What this does, and what it refuses to do
Every number on this page comes from data you supply: the same material run at three or more heating rates. Nothing is trained, predicted from structure, or looked up. That is deliberate: isoconversional analysis is one of the few places in thermal analysis where the honest answer is pure inference from the measurement in front of you.
The single most common error in decomposition kinetics is to run Kissinger on one peak temperature per heating rate, get one activation energy with an excellent r², and carry it into a service-life prediction. That is only valid if the decomposition is a single elementary step. Most polymer and energetic decompositions are not. This tool computes Ea at each conversion by five methods and tells you when a single number is meaningless, which is the part a straight line through four points cannot.
1. Your runs
Paste one run per box: two columns, temperature and signal, separated by a tab, comma, or spaces. Temperature in °C or K (detected automatically). The signal can be mass %, mass in mg, or conversion 0–1, and it is normalised to conversion α between its own first and last value. Three heating rates is the practical minimum; four or five is better, and they should span at least a factor of five.
2. Activation energy against conversion
3. Kissinger, for comparison
Kissinger uses only the peak of each run, so it yields one activation energy for the whole process and cannot show Ea varying. It also assumes the conversion at the peak is the same at every heating rate, which holds for a first-order process and not in general. It is a reasonable first look and a poor last word.
4. Does the implementation actually work?
A kinetics tool that cannot be checked is worth nothing, so the check ships with it and runs in your browser right now. Synthetic first-order curves are generated from a known activation energy, then analysed as though they were measured. Every method must return the number it was given. The acceptance criterion was fixed before the result: recovery within 2% for the unbiased methods.
The residual OFW offset is not a bug. Ozawa–Flynn–Wall linearises the temperature integral through Doyle's approximation, whose slope carries a factor of 1.0518; the method is systematically high by about 1–2% even on perfect data. Seeing that offset here is what tells you the correction is applied in the right direction. Inverted, it would read about 168 kJ/mol against a true 150.
Reading the result
If Ea(α) is flat to within roughly 10% across the reliable conversion range, a single activation energy is defensible and every method should agree to within a few percent.
If Ea(α) drifts or steps, the process has more than one rate-limiting step and no single Ea exists. A rising Ea commonly means an easy process finishing and a harder one taking over. Plasticiser or solvent loss giving way to backbone scission is the everyday polymer case. What Kissinger returns for such a process is a conversion-weighted average that no theory supports, however good its r² looks.
The ends are unreliable. Below about α = 0.05 and above 0.95 the conversion is dominated by where you put the baseline rather than by chemistry, which is why the default range excludes them.
On lifetime extrapolation. This tool deliberately does not offer a one-click shelf-life number. Taking an Ea measured near 300 °C down to 25 °C storage is an extrapolation of many orders of magnitude in rate, and it assumes the mechanism is unchanged across that entire span, an assumption the measurement cannot test. For an energetic formulation a wrong answer there is a safety consequence, not a plotting error.
Method background: the isoconversional principle, the differential (Friedman) and integral (KAS, OFW, Starink) forms, and Vyazovkin's nonlinear formulation are standard; the ICTAC Kinetics Committee recommendations (Vyazovkin et al., Thermochimica Acta 520, 1–19, 2011, and the 2014 companion on complex processes) are the usual reference for how many heating rates to use, which conversion range to trust, and when Ea variation means a multi-step process. Consult them before publishing a number from any tool, this one included.
Related tools
Generate the thermograms this analysis consumes, and see how a formulation's steps superpose, on the Thermal Analysis page. Every term here (conversion, activation energy, heating rate) is defined in the Glossary.