Two Techniques for the Analysis of WS-3 With Potential Application to the Analysis of Other Cooling Agents *

The cooling agents were obtained from Vigon Intern. Inc. (East Stroudsburg, PA, USA). The HCOOH, methanol and nicotinamide (vitamin B3) were obtained from Sigma-Aldrich (St. Louis, MO, USA). Water 18.2 mΩ/cm was obtained from a Barnstead Nanopure unit (Thermo Scientific Rockford, IL, USA). For the filtration of extracts, 0.45 µm PVDF filters were used (Whatman Autovial, GE Healthcare, Little Chalfont, UK). Scintillation vials of 20 mL with screw top caps and GC vials of 2 mL with screw top caps and septa were also utilized.

The GC separation with MS detection was performed on a 7890B/5977A GC/MS instrument from Agilent (Wilmington, DE 19808, USA) equipped with a VF-1701ms column 30 m × 0.25 mm with 0.25 µm film from Agilent. Data interpretation was performed using MassHunter Qualitative Analysis B.07.00 and MassHunter Quantitative Analysis B.08.00. The extraction of samples was performed on a wrist action shaker (Burrell Co., Pittsburgh, PA, USA).

The HPLC separation with MS detection was performed on an Acquity I UPLC system that consisted of a binary pump, an autosampler with cooling capability, a QDa and a PDA detector from Waters (Milford, MA, USA). Data interpretation was performed using MassLynx version 4.1 program from Waters. The HPLC chromatographic separation was achieved on an Acquity UPLC^® BEH C18 column 1.7 µm particles, 2.1 × 50 mm dimensions from Waters.

Four solutions in methanol containing 125 µg/mL of WS-3, WS-23, Evercool 180, and Evercool 190, respectively, were prepared as stock solutions. Equal volumes from these solutions were mixed to obtain the Standard 1 that contained 25 µg/mL of each cooling agent. From this solution, 10 more standards were prepared, using repeated dilutions 50% previous standard 50% methanol, with the lowest standard containing 0.022 µg/mL. The standards were analyzed by both the GC/MS and the LC/MS procedures. For the GC/MS analysis only Standard 1 (25 µg/mL) to Standard 4 (1.56 µg/mL) were used for the calibration, and single-compounds solutions containing 125 µg/mL and 62.5 µg/mL coolant were also used. For the LC/MS analysis, only Standard 2 (12.5 µg/mL) to Standard 11 (0.024 µg/mL) were utilized. The GC/MS analysis could be used for more concentrated solutions of the analytes while the LC/MS procedure was more sensitive and could be used for diluted solutions. For the analysis, to 1 mL solution of standard was added 20 µL of a solution of an internal standard that consisted of 625 µg/mL nicotinamide in methanol to obtain in each standard a final concentration of 12.5 µg/mL I.S. (reported to 1 mL, not corrected to a final volume of 1.02 mL).

Only various samples containing WS-3 cooling agent were available for analysis. The analyzed samples were extracted with methanol, the amount of sample and that of methanol were adjusted to obtain a final concentration in the range of calibration standards. The methanol extracts were analyzed directly and no cleanup was applied. The same extraction procedure used for the GC/MS analysis was used for the LC/MS analysis but for the LC/MS analysis the first extract was further diluted to bring the concentration in the range of calibration. In addition, two tobacco samples were also extracted on a Dionex-ASE-350 (Thermo-Fisher, Waltham, MA, USA) for proving the extraction efficiency.

For the tobacco samples, 400 mg tobacco were extracted with 10 mL methanol for 30 min on a wrist-action shaker. The solutions were filtered through 0.45 µm PVDF filters. For the GC/MS analysis the filtered solutions were directly used for analysis. For the LC/MS analysis a further dilution was necessary and 100 µL filtered tobacco extract was placed in a 2 mL vial and 900 µL of methanol were added to obtain 1 mL final solution for analysis (1:10 dilution).

For the nicotine pouches analysis, 400 mg pouch material was extracted with 10 mL methanol for 30 min. The solution was filtered through a 0.45 µL PVDF filter and further diluted 1:5 with methanol for LC/MS analysis. For the GC/MS, the 1:5 dilution is not necessary.

For the analysis of WS-3 in smoke, the total particulate matter (TPM) from three cigarettes was collected on a 44-mm Cambridge pad. Two smoking conditions were utilized, ISO (International Standard Organization) (9, 10) and HCA (Health Canada Intensive) (11). The smoking was performed at Eurofins Lancaster Lab at R.J. Reynolds Tobacco Co. The Cambridge pads containing the TPM were placed in 20-mL scintillation vial and extracted with 10 mL methanol for 30 min on a wrist action shaker. The extracts were filtered through a 0.45-µm PVDF filter and further diluted 1:10 for LC/MS analysis.

For the analysis of flavor solutions, 30 mg flavor solution were dissolved in 10 mL methanol. The obtained solution was further diluted 50 times by using 0.2 mL flavor solution and 9.8 mL methanol.

The LC/MS analysis of the coolants with the separation performed on the Acquity UPLC® BEH C18 column in gradient conditions. Solution A for the gradient was 80% water, 20% methanol 0.1% HCOOH, and the solution was brought to pH 3.33 with a few drops of NH₄OH. Solution B was methanol. The flow rate of the mobile phase was 0.3 mL/min and the linear gradient conditions are indicated in Table 1.

The detection was performed in positive mode with the following ions monitored: m/z = 123.1 for nicotinamide (I.S.), m/z = 172.1 for WS-23, m/z = 212.2 for WS-3, m/z = 289.2 for Evercool 190 and m/z = 299.2 for Evercool 180. The QDa instrument had a cone voltage 15 V, gain = 1, temperature setting at 600 °C, capillary voltage 0.9 kV.

In the conditions for the LC described in Table 1 and for the QDa indicated above, a typical chromatogram for a standard containing 0.78 µg/mL of each analyte and 12.5 µg/mL internal standard is shown in Figure 2.

Figure 2.

From the four coolants, WS-23, WS-3 and Evercool 180 are amides and their elution time is not much affected by the pH of the mobile phase. However, the pyridine moiety from the Evercool 190 makes the retention time of this compound sensitive to the mobile phase pH. For a mobile phase at pH 2.9, Evercool 190 elutes very close to WS-23, and for pH = 3.8 it coelutes with WS-3. For this reason, the pH of the mobile phase must be precisely controlled.

Based on a specific retention time for each analyte, of the specific mass detected by the MS detector and also on the specific conditions of detection (settings of the MS instrument) it can be concluded that the method has a very good selectivity.

The quantitation of coolants based on the LC/MS analysis was also performed using calibrations analyte as a function of peak areas but the experimental data were better fit by dependencies of the form

Y=aX2+bX+c $$Y = a\,{X^2} + b\,X + c$$

where Y is the concentration of the cooling agent in mg/mL and X is the peak area of the specific ion of each analyte. The internal standard was used only to verify the reproducibility of the chromatographic process (was used as chromatographic standard) and was not applied for normalization of analyte peak areas. The use of normalized peak areas for quantitation did not lead to better precision of the quantitation. The parameters a, b, and c used for calibration for each cooling agent as well as the corresponding correlation coefficient R² are given in Table 2.

The calibrations were used to back-calculate the amount of cooling agent in the standards, and the limit of quantitation LOQ for each coolant was selected as equal with the lowest calibration standard that generated by back-calculation a result within 90–110%-range of the correct result for the average of two measurements with an RSD lower than 5–6%.

This procedure generated the value LOQ = 0.049 µg/mL for all four analytes. By considering that the limit of detection LOD is typically 1/3 LOQ, the technique can be considered as having LOD = 17 ng/mL.

In addition, the low RSD even for the lowest standard (lower than 5%) indicated that this procedure has a very good precision. For the analyzed samples, the RSD% (Tables 8–12) was also lower than 5% indicating a good precision for the whole procedure.

The extraction efficiency was evaluated only for WS-3 cooling agent. For the verification of extraction efficiency the tobacco samples of 1 g were spiked with a solution in propylene glycol of WS-3 containing 12.37% WS-3 by weight to obtain levels of WS-3 on tobacco of 0.2%, 0.1% and 0.05% WS-3. Each level was applied on one sample followed by triplicating the analysis. The results of the analysis of the spiked samples are indicated in Table 3.

The results from Table 3 indicate very good recovery of the WS-3. Further proof of efficiency of simple methanol extraction for the analysis of WS-3 was obtained by extracting 300 mg tobacco with 17.1 mL methanol in an ASE-350 accelerated solvent extractor, using a temperature of 100 °C with extraction for 10 min. The results for the level of WS-3 in two tobacco samples were within 97–103% range as for the simple methanol extraction.

The result for WS-3 is a good indication that the recovery of other cooling agents will be also good by the simple methanol extraction. However, the recovery for the other cooling agents was not tested.

The accuracy of the LC/MS procedure was proven by obtaining the correct results from the back-calculation of the levels of the analyzed cooling agents in the standard samples, from the excellent results obtained from the recovery data as indicated in Table 3, and also by obtaining the same values for WS-3 in the analyzed samples by the LC/MS procedure and by the GC/MS procedure as further shown in Tables 8 to 10.

The GC/MS analysis of the coolants with the separation performed on the VF-1701ms column was performed in conditions indicated in Table 4.

Using the conditions described in Table 4, a typical chromatogram for a standard mix containing 25 µg/mL of each cooling agent and the internal standard at 12.5 µg/mL is shown in Figure 3.

Figure 3.

For the quantitation of cooling agents using the GC/MS analysis, we used linear calibration curves of the type

Y=aX+b $$Y = a\,X + b$$

where Y is the concentration of the cooling agent in µg/mL and X is the peak area of the specific extracted ion of each analyte.

The internal standard was used only to verify the integrity of the chromatogram (I.S. was used as chromatographic standard) and was not used for normalization of other peak areas since this procedure did not lead to better results. The ion used to generate the extracted peak in the chromatogram, the parameters a and b used for calibration for each cooling agent as well as the corresponding correlation coefficient R² are given in Table 5.

The calibrations were used to back-calculate the amount of cooling agent in the standards, and the LOQ for each coolant was selected as equal with the lowest calibration standard that generated by back-calculation a result within 90–110% range of the correct result for the average of two measurements with an RSD lower than 5–6%. As a result, the LOQ values for the analyzed cooling agents are those indicated in Table 6.

Based on the calibration range and using the same criterium as for establishing LOQ in LC/MS method (back-calculation results within 90–110% range of the correct value for the average of two measurements with an RSD % lower than 5–6%) the value for LOQ for the GC/MS method were those indicated in Table 6.

A number of samples containing WS-3 were analyzed by the LC/MS procedure, and a few of them also by the GC/MS procedure. The selection of the samples was made to include a variety of matrices such as tobacco, material from nicotine pouches, cigarette smoke and several flavors. The samples description is given in Table 7. The experimental cigarettes were made at R.J. Reynolds Tobacco Co. The commercial cigarettes were purchased from the market.

The levels of WS-3 in tobacco from cigarettes is indicated in Table 8. Each measurement was performed in triplicates.

The levels of WS-3 in four nicotine pouches using the LC/MS analysis are indicated in Table 9. Each measurement was performed once and the RSD % is reported to different analyzed samples.

Sample (1) was also analyzed using the GC/MS procedure, also in triplicate. The levels of WS-3 in Sample (1) by GC/MS analysis are indicated in Table 10.

As indicated in Table 9, the LC/MS procedure for WS-3 analysis provides results with very good reproducibility and very close with the results for GC/MS analysis indicated in Table 10. This is a proof of the accuracy of the two methods.

The results for the levels of WS-3 in cigarette smoke are indicated in Table 11.

For the analysis of flavor solutions, the problem was the large difference between the sample concentration in WS-3 (about 12%) and the range of calibration for the LC/MS analysis (48.8 ng/mL to 12.5 mg/mL). For this reason, the initial samples were diluted 16,666 times. The results for the WS-3 in the analyzed flavor solutions are indicated in Table 12.

The relatively low RSD% for the results from Table 12 prove the applicability of the LC/MS procedure for the analysis of flavor samples.

The application of the LC/MS analysis of WS-3 demonstrated the utility of the procedure for a variety of samples, showing that the matrix (tobacco, nicotine pouches, smoke, flavor solutions) does not affect adversely the analysis.