Engineering and Characterization of Polymer Surfaces

for Biomedical Applications

Hans Jörg Mathieu1 · Yann Chevolot2 · Laurence Ruiz-Taylor3 · Didier Léonard4
1 Materials Institute, École Polytechnique Fédérale de Lausanne (EPFL), 1015 Lausanne EPFL,
Switzerland. E-mail: [email protected]
2 Laboratoires Goëmar, UMR 1931 CNRS/Laboratoires Goëmar, Station biologique, 29660
Roscoff. France. E-mail: [email protected]
3 Zyomyx Inc., 26101 Research Road, Hayward CA 94545, USA.
E-mail: [email protected]
4 Analytical Technology, Microanalysis group, GE Plastics Europe, NL-4600 AC Bergen op
Zoom, The Netherlands. E-mail: [email protected]

The application of synthetic polymers in the growing field of materials for medical applica-
tions is illustrated by examples from recent work at the Materials Institute of the Swiss Federal
Institute of Technology in Lausanne. The review highlights the need for functionalization and
chemical control of material surfaces at a molecular/functional level. After a brief introduc-
tion into the surface chemical analysis tools, i.e., X-ray Photoelectron Spectroscopy (XPS) and
Time-of-Flight Secondary Ion Mass Spectrometry (ToF-SIMS) combined with contact angle
measurements, phosphorylcholine biomimicking polymers as well as immobilization of car-
bohydrates on polystyrene are presented.

Keywords: Polymers, Surface analysis, Functionalization, Immobilization, Glycoengineering

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 Methods for surface characterization . . . . . . . . . . . . . . . . . 3

2.1 Surface Chemical Analysis . . . . . . . . . . . . . . . . . . . . . . . . 4

2.1.1 X-ray Photoelectron Spectrometry (XPS) . . . . . . . . . . . . . . . 4
2.1.2 Time-of-Flight Secondary Ion Mass Spectrometry (ToF-SIMS) . . . 8
2.2 Contact Angle Measurements . . . . . . . . . . . . . . . . . . . . . . 11

3 Phosphorylcholine Functional Biomimicking Polymers . . . . . . 13

4 Surface Glycoengineering of Polystyrene . . . . . . . . . . . . . . . 23

5 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

Advances in Polymer Science, Vol. 162

© Springer-Verlag Berlin Heidelberg 2003
2 H. J. Mathieu · Y. Chevolot · L. Ruiz-Taylor · D. Léonard

List of Abbreviations and Symbols

cA Atomic concentration of element A (at %)
Da Dalton
e Electron charge
Eb Binding energy
Ecin Kinetic energy
EK Core level binding energy
h Planck’s constant
v Frequency of light
fA Isotopic abundance of element A
h Height of sessile drop (angle of contact)
IA Intensity of XPS signal of element A
I±A Intensity of positive or negative ions at mass A
Ip Flux of primary ions
Is Flux of X-ray source
k Instrumental variable (XPS)
Ka X-ray transition
l Diameter of droplet
L Flight distance
m Mass
M Matrix
n Number
SA Sensitivity factor of element A
Si Sensitivity factor of element i
T Flight time
v Speed
Vacc Acceleration potential
z Depth
Ytot Total sputter yield
Y±A(M) Total positive or negative ion yield of element A in matrix M
z Charge of molecule
DE Energy resolution
FA Work function of analyzer (XPS)
g±A(M) Positive or negative ionization probability of element A in matrix M
g lv Liquid-vapor interfacial tension
g sv Solid-vapor interfacial tension
g sl Solid-liquid interfacial tension
l Inelastic mean free path of photoelectrons
q Angle
qA Advancing angle
L Escape depth of photoelectrons
hS Analyzer constant (ToF-SIMS)
Engineering and Characterization of Polymer Surfaces for Biomedical Applications 3

1
Introduction
Materials used in biomedical devices (so-called biomaterials) should fulfill two
major requirements. First, they should possess the mechanical and physical prop-
erties that allow them to replace faulty functions of the body. Second, as they are
interfacing biological systems they should be at least bio-inert (no undesired reac-
tions) and, even better, trigger positive responses from these biological systems. In
the latter case, responses are mainly governed by interfacial interactions, i.e., by the
surface properties of the material such as surface energy, surface roughness, and
surface chemical composition. Consequently, analytical methods are of primary
importance to guide surface modifications of materials to lead to biospecific sur-
face properties and also to understand the relationship between surface chemis-
try/roughness/energy and the biological response.
Although historically metals were the first biomaterials, polymers have gained a
large application in this field and great efforts have been devoted to design poly-
meric materials with the right physical and interfacial properties [1]. This work in-
tends to illustrate the application and importance of surface modifications and the
need of surface analytical tools in the field of polymeric biomaterials [2, 3]. The
examples given below stem from the work of several former PhD students and
post-docs of our laboratory and illustrate how surface analysis and biological as-
says are used in a complementary manner to design successfully new soft bioma-
terials. The first example (Sect. 3) shows how one can synthesize a polymer whose
surface properties reduce non-specific protein adsorption and consequently unde-
sired biological reactions. The second example (Sect. 4) copes with the oriented
immobilization of biologically active molecules (carbohydrates) at materials sur-
faces.

2
Methods for Surface Characterization [4]
In this first part, surface analytical methods such as X-ray Photoelectron Spectros-
copy (XPS) and Time-of-Flight Secondary Ion Mass Spectrometry (ToF-SIMS) as
well as contact angle measurements are briefly introduced. The first two tech-
niques give chemical information on the first monolayers of a solid surface while
the latter provides information related to the surface energy. In the following sec-
tion, the basic principle of these analytical tools is discussed as well as the typical
information they give and their limitations.
4 H. J. Mathieu · Y. Chevolot · L. Ruiz-Taylor · D. Léonard

source:
photons detector spectra
ions
signal:
electrons
ions
in-depth
analyzer profiles

images

Fig. 1. Principle of XPS and ToF-SIMS surface analysis

2.1
Chemical Surface Analysis

Chemical analysis of surfaces and in particular of polymers is made possible by

probing with photons (soft X-rays) or ions [5, 6]. The principle of chemical sur-
face analysis is illustrated schematically in Fig. 1. A primary source is directed to-
wards the surface of a solid sample and the spectrometer measures properties of
emitted particles.
The emitted secondary particles (electrons or ions) carry information on the
composition of the top-surface and underlying layers. In addition, the imaging ca-
pabilities make it possible in certain cases to identify heterogeneity in surface
chemical composition. Also, the number of detected particles can be used for
(semi-)quantification directly from the measured spectra, in-depth profiles, or
images. More precisely XPS (probing with photons) and ToF-SIMS (probing with
ions) allow one to access depths from 1 to 10 nm. However, due to their very high
surface sensitivity, these methods are subject to contamination effects at the sur-
face, requiring then a well-controlled preparation of the sample and ultra-high
vacuum (UHV) conditions for analysis, i.e., pressures below 10–6 Pa. In the follow-
ing a short description of the two analytical tools is presented.

2.1.1
X-ray Photoelectron Spectrometry (XPS)

XPS is a technique based upon photo-electronic effect. Under X-ray (photon) ex-
posure, electrons are emitted with energy values characteristic of the elements
present at the surface. Figure 2 illustrates the complete photo-ionization process
including excitation and relaxation steps.
Let us assume that the primary X-ray with energy hn creates a photoelectron at
the core energy level EK (a). The Einstein equation gives the relation between exci-
Engineering and Characterization of Polymer Surfaces for Biomedical Applications 5

a) excitation
EF

EV

Eb
EL
hv e–ph

EK

relaxation e–Auger

b) c)
EF

EV
Eb

EL

X-ray

EK

Fig. 2. Principle of the photo-electronic effect: the excitation and relaxation processes are
shown indicating schematically the different binding states with the Fermi energy (EF) as the
reference level (=0). The valence band energy is EV followed by a discrete level of EL and the
core level EK; after [7]

tation energy (hn), kinetic energy Ekin of the emitted photoelectron and its binding
energy Eb:

Eb = hv – Ekin – F A (1)
where FA is the work function of the analyzer detector. An XPS spectrometer
measures the kinetic energy Ekin of a core photoelectron. From Eq. (1), Eb is deter-
mined. The excitation process is exploited to identify solid elements from lithium
(atomic number Z=3). The sensitivity limit in XPS is approximately 0.1% of a
monolayer corresponding to 1015 particles/cm2. An illustration is given in Fig. 3,
which shows a survey spectrum of polystyrene (PS) after a radiofrequency oxygen
plasma treatment. Typical reference energies for binding energy calibration are ac-
cording to the ISO standard [8]:
Au 4f7/2 84.0 eV
C 1 s 285.0 eV
In Fig. 3, one observes the core level transitions of C1s and O1s as well as the Auger
transition of oxygen, OKVV. Indeed, the photoemission process is followed by a re-
laxation process either (b) or (c) as shown in Fig. 2. In the process (c) a third elec-
6 H. J. Mathieu · Y. Chevolot · L. Ruiz-Taylor · D. Léonard

Fig. 3. XPS survey spectrum of a radiofrequency oxygen plasma treated polystyrene (PS) ob-
tained with monochromatic Al primary radiation. The intensity I (cps) is shown as a function
of binding energy (eV)

tron called the Auger electron is emitted after a transition of an electron from a
lower level (for example the valence band level EV) to the core level EK. In the proc-
ess (b), a secondary X-ray is created after filling the hole at the EK level by a valence
band electron. This photon is not measured in XPS experiments.
EK can exhibit dependence upon the oxidation state of the element. Narrow
scans around an element of interest allow one to determine quantitatively the var-
ious binding states of this element. In particular for polymers, carbon and oxygen
binding states can be identified as illustrated in Fig. 4 for the C1s transition of PM-
MA.
Figure 4 shows the different functional groups and their respective relative areas
are given in the caption. The theoretical relative intensities of the different func-
tional groups are for carbon -C-C- plus -C-H (60%), -C-O-C- (20%), -C=O
(20%) and oxygen -O-C- (50%) and -O=C- (50%), respectively.
XPS peak intensities (areas) IA are a means of quantification. The relation be-
tween IA and the atomic concentration cA of an element or chemical component
at a depth z is
L
Ê z ˆ
I A = kI s S A Ú c A (z)expÁ – ˜dz (2)
Ë l cos q ¯
0

where k is an instrument variable, Is the primary beam flux, SA the elemental sen-
sitivity, and l the inelastic mean free path of the photoelectron trajectory multi-
plied by cos q where q is the take-off angle of the emitted electron with respect to
Engineering and Characterization of Polymer Surfaces for Biomedical Applications 7

Fig. 4. XPS high resolution spectrum of C1s of Poly(methyl metacrylate) (PMMA). The four
components correspond to C-H, C-C, C-O, and C=O functional groups of PMMA; after [9]

the surface normal. Due to the exponential term in Eq. (2) a reasonable upper lim-
it for integration is

L = 3l cos q (3)
Typical values of the escape depth L for polymers are between 5 and 10 nm in-
dicating the shallow information depth of XPS [10]. Quantification is performed
by application of the simple formula

IA
SA
cA = (4)
n
Ii
 Si
i =l
applying elemental sensitivity factors from the literature (for example [5]) or those
provided by the spectrometer manufacturers and summing over the number of el-
ements taken into account. As for polymers accuracy of a few percent is typically
obtained by use of Eq. (4). From Eq. (3) one can see that measurements at differ-
ent take-off angles allow probing sample composition at different depths. Thus
such angle resolved XPS (ARXPS) is an elegant way of obtaining a depth profile in
a non-destructive manner.
When analyzing polymers, charging effects and possible degradation have to be
taken into account [11–15]. Emitted photoelectrons carry a negative charge and
may lead to a positive charge build-up. This effect can be compensated by supply-
ing the sample surface with low energy electrons for charge neutralization. Degra-
8 H. J. Mathieu · Y. Chevolot · L. Ruiz-Taylor · D. Léonard

dation of polymer under X-ray is described in the literature. The reader is referred
to [9–16] for complementary information.
Imaging is possible with a lateral resolution limit of a few microns for state-of-
the-art spectrometers. All measurements are carried out under ultra high vacuum
conditions (UHV). A fast entry lock is usually available for a transfer of a sample
within minutes from atmospheric pressure to 10–6 Pa. Commonly two types of
sources are used in XPS, either MgKa1.2 or Al Ka1.2 radiation with an energy of
1253.6±0.7 or 1486.6±0.85 eV, respectively. Al Ka1.2 radiation is often monochro-
matized by elimination of the Ka2 ray to result in a better defined energy spread of
the incoming X-rays, i.e., a smaller full width at half maximum (FWHM) of the Ka
line allowing higher energy resolution DE down to 0.5 eV of the emitted photo-
electrons. For more detailed information the reader is referred to the literature
[17].
In conclusion, XPS and ARXPS are valuable tools for quantitative elemental
analysis and identification of functional chemical groups within the first few na-
nometers of the surface at relatively high sensitivity.

2.1.2
Time-of-Flight Secondary Ion Mass Spectrometry (ToF-SIMS)

Static Secondary Ion Mass Spectrometry (S-SIMS) is a more sensitive (sensitivity

of 109 atoms/cm2) surface analysis technique than X-ray Photoelectron Spectros-
copy (1012 atoms/cm2) [6, 18]. Under primary bombardment with a focused ion
beam the solid surface emits secondary particles. As illustrated in Fig. 5, the bom-
bardment with primary ions such as Ga+, Ar+, or other molecular ion sources like
SF5+ [19–21] or C60 [22], which should enhance molecular ion formation at high
masses vs fragmentation, provokes the emission of neutral, positively, or negative-
ly charged fragments and clusters.

±
Fig. 5. Principle of Secondary Ion Mass Spectrometry
Engineering and Characterization of Polymer Surfaces for Biomedical Applications 9

In SIMS, one distinguishes between the static and the dynamic modes. In static
SIMS only a fraction of the first monolayer of the surface layer is perturbed. This
depends on the flux of the primary ion beam, which is kept well below 10–12 par-
ticles/cm2. This number corresponds to 1‰ of the number of particles of the first
monolayer, which is approximately 1015 particles. Past this limit, degradation sig-
natures can be detected. During further bombardment the top-surface will then be
completely destroyed, leading to a depth profile type of information (dynamic
mode).
The intensity IA± for emitted positively or negatively charged secondary ions
measured for a given target is described by the following equation:

I ±A = I pYtot g ±A(M ) f Ac AhS (5)

with

±
YM = Ytot g ±A(M ) (6)

Here Ip is the primary ion current, Ytot the total neutral sputter yield of species
A in the matrix M, gA±(M) the ionization probability of species A in the matrix M,
fA the isotopic abundance of element A, cA the atomic concentration, and hS an in-
strumental constant. IA± may vary over several orders of magnitude. Due to its
high sensitivity, surface contamination may influence it strongly. Furthermore, IA±
is very dependent upon matrix effects that influence strongly the number of emit-
ted ions compared to emitted neutrals. Their ratio is typically 10–4 or smaller for
polymers and biomaterials [23]. The influence of the matrix on the ionization rep-
resents the severest limitation of this technique for quantitative analysis of both in-
organic and organic materials.
In modern static SIMS instruments, the most efficient spectrometer is the time
of flight spectrometer. It has typical current densities of the order of 1 nA/cm2,
which corresponds to approximately 1010 particles/cm2/s [23, 24]. The emitted
secondary ions are separated according to their mass in this spectrometer. Figure 6
shows schematically the principle of the time of flight measurement.
The relation between acceleration voltage, Vacc, and kinetic energy, Ekin, of the
secondary ions is developed into

m 2 (7)
eVacc = v
2
where e is the electronic charge, m the mass, and v the velocity of the accelerated
ion. This leads to a simple relation between time of flight, t, and the square root of
the ion mass:
10 H. J. Mathieu · Y. Chevolot · L. Ruiz-Taylor · D. Léonard

Vacc

Ippulsed

S D

IL
L

Fig. 6. Principle of a time-of-flight measurement: S sample, Vacc acceleration voltage, Ip pulsed

ion beam, L flight length, IL immersion lens, D detector

L L (8)
t= = m
v 2eVacc

where L is the flight distance being typically 1–2 m for commercial ToF-spectrom-
eters. In ToF instruments the primary ion source is pulsed and the mass resolution
will depend on the pulse width. The flight time for an ion with mass 100 Da is sev-
eral microseconds and therefore picosecond pulses are required to obtain a relative
mass resolution of m/Dm>5000 at m=28 Da. Liquid Metal Ion Guns (i.e., Ga+ pri-
mary ions) allow one to focus the ion beam down to submicron beam diameter
and then to perform SIMS imaging (mapping of elemental and molecular second-
ary ions). ToF-SIMS spectrometers are equipped with a fast entry lock allowing
one to introduce a sample in the ultra-high vacuum range (UHV) within a few
minutes.
Figure 7 shows schematically a spectrum acquired with a time-of-flight spec-
trometer. It shows a typical fragment of the maleimide molecule fragment
C4H2NO2– at 96.009 Da. Due to the mass resolution of m/Dm=6900 it can clearly
be distinguished from the sulfate molecule SO4– at 95.952 Da. It illustrates that ion
fragments can be measured with high mass resolution. This allows a unique deter-
mination of their chemical composition.
Examples of fingerprint spectra are shown in Sect. 3 (Figs. 11 and 12). Other
detailed information is found elsewhere [17].
Engineering and Characterization of Polymer Surfaces for Biomedical Applications 11

2.2
Contact Angle Measurements

Contact angle measurements are used to assess changes in the wetting characteris-
tics of a surface to indicate changes in surface wettability. Information that one ob-
tains largely depends on the interpretation of contact angle in terms of the Young
equation [25, 26]:

g lv cos q = g sv – g sl (9)
where glv is the liquid-vapor, gsv the solid-vapor, and gsl the solid-liquid interfacial
tension, respectively, and q the measured angle with respect to the surface, as illus-
trated schematically in Fig. 8.
The Young equation implies that the equilibrium contact angle is unique for a
given solid-vapor-liquid system on a flat and rigid surface. In addition, it is gener-
ally accepted that the vapor spreading pressure can be neglected for q>10° [27].
However, in many practical cases the experimentally observed contact angle of a
given system is not uniquely determined by the surface tensions of the solid and

O 96.009 Da

30
N

–
O C4H2NO2
Intensity (cps)

m
20 = 6900
Dm

–
SO4
10 95.952 Da

0
95.8 95.9 96.0 96.1 96.2
m/z

Fig. 7. Time-of-flight mass spectra of negative ions of C4H2NO2– at 96.009 Da which is clearly
separated from SO4– at 95.952 Da thanks to the mass resolution of m/Dm=6900
12 H. J. Mathieu · Y. Chevolot · L. Ruiz-Taylor · D. Léonard

γlv
vapor
γsv θ liquid

γsl
solid

Fig. 8. Equilibrium sessile drop system; glv is the liquid-vapor, gsv the solid-vapor, and gsl the
solid-liquid interfacial tension, respectively and q the measured angle with respect to the sur-
face

liquid, but also by other parameters such as chemistry, inhomogeneity, roughness,

surface deformation, surface reorganization, and chemical contamination [28,
29]. The current understanding of the drop size dependence of contact angles is
discussed in a recent review by D. Li [29]. Surfaces are classified into high- and
low-energy surfaces regarding their interfacial properties. For high-energy surfac-
es adsorption occurs easily, for low-energy surfaces not. Liquids and soft organic
solids such as polymers exhibit surface energies below 100 mN/m, while for hard
solids like metals it is around 500–5000 mN/m. A highly water- and oil repellent
surface exhibits a very small critical surface energy as defined by Zisman [30]. Gen-
erally, surface energy decreases with increasing temperature and ambient pressure
and rises with increasing salt concentration. A small contact angle results for wet-
table surfaces, if the interfacial energy is smaller that the surface energy of the pure
material at the interface to air or vacuum.
Kovats et al. carried out substantial experimental work on contact angles and
surface energy [31–35]. Different angle measuring methods were compared and
surface tensions of 83 organic liquids determined. The importance of a reliable ref-
erence surface applying the Zisman concept is discussed in a theoretical contribu-
tion by Swain and Lipowsky [26] presenting a general form of the Young equation.
Reviews of the critical surface energy determination are found elsewhere [36–38].
The influence of topography on wettability was discussed recently by Öner and
McCarthy [39]. The hydrophobicity of a water repellent surface is discussed. In
such a case the static contact angle is irrelevant, and the dynamic wettability has to
be addressed by measuring the hysteresis, i.e., the difference between advancing
and receding angle. The influence of roughness of ultrahydrophobic polymer sur-
faces (polypropylene and poly(tetrafluoroethylene) (PTFE)) exposed to a radiof-
requency-plasma was discussed elsewhere [40] using XPS and Atomic Force Mi-
croscopy to determine size scale and topology of the roughness. Their most hydro-
phobic surfaces exhibited advancing (A) and receding (R) contact angles of qA and
qR=172° and 169°, respectively.
Engineering and Characterization of Polymer Surfaces for Biomedical Applications 13

h
q

Fig. 9. Contact angle calculation; after [43]

Contact angle measurements of the advancing or receding angle can be per-

formed under a microscope equipped with a CCD camera and a goniometer deter-
mining the slope of the droplet of a given volume of ultrapure water (=
1015 MWcm). The wettability reported in Sect. 3 of this review was evaluated by a
contact angle method using the sessile drop test based on the semi-empirical meth-
od proposed by the literature [41, 42]. The purity of the wetting agent was verified
by measuring the liquid surface tension gLV using the Wilhelmy technique [28]
(platinum plate, KSV sigma 70 Wilhelmy balance) and comparing the obtained
value with the literature (gLV=78.8 mJ/m2) [28, 43]. The height and the contact di-
ameter l of a 1-µl drop of deionised water (grade nanopure milliQ, 17 MW) were
determined after depositing the drop and taking a picture with a CCD camera [28].
The advancing angle q was calculated using the following equation [43]:

Ê hˆ (10)
q A = 2 arctanÁ 2 ˜
Ë l¯

Height h and diameter l are illustrated in Fig. 9. They were determined after dep-
osition of the drop and taking a CCD picture. The purity of the wetting agent was
verified by measuring the liquid surface tension gLV using the Wilhelmy technique
(platinum plate, KSV sigma 70 Wilhelmy balance) and comparing the obtained
value with the literature (gLV=78.8 mJ/m2) [28, 43].

3
Phosphorylcholine Functional Biomimicking Polymers
This section illustrates results obtained in our laboratory in the context of the de-
sign and control of the surface properties of polymers used as biomaterials [43].
Indeed, interactions between biomaterials and tissues occur via a layer of proteins
adsorbed at the surface of any implant [44]. Such protein adsorption is the imme-
14 H. J. Mathieu · Y. Chevolot · L. Ruiz-Taylor · D. Léonard

a) PCPUR
O
O O O O O O
O O C HN CH 2 NH C O P O N
O 6
w x y O z

b) P(MMA:MA:APC)

x y z
O O O O
OM e OM e
O P N
O O
O

Fig. 10a,b. Schematic structures of the PC copolymer synthesized: a PCPUR; b

P(MMA:MA:APC); after [56, 58]

diate event occurring on its first contact with biological fluids and tissues [3]. One
major restriction of polymers is their tendency to exhibit thrombogenic proper-
ties.
The response of biomolecules and cell membranes is determined by many fac-
tors, some of which are the chemical composition and conformation of the mole-
cules, the surface energy, and topography of the top surface layers which are in
contact with biological systems, i.e., body fluids and cells [45]. The work illustrat-
ed here consisted in designing new polymers with functional properties capable of
promoting the attachment of specific cells. The first step consisted in a polymer
system which surface inhibits non-specific cell attachment. This strategy is based
on the incorporation of cell membrane constituents such as phosphorylcholine
(PC) or phospholipid analogues into polymers [46–51].
Since polyurethanes have traditionally proved to be reasonably bio- and hemo-
compatible materials and have therefore been widely used for biomedical applica-
tions such as vascular prosthesis, artificial organs, blood contacting devices, pe-
ripheral nerve repair, or other prosthetic devices [52], our first PC copolymer sys-
tems were initially based on the polyurethane chemistry. We had earlier demon-
strated that the presence of PC groups in poly(urethane) strongly reduces cell at-
tachment at the surface even in protein enriched media [53] and work from Coop-
er and coworkers [54, 55] also showed that phosphorylcholine containing poly-
urethane limited neutrophil and bacterial adhesion. Two copolymers, PCPUR189
and PCPUR167, were synthesized with different concentration of PC moieties
[56]. The final PC concentration in the bulk of the copolymers was 3.4 mol% and
4.3 mol% for the PCPUR189 and PCPUR167, respectively. Schematics of struc-
ture of the phosphorylcholine containing polyurethane (PCPUR) copolymers
synthesized is given in Fig. 10.
The second copolymer system was chosen because of the higher synthetic flex-
ibility offered by acrylate chemistry. Our choice was directed towards an acrylic
Engineering and Characterization of Polymer Surfaces for Biomedical Applications 15

terpolymer system, with methyl methacrylate (MMA) and methyl acrylate (MA)
as principal components. The purpose was to enable the adaptation of the matrix
mechanical properties by adjusting the glass transition temperature of the system
[57]. 2-Acryloyloxyethyl phosphorylcholine (APC) was synthesized and added to
MMA and MA through copolymerization to control the surface properties of the
P(MMA:MA:APC) terpolymer with respect to cell attachment [58]. The final PC
concentration in the bulk of the P(MMA:MA:APC) terpolymer was 1.7 mol%. A
schematic representation of the structure of the terpolymer is given in Fig. 10.
A major part of these projects was also dedicated to understanding the bulk and
solution structural organization of such polymers and here we would like to refer
the reader to appropriate literature [56, 58]. While it was shown that the am-
phiphile nature of the terpolymer system, and in particular the phosphorylcholine
(PC) groups, played an important role in the structure organization and molecular
mobility of the copolymers, the results displayed in the present discussion focus on
how the presence of these PC groups impacted on the surface properties of the co-
polymer synthesized. With this aim in mind, surface analytical techniques such as
XPS and ToF-SIMS were used and complemented with contact angles and biolog-
ical in-vitro assays to characterize the dynamics of the copolymers surface and as-
sess the extent of the resistance to the non-specific attachment of cells on samples
coated with the PC copolymers (for experimental details, refer to [56, 58]).
ToF-SIMS analysis of the 2-acryloyloxyethyl phosphorylcholine (APC) mono-
mer led to the typical positive fragmentation pattern of the phosphorylcholine
group in agreement with literature [59, 60]. Figure 11a illustrates the high relative
intensity of some of these typical positive fragments such as C5H13N+O3P
(166 Da) and C5H15N+O4P (184 Da). A more extensive list of the major positive
ion-fragments detected is found elsewhere [58].
Figure 11b presents the same mass range of the positive SIMS spectrum obtained
for the P(MMA:MA:APC) terpolymer. None of the typical fragments of the PC
group can be identified despite the high surface sensitivity of the technique. In con-
trast, characteristic fragments from P(MMA:MA) matrix were observed in agree-
ment with the literature [61], such as C8H9O3+ (153 Da) as well as C2H3O2+
(59 Da), C4H5O+ (69 Da), and C5H9O2+ (101 Da) (not displayed in Fig. 11b). Neg-
ative ions spectra confirmed these results. Indeed, although the PO3– fragment could
be detected, its intensity was too low to consider this as evidence of PC presence.
In contrast, investigations of PCPUR polymers using ToF-SIMS show that both
polymers exhibit similar positive and negative mode mass spectra. In this case, ma-
jor secondary ions characteristic of the PC polar head group were observed (PO3–
(79 Da), C5H12N+ (86 Da), C5H13NO2P+ (150 Da), C5H13NO3P+ (166 Da), and
C5H15NO4P+ (184 Da) as shown in Fig. 12a,b.
Furthermore, the characteristic fragment of GPC containing coatings,
C8H19NO4P+ (224D), was also detected (data not shown). The comparison of the
16 H. J. Mathieu · Y. Chevolot · L. Ruiz-Taylor · D. Léonard

a) APC
184
166
O
P N O
O O
P N
HO O
OH

150 160 170 180 190 200

153

C b) P(MMA-MA-APC)
O O O

150 160 170 180 190 200

m/z [a.m.u.]

Fig. 11a,b. Comparison between the positive ToF-SIMS mass spectra of the APC: a monomer;
b terpolymer in the 150–200 Da mass range; after [58]

3.5 104
SIMS Intensity (a.u.)

x 20 a)
*
*
** * *
0
0 20 40 60 80 100 120 140 160 180 200
3.5 104
b)
SIMS Intensity (a.u.)

x 20

0
0 20 40 60 80 100 120 140 160 180 200

Fig. 12a,b. Comparison of the positive mode ToF-SIMS spectra of: a PC-PUR189; b PC-
PUR167 in the mass range 0–200 Da, after [56]; * denotes the position of the characteristic PC
fragments

positive mode ToF-SIMS spectra shows, however, that relative normalized intensi-
ties of the characteristic PC fragments are higher by approximately 50% for
PCPUR167 compared to PCPUR189.
Studies regarding the conformation of lecithins have allowed the determination
of the size of the phosphorylcholine polar group. This dimension derived from X-
Engineering and Characterization of Polymer Surfaces for Biomedical Applications 17

Table 1. AR-XPS surface elemental concentrations of PCPUR and P(MMA:MA:APC) copoly-

mers. Comparison between two emission angles of 20 and 80 ° with respect to the sample sur-
face, corresponding to an information depth of 3–4 nm and 8–11 nm, respectively; compiled
after [56, 58]
Emission angle (information depth) C [at%] O [at%] N [at%] P [at%]
PCPUR189 20 ° (3–4 nm) 67.1±0.5 21.4±0.8 11.5±0.5 0.3±0.09
80 ° (8–11 nm) 66.1±0.0 21.6±0.3 11.7±0.2 0.6±0.06
PCPUR167 20 ° (3–4 nm) 66.7±0.6 21.6±0.7 11.4±0.3 0.6±0.06
80 ° (8–11 nm) 65.9±0.8 21.8±0.7 11.7±0.3 0.8±0.06
P(MMA:MA:APC) 20 ° (3–4 nm) 70.0±0.6 29.8±0.6 0.0 0.2±0.02
80 ° (8–11 nm) 70.0±0.0 29.4±0.0 0.3±0.06 0.3±0.03

ray diffraction studies is 11 Å for both the monohydrated or fully hydrated leci-
thins [62] and is in agreement with the 12 Å obtained with space-filling models for
a fully extended PC group parallel to the fatty acid chains and with the zwitterion
in the gauche conformation about the O-C-C-N bonds [63]. Hence, the typical
fragments of the phosphorylcholine groups should be detectable if they were
present at the uppermost surface of the P(MMA:MA:APC) terpolymer. Possible
matrix effects have already been reported in organic systems [64, 65]. Although
they may affect the emission probabilities of typical PC secondary ions, it is very
unlikely that they should inhibit completely the emission of all the PC fragments.
These SIMS observations therefore suggest that under the UHV conditions re-
quired for the analysis, the extreme surface of the P(MMA:MA:APC) terpolymer
is depleted in PC groups. This burying effect can directly be understood by the de-
sire of the system to reorganize so as to minimize surface energy in UHV environ-
ment.
XPS analyses were performed on all copolymers. The atomic concentrations of
carbon (C1s), oxygen (O1s), nitrogen (N1s), and phosphorus (P2p) were compared
between the two depths of 3–4 nm and 8–11 nm and are reported in Table 1.
Contrasting with ToF-SIMS analysis, nitrogen and phosphorus were detected
on the P(MMA:MA:APC), although in the case of nitrogen we were close to the de-
tection limit. As it can also be seen from the high resolution elemental scans re-
ported in Fig. 13a,b, the nitrogen and phosphorus atomic concentrations meas-
ured are very low; for an emission angle of 20° no nitrogen is even detected where-
as the phosphorus concentration is about 0.2 at%. This can be understood by con-
sidering the difference of electron inelastic mean free paths (imfp) between both
elements. Indeed, calculations based on the estimation of the imfp made by Tanu-
ma et al. [66] for PMMA as well as on the relation derived by Seah and Dench [67]
for the imfp of organic compounds allow the estimation of the escape depth of the
XPS information. Both approaches give similar results. In the case of nitrogen, the
information depth is approximately 3 nm and 9 nm whereas for phosphorus it
18 H. J. Mathieu · Y. Chevolot · L. Ruiz-Taylor · D. Léonard

Escape depth 3 nm (a) N1s

Escape depth 9 nm

N(E)*E

410 405 400 395

BINDING ENERGY [eV ]

Escape depth 4 nm b) P2p

Escape depth 11 nm
N(E) *E

140 135 130

BINDING ENERGY [eV]

Fig. 13a,b. High-resolution elemental scans as a function of depth of: a the nitrogen N1s; b the
phosphorous P2p of the P(MMA-MA-APC) copolymer; after [58]

ranges from 4 nm to 11 nm for the 20° and 80° emission angles, respectively. As the
information depth is augmented, an increase in the nitrogen (from 0 to 0.3 at%)
and phosphorus (from 0.2 to 0.3 at%) atomic concentrations is noticeable.
Similar comments can be made concerning the PCPUR copolymers, for which
elemental concentrations are comparable for both copolymers at both emission
angles, except in the case of phosphorus. Indeed, although phosphorus concentra-
tions detected are very low, they are significantly different (outside of the experi-
mental deviation) and show that the concentration is higher in PCPUR167
(Table 1) in agreement with the ToF-SIMS results described earlier. Moreover, it
can be seen that the phosphorus P2p photoelectron intensity (and consequently
Engineering and Characterization of Polymer Surfaces for Biomedical Applications 19

the phosphorus atomic concentration) also increases with the depth analyzed for
these PCPUR copolymers.
This suggests that under UHV conditions, for both copolymer systems (acr-
ylates or polyurethanes), there is an increasing concentration gradient of the phos-
phorylcholine groups from the surface to the bulk of the polymer. The copolymer
system tends to lower its interfacial free energy by burying the PC group under-
neath the extreme surface.
In-vitro assays were performed in serum free (SFM) and serum containing me-
dia (SCM) to evaluate the cell attachment properties on both PC-containing and
PC-free copolymers together with SiO2 as reference [56, 58]. Results obtained after
4 h of incubation (Figs. 14 and 15) show that the cell attachment and differentia-
tion levels on the PC containing copolymers were strongly reduced in both media
compared to PC free system and the SiO2 reference. Hence, a small concentration
of PC groups (~1.7 mol% for the P(MMA:MA:APC) and ~3.4 mol% for
PCPUR189) is already efficient in reducing the non-specific attachment of cells by
roughly 70% for both media. Cell attachment is even further reduced as the PC
content of the copolymer increases (~4.3 mol% for PCPUR167), since a 90% de-
crease is obtained for PCPUR167 in both media when compared to SiO2.
Extension of the culture time up to, respectively, three days for the
P(MMA:MA:APC) copolymer and four days for the PCPUR copolymers, showed
no evolution in the cell attachment properties on the PC containing surfaces, as
opposed to the PC-free surfaces (SiO2 reference as well as P(MMA:MA) polymer).
Indeed, after 72 h in SCM, a large population of cells were attaching and differen-
tiating on P(MMA:MA) as can be seen from Fig. 16a. In contrast, on the

140 SiO2
P(MMA:MA)
P(MMA:MA:APC)
120
Mean celladhesion [cells/mm2]

100

80

60

40

20

0
SFM SCM

Fig. 14. Comparison of the cell attachment level after 4 h incubation, on the APC free and the
APC containing copolymers as a function of time in serum free (SFM) and serum containing
(SCM) media; after [58]
20 H. J. Mathieu · Y. Chevolot · L. Ruiz-Taylor · D. Léonard

200
SiO2 PC-PUR189 PC-PUR167

Cell attachement level [cells/mm2]

150

100

50

0
SFM SCM

Fig. 15. Comparison of the cell attachment level on the SiO2 reference, PC-PUR189 and PC-
PUR167 after 4 h in SFM and SCM (*p<0.01 PCPUR190/PCPUR167 Student t-test, n=15 and
**p<0.001, PCPUR189/PCPUR167 Student t-test, n=15) ; after [56]

a) b)

50 µm 50 µm

SiO 2
c)

50 µm
P(MMA:MA:APC)

Fig. 16a–c. Extent of cell attachment and differentiation on: a P(MMA:MA);

b P(MMA:MA:APC); c SiO2, after 72 h in serum-containing media [58]

P(MMA:MA:APC) surface, cells rather aggregate to form clusters than adhere to

the terpolymer surface (Fig. 16b), or preferentially grow on the uncovered silicon
part of the sample (Fig. 16c). This last observation gives additional information
about the absence of any toxic leakages from the acrylic terpolymer film. The cell
attachment level was comparable in both SFM and SCM suggesting that the pres-
Engineering and Characterization of Polymer Surfaces for Biomedical Applications 21

(a) (b)

(c)

100 µm

Fig. 17. Extent of cell attachment after 96 h in serum containing media, on: a SiO2 reference
substrate; b PCPUR189 in SCM; c PCPUR167 in SCM. Note the extensive reduction in cell
attachment on PCPUR polymers compared to the reference substrate; after [56]

ence of binding proteins did not significantly influence the cell attachment level on
the acrylic terpolymer surface.
Figure 17 illustrates the extent of cell attachment on the PCPUR copolymers af-
ter 96 h in the protein rich media (SCM). Although few cells are attaching on
PCPUR189 (Fig. 17b) and have adopted this typical triangular morphology indic-
ative of good attachment, neurite extension on this substrate is restricted com-
pared to the SiO2 reference (Fig. 17a). Furthermore, on PCPUR167 samples, al-
most no adhering cells remain, except on a small area of the sample where the
PCPUR167 coating was accidentally removed. Cells are then growing on the bare
SiO2 substrate along the scratched portion of the coating (Fig. 17c).
Non-specific adsorption of various proteins to these copolymers was also stud-
ied [56] and followed essentially the same trend as the one observed for the non-
specific attachment of cells, i.e., the non-specific protein adsorption was lower on
the copolymer with the highest PC content. Hence, this inhibition of cell attach-
ment on PCPUR polymer surfaces (and non-specific protein adsorption) can be
directly related to the presence of PC groups. This implies that under aqueous con-
ditions the PC groups were now present at the extreme surface for the copolymers
to display such non-permissive properties.
Static water wettability measurements performed on PCPUR189 and
PCPUR167, showed that the mean contact angle of a drop of water was 20° smaller
22 H. J. Mathieu · Y. Chevolot · L. Ruiz-Taylor · D. Léonard

on PCPUR167 (60±3.5°) compared to PCPUR189 (81±2.5°). This suggests an in-

creased hydrophilicity of PCPUR167 (enhanced water-polymer interaction),
which is in agreement with the higher phosphorylcholine groups concentration.
In addition, in contrast to PCPUR189, the drop was observed to spread and flatten
at the surface of the PCPUR167 polymer. Such a phenomenon can be attributed to
both swelling and chain reorientation [68, 69], which appears easier in the case of
PCPUR167 in agreement with the water weight gain and DSC observations [56].
Results of XPS and ToF-SIMS characterizations correlated to in-vitro assays and
contact angles obtained for both phosphorylcholine containing polymer systems
(PC-PURs and P(MMA:MA:APC)) have shown that the amphiphile nature of
those copolymers and in particular the presence of the PC moieties generates a
highly mobile outermost surface capable of quickly reorganizing as a function of
the polarity of the environment it is subjected to. Under UHV conditions, both
systems tend to reorganize so as to minimize their interfacial free energy by bury-
ing the PC groups, which have no affinity for the UHV environment and hence
lowering the PC groups surface concentration. Many similar examples of surface
reorganization of polymer systems upon the polarity of the environment have
been reported [69–73]. In the case of the acrylate system, the extreme surface an-
alyzed with ToF-SIMS was even depleted in PC groups. A difference in matrix ef-
fects is possible between the polyurethane and acrylate systems and can make a
quantitative comparison of the PC content difficult. However, it is still not expect-
ed that different ions should be observed for both systems. Indeed, these fragments
follow the same fragmentation process, regardless of the matrix they are emitted
from; they could be observed for GPC, PC-PURs, APC, as well as for phospholip-
ids [56, 59, 60]. ToF-SIMS results show that the relative intensity of the PC frag-
ments decreases, as the PC bulk concentration decreases from 4.3 mol% in PC-
PUR167 to 3.4 mol% in PC-PUR189, until fragments are no longer detectable at
the outermost surface when the PC bulk concentration is 1.7 mol% in the acrylate
system. On the other hand, AR-XPS measurements have shown that the PC groups
concentration augments as the depth of analysis increases.
In the case of the acrylate system, the PC depletion of the outermost surface of
the terpolymer suggests that there can be a PC threshold concentration below
which, under UHV conditions, PC groups can be completely buried/hidden from
the surface. However, this PC threshold concentration may not only be depending
on the PC bulk concentration of the polymer. Indeed, the difference in polarity be-
tween the matrix and the PC groups is greater in the P(MMA:MA:APC) acrylate
system than in polyurethane polymers. The driving force for surface reorganiza-
tion is, therefore, greater in the P(MMA:MA:APC) terpolymer since the polar PC
groups have no affinity for the UHV environment and are masked by hydrophobic
polymer chains. On the contrary, in response to a polar environment, the terpoly-
mer surface rearranges to present the phosphorylcholine functionalities at the top
surface, hence minimizing interfacial energy and exhibit strong non-permissive
Engineering and Characterization of Polymer Surfaces for Biomedical Applications 23

properties with respect to cell attachment. The amphiphile nature of the terpoly-
mer constitutes by itself the main driving force for surface reorganization. It gen-
erates a highly mobile polymer surface, with quick self-reorganization properties
upon the environment. Besides, cohesion energy of the different constituents of
the acrylate terpolymer (and especially of the hydrocarbon backbone) is always
lower than the energy barrier to rotation of a single C-C bond (12.6 kJ/mol) [74]
meaning that chains are free to rotate, hence promoting surface reorganization.
In the case of PC-PURs polymers, surface reorganization certainly also occurs
although, even if as suggested by the wettability behavior of the polymers, it was
not as obvious as for the acrylate system. A PC group concentration difference
close to 50% could already be detected by ToF-SIMS at the outermost surface be-
tween PC-PUR189 and PC-PUR167. Furthermore, the PC concentration, as ob-
served using AR-XPS, increased with the analyzed depth. Nevertheless, the driv-
ing force for surface reorganization is certainly smaller, due to lower polarity dif-
ferences between PC groups and the polyurethane matrix. The large amount of
water retained in this polymer structure [56] also leads to an apparently more po-
lar matrix. Besides, the ability of the urethane and urea segments of the matrix to
form hydrogen bonds between N-H and C=O groups increases the cohesion en-
ergy of the system, thus restricting the molecular mobility of the polymer chains
[74].
In-vitro experiments correlated to bulk and surface characterization results
have demonstrated that an increase of PC groups concentration such as in the
polyurethane system leads to a reduction of the cell attachment level in both serum
containing and serum free media. The presence of phosphorylcholine groups, as
demonstrated on the acrylate system, is clearly responsible for the non-permissive
properties exhibited by the PC containing surfaces with respect to cell attachment.
Moreover, a low PC concentration (ª1.7 mol%) is already sufficient to reduce
strongly cell adhesion at the surface of the terpolymer. Bulk organization (due to
higher polarity differences) between the phosphorylcholine groups and a "flexible"
polymer matrix seems to be the determining factor in providing a very compliant
surface that can easily adapt to its environment. Hence, the high compliance of the
surface and the dynamic flip-flop behavior of the PC groups are certainly playing
a crucial role in the prevention of the non-specific adsorption of proteins and cells
on such PC containing copolymers.

4
Surface Glycoengineering of Polystyrene
A derivatization can be performed to induce a selective attachment of target cells
either on new polymers designed to control non-specific cell attachment (Sect. 3)
or on any other substrate intended for biomaterial applications. Photoimmobili-
zation of peptidic sequences was performed on newly synthesized PCPUR con-
24 H. J. Mathieu · Y. Chevolot · L. Ruiz-Taylor · D. Léonard

taining polymers described in the section Phosphorylcholine Functional Biomim-

icking Polymers above [56]. This section of this review illustrates the work done in
our laboratory on the surface glyco-engineering [75], i.e., the specific covalent at-
tachment of carbohydrates at the surface of substrates via photo-addressable new-
ly synthesized molecules. It is wished to emphasize how surface analytical tools
and biological assays give complementary information on (1) the immobilization
of intact molecules and (2) the availability of active molecules for recognition.
Biological systems make considerable use of surface glycosylation. Indeed glyc-
osylated molecules are involved in, among others, molecule trafficking, cell recog-
nition (including species discrimination), blood group typing, and blood coagu-
lation cascade [76–79].
Carbohydrate immobilization is reported in the literature for galactose [80],
melibiose [81], mannose [81, 82], lactose [82] starch [83], lactosaminide [84], and
heparin [85–89]. This has been achieved mainly by thin film adsorption (weak in-
teractions) or by covalent binding. In our work, light was used as a means for ori-
ented covalent immobilization of carbohydrates. Orientation and availability are
key parameters to allow the interaction of the cell receptor with the immobilized
molecules. Photoimmobilization provides a versatile tool with respect to the sub-
strate (organic and inorganic) and allows one to create easily microdomains of bi-
orecognition with addressable printing, mask-assisted lithography techniques.
The photoreagents most often used for photoimmobilization of biomolecules are
arylazides, trifluoromethyl-aryl diazirines, and benzophenones [90]. These rea-
gents generate very reactive intermediates upon light activation. Their interaction
with the support material leads to the formation of covalent bonds [90]. Thus, the
synthesis of molecules containing carbohydrate and photoactivatable domains
makes it possible to achieve covalent binding of biologically active carbohydrates.
Utilizing diazirine as a photoactivatable function, a reactive carbene is generated
by thermochemical or light activation (350 nm). A covalent bond is generated
with the surface provided that close contact between the surface and the carbene
is obtained. Reaction of aryl diazirine with various substrates has previously been
described [91–94]. The carbene may insert into C-H, C-C, C=C, N-H, O-H, S-H
bonds [90]. Chevolot et al. [95] reported the synthesis and the immobilization on
a CVD deposited diamond surface of an aryl diazirine containing a galactose pho-
toreagent {4-(3-D-galactopyranosylsulfanyl-2,5-dioxopyrrolidin-1-yl)-N-(3-(3-
trifluoromethyl-3H-diazirin-3-yl)phenyl}butyramide (MAD-Gal) (Scheme 1)
[95]. Based on XPS and ToF-SIMS data it was concluded that grafting of the pho-
toactivatable reagent MAD-Gal is possible using glycosylated aryldiazirines. Using
a masking technique [96] a specific pattern of immobilized carbohydrate was laid
down on diamond and verified with ToF-SIMS. This is illustrated in Fig. 18.
Léonard et al. [94, 96, 97] demonstrated that the molecule was immobilized as
a whole without degradation due to the immobilization process. They also dem-
onstrated that insertion on diamond was indeed carbene mediated. This was evi-
Engineering and Characterization of Polymer Surfaces for Biomedical Applications 25

Fig. 18. Negative-mode ToF-SIMS of F- after mask-assisted patterning on thin diamond films
– size of the image 115•115 µm; after [75]

OH OH
O
O
H
HO N
OH N

O
O

F3C N
N

Scheme 1. Structure of MAD–Gal

denced in the case of oxygenated sites at the surface of CVD deposited diamond
but other functional sites could also be involved.
In the following discussion, it is intended to compare the immobilization of two
photoactivatable carbohydrates on polystyrene (PS) and to illustrate how their bi-
ological activity was tested. The same galactose aryl diazirine as the one described
for diamond modification (see Scheme 1) was used and compared to the also new-
ly synthesized lactose aryl diazirine displayed in Scheme 2.
The immobilization on polystyrene (the material most often used in biological
assays) was controlled with surface analysis spectroscopic methods XPS and ToF-
26 H. J. Mathieu · Y. Chevolot · L. Ruiz-Taylor · D. Léonard

RO OR OR
H
O O N
O S
RO RO
OR OR O

F3C N
N

Scheme 2. Chemical structure of lactose aryl diazirine (R=H)

SIMS. The biological activity of the polystyrene-modified surfaces was probed

with the lectin Allo A, primary rat hepatocytes, and a-2,6-sialyltransferase.
Allo A lectin was reported to have a specific affinity for galactose residues espe-
cially as part of lactose (3.1 mmol/l inhibits hemoglutinin activity) and O-nitro-
phenyl b-D-galactopyranoside (inhibitory at 12.5 mmol/l) [98]. Hepatocytes ex-
press on their surface the asialoglycoprotein receptor which is responsible for the
clearance of abnormal galactose-terminated serum glycoproteins [99, 100]. Most
serum glycoproteins carry terminal sialic acid residues and a penultimate galactose
residue. When desialylated, the exposed galactose residues of the glycoprotein can
interact with the asialoglycoprotein receptor, initiating removal of the glycopro-
tein from the circulation by endocytosis. Subsequently, the incorporated protein is
hydrolyzed in lysosomes [100–103]. The transfer of sialic acid by a-2,6-sialyltrans-
ferase [104] to galactose attached to a solid support was studied to develop the ap-
proach to solid phase semi-synthesis [105].
Immobilization on PS was carried out as described by Léonard et al. [94, 96].
The samples were washed four times in water (HPLC grade). In the following ex-
perimental discussion, samples are referred to as A, B, and C. Sample A is the pris-
tine polystyrene. Sample B is polystyrene with photoactivatable carbohydrates af-
ter glycosilation and washing. This sample treatment does not include light expo-
sure. Physisorbed molecules should be removed by washing. Sample C is the pol-
ystyrene surface with molecules deposited on it, then photoactivated with light of
350 nm wavelength and finally washed leaving an expectedly modified PS surface.
When comparing samples A, B and C, simple signatures of the molecules such
as fluorine, CF3 among others, should allow one to test easily the efficiency of their
immobilization on PS with XPS and ToF-SIMS. Table 2 illustrates the case of lac-
tose aryl diazirine. The residual level of molecules observed on sample B by ToF-
SIMS was below the XPS instrument detection limit. It confirms that the washing
procedure was able to remove physisorbed molecules. For sample C, after washing,
covalently attached molecules were expected to remain bound to the surface, in
contrast with sample B [94, 96]. XPS fluorine atomic percentage, as well as CF3–
and F– normalized intensities (Table 2), illustrate in the same way the successful
immobilization of the molecule at the surface of polystyrene. These results are
Engineering and Characterization of Polymer Surfaces for Biomedical Applications 27

Table 2. XPS and ToF-SIMS analysis of lactose aryl diazirine grafted PS. XPS atomic percent-
ages are displayed for samples A, B and C. Sample A corresponds to as received polystyrene.
Sample B corresponds to polystyrene on which a methanolic solution of lactose aryl diazirine
was deposited, and no light activation was performed before the final washing. This surface
should be similar to sample A. Sample C corresponds to polystyrene on which a methanolic
solution of lactose aryl diazirine was deposited and light activation was performed before the
final washing. The molecule should be present at the surface of the material. Three areas were
analyzed per sample, after [105]
XPS percentages (at %) Sample A Sample B Sample C
N bdl 0.34±0.20 0.51±0.49
F bdl bdl 1.18±0.4
S bdl bdl 0.18±0.16
C 97.79±0.36 94.53±1.27 88.77±0.54
O 2.27±0.47 5.13±1.08 9.35±0.49
ToF-SIMS
Corrected total intensity (counts) •104 16.3±4.5 49.6±5.4 43.7±10.7
F- normalized intensity ‰ 1.10 2.72±0.42 63.17±21.78
CF3– normalized intensity ‰ 0.13 0.20±0.08 0.65±0.07

similar to those of the immobilization of MAD-Gal on diamond [96] and on PS

[106].
Lactose and galactose residue bioavailability at the surface of polystyrene was
probed with the b-galactose specific lectin Allo A through a biotinylated Allo A
(Fig. 19). The amount of surface lectin was measured with 35S streptavidin. 100%
radioactivity relates to the total amount of radioactivity of the deposited radiola-
beled streptavidin before washing. This represents an arbitrary standard but is
used to indicate the relative levels of radioactivity remaining after the various
treatments of the modified surface. The results are illustrated in Figs. 19 and 20 for
lactose and galactose modification experiments of the polystyrene surface, respec-
tively. For the two molecules, the percentages of radioactivity for samples A were
similar (1.2 and 2.5%). For samples B (lactose aryl diazirine), the signal remained
very low. For samples C (lactose aryl diazirine), the radioactivity reached the value
of approximately 3%, which is seven times lower than for the corresponding sam-
ple C in the case of MAD-Gal experiments as shown in Fig. 20.
Furthermore, inhibition with asialofetuin of lectin binding on immobilized ga-
lactose (MAD-Gal) was observed [105]. It confirmed that the interaction was spe-
cific and not due to a variation in physisorption, which in turn is related to the sur-
face energy.
In addition, XPS demonstrated that the surface concentrations of the two mol-
ecules (MAD-gal, lactose aryl diazirine) were almost similar (see Table 2).
Figure 21 indicates quantitative XPS and normalized ToF-SIMS data of MAD-Gal
28 H. J. Mathieu · Y. Chevolot · L. Ruiz-Taylor · D. Léonard

3.5

3.0

2.5
% of radioactivity

2.0

1.5

1.0

0.5

0
A B C

Fig. 19. Binding of biotinylated Allo A lectin to aryl lactose derivatized PS surfaces. Derivatized
surfaces were incubated with Allo A lectin and then extensively washed to remove physisorbed
lectins. [35S] streptavidin was incubated and the surfaces were again rinsed to remove excess
streptavidin. Finally, radioactivity was measured by scintillation counting. The radioactivity
is higher on sample C; after [105]

25

20
% of radioactivity

15

10

0
A B C 1 C1 C2

Fig. 20. Binding of biotinylated Allo A lectin to MAD-Gal derivatized PS surface. The deriva-
tized PS surfaces were incubated with Allo A lectin and then extensively washed to remove
physisorbed lectins. [35S] streptavidin was added and the surfaces were again washed to re-
move excess streptavidin. Finally, radioactivity was measured by scintillation counting. The
radioactivity is higher on sample C and increases with the concentration of the MAD-Gal so-
lution (C, C1 and C2 corresponding to 0.25, 0.025 and 0.0025 mmol/l, respectively) used for
immobilization. This illustrates the higher binding of streptavidin to the surface C. I corre-
sponds to incubation of MAD-Gal grafted PS with asialofetuin and Allo A lectin. 100% corre-
sponds to the total added radioactive streptavidin; after [105]
Engineering and Characterization of Polymer Surfaces for Biomedical Applications 29

120 0.7

0.6
100
ToF-SIMS F normalised intensity ‰
0.5

XPS F atomic percentage %

80
F % (XPS)
0.4
60
0.3
–
F (ToF-SIMS)
–

40
0.2

20 0.1

0 0
0 0.05 0.1 0.15 0.2 0.25 0.3
Applied MAD-Gal (mM)

Fig. 21. F– normalized intensity values (absolute intensity of F–/(total intensity – H– intensity))
and fluorine atomic percentages are displayed as a function of the MAD-Gal concentration
used for immobilization. The intensity of surface characteristic signatures increases with in-
creasing density of MAD-Gal; after [105]

immobilized on PS as a function of the MAD-Gal concentration used for immo-

bilization.
Therefore, the difference in lectin binding can be attributed to differences in af-
finity of the lectin for the different terminal galactose residues at the surface. Ac-
cording to Yamashita [98] the affinity of the lectin was observed to decrease for
various substrates such as phenyl b-D-galactopyranoside, O-nitrophenyl b-D-ga-
lactopyranoside, lactose, and lactulose. It is therefore possible that in this case and
as for phenyl b-D-galactopyranoside and O-nitrophenyl b-D-galactopyranoside,
the succinimidyl group enhanced the affinity of the lectin towards the galactose of
MAD-Gal relative to the galactose residue of the immobilized lactose.
On the cellular level, functionality of the galactose and lactose residues was test-
ed with primary rat hepatocytes. The MAD-Gal coated surface altered hepatocel-
lular function. MTT (3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bro-
mide) and Neutral Red (NR) uptake (Fig. 22) increased for hepatocytes cultured
on MAD-Gal coated polystyrene surfaces and the increase correlated with the den-
sity of galactose molecules on the surface. MTT formation and NR uptake in-
creased up to a concentration of 2.5 mmol/l galactose and showed no saturation
effect. CMF stands for crude membrane fraction which is taken as a reference.
NR uptake is related to lysosomal activity while MTT is related to the bioreduc-
tive uptake, when lectins also specific of galactose residues (RCA) were added most
probably because lectins bind to the terminal galactose residues of MAD-Gal. In
the case of lactose modified polystyrene, no increase in the NR uptake was ob-
served (data not shown). On the other hand, when surfaces are then incubated
30 H. J. Mathieu · Y. Chevolot · L. Ruiz-Taylor · D. Léonard

150

140
% Neutral red uptake

130

120

110

100

90
C incubated
C0 C C1 C2 CMF Asialof. A
with lectin

Fig. 22. Neutral Red (NR) uptake is related to the lysosomal activity. The NR uptake decreased
with the decreasing of surface galactose residues (C0, C, C1, C2 corresponding to 2.5, 0.25,
0.025, and 0.0025 mmol/l). The NR uptake was similar to that of surface A (plain polystyrene)
when the surface was incubated with a lectin (RCA) that protects the galactose residues. Hepa-
tocytes possess higher NR uptake on asialofetuin coated surfaces than on surface A, but lower
than on MAD-Gal grafted PS. 100% corresponds to surface A; after [105]

with 14C CMP-NeuAc and a-2, 6-sialyltransferase, radioactivity was only detected
on lactose modified PS (not illustrated).
It was also demonstrated that the radioactivity on the surface increased with in-
cubation time. When incubated without the enzyme, radioactivity was similar to
the background level. These results illustrate that the enzyme was able to recognize
lactose on PS surface as a substrate for the transfer of sialic acid and to catalyze this
transfer [105].
Improvement of the biological interactions will require more complex struc-
tures (cluster, spacer design, oligosaccharides). For example, as in the case of
asialoglycoprotein receptor, a tri-anternary structure of the galactose residues in-
creases the affinity of the receptor for the blood circulating glycoprotein. It was
also demonstrated that the distance between the residues of the tri-anternary
structure influences the affinity [105]. In a similar manner surface glycoengineer-
ing will require such an architectural design.
From the presented data it is concluded that diazirine-containing photorea-
gents containing mono- and disaccharides can be synthesized and immobilized on
polystyrene. Biological activity of the modified polystyrene was demonstrated
with Allo A lectin, primary rat hepatocytes and a-2, 6-sialyltransferase (not shown
here) [105]. It was also demonstrated that the biological activity was not only de-
Engineering and Characterization of Polymer Surfaces for Biomedical Applications 31

pendent on the terminal residue but probably also on the spacer structure. Never-
theless, the combination of chemistry, surface analytical tools, and biology is
found to be a powerful feedback system for the design of bioactive glycoengineered
surfaces.

5
Concluding Remarks
This review illustrates the efficiency of the surface modification of polymers with
bio-specific properties allowing key-lock interactions with cells. An advanced
chemistry is needed to design new polymers exhibiting control of non-specific cell
attachment allowing further surface bio-specific surface interactions. New compo-
nents can be designed permitting bio-active surface modifications like photoad-
dressable carbohydrate compounds. Furthermore, surface analytical tools like XPS
and ToF-SIMS are important for the preparation and control of the intended sur-
face properties. Their spectra and images give insight into the top-surface compo-
sition with a high sensitivity. They can even be directly correlated to biological re-
sults such as the increase of MAD-Gal surface density correlated with the increase
in the observed biological response. The complementarity of both techniques is
also illustrated by these examples. The difference in probing depth allows high-
lighting compositional differences at the surface as well as gaining information on
functional group orientation. This review illustrates that surface analysis tech-
niques are powerful analytical means for the control of bio-active polymer engi-
neering. The high sensitivity (ToF-SIMS) and the possibility to quantify data
(XPS) identify these techniques as a major tool for further developments of bio-
materials.

Acknowledgements. The authors would like to thank N. Xanthopoulos, G. Coullerez,

and X. Gao for help with some of the experiments and discussions. Financial sup-
port is acknowledged from the Swiss Priority Program of Materials and the Com-
mon research program in Biomedical Engineering 1999–2002 between the Univer-
sity Hospitals of Geneva and Lausanne (HUG and CHUV), the Universities of Ge-
neva and Lausanne, and the Swiss Federal Institute of Technology Lausanne
(EPFL) as well as from the Bundesamt für Berufsbildung und Technologie
(BBT/KTI), Bern, contract no. CTI-5170.1 MTS.

References
1. Ratner BD, Chilkoti A, Lopez GP (1990) Plasma deposition and treatment for biomate-
rial applications. In: d’Agostino R (ed) Plasma deposition, treatment and etching of pol-
ymers. Academic Press, San Diego, pp 463–516
2. Williams DF (1992) Biofunctionality and biocompatibility. In: Williams DF (ed) Medical
and dental materials. VCH, Weinheim, pp 1–27
32 H. J. Mathieu · Y. Chevolot · L. Ruiz-Taylor · D. Léonard

3. Hubbell JA, Langer R (1995) Chem Eng News 73:42–54

4. Mathieu HJ (2001) Surf Interface Anal 32:3–9
5. Briggs D, Seah MP (1990) (ed) Practical surface analysis, vol I – Auger and X-ray photo-
electron spectroscopy, 2nd edn, vol 1. Wiley, Chichester
6. Briggs D, Seah MP (1992) (ed) Practical surface analysis, vol II – ion and neutral spec-
troscopy, vol 2. Wiley, Chichester
7. Mathieu HJ (2001) Elemental analysis by AES, XPS and SIMS. In: Alfassi ZB (ed) Non-
destructive elemental analysis. Blackwell Science, Oxford, pp 201–231
8. ISO (2001) Surface chemical analysis – X-ray photoelectron spectrometers – calibration
of energy scales. In: International Standard 15472, 2001
9. Beamson G, Briggs D (1992) High resolution XPS of organic polymers. The Scienta
ESCA300 database, Wiley, Chichester
10. Powell CJ, Jablonski A (1999) J Phys Chem Ref Data 28:19–62
11. Frydman E, Cohen H, Maoz R, Sagiv J (1997) Langmuir 13:5089–5106
12. Buchwalter LP, Czornyj G (1990) J Vac Sci Tech A 8:781–784
13. Chaney R, Barth G (1987) Fresenius J Anal Chem 326:143
14. Clark DT, Brennan WJ (1986) J Electron Spectrosc 41:399–410
15. Storp S (1985) Spectrochim Acta B 40:745–756
16. Coullerez G, Chevolot Y, Léonard D, Xanthopoulos N, Mathieu HJ (1999) J Surf Anal
5:235–239
17. Vickerman JC (ed) (1997) Surface analysis – the principal techniques. Wiley, Chichester
18. Briggs D (1992) Static SIMS-surface analysis of organic materials. In: Briggs D, Seah MP
(eds) Practical surface analysis, vol 2 – ion and neutral spectroscopy. Wiley, Chichester,
pp 367–423
19. Benninghoven A, Stapel D, Brox O, Burkhardt H, Crone C, Thiemann M, Arlinghaus HA
(1999) Static SIMS with molecular primary ions. In: SIMS XII. Elsevier, Brussels
20. Stapel D, Thiemann M, Hagenhoff B, Benninghoven A (1999) Secondary ion emission
from LB-layers under molecular primary ion bombardment. In: SIMS XII. Elsevier,
Brussels
21. Hagenhoff B, Cobben PL, Bendel C, Niehuis E, Benninghoven A (1997) Polymers under
SF5 bombardment – a systematic investigation. In: SIMS XI. Wiley, Orlando, Florida
22. Justes DR, Harris RD, Stipdonk MJV, Schweikert AE (1997) A comparison of Cs and C60
primary projectiles for the characterization of GaAs and Si surfaces. In: SIMS XI. Wiley,
Orlando, Florida
23. Vickerman JC, Briggs D (eds) (2001) ToF-SIMS: surface analysis by mass spectrometry.
IM Publications and Surface Spectra, Chichester
24. Vickerman JC, Swift AW (1997) Secondary ion mass spectrometry. In: Vickerman JC
(ed) Surface analysis – the principal techniques. Wiley, Chichester, pp 135–214
25. Young T (1805) Philos Trans R Soc London 96:65
26. Swain PS, Lipowsky R (1998) Langmuir 14:6772–6780
27. Garbassi F, Morra M, Occhiello E (eds) (1994) Polymer surfaces from physics to technol-
ogy. Wiley, Chichester
28. Good RJ (1993) Contact angle, wetting, and adhesion: a critical review. In: Mittal KL
(ed) Contact angle, wettability and adhesion. VSP-Wiley, Weinheim, pp 3–36
29. Li D (1996) Colloid Surf A 116:1–23
30. Zisman WA (1964) Relation of equilibrium contact angle to liquid and solid constitu-
tion. In: Fowkes FM (ed) Contact angle, wettability, and adhesion. American Chemical
Society, Washington, D.C., pp 1–51
31. Boksanyi L, Liardon O, Kovats E (1976) Adv Colloid Interface Sci 6:95–137
32. Erard J-F, Nagy L, Kovats E (1983) Colloid Surf 9:109–132
33. Gobet J, Kovats E (1984) Adsorpt Sci Technol 1:111–122
34. Korösi G, Kovats E (1981) J Colloid Surf 2:315–355
35. Riedo F, Czencz M, Liardon O, Kovats E (1978) Helv Chim Acta 61:1912–1941
36. de Gennes PG (1985) Rev Mod Phys 57:827–863
Engineering and Characterization of Polymer Surfaces for Biomedical Applications 33

37. Fowkes FM (1964) Contact angle, wettability, and adhesion. Advances in Chemistry Se-
ries, vol 43. American Chemical Society
38. Holloway PJ (1970) Pestic Sci 1:156–163
39. Öner D, McCarthy TJ (2000) Langmuir 16:7777–7782
40. Youngblood JP, McCarthy TJ (1999) Macromolecules 32:6800–6806
41. Walliser A (1992) Caracterisation des interactions liquide-fibres élémentaires par mouil-
lage. Université de Haute Alsace (F), 92-MUHL-0248
42. Connor M (1995) Consolidation mechanisms and interfacial phenomena in thermo-
plastic powder impregnated composites. EPFL Thesis No 1413
43. Ruiz L (1997) Thesis: synthesis and characterisation of phosphorylcholine containing
polymers designed to promote specific cell attachment via surface derivatisation. École
Polytechnique Fédérale de Lausanne: Lausanne
44. Marieb EN (1998) Human anatomy and physiology, 4th edn. Addison-Wesley Publish-
ing
45. Williams DF (1987) Definitions in biomaterials, vol 4. Progress in biomedical engineer-
ing. Elsevier
46. Kadoma Y, Nakabayashi N, Masuhara E, Yamauchi J (1978) Kobunshi Ronbunshu
35:423–427
47. Chapman D (1993) Langmuir 9:39–45
48. Hayward JA, Chapman D (1984) Biomaterials 5:135–142
49. Ishihara K, Aragaki R, Ueda T, Watenabe A, Nakabayashi N (1990) J Biomed Mater Res
24:1069–1077
50. Ishihara K, Aragaki R, Yamazaki JI, Ueda T, Watenabe A, Nakabayashi N (1990) Seitai Za-
iryo 8:231–236
51. Ishihara K, Oshida H, Endo Y, Ueda T, Watenabe A, Nakabayashi N (1992) J Biomed
Mater Res 26:1543–1552
52. Lelah MD, Cooper SL (1986) Polyurethanes in medicine. CRC Press
53. Ruiz L, Johnston DS, Makohliso SA, Aebischer P, Mathieu HJ (1995) Biomimetic coat-
ings on silicon wafers: synthesis and characterisation in ECASIA 95 Montreux, Switzer-
land. Mathieu HJ, Reike B, Briggs D (eds.) Wiley, Chichester
54. Yung LL, Cooper LS (1998) Biomaterials 19:31
55. Baumgartner JN, Yang CZ, Cooper SL (1997) Biomaterials 18:831
56. Ruiz L, Fine E, Vörös J, Makohliso SA, Léonard D, Johnston DS, Textor M, Mathieu HJ
(1999) J Biomater Sci Polym E 10:931–955
57. Peyser P (1989) Glass transition temperatures of polymers. In: Brandrup J, Immergut EH
(eds) Polymer handbook. Wiley, pp VI 209–277
58. Ruiz L, Hilborn JG, Léonard D, Mathieu HJ (1998) Biomaterials 19:987–998
59. Davies MC, Lynn RAP (1990) Crit Rev Biocompat 5:297–341
60. Yianni YP (1992) Biocompatible surfaces based upon biomembrane mimicry In: Quinn
PJ, Cherry RJ (eds) Structural and dynamic properties of lipids and membranes. Port-
land Press, London, UK, pp 187–216
61. Hearn MJ, Briggs D (1988) Surf Interface Anal 11:198–213
62. Phillips MC, Finer EG, Hauser H (1972) Biochim Biophys Acta 290:397–402
63. Hauser H (1975) Phospholipid model membranes: demonstration of a structure -activ-
ity relationship in chemoreception. Infr Retr Ltd, London
64. Briggs D, Hearn MJ (1988) Surf Interface Anal 13:647–669
65. Chilkoti A, Castner DG, Ratner BD, Briggs D (1990) J Vac Sci Technol A8:2274
66. Tanuma S, Powell CJ, Penn DR (1993) Surf Interface Anal 21:165–176
67. Seah MP, Dench WA (1979) Surf Interface Anal 1:2–11
68. Garbassi F, Morra M, Occhiello E (1994) Surface energetics and contact angle. In: Poly-
mer surfaces from physics to technology. Wiley, Chichester, chap 4, pp 161–219
69. Holly FJ, Refojo MF (1975) J Biomed Mater Res 9:315–326
70. Lavielle L, Schultz J (1985) J Colloid Interface Sci 106:438–445
71. Lavielle L, Schultz J, Sanfeld A (1985) J Colloid Interface Sci 106:446–451
34 H. J. Mathieu · Y. Chevolot · L. Ruiz-Taylor · D. Léonard

72. Andrade JD (1988) Polymer surface dynamics. Plenum Press, New York London
73. Ratner BD, Weathersby P, Hoffman AS, Kelly MA, Scharpen LH (1978) J Appl Polym Sci
22:643–664
74. Wirpsza Z (1993) Polyurethanes: chemistry, technology and applications. Polymer sci-
ence series. Ellis Horwood PTR Prentice Hall
75. Chevolot Y (1999) Thesis: Surface photoimmobilisation of aryl dyazirine containing car-
bohydrates – tools toward surface glycoengineering. École Polytechnique Fédérale de
Lausanne, Lausanne
76. Lee YC, Lee RT (1995) Acc Chem Res 28:321–327
77. Monsigny M (1995) Biofutur 142:27–32
78. Petrak K (1994) Adv Drug Deliver Rev 13:211–213
79. Varki A (1995) Glycobiology 3:97–130
80. Hatanaka K, Takeshige H, Akaike T (1994) Carbohydr Chem 13:603–610
81. Hermanson GT, Mallia AK, Smith PK (1992) Immobilized affinity ligand techniques. Ac-
ademic Press
82. Kobayashi K, Kobayashi A, Akaike T (1994) Method Enzymol 247:409–419
83. Onyiriuka EC (1990) Appl Spectrosc 44:808–811
84. Kobyashi K, Akaike T, Usui T (1994) Method Enzymol 242:226–235
85. Guire PE (1990) Method of improving the biocompatibility of solid surfaces. Pat 4–973–
493, Bio-Metric Systems, Inc, US
86. Guire PE (1990) Biocompatible Coating for Solid Surfaces. Pat 4–979–959, Bio-Metric
Systems, Inc, US
87. Erdtmann M, Keller R, Baumann H (1994) Biomaterials 15:1043–1049
88. Anderson AB, Tran TH, Hamilton MJ, Chudzik SJ, Hasting BP, Melchior MJ, Hergen-
rother RW (1996) Am J Neuroradiol 17:859–863
89. Anderson AB, Enrico LS, Melchior MJ, Pietig JA, Tran LV, Tran TH, Duquette PH (1994)
Photochemical immobilization of heparin to reduce thrombogenesis. Tewntieth Annual
Meeting of the Society for Biomaterials, Boston, MA, USA
90. Sigrist H, Collioud A, Clémence J-F, Gao H, Luginbühl R, Sänger M, Sundarababu G
(1995) Opt Eng 34:2339–2347
91. Gao H, Sanger M, Luginbühl R, Sigrist H (1995) Biosens Biolectron 10:317–328
92. Gao H, Luginbühl R, Sigrist H (1997) Sens Actuators B 38/39:38–41
93. Collioud A, Clémence J-F, Sänger M, Sigrist H (1993) Bioconjugate Chem 4:528- 536
94. Léonard D, Chevolot Y, Bucher O, Haenni W, Sigrist H, Mathieu HJ (1998) Surf Interface
Anal 26:783–792
95. Chevolot Y, Bucher O, Léonard D, Mathieu HJ, Sigrist H (1999) Bioconjugate Chem
10:169–175
96. Léonard D, Chevolot Y, Bucher O, Sigrist H, Mathieu HJ (1998) Surf Interface Anal
26:793–799
97. Léonard D, Chevolot Y, Heger F, Martins J, Crout DHG, Sigrist H, Mathieu HJ (2001)
Surf Interface Anal 2001:457–464
98. Yamashita K (1989) Method Enzymol 179:331
99. Stockert RJ, Morell AG, Ashwell G (1991) 12:441
100. Baenziger JU, Maynard Y (1980) J Biol Chem 255:4607–4613
101. Schwartz AL (1990) Annu Rev Immunol 8:195–229
102. Ashwell G, Harford J (1982) Annu Rev Biochem 51:531–554
103. Kawasaki T, Ashwell G (1976) J Biol Chem 251:12
104. Malissard M, Zeng S, Berger EG (1999) Bioconjugate Chem 16:125
105. Chevolot Y, Martins J, Milosevic N, Léonard D, Zeng S, Malissard M, Berger EG, Maier
P, Mathieu HJ, Crout DHG, Sigrist H (2001) Bioorgan Med Chem 9:2943–2953
106. Sigrist H, Chevolot Y, Crout D, Martins J, Mathieu HJ, Lohmann D (2000) European Pat-
ent Application 99112422.3–2110