[Ag9(1,2-BDT)6]3–: How Square-Pyramidal Building Blocks Self-Assemble into the Smallest Silver Nanocluster

The emerging promise of few-atom metal catalysts has driven the need for developing metal nanoclusters (NCs) with ultrasmall core size. However, the preparation of metal NCs with single-digit metallic atoms and atomic precision is a major challenge for materials chemists, particularly for Ag, where the structure of such NCs remains unknown. In this study, we developed a shape-controlled synthesis strategy based on an isomeric dithiol ligand to yield the smallest crystallized Ag NC to date: [Ag9(1,2-BDT)6]3– (1,2-BDT = 1,2-benzenedithiolate). The NC’s crystal structure reveals the self-assembly of two Ag square pyramids through preferential pyramidal vertex sharing of a single metallic Ag atom, while all other Ag atoms are incorporated in a motif with thiolate ligands, resulting in an elongated body-centered Ag9 skeleton. Steric hindrance and arrangement of the dithiolated ligands on the surface favor the formation of an anisotropic shape. Time-dependent density functional theory based calculations reproduce the experimental optical absorption features and identify the molecular orbitals responsible for the electronic transitions. Our findings will open new avenues for the design of novel single-digit metal NCs with directional self-assembled building blocks.


Table of Contents
ESI mass spectra were recorded using a Bruker MicroTOF-II.The single crystals of nanoclusters were dissolved in a sovent mixture of DMF and acetonitrile (HPLC grade) and the solution was electrosprayed at 300 µL/min flow rate in both positive and negative modes.The instrument parameters were maintained as follows: mass range: 100-10000 Da, capillary voltage: 2.5 kV, nebulizer: 0.1 bar, dry gas: 1.2 L/min at 80-100 o C.

Single-crystal X-ray diffraction data
Crystals of [Ag 9 (1,2-BDT) 6 ](TOA) 3 was collected on a Bruker X8 PROSPECTOR APEX2 CCD diffractometer at room temperature using Cu Kα radiation (λ = 1.54178Å).Indexing was performed using APEX3 (Difference Vectors method). 1 Data integration and reduction were performed using SaintPlus 8.34A. 2 Absorption correction was performed by analytical method implemented in SADABS. 3Space group was determined using XPREP implemented in APEX2. 1 Structure was solved using SHELXS-97 (direct methods) and refined using SHELXL-2014 (full-matrix least-squares on F 2 ) contained in WinGX. 4Crystal data and refinement conditions are shown in Table S1.A full list of restraints and constraints is contained within the CIF file.Due to significant thermal motion and disorder, a set of restraints and constraints was applied to make both geometry and thermal parameters of alkyl chains reasonable (mainly DFIX 1.50(1) and 2.45(2), 1.52(1) and 2.46(2) for 1,2-and 1,3-C-C distances; SIMU 0.02 and RIGU 0.002 for U ani (C)).Occupancies of the disordered parts of the alkyl chains were fixed at 0.5.

Steady state and time resolved PL Spectrometer:
Steady state PL measurement was performed using FluoroMax ® -4 spectrometer.Steady state PL measurement a continuous output of 150-W from xenon lamp and single grating excitation and emission monochromators having resolution 0.3 nm, maximum scan seep 80 nms -1 , accuracy S-4 ±0.5 nm, step size 0.625-100 nm, range 0-950 nm are used inside spectrometer.A calibrated photodiode (R928P) is used to detect the emission photons from the range of 200-850 nm with a linearity of 2 X 10 16 counts s -1 (<100 dark counts s -1 ).Time resolved PL measurement, the instrument works on the principle of time correlated single photon counting (TCSPC).In the present work, 450 nm laser pulses was used as the excitation light sources and a TBX-04 detection module coupled with a special Hamamatsu PMT was used for photons detection.The PL decays were detected at 790 nm (instrument detection limit).The decay trace was fitted using the exponential equation   =  !!! !!/! !where, I (t) is the total intensity remaining at time t.
Where,  ! and τ i are the amplitude and decay time of i th component, respectively.The average lifetime of the measured samples is calculated using equation.

fs-TA spectroscopy
The fs-TA spectroscopy were performed on timescales of 0.1 ps to 6 ns, which is based on a multipass amplified Ti:sapphire laser (800 nm laser pulses of 7 mJ/pulse energy of ~100 fs pulse width having 1 kHz repetition rate, Astrella from Coherent), and in conjunction with Helios spectrometers.The excitation pump pulses at 450 were generated after passing through a fraction of 800 nm beam into the spectrally tunable (240−2600 nm) optical parametric amplifier (Newport Spectra-Physics).The pump fluence of the excitation laser source was adjusted by using neutral density (ND) filter to avoid the multiple charge carriers generation.The probe pulses (UV visible and NIR wavelength continuum, white light) were generated by passing another fraction of the 800 nm pulses through the 2-mm thick calcium fluoride (CaF 2 ) crystal.
Before white light generation, the 800 nm amplified pulses were passed through a motorized delay stage.Depending on the movement of delay stage, the transient species were detected following excitation at different time scales.The white light was split into two beams (named as signal and reference) and focused on two fiber optics for the improvement of better signal to noise ratio.The excitation pump pulses were spatially overlapped with the probe pulses on the samples after passing through a synchronized mechanical chopper (500 Hz), which blocked an S-5 alternative pump pulses.The absorption changed (ΔA) was measured with respect to the time delay and wavelength (λ).All spectra were averaged over a time period of 2 s for each time delay.
Note: It is advised to use to wear appropriate eye-protection during the laser experiments.

Figure S9 .Figure S10 .
Figure S9.Additional molecular orbitals involved in the prominent optical transitions.

Table S2 .
The calculated optical transitions with the highest contributions to the optical absorption features.