The optimal calibration curve is calculated by matching the curve to the calibration points. The curve calculation is based on a least squares fit (LSQ), which minimizes the sum of the residual squares. The curve type is applied to weighted calibration points. The calculation depends on the definition of the response factor (RF definition, see Response type and response factor).
With RF defined as Response per amount:
Σ(wt * (CalPointArea - CalculatedArea)2) = min
where | |
Σ | Sum over the calibration points (levels) |
CalculatedArea | The area read from the curve at calibration level amount |
CalPointArea | Calibration level area |
wt |
Calibration point weighting factor |
Σ(wt * (CalPointHeight - CalculatedHeight)2) = min
where | |
Σ | Sum over the calibration points (levels) |
CalculatedHeight | The height read from the curve at calibration level amount |
CalPointHeight | Calibration level height |
wt |
Calibration point weighting factor |
With RF defined as Amount per response:
Σ(wt * (CalPointAmount - CalculatedAmount)2) = min
where | |
Σ | Sum over the calibration points (levels) |
CalculatedAmount | The amount read from the curve at calibration level area or height |
CalPointAmount | Calibration level amount (corrected amount, calculated from compound amount in sample, sample-specific compound multipliers, and sample-specific compound dilution factors) |
wt |
Calibration point weighting factor |