Limits of Quantification and Detection

Every COA indicates a limit of quantification and/or a limit of detection, but what are they and why do they matter? These limits are important because they provide an understanding of what an analytical lab is actually able to measure. This is especially important when purchasing products marketed as “THC-free”.

Imagine you are taking a photo of an apple tree and a friend asks, “How many apples are on that tree?” How do you answer? And how could you provide a more accurate answer to your cider-loving friend? If you took the photo on your iPhone from a distance so you could capture the sunset behind the tree, an accurate answer is going to be nearly impossible. Perhaps you could zoom in but at some point distinguishing apples from leaves in a pixelated mess will only get you so far. Your ability to distinguish leaves from apples is the limit of detection. Maybe with this photo you could make a guess at say twenty apples. 

Now imagine instead of your iPhone, you set up a tripod and used a fancy zoom lens to take an immaculately detailed photo of just the top of the tree. You would be able to count exactly how many apples are visible to you from the photo, let’s say 42. How many apples you can individually count from this vantage point is the limit of quantification. The second photograph gives you a far more accurate view of how many apples there really are, but short of cutting the tree down and harvesting every apple - they are both just guesses limited by your view. 

This is how analytical chemistry and the quantification of compounds like cannabinoids work. The answer very much depends on the equipment you use (an iPhone vs a telephoto lens) and how accurate you need the answer to be (a ballpark using a photo vs dead on accuracy by physically harvesting and counting every actual apple). When thinking about “THC-free” products in the hemp industry, it is useful to think along similar lines. How accurate of a result do you need? How detailed (or zoomed in) do you need to go?

DEFINITIONS

Limit of Detection (LOD)

• Scientific: the smallest amount of analyte that can be distinguished from zero
• Translated To Hemp: the lowest amount of a cannabinoid that can be observed with a lab’s instruments

Limit of Quantification (LOQ)

• Scientific: the lowest amount an analytical method can measure with precision
• Translated To Hemp: the lowest amount of a given cannabinoid that a lab can accurately measure 

Where To Find Them On A COA

How This Applies To THC-Free

Let’s take a look at an example using hemp extracts and how THC is reported on a certificate of analysis (COA). Many times a COA will state 0.0% THC. In these cases it is important to reference the reported LOD or LOQ to truly understand how much THC could exist.

If the reported LOQ is 0.2%, you know that the method employed can only measure THC with precision at or above a level of 0.2%. This means the sample could contain 0.19% THC but it would show as 0.0% on the COA because the level of THC is below the LOQ. This also means if the exact same sample is tested with a different method capable of quantifying THC at a level of 0.1%, the COA would show a value for THC at 0.19%. In this case, the lab with an LOQ of 0.2% reports 0.0% THC, while the lab with an LOQ of 0.1% reports 0.19%. Is this a “T-free” product? 

In another scenario, let’s say a sample containing 0.07% THC is submitted to both of the previously mentioned labs. Now both COAs would say 0.0% THC. Case closed right? Nope. Now a third lab with an LOQ of 0.05% is used. This third lab reports THC at a level of 0.07%. So once again - is this a “T-free” product? These examples may seem pedantic or far fetched, but at Treehouse we see these very issues every day and it is one of our most commonly addressed questions.

When viewing product COAs it is imperative to evaluate the reported LOQs in order to better understand THC levels in products. And remember, any hemp supplier can only confidently say there is less THC than the lab’s reported LOQ. 

With all of this said, the question remains does a T-Free product even exist? Treehouse would say - yes. When thinking about THC tests at these extremely low levels, a consumer can think of the results as essentially 0%. For example, if a 30 mL tincture is formulated with 1000 mg of Elemental Extract: E3 with undetectable THC at an LOQ of 0.1% then the THC content in the tincture will be 0.003% at the most. This means each 1 mL dropper from the tincture will contain 0.03 mg of THC or 30 micrograms. To think of this a different way, this is the equivalent of one grain of rice in 30,000 grains of rice. Again, for all intents and purposes this is 0%. 

Though it is possible to create products with even less THC than this, the real question becomes at what percentage can we agree it is essentially 0%? Is it 0.004% or 0.000004%? 

Diving Deeper: Accumulation of Error

Variance in reported values is another fundamental component of any chemical analysis. This is because small imperfections in measurements (errors) accumulate throughout the analytical process. No matter how careful or competent the operator is, there is inherent error in every measurement. These errors can never be fully eliminated, and the accumulation of said errors results in statistical fluctuations of the measured value around the true number. (Think back to your stats class from high school. Remember the “normal distribution” or “bell curve” that was constantly referenced? Yep, the same thing applies here!) 

When measuring THC levels in an oil, for example, the operator must weigh out an accurate amount of the sample. The scale and human operator, being close to perfect, but not quite, contribute small errors to the measurement. Next the sample is diluted in solvent (say 10 mL), which is typically measured with a pipette. Small inaccuracies and inconsistencies from the pipette contribute more error. A smaller amount of this diluted sample (say 0.1 mL) is then taken up with a syringe (which contributes more error) and further diluted with even more solvent (again, a source of error derived from measurement). This sample is then injected on the instrument (and now error from the injector joins in). Finally the results from the instrument are measured against known standards (which also come with a slew of built in error from their separate preparation). Each of these sources of error may be so small that, alone, they don’t contribute to a relevant amount of variance in the end result.  But each individual error propagates down the entire analytical process, and taken together the accumulated error produces perceptible variations in the end results. This explains why the same homogeneous batch of CBD isolate may test at 100% purity one day and at 98% the next. This also explains why results over 100% are possible. Accumulation of error can also result in a measurement at 102%. This matters when it comes to THC because the amount of THC in your sample could flirt with the limit. That is one of many reasons why a lab could report detectable THC while another lab could report undetectable THC in the same sample. 

When digesting all of this, try and consider just how impressive a feat this all is. Analytical chemistry gives us the ability to determine the purity or content of dozens of different molecules, with a mixture containing trillions of molecules, and thousands of different compounds (all of which are far too small to be viewed by even the most sensitive microscopes!) So a variance of just a few percentage points is pretty darn astounding all things considered!