Cooperative Gold Nanoparticle Stabilization by Acetylenic Phosphaalkenes

Acetylenic phosphaalkenes (APAs) are used as a novel type of ligands for the stabilization of gold nanoparticles (AuNP). As demonstrated by a variety of experimental and analytical methods, both structural features of the APA, that is, the P=C as well as the C≡C units are essential for NP stabilization. The presence of intact APAs on the AuNP is demonstrated by surface-enhanced Raman spectroscopy (SERS), and first principle calculations indicate that bonding occurs most likely at defect sites on the Au surface. AuNP-bound APAs are in chemical equilibrium with free APAs in solution, leading to a dynamic behavior that can be explored for facile place-exchange reactions with other types of anchor groups such as thiols or more weakly binding phosphine ligands.


General Procedure Gold Nanoparticle Synthesis
All glassware and stirring bars are thoroughly clean using a aqua regia, rinsed with deionized water and dried at 130 °C for 24 h. Solvents are freshly distilled used immediately. The ligand (0.03 mmol) and HAuCl 4 *3(H 2 O) are dissolved in 5-10 ml THF. The tetrachloro auric acid is added to the ligand solution and stirred in the dark for the given mixing time. Subsequently, reducing agent (e.g. triethyl silane) is added over a period of 30 sec. and stirred for 1-24 h. The reaction mixture is filtered through normal filter paper and the filtrate is collected. All volatiles can be removed at slightly elevated temperatures (35 °C) under vacuum and the residue is washed with cold methanol. Solutions can also directly be used for NMR investigations or TEM imaging. Crystallographic data sets were collected from single crystal samples mounted on a loop fiber and coated with N-paratone oil (Hampton Research). The collection was performed using a Bruker SMART APEX diffractometer equipped with an APEXII CCD detector, a graphite monochromator and a 3-circles goniometer. The crystal-to-detector distance was 5.0 cm, and the data collection was carried out in 512 x 512 pixel mode. Cell refinement and data reduction were performed with SAINT V7.68A (Bruker AXS) on the final data. Absorption correction was done by multi-scan methods using SADABS96 (Sheldrick). The structure was solved by direct methods and refined using SHELXL97. All non-H atoms were refined by full-matrix least-squares with anisotropic displacement parameters while hydrogen atoms were placed in idealized positions.

Raman spectroscopy:
Measurements were performed on a performed on a Renishaw micro-Raman spectrometer (Modell 1000) equipped with a 514 nm Ar laser and a 1200 lines/mm grating. The spectra are reference with a Si-reference sample. The free ligands were recorded as solid samples on an amorphous glass substrate with exposure times of 10 seconds, a laser power of 25% and accumulation of 2 scans. Longer exposure times, more scans or higher laser power were precluded due to sample decomposition. The spectra were baseline corrected with a 5 th order polynomial. The AuNPs are measured as solid samples after extensive purification as described in the synthetic procedures. In total 3 accumulated scans à 10 seconds with a laser power of 100% compose the final spectra. The spectra were processed with the WIRE2 smooth (version 2.  basis set for electrons in the molecule and a single-zeta plus polarization orbitals (SZP) basis set for electrons in the gold particle. The GGA of Perdew-Burke-Ernzerhof was used for the exchange-correlation functional. 6 The reciprocal cell was sampled with a 5×5×3 k-mesh. We have additionally considered a more densely packed film of molecules either in a laterally smaller cell or with two molecules per cell, but the additional stabilization of the system by intermolecular interactions is quite similar for all molecules and does not explain experimentally observed behavior for different compounds. In these calculations we heave also checked convergence of our calculations with respect to the slab thickness and basis size (thicker gold slab and DZP basis for surface gold atoms were tested) and found it satisfactory.
Here we started by modeling the AuNPs by a flat gold (111)-surface resulting in a Au-P bond distance of more than 4 Å for the different calculations indicating a very weakly bond system. Adsorption energies are slightly negative indicating that the systems are not adsorbed at all, but theses systems should be weakly physisorbed and the dispersive forces are not fully captured in our approach.
The AuNPs can be expected to contain a number of defect sites of the perfect Au(111) surfaces. To model these defects, we consider a gold ad atom on the previous used surface. In these cases it is possible for the phosphorus lone pairs to coordinate with the defect gold atom, forming a stronger bond. We calculate an Au-P bond distance of about 2.35 Å, more in agreement with a chemisorbed (coordinated) molecule. In agreement, we also get adsorption energies of about 1 eV.
To elucidate the difference in the adsorption mechanism, we study Crystal Orbital Hamilton population (COHP). COHP is a way to quantify the density of states (PDOS) from the DFTcalculations into bonding, anti bonding and non-bonding states, and opens the possibility to calculate the bond strength. 7 By integrating the COHP-curves up to the Fermi level (E F ) it is possible to obtain the bond energy for the given interaction (or actually the change to the band-structure energy due to this interaction). Positive peaks in the COHP corresponds to anti-bonding and negative to bonding interactions. 3  Interaction between the P atom and the underlying gold is more complicated, as shown on Fig. 3. Here the P-atom's interaction is bonding to the underlying flat gold surface. Adding the extra gold makes phosphor interaction anti-bonding to the bulk gold.
The interaction between the P atom and the extra defect on gold -ad-atom demonstrate a significantly stronger interaction (factor 10 scaling is applied to fit into the same graph) of about -3.86 kcal/mol. Representative NMR and UV data of selected samples: Comparison of purified and crude AuNPs. The crude AuNPs solutions were gently dried vacuum and redissolved in C 6 D 6 . Purification of the AuNPs after removal of all volatiles and subsequent extensive washing cold methanol. The residue was redissolved in C 6 D 6 .
In the crude samples the 31 P and 1 H-NMR data suggest he presence of intact phosphaalkene species in solution, i.e. not bound to the gold surface. After purification almost all dissolved ligands were removed, however no additional species were detected that could be assigned to surface bound phosphaalkenes.
UV/data of AuNPs before and after purification indicating no changes of the AuNPs during the purification procedure NMR data for the previsouly non reported gold complexes and [AuCl*1]. The 31 P-NMR resonance shifts to higher fields as compared to the uncomplexed ligand (i.e. δ 31 P (1) = 270 ppm) typical for π-backbonding metals. For the broadening of the 31 P signal we could not find a rational explanation.
NMR data of previously non-reported gold complex [AuCl*4]: 13 C-NMR data could not be recorded due to broadening of the lines, which is in line with the low intensity and the significant broadening of the resonance observed in 31 P-NMR.