Fig. 3: Theoretical model.

a Schematic representation of the model. Orange blocks represent different parts of the model: SDW, PLD, and a thermal bath. b, c Schematic evolutions of \(y\) and \(L\) on their respective potential surfaces. Features of the potential surfaces \(E(y,t)\) and \(E(L,t)\) are exaggerated, and the color accentuates the surface shape, for illustration purposes. d Map of the simulated amplitude of the acoustic mode, depending on pump–pump, and pump–probe delay. The faint row, highlighted by the blue arrow, at 0 ps pump–pump delay demonstrates the weaker excitation by a single high-fluence pulse. Higher amplitudes are achieved at higher pump–pump delay times. e Typical example of the fit achieved by the model. Purple line is the experimental data, vertical bars show uncertainties, estimated as a standard deviation \({\sigma }_{{{\mathrm{FEL}}}}\) of the PLD amplitude at t < 0 and as \({\sigma }_{{{\mathrm{full}}}}=\sqrt{{\sigma }_{{{\mathrm{FEL}}}}^{2}+{\sigma }_{{{\mathrm{ex}}}}^{2}}\) at t > 0, where \({\sigma }_{{{\mathrm{ex}}}}\) is a standard deviation of the minimum reached PLD amplitude after the first excitation between different scans with identical fluence \(P\). Orange line is the theoretical model. The vertical red dashed line marks the time of second pulse arrival.