How We Measure Markers: HPLC, HPTLC, UV & GC Explained

A standardised percentage on a Certificate of Analysis looks like a fact. It behaves like one in purchase orders, in formulation calculations, in label claims. But every such figure is the output of an analytical procedure, and procedures differ enormously in what they can resolve, what they respond to, and how defensible their answer is. Two laboratories analysing the same drum can return different numbers — not because one is dishonest, but because they asked the material a different question. For R&D formulators, QC managers and procurement teams, understanding the method behind the marker is the difference between comparing equivalent specifications and comparing labels that merely share a unit.

This article walks through the analytical toolkit used to characterise and quantify botanical extracts: high-performance liquid chromatography (HPLC), high-performance thin-layer chromatography (HPTLC), UV-Vis spectrophotometry, gas chromatography (GC and GC-MS), and the mass-spectrometric and classical methods that fill the gaps. It closes with the validation vocabulary you should expect a supplier to speak fluently, and a short list of questions that turn a glossy spec sheet into a trustworthy one.

Why the method behind the number matters

Selectivity is the heart of the matter. A method's selectivity is its ability to measure the analyte you care about without interference from everything else in the extract. A chromatographic method physically separates constituents before measuring them, so it can report on one named compound. A spectrophotometric method measures a bulk optical property of the whole solution, so it reports on a class of compounds — or on anything that happens to absorb at the chosen wavelength. The same plant material can therefore yield a high number by a non-selective method and a lower, more honest number by a selective one.

This is why a like-for-like price comparison can be quietly misleading. If one supplier reports total flavonoids by a UV colorimetric assay and another reports a specific glycoside by HPLC, the figures are not on the same scale and the cheaper material may carry less of the active you actually want. The method tells you which scale you are reading.

HPLC: the defensible workhorse

High-performance liquid chromatography is the default for quantifying specific markers in botanical extracts, and for good reason. The sample is dissolved and pumped under high pressure through a column packed with a stationary phase; constituents partition between that stationary phase and the flowing mobile phase, and so emerge separated in time. Each separated peak can be quantified against a reference standard run under identical conditions. Because the constituents are physically resolved before detection, HPLC can report on one molecule among many — the property that makes it the number you can defend to an auditor or a regulator.

Reversed-phase and the C18 column

Most botanical assays run in reversed-phase mode (RP-C18): a non-polar stationary phase (octadecyl-silica, C18) with a more polar mobile phase, typically water or buffer mixed with methanol or acetonitrile. In this mode, more polar constituents elute earlier and more lipophilic ones later — a good fit for the polyphenols, glycosides, alkaloids and lactones that dominate standardised extracts. The column chemistry, particle size and dimensions are part of the method; change them and retention times and resolution change with them.

Detection: UV and the diode-array detector

The most common detector is UV, and increasingly the diode-array detector (DAD or PDA), which records a full UV-Vis spectrum across the whole run rather than absorbance at a single wavelength. DAD adds a powerful identity check: a peak that elutes at the right time and also shows the expected spectral shape (and spectral purity, with no co-eluting interferent) is far more convincingly the target compound. Note the important distinction this introduces — HPLC-UV uses UV as a detector after separation, which is not the same as standalone UV-Vis spectrophotometry of an unseparated extract, covered below.

Gradient versus isocratic elution

An isocratic method holds the mobile-phase composition constant throughout the run — simple, robust and well suited to a single marker or a small group eluting in a narrow window. A gradient method changes the mobile-phase composition over time (for example, increasing the organic fraction), which is what you need to resolve a complex mixture spanning a wide polarity range within a reasonable run time. Gradient methods give superior resolution for fingerprinting and multi-marker assays but demand more controlled instrumentation and re-equilibration between injections.

Reference standards and system suitability

An HPLC result is only as trustworthy as the reference standard it is calibrated against and the system used to generate it. Two checks matter most. First, the reference standard should be authenticated and of known purity — ideally pharmacopoeial (USP, EP) or a characterised, certified standard — because the assay quantifies the unknown by comparison to it. Second, a system suitability test is run before the assay to prove the instrument is fit for purpose on the day: it verifies parameters such as resolution between critical peak pairs, theoretical plate count (efficiency), peak tailing/asymmetry, and the repeatability of replicate injections. If system suitability fails, the run is invalid before any sample number is reported.

HPTLC: fingerprinting and identity

High-performance thin-layer chromatography separates constituents on a coated plate rather than in a column. A precisely metered band of sample is applied near the bottom of a silica plate, the plate is developed in a closed chamber as solvent rises by capillary action, and the separated zones are then visualised — under white light, under UV (often 254 nm and 366 nm), or after derivatisation with a reagent that produces characteristic colours. The result is a visual pattern of bands: a chromatographic fingerprint of the material.

That fingerprint is HPTLC's strength. Comparing the band pattern of an incoming lot against an authenticated reference and against pharmacopoeial monographs is a powerful, cost-effective identity and authenticity test — it can flag adulteration, substitution of a related species, or a profile that simply does not match what the botanical should show. Many pharmacopoeias specify HPTLC identity methods for exactly this reason. It is also well suited to screening many samples in parallel on a single plate.

Densitometry and semi-quantitation

HPTLC can move beyond identity toward quantitation using densitometry: a scanning densitometer measures the optical density (absorbance or fluorescence) of each band, and band intensity is related to the amount present by calibrating against reference standards applied to the same plate. This makes HPTLC genuinely semi-quantitative. In careful, validated hands it can approach HPLC for some assays, but in general it is less precise than HPLC for an exact assay value — its enduring role is confirming identity and profile rather than producing the single defensible potency figure.

UV-Vis spectrophotometry: totals, not molecules

UV-Vis spectrophotometry measures how much light a solution absorbs at chosen wavelengths, with absorbance proportional to concentration (the Beer-Lambert relationship). Crucially, it does this without separating the extract first — the instrument sees the sum of everything in the cuvette that absorbs at that wavelength. That makes UV-Vis fast, inexpensive and ideal for total-class assays, where the goal is to estimate a whole family of related compounds rather than one molecule.

Common examples are total phenolics by the Folin-Ciocalteu assay (reported as gallic acid equivalents), total flavonoids by the aluminium-chloride colorimetric method (as quercetin or rutin equivalents), and total tannins or total saponins by class-specific colour reactions. The defining feature, and the defining limitation, is in the word total: the result is a class estimate expressed against a chosen reference compound, not a measurement of any specific marker.

None of this makes UV-Vis a lesser technique — it is the right tool when a class total is genuinely what you want, when no single marker defines the material, or when no suitable chromophore-specific method exists. The error is using a class total as if it were a specific-marker assay.

GC, mass spectrometry and the classical methods

GC and GC-MS for volatiles and essential oils

Gas chromatography separates compounds in the vapour phase and is therefore the method of choice for volatile and semi-volatile constituents — the components of essential oils such as terpenes and terpenoids, and other compounds that can be vaporised without decomposing. Coupled to a mass-spectrometric detector (GC-MS), it both separates and identifies, matching each peak's mass spectrum against libraries to confirm component identity. GC-MS is the standard for characterising an essential oil's composition and for detecting volatile adulterants. GC also underpins two safety-panel tests: residual solvents (typically by headspace GC against ICH Q3C class limits) and, with tandem MS (GC-MS/MS), part of multi-residue pesticide screening.

LC-MS/MS for trace levels and confirmation

Liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) combines chromatographic separation with the specificity and sensitivity of mass detection. It excels where targets are present at trace levels or where unambiguous identity is required: confirming a marker's identity beyond what UV detection can prove, quantifying potent compounds at very low concentrations, screening for non-volatile pesticide residues, and detecting contaminants or adulterants. It is more specialised and more costly than HPLC-UV, and is deployed where that extra specificity and sensitivity earn their keep.

Titrimetric and gravimetric methods

Classical wet-chemistry methods still have their place. Titrimetric (volumetric) assays quantify a constituent by reacting it to a measurable endpoint and are used for some well-defined actives and for parameters such as acid or alkaloid content where a robust stoichiometric reaction exists. Gravimetric methods determine the mass of a fraction after extraction, precipitation or drying — for instance a crude “total saponins” fraction, or loss on drying for moisture. Both are simple and inexpensive but inherently low in selectivity: co-extracted or co-precipitated matrix is counted in the result, so a gravimetric total is easily overstated relative to a chromatographic assay of the true active.

Methods at a glance

Reference standards and method validation

A quantitative result rests on two foundations: the reference standard it is compared against, and the validation that proves the method does what it claims. On standards, the question to ask is what the marker is calibrated against — a pharmacopoeial or certified reference standard of known identity and purity, or an uncharacterised in-house material. The quality of the standard sets a ceiling on the quality of every result derived from it.

On validation, the internationally recognised framework is ICH Q2, which defines the performance characteristics a method must demonstrate. You do not need to run these tests yourself, but you should recognise the vocabulary and expect a supplier to be able to discuss them:

• Specificity — the method measures the intended analyte without interference from other constituents, impurities or the matrix. This is what separates a marker assay from a class total.
• Linearity — the response is proportional to concentration across the working range, so the calibration is sound (assessed by the correlation of a calibration curve).
• Accuracy — results agree with the true value, typically demonstrated by recovery of known spiked amounts.
• Precision — results are repeatable. Repeatability covers one analyst/instrument over a short period; intermediate precision covers different days, analysts or instruments within a lab.
• Range — the interval between the upper and lower concentrations for which the method has acceptable accuracy, precision and linearity.
• LOD (limit of detection) — the lowest amount reliably detected (distinguished from background), though not necessarily quantified.
• LOQ (limit of quantitation) — the lowest amount that can be measured with acceptable accuracy and precision; the practical floor for reporting a number.
• Robustness — the method withstands small, deliberate variations (mobile-phase composition, temperature, column lot) without losing reliability.

What to ask your supplier

You can interrogate any potency claim with a short, disciplined set of questions. They cost nothing to ask and they quickly separate a controlled, evidence-backed specification from a marketing figure.

1. 1What method produced this number — HPLC, HPTLC, UV-Vis, GC-MS, LC-MS/MS or a classical assay? A percentage without a method is not a specification.
2. 2Is the marker a specific named compound or a total class, and what exactly is summed in any total? Get the compound list behind a total.
3. 3What reference standard is the assay calibrated against — pharmacopoeial (USP/EP), certified, or characterised in-house — and of what stated purity?
4. 4Is the assay method validated to ICH Q2 (specificity, linearity, accuracy, precision, LOD/LOQ), and can you share the validation summary?
5. 5Does the COA state the method against each result, alongside the acceptance limit — or does it just print a number?
6. 6Can you provide a representative chromatogram (HPLC) or fingerprint (HPTLC) for the lot, and a system suitability record?
7. 7For solvent-extracted material, what is the residual-solvent result by headspace GC against the relevant ICH Q3C class limit?
8. 8If we second-source, will the new lab use a comparable method and the same marker definition — or will the number shift even at the same nominal percentage?

A supplier that answers these readily — naming methods, standards and validation status, and producing chromatograms on request — is telling you the percentage rests on evidence. Friction at this stage is itself information.