Development of ultrahigh‐performance liquid chromatography/mass spectrometry and ultrahigh‐performance supercritical fluid chromatography/mass spectrometry assays to determine the concentration of Bitrex™ and sodium saccharin in homemade facemask fit testing solutions

Rationale Fast and easily transferable chromatography/mass spectrometry assays were required to detect and quantify the amount of Bitrex™ and sodium saccharin in homemade facemask fit testing solutions. Methods Bitrex™ solutions were analysed using reversed‐phase ultrahigh‐performance liquid chromatography coupled with positive ion electrospray ionisation mass spectrometry (UHPLC/ESI‐MS). Separation was achieved using a mobile phase gradient with an Acquity BEH C18‐packed column. Sodium saccharin solutions were analysed using ultrahigh‐performance supercritical fluid chromatography coupled with negative ion electrospray ionisation (UHPSFC/ESI‐MS). Separation was achieved using isocratic elution with an Acquity UPC2 Torus Diol packed column and a methanol (25 mM ammonium acetate) co‐solvent. Results The calibration curves obtained using the ratio of the active compound to an internal standard generated linear regression values (R2) >0.99. Samples analysed prior to and after an autoclave sterilisation process and bottling gave repeatable measurements within 10% of the expected concentration. Conclusions The two assays afford a fast robust and quantitative analytical method for the detection of the active components used to test the efficacy of the homemade facemask testing solutions.


| INTRODUCTION
In extreme environments where airborne pollutants may be present, e.g. toxic gases, fumes, vapours, and other harmful pollutants, it is essential that personal protective equipment is worn and fitted correctly. In hospitals and care homes the use of respirators and face protection ensures that contaminants do not cause potential harm to the wearer. Filtering facepiece (FFP)1 masks work to protect individuals from spreading their own saliva and bodily fluids into the wider environment, while FFP2 masks are respirator face masks that filter out noxious and toxic chemicals/odours, vapours, and other harmful pollutants from the air. 1 All protective equipment must be fitted correctly and adjusted individually to the wearer's face. A Face Fit Test is normally undertaken prior to the respirator or facemask being worn and the fit needs to be checked regularly. There are two tests as defined by the Occupation Health and Standard Association (OSHA) in 29 CFR 1910.134. 2 The Qualitative Fit Test (QLFT) is a simple pass/fail test that measures the user's response to a test solution. The Quantitative Fit Test (QNFT) is an assessment of the adequacy of fit of the respirator by measuring the amount of leakage into the respirator. 3 Two different solutions are commonly used to assess disposable and reusable half-masks: denatonium benzoate, otherwise known by the brand name Bitrex™, 4 and sodium saccharin 5 (for structures, see Figure 1). The recent outbreak of COVID-19 in the UK has led to shortages of these commercial spray solutions, as demand has significantly increased in line with increased usage of masks and respirators. Chemists at the University of Southampton have produced homemade solutions 6  An analytical quality control (QC) method was required to speedily assess the quantity of taste test compound in each prepared batch solution. This was required immediately after batch preparation and following an autoclave sterilisation (>120 C) and subsequent bottling process. A minimum of 10% of the bottles were tested prior to distribution.
The initial aim of the analytical QC methodology was to develop fast, easily transferable, fit-for-purpose tests that could be undertaken across many analytical laboratories. A reversed-phase (RP) high-or ultrahigh-performance liquid chromatograph is one of the most common instruments available to a QC laboratory; hence, this separation technique coupled to mass spectrometry was initially attempted for the analysis of both solutions. The methods used were adapted from several previously published articles. [9][10][11][12][13][14] As Bitrex™ is a quaternary ammonium compound positive ion electrospray ionisation (ESI) is required for its analysis, while is the primary mobile phase used and is a supercritical fluid above its critical point of 31.1 C and 73.8 bar, where it has properties intermediate between those of a gas and a liquid. [15][16][17][18] For many reasons scCO 2 is the most commonly used mobile phase; its critical point is easily obtainable, and it is readily available, inexpensive, considered green, and relatively safe to use. 19 UHPSFC can be easily coupled to a mass spectrometer using an atmospheric pressure ionisation source with a flow splitter. A makeup solvent, such as methanol with formic acid, is delivered to the mass spectrometer via a splitter configuration to promote ionisation and ensure a stable spray.

| Stock solutions
The standard compounds (Bitrex™ and sodium saccharin) were prepared volumetrically at a concentration of 10 mg made up to 10 mL (methanol) to give stock solutions of 1 mg/mL.

| Internal standard preparation
The internal standard compounds (oxybutynin chloride and vanillic acid) were initially prepared at a concentration of 1 mg/mL in methanol and then diluted using volumetric dilution to the appropriate concentration (for structures, see Figure 2).

| Calibration preparation
Standard calibration solutions were prepared volumetrically for Bitrex™ containing nominally 20, 50, 100, 150 and 200 ng/mL with oxybutynin chloride as the internal standard at 100 ng/L. Standard calibration curves were prepared for sodium saccharin containing nominally 5, 10, 15, 20, 32.5 and 50 μg/mL with vanillic acid as the internal standard at 10 μg/mL. The concentration of the standards was selected to ensure that the calibration curves correspond to the linear ionisation response region of each instrument.

| Sample preparation
Each batch of Bitrex™ and sodium saccharin was prepared at two concentrations, one named as SENSE (to test the response of an individual) and one named as TEST (to test the facemask once it has been fitted to the individual). A volume of 1 mL of each batch solution was removed from the bulk and prepared for QC analysis. The Bitrex™ SENSE solution was prepared at a concentration of 135 mg/mL and this was serially diluted using methanol containing the oxybutynin chloride internal standard. The first step was a 1:1 dilution using the internal standard at 200 ng/mL followed by dilutions using the internal standard at 100 ng/mL to a concentration of 67.5 ng/mL. (×2000 dilution). The TEST solution concentration was 1.69 g/L and this solution was serially diluted using methanol containing the same internal standard to a concentration of 84.7 ng/mL (×20,000 dilution). The saccharin SENSE solution was prepared at a concentration of 8.3 mg/mL and was serially diluted using methanol containing the vanillic acid internal standard. The first step was a 1:1 dilution using the internal standard at 20 μg/mL followed by dilutions using the internal standard at 10 μg/mL to a concentration of 10.4 μg/mL (×800 dilution). The TEST solution concentration was 536 mg/mL and this solution was serially diluted using methanol containing the same internal standard to a concentration of 26.8 μg/mL (×20,000 dilution).
The column was held at 50 C in a column oven and 2.0 μL of each sample was injected. Solvent A, water + 0.2% FA, and solvent B, acetonitrile + 0.2% FA, were used for separation at a flow rate of 0.6 mL/min. A gradient elution was performed using the method reported in Table 1 and a 1-min isocratic pre-run was used for column equilibration.
Separations were performed for the sodium saccharin samples using an Acquity ultrahigh-performance convergence chromatograph (UPC 2 , Waters) with a Torus Diol packed column (1.7 μm particle size, 3 × 100 mm; Waters). The column was held at 40 C in a column oven and 2.0 μL of each sample was injected. scCO 2 with methanol (25 mM ammonium acetate) co-solvent was used for separation at a flow rate of 1.5 mL/min. The scCO 2 back pressure of the system was set to 150 bar. An isocratic elution was performed using co-solvent B at 40% for 1.5 min.

| Mass spectrometry
Positive ion ESI mass spectra were recorded using a triple quadrupole mass spectrometer (Xevo TQD) and an ESCi multi-mode ionisation source (both from Waters) with the following conditions: capillary

| Method development
In the first instance, commonly available RP-UHPLC/MS assays requiring minimal modification and customisation were used to facilitate the QC analysis for both compounds. Generic gradients were used, starting at 5% solvent B increasing to 100% solvent B over 5 min with a standard BEH C18 column. The first task was to identify a suitable internal standard for each assay. Oxybutynin chloride was selected for the Bitrex™ assay since this was readily available in the laboratory and is routinely used as a component of the system suitability test used for RP-UHPLC/MS (for structure, see Figure 2A). It has a similar retention time to Bitrex™ and its positive ion electrospray ionisation efficiency closely matched that of Bitrex™. To accelerate the analysis, the generic gradient was optimised to 20% solvent B increasing to 100% solvent B in 1.5 min with a 1-min isocratic pre-run ( Figure 3). The linear response for Bitrex™ using this assay was defined using solutions of standard Bitrex™ ranging from 1 ng/mL to 1000 ng/mL. SIM for the molecule was used to improve peak profile and ensure that the methods could be transferred to other analysers. Once the linear response region had been defined, calibration solutions were prepared (20-200 ng/mL) containing the internal standard at a concentration within the middle of the curve (100 ng/mL). A five-point calibration curve was constructed automatically with TargetLynx using the ratio of Bitrex™ to the internal standard, and gave R 2 values of >0.99 ( Figure 4).The same assay was attempted for the analysis of sodium saccharin using negative ion ESI; however, the peak shape for sodium saccharin was initially poor. This was somewhat improved by changing to water as the internal standard diluent to give a fit-for-purpose assay. The saccharin analysis was developed using an in-house UHPSFC/MS assay, 10% co-solvent (methanol 25 mM ammonium acetate) increasing to 40% co-solvent over 5 min using a Torus Diol column.
Vanillic acid was selected as the internal standard for this assay since this was readily available in the laboratory (for structure, see Figure 2B). Its negative ion ESI efficiency closely matched that of sodium saccharin and it gave a similar good chromatographic peak shape. To accelerate the analysis, the assay was optimised using an isocratic 40% co-solvent method for 1.5 min ( Figure 5), hence removing the need for a pre-run column equilibration method. The linear response for sodium saccharin using this assay was defined using solutions of standard sodium saccharin ranging from 1 μg/mL to 1,000 μg/mL using SIM for the deprotonated molecule, to improve the peak profile. Once the linear response region had been defined, calibration solutions were prepared (5-50 μg/mL) containing the internal standard at a concentration in the middle of the curve (20 μg/mL). A five-point calibration curve was constructed automatically with TargetLynx using the ratio of sodium saccharin to the internal standard, and gave R 2 values of >0.99.