Last updated on August 11, 2026 · Originally published March 15, 2019
THC and CBD numbers tell you how much of the major cannabinoids a product contains. They tell you very little about why two batches with near-identical potency can smell nothing alike. That takes terpene analysis.
But a terpene profile is not sitting inside the flower waiting to be read off. What comes back on the certificate depends on which compounds the laboratory decided to look for, how the sample was prepared, how well those compounds separated, and what detector was watching when they arrived.
Why terpenes are difficult to measure
Analyzing terpenes is no straightforward matter, even by the lofty standards of cannabis science.
Many of the major targets are chemically similar, and several are structural or geometric isomers of one another. Some can leave the chromatographic column at nearly the same moment and produce similar fragments in a mass spectrometer, which makes both clean separation and confident identification harder than the instrument list suggests.
Ocimene is a useful example. A method-validation study published in 2025 did successfully quantify its cis and trans isomers — but the final method still contained several coeluting pairs involving ocimene, and needed additional standards and selected ions to tell them apart. A method has to be designed around what the column can actually separate, not around what you would like it to.
There is a second problem underneath that one. The flower is a crowded place. Beyond the analytes of interest sits everything else the plant makes, and that background — the matrix — behaves like a smokescreen, making the compounds you care about harder to pick out.
First, the laboratory needs a representative sample
Before any of that, the batch has to become a test portion.
Which buds get selected, how much material is combined, and how thoroughly it is homogenized all affect what comes back. The instrument may see only a fraction of a gram, and that fraction has to stand for everything it was taken from. No amount of analytical precision downstream can rescue a sample that was not representative to begin with.
Homogenization improves repeatability, but it also becomes part of the method. Terpenes are volatile, and every additional handling step is another opportunity for the sample to change before it reaches the instrument.
Getting the sample into the instrument
This step gets skipped in most discussions and it shapes the result as much as the detector does.
With liquid injection, the laboratory first extracts the target compounds into a solvent. A small aliquot goes into a heated inlet, vaporizes there, and enters the column. This can capture a broad range of targets, including the less volatile terpenoids, but it carries more of the extracted matrix into the instrument along with them.
Static headspace sampling leaves most of that behind. The sample is sealed in a vial and left to equilibrate under controlled conditions; volatile compounds partition into the gas above it, and a portion of that gas is what gets injected. Headspace solid-phase microextraction, or HS-SPME, goes a step further, using a coated fibre to concentrate the volatiles from that gas before injection.
Temperature, equilibration time and preparation all determine which compounds make it into the instrument and in what proportion. This is not a theoretical concern: one study comparing four introduction techniques on the same material found that direct-immersion SPME Arrow gave responses averaging six times those of headspace SPME Arrow, and that headspace SPME precision varied wildly — relative standard deviations reaching 76%, against 15% or better for direct immersion. Our coverage of accelerated solvent extraction and injection techniques goes through that work in more detail.
Headspace is not merely a cleaner route into the GC. It is part of the measurement.
How gas chromatography separates the mixture
Terpenes are volatile — readily turned to vapor — which is what makes gas chromatography the standard approach.
The vaporized sample is picked up by a carrier gas, usually helium, hydrogen or nitrogen, and pushed through a long, narrow column. Compounds travel at different rates depending on how strongly they interact with the column’s coating, so what enters as a mixture leaves as a sequence. More volatile compounds generally emerge earlier and larger, higher-boiling ones later, though the exact order also depends on the stationary phase.
The detector sits at the far end. Everything that follows is a question of what you put there.
GC-FID
Flame-ionization detection is the most established and economical route to routine terpene quantification. As compounds leave the column they are burned in a hydrogen–air flame; the ions produced generate an electrical signal, and within the method’s calibrated range the area under a given compound’s peak tracks how much of it reached the detector. FID does not respond identically to every compound per unit mass, which is exactly why each target needs its own calibration.
FID is sensitive, predictable and linear across a wide concentration range. Its limitation is not detection — it is that it tells you almost nothing about what it detected. A laboratory identifies a peak mainly by matching its retention time against an authentic reference standard.
That puts the weight on the separation. If two terpenes coelute, FID may record them as a single combined peak, and the crowded cannabis matrix can introduce interferences of its own. Sample preparation, column choice and method validation matter as much as the detector.
GC-MS
Mass spectrometry adds a second source of information. After separation, arriving molecules are ionized and broken into characteristic fragments, sorted by mass-to-charge ratio. The resulting spectrum can be compared against reference libraries.
That gives MS considerably more identifying power than FID, particularly in full-scan analysis, where an unexpected compound may produce a searchable spectrum. But a library match is evidence, not proof. Closely related terpenes can produce similar spectra, and confident quantitative work still leans on reference standards, retention data and a validated method.
GC-MS/MS
Tandem mass spectrometry adds another layer of selectivity. In a typical triple-quadrupole instrument, the first mass analyser selects a particular precursor ion from the compound’s spectrum. A collision cell fragments it. The final analyser watches for one or more characteristic product ions.
Retention time plus a selected precursor-to-product transition is a much harder thing for an interfering compound to imitate. It does not make chromatographic separation irrelevant, but it makes difficult coelutions more manageable.
The payoff can be substantial. One validated static-headspace GC-MS/MS method simultaneously quantified 93 terpenoids in dried cannabis inflorescences and extracts, and was applied across sixteen medical cannabis chemovars.
How a peak becomes a number
Identifying a peak is only half the job, and the half that gets discussed.
To quantify it, the laboratory runs reference standards containing known concentrations of each target terpene and builds a calibration curve relating concentration to detector response. An internal standard — a known compound added at the same concentration to both standards and samples — helps correct for variation in preparation and injection.
What that looks like in practice: a validated GC-FID method for ten major cannabis terpenes used n-tridecane as its internal standard, calibrated from 1 to 100 µg/mL, and established limits of detection and quantification of 0.3 and 1.0 µg/mL. Recovery ran between 89 and 111% across placebo material and high-THC chemovars, with relative standard deviations under 10%. Those numbers are the method — not the instrument model.
A validated method also has to establish its own precision, accuracy, recovery, linear range, limit of detection and limit of quantification. Before anything is reported as a percentage or in milligrams per gram, results may need correcting for dilution, sample mass and moisture content — and, where the validated method specifies it, recovery.
This is why two laboratories running instruments with the same initials can report different numbers for the same material. Their target lists, standards, sample preparation, columns, calibration models and reporting bases may all differ.
Can terpenes be measured by LC-MS?
Yes, though it remains an unconventional route for volatile compounds.
A method published in 2024 used LC-APCI-MS/MS to quantify sixteen terpenes alongside seven cannabinoids — acidic forms included — in a single cannabis analysis. Extraction was a simple ethanolic procedure, separation took about twenty-five minutes on a reversed-phase C18 column, and the method was validated against international guidelines before being applied to fifty-five authentic samples.
Fifteen of those samples were also run by a validated GC-FID method. The terpene profiles were broadly comparable overall, with one notable exception: limonene concentrations differed by more than twofold between the two techniques, possibly through coelution or differences in detector response. Bias stayed below 15% for the validated analytes, and terpene recovery fell between 80 and 120%.
The appeal is obvious: one run, both compound classes, no separate derivatization step. That is something conventional GC methods cannot easily match. But the limonene result is also why LC-MS has not displaced GC for routine terpene work. It is better understood as a different analytical strategy than as a straightforward replacement.
What the result on a COA actually means
Which is worth remembering when reading a certificate of analysis. The number beside a terpene’s name is not a fact about the plant so much as a fact about the plant and the method used to look at it.
A terpene result is a measurement of a particular sample, prepared a particular way, against a particular target panel. A certificate listing twelve terpenes and one listing ninety-three are not necessarily contradicting each other. They may simply have asked different analytical questions.
For the broader context on these compounds, see our guide to cannabis terpenes.
Originally published March 15, 2019. Updated August 2, 2026. This article was substantially expanded from its original 2019 version. New sections cover representative sampling and homogenization, how samples are introduced to the instrument, tandem mass spectrometry, LC-MS, and how a chromatographic peak is converted into a reported concentration. The description of tandem MS has been corrected, and material has been incorporated from “Measuring Terpenes with Analytical Instruments” by Lance Griffin, published December 16, 2020. Six primary sources have been added.
Reviewed by Nani Frenkel, Chief Editor
Research for this article was compiled using DeepWeed, T&T’s cannabis research database — explore the underlying studies there.
Sources
- Shapira A, Berman P, Futoran K, Guberman O, Meiri D. Tandem mass spectrometric quantification of 93 terpenoids in cannabis using static headspace injections. Anal Chem. 2019;91(17):11425–11432. doi:10.1021/acs.analchem.9b02844
- Atkins PL. Sample processing and preparation considerations for solid cannabis products. J AOAC Int. 2019;102(2):427–433. doi:10.5740/jaoacint.18-0203
- Myers C, Herrington JS, Hamrah P, Anderson K. Accelerated solvent extraction of terpenes in cannabis coupled with various injection techniques for GC-MS analysis. Front Chem. 2021;9:619770. doi:10.3389/fchem.2021.619770
- Ibrahim EA, et al. Quantitative determination of cannabis terpenes using gas chromatography-flame ionization detector. Cannabis Cannabinoid Res. 2023;8(5):899–910. doi:10.1089/can.2022.0188
- Raeber J, et al. Simultaneous quantification of terpenes and cannabinoids by reversed-phase LC-APCI-MS/MS in Cannabis sativa L. samples combined with a subsequent chemometric analysis. Anal Bioanal Chem. 2024;416:4193–4206. doi:10.1007/s00216-024-05349-y
- A validated GC-MS method for major terpene quantification in hydrodistilled Cannabis sativa essential oil. Phytochem Anal. 2025. doi:10.1002/pca.3526
