You are on page 1of 9

Liquid phase deposition of titania onto nanostructured

poly-p-xylylene thin lms


Niranjan Malvadkar,
a
Walter J. Dressick
*
b
and Melik C. Demirel
*
a
Received 11th February 2009, Accepted 23rd April 2009
First published as an Advance Article on the web 27th May 2009
DOI: 10.1039/b902882j
We describe a simple, solution-based, two-step process for the fabrication of titaniaparylene
composite lms as a prerequisite for their evaluation as materials for bioimplant applications. In the
rst step, a ligand capable of binding titania, such as phenylphosphonic acid, is physisorbed onto
a nanostructured poly-p-xylylene thin lm previously prepared via surface-promoted oblique angle
polymerization of radicals formed during vapor-phase pyrolysis of [2.2]-p-cyclophanes. The adsorbed
ligand templates conformal growth of titania on the polymer surface in the second step via a liquid
phase deposition process involving the controlled hydrolysis of (NH
4
)
2
TiF
6
in the presence of H
3
BO
3
in
pH 2.88 aqueous solution at 50

C. SEM and AFM analyses support a deposition mechanism that
includes direct growth of titania on the ligand-impregnated parylene surface, as well as incorporation of
titania nanoparticles nucleated in solution into the growing lm. XPS and XRD results show that the
as-deposited titania contains both amorphous and nanocrystalline anatase phases, with the latter
readily consolidated by annealing at 200

C without destruction of the underlying parylene polymer.
Titania adhesion can be tuned by proper choice of the ligand, with ligands such as phenylphosphonic
acid that strongly bind to titanium dioxide leading to deposition of titania lms that pass the Scotch

tape adhesion test.


Introduction
Polymerceramic interfaces are important for many technologi-
cally signicant applications, including bone tissue engineering
1
and bioimplants.
2
Ceramics such as titania and alumina are
generally favored for these applications due to their biocom-
patibility and osteointegration with osteoblasts (bone forming
cells). However, the highly brittle nature of conventional forms
of these ceramics makes it impossible to mimic the mechanical
properties of bone, hindering their use.
3
One means to improve the mechanical properties of the
ceramic is to reduce grain size to the nanometre regime (i.e., <100
nm in at least one dimension), thereby better mimicking the
surface of natural bone.
4
Because osteoblast attachment to
nanophase ceramics is superior to conventionally prepared
ceramics,
5
an improved interface between natural bone and the
bioimplant is also obtained. However, bending modulus
mismatches between ceramic and bone, even for nanophase
ceramics,
5
remain sufciently large to adversely affect the
durability, viability, and performance of the implant. Polymer
ceramic composites,
5
fabricated primarily via solgel
4
or chem-
ical vapor deposition
6
methods, represent a potential solution to
this problem, provided that cost, adhesion, and related issues
associated with the composite materials interface and mechan-
ical properties can be adequately addressed.
Poly-p-xylylenes (i.e., parylenes, PPXs) are a class of polymers
readily formed via surface-promoted polymerization of radicals
formed during the vapor phase pyrolysis of substituted [2.2]-p-
cyclophanes, as shown in Fig. 1A.
7,8
The biocompatibility of
various functionalized PPX lms is well recognized,
9,10
with
current applications ranging from coatings for retinal/ocular
11
or
neural
12
implants to antifouling coatings for stents.
13,14
These
applications generally exploit the ability to tune cellular adhe-
sion
15
or selectively attach materials for therapeutic release
13,14
to
the parylene lm by controlling its surface chemistry through
judicious choice of the [2.2]-p-cyclophane substituent groups.
Control of parylene lm morphology provides an alternate
means to address issues such as composite adhesion or wetta-
bility that complements surface chemistry approaches. For
example, adhesion of disparate materials is often enhanced if the
surface of one is roughened prior to deposition of the second due
to mechanical interlocking at the composite interface. Such
roughening can be accomplished for parylene lms via
conventional methods as diverse as photolithography
16,17
or
chemical mechanical polishing.
18
However, these techniques
require additional processing steps that increase costs or alter
lm surface chemistry, while providing roughness features of
limited resolution or reproducibility, respectively.
Recently, we described a straightforward and simple means to
precisely control the lateral and surface morphology of vapor-
deposited parylene lms at the nanoscale during lmformation.
8,19
The process is illustrated in Fig. 1Band involves radical deposition
and surface polymerization at an oblique angle (a 90

), rather
than normal, to the substrate surface. Surface diffusion and
a
Materials Research Institute and Engineering Science, College of
Engineering, The Pennsylvania State University, University Park, PA,
16802. E-mail: mdemirel@engr.psu.edu
b
Naval Research Laboratory, Code 6910, 4555 Overlook Avenue, S.W.,
Washington, DC, 20375. E-mail: walter.dressick@nrl.navy.mil
Electronic supplementary information (ESI) available: Fig. S1, S2 and
S3. See DOI: 10.1039/b902882j
4796 | J. Mater. Chem., 2009, 19, 47964804 This journal is The Royal Society of Chemistry 2009
PAPER www.rsc.org/materials | Journal of Materials Chemistry
geometric shadowing effects
7,20
occurring during oblique angle
polymerization (OAP) lead to parylene growth as nanostructured
rods onthe substrate, as showninFig. 1C(e.g., a10

; u0s
1
).
19
The inclination, surface density, length, and diameter of the par-
ylene nanorods are determined by the oblique angle (a), as well as
factors suchas the chemical species initially present onthe substrate
surface, substrate and vapor pyrolysis temperatures, deposition
time, and vapor pressure within the deposition chamber. In addi-
tion, rotation of the substrate about the normal (i.e., u > 0 s
1
)
during OAP permits the fabrication of lms exhibiting more
complex morphologies, such as helices or chevrons.
8
Nanostructured parylene lms formed via OAP exhibit excel-
lent biocompatibility and bioactivity, fostering broblast cell
adhesion and proliferation in vitro.
19
The roughened surface
afforded by the nanostructured polymer permits control of lm
wettability
21
and provides a template for the reproducible vapor
phase deposition of adherent Au and Ag metal lms useful as
substrates for surface-enhanced Raman spectroscopy (SERS).
22
Metal adhesion can be further improved by incorporation of
a chemical adhesive component complementing the mechanical
component provided by the nanostructured morphology. For
example, we have recently prepared parylenemetal interfaces
exhibiting enhanced metal adhesion via an electroless process in
which a pyridine ligand is rst physisorbed onto the nano-
structured parylene surface and is stabilized by favorable p and/
or van der Waals interactions with parylene aromatic groups.
23
The adsorbed pyridine covalently and selectively binds colloidal
Pd(II) species
2426
that subsequently catalyze electroless deposi-
tion of adherent, conformal Ni
23
or Co
27
lms onto the nano-
structured surface.
In this work, we generalize our ligand adsorbate concept by
extending it to the fabrication of nanostructured parylenetitania
composites. We describe a liquid phase deposition (LPD)
method for conformal deposition of titania onto a nano-
structured parylene thin lm and characterize various physical
and chemical properties of the resulting composite. Our efforts
constitute a necessary rst step in developing an understanding
of the factors affecting adhesion, wettability, and mechanical
properties of these new composite materials, important for their
subsequent evaluation as potential bioimplant materials and
coatings.
Experimental
Materials and nanostructured lm preparation
Deionized water (18.2 MU cm
1
) from a Barnstead NanoPure II
deionizer was used for all experiments. All materials were used as
received. The ligands phenylphosphonic acid (PPA), thiophenol
(TPH), pyridine (PYR) and the ammonium hexauorotitanate
and boric acid were all ACS reagent grade or better from Aldrich
Chemical Co. Parylene C (i.e., dichloro-[2.2]-p-cyclophane;
Parylene Distribution Services, Inc.; Lot No. 060514; note
Fig. 1A) was used for the preparation of all nanostructured
poly(chloro-p-xylylene) (i.e., PPX-Cl) lms. The OAP process for
the preparation of the PPX-Cl lms has been described in detail
elsewhere.
20
Briey, a 0.3 g portion of dichloro-[2.2]-p-cyclo-
phane, vaporized at 175

C, was subsequently pyrolyzed at 690

C to form free radicals according to Fig. 1A in a modied


vacuum chamber.
28
Styrylethylsiloxane self-assembled mono-
layer (SAM) coated p-type Si(111) wafers (Silicon Quest Inter-
national, NV) were used as substrates for PPX-Cl deposition. As
shown in Fig. 1B, the substrate was held xed (i.e., u 0 s
1
) at
a constant angle of a 10

to the radical ux to initiate the


surface polymerization leading to formation of the PPX-Cl
nanorods (note Fig. 1C).
20
Deposition was carried out under low-
vacuum conditions (10 mbar) for 10 minutes. Planar deposited
PPX-Cl lms were prepared under the same conditions without
using the angular ux (i.e., u 0 s
1
; a 90

).
Titania deposition
PPX-Cl coated Si wafers (OAP and planar methods) were incu-
bated in 0.5 M aqueous phenylphosphonic acid (PPA), 0.5 M
aqueous pyridine (PYR), or 0.5 M thiophenol (TPH) in ethanol
solutions in closed glass containers in a fume hood for 48 h. The
wafers were then rinsed in their respective solvents and trans-
ferred to the titania bath. The LPD bath
29
was prepared imme-
diately prior to use and consisted of an aqueous solution
containing 0.05 M (NH
4
)
2
TiF
6
and 0.15 M H
3
BO
3
. Dilute
aqueous hydrochloric acid, HCl, was added to adjust the pH to
2.88. The PPX-Cl coated wafers were submerged vertically in the
bath held at 50

C for 3 to 24 h to deposit the titania lms. The
wafers were then removed from the bath, sonicated in water for
30 s (VWR Scientic Model 75 HT Aquasonic Bath), rinsed in
fresh water, and dried in a streamof N
2
gas (compressed N
2
tank,
water-pumped). Wafers were stored in sealed Fluoroware

containers until needed for characterization.


Fig. 1 (A) Structures of materials used and pyrolysis reaction; (B)
polymerization method; (C) PPX-Cl nanorods prepared via the scheme in
part B (u 0 s
1
; a 10

; b 55

).
This journal is The Royal Society of Chemistry 2009 J. Mater. Chem., 2009, 19, 47964804 | 4797
Characterization
An atomic force microscope (AFM, Nanoscope

E, Veeco Inc.)
equipped with an air chamber and silicon nitride cantilevers
operated in contact mode and a JEOL 6700F eld-emission
scanning electron microscope (FESEM) operated at an acceler-
ating voltage of 3 kV were used to characterize the morphology
of the deposited titania lms. Film roughness values were
measured by AFM using methods described previously.
23
The
compositions of the titaniaPPX-Cl composites were determined
using the energy dispersive X-ray (EDX) attachment (Ametex) of
a Philip XL-30 scanning electron microscope (SEM) and an Axis
Ultra XPS X-ray photoelectron spectrophotometer (Kratos)
equipped with a monochromatic Al Ka X-ray source and oper-
ated at a 20 eVpass energy (700 mm300 mmhybrid sample spot
size) under high vacuum conditions (10
9
torr). X-Ray photo-
electron spectra were collected at a take-off angle of 90

and
analyzed using CasaXPS v. 2.312Dev9 spectral tting software
with the C(1s) peak at 284.6 eV taken as the reference. As-
prepared titania samples were subsequently calcined under Ar
atmosphere at 200

C for 24 h and characterized by X-ray
diffraction (XRD). Care was taken to avoid heating the samples
above the PPX-Cl melting temperature of 290

C
30
at any time
during the calcination process. The XRD data were obtained
using a Scintag X2 X-ray diffractometer with a Cu Ka radiation
source and a Si(Li) Peltier cooled detector. The diffraction
patterns were taken using grazing angle incidence (incidence
angle of 2 degrees) while the detector was scanned from 5 to 65
degrees at a rate of 0.02 degree s
1
.
Adhesion tests were performed using the Scotch

tape test.
29
Equal pressure was applied during the attachment and removal
of the tape for all composite lms. Titania coverage was calcu-
lated from 5 low-magnication SEM images taken before and
after the Scotch

tape test. Static contact angle with DI water (10


mL droplet) was recorded using a contact angle measuring
instrument (FTA 1000B, First Ten Angstroms Inc. Portsmouth,
VA). Film thickness was measured before and after LPD titania
deposition using a prolometer (Tencor P-10) with a 1 mg stylus
force and 20 mm s
1
scan rate. Transmission electron microscope
(TEM) samples were prepared by scraping the titania membrane
off the PPX surface, suspending it in ethanol, and sonicating for
20 minutes. A drop of this suspension was added to a carbon-
coated copper grid and allowed to dry in air. TEM images and
selected-area electron diffraction (SAED) were obtained using
a Philips (FEI) EM420T TEM with an accelerating voltage of
120 kV. Titania nanoparticles were identied using selected-area
EDX before taking the TEM image.
Results and discussion
LPD is a process in which controlled hydrolysis of metal ions in
aqueous solution generates oxide or related precursor species
that are subsequently deposited onto various substrate
surfaces.
31
The process differs from electroless deposition in that
no change in redox state occurs during the deposition process so
that no reductant or catalyst is required. A variety of metal
oxides, including titania,
3238
can be deposited in this manner,
with oxide composition and morphology determined primarily
by the precursor metal species, hydrolysis conditions, and
substrate surface chemistry.
31
Oxide nucleation and growth can
occur directly at the substrate surface through a heterogeneous
nucleation and growth process to form a conformal oxide
coating and within the aqueous solution to generate oxide
nanoparticles.
29
The latter process is increasingly favored in
highly supersaturated solutions and ultimately leads to inclusion
of the resulting nanoparticles in the oxide lm growing at the
substrate surface.
29,31
For titania lms, controlled hydrolysis of TiCl
n
(OC
2
H
5
)
4n
(n
0, 2, or 4)
34,35
or peroxotitanate
39
species leading to deposition
of amorphous titania lms onto bare silica or silica modied with
chemisorbed amino-, phenyl-, and octadecylsiloxane self-
assembled monolayers (SAMs) has been reported. In contrast,
analogous work using TiF
6
2
species usually leads to deposition
of anatase,
36,37
with a preference for c-axis crystal orientation
perpendicular to the surface for certain substrates.
29,31
In this
work, we have utilized aqueous hexauorotitanate solutions
containing boric acid as a uoride ion scavenger at pH 2.88 and
50

Cto control the TiF


6
2
hydrolysis rate, as illustrated in eqn.
(14), with the net oxide deposition reaction shown in eqn. (5):
TiF
6
2
+ n H
2
O 4[TiF
6n
(OH)
n
]
2
+ n HF (1)
[TiF
6n
(OH)
n
]
2
+ OH

/[TiF
5n
(OH)
n+1
]
2
+ F

///[Ti(OH)
6
]
2
(2)
H
3
BO
3
+ 4 HF /HBF
4
+ 3 H
2
O (3)
[Ti(OH)
6
]
2
+ D /TiO
2
(s) + 2 H
2
O + 2 OH

(4)
2 (NH
4
)
2
TiF
6
+ 3 H
3
BO
3
+ HCl /2 TiO
2
(s) + 5 H
2
O
+ NH
4
Cl + 3 NH
4
BF
4
(5)
Under these deposition conditions, the solution is highly
supersaturated and titania nucleation and growth both directly
at the substrate surface and in the solution occur, with subse-
quent incorporation of titania nanoparticles formed in the
solution into the surface titania lm.
29
Fig. 2 depicts the simple two-step scheme used for deposition
of titania lms onto our nanostructured PPX-Cl substrates. In
the rst step, the nanostructured PPX-Cl lm surface is
impregnated with a ligand capable of binding titania or its
precursor species (e.g., eqn. (14)) via simple immersion into
a ligand solution. Ligand physisorption occurs through forma-
tion of favorable p and/or van der Waals interactions between
hydrophobic portions of the ligand structure and aromatic resi-
dues of the parylene chains in a process analogous to that
described for ligand adsorption into solvent-templated nano-
cavities in aromatic organosiloxane SAMs.
4045
For the nano-
structured PPX-Cl lms, however, the corresponding
nanocavities are templated by the anisotropic growth condi-
tions present during the vapor-phase OAP process used to
deposit the parylene, rather than a solvent. Ligand adsorption is
facilitated by the amorphous nature and high curvature of the
parylene nanorods formed, while corresponding planar parylene
lms exhibit negligible ligand binding.
23
The three-dimensional
nature of the parylene polymer chains in nanostructured PPX-Cl
lms provides added conformational exibility in binding ligand
compared to the more rigid two-dimensional aromatic siloxane
SAMs, permitting dissolution and successful binding of a wider
4798 | J. Mater. Chem., 2009, 19, 47964804 This journal is The Royal Society of Chemistry 2009
range of ligands via the use of small hydrophilic alcohols such as
ethanol, in addition to water, as ligand solvents.
We have tested three species, namely phenylphosphonic acid
(PPA), thiophenol (TPH) and pyridine (PYR), which spans
a range of titania binding strengths, as ligands. Because Ti
IV
preferentially binds oxygen-, rather than nitrogen-containing
ligands, it binds strongly to phosphonate-containing species such
as PPA but interacts only weakly with aromatic N-containing
species such as PYR.
46
The selection of TPH represents a ligand
of intermediate binding strength due to the presence of the SH
site, rather than the more strongly interacting OH site of the
PPA. After incorporation of ligand into the nanostructured
PPX-Cl lm, treatment with the aqueous hexauorotitanate-
boric acid LPD bath in the second process step of Fig. 2 initiates
titania lm deposition at the ligand sites.
Table 1 summarizes our titania deposition results as functions
of (1) LPD bath treatment time for PPA-impregnated PPX-Cl
lms, and (2) type of ligand for PPX-Cl lms separately treated
by each ligand (48 h) and the LPD bath (24 h). For the time-
dependent study, SEMs and AFM surface maps are provided in
Fig. 3 as functions of the LPD bath treatment times for the PPA-
impregnated nanostructured PPX-Cl lms. The quantity of
titania deposited onto the PPA-treated PPX-Cl surface increases
monotonically with time, as shown by the decreasing C and
increasing Ti EDX signals in Table 1. Only traces of titania are
observed on the PPX-Cl surface in the absence of ligand treat-
ment, as expected for the model of Fig. 2.
Titania initially deposits as a conformal lm onto the PPA-
impregnated PPX-Cl nanorods, as shown in Fig. 3A, covering
50% of the substrate surface after only 3 h of deposition. After
longer times, however, the distinct PPX-Cl nanorods are
increasingly covered as fronts of growing titania lms merge,
lling the pores between the PPX-Cl nanorods. At the same time,
titania nodules appear and persist on the surface as shown in the
SEM images of Fig. 3B and 3C for 6 h and 24 h LPD bath
treatments, respectively. Signicant changes in surface roughness
simultaneously occur, as shown by the AFM maps in Fig. 3 and
summarized in Table 1. Initial PPX-Cl lm roughness (i.e., 46.3
5.0 nm) decreases markedly during the rst 3 h of titania
deposition (i.e., 19.5 1.2 nm), followed by a slower increase to
31.6 1.9 nm after 24 h. The changes in morphology shown in
the SEMs of Fig. 3, together with the roughness variations
documented in Table 1, are consistent with a model in which
initial nucleation and conformal growth of titania on the PPA-
treated PPX-Cl nanorods is supplemented at longer times by
incorporation of titania nanoparticles nucleated in the solution
into the growing titania lm, as expected for LPD from
a supersaturated solution under our conditions (note experi-
mental section).
29,31
Ligand type also signicantly affects the deposition of titania
onto the PPX-Cl lm. Titania coverage decreases from >95%
when PPA ligand is present to 5090% for TPH and <1020%
for PYR in Table 1. The Ti EDX signal of 19.8% noted for PPA
is reduced to 14.7% for TPH and just 1% for PYR. Both trends
correspond to the expected titanialigand binding strength of
PPA > TPH [PYR. Although PYR adsorbs well to the PPX-
Cl nanostructures,
23
its weaker interactions with Ti
IV
species
severely limits titania deposition. However, adsorbed pyridyl N
sites typically exhibit surface pK
a
s of 3.35.9
47,48
and will be
substantially protonated under our titania deposition conditions
(i.e., 72.599.9% protonation, respectively, at pH 2.88). There-
fore, the minimal titania deposition observed in this case more
likely results from weaker electrostatic, rather than covalent,
interactions between anionic titania precursors shown in eqn. (1
3) and accessible protonated pyridyl sites on the PPX-Cl surface.
The growth of titania is summarized in Table 1 (thickness
column) and depends on the type of ligand present in the PPX-Cl
lm. While PPA and TPH treated PPX-Cl lms show the highest
growth rates, the PYR treated PPX-Cl lm shows sluggish
growth. The growth of the untreated PPX-Cl lm is extremely
poor and takes place primarily due to the physical entrapment of
the titania nanoparticles in the porous PPX-Cl. The initial rapid
titania growth rate eventually slows at longer bath treatment
times, as shown in Fig. 3D for the PPA-impregnated PPX-Cl
lm. This behavior reects the diminished surface area available
for titania deposition as lling of the interstitial regions between
the PPX-Cl nanorods by titania is completed and the titania
completely covers the PPX-Cl lm (note Fig. 3B and 3C).
The corresponding contact angles of the titaniaPPX-Cl
composites from Table 1 also drop steadily as the deposition
proceeds. The PPX-Cl lm possesses a hydrophobic surface with
a contact angle of 124 5

due to the inherent lowsurface energy


of the PPX-Cl combined with the enhanced hydrophobicity due
to the nanostructured morphology. As the titania covers the
nanostructured PPX-Cl, the composite lm exhibits a higher
surface energy and therefore a lower contact angle. It is clear that
after 6 h of titania deposition (Fig. 3B), the PPX-Cl surface is
essentially completely covered with titania and therefore exhibits
superhydrophilicity. In general, higher coverage and thickness of
the titania layer lowers the contact angle for the composite lm,
consistent with expectations based on the model of Fig. 2 and our
titania growth rate and coverage observations.
Fig. 2 Titania deposition scheme and ligand structures.
This journal is The Royal Society of Chemistry 2009 J. Mater. Chem., 2009, 19, 47964804 | 4799
Surface effects are also manifested in the adhesive behavior of
the as-deposited titania lms. Fig. 4 and Fig. S1 (see ESI)
illustrate Scotch

tape adhesion test results for titania lms


deposited onto the PPA- and TPH-impregnated nanostructured
PPX-Cl lms. Although some faint cracks appear after removal
of the tape, negligible quantities (<5%) of the titania lm are
removed from the surface when PPA ligand is present during the
LPDprocess. TEMimages (see Fig. S2 in ESI) clearly reveal the
presence of the 100150 nm diameter titania nanoparticles
preferred for enhanced adhesion properties required for bioim-
plant composite applications.
4
In addition, coiling of the PPX-Cl
nanorods during the titania deposition, leading to enhanced
adhesion via mechanical interlocking of the titania and the PPX-
Cl components, is also observed. The excellent adhesion and high
coverage of the titania in this case is consistent with its strong
binding towards phosphonic acid OH groups over a large pH
range (pH 0 to pH 9).
49,50
In contrast, at least 50% of the titania
lm (light areas, Fig. 4) is removed from the PPX-Cl surface
when the TPH ligand is used. Unfortunately, the incomplete and
variable initial titania coverage (i.e., 5090%) associated with the
TPH-impregnated PPX-Cl lm precludes a more quantitative
comparison of its adhesive properties at this time. Nevertheless,
Table 1 Titania deposition on ligand-treated nanostructured PPX-Cl lms
Ligand type
a
LPD bath
c
time/h
Initial titania
coverage (%)
d
% C (EDX) % Ti (EDX)
Roughness/
nm
e
Titania
thickness/mm
Contact
angle/degrees
e
PPA 1 0.21 0.1
PPA 3 50% 74.3% 4.3% 19.5 1.2 0.76 0.1
PPA 6 >95% 53.0% 9.2% 26.6 2.1 1.19 0.5
PPA 24 >95% 19.2% 19.8% 31.6 1.9 2.26 0.4
TPH
b
24 5090% 34.8% 14.7%
f
2.03 0.2
PYR 24 <1020% 67.5% 1.0%
f
0.69 0.2
None 24 <5%
f
<0.15
Control
polymer
0 0% 46.3 5.0 0
a
Treatment with 0.5 M aqueous ligand solution for 48 h unless noted otherwise.
b
Ethanol solution.
c
pH 2.88 at 50

C.
d
Visually estimated from
scanning electron micrographs of substrates after titania deposition.
e
Average 3s (6 measurements).
f
Roughness measurement was not obtained
by AFM due to non-uniformity of the surface.
4800 | J. Mater. Chem., 2009, 19, 47964804 This journal is The Royal Society of Chemistry 2009
the results shown in Fig. 4 clearly illustrate the importance of
proper ligand selection in controlling the adhesion of the as-
deposited titania lm.
XPS analyses provide further insight concerning the nature of
the titania deposits and support for the LPD scheme shown in
Fig. 2. For example, the XPS spectra shown in Fig. 5 illustrate
the incorporation of the PPA ligand into the nanostructured
PPX-Cl lmafter treatment with aqueous PPAsolution but prior
to titania lm deposition. The single P(2p) peak at 133.5 eV
(100.0%) in Fig. 5A is identical to that observed elsewhere for
analogous aromatic phosphonate species.
51
The corresponding
O(1s) peaks due to P]O and POH are observed after decon-
volution in Fig. 5B at 531.6 eV (46.0%) and 533.1 eV (54.0%),
respectively. These are in excellent agreement with values of
531.7 eV and 533.2 eV observed by Adolphi and coworkers
52
and
531.3 eV and 533.0 eV reported by Cabeza and coworkers,
53
respectively, for analogous phosphonic acid species.
Fig. 6 shows the XPS results for the titania lm of Fig. 3C
prepared by LPD using the PPA-impregnated nanostructured
PPX-Cl lm analyzed in Fig. 5. The Ti(2p
3/2
) signal shown in
Fig. 6A comprises two components after deconvolution at
energies of 459.1 eV (87.7%) and 460.1 eV (12.3%). Identical
values (0.1 eV uncertainty) of 459.1 eV (88.5%) and 460.2 eV
(11.5%) are observed for titania lms deposited onto the TPH-
treated PPX-Cl substrate (note Fig. S3A in ESI). These peak
components occur at or just beyond the edge of the 458.4459.0
eV Ti(2p
3/2
) energy range usually associated with crystalline
anatase and rutile titanium dioxide,
32,54,55
complicating the
interpretation of the spectra. In fact, our spectra are consistent
Fig. 3 SEM and AFM images of PPA-treated PPX-Cl lms after
different LPDbath treatment times: (A) 3 h; (B) 6 h; (C) 24 h. AFMmaps
represent 2 mm 2 mm scan areas (h height). (D) Growth of the LPD
titania layer measured using a prolometer.
Fig. 4 Titania adhesion tape test results. SEMs are shown for titania
lms deposited onto nanostructured PPX-Cl lms bearing PPA or TPH
ligand before and after tape removal. Samples were treated with ligand
for 48 h followed by titania LPD for 24 h (note experimental section).
Fig. 5 XPS analysis of PPA-treated PPX-Cl lm. (A) P(2p); (B) O(1s).
Key: black line XPS spectrum; red solid line baseline; red dotted line
deconvoluted XPS peaks; red dashed line tted XPS spectrum.
This journal is The Royal Society of Chemistry 2009 J. Mater. Chem., 2009, 19, 47964804 | 4801
with the presence of both crystalline and amorphous TiO
x
(x #
2) phases. For example, mixed titanium oxides prepared using
a solgel method
56
display a Ti(2p
3/2
) peak at 459 eV analogous
to the major peak component at 459.1 eV in Fig. 6A. In addition,
Bhaumik and coworkers
51
have previously observed a Ti(2p
3/2
)
peak at 460.4 eV, assigned to tetragonal coordination of Ti
IV
by
oxygen in a local crystalline environment, in porous disordered
open-framework titanium oxophosphonate compounds. The
appearance of a similar peak at 460.1 eV in Fig. 6A (and 460.2
eV in Fig. S3A) is consistent with the presence of a corre-
sponding crystalline component in our material. At least two
mechanisms exist for incorporation of such material during our
deposition process, including: (1) capture of crystalline titania
nanoparticles independently nucleated in solution by the
growing titania surface lm, and (2) ligand-templated tetragonal
coordination of Ti
IV
, at least during the direct nucleation and
growth of titania from the ligand-modied PPX-Cl surface that
predominates during the initial stages of the LPD process.
Unfortunately, our data are currently insufcient to ascertain
which of these pathways, if any, contributes to the structure of
our titania lms.
The O(1s) spectrum of the as-deposited titania lm in Fig. 6B
exhibits component peaks after deconvolution at 530.4 eV
(50.9%) and 532.3 eV (49.1%). Analogous O(1s) components are
observed for titania lms deposited onto TPH-impregnated
PPX-Cl lms (i.e., 530.4 eV (57.9%) and 532.1 eV (42.1%), note
Fig. S3B in ESI). The differences between the O(1s) spectra of
Fig. 6B and S3B (i.e., the +0.2 eV shift of the higher energy O(1s)
component for titania lm deposited using PPA compared to
TPH and its larger area relative to its 530.4 eV component) likely
reect minor differences in composition and/or structure between
the deposited titania lms, rather than contributions due to the
phosphonate oxygen atoms of the PPA ligand for Fig. 6B. The
latter are excluded because no Cl(2p) signal due to the underlying
PPX-Cl is observed during XPS analysis, indicating that the
titania lm of Fig. 6B is sufciently thick to attenuate any O(1s)
signal due to physisorbed PPA ligands on the parylene surface.
Examination of the O(1s) spectra further supports the presence
of both crystalline and amorphous components in the as-
deposited titania lms. For example, Gonbeau and coworkers
55
attribute a Ti]O O(1s) peak at 530.4 eV to tetragonally coor-
dinated Ti
IV
in an ionic O
2
crystalline environment, with TiOH
O(1s) peaks at 532.1532.3 eV assigned to Ti
IV
in a formal O
2
environment comprising weakly adsorbed species and/or
subsurface oxide deciencies more characteristic of amorphous
TiO
x
(x # 2) lms. In similar fashion, the component peak
positions in Fig. 6B (and Fig. S3B in ESI) are in good agreement
with those observed at 530.4 eV for Ti]O and 532.3 eV for Ti
OH in amorphous TiO
x
(x # 2) lms previously prepared using
peroxotitanate solutions.
39
XRDresults shown in Fig. 7 conrmthe mixed structure of the
titania lms, supporting our interpretation of the XPS spectra in
Fig. 6. The XRD spectrum of the as-deposited titania lm from
Fig. 3Cis illustrated in Fig. 7A. The absence of a diffraction peak
at 2q 22

associated with the underlying PPX-Cl


7
substrate is
consistent with the essentially non-porous, conformal nature of
the titania lm. The small diffraction peak observed at 2q 25

in an otherwise featureless spectrum is characteristic of anatase,


indicating deposition of a titania lm containing both crystalline
and amorphous phases under our LPDconditions. This behavior
contrasts with the deposition of crystalline anatase lms usually
noted for TiF
6
2
-based LPD baths
29,31
and clearly illustrates the
ability of our ligand-impregnated nanostructured PPX-Cl
substrates to inuence the structure of the deposited titania.
Crystallization of amorphous titania can be induced via
thermal annealing at temperatures > 400

C.
57,58
However,
these temperatures clearly exceed the 290

C melting point of
our PPX-Cl lm.
30
Therefore, we have examined the annealing
behavior of our composites at a temperature of 200

C. The
XRD result in Fig. 7B indicates growth of XRD peaks charac-
teristic of anatase at 25

and 37

after annealing the sample of


Fig. 7A at 200

C for 24 h. Because direct transformation of
amorphous titania to anatase is inhibited at this temperature,
57,58
our results suggest a mechanism involving consolidation of
nanocrystalline domains within the composite during the
Fig. 6 XPS analysis of PPA-treated PPX-Cl lm after titania LPD. (A)
Ti(2p); (B) O(1s). The key is identical to that used in Fig. 5.
Fig. 7 XRD spectra of titania lms: (A) as-deposited; (B) after
annealing (200

C, 24 h, Ar). Vertical bars identify the calculated posi-
tions and intensities for X-ray reections due to anatase.
4802 | J. Mater. Chem., 2009, 19, 47964804 This journal is The Royal Society of Chemistry 2009
annealing step. Although the PPX-Cl melting point certainly
limits the annealing temperature in this work, parylene deriva-
tives exhibiting melting points as high as 420

C
30
are available
via changes in the nature and number of functional groups
present in the [2.2]-p-cyclophane precursor, permitting the use of
even higher temperatures when necessary for annealing such
composites.
Conclusions
In conclusion, we have described a new procedure for the fabri-
cation of polymerceramic composite materials comprising par-
ylenetitania thin lms. The method utilizes nanostructured
parylene thin lms, fabricated via direct vapor-phase pyrolysis
and surface-induced oblique angle polymerization of [2.2]-p-
cyclophane precursors, as substrates for subsequent liquid phase
deposition
29,31
of conformal titania lms under aqueous process-
ing conditions amenable for use in a manufacturing environment.
The key process step involves selective physisorption onto the
nanostructured parylene surface of a titania-binding ligand
species prior to titania deposition. The presence of ligand at the
interface between the nanostructured parylene and titania lms
profoundly inuences the physicochemical properties of the
resulting composite, providing several advantages for our process.
For example, the use of a ligand permits control of the titania
adhesion strength in the composite by introducing a tunable
bonding pathway to complement the mechanical adhesive
component provided by the surface roughness associated with
the nanostructured parylene. Titania adhesion to the underlying
parylene is signicantly improved through the use of ligands that
covalently bind titania, with strongly binding ligands such as
phenylphosphonic acid providing titania lms that readily pass
the Scotch

tape adhesion test. The presence of non-covalently


adsorbed ligand on the nanostructured parylene does not
compromise adhesion in this case, consistent with known coop-
erative bonding effects for ensembles of non-covalently inter-
acting species
5961
and results from previous studies of analogous
metalSAM composites.
45,62,63
In fact, the use of a non-covalent
physisorption process to bind the ligand to the nanostructured
parylene surface also avoids the extra process step(s) and harsh
reaction conditions normally associated with chemical grafting
of ligands that can compromise the desirable thermal, electrical,
chemical, and mechanical properties of the parylene
30
required
for various applications.
The presence of a physisorbed ligand, in combination with the
nanostructured morphology of the parylene lm, also inuences
the structure of titania lm deposited. Titania lms containing
both crystalline anatase and amorphous TiO
x
(x # 2) phases are
observed using the hexauorotitanate-boric acid LPD bath
employed in our experiments. This behavior contrasts with the
deposition of anatase lms typically observed on SAMs
29,31
using
the same bath. The ability to consolidate the nanocrystalline
components of our titania lms via thermal annealing, which is
facilitated by the thermal stability of the underlying parylene
lms, without destruction of the composite represents an
important prerequisite for fabrication of materials potentially
useful for bioimplant, biomedical coating, biosensing, and tissue
culture and growth applications. We are continuing our studies
to measure and better understand and control the
physicochemical properties of these composites with the goal of
preparing materials having properties appropriate for realizing
such applications.
Acknowledgements
We thank the Ofce of Naval Research for nancial support for
this work through a 6.2 Research Program Award
(N000140710801) and Young Investigators Program Award for
MCD and the Naval Research Laboratory Core 6.1 Research
Program for WJD.
References
1 S. W. Kang, H. S. Yang, S. W. Seo, D. K. Han and B. S. Kim, J.
Biomed. Mater. Res., Part A, 2008, 85A, 747756.
2 B. Feddes, J. G. C. Wolke, W. P. Weinhold, A. M. Vredenberg and
J. A. Jansen, J. Adhes. Sci. Technol., 2004, 18, 655672.
3 M. Y. Shareef, P. F. Messer and R. Vannoort, Biomaterials, 1993, 14,
6975.
4 S. Kay, A. Thapa, K. M. Haberstroh and T. J. Webster, Tissue Eng.,
2002, 8, 753761.
5 T. J. Webster, C. Ergun, R. H. Doremus, R. W. Siegel and R. Bizios,
Biomaterials, 2001, 22, 13271333.
6 S. C. Huang, T. F. Lin, S. Y. Lu and K. S. Chou, J. Mater. Sci., 1999,
34, 42934304.
7 M. Cetinkaya, N. Malvadkar and M. C. Demirel, J. Polym. Sci., Part
B: Polym. Phys., 2008, 46, 640648.
8 M. C. Demirel, Colloids Surf., A, 2008, 321, 121124.
9 D. Klee, J. Boing and H. Hocker, Materialwiss. Werkstofftech., 2004,
35, 186191.
10 D. Wright, B. Rajalingam, J. M. Karp, S. Selvarasah, Y. B. Ling,
J. Yeh, R. Langer, M. R. Dokmeci and A. Khademhosseini, J.
Biomed. Mater. Res., Part A, 2008, 85A, 530538.
11 P. J. Chen, D. C. Rodger, R. Agrawal, S. Saati, E. Meng, R. Varma,
M. S. Humayun and Y. C. Tai, J. Micromech. Microeng., 2007, 17,
19311938.
12 J. M. Hsu, L. Rieth, S. Kammer, M. Orthner and F. Solzbacher, Sens.
Mater., 2008, 20, 87102.
13 J. Lahann, D. Klee, W. Pluester and H. Hoecker, Biomaterials, 2001,
22, 817826.
14 E. M. Robinson, R. Lam, E. D. Pierstorff and D. Ho, J. Phys. Chem.
B, 2008, 112, 1145111455.
15 J. Lahann, H. Hocker and R. Langer, Angew. Chem., Int. Ed., 2001,
40, 726728.
16 K. M. Vaeth and K. F. Jensen, Chem. Mater., 2000, 12, 13051313.
17 J. M. Moran-Mirabal, C. P. Tan, R. N. Orth, E. O. Williams,
H. G. Craighead and D. M. Lin, Anal. Chem., 2007, 79, 11091114.
18 G. R. Yang, Y. P. Zhao, J. M. Neirynck, S. P. Murarka and
R. J. Gutmann, J. Electrochem. Soc., 1997, 144, 32493255.
19 M. C. Demirel, E. So, T. M. Ritty, S. H. Naidu and A. Lakhtakia, J.
Biomed. Mater. Res., Part B, 2007, 81B, 219223.
20 M. Cetinkaya, S. Boduroglu and M. C. Demirel, Polymer, 2007, 48,
41304134.
21 S. Boduroglu, M. Cetinkaya, W. J. Dressick, A. Singh and
M. C. Demirel, Langmuir, 2007, 23, 1139111395.
22 P. Kao, N. A. Malvadkar, M. Cetinkaya, H. Wang, D. L. Allara and
M. C. Demirel, Adv. Mater., 2008, 20, 35623565.
23 M. C. Demirel, M. Cetinkaya, A. Singh and W. J. Dressick, Adv.
Mater., 2007, 19, 44954499.
24 W. J. Dressick, C. S. Dulcey, J. H. Georger, G. S. Calabrese and
J. M. Calvert, J. Electrochem. Soc., 1994, 141, 210220.
25 W. J. Dressick, L. M. Kondracki, M.-S. Chen, S. L. Brandow,
E. Matijevic and J. M. Calvert, Colloids Surf., A, 1996, 108, 101111.
26 D. I. Ma, L. Shirey, D. McCarthy, A. Thompson, S. B. Qadri,
W. J. Dressick, M.-S. Chen, J. M. Calvert, R. Kapur and
S. L. Brandow, Chem. Mater., 2002, 14, 45864594.
27 N. Malvadkar, S. Park, M. Urquidi-MacDonald, H. Wang and
M. C. Demirel, J. Power Sources, 2008, 182, 323328.
28 W. F. Gorham, J. Polym. Sci., Part A1, 1966, 4, 30273039.
This journal is The Royal Society of Chemistry 2009 J. Mater. Chem., 2009, 19, 47964804 | 4803
29 K. Koumoto, S. Seo, T. Sugiyama, W. S. Seo and W. J. Dressick,
Chem. Mater., 1999, 11, 23052309.
30 J. Frados, Modern Plastics Encyclopedia, McGraw Hill, New York,
vol. 45, (no 1A), 1968.
31 Y. F. Gao and K. Koumoto, Cryst. Growth Des., 2005, 5, 19832017.
32 D. Huang, Z. D. Xiao, J. H. Gu, N. P. Huang and C. W. Yuan, Thin
Solid Films, 1997, 305, 110115.
33 H. Shin, M. Agarwal, M. R. De Guire and A. H. Heuer, Acta Mater.,
1998, 46, 801815.
34 Y. Masuda, T. Sugiyama, H. Lin, W. S. Seo and K. Koumoto, Thin
Solid Films, 2001, 382, 153157.
35 Y. Masuda, Y. Jinbo, T. Yonezawa and K. Koumoto, Chem. Mater.,
2002, 14, 12361241.
36 Y. Masuda, T. Sugiyama and K. Koumoto, J. Mater. Chem., 2002,
12, 26432647.
37 Y. Masuda, T. Sugiyama, W. S. Seo and K. Koumoto, Chem. Mater.,
2003, 15, 24692476.
38 S. Yamabi and H. Imai, Chem. Mater., 2002, 14, 609614.
39 Y. F. Gao, Y. Masuda and K. Koumoto, Chem. Mater., 2004, 16,
10621067.
40 W. J. Dressick, M.-S. Chen and S. L. Brandow, J. Am. Chem. Soc.,
2000, 122, 982983.
41 B. D. Martin, S. L. Brandow, W. J. Dressick and T. L. Schull,
Langmuir, 2000, 16, 99449946.
42 W. J. Dressick, P. F. Nealey and S. L. Brandow, Proc. SPIEInt. Soc.
Opt. Eng., 2001, 4343, 294305.
43 S. L. Brandow, T. L. Schull, B. D. Martin, D. C. Guerin and
W. J. Dressick, Chem.Eur. J., 2002, 8, 53635367.
44 M.-S. Chen, W. J. Dressick, T. L. Schull and S. L. Brandow, Proc.
SPIEInt. Soc. Opt. Eng., 2002, 4608, 155159.
45 S. L. Brandow, M.-S. Chen, C. S. Dulcey and W. J. Dressick,
Langmuir, 2008, 24, 38883896.
46 F. A. Cotton, G. Wilkinson, C. A. MurilloandM. Bochmann, Advanced
Inorganic Chemistry, John Wiley & Sons, Inc., New York, 1999.
47 H. X. He, H. A. Zhang, Y. C. Wang and Z. F. Liu, Mol. Cryst. Liq.
Cryst., 1999, 337, 301304.
48 T. Z. Mengistu, V. Goel, J. H. Horton and S. Morin, Langmuir, 2006,
22, 53015307.
49 P. Pechy, F. P. Rotzinger, M. K. Nazeeruddin, O. Kohle,
S. M. Zakeeruddin, R. Humphry-Baker and M. Gratzel, J. Chem.
Soc., Chem. Commun., 1995, 6566.
50 X. J. Zhang, T. Y. Ma and Z. Y. Yuan, J. Mater. Chem., 2008, 18,
20032010.
51 K. Sarkar, S. C. Laha, N. K. Mal and A. Bhaumik, J. Solid State
Chem., 2008, 181, 20652072.
52 B. Adolphi, E. Jahne, G. Busch and X. D. Cai, Anal. Bioanal. Chem.,
2004, 379, 646652.
53 A. Cabeza, M. D. Gomez-Alcantara, P. Olivera-Pastor,
I. Sobrados, J. Sanz, B. Xiao, R. E. Morris, A. Cleareld and
M. A. G. Aranda, Micropor. Mesopor. Mater., 2008, 114, 322
336.
54 C. D. Wagner, W. M. Riggs, L. E. Davis and J. F. Moulder,
Handbook of X-ray Photoelectron Spectroscopy, The Perkin-Elmer
Corporation, Eden Prairie, Minnesota, 1979.
55 J. C. Dupin, D. Gonbeau, P. Vinatier and A. Levasseur, Phys. Chem.
Chem. Phys., 2000, 2, 13191324.
56 J. E. Goncalves, S. C. Castro, A. Y. Ramos, M. C. M. Alves and
Y. Gushikem, J. Electron Spectrosc. Relat. Phenom., 2001, 114116,
307311.
57 C. D. Terwilliger and Y. M. Chiang, J. Am. Ceram. Soc., 1995, 78,
20452055.
58 D. C. Hague and M. J. Mayo, J. Am. Ceram. Soc., 1994, 77, 1957
1960.
59 J. Elemans, A. E. Rowan and R. J. M. Nolte, J. Mater. Chem., 2003,
13, 26612670.
60 E. A. Meyer, R. K. Castellano and F. Diederich, Angew. Chem., Int.
Ed., 2003, 42, 12101250.
61 M. J. Serpe and S. L. Craig, Langmuir, 2007, 23, 16261634.
62 M.-S. Chen, S. L. Brandow, T. L. Schull, D. B. Chrisey and
W. J. Dressick, Adv. Funct. Mater., 2005, 15, 13641375.
63 W. J. Dressick, M.-S. Chen, S. L. Brandow, K. W. Rhee, L. M. Shirey
and F. K. Perkins, Appl. Phys. Lett., 2001, 78, 676678.
4804 | J. Mater. Chem., 2009, 19, 47964804 This journal is The Royal Society of Chemistry 2009

You might also like