Simultaneous mapping of the ultrafast time

and fluence dependence of the laser-induced

insulator-to-metal transition in magnetite

J. O. Schunck1,2, P. S. Miedema1, R. Y. Engel1,2, S. Dziarzhytski1, G. Brenner1, N. Ekanayake1, C.-F. Chang4, P.

Bougiatioti5, F. Döring5, B. Rösner5, C. David5, C. Schüßler-Langeheine6 and M. Beye1,2,3,*
1
Deutsches Elektronen-Synchrotron DESY, Notkestr. 85, 22607 Hamburg, Germany
2
Physics Department, Universität Hamburg, Luruper Chaussee 149, 22761 Hamburg, Germany
3
Department of Physics, AlbaNova University Center, Stockholm University, SE-10691 Stockholm, Sweden
4
Max Planck Institute for Chemical Physics of Solids, Nöthnitzer Straße 40, 01187 Dresden, Germany.
5
Paul Scherrer Institut (PSI), Forschungsstraße 111, 5232 Villigen, Switzerland
6
Helmholtz-Zentrum Berlin für Materialien und Energie, Albert-Einstein-Straße 15, 12489 Berlin, Germany
*
[email protected]

Abstract
Pump-probe methods are a ubiquitous tool in the field of ultrafast dynamic measurements. In recent years, X-ray
free-electron laser experiments have gained importance due to their ability to probe with high chemical selectivity
and at atomic length scales. To obtain the complete dynamic information, measurements are typically repeated many
thousands of times with varying delay and/or fluence settings. This generally necessitates that the sample fully
recovers before the subsequent excitation and that probe pulses and the induced dynamic evolution are comparable
to each other. These conditions present a significant challenge when the sample fluctuates between different initial
states or when it is susceptible to damage. Also, source fluctuations are normally intrinsic to free-electron laser
pulses. Here, we present a time-to-space mapping imaging scheme that enables us to record a delay range of several
picoseconds as well as a laser fluence range of about one order of magnitude in every single shot of the x-ray probe.
This approach can circumvent the aforementioned preconditions. We demonstrate the use of this scheme by
mapping the femto- and picosecond dynamics of the optically induced insulator-to-metal Verwey transition in a
magnetite thin film. We employ resonant diffraction at the free-electron laser FLASH for probing. The dynamics
of the magnetite thin film are found to follow a biexponential decay in line with earlier studies on bulk crystals. By
extrapolating our results towards the conditions found at X-ray free-electron lasers with higher energy, we
demonstrate that the presented data could be recorded in a single shot.

1
1 Introduction Here, we present the results of setup that enables the
simultaneous recording of a delay range of several
The intriguing properties of complex, functional and picoseconds and a fluence range of about one order of
quantum materials, like high-temperature magnitude. We use an X-ray optical Fresnel zone plate
superconductivity (1, 2), colossal magnetoresistance to image the iron L3-edge resonant soft X-ray
(3, 4), strange metallic behavior (5, 6), and insulator- diffraction (RSXD) signal from a magnetite sample. A
metal phase transitions (7) have attracted researchers’ non-collinear pump-probe geometry with an angle of
interest on their quest to engineer and harvest these approximately 70° between the pump and probe beams
phenomena for several decades. Excitation with allows the pump-probe delay to be mapped onto a
ultrashort laser pulses is a particularly powerful tool to spatial axis on the detector (see Figure 1). The
manipulate equilibrium phases or create transient states orthogonal axis images the signal across the laser beam
which do not occur under equilibrium conditions (8– spot onto the detector, essentially mapping the pump
11). Usually, time-resolved studies are performed using laser fluence distribution. With this setup, a large
two ultrashort pulses: the first pulse (the pump) starts a parameter range can be acquired in a static geometry
dynamic process, and the subsequent pulse (the probe) without scanning. Given a sufficiently high signal, such
is used to observe the change induced in the sample. a setup even allows capturing a complete dataset in a
This pump-probe process is repeated thousands or single shot.
millions of times while varying the temporal separation
between the two pulses (the delay) and/or the fluence As a sample, we chose a thin magnetite (Fe3O4) film.
of the pump pulse. Being the first material in which a temperature-driven
insulator-metal transition was discovered (21–24),
X-ray free-electron lasers (FELs) offer the opportunity magnetite continues to serve as a model system for how
to use X-ray pulses for element-selective and atomic- charge, orbital and lattice degrees of freedom shape the
scale probing with high signal levels and femtosecond Verwey insulator-metal transition (25–34): Below the
temporal resolutions. Many FELs today generate transition temperature of 124 K, the resistivity
pulses using the process of self-amplified spontaneous increases by two orders of magnitude due to an order-
emission (SASE) (12–14), which results in fluctuating to-disorder transition (22). The high-temperature
pulse properties, like intensity, pulse duration and metallic phase has a cubic inverse-spinel crystal
spectral content. This characteristic motivated the structure in which Fe3+-ions occupy the tetrahedral (A)
development of techniques for recording a complete set sites, whereas octahedral (B) sites are occupied by both
of information for every single shot. In the post- Fe2+ and Fe3+ ions (26). When magnetite transitions to
analysis, the data can then be sorted and binned the low-temperature phase, the crystal unit cell doubles
appropriately. In addition, such an approach also helps to a distorted monoclinic structure (35), characterized
to detect non-reproducibility in the sample response by charge and orbital order (COO) (23, 27–30) within
itself, caused, e.g., by samples that fluctuate between the B-site Fe network. Emerging superstructure
different initial states (15, 16). Here, simultaneous reflections can be accessed with RSXD e.g. at the iron
detection of a range of parameters offers the L2,3-edges using the (00½) peak (the notation refers to
opportunity to assign recorded data to its respective the high-temperature unit cell) to probe the low-
state, thereby facilitating a study of such systems in the temperature order (29–32, 36). Accordingly, the
first place. In repeated pump-probe experiments in reflections vanish in the high-temperature structure.
general, one needs to take care that the sample is given
enough time to return to the ground state and that the In the parallel detection setup, both the FEL and the
pulse energies are not too high to cause permanent probe laser illuminate an extended area on the order of
damage to the sample. Alternatively, if the sample is one square millimeter on the sample, thus a
destroyed by every shot, a new sample or spot on the homogeneous sample with flat high-quality surface is
sample needs to be used for every measurement (17), required. Here, a high-quality magnetite thin film was
posing requirements on the available amount or size of used. Besides being flat, thin films epitaxially grown
sample. Many of these challenges can be alleviated on spinel Co2TiO4 substrates that closely match the
using approaches that can do a complete experiment, crystal structure and lattice constant of magnetite have
i.e., measure the required delay (and potentially been shown to have excellent properties (37): In
fluence) dependence in a single shot. One example of contrast to widely used magnesium oxide (MgO)
this approach is the X-ray streaking method (18–20), substrates (38–40), the resistance hysteresis width of
used at X-ray free-electron lasers to probe a delay range the Verwey transition matches that of the bulk, the
of approximately 1.5 ps in a single measurement. correlation length of the COO is much larger than for
films grown on MgO and comparable to bulk single

2
crystals, and the Verwey transition temperature even
surpasses that of the bulk by 3 K.
As earlier experiments have shown, the excitation of
the magnetite low-temperature phase with ultrashort
infrared (IR) laser pulses destroys the COO
superstructure and drives the system into the high-
temperature structure within several picoseconds (41–
46). Previous time-resolved IR pump, RSXD probe
studies (41, 42) on magnetite bulk single crystals
observed a two-step process on femto- and picosecond
time scales, characterized by an initial sub-picosecond
reduction of the scattering intensity caused by direct Figure 1: Time-to-space mapping setup. Soft X-
excitation of charge transfer between charge ordered ray pulses from FLASH resonant to the Fe L3 edge
sites (25, 43). The character of the ensuing evolution on (blue) probe the resonant scattering of Fe3O4 upon
the picosecond time scale depends on the pump fluence pumping with 800 nm pulses (red). The horizontal
(42): For fluences below a temperature-dependent scattering plane is shown in yellow. The incidence
threshold (e.g. approximately 1.3 mJ/cm² at 80 K), and scattering angle θ is 32.75° with respect to the
only a few holes in the charge order are induced. The sample surface and the pump laser is incident onto
deposited energy is not sufficient to overcome the the sample under an angle α of 16° with respect to
threshold of the phase transition so that the long-range the sample surface normal. Due to the relative angle
order recovers within picoseconds. At fluences above between FEL and pump, a delay range of several
this threshold, a transient non-equilibrium situation picoseconds is probed simultaneously. An
occurs with coexistence of insulating and metallic additional delay Δ𝐷𝑆 can be set by a mechanical
regions. Within several picoseconds, the volume delay stage. An off-axis Fresnel zone plate images
fraction of the metallic phase grows, and the system the resonant scattering of the unfocused FEL beam
evolves towards a homogeneous high-temperature from the sample onto a two-dimensional CCD
phase. In this work, we observe a behavior of the thin detector.
film that is consistent with these findings, while
acquiring the time- and fluence dependence resolution of the experiment was ∼120 fs (see
simultaneously with every shot. supplementary material). The laser spot measured
approximately 2.6 mm (horizontal, FWHM) by
0.2 mm (vertical, FWHM) and was focused such that a
2 Experimental setup flat-top region formed in the horizontal direction (see
supplementary material, Figure S1). The sample was
Measurements with the time-to-space mapping setup horizontally smaller than the laser footprint, extending
were performed at beamline BL2 (47–49) of the free- only in the flat-top region of the laser focus. The pulse
electron laser (FEL) FLASH at DESY in Hamburg, repetition rate of the FEL was 10 Hz, while that of the
Germany using the ultrahigh-vacuum diffractometer pump laser was 5 Hz so every other FEL shot probed
MUSIX (50). The schematic experimental setup is the unpumped sample, serving as a reference for
shown in Figure 1. To record resonant diffraction from normalization of the signal. We verified that the sample
a magnetite sample at the iron L3-edge at FLASH, third was fully recovered in between shots. Due to the small
harmonic radiation with a central photon energy of third harmonic fraction of the FEL beam, the scattering
706 eV was used. The full width at half-maximum cross-section of the sample and the large FEL footprint,
(FWHM) bandwidth of the FEL pulses was the count rate for the data set was on average
approximately 3.6 eV. The pulse energy of the FEL 0.54 counts/shot. In total, the data shown here
fundamental was on average 20 µJ and the third comprises 12 h of acquisition time and approximately
harmonic emission is usually two to three orders of 230,000 counts.
magnitude weaker (51–53).
The pump laser beam was incident under an angle of
The unfocused FEL beam had a diameter of 16° with respect to the sample surface normal (see
approximately 4.2 mm at the sample position. The Figure 1). The sample was a 40 nm thin magnetite
pump laser (54) wavelength was centered around (Fe3O4) film, grown on Cr2TiO4 (001) (37), measuring
800 nm, the pulse duration was 57 fs FWHM and the approximately 1.9 mm (horizontal) by 8 mm (vertical).
pulse energy was 16.8 µJ. The combined temporal The incidence angle θ of the FEL was 32.75° with

3
respect to the sample surface to fulfill the Bragg relative pump-probe delay Δ𝐷𝑆 (see Figure 1) back and
condition for the (00½) superstructure peak. The forth by 3 ps using a movable delay stage, thereby
sample was cooled to a base temperature of 100 K for moving the entire 7 ps delay window horizontally over
the entire duration of the measurements. An off-axis the sample surface. This scan also served as a cross-
Fresnel zone plate (FZP) imaged the diffraction signal calibration of the delay axis (see below and Figure 2).
onto a two-dimensional charge-coupled device (CCD), Scanning the pump laser spot in the vertical direction
thereby magnifying the diffraction signal from the is not that simple but generally possible; also, the
sample by a factor of 4.1. Further details on the overall laser intensity could be varied. But even for a
experimental setup can be found in the supplementary static laser spot, the correlation between fluence and
material. vertical sample position can be checked, since the
highest fluence occurs in an enlarged region around the
The non-colinear geometry of FEL and pump laser was
center of the beam, while all other fluence regions
used as a time-to-space mapping setup (see Figure 1):
appear twice in our data on both sides of the most
Due to the angle between the two beams, the relative
intense part that can be compared. The detection of
pulse front arrival time of both pulses varied along the
inhomogeneities in the dynamic response can occur
horizontal dimension, thereby mapping a time (i.e.
within the spatial resolution of the setup, which we
relative delay) axis to the horizontal space coordinate
estimate to be around 3 µm to 4 µm for the present
of the sample. With the FZP, the resonant diffraction
magnification (55).
signal from the sample, including this delay axis, was
spatially resolved by imaging onto the detector. The
delay range which can be covered in a single shot was 3 Results
determined by the geometry of the experiment
(incident angles of FEL and laser beams), as well as the As described above, we obtained data as a function of
magnification of the FZP and the properties of the the horizontal sample position, x, the vertical sample
detector. In the present setup, the delay axis was position, y, and the delay stage position, Δ𝐷𝑆 . How we
mapped on the CCD as 23.6 fs/pixel. In principle, this obtain the calibration of the delay axis on the detector
would allow us to record a delay range of up to 50 ps, coordinates from this multidimensional data set, as well
but in our experiment, the simultaneously recorded as data treatment steps used to obtain fluence-
delay range was limited by the horizontal sample size dependent delay traces, are visualized in Figure 2.
to approximately 7 ps. Like in a conventional pump- Small regions of the sample imaged at delay stage
probe scheme, the recorded delay range can be further positions Δ𝐷𝑆 of 0 ± 0.5 ps and 2 ± 0.5 ps are shown in
extended by scanning the pump laser delay stage (Δ𝐷𝑆 panels (a) and (b), respectively. In both panels, the
in Figure 1). depicted pump signal has been normalized to the
unpumped signal. The part of the sample where the
In order to keep the laser profile rather homogeneous
COO was pumped away by the laser is visible as blue
along the probed horizontal direction, which is used as
area. The horizontal position where both pulses cross
delay axis, we used a cylindrical lens to focus the
the surface at the same time (time overlap, t0) can be
optical laser to a horizontal line on the sample. In the
seen to be the x position where the signal transitions
vertical direction, the laser is focused much stronger,
from high to low scattering intensity (black dashed
well below the size of the probing FEL beam, such that
the full vertical spatial laser profile was probed lines). We chose the origin of the Δ𝐷𝑆 scale such that t0
simultaneously. In the data analysis, we could thus is in the horizontal center of the detector window
separate regions in the center of the vertical laser spot (x = 150 px), see Figure 2(a).
profile, which were pumped with a high local fluence, The vertical profile of the signal which is averaged in
from more weakly pumped regions further away. the pumped area of panel (b) along the x-axis is shown
Our imaging setup recorded two spatial dimensions as turquoise data in panel (c) of Figure 2. This profile
simultaneously. We used our special experimental maps the local strength of the pump-probe effect and
configuration to map the spatial coordinates onto one follows the local fluence profile of the laser, which we
temporal and one local fluence axis. As this scheme show overlaid to the pump laser profile as extracted
introduces a correlation of signals recorded with from an independent knife edge scan (green).
different delays and fluences with sample position, it We now focus on calibrating the delay scale. For this,
works best with samples that are spatially we use the region with the largest pump effect in the
homogeneous over the probed area. But the scheme is vertical center of the laser spot profile. We average the
also able to detect spatial inhomogeneities in the x-traces with pump fluences higher than 4 mJ/cm² and
dynamical response: to this end, we scanned the
4
 Figure 2: Pump-probe data treatment. The pump-probe delay varies along the horizontal x-axis
because of the different incidence angles of FEL and pump laser. Panels (a), (b): Normalized images of
the pump effect on the sample obtained by dividing pumped by unpumped events for delay stage position
ranges of Δ𝐷𝑆 = (0 ± 0.5) ps and Δ𝐷𝑆 = (2 ± 0.5) ps, respectively. One positional unit on the detector
corresponds to 13.5 µm. Panel (c) shows the laser beam profile (green, upper axis) and the corresponding
profile of the diffracted signal (cyan, lower axis), obtained by averaging only the pumped signal in panel
(b) along the horizontal axis. Panel (d): By scanning the pump laser delay stage position Δ𝐷𝑆 (see also
Figure 1) and plotting the average of detector y positions pumped with the highest fluences against the
delay stage position (data shown along the red lines are extracted from panels (a) and (b)) allows one to
observe how temporal overlap moves along x (black dashed line). The slope of the black dashed line
can be used to translate every row onto a common delay axis. The x-axis and y-axis have been binned
by a factor of 8 and 6, respectively. Panel (e) then shows the data set on the common horizontal delay
axis, while the vertical axis still shows the mechanical delay stage position. Pump-probe traces as shown
in Figure 3 are finally derived by averaging the data in panel (e) along the vertical dimension.
plot these for the different delay stage positions as a With this conversion, we can align each row to a
map (Figure 2(d)). t0 extends along a diagonal (dashed common delay axis, see Figure 2(e). A common delay
line) reflecting the linear influences of both x and Δ𝐷𝑆 trace using all available data is then obtained by
on the delay. The slope of this diagonal line yields the averaging the data in Figure 2(e) along the y axis. We
delay mapping factor on our detector (here 23.6 fs/px). note that there is a hint towards a small sample position
This value depends on the relative incidence angle dependence of the observed pump effect; it appears that
between FEL and pump laser, the sample and the sample region that is mapped onto x ≈ 90 exhibits a
observation angles, the magnification of the imaging somewhat slower pump dynamics, which can be seen
zone plate and the pixel size of the detector. By as a comparatively higher intensity around a delay
choosing these parameters, the setup-limited temporal value of 1 ps for delay stage positions of approximately
resolution can be tuned to match the experimental 2.5 ps. However, the present data set does not have
requirements.
5
 Usually, the pump-fluence dependence of the ultrafast
response of the sample to IR excitation is determined
by repeating the pump-probe experiment with different
pump laser attenuation settings. In our setup, we can do
this by sorting different rows of the detector image
according to the pump fluence at that position, see
Figure 2(c). We chose to sort the data into five fluence
bins (see supplementary material; Figure S2). Thus, we
obtain five delay traces spanning a fluence range
between 0.7 mJ/cm2 and 4.1 mJ/cm2 without actually
changing the incident laser pulse energy. The resulting
fluence-dependent IR pump, RSXD probe traces are
depicted in Figure 3(a), showing a double exponential
decay characterized by a fast and a slow component.
The resulting traces were fitted using a double
exponential decay function (solid lines) convolved with
a Gaussian of a width corresponding to the temporal
resolution of the experiment (120 fs FWHM) as a fit
model. The shaded regions in Fig. 3(a) show the
experimental one-sigma standard deviation calculated
Figure 3: Ultrafast fluence-dependent response of for each data point during the averaging of multiple
a 40 nm thin magnetite film to 800 nm laser traces and binning on the x-axis. During the weighted
excitation. Panel (a): The intensity of the (00½) least-squares fitting, the inverse experimental error is
superstructure peak decreases on a few-picosecond used as weights for the respective data points. We find
time scale upon irradiation with the pump laser for that following the IR excitation, a fast and a slow decay
fluences between 0.7 mJ/cm² and 4.1 mJ/cm². The process, respectively named A and B, occur, matching
data was normalized to 1 for negative delays and fit what has previously been observed for single crystals
curves were offset by 0.5 relative units (rel. u.) for (41, 42, 56). Within our experimental uncertainty, there
clarity, and the dashed lines to the right show 0 are no indications of a recovery in the measured delay
intensity for the data of the respective fluence. The range.
shaded area is the one-sigma measurement In panels (b) through (e) of Figure 3, the fluence
uncertainty of the data, calculated as the standard dependence of the fit results for the slow decay time
deviation for each delay value upon averaging constant 𝜏𝐵 (b), the amplitudes A (c) and B (d) and the
several traces of similar fluence and binning on the total long-time decrease A+B (e) are shown. Error bars
delay axis. Full lines show fits with a double represent the one-sigma standard deviations. The fast
exponential decay convolved with a Gaussian decay is limited by the temporal resolution of the
profile with a width corresponding to the experiment to 120 fs. In the fitting procedure, this is
experimental resolution, shown below in black at a realized by the convolution with the resolution-limited
delay of 0 ps. The fast time constant 𝜏𝐴 is set to be Gaussian function, therefore the value of 𝜏𝐴 was fixed
much faster than the temporal resolution. The to a value of 10 fs, much faster than the temporal
fluence dependence of the slower time constant 𝜏𝐵 resolution. Both amplitudes, A and B, show a linear
is shown in panel (b). Panels (c) and (d): Fluence increase with fluence of 0.08 rel. u./(mJ/cm2) and
dependence of the amplitudes 𝐴 (fast process) and 0.16 rel. u./(mJ/cm2), respectively (black dashed lines
𝐵 (slow process) alongside linear trend lines with in Figure 3(b) and (c)). The slow decay time constant
slopes 0.08 rel. u./(mJ/cm2) and 𝜏𝐵 ranges from 1 ps to 2 ps, generally matching
2
0.16 rel. u./(mJ/cm ), respectively. Panel (e): The previously reported values (42). See the supplementary
total decay amplitude (i.e. sum of A and B). See material for more details on the fitting procedure.
supplementary material for details on the fitting
procedure.
4 Discussion
sufficient quality to study this rather weak effect in
more detail. The IR pump, RSXD probe results obtained on
magnetite thin films qualitatively match those

6
previously reported for single crystals (41, 42, 56): we one would ideally study both sample types in the same
find a biexponential decay with a fast time constant 𝜏𝐴 setup. As intrinsic sample properties are concerned,
limited by the temporal resolution and a slow time differences between our magnetite thin film and a bulk
constant on the picosecond scale. In the earlier studies, magnetite crystal are expected to be small: A study
the fast process 𝐴, induced by the 800 nm pump, has directly comparing static RSXD and X-ray absorption
been connected to electron transfer excitations from an data of thin films and bulk single crystals only finds
Fe2+ 3+
B to an FeB ion (41–43, 46). This process directly minor spectral differences (36). In that study, MgO was
alters the charge distribution and moves electrons used as a substrate for the thin film (as compared to
between orbitals, which rapidly destroys charge and Co2TiO4 for our study), which is arguably even less
orbital order. We corroborate this within our resolution similar to bulk magnetite, judging by the very broad
of 120 fs. The linear fluence dependence of amplitude resistance hysteresis (37).
𝐴, previously reported for single crystals (42), is
We can further rule out the limited film thickness as an
confirmed by our measurements. We find, however,
explanation for this behavior, since the X-ray probing
quantitative differences beyond these generally
depth of around 10 nm to 40 nm (41, 42) is at most
agreeing results: (I) We do not observe two distinct
around the same as the film thickness. Therefore, even
fluence regimes with a ps recovery for low fluences and
a much thicker bulk piece of magnetite is only probed
(II) the pump for the fast decay process seems to be less
at a depth corresponding roughly to the film thickness.
efficient than in earlier studies from bulk samples; the
Different thermal conductivities of the film substrate as
fluence dependence of the fast decay amplitude 𝐴 has
compared to the bulk sample are expected to influence
been reported to scale as 0.3 rel.u./(mJ/cm2) for single
the evolution only on longer timescales than studied
crystals, while we find 0.08 rel.u./(mJ/cm2) for the thin-
here.
films studied here. A saturation of this process is not
apparent in our data, either. We discuss these Finally, the saturation of the fast process in single
differences in the following. crystals has been observed at fluences above 2 mJ/cm2,
whereas we do not see saturation up to 4.1 mJ/cm2.
The two different fluence regimes (I) have been
However, since the total decay amplitude is already
reported beyond a threshold fluence of approximately
very close to 1 at our highest fluence, we deduce from
1.3 mJ/cm² for an initial sample temperature of 80 K
Figure 3(e) that we just missed the onset of saturation,
(42). The equivalent energy of this threshold
which we expect to start at fluences around 4.3 mJ/cm2
corresponds to the heat, that would be necessary to
for our experiment.
drive the sample from its initial temperature through
the Verwey transition temperature at 124 K under In principle though, the nature of the thin film could
equilibrium conditions. Since we used a higher initial support a different fluence dependence: In thin films,
sample temperature of 100 K, the fluence threshold the presence of the Co2TiO4 substrate could influence
required to reach the high-fluence regime is expected the initial electron excitation by the IR pulse via strain
to decrease according to the altered temperature which stabilizes the ordered low-temperature phase
difference between the base temperature and the because Co2TiO4 has the same inverse spinel crystal
Verwey transition temperature to approximately structure as magnetite and only a small lattice
0.7 mJ/cm² (41, 42, 57). While in principle we can mismatch of +0.66%, resulting in the increased
extract data for very low fluence from the far wings of transition temperature in these films as compared to the
the vertical laser profile, the statistics of these data did bulk (37). The mechanism behind this stabilization is
not allow a quantitative analysis. The lowest fluence thought to be tensile strain imposed by the substrate,
data that we could analyze corresponds to the expected resulting in increased lattice constants in the film plane.
threshold fluence, which explains the absence of Enlarged inter-atomic distances facilitate the formation
recovery in the data set. of the ordered state, hence an increased transition
temperature is observed (37). As a consequence of this
Concerning the different fluence dependencies of the
stabilization of the ordered phase, the electron transfer
amplitude 𝐴 (II), we emphasize that a quantitative
from Fe2+ 3+
B to FeB sites, which is the process of the
comparison of fluences in different experiments is
initial destruction of the order, may be energetically
challenging. Small deviations coming from differing
more expensive. Per incident IR photon, fewer charge
calibrations in spot sizes and laser pulse energies,
transfer excitations could take place leading to a shift
different estimates of sample reflectivity (that also
of saturation to higher fluences. However, the observed
relate to surface quality) as well as using different
differences in fluence at which saturation of process 𝐴
incidence angles and polarizations may add up to
occurs and the slope of amplitude 𝐴 with fluence can
sizeable differences. For a quantitative comparison,
likely not be entirely explained by the influence of the
7
substrate, as the Co2TiO4substrate leads to a experimental conditions with a potential minor
comparably minor increase in transition temperature contribution from the influence of the thin-film
from 124 K (bulk) to 127 K. substrate.
Therefore, we attribute most of the quantitative Excitingly, our setup can be used to record full datasets
differences in the fluence dependence to differing on laser-induced dynamics in a single shot with
experimental conditions and according uncertainties in complete information on time- and fluence
cross-calibrating experimental variables. A potential dependences. Optimized experiments can achieve a
stabilization of the ordered phase in the thin film, dramatically increased signal level which exceeds the
originating from the substrate, could enhance the one in our demonstration by more than six orders of
observed difference, but typical reproducibility errors magnitude as we discuss in the following: The used
in such setups do not allow for a more rigorous pulses from FLASH had 20 µJ pulse energy in the
conclusion. fundamental with an expected contribution at the Fe L-
edge of about 20 nJ. Higher-energy FELs, like the
European XFEL, PALFEL, SwissFEL or LCLS, can
5 Conclusion and Outlook typically produce mJ level pulse energies at these
photon energies (58), which yields an intensity gain of
In this study, we present results of a time-to-space five orders of magnitude. Additionally, focusing the
mapping setup for time-resolved pump-probe FEL beam to a line measuring 2 × 0.2 mm2 and
experiments. This setup is capable of recording an therefore matching the studied area yields another gain
extended delay range (dependent on the dimensions of of more than an order of magnitude compared to the
the sample and the laser and FEL beams) and a fluence round spot with about 4.2 mm diameter we used here.
range of approximately one order of magnitude in a With these improvements, the signal level of a single
static setup, without the need for any scans. This is shot has the potential to substantially surpass that of the
achieved within a range of a few millimeters on the full data set presented here. We show here results from
sample where the pump laser and FEL overlap, both 12 h acquisition with an average of about
incident under a relative angle of about 70°. Due to this 0.5 counts/shot, equal to averaging about 230,000 shots
angle, the relative delay of both pulses is imprinted into from the FEL. These improvements can yield an
the horizontal dimension on the sample and imaged increase in signal of more than six orders of magnitude,
onto the detector with a Fresnel zone plate. In the demonstrating that single shot acquisition of a
vertical direction, imaging the laser spot profile allows complete dataset on ultrafast dynamics of sensitive
for the simultaneous probing of regions of the sample samples including delay and fluence dependences is
that have been pumped by higher (in the center of the feasible.
spot) or lower (at the edges of the spot) fluences.
We note though that the conditions described above
With this setup, the ultrafast phase transition dynamics result in probe fluences of about 500 mJ/cm2, which is
of high-quality magnetite (Fe3O4) thin films was certainly beyond the damage threshold of the majority
investigated, induced by 800 nm laser pulses using of solids. Consequently, single-shot acquisition
resonant soft X-ray diffraction at the iron L3-edge as becomes a necessity in a measure-before-destroy
probe. The induced dynamics generally match those approach. In order to reduce the probe fluence to a level
previously observed in single crystals: A biexponential of <1 mJ/cm2, which should be tolerable and non-
decay of charge and orbital order constituted of a fast disturbing to most samples (59), the FEL pulse could
process (sub 120 fs) and a slower process (1 ps to 2 ps). be attenuated by about one order of magnitude while
The former has been assigned to charge transfer within still yielding a data set of comparable quality to that
the B-site Fe network, while the latter has been shown here. The remaining about two orders of
attributed to the reorganization of remaining charge- magnitude can be recovered by increasing the probe
ordered islands within an otherwise metallic spot size optimized to the studied delay and pump
surrounding. The amplitudes of both decay processes fluence ranges and tolerable probe fluence.
are found to linearly depend on the fluence up to a value Consequently, the spot size of the optical pump laser
of 4.1 mJ/cm2. Although saturation is not observed at would need to be increased accordingly, necessitating
the highest fluence, we expect that saturation will occur an increase in pulse energy to adequately excite the
around 4.3 mJ/cm2. sample.
We ascribe the quantitatively different dynamic
behavior of the fluence dependence in comparison to
single crystals to mainly originate from different
8
6 Supplementary Material Author contributions

The supplementary material provides some more J.O.S. Formal Analysis (lead), Investigation (equal),
detailed information on the calibration of the pump Visualization (lead), Writing/Original Draft
laser fluences, experimental parameters (temporal Preparation (equal), Writing/Review & Editing (lead).
resolution of the setup, Fresnel zone plate optic and P.S.M. Investigation (equal), Writing/Review &
detector) and the fitting procedure of the pump-probe Editing (equal). R.Y.E. Formal Analysis (supporting),
delay traces. Investigation (equal), Writing/Review & Editing
(equal). S.D. Investigation (equal), Writing/Review &
Editing (equal). G.B. Investigation (equal),
7 Acknowledgments Writing/Review & Editing (equal). N.E. Investigation
(equal), Resources (equal), Writing/Review & Editing
The setup as well as MB, PSM, RYE and JOS are (equal). C.-F.C. Resources (equal), Writing/Review &
funded through a grant to MB for a Helmholtz young Editing (equal). P.B. Investigation (equal),
investigator group under contract number VH-NG- Writing/Review & Editing (equal). F.D. Resources
1105. CSL acknowledges funding by the Deutsche (equal), Writing/Review & Editing (equal). B.R.
Forschungsgemeinschaft (DFG, German Research Investigation (equal), Resources (equal),
Foundation)– Project-ID 328545488 – TRR 227. We Writing/Review & Editing (equal). C.D. Resources
acknowledge DESY (Hamburg, Germany), a member (equal), Writing/Review & Editing (equal). C.S-L.
of the Helmholtz Association HGF, for the provision of Conceptualization (equal), Investigation (equal),
experimental facilities. Parts of this research were Writing/Original Draft Preparation (equal),
carried out at FLASH and we would like to thank Sven Writing/Review & Editing (equal). M.B.
Toleikis for helpful discussions and assistance in using Conceptualization (lead), Funding Acquisition (lead),
beamline BL2, as well as the DESY machine and Investigation (equal), Project Administration (lead),
photon operators for assistance in using the FEL beam. Supervision (lead), Writing/Original Draft Preparation
Beamtime was allocated for proposals F-20170534 and (equal), Writing/Review & Editing (equal).
F-20181193.

9 Data Availability Statement

8 Author Declarations
The data that supports the findings of this study are
Conflicts of interest available from the corresponding author upon
reasonable request.
The authors have no conflicts to disclose.

10 References

1. J. G. Bednorz, K. A. Müller, Possible High Tc Superconductivity in the Ba-La-Cu-O System. Zeitschrift

für Phys. B Condens. Matter 64, 189–193 (1986).
2. B. Keimer, S. A. Kivelson, M. R. Norman, S. Uchida, J. Zaanen, From quantum matter to high-
temperature superconductivity in copper oxides. Nature 518, 179–186 (2015).
3. E. Dagotto, T. Hotta, A. Moreo, Colossal magnetoresistant materials: the key role of phase separation.
Phys. Rep. 344, 1–153 (2001).
4. E. Dagotto, Complexity in strongly correlated electronic systems. Science. 309, 257–262 (2005).
5. H. Takagi, B. Batlogg, H. L. Kao, J. Kwo, R. J. Cava, J. J. Krajewski, W. F. Peck, Systematic evolution of
temperature-dependent resistivity in La2−xSrxCuO4. Phys. Rev. Lett. 69, 2975–2978 (1992).
6. P. W. Phillips, N. E. Hussey, P. Abbamonte, Stranger than metals. Science 377 (2022).
7. M. Imada, A. Fujimori, Y. Tokura, Metal-insulator transitions. Rev. Mod. Phys. 70, 1039–1263 (1998).
8. J. Orenstein, Ultrafast spectroscopy of quantum materials. Phys. Today 65, 44–50 (2012).

9
9. M. Mitrano, Y. Wang, Probing light-driven quantum materials with ultrafast resonant inelastic X-ray
scattering. Commun. Phys. 3, 1–9 (2020).
10. A. de la Torre, D. M. Kennes, M. Claassen, S. Gerber, J. W. McIver, M. A. Sentef, Colloquium:
Nonthermal pathways to ultrafast control in quantum materials. Rev. Mod. Phys. 93, 041002 (2021).
11. A. Zong, B. R. Nebgen, S.-C. Lin, J. A. Spies, M. Zuerch, Emerging ultrafast techniques for studying
quantum materials. Nat. Rev. Mater. 8, 224–240 (2023).
12. A. M. Kondratenko, E. L. Saldin, Generation of Coherent Radiation by a Relativistic Electron Beam in an
Ondulator. Part. Accel. 10, 207–216 (1980).
13. S. V. Milton, E. Gluskin, N. D. Arnold, C. Benson, W. Berg, S. G. Biedron, M. Borland, Y.-C. Chae, R. J.
Dejus, P. K. Den Hartog, B. Deriy, M. Erdmann, Y. I. Eidelman, M. W. Hahne, Z. Huang, K.-J. Kim, J.
W. Lewellen, Y. Li, A. H. Lumpkin, O. Makarov, E. R. Moog, A. Nassiri, V. Sajaev, R. Soliday, B. J.
Tieman, E. M. Trakhtenberg, G. Travish, I. B. Vasserman, N. A. Vinokurov, X. J. Wang, G. Wiemerslage,
B. X. Yang, Exponential Gain and Saturation of a Self-Amplified Spontaneous Emission Free-Electron
Laser. Science 292, 2037–2041 (2001).
14. J. Rönsch-Schulenburg, E. Hass, N. M. Lockmann, T. Plath, M. Rehders, J. Roßbach, G. Brenner, S.
Dziarzhytski, T. Golz, H. Schlarb, B. Schmidt, E. Schneidmiller, S. Schreiber, B. Steffen, N. Stojanovic, S.
Wunderlich, M. Yurkov, Operation of FLASH with short SASE-FEL radiation pulses. Proc. 36th Int. Free
Electron Laser Conf. FEL 2014, 342–345 (2014).
15. T. Wang, D. Zhu, B. Wu, C. Graves, S. Schaffert, T. Rander, L. Müller, B. Vodungbo, C. Baumier, D. P.
Bernstein, B. Bräuer, V. Cros, S. de Jong, R. Delaunay, A. Fognini, R. Kukreja, S. Lee, V. López-Flores, J.
Mohanty, B. Pfau, H. Popescu, M. Sacchi, A. B. Sardinha, F. Sirotti, P. Zeitoun, M. Messerschmidt, J. J.
Turner, W. F. Schlotter, O. Hellwig, R. Mattana, N. Jaouen, F. Fortuna, Y. Acremann, C. Gutt, H. A. Dürr,
E. Beaurepaire, C. Boeglin, S. Eisebitt, G. Grübel, J. Lüning, J. Stöhr, A. O. Scherz, Femtosecond Single-
Shot Imaging of Nanoscale Ferromagnetic Order in CoPd Multilayers Using Resonant X-Ray Holography.
Phys. Rev. Lett. 108, 267403 (2012).
16. C. Klose, F. Büttner, W. Hu, C. Mazzoli, K. Litzius, R. Battistelli, S. Zayko, I. Lemesh, J. M. Bartell, M.
Huang, C. M. Günther, M. Schneider, A. Barbour, S. B. Wilkins, G. S. D. Beach, S. Eisebitt, B. Pfau,
Coherent correlation imaging for resolving fluctuating states of matter. Nature 614, 256–261 (2023).
17. S. M. Vinko, O. Ciricosta, B. I. Cho, K. Engelhorn, H.-K. Chung, C. R. D. Brown, T. Burian, J.
Chalupský, R. W. Falcone, C. Graves, V. Hájková, A. Higginbotham, L. Juha, J. Krzywinski, H. J. Lee, M.
Messerschmidt, C. D. Murphy, Y. Ping, A. Scherz, W. Schlotter, S. Toleikis, J. J. Turner, L. Vysin, T.
Wang, B. Wu, U. Zastrau, D. Zhu, R. W. Lee, P. A. Heimann, B. Nagler, J. S. Wark, Creation and
diagnosis of a solid-density plasma with an X-ray free-electron laser. Nature 482, 59–62 (2012).
18. C. David, P. Karvinen, M. Sikorski, S. Song, I. Vartiainen, C. J. Milne, A. Mozzanica, Y. Kayser, A. Diaz,
I. Mohacsi, G. A. Carini, S. Herrmann, E. Färm, M. Ritala, D. M. Fritz, A. Robert, Following the dynamics
of matter with femtosecond precision using the X-ray streaking method. Sci. Rep. 5, 7644 (2015).
19. M. Buzzi, M. Makita, L. Howald, A. Kleibert, B. Vodungbo, P. Maldonado, J. Raabe, N. Jaouen, H.
Redlin, K. Tiedtke, P. M. Oppeneer, C. David, F. Nolting, J. Lüning, Single-shot Monitoring of Ultrafast
Processes via X-ray Streaking at a Free Electron Laser. Sci. Rep. 7, 7253 (2017).
20. E. Jal, M. Makita, B. Rösner, C. David, F. Nolting, J. Raabe, T. Savchenko, A. Kleibert, F. Capotondi, E.
Pedersoli, L. Raimondi, M. Manfredda, I. Nikolov, X. Liu, A. E. D. Merhe, N. Jaouen, J. Gorchon, G.
Malinowski, M. Hehn, B. Vodungbo, J. Lüning, Single-shot time-resolved magnetic x-ray absorption at a
free-electron laser. Phys. Rev. B 99, 144305 (2019).
21. E. J. W. Verwey, Electronic Conduction of Magnetite (Fe3O4) and its Transition Point at Low
Temperatures. Nature 144, 327 (1939).
22. E. J. W. Verwey, P. W. Haayman, Electronic conductivity and transition point of magnetite (“Fe3O4”).
Physica 8, 979–987 (1941).
23. F. Walz, The Verwey transition - A topical review. J. Phys. Condens. Matter 14, 285–340 (2002).
24. J. García, G. Subías, The Verwey transition—a new perspective. J. Phys. Condens. Matter 16, R145–R178
10
 (2004).
25. S. K. Park, T. Ishikawa, Y. Tokura, Charge-gap formation upon the Verwey transition in Fe3O4. Phys. Rev.
B 58, 3717–3720 (1998).
26. J. P. Wright, J. P. Attfield, P. G. Radaelli, Long Range Charge Ordering in Magnetite Below the Verwey
Transition. Phys. Rev. Lett. 87, 266401 (2001).
27. J. P. Wright, J. P. Attfield, P. G. Radaelli, Charge ordered structure of magnetite Fe3O4 below the Verwey
transition. Phys. Rev. B 66, 214422 (2002).
28. I. Leonov, A. N. Yaresko, V. N. Antonov, M. A. Korotin, V. I. Anisimov, Charge and Orbital Order in
Fe3O4. Phys. Rev. Lett. 93, 146404 (2004).
29. D. J. Huang, H.-J. Lin, J. Okamoto, K. S. Chao, H.-T. Jeng, G. Y. Guo, C.-H. Hsu, C.-M. Huang, D. C.
Ling, W. B. Wu, C. S. Yang, C. T. Chen, Charge-Orbital Ordering and Verwey Transition in Magnetite
Measured by Resonant Soft X-Ray Scattering. Phys. Rev. Lett. 96, 096401 (2006).
30. J. Schlappa, C. Schüßler-Langeheine, C. F. Chang, H. Ott, A. Tanaka, Z. Hu, M. W. Haverkort, E.
Schierle, E. Weschke, G. Kaindl, L. H. Tjeng, Direct Observation of t2g Orbital Ordering in Magnetite.
Phys. Rev. Lett. 100, 026406 (2008).
31. A. Tanaka, C. F. Chang, M. Buchholz, C. Trabant, E. Schierle, J. Schlappa, D. Schmitz, H. Ott, P. Metcalf,
L. H. Tjeng, C. Schüßler-Langeheine, Symmetry of Orbital Order in Fe3O4 Studied by Fe L2,3 Resonant X-
Ray Diffraction. Phys. Rev. Lett. 108, 227203 (2012).
32. A. Tanaka, C. F. Chang, M. Buchholz, C. Trabant, E. Schierle, J. Schlappa, D. Schmitz, H. Ott, P. Metcalf,
L. H. Tjeng, C. Schüßler-Langeheine, Analysis of charge and orbital order in Fe3O4 by Fe L2,3 resonant x-
ray diffraction. Phys. Rev. B 88, 195110 (2013).
33. M. Hoesch, P. Piekarz, A. Bosak, M. Le Tacon, M. Krisch, A. Kozłowski, A. M. Oleś, K. Parlinski,
Anharmonicity due to Electron-Phonon Coupling in Magnetite. Phys. Rev. Lett. 110, 207204 (2013).
34. A. Bosak, D. Chernyshov, M. Hoesch, P. Piekarz, M. Le Tacon, M. Krisch, A. Kozłowski, A. M. Oleś, K.
Parlinski, Short-Range Correlations in Magnetite above the Verwey Temperature. Phys. Rev. X 4, 011040
(2014).
35. M. S. Senn, J. P. Wright, J. P. Attfield, Charge order and three-site distortions in the Verwey structure of
magnetite. Nature 481, 173–176 (2012).
36. C. F. Chang, J. Schlappa, M. Buchholz, A. Tanaka, E. Schierle, D. Schmitz, H. Ott, R. Sutarto, T. Willers,
P. Metcalf, L. H. Tjeng, C. Schüßler-Langeheine, Intrinsic and extrinsic x-ray absorption effects in soft x-
ray diffraction from the superstructure in magnetite. Phys. Rev. B 83, 073105 (2011).
37. X. Liu, C.-F. Chang, A. D. Rata, A. C. Komarek, L. H. Tjeng, Fe3O4 thin films: controlling and
manipulating an elusive quantum material. npj Quantum Mater. 1, 16027 (2016).
38. W. Eerenstein, T. T. M. Palstra, T. Hibma, S. Celotto, Origin of the increased resistivity in epitaxial Fe3O4
films. Phys. Rev. B 66, 201101 (2002).
39. S. K. Arora, R. G. S. Sofin, I. V. Shvets, Magnetoresistance enhancement in epitaxial magnetite films
grown on vicinal substrates. Phys. Rev. B 72, 134404 (2005).
40. X. H. Liu, A. D. Rata, C. F. Chang, A. C. Komarek, L. H. Tjeng, Verwey transition in Fe3O4 thin films:
Influence of oxygen stoichiometry and substrate-induced microstructure. Phys. Rev. B 90, 125142 (2014).
41. N. Pontius, T. Kachel, C. Schüßler-Langeheine, W. F. Schlotter, M. Beye, F. Sorgenfrei, C. F. Chang, A.
Föhlisch, W. Wurth, P. Metcalf, I. Leonov, A. Yaresko, N. Stojanovic, M. Berglund, N. Guerassimova, S.
Düsterer, H. Redlin, H. A. Dürr, Time-resolved resonant soft x-ray diffraction with free-electron lasers:
Femtosecond dynamics across the Verwey transition in magnetite. Appl. Phys. Lett. 98, 182504 (2011).
42. S. de Jong, R. Kukreja, C. Trabant, N. Pontius, C. F. Chang, T. Kachel, M. Beye, F. Sorgenfrei, C. H.
Back, B. Bräuer, W. F. Schlotter, J. J. Turner, O. Krupin, M. Doehler, D. Zhu, M. A. Hossain, A. O.
Scherz, D. Fausti, F. Novelli, M. Esposito, W. S. Lee, Y. D. Chuang, D. H. Lu, R. G. Moore, M. Yi, M.
Trigo, P. Kirchmann, L. Pathey, M. S. Golden, M. Buchholz, P. Metcalf, F. Parmigiani, W. Wurth, A.

11
 Föhlisch, C. Schüßler-Langeheine, H. A. Dürr, Speed limit of the insulator–metal transition in magnetite.
Nat. Mater. 12, 882–886 (2013).
43. F. Randi, I. Vergara, F. Novelli, M. Esposito, M. Dell’Angela, V. A. M. Brabers, P. Metcalf, R. Kukreja,
H. A. Dürr, D. Fausti, M. Grüninger, F. Parmigiani, Phase separation in the nonequilibrium Verwey
transition in magnetite. Phys. Rev. B 93, 054305 (2016).
44. E. Baldini, C. A. Belvin, M. Rodriguez-Vega, I. O. Ozel, D. Legut, A. Kozłowski, A. M. Oleś, K.
Parlinski, P. Piekarz, J. Lorenzana, G. A. Fiete, N. Gedik, Discovery of the soft electronic modes of the
trimeron order in magnetite. Nat. Phys. 16, 541–545 (2020).
45. W. Wang, J. Li, Z. Liang, L. Wu, P. M. Lozano, A. C. Komarek, X. Shen, A. H. Reid, X. Wang, Q. Li, W.
Yin, K. Sun, I. K. Robinson, Y. Zhu, M. P. M. Dean, J. Tao, Verwey transition as evolution from
electronic nematicity to trimerons via electron-phonon coupling. Sci. Adv. 9, eadf8220 (2023).
46. B. Truc, P. Usai, F. Pennacchio, G. Berruto, R. Claude, I. Madan, V. Sala, T. LaGrange, G. M. Vanacore,
S. Benhabib, F. Carbone, Ultrafast generation of hidden phases via energy-tuned electronic photoexcitation
in magnetite. Proc. Natl. Acad. Sci. 121, e2316438121 (2024).
47. K. Tiedtke, A. Azima, N. von Bargen, L. Bittner, S. Bonfigt, S. Düsterer, B. Faatz, U. Frühling, M.
Gensch, C. Gerth, N. Guerassimova, U. Hahn, T. Hans, M. Hesse, K. Honkavaar, U. Jastrow, P. Juranic, S.
Kapitzki, B. Keitel, T. Kracht, M. Kuhlmann, W. B. Li, M. Martins, T. Núñez, E. Plönjes, H. Redlin, E. L.
Saldin, E. A. Schneidmiller, J. R. Schneider, S. Schreiber, N. Stojanovic, F. Tavella, S. Toleikis, R.
Treusch, H. Weigelt, M. Wellhöfer, H. Wabnitz, M. V. Yurkov, J. Feldhaus, The soft x-ray free-electron
laser FLASH at DESY: beamlines, diagnostics and end-stations. New J. Phys. 11, 023029 (2009).
48. J. Feldhaus, FLASH—the first soft x-ray free electron laser (FEL) user facility. J. Phys. B At. Mol. Opt.
Phys. 43, 194002 (2010).
49. S. Toleikis, The FLASH facility current status in 2018 and future upgrade plans. AIP Conf. Proc. 2054,
030015 (2019).
50. M. Beye, R. Y. Engel, J. O. Schunck, S. Dziarzhytski, G. Brenner, P. S. Miedema, Non-linear soft x-ray
methods on solids with MUSIX—the multi-dimensional spectroscopy and inelastic x-ray scattering
endstation. J. Phys. Condens. Matter 31, 014003 (2019).
51. E. L. Saldin, E. A. Schneidmiller, M. V. Yurkov, Properties of the third harmonic of the radiation from
self-amplified spontaneous emission free electron laser. Phys. Rev. Spec. Top. - Accel. Beams 9, 030702
(2006).
52. S. Düsterer, P. Radcliffe, G. Geloni, U. Jastrow, M. Kuhlmann, E. Plönjes, K. Tiedtke, R. Treusch, J.
Feldhaus, P. Nicolosi, L. Poletto, P. Yeates, H. Luna, J. T. Costello, P. Orr, D. Cubaynes, M. Meyer,
Spectroscopic characterization of vacuum ultraviolet free electron laser pulses. Opt. Lett. 31, 1750 (2006).
53. W. Ackermann, G. Asova, V. Ayvazyan, A. Azima, N. Baboi, J. Bähr, V. Balandin, B. Beutner, A. Brandt,
A. Bolzmann, R. Brinkmann, O. I. Brovko, M. Castellano, P. Castro, L. Catani, E. Chiadroni, S. Choroba,
A. Cianchi, J. T. Costello, D. Cubaynes, J. Dardis, W. Decking, H. Delsim-Hashemi, A. Delserieys, G. Di
Pirro, M. Dohlus, S. Düsterer, A. Eckhardt, H. T. Edwards, B. Faatz, J. Feldhaus, K. Flöttmann, J. Frisch,
L. Fröhlich, T. Garvey, U. Gensch, C. Gerth, M. Görler, N. Golubeva, H.-J. Grabosch, M. Grecki, O.
Grimm, K. Hacker, U. Hahn, J. H. Han, K. Honkavaara, T. Hott, M. Hüning, Y. Ivanisenko, E. Jaeschke,
W. Jalmuzna, T. Jezynski, R. Kammering, V. Katalev, K. Kavanagh, E. T. Kennedy, S. Khodyachykh, K.
Klose, V. Kocharyan, M. Körfer, M. Kollewe, W. Koprek, S. Korepanov, D. Kostin, M. Krassilnikov, G.
Kube, M. Kuhlmann, C. L. S. Lewis, L. Lilje, T. Limberg, D. Lipka, F. Löhl, H. Luna, M. Luong, M.
Martins, M. Meyer, P. Michelato, V. Miltchev, W. D. Möller, L. Monaco, W. F. O. Müller, O. Napieralski,
O. Napoly, P. Nicolosi, D. Nölle, T. Nuñez, A. Oppelt, C. Pagani, R. Paparella, N. Pchalek, J. Pedregosa-
Gutierrez, B. Petersen, B. Petrosyan, G. Petrosyan, L. Petrosyan, J. Pflüger, E. Plönjes, L. Poletto, K.
Pozniak, E. Prat, D. Proch, P. Pucyk, P. Radcliffe, H. Redlin, K. Rehlich, M. Richter, M. Roehrs, J.
Roensch, R. Romaniuk, M. Ross, J. Rossbach, V. Rybnikov, M. Sachwitz, E. L. Saldin, W. Sandner, H.
Schlarb, B. Schmidt, M. Schmitz, P. Schmüser, J. R. Schneider, E. A. Schneidmiller, S. Schnepp, S.
Schreiber, M. Seidel, D. Sertore, A. V. Shabunov, C. Simon, S. Simrock, E. Sombrowski, A. A. Sorokin,
P. Spanknebel, R. Spesyvtsev, L. Staykov, B. Steffen, F. Stephan, F. Stulle, H. Thom, K. Tiedtke, M.
Tischer, S. Toleikis, R. Treusch, D. Trines, I. Tsakov, E. Vogel, T. Weiland, H. Weise, M. Wellhöfer, M.
Wendt, I. Will, A. Winter, K. Wittenburg, W. Wurth, P. Yeates, M. V. Yurkov, I. Zagorodnov, K. Zapfe,
12
 Operation of a free-electron laser from the extreme ultraviolet to the water window. Nat. Photonics 1, 336–
342 (2007).
54. H. Redlin, A. Al-Shemmary, A. Azima, N. Stojanovic, F. Tavella, I. Will, S. Düsterer, The FLASH pump–
probe laser system: Setup, characterization and optical beamlines. Nucl. Instruments Methods Phys. Res.
Sect. A Accel. Spectrometers, Detect. Assoc. Equip. 635, S88–S93 (2011).
55. J. O. Schunck, F. Döring, B. Rösner, J. Buck, R. Y. Engel, P. S. Miedema, S. K. Mahatha, M. Hoesch, A.
Petraru, H. Kohlstedt, C. Schüßler-Langeheine, K. Rossnagel, C. David, M. Beye, Soft x-ray imaging
spectroscopy with micrometer resolution. Optica 8, 156 (2021).
56. N. Pontius, M. Beye, C. Trabant, R. Mitzner, F. Sorgenfrei, T. Kachel, M. Wöstmann, S. Roling, H.
Zacharias, R. Ivanov, R. Treusch, M. Buchholz, P. Metcalf, C. Schüßler-Langeheine, A. Föhlisch, Probing
the non-equilibrium transient state in magnetite by a jitter-free two-color X-ray pump and X-ray probe
experiment. Struct. Dyn. 5, 054501 (2018).
57. E. F. Westrum, F. Grønvold, Magnetite (Fe3O4) Heat capacity and thermodynamic properties from 5 to 350
K, low-temperature transition. J. Chem. Thermodyn. 1, 543–557 (1969).
58. T. Tschentscher, C. Bressler, J. Grünert, A. Madsen, A. Mancuso, M. Meyer, A. Scherz, H. Sinn, U.
Zastrau, Photon Beam Transport and Scientific Instruments at the European XFEL. Appl. Sci. 7, 592
(2017).
59. R. Y. Engel, O. Alexander, K. Atak, U. Bovensiepen, J. Buck, R. Carley, M. Cascella, V. Chardonnet, G.
S. Chiuzbaian, C. David, F. Döring, A. Eschenlohr, N. Gerasimova, F. de Groot, L. Le Guyader, O. S.
Humphries, M. Izquierdo, E. Jal, A. Kubec, T. Laarmann, C.-H. Lambert, J. Lüning, J. P. Marangos, L.
Mercadier, G. Mercurio, P. S. Miedema, K. Ollefs, B. Pfau, B. Rösner, K. Rossnagel, N. Rothenbach, A.
Scherz, J. Schlappa, M. Scholz, J. O. Schunck, K. Setoodehnia, C. Stamm, S. Techert, S. M. Vinko, H.
Wende, A. A. Yaroslavtsev, Z. Yin, M. Beye, Electron population dynamics in resonant non-linear x-ray
absorption in nickel at a free-electron laser. Struct. Dyn. 10, 054501 (2023).

13