不存在稳定岛——这是一个可证伪的硬预测
No island of stability — this is a falsifiable hard prediction
元素周期表的结构——为什么是2、8、18、32个元素一行，为什么特定轨道有特定容量——在量子力学里是推导出来的，但用的是经验型势函数（Woods-Saxon势），不是第一性原理。
SAE从DD结构推导壳层容量：
1DD贡献n_dual = 2（自旋，即L/R对偶）；2DD贡献角量子数l的允许取值数n；3DD贡献角向简并度(2l+1)。因此subshell容量 = 2(2l+1)，shell容量 = 2n²。
Madelung填充规则（为什么4s先于3d填充）通常以"n+l最小者先填"的经验规则描述。SAE的解释：n+l = 2DD加法成本 + 3DD乘法成本 = 总拓扑成本。不需要经验输入。
Cr和Cu的Madelung例外（预期d⁴s²却实际是d⁵s¹）来自d⁵ = 5个doublet各一个电子时的完美对称（n_doublets共振）。Gd的例外来自f⁷的n_shells+1共振。SAE实现零未解释例外。
铅(Pb, Z=82)是自然界中最重的稳定元素。为什么是82？
82 = n_dual × (N_1DD − 1)：双侧对外泄漏通道饱和。铅之后的Po(84) = n_dual × N_1DD：含自泄漏的全部通道满额。Hs(108) = d × n_axes³ = 67+41：全部泄漏通路饱和。
Magic numbers的DD身份：核物理的幻数{2, 8, 20, 28, 50, 82, 126}的差值序列{6, 12, 8, 22, 32, 44}全部有DD身份，其中11 = N_blocks − 1（泄漏给其他11个block）。
关键点：幻数不是壳层势函数的输出（可以外推），而是DD泄漏通道的饱和序列（有硬边界）。
核物理学界长期预测，在Z≈114-126（超重元素区）存在一片"稳定岛"——因为传统核壳层模型预测这里有幻数，稳定性应该回升。
SAE给出一个与标准核物理直接对立的硬预测：Z>108不存在任何自持稳定的同位素（半衰期>1年）。所谓的稳定岛不存在。
理由：幻数不是传统核壳层势的输出，可以无限外推；幻数是DD泄漏通道的饱和序列，在Z=108处（全部泄漏通路饱和）有硬边界。超过这个边界，不再有结构性稳定机制。
这是一个可证伪的预测。目前Z=115(Mc)最长寿同位素半衰期约220毫秒，Z=117(Ts)约51毫秒——远低于1年。如果实验室合成出Z>108的稳定同位素，SAE的泄漏通道饱和理论被证伪。
预测还未被证伪。但它随时可以被证伪。这是好的科学。
The structure of the periodic table — why 2, 8, 18, 32 elements per row, why specific orbitals have specific capacities — is derived in quantum mechanics, but using an empirical potential function (the Woods-Saxon potential), not first principles.
SAE derives shell capacities from DD structure:
1DD contributes n_dual = 2 (spin, the L/R duality); 2DD contributes the number of allowed angular quantum number l values; 3DD contributes the angular degeneracy (2l+1). Therefore subshell capacity = 2(2l+1), shell capacity = 2n².
The Madelung filling rule (why 4s fills before 3d) is usually stated as the empirical rule "lowest n+l first." SAE's explanation: n+l = 2DD additive cost + 3DD multiplicative cost = total topological cost. No empirical input needed.
The Madelung exceptions for Cr and Cu (expected d⁴s², actual d⁵s¹) arise from perfect symmetry when d⁵ has exactly one electron per doublet (n_doublets resonance). Gd's exception comes from f⁷'s n_shells+1 resonance. SAE achieves zero unexplained exceptions.
Lead (Pb, Z=82) is the heaviest stable element in nature. Why 82?
82 = n_dual × (N_1DD − 1): bilateral external leakage channels saturated. The next element, Po(84) = n_dual × N_1DD: all channels including self-leakage at full capacity. Hs(108) = d × n_axes³ = 67+41: all leakage pathways saturated.
DD identity of magic numbers: the difference sequence {6, 12, 8, 22, 32, 44} of nuclear magic numbers {2, 8, 20, 28, 50, 82, 126} all have DD identities, where 11 = N_blocks − 1 (leakage to the other 11 blocks).
Key point: magic numbers are not outputs of shell potential functions (which can be extrapolated indefinitely), but saturation sequences of DD leakage channels (with a hard boundary).
Nuclear physicists have long predicted a "island of stability" around Z≈114-126 — because the traditional nuclear shell model predicts magic numbers there, suggesting stability should increase.
SAE makes a hard prediction directly opposed to standard nuclear physics: No self-sustaining stable isotope (half-life >1 year) exists for Z>108. The island of stability does not exist.
The reason: magic numbers are not outputs of a traditional nuclear shell potential that can be infinitely extrapolated; they are the saturation sequence of DD leakage channels, with a hard boundary at Z=108 (all leakage pathways saturated). Beyond this boundary, no structural stability mechanism remains.
This is a falsifiable prediction. Currently, Z=115 (Mc)'s longest-lived isotope has a half-life of ~220 milliseconds; Z=117 (Ts) is ~51 milliseconds — far below one year. If a stable isotope with Z>108 is synthesized, SAE's leakage channel saturation theory is falsified.
The prediction has not yet been falsified. But it can be falsified at any time. That is good science.