Energy partitioning in the femtosecond-laser-induced associative ${\mathrm{D}}_2$ desorption from Ru(0001)

Energy transfer to different degrees of freedom during the femtosecond-laser-induced recombinative desorption of ${\mathrm{D}}_2$ from a deuterium-covered Ru(0001) surface $({\mathrm{D}}_{\mathrm{ads}}+{\mathrm{D}}_{\mathrm{ads}}∕\mathrm{Ru}\to {\mathrm{D}}_{2,\text{gas}}+\mathrm{Ru})$ has been investigated. $(1+1^{\prime })$-resonance-enhanced multiphoton photoionization (REMPI) and time-of-flight (TOF) measurements are utilized to provide information on the internal and external (translational) energy content, respectively. Rovibrational population distributions of the reaction product are detected via various $B{}^1\Sigma _u^{+}\leftarrow X{}^1\Sigma _g^{+}$ Lyman bands using tunable vacuum ultraviolet laser radiation in the resonant excitation step. Rotational quantum state populations in the vibrational ground state and the first excited state are measured yielding average rotational energies of $⟨E_{\mathrm{rot}}⟩∕k_{\mathrm{B}}=800$ and $1500\mathrm{K}$, respectively, for an absorbed laser fluence $⟨F⟩$ of $85\mathrm{J}∕{\mathrm{m}}^2$. In addition, a mean vibrational energy of the desorbing molecules is extracted which amounts to $⟨E_{\mathrm{vib}}⟩∕k_{\mathrm{B}}=1200\mathrm{K}$. Extensive TOF measurements enable complete energy balancing with $⟨E_{\mathrm{trans}}⟩∕2k_{\mathrm{B}}=2500\mathrm{K}$ at $⟨F⟩=85\mathrm{J}∕{\mathrm{m}}^2$ and underline the nonthermal and unequal energy partitioning between the different degrees of freedom within the reaction product. The effects of multidimensional electronic friction between substrate and adsorbate layer and peculiarities of the potential energy landscape governing the ${\mathrm{D}}_2$ recombination are discussed.

Figure 1

Schematic diagram of the experimental setup combining a high-power femtosecond-laser system with tunable narrow-bandwidth vuv generation based on a solid-state laser-pumped dye laser. (SHG, THG, and FHG indicate second, third, and fourth harmonic generation, respectively.) Experiments are performed under ultrahigh vacuum conditions, where the ${\mathrm{D}}_2$ molecules are formed and analyzed after desorption from the Ru surface.

Figure 2

(Color online) TOF distributions of fs-laser-induced ${\mathrm{D}}_2$ for various laser fluences $⟨F⟩$ ranging from $50\text{to}140\mathrm{J}∕{\mathrm{m}}^2$. Modified Maxwell-Boltzmann distributions (solid lines) fit the experimental spectra. Mean translational energies expressed in kelvin are obtained from the second moments of the discrete experimental data points. Note the clear shift to shorter flight times, i.e., higher translational energies, with increasing laser fluence.

Figure 3

(Color online) Translational energies $⟨E_{\mathrm{trans}}⟩∕2k_{\mathrm{B}}$ of ${\mathrm{D}}_2$ and ${\mathrm{H}}_2$ desorbing from Ru(0001) after fs-laser excitation as a function of the yield-weighted laser fluence $⟨F⟩$. Solid and dashed lines show the yield-weighted adsorbate temperature for a D and H layer, respectively (see text). Both experiment and model show a clear isotope effect. The mean translational energies expressed in kelvin are by a factor of $\sim 1.35$ larger than the calculated yield-weighted adsorbate temperature values. Note extrapolation of a linear fit to experimental ${\mathrm{D}}_2$ data results in an intercept of the $⟨E_{\mathrm{trans}}⟩∕2k_{\mathrm{B}}$ axis at $1250\mathrm{K}$.

Figure 4

(Color online) Ionization spectrum of ${\mathrm{D}}_2$ desorbing from $\mathrm{D}∕\mathrm{Ru}\left(0001\right)$ after fs-laser irradiation with $⟨F⟩=85\mathrm{J}∕{\mathrm{m}}^2$ versus the vuv wavelength of the REMPI excitation step. The line notation of the resonances is explained in Ref. 35.

Figure 5

(Color online) Rovibrational population distributions $N(v^{″},J^{″})$ of ${\mathrm{D}}_2$ desorbing from Ru(0001) after fs-laser irradiation $(⟨F⟩=85\mathrm{J}∕{\mathrm{m}}^2)$. Boltzmann plots for the vibrational ground and first excited state are shown versus the rotational energy (top panel) and the total internal energy $E_{\mathrm{rot}}+E_{\mathrm{vib}}$ (bottom panel), respectively. Filled circles, ${\mathrm{D}}_2(v^{″}=0)$; open squares, ${\mathrm{D}}_2(v^{″}=1)$.

Figure 6

(Color online) Time profiles of electronic and phononic temperatures $T_{\mathrm{el}}$ and $T_{\mathrm{ph}}$ of the Ru substrate together with the adsorbate temperature $T_{\mathrm{ads}}$ of a D adlayer and its respective reaction rate $R\left(t\right)$ after fs-laser excitation with $⟨F⟩=85\mathrm{J}∕{\mathrm{m}}^2$. In the calculations based on the two-temperature (Ref. 33) and the modified electronic friction (Ref. 34) model, the previously obtained values for the electronic coupling time and the activation energy were used (Ref. 10). The basics of these models are illustrated in the inset.

Figure 7

(Color online) Illustration of the ${\mathrm{D}}_2$ desorption process after fs-laser excitation. Coordinates denote the center-of-mass distance $z$ of the ${\mathrm{D}}_2$ molecule from the surface and the D-D interatomic distance $d_{\mathrm{DD}}$. Note the qualitative character of the potential energy landscape and the exemplary trajectory demonstrating a possible route which results in predominantly translational energy and significantly less vibrational excitation. The ground PES is adopted from Ref. 52 in which the ${\mathrm{H}}_2$ interaction with a clean Ru(0001) surface was calculated.