Reviewed by Nani Frenkel, Chief Editor
Anyone who has decarboxylated cannabis at home knows the smell. You put ground flower in a low oven, and within twenty minutes the entire house knows about it. People usually treat that as an inconvenience, something to manage with an open window.
It is worth understanding what that smell actually is. Much of what you are smelling is the volatile fraction leaving — terpenes, along with other aroma compounds. The aroma filling your kitchen is aroma that is no longer in your material, and it is not coming back.
That is one half of the trade at the center of decarboxylation. The other half is that you do not get back as much as you put in, and most guides skip both.
What decarboxylation is
Fresh cannabis carries most of its potential THC as THCA, tetrahydrocannabinolic acid, with much smaller amounts already present as neutral THC. THCA has an extra carboxyl group attached and does not produce intoxication. Heat removes that group, releasing carbon dioxide and converting THCA to THC. The same reaction turns CBDA into CBD, and CBGA into CBG.
This is why eating raw flower generally produces little intoxication, and why cannabis is smoked, vaporized, or heated before infusion. Smoking and vaping cause rapid decarboxylation during use, through sheer intensity of heat. Making edibles or oils generally means doing it deliberately, either as a separate step beforehand or during extraction and infusion.
Time and temperature work together
There is no single correct setting, and under commonly studied conditions decarboxylation behaves as a first-order or pseudo-first-order reaction: its rate depends strongly on temperature, while the amount converted depends on how long the material is held there. Higher heat for less time, or lower heat for longer, can reach the same endpoint.
A widely cited kinetic study heated cannabis extract in a vacuum oven at five temperatures between 80°C and 145°C for up to an hour, tracking acidic and neutral forms by supercritical fluid chromatography. It observed first-order behavior through most of that range, though the reaction at 145°C ran too fast for its order to be determined reliably. More useful was a difference between compounds: the rate constants for THCA were roughly twice those of CBDA and CBGA. The acids do not convert at the same speed. A protocol that fully converts THCA in a THC-dominant extract may therefore leave more CBDA unconverted in a CBD-dominant one.
This is also why published temperatures disagree so much. That study used dried extract in individual vials under vacuum. Ground flower in an ordinary domestic oven is a different matrix under different conditions, so its numbers do not transfer directly.
- The oven’s setting is not the material’s temperature, particularly in the first several minutes.
- Flower, extract and oil behave differently.
- Open and closed systems behave differently — an open system loses volatiles continuously.
- Different cannabinoid acids convert at different rates.
“240°F for 40 minutes” is a serviceable rule of thumb, not a chemical constant.
Conversion has an endpoint
The intuition that more heat means more THC is wrong, and it is worth being explicit about where it breaks.
Too little heat or too little time leaves part of the cannabinoid pool in its acidic form. For a product intended to deliver neutral THC, CBD or CBG, that means incomplete conversion.
But once conversion is close to its practical maximum, additional heat produces diminishing returns while increasing the opportunity for evaporation, oxidation and other transformations. The optimum is not the hottest or the longest treatment. It is the point at which the conversion you wanted has happened, at a level of loss you find acceptable.
Which losses, and how large, depends on what you started with. The vacuum-oven study found THCA conversion comparatively straightforward under its oxygen-limited conditions, while CBDA and CBGA showed larger unexplained losses. Thermal behavior is not identical across cannabinoids: under those particular conditions, losses ran to roughly 8% for the THCA and THC total, 18 to 25% for CBDA and CBD, and 53% for CBGA and CBG.
Why THCA does not become an equal weight of THC
Here is the part most decarboxylation guides never mention, and it changes how you should read a lab report.
The reaction works by throwing away a piece of the molecule. THCA has a molecular weight of about 358 g/mol; the carbon dioxide it sheds accounts for around 44 of that, roughly 12.3%. What remains is THC at about 314 g/mol. One gram of THCA cannot become one gram of THC — it can at best become about 0.877 grams.
That figure is why the standard total-THC calculation looks the way it does:
Total THC = THC + (THCA × 0.877)
The same factor applies to CBDA converting to CBD, since the carboxyl group carries almost identical relative mass in both.
In practice: flower testing at 20% THCA will not give you 20% THC. Complete conversion of that THCA yields roughly 17.5%, before adding whatever neutral THC was already present and before subtracting anything lost to evaporation or degradation along the way. The 12.3% is not a processing inefficiency you can engineer away. It is the reaction working correctly.
What heat costs the volatile profile
The temperatures that convert cannabinoids also drive terpenes out of the material, and the terpenes go first.
A 2023 review in Cannabis and Cannabinoid Research worked through this using vapor pressure — the property governing how readily a compound leaves a liquid or solid and becomes a gas. Higher vapor pressure at a given temperature means earlier evaporation. Monoterpenes have considerably higher vapor pressure than cannabinoids, so as material heats, they leave first. The authors found composition shifting during drying and curing, well before deliberate heat, and becoming extensively altered through decarboxylation. Their conclusion about the results is blunt: commercial cannabis products are substantially depleted in terpenes relative to the plants they came from, which makes labels like “full spectrum” and “whole plant” difficult to defend.
The loss is not even, and that is the interesting part.
Monoterpenes are ten-carbon molecules and light. Myrcene has a normal boiling point around 167°C, while ocimene and limonene boil around 176°C. Those figures sit above ordinary home-decarboxylation temperatures — but boiling is not the threshold for evaporation. Volatile compounds escape steadily well below it, which is the whole reason you can smell them. Sesquiterpenes are fifteen-carbon molecules and considerably less volatile: β-caryophyllene boils near 263°C. That does not make it immune, but it leaves far more slowly.
So the profile does not thin evenly — it shifts. As the light compounds go, what remains is relatively enriched in the heavy ones. Bright citrus and herbal notes are the first casualties; the peppery, woody character of the heavier fraction becomes more prominent, not because it grew but because everything above it shrank. If a decarbed batch smells darker and flatter than the flower it came from, that is not imagination.
How processors limit the losses
Processors can mitigate this, though no single approach is universal. Broadly there are two levers: recover the volatile fraction separately under mild conditions and recombine it afterwards, or reduce thermal exposure in the first place.
A 2020 paper in the Journal of Pharmaceutical and Biomedical Analysis tested six protocols for medical cannabis oils, looking specifically for methods preserving cannabinoids and terpenes together. Two performed well. One distilled the essential oil off first, then extracted the residual material and recombined the fractions. The other used room-temperature maceration starting from partially decarboxylated material — conversion treated as a dial rather than a switch, taken as far as the product requires and no further. Live resin production addresses the earlier losses instead, flash-freezing fresh material to avoid much of what drying and curing would take.
What is removed can also be replaced. Reintroduced terpenes are either cannabis-derived, recovered from the plant itself, or sourced from other botanical material — the limonene in a cannabis vape may well have come from an orange. When the goal is to reproduce a specific cultivar, laboratory analysis provides the target, though in practice many blends are standardized or strain-inspired rather than reconstructions of any particular flower. Either way, you can rebuild only what you measured, and commercial panels capture only part of cannabis’s volatile complexity. A reconstructed profile is a sketch, not a copy.
A note on laboratory measurement
Decarboxylation also shapes how potency gets measured. When gas chromatography is performed without derivatization, the heated inlet converts acidic cannabinoids during the analysis itself, so an estimate of total THC or CBD depends on how completely that conversion occurs. A 2025 study in Pharmaceutics found conversion rates of only about 50 to 60% for THCA and CBDA in prepared test samples, showing that the method can underestimate total cannabinoid content when appreciable acidic cannabinoids remain.
Liquid chromatography avoids this particular problem by measuring acidic and neutral forms separately, and derivatized GC methods can do so as well. In non-derivatized GC testing, the reaction that defines this article can therefore influence the number printed on the certificate of analysis.
The short version
Decarboxylation is not a preparation step that happens before the interesting chemistry. It is the interesting chemistry: cannabinoid acids lose carbon dioxide and become their neutral counterparts, while the process can also reduce cannabinoid recovery and strip away much of the volatile aromatic fraction.
That aromatic cost is paid disproportionately by the monoterpenes, leaving the surviving profile weighted toward heavier, less volatile compounds. Knowing the trade does not let you avoid it. It does let you decide how much of it to make.
Research for this article was compiled using DeepWeed, T&T’s cannabis research database — explore the underlying studies there.
Image Source: Kindel Media, Pexels
Sources
- Decarboxylation study of acidic cannabinoids: a novel approach using ultra-high-performance supercritical fluid chromatography/photodiode array-mass spectrometry. Cannabis Cannabinoid Res. 2016;1(1):262–271. doi:10.1089/can.2016.0020
- Eyal AM, et al. Vapor pressure, vaping, and corrections to misconceptions related to medical cannabis’ active pharmaceutical ingredients’ physical properties and compositions. Cannabis Cannabinoid Res. 2023. doi:10.1089/can.2021.0173
- Innovative methods for the preparation of medical Cannabis oils with a high content of both cannabinoids and terpenes. J Pharm Biomed Anal. 2020;186:113296. doi:10.1016/j.jpba.2020.113296
- Incomplete decarboxylation of acidic cannabinoids in GC-MS leads to underestimation of the total cannabinoid content in cannabis oils without derivatization. Pharmaceutics. 2025;17(3):334. doi:10.3390/pharmaceutics17030334
