Light-induced molecular processes on ice Grecea, Mihail Laurentiu

Grecea, Mihail Laurentiu

Citation

Grecea, M. L. (2006, February 23). Light-induced molecular processes on ice. Retrieved from https://hdl.handle.net/1887/4322

Version: Corrected Publisher’s Version

License: Licence agreement concerning inclusion of doctoralthesis in the Institutional Repository of the University of Leiden

Downloaded from: https://hdl.handle.net/1887/4322

Chapter 4

Ultrafast laser-induced

desorption of water from

amorphous solid water

4.1 Introduction

The mechanisms of energy and charge transport through water are of interest from both a fundamental and technological point of view. The fundamental interest arises from processes which are of importance for e.g. biology, environmental sciences and astrochemistry. For the latter, for example, ionizing radiation in the interstellar medium may trigger chemistry at the surface of ice particles. The radiation will be predominantly absorbed in the bulk of the material, whereas many of the important chemical processes occur on the particle surface. Transport of energy or electrons is therefore important in these processes. On the other hand, our understanding of electron-mediated processes in aqueous systems is important for technological applications as diverse as electrochemistry, radiation processing and medical diagnosis and therapy. As such, there has been much effort trying to understand the fundamentals of electron-initiated processes in water and aqueous systems. An excellent recent review is provided by ref. [1].

A class of very fundamental studies consists of well-defined nanometer water layers in UHV that are supported on metal substrates, which are exposed either directly to (low energy) electrons (see, e.g., [2] and references therein), or to laser radiation that liberates electrons from the support, after which these are injected into the ice film [3-7]. These and other studies have provided detailed insights into the dynamics of excited states, water-substrate interactions, the role of water film morphology and the solvation dynamics of energetic electrons. One important aspect that has not yet been studied in much detail is the mechanism of water transport through the supported ice layers, following electron-induced excitation of water molecules, resulting in desorption.

well-defined water layer, whose thickness can be controlled up to a few angstroms. Desorption occurs at the water-vacuum interface.

The scheme of the experiments is presented in Figure 4.1: electrons are injected into the water layer from a Pt surface after irradiation with a femtosecond laser pulse. The laser pulse temporarily heats the electrons to ~1000’s of K (see inset of Figure 4.1), so that these have sufficient energy to excite water molecules which can then desorb. By using immiscible layers of D2O on top of H2O or vice versa, we can determine that the water does not solely desorb from the outermost part of the layer, but from inside as well. We observe that water molecules can penetrate a water film: a water molecule has a mean free path length corresponding to ~2.5 water bilayers, i.e., a distance in excess of 1 nm.

Figure 4.1. Schematic representation of the experimental approach: (1) The laser

excites the electrons in the Pt(533) substrate, which (2) are injected into the thin ASW film, where (3) sufficient energy is released during the relaxation process that (4) water desorbs that is detected with a QMS in TOF experiments. The inset shows the time evolution of the electron and phonon (lattice) temperature for a typical excitation fluence of ~90 Jm-2_{. The electrons are heated to ~3000 K, whereas the phonons remain} relatively cold. 2 Pt(533) e -1 3 4 QMS

h

ν

time after excitation (ps)

T

T

elec

4.2 Experimental

The experiments are performed in a UHV chamber in combination with an amplified Ti:Sa femtosecond laser system (~2.7 mJ, 130 fs, 1 kHz). Roughly 2 mJ of the 800 nm laser output is used to induce desorption of water. Water layers (H2O or D2O) are grown by an effusive molecular beam with the sample at 100 K resulting in highly compact ASW films [8]. Desorbed water molecules are detected with a mass spectrometer after a 70 mm flight path along the surface normal. After amplification with a fast amplifier the molecules are counted with a multichannel scaler as a function of their flight time. Desorption of water clusters as well as photoproducts, e.g., H2, could not be observed. Also in thermal desorption experiments after irradiation no photoproducts are observed. Direct excitation of water by the laser pulse is negligible as water does not absorb light with a wavelength of 800 nm. Laser-induced crystallization of the ASW phase was not observed [6]. Experiments are single pulse, enabled through the combination of a mechanical chopper and a shutter. The film thickness is obtained from temperature-programmed desorption experiments and expressed in monolayer (ML). Note that 1 ML corresponds to a layer of two water molecules; see Chapter 3 for definition of ML.

4.3 Results and discussion

The inset of Figure 4.2 shows a time-of-flight (TOF) spectrum of water desorbing from a 5 ML thick ASW film averaged over 20 shots with each shot on a fresh spot on the solid water film. A Maxwell-Boltzmann fit reveals a translational temperature of ~2300 K. As the translational temperature of the desorbing water exceeds the phonon temperature (and indeed the melting temperature of Pt), the desorption process can be concluded to be electron-mediated. The first-shot desorption yield is obtained by integration of the corresponding TOF spectrum and is plotted in Figure 4.2 along with the water (D2O) translational temperatures as a function of excitation fluence for a 15 ML thick film of ASW. The strongly non-linear fluence dependence of the yield (Yield ~ Φ6_{, where} Φ _{is the laser yield-weighted fluence [10]) is also reminiscent} of electron-mediated desorption.

phase. The electron energy distribution for a given electron temperature Telec is stated by the Fermi-Dirac distribution:

( ) / 1

( ;

) (

1)

elec

n E T

=

e

−

For one laser shot, Telec depends on time in the manner exemplified in Figure 4.1. The desorption probability is assumed to be proportional to the total number of electrons above 2 eV created with one laser shot:

2

(

2

)

( ;

)

P

n E

eV

dt

dE n E T

∝

>

=

∫

∫

where the temporal integration is carried out from -10 ps to 1 ns in the numerical procedure used to reproduce the data. This range is more than sufficient, as electron cooling occurs very rapidly. A measure of the translational energy can be obtained from weighting the instantaneous electron temperature with the number of sufficiently energetic electrons, and normalizing to the latter:

1 1

10 2 10 2

( ; ) ( ; )

F F

ns ns

translational elec elec elec

channels are included in this simple model. Note, however, that the trend is clearly reproduced. Hence, we conclude that hot electrons injected from the Pt interface trigger the observed desorption events. Increasing the threshold for the required electron energy to 3 eV rather than 2 eV, changes very little in the calculated results: the translational temperature increases by ~150 K over the whole fluence range studied here, while the calculated fluence dependence describes the data even better compared to the 2 eV cutoff.

Figure 4.3 reveals that both the desorption yield and the translational temperature decrease with increasing layer thickness. The fluence is chosen such that for an initial coverage of one monolayer, roughly 50% of the water desorbs

co

unts (a.u.)

time of flight (µs)

T

(K

)

absorbed laser fluence (Jm

)

0.10 0.08 0.06 0.04 0.02 0.00

D

O des

orption (ML/s

hot)

Figure 4.2. The first shot yield and the translational temperature as a function of

in a single shot, as observed in depletion curves (the desorption yield as a function of the laser shot number on one place; data not shown). For thicker layers (>5 ML) no significant depletion (<10%) is observed within the first 10 laser shots. The decay of the desorption intensity with layer thickness is more or less exponential, with a ~5 ML exponential length scale. The electron mean free path of 1 eV electrons, has been demonstrated to be characterized by a 2.3 ML exponential length [13], so that the decay observed here is appreciably slower than that expected from the electron penetration depth. This implies that either additional water transport or excitation transport occurs through the film, prior to desorption. 1 10 100 1000 10000 H2 O yie ld (a . u.) 40 30 20 10 0 H₂O layer thickness (ML) 5000 4000 3000 2000 1000 T tr ansl a ti onal (K) Yield T_{translational} Model

Figure 4.3. First shot yield and translational temperature as a function of the layer

Transport through the water film is investigated in more detail by performing experiments on mixed isotope H2O/D2O layers. In these experiments, we detect at mass 20 (D2O), as mass 18 could be H2O as well as OD fragments resulting from dissociation in the mass spectrometer. We first dose a layer of D2O and cover it with a layer of H2O (referred to as an H2O capping layer) or vice versa (spacer layer). Schematic drawings of the two cases are given as an inset in Figure 4.4. It has been shown that the diffusivity of water molecules is exceedingly small at 100 K (D << 10-20 _cm2_s-1_{), so that mixing cannot occur [14].} This is further corroborated by the TPD spectra shown in Figure 4.4, which demonstrate that the outer isotope always desorbs first. This is in agreement with previous observations that demonstrate that the diffusivity of water does not become significant until ~150 K, and the desorption rates become equal for the two isotopes at higher temperatures [14].

2000 1600 1200 T tr ans lat ional (K ) 15 10 5 0 H₂O cap/spacer layer (ML) MODEL 10 100 1000 D2 O yield (a. u. ) 15 10 5 0 H₂O cap/spacer layer (ML) MODEL H₂O cap H₂O spacer 2000 1600 1200 T tr an sl at iona l _(K ) EXPERIMENT 10 100 1000 D2 O yield ( a . u.) EXPERIMENT H₂O cap H₂O spacer

a

c

b

d

Figure 4.5. The D2O first shot yield (a) and translational temperature (b) after laser-induced desorption from 5 ML D2O layers on Pt covered with H2O layers of various thickness (capping layer) and from 5 ML D2O layers of various thickness on Pt spaced by H2O layers of various thickness. The inset of Figure 4.5c shows the two conditions. (c) and (d) The corresponding calculated desorption yield and translational temperature, respectively.

The laser-induced desorption yield is plotted in Figure 4.5a for layers with 5 ML of D2O and a variable thickness in the H2O spacer or capping layer. In both cases D2O is detected and the yield decreases with increasing spacer or capping layer thickness. Surprisingly, significant amounts of D2O can penetrate the H2O

It is important to note that only the first shot yield is measured. After the first shot we can not exclude that the layers are locally intermixed due to transient laser heating. The data points are averaged over typically 8 fresh spots on the sample. The translational temperature of the desorbing D2O molecules is depicted in Figure 4.5b. Remarkably, the temperature for D2O molecules that have to penetrate the H2O layer is significantly and reproducibly higher than for D2O desorbing from top of the layer; water close to the Pt substrate displays higher translational temperatures.

From our data we can conclude that water molecules can penetrate through the ASW layer without loosing significant amounts of kinetic energy. To extract the mean free path of a water molecule, we model our data in the following way. Electrons from the Pt are injected in the water layer and can excite a water molecule. The electron penetration into the water layer is characterized by an exponential decay, which has previously been determined to be 2.3 ML for 1 eV electrons [13]. Electron-excited water molecules can subsequently traverse the remaining water layer across to the vacuum, allowing the molecule to desorb. The desorption yield is thus determined by the product of the probability that a certain water molecule is excited by an electron from the Pt and its subsequent transmission probability through the water layer. In the model we postulate that the excitation probability Pexc depends exponentially on the distance from the Pt surface and is determined by the electron mean free path, while the molecular escape probability Pesc depends exponentially on the distance from the excitation location to the free surface.

obtain an electron mean free path, which determines the spatial variation of Pexc, of 4 ML and a water mean free path, which determines the spatial variation of Pesc, of 2.5 ML. The inferred electron mean free path is larger than the value of 2.3 ML [12] reported in literature for 1 eV electrons, indicating that either the electron mean free path is slightly higher for the laser-excited hot electrons, which exhibit a distribution of energies, or secondary processes enhance this excitation length. Preliminary molecular dynamics simulations indicate that the latter effect may indeed contribute: the excitation can be transferred from one water molecule to a nearest neighbor. The translational temperature as a function of layer thickness is best described with an exponential function with a very slow decay (~40 ML) due to energy losses in the excitation exchange in between water molecules. The result is plotted in Figure 4.3 and 4.5d, taken into account the isotope effect. With this simple model, the trend that the temperature for the capping experiment is higher than for the spacer experiment can be reproduced.

1.0 0.8 0.6 0.4 0.2 0.0 nor malize d pr obab ili ty 14 12 10 8 6 4 2 0

position within the layer (ML)

2000 1500 1000 500 0 T tr an slat io na l _(K ) P_exc P_esc

Pt surface vacuum interface

Figure 4.6. Summary of the model parameters resulting from fits to the data

presented in Figures 4.3 and 4.5 exemplified for a 15 ML thick water film: The electron injection length is determined by parameter Pexc, which decays on ~4 ML length scales; the water escape probability Pesc, determined by the mean free path of a water molecule through ASW decays from the water-vacuum interface on a 2.5 ML length scale. The translational temperature decays slowly from the Pt interface (gray line).

4.4 Conclusions

Bibliography

[1] B.C. Garrett, D.A. Dixon, D.M. Camaioni, D.M. Chipman, M.A. Johnson, C.D. Jonah, G.A. Kimmel, J.H. Miller, T.N. Rescigno, P.J. Rossky, S.S. Xantheas, S.D. Colson, A.H. Laufer, D. Ray, P.F. Barbara, D.M. Bartels, K.H. Becker, H. Bowen, S.E. Bradforth, I. Carmichael, J.V. Coe, L.R. Corrales, J.P. Cowin, M. Dupuis, K.B. Eisenthal, J.A. Franz, M.S. Gutowski, K.D. Jordan, B.D. Kay, J.A. LaVerne, S.V. Lymar, T.E. Madey, C.W. McCurdy, D. Meisel, S. Mukamel, A.R. Nilsson, T.M. Orlando, N.G. Petrik, S.M. Pimblott, J.R. Rustad, G.K. Schenter, S.J. Singer, A. Tokmakoff, L.S. Wang, C. Wittig, T.S. Zwier, Role of water in electron-initiated

processes and radical chemistry: issues and scientific advances, Chem. Rev. 105

(2005) 355-89.

[2] N.G. Petrik, G.A. Kimmel, Low-energy electron-stimulated luminescence of thin

H2O and D2O layers on Pt(111), J. Phys. Chem. B 109 (2005) 15835-41.

[3] U. Bovensiepen, Ultrafast electron transfer, localization and solvation at ice-metal

interfaces: correlation of structure and dynamics, Progr. Surf. Sci. 78 (2005)

87-100.

[4] U. Bovensiepen, C. Gahl, J. Stahler, M. Wolf, Ultrafast electron dynamics in

amorphous and crystalline D2O layers on Ru(001), Surf. Sci. 584 (2005) 90-97.

[5] U. Bovensiepen, C. Gahl, M. Wolf, Solvation dynamics and evolution of the spatial

extent of photoinjected electrons in D2O/Cu(111), J. Phys. Chem. B 107 (2003)

8706-15.

[6] D. Chakarov, B. Kasemo, Photoinduced crystallization of amorphous ice films on

graphite, Phys. Rev. Lett. 81 (1998) 5181-84.

[7] C. Gahl, U. Bovensiepen, C. Frischkorn, M. Wolf, Ultrafast dynamics of electron

localization and solvation in ice layers on Cu(111), Phys. Rev. Lett. 89 (2002)

107402.

[8] K.P. Stevenson, G.A. Kimmel, Z. Dohnalek, R.S. Smith, B.D. Kay, Controlling the

morphology of amorphous solid water, Science 283 (1999) 1505-07.

[9] E.H.G. Backus, Driving and probing surfaces with light, Ph.D. Thesis, 2005. [10] S. Funk, M. Bonn, D.N. Denzler, C. Hess, M. Wolf, G. Ertl, Desorption of CO from

Ru(001) induced by near-infrared femtosecond laser pulses, J. Chem. Phys. 112

(2000) 9888-97.

[11] S.I. Anisimov, B.L. Kapeliovich, T.L. Perel'man, Electron-emission from surface of

metals induced by ultrashort laser pulses, Sov. Phys. JETP 39 (1974) 375-80.

[12] M.A. Henderson, The interaction of water with solid surfaces: fundamental aspects

revisited, Surf. Sci. Rep. 46 (2002) 5-308.

[13] S.K. Jo, J.M. White, Low energy (<1 eV) electron transmission through condensed

layers of water, J. Chem. Phys. 94 (1991) 5761-64.