Probing resistance to protein adsorption of oligo(ethylene glycol)-terminated self-assembled monolayers by scanning force microscopy

Article abstract of DOI:10.1021/ja991049b

Functionalized scanning force microscope (SFM) probes were used to investigate and to mimic the interaction between fibrinogen and self-assembled monolayers (SAMs) of methoxytri(ethylene glycol) undecanethiolates -S(CH₂)₁₁(OCH₂CH₂) ₃OCH₃ (EG3-OMe) on gold and silver surfaces. The SAMs on gold are resistant to protein adsorption, whereas the films on silver adsorb variable amounts of fibrinogen. Experiments were performed with both charged and hydrophobic tips as models for local protein structures to determine the influence of these parameters on the interaction with the SAMs. A striking difference between the two monolayers was established when the forces were measured in an aqueous environment with hydrophobic probes. While a long-range attractive hydrophobic interaction was observed for the EG3-OMe on silver, a repulsive force was measured for EG3-OMe on gold. The strong dependence of the repulsive force for the EG3-OMe-gold system upon the solution ionic strength suggests that this interaction has a significant electrostatic contribution. The observed differences are attributed to the distinct molecular conformations of the oligo-(ethylene glycol) tails on the gold-supported (helical) and silver-supported ("all-trans") monolayers. A comparison of the force/distance curves for the EG3-OMe SAMs with those measured under identical conditions on end-grafted poly(ethylene glycol) (PEG 2000) on gold further emphasizes that the nature of the repulsive forces originating from the short-chain oligomers is unique and not related to a "steric repulsion" effect.

Full text of DOI:10.1021/ja991049b

10134
J. Am. Chem. Soc. 1999, 121, 10134-10141
Probing Resistance to Protein Adsorption of Oligo(ethylene
glycol)-Terminated Self-Assembled Monolayers by Scanning Force
Microscopy
K. Feldman,† G. Ha1hner,*,† N. D. Spencer,† P. Harder,‡ and M. Grunze*,‡
Contribution from the Laboratory for Surface Science and Technology, Department of Materials,
ETH Zu¨rich, Sonneggstrasse 5, CH-8092 Zu¨rich, Switzerland, and Angewandte Physikalische Chemie,
UniVersita¨t Heidelberg, INF 253, 69120 Heidelberg, Germany
ReceiVed April 1, 1999
Abstract: Functionalized scanning force microscope (SFM) probes were used to investigate and to mimic the
interaction between fibrinogen and self-assembled monolayers (SAMs) of methoxytri(ethylene glycol)
undecanethiolates -S(CH2)11(OCH2CH2)3OCH3 (EG3-OMe) on gold and silver surfaces. The SAMs on gold
are resistant to protein adsorption, whereas the films on silver adsorb variable amounts of fibrinogen. Experiments
were performed with both charged and hydrophobic tips as models for local protein structures to determine
the influence of these parameters on the interaction with the SAMs. A striking difference between the two
monolayers was established when the forces were measured in an aqueous environment with hydrophobic
probes. While a long-range attractive hydrophobic interaction was observed for the EG3-OMe on silver, a
repulsive force was measured for EG3-OMe on gold. The strong dependence of the repulsive force for the
EG3-OMe-gold system upon the solution ionic strength suggests that this interaction has a significant electrostatic
contribution. The observed differences are attributed to the distinct molecular conformations of the oligo-
(ethylene glycol) tails on the gold-supported (helical) and silver-supported (“all-trans”) monolayers. A comparison
of the force/distance curves for the EG3-OMe SAMs with those measured under identical conditions on end-
grafted poly(ethylene glycol) (PEG 2000) on gold further emphasizes that the nature of the repulsive forces
originating from the short-chain oligomers is unique and not related to a “steric repulsion” effect.
Introduction
chains in the near-surface region. This theory is not appropriate,
however, to explain the observation that protein resistance is
also inherent in oligo(ethylene glycol)-terminated alkanethiol
SAMs,6 where the conformational freedom of the OEG tails is
restricted by packing forces.
Self-assembled monolayers (SAMs) of alkanethiolates on
metal surfaces constitute a class of molecular assemblies formed
by the spontaneous chemisorption of long-chain functionalized
molecules on the surface of solid substrates.1,2 Due to their ease
of preparation, long-term stability, controllable surface chemical
functionality, and high, crystal-like, two-dimensional order,
SAMs represent suitable model surfaces to study molecular
adsorption, adhesion, wetting, lubrication, and the interaction
of proteins and cells with artificial organic surfaces. The latter
phenomena are of crucial importance to the fields of biomate-
rials, biosensors, and medical devices.
The outstanding protein-resistant properties of poly(ethylene
glycol) (PEG)-containing surfaces have been recognized for a
long time and extensive experimental and theoretical work has
been carried out to elucidate the physics underlying these
properties.3-5 When interpreted in terms of the “steric repulsion”
theory,3-5 the protein resistance of PEG is associated with a
high conformational freedom and hence entropy of its solvated
The protein adsorption characteristics of the methoxytri-
(ethylene glycol) undecanethiolate (EG3-OMe) SAMs are
strongly sensitive to the substrate used.7 While the monolayers
self-assembled on Au are protein resistant, those self-assembled
on Ag are not. Fourier transform infrared reflection-absorption
spectroscopic (FTIRAS) experiments show that the SAMs on
Au and Ag differ in the conformation of the OEG tail.7 On
gold a conformation with spectral characteristics typical of
helical and amorphous PEG is found, whereas the spectra
observed on the Ag substrates resemble those for the planar
all-trans conformation of stretched PEG samples (Figure 1). The
different conformation and higher packing density in the Ag-
supported monolayer is due to the formation of an incom-
mensurate solid phase with nearly upright chains and a
perceptibly smaller lattice spacing, as confirmed by calculation
of the lowest-energy monolayer configurations on Au and Ag.8
Good agreement is found between the experimental and
calculated vibrational spectra of the lowest-energy SAM con-
figurations on Au and Ag using our structural models.9
* Corresponding authors.
† ETH Zu¨rich.
‡ Universita¨t Heidelberg.
(1) Ulman, A. Chem. ReV. 1996, 96, 1533-1554.
(2) Ulman, A. An Introduction to Ultrathin Organic Films: From
Langmuir-Blodgett to Self-Assembly; Academic Press: Boston, 1991.
(3) Harris, J. M. Poly(Ethylene Glycol) Chemistry; Plenum: New York,
1992.
(4) Jeon, S. I.; Andrade, J. D. J. Colloid Interface Sci. 1991, 142, 159-
166.
(6) Prime, K. L.; Whitesides, G. M. J. Am. Chem. Soc. 1993, 115,
10714-21.
(7) Harder, P.; Grunze, M.; Dahint, R.; Whitesides, G. M.; Laibinis, P.
E. J. Phys. Chem. B 1998, 102, 426-436.
(5) Taunton, H. J.; Toprakcioglu, C.; Fetters, L. J.; Klein, J. Nature 1988,
332, 712-714.
(8) Pertsin, A. J.; Grunze, M.; Garbuzova, I. A. J. Phys. Chem. B 1998,
102, 4918-4926.
10.1021/ja991049b CCC: $18.00 © 1999 American Chemical Society
Published on Web 10/15/1999

Oligo(ethylene glycol)-Terminated Monolayers
J. Am. Chem. Soc., Vol. 121, No. 43, 1999 10135
onto polymer surfaces22 have also been reported in the literature.
However, the importance of the different chemical regions
(neutral or charged) in a protein with respect to the net
interaction of the macromolecule with a surface is still being
debated.23
In the following, we first describe force-distance measure-
ments between fibrinogen-modified probes and the surfaces of
tri(ethylene glycol) (EG3-OMe)-terminated alkanethiolate mono-
layers on gold and silver, obtained with a commercial AFM to
demonstrate the good correlation between these experiments and
fibrinogen adsorption data obtained by FTIR. Subsequently,
results obtained with charged (oxidized/hydroxylated) or hy-
drophobized (hexadecanethiol-derivatized) probes are reported.
These experiments are conducted to elucidate qualitatively the
different contributions and relevance of charged or hydrophobic
patches in fibrinogen (the protein used in this study) to the
resulting interactions with tri(ethylene glycol)-terminated SAMs
on gold or silver.
Figure 1. Molecular cross-sections for the helical EG3-OMe on gold
with a ∼30° tilt of the alkyl chain and a perpendicular orientation of
“all-trans” EG3-OMe on silver. For details see ref 7.
Experimental Section
Materials. EG3-OMe (1-mercaptoundec-11-yl)tri(ethylene glycol)
methyl ether was prepared according to a general procedure developed
by Prime and Whitesides.6 11-Bromoundec-1-ene was added dropwise
to a solution of tri(ethylene glycol) methyl ether in THF with 2 equiv
of NaH and the solution was stirred overnight. Nonconsumed NaH was
reacted with 2-propanol. The solvent was removed and the remaining
product was purified by column chromatography. A solution of the
olefin in methanol containing 4 equiv of thiolacetic acid and 10 mg of
AlBN was irradiated for 6 h under an atmosphere of nitrogen with a
450 W, medium-pressure mercury lamp. Concentration of the reaction
mixtures by rotary evaporation at reduced pressure followed by
purification by chromatography on silica gel gave the thioacetate. The
latter was refluxed overnight in 0.5 M HCl in methanol and the
concentrated solution was purified by column chromatography. The
purity of the thiol was checked with NMR and mass spectroscopy.
The syntheses of EG6-OH and EG[3,1]-OMe have been described
elsewhere.24 (1-Mercaptoundec-11-yl)poly(ethylene glycol) methyl ether
(MW 2000) was prepared analogous to a general procedure for
mercaptoundecyl oligo(ethylenglycol) described by Prime and White-
sides.24 The following changes were made: PEG 2000 monomethyl
ether as starting material was dried over molecular sieve 48 h before
use. For the coupling reaction to the alkyl chain, 2 equiv of 11-
bromoundec-1-ene was added to a solution of PEG 2000 monomethyl
ether in dry THF with 2 equiv of NaH and the mixture was stirred
overnight. Purification of the different synthetic steps was carried out
by column chromatography on silica gel with chloroform/methanol 1:3
as eluent. Purity and molecular weight distribution of the thiol were
checked by NMR and MALDI mass spectrometry (Mn (number average)
) 2199, Mw (weight average) ) 2224; Mw/Mn ) 1.01).
To interpret the protein-resistance properties of the SAMs
containing the helical conformers, the presence of a stable
interphase water layer preventing the SAM from direct contact
with protein molecules was postulated from ab initio 6-31G
SCF calculations of the microscopic structure of the SAM/water
interphase region.10 The observed difference in protein adsorp-
tion between the helical and planar SAM phases is assumed to
be caused by a difference in the structural organization of water
near the SAM surface. Calculations show that the helical SAM
phase easily accommodates water molecules, which act as a
template for further adsorption of water via a hydrogen bridge-
bonding network. The resulting interphase water layer is tightly
bound to the helical or amorphous SAM surface, thereby
preventing direct contact between the surface and the protein.
In contrast, the denser SAM phase on silver does not allow water
molecules to form strong hydrogen bonds with oxygen atoms
in the EG3-OMe strands. As a result, no interphase water layer
capable of hindering protein adsorption is formed. Similar
explanations have been given in the literature11 to correlate the
inertness of some hydrophilic surfaces toward protein adsorption
with the contact angles of water.
Force measurements have been performed with scanning
probe techniques to study protein properties and protein-surface
interactions. Apart from specific recognition and specific
interactions,12-18 single-molecule force spectroscopy,19 adhesion
forces between ligand-receptor pairs,20,21 and protein adsorption
SAM Preparation. Polycrystalline gold (99.99%, Balzers Materials,
Liechtenstein) and silver (99.99+%, Aldrich Chem. Co., Milwaukee,
WI) substrates were prepared by thermal evaporation of these metals
onto plasma-cleaned pieces of singly polished silicon (100) wafers
(MEMC Electronic Materials, Inc., St. Peters, MO) in a BAL-TEC
(Balzers, Liechtenstein) MED-020 coating system operated at (3-5)
× 10-5 mbar. Evaporation of a 5-nm chromium adhesion layer was
followed by deposition of a 100-nm layer of gold or silver at a rate of
0.5 nm/s. Coated substrates were immediately immersed into 2 mmol
thiol solutions in ethanol. Upon removal from the thiol solution, the
SAMs were rinsed with pure ethanol and dried with nitrogen.
(9) Pertsin, A. J.; Grunze, M. Langmuir 1994, 10, 3668-74.
(10) Wang, R. L. C.; Kreuzer, H. J.; Grunze, M. J. Phys. Chem. B 1997,
101, 9767-9773.
(11) Vogler, E. A. AdV. Colloid Interface Sci. 1998, 74, 69-117.
(12) Florin, E. L.; Moy, V. T.; Gaub, H. E. Science 1994, 264, 415.
(13) Ludwig, M.; Moy, V. T.; Rief, M.; Florin, E. L.; Gaub, H. E.
Microsc. Microanal. Microstrucrt. 1994, 5, 321.
(14) Lee, G. U.; Kidwell, D. A.; Colton, R. J. Langmuir 1994, 10, 354-
357.
(15) Florin, E. L.; Rief, M.; Lehmann, H.; Ludwig, M.; Dornmair, C.;
Moy, V. T.; Gaub, H. E. Biosens. Bioelectron. 1995, 10, 895-901.
(16) Allen, S.; Chen, X. Y.; Davies, J.; Davies, M. C.; Dawkes, A. C.;
Edwards, J. C.; Roberts, C. J.; Sefton, J.; Tendler, S. J. B.; Williams, P. M.
Biochemistry 1997, 36, 7457-7463.
(17) Chowdhury, P. B.; Luckham, P. F. Colloids Surf. A 1998, 143, 53-
57.
(18) Bowen, W. R.; Hilal, N.; Lovitt, R. W.; Wright, C. J. J. Colloid
Interface Sci. 1998, 197, 348-352.
(19) Rief, M.; Oesterhelt, F.; Heymann, B.; Gaub, H. E. Science 1997,
275, 1295-1297.
(20) Lee, G. U.; Chrisey, L. A.; Colton, R. J. Science 1994, 266, 771.
(21) Moy, V. T.; Florin, E. L.; Gaub, H. E. Colloids Surf. A 1994, 93,
343.
(22) Chen, X.; Davies, M. C.; Roberts, C. J.; Tendler, S. J. B.; Williams,
P. M.; Davies, J.; Dawkes, A. C.; Edwards, J. C. Langmuir 1997, 13, 4106-
4111.
(23) Horbett, T. A., Brash, J. L.; Horbett, T. A.; Brash, J. L., Eds.
American Chemical Society: Washington, DC, 1995; pp 11-14.
(24) Prime, K. L.; Whitesides, G. W. J. Am. Chem. Soc. 1991, 113, 12.

10136 J. Am. Chem. Soc., Vol. 121, No. 43, 1999
Feldman et al.
Characterization of SAMs involved static water-contact-angle measure-
ments with a contact-angle goniometer (Rame´-Hart, Inc., Mountain
Lakes, NJ) and film-thickness measurements by ellipsometry (Type
L-116C, Gaertner Science Corp., Chicago, IL) with a 70° angle of
incidence of a He-Ne laser. Results of water-contact-angle measure-
ments were in agreement with those reported earlier:7 63° ( 2° for the
EG3-OMe SAMs on both gold and silver. Representative films were
checked by FTIR with respect to their overall quality and molecular
conformation.
A sample of PEG 2000-thiolate was prepared by immersion of a
gold substrate into an ethanolic 1 mmol thiol solution for 48 h;
characterization of the sample revealed that the packing density of the
molecules (distance between the binding sites) is slightly less than the
unperturbed radius of gyration for the given molecular weight distribu-
tion (Rg ∼ 3.6 nm).
To obtain charged SFM probes we used “sharpened” Si3N4 Microlevers
(Park Scientific Instruments, Sunnyvale, CA) with a nominal radius of
curvature of 20 nm, and treated them for 45-60 s in a RF-plasma
cleaner (Harrick Scientific Corp., Ossining, NY), operated at 40 W
with a water-enriched oxygen feed to ensure both removal of
contaminants and hydroxylation of the probe-tip surface. The plasma-
treated tips were stored either in deionized water or in PBS buffer
solution.
The largest dimension of a fibrinogen molecule in the native state
is on the order of 470 Å.25,28 The size of the charged RC-domains of
this protein and of its hydrophobic D-domains is quite comparable to
the contact area between the SFM probes and the surface.28
SFM Measurements. Force-distance measurements and imaging
were performed with a Nanoscope IIIa scanning force microscope
(Digital Instruments, Inc., Santa Barbara, CA) equipped with a liquid
cell. We monitored the temperature inside the cell with a K-type
thermocouple. The temperature was in the range of 27-30 °C during
our measurements. Force-distance curves were collected with a cycle
frequency of 0.3-0.5 Hz. All liquids introduced into the liquid cell
were filtered with 0.22 µm “millex-GV”, low-protein-binding filters
(Millipore, Bedford, MA). Manufacturer-provided nominal values of
the cantilevers’ spring constants, k0, and frequencies, ω0, were used to
calculate the actual spring constant, k, by measuring resonant frequen-
cies, ω, of the probes according to the equation:
Probes for SFM Measurements. It is known that formation of
fibrinogen clusters on a hydrophobic surface is due to strong intermo-
lecular interactions. These appear to be less significant in the presence
of a mica surface.25,26 Hence, to avoid clustering, we decided to mainly
use oxygen-plasma-treated Si3N4 probes to adsorb fibrinogen. Similarly
to mica, these probes acquire a negative charge at neutral pH. It should
be noted, however, that similar results to those reported in the following
were also obtained using hydrophobic (C16-functionalized) probes with
adsorbed fibrinogen.
“Hydrophobic” probes (referred to below as C16-probes) were
prepared by vapor-phase deposition of hexadecanethiol (Aldrich
Chemical Co., Inc., Milwaukee, WI) onto gold-coated Si3N4 probes
(Digital Instruments, Inc., Santa Barbara, CA) with a nominal radius
of curvature of 30 nm. Both sides of the plasma-cleaned probes were
coated with a 5-nm chromium layer followed by 50 nm of gold. After
gold deposition, the probes were placed in a metal desiccator containing
200 mL of hexadecanethiol, the desiccator was then evacuated to ∼0.1
mbar and kept under vacuum overnight. To ensure the quality of SAMs,
a piece of a silicon wafer was gold coated and functionalized together
with the probes and analyzed by ellipsometry, static contact angle, and
XPS. The analyses did not reveal any notable differences between the
SAMs of hexadecanethiol prepared via vapor-phase deposition and those
prepared by immersion into a 5 mM solution in ethanol and therefore
we assume a similar quality of the films on the Au-coated Si3N4 tips
used in our force measurements.
Adsorption of fibrinogen onto SFM probes was conducted according
to the following protocol: Si3N4 probes, previously cleaned with water-
enriched oxygen plasma, or C16-probes (Digital Instruments, Inc., Santa
Barbara, CA) were placed in phosphate-buffered saline solution (PBS)
containing 0.01 M phosphate buffer, 0.0027 M KCl, and 0.137 M NaCl
(solution obtained by dissolving PBS tablets, P-4417, Sigma Chem.
Co., St. Louis, MO). Fraction I fibrinogen from human plasma (F-
4883, Sigma Chem. Co., St. Louis, MO) was dissolved in PBS solution
at a concentration of 2 mg/mL. This solution was then added to that
containing SFM probes to achieve a final concentration of fibrinogen
of approximately 1 mg/mL. After 1 h of adsorption, more PBS solution
was added, followed by aspirating the liquid with a vacuum line to
remove any fibrinogen film that may have formed at the liquid-air
interface. This procedure was repeated several times, to ensure that
upon removal of the probes from the solution no fibrinogen film was
transferred from the air/water interface onto the probes.
Each set of samples was measured with only one probe to ensure
that observed changes in the tip-surface interaction were not due to
variability in cantilever stiffness or probe-tip radius, although the latter
might be affected by the number of loading-unloading cycles. The
force-distance measurements with the fibrinogen-preadsorbed probes
were conducted in the following manner: at least 64 force-distance
curves were collected in PBS solution, both at the same point and at
adjacent locations.
To gain further insight into the nature of the distinctly different
protein-resistance behavior of the two monolayers, we employed probes
with better-defined surface compositions to “mimic” the nonspecific
interaction of the fibrinogen macromolecule with monolayers of EG3-
OMe. Since proteins contain both hydrophobic and charged domains,27
we used hydrophobic (C16) and oxygen-plasma-treated Si3N4 probes,
which acquire a net negative charge at biologically relevant pH values.
k ) k0*(ω/ω0)2
assuming that the effective mass of the cantilevers is constant. Prior to
force-versus-distance measurements, 100- and 500-nm z-calibration
gratings (TGZ-type, NT-MDT, Zelenograd, Russia) were scanned in
contact mode to ensure proper calibration of the z-piezo. Piezo-
displacement cantilever-deflection curves were converted into force-
distance curves according to the procedure described in ref 29. Zero
separation corresponds to a hard-wall potential, i.e., there is no absolute
measure for the distance between tip and surface. Force-versus-distance
measurements of each series of samples were performed with the same
probe to minimize the error in distance due to variability in the spring
constant value. It is important to note that only semiquantitative analysis
of the force-distance data is possible due to the unknown precision of
spring-constant and probe radii values. In addition, the probe radius
may change due to blunting caused by many repetitions of the
measurements.
Results
In establishing possible differences between the EG3-OMe
SAMs on gold and silver, we first recorded 3 × 3 µm2 SFM
images of the two surfaces in TappingMode with a 5% reduction
in the set-point amplitude. A surface-roughness analysis revealed
no significant differences between the root-mean-square (RMS)
roughness values of the two surfaces, which were both found
to be around 1.6 nm. Hence, the differences in the force-
distance curves on gold and silver described below cannot be
related to differences in surface topography.
Figure 2 shows representative force-separation curves mea-
sured with a fibrinogen-modified probe in pure PBS solution
on a series of surfaces including mica (Figure 2a), hexade-
canethiol-covered gold (Figure 2b), a protein-resistant EG3-OMe
monolayer on gold (Figure 2c), and a protein-adsorbing film
on silver (Figure 2d). According to FTIR measurements, the
latter adsorbed about 20% of a fibrinogen monolayer.
(25) Erlandsson, R.; Olsson, L.; Bongrand, P.; Claesson, P. M.; Curtis,
A. S. G., Eds.; Springer: Berlin, 1994, Chapter 4.
(26) Ta, T. C.; Sykes, M. T.; McDermott, M. T. Langmuir 1998, 14,
2435-2443.
(27) Andrade, J. D.; Horbett, T. A.; Brash, J. L., Eds. ACS Symp. Series
No. 343; American Chemical Society: Washington, DC, 1987; Chapter 1.
(28) Feng, L.; Andrade, J. D.; Horbett, T. A.; Brash, J. L., Eds. ACS
Symp. Series No. 602; American Chemical Society: Washington, DC, 1995;
Chapter 5.

Oligo(ethylene glycol)-Terminated Monolayers
J. Am. Chem. Soc., Vol. 121, No. 43, 1999 10137
Figure 2. Representative force-versus-separation curves obtained with a fibrinogen-modified Si3N4 probe (k ) 0.03 N/m) in PBS solution on (a)
freshly cleaved mica surface, (b) hexadecanethiol SAM on gold, (c) EG3-OMe SAM on gold, and (d) EG3-OMe on silver showing attractive
interaction and multiple pull-offs (representative of 20% of the force-distance data taken).
The hydrophilic (mica) and the hydrophobic (C16) surfaces
are intended to serve as a reference for strong fibrinogen binding.
Both surfaces show an attractive interaction with the fibrinogen-
modified probe upon approach and strong adhesion upon
retraction, indicating that the proteins establish contact to both
probe and sample surfaces. Note the difference in the extent of
the attractive interaction of fibrinogen between the two surfaces.
Clearly, the attractive force upon approach, the pull-off forces,
and the adhesion hysteresis (the area between the approaching
and retracting curves) were all greatest for the fibrinogen-C16
system, which is in good agreement with results reported in
the literature.26
forces for the SAM were determined to be 11.7 ( 1.4 nN on
gold and 10.6 ( 0.3 nN on silver. The broader distribution of
the pull-off forces from the EG3-OMe on gold might be due to
the variations of the elastic modulus of the SAM on gold, which
would, in turn, depend on the degree of crystalline order in the
monolayer. The similarity of the pull-off forces on gold and
silver in air measured with hydrophilic, largely hydroxylized
probes is consistent with the similar water contact angles
measured for the two monolayers.7 Similar water-contact angles
correspond to similar surface interfacial energies with water and,
therefore, to similar values of the work of adhesion which, in
turn, is reflected in the pull-off forces.
As can be seen in Figure 2c, there are no attractive forces
between a fibrinogen-modified probe and an EG3-OMe mono-
layer on gold. However, there is a reproducible loading-
unloading hysteresis, which did not change when pure buffer
was replaced by fibrinogen solution. Two hundred and fifty six
force curves consecutively recorded in pure PBS and fibrinogen
solution gave no evidence of attractive or adhesive forces. This
was reproduced on several samples.
Similar observations have been reported by Sheth and
Leckband by means of a surface forces apparatus.30 These
authors, however, measured an attractive interaction after
pressing streptavidin onto PEG. The very small attractive forces
found in their experiment might also be present here, but the
effect is too small to be unambiguously detected.
Force-versus-distance measurements were also performed
with an oxygen-plasma-treated Si3N4 probe (k ) 0.01 N/m) in
deionized (DI) water (18.2 MΩ‚cm) and PBS buffer (Figure
3). We previously found32 that probes treated with water-rich
oxygen plasma have an isoelectric point at pH ≈3, which implies
that at higher pH values the probes acquire a negative surface
charge due to deprotonation of hydroxyl groups. Our measure-
ments of the EG3-OMe SAMs on gold and silver in DI water
with a negatively charged probe showed a long-range repulsion
followed by a short-range attraction, similar to a DLVO-type
interaction.33 The repulsive force was found to be stronger for
the SAM on gold, while the pull-off forces were greater for
that on silver. Measurements in PBS buffer did not reveal any
significant differences between the two surfaces. As expected,
the ion concentration had a significant effect on the observed
interaction: for both surfaces attractive forces are observed in
For the EG3-OMe monolayer on silver, we observed multiple
pull-off events indicating adhesion of the fibrinogen-coated
probe, in agreement with the FTIR protein-adsorption measure-
ments (Figure 2d).
(29) Ducker, W. A.; Senden, T. J.; Pashley, R. M. Nature 1991, 353,
239.
Measurements with oxygen-plasma-treated probes were per-
formed in air, deionized water, and the same PBS buffer as used
in the protein-adsorption experiments. Force-versus-distance
curves in air with an oxygen-plasma-treated probe (k ) 0.03
N/m) revealed no significant differences between the EG3-OMe
SAMs on gold and silver. A large adhesion hysteresis primarily
due to capillary forces31 was found for both surfaces. Pull-off
(30) Sheth, S. R.; Leckband, D. Proc. Natl. Acad. Sci. 1997, 94, 8399-
8404.
(31) Grigg, D. A.; Russel, P. E.; Griffith, J. E. J. Vac. Sci. Technol. A
1992, 10, 680.
(32) Feldman, K.; Tervoort, T.; Smith, P.; Spencer, N. D. Langmuir 1998,
14, 372-378.
(33) Israelachvili, J. Intermolecular and Surface Forces, 2nd ed.;
Academic Press: San Diego, CA, 1992.

10138 J. Am. Chem. Soc., Vol. 121, No. 43, 1999
Feldman et al.
Figure 3. Representative force-versus-separation curves obtained with an oxygen plasma-treated Si3N4 probe (k ) 0.02 N/m) measured on (a)
EG3-OMe on gold in deionized water, (b) EG3-OMe on silver in deionized water, (c) EG3-OMe on gold in PBS solution, and (d) EG3-OMe on
silver in PBS solution.
the PBS experiment, with a significantly smaller range of
interaction compared to measurements in DI water.
The most interesting results were obtained with hydrophobic
probes. Measurements in pure water with C16-probes revealed
a long-range repulsion for the EG3-OMe monolayer on gold
and a long-range attraction for that on silver (Figure 4). When
DI water was replaced by PBS solution, the extent of the
repulsive interaction diminished on gold, but did not change
sign. For silver, however, only very small changes were found
in PBS buffer as compared to pure water (Figure 4). The
experiments were repeated many times with different samples
of EG3-OMe and several batches of C16-probes to confirm their
correlation with the FTIR protein-adsorption measurements. We
also investigated two other 1-undecylthiols, either with a
hydroxyl-terminated hexa(ethylene glycol) (EG6-OH) or a
methoxy-terminated tri(ethylene glycol) (EG3-OMe) end group
and a -CH2OCH3 side chain at the C-12 atom (EG[3,1]-OMe).
Force-distance measurements on monolayers of either thiol on
gold or silver established the same correlation for all investigated
films. Whenever a repulsive force was observed upon approach
of a hydrophobic probe in PBS solution, that surface resisted
fibrinogen adsorption as determined by the FTIR measurements.
The repulsive curves measured upon approach to EG3-OMe
on gold showed some variability with respect to the range of
interaction. Some of them were purely repulsive, while oc-
casionally a jump-to-contact at distances of ∼5 nm before
contact appeared. Whenever a jump-to-contact occurred, this
was accompanied by a hysteresis in the loading-unloading
Figure 4. Advancing force-versus-separation curves measured with a
hydrophobic C16-probe (k ) 0.12N/m) in deionized water and PBS
solution on the SAMs of EG3-OMe on gold and silver.
cycle, while no hysteresis was found in the purely repulsive
curves. This variability of the SFM data in terms of the range
of the forces and the presence of an adhesion hysteresis might
be attributed to the structural inhomogeneity of the monolayers,
i.e., the presence of different phases with dissimilar mechanical
properties in the EG3-OMe layer.
Force-distance measurements with a C16-probe and an EG3-
OMe SAM on gold in DI water and PBS solution revealed a
strong dependence of the range of the repulsive force upon the
ionic strength of the solution; the same measurements for the
monolayer on silver showed a long-range attractive interaction
with very little dependence on the ionic strength. These results
suggest that, due to different conformations of the EG3-OMe
molecules on gold and silver, the observed interactions may
well be of a distinct physical nature, i.e., electrostatic for EG3-
OMe on gold and hydrophobic for that on silver. To further
elucidate this hypothesis, we performed force-distance mea-

Oligo(ethylene glycol)-Terminated Monolayers
J. Am. Chem. Soc., Vol. 121, No. 43, 1999 10139
Figure 5. Advancing force-versus-separation curves measured with a hydrophobic C16-probe (k ) 0.06 N/m) in aqueous solutions of KNO3 on
(a) the SAMs of EG3-OMe on gold and silver and (b) PEG2000-thiolate on gold. Solid lines represent fitted data.
surements with a C16-probe on the two monolayers in aqueous
solutions of KNO3 of various ionic strengths (Figure 5a), and,
once again, detected repulsive forces for the EG3-OMe on gold
that displayed a strong dependence upon the ionic strength of
the solution and attractive long-range forces for that on silver
that were less influenced by the presence and concentration of
ions.
We also performed measurements on the end-grafted PEG
2000 brush on gold with a C16-probe (Figure 5b) to demonstrate
that the long-range effects observed on packed, ordered mono-
layers of EG3-OMe are not detected in grafted polymer brushes.
Also, in marked contrast to the behavior of the shorter
molecules, end-grafted PEG scarcely revealed any ionic-strength
dependence.
affinity of fibrinogen. However, it is helpful to be able to exclude
surface roughness as a possible variable in our experiments.
The presence of a hysteresis in the repulsive part of the force
curve measured with a fibrinogen-coated C16-probe for EG3-
OMe on gold (Figure 2c) indicates the presence of energy-
dissipating processes, which can be attributed to viscoelastic
conformational changes in the protein layer induced by the
loading pressure. Although pressure-induced changes in protein
conformation during loading and unloading might lead to
aggregation of protein in the contact area, giving rise to long-
range repulsive forces,34 successive measurements at a single
surface location showed no changes in the interaction. This also
confirms that the protein layer recovers upon the retraction of
the probe and that no plastic deformation is induced.
According to previous theoretical work on these systems,10
strong interaction of water molecules with the helical conformer
of the EG3-OMe SAM on gold is a necessary condition for
protein resistance, because oligo(ethylene glycol) per se is not
intrinsically protein resistant. Water cannot associate with the
all-trans structure of the EG3-OMe SAM on silver, and hence
this system exhibits protein-adsorption behavior that is char-
acteristic for a slightly hydrophobic surface. To demonstrate
unambiguously that the solvent plays a crucial role in explaining
the difference in the adsorption characteristic of the helical and
all-trans conformers, force-distance measurements with a
hydrophobic probe were performed in perfluorodecalin C10F18
(Fluorochem, U.K.), which is a nonpolar, non-hydrogen-bonding
liquid with a low dielectric constant (ꢀ ∼ 2.0). In agreement
with the theoretical arguments, we did not observe any long-
range forces in perfluorodecalin on the two SAMs (Figure 6).
Instead we found attractive interactions on both surfaces.
The interaction between charged bodies in electrolyte solu-
tions can be described by the DLVO theory, which takes into
account both electrostatic and van der Waals forces.33 For low
surface potentials (ψ < 25 mV) the electrostatic force per unit
area between the two bodies bearing charge densities σ1 and σ2
can be described by the following equation:35,36
2
ꢀꢀ0
Fel ) [(σ12 + σ22)e-2κD + σ1σ2e-κD
]
(1)
For larger separations, electrostatic force between a sphere
of radius R and a flat surface becomes:36,37
4πRσ1σ2
Fel )
e-κD
(2)
ꢀꢀ0κ
where 1/κ is the so-called Debye length and D is the separation
between the two bodies. The Debye length, being a solution
property, can be calculated from the following equation:33
Discussion
The images recorded for EG3-OMe on gold and silver did
not reveal any differences in roughness. Since the protein-
adsorption experiments were conducted using fibrinogen, which
is known to have an extremely high surface activity,28 one would
not expect roughness to be a factor influencing the surface
(34) Claesson, P. M.; Blomberg, E.; Froberg, J. C.; Nylander, T.;
Arnebrant, T. AdV. Colloid Interface Sci. 1995, 57, 161-227.
(35) Parsegian, V. A.; Gingell, D. Biophys. J. 1972, 12, 1192-1204.
(36) Butt, H.-J. Biophys. J. 1991, 60, 777-785.
(37) Butt, H.-J.; Jaschke, M.; Ducker, W. Bioelectrochem. Bioenerg.
1995, 38, 191.

10140 J. Am. Chem. Soc., Vol. 121, No. 43, 1999
Feldman et al.
1/2
F∞ie2z2i
lengths of 11 nm in pure water and 6 nm in PBS buffer (Figure
4). These are typical decay lengths for hydrophobic forces.
Although the molecular origin of hydrophobic forces is not well
understood43 and the extent of the interaction is often ques-
tioned,44 there is substantial experimental evidence of a long-
range attractive interaction in an aqueous environment between
hydrophobic surfaces with the attractive forces sometimes being
sensed at separations of 50 nm and more.41,45
In contrast to the attractive hydrophobic interaction between
EG3-OMe film on silver and the C16-probe, a strong depen-
dence of the repulsiVe force on the ion concentration observed
on the EG3-OMe monolayer on gold in DI water and PBS buffer
suggests that this interaction involves a significant electrostatic
contribution. Furthermore, since the helical or amorphous
structure of the OEG tails of the SAM on gold allows a strong
interaction of water molecules via hydrogen bonding with
oxygen atoms of the OEG, there might be an additional repulsive
contribution of hydration or steric-protrusion forces, which decay
roughly exponentially with a distance of 0.2-0.4 nm.42,45 This
short-range, exponentially decaying force observed in PBS
solution looks similar to data obtained on lipid bilayers in pure
water with SFM46 and the surface forces apparatus (SFA),47
and is assigned to steric/hydration forces there.33,46,47
κ )
m-1
(3)
∑
(
)
ꢀꢀ0kBT
i
where F∞i is the bulk ionic concentration, zi is the ionic charge,
kB is Boltzmann’s constant, e is the electronic charge, and T is
the absolute temperature.
One important implication of eq 1 is that the overall
electrostatic interaction is not eliminated when one of the surface
charge densities is set to zero, e.g. σ1 ) 0 and σ2 * 0. It is
therefore not necessary for the both surfaces to carry charges.
The issue of surface charges on the EG3-OMe monolayer
on gold is also important. The tails of EG3-OMe on gold can
easily accommodate water molecules and seem to be present
in both helical and amorphous states; for such “soft”, permeable
interfaces with polar tails it has been shown that the effectiVe
surface charge density, σeff, depends on the surface charge
density, σ, and the dipole density, ν:38,39,40
σeff ) [σ cosh(κl) + νκ sinh(κl)]e-κl
(4)
where l is the thickness of the “soft” polar region. Therefore, if
there are no surface charges (or the net surface charge is zero)
but there is a dipole field, a surface would bear an effective
non-zero charge, which would give rise to an electrostatic
interaction.
Study of the ionic concentration effect on the range of forces
measured between the EG3-OMe SAMs and C16-probes (Figure
5a) provided further evidence that the overall interaction with
the monolayer on gold had a significant electrostatic component.
Data for EG3-OMe on gold shown in Figure 5a were fitted with
an electrostatic force law for a case where the second surface
(in our case, that of the C16-probe) does not carry a surface
charge (see eq 1). Measurements of the symmetric system C16-
probe/C16-surface did not show any repulsive interaction in
water and salt solutions indicating that there were no charges
present on these surfaces and the interaction was found to be
strongly attractive and occurred at separations of about 20 nm.
Long-range repulsive forces followed by short-range attraction
were measured in DI water for the EG3-OMe monolayers on
gold and silver substrates with a negatively charged oxygen-
plasma-treated Si3N4 probe (Figure 3). When comparing the
force curves in DI water and PBS buffer it becomes clear that
the ion concentration has a pronounced effect on the tip/substrate
interaction. The two substrates differ only in the strength of
the repulsive force. This difference may be explained if we now
assume that, due to conformational differences, the EG3-OMe
monolayer on gold carries an effective surface charge (eq 4),
while that on silver does not. There will still be a repulsive
electrostatic interaction on silver (see eq 1, with σ2 ) 0) but it
will be weaker than that for gold where σ1 * 0 and σ2 * 0.
Upon introduction of the PBS solution, the range of the
electrostatic force is, as expected, dramatically reduced. The
screening of the electrostatic repulsion is due to the high ionic
strength of the PBS solution (we calculated the Debye length
of the PBS solution to be 1/κ ) 0.76 nm at room temperature
compared to ∼1 µm for DI water) and only the attractive van
der Waals forces are clearly observed (Figure 3). In addition,
the type of interaction changed from repulsive in pure water to
attractive in PBS buffer, the range was found to be much shorter
in the case of high ion concentrations, and almost no hysteresis
was found for measurements performed in PBS buffer, although
the hysteresis might also scale with the Debye length and hence
be too small to be detectable here.
-
Note that traces of ions (in particular, HCO3 from dissolved
CO2) present in the water may reduce the Debye length of dilute
solutions significantly.
Although for the case of charge on one side only, eq 1 reduces
to Fel e-2κD, we found that for most measurements and higher
ionic strengths the experimental data were best fitted with Fel
e-κD. We performed measurements using a C16-probe and a
bare SiOx surface in aqueous solutions of KNO3salso a “charge-
on-one-side-only” systemsand also observed in this case that
the data were best fitted with the e-κD coefficient. We believe
that this is an image force effect, which comes into play for
counterions in water that are surrounded by surfaces of higher
dielectric constant.33
Results of the fits to the EG3-OMe data are shown in Table
1. The radius of the probe used was checked by scanning over
the ridges of a SrTiO3 single crystal48 and was found to be ∼110
nm. Dipole moments per molecule were calculated from dipole
density, ν, assuming l ) 1 nm in eq 4 and packing density of
21.3 Å2/molecule.7 Data for EG3-OMe on silver were fitted with
The EG3-OMe monolayer on silver consists of molecules that
are tightly packed and expose their hydrophobic methyl groups
at the film/water interphase. The observed long-range attractive
forces between the hydrophobic C16-probe and the EG3-OMe
SAM on silver in water and PBS solution (Figure 4) point to
the presence of hydrophobic interaction.41,42 Indeed, these force
curves are well reproduced by exponential fits with decay
(41) Yoon, R. H.; Flinn, D. H.; Rabinovich, Y. I. J. Colloid Interface
Sci. 1997, 185, 363-370.
(42) Israelachvili, J.; Wennerstro¨m, H. Nature 1996, 379, 219-225.
(43) Tsao, Y.-H.; Evans, D. F.; Wennerstro¨m, H. Science 1993, 262,
547-550.
(44) Wood, J.; Sharma, R. Langmuir 1995, 11, 4797-4802.
(45) Claesson, P. M.; Bongrand, P.; Claesson, P. M.; Curtis, A. S. G.,
Eds. Springer: Berlin, 1994, Chapter 2.
(38) Bell, G. M.; Levine, P. L. J. J. Colloid Interface Sci. 1980, 74,
530-548.
(39) Belaya, M.; Levadny, V.; Pink, D. A. Langmuir 1994, 10, 2010-
(46) Dufreˆne, Y. F.; Barger, W. R.; Green, J. B. D.; Lee, G. U. Langmuir
1997, 13, 4779-4784.
(47) Marra, J.; Israelachvili, J. Biochemistry 1985, 24, 4608.
(48) Sheiko, S. S.; Mo¨ller, M.; Reuvekamp, E. M. C. M.; Zandbergen,
H. W. Phys. ReV. B 1993, 48, 5675.
2014.
(40) Xu, W.; Blackford, B. L.; Cordes, J. G.; Jericho, M. H.; Pink, D.
A.; Levadny, V. G.; Beveridge, T. Biophys. J. 1997, 72, 1404-1413.

Oligo(ethylene glycol)-Terminated Monolayers
J. Am. Chem. Soc., Vol. 121, No. 43, 1999 10141
Table 1. Fitting Results of the Data Shown in Figure 5a
EG3-OMe on gold
KNO3 1/κ calcd, 1/κ fitted, dipole moments/ EG3-OMe on Ag
concn, M
nm
nm
molecule, D
1/λ, nm
10-4
10-3
10-2
10-1
30.7
9.7
3.1
14.5
9.3
2.7
4.6
2.5
2.2
1.4
9.6
8.3
5.1
3.1
1.0
0.7
empirical formula41 of the form Fh exp(-λD) to demonstrate
that, unlike the situation on gold, the constant 1/λ does not
correlate with the calculated Debye length of the solution.
Apparently, while the molecular conformers of OEG on the gold
surface generate sufficiently strong dipolar fields to cause a
screenable electrostatic interaction with the C16-probe, the
interaction on silver is dominated by nonelectrostatic forces. A
detailed discussion on the difference in dipolar moments for
the helical, amorphous, and planar “all trans” conformers in
contact with water will be presented elsewhere.49
Experiments with end-grafted PEG 2000 show a clear
distinction between PEG brush behavior and that of EG3-OMe
monolayers. The brushes exhibit no long-range interactions and
little ionic strength dependence (Figure 5b). In fact, the force
was found to become slightly more repulsive with increasing
ionic strength, since the 0.1 M KNO3 solution represents good
solvent conditions.33 The lack of long-range interactions for the
polymer brush demonstrates that the repulsive interactions found
for the short-chain oligomers is unique and not related to the
steric repulsion effect responsible for the protein resistance for
the end-grafted polymers.
Figure 6. Representative force-versus-separation curves measured with
the C16-probe (k ) 0.12 N/m) in perfluorodecalin on (a) EG3-OMe
on gold and (b) EG3-OMe on silver.
Finally, the data recorded for EG3-OMe on Au and Ag under
perfluorodecalin (Figure 6) illustrate the fact that the difference
in the interfacial force does not only depend on the molecular
conformation, but also on the solvent, and is significant only
in an aqueous environment. Hence, to understand resistance to
protein adsorption of biological or organic surfaces, it is
important that the complete system, i.e., film and solvent, be
considered.
to be hydrophobic in nature, the long-range repulsive force
measured on the gold system displayed a strong ionic strength
dependence, implying an important electrostatic component. The
difference in sign of the forces can be correlated with the distinct
molecular conformations of the OEG tails and the resulting
difference in their dipolar fields. This effect dominates the
interaction between the OEG-derivatized SAM and the SPM
probe in the case of the helical and amorphous conformers on
gold but is too small to contribute significantly to the total
interaction in the case of the planar, “all trans” conformers on
silver.
Conclusions
By means of a scanning force microscope with well-defined,
functionalized tips, we were able to elucidate the contributions
that are relevant to protein adsorption of fibrinogen on SAMs
of EG3-OMe on gold and silver.
When using a hydrophobic tip to mimic hydrophobic regions
of the protein, interactions of opposite sign were observed on
the EG3-OMe-Au and EG3-OMe-Ag systems in aqueous media.
While the attractive force observed on the silver system appeared
Acknowledgment. We thank the Deutsche Forschungsge-
meinschaft, the Office of Naval Research, and the Council of
the Swiss Federal Institutes of Technology (PPM and MINAST
Program) for their financial support of this work. We thank W.
Eck and S. Herrwerth for the EG3-OME and PEG 2000
materials.
(49) Zolk, M.; Buck, M.; Eisert, F.; Grunze, M.; Wang, R.; Kreuzer, H.
J. In preparation.
JA991049B