The chirped-pulse amplification (CPA) technique [1] proposed in 1985 has been used to enable the rapid development of ultra-intense and ultra-short lasers, and extremely high-field ultrafast lasers can create unprecedented comprehensive laboratory-scale extreme physics. Until now, the petawatt (PW)-level laser systems using CPA and the optical parametric chirped pulse amplification (OPCPA) techniques [2] have been realized in several laboratories and companies [3 –10]. Many laboratories in the world are engaging in the development of a 10-PW-level laser system based on CPA and OPCPA techniques, such as ELI [11], Vulcan-10PW [12], SIOM-10PW [10], and PEARL-10PW [13]. Compared with the OPCPA technique, the CPA technique particularly using Ti:sapphire (Ti:S) CPA systems is still the main method to achieve PW and 10 PW-levels laser pulses for its high efficiency and stability. As a single-beamline prototype for the ELI project, the APOLLON 10-PW facility in progress aims at generating a pulse of 150 J/15 fs with a power of 10 PW is based on the Ti:S CPA technique [14]. The Ti:S CPA technique is more mature for obtaining a broadband and energy-stable laser pulse with a high conversion efficiency in the output range of several-tens-joule.

Theoretically, the amplified output energy can be greatly improved with the growth of the large-aperture Ti:S crystal [15]. However, the transverse amplified spontaneous emission (TASE) and parasitic lasing (PL) within the booster amplifier volume are more easy to clamp the amplified pulse energy when the larger-aperture Ti:S crystals are pumped at higher pump fluence and energy [16]. With the increase of diameter of the Ti:S crystals, this is the main barrier for realizing high energy Ti:S CPA amplifier. Currently, the techniques for suppressing transverse PL are in two major ways. First, the passive technique including the matched-index cladding technique [5,17 –19] is used to increase the losses of spontaneous emission. Second, the active technique is to control the transverse gain, including optimization of the time delay [20,21], the use of lightly doped Ti:S crystal [5], and so on. Based on these two manners, the maximum output energy reported from a femtosecond PW Ti:S laser system is 72.6 J based on a 100-mm-diameter Ti:S [8], which is far from the desirable output energy. Therefore, to satisfy the energy requirement of a 10-PW laser, a higher energy output must be implemented using larger-aperture Ti:S crystals. Accordingly, the PL suppression of the Ti:S CPA system should be tested using the existing techniques.

To obtain a higher energy output, we investigated the output properties of the Ti:S laser system pumped at a higher pump energy generated by frequency-doubled Nd:glass disk amplifiers at 527 nm. On the basis of matched-index cladding and optimization of the time delay techniques, we used a lightly doped Ti:S crystal to further suppress the PL. To enhance the effect of the time delay on suppression of the PL, we increased the injected seed energy and compared the output energy from the booster amplifier with different injected seed energies of 28 and 35 J. By adjusting the arrival time of the two pump pulses at the Ti:S surface and enhancing the energy of injected seed pulse, the PL was effectively suppressed. In this Letter, we present experimental results from the first hundred-joule level Ti:S CPA system, demonstrating what the authors believe to be the highest-energy amplified pulse produced from a Ti:S CPA system. The system produced an uncompressed output pulse with an energy of 192.3 J and demonstrated compression to 27.0 fs, indicating a potential peak power of 5.13 PW.

The basic design of the system (Fig. 1) relies on a six-stage Ti:S CPA amplifier. The structure was similar to that in Ref. [8], with the addition of a four-pass amplifier to enhance energy stability. The output energy from the front end operating at a repetition rate of 5 Hz was 3.5 J, with a temporal duration of 1.4 ns (FWHM) and a spectral range from 745 to 830 nm. The following power and booster amplifiers were pumped by a frequency-doubled Nd:glass laser operating in a single shot mode. A 80-mm-diameter Ti:S was used in the four-pass power amplifier with an output energy of about 30 J and a beam size of 65 mm in diameter. Then, the amplified laser pulse was up-collimated to a diameter of 110 mm for the final injection into a four-pass 150-mm-diameter Ti:S booster amplifier, as shown in the red shaded box in Fig. 1. The pump laser, shown in the green shaded box in Fig. 1, was seeded from the front-end, which comprised regenerative and rod amplifiers, similarly to the case in Ref. [8]. The rod amplifiers were split into two beam lines. One was used to pump the 80-mm-diameter Ti:S, and the other was also split into two beam lines for injection into disk amplifiers. Each amplifier chain comprised a four-disk amp and a six-disk amp, and the 1053 nm pulse can be amplified to a maximum of 400 J with a 105 mm beam. After the frequency doubling, the two pump beams were up-collimated to 120 mm, and were spatially filtered, with the image relayed to the Ti:S crystal surface with a maximum output energy of 190 J at 527 nm for each channel.

 

Fig. 1. Schematics of the CPA experimental setup.

Download Full Size | PDF

To suppress the PL in the final amplifier, we used the Cargille Series M refractive index liquid doped with an absorber (IR 140) as the cladding material. Ti:S is known to be a uniaxial crystal. According to the Sellmeier equation of Ti:S [22], the refractive index curves of o-light and e-light in Ti:S and refractive index liquid are shown in Fig. 2. Although the index difference between the matching liquid and e-light is very small at the central wavelength, there is an offset at the edge of the spectrum, especially the difference with o-light in Ti:S. Based on the previous experimental results, the difficulty in realizing exact index matching restricts the diameter of the pump area to $\sim 6\mathrm{cm}$ corresponding to an extracted energy of about 30 J from Ti:S. Therefore, the index-matching cladding technique is inevitably limited for large-aperture Ti:S crystals pumped at a high pump fluence and energy.

 

Fig. 2. Refractive index curve of o-light (solid blue line) and e-light (solid red line) in Ti:S and a refractive index liquid (solid green line).

Download Full Size | PDF

Additionally, we used a lightly doped Ti:S crystal (GT Crystal Systems LLC) with 150 mm in diameter and 46.7 mm in thickness. In the certificate of compliance of the 150-mm-diameter Ti:S, the flatness and TWE are specified to be $\lambda /10\mathrm{a}\mathrm{t}632\mathrm{nm}$ and $\le \lambda /2\mathrm{a}\mathrm{t}632\mathrm{nm}$ across the clear aperture. The pump absorption coefficient of the Ti:S crystal is about 0.66/cm. Compared to the highly doped Ti:S crystals, the gain distribution in the Ti:S crystal is more even for lightly doped Ti:S crystals. The time delay between the seed and pump pulses was precisely manipulated by a master clock (THALES Optronique SA) circuit with a time jitter less than 50 ps, which has little influence on energy stability. The time intervals between passes in the Ti:S amplifier were 11 ns ($\mathrm{\Delta }{T}_{12}$), 12 ns ($\mathrm{\Delta }{T}_{23}$), and 13 ns ($\mathrm{\Delta }{T}_{34}$). On the basis of the time interval between the first and second passes, we set the arrival time at the Ti:S surface of the two pump beams by changing the optical distance difference between the two pump pulses (Fig. 3). In this design, the time interval (11 ns) between the first pass and the front edge of the first pump pulse is equal to the time interval between the second pass and the front edge of the second pump pulse [as shown in Fig. 3(b)]. This design keeps the inverted population density in the Ti:S crystal at low levels during the entire amplification process by optimizing energy extraction of the seed pulse during amplification. The pulse duration of the pump beam was 15 ns for FWHM and $\sim 26\mathrm{ns}$ for $1/e^2$ full width [green dashed cure in Fig. 3(b)].

 

Fig. 3. (a) Schematic sketch for the time delay of two pump pulses and (b) Illustration of the time delay.

Download Full Size | PDF

The energy of the injected seed pulse has a significant influence on transverse gains. Because the transverse gain of two surfaces is higher than the inside area, we calculated the transverse gains of the front surface with the pumping time at different injected energies using the theoretical model in [20]. The parameters used in our simulation are based on the above experimental values. Figure 4 shows the surface transverse gains with injection of 28 and 35 J, respectively. The time delay is 16 ns, and the pump energy is 280 J. The results indicate that the surface transverse gains are further reduced at injection of 35 J compared to 28 J.

 

Fig. 4. Surface transverse gain with pumping time at different injected seed pulse energies: 28 J (blue solid curve), 35 J (red solid curve). The green dashed curve is the temporal pump profile.

Download Full Size | PDF

Based on the aforementioned passive and active PL-suppression techniques, we measured the output energy [Fig. 5(a)] and conversion efficiency [Fig. 5(b)] with respect to the pump energy. For an injected seed energy of 28 J (black rectangles), the output energies and conversion efficiencies increased with the pump energy, and the maximum amplified output energy of 154.3 J was obtained at pump energy of 262 J. The optimized time delay was 19 ns. When the pump energy increased further up to about 275 J, the PL occurred strongly, and the output energy was decreased to 138.8 J. The output energy was not improved significantly, although we continued to change the time delay at a pump energy of 275 J. To suppress PL, we further improved the injected seed energy of the final amplifier based on our simulation. The amplified output energy from the power amplifier was improved by 25% from 28 to 35 J by increasing the pump energy to 75 J, and maintaining other parameters unchanged. With the injection of 35 J (red circles), the output energy was also not improved if the time delay remained at 19 ns with the pump energy of 275 J. When the time delay was optimized to 16 ns, a maximum amplified output energy of 192.3 J was obtained at pump energy of 312 J, which corresponds to a conversion efficiency of 50.4%. The experimental results confirmed that the improvement of seed injection made the PL suppressed more effectively. After the final booster amplifier, the near field of the amplified laser beam was measured, as shown in the inset figure of Fig. 5(b). One sees a flat-top spatial beam profile due to gain saturation, with some intensity modulations. Relative RMS intensity modulations are 22 and 18% for the horizontal and vertical directions, respectively, which are acceptable because large-aperture, thick Ti:S crystals are difficult to be made very homogeneous.

 

Fig. 5. Measured: (a) amplified output energy and (b) conversion efficiency with respect to pump energy for injection seed energy of 28 J (black rectangles) and 35 J (red circles), respectively. The inset figure in (b) shows the near field spatial profiles of the amplified laser beam after the final amplifier.

Download Full Size | PDF

The evolution of the laser spectra across the whole system is shown in Fig. 6. Through adjusting the spectral filter in the regenerative amplifier (RA), the spectrum after RA (solid green line) was shaped to be preweighted on the blue side. As shown in Fig. 6, the redshift and gain narrowing of the amplified spectrum were well controlled. The spectral width of the amplified output pulse from the final booster amplifier can support 22.4 fs transform limited pulse. For optimization of the pulse compression, part of the amplified pulse was recompressed by a small size grating compressor consisting of four 1480 groove/mm gold-coated holographic gratings. Autocorrelation traces of amplified pulse were measured and shown in Fig. 7. Assuming a deconvolution factor of 1.414, the duration of the compressed pulse was 27.0 fs. The compressor throughput efficiency was measured to be 72%. Assuming similar throughput efficiency for gratings large enough for the full pulse energy, the compressed pulse energy would be 138.5 J. With 27.0 fs pulse envelope, this would correspond to a maximum peak power of 5.13 PW.

 

Fig. 6. Spectral evolution throughout amplification, RA (solid green line), multipass amplifier 1 (M1, dashed gray line), multipass amplifier 2 (M2 dotted blue line), multipass amplifier 3 (M3, solid pink line), 80-mm-diameter Ti:S output (solid cyan line), 150-mm-diameter Ti:S output (solid red line).

Download Full Size | PDF

 

Fig. 7. Measured autocorrelation traces of a compressed amplified pulse.

Download Full Size | PDF

A wavefront sensor (Phasics, SID-4) was used to measure the wavefront of the output beam. The measured peak-to-valley (PTV) value was $2.5\lambda$, and the RMS value was $0.23\lambda$, which could be corrected by an adaptive optics system [23]. The further wavefront correction of the whole laser system and measurement of focusability will be implemented after the large-size compressor gratings and deformable mirror are assembled. Based on a similar front end to that in the 2.0 PW laser system [8], the pulse contrast was measured to be $1.5\times {10}^{-11}$ (100 ps before the main pulse) at 5 Hz repetition rate without pumping the power amplifier and final amplifier.

In conclusion, we experimentally achieved an amplified output energy of 192.3 J and a conversion efficiency of 50% in a 150-mm-diameter Ti:S CPA amplifier, which is, to the best of our knowledge, the first time to break through hundred-joule level in Ti:S laser systems. We investigated the output characteristics of the final amplifier at high pump energy with different injected signal energies of 28 and 35 J. Meanwhile, the arrival time of the two pump pulses at Ti:S surface was adjusted to enhance the effectiveness of the pump-seed time delay optimization. By optimizing the pump-seed time delay and increasing the injected seed energy, the PL in the final amplifier pumped at 312 J was suppressed effectively. Based upon the pulse duration and efficiency achieved in compressing a small portion of the total output pulse energy, the system can support the highest peak power of 5.13 PW. The experimental results provide an experimental basis for implementing larger aperture Ti:S CPA laser systems with 10 PW ultra-intense and ultra-short pulse output.