In situ studies of the adsorption kinetics of 4nitrobenzenediazonium salt on gold
Dilushan R. Jayasundara,a Ronan J. Cullen,a Laura Soldi,a
and Paula E. Colavita.a, b *
a - School of Chemistry, University of Dublin Trinity College, College Green, Dublin 2, Ireland.
b - Centre for Research on Adaptive Nanostructures and Nanodevices (CRANN), Trinity College
Dublin, Dublin 2, Ireland
*
Corresponding author. E-mail: colavitp@tcd.ie.
1
Abstract
Self-assembled organic layers are an important tool for modifying surfaces in a range of applications in
materials science. Covalent modification of metal surfaces with aryldiazonium cations has attracted
much attention primarily because this reaction offers a route for spontaneously grafting a variety of
aromatic moieties from solution with high yield. We have investigated the kinetics of this process by
performing real-time, in situ nanogravimetric measurements. The spontaneous grafting of 4nitrobenzene diazonium salts onto gold electrodes was studied via quartz crystal microbalance (QCM)
from aqueous solutions of the salt at varying concentrations. The concentration dependence of the
grafting rate within the first 10 min is best modelled by assuming a reversible adsorption process with
free energy comparable to that reported for arylthiols self-assembled on gold. Multilayer formation was
observed after extended grafting times and was found to be favoured by increasing bulk concentrations
of the diazonium salt. Modified gold surfaces were characterized ex situ with cyclic voltammetry,
Infrared Reflection Absorbance Spectroscopy and X-ray Photoemission Spectroscopy. Based on the
experimentally determined free energy of adsorption and on the observed grafting rates we discuss a
proposed mechanism for aryldiazonium chemisorption.
Keywords: functionalization, diazonium, aryldiazonium, gold, QCM, Langmuir, adsorption,
chemisorption.
2
1. Introduction
The ability to form molecular organic coatings bearing specific functional groups via self-assembly
has been the subject of intense investigation for more than two decades. Molecular self-assembly on
metal substrates has allowed researchers to prepare well defined surfaces for numerous applications
ranging from chemical sensors to organic electronic devices.1-5 Furthermore, self-assembled
organic/metal interfaces constitute excellent model systems for fundamental studies of reactions at
surfaces, adhesion, adsorption and charge transfer.6-8
Covalent modification of metal surfaces with aryldiazonium salts has emerged as an important
reaction for the immobilization of a variety of aromatic chemical moieties from solution with high
yield.9-12 Aryldiazonium salts can be grafted via immersion of a metallic surface in a solution of these
compounds, in either aqueous or organic solvents. The effectiveness of this metal functionalization
strategy has led to an increased interest in its use for nanomaterial synthesis and modification.13-16 In the
case of gold substrates it has recently been demonstrated that spontaneous grafting occurs via reduction
of aryldiazonium by the gold surface17 and leads to formation of Au—C bonds18 according to scheme I,
as initially hypothesized when the reaction was first reported.19
The simplicity of aryldiazonium spontaneous grafting and the robustness of the resulting covalent
Au—C anchoring bond20 offer considerable practical advantages for applications in micropatterning,21
sensors or surface modification/capping of gold nanomaterials.13, 18 For many of these applications, it is
critical to achieve control over molecular coverage and surface defect density, which in turn requires a
fundamental knowledge of experimental factors controlling kinetics and reaction yields. In the case of
aryldiazonium grafting this is of particular concern, since aryl radicals formed during the reaction are
known to graft on surface-bound aryl groups, resulting in multilayer formation and poorly defined,
inhomogeneous films.22 Recent work by Downard and co-workers demonstrated that the structure of
organic layers obtained via spontaneous grafting on metals varies considerably depending on deposition
time and solvent.17 The study was carried out in weakly acidic solutions and provided evidence for the
existence of two stages in the growth rate or aryldiazonium layers: a fast initial deposition followed by a
3
slower increase in coverage that saturates at approximately 3 molecular layers. However, very little is
still known about the kinetics of aryldiazonium chemisorption on gold and, importantly, whether it is
possible to use kinetics to control the final structure/coverage of the organic layer.
We have carried out an investigation of the kinetics of adsorption of 4-nitrobenzenediazonium
(pNBD) on gold in aqueous solutions using Quartz Crystal Microbalance (QCM) techniques. The ability
of a QCM to detect mass changes at a surface with nanogram resolution allowed us to monitor the
adsorption process by directly determining the change in deposited mass during pNBD layer growth.
QCM experiments in liquids have been used in order to monitor the kinetics of other adsorption
reactions, such as those of thiolated molecules on gold, and have provided valuable information for
determining adsorption rates, thermodynamics of adsorption and the mechanism of self-assembly.23-31
Quantitative estimates of the surface coverage of pNBD layers are typically obtained in the literature by
integration of the voltammetric signal associated with the reduction of nitrophenyl groups; this
technique has been shown to be often unreliable because not all surface bound nitrophenyl groups are
necessarily also electrochemically accessible.11, 32, 33 Nanogravimetric measurements, on the other hand,
provide a direct quantitative estimate of the accumulated mass at the surface, in situ and in real time and
are therefore ideally suited to providing kinetic information on pNBD grafting reactions.
We monitored adsorption reactions in aqueous solution via QCM at varying pNBD concentrations,
and found that the grafting rate at early times could be modelled as a reversible Langmuir adsorption
process. Adsorption rate constants were used to calculate a free energy of adsorption for pNBD on gold.
These results were supported also by structural characterization of pNBD layers using a combination of
electrochemistry, infrared spectroscopy and X-ray Photoemission Spectroscopy (XPS).
2. Experimental Section
2.1 Chemicals and Materials. Acetonitrile (ACN, HPLC grade, Fisher), sodium perchlorate (Sigma)
and absolute ethanol (EtOH), methanol (semiconductor grade, Aldrich), hydrogen peroxide (35%,
Sigma), sulfuric acid (Aldrich), acetone (HPLC grade, Aldrich) and 4-nitrobenzenediazonium
4
tetraflouroborate (pNBD, Aldrich) were used as received. Deionized water was used for all aqueous
solutions. All glassware and QCM cells were cleaned with piranha solution (3:1, H2SO4 to H2O2) before
use (WARNING: Piranha solution should be handled with caution; it is a strong oxidant and reacts
violently with organic materials. It also presents an explosion danger. All work should be performed
under a fume hood). pNBD solutions were made to the required concentration and deareated with Ar;
pNBD solutions were used for grafting within 1 h of being prepared to ensure that the salt was not
hydrolyzed significantly during experiments.34
2.2 Sample characterization. Quartz Crystal Microbalance (QCM) was used to monitor the kinetics
of spontaneous grafting on gold surfaces. 10 MHz crystals with 100nm thick vapour-deposited gold
electrodes were used in this study (International Crystal Manufacturing). Crystals were
electrochemically cleaned in a 0.1 M H2SO4 solution: the potential of the gold crystal was repeatedly
cycled between -0.3 and 1.5 V (vs.Ag/AgCl) until no changes in the Au oxidation/reduction peaks were
observed.35 The electrode was then washed with copious amounts of deionized water and dried in a
stream of Ar before being inserted into a static cell (International Crystal Manufacturing).
The QCM setup consists of a static Teflon reaction cell, a lever oscillator and a frequency counter
(SR620, Stanford Research) connected to a computer for data recording using LabVIEW software. The
crystal was clamped in the static cell with O-rings on both sides resulting in only one face being
immersed in the liquid with a geometric area of 0.205 cm2. The cell was placed inside a home-built
temperature-controlled box equipped with Peltier cooling units that maintained temperature at 20  0.5
C. The box also served as a Faraday cage in order to minimize electrical noise. Frequency was recorded
initially on the dry crystal under Ar atmosphere; deionized and Ar-purged water was then injected into
the cell to a volume of 4.550 mL. Once the system had reached frequency stability to  1 Hz
(approximately 3-4 h), the contents of the cell were stirred for 40 s and, immediately afterwards, 50 L
of pNBD stock solution were injected.36 The above method was adopted in order to ensure mixing and
minimize temperature and viscosity changes introduced by solution injections. Control tests carried out
with only stirring or with 50 L water injections showed that this procedure preserves the frequency
5
stability of the QCM cell. The frequency was recorded continuously after injection in order to monitor
the deposition rate. After reaction with pNBD, crystals were rinsed with copious amounts of ACN and
dried in Ar before any subsequent characterization. f vs. time curves were obtained by correcting the
raw frequency data using the frequency baseline prior to injection and removing electrical noise spikes
from the data (Igor Pro 6.04).
XPS characterization was performed on an Omicron ultrahigh vacuum system at 1x10-10 mbar base
pressure, equipped with a monochromatized Al K source (1486.6 eV) and a multichannel array
detector. Spectra were recorded with an analyzer resolution of 0.5 eV at 45 take-off angle. Atomic area
ratios were determined by fitting to Voigt functions after background correction using commercial
software (Igor Pro 6.04).
Cyclic voltammetry (CV) was performed on a potentiostat (CHI660C) using a three-electrode setup
with Pt wire and Ag/AgCl (IJ Cambria) as counter and reference electrodes, respectively. CV was
carried out on the QCM static cell, with the Au-coated crystal as the working electrode. Argon purged
solutions of supporting electrolyte containing 0.1 M NaClO4 in 1:9 EtOH:H2O were used for all
electrochemical experiments.
Infrared-Reflection Absorption Spectroscopy (IRRAS) was performed on a Fourier Transform
Infrared (FTIR) spectrometer (Bruker Tensor 27) using a mercury-cadmium-telluride (MCT) detector
and a VeeMaxII variable angle specular reflectance accessory with wire grid polarizer. Spectra were
collected using p-polarized light at 80 incidence from the surface normal; 256 scans at 4 cm-1 resolution
were collected for both sample and background.
3. Results
3.1 Nanogravimetric monitoring of pNBD adsorption
Nanogravimetric experiments using QCM rely on the calculated mass change at a quartz crystal by
measuring the frequency change as expressed by the Sauerbrey equation:37
6
f  
2 f 02
m
A 
(1)
Where, f0 is the resonance frequency of the fundamental mode of the QCM in air, A is the effective
surface area of the electrodes, and  and  are the density and shear modulus of quartz. This equation is
valid for crystal oscillations in air and for mass changes arising from rigidly coupled masses. When
crystal oscillations occur in liquids, as in the case of our measurements, it is necessary to account for
additional contributions to the resonant frequency:38, 39
f  f m  f y  f a  f x
(2)
where, fm is due to the adsorbed mass, fy is due to viscous damping, fa is due to surface stress and
fx arises from non-shear coupling. The frequency change due to viscous damping is negligible in our
case due to the extremely low concentrations used and the small amount of liquid injected into the
reaction cell (1% of the total volume). Viscous damping arising from deposition of an organic layer has
been shown to be negligible in the case of alkylthiol self-assembled monolayers on gold.38, 39 We can
assume that viscous damping is also negligible in the case of pNBD layers, since these have been shown
to typically grow to approximately 2 nm thickness under similar conditions.17, 18, 20 Finally, fa and fx
contributions, although they may vary between experimental runs, are typically found to be time
independent during adsorption processes from solution.38, 39 Therefore, frequency changes of the QCM
crystal can be satisfactorily approximated with equation (1), under the experimental conditions used for
our studies. The resonant frequency of the QMC prior to injection of pNBD into the cell is considered as
the baseline in determining the frequency shifts and thereby the mass shifts for all of our experiments.
Figure 1a, shows the frequency change as a function of time obtained using four different pNBD
concentrations of 5 M (trace A), 10 M (trace B), 100 M (trace C) and 500 M (trace D) . Control
tests carried out by injecting water instead of pNBD stock solution, showed that the observed decrease
in frequency can only be attributed to pNBD surface adsorption. Considering the area of the QCM
crystal exposed to solution and a theoretical maximum coverage for pNBD40 of 12  10-10 mol/cm2, it is
7
possible to apply equation (1) in order to convert the frequency curves in Figure 1a to surface coverage
() curves for pNBD.
Figure 1b shows surface coverage as a function of time obtained for 5 M (trace A), 10 M (trace B),
100 M (trace C), and 500 M (trace D) pNBD concentrations. At the highest concentration the
deposition curve suggests that the deposited mass rapidly increases to values higher than those expected
for a pNBD monolayer and appears to saturate within 60 min to the equivalent of three pNBD layers.
This result is in agreement with previously reported thickness measurements of spontaneously grafted
layers on gold, that indicate that pNBD films grow up to 2-3 layers when pNBD bulk concentrations are
> 0.001 M.17, 20 For the lowest concentrations ( 100 M) it is possible to identify two regimes in the
deposition curves: at early times there is a rapid increase in coverage followed by a slower deposition
rate at longer times. At early times (<10 min) and low concentrations ( 100 M) the coverage remains
within 1 ML; we therefore decided to investigate whether analytical models of adsorption could be
applied to modelling of pNBD deposition curves within this sub-monolayer region.
3.2 Modelling of deposition curves
The deposition rate at early times was found to be concentration dependent; Figure 2 shows the
deposition curves during the first 10 min for all of the concentrations used for the analysis. Modelling of
the deposition curves shown in Figure 2 can provide information on the mechanism of adsorption of
pNBD on gold as well as be used to obtain kinetic and thermodynamic information on the reaction.
In quiescent solutions, pNBD molecules must first diffuse towards the gold surface in order to be
either chemisorbed directly at an empty surface site from solution, or adsorbed onto the surface; both of
these processes would lead to an increase of the measured mass. In the case of adsorption, once pNBD is
at the surface it can subsequently chemisorb via release of N2 and formation of Au—C bonds, or it can
desorb back into solution. Scheme II shows a summary of reactions describing these two possible
pathways resulting in mass accumulation at the surface. Scheme II assumes that no surface diffusion or
multilayer adsorption takes place; we will also assume in the following analysis that all surface sites are
8
equivalent and that interactions between molecules can be neglected so that rate constants ki are
coverage-independent.
If the reaction followed the first pathway, under the assumption that diffusion is a fast process (i.e. the
deposition rate is not limited by mass transport), the coverage should increase according to the following
rate equation and associated integrated rate law:41
d
 k sol (1   )c
dt
(3)
 (t )  1  exp( k ct )
sol
(4)
where the concentration c can be considered equivalent to the bulk concentration and the rate constant
ksol is concentration independent.
If we assume that the reaction follows the second pathway, always under the assumption that diffusion
to the surface is fast, it is useful to consider two extreme cases depending on the relative value of kch and
kd. If kch >> kd then the adsorption of pNBD onto the Au surface constitutes the rate determining step.
This assumption would result in a time evolution of  according to the equivalent of (4) but with rate
constant ka; the overall process could be described as an irreversible 1st order adsorption. The time
evolution of surface coverage under this mechanistic hypothesis would be indistinguishable from that of
pathway 1, since we would expect to observe exponential behaviour with a concentration-dependent
time constant and concentration-independent limiting coverage.
If kch << kd then the adsorption equilibrium would be established over a shorter timescale than that
needed for chemisorption to occur (pre-equilibrium approximation) and the rate equations would be
those of a reversible Langmuir adsorption process:41
d
 k a (1   )c  k d 
dt
(5)
 (t )  A[1  exp( k obst )]
(6)
with, A     
c
and k obs  k a c  k d
c  (k d k a )
9
The rate law in equation (6) predicts an exponent that varies linearly with concentration and a
concentration-dependent limiting coverage.
We modelled our experimental  vs. time curves using both equation (4) and equation (6). Fits carried
out using equation (4) were found to be unsatisfactory, thus suggesting that neither an irreversible
adsorption model nor a direct reaction from solution model are suitable for explaining the observed
deposition rates. Equation (6), on the other hand, was found to satisfactorily fit all of the curves for 5,
10, 50 and 100 M concentrations. In order to minimise any potential contributions to the total mass
arising from multilayer growth (which would not be modelled by eq. 6), these fits were carried out by
fitting curves up to deposition times that did not result in a coverage greater than 0.60 ML. Figure 3a
shows an example of fits to these two equations for 10 M pNBD concentrations over the first 10 min of
deposition; Table 1 shows the values of kobs and () obtained for all four concentrations investigated
in our work. Both kobs and () were shown to vary with bulk pNBD concentrations, in particular, ()
was found to increase with increasing c, as expected from a reversible Langmuir adsorption process.
Figure 3b shows a plot of kobs values as a function of concentration calculated from multiple
measurements and fits. A linear fit of kobs vs. c yields values for ka and kd of 97  26 M-1s-1 and (1.6 
0.7) 10-3 s-1, respectively. The concentration dependence of () and a non-zero value obtained for kd
indicate that the deposition curves at early times and at low bulk concentrations can be modelled as a
Langmuir adsorption equilibrium system.
These results were obtained under the assumption that the observed deposition rates were under
adsorption control, not under mass transport control. Although the good fits obtained using equation (6)
suggest that this is a correct assumption, we also attempted fits with two diffusion controlled models
that are known to lead to analytical solutions and have been previously used for adsorption studies in
solution. First, experimental curves were modelled as a purely diffusion controlled irreversible process,
without a limit on the number of surface sites, by fitting data to equation 7:36, 42, 43
 (t )  Ac t
(7)
10
where A is a constant and c is the bulk concentration. Second, we tested a diffusion-limited Langmuir
adsorption model as developed by Rahn and Hallock,36, 44 with a fit to the expression:
 (t )  A1  exp(  t 0.5 )
(8)
Neither of these two diffusion-controlled models was found to fit our data satisfactorily, thus
supporting the assumption that, under our experimental conditions, the observed deposition rates are
adsorption controlled. Figure 3a shows an example of a fit to the diffusion-controlled model of equation
(8), compared to those obtained for adsorption-controlled models described by eq. (4) and (6).
From the experimentally obtained values of ka and kd it is possible to calculate a free energy of
adsorption ΔGa for the pNBD-gold system:30, 41
K eq 
ka
kd
(9)
Gad   RT ln K eq (10)
For the pNBD adsorption on gold we obtain a Keq of 60625  31105 M-1 which leads to a ΔGa value of
-26.8  1.2 kJ mol-1.
3.3 Characterization of pNBD layers on gold
QCM results provide real time information on mass density at the surface, however, they do not
provide structural information on the deposited films. For this reason we carried out a characterization
of pNBD layers on gold using electrochemical methods, IRRAS and XPS.
After pNBD grafting it is possible to use cyclic voltammetry in order to estimate the surface coverage
of nitrophenyl groups at the surface. Figure 4a and b show cyclic voltammograms (CVs) obtained using
a quartz crystal as working electrode in 0.1 M NaClO4 in EtOH/H2O at 0.100 V s-1,17 after deposition
from 10 M pNBD aqueous solutions for 2 and 60 minutes followed by rinsing in water; both first and
second scans are shown in the figure. Both CVs show the characteristic irreversible electroreduction
peak arising from the reduction of Ph—NO2 to Ph—NH2 at approximately -0.8 V (vs. Ag/AgCl), thus
indicating that the presence of nitrophenyl groups can be detected even after 2 min in the deposition
solution. The return scan for the layer deposited for 60 min also displays a quasi-reversible redox wave
11
at -0.3 V (vs. Ag/AgCl) that originates from a 2-electron oxidation of Ph—NHOH to Ph—NO. The
presence of this peak indicates that some of the nitrophenyl groups undergo only a partial reduction to
hydroxylaminophenyl instead of being fully reduced to aminophenyl groups; this is in agreement with
previous reports of the electrochemistry of pNBD grafted layers.17,45,46 The hydroxyaminophenyl
oxidation peak is absent in the CVs obtained after 2 min deposition (Figure 4a) suggesting that
electroreduction of nitrophenyl to aminophenyl goes to completion in the case of samples prepared over
short deposition times. Layers at the early stages of deposition are likely to allow for greater
electrochemical access to all of the nitrophenyl groups within the pNBD layer,32 thus facilitating their
complete reduction. This is further supported by the presence of an oxygen electroreduction peak at
-0.25 V (vs. Ag/AgCl) on the 2 min sample (absent in the 60 min sample). This peak is typically
observed in perchlorate solutions at bare Au47-49 and therefore indicates that gold is exposed to the
electrolyte solution and pNBD molecules do not passivate its surface at early deposition times.
The total electroreduction charge was integrated in order to obtain an estimate of Ph—NO2 surface
coverage,17,
32
yielding values of 0.8 10-10 and 9.210-10 mol/cm2 after 2 and 60 min, respectively.
These values correspond to surface densities of 0.06 ML and 0.77 ML at 2 and 60 min. This difference
in coverage is also reflected in the IRRAS spectra of these layers, as shown in Figure 4c. After 2 min of
immersion in 10 M pNBD aqueous solutions followed by rinsing in water, the IRRAS profile does not
display the characteristic peaks of nitrophenyl groups (bottom trace). When gold surfaces are immersed
for 60 min (top trace) it is possible to detect the symmetric and antisymmetric N—O stretching peaks at
1350 and 1524 cm-1, respectively, associated to ArNO2 moieties.
Surface coverage values obtained ex situ via CV are at least 65% lower after 2 min, and 50% lower
after 60 min, than those obtained in situ via QCM (see Figure 1 and 2). The difference in estimates of
surface coverage at early deposition times can be ascribed to removal of physisorbed pNBD after rinsing
and immersion in aqueous electrolyte solution prior to CV characterization. Therefore, it is reasonable to
assume that the majority of pNBD present at gold surfaces during the initial stages of deposition is
physisorbed, whereas chemisorption is a much slower process. The presence of physisorbed nitrophenyl
12
moieties thus supports a proposed mechanism for spontaneous grafting involving reversible adsorption
at early deposition times.
XPS spectra were taken on pNBD adsorbed on gold in order to investigate the chemical composition
of the films. Figure 5 shows the N 1s region of a bare Au-coated quartz crystal (trace A) and of samples
prepared from 10 M pNBD aqueous solutions after deposition for 2 (trace B) and 60 min (trace C),
followed by rinsing. After 60 min the XP spectrum displays two peaks with maxima at 405.5 and 399.0
eV. The sharp distinct peak centred at 405.5 eV is assigned to the N 1s of the –NO2 group based on
previous reports of nitrophenyl grafting on metals and carbon.10, 17, 20, 50-55 The broader peak at 399.0 eV
has been attributed to the reduction of –NO2 groups during XP measurements10, 50 or, alternatively, to
the presence of –N=N– moieties.20,
51, 52
Our XP measurements in the N 1s region were taken
immediately after insertion in the analysis chamber and were collected using a low number of repeat
scans to minimise reduction processes under photoelectron emission; we therefore propose that the
dominant contribution to the 399.0 eV peak in Figure 5 arises from azo moieties in the organic layer.
Finally, the absence of a peak at 403.8 eV, typically attributed to diazo groups suggests that the terminal
diazonium cation of pNBD molecules is not present in these layers.12
The area ratio A399 : A405 was found to be 1.3 : 1; this value can be used to estimate the proportion of
molecules chemisorbed via robust Au—C bonds. Since purely physisorbed molecules are expected to
yield AlowBE : AhighBE ratios of 2 : 1, our observed value indicates that approximately 35% of adsorbed
pNBD molecules in the layer have dissociated dinitrogen during the adsorption process, whereas 65% of
them display azo groups in their structure. After only 2 min of deposition it is possible to discern the
appearance of the same two peaks in the N 1s region, however, these two peaks are at the limit of
detection and it is not possible to provide a quantitative estimate of their relative contributions.
In summary, electrochemical characterization of pNBD layers reveals a significantly lower coverage
than that measured in situ, at comparable deposition times, via QCM. This difference likely arises from
the removal of physisorbed pNBD prior to ex situ characterization, since it is observed even at very low
coverages, when all nitropheny groups are likely to be accounted for by the electroreduction charge. XP
13
spectra of pNBD spontaneously assembled on Au show that a significant proportion of these molecules,
approximately 65% of them, do not undergo dediazotization. This finding suggests that molecules
bound via Au—C bonds might constitute a third of the molecules found at Au surfaces under our
deposition conditions.
4. Discussion
QCM techniques allowed us to monitor the adsorption/grafting of pNBD from aqueous solutions on
Au surfaces in situ and in real time. Our results show that at concentrations higher than 100 M these
organic layers can grow rapidly into multilayers with saturation coverages of approximately 2-3 layers,
in agreement with previous reports of spontaneous pNBD grafting on Au.17,
20
However, when
concentrations are kept below this threshold, pNBD coverages remain in the sub-monolayer regime over
prolonged deposition times. We also observed that at early times the grafting rate is strongly
concentration-dependent, with increasing concentrations leading to higher deposition rates. This is not
the case after longer deposition times:  vs. time curves showed considerable irreproducibility across
samples after the first 30 min of deposition; the slow growth rates observed in this pNBD deposition
regime could not be correlated to bulk pNBD concentrations in any clear manner. It is likely that these
later stages of deposition involve significant contributions to the coverage that originate from multilayer
deposition. We therefore decided to investigate whether an analytical model of adsorption kinetics could
provide mechanistic insights into the early stages of pNBD deposition, where sub-monolayer deposition
takes place.
Langmuir adsorption kinetics was found to satisfactorily model  vs. time curves at early times for all
four concentrations investigated. This mechanism assumes that a first step in the formation of pNBD
layers on gold involves physisorption of the nitrobenzenediazonium molecule onto Au. This adsorbed
molecule would then convert to a covalently bound species through a much slower process, as shown in
pathway 2 in Scheme II. This mechanism had been previously proposed for pNBD grafting on carbon
nanomaterials, such as graphene and nanotubes. Strano and co-workers have shown that spontaneous
14
aryldiazonium grafting on carbon nanotubes involves a charge-transfer mediated adsorption at the
surface, followed by a slow conversion of the aryldiazonium to a chemisorbed species.56-58 Similarly,
Stark and co-workers recently demonstrated spectroscopically that the first step of pNBD grafting on
graphene is physisorption, leading to the formation of a charge-transfer complex with the diazo group
still attached to the molecule.59 The good agreement observed in our results between experimental
deposition curves and a reversible Langmuir adsorption process indicates that this mechanism might be
valid for the spontaneous grafting of pNBD on Au.
Lehr and Downard showed that, in acidic solutions, grafting on gold leads to an increase in the open
circuit potential, thus suggesting that charge-transfer is also an important step for diazonium reactions
on Au.17 However, there have been no reports suggesting a structure for this charge transfer complex.
Interestingly, XP spectra of spontaneously grafted pNBD layers show N 1s profiles with a large
contribution at 399 eV, a lower binding energy relative to the peak at 405.5 eV arising from -NO2
groups and lower than peaks at 403 eV arising from –N2+ end groups in pNBD.45 The peak at 399 eV
has a binding energy characteristic of azocompounds and has been previously attributed to Ph—N=N—
Ph in multilayers or to Metal—N=N—Ph groups within the layer.52 Its presence therefore indicates that
a high proportion of molecules at the Au surface do not loose dinitrogen during the course of pNBD
deposition. The significant contribution of this peak even at early deposition times and sub-monolayer
coverages, however, suggests that azo-moieties in spontaneously grafted layers might not be necessarily
associated to multilayer growth. Formation of Au—N=N—Ph groups is therefore a more likely
explanation, although from our present results it is difficult to determine whether this structure is an
intermediate or a secondary product of pNBD grafting on gold.
Measurements of deposition curves at different concentrations allowed us to provide an estimate for
the free energy of adsorption, which was found to be -26.8 kJ/mol. This value is slightly more negative
than reported free energies of adsorption for short linear alcohols,60-62 and comparable to those of
alkylamines,63 alkylthiols30, 43, 64 and arylthiols65, 66 on gold. It is important to notice that our calculated
Ga is significantly lower in absolute value than bond energies for covalent immobilization on Au
15
through Au—C bonds, which are computationally estimated at 100 kJ/mol.67 Therefore, it is unlikely
that the adsorption/desorption equilibrium identified as the first growth regime (<10 min) involves the
formation of covalent bonds between the substrate and pNBD molecules. This is also the case for
assembly of thiols on gold, which display free energies of adsorption ~20-25 kJ/mol while the Au—S
bond energy is ~160 kJ/mol.65 For thiols, in fact, it has been shown that a first adsorption step is
followed by a second step with slower kinetics68,
69
and that, similarly to our experiments, ex situ
determinations of surface coverage result in considerably lower values than in situ ones.68
5. Conclusions
We have carried out an in situ investigation on the kinetics of adsorption of 4-nitrobenzenediazonium
salts on Au surfaces using QCM techniques. We found that the early stage of pNBD grafting involves a
reversible adsorption step, similarly to proposed mechanisms for the covalent grafting of pNBD on
carbon nanomaterials. This first kinetic step is obscured when bulk concentrations are higher than
approximately 100 M, in which case multilayer growth is observed to rapidly occur. Chemisorption
leading to Au—C bond formation appears to be a slower process that takes place once a physisorbed
pNBD precursor is formed.
The spontaneous assembly of aryldiazonium has recently emerged as an excellent strategy for the
functionalization and stabilization metal nanomaterials as well as thin films and macroscopic samples.
Our results offer mechanistic insights on this reaction as well as practical guidelines for controlling
rates of deposition and coverage of these organic layers.
Acknowledgements. This publication has emanated from research conducted with the financial
support of Science Foundation Ireland under Grant Number 09/RFP/CAP2174 and of the Irish Council
for Science Engineering and Technology (IRCSET) under the Postdoctoral Fellowship Scheme. The
16
authors are grateful to Dr. Cormac McGuinness and Dr. Michael E.G. Lyons for granting access to XPS
and electrochemical instrumentation respectively.
References
1.
Shenhar, R.; Norsten, T. B.; Rotello, V. M., Adv. Mater. 2005, 17, 657-669.
2.
Chaki, N. K.; Vijayamohanan, K., Biosens. Bioelectron. 2002, 17, 1-12.
3.
Huie, J. C., Smart Mater. Struct. 2003, 12, 264.
4.
Fendler, J. H., Chem. Mater. 2001, 13, 3196–3210.
5.
Gooding, J. J.; Mearns, F.; Yang, W.; Liu, J., Electroanalysis 2003, 15, 81-96.
6.
Mrksich, M.; Whitesides, G., Annu. Rev. Biophys. Biomol. Struct. 1996, 25, 55-78.
7.
Chaudhury, M. K., Curr. Opin. Colloid Interface Sci. 1997, 2, 65-69.
8.
Heath, J. R., Annu. Rev. Mater. Res. 2009, 39, 1-23.
9.
Combellas, C.; Delamar, M.; Kanoufi, F.; Pinson, J.; Podvorica, F. I., Chem. Mater. 2005, 17,
3968-3975.
10.
Adenier, A.; Cabet-Deliry, E.; Chausse, A.; Griveau, S.; Mercier, F.; Pinson, J.; Vautrin-Ul, C.,
Chem. Mater. 2005, 17, 491-501.
11.
Adenier, A.; Barre, N.; Cabet-Deliry, E.; Chausse, A.; Griveau, S.; Mercier, F.; Pinson, J.;
Vautrin-Ul, C., Surf. Sci. 2006, 600, 4801-4812.
12.
Stewart, M. P.; Maya, F.; Kosynkin, D. V.; Dirk, S. M.; Stapleton, J. J.; McGuiness, C. L.;
Allara, D. L.; Tour, J. M., J. Am. Chem. Soc. 2004, 126, 370-378.
13.
Mirkhalaf, F.; Paprotny, J.; Schiffrin, D. J., J. Am. Chem. Soc. 2006, 128, 7400-7401.
14.
Griffete, N.; Herbst, F.; Pinson, J.; Ammar, S.; Mangeney, C., J. Am. Chem. Soc. 2011, 133,
1646-1649.
15.
Ghosh, D.; Chen, S. W., Chem. Phys. Lett. 2008, 465, 115-119.
16.
Ghosh, D.; Pradhan, S.; Chen, W.; Chen, S. W., Chem. Mater. 2008, 20, 1248-1250.
17.
Lehr, J.; Williamson, B. E.; Flavel, B. S.; Downard, A. J., Langmuir 2009, 25, 13503-13509.
18.
Laurentius, L.; Stoyanov, S. R.; Gusarov, S.; Kovalenko, A.; Du, R. B.; Lopinski, G. P.;
McDermott, M. T., ACS Nano 2011, 5, 4219-4227.
19.
Fan, F. R. F.; Yang, J. P.; Cai, L. T.; Price, D. W.; Dirk, S. M.; Kosynkin, D. V.; Yao, Y. X.;
Rawlett, A. M.; Tour, J. M.; Bard, A. J., J. Am. Chem. Soc. 2002, 124, 5550-5560.
17
20.
Shewchuk, D. M.; McDermott, M. T., Langmuir 2009, 25, 4556-4563.
21.
Lehr, J.; Garrett, D. J.; Paulik, M. G.; Flavel, B. S.; Brooksby, P. A.; Williamson, B. E.;
Downard, A. J., Anal. Chem. 2010, 82, 7027-7034.
22.
Pinson, J.; Podvorica, F., Chem. Soc. Rev. 2005, 34, 429-439.
23.
Rodenhausen, K. B.; Duensing, B. A.; Kasputis, T.; Pannier, A. K.; Hofmann, T.; Schubert, M.;
Tiwald, T. E.; Solinsky, M.; Wagner, M., Thin Solid Films 2011, 519, 2817-2820.
24.
Newman, J. D. S.; Blanchard, G. J., Phys. Chem. Chem. Phys. 2007, 9, 6142-6149.
25.
Tamerler, C.; Duman, M.; Oren, E. E.; Gungormus, M.; Xiong, X. R.; Kacar, T.; Parviz, B. A.;
Sarikaya, M., Small 2006, 2, 1372-1378.
26.
Bieri, M.; Burgi, T., Phys. Chem. Chem. Phys. 2006, 8, 513-520.
27.
Bieri, M.; Burgi, T., J. Phys. Chem. B 2005, 109, 22476-22485.
28.
Liao, S.; Shnidman, Y.; Ulman, A., J. Am. Chem. Soc. 2000, 122, 3688-3694.
29.
Lee, Y. J.; Jeon, T. C.; Paik, W. K.; Kim, K., Langmuir 1996, 12, 5830-5837.
30.
Karpovich, D. S.; Blanchard, G. J., Langmuir 1994, 10, 3315-3322.
31.
Thomas, R. C.; Sun, L.; Crooks, R. M.; Ricco, A. J., Langmuir 1991, 7, 620-622.
32.
Ceccato, M.; Nielsen, L. T.; Iruthayaraj, J.; Hinge, M.; Pedersen, S. U.; Daasbjerg, K., Langmuir
2010, 26, 10812-10821.
33.
Podvorica, F. I.; Kanoufi, F.; Pinson, J.; Combellas, C., Electrochim. Acta 2009, 54, 2164-2170.
34.
Canning, P. S. J.; McCrudden, K.; Maskill, H.; Sexton, B., J. Chem. Soc.-Perkin Trans. 2 1999,
2735-2740.
35.
Fischer, L. M.; Tenje, M.; Heiskanen, A. R.; Masuda, N.; Castillo, J.; Bentien, A.; Émneus, J.;
Jakobsen, M. H.; Boisen, A., Microelectron. Eng. 2009, 86, 1282-1285.
36.
Hibbert, D. B.; Gooding, J. J.; Erokhin, P., Langmuir 2002, 18, 1770-1776.
37.
Sauerbrey, G. Z., J. Physik 1959, 155, 206-212.
38.
Yang, M.; Thompson, M., Langmuir 1993, 9, 1990-1994.
39.
Yang, M. S.; Thompson, M.; Duncanhewitt, W. C., Langmuir 1993, 9, 802-811.
40.
Liu, Y.-C.; McCreery, R. L., J. Am. Chem. Soc. 1995, 117, 11254-11259.
41.
Masel, R. I., In Principles of Adsorption and Reaction on Solid Surfaces; John Wiley & Sons:
New York, 1996; pp 235-353.
42.
Motschmann, H.; Stamm, M., Macromolecules 1991, 24, 3681-3688.
43.
Schlenoff, J. B.; Li, M.; Ly, H., J. Am. Chem. Soc. 1995, 117, 12528–12536.
18
44.
Rahn, J. R.; Hallock, R. B., Langmuir 1995, 11, 650-654.
45.
Delamar, M.; Désarmot, G.; Fagebaume, O.; Hitmi, R.; Pinsonc, J.; Savéant, J. M., Carbon
1997, 35, 801-807.
46.
Brooksby, P. A.; Downard, A. J., J. Phys. Chem. B 2005, 109, 8791-8798.
47.
Richer, J.; Stolberg, L.; Lipkowski, J., Langmuir 1986, 2, 630-638.
48.
Shao, M. H.; Adzic, R. R., J. Phys. Chem. B 2005, 109, 16563–16566.
49.
Strbac, S.; Adzic, R. R., Electrochim. Acta 1996, 41, 2903-2908.
50.
Mendes, P.; Belloni, M.; Ashworth, M.; Hardy, C.; Nikitin, K.; Fitzmaurice, D.; Critchley, K.;
Evans, S.; Preece, J., ChemPhysChem 2003, 4, 884-889.
51.
D'Amours, M.; Belanger, D., J. Phys. Chem. B 2003, 107, 4811-4817.
52.
Laforgue, A.; Addou, T.; Belanger, D., Langmuir 2005, 21, 6855-6865.
53.
Chamoulaud, G.; Belanger, D., J. Phys. Chem. C 2007, 111, 7501-7507.
54.
Ortiz, B.; Saby, C.; Champagne, G. Y.; Belanger, D., J. Electroanal. Chem. 1998, 455, 75-81.
55.
Sinitskii, A.; Dimiev, A.; Corley, D. A.; Fursina, A. A.; Kosynkin, D. V.; Tour, J. M., ACS Nano
2010, 4, 1949-1954.
56.
Nair, N.; Kim, W.-J.; Usrey, M. L.; Strano, M. S., J. Am. Chem. Soc. 2007, 129, 3946-3954.
57.
Usrey, M. L.; Lippmann, E. S.; Strano, M. S., J. Am. Chem. Soc. 2005, 127, 16129-16135.
58.
Strano, M. S.; Dyke, C. A.; Usrey, M. L.; Barone, P. W.; Allen, M. J.; Shan, H.; Kittrell, C.;
Hauge, R. H.; Tour, J. M.; Smalley, R. E., Science 2003, 301, 1519-1522.
59.
Koehler, F. M.; Jacobsen, A.; Ensslin, K.; Stampfer, C.; Stark, W. J., Small 2010, 6, 1125-1130.
60.
Beltowska-Brzezinska, M.; Luczak, T.; Holze, R., Surf. Sci. 1998, 418, 281-294.
61.
Brzostowska-Smolska, M.; Krysinski, P., Colloids Surf., A 1998, 131, 39-44.
62.
Richer, J.; Lipkowski, J., J. Electroanal. Chem. 1988, 251, 217-234.
63.
Luczak, T., Colloids Surf., A 2006, 280, 125-129.
64.
Jung, L. S.; Campbell, C. T., J. Phys. Chem. B 2000, 104, 11168-11178.
65.
Kolega, R. R.; Schlenoff, J. B., Langmuir 1998, 14, 5469-5478.
66.
Jakubowicz, A.; Jia, H.; Wallace, R. M.; Gnade, B. E., Langmuir 2004, 21, 950-955.
67.
Jiang, D.-e.; Sumpter, B. G.; Dai, S., J. Am. Chem. Soc. 2006, 128, 6030-6031.
68.
Peterlinz, K. A.; Georgiadis, R., Langmuir 1996, 12, 4731-4740.
19
69.
Xu, S.; Cruchon-Dupeyrat, S. J. N.; Garno, J. C.; Liu, G.-Y.; Jennings, G. K.; Yong, T.-H.;
Laibinis, P. E., J. Chem. Phys. 1998, 108, 5002-5012.
20
FIGURE CAPTIONS
Figure 1. (a) Time evolution of QCM resonant frequency after injection of pNBD aqueous solutions to a
total concentration of 5 M (trace A), 10 M (trace B), 100 M (trace C) and 500 M (trace D). (b)
Surface coverage curves as a function of time obtained from curves in (a) using the Sauerbrey equation.
Figure 2. (a) Surface coverage as a function of time up to 10 min for 5 M (trace A), 10 M (trace B),
50 M (trace C) and 100 M (trace D) pNBD concentrations.
Figure 3. (a) Example of fits to a deposition curve obtained for a 10 M pNBD concentration (exp) to
Langmuir reversible adsorption (LA), to diffusion-controlled Langmuir adsorption (DCL)44 and to an
irreversible 1st order adsorption process (IA). (b) Plot of constant kobs vs. bulk concentration obtained
from fits to reversible Langmuir adsorption behaviour; error bars represent 80% C.I.
Figure 4. (a, b) Cyclic voltammograms in 0.1 M NaClO4 in EtOH/H2O at 0.1 V s-1 obtained from gold
coated quartz crystals that had been grafted in 10 M pNBD solutions for (a) 2 min and (b) 60 min. (c)
IRRAS spectra of pNBD layers obtained from the same solution after 2 min and 60 min; spectra have
been offset in order to facilitate the comparison.
Figure 5. XP spectra of Au-coated quartz crystals in the N 1s region after cleaning (bottom trace) and
after deposition of pNBD from 10 M solutions for 2 min (middle trace) and 60 min (top trace).
21
kobs  80% C.I.
(103 s-1)
5.5  2.6
()
.
0.23  0.09
10
1.4  1.4
0.38  0.12
50
7.1  3.3
0.75  0.01
100
10.2  8.1
0.86  0.06
[pNBD]
(µM)
5
Table 1. Summary of values of observed rate constant and limiting coverage obtained from fits of
deposition curves to a reversible Langmuir adsorption model as shown in equation (6).
22
Scheme I. Net chemisorption reaction for 4-nitrobenzenediazonium salts.17,18
Scheme II. Possible mechanisms of covalent grafting for pNBD (M).
23
Figure 1. (a) Time evolution of QCM resonant frequency after injection of pNBD aqueous solutions to a
total concentration of 5 M (trace A), 10 M (trace B), 100 M (trace C) and 500 M (trace D). (b)
Surface coverage curves as a function of time obtained from curves in (a) using the Sauerbrey equation.
24
Figure 2. (a) Surface coverage as a function of time up to 10 min for 5 M (trace A), 10 M (trace B),
50 M (trace C) and 100 M (trace D) pNBD concentrations.
25
Figure 3. (a) Example of fits to a deposition curve obtained for a 10 M pNBD concentration (exp) to
Langmuir reversible adsorption (LA), to diffusion-controlled Langmuir adsorption (DCL)44 and to an
irreversible 1st order adsorption process (IA). (b) Plot of constant kobs vs. bulk concentration obtained
from fits to reversible Langmuir adsorption behaviour; error bars represent 80% C.I.
26
Figure 4. (a, b) Cyclic voltammograms in 0.1 M NaClO4 in EtOH/H2O at 0.1 V s-1 obtained from gold
coated quartz crystals that had been grafted in 10 M pNBD solutions for (a) 2 min and (b) 60 min. (c)
IRRAS spectra of pNBD layers obtained from the same solution after 2 min and 60 min; spectra have
been offset in order to facilitate the comparison.
27
Figure 5. XP spectra of Au-coated quartz crystals in the N 1s region after cleaning (bottom trace) and
after deposition of pNBD from 10 M solutions for 2 min (middle trace) and 60 min (top trace).
28
TOC Figure
29