REVIEW OF SCIENTIFIC INSTRUMENTS 82, 033703 (2011)
Time-resolved imaging of laser-induced refractive index changes
in transparent media
Alexandre Mermillod-Blondin,1,a)Cyril Mauclair,1,2Jörn Bonse,3Razvan Stoian,2
Eric Audouard,2Arkadi Rosenfeld,1and Ingolf V. Hertel1,b)
1Max-Born-Institut für Nichtlineare Optik und Kurzzeitspektroskopie, Max-Born-Straße, D-12489 Berlin,
2Laboratoire Hubert Curien (UMR 5516 CNRS), Université de Lyon, Université Jean Monnet, 42000 Saint
3BAM Bundesanstalt für Materialforschung und-prüfung, Fachgruppe VI.4 Oberflächentechnologien, Unter
den Eichen 87, D-12205 Berlin, Germany
(Received 8 October 2010; accepted 23 November 2010; published online 2 March 2011)
We describe a method to visualize ultrafast laser-induced refractive index changes in transparent
materials with a 310 fs impulse response and a submicrometer spatial resolution. The temporal
profile of the laser excitation sequence can be arbitrarily set on the subpicosecond and picosec-
ond time scales with a pulse shaping unit, allowing for complex laser excitation. Time-resolved
phase contrast microscopy reveals the real part of the refractive index change and complemen-
tary time-resolved optical transmission microscopy measurements give access to the imaginary part
of the refractive index in the irradiated region. A femtosecond laser source probes the complex
refractive index changes from the excitation time up to 1 ns, and a frequency-doubled Nd:YAG
laser emitting 1 ns duration pulses is employed for collecting data at longer time delays, when
the evolution is slow. We demonstrate the performance of our setup by studying the energy re-
laxation in a fused silica sample after irradiation with a double pulse sequence. The excitation
pulses are separated by 3 ps. Our results show two dimensional refractive index maps at differ-
ent times from 200 fs to 100 μs after the laser excitation. On the subpicosecond time scale we
have access to the spatial characteristics of the energy deposition into the sample. At longer times
(800 ps), time-resolved phase contrast microscopy shows the appearance of a strong compression
wave emitted from the excited region. On the microsecond time scale, we observe energy transfer
outside the irradiated region. © 2011 American Institute of Physics. [doi:10.1063/1.3527937]
Femtosecond laser-induced modifications of transparent
materials have triggered a considerable amount of theoreti-
cal and experimental work since the advent of stable fem-
tosecond laser sources in the 1990s. The popularity of this
research theme arises from the fact that femtosecond laser-
induced modification of transparent media implies very chal-
lenging fundamental aspects but also opens the door to a num-
ber of exciting applications. The fundamental aspects cover
the whole interaction scenario from free carrier generation to
structural modifications. Free carrier generation is still nowa-
days a very active research field, experimentally as well as
theoretically.1–4The decay of free carriers into localized point
defects has been studied since decades,5–10but their contri-
bution is not sufficient to explain the magnitude of the re-
fractive index changes observed. Local densification due to
a change in the structure of fused silica upon heating also has
to be taken into account.11Very promising applications were
reported in glass when the femtosecond laser beam is used
as a microstucturing tool. In particular, direct laser writing
of waveguides12–15including active waveguide devices,16mi-
crochannel drilling,17,18diffraction gratings,19,20lenses,21,22
a)Electronic mail: email@example.com.
b)Also at Department of Physics, Free University of Berlin, 14195 Berlin,
or more complex structures23–25paving the way for laser
sculpted lab on chip devices.26,27Some research groups have
explored the possibility of microprocessing bulk materials
with temporally shaped pulses.28–31Using temporally shaped
pulses extends micromachining capabilities but also brings on
new theoretical challenges.
All the applications aforementioned involve tight focus-
ing. However, despite of strong localization, the dynamic spa-
tial redistribution and relaxation of energy in excited matter is
important. Studying the spatiotemporal dynamics of the mate-
rial modification where the laser beam size is focused down to
a size of a few micrometers is a very challenging task. Though
time-resolved techniques exist for a large range of applica-
tions, we present below the most popular methods that were
recently developed to monitor the temporal evolution of laser-
induced complex refractive index changes. In addition to the
following brief review, the reader is also encouraged to con-
sult Ref. 32. Time-resolved optical transmission microscopy
(OTM) (also called shadowgraphy or microshadowgraphy) is
a method which provides direct information about the dy-
namic absorption in the plasma region. It has been employed
for monitoring laser-induced plasmas in various transparent
media under tight focusing conditions, for instance, in air,33
water,34,35or glassy materials.9,36,37In addition to transient
absorption, complementary methods revealing the transient
evolution of the real part of the laser-induced refractive index
0034-6748/2011/82(3)/033703/8/$30.00© 2011 American Institute of Physics
033703-2Mermillod-Blondin et al.Rev. Sci. Instrum. 82, 033703 (2011)
changes were also reported. Those methods mostly rely on
interferometric measurements.9,33,34,38As a downside, inter-
ferometry delivers sets of fringes patterns that have to be ana-
lyzed in order to retrieve the refractive index distribution. The
same limitation applies for in-line holography39and transient
quantitative phase microscopy40where various images have
to be analyzed in order to retrieve the relevant information by
solving light transport equations, but with the advantage of
a high temporal resolution. Conversely, phase-contrast-based
methods directly supply images where the intensity distri-
bution is proportional to the phase changes experienced by
the probe beam. The proportionality between refractive index
change and intensity in the image plane is limited to the case
of optical path differences between bulk and modified mate-
rial much smaller than the observation wavelength.41How-
ever, because the transverse dimension of the laser-induced
modifications lies in the micrometer range and the expected
refractive index changes are modest (≈10−3in fused silica),13
the small optical path difference condition is always fulfilled
in our experiments. The idea of exploiting the principles of
phase contrast for the observation of transient phase objects
was already developed earlier42,43for probing plasma density
fluctuations. In both cases, macroscopic objects were studied.
In the previous work,44we presented results obtained with a
time-resolved phase contrast microscopy (PCM) setup in its
early development phase with limitations for probing at time
delays longer than 1 ns. Moreover, our setup did not offer the
possibility to work with temporally shaped pump pulses.
In this paper, we present and characterize an extended
time-resolved microscopy apparatus providing snapshots of
the laser-induced transient refractive index distribution. It per-
forms a double function: dynamic phase contrast and dy-
namic optical transmission microscopy. The optical transmis-
sion configuration allows monitoring the dynamic absorption
in the region of laser-matter interaction. The laser excitation
sequence can be programmatically modulated with a tempo-
ral pulse shaping unit based on spectral phase synthesis. In
Sect. II, we describe the experimental arrangement in detail
and characterize the spatiotemporal resolution of our setup.
The spatial resolution is estimated by employing the slanted
edge method. The measurement of the impulse response is
performed by deconvolution. In Sec. III, we demonstrate the
capabilities of time-resolved phase-contrast microscopy and
optical transmission microscopy by investigating the dynamic
response of an amorphous silica sample to a double pulse ir-
radiation sequence. We also look into the problem of image
representation. Finally, we discuss and comment the exper-
imental results. Some limitations of our setup are presented
with possible workarounds.
II. MATERIALS AND METHODS
A. Setup overview
This pump-probe setup (see Fig. 1) is structured in three
blocks: the pump pulse shaper, the phase contrast microscope,
and the microscope illumination.
We employ an amplified femtosecond laser source (150
fs pulse duration, central wavelength 800 nm) as the excita-
FIG. 1. (Color online) Sketch of the experimental setup. EMCCD: electron
multiplying charged coupled device, DG: diffraction grating, SLM: spatial
light modulator, PC: phase contrast, OTM: optical transmission microscopy,
SHG: second harmonic generation, and BD: beam diffuser.
tion source. Aside from natural temporal profile of the laser
pulses (i.e., quasi-Gaussian), the sample can be excited with
arbitrary temporal envelopes by using the pulse shaping unit.
The pulse shaping unit is described in Fig.1. A programmable
spatial light modulator placed in the Fourier plane of a zero
dispersion stretcher induces a controlled dephasing over the
spectral components of the input pulse.45This spectral phase
When the spectral dephasing is constant over the spectrum (or
when no signal is applied on the modulator), the output pulse
is identical to the input pulse, offering the possibility to oper-
ate with an unshaped pulse. We focus the output laser pulse
inside the sample with a microscope objective of numer-
ical aperture of 0.45 (50×). In order to limit wavefront
distortions,46we set the geometrical focal plane at a distance
of 200 μm from the air dielectric interface.
The phase contrast/transmission microscope is based on
a commercially available optical microscope (BX41 from
Olympus) equipped with a phase contrast module. The phase
contrast module consists of two elements: a substage con-
denser and a set of phase contrast microscope objectives.
The substage condenser concentrates the illumination source
onto the sample. In phase contrast mode (PCM), the en-
trance pupil of the condenser is annular. A positive high phase
contrast objective (20× magnification) including a phase-
retarding plate forms the image of the sample on the sensor
of an electron multiplying CCD (EMCCD) camera (Andor
iXon) through a 2× field expander. In optical transmission
microscopy (OTM), the phase contrast microscope objective
is replaced by a conventional microscope objective. Although
in OTM the shape of the entrance pupil is traditionally circu-
lar, we decided in favor of the same annular entrance pupil as
in the phase contrast mode in order to keep the illumination
conditions and the temporal characteristics of the probe beam
as similar as possible in both modes (PCM and OTM).
The microscope design is modified so that a pulsed laser
beam can be used as the illumination source. We opted for
033703-3Mermillod-Blondin et al. Rev. Sci. Instrum. 82, 033703 (2011)
critical illumination. As a major upside, critical illumination
is well adapted to parallel illumination sources and requires
a minimal amount of optical components which lead to ultra-
fast pulse elongation. More specifically, there is no collector
lens nor field lens in this illumination scheme. Two lasers are
alternatively employed. For snapshots with a subpicosecond
resolution, a beam splitter divides the beam before the pulse
shaping unit as depicted in Fig. 1. The part of the beam which
does not enter the shaper is used for illumination purposes. By
optically delaying the arrival time of the laser pulses into the
sample, we are able to cover a time span from the subpicosec-
ond up to the nanosecond range. In order to discriminate the
photons originating from the probe, we perform a frequency
doubling to 400 nm before homogenizing the beam. For probe
pulse delays longer than 10 ns, a frequency-doubled Nd:YAG
(1 ns pulse duration, 532 nm wavelength) is synchronized and
delayed electronically with respect to the femtosecond pump
pulse. A 5 nm wide optical bandpass filter centered on the
illumination wavelength (532 or 400 nm) rejects light that
does not carry image information. This parasitic light has two
origins, the laser-induced electron hole plasma luminescence
and the scattering of the trailing edge of the laser pulse by
the transient and structural modifications created by the lead-
ing edge of the pulse. Using a laser beam as an illumination
source is a challenging task, mainly because of its coherence.
Any scattering center or imperfect optical surface results in
complex interferences patterns deteriorating dramatically the
image quality.47Therefore, the spatial coherence of the laser
must be reduced. Moreover, our laser sources deliver gaussian
beams, which implies that the beam also has to be homoge-
nized. For those purposes, we use a pair of beam diffusers
BD1 and BD2. When the nanosecond laser is used as an illu-
mination source, BD1 is a moving optical fiber48and BD2 is a
ground glass diffuser. When using femtosecond pulses for the
illumination, the optical fiber is replaced by a rotating glass
diffuser to preserve a short pulse duration. Beam diffusers
generate speckle. The detrimental influence of the speckle is
minimized by taking 20 snapshots of the same event for dif-
ferent positions of the beam diffusers.
B. Spatiotemporal characteristics
1. Impulse response
The impulse response is determined by the duration of
the probe pulse in the focal plane of the condenser. A direct
measurement is not possible at least for two reasons. First, the
probe pulse is a 400 nm beam. This precludes using a method
based on the two-photon response of a GaAsP photodiode.49
Second, we are interested in the probe pulse duration in the
plane of the laser-induced modification, i.e., inside the sam-
ple. Therefore, we apply a method based on temporal signal
deconvolution. We search for the gating function Ig(t) such as
Aexpt.(t) = Ath(t) ⊗ Ig(t),
where ⊗ denotes the convolution product, Ath(t) is the theo-
retical electron gas absorbance, and Ig(t) represents the inten-
sity profile of the probe pulse. The experimental absorbance
Aexpt.(t) is measured by using our setup in OTM. We focus the
laser pump pulse in a fused silica sample with a microscope
objective having numerical aperture of 0.12. The energy per
to the laser direction so that each laser pulse is focused on a
fresh spot. The electron gas absorbance Ath(t) and its density
ne(t) are related through the relation9
where Ipd and Ip0correspond to the probe intensity atten-
uated by the electron plasma and the background intensity,
respectively. The term σ is the absorption cross section for
inverse bremsstrahlung and x denotes the propagation direc-
tion of the probe beam. The integration limits correspond to
the depth of field (DOF) of the recorded image about 2.9 μm
in our experimental conditions. We perform the analysis in
a region where the electron gas thickness is slightly greater
than the DOF. This allows making the approximation that the
free electron density is constant over the DOF and there are
no contributions from regions out of focus. In this case, Eq.
(2) shows that Ath(t) and ne(t) are proportional. A theoretical
estimate of ne(t) is obtained by solving the free electron gen-
eration and continuity equation in a laser field involving photo
and collisional ionization (see, for instance, Ref. 50 for a de-
tailed description of the free electron generation and continu-
ity equation). The continuity equation was solved employing
a trapping time of 150 fs for the laser-induced free electrons
as estimated by Audebert and co-workers.51We emphasize
that the trapping time strongly depends on the laser intensity.
Therefore, we ensured that the laser intensity is ≈2 × 1013
W/cm2, close to the 2.7 × 1013W/cm2reported in Ref. 51, in
the region where Ipdis measured.
After normalization of Athand Aexpt., we retrieve Ig(t)
by solving Eq. (1). A solution of Eq. (1) is hard to obtain
by direct numerical deconvolution for stability reasons. We
overcome this difficulty by assuming that Ig(t) has a gaus-
sian shape. This assumption implies that the major source of
alteration experienced by the probe pulse is a temporal broad-
ening, consistently with the existing theoretical models.52–54
The only parameter to be found is then the full width at half
maximum (FWHM) of the probe pulse. The optimal FWHM
is determined by iterative deconvolution. We define a range
of FWHMs to be tested. For each FWHM value, we compare
the result of Ath(t) ⊗ Ig(t) to Aexpt.(t). The probe pulse has
the FWHM that minimizes the sum of the square residuals in
a least-squares fit.
Figure 2 shows the time resolution obtained for a circular
condenser (the entrance pupil of the condenser is a disk) and
for an annular condenser (the entrance pupil is the annulus
used in phase contrast mode). We find a resolution of 720 fs
with a circular condenser and 310 fs when using an annular
condenser. We essentially attribute the probe pulse elongation
to temporal broadening in the condenser. The condenser is
made of a thick lens of B270 Schott glass. In addition to group
velocity dispersion, chromatic aberration leads to a difference
in the arrival time of the different rays in the focal plane of
the condenser, translating into a pulse elongation.52Using
an annular entrance pupil brings significant improvements.
Ath(t) = lnIp0
033703-4Mermillod-Blondin et al. Rev. Sci. Instrum. 82, 033703 (2011)
FIG. 2. (Color online) (a) Estimate of the probe pulse duration in the focal
plane of the condenser. The circles and the squares represent the absorbance
recorded for the same excitation conditions in fused silica at different pump-
probe time delays with a circular and annular condenser, respectively. The
solid lines represent the result of the theoretical absorbance convoluted by a
720 fs FWHM gaussian pulse (red curve) and by a 310 fs FWHM gaussian
pulse (blue curve). (b) Time-resolved observations in optical transmission
microscopy in fused silica for various pump-probe delays τ in the plane of
incidence of the laser with a circular condenser. The laser is focused by a
microscope objective with a 0.12 numerical aperture. The absorbance values
presented in (a) were calculated from the regions Ipd and Ipo (see text) shown
in the figure. The laser intensity is ≈2 × 1013W/cm2in the region Ipd. (c)
Same as (b) for an annular condenser.
Theoretically, the pulse duration can even be reduced down to
the same value that would be expected when using monochro-
matic light by choosing the proper annular aperture.53How-
ever, in our case, the dimensions of the annular aperture are
in the phase contrast objective, preventing us from reaching a
pulse duration better than 310 fs.
In conclusion, we emphasize two points. First, the time
response of our setup is currently limited to approximately
300 fs (about twice the probe pulse duration). The time re-
sponse can be improved when using a shorter pulse duration
for illumination. Second, to be able to properly correlate PCM
and OTM pictures on the femtosecond and picosecond time
scales, we must use the same annular pupil for the condenser.
2. Spatial resolution
The spatial resolution indicates the size of the smallest
features visible in the acquired microscopy images. As estab-
lished in the previous paragraph, we use an annular condenser
in OTM and PCM. According to Ref. 55, employing an an-
nular condenser has consequences on the spatial resolution.
Moreover, we use illumination sources with different wave-
lengths (400 and 532 nm) and different coherence properties.
This certainly has consequences on the point spread function
FIG. 3. (Color online) Estimate of the spatial resolution. The green and blue
solid lines represent the MTF values provided by the slanted edge method
when the microscope illumination is performed with a pulsed nanosec-
ond laser (λ = 532 nm) and a frequency doubled femtosecond laser (λ
= 400 nm), respectively. The circles and the crosses show the correspond-
ing MTF values directly measured with a 1951 USAF resolution target. The
dashed lines correspond to the spatial resolution according to the Rayleigh
criterion. The spatial resolution is ≈670 nm (1.49 line pairs/μm) when the
illumination is performed with the nanosecond laser and ≈610 nm (1.64 line
pairs/μm) when the illumination is performed with the femtosecond laser.
of our microscopy system. Therefore, it is relevant to proceed
to a careful study of the spatial resolution of our microscope.
A good indicator for the spatial resolution is the mod-
ulation transfer function (MTF). The MTF is defined as the
real part of the Fourier transform of the point spread func-
tion. We measured the MTF in OTM mode by applying the
slanted edge method.56The edge corresponds to the lower
side of the greater square from a high resolution USAF 1951
target. Because of the speckled illumination, we could not
process the slanted edge pictures directly in the Fourier do-
main. Instead, we fitted the edge spread function with an er-
ror function.57Figure 3 shows the results for an annular con-
denser and the two illumination sources. The MTF values at
the origin (for a zero spatial frequency) correspond to the
maximum contrast C = (Ibkg− Iblack)/(Ibkg+ Iblack) reach-
able, where Ibkgand Iblackare the mean value of all pixels
in the field of view in the absence of sample and when the
illumination is obstructed, respectively. We also reported in
Fig. 3 the measurements obtained with patterns of three bars
from a high resolution USAF 1951 target. These data points
are provided to compare the slanted edge methods results to
a direct measurement technique. Nevertheless, we are aware
that bar patterns measurements have to be interpreted cau-
tiously. First, bar patterns do not directly give MTF values, as
they do not contain single spatial frequencies. Rigourously,
those targets provide a measurement of the contrast transfer
function (CTF). At low frequencies, the CTF has higher val-
ues than the MTF. As the data points obtained with the USAF
target are comparable to the slanted edge method results, we
conclude that we may overestimate the real MTF values by a
few percent when using the slanted edge in combination with
a fit procedure. The relatively weak contrast of low frequency
objects (47% with the femtosecond laser and 73% with the
nanosecond laser, see Fig. 3) is attributed to detector noise,
but also to the presence of speckle. Hence, further optimiza-
tion of the beam diffusers may allow improving the overall
033703-5Mermillod-Blondin et al.Rev. Sci. Instrum. 82, 033703 (2011)
A useful figure of merit when describing the spatial re-
sponse of an optical system is the Rayleigh criterion. If we
suppose that there are no significant aberrations in our optical
system, the Rayleigh criterion corresponds to the spatial fre-
quency at which the MTF is 9%, i.e., 1.64 line pairs per μm
(≈610 nm) when using the 400 nm laser beam and 1.49 line
pairs per μm and (≈670 nm) when using the 532 nm laser
beam. These results reflect the performance of our whole mi-
croscopy system including illumination, optics, and EMCCD
detector. Therefore, the spatial resolution lies significantly be-
low the diffraction limit (444 and 591 nm for 400 and 532 nm
illumination wavelengths, respectively).
A. Experimental results
We demonstrate the capability of our setup of analyzing
complex excitation effects by studying the laser energy cou-
pling and relaxation in a fused silica sample under single shot
irradiation with a double pump pulse sequence generated by
the pulse shaper. The choice of a double pulse sequence is ar-
bitrary. In principle, any pulse shape that can be produced by
the pulse shaping unit might be employed. However, double
pulse sequences appear in a number of laser micromachining
studies,29,58,59making it a relevant example of pulse shaping.
The double pulse sequence we used had the energy of 4 μJ
(measured after the focusing objective). It is generated by ap-
plying a square pattern on the SLM serving as a spectral phase
modulation with an amplitude close to ±π. The frequency of
the square pattern is such as a double pulse with a 3 ps separa-
tion time is obtained. By arbitrary convention, the first pulse
arrives at −1.5 ps and the second pulse arrives at 1.5 ps. A
rectangular phase function usually provides regularly spaced
replica at multiple of the separation distance that were mini-
mized by a careful calibration of the phase. The fused silica
sample is continuously scanned transversally with respect to
the laser direction so that every laser double pulse sequence
interacts with a pristine material.
The results are presented in Fig. 4 for pump-probe delays
τ of −1.2 ps (i.e., approximatively 300 fs after the arrival of
the first pulse), 1.8 ps (i.e., approximatively 300 fs after the
arrival of the second pulse), 800 ps, 20 ns, 10 μs, and 100 μs.
We show pictures obtained in PCM and OTM modes when
accumulating 20 pictures of irradiation events obtained under
identical conditions. The laser beam comes from the left side
and the plane of incidence corresponds to the plane in which
the pictures were recorded.
B. Data representation
PCM and OTM data visualization deserves a special at-
tention. The EMCCD camera delivers 14 bit gray scale im-
ages, where the pixel gray values are proportional to the inci-
dent light intensity. In order to reveal fine details, a uniform
gray scale representation is not optimal, and it is of common
practice to use a gradient color scale. As a major difficulty,
gradient color scales usually only contain 256 distinct colors.
The challenge is hence to represent 214= 16384 values with
only 256 colors in an optimal manner. For this purpose, we
apply a transformation algorithm to perform a 14 to 8 bit con-
version without significant loss of information. In the OTM
mode, the information lies in the absorption region, and in
fractive index change varies. When pictures are recorded with
a sufficient dynamic range, the gray value histograms corre-
sponding to image information and image background show
very little overlapping. In this case, it is possible to optimize
data representation by assigning different color depths to the
information and the background. For example, in Fig. 4(a), in
OTM mode, the background information is represented using
25 distinct color bins, whereas the other 230 bins are dedi-
cated to the representation of the absorption region. A linear
conversion of 14 to 8 bits would result in representing the
background with 187 values, leaving only 68 bins for repre-
senting the region containing physical information.
A. Optical transmission microscopy
The OTM pictures show the temporal absorption in the
interaction region. The local attenuation of the probe light
is attributed to two factors: inverse bremsstrahlung and ab-
sorption by point defects. Inverse Bremsstrahlung absorption
by the laser-induced electron hole plasma plays a dominant
role immediately after laser irradiation, when electrons have
just been injected into the conduction band and did not start
decaying into point defects. Some point defects have energy
levels close to the bandgap (less than 3 eV) that can absorb
probe beam photons.6,7
In Fig. 4, for τ = −1.2 ps, we observe a conically shaped
electron gas. This electron gas is generated by the first pulse
through the combined effect of multiphoton and avalanche
ionization. Although the energy is equally distributed
in the direction of the laser upon irradiation with the second
pulse. This reveals that precursors were generated by the first
pulse in the vicinity of the absorption region. Elucidating the
origin of those precursors is beyond the scope of this article
and will be performed in the future.
At this point, we emphasize that the absorption magni-
tude does not reflect the density of free electrons nor the de-
fect density, but rather the product of the linear attenuation
along the optical path. This prevents from leading quantita-
tive analysis directly from the raw picture. Extracting quanti-
tative values for the free electron density ne(t) in the region
of permanent modification is a complex problem involving
assumptions on the collisional scattering time and optical car-
rier masses. In this region, the depth of field is larger than the
extent of the electron gas. In this case, it is inappropriate to
neglect the radial dependence of ne(t) in Eq. (2) as we previ-
ously did, and it is necessary to perform an Abel inversion.
The absorption region vanishes at τ = 10 μs. This indi-
cates that the electron gas and the majority of the point defects
with an energy level close to the conduction band decayed
within this time frame. Nevertheless, structural modifications
with strong scattering or absorbing properties persist.
033703-6Mermillod-Blondin et al.Rev. Sci. Instrum. 82, 033703 (2011)
FIG. 4. (Color online) Time-resolved observations in PCM and optical transmission microscopy in fused silica for various pump-probe delays τ in the plane of
incidence of the laser. The optical excitation sequence is a double pulse with a 3 ps separation and an overall energy of 4 μJ. The arrow (OTM, 800 ps) indicates
that the laser comes from the left side. The dashed line shows an approximation of the propagation envelope in the absence of nonlinear effects. The dash-dotted
line corresponding to the estimated focal plane is drawn in (a) to help visualizing the expansion of the interaction region. (a) The illumination wavelength λ is
400 nm and the full duration at half maximum td of the illumination pulse is 150 fs. (b) λ = 532 nm and td = 1 ns.
B. Phase contrast microscopy
The time-resolved PCM technique is sensitive to the tran-
sient evolution of the real part of the refractive index. In this
experiment, we expect the refractive index variations to arise
as a consequence from different factors such as the existence
of a laser-induced free electron gas, the creation of point de-
fects, the propagation of pressure waves, and thermal diffu-
At the earliest delay times (τ = −1.2 ps and τ = 1.8 ps),
the decrease of the refractive index is presumably due to the
contribution of the free electrons, as predicted by the Drude
model.60Although the results presented here reflect well the
expected variation of the refractive index, the time-resolved
phase contrast microscopy method does not always provide
directly exploitable data. In particular, the influence of the
trast microscopy is, in principle, suitable for pure phase ob-
jects only. When the studied object is not purely transparent,
the contribution of the absorption must be removed to obtain
a correct qualitative estimate of the real part of the refrac-
tive index variation. Otherwise, the refractive index change
is overestimated (in a positive phase contrast microscopy
arrangement) wherever absorption takes place. This phe-
nomenon is particularly visible in Fig. 4(a) for τ = −1.2 ps
and τ = 1.8 ps. In order to remove the absorption contribu-
tion, we apply a procedure in three steps. First, we generate a
binary mask by thresholding the transient absorption picture.
This simple segmentation operation defines two domains: the
absorption region and the background. Second, we compute
a normalized absorption picture. This picture has the same
size as the original picture. An arbitrary value of 1 is assigned
to all the background pixels. The pixels belonging to the ab-
sorption region take the value α = Ipd/Ip0, where Ip0is the
average gray level of the background. As a result, the nor-
malized absorption picture is composed of pixels having val-
ues between zero (extreme case, where no light is transmit-
ted) and one (where there is no light attenuation). In a third
step, the phase contrast picture is divided by the normalized
absorption picture. As an example, a phase contrast picture
corrected from absorption is presented in Fig. 5. Note the ap-
pearance of numerous features in the focal volume compared
to the noncorrected picture, as shown in Fig. 4. This simple
correction procedure assumes that the dephasing experienced
by the probe beam φ fulfills φ ? 2 ∗ π. It is possible to per-
this would necessitate an additional image acquired in nega-
tive phase contrast mode.41
The observation of pure phase objects can be carried out
directly from the noncorrected PCM pictures. For instance, at
τ = 800 ps [Fig. 4(a), right column], a pressure wave is visi-
ble at a distance of ≈4 μm from the optical axis (20 pixels).
This corresponds to a velocity of 5 km/s, close to the the-
oretical value of 5.54 km/s for the speed of sound in fused
silica. Another example is provided by the picture taken at
τ = 10 μs [Fig. 4(b), middle column]. The dark cloud around
the optical axis is a pure phase object with a refractive in-
dex higher than the surrounding material. This is consistent
with the fact that a temperature elevation results in a posi-
tive refractive index change in fused silica.61No noticeable
FIG. 5. (Color online) Time-resolved observation in PCM in fused silica for
a pump-probe delay τ = 1.8 ps. The contribution of the electron gas absorp-
tion was removed. Otherwise, as shown in Fig. 4.
033703-7Mermillod-Blondin et al. Rev. Sci. Instrum. 82, 033703 (2011)
transient change could be detected for delays greater than
τ = 100 μs.
The complementary time-resolved phase contrast and op-
tical transmission microscopy method we have presented is
able to record the evolution of the complex refractive in-
dex in a transparent medium following laser excitation on
different time scales. The real part of the refractive index
is observable in phase contrast microscopy whereas opti-
cal transmission microscopy gives insights into the imagi-
nary part of the refractive index. Our setup can be used with
any transparent medium including gas, liquids, and biological
samples. The complex refractive index evolution of the ir-
radiated target can be tracked from the subpicosecond time
scale, before thermal mechanism takes place, up to the mil-
lisecond range, when the sample reaches its final state.
As an example, for the material fused silica, we were
able to reveal the laser-generated free electron plasma hun-
dreds of femtoseconds after the ultrashort laser pulse exci-
tation and its relaxation, which includes thermal effects and
the release of pressure waves, before forming a permanent
material modification within 10 μs and its subsequent cool-
ing within 100 μs. The smallest visible objects resolvable in
our current setup have a size of 610 and 670 nm, depend-
ing on the illumination source employed. The temporal enve-
lope of the laser excitation can be programmed by the user.
We demonstrated the performance of our method with a
double pulse sequence (3 ps). However, the temporal pro-
file of the excitation is only limited by our pulse shaping
In conclusion, our setup allows to excite any transparent
medium in a customized manner due to the potential offered
by femtosecond pulse shaping, and to directly observe the as-
sociated fast (subpicosecond) and slow (microsecond) energy
relaxation channels with submicrometer resolution.
The authors would like to thank M. L. Boyle for his con-
tribution to the early developments of the setup, F. Noack for
his support with laser maintenance, and M. Teichmann for his
help with software development under Python. This work has
been partially funded by the DFG projects RO2074/5-3, 5-1
of the l’Agence Nationale de la Recherche and the LASUR—
Réseau des Technologies Femtoseconde.
1V. V. Temnov, K. Sokolowski-Tinten, P. Zhou, A. El-Khamhawy, and D.
von der Linde, Phys. Rev. Lett. 97, 237403 (2006).
2P. P. Rajeev, M. Gertsvolf, P. B. Corkum, and D. M. Rayner, Phys. Rev.
Lett. 102, 083001 (2009).
3D.Grojo, M.Gertsvolf, S. Lei, T. Barillot,D. M.Rayner,and P.B.Corkum,
Phys. Rev. B 81, 212301 (2010).
4B. Rethfeld, O. Brenk, N. Medvedev, H. Krutsch, and D. Hoffmann, Appl.
Phys. A: Mater. Sci. Process. 101, 19 (2010).
5H. Nishikawa, T. Shiroyama, R. Nakamura, Y. Ohki, K. Nagasawa, and Y.
Hama, Phys. Rev. B 45, 586 (1992).
6P. N. Saeta and B. I. Greene, Phys. Rev. Lett. 70, 3588 (1993).
7G. Petite, P. Daguzan, S. Guizard, and P. Martin, Appl. Surf. Sci. 109/110,
8H. Sun, S. Juodkazis, M. Watanabe, S. Matsuo, H. Misawa, and J. Nishii,
J. Phys. Chem. B 104, 3450 (2000).
9Q. Sun, H. Jiang, Y. Liu, Z. Wu, H. Yang, and Q. Gong, Opt. Lett. 30, 320
10V. V. Kononenko, V. P. Pashinin, M. S. Komlenok, and V. I. Konov, Laser
Phys. 19, 1294 (2009).
11J. W.Chan,T. Huser, S. Risbud, and D. M.Krol,Opt. Lett. 26,1726 (2001).
12K. M. Davis, K. Miura, N. Sugimoto, and K. Hirao, Opt. Lett. 21, 1729
13S. Nolte, M. Will, J. Burghoff, and A. Tuennermann, Appl. Phys. A: Mater.
Sci. Process. 77, 109 (2003).
14L. Shah, A. Arai, S. Eaton, and P. Herman, Opt. Express 13, 1999 (2005).
15C. Mauclair, A. Mermillod-Blondin, N. Huot, E. Audouard, and R. Stoian,
Opt. Express 16, 5481 (2008).
16M. Ams, G. D. Marshall, P. Dekker, J. A. Piper, and M. J. Withford, Laser
Photonics Rev. 3, 535 (2009).
17Y. Li, K. Itoh, W. Watanabe, K. Yamada, D. Kuroda, J. Nishii, and Y. Jiang,
Opt. Lett. 26, 1912 (2001).
18F. He, Y. Cheng, Z. Xu, Y. Liao, J. Xu, H. Sun, C. Wang, Z. Zhou, K.
Sugioka, K. Midorikawa, Y. Xu, and X. Chen, Opt. Lett. 35, 282 (2010).
19L. Sudrie, M. Franco, B. Prade, and A. Mysyrowicz, Opt. Commun. 171,
20K. Zhou, Z. Guo, W. Ding, and S. Liu, Opt. Express 18, 13640 (2010).
21K. Yamada, W. Watanabe, Y. Li, K. Itoh, and J. Nishii, Opt. Lett. 29, 1846
22D. Grojo, M. Gertsvolf, H. Jean-Ruel, S. Lei, L. Ramunno, D. M. Rayner,
and P. B. Corkum, Appl. Phys. Lett. 93, 243118 (2008).
23F. He, Y. Cheng, L. Qiao, C. Wang, Z. Xu, K. Sugioka, K. Midorikawa, and
J. Wu, Appl. Phys. Lett. 96, 041108 (2010).
24C.-H. Lin, L. Jiang, H. Xiao, Y.-H. Chai, S.-J. Chen, and H.-L. Tsai, Opt.
Lett. 34, 2408 (2009).
25G. D. Valle, R. Osellame, and P. Laporta, J. Opt. A: Pure Appl. Opt. 11,
26R. M. Vazquez, R. Osellame, M. Cretich, M. Chiari, C. Dongre, H. Hoek-
stra, M. Pollnau, H. van den Vlekkert, R. Ramponi, and G. Cerullo, Anal.
Bioanal. Chem. 393, 1209 (2009).
27K. Sugioka, Y. Hanada, and K. Midorikawa, Laser Photonics Rev. 4, 386
28L. Englert, M. Wollenhaupt, L. Haag, C. Sarpe-Tudoran, B. Rethfeld, and
T. Baumert, Appl. Phys. A: Mater. Sci. Process. 92, 749 (2008).
29T. Nagata, M. Kamata, and M. Obara, Appl. Phys. Lett. 86, 251103 (2005).
30A. Mermillod-Blondin, C. Mauclair, A. Rosenfeld, J. Bonse, I. V. Hertel,
E. Audouard, and R. Stoian, Appl. Phys. Lett. 93, 021921 (2008).
31A. Mermillod-Blondin, I. M. Burakov, Y. P. Meshcheryakov, N. M. Bul-
gakova, E. Audouard, A. Rosenfeld, A. Husakou, I. V. Hertel, and R.
Stoian, Phys. Rev. B 77, 104205 (2008).
32A. Horn, Ultra-Fast Material Metrology (Wiley-VCH, NewYork, 2009).
33S. Garnov, V. Konov, A. Malyutin, O. Tsarkova, I. Yatskovsky, and F.
Dausinger, Laser Phys. 13, 386 (2003).
34E. Abraham, K. Minoshima, and H. Matsumoto, Opt. Commun. 176, 441
35C. Schaffer, N. Nishimura, E. Glezer, A. Kim, and E. Mazur, Opt. Express
10, 196 (2002).
36D. G. Papazoglou, I. Zergioti, and S. Tzortzakis, Opt. Lett. 32, 2055
37W. Gawelda, D. Puerto, J. Siegel, A. Ferrer, A. R. de la Cruz, H. Fernandez,
and J. Solis, Appl. Phys. Lett. 93, 121109 (2008).
38A. Horn, D. Wortmann, A. Brand, and I. Mingareev, Appl. Phys. A: Mater.
Sci. Process (to be published).
39D. G. Papazoglou and S. Tzortzakis, Appl. Phys. Lett. 93, 041120 (2008).
40A. Horn, I. Mingareev, A. Werth, M. Kachel, and U. Brenk, Appl. Phys. A:
Mater. Sci. Process. 93, 165 (2008).
41T. Noda and S. Kawata, J. Opt. Soc. Am. A 9, 924 (1992).
42H. M. Presby and D. Finkelstein, Rev. Sci. Instrum. 38, 1563 (1967).
43H. Weisen, Infrared Phys. 25, 543 (1985).
44A. Mermillod-Blondin, J. Bonse, A. Rosenfeld, I. V. Hertel, Y. P. Meshch-
eryakov, N. M. Bulgakova, E. Audouard, and R. Stoian, Appl. Phys. Lett.
94, 041911 (2009).
45A. M. Weiner, Rev. Sci. Instrum. 71, 1929 (2000).
46C. Hnatovsky, R. S. Taylor, E. Simova, V. R. Bhardwaj, D. M. Rayner, and
P. B. Corkum, J. Appl. Phys. 98, 013517 (2005).
47R. Hard, R. Zeh, and R. Allen, J. Cell Sci. 23, 335 (1977).
033703-8Mermillod-Blondin et al.Rev. Sci. Instrum. 82, 033703 (2011) Download full-text
48H. Ambar, Y. Aoki, N. Takai, and T. Asakura, Appl. Phys. B: Lasers Opt.
38, 71 (1985).
49A. Millard, D. Fittinghoff, J. Squier, M. Muller, and A. Gaeta, J. Microsc.
193, 179 (1999).
50I. M. Burakov, N. M. Bulgakova, R. Stoian, A. Mermillod-Blondin, E.
Audouard, A. Rosenfeld, A. Husakou, and I. V. Hertel, J. Appl. Phys. 101,
51P. Audebert, P. Daguzan, A. Dos Santos, J. C. Gauthier, J. P. Geindre, S.
Guizard, G. Hamoniaux, K. Krastev, P. Martin, G. Petite, and A. Antonetti,
Phys. Rev. Lett. 73, 1990 (1994).
52Z. Bor, Opt. Lett. 14, 119 (1989).
53M. Kempe, U. Stamm, and B. Wilhelmi, Opt. Commun. 89, 119 (1992).
54M. Kempe and W. Rudolph, Opt. Lett. 18, 137 (1993).
55C. J. R. Sheppard and T. Wilson, Appl. Opt. 18, 3764 (1979).
56M. Estribeau and P. Magnan, Proc. SPIE 5251, 243 (2004).
57B. D. Price, C. J. Esbrand, A. Olivo, A. P. Gibson, J. C. Hebden, R. D.
Speller, and G. J. Royle, Rev. Sci. Instrum. 79, 113103 (2008).
58I. H. Chowdhury, X. Xu, and A. M. Weiner, Appl. Phys. Lett. 86, 151110
59D. Wortmann, M. Ramme, and J. Gottmann, Opt. Express 15, 10149
60K. Sokolowski-Tinten, J. Bialkowski, and D. von der Linde, Phys. Rev.
B 51, 14186 (1995).
61T. Toyoda and M. Yabe, J. Phys. D: Appl. Phys. 16, L97 (1983).