To observe the 2nd THz-wave generated in the cascade process, we focused on the angular difference between the 1st and 2nd THz-waves. In is-TPG, the 2nd THz-wave has a different generation angle from the 1st THz-wave due to the noncollinear phase matching condition. For example, when the 1st THz-wave is 1.37 THz, the 2nd THz-wave is spatially separated from the 1st THz-wave and focused through a lens because of the angular difference of about 0.6° inside the crystal and about 3.7° outside the crystal. Therefore, as shown in Fig. 1 [c(I)], a pyroelectric detector with a ⌀ 1 mm aperture was raster-scanned in the y-z plane at the focal point to obtain the spatial distribution and to observe the 1st and 2nd THz-waves. The measurement results are shown in Fig. 2. Two beam components were observed 6.8 mm apart in the y-axis direction. The difference of the generation angles of the two beam components was 3.9°, which is almost consistent with the theoretical value of 3.7° derived from the phase matching condition.
However, this result alone leaves open the possibility that the THz-wave with different generation angle is not 2nd THz-wave but parametric fluorescence or stray light of 1st THz-wave. Parametric fluorescence is a type of quantum noise which is the THz component without seeding. Since this parametric fluorescence is broadband at angles that satisfy the noncollinear phase matching condition, it is spatially separated from the 1st THz-wave in the setup shown in Fig. 1. On the other hand, 2nd THz-wave has the same frequency as the 1st THz-wave because it is generated by the 1st THz-wave as injected seed [18]. Therefore, to confirm whether the THz-wave with different generation angle is 2nd THz-wave or parametric fluorescence, the frequency spectrum was measured using a scanning Fabry-Perot etalon with two metal meshes as shown in Fig. 1 [c (II)]. By sweeping the metal mesh spacing by an automated stage and obtaining interference waveforms, the frequency can be measured from the spacing of the interference peaks. In this setup, both 1st THz-wave and "THz-wave with different generation angle" were injected into the scanning etalon and measured with the same pyroelectric detector. Two identically spaced interference waveforms were observed as shown in Fig. 3. The frequency of the two components calculated from the peak spacing was confirmed to be the same, 1.37 THz. The larger amplitude of the two observed interference waveforms is considered to be 1st THz-wave, while the smaller amplitude is considered to be THz-wave with a different generation angle. This result confirms that the THz-wave with different generation angle is not a parametric fluorescence, since it is narrowband and has the same frequency as 1st THz-wave.
Next, we confirmed that THz-wave with different generation angle is not stray light of the 1st THz-wave. Since stray light of the 1st THz-wave due to some kind of diffraction or refraction has a different generation angle and the same frequency as the 1st THz-wave, it cannot be reliably distinguished from the 2nd THz-wave in the experiment described above. Therefore, we focused on the oscillation threshold of the THz-wave. Since the 1st and 2nd THz-waves are generated in pairs with the 1st and 2nd Stokes beams, respectively, we compared the input-output characteristics of the THz and Stokes generation. Figure 4 shows the input-output characteristics measured after spatially separating the THz-waves in the setup of Fig. 1 [c (III)]. The THz-wave was observed with a pyroelectric detector and the Stokes beam was observed with a near-infrared energy meter. Figure 4 (a) and (b) show the input/output characteristics of the Stokes beam and THz-wave, respectively. The 2nd Stokes and "THz-waves with different generation angle" have close oscillation thresholds.
From these results, it was confirmed that the THz-wave with different generation angle is 2nd THz-wave, and it was successfully observed for the first time. Furthermore, as shown in Fig. 5, the combined extraction of the 1st and the 2nd THz-waves improves the output by about 20%, i.e., higher output power due to the higher-order THz-wave was confirmed. This 20% value is consistent with the result predicted from the Stokes intensity in a previous study [11]. The 1st and the 2nd THz-waves can be focused to almost the same point by changing the design of the Si prism coupler. Furthermore, the 2nd THz-wave can be used as a reference for calibrating the 1st THz-wave because the pulse intensity of the 2nd is proportional to that of the 1st.