Terpene Reintroduction and Strain-Profile Reconstruction

Why refined extract has no character, and why that is the point human data

Distillation separates by volatility, and the cannabinoids are the least volatile thing of interest in the extract. A short-path or wiped-film run is therefore designed to drive off everything lighter than the cannabinoid fraction first: water, residual solvent, and the entire mono- and sesquiterpene complement. That is not a defect of the process, it is the process working. The consequence is that the material that comes off the main body is potent, pale, chemically simple and organoleptically dead. It has no cultivar identity because the molecules that carry cultivar identity have been removed several stages earlier. Terpene reintroduction is the formulation step that gives the product back an aroma, a flavour and a market position, and in a vape it also changes the rheology of the fill in ways that matter more than the flavour does.

• The volatile fraction is a small mass share of the flower and a very small mass share of the concentrate, but it carries essentially all of the aroma.
• A terpene recovery step during processing can capture some of the original fraction for later reintroduction; most operations instead buy stocks and blend.
• Aroma is the first thing a buyer evaluates and the only property they can assess before purchase, which is why this step gets disproportionate commercial attention.

Sources: Booth JK 2019* · Sommano SR 2020* · Brenneisen R 2007* · Hazekamp A 2007*

Inclusion rates by product type contested industry practice, not published data

Inclusion rate is bounded at both ends and the bounds are different for every route. In a vape cartridge the usual working band is about 2 to 5 percent terpene by mass of the finished fill. Below roughly 1 to 2 percent the draw is harsh and flavourless, because the terpene fraction is doing double duty as the flavour and as the thing that softens the aerosol. Above roughly 8 to 10 percent two separate problems arrive together: the fill becomes thin enough to leak past the wick and the seals of a cartridge designed for a thicker liquid, and the terpene load itself becomes a meaningful share of what is being inhaled and thermally decomposed. In edibles and tinctures the constraint is not rheology but the palate, and useful inclusion is a fraction of a percent; terpenes are aggressive flavours at a level far below where they change viscosity. In topicals the constraint is skin tolerance, since several common terpenes are documented dermal irritants and sensitisers at high load, so finished topicals typically sit under a couple of percent.

Contested — caveat. These bands are working commercial practice compiled from formulator and processor use, not values from a published study. They vary with hardware, cannabinoid profile and the specific terpene stock. Treat them as starting points to be confirmed empirically in your own hardware, not as specifications.

Sources: MELEK hemp-science shelf 2026* · Sommano SR 2020* · Meehan-Atrash J 2019*

Solubility, miscibility and why a blend can come apart industry practice, not published data

Terpenes are lipophilic hydrocarbons and hydrocarbon derivatives, and cannabinoid distillate is a lipophilic resin, so at the crude level the two are freely miscible and mixing is not a solubility problem. The differences that matter are within the terpene fraction itself. Individual terpenes span a wide range of polarity and a wider range of volatility: an oxygenated monoterpene such as linalool or 1,8-cineole is measurably more polar than a pure hydrocarbon monoterpene such as alpha-pinene, and the sesquiterpenes boil 60 to 100 °C higher than the monoterpenes. Three practical consequences follow. A blend that looks homogeneous immediately after mixing can stratify slowly if a component is near its solubility limit in the resin, particularly where a polar oxygenated component is loaded heavily. A blend exposed to warmth or open headspace loses its lightest components preferentially, so the ratio drifts toward the sesquiterpenes over time and the product smells different at the end of a production run than at the start. And a blend introduced into hot distillate flashes off exactly the monoterpenes that were the most expensive part of the stock.

• Blend into warm, not hot, material. Gentle warming to roughly 40 to 60 °C drops the viscosity of distillate enough to mix without driving off the light ends.
• Every 20 °C above that costs monoterpenes you have already paid for, and the loss is selective, so it changes the profile and not just the total.
• Mix in a closed or headspace-minimised vessel, and mix to full homogeneity before filling. A cartridge filled from a partially mixed batch is a concentration gradient with a label on it.
• Stratification is a QC question, not a theoretical one: pull from top and bottom of a held vessel and assay, rather than assuming a blend stays blended.
• Terpene stocks themselves oxidise. Alpha-pinene and limonene form oxidation products on storage in air, so stock age and headspace history are part of the formulation record.

Sources: Sommano SR 2020* · Booth JK 2019* · MELEK hemp-science shelf 2026*

Strain-profile reconstruction from a GC panel: the arithmetic industry practice, not published data

The trade practice of rebuilding a named cultivar profile from its laboratory terpene panel is straightforward arithmetic and it is worth spelling out because the honest limits live in the arithmetic rather than in the chemistry. A GC terpene panel reports each target compound as a mass percentage of the sample, typically flower. Step one is to sum the reported terpenes to get the total reported volatile fraction for that sample; that sum is usually somewhere between about 0.5 and 3 percent of flower mass. Step two is to divide each individual compound by that sum, which normalises the panel to relative percentages of the reported fraction and removes the effect of how terpene-rich that particular harvest was. Step three is to treat those normalised relative percentages as the recipe: to make 100 g of terpene blend you weigh out that many grams of each individual terpene stock. Step four is to decide the inclusion rate of the finished blend into the product, which is an independent decision from the ratio and is governed by the table above. A worked example: a panel reporting 0.62 percent myrcene, 0.31 percent beta-caryophyllene, 0.21 percent limonene, 0.14 percent alpha-pinene and 0.12 percent linalool sums to 1.40 percent; normalised, that is 44.3, 22.1, 15.0, 10.0 and 8.6 percent of the blend, and at a 4 percent inclusion rate a 1000 g batch of finished formulation takes 40 g of blend made from 17.7 g myrcene, 8.8 g beta-caryophyllene, 6.0 g limonene, 4.0 g alpha-pinene and 3.4 g linalool.

• Normalise, do not copy. The absolute percentages in a panel describe how much terpene that harvest had; only the ratios describe what it smelled like.
• Panels are usually reported on flower. A panel from concentrate has already been through a process that changed the ratio, so it describes the concentrate rather than the cultivar.
• Chirality is invisible in most panels. Alpha-pinene has two enantiomers with recognisably different odour, and a panel reporting a single alpha-pinene figure does not tell you which one, or in what ratio, the plant made.
• Weigh by mass, not by volume. Terpene densities differ enough that a volumetric shortcut moves the ratio.

Sources: Giese MW 2015* · Hazekamp A 2012* · Sommano SR 2020* · MELEK hemp-science shelf 2026*

What a reconstruction cannot recover human data

The reconstruction is faithful to the panel, and the panel is not faithful to the plant. A GC terpene method quantifies a selected target list, because quantification requires a reference standard for every analyte and a laboratory stocks standards for the compounds its customers ask about. A cultivar volatile profile, as characterised in the research literature, contains well over a hundred compounds, and the trace constituents include classes a cannabis terpene panel does not look for at all — most notably volatile sulfur compounds, which are present at parts-per-billion levels and are the documented source of some of the most distinctive and commercially prized aromas in modern cultivars. A blend rebuilt from a 20-compound panel therefore reproduces the ratio of the twenty compounds that were measured and omits everything that was not on the list, which is disproportionately the material that made the cultivar recognisable. The result usually smells like a competent generic version of the cultivar and not like the cultivar. This is a limitation of the method, honestly stated; it is not a reason not to do it, and it is the reason a reconstructed blend and a recovered native fraction are not the same product.

• You cannot rebuild what was never measured. The gap between a target list and a full volatile profile is the whole gap between a reconstruction and the original.
• Non-terpene volatiles — thiols and other sulfur compounds, esters, aldehydes — are frequently outside the panel scope and frequently the signature.
• Matrix matters. The same blend on distillate, on a live-resin base and on flower does not present the same way, because the base contributes and because volatilisation in hardware is not the same event as combustion.
• Honest labelling: a reconstructed profile is a cultivar-inspired blend. Describing it as the cultivar is a claim the arithmetic does not support.

Sources: Giese MW 2015* · Booth JK 2019* · Sommano SR 2020* · Gilbert AN 2018*

Botanical versus cannabis-derived terpenes industry practice, not published data

When a botanical stock and a cannabis-derived fraction contain the same molecule, they contain the same molecule: beta-caryophyllene from black pepper or clove is the same compound as beta-caryophyllene from cannabis, and no analysis distinguishes them at the single-compound level except by isotope ratio work that nobody in this trade does. That settles the chemistry and leaves four real differences. Cost and availability: single botanical terpenes are commodity flavour and fragrance materials produced at industrial scale and priced accordingly, while a cannabis-derived terpene fraction is a low-yield co-product of a cannabinoid process and costs one to two orders of magnitude more per unit mass. Purity and specification: a food- or fragrance-grade botanical isolate comes with a specification, a lot number and often a documented purity that a cannabis-derived fraction does not. Composition: this is the substantive one — a cannabis-derived fraction is not a set of terpenes, it is everything volatile that came off that material, which includes minor cannabinoids carried over in the fraction, oxidation products, and the trace non-terpene volatiles a botanical blend cannot contain by construction. And regulation: botanical terpene stocks sit inside established flavour and fragrance regulatory frameworks with defined food-use status, whereas a cannabis-derived fraction inherits the regulatory status of the cannabis it came from, which is jurisdiction-dependent and can make an otherwise identical product a controlled item.

Sources: Sommano SR 2020* · Booth JK 2019* · MELEK hemp-science shelf 2026* · Rowe RC 2020*

Flavour versus effect, stated honestly contested in vitro

Two claims travel together in the trade and they have completely different evidential standing. The flavour and aroma claim is immediate, large, reproducible and uncontroversial: adding a terpene blend to a characterless distillate changes what the product smells and tastes like, a user can detect the change reliably, and the effect is present at inclusion levels of a couple of percent. Nobody disputes this and it does not need a citation to a receptor study. The effect claim — that the ratio of terpenes in a product steers the character of the cannabinoid experience, so that one blend is sedating and another is energising — is popular, commercially load-bearing, and weakly supported in humans at these inclusion levels. The pharmacological arguments on the sceptical side are concrete: the terpene mass delivered in a few puffs of a 4 percent blend is small, the plasma concentrations that would result are far below the concentrations at which most terpenes show activity in vitro, and a direct functional study of common cannabis terpenoids at human CB1 and CB2 found no modulation of Δ9-THC signalling. The arguments on the other side are also real: some terpenes have documented activity at non-cannabinoid targets, beta-caryophyllene is a genuine CB2 agonist in its own right rather than a modulator, and a rodent study reported cannabimimetic and cannabinoid-enhancing behaviour for several cannabis terpenes. What is missing is the human dose-response work at realistic product inclusion levels. Until that exists, the correct position is that the flavour effect is established and the effect-steering claim is a hypothesis that the market has already priced as if it were settled.

• Established: terpene reintroduction changes aroma and flavour, and changes the harshness of a vape draw.
• Established: beta-caryophyllene binds and activates CB2 directly, which is a pharmacological fact about that one compound, not evidence for ratio-steering generally.
• Not established: that a 2 to 5 percent blend at typical use volumes reaches concentrations that modulate the subjective cannabinoid experience in humans.
• Confound: expectation. A product that smells strongly of a cultivar the user associates with a given effect is not a blinded comparison, and almost no trade evidence is blinded.

Contested — caveat. The effect-modulation claim is genuinely disputed. Russo 2011 set out the entourage hypothesis and remains the most-cited case for it; Santiago 2019 found no modulation of Δ9-THC activity at human CB1 or CB2 by common cannabis terpenoids in a functional assay; LaVigne 2021 reported cannabimimetic and enhancing effects in mice. In-vitro and rodent findings at applied concentrations do not establish a human effect at product inclusion levels. The aroma and flavour claim is not contested; only the effect claim is.

Sources: Russo EB 2011* · Santiago M 2019* · LaVigne JE 2021* · Gilbert AN 2018* · Sommano SR 2020*

Thermal fate: what happens to the blend in the hardware in vitro

A terpene blend is not inert in use. At vaporiser coil temperatures and at dab temperatures terpenes do not merely evaporate; they partially degrade, and the degradation chemistry is documented. Monoterpenes in particular are established precursors of aromatic degradation products under the thermal conditions of dabbing, and the same chemistry operates in a cartridge driven hard or run dry. This has two formulation consequences that are usually ignored. First, the aerosol a user inhales is not the blend you formulated, so a blend chosen purely for how it smells cold is a partial description of the product. Second, terpene load and device power interact: raising inclusion rate and raising coil temperature push in the same direction on degradation-product formation. This is the same reasoning that makes the diluent question on the viscosity page a safety question and not a rheology question — anything added to an inhaled formulation is being chosen as a pyrolysis feedstock whether the formulator thinks of it that way or not.

• Terpene degradation products documented from cannabis concentrate thermal studies include aromatic hydrocarbons and reactive carbonyls; formation rises with temperature.
• Running a cartridge dry or at high power is the worst case, because the residual film sees the highest temperature.
• A blend should be evaluated in the hardware it will ship in, hot, and not only in the jar, cold.

Sources: Meehan-Atrash J 2017* · Meehan-Atrash J 2019* · Sleiman M 2016*