Run any Skill in Manus with one click

apt-reconstruct

Stars7

Forks5

UpdatedJune 24, 2026 at 13:55

Reconstruct 3D atom positions (x, y, z) from detector hit coordinates and the voltage curve using the Gault et al. straight-flight-path algorithm. Use when the user wants to reconstruct, build, or recompute the 3D point cloud from raw detector data.

Installation

Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.

Run Skill in Manus

Source

peterfelfer

peterfelfer/Atom-Probe-Toolbox

View GitHub Repository View Creator Repositories

Download

Run Skill in Manus

Related occupationsSOC

Based on SOC occupation classification

Biological Scientists, All OtherLife, Physical, and Social Science Occupations·SOC 19-1029

SKILL.md

readonly

name	apt-reconstruct
description	Reconstruct 3D atom positions (x, y, z) from detector hit coordinates and the voltage curve using the Gault et al. straight-flight-path algorithm. Use when the user wants to reconstruct, build, or recompute the 3D point cloud from raw detector data.
argument-hint	[element]

Reconstruct the 3D atom positions from detector hit coordinates and the voltage curve, following Gault et al., Ultramicroscopy 111 (2011) 448-457. Use the MATLAB MCP server.

Prerequisites

The MATLAB workspace must contain:

pos — atom probe data with detector coordinates and the voltage curve: columns detx, dety (detector hit position in nm) and VDC (standing voltage). These exist only in extended formats (.epos / .apt / pyccapt HDF5), not in plain .pos. If only a .pos is loaded, tell the user to reload an .epos/.apt/HDF5 file via /apt-load first.

This skill overwrites pos.x, pos.y, pos.z with the newly computed reconstruction.

Steps

1. Set reconstruction parameters

Instrument geometry and detection efficiency come from the instrument the data was acquired on. If cfg is available you can read defaults from it (cfg = APTConfig.getInstance(); cfg.reconstruction.detectionEfficiency, cfg.reconstruction.flightPath), otherwise ask the user which instrument was used and use the toolbox presets (LEAP4000XHR detEff 0.37 / flight 110, LEAP5000XR 0.52 / 120, LEAP5000XS 0.80 / 120, EIKOS 0.50 / 100).

reconPara.detEff      = 0.36;   % detector efficiency (0..1)
reconPara.flightLength = 110;   % flight path in mm
reconPara.evapF       = 32.49;  % evaporation field in V/nm
reconPara.kf          = 3.3;    % field factor
reconPara.ICF         = 1.4;    % image compression factor
reconData = table;

evapF, kf and ICF are material/calibration dependent; the values above are the workflow defaults. Ask the user if they have calibrated values, otherwise keep these.

2. Average atomic density

Either set it directly (atoms/nm^3) or derive it from the main element of the sample. $ARGUMENTS is the element symbol (default Fe):

load isotopeTable_naturalAbundances.mat;
element = "Fe";   % use $ARGUMENTS if provided
reconPara.avgDens = mean(isotopeTable.atomDensity(isotopeTable.element == element));
clearvars element isotopeTable

If the user knows the density directly, skip this and set reconPara.avgDens = 85; (example value).

3. Detector polar coordinates and radius evolution

[reconData.ang, reconData.rad] = cart2pol(pos.detx, pos.dety);
reconPara.areaDet = (max(reconData.rad))^2 * pi();      % effective detector area
reconData.radEvol = pos.VDC/(reconPara.kf * reconPara.evapF);   % specimen radius (nm)

4. Calculate x and y coordinates

reconData.thetaP = atan(reconData.rad / reconPara.flightLength);
reconData.theta  = reconData.thetaP + asin((reconPara.ICF - 1) * sin(reconData.thetaP));
[reconData.zP, reconData.distance] = pol2cart(reconData.theta, reconData.radEvol);
[pos.x, pos.y] = pol2cart(reconData.ang, reconData.distance);   % nm

5. Calculate z coordinate

reconData.zP   = reconData.radEvol - reconData.zP;            % z shift from cap top
reconPara.omega = 1 ./ reconPara.avgDens;
reconData.dz   = reconPara.omega * reconPara.flightLength^2 * reconPara.kf^2 * reconPara.evapF^2 ...
                 / (reconPara.detEff * reconPara.areaDet * reconPara.ICF^2) * pos.VDC.^-2;
reconData.cumZ = cumsum(double(reconData.dz));               % accumulative depth
pos.z = reconData.cumZ + reconData.zP;                       % nm

Report

Tell the user the reconstruction is complete and report the new coordinate ranges (min/max of pos.x, pos.y, pos.z) and the parameters used (detEff, flightLength, evapF, kf, ICF, avgDens). The intermediate values live in the reconData table and reconPara struct if they want to inspect them.

Suggest follow-ups:

Visualize the result: /apt-visualize
Build the mass spectrum: /apt-spectrum
Export the reconstructed data: posExport(pos, 'reconstructed.pos')
Adjust kf / ICF and re-run if the aspect ratio or feature spacing looks wrong (these are the main tuning knobs).

Toolbox functions used

cart2pol, pol2cart, cumsum (MATLAB built-ins); reconstruction is implemented inline per the Gault algorithm. Optional: APTConfig.getInstance, posExport, posLoad.

More from this repository

same repository

apt-cluster-voronoi

peterfelfer/Atom-Probe-Toolbox

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).

2026-06-247

apt-colorscheme

peterfelfer/Atom-Probe-Toolbox

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.

2026-06-247

apt-concentration-kernel

peterfelfer/Atom-Probe-Toolbox

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.

2026-06-247

apt-data-quality

peterfelfer/Atom-Probe-Toolbox

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.

2026-06-247

apt-fdm

peterfelfer/Atom-Probe-Toolbox

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.

2026-06-247

apt-interfacial-excess

peterfelfer/Atom-Probe-Toolbox

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.

2026-06-247

name	apt-reconstruct
description	Reconstruct 3D atom positions (x, y, z) from detector hit coordinates and the voltage curve using the Gault et al. straight-flight-path algorithm. Use when the user wants to reconstruct, build, or recompute the 3D point cloud from raw detector data.
argument-hint	[element]

Reconstruct the 3D atom positions from detector hit coordinates and the voltage curve, following Gault et al., Ultramicroscopy 111 (2011) 448-457. Use the MATLAB MCP server.

Prerequisites

The MATLAB workspace must contain:

pos — atom probe data with detector coordinates and the voltage curve: columns detx, dety (detector hit position in nm) and VDC (standing voltage). These exist only in extended formats (.epos / .apt / pyccapt HDF5), not in plain .pos. If only a .pos is loaded, tell the user to reload an .epos/.apt/HDF5 file via /apt-load first.

This skill overwrites pos.x, pos.y, pos.z with the newly computed reconstruction.

Steps

1. Set reconstruction parameters

reconPara.detEff      = 0.36;   % detector efficiency (0..1)
reconPara.flightLength = 110;   % flight path in mm
reconPara.evapF       = 32.49;  % evaporation field in V/nm
reconPara.kf          = 3.3;    % field factor
reconPara.ICF         = 1.4;    % image compression factor
reconData = table;

evapF, kf and ICF are material/calibration dependent; the values above are the workflow defaults. Ask the user if they have calibrated values, otherwise keep these.

2. Average atomic density

Either set it directly (atoms/nm^3) or derive it from the main element of the sample. $ARGUMENTS is the element symbol (default Fe):

load isotopeTable_naturalAbundances.mat;
element = "Fe";   % use $ARGUMENTS if provided
reconPara.avgDens = mean(isotopeTable.atomDensity(isotopeTable.element == element));
clearvars element isotopeTable

If the user knows the density directly, skip this and set reconPara.avgDens = 85; (example value).

3. Detector polar coordinates and radius evolution

[reconData.ang, reconData.rad] = cart2pol(pos.detx, pos.dety);
reconPara.areaDet = (max(reconData.rad))^2 * pi();      % effective detector area
reconData.radEvol = pos.VDC/(reconPara.kf * reconPara.evapF);   % specimen radius (nm)

4. Calculate x and y coordinates

reconData.thetaP = atan(reconData.rad / reconPara.flightLength);
reconData.theta  = reconData.thetaP + asin((reconPara.ICF - 1) * sin(reconData.thetaP));
[reconData.zP, reconData.distance] = pol2cart(reconData.theta, reconData.radEvol);
[pos.x, pos.y] = pol2cart(reconData.ang, reconData.distance);   % nm

5. Calculate z coordinate

reconData.zP   = reconData.radEvol - reconData.zP;            % z shift from cap top
reconPara.omega = 1 ./ reconPara.avgDens;
reconData.dz   = reconPara.omega * reconPara.flightLength^2 * reconPara.kf^2 * reconPara.evapF^2 ...
                 / (reconPara.detEff * reconPara.areaDet * reconPara.ICF^2) * pos.VDC.^-2;
reconData.cumZ = cumsum(double(reconData.dz));               % accumulative depth
pos.z = reconData.cumZ + reconData.zP;                       % nm

Report

Suggest follow-ups:

Visualize the result: /apt-visualize
Build the mass spectrum: /apt-spectrum
Export the reconstructed data: posExport(pos, 'reconstructed.pos')
Adjust kf / ICF and re-run if the aspect ratio or feature spacing looks wrong (these are the main tuning knobs).

Toolbox functions used

cart2pol, pol2cart, cumsum (MATLAB built-ins); reconstruction is implemented inline per the Gault algorithm. Optional: APTConfig.getInstance, posExport, posLoad.