尚未检测到更新。GitHub 发布版本后会显示在这里。
个人 GitHub 账号下公开仓库的所有者可以登录并验证所有权,自行管理介绍,无需管理员批准。
登录并生成验证文件,在 30 分钟内将其提交到此仓库的默认分支。其他分支或分叉仓库中的文件无效。然后在下方检查文件,即可无需管理员批准直接管理项目。请保留文件以维持管理权限。
如需登录,将前往登录页面。登录后请回到此处,再次点击所选的验证按钮。
在项目文件夹中运行 /nomorevibe verify,即可验证所有权并自行管理介绍。
一分钟完成验证Interactive energy level (Grotrian) diagrams of neutral atoms: wavelengths, dipole matrix elements, lifetimes, hyperfine data from NIST and measurements
翻译准备中
取自仓库 README
正在加载完整 README。 以下为已保存的摘录。
Interactive Grotrian diagrams for the periodic table. The landing page is a clickable periodic table: 90 elements have a page built from the NIST Atomic Spectra Database (strongest classified lines below 2 µm, with wavelengths and dipole matrix elements derived from NIST transition rates). Twelve elements used in cold-atom, tweezer, clock and ion-trap labs (Li, Be, Na, Mg, K, Ca, Rb, Sr, Cs, Ba, Yb, Dy; 22 isotopes) additionally carry measured lifetimes, transition rates, hyperfine constants and isotope shifts with citations.查看完整 README ↗
Interactive site: https://muuuun.github.io/atom-energy-levels/
Interactive Grotrian diagrams for the periodic table. The landing page is a clickable periodic table: 90 elements have a page built from the NIST Atomic Spectra Database (strongest classified lines below 2 µm, with wavelengths and dipole matrix elements derived from NIST transition rates). Twelve elements used in cold-atom, tweezer, clock and ion-trap labs (Li, Be, Na, Mg, K, Ca, Rb, Sr, Cs, Ba, Yb, Dy; 22 isotopes) additionally carry measured lifetimes, transition rates, hyperfine constants and isotope shifts with citations. On every page: hover a line or a level to isolate it, click to pin, filter by transition type (E1, intercombination, M1, E2, M2, Rydberg).
python3 fetch_all.py # download / cache the NIST ASD tables of every neutral atom
python3 compute.py all # (or: curated | auto | <key> ...) NIST ASD + data/literature/*.json (+ ARC, pairinteraction for alkalis) -> data/<key>/atom.json, *.csv
python3 plot.py all # -> docs/<slug>/diagram.svg|pdf, preview.png, data.json
python3 build_site.py # -> docs/**/index.html, sitemap.xml
python3 stats.py # -> data/provenance.md (which source supplied how many numbers)
species.py lists the curated species; every other element gets an automatic NIST-only entry. Adding a curated one is a dictionary entry: kind="nist" needs only the element symbol, an energy
cut and the isotope; kind="alkali" additionally uses ARC so that every E1-allowed pair gets a matrix element.
Priority for every number: measurement (data/literature/<El>.json, each value with citation and URL) > NIST ASD >
high-accuracy theory quoted in the literature > ARC model potential (alkalis only). Each value carries its tier in the
CSV files and on the site (measured, NIST, theory, model calc.); on the diagrams * marks theory and ≈ a model value.
| Quantity | Source |
|---|---|
| Level energies, wavelengths, frequencies | NIST ASD level energies; measured isotope-specific frequencies where available |
| Transition rates, matrix elements | measured > NIST ASD > all-order theory (UDel portal, Safronova et al.) > ARC |
| Lifetimes | measured > theory > ARC sum of rates (alkalis) |
| Hyperfine constants, isotope shifts | measurements (Allegrini et al. 2022 survey, Steck, original papers) |
| Rydberg term energies | pairinteraction (Rb); ARC quantum defects hung from the NIST ionisation limit (other alkalis) |
Current counts are in data/provenance.md; per-species cross-checks in data/<key>/validation.txt.
Known caveats: NIST level energies refer to the natural isotope mixture, so wavelengths of minority isotopes are off by the
isotope shift (listed where measured); conflicts between sources are recorded in the notes / conflicts fields of the
literature files; some old lifetime measurements (Cs n ≥ 8) are probably less accurate than modern theory but are kept as primary.