Posted on

How can a substance have a purity over 100 percent??

At Supplements for Work, we strive to bring our customers the highest purity compounds possible. This means that all our compounds are tested to ensure that they behave in the way we would expect the compound to behave and are pure. We employ a third party chemical testing organization to ensure that the data we receive is accurate and truthful. Below is an example of one of those reports for a batch of Tianeptine Sodium.

A COA that we received from a third-party testing company.

Notice anything odd about the results? According to this analysis, our Tianeptine Sodium is 101.7% pure, which, while flattering, is physically impossible. I will attempt to explain how this result came to be and what our customers should take away from a Certificate Of Analysis (COA) like this.

To understand how it is possible to have an analytical purity above 100%, we first must understand how the purity is determined. In this case, High-Performance Liquid Chromatography (HPLC) is used. HPLC is a technique used to separate, identify, and quantify all the components of any given mixture. Most modern HPLC procedures use an all-in-one, self-contained machine.

Courtesy of
A diagram showing all the individual parts of an HPLC machine.

The basic set up of an HPLC is as follows: There are two reservoirs of solvents, one polar and one non-polar. These are mixed in different amounts over the 30-minute runtime. For example, a procedure might start out being 100% solvent A and slowly transition to being 100% solvent B after 25 minutes. The middle contains every variation in between those two extremes. This creates a gradient that pulls different compounds within a sample at different rates depending on how they interact with each solvent. These flow at very high pressure to a specialized analytical column designed to retain compounds within the column, creating a greater separation between the multiple compounds within a given sample.  On the other end of the column is a detector that scans the liquid that leaves the column. The detector is connected to a computer that compiles the data and outputs a graph like the one below.

An example of what an HPLC output looks like. Note that this HPLC protocol could be further improved to provide greater separation between the two tallest peaks.

The Y-axis represents the intensity of the signal the detector picked up (corresponding to concentration) while the X-axis shows the time that the compound left the column (corresponding to polarity).

To use this data to find out the purity of a substance you need to have a sample standard. This is a virtually pure sample of the compound that can be used to judge other samples against. Making something this pure is incredibly difficult and requires time-intensive work, so standards are notoriously expensive. For example, 5mg of Tianeptine Sodium Salt standard from Millipore Sigma is $254, and that is still only 98% pure.

The sample standard is dissolved at a known concentration (1 mg/ml) and run using the method that the experimental sample will be run at. The experimental sample is then dissolved at the same concentration (1 mg/ml) and run. An example of how this might look is below.

An example of a comparison between a pure reference standard and an impure compound.

By calculating the area under each peak and dividing the experimental peak by the standard peak we can find the purity of our compound. So how then can we get values greater than 100%? When dealing with weight values as low as 1 mg the accuracy of our equipment begins to break down, so it is possible when weighing out the 1 mg there was error intrinsic to the sample. For example, if 0.98 mg was weighed out instead of 1 mg/ml and our experimental sample was overweighed to 1.01 mg the sample would appear to be 103% pure. To combat this, liquid dilutions are sometimes used to spread the error over much more solution resulting in a lower error in the end results. However, this can be a problem when the material is prohibitively expensive as is the case with Tianeptine standards.

Even if the initial solution is a perfect 1.0000 mg/ml the equipment that samples the solution is not immune to malfunctioning and may take 19.99 μL instead to the programmed 20 μL, which can create a similar error as detailed above. This type of equipment malfunction can be caused by static electricity, miscalibration, or power surges.

The discrepancy can also be explained by an old standard solution of Tianeptine. Standard solutions are usually made very shortly in advance of the testing and are stored at -20 C or lower. If the solution is made too far in advance, or not stored properly in the run-up to the tests, then small amounts of compound degradation can occur. Then when this is compared to the other samples, it will make it appear to be a higher purity. This form of error can appear even with no fault of the chemist, as some compounds are just inherently unstable in solution and will begin to break apart almost immediately until an equilibrium of broken and unbroken molecules is reached.

Analytical chemistry is a complex field that requires immense precision on a level that most people are not used to thinking about. Experimental error is a part of the everyday reality of dealing with highly specialized and finicky equipment such as milligram balances, HPLC samplers, and HPLC columns. For the most part analytical chemistry succeeds in providing highly accurate information about compounds identity and purity, however, this is an uphill battle against the ever-present invasion of error into the scientific process. All we can do is try to minimize this error to the greatest extent possible and live with results knowing they are as close to the truth as we are able to get at this point.