If cannabis were regulated like other medicines, the product you obtain at a dispensary would contain a label with an accurate statement of the amount of active pharmaceutical ingredient, a list of the inactive ingredients, and an expiration date. I present data below that shows up to 77% of dispensary products are mislabeled. This is alarming since it means cannabis patients are receiving the wrong dose of medicine, and cannabis consumers are being ripped off. To solve this problem, cannabis dispensaries need to do their own testing. I discuss appropriate technologies for this.

Think about your last trip to a pharmacy. What did you see? Everything imaginable on sale from aspirin to groceries to office supplies perhaps. And what did you see behind the counter as you were waiting to get your prescription filled? Probably rows of shelves with bottles of pills in them, and pharmacists and their assistants clad in nice white laboratory coats, looking for all the world like analytical chemists. But what was missing behind that counter? The fume hoods, chromatographs, and spectrometers you would normally see in an analytical chemistry laboratory. Why is this? Because thanks to the US Food and Drug Administration (FDA) every bottle behind that pharmacy counter contains a known amount of active pharmaceutical ingredients, a list of inactive ingredients (excipients), and an expiration date.

For example, a bottle of pain reliever I bought recently from my local pharmacy clearly states each tablet contains 200 mg ibuprofen, and there is a list of inactive ingredients in descending order of concentration. The label also says, “Store at 20 °C to 25 °C . . . avoid excessive heat above 40 °C.” The label also has a batch number on it, instructions for use, warnings, indications, contra-indications, and a sell by date.

How do we know what is in each tablet? How do we know the best storage conditions? How was the sell by date determined? Because the FDA compels pharmaceutical companies to perform in-house testing to make sure every bottle produced contains what is advertised. Additionally, each label will have a clear sell by date on it because chemistry tells us medicines degrade over time, and again thanks to the FDA stability studies were performed so we know how long it is safe take a specific medicine.

Contrast this with your last trip to your local cannabis dispensary, assuming they are legal where you live and you actually visit them. The lack of analytical equipment is similar to a normal pharmacy, but that is where the similarity ends. Pick up a cannabidiol (CBD) tincture bottle. It may list a few ingredients such as CBD extract, medium chain triglycerides (MCT) oil, and “flavoring” but that is probably it. A list of excipients is typically not there, nor should you expect to see a sell by date.

Cannabis medicines contain more than CBD and tetrahydrocannabinol (THC), they contain other cannabinoids, terpenes, and who knows what else. An extractor may formulate their tinctures with a distillate that is 90% cannabinoids, but what is in that other 10%? Fats? Waxes? Chlorophyll? Pesticides and heavy metals? Is any of what is in that 10% harmful to human health? We have no idea since there is little scientific work on the analysis of the noncannabinoid components of cannabis extracts and distillates. This means we are formulating cannabis medicines with poorly characterized ingredients, putting our patient’s health at risk.

The Dose Makes the Poison . . . and the Cure

The old adage “the dose makes the poison” (1) translated into chemical terms means that concentration matters. A little of something may be benign or even beneficial, whereas in large doses it can be fatal. For example, acetaminophen is an effective pain reliever and fever reducer, but in high doses is a liver poison (2). At minimum then, the label for any medicine needs to have an accurate statement of the amount of active pharmaceutical ingredient (API) present in each dose. And yet in the cannabis industry mislabeled medicines are an ongoing problem (3–5). In a recent paper, 62% of commercially available samples had incorrect CBD amounts on their labels (5). This same paper found that CBD degrades measurably over the course of 30 days, and that light and heat accelerate the degradation (5). This means the common sell by date of one year on cannabis products is probably wrong.

To supplement the data on CBD products, I performed a study on the labeling of marijuana buds. This involved purchasing 1/8 oz of buds of different strains at local dispensaries, noting the total THC value on the label, and then having the potency tested by high performance liquid chromatography (HPLC) at a California state licensed, International Organization for Standardization (ISO) certified laboratory (6). The results for the 14 different marijuana strains are shown in Table I.

The relative difference for each strain was calculated by subtracting the label value from the laboratory value and dividing by the laboratory value. The state of California considers a cannabis product mislabeled if the laboratory value is more than 10% relative different than the label claim. Other workers have used this same standard (5). Using this criterion, 10 out of 13 or 77% of the samples examined here were mislabeled. Note that for all strains the third party laboratory value is lower than the label claim. On average, the laboratory value is lower than the label value by 4.55 wt.% total THC, and on average the relative percent difference was 25.2%.

The fact that all strains studied are significantly lower than advertised is distressing. This means cannabis patients are receiving the wrong dose of medicine, and that cannabis consumers are not getting what they are paying for. All the samples had been harvested and packaged at least two months before this study. A possible explanation for these results is that the marijuana lost potency over time under the storage conditions used. Cannabis potency decrease over time has been observed before (7-10). Another possible explanation for these results is inter-laboratory error, where different cannabis laboratories obtain different numbers on the same samples. I have written on this problem extensively in previous columns and papers (11). If the vendor's in this study used a different third party laboratory than I did, it makes sense that we might get different results. Ultimately though these results are a condemnation of the state of California's laboratory testing certification program. This obviously needs to be reformed to prevent cannabis patients from receiving incorrect doses and cannabis consumers from being ripped off.

Imagine the uproar if 77% of the bottles in a batch of ibuprofen were mislabeled? There would be a hue and cry, gnashing of teeth, reams of bad publicity, calls for investigations, sanctions, and possible criminal prosecution. And yet this same exact scenario takes place regularly in the cannabis industry. Why do we tolerate this?

Why Do Dispensaries Need In-House Testing?

To prevent these problems dispensaries need to do their own in-house testing for these reasons.

 Dispensaries are clearly selling mislabeled products. Medicines need to be labeled properly so patients get the proper dose. Dispensaries should test their stock before sale to make sure it is labeled properly.

 An incorrectly labeled product could cause harm to a consumer. If a dispensary tested its stock itself, it can make sure only correctly labeled medicines are sold, reducing liability.

3. Prevent Customers from Being Ripped Off:

 Cannabis consumers are entitled to get what they are paying for. If a jar of buds says it is 25 wt.% total THC, that is what the consumer should get. Based on the data above this is clearly not the case. If dispensaries tested their products before sale they could insure that only properly labeled products are sold, insuring consumers get what they pay for.

 Dispensaries set the price they pay suppliers based in part on potency. In many cases suppliers will provide certificates of analysis (COAs) from licensed laboratories to help determine price. However, there is nothing preventing suppliers from accidentally or intentionally giving the wrong COA to a potential buyer—and yes, I know this happens. Dispensaries need to protect themselves from being ripped off by doing on the spot potency tests of products before they buy them so they know what they are getting.

: Since we know that cannabinoids degrade over time (7–10), it makes sense for dispensaries to monitor what is on their shelves. This will prevent them from selling mislabeled or tainted products to their customers.

: By doing their own analyses, dispensaries can assure the public they are getting what they pay for and the products being purchased are safe. Our industry will not realize its full potential until the public can have full confidence that cannabis medicines are safe and effective.

Should Dispensaries Set Up Their Own Testing Laboratories?

Setting up a cannabis analysis laboratory with all the equipment needed to test for pesticides, heavy metals, terpenes, and potency is a million dollar plus proposition. This is probably not practical for most dispensaries. However, given that in most instances the API is THC or CBD, having the ability to measure these analytes would be the best way to insure correct dosages and accurate label claims. Many third party laboratories use HPLC or gas chromatography (GC)to measure potency (11–14).

I have pointed out in previous columns that amongst the criteria to use when judging an analytical method are speed, accuracy, and cost (11). I have also pointed out that chromatography is accurate, but can be slow and expensive (11). Infrared (IR) spectroscopy on the other hand can offer equivalent accuracy but will always be faster, cheaper, and easier than chromatography (11). There exist IR spectroscopy-based cannabis analyzers that work on flower (15,16), extracts (17), distillates (18), and tinctures (19) that cost around $30k (20). This is not cheap, but probably much more within the budget of a typical cannabis dispensary than the cost of a chromatograph.

Mislabeled cannabis medicines are an ongoing problem in this industry. We found 77% of the products purchased at dispensaries are mislabeled. Mislabeled medicines mean patients are receiving incorrect doses, consumers are being ripped off, and dispensaries are exposed to significant litigation. For these reasons, dispensaries need to test their products before they sell them. Suggested potency testing technologies were discussed.

https://en.wikipedia.org/wiki/The_dose_makes_the_poison

https://en.wikipedia.org/wiki/Paracetamol

https://mjbizdaily.com/nearly-a-fifth-of-california-marijuana-products-failing-testing-standards/

M.O. Bonn-Miller, M.J.E. Loflin, B.F. Thomas, J.P. Marcu, T. Hyke, and V. Ryan, 

Journal of the American Medical Association

C. Mazzetti, E. Ferri, M. Pozzi, and M. Labra, 

J. Fairbairn, J. Liebmann, and M. Rowan, 

I. Trofin, G. Dabija, D. Vaireanu, and L. Filipescu, 

B.C. Smith, Terpenes & Testing Magazine, Nov./Dec.(6), 48–51 (2017).

M.W. Giese, M.A. Lewis, L. Giese, and K.M. Smith, 

T. Ruppel and M. Kuffel, "Cannabis Analysis: Potency Testing Identification and Quantification of THC and CBD by GC/FID and GC/MS," PerkinElmer Application Note (2013).

B.C. Smith, M. Lewis, and J. Mendez, “Optimization of Cannabis Grows Using Fourier Transform Mid-Infrared Spectroscopy,” PerkinElmer Application Note (2016).

B.C. Smith, C.A. Fucetola, K. Ehrmantraut, and T. Hagan, 

, PHD, is Founder, CEO, and Chief Technical Officer of Big Sur Scientific. He is the inventor of the BSS series of patented mid-infrared based cannabis analyzers. Dr. Smith has done pioneering research and published numerous peer-reviewed papers on the application of mid-infrared spectroscopy to cannabis analysis, and sits on the editorial board of 

. He has worked as a laboratory director for a cannabis extractor, as an analytical chemist for Waters Associates and PerkinElmer, and as an analytical instrument salesperson. He has more than 30 years of experience in chemical analysis and has written three books on the subject. Dr. Smith earned his PhD on physical chemistry from Dartmouth College. Direct correspondence to: 

Articles

Cannabis Analysis 

Mislabeled cannabis medicines are an on-going problem; here we discuss appropriate technologies for cannabis dispensaries to do their own testing.

Why Dispensaries Need In-House Testing

Ethanol has maintained its position as one of the most scalable extraction methods because of its simple methodology and solvent properties. Over time, the technology used in ethanol extraction has increased the selectivity of the method and resulted in shortened post-processing times and increased purity. In this article, we will walk through the evolution of the extraction method within the cannabis industry, where it stands today, and the other alcohols that are being introduced to the cannabis extraction space.

Ethanol has been used for centuries as an extraction method and an ingredient to produce perfumes, food colorings and flavorings, medicinal bases, and essential oils. The US Food and Drug Administration (FDA) has found production of consumer goods using food-grade ethanol to be safe for human use and consumption, which cannot be said of other alcohols. It is the second most popular solvent behind water. It is also the least toxic of all alcohols, making it one of the most widely used solvents in consumer goods. Because of all the aforementioned characteristics, it is no surprise that ethanol has also maintained its position in the cannabis industry as one of the most widely used solvents for extraction of cannabinoids.

Ethanol is a polar solvent, but can have both polar and nonpolar properties. It attracts polar and ionic molecules, through its hydroxyl group, and can attract nonpolar molecules because of the nonpolar nature of the ethyl group. Hydrogen bonding of ethanol and water with the hydroxyl group can be seen in Figure 1.

Tetrahydrocannabinolic acid (THCA) and cannabidiolic acid (CBDA) are both polar compounds due to their acidic hydroxyl group. Therefore, they are both easily extracted with ethanol at room temperature. Chlorophyll is an undesired polar compound that easily coextracts with the cannabinoids during most methods of ethanol extraction. This is why crude ethanol extracts, such as modern products listed as Rick Simpson oil (RSO), have a dark green color and concentrations in the 40–60% range. Advancements have been made in methodology to optimize for the extraction of cannabinoids and exclude chlorophyll and waxes.

This method has been used for centuries and is considered a “traditional” medicinal preparation method for phytochemicals (2).

Plant material, like cannabis, is soaked in ethanol.

The plant material is then filtered to separate the solid from the liquid solvent-solute solution.

The solution is processed using rotary evaporation or falling film evaporation to remove the solvent. This results in a crude botanical extract.

Products, such as alcohol-based tinctures, can be produced directly from the extraction using the solute-solvent solution.

Minimal control over extracted target compounds

Can be a long extraction time depending on the input material, proof of ethanol, and temperature at which the extraction occurs.

With respect to cannabis extraction, the inability to control the temperature of the extraction and therefore the polarity of the solvent, results in an extract containing chlorophyll, water, and other compounds. If the goal of the extraction is to purely extract cannabinoids, this method produces an extract that is not.

This method has been used for decades and is considered a “traditional” medicinal preparation method for botanicals.

Plant material, like cannabis, is packed into a column that is positioned between the boiling flask containing ethanol and a
condensing column.

The ethanol is heated to produce a vapor, which will then interact with the plant material such that an extraction occurs. The resulting solution is a mixture of solvent and compounds that are soluble in ethanol at the vapor temperature of ethanol.

The ethanol-extract solution then collects in the boiling flask.

Plant material is contained separately from ethanol-extract solution, eliminating the need to filter the plant material from the solution.

As with the maceration method, the result is a cannabis extract that is far from pure.

This method has developed over the past decade to improve the selectivity of ethanol and optimize extraction of cannabinoids.

Plant material is loaded into a mesh bag or basket and placed in the extraction chamber.

Ethanol is chilled to -40 °C then introduced to the extraction chamber soaking the plant material and the solute is extracted with the ethanol that flows through the plant material.

The ethanol-extract solution is then pumped out of the extraction chamber to a collection vessel to be processed further.

The solvent is evaporated using rotary evaporation or falling film evaporation.

The extract can then be used to formulate final products or distilled further for higher purity of cannabinoids.

Increased solvent selectivity, resulting in a more pure-cannabinoid extract

High electricity usage to chill the solvent

Liquid-Solid Mix and Separation Centrifuge Extraction with Cold Ethanol

This is the most common method used in cannabis and hemp ethanol
extraction today.

Plant material is ground to a uniform size.

Plant material is added to a mesh bag and placed inside a centrifuge basket.

Ethanol is chilled to -40 °C then introduced to the centrifuge chamber where the ethanol soaks the cannabis.

The centrifuge is powered on and churns like a washing machine to mix the solvent and plant material while the solute is extracted.

After the extraction, the extracted solution is pumped out of the centrifuge basket into a collection vessel.

The solution is then introduced to a solvent removal process, like falling film evaporation, which leaves the final extract complete and ready to formulate.

Separation of liquid and solid occurs simultaneously with the extraction, eliminating a filtration step

Equipment maintenance—ethanol can easily cause fast wear on bearings that are integral to high speed centrifuges

Highly laborious to continuously remove and refill the centrifuge with plant material

Note: As the process scales up, a noticeable decrease in yield has been acknowledged as a result of more ethanol being used and remaining within the plant material. This reduces the yield and increases processing time in the evaporation stage.

Other alcohols, like isopropyl or methanol, have been utilized in botanical extraction for many years. Isopropyl alcohol and methanol are significantly lower in cost than ethanol, but have their own pros and cons including being toxic to inhale or ingest at far lower concentrations than ethanol (6). Methanol has a lower boiling point, which also indicates a lower polarity. Being less polar than ethanol, methanol is less efficient at extracting THCA and CBDA. Isopropyl alcohol has a higher boiling point and higher polarity, this results in higher yields of cannabinoids in a shorter amount of time, but also extracts higher quantities of chlorophyll. The evaluation of solvents is heavily focused around the cost of a solvent, a company’s ability to remove all of the residual solvent to minimize consumer risk, and post-processing time and methods. These alternative alcohols are strong candidates for any business that further refines the extract using distillation. Wiped or thin-film distillation can produce a pure cannabinoid distillate, reducing the need for purity in the crude form. 

Extraction Science

A review of the evolution of the ethanol extraction method within the cannabis industry, where it stands today, and the other alcohols that are being introduced to the cannabis extraction space.

The Evolution of Ethanol Extraction Methods in Cannabis

The analytical scientist is often tasked with a difficult job of being accurate, efficient, and expedient in their work. Often these focused goals do not allow a lot of time for research and fine tuning of their workflow. Many chromatographers use methods they find from manufacturers or technical sources and adapt them to their situations and analyses. Sometimes the methods get adjusted or “tweaked” to improve the fit or increase the output but, unless the scientist is doing research or creating a totally new method approach, they build upon the backs of other methods. In this column, we take a deeper look into the chemistry, physics, and methodology of high performance liquid chromatography (HPLC) methods. We will look at how columns function and what changes can be made to increase resolution, efficiency, and separation to reach a laboratory’s chromatographic goals.

The start of method development is identifying and understanding the physical and chemical nature of the analytical targets and the scope of analytical instruments. Methods are a complete package of the chemistries, modalities, and functionality of all the targets, sample preparation, instrumentation, chromatographic phases, and parameters that allow for the separation, identification, and quantitation of the analytes of interest. 

Often, ambitious analysts try to create a single method for all their analytes under one process and one instrument and end up with lots of mediocre data. Equally frustrating is setting up a method with standard conditions and settings without fine tuning it to the samples and analytes then becoming disappointed when the method “doesn’t work.” It is important to recognize that there will be instances where the sample preparation or the analysis method will require separate processes or “tweaking” to report all the target analytes efficiently and accurately. There is no one size fits all or straight out of the box solution. 

I am sorry to say there is no one piece of instrumentation that will fill all needs. Some technologies have a wider range of targets, such as liquid chromatography–mass spectrometry (LC–MS) versus gas chromatography (GC)–MS, but each technique has its limitations and uses. Instrument choice is often dependent on the chemistry of target analytes and their potential analytical concentration. For instance, in cannabis there are several classes of organic analytes that are routinely examined including cannabinoids for potency; terpenes and flavonoids for identity, flavor, fragrance, and chemical fingerprinting; and pesticide residues or mycotoxins as potential contaminants. By their very nature all of these compounds occur in vastly different concentrations. Trace analysis (low parts-per-million [ppm] or parts-per-billion [ppb]) is the range for any method created to quantify potentially dangerous contaminants such as pesticides or mycotoxins and requires systems such as GC–MS, LC–MS and LC–tandem MS (MS/MS), which have sensitivity in those low ranges.

Cannabinoids and terpenes on the other hand can occur in the high ppm level up to the percent level and require a different range of analysis such as what is found in high performance liquid chromatography (HPLC) coupled with ultraviolet-visible (UV-vis) detection and GC coupled with flame ionization detection (FID). These two ranges of analysis often cannot be produced and quantitated by a single method of an instrument without significant changes to the samples.

The accuracy and ability to quantify analyte concentrations depends on the instrument's analytical specification levels and dynamic range often bracketed by the level of detection (LOD) and level of linearity (LOL)(Figure 1). The lowest limit of an analytical system is the 

(LOD), this is the point where a target can be differentiated from a blank or noise with a high degree of confidence (usually over three standard deviations from the noise or blank response). The highest level of accurate quantitation ends with the LOL, where the linearity of the system starts to skew often due to detector saturation. Peaks that reach LOL appear broad, flatten at their apex, or are cut off before their apex. 

The range of the most accuracy (dynamic range) is between the LOD and LOL starting at the limit of quantitation (LOQ). The 

 (LOQ) is the lower limit of a method or system, which the target analyte can be reasonably calculated (over 10 standard deviations from the blank or baseline response). 

A simple method to determine if a response peak reaches the cutoff for LOD or LOQ is to look at the ratio of signal-to-noise (S/N). A blank baseline in chromatographic systems is rarely flat and straight. The lowest points of the chromatogram are a combination of the true baseline and system noise. 

 is the sum of all the random variations (electrical, temperature, and so on) and contamination or interference from the chemical components. 

To determine if a peak can be quantified, one can either compare relative heights or relative areas. In comparing relative height, the analyst averages the mean height of the noise and compares it to the height of the target peak from the noise mean height. To compare areas, one or more “peaks” in the noise are integrated with similar width to the target peak and the areas are compared. If the ratio is greater than three then it qualifies as within LOD and if the ratio is greater than 10, then that peak can be used for quantitation (LOQ).

The best practice is to integrate the noise at the baseline at several points and average the baseline noise responses then compare to the integrated peak of interest (Figure 2). For example, if the chromatographer is interested in peak A, they should integrate areas of baseline near the peak of similar peak widths to the target or measure its height compared to the average range of peak heights found in the noise. If the average of those baseline noise peaks is 100 units (height or area) then peak A must be at least 300 units to meet the LOD criteria of 3X. If peak A is only 200 units then it fails and cannot be used for either identification or quantification. If peak B is the peak of interest and has a S/N >3 but <10, then it can be used for detection (LOD) and possibly identification but should not be used for quantification (LOQ). Finally, a peak such as C can be used for both identification and quantification because its S/N value is higher than 10. 

Instrument sensitivity (represented as S/N) can be increased by decreasing baseline noise without increasing the target response. Noise in a system can be created by matrix from the extracted sample, contamination of the sample, and contamination of either the stationary phase or mobile phase. Proper sample clean-up and processing can sometimes reduce baseline noise, so the target peak is not “lost in the weeds” of the baseline. Often it is believed that one can get a better response by injecting a larger sample aliquot, but if the sample matrix is a contributor to the noise, then a larger sample means more matrix as well and will not necessarily help with the issues of S/N. 

As for the mobile phases, impurities in mobile phase can directly affect baseline noise. The wrong grade of gases in GC can create high baselines while HPLC mobile phases can accumulate contamination by exposure to the laboratory environment. Replacing old solvents with fresh solvent can dramatically lower HPLC baseline noise. In some cases, especially in LC–MS and ppb analysis, the use of highly filtered LC–MS solvents can also play a role in reducing baseline noise. 

Solid phase contamination and build up can play a role in baseline noise. As columns age, the backbone materials—silanes, siloxanes, and so forth—can break down or lose protective end capping, which increases noise. Harsh or acidic HPLC mobile phases can strip column phases and promote column breakdown. By examining the chromatographic baseline and cleaning up the contributing factors it can ensure that more target peaks fall into the LOQ needed for many analyses. 

HPLC is a powerful analytical tool that needs extensive adjustments to maximize its accuracy and efficiency. All the adjustments and fine tuning of parameters ultimately are based on the targets of interest. Generally, chromatographers group targets as either polar or nonpolar analytes, with some fluctuating between the two classifications. Nonpolar analytes such as alkanes and alike often are targets for normal phase chromatographic methods while polar analytes such as carboxylic acids, will be examined with reversed-phase chromatographic methods. The majority of modern HPLC analysis falls under the classification of reversed-phase LC.

The parameters of reversed-phase LC dictate that the mobile phases used with be polar solvents while the stationary phase (column) will be nonpolar. Target analytes that dissolve in polar solvents and water will then use either ion-exchange columns or reversed-phases columns depending on their chemistry (Figure 3). The most common reversed-phase LC columns are the C18 or C8 columns since they service a wide variety of chemistries. Column manufacturers may change or supplement the chemistries of their standard C18 columns with different modifiers, which can assist with particular issues such as resistance to highly aqueous methods or better retention of highly polar compounds.

The chemical properties (in addition to the concentration) of your targets will dictate—to a degree—the type of instrumentation that will be needed for analysis. As was stated previously, the concentration of analytes can dictate instrumentation such as UV over MS. But the chemical nature of the compounds will play an important role not only in selecting the right column but the right detector. Compounds that are easily ionized can be detected using electrospray ionization (ESI) with LC–MS. But, if your compounds are not as easily ionized then other atmospheric pressure ionization sources, such as chemical and photoionization (APCI and APPI) need to be considered if LC–MS is the method of detection. Molecules with chromophores can utilize UV-vis detectors, while fluorescent molecules can utilize other chemiluminescent detectors. 

Once the chromatographer understands their target analyte’s range requirements, selected instruments, detectors, and targeted possible column chemistries it is time to start building or refining an analysis method.

One of the first questions asked when researching a new method is: Has anyone already created a method for my analysis? The good news is that most of the time someone, somewhere, has performed a similar analysis on a similar class of compounds or a similar type of material. The bad news is that the methods you find usually are a starting point and not an out-of-the box solution. Methods that are found in journals and websites are not necessarily a perfect fit for your laboratory’s instrument configurations nor are they all validated for the purpose you might need (if they are validated at all). 

In some cases where there are different size columns, tubing, particle, or pore size, it is helpful to use an HPLC method translation tool to help alter flow rates and gauge retention time. One free method translator tool can be found at: 

https://www.acdlabs.com/resources/freeware/translator

. A downloadable excel calculator can be found at: 

https://ispso.unige.ch/labs/fanal/hplc_calculator:en

. Method translation tools do not always take into consideration what may be practical for your system. They are usually just sophisticated calculators. The resulting parameters may exceed the capacity for your system (that is, the back pressure created will be too high, or the injection volume cannot be achieved with the system). Again, these parameters are meant to be a starting point for optimization and must be examined with a critical eye.

In initial method development, it is best to start by separating targets from either a standard or a well-characterized sample to achieve separation, then use various concentrations and matrices to improve sensitivity and response. An initial chromatogram may have multiple unretained or overlapping (coeluting) analytes. It is also possible to have peaks retained on the column past the method run time (Figure 4). It is also possible that “negative peaks” (peaks or dips below the baseline) can be seen in the chromatogram, usually appearing right before the unresolved peaks. 

Negative peaks are a visual representation of the difference or change in detector response that result in response below the baseline or set point. These peaks can be caused by a number of issues including the movement of internal values such as when the injector values switch back and forth from injection to normal flow path. Another way negative peaks can be observed is when the solvent used to dilute your sample differs from the solvent composition of the HPLC system. For example, if your extracted sample of cannabis is in 100% ethanol, but the HPLC mobile phase starts as 50:50 water–acetonitrile; then you may have a significant negative peak as the ethanol hits the detector due to changes in absorbance, refractive index,
or conductivity. 

In any method development it is best to remember the old adage: “Slow and steady wins the race.” Change one parameter at a time and then move on. It is very tempting to change multiple parameters (mobile phase, additives, flow rate, temperature) all at once to save time and effort. Unfortunately, often it dilutes the effort, and you cannot track which changes give you the
best resolution.

There will be times that there are only general methods available that are remotely similar to your intended analysis; in that case basic parameters become the starting point on which to build a method that provides resolution of target analytes. 

—the ability of a column to separate peaks in a chromatogram—is a complex interaction of forces and factors observed as adjacent peaks merging (coeluting) to a single peak or separating (resolving) into two or more peaks. Resolution is the relationship between the differences in retention times of two adjacent peaks divided by the sum of the peak widths of the two target peaks at the baseline. 

 or Rt is the time between injection and the appearance of the peak maximum or apex. The first step in evaluating another analysts’ method is to try and duplicate all the conditions, instrumentation, columns, mobile phase, and settings described then run a known test sample or standard to obtain a starting chromatogram. 

 (w) is the width of the peak or chromatographic band measured at the baseline by drawing tangent lines from the inflection points of the curve that represents the peak. The resolution between two adjacent peaks of similar heights can be calculated using the resolution equation (Equation 1): For example, if an analyst is separating cannabis terpenoids and there are two terpenes that resolve at 7.5 min and 8.0 min with peak widths of 0.5 min and 0.75 min would have R = 2(8.0 – 7.5)/(0.5+0.75) = 0.8

Peak resolution needs to have R >0.6 to be able to discern separation between two peaks of equal height. Two peaks with values of R >1.5 are considered to have good resolution and peaks of R >2 are fully or baseline resolved. In our example, the two terpenes would be coeluting with some minimal separation at the apexes (Figure 5).

For a cannabis analyst who needed to separate all the terpenes, cannabinoids, pesticides, and so on, effort to increase resolution would have to begin. According to the first resolution equation, resolution is affected by peak width and retention time. To improve resolution the analyst would have to either increase the retention time of one of the peaks (meaning that they would have to find a way to keep the peak partitioned in the stationary phase for a longer time) or decrease the peak width. There are three basic HPLC parameters that the chromatographer can use to change conditions of their method: mobile phase composition, solid phase chemistry, and flow rate. The usual first step to increase the amount of time an analyte spends in either phase is to change the composition of the mobile phase, usually by changing the gradient program. 

Peaks that elute early in the chromatogram and spend too much time in the mobile phase and are not being retained well by the stationary phase. These compounds are more polar or are hydrophilic. Most reversed-phase LC gradient methods start with a high concentration of water (50–80%) and a lower concentration of organic mobile phase. The general rule of thumb for reversed-phase gradient methods is to start organic solvents low and increase to high. If you have early eluting compounds it means you either have to start with slightly less water mobile phase or change the stationary phase chemistry (column). Changes to the flow will not help unretained peaks and very early eluting compounds.

Peaks that elute late in the chromatogram spend too much time with the stationary phase and only elute when the organic mobile phase concentration is high enough to force them from the column. These compounds are less polar and more hydrophobic. To speed up late eluting compounds the organic phase can be increased more rapidly by increasing the gradient slope. Changes in the stationary phase chemistry can also be used to resolve late eluters. Sometimes increasing flow rate can help stop late eluting peaks from being lost if they are retained on the column until after the program ends. Some instruments allow for changes in flow rate over the method run and increasing flow rate may assist with late eluting peaks in this approach.

Generally, flow rate is not the first choice for peak resolution. Resolution is a function of peak width and time, which we observed in equation 1, so by increasing flow you do decrease the peak width, but you also decrease the difference in retention times, which has a bigger impact on resolution. But resolution is also a function of three other factors: retention (or capacity), separation (or selectivity), and efficiency characterized in Equation 2 for isocratic methods:

 (k) is the time that a sample analyte resides in the stationary phase relative to the time it spends in the mobile phase. It is calculated from time difference between the analyte peak (tR) and the unretained peak divided by the time of the unretained peak (t0)
(Equation 3).

 Capacity factors or retention values less than five have the biggest effect on resolution. For example, if k = 10 then in the resolution equation 2 – the factor for retention is 10/(1+10) which is close to 1; but if the k = 1, then the retention factor is 1/(1+1) or ½ and will have a larger impact
on retention.

 (α) is the ability of a chromatography system to distinguish the chemical differences between compounds. It is measured as a ratio capacity factors (k) of the two peaks being examined (Equation 4).

 (N or H) is a measure of theoretical plates in a chromatographic system. Theoretical plates are a concept described by the Nobel prize winners Martin and Synge in which they related the theories of distillation to chromatography. In distillation columns for crude petroleum, there are actual separator plates to help separate and isolate petroleum fractions. Early petroleum distillation used different length columns that related to its efficiency. Chromatography columns do not have physical plates that can be measured, so the plates are theoretical to describe the efficiency of the column. In this regard, the length of the column is called a height equivalent to a theoretical plate (HETP) and is measured as column length (L) divided by theoretical plates (N). Efficiency or theoretical plates are calculated as by the relationship between peak retention time (tR) and the peak width at baseline (wb) (Equation 5).

If peak width is taken at the half height point in the peak it is called 

 (w1/2) and is multiplied by 5.54 instead of 16. Efficiency (N) can also be calculated as a relationship of retention time (tR), peak height (hp), and peak area (A) (Equation 6). Finally, efficiency can be measured as a relationship between separation factor (α), column length (L), and particle size (dp) (Equations 7 and 8).

If the chromatographer wants to increase resolution, then they must find a way to increase one of the three factors: retention, selectivity, or efficiency (Figure 6). In many cases, the first place to start is altering the mobile phase instead of changing columns (stationary phase). The goal is to increase the time the peaks spend in the mobile phase. Increasing retention will result in peaks spending more time in the solid phase, moving the peaks apart but it can result in 

 (increased peak width over time of analysis) and this approach only works best on early eluting peaks (which the retention factor is less than five). One way to increase retention (k) is to use weaker solvents or ones with lower elution strengths. 

 (ε0) was introduced. Eluotropic value (also known as elution strength) describes the ability of a solvent to pull an analyte from the solid phase. The larger the ε0 of the solvents of the mobile phase relates to less retention (Table I). Water has a high ε0, so using a lower concentration in the mobile phase can increase retention. Switching solvents for lower ε0 solvents such as changing methanol (ε0=0.95) for acetonitrile (ε0=0.65) is the best way to increase retention, but will increase in method analysis time. The customary approach for developing methods for a sample with many types of analytes is to increase k for early eluting compounds and decrease k for late eluting compounds.

Since selectivity is related to retention, many of the same approaches such as changing solvents is effective as a method to change selectivity. Selectivity can also be changed by chemical effects to solvent through additives and pH modifiers. The chemical nature of the target analytes become important, especially in the understanding of pH, pKa, and pKb which are the basis for all acid-base interactions. The letter p means that value or measurement is based on a logarithmic (-log) value or measurement. Then pH is the measurement of hydrogen ion concentration of a compound in an aqueous solution, where pH = -log [H+] and equates to relative acidity or alkalinity of a compound. The scale ranges from 0 to 14, where low values equate to acidic compounds and high values are alkaline. Neutral for the scale is pH of 7. The pH of a solution indicates an acid or base, but not necessarily the true strength of that acid or base. To measure the strength of an acid or base, one needs to look at the pKa or pKb. 

 are the -log of Ka and Kb (acid and base dissociation constants) that predict if a compound will donate or accept protons at a specific pH. These values are true indicators for the strength of an acid or base since they are unaffected by the amount of water added to the solution. For acids, smaller pKa values indicate stronger acids while in bases smaller pKb indicate stronger bases. By adding buffers to mobile phase, the chromatographer can change resolution (Table II). As methods are being developed, always start with weaker acids at the lowest functional concentration possible (~ 0.01 M) and increase in acid strength and concentration until the desired results are achieved. High levels of buffer (>0.1 M) since extreme pH (very acidic or very basic) can damage columns and increase they viscosity of the mobile phase thereby increasing pressure in the system. Most columns are safe in the range between pH 2 to 7.5 (but consult column specifications). For basic or mixed sample types, mobile phase pH near neutral increases retention, while acidic sample are aided by acidic mobile phase.

If an acid analyte is in a mobile phase whose pH is similar to the target analyte’s pKa, then there is lower degree of ionization and an increase of retention can be observed. When mobile phase pH is significantly over or under an analytes pKa or pKb then the degree of ionization increases and reduces retention. So, protonated basic analytes may have low retention at acidic pH while acidic analytes will have increased retention. 

The reverse can also be true that basic analytes in a high pH mobile phase will increase retention where acid acids will decrease in retention. Differences between target pKa (or pKb) and mobile phase pH and pKa are of even more importance in LC–MS where ionization is critical to detection. Keeping mobile phase (and buffers) pH and pKa similar to the target improves resolution because it moves the ionizable target to a more neutral state, but for LC–MS to be effective the compound needs ionic character or to have the ability to be easily ionized. Therefore, the rule of thumb for pH, pKa, and mobile phase is to aim for mobile phase pH 1–2 units from the target analytes pKa within the specification range of the column.

It is important to remember that not all buffers are appropriate for all applications. Salts can precipitate out of solution as mobile phase composition changes. Salts are also problematic for LC–MS analysis since they can inhibit ionization. If solids are dissolved into mobile phase it is important to remember to filter the mobile phase and, in some cases, the mobile phase may need to be heated to dissolve solids completely. Changes in column compartment temperature may aid in keeping buffers in solution.

In some cases, temperature of the column compartment can be used as a tool to alter the retention and selection of a method. Keeping a constant temperature can often help with fine tuning a retention time by eliminating external temperature changes. Some reduction in retention (k) can be achieved by increasing temperature allowing for peaks to elute earlier. Many changes to resolution (retention, selectivity, and efficiency) can be made by the manipulation of the mobile phase chemistry (Figure 7).

There will be times when all the changes to mobile phase have not resulted in the desired resolution (retention, selectivity, and efficiency) so then it is time to consider a different column chemistry or different physical column parameters. 

In reversed-phase LC, the most retention is usually found in C18 columns followed in descending order to C8, C2, and so forth. The lowest retention is found in phenyl and cyano columns. Most reversed-phase LC systems have some form of C18 column as part of their standard materials package from a manufacturer upon set up and is often the most used or tested column to start method development. The specifications for the column should designate its internal diameter (mm), length (mm), particle size (µm), and pore size in Angstroms (Å ). A fairly standard size for a starting column is 4.6 mm x 100–150 mm, 5 µm and 80–100 Å.

The physical column characteristic can have a large effect on the resolution, system pressure and analysis time. First, let us look at column length. Increasing the length of the column has a direct effect on efficiency from Equation 7. Doubling the length of the column does give you a resolution increase of about 1.4x, but it also doubles the pressure and analysis time. Resolution is increased at the expense of time and pressure. So conversely, if you want to decrease back pressure and decrease analysis time then a shorter column is needed, but that will reduce resolution.

The second parameter of the resolution from equations 7 and 8 is particle size. As the particle size increases (without any change in column length), resolution, and efficiency (theoretical plates) will decrease. Smaller particles increase resolution, efficiency, and retention but they also will exponentially increase pressure of the system: P α (1/dp)2. Modern HPLC systems have increased their ability to handle higher pressures and often can handle increased pressure from smaller particles. If smaller particles are needed to maximize resolution, then the chromatographer can reduce column length to reduce pressure and not effect efficiency dramatically since efficiency is inversely proportional to particle size and directly proportional to length. The benefits of this approach are shorter run times, less solvent use since column volume is reduced, and less backpressure with increased resolution. 

The diameter of a column becomes important in the discussion of pressure, flow, and column volume. Columns of diameters between 3.9–4.6 mm range are considered to be standard bore. Diameters in the range of 2.0–3.2 mm are called narrow bore. Columns with diameters of 1.0 or less are considered microbore or capillary columns. The diameter of the column effects peak height, which can also affect efficiency as seen in Equation 6. Changes in column diameter must also consider changes in 

 (which is the volume of mobile phase divided by time) and 

 (µ), which is the distance mobile phase travels over time. By adjusting flow rate to maintain linear velocity, peak height is increased, and band broadening is decreased allowing for improved retention.

One of the final dimensions of a column is pore size. The stationary phase is composed of particles that are riddled with spaces and openings. The spaces allow for increased surface area and interaction between the target analytes and the stationary phase. Typical pore sizes can range from 50 Å to over one million Å. The choice of pore size is dictated by the approximate range of the molecular weights of the various target analytes. Most small to mid-range molecular weight analysis (<1000 mw) can utilize 50–100 Å pores. Some specialty molecules with stereochemistry considerations could require larger pore sizes to accommodate steric interactions with the stationary phase. Small pores have larger amounts of surface area than larger pores and allow for much more interaction between molecules and the stationary phase. 

Navigating the Lab

A deeper look into the chemistry, physics, and methodology of HPLC methods.

Looking with Light: Breaking Down Liquid Chromatography Method Development

The National Institute of Standards and Technology (NIST) produces thousands of standard reference materials and has a long history of developing quality assurance programs for improving analytical measurements. Past programs have helped laboratories accurately measure compounds in vitamins, dietary supplements and environmental contaminants. In this month’s “Cannabis Crossroads,” I interview Dr. Walter Wilson, who is coordinating the Chemical Services Division cannabis research program at NIST, with a focus on developing cannabis reference materials and a quality assurance program (CannaQAP). Dr. Wilson is also working on the development of chromatographic methods to separate cannabinoids in complex cannabis matrices such as dried plants, extracts, concentrates, edibles, and so forth. Here, Dr. Wilson shares his progress with us.

It was recently announced that NIST will help laboratories distinguish hemp from cannabis. Can you please tell us more about the study that NIST is inviting laboratories to participate in?

NIST has developed an integrated measurement services program for forensic and cannabis (hemp and marijuana) laboratories to help ensure the quality of routine analysis. The tools developed by this cannabis program will support measurements establishing legal, commerce, and safety claims through the development of fit-for-purpose analytical methodologies, production of fit-for-purpose reference materials (RMs), and implementation of a cannabis quality assurance program (CannaQAP).

CannaQAP was developed to help forensic and cannabis testing laboratories demonstrate and improve measurement comparability and competence. CannaQAP is a perpetual inter-laboratory study mechanism that is similar to a proficiency testing (PT) scheme but without the pass/fail grade. The first exercise of CannaQAP focused on the determination of tetrahydrocannabinol (THC), tetrahydrocannabinolic acid (THCA), total THC, and up to 15 additional cannabinoids in two hemp oil samples. This exercise was completed in November 2020 with 116 participants and the final report will be publicly available soon. The second exercise of CannaQAP is focusing on the determination of cannabinoids, toxic elements, and moisture in multiple cannabis ground plant samples (hemp
and marijuana).

Is this study for cannabis QC laboratories, forensics laboratories, academic laboratories, or everyone? Can you share with us the goals and objectives of this study and exactly how it will work?

The call for sign-ups closed on March 5th with approximately 230 participants, including 22 international laboratories, with the goal of improving the measurement comparability and competence across both the forensic and cannabis communities. Exercise two participants consist of both hemp and marijuana testing laboratories (60%), forensic laboratories (21%), regulators (7%), instrument manufacturers (≈5%), academic laboratories (≈5%), and RM/PT providers (≈2%). We will ship samples to participants in April 2021 and accept results at our 

 through May 28, 2021. NIST will provide the participants with a preliminary data summary including NIST results in mid-June, which will give participants an opportunity to review and revise their submissions to correct any errors. The final report will be peer-reviewed and made publicly available as a NIST Internal Report (IR), however the identities of individual laboratories are anonymized and known only to them and NIST.

Although the sign-up period has already passed for exercise two, CannaQAP does plan to continue to conduct two exercises a year and include different analytes (such as pesticides and terpenes) and matrices (for example, extracts and edibles) in future studies.

Cannabis QC laboratories are now analyzing 21 or more cannabinoids and many contaminants. Which compounds will be analyzed in this NIST study?

The first exercise of CannaQAP focused on the measurement of up to 17 cannabinoids (see Table I), total THC, and total CBD in two hemp oils, which were selected based on the available commercial standards at the start of exercise one. To keep a link between exercise one and exercise two, the same cannabinoids are being measured in the six different cannabis plant samples. In addition, toxic elements (As, Cd, Pb, Hg, Be, Co, Cr, Mn, Mo, Ni, Se, U, and V), and moisture studies were included in exercise two.

When is the deadline to submit and how long will the study run? Can you point us to a URL or share more details on this project?

The registration for exercise two has already closed. Updates about the CannaQAP are posted on our 

 and the CannaQAP website. NIST recommends for any laboratory interested in being a part of future CannaQAP studies, please create an account on the CannaQAP website. There is no charge for participation in the program, but during the account registration process laboratories are required to provide a shipping account number to which future sample shipping costs will be charged.

In July 2021, exercise three sign up will be announced via email to all CannaQAP accounts and sign up will remain open for approximately one month. Participants who wish to receive cannabis samples with total THC mass fraction greater than 0.3% will be required to provide NIST with a Drug Enforcement Administration (DEA) 222 form (domestic) or import permit (international). NIST will ship samples to the participants in October 2021 with testing results due in
November 2021.

How will samples be sent over state lines if this is federally illegal?

NIST has the proper DEA registration to permit the handling of legal hemp or controlled marijuana materials as research samples and transfer of these samples to other DEA registered laboratories. NIST will verify the THC mass fraction of all samples via analytical measurement prior to shipment and provide documentation to ensure all laboratories will receive only materials that they are legally permitted to receive. After consultation with the DEA, NIST requires that all laboratories that wish to receive cannabis samples with THC mass fraction greater than 0.3% must provide NIST with a completed DEA 222 form, which allows for the transfer of control substances between DEA registered laboratories. All submitted DEA forms are reviewed by NIST and DEA for completeness
and accuracy.

There are a wide variety of methods used to determine cannabinoid concentration such as high performance liquid chromatography (HPLC), ultrahigh-pressure liquid chromatography (UHPLC), gas chromatography (GC), mid-infrared (IR) spectroscopy, liquid chromatography-mass spectrometry (LC–MS), and so on. What technologies will be accepted in this study and will data be compared to peer groups (akin to a College of American Pathologists [CAP] survey in clinical laboratories)? What technologies is NIST using in your own laboratories? Is there a “gold standard” technique?

CannaQAP has laboratories use their normal in-house analytical methods to test program samples. NIST also provides a list of several analytical methods currently under development by standards organizations such as ASTM International and AOAC International.

NIST will measure cannabinoids in all CannaQAP samples using liquid chromatography-ulraviolet (LC-UV), LC-tandem mass spectrometry (MS/MS), or GC–MS, which will also be used for assigning mass fraction (%) values to future cannabis reference materials.

NIST is a leader in providing high quality reference standards. Can you tell us more about your efforts to develop hemp reference material? When can we expect to see these reference materials available to laboratories?

NIST has started working on a hemp plant reference material for the determination of cannabinoids, toxic elements, and moisture with a target release date in early 2022. NIST does plan to expand beyond the initial reference material to focus on additional analytes (such as, pesticides and terpenes) and sample matrixes (such as oils, extracts, and edibles). NIST began with plant reference material development to help address the analytical challenges in confidently distinguishing legal hemp from illegal marijuana samples in forensic and cannabis testing laboratories. In both cases, the laboratories need mechanisms to demonstrate competence in making these measurements, which participation in CannaQAP or use of reference materials can provide. The production of reference materials is a lengthy process and we hope that samples provided to participants through CannaQAP can help fill the gap until NIST is able to provide reference materials beyond the initial ground plant material.

Is there anything else that you would like to share about this exciting project?

At the end of 2020, the same cannabis team was awarded a $350,000 grant by the National Institutes of Justice for 2021 for the development of a GC–MS method and evaluation of multiple portable IR spectrometers to evaluate THC content of seized cannabis samples in forensic laboratories. Before the passage of the 2018 Farm Bill, forensic laboratories were only required to confirm the presence of cannabis through microscopic evaluation and GC–MS confirmation of the presence of THC in a seized cannabis sample. Forensic laboratories are now being asked to quantify THC to determine if seized cannabis samples are legal hemp or illegal marijuana, but they do not have the necessary tools to make these measurements. This project focuses on development of these methods at NIST and their transfer into forensic laboratories with standard operating procedures.
CannaQAP can then be used by these laboratories as an analytical test to help them validate these methods and their continued performance.

Will you or your colleagues at NIST be willing to share study results at a future Cannabis Science Conference?

All the results from CannaQAP are published as NIST IRs and will be publicly available through the CannaQAP website. In addition, NIST will be presenting results from these studies at many conferences in the future and would be happy to include the Cannabis Science Conference at some point in the future.

 is the columnist and editor of “Cannabis Crossroads” and a contributing editor to 

 magazine. Crossney is also the president and CEO of Cannabis Science Conference. Direct correspondence to:
Josh@JoshCrossney.com

Cannabis Crossroads

Dr. Walter Wilson discusses NIST's focus on developing cannabis reference materials and a quality assurance program (CannaQAP).

Is This Cannabis or Hemp—NIST Helps Cannabis Laboratories Achieve High-Quality Measurements

When using 3M rapid yeast and mold petrifilm for the analysis of dried cannabis flower

(Cannabis spp.), the manufacturer suggests two different incubation temperature ranges, 25 °C ± 1 °C or 28 °C ± 1 °C. In this study, dried cannabis flower found to have yeast and mold during compliance screening were randomly chosen to be plated at both temperatures to compare total amount of yeast and mold growth. These plates were incubated between 60–72 h in accordance with manufacturer guidelines for microbial testing. Overall, it was found that all plates that exhibited growth showed higher colony forming unit (CFU) counts at 25 °C and some observed to be statistically significant in their difference in favor of 25 °C. I urge all cannabis microbiologists and state compliance boards to require 3M rapid yeast and mold petrifilm, when testing dried cannabis flower, to be incubated at 25 °C to ensure accuracy in testing and safety of the consumers.

There is such a diversity of microorganisms on dried cannabis flower (

) that it has led us to selectively test for the ones harmful to humans. These guidelines are a pass–fail system that are regulated on a state-to-state basis, and can be set up as either a colony forming unit (CFU) threshold or through presence and absence testing (1). One area of concern for cannabis has been yeast and mold contamination of the flower through the growing and drying process. Several different types of yeast and mold have been found to find a suitable environment in cannabis production conditions through a speciation process (2). Cultivators have also now started adding different microorganisms, including fungi, to help facilitate the growth of their product in a way to be more natural in their growing situations (3). The problem stems from that some of the microorganisms, in large enough quantities, can lead to human harm if not properly tested for (1).

The overall purpose of this study is to compare CFU counts at two different manufacturer recommended temperatures to determine the optimum temperature for measuring yeast and mold on dried cannabis flower. Dried cannabis flower is defined as the crude flowering part of the cannabis plant (4). This is generally the main part of the cannabis plant that is tested. The manufacturer of petrifilm, 3M, recommends two different temperatures for the incubation period of their rapid yeast and mold product Cat: 6475/7477. These temperatures are 28 °C ±1 °C or 25 °C ± 1 °C. After reaching out to 3M their response stated that there were two different ranges because some yeast and mold organisms are better recovered at 25 °C while others grew better at 28 °C. Following this inquiry, the question was asked is there a difference between 28 °C ±1 °C or 25°C ± 1 °C for incubation temperature? Secondly, if there is a difference in temperature, which of the two temperatures would be the optimal temperature for dried cannabis flower?

Five client samples that were found to have a total yeast and mold count under full panel compliance testing were chosen at random to be used in this temperature comparison. These samples were crushed to form a homogeneous cannabis mix in a sterile bag and transferred to the weighing area. Next, 1 g from each sample was weighed out aseptically into a sterile filtered bag to which 99 mL of sterile butterfield’s diluent was added. Following that, 1 mL of this solution was transferred to petrifilm and pressed. This process was repeated 50 times on to 3M rapid yeast and mold petrifilm for each of the five samples. This was done in accordance with the instructions given by the manufacturer and Association of Official Agricultural Chemists (AOAC) recommendations (2,4,5). Petrifilm is a simple, ready to use plate system with specialized antibiotics to facilitate the growth of only yeast and mold on the plate in a rapid manner (5). These 50 replicates were then equally divided randomly to either be incubated at 28 °C ±1 °C or 25 °C ±1 °C. Following the state of Nevada and 3M guidelines for incubation, these plates were incubated between 60–72 h along with a blank and a laboratory control spike (LCS). This LCS was made from a pure culture of 

 (NCPF 3179/ATCC10231) to confirm fungi growth can occur under present conditions. The blank was plated to confirm that no cross-over contamination occurred throughout the process. Due to client confidentiality, the names of the clients will be hidden and will be given a four-digit number to associate the sample with. These numbers are as follows: 5205, 5124, 5251, 5308, and 5166.

Samples 5124 and 5251 CFU counts were found to be significantly different at the 95% confidence interval (Table I). Samples 5308, 5166, and 5202 were found to be nonsignificant even though the average CFU count was higher at the lower temperature (Figure 1). To determine whether a significant difference between samples existed, a two-sample t-test with df=48 was conducted, with alpha set at the 0.05 level. In all four samples that experienced growth, all difference of means were negative again showing that high counts were observed at the lower temperature recommendation. Sample 5308 showed no growth at the dilution. From this data we can deduce that the lower incubation temperature is likely more ideal for yeast and molds commonly found on dried cannabis.

When testing dried cannabis flower for total yeast and mold, utilizing 3M yeast and mold rapid petrifilm, 25 °C offers statistically higher counts in comparison to 28 °C. Sample 5251 and 5124 showed significant difference between temperatures at the 95% confidence interval (Table I). All samples that showed growth yielded higher counts at the lower temperature (Figure 1). Along with higher counts, colonies are easier to enumerate because of the increase in morphological diversity observed on the plates. These morphological differences mainly pertain to the shade of color of the colonies found on the plate. The remaining sample that showed no growth was determined to be plated at a dilution too high to see the true CFU count. Secondly, samples were in refrigerated storage prior to research and development for more than 30 days due to state guidelines of sample disposal. This could have also played into the lack of growth for sample 5308. 

This preliminary research has areas of possible expansion in many ways. For one a selective agar, such as Dichloran Rose-Bengal Chloramphenicol agar (DRBC), could be used to compare the counts of different temperatures to a single count. Secondly, temperatures outside of manufacturers specifications could be implemented because of the complex matrix of the dried cannabis flower. This matrix specifically on 3M rapid yeast and mold petrifilm can lead to misidentification to the untrained eye due to the cannabis matrix causing a blue color to appear, but these small matrix interferences do not match the morphology of other yeast and mold (6). With continuing innovations in temperature, media, and measurement of contaminate we can hope to minimize the effect of matrix interference in the sample measurement. Ultimately, the goal is to not only supply cultivations and manufacturers of cannabis with honest and accurate results so that they can better remediate their product, but to also not allow harmful product to reach the marketplace. In late 2020, the AOAC developed a task force to evaluate yeast and mold determination in cannabis, but as microbiologists in this new industry we are responsible for validating these methods to protect the consumers of the marketplace (7). Based on these preliminary results we have identified 25 °C to achieve the most reliable growth for rapid yeast and mold petri. I urge governing bodies of cannabis and method developers to require that laboratories that utilize the 3M rapid yeast and mold petrifilm to incubate at 25 °C or to research how these temperatures may affect growth in different states. This will maximize the potential for CFU growth for samples and limit the possible health concern of allowing contaminated product to enter the marketplace.

I would personally like to thank RSR Analytical Laboratories for help fund this research, Tao Zhong for aid in sample preparation, and Wyatt J. Tarter for statistical analysis.

R. Upton, L. Craker, M. ElSohly, A. Romm, E. Russo, and M. Sexton, 

Standards of Identity, Analysis, and Quality Control

 (American Herbal Pharmacopoeia, Scotts Valley, California, 2014).

Archives of Phytopathology and Plant Protection

Z.K. Punja, D. Collyer, C. Scott, S. Lung, J. Holmes, and D. Sutton, 

3M Petrifilm Rapid Yeast and Mold Count Plate Interpretation Guide. 3M, 

https://multimedia.3m.com/mws/media/902046O/3m-petrifilm-rapid-ym-countplate-interpretation-guide.pdf. Accessed 2 February 2021

P. Bird, J. Flannery, E. Crowley, J. Agin, D. Goins, and R. Jechorek, 

N. Stolze, C. Bader, C. Henning, J. Mastin, A.E. Holmes, and A.L. Sutlief, 

 105681 doi:10.1016/j.mimet.2019.105681 (2019).

In this study, dried cannabis flower found to have yeast and mold during compliance screening were randomly chosen to be plated at two different incubation temperatures to compare total amount of yeast and mold growth.

April 2021

Why Dispensaries Need In-House Testing

The Evolution of Ethanol Extraction Methods in Cannabis

Looking with Light: Breaking Down Liquid Chromatography Method Development

Is This Cannabis or Hemp—NIST Helps Cannabis Laboratories Achieve High-Quality Measurements

Temperature Comparison of 3M Rapid Yeast and Mold Petrifilm Utilizing Manufacturer’s Suggested Temperatures on Dried Cannabis Flower (Cannabis spp.)

The Environmental Implications of Energy Consumption in Cannabis Cultivation

The Fields of Science and Technology Would Not Exist Without the Use of Buffers—Why Aren’t They Used with Cannabis Extractions?

Gaining Deep Knowledge About Cannabis Cultivation: How and Why