AI 中文总结
本研究通过GW/BSE方法探究不同尺寸CdSe纳米团簇的准粒子与激子性质，验证了第一性原理衍生的紧束缚模型，揭示光谱蓝移的成因，为半导体纳米结构光学预测提供多尺度路线图。
AI 中文摘要
化学计量比$\boldsymbol{\text{Cd}_n\text{Se}_n}$纳米团簇中的量子限域效应会大幅削弱电子屏蔽，驱动准粒子自能修正（$\boldsymbol{\triangle_{\text{QP}}}$）与激子结合能（$\boldsymbol{E_b}$）之间产生微妙的尺寸依赖竞争关系。本文针对具有代表性的尺寸系列（$\boldsymbol{n = 3, 6, 13, 33}$）开展$\boldsymbol{GW/BSE}$研究，并利用该研究结果验证了一种源自第一性原理的可扩展原子紧束缚（$\boldsymbol{TB}$）框架。研究结果表明，文献中先前报道的1-2 eV光谱蓝移源于单粒子$\boldsymbol{GW}$收敛伪影，而非电子-空穴核的缺陷。研究显示，随着团簇体积增大，$\boldsymbol{\triangle_{\text{QP}}}$与$\boldsymbol{E_b}$之间近乎完美的抵消关系会被打破，原因是介电屏蔽快速出现，使$\boldsymbol{E_b}$的衰减速度快于$\boldsymbol{\triangle_{\text{QP}}}$，进而与平均场预测产生显著偏差。对电子结构的空间逆参与率分析表明，基频预峰的光学抑制源于局域价轨道与离域导态之间的严重空间不匹配。最后，研究证实，限域诱导的准粒子带隙和光学 onset 的标度关系可通过经剪刀修正、DFT参数化的$\boldsymbol{TB}$模型准确复现。因此，本研究为将有效多体效应核嵌入计算高效模型提供了定量多尺度路线图，可对包含多达数千个原子的真实半导体纳米结构进行可靠的光学预测。
英文摘要
Quantum confinement in stoichiometric $\mathrm{Cd}_n\mathrm{Se}_n$ nanoclusters dramatically attenuates electronic screening, driving a delicate, size-dependent competition between quasiparticle self-energy corrections ($Δ_{\mathrm{QP}}$) and exciton binding energies ($E_b$). Here, we present a $GW$/BSE study across a representative size series ($n = 3, 6, 13, 33$) and leverage it to validate a scalable atomistic tight-binding (TB) framework derived from first principles. Our results demonstrate that 1-2 eV spectral blueshifts previously reported in the literature arise from single-particle $GW$ convergence artifacts rather than deficiencies in the electron--hole kernels. We show that the near-perfect cancellation between $Δ_{\mathrm{QP}}$ and $E_b$ breaks down as cluster volume increases, driven by the rapid onset of dielectric screening attenuating $E_b$ faster than $Δ_{\mathrm{QP}}$ and leading to a pronounced divergence from mean-field predictions. Spatial inverse participation ratio analysis of the electronic structure reveals that optical suppression of fundamental pre-peaks stems from a severe spatial mismatch between localized valence orbitals and delocalized conduction states. Finally, we demonstrate that the confinement-induced scaling of the quasiparticle gap and the optical onset is accurately reproduced by a scissor-corrected, DFT-parameterized TB model. As such, this work provides a quantitative multiscale roadmap for embedding effective many-body effects kernels into computationally efficient models, enabling reliable optical predictions for realistic semiconducting nanostructures containing up to thousands of atoms.
