Open Access
Add to BibSonomy
Abstract
We demonstrated stable intermittent operation of a 266-nm picosecond pulsed light source with an average power of 20 W. The 266-nm beam, which had a maximum average power of 35.5 W, was generated by frequency conversion of a 1064-nm laser with an LiB3O5 crystal and a CsLiB6O10 (CLBO) crystal. The 1064-nm laser had a repetition rate of 600 kHz and an average power of 130 W and was capable of intermittent operation with an acousto-optic modulator in the fundamental laser section. By investigating the crystal temperature rise caused by the 266-nm light absorption in the CLBO crystal, we found that the crystal temperature rise caused by nonlinear absorption must be suppressed to achieve stable intermittent operation. The countermeasures allowed stable-intermittent operation at an average power of 20 W to be achieved, with a response time of 1.1 s for the 10%–90% rise conditions and a stability of 2%p-p for the average power fluctuation from 2 to 120 s. These results show that deep-ultraviolet picosecond pulses with an average power of 20 W can be used for industrial applications that require stable intermittent operation.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Deep-ultraviolet (DUV) beams have high photon energy, high absorption for many materials, and high scattering intensity. Because of these characteristics, they are used for laser processing of silicon, glass, and organic materials [1–3], high-brilliance γ-ray generation [4], and as light sources for inspection [5]. DUV solid-state lasers, which consist of a solid-state laser and a frequency converter, can achieve high beam quality and can be focused into a few micrometers by using lenses. This feature makes DUV solid-state lasers suitable for micro processing applications. Some laser machining processes, such as laser drilling of high density printed circuit boards require stable ON/OFF operation of laser pulses (intermittent operation) and intensity modulation in addition to high average power and long-term stability [6–8]. Although there are many studies of the output power and reliability, no studies of intermittent operation have reported. In this paper, we investigated the problems in achieving intermittent operation of high-power DUV solid-state lasers and propose countermeasures.
The development of the CsLiB6O10 crystal (CLBO) [9] in 1995 DUV solid-state lasers with higher output power have been developed, and by 2003, an average power of up to 42 W had been reported using nanosecond pulsed lasers [10–12]. However, the long-term stability was still insufficient, and an average power of 20 W could only be maintained for about 100 h [12]. In 2013, we developed a 266-nm light source with an average power of 4.5 W using a picosecond pulsed laser featuring a sufficiently narrower linewidth compared with a spectral acceptance bandwidth of a CLBO crystal to solve long term power degradation [13]. Subsequently, reports began to exceed the average power of 1 W in the picosecond pulse regime, up to an average power of 7.6 W by 2019 [14–16], and in 2020 we reported stable generation for 5000 h at an average power of 10 W using a large-aperture CLBO crystal [17]. From 2020 to 2021, an average power of 20 W was reported with laser processing in mind [3,18]. In 2022, we reported stable generation for 10,000 h at an average power of 20 W, showing that high-power and long-term stability [19].
To realize intermittent operation, it is necessary to use intensity modulators, such as electro-optic modulators (EOMs) and acousto-optic modulators (AOMs) [20]. Potassium dihydrogen phosphate and beta-barium borate crystals are used as materials for EOM elements, whereas quartz glass and quartz crystals are used for AOM elements. Because reliable intensity modulators are commercially available only in the near infrared to ultraviolet regions, if the laser wavelength is in the DUV region, there are no materials and antireflective (AR) coatings that resist the degradation caused by high-power DUV light during stable long-term operation, the generated DUV light needs to be controlled via intermittent operation or intensity modulation of the input fundamental laser. However, during high-power DUV generation, the crystal temperature shifts from the optimum operating point due to self-heating effect caused by absorption of the DUV beam generated in the nonlinear optical crystal. If the laser wavelength after frequency conversion is in the DUV range below 300 nm, linear and nonlinear absorption [21–23] of the DUV light generated in the nonlinear optical crystal becomes non-negligible. This causes temperature fluctuations in response to the DUV power generated in the nonlinear optical crystal, and the DUV power responds with a delay relative to the fundamental because the conversion efficiency depends on these temperature fluctuations. Thus, it is difficult to control DUV pulse trains just by adjusting the amplitude of the input pulse trains. Therefore, self-heating effect must be suppressed to shorten the response time and achieve stable intermittent operation.
In this study, we investigated a method to suppress self-heating effect with the aim of achieving stable intermittent operation. In addition, we have been developing a narrow-linewidth, high-peak-power fundamental laser and a frequency converter to address the issue of long-term stability of DUV light [13,17,19]. In the present study, we extended our previous work [19], which briefly mentioned response time, and discuss it in the details.
First, we investigated the dependence of the CLBO crystal holder temperature on the output power at a wavelength of 266 nm using the input beam diameter and input power to the CLBO crystal as parameters to determine the degree of self-heating effect. Then, the change in crystal holder temperature and temperature acceptance bandwidth for the power at 266 nm were investigated. Next, we examined the response time and stability during intermittent operation. In addition, we measured the output end surface temperature of a CLBO crystal using a thermal imager to confirm the temperature rise that depends on the beam diameter and peak power density. The temperature rise was suppressed by increasing the beam diameter and decreasing the power density at 266 nm in the crystal. Finally, we discussed the decrease in efficiency at high power and confirmed that the effective length contributing to frequency conversion decreased as the temperature increased. Furthermore, the power density for intermittent operation was lower than that for long-term stability, and that long-term stability could be simultaneously achieved by using a CLBO crystal under operating conditions that allow intermittent operation. To the best of our knowledge, we have realized the first DUV solid-state laser with both stable intermittent operation and long-term stability at an average power of 20 W.
2. Experimental setup
Figure 1(a) shows the experimental setup for DUV generation at a wavelength of 266 nm, consisting of a narrow-linewidth, high-peak-power fundamental laser source section and a frequency conversion section developed in a previous study [19].
Fig. 1. Configuration of the experimental setup for measuring (a) DUV power and (b) temperature distribution. AOM: acousto-optic modulator; M1: HR 1064 nm mirror; A1: water-cooled aperture; DM1: HR 532/AR 1064 nm dichroic mirror; L1–L3: lenses; HWP: half waveplate; PBS: polarization beam splitter; ATT: optical attenuator; W1: AR 532 nm window; DM2: HR 532/AR 266 nm dichroic mirror; DM3: HR 266/AR 532 nm dichroic mirror; BW: CaF2 Brewster window; CDA: clean dry air; PD: photodiode.
The fundamental laser source used a distributed feedback (DFB) semiconductor laser as the seed source. The laser source generated an average power of 150 W and a pulse duration of 14 ps with a linewidth of 0.15 nm at a repetition rate of 600 kHz by multi-stage amplification using fiber amplifiers and Nd:YVO4 amplifiers. A water-cooled AOM with an active aperture of 6.5 mm diameter was installed after the final amplifier to generate a first-order diffraction beam with an average power of 130 W, which could be synchronized with the seed laser pulse to enable ON/OFF control (intermittent operation) of picosecond pulse trains. The AOM device was made of a quartz crystal with a 1064 nm AR coating. The diffracted beam was launched into a temperature controlled LiB3O5 (LBO) crystal at 56 °C with an aperture of 6 × 6 mm and a length of 20 mm (θ = 90.0°, φ = 10.4°, critically phase-matched) and a beam diameter of 2.0 mm to generate a 532-nm beam with an average power of 90.2 W, a pulse duration of 11 ps, and a linewidth of 0.042 nm.
A beam expander consisting of three plano-concave, plano-convex lenses (L1–L3) were used to adjust and collimate the beam diameter of the 532-nm beam. The power of the 532-nm beam was adjusted by an optical attenuator consisting of a half waveplate and a polarizing beam splitter (PBS), and then launched into a CLBO crystal with an aperture of 16 × 16 mm and a length of 15 mm (θ = 62.0°, φ = 45.0°, critically phase-matched) to generate a 266-nm beam. The CLBO crystal was wrapped in a 0.2-mm-thick TIN sheet and held in a copper holder with a built-in heater and temperature sensor. The CLBO crystal was placed in a chamber purged with clean dry air and was temperature controlled at 150 °C. Furthermore, the phase matching angle was adjusted with a 266-nm power at 0.1W. The 266-nm beam was reflected by two dichroic mirrors to remove the residual 532-nm beam, and the average power was measured with a powermeter. The powermeter was a thermopile sensor (F150A-BB-26, Ophir) with a specified 0%-95% rise time of 1.5 s. The CLBO crystal was wedged with 1° at both ends to suppress nearly coaxial reflected light from the surfaces and the AR coating (532 and 266 nm) was applied only on the input surface. The output surface was uncoated to ensure stable long-term generation for more than 10,000 h. The input beam diameter on the CLBO crystal was adjusted by the beam expander (L1–L3) to 4.2, 6.0, and 8.0 mm. The upper limit of the collimated beam diameter was 8.0 mm, which was half of the crystal aperture, to prevent direct heating of the CLBO crystal holder by the edge component of the Gaussian beam. The lower limit was a diameter of 4.2 mm, which was limited by the lens system. The generated 266-nm pulses had diameters of 3.1, 5.0, and 6.4 mm, respectively, and a pulse duration of 8 ps. The power was sampled at 0.1 s intervals when the 266-nm pulse train was intermittently operated with the fundamental laser pulse train switched on and off by the AOM installed in the fundamental laser. The pulse trains were also observed by a photodiode (Model 1621, New Focus) with a rise time of 1 ns and an oscilloscope with a sampling rate of 1 GS/s. An ultraviolet transmissive and visible range absorption filter (UL-340, OMG) of 2 mm thickness was placed just before the photodiode. For the 1064-nm laser, no response delay or output power instability during intermittent operation was observed. The 532-nm beam also achieved the same intermittent operation following the 1064-nm laser owing to a water-cooled aperture just before the LBO crystal, and the 10%–97% rise time was 1.5 s, as described below.
Figure 1(b) shows the experimental setup for measuring the surface temperature of the CLBO crystal. To observe the temperature distribution on the output end surface and side surface during the generation of 266-nm light, the CLBO crystal was removed from the CLBO chamber and the crystal was placed on the rotating stage. A half waveplate was added to the 532-nm beam passing through the PBS in Fig. 1(a) to make the polarization of the 532-nm beam vertical, and only the phase-matching angle was adjusted. A 266-nm beam was generated at room temperature by adjusting phase-matching angular direction θ with the rotating stage. Temperature distributions at the output end and side surfaces during generation of the 266-nm beam was measured using a thermal imager (885, Testo) with a temperature resolution of 0.03 °C. The dumper and powermeter were placed at least 1 m away from the CLBO crystal so that the heat that they generated in the latter stage would not affect the temperature measurement on the CLBO crystal.
3. Experimental result and discussion
3.1. Frequency conversion characteristics
First, the dependence of the 266-nm power on the CLBO crystal holder temperature was investigated to clarify the heat generation in the CLBO crystal. Frequency conversion using nonlinear optical crystals requires the phase-matching conditions to be satisfied, and the conversion efficiency is highly sensitive to angle, wavelength, and temperature. In this experiment, the angle and wavelength were set as constant, and only the temperature dependence was observed. A 532-nm beam was launched into the CLBO crystal, and the crystal holder temperature was decreased from 152 to 144 °C at 0.5 °C intervals on the hot side while the 266-nm light was generated. The reason for starting from the hot side was to obtain the maximum power when the heat generation by absorption of the 266-nm beam is strong [24], and the maximum power is not obtained if stared from the cold side. At each set temperature, the power was recorded after waiting for 6 min until the temperature inside the CLBO crystal was sufficiently stable and the 266-nm power fluctuation less than 0.5%/min. The average power of the 532-nm beam launched into the CLBO crystal was varied from 22.2 to 90.2 W using an optical attenuator consisting of a waveplate and PBS, and the power of the 266-nm beam generated was measured. In addition, measurements were held for input beam diameters of 4.2, 6.0, and 8.0 mm. The holder temperature and attenuator were programmed for automatic measurement of 266-nm average power. The results are shown in Fig. 2.
Fig. 2. Characteristics of frequency conversion with input beam diameter of (a) 4.2, (b) 6.0, (c) and 8.0, and (d) dependence of the maximum average 266-nm power on input beam diameter.
Figures 2(a), (b), and (c) show the crystal holder temperature dependence of the average 266-nm power for input beam diameters of 4.2, 6.0, and 8.0 mm, respectively. The phase matching temperature at low power differs for each beam diameter because the beam expander adjustment has changed the optical axis. For an input beam diameter of 4.2 mm, the holder temperature at which the maximum 266-nm power was reached shifted to the low-temperature side as the input power increased. With an input power of 22.2 W, the maximum power was 6.2 W at a holder temperature of 150 °C. With an input power of 90.2 W, the maximum output power was 34.6 W at 145.0 °C. After the automatic measurement, we re-measured manually around 145 °C. At 145.0 °C, 34.5 W of 266-nm power was obtained, but at 144.8 °C, the 266-nm power decreased rapidly to 3 W. This phenomenon is caused by the breaking down of the phase matching condition in the CLBO crystal, which was maintained at the temperature generated by the sum of the heat from the heater and the heat from the absorption of the 266-nm beam, as shown in our previous work [24]. The dependence of the CLBO crystal holder temperature on the 266-nm power followed the Sinc function [20,25] and was generally symmetric. A curve along high symmetry obtained for all input powers for a beam diameter of 8.0 mm, although a slight asymmetry was observed for the beam diameter of 6.0 mm, and a large asymmetry for the beam diameter of 4.2 mm. These observations indicated that self-heating effect occurred in the CLBO crystal. In addition, because the absorption of the 532-nm beam in the CLBO crystal was negligible, the effect of self-heating effect due to the absorption of the 266-nm light that occurred during the frequency conversion process was observed [24].
Figure 2(d) shows the beam diameter dependence of the maximum 266-nm power for each input power. The 266-nm power increased with decreasing beam diameter for 532 nm input powers of 35.2 and 62.4 W. However, for the input power of 90.2 W, the 266-nm power reached a maximum at an input beam diameter of 6.0 mm and decreased at 4.2 mm. The 266-nm power and conversion efficiencies were 35.5 W and 39.4% at a diameter of 6.0 mm and 34.6 W and 38.4% at a diameter of 4.2 mm, respectively.
Figure 3(a) shows the dependence of the crystal holder temperature shift on the power at 266 nm where the power reached its maximum. The temperature shift was based on the temperature at which the maximum 266-nm power was reached at the minimum input for each beam diameter. The smaller the input beam diameter was, the larger the shift in holder temperature for phase matching. Figure 3(b) shows the dependence of the temperature acceptance bandwidth (half width at half maximum [HWHM]) on 266-nm power on the low-temperature side (left side). The bandwidth decreased as the 266-nm power increased, and the rate of change was larger for smaller input beam diameters. The increase in the rate of change was an increase in self-heating effect due to the increase in absorbed power density that contributed to heat generation.
Fig. 3. (a) Temperature difference as a function of the average 266-nm power and (b) thermal bandwidth of the left side (HWHM) as a function of average 266-nm power.
Figure 3(b) shows that the temperature acceptance bandwidth (HWHM) was 2.2 °C at an average power of 5 W for an input beam diameter of 8.0 mm. The value calculated from the Sellmeier equation and the temperature coefficient for CLBO crystal [26] was 3.2 °C cm (HWHM), which was 2.1 °C for a length of 15 mm, in good agreement with the observed value. The temperature acceptance bandwidth of the phase matching was defined as the temperature at which the 266-nm power was halved. To keep the power fluctuation due to crystal temperature change within 3%, which is required for laser processing, it is necessary to keep the temperature change within approximately a fifth of the temperature acceptance bandwidth (HWHM) [20,26]. Therefore, the power at which the temperature shift of the CLBO crystal holder due to self-heating effect was less than 0.44 °C is the power that can be used for practical applications. Figure 3(a) shows that the practical 266-nm powers were about 10, 17, and 20 W for input beam diameters of 4.2, 6.0, and 8.0 mm, respectively, and a diameter of 8.0 mm, which had the lowest maximum power, would be suitable.
Because the temperature acceptance bandwidth depends on the crystal length, the size of the allowable temperature shift range can be increased by shortening the crystal length. However, conversion efficiency decreases by shortening the crystal length, it is necessary to reduce the input beam diameter to increase the power density in the crystal so that conversion efficiency drop is compensated. Consequently, the temperature shift due to self-heating effect becomes much larger than the increased the allowable temperature shift range by shortening the crystal length, and the practical power cannot be increased [24]. Therefore, in this study, the length of the CLBO crystal was set to 15 mm. However, the use of such a long crystal results in a narrower spectral acceptance bandwidth, 0.085 nm for a 15-mm-long CLBO crystal [26]. Therefore, the linewidth of 532-nm light must be less than half of the bandwidth to achieve efficient conversion.
3.2. Rising characteristics of 266-nm power during intermittent operation
The rising characteristics were measured to determine the response time of the 266-nm power. The attenuator placed in 532 nm section was adjusted so that the 266-nm power was set to 20 W for beam diameters of 4.2, 6.0, and 8.0 mm, and the 532 nm input power and CLBO crystal holder temperature were set to 49.1 W and 148.7 °C, 62.0 W and 149.0 °C, and 77.9 W and 149.5 °C, respectively as shown in Table 1.
Table 1. Condition of intermittent operation
The crystal holder temperature was set to the temperature at which the time required to reach 97% from 10% of the specified 266 power (response time) is the shortest for each beam diameter, not to the temperature at which maximum power can be obtained as shown in Fig. 2. Under each condition, the ON/OFF time of the fundamental laser pulse train was controlled by the AOM installed in the fundamental laser section, and the 266-nm pulse was operated intermittently. Because the temperature response time was on the order of seconds, power measurements were sampled at 0.1 s intervals using a powermeter with a specified 0%-95% rise time of 1.5 s. The actual measured time using the fundamental are shown in Table 2. The OFF and ON periods were both set as 120 s. Figure 4 shows the results of the measurements.
Fig. 4. Rising characteristics of average 266-nm power with input beam diameters of (a) 4.2, (b) 6.0, and (c) 8.0 mm.
Figure 4 shows that intermittent operation was realized at 20 W for each beam diameters, but that the rise behavior varied. The peak-to-peak stabilities of the average power from 2 to 120 s after turning on the fundamental pulse train were 20, 6, and 2%p-p, respectively. Because laser processing applications require a stability of less than 3%p-p, the only characteristic that met the requirement was at a diameter of 8.0 mm. Figure 5(a) shows an enlarged view of normalized by the steady-state power of 120 s after the first rise of each beam diameter. The enlarged figure in Fig. 5(a) shows that there was a large difference in response for the 266-nm power. For input beam diameters of 4.2, 6.0, and 8.0 mm, the beam diameters of the 266-nm beam generated were 3.1, 5.0, and 6.4 mm, respectively, and thus the peak power density and average power density of the 266-nm beam at the output end surface of the CLBO crystal were 112 MW/cm2 and 537 W/cm2, 42.5 MW/cm2 and 204 W/cm2, and 25.9 MW/cm2 and 124 W/cm2, respectively as shown in Table 2.
Fig. 5. (a) Close-up of rising characteristics of average 266-nm power and (b) output waveform of photodiode with an input beam diameter of 8.0 mm.
Table 2. Measured rise time for each wavelength and input beam diameter
The rise time at 1064-nm wavelength shows the characteristics of the powermeter. The rise time at 532-nm wavelength is the same as that at 1064 nm, indicating that no self-heating effect occurs in the frequency conversion process to a 532 nm. For the 266-nm wavelength the 10%-90% rise times at each input beam diameter were 3.5, 1.3, and 1.1 s, respectively. The 10%–97% rise times to within 3% of the steady-state power were 7.7, 3.8, and 1.6 s, respectively. The 10%-90% and 10%-97% rise times of the both 1064-nm and 532-nm beams were 1.1 and 1.5 s, indicating that the rise times for the 266-nm beam at input diameters of 4.2 and 6.0 mm were longer than the input beam by the self-heating effect. For an input beam diameter of 8.0 mm, each rise time are 1.1 and 1.6 s as almost same as input beam, results in due to the suppression of the self-heating effect. Figure 5(b) shows the results of the photodiode observation of the pulse train at the first rise with a beam diameter of 8.0 mm. The crest values of the 266-nm pulses equivalent to the steady-state value were obtained immediately after the pulse generation.
For input beam diameters of 4.2 and 6.0 mm, the conversion efficiencies from 532 to 266 nm were 40.7% and 32.3%, respectively, enabling the generation of a 266-nm light with high efficiency. However, even with further adjustment of the CLBO crystal holder temperature, it was difficult to reduce the rise time. For an input beam diameter of 8.0 mm, the peak and average power densities were 25.9 MW/cm2 and 124 W/cm2, respectively, and although the conversion efficiency was only 25.7%, suitable rise time characteristics were achieved and the crest values of the 266-nm pulses equivalent to the steady-state value were obtained immediately after the pulse generation. This peak power density was lower than the power densities of 45 [27] and 34 MW/cm2 [19] required for long-term stable operation. In other words, in our configuration, the threshold for unstable operation due to self-heating effect was lower than the DUV-induced damage threshold of the CLBO crystal, and the DUV light that achieved intermittent operation also simultaneously satisfied long-term stability.
3.3. Temperature rises due to absorption in the CLBO crystal
To observe the crystal temperature-rise that occurred in the CLBO crystal directly, the output end surface of the CLBO crystal was measured with the thermal imager in the setup shown in Fig. 1(b). During the generation of the 266-nm beam, the phase-matching angle was repeatedly adjusted by the rotation stage to maximize the 266-nm power, and the input power was repeatedly adjusted by the attenuator to obtain a 266-nm power of 20 W. The phase-matching angle was adjusted at intervals of about 2 seconds for 60 s after 266 nm generation and at intervals of about 30 seconds after 60 s. Figure 6 shows the output end surface temperature distributions 120 s after the generation of the 266-nm beam with input beam diameters of 4.2 and 8.0 mm.
Fig. 6. Temperature distribution on the output end surface at 20 W for the 266 nm beam with input beam diameters of (a) 4.2 and (b) 8.0 mm.
Heat generation and temperature distribution were observed for input beam diameters of 4.2 and 8.0 mm. For the 4.2-mm beam diameter, the maximum temperature was 29.4 °C and the temperature difference between the center and edge was approximately 2.8 °C. For the 8.0-mm beam diameter, the maximum temperature was 26.5 °C and the temperature difference was approximately 0.7 °C. For the 4.2-mm beam diameter, the self-heating effect was higher than for the 8.0-mm beam diameter, despite the average 266-nm power remaining constant. The 532-nm beam had an input power of 90.2 W with a beam diameter of 4.2 mm, and the polarization was rotated by 90° and the 266-nm light was not generated. The temperature difference at the center and edge of the CLBO crystal was less than 0.1 °C, confirming that the absorption at 532 nm was sufficiently small.
Then, we measured the temperature difference on the output surface at 20 W for the 266-nm beam as a function of peak power density at an input beam diameter of 4.2 mm. The oscillator repetition rate was varied from 600 to 1700 kHz, and the 532-nm input power was adjusted using an attenuator to keep the 266-nm power constant at 20 W. No major change in the input beam diameter was observed under these conditions, and the beam diameter was treated as constant. The average power density of the 266-nm beam was constant at 537 W/cm2 and the peak power density varied from 38 to 112 MW/cm2. The measurement results are shown in Fig. 7.
Fig. 7. Temperature difference on the output surface at 20 W for the 266-nm beam as a function of peak power density.
Figure 7 shows that the temperature difference at the output end surface of the CLBO crystal was 1.6 °C at a peak power density of 38 MW/cm2 and 3.0 °C at 112 MW/cm2, even though the average 266-nm power at the output end surface was 20 W and the average power density was constant. The slope of the linear approximation in this measurement range was 0.0183 °C/(MW/cm2), indicating a temperature change of 1.8 °C at 100 MW/cm2. The higher peak power density of the 266-nm beam increased the temperature of the CLBO crystal output end surface, confirming that nonlinear absorption in addition to linear absorption in the CLBO crystal caused self-heating effect in the CLBO crystal. Furthermore, the temperature rise due to nonlinear absorption was not negligible compared with the temperature acceptance bandwidth of 2.1 °C for CLBO crystals. Therefore, it is necessary to suppress the self-heating effect that occurs in the CLBO crystal to achieve stable intermittent operation, and it is important to convert the frequency at a low peak power density for the 266-nm beam.
Next, we discuss the cause of the lower conversion efficiency for a beam diameter of 4.2 mm in Fig. 2(d). Because we expected the temperature gradient in the longitudinal direction (propagation direction) to be the cause, we observed the CLBO crystal from the side surface using the thermal imager during the generation of the 266-nm beam in the setup shown in Fig. 1(b). Figure 8 shows the temperature distribution of the CLBO crystal 120 s after the generation of a 266-nm light with an average power of 35W. The input beam diameter of the 532-nm beam was set to 4.2 mm and an average power of 90.2 W was propagated from left to right into the CLBO crystal. The 532 nm input beam was launched into 4 mm away from the side surface of the CLBO crystal.
Fig. 8. Temperature distribution along the length of a CLBO crystal after 120 s from the generation of a 266 nm beam with an average power of 35 W and peak power density of 196 MW/cm2. (b) Temperature distribution along at the center height of the CLBO crystal.
The crystal temperature increased in the output direction and reached a maximum near the output end surface. The phase-matching angle was fine-tuned continuously until the temperature reached a steady state to maintain the power of 35W of the 266-nm beam. The 266-nm peak power density at the output end surface was calculated to be 196 MW/cm2.
The temperature rise in the crystal was caused by linear and nonlinear absorption of the 266-nm beam. Therefore, the higher the 266-nm power, the higher the temperature. Figure 8(b) shows that the temperature saturates about 5 mm before the output end. In other words, the 266 nm output was also saturated. Therefore, this result means only 10 mm of the 15-mm-long CLBO crystal contributed to the frequency conversion. As the beam diameter was reduced, the power at 266 nm grows from the beginning of the crystal length due to the increasing in peak power density, on the other hand, nonlinear absorption also increased and resulted in inducing strong self-heating effect (temperature rise). Therefore, toward the end of the crystal length, where the 266-nm power was higher, the crystal temperature increased above the temperature acceptance bandwidth of the CLBO crystal, and frequency conversion did not proceed. As a result, the effective crystal length was shortened, resulting in lower conversion efficiency. This effect would have resulted in a lower 266-nm power with an input of 90.2 W and a beam diameter of 4.2 mm compared with a beam diameter of 6.0 mm.
Several methods of controlling the temperature gradient along the length of the crystal has been demonstrated [28,29], and those methods were expected to improve the conversion efficiency. However, the temperature rise itself must be suppressed to achieve stable intermittent operation without response delay. The technical difficulty is particularly challenging in the high-power range because the absolute absorbed power becomes large. It is difficult to reduce the absorption coefficient of nonlinear optical crystals dramatically from the current state. Therefore, to achieve stable intermittent operation, nonlinear optical crystals require frequency conversion with a lower peak power density.
To solve this problem, we developed a 532-nm beam with a narrow linewidth of 0.042 nm which sufficiently narrower linewidth compared with a spectral acceptance bandwidth of a CLBO crystal, a high peak power of 13.7 MW, and an average power of 90.2 W by using a DFB semiconductor laser as the seed laser, a solid-state amplifier as the main amplifier, and an LBO crystal. As shown in Section 3.1, this laser was collimated to 8.0 mm in diameter and launched into a CLBO crystal with an aperture of 16 mm and length of 15 mm to generate a 25.9 W, 266-nm beam. Furthermore, as shown in Section 3.2 the average power was then set to 20 W by setting the input power to 77.9 W, the peak power density at the output end surface of the CLBO crystal to 25.9 MW/cm2, and the average power density to 124 W/cm2. Consequently, stable intermittent operation at a practical conversion efficiency and stable operation for 10,000 h were both achieved. For a DUV light achieves intermittent operation through the fundamental laser, only the required number of pulses need to be generated at the required timing, which prevents consumption of the CLBO crystal and other DUV optical components, and further long-term stable operation is expected.
4. Conclusions
We demonstrated stable intermittent operation of a 266-nm picosecond pulsed light source with an average power of 20 W. A DFB semiconductor laser was used as a picosecond pulsed seed source to generate a 1064-nm laser operating at a 600 kHz repetition rate, 130 W average power, 0.15 nm linewidth, 14 ps pulse duration, and 15.5 MW peak power with an AOM in the fundamental amplifier. Frequency conversion using an LBO and a CLBO crystal generated a 532-nm beam with an average power of 90.2 W at a conversion efficiency of 69.4% and a 266-nm beam with an average power of 35.5 W at a conversion efficiency of 39.4%, respectively. Furthermore, the effect of the power density of the 266-nm beam generated in the CLBO crystal on the frequency conversion characteristics and response time during intermittent operation was investigated. To address this issue, the input beam diameter of the 532-nm beam on the CLBO crystal was increased to 8.0 mm, and stable intermittent operation at an average power of 20 W was achieved by setting the peak power density at 266 nm and the average power density on the output end surface of the crystal to 25.9 MW/cm2 and 124 W/cm2, respectively. A response time of 1.1 s, which was equal to the response time of the input laser for the 10%–90% rise, and a pulse crest value equal to the steady-state value immediately after the pulse generation were realized. A 2%p-p stability of the average power from 2 to 120 s was obtained after turning on the fundamental pulse train. Furthermore, the temperature rise of the CLBO crystal was directly measured using a thermal imager, and it was clarified that it is important to suppress the temperature rise due to nonlinear absorption to achieve stable intermittent operation. The efficiency decrease that occurred at high peak power density was caused by decrease in the effective length of the CLBO crystal as the temperature increased due to self-heating effect. In addition, the peak power density of the 266-nm beam for stable intermittent operation was lower than the peak power density for long-term stability. This DUV light source, which enables intermittent operation, only generates the necessary number of pulses at the necessary timing, preventing consumption of the CLBO crystal and other DUV optical components, and allowing further long-term stable operation. These results show that DUV picosecond pulsed light source with an average power of 20 W can be used for industrial applications that require stable intermittent operation.
Funding
New Energy and Industrial Technology Development Organization (P16011).
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. Y. Imamiya, S. Akama, Y. Fujita, and H. Niitani, “Development of Microfabrication Technology using DUV Laser,” Mitsubishi Heavy Industries Tech. Rev. 53(4), 49–54 (2016).
2. Y. Kawasuji, Y. Adachi, A. Suwa, J. Fujimoto, K. Kakizaki, and M. Washio, “Pulse Width Dependence of Ablation Threshold and Ablation Rate,” Micro Session 1 of Laser Material Microprocessing of ICALEO2020, 0442_0638_000138, (2020).
3. S. Häfner, C. Wagner, B. Shnirman, M. Ginter, J. Brons, M. Sailer, A. Fehrenbacher, D. Grossmann, D. Flamm, K. Janami, S. Ruebling, U. Quentin, A. Budnicki, I. Zawischa, and D. H. Sutter, “Deep UV for materials processing based on the industrial TruMicro Series of ultrafast solid-state laser amplifiers,” Proc. SPIE 11670, 49 (2021). [CrossRef]
4. N. Muramatsu, Y. Kon, S. Daté, Y. Ohashi, H. Akimune, J. Y. Chen, M. Fujiwara, S. Hasegawa, T. Hotta, T. Ishikawa, T. Iwata, Y. Kato, H. Kohri, T. Matsumura, T. Mibe, Y. Miyachi, Y. Morino, T. Nakano, Y. Nakatsugawa, H. Ohkuma, T. Ohta, M. Oka, T. Sawada, A. Wakai, K. Yonehara, C. J. Yoon, T. Yorita, M. Yosoi, and LEPS Collaboration, “Development of high intensity laser-electron photon beams up to 2.9 GeV at the SPring-8 LEPS beamline,” Nucl. Instrum. Methods Phys. Res., Sect. A 737, 184–194 (2014). [CrossRef]
5. H. Nakao, M. Morita, Y. Kaneda, A. Miyamoto, T. Tago, T. Sasa, M. Sasaura, and Y. Furukawa, “High power 4th harmonic generation with optimized enhancement cavity,” in 2017 European Conference on Lasers and Electro-Optics and European Quantum Electronics Conference, (Optical Society of America, 2017), paper CA_P_16.
6. K. Deki, T. Yokota, J. Sakuma, and Y. Osako, “Thermally stable third harmonic generation using type 1 LBO,” Tech. Digest of CLEO 2000, CTuA16, (2000).
7. H. Ikarashi, T. Sato, S. Teraki, M. Yoshida, and H. Ozaki, “Low Dk / Df Dielectric Material for 5 G Applications,” ICEP2022, TE3-4 (2022).
8. M. Koh, K. Yoneda, K. Nakada, S. Sekiguchi, S. Mishima, N. Ishikawa, and T. Ogata, “Novel Thermosetting Low Dk/Df Film and Its Performance,” ICEP2022, TC2-2 (2022).
9. Y. Mori, I. Kuroda, S. Nakajima, T. Sasaki, and S. Nakai, “New nonlinear optical crystal: Cesium lithium borate,” Appl. Phys. Lett. 67(13), 1818–1820 (1995). [CrossRef]
10. T. Kojima, S. Konno, S. Fujikawa, K. Yasui, K. Yoshizawa, Y. Mori, T. Sasaki, M. Tanaka, and Y. Okada, “20-W ultraviolet-beam generation by fourth-harmonic generation of an all-solid-state laser,” Opt. Lett. 25(1), 58–60 (2000). [CrossRef]
11. M. Nishioka, S. Fukumoto, F. Kawamura, M. Yoshimura, Y. Mori, and T. Sasaki, “Improvement of laser-induced damage tolerance in CsLiB6O10 for high-power UV laser source,” in Conference on Lasers and Electro-Optics/Quantum Electronics and Laser Science Conference, Technical Digest (Optical Society of America, 2003), paper CTuF2.
12. T. Kojima, S. Konno, S. Fujikawa, K. Yasui, T. Kamimura, M. Yoshimura, Y. Mori, T. Sasaki, M. Tanaka, and Y. Okada, “100-hour operation of an all-solid-state 20-W 266-nm UV laser by using high-quality CLBO crystal,” ASSL, 68, Trends in Optics and Photonics Series, paper WC2 (2002).
13. Y. Orii, Y. Takushima, M. Yamagaki, A. Higashitani, S. Matsubara, S. Murayama, T. Manabe, I. Utsumi, D. Okuyama, and G. Okada, “High-energy 266-nm picosecond pulse generation from a narrow spectral bandwidth gain-switched LD MOPA,” Tech. Digest of CLEO 2013, JTh2A.64, (2013).
14. C. Chang, K. Hong, P. Krogen, H. Liang, G. Stein, J. Moses, C. Lai, J. Siqueira, L. Zapata, and F. Kärtner, “Multi-mJ, kHz intense picosecond deep ultraviolet source based on a frequency-quadrupled cryogenic Yb:YAG laser,” in Advanced Solid State Lasers, OSA Technical Digest (online) (Optica Publishing Group, 2014), paper ATu3A.3.
15. O. Novák, H. Turčičová, M. Smrž, T. Miura, A. Endo, and T. Mocek, “Picosecond green and deep ultraviolet pulses generated by a high-power 100 kHz thin-disk laser,” Opt. Lett. 41(22), 5210–5213 (2016). [CrossRef]
16. H. Turcicova, O. Novak, L. Roskot, M. Smrz, J. Muzik, M. Chyla, A. Endo, and T. Mocek, “New observations on DUV radiation at 257 nm and 206 nm produced by a picosecond diode pumped thin-disk laser,” Opt. Express 27(17), 24286–24299 (2019). [CrossRef]
17. K. Kohno, Y. Orii, H. Sawada, D. Okuyama, K. Shibuya, S. Shimizu, M. Yoshimura, Y. Mori, J. Nishimae, and G. Okada, “High-power DUV picosecond pulse laser with a gain-switched-LD-seeded MOPA and large CLBO crystal,” Opt. Lett. 45(8), 2351–2354 (2020). [CrossRef]
18. K. Liu, H. Li, S. Qu, H. Liang, Q. J. Wang, and Y. Zhang, “20 W, 2 mJ, sub-ps, 258 nm all-solid-state deep-ultraviolet laser with up to 3 GW peak power,” Opt. Express 28(12), 18360–18367 (2020). [CrossRef]
19. Y. Orii, K. Kohno, H. Tanaka, M. Yoshimura, Y. Mori, J. Nishimae, and K. Shibuya, “Stable 10,000-hour operation of 20-W deep ultraviolet laser generation at 266 nm,” Opt. Express 30(7), 11797–11808 (2022). [CrossRef]
20. W. Koechner, “Solid-State Laser Engineering,” Springer, sixth revised and updated edition, p499–p521. (2005).
21. G. Kurdi, K. Osvay, J. Klebniczki, M. Divall, E. J. Divall, Á. Péter, K. Polgár, and J. Bohus, “Two-photon-absorption of BBO, CLBO, KDP and LTB crystals,” in Advanced Solid-State Photonics, Technical Digest (Optical Society of America, 2005), paper MF18.
22. T. Kamimura, R. Nakamura, H. Horibe, M. Nishioka, M. Yamamoto, M. Yoshimura, Y. Mori, T. Sasaki, and K. Yoshida, “Characterization of two-photon absorption related to the enhanced bulk damage resistance in CsLiB6O10 crystal,” J. Appl. Phys. 44(21), L665–L667 (2005). [CrossRef]
23. A. Dubietis, G. Tamošauskas, A. Varanavičius, and G. Valiulis, “Two-photon absorbing properties of ultraviolet phase-matchable crystals at 264 and 211 nm,” Appl. Opt. 39(15), 2437–2440 (2000). [CrossRef]
24. Y. Orii, H. Sawada, D. Okuyama, M. Yamagaki, H. Izumi, K. Shibuya, and G. Okada, “A High-Power Deep-UV Picosecond Laser,” The Rev. Laser Eng. 45(9), 580–586 (2017). [CrossRef]
25. V. G. Dmitriev, G. G. Guradyan, and D. N. Nikogosyan, “Handbook of nonlinear optical crystals,” Springer, third revised edition, p50, 51. (1999).
26. N. Umemura, K. Yoshida, T. Kamimura, Y. Mori, T. Sasaki, and K. Kato, “New data on the phase-matching properties of CsLiB6O10,” ASSL Vol. 26 of OSA Trends in Optics and Photonics (Optical Society of America, 1999), paper PD15.
27. K. Takachiho, M. Yoshimura, Y. Takahashi, M. Imade, T. Sasaki, and Y. Mori, “Ultraviolet laser-induced degradation of CsLiB6O10 and β-BaB2O4,” Opt. Mater. Express 4(3), 559–567 (2014). [CrossRef]
28. Y. K. Yap, K. Deki, N. Kitatochi, Y. Mori, and T. Sasaki, “Alleviation of thermally induced phase mismatch in CsLiB6O10 crystal by means of temperature-profile compensation,” Opt. Lett. 23(13), 1016–1018 (1998). [CrossRef]
29. K. Tsubakimoto, H. Yoshida, and N. Miyanaga, “600 W green and 300 W UV light generated from an eight-beam, sub-nanosecond fiber laser system,” Opt. Lett. 42(17), 3255–3258 (2017). [CrossRef]
