Excimer probe of the binding of alkyl disulfides to gold nanoparticles and subsequent monolayer dynamics

A symmetric alkyl disulfide bearing two pyrenyl chromophores allows for fluorescence monitoring of its binding to gold nanoparticles and the subsequent diffusion of the geminate thiolates on the particle surface.

Metal nanoparticles protected by self-assembled monolayers (monolayer-protected clusters, MPCs) attract attention from a wide range of disciplines.1–4

These hybrid systems indeed show a number of interesting features.

The protecting monolayer can be chemically modified, making these clusters suitable platforms for combining different functions into one entity.

Another aspect of MPCs is the possibility of observing special photonic effects on attached chromophores, such as enhanced optical absorption and/or emission.5,6

These effects are associated with the plasmon resonance of the electrons in the metal core.7

Functionalised MPCs offer the potential to study such effects in chemically and structurally well-defined systems in solution using a variety of spectroscopic techniques.

In search of plasmon-related effects on chromophores on the surface of MPCs, we are interested in developing simple and flexible procedures to insert fluorescent chromophores as dopants into n-alkanethiol monolayers assembled on the surface of different types of metal nanoparticles.

The nanoparticles that we chose to begin with are “naked particles”,8 a suspension of ∼5 nm diameter gold nanoparticles in toluene, stabilised by 0.04 M tetra-n-octylammonium bromide (TOAB).

The suspension is obtained by the classical Brust method,9 but omitting the thiol.

Naked particles can be functionalised afterwards by suitable thiols and disulfides, which are not exposed to the reaction conditions under which the metal particles themselves are formed.

Because of their relatively large size, naked particles display a clear plasmon resonance (λmax (toluene) = 528 nm) in their optical absorption spectrum.

The suspensions have a deep red colour.

As a model fluorescent dopant for self-assembled n-alkanethiol monolayers, we have synthesised disulfide 1 using commercially available 1-pyrenemethanol. A concomitant advantage of the pyrenyl chromophore is that it offers the possibility to probe the binding of disulfides to nanoparticles and subsequent monolayer dynamics through its excimer emission.

Pyrene-coated gold nanoparticles have been synthesised recently by the Brust method in the presence of pyrenyl-modified alkanethiols.10–13

In these cases, the pyrene chromophores are present in high concentrations on the particle surface, potentially complicating their photophysics.

Very recently, Montalti et al13. studied the displacement of thiols from the particle surface by observing the recovered pyrene monomer emission of fluorophores liberated from the surface.

In the present work, we show that the emission of symmetric disulfide 1 (Scheme 1) provides a new way of studying the binding of molecules to metal surfaces and their subsequent dynamics on these surfaces.

Disulfide 1 itself shows, in addition to pyrene monomer emission, a distinct intramolecular excimer band in its emission spectrum (Fig. 1, solid line).

The intramolecular nature of this band is confirmed by its presence even at very low concentrations (10–7 M) and by the fact that its relative intensity does not vary with concentration.

The excimer band disappears in vitrified media (e.g. in methylcyclohexane at 77 K), and also when 1 is exposed to dithiothreitol, a well-known disulfide cutting reagent.14

We studied the binding of 1 to “naked” gold nanoparticles by fluorescence spectroscopy. The molecule was mixed with larger amounts of di-n-octyldisulfide (typical ratio 5 ∶ 95 12) to have it inserted as a dilute dopant in the particle protecting monolayer.

Starting from a naked particle solution, we synthesised MPCs coated either with a monolayer of 100% of di-n-octyl disulfide152, Aun(2)1.00m, or with a monolayer consisting of 95 mol% of 2 and 5 mol% 1, Aun(2)0.95m(1)0.05m, by adding a tenfold excess of disulfides.

After one week, the volume of the solution was reduced by evaporating the toluene.

The functionalised particles were precipitated using isopropanol.

After centrifugation, the supernatant was replaced with fresh isopropanol, and centrifugation was continued.

This was repeated four times to wash away any remaining free ligand and TOAB.

In the case of Aun(2)0.95m(1)0.05m complete removal of excess ligand was apparent from the absence of fluorescence in the two last isopropanol washings.

After drying, the particles were redissolved in toluene and stored as such.

The UV/Vis spectra (see ESI) of Aun(2)1.00m and Aun(2)0.95m(1)0.05m are almost identical, except for the presence of typical pyrene related peaks (e.g. at 346 nm) in the spectrum of the latter particles.

We observe a broad band (λmax (toluene) = 528 nm) which originates from the surface plasmon resonance of the nanoparticles and whose shape is typical of thiol-modified gold nanoparticles.

Compared to the TOAB-stabilised naked particles, the plasmon of the passivated particles is less pronounced and located at slightly shorter wavelength, which agrees with the presence of thiols on the particle surface.16

Interestingly, the fluorescence emission spectrum of Aun(2)0.95m(1)0.05m displays (Fig. 1, dashed line) a significantly altered excimer/monomer emission intensity ratio with respect to parent compound 1.

This may have several causes.

The excimer emission might be selectively quenched by the metal particle core, or excimer formation may be suppressed in the monolayer on the particle surface.

It might also be that 1 dissociates after chemisorption and that the geminate thiolates migrate over the particle surface, eventually yielding isolated pyrene chromophores that do not form excimers.

To gain more insight into this, we carried out titration experiments in which small aliquots of naked particle solution were added to solutions containing a mixture of 1 and 2 (ratios 1 ∶ 99–5 ∶ 95, total disulfide concentration 1.5 µM).

The series of spectra in Fig. 2 shows that immediately after each addition of naked particles both monomer and excimer emission intensities are diminished.

The pyrene singlet excited state is quenched to a certain extent upon binding of 1 to gold nanoparticles, diminishing monomer fluorescence as well as the formation of excimers from this state.

The reduced amount of excimers formed, then, is further quenched by the nanoparticles diminishing the excimer emission intensity even more than the monomer emission.

Pyrenyl chromophores may be quenched by charge transfer to the metal particle core,11 but energy transfer is also likely to play a role in the fluorescence quenching near metal nanoparticles.17

The titration curve, obtained by plotting the integrated excimer intensity (470–670 nm) as a function of nanoparticle concentration, first descends with constant slope and then appears to slowly level off as the particle concentration is increased.

Since all disulfides are expected to be rapidly and irreversibly chemisorbed by the added nanoparticles,18 one would expect a curve that decreases with constant slope until all disulfides are bound, and then abruptly changes into a horizontal line.

The observed leveling-off suggests that the rapid chemisorption step is followed by a secondary, slower process.

The time between two additions (and hence between points in the titration curve) is about 10 minutes.

If the secondary process is not complete within this time, the shape of the titration curve may be altered.

Fig. 3 confirms that such a process is indeed present.

Immediately after addition of 1 equivalent of particles, both monomer and excimer emission intensities are diminished, but as time goes by, an increase in monomer intensity is observed, accompanied by a decreasing excimer intensity.

A plausible mechanism for the increasing monomer emission at the expense of excimer emission is a migration of the pyrenyl chromophores over the surface of the nanoparticles (Scheme 2).

Such a mechanism is in accordance with observations by Boal and Rotello19 who were able to use the migration of thiolates on gold nanoparticles to obtain self-optimised binding sites on the particle surface.

On flat gold surfaces, diffusion of thiolates has also been observed, mainly by scanning tunneling microscopy,20–22 and more recently by voltammetry.23

The present work indicates that excimer probes such as 1 might provide a rapid, simple and non-destructive method for gaining more information on thiolate diffusion on metal surfaces.

A question remains why there is still some excimer fluorescence present in the emission spectrum of the 5 mol%-doped Aun(2)0.95m(1)0.05m particles.

To verify our guess that this is due to a fraction of the pyrenyls still being statistically close enough to form excimers, we examined nanoparticles covered by protecting monolayers containing different doping levels of pyrenyl-modified thiolates.

At the 1 mol% doping level, no observable excimer fluorescence remains, but as the surface concentration increases excimer fluorescence shows up (Fig. 4).

The pyrenyl-thiolates still do migrate after chemisorption but this does not prevent them having other pyrenyl groups in their vicinity at high doping concentration.

The pyrene excimer is a well-known probe of molecular dynamics, and here we present indications that it may be used in the study of the dynamics of self-assembled monolayers on metallic nanoparticles in solution.

The excimer emission of 1 enables us to probe migration on the molecular scale, averaged over a large number of copies of the same system.

This averaging makes it a valuable tool to obtain kinetic data on the dynamics of thiolates on metallic surfaces.

The method may be extendable to planar metal surfaces, where the data could perhaps be combined with single-molecule data from scanning tunneling microscopy.

The migration of pyrenyl-labelled thiolates suggests that the disulfide bond is cleaved upon chemisorption to the metal surface.

Evidence from surface-enhanced Raman (SERS) studies of disulfides on silver is in line with cleavage of the S–S bond.24

In contrast, grazing incidence X-ray diffraction studies suggest that chemisorbed thiols and disulfides exist as disulfides on the metal surface.25

Such bound disulfides, however, need not have the same chemical properties as disulfides in solution.

Exchange between neighbouring disulfide pairs on the surface may be a mechanism leading to thiol migration.

There exists considerable controversy on the way thiols and disulfides are bound to metal surfaces.

Studies on metal nanoparticles in solution (such as those in ref. 26) may give more insight on the chemical nature of gold–sulfur bonds and their dynamics.

In the perspective of putting isolated fluorophores at a well-defined distance from the surface of metallic particles, an interesting result is that starting from symmetric disulfides bearing two chromophores it is still possible to obtain isolated chromophores at the surface.

However, at doping levels as low as 5 mol% interchromophoric interactions, in this case leading to excimer emission, already start to show up for dopant molecule 1.

Further work elaborating on the present findings is in progress.