Mass spectrometry is the instrument that finds trace molecules at concentrations low enough to fail a cannabis batch. It is how a laboratory detects a pesticide at parts-per-billion concentrations, determines whether a mycotoxin is below its reporting limit, and investigates compounds no one put on the label. It is also routinely oversold, as though every number on a certificate of analysis came out of one. Most potency numbers do not. This page explains what a mass spectrometer measures, which kinds cannabis labs use and for what, and where the technique runs out.
What does a mass spectrometer actually measure?
One thing: the mass-to-charge ratio (m/z) of ions. Everything else is inference. Every mass spectrometer has three parts. An ion source gives molecules an electric charge. Common examples are electron ionization after gas chromatography, which hits molecules with electrons in a vacuum, and electrospray or atmospheric-pressure chemical ionization (APCI) after liquid chromatography. A mass analyzer sorts those ions by m/z. A detector counts them. The output is a spectrum: which m/z values arrived, and how many of each.
That spectrum can do two different jobs, and it helps to keep them apart. Identification asks what a compound is, from its mass, its fragments, and when it came off the chromatography column. Quantitation asks how much is there, by comparing the signal with calibration standards of known concentration. A mass spectrometer is excellent at the first and good at the second, provided the lab has the right reference standards and controls for the sample matrix, which in cannabis is a large proviso (see below).
In a testing lab the spectrometer almost never works alone. It sits at the end of a chromatograph that first separates the mixture over time, so that the spectrometer sees a few compounds at once rather than hundreds. GC-MS pairs it with gas chromatography; LC-MS pairs it with liquid chromatography. Our testing methods guide sets out how the techniques fit together.
Which mass spectrometers do cannabis labs use, and for what?
Single-quadrupole GC-MS. A workhorse for volatile and semivolatile compounds that can survive the temperatures used in gas chromatography. Electron ionization produces reproducible fragment patterns that can be matched against reference libraries, which is how labs screen terpenes, residual solvents and unexpected peaks. Library matching can suggest an identity; confirmation normally requires a reference standard and a retention-time match. A spectrum library is only as useful as its entries; the NIST library’s expansion to 121 cannabinoids mattered for exactly this reason.
Triple-quadrupole (tandem) MS, written MS/MS. The instrument behind contaminant testing. The first quadrupole selects an ion of one mass, a collision cell breaks it apart, and the third quadrupole selects one specific fragment. Monitoring that precursor-to-fragment pair cuts through background so effectively that a triple-quad can quantify dozens of target compounds, in some methods more than a hundred, in one run at parts-per-billion levels. It is a targeted technique: in a targeted run, the laboratory generally sees only the compounds and transitions programmed into the method.
High-resolution MS (Q-TOF and Orbitrap). These measure mass to within a few parts per million, enough to narrow the possibilities to one or more plausible elemental formulas from accurate mass and isotope pattern.[1] Fragmentation data and, ideally, a reference standard are still needed to establish the structure. That is the tool for unknowns: degradation products, synthesis by-products, adulterants. It is less common in routine compliance labs than in research and investigative work, because it costs more and the targeted triple-quad is better suited to a fixed list of regulated analytes.
When is mass spectrometry the practical standard, and when is it optional?
Pesticides: MS/MS is the practical standard. California currently sets action levels for 66 pesticides. The 21 compounds that formerly failed a sample whenever they were detected now carry numerical limits of 0.10 µg/g, or 100 parts per billion; the list expands to 80 pesticides on April 1, 2028.[2] The regulation specifies action levels and the limits of quantitation laboratories must achieve, not the instrument they must use. In practice, LC-MS/MS and GC-MS/MS are the principal ways to quantify so many compounds at these concentrations in cannabis. Even then, one ionization mode is not always enough. In one validated hemp method, 62 of California’s 66 pesticides were measured by LC-MS/MS with electrospray, but four very non-polar compounds, including chlordane and pentachloronitrobenzene, ionize poorly that way and were measured with APCI instead, showing that one LC-MS/MS platform could cover all 66; those four are conventionally run on GC-MS/MS.[3] The GC pesticide-screening article covers that side.
Mycotoxins: usually LC-MS/MS. California fails a sample if the combined concentration of aflatoxins B1, B2, G1 and G2 exceeds 20 µg/kg, or if ochratoxin A separately exceeds 20 µg/kg.[4] LC-MS/MS is commonly used because it can measure all five selectively at trace levels, although the regulation does not prescribe an instrument.
Residual solvents: GC, often with MS. California lists 6 solvents in its strictest category (benzene, for example, at 1 µg/g) and 14 in the second (butane, propane and ethanol at 5,000 µg/g).[5] The rule specifies the limits, not the instrument. Labs typically use headspace gas chromatography, with either a mass spectrometer or a flame ionization detector; the MS adds confirmation of identity, which matters when an unexpected peak appears.
Potency: usually not MS. Most THC and CBD numbers come from liquid chromatography with a UV detector. Cannabinoids are present at percent levels in flower, so UV has sensitivity to spare, and LC, unlike GC, measures THCA and THC separately because it does not heat the sample. Gas chromatography heats the sample, causing acidic cannabinoids to decarboxylate. Without derivatization it therefore cannot measure THCA and THC separately as they existed in the original sample, and incomplete conversion can underestimate total cannabinoid content.[1,6] MS earns its place in potency work for minor cannabinoids, for low-dose edibles and beverages, and for telling apart compounds that UV cannot: CBG and CBD, for instance, have UV spectra too similar for the detector to distinguish if they co-elute.[1] Why two labs can still disagree on THC is the subject of our potency article.
The broader picture is that “required” depends on where you are. An analysis of 36 states and the District of Columbia found 679 different regulated contaminants, 551 of them pesticides, with action levels for the same compound differing by up to four orders of magnitude between jurisdictions.[7] Our contaminant testing pillar follows that patchwork.
Why is cannabis such a hard sample?
Because the things labs are looking for are vanishingly small and the things they are not looking for are enormous. A pesticide at 100 parts per billion shares the extract with cannabinoids that can make up a fifth of the flower’s weight and terpenes at up to several percent,[3] a concentration gap of roughly six orders of magnitude. That causes two problems.
The first is ion suppression. In electrospray, everything leaving the column at the same moment competes for charge. When a large co-eluting peak takes most of it, the trace analyte produces less signal than it would in clean solvent, and the result reads low. A 2018 review called this matrix effect “the major drawback” of MS for cannabinoid analysis and said that without isotope-labeled internal standards, which closely track the analyte and experience similar extraction losses and ion suppression, only unreliable values are obtained.[1] The hemp pesticide method above used 20 labeled internal standards for its 66 pesticides, many of them standing in for compounds without an exact labeled match, plus matrix-matched calibration.[3]
The second is interference: a matrix compound that happens to share the analyte’s mass and fragment. In the same method, hemp produced so much interference at the usual transition for the fungicide propiconazole that the authors switched to a transition based on a heavier chlorine isotope, cutting the lowest quantifiable level from 0.5 to 0.02 µg/g.[3] That is good method development. It is also a reminder that a contaminant result is only as good as the validation behind it, and that a method validated on flower is not automatically valid for a gummy or a vape oil.
What can’t mass spectrometry tell you?
Isomers. In routine LC-MS/MS cannabinoid methods, delta-8-THC and delta-9-THC share the same molecular mass and many of the same product ions. Reliable identification and quantitation therefore depend on chromatographic separation.[1] If they co-elute, the method may confuse them, combine their signals or fail its confirmation criteria; the exact outcome depends on the method. The same challenge applies to delta-10-THC and the growing list of semisynthetic isomers; our page on THC 8, 9 and 10 explains the chemistry.
What it was not asked. A targeted triple-quad method reports on its list. A product can pass every regulated test and still contain something no one screened for. When researchers analyzed 27 delta-8-THC vape products with NMR, GC-MS and metals analysis, every one contained reaction by-products, including an isomer of THC and a previously undescribed cannabinoid; 11 contained unlabeled cutting agents; and none matched its labeled delta-8 content.[8] Untargeted screening, increasingly with high-resolution instruments, is how such compounds get found, and it is not part of the routine compliance panels described above.
Whether the sample represents the batch. No instrument corrects for a sample that was not representative, a reference standard that was degraded, or a lab that chose not to report a result. Mass spectrometry is a measurement, not a guarantee. That is the point Cindy Orser makes about measurement uncertainty in our interview with her, and it applies with extra force at parts per billion.
What should a reader take from a certificate of analysis?
Look at which method produced which number. Potency by HPLC-UV, pesticides and mycotoxins by LC-MS/MS and GC-MS/MS, and solvents by headspace GC is the normal pattern. A “pass” on pesticides means none of the listed compounds exceeded its action level in that state, using that lab’s validated method; it does not mean the product contains no pesticides, and it says nothing about compounds that are not on the list. A “not detected” result does not mean zero. It means the signal fell below the laboratory’s stated reporting threshold, which may be its limit of detection or its limit of quantitation; the certificate or its accompanying definitions should say which.
The 2017 version of this page said mass spectrometry was the best way to find pesticides in cannabis. That is still true, and the regulations have since made tandem mass spectrometry the practical standard for contaminant testing. What has changed is the list of things worth finding, and the evidence that a clean certificate depends as much on what a lab was asked to look for as on how good its instrument is. For the whole testing picture, start at our cannabis testing hub.
References
- Citti C, Braghiroli D, Vandelli MA, Cannazza G. Pharmaceutical and biomedical analysis of cannabinoids: a critical review. Journal of Pharmaceutical and Biomedical Analysis. 2018;147:565–579.
- California Code of Regulations, Title 4, §15719. Residual pesticides testing, as amended by Department of Cannabis Control rulemaking DCC-2025-03-R (filed July 27, 2026; effective October 1, 2026).
- Dalmia A, Cudjoe E, Jalali J, et al. A LC-MS/MS method with electrospray ionization and atmospheric pressure chemical ionization source for analysis of pesticides in hemp. Journal of Cannabis Research. 2021;3:50.
- California Code of Regulations, Title 4, §15721. Mycotoxin testing.
- California Code of Regulations, Title 4, §15718. Residual solvents and processing chemicals testing.
- Franzin M, Di Lenardo R, Ruoso R, et al. 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.
- Jameson LE, Conrow KD, Pinkhasova DV, et al. Comparison of state-level regulations for cannabis contaminants and implications for public health. Environmental Health Perspectives. 2022;130(9):097001.
- Meehan-Atrash J, Rahman I. Novel Δ8-tetrahydrocannabinol vaporizers contain unlabeled adulterants, unintended byproducts of chemical synthesis, and heavy metals. Chemical Research in Toxicology. 2022;35:73.
Originally published March 27, 2017. Rebuilt October 2026 with current instruments, the regulations that have made tandem mass spectrometry the practical standard for contaminant testing, and its limits.
