Silanization of solid surfaces via mercaptopropylsilatrane: A new approach of constructing gold colloid monolayers

Article abstract of DOI:10.1039/c4ra05583g

Mercaptopropylsilatrane (MPS) was investigated as a novel self-assembled film on silica surfaces and also as a novel adhesive layer for the construction of a gold colloid monolayer on silica surfaces. We compare the preparation procedure and film quality of the MPS films to those of (3-mercaptopropyl)trimethoxysilane (MPTMS), which is commonly used for anchoring of gold nanoparticles on silica surfaces. The films were characterized by Fourier transform infrared spectroscopy, contact angle measurements, X-ray photoelectron spectroscopy, atomic force microscopy, and Ellman's reagent to determine surface mercaptan concentration. The process in preparing the MPS films involves more environmentally friendly aqueous or polar organic solvents, takes significantly shorter time (<30 min), and results in more uniform and reproducible films due to its insensitivity to moisture. The MPS films also have higher mercaptan surface density than that of the MPTMS films, resulting in higher saturation coverage of the gold colloid monolayers on the MPS-coated substrates. The higher ambient stability of the MPS films as compared to the MPTMS films is important for applications where sufficient durability of the self-assembled films under ambient conditions is required. Thus, mercaptosilatrane may have the potential to replace mercaptosilane for surface modification and as an adhesive layer for the construction of a noble metal colloid monolayer on oxide surfaces.

Full text of DOI:10.1039/c4ra05583g

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Silanization of solid surfaces via
mercaptopropylsilatrane: a new approach of
constructing gold colloid monolayers
Cite this: RSC Adv., 2014, 4, 46527
Wen-Hao Chen,a Yen-Ta Tseng,a Shuchen Hsieh,bc Wan-Chun Liu,a
Chiung-Wen Hsieh,b Chin-Wei Wu,a Chen-Han Huang,a Hsing-Ying Lin,a
a
b
a
*
Chao-Wen Chen, Pei-Ying Lin and Lai-Kwan Chau
Mercaptopropylsilatrane (MPS) was investigated as a novel self-assembled film on silica surfaces and also as
a novel adhesive layer for the construction of a gold colloid monolayer on silica surfaces. We compare the
preparation procedure and film quality of the MPS films to those of (3-mercaptopropyl)trimethoxysilane
(MPTMS), which is commonly used for anchoring of gold nanoparticles on silica surfaces. The films were
characterized by Fourier transform infrared spectroscopy, contact angle measurements, X-ray
photoelectron spectroscopy, atomic force microscopy, and Ellman's reagent to determine surface
mercaptan concentration. The process in preparing the MPS films involves more environmentally friendly
aqueous or polar organic solvents, takes significantly shorter time (<30 min), and results in more uniform
and reproducible films due to its insensitivity to moisture. The MPS films also have higher mercaptan
surface density than that of the MPTMS films, resulting in higher saturation coverage of the gold colloid
monolayers on the MPS-coated substrates. The higher ambient stability of the MPS films as compared to
the MPTMS films is important for applications where sufficient durability of the self-assembled films
under ambient conditions is required. Thus, mercaptosilatrane may have the potential to replace
mercaptosilane for surface modification and as an adhesive layer for the construction of a noble metal
colloid monolayer on oxide surfaces.
Received 11th June 2014
Accepted 19th September 2014
DOI: 10.1039/c4ra05583g
www.rsc.org/advances
strong intramolecular donor–acceptor interaction between
nitrogen and silicon atoms, silatranes are chemically more
stable to hydrolysis than trialkoxysilanes.14,15 As such, the low
Introduction
Monolayers exposing the mercapto group are interesting
because of their ability to coordinate to noble metals. For
example, (3-mercaptopropyl)trimethoxysilane (MPTMS) is oen
used to construct gold colloid monolayers on silica surfaces for
important applications in biosensing,1,2 surface-enhanced
Raman scattering,3 catalysis,4 and so on. Unfortunately,
MPTMS is very sensitive to moisture which leads to homoge-
neity and reproducibility problems.5–8 Although vapor deposi-
tion of MPTMS9–12 would be a good way to limit polymerization
processes and prevent the effects of moisture, the reaction time
is usually greater than that by liquid-phase silanization and the
surface coverage is not complete.12
rate of hydrolysis of the silatrane molecule prevents formation
of large polymer clusters on a surface.15 Hsieh and coworkers
were the rst to use aminopropylsilatrane as an adhesive lm
for anchoring gold nanoparticles (AuNPs) on silicon surfaces.16
However, the binding of amino group to gold is weaker than
that of mercapto group17,18 and thus imposes limitations on the
applications of xing noble metal nanoparticles on a substrate.
For example, AuNPs on amino surface can easily be displaced
when the AuNP surface is further modied with a self-
assembled monolayer (SAM), which is typically based the Au–
S bond.
We report here a new approach to perform silanization of
oxide surfaces via mercaptopropylsilatrane (MPS) to form a self-
assembled lm and subsequently the construction of gold
colloid monolayer on the self-assembled lm. To the best of our
knowledge, MPS has not been used to perform silanization of
oxide surfaces, nor has it been used as an adhesive lm for
anchoring noble metal nanoparticles on oxide surfaces. In this
work, we have explored the chemistry in detail by varying
solvent, concentration, and reaction time, characterized the
self-assembled lms using Fourier transform infrared (FTIR)
Recently, silatranes are receiving attention from the point of
view of their structural features and applications.13 Due to a
aDepartment of Chemistry and Biochemistry and Center for Nano Bio-Detection
(AIM-HI), National Chung Cheng University, Chiayi 62102, Taiwan. E-mail: chelkc@
ccu.edu.tw; Fax: +886 5 2721040; Tel: +886 5 2720411 ext. 66411
bDepartment of Chemistry and Center for Biomedical Sensor Research, National Sun
Yat-Sen University, Kaohsiung 804, Taiwan
cSchool of Pharmacy, College of Pharmacy, Kaohsiung Medical University, Kaohsiung
80708, Taiwan
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 46527–46535 | 46527

View Article Online
RSC Advances
Paper
spectroscopy, contact angle measurements, X-ray photoelectron was at about 520 nm and the peak absorbance was about 2.6.
spectroscopy (XPS), atomic force microscopy (AFM), and Ell- For preparation of gold colloid monolayers, the stock AuNP
man's reagent to determine surface mercaptan concentration, solution was diluted by 1% sodium citrate solution until the
and examined the ambient stability of the lms. Furthermore, absorbance was 2.0. As such, the concentration of the diluted
we have explored the use of MPS as an adhesive lm for AuNP solution is about 10 nM.20 Histograms of the images
anchoring gold nanoparticles on silica surfaces, and compare derived from transmission electron microscopy (TEM) revealed
its performance to that of MPTMS.
that the mean diameter of the AuNPs was 13 ꢂ 1 nm.
Experimental
Materials
Preparation of mercaptan-functionalized substrates
Glass slides with dimensions 5.5 cm ꢃ 0.8 cm were cleaned by
soap and then sonicated sequentially in methanol, acetone,
puried water, methanol, acetone, and puried water, with
sonication time of 10 min in each step. The glass slides were
dried in an oven at 70 ꢀC and then stored at room temperature.
Before use, the glass slides were treated by oxygen plasma (0.6
torr, Harrick Scientic Products, Inc.) for 20 min.
The MPTMS-modied substrates were prepared as follows. A
solution of 5% MPTMS in toluene (v/v, about 0.25 M) was
allowed to prehydrolysis for 12 h. Then the glass slides were
immersed into a 5% MPTMS solution for 12 h. Subsequently,
the modied glass slides were sonicated sequentially in meth-
anol–toluene (1 : 1, v/v), methanol, and puried water for 15
min in each step, and then dried under a nitrogen stream.
The MPS-modied substrates were prepared as follows. The
stock MPS solution was diluted to 1 : 1000 (v/v, about 0.46 mM)
in methanol and the glass slides were immersed into the solu-
tion for 30 min. Subsequently, the modied glass slides were
rinsed by methanol and puried water and then dried by
nitrogen stream.
The following chemicals, (3-mercaptopropyl)trimethoxysilane
(MPTMS, Sigma Aldrich), tetrachloroauric acid (Sigma Aldrich),
¨
triethanolamine (Riedel-de Haen), toluene (Macron), dichloro-
methane (Macron), sodium citrate (J.T. Baker), methanol (J.T.
Baker), pentane (J.T. Baker), potassium carbonate (Shimakyu),
sodium sulfate (Shimakyu), acetonitrile (Tedia), dimethyl sulf-
oxide (DMSO, Acros), and 5,50-dithio-bis(2-nitrobenzoic acid)
(DTNB, Aldrich) were used as received. All aqueous solutions
were prepared with water that had been puried by using a
Millipore Milli-Q water purication system (Millipore) with a
specic resistance of 18.2 MU cm.
Synthesis of MPS
MPS was prepared using a synthetic method reported earlier19
but with slight modications. A solution of MPTMS (4.693 g,
20.3 mmol) and triethanolamine (3.2 g, 21.4 mmol) in 30 mL
toluene was heated at 110 ꢀC in a round-bottomed ask and
then reuxed for about 30 h. The progress of the reaction was
followed by thin layer chromatography. Aer cooling to room
temperature, the resulting product was washed by cold n-
pentane and then dried by a rotary-vacuum evaporator. The
dried powder was re-dissolved in 1 : 1 dichloromethane/n-
pentane. Aer re-crystallization twice, a white solid 2.30 g MPS
(9.2 mmol) with a yield of 45.5% was obtained. The deposit was
re-dissolved in DMSO (20 mL) as a stock solution for storage.
Fourier transform infrared spectroscopy (FTIR) and
hydrogen nuclear magnetic resonance (1H NMR) were used to
Quantication of mercaptan surface density
Samples of MPS and MPTMS lms on glass slides were
immersed in a cuvette containing 1 mL of Ellman's reagent, (4
mM DTNB in 0.1 M pH 8.0 phosphate buffer). The reaction was
allowed to proceed for 12 h at room temperature and the
absorbance was measured at the wavelength of 412 nm.
Controlled samples were elaborated with non-modied glass
slides. The amount of mercaptan moieties was calculated from
the corresponding standard curve elaborated between 2 ꢃ 10ꢁ8
M and 5 ꢃ 10ꢁ6 M of mercaptoethanol.
1
examine the purity and structure. H 200 MHz NMR (CDCl3):
0.285 (SiCH2, 2H, t, J ¼ 16 Hz), 1.22 (HS, 1H, t, J ¼ 12 Hz), 1.55
(CH2, 2H, m, J ¼ 42.6 Hz), 2.64 (SCH2, 2H, t, J ¼ 16 Hz), 2.788
(OCH2, 6H, t, J ¼ 11.2 Hz), 3.647 (NCH2, 6H, t, J ¼ 11.2 Hz). FTIR
(KBr, cmꢁ1): 2928 and 2876 (vs., nCH2), 2545 (w, nSH), 1454 and
1277 (m, dCH2), 1124 and 1100 (vs., nCO), 939 (s, nCC), 909 (s, nCN),
776 (s, n(as)SiO), 616 (m, n(s)SiC), and 582 (m, nN/Si).14,19
Ambient stability test
The modication of the substrates by MPTMS (toluene 16 h) or
MPS (MeOH 30 min) was under dark and room temperature
Preparation of AuNPs
AuNPs were synthesized according to the following procedure. conditions according to the procedures as described above.
HAuCl4$3H2O (0.88 mM, 20 mL) was heated to boiling with Once a substrate had nished the modication procedures, the
vigorous stirring. A fresh 1% sodium citrate solution (2.4 mL) stability of the modied substrate is tested immediately under
was added in the boiling solution rapidly. The color of the room light and room temperature conditions for standing time
solution turned from yellow to colorless, then very dark purple, periods of 0, 1, 3, 24, and 48 h. Then the modied substrate was
then ruby-red. The resulting solution was kept boiling for 20 immersed in a diluted AuNP solution to anchor AnNPs on the
min, then cooled down to room temperature and stored at 4 ꢀC modied surface. The result of the stability test was observed by
as a stock solution. The stock AuNP solution was characterized measuring the UV-visible spectra of the AuNPs anchored on the
by a UV-visible spectrometer. Typically, the peak wavelength modied substrate.
46528 | RSC Adv., 2014, 4, 46527–46535
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
Instrumentation
UV-visible absorption spectra were obtained with full spectrum
from 400 to 800 nm by a Jasco V-570 spectrometer. TEM images
were recorded using a JEOL 2010 HRTEM operating at an
accelerating voltage of 200 kV. Scanning electron microscopy
(SEM) images for determination of the coverage of the gold
colloid monolayers were obtained by a Hitachi 4800I eld-
emission scanning electron microscope (FE-SEM) operated at
15 kV. The surface morphology of the lms was characterized
using an atomic force microscope (MFP-3D™, Asylum
Research, Santa Barbara, CA) under ambient conditions. For
AFM experiments, a silicon cantilever (Olympus AC240TS) with
a normal spring constant of 2 N mꢁ1 (thermal method)
(Proksch, Schaffer, Cleveland, Callahan, & Viani, 2004) was
used. All AFM images were acquired in AC (tapping) mode with
an image resolution of 512 ꢃ 512 pixels.
Fourier transform infrared (FTIR) spectroscopy in the
transmission mode was used to characterize the bulk MPTMS
and MPS samples and the molecular packing of the MPTMS and
MPS lms. To collect IR spectra of bulk samples, MPTMS or
MPS was casted onto a KBr pellet and FTIR spectra was recorded
on a Bruker Alpha FTIR spectrometer (4 cmꢁ1 resolution, 2048
scans) equipped with a deuterated triglycine sulfate (DTGS)
detector. To collect IR spectra of thin lms, a MPTMS lm or a
MPS lm was prepared on a silicon wafer and FTIR spectra was
recorded on a Bruker 66 v/s FTIR spectrometer equipped with a
liquid nitrogen-cooled mercury–cadmium–telluride (MCT)
detector (8 cmꢁ1 resolution, 256 scans, sample compartment
vacuum pressure was 0.12 hPa).
Equilibrium water contact angles were measured on modi-
ed silicon wafer surfaces and on a bare silicon wafer substrate
using a CAM-M200 contact angle meter (Creating Nano Tech-
nologies, Inc., Taiwan) at room temperature. All the contact
angles measured in this study were undertaken with 10 mL
volume at different locations on each sample surface.
XPS was measured with a Kratos Axis Ultra DLD spectrom-
eter using a Mg/Al achromatic source 450 W. Each sample was
analyzed at photoelectron take-off angle of 45ꢀ. The take-off
angle is dened as the angle between the analyzer and the
substrate surface. The silica substrate was used an internal
standard for XPS positions and a binding energy of 99.15 eV was
used for elemental silicon.21 The peak deconvolution was per-
formed aer baseline subtraction by assuming that the
constituent peaks have Gaussian shape.
Fig. 1 Schematic illustration showing the procedures for surface
functionalization by MPTMS or MPS and anchoring of AuNPs on those
functionalized substrates.
Literature review shows that the MPTMS system usually uses
a concentration of 0.1% v/v–5% v/v and takes 0.5 h–24 h in
anhydrous organic solvent, such as toluene and benzene, to
form a well-ordered lm.5,6,21 However, MPTMS is very sensitive
to moisture which leads to agglomeration and polymerization
on a substrate during deposition, generally resulting in multi-
layer deposition and irregular surface morphology.5–8 On the
other hand, silatranes are relatively stable to moisture13,16 and
thus may be applicable in more polar solvents. In this aspect,
using the capability of the lms to anchor AuNPs as an indicator
to monitor the lm formation process, we found that lms
formed from MPS dissolved in nonpolar solvents (e.g., toluene)
or polar solvent (e.g., methanol) have similar activity and
kinetics (data not shown). Surprisingly, an aliquot of stock MPS
solution diluted by water also show good activity towards
anchoring of AuNPs. The lms were then characterized using
FTIR spectroscopy, contact angle measurements, XPS, AFM,
and Ellman's reagent.
Infrared spectroscopy
FTIR spectroscopy was performed to conrm the presence of
the MPS and MPTMS lms on silicon wafer samples and
evaluate their structures. Fig. 2 shows the spectra in the
region between 700 and 3300 cmꢁ1 that contain the vibra-
tional modes of CH2 stretching, CH3 stretching, SH stretch-
ing, and Si–O–Si stretching. Fig. 2a illustrates the IR
Results and discussion
The mercapto group is known to have specic interaction with spectrum of bulk MPS with peaks of 2928, 2876 cmꢁ1 for
many metals, including gold, silver, and copper. For this asymmetric and symmetric CH2 stretching, and 2545 cmꢁ1
purpose, MPTMS is one of the popular precursors to prepare a for SH stretching, respectively. A characteristic peak is also
mercaptan-functionalized surface for anchoring of noble metal found at 582 cmꢁ1 (data not shown), which has been assigned
nanoparticles. Here, we explore the use of MPS to prepare to the vibration of the coordination bond (N / Si) in sila-
mercaptan-functionalized surfaces and compare its perfor- tranes.14,19 For the IR spectrum of the MPS lm shown in
mance to that of MPTMS. The procedures used to prepare MPS- Fig. 2a, we have observed CH2 stretching modes at 2927 and
or MPTMS-modied surfaces are illustrated schematically in 2860 cmꢁ1, and SH stretching mode at 2545 cmꢁ1. These peak
the rst step shown in Fig. 1.
positions and assignments are in agreement with
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 46527–46535 | 46529

View Article Online
RSC Advances
Paper
1400ꢁ400 cmꢁ1 range reect the structural orderedness of
the silica skeleton.30,31 By considering long-range interaction
provided by the Coulomb elds in the silica skeleton,
Galeener and Lucovsky have accounted for three pairs of
longitudinal optical (LO) and transverse optical (TO) vibra-
tional modes in the vibrational spectra of vitreous silica.30 As
shown in Fig. 2c, the IR spectra in the 1400ꢁ700 cmꢁ1 range
show that the MPS lm has two peaks at 1136 and 1073 cmꢁ1
,
which are respectively assigned to the LO and TO modes of
the Si–O–Si asymmetric stretching vibration, while the
MPTMS lm has only one peak at 1037 cmꢁ1. The LO–TO
splitting exhibited in the IR spectrum of the MPS lm suggests
that the silica skeleton in the MPS lm is sufficiently in order.
Contact angle measurements
To further characterize the lms, water contact angle
measurements were performed on the surfaces of the MPS- and
MPTMS-modied silicon wafer samples. The contact angle for
ethanol cleaned silicon wafers was 24.6ꢀ ꢂ 2.3ꢀ, which is
indicative of a hydrophilic surface. The contact angle for the
MPS and MPTMS lms was much higher, at 57.3ꢀ ꢂ 1.1ꢀ and,
57.3ꢀ ꢂ 4.2ꢀ, respectively. Such values are consistent with values
reported in literature21,24,32,33 and further conrm the presence
of the respective lms on the silicon surfaces. The better
reproducibility of the contact angle for the MPS lm also
suggests that the MPS lm is more uniform than the MPTMS
lm.
Fig. 2 FTIR spectra in the 3300ꢁ2200 cmꢁ1 range for (a) the MPS film
and bulk and (b) the MPTMS film and bulk; (c) FTIR spectra in the
1400ꢁ700 cmꢁ1 range for the MPS and MPTMS films.
X-ray photoelectron spectroscopy
Fig. 3a shows representative XPS spectra of a MPS lm (line I)
and a MPTMS lm (line II) on glass substrate. Both spectra are
very similar and conrm the presence of four elements (Si, C, O
and S) in the lms. Such a spectral similarity suggests that the
atomic composition of the end product of both lms is similar.
The most signicant feature in the XPS spectrum of the MPS
lm is the absence of N signal (ꢄ400 eV), which agrees with a
proposed mechanism of partial hydrolysis of the silatrane fol-
lowed by the reaction with a silica surface15 and our proposed
mechanism as shown in Fig. 4. The absence of the N peak also
conrms that, aer the hydrolysis, no triethanolamine was le
as a byproduct on the substrate.
Fig. 3b shows XPS C(1s) spectra of the MPS lm (line I) and
the MPTMS lm (line II) on glass substrate. The C(1s) XPS
spectrum of bare silica substrate shows only one band at
284.7 eV, whereas those of the MPS and MPTMS lms on silica
substrate show two contribution bands at 284.7 eV and
286.3 eV. The band at 284.7 eV corresponds to hydrocarbon and
carbon bonded to silicon and the band at 286.3 eV corresponds
to carbon bonded to oxygen and sulfur.5 The appearance of a
new band at 286.3 eV demonstrates indeed both MPS and
MPTMS react with the SiO2 surface.
literature.19,22 Comparatively, the IR spectrum of bulk
MPTMS shown in Fig. 2b indicates peaks at 2942 cmꢁ1 for
asymmetric CH3 stretching and 2841 cmꢁ1 for symmetric
CH3 stretching of the methoxyl groups, and 2565 cmꢁ1 for SH
stretching.23–25 The IR spectrum of the MPTMS lm shown in
Fig. 2b indicates CH2 stretching modes at 2927 and 2870
cmꢁ1, and SH stretching mode at 2545 cmꢁ1. The disap-
pearance of the CH3 stretching modes suggests that hydrolysis
and condensation occurred within the MPTMS lm.
The positions of the peak frequencies for CH2 stretching,
SH stretching, and Si–O–Si stretching modes provide insight
into the intermolecular environment in these lms.26,27 First,
the relatively high CH2 vibrational frequencies suggest that
the propyl chains in both the MPS and MPTMS lms are more
liquid-like. On the other hand, the lower CH2 symmetric
stretching frequency for the MPS lm at 2860 cmꢁ1 as
compared to that of MPTMS lm at 2870 cmꢁ1 suggests that
the self-assembled MPS molecules have a better packing in
the lm. Second, it has been reported that the SH stretching
mode free from hydrogen bonding is at 2565 cmꢁ1 while SH
involved in hydrogen bonding has the SH stretching mode at
lower frequency.25,28,29 As such, it suggests that the SH group
in bulk MPTMS is free from hydrogen bonding while the SH
group in the MPTMS lm, bulk MPS, and MPS lm probably
has signicant SH/S interaction.28,29 Third, peak positions
and widths of the Si–O–Si vibrational modes in the
Fig. 3c shows XPS S(2p) spectra of the MPS lm (line I) and
the MPTMS lm (line II) on glass substrate. The S(2p) core-
level band of both MPTMS and MPS lms can be decom-
posed into three peaks at binding energies of 162.1, 165.4,
169.5 eV and 162.1, 165.5, 169.6 eV, respectively. Comparison
46530 | RSC Adv., 2014, 4, 46527–46535
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
Fig. 4 Schematic representation of the mechanism proposed for the
reaction between MPS and surface silanol.
disulde (–SS–) species.32,34 XPS alone cannot distinguish
between the different kinds of reduced sulfur, mercaptan, or
disulde.34 The other peak at 169.5 eV or 169.6 eV is consistent
with oxidized sulfur (–SOxH).32,35,36 The area ratios of the
oxidized sulfur band with respect to the total sulfur curve for the
MPS lm and the MPTMS lm are 56.1% and 74.2%, respec-
tively, suggesting that the mercapto group in the MPS lm is
less susceptible to oxidation. It should be noted that oxidation
may take place during storage of the sample until XPS
measurement37 and X-rays can have a damaging effect on the
surface structures.38 Nevertheless, considering that both lms
were prepared and analyzed at the same time, these two factors
should affect both lms to the same extent.
Quantication of mercaptan surface density
Surface density of the mercapto group is of great concern,
because it determines the density of self-assembled mole-
cules on top of the lms and hence the reactivity of the lms
toward anchoring of AuNPs. The Ellman's reaction was rst
developed for the determination of mercapto group in bio-
logical materials39 and later on also applied to the determi-
nation of surface mercapto group.40 In this study,
quantication of the number of accessible mercapto group of
the MPS and MPTMS lms was performed by using the Ell-
man's reagent, 5,50-dithio-bis(2-nitrobenzoic acid) (DTNB).
The optimum reaction time between the lms and DTNB was
investigated by monitoring the absorbance at 412 nm over a
time period of 2 to 12 h. Beyond 10 h of reaction time, the
absorbance value level off and hence a reaction time of 12 h
was used for subsequent analyses. The mean mercaptan
surface density of the MPS and MPTMS lms was 5.4 ꢂ 0.4
nm2 and 4.1 ꢂ 0.4 nm2, respectively. The mercaptan surface
density of the MPS lm is consistent with the average number
of hydroxyl group of ꢄ5 nmꢁ2 typically found on silica
surfaces41,42 and is also reasonable as compared to a total of
about 8 surface Si atom per nm2.43 Furthermore, the
Fig. 3 (a) The full XPS spectra, (b) the C(1s) XPS spectra, and (c) the
S(2p) XPS spectra of a MPS film (line I) and a MPTMS film (line II). Curve-
fitting of the C(1s) XPS spectra and the S(2p) XPS spectra are also
shown.
of the area integration of the sulfur peaks for MPS : MPTMS
is 1.42 : 1, suggesting that the MPS lm has a higher surface
sulfur density than that of the MPTMS lm. This is consistent
with the FTIR results where the MPS modied substrate
shows a larger SH stretching peak at 2545 cmꢁ1. The S(2p3/2,1/2
)
doublet feature of both lms at 162.1/165.4 eV and 162.1/
165.5 eV are attributed to either reduced sulfur (–SH) or to
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 46527–46535 | 46531

View Article Online
RSC Advances
Paper
estimated mercaptan surface density agrees with the
approximate area (0.20–0.21 nm2) projected by each alkyl
chain in the plane of close-packed alkylsiloxane mono-
layers.44,45 In addition, the higher mercaptan surface density
of the MPS lm as compared to that of the MPTMS lm is
consistent with the XPS and FTIR results, suggesting that the
self-assembled MPS molecules have a better packing density
in the lm and/or the mercapto group in the MPS lm is less
susceptible to oxidation.
Kinetics of lm formation
To investigate the formation kinetics of both the MPS and
MPTMS lms, the high-extinction surface plasmon resonance
(SPR) of gold colloid allows us to monitor the lm formation
process by following the UV-vis spectra of the anchored particles
with increasing lm formation time. In this study, the absor-
bance of two sets of glass slides was measured aer the glass
slides in each set were immersed separately into a MPS solution
or a MPTMS solution for various time periods. The MPTMS- and
MPS-coated glass slides were then immersed in a 10 nM AuNP
solution for 30 min (see below). Fig. 6 shows the trends of
absorbance at 530 nm versus immersion time for the two set of
glass slides. As shown in previous studies46,47 and in this study
(curve a of Fig. 6), MPTMS needs a long time (>8 h) to form a
complete layer with optimum mercaptan surface density. In
contrast, as shown in curve b of Fig. 6, MPS requires just 25 min
to form a complete layer with optimum mercaptan surface
density. The faster formation rate of the MPS lm is probably
due to our proposed mechanism as shown in Fig. 4. First, there
is likely an attractive interaction between the electron lone pair
on the N atom of the silatrane and the silanol group on the silica
surface so that silatrane molecules reach the silica surface
much faster. Second, ab initio calculations show that the
transannular Si–N interaction leads to a larger proton affinity of
the oxygen atoms in silatranes than that in trimethoxysilanes
and thus facilitates the condensation reaction at the silica
surface.48
Atomic force microscopy
Fig. 5a and b shows a typical series of AFM 2D and 3D images
displaying the surface of a MPTMS-modied silicon substrate.
The size of the AFM scans is 1 ꢃ 1 mm2. The images show the
surface morphology of the MPTMS lm, a few of small protru-
sions appear on the surface. The bright spots in the image
indicate where aggregates have formed. Such an observation is
consistent with previous reports and is attributed to the
competition between self-polymerization and condensation
reaction with surface silanol.5,8 The surface roughness is 137 pm
as calculated from Fig. 5a. Fig. 5c and d show a typical series of
AFM 2D and 3D images displaying the surface of a MPS-
modied silicon substrate. The surface morphology of the
MPS-modied silicon substrate shows that the lm is homo-
geneous and smooth. The calculated surface roughness is 110
pm. The AFM results show that the MPS lm is smoother than
the MPTMS lm. This is consistent with the better reproduc-
ibility of contact angle measurement for the MPS modied
surface. The higher roughness of the MPTMS lm has been
attributed to unhydrolyzed Si–OCH3 bonds which make lateral
cross-linking between neighbouring adsorbed MPTMS mole-
cules more difficult.8 Incomplete lateral cross-linking may
result in a lower surface coverage and such a proposition is
consistent with the results of surface mercaptan concentration
determination.
Anchoring gold nanoparticles
Many processes are involved in the formation of a gold colloid
lm on mercaptan-functionalized surfaces. Natan and
coworkers found that gold colloid monolayer coverage is
limited at early times by diffusion and at later times by inter-
particle repulsion.46 When the MPTMS- and MPS-coated glass
slides were immersed in a solution of 13 nm diameter AuNPs
Fig. 5 AFM 2D and 3D images. (a) 3D image and (b) 2D image of Fig. 6 Absorbance at 530 nm versus immersion time for two sets of
MPTMS modified silicon surface (0.25 M in toluene, 12 h). (c) 3D image glass slides with each set of glass slides immersed separately into a
and (d) 2D image of MPS modified silicon surface (0.46 mM in meth- solution of (-) MPS (0.46 mM in MeOH) or (B) MPTMS (0.25 M in
anol, 30 min).
toluene) and then immersed in a 10 nM AuNP solution for 30 min.
46532 | RSC Adv., 2014, 4, 46527–46535
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
(AuNP concentration ¼ 10 nM, absorbance ¼ ꢄ2.0 A.U.) with
various immersion times, we found that both the MPTMS- and
MPS-modied substrates began to reach the saturation
coverage at about 30 min (data not shown). Hence, we used a
reaction time of 30 min for preparation of gold colloid lms.
Fig. 7 shows the typical UV-vis spectra and FE-SEM images
of the gold colloid monolayer on a MPTMS-coated substrate
and a MPS-coated substrate prepared at optimized condi-
tions. As can be seen in the FE-SEM images, the mean
diameter of the AuNPs is still about 13 nm and particle
aggregation is almost not present in carefully prepared
surfaces. Similar conclusions are also obtained by AFM (data
not shown). It should be noted that the saturation coverage of
the gold colloid monolayer on the MPS-coated substrates
(26.5%) is higher that on the MPTMS-coated substrates
(22.6%), while the full-width half-maximum of the SPR band
is similar (82.1 nm versus 93.2 nm). Since gold colloids
possess mobility and might aggregate to larger colloids, UV-
vis spectra and FE-SEM images of the gold colloid monolayer
aer longer period of storage time were traced. In a period of
about 2 months, no signicant difference in UV-vis spectra
and FE-SEM images were observed. As AuNPs are important
sensor materials today, reproducibility of the sensor fabri-
cation process is important in order to meet analytical
performance standards. In this regard, we explore the
reproducibility of using both MPTMS and MPS as linkers to
prepare gold colloid monolayers on glass slides. Results show
that the average coefficients of variation (CV) of the peak
absorbance for batch-to-batch runs (n ¼ 7) with MPTMS and
MPS are 2.3% and 1.6%, respectively, while the coefficients of
variation (CV) of the peak absorbance for day-to-day runs (n ¼
7) with MPTMS and MPS are 5.7% and 3.4%, respectively.
Although the reproducibility of using MPTMS as a linker is
acceptable, the reproducibility of using MPS as a linker is
even better. Together with the shorter preparation time
required, we believe MPS will benet the development of
sensors based on gold colloid technology.
Ambient stability
The ultimate utility of mercaptan-functionalized lms will be
critically dependent on their stability under ambient condi-
tions. Organothiol monolayers on Au exposed to air for pro-
longed periods has been shown to be oxidized to sulnates and
sulfonates, indicating their nite stability.37,49–51 To explore the
stability of MPTMS and MPS lms under ambient laboratory
conditions, the MPTMS- and MPS-coated glass slides were
exposed in ambient air under room light at various standing
times and then immersed in a solution of AuNPs for 30 min.
Then we dene the lms with negligible exposure to ambient air
under room light (i.e., zero standing time) have 100% activity. In
other words, activity is dened as Ai/A0 where A0 is the absor-
bance at zero standing time, Ai is the absorbance at a nite
standing time, and the absorbance is measured at 530 nm.
As shown in Fig. 8, the activity of the MPTMS lms at a
standing time of 1 h remains 20%, while the activity at a
standing time of 25 h remains just 10%. On the contrary, there
is not observable loss in activity for the MPS lms at a standing
time of 1 h. At a standing time of 25 h, the activity of the MPS
lms remains 82%, while the activity at a standing time of 48 h
still remains more than 75%. This is not a trivial result as the
end products of both lms are expected to be similar. Senkevich
and coworkers have demonstrated that the mercapto group in
the starting MPTMS molecule is stable in liquid form over a
period of 8 months, suggesting that the oxidized mercaptan
molecular layer is not due to the oxidation of the starting
molecule but rather due to the chemistry residing on the
surface.34 It is well-known that the hydrolysis of the MPTMS
molecule in the lm form is oen incomplete and thus a
MPTMS lm oen retains its methoxy groups.21 As the pKa of
propyl mercaptan (pKa ¼ 10.6)52 is lower than that of water (pKa
¼ 15.7), we propose a reaction mechanism as shown in Fig. 9 to
account for the lower ambient stability of the MPTMS lm.
Since the mercapto group is more acidic than water molecules
from moisture, it will accelerate the hydrolysis of a neighbour-
ing methoxy group in a MPTMS lm and form a thiolate itself.
Fig. 7 Absorbance spectra and FE-SEM images of AuNPs on a MPS- Fig. 8 Activity of the (-) MPS and (B) MPTMS films after exposure in
coated glass slide (line I) and a MPTMS-coated glass slide (line II).
ambient air under room light at various standing times (n ¼ 3).
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 46527–46535 | 46533

View Article Online
RSC Advances
Paper
Acknowledgements
This work was supported by Ministry of Science and Technology
(Taiwan) through grant number NSC 102-2113-M-194-003-MY3.
References
1 From Bioimaging to Biosensors - Noble Metal Nanoparticles in
Biodetection, ed. L.-K. Chau and H.-T. Chang, Pan Stanford
Publishing, Singapore, 2013.
2 B. Sepulveda, P. C. Angelome, L. M. Lechuga and L. M. Liz-
Marzan, Nano Today, 2009, 4, 244–251.
3 R. G. Freeman, K. C. Grabar, K. J. Allison, R. M. Bright,
J. A. Davis, A. P. Guthrie, M. B. Hommer, M. A. Jackson,
P. C. Smith, D. G. Walter and M. J. Natan, Science, 1995,
267, 1629–1632.
4 D. T. Thompson, Nano Today, 2007, 2, 40–43.
5 M. Hu, S. Noda, T. Okubo, Y. Yamaguchi and H. Komiyama,
Appl. Surf. Sci., 2001, 181, 307–316.
6 C. M. Halliwell and A. E. G. Cass, Anal. Chem., 2001, 73, 2476–
2483.
7 A. V. Krasnoslobodtsev and S. N. Smirnov, Langmuir, 2002,
18, 3181–3184.
Fig. 9 Schematic representation of the mechanism proposed for the
oxidation of the mercapto group in MPTMS films.
8 S.-R. Yang and B. O. Kolbesen, Langmuir, 2008, 255, 1726–
1735.
9 E. Pavlovic, A. Quist, U. Gelius and S. Oscarsson, J. Colloid
Interface Sci., 2002, 254, 200–203.
10 P. Doppelt, N. Semaltianos, C. D. Cavellin, J. L. Pastol and
D. Ballutaud, Microelectron. Eng., 2004, 76, 113–118.
11 E. Finocchio, E. Macis, R. Raiteri and G. Busca, Langmuir,
2007, 23, 2505–2509.
The thiolate thus formed is susceptible to oxidation,53 for
example, reacts with oxygen to form sulfenic acid. Sulfenic acid
may then be further oxidized to sulnic acid and sulfonic acid.
On the contrary, the MPS lm is relatively stable because of the
complete hydrolysis of the Si–O–C bonds in the silatrane, as
suggested by the XPS results.
12 B. Pattier, J.-F. Bardeau, M. Edely, A. Gibaud and N. Delorme,
Langmuir, 2008, 24, 821–825.
13 J. K. Puri, R. Singh and V. K. Chahal, Chem. Soc. Rev., 2011,
40, 1791–1840.
Conclusions
In this study, we have shown that MPS can be used to form a
self-assembled lm on silica surfaces and also as an effective 14 M. G. Voronkov, Pure Appl. Chem., 1966, 13, 35–60.
adhesive layer for the construction of a gold colloid monolayer 15 L. S. Shlyakhtenko, A. A. Gall, A. Filonov, Z. Cerovac,
on silica surfaces. As silatranes are relatively stable to moisture,
MPS is applicable in more environmentally friendly aqueous
A. Lushnikov and Y. L. Lyubchenko, Ultramicroscopy, 2003,
97, 279–287.
and polar organic solvents while MPTMS usually works better in 16 S. Hsieh, W.-J. Chao and C.-W. Hsieh, J. Nanosci.
anhydrous organic solvents. In comparison with MPTMS, the Nanotechnol., 2009, 9, 2894–2901.
MPS lms are smoother and more uniform, and are completely 17 C. A. Martin, D. Ding, H. S. van der Zant and J. M. van
free of molecular aggregates. MPS also requires a signicantly Ruitenbeek, New J. Phys., 2008, 10, 065008.
shorter time than MPTMS to form a complete layer and exhibits 18 D. V. Leff, L. Brandt and J. R. Heath, Langmuir, 1996, 12,
a higher mercaptan surface density in the lm. Such a higher 4723–4730.
mercaptan surface density in the MPS lm facilitates the 19 R. Spennato, M.-J. Menu, M. Dartiguenave and
anchoring of gold nanoparticles on the lm. As a result, the Y. Dartiguenave, Transition Met. Chem., 2004, 29, 830–839.
saturation coverage of the gold colloid monolayer on the MPS- 20 X. Liu, M. Atwater, J. Wang and Q. Huo, Colloids Surf., B,
coated substrates is higher that on the MPTMS-coated 2007, 58, 3–7.
substrates. Surprisingly, the MPS lms also exhibit a higher 21 J. J. Senkevich, G.-R. Yang and T.-M. Lu, Colloids Surf., A,
ambient stability as compared to the MPTMS lms. This char- 2002, 207, 139–145.
acteristic is benecial to many applications where sufficient 22 Y. Yu, J. Addai-Mensah and D. Losic, J. Nanosci. Nanotechnol.,
durability of the self-assembled lms under ambient conditions 2011, 11, 1–8.
is important. Thus, mercaptosilatrane as an efficient and 23 Y.-S. Li, Y. Wang, T. Tran and A. Perkins, Spectrochim. Acta,
environmental-amiable molecular linker may have a potential
Part A, 2005, 61, 3032–3037.
to replace mercaptosilane for surface modication.
24 D. G. Kurth and T. Bein, Langmuir, 1993, 9, 2965–2973.
46534 | RSC Adv., 2014, 4, 46527–46535
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
25 M. Hayashi, Y. Shiro and H. Murata, Bull. Chem. Soc. Jpn., 40 J. D. Bass and A. Katz, Chem. Mater., 2006, 18, 1611–1620.
1966, 39, 112–117. 41 L. T. Zhuravlev, Langmuir, 1987, 3, 316–318.
26 R. G. Snyder, H. L. Strauss and C. Elliger, J. Phys. Chem., 42 A. Tuel, H. Hommel and A. P. Legrand, Langmuir, 1990, 6,
1982, 86, 5145–5150. 770–775.
27 M. D. Porter, T. B. Bright, D. L. Allara and C. E. D. Chidsey, J. 43 M. H. Du, A. Kolchin and H. P. Cheng, J. Chem. Phys., 2004,
Am. Chem. Soc., 1987, 109, 3559–3568. 120, 1044–1054.
28 B. A. Kolesov, V. S. Minkov, E. V. Boldyreva and 44 S. R. Wasserman, G. M. Whitesides, I. M. Tidswell,
T. N. Drebushchak, J. Phys. Chem. B, 2008, 112, 12827–12839.
29 V. S. Minkov, S. V. Goryainov, E. V. Boldyreva and
C. H. Gorbitz, J. Raman Spectrosc., 2010, 41, 1748–1758.
30 F. L. Galeener and G. Lucovsky, Phys. Rev. Lett., 1976, 37,
1474–1478.
B. M. Ocko, P. S. Pershan and J. D. Axe, J. Am. Chem. Soc.,
1989, 111, 5852–5861.
45 I. M. Tidswell, T. A. Rabedeau, P. S. Pershan, S. D. Kosowsky,
J. P. Folkers and G. M. Whitesides, J. Chem. Phys., 1991, 95,
2854–2861.
31 R. B. Laughlin and J. D. Joannopoulos, Phys. Rev. B: Solid 46 K. G. Grabar, P. C. Smith, M. D. Musick, J. A. Davis,
State, 1977, 16, 2942–2952.
D. G. Walter, M. A. Jackson, A. P. Guthrie and M. J. Natan,
32 J. Liu and V. Hlady, Colloids Surf., B, 1996, 8, 25–37.
J. Am. Chem. Soc., 1996, 118, 1148–1153.
33 J. J. Cras, C. A. Rowe-Taitt, D. A. Nivens and F. S. Ligler, 47 J.-Y. Tseng, M.-H. Lin and L.-K. Chau, Colloids Surf., A, 2001,
Biosens. Bioelectron., 1999, 14, 683–688. 182, 239–245.
34 J. J. Senkevich, C. J. Mitchell, G.-R. Yang and T.-M. Lu, 48 A. Yoshikawa, M. S. Gordon, V. F. Sidorkin and
Langmuir, 2002, 18, 1587–1594. V. A. Pestunovich, Organometallics, 2001, 20, 927–931.
35 N. Balachander and C. N. Sukenik, Langmuir, 1990, 6, 1621– 49 Y. Li, J. Huang, R. T. McIver and J. C. Hemminger, J. Am.
1627. Chem. Soc., 1992, 114, 2428–2432.
36 N. G. Semaltianos, Surf. Coat. Technol., 2007, 201, 7327– 50 M. J. Tarlov and J. G. Newman, Langmuir, 1992, 8, 1398–1405.
7338.
51 M. H. Schoensch and J. E. Pemberton, J. Am. Chem. Soc.,
1998, 120, 4502–4513.
37 T. M. Willey, A. L. Vance, T. van Buuren, C. Bostedt,
L. J. Terminello and C. S. Fadley, Surf. Sci., 2005, 576, 188– 52 M. M. Kreevoy, E. T. Harper, R. E. Duvall, H. S. Wilgus and
196. L. T. Ditsch, J. Am. Chem. Soc., 1960, 82, 4899–4902.
38 K. Heister, M. Zharnikov, M. Grunze, L. S. O. Johansson and 53 H. Rieley, G. K. Kendall, F. W. Zemicael, T. L. Smith and
A. Ulman, Langmuir, 2001, 17, 8–11.
S. Yang, Langmuir, 1998, 14, 5147–5153.
39 G. L. Ellman, Arch. Biochem. Biophys., 1959, 82, 70–77.
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 46527–46535 | 46535