Calibration Curve Calculator

A calibration curve is the line your standards drew for your instrument. Once that line exists, an unknown sample is no longer a raw reading: it can be turned into an amount, and the number you report is simply where the sample's response crosses the line.

The line takes four numbers: the sensitivity (its slope), the background (its intercept), the signal your instrument produced, and the concentration behind that signal. Type any three and the page solves the fourth, so the same tool answers three different questions: what concentration does this reading correspond to, what response should a standard of this strength give, and what slope and intercept did my standards actually produce.

No unit is imposed anywhere, deliberately. An instrument may report absorbance, peak area, counts, current or millivolts, and a concentration may be molar, mass-based or parts per million. The arithmetic asks only that the standards, the sample and the answer you report are on one common scale.

That makes the page useful to anyone who measures a response against known amounts: cuvette and microplate assays, chromatographic and atomic-spectrometric runs, electrochemical probes, colorimetric kits, and teaching laboratories where the point is to build the line rather than to use one.

One habit worth keeping: compare the answer with the range of concentrations your standards covered. Inside that range the line is supported by measurements; outside it the arithmetic still runs, but it is predicting rather than interpolating.

Everything on this page is one statement rearranged. The line is

The four rearrangements below are all the calculator does. Each one corresponds to one empty box:

The first is the one most people come for. The subtraction is the part worth understanding: the background is not a correction bolted on at the end, it is the response the matrix, the solvent and the reagent blank produce on their own. Dividing only the response that sits above that level by the sensitivity is what turns an instrument reading into an amount of analyte.

What the page will not accept

Four boundaries follow from what the numbers mean, and the page enforces all of them. A sensitivity of zero or below is refused, because an instrument that gains nothing per unit of concentration has no calibration to invert. Background, concentration and signal cannot be negative: a background signal is a response, a concentration is an amount, and a measured response is never below nothing.

Those rules also catch impossible results rather than hiding them. If the entries push a solved value onto the wrong side of zero, the value stays on screen and is flagged, so you can see what the arithmetic produced and decide which entry was wrong:

A negative concentration from a sample whose signal sits below the background means the reading cannot belong to that line. Either the blank or matrix correction is wrong, or the sample is simply below what this calibration can resolve. The other common flag is an empty box with a message instead of a number: that is the page telling you the division it would need has a zero in the denominator, most often a concentration of zero while you are asking for a sensitivity.

Building the line from two standards

If you have run standards but not fitted them, two points are enough to fill the line

Take standards at $x_1 = 1$ reading $y_1 = 0.35$ and at $x_2 = 5$ reading $y_2 = 1.75$:

The intercept comes out at zero, which is a sensible answer for a method with no measurable matrix contribution. Feeding the line back in, a sample that reads $y = 0.98$ returns $x = 2.8$.

Two situations that are not the usual “what is in my sample” question, because that one is already covered above.

When a result will be reported or acted on, have the responsible analyst review it against the current calibration and the run's quality-control data. The page can tell you what the arithmetic says about the numbers you entered, and that is all it can tell you.