martin wolf department of physics, freie universität berlin€¦ · department of physics, freie...

51
structure kinetics dynamics E. Muybrigde 1887 Martin Wolf Department of Physics, Freie Universität Berlin Application of femtosecond laser spectroscopy “Modern Methods in Heterogeneous Catalysis Research“ - WS 04/05 Time-resolved Studies of Surface Reactions Goal: Microscopic understanding of coupling between electronic excitations and nuclear degrees of freedom

Upload: others

Post on 20-Jul-2020

3 views

Category:

Documents


0 download

TRANSCRIPT

  • structure

    kinetics

    dynamics

    E. Muybrigde 1887

    Martin WolfDepartment of Physics, Freie Universität Berlin

    Application of femtosecond laser spectroscopy

    “Modern Methods in Heterogeneous Catalysis Research“ - WS 04/05

    Time-resolved Studies of Surface Reactions

    Goal: Microscopic understanding of coupling between electronic excitations and nuclear degrees of freedom

  • Timescales of chemical reactions

    Typical timescale: 10-100 fs

    Temporal evolution from reactants to products:

    Dynamics of the transition state

    Key concept: Dynamics on Born-Oppenheimer potential energy surface

    Non-adiabatic coupling between electronic states near (avoided) crossings

  • QMS

    Ru(0001) T = 95 KCO

    Ru(0001)

    O

    T=95K

    CO

    800 nm, 120 fs~ 50 mJ/cm2

    CO2

    2 CO + O → CO ↑adsads

    Motivation: fs-laser-induced chemistry

    Example: CO Oxidation on Ru(001) competes with CO desorption (UHV conditions)

    New reaction pathway is „switched on“ upon fs excitation

    hν2 CO + O → CO ↑ + CO2↑adsads

    (1x2) - O/Ru(001)+ CO saturation

    Flux

    First ShotYield

    200150100500Tim e -o f-F lig h t (µs )

    CO2 : 1590 K

    Etrans =

    CO : 640 K

    Flux

    [arb

    . uni

    ts]

    fs-induced reaction pathways

  • Outline

    IntroductionNon-adiabatic processes at surfaces

    Surface femtochemistry

    Electron and phonon mediated pathways

    Isotope effects and electronic friction model

    Electron thermalization in metals

    Test of the two-temperature model

    metal semiconductor

    e-h pair

    e-

    Non-linear optics as a probe of surface dynamics

    Vibrational sum-frequency generation spectroscopy

    log

    (2PP

    E In

    tens

    ity)

    0.60.30.0-0.3

    E-EF (eV)

    Ru(001)

    non-thermalizedelectrons

    Electron dynamics thin ice layers on metals Ru(001)

    SFG

  • Gas surface interaction

    Role of non-adiabatic processes in surface reactions ?

    Example: Adsorption at a metal surface

    E

    R

    „forced-oscillator model“

    Energy dissipation via phonon excitation

    Coupling to electron-hole pairexcitations in the substrate

    Exothermic reactions at metal surfaces

    Question: Coupling between nuclear(vibrational) motion of adsorbate and electronic excitations ?

  • Chemicurrents

    See review article by H. Nienhaus, Surf. Sci. Rep. 45, 3 (2002)

    Evidence for non-adiabaticcoupling between adsorbatemotion electronic excitations

    Direct observation of e-h pair excitation

    Adsorption on thin metal film

    Charge separation acrosssmall Schottky barrier

    Chemicurrent

    Current provides lower limitof excitation probability

  • Chemicurrents: Mechanism

    Newns-Anderson model

    Chemiluminescence Exoelectron emission

    Unoccupied affinity level ispulled down by adsorbatesurface interaction

    Filling of hole below EFermiby substrate electrons

    broadening of resonance linewidth

    adsorbate substratecharge transfer

    e-h excitation:- chemicurrent- exoemission

    luminescence

  • Electronic excitations at surfaces

    metal semiconductor

    e-h pair

    e-

    Gergen et.al., Science 294, 2521 (2001)

    Chemicurrent observed forvarious adsorbates

    chemical reactions drivenby electronic excitations

    Energy dissipation via e-h pair excitation („electronic friction“)

    Feasibility of reverse process?

    electronic excitations playmaior role in gas surfaceinteraction

  • Laser induced surface dynamics

    Question: How does laser excitation induce surface reactions ?

    Electronic excitations and charge transfer

    Chemical bonding implies electronic couplingbetween „adsorbate“ and substrate

    Vibrational relaxation at metalsby electron-hole pair excitation

    hot e-

    E

    EF

    e-transfer

    substrate adsorbate

    HOMO

    LUMOElectron stimulated desorptionand surface photochemistry

    Energy transfer by coupling betweenelectronic and nuclear degrees of freedom

    Thermal activation

    Phonon driven process (in electronic ground state)

    kBT

  • metal

    Optical excitation of metal surfaces

    Relaxation to phonon and adsorbate vibrational excitation

    Mechanism and timescales of energy transfer after optical excitation

    Photoabsorption in a metal substrate creates a transientnon-equilibrium electron distribution

    Primary step: Electron thermalization and cooling

    ~0.1-1 ps

    ~1 pselectrons

    Tel

    direct absorptionfs-excitation reaction

    >1 ps

    ~100 fs

    phononsTphm

    etal

    adsorbate vibrations Tads

  • Electron thermalization dynamics in metals: Test of the two-temperature model using 2PPE

    Electron dynamics in metals following optical excitation

  • The two-temperature model

    Nearly thermal (Fermi-Dirac)distribution for electrons

    Model assumes two heat baths for electrons and phonons: Cel >> Cph

    PdCoupled diffusion equationfor Tel and Tph

  • Time-resolved two-photon photoemission

    2kinF hEE-E ν−Φ+=

    EF

    Evac

    Ekin

    hν1

    hν2

    e-

    Φ

    ϕ

    /mE2sink kin|| hϕ=

    TOF

    e-

    energy- and time-resolution:

    pulseduration:65 fs

    momentum resolution:

    e-UV probe pulse

    pump pulse

    e-e scattering

    Electron thermalization dynamics:

    How to probe electron thermalization and cooling ?

    hot electroncascades

  • Experimental setup

    Femtosecond time-resolved two-photon photoemission spectroscopy

    µJ

  • W. S. Fann, R. Storz, H.W.K. Tom, J. Bokor,Electron thermalization in gold, Phys. Rev. B 46, 13592 (1992)

    Method: time-resolved photoemissionwith hνprobe > Φ

    Pump fluence: 120 µJ/cm2

    300 µJ/cm2

    Electron thermalization in gold

    Electron thermailization occurs fasterwith increasing excitation density

    Fermi-Dirac distribution

  • Time-resolved 2PPE from Ru(001)

    0.0001

    0.001

    0.01

    0.1

    1

    2PP

    E in

    tens

    ity

    1.51.00.50.0

    E - EF (eV)

    0.0001

    0.001

    0.01

    0.1

    1

    0.0001

    0.001

    0.01

    0.1

    1

    580 µJ/cm2

    230 µJ/cm2

    40 µJ/cm2∆ t / fs

    0100200300400500

    Ru(001)

    e-UV probe800 nm pump

    Pump-probe scheme:

    probing electron dynamics around EFv

    Lisowski et al., Appl. Phys. A 78, 165 (2004)

    e-e scattering

    1) Electron thermalization:

    2) Relaxation by e-ph scattering

    3) Energy transport in the bulk

  • Prediction from 2TM

    Analysis of electron distribution

    Comparison with 2TM:

    Assume thermalized distribution

    for electrons and phonons

    Thermalized electrons carry onlyminor fraction of excess energy

    Non-eqilibrium carrier distribution:

    non-thermalized ‚hot‘ electrons

    thermalized Fermi distribution

    measured Tel

    2TM over-estimates final Tel

  • DOS of Ru and Cu

    -0.05

    0.00

    0.05

    norm

    aliz

    ed p

    ump-

    indu

    ced

    varia

    tion

    0.60.30.0-0.3E-EF(eV)

    delay / fs0100200300500

    Dynamics of electrons and holes

    e-probe

    Pump induced changes of electron distribution:

    Highly symmetric electron and hole distributions for all delaysdue to nearly constant DOS around EF

    Dynamics of electrons and holes can be treated equivalently

    Excess energy of electrons + holes

    electrons

    holes

    pump

    Ultrafast relaxation (much faster compared to noble metals, e.g Au: 1-2 ps)

  • Extended heat bath model

    efficient ballistic transportof non-thermal electrons

    Goal: Understanding the role of non-thermalized electrons

    however, peak energy is describedreasonable well by both models

    2TM underestimates rate of energy flow into the bulk

    2

    4

    68

    10-6

    2

    4

    68

    10-5

    2

    4

    68

    elec

    tron

    ener

    gy[J

    /cm2

    ]

    9006003000-300time delay (fs)

    Extended modelincluding ballistic transport

    2TM

    2TM

    Extension of two temperature modelAppl. Phys. A 78, 165 (2004)

  • Femtochemistry at metal surfaces

    substrate-mediated excitation mechanism dominates

    Anisimov, et al. Sov. Phys. JETP , 375 (1974)39

    120 fs50 mJ/cm2

    phonon

    electron

    Time (ps)

    Tem

    per

    atu

    re (

    K)

    T

    T

    1 20

    Mechanism and timescales of energy transfer after optical excitation

    transient non-equilibrium between electrons and phonons: Tel >> Tph

    ~0.1-1 ps

    ~1 pselectrons

    Tel

    direct absorptionfs-excitation reaction

    >1 ps

    ~100 fs

    phononsTphm

    etal

    adsorbate vibrations Tads

    new reaction mechanism?by non-thermal activationby separation of time scales of energy flow

  • Experimental Investigations

    • NO and O2/Pd• CO/Cu • CO/Pt• O2/Pt •CO/NiO• CO + O2/Pt• CO + O/Ru, H+H/Ru

    Misewich, Loy, HeinzTom, PrybylaHo MazurStephenson, Richter, CavanaghBonn, Wolf, ErtlZacharias, Al-ShameryDomen

    YieldFluence dependenceTranslational energyVibrational energy Rotational energyUltrafast dynamicsInfluence of laser pulse durationInfluence of photon energyDependence on adsorption siteDependence on coverageCompetitive reaction pathways

    Surface femtochemistry

    Systems

  • QMS

    Ru(0001) T = 95 KCO

    Ru(0001)

    O

    T=95K

    CO

    800 nm, 120 fs~ 50 mJ/cm2

    CO2

    2 CO + O → CO ↑adsads

    Example I: Desorption and oxidation of CO on Ru(001)

    Oxidation of CO impossible under equlilibrium conditions

    200150100500Time-of-Flight (µs)

    CO: 640 K

    CO2: 1590 KEtrans =

    4003002001000

    Absorbed Fluence (yield weighted ) [J/m2]

    Y(CO2) ~ F3.2

    Y(CO) ~ F3.1

    x 10F

    lux

    Firs

    t Sho

    t Yie

    ld

    New reaction pathway upon fs excitation

    hν2 CO + O → CO ↑ + CO2↑adsads

    (1x2) - O/Ru(001)+ CO saturation

  • Example II: Recombinative desorption of H2 on Ru(001)

    H2 formation induced by fs laser excitation of H/Ru(001)

    Pronounced isotope effects in H2/D2 yield

    High translational temperature of desorbing species

    Full monolayerof H/Ru(001)

    QMS

    H2

    H

    Ru(001)

    H

    = 10 ± 2.4Yield (H2)Yield (D2)

    H2: (2110 ± 400) K

    flight time [µs]

    flux[a.u.]

    60

    (x 10)

    40200

    0

    D : (1720 ± 290) K2

    / 2kkin B

  • Reaction mechanisms

    Electron-mediatedPhonon-mediated

    How to distinguish?

    Desorption Inducded byMultiple Electronic Transitions

    DIMET -

    substrate

    adsorbateE

    reaction coordinate

    phonons Edes

    hot e-

    E

    EF

    τ ~1-10 fs

    ground state

    stateexcited

    Edes

    e-transfer

    reaction coordinatesubstrate

    kBTph

  • Temperature profiles and 2-pulse correlation

    TphTel

    6000

    5000

    4000

    3000

    2000

    1000

    020151050

    time [ps]

    surfa

    cete

    mpe

    ratu

    re[K

    ]

    pulse 1 pulse 2

    FWHM ~ 10 ps

    FWHM ~ 1 ps

    pulse-pulse delay [ps]re

    actio

    nyi

    eld

    [a.u

    .]

    -200 -100 0 100 200

    phonon-mediated: slow response

    electron-mediated:fast response

    Goal: Discerning electron and phonon driven mechanisms

    Two-pulse correlation measurement of the reaction yield allowsto distinguish between the two reaction mechanisms!

    QMS

    CO

    Ru(001)

    CO

    O

    CO2delay

  • Experimental setup

    Ti:sapphire oscillator + amplifier 4 mJ 800 nm, 110 fs, 20-400 Hz,

    2 equally intensepulses

    Ru(001)

    QMS

    H2/D2CCD

    pulse profilediagnostics

    /

    UHV

    “Feulner cup“ detector

    80 mm

    O = 33 mm/

    Ru(001)

    QMSionisationvolume

    H2

    CO

  • 2-pulse correlation measurements

    120 fs50 mJ/cm2

    phonon

    electron

    Time (ps)

    Tem

    per

    atu

    re (

    K)

    T

    T

    1 20

    CO

    O

    ∆t

    Ru

    Tel Tph

    CO

    CO2

    800 nm120 fs

    -200 -100 0 100 200

    x24 CO2

    COfir

    stsh

    otyie

    ld(a

    .u.)

    pulse-pulse delay (ps)

    FWHM=3 ps

    FWHM=20 ps

    2-pulse correlation:

    Firs

    t Sho

    t Yie

    ld

    Goal: discriminate between electron and phononmediated mechanism

    : Oxidation hot electron-mediated

    : Desorption phonon-mediated

  • DFT calculations C. Stampfl, M. Scheffler, FHI Berlin

    unoccupiedanti-bonding

    orbital ~1.7eVabove EF

    increasing Tel leads to

    weakening ofRu-O-Bond

    Ru(001)

    O

    Hot electron-induced activationof the O-Ru bond

  • Mechanism for CO oxidation

    M. Bonn et al, Science , 1042 (1999)285

    hotelectrons

    EFermi

    E O-level

    anti-bonding

    reaction coordinate

    electronshot

    Ru(001)

    E

    CO + O

    0.3 eV

    4.7 eV

    1.7eV

    from DFT

    V = 1V = 2

    ¬

    O + CO

    CO

    CO + O

    ad ad

    ad

    2,g

    g g

    gad

    ¬

    ¬

    E ~1.8 eVa

    Hot electron mediated vibrational excitation of the O-Ru bond

    Non-adiabatic energy transfer from electronic excitations to adsorbate coordinate

  • Discussion

    E

    reaction coordinate

    O

    O

    C

    O*thermal excitation

    O

    0.8 eV

    CO

    O

    O

    CO

    O

    CO*

    O

    C

    fs-laser excitation

    1.8 eV

    1.0 eV1.8 eV

    e-

    kT

    Under thermal excitation the CO has desorbed before oxygen is activated

    Reaction pathways for thermal versus femtosecond excitation

    Seperation of time scales opens new reaction pathway

  • 2PC of laser-induced H2 formation

    Ultrafast response indicatescoupling to hot electron transient Coupling rates: τel τph τel-ph-1 -1 -1 ., ,

    met

    al

    τphτel-ph Tph

    τel

    elTheat diffusion

    adsT

    ads-E /k Ta bRate ~ e

    Two-Temperature Model

    Coupled heat baths: el ph adsT , T , T

    first

    shot

    yiel

    d[a

    .u.]

    pulse-pulse delay [ps]

    H2exp. data

    FWHM = 1.1 ps

    0-10 -5 0 5 10

    model

    Brandbyge et al., PRB 52, 6042 (1995)

    Electronic friction model yields: Ea = 1.35 eV and τel = 180 fs(τel = 360 fs for D2)

    for H2

  • Isotope effect in recombinative desorption

    H2 formation induced by fs laser excitation of H/Ru(001)

    Pronounced isotope effects in H2/D2 yield

    Full monolayerof H/Ru(001)

    QMS

    H2

    H

    Ru(001)

    H

    = 10 ± 2.4Yield (H2)Yield (D2)

    H2: (2110 ± 400) K

    flight time [µs]

    flux[a.u.]

    60

    (x 10)

    40200

    0

    D : (1720 ± 290) K2

    / 2kkin B

    Isotope effect

  • Energy transfer by coupling betweenadsorbate vibration and metal electrons

    Electronic friction:

    hotelectrons

    EFermi

    E

    leveladsorbate

    electronshot

    metal

    Coupling rate η = ~ 11 Melτ

    Isotope effect and electronic friction model

    Brandbyge et al., PRB 52, 6042 (1995)

    pote

    ntia

    lene

    rgy

    Ru-H/D distance

    relectronicallyexcited PES

    ground state of substrate-adsorbate-complex

    Origin:- mass-dependent distance

    traversed on excited PES

    - lighter adsorbate starts movingmore rapidly

  • Fluence dependence: isotope ratio and first shot yield

    Electronic friction model* simultanouslydescribes: • 2-pulse correlation

    • fluence depedence of yield and isotope ratio

    * Brandbyge et al., PRB 52, 6042 (1995)

    exp. data model power law fit

    Y(H2) ~ 2.8 and Y(D2) ~ 3.2

    0

    160140120100806040200

    0

    403020100laser shot #

    ~75J/m2 Hcounts

    [a.u.]2

    first

    shot

    yiel

    d[a

    .u.]

    isotoperatio

    20

    10

    2001601208040

    experimental data

    5

    15

    25

    0

    isot

    ope

    ratio

    Y(H

    2)/Y

    (D2)

    electronic friction model

  • ~0.1-1 ps

    ~1 pselectrons

    Tel

    direct absorptionfs-excitation reaction

    >1 ps

    ~100 fs

    phononsTphm

    etal

    adsorbate vibrations Tads

    Real-time probing of vibrational dynamics

    So far probed only the (desorbing) reaction products

    Goal: Probe suface reactions directly in the time domain

    Use surface sensitive non-linear optics (SHG, SFG) to obtain a direct look inside the excited adlayer

  • Vibrational spectroscopy: Sum-frequency-generation

    Principle of sum-frequency generation (SFG):

    Pot

    . Ene

    rgy

    C-O distance

    IRv=0

    v=1

    v=2

    SFGVIS

    CO

    C-O stretch

    P E E= χ ω(2) (2)

    IR ωVISω ωSFGSFG

    ωSFG = ωIP + ωVIS

    Surface sensitive method:SFG is symmetry forbidden in isotropic (bulk) media (in dipole approximation)

    Second order non-linearoptical process

    Time-resolved SFG as molecule specific probe of surface reactions

  • Broadband IR vibrational SFG spectroscopy

    SFG spectrum ( 3x 3)-CO/Ru(001) √ √

    SFG wavelength [nm]690 688 686 684694 692

    SFG

    inte

    nsity

    IR

    20001950

    FWHM9 cm-1

    1900 2050IR wavenumber [cm ]-1

    2100 2150

    200 K

    SFG

    IR intensity

    IR

    VIS SFG

    IR

    VIS

    Ru

    Vibrational spectroscopy without tuning the IR frequency

    SFG

    IRIR

    VIS

    VIS

    Γn

    ΓSFG

    01

    2

    ~5 cm-1

    ~150 cm-1

    Use spectrally broad pulse with spectrally narrow upconversion pulse

    fs- IRVIS

    L.J. Richter et al., Opt. Lett. 23, 1594 (1998).

  • SFG experimental setup

    Ti:sapphire oscillator + amplifier4 mJ, 800 nm , 110 fs, 20-400 Hz,

    OPG/OPA (TOPAS), 20-30 J µ2.5-10 mµ

    tunableIR pulse

    pump pulse

    spectrometerCCD camera

    multi-channel detectionRu(001)

    QMS

    SFGprobe

    pulse shaperFWHM 7 cm , 7 J µ-1

    narrow bandVIS pulse

  • SFG

    inte

    nsity

    (a.u

    .)

    21002050200019501900IR wavenumber (cm-1)

    3 ps

    23 ps

    0.5 ps×8

    ×8

    ×8

    Pump-Probe delay –4.5 ps–2.0 ps–1.5 ps–1.0 ps–0.5 ps0.5 ps3.0 ps

    23.0 ps68.0 ps

    168.0 ps

    CO

    yiel

    d(a

    .u.)

    100shot number

    210020001900

    T =340 K, F=55 J/m02

    CO/Ru(001)

    Time-resolved SFG of CO/Ru during desorption

    IR

    fs pump

    CO

    Ru(001)

    VIS SFG

    M. Bonn Phys. Rev. Lett , (2000) 4653et al, 84

    Spectroscopic snapshots of the CO stretch mode

    opticalprobe

    Pronounced transient red shift, linewidthbroadening and decrease of intensity

    Problems:Dipole-dipole coupling in the adlayerand small concentration of „products“

  • Real-time probing of surface reactions: 2PPE Snapshots of excited electronic states during evoltution of wavepacket

    Time-resolved 2PPE spectra ‚Desorption‘ of Cs/Cu(111)

    photoemissionprobe

  • Electron localization in adsorbate layers

    Electron localization and solvation in polar adsorbate overlayers

  • H2O: 1.7.eV

    NH3: 0.8.eV

    Optical absortion spectrumSchindewolf, Angew. Chemie 80, (1968) 165

    Solvated electrons in polar liquids

    Courtesy: C.P. Schulz, MBI Berlin

    Brief history:

    Davy (1808) reports blue ammoniacontaining compounds

    Weyl (1863) discovers a bluesolution of Na in liquid ammonia

    Kraus (1908) attributes the bluecolor to trapped electronssurrounded by NH3 molecules

    Ogg (1940) develops firstmodel for solvated electrons

    Hart & Boag (1962) discoverthe hydrated electron in water

  • Time-resolved two-photon photoemission

    2kinF hEE-E ν−Φ+=

    EF

    Evac

    Ekin

    hν1

    hν2

    e-

    Φ

    ϕ

    /mE2sink kin|| hϕ=

    TOF

    e-

    energy- and time-resolution:

    pulseduration:65 fs

    momentum resolution:

    4.0

    3.5

    3.0

    2.5

    4.0

    3.5

    3.0

    2.5

    E-E

    F[eV]

    2PPE intensity

    IS

    SS

    solvatedelectrons

    2PPE

    inte

    nsity

    4002000-200pump-probe delay [fs]

    SS

    solvated electrons

    1 BLD O/Cu(111)2

    E-E

    F[e

    V]

    2PPE spectra as a function of time delay

    analysis of solvation and localization dynamicsusing time- and angle-resolved 2PPE

  • 2PPE experiment

    e-

    TOFUHV-chamber

    QMS

    ∆t=

    -platesλ 2

    CCD-camera

    Pinhole-doser

    gas doser

    Probe

    ∆xc

    variabletime delay

    tunable

  • Part I: Electron transfer dynamics in D2O/Cu(111)

    Dynamics of photoinjected electrons in multilayers of amorphous ice

    eS

    eCB

    hνpump= 3.92 eVhνprobe= 1.96 eV

    D2O/Cu(111)4 BL

    τp = 65 fs

    E - EF [eV]

    delay [fs]

    2PP

    E in

    tens

    ity

    4.03.63.22.82.4

    0

    200

    400

    600

    time-resolved spectra

    Ultrafast relaxation within experimental time resolution (eCB -> eS)

    Stabilisation of localized electronic state (eS) on time scale of 0.1-1 ps

    two-photon-photoemission

    E-E

    [ eV]

    4.0

    3.5

    3.0

    2.5

    E-E

    F[e

    V]

    6004002000delay [fs]

    eS

    eCB

    solvatedelectrons

    conductionband

  • Energetic stabilization and population decay

    3.0

    2.8

    2.6

    e pe

    ak m

    axim

    um: E

    - E

    SF[

    eV]

    150010005000delay [fs]

    5.0 BL3.8 BL3.0 BL2.9 BL

    Coverage

    eS

    peak shift

    delay [fs]

    1

    10

    100

    2PPE

    inte

    nsity

    8004000

    τ =110 fs

    eSeCB

    population

    eS: energetic stabilization 300 meV/ps,non-exponential decay t >100fs

    eCB: population decay within pulse width

    Isotope effects:very similar stabilization ratefor H2O and D2O

    ~50 meV shift due to differentband gap/ orientational disorder

    isotope effect

    delay [fs]

  • Dispersion and localization dynamics

    Angle-resolved 2PPE spectra hν2

    mek||= 2 Ekin / h2 sin(α)

    TOF

    e-

    α

    z

    first moment

    eS

    eCB

    3.0

    2.9

    2.8

    2.7

    E-

    EF

    [eV]

    0.10.0

    k|| [Å-1]

    3.53.02.5

    E - E F [eV]

    2PP

    EIn

    tens

    ity

    α0°8°

    14°18°

    ∆t= 0 fs

    ∆t= 200 fs

    ∆t = 0 fs

    50 fs

    100 fs

    200 fs

    300 fs

    1) Electron injection intodelocalized state witinexperimenal time resolution

    2) Localization within

  • collectivesolvation coordinate

    (0)

    q

    E-E

    F

    (eV

    )0

    5

    q1

    V k qCB || 1( , )

    V qS 1( )

    Mechanism

    Electron transfer, localization and solvation dynamics in H2O/Cu(111)

    hω2

    z

    E-E

    F

    (eV

    )

    0

    -5

    5

    VB

    CB

    Cu(111) H O2

    Evac

    hω1 Egap

    real space

    eSV qS 2( )

    Electron transfer into theconduction band (CB) ofthe ice layer

    Relaxation to the bottomof CB (τ

  • Summary

    IntroductionNon-adiabatic processes at surfaces: Chemicurrents

    Electron thermalization in metals

    Test of the two-temperature model

    Electron dynamics thin ice layers on metals

    Electron injection, localization and solvation

    Ru(001)

    SFG

    0.0001

    0.001

    0.01

    0.1

    1

    2PPE

    inte

    nsity

    1.51.00.50.0

    E - EF (eV)

    ∆t / fs0100200300400500

    Ru

    2PP

    E In

    tens

    ity

    2PPE

    e-+ +-

    Cu

    D O2

    2PPE

    Surface femtochemistry

    Electron and phonon mediated pathways

    Isotope effects and electronic friction model

    Vibrational sum-frequency generation spectroscopy

  • thanks to

    C. Gahl, U. Bovensiepen, J. Stähler, M. Lisowski, P. Loukakos,

    D. Denzler, S. Funk, C. Hess, and G. Ertl, Fritz-Haber-Institut Berlin

    M. Bonn, AMOLF, The Netherlands

    S. Wagner, and C. Frischkorn, Freie Universität Berlin

  • H. Petek and S. Ogawa, Femtosecond time-resolved two-photon photoemission studiesof electron dynamics in metalsProgress in Surface Science 56, 239-311 (1998).

    H. L. Dai and W. Ho, Laser Spectroscopy and Photochemistry at Metal SurfacesWorld Scientific (1995).

    Literature:

    P.M. Echenique, R. Berndt, E.V. Chulkov, Th. Fauster and U. Höfer, Decay of electronic excitations at metal surfacesSurface Science Reports 52, 219-318 (2004).

    H. NienhausElectronic excitations by chemical reactions at metal surfacesSurface Science Reports 45, 1-78 (2002).