1
Adsorption kinetics of charged thiols on gold nanoparticles

2
Colloidal Au nanoparticles have been functionalized with mercaptopropane sulfonate (MPS) and mercaptoethylamine (MEA) short alkanethiols in aqueous solutions.

3
Adsorption kinetics of these charged thiols onto gold nanoparticles has been followed by the decrease in the surface plasmon band at 530 nm due to the red shift caused by particle-to-particle aggregation.

4
Titration curves resulted from plotting the 530 and 800 nm-absorbance versus the added amount of thiol in solution, and yielded the mercaptane coverage on Au.

5
While the adsorption of positively charged MEA on negatively charged Au colloid is very rapid, slower and more complex adsorption kinetics have been found for the negatively charged MPS.

Introduction

6
Metal nanoparticles have attracted great interest because of their possible use in “bottom-up” construction of nanoelectronic devices with spherical nanoparticles and cylindrical nanorods1–4 since they possess unique optical, electrical, magnetic and catalytic properties.

7
Gold nanoparticles synthesised by citrate reduction of HAuCl4 solutions are nearly monodisperse spheres of a sized controlled by the preparation conditions, 20 nm in our case.5,6

8
They have a negative surface charge as a consequence of bond citrate and present a characteristic surface plasmon mode absorbance at 520–530 nm for 15–20 nm Au nanoparticles.

9
Interesting optical properties such as surface enhanced Raman scattering (SERS) by as much as 1013–1014 fold7 and visible luminiscence8,9 have been reported for monolayer-protected gold clusters.

10
A red-shift in reflection mode (red to blue colour change of the transmitted light) is observed when the interparticle distance is shortened.10,11

11
This shift is observed for the floculation of functionalized Au nanoparticles for instance when high ionic strength electrolyte solutions are added to the particle colloidal solutions due to screening of the negative charge particle repulsion leading to shorter interparticle distances12 or when DNA,13,14 or heavy metal ion15 linkers are used to form larger linked-particle assemblies.

12
Close proximity of Au nanoparticles in self-assembled films also causes a red-shift phenomena.16–19

13
Theoretical interpretation of light extinction (absorption and scattering) of metal spherical colloids dates back to Mie.20

14
Quinten and Kreibig developed theoretical models for the optical properties of different non-linear aggregates.21

15
Two-dimensional arrays of varying diameter have been obtained after thiol modification of nanocrystals with alkanethiols of different chain length.22

16
The surface plasmon absorption band of metal nanoclusters is sensitive to the adsorbed ions or surface complexed molecules, which dampen the surface plasmon band as shown for colloidal silver particles23,24 with I and SH, and for Au nanoparticles capped with tetraoctylammonium bromide.25

17
Whitesides and co-workers26 have described the change in the floculation of gold nanoparticle dipersions upon chemisorption of alkanethiols bearing terminal functional groups by observation of the spectral changes in the UV-Visible range (red shift).

18
However, a detailed kinetic study of these changes has not been reported.

19
In the present study we analyse the adsorption kinetics of positively and negatively charged short alkanethiols on 20 nm average diameter citrate stabilized gold colloidal nanoparticles5 by the evolution of the optical absorption in the 400–800 nm region.

20
From “titration” experiments of the colloidal suspension with diluted thiol solutions we obtain the moles required for colloidal surface saturation following the changes of the absorption spectra.

21
Kinetic information has been obtained monitoring the attenuation of the plasmon band in presence of large excess of thiol with respect to the concentration required for thiol surface saturation.

Experimental

22
The following materials were obtained from Aldrich: HAuCl4 3H2O (99.999%), trisodium citrate dihydrate (99,9%), 3-mercapto-1-propane sulfonate (MPS), 2-mercaptoethylamine hydrochloride (MEA), 11-mercaptoundecanoic acid (MUA), 6-mercaptohexanol.

23
All the solutions were prepared from Milli Q water.

24
All glassware used in the preparation and derivatisation of colloidal Au was treated with aqua regia (3 parts HCl, 1 part HNO3), rinsed in Milli-Q water, and oven-dried prior to use.

25
Colloidal gold was prepared following the procedure reported by Natan and co-workers.5

26
Briefly, in a 250 ml round-bottom flask equipped with a condenser, 100 ml of 0.01% HAuCl4 was brought to a boil with vigorous stirring; and to this solution was added 1.5 ml of 1% sodium citrate.

27
The solution turned blue within 25 s and the final colour change to red-violet 70 s later.

28
Boiling continued for an additional 10 min, the heating source was removed, and the colloidal was stirred for another 15 min.

29
The colloidal dispersions were filtered through 0.2 µm filter to trap any large particles.

30
Particle size was verified by TEM indicated an average diameter of 20 ± 1 nm.

31
Gold nanoparticles in aqueous dispersions were imaged in a Phillips CM200 Transmission Electron Microscope after placing drops of a dispersion onto gold grids, formwar of 400 mesh (SPI, PA, USA), and allowing the liquid to dry in air at room temperature.

32
Confirmation by light scattering (90Plus/BI-MAS multi angle particle sizing option, Brookhaven Instuments Corp, NY) yielded 21 ± 0.4 nm.

33
Derivatisation was performed by adding thiol solution to a stirred colloidal suspension at room temperature.

34
The effect of adding charged thiols on the red-shift in colloidal suspensions was monitored by the absorbance spectrum in the 400–800 nm region with a Shimadzu UV–1603 spectrophotometer in standard quartz cuvettes.

35
The citrate stabilised particles were further derivatised with 3-mercapto-1-propane sulfonate (MPS, negatively charged end) and 2-mercapto ethyl amine hydrochloride (MEA, positively charged end) respectively and the progress of the thiol adsorption was followed in a UV-visible spectrophotometer.

36
Post derivatisation of the citrate stabilized nanoparticles was performed by successive steps, each one consisting of small volume additions of 0.4 mM thiol solution (volume ratio 1/300), to a stirred colloidal suspension and recording the UV–vis stationary spectra.

37
In the case of MPS the reaction was so slow that it was necessary to wait ca. 20 min to satisfy this requirement.

38
In situ thiolation during particle formation by the citrate reducing agent was accomplished by the method described by Kunitake et al.27.

Results and discussion

39
It is well known that stability of aqueous colloidal suspensions of alkanethiols derivatised gold nanoparticles is strongly dependent on the way of synthesis.

40
In fact, further addition of thiols to stable suspensions usually leads to instability and the kinetic evolution of the resulting system also depends on the way the addition is carried out.

41
Yoneyama and Kunitake27 have shown that gold nanoparticles are stabilized by alkanethiols if added to citrate AuCl4 solution during the synthesis and the particle size can be controlled by thiol concentration/gold ratio.

42
In Fig. 1 we compare the uv-visible absorbance spectra for freshly prepared aqueous dispersions of mercapto propane sulfonate (MPS) stabilized gold nanoparticles (a) prepared by reduction of AuCl4 in solutions containing both citrate and the thiol of different MPS to total Au concentration ratio (b–d).

43
A single absorption peak at 530 nm is observed with absorbance decreasing at increasing thiol concentration.

44
TEM analysis has shown individual particles with no aggregation apparent, and size decreasing at increasing thiol concentration.

45
If the same thiols are added once the citrate stabilized Au nanoparticles have been formed, aggregation and further floculation occurs while a red shift in the absorbance spectrum is observed.

46
Whitesides26 has reported changes in floculation of Au nanoparticle dispersions upon chemisorption of alkane-thiols bearing functional groups by observation of the spectral changes (red shifts).

47
Fig. 2 shows the absorption spectra of alkanethiol stabilized citrate Au nanoparticles after addition of different thiols.

48
It should be noticed that the spectra depends strongly on the alkane chain length and the nature of the functionalized end group.

49
Two absorption peaks are apparent at 530–535 nm and 630–660 nm respectively.

50
For long enough alkane thiol chains such as MUA (mercaptoundecanoic acid) only the 535 nm absorption band is observed (Similar results have been shown by Mulvani et al.11,19. with SiO2 coated Au nanoparticles) and a stable suspension is obtained (Fig. 2b).

51
For short alkanethiols as MPS (Fig. 2c) and MEA (Fig. 2e), on the other hand, kinetic effects are observed and the spectra also depend on the time elapsed after thiol addition.

52
The different behaviour should be related to the replacement of adsorbed citrate anions on the nanoparticle by thiol upon addition of thiol solution.

53
After addition of thiol solution to stabilized citrate Au nanoparticles a decrease of relative absorbance at 530 nm ΔAr = [A(τ) – A(∞)]/[A(0) – A(∞)] is observed.

54
This is a consequence of a decrease on the individual nanoparticles concentration.

55
An intriguing feature of the thiol adsorption on nanoparticles is the adsorption kinetics: While functionalization of the Au nanoparticle by the positively charged thiol (MEA) is very fast as can be seen in Fig. 3a, for negatively charged MPS on the other hand, the process is quite slow (see Fig. 3b).

56
In these experiments the thiol is in large excess with respect to the concentration required for surface saturation as it is shown later.

57
The absorbance decay is the result of two successive processes: adsorption of thiols and aggregation of derivatised particles.

58
Transmission electron microscopy (TEM) examination of the resulting red-shifted dispersions demonstrates the floculation of the thiol modified nanoparticles as shown in Fig. 4a and 4b for MPS and MEA respectively.

59
Different results are obtained from experiments where the amount of thiol is very small with respect to the concentration required for surface saturation (see below).

60
Figs. 5 and 6 show the evolution of the steady state absorption electronic spectra of Au nanoparticles after addition of increasing amounts of thiol to the particle dispersion for MPS (Fig. 5) and for MEA (Fig. 6).

61
Three relevant features are apparent: (a) the surface plasmon energy is constant; this is due to the fact that the radii of the individual nanoparticles is practically unmodified when the number of Au–S bonds or thiol surface coverage increases, (b) The absorbance at the surface plasmon resonance frequency (λ = 530 nm) decreases with the increase in the amount of the added thiol, this effect indicates a loss of individual nanoparticles density, and c) at higher wavelengths (600–800 nm) a new absorption band appears which increases with the amount of adsorbed thiol.

62
This results in a colour shift at the naked eye from red to blue (by transmitted light) which has been ascribed by several authors to shortening of the particle-to-particle distance upon colloid aggregation.10–21

63
Longer chain alkanethiols also decrease the surface plasmon band absorbance but the long wavelength red shift is absent for mercaptodecane.

64
Figs. 7 and 8 depict plots of the steady state absorption at 530 and 800 nm respectively as a function of the total number of moles of added thiols.

65
For both thiols we observe a similar trend in the absorption dependence on the number of moles of mercaptane added which decreases for 530 nm while increases for 800 nm until a similar final absorption value is reached, ca. 0.7 and 0.8 for MEA and MPS respectively.

66
Extrapolation of the absorption titration curves in Figs. 7 and 8 to the final value results in a similar value of added thiol, ca. 8 nmol.cm–3.

67
This quantity can be correlated with the surface concentration of MPS or MEA adsorbed on the colloidal particles: The concentration of gold nanoparticle suspension is 1.6 × 10–9 molar.

68
This value was obtained considering that the mass of HAuCl4 employed for colloidal suspension preparation was completely transformed into colloidal particles with an average diameter of 20 nm which has been confirmed by STM measurement on HOPG, TEM and light scattering.

69
The area per particle calculated from this diameter, 1.02 × 10–11 cm2, leads to a total surface colloid area of 9.7 cm2 per cubic centimetre of suspension.

70
Under the assumption that the 8 nmol cm–3 involved in the titration process were totally adsorbed, a surface concentration of 8.2 × 10–10 mol cm–2 has been obtained.

71
This value is coincident with the coverage reported for a totally covered Au surface by alkanethiol, c.a. 10–9 mol cm–2.28

72
This evidence strongly suggests that the final value obtained from the extrapolation in Figs. 7 and 8 is equivalent to a stoichometric point of a titration of the colloid surface.

73
When derivatisation is completed the surface of gold nanoparticles are covered by a monolayer of thiol adsorbed molecules with their sulfonate or amino end-groups respectively pointing towards the solution and as a result of this a well characterised colloid surface is obtained.

74
It is worth noticing that in either case the negatively charged Au nanoparticles have been functionalized with positively or negatively end charged short chain alkanethiols.

75
So far we have discussed the steady state absorbance of mercaptane protected Au clusters.

76
Using the experimental approach described above it is possible now attempt an explanation for the kinetic evolution of their optical spectra either in the case of MPS and MEA.

77
Whitesides has discussed the mechanism for SAM formation on Au nanoparticles, their stability and floculation.26

78
The optical density of non-aggregated particles is expected to be proportional to relative absorbance ΔAr, [A(t) – A(∞)]/[A(0) – A(∞)].

79
If the rds were the aggregation step, dA/dt should be proportional to the square of the non-aggregated particle density, and consequently proportional to ΔA2r at the beginning of the process.29

80
This is observed in the case of MEA: the reciprocal of ΔAr shows a linear dependence with time for about 20 s (Fig. 9).

81
Conversely, in the case of MPS this dependence is far from being observed.

82
As depicted in Fig. 3 for different MPS initial concentrations the time evolution of the absorbance at 530 nm indicates very slow and concentration dependent adsorption kinetics.

83
As expected, the higher the thiol concentration the faster the mercaptane uptake by the gold nanoparticle.

84
It should be noticed that the result observed is not due to an increase of the solution ionic strength since no effect was observed upon addtion of NaCl of the same concentration as the sodium mercaptopropanesulfonate, but a high concentration of salt (ca. 1 M) produces a red-shift of the colloidal solution.

85
The rate of the process is reflected in the shape of the Avs.t dependence at 530 nm.

86
It is noteworthy that the shape of the transients depicted in Fig. 3 are not indicative of simple kinetics.

87
A good curve fit could be obtained with the following empirical equation, as depicted in Fig. 10 (solid lines).where C is the thiol concentration, k an apparent rate constant and b an empirical parameter.

88
Eqn. (1) may be explained if the adsorption process is considered to be the rds step.

89
It has to be taken into account that at the beginning this process is very slow due to the electrostatic repulsion of a negatively charged thiol with a negatively charged colloid30 and also that a subsequent acceleration of the reaction is observed when the thiol is adsorbed.

90
A suitable expression to account for this experimental condition is: where S is the fraction of free surface of the colloidal particles.

91
To a first approximation, the adsorption rate is expected to be proportional to S and C.

92
The reaction order ½ on thiol concentration is assumed for a dissociative adsorption isotherm or for the involvement of two binding sites per thiol molecule.

93
The reaction:HSRSO3 + 2Au – X = AuHads + AuSRSO3 ads + 2Xwhere X could be the chloride or the citrate anion30 accounts for the mercaptane surface modification and is consistent with the reaction order and the change in the surface negative charge.

94
Brust and co-workers have provided NMR evidence for the adsorption of RSH or adsorbed hydrogen with the thiol on the Au nanoparticles.31

95
The term (1 – S) is introduced in the kinetic law in order to take into account that an increment of thiol surface concentration facilitates the adsorption process; this is similar to the effect observed for autocatalytic reactions and would indicate that the reaction proceeds via a thiol associative mechanism.32

96
The exponential term in eqn. (2) corresponds to the contribution of the electrostatic work to the free energy of activation due to a decrease in the colloidal particles charge surface density as a result of the adsorption process (exchange of surface anions by the chemisorbed thiol).

97
The electrostatic energy involved in bringing a charged thiol ion from infinite separation to the equilibrium distance at the colloidal particle surface (activated complex), r, is given by: where a is the area of a colloidal particle, e the electron charge, ε0 the electrical permittivity of free space, ε the static dielectric constant and σ is the surface density of charge of a colloidal particle.

98
σ may be written in terms of the relative uncovered surface, S, as: σ = σ0S + σ1(1 – S) = (σ0 – σ1)S + σ1where σ0 and σ1 are the surface charge densities corresponding to S = 1 and S = 0 respectively.

99
Then eqn. (4) may be rewriting as: we = α + βSwhereFinally the adsorption rate constant is given by:k′ = k0exp(–we/RT) = k′0exp(–βS)Titration curves in Figs. 7 and 8 show that ΔAr is linearly dependent on the number of adsorbed moles, n, before the titration end point, n0.

100
This may be indicated as:and considering that n is proportional to the coverage, i.e. (1 – S), it is easily found that ΔAr = S.

101
Replacing this ratio in eqn. (2), we find the empirical eqn. (1).

Conclusions

102
The increase in the high wavelength absorption of functionalized nanoparticles has been related to the decrease in the particle to particle distance, this has been further proved by monolayer protection of Au nanoparticles with 6-mercaptohexanol (MH) and 11-mercaptoundecanoic acid (MUA).

103
In the present case a shorter nanoparticle average separation is consistent with a decrease in the charge surface density upon thiol binding to the Au nanoparticle.

104
While the change in the surface plasmon absorption intensity is given by the interaction of each individual Au nanoparticle with the electromagnetic radiation, the red shift corresponds to a cooperative association of colloidal particles.

105
For short length particle-to-particle spacers such as short alkanethiols the coupled dipole interacting with the light results in a red shift while for longer spacers such as MUA the dipolar interaction is very weak or nonexistent and only the plasmon absorbance band is observed.

106
For intermediate thiol lengths, such as for mercaptohexanol, two distinct absorbance bands are apparent at 530 and 670 nm.

107
Both short chain alkanethiols, irrespective of their end charge, lead to a red shift and a decrease in the plasmon band intensity of the citrate-formed Au nanoparticles.

108
The end charge, however, determines the kinetics of the whole derivatisation process, i.e. adsorption–aggregation, because of the electrostatic interactions with the negatively charged citrate formed Au nanoparticles.

109
In the case of MPS, our results show that the kinetics are determined by the adsorption step.

110
Conversely in the MEA case the adsorption step is so fast due to electrostatic attraction and the rate determining step is the aggregation step.

111
With the present experimental evidence it is not possible to exclude that the aggregation process following the adsorption step does not influence the adsorption kinetics by either positive or negative feedback.

112
A quantitative analysis of the absorption time dependence, based on a repulsive interaction model, is consistent with the experimental kinetic evidence.

113
It should be noticed that the present analysis is only valid if the adsorption kinetics is rate determining step and if the system aggregates.

114
Examples of thiol-modified oligonucleotides, peptides, aminodextrans, etc. that give stable colloidal dispersions have been reported and it is not clear at present why aggregation occurs in some cases and not in others33.