High-energy nanosecond lasers operating at ∼2µm are of great significance in serving as candidate drive sources of high-energy extreme ultraviolet (EUV) in lithography applications owing to their compact architecture, excellent stability, and pulse controllability, as well as a comparable conversion efficiency (CE) of EUV to that driven by 10.6 µm CO2 lasers^[1–3]. In addition, 2 µm lasers can be used in medical treatment, atmospheric molecular detection, and lidar applications as the absorption band of H2O and CO2 lies around 2 µm^[4–6]. Furthermore, the 2 µm lasers have been proven to be ideal pump sources for mid-infrared optical parametric oscillation (OPO) and optical parameter amplification (OPA), which can generate laser radiation in the longer wavelength region covering the 3–5 µm band^[7,8]. Two schemes are commonly employed for generating 2 µm lasers. Scheme one involves utilizing Ho3+-doped or Tm3+-doped materials as the gain media, which can be pumped by Tm lasers or laser diodes (LDs), respectively. In 2023, Wang et al. demonstrated a Tm:YAP-laser-pumped Ho:YLF master oscillation power amplifier (MOPA) and obtained a pulse energy of 84.7 mJ at a 100 Hz repetition rate with a pulse width of 18 ns^[9]. Another scheme entails generating 2 µm lasers through OPO pumped by 1 µm lasers. Compared with scheme one, this scheme offers the benefits of broad spectral tunability and mild thermal effects. Potassium titanyl phosphate (KTP) has emerged as one of the most feasible nonlinear crystals for OPO due to its large nonlinear coefficient and high damage threshold, while allowing a wide range of phase matching angles and a spectral tuning range covering 1.57–3.39 µm when pumped by a 1064 nm laser^[10–12]. For the KTP-OPO, type II phase matching is predominantly employed since type I phase matching encounters a smaller nonlinear coefficient and thus limited practical applicability. In 2010, Wei et al. reported a degenerate 2128 nm intracavity KTP-OPO with a pulse energy of 107 mJ and an optical CE of 10.5% at the repetition rate of 30 Hz, in which two KTP crystals were positioned in a walk-off compensation arrangement^[13]. In 2015, He et al. reported a 2 µm KTP-OPO with a high average power of 18 W and a pulse width of 70 ns. They used a 1064 nm Nd:YAG laser as the pump source with a repetition rate of 7 kHz and a pump power of 410 W. The central wavelengths of the signal and idler were 2101.4 and 2155.1 nm, respectively^[14]. In 2016, Mei et al. demonstrated a widely tunable KTP-OPO pumped by an acousto-optically Q-switched Nd:YVO4 laser, producing a maximum output power of 3.65 W with a wavelength tuning range of 2.088–2.133 µm (signal) and 2.122–2.171 µm (idler)^[15].

For the 2 µm laser-driven EUV experiments, a high-energy drive source is generally required. In 2021, Behnke et al. scattered tin droplet targets utilizing a 2 µm laser with a total pulse energy of several 100 mJ and produced a high EUV CE of 3.1% and a spectral purity (SP; the proportion of in-band EUV energy to the total energy in the range of approximately 5.5–25.5 nm) of 7.4%, surpassing the CE (1.7%) and SP (4.2%) obtained with the use of 1-µm-drive lasers with the same conditions^[16]. They combined the KTP-OPO and OPA (KTP-MOPA) to generate 2 µm lasers employing a total of four KTP crystals. The 2 µm pulse energy of the OPO stage is about 3.6 mJ, and the optical CE of the oscillator and amplifier are both less than 25%. In 2023, Lars Behnke et al. reported a high-energy 2 µm KTP-MOPA light source for EUV plasma applications, and they obtained 2 µm pulse energy of 17 mJ in the OPO stage with an optical CE of 24% and a root mean square (RMS) energy stability of 2%^[17].

In this work, we developed a 2 µm nanosecond laser with a total pulse energy of 162.6 mJ and an optical CE of 37.9% using a plane–plane cavity OPO, which employed two KTP crystals positioned opposite to compensate for the walk-off effect. The 2 µm OPO features an excellent RMS energy stability of ∼0.4%. To the best of our knowledge, this is the highest 2 µm pulse energy that is achieved from just a single KTP-OPO scheme without any additional OPA stages. Additionally, the effects of the walk-off, crystal lengths, pump trips, output coupling (OC) transmittance, and cavity lengths on the output performance were studied in detail.

The schematic of the experimental setup is shown in Fig. 1. The pump source is an electro-optically Q-switched nanosecond Nd:YAG laser with a central wavelength of 1064 nm, a repetition rate of 10 Hz, and a pulse width of 6.3 ns. A half-wave plate 1 (HWP₁) and a thin-film polarizer 1 (TFP₁) are utilized to obtain a horizontally polarized pump light, and the pump energy is adjustable by rotating the optical axis angle of HWP₁. A Faraday rotator (FR) cooperating with HWP₂ and TFP₁ constitutes an optical isolator to prevent the feedback light from damaging the pump source. The TFP₂ can further improve the polarization degree of the pump light, and both TFPs are placed at a Brewster angle of 56.6°. The focal lengths of the lenses L₁ and L₂ are 300 and −240mm, respectively, and the spot diameter of the pump after two lenses is 6.5 mm. The crystals utilized in the OPO are KTP, which is a biaxial crystal with different refractive indices in three principal axes. In the x–z principal plane, KTP has a larger gain coefficient and nonlinear coefficient than the other principal planes; hence, more efficient phase matching can be achieved. The 2 µm laser output can be achieved at a phase-matching angle of 51.4° according to the calculation; therefore, the KTP crystals employed in the experiment are cut based on type II phase matching with an angle of θ=51.4° and φ=0°, and then they are anti-reflective (AR) coated for 1.064 and 2.1 µm in both end faces. KTP crystals with three different lengths are tested, which are 15, 20, and 25 mm. The angle between the pump wave vector and the z-axis of KTP is less than 90°, inevitably causing a spatial separation between the extraordinary light (e-light) and the ordinary light (o-light) under critical phase matching^[18]. Two KTP crystals are deployed within the oscillator with their optical axes opposite each other to mitigate the walk-off effect, consequently, obtaining a double parametric gain (ignoring the surface reflection of KTP crystals) while eliminating the spatial separation of the signal and the idler^[19]. The plane end mirror M₁ is AR-coated at 1.064 µm and highly reflective (HR) coated at 2.1 µm. Two sorts of OC mirrors M₂ that are AR- and HR-coated at 1064 nm are utilized in the OPO, which corresponds to the single- and double-pass pump configurations, respectively. Three output coupling transmissions of 10%, 20%, and 30% are tested to optimize the output performance. A dichroic mirror (DM) that is AR-coated at 2.1 µm and HR-coated at 1.064 µm (AOI=45°) is applied to filter the residual pump light. Subsequently, the signal and the idler are separated by TFP₃.

We first investigated whether the compensation of spatial walk-off has a notable impact on the output performance of 2 µm OPO, and the results are illustrated in Fig. 2. The total energy of the signal and idler are measured behind the DM (detected by Ophir-PE50BF). It is found that the pulse energy and efficiency with compensation are higher than those without compensation at a high pump energy. In the absence of compensation, there exists a distinct spatial separation between the signal and the idler; consequently, the pulse energy cannot be spatially concentrated and sufficiently exploited, as shown in the inset of Fig. 2. The pump intensity is 206.1MW/cm2 at the maximum pump energy of 431 mJ, which is lower than the general damage threshold of 500MW/cm2^[20] for the coatings of the KTP crystal. The pump threshold measured in the experiment is 26.6MW/cm2 for the single-pass pump configuration with a cavity length of 62 mm and an OC transmission of 20%. The calculated pump threshold is 28.5MW/cm2^[21] with equivalent parameters, which elucidates that the experimental design aligns closely with theoretical expectations.

The effect of the pump trips, crystal length, and OC transmittance on the 2 µm output performance is investigated with the OPO cavity length maintained at 62 mm in all cases. As shown in Fig. 3, apparently, higher 2 µm pulse energy can be obtained when using an OC with higher transmittance. The comparison between Figs. 3(a) and 3(b) demonstrates that the pump threshold of OPO can be reduced by increasing the pump trips at the same crystal length of 20 mm. Higher energy can be obtained in the double-pass pump configuration over that of the single-pass case at the OC transmittances of 10% and 20%. The results are consistent with the qualitative analysis of enhancing the nanosecond OPO CE and reducing the OPO threshold in Ref. [21]. However, when using the OC of T=30%, the case of the single-pass pump configuration can emit higher maximum pulse energy than that of the double-pass case, with a higher pump threshold than that of the T=10% and 20% cases. A similar phenomenon is more obvious when applying a 25-mm-long KTP crystal, as shown in Figs. 3(c) and 3(d).

In addition, the optical CE and pump threshold exhibit negligible alteration with the extension of the crystal length in the double-pass pump configuration. For 20 and 25 mm KTP crystals, the maximum optical CEs of 29.16% and 29.86% and the minimum pump thresholds of 19.8 and 16.8MW/cm2 are obtained, respectively. Nevertheless, for the single-pass configuration, the crystal length has a more pronounced effect on the output performance, with maximum efficiencies of 33.8% and 36.6% and minimum pump thresholds of 52.1 and 26.6MW/cm2 for the 20 and 25 mm KTP crystals, respectively.

Although extending the crystal length can increase the effective gain in the cavity, it simultaneously leads to a larger total intrinsic loss introduced by the KTP crystals. We then utilized a 15-mm-long KTP crystal to exemplify this phenomenon observed in the aforementioned circumstance, as shown in Figs. 4(a) and 4(b). With the same cavity length of 62 mm, it is evident that the 2 µm pulse energy obtained using the double-pass pump configuration is higher than that of the single-pass case in each OC transmittance. In addition, the maximum pulse energy of the double-pass case is 135.9 mJ, surpassing the pulse energy of 124.8 and 125.8 mJ when employing the 20 and 25 mm crystals in the double-pass case, which implies that the intrinsic loss effect can be mitigated by appropriately decreasing the gain length. Different pump passes correspond to different pump optical pathlengths in the crystals. Table 1 summarizes the maximum output energy under various pump pathlengths. It is obvious that the output energy first increases and then decreases with the pump pathlength. To summarize the above-mentioned experiments, the highest 2 µm pulse energy was obtained in the single-pass pump configuration utilizing the 25-mm-long KTP crystal with the OC of T=30%.

Furthermore, we also optimized the output performance of the KTP-OPO by adjusting the cavity length, and the results are shown in Fig. 4(c). The cavity lengths of 115, 87, 75, and 58 mm were evaluated, and it is evident that the shorter cavity length embodies the higher pulse energy and lower OPO pump threshold. As the cavity length increases, the diffraction loss in the OPO also increases, resulting in a decrease in the total output energy. The shortest cavity length is 58 mm due to the limitation of the mechanical structure. The maximum 2 µm pulse energy reaches 162.6 mJ, with an optical CE of 37.7% and a slope efficiency of 45.2%. For specific application scenarios, the cavity length can be adjusted appropriately to improve the beam quality.

The signal (o-light) and the idler (e-light) feature disparate polarizations owing to type II phase matching and thus can be separated by TFP₃. The maximum pulse energy of the signal is 81.9 mJ, and the slope efficiency is 26.3%, as shown in Fig. 5(a), which is similar to the idler (79.5 mJ, 25.4%). The central wavelengths of the signal and the idler are 2109 and 2149 nm, respectively, as shown in Fig. 5(b). The spectral full widths at half-maximum (FWHMs) of the signal and the idler are 10.6 and 10.8 nm, respectively, and the actual value should be smaller than the measured value due to the limited resolution of the spectrometer, which is only 3 nm. Typical pulse profiles of the signal and idler are measured by a fast photodetector (rise time <28ps, EOT-et5000), as shown in Figs. 6(a) and 6(b). The pulse widths are 2.28 and 2.19 ns, respectively. The energy stabilities of the pump, signal, idler, and total 2 µm laser are continuously monitored for 30 min. The pump light is measured in front of the end mirror M₁, which has an RMS energy stability of 0.13%, as shown in Fig. 7(a). The OPO also exhibits excellent energy stability with a total 2 µm laser RMS value of 0.40%, signal RMS of 0.35%, and idler RMS of 0.31% [see Figs. 7(b)–7(d)]. To achieve a superior spatial concentration of the pulse energy, we combine the two beams and focus them after precise tuning. A lens with a focal length of 150 mm is placed 280 mm away from M₂, and the inset of Fig. 7(d) presents the far-field beam spot after the combination of two beams. The beam exhibits a good Gaussian distribution in both horizontal and vertical directions, measuring 1.92mm×2.54mm in size, which is a little elliptical due to the spatial walk-off effect. The reception angles of the KTP crystal differ in the vertical and horizontal directions; thus, the beam spot is stretched in the vertical direction due to the larger acceptance angle. The smaller acceptance angle in the horizontal direction limits the size of the transverse beam during oscillation, thus enhancing the horizontal beam quality^[21].

In conclusion, we have demonstrated a high-energy and high-efficiency 2 µm nanosecond laser based on a dual-KTP-OPO. The KTP-OPO is pumped by a 1064 nm Nd:YAG laser with a repetition rate of 10 Hz and can produce a maximum pulse energy of 162.6 mJ, which is the highest 2 µm pulse energy achieved solely with a single KTP-OPO without any additional amplification stages. A maximum optical CE of 37.7% and a slope efficiency of 45.2%, as well as an RMS energy stability of 0.4%, are achieved. Additionally, we characterized the effects of various factors on the output performance of KTP-OPO and found that higher pulse energy could be obtained with the compensation of spatial walk-off and the use of relatively larger OC transmission and shorter cavity lengths. The total intrinsic in the KTP crystals cannot be negligible; thus, appropriately decreasing the crystal length and pump trips can facilitate the improvement of 2 µm pulse energy. In the future, with further optimization of cavity mode matching and mirror coatings, we believe that higher pulse energy and better beam quality should be achievable, which can be exploited as an excellent laser source for demonstration experiments to drive EUV radiations.