Directing Intrinsic Chirality in Gold Nanoclusters: Preferential Formation of Stable Enantiopure Clusters in High Yield and Experimentally Unveiling the “Super” Chirality of Au144

Chiral gold nanoclusters offer significant potential for exploring chirality at a fundamental level and for exploiting their applications in sensing and catalysis. However, their widespread use is impeded by low yields in synthesis, tedious separation procedures of their enantiomeric forms, and limited thermal stability. In this study, we investigated the direct synthesis of enantiopure chiral nanoclusters using the chiral ligand 2-MeBuSH in the fabrication of Au25, Au38, and Au144 nanoclusters. Notably, this approach leads to the unexpected formation of intrinsically chiral clusters with high yields for chiral Au38 and Au144 nanoclusters. Experimental evaluation of chiral activity by circular dichroism (CD) spectroscopy corroborates previous theoretical calculations, highlighting the stronger CD signal exhibited by Au144 compared to Au38 or Au25. Furthermore, the formation of a single enantiomeric form is experimentally confirmed by comparing it with intrinsically chiral Au38(2-PET)24 (2-PET: 2-phenylethanethiol) and is supported theoretically for both Au38 and Au144. Moreover, the prepared chiral clusters show stability against diastereoisomerization, up to temperatures of 80 °C. Thus, our findings not only demonstrate the selective preparation of enantiopure, intrinsically chiral, and highly stable thiolate-protected Au nanoclusters through careful ligand design but also support the predicted “super” chirality in the Au144 cluster, encompassing hierarchical chirality in ligands, staple configuration, and core structure.

h, the white precipitate was removed and the solution was washed with diluted HCl and water.After extraction of the aqueous phase with DCM, the combined organic phases were dried over Na2SO4.Removal of DCM by rotary evaporation yielded a clear oil, which was purified by silica column chromatography in 1:3 hexane:EtOAc (EtOAc = ethylacetate).The intermediate (14.50 g, 59.9 mmol) and 4.56 g (59.9 mmol) of thiourea were dissolved in 85 ml of ethanol (EtOH).After refluxing at 80 °C for 72 h, 60 ml of 20% NaOH were added and the reaction mixture kept at 80 °C for another 60 min.The solution was subsequently cooled to room temperature and acidified with 100 ml of 10% HCl.The organic phase was extracted with hexane and dried over Na2SO4.Removal of the solvent and impurities was achieved by distillation at 140 °C.The product, which was obtained as a clear oil, was characterized by nuclear magnetic resonance spectroscopy (NMR; see Figures S2-S3).

Au38(2-MeBuS)24
Au38(2-MeBuS)24 was synthesized by adapting the protocol reported by Stellwagen and co-workers. 30 mg (0.15 mmol) HAuCl4•3H2O and 155 mg (0.5 mmol) of L-glutathione (GSH) were dissolved in 8 mL MeOH and 3.5 mL H2O, yielding a white suspension.After cooling the mixture to 0 °C, 47 mg (1.2 mmol) NaBH4 suspended in 2.4 mL ice-cold water were added, which resulted in the formation of a black precipitate.The reaction was stirred at 0 °C for 1 h.After separating the black precipitate by centrifugation, the solid was dissolved in 2.4 mL H2O, 1.5 mL acetone and 2 mL (15.6 mmol) 2-MeBuSH.The reaction was continued at 80 °C for 16 h.The phases were separated and the aqueous phase washed with DCM.The organic phase was dried and the resulting black precipitate washed with ethanol for purification (yield with respect to HAuCl4•3H2O was 78%).

Au38(2-PET)24
Au38(2-PET)24 was prepared by following the procedure published by Pollitt and co-workers. 4First, 1 g (2.9 mmol) HAuCl4•3H2O and 3.17 g (10.3 mmol) GSH were dissolved in 100 ml acetone, giving a yellow suspension, which was stirred at 0 °C for 30 min.Next, an ice-cold solution of 30 ml H2O and 0.98 g (25.9 mmol) NaBH4 was poured in carefully, resulting in the formation of a black precipitate.After decanting the solvent and drying the solid, 6 ml EtOH, 10 ml toluene, 30 ml water and 10 ml (74.7 mmol) 2-PET were added.The mixture was stirred at 80 °C for 4 h and subsequently cooled to room temperature.50 ml hexane were added and the black precipitate removed by filtration.After washing several times with MeOH, the crude product was redissolved in DCM and dried by rotary evaporation at 30 °C.For purification of the crude product, SEC (THF, Bio-Beads SX-1 support) was performed (yield of final product was 6% with respect to HAuCl4•3H2O).

Au38(S-Bu)24
The synthesis was performed in an analogous fashion to the procedure presented in Section 1.3.3, with the difference that 20 ml (185.6 mmol) of SH-Bu were added instead of the 2-MeBuSH.Furthermore, after stirring at 80 °C for 16 h, the cluster was precipitated with 88 ml of a 1:10 H2O:MeOH solution.The black/violet solid was then filtered, washed with EtOH, and subsequently extracted with toluene and dried.Yield with respect to HAuCl4•3H2O was 26%.

Au144(2-MeBuS)60
Au144(2-MeBuS)60 nanoclusters were prepared by modifying the procedure presented by Qian et al. 5 236 mg (0.6 mmol) HAuCl4•3H2O were mixed with 380 mg (0.7 mmol) TOAB and dissolved in 30 mL MeOH.The red solution was stirred for 15 minutes at room temperature, after which 394 µL (3.2 mmol) 2-MeBuSH were added, giving a white suspension.After stirring for 15 min at room temperature, the polymer suspension was reduced using a cooled solution of 227 mg (6 mmol) NaBH4 dissolved in 12 mL water, yielding a black precipitate.The black solution was stirred for another 5 h at room temperature.Subsequently, the black precipitate was separated by centrifugation and washed several times with methanol.The crude product was purified by SEC (THF, Bio-Beads S-X1 support).Au144(2-MeBuS)60 eluted as the first fraction (black), followed by another black fraction (identified as Au38(2-MeBuS)24) and Au25(2-MeBuS)18 in anionic (reddish-brown) and neutral (greenish) charge state.The yield of Au144(2-MeBuS)60 with respect to HAuCl4•3H2O was 73%.

Au144(S-Bu)60
The synthesis protocol was analogous to the one described in Section 1.3.6,except for the addition of 343 µl (3.2 mmol) HS-Bu instead of 2-MeBuSH.The yield of Au144(S-Bu)60 with respect to HAuCl4•3H2O was 38%.

NMR Spectra of the 2-MeBuSH Ligand
Nuclear magnetic resonance (NMR) spectroscopy was measured on a Bruker Avance 400 MHz NMR spectrometer.Samples were dissolved in CDCl3 and the solvent signal was used as internal reference.Chemical shifts relative to trimethylsilane (TMS) are reported.

Model Structure and Properties of Au25 and Au38 Clusters
For the DFT calculations of Au25 and Au38 clusters, the electronic structure calculations were carried out using the Amsterdam Modeling Suite (AMS) 2021.1 package. 6The initial structures before optimization were based on previously reported structures [7][8][9] and the ligands were replaced with a gasphase optimized structure of 2-MeBuSH.1] A double zeta polarized basis set (DZP) was used. 124] The gradient convergence was set to 10 -3 and the energy convergence to 10 -4 .For Au38, after a first optimization, further isomers of each structure were created by either editing the ligand positions using the MacMolPlt 15 software or by sampling structures from classical molecular dynamics simulations (see details in section 6.3).Geometry optimizations were performed on all newly created structures.Optical absorption and circular dichroism (CD) spectra were obtained after a linear response TD-DFT+TB 16 calculation and subsequent convolution of the excitation energies into a spectrum by applying a Gaussian fit with a full width half-maximum (FWHM) of 30 nm (15 nm only for the pure (S)-MeBuSH ligand in Figure S28).TD-DFT+TB is an approximate TD-DFT method that builds a tightbinding-like excited state calculation on a standard DFT ground state calculation; this method does not require tight-binding parameters, but greatly reduces the required computational time.All TD-DFT+TB calculations were performed at the BP86/DZP level of theory.The CD spectrum of the (S)-2-MeBuSH ligand was obtained by employing a Gaussian fit with a FWHM of 15 nm.8] The calculated excitation energies were converted to wavelength units (nm) prior to fitting in order to compare with the experimental spectra.To confirm that TD-DFT+TB was applicable for the calculations of the nanoclusters in question, some spectra were also simulated employing time dependent-density functional theory (TD-DFT) (see Figure S13). 19

Model structure and properties of Au144(2-MeBuS)60 cluster
The electronic, optical, and chiral properties of the Au144(2-MeBuS)60 cluster were calculated using DFT as implemented in the software GPAW. 20GPAW uses real-space grids and scalar-relativistic corrections for metal atom setups.As a starting structure for modeling, we used the crystal structure of Au144(SCH2Ph)60 cluster 21 and replaced the original SCH2Ph with 2-MeBuS ligands.We selected the right-handed enantiomer of the cluster and the corresponding chirality of the ligand.The symmetry of the cluster structure regarding the metal core, metal-ligand interface, and the bonding directions of the ligands within the protecting SR-Au-SR units was fixed during replacement.At first, the symmetrical model structure was optimized using the Perdew-Burke-Ernzerhof (PBE) xc-functional 22 and a 0.2 Å grid spacing.Optimization was continued until the maximum forces acting on atoms were below 0.05 eV/Å.The electronic structure was analyzed by projecting the density of states to spherical harmonic functions centered at the center of the mass of the cluster with a cutoff radius of 15.0 Å. 23 The results of the analysis were shown together with analyzing the origin of the chirality.Optical absorption and CD spectra were calculated using linear response time-dependent DFT and PBE as a kernel. 24Spectra were calculated using a 0.3 Å grid spacing for better computational efficiency.6] The analysis was done just in one direction, along the main principal axis of moments of inertia of the cluster, which is adequate for this nearly spherical system.

Molecular dynamics simulations and essential dynamics analysis
1] For all MD simulations, periodic boundary conditions and the minimum image convention were employed.The 'v-rescale' thermostat was used to control the temperature with a time constant of 0.1 ps. 32For the NPT simulations, isotropic pressure coupling with a reference pressure of 1 bar was achieved using the Berendsen barostat with a time constant of 1.0 ps. 33Prior to the production MD simulations, the energy of the solvated system was minimized using the steepest descent method followed by a short equilibration consisting of 10 ns NVT (constant number of particles, volume, and temperature) at 200 K followed by 10 ns NPT (constant number of particles, pressure, and temperature) at 298.15 K and 1 bar pressure using the v-rescale thermostat and Berendsen barostat. 33During the equilibrations, the heavy atoms of the nanocluster were positionrestrained using a harmonic potential with a force constant of 1000 kJ/mol/nm 2 to allow the solvent to relax around the nanocluster.For improved performance, bonds involving hydrogens were constrained using the LINear Constraint Solver (LINCS) algorithm. 34The use of bond constraints allowed for a 2-fs time step to be used for the integration of the equations of motion, which was performed using the Leap-Frog 35 algorithm.The PME technique 36 was used to calculate electrostatic interactions with a cutoff distance of 1.0 nm for the real space contributions, cubic interpolation, a maximum fast Fourier transform grid spacing of 0.12 nm for the reciprocal space sum, and tinfoil boundary conditions.Lennard-Jones interactions were cut-off at 1.0 nm and long-range dispersion corrections were applied for the energy and pressure.The Verlet cut-off scheme 37 was used, with the allowed energy error due to the Verlet buffer set to the Gromacs default of 0.005 kJ/mol/ps/atom.For the Au38 cluster, we chose two lower energy isomers as the starting structure to perform classical molecular dynamics (MD) simulations, and GAFF 38 was applied for dichloromethane molecules.The gold thiolate nanocluster system was simulated in a cubic simulation box (side length = 10 nm) with 8234 dichloromethane solvent molecules.Production simulations were carried out for a total of 10 ns with the NPT ensemble at 298.15 K and 1 bar, continuing from the equilibrated structures and velocities.Structures were extracted every 1 ns from both trajectories.
For the Au144 cluster, the DFT-optimized model of the Au144(2-MeBuS)60 cluster was solvated in a periodic cubic box of methanol (997 nm 3 ).300 ns of production MD was carried out by keeping the temperature at 300 K with the velocity-rescale thermostat 39 and pressure at 1 bar using Parinello-Rahman barostat 40 with a period of 2.0 ps.2] ED analysis represents the principal motion directions of the system.By using a covariance matrix constructed from atomic coordinates, the most important elements of the position fluctuations in the MD trajectory are extracted (eigenvalue-eigenvector decomposition).Thus, we used this analysis to find out the most relevant structural conformations of the cluster and their basins seen during the MD simulation.We selected one representative snapshot structure (MD-frame 3555) around the observed minimum regions of the free energy landscape and compared its properties with the symmetrical model structure.Therefore, we repeated the calculations of the optical and CD spectra for the optimized snapshot structure the same way as for the symmetrical model structure.In addition to ED analysis, we studied the behavior of the radius of gyration (Rg) during the MD simulation using VMD software 43 and analyzed the flipping of the ligands within the protecting units.Ligand conformations were analyzed using the dihedral angle between the vectors defined along the S-C bond at both ends of each unit.In practice, 0 degrees means ligands pointing to the same side of the unit, while 100 and -100 degrees mean ligands pointing to different sides of the unit Possible angles are restricted by spatial effects on the cluster surface which makes the angle range deviate from the ideal from +180 to -180 degrees.

Additional Information and Results for the Calculations of Au38(2-MeBuS)24
In the following section, isomer 1 corresponds to structures obtained after modification of the experimental Au38 crystal structure, 7 whereas isomer 2 was obtained starting from the lowest energy structure found by Lopez-Acevedo et al. (denoted as JACS2010 structure). 8The appendix a corresponds to the anti-clockwise and the appendix b to the clockwise enantiomer.Figure S11 shows a comparison of the two different anti-clockwise isomers 1a and 2a.For isomer 1a, the ligands in the monomeric staple units (see side view in (a)) are mostly facing outwards, whereas the ones of isomer 2a are also a slightly tilted up-and downwards, respectively.This is due to the different orientation of the -S(R)-Au-S(R)-units, which restricts the orientation of the hydrocarbon framework of the 2-MeBuSH ligand in isomer 1a.Furthermore, the top view of the nine ligands in the upper dimeric staple units show that isomer 2a has a very symmetric arrangement of these nine ligands, with each subset (i.e. the three top, middle and bottom ligands in line of sight) mostly following the idealized D3 symmetry of the cluster.For isomer 1a, the same nine ligands are arranged in a much less symmetric fashion.Whereas the bottom three ligands still take symmetrical positions with respect to each other, the topmost ones as well as the ones in the middle of the staple units do not.This implies that the symmetry of isomer 1a is reduced as compared to 2a, which might be related with higher energy.However, besides symmetry, other structural factors (for example the different arrangement of the monomeric staples and its implication for the orientation of the surrounding ligands) will affect that as well.Table S2.Positions of the features A-H in the measured and in the calculated CD spectra as labeled in the Figure 4.          Table S3.Positions of the features A-H in the measured and in the calculated CD spectra as labelled in the Figure 5.The UV-Vis and CD spectra were also calculated for the intrinsically achiral cluster [Au 25 (2-MeBuS) 18 ] -after optimization at BP86/DZP level of theory.This cluster has an achiral arrangement of its dimeric staple units, thus, its chiral properties are solely due to induction by the chiral (S)-2-MeBuSH ligand (see ligand spectra in Figure S28).As can be seen upon comparing to the experimental spectra (Figure S27), the energies of the optical absorption bands are significantly underestimated by TD-DFT+TB at this level of theory.This has been reported for Au 25 calculated with the BP86 functional before. 44,45For better visualization, the theoretical spectrum was shifted by +0.5 eV and then shows acceptable agreement with the experimental spectrum, especially considering the shape of the bands.However, for the CD spectra (Figure S27b), significant deviations are noticed as well, which cannot be corrected by a shift of the energy axis only.This shows that the current cluster model is not sufficient for comparison to the experiment.Further refinement, for example at a different level of theory, would be required, but lies outside the scope of this publication.

Figure S11 :
Figure S11: Structures of isomer 1a and 2a: (a) side view and (b) top view.Note that in (b), the hydrocarbon framework of all but the 9 ligands in the dimeric staples on top is not shown to allow for better visualization.

Figure S14 :
Figure S14: Experimental UV-Vis spectrum of Au38(2-MeBuS)24 and theoretical optical absorption spectrum including oscillator strength of isomer 2a in gas phase.Note that the experimental UV-Vis spectrum is offset on the intensity axis from the theoretical one to allow for better visualization.

Figure S15 :
Figure S15: Structures of isomer 2a and 2b: (a) side view and (b) top view.Note that in (b), the hydrocarbon framework of all but the 9 ligands in the dimeric staples on top is not shown to allow for better visualization.

Figure S16 :9
Figure S16: Theoretical UV-Vis (a) and CD spectra (b) of isomers 2a and 2b in gas phase.

Figure S18 .
Figure S18.Experimental UV-Vis spectrum of Au144(2-MeBuS)60 cluster (red curve) compared to calculated optical absorption spectrum of the symmetrically built model cluster shown in Figure S17 (black curve).Observed features are labelled from 1-5.

Figure S19 .
Figure S19.Rotatory strength transition contribution map (RTCM) of the lower energy peaks from A to D. Contour plot shows the negative (blue) and the positive (red) contributions to the total rotatory strength of the system as decomposed to transitions between Kohn-Sham states.Projected density of states to spherical harmonics functions centered at the center of mass of the cluster is shown at the lower left panel for occupied states and at the right upper panel for the unoccupied states.Lower right panel shows the calculated CD spectrum with a small arrow labelling the position of the analyzed peak.Position of the peak is also denoted in wavelength units in the corner of each contour plot panel.Analysis is done with respect to the electric field at the direction of the main principal axis of moments of inertia of the system.

Figure S20 .
Figure S20.Rotatory strength transition contribution map (RTCM) of the lower energy peaks from E to H.

Figure S21 .
Figure S21.a) Radius of gyration (Rg) of the Au144(2-MeBuS)60 cluster and b) root mean square deviation (RMSD) of atomic positions as a function of simulation time.

Figure S22 .
Figure S22.Essential dynamics (ED) analysis showing the conformational free energy for the structures seen during the last 200 ns of 300-ns MD simulation in units of kJ/mol.Projections are made with respect to the two most important principal components given by the analysis.The selected representative snapshot structure is labelled by the green circle, whereas the region for the lowest free energy values is shaded with blue.Zero level of conformational free energy is set to zero, and only the variations below 5 kJ/mol are shown.

Figure S23 .
Figure S23.Flipping of ligands within the protecting units from 100 ns to 300 ns during MD simulation determined by the dihedral angle between the S-C bonds at both ends of each protecting unit.Each sub-panel visualizes the conformational fluctuations for one protecting unit.As visualized on top, 0 degrees refers to ligands pointing at the same direction, and 100 and -100 degrees to different direction.

Figure S24 .
Figure S24.Comparison of the calculated a) optical absorption and b) CD spectra of the symmetrical initial model structure (brown line) and of the MD-snapshot structure (red line).

Table S1 :
8elative energies of the lowest energy calculated Au38(2-MeBuS)24 isomers.A=anticlockwise and C=clockwise staple rotation.Isomers 1 are crystal structure7based and isomers 2 were obtained starting from the calculated structure by Lopez-Acevedo et al.8