Proceedings Article | 18 October 2019
KEYWORDS: Frequency conversion, Heteroepitaxy, Laser sources, Mid-IR, Long wavelength infrared, Nonlinear optics, Infrared radiation, Gallium arsenide, Crystals, Missiles
Due to advances in the heat seeking technology combat loses from heat seeking missiles have exceeded 95%. That is why today the need for compact, broadly tunable mid and longwave IR sources for IR countermeasures is more urgent than ever. Laser radar, high speed IR communications and remote sensing of chemical and biological agents are other military and security applications of such sources. Many other commercial applications in areas such as environmental sensing, industrial production, spectroscopy and medicine also encourage their development. The existing direct laser sources in these wavebands have low output power, limited wavelengths and poor tunability. That is why efforts into development of frequency conversion devices, especially those that are based on quasi-phase matching have been made to provide alternatives. In non-polar materials frequency conversion can be achieved in thick periodic structures grown on OP templates. In this work we report several studied OP materials, the most mature of which is OPGaAs where we have achieved 60% conversion efficiency. We show how these materials achieve their limits bringing us to the idea of combining them by heteroepitaxy (especially when native substrates are unavailable) or by growing ternaries attempting to combine their best NL properties. For example, GaAs exhibits strong two-photon absorption (2PA) in the convenient pumping range of 1–1.7 µm, which easily deteriorates the OPGaAs device performance. Replacing GaAs with GaP, which nonlinear susceptibility is similar but its 2PA in the same frequency range is with 3 orders of magnitude lower, is soon followed by disappointments related to the high price and low quality of the commercially available GaP substrates. Fortunately, the attempts to grow heteroepitaxially GaP on GaAs and OPGaP on the high quality OPGaAs templates led unexpectedly to good results in spite of the relatively large lattice mismatch (-3.26%) between GaP and GaAs. More heteroepitaxial cases (some more favorable) such as ZnSe/GaAs (+0.238 %), α-GaSe/GaP (-0.607 %), etc. were also attempted, leading by the fact that in some of these cases large enough crystalline substrates are not available. Although that the state-of-art results – smooth surface morphology (1-2 nm in 5x5 µm AFM scanning spot), high crystalline quality (XRD: FWHM within 50-60 arcsec) and excellent domain fidelity for the growth on OP-templates were achieved with the GaP/GaAs growths, growth of ZnSe/GaAs may also be considered as successful (with the remark that the high vapor pressure of the Zn-source restricts the duration of growth. There is much more to ask for the growths of GaSe where the growth of the cubic α-phase GaSe/GaP were more successful than the growth on the hexagonal β-phase GaSe/GaN. High resolution TEM images of the interface allowed to learn more about the growth mechanisms. For example, during fast growth processes (HVPE) the strain built due to the lattice and the thermal mismatch relaxes faster by formation of voids or roughening of surfaces, which may postpone the formation of the typical misfit dislocations observable during slow growths (MOCVD, MBE). All this allowed to predict success for other heteroepitaxial cases – for instance ZnTe/GaSb.
