Correct a run
Upload one sequence as Isodat exported it. The file is processed on the server and discarded; nothing is stored. Every correction is computed from the run’s own standards, and each choice you make on the steps that follow is recalculated in this browser.
- {{k.t}}
Columns
{{up.mapNote}}{{up.mapFoot}}
The page keeps the export in this tab only, to send it again when you download the workbook or the results table. Each request is processed in a temporary folder that is removed before the answer comes back.
Standards and roles
Each identifier is matched against the standards table by its text, ignoring case; where two rows match, the lower one wins. Positions that match no row are samples.
Blank
Empty tin capsules show what the capsule and the system add to every peak. On Auto the blank is subtracted only when at least three capsules gave a peak and their δ values agree within 2‰ (1 SD). {{text.blankFlag}}
Linearity
{{text.linSeries}} shows whether δ depends on peak size. On Auto a slope is applied only when it differs from zero at p < {{text.alpha}}. It pivots on the series’ mean peak area, so a peak of that size does not move; peaks outside the series are extrapolated and flagged.
Drift
{{text.monitors}} placed through the run show whether the instrument moved. On Auto a slope is applied only when it differs from zero at p < {{text.alpha}}. Select a monitor on a chart to leave it out of the fit, and again to put it back.
Calibration
{{text.anchors}} put the run on the AIR and VPDB scales. Accepted values are regressed on measured ones and the line is applied directly, never inverted. Samples outside the anchors are extrapolated, and their uncertainty grows with the distance.
Elemental content
A k-factor fitted to every weighed standard turns peak area into micrograms of N and C. When sensitivity moves through a run, the drift monitors read high or low in %N and %C; the correction divides each position by the monitors’ factor there, interpolated between monitors and held flat beyond the first and last. On Auto it applies when a monitor is off by more than {{text.elemGate}}%. Select a monitor to leave it out.
Uncertainty
Each value carries a standard uncertainty u from the anchors’ certified values, the standard errors of their measured means, and single-analysis repeatability, propagated through the calibration line as a Kragten spreadsheet would. U = {{text.k}}u. It does not include the blank, linearity, or drift corrections, or how one run compares with another.
Quality checks
Checks that change no value but say how far to trust them. Caution when reversing a linearity or drift choice would move a sample by more than U, or when a monitor used for elemental content is off by more than 10%; Note for anything else worth a look.
Results
Corrected values for every sample, with U (k = {{text.k}}), %N, %C, and C:N. Select a row or a point to see how its value was reached.
The sequence Bar height is the N2 peak area on a log scale; select a bar to inspect it.
{{c.title}}
{{c.sub}}{{c.title}}
{{c.sub}}Take it with you
Methods paragraph, written from the choices above
{{res.methods}}
The numbers
{{flow.effect.title}}
{{l.t}}
Where U comes from {{flow.budget.title}}, share of u²
{{flow.insp.title}}
{{flow.insp.sub}}- {{f.k}}
- {{f.v}}
- {{f.k}}
- {{f.v}}