1. Introduction

Figure 1. Working principle of the e-cig topography sensor. (a) Schematic of a "smart e-cigarette" equipped with a built-in topography sensor comprised of a photometric sensor, a pressure sensor, and a micro-controller unit (MCU). Example of (b) optical signal and (c) inhalation pressure collected from the photometric sensor and the pressure sensor, respectively. (d) Schematic of the photometric sensor for detecting an e-cigarette aerosol concentration.

Figure 1. Working principle of the e-cig topography sensor. (a) Schematic of a "smart e-cigarette" equipped with a built-in topography sensor comprised of a photometric sensor, a pressure sensor, and a micro-controller unit (MCU). Example of (b) optical signal and (c) inhalation pressure collected from the photometric sensor and the pressure sensor, respectively. (d) Schematic of the photometric sensor for detecting an e-cigarette aerosol concentration.

Figure 2. A prototype smart e-cigarette with an e-cig topography sensor. (a) Integration of the e-cig topography sensor assembly with a commercial e-cigarette module. (b) Photograph of the e-cig topography sensor assembly. (c) Photographs of the core components of the sensors, compared to the size of a U.S. penny coin. (d)(e) Photographs of the constructed smart e-cigarette prototype.

Figure 2. A prototype smart e-cigarette with an e-cig topography sensor. (a) Integration of the e-cig topography sensor assembly with a commercial e-cigarette module. (b) Photograph of the e-cig topography sensor assembly. (c) Photographs of the core components of the sensors, compared to the size of a U.S. penny coin. (d)(e) Photographs of the constructed smart e-cigarette prototype.

2. E-cig Topography Sensor Working Principle

3. Materials and Methods

3.2. Sensor Signal Processing

Figure 1d shows the schematic of using MAX30105 sensor for conducting the photometric measurement of an e-cigarette aerosol. The three LEDs emit light at center wavelengths of 527 nm (green LED), 660 nm (red LED), and 880 nm (near-infrared LED), respectively. They are switched on/off in an alternating fashion such that only one color of light is detected by the photodiode at a time. The light absorbed in the active area of the photodiode generates a photocurrent, which is proportional to the optical power of light. The photocurrent is converted into a voltage signal, which is then digitized into an integer value using 14-bit analog-to-digital conversion. As a result, the output of the MAX30105 is a number which is proportional to the optical power absorbed by the photodiode. All three wavelengths have been individually tested in our experiments and the results from the near-infrared wavelength have been found to perform the best. Therefore, in this work, only the near-infrared wavelength is used.

Figure 3a shows one raw optical signal, $S_{raw}\left(t\right)$, acquired from the sensor when testing the smart e-cigarette with the vaping machine. The data sampling rate of the optical signal is 50 Hz, and the sampling interval is 20 ms, which is sufficient to resolve the rising and falling optical signal triggered by the generated e-cigarette aerosol. For all experiments in this work, sensor signals were acquired for a duration of 60 sec. This duration is much longer than the practical puff duration of a regular e-cigarette puff, however, it is necessary for the purpose of synchronizing data acquisition sequence of the e-cig topography sensor with the control sequence of the vaping machine. At the beginning of the raw optical signal, there was no e-cigarette aerosol, and the baseline signal, $S_{base}$, was calculated by the average of $S_{raw}\left(t\right)$ during the window between $t=$ 5 sec and $t=$ 10 sec (marked on Figure 3a). The actual optical signal, $S\left(t\right)$, shown in Figure 3b, was calculated by subtracting the baseline signal from the raw signal, i.e.$S\left(t\right)=S_{raw}\left(t\right)-S_{base}$, in order to correct the sensor’s response to zero aerosol.

Similarly, the signal from the pressure sensor, $P\left(t\right)$, is also plotted in Figure 3a. The BME680 detects the air pressure through piezoresistors which can measure the mechanical stress or strain induced by a membrane displaced by the applied air pressure. The data sampling rate of pressure signal is 5 Hz, and the sampling interval is 200 ms. Since the sampling rate for pressure is lower than that for the optical signal, the acquired pressure signal was interpolated at the time steps of the optical signal. A negative pressure can be clearly observed when the aerosol is being drawn out of the device. The ambient pressure, $P_{amb}$ was calculated by the average of $P\left(t\right)$ during the window between $t=$ 50 sec and $t=$ 55 sec (marked on Figure 3a) after the pressure has reached the stable ambient pressure after the puff. The inhalation pressure, $\Delta P\left(t\right)$, calculated as $\Delta P\left(t\right)=P_{amb}-P\left(t\right)$, is plotted in Figure 3b.

The e-cigarette aerosol is detected only when $S\left(t\right)$ is above a threshold, as illustrated in the magnified view of signals in Figure 3c. In this work, a threshold of 300 is applied in all signal processing. Each puff lasts from time stamp $t_0$ to $t_1$, which is located by the threshold. The puff duration is given by $t_p=t_1-t_0$. The puff duration measured in this approach is longer than the activation time of atomizer because the aerosol still lasts for a brief duration after the atomizer begins to cool down. Our definition of puff duration is slightly different from that given by CORESTA [43], but the insights that can be derived from the measured values are the same.

In the signal processing, the numerical calculation of mass of TPM, according to Equation (4), is carried out as

$$M_{puff}=\alpha {\int }_{t_0}^{t_1}S\left(t\right)\sqrt{\Delta P\left(t\right)}dt=\alpha AUC$$

(5)

where $AUC$ is the area under the curve (numerical integral) of $S\left(t\right)\sqrt{\Delta P\left(t\right)}$, as plotted in Figure 3d.

For all experiments in this work, this algorithm of signal processing was implemented into the MCU of the e-cig topography sensor through the Arduino C code. The measurement results were automatically displayed on the screen of the smart e-cigarette after each acquisition. The same algorithm was also applied when post-processing the collected sensor data using the MATLAB script.

Figure 3. Algorithm for sensor signal processing. (a) Raw signals collected from the photometric sensor and the pressure sensor. (b) Baseline-subtracted optical signal and inhalation pressure. (c) Determination of the puff duration with a threshold on the optical signal. (d) Calculation of the numerical integral of $S\left(t\right)\sqrt{\Delta P\left(t\right)}$ during the identified puff.

Figure 3. Algorithm for sensor signal processing. (a) Raw signals collected from the photometric sensor and the pressure sensor. (b) Baseline-subtracted optical signal and inhalation pressure. (c) Determination of the puff duration with a threshold on the optical signal. (d) Calculation of the numerical integral of $S\left(t\right)\sqrt{\Delta P\left(t\right)}$ during the identified puff.

3.3. Vaping Machine and Reference Measurement

In order to calibrate and validate the response of our e-cig topography sensor, the mass of TPM in the e-cigarette aerosols need to be measured from both the e-cig topography sensor and a reference aerosol setup, simultaneously. In this work, we implemented a home-made vaping machine and a commercial aerosol monitor, DustTrak II 8530 (manufactured by TSI), in the reference measurement setup, as shown in Figure 4. The upper limit of the working range of the aerosol monitor (AM) is 400 mg/m${}^3$. Since the concentrations of e-cigarette aerosols are much higher than this limit, the generated e-cigarette aerosols must be diluted before the measurement. We implemented a dilution box, with a volume $V_{box}$ = 0.0946 m${}^3$, to dilute the aerosol by a factor ranging from 300 to 1000. For a given puff, all the PM in the generated e-cigarette aerosol was captured and homogenized inside the dilution box. The concept of our reference measurements based on aerosol dilution shares similarities with the experimental setup used by Sousan et al. [56]. Multiple fans were installed inside the dilution box to create chaotic flows to quickly homogenize the e-cigarette aerosol within 20 seconds, which was verified by readings from multiple miniature optical particle counters (PMSA003, manufactured by Plantower Technology) installed at different locations inside the dilution box.

The mass concentration of the diluted aerosol, $C_{AM}$, was measured by the aerosol monitor. Accordingly, the reference measurement result of mass of TPM in each puff, $M_{Ref}$, is calculated by

$$M_{Ref}=C_{AM}{V}_{box}$$

(6)

A PM${}_{1.0}$ impactor plate was installed at the inlet of the aerosol monitor to measure particles smaller than 1.0 $\mu$m. According to existing studies, the PM in e-cigarette aerosols, when evaluated by mass concentration, are mostly fine particles [26,39,40,41,42,56]. In our experiments, we have tested both PM${}_{1.0}$ impactor and PM${}_{2.5}$ impactor, and found that PM${}_{1.0}$ is at least 97% of PM${}_{2.5}$, which is consistent with the reported trends [42,56]. The aerosol monitor used in this work can detect particles as small as 0.1 $\mu$m. Therefore, our reference measurements can effectively measure the mass concentration of PM in the particle size range of 0.1 $\mu$m to 1.0 $\mu$m. Based on the reported results, our detection range can effectively cover the total mass of PM in the e-cigarette aerosols [26,39,41,42,56]. Additional correction coefficients, which can be attained from gravimetric methods or more advanced aerosol instruments (such as SMPS), may be applied on our reference measurement results to further improve the accuracy. However, such improvements are expected to be marginal and the insights will not deviate from our findings based on this setup.

In our vaping machine, the switch button of the e-cigarette was automatically activated by a solenoid button pusher (BP), with a 3D-printed housing to fix the smart e-cigarette and the button pusher together. All the valves, pumps, and the button pusher were automatically controlled through relays, which were further controlled by a programmed Arduino Mega2560 MCU. In order to minimize the loss of e-cigarette aerosols in the tubing and valves, motorized ball valves (MBVs), with a wide inner diameter (ID) of 0.5 inch, were used as valves to pass aerosols, such as MBV1 and MBV2 marked in Figure 4a. For valves passing clean air, regular solenoid valves (SVs) with narrow passages were used, such as SV1 and SV2. Air pumps, P1 and P2, were installed to drive the flow of clean air, which had been filtered with HEPA zero filters (ZFs), such as ZF1 and ZF2.

A photograph of the experimental setup is shown in Figure 4b, with key components marked in the photograph. Figure 4c shows a flow chart of the control sequence used for every reference measurement. Before the measurement, clean air was filled inside the dilution box through ZF1, P1 and SV1 to purge all PM out of the box. When an acquisition sequence was activated by the Arduino Mega2560 MCU, valves MBV1 and SV2 opened, and P2 turned on to pump the air out of the dilution box to create a negative pressure in the dilution box. This negative pressure serves as the inhalation pressure mimicking a user’s inhalation. The peak inhalation pressure $\Delta P_{max}$ was controlled by the duration of pumping, $t_{vac}$, in this step. When $t_{vac}$ was varied from 28 sec to 112 sec, $\Delta P_{max}$ changed from about 280 Pa to 600 Pa. Near the end of the pressurization, the button pusher was activated to push the switch button of the e-cigarette for the preset duration, $t_{btn}$. The generated e-cigarette aerosol was first measured by the e-cig topography sensor inside the e-cigarette device, and then passed through MBV1 to enter the dilution box. There was a delay of 1 sec from the end of the button pusher activation to the beginning of closing MBV1 to allow sufficient time to capture all generated aerosols in the dilution box. After all aerosol has been captured inside the dilution box, MBV1 and SV2 closed and the aerosol was mixed with clean air to get homogenized in the box for 20 sec. After the aerosol got homogenized in the dilution box, SV1 and MBV2 opened, and the aerosol monitor (AM) switched on to measure the mass concentration of the diluted aerosol at a flow rate of 3 L/min. After AM collected the data, valves MBV2 and SV1 closed, and the box was thoroughly cleaned by removing all aerosols inside the box. The cleanness of the air in the box was confirmed by the reading of particle counters inside the box. Then, the vaping machine was ready for the next measurement. After a few experiments, residues were observed on the inner wall of the box, however, these residues adsorbed on the wall will not affect the PM of the aerosol inside the box, which was also confirmed by the particle counters inside the box.

This control sequence for reference measurements was activated by the Arduino Mega2560 MCU while the data acquisition sequence of the e-cig topography sensor was triggered by the MCU on the sensor. In our experiments, these two sequences were started separately and manually. As a result, there can be a random time shift, up to 15 sec, between the two sequences. Ideally, the aerosol should begin to appear in the optical signal around the same time stamp for all experiments. However, due to this artifact of random time shift between two sequences, the signals can occur at a shifted time stamp, as can be seen from certain graphs of this article. To deal with this time shift, we acquired signals for a very long duration of 60 sec for each puff, which can ensure that the sensor signals are completely collected when the aerosol is generated through the vaping machine during this long time frame. It should be noted that such a long duration of acquisition is only necessary for synchronizing the two sequences to compare the sensor’s reading with the reference measurement. For practical applications of the e-cig topography sensor alone inside an e-cigarette, the duration of signal acquisition can be shortened into a few seconds or automatically adapted according to the user’s actual puff duration.

Figure 4. Experimental setup for reference measurements of mass of TPM in the puff. (a) Schematic of the experimental setup. (b) Photograph of the constructed setup with a home-made vaping machine. (c) Flow chart of the control sequences used in each measurement of e-cigarette aerosols.

Figure 4. Experimental setup for reference measurements of mass of TPM in the puff. (a) Schematic of the experimental setup. (b) Photograph of the constructed setup with a home-made vaping machine. (c) Flow chart of the control sequences used in each measurement of e-cigarette aerosols.

3.6. Effects of Atomizer Power

Atomizer power and coil resistance are two of the main factors influencing the PM generated by e-cigarettes. Studies have shown that higher powers produce aerosols with larger mass of TPM per puff [34,38,39,40,41]. To study the effects of atomizer power on the aerosol output, the e-cig topography sensor has been tested by varying the atomizer power from 15 W to 40 W. Other conditions were kept at the default setting: 2.0 sec button pusher duration, 54 sec pressurization time of the dilution box. Three experiments were carried out for each configuration and all results are plotted in Figure 5 and summarized in Table A3.

As shown in the curves plotted in Figure 5a, using a higher atomizer power, a stronger optical signal was attained, which means an e-cigarette aerosol of a higher concentration. In addition, the transient details of the optical signal also suggest that with a higher atomizer power, the aerosol concentration grows to the peak at a faster rate, which is consistent with the expectations as a higher atomizer power gives a higher rate of thermal energy transfer to vaporize the e-liquid. For each condition, the relation between the optical signal and inhalation pressure validates that our home-made vaping machine has consistent and precise controls on the button pusher and the inhalation pressure.

Figure 5b shows the mass of TPM measured from our e-cig topography sensor, compared with the reference measurements. Overall, the results of our sensor match very well with the trend of the reference measurements. For atomizer power lower than or equal to 30 W, the mass of TPM grows almost linearly with increasing atomizer power, which is consistent with the trend reported by Floyd et al. [38]. Particularly, in this range, the relative errors of individual measurements are mostly less than 9%. As listed in Table A3, only one data point, trial No. 3, has a relative error higher than 9%. For atomizer power above 30 W, both the reference and e-cig topography sensor show a nonlinear relation of the mass of TPM versus the power, which is consistent with the logarithmic relation reported by Pourchez et al. [41]. For this range, the relative errors of the e-cig topography sensor have increased into 11%-17%. We attribute this higher error to the optical signals from an aerosol concentration outside the calibrated range. The calibration of our sensor was essentially a single-point linear calibration, carried out at 30 W atomizer power, and the peak optical signal acquired during calibration was around 7500 (Table A2). When testing the atomizer power above 30 W, the peak optical signal has reached higher than 8000 (Table A3), and the optical signal has deviated from the linear dependence versus the aerosol concentration due to multiple scattering of light from high-concentration particles [53].

Figure 5. Effects of atomizer power on the e-cigarette aerosol output. (a) Example signals acquired from the e-cig topography sensor. Plots of sensor signals for different atomizer powers specified by line styles. Solid line: 20 W; dashed line: 30 W; dotted line: 40 W. (b) Mass of TPM in the puff measured from the e-cig sensor (red triangles) and from the reference setup (black circles) versus atomizer power.

Figure 5. Effects of atomizer power on the e-cigarette aerosol output. (a) Example signals acquired from the e-cig topography sensor. Plots of sensor signals for different atomizer powers specified by line styles. Solid line: 20 W; dashed line: 30 W; dotted line: 40 W. (b) Mass of TPM in the puff measured from the e-cig sensor (red triangles) and from the reference setup (black circles) versus atomizer power.

3.7. Effects of Puff Duration

Puff duration affects the e-cigarette aerosol output significantly, as a longer puff duration can activate the atomizer for a longer period and generate more aerosol [26,32]. To study the effects of puff duration on the e-cigarette aerosol output, the e-cig topography sensor was tested by varying button pusher duration from 1.5 sec to 2.5 sec. Other conditions were kept at the default setting: 30 W atomizer power, 54 sec pressurization time of the dilution box. Three experiments were carried out for each configuration and all results are plotted in Figure 6 and summarized in Table A4.

As shown in Figure 6a, with a longer button pusher duration, a longer puff duration can be clearly observed from the wider peak in optical signal. From the data points plotted in Figure 6b, the mass of TPM almost grows linearly with longer button pusher duration, which is consistent with existing studies [26,32]. The data listed in Table A4 show that the relative errors of measurements are mostly less than 6%. Only one data point for 2.5 sec button pusher duration gives a relative error about 11%. This increased error is again related to the higher aerosol concentrations when the button was pressed for a longer duration, as can be seen from the peak optical signal.

Figure 6. Effects of puff duration on the e-cigarette aerosol output. (a) Example signals acquired from the e-cig topography sensor. Plots of sensor signals for different button pusher durations specified by line styles. Solid line: 1.5 sec; dashed line: 2.0 sec; dotted line: 2.5 sec. (b) Mass of TPM in the puff measured from the e-cig topography sensor (red triangles) and from the reference setup (black circles) versus button pusher duration.

Figure 6. Effects of puff duration on the e-cigarette aerosol output. (a) Example signals acquired from the e-cig topography sensor. Plots of sensor signals for different button pusher durations specified by line styles. Solid line: 1.5 sec; dashed line: 2.0 sec; dotted line: 2.5 sec. (b) Mass of TPM in the puff measured from the e-cig topography sensor (red triangles) and from the reference setup (black circles) versus button pusher duration.

3.8. Effects of Inhalation Pressure

Given an e-cigarette device, the flow rate is determined by the inhalation pressure and the air flow resistance of the device [36,54]. To study the effects of flow rate on the e-cigarette aerosol output, the e-cig topography sensor has been tested with varying inhalation pressure from 300 Pa to 600 Pa. This range has been decided in order to match one user’s inhalation pressure when testing the smart e-cigarette device. The inhalation pressure was controlled through the aforementioned pressurization duration on the dilution box, from 28 sec to 112 sec. Under this range of testing conditions, the volumetric flow rate, measured using a flow meter installed inline with pump P2 in Figure 4a, was found to range roughly from 4.5 L/min to 7.5 L/min. Other conditions were kept at the default setting: 30 W atomizer power, 2.0 sec button pusher duration. Three experiments were carried out for each configuration and all results are plotted in Figure 7 and summarized in Table A5.

As shown in Figure 7a, low pressure (around 300 Pa), medium pressure (around 450 Pa), and high pressure (around 600 Pa), were applied to draw the puffs from the e-cigarette mouthpiece. With a lower pressure and therefore a lower flow rate, the optical signal was stronger, which suggests a higher aerosol concentration, because the generated e-cigarette aerosol was extracted at a lower flow rate. From the reference measurements plotted in Figure 7b, the mass of TPM increases with higher pressure (higher flow rate), which matches the reported trends related to flow rate [35,37]. As shown in Figure 7b and Table A5, for low and medium pressure, the readings of our e-cig topography sensor matched very closely with the reference measurements with relative error lower than 7%. For high pressure (high flow rate), the relative error increased into 9%-13% and the reading of our e-cig topography sensor deviates from the reference measurements. In this work, a simplified relation between flow rate and inhalation pressure, in Equation (3), was used. For higher pressure, this equation requires further corrections to improve the accuracy, which can be achieved by more precise data fitting in the calibration procedure [36].

Figure 7. Effects of inhalation pressure on the e-cigarette aerosol output. (a) Example signals acquired from the e-cig topography sensor. Plots of sensor signals for different inhalation pressures specified by line styles. Solid line: low pressure (around 300 Pa); dashed line: medium pressure (around 450 Pa); dotted line: high pressure (around 600 Pa). (b) Mass of TPM in the puff measured from the e-cig topography sensor (red triangles) and from the reference setup (black circles) versus peak inhalation pressure.

Figure 7. Effects of inhalation pressure on the e-cigarette aerosol output. (a) Example signals acquired from the e-cig topography sensor. Plots of sensor signals for different inhalation pressures specified by line styles. Solid line: low pressure (around 300 Pa); dashed line: medium pressure (around 450 Pa); dotted line: high pressure (around 600 Pa). (b) Mass of TPM in the puff measured from the e-cig topography sensor (red triangles) and from the reference setup (black circles) versus peak inhalation pressure.

3.9. Effects of Cold Atomizer Coil

When an e-cigarette operates from a condition of a cold atomizer coil (at ambient temperature), referred to as "initial state", the aerosol output is significantly lower than that from a coil that has already been warmed up after a few puffs, known as "steady-state" [6]. To study the effects of the atomizer condition, the smart e-cigarette was tested starting from a condition when the atomizer coil was cold, and 6 puffs were experimented continuously with inter-puff interval of 3.5 min. The following setting was used for all the 6 puffs: 30 W atomizer power, 1.5 sec button pusher duration, 54 sec pressurization time of the dilution box. To investigate the temperature-related trends, the aerosol temperature was also collected from the BME680 sensor, using the approach as demonstrated in our previous work [45].

The sensor signals acquired from the first three puffs are shown in Figure 8a and the mass of TPM are given in Figure 8b. Starting from the cold atomizer coil, as more puffs were activated from the e-cigarette, the rising edges of the optical signals became steeper and reached higher amplitudes, as shown in the marked arrows. The mass of TPM in the first puff (for cold atomizer) was noticeably lower than that of the second puff and so on. The aerosol temperature also increased as the atomizer got warmer with more puffs. Both the trend of aerosol output and temperature are consistent with the reported studies [6]. The aerosol output reached the steady state after 5 or 6 puffs. Compared with the reference measurements, the e-cig topography sensor has clearly captured the trends as the atomizer coil gets warmed up. The relative error of the e-cig topography sensor’s reading is lower than 7%, in a level similar with the results in aforementioned tests. To test the repeatability of the trend, the experiments were repeated in another three cycles and the results were plotted in Figure 8c. The trend of increasing e-cigarette aerosol PM output as the atomizer coil gets warmed up can be clearly verified. For the cold atomizer (first puff), the sensor’s reading deviates from the reference measurements more than later puffs, probably due to the effects of air temperature inside the e-cigarette aerosol passage which is experienced by the optical sensor. The sensor was calibrated at the steady state, in which the optical sensor was in a warm environment. For the initial state (first puff), the temperature at the optical sensor was lower than that for the steady state (later puffs), which can cause the optical response to drift a little.

Figure 8. Effects of atomizer condition (cold or warmed-up coil) on the e-cigarette aerosol output. (a) E-cig topography sensor signals acquired during the first three puffs from the smart e-cigarette starting to operate from a cold atomizer coil. The rising edges of the optical signals are marked with arrows. (b) Mass of TPM in the puff measured from the e-cig topography sensor (red triangles), from the reference setup (black circles), and aerosol temperature (blue squares), versus puff number as the coil gets warmed up in one trial. (c) Mass of TPM versus puff number as the coil gets warmed up with experiments repeated in three cycles. Statistic quantities (mean and standard deviation), are plotted.

Figure 8. Effects of atomizer condition (cold or warmed-up coil) on the e-cigarette aerosol output. (a) E-cig topography sensor signals acquired during the first three puffs from the smart e-cigarette starting to operate from a cold atomizer coil. The rising edges of the optical signals are marked with arrows. (b) Mass of TPM in the puff measured from the e-cig topography sensor (red triangles), from the reference setup (black circles), and aerosol temperature (blue squares), versus puff number as the coil gets warmed up in one trial. (c) Mass of TPM versus puff number as the coil gets warmed up with experiments repeated in three cycles. Statistic quantities (mean and standard deviation), are plotted.

3.10. Tracking a User’s Puff Topography

One goal of the e-cig topography sensor is to measure the aerosol output and nicotine yield in each puff, regardless of the device setting of the e-cigarette, the atomizer condition, and the user’s vaping habits. As a proof of concept, the smart e-cigarette was directly tested by a user, one author of this article, with the user’s consent, using the user’s preferred setting of 20 W and 25 W atomizer power, as shown in Figure 9a. For each power setting, the user carried out 6 puffs with the user’s own comfortable puff duration and inhalation pressure. After each puff, the e-cig topography sensor’s reading was displayed on the device, showing the puff duration, the mass of TPM in the puff, the estimated nicotine yield, and the aerosol temperature, as shown in Figure 9b.

Figure 9c shows the sensor signals acquired for the first puff under 20 W atomizer power. It should be noted that the profiles of the signals are different from those acquired when driven by the vaping machine, because of the different fashions in applying the inhalation pressure from the user, compared with that from the vaping machine. Figure 9d shows the aerosol output and nicotine yield for all the puffs. The nicotine yield was estimated based on the weight concentration of nicotine in the e-liquid using the aforementioned approach. For 20 W and 25 W atomizer power respectively, the statistic quantity of the attained 6 puffs was analyzed and plotted in Figure 9e. It can be clearly observed that, when the user applied a higher atomizer power, the aerosol output increased significantly. Such a simplified statistical analysis can provide valuable information on the user’s puff topography. According to the puff duration and peak pressure plotted in Figure 9f,g, the trends of aerosol output are consistent with the aforementioned results attained from the vaping machine, which validated the effectiveness of our e-cig topography sensor in tracking the user’s puff topography. One can observe that, the parameters that are controlled by the same user, including the puff duration and peak pressure, can vary over a wide range, puff after puff. In these tests, the user simply took all puffs naturally and didn’t intend to change the vaping pattern. Despite these variations, using our e-cig topography sensor, the mass of TPM and nicotine yield were conveniently tracked for every puff, which provided crucial information for evaluating the health risks in the user’s vaping habits.

Figure 9. Tracking a user’s puff topography with the e-cig topography sensor. Photograph of (a) the user testing the smart e-cigarette and (b) the display on the device showing the results after one puff. (c) Sensor signals acquired during the first puff for 20 W atomizer power. The data point for this acquisition is marked as dashed circles on the curves in (d)(f)(g). (d) Mass of TPM in the puff and estimated nicotine yield from multiple puffs using 20 W and 25 W atomizer power. (e) Statistical quantity (mean, and standard deviation) of user’s puff topography for using 20 W and 25 W atomizer power, respectively, attained from the puffs shown in (d). (f) Puff duration and (g) peak inhalation pressure acquired from the user’s puffs using 20 W and 25 W atomizer power.

Figure 9. Tracking a user’s puff topography with the e-cig topography sensor. Photograph of (a) the user testing the smart e-cigarette and (b) the display on the device showing the results after one puff. (c) Sensor signals acquired during the first puff for 20 W atomizer power. The data point for this acquisition is marked as dashed circles on the curves in (d)(f)(g). (d) Mass of TPM in the puff and estimated nicotine yield from multiple puffs using 20 W and 25 W atomizer power. (e) Statistical quantity (mean, and standard deviation) of user’s puff topography for using 20 W and 25 W atomizer power, respectively, attained from the puffs shown in (d). (f) Puff duration and (g) peak inhalation pressure acquired from the user’s puffs using 20 W and 25 W atomizer power.

4. Discussion

5. Conclusions

Abbreviations

References

1. Grana, R.; Benowitz, N.; Glantz, S.A. E-cigarettes: a scientific review. Circulation 2014, 129, 1972–1986. [Google Scholar] [CrossRef] [PubMed]
2. Brown, C.J.; Cheng, J.M. Electronic cigarettes: product characterisation and design considerations. Tobacco Control 2014, 23, ii4–ii10. [Google Scholar] [CrossRef] [PubMed]
3. Chapman, S.L.C.; Wu, L.T. E-cigarette prevalence and correlates of use among adolescents versus adults: a review and comparison. Journal of Psychiatric Research 2014, 54, 43–54. [Google Scholar] [CrossRef] [PubMed]
4. Cameron, J.M.; Howell, D.N.; White, J.R.; Andrenyak, D.M.; Layton, M.E.; Roll, J.M. Variable and potentially fatal amounts of nicotine in e-cigarette nicotine solutions. Tobacco Control 2014, 23, 77–78. [Google Scholar] [CrossRef]
5. Breland, A.; Soule, E.; Lopez, A.; Ramôa, C.; El-Hellani, A.; Eissenberg, T. Electronic cigarettes: what are they and what do they do? Annals of the New York Academy of Sciences 2016, 1394, 5–30. [Google Scholar] [CrossRef]
6. Sleiman, M.; Logue, J.M.; Montesinos, V.N.; Russell, M.L.; Litter, M.I.; Gundel, L.A.; Destaillats, H. Emissions from electronic cigarettes: key parameters affecting the release of harmful chemicals. Environmental Science & Technology 2016, 50, 9644–9651. [Google Scholar] [CrossRef]
7. Hess, C.A.; Olmedo, P.; Navas-Acien, A.; Goessler, W.; Cohen, J.E.; Rule, A.M. E-cigarettes as a source of toxic and potentially carcinogenic metals. Environmental Research 2017, 152, 221–225. [Google Scholar] [CrossRef]
8. Dunbar, Z.R.; Das, A.; O’Connor, R.J.; Goniewicz, M.L.; Wei, B.; Travers, M.J. Brief report: lead levels in selected electronic cigarettes from Canada and the United States. International Journal of Environmental Research and Public Health 2018, 15, 154. [Google Scholar] [CrossRef]
9. Gaur, S.; Agnihotri, R. Health effects of trace metals in electronic cigarette aerosols—a systematic review. Biological Trace Element Research 2019, 188, 295–315. [Google Scholar] [CrossRef]
10. Holliday, R.; Chaffee, B.; Jakubovics, N.; Kist, R.; Preshaw, P. Electronic Cigarettes and Oral Health. Journal of Dental Research 2021, 100, 906–913. [Google Scholar] [CrossRef]
11. Keith, R.; Bhatnagar, A. Cardiorespiratory and Immunologic Effects of Electronic Cigarettes. Current Addiction Reports 2021, 8, 336–346. [Google Scholar] [CrossRef] [PubMed]
12. Cao, Y.; Wu, D.; Ma, Y.; Ma, X.; Wang, S.; Li, F.; Li, M.; Zhang, T. Toxicity of electronic cigarettes: A general review of the origins, health hazards, and toxicity mechanisms. Science of The Total Environment 2021, 772, 145475. [Google Scholar] [CrossRef] [PubMed]
13. Ween, M.P.; Moshensky, A.; Thredgold, L.; Bastian, N.A.; Hamon, R.; Badiei, A.; Nguyen, P.T.; Herewane, K.; Jersmann, H.; Bojanowski, C.M.; Shin, J.; Reynolds, P.N.; Crotty Alexander, L.E.; Hodge, S.J. E-cigarettes and health risks: more to the flavor than just the name. American Journal of Physiology-Lung Cellular and Molecular Physiology 2021, 320, L600–L614. [Google Scholar] [CrossRef] [PubMed]
14. Williams, M.; Villarreal, A.; Bozhilov, K.; Lin, S.; Talbot, P. Metal and Silicate Particles Including Nanoparticles Are Present in Electronic Cigarette Cartomizer Fluid and Aerosol. PLOS ONE 2013, 8, 1–11. [Google Scholar] [CrossRef] [PubMed]
15. Goniewicz, M.L.; Knysak, J.; Gawron, M.; Kosmider, L.; Sobczak, A.; Kurek, J.; Prokopowicz, A.; Jablonska-Czapla, M.; Rosik-Dulewska, C.; Havel, C.; others. Levels of selected carcinogens and toxicants in vapour from electronic cigarettes. Tobacco Control 2014, 23, 133–139. [Google Scholar] [CrossRef] [PubMed]
16. Lerner, C.A.; Sundar, I.K.; Watson, R.M.; Elder, A.; Jones, R.; Done, D.; Kurtzman, R.; Ossip, D.J.; Robinson, R.; McIntosh, S.; Rahman, I. Environmental health hazards of e-cigarettes and their components: Oxidants and copper in e-cigarette aerosols. Environmental Pollution 2015, 198, 100–107. [Google Scholar] [CrossRef]
17. Williams, M.; To, A.; Bozhilov, K.; Talbot, P. Strategies to reduce tin and other metals in electronic cigarette aerosol. PLoS One 2015, 10, e0138933. [Google Scholar] [CrossRef]
18. Mikheev, V.B.; Brinkman, M.C.; Granville, C.A.; Gordon, S.M.; Clark, P.I. Real-time measurement of electronic cigarette aerosol size distribution and metals content analysis. Nicotine & Tobacco Research 2016, 18, 1895–1902. [Google Scholar] [CrossRef]
19. Palazzolo, D.L.; Crow, A.P.; Nelson, J.M.; Johnson, R.A. Trace metals derived from electronic cigarette (ECIG) generated aerosol: potential problem of ECIG devices that contain nickel. Frontiers in Physiology 2017, 7, 663. [Google Scholar] [CrossRef]
20. Olmedo, P.; Goessler, W.; Tanda, S.; Grau-Perez, M.; Jarmul, S.; Aherrera, A.; Chen, R.; Hilpert, M.; Cohen, J.E.; Navas-Acien, A.; Rule, A.M. Metal Concentrations in e-Cigarette Liquid and Aerosol Samples: The Contribution of Metallic Coils. Environmental Health Perspectives 2018, 126, 027010. [Google Scholar] [CrossRef]
21. Williams, M.; Li, J.; Talbot, P. Effects of model, method of collection, and topography on chemical elements and metals in the aerosol of tank-style electronic cigarettes. Scientific Reports 2019, 9, 13969. [Google Scholar] [CrossRef]
22. Seiler-Ramadas, Radhika andSandner, I.; Haider, S.; Grabovac, I.; Dorner, T.E. Real-time measurement of electronic cigarette aerosol size distribution and metals content analysis. Wiener klinische Wochenschrift 2021, 133, 1020–1027. [CrossRef]
23. Schober, W.; Szendrei, K.; Matzen, W.; Osiander-Fuchs, H.; Heitmann, D.; Schettgen, T.; Jörres, R.A.; Fromme, H. Use of electronic cigarettes (e-cigarettes) impairs indoor air quality and increases FeNO levels of e-cigarette consumers. International Journal of Hygiene and Environmental Health 2014, 217, 628–637. [Google Scholar] [CrossRef] [PubMed]
24. Cho, J.H.; Paik, S.Y. Association between electronic cigarette use and asthma among high school students in South Korea. PloS One 2016, 11, e0151022. [Google Scholar] [CrossRef] [PubMed]
25. Talih, S.; Balhas, Z.; Eissenberg, T.; Salman, R.; Karaoghlanian, N.; El Hellani, A.; Baalbaki, R.; Saliba, N.; Shihadeh, A. Effects of User Puff Topography, Device Voltage, and Liquid Nicotine Concentration on Electronic Cigarette Nicotine Yield: Measurements and Model Predictions. Nicotine & Tobacco Research 2014, 17, 150–157. [Google Scholar] [CrossRef]
26. Fuoco, F.; Buonanno, G.; Stabile, L.; Vigo, P. Influential parameters on particle concentration and size distribution in the mainstream of e-cigarettes. Environmental Pollution 2014, 184, 523–529. [Google Scholar] [CrossRef]
27. Li, Y.; Burns, A.E.; Tran, L.N.; Abellar, K.A.; Poindexter, M.; Li, X.; Madl, A.K.; Pinkerton, K.E.; Nguyen, T.B. Impact of e-Liquid Composition, Coil Temperature, and Puff Topography on the Aerosol Chemistry of Electronic Cigarettes. Chemical Research in Toxicology 2021, 34, 1640–1654. [Google Scholar] [CrossRef]
28. Zhao, D.; Ilievski, V.; Slavkovich, V.; Olmedo, P.; Domingo-Relloso, A.; Rule, A.M.; Kleiman, N.J.; Navas-Acien, A.; Hilpert, M. Effects of e-liquid flavor, nicotine content, and puff duration on metal emissions from electronic cigarettes. Environmental Research 2022, 204, 112270. [Google Scholar] [CrossRef]
29. Sharpless, N. How FDA is Regulating E-Cigarettes. https://www.fda.gov/news-events/fda-voices-perspectives-fda-leadership-and-experts/how-fda-regulating-e-cigarettes.
30. Behar, R.Z.; Hua, M.; Talbot, P. Puffing Topography and Nicotine Intake of Electronic Cigarette Users. PLOS ONE 2015, 10, 1–18. [Google Scholar] [CrossRef]
31. Cunningham, A.; Slayford, S.; Vas, C.; Gee, J.; Costigan, S.; Prasad, K. Development, validation and application of a device to measure e-cigarette users’ puffing topography. Scientific Reports 2016, 6, 35071. [Google Scholar] [CrossRef]
32. Zhao, T.; Shu, S.; Guo, Q.; Zhu, Y. Effects of design parameters and puff topography on heating coil temperature and mainstream aerosols in electronic cigarettes. Atmospheric Environment 2016, 134, 61–69. [Google Scholar] [CrossRef]
33. Williams, M.; Bozhilov, K.; Ghai, S.; Talbot, P. Elements including metals in the atomizer and aerosol of disposable electronic cigarettes and electronic hookahs. PloS One 2017, 12, e0175430. [Google Scholar] [CrossRef] [PubMed]
34. Farsalinos, K.; Poulas, K.; Voudris, V. Changes in Puffing Topography and Nicotine Consumption Depending on the Power Setting of Electronic Cigarettes. Nicotine & Tobacco Research 2017, 20, 993–997. [Google Scholar] [CrossRef]
35. Floyd, E.; Greenlee, S.; Oni, T.; Sadhasivam, B.; Queimado, L. The Effect of Flow Rate on a Third-Generation Sub-Ohm Tank Electronic Nicotine Delivery System—Comparison of CORESTA Flow Rates to More Realistic Flow Rates. International Journal of Environmental Research and Public Health 2021, 18, 7535. [Google Scholar] [CrossRef]
36. Floyd, E.; Oni, T.; Cai, C.; Rehman, B.; Hwang, J.; Watson, T. Validation of a High Flow Rate Puff Topography System Designed for Measurement of Sub-Ohm, Third Generation Electronic Nicotine Delivery Systems. International Journal of Environmental Research and Public Health 2022, 19. [Google Scholar] [CrossRef]
37. Korzun, T.; Lazurko, M.; Munhenzva, I.; Barsanti, K.C.; Huang, Y.; Jensen, R.P.; Escobedo, J.O.; Luo, W.; Peyton, D.H.; Strongin, R.M. E-cigarette airflow rate modulates toxicant profiles and can lead to concerning levels of solvent consumption. ACS Omega 2018, 3, 30–36. [Google Scholar] [CrossRef]
38. Floyd, E.L.; Queimado, L.; Wang, J.; Regens, J.L.; Johnson, D.L. Electronic cigarette power affects count concentration and particle size distribution of vaping aerosol. PloS One 2018, 13, e0210147. [Google Scholar] [CrossRef]
39. Mulder, H.A.; Patterson, J.L.; Halquist, M.S.; Kosmider, L.; Turner, J.B.M.; Poklis, J.L.; Poklis, A.; Peace, M.R. The effect of electronic cigarette user modifications and e-liquid adulteration on the particle size profile of an aerosolized product. Scientific Reports 2019, 9, 1–8. [Google Scholar] [CrossRef]
40. Lechasseur, A.; Altmejd, S.; Turgeon, N.; Buonanno, G.; Morawska, L.; Brunet, D.; Duchaine, C.; Morissette, M.C. Variations in coil temperature/power and e-liquid constituents change size and lung deposition of particles emitted by an electronic cigarette. Physiological Reports 2019, 7, e14093. [Google Scholar] [CrossRef]
41. Pourchez, J.; Parisse, S.; Sarry, G.; Perinel-Ragey, S.; Vergnon, J.M.; Clotagatide, A.; Prévôt, N. Impact of power level and refill liquid composition on the aerosol output and particle size distribution generated by a new-generation e-cigarette device. Aerosol Science and Technology 2018, 52, 359–369. [Google Scholar] [CrossRef]
42. Alderman, S.L.; Song, C.; Moldoveanu, S.C.; Cole, S.K. Particle Size Distribution of E-Cigarette Aerosols and the Relationship to Cambridge Filter Pad Collection Efficiency. Contributions to Tobacco & Nicotine Research 2015, 26, 183–190. [Google Scholar] [CrossRef]
43. CORESTA. Cooperation Centre for Scientific Research Relative to Tobacco (CORESTA). [https://www.coresta.org].
44. Dunkhorst, W.; Lipowicz, P.; Li, W.; Hux, C.; Wang, Q.; Koch, W. In-situ characterization of e-cigarette aerosols by 90°-light scattering of polarized light. Aerosol Science and Technology 2018, 52, 717–724. [Google Scholar] [CrossRef]
45. Jiang, H. Smart Electronic Cigarettes with Built-in Aerosol Sensors. 2022 IEEE Sensors, 2022, pp. 1–4. [CrossRef]
46. Soulet, S.; Duquesne, M.; Toutain, J.; Pairaud, C.; Lalo, H. Experimental Method of Emission Generation Calibration Based on Reference Liquids Characterization. International Journal of Environmental Research and Public Health 2019, 16. [Google Scholar] [CrossRef] [PubMed]
47. Gillman, I.; Kistler, K.; Stewart, E.; Paolantonio, A. Effect of variable power levels on the yield of total aerosol mass and formation of aldehydes in e-cigarette aerosols. Regulatory Toxicology and Pharmacology 2016, 75, 58–65. [Google Scholar] [CrossRef]
48. Wu, J.; Yang, M.; Huang, J.; Gao, Y.; Li, D.; Gao, N. Vaporization characteristics and aerosol optical properties of electronic cigarettes. Environmental Pollution 2021, 275, 116670. [Google Scholar] [CrossRef]
49. Lampos, S.; Kostenidou, E.; Farsalinos, K.; Zagoriti, Z.; Ntoukas, A.; Dalamarinis, K.; Savranakis, P.; Lagoumintzis, G.; Poulas, K. Real-Time Assessment of E-Cigarettes and Conventional Cigarettes Emissions: Aerosol Size Distributions, Mass and Number Concentrations. Toxics 2019, 7. [Google Scholar] [CrossRef]
50. Wasisto, H.S.; Zhang, Q.; Merzsch, S.; Waag, A.; Peiner, E. A phase-locked loop frequency tracking system for portable microelectromechanical piezoresistive cantilever mass sensors. Microsystem Technologies 2014, 20, 559–569. [Google Scholar] [CrossRef]
51. Dunkhorst, W.; Lipowicz, P.; Koch, W. Characterization of highly concentrated organic aerosols by optical extinction in the mid infrared regime: Application to e-cigarettes. Journal of Aerosol Science 2016, 94, 33–42. [Google Scholar] [CrossRef]
52. Lee Black, D.; McQuay, M.Q.; Bonin, M.P. Laser-based techniques for particle-size measurement: A review of sizing methods and their industrial applications. Progress in Energy and Combustion Science 1996, 22, 267–306. [Google Scholar] [CrossRef]
53. Xu, R. Particle characterization: light scattering methods; Vol. 13, Springer Science & Business Media, 2001. [CrossRef]
54. Huynh, B.K.; Chen, Y.; Fletcher, D.F.; Young, P.; Zhu, B.; Traini, D. An Investigation into the Powder Release Behavior from Capsule-Based Dry Powder Inhalers. Aerosol Science and Technology 2015, 49, 902–911. [Google Scholar] [CrossRef]
55. Lim, M.; Lee, B. Development of a Portable Monitoring System for Indoor E-Cigarettes Emission. 2020 IEEE SENSORS, 2020, pp. 1–4. [CrossRef]
56. Sousan, S.; Pender, J.; Streuber, D.; Haley, M.; Shingleton, W.; Soule, E. Laboratory determination of gravimetric correction factors for real-time area measurements of electronic cigarette aerosols. Aerosol Science and Technology 2022, 56, 517–529. [Google Scholar] [CrossRef] [PubMed]
57. Chen, C.; Jiang, M.; Luo, X.; Tai, H.; Jiang, Y.; Yang, M.; Xie, G.; Su, Y. Ni-Co-P hollow nanobricks enabled humidity sensor for respiratory analysis and human-machine interfacing. Sensors and Actuators B: Chemical 2022, 370, 132441. [Google Scholar] [CrossRef]
58. Gai, Y.; Jiang, Y.; Li, Z. Advances in health rehabilitation devices based on triboelectric nanogenerators. Nano Energy 2023, 116, 108787. [Google Scholar] [CrossRef]