iron oxide magnetic nanoparticles used as probing agents to study the nanostructure of mixed...
TRANSCRIPT
Iron oxide magnetic nanoparticles used as probing agents to study thenanostructure of mixed self-assembled monolayers
Benoit P. Pichon,*a Gregory Barbillon,b Pascal Marie,c Matthias Paulya and Sylvie Begin-Colina
Received 5th July 2011, Accepted 25th August 2011
DOI: 10.1039/c1nr10729a
Self-assembled monolayers (SAMs) of organic molecules are of exceptional technological importance
since they represent a convenient, flexible, and simple system for tuning the chemical and physical
properties of surfaces. The fine control of surface properties is directly dependent on the structure of
mixed SAMs which is difficult to characterize at the nanoscale with usual techniques such as scanning
probe microscopies. In this study, we report on a general method to investigate at the nanoscale the
structure of molecular patterns which consist in SAMs of two components. Iron oxide nanoparticles
(NPs) have been used as probing agents to study indirectly the structure of mixed SAMs. Mixed SAMs
were prepared by the replacement of mercaptododecane (MDD) adsorbed by mercaptoundecanoic
acid (MUA) molecules on gold substrates. Therefore, the SAM surface displays both chelating
carboxylic terminal groups and non-chelating methylene terminal groups. As NPs have been previously
demonstrated to specifically interact with carboxylic acid groups, the increasing density in NPs was
correlated with the evolution of the COOH/CH3 terminal groups ratio. Therefore the structure of
mixed SAMs was studied indirectly as well as the kinetic of the replacement reaction and its mechanism.
With this aim, we took advantage of the SPR properties of the gold substrate and of the high refractive
index of iron oxide nanoparticles to follow their assembling on mixed SAMs as a time resolved study.
The high sensitivity and tuning of the SPR signal over a wide range of wavelengths are correlated with
the NP density. Furthermore, SEM combined with image analysis has allowed studying the
replacement rate of MDD by MUA in SAMs. We took also advantages of the magnetic properties of
NPs to evaluate qualitatively the replacement of thiol molecules.
Introduction
Self-assembled monolayers (SAMs) of organic molecules are of
exceptional technological importance for a wide range of appli-
cations depending on surface modifications.1,2 They represent
a convenient, flexible, and simple system for processing and
modifying the chemical and physical properties of surfaces
(hydrophilic/phobic, self-cleaning, antibacterial, etc.). SAMs are
usually prepared by the spontaneous adsorption of alkanethiols
on gold substrates. A wide range of molecules has been used to
generate well-defined organic surfaces which display functional
groups. An interesting achievement consists in mixed SAMs of
multiple components which offer a high tunability of the surface
activity as well as patterning down to the nanoscale.3 Mixed
SAMs are usually prepared by the co-adsorption of two different
molecules which is difficult to control since the molar ratio of
both components adsorbed on surfaces usually differs from that
of the solution.1 Indeed, the relative solubility of the molecules,
their relative interaction with solvent and the substrate and the
intrinsic properties of the head and tail groups will induce pref-
erential molecular interactions between adsorbates of the same
type.4,5 Therefore the composition and the structure of such
SAMs often result in phase segregation.4,6,7 Nanodomains have
been shown several times in mixed SAMs by STM7–10 and
AFM.7,11 While phase segregation results in the patterning of
surfaces, its formation process is still unclear and usually
proceeds to intermediate and partial phase separation which is
difficult to control. Therefore, the SAM structure is difficult to
predict since many parameters have to be taken into account.
Another method to pattern surfaces by using mixed SAMs
consists in taking advantage of the reversible absorption of thiol
molecules on gold surfaces.12–15 Their preparation consists in the
gradual replacement of self-assembled molecules in the mono-
layer by others and is reported to occur preferentially at
boundaries and defect sites. In contrast to the co-adsorption
aInstitut de Physique et de Chimie des Mat�eriaux de Strasbourg (UMRCNRS-UdS 7504), 23 rue du Loess – BP43, 67034 Strasbourg Cedex 2,France. E-mail: [email protected]; Fax: +33 (0)3 88 10 72 47;Tel: +33 (0)3 88 10 71 33bLaboratoire Charles Fabry de l’Institut d’Optique (CNRS UMR 8501),Universit�e Paris-Sud, Campus Polytechnique, RD128, 2 avenue AugustinFresnel, 91127 Palaiseau Cedex, FrancecInstitut Charles Sadron, Universit�e de Strasbourg, CNRS UPR 22, 23 ruedu Loess – BP 84047, 67034 Strasbourg Cedex 2, France
4696 | Nanoscale, 2011, 3, 4696–4705 This journal is ª The Royal Society of Chemistry 2011
Dynamic Article LinksC<Nanoscale
Cite this: Nanoscale, 2011, 3, 4696
www.rsc.org/nanoscale PAPER
method which is ruled by the competitive adsorption of mole-
cules, the replacement method enables to direct the surface
patterning more precisely by the spatial arrangement of defect
sites and the replacement time. The replacement of hex-
adecanethiol (HDT) monolayers with 12-mercaptododecanoic
acid (MDDA) has been shown to proceed domain wise and to
result in phase separation in the SAM into sizable domains.14
Nevertheless, the structure of mixed SAMs is very difficult to
investigate precisely. Unless analysis techniques such as XPS
(composition) and scanning probe microscopies (spatial
arrangement) have greatly enhanced the understanding and the
optimization of the patterning of SAMs at the nanoscale, they
still remain difficult to process.10 STM or AFMoperates at a very
local scale and molecules with similar lengths are difficult to
discriminate when absorbed on a surface. An alternative is the
use of probing agents to study indirectly the structure of mixed
SAMs. Recently, iron oxide nanoparticles were found to be
useful to study the patterning of surfaces by mixed SAMs
prepared by the absorption method.16 The spatial arrangement
of chelating carboxylic acid and non-coordinating methylene
terminal groups at the SAMs surface was investigated as
a function of the NPs assembling. Indeed, nanoparticles were
observed to assemble only on specific regions of SAMs decorated
with carboxylic acid.
Nanoparticles may be used as probing agents which enable to
study indirectly the structure of mixed SAMs on gold substrates.
Furthermore, the use of gold substrate may enable to study the
kinetic of the NPs assembling by the Surface Plasmon Resonance
(SPR) technique. In recent decades, the SPR technique became
very attractive to probe and characterize chemical modifications
on noble metal surfaces such as gold.17 The SPR is very sensitive
to surrounding medium and, in particular, to small local changes
of the refractive index. Therefore, SPR was used to characterize
surface modifications resulting from the adsorption of organic
molecules.18 Gold nanoparticles, also featured by localized
surface plasmon (LSPR), were naturally used to improve the
sensitivity of this technique by extending the specific surface of
the film.19–26 They were assembled onto gold substrates func-
tionalized by SAMs of thiol molecules. However, the use of non-
metallic nanoparticles to study the changes of the SPR signal has
been poorly reported.22,27 Among them, iron oxide nanoparticles
have been used to detect very low concentration of biological
materials by the SPR technique with a high sensitivity and
selectivity,28–32 thanks to their high refractive index (2.42 for
magnetite).33
In this study, we report on a general method to investigate the
structure of molecular patterns at the nanoscale. Iron oxide
nanoparticles are used as probing agents to perform indirect and
time resolved studies of the structure and of the formation
mechanism of mixed SAMs. Mixed SAMs were prepared by the
gradual replacement of molecules containing non-chelating
methylene terminal groups by molecules displaying chelating
carboxylic terminal groups at their surface. The composition of
the SAMs and the spatial arrangement of methylene and
carboxylic acid terminal groups are investigated as a function of
the replacement time. As nanoparticles interact specifically with
carboxylic acid groups, their assembling on mixed SAMs enables
to determine the patterning of the surface as a function of the
replacement time. The assembling of NPs was studied by using
the SPR technique which was combined with SEM image anal-
ysis. The use of high refractive index iron oxide nanoparticles
results in the enhancement of the variation of the SPR properties
of gold substrates to determine quantitatively the NPs density.
We took also advantages of the magnetic properties of NPs to
evaluate qualitatively the replacement of organosulfur
molecules.
Experimental section
Chemicals and materials
11-Mercaptoundecanoic acid (MUA) and mercaptododecane
(MDD) were purchased from Aldrich. Absolute ethanol and
tetrahydrofurane were used as received.
Gold film preparation
Glass substrates are cleaned in a freshly prepared piranha solu-
tion (3 : 1 H2SO4 (98%), H2O2 (30%)) for 30 min. Once cooled,
the glass substrates are rinsed abundantly with deionized water
and dried with N2 gas. The last step is to evaporate a gold thin
layer (48 nm) in order to realize the metallic film. Previously, an
adhesion layer (Cr) for gold is evaporated (2 nm).
Preparation of self-assembled monolayers (SAMs)
Freshly cleaned ion sputtered gold substrates under O2/Ar
plasma (2 min) were used for all preparation. Mixed SAMs were
prepared by the replacement method which occurs in two
steps.14,34,35 First, MDD-SAMs were prepared by soaking gold
substrates at room temperature for 24 hours in 10 mmol etha-
nolic solutions of mercaptododecane (MDD). Second, freshly
washed and dried MDD-SAMs were then immersed in
a 10 mmol ethanolic solution of 11-mercaptoundecanoic acid
(MUA) at room temperature for different times (30 minutes, 2, 4,
7, 24, 96, 170 and 288 hours). SAM substrates were then washed
extensively with absolute ethanol to remove unbound thiol
molecules and dried under N2 stream. The corresponding SAMs
were respectively named SAM-0.5H, SAM-2H, SAM-4H, SAM-
7H, SAM-24H, SAM-96H, SAM-170H and SAM-288H. Both
molecules have similar chain lengths to maximize the reciprocal
influence of the chelating and non-chelating abilities of carbox-
ylic and non-polar methylene terminal groups towards the mean
surface activity and mixture.
Preparation of iron oxide nanoparticles (NPs)
Iron oxide NP were produced by the thermal decomposition
method and following the procedure that is detailed in ref. 36–38.
It consists in the preparation of an iron(III)/oleate complex
(Fe(oleate)3), which is thermally decomposed in a high boiling
solvent in the presence of oleic acid. Fe(oleate)3 was prepared
from FeCl3$H2O (10.8 g, 40 mmol, 97%, Aldrich) which was
dissolved in 60 mL H2O (Milli-Q) and 80 mL ethanol. This
solution was mixed with a solution of sodium oleate (36.5 g,
120 mmol, 82%, Riedel-de Ha€en) dissolved in hexane (140 mL)
and refluxed at 70 �C for 4 h. The organic phase containing the
iron oleate complex was separated, washed three times with
distilled water (30 mL) to extract salts, dried using MgSO4, and
This journal is ª The Royal Society of Chemistry 2011 Nanoscale, 2011, 3, 4696–4705 | 4697
finally hexane was evaporated. The resulting iron oleate complex
was a reddish-brown viscous solution and stored at 4 �C. A
combination of Fe(oleate)3 (2 g, 2.2 � 10�3 mol), oleic acid
(1.24 g, 3.3 � 10�3 mol) and octyl ether (20 mL) was stirred for
1 h to dissolve the reactants. The temperature was carefully
raised to reflux with a heating rate of 5 �C min�1 without stirring
for 120 min under air. After cooling down to room temperature,
the black suspension of nanocrystals was washed 3 times by
addition of ethanol and centrifugation (8000 rpm, 10 min). The
obtained nanocrystals could be easily suspended in various
organic solvents to raise a highly stable suspension which can be
stored for several months. The size monodispersity of NPs was
improved by applying a size selection precipitation process.39 The
nanoparticles were suspended in hexane at a concentration of
1 mg mL�1 and precipitated by adding the same volume of
acetone followed by centrifugation. The precipitate was redis-
persed in tetrahydrofurane (THF) to prepare a highly stable
suspension of coated nanoparticles with a specific concentration
of 3.7 mg mL�1.
Assembling of iron oxide nanoparticles on SAMs
Replaced SAMs were immersed directly after preparation in the
suspension of oleic acid coated nanoparticles in THF at room
temperature for only 10 minutes. Each substrate was then
extensively rinsed with THF and placed in an ultrasonic bath in
THF for 10 seconds to remove any physisorbed nanoparticles
and finally dried under a stream of nitrogen.
Characterization of the samples
Scanning electron microscopy was performed using a JEOL 6700
microscope equipped with a field emission gun (SEM-FEG)
operating at an accelerating voltage of 3 kV. Atomic Force
Microscopy (AFM) was performed using a Digital Instrument
3100 microscope coupled to a Nanoscope IIIa recorder.
Measurements were done in the tapping mode onto substrates
before and after exposition to the suspension of nanoparticles.
Collected data were analyzed with WSXM software.40
Image analysis
The digitized SEM images were analyzed with the Visilog�
software (Noesis, France) following an edge detection procedure
to determine the density of nanoparticles and their position.
Several techniques were combined such as filtering (background
correction and Gaussian smoothing for removing detail and
noise), edge detection methods (Laplacian of Gaussian and zero-
crossing operators), and adaptive threshold to locate the
boundary (or edge) of each particle. The nanoparticles were
analyzed from SEM pictures as we reported previoulsy16 so as to
extract their number and positions within the image (X and Y
coordinates of the centre of mass). Densities in NPs were
calculated on each mixed SAMs from the number of nano-
particles on a precise area. The surface of the SAMs covered by
NPs was deduced according to the size, the spherical shape and
the density in NPs. NPs can be assimilated to circles with
a maximum compacity of 0.9 if they are ideally organized in a 2D
hexagonal closed packed (hcp) lattice. Therefore the effective
surface covered by NPs is compared to the maximum value of
90%.
Surface Plasmon Resonance measurements
The used setup is the Kretschmann configuration with
a TM-polarized white light source and a fixed angle of illumi-
nation (q ¼ 57�) (Scheme 1). The wavelength range used in our
experiments is 600–900 nm. An equilateral SF10 glass prism has
been used, and for glass substrates, the refractive index is similar
to the glass prism. The plane face of the prism was coupled to the
SF10 glass substrates via index matching oil (Cargille Labora-
tories Inc.). The water is used as reference, because the spectral
shifts due to the adsorption of magnetic nanoparticles will
remain in the 600–900 nm range, which is our work range. The
spectral measurements were chosen in order to use a simple
model described by Campbell et al. (see the Results and discus-
sion section) to determine the effective thickness of the adsorbed
layer from the spectral shifts and thus to calculate the paving
density of magnetic nanoparticles. With the value of effective
thickness d, we determine the paving density by dividing this
value d by the magnetic nanoparticles size (diameter), which is
12 nm. Then, this value is divided by 0.9 corresponding to the
compacity of the hexagonal 2D lattice, and we thus obtain the
paving density value expressed as a percentage of the maximum
value of paving density (here 0.9).
Magnetic measurements
The magnetic properties of nanoparticles assembled on SAMs
were investigated by using a superconducting quantum inter-
ference device (SQUID) magnetometer (Quantum Design model
MPMS-XL). The magnetization was recorded at 5 K as a func-
tion of the applied magnetic field. The substrates containing the
assemblies of nanoparticles were placed parallel to the applied
magnetic field. The magnetization values measured for nano-
particles assembled on SAMs were normalized according to the
weight of each substrate.
Results and discussion
The assembling of NPs in 2D arrays can be controlled by the
surface patterning generated by SAMs displaying different
functional head groups.16,41 Nanoparticles have been demon-
strated to assemble specifically on SAM surfaces decorated with
carboxylic acid head groups and not with methylene head
Scheme 1 Simplified scheme of principle of the surface plasmon reso-
nance measurements.
4698 | Nanoscale, 2011, 3, 4696–4705 This journal is ª The Royal Society of Chemistry 2011
groups. Therefore mixed SAMs of 11-mercaptoundecanoic acid
(MUA) and mercaptododecane (MDD) were prepared by the
solution phase technique.1 Both molecules have similar chain
lengths to maximize the reciprocal influence of the chelating and
the non-coordinating abilities of carboxylic acid and methylene
groups, respectively on the mean surface activity and mixture. In
this study, the fine tuning of the chemical nature of surfaces was
performed by preparing mixed SAMs according to the replace-
ment method which consists in a stepwise deposition method
(Scheme 2).14,34,35 First, MDD-SAMs were prepared by the
immersion of gold substrates for 24 hours in an ethanolic solu-
tion of MDD molecules (10 mM). Second, the MDD molecules
were progressively exchanged by dipping the MDD-SAMs in
a solution of MUA molecules (10 mM) for times from 0.5 to
288 hours. SAMs were named as a function of the replacement
time, respectively SAM-0.5H, SAM-2H, SAM-4H, SAM-7H,
SAM-24H, SAM-96H, SAM-170H and SAM-288H. The modi-
fication of the surface chemistry was confirmed by contact angle
measurements with water (Table 1). The contact angle measured
for the SAM-MDD fits perfectly with a highly hydrophobic
surface.42 This value decreases with the replacement time to
a value which is similar to SAM-MUA43 and demonstrates the
gradual replacement of the methylene groups by the carboxylic
acid groups at the SAM surface.43 The use of methylene and
carboxylic acid groups as terminal functional groups induces
a large difference in the wetting properties of the corresponding
mixed SAMs. Nevertheless, contact angle measurements were
performed onto relatively large surfaces, and this technique is not
suited to distinguish the arrangement of the functional head
groups at the nanoscale. Therefore, iron oxide NPs have been
used as probing agents to study the formation of mixed SAMs
and the spatial arrangement of carboxylic acid groups at their
surface.
Stable suspensions of nanoparticles with narrow size distri-
bution and no aggregation have to be prepared with the aim to
correlate the evolution of the SAM structure with the changes of
the SPR signal and with the NP density. 12 nm sized iron oxide
nanoparticles which crystallize in the spinel structure (Fe3–dO4)
have been synthesized following the thermal decomposition
method.37,44,45 The coating of nanoparticles with oleic acid leads
to highly stable suspensions in many organic solvents and avoids
their agglomeration. The assembling process was performed by
dipping each mixed SAM for 10 minutes in the suspension of
nanoparticles in THF.16,41 Specific interactions between carbox-
ylic acid terminal groups at the SAM surfaces and the NP surface
occurred through a ligand exchange process. Then the SAMs
were extensively rinsed with THF to remove the physisorbed NPs
which may remain attached onto the first NPs layer.
The assembling of iron oxide NPs on each mixed SAMs has
been studied by the SPR technique. The increasing density in
NPs immobilized onto mixed SAMs was expected to strongly
influence the optical properties of the gold film. Therefore the
quantitative study of SPR changes is correlated with the amount
of non-plasmonic nanoparticles assembled on SAMs. The
reflectivity is measured as a function of the wavelength of the
light source (i.e. at a fixed angle of incidence), which is directly
Table 1 Contact angle values (degrees) measured with water on mixed SAMs as a function of the replacement time
SAM-MDD SAM-0.5H SAM-2H SAM-4H SAM-24H SAM-96H SAM-288H SAM-MUA
103 � 1.2 82 � 1.9 77 � 1.6 75 � 1.6 73 � 2.4 69 � 2.3 54 � 0.8 49 � 1.3
Scheme 2 Preparation of mixed SAMs by the replacement method and assembling of nanoparticles on the SAMs.
This journal is ª The Royal Society of Chemistry 2011 Nanoscale, 2011, 3, 4696–4705 | 4699
related to the modification of the metal surface. At the minimum
of reflectivity, the wavelength is directly related to the refractive
index and to the thickness of the adsorbed layer. On the basis of
these considerations, the high refractive index of iron oxide
nanoparticles (2.42 for magnetite)33 is expected to enhance the
shift of the wavelength.
Fig. 1 shows the evolution of the SPR by monitoring the
minimum of the reflectivity as a function of the wavelength. The
reflectivity curve with a minimum centred at the lowest SPR
wavelength (lAu ¼ 679 nm) represents the SPR peak of the gold
film without functionalization. The second curve displays
a minimum which is shifted to a higher wavelength lAu+MDD of
683 nm. This is consistent with the adsorption of MDD mole-
cules on the gold film.46 The next reflectivity curves represent the
peaks of surface plasmon resonance after each replaced SAM
was immersed in the NP suspension. All curves display
a minimum in reflectivity which is shifted to higher wavelengths
when the replacement time increases. Shifts increase from 19 to
181 nm with respect to the wavelength corresponding to the
SAM-MDD (Table 2). This observation demonstrates that the
SPR reflectivity is modified and depends on the density of NPs
immobilized on SAMs. Moreover, a pronounced and regular
increase of the minimum reflectivity was observed with the SAM
replacement time. This phenomenon has already been assigned
to the 2D assembling of gold NPs on a gold substrate by Tamada
et al.23,24 These experiments show the large tuning of the SPR
signal over a range of almost 200 nm. The spectral shift is far
larger in comparison to studies that relate on the adsorption of
molecules onto gold surfaces. We ascribe such a variation to the
high refractive index of magnetite, which agrees with the simple
immobilization of nanoparticles onto gold surfaces functional-
ized by mixed SAMs.
The wavelength shiftDl of the minimum in reflectivity for each
mixed SAM covered by NPs has been determined from the
minimum of the MDD-SAM which is uncovered by NPs
(Table 2). From these values, we calculated the effective thick-
ness d of the deposited layer of NPs according to the simple
model described by Campbell et al.:47
Dl ¼ mDn [1 � exp (�2d/ld)]
where Dl is the wavelength shift, m is the sensitivity of our gold
film to the local refractive index, Dn is the change in refractive
index induced by an adsorbate (Dn ¼ nadsorbate � nwater), d is the
effective adsorbate layer thickness and ld is the characteristic
evanescent electric field decay length. The values ofm and ld were
calculated by the Finite Difference Time Domain (FDTD).24,25
We found the following values for our gold film: m ¼ 3330 nm
per refractive index unit (RIU) and ld ¼ 253 nm. The change in
the refractive index Dn is determined from a reference, which is
the refractive index of the medium in which the experiment is
performed, i.e. water (n ¼ 1.33). Iron oxide magnetic nano-
particles are coated with oleic acid, which represent a thickness of
about 1.5 nm. Therefore, the refractive index of both compo-
nents (nFe3O4¼ 2.42, nOA ¼ 1.45) has to be taken into account to
calculate the refractive index of these hybrid nanoparticles. It is
reasonable to consider that the refractive indices of these hybrid
nanoparticles (nNP ¼ 2.12) result from the combination of the
refractive index of each component in the proportion of their
volume fraction, respectively about 0.7 and 0.3. The effective
thickness values d corresponding to each spectral shift increase
regularly to 8.9 nm with the replacement time (Table 2). It should
be noticed that after 288 hours of replacement time, this value
remains lower than the diameter of 12 nm sized NPs. Therefore,
the corresponding paving density was determined from these
values of d by taking into account the hexagonal compact
packing of 12 nm sized NPs with a spherical shape and coated by
oleic acid. We observe the paving density to increase regularly
from 8.4% on SAM-0.5H to 82.0% on SAM-288H. The latter is
in good agreement with earlier results41 showing that NPs cover
Fig. 1 SPR measurements on replaced SAMs after exposure to the NP suspension.
4700 | Nanoscale, 2011, 3, 4696–4705 This journal is ª The Royal Society of Chemistry 2011
89% of the surface of a MUA-SAM displaying only COOH
terminal groups at the surface.
To confirm the efficiency of SPR analysis as a quantitative
technique, NP assemblies onto replaced SAMs were also studied
by scanning electron microscopy (SEM) (Fig. 2). SAM-MDD
was used as a reference and as previously observed,41 no NPs
were immobilized after exposure to the NP suspension, only gold
grains were observed. In contrast, assemblies of NPs were clearly
observed on every replaced SAMs. The density in NPs clearly
increases as long as the initial MDD-SAM has been previously
immersed in the solution of MUA molecules. On the basis that
the carboxylic acid terminal groups at the SAM surface are
responsible for the immobilization of iron oxide nanoparticles,
we can conclude that the amount of COOH terminal groups
increases with the replacement times from 0.5 to 288 hours.
Moreover, the NP density remains unchanged after an increase
of the immersion time from 10 minutes to one hour. Taking into
consideration the rather fast kinetic of adsorption (below
10 minutes) of iron oxide NPs on a SAM fully decorated with
carboxylic acid group,41 the increasing density in nanoparticles is
thus correlated with the increasing amount of these groups on the
SAM surface.
The density in NPs immobilized on SAMs was calculated by
using the Visiolog� software (see the Experimental section). The
values are plotted as a function of the replacement time of MDD
Table 2 Effective thicknesses d and paving density of magnetic nano-particle layers calculated from the spectral shifts Dl of the minimum inreflectivity of SPR curves. Values are given as a function of the replace-ment time
Replacement time/hSpectralshift Dl/nm
Effectivethickness d/nm
Pavingdensity (%)
0.5 19 0.91 8.42 33 1.6 154 53 2.54 247 75 3.61 3324 99 4.79 4496 133 6.5 60170 170 8.34 77288 181 8.9 82
Fig. 2 SEM images of NPs assembled on replaced SAMs. (a) SAM-MDD, (b) SAM-0.5H, (c) SAM-2H, (d) SAM-4H, (e) SAM-7H, (f) SAM-24H, (g)
SAM-96H, (h) SAM-170H and (i) SAM-288H.
This journal is ª The Royal Society of Chemistry 2011 Nanoscale, 2011, 3, 4696–4705 | 4701
by MUA in Fig. 3a. The density in NPs continuously increases
with the replacement time and thus with the amount of chelating
carboxylic acid groups. Gold grains which are clearly observed
on SAM-MDD tend to be hidden as the density of NPs assem-
bled on mixed SAMs increases. More precisely, it increases
linearly and very fastly from 650 � 115 NP mm�2 on the SAM-
0.5H to 1611 � 79 NP mm�2 on the SAM-7H. Taking into
account the surface of 12 nm sized NPs shelled by a 1.5 nm thick
layer of oleic acid38 which corresponds to an occupied area of
176 nm2, these densities correspond to 13� 1.4% to 33� 1.6% of
the surface coverage, respectively. In contrast, the NP density on
SAM-24H to SAM-288H tends to level off from 3355 � 150 NP
mm�2 to 4023 � 101 NP mm�2. On SAM-288H, the surface
covered by NPs reaches a maximum of 80.8 � 2.0% which
remains slightly lower than in the case of NPs assembled onto
a SAM-MUA which is fully decorated with chelating carboxylic
acid terminal groups.41 It shows that the replacement of MDD
molecules by MUA molecules tends to reach the equilibrium
despite the excess of MUA molecules in the solution during the
preparation of mixed SAMs. In addition, if NPs would have been
assembled into the ideal 2D-hexagonal compact packing (hcp)
with a compacity of 0.9, NPs would cover almost 90%. This
difference is due to the random deposition of NPs which results
in some uncovered areas. Fig. 3b demonstrates that the density
values measured from SEM images are perfectly correlated with
the SPR shift (Dl) at the whole scale of each sample. Therefore,
the SPR technique is a very efficient way to quantify the amount
of NPs assembled on a gold surface at the sample scale.
SEM pictures also show that NPs initially assembled at the
gold grain junction. By increasing the replacement time, they
form domains which increase in size. The spatial distribution of
NPs assembled on SAM surfaces was evaluated by plotting the
pair correlation function (PCF) for each sample. This function
determines the average distance between each NP and its nearest
neighbours. It can be interpreted as an averaged probability of
finding the centre of a particle at a given distance, r, from the
centre of another particle normalized to the uniform probability
at large distances.48,49 Fig. 4 shows the PCFs for NPs assembled
on replaced SAMs. We note that for all SAMs, the PCFs are very
similar and reveal the typical features of NPs assembled with
short-range topological ordering. The g(r) function exhibits, in
all cases, a well pronounced primary maximum at the distance r
between the particles in the range between 12.0 and 13.6 nm.
These values are in good agreement with 12 nm sized NPs shelled
by a 1.5 nm thick layer of oleic acid which are tightly packed in
the assemblies. Moreover, the PCF is almost uniform for the
distribution of particles. Indeed, g(r) approaches unity for r > 25
nm which is significant of particles randomly located (jammed
disordered packing), they have essentially no structure. Never-
theless very weak peaks are observed at about 22 and 33 nm after
96 h of replacement time which may correspond to a short range
ordering of NPs. However, this accounts much more from the
high density in NPs on SAM-96H and SAM-288H than from
a real organization of NPs. These results indicate that the
structure of the NPs assemblies is not changing appreciably
although the chemical nature of the SAMs and the NPs density
change.
In a complementary study, the surface morphologies and
roughnesses of SAM-7H and SAM-96H after exposure to the NP
suspension were analyzed by AFM and compared to a SAM
before being exposed to the NP suspension (Fig. 5). The surface
coverage and the height of the covered areas were measured
assuming the length of both MDD and MUA to be identical.
NPs are clearly observed on both SAMs according to the cross-
section profiles which show average height differences of about
8.7 nm corresponding roughly to the NP size. These results are in
Fig. 3 (a) Wavelength shifts at the minimum of reflectivity (closed circles) measured by the SPR technique and NP density (opened circles) measured by
image analysis of SEM images on replaced SAMs after exposure to the NP suspension. (b) Wavelength shifts at the minimum of reflectivity plotted as
a function of the NP density.
Fig. 4 Pair correlation function of NPs assembled on replaced-SAMs.
4702 | Nanoscale, 2011, 3, 4696–4705 This journal is ª The Royal Society of Chemistry 2011
good agreement with SEM images and confirm the higher NP
density on SAM-96H than on SAM-7H. In addition, the root
mean roughness in an area of 5 � 5 mm2 was determined to be
twice higher on SAM-96H (2.4 nm) than on SAM-7H (1.2 nm).
The increase of this value is consistent with the increasing density
of NPs when compared to SAMs before exposition to NPs which
are rather flat with low roughness (0.8 nm). As reported in
previous studies,41 gold substrates have a similar roughness
(0.9 nm) due to their topography which consists in gold grains.
The magnetic properties of iron oxide NPs were also used to
determine the amount of NP immobilized on mixed SAMs. The
magnetization was measured as a function of an external
magnetic field at 5 K (Fig. 6). TheM(H) curves of NPs assembled
on SAM-7H and SAM-96H are very similar and show hystereses
which correspond to a ferrimagnetic behaviour. These results
agree well with such 12 nm sized Fe3�dO4 nanoparticles prepared
by the decomposition technique.36,38,41,50 Both M(H) curves
exhibit a saturation magnetization (Ms) which corresponds to the
alignment of all the magnetic moments of the NPs in the direc-
tion of the applied field. Ms is directly and proportionally
dependent on the quantity of NPs. After dipping in the NP
suspension, SAM-96H exhibits a higherMs value than SAM-7H,
which is in good agreement with the higher NPs density on the
former as measured by the SPR technique and SEM images
analysis.
Discussion
The immobilization of nanoparticles on mixed SAMs leads to
interesting information on the nanostructure of mixed SAMs
and on the mechanism of the replacement reaction of organo-
thiol molecules adsorbed on a gold surface. On the basis of our
previous work,16,41 we know that NPs assemble on specific areas
of mixed SAMs containing carboxylic acid terminal groups.
Therefore the amount and the spatial arrangement of NPs
assembled on mixed SAMs are directly related to the amount and
the spatial arrangement of carboxylic acid terminal groups at the
SAM surface, i.e. the SAM structure.
The NP density was measured for each mixed SAMs by both
SPR technique and SEM image analysis. It increases as a func-
tion of the replacement time on mixed SAMs from 650 �
115 NP mm�2 up to 4023 � 101 NP mm�2 which is perfectly
correlated with the shift of the minimum in reflectivity from
683 nm to 864 nm. Such a wide shift accounts for an overall
tuning of more than 180 nm with respect to the curve of the
MDD-SAM and shows the high sensitivity of the SPR technique.
The SPR technique is also very suitable to measure the average
thickness of the NP layer. It reaches a maximum of 8.9 nm on
SAM-288H which is lower than the NPs diameter of 12 nm.
These different values can be explained by the spherical shape of
NPs which induces an inhomogeneous thickness. SEM andAFM
also show that the NP layer on SAM-288H is not fully covered by
the NPs (80.8 � 2.0%).
The assembling of NPs on mixed SAMs prepared with
different replacement times enabled us to perform a time resolved
study. Indeed, we followed the evolution of the NP density as
a function of the replacement of MDD by MUA molecules.
Therefore we investigated the kinetic of the replacement reaction
Fig. 5 AFM images of NPs assembled on SAM-7H (a, b, c and d) and SAM-96H (e, f, g and h). Topographic height images (a and e), phase images (b
and f), 3D images (c and g) and cross-section profiles (d and h) of panels observed at the horizontal line in topographic height images.
Fig. 6 Magnetization recorded as a function of an applied magnetic field
recorded at 5 K for SAM-7H and SAM-96H.
This journal is ª The Royal Society of Chemistry 2011 Nanoscale, 2011, 3, 4696–4705 | 4703
which clearly occurs very fast within the first 24 hours, while it
slows down for longer times as reported previously for
organothiols.51–53
The spatial arrangement of NPs on mixed SAMs also gives
some interesting information on the replacement mechanism.
Indeed, NPs assemble initially and preferentially in small
domains which correspond to the gold grain junctions. By
increasing the replacement time, NPs were immobilized in
domains with sizes increasing gradually from about 20 nm2
(SAM-0.5H) to 200–300 nm2 (SAM-24H). It shows that areas
occupied by carboxylic acid groups are initially located at the
gold grain junctions and extend at the surface of the SAMs.
Therefore, the replacement unequivocally takes place domain
wise and not in a random fashion. At longer replacement times,
some areas which correspond to gold terraces still remain
uncovered of NPs on SAM-288H and agree with a slower kinetic
of the replacement reaction in these areas. In addition, NPs tend
to be assembled individually on these areas, which agree with the
formation of small domains decorated by carboxylic acid groups
in larger methylene rich domains. This may account from some
rearrangement of MDD and MUA molecules in the SAM.
The phase segregation in mixed SAMs prepared by the co-
adsorption method is controlled by the affinity of organothiols
with solvents and the gold surface.16 In the case of the replace-
ment method, it is mainly controlled by the structural non-
homogeneity of the initial SAMs. Indeed the compactness of the
SAMs controls the kinetic of the exchange reaction. The different
miscibility of MUA and MDD molecules with regard to inter-
molecular interactions such as hydrogen bonding between
carboxylic acid groups against van der Waals interactions
between alkylene chains has to be taken into account.39,40 In
a first step, the replacement occurs preferentially in the domain
boundaries or at defect sites. Indeed gold grain junctions contain
steps, kinks, and other defects in contrast to gold flat terraces.
The relief induced by gold grain junctions is expected to dis-
favour the average coordination of gold atoms and to favour the
desorption of Au–organothiol complexes.52 It also favours the
disordering of alkyl chains of MDD molecules which results in
weaker interactions and in a higher solubility. In a second step,
the replacement reaction becomes much slower because defect
sites and domain boundaries decrease while crystalline domains
of tightly packed molecules are favoured by flat and large gold
terraces. The replacement reaction only occurs at the interface
between MDD and MUA domains. In these areas, the disorga-
nization of MDD molecules is favoured by the growing domains
ofMUAwhose formation is ruled by hydrogen-bonding between
carboxylic acid terminal groups.54 These interactions are
strengthened between the new adsorbates but disturb the van der
Waals interactions between the MDD molecules. As far as the
surface occupied by non-polar regions decreases, such an inter-
face between domains reduces and the replacement reaction is far
slower and tends to the equilibrium. This phenomenon is
concomitant to the reduction of the domain size and to the fact
that NPs tend to become independent after 170H of replacement
time. These observations may correspond to the rearrangement
of MUA molecules with the remaining MDD molecules within
the SAM. Nevertheless, due to the tight packing of the MDD
molecule and its very limited amount in the SAM, it cannot result
from the desorption/readsorption process. A surface diffusion
limited mechanism is much more possible and results in the
spatial rearrangement of the molecules to minimize the system
energy and reach the thermodynamic equilibrium.10
Conclusions
The assembling of iron oxide nanoparticles has been demon-
strated to be highly efficient to investigate the structure of mixed
SAMs at the nanoscale. The preparation of mixed SAMs by the
replacement method offers interesting possibilities to study the
structure of SAMs as a time resolved study. We took advantage
of both the SPR of gold substrate and the high refractive index of
iron oxide nanoparticles to determine precisely the density in
NPs at the scale of the samples by a simple and alternative
technique such as the surface plasmon resonance (SPR) spec-
troscopy. We showed the high sensitivity of the SPR of the gold
surface to the deposition of iron oxide (Fe3O4–d) NPs on mixed
SAMs. The minimum in reflectivity of the SPR signal has been
shifted on a large spectral range which is correlated quantita-
tively with the increasing density in NPs. These results were
confirmed by the density values calculated by SEM image anal-
ysis and by the qualitative evolution of the saturation magneti-
zation value which also depends on the amount of NPs
assembled on the mixed SAMs. These results were directly
correlated with the amount of carboxylic acid terminal groups at
the SAM surface. SEM and AFM also provided some informa-
tion on the spatial arrangement of NPs as a function of the
replacement time and thus on the mechanism of the replacement
of MDD byMUAmolecules. The replacement reaction of MDD
by MUA on the gold surface occurs in two distinctive steps: (i)
very fastly at defect and default sites in the SAM which are
favoured at the gold grain junctions and then level off at flat gold
terraces and (ii) by the diffusion of MUA in MDD rich domains
which remains on gold terraces. Finally, mixed SAMs prepared
by the replacement method result in a different patterning of the
surface in comparison to the co-adsorption method. The fine
adjustment of the surface activity of replaced SAMs by
controlling the immersion times of the initial MDD-SAM in the
solution of MUA molecules should be of primary interest to
study the magnetic properties of NP assemblies.
Acknowledgements
Financial supports were provided by the agence national pour la
recherche (ANR) and direction g�en�erale de l’armement (DGA).
Notes and references
1 J. C. Love, L. A. Estroff, J. K. Kriebel, R. G. Nuzzo andG. M. Whitesides, Chem. Rev., 2005, 105, 1103.
2 R. L. Grimm, N. M. Barrentine, C. J. H. Knox and J. C. Hemminger,J. Phys. Chem. C, 2008, 112, 890.
3 R. K. Smith, P. A. Lewis and P. S. Weiss, Prog. Surf. Sci., 2004, 75, 1.4 C. D. Bain, J. Evall and G. M. Whitesides, J. Am. Chem. Soc., 1989,111, 7155–7164.
5 J. P. Folkers, P. E. Laibinis and G. M. Whitesides, Langmuir, 1992, 8,1330–1341.
6 C. D. Bain and G.M.Whitesides, J. Am. Chem. Soc., 1989, 111, 7164.7 S. Chen, L. Li, C. L. Boozer and S. Jiang, Langmuir, 2000, 16, 9287.8 S. J. Stranick, A. N. Parikh, Y. T. Tao, D. L. Allara and P. S.Weiss, J.Phys. Chem., 2002, 98, 7636–7646.
9 I. Choi, Y. Kim, S. K. Kang, J. Lee and J. Yi, Langmuir, 2006, 22,4885–4889.
4704 | Nanoscale, 2011, 3, 4696–4705 This journal is ª The Royal Society of Chemistry 2011
10 S. J. Stranick, A. N. Parikh, Y. T. Tao, D. L. Allara and P. S.Weiss, J.Phys. Chem., 1994, 98, 7636.
11 K. Tamada, M. Hara, H. Sasabe and W. Knoll, Langmuir, 1997, 13,1558.
12 P.-H. Lin and P. Guyot-Sionnest, Langmuir, 1999, 15, 6825.13 I. Takao, M. Wataru, T. Hiroshi, C. Nami, A. Uichi and
F. Masamichi, J. Vac. Sci. Technol. A, 2000, 18, 143714 T. Kakiuchi, K. Sato, M. Iida, D. Hobara, S.-i. Imabayashi and
K. Niki, Langmuir, 2000, 16, 7238.15 P. Gupta, A. Ulman, S. Fanfan, A. Korniakov and K. Loos, J. Am.
Chem. Soc., 2004, 127, 4.16 B. P. Pichon, M. Pauly, P. Marie, C. Leuvrey and S. Begin-Colin,
Langmuir, 2011, 27, 6235–6243.17 H. Raether, Surface Plasmons on Smooth and Rough Surfaces and on
Gratings, Heidelberg, 1998.18 E. Ostuni, L. Yan andG.M.Whitesides,Colloids Surf., B, 1999, 15, 3.19 L.Aihua,A.GopalandM.Chuanbin,Adv.Mater., 2009,21, 1001–1005.20 G. Barbillon, J. L. Bijeon, J. Plain and P. Royer, Thin Solid Films,
2009, 517, 2997.21 E. Golub, G. Pelossof, R. Freeman, H. Zhang and I. Willner, Anal.
Chem., 2009, 81, 9291.22 E. Hutter, J. H. Fendler and D. Roy, J. Phys. Chem. B, 2001, 105,
11159.23 M. Ito, F. Nakamura, A. Baba, K. Tamada, H. Ushijima,
K. H. A. Lau, A. Manna and W. Knoll, J. Phys. Chem. C, 2007,111, 11653.
24 X. Li, K. Tamada, A. Baba, W. Knoll andM.Hara, J. Phys. Chem. B,2006, 110, 15755.
25 M. Riskin, R. Tel-Vered, O. Lioubashevski and I. Willner, J. Am.Chem. Soc., 2009, 131, 7368.
26 Y. Yong, et al., Nanotechnology, 2006, 17, 2821.27 P. J. Cameron, X. Zhong and W. Knoll, J. Phys. Chem. C, 2007, 111,
10313.28 Y. Teramura, Y. Arima and H. Iwata,Anal. Biochem., 2006, 357, 208.29 J. Wang, A. Munir, Z. Zhu and H. S. Zhou, Anal. Chem., 2010, 82,
6782.30 Y. Sun, D. Song, Y. Bai, L.Wang, Y. Tian andH. Zhang,Anal. Chim.
Acta, 2008, 624, 294.31 K. S. Lee, M. Lee, K. M. Byun and I. S. Lee, J. Mater. Chem., 2011,
21, 5156.32 S. D. Soelberg, R. C. Stevens, A. P. Limaye and C. E. Furlong, Anal.
Chem., 2009, 81, 2357.33 D. Grigoriev, D. Gorin, G. B. Sukhorukov, A. Yashchenok,
E. Maltseva and H. M€ohwald, Langmuir, 2007, 23, 12388.
34 P.-H. Lin and P. Guyot-Sionnest, Langmuir, 1999, 15, 6825–6828.35 F. v. Wrochem, F. Scholz, A. Schreiber, H.-G. Nothofer, W. E. Ford,
P. Morf, T. Jung, A. Yasuda and J. M. Wessels, Langmuir, 2008, 24,6910.
36 A. Demorti�ere, B. Pichon, P. Panissod, B. Donnio, G. Pourroy,D. Guillon and S. B�egin-Colin, Nanoscale, 2011, 3, 225–232.
37 J. Park, K. An, Y. Hwang, J.-G. Park, H.-J. Noh, J.-Y. Kim,J.-H. Park, N.-M. Hwang and T. Hyeon, Nat. Mater., 2004, 3, 891.
38 M. Pauly, B. P. Pichon, A. Demorti�ere, J. Delahaye, C. Leuvrey,G. Pourroy and S. B�egin-Colin, Superlattices Microstruct., 2009, 46,195.
39 W. L. Wilson, P. F. Szajowski and L. E. Brus, Science, 1993, 262,1242–1244.
40 I. Horcas, R. Fernandez, J. M. Gomez-Rodriguez, J. Colchero,J. G�omez-Herrero and A. M. Baro, Rev. Sci. Instrum., 2007, 78,013705.
41 B. P. Pichon, A. Demortiere, M. Pauly, K. Mougin, A. Derory andS. Begin-Colin, J. Phys. Chem. C, 2010, 114, 9041.
42 P. E. Laibinis and G. M. Whitesides, J. Am. Chem. Soc., 1992, 114,1990–1995.
43 N. Faucheux, R. Schweiss, K. L€utzow, C. Werner and T. Groth,Biomaterials, 2004, 25, 2721.
44 A. Demortiere, P. Panissod, B. P. Pichon, G. Pourroy, D. Guillon,B. Donnio and S. Begin-Colin, Nanoscale, 2011, 3, 225.
45 B. P. Pichon, O. Gerber, C. Lefevre, I. Florea, S. Fleutot, W. Baaziz,M. Pauly, M. Ohlmann, C. Ulhaq, O. Ersen, V. Pierron-Bohnes,P. Panissod, M. Drillon and S. Begin-Colin, Chem. Mater., 2011,23, 2886–2900.
46 C. Altavilla, E. Ciliberto, A. Aiello, C. Sangregorio and D. Gatteschi,Chem. Mater., 2007, 19, 5980.
47 L. S. Jung, C. T. Campbell, T. M. Chinowsky, M. N. Mar andS. S. Yee, Langmuir, 1998, 14, 5636.
48 Z. Adamczyk, Irreversible Adsorption of Particles, in AdsorptionTheory, Modeling and Analysis, ed. J. Toth, Marcel-Dekker, NewYork, 2002.
49 B. Senger, J. C. Voegel and P. Schaaf, Colloids Surf., A, 2000, 165,255.
50 B. P. Pichon, P. Louet, O. Felix, M. Drillon, S. Begin-Colin andG. Decher, Chem. Mater., 2011, 23, 3668–3675.
51 C. E. D. Chidsey, C. R. Bertozzi, T. M. Putvinski and A. M. Mujsce,J. Am. Chem. Soc., 1990, 112, 4301.
52 D. M. Collard and M. A. Fox, Langmuir, 1991, 7, 1192.53 J. B. Schlenoff, M. Li and H. Ly, J. Am. Chem. Soc., 1995, 117, 12528.54 E. Cooper and G. J. Leggett, Langmuir, 1999, 15, 1024–1032.
This journal is ª The Royal Society of Chemistry 2011 Nanoscale, 2011, 3, 4696–4705 | 4705