Viewpoint | Published:


Challenges and opportunities in attosecond and XFEL science

Nature Reviews Physicsvolume 1pages107111 (2019) | Download Citation

Seven scientists share their views on some of the latest developments in attosecond science and X-ray free electron lasers (XFELs) and highlight exciting new directions.

Time for excursions into the time domain

Eva Lindroth

The possibility to generate light pulses with a duration of fewer than a hundred attoseconds has opened new perspectives for studying the electron dynamics in atoms and molecules and promises a deeper understanding of ionization and charge transfer processes, with implications also for chemistry. One of the most interesting applications of attosecond technology is the possibility to measure the phase of quantum mechanical wave packets and thus completely characterize them. The derivative of this phase with respect to energy gives the group the delay acquired by the electronic wave packet when it propagates through an atomic potential. During the past few years, the study of the delay in photoionization has emerged as a prominent example of what can be achieved in the time domain, with highlights such as the temporal reconstruction of the build-up of an autoionizing resonance1.

Experimentally measured delays include not only the physically interesting photoionization delay but also contributions from the ionizing field and from the measurement process. This is a challenge, but it can be circumvented by exploiting relative delays to remove the effect from the light field. A number of different approaches have already been successfully used in this direction. The measurement-induced contribution may be corrected for theoretically, at least in the general case, as it depends on the long-range potential and not on the details of the multi-body environment. In systems that lack a long-range potential, such as negative ions, measurement-induced contributions vanish completely for more energetic photoelectrons. It is thus possible to imagine strategies to control the measurement effects with a combination of theoretical and experimental efforts.

The possibility of using a stopwatch for events on the atomic scale can be used to address questions around charge transfer and ionization in a variety of systems from isolated atoms to solids. But it is also possible to address fundamental questions concerning time and time-order in quantum mechanics. A still-debated point is whether and when, from a fundamental point of view, it is possible to say that one quantum mechanical event occurs before another. Ionization from inner orbitals provides a suitable test case. After such an event, the relaxation processes typically involve the ejection of a secondary electron in an Auger process. It is well known from frequency domain studies that when a slow photoelectron is followed by a fast Auger electron this gives rise to specific spectral features called post-collisional effects. In a classical picture, the situation corresponds to the Auger electron overtaking the photoelectron, meaning that the potential landscape through which the two electrons are moving suddenly changes. How is this temporal interplay between the electrons manifested in the ionization delay and can delay measurement be used to study the time-development of entanglement and its connection to the emergence of time-order?

Attosecond technology for imaging the electron timescale in photochemistry

Francesca Calegari

Watching a chemical reaction in progress has long been a dream of chemists. In 1999, Ahmed Zewail was awarded a Nobel Prize for the real-time investigation of the formation and breaking of chemical bonds, heralding the beginnings of femtochemistry. Several years later, the advent of attosecond technology dramatically improved the resolution of time-resolved measurements, ultimately allowing direct access to the electronic timescale in matter.

In its infancy, most attosecond science was devoted to the investigation of fundamental phenomena in atoms and small molecules, for which increasingly accurate theoretical models have been developed. However, real progress occurs when a field is pushed towards new challenges. I believe this is exactly what happened to attosecond science after two theoretical works suggested that the sudden ionization of a large molecule could induce attosecond charge flow along the molecular backbone2,3. These works triggered the investigation of more complex targets. Scientists immediately realized that attosecond technology could be used to track this ultrafast charge motion, known as charge migration, and ultimately control the outcome of a chemical reaction by acting at the electronic timescale. In the past few years, several attempts have been made to capture charge migration in biochemically relevant molecules, such as aromatic amino acids, with preliminary but encouraging results. Much effort has also been dedicated to improving the theoretical modelling to understand the basics of this process.

Today, there remain challenges that make it difficult to predict the full potential of attochemistry. First, the role of complex non-adiabatic couplings between the electronic and nuclear degrees of freedom still needs to be clarified. There are a few theoretical predictions that suggest that nuclear motion could ‘kill’ electronic coherences, even in the first few femtoseconds after light absorption. This could strongly limit our ability to drive coherent electronic motion and thereby control the reactivity of a system. Second, it is still unknown how the environment affects the dynamics initiated in the molecule. Until now, experiments have generally been carried out in isolated molecules in the gas phase. However, a more realistic environment needs to be considered, as most bio-relevant molecules are typically found, for instance, in water. Third, a strong technological effort is required to improve the signal-to-noise ratio of attosecond time-resolved experiments in biomolecules. The relatively low conversion efficiency of the attosecond pulse generation process, especially at high photon energies, combined with the extremely low density of the complex molecular target, make the ongoing experiments challenging. In this context, tremendous progress could be made by replacing table-top attosecond sources with free-electron lasers (FELs). However, whether FELs will soon be ready to deliver attosecond light pulses with reproducible time profile and intensity is yet to be demonstrated.

I cannot predict whether there will be a bright future for attochemistry, but to me one thing is clear: imaging and controlling the chemical reactivity of a molecule on the electronic timescale will open up new and important perspectives into photochemistry and photobiology, with tremendous prospects for understanding fundamental problems across physics, chemistry and biology.

3D dynamic imaging of complex systems with XFELs

Linda Young

In 2009, the world’s first hard XFEL provided a 10-billion-fold gain in X-ray peak brightness. This fuelled the avidly pursued, but still elusive, dream of single-shot imaging of non-crystalline samples at near-atomic resolution. The overarching concept of ‘diffract-before-destroy’ imaging is that a powerful burst of ultrafast X-rays can produce a coherent diffraction pattern of a nanoscale object — before the object explodes owing to Coulomb repulsion caused by a massive parallel ionization of the atoms in the sample. A collection of snapshots of identical objects in random orientations then provides information from which a 3D structure can be retrieved.

Underlying the development of such 3D imaging is an understanding of the fundamental interactions of X-rays with matter at extreme intensities, up to 1020 W cm−2. Since 2009, much has been learnt about ultrafast, ultra-intense X-ray interactions, using atoms, molecules and rare gas clusters as targets. For instance, sequential multiphoton absorption dominates, and nonlinear field-driven phenomena that characterize intense optical laser interactions are of minor importance. Now, theoretical modelling can point to optimal X-ray laser parameters for ultrafast atomistic imaging and warn about ‘hot spots’ for damage. This knowledge base for flash imaging of molecular structures has been built with the first generation of XFELs, which operate near 100 Hz primarily with stochastically varying self-amplified spontaneous-emission pulses. These powerful but crude pulses have enabled the successful and popular application of serial femtosecond crystallography for biomolecule structure determination, as in this application the details of the temporal pulse structure are of relatively minor importance.

However, with the next generation of high-repetition-rate XFELs one can envisage improved X-ray temporal profiles, stability and synchronization with external laser pulses, in addition to multicolour, multipulse X-ray pump–probe modalities using sub-femtosecond pulses. I believe that the improved X-ray characteristics can lead to a new era of X-ray quantum control empowered by the ability to capture all ejecta, photons, ions and electrons from individual interaction events, accumulate this information over many events and, subsequently, extract insights on the 3D dynamic behaviour of complex systems at the atomic and electronic scale.

Exploring the time dependence of phase stability

Marion Harmand

Understanding how matter behaves under the most extreme conditions is very challenging. It is of great importance for the study of planetary interiors or to unravel the processes taking place during accretion, giant and high-velocity impacts of astrophysical bodies. Laser compression is a unique tool that enables us to obtain pressures and temperatures far above the current limits of diamond anvil cell technology (several megabar and thousands of Kelvin). It consists of generating compression waves in a material from the rapidly expanding plasma created on its surface upon irradiation with a pulsed laser beam. However, the ultrafast timescales involved in laser compression (picosecond and nanosecond timescales) combined with a limited number of ultrafast diagnostics has so far prevented full characterization of the material sample under these extreme conditions, in particular its structural physical properties. Furthermore, the very short timescale of the measurements has always cast some doubt on the relevance of the technique for the long timescales in planetary science.

XFEL facilities brought a new range of ultrafast X-ray diagnostics for time-resolved experiments. X-ray pulses can be combined with optical laser compression in pump–probe schemes: a pump optical laser creates a shock wave going through the sample, which is probed after a controlled time delay with an XFEL probe. Therefore it is now possible to study matter under laser compression on the picosecond timescale. The high brilliance of XFELs enables high-quality X-ray diffraction, scattering and absorption spectroscopy with laser compression. In this way, it is possible to investigate ultrafast changes under conditions beyond the limits of static compression (with diamond anvil cell technology), and study electronic and atomic structures in systems whose intrinsic interplay gives rise to exotic states of matter, such as anomalous melting or metastable phases. Furthermore, in recent laser compression experiments the higher repetition rate allows statistically significant amounts of data to be acquired for the first time.

XFELs are now changing our perception of laser compression physics by providing access to ultrafast dynamics and out-of-equilibrium processes, and by offering new opportunities to relate measurements of such extreme conditions to the properties of matter in planetary interiors. However, for the analogy to work, one needs to understand how to relate ultrafast experimental timescales to planetary timescales, through the concept of stability. Traditional studies using static compression are implicitly based on the concept of phase stability, assuming a free energy minimum to derive the phase diagrams and physical properties. However, dynamic compression experiments coupling XFELs and high-energy lasers call for a different paradigm. How can we explore phase stability or instability by accessing temporal behaviour and dynamics? What information can be extracted from the temporal behaviour and out-of-equilibrium processes? How shall we define our experimental protocols to adapt to this new paradigm and think about the notion of stability in a physical system under extreme perturbations?

In this context, I think it is of the utmost importance to redefine the notion of temporal stability relative to the energetic stability of a system. In practice, I identify three important aspects to be considered experimentally. First, we need to explore the nature of the accessible phase transition and associated limitations when using laser compression by varying the accessible sample parameters, such as preferred crystal orientations and grain sizes, and modifying the compression schemes. Second, we need to define experimental protocols to disentangle the ultrafast changes in the thermodynamic conditions from the phase transition processes and kinetics. Third, we need to explore the probabilities related to the observation of the onset of a phase transition and its stability. In other words, look into the changes that would occur in a system evolving towards a phase before it fully reaches that phase. These ideas will certainly require close collaboration between theorists, modellers and experimentalists.

High-harmonic generation interferometry

Nirit Dudovich

Attosecond science enables the observation of multi-electron dynamics in atoms, molecules and solids. One of the most exciting advances in this field is high harmonic generation (HHG) spectroscopy. It combines sub-angstrom spatial resolution with attosecond temporal resolution, making it possible to dynamically resolve the structure of electronic wavefunctions as they evolve. HHG spectroscopy exploits a built-in pump–probe process: driven by the strong laser field, the liberated electron wavepacket returns to the parent ion and probes the hole through radiative recombination. In turn, this leads to the emission of higher-order harmonics of the driving laser field. The emitted attosecond burst of light takes a snapshot of the system, probing the evolving wavefunction, in an attosecond movie in which each harmonic order serves as one frame.

During the past two decades, HHG spectroscopy has been successfully applied to resolve fundamental strong-field phenomena such as field-induced tunnelling in atoms, hole dynamics or charge migration in molecules and field-induced currents in solids. However, whereas HHG spectroscopy holds great promise for both the measurement and control of matter, the understanding of the processes involved and the implementation of the technique still pose significant challenges. The extreme nonlinear nature of the strong-field interaction offers numerous channels — strongly coupled by the laser field — in which electronic dynamics can evolve.

One of the most important aspects of HHG spectroscopy lies in its coherent nature. The strong-field interaction directly transfers the coherence of the laser field into the coherent properties of the electronic wavefunction and then, upon re-collision, into the optical properties of the emitted harmonics. The coherence is encoded in the amplitudes, phases and polarization state of the harmonics and is key to reconstructing the internal dynamics. But resolving the internal coherence is one of the primary challenges in HHG spectroscopy. As in many other branches of physics, the presence of coherence is determined through interferometry. Synchronizing several fields produces the interferometric measurement during the strong-field interaction itself. By using a combination of strong and weak driving laser fields, one can perturb the interaction on sub-optical-cycle timescales. This perturbation induces a temporal interferometer whose two arms are represented by two attosecond bursts originating from consecutive laser half-cycles. The control of the two-colour delay manipulates the relative phase between the two arms and enables resolution of the coherent properties of the electronic wavefunction, represented by each arm. This scheme was applied to resolve field-induced tunnel ionization in atoms, multichannel dynamics in molecules or interband dynamics in solid state systems.

A more advanced scheme combines attosecond extreme ultraviolet (XUV) pulses with a strong infrared field. In this scheme the initial excitation is induced by an attosecond XUV pulse, rather than field-induced tunnelling. The integration of attosecond pulses with re-collision combines the simplicity of single XUV photon excitation with the accuracy provided by the re-collision process. In this case one creates an internal interferometer whose arms are the single photoionization channel and the re-collision path. This scheme enables direct control over the interferometer, whose beam splitter, in this case, the photoionization, can be fully tuned both in time and in energy.

Finally, the HHG coherence can be directly resolved via an all-optical approach. In this case, the internal interferometer, induced by multiple quantum path interference, is replaced by an external optical interferometer. Optical extreme ultraviolet interferometry has been realized by inducing two or more phase-locked HHG sources. This scheme has been applied to reveal re-collision dynamics, the contribution of multiple molecular orbitals or as a direct measurement of the spectral phase associated with photoionization.

Inducing strong-field interferometers increases the dimensionality of both the measurement and the control of attosecond-scale processes. Multidimensional HHG spectroscopy has the potential to isolate the numerous degrees of freedom involved in many strong-field phenomena, thus revealing their underlying mechanisms. The role of coherence in systems with more complex coupled electronic and nuclear dynamics and the possibility to laser-control these dynamics on the sub-cycle timescale set some of the most important challenges of the field.

Towards attosecond XFELs

Nora Berrah

XFELs4 have unprecedented brightness and short pulses of the order of a few femtoseconds. XFELs are relatively new tools but they have already enabled new science opportunities in physical chemistry and biology; in the study of nonlinear processes in atoms, molecules and solids; and in understanding ultrafast molecular dynamics.

High-energy X-ray photons make it possible to flag element-specific and site-specific physical and chemical processes, which is not possible with optical photons. Time-resolved X-ray scattering techniques make it possible to take snapshots of fundamental processes in solids, films and molecules and to reconstruct images to determine the sample structure. Furthermore, targeting the inner shells in atoms creates localized charge distributions that can be tracked in real time. This ability is useful in spectroscopic studies of photoinduced processes such as photosynthesis and vision, which take place in molecules through coupled nuclear and electronic dynamics on femtosecond and sub-femtosecond timescales. Charge transfer and charge migration also play a decisive role in chemical bond cleavage. In photoinduced processes, a superposition of states evolves on attosecond timescales, leading to electronic motion governed by electron correlations. Such coherent electronic and nuclear dynamics are challenging to capture with ab initio theories, or with attosecond experiments with table-top lasers, which recently started to measure the electronic motion created by photoinduced excitation. These measurements are limited by the low intensity of the attosecond pulses generated by the table-top lasers — which are typically strong near-infrared fields or a combination of low-intensity attosecond and strong femtosecond pulses — restricting the range of phenomena that can be investigated.

The limitation of low-intensity attosecond pulses will be overcome by attosecond XFELs. The upcoming Linac Coherent Light Source (LCLS-II)5 is part of ongoing efforts to reach the attosecond regime and will provide attosecond pulses. The current LCLS has electrons accelerated in copper pipes that operate at room temperature, allowing the generation of 120 X-ray laser pulses per second. The upgraded LCLS-II will add a superconducting accelerator, generating an almost continuous X-ray laser beam. The attosecond pulses are created by manipulating the electron bunch. The LCLS-II facility will, in addition, permit increased access to the FEL, enabling simultaneous experiments during beamtime operation (currently, it is single-user only).

X-ray-based attosecond research will provide amazing opportunities, but not without technical challenges. The self-amplified spontaneous-emission-based FELs, which are built by generating spontaneous lasing by accelerating a high-energy electron beam close to the speed of light, are inherently noisy and present jitter. However, future FELs, including LCLS-II, will use self-seeding and other upgrades to improve the ultrafast pulse profile and generate pulses with high transverse and longitudinal coherence.

Controlling, shaping and imaging quantum matter with intense light

Olga Smirnova

The detection of light shifts of atomic energy levels heralded a new field of shaping matter with light. Light shifts enable optical traps, lattices and tweezers, and the creation of Bose–Einstein condensates and other exotic states of quantum matter. Increasing the intensity of the light fields reveals other facets of this phenomenon: the Autler–Townes splitting of states and, more generally, the appearance of the Floquet ladder, that is, additional light-induced states separated by photon energy. In turn, this has led to new ideas from electromagnetically induced transparency to the Floquet engineering of quantum matter. These concepts involve the response of matter to light on a timescale far longer than a laser cycle, averaging the response over many laser oscillations.

Upon increasing the light intensity further, the utility of the frequency-domain photon counting picture subsides: too many photons are absorbed or emitted, making photon counting cumbersome. Instead, the time-domain picture becomes particularly insightful as it goes beyond the Floquet ladder by revealing and exploiting the time-domain, sub-laser-cycle response of matter to intense light fields — the mark of strong-field and attosecond physics. Strong-field and attosecond physics bring and exploit truly powerful tools for controlling, shaping and imaging matter with light owing to virtually unlimited opportunities to sculpt individual light oscillations offered by modern technology. These capabilities open several exciting opportunities.

First, in strong light fields, optical transitions are localized near the peaks of the oscillating laser field, a feature encountered in atoms, molecules and solids. This feature enables the exquisite temporal control over electron injection or emission using tailored, multicolour, polarization-controlled laser pulses, in which the phase of each oscillation is locked to the pulse envelope. In solids, this creates opportunities for controlling the excited electronic density in the Brillouin zones. Achieving such control could enable manipulation of currents in solids at sub-laser-cycle timescales, that is, many orders of magnitude faster than possible in conventional electronics.

Second, the sensitivity of the strong-field response to orbital momentum and spin on the sub-laser-cycle timescale, now demonstrated in atoms, is a gateway to sub-laser-cycle imaging of magnetic, chiral and topological phenomena. Thus, strong-field and attosecond physics may offer new tools to explore new phases of materials. Since strong light fields can inject spin-polarized electrons in the atomic and molecular continua, they should be able to do the same for the conduction bands of solids. Controlling the flow of spin in solids at the sub-cycle timescale is a new opportunity for ultrafast spintronics.

Third, very recent work6 shows that the sub-cycle strong-field response is sensitive to topological phases of matter, distinguishing between trivial and topological insulators that have identical band structure. This sub-cycle response can be detected via high harmonic spectroscopy whose sensitivity to chiral and correlation-driven electron dynamics has now been well documented in molecules. Application of high harmonic spectroscopy to solids opens a challenging but exciting opportunity for time-resolved imaging of new phases and phase transitions.

An interesting example of shaping matter with intense light is the so-called Kramers–Henneberger atom, which has been recently observed experimentally7. It is an example of strong-field-built electronic structure, in which the electron motion is dominated by the laser field, but the electron is still bound to the atomic core. Applying this concept to solids could lead to new ways of controlling the band structure beyond Floquet engineering, that is, by exploiting the highly non-perturbative interaction with light involving a vast number of low-frequency photons. Furthermore, this could lead to the ambitious goal of creating new phases of matter by reshaping the electronic structure with strong low-frequency laser fields.

Finally, imaging and controlling ultrafast electron dynamics in chiral molecules is an exciting challenge. The aim is to achieve ultimate control over chiral electronic response, that is, to generate such response only in species of desired handedness, while silencing the response of others. To make it happen, a special type of laser electric field is required. Technically speaking, such laser electric field should be chiral in the dipole approximation, that is, locally chiral at every point in space. It should also maintain its handedness globally in space, across the whole interaction region. Recent work shows how such fields can be generated by combining several phase-locked and polarization-controlled light pulses, and how highly nonlinear optical response can be induced only in molecules with desired handedness8.

These challenges and opportunities point us to several exciting directions, unified by the desire to reveal new properties of matter associated with highly non-equilibrium ultrafast electronic response and create new ways of shaping, imaging and controlling matter with light.

Additional information

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Change history

  • 01 February 2019

    The emails of all corresponding authors have been added to this article.


  1. 1.

    Gruson, V. et al. Attosecond dynamics through a Fano resonance: monitoring the birth of a photoelectron. Science 354, 734–738 (2016).

  2. 2.

    Cederbaum, L. & Zobeley, J. Ultrafast charge migration by electron correlation. Chem. Phys. Lett. 307, 205–210 (1999).

  3. 3.

    Remacle, F. & Levine, R. D. An electronic time scale in chemistry. Proc. Natl Acad. Sci. USA 103, 6793–6798 (2006).

  4. 4.

    Huang, S. et al. Generating single-spike hard X-ray pulses with nonlinear bunch compression in free-electron lasers. Phys. Rev. Lett. 119, 154801 (2017).

  5. 5.

    Hartmann, N. et al. Attosecond time-energy structure of X-ray free-electron laser pulses. Nat. Photon. 12, 215–220 (2018).

  6. 6.

    Silva, R. E. F. et al. All optical ultrafast probe of a topological phase transition. Preprint at arXiv (2018).

  7. 7.

    Matthews, M. et al. Amplification of intense light fields by nearly free electrons. Nat. Physics 14, 695–700 (2018).

  8. 8.

    Ayuso, D. et al. Locally and globally chiral fields for ultimate control of chiral light matter interaction. Preprint at arXiv (2018).

Download references

Author information


  1. Department of Physics, Stockholm University, Stockholm, Sweden

    • Eva Lindroth
  2. Attosecond Science Division at Deutsches Elektronen-Synchrotron (DESY), Hamburg, Germany

    • Francesca Calegari
  3. Department of Physics, Hamburg University, Hamburg, Germany

    • Francesca Calegari
  4. Argonne National Laboratory and The University of Chicago, Lemont, IL, USA

    • Linda Young
  5. Sorbonne Université, Muséum National d’Histoire Naturelle, UMR CNRS 7590, IRD, Institut de Minéralogie, Physique des Matériaux et de Cosmochimie (IMPMC), Paris, France

    • Marion Harmand
  6. Weizmann Institute of Science, Rehovot, Israel

    • Nirit Dudovich
  7. Department of Physics, University of Connecticut, Storrs, CT, USA

    • Nora Berrah
  8. Max Born Institute for Nonlinear Optics and Short Pulse Spectroscopy, Berlin, Germany

    • Olga Smirnova
  9. Technische Universität Berlin, Berlin, Germany

    • Olga Smirnova


  1. Search for Eva Lindroth in:

  2. Search for Francesca Calegari in:

  3. Search for Linda Young in:

  4. Search for Marion Harmand in:

  5. Search for Nirit Dudovich in:

  6. Search for Nora Berrah in:

  7. Search for Olga Smirnova in:

Competing interests

The authors declare no competing interests.

Corresponding authors

Correspondence to Eva Lindroth or Francesca Calegari or Linda Young or Marion Harmand or Nirit Dudovich or Nora Berrah or Olga Smirnova.

About this article

Publication history



Newsletter Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing