Open Access
Add to BibSonomyAbstract
In high average power multi-pass amplifier systems, Pockels cell, used for isolating and controlling number of passes, encounters both limitation of aperture and thermo-effects. We propose and demonstrate for the first time, as far as we know, a reflecting Pockels cell (RPC) which is longitudinally excited based on KD*P utilizing matched a discharge chamber and a copper plate as electrodes. In the RPC, electro-optic crystal can be longitudinally conduction-cooled. This device, with a 40mm × 40mm clear aperture, can be scaled to larger, and driven by one low voltage pulse. Excellent switching efficiency, high static extinction ratio, and negligible thermo-effects have been achieved.
©2010 Optical Society of America
1. Introduction
Many high-average-power laser systems under development for studying of inertial fusion energy (IFE) and high energy intensity physics use a multi-pass amplifier architecture to reduce costs and physical size of the facility [1]. In these amplifier systems, the combination of repetition-rate Pockels cell and a thin-film polarizer forms an optical switch, which is key to control the number of passes that laser-pulse train make through the amplifier cavity. Existing Pockels cells encounter both limitation of scalable aperture and thermo-effects, because the dimension of laser beam is commonly as large as several centimeters, in addition to its high average-power load. The optical absorption, even weak in the case of a Pockels cell, can lead to thermal effects which degrade its performance. These include wave-front distortion, thermal stress, and stress-induced depolarization. Kurtev carried out some experiments to observe thermally induced depolarization loss of Q-switch and its effects on the output energy of a laser oscillator [2]. In the experiment, a Pockels cell based on the longitudinal electro-optical effect in crystal of KD*P was used, and its size was Φ10mm × 25mm. The Nd:YAG laser repetition rate was from 10 to 100Hz. Under these experimental conditions, the depolarization in the Pockels cell began to take place at output power of 5W. While the output power reached 10W, the depolarization loss increased to 12%. The undesirable effects must be carefully controlled in a high performance laser system.
To reduce thermo-optic depolarization, the direct way is to use a thin crystal. Unfortunately, ring-electrode Pockels cells would need to have a crystal thickness of the order of the transverse aperture for the electric field to be homogenous everywhere, thus the large aperture leads to serious absorption. For a transversally excited Pockels cell, the electric field uniformity is independent on the crystal length, but propagation of the optical beam perpendicular to the optic axis exposes the beam to the full intrinsic birefringence of the crystal. For high average power application, it is necessary to use two pieces of identical crystal to compensate the thermal depolarization loss induced by natural birefringence. The birefringence compensated switch developed by LLNL can sustain 300W laser in an aperture of 3.25cm × 6cm [3]. Because the half-wave voltage is inverse proportional to the aspect ratio of the crystal, a large clear aperture and thin crystal leads to very high half-wave voltage and surface electric field. That means it is hard to scale the transversally excited switch up in aperture. The plasma-electrodes Pockels’ cell (PEPC) can adopt thin crystal, and be scaled to large aperture. However, the thermal load in the electro-optic crystal can’t be carried off efficiently because of the low pressure environment [4,5]. In this paper, we demonstrate in detail the design and performance of a reflecting Pockels cell (RPC) for high average power laser systems application. In this device, the electro-optic crystal is released from the low pressure environment, so that it can be longitudinally conduction-cooled, which makes heat transfer path shorter and temperature gradient in the KD*P smaller.
2. Description of the RPC
For an output average power in excess of 200 W (20 J, 10 Hz) at 532 nm, the laser must deliver up to 400 W (40 J, 10 Hz) at 1064 nm, which implies that the Pockels cell in the multi-pass amplifier system was required to handle up to multi-kilowatt in an aperture of 30mm × 30mm. To control the amplified passes that laser pulses make through the cavity, the Pockels cell is required to provide a contrast 100:1 averaged over the clear aperture, or >50:1 locally. The switch efficiency must be high enough (>99%) to allow high extraction efficiency, and the optical aberration must be less than λ/2 at the operating point for sufficient output beam quality [6]. Large aperture requirements limit the electro-optic material to KDP and its deuterated isomorph, KD*P, which are the only materials available at the scale. The lower optical absorption and higher electro-optic coefficients of KD*P make it the best material of choice for high average power application up to now.
The longitudinally excited RPC is a large aperture electro-optic device that is used to rotate the polarization of an incident linearly polarized laser beam of large transverse area. The configuration of the RPC is illustrated in Fig. 1 . The antireflection coated KD*P, with a volume of 40 mm × 40 mm × 5 mm, is covered respectively with a copper-substrate mirror and a discharge chamber on each side. The copper-substrate mirror, also used for heat-sink and grounded electrode, is designed to be in contact with KD*P crystal. The cooling efficiency lies on the mechanical contact status between the copper-sink and the KD*P. On the other side, if the air pressure between the copper and the KD*P is left at normal atmosphere, about 105 Pa pressure would be loaded on the KD*P due to low pressure environment in the discharge chamber. That would result in stress distribution in the KD*P, even break the crystal. To overcome these difficulties, a vacuum airproofed technics is developed to enlarge the contact area by decreasing of the air filling ratio. The chamber, covered with a piece of K9 glass which is designed as a cylinder in view of contrast ratio requirement [6], is filled with neon gas. An alloy with work function of 2.42 eV is chosen as a material of the discharge electrodes. These electrodes are processed into pyramidal structure.
Fig. 1 A photograph of the RPC, installed on an adjustable mechanical support (a). A schematic cut is also displayed which points out the main parts of the RPC (b).
To rotate by 90° the polarization of a linearly polarized laser beam by using Pockel’s effect in the RPC, a minus voltage Vπ /2 would need to be applied across the KD*P through the copper-electrode and the highly conductive transparent plasma formed in the discharge chamber. The linearly polarization of an incident laser beam becomes the rotatory polarization after passing through the excited KD*P for the first time. Then the laser beam is reflected by the copper-substrate mirror, passing through the excited KD*P for the second time. Now the polarization of the incident laser beam has been rotated by 90°. The so-called quarter-wave voltage is given by:
Where λ is the laser wavelength, no and r63 are, respectively, the ordinary optic index and the longitudinally electro-optic coefficient of the crystal. For the wavelength (1064 nm) and for KD*P, Vπ/2 = 3.0 kV. The extra capacitance comes from the plasma sheath that forms at the interface between the plasma and the crystal. The voltage divider expression is [7]:Where Vsw = 5.4 kV is the switch-pulse voltage required for 90° rotation, Csheath = 177.0 pF is the sheath capacitance, and CKD*P = 141.6 pF is the capacitance of the 40mm × 40mm × 5mm KD*P crystal. However, helium in the discharge chamber breakdowns at the voltage equal to or greater than 11 kV based on our previous experience. To realize one pulse process driven RPC, the idea of capacity dividing voltage, we presented earlier, is employed [8]. By adjusting C0 and output voltage of the switch pulse generator, both higher discharging voltage and Vsw between the plasma electrode and copper-substrate can be satisfied.3. Experimental setup
Figure 2 shows the experimental setup used to evaluate the switching performance of the RPC. A Q-switched laser (10 ns pulse duration at 1064 nm) is employed to analyze the RPC. After traversing a Glan prism, a small portion of the horizontal polarized laser beam is diverted into a reference photodiode detector. After traversing the RPC, the reflected laser beam is diverted into the part B by a splitter, and detected by a photodiode. As the dynamic range of the laser energy is limited to several mJ, the absorbing plates are employed to enlarge dynamic range. Two Glan prisms, with 105 extinction ratios, are used to analyze the extinction ratio and the switching efficiency of the RPC point by point across the surface of the crystal shown in Fig. 3 . A power supply, which can output a voltage pulse with bottom width 360 ns and maximal flat voltage −15 kV, is employed to activate the RPC. The switch-pulse voltage Vsw is also monitored with a high voltage probe (TEK P3016A), in order to control the electrical behavior of the RPC.
Fig. 2 The optical bench set up in the laboratory to analyze the RPC. The part A is the one used to detect the incident laser beam, whereas the part B is the one used to measure the reflective laser beam.
The static extinction ratio is the ratio of the detected intensity with the Glan prisms aligned to the intensity with the Glan prisms crossed without activating the RPC. The static extinction ratio is given by:
Where IBP and IAP is the photovoltaic value respectively detected by the part B and the part A while two Glan prisms are aligned, IBV, IAV is that while two Glan prisms are crossed, and χ is the attenuation multiple. The switch efficiency is given by:Where and is also measured with two Glan prisms aligned, but with RPC activated before the laser pulse arriving.4. Experimental results
For discussion of the experimental results, we define the system extinction ERsys, which is the ratio of the detected intensity with the polarizer aligned to the intensity with the polarizer crossed without activating the RPC. ERsys is a measure of depolarization errors in the various optical components. With the RPC replaced by a mirror, we find that ERsys > 1.0 × 104. Before testing the performance of the RPC, we explored the operating point of the RPC, including the optimal pressure in the discharge chamber and the required V sw. In Fig. 4 , we present a neon gas discharging photograph obtained with a CCD. With the air pressure 3500 Pa at far end of the inlet and 12.5 Pa at far end of the outlet, neon gas breakdowns stably while the output voltage of the power supply greater than 11 kV. The discharge chamber is filled with plasma in the whole clear aperture. All the following results were obtained at this pressure. This pressure corresponds to the best efficiency in view of smallest breakdown flutter and lowest breakdown voltage. Figure 5 shows the chopped wave as well as the switch-pulse voltage wave during the gas discharging. The blue curve exhibits the typical voltage shape of one pulse driven RPC. When the switch pulse voltage drops to about −12.8 kV, neon gas in the discharge chamber begins to breakdown, and then the voltage rises to −5 kV before reaching a flat bottom constant value of −5.6 kV lasting roughly 200 ns. 3.0 kV voltage is applied across the KD*P, which rotates by 90° the polarization of the incident linearly polarized laser beam. The green curve presents the chopped wave from a 1064 nm CW-laser.
By adjusting the output voltage of the power supply, the dependence of switch efficiency on switch-pulse voltage is measured, which is shown in Fig. 6 . The Vsw, corresponding to the maximal switch efficiency, is 5.6 kV, slightly greater than the theory value 5.4 kV. All the following results are measured with Vsw = 5.6 kV.
To estimate the performance of the RPC, both the static extinction ratio and switch efficiency are measured at six typical points located in the clear aperture, shown in Table 1 . The lowest extinction ratio is equal to 971:1 appearing at the corner of the KD*P, probably induced by residual stress in the KD*P. The switching efficiency of the RPC, measured by employing the Nd:YAG pulse laser, is better than 99.6% in the whole clear aperture.
Table 1. The static extinction ratio and switch efficiency at six typical points
5. Thermo-optic effects in the RPC
For high average power application, temperature in the electro-optic crystal will rise because of its laser absorption. If the absorption coefficient is α, the volume heat productivity is
Here, I(x,y,z) is power density of the laser beam and z is parallel to the optical axis. For the longitudinally conduction-cooled RPC, we study the thermo-effects by finite-element method. The incident laser flux is 5 J/cm2 with repetition rate 10 Hz, in wavelength 1064 nm, and the beam dimension 30mm × 30mm. For 4 passes amplifier systems, each laser pulse makes through the Pockels cell for 4 times, so the Pockels cell is required to handle up to 200 W/cm2. The absorption coefficient of 98% deuterated KD*P is 0.002 cm−1, leading to an average power of 0.4 W/cm2 on the KD*P 3cm × 3cm clear aperture. The heat transfer coefficient between the copper-substrate and the KD*P is a key factor in view of cooling efficiency. Shigeke Tokita measured the heat transfer coefficient between the copper-substrate and the sapphire, which is larger than 4 W/cm2K [9]. Here, the heat transfer coefficient supposes to be 0.5 W/cm2K, to say the least. The initial temperature is 20 °C, and the copper heat-sink is water-cooled. Figure 7 . shows the steady-state temperature distribution in KD*P and copper heat-sink. By longitudinally face-cooling, with high efficiency, both the temperature rise and the transverse temperature gradients are small.Adopting thermo-elasticity model, the thermo-stress distribution is numerically calculated. End-face deformation will induce wave-front distortion. Furthermore, the stress birefringence will cause depolarization due to elastic-optical effect [10]. The depolarization and wave-front distribution at thermal balance state is given respectively in Fig. 8 and Fig. 9 . The maximum depolarization decreases to 0.82%, and the average value for the whole beam area is 0.13%. For wave-front distortion, the PV (peak-to-valley) value is 0.24 λ (λ = 1064 nm). From section 2, we can see that these perturbations are acceptable to ensure the laser performance.
6. Conclusion
We have described in detail the design and performance of a reflecting plasma-electrode Pockels cell. By adopting plasma-electrode technology makes the RPC available in large clear aperture. On the other side, designing the Pockels cell reflecting, one face of the electro-optic crystal is released from the low pressure environment. Therefore the crystal can be longitudinally face-cooled by the copper-substrate with high efficiency to manage thermo-effects. The measured results indicate: the static extinction ratio is greater than or equal to 971:1 locally in the clear aperture, and the dynamic switching efficiency is greater than 99.6%. The simulation results show: For 200 W/cm2 laser power densities, the average depolarization in the RPC is 0.13%, and the PV value is 0.24 λ. These perturbations are acceptable to ensure the laser performance according to section 2. We can see from the said above that the RPC is a potential Pockels cell for isolating and controlling number of pass in high average-power amplifier systems with a large clear aperture.
References and links
1. J. Caird, A. Bayramian, et al., “Mercury: A high repetition rate laser for high energy density physics, ” 29th European Conference on Laser Interaction with Matter, Madrid, Spain, UCRL-PRES-221983 (2006).
2. S. Z. Kurtev, O. E. Denchev, and S. D. Savov, “Effects of thermally induced birefringence in high-output-power electro-optically Q-switched Nd:YAG lasers and their compensation,” Appl. Opt. 32(3), 278 (1993). [CrossRef] [PubMed]
3. L. F. Weaver, C. S. Petty, and D. Eimerl, “Multikilowatt Pockels cell for high average power laser systems,” J. Appl. Phys. 68(6), 2589–2598 (1990). [CrossRef]
4. J. Goldhar and M. A. Henesian, “Electro-optical switches with plasma electrodes,” Opt. Lett. 9(3), 73–75 (1984). [CrossRef] [PubMed]
5. G. Gardelle and E. Pasini, “A simple operation of a plasma electrode Pockels cell for the laser Megajouls,” J. Appl. Phys. 91(5), 2631–2636 (2002). [CrossRef]
6. B. E. Kruschwitz, J. H. Kelly, M. J. Shoup Iii, L. J. Waxer, E. C. Cost, E. T. Green, Z. M. Hoyt, J. Taniguchi, and T. W. Walker, “High-contrast plasma-electrode Pockels cell,” Appl. Opt. 46(8), 1326–1332 (2007). [CrossRef] [PubMed]
7. X. Zhou, G. Wenqiong, Z. Xiongjun, S. Zhan, and W. Dengsheng, “One-dimensional model of a plasma-electrode optical switch driven by one-pulse process,” Opt. Express 14(7), 2880–2887 (2006). [CrossRef] [PubMed]
8. X. Zhang, D. Wu, J. Zhang, H. Yu, J. Zheng, D. Cao, and M. Li, “One-pulse driven plasma Pockels cell with DKDP crystal for repetition-rate application,” Opt. Express 17(19), 17164–17169 (2009). [CrossRef] [PubMed]
9. S. Tokita, J. Kawanaka, and Y. Izawa, “Sapphire cooling at both faces of high-power cryogenic Yb:YAG disk laser,” 2nd International Workshop on High Energy Class Diode Pumped Solid State Lasers, Jena, Germany, 10–12 June (2005).
10. D. Eimerl, “High average power harmonic generation,” IEEE J. Quantum Electron. 23(5), 575–592 (1987). [CrossRef]
