Kim 2023

Download as pdf or txt
Download as pdf or txt
You are on page 1of 22

Article

Attosecond field emission


https://doi.org/10.1038/s41586-022-05577-1 H. Y. Kim1, M. Garg2, S. Mandal1, L. Seiffert1, T. Fennel1 & E. Goulielmakis1 ✉

Received: 25 June 2022

Accepted: 18 November 2022 Field emission of electrons underlies great advances in science and technology,
Published online: 25 January 2023 ranging from signal processing at ever higher frequencies1 to imaging of the atomic-
scale structure of matter2 with picometre resolution. The advancing of electron
Open access
microscopy techniques to enable the complete visualization of matter on the native
Check for updates spatial (picometre) and temporal (attosecond) scales of electron dynamics calls for
techniques that can confine and examine the field emission on sub-femtosecond time
intervals. Intense laser pulses have paved the way to this end3,4 by demonstrating
femtosecond confinement5,6 and sub-optical cycle control7,8 of the optical field
emission9 from nanostructured metals. Yet the measurement of attosecond electron
pulses has remained elusive. We used intense, sub-cycle light transients to induce
optical field emission of electron pulses from tungsten nanotips and a weak replica of
the same transient to directly investigate the emission dynamics in real time. Access
to the temporal properties of the electron pulses rescattering off the tip surface,
including the duration τ = (53 as ± 5 as) and chirp, and the direct exploration of
nanoscale near fields open new prospects for research and applications at the interface
of attosecond physics and nano-optics.

The interaction of atoms and molecules with intense laser fields gives their generation and their measurement methodologies. On the genera-
rise to attosecond electron pulses10 that can study the structure and tion side, the driving laser pulses should be both short and intense so as
dynamics of these systems on recollision with their parent ion11. Atto- to confine the tunnelling of electrons into a sub-femtosecond window
second techniques12 can now gain access to the temporal profile of the (<1 fs), as well as to impart the recolliding electron pulses with a de Broglie
recolliding electron pulses and concomitant structural dynamics13,14 wavelength (<2.74 Å, >20 eV) that allows atomic-scale investigation
in their parent ions by measuring the transient properties of high of the parent surface28. On the measurement side, in situ attosecond
harmonics15 emitted during the interaction. Studies of the interac- metrologies24–27 shall be extended to incorporate temporal gating of
tion of intense laser fields with nanostructured metals over the past the optical field emission without relying on the concomitant high
two decades have suggested that the semiclassical concepts3–9,16,17 harmonic radiation. Measurements of this kind have so far permitted
developed earlier to describe electron dynamics in atoms can afford a access to the driving-field waveform of light waves by tracking the spec-
central role in the understanding of the optical field electron emission. trally integrated currents induced in the bulk of solids29,30 or the cutoff
In analogy to atoms, electrons set free from the apex of a nanotip at the energy variation of rescattered electrons in atoms31 and nanotips32, but
field crest of an intense laser pulse should also form ultrashort electron a direct time-resolved measurement of attosecond electron pulses in
pulses (Fig. 1a, inset), which—on recollision with the tip surface about the optical field emission has remained beyond reach.
three-quarters of the laser period (T ≈ 2 fs) later—could examine both Guided by the above provisions, we studied the field emission in
dynamics and structure. Owing to the ultrashort time interval between tungsten nanotips (work function ϕ ≈ 4.5–5.0 eV) using intense (around
emission and recollision events, and in contrast to other emerging 1013 W cm−2), sub-cycle (about 1.9 fs) optical transients (centroid
electron-pulse technologies18–21, the electron-pulse wave packet shall energy around 1.8 eV) generated in a light-field synthesizer33,34. The
undergo a negligible temporal spread, allowing its confinement to experiments were performed in a multifunctional experimental setup
sub-cycle timescales. (see Methods and Extended Data Fig. 1) that combines photoemission
However the real-time tracking of electron pulses generated in the spectroscopy of atoms and solids, optical pump–probe measurement
optical field emission has remained challenging. Although ordinary methodologies and EUV attosecond streaking35 for the sampling of the
attosecond streaking techniques can be used to map the temporal driving-field waveforms.
structure of the extreme ultraviolet (EUV) electron emission from metal We began our experiments by analysing the nonlinearity of the elec-
surfaces22 and nanotips23, they cannot directly investigate electron tron yield to the driving-field intensity. A plate detector (see Extended
pulses emerging in the optical field emission. Furthermore, the absence Data Fig. 2) recorded the total number of emitted electrons from the
of high harmonic emission from laser-driven nanotips constrains electrically grounded nanotip (yellow points, Fig. 1b) for a range of
the applicability of in situ attosecond techniques24–27 for examining peak intensities of the impinging laser transient. For the highest peak
the structure of the electron pulses in these systems. intensity of around 42 TW cm−2, we recorded up to about 1,000 elec-
Harnessing electron pulses emerging in the optical field emission for trons per pulse. Evaluation of the slope of the total electron yield versus
realizing new spectroscopies that combine attosecond temporal and peak intensity (purple line in Fig. 1b) showed an emission nonlinearity
nanometre spatial resolution calls for essential advancements on both of about 1.18 ± 0.09, which is well below the multiphoton threshold

Institut für Physik, Universität Rostock, Rostock, Germany. 2Max Planck Institute for Solid State Research, Stuttgart, Germany. ✉e-mail: e.goulielmakis@uni-rostock.de
1

662 | Nature | Vol 613 | 26 January 2023


a W b fitting (black and grey dashed lines in Fig. 1c) of the experimental data
(exp)
yielded s (exp)

Number of electrons per pulse


W,high = dEc /dUp = 118± 5.1 and s W,low = dEc /dUp = 24.1± 1.35,
(i) 103 Slope = 1.18 ± 0.09 1,000
800
(ii)
Nickel-coated
600 respectively. Here Up is the ponderomotive energy. By taking the ratio
400
dual mirror between high and low cutoff energy slopes (about 4.91 ± 0.35), we find
{

Electron
pulse
200 that it is compatible with that anticipated for backscattered and direct
(iii)
lse 102 100 electrons (10Up/2Up ≈ 5) in the single-electron approximation36.
pu
ptic
al An experimental assessment of the near-field enhancement in
cleo the vicinity of the nanotip and its comparison with the theoretical
-cy
gle 5 10 20 30 40
Sin Intensity (TW cm–2) predictions could further attest to the validity of single-electron
c
Number of electrons per pulse semiclassical pictures to describe emission under our experimental
200 400 600 800 1,000
250 0 conditions. To this end, we compared the emission spectra of electrons
Electron
spectrometer 200 W nanotip
–0.5 from the nanotip with those in a low-density neon gas. Figure 1d con-
W nanotip
–1.0 trasts electron spectra recorded from W nanotip (red curve) and neon
Energy (eV)

150 –1.5 atoms (blue curve) under identical field waveforms. Figure 1e shows
d 100
–2.0
electron spectra emanating from Ne over a broad range of driving-pulse
100 W nanotip intensities. A linear fitting (Fig. 1e, dashed line) of the cutoff energy
50
Spectral intensity (a.u.)

Ne gas (Fig. 1e, points) of the neon spectra versus intensity yielded a slope
10 20 30 40 s (exp)
Ne = 9.88 ± 0.34 that agrees well with the semiclassical predictions
e Intensity (TW cm–2)
10–1
100
under our driving waveforms (sNe = 10.8) (see Methods). The field-
Energy (eV)

80 Ne gas enhancement factor in the vicinity of the tungsten nanotip, which is


60
(exp)
40 sW,high
20 now evaluated as f = = 3.46± 0.10, shows a fair agreement with
10–2 s(exp)
Ne
50 100 150 200 250 40 60 80 100
Energy (eV) Intensity (TW cm–2) the theoretical prediction, fth ≈ 3.8 (Methods). This comparison further
attests to the compatibility of the emission processes from the nanotip
Fig. 1 | Optical field emission by intense, sub-cycle optical transients.
with single-electron, semiclassical concepts and suggests that multi-
a, Simplified schematic of the experimental setup. A sub-cycle pulse (orange
curve) is spatially separated and focused by a dual concave Ni mirror module.
electron charge interactions remain negligible.
A time delay between the pulses reflected by the inner and outer mirrors is Detailed semiclassical simulations based on the experimentally
introduced by a piezo stage. Tungsten nanotips (apex radius around 35 nm) or derived quantities (see Methods) further support the above perspec-
a gas jet of neon atoms can be positioned in the laser focus. Emitted electron tive. The simulations accurately reproduce the experimental spectra
spectra are recorded by a time-of-flight spectrometer (acceptance angle about (Extended Data Fig. 4) and the association of high and low cutoff ener-
6°) placed roughly 3 mm downstream of the electron source and aligned along gies of the emission to backscattered and direct electrons, respectively.
the laser polarization axis. The inset shows that electrons marked by the green The inclusion of multielectron interactions (Extended Data Fig. 6)
shaded curve are (i) set free and accelerated by the intense laser field to form showed the presence of weak-only effects of space–charge distortions
an electron pulse, which, on recollision with the nanotip surface, (ii) can on the emitted spectra and the spectral phase of the electron emission
investigate both dynamics as well as structure. On backscattering off the tip further corroborating the appropriateness of single-electron models
surface, (iii) the electron pulse is further accelerated by the laser to escape the
to describe the emission process.
interaction area. b, Total electron yield per pulse as a function of increasing
The absence of discernible cutoffs at intermediate energies (60–
peak intensity of the driving laser pulses (yellow points) and its linear fitting
120 eV) in both experiments (Fig. 1c) and simulations (Extended Data
(purple line) on a logarithmic scale. c, Electron spectra from the tungsten
nanotip (false colour in logarithmic scale) versus peak intensity. Stars and dots
Figs. 4 and 6) supports the notion that, under sub-cycle driving, the
denote the cutoff energies. Black and grey dashed lines show the linear fitting high-energy part of the emitted spectrum is associated with the recol-
of the cutoff energy versus incident peak intensity of the laser pulse. d, Optical lision of a single electron pulse at the tip surface. A weak, low-energy
emission electron spectra from the tungsten nanotip (red curve) and neon backscattered emission indicated by the simulations (Extended Data
atoms (blue curve) for nearly identical peak intensity (about 40 TW cm−2). Fig. 4a) is not directly resolved in the experiments (Fig. 1c). Yet the
e, Same as in c but for neon atoms. presence of such emission channels will become apparent later in this
work by time resolving the optical field emission.

ϕ/ħωL ≈ 3 and supports the notion that field-driven electron tunnel- Homochromatic attosecond streaking
ling dominates the ionization of tungsten over the entire range of the As vacuum is dispersive to electron pulses, a temporal characteriza-
studied intensities. tion has a concrete meaning at a specific point in space. Because the
Next we conducted a spectral-domain study of the emission under generated electron pulses scrutinize the ‘sample’ during the recol-
precisely characterized driving fields and interrogated compatibility lision with the parent surface as to their temporal structure at this
of our findings with the predictions of semiclassical, single-electron point, it is mostly relevant for harnessing the power of these pulses in
models. Given the multielectron nature of the emission (Fig. 1b), this time-resolved applications.
step is essential for applying semiclassical single-electron concepts To understand how, we revisit the process of strong field recollision
later in this work to scrutinize the temporal structure of the electron of an electron wave packet under an intense optical waveform (Fig. 2).
emission. As space–charge interactions can substantially distort the Set free around the peak of a laser field crest, an electron pulse will
emission spectra, in comparison with semiclassical predictions, a recollide with the tip surface at an instance tr (Fig. 2a) with an energy
detailed interrogation of the cutoff energies of the emission can shed of about 3Up (refs. 11,37). An attosecond measurement of the electron
light onto potential multielectron contributions. pulse entails access into its waveform ψr(t) at the surface of the nanotip

Electron spectra recorded as a function of the peak intensity of the or equivalently, into its associated complex spectral amplitude ψr(p),
driving pulse (Fig. 1c) showed the formation of two well-discernible in which p is the recollision momentum of the electron. However, as

cutoffs (black points, grey points and false-colour plot in Fig. 1c, respec- ψr(p) is not directly accessible in measurements, it is important to link
tively), whose energies scale linearly with peak intensity. Evaluation of it to other measurable quantities. Following backscattering off the
the corresponding slopes for high and low cutoff energies (Ec) by linear tip surface, the wave packet acquires further phase from both its

Nature | Vol 613 | 26 January 2023 | 663


Article
a a b
~
–e[Ap(t) + AHAS(t + W)] |\t(p, W)|2 5.0

Vector potential (×10–2 V fs Å–1)


15
200
–eAp(t) –2 –1 0
10
Momentum variation

~TL 2.5
–eAHAS(t + W) 150 5

Energy (eV)
~
|\t(p)|2

Momentum
0 0
100
–5
–eAp(t) –2.5
50 –10
HAS
–15 EUV-AS
~ –5.0
|\r(t)|2 |\r(p)|2
–6 –4 –2 0 2 4 6 –6 –4 –2 0 2 4 6
Delay (fs) Time (fs)
c d

Near field (V Å–1)


5
Time Intensity 200 200
b –2 –1 0 0
~TL
200 150 150 –5

Energy (eV)

Energy (eV)
0 0.5 1.0
Energy (eV)

0
150 100 100
–1
50 50
100
–2
–5 0 5 –5 0 5 –5 0 5
Delay (fs) Delay (fs) Delay (fs) –6 –4 –2 0 2 4 6 –6 –4 –2 0 2 4 6
Delay (fs) Time (fs)

Fig. 2 | Homochromatic attosecond streaking. a, An electron wave packet


Fig. 3 | Homochromatic attosecond streaking of the optical field emission.
ψr(t), sampled on recollision at the nanotip surface, is represented by its
∼ ∼ 2 a, Homochromatic attosecond streaking (HAS) spectrogram of the optical
momentum distribution ψr (p ) and corresponding spectral intensity ∣ψr (p )∣ . field emission from a tungsten nanotip consisting of 60 individual electron
Following backscattering, further acceleration of the wave packet by the spectra recorded as a function of the delay (step size of 200 as) between an
vector potential of the pump field Ap(t) (red curve) gives rise to a terminal intense (Ip ≈ 25.4 TW cm−2) single-cycle pulse and a 6.3 × 10 −3 times weaker gate
∼ ∼ 2
wave packet ψt(p ) with corresponding spectral intensity I (p ) = ∣ψt(p )∣ . The pulse. b, Vector potential Ag(t) evaluated by the photoelectron energy cutoff
coherent superposition (dashed yellow curve) of the pump field with a weak, variation in the HAS spectrogram in panel a as a function of the delay (blue)
time-delayed (τ) replica (the gate) results in a momentum shift, −eAHAS(t + τ) of and from an ordinary streaking spectrogram (red) (see Extended Data Fig. 9)
∼ 2
the final momentum distribution I (p , τ ) = ∣ψt(p , τ )∣ . b, Simulated delay- in absolute units. c, Simulated HAS spectrogram using the experimentally
dependent homochromatic attosecond streaking (HAS) spectrograms for recorded field waveforms of pump and gate pulses. The dashed vertical lines
electron pulses without chirp (left panel), positive chirp of 3.5 × 103 as2 (middle in a and c denote different rescattering events. d, Terminal energies of direct
panel) and third-order chirp of 1 × 105 as3 (right panel). The dashed white curves (green) and backscattered electron emissions (purple) associated (dashed
highlight the resulting energy-dependent shifts of the intensity modulation. horizontal lines) with corresponding cutoff energies in the spectrogram (c),
as well as release times (dashed vertical lines) near-field pump waveform (red
curve). EUV-AS, extreme ultraviolet attosecond streaking.

interaction with the driving field (Volkov phase) as well owing to its
free-space propagation. If we define an auxiliary terminal wave packet
∼ ∼
ψt(p), the spectral intensity I (p) = ∣ψt(p)∣2 (refs. 38–40) (as marked in

Fig. 2a) can be expressed as (in atomic units): in which ψt(t) is the inverse Fourier transform of ψt(p) and AHAS(t + τ)
represents an effective vector potential, which is explicitly related to
2
∼ ∞  p2  the incident vector potential Ag(t) of the gate pulse as shown in Methods
I (p) = |ψt(p)|2 = ∫−∞ ψr(tr)expi 2 tr exp[−iS(p, tr ; Ap (t))]dtr (1)
and which accounts for the momentum an electron accumulates from
 
the instance of its birth to the detection. Equation (2) implies: (i) a vari-
ation of τ permits the composition of a streaking-like spectrogram whose
∞ 1
in which S(p, tr; Ap (t)) = ∫  2 [p + Ap (t)]2 − 2 p 2  dt denotes the Volkov
1 ∼
tr   reconstruction can allow retrieving the phase of ψt(p) and (ii) the momen-
phase imparted to the electron wave packet by the vector potential Ap(t) tum variation of the electron distribution follows AHAS(t). Whereas impli-
of the intense driving pulse (hereafter referred to as pump, see red curve cation of (ii) (see Methods) allows sampling of the waveform of a light
in Fig. 2a) after rescattering at time tr. Hence reconstruction of the recol- pulse31, (i) is essential for mapping the dynamics of the field emission.
liding wave packet ψr(t) should be possible if, other than I(p), which is To distinguish from conventional attosecond streaking, we refer to this
a directly measurable quantity (that is, the spectrum of the electron approach as homochromatic attosecond streaking (HAS), highlighting

emission), the phase of ψt(p) as well as the waveform Ap(t) are accessed. that the carrier frequency of the pump and gate fields is identical.

Access to the phase of ψt(p) is possible by temporally gating the elec- Shown in Fig. 2b are simulated HAS spectrograms under conditions
tron emission with a weak replica of the pump pulse (hereafter referred pertinent to the experiments presented here. Notably, and in close anal-
to as gate with vector potential Ag(t)) when η ≡ ∣Ag (t)∣2 / ∣Ap (t)∣2 ≪ 1 ogy to ordinary attosecond streaking41,42, different types of chirp of the
(refs. 27,31). In this limit, the pump pulse is solely responsible for releas- recolliding electron pulse ψr(t) yield distinct visual signatures in the spec-
ing the electron wave packet, whereas the gate pulse primarily alters trogram, manifested as shifts of the spectral intensity modulation versus
its phase. This is manifested by the shift and reshaping of the terminal delay and energy, as highlighted by the dashed white curves in Fig. 2b.
photoelectron spectra at the end of the driving pulse (Fig. 2a). If the In our experiments, we derived pump and gate pulses by the
delay τ between pump and gate pulses is varied (see Methods), the spatiotemporal division of the sub-cycle optical transients using a

terminal spectral distribution of the released electron, I (p, τ) = ∣ψt(p, τ)∣2 dual-mirror module as shown in Fig. 1a. Figure 3a plots a HAS spectro-
can be approximated as: gram recorded by our apparatus. A ratio η ≈ 6.3 × 10−3 warrants that the
gate pulse is sufficiently weak to serve as a phase gate (see Methods),
2 whereas a remaining weak amplitude modulation (5–10%) of the spec-
∞  p2  trogram versus delay was useful for evaluating the timing between
I (p, τ) ∝ ∫−∞ ψt (tr )exp i tr  exp[− iS (p, tr ; AHAS (t))]dtr (2)
 2  pump and gate pulses, and therewith to clock the recollision instance
with respect to the waveform of the driving pump pulse.

664 | Nature | Vol 613 | 26 January 2023


a b c d
Experiment Reconstruction
0 0.5 1.0 0 0.5 1.0 1.5 4
1.0 1.0
200 1.0 2
0.8 0.8

Intensity (a.u.)

Phase (rad)
Phase (rad)
Intensity (a.u.)
Energy (eV)

0
0.6 0.5 0.6 53 ± 5 as
150 –2
0.4 0 0.4
–4
0.2 0.2
100 –0.5
–6
0 0
–6 –4 –2 0 2 4 6 –6 –4 –2 0 2 4 6 20 40 60 80 –0.4 –0.2 0 0.2 0.4
Delay (fs) Delay (fs) Energy (eV) Time (fs)
e f
100

Intensity (a.u.)
5.0 1.0
0 0.5 1.0
80
Near field (V Å–1)

2.5 0.5

Energy (eV)
60
0 0
40
–2.5
20
–5.0
0
–6 –4 –2 0 2 4 6 –0.5 –0.4 –0.3 –0.2 –0.1 0 0.1 0.2 0.3 0.4 0.5
Time (fs) Time (fs)

Fig. 4 | Measurement of attosecond electron pulses in the optical field (magenta fill). f, Time–frequency analysis of the attosecond electron pulse
emission. a,b, Experimentally recorded (a) and reconstructed (b) HAS (false colour is intensity in arbitrary units) and retrieved release times (black
spectrograms from a W nanotip. c, Retrieved backscattered electron-pulse curve). The dashed red curve denotes release times calculated semiclassically
spectrum (magenta fill) and its spectral phase (red curve). d, Intensity profile for the near-field light waveform in e. Error bars represent standard deviations
of the electron pulse (magenta fill) and its temporal phase (red line). e, Near of the corresponding values indicated.
field (red curve) and its timing with respect to the attosecond electron pulse

The evaluated Ag(t) (blue curve in Fig. 3b) by HAS exhibits an excel- reconstruction of data in Fig. 4a based on equations (1) and (2), the
lent waveform matching that derived by EUV attosecond streaking (red retrieved field parameters Ap(t) and Ag(t) (Fig. 3b), the absolute time
curve in Fig. 3b), as attested by the degree of similarity43 (around 0.95). delay τ (Extended Data Fig. 8) and the numerical algorithm detailed
In this case, the HAS-evaluated vector potential represents the near field in Methods.
of the gate pulse in the vicinity of the tip. Moreover, the ratio between The retrieved electron-pulse profiles in spectral and time domains
the absolute amplitudes of the vector potentials evaluated by the two are shown in Fig. 4c,d, respectively. The spectrum extends over the
methods provides a direct measurement of the field-enhancement range 20–80 eV (Fig. 4c, magenta-shaded curve) and is temporally
factor f = 3.73 ± 0.25, which is in close agreement with the result of the confined to about 53 ± 5 as, as measured at the full width at half maxi-
methodology highlighted in Fig. 1c (conducted with a different nanotip) mum of its intensity profile (Fig. 4d, magenta-shaded curve). Figure 4e
and the theoretical estimations. compares the retrieved electron pulse and the near-field profiles, and
How time-domain measurement of electron emission in nanostruc- shows that the electron wave packet revisits the surface of the nano-
tured materials benefits the intuitive understanding of the process and tip (magenta-shaded curve) at times close to the zero transition of
its real-time clocking can be best appreciated by a close inspection of the the driving field (red curve). This observation is compatible with the
oscillating phases of high and low cutoff energies in Fig. 3a. For instance, semiclassical understanding of strong-field rescattering in atoms11,36.
an apparent delay of the maxima of the corresponding oscillations A closer inspection of the retrieved spectral and temporal phases
(highlighted by red and blue dashed curves, respectively, in Fig. 3a) (red curves in Fig. 4c,d) also shows a negligible temporal spread of
versus delay indicates a retardation in the electron emission at lower the electron pulse compared with its Fourier-limited duration (about
energies by approximately a laser cycle (TL ≈ 2.3 fs). This feature, also 50 as). The compatibility of this finding with the semiclassical model
well reproduced in our simulation (Fig. 3c), shows that the low-energy can be best appreciated by comparing the group delay of the electron
emission consists of a mixture of direct electrons emerging within the release (solid black line in Fig. 4f) as evaluated by the time–frequency
main half-cycle of the driving field (green dots in Fig. 3d) and low-energy analysis (Fig. 4f, false colour) of the attosecond electron-pulse wave-
backscattered electrons generated approximately a cycle of the driving form in Fig. 4d with the semiclassical release times (dashed red curve
field later (purple dots in Fig. 3d). Moreover, the uniform amplitude and in Fig. 4f) calculated using the near-field waveform of the driving pulse
energy modulation of the photoelectron spectrum (Fig. 3a) over a broad (red curve in Fig. 4e). The comparison highlights that, as anticipated
range of energies (50–150 eV) offers compelling evidence that emission by the semiclassical recollision model, electrons of energy close to the
is confined to a single electron pulse generated within a field half-cycle. cutoff, that is, where short and long trajectories merge, carry a minor
For the reconstruction of the electron wave packet, we are primarily temporal chirp. This feature, common in near-cutoff high harmonic
interested in the properties of the recolliding electron pulses at ener- emission in atoms24,41, further strengthens the link between strong-field
gies typically higher than 20 eV, that is, where this pulse could serve optics in atoms and nano-optics systems.
high-resolution/atomic-scale, spatial analysis of surfaces28. Considering
the ratio of the terminal energy of electrons (about 10Up) and that at
the recollision instance (about 3.17Up), we conclude that the relevant Conclusion
information is encoded at the high-energy end (>80 eV) of the spec- Direct measurement of attosecond electron pulses in the optical field
trogram of Fig. 3a, as isolated in Fig. 4a. Figure 4b shows the numerical emission extends the repertoire of ultrafast-science tools. The technique

Nature | Vol 613 | 26 January 2023 | 665


Article
presented here could offer a broad range of capabilities for tracking 21. Priebe, K. E. et al. Attosecond electron pulse trains and quantum state reconstruction in
ultrafast transmission electron microscopy. Nat. Photon. 11, 793–797 (2017).
ultrafast collective or correlated electron dynamics in various materials 22. Cavalieri, A. L. et al. Attosecond spectroscopy in condensed matter. Nature 449, 1029–1032
deposited or attached to tungsten nanotips44,45, as well as in gas-phase (2007).
systems. The recolliding electron pulses shall also enable attosecond- 23. Förg, B. et al. Attosecond nanoscale near-field sampling. Nat. Commun. 7, 11717 (2016).
24. Dudovich, N. et al. Measuring and controlling the birth of attosecond XUV pulses. Nat.
resolved, laser-induced electron diffraction46 experiments at nanotips, Phys. 2, 781–786 (2006).
as well as nanodiffraction ex situ experiments in specimens placed at 25. Shafir, D. et al. Resolving the time when an electron exits a tunnelling barrier. Nature 485,
nanometric distance from the tip apex47. Such possibilities could enable 343–346 (2012).
26. Kim, K. T. et al. Manipulation of quantum paths for space–time characterization of
new routes for exploring the structure and dynamics of condensed attosecond pulses. Nat. Phys. 9, 159–163 (2013).
matter in four dimensions. 27. Kim, K. T. et al. Petahertz optical oscilloscope. Nat. Photon. 7, 958–962 (2013).
28. Van Hove, M. A., Weinberg, W. H. & Chan, C.-M. Low-Energy Electron Diffraction:
Experiment, Theory and Surface Structure Determination Vol. 6 (Springer, 2012).
29. Schiffrin, A. et al. Optical-field-induced current in dielectrics. Nature 493, 70–74 (2013).
Online content 30. Rybka, T. et al. Sub-cycle optical phase control of nanotunnelling in the single-electron
Any methods, additional references, Nature Portfolio reporting summa- regime. Nat. Photon. 10, 667–670 (2016).
31. Kim, Y. H. et al. Attosecond streaking using a rescattered electron in an intense laser field.
ries, source data, extended data, supplementary information, acknowl- Sci. Rep. 10, 22075 (2020).
edgements, peer review information; details of author contributions 32. Hoff, D. et al. Tracing the phase of focused broadband laser pulses. Nat. Phys. 13, 947–951
and competing interests; and statements of data and code availability (2017).
33. Wirth, A. et al. Synthesized light transients. Science 334, 195 LP–200 (2011).
are available at https://doi.org/10.1038/s41586-022-05577-1. 34. Hassan, M. T. et al. Optical attosecond pulses and tracking the nonlinear response of
bound electrons. Nature 530, 66–70 (2016).
35. Goulielmakis, E. et al. Direct measurement of light waves. Science 305, 1267–1269
1. Han, J.-W., Moon, D.-I. & Meyyappan, M. Nanoscale vacuum channel transistor. Nano Lett. (2004).
17, 2146–2151 (2017). 36. Paulus, G. G., Becker, W., Nicklich, W. & Walther, H. Rescattering effects in above-threshold
2. Brodusch, N., Demers, H. & Gauvin, R. Field Emission Scanning Electron Microscopy: New ionization: a classical model. J. Phys. B At. Mol. Opt. Phys. 27, L703–L708 (1994).
Perspectives for Materials Characterization (Springer, 2018). 37. L’Huillier, A. et al. High-order harmonic-generation cutoff. Phys. Rev. A 48, R3433–R3436
3. Hommelhoff, P. & Kling, M. Attosecond Nanophysics: From Basic Science to Applications (1993).
(Wiley, 2015). 38. Lewenstein, M., Balcou, P., Ivanov, M. Y., L’Huillier, A. & Corkum, P. B. Theory of
4. Dombi, P. et al. Strong-field nano-optics. Rev. Mod. Phys. 92, 25003 (2020). high-harmonic generation by low-frequency laser fields. Phys. Rev. A 49, 2117–2132 (1994).
5. Hommelhoff, P., Sortais, Y., Aghajani-Talesh, A. & Kasevich, M. A. Field emission tip 39. Lewenstein, M., Kulander, K. C., Schafer, K. J. & Bucksbaum, P. H. Rings in above-threshold
as a nanometer source of free electron femtosecond pulses. Phys. Rev. Lett. 96, 77401 ionization: a quasiclassical analysis. Phys. Rev. A 51, 1495–1507 (1995).
(2006). 40. Milošević, D. B., Paulus, G. G., Bauer, D. & Becker, W. Above-threshold ionization by
6. Ropers, C., Solli, D. R., Schulz, C. P., Lienau, C. & Elsaesser, T. Localized multiphoton few-cycle pulses. J. Phys. B At. Mol. Opt. Phys. 39, R203–R262 (2006).
emission of femtosecond electron pulses from metal nanotips. Phys. Rev. Lett. 98, 43907 41. Goulielmakis, E. et al. Single-cycle nonlinear optics. Science 320, 1614–1617 (2008).
(2007). 42. Hofstetter, M. et al. Attosecond dispersion control by extreme ultraviolet multilayer
7. Krüger, M., Schenk, M. & Hommelhoff, P. Attosecond control of electrons emitted from a mirrors. Opt. Express 19, 1767–1776 (2011).
nanoscale metal tip. Nature 475, 78–81 (2011). 43. Hecht, E. Optics (Pearson, 2017).
8. Piglosiewicz, B. et al. Carrier-envelope phase effects on the strong-field photoemission of 44. Tafel, A., Meier, S., Ristein, J. & Hommelhoff, P. Femtosecond laser-induced electron
electrons from metallic nanostructures. Nat. Photonics 8, 37–42 (2014). emission from nanodiamond-coated tungsten needle tips. Phys. Rev. Lett. 123, 146802
9. Bormann, R., Gulde, M., Weismann, A., Yalunin, S. V. & Ropers, C. Tip-enhanced strong- (2019).
field photoemission. Phys. Rev. Lett. 105, 147601 (2010). 45. Li, C. et al. Extreme nonlinear strong-field photoemission from carbon nanotubes. Nat.
10. Niikura, H. et al. Sub-laser-cycle electron pulses for probing molecular dynamics. Nature Commun. 10, 4891 (2019).
417, 917–922 (2002). 46. Meckel, M. et al. Laser-induced electron tunneling and diffraction. Science 320, 1478–1482
11. Corkum, P. B. Plasma perspective on strong field multiphoton ionization. Phys. Rev. Lett. (2008).
71, 1994–1997 (1993). 47. Garg, M. & Kern, K. Attosecond coherent manipulation of electrons in tunneling
12. Krausz, F. & Ivanov, M. Attosecond physics. Rev. Mod. Phys. 81, 163–234 (2009). microscopy. Science 367, 411–415 (2020).
13. Baker, S. et al. Probing proton dynamics in molecules on an attosecond time scale.
Science 312, 424–427 (2006). Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
14. Wolter, B. et al. Ultrafast electron diffraction imaging of bond breaking in di-ionized published maps and institutional affiliations.
acetylene. Science 354, 308–312 (2016).
15. Ferray, M. et al. Multiple-harmonic conversion of 1064 nm radiation in rare gases. J. Phys. Open Access This article is licensed under a Creative Commons Attribution
B At. Mol. Opt. Phys. 21, L31–L35 (1988). 4.0 International License, which permits use, sharing, adaptation, distribution
16. Zherebtsov, S. et al. Controlled near-field enhanced electron acceleration from dielectric and reproduction in any medium or format, as long as you give appropriate
nanospheres with intense few-cycle laser fields. Nat. Phys. 7, 656–662 (2011). credit to the original author(s) and the source, provide a link to the Creative Commons licence,
17. Herink, G., Solli, D. R., Gulde, M. & Ropers, C. Field-driven photoemission from and indicate if changes were made. The images or other third party material in this article are
nanostructures quenches the quiver motion. Nature 483, 190–193 (2012). included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
18. Baum, P. & Zewail, A. H. Attosecond electron pulses for 4D diffraction and microscopy. to the material. If material is not included in the article’s Creative Commons licence and your
Proc. Natl Acad. Sci. 104, 18409–18414 (2007). intended use is not permitted by statutory regulation or exceeds the permitted use, you will
19. Hassan, M. T., Baskin, J. S., Liao, B. & Zewail, A. H. High-temporal-resolution electron need to obtain permission directly from the copyright holder. To view a copy of this licence,
microscopy for imaging ultrafast electron dynamics. Nat. Photon. 11, 425–430 (2017). visit http://creativecommons.org/licenses/by/4.0/.
20. Morimoto, Y. & Baum, P. Diffraction and microscopy with attosecond electron pulse
trains. Nat. Phys. 14, 252–256 (2018). © The Author(s) 2023

666 | Nature | Vol 613 | 26 January 2023


Methods robustness of the nanotip on the timescale of typical measurements
(a few minutes) as well as over several hours.
Experimental
Attosecond EUV streaking. For attosecond EUV streaking measure- One-dimensional, semiclassical simulations of the optical field
ments (Extended Data Fig. 1a), the sub-cycle transients are focused emission
onto the neon gas jet to generate EUV pulses by high-harmonic genera- The time-dependent ionization probability from a tungsten nanotip
tion. The collinearly propagating EUV and optical pulses are spatially was approximated by the Fowler–Nordheim formula as5,17,49,50:
separated by a Zr filter, which also acts as an EUV high-pass spectral
filter (>70 eV), enabling the isolation of a single attosecond pulse. The  4 2m ϕ 3/2 
p(E (t)) = Nθ(− E (t)) ∣E (t)∣2 exp  − , (3)
EUV and optical pulses are reflected off a dual-mirror assembly, which  3ħe ∣E (t)∣ 
consists of a Mo/Si inner mirror (centred at around 85 eV) and a nickel
outer mirror, respectively. Inner and outer mirrors can be delayed with in which E(t) is the electric-field waveform of the driving pulses, ϕ is the
nanometric resolution (Extended Data Fig. 1a). EUV and optical pulses work function of the metal and m, ħ and e are electron mass, reduced
are focused onto a second Ne gas jet. Photoelectron spectra recorded Planck’s constant and electron charge, respectively. We calculated
as a function of the delay between the inner and outer mirrors allow electron trajectories using the classical equations of motion in the
the composition of attosecond streaking spectrograms, which allow single-electron limit as7,8,11,17,36:
the detailed characterization of the attosecond EUV pulse and, notably
for these experiments, the field waveform of the optical pulse. Details d2 zi
m = − ef0 E (t) (4)
about the relevant techniques can be found in refs. 35,41,48. dt 2

HAS. HAS measurements are performed on the same setup by means Here i is the index of each individual trajectory and f0 is the field-
of (1) automated removal of the Ne gas used to generate high harmon- enhancement factor. At the end of the pulse, an electron spectrum
ics and the Zr filter, (2) automated replacement of the inner mirror in is evaluated by a spectral binning of the energies of all trajectories
the dual-mirror module by a Ni-coated one of the same focal lengths weighted by the ionization rate at the instances of their births.
(Extended Data Fig. 1b) and (3) the streaking gas nozzle is replaced by an For the experiments described here, we simulated electron spectra
electrically grounded tungsten nanotip. The above setup modifications from tungsten (ϕ = 4.53 eV) versus peak field intensity of the driving
are executed in a fraction of a minute and warrant identical conditions pulse (Extended Data Fig. 4). The driving field (red curve in Extended
for all relevant measurements. In the HAS configuration of the setup Data Fig. 4a) used in our simulations was measured by an EUV attosecond
in Extended Data Fig. 1b, Ni-coated inner and outer mirrors spatially streaking setup35,48 integrated in our apparatus. A field-enhancement
divide the optical pulse into pump (inner mirror beam) and gate (outer factor of f0 = 3.46 used in these simulations was derived experimen-
mirror) pulses. A delay between the pulses reflected off the inner and tally as described in the main text. The quiver length of electrons in
outer mirrors, respectively, is introduced by a precision transitional our intensity range (the longest length of about 1.7 nm at the highest
stage (see inset in Extended Data Fig. 1b). The HAS measurements as intensity of about 41 TW cm−2) is considerably shorter than the decay
well as the intensity-scaling measurements of the electron spectrum length of the near field (about 30 nm). Hence the released electrons
and yields were performed with driving-field intensities below a criti- experience a nearly homogeneous near field and therefore quenching
cal intensity, at which the irreversible optical damage of the nanotips effects17 owing to the near-field decay can be neglected.
occurs. This upper limit of intensity was experimentally identified by In agreement with the data of Fig. 1c, the simulated electron spectra
observing sudden and irreversible drops of electron cutoff energy in exhibit two well-discernible energy cutoffs (red and blue dashed lines
intensity-scaling measurements. in Extended Data Fig. 4b) associated with the backscattered (purple
line in Extended Data Fig. 4a) and the direct (green line in Extended
Measurement of the absolute electron yield in the optical field Data Fig. 4a) electrons. The slopes of high ( s (th)
W,high = dEc /dUp = 130.3 )
emission. For the measurement of the total electron counts per pulse and low (s (th)
W,low = 26.1 ) cutoff energies agree well with those in our meas-
generated in our setup, a thin electrode (size roughly 5 mm × 5 mm) is urements (Fig. 1c). The theory shows further emission cutoffs at ener-
introduced about 2 mm above the nanotip (Extended Data Fig. 2). This gies lower than that of the direct electrons. Because these are relatively
configuration allows detection of released electrons over a solid angle weak, they do not leave any direct signatures in the photoelectron
Ω > π steradians. The induced voltage on the thin plate is measured by spectra. Yet such contributions become visible in HAS spectrograms,
a lock-in amplifier at the reference frequency of the repetition rate of as discussed in Fig. 3.
the driving laser (about 3 KHz). The electronic current is evaluated by
dividing the induced voltage by the system impedance (10 MΩ). The FDTD simulations of the field enhancement
total electron count per pulse is in turn obtained by dividing the current To theoretically estimate the near-field enhancement in the vicinity of
by the repetition rate of the laser and the electron charge. the tungsten nanotip, we numerically solved Maxwell’s equations by
means of three-dimensional finite-difference time-domain (FDTD) sim-
Sample robustness and measurement stability study. Tungsten ulations. The nanotip was modelled as shown in Extended Data Fig. 5a
nanotips were exposed to intensities of up to I ≈ 45 TW cm−2 without as a sphere with radius of 35 nm that smoothly transitions to a cone with
noticing any damage. This was verified by performing the intensity– an opening angle (single side) of 15° and considering optical properties
cutoff energy study shown in Fig. 1b, of both increasing as well as de- for tungsten51. The simulations predict a peak field-enhancement factor
creasing intensity, and recorded identical curves. When the critical of about 3.8 close to the surface at the tip apex. For comparison, the
intensity reached I ≈ 45 TW cm−2, the tip is damaged and the cutoff spatial distribution of the enhancement at a respective tungsten nano-
energy irreversibly drops to much lower values without the possibility sphere (that is, excluding the cone) is shown in Extended Data Fig. 5b,
for recovery unless a new tip is installed. with a slightly lower peak enhancement factor of approximately 2.7.
To explore short-term and long-term stability of our system, we
recorded electron spectra as a function of time under identical con- Three-dimensional, semiclassical trajectory simulation
ditions and for intensities typically higher than those used in the HAS including charge interaction
measurements. Data shown in Extended Data Fig. 3 suggest an excellent To inspect whether charge interaction substantially affects the elec-
stability of cutoff energy and electron yield, implying the structural tron emission dynamics for the considered parameters, we performed
Article
semiclassical trajectory simulations using the Mie–mean field–Monte To investigate how this energy shift could potentially affect the
Carlo (M3C) model52. The latter has been used extensively for the study retrieval of the electron pulse at the instance of recollision, we applied
of strong-field ionization in dielectric nanospheres53–55 and has recently the simulated phase effects on our experimental data (see the section
also been adopted for the description of metallic nanotips56. In brief, we ‘HAS reconstruction methodology’) and evaluated once again the spec-
mimic the apex of the nanotip as one-half of a sphere with correspond- tral and temporal properties of the recolliding attosecond electron
ing radius. The near field is evaluated as the combined linear polariza- pulse. Key observations associated with space–charge interactions
tion field owing to the incident pulse (evaluated by means of the Mie include a weak, uniform shift of the central energy of the recolliding
solution of Maxwell’s equations) and an extra nonlinear contribution electron by a few electronvolts (Extended Data Fig. 6d) and a subtle
owing to charge interaction treated as a mean field in electrostatic change in the temporal profile of the electron pulse (Extended Data
approximation (evaluated by high-order multipole expansion). The Fig. 6e), resulting in an approximately 4-as elongation of its duration at
latter includes Coulomb interactions among the emitted electrons as the full width at half maximum, which is within the error of the experi-
well as their image charges (that is, an extra sphere polarization caused mental reconstruction (about 5 as).
by the free electrons). Photoelectron trajectories are generated by
Monte Carlo sampling of ionization events at the sphere surface, at Mathematical formulation of HAS
which we evaluate tunnelling probabilities within WKB approximation The key objective of HAS is to retrieve the temporal structure of an
by integration through the barrier provided by the local near field. attosecond electron pulse wave packet ψr(t) at the moment of its recol-
Trajectories are propagated in the near field by integration of classi- lision on its parent surface. As this wave function is not directly acces-
cal equations of motion and accounting for electron–atom collisions sible, it is necessary to link it to other quantities that are either directly
through respective scattering cross-sections for electrons moving measured in the experiments (such as the terminal spectral intensity

within the material. To mimic the slightly higher peak enhancement I (p) = ∣ψt(p)∣2 at a detector) or can be reconstructed from the experi-
of the linear response near field at a tungsten tip (≈3.8) as compared mental data.
with a sphere (≈2.7), see Extended Data Fig. 5, we rescaled the incident
laser intensity by a factor of 1.4. The performed M3C simulations predict Description of strong-field electron emission. Considering an elec-
about three times fewer emitted electrons than the experiments, which tron released from and driven back to a surface by a strong pump field
we attribute to contributions of slow electrons originating from the Ep(t), its recolliding wave packet ψr(t) can be linked to its terminal spec-

shank of the nanotip. This is substantiated by comparison of the total tral amplitude ψt(p) at the end of interaction with the driving pulse.
electron yields predicted for the nanotip and the half-sphere, obtained The time-dependent recollision wave packet ψr(t) is expressed
∼ ∞ 2
through integration of the local ionization rates over the respective by means of its Fourier representation ψr(p) = ∫ −∞ ψr(tr)exp[i p2 tr ]dtr and,
surface areas and the pulse duration. However the charge densities following the recollision, the spectral amplitude is transformed to
at the pole of the sphere and the tip apex are comparable, enabling to the terminal form38–40,58 (in atomic units):
inspect the impact of charge interactions within the simplified simu-
lation model. Extended Data Fig. 6a compares multielectron spectra ∼ ∞  p2 
simulated with (solid curves) and without (dashed curves) accounting ψt(p) ∝ i ∫ ψr(tr)exp i tr exp[−iS(p, ∞, tr; Ap (t))]dtr (5)
−∞
 2 
for charge interactions among the individual electrons and for four
representative settings of the peak intensities (and corresponding Here S is the Volkov phase imparted to the electron wave packet only
electron yields) of the driving field, as indicated in the legend. The by the vector potential Ap(t) of the pump pulse after recollision at an
presence of charge interaction is primarily manifested by a noticeable instance tr, at which the general form of the Volkov phase accumulated
decline of the yield of direct electrons (<50 eV), which—in accordance from a time instance t1 to a later instance t2 by an electric field with vec-
with earlier works52,57—can be attributed to a partial trapping of these tor potential A(t) is expressed as59:
electrons in the vicinity of the tip. Quasi-static electric fields generated
by the trapping of low-energy electrons in turn affect the dynamics of t2
1 1 
S(p, t2, t1; A(t)) = ∫  [p + A(t)] − p 2  dt
2
the recolliding electrons and give rise to an increase of the terminal (6)
t1  2 2 
electron-energy cutoff 52,57. For the highest intensity and corresponding
electron yield, this shift is approximately 8% in energy. Note that equation (6) excludes free-space propagation, that is,
Although transparent manifestations of such effects cannot be dis- it vanishes in the absence of the field, and equation (5) reflects the
cerned in our experiments, it is useful to understand potential impli- momentum-dependent wave function at the surface, including phases
cations of charge interactions on the characterization of attosecond accumulated only by the pump field.
electron pulses using HAS. To this end, we extended our analysis to the Earlier semiclassical theories of strong-field emission39,40,58 in atoms
time domain for pulses of intensity (about 31 TW cm−2) and electron have suggested that the recolliding wave packet ψr(t) can be expressed
yield (about 600 electrons per pulse), mimicking the experimental by integration over ionization instance t′ before recollision at time tr
conditions in our HAS measurements. and over canonical momenta p′ in terms of the ionization amplitude,
Extended Data Fig. 6b shows the recollision energy distribution of the dictated by the dipole transition Ep(t′)d(p′ + Ap(t)), the scattering ampli-
electron ensemble versus release times without (top) and with (bottom) tude typically described as g(p′ + Ap(tr)) and the Volkov phase that the
inclusion of charge interactions. The instantaneous release energy is electron accumulates from t′ to tr as:
evaluated by taking the first momentum of the time-resolved spectra
tr
(black dashed and solid curves) and the temporal phases (blue dashed ψr(t r) = ∫ −∞ ∫ g(p ′ + Ap (tr))Ep(t ′)d(p ′ + Ap (t ′))
and solid curves) by the temporal integration of the instantaneous
  (7)
energy. Their comparison, shown in Extended Data Fig. 6b, suggests p2
× exp  − iS(p ′ , t r, t ′; Ap (t )) − i (t r − t ′) + iϕt ′dp ′ dt ′
marginal differences and, thus, corresponding negligible effects on the  2 
temporal characteristics of the electron pulse at the time of recollision.
When the time-domain analysis is extended to the terminal ener- Here d and g are the dipole and scattering matrix element, respec-
gies of the recolliding electrons versus release time, the space–charge tively, as defined in refs. 39,40,58, and eiϕt reflects the extra phases acquired
interactions are manifested by a uniform upshift of terminal energies during the time evolution of the bound state before ionization. In our
by about 8% but leave the temporal phase of the terminal electron wave experimental setting of HAS in which the Keldysh parameter of γ ≈ 0.38
packet unaffected (Extended Data Fig. 6c). suggests the tunnelling regime60, those three processes (ionization,
propagation and backscattering) can be independently treated without Note that if the last phase term e−iΔS was missing, the expression would
non-adiabatic corrections on the Volkov phase, S (refs. 61,62). be identical to equation (7) and, thus, the gate-free recollision wave
function ψr(tr). Hence it would obviously be convenient to remove the
Description of electron wave packets under addition of a weak gate e−iΔS term from the integrals, as this would enable to express the recol-

field. Equation (5) implies that access to ψr(t) is possible if ψt(p) and liding electron wave packet ψ (g) r
(tr, τ) in equation (8) through the unper-
Ap(t) are known. Therefore our goal is to describe how these quantities turbed counterpart ψr(tr) and an extra phase. To proceed further with
can be accessed using a phase-gating process of the optical field emis- this idea, we consider two approximations.
sion by a weak replica of the driving pulse (gate pulse). Now we inspect First, following the famous saddle-point approximation, the domi-
the effects of adding the gate pulse on the dynamics of the electron nant contribution in the integration over the canonical momenta
described in equations (5)–(7). We define the gate pulse by a field p′ is provided by the kinetic momentum p′ + Ap(t) that equals the
Eg(t + τ) and its vector potential Ag(t + τ), in which τ is the delay between kinetic momentum pr of the recolliding electron at the surface.
the pump and gate pulses as described above. By replacing the pump Hence the kinetic momentum term in ΔS may be approximated as
fields and its vector potentials by the superposition of pump and gate ΔS(p ′ + Ap(t), tr, t ′; Ag (t + τ)) ≈ ΔS(pr , tr, t ′; Ag (t + τ)).
pulses in equations (5)–(7) as Ep(t) → Ep(t) + Eg(t + τ) and Ap(t) → Ap(t) + Second, because the exponent term of the extra phase e−iΔS is oscil-
Ag(t + τ), the terminal spectral amplitude perturbed by the gate can be lating slowly compared with e−iS in the time integration over t′ in equa-
rewritten as: tion (11), the extra phase ΔS can be approximated by a time average ΔS
within a time window Δt:
∼ ∞  p2 
ψt(p, τ) ∝ i ∫ ψ (g)(tr, τ)exp i tr exp 1 tr
−∞ r
 2  (8) ΔS =
∣Δt∣
∫ t −Δt ΔS(pr , tr, t ′, Ag (t + τ))dt ′
r
(12)
[−iS (p, ∞, tr; Ap (t) + Ag (t + τ)]dtr
As our main goal is to reconstruct attosecond electron wave packets
in which ψ (g) r
(tr, τ) denotes the recolliding electron wave packet per- that contribute to the spectral cutoff, we choose Δt as the time interval
turbed by the extra gate pulse as marked by the superscript (g) to be between ionization and recollision of the classical backscattering tra-
distinguished from the gate-free counterpart ψr(t) (see equation (7)). jectory that results in the highest final kinetic energy. To evaluate the
Because the gate-free quantity ψr(t) is of interest in this discussion, the averaged phase ΔS and thereby simplify the analytical form of equa-

subject in this section is how to express ψt(p, τ) in terms of ψr(t) with tion (12), ΔS can be expressed as:
phase terms introduced by the gate.
tr ∞
First we investigate the perturbed recolliding electron wave packet ΔS (pr , t r, t ′, Ag (t + τ )) = ∫ pr Ag (t + τ )dt = − ∫ pr Ag (t + τ )dt
ψ (g)
r
(tr, τ) in equation (8) and how to link it with the gate-free electron t′ tr
(13)

wave packet ψr(t). If the gate field is sufficiently weak, that is, + ∫ pr Ag (t + τ )dt
t′
η = ∣Ag (t )∣ 2/ ∣Ap (t )∣ 2 ≪ 1, the dipole transition and scattering amplitudes
can be considered invariant, that is, [Ep(t ) + Eg(t + τ )] d (p + Ap (t ) +
Ag (t + τ )) ≈ Ep(t )g (p + Ap (t )) and g (p + Ap (t ) + Ag (t + τ )) ≈ g (p + Ap (t )) in the By inserting equation (13) into equation (12), the effective (averaged)
expression of the recolliding electron wave packet (equation (7)). phase variation ΔS can now be evaluated:
In such case, the gate only modifies the phase imparted on the wave
packet between ionization and recollision:
∞ ∞
 1 0 
ΔS = − ∫ pr Ag (t + τ )dt + ∫ pr  ∫ A (t + t ′ + τ ) dt ′ dt
tr tr Δt −Δt g
tr 
ψ(g)
r
(t r, τ ) ≈ ∫ −∞ dt ′ ∫ dp ′g(p ′ + Ap (tr))Ep(t ′)d(p ′ + Ap (t ′)) (b )
≡ A g (t + τ )
(14)

 (9) = − ∫ pr Ag (t + τ ) − A g (t + τ ) dt
(b )
 p2
× exp  −iS(p ′ , t r, t ′; Ap (t ) + Ag (t + τ )) − i (t r − t ′) + iϕt ′ tr  
 2 
(
= −ΔS pr ,∞, tr; Ag (t + τ ) − A g (t + τ )
(b )
)
In view of these considerations, the variation of the phase (see equa-
(b)
tion (6)) can be expressed as: in which A g (t) is defined as:
0
(
S p, t2, t1 ; Ap (t) + Ag (t + τ) ) (b)
A g (t ) =
1
Δt
∫ −Δt Ag (t + t ′)dt ′ (15)
t2
1 2 1 
= ∫  p + Ap (t) + Ag (t + τ ) − p 2 dt
t1 2 2  (10) Using the above-described approximations now enables to pull the
t2 t2
1 2 1 
≈ ∫  p + Ap (t) − p 2 dt + ∫ p + Ap (t)Ag (t + τ )dt extra phase term out of the integrations in equation (11) and, consider-
t1 2 2   
t1 ing the sign flip ΔS → −ΔS at the backscattering instance, the perturbed
 
( )
= S p, t2 , t1 ; Ap (t ) ( )
≡ΔS p + Ap (t ), t2 , t1 ; Ag (t + τ ) recolliding electron wave packet can be expressed through the gate-free
wave packet and the extra averaged phase term ψ (g) r
(tr, τ) ≈ ψr(tr)eiΔS in
Here the square term of Ag(t + τ) is ignored, as its contribution is equation (8).
negligible compared with the other terms. Equation (10) implies that We now move on to discuss how the electron wave packet can be
the gate field introduces an extra phase of ΔS(p + Ap (t), t2, t1; Ag (t + τ)) described at the end of the interaction in the presence of the pump
to the gate-free case. As a result, equation (9) can be rewritten as: and gate fields (equation (8)). Taking the results of equations (11)–(14)
and restoring pr with the kinetic momentum p + Ap(t), equation (8) can
tr
be rewritten as:
ψ (g)
r
(tr , τ ) ≈ ∫ dt ′ ∫ dp ′ g (p ′ + Ap (tr))Ep(t ′)d (p′ + Ap (t ′))
−∞

 p2  ∼ ∞  p2 
× exp −iS(p′, tr, t ′; Ap (t)) − i (tr − t ′) + iϕt ′ exp (11) ψt(p, τ) ∝ i ∫ ψr(tr)exp i tr  exp[−iS(p, ∞, tr; Ap (t) + Ag (t + τ ))]
 2 
−∞
 2  (16)
(b)
[−iΔS(p′ + Ap (t), tr, t ′; Ag (t + τ))] × exp[−iΔS(p + Ap (t), ∞, tr; Ag (t + τ ) − A g (t + τ ))]dtr
Article
Because the integration ranges for S and ΔS are identical (from tr to ∞), the momenta translated by the Volkov propagation with the exponent
the two phases can be merged into a single equation (S′ = S + ΔS), exp[−iS(p, ∞, tr; Ap (t))] (see equations (1) and (5)), but without the
phase from space propagation to the detection. Using the terminal

1 1  electron wave packet ψt(t) (equation (21)), the terminal electron spec-
∫ t  2 p + Ap (t ) + Ag (t + τ)
2
S ′(p, ∞, tr , τ ) = − p 2 dt tral amplitude (equation (20)) can be further simplified as:
r 2 

+ ∫ t (p + Ap (t ))(Ag (t + τ) − A (gb)(t + τ))dt
r
∼ ∞  p2 
ψt(p, τ) ∝ i ∫ ψt(tr)exp i tr exp[−iS(p, ∞, tr; AHAS (t + τ))]dtr (22)
  2
−∞
 2 
∞

1 
∫ (b)
≈  p + Ap (t ) + 2Ag (t + τ ) − A g (t + τ ) 
tr  2 
(17) under the condition that the variation of the vector potential is weak
  during the time window of the recollision. The HAS spectrogram equa-
  ≡ AHAS (t + τ )
tion then reads:

1 2 ∼
I (p, τ) = ∣ψt(p, τ)∣2
(
− p dt = S p, ∞, tr; Ap (t ) + AHAS (t + τ )
2 )
 2
(23)
 ∞  p2 
∝ ∫ −∞ ψt(tr)exp i 2 tr exp[−iS(p, ∞, tr; AHAS (t + τ))]dtr
 
in which AHAS(t) is hereafter referred to as the effective HAS vector
potential and reads: Equation (23) describes a spectrogram whose reconstruction allows
access to the final electron wave packet ψt(t) and, correspondingly,
(b)
(18) ∼
AHAS (t) = 2Ag (t) − A g (t) ψt(p) as well as AHAS(t).

This expression of the effective HAS vector potential is compatible The effective HAS vector potential AHAS(t). An explicit relationship
with the classical momentum accumulation during the excursion between the incident gate vector potential Ag(t) and the effective HAS
from the ionization to the detection, Δp = −e[2A(tr) − A(tr − Δt)], under vector potential AHAS(t) can be best understood in the Fourier domain.
t ∞ ∼
the rescattering condition, ∫ r Ag (t)dt = ΔtAg (tr − Δt) (refs. 11,36,63). Using the Fourier expansion, Ag (t) = ∫ dωAg (ω)eiωt, the effective HAS
t r−Δt
∼ vector potential can be expressed as,
−∞
Using equation (17), the terminal electron amplitude ψt(p, τ) can be
expressed as: ∞
∼ 1 0 ∞

AHAS (t) = 2 ∫ Ag (ω)eiωt dω − ∫ ∫ A (ω) eiω(t + t ′)dωdt ′
∼ ∞  p  2 −∞ Δt −Δt −∞ g
ψt(p, τ) ∝ i ∫ ψr(tr) exp i tr exp[ − iS(p, ∞, tr; Ap (t) ∞
∼  i 
−∞
 2  (19) = ∫ Ag (ω) 2 − (e−iωΔt − 1) eiωt dω (24)
−∞  ωΔ t 
+ AHAS (t + τ))]dtr ∞
∼ ∼
= ∫ Ag (ω) g (ω)e dωiωt
−∞
Equation (19) implies that the gate also contributes to the terminal
momentum of the electron wave packet by AHAS(tr + τ), which depends in which the newly introduced multiplier ∼
g (ω) is defined as:
on the time delay τ. As described in equation (5), the momentum
contribution Ap(t) of the pump field is already incorporated in the ∼  i 
∼ g (ω) = 2 − (e−iωΔt − 1) (25)
gate-free terminal electron spectral amplitude ψt(p), whose intensity  ωΔ t 
is directly accessible in experiments. Therefore it is convenient for
the analysis of HAS data to express equation (19) with the terminal As shown in equation (24), the Fourier components of the effective
∼ ∼
form of ψt(p). We decompose the phase in equation (19) by means HAS vector potential AHAS (ω) is related to those of the incident gate

of S(p, t 2, t1; Ap (t ) + Ag (t + τ )) ≈ S(p, t 2, t1; Ap (t )) + ΔS(p + Ap (t ), t 2, t1; AHAS (t + τ )) vector potential Ag (ω) by multiplication of ∼
g ( ω)
and rewrite equation (19) as: ∼ ∼
AHAS (ω) = Ag (ω)∼
g ( ω) (26)
∼ ∞  p  2
ψt(p, τ) ∝ i ∫ ψ (tr)exp i tr exp[−iS(p, ∞, tr; Ap (t))] ∼
−∞ r
 2  (20) The gate multiplier ∼g (ω) is independent from Ag (ω). This allows the
possibility of the complete characterization of Ag(t) from AHAS(t)
× exp[−iΔS(p + Ap (t), ∞, tr; AHAS (t + τ))]dtr
imprinted in a HAS spectrogram.
To better visualize the concept of AHAS(t) and to verify the validity
In analogy to equation (11), if the e−iΔS term vanishes, the above equa- of the assumptions used in the above derivation, a semiclassical
tion is identical to equation (5), which links the recolliding electron simulation of a HAS spectrogram was performed using single-cycle

wave packet ψr(t) to the terminal spectral amplitude ψt(p). Here, simi- pulses. The photoelectron spectrum cutoff energy variation evaluated
lar to the Fourier representation of the recollision wave packet, we by the HAS spectrogram is compared with the effective HAS vector
define the Fourier pair of the terminal electron wave packet as: potential AHAS(t) calculated using equation (24) (Extended Data Fig. 7).
The multiplier ∼ g (ω) depends on the excursion time Δt between ioni-
∼ ∞  p2  zation and the backscattering event of the highest-energy electron.
ψt(p) ≡ ∫ −∞ dt ψt(t) expi 2 t , On the basis of the well-established recollision model, 0.685 times
 
(21) the central excursion period64,65 (central period of E (ω)/ω 2)), which
∞  p2 
ψt(t) ≡ ∫ −∞ pdp ψ∼t(p) exp−i 2 t  corresponds to about 0.85TL, was used for Δt to evaluate ∼ g (ω). Here
  TL is the centroid period of the laser pulse. Extended Data Fig. 7b–d
shows that the cutoff energy variation in a HAS spectrogram closely
Note that the terminal electron wave packet ψt(t) is an auxiliary elec- follows AHAS(t) (black curve), as calculated by the unmodified vector
tron wave packet that contains time–structure information of the potential of the incident pulse (dashed red curve), regardless of the
recolliding electron wave packet ψr(t) at the recollision surface with carrier-envelope phase.
Retrieval of the vector potential Ag(t) from a HAS spectrogram As motivated in the main text, we isolated a spectral area of interest
The above discussion suggests that, by tracing the variation of the (AOI) from 80 to 230 eV (Extended Data Fig. 11a,b). With this region of
cutoff energy in a HAS spectrogram, we can obtain AHAS(t) (red curves interest, the terminal wave packet can be expressed as:
in Extended Data Fig. 8a,b). Therefore access to the Fourier components

∼(AOI)
of the effective HAS vector potential allows the characterization of the
∼ ∼ ψ (AOI)(t ) = ∫ ∣ψ t (ω)∣ e−iφ (ω) eiωt dω (27)
vector potential of the incident gate: Ag (ω) = ∼
−1 t −∞
g (ω)AHAS (ω) (Extended
Data Fig. 8c,d). The retrieved incident vector potential Ag(t) is shown
in blue in Extended Data Fig. 8b. Here φ(ω) is the spectral phase of the electron wave packet modelled
as a polynomial series up to the sixth order,
Identification of the absolute zero delay in a HAS spectrogram.
Identification of the zero delay between pump and gate pulses in N =6

a HAS spectrogram can be obtained with various methods. Here φ (ω) = ∑ Dn(ω − ωc)n (28)
n
we opted for a method that allows the absolute delay to be derived
directly from the HAS spectrogram. Even though the difference
between the intensities of pump and gate pulses is more than two in which Dn and ωc are the nth order dispersion and central frequency,
orders of magnitude, discernible modulations (approximately respectively. The reconstruction is based on a least-squares algorithm
5–10%) of the spectral amplitude of the spectrogram remain. In a written in MATLAB, which aims at the total minimization of the dif-
HAS spectrogram, the total photoelectron yield variation can be ference among the experimental (Fig. 4a and Extended Data Fig. 11a)
evaluated by spectral integration at each delay point (Extended Data and reconstructed spectrogram (Fig. 4b and Extended Data Fig. 11b).
Fig. 8e). The absolute zero-delay point can be found as the delay To further increase the fidelity of the reconstruction, we also simul-
point at which the yield is maximally varied (vertical dashed line in taneously fit the differential map D(E, τ) of a HAS spectrogram I(E, τ),
Extended Data Fig. 8e). which is defined as:

Benchmarking HAS through EUV attosecond streaking. EUV


∂I (E , τ)
attosecond streaking provides access to the detailed field waveform D(E , τ) = ∫ dτ (29)
of a pulse33–35. Because this technique of field characterization is inte- ∂τ
grated in our experimental setup, it allows us to benchmark HAS as a
field-characterization method. The differential map is useful because it can eliminate the unmodu-
Extended Data Fig. 9a,b shows the HAS and EUV attosecond streak- lated intensity along the delay axis and allows the retrieval algorithm
ing measurements, respectively. The vector potential waveform of the to reconstruct fine details of the experimental trace (Extended Data
incident gate pulse retrieved from the cutoff analysis in HAS (red curve Fig. 11c–e). As an initial guess for the phase, zero phase was used. The
in Extended Data Fig. 9c) and that from EUV attosecond streaking (blue retrieval of the terminal electron wave packet is shown in Extended
curve in Extended Data Fig. 9c) show excellent agreement, as verified Data Fig. 11f,g.
by the degree of similarity of about 0.95 (ref. 43), and support the notion In a next stage of the reconstruction, the recolliding electron pulse,
that the gate pulse indeed acts as a phase gate. which is the key quantity in this work, is evaluated by the inverse Volkov
propagation of the retrieved terminal electron wave packet as:
The gate pulse as a phase gate
The compact description of HAS as a spectrogram implied by equa- ∞
∼(AOI)  p2 
tions (2) and (23) assumes that the weak replica of the pump field acts ψ (AOI)
r
(t ) = ∫ dp ψ t (p)exp −i t  exp[iS(p, ∞, t ; Ap (t))] (30)
−∞
 2 
as a nearly pure phase gate on the electrons released by the pump.
In other words, it can modify the momentum of electrons released by
the pump field but does not greatly influence the process of electron The Volkov basis is reconstructed by precisely measuring the
ionization. Yet, unless the ionization nonlinearities are well understood pump-field waveform and its timing with respect to the emission. The
(for instance, in atoms), a theoretical estimate of the required ratio retrieved recolliding electron pulse is shown in Fig. 4.
between pump and gate pulses for attaining a sufficiently pure phase
gate requires experimental validation.
To identify safe limits within which the above condition is met, we Data availability
performed HAS measurements under different gate strengths and The data supporting the conclusions in the paper are available from
compared the vector potential waveforms extracted from HAS to those the corresponding authors on reasonable request. Source data are
characterized by EUV attosecond streaking. As shown in Extended provided with this paper.
Data Fig. 10a,b, the vector potential waveforms from two techniques
achieve best agreement at low gate/pump intensity ratio (η < 10−2). At 48. Schultze, M. et al. State-of-the-art attosecond metrology. J. Electron. Spectros. Relat.
higher intensity ratios, we observe a gradually increasing disagreement Phenom. 184, 68–77 (2011).
between the reconstructed waveforms with the two methods (Extended 49. Fowler, R. H. & Nordheim, L. Electron emission in intense electric fields. Proc. R. Soc.
Lond. A Math. Phys. Eng. Sci. 119, 173–181 (1928).
Data Fig. 10c,d), implying that the gate pulse no longer serves as a weak 50. Yalunin, S. V., Gulde, M. & Ropers, C. Strong-field photoemission from surfaces: theoretical
perturbation. These measurements suggest that, for the studied sys- approaches. Phys. Rev. B 84, 195426 (2011).
tem, HAS measurements require a gate pulse whose intensity is about 51. Rakić, A. D., Djurišić, A. B., Elazar, J. M. & Majewski, M. L. Optical properties of metallic
films for vertical-cavity optoelectronic devices. Appl. Opt. 37, 5271–5283 (1998).
10−2 lower than the pump intensity. 52. Seiffert, L., Zherebtsov, S., Kling, M. F. & Fennel, T. Strong-field physics with nanospheres.
Adv. Phys. X 7, 2010595 (2022).
HAS reconstruction methodology 53. Süßmann, F. et al. Field propagation-induced directionality of carrier-envelope phase-
controlled photoemission from nanospheres. Nat. Commun. 6, 7944 (2015).
At the first stage of the reconstruction of the HAS spectrogram, the 54. Seiffert, L. et al. Attosecond chronoscopy of electron scattering in dielectric nanoparticles.
terminal electron wave packet ψt(t) is retrieved, as its spectral intensity Nat. Phys. 13, 766–770 (2017).
∼ 55. Liu, Q. et al. Ionization-induced subcycle metallization of nanoparticles in few-cycle
∣ψt(p)∣2 can be directly obtained by a gate-free photoelectron spectrum.
pulses. ACS Photonics 7, 3207–3215 (2020).
Therefore the reconstruction problem is reduced to retrieval of the 56. Schötz, J. et al. Onset of charge interaction in strong-field photoemission from nanometric
spectral phase. needle tips. Nanophotonics 10, 3769–3775 (2021).
Article
57. Seiffert, L. et al. Trapping field assisted backscattering in strong-field photoemission from Acknowledgements We thank J. Apportin and B. Kruse for support with the FDTD
dielectric nanospheres. J. Mod. Opt. 64, 1096–1103 (2017). simulations. This work was supported by the Deutsche Forschungsgemeinschaft (DFG,
58. Suárez, N., Chacón, A., Ciappina, M. F., Biegert, J. & Lewenstein, M. Above-threshold German Research Foundation) SFB 1477 ‘Light-Matter Interactions at Interfaces’, project
ionization and photoelectron spectra in atomic systems driven by strong laser fields. number 441234705.
Phys. Rev. A 92, 63421 (2015).
59. Wolkow, D. M. Über eine Klasse von Lösungen der Diracschen Gleichung. Z. Phys. 94, Author contributions H.Y.K., M.G. and S.M. performed experiments and analysed the
250–260 (1935). experimental data. H.Y.K., M.G., L.S. and T.F. performed theoretical modelling and simulation.
60. Keldysh, L. V. Ionization in the field of a strong electromagnetic wave. Sov. Phys. JETP 20, E.G. conceived and supervised the project. All authors contributed to the preparation of the
1307–1314 (1965). manuscript.
61. Pedatzur, O. et al. Attosecond tunnelling interferometry. Nat. Phys. 11, 815–819
Funding Open access funding provided by Universität Rostock.
(2015).
62. Kneller, O. et al. A look under the tunnelling barrier via attosecond-gated interferometry. Competing interests The authors declare no competing interests.
Nat. Photon. 16, 304–310 (2022).
63. Scrinzi, A., Ivanov, M. Y., Kienberger, R. & Villeneuve, D. M. Attosecond physics. J. Phys. B Additional information
At. Mol. Opt. Phys. 39, R1–R37 (2005). Supplementary information The online version contains supplementary material available at
64. Paulus, G. G., Becker, W. & Walther, H. Classical rescattering effects in two-color above- https://doi.org/10.1038/s41586-022-05577-1.
threshold ionization. Phys. Rev. A 52, 4043–4053 (1995). Correspondence and requests for materials should be addressed to E. Goulielmakis.
65. Becker, W., Goreslavski, S. P., Milošević, D. B. & Paulus, G. G. The plateau in above- Peer review information Nature thanks Qing Dai, Michael Krüger and Peng Zhang for their
threshold ionization: the keystone of rescattering physics. J. Phys. B At. Mol. Opt. Phys. 51, contribution to the peer review of this work.
162002 (2018). Reprints and permissions information is available at http://www.nature.com/reprints.
Extended Data Fig. 1 | Experimental setup. EUV (a) and HAS (b) configurations. The schematics of the adjustable iris and the dual inner mirror assembly are
reproduced with permission by SmarAct GmbH.
Article

Extended Data Fig. 2 | Module for the measurement of the absolute


electron yield. Electrons emitted from a tungsten nanotip exposed to
single-cycle pulses were captured by a thin electrode placed about 2 mm above
the tip. The induced voltage on the electrode was recorded by a lock-in
amplifier.
Extended Data Fig. 3 | Stability of electron emission from tungsten individual spectra in the left panel (middle). Individual spectra superimposed
nanotips. a, Consecutive measurements of photoelectron spectra recorded for comparison (right). b, The same measurement as in a but over a time period
under identical conditions and over a time range of about 100 s (left). The grey of about 80 min.
line indicates the cutoff energy. Electron yield evaluated by integration of the
Article

Extended Data Fig. 4 | Semiclassical simulations of the optical field terminal energies of backscattered and direct electrons as functions of birth
emission from tungsten nanotips in the single-cycle limit. a, Terminal time, for which the ionization rate is encoded on the line shading in logarithmic
electron energy as a function of the ionization time under the electric field of scale. b, Electron spectra plotted as a function of the peak field intensity. Red
the incident single-cycle pulses (red curve). A peak field intensity of 41 TW cm−2 and blue dashed lines indicate the high and low cutoff energies, respectively.
was used in accordance with the experiments. Purple and green lines indicate
Extended Data Fig. 5 | FDTD simulations of the near-field enhancement. Evaluated spatial distribution of the field-enhancement factor at the apex of a
tungsten nanotip with an apex radius of 35 nm and opening angle of 15° (a) and a similarly sized tungsten sphere (b).
Article

Extended Data Fig. 6 | Impact of charge interactions on electron emission respectively. c, The same analysis as in b but for the terminal energy of the
dynamics. a, Electron spectra obtained from three-dimensional semiclassical electron ensemble. The charge interaction free instantaneous energy (dashed
M3C trajectory simulations for the experimental parameters without (dashed black curve) is virtually identical to that modified by charge interactions (solid
curves) and with (solid curves) the inclusion of charge interactions for black curve), except for a uniform shift (dashed grey curve) along the energy
representative pulse peak intensities. b, Recollision energy distribution as a axis. d, Retrieved spectral intensity (blue) and phase (red) of the recolliding
function of the release time without (top) and with (bottom) charge interactions electron pulse with (dashed curves) and without (solid curves) inclusion of
(false colour denotes spectral intensity). Black and blue curves are the space–charge effects on the terminal electron pulse. e, The same analysis as in
instantaneous energy and the temporal phase, respectively, whereas dashed d but in the temporal domain.
and solid curves mark the cases without and with the charge interactions,
Extended Data Fig. 7 | Incident and effective HAS gate vector potentials. of the driving pulse. The red dashed and black solid curves indicate waveforms
a, Gate multiplier function g͠ (ω ) as defined in the text. b–d, Semiclassically of the incident gate vector potential Ag(t) and effective HAS gate vector
simulated HAS spectrograms at different carrier-envelope phase (CEP) settings potential A HAS(t) calculated using equation (24), respectively.
Article

Extended Data Fig. 8 | Retrieval of the vector potential of the incident gate incident Ag(t) (blue) vector potentials. e, Identification of timing between gate
field. a, Measured HAS spectrogram (Fig. 3a). b, AHAS(t) (red) directly extracted and pump pulses in a HAS spectrogram. The variation of electron yields is
by tracing the cutoff energy variation in the spectrogram and the retrieved evaluated by the spectral integration of the HAS spectrogram (blue curve) for
incident vector potential Ag(t) (blue). c, Amplitude (blue) and phase (red) of the each delay. The absolute zero delay between gate and pump pulses is identified
−1
inverse of the multiplier g͠ (ω ) . d, Spectral intensity of AHAS(t) (red) and as the point at which the yield variation is maximized (dashed red line).
Extended Data Fig. 9 | Benchmarking of HAS through EUV attosecond streaking. HAS (a) and EUV attosecond streaking (b) spectrograms under identical
driving-pulse field. c, Comparison of the vector potential of the gate pulse evaluated from HAS (red) and EUV attosecond streaking (blue) techniques.
Article

Extended Data Fig. 10 | Benchmarking HAS. Top panels, HAS spectrograms potential waveform characterized by EUV attosecond streaking (red curve) and
recorded at different gate/pump intensity ratios, η ≈ 4.5 × 10 −3 (a), 8.3 × 10 −3 HAS (blue curve) for each gate/pump intensity ratio.
(b), 2.4 × 10 −3 (c) and 1.4 × 10 −3 (d). Bottom panels, comparison of the vector
Extended Data Fig. 11 | Retrieval of the terminal electron wave packet. reconstructed (red) spectrograms. Retrieved terminal electron wave packet in
Measured (a) and reconstructed (b) HAS spectrograms in the energy range spectral (f) and temporal (g) domains. The purple and red curves denote the
from 80 to 230 eV. Differential maps of measured (c) and reconstructed intensity and phase, respectively.
(d) spectrograms. e, The phase of the intensity shift in measured (blue) and

You might also like