Quantifying residual solvents in active pharmaceutical ingredients

Detecting residual solvents in pharmaceutical processes is crucial for pharmacopeial purposes. First, it ensures product quality by confirming that harmful or unwanted solvents are removed or below risk levels, preventing potential health hazards. Second, regulatory compliance often mandates limits on residual solvents to ensure product safety. Additionally, understanding and controlling solvent levels contribute to process efficiency and cost-effectiveness in industrial settings.

Nuclear Magnetic Resonance (NMR) spectroscopy is a powerful analytical tool. It provides detailed information about the molecular structure, dynamics, and interactions of organic compounds without the need for extensive sample preparation. NMR is non-destructive, allowing for repeated measurements on the same sample. It’s versatile, providing insights into chemical environments, connectivity, and three-dimensional structures. Additionally, NMR is quantitative, offering information about the concentration of different nuclei in a sample. One of its limitations concerns the resolution of all signals of choice especially for 1H NMR. Even with an outstanding field strength, complex matrices within process samples or broad product signals (e.g., polymers with short T2 times) cover huge ranges of chemical shifts and prevent the detection of such residual solvent traces.

In collaboration with Aspen Oss B.V., we developed an NMR protocol on our Spinsolve benchtop NMR systems to lift the curtain of unwanted matrix signals covering the residual solvent signals of interest consequently being able to quantify the latter. Within this project, Aspen Oss B.V. was interested to identify and quantify isopropyl alcohol and diethyl ether contents within an API matrix. Figure 1 compares the standard 1D proton spectrum (blue) with the outcome obtained when the relaxation filter is included before the acquisition of the NMR signal (black).  After applying the relaxation filter, the quartet of the diethyl ether at 3.25 ppm, the septet at around 3.80 ppm of the isopropyl alcohol, and the methyl signals of both compounds at 1 ppm are clearly visible and fully baseline-separated, which is the requirement for precise quantification.