Calibration Curve
Standards → linear regression with R², plus Beer-Lambert ε computation.
How to use this tool
Turn a set of standards into a straight-line calibration, then read an unknown's concentration off that line. Built for absorbance (Beer-Lambert) curves, but works for any linear instrument signal.
What to enter
- Conc. (x): the known concentration of each standard, in whatever units you work in.
- Signal (y): the instrument reading for that standard (absorbance, peak area, mV…). Add or remove rows with the buttons.
- Unknown signal (y): the reading from your sample; the tool back-calculates its concentration.
Reading the result
The headline is the unknown's concentration. Below it sits the line's equation y = m·x + b and R², the linearity score: aim for R² ≥ 0.995. Keep your unknown inside the range of the standards, extrapolating past them is unreliable.
Worked example
The five default standards fit y = 0.106·x + 0.003 with R² ≈ 0.99997; an unknown reading 0.5 back-calculates to a concentration of about 4.68.
Regression
The unknown is read off the fitted line y = m·x + b. R² scores linearity, aim for ≥ 0.995; a lower value usually means one standard is off. Keep the unknown's signal inside the range of your standards, since the line isn't reliable beyond them. For absorbance data the slope m equals ε·ℓ.
Methodology
An ordinary least-squares fit of signal (y) against concentration (x) yields y = m·x + b. The coefficient of determination R² reports linearity. An unknown's concentration is back-calculated as x = (y − b) / m. For absorbance data this is the Beer–Lambert working curve; the slope equals ε·ℓ (molar absorptivity × path length).
Acceptance
- R² ≥ 0.995 is the common bar for quantitative calibration.
- Keep unknowns inside the calibrated range, extrapolation is unreliable.