在能带之间， 重新定义价值。 Engineering the gap.

2026 年起，仅中国每年产生 50 万吨退役光伏组件；2040 年达到 200 万吨；2050 年全球累计 1.6 亿吨。 From 2026, China alone generates 0.5M tons of retired PV panels per year, rising to 2M by 2040 and 160M cumulative globally by 2050.

现有处置只有两条路：拆解卖材料（毛利 < 5%、能耗高、化学品污染）或低价梯次利用（接受度极低、电站运营商拒绝采购）。本章对这两条路径分别建立失效模型，定量估计其经济边界。 Existing pathways — material recovery (margin <5%, high-energy, chemical-heavy) and cascade reuse (rejected by operators) — are formally modelled with quantitative failure boundaries.

跳出"工艺成本博弈"，把退役光伏的高值化重新定义为能带间隙工程问题。这一物理学起点贯穿后续所有章节，是整个研究范式的逻辑锚点。 Step outside the "process-cost game" and reframe high-value remanufacturing as a bandgap engineering problem. This physics-first starting point anchors every subsequent chapter.

本章首先建立失效模式与能带特征演化的对应关系：PID 与 EVA 黄化会分别导致少子寿命和 380-500 nm 反射率出现下降，热斑/断栅/微裂纹局部影响 EQE。 The chapter first maps failure modes to bandgap signature evolution: PID and EVA yellowing respectively reduce minority-carrier lifetime and 380-500 nm reflectance, hotspots/broken fingers/microcracks locally compromise EQE.

规划构建 BgDB v1.0：目标覆盖多省多家电站 3,000+ 块退役组件，每块含 5 模态原始数据（PL、EL、反射光谱、椭偏、IR）。计划 12 个月内完成首批采样，成为国内首个公开的退役晶硅组件能带特征数据集。 Planned construction of BgDB v1.0: target coverage of 3,000+ retired panels from multiple stations across multiple provinces, each with 5 modalities (PL, EL, reflectance, ellipsometry, IR). First-batch sampling targeted within 12 months, aiming to become China's first publicly available bandgap-feature dataset for retired c-Si modules.

BgClassNet 准确率 ≥90%（目标）。 BgClassNet accuracy ≥90% (target).

本章合成黄、橙、红三系列全无机钙钛矿量子点。 Three all-inorganic perovskite QD series synthesised (yellow, orange, red).

研究局限性诚实披露：BgDB 数据集主要覆盖国内 4 省，海洋性 / 热带湿热气候数据不足；薄膜涂层主要针对反射型，透射型 BIPV 设计未覆盖；钙钛矿真实户外稳定性需实地长期监测；产线良率提升与跨气候带应用扩展仍待实践检验。 Honest limitations: BgDB covers 4 provinces, missing maritime and tropical climates; coatings target reflective only, not transmissive BIPV; real-world QD durability still needs field validation; mainnet, liquidity bootstrap, and DAO transition remain unproven in practice.

四个未来研究方向：(1) 多机制协同——钙钛矿叠层 + QD + 薄膜协同；(2) 智能化——基于 BgDB 的扩散模型直接生成工艺参数；(3) 材料——无铅替代方案；(4) 社会经济——量化对中国千万吨级退役潮的政策协同效应。 Four future directions: (1) multi-mechanism synergy — perovskite tandem + QD + coatings; (2) AI-driven process generation via diffusion models trained on BgDB; (3) Pb-free materials; (4) quantified social-economic impact at China's tens-of-millions-of-tons scale.