Supporting Information for: Superatom State-Resolved Dynamics of the Au 25 (SC 8 H 9 ) - 18 Cluster from Two-Dimensional Electronic Spectroscopy Tatjana Stoll, 1 Enrico Sgrò, 1 Jeremy W. Jarrett, 2 Julien Rehault, 1,3 Aurelio Oriana 1, Luca Sala, 1 Federico Branchi, 1 Giulio Cerullo, 1 Kenneth L. Knappenberger, Jr. 1,2,* 1 IFN-CNR, Dipartimento di Fisica, Politecnico di Milano, Piazza Leonardo da Vinci 32, 20133 Milano, Italy 2 Department of Chemistry and Biochemistry, Florida State University, Tallahassee, Florida 32306-4390, United States 3 Paul Scherrer Institute, CH-5232 Villigen PSI, Switzerland Table of Contents Overview of 2DES... S2 Synthesis and Structural Characterization of [Au 25 (SC 8 H 9 ) 1 18 ][N(C 8 H 17 ) 4 ) + ].... S3 Experimental Setup for Two-Dimensional Electronic Spectroscopy... S6 Data Fitting Procedure... S8 Supplementary data... S9 References:... S11 S1
Overview of 2DES Femtosecond two-dimensional electronic spectroscopy (2DES) can be treated as an extension of conventional pump-probe spectroscopy. In ultrafast pump-probe spectroscopy, a sample is excited by a temporally short and spectrally broad laser pulse, which often results in simultaneous excitation of multiple electronic and vibrational states. Therefore the detected transient absorption signal is the summation of contributions from all of these states. The resultant spectral overlap often precludes achieving state-resolved electron dynamics for systems characterized by a high density of states, such as the monolayer-protected clusters studied here. 2DES preserves the high temporal resolution of femtosecond pump-probe spectroscopy while also making use of the broad laser spectral bandwidth for stateresolved electronic relaxation studies. This is achieved by expanding nonlinear transient absorption signals onto a 2D spectrum that plots both the corresponding excitation and detection frequencies, as shown in Figures 2-4 of the main manuscript. The system nonlinear response is generated by excitation with three consecutive laser pulses separated by controllable delays t 1 (coherence time) and t 2 (population time); the population time corresponds to the pump-probe time delay for conventional pump-probe measurements. This three-pulse sequence builds up a macroscopic third-order nonlinear polarization that emits a signal field following the third pulse at time delay t 3. The signal field is fully measured in amplitude and phase using a fourth time-delayed laser pulse (which is sometimes the third pulse itself) that serves as a local oscillator. Fourier transformation of t 1 and t 3 for a fixed t 2 population time provides the excitation (ω 1 ) and detection (ω 3 ) axes that form the 2D map. For a description of experimental implementations of 2DES, see reference S1. S1 Cross peak detection at various ω 1 -ω 3 combinations allows for examination of state-to-state energy flow in complex photonic systems. The 2DES maps described in the main manuscript illustrate the effectiveness of the technique for providing state-resolved description of electronic relaxation dynamics for monolayer-protected clusters. S2
Synthesis and Structural Characterization of [Au 25 (SC 8 H 9 ) 1 18 ][N(C 8 H 17 ) 4 ) + ]. [Au 25 (SC 8 H 9 ) 1 18 ][N(C 8 H 17 ) 4 ) + ] was synthesized with small modification of a previously reported method. S2 Here, [N(C 8 H 17 ) 4 ) + ] serves as a counter ion. [Au 25 (SC 8 H 9 ) 1 18 ][N(C 8 H 17 ) 4 ) + ] was synthesized by adding 1.00 g (2.54 mmol) of HAuCl 4 and 1.56 g (2.85 mmol) of tetraoctylammonium bromide (TOAB) to 70 ml of THF in a 1 L erlenmeyer flask. This solution was stirred for 15 min. Next 1.8 ml (13.4 mmol of phenylethanethiol was added to the solution and stirred until the solution became colorless, about 3 hours. In a separate flask a solution was prepared by dissolving 925 mg (25.5 mmol) of NaBH 4 in 24 ml of ice cold water. This solution was than rapidly added to the initial solution under vigorous stirring. After about ten minutes the flask was covered with aluminum foil and the reaction was stirred for 48 hours. The resulting mixture was allowed to settle. Excess water was pipetted off the bottom of the flask and the THF solution, which contains the product, was dried under vacuum to a volume of 5 10mL. To this a large excess (70-80 ml) of 200 proof but not anhydrous ethanol was added to the solution and put back on the rotovap until little or no THF can be detected by scent. The use of non-anhydrous ethanol/methanol is important because [Au 25 (SC 8 H 9 ) 1 18 ][N(C 8 H 17 ) 4 ) + ] is somewhat soluble in these solvents when they are anyhdrous. Orange color in the ethanol washes indicates that [Au 25 (SC 8 H 9 ) 18 1 ][N(C 8 H 17 ) 4 ) + ] is being solubilized with the thiol. The suspended solution was collected by centrifugation. The pelleted product was washed again by resuspension in methanol, followed by product collection by centrifugation. The methanolic supernatant was discarded and the product was solubilized in dichloromethane, followed by centrifugation. The supernatant was collected and the product was extracted from the pellet a second time. Methanol was then added to the dichloromethane solution to precipitate the product and the dichloromethane was removed under vacuum. The product was suspended in methanol and then centrifuged and the supernatant discarded. If the odor of thiol remains, additional washes may be necessary. Once excess thiol has been removed any remaining methanol was dried off and the product, which consist largely of [Au 25 (SC 8 H 9 ) 1 18 ][N(C 8 H 17 ) 4 ) + ], was dissolved in a S3
minimal amount of a toluene solution containing TOAB. Next fractional precipitation was preformed by adding a small amount of either anhydrous or non-anhydrous ethanol to the solution. The solution was than centrifuged. The initial precipitates are characterized by UV/Vis and are discarded if it does not have a significant amount of [Au 25 (SC 8 H 9 ) 1 18 ][N(C 8 H 17 ) 4 ) + ]character. This process is repeated until the precipitate appears to be [Au 25 (SC 8 H 9 ) 1 18 ][N(C 8 H 17 ) 4 ) + ]by UV/Vis. Finally x-ray quality crystals can be grown by slowly cooling the supernatant in a freezer. After crystallization if a significant amount [Au 25 (SC 8 H 9 ) 1 18 ][N(C 8 H 17 ) 4 ) + ] remains in solution than more ethanol is added and the supernatant is put back into the freezer until the supernatant no longer appears to be [Au 25 (SC 8 H 9 ) 1 18 ][N(C 8 H 17 ) 4 ) + ] by UV/Vis. The synthesis of [Au 25 (SC 8 H 9 ) 1 18 ][N(C 8 H 17 ) 4 ) + ] was confirmed by square wave voltammetry and linear absorption measurements as shown in Figure S1 and S2. Square wave voltammograms were performed on a BAS 100 potentiostat in a DCM solution containing 0.1 M tetrabutylammonium hexafluorophosphate. Figure S1. The absorbance spectrum of [Au 25 (SC 8 H 9 ) 18 1 ][N(C 8 H 17 ) 4 ) + ] S4
Figure S2. A square wave voltammogram of the final product. S5
Experimental Setup for Two-Dimensional Electronic Spectroscopy. The experimental set-up for two-dimensional electronic spectroscopy (2DES) measurements in the visible is shown in Figure S3. The optical layout and experimental procedure for acquiring femtosecond time-resolved two-dimensional data has been previously described in detail. S3 The pump frequency axis is created using a time-delayed and phase-locked excitation pulse pair generated by Translating-Wedge-Based Identical Pulses encoding System (TWINS) technology. The TWINS principle has been described in detail. S3,S4 The t 1 calibration and phasing procedures used to generate the 2DES maps described in the main text were the same as previously described. S3 The 2DES system is built around an amplified Ti:sapphire laser system (Libra, Coherent), which delivered 4-mJ, 100-fs pulses at 800 nm, and 1-kHz repetition rate. Portions of the laser with 300 µj energy each were used to pump two synchronized noncollinear optical amplifiers (NOPAs), which formed the pump and probe arms of the 2DESsetup. The first NOPA (NOPA1 in Figure S2) delivered visible pump pulses with a spectrum spanning from 2.4 ev to 1.8 ev nm, as shown in Figure 1a. These pump pulses were compressed to sub-30-fs duration by multiple reflections from custom-designed double-chirped mirrors (DCMs). The second NOPA (NOPA2 in Figure S2) delivers visible probe pulses with a spectrum spanning between 2.3 ev and 1.7 ev, compressed to sub-20-fs duration by multiple reflections from custom-designed DCMs. The phaselocked pump-pulse pair is generated by the TWINS setup, while the probe pulse is delayed by t 2 via a conventional translation stage. As previously described, the t 1 time delays can be controllably changed from 10 as to 500 fs. A portion of the pump beam is redirected from the pump propagation path and sent to a photodiode to monitor the t 1 interferogram of the pump pulse pair, which is used to determine time zero and properly phase the 2DES spectra. The t 2 delays spanned from -150 fs to 10 ps, with step sizes of as small as 10 fs. The t 2 time is determined by the pathlength of a linear translation stage that temporally delays the pump and probe laser pulses. Pump and probe pulses were non-collinearly focused on the sample to spot sizes of 110 µm and 100 µm, respectively. The sample was contained in a 200-µm pathlength cuvette. The transmitted probe light was dispersed on a spectrometer. Rapid scanning of the t 1 time delays allowed for a typical 2DES map at a given t 2 delay to be acquired within 10 seconds; the set of 2DES S6
measurements at different t 2 delays was then repeated multiple times until a satisfactory data quality was reached. Dispersion introduced to the pump pulse replicas by the TWINS optics was compensated by a suitable number of reflections on a DCM pair, and spectral phase correction was verified using a Spatially Encoded Arrangement for Temporal Analysis by Dispersing a Pair Of Light E-fields (SEA-TADPOLE) setup. S5 Figure S3. Experimental Set-up for femtosecond time-resolved 2DES. NOPA: noncollinear optical parametric amplifier; DCM: double chirped mirror; PD: photodiode; L: converging lens; t 1 : linear translation for t 1 pump pulse pair delay; t 2 : linear translation for t 2 pump-probe delay; BS: beam splitter; P N : polarizer number N; M N : motorized delay N; TWINS: Translating-Wedge-Based Identical Pulses encoding System; VA: variable pulse attenuator; SM: spherical mirror. S7
Data Fitting Procedure. Dynamics traces were extracted from 2D maps at specified excitation-detection energies at each t 2 pump-probe delay. These data were fit with an in-house program that uses an iterative least-squares approach to fit the data with the following equation: S( t) = g t ( ) A i exp t /τ i i ( ) + A inf Where the symbol stands for convolution and g(t) is the instrument response function (IRF), i.e. the cross-correlation of Gaussian pump and probe laser pulses, A i is the amplitude coefficient of the i th component, τ i is the time constant of the i th component, and A inf is a non-decaying plateau function. The dynamics data in this study were treated with two exponential functions (i = 2) one growth component and one decay component. Table S1: Exponential fitting results obtained from dynamics from the regions traces presented in Figures 3c and 3d. Pump Probe τ 1 Growth τ 2 Decay Energy Energy (fs) (fs) (ev) (ev) Growth Decay Plateau Amplitude, Amplitude, Amplitude, A 1 A 2 A inf 2.21 1.84 10 ± 16 200 ± 9 0.592-0.328 0.080 2.21 1.94 15 ± 5 310 ± 14 0.644-0.325-0.031 2.21 2.07 73 ± 8 280 ± 6 0.476-0.384-0.139 2.21 2.18 201 ± 15 285 ± 19 0.473-0.493-0.034 1.95 2.07 85 ± 15 390 ± 16 0.338-0.375-0.287 S8
Supplementary data Figure S4. Time-dependent differential transmission signal resulting from broad bandwidth excitation of Au 25 (SC 8 H 9 ) - 18 using a conventional pump-probe experimental geometry. Panel (a) portrays the wavelength-resolved signal recorded by an optical multichannel analyzer plotted versus pump-probe time delay. Here, a red false color image represents positive ΔT/T signals, which corresponded to transient bleaching. Blue represents negative ΔT/T signal amplitudes, reflecting excited state absorption signal. Panel (b) shows the ΔT/T spectra recorded for several pump-probe time delays as specified in the panel insert. A time dependent blue shift of the bleach signal was observed as the pump probe time delay is increased from 200 fs to 1000 fs. Panel (c) shows time-dependent ΔT/T signals recorded at specific probe wavelengths given in the figure inset and identified by horizontal lines in panel (b). The time-dependent data showed an inversion in the sign of the signal recorded at 1.84 ev probe energy (from negative to positive ΔT/T) as the pump-probe time delay was increased. S9
Figure S5. Slices of 2D maps along the probe axis at excitation energy of 2.21 ev. The color scale maps 150 to 800 fs of t 2 delay from red to violet. The data show a 65 mev spectral blue shift and 170 mev peak narrowing as t 2 increases. Figure S6. ESA peak maximum location (black) and ESA peak width (red) as a function of t 2 delay. These data are results from fitting the ESA peak with a Gaussian function to determine the peak width and location. They depict a t 2 -dependent a 65 mev blue shift and a 170 mev peak narrowing. S10
References: (S1) Fuller, F. D.; Ogilvie, J. P. Annu. Rev. Phys. Chem. 2015, 66, 667. (S2) Parker, J. F.; Weaver, J. E. F.; McCallum, F.; Fields-Zinna, C. A.; Murray, R. W. Langmuir 2010, 26, 13650. (S3) Rehault, J.; Maiuri, M.; Oriana, A.; Cerullo, G. Rev. Sci. Instrum. 2014, 85, 10. (S4) Maiuri, M.; Rehault, J.; Carey, A. M.; Hacking, K.; Garavelli, M.; Luer, L.; Polli, D.; Cogdell, R. J.; Cerullo, G. J. Chem. Phys. 2015, 142, 10. (S5) Bowlan, P.; Gabolde, P.; Shreenath, A.; McGresham, K.; Trebino, R.; Akturk, S. Opt. Express 2006, 14, 11892. S11