Suppression over a high dynamic range of ase at the rising edge of ultra-intense femtosecond pulses

The autocorrelator to measure the duration of frequency doubled pulses of a Ti:Sa laser developed in MBI uses a new principle, namely self-diffraction in a material with a high polarisation vector /3 like sapphire. The fundamental beam (800nm) is split into two parts and then focused in non-linear medium a thin sapphire plate. Temporally and spatially overlapped, the beams interfere producing such an intensity dependent phase diffraction grating, which scatters the beams. The intensity of the diffracted signal is recorded depending on the delay between the 2 beams.

Acting as LOA sub contractor, Fastlite company was in charge of developing one prototype of ultrafast Pockels cell able to switch an optical gate in less than 200 picoseconds with a contrast ratio better than 10. Principle: This cell is composed by 2 RTP crystals inserted in ceramic transmission lines exhibiting a dielectric constant very close to the RTP one. Crystal is surrounded by 2 photo-conductive switches in GaAs (Gallium Arsenide) allowing a very fast commutation. The principle consists in the discharge of one part of the transmission line (RTP crystal) in the 2 lines surrounding this crystal. The switching time is very fast (<1ps) and generate a negative pulse with an amplitude half than the load voltage applied on both photo-conductive switch. These 2 pulses propagate in the transmission line at a speed c/Õr (about 12 107m/s). The size of RTP crystals is 6 x 6 x 7mm. So the propagation time through the crystals is about 50ps. The switching time of the Pockels cell is also linked to the light velocity in the crystal (40ps).

The development of high-energy seeding pulses from a Kerr-lens mode-locked oscillator is one of the techniques explored in the SHARP program in order to reduce the level of the Amplified Spontaneous Emission (ASE) at the output of the laser chain. By seeding the preamplifier with energy two orders of magnitude higher than it is usually done, we will reduce the overall gain by the same amount and therefore decrease the level of ASE by at least two orders of magnitude. Conventional Kerr-lens mode-locked oscillators delivers very short pulses (10fs) at a repetition rate of 100MHz with an energy of few nJ. The advantages of cavity-dumping Kerr-lens mode-locked oscillators is to keep the temporal properties of the pulse (short pulse duration and cleanliness on a high dynamic range) while being able to extract the intra-cavity energy which is largely higher than the one coming from the output coupler. The output energy of cavity-dumped oscillators depends on the technique that is used to extract the pulse from the cavity. Usually acousto-optic modulators are used, but nowadays the available energy is limited to 30nJ. Our oscillator is based on electro-optic cavity dumping with a Pockels cell. That one is able to work with a beam size up to 4 mm which allows the inside cavity energy to be largely higher than with an acousto-optic device.

Double CPA with non linear induced birefringence rotation: The principle of the double CPA concept is to build a first CPA laser system in order to obtain femtoseconde pulses at a millijoule level. Those pulses are filtered thanks to a nonlinear filter in order to obtain pulses of a few hundred of microjoules with a high contrast. The laser beam is then amplified again in a second CPA laser chain. The nonlinear filter presented here is based on nonlinear induced birefringence in air. In this experiment, the laser is focus in the air. At the focal point, the intensity of the femtosecond part of the laser pulse is high enough to make some nonlinear effects, which turn the polarisation of the laser beam. The intensity of the ASE part or of the satellite pulses does not generate such effects. Thanks to a polarizer, it is then possible to discriminate the femtosecond pulse from the ASE background. The pulse cleaning depends of the pulse energy, the beam diameter at the focal point, the chirp of the pulse before the filter. The best results we obtained were 125 µJ, 55fs pulses with an improvement of the contrast by 3 orders of magnitude. Thanks to the possibility to make a spatial filtering during the process, the spatial quality of the laser beam is good. Those results are very sensitives to the experimental parameters which have to be carefully adjusted.

In CPA lasers, amplified spontaneous emission (ASE) forms a pedestal underneath the main pulse. The ASE originates in the front-end amplifiers. By using appropriately gated Pockels cell (PC) shutters, its duration can be restricted to less than 10ns. This is still long even versus the duration of the stretched main pulse that is typically a few hundred ps long. Due to its incoherence, the ASE cannot be compressed and hence leaves the compressor unchanged. In multi-TW titanium:sapphire lasers, the ratio of the peak intensity of the compressed main pulse and the ASE intensity is of the 1019W/cm², the ASE intensityorder of 107. When the main pulse is focused to reaches 1013W/cm². For solid targets, this value is sufficiently large to create a preplasma with which the main pulse interacts. An important parameter in this context is the scale length of the preplasma. In order to investigate its effect, the ASE preceding the main pulse has to be controlled. Following the philosophy of minimal invasion into the laser facility, this can be achieved in a relatively simple manner by incorporating an ultrafast PC shutter into the beamline. Since the rise time of a PC scales with the diameter of the KDP crystal, the diameter of the main pulse should be small. To meet this requirement, the shutter is best placed after the amplifier in which the main contribution to the ASE originates. In our ATLAS (Advanced Titanium: Sapphire Laser) facility, this is the regenerative amplifier. After this amplifier, the beam diameter is expanded from 2mm to 6mm to avoid damage in the Glan prisms as well as in the slow and fast PCs. The first slow shutter consists of the first two small Glan prisms with the PC3 placed in-between and has a 12-ns long transmission window. Hence the ASE can be maximally 12 ns long. However, actually it is shorter because the main pulse is positioned a few ns behind the leading edge of the window.

In CPA lasers, amplified spontaneous emission (ASE) forms a pedestal underneath the main pulse. The ASE originates in the front-end amplifiers. By using appropriately gated Pockels cell (PC) shutters, its duration can be restricted to less than 10ns. This is still long even versus the duration of the stretched main pulse that is typically a few hundred ps long. Due to its incoherence, the ASE cannot be compressed and hence leaves the compressor unchanged. In multi-TW titanium:sapphire lasers, the ratio of the peak intensity of the compressed main pulse and the ASE intensity is of the order of 107. When the main pulse is focused to 1019W/cm², the ASE intensity reaches 1013W/cm². For solid targets, this value is sufficiently large to create a preplasma with which the main pulse interacts. An important parameter in this context is the scale length of the preplasma. In order to investigate its effect, the ASE preceding the main pulse has to be controlled. Following the philosophy of minimal invasion into the laser facility, this can be achieved in a relatively simple manner by incorporating an ultrafast PC shutter into the beamline. Since the rise time of a PC scales with the diameter of the KDP crystal, the diameter of the main pulse should be small. To meet this requirement, the shutter is best placed after the amplifier in which the main contribution to the ASE originates. In our ATLAS (Advanced Titanium: Sapphire Laser) facility, this is the regenerative amplifier. After this amplifier, the beam diameter is expanded from 2mm to 6mm to avoid damage in the Glan prisms as well as in the slow and fast PCs. The first slow shutter consists of the first two small Glan prisms with the PC3 placed in-between and has a 12-ns long transmission window. Hence the ASE can be maximally 12 ns long. However, actually it is shorter because the main pulse is positioned a few ns behind the leading edge of the window.

A new kind of nonlinear filtering technique is studied, based on a based on cross-polarized wave generation in non linear centro-symmetric crystals. Cross-polarized wave (XPW) is generated by four wave mixing governed by the anisotropy of the third order nonlinearity. We have selected this process because of the perfect group velocity matching. We used both BaF2 and YVO4 0.9mm long crystals. A full description of this process applied to cubic and tetragonal crystals has been previously detailed (N. Minkovski, G. I. Petrov, S. M. Saltiel, O. Albert and J. Etchepare, J. Opt. Soc. Am. B, vol. 21, p. 1659 (2004)). The contrast enhancement has been demonstrated to be only limited by the extinction ratio of the output polarizer (4.6dB in our case) by measuring the temporal profiles of the pulse before and after filtering with a third order cross correlator. A contrast improvement better than 4 orders of magnitude has been recorded. The energy converted from the femtosecond part of the main pulse to the XPW is about 10%. These very promising results have been published in Opt. Lett., vol. 30, n°8, p. 920 (April 2005). Work is in progress by investigation of different non-linear crystals to increase XPW generation efficiency.

Summary -Early work (2001-2002): detailed experimental and theoretical study of plasma mirror. - First application (one stage plasma mirror) to harmonic generation on solid target at CEA 2002 - Two stages plasma mirror configuration at LOA 2003-2004 - Early work on a two stages plasma mirror at CEA 2004 - Conclusion. - Main publications. - Deliverables. - Detailed experimental and theoretical study of plasma mirror: In this work we demonstrate that, using single plasma mirror with anti-reflection coating the pulse contrast increases by more than 102, with efficiency higher than 60%, and a good beam quality. More detail can be obtained from ref(1). In order to get a good description of the plasma mirror dynamic a hydrodynamic code was build. This code was validated by comparison with our experimental data as an accurate benchmark. We can now assess the performance of PM mirror and determine the optimum operational fluence on the mirror. The user-friendly version of the hydro-code on Labwiew platform is available on request. Several plasma mirror regimes have been identified, the most interesting is the �robust regime�: Pulses shorter than 600 fs and fluence between the triggering threshold (5J/cm) up to 500 joules/cm2. At this condition, the reflectivity is mostly insensitive to the fluence, and is around 70 %. - First application to harmonic generation at CEA with one stage plasma mirror: The experience has been performed on the UHI laser of the Saclay Laser Interaction Center (SLIC). This Ti.Sapphire laser produces pulses of 600mJ and 60fs duration. The laser beam was focused by an f/6 off-axis parabola (f=500mm). By setting anti-reflection coated plasma mirror 12mm before the focus the fluence on the PM is about 60 J/cm2. The beam was S polarized with respect to PM surface. Contrast measurement The first part of experience aimed at measuring the improved temporal contrast and beam quality. After the PM and after focalization, the pulse was attenuated and collected by an image relay system. The observed focal spot shown no degradation compare to configuration without PM. The temporal contrast has been measured using a SEQUOIA3 third order correlator. This measurement show that the contrast has been improved by about two orders of magnitude, as expected. This improvement extent from ASE to coherent pedestal. Harmonics generation. The second part of experience aimed at using the PM to study the interaction of the focalized pulse on solid target.. The peak intensity on the target was estimated to be 3 1018 W/cm2.. The light reflected from the target was send to a flat field XUV spectrometer. High contrast pulse allows interaction with a steep electronic density gradient. As expected a well-collimated beam of even and odd harmonics was clearly observed up to the 20th order. Paper on this work as been published on Appl.Opt.(ref 2) To extend this study in the truly relativistic regime I >> 1018 W/cm2, the contrast should be increase more and that requires the use of two successive plasma mirrors. - Double plasma mirror configuration at LOA: A double plasma mirror has been implemented one the �Laser salle jaune� of the LOA at ENSTA. This Ti.Sa. laser produces 1 Joules, 35 fs, pulses. The main goal was a 104 contrast increasing, using a user-friendly DPM configuration. Of course this DPM should be easily removable. The plasma mirror should be put at the laser front end, before the experimental chamber. At least on plasma mirror is working in the intermediate field where the amplitude distortions are larger for actual beam. In order to optimize this type of configuration and predict actual performance, we have build a time dependent 2D optical propagation code. This codetake into account local reflectivity on plasma mirrors and propagation between mirrors and to the final target. This open source code has been build with Labview platform and is easily transportable As conclusion using this new installation, Contrast greater than 1011 is obtain, Output beam can be focused in a very nice focal spot, More than 50% efficiency obtained on target. - Early work on a two stages plasma mirror at CEA.-2004: Construction of this double plasma mirror is extrapolated from the LOA configuration in a more compact scheme. The overall length is reduced to 2 meter at the expense of a reduced number of shot (100), before changing the plasma mirror.. - Main publications : -- Phys. Rev. E 69, 026402 (2004) -- Optics Letters 15, / 2004 / vol. 29, No. 8 - Deliverables: Construction of Single plasma mirror at CEA , Construction of double plasma mirror LULI and CEA, 2D simulation code for multiple plasma mirror (Open source on Labview platform).

This result consists in using a CS² cell acting as a Kerr gate pumped by one part of the amplified beam. Instead of using a liquid CS² cell , we can use a solid state non linear medium exhibiting a faster time response than the liquid medium. Intrinsic overall efficiency of this type of pulse cleaner can reach 80%, but if we take in account the fact that 80% of the laser beam itself is used as the pump beam the real efficiency is dropping down to 10 to 15%.

Double CPA concept with non linear filtering : The principle of the double CPA concept is to build a first CPA laser system in order to obtain femtoseconde pulses at a millijoule level. Those pulses are filtered thanks to a nonlinear filter in order to obtain pulses of a few hundred of microjoules with a high contrast. The laser beam is then amplified again in a second CPA laser chain. The nonlinear filter presented here is based on a sagnac interferometer. On one arm of the interferometer, a non linear medium is introduced (a block of BK7 in our case). The femtosecond part of the laser pulse generate strong nonlinear effets in the medium which change the phase of this part of the pulse whereas it is not the case of the ASE part and of the satellite pulses of the laser beam (there are few nonlinear effets). The value of the phase difference is a fonction of the laser beam diameter, the pulse energy and the length of the nonlinear medium. When those parameters are well suited, it is possible to obtain a very clean pulse at the output of the interferometer. The best results obtained were 200µJ, 50fs pulses with a contrast improvement around 4 order of magnitude. Unfortunately, due to the important B integration inside the interferometer, the spatial profile of the output beam is disturbed. During the filtering, the spatial defaults of the input beam are amplified and some hot spots could appear.

The high dynamic correlator is based on an efficient all reflective attenuation system coupled with high efficiency thin nonlinear frequency conversion BBO crystals to keep ultra-short duration pulses with high intensity at the zero delay. The system allows to measure a contrast dynamic of 10{12}, covers a temporal range of one nanosecond, which matchs measurements achieved with fast photodiode , and has a temporal resolution of 80fs. Removable off-axis parabolic mirrors were added to allow direct measurements of the temporal contrast of femtosecond Titanium sapphire oscillators by focusing the beams onto the non-linear crystals.

A compact photolytical XeF(C-A) amplifier for direct amplification of high-power visible femtosecond laser pulses has been developed. This device uses a low density gaseous amplifier medium, attractive for ultrafast laser amplification due to its low non-linear index of refraction making possible direct amplification, without pulse stretching, of high-power ultrashort pulses, its high breakdown threshold and scalability to very large volumes. Among all gas laser media, application of the photolytical XeF(C-A) laser for high energy amplification is attractive for the development of ultra-high power laser systems up to the petawatt power level due to the XeF(C-A) broad amplification bandwidth (80nm FWHM centered near 475nm) and a rather high saturation fluence (0.05 J.cm{-2}), as well as a very low level of Amplified Spontaneous Emission. In that context, a XeF(C-A) amplifier cavity of 50cm long with a clear aperture of 20 x 4.5cm{2} has been designed and put into operation. The XeF(C-A) active medium, consisting of XeF2/1:N2/37:Ar/730 gas mixture under 1 bar total pressure, is photolytically pumped by two VUV optical planar pumping sources located in parallel oppositely to each other and having the following parameters: a large radiating area (hundreds of cm2) and a high brightness temperature (> 20 000 K) of the discharge plasma, a rapid discharge circuit with submicrosecond light-pulse duration and a low jitter of initiation (<50ns). Measurements of optical characteristics (pump power, gain, ASE, etc.) of the amplifier have been performed. A pilot experiment to test the capability of XeF(C-A) medium for femtosecond amplification has been realized. A commercial femtosecond laser system delivered linearly-polarized 150fs seed pulses with energy of several microjoules at 480nm to the amplifier cavity. The seed pulse entered the amplifier through a CaF2 window and made a dozen roundtrips inside, reflecting from the mirrors. The medium was pumped by a microsecond pulse of VUV radiation from a multi-channel sliding discharge consisting of 50 parallel discharge channels simultaneously initiated on the area of 40 x 18 cm{2} along one side of the rectangular amplifier cavity. This pumping configuration corresponds to a one-side pumping scheme (two sides are today operational). Amplification of the seed pulse energy by a factor of 5 was registered corresponding to a small-signal gain of 0.0016 per cm. The amplified pulse spectrum exhibits narrow-band absorption features but much weaker than for a e-beam pumped medium. This is of prime importance for amplification of femtosecond pulses as any distortion in the amplified spectrum leads to an increase of the final pulse duration. These weak narrow-band absorption lines, corresponding to the Rydberg series of transitions from the Xe (6s3P0) excited state, saturate at several times smaller energy density than XeF(C-A) one, thus allowing to avoid pulse width distortions during the amplification process. This very particular gas medium also shows a low level of ASE due to its low small signal gain and a relatively long radiative lifetime (100ns). Moreover, it is important to underline that the pumping used (photodissociation by a VUV radiation) is the most suitable for preserving the spatial and temporal phase of the amplifying beam. As a conclusion, this makes this photolytically XeF(C-A) amplifier device very promising for obtaining high contrast high power femtosecond pulses. The high contrast is obtained through frequency conversion of Ti:Sa ultrafast laser sources ensuring laser pulse cleaning at low energy and direct amplification in a gaseous medium having a very low non-linear index of refraction and generating a very small ASE pedestal (< 1 W.cm{-2}). A further advantage of the laser chain resides in the suppression of the vacuum grating compressor at the end of the laser chain thus eliminating a source of pre-pulses and significant energy loss and greatly reducing complexity and cost of the laser system. This should allow the development of multiterawatt femtosecond laser with > 10^10 temporal contrast. This particular amplifier might then replace or be an alternative to conventional solid-state power amplifiers in multiterawatt and even petawatt ultrafast laboratory facilities. Due to its emission bandwidth in the visible, this power amplifier shall also provide very high energy ultrafast laser pulses in the blue-green region outclassing the best performances of conventional infra-red frequency-doubled ultrafast lasers in this spectral region. Applications of XeF(C-A) amplifier should concern upgrade of high energy ultrafast laboratory facilities and related high-field physics experiments concerning time-resolved dense plasma diagnostics, generation of laser plasma and X-ray sources, ignition of photonuclear reactions, particle acceleration or to explore the relativistic regime of the interaction of radiation with matter.

FORTH has proceeded with the evaluation and set-up of second and third order background free autocorrelator units as sampling autocorrelators with high dynamic range for the UV fs excimer amplified pulses in collaboration with TUC in Chania. A streak camera time profiling with ps resolution of the amplified pulse has been carried out. In the first year of the project we proceeded on the construction and testing of a non-collinear VUV third harmonic generation in noble gasses Michelson interferometer. The setup is a general-purpose unit consisting of a two arm Michelson the output of which is passed through a noble gas cell for the generation of the third harmonic and its detection as a function of the relative delay of the two beams. Additionally a similar Michelson interferometer was built in combination with static NO cell detector. The two-photon ionization of NO from 248nm radiation provides the second order autocorrelation measurement of the UV beam as a function of the delay of the two beams. The two autocorrelators have been assembled and tested. The 3rd order autocorrelator with the fundamental of an amplified Ti:Sa system and the 2nd order autocorrelator with a subpsec KrF system. The autocorrelators will be adapted to the operational parameters of a Xe based excimer in collaboration with the TUC partner, once available. The contrast of the amplification process has been further studied and quantified by ps streak camera work with the amplified pulse.

Leistungen

Diese Seite teilen

Herunterladen