Compared to other visible spectral regions, a laser source emitting in the yellow to orange band, i.e., with a wavelength roughly around 570–620 nm, is relatively difficult to access, at least when high output power, beam quality, and power efficiency are required. However, orange and yellow laser sources are applied, e.g.,  for laser guide stars (sodium laser beacons) and in medical therapies (e.g.,  photocoagulation in ophthalmology). During the past years, researchers have experimented with laser output in this specific band using various methods, including dye lasers at about 575 nm^[1], copper vapor lasers emitting at 578 nm^[2], InGaP/InAlGaP-based laser diodes operating at 610 nm^[3], Pr3+, Yb3+-codoped upconversion fluoride fiber lasers at 605 nm^[4], intracavity frequency-doubled lasers based on gain medium of Cr4+:MgSiO4 (forsterite) crystal^[5], and GaInNAs or InGaAs quantum wells^[6,7]. Some yellow or orange laser sources are based on sum-frequency generation. For example, mixing the outputs of two Nd:YVO4 lasers emitting at 1064 and 1342 nm, respectively, results in orange light with 593.5 nm^[8]. There are Raman lasers, often based on Raman-active bulk crystals (e.g.,  tungstate crystals), which can either generate orange or yellow lasers from green pump light^[9] or generate light with wavelengths around 1.1–1.2 µm with a 1 µm pump source^[10], so that subsequent frequency doubling or sum-frequency generation leads to orange or yellow lasers. Overall, these methods can be classified into two categories, namely direct and indirect methods. Moreover, it can be seen that these direct methods are either inefficient, cumbersome, or expensive. In addition, these indirect methods make it difficult to overcome the problems of complexity and inefficiency.

Recently, research on directly generating yellow/orange lasers based on fiber lasers and solid-state lasers has attracted much attention. For example, Zou et al.^[11] reported a blue diode laser pumped a Dy3+-doped yellow fiber laser with a laser wavelength of 575 nm and a maximum output power of 1.12 W. Regarding orange lasers, in fact, rare Earth Pr3+ ions have rich emissions in the visible spectral region. Taking the most widely studied Pr:YLF crystal as an example, this crystal has two significant emission peaks in the orange band peaking at about 607 and 604 nm when pumped by a blue diode laser. This orange laser is directly obtained based on solid-state laser technology and has the advantages such as high efficiency, stability, and low cost. Moreover, orange light when operating in single-frequency mode is considered to have important applications in fields such as flow cytometry^[12] and the treatment of epidermal pigmented lesions^[13]. Nevertheless, one fact is that current research methods for single-frequency orange lasers are very limited, and there are few reports on orange single-frequency lasers. One example is that, in 2000, Spiekermann et al.^[14] reported a 596 nm single-frequency orange laser with a maximum output power of 16.7 mW by sum-frequency mixing two single-frequency Nd:YAG infrared lasers at 1064 and 1357 nm. Obviously, this is also an indirect method of obtaining orange lasers, and it is quite complex, resulting in a low-power single-frequency laser. By inserting two etalons into a diode-pumped Pr:YLF laser cavity with a simple linear cavity configuration, in 2020, we achieved the first all-solid-state Pr-doped single-frequency laser in the orange spectral region^[15]. Although there has been a significant increase in output power compared to the aforementioned orange single-frequency laser, the insertion of two etalons still restricts the maximum output power of the orange single-frequency laser to only 175 mW. Moreover, this method of inserting etalons into the resonant cavity makes it difficult to suppress multi-longitudinal-mode laser emission at a high pump power, so it is only suitable for low-power single-frequency laser output. Very recently, research on Pr-doped single-frequency lasers based on the ring cavity has received attention^[16–19]. For instance, a near 4.0 W single-frequency red laser has been successfully demonstrated^[16], which confirms the effectiveness of the ring cavity in the development of high-power Pr-doped single-frequency lasers. To address this issue and further enhance the output power of the Pr3+ orange single-frequency laser, in this present investigation, we have applied the ring cavity technology to the study of the Pr3+ orange single-frequency laser and ultimately achieved watt-level output. In addition, we also conducted other extensive research on orange single-frequency lasers, including wavelength tuning and simultaneous dual-wavelength operation.

Experimental setup of the all-solid-state single-frequency orange laser is schematically shown in Fig. 1. Two InGaN blue diode lasers LD1 and LD2 were used for dual-end pumping of the laser gain medium, which both emit at 444 nm with maximum output powers of about 12 W. The focal lengths of the two focusing lenses f1 and f2 are both 60 mm. The two concave input mirrors M1 and M2 are coated with high transmittance (>96%) at the pumping wavelength and high reflectivity (>99.9%) at the laser wavelengths. The curvature radii of M1 and M2 are both 50 mm. The concave folding mirror M3 is coated with high reflectivity (>99.9%) at laser wavelengths and high transmittance of 90% around 639.5 nm, which is the highest emission peak of the laser crystal. The concave mirror M4 is coated with a partial transmittance of about 2.7% at the orange wavelength range, acting as the output coupler. The curvature radii of M3 and M4 are both 100 mm. An optical diode (OD) consisting of a half-wave plate (HWP) and a Faraday rotator was inserted into the ring cavity for single-frequency operation. It should be pointed out that originally the available OD was coating designed for the operation of single-frequency 640 nm red lasers^[16]. With respect to the orange light, the transmittances of the HWP and Faraday rotator (FR) are reduced to 99.6% and 99.3%, respectively, according to our measurements. These coating losses will undoubtedly affect the output performance of orange single-frequency lasers.

The used gain medium is a 15-mm-long a-cut Pr:YLF crystal. The crystal is AR-coated and has a dopant concentration of 0.2% (atom fraction). Under this situation, we measured the absorption ratio of the total pump power by this crystal to be 79%. To reduce the thermal lensing effect, the crystal was wrapped with indium foil and enclosed with a copper block, which was water-cooled by a chiller with temperature set at 14°C. The output beam was split into three parts and directed to a Fabry–Perot (F–P) scanning interferometer, an optical spectrum analyzer (OSA), and a power meter, respectively, to monitor the longitudinal mode, wavelength, and power at the same time.

According to the standard ABCD law, we designed and optimized the total physical length of the laser cavity to be about 435 mm, which was short enough for single-frequency operation and long enough for the insertion of the OD and the etalon. The folding angles of the four mirrors were designed to be about 5°, small enough for neglecting the astigmatism. This laser resonator is designed to have a structure with good symmetry. The distance between the M1 mirror and laser crystal is approximately 24 mm, which is the same as the distance between the M2 mirror and crystal; the distances of M1–M4 and M2–M3 are also about the same, which are both 110 mm; the distance between M3 and M4 is approximately 140 mm. Under this situation, we can calculate that the beam waist size in the Pr:YLF crystal is about 50 µm. Additionally, a 0.15-mm-thick uncoated F–P etalon was inserted into the cavity for wavelength tuning when needed.

YLF crystal is a typical uniaxial fluoride crystal, therefore Pr3+-doped YLF crystal has two orthogonal polarization directions, as shown in Fig. 2, namely the σ-polarized 607 nm emission line and the π-polarized 604 nm emission line. The orange emission studied in this work is based on the P30→H36 transition of Pr3+ ions, and there is a stronger emission of red emission in the long wavelength about 35 nm away from the orange emission. Therefore, in the design of this experiment, we suppressed the red emission by appropriately coating the laser cavity mirror. In addition, since 607 nm is stronger than 604 nm from the perspective of emission intensity, in the absence of particular polarization selection components in the laser cavity, the 607 nm emission line should be stimulated first, thereby suppressing the slightly weaker 604 nm emission.

The result of the first laser experiment was obtained without the presence of the etalon. At a lower pump power, we first observed 607 nm single-wavelength laser emission. As the pump power increased to approximately 14 W, we also observed the appearance of a 604 nm wavelength, which means simultaneous dual-wavelength laser emission. However, in this case, we found that the laser did not operate in the single-frequency mode, but instead in the few-longitudinal mode. This phenomenon indicates that under high-power pumping, the gain level inside the laser cavity is also very high, resulting in the inability to suppress the low-intensity 604 nm emission line. However, by finely tuning the OD, the 604 nm oscillation can be finally suppressed, and a stable single-frequency laser at 607 nm was then achieved. This can be explained by the OD-induced extra intracavity loss difference for these two wavelengths. Figure 3 shows the output power of the 607 nm single-frequency laser. In comparison to the previous results, the maximum output power achieved in this work was significantly improved to 1.19 W, and the slope efficiency with respect to the pump power was fitted to be about 8.6%. No power saturation was observed at the maximum pumping power, indicating that there was potential for further improvement in the output power. For the Pr:YLF orange laser, we have ever achieved a 42% high-efficiency laser output using a simple plane-concave cavity^[20]. The main reasons for the present low-efficiency operation in this work can be analyzed as follows. First, a relatively complex ring cavity was used, and it operated unidirectionally, which led to increased cavity losses and a direct halving of output power. Second, in this work, we used a high-power pump source with low beam quality and a longer laser crystal. According to our estimation, this directly leads to a mode matching of about 59% between the pump beam and cavity mode in the Pr:YLF crystal in the current study. According to the formula for laser efficiency, this almost further reduces the laser efficiency by half. The optical spectrum was measured with an ANDO AQ-6315 E OSA, and the result is shown in the inset of Fig. 3. When the resolution of the OSA was set to 0.05 nm, the peak wavelength was measured to be about 607.42 nm. The spectral characteristic of the single-frequency laser at its maximum output power was analyzed with an F–P confocal scanning interferometer (SA200-5B with a free spectral range (FSR) of 1.5 GHz, a resolution of 7.5 MHz, and a finesse of >200, Thorlabs), and the result is shown in Fig. 4, which clearly indicates that the laser operated in the single longitudinal mode (SLM). The calculated linewidth of the single-frequency laser was about 20.3 MHz. The power stability of the 607 nm single-frequency laser in 20 min is shown in Fig. 5. The average output power was 1.18 W, and the root-mean-square (RMS) power stability relative to the average power during this period was found to be 0.41%, indicating good power stability. The beam quality was also evaluated by measuring the beam spot size along its propagation axis after a 75 mm (focal length) focusing lens. By fitting the data, we can estimate that the 607 nm single-frequency laser has M2 factors of 1.48 and 1.57 in x and y directions, respectively, as shown in Fig. 6. Basically, the beam quality can be referred to as near the diffraction limit, but we still hope to further optimize the output beam quality in future work, making it closer to TEM00 mode, which should have a positive effect on improving the power stability of the laser.

In the experiment, we found that aligning the OD can produce single-frequency laser output at 604 nm too, but the oscillation was not stable, and oscillation at 607 nm showed up occasionally. That is to say, in this situation, it is not possible to suppress the laser emission of 607 nm solely by adjusting the OD. In order to realize pure 604 nm output, a 0.15-mm-thick etalon was inserted into the cavity to filter the undesired oscillation. Transmission of an etalon can be written as^[15]T=(1−R)2(1−R)2+4Rsin2(2πnlecosθλ),(1)where R is the reflectivity of the etalon, n is the refractive index, le is the thickness of the etalon, λ is the wavelength, and θ is the refraction angle inside the etalon. We plot the transmissions of the etalon at 607 nm and 604 nm and their difference versus the tilt angle in Fig. 7, from which one can see that the insertion of the etalon enhances the wavelength discrimination. For example, in the experiment, we completely suppressed the oscillation at 607 by optimizing the position and tilting the angle of the etalon between about 1° to 1.5°. In Fig. 8, the output power of the 604 nm single-frequency laser is shown. The maximum output power was reduced to 0.69 W, and the slope efficiency was fitted to be about 5.2%. Compared to the 607 nm laser, there is a significant attenuation in the output power of this 604 nm laser, which is partly due to the weak gain of this emission line itself, and partly due to insertion loss of the used etalon, which can be estimated to be about 0.2%^[21]. The optical spectrum is shown in the inset of Fig. 8, and the peak wavelength is 604.22 nm. The single-frequency operation of the laser at its maximum output power is confirmed using the F–P scanning interferometer, as shown in Fig. 9. The calculated linewidth was estimated to be about 16.7 MHz. Figure 10 shows the power stability of the single-frequency laser at 604 nm within 20 min to be about 0.56%, slightly decreased compared to the 607 nm single-frequency laser. We attribute the stability deterioration to the insertion of the etalon, which led to degradation in mechanical stability.

Moreover, the inserted etalon has not been temperature controlled, so the temperature stability also has a negative impact on the output power stability of the single-frequency laser.

Pr:YLF crystal has relatively wide emission in the orange spectral region, and if the wavelength of a single-frequency laser is tunable, it will undoubtedly enhance the applicability of the single-frequency laser. Driven by this motivation, we conducted wavelength-tuning experiments on the two orange single-frequency lasers with the help of the etalon. As shown in Fig. 11, for the 604 nm single-frequency laser, its wavelength can be tuned from 603.99 to 605.02 nm, respectively, with output powers of 671 and 589 mW. Between the two wavelengths, by finely adjusting the etalon, we also obtained nine other single-frequency lasers, which show good continuity of wavelength tuning. For the 607 nm single-frequency laser, we found its wavelength tuning from 607.16 to 607.61 nm. It should be pointed out that, in order to ensure that these tuned lasers all operate in single-frequency mode, we need to slightly tilt the angle of the etalon and at the same time gently orientate the FR.

In order to achieve simultaneous laser operation of the two orange lasers and ensure that they operate in a single-frequency laser state, we also took into account fine tuning of the HWP because in reality the thickness of the HWP is only 60 µm, which is thinner than the used etalon. Compared to the vertical incidence state maintained by the HWP during the above experiments, we found that simultaneous laser operation of the two orange single-frequency lasers can be achieved when the incident angle of the HWP is about 5°. Through a Glan–Taylor polarizer, we first discovered that the polarization directions of the two single-frequency lasers are orthogonal, namely π polarization for the 604 nm laser and σ polarization for the 607 nm laser, respectively. Moreover, when we observed that the spectral intensities of the two orange lasers were comparable to the OSA, we recorded that the maximum power of the orthogonal polarized dual-wavelength orange single-frequency laser was about 0.97 W, as shown in Fig. 12, which plots the input–output power relationship. Reducing the pump power resulted in a decrease in output power but still maintained the dual-wavelength single-frequency state monitored with the F–P scanning interferometer. The inset in Fig. 12 is the laser spectrum of the dual-wavelength orange laser at its maximum power, showing peak wavelengths at 607.36 and 604.64 nm. Figure 13 shows the result of the F–P scanning interferometer with an estimated linewidth of 9.7 MHz for the 607 nm laser and 9.3 MHz for the 604 nm laser. Comparing the single-frequency lasers of 607, 604, and (607 + 604) nm obtained in the experiments, we found that, with the insertion of the etalon, the linewidth of the laser decreased, which may be due to the further narrowing of the linewidth by the etalon.

For the laser wavelength measured by the OSA in the experiment, we did not observe any wavelength changes, which means that at the current resolution of the OSA, all laser wavelengths are stable. However, since we did not adopt active frequency stabilization measures in the current experiment, we observed weak time-domain jitter when using the scanned confocal F–P interferometer as a high-resolution measurement device. Therefore, these reports are only representative measurement results of F–P interferometers. In addition, this jitter undoubtedly leads to inaccuracy in the linewidth determination of these single-frequency lasers. Based on our estimation, the measured frequency drift was less than 10 MHz.

In conclusion, single-frequency laser operation in the orange spectral region has been demonstrated based on all-solid-state Pr:YLF ring lasers. A σ-polarized 607 nm single-frequency laser has been obtained with a maximum output power of 1.19 W and a linewidth of 20.3 MHz. The lasing wavelength can be tuned from 607.16 to 607.61 nm. A π-polarized 604 nm single-frequency laser is also achieved with a maximum output power of 0.69 W and a linewidth of 16.7 MHz, and the wavelength can be tuned from 603.99 to 605.02 nm. The two orthogonally polarized orange lasers can also operate simultaneously as a single-frequency laser with linewidths of 9.7 and 9.3 MHz, respectively.

In this work, we did not find any phenomenon of output power saturation, so it may be feasible to increase the pump power to improve the output power. In addition, the OD system we used in the experiments was not designed for orange lasers, so there are additional losses, and the transmittance of the output coupler is also low. We hope that in the near future, after optimizing the above aspects, the output power of orange single-frequency lasers can be further significantly improved.