Measuring electron temperature dynamics of gold nanoparticles by ultrafast photoemission spectroscopy

The dynamic processes and the energy redistribution of the charges following the photoexcitation of metallic nanoparticles (NPs) lie at the very heart of some of the most interesting light-induced physical phenomena, such as photocatalysis and solar energy conversion. In fact, the study of evolution of the states of nanostructured matter following impulsive photoexcitation is a fascinating subject that has a huge fundamental and technological impact. These spatially-confined systems attract great interest for their applications in nanophotonics, hyperthermia, and photocatalysis thanks to their peculiar optical properties that promote hot-electron-mediated phenomena and nonlinear effects. As an example, it has been observed that the incorporation of metallic NPs in photovoltaic cells significantly improves their performances. 
In order to investigate the mechanisms involved into electromagnetic-energy conversion we need to study the fundamental processes occurring following the interaction of light with matter. In this work, thanks to the strong collaborative effort of different research groups, we were able to report for the first time a direct measurement of the ultrafast electron-temperature dynamics in plasmonic gold NPs in the first femtoseconds after impulsive photoexcitation. We measured the ultrafast electron-temperature dynamics within ensembles of NPs deposited onto a transparent conductive oxide, by means of ultrafast pump-probe photoemission experiments performed at the SPRINT laboratory of the CNR-IOM institute. The NPs were excited by means of an ultrashort pulse with wavelength at 650 nm (close to their localized surface plasmon resonance, λ_LSPR=600 nm), and ultrafast time-resolved photoemission spectra were collected as a function of the delay time after the excitation exploiting extreme-ultraviolet ultrashort pulses obtained by high-harmonic generation (HHG, photon energy 16.9 eV, Figure 1).
 

Figure 1.  Time-resolved pump-probe spectra of Au NPs at the Fermi edge (markers). Best fits according to Fermi-Dirac distribution (solid lines). The curves have been offset in energy and intensity for the sake of clarity. The inset shows spectra acquired in correspondence of three representative delay values (-10, 0.75, and 100 ps, respectively).

We were able to detect tiny variations on the Fermi edge of gold NPs, ascribed to the ultrafast heating and relaxation of the electron gas. By performing the appropriate fits on the Fermi edge we extracted the temperature of the electron gas as a function of the delay time elapsed from the moment of excitation (Figure 2). After excitation, the electronic temperature quickly increased and reached the maximum value after several hundrend of femtoseconds (~800 fs) and then gradually relaxed towards the environment temperature.  Theoretical simulations were performed in order to confirm the effect, yielding a temperature dynamics in good agreement with observations (Figure 2). 
The results reported here represents a significant progress for future experimental and theoretical investigationin the field, opening exciting perspectives for direct and quantitatively accurate studies of the electrodynamics of metallic bulk and nanostructured systems.

Figure 2.  Time-resolved pump-probe spectra of Au NPs at the Fermi edge (markers). Best fits according to Fermi-Dirac distribution (solid lines). The curves have been offset in energy and intensity for the sake of clarity. The inset shows spectra acquired in correspondence of three representative delay values (-10, 0.75, and 100 ps, respectively).Electron temperature values extracted from the experimental time-resolved pump-probe spectra (red markers). Red solid line: fit of the experimental data by means of the sum of an exponential rise function, an exponential decay function and a step function convolved with a Gaussian simulating the instrumental response function of the system. Green line: simulated electron temperature according to the three-temperature model (3TM). The simulated temperature is rescaled by a factor of 0.52 for better comparison with the measurements in terms of temporal dynamics. Black line: Gaussian profile of the driving term in the 3TM (FWHM equal to 120 fs).

 

This research was conducted by the following research team:

Maria Sygletou,¹ Stefania Benedetti,² Marzia Ferrera,¹ Gian Marco Pierantozzi,³ Riccardo Cucini,³ Giuseppe Della Valle,⁴ Pietro Carrara,⁵ Alessandro De Vita,⁵
Alessandro di Bona,² Piero Torelli,³ Daniele Catone,⁶ Giancarlo Panaccione,³ Maurizio Canepa,¹and Francesco Bisio⁷

¹ OptMatLab, Dipartimento di Fisica, Università di Genova, Genova, Italy
² CNR-Istituto Nanoscienze, Modena, Italy
³ Istituto Officina dei Materiali-CNR, Laboratorio TASC, Trieste, Italy
⁴ Dipartimento di Fisica, IFN-CNR, Politecnico di Milano, Milano, Italy
⁵ Dipartimento di Fisica, Universita degli Studi di Milano, Milano, Italy
⁶ Istituto di Struttura della Materia - CNR (ISM-CNR), EuroFEL Support Laboratory (EFSL), Rome, Italy
⁷ CNR-SPIN Istituto Superconduttori Materiali Innovativi e Dispositivi, Italy

Contact persons:

Maria Sygletou, email:
Francesco Bisio, email: