Characterization of film failures by bismuth electrodeposition-Application to thin deformed fluorocarbon films for stent applications

Article abstract of DOI:10.1016/j.electacta.2009.09.079

A new approach for the detection and visualization of nano-scaled film defects, such as fissures and holes is described. The procedure is based on a selective electrochemical deposition of bismuth onto uncovered parts of stainless steel substrates due to film cracking after deformation. Cyclic voltammetry experiments enabled the identification of the bismuth redox signals, necessary for the subsequent potentiostatic electrodeposition. A combination of XPS and ToF-SIMS analyses demonstrated that deposition with commonly used acidified electrolytes led to a significant film degradation, by defluorination and chain scission. When an identical procedure was performed in neutral conditions, which was realized by the addition of a bismuth chelating agent, no influence onto the integrity of the fluorocarbon coatings was observed. By SEM, XPS and ToF-SIMS analyses, the deposition of bismuth on the film defects was clearly demonstrated and the failures were evidenced. The identification of nano-scaled coating failures, which was extremely difficult before using common characterization tools, was now realized. This work makes part of a strategy to cover stent materials with a protective fluorocarbon layer and to examine the influence of stent expansion onto the film cohesive properties.

Full text of DOI:10.1016/j.electacta.2009.09.079

Electrochimica Acta 55 (2010) 1042–1050
Contents lists available at ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Characterization of film failures by bismuth electrodeposition—Application to
thin deformed fluorocarbon films for stent applications
Servaas Holvoeta,1, Paula Hornya,1, Stephane Turgeona, Pascale Chevalliera, Jean-Jacques Pireauxb,
Diego Mantovania,∗
a Laboratory for Biomaterials and Bioengineering, Department of Materials Engineering & University Hospital Research Center, Laval University, Quebec City, QC, G1K 7P4, Canada
b Laboratoire Interdisciplinaire de Spectroscopie Electronique, FUNDP U. Namur, B-5000 Namur, Belgium
a r t i c l e i n f o
a b s t r a c t
Article history:
A new approach for the detection and visualization of nano-scaled film defects, such as fissures and
holes is described. The procedure is based on a selective electrochemical deposition of bismuth onto
uncovered parts of stainless steel substrates due to film cracking after deformation. Cyclic voltammetry
experiments enabled the identification of the bismuth redox signals, necessary for the subsequent poten-
tiostatic electrodeposition. A combination of XPS and ToF-SIMS analyses demonstrated that deposition
with commonly used acidified electrolytes led to a significant film degradation, by defluorination and
chain scission. When an identical procedure was performed in neutral conditions, which was realized by
the addition of a bismuth chelating agent, no influence onto the integrity of the fluorocarbon coatings
was observed. By SEM, XPS and ToF-SIMS analyses, the deposition of bismuth on the film defects was
clearly demonstrated and the failures were evidenced. The identification of nano-scaled coating failures,
which was extremely difficult before using common characterization tools, was now realized. This work
makes part of a strategy to cover stent materials with a protective fluorocarbon layer and to examine the
influence of stent expansion onto the film cohesive properties.
Received 23 April 2009
Received in revised form
30 September 2009
Accepted 30 September 2009
Available online 6 October 2009
Keywords:
Plasma deposited fluorocarbon coatings
Film cracking
Potentiostatic deposition
Bismuth
Surface analyses
© 2009 Elsevier Ltd. All rights reserved.
1. Introduction
barrier effectiveness, durability and stability) need to be evaluated
and preserved after its expansion, which could induce a local plas-
Coronary stents are medical devices widely used to open arter-
ies partially closed by atherosclerosis [1]. Coronary stents, mostly
made from 316L stainless steel, were shown to corrode in blood
after their insertion by balloon angioplasty in the human body [2].
In addition, they have proved to release potentially toxic ions, such
as nickel and chromium [3,4]. Therefore, coated coronary stents
have been introduced in order to promote their biocompatibility
and are nowadays approved by the Food and Drug Administration
(FDA) [5]. However, plastic deformations induced on the coated
stents during their clinical expansion have been shown to generate
coating defects, such as delamination and fissures [6,7]. As a conse-
quence, the FDA has provided guidance to the stent industry, which
implies that certain properties of the coating (such as adhesion,
tic deformation up to 25% [8–11]. However, the characterization of
nano-scale cracks and holes with common characterization tools is
not a trivial task and no specific tests have been proposed by the
FDA.
In earlier work, a multi-step process was previously developed
to cover the stainless steel surface with an ultra-thin uniform and
cohesive plasma polymerized fluorocarbon (Teflon-like) coating
[12]. Defects were extremely difficult to observe with scanning
electron microscopy (SEM) due to their low contrast and nanome-
ter size and moreover, their existence could not be proved with
X-ray photoelectron spectroscopy (XPS) due to the low sensitivity
of the latter technique [13]. Moreover, the formation of nanoscopic
defects at the interface and in the film could not be derived
directly by these surface analysis techniques [13]. However, time-
of-flight secondary ion mass spectroscopy (ToF-SIMS) and X-ray
photo-emission electron microscopy (XPEEM) experiments on the
deformed thin films yielded evidence of film failures (by the detec-
tion of metallic residues), but the lateral spatial resolution of the
techniques was insufficient for these defects to be visualized [13].
The paper proposes a new method to characterize the cracks
which may have been induced in the coating after deformation. The
method is based on an electrochemical reaction between metallic
∗
Corresponding author at: Laboratory for Biomaterials and Bioengineering,
Department of Mining, Metallurgy and Materials Engineering, Pav. Adrien-Pouliot,
1745-E, Laval University, Quebec City, QC, G1K 7P4, Canada.
Tel.: +1 418 656 2131x6270; fax: +1 418 656 5343.
E-mail address: Diego.Mantovani@gmn.ulaval.ca (D. Mantovani).
URL: http://www.lbb.gmn.ulaval.ca/ (D. Mantovani).
Both authors contributed equal to this paper.
1
0013-4686/$ – see front matter © 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2009.09.079

S. Holvoet et al. / Electrochimica Acta 55 (2010) 1042–1050
1043
Fig. 1. Working principle of selective deposition of bismuth through a hole formed
in the film.
Fig. 2. The mounting of a (deformed) specimen into the electrochemical cell.
ions and the metallic surface of uncoated stents, as schematized
in Fig. 1. Because the ions will preferentially interact with metallic
compounds rather than with an insulating surface, deposition of a
metallic compound onto a cracked film would allow the detection
of holes and fissures. Baumert et al. have previously demonstrated
the occurrence of cracks after the plastic deformation of silane
plasma-coated galvanized steel by reduction of cupper in acidic
solution [14,15]. Therefore, this Cu deposition enabled the char-
acterization of defect formation by SEM imaging which had not
been detected by surface analytical tools. However, in the case of
stainless steel as in the present project, the poor electrical proper-
ties of the chromium oxides which form the top layers of stainless
steel would not allow the reduction of any metallic compound. It is
therefore suggested to use an electrodeposition technique instead
of a reduction-oxidation reaction as performed by Baumert et al.
[14,15].
whereas in the case of Baumert et al. the reaction could only occur in
acidic medium which could lead to film degradation and oxidation.
In this work, cyclic voltammetry measurements on electropol-
ished steel enabled the identification of the bismuth redox signals,
in order to assess the optimal conditions for the subsequent poten-
tiostatic electrodeposition. Subsequently, a combination of XPS and
ToF-SIMS analyses was applied in order to study the influence of the
deposition parameters onto the film integrity. Finally, film failures
caused by plastic deformation were visualized by SEM by means of
the described electroplating method.
2. Materials and methods
2.1. Equipment
Bismuth ions have been chosen as the metallic component of the
electrolyte because of its high atomic sensitivity factor for XPS (9.1
versus 2.9 for Fe and 0.7 for O in our system), meaning that lower
concentrations of bismuth could be more accurately detected [16].
The evaluation of the amount of deposited bismuth would enable
quantification of the exposed substrate area due to film failures. For
SEM, the high differences in atomic number between the metallic
substrate (Z(Fe) = 26), the coating (Zaverage ∼7) and bismuth (Z = 83)
would enhance the contrast between a small volume of deposited
bismuth and the background.
Up to date, bismuth electrodeposition has been largely inves-
tigated because of its unusual electronic properties [17]. Many
authors have reported the electroplating of bismuth or bismuth
oxide thin films on various substrates in acidic electrolyte solu-
tions [18–28]. The choice of the Bi3+ concentration and the pH of the
solution have been made in all of these works considering the sta-
bility of the Bi ions in solution according the Pourbaix diagram [29].
Indeed, the Bi salt (Bi(NO3)3·5H2O) can be dissolved only in nitric
acid and precipitates at pH exceeding 4.5. Furthermore, almost all
reported works in the literature were carried on in acidic solutions
but with no mention of the pH [18,22,24–26,28,30] except three
works: Gujar et al. at pH 2.0 [30] and Wang and Lu as well as Hut-
ton et al. at pH 4.5 [20,31,32]. Moreover, two works were carried out
using chelating agents of Bi3+ ions, such as triethanolamine (TEA)
[22] andethylenediaminetetraaceticacid (EDTA) [21] but no details
onto the pH were given.
The plasma films were deposited onto 12.7 and 0.5 mm
thickness disk-shaped SS316L electropolished substrates using
a custom-built plasma reactor [33]. Briefly, the electropolished
sample was introduced in a radio-frequency (13.56 MHz) plasma
reactor at a distance of 6 cm below the discharge zone for a
pulsed H2 plasma etching using 100 W pulses of 25% duty cycle
(ton = 100 ms; toff = 300 ms), at a pressure of 0.665 mbar and a gas
flow rate of 10 sccm during 100 s. Subsequently, pulsed afterglow
plasma polymerization was carried out in the same reactor at 11 cm
below the discharge, produced by 150 W pulses of 5% duty cycle
(ton = 5 ms; toff = 90 ms), at a pressure of 0.930 mbar, a gas flow rate
of 20 sccm (94% C2F6 and 6% H2) and a depositon time of 5 min
[12,13].
The thickness of the film before deformation was evaluated by
ellipsometry at 35 nm [13]. The coated samples were plastically
deformed at 25% using a small-punch test device mounted on a
SATEC T20000 testing machine (Instron, Norwood, MA, USA) at
room temperature at a displacement rate of 0.05 mm s−1 and a load
of 2200 N [12]. All subsequent analyses, SEM, XPS and ToF-SIMS,
were performed on the topmost part of the samples. The specimen
is imbedded in Teflon mounting to expose only a small part of the
surface. The exposed area is determined by a 4 mm I.D. Viton O-ring
(Fig. 2).
All electrochemical measurements and treatments were per-
formed with a 273A and a Versastat V3 potentiostat (Princeton
Applied Research, Oak Ridge, TN, USA) in connection with a per-
sonal computer. A coiled platinum wire (15 cm length and 0.3 mm
in diameter) was used as the working electrode and an Ag/AgCl–KCl
saturated electrode as the counter electrode (+197 mV versus SHE).
All potentials were given in reference to the counter electrode.
Electrodeposition was performed in 500 mL of electrolyte in a
custom-built cell with a 10 min argon purge before the electrode-
position.
Considering all these parameters, our original approach com-
pared to Baumert et al. is to use Bi electrodeposition as an easy
tool in order to characterize nanodefects or/and nanofailures of
films deposited onto any metallic substrates. Indeed the technique
of Baumert et al. cannot be applied onto stainless steel substrate
while it was based onto a reduction–oxidation process. Due to the
reduction–oxidation potential of iron, cupper can not be reduced
by it. Chromium being already fully oxidized will not reduce cup-
per either. Furthermore, the pH of the complexed Bi solution can
be modified in order to be applied to different types of coating,
Surface imaging was carried out by scanning electron
microscopy (JEOL JSM-35CF and JEOL JSM6360LV, Jeol, Tokyo,

1044
S. Holvoet et al. / Electrochimica Acta 55 (2010) 1042–1050
Japan) using secondary electron images collected by an
Everhart–Thornley detector and at beam energies ranging between
5 and 15 keV. The computation of the area fraction of the white Bi
crystals was carried on using the ImageJ software [34,35].
and stirred for 1 h. The volume of the solution was increased up to
900 mL, by adding DI water. Subsequently, a 30% NH4OH solution
(Baker Chemical Co., Phillipsburg, NJ, USA) was added drop wise till
the pH reached the pKa region of NH3/NH4+ (pH ∼9.25). Finally the
pH of the solution was brought back to 7.4 by the drop wise addition
of diluted HNO3 and the total volume was adjusted to 1 L with DI
water, yielding a clear, fully transparent Bi3+(EDTA) solution.
The surface chemical composition was analyzed by X-ray photo-
electron spectroscopy. Three samples per treatment were analyzed
and three different spots were examined on each sample. These
analyses were performed using an X-ray photoelectron spectrom-
eter (XPS-PHI 5600-ci Spectrometer-Physical Electronics, Eden
Prairie, MN, USA). Charge compensation was not necessary. XPS
spectra were acquired at a detection angle of 45◦, using the K␣ line
of standard aluminium (K␣ = 1486.6 eV) X-ray source for the survey
analyses and the magnesium one (K␣ = 1253.6 eV) for high resolu-
tion analyses, both were operated at 300 W. The survey data were
obtained from Multipak program (Matlab) by integrated each peak
area and taking into account the sensibility factor of each element:
C 1s 0.296, F 1s 1.00, O 1s 0.711 and Bi 4f 9.14. The curve fitting for
the high resolution C 1s peaks was determined by means of a least
squares peak fitting procedure with a Shirley-type background.
Five spectral components were used, representing the five major
types of chemical environments of a carbon atom in a fluorocarbon
structure: C–C/C–H (hydrocarbon), C–CFx, CF, CF2 and CF3. All com-
ponents were assigned with the same full width at half-maximum.
Their respective positions were not fixed during the whole fitting
process. Assuming the BE of C–C/C–H to be 286.0 eV, the mean BEs
obtained for the different C 1s components were 287.2 eV (C–CFx),
289.4 eV (C–CFx), 292.0 eV (CF2) and 294.0 eV (CF3). These values,
and the corresponding assignments, agreed well with data pub-
lished in the literature [13,36]. Imaging of the 4f peak of Bi was
acquired with a standard Al K␣ X-ray source at a power of 300 W.
The detection angle was 45◦ and the pass energy 30 eV. The image
area was 0.5 × 0.5 mm2, with a spatial resolution of 30 ␮m and a
numeric resolution of 128 × 128 pixels (∼15 ␮m/pixel).
3. Results and discussion
Considering the literature review on the pH of Bi3+ solutions
[18–28], three different pH were considered in this study: one at
pH 1.0 (highly acidic solution), one at pH 4.0 with acetate buffer and
a last one at pH 7.4 with EDTA as chelating agent. In preliminary
tests however, the highly acidic electrolyte (pH 1.0) was extremely
harmful to the ultra-thin fluorocarbon film as XPS analyses exhib-
ited metallic compounds. So only the two Bi solutions at higher pH,
4.0 and 7.4, respectively, have been selected.
Another important factor in electrodeposition process is the
electrolyte concentration: too low concentrations led to few nucle-
ations whereas higher concentrations caused three-dimensional
growth. Sadale and Patil [22] have proposed a mechanism of depo-
sition steps for the formation of a Bi thin film by electrodeposition.
The analysis of the chronoamperometry curves showed that below
0.1 mM, the region corresponding to critical nuclei formation was
delayed or missing due to the ohmic drop of the solution with
decreasing concentration, which therefore significantly decreased
the nucleation rate. Above 10 mM, 3D growth is controlled by the
diffusion which was not the case for lower concentrations [22].
The electrodeposition potential is also a significant parameter
to obtain the expected Bi deposit. In the reported works, most
electrodepositions have been carried on between −255 mV [22,25],
−455 mV [21], up to −455 mV [26] or at −1 V [20]. In this work, the
choice of the potential was determined by considering the reduc-
tion potential of Bi observed by cyclic voltammetry, the oxidation
potential of Bi [29], as well as the density of nuclei and their growth,
in order to cover the defects as densely as possible but to remain
within their edges. The corrosion potential of the substrate was also
considered to avoid pickling formation.
ToF-SIMS analyses were realized with a ToF-SIMS IV (ION-ToF
GmbH, Münster, Germany) with a 25 keV Ga+ analysis gun. The gal-
lium current was 1 pA, and the beam hit the target at an angle of 45◦,
and the primary ion dose density was below the static SIMS limit of
1013 ions/cm2 [37,38]. Ions used for the mass calibration were H+
(m/z = 1), C+ (m/z = 12) and C2F5 (m/z = 119). For spectral acquisi-
+
tion, positive ion mass spectra were acquired from a 500 × 500 ␮m2
area, and mass fragment ratios were directly calculated from mass
intensities. ToF-SIMS in negative mode is not detailed in this paper
as degradation/defluorination processes which could be detected
by CxFy moieties are very well detected in the positive mode [39].
Therefore, in this work, the appropriate Bi electrodeposition
parameters such as pH, electrolyte concentrations, and deposi-
tion potential have been chosen considering this literature review.
Moreover, the deposition time has been determined considering
the stability of the polymer film in the electrolyte as well as the
nucleation rate and the crystal growth.
2.2. Products
Bismuth was electrodeposited from aqueous solutions of bis-
muth nitrate (Bi(NO3)3·5H2O, ACS, 98%, Alfa Aesar, Ward Hill, MA,
USA) of different concentrations (10, 1 and 0.1 mM).
3.1. Bi3+ versus Bi3+(EDTA) solutions
The conditions of electrodeposition (pH and electrolyte concen-
tration) were determined from the consideration of the polymer
film stability, of the literature and of the Pourbaix diagram of bis-
muth in order to avoid the formation of undesirable components
such as bismuth hydrides and/or bismuth oxides [29,40]. The solu-
bility of Bi3+ cations is limited to acidified solutions only, therefore
a chelating agent was added in order to complex it and, as a conse-
quence, dissolve the Bi3+ at higher pH values. In literature, the use
of triethanolamine and ethylenediamine tetraacetic acid (EDTA) is
described for complexing metallic cations, such as Bi3+, in order
to vary the morphology of the electroplated metal [21,22]. In this
work, EDTA was selected to complex the Bi3+ and its chelating prop-
erties were applied to prepare Bi3+ solutions around neutral pH. At
pH 7.4, EDTA exists mainly in its zwitterionic form, with the two
tertiary amines fully protonated. In this way, the nucleophilic char-
acter of the EDTA has seriously decreased, which was assumed not
to cause any potential harm to the fluorocarbon coatings. In what
A 0.1 M Bi3+ stock solution was obtained by dissolving 4.9 g
(10 mmol) of the salt into 100 mL of diluted nitric acid (HNO3, ACS,
70%, Laboratoire MAT, Quebec, QC, Canada) (17% HNO3/DI water
(v/v) %) [19]. The mixture was stirred until complete dissolution of
the salts. For electrodeposition from acidified Bi3+ solutions, this
stock solution was then further diluted with DI water and a 1 M
acetate buffer to obtain the 10, 1 and 0.1 mM Bi3+ solutions at pH 4
(with a final buffer concentration of 0.1 M). The 1 M acetate buffer
was prepared by dissolving 31.73 g of NaOAc (ACS, Fisher Scientific,
Fair Lawn, NJ, USA) into 49 mL of glacial HOAc (ACS, ACS, Fisher Sci-
entific, Fair Lawn, NJ, USA) and the final volume was adjusted to 1 L
with DI water (pH 4.0).
The Bi3+(EDTA) solution was prepared from the initial 0.1 M Bi3+
stock solution. 1.01 equivalent of ethylenediamine tetraacetic acid
disodium salt (EDTA H2Na2, Baker Chemical Co., Phillipsburg, NJ,
USA) (0.101 mol, 3.76 g) was added to the acidified Bi3+ solution

S. Holvoet et al. / Electrochimica Acta 55 (2010) 1042–1050
1045
Fig. 3. Cyclic voltammetry curves on non-coated electropolished stainless steel for an acidified Bi3+ solution (left) and a neutral Bi3+(EDTA) solution (right), both at 10 mM.
follows the two solutions, i.e. Bi3+ (pH 4.0) which is the maximum
allowable pH without chelating the Bi cations and Bi3+(EDTA) (pH
7.4) were applied for electroplating purposes.
face showed a high density of pickling indicating the corrosion of
the substrate above −0.75 V. Below −0.9 V, reduction of water was
dominant. Moreover, by considering the smallest Bi grain shape
and highest nuclei density, as well as the lowest fluorocarbon
film damage, the choice of −0.85 V as deposition potential for the
Bi3+ electrolyte was permitted. Moreover, the SEM study of the
fluorocarbon film with Bi deposition times ranging between 10 s
and 3 min as well as the exposed metallic atomic fraction by XPS
allowed to estimate that a 10 s deposition was sufficient to favor Bi
grain formation without altering the defect shape. Thus, the elec-
trodeposition parameters were fixed at a deposition time of 10 s
and a potential of −0.85 V.
3.2. Cyclic voltammetry
As explained above to favor the nucleation process rather than
the 3D growth, the maximum Bi concentration should be 10 mM.
Whereas if the concentration is to low (<0.1 mM), the nucleation is
too delayed. Therefore the electrochemical study was carried on at
the maximum and minimum concentrations of 0.1 and 10 mM of
Bi3+ ions.
Potentiostatic experiments were performed on electropolished
flat specimens, changing the solution and bath concentration as
variables. The potentiostatic curves (not shown) proved that both
critical nuclei and cluster formation occurred within the first sec-
onds and were diffusion controlled [22]. After a few seconds,
the current reached a constant value, indicating that ion migra-
tion and electrodeposition were in dynamic equilibrium [21].
For 10 mM Bi3+ electrolytes, bismuth growth was diffusion con-
trolled, whereas for lower concentrations (0.1 mM) the deposition
of bismuth was concentration controlled, due to the loss of ionic
availability [22].
After XPS survey analyses, the XPS Bi 4f signal was applied
in order to obtain the amount of deposited bismuth. Fig. 4 dis-
plays the percentages of bismuth present on the steel surfaces,
combined with the charges which were consumed during the
reduction process (from Bi3+(EDTA) solutions). In order to observe
a trend, a third intermediate concentration at 1 mM Bi3+ has been
added to the 10 mM Bi3+ and 0.1 mM Bi3+ studied concentrations.
Fig. 3 shows the cyclic voltammetric curves obtained on
non-coated electropolished steel for Bi3+ at pH 4.0 (a) and its
corresponding EDTA complex at pH 7.4 (b), both at 10 mM. The
voltammograms at 0.1 mM did not present any peak due to the low
concentration of the cations (not shown) [26]. The voltammograms
depicted the cathodic and anodic peaks from 0.5 to −1.0 V and back
to 0.5 V at a constant sweep rate of 0.01 V/s. The bismuth oxidation
signal was rather similar for both solutions. For non-complexed bis-
muth, two clear peaks were localized at the reduction potential of
bismuth, at 0.1 V for the Bi3+ to Bi+ reduction and at −0.23 V for the
Bi+ to Bi reduction [21]. The occurrence of the cross-over between
the anodic and cathodic curves is characteristic of the formation of
bismuth nuclei [27]. For Bi3+(EDTA) complexes, only one reduction
signal occurred at −0.8 V, and in addition, the reduction current
had decreased slightly. The addition of EDTA shifted the reduction
of Bi3+ to a more negative potential. This phenomenon was also
observed in the work of Jiang et al., where the addition of EDTA
caused the reduction potential to drop to more negative values [21].
This negative shift of the reduction potential was ascribed to the
change in conditional electrode potential caused by the formation
of the Bi3+ complex with EDTA. It could be argued that the −0.8 V
signal should rather be assigned to the reduction of hydrogen than
to bismuth reduction. However, as the proton concentration in this
solution did not exceed 10−7 M, this degree of acidity is too low to
cause a reduction signal with such intensity, as compared to the
bismuth oxidation signal (CBi = 10−2 M). In addition, as the latter
solution is a non-buffered medium, a change in proton concentra-
tion would translate into a change in pH after the CV experiments,
an observation which could not be made by subsequent pH analy-
ses. Another hypothesis in which water was reduced into hydroxyl
anions at −0.8 V is rejected as well, for the same reason mentioned
before. No increase in pH was observed after electrodeposition.
3.3. Potentiostatic deposition
Fig. 4. Percentage of Bi deduced from XPS survey and charge data for the potentio-
static electrodeposition onto non-coated electropolished stainless steel substrates
from Bi3+(EDTA) electrolyte solutions at different concentrations.
Potentiostatic electrodepositions were performed at potentials
ranging from −0.4 to −0.9 V. The SEM observation of the steel sur-

1046
S. Holvoet et al. / Electrochimica Acta 55 (2010) 1042–1050
Fig. 5. SEM images of deposited bismuth onto non-coated electropolished stainless steel from a 10 and 0.1 mM electrolyte solution.
Increasing the initial bismuth concentration made the amount of
electroplated bismuth increase as well. This was confirmed by the
charge data, where more milliCs were consumed with an increasing
electrolyte concentration. Non-chelated Bi3+ solutions displayed
similar trends (not shown).
to 3.6 0.1 at% Bi when the detection angle was decreased to 15◦
(grazing angle) during the XPS survey (depth of analysis for XPS
is a few nanometers). In these specific electroplating conditions,
the formation of a very thin and continuous metallic bismuth film
was therefore assumed. Extra evidence for the latter was yielded
by covering the half of an electropolished sample with polyimide
tape, which was removed after the potentiostatic deposition. Sub-
sequently, after cleaning the specimen with acetone, XPS imaging
along the deposition edge was performed. The distribution of the
Bi 4f signal over the surface, as displayed in Fig. 5, indicated the
presence of a metallic bismuth film, which could not be visualized
with SEM due to the thin character of the latter.
Since the objective of the electrodeposition consists of a com-
plete filling of the cracks in the coating, a high density of nuclei
was preferred to growth. For this reason, all further experiments
were carried out with 0.1 mM solutions. Fluorocarbon polymer
coated specimens, flat and 25% deformed, were subjected to a
potentiostatic deposition (for 10 s) using the two bismuth solutions,
The SEM micrographs in Fig. 5 show the morphology of the
metallic bismuth, deposited in different conditions. For the more
concentrated 10 mM Bi3+ solutions, both chelated and not chelated
with EDTA, the growth was preferentially three-dimensional, as
could be observed in the SEM image at 10,000× magnification. On
the other hand, germination of nuclei was promoted at lower con-
centrations of electrolyte (0.1 mM). For non-complexed Bi3+, the
stainless steel was covered with a dense pattern of circular germs,
as visible in the 18,000× magnification image. These nuclei could
not be visualized for the Bi3+(EDTA) specimens, even at higher
magnifications (depth of analysis for SEM is 1 ␮m). However, XPS
analysis proved the bismuth to be electroplated onto the steel at a
concentration of 0.9 0.2 at% Bi, a value which even increased up

S. Holvoet et al. / Electrochimica Acta 55 (2010) 1042–1050
1047
the plastic deformation indicating that the surface chemistry has
remained stable, in agreement with earlier observations [12]. Depo-
sition from an acidified Bi3+ solution however yielded the extra Bi
4d and Bi 4f photoelectron signals onto the survey (1.0 0.2 at%
Bi), which would seem to indicate a highly exaggerated exposed
surface area. In addition, when focusing onto the background, its
behaviour at higher binding energies (lower kinetic energies) was
different from the other films, which indicated an increased emis-
sion of photoelectrons from the underlying metallic stainless steel
substrate, as in Tougaard model [41]. Indeed the Tougaard model
demonstrated that the intensity of the background of inelastically
scattered electrons depends strongly on the thickness of the over-
lay. This decrease of film thickness would be due to the rather
aggressive deposition conditions, acidic medium pH 4.0, of the lat-
ter sample [42]. In addition, the F/C ratio calculated for this sample
dropped significantly compared to all other samples. This phe-
nomenon of defluorination could be due to different degradation
mechanisms such as cross-linking and/or coating oxidation [43].
Extra evidence for the fact that film oxidation has occurred, is the
presence of oxygen (9.5 at%), clearly visible on the survey spectrum.
It should be mentioned that no nitrogen was detected into or onto
the fluorocarbon films, in this way providing the assumption that
the EDTA auxiliary complex did not precipitate during the reduc-
tion process. Finally, the survey obtained for electroplated samples
from Bi3+(EDTA) on a 25% deformed sample resulted into the detec-
tion of bismuth onto the coating (0.2 0.0 at% Bi), indicating that
approximately 0.2% of the substrate was exposed by film defects.
Apparently, the chemical composition of the film was not influ-
enced by the electrochemical procedure, as the composition was
similar to that of a non-treated film.
Fig. 6. XPS survey data for 35 nm plasma deposited fluorocarbon coatings. XPS data
are presented in arbitrary units (a.u.).
chelated and non-chelated (0.1 mM). The chemical composition of
the nano-thin films after the electrochemical procedure was stud-
ied with XPS and ToF-SIMS.
3.4. XPS analysis
Fig. 7 presents the C (1s) high resolution scans, corresponding
with the C (1s) signal from the surveys displayed in Fig. 6. The
regions in the C (1s) spectra were fitted with five spectral com-
ponents assigned to –CF3 (294 eV), –CF2 (292 eV), –CF (289.4 eV),
–C–CF (287.2 eV) and C–C (286 eV) and their atomic contributions
(in percent) are displayed above each signal [44].
Whereas after plastic deformation, no variation in the initial
coating composition was observed, in agreement with previ-
ous works [12], deposition of Bi3+ in acidified medium caused a
dynamic change in C (1s) signal distribution. Indeed, an important
decrease in CF moieties (CF3, CF2) was observed, and a new band
assigned to C–C/C–H (at 286 eV) became the major C component
(around 40%). Moreover, the slight increase of the band at 287.2 eV
correlated with the 9.5% of O in survey can be explained by the for-
mation of C–O moieties due to film oxidation. These C–C (and C–H)
moieties were not detected by using the less aggressive Bi3+(EDTA)
solution. The absence of degradation/oxidation in these latter elec-
XPS characterizations were performed in order to study and
quantify in first hand the deposition of bismuth into the fluo-
rocarbon film defects. In second hand, XPS could yield valuable
information about the influence of the electrochemical procedure
onto the surface composition of the films. This latter issue was of
huge importance as different experimental parameters (pH, volt-
age and deposition time) could be expected to be responsible
for film cracking and/or delamination during the entire proce-
dure, which could lead to an over-estimation of the original
defects.
Typical XPS surveys on 35 nm coatings are displayed in Fig. 6
before and after deformation, and after bismuth electroplating with
the two solutions. In all spectra, no metallic compounds (Cr, Fe)
were observed and due to the XPS sensitivity it could be esti-
mated that less than 1% of the substrate was exposed through
defects in the films. The first two spectra do not show any signifi-
cant change in chemical composition of the fluoropolymer due to
Fig. 7. XPS C 1s high resolution data for 35 nm plasma deposited fluorocarbon coatings.

1048
S. Holvoet et al. / Electrochimica Acta 55 (2010) 1042–1050
Table 1
ToF-SIMS signal ratios for 35 nm plasma deposited fluorocarbon coatings.
Flat
25% Deformed
25% Deformed
Bi3+ (pH 4.0)
25% Deformed
Bi3+ (EDTA) (pH 7.4)
End group/end group
C2F5/CF3
C3F7/CF3
0.7
0.4
0.6
0.3
0.2
0.0
0.7
0.4
End group/polymer fragment
CF3/C4F7
C2F5/C4F7
C3F7/C4F7
4.7
3.2
1.6
6.0
3.4
1.6
49.0
8.0
2.0
4.4
3.2
1.6
acquired spectra, yielded valuable information about the polymeric
nature of the surface and the variation in film composition during
the treatments (Table 1). When focusing on the ratio of different
end groups (CnF2n+1) to CF3 (as shortest possible chain termination),
more C2F5 fragments were measured compared to C3F7 fragments
for all different samples, in agreement with the statistic-structural
composition of plasma films. These ratios for flat specimens did not
show large variation after 25% plastic deformation, as confirmed
in earlier work [13]. After deposition from the acidified Bi3+ solu-
tion, these ratios dropped significantly indicating an increase in the
percentage of CF3 terminations, probably through the breakage of
longer CF2 chains. Electroplating from the Bi3+(EDTA) solution did
not prove any significant chain scission to have occurred. When
the ratios of different end groups (CnF2n+1) to the long polymer
chain fragment (C4F7 = CnF2n−1) were compared, similar conclu-
sions could be drawn. Only in case of the more aggressive Bi3+
electroplating solution, the presence of the high mass fragment
had decreased seriously, results which were consistent with the
ratios discussed above and could be explained by breakage of the
polymeric chains during the electrochemical deposition. In this
way, the ToF-SIMS trends thus confirmed the XPS results discussed
earlier.
Fig. 8. ToF-SIMS spectra for 35 nm plasma deposited fluorocarbon coatings.
troplating conditions could probably be correlated directly to the
lower degree of acidity.
3.5. ToF-SIMS analysis
3.6. Characterization of film defects induced by plastic
deformation
Fig. 8 presents some typical ToF-SIMS spectra obtained from the
35 nm thin fluorocarbon films before and after plastic deformation,
and after the bismuth electroplating procedure. All spectra con-
tained a series of signals at m/z = 31, 81, 131 and 181 which could
be assigned to CnF2n−1 mass fragments, such as CF, C2F3, C3F5, and
C4F7, respectively. These fragments were due to scission of CF2
chains as described earlier [13,45]. Plasma deposited fluoropoly-
mers tend to contain a mixture of components due to cross-linking
within the structure. Aligned long chains of polymer molecules
were evidenced through the mass fragments relating to CnF2n+1
such as m/z = 69, 119, 169 and 219 assigned to CF3, C2F5, C3F7,
and C4F9, respectively (end groups). Additionally, more complex
fragments were present relating to other fluorocarbon molecules
and hydro fluorocarbon molecules. Both sets of mass fragments
were visible in the four spectra, however specific varieties in inten-
sity could be observed among them. In case of bismuth-treated
specimens, where deposition was performed at low pH (pH 4.0),
the absence of higher molecular weight polymer fragments was
observed, indicating serious degradation due to the procedure.
When, however, bismuth complexes were applied, the signal distri-
bution displayed a similar trend compared to the flat, non-treated
sample, pointing out that the film stayed rather intact by the elec-
trochemical treatment. It must be emphasized that for both two
solutions, bismuth had deposited onto the coatings into its metallic
form (at m/z = 209). Analogously to the XPS results, higher quanti-
ties of bismuth were measured onto films electroplated from pH
4.0, indicating again a highly exaggerated exposed surface area.
More quantitatively, the analysis of the relative signal intensities
and ratios of the different coating fragments, calculated from the
Potentiostatic bismuth deposition was applied on 35 nm thin
coated specimens, before and after a 25% plastic deformation, with
the two different bismuth solutions. For both electrodeposition
solution, acidic (pH 4.0) and neutral (pH 7.4), all optimized parame-
ters such as the reduction voltage, Bi concentration and deposition
time were kept constant: −0.85 V, 0.1 mM and 10 s respectively.
Bismuth deposition was visualized by SEM, as presented in Fig. 9.
Before deformation, no coating defects such as holes were visible
onto the surface, independent from the applied solution. XPS anal-
yses on these samples did not detect any bismuth (<0.1 at% Bi).
After plastic deformation, the slip bands were clearly visible and
no delamination of the films was observed. For both solutions, the
deposition of bismuth onto deformed samples had occurred along
the slip bands as observed in Fig. 9.
The choice of the deposition parameters, and deposition time
in specific (10 s), had proved to be sufficient in order to reduce
bismuth into the nano-scaled film failures. The three-dimensional
growth of the Bi grains was rather limited, as was initially postu-
lated, and a high density of Bi nuclei was obtained into the cracks
of the deformed film.
For Bi electrodeposition in acidic medium, the coverage of the
slip bands by Bi grains was evident, whereas for Bi3+(EDTA) treated
films the Bi was only detected after screening the surface for
defects. The density of deposited bismuth was thus much lower
in neutral electroplating conditions (as measured with XPS: 0.2 at%
versus 1.0 at% at pH 4.0), indicating that the fissures present on
pH 4.0 treated specimens was increased by degradation due to the

S. Holvoet et al. / Electrochimica Acta 55 (2010) 1042–1050
1049
Fig. 9. SEM micrographs for bismuth-electroplated plasma deposited fluorocarbon coatings.
applied solution. In addition to the XPS results, bismuth quantifica-
tion was attempted directly from the 2500× SEM images in Fig. 9, by
computing the area fraction of white bismuth crystals in the entire
image area. Bi3+(EDTA) treated films yielded 0.4% of the area cov-
ered with bismuth grains versus 2.1% for films manipulated with
the acidified solution, which confirmed the ratio obtained with XPS
between the two methods. The latter SEM percentages are obvi-
ously an over-estimation of the amount of exposed steel surface,
as a zoom had to be made onto the regions of interest in order for
the bismuth grains to cover more then a few pixels. Moreover, diffi-
culty in the differentiation of bismuth grains and topographic sharp
features could not be completely overcome.
35 nm films might be due to the decrease in film thickness on these
slip bands where dissolution of the coating occurred which was
complete for the thinnest regions (slip bands). In the case of elec-
troplating performed in conditions which had no influence onto the
chemical composition, as proven with XPS and ToF-SIMS earlier,
the deposited bismuth could be directly assigned to the presence
of film defects induced by the plastic deformation. Thus due to the
latter, we can now affirm that around 0.2% of the stainless steel
surface had become uncovered, whereas we previously could only
affirm that the uncovered area was <1%. This new method enabled
the visualization of nano-scale defects onto a surface, which could
hardly be detected with common characterization tools.
Note that no bismuth deposition was observed on the grain
boundaries for these thin films. Apparently, these coatings were
able to follow the plastic deformation onto these high deformation
areas, by reducing the film thickness onto the slip bands, but not in
that extent to fully uncover the substrate. In the case of acidic elec-
trodeposition, the insufficient resistance to film failures for these
Current works involves the application of the presented method
onto films differing in film thickness (up to a few hundreds of
nanometers). In addition, this technique could be helpful to elu-
cidate the mechanism of film cracking by a bismuth electroplating
study on specimens with increasing orders of deformation magni-
tude.

1050
S. Holvoet et al. / Electrochimica Acta 55 (2010) 1042–1050
4. Conclusions
[8] C. Dumoulin, B. Cochelin, J. Biomech. 33 (2000) 1461.
[9] F. Migliavacca, L. Petrini, M. Colombo, F. Auricchio, R. Pietrabissa, J. Biomech.
35 (2002) 803.
This work presented a new method for the characterization of
nano-scale defects into thin fluorocarbon films, as part of a strategy
in the development of new adherent stent coatings. The approach
consisted in a selective electrochemical bismuth deposition onto
the exposed stainless steel substrates through the coatings fail-
ures. A rational screening of different deposition parameters such
as pH, electrolyte concentration, deposition voltage and time was
presented in order to obtain a non-destructive deposition proce-
dure during which nucleation of bismuth onto the stainless steel
was optimal. Using a combination of XPS and ToF-SIMS analysis, it
was demonstrated that bismuth electrodeposition from a neutral
pH solution had no influence onto the surface composition of the
films. This was not the case when deposition was performed from
an acidified solution, as XPS and ToF-SIMS fragments indication
chain scission and defluorination.
[10] F. Migliavacca, L. Petrini, V. Montanari, I. Quagliana, F. Auricchio, G. Dubini,
Med. Eng. Phys. 27 (2005) 13.
[11] J. Stolpmann, H. Brauer, H.J. Stracke, R. Erbel, A. Fischer, Materialwiss. Werk-
stofftechn 34 (2003) 736.
[12] F. Lewis, P. Horny, P. Hale, S. Turgeon, M. Tatoulian, D. Mantovani, J. Phys. D:
Appl. Phys. 41 (2008) 045310.
[13] P. Hale, S. Turgeon, P. Horny, F. Lewis, N. Brack, G. Van Riessen, P. Pigram, D.
Mantovani, Langmuir 24 (2008) 7897.
[14] B. Baumert, M. Stratmann, M. Rohwerder, Zeitschrift Metall. 95 (2004) 447.
[15] B. Baumert, M. Stratmann, M. Rohwerder, Mater. Res. Soc. Symp. Proc. 795
(2003) u.5.23.1.
[16] C.D. Wagner, L.E. Davis, M.V. Zeller, J.A. Taylor, R.H. Raymond, L.H. Gale, Surf.
Interface Anal. 3 (1981) 211.
[17] G.E. Smith, J.M. Rowell, G.A. Baraff, Phys. Rev. 135 (1964) A1118.
[18] J. Ismail, M.F. Ahmed, P.V. Kamath, J. Electroanal. Chem. 354 (1993) 51.
[19] S.I. Córdoba deTorresi, I.A. Carlos, J. Electroanal. Chem. 414 (1996) 11.
[20] E.A. Hutton, B. Ogorevc, S.B. Hocevar, F. Weldon, M.R. Smyth, J. Wang, Elec-
trochem. Commun. 3 (2001) 707.
[21] S. Jiang, Y.H. Huang, F. Luo, N. Du, C.H. Yan, Inorg. Chem. Commun. 6 (2003)
781.
35 nm thin coatings proved to be sufficiently adherent in order
to follow the plastic deformation onto high deformation areas such
as the grain boundaries. The film thickness onto the slip bands was
however reduced, were small film cracks occurred. Defects into
the film, induced by a plastic deformation and which could not be
measured by common characterization tools, could now easily be
visualized by SEM due to the high chemical contrast of the electro-
plated bismuth. In addition, failure percentages can now directly
be quantified by XPS. In case of our fluorocarbon films, the amount
of exposed surface through fissures raised up to 0.2%. This method
presents a new and original electrochemical procedure, which is
widely applicable in coating and thin film sciences.
[22] S.B. Sadale, P.S. Patil, Solid States Ionics 167 (2004) 273.
[23] V. Fruth, M. Popa, D. Berger, R. Ramer, A. Gartner, A. Ciulei, A. Zaharescu, J. Eur.
Ceram. Soc. 25 (2005) 2171.
[24] S. Michel, S. Diliberto, C. Boulanger, N. Stein, J.M. Lecuire, J. Crystal Growth 277
(2005) 274.
[25] P.M. Vereecken, K. Rodbell, C.X. Ji, P.C. Searson, Appl. Phys. Lett. 86 (2005)
121916.
[26] M.L. Yang, Z.B. Hu, J. Electroanal. Chem. 583 (2005) 46.
[27] V. Richoux, S. Diliberto, C. Boulanger, J.M. Lecuire, Electrochim. Acta 52 (2007)
3053.
[28] E. Sandnes, M.E. Williams, U. Bertocci, M.D. Vaudin, G.R. Stafford, Electrochim.
Acta 52 (2007) 6221.
[29] J. van Muylder, M. Pourbaix, Bismuth, in: Atlas of electrochemical equilibria in
aqueous solutions, National Association of Corrosion Engineers, Houston, 1974,
p. 533.
[30] T.P. Gujar, V.R. Shinde, C.D. Lokhande, R.S. Mane, S.H. Han, Appl. Surf. Sci. 252
(2006) 2747.
[31] J. Wang, J. Lu, Electrochem. Commun. 2 (2000) 390.
[32] E.A. Hutton, J.T.v. Elteren, B. Ogorevc, M.R. Smyth, Talanta 63 (2004) 849.
[33] M. Haidopoulos, S. Turgeon, C. Sarra-Bournet, G. Laroche, D. Mantovani, J. Mater.
Sci.: Mater. Med. 17 (2006) 647.
[34] M.D. Abramoff, P.J. Magelhaes, S.J. Ram, Biophoton. Int. 11 (2004) 36.
[35] W.S. Rasband, ImageJ, B.U.S. National Institutes of Health, Maryland, USA,
1997–2009, http://rsb.info.nih.gov/ij/.
[36] Favia S P., Plasma deposition of fluoropolymer films in different glow discharge
regimes, in: H. Biederman (Ed.), Plasma Polymer Films, Imperial College Press,
London, 2004, p. 25.
Acknowledgment
This work was partially supported by Natural Science and Engi-
neering Research Council of Canada, and from collaborative grants
between Quebec and Wallonie/Bruxelles (Ministry of International
Relations). The authors are grateful to Nicolas Miné, researcher at
the LISE, for his assistance during the ToF-SIMS experiments. The
authors would like to thank Francois Lewis for his precious advices
and comments all along this work.
[37] D. Briggs, Surf. Interface Anal. 14 (1989) 209.
[38] S. Polizu, M. Maugey, S. Poulin, P. Poulin, L.H. Yahia, Appl. Surf. Sci. 252 (2006)
6750.
References
[39] D. Leonard, P. Bertrand, M.K. Shi, E. Sacher, L. Martini, Plasmas Polymers 4
(1999) 97.
[1] K. Lau, A. Johan, U. Sigwart, J. Hung, Singapore Med. J. 45 (2004)
305.
[2] D. Stoeckel, A. Pelton, T. Duerig, Eur. Rad. 14 (2004) 292.
[3] M. Costa, Z. Zhuang, X. Huang, S. Cosentino, C.B. Klein, K. Salnikow, Sci. Total
Environ. 148 (1994) 191.
[4] B.H. Chew, J.D. Denstedt, Nat. Clin. Pract. Urol. 1 (2004) 44.
[5] Food and Drug Administration, Guidance for Industry and FDA Staff, Non-
Clinical Tests and Recommended Labeling for Intravascular Stents and
Associated Delivery Systems, 2005.
[6] H. Wieneke, T. Sawitowski, S. Wnendt, A. Fischer, O. Dirsch, I.A. Karoussos, R.
Erbel, Herz 27 (2002) 518.
[40] B. Lovrecˇek, I. Mekjavic´, M. Metikosˆ-Hukovic´, Nitrogen, phosphorus, arsenic,
antimony, and bismuth, in: A.J. Bard, R. Parsons, J. Jordan (Eds.), Standard Poten-
tials in Aqueous Solution, Marcel Dekker, New York, 1985, 180.
[41] F. Yubero, S. Tougaard, Phys. Rev. B 46 (1992) 2486.
[42] L. Kover, S. Tougaard, J. Toth, L. Daroczi, I. Szabo, G. Langer, M. Menyhard, Surf.
Interface Anal. 31 (2001) 271.
[43] M. Touzin, P. Chevallier, F. Lewis, S. Turgeon, G. Laroche, D. Mantovani, Surf.
Coat. Technol. 202 (2008) 4884.
[44] M.B.J. Wijesundara, G. Zajac, E. Fuoco, L. Hanley, J. Adhes. Sci. Technol. 15 (2001)
599.
[45] D.G. Castner, K.B. Lewis, D.A. Fischer, B.D. Ratner, J.L. Gland, Langmuir 9 (1993)
537.
[7] T.C. Woods, A.R. Marks, Ann. Rev. Med. 55 (2004) 169.