Here are some info I have found. Below are the statements in the articles.

As an important part of modern optics and photonics, nonlinear optics is of great significance for photonic applications such as optical harmonic generation [1,2], ultrafast light switching [3], optical imaging [4] and optical data storage applications [5,6]. Yet, the inherent nonlinear response of materials is usually weak and difficult to control. The existing nonlinear optical crystals improve the nonlinear conversion efficiency by increasing the length of the interaction. Researchers have also taken a series of methods to improve the nonlinear properties of materials, but the designs are often more complex and less effective. It is urgent to look for materials with high stability and huge nonlinear optical (NLO) response. https://doi.org/10.1016/j.optmat.2021.110841

Another example: if the optical properties of nonlinear oxides are repeatable, this high level of accuracy would be unnecessary. It is shown that MBE affords unique advantages, particularly in the crucial areas of stoichiometry control and deposition uniformity, while offering significant challenges in areas such as growth rate and, in limited cases, interfacial chemistry control. https://doi.org/10.1116/1.1926294

The observation that with increasing particle size SA becomes more pronounced is consistent with earlier measurements on Ag and Au nanoclusters and NPs. (22, 34, 62) This effect is understood in terms of size dependence of the electronic structure of the particles. For very small particles the discrete nature of their electronic states hinders plasmonic excitations, reducing considerably their linear absorption and, thus, corresponding to very high saturation irradiance for the particle’s first excited state. https://doi.org/10.1021/acs.jpcc.7b09017