HPLC vs GC for Cannabinoid Testing

What the two methods actually do

Both High-Performance Liquid Chromatography (HPLC) and Gas Chromatography (GC) separate a mixture into its components so each can be measured. The difference is the mobile phase and the temperature.

In HPLC, a liquid solvent pushes the dissolved sample through a packed column at room temperature or modest heat (typically 25–40 °C). Cannabinoids are detected by their UV absorbance, usually with a diode-array detector (DAD) ^[1][2].

In GC, the sample is vaporized in a hot inlet (often 250–300 °C) and carried through the column by an inert gas like helium. Detection is typically by flame ionization (FID) or mass spectrometry (MS) ^[3].

That inlet temperature is the whole story. Cannabinoids in living plants exist primarily as carboxylic acids — THCA, CBDA, CBGA — not as the neutral forms (THC, CBD, CBG) you feel when you smoke. Heat drives off the carboxyl group as CO₂ in a reaction called decarboxylation ^[4].

Why GC inflates 'THC' unless you derivatize

Running raw cannabis on a standard GC method converts most THCA in the sample to THC inside the injector before it ever reaches the column. The chromatogram then shows a single THC peak that represents the sum of (original THC) + (THC formed from THCA) Strong evidence ^[4][5].

This isn't necessarily wrong — it's just answering a different question. A GC potency result approximates 'total THC available after smoking or vaping,' because combustion also decarboxylates the acids. But it cannot tell you how much of the flower was THCA versus THC to begin with, which matters for edibles, tinctures, raw juicing, and regulatory labeling that distinguishes the two.

Decarboxylation in the GC inlet is also incomplete and variable — studies have shown losses of roughly 30% or more of the theoretical maximum, depending on inlet temperature, residence time, and matrix ^[5]. That introduces method-dependent error.

Workarounds exist. Derivatizing the sample (e.g., with a silylating agent like BSTFA) protects the carboxylic acid group and allows GC to resolve THCA and THC separately ^[6]. But derivatization adds labor, cost, and another opportunity for error, which is why most potency labs simply switched to HPLC.

Why HPLC became the potency standard

Because HPLC never heats the sample past the boiling point of the solvent, THCA stays as THCA and CBDA stays as CBDA. A typical reverse-phase HPLC-DAD method resolves at least 10–17 cannabinoids in a single run, including THCA, THC, CBDA, CBD, CBN, CBG, CBGA, CBC, THCV, and CBDV ^[1][2][7].

This matters for several reasons:

• Regulatory compliance. In jurisdictions where 'hemp' is defined as ≤0.3% Δ9-THC (US Farm Bill) or where total THC includes a conversion factor (THC + 0.877 × THCA), labs need the acid and neutral forms reported separately ^[8].
• Product labeling. A raw flower might be 22% THCA and 0.5% THC. Smoked, that converts to roughly 19–20% THC. A label needs both numbers honest.
• Stability monitoring. Tracking the ratio of CBN to THC over time tells you about oxidative aging Strong evidence ^[4].

The AOAC and ASTM have both adopted HPLC-based reference methods for cannabis potency ^[7].

Where GC still wins

GC isn't obsolete in a cannabis lab — it's the right tool for volatile analytes:

• Residual solvents in extracts (butane, ethanol, hexane, etc.) are analyzed by headspace GC-MS or GC-FID. HPLC cannot do this well ^[9].
• Terpene profiling is dominated by GC-FID and GC-MS because terpenes are volatile, thermally stable, and don't absorb much UV light. See Terpene Analysis Methods.
• Pesticide screening often combines LC-MS/MS and GC-MS/MS to cover different chemistries.

So a modern cannabis testing lab typically runs HPLC for cannabinoid potency, GC for terpenes and solvents, and LC-MS/MS or GC-MS/MS for pesticides and mycotoxins.

What to look for on a Certificate of Analysis

When you read a COA, check:

1. Method. It should specify HPLC (sometimes written HPLC-UV, HPLC-DAD, or UHPLC) for cannabinoids. If it says GC without mentioning derivatization, treat the acidic cannabinoid columns skeptically.
2. Separate THCA and THC values. A legitimate HPLC report lists both.
3. 'Total THC' formula. Usually reported as THC + (THCA × 0.877). The 0.877 factor accounts for the mass of CO₂ lost during decarboxylation ^[8].
4. Limit of quantitation (LOQ). Trace cannabinoids like CBC or THCV may be present but below the method's detection floor.

If a COA shows '0.00% THCA' and a large THC number on raw flower, the lab is almost certainly using a non-derivatized GC method and reporting only post-decarb totals Strong evidence.

Bottom line

HPLC and GC are complementary, not competing. For potency labeling and any question about the acid/neutral balance in the actual product you're buying, HPLC is the appropriate method. For terpenes and residual solvents, GC is appropriate. Anyone selling you a 'GC is more accurate' or 'HPLC is more accurate' narrative without specifying the analyte is oversimplifying.

For more on what the numbers on a label mean in practice, see Decarboxylation and Reading a Cannabis COA.

Sources

How this page was made