Mater. Res. Soc. Symp. Proc. Vol. 1569 © 2013 Materials Research Society
D
O
I
:
1
5
5
7
/
o
p
0
1
3
0
.
1
l
.
2
.
Feasibility Study of Carbon Nanotube Microneedles for Rapid Transdermal Drug Delivery
Bradley J. Lyon
1
, Adrianus I. Aria
1
, and Morteza Gharib
1
1
Graduate Aerospace Laboratories, California Institute of Technology, Pasadena, CA 91125,
U.S.A.
ABSTRACT
We i
ntroduce a new approach for fabricating hollow microneedles using
vertically
-
aligned
carbon nanotubes (
VA
-
CNT
s
) for rapid transdermal drug delivery. Here, we discuss the
fabrication of the
microneedles
emphasizing the overall simplicity and flexibility of
the method
to allow for potential industrial application
. By capitalizing on the nanoporosity of the CNT
bundles, uncured polymer
can be wicked into the needles
ultimately
creating a high strength
composite of aligned nanotubes and polymer. Flow through th
e microneedles as well as
i
n vitro
penetration of the
microneedles
in
to
swine skin is demonstrated.
Furthe
rmore, we present a trade
study
comparing the
difficulty and complexity of the fabrication process of our CNT
-
polymer
microneedles with other standard
microneedle fabrication approaches.
INTRODUCTION
Microneedles are envisioned to provide a painless, self
-
administered alternative to
standard hypodermic injection.
Specifically, hollow microneedles allow for a delivery
architecture that is more flexi
ble than other
microneedle designs such as solid microneedles,
drug-
coated microneedles, or dissolving microneedles. This is because hollow microneedles
allow for variable delivery rate from rapid injection to emulate a hypodermic injection to slow
steady
delivery to mimic intravenous drug therapy. Additionally, since the hollow microneedle is
designed to be inert with respect to the
drug
,
many
current drugs that are delivered
into the skin
can be directly used i
n the
hollow
microneedle architecture
[
1
]
.
Previous studies on the fabrication of hollow microneedles ha
ve
focused on top-
down
fabrication approaches using either silicon,
metal
, or glass
.
While these approaches have yielded
functioning microneedles, their fabri
cation involves iterated etching or micromachining
techniques that ultimately add complexity to
the fabrication and limit
industrial
-
scale application
[1]
. Here, we introduce a new perspective on forming hollow microneedles by fabricating
microneedles usin
g a bottom
-
up approach by using vertically
-
aligned carbon nanotubes (VA
-
CNT
s
)
. VA
-
CNT
s
are first used
as a scaffold for forming the shape of the microneedle and then
as a fiber component in a CNT
-
polymer composite to create a high strength material capable
of
penetrating the skin.
By utilizing standard methods to grow VA
-
CNTs
,
including catalyst patterning and
thermal chemical vapor deposition (CVD), we directly produce a hollow microneedle
.
In
designing
the CNT
-
polymer composite microneedle, the final pr
oduct must achieve four
mechanical
objectives: (i)
have
high mechanical strength under compression to achieve skin
penetration, (ii) conformally coat the VA
‐
CNT
s
with polymer allowing the
microneedle
to retain
its original shape
from catalyst patterning
,
(iii) anchor
microneedles
to a common polymer base
for easy transfer
from the growth substrate to drug delivery platforms, and (iv) maintain an
unobstructed hollow cavity for drug delivery.
8
0
3
These objectives are interdependent and thus careful choice and a
pplication of polymer
must be applied to create a functioning microneedle. Here
, we demonstrate our approach using
the negative photoresist
SU8
-2025 (MicroChem, Newton, MA) as a candidate polymer. The
high UV absorbance of CNT
s allow us to selectively cu
re the microneedle device. Thus, the
polymer base can
be cured while leaving the polymer within the microneedle cavity uncured
allowing for removal
of the polymer
in later steps
. SU8
-2025 is chosen due to its property of
forming thick films through spin c
oating (typically 25
μm at 3000 rpm) and high elastic modulus
of 3 GPa
[ 2] . A p
revious stud
y has
shown that low viscous SU8, such as SU8
-2002 (MicroChem,
Newton, MA)
, can be spin coated on VA
-CNT
s t o create a composite [
3] . However, the thin
film created by this resist
through spin coating
(2μm at
3000 rpm) i
s too fragile for making a
structurally
supportive base for the microneedles
[2]
. After proving that the CNT
-polymer
composite microneedle is mechanicall
y feasible with SU8
-2025, future studies will focus on
broadening the number of polymers that can be employed in this technique.
EXPERIMENT
Vertically aligned CNTs are fabricated via thermal CVD on a silicon substrate patterned
with catalyst in the sha
pe of hollow circle
s which act
as the microneedle template
. To obtain this
pattern, the silicon wafer is masked with photoresist patterned
through photolithography
into the
desired microneedle geometry. A
lumina and iron of thickness 10nm and 1nm respective
ly are
then deposited on the wafer via electron beam deposition. After deposition, photoresist is
stripped from the wafer to finalize the substrate preparation. For the CVD process
, ethylene and
hydrogen gas
is flowed
across
the substrate at 750°C, 600 tor
r and 490 sccm and 210 sccm
respectively for an hour to achieve VA
-CNT growth
.
At the end of CVD, the VA-
CNTs are patterned into hollow cylinders
matching the
catalyst pattern
of 150 μ
m outer diameter and inner cavity diameter of 25
μm (Figure 1, Step 1)
.
The height of the microneedles can be varied
from 150 μm to 400 μm. Patterned VA-
CNT
s
alone cannot act as a microneedle.
VA
-CNT
s are
prone to buckling and have
an exceedingly low
modulus under compression of only 550 kPa
[ 4] . Additionally, the VA
-CNT
s must be connected
to a common base other than the growth substrate to allow for the microneedles to be transferred
onto a device platform for drug delivery.
Starting with the patterned VA
-CNTs, we incorporate SU8
-2025 to create the composite
microneedle. SU8
-2025 is dropcast
ed onto the V
A-CNT sample
which is then passively wicked
into the interspacing of the VA
-CNTs
. Next, the sample is
spin coated at 3000 rpm for 60
seconds
to
remove the excess material and
simultaneously create a base that is thick enough
(nominal 25 μ
m) to allow for the microneedles to be easily removed from the silicon substrate in
later steps with minimal damage to the microneedles and the polymer base
(Step 2
). However,
the high viscosity of the SU8-
2025 causes the resist
to pool in the inner cavity of the
microneedle. After spin coating, the sample is soft baked at 95°C for four minutes.
The p
hotoresist is selectively cured under oblique i
ncidence UV light
( UV/Visible Light
Exposure Chamber
, MTI Corp., Richmond, CA)
at 30 mW/cm
2
for up to 1 minute
( Step 3
). In
this arrangement, the polymer base is cured along with the SU8 embedded within the VA-
CNT
interspacing
. T
he SU8 in the inner cavity remains uncured due to the oblique incidence of the
UV light preventing direct exposure of the inner cavity
. Additionally, the
VA-
CNT
s in the CNT
-
SU8 composite
attenuat
e the amount of UV light
reaching the inn
er cavity through the
composite. Following UV exposure, the sample goes through a post exposure bake
1. Patterned VA
-CNT
Fabrication
2. Spin Coat
3
.
UV Exposure
4. Liquid Development
5
. Thermal Cure
6
. Substrate
Separation
7.
Attach to
Delivery Platform
Developer
UV
Figure 1
-
Microneedle fabrication process starting with CVD fabrication of patterned VA
-CNTs
.
for 3 minutes at 95°C. The m
icro
needles are then submerged in SU8 D
eveloper (MicroChem,
Newton, MA) to clear the uncured photoresist from
the inner cavity
( Step 4
). The submerged
sample is placed on a shaker table set to 150
rpm for 10 minutes. Following development, the
sample is rinsed
in isopropanol and cured in a vacuum oven at 150 °C for 20 minutes
( Step 5
).
At this point
, the
CNT
-SU8 composite
microneedle is fully formed with a
clear inner
cavity and a common SU8 base. By taking advantage of the poor adhesion between SU8 and the
silicon substrate as well as the thick SU8 base layer, the microneedles can be removed
mechanically with a razor blade or tweezers
( Step 6
). The device is then transferred onto a
delivery platform that connects
the inner cavity of the microneedles to a liq
uid reservoir for drug
flow
( Step 7
).
RESULTS & DISCUSSION
CNT
-SU8 Microneedle
The
CNT
-SU8
microneedles are spaced on a 1
mm grid to ensure that each needle
independently penetrates the skin (Figure 2
a,b).
Ideal microneedle height was found to be in
the
range of 200 μm to 250
μm. Microneedles above
this range are more susceptible to buckling
failure. Microneedle heights below 200 μm
were found to be too short to achieve consistent skin
penetration. SU8 conformally coats the microneedle allowing the f
inal product to retain the
original pattern defined by catalyst patterning and CVD growth of the VA
-CNT
s. Comparing the
VA-
CNT structure to the CNT
-SU8 composite structure, we see that the SU8 fully envelops the
interspacing between the nanotubes creating
a single solid
composite
structure (Figure 2
c,d
).
SEM imaging of the underside of the needle confirms that the inner cavity of the needle is clear
of polymer after fabrication (Figure 2
e).
Liquid flow through the microneedle is demonstrated
by connecting a microneedle array
to a water filled syringe
. Through hand actuation, the microneedle is capable of expelling water
into the air
at rates of up to 600 μL/min per needle (Figure 3a). Despite the relatively thin
polymer base of approximately 25 μ
m, the ba
se shows no signs of fatigue or cracking under high
flow rates.
Achieving
flow greater than 100μL/min per needle under minimal actuation pressure
demonstrates the low hydraulic resistance of the device which ultimately lowers
the work needed
to flow liquid
through the microneedles at any flow rate.
For
the current microneedle height
, we
anticipate
for
in vitro
drug delivery using delivery rates
of about 1 to 10 μL/min per needle
and
volumes of about 100μL
due to the restrictive permeability of the skin.
Lon
ger microneedles
penetrating deeper into the skin may enable higher flow rates and may lead to future applications
of the microneedle as an auto
-injector for rapid delivery of rescue medication.
(
d)
(c)
(b
)
(a)
(e)
Figure 2
- (a),
(b)
CNT
-SU8 composite microneedle array. Outer diameter 150
μ m with 25 μ
m
inner diameter. Structure of microneedle
(c)
before and
(d)
after SU8 incorporation.
(
e)
Underside of
microneedle after fabrication showing the
inner cavity
is clear of polymer
.
In vitro
skin penet
ration
is demonstrated on thin (<0.5 mm) samples of dorsal swine skin
prepared
by using a dermatome to cut full thickness skin samples
. Microneedles were pressed by
hand into the swine skin and achieved penetration at tip pressures in the range of 60MPa to
90MPa. Prior to penetration, the microneedles are coated with dry methylene blue powder. Upon
contact with the interstitial fluid in the skin, the methylene blue dye is passively released from
the needle marking the point of contact with the skin (Figure 3b). The clear pattern of the 2 x 2
microneedle array
indicates
positive penetration and ultimately demonstrates that the CNT
-SU8
composite has sufficient strength to achieve skin penetration. To present, all of the mechanical
objectives of the microneedle
outlined previously
have been achieved. Ongoing work is
now
looking to build upon these results by characterizing the microneedle’s
in vitro
liquid delivery
into the skin as well as broadening the method to demonstrate incorporation of other polymers.
(
a)
(b)
1 mm
1
cm
Figure 3
- (a)
Water jets from the microneedle array with exit velocity of about 600 μL/min per
needle
. (b)
In vitro
swine skin penetration marks by microneedles coated in methylene blue dye.
Fabrication Method Comparison
A t
rade study was conducted by comparing
the laboratory difficulty in fabricating the
CNT
-SU8 composite microneedle to
the current
primary approaches for microneedle fabrication
including silicon microneedles and micromolding
. Fabrication of both silicon and CNT
-SU8
microneedles
offer
similar advantages in terms of geometry customization and parallel
processing of large numbers of microneedles. In characterizing the fabrication process for silicon
microneedles, reactive ion etching (
RIE
) is presumed
to be the primary
method for defining the
microneedle shape with additional processes such as wet etching or micromachining used as
secondary methods to optimize the microneedle geometry
[1, 5-
7] .
For each process used in the
fabrication method, a difficulty factor between 1 through 3 is assigned based on the difficulty of
the process on the laboratory
scale (Table I).
Ta
ble I
: Trade s
tudy of
fabrication difficulty
for CNT
-SU8 composite microneedles and silicon
microneedles.
CNT-SU8 Composite
Silicon (Dry Etched)
Photolithography
Physical Vapor
Deposition
Nanotube
Fabrication
Polymer
Incorporation
Mask Deposition
*
Mask Removal
*
Reactive Ion
Etching
Wet Etching
Process Difficulty
Technique
Difficulty
Techniques
Number of Times Process Performed
Micromachining
1
1
1
1
—
—
—
—
0-1
2-3
2-3
2-3
2-3
0-2
0-1
—
—
—
1
2
1
1
1
1
2
2
3
10-22
5-8
Technique
Difficulty
*
Metal or Oxide
Mask
CNT
-SU8 composite
has a smaller total technique difficulty than
silicon showing that the
CNT
-polymer method has the potential to be significantly
simpler than the silicon approach. This
is because
creating
a hollow cavity in the silicon approach requires iterative etching steps which
in turn require
iterative processing
steps.
In contrast, the entire geometry of the CNT
-SU8 needle
is defined from th
e sequential steps of catalyst patterning and nanotube fabrication. Another
important consideration in the CNT
-SU8 approach is that
incorporation of
SU8
on the VA
-CNTs
relies on a combination of simple and passive mechanisms such as capillary action and UV
exposure
that can be easily executed
on both the laboratory and industrial level
. The primary
fabrication challenge for the CNT
-polymer microneedle is the fabrication of VA
-CNTs.
However, the recent increase of commercial fabrication options should lower the overall
fabrication difficulty in the long run.
To present, only
a hollow cylinder
geometry has been considered for our CNT
-SU8
microneedle. In comparison with other microneedles, a tapered shape is typically preferred to
minimiz
e the tip area as the r
equired penetration force scales linearly with tip area [
8] . Previous
work has demonstrated that drug release from the top of the microneedle is susceptible to skin
occlusio
n which increases the hydraulic resistance of delivery. A
suggested alternative is to
release the drug
from the side of
the microneedle to
optimize the delivery by
lower
ing the
hydraulic resistance [
1, 5]
. The fabrication of CNT
-SU8 microneedles can be amended to
incorporate these more complex geometries if desired.
T he microneedles can be modified after
polymer incorporation via micromachining or before polymer incorporation using techniques
such as capillography
[ 3] or focused ion beam milling
. Further study of the
in vitro
drug release
profile for our current microneedle is necessa
ry before determining if the potential benefit of
optimizing the shape of our microneedles outweighs the cost of increased fabrication difficulty.
Micromolding is a
nother
common approach to producing hollow microneedles by
electroplat
ing
micromold
s to produce a hollow metallic microneedle. Typical master structures
for micromolding are fabricated by micromachining or laser drilling of a bulk material which can
be time consuming and ultimately limits the minimum feature size of the needle
[ 1, 9]
. However
after the master structure is complete, the micromolding process allows for fast fabrication of
large numbers of mic
roneedles.
Both the silicon and CNT
-SU8 needle can achieve much finer
feature size and thus may both be potentially used as a master structure for micromolding. A
previous study has demonstrated the successful use of CNT
-SU8 composites for general
micromo
lding applications
[3].
CONCLUSIONS
A new approach to fabricating hollow microneedles has been shown using vertically
-
aligned carbon nanotubes and SU
8- 2025. By taking advantage of self
-assembly, VA
-CNTs can
be simply adapted as a hollow microneedle by i
ncorporating SU
8- 2025. SU
8- 2025 allows for the
creation of a strong composite with VA
-CNT
s while simultaneously creating a
supportive
base
for the microneedle array
. Initial experiments have
shown that the CNT
-SU8 needles can
penetrate the skin
in vitro
an
d can structurally support high flow rates of up to 600μ
L/min per
needl
e. Further studies will investigate expanding the number
of polymer
s that can be used in
this architecture as well as
in vitro
characterization
of the microneedle’s delivery performance.
In comparing the
fabrication of CNT
-SU8 microneedles with other approaches
, we find
that the CNT
-SU8 fabrication method is potentially
simpler than that of silicon microneedles
.
CNT
-SU8 composites may also be incorporated
as master structures in
micromolds to allow for
feature size on the order of several microns.
Laboratory scale fabrication represents only a single
aspect for comparing microneedles. Future trade studies will need
to take into account the
performance, cost, and scalability aspects of the
CNT
-polymer microneedle to properly identify
future applications and markets for this technology.
ACKNOWLEDGMENTS
We acknowledge Zcube s.r.l for their financial support of this work. We
also
acknowledge the Kavli
Nanoscience Institute and the Geology a
nd Planetary Sciences Analytical
Facility for their
support in running experiments.
REFERENCES
1.
Y.-C. Kim, J.-
H. Park and M. R. Prausnitz, Advanced Drug Delivery Reviews
64
(14),
1547-
1568 (2012).
2.
Microchem, SU
-8 2000 Processing
Guidelines.
3.
M. De Volder, S. H. Tawfick, S. J. Park, D. Copic, Z. Zhao, W. Lu and A. J. Hart,
Advanced materials
22
(39), 4384-
4389 (2010).
4.
L. Ci, J. Suhr, V. Pushparaj, X. Zhang and P. M. Ajayan, Nano letters
8
(9), 2762-
2766
(2008).
5.
H. J. G. E. Gardeniers, R. Luttge, E. J. W. Berenschot, M. J. De Boer, S. Y. Yeshurun, M.
Hefetz, R. van't Oever and A. van den Berg, Microelectromechanical Systems, Journal of
12
(6), 855-
862 (2003).
6.
B. Ma, S. Liu, Z. Gan, G. Liu, X. Cai, H. Zhang and Z. Yang, Mic
rofluidics and
Nanofluidics
2
(5), 417
-423 (2006).
7.
L. M. Yu, F. E. H. Tay, D. G. Guo, L. Xu and K. L. Yap, Sensors and Actuators A:
Physical
151
(1), 17
-22 (2009).
8.
S. P. Davis, B. J. Landis, Z. H. Adams, M. G. Allen and M. R. Prausnitz, Journal of
biomechanics
37
(8), 1155-
1163 (2004).
9.
J. J. Norman, S. O. Choi, N. T. Tong, A. R. Aiyar, S. R. Patel, M. R. Prausnitz and M. G.
Allen, Biomedical microdevices
15
(2)
, 203-
210 (2013).