Billerica, MA). The inline design removed the need to prefill the module and also allowed
for simpler integration with the drainage system, eliminating the extra 3-way valve
component. In both modules, the integrated FFC was used as the connection scheme to the
impedance measurement system.
Initially, four electrodes having different surface areas were fabricated on a single
device packaged within the cap module (v1) to evaluate effects of electrode size on sensor
159
performance. Following design down-select, a single electrode size packaged in a luer lock
module (v2) was used for all subsequent testing and characterization.

Figure 4-5. (a) Optical micrograph of v1 Parylene device with 4 electrode designs to see
the effect of electrode size on performance. (b) Electrically packaged v1 Parylene device
using a ZIF connector and integrated flat flexible cable (FFC). (c) Electrically packaged v2
Parylene device with single electrode design and biocompatible epoxy (yellow) for
encapsulation.

Figure 4-6. Fluidically packaged sensors in (a) cap and (b) inline modules used for
benchtop testing. The inline module was designed for sensor integration with external
ventricular drainage systems for clinical validation studies. In this form factor, sensors are
curled within the lumen to permit uninterrupted flow.

160
4.4 Sensor characterization
4.4.1 Sensor characterization setup
For benchtop testing, blockages of the catheters were simulated by mock silicone
catheters of varying numbers of holes, with 16 holes simulating 100% open (a 4-holed
catheter would be classified as 75% blockage, 8 holes as 50%, etc.) (Figure 4-7a) [23].
Mock catheters were constructed by sealing one end of a silicone tube (1.5 mm ID) with
cured silicone and manually punching holes with a 15 gauge coring needle to create 1 mm
diameter holes (Figure 4-7b), similar to those used in a proximal catheter of a Medtronic
hydrocephalus shunt (Becker® EDMS Ventricular Catheter, Medtronic, Minneapolis,
MN).

Figure 4-7. (a) Mock silicone catheters (ID = 1 mm) with varying number of holes used
for benchtop testing. (b) Magnified image of manually punched hole (Ø = 1 mm) using a
15 gauge coring needle.

The catheter was then placed within a beaker of aCSF (ionic formulation given in
Table 4-1) and connected to the sensor module using a 1/16” barb-to-luer connector. The
assembly was filled via a syringe or peristaltic pump (for static and flow conditions,
161
respectively) prior to testing (Figure 4-8). A platinum wire electrode was placed within the
beaker to close the circuit and complete the sensing setup. Impedance measurements were
acquired using a Gamry R600 potentiostat for experiments requiring measurement over a
frequency range for initial characterization, or an Agilent e4980 (1 Vp-p., 10 kHz; Agilent
Technologies, Santa Clara, CA) or HP/Agilent 4285a (1 Vp-p., 75 kHz; Agilent
Technologies, Santa Clara, CA) for measurements at a single frequency.

Table 4-1. Ionic formulation of artificial CSF used in benchtop experiments. Compound
weights are given for mixture into 1 L of DI water.
Compound Weight (g)
NaCl 8.66
KCl 0.224
CaCl2 · 2H2O 0.206
MgCl2 · 6H2O 0.163
Na2HPO4 · 7H2O 0.214
NaH2PO4 · H2O 0.027

Figure 4-8. (a) Image of experimental setup demonstrating fluidic connection between the
catheter and sensor module. (b) Experiments were conducted using either the (i) cap or
inline module with a (ii) syringe or peristaltic pump for static and flow conditions,
respectively. Impedance was measured between the sensor electrode and platinum ground
electrode in the beaker.

A series of experiments were conducted for benchtop characterization of the
Parylene patency sensor. First, the fmeas that isolates the solution resistance within the
162
sensor’s impedance response was determined in order to obtain the optimal sensing
performance for patency. Then, sensor sensitivity measurements were performed to assess
the relationship between measured electrochemical impedance and number of holes (i.e.
percent shunt blockage). These experiments were conducted using four electrode sizes all
fabricated on a single device (v1) to evaluate the effects of electrode size on sensor
performance and for design down-selection.
Following down-selection of the ideal electrode size, a new device was fabricated
(v2) with the chosen electrode size for further characterization in simulated in vivo
conditions. As these devices are to be used within the clinic, the effects of clinically
relevant patient temperatures as well as possible flow rates within the shunt systems were
evaluated. The effect of temperature on sensor performance was evaluated by using a hot
plate to maintain the beaker of aCSF at a constant temperature between 32–44°C. A
temperature probe was placed within the aCSF solution to record the temperature. Flow
was generated in the system by connecting the inline module to a peristaltic pump (WM
120U DM3, Watson Marlow, Wilmington, MA). A 0.38 bore tubing was chosen for use
with the peristaltic pump to provide physiologically relevant flow conditions: 0.03 ml/min
– 0.6 ml/min at 120 pulses/min [22]. Functionality of the devices following hydrogen
plasma (H2O2) sterilization, a commonly used sterilization technique used by hospitals,
performed with a Sterrad 100 NX system (Advanced Sterilization Products, Irvine, CA)
was also assessed to examine performance post-sterilization. Additional benchtop studies
of sensor drift were also conducted using a “bottle brain” benchtop model consisting of a
closed bottle system with external pressure control, and ports for catheter input and a Pt
ground wire.

4.4.2 V1 sensor characterization
Impedance measurements were conducted at frequencies between 0.1 Hz – 1 MHz
and the magnitude and phase as a function of frequency are presented in Figure 4-9a and
Figure 4-9b, respectively. An optimal measured fmeas was determined for each sensor size
163
(Table 4-2). Frequency ranges >fmeas, corresponding to where the solution resistance
dominates the impedance response, correlated well with simulated catheter blockage.

Figure 4-9. (a) Electrochemical impedance spectroscopy results of the impedance
magnitude of the E1 electrode indicating variations for measurements frequencies > 10
kHz between different catheter blockages. (b) Electrochemical impedance spectroscopy
results of the impedance phase of the four electrode sizes demonstrating varying optimal
measurement frequencies (where phase = 0°). Dotted line indicates 10 kHz, the fmeas
observed for the E1 electrode.

Table 4-2. Obtained optimal impedance measurement frequencies and sensitivities for
electrodes of the Parylene-based EC-MEMS patency sensor.
Electrode
design
Surface
area (µm
2
)
fmeas (kHz)
Sensitivity
(%ΔZ / %Blockage)
E1 300,000 10 0.183
E2 20,000 30 0.157
E3 20,000 30 0.168
E4 17,320 30 0.161

By analyzing the data for each electrode at its corresponding optimal measurement
frequency, calibration curves (“patency curves”) were generated and results indicated that
the impedance varied inversely with the number of open holes (percent blockage) of the
catheter (Figure 4-10). These results suggest that increasing catheter hole blockages alters
the cross sectional area term of equation 1; reducing the number of open holes reduces the
available conduction paths between the pair of electrodes and thus increases the impedance.
164
In analyzing the results, though all electrode sizes were similar in performance, the largest
electrode (E1) was chosen as the final design moving forward as: (1) the sensitivity was
slightly higher than the others (0.183 %Δimpedance/% blockage) and (2) a larger electrode
surface area has been shown to reduce noise and drift for electrochemical impedance
measurements [51]. In all subsequent experiments, fmeas was set as 10 kHz for the E1
electrode.

Figure 4-10. Representative patency curve obtained for the E1 electrode within the cap
module indicating an inverse relationship between measure impedance magnitude and the
number of open holes (percent blockage).

4.4.3 V2 sensor characterization
Characterization experiments with the inline module packaged device determined
that the newly designed and packaged patency sensor performed identically to v1 of the
device. The cured biocompatible epoxy used to seal the Parylene device within the module
was qualitatively tested under pressure for air or liquid leaks by subjecting the inline
module to nitrogen pressure in empty and liquid filled cases. No air leaks were observed
up to 1500 mmHg (limit of the experimental setup) and no liquid leaks were observed up
to 100 mmHg, both considerably higher than the expected pressure ranges within the brain
(0-25 mmHg).

165
Patency curves for these devices retained an inverse relationship and a similar
sensitivity observed in previous experiments, demonstrating an increase of ~27%
impedance magnitude increase for a 87.5% (2 holes open) blockage (Figure 4-11). In an
analysis of the sensor response, the variation in measurement was determined to be 100-
200 Ω, which is within the resolution of the measurement system. Using the high precision
LCR meters for benchtop testing, the devices can resolve 0.2% of the baseline impedance
(100-200 Ω), which correlates to 2-3% obstruction. For clinical application with external
ventricular drains, a data logger board was developed that utilized a microcontroller to take
impedance measurements with the sensor for storage to a µSD card; this system has shown
a resolution of 3% of the baseline, but impedance changes correlated to obstruction events
of the catheter are expected to be greater than a 3% impedance change. Further in vivo
testing is necessary to correlate a patency sensor measurement with what would qualify as
a clinical shunt obstruction failure. But currently, complete obstruction of the catheter
results in a sensor reading of 3.584 MΩ, which is >8000% increase in measured impedance.

Figure 4-11. Patency curve obtained for v2 sensor within inline module confirming a
similar shape to v1 results with the E1 electrode within the cap module.

4.4.4 Hole position dependency
When simulating catheter obstruction in benchtop experiments, mock catheters
were fabricated to model hole obstruction in a specific direction, with simulated blockages
166
occurring from the top (i.e. “proximal” to the sensor) downwards (i.e. “distal” to the sensor)
(Figure 4-12a). In an experiment where the order of obstruction was reversed, such that the
blockage occurred from the distal-end proximally (i.e. bottom-up) (Figure 4-12a),
decreased sensitivity was observed even though the total number of open holes remained
the same (Figure 4-12b). Closer analysis of the results indicated that for the 2, 4, 6, and 8-
holed catheters within the “Reversed” catheter set, measured impedances demonstrated
similar values as long as the most proximal hole was in the same position and was patent.
This follows from a “path of least resistance” effect, where a majority of the ionic
conduction path between the two electrodes is contributed across the nearest opening (i.e.
the most proximal hole). This result revealed that the sensor predominantly monitors the
patency (or position) of the hole closest to it.

Figure 4-12. (a) Two different catheter hole orientations (normal and reversed) used to
assess the transduction mechanism of the sensor. (b) Resultant patency curves of the two
catheter orientations indicate that the sensor was not measuring the number of open holes,
but rather the position of the first open hole, following a “path of least electrochemical
resistance” effect.

This result was confirmed in another experiment exploring two different catheter
sets with varying hole orientations. In the first set (“Normal”), mock catheters were
constructed such that the total number of open holes were 2, 4, 8, and 16 (Figure 4-13a).
In the second set (“Alternate”), mock catheters were constructed with a total number of
open holes of 2, 6, 14, and 16, as shown in Figure 4-13b. It is important to note that when
167
comparing the 4-holed catheter from the Normal set and the 6-holed catheter from the
Alternate set, while the total number of open holes is different, the top-most hole is in the
same position. A similar comparison can be made for the 8 and 14-holed catheters from
the Normal set and Alternate set, respectively. When looking at the sensor patency curves
for these two sets of catheters, a similar value is recorded between the catheters with
identical top hole positions (described previously) due to the “path of least electrochemical
resistance” effect (Figure 4-13b).

Figure 4-13. (a) Two different catheter sets (normal and alternate) used to assess the
dependency of the sensor response on the first hole. Total number of open holes are given
above each catheter drawing. Note that each catheter between the sets has the same top-
most open hole position but a different number of open holes (e.g. 4 holed-catheter of the
Normal set has the same top-most hole position of the 6-holed catheter in the Alternate
set). (b) Resultant patency curves of the two catheter sets indicate that even with varying
number of total open holes, the measured impedance is dependent only on the top most
hole, following a “path of least electrochemical resistance” effect, and thus a more linear
response in the case for blockages from the top of the catheter downwards.

It follows that the patency curves for the sensor depend largely on the obstruction
model for the mock catheters; if the obstruction occurs in a linear manner specifically from
the top hole downwards, then the patency curve is linear. This conclusion requires adjusting
our model of the sensors, such that the variable parameter in Equation 1 is the l, or the
distance between the electrodes, and not the A term— in considering the sensing
168
mechanism of the patency sensor for these catheters, the only hole of importance is the top-
most hole, the remaining holes have marginal impact on electrochemical impedance. Thus
blockages of the catheter from the top down merely extend the ionic conduction path
between the two electrodes.
Following from the previous discussion, if the obstruction model for the
hydrocephalus catheter does not follow a linear relation from the top hole downwards, then
the sensor becomes insensitive to specific blockages and requires a new sensor orientation.
For these cases, four sensor electrode orientations (Figure 4-14) utilizing Pt wire electrodes
were investigated to see the possible effects on improving the hole specificity of the sensor.
In orientations A and C, the ground electrode is placed in between the hole pairs with the
hopes to improve specificity while keeping all electrodes within the catheter, while B and
D are designed to improve hole specificity while maintaining a similar orientation to the
normal device as the ground electrode is placed outside of the catheter.

Figure 4-14. 4 electrode orientations designed to improve the hole specificity of the sensor.

For the normal electrode orientation within the inline modules, the impedance
magnitudes measured when a total of 6 holes are patent for varying blocked hole
orientations is shown in Figure 4-15. Note that for positions 2-4, the values remain
consistent while the hole position 1 (when the top most hole is blocked) the impedance is
increased, resulting in measurement variability. In analyzing the sensor responses of the
169
new electrode orientations (Figure 4-16), we can see benefits in improving electrode
blockage specificity as the standard deviations for the measured impedance for the 6 hole
catheters of varying positions is considerably reduced compared to the baseline (Table 4-3).
However, for all new electrode orientations, though the hole specificity was improved, the
sensitivity was severely decreased to an impedance change of ~10% for a blockage of
87.5% (compared to ~27% obtained in our normal orientation). In addition, these
orientations require additional electrode integration methods to modify current catheter
systems for clinical implementation, which was recommended against by clinical design
advisors— quicker device adoption among neurosurgeons would be facilitated by less
substitutions with their currently used hardware [54]. However, moving forward, these
designs can be considered for future sensor orientations.

Figure 4-15. (a) Impedances measured using the inline module for (b) a 4 catheter set with
six open holes of varying positions. Note that impedances are constant for catheters 2-4
while catheter 1 has a large increase in impedance due to the patency of the top-most hole.

170

Figure 4-16. Patency curves obtained for the 4 electrode orientations: (a) orientation A,
(b) orientation B, (c) orientation C top pair, (d) orientation C bottom pair, (e) orientation D
top pair, and (f) orientation D bottom pair. These plots highlight the measured impedances
for the 6-holed catheter set given in Figure 4-15b.

171
Table 4-3. Measured standard deviations (n = 4) of 6-holed catheter impedance
measurements (varying blocked hole positions) to compare efficacy of electrode design in
improving hole specificity.
Electrode
Orientation
Pair SD of |Z| of 6-holed
catheters
Normal - 2581.70
A - 140.40
B - 185.53
C
Bottom 977.38
Top 928.71
D
Bottom 597.34
Top 518.21

Fortunately for the normal orientation, high patency sensitivity for the top-most
hole is preferred, as reported by benchtop studies in the literature that suggest 80% of the
flow through the catheters occurs at these top holes [18, 26]. The high flow [26] and greater
shear stresses [22, 23, 55] that occur across this hole greatly increases the likelihood of
obstruction at this site (Figure 4-17). These results were confirmed in our own benchtop
flow studies with colored dye and our mock catheter system; bulk flow only through the
top most hole was qualitatively confirmed (Figure 4-18). Though these results may vary
for chronically implanted catheters, these benchtop results are promising. Monitoring the
top hole may be critical if not sufficient for assessing shunt patency. In addition, this
electrode orientation is optimal for tracking other clinical phenomena that may occur
during the implantation or lifetime of the shunt catheter, namely assessing improper
placement of the shunt where the top-most hole is placed right next to the ventricular wall
[18, 24, 26], or ventricular size changes during treatment and subsequent catheter
movement (e.g. slit or borderline slit ventricles) [19, 20]. Both have been shown to
correlate with the top-most hole being obstructed in explanted catheters.

172

Figure 4-17. Images of explanted proximal catheters that indicate the tissue obstruction
within the top most holes. (Reproduced with permission from the Journal of Neurosurgery
[26])

Figure 4-18. (a) Time-lapse images (1→4) of a dye drop experiment into a 16-holed
catheter within aCSF under flow (0.3 mL/min). Note that though the dye was placed at the
bottom holes, a blue dye stream is not observed until the diffused dye reaches the top hole
(image 3). Even in image 4, though the dye is within the proximity of the other holes, dye
flow is still only observed through the most proximal hole. (b) Magnified image of catheter
illustrating the blue flow stream into the catheter only at the most proximal hole (white
circle).

4.4.5 Thermal effects
Patency curves for the inline module tested in aCSF solution temperatures between
32-44°C indicated a reduction in baseline impedances largely due to increased ionic
173
mobility (i.e. increased conductivity of the solution) at elevated temperatures (Figure
4-19a). To better understand the dependence of solution conductivity on temperature, it is
useful to consider Einstein’s relation between a particle’s diffusion coefficient and its
mobility:

D = µ𝑘 𝐵 𝑇 (4-2)

where D is the diffusion coefficient, µ is the particle’s mobility, kB is Boltzmann’s constant,
and T is the absolute temperature.
Specifically for ionic solutions, the Stokes-Einstein relation (Equation 4-3) is used
to assess the diffusion coefficient for spherical particles (i.e. ions) traveling through a
liquid. By combining the Stokes-Einstein relation with the Nernst-Einstein equation
(Equation 4-4), the conductivity (σ) of a model electrolyte solution is defined (Equation 4-
5):

D =
𝑘 𝐵 𝑇 6𝜋𝑟𝜂 (4-3)

σ =
(𝑧𝑒 )
2
𝑐 𝑘 𝐵 𝑇 𝐷 (4-4)

σ =
(𝑧𝑒 )
2
𝑁 6𝑉𝜋𝑟𝜂 (4-5)

where r is the ionic radius, η is the viscosity of the solution, z is the valence number, e is
the unit charge, c is concentration, V is volume, and N is the number of charge carriers.
According to Equation 4-5, we see that the conductivity of the solution is proportional to
the number of charge carriers (N) and inversely proportional to the viscosity of the solution
174
( η). Viscosity varies with temperature following an Arrhenius relationship (Equation 4-6;
R is the universal gas constant, T is temperature, A is a proportionality constant, and E is
the activation energy for viscous flow), an increase in temperature would cause a
subsequent decrease in viscosity, leading to an increase in conductivity and thus a decrease
in the measured impedance. It has been shown within literature that for ultrapure water,
even a variation in temperature of 1°C causes an increase in conductivity of 5.55% [56].

𝜂 = 𝐴 𝑒 [
𝐸 𝑅𝑇
]
(4-6)

The temperature coefficient of resistance (TCR) calculated for the Parylene patency
sensor was -287.7 Ω/°C within this benchtop system. Experiments that explored more
clinically relevant hydrocephalic patient temperatures (34-38°C [54]) indicated no large
variations for baselines for the patency curves for the inline module (Figure 4-19b). These
devices are largely insensitive to patient temperature swings, as expected impedance
increases due to shunt obstructive failure is much higher than a slight decrease in baseline.

Figure 4-19. (a) Patency curves obtained for the inline modules for varying temperatures
32–44°C, indicating a decrease in baseline, but retaining a similar sensitivity. (b) Within
clinically relevant temperatures between 34-38°C, no large variations were recorded.
175

4.4.6 Flow effects
Similar to observed effects for increased temperatures, the presence of flow within
the system (and thus across the sensor electrodes) reduced the baseline impedance (~8%)
largely due to increase ionic mobility (i.e. µ in Equation 4-2) (Figure 4-20). The use of
electrochemical impedance to measure fluid flows has been demonstrated within literature
and the sensitivities of impedance to flow largely depend on the electrode orientations [57].
For the inline module, the specific electrode orientation as parallel to the fluid flow as well
as the relatively large separation (>7 cm apart) resulted in similar sensor responses at
clinically relevant flow ranges 0.03 ml/min – 0.60 ml/min [33] (Figure 4-20). Thus the
sensor is also largely insensitive to expected variation in flow speeds, but is sensitive to the
presence of flow vs. no flow. Fortunately, because of the redundancy of the measurements
with the other two sensors (pressure and flow) that are included within the multi-sensor
module, analysis that includes the measurements of all three sensors can minimize any
false-positives in impedance increase due to the absence of flow.

Figure 4-20. Patency curves measured for inline modules indicate that there is a decrease
in impedance of ~8% with the addition of flow, but no variation among flow rates within
the range of 0.03-0.6 ml/min.

176
4.4.7 Drift experiments
Long term drift experiments were carried out using the bottle brain system for two
weeks at room temperature to look for any prevailing trends in the baseline signal of the
devices. A 16-holed mock silicone catheter, placed inline with the sensor module using a
1/16” barb connector, was situated within the bottle brain along with a ground Pt wire
electrode. Flow out of the catheter situated in the bottle brain (20°C) was established using
a peristaltic pump set at 0.3 ml/min. Impedances were measured in 24 hour increments for
14 days. An experiment end time of 14 days was chosen to validate the sensor response
over the planned time for future clinical validation studies with the EVD system.
Impedance varied ±3.5% during the 14 days, largely due to the presence of bubbles within
the system on the third and fourth days (Figure 4-21). These results are promising, and this
minor variation compares favorably to the expected response due to obstruction, ~30%.
Future in vivo studies will confirm the stability of the sensor reading for chronic
applications.

Figure 4-21. Measured impedance of patency sensor for a 16-holed catheter in a bottle
brain within flow for 14 days demonstrating low drift of the sensor response.

4.4.8 Sterilization effects
H2O2 plasma sterilization of inline module packaged device was conducted at
177
Children’s Hospital Los Angeles (CHLA) per institutional protocols for sterilization of
their medical hardware using a Sterrad 100 NX system. Patency curves following
sterilization indicated that sensor performance remained unaltered (Figure 4-22). Electrode
characterization using electrochemical impedance spectroscopy and electrochemical
impedance spectroscopy (EIS; 1x PBS, 1-1,000,000 Hz, Ag/AgCl reference) and cyclic
voltammetry (CV; 1x PBS, -0.2 to 1.2V, scan rate of 250 mV/s for 30 cycles, Ag/AgCl
reference) also indicated no changes in electrode surface area or properties following
sterilization, demonstrating the compatibility of the sterilization process for Parylene-based
free film devices. External studies of sterilization efficacy (HIGHPOWER Validation
Testing & Lab Services Inc., Rochester, NY) validated the sterilization process of the inline
modules using the Sterrad system and approved the devices for use in the clinic.

Figure 4-22. Hydrogen peroxide plasma sterilization had no effect on the patency curve of
the inline module.

178

Figure 4-23. (a) Electrochemical impedance spectroscopy and (b) cyclic voltammetry
measurements of patency electrode packaged in inline module pre and post H 2O2 plasma
sterilization. Results indicate to changes to the electrode surface properties following
sterilization.

4.5 Benchtop testing
4.5.1 Dynamic obstruction studies
The ability of the sensor to measure dynamic blockages in the benchtop system was
assessed using three different benchtop models of dynamic blockage to mimic expected
obstruction phenomena in vivo. Initial dynamic blockage experiments were carried out by
pinching the catheter line to effectively obstruct the lumen of the catheter using an external
stylus probe (Figure 4-24). These experiments not only illustrated the dynamic response of
the sensor, but also demonstrated the sensor response to blockages of the lumen of the
catheter, one of the predominant causes for shunt obstructive failure. For this phenomenon,
the response of the sensor was found to model the 1/A of Equation 1 (due to changes in the
cross sectional area of the ionic conductive path), with the impedance changes increasing
as more of the lumen is obstructed (Figure 4-25).

179

Figure 4-24. Image of catheter pinching experiment to model obstruction of the lumen of
the catheter using a stylus.

Figure 4-25. (a) Observed sensor response to catheter deflection using a stylus to close off
the lumen. (b) Results indicate that the sensor response varies inversely due to changes in
the cross sectional area between the electrodes due to lumen obstruction.

For a more surface driven obstruction study, a 16-holed catheter was placed within
a beaker of aCSF, and transient blockages were simulated by sheathing/unsheathing the
drainage ports using larger diameter tubing while patency was continuously measured.
Results indicated that the sensor was capable of repeatedly measuring blockage events over
time (Figure 4-26). Time constants observed in the settling of the measured impedance is
due to the sheathing mechanics of the larger diameter tubing being placed over the 16 holes
to impede the conduction paths between the pair of electrodes.

180

Figure 4-26. (a) Transient blockage experiments of sheathing/unsheathing a 16-holed
catheter illustrated (b) real-time measurement capabilities of the sensors. Obstruction
events are labeled with an X.

Finally, a “reverse obstruction model” was explored using PEG to obstruct specific
holes of a 16-holed mock catheter, then tracking changes in patency during the dissolution
of the PEG within aCSF. This technique offered effective control over blockage
occurrences along the different holes of the catheters. In the first experiment, all holes of
the mock catheter except for the bottom-most hole were covered with a dip-coat of PEG
and was placed within a beaker of aCSF; a plot of the time varying patency signal as well
as corresponding images can be seen in Figure 4-27a. For this experiment, impedance
maintained a high magnitude (due to the PEG blockage of catheter holes) until the initial
dissolution of the top hole. Dissolution at the top-hole occurred first due to the dip coating
method of PEG, producing a thinner coating near the top, which was confirmed through
the images. This dissolution created a sharp decrease in the impedance magnitude (region
(ii)), until the PEG was fully dissolved at the top hole (region (iii)). Following this, the
181
impedance maintained a constant value even after all the holes were opened (region (iv)),
indicating the dependence of the sensor response on the patency of the first hole as
discussed in the previous sections. This result is further evidenced by a second experiment
in which only the top hole was covered with PEG and the rest were left open (Figure 4-27b).
A similar shape in the impedance response to this dissolution study illustrates that the top-
most hole is the most critical for the sensor response, and the remaining holes are nearly
nonexistent.

Figure 4-27. Dynamic obstruction study using PEG coated mock catheters in aCSF,
demonstrating tracking of PEG dissolution (“reverse blockage”) using sensor. (a) In the
first experiment, all holes of the catheter except for the bottom-most hole were covered
with PEG and sensor response was measured during PEG dissolution. (i) PEG coated
catheter is placed in aCSF, (ii) PEG dissolves from the top-most hole, (iii) top hole is open,
(iv) all holes are open. These results confirm the dependence of the sensor on the first hole.
(b) The second experiment revealed a similar response although only the top most hole was
covered with PEG.

182
4.5.2 Human CSF studies
Sensor characterization was also carried out in de-identified, discarded human CSF
obtained from patients from Los Angeles County and University of Southern California
(USC) Hospital. Human CSF was obtained using USC Institutional Review Board (IRB)
approved protocols and all experiments were conducted per USC Institutional Biosafety
Committee (IBC) approved procedures (IRB# HS-14-00608, USC IBC# BUA-15-00026).
In these experiments, patency curves were obtained with the human CSF, which was heated
to 37°C while stirred to maintain solution homogeneity. Qualitatively, the obtained human
CSF sample was comprised of CSF, blood components (due to subarachnoid hemorrhage),
as well as cellular and tissue debris. Baseline impedance measurements for a 16-holed
catheter was slightly higher than that observed in aCSF likely due to the increase
concentration of blood proteins and tissue particles. Sensor response was identical to
characterization of the sensor using aCSF, demonstrating a 35% impedance increase for
87.5% blockage (Figure 4-28a), which validated the ionic make-up of the aCSF used in
previous characterization experiments as an electrochemical model for CSF and is
promising in moving towards in vivo experiments in animals as well as within the clinic
with the EVDs. Short term drift studies (5 hours, no flow) indicated that there was a sharp
increase in impedance within the first 30 minutes, likely due to cooling of the CSF within
the inline module, which is positioned away from the heated solution, and a subsequent
slow decrease in impedance over time (Figure 4-28b). Because of the makeup of the human
CSF (traces of blood and tissue particles suspended in CSF), the decrease in impedance
can stem from a variety of sources. These results need to be further validated in in vivo
experiments.

183

Figure 4-28. (a) Patency curve for device S12 tested in human CSF solution indicating
preservation of performance. (b) 5 hr drift measurement of sensor response in configuration
for use with EVD systems indicating a slow downward drift of 5% over 5 hours. Additional
studies are required for further validation of sensor performance

4.6 Future studies
4.6.1 EVD validation studies
Moving forward, there is still an additional study that must be completed prior to
clinical implementation with the EVD package for initial device validation. Though the
sensor response within expected clinical environments and sterilization efficacy has been
proven, the devices still must be tested for compatibility with MRI imaging systems. A
MRI Conditional label must be obtained so that there are no large heating, image artifact,
or induced movement issues when the devices are placed within the high magnetic field
during patient imaging. It is expected that with the small amount of metal (non-
ferromagnetic) and the largely polymer composition of the packaged device, that the device
will be MRI compatible. However, testing must be carried out to validate this hypothesis.
Following this study, validation experiments with external ventricular drains will
be carried out in enrolled patients within hospitals in the Los Angeles area that require
EVDs for time periods of 2 days to 2 weeks. The module will be incorporated into the
standard of care EVD system, and will measure ICP, CSF flow, and shunt patency hourly
(Figure 4-29). For the electrode configuration of the patency sensor for the EVD, it is
184
envisioned that the ground electrode would be an externally placed on the skin. Data will
be compared with the current standard of care (a built in pressure sensor) as well as total
collected CSF volume for comparison. Sensors will collect data hourly using the
aforementioned data logger board and saved to a micro SD card until the EVD is removed.
Results from this study will help in providing validation data on the in vivo performance
of our device, and will aid in sensor design changes within the implantable package for use
with VP shunts.

Figure 4-29. (a) Cartoon schematic and (b) image of inline module packaged sensor for
integration with external ventricular drains (EVDs) within clinics.

4.6.2 Implantable package
Towards the development of an implantable package for the sensor, a SolidWorks
model was drawn to begin the design process and the initial design was 3D printed. Per
design advice given from Dr. Eisha Christian, the implantable module was designed to be
another isolated unit within the shunt system, without modifying current hardware to
encourage faster surgeon adoption. As observed in Figure 4-30, the module consists of two
parts: the body and the cap. The body is comprised of a main fluidic channel (1 mm
diameter) that allows for CSF flow from one barbed fitting (connection to catheter) to the
other barbed fitting (for connection to valve). A reservoir is built around the flow channel
to house the electronics for wireless power and data communication, as well as a slit within
185
the reservoir so that the Parylene device can be inserted into the flow channel while
maintaining connections to the electronics. Two cavities within the reservoir act as storage
areas to fit electronic parts as well as ferrite cores if necessary to improve inductive
coupling performance. An integrated lip around the top portion of the body is designed to
wind an inductive powering coil. The cap is designed to provide a smooth top to the body
of the module, while enclosing all of the electronics. It is also envisioned that a Pt disc
electrode will be integrated on the top or bottom of the implantable module to act as the
ground electrode for the patency sensor. A combination of epoxy filled encapsulation
within the reservoir of the module and subsequent Parylene C coating will aid in improving
long-term performance within the body. Following sensor characterization within the
implanted module, in vivo validation studies within a rat model will provide data on the
chronic implanted performance of the sensors.

186

Figure 4-30. (a) 3D model of implantable packaging envisioned for the sensor for
placement in between the proximal catheter and valve. (b) Cross sectional view of the
implantable packaging. (c) Images of 3D printed implantable module to scale for initial
testing.

4.7 Summary
In summary, the Parylene-based patency sensor provides an integrative solution to
directly monitor shunt patency over the treatment duration. Insensitivity to both clinically
relevant temperatures and flow conditions are advantageous to in vivo performance. The
sensing method itself is such a simple technique to assess shunt dynamics and efficacy, and
with the proper integration strategies, can be applied to a variety of shunt-based devices
beyond those used for hydrocephalus. The clarification of sensor mechanism as a
measurement more of the patency of the closest hole to the electrode is also key in
demonstrating this type of “least path of resistance” effect for electrochemical impedance,
which can possibly be leveraged in the development of future impedance-based sensors.
187
There is still additional work to be done prior to clinical implementation of these
devices, namely the MRI compatibility and in vivo animal model studies, but the initial
benchtop results are promising. The compatibility of these packaged devices with the H2O2
sterilization process and definite confirmation of sterilization via rigorous tests not only
approve our devices for use within the clinic, but also affirms the procedure for implantable
Parylene devices as a whole without affecting device properties. Additionally, the low drift
measured with these sensors over a 14 day period is encouraging on reliable performance
for these clinical studies. The positive results obtained with experiments with human CSF
gives further assurance of the in vivo potential of the device.
Moreover, the concept of a multi-sensor device that integrates patency, pressure,
and flow sensing into a single module is one step towards the aim of a “smart shunt”
technology, which “[s]hort of a cure, … would be one of the most exciting and impactful
developments in the treatment of hydrocephalus,” according to Dr. Lutz of the University
of Washington, a leading expert in the development of smart hydrocephalus shunts [28]. It
is the author’s hope that with the current development of this Parylene-based technology,
the multi-sensor module will provide a sense of fortified optimism and confidence for
neurosurgeons and patients alike in the treatment of hydrocephalus, and possibly even for
the first time demonstrate the first commercialized Parylene-based sensor device within the
clinic.

4.8 References
[1] Hydrocephalus Association. Classifications and Causes. Available:
http://www.hydroassoc.org/classifications-and-causes/
[2] I. K. Pople, "Hydrocephalus and shunts: what the neurologist should know,"
Journal of Neurology, Neurosurgery & Psychiatry, vol. 73, pp. i17-i22, 2002.
[3] S. Jetzki, et al., "A multisensor implant for continuous monitoring of intracranial
pressure dynamics," Ieee Transactions on Biomedical Circuits and Systems, vol. 6, pp.
356-365, 2012.
188
[4] S. Hakim, et al., "The physics of the cranial cavity, hydrocephalus and normal
pressure hydrocephalus: mechanical interpretation and mathematical model," Surgical
Neurology, vol. 5, pp. 187-210, 1976.
[5] Hydrocephalus Association. Normal Pressure Hydrocephalus. Available:
http://www.hydroassoc.org/normal-pressure-hydrocephalus/
[6] A. Farin, et al., "Endoscopic third ventriculostomy," Journal of Clinical
Neuroscience, vol. 13, pp. 763-770, 2006.
[7] J. S. Baru, et al., "John Holter’s shunt," Journal of the American College of
Surgeons, vol. 192, pp. 79-85, 2001.
[8] International Society for Pediatric Neurosurgery. Medtronic PS Medical Delta
Valves. Available:
http://ispn.guide/book/Parking%20Lot/The%20ISPN%20Shunt%20Manual/medtronic-
ps-medical-deltasup%C2%AEsup-valves
[9] J. Drake, et al., "CSF shunts 50 years on–past, present and future," Child's Nervous
System, vol. 16, pp. 800-804, 2000.
[10] S. R. Browd, et al., "Failure of cerebrospinal fluid shunts: part I: obstruction and
mechanical failure," Pediatric neurology, vol. 34, pp. 83-92, 2006.
[11] A. V. Kulkarni, et al., "Cerebrospinal fluid shunt infection: a prospective study of
risk factors," Journal of Neurosurgery, vol. 94, pp. 195-201, 2001.
[12] J. M. Drake, et al., "Randomized trial of cerebrospinal fluid shunt valve design in
pediatric hydrocephalus," Neurosurgery, vol. 43, pp. 294-303, 1998.
[13] J. Kestle, et al., "Long-term follow-up data from the Shunt Design Trial," Pediatric
neurosurgery, vol. 33, pp. 230-236, November 2000 2000.
[14] S. R. Browd, et al., "Failure of Cerebrospinal Fluid Shunts: Part I: Obstruction and
Mechanical Failure," Pediatric neurology, vol. 34, pp. 83-92, 2006.
[15] E. J. Haberl, et al., "Experiences with a gravity-assisted valve in hydrocephalic
children: Clinical article," Journal of Neurosurgery: Pediatrics, vol. 4, pp. 288-293, 2009.
[16] H. L. Brydon, et al., "Protein adsorption to hydrocephalus shunt catheters: CSF
protein adsorption," Journal of Neurology, Neurosurgery & Psychiatry, vol. 64, pp. 643-
647, 1998.
189
[17] R. D. Dickerman, et al., "Failure rate of frontal versus parietal approaches for
proximal catheter placement in ventriculoperitoneal shunts: revisited," Journal of Clinical
Neuroscience, vol. 12, pp. 781-783, 2005.
[18] U. W. Thomale, et al., "Perforation holes in ventricular catheters—is less more?,"
Child's Nervous System, vol. 26, pp. 781-789, 2010.
[19] C. A. Harris and J. P. McAllister, "What we should know about the cellular and
tissue response causing catheter obstruction in the treatment of hydrocephalus,"
Neurosurgery, vol. 70, pp. 1589-1602, 2012.
[20] C. Blegvad, et al., "Pathophysiology of shunt dysfunction in shunt treated
hydrocephalus," Acta Neurochirurgica, vol. 155, pp. 1763-1772, 2013.
[21] F. Lundberg, et al., "Presence of vitronectin and activated complement factor C9
on ventriculoperitoneal shunts and temporary ventricular drainage catheters," Journal of
Neurosurgery, vol. 90, pp. 101-108, 1999.
[22] C. A. Harris, et al., "Mechanical contributions to astrocyte adhesion using a novel
in vitro model of catheter obstruction," Experimental Neurology, vol. 222, pp. 204-210,
2010.
[23] C. A. Harris and J. P. McAllister II, "Does drainage hole size influence adhesion
on ventricular catheters?," Child's Nervous System, vol. 27, pp. 1221-1232, 2011.
[24] B. A. Kaufman and T. Park, "Ventricular anatomy and shunt catheters," Pediatric
neurosurgery, vol. 31, pp. 1-6, 1999.
[25] S. Tuli, et al., "Change in ventricular size and effect of ventricular catheter
placement in pediatric patients with shunted hydrocephalus," Neurosurgery, vol. 45, p.
1329, 1999.
[26] J. Lin, et al., "Computational and experimental study of proximal flow in
ventricular catheters: Technical note," Journal of Neurosurgery, vol. 99, pp. 426-431, 2003.
[27] J. M. Wong, et al., "Patterns in neurosurgical adverse events: cerebrospinal fluid
shunt surgery," Neurosurgical Focus, vol. 33, p. E13, 2012.
[28] B. R. Lutz, et al., "New and improved ways to treat hydrocephalus: Pursuit of a
smart shunt," Surg Neurol Int, vol. 4, p. S38, 2013.
[29] J. S. Kroin, et al., "Long-term testing of an intracranial pressure monitoring
device," Journal of Neurosurgery, vol. 93, pp. 852-858, 2000.
190
[30] H. J. Yoon, et al., "Micro devices for a cerebrospinal fluid (CSF) shunt system,"
Sensors and Actuators A: Physical, vol. 110, pp. 68-76, 2004.
[31] A. Ginggen, et al., "A telemetric pressure sensor system for biomedical
applications," Biomedical Engineering, IEEE Transactions on, vol. 55, pp. 1374-1381,
2008.
[32] S. Neff, "Measurement of flow of cerebrospinal fluid in shunts by transcutaneous
thermal convection: Technical note," Journal of Neurosurgery: Pediatrics, vol. 103, pp.
366-373, 2005.
[33] T. Bork, et al., "Development and in-vitro characterization of an implantable flow
sensing transducer for hydrocephalus," Biomedical Microdevices, vol. 12, pp. 607-618,
2010.
[34] J. Burger, et al., "In-vitro characterization of an implantable thermal flow sensor
for hydrocephalus," in World Congress on Medical Physics and Biomedical Engineering,
September 7-12, 2009, Munich, Germany, 2010, pp. 265-268.
[35] T. Clark, et al., "Towards diagnostic flow measurement from a catheter tip pressure
sensor," in Biomedical Circuits and Systems Conference (BioCAS), 2014 IEEE, 2014, pp.
356-359.
[36] S. Rajasekaran, et al., "Mechanism for measurement of flow rate of cerebrospinal
fluid in hydrocephalus shunts," in Engineering in Medicine and Biology Society (EMBC),
2014 36th Annual International Conference of the IEEE, 2014, pp. 2153-2156.
[37] P. K. Eide and A. Bakken, "The baseline pressure of intracranial pressure (ICP)
sensors can be altered by electrostatic discharges," Biomedical engineering online, vol. 10,
p. 75, 2011.
[38] I. M. Elixmann, et al., "Case study of relevant pressures for an implanted
hydrocephalus valve in everyday life," in Engineering in Medicine and Biology Society
(EMBC), 2012 Annual International Conference of the IEEE, 2012, pp. 1635-1638.
[39] M. A. Poca, et al., "Prospective study of methodological issues in intracranial
pressure monitoring in patients with hydrocephalus," Journal of Neurosurgery, vol. 100,
pp. 260-265, 2004.
[40] I. Elixmann, et al., "Single pulse analysis of intracranial pressure for a
hydrocephalus implant," in Engineering in Medicine and Biology Society (EMBC), 2012
Annual International Conference of the IEEE, 2012, pp. 3939-3942.
191
[41] S. S. Basati, et al., "Dynamic brain phantom for intracranial volume
measurements," Biomedical Engineering, IEEE Transactions on, vol. 58, pp. 1450-1455,
2011.
[42] S. Basati, et al., "Cerebrospinal fluid volume measurements in hydrocephalic rats:
Laboratory investigation," Journal of Neurosurgery: Pediatrics, vol. 10, pp. 347-354, 2012.
[43] I. M. Elixmann, et al., "In-Vitro Evaluation of a Drainage Catheter with Integrated
Bioimpedance Electrodes to Determine Ventricular Size," Biomedical
Engineering/Biomedizinische Technik, 2013.
[44] I. Elixmann, et al., "Control of an Electromechanical Hydrocephalus Shunt-a New
Approach," 2014.
[45] S. Basati, et al., "Impedance Changes Indicate Proximal Ventriculoperitoneal Shunt
Obstruction In-vitro," 2014.
[46] L. Yu and E. Meng, "A microbubble pressure transducer with bubble nucleation
core," in Micro Electro Mechanical Systems (MEMS), 2014 IEEE 27th International
Conference on, 2014, pp. 104-107.
[47] L. Yu, et al., "An implantable time of flight flow sensor," in Micro Electro
Mechanical Systems (MEMS), 2015 28th IEEE International Conference on, 2015, pp.
620-623.
[48] B. J. Kim, et al., "MEMS electrochemical patency sensor for detection of
hydrocephalus shunt obstruction," in Micro Electro Mechanical Systems (MEMS), 2015
28th IEEE International Conference on, 2015, pp. 662-665.
[49] C. A. Gutierrez and E. Meng, "Impedance-Based Force Transduction Within Fluid-
Filled Parylene Microstructures," Journal of Microelectromechanical Systems, vol. 20, pp.
1098-1108, Oct 2011.
[50] B. J. Kim, et al., "Parylene-based electrochemical-MEMS force sensor array for
assessing neural probe insertion mechanics," in Micro Electro Mechanical Systems
(MEMS), 2012 IEEE 25th International Conference on, 2012, pp. 124-127.
[51] R. Sheybani, et al., "Design, fabrication, and characterization of an
electrochemically-based dose tracking system for closed-loop drug delivery," in
Engineering in Medicine and Biology Society (EMBC), 2012 Annual International
Conference of the IEEE, 2012, pp. 519-522.
192
[52] Y. Nam, "Material considerations for in vitro neural interface technology," MRS
bulletin, vol. 37, pp. 566-572, 2012.
[53] C. A. Gutierrez, et al., "Epoxy-less packaging methods for electrical contact to
parylene-based flat flexible cables," in Solid-State Sensors, Actuators and Microsystems
Conference (TRANSDUCERS), 2011 16th International, 2011, pp. 2299-2302.
[54] Personal Communication, E. Christian. 2015
[55] C. A. Harris, et al., "Effects of surface wettability, flow, and protein concentration
on macrophage and astrocyte adhesion in an in vitro model of central nervous system
catheter obstruction," Journal of Biomedical Materials Research Part A, vol. 97, pp. 433-
440, 2011.
[56] J. J. Barron and C. Ashton, "The effect of temperature on conductivity
measurement," TSP, vol. 7, 2005.
[57] H. E. Ayliffe and R. D. Rabbitt, "An electric impedance based
microelectromechanical system flow sensor for ionic solutions," Measurement science and
Technology, vol. 14, p. 1321, 2003.

193

“Knowing is not enough, we must apply. Willing is not enough, we must do.” – Bruce Lee

As the research community endeavors towards the next generation of implantable
devices, Parylene-based efforts are slowly progressing into the limelight. Current implant
technologies rely on complex packaging to maintain hermeticity and are often constructed
on rigid materials with the hope of some integration with the body’s soft tissues. As an
improved paradigm, the flexible and biocompatible qualities of Parylene highlight the
suitability of the material for better in vivo integration not only within the tissues of the
body, but also as an add-on component for implanted hardware for supplementary
functionality. The compatibility of Parylene with micromachining processes also
demonstrates the capacity of the material as a substrate for precisely designed and batch-
fabricated microdevices. Specifically for the two cortical applications described within this
work, Parylene-based implantable devices are exceptional solutions for the needs within
intracortical electrode and hydrocephalus treatment research.
In chapter 2, a Parylene-based sensor was presented that aimed to improve the
reliability of implanted intracortical electrodes through a data-driven approach. The
flexibility and low-profile nature that comes with the use of a Parylene substrate was key
in the integration of these sensors onto ceramic probes to measure interfacial mechanics
during insertion into agarose brain phantoms. For the first time, forces encountered along
the length of the probe were measured in situ, quantitatively confirming studies that were
hypothesized via modeling efforts that the majority of interfacial forces are within the first
1 mm of the probe. Additional studies with sensor integrated neural probes would help

CONCLUSION
194
establish design parameters for probe size, shape, and electrode placement for more
improved solutions for chronically implanted intracortical electrodes.
In chapter 3, a secondary approach in improving the reliability of chronically
implanted cortical electrodes was presented, in the development of an electrode technology
constructed on the flexible Parylene substrate. As a solution to the mechanical mismatch
between the state-of-the-art, rigid metal or silicon probes and the surrounding neural tissue,
a Parylene-based approach was demonstrated to have reliable performance for up to a year
implanted within rat models. In addition, the complementary design strategies of a 3D
conical structure for neuronal ingrowth as well as the use of neurotrophic and anti-
inflammatory coatings validated a multi-prong approach to further improve device
performance. The thermoforming of planar Parylene devices to construct 3D structures was
also demonstrated, and the effects of the process on the material and chemical properties
of the polymer were elucidated to present a new approach in the field of constructing 3D
Parylene devices. To facilitate the development of Parylene-based devices as a whole,
electrical packaging and implant strategies was also described to create a basis for
implementation of these flexible devices for in vivo applications.
Hydrocephalus is a condition that affects 1 in 500 of children born today, but
research in improved treatment technologies has largely been stagnant over the past 50
years. In chapter 4, a Parylene-based patency sensor designed to integrate with shunt
hardware was presented. The sensor adds continuous monitoring functionality to current
treatment methods to assess shunt dynamics and treatment efficacy. As the sensor utilizes
a simple impedance-based transduction method and is built on a Parylene substrate, the
device is easy to integrate with both external and implanted shunt systems. Sensor
characterization at physiologically relevant temperature and flow conditions not only
elucidated sensor responses in vivo, but also expounded these effects on any Parylene-
based impedance sensor integrated within a fluidic channel. Clarification of the sensing
mechanism and determination of a “least resistive path” effect for this sensor also marked
a sensing paradigm that may be useful for future devices. Experiments conducted with
195
hydrogen peroxide plasma sterilization of the devices indicated that the process is
compatible with Parylene-based devices, which is noteworthy as many clinical devices in
hospitals are sterilized using this method. Overall, this work set the groundwork for clinical
packaging and testing for this Parylene-based device that can be extended to any device for
use within the clinic.
To conclude, the author hopes that the previous chapters detailing the development
of implantable Parylene-based devices for cortical applications have convinced the reader
of the potential and suitability of Parylene as a substrate for implantable devices. There is
still much work to be done however, and the maturation of Parylene as a bioMEMS
material is a growing and integrative effort among the MEMS community. It is the author’s
vision and hope that the use of Parylene as a substrate material for implantable devices
would extend beyond academic laboratories and towards clinical implementation. For, as
Dr. Christian Gutierrez wrote, “…it is through its impact on society that the true measure
of a technology’s innovation is made,” and Parylene is more than capable of filling the
gaps of today’s insufficient technologies to impact society.

196
APPENDIX A: INTRACORTICAL FORCE SENSOR PROCESS FLOW

1. Bake clean 3” silicon wafer to remove moisture 140 °C, > 10 mins

2. Deposit Parylene (5 m)

3. Pattern AZ 5214-IR for lift-off ( 2 m thick) (Cortical array – IR mask 1)
Pre spin 5 sec, 500 rpm
Spin 45 sec, 2 krpm
Softbake 90 °C , 1:10 min
Exposure 50 mJ/cm
2
(10 mW/cm
2
, 5 sec)
IR bake 120 °C , 45 sec
Global exposure 300 mJ/cm
2
(10 mW/cm
2
, 30 seconds)
Development (AZ 351 1:4 dilution) 20-22 seconds

4. Descum, O 2 plasma 100 W, 100 mTorr, 1 min

5. Metal deposition (Pt) 2000 Å (in 3 runs of 666 Å )

6. Lift-off in acetone (gentle scrub if necessary)

7. Descum, O 2 plasma 100 W, 100 mTorr, 1 min

8. Deposit Parylene (2 m)

9. Pattern AZ 4400 etch mask ( 4.2 4.3 m thick) (Cortical array – mask 2)
Pre spin 5 sec, 500 rpm
Spin 45 sec, 4 krpm
Softbake 90 °C , 2 min
Exposure 150 mJ/cm
2
(10 mW/cm
2
, 15 sec)
Development (AZ 351 1:4 dilution) 45-50 seconds

10. Reactive Ion Etching (Oxygen plasma)
100 W, 100mTorr, 5 mins

11. Clean Acetone, IPA DI water

12. Pattern AZ 4620 (double layer) sacrificial channel mold (19.5-21 m thick) (Cortical array –
mask 3)
LAYER 1
Pre spin 5 sec, 500 rpm
Spin 45sec, 2.25 krpm
Softbake 90 °C, 6 minutes
LAYER 2
197
Pre spin 5 sec, 500 rpm
Spin 45 sec, 2.25 rpm
Softbake 90 °C, 7 minutes
Exposure 600 mJ/cm
2
(10 mW/cm
2
, 60 sec)
Development 1:45 - 2 minutes

13. Deposit Parylene (5 m)

14. Pattern AZ 4620 etch mask (11-12 m thick) (Channel ports and contact pad etch) (Cortical
array – mask 4)
Pre spin 5 sec, 500 rpm
Spin 45 sec, 2 krpm
Softbake 90 °C , 6 min
Exposure 550 mJ/cm
2
(10 mW/cm
2
, 55 sec)
Development (AZ 351 1:4 dilution) 1:10 min

15. Reactive Ion Etching (Oxygen plasma)
100 W, 100mTorr, 5 mins
Rotate 3x

16. Pattern AZ 4620 etch mask ( 11-12 m thick) – Spin on top of remaining mask (Cortical
array – mask 5)
Pre spin 5 sec, 500 rpm
Spin 45 sec, 2 krpm
Softbake 90 °C , 6 min
Exposure 550 mJ/cm
2
(10 mW/cm
2
, 55 sec)
Development (AZ 351 1:4 dilution) 1:15 seconds

17. Reactive Ion Etching (Oxygen plasma)
100 W, 100mTorr, 5 mins
Rotate 3x

Release Clean surface with Q-tip and acetone.
Peel carefully while immersed in water

198
APPENDIX B: PARYLENE SHEATH ELECTRODE ARRAY PROCESS
FLOW

1. Dehydration bake 4” Si wafer to remove moisture 140 °C, > 10 mins

2. Deposit Parylene (~5 m)

3. Pattern AZ 5214-IR for lift-off (2 m thick) (Mask 1)
Pre spin 5 sec, 500 rpm
Spin 45 sec, 2 krpm
Softbake 90 °C, 1:10 min
Exposure 50 mJ/cm
2
(10 mW/cm
2
, 5 sec)
IR bake 120 °C, 45 sec
Global exposure 300 mJ/cm
2
(10 mW/cm
2
, 30 sec)
Development (AZ 351 1:4 dilution) 20-22 sec

4. Descum, O 2 plasma 100 W, 100 mTorr, 1 min

5. Metal deposition (Pt) 2000 Å (in 3 runs of 666 Å )
6. Lift-off in acetone (gentle scrub)

Base Perforation Etch
7. Pattern AZ 4620 etch mask (10 µm thick) (Mask 2)
Pre spin 5 sec, 500 rpm
Spin 45 sec, 2 krpm
Softbake 90 °C, 5 min
Exposure 400 mJ/cm
2
(10 mW/cm
2
, 40 sec)
Development (AZ 351 1:4 dilution) 50–60 sec

8. Reactive Ion Etching (Oxygen plasma)
100 W, 100mTorr, 5 mins
Rotate 5x
9. Wash off PR mask with triple acetone & IPA bath with cotton swabbing

10. Descum, O 2 plasma RIE 100 W, 100 mTorr, 1 min

11. Deposit Parylene (2 m)

Insulation Layer Etch
12. Pattern AZ 4620 etch mask (10 µm thick) (Mask 3)
Pre spin 5 sec, 500 rpm
Spin 45 sec, 2 krpm
199
Softbake 90 °C, 5 min
Exposure 400 mJ/cm
2
(10 mW/cm
2
, 40 sec)
Development (AZ 351 1:4 dilution) 50–60 sec

13. Reactive Ion Etching (Oxygen plasma)
100 W, 100mTorr, 5 mins
Rotate 3x

14. Wash off PR mask with triple acetone & IPA bath with cotton swabbing

Sacrificial PR for cone
15. Pattern AZ 4620 etch mask (10 µm thick) (Mask 4)
Pre spin 5 sec, 500 rpm
Spin 45 sec, 2 krpm
Softbake 90 °C, 5 min
Exposure 400 mJ/cm
2
(10 mW/cm
2
, 40 sec)
Development (AZ 351 1:4 dilution) 50–60 sec

16. Descum, O 2 plasma 100 W, 100 mTorr, 1 min

17. Deposit Parylene (5 m)

Channel Release/Partial Cutout Etch
18. Pattern AZ 4620 etch mask (10 µm thick) (Mask 5)
Pre spin 5 sec, 500 rpm
Spin 45 sec, 2 krpm
Softbake 90 °C, 5 min
Exposure 400 mJ/cm
2
(10 mW/cm
2
, 40 sec)
Development (AZ 351 1:4 dilution) 50–60 sec

19. Reactive Ion Etching (Oxygen plasma)
100 W, 100mTorr, 5 mins
Rotate 5x

DO NOT CLEAN OFF MASK

Cutout Etch
20. Pattern AZ 4620 etch mask on top of previous PR (2x10 µm thick) (Mask 6)
Pre spin 5 sec, 500 rpm
Spin 45 sec, 2 krpm
Softbake 90 °C, 5 min
Pre spin 5 sec, 500 rpm
Spin 45 sec, 2 krpm
200
Softbake 90 °C, 6 min
Exposure 540 mJ/cm
2
(10 mW/cm
2
, 54 sec)
Development (AZ 351 1:4 dilution) 50–60 sec

21. Reactive Ion Etching (Oxygen plasma)
100 W, 100mTorr, 9 mins
Rotate 4x

Release Clean surface with Q-tip and acetone.
Peel carefully while immersed in water

Thermoforming
1. Make sure all probes are clean: Acetone/IPA/DI H 2O spray down
2. Place PSE devices on a glass slide and insert the microwires into the PSE sheath using the
stainless steel or tungsten microwires under the stereoscope

3. Place samples in vacuum oven
4. Turn vent off completely
5. Turn vacuum on slowly until it reaches 0 inHg

[Do steps 6-11 if doing nitrogen backflow]
6. Turn vacuum off
7. Open the gas knob on the oven
8. Set flow of the N 2 to 10 SCFH (standard cubic feet/hr)
9. Wait until N 2 fills the entire chamber
10. Turn vacuum on again until reach 0 inHg
11. Repeat steps 4-7 to repeat the N 2 chamber flush

12. Press EZ1 program (Ramp to 200°C, hold for 48 hrs, return to 20°C)
13. Following end of process, wait overnight (~15 hrs) for samples to cool down

14. Remove samples carefully
15. Remove the microwires carefully and peel the devices off the glass slide. Immersion in DI
water can aid in process
16. Dry devices

201
APPENDIX C: PARYLENE-BASED PATENCY SENSOR PROCESS
FLOW

1. Dehydration bake 3” Si wafer to remove moisture 140 °C, > 10 mins

2. Deposit Parylene (~12 m)

3. Pattern AZ 5214-IR for lift-off (2 m thick) (Mask 1)
Pre spin 5 sec, 500 rpm
Spin 45 sec, 2 krpm
Softbake 90 °C, 1:10 min
Exposure 50 mJ/cm
2
(10 mW/cm
2
, 5 sec)
IR bake 120 °C, 45 sec
Global exposure 300 mJ/cm
2
(10 mW/cm
2
, 30 sec)
Development (AZ 351 1:4 dilution) 20-22 sec

4. Descum, O 2 plasma 100 W, 100 mTorr, 1 min

5. Metal deposition (Pt) 2000 Å (in 3 runs of 666 Å )

6. Lift-off in acetone (gentle scrub)

7. Descum, O 2 plasma RIE 100 W, 100 mTorr, 1 min

8. Deposit Parylene (12 m)

Insulation Layer Etch
9. Pattern AZ 4620 etch mask (15 µm thick) (Mask 5)
Pre spin 5 sec, 500 rpm
Spin 45 sec, 1.2 krpm
Softbake 115 °C, 4 min
Exposure 550 mJ/cm
2
(10 mW/cm
2
, 55 sec)
Development (AZ 351 1:4 dilution) 90 sec
10.
11. Reactive Ion Etching (Oxygen plasma)
100 W, 100mTorr, 5 mins
Rotate 12x

12. Wash off PR mask with triple acetone & IPA bath with cotton swabbing

Sacrificial PR for pressure sensor
13. Pattern AZ 4620 etch mask (8 µm thick)
Pre spin 5 sec, 500 rpm
202
Spin 45 sec, 3.5 krpm
Softbake 110 °C, 3 min
Exposure 400 mJ/cm
2
(10 mW/cm
2
, 40 sec)
Development (AZ 351 1:4 dilution) 80 sec

14. Descum, O 2 plasma 100 W, 100 mTorr, 1 min

15. Deposit Parylene (4 m)

Pressure Sensor Channel Opening/Partial Cutout Etch
16. Pattern AZ 4620 etch mask (15 µm thick) (Mask 5)
Pre spin 5 sec, 500 rpm
Spin 45 sec, 1.2 krpm
Softbake 115 °C, 4 min
Exposure 550 mJ/cm
2
(10 mW/cm
2
, 55 sec)
Development (AZ 351 1:4 dilution) 90 sec

17. Reactive Ion Etching (Oxygen plasma)
100 W, 100mTorr, 5 mins
Rotate 4x

DO NOT CLEAN OFF MASK

Cutout Etch
18. Pattern AZ 4620 etch mask on top of previous PR (15 µm thick) (Mask 6)
Pre spin 5 sec, 500 rpm
Spin 45 sec, 1.2 krpm
Softbake 115 °C, 4 min
Exposure 550 mJ/cm
2
(10 mW/cm
2
, 55 sec)
Development (AZ 351 1:4 dilution) 90 sec

19. Reactive Ion Etching (Oxygen plasma)
100 W, 100mTorr, 5 mins
Rotate 24x

Release Clean surface with Q-tip and acetone.
Peel carefully while immersed in water

203
APPENDIX D: PARYLENE DEVICE RELEASE AND PACKAGING
FOR ZIF CONNECTOR

1. First clean off the top mask of PR that was used for the cut out etch using a cleanroom grade
q-tip and acetone

2. Immediately submerge the wafer in a large petri dish filled with DI water to allow for release
of the devices off the wafer

3. Use a pair of fine tweezers to work non-signficant corners to encourage the release off the
wafer. The devices should lift off the Si wafer easily

4. Place released devices on a TexWipe for them to dry

[Do steps 5-8 if sacrificial PR structures are present]
5. Place the devices in a glass petri dish with acetone to remove the sacrificial photoresist. Soak
for 10 minutes (or however long to remove the PR)

6. Move the devices to another glass petri dish with IPA. Soak for 10 minutes

7. Move the devices to another petri dish with DI water. Soak the probes for 5 minutes

8. Dry the probes by putting them on a TexWipe

9. For each device, glue a ~254 µm thick PEEK backing to the connector end using cyanoacrylate
(Krazy Glue) and cut the PEEK backing with scissors to match the outline of the Parylene cable

Abstract (if available)

Abstract The development of implantable devices for biomedical applications has been heavily galvanized by the introduction of microelectromechanical systems (or MEMS) technologies. Stemming from efforts within the semiconductor industry, MEMS enables the miniaturization of devices while maintaining, and in some cases, improving their functionality. MEMS are ideal for implantable applications, as smaller devices can reduce the foreign body response once implanted and allow for better integration within the body. However, though these micro-devices are a significant improvement over previous implant technologies, there is still much room for improvement. ❧ Current implants are marred by rigid substrates such as metals, ceramics, and silicon that create a mechanical mismatch between the implant and the surrounding tissue and shortens the implant lifetime. In addition, typical MEMS devices are sensitive to wet, saline in vivo environments due to the materials used, which can affect chronic performance within the body. The following work centers on the development of Parylene C-based technologies as an improved paradigm for implantable devices due to their superior material properties and integratable capabilities. Parylene C is a USP class VI polymer (highest possible rating of biocompatibility for plastics) and is compatible with micromachining processes, which enables the construction of MEMS devices using a Parylene C thin film substrate and architecture. In addition, Parylene C’s low Young’s modulus (~2-3 GPa) and thus flexibility compared to silicon or metal allows for improved tissue integration as well as simple integration with current implant hardware to add functionality. ❧ The following work details Parylene C-based approaches to filling the needs for two thrusts within the cortical research field: namely, improving chronically implanted intracortical electrode technologies for neural prosthetics and improving current hydrocephalus treatment. Chapters 2 and 3 delineate two Parylene C-based approaches in improving the chronic reliability of implanted intracortical electrode technologies. A Parylene C-based force sensor array is demonstrated in chapter 2 that leverages the flexibility of Parylene C for use as a deformable membrane to measure tissue forces enacting on the implanted electrode. The thin-film and low-profile nature of the sensor array also allows for easy integration of the sensor array on the face of these intracortical probes to assess implantation and in vivo mechanics once implanted. These sensors will enable, for the first time, quantitative measurement of interactional forces between the probe and tissue in situ. It is the hope that studies with these instrumented sensors will help establish design parameters for probe size, shape, and electrode placement for more improved solutions for chronically implanted intracortical electrodes. ❧ In chapter 3, a secondary approach in improving the reliability of chronically implanted cortical electrodes is presented, in the development of an electrode technology fabricated on a Parylene C substrate. As a solution to the mechanical mismatch between the state-of-the-art, rigid metal or silicon probes and the surrounding neural tissue, a flexible, thin-film Parylene-based approach is presented and implemented in rat models to assess chronic performance. In the development of this device, a thermoforming process is introduced where three dimensional (3D) structures are constructed from planar Parylene devices using thermal shaping. The effects of the process on the material and chemical properties of the polymer are also presented to further characterize this post-process as a new technique in establishing 3D Parylene structures. Electrical packaging and implant strategies are also discussed to create a basis for implementation of these flexible devices for in vivo applications. ❧ Chapter 4 discusses a Parylene C-based approach in improving current hydrocephalus treatment. Hydrocephalus is a chronic condition of the excess accumulation of fluid within the brain, usually treated by implanting a catheter system (“shunt”) that is marred by a failure rate of 40% within the first year due to catheter obstruction. Though failure rates are high, there are no current practical methods to assess shunt patency

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Strategies for improving mechanical and biochemical interfaces between medical implants and tissue

PDF

The electrochemical evaluation of Parylene-based electrodes for neural applications

PDF

Parylene-based implantable interfaces for biomedical applications

PDF

Parylene C bioMEMS for implantable devices with electrochemical interfaces

PDF

Implantable bioMEMS drug delivery systems

PDF

Development of fabrication technologies for robust Parylene medical implants

PDF

Development of micromachined technologies for neural interfaces

PDF

A wireless implantable MEMS micropump system for site-specific anti-cancer drug delivery

PDF

Parylene-based biomems sensors for multiple physiological systems

PDF

Thin-film impedimetric sensors for chronic in vivo use: design and application to hydrocephalus treatment

PDF

Penetrating parylene neural probe array for dense, in vivo, chronic recordings

PDF

Fabrication and packaging of three-dimensional Parylene C neural interfaces

PDF

Integrated wireless piezoelectric ultrasonic transducer system for biomedical applications

PDF

High-resolution data acquisition with neural and dermal interfaces

PDF

Multi-region recordings from the hippocampus with conformal multi-electrode arrays

PDF

Electronics design and in vivo evaluation of a wirelessly rechargeable fetal micropacemaker

PDF

Piezoelectric ultrasonic and acoustic microelectromechanical systems (MEMS) for biomedical, manipulation, and actuation applications

PDF

Liquid flow rate sensing via solution resistance for in vivo applications

PDF

A percutaneously implantable wireless neurostimulator for treatment of stress urinary incontinence

PDF

Scale-up of vapor-phase deposition of polymers: towards large-scale processing

Asset Metadata

Creator Kim, Brian Jung (author)

Core Title Development of implantable Parylene-based MEMS technologies for cortical applications

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Biomedical Engineering

Publication Date 07/21/2016

Defense Date 05/21/2015

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag bioMEMS,implants,MEMS,microelectrodes,micromachining,OAI-PMH Harvest,Parylene C,polymers,sensors

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Meng, Ellis (committee chair), Gupta, Malancha (committee member), Weiland, James D. (committee member)

Creator Email bjungkim@gmail.com,brianjk@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-602341

Unique identifier UC11300592

Identifier etd-KimBrianJu-3624.pdf (filename),usctheses-c3-602341 (legacy record id)

Legacy Identifier etd-KimBrianJu-3624.pdf

Dmrecord 602341

Document Type Dissertation

Format application/pdf (imt)

Rights Kim, Brian Jung

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA