Dynamic position shifts of X-ray emission from a water film induced by a pair of time-delayed femtosecond laser pulses

Femtosecond double-pulsed laser excitation of a water film in air showed enhancements of X-ray intensity as compared with single pulse irradiation. The position of the highest yield of X-rays strongly depends on temporal separation between the pre-pulse and the mainpulse (energy ratios where ∼ 1 : 10). The strongest X-ray emission was observed at 10-15 ns delay of the main-pulse. Nanoscale roughening of water surface can account for the observation. c © 2017 Optical Society of America OCIS codes: (140.7090) Ultrafast lasers; (160.4236) Nanomaterials; (240.6680) Surface plasmons; (320.2250) Fem-


Introduction
Intense femtosecond (fs-) laser interaction with condensed matter involves linear and nonlinear absorption [1], which results in series of processes including plasma formation [2], ablation [3], and X-ray pulse emission [4]. The characteristics of such X-ray pulses have been considered as an important research topic for applications of time-resolved X-ray diffraction [5] and X-ray absorption fine structure (XAFS) measurements [6]. When the X-ray pulses are applied to the above measurements, the X-ray intensity is one of the most significant parameters. It has become an important subject to enhance X-ray intensity with a constant amount of fs-laser photon flux. The enhancement of X-ray intensity is attributed to an effective increase of the absorption efficiency and an enhanced generation of high-energy electrons from the highly ionized plasma [7]. Solid targets such as metals and transparent glasses [8,9] with flat surfaces have been used for X-ray generation under intense fs-laser irradiation. Such X-ray generation is considered to be due to surface plasmon resonance effects [10]. The enhancements of X-ray intensity from solid targets by changing the structure of the layered surfaces have been reported [11,  Fig. 1. Setup used for X-ray detection using two fs-laser pulses with controlled time separation, ∆t; HWP is the half-wave plate, PBS is the polarized beam splitter, RR is the retro-reflector, ODL is the optical delay line, OAPM is the off-axis parabolic mirror, S is the sample solution, GM is the Geiger Müller counter, APS is the automatic positioning system, DS is the displacement sensor. 12]. In contrast to the solid targets, liquids/solutions are promising since they can be reused continuously by a circulative pump. Moreover, the X-ray emission spectra of aqueous solution can be easily controlled by changing the concentrations or species of solutes which include electron and metallic suspensions [13,14]. These features are advantageous for time-resolved XAFS measurements since long-time count accumulation is indispensable and the characteristics of solutes can be used as wavelength standards [15].
X-ray intensity enhancements under the double-pulsed excitation to aqueous solution surfaces, with the pre-pulse and the main pulse irradiations, have been reported elsewhere [16]. The pre-pulse irradiation with relatively low intensity results in various time-dependent phenomena such as plasma formation/decay in the picosecond range or ablation in the nanosecond range. Such various phenomena change the solution surface conditions, then we expect that the optimum position for X-ray emission of the solution surface to the laser focus changes as the delay time increases. As practical developments of bench-top X-ray sources, this knowledge is indispensable.
In this study, we perform the axial-scan of the water film exposure for the X-ray intensity measurements using the automatic positioning system [17] under the double-pulsed excitation conditions. Controlled time separation from tens-of-fs to tens-of-ns was investigated for the fixed ratio of pre-pulse and the main pulse.

Samples and procedures
The experimental setup is shown in Fig. 1. Transform-limited fs-laser pulses (t p = 40 fs, λ = 800 nm, 1 kHz, horizontally-polarized, Mantis, Legend Elite, HE USP, Coherent, Inc.) were separated and combined collinearly as double pulses with half-wave plates (65-906, Edmund Optics) and polarization beam splitters (47-048, Edmund Optics). Vertically and horizontally polarized pulses are defined as a pre-pulse and a main-pulse, respectively. The vertically-polarized pre-pulse intensity is well low and its irradiation does not induce X-ray emission by itself therefore the obtained X-ray emission under the double pulses irradiation conditions is always a single X-ray pulse. For the delay time of the main pulse relative to the pre-pulse, there are three time zones, I: 0 -5 ns, II: 5 -10 ns and III: 10 -15 ns [ Fig. 1], set in the optical path of the pre-pulse. The optical delay for each time zone is obtained with an 80 cm-long variable  (c) Slices of the X-ray intensity spatial profiles for the time delays from -48.4 ps to 1 ns. (d) A high z-axis resolution plot of the X-ray intensity at ∆t = 1.6 ps at which the local maximum was observed in the picosecond range. The double peak is deconvoluted into two Gaussian contributions. X-ray emission without the pre-pulse was more than twice lower in the peak amplitude. delay line (SGSP 46-800, Sigma Koki). Change of the time zones from I to II or III can be done with two independent fixed retro-reflectors. The optical path of the main-pulse has the same optical length as the longest optical path of the pre-pulse. After combination, the pulses were tightly focused onto a solution film with an off-axis parabolic mirror (the effective focus length of f = 50.8 mm, the reflection angle of 90 degrees, and the numerical aperture N A = 0.25, 47-097, Edmund Optics) in air. At these focusing conditions, the theoretical diameter of the focal spot is 2w 0 = 1.22λ/N A = 3.9 µm and the axial extent along the propagation estimated as two Rayleigh lengths 2z R = 2n λ N A 2 = 25.5 µm where n = 1 is refractive index of air. Noteworthy, the focal region is longer than spatial extent of the pulse ct p = 12 µm; c is the speed of light.
The solution film was formed by a solution flow system [18] and its position was controlled by 3D-automatic (KS701-20LMS, Suruga Seiki) and rotation stages (KS401-60, Suruga Seiki) [17]. Two orthogonal plastic pipette nozzles (inner diameter of 0.46 mm), which are connected to a circulation pump (PMD-211, Sanso), make two colliding solution flows (flow rate ∼180 mL/min). Such flow collisions result in the formation of a thin solution film. The solution film is oriented perpendicularly to the solution flows from the nozzles and moves at ∼ 5.8 m/s or nm/ns rate. The estimated thickness of the solution film is ∼ 20 µm. The solution film was set on the automatic Fig. 3. (a) Dependence of the solution surface position of the ∆t time-delayed main-pulse for the maximum intensity of X-rays from a 20-µm-thick water film (thickness of film is marked by shaded area). Error bars are plotted as standard deviation for 20 pulse measurement. Prepulse energy was 80 µJ, the main-pulse was 700 µJ. Inset schematically shows contributions of the main three peak positions at which X-ray emission is maximized. (b) Schematics of two-pulse irradiation of a solution film for the three typical positions of the solution surface when the maximum of X-rays is generated; focusing position is at the main pulse. The width of film and axial extent of focal region for low intensity focus are comparable. (c) Intensity profiles for N A = 0.25 focusing in air; length of pulse ct p 12 µm. Intensity is presented on logarithmic scale, I = 0.5 at z R is marked by a contour line (corresponds to the depth of focus). For comparison, the width of water film, h, is marked; in experiments, the irradiation angle was θ = 60 • and the effective width of water film was twice larger h/ cos θ.
positioning system and the incident angle of the laser pulse was at 60 degrees to the normal of the solution film for the optimization of X-ray emission by change the position of solution film surface to the laser focus. At this close to Brewster angle position, the reflectivity of water surface for the p-polarized main-pulse is minimum R p → 0 and favors coupling of light energy into pre-excited water film by the preceding pre-pulse. The operation detail of the automatic positioning system with Labview code was reported elsewhere [17].
In this study, for higher precision of the solution film position, a displacement sensor (precision of 0.2 µm, LK-G80, KEYENCE) was set behind the automatic positioning stage and its signal which shows an absolute value of distance from the stage was used as feedback parameter of automatic positioning system. A Geiger counter (SS315, Southern Scientific) was used for X-ray measurements. The major gas component of the Geiger counter is helium (>95%, ∼0.5 bar) with some halogens as quenching agent. The input window material is mica (1.6 mg/cm 2 ) and its thickness is ∼ 6 µm. The distance between the laser focus (the X-ray point source) and the Geiger counter was kept constant at 15 cm in the entire experiment. The solid angle of the detector was 1.39 × 10 −4 steradians. All the experiments were carried out in air under atmospheric pressure (1 atm) at room temperature (296K). Therefore, it is certain that the Geiger counter detects only X-ray, not α nor β-rays. Figure 2 shows X-ray intensity from the double pulse-irradiated water film at different time delays of the main-pulse. At shorter time scale from 0.6 to 1.6 ps the position of the solution surface had to be shifted up-stream (towards incoming fs-pulse) for the most intense X-ray   Fig. 2(a)]. The position z = 0 is the Z position at the highest X-ray intensity when only the main pulse was irradiated. On a longer time scale up to several nanoseconds, a tendency was observed of down-stream shift of intense X-ray emission [ Fig. 2(b)]. Figure 2(c) represents slices of the X-ray intensity spatial profiles for the time delays from -48.4 ps to 1 ns. Figure 2(d) represents the X-ray intensity profile at ∆t = 1.6 ps, which is with two different Gaussian fitting curves. Figure 3 shows the location of the solution surface position along the axis of the laser propagation should be set for the maximum of X-ray emission depending on the time separation between pre-pulse and the main-pulse. The maximum of X-ray emission was achieved for a ∆t 1.6 ps delay of the main-pulse. This enhancement at the up-stream position (towards the incoming laser pulse) was short lived and another maximum was developed after the location of nanosecond time scale. Another location of intense X-rays was located down-stream and evolved on a ns-time frame.

X-rays out of water irradiation
The exact geometrical focusing of a Gaussian beam corresponding to the used N A = 0.25 off-axis parabolic mirror is shown in Fig. 3(c), where the outline of the outer intensity contour Fig. 6. FDTD modeling of 800 nm wavelength light (plane wave) traversing water film with several generic surface ripple patterns expected to develop under pulsed excitation. The patterns are: the period Λ = λ on the front surface and depth profile ∼ λ/2, the back-side period λ/n, random pattern of similar size protrusions caused by light filamentation pattern, and initial stages of departing droplets. Insets show schematically the pattern simulated. Thickness of film is taken 10 µm to reduce simulation time; Intensit y ∝ E 2 .
is following the waist defining profile w(z) = w 0 1 + (z/z r ) 2 with z R = πw 2 0 /λ. The axial extent of the ultra-short laser pulse, ct p , is almost twice shorter as the depth-of-focus, 2z R . This defines favorable conditions to reach higher intensity on the target surface before onset of strong air ionization. The molecular number density of air is ∼ 2.5 × 10 19 cm −3 which cause an intensity clamping during filamentation in air breakdown by ultra-short laser pulses [2]; water has 1.34 × 10 3 larger molecular number density. It was shown that for a higher-N A the intensity of the filament clamping in air is increasing and for N A = 0.12 is I = (2 − 3) × 10 14 W/cm 3 as determined from side imaging and shadowgraphy of the focal region [2], where the plasma density of ∼ 3 × 10 19 cm 3 is reached (a full ionization of air). Larger clamping intensity can be predicted for N A = 0.25 used in our study, however, due to a very short pulse duration the localization of filamentation and air breakdown should be close to focus. Estimation of irradiance from geometrical focus [ Fig. 3(c)] gives I p = 6.6 × 10 14 W/cm 2 for a E p = 100 µJ pre-pulse (for the N A = 0.12 used in [2], Liu, et al. with clamping at I = (2 − 3) × 10 14 W/cm 2 the theoretical average irradiance would reach I p = 1.5 × 10 14 W/cm 2 and the peak intensity is twice larger). Evidently, in the case of a pre-pulse excitation of water with ∼ 10 3 larger molecular number, it is expected that larger plasma densities can be created. Figure 4 shows optimization of X-ray intensity by axial positioning of the solution surface at temporal separation between the pulses ∆t at 5 and 15 ns. The Gaussian fits show strong localization of X-ray intensity peak at the down-stream with 50 ± 10 µm precision. Up to 75-90% X-ray intensity peak enhancement was observed at 15 ns delay for the focus position on the film (z = 0). Approximately twice more intense X-ray emission was observed during optimized main-pulse placement after 10-15 ns delay as compared with the first spike of X-ray emission during 1-2 ps window [ Fig. 5]. To visualize light intensity distribution on a micro-film of water film under excitations which corresponds to plasma formation at close-to-critical densities several patterns on water film were simulated [ Fig. 6]. The modeling results have a qualitative character since only take into account a linear light-matter interaction, however, is informative to reveal possible field enhancement effects due to nanotexture/roughness of surfaces. A thin skin-layer of plasma is formed on the surface of dielectric [19] (e.g., water) which can support surface wave at the plasma-air interface with the period close to the wavelength of light in air, Λ = λ, while on the inner interface plasmawater surface excitation has period Λ = λ/n/2, where n = 1.33 is the refractive index of water. This plasma wave model is consistent with experimental observation of sub-wavelength ripples' formation on surface and inside dielectrics [19,20]. These two cases qualitatively represented with larger and smaller period ripples on the two different sides of film. The random pattern of water nano-protrusions and droplet formation are also shown (noteworthy, the droplets are formed on much longer time scales of tens-to-hundreds nanoseconds). Surface capillary wave is a candidate for explanation of surface roughening, film, and droplet formation. The size and time scale of those evens are important to estimate. The period of capillary Λ = 4 σh/ρ √ 2πt w , where σ is surface tension, ρ is the mass density, τ w is the lifetime of the wave (e.g., lifetime of the molten phase on solid materials [21]), h is the height of the wave (molten phase).

Nanosecond time delays
The finite-difference time-domain (FDTD) simulations (Lumerical) show light field intensity enhancement up to 3 times for several typical expected surface morphologies of pre-excited water films [ Fig. 6]. Even larger light localization was observed in the case of single droplets departing from film, however, such events would be evolving at µs time scale beyond the tested time window. Strong Talbot pattern of refocusing occurs inside the water film and can be instrumental to reach more efficient excitation and heating of water.

Discussion
As a development of bench-top X-ray sources, emission spectra, pulse width, and X-ray photon numbers are basic and indispensable characteristics. Hatanaka, et al. reported typical X-ray emission spectra from various aqueous solution [13-15, 22, 23], including distilled water [13,15] as in the current case, which shows structureless spectra in the energy region of 3 -20 keV (0.4 -0.06 nm) indicating bremsstrahlung without characteristic X-ray lines in this energy region because only the light elements such as hydrogen and oxygen are under laser excitation. Even under the double-pulsed excitation conditions with aqueous solutions, X-ray emission spectra have been reported [16], which show the X-ray intensity enhancements and the intensity increase in the higher photon energy region. As for the pulse-width measurements, there are limited numbers of reports on synchrotron radiation, X-ray free-electron laser, and femtosecond