Ultrafast Vibrational Dynamics of CO Ligands on RuTPP/Cu(110) under Photodesorption Conditions

Round 1

Reviewer 1 Report

This work reports laser induced desorption of CO molecules from Ru-TPP adsorbed on a Cu surface. The desorption yield of CO ligands bonded to Ru-TPP was significantly larger than those directly bound to Cu surface irrespective with the excitation photon energies ranging from 1.55 to 3.1 eV. In addition, time-resolved vibrational spectroscopy has been applied to reveal ultrafast frequency shift of C-O IS mode with visible femtosecond laser excitations. The transient frequency of IS mode exhibits a blue-shift in picosecond time region upon interaction with substrate hot-electrons which is in a stark contrast with related previous works in which transient shifts showed red-shifts concomitant with excitation of the frustrated modes.

Although the mechanism of the photo-detachment of the ligand CO is not fully resolved, the reported novel findings are interesting enough and merits publication in Surfaces as an article. I would like to suggest the author(s) consider the following points concerning the desorption mechanism prior to the publication.

1) The marked polarization dependence in Figure 3 suggests a surface localized transition is responsible for the photo-detachment in which adsorbate-substrate hybrid state is involved. The absence of the facile desorption with a CW light excitation indicates that a multiple electronic transition by the intense irradiation occurs. This contradicts to the prevailing hot-electron induced mechanism: the memory of the photon polarization is expected to be lost prior to the agitation of the vibrations by hot-electrons. Therefore, the fact that the nonlinear behaviour shows a significant polarization-dependence is something new, and an additional discussion on the desorption mechanism based on the novel finding would be possible. The higher yield of the phtodesorption for CO-RuTPP than CO-Cu indicates that the lifetime of the adsorbate induced electronic states is much longer in the former case than in the latter.

The authors refuted the contribution of the direct excitation because the “facile laser desorption was observed for all wavelength used (400, 532, and 800 nm)”. However, this is not enough to refute the participation of surface localized excitation because the resonance states would be broadened by a hybridization with substrate; the authors should discuss the photon energy dependence (and polarization dependence) of the absolute cross section. The participation of the direct excitation can be excluded only in the case that the cross section follows the substrate absorptance.

The strong hot-band signals of the IS mode observed in Figure 2B indicates that the vibrational lifetime of IS mode for CO-RuTPP is much longer than that of CO-Cu. Because the damping of IS mode is dominated by substrate electron-hole pair excitation, this means that the non-adiabatic coupling between CO vibration and substrate electron-hole pair excitation is much weaker for CO-RuTPP than that in CO-Cu. This is reasonable because the strength of the non-adiabatic coupling is determined by the amount of charge flow between substrate and adsorbate as a function of the nuclear displacement: in the case of CO-RuTPP, the IS mode would be relatively decoupled from substrate electrons because CO binds not directly to Cu substrate but to the Ru atom in TPP molecule. This would be also the case not only for IS mode but also for other frustrated external modes, reducing damping rates of the frustrated modes those are excited by the electronic transitions and helping the effective desorption to occur.

 

2) Even though the IS mode frequency shows a monotonous red-shift by the static temperature increase both for CO-RuTPP and CO-Cu (Figure 4), the transient shifts show a stark contrast between these two systems, particularly in a few ps time-range where hot-electrons play a role. This would be also linked to that the excitation in CO-RuTTP involves surface local transition in which CO-Ru anti-bonding state is populated: at the higher fluences, the direct transition can occur in a multiple fashion within the laser pulses leading to an accumulation of the vibrational energy within the CO-Ru potential surface which cannot be attained by a conventional static heating.

The red-shift observed in most of the previous works on CO- metal surfaces roots in the fact that the lateral displacements of CO position in any direction from the equilibrium site bring the molecule toward the bridge or hollow sites which exhibit much lower IS mode frequencies. This would not be the case for CO-RuTPP: CO binds to the Ru atom in TPP molecule which possesses a strongly directional d-orbital; any displacement of CO centre-of-mass from the equilibrium position would lead to a reduction in the degree of ligation that would result in a blue-shift of the IS mode.

 

3) A minor comment: it is strange that the red-shift peak occurs at a negative-delay time in Figure 5A, while that appears at positive time in Figure 6. Are the horizontal axes correct?

Author Response

Please see the attached word file.

Author Response File: Author Response.docx

Reviewer 2 Report

This manuscript presents the investigation of ultrafast vibrational dynamics of CO adsorbed on RuTPP/Cu(110) using pump-probe sum frequency generation to explain CO desorbing mechanism from the RuTPP molecules in comparison to CO on bare Cu(110). An easier laser assisted desorption from RuTPP is found while CO-RuTPP desorption reveal a higher desorption energy. CO spectroscopy and dynamics are considerably modified by the introduction of the RuTPP monolayer between CO and Cu(110). They propose that the unusual transient CO frequency blueshift can be explained by a stronger anharmonic coupling between CO frustrated rotation and internal stretch that is modified by Ru-CO charge transfer and electron mediated phonon-phonon coupling.

This work is in the continuation of the already well studied CO ultrafast dynamics on various metallic surfaces, but clearly brings new insight by investigating CO coupling with organic molecules. This paper is publishable but subject to minor revisions noted below:

1) Introduction, line 51-52: Authors give a complete overview of the literature concerning the sub-ps dynamics of CO/metal surfaces, but it might be interesting to have more details about the interest of decoupling CO from the substrate. Is it a way to tune CO/surface coupling for a better understanding of CO photodesorption mecanisms or it may have an interest for applications?

2) Materials and methods, line 66: How spectra are normalized over more than 200 cm-1 (see figure 2b) considering the 200 fs duration of the IR pulse (<100 cm-1)?

3) Materials and Methods, line 68: Removing the NR in SFG signal is understandable to probe CO without any ambiguity as demonstrated by several groups. But in the presence of several close vibrations, time beating leads to modulations in the decay of the vibrational decoherence, that requires careful spectro-temporal data deconvolution to extract the proper frequency, bandwidth and intensity. In addition, considering the CO decoherence time on metals of about 0.5-2 ps, the time shifting by 1.3 ps, in order to suppress the NR signal, can considerably reduce the CO signal. How the authors have chosen this value of 1.3 ps and have they evidenced relative intensity changes between CO bands as a function of the time-shift of the visible pulse?

4) Results, line 80-81: It would have been interesting to see a STM image of CO/Cu(110) and RuTPP/Cu(110) prior CO adsorption.

5) Results, line 83: How authors have evidenced CO adsorption between RuTPP? I would write: “Between the RuTPP islands remains space for CO adsorption on Cu(110).”

6) Figure 1 (page 3): A scale is needed to figure out the size of objects on the STM image. I also suggest to add the size of the scanned area in the caption. The superimposed lattice lines do not help the reader. They could be displayed only on one part of the STM image (upper-left corner for example). Do the authors have an estimation of the coverage of RuTPP? How to explain the fact that some terraces are not covered by Ru-TPP?

7) Figure 2 (page 4): How SFG spectra are fitted? What are the experimental conditions to acquire SFG spectra, considering the laser induced desorption presented in Figure 3 ? At 10-8 mbar or for a given CO exposure?

8) Results, line 203-205: The -1.5 ps shift in time scale should be explained.

9) Results, line 206-208: According to STM image, Ru-TPP coverage is high on some terrace and not present on other. A high coverage of Ru-TPP is not compatible with high CO coverage in between. CO is more likely able to reach high coverage on terraces free on Ru-TPP. This second scenario is also compatible with TPD measurements. This point of view should be considered by the authors.

10) Figure 5 (page 7): (a) How to explain the minimum of CO frequency shifted at -1.5 ps which is at +0.5 ps in Figure 6? (b) Have authors evidenced a similar behavior for the V0-2 transition upon pump irradiation? (c) and line 222-223: The black curve is a parabolic fit, as usually expected for a second ordered driven optical process, but it is in contradiction with the statement that a different behavior occurs a higher fluence. I would say that as well as for desorption, a threshold fluence need to be reached. Can the author address this issue?

11) Results, line 220-222: At low fluence CO-RuTPP dynamics show a -1 cm-1 frequency shift in a time range (-3 to -1 ps) similar to CO/Cu(110) which is mediated by hot electrons rather than phonon. The phonon mediated regime starts after -1 ps according to author's measurements and well-known CO dynamics on metal. At high fluence, a fast blueshift is observed then followed by a slower regime up to +5ps more compatible with a phonon assisted dynamics. Authors should clarify this point.

12) Figure 6 (page 8): Curve for saturated and 0.1 ML CO are not presented in the text, even if there are used in the discussion.

13) Discussion (page 8-10): The authors tentatively assign the transient blueshift of CO frequency to the displacement or bending of CO to induce different anharmonic coupling between low frequency modes and the internal stretch of CO (adsorbed on Ru-TPP) as well as electron mediated phonon-phonon coupling. Could the coupling of Ru-TPP with hot electrons be responsible for the observed blueshift, by changing charge transfer between Ru and CO. Authors talked about Ru-CO bond length change? Is there evidence in the literature of a strong change of Ru-TPP LDOS? Globally it appear to me the impact hot electrons on Ru-TPP should be studied to fully understand CO desorption in the system.

14) Discussion, line 237: Do the authors talk about CO/RuTPP or CO/Cu(110)?

15) Discussion, line 247-248: Do the authors mean their experiments specifically?

16) Discussion, line 262-264: As far as I know, pump-probe SFG experiments of CO/Pd(100) surface and Pd NPs where CO adsorbs on bridge sites, have only shown a redshift of the frequency as well as for CO atop on Pd or Pt(111) surface.

17) Discussion, line 276-278: It is not visible in Figure 6.

18) Discussion, line 324: If Ru-CO bond lengthening play an important role for a more facile laser induced desorption (due to energy states of CO-RuTPP close to Fermi level) as well as to the transient blueshift, then this point should be highlighted.

 

Minors corrections:

Line 58: I suggest to add “than CO/Cu(110)” after “more stable CO-RuTPP/Cu(110).

Line 83: Mentioning the surface for CoTPP would help.

Line 147: Misspelling of “p-polarization”.

Line 201: “Figure 5”, not “Figure 3”.

Figure 5: (a) Why not adding a red line to guide eye as well? Caption: It might help the reader to remind that the frequency shifts are measured on CO vibrational band at 1957 cm-1 for CO-RuTPP and at 2093 cm-1 for CO/Cu(110).

Line 250: "where" instead of "so"

Line 328: I suggest to add “pump-probe” before “SFG”.

Author Response

please see attached word document

Author Response File: Author Response.docx