メニュー
Calculate the composition (concentration) of atom probe data. Handles range extraction, ion allocation, and concentration calculation. Use when the user asks about composition, concentration, or chemical analysis.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
| name | apt-composition |
| description | Calculate the composition (concentration) of atom probe data. Handles range extraction, ion allocation, and concentration calculation. Use when the user asks about composition, concentration, or chemical analysis. |
| argument-hint | [instrument] |
Calculate the bulk composition from the current mass spectrum and pos data. Use the MATLAB MCP server.
The MATLAB workspace must contain:
pos — loaded atom probe data (from /apt-load)spec — mass spectrum figure with ions and ranges defined (from /apt-spectrum + rangeAddAll)If ranges haven't been defined yet, tell the user they need to define ranges first:
rangeAddAll(spec, colorScheme, 0.04, true);
rangeTable = rangesExtractFromMassSpec(spec);
disp(rangeTable);
pos = posAllocateRange(pos, rangeTable, 'decomposed');
Use 'decomposed' mode to break molecular ions into constituent atoms. If the user specifically wants molecular ion identities, use 'direct' instead.
If the user provided an instrument as $ARGUMENTS, map it to detection efficiency:
Otherwise, read it from the config:
cfg = APTConfig.getInstance();
detEff = cfg.reconstruction.detectionEfficiency;
If the detection efficiency is still the default (0.37) and the user hasn't confirmed their instrument, ask which instrument was used.
conc = posCalculateConcentrationSimple(pos, detEff, {'unranged'}, '', 'mode', 'atomic');
disp(conc);
Exclude 'unranged' atoms by default. If the user wants to include them or exclude other species, adjust the exclude list.
unc = concentrationUncertainty(conc.counts);
Present results as a formatted table showing:
Ask if the user wants to:
writetable(conc, 'composition.csv')Voronoi-cell cluster determination (Felfer et al. 2015) — detect and extract clusters via Voronoi-cell volumes, Delaunay triangulation, and a Kolmogorov-Smirnov test against a randomised reference. Use when the user wants the Voronoi/clusterDetermination method specifically (this is NOT DBSCAN — for density-based DBSCAN use /apt-cluster instead).
Create a new ion color scheme or add ions/colors to the existing one. Produces/updates the `colorScheme` workspace variable used by ranging and 3D visualization everywhere else. Use when the user wants custom ion colors, a new color scheme, or to add a missing ion color.
Per-vertex concentration mapping onto a mesh/interface — compute atom-to-mesh distances, clip a slab around the interface, voxelise by nearest vertex, and apply a concentration kernel to colour each vertex. Use when the user wants a concentration map painted onto an isosurface or modelled grain boundary mesh.
Run quality-control checks on an atom probe dataset before analysis — basic stats, voltage curve, detector hit map (FDM), mass spectrum sanity, coordinate ranges, multi-hit fraction, and automated quality metrics. Use when the user wants to assess data quality or sanity-check a reconstruction.
Create a field desorption map (detector hit-density / concentration image) and optionally a movie of detector activity over the course of the experiment. Use when the user wants an FDM, desorption map, detector image, or to watch features evolve during the measurement.
Quantify interfacial excess (Gibbsian excess of solute, in at/nm^2) across an interface — either a single value via the Krakauer-Seidman integral method, or a 2D excess map over the surface. Use when the user wants the amount of an element segregated to an interface or grain boundary.