Optical and Miller Model Datasets for 63 NLO Crystals and KCuMoS4

Nonlinear optical (NLO) phenomena play a pivotal role in materials research and technological advancements, particularly for optoelectronic modulation and advanced photonic devices. Accurate evaluation of intrinsic NLO responses is crucial for targeted material discovery and performance optimization. However, conventional NLO coefficients (χijk(2)) exhibit strong bandgap (Eg) dependence, fundamentally distorting their correlation with actual conversion efficiency (η). This limitation not only impedes fair performance comparisons across materials with different Eg values but also artificially amplifies the NLO capabilities of narrow-Eg systems. To address this critical challenge, we propose an intrinsic NLO metric φijk=δijk·χii(1), incorporating a static Miller dispersion correction (δijk) that enables cross-Eg evaluations and reveals structural limits of achievable η through fundamental parameter optimization. Our first-principles calculations, performed using our in-house NLOtools, demonstrate the advantage of φijk over χijk(2) in decoupling Eg effects while establishing a robust framework for rational design. This framework is demonstrated through the rational design of KCuMoS4, where targeted skeletal reconstruction and elemental substitution yielded a superior intrinsic NLO potential, confirmed by its concurrently high φijk and normalized Φijk values. The favorable performance of KCuMoS4 illustrates a route to mitigating the classic trade-off between χijk(2) and Eg, providing a practical methodology for the rational design of high-performance NLO materials.