New probing techniques of radiative shocks

Transcription

1 New probing techniques of radiative shocks Chantal Stehlé a, Michaela Kozlová b, Jean Larour c, Jaroslav Nejdl b,d, Norbert Champion a, Patrice Barroso e, Francisco Suzuki-Vidal f, Ouali Acef g, Pierre-Alexandre Delattre a, Jan Dostál b, Miroslav Krus b,d, Jean-Pierre Chièze h a LERMA, UMR 8112, Observatoire de Paris, UPMC, CNRS, 5 Place J. Janssen, Meudon, France b Institute of Physics of ASCR, Na Slovance 2, Prague 8, Czech Republic c Laboratoire de Physique des Plasmas, UMR 7648, Ecole Polytechnique, UPMC, CNRS, Palaiseau, France d Dept. of Physical Electronics, Faculty of Nuclear Sciences and Physical Engineering, Czech Technical University in Prague, Brehova 7, Prague 1, Czech Republic e GEPI, UMR 8111, Observatoire de Paris, Paris Diderot, CNRS, 5, place Jules Janssen, Meudon, France f The Blackett Laboratory, Imperial College, Prince Consort Road, London SW7 2AZ, UK g SYRTE, UMR 8630, Observatoire de Paris, UMPC, CNRS, 61, Av. de l Observatoire, Paris, France h IRFU/Service d Astrophysique, CEA-Saclay, Gif sur Yvette, France Abstract Radiative shock waves propagating in xenon at a low pressure have been produced using 60 Joules of iodine laser ( λ=1.315 µm) at PALS centre. The shocks have been probed by XUV imaging using a Zn X-ray laser (λ= 21 nm) generated with a 20-ns delay after the shock creating pulse. Auxiliary high-speed silicon diodes allowed performing space- and time-resolved measurement of plasma self-emission in the visible and XUV. The results show the generation of a shock wave propagating at 60 km/s preceded by a radiative precursor. This demonstrates the feasibility of radiative shock generation using high power infrared lasers and the use of XRL backlighting as a suitable diagnostic for shock imaging. Preprint submitted to Optics Communications October 24, 2011

2 Keywords: X-ray lasers - Shock waves - Plasmas 1. Introduction The study of strong radiative shock waves in gases has been stimulated within the last 10 years by the availability of kj laser installations, which allow to launch shock waves with velocities larger than 60 km/s and then to study the coupling between radiation and hydrodynamics in simple geometries [1, 2, 3, 4]. This topic is very important for understanding the physics of astrophysical shock waves in processes occurring during stellar formation. For instance, shock waves are produced during the gravitational collapse from a dense molecular cloud. Later on, in young stars, an accretion shock is formed when matter falls from the circumstellar disk into the stellar photosphere. In addition, shocks are also observed at the head of high velocity protostellar jets. A radiative shock is a very strong shock wave, in which the shock front is heated at a very high temperature and thus presents a strong self-emission. Depending on the opacity, the photons emitted from this zone can be reabsorbed by the cold gas in which the shock propagates. This generates a noticeable ionisation wave, also called a radiative precursor [5, 6, 7]. The complexity of the physical description of these flows, which results from radiation effects [8], is enhanced by the non ideal effects in the gas (e.g. ionisation, equation of state...) [9] and by the geometrical effects even in cylindrical or parallelepipedic shock tubes [10, 11]. Most of the experiments have focused on the study of the radiative precursor, e.g. by using optical laser probing (shadowgraphy and interferometry) 2

3 [1, 4, 11]. Hard X-ray back-lighting has been used at Rochester [2] to image the front shock. From these types of studies, several points can be highlighted: (i) radiation cooling strongly affects the geometry and temperature of the precursor, (ii) the interactions between the shock and the walls may be appreciable [12], (iii) the shock wave may be observed at late times ( 40 ns) in a regime where the velocities of the ionisation and shock fronts are comparable [4]. However, none of these experiments were able to probe the whole shock structure, i.e. the precursor, the shock front and the postshock. Moreover, the study of the radiative signatures of these shocks remains very challenging [4]. The experiments presented here aim at testing for the first time the feasibility of two new diagnostics: instantaneous imaging of the whole shock structure with an auxiliary X-ray laser (XRL) at 21.2 nm and a photometric, spatially resolved, study of the plasma self-emission which will also be useful to follow the shock chronometry. 2. Experimental setup The experiments were conducted at the Prague Asterix Laser System (PALS) [13]. The core of the experimental setup (Figure 1) consists of two beams from the iodine laser (λ = 1.315µm), with a pulse duration of 0.3 ns. The first beam (hereafter called AUX beam, 60 J) drives the shock-wave inside a target filled with xenon at low pressure ( 0.3 bar, i.e g/cm 3 ). The second beam (hereafter called MAIN beam, 500 J) generates a laser beam at 21.2 nm ( λ = 21.2 nm, hereafter called XRL beam), with a pulse duration of 0.15 ns, on a Zn slab. In terms of diagnostics, the XRL beam is used as a backlighter of the shock wave with a controlled delay of 20 3

4 Figure 1: Schematic experimental setup showing the two main InfraRed laser beams (MAIN and AUX) which drive respectively the XRL probe beam and a radiative shock inside a Xe-filled gas tube. The XUV imaging and diode diagnostics are also shown in the figure. ns after AUX beam allowing an almost instantaneous imaging of the shock (For the shock velocity of 60 km/s, the blurring effect is 9 µm which is of the order of magnitude of the thickness of the shock front). The second main diagnostic consists of a set of high-speed silicon diodes recording the plasma self-emission at a given position of the shock tube. Finally, an X-ray pinhole camera monitors the energy distribution of the AUX laser beam on the target. The targets used for the shock generation consist of a miniaturized shock tube of section 0.4 x 0.4 mm 2, and length 6 mm, closed by a foil, 4

5 made of a gilt polystyrene (CH and Au thickness respectively 10 µm and 500nm) (Figure 2). The AUX beam is focused on this foil by a f =100 cm lens, yielding an irradiance of 1-2 x W/cm 2 on a spot of diameter 0.4 mm. The polystyrene ablation by the AUX laser propels by rocket effect this small section of the Au-CH foil, acting as a piston, inside the shock tube. At this irradiance, 1D simulations and earlier experiments [1, 4, 11] foresee a shock velocity of km/s. The 500-nm gold layer deposited on the back side of the polystyrene foil, is used as a radiative shield against the x-ray emission generated by the ablation of the plastic. For the XUV back-lighting, as for plasma self emission diagnostics, we used 150 nm thick Si3N4 windows with a length of 3 mm and width 0.4 mm. A post-mortem inspection of the targets allowed to control the centering of the AUX beam on the target. The targets were filled outside the chamber with xenon at a fraction of bar and then placed inside the chamber. The precise design of the targets and target holder resulted in a reproducibility better than 100 µm in the target positioning before the shot. Using 1D simulations [21], it is possible to infer the typical characteristics of the shock wave. For an initial gas pressure of 0.3 bar, and with the previous irradiance conditions on the CH-Au piston, a mean average shock wave velocity of 60km/s is expected, from 3 ns to 30 ns after the time of AUX arrival on the target. This is in agreement with previous experiments performed at PALS in the same range of laser irradiance but with the laser operating at a shorter wavelength (438 nm instead of µm in the present work) [4, 11]. The XRL beam at 21.2 nm is generated in a spherical vacuum chamber from a plasma column with a length of 30 mm by the sequence of a 5

6 Figure 2: a) Image of the target with the gilt polystyrene piston, an Si3N4 window (0.4 x 4 mm 2 ) on its Si frame, the Al structure and the two filling capillaries ; (b) scheme of the target. The AUX laser, represented by a triangle, comes from the top and is focused on the piston (yellow), whereas an XRL laser passes through Si3N4 windows and the channel filled with Xe. The opaque Si frame is shown in green and the Al structure in grey. The Xe gas is contained within a 0.4 x 0.4 x 6 mm 2 channel in the target center, delimited by the Al structure on two sides and by Si3N4 windows on the two perpendicular sides. 6

7 Figure 3: Relative positions of the PERP and OBL diodes relative to the target and XRL beam. low intensity infrared prepulse (loosely focused 700 µm) and the strong MAIN pulse (tightly focus 150 µm) on a Zn slab. By using the half-cavity configuration, the XRL is able to deliver up to 10-mJ in a 150-ps pulse with a divergence of about 4 x 6 mrad 2 [14, 15]. The XRL beam propagates trough the shock - target, and is reflected by a spherical Mo/Si multilayer mirror (f= 300 mm); it passes through a pinhole located in the focal plane of the mirror and reaches a cooled CCD camera located at the position of the image of the shock tube through this mirror. A pixel on the CCD (13.5 µm) corresponds to 0.18 µm on the target (magnication of 8.2). As explained by Kozlová et al. [16], the iron pinhole (500 µm diameter) acts as a spatial filter to improve the contrast on the CCD of the XRL versus the strong plasma selfemission. Moreover, the intense visible stray light of the plasma is filtered by a 0.4 µm thick self-supported Al foil. Two high-speed 1-mm 2 silicon diodes were located side-on to record the time dependent self-emission of the shock wave. Both diodes (named hereafter PERP and OBL) were positioned 7

8 Figure 4: AXUV-HS5 IRD diodes detectivity (in A/W) in the ev range (black dashed line, data taken from ). The transmission of the 150 nm thick Si3N4 window is reported in black (bold line). The position of the XRL laser energy is reported by a red arrow. Typical blackbody spectra (40 B(hν)/σT 4, no unit) are also plotted for two temperatures: 10 ev (in blue solid line) and 20 ev ( in solid dashed line). These temperatures are representative of the precursor and shock front respectively. 8

9 at about 55 mm from the lateral window of the target with different angles of sight (see Figure 3). The diodes were positioned at an angle of 30 degrees with respect to the horizontal plane in order to allow the XRL beam to pass through the windows. One of the diodes (PERP) is in the plane perpendicular to the longitudinal axis of the shock tube, and the second one (OBL) makes an angle of 110 degrees (Figure 3). The spatial resolution is ensured by a 50-µm thick tungsten slit (respectively x 0.4 mm 2 and 0.3 x 0.4 mm 2 for PERP and OBL) located at 9.5 mm and 17 mm from the window respectively.the slit orientation is perpendicular to the longitudinal axis of the channel in which the shock propagates. The detector response and the transmission of the 150-nm Si3N4 window are presented in Figure 4. The curves of the blackbody emissivity at two temperatures (10 and 20 ev) of the shock structure are reported to indicate the spectral range expected in the experiment. The rise-time of the diodes is given to be 0.7 ns, which is compatible with high quality coaxial cables, 3-GHz bandwidth bias tees and attenuators. A careful time compensation test of the detection chains gave a 0.25 ns dispersion between the corresponding two channels on a 1-GHz 10-GS/s oscilloscope. The time response of the detection chain is adapted to the study of a shock traveling from 20 to 60km/s. Assuming a shock velocity of 20 to 60 km/s, the time window associated to the PERP diode will be 10 to 4 ns, whereas it will be 22 to 7 ns for OBL. 3. XUV radiography of the shock Monochromatic XRL radiography offers several advantages compared to other techniques: the number of photons, a parallel beam, and the possibility 9

10 to probe dense matter up to a fraction of the electron critical density (N cr [cm 3 ]= λ 2 [nm], i.e cm 3 at λ =21 nm). In our case, this technique will provide an image of the whole shock structure. It will allow to probe the dense part of the shock, but also the heated low-density precursor, where absorption occurs through the rich spectral signatures of the moderately charged Xe ions (Xe VII - XII) in the XUV [17, 18, 20, 19]. The expected structure of the plasma at 20 ns, during the 0.15 ns XRL pulse is reported in Figure 5. These results were obtained using a 1D radiative hydrodynamical simulation [21], showing a maximum electron density up to cm 3 in xenon and cm 3 in gold. An estimate of the compression and the pressure in the shocked xenon results in 18 and bar respectively. The thickness of the shocked xenon layer is approximately equal to 10 µm, which, taking into account the pixel size of 13.5 µm and the magnification of the imaging setup, corresponds to 6 pixels on the CCD. Figure 6 shows two consecutive measurements, with and without the presence of a shock wave respectively. The initial position of the piston is localized close to the right border of the image and the AUX laser (66 J) is coming from the right. The relatively low number of counts ( after subtraction of the image background) is a consequence of four factors: (i) the geometric losses (only a fraction of the XRL beam passing through the target), (ii) the strong absorption of XRL laser light by the two Si3N4 windows closing the target (transmission of 0.15%), (iii) losses at the Mo-Si mirror ( 30% at 21 nm ) and (iv) through the Al filter (50%) shielding the CCD (absorptions taken from [22]). The unexpected static speckles visible on the records are attributed to an insufficient surface quality of the commercial substrate of 10

11 Figure 5: Snapshot of the plasma at 20 ns from 1D simulations [21]. The initial position of the piston is at x = 0.5 cm and the laser is coming from the right. The position of the shock front in xenon is located at 0.37 cm. At this position, the electron density (black curve) peaks at cm 3, the electron temperature (blue) at 26 ev and the mean ion stage (red) at 13. The Xe-Au interface is at cm ( the electron density reaches cm 3 in this gold layer) and the CH-Au interface is located at cm. 11

12 Figure 6: Experimental results with and without the presence of a shock (top and bottom respectively), as seen through the Si3N4 windows. The delay between the XRL and the AUX beam was set to 20 ns. The instrumental background was subtracted. Dimensions are 2048 x 400 pixels. The initial position of the piston is located on the right vertical edge. The shock (top image) is coming from the right. The dimensions of the target (in mm) are presented in Figure 7. The black regions correspond to the frame supporting the Si3N4 membrane. The frame covers a small section ( 0.6 mm) of the tube close to the piston (see Figure 2). the spherical imaging MoSi XUV mirror. It is important to note that, beside the external upper and lower parts of the images, these two images cover the same section of the MoSi mirror. The differences in the visible area come from slight misalignments of the Si3N4 windows during the gluing process. The noisy patterns present a negligible shot-to-shot displacement. Thus, despite these difficulties, this allowed to perform the division of the two images. The resulting synthetic image, presented in Figure 7, is thus the transmission of the plasma at 21.2 nm and at 20 ns after the arrival of AUX on the target. The shock wave is located where the transmission is at minimum, i.e. at 1.5 mm from the initial position (around pixel 1100). This is slightly larger by 0.15 mm than the position expected from the simulation, which remains reasonable. Due to the small scale structures of the image, the 12

13 Figure 7: Top image: ratio of the two images shown in Figure 6, smoothed over 20 pixels. The two initial images have been normalized, in such a way that the transmission in the left part of the tube (unshocked cold gas, close to pixel 1) is fixed to unity. Bottom image: profile variation of the image along horizontal distance on the CCD (in pixels), with and without smoothening (red and black curves respectively). The position along the target (in mm) is also given, with x = 0 mm corresponding to the initial position of the piston. The shock wave is moving from right to left. The image also shows the presence of small structures described in the text. darker fine structures (thickness 10 pixels) close to this position cannot be guaranteed as coming from the shocked xenon nor gold. It is interesting to note that the hottest part of the precursor (between pixels x = 800 and 1100, 13

14 i.e. about 100 µm in the target) absorbs significantly the XRL radiation. Behind the shock (pixel number larger than 1100), i.e. in the CH plasma, the image remains dark. The darkening is here more pronounced near the window horizontal boundaries. This could be attributed to the walls being heated by the shock. XUV imaging of radiative shocks allows to estimate the position of the shock front, and thus the shock timing (independently measured with our time resolved diode diagnostics). For precise quantitative studies it is necessary to know the opacity which varies both with density and temperature. The fact that the density in the radiative precursor is almost constant and that the fluid velocity is zero allows to use the diagnostics as a thermometer of the shock, and to prove the presence of a radiative precursor. 4. Result from diodes The shock velocity estimated from the XRL probing is confirmed by the plasma self-emission registered by the two fast diodes, PERP and OBL. Both of them recorded the plasma self-emission coming from a section of the shock tube located at 2.2 mm from initial position of the piston, and with an emitting surface of about 0.2 x 0.4 mm 2. The diodes were not filtered and no analogical attenuators were used on them. The resulting record is reported in the figure 8. In both cases, we observe a strong peak at time t = 0, which is attributed to stray light during the intense interaction between the AUX beam and the CH-Au foil. Although special care was taken to minimize this stray light, (placing several radiative shields), we did not succeed to suppress this effect. The observed saturation of the signal at these early times affects 14

15 Figure 8: Signal of the PERP (red) (multiplied by 5) and OBL(blue) photodiodes.the arrows indicate examples of speed for shocks passing at the center of the diodes observation zones. mostly PERP diode. As a consequence an electrical echo is observed at 15 ns, due to an impedance mismatch of the coaxial feedthrough at the vacuum chamber wall. The signal from the 2 diodes increases after 20 ns, reaching a maximum at 40 ns, which is more pronounced in diode OBL. This maximum is followed by a decrease (almost a plateau) for diode PERP and a secondary peak at 80 ns. The calculated time dependent local values of mean ionic stage T(eV) and ionic charge < Z >, at the center of the 0.2 x 0.4 mm 2 observational window recorded by the diodes, are reported on Figure 9. The plot shows a strong increase of the temperature with time, which is associated to the 15

16 Figure 9: Time variation of mean ionic stage < Z > and temperature T in ev in the plasma at a distance of 2.2 mm, from the initial position of the piston. Figure 10: Estimation of the current generated by the OBL diode, looking at 2.2 mm behind the initial position of the piston, supposing a blackbody emission through the Si3N4 windows from a section of the heated plasma corresponding to 0.2 x 0.4 mm 2. 16

17 observation of the precursor through the window. The peak is reached when the shock front appears in the window, with a temperature of 24 ev from 1-D simulations. Then the temperature decreases abruptly. The expected corresponding theoretical signal (blackbody radiation assumption) is plotted in Figure 10. It shows a peak in the emissivity at 36 ns. This timing is close to the experimental value of 40 ns obtained from diode OBL (Figure 8). Concerning the peak intensity (the shock front), results from XUV imaging at 20 ns and self-emission diagnostics at mm are consistent with 1-D simulations, i.e. a shock wave propagating at 60 km/s. Strong differences between theory and experiment however, are observed after the front shock has passed through the observational window (t > 40 ns), as the target presents a strong residual emission. The reasons for this discrepancy are not yet understood. The possibility of a xenon trail after the gold layer and the heating of the walls will be investigated in future experiments. In addition, the differences between diodes PERP and OBL at late times are not understood yet. Finally, the computed peak value for the current in the diode of 60 A, results in a probe output 3 orders of magnitude larger than observed. However the theoretical assumption used for this computation is very crude in terms of radiative flux modelling. Moreover additive losses due to misalignment, absorption in the pinhole, cable attenuation, radiative cooling in the shock wave as radiation losses at the windows are certainly very important. A more detailed, 3-D modelling of the radiative shock, including effects such as the wall albedo and a 3-D radiative transfer post-processing, will be the object of future studies. 17

18 5. Conclusions We present the first experimental study of radiative shock waves in gases by simultaneous XRL imaging and spatially resolved plasma self-emission. The results obtained by both diagnostics are consistent and in qualitative agreement with the chronometry expected from 1-D simulations. XRL backlighting is particularly interesting as it allows probing not only the shock front but also a part of the radiative precursor. Using high quality imaging mirrors this XRL diagnostic will allow to visualize departures from 1D behavior, as for instance the precursor bending or the heating of the shock tube. As the mass density in the radiative precursor is equal to that of the unperturbed xenon, it is important to point out that the XRL probing would theoretically allow to measure the temperature, supposing the opacity at 21 nm correctly known from theory or from other dedicated experimental studies. In addition, complementary diagnostic linked to ionization, for instance through XUV laser interferometry [23], is required to deduce both the density and temperature. Concerning the study of the plasma self-emission using diode diagnostics, the present result opens a door to more detailed studies, using for instance diodes with various filters to improve the evaluation of the temperature of the emitting plasma. Future studies will take benefit from 2D hydro-simulations, which were already successful in the interpretation of former experiments with larger channel sections. Diminishing the section of the tube will decrease the optical depth and thus increase the radiation losses, leading to shock curvature. Moreover, the interaction between the plasma and the tube itself would be interesting to study in support to these experiments. 18

19 6. Acknowledgements The authors acknowledge the vital contributions of the PALS technical staff, the target fabrication staff of Observatoire de Paris, the important contribution of V. Petitbon at IPN for the piston manufacturing, and illuminating discussions with F. Delmotte and S. de Rossi (Institut doptique, University Paris XI), B. Rus (Institute of Physics in Prague), F. Delahaye, A. Ciardi (LERMA), M. Gonzalez (IRFU). The work was supported by LASER- LAB access program, French ANR, under grant 08-BLAN , PICS4343 of CNRS, by University Pierre et Marie Curie and Observatoire de Paris. This work was partially supported by the Academy of Sciences of the Czech Republic (Project No. M ), by the Czech Ministry of Education, Youth and Sports (Project Nos. 7E08099 and 7E09092), and by the Czech Science Foundation (Grant No. 202/08/1734). References [1] S. Bouquet, C. Stehlé, M. Koenig, J.P. Chièze, A. Benuzzi-Mounaix, D. Batani, S. Leygnac, X. Fleury, H. Merdji, C. Michaut, F. Thais, N. Grandjouan, T. Hall, E. Henry, V. Malka, J.P.J. Lafon, Phys. Rev. Lett. 92 (2004) [2] A.B. Reighard, R.P. Drake, K.K. Dannenberg, D.J. Kremer, M. Grosskopf, E.C. Harding, D.R. Leibrandt, S.G. Glendinning, T.S. Perry, B.A. Remington, J. Greenough, J. Knauer, T. Boehly, S. Bouquet, L. Boireau, M. Koenig, T. Vinci, Physics of Plasmas, 13 (2006) [3] M. González, E. Audit, C. Stehlé, A&A, 497 (2009)

20 [4] C. Stehlé, M. Gonzalez, M. Kozlova, T. Lanz, B. Rus, T. Mocek, O. Acef, J.P. Colombier, N. Champion, K. Jakubzak, J. Polan, P. Barroso, D. Bauduin, E. Audit, M. Stupka, Laser Part. Beams, 28 (2010) 253. [5] Y.B. Zeldovich, Y.P. Raizer, Physics of shock waves and high temperature hydrodynamic phenomena, New York: Academic Press, [6] A. Penzo, J. Tassart, Journal de Mécanique Théorique et Appliquée, 3 (1984) 381. [7] Yu N. Kiselev, I.V. Nemchinov, V.V. Shuvalov, Computational Mathematics And Mathematical Physics, 31 (1991) 87. [8] S. Bouquet, R. Teyssier, J.P. Chièze, Astroph. J Supp., 127 (2000) 245. [9] C. Michaut, C. Stehlé, S. Leygnac, T. Lanz, L. Boireau, Eur. Phys. J. D, 28 (2004) 381 [10] S. Leygnac, L. Boireau, C. Michaut, T. Lanz, C. Stehlé, C. Clique, S. Bouquet, Phys. of Plasmas, 13 (2006) [11] M. González, C. Stehlé, E. Audit, M. Busquet, B. Rus, F. Thais, O. Acef, P. Barroso, A. Bar-Shalom, D. Bauduin, M. Kozlova, T. Lery, A. Madouri, T. Mocek, J. Polan, J. Laser Part. Beams, 24 (2006) 535. [12] F.W. Doss, H. F. Robey, R.P. Drake, C.C. Kuranz, Phys. of Plasmas, 16 (2009) [13] K. Jungwirth, A. Cejnarova, L. Juha, B. Kralikova, J. Krasa, E. Krousky, P. Krupickova, L. Laska, K. Masek, T. Mocek, M. Pfeifer, 20

21 A. Präg, O. Renner, K. Rohlena, B. Rus, J. Skala, P. Straka, and J. Ullschmied, Physics of Plasmas, 8 (2001) [14] B. Rus, T. Mocek, A.R. Präg, M. Kozlová, M. Hudecek, G. Jamelot, A. Carillon, D. Ros, J-C. Lagron, D. Joyeux, D Phalippou, Plasma Phys. Control. Fusion, 44 (2002) B207 [15] B. Rus, T. Mocek, A.R. Präg, M. Kozlová, G. Jamelot, A. Carillon, D. Ros, D. Joyeux, D Phalippou, Phys. Rev. A 66 (2002) [16] M. Kozlová, B. Rus, T. Mocek, J. Polan, P. Homer, D. Snopek, K. Jakubczak, M. Fajardo, A. Barszczak Sardinha, X-Ray Lasers 2008, Proceedings Book Series: Springer Proceedings in Physics Volume: 130 (2009) 417. [17] H. Fiedorowicz, A. Bartnik, M. Szczurek, H. Daido, N. Sakaya, V. Kmetik, Y. Kato, M. Suzuki; M. Matsumura, J. Tajima, T. Nakayama, T. Wilhein, Optics Comm., 163 (1999) 103. [18] A. Sazaki, A. Sunahara, K. Nishihar, T. Nishikawa, F. Koike, H. Tanuma, The Review of Laser Engineering Supplemental Volume 36 (2008) [19] E.R. Kieft, K. Garloff, J. J. A. M. Van der Mullen, V. Banine, Phys. Rev. E 71 (2005) [20] J. Blackburn, P.K. Carroll, J. Costello, G. OSullivan, J. Opt. Soc. America, 73 (1983)

22 [21] R. Ramis, R. Schmalz, J. Meyer-Ter-Vehn, Comp. Phys. Comm., 49 (1988) 475. [22] B.L. Henke, E.M. Gullikson, and J.C. Davis, Atomic Data and Nuclear Data Tables, 54 (1993) 181, also at constants/. [23] L. M. R. Gartside, G. J. Tallents, A. K. Rossall, E. Wagenaars, D. S. Whittaker, M. Kozlová, J. Nejdl, M. Sawicka, J. Polan, M. Kalal, and B. Rus, Opt. Let., 35 (2010)