Alexis St-Gelais, chimiste – Popularization
One of PhytoChemia’s most renowned assays in Canada is our detailed cannabis terpenes analysis. This assay can have different applications and comes in two declinations, “main” and “full” terpenes. Let’s have a tour on the ins and outs of our analytical service, and how it can be of use to your business.
|
Quick highlights
|
In this post, we will discuss the outline and principles of our assay, define the concepts of full/main terpenes and anhydrous/as is results, and warn of the potential pitfalls of the notion of “total terpenes”.
Terpenes (and volatiles) in cannabis and hemp
Besides the very well-known cannabinoids, Cannabis sativa is also an aromatic plant species. The trichomes of the inflorescences, in particular, produce not only the cannabinoids but also smaller, volatile molecules, most of which are terpenes. The term “terpene” designates all compounds sharing a common origin in the plant metabolism, deriving from isoprene units assemblies. In addition to those terpenes, a few other non-terpenic but nevertheless volatile molecules can be found, like hexyl hexanoate which is a fatty acid ester.
Terpenes are far from being limited in occurence to cannabis and hemp, and not all terpenes are volatile compounds, either. For example, lycopene, the red pigment in tomato, is a terpene, and it is not volatile. It is therefore important to keep in mind that the term “terpene” is used in the field of cannabis as a linguistic shortcut for its volatile constituents, but that the overlap is not perfect between those categories. Below, we will use the commonplace meaning of terpenes* with regards to cannabis, i.e., the volatile compounds (boiling point lower than that of phytol) encountered in the plant, mostly comprising terpenes.
Cannabis terpenes are relatively abundant. For dried buds, typical concentrations can be as low as 2-3 mg/g up to slightly over 60 mg/g, depending on strains. It therefore is not surprising that this wealth of compounds attracts interest in terms of quality control, strain breeding and pheno-hunting.
How we test for cannabis terpenes
Extracting the terpenes
Generally speaking, terpenes have good affinity for non-polar organic solvents, like dichloromethane, hexanes and pentane. PhytoChemia uses pentane as an extraction solvent, because it can be purchased in high purity for gas chromatography (see below). The extraction of terpenes can therefore be relatively simply effected by grinding the cannabis as finely as possible. A weighed amount of the plant material is then mixed with pentane. At this point, a known amount of methyl octanoate is added to the extraction vessel – more on that below. The tube is then sealed and subjected to vigorous agitation for several minutes. Finally, the solvent can then be drawn and filtered to obtain a liquid terpenes extract that is ready for analysis.
Comparison of analytical approaches
Being volatile molecules, cannabis terpenes are highly suited for gas chromatographic (GC) analyses. GC refers to the separation technique used to discriminate the different terpenes from each other and get an individual reading for each. A general explanation on GC can be found here.
The most common practice to monitor terpenes is to quantitate a predefined array of compounds individually against calibration curves. While valid on a compound-by-compound basis, this approach has the drawback of giving information only for target molecules and disregarding any other. A large number of terpenes found in cannabis are not available as pure standards for purchase, and therefore are not amenable to this technique. In other terms, with a classical targeted quantitation, you can miss a fair part of the bigger picture.
This challenge is not unique to cannabis. It also is highly relevant to the essential oils and fragrances industries, where targeting compounds one by one in complex mixtures is either impossible owing to lack of standards or prohibitive in terms of cost and labour. To circumvene this problem, a group of experts from the International Organization of the Flavor Industry (IOFI) have worked for years on finding a way to robustly quantitate virtually any volatile compound against a single reference molecule. The result of their work was published in 2016 [1], and our laboratory quickly became a convinced user of this approach. All of our reports refer to the paper in question. Let’s see how it works in theory.
Gas chromatography coupled to flame ionization detection (GC-FID)
The IOFI approach is based on gas chromatography coupled to flame ionization detection (GC-FID). With both mass spectrometry (MS) and FID, the structure of a molecule can affect how efficiently it generates a signal at the detector. However, FID is significantly more constitent accross different compounds, and also from one laboratory to another. You can read more on that topic in one of our previous blogposts. The FID detector is based on the combustion of carbon in a flame, where it is oxidized and generates charged particles that allow a current to move accross the flame. The more carbon oxidized, the more particles, and the higher the conductivity of the flame becomes. Hence, the electrical current reading is correlated to the mass of carbon-containing molecules being burnt at a given time during the analysis.
Many volatile molecules are not only made of carbon but also comprise other atoms. Those atoms can affect the oxidization process of the carbon of the molecule within the flame, if only because e.g., oxygen is already bound to some of the carbon atoms. Therefore, the signal intensity obtained within the FID detector for two different molecules burnt in the same amount (mass) can be slightly different. This is why, formally, each compound should be quantitated according to its own calibration curve. This is where the IOFI has proposed a clever alternative.
The IOFI predicted reponse factors
Cachet et al injected almost 500 individual volatile molecules on GC-FID instruments at various concentrations. By studying the intensity of their FID response according to their concentration, they noticed that it was affected by the molecular formula of the compound (number of carbon, hydrogen, oxygen, nitrogen, sulfur, fluorine, chlorine, bromine, iodine and silicium atoms), the number of benzene rings, and its molecular weight. Therefore, when the structure is known, it theoretically becomes possible to predict the response behaviour of a molecule, even when this molecule is not available to draw a calibration curve.
To take advantage of this feature of the FID response of molecules, the IOFI working group uses an internal standard of methyl octanoate. This substance is not naturally present in cannabis, and therefore suitable for this purpose. Methyl octanoate can be introduced in a known amount into a sample, and it is then monitored on the GC-FID chromatogram during analysis. This produces a methyl octanoate peak for which intensity (i.e., peak area) can be directly correlated to mass of standard in the sample injected.
All other peaks in the sample can then be quantitated by comparison with this reference peak. As mentionned above, this would be faulty since not all molecules will produce the same signal as methyl octanoate in the FID detector. A predicted response factor is therefore applied to the result to correct for this effect, based on a formula proposed by the IOFI. This formula allows to correct the result of a given peak, quantitated in equivalents of methyl octanoate, into a concentration value that takes into account the molecular formula, molecular weight and number of benzene rings of the molecule considered. Using this approach, any volatile compound can be quantitated against the methyl octanoate reference with a reasonably small calculation error. PhytoChemia has used the IOFI approach to participate to several interlaboratory assays of terpenes in cannabis, and the results have always been compliant with the expected values, showing the efficiency of the approach. The methodology has been validated and our validation report can be provided to customers upon request.
Identifying the peaks
One must at that point know what the structure of the molecule producing each peak is, to properly compute the predicted response factor for each molecule. Since we are not bound to a set of terpenes that are available for purchase, we can broaden the perspective and use non-targeted chromatographic profile interpretation. We therefore examine each peak of the chromatogram looking for a match. To this end, we use a combination of two techniques: dual columns retention indices, and complementary mass spectrometry whenever needed (GC-MS). More can be read on this process in this example blogpost.
Throughout this process, and inevitably in the field of phytochemistry (the chemistry of plants), some molecules will remain unidentified. This does not make the signal irrelevant, but it simply means that the compound’s structure and name are not known. Major unknown constituents can be reported within analyses too, both for future reference and to track them as one would track any other known constituent in the course of batch-to-batch quality control. Since the IOFI calls for known molecular formulas to establish predicted response factors, unknown compounds are expressed directly in terms of equivalents of methyl octanoate, without correction factor.
Bottom line
By combining pentane extraction, an internal standard, the IOFI approach and careful GC-FID analyses coupled to extensive retention indices and GC-MS databases, PhytoChemia can provide a non-targeted screening of terpenes in a cannabis or hemp sample. This means that we can list and quantitate all terpenes (and other volatiles) present in a detectable amount in an efficient manner, instead of being limited to those substances available as calibration standards – and avoiding looking for substances that are not expected in cannabis despite being included in calibration mixes. With our assay, you get the whole picture!
Practical considerations of terpenes analysis (also frequently asked questions!)
“Full terpenes” vs “Main terpenes”
PhytoChemia offers its terpenes screening in two variants: full and main terpenes. It must imperatively be kept in mind that regardless of the option, the analytical approach we use is identical: same extraction, same internal standard, same instrument.
The difference between those two services is not based on scientific considerations, but rather on practical ones. Indeed, a notable limiting factor for terpenes analysis is data interpretation, which takes time – and not all applications require the same level of detail in analysis. Full and main terpenes are therefore two sides of the same coin, taylored by PhytoChemia to suit different needs.
Full terpenes simply mean that we will try to identify as many peaks as reasonably feasible on the chromatogram. This will produce a list that can comprise up to 150-180 compounds, although it can also be less – this is dependant on the sample itself and what it yields upon injection. Many of these compounds will be found in very small amounts only, and some will be unknowns that are however recurring in the species. We generally also include a short sentence with salient features of the profile as a conclusion for full terpenes analysis reports. Since data interpretation takes more time for our team, and therefore high-end, this service is offered at a higher price.
Main terpenes are a collection of a bit over 80 volatile compounds (mostly terpenes) whose list was assembled by PhytoChemia. Those terpenes were chosen because they either:
- ∙ Are regularly quantitatively important (like myrcene, caryophyllene, or selina-3,7(11)-diene);
- ∙ Are characteristic of cannabis (like pinene hydrates and ipsdienol, as well as two recurring unknowns);
- ∙ Feature a presence/absence behavior that can differentiate strains (like hexyl butyrate, bulnesol or delta-guaiene and a third unknown strongly associated to terpinolene);
- ∙ Drew particular interest to customers for various reasons (like hashishene and phytol)
Since the detailed minor constituents are not as extensively characterized, this service takes less data analysis time and can therefore be offered for a lower price. This is an entry level service that is therefore offered at a lower price tag.
It also is worth mentionning that all main terpenes** are included in the full terpenes screen. Finally, keep in mind that regardless of the option you picked, if your sample is poor in terpenes and yields few peaks, the number of identified compounds can be lower in both cases.
Quick guide to choosing full or main terpenes:
- ∙ If you work on a budget, pick main terpenes. Since they require less data interpretation, they are offered at a lower price.
- ∙ If you need a high level of detail, head for full terpenes. These are suited for research purposes, or pheno-hunting.
- ∙ If you produce an extract, like hash or wax, choose full terpenes. Your specific extraction conditions could trigger peculiar results that are more likely to be caught by this more extensive screen.
The total terpenes concept: use with proper precautions
Total terpenes content is a metric that should be used with high care, because of its very nature: it is a sum of what laboratories are looking at! If one tests for 10 terpenes, the “total terpenes” is merely the sum of these 10 targets, and discounts any other terpenes that might be present. It therefore is quite easy to end up comparing things that are completely different but listed under the same concept.
Considering that PhytoChemia is looking at a fairly higher number of terpenes than most laboratories, it also makes sense that the sums we report are also higher – if only for example because we consider selinadiene-type sesquiterpenes, often abundant but mostly overlooked by “traditionnal” target-based quantitation methodologies. Even if you compare between reports of our own laboratory prepared for other people, be mindful: those taking the “main terpenes” option have a report listing typically around 80 terpenes, which together will typically account between 90 and 95% of what a “full terpenes” screen would have yielded. This is a small difference, but it should be kept in mind when performing comparisons.
Therefore, comparisons between our reports should be reliable, provided you do not compare “main terpenes” with “full terpenes” carelessly. If you want to do that, we suggest to first sort the “full terpenes” to only consider the compounds found in the “main terpenes” reports and exclude the rest. Cross-comparisons with results from other laboratories will possibly generate noise and discrepancies, since the lists of covered terpenes will not match.
Another parameter to take into account is residual moisture – see anhydrous vs as is terpenes concentrations in the next section.
Hence, to be relevant, a “total terpenes” metric should ideally be accompanied by a definition of what is included within.
As for typical total terpene values (using our screens), the bulk of samples we receive would be in the 18-60 mg/g range. Values below 10 mg/g are more typical of industrial hemp, i.e. that has not been selected over time for secondary metabolites, while they are not so common for cannabis (at least based on our experience). We have seen above 60 mg/g in some rarer cases. Extracts can score much higher, depending on their nature, because they can be enriched in terms of terpenes. An essential oil of cannabis or hemp that would be devoid of significant amounts of cannabinoids would close on a terpenes content of 1000 mg/g, i.e. 100%.
“Anhydrous” and “as is” terpenes concentrations
When dealing with plant material, the question of water content or residual moisture is never very far. Very often, monographies from pharmacopoeias setting regulatory content of signature molecules in medicinal plant material will do so on dry (anhydrous) basis. This means that the concentration of e.g., curcumin in turmeric, should be expressed on the mass of turmeric tested after correcting it for the amount of residual water. For instance, 1 g of turmeric powder containing 10% residual moisture is in fact 0.9 g of anhydrous turmeric. This helps ensuring lab-to-lab consistency, because the residual moisture content can vary depending on local conditions. For the same original turmeric sample, if lab A stores it in a cold chamber with significant humidity, and lab B in a dry container with silica dessicant, after a few days, the residual moisture will be different. Correcting for humidity removes this source of error.
On the other hand, for cannabinoids, Health Canada tends to promote the expression of results on the material as received (“as is”). Concentrations are therefore expressed in terms of the material weighed exactly as it was received, without correcting for humidity. That being said, Health Canada does not require nor regulate terpenes analysis, which leaves this question unresolved. Faced with both the usual pharmacopoeias approach and Health Canada’s inclination, and the fact that different customers had different preferences in that regard, PhytoChemia chose to report both values – anhydrous and as is. It therefore is the end user’s responsibility to pick the proper metric according to their own specifications. Also, for comparisons purposes, one should never compare anhydrous vs as is results, because they refer to two different metrics.
Anhydrous concentrations are always higher. They offer a way to compare different strains, although their residual moisture may differ. On the other hand, as is concentrations are lower (because the water content “dilutes” the terpenes with reference to the overall plant material mass) and allow the end user to figure out how much terpenes are contained within a given amount of their cannabis product, since the latter is not used in its anhydrous form.
For extracts or transformed products, there is no such distinction. Plant material never reaches near 0% residual moisture unless it is oven-dried or freeze-dried (neither of which is convenient for cannabis producers), but extracts can be fully dried. Therefore, moisture content is not a relevant question in those cases and the terpenes concentrations are reported as a single value.
Notes
*It is commonplace to talk about terpenes for cannabis, by virtue of convention. That being said, if you used steam distillation with cannabis in a similar fashion as you would for e.g., peppermint, you would also be able to condensate a fraction of organic and fragrant liquid that would contain those volatile molecules (mostly terpenic in nature). Such a liquid would be called an essential oil, and that term is what is generally used for most other aromatic plants beyond cannabis.
**The list of terpenes included in our main terpenes service is as follows: hexanol, hashishene, α-thujene, α-pinene, camphene, α-fenchene, β-pinene, sabinene, myrcene, α-phellandrene, Δ3-carene, α-terpinene, para-cymene, limonene, β-phellandrene, 1,8-cineole, (Z)-β-ocimene, (E)-β-ocimene, γ-terpinene, cis-sabinene hydrate, octanol, fenchone, terpinolene, para-cymenene, trans-sabinene hydrate, linalool, endo-fenchol, trans-pinene hydrate, cis-pinene hydrate, camphene hydrate, epoxyterpinolene, ipsdienol, borneol, terpinen-4-ol, para-cymen-8-ol, α-terpineol, hexyl butyrate, citronellol, (4Z)-decenol, geraniol, decanol, α-cubebene, α-ylangene, hexyl hexanoate, β-caryophyllene, α-santalene, γ-elemene, α-guaiene, trans-α-bergamotene, α-humulene, allo-aromadendrene, (E)-β-farnesene, β-selinene, valencene, α-selinene, δ-guaiene, β-bisabolene, (3E,6E)-α-farnesene, spirovetiva-1(10),7(11)-diene, eremophila-1(10),7(11)-diene, selina-4(15),7(11)-diene, selina-4,7(11)-diene, selina-3,7(11)-diene, (E)-α-bisabolene, germacrene B, eudesma-5,7(11)-diene, (E)-nerolidol, caryophyllene oxide, guaiol, humulene epoxide II, 10-epi-γ-eudesmol, selin-6-en-4α-ol and an isomer thereof, γ-eudesmol, β-eudesmol, α-eudesmol, bulnesol, (3Z)-caryophylla-3,8(13)-dien-5β-ol, α-bisabolol, juniper camphor, aromadendra-4,10-diol, (2E,6E)-farnesol, cryptomeridiol, meta-camphorene and phytol, as well as up to three characteristic unknown constituents of cannabis (two sesquiterpenes and an oxygenated monoterpene).
