1 Introduction

The lasers with output wavelengths around 550–560 nm have important applications in medical treatment and molecular biology, because some yellow-green lights among this range locate at the absorption peak of several popular fluorescent dyes and in hemoglobin [1, 2]. In addition, these sources in this spectral region have excellent penetration through fluid and pigmentary disturbances, less dispersion of energy in the neurosensory retina, less discomfort to the patient, greater margin of safety than other sources [3]. Moreover, the wavelength range from 550 to 560 nm is most sensitive to human eyes and suitable for laser display and lighting [4]. Generally, these sources are obtained by methods such as sum-frequency mixing (SFM) of Nd:YAG or Nd:YVO4 lasers [5, 6], Raman shifting pumped with a 532 nm laser [7], direct frequency doubling of Nd:YAG lasers at near 1.1 μm [8–13]. For the sum-frequency process and Raman shifting, complex components and complicated operating are required, and for the doubling frequency, only the radiations at 556 and 561 nm were obtained by frequency doubling of Nd:YAG lasers at 1,112 and 1,123 nm [3, 14–16].

We focused in this work on the case of another well-known neodymium-doped material, Nd:GGG (Gadolinium Gallium Garnet). GGG is one of the best crystals suitable for Nd3+ doping. It has a number of advantages like the possibility to be grown in larger size [17], a good thermal capacity of 0.38 J/(g K) [18] and a weak concentration quenching offering the possibility of higher doping levels, around 4 % in contrast to about 1.5 % in the case of YAG [19, 20]. Consequently, Nd:GGG crystals have been used as the best host for the solid state heat capacity laser (SSHCL) [18, 21]. However, it can also be used to create high-power lasers with self-phase conjugation [22]. The characteristics of the Nd:GGG lasers at common spectral lines have been widely investigated, e.g., 1,062 nm [23, 24], 1,331 nm [25, 26], and 938 nm [27, 28]. Besides these three lines mentioned above, other wavelengths are expected to be generated with Nd:GGG crystal as laser material according to the fluorescence spectra of Nd:GGG crystal [20], i.e., near 1.1 μm. It is expected to obtain lasers in the range from 550 to 560 nm by frequency doubling technology. In 2011, Zhang et al. [29] reported the dual-wavelength (1,110 and 1,105 nm) continuous wave (CW) Nd:GGG laser for the first time. The total output power of 13.2 W was obtained under the dual-wavelength CW operation condition. As far as we know, as for Nd:GGG, 1,105 nm single wavelength operation and the corresponding frequency-doubled 552 nm yellow-green lasers have not been reported.

In this paper, we report a high-power diode-side-pumped 1,105 nm Nd:GGG laser and a laser at 552 nm based on intracavity frequency doubling of 1,105 nm laser for the first time. The coatings of the cavity mirrors were carefully designed to optimize the performance of the laser, a single wavelength at 1,105 nm by suppressing the oscillation at 1,110 nm was achieved. And 26.8 W CW laser output at 1,105 nm was achieved when the pumping power of the laser diodes reached 170 W. Using a noncritical type phase matching LBO crystal, the maximum power of the frequency-doubled output at 552 nm was found to be as high as 7.3 W with a pulse repetition rate (PRR) of 10 kHz. This corresponds to an optical-to-optical conversion efficiency of about 4.3 %. To the best of our knowledge, these are the highest output powers for both 1,105 nm and 552 nm lasers ever reported.

2 Experimental setup

The experimental setup of the intracavity frequency doubling Nd:GGG/LBO laser at 552 nm is shown in Fig. 1. The laser head (Northrop Grumman, USA) was consisted of an Nd:GGG rod (1.0 at. %, ∅3 mm × 62 mm), a cooling sleeve, a diffusive optical pump cavity and three diode array modules operating at 808 nm. The total pump power for laser head was 180 W. Because the branching ratios at 1,105 nm line is smaller than that of the 1,062 and 1,331 nm lines, the coatings of resonator mirrors should be designed to suppress the parasitic oscillations at 1,064 and 1,331 nm [20, 29, 30]. Suppression of the above two transition lines can be done using high transmission coatings on the cavity mirrors. In our work, the rear mirror (RM) was coated for high-reflectance (HR) at 1,105 nm (R > 99.8 %), 1,110 nm (R > 99.6 %) and partial-reflection (PR) at 1,062 and 1,331 nm (T > 60 %) to suppress the 1,062 and 1,331 nm lines. However, the suppression of 1,110 nm without hurting the output of 1,105 nm is difficult to realize because 1,110 nm line is so close to the 1,105 nm line. According to Ref. [11], with the output coupler (OC) transmissions at 1,110 nm slightly higher (>1.9 %) than that at 1,105 nm, the laser can operate at 1,105 nm more efficiently, suppressing 1,110 nm laser. The OC in our experiment was a flat mirror with entrance face HR coated at 1,105 nm (R > 99.9 %) and PR at 552 nm (T > 90 %). The other face was HR at 1,110 nm (R > 97.5 %). So, the OC coating was carefully designed to place the 1,105 nm line at the edge of the transmission curve to introduce higher loss to 1,110 nm lines and to achieve efficient operation at 1,105 nm. Compared to flat–flat and concave–flat resonator, the employment of the convex–flat resonator increases the diffraction loss of high order transverse modes of the fundamental wave, which leads to better beam quality of the fundamental wave [31, 32]. So the laser cavity adopted a convex-flat structure, and the radius of curvature of RM was 1,000 mm. Using ABCD matrix method, the radius of mode size on the laser crystal and doubling crystal were calculated to be 420 and 382 μm, respectively. In order to collect 552 nm radiation in both directions, a coupling mirror (M1) coated HR at 552 nm (R > 99.8 %), high transmission (HT) at 1,105 nm (T > 99.6 %) and HT at 1,110 nm (T > 99.5 %) was employed in our experiment.

Fig. 1
figure 1

Schematic diagram of the intracavity frequency doubling Nd:GGG/LBO laser at 552 nm. RM rear mirror, OC output coupler, M1 coupling mirror, AO acousto-optical Q-switch, LBO LiB3O5 crystal

Full size image

The 46-mm long acousto-optical (AO) Q-switch (QS27-4S-B, Gooch and Housego, UK) had AR (R < 0.2 %) coatings at 1,060–1,130 nm on both faces and was driven with 15 W of rf power. The Nd:GGG rod was AR coated at 1,105 nm (R < 0.2 %) on both end faces. The LBO crystal (θ = 90°, ϕ = 8.7°) with a size of 2 × 2 × 10 mm3 was used as the frequency doubler to realize type II noncritically phase matching, and it was AR (R < 0.2 %) coated at 1,105 and 552 nm on both surfaces. The Nd:GGG laser module and the Q-switch were water cooled to be 20 °C. To achieve optimal phase matching, the LBO crystal was wrapped with indium foil and mounted in water-cooled copper blocks, and the water temperature was maintained at 18 °C. In order to reduce the diffraction loss, all the elements were placed as close as possible to form a compact cavity. As a result, the overall cavity length was 235 mm, the length of second harmonic generation (SHG) cavity was 38 mm.

The laser output power was measured by a power meter (EPM 2000. Coherent Inc., USA), and temporal behaviors of the Q-switched laser were recorded by a TDS 5052B digital oscilloscope (500 MHZ bandwidth, 5 G Samples/s, Tektronix Inc.). With an optical spectrum analyzer (Yokogawa AQ6315A, Japan), the emission spectra of the laser were measured. The beam quality was studied with a NanoScan beam analyzer (NS-PYRO/9/5, Photons Inc.) and a precision linear stage (Zolix, Inc.).

3 Experimental results

First, we investigated the operation of 1,105 nm Nd:GGG laser. The LBO crystal and mirror M1 in Fig. 1 were removed from the resonator and a flat mirror with entrance face HR coated at 1,105 nm (R > 98.9 %) and the other face HR coated at 1,110 nm (R > 96.5 %) was used as the output coupler. The RM mirror was coated the same to the RM mirror of frequency doubling experimental setup shown in Fig. 1. Figure 2 shows the average output power at 1,105 nm in CW mode and Q-switched mode as a function of the incident pump power. The output power increases monotonically as the pump power increases. The maximum CW output power of 26.8 W was obtained at an incident pump power of 170 W, corresponding to a slope efficiency of 22.4 % and a threshold power of 45.4 W. The highest average output power of 18.1 W was measured at 170 W of incident pump power in a Q-switched regime at a PRR of 15 kHz, corresponding to an optical conversion efficiency of 10.6 %. With the optical spectrum analyzer, the laser spectrum was recorded, which is shown in Fig. 3. We observed that, during all the CW and Q-switched processes, with 0.05 nm resolution of the optical spectrum analyzer, the peak wavelength was located at 1,104.9 (±0.05) nm with FWHM of 0.06 nm, and no other wavelength was detected.

Fig. 2
figure 2

Output power of 1,105 nm laser versus incident pump power

Full size image
Fig. 3
figure 3

Optical spectrum of 1,105 nm laser at the highest output power

Full size image

Using the 552 nm output coupler, we demonstrated the performance of the 552 nm laser. In order to determine the optimal PRR of the device working at 552 nm, we measured the output power of the 552 nm laser at different PRRs. The output power of the 552 nm laser versus PRR is shown in Fig. 4. To avoid damage to the optical elements, the optimization process was carried out at a pump power of 130 W. From Fig. 4, we can see the maximum output power was obtained at the PRR of 10 kHz. When the phase matching condition is satisfied through careful adjustment of the orientation of the LBO, a strong laser output at 552 nm generated by frequency doubling the 1105 nm laser can be obtained. The dependence of the output power at 552 nm on the incident pump power is shown in Fig. 5. It can be seen that the output power increased with the increasing pump power and the pumping threshold of 552 nm yellow laser was about 45.9 W. With the incident pump power of 170 W and a PRR of 10 kHz, the output power of 7.3 W at 552 nm was obtained, corresponding to an optical-to-optical conversion efficiency of 4.3 %. The output optical spectrum at the highest output power condition is shown in Fig. 6, with 0.05 nm resolution of the optical spectrum analyzer, the frequency doubling peak wavelength determined could be 552.6 (±0.05) nm with FWHM of 0.04 nm.

Fig. 4
figure 4

Output power at 552 nm laser versus PRRs

Full size image
Fig. 5
figure 5

Output power at 552 nm versus the incident pump power at a PRR of 10 kHz

Full size image
Fig. 6
figure 6

Optical spectrum of 552 nm laser at the highest output power

Full size image

The pulse temporal behaviors were recorded by an oscilloscope at the pump power of 170 W. The fundamental frequency pulses were detected by an InGaAs photodiode (1.5 GHZ bandwidth, 850–1,650 nm response spectrum, and 0.3 ns response time) and the SHG pulses by a silicon photodetector (400–1,100 nm response spectrum, and 1 ns response time). A typical oscilloscope trace of the Q-switched pulse train of the 552 nm laser is presented in Fig. 7. The pulse-to-pulse amplitude fluctuation of the pulse train was measured to be less than 4 %. The pulse waveform of 1,105 and the 552 nm is shown in Fig. 8, and the pulse widths were 219 and 181 ns, respectively. The corresponding pulse energy and peak power of 552 nm laser were 0.73 mJ and 4.03 kW, respectively. From Fig. 8 we can see under the same incident pump power, the pulse width of SHG was shorter than the fundamental frequency wave. This is because the SHG photon is excited by the fundamental frequency photon, and the loss of the SHG wave is larger than that of fundamental wave. Larger loss reduced narrower pulse width of the SHG wave.

Fig. 7
figure 7

Pulse train of 10 kHz PRR at the pump power of 170 W

Full size image
Fig. 8
figure 8

Typical Q-switched pulse at the pump power of 170 W

Full size image

Figure 9 shows the pulse widths versus input pump power at a PRR of 10 kHz. As pump power increased, the pulse width decreased. The relation between pulse width and input pump power was consistent with that between the fundamental frequency wave and input pump power. When the pump power increased from 45.9 to 170 W, the pulse width of SHG decreased from 432 to 181 ns.

Fig. 9
figure 9

Pulse widths versus pump power at a PRR of 10 kHz

Full size image

The beam quality of 552 nm beam was studied with a NanoScan beam analyzer (NS-PYRO/9/5, PHOTON Inc.) and a precision linear stage (Zolix, Inc.). Through the experiment, we found that the M 2 factors increased with the increasing pump power. The beam quality became worse as pump power increased. This is because higher pump power induces more serious thermal lensing effect [33]. The typical beam profiles of the 552 nm beam at the maximum output power in our experiment are shown in Fig. 10, where (a) and (b) show the three- and two-dimensional distributions, respectively. By focusing the beam with a lens (f = 200 mm), the M 2 factors of the 552 nm beam were measured to be 19.8 ± 0.5 and 17.6 ± 0.5 in the horizontal and vertical directions, respectively. We have done further experiments, we compared the convex–flat (the radius of curvature of RM was 1,000 mm), flat–flat and concave–flat (the radius of curvature of RM was 1,000 mm) structure. The cavity length was kept constant, with the equal pump power (100 W), the values of M 2 factor using convex–flat cavity, flat–flat cavity and concave–flat cavity were about 4.5, 6.3 and 10.6, respectively. The convex–flat cavity indeed improved the beam quality comparing with flat–flat cavity and concave–flat cavity. In the future, the aperture inserted to the cavity will limit the higher order transverse mode and improve the beam quality. Moreover, with the equal pump power, as cavity length increases, the beam quality would be improved [33].

Fig. 10
figure 10

Intensity distribution of the 552 nm beam at output power of 7.3 W. a the three-dimensional distributions, b two-dimensional distributions

Full size image

4 Conclusions

A diode-side-pumped frequency-doubled Q-switched Nd:GGG/LBO laser at 552 nm with a compact convex–flat straight cavity has been demonstrated. The maximum output power of 7.3 W was achieved at the repetition rate of 10 kHz and pump power of 170 W, corresponding to optical-to-optical conversion efficiency of 4.3 %. The pulse width was 181 ns at the maximum output power, and the M 2 factors were measured to be 19.8 and 17.6 in the two orthogonal directions, respectively.