Science - USA (2020-09-25)

SCIENCE sciencemag.org

GRAPHIC: C. BICKEL/

SCIENCE

age), but stimulated emission of a photon

into that same x-ray pulse immediately fills

the hole with another outer electron from

the same atom, essentially instantly “heal-

ing” the atom. This process creates a mildly

excited neutral atom that has obtained a soft

kick or push along the laser direction, as if

it had only interacted with a n on-ionizing

photon of much lower energy. Atoms under-

going other natural (i.e., statistical) decay

processes, such as spontaneous emission of

photons or electrons, will suffer a large mo-

mentum kick, and thus will be displayed as

a diffuse halo on the detector. The nearly

negligible momentum transfer of the x-

ray–stimulated Raman process, by contrast,

arises from the almost perfect compensation

of momenta of the c o-propagating absorbed

and re-emitted photons in the combined

nonlinear process of coherent excitation and

stimulated emission, respectively.

Stimulated x-ray Raman scattering in

a dense medium has previously been ob-

served by a collective (propagation-based)

experiment ( 3 ). By contrast, Eichmann et

al. experimentally isolate, extract, and de-

tect exclusively the individual “survivor

particles” that were excited but remained

neutral in the intense x-ray flash. In the

previous approach, the signal light was co-

propagating with the intense x-rays and

thus limited the detection of weak signals,

whereas neutral-momentum imaging al-

lows a background-free detection with sin-

gle-particle sensitivity.

One key to the success of this experiment

was the high repetition rate of what is cur-

rently the world’s most powerful x-ray laser,

the European XFEL (EuXFEL) at Schenefeld

near Hamburg. With effectively 1000 shots

per second and their method to isolate just

the atoms undergoing the x-ray–stimulated

Raman process, Eichmann et al. could op-

timize the exciting x-ray–atom interaction.

The extreme sensitivity of their recoil-

imaging approach allowed the fine-tuning

of the x-ray properties to maximize the de-

sired “gentle” nonlinear x-ray excitation rel-

ative to competing spontaneous decay chan-

nels, such as fluorescence or Auger electron

decay. By varying the photon energy and

intensity of EuXFEL, they obtained an al-

ready impressive relative yield of ~10% for

the stimulated versus spontaneous Raman

processes, which was also confirmed and

understood by theoretical modeling. This

understanding also showed that the limita-

tions are “just” machine parameters of the

FEL. Efficiency could be much improved,

such that the stimulated Raman process is

favored over the spontaneous Raman pro-

cess by a factor of >10. What’s still missing

are suitable ways of controlling x-rays that

are partly in place at lower-energy FELs

( 4 ). Their translation to hard–x-ray FELs

would then even allow coherent-control

approaches for quantum-state manipula-

tion, such as stimulated Raman adiabatic

passage (STIRAP) ( 5 ), while also addressing

and selectively steering all electrons of an

atom down to its core.

A range of opportunities can be expected

from this fundamental-physics result. The

observed interaction merges x-ray element

specificity from core electrons ( 6 ) with the

chemically relevant excitations of valence

orbitals and thus provides a powerful tool

for analytical spectroscopy. The authors

calculate that signals can be massively

enhanced by using two-color modes of op-

eration with few-femtosecond pulse dura-

tions, and indeed x-ray FELs are currently

being developed in this direction.

The use of two sets of two-color pulses

[in the best case with attosecond duration

( 7 )] that take advantage of FEL coherence

effects in resonant atomic absorption ( 8 )

would allow time-resolved probing of elec-

trons excited at a given atom on one end

of a molecule and traveling to a “target”

atom on the other end of the molecule.

This process is predicted to occur on time

scales of only a few femtoseconds ( 9 , 10 ).

The method would implement two-dimen-

sional spectroscopy with x-rays—as con-

ceived quite some time ago ( 11 )—but, as is

less commonly known, the signal could con-

sist of massive particles rather than light,

as a proof-of-principle experiment at much

lower frequencies recently showed ( 12 ).

Our improved understanding of funda-

mental x-ray light–matter interaction keeps

driving another revolution, namely com-

prehensive control and engineering of well-

defined quantum states and dynamics of

matter on a (sub-)atomic level. The work of

Eichmann et al. is part of an evolving skill

set for taming the darker side of x-rays. The

result will be sharper tools to control their

exciting properties. j

REFERENCES AND NOTES

1. W. C. Roentgen, Nature 53 , 274 (1896).


2. U. Eichmann et al., Science 369 , 1630 (2020).


3. C. Weninger et al., Phys. Rev. Lett. 111 , 233902 (2013).


4. E. Ferrari et al., Nat. Commun. 7 , 10343 (2016).


5. K. Bergmann, H. Theuer, B. W. Shore, Rev. Mod. Phys. 70 ,
   1003 (1998).


6. K. Siegbahn, Science 217 , 111 (1982).


7. J. T. O’Neal et al., Phys. Rev. Lett. 125 , 073203 (2020).


8. T. D i n g et al., Phys. Rev. Lett. 123 , 103001 (2019).


9. L. S. Cederbaum, J. Zobeley, Chem. Phys. Lett. 307 , 205
   (1999).


10. F. Remacle, R. D. Levine, Proc. Natl. Acad. Sci. U.S.A. 103 ,
    6793 (2006).


11. I. V. Schweigert, S. Mukamel, Phys. Rev. Lett. 99 , 163001
    (2007).


12. S. Roeding, T. Brixner, Nat. Commun. 9 , 2519 (2018).

10.1126/science.abd6168

25 SEPTEMBER 2020 • VOL 369 ISSUE 6511 1569

Damaged atoms

When a “normal” x-ray (single photon) is absorbed by an atom, it knocks out core

electrons. This process causes further ionization through cascaded decays.

Beam of atoms

More intense but gently exciting

Intense x-rays can move two electrons through a “nonlinear” two-photon process.

Concerted absorption and emission of x-ray photons leads to only mild (valence)

excitation of the neutral atom.

Imaging detector

A spatially resolving detector measures the momentum “kicks” of mildly

excited neutrals. These kicks are almost nonexistent for the coherent x-ray

(stimulated Raman) process; the detector image thus refects the

unperturbed “I-shaped” intersection of the x-rays and the atomic beam.

Electron Momentum

recoil “kick”

Detector

No signal or diNuse hits for

statistically excited neutrals

Ionized atom

(damaged)

Almost no

recoil

Excited atom

Intense

pulsed

x-rays

Wea k x- rays

Different x-ray responses

X-rays normally ionize or highly excite atoms out of their core shells. Eichmann et al. show how intense x-ray

pulses can create only slightly excited neutral atoms formed by promotion of core electrons to high valence

states, a key step for site-selective triggering of chemical reaction dynamics.