What is Chemical Analysis?
Chemical analysis is the process of measuring the chemical constituents of a sample. Chemical analysis in the cannabis industry involves having your flower, oil, and edibles tested for cannabinoids, terpenes, and toxic substances, such as pesticides, solvents, and heavy metals.
Many members of the cannabis community may be familiar with chemical analysis when they have cannabis flower tested, and receive an amount of THCA as a percent. Results for THCA vary but in flower around 20% THCA is common. Terpenes can be reported in percent but are more often reported as mg/g. Pesticides, solvents, and metals are reported in ppm and ppb. Although these units can be confusing, they are all basically the measuring the same thing: How much of the chemical of interest is there as a ratio of the sample? Each unit has a slightly different way of describing that ratio. A unit of percent is worth 1/100th of the total sample. A mg/g is worth 1/1,000th of the total sample. A ppm is worth 1/1,000,000th of a sample, which is where the acronym comes from (part per million). We have summarized these units in a chart below:
These units reflect how much of the analyte is in the sample analyzed. If a sample of cannabis flower is 20% THCA, that means that 1,000 mg of ground flower have 200 mg of THCA, because 200 is 20% of 1,000. But if the same batch has 20 ppm myclobutanil in it, a toxic pesticide, then a different calculation is needed. Out of 1,000 mg of ground flower, 0.02 mg of myclobutanil is present.
Regardless of whether scientists are analyzing something present at high or very low concentrations, the method is the same. The chemical is isolated from the sample, and measured relative to the original sample. In the early days of analytical chemistry, chemists had to physically isolate a compound, often by complex techniques like distillation or precipitation, and weigh it to know what percent the compound comprised the original sample. With these older techniques, measuring below a few percent was impossible, requiring more modern and sensitive analytical techniques.
Before discussing more modern techniques, let’s review the concept of separation of compounds necessary for analyzing and quantifying chemical constituents in the sample. Consider a sample of cannabis hash oil that is contaminated with the pesticide myclobutanil. Almost all of the sample is composed of the molecule THCA, but some of it is composed of myclobutanil. To help illustrate this concept, we have magnified the molecules many times.
Image01: Diagram illustrating the need to separate compounds in a sample to quantify the relative
amounts of each compound in the original sample. Illustrated here is a cannabis hash oil sample
contaminated with the pesticide myclobutanil in the center with a scale indicating total sample mass.
On the left isolated THCA is shown being weighed on a scale, and on the right isolated myclobutanil is
being shown being weighed on a scale.
Image 1 shows that separating individual compounds facilitates the analysis and quantification of the isolated chemical constituents. Using the above example, a 7-gram oil sample has 6 grams of THCA in it, or about 85.71% THCA. Alternatively, the myclobutanil represents 1 gram out of the original 7 grams, or about 14.29% myclobutanil. With the ability to separate compounds, quantifying and reporting chemical content is simpler to understand and perform.
How is the Analysis Performed?
Modern analytical chemistry performs separations using chromatography, a process of separating compounds by flowing them through a very tight space called a column. The flowing material is called the mobile phase, and may be liquid, gas, or a supercritical fluid. The column, called the stationary phase, is what interferes with the passage of compounds allowing them to separate.
Gas chromatography (GC), often used in the cannabis industry for terpene, residual solvent, and pesticide analysis, uses a gas mobile phase like helium or hydrogen, and passes compounds through a column about 250 μm wide about 30 meters long. Liquid chromatography (LC), often used in the cannabis industry for cannabinoid, pesticide, and mycotoxin analysis, uses a liquid mobile phase like water, methanol and acetonitrile, and passes through a column about 5 mm wide and about 100 mm long.
Image 2 illustrates the concept behind chromatography by showing four compounds mixed together in a solution. A chromatographic run is a cycle of injecting a mixture at the upstream point of the column, and waiting until all compounds have exited the column.
Image02: A diagram illustrating the process of chromatographic separation of four compounds, yellow,
blue, green and red. At the top we see the initial sample, a mixture of all compounds. Below we see the
compounds separating at the right of the column in the early-run separation. As the compounds travel
through the column, they separate to larger degrees until in the late-run separation they elute from the
column completely separated from each other.
With an ability to separate, detect, and measure compounds, analyzing complex samples and generating reports is possible. Yet the quality of the data reported can vary considerably. To assure high quality and accurate results, one must consider factors that may cause errors in the reported results.
What Can Cause Inaccurate Results?
Errors in reported results can be caused by a number of factors. The two most common issues in cannabis analysis are related to the nature of the extraction procedure: extraction efficiency and matrix interferences.
The sample is typically a complex solid containing various chemical constituents called a matrix. Because the matrix cannot be directly injected into a chromatography column, the compound of interest must be extracted from the matrix in a solvent that can be injected. For the measurement made by the detector to accurately reflect the original sample’s chemical content, the extraction solvent must contain close to 100% of the compound of interest dissolved in it.
The need for close to full extraction efficiency is why cannabis labs analyzing flower use tissue homogenizers, which are medical grade grinders, and ultrasonic baths to fully disrupt the plant material and fully dissolve all compounds of interest. More complex matrices, like certain types of edibles containing uncommon oils and butters, may have issues dissolving in the standard extraction solvents used to dissolve cannabinoids thereby requiring special solvents.
Matrix interferences are errors in quantification of a compound caused by interfering matrix compounds present in the extraction solvent. When an extraction is performed on flower for cannabinoids or pesticides, the extract contains the compounds of interest but also includes abundant, diverse biochemical compounds present in the matrix. These matrix compounds can cause interferences with the detection of compounds of interest by overlapping with the peaks in the chromatogram.
Image03: An illustration of the relationship between chemical concentration and chromatogram. In
each diagram, the concentration of molecules is illustrated by blue dots below the x-axis and by a red
function tracing the chromatogram above the x-axis. The height of the peak in the chromatogram is
proportional to the concentration of the compounds passing through the detector. As the group of
molecules come closer together, the peak becomes taller and sharper.
In chromatography, the exiting process at the downstream end of the column is called elution. In image 2, the compounds enter the column, still nearly mixed, but as they travel through the resistance of the stationary phase, the individual compounds separate. At the end of the column, they elute one at a time, where they can be measured and recorded by a calibrated detector. This method enables analytical chemists to separate compounds of interest and determine the concentration to generate the data found on lab reports.
To better understand how the detector reads the compounds as they elute from the column, we need to better understand the chromatogram. A chromatogram is a visual representation of the concentration of analytes that illustrates clusters of compounds as peaks (Image 3). A variety of detectors can produce these chromatograms: ultraviolet-visible light detector (PDA or DAD), flame-ionization detectors (FID), or mass spectrometers (MS). These detectors record the signal generated by a compound with respect to the time it is retained on the column. The resultant peaks are plotted and the integrated peak area is proportional to the total amount of compound detected.
Image 4: Illustration of the same chromatographic method separating the same three compounds in pure solvent (above) vs. in extracted solvent (below). The peaks corresponding to compounds A, B, and C are indicated.
Image 4 reveals that even though a method may be able to separate compounds from each other, it may not be able to accurately detect those same compounds in an extracted solvent. The effect of matrix interferences renders the detection of compound C impossible, and interferes with the accurate measurement and quantification of compound B. This would cause a lab report to erroneously indicate that no compound C was detected even though it was present in the sample, a phenomenon called a false negative. Additionally, the lab would report a much lower amount of compound B that was present in the sample. Both the effects on compound B and C are considered matrix interference effects. Compound A, because it is not affected by the matrix interferences, would be reported accurately in this example.
Another example of a lab reporting error is a false positive triggered by a matrix interference compounds. A false positive is when a laboratory reports a compound that was not present in the sample. This can happen when a matrix interference compound elutes from the separation column at a similar time as the compound of interest and reacts in a similar way with the detector.
What Quality Checks Can Be Used to Avoid Reporting Errors?
There are several quality control checks that a laboratory can perform to assure a lack of interferences. To avoid contamination, a laboratory can run blank samples to ensure the instrument is not detecting compounds of interest when none are introduced. Additionally, a laboratory should measure matrix blank samples to confirm that the matrix is not causing false positive detection. A matrix blank sample is a matrix that does not contain the compound of interest. A good example of a matrix blank for a cannabinoid analysis is hops, because it is genetically closely related to cannabis but has no THC or other cannabinoids. The blank QC samples help to disprove that the lab is reporting false positives.
Aside from running blanks, it is also important to run spikes. A spike is when a sample is prepared as a blank, and the laboratory intentionally adds in, or “spikes in” a known amount of analyte. This exercise of spiking a sample with a known concentration of analyte and measuring the result is called a spike-recovery study. Spike-recovery samples validate extraction efficiency, showing that the extraction procedure is recovering close to 100% of the spiked-in compounds. Additionally, spike-recovery samples show that the matrix interference effects are not obscuring signal from compounds of interest and interfering with quantification.
Because a known amount of analyte was spiked into the sample, data can be evaluated regarding the lab’s performance. By measuring the amount detected by the analysis and comparing it to the amount spiked into the sample, the lab’s accuracy can be converted into a percent often called a percent recovery. A percent recovery of 90% would mean the lab’s results are about 10% lower than the true amount in their sample, and 110% would mean the lab’s results are about 10% higher. Close to 100% means the amount reported by the lab is very close the amount in the sample.
In addition to spike-recovery samples allowing for the measurement of accuracy in the form of percent recovery numbers, the precision of the lab can be measured in relative percent difference (RPD). Often analytical QC programs have requirements that two spike-recovery samples must be run side-by-side and the RPD between them is measured and must be below a certain value. This number provides the variation, as a percent, when two samples possessing equivalent concentrations of analyte are present. The smaller this percent, the more reproducible and therefore precise the reported laboratory data.
To read more on analytical laboratory quality control procedures, please look to the references for further information. Many chemical quality control bodies mandate spike recovery and matrix of method blank samples in the QAQC program of an analytical lab, including the ICH, the FDA, the EPA, the California Bureau of Cannabis Control (BCC), and several academic publications[3, 5].
How Do I Measure the Quality of a Laboratory’s Reported Data?
The best way to use laboratory QC data is to measure a lab’s performance and therefore the trustworthiness of the data they are reporting to their clients. Below we have outlined the things to look for and request for each specific concern that a client may have with a laboratory. This data that will be mentioned below should be found in a laboratory’s quarterly QC reports, which show the results of all QC samples performed daily by the laboratory. Looking through this data will provide a sense of both how accurate and how precise the laboratory is at analyzing each compound of interest on their report. Remember, a laboratory may be reporting accurate numbers for nearly every compound on a report but may still be reporting erroneous results on a single compound. It is for this reason it is important to request all QC data from your testing laboratories, especially if you suspect there is an error in the data they are reporting.
If you suspect your lab is reporting false positives…
You’ll want to check their QC data for blanks and method or matrix blanks. If you are seeing that the levels detected by the laboratory in their own blanks are not stable well below the amounts being reported, then it is likely laboratory contamination may be causing false positives.
You’ll want to pay particular attention to their data on matrix blanks, what they used as a matrix blank, and how closely they matched it to your samples.
When comparing laboratories, see whose baseline blank value is lowest, and see whose baseline blank value is the most stable over time, as these are both signs of high quality analysis.
If you suspect your lab is reporting false negatives…
You’ll want to check their spike recovery data to be sure that your laboratory is able to fully recover compounds of interest from the matrix in question and are able to properly detect those compounds when extracted from matrix with interfering compounds.
You’ll want to pay particular attention to matrix-based spike-recoveries down near the detection limit, quantification limit, or reporting limit (sometimes called LOD, LOQ, RL, or MRL). This is important if you suspect false negatives because these spikes represent the lowest level that can be seen. Strong accuracy and precision at that level would indicate high quality analysis.
When comparing laboratories, see who has the percent recoveries closet to 100% and the lowest RPD values, as these are signs of accuracy and precision respectively.
If your lab results are not consistent and/or illogical…
You’ll want to check their spike recovery data check to gauge their accuracy. If this number is not near 100% + 10% this is an indicator of a poor quality analysis.
Also check their RPD data to make sure the reported results are reproducible. If this number is not near 0% + 10% this is an indicator of a poor quality analysis.
Other indicators of lab data quality
Compare laboratory instrument reporting limit (IRL) and see who can report and detect to the lowest level, indicating the highest sensitivity. This is generally an indicator of higher laboratory quality.
Use common sense practices like summing the total percentages to be sure nothing is over 100%. Although values slightly above 100% can occur due to rounding errors, there is no scientifically justifiable reason to have a report on a sample add up to more than 102%.
Guideline, ICH Harmonised Tripartite. "Validation of analytical procedures: text and methodology Q2 (R1)." International Conference on Harmonization, Geneva, Switzerland. 2005.
US Food and Drug Administration. “Analytical Procedures and Methods Validation for Drugs and Biologics.” July 2015.
Lesnik, Barry “Guidance for Methods Development and Methods Validation for the RCRA Program”. US Environmental Protection Agency. April 6, 1992
US Environmental Protection Agency. “SW-846 Update V: Chapter 1” July 2014
US Food and Drug Administration “Ora Laboratory Procedure: Assuring the Quality of Test Results” October 1st, 2003.
Cal. Code Proposed Regs. Tit. 16, Division 42, § 5730