Our first setup is based on a DPSS self-Q-switched Nd:YAG laser from TEEM-Photonics delivering 7 μJ nanosecond pulses at a rate of 20 kHz, we also investigate a more powerful model (100µJ) but with a lower repetition rate (1KHz). We are now considering a new model from TEEM photonics with an higher repetition rate.
The beam waist inside the laser crystal was estimated to be 50 μm. For the SHG non-linear crystal, instead of Lithium Triborate (LBO), usually used in commercial systems, we have employed a Potassium Titanyl Phosphate (KTP) crystal (EKSMA-Optics, Vilnius, Lithuania), with a moderate focusing of the fundamental beam in order to minimize the “gray tracking” effect. The crystal was cut for Type II excitation (Θ = 90 and φ = 23.5). This phase-matching condition has the advantage of being very close to non-critical phase matching at room temperature, while still keeping a good conversion efficiency. With a 1 cm-crystal length and a 330 μm fundamental beam waist we obtained 20 mW of 532 nm Gaussian beam. The walk-off between the fundamental and the SHG beams is virtually undetectable. To perform SFG by mixing the fundamental and the SHG beams, non-critical phase matching conditions cannot be reached. In addition the choice of commercially available crystals is very limited, so we used a 1 cm-long Type II LBO crystal (Θ = 42.2 and φ = 90) and we focused the fundamental beam to a 100 μm waist. The beam at 355 nm has a power of 2 mW with a near Gaussian profile (M2 = 1.1), while the 532 nm beam is unaffected in power, profile and polarization. Unfortunately this crystal introduces a small lateral shift between the 355 and the 532 nm beams, due to walk-off, of approximately 40 μm which is enough to prevent the two colors from perfectly overlapping in the objective focal plane. Hopefully, the large difference of wavelengths between the two beams and the spectral dispersion of refractive index of the silica allow the compensation of the lateral shift by a simple silica slab as shown in Figure. It is easy to calculate that, at Brewster incidence, a 5 mm-thick slab is enough to compensate the 40 μm lateral shift between the 532 and 355 nm beams.