Thiol-stabilized Gold Nanoparticles: new ways to displace thiol-layers using Yttrium or Lanthanide Chlorides

. We use the aurophilic interactions shown by lanthanides to overcome the sulfur-gold interaction. UV-Vis and XPS spectroscopy confirm that Yttrium or lanthanide chlorides easily displace sulfur ligands from the surface of thiol-stabilized gold nanoparticles.


INTRODUCTION
The modification of gold surfaces with thiol derivatives has been extensively explored 1,2,3,4 in the past two decades since the first report in 1993 by Mulvaney and Giersig 5 and the very wellknown Brust-Schiffrin bi-phasic method developed in 1994. 6 The thiolate bond resulting from sulfur-gold interaction has proven even stronger than the gold-gold metallic bond 7 and therefore the modification of gold surfaces, particularly gold nanoparticles (AuNPs), with thiolate ligands is considered the most stable passivation method for AuNPs. 8,9,10 The exchange of ligands at the AuNPs surface has been studied as many applications need to recover either the ligand, 11 the AuNP or both. 12,13 Aurophilic interactions between the surface of gold nanoclusters (AuNCs) and lanthanides, namely Ln 3+ -Au + , have recently been reported. 14 However, the relative strength of the Ln 3+ -Au + interaction is still unknown, especially compared to the well-known stability of the sulfur-gold bond. Here, we use lanthanides as a new chemical strategy for sulfur ligand displacement from thiol-stabilized AuNPs.
Nowadays, the use of lanthanides is mainly related to the formation of up-conversion nanoparticles for bioanalytical applications. 15,16 The inner electronic configuration of these elements is suitable for electronic excitation that gives rise to the photon-upconversion process.
It is known that some lanthanides can form complexes with carbonyl groups, and to the best of our knowledge no interactions with sulphur moieties have been reported. 17,18,19,20 Herein we present our efforts to demonstrate that thiol-capped AuNPs can undergo ligand release in the presence of a RE salt. UV-vis and XPS analysis prove that the sulfur material is released from the surface of the Au nanostructure upon addition of RE salts, possibly through the mechanism proposed in Scheme 1. Scheme 1. Thiol-capped Au nanostructures undergo ligand release in the presence of rare earth ions (RE n+ ) inducing particle agglomeration and possible concomitant Au-Au bond rupture.

EXPERIMENTAL SECTION
Synthesis of AuNPs@thiol. Thiol-capped AuNPs were prepared by thiol ligand exchange of the precursor AuNPs capped with citrate. Thus, citrate-capped AuNPs (AuNPs@C) were synthesized from HAuCl 4 according to a reported procedure 21  Ionic strength effect. In order to rule out the effect of the ionic strength change, the AuNPs@thiols (or AuNPs@citrate) were mixed with NaCl solutions of different concentrations reaching the same or higher ionic strength than in the case of RE salts. The ionic strength was

RESULTS AND DISCUSSION
As is well known, thiol compounds have strong affinity for Au surfaces due to the formation of a strong Au−S bond; 7, 21 however, we show here that the Au surface thiol passivation can be defeated in the presence of RE ions. In order to test this hypothesis, three thiols carrying different moieties were used to passivate Au nanostructures: 2-mercaptoethanol (S-ETH), 7-mercapto-4methylcoumarin (S-COU) and 11-mercaptoundecanoic acid (S-MUA) (Scheme 2). The diverse chemical properties of the thiols selected would help to show the extent of this methodology.

Scheme 2. Thiols used in this work.
Thiol-capped gold nanomaterials were prepared in our lab as described in the experimental section. Briefly, thiol-aged gold NPs were produced by overnight incubation of citrate-stabilized AuNPs with the corresponding thiol in order to ensure complete ligand exchange (i.e. citrate replaced by thiol) maintaining the stability of the colloid. Residual RSH were removed by several centrifugation and washing cycles. Figure S1 shows the UV−vis spectra of the three AuNPs@thiolate studied here. Plasmon resonance band at maximum around 528 nm and 530 nm are identified for AuNPs@S-COU and AuNPs@S-MUA, respectively. AuNPs@S-ETH presents two plasmon resonance bands around 530 nm and 693 nm, most likely because of the change of interspacing distance between AuNPs due to the small dimensions of this ligand. 22 Figure 1B-D), probably due to lower surface accessibility associated with the length of the ligand chain 23 which hinders the particle surface from interacting with incoming ligands. Also, the ligand carbonyl groups near the surface could complex with lanthanides as previously demonstrated. 17,18 These phenomena can also be followed by the changes on the solution color, gradually shifting from red to blue (Figure 1 insets). Additionally, the AuNPs@C, used as the precursor for the other thiolated-AuNPs, were also mixed with lanthanides salts as a control experiment. Interestingly, citrate coated nanoparticles are very stable upon addition of RE salts, concentrations around 10 times higher than those needed in order to cause similar changes in the UV-Vis spectra (Figure 2, Figure S6 and Table 1). Interaction between citrate and RE salts are known 24     In order to confirm that the drastic change on the plasmon resonance band of the thiol-capped AuNPs is due to the displacement of thiol-ligands, we analysed the resulting NPs by High-Resolution X-ray Photon Spectroscopy (HR-XPS). First we analyse the Au 4f region, in order to determine the oxidation state of the Au species. Figure 3 shows that Au 4f signals shift towards binding energies (BE) that are higher (more oxidized Au species) when AuNPs surface is surrounded by Y or S-ETH (See tables 2 and S1) rather than citrate. Thus, increasing BE are expected for capping agents ranging from citrate to thiols to RE ions. When the particles are treated with YCl 3 (similar results where found with YbCl 3 and TmCl 3 ), they are subjected to centrifugation and several washes before XPS analysis. Figure 4 shows the Au 4f HR XPS spectra obtained before and after addition of the RE salts, for the resulting pellet (AuNPs precipitate) and supernatant. As expected for Au-RE interactions, the Au signals corresponding to the AuNPs precipitate shift toward higher BE. When AuNPs@thiols are treated with YbCl 3 , Au species can be also detected in the supernatant. Interestingly enough, these species show higher oxidation than the Au species found in the pellet, suggesting that these Au species are different from those found in the AuNPs. This could support the mechanism described in Scheme 1 where Au atoms are also displaced from the AuNPs surface. When YCl 3 or TmCl 3 are used, gold species are not found in the supernatant, although the presence of trace amounts cannot be ruled out. The characteristic S 2p signals (S 2p3/2 about 162.05 eV) displayed for each thiol-stabilized AuNPs are in agreement with gold-thiolate bonding ( Figure S9). 26,27,28,29 In the case of AuNPs@S-ETH and AuNPs@MUA, BE signals above 164 eV account for the presence of more oxidized sulfur compounds (not seen in the case of AuNPs@COU; Figure S9B). When treated with lanthanide ions, the thiol-stabilized AuNPs release more reduced sulfur species. Figure 5 shows the HR XPS spectra in the S 2p region found for AuNPs@S-ETH upon treatment with Y, Yb or Tm chlorides. Although sulfur species can still be found in the AuNPs precipitate, there is no doubt about the presence of a large amount of released sulfur species in the supernatant. Interestingly enough, the S 2p signals found in the latter correspond to more reduced sulfur species as can be inferred from the more intense peak found around 164 eV. Evidence for the presence of Yttrium in both the pellet and the supernatant is obtained by looking at the characteristic Y 3d BE signals between 150-162 eV (notice that their coincides with the S 2p region, Figure 5A). 30,31 The same was also found for Yb and Tm (Table S1). Table 2 summarizes the XPS results obtained for each particle after addition of YCl 3 (See Table S1 for YbCl 3 and TmCl 3 ). It is important to note that the presence of sulfur compounds in the supernatant was confirmed in the cases of S-ETH and S-COU, but cannot be ruled out when S-MUA was used, more likely due to a lower amount of displaced ligand. This is in agreement with the lower interaction seen by UV-vis spectroscopy when AuNPs@SMUA are used.    Binding energy (eV) for the principal elements in eV; referenced to aliphatic carbon (C 1s 284.8 eV). P: pellet. SN: supernatant. ND: No detected.

CONCLUSIONS
We were able to demonstrate that thiol-capped AuNPs can experience ligand displacement in the presence of rare earth salts, such as YCl 3 , YbCl 3 or TmCl 3 . To the best of our knowledge, this is the first attempt using lanthanide or Yttrium salts to displace thiols from gold surfaces. Although the detailed nature of this phenomenon still remains unclear, we believe these findings create many opportunities in the field of bioimaging and bioanalytical applications in which RE complexes are very well known.
Further, the optical changes resulting from the ligand displacement show a notable spectral threshold that may also find applications in the analytical detection of lanthanides. The length of the ligand as well as its nature affect the efficiency of the displacement, a more detailed analysis and quantification are under study now in our lab.