Determination of Nicotine Protonation State in E-Liquids by Low-Resolution Benchtop NMR Spectroscopy

Abstract

Over several years, e-liquids with “nicotine salts” have gained considerable popularity. These e-liquids have a low pH, at which nicotine occurs mostly in its monoprotonated form. Manufacturers usually accomplish this by the addition of an organic acid, such as levulinic acid, benzoic acid, or lactic acid. Nicotine in its protonated form can be more easily inhaled, enhancing the addictiveness and attractiveness of products. Several techniques have been described for measuring the protonation state of nicotine in e-liquids. However, nuclear magnetic resonance (NMR) spectroscopy is particularly suited for this purpose because it can be performed on unaltered e-liquids. In this article, we demonstrate the suitability of a benchtop NMR (60 MHz) instrument for determining the protonation state of nicotine in e-liquids. The method is subsequently applied to measure the protonation state of 33 commercially available e-liquids and to investigate whether the vaping process alters the protonation state of nicotine. For this purpose, the protonation state in the condensed aerosol obtained by automated vaping of different e-liquids was compared with that of the original e-liquids. Two distinct populations were observed in the protonation state of nicotine in commercial e-liquids: free-base (fraction of free-base nicotine α_fb > 0.80) and protonated (α_fb < 0.40). For 30 e-liquids out of 33, the information on the packaging regarding the presence of nicotine salt was in agreement with the observed protonation state. Three e-liquids contained nicotine salt, even though this was not stated on the packaging. Measuring the protonation state of nicotine before and after (machine) vaping revealed that the protonation state of e-liquids is not affected by vaping. In conclusion, it is possible to determine the nicotine protonation state with the described method. Two clusters can be distinguished in the protonation state of commercial e-liquids, and the protonation state of nicotine remains unchanged after vaping.

1. Introduction

A relatively recent development in e-cigarette product design has been the introduction of so-called “nicotine salts”: nicotine in a protonated state. Over the last couple of years, the prevalence of use of nicotine salt-containing e-liquids has increased.² When inhaled, the “throat hit” of nicotine salts is decreased when compared with free-base nicotine (Nic), at least at high concentrations.³⁻³⁵⁴ Throat hit is a term colloquially used for a slightly irritating sensation in the throat caused by components of the aerosolized e-liquid, in particular nicotine, although other components may also contribute. While most users find a mild throat hit pleasurable, a strong throat hit is experienced as unpleasant, especially by novel users of nicotine products.⁴ Nicotine salts can therefore increase the addictiveness of e-cigarettes because they allow users to inhale higher concentrations of nicotine while maintaining a tolerable throat hit. In particular, for nonsmokers, it may make e-cigarettes with high concentrations of nicotine more palatable. Notably, the European Tobacco Products Directive (2014/40/EU) prohibits the use of additives in tobacco products that facilitate inhalation or nicotine uptake (article 20.3).⁵ These effects may therefore be of interest to regulators. Thus, it would be useful to develop an analytical method for determining the protonation state of nicotine in e-liquids.

Nicotine is a dibasic compound (pK_a’s of NicH₂²⁺ and NicH⁺ are 3.10 and 8.01, respectively, in water at 25 °C).⁶ The different forms are depicted in Figure 1.

Figure 1.

Free-base and protonated forms of nicotine.

In the pH range of e-liquids reported in the literature (>4.78⁷⁻⁸), the concentration of NicH₂²⁺ in e-liquids is negligible. The fraction of free-base nicotine (α_fb) can therefore be calculated according to eq 1.

1

Several different approaches have already been described for measuring the protonation state of nicotine in e-liquids, based on HS-SPME-GCMS,⁹ proton nuclear magnetic resonance (¹H NMR),⁹⁻¹² liquid–liquid extraction,^13,14 and pH measurements.^8,14−16 To allow determination of the protonation state of nicotine as it exists in e-liquid, it is essential to apply an analytical technique that does not require modifying the e-liquid (e.g., dilution or extraction) because any change in composition may have an effect on the protonation state.

Duell et al. have published an NMR-based method to determine the protonation state of nicotine in e-liquids,¹⁰ which satisfies that criterion. However, the important disadvantages of a high-resolution NMR instrument, such as that used by Duell et al., are its large dimensions and cost, which is likely to be prohibitive for many laboratories. In contrast, low-resolution NMR instruments have become relatively affordable and compact (benchtop). It remained unclear whether the methodology can be adapted to work on a low-resolution benchtop NMR instrument.

In this research, we show that a low-resolution benchtop NMR instrument is adequate for measuring the protonation state of nicotine in e-liquids. The results are compared with those from (simpler) pH measurements,^8,14−1550 to verify whether the increased cost and complexity of NMR measurements is warranted. We subsequently determined the α_fb values of 33 commercially available e-liquids. Furthermore, we show that vaping does not affect nicotine protonation: α_fb values found are the same in aerosol as they are in e-liquid before vaping, in agreement with the finding reported earlier by Duell et al.¹⁰

2. Experimental Procedures

2.1. Materials

Thirty-three e-liquids were purchased online from a Dutch vendor of e-liquids, selected for variation in flavor, brand, propylene glycol (PG)/glycerol (Glc) ratio, and nicotine in free-base or salt form (as indicated on the packaging). Control e-liquids were prepared by combining PG, Glc, nicotine, and, if applicable, acetic acid, benzoic acid, levulinic acid, lactic acid, tert-butylamine, and sodium 3-(trimethylsilyl)-1-propanesulfonate (DSS). All chemicals were purchased from Sigma-Aldrich (Amsterdam, The Netherlands) and were used without further purification.

2.2. Nuclear Magnetic Resonance (NMR) Spectroscopy

¹H NMR spectra were recorded at ambient temperature on a Magritek Spinsolve (60 MHz; Aachen, Germany) at room temperature with the following settings: scans: 16, acq. time: 6.4 s, rep. time: 30 s, pulse angle: 60°. The two calibration samples for each test sample were prepared by addition of either 3 equiv of acetic acid (to ensure all nicotine is monoprotonated) or 1 equiv of tert-butylamine (to ensure the presence of free-base nicotine). All peaks were referenced to the DSS standard (20 mM).¹¹ Sample volumes of 0.5 mL were measured in Wilmad (type WG-1000-7) NMR tubes. Peak assignment was based on literature data.¹⁰ The nicotine peaks in the spectra were found by comparing the spectrum of the sample with spectra of nicotine in water and nicotine in a PG/Glc mixture (Figure 2). Further confirmation was obtained by verifying that these peaks shifted as described in the literature when the pH was changed. The chemical shift of each peak was considered to be at the highest intensity of the peak. Peaks H_a and H_d overlapped in the spectra due to the low resolution of the benchtop NMR spectrometer. As such, they were treated as one peak (designated H_a+d), which can be found between 8.5 and 8.8 (ppm).

Figure 2.

Structure of nicotine and ¹H NMR spectra of nicotine (20 mg/mL) in D₂O (top), PG/Glc (1:1 v/v) diluted 10× with D₂O (middle), and nicotine in PG/Glc (1:1 v/v) (bottom). Peak assignments are shown in the bottom spectrum and based on literature data.¹⁰

2.3. pH

pH values were recorded at ambient temperature using a Metrohm 781 pH/ion meter. Samples were diluted 10 times with deionized water, in accordance with the literature.¹ pH measurements were performed in duplicate. From the pH, the concentration of free-base nicotine was calculated using eqs 2 and 3, which are derived from the Henderson–Hasselbalch equation.

2

3

2.4. α_fb Calculation

For each individual e-liquid, two calibration samples (called “minimum” and “maximum”) were prepared with free-base and monoprotonated nicotine, respectively, as described by Duell et al.¹⁰ The chemical shift of the H_a+d peak was measured in the test sample and in these two calibration samples. α_fb was then calculated according to eq 4.

4

in which the symbols δH_a+d maximum and δH_a+d minimum are the chemical shifts of the H_a+d peak in the two calibration samples (i.e., of monoprotonated and free-base nicotine).

Whereas Duell et al. also incorporated the chemical shift of peak H_b in their calculations, this peak was often too broad and of too low intensity in the spectra acquired with the benchtop NMR spectrometer and therefore disregarded. Similarly, peak H_e cannot be reliably located in all e-liquids due to its occasional low intensity and overlap with other peaks. Supporting Information 2 describes how to utilize peak H_e (when detectable, and in addition to H_a+d) to calculate α_fb and describes a comparison of the two methods. To calculate the error in the determination of α_fb, duplicate measurements of 21 commercial liquids were performed, and the error was calculated as the average of the ranges of these measurements.

2.5. Machine-Generation of Aerosol

E-Cigarette aerosol was collected using the method described by Stephens et al.,¹⁷ with one modification: 0.75 g of cotton wool (Boom, Meppel, The Netherlands) was used. A Borgwaldt LM4E modular vaping machine (Körber Technologies, Hamburg, Germany) was used for automated puffing. 200 puffs were taken according to the ISO 20768:2018 puff regime (puff volume: 55 mL, puff duration: 3 s, puff frequency: 2 puffs/min).¹⁸ The e-cigarettes were mounted at a 20° angle in the vaping machine. For the e-cigarettes, Just Fog Q16 clearomizers (fitted with 1.6 Ω coils) and Eleaf iStick Pico batteries were used. The batteries were set to 11 W (using the constant power mode). The power setting was in accordance with manufacturers’ recommendation for this clearomizer (6.4–12.1 W) and was previously verified by a panel of human volunteers not to result in “dry puffs.”¹⁹ The clearomizer airflow intake opening was set to 1.5 mm using a metal spacer. E-Cigarettes were filled with e-liquid at least 1 h before each experiment.

3. Results

3.1. Method Development

Figure 2 shows that the peaks of the pyridine group of nicotine can be distinguished from the ¹H NMR spectra. Even though the peaks are relatively broad in PG/Glc (1:1 v/v) matrix, they are sufficiently well resolved to assess the protonation state of nicotine because the pyridine proton peaks (H_a, H_b and H_c) are located at a much higher shift than the large PG and Glc peaks (the methylpyrollidine protons (H_e) are not always well separated from the large PG and Glc peaks). The peaks in the ≥90% (v/v) D₂O spectra are significantly sharper, and they are even sharper, even in the absence of PG/Glc.

We compared our method for assessing the protonation state of nicotine with the methodology that relies on pH measurements (which requires diluting the sample with water). We prepared e-liquids containing a fixed amount of nicotine (20 mg/mL) and different amounts of acetic acid and PG/Glc ratios and compared the α_fb values measured with the two methods, for each liquid (see Figure 3). Higher values for α_fb were found with NMR, compared to α_fb calculated from pH measurements of the diluted samples.

Figure 3.

α_fb values of e-liquids with different PG/Glc ratios and different amounts of acetic acid, determined by ¹H NMR (filled symbols, solid lines) and pH measurements (open symbols, dotted lines) (Henderson–Hasselbalch method).¹

The PG/Glc ratio does not have a very strong effect, although the difference between the two methods appears to be smaller for the liquid with a PG/Glc ratio of 1:1.

3.2. Protonation State of Commercially Available E-Liquids

Next, we tested the feasibility of using a 60 MHz benchtop NMR instrument for commercially available e-liquids. Table 1 shows the fraction of free-base nicotine of 33 e-liquids.

Table 1. α_fb Values of 33 Commercially Available E-Liquids as Determined Using the Described Method.

	e-liquid name	PG percentagea	Nic. salt (Y/N)a	αfb
1	Dinner Lady Blackcurrant Ice	30	N	0.39
2	Dinner Lady Salt Nic Blackcurrant Ice	50	Y	0.05
3	Vampire Vape Heisenberg	60	N	0.91
4	Vampire Vape Nic Salts Heisenberg	50	Y	0
5	Cirkus The Authentics—Arctic Mint E-Liquid	50	N	1
6	Cirkus Nic Salt E-Liquid—Arctic Mint	50	Y	0.08
7	Edge E-Liquid—Very Menthol	55	N	1
8	Edge Nic Salt E-Liquid—Very Mentol	60	Y	0.22
9	Element E-Liquids—555 Tobacco	25	N	0.18
10	Element Nic Salt E-Liquid—555 Tobacco	35	Y	0
11	Element E-Liquids—Candy Punch	25	N	0.04
12	Element Nic Salt E-Liquids—Candy Punch	35	Y	0
13	Halo Nic Salt E-Liquid—Tribeca	70	Y	0.23
14	Halo E-Liquid Tribeca	70	N	0.88
15	Six Lick Love Bite	50	Y	0.04
16	T-Juice Clara—T Nic Salt	50	Y	0.04
17	T-Juice Green Kelly Nic Salt	50	Y	–0.05
18	Vampire Vape Bat Juice	60	N	0.8
19	Veep Menthol	20	N	1
20	Veep Spearmint	20	N	1
21	VGOD SaltNic Cubano Silver	45	Y	0.32
22	VGOD SaltNic Luscious	45	Y	0.21
23	ZAP! Starfruit Burst	30	N	1
24	Aramax—Classic Tobacco	50	N	0.96
25	IVG Nic Salt E-Liquid Bubblegum	50	Y	0.31
26	IVG Nic Salt E-Liquid—Cinnamon Blaze Chew	50	Y	0.11
27	Millers Nic Salt E-Liquid—Desert	70	Y	0
28	Zap! Aisu Nic Salt E-Liquid—Cucumber	50	Y	0.1
29	Zap! E-Liquid—Melonade	30	N	0.96
30	Zap! Nic Salt E-Liquid—Melonade	not specified	Y	0.11
31	Zap! Nic Salt E-Liquid −Passionfruit Zest	50	Y	0.36
32	Zap! E-Liquid—Passionfruit Zest	30	N	0.96
33	Zap! Nic Salt E-Liquid—Snow Pear	50	Y	0.22

^aThe Columns “PG Percentage” and “Nic Salt Y/N” List Information Indicated on the Product Packages.

Figure 4 shows the distribution of α_fb values. Two populations can be distinguished: the mostly free-base nicotine-containing e-liquids in the top band (α_fb > 0.80) and mostly protonated nicotine-containing e-liquids (α_fb < 0.40) in the lower band. Remarkably, three e-liquids, which are not advertised as nicotine salts, do show high protonation grades #1 (Dinner Lady Blackcurrant Ice, α_fb = 0.39), #9 (Element E-Liquids—555 Tobacco, α_fb = 0.18), and #11 (Element E-Liquids—Candy Punch, α_fb = 0.04). To calculate the error in the determination of α_fb, as shown in Figure 4, duplicate measurements of 21 of these commercial liquids were performed. The error was subsequently calculated as the average of the ranges of these measurements and was found to be 0.07.

Figure 4.

Measured α_fb values for 33 commercial e-liquids. The color of the symbols indicates whether the presence of nicotine salt was indicated on the packaging (orange) or not (blue). Numbers in labels refer to the e-liquids and their entry number in Table 1. Error bars indicate the error of the method (±0.035).

3.3. Effect of Vaping E-Liquids on Their Protonation State

To investigate if α_fb of e-liquid is affected by the processes that occur during vaping (e.g., heating, evaporation, condensation), the protonation state of nicotine was measured in e-liquids, and their condensed aerosol samples were collected after (automated) vaping. E-Liquids containing nicotine salts were prepared with three different organic acids (lactic acid, levulinic acid, and benzoic acid) at different concentrations (0, 0.5, and 1.0 mol equiv). All liquids contained 20 mg/mL nicotine and had a 50/50 (v/v) PG/Glc ratio. The measurements showed that for all of the liquids examined, α_fb was essentially unchanged in the aerosol samples collected after vaping, compared to the e-liquid before vaping (the average difference in α_fb before/after vaping was 3.2 ± 2.2%; measurements are in Supporting Information 3).

4. Discussion

Our results show that NMR is an excellent technique for measuring the protonation state of nicotine in e-liquids and that a cost-effective benchtop 60 MHz NMR instrument is adequate for this purpose. The observation that vaping does not affect the protonation state of nicotine implies that user-relevant information can be obtained directly from measurements of e-liquid samples, without the need for (time-consuming) automated vaping.

Two clusters can be clearly distinguished in the 33 e-liquids measured in this study: liquids containing predominantly protonated and free-base nicotine, respectively. Three liquids were found to contain nicotine salts, while this was not indicated on the packaging.

The NMR method was compared with a (simpler) method based on pH measurements of diluted e-liquid samples. Different results were obtained with these two methods, strongly suggesting that the dilution step necessary during sample preparation for the pH measurements affects the protonation state of nicotine. This stresses the importance of using a method that does not require modifications to the e-liquid, such as NMR, even if it is more time-consuming and more expensive.

We used standard NMR tubes and added DSS to the samples as a calibrant, in contrast to Duell et al.,¹⁰ who used coaxial NMR tubes to allow use of an external lock solvent. Pankow et al. have previously determined that DSS can be added in these types of samples without affecting the protonation state of nicotine,¹¹ and we also verified this experimentally (see Supporting Information 1) using pH measurements and NMR experiments.

Higher α_fb values were found with the NMR method compared to pH measurements of diluted samples. Assuming there is no effect of the e-liquid components on the pH meter, it could be concluded that the protonation of nicotine increases upon dilution of the e-liquids. In other words, methods that rely on measurements of diluted samples will be biased (measuring higher levels of protonation than those existing in the unmodified liquid). This finding is in accordance with the literature and underlines the importance of avoiding sample preparation (e.g., diluting (with water or other solvents) or extractions).^10,20

In principle, the shift of other peaks, such as H_e and H_b, can be used (in addition to H_a+d), which might be expected to yield more accurate results (because more information contained in the NMR spectrum is used). However, peaks H_e and H_b cannot be located reliably for all e-liquids due to the presence of intense solvent peaks and the broadening of the peaks in the viscous solvent. Therefore, we felt that it is preferable to consistently use the same method for all samples (and over time) to better allow comparisons of results. Furthermore, as described in Supporting Information 2, the results are not significantly different.

Because differences in the matrix effect between individual e-liquid samples were generally small, it could be considered to calculate average calibration values from a large set of e-liquids, instead of preparing two calibration samples (monoprotonated and free-base) for each individual sample. However, because a few liquids (three in our set of n = 33) were encountered that did exhibit significant matrix effects (difference in α_fb > 0.09), additional work and larger sample volumes required for preparing individual calibration samples are warranted.

We have tested the effect of vaping on the protonation state of nicotine in e-liquids using nicotine salts prepared with three different organic acids (commonly used in commercial e-liquids) and found no significant effect. This is in agreement with the findings of Duell et al., who also found a high degree of correlation of α_fb between collected aerosols and their parent e-liquids.¹⁰ It cannot be excluded that differences exist in this regard for other organic acids. Note that the measurements after vaping were performed on samples of condensed aerosol, so the assumption was made that the protonation state in vape does not change upon condensation.

5. Conclusions

We have shown that a low-resolution (60 MHz) benchtop NMR instrument can be used to determine the fraction of free-base nicotine in e-liquids. With the developed method, it was shown that vaping of e-liquids does not seem to affect their protonation state. The method was applied to analyze the fraction of free-base nicotine in 33 commercially available e-liquids. The e-liquids clustered into two groups, with predominantly protonated and free-base nicotine, respectively. All of the products with nicotine salts listed in their ingredients appear in the protonated cluster, and most of the products without declared nicotine salts appear in the free-base nicotine cluster. The method described here offers a convenient and fast way to detect nicotine salts for regulatory and research purposes.

Glossary

Abbreviations

NMR

nuclear magnetic resonance

Nic

nicotine

α_fb

fraction of unprotonated (free-base) nicotine

PG

propylene glycol

Glc

glycerol

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.chemrestox.3c00417.

• Influence of DSS on the protonation state of nicotine; additional use of peak He for determining α_fb; and nicotine protonation state before and after vaping (PDF)

Author Contributions

CRediT: Arnout P. T. Hartendorp conceptualization, data curation, formal analysis, investigation, methodology, project administration, supervision, validation, visualization, writing-original draft, writing-review & editing; Imane Ahlal data curation, formal analysis, investigation, methodology, writing-review & editing; Wouter F. Visser conceptualization, formal analysis, methodology, visualization, writing-original draft, writing-review & editing; Ernesto P. Baloe investigation, methodology, writing-review & editing; Daan G. W. Lensen investigation, methodology, writing-review & editing; Max J. van Alphen investigation, methodology; Hetty Nagtegaal investigation, methodology, writing-review & editing; Wilbert de Ruijter investigation, methodology, writing-review & editing; Walther N. M. Klerx: investigation, methodology, writing-review & editing; Reinskje Talhout conceptualization, funding acquisition, investigation, methodology, supervision, writing-review & editing.

This research was funded by The Netherlands Food and Consumer Product Safety Authority (NVWA), project 9.7.1.

The authors declare no competing financial interest.