Attosecond spectroscopy on solids Reinhard Kienberger Technische Universität München Max-Planck Institut für Quantenoptik, Garching X-Ray Science in the 21 st Century KITP 06/Aug/2010
Overview Needs for generating isolated attosecond x-ray pulses via HHG: Phase-stable few-cycle driver pulses at long wavelengths attosecond measurement technique: streaking What to do with attosecond pulses? Pulse duration measurement at LCLS
The Laser System Hollow Fiber/ Chirped Mirror Pulse Compressor <3.5 fs pulse duration 400 µj energy Experiment Booster: 4 khz, 4 mj, 28 fs CPA Amplifier 20 fs pulse duration 1.3 mj energy 3 khz repetition rate Ti:Sapphire Oscillator sub-10 fs pulse duration 4.6 nj energy 82 MHz repetition rate
key tools of attosecond technology: synthesized few-cycle wave & synchronized sub-fs xuv pulse hollow fiber neon gas intensity 10 4 10 2 spectrum normalized intensity 10 0 400 700 1000 8 7 6 5 4 3 2 pulse duration < 3.5 fs pulse energy > 0.4 mj < 1.5 cycles > 0.1 TW λ [nm] A. L. Cavalieri et al, New J. Phys. 9, 242 (2007); E. Goulielmakis et al, Science 317, 769 (2007) 1 0-20 -15-10 -5 0 delay [fs] 5 10 15 20
Few-Cycle Laser Pulses Control of the Waveform E(t) = E a (t)cos(ω L t + φ) Cosine waveform φ = 0 Sine waveform φ = π/2 T 0 /4 625 as (@ λ 0 0.75 µm) T 0 = λ 0 /c 2.5 fs Requires measurement & control of φ
Detection of the absolute phase Electron detection: MCP Shielded drift tube Trigger:PD Left-right ATI spectra Ionized gas: Xenon slit MCP Laser beam Phase change by glass wedges Previous results: average of 10000 shots Paulus et al. Nature, 414 182 (200
CEP of consecutive phase-stabilized pulses consecutive time-of-flight ATI spectra of rescattered electrons (laser rep. rate = 3kHz) phase distribution of CEP-stabilized shots (4500 shots; 3kHz) Std single shot = 278mrad shot-to-shot evolution of the CEP
Parametric (Lissajous-like) representation Wittmann et al. Nature Physics 4 (200 single shot left-right TOF spectra CEP=0 CEP=π Non-phase-stabilized single shots- random distribution asymm 2 7π/6 π 5π/6 4π/3 Left- Left+Righ Right t phase scan- limited by fluctuations asymm. 2 11π/6 5π/3 3π/2 mean std asymm 1 π/2 2π/3 High precision asymm. 1 2π π/6 π/3 No phase ambiguit y
Non-phase-stabilized laser shots as they re arriving non-stabilized shots has random phase distribution first principle calibration: phase difference between shots is immediately given real CEP retrieved by TDSE Wittmann et al. Nature Physics 4 (200
Phase-scan with a stabilized laser Wittmann et al. Nature Physics 4 (200
High-order Harmonic Generation in the gas phase Step 1 Optical field ionization E(t) L e - Step 2 Step 3 XUV emission on recollision e - E(t) L Spectral Intensity - e Acceleration E(t) L e - P.B. Corkum, PRL 71 (1993)
Extending the cutoff using an IR driving field
Comparison VIS / IR driver V. Tosa and V. Yakovlev
2.1 μm few-cycle OPCPA system New pump laser Disc regen 3 khz / 1.5 12 ps mj/ 27 mj > 0.9 2mJ mj,.?? 15 fs
IR driven HHG Cut-off ~1.6 kev
Recombination Emission from Strongly-Driven Atoms Multi-cycle driver pulse : τ p» T o High-order odd harmonics of the driver laser Spectral intensity 25 nm 12.5 nm 7.5 nm 1 0.1 21 51 81 111 Harmonic order Cut-off harmonics: train of attosecond bursts L Huillier, Balcou, 1993, PRL 70, 774 Macklin et al, 1993, PRL 70, 766 Paul et al, Science 292, 1689 (2001) Tsakiris, Charalambidis et al, 2003
steering bound electrons with controlled light fields: the birth of an attosecond pulse A. Baltuska et al, Nature 421, 611 (2003) xuv-filter 1000 intensity [a.u.] E L (t) E L (t) E L (t) E L (t) E L (t) 0 50 60 70 80 90 100 110 photon energy [ev] E L (t) ħω x electron trajectories cosine wave
attosecond xuv/x-ray pulse generation 1000 intensity [counts] 0 60 70 80 90 100 110 photon energy [ev] sine wave A. Baltuska et al., Nature 421, 611 (2003)
attosecond pulse generationand measurement time-of-light electron spectrometer few-femtosecond, femtosecond, few-cycle laser pulse λ L 750 nm τ L = 3.5-5 fs W L = 0.3-0.5 mj Ne gas xuv pulse knocks electrons free in the presence of the few-cycle laser field 1000 intensity [a.u.] 0 50 atomic gas XUV spot near- diffraction-limited xuv/soft-x-ray beam 60 70 80 90 100 110 Drescher et al, Science 291, 1923 (2001) Hentschel et al, Nature 414, 509 (2001) Kienberger et al, Science 297, 1144 (2002)
Ionization with an Isolated Attosecond Pulse Ne Δp() t = r e t r E L () t dt Mo-Si multilayer mirror delay possible Time-of-flight spectrometer Δp p f p i Nozzle Detection as in: Kienberger et al., Science 297, 1144 (2002) XUV cut-off energy: ~100 ev Mirror reflectivity bandwidthup to: 30 ev (FWHM)
Ionization at different instants of time
final momentum of photoelectrons depending on the release time final momentum of the electron time +Δp max 0 -Δp max
Laser electric field Mapping Time to Energy Change in electron energy ΔW(t 7 ) ΔW(t 6 ) ΔW(t 5 ) t 1 t 2 t 3 t 4 t 5 t 6 t 7 instant of electron release ΔW(t 4 ) ΔW(t 3 ) ΔW(t 2 ) Incident X-ray intensity ΔW(t 1 ) ħω x -500 as 0 500 as
Optical Streak Camera, 1834
sampling field oscillations ΔW +10 ev 0-10 ev ħω x t D R. Kienberger et al, Nature 294 (2004)
85 75 65 a streaking trace delay [fs] -4 0 4 8 E. Goulielmakis et al, Science 305, 1267 (2004 electron kinetic energy [ev]
Goulielmakis et al, Science 320, 1614 (2008) isolated sub-100-as pulses
AS beamline
real-time observation of multi-electron dynamics (excitation, relaxation, correlation) in atoms, molecules & solids? by means of strong-field-induced free-free transitions: streaking kinetic energy 0 unocc. valence occupied valence final charge state binding energy photo- emission & shake up +1 valence photo- emission +1 Auger decay +2 core-level photo- emission +1 Auger decay & shake up +2 core orbital
Ionization of electrons in a solid sample TOF NIR XUV A () t l Al (') t sample system
proof of principle: attosecond streaking of core-level (4f) attosecond spectroscopy in solids? & conduction-band (Fermi-edge) electrons in tungsten 800 E L (t) electrons from 4f band Δτ = 110 Fermi ± 70 edge as energy 0 delay delay [fs] 400 1 0 65 0 5.0 63 61-0.5 4.5 59 3.0 57 55-1 1.5 E X (t) 0 τ x = 300 as ħω x = 95 ev τ L = 4f 5 fs band λ L = 750 nm valence core-level Auger photo- Auger electrons from -1.5 photo- photo- the decay emission 4f band reach the surface decay 0 20 40 60 80 100 120 140 emission ~ emission 110 as later than & those & from the Fermi- shake up shake up photoelectron energy [ev] +1 edge +1 +2 +1 energy shift [ev] 0.5 +2-6 Fermi edge -4-2 0 2 4 delay [fs] 91 89 87 20 40 85 60 80 100 120 140 83 A. L. Cavalieri et al., Nature 449, 1029 (2007)
as spectroscopy of clean metal surface Re (0001) Samples: Re(0001) and W(110) Photon energies: 90, 120, 130 ev Core level electrons delayed by 55 90 as. Δt CB 4f Spectrogram analysis: retrieval algorithm based on analytical solution of TDSE Delay of 4f core levels relative to CB: 64 ± 10 as
Origin of delay? Propagation effect Assumption: streaking acts outside solid Different inelastic mean free path different depth profile Different group velocity (final-state effect?) 4f emitted after CB E E, vac, solid 16 Theory: A. Kazansky et al., PRL 102, 177401 (2009). C. Lemell et al., PRA 79, 62901 (2009). C.-H. Zhang et al., PRL 102, 123601 (2009). J.C. Baggesen et al., PRA 78, 32905 (2008).
Current Effort: Absolute Emission Time..blow gas on the sample.. geometric effects.
What about the reference? Ne 2s / 2p gas phase 100 100 kinetic energy 0 2p 2s -5.5 1 0 delay [fs] intensity [arb. units] 5.5 0 80 80 60 40 60 40 20 20 1.0 photoelectron energy energy [ev] [ev] binding energy E X (t) delay -300 0 300 emission of low-energy photoelectron is delayed by about 20 as 0.5 0.0 M. Schultze et al, Science 328,, 1658 (2010).
Duration of photoemission? Idea: Introduce electronic reference state localized at the surface time zero marker Baggesen et al., PRA 78, 32905 (2008).
Xe monolayer on Re Re CB Xe 5p Re 4f Xe 4d
Xe monolayer on Re Order Time delay of appearance Re CB Xe 5p Re 4f 46 as 65 as 6 as ± 11 as Xe 4d NIR intensity ~ 5.5 10 11 W/cm 2
Comparison atomic vs. solid xenon atomic xenon solid xenon (50 layers) +22 ± 12 as 4d is first electron propagation effects 20 ± 3 as 5p is first
Electron transport through defined adlayers 1 3 metallic layers substrate and electron states
Electron transport through defined adlayers IR XUV electron transport IR refraction and penetration dielectric multilayer substrate localized and delocalized electron states
Outlook Generation of UV pulses for excitation of molecules: UV-pump/XUV-probe Charge-transfer in molecules and soldis
Attosecond 2PPE Goal: study excited state dynamics with as to few-fs time resolution Requirements: few-cycle UV pump pulses (hν= 4-6 ev) synchronized with Isolated XUV probe pulse (hν= 80-130 ev) TOF E ΔE hν(uv) variable Δt
Intense pulses in the deep ultraviolet Gas target, Ne 1 9 bar SD- FROG setup U. Graf et al., Optics Express 16,18956 (2008)
Emission of UV light at different target pressures
Intense (>2 μj) pulses in the deep ultraviolet U. Graf et al., Optics Express 16,18956 (2008)
Autocorrelation trace on Kr ion yield Graf et al. Opt. Exp. accepted
delay chamber
Collinear generation of UV and XUV NIR/VIS τ < 4 fs 400 μj Neon ca 0.2-0.3 bar Neon ca 3-5 bar Argon ca 1 bar XUV τ 80 as UV τ 4 fs E 1.5μJ Isolated XUV attosecond pulses 1000 Intensity [counts] 0 60 70 80 90 100 110 Photon energy [ev] A. Baltuska et al., Nature 421, 611 (2003) E. Goulielmakis et al., Science Vol. 320 (2008) Intense sub-4 fs pulses in the deep UV Intensity Phase Intensity Phase Time [fs] Wavelength [nm] U. Graf et al., Optics Express 16, 18956 (2008)
Collinear generation of UV and XUV NIR/VIS NIR/VIS + UV + XUV Neon ca 3-5 bar Argon ca 1 bar Neon ca 0.2-0.3 bar
Collinear generation of UV and XUV up to 1 µj Neon Neon 230 mbar up to 10 7 ph/pulse E. Bothschafter et al., Optics Express (2010)
Exciting an electron wavepacket in a molecule Sampling the evolution with XUV AS-pulses First candidate: ozone
Oriented molecules on a surface uv- or SAM
probing charge transfer/migration in complex (bio-)molecules? F. F. Remacle, Remacle, R. R. D. D. Levine, Levine, PNAS PNAS 103, 103, 6793 6793 (2005) (2005) 80 82 84 86 photoelectron energy [ev] 88 90 92 94 0.5 0.75 1 1.25 1.5 time delay[fs]
Colleagues & Cooperations F. Krausz R. Ernstorfer M. Schultze Y. Deng A. Schiffrin N. Karpowics E. Goulielmakis A. Cavalieri X. Gu G. Marcus I. Grguras Jobst M. Fiess W. Helml T. Paasch M. F. Reiter B. Horvath E. Magerl T. Metzger B. Dennhardt M. Hassan T. Wittmann E. Bothschafter A. Schwarz theory: J. Gagnon, V. Yakovlev MPQ Garching, Germany P. Echenique San Sebastian, Spain XUV optics & spectroscopy: surface dynamics: CEP measurement: M. Hofstetter, U. Kleineberg Ludwig Maximilians Univ. Munich, Germany S. Neppl, P. Feulner, J. Barth TU Munich, Germany G. G. Paulus Univ. Jena, Germany
X-ray Pulse duration measurement at LCLS In collaboration with: DESY, CFEL, Ohio State U., Dublin U., SLAC
Laser-field dressed photoelectrons Electron energy E L (t) I x (t) hν L hν x τ x >> T 0 /2 long X-ray pulse W binding
Laser-field dressed photoelectrons λ IR = 2 μm T 0 /2 = 3.3 fs I x (t) Electron energy E L (t) τ x < T 0 /2 Detector short X-ray pulse hν x W binding 0
taking into account: - photon energy jitters by 1 % - bandwidth jitters (7 ev + 3 ev) single-shot streaking measurement of τ x-ray E el Maximum shift! +ΔE ħν-i p Maximum broadening! A Laser - ΔE time-jitter Retrieval of A max ħω x Retrieval of τ x-ray
spectra sorted acc. to phase cavity timer long pulse, λ Laser = 2 μm photoelectron energy (TOF) delay FEL IR
spectra sorted acc. to phase cavity timer short pulse, λ Laser = 2 μm toelectron energy (TOF) delay FEL IR
g{x XÇw