A revision of the multiple‐path particle dosimetry model focusing on tobacco product aerosol dynamics

To assess the health impact of inhaled aerosols, it is necessary to understand aerosol dynamics and the associated dosimetry in the human respiratory tract. Although several studies have measured or simulated the dosimetry of aerosol constituents, the respiratory tract focus areas have been limited. In particular, the aerosols generated from tobacco products are complex composites and simulating their dynamics in the respiratory tract is challenging. To assess the dosimetry of the aerosol constituents of tobacco products, we developed a revised version of the Multiple‐Path Particle Dosimetry (MPPD) model, which employs (1) new geometry based on CT‐scanned human respiratory tract data, (2) convective mixing in the oral cavity and deep lung, and (3) constituent partitioning between the tissue and air, and clearance. The sensitivity analysis was conducted using aerosols composed of four major constituents of electronic cigarette (EC) aerosols to investigate the parameters that have a significant impact on the results. In addition, the revised model was run with 4 and 10 constituents in ECs and conventional cigarettes (CCs), respectively. Sensitivity analysis revealed that the new modeling and the physicochemical properties of constituents had a considerable impact on the simulated aerosol concentration and dosimetry. The simulations could be carried out within 3 min even when 10 constituents of CC aerosols were analyzed simultaneously. The revised model based on MPPD is an efficient and easy‐to‐use tool for understanding the aerosol dynamics of CC and EC constituents and their effect on the human body.


| INTRODUCTION
Understanding the dynamics of aerosols in the human airway is essential for investigating their influence on the body because the biological impact of inhaled substances at the cellular level depends on their local concentrations in the airway. 1 In particular, aerosols of tobacco products such as conventional cigarettes (CCs), electronic cigarettes (ECs), and heated tobacco products are composed of a wide variety of constituents with different properties, 1,2 and it is difficult to understand their behavior in respiratory systems.For example, nicotine, which is a representative constituent in CCs and ECs, can easily evaporate from droplets. 3Previous studies have reported methods for measuring the deposition and uptake amount of the constituents in tobacco product aerosols. 4used positron emission tomography imaging methods to visualize and analyze inhaled isotope-labeled nicotine, and Ishikawa et al. 5 estimated the inhaled amounts of aerosols by analyzing the exhaled aerosols.These methods can be used for broad measures of the amount of aerosol deposition and uptake, such as total, organ-specific, or lung lobe-specific dosimetry; however, other approaches are necessary for analyzing local concentrations of aerosols.7][8][9] These days, simulated dosimetry is practically employed for toxicological assessment.For example, the Organization for Economic Co-operation and Development (OECD) showed a case study in which human equivalent concentration was estimated by calculation based on computational fluid dynamics (CFD) modeling. 10In the tobacco product research area, there are some CFD models that can be used to simulate the deposition amounts of aerosols and individual chemical constituents by solving equations for continuous aerosol flow in the airways. 11,12Although CFD models enable the visualization of aerosol dynamics, a large amount of computational power would be required when considering the fate of multiple constituents throughout the entire airway.Therefore, the CFD approach is often used for site-specific modeling.Possible solutions for whole-airway modeling are semi-empirical models, 13 as well as deterministic mathematical models.In particular, the Multiple-Path Particle Dosimetry (MPPD) model, which employs a multiple-path deterministic mathematical approach, is widely used as a particle dosimetry model. 14,15arious modifications of the MPPD model have been proposed to simulate the deposition amount of tobacco product aerosols.Although MPPD v 3.04 (Applied Research Associates, Albuquerque, NM) is suitable for modeling the deposition of inhalants and air pollutants, there are limitations when considering the characteristics of aerosols from tobacco products.For example, this model does not consider phase changes of aerosol constituents in the calculation.In tobacco product aerosols, constituents with medium vapor pressures (approximately 1 Pa-5 kPa at 37 C), such as nicotine, change phase between the droplet and vapor. 16,17Thus, these dynamics should be considered when carrying out simulations for tobacco products.9][20] However, there is still room for improvement in these models.The geometry in the current MPPD models is based on a silicone rubber cast model that was made during an autopsy in which the geometry may be different from that of a live person. 21In addition, current models do not necessarily reflect tobacco product use behavior that causes mixing with diluted air in the body.When a person withdraws a puff of tobacco product, the puff is held in the mouth.After the mouth hold, dilution air enters the oral cavity and carries the puff into the deep lung, mixing the puff in the oral cavity. 22Once the puff reaches the alveolar region, additional mixing by convection occurs when the inhaled puff and residing air merge.In addition, in the current model, constituents in the gas phase are estimated to be absorbed all at once upon adhering to the surface of lung tissue.In practice, they distribute between the tissue surface and the air and then they are cleared from the surface. 23,24o fill the gap between the real-life situation and the mathematical model, we developed a revised model in which the geometry is updated with a minimally invasive method, a computed tomography (CT) scan, and the modeling accounts for the physical processes of tobacco product consumption and the tissue-air partitioning and clearance rate of aerosol constituents.With the revised model, we performed a sensitivity analysis by comparing the current and revised models, and by measuring the relationships between model inputs and outputs to investigate the parameters that have a significant impact on the modeling results.The sensitivity analysis was conducted using aerosols composed of four constituents, water, nicotine, propylene glycol (PG), and glycerol, which are the major constituents of EC aerosols.We also confirmed whether the results of sensitivity analysis were consistent with the well-known thermodynamic properties of aerosols.Finally, we performed deposition simulation with an increased number of chemical constituents that are typically found in CCs to investigate the applicability of the model for more complex aerosols.

| Revision of the MPPD model
The revised MPPD model for tobacco products was developed based on previous studies [18][19][20]25 with the support of Applied Research Associates. To eflect puff dynamics in the oral cavity, puff convective mixing with dilution air and reserved air in the alveoli, and the partitioning of constituents with a medium vapor pressure at the air-tissue interface in the lung, the following calculations for each constituent were incorporated into the revised model.

| Geometry
The geometry of the oral cavity was created with the same method reported by Asgharian, Price, et al., 18 based on a replica cast made by Cheng et al. 26 The upper respiratory tract (URT) geometry included the oral cavity, oral pharynx, and larynx down to the beginning of the trachea.For computational purposes, the geometry was divided into 52 sections of 3-mm length.Each section was assumed to have a cylindrical geometry with a diameter equal to the hydraulic diameter of the entire section.Thus, similar to the lower respiratory tract, the oral cavity can be thought of as a typical 52-path generation model.
The geometry data for the entrance to the lower respiratory tract up to airway generation 16 were obtained from CT and converted into an asymmetric model.The CT data were originally obtained in the previous study. 27The study was approved by the Institutional Review Board at the University of Washington (Seattle, WA) where the volunteer recruitment and CT imaging was performed.The University of Washington IRB approval number is 40712.Informed consent was obtained from each participant.The CT data was then processed by following procedure because information about the deep lung, including the branching angles, branch lengths, and branch diameters, cannot be obtained from a standard CT scan owing to resolution limitations.To reach down to airway generation 16, the arteries that travel alongside the airway branches were used as a surrogate for the airway paths.In this process, the resolved airways and arterial branches were tracked separately, and then they were joined together to form a unified geometrical representation of the lung.Conversion of CT data with deep lung geometry into usable (i.e., standard triangulated language) format was performed and provided by Battelle Memorial Institute (Columbus, OH) and the Pacific Northwest National Laboratory (Richland, WA).The deep lung geometry after 17 generations was estimated as in a previous study. 28

| Convective mixing of the puff with the dilution air
Tobacco users hold the puff in the oral cavity following the puff withdrawal.After the mouth hold, dilution air is inhaled and mixed with puff carrying the mixture into the deep lung.In an idealistic scenario, the entire amount of dilution air completely and instantaneously mixes with the puff in the oral cavity before entering the lower respiratory tract.The idealistic mixture concentration is found from a simple mass balance: where C mh is the average concentration of the vapor in the oral cavity at the end of the mouth hold, V o is the volume of the oral cavity, C v is the average vapor concentration of a constituent, q d is the dilution airflow rate, and t inh , t pw , and t mh are the times for the total inhalation, puff withdrawal, and mouth hold, respectively.Previous studies have adopted this idealistic scenario 19 ; however, mixing during the puff withdrawal of tobacco products is more complex.The mouth is filled with the puff at the end of the mouth hold.The inhaled dilution air enters the oral cavity, mixes with the puff changing the puff concentration over time, and carries it into the lungs.We adopted a new mixing model of the oral cavity (Figure 1) to simulate the tracheal inlet concentration during inhalation.In this model, the oral cavity with an initial uniform concentration is given by where C mh is the local concentration of the vapor in the puff throughout the oral cavity and V is the local volume of each oral cavity section.A mass balance is performed for the oral cavity for droplets and vapor constituents: where C v is the vapor concentration of a constituent, V p is the puff volume, and t is the elapsed time.The concentration leaving the oral cavity and entering the trachea (C inh ) is found from the solution of the above equation and is given by The constituent concentration leaving the oral cavity and entering the trachea is found from solving Equation (3) and is time dependent.To simplify the computations, we find an average concentration at the entry to the lower respiratory tract (trachea) by averaging the concentration over total inhalation time t inh ð Þ.The mass of inhaled materials entering the lower respiratory tract following the mouth hold m inh ð Þis given by where C inh is the average concentration of vapor leaving the oral cavity and entering the trachea.Thus, the average concentration entering the lower respiratory tract normalized by average concentration at the end of the mouth hold is represented as

| Convective mixing of the puff with the reserve air
The puff in the tracheobronchial (TB) region (airway generations 0-16) travels down into the deep lung by mixing with and pushing the existing air to the more distal airways.There is relatively little convective and diffusive mixing in the TB region because the airway expansion is small and the flow convection is high.Once the puff reaches the alveolar region, the axial convection is also small; however, lateral mixing via the expansion of the alveolar region occurs.Convective mixing of the puff with the reserve air in the deep lung occurs while the vapor is being transported through the airway and is being taken up by the airway tissue.Hence, mixing is coupled with uptake.Modeling details are provided below when introducing the transport model for vapors.

| Modeling of coupled puff transport equations at the airway surface
The convective mixing of vapors in an airway composed of respiratory bronchioles and alveoli can be accounted for in the mass balance equation for vapors in a simplified model in which the airway contains a lumen (respiratory duct) and unsmooth space, as shown in Figure 2. In Figure 2, A is the duct cross-sectional area and A 0 is the total crosssectional area, which includes the space in a single respiratory bronchiole.After reaching the walls, gases are assumed to partition between the air and the lung tissue according to Henry's law. 29,30As the gas concentration increases in the tissue, its absorption tends to slow in proportion to its saturation vapor pressure (Raoult's law). 31,32The full set of transport equations during inhalation and exhalation are solved to determine the concentrations of vapor in the air and lung tissue during a smoking session.The mass balance equation for the inhaled vapor in the airway of radial dimension r is given by: where in which q is the airflow rate through the airway section, z is the axial dimension, D v is the vapor diffusion coefficient in air, F is the mass fraction of a constituent in the droplet, C d is the droplet concentration in air, m d is the mass of a single droplet, and k g , k t , and k i indicate the air and tissue phase mass transfer coefficients, which are determined by, for example, the tissue-air partition coefficient PC ta ð Þ and clearance rate. 25,33Equation (7A) holds during both inhalation and exhalation.This equation is also applicable in the TB airways by letting A ¼ A 0 .The solution to Equation (7A) is obtained numerically for the entire respiratory tract during inhalation, lung hold, and exhalation.
During lung hold, the following equation is used, which drops the convective term in Equation ( 7): The dosimetry model for vapor constituents is coupled with the droplet transport model.The transport equation for the droplet concentration C d ð Þ in the air is given as: where D d is the vapor diffusion coefficient in air, λ d C d is the number of droplets lost to the walls per unit time per volume, and β is the coagulation kernel.The model is simplified to the following form for lung hold: Equations (7A)-(7C)-( 10) are used to calculate the concentration of an inhaled puff in the air and tissue throughout the respiratory tract during a full puff process, which consists of inhalation, lung hold, and exhalation.The deposition and uptake fractions in the lung airways are subsequently determined from the simulated concentrations.Additional details of the modeling approach are given in Asgharian, Price, et al. 18

| Sensitivity analysis and calculation with EC and CC conditions
We conducted sensitivity analysis by assuming the use of an EC whose puff mixture of vapors and droplets initially consists of water (1%), nicotine (3%), PG (48%), and vegetable glycerol (48%).The physicochemical properties of these constituents at 37 C were simulated with the thermodynamics calculator, ProPhyPlus (ProSim, Toulouse, France) and the EPA On-line Tools for Site Assessment Calculation 34 (Table 1).Regarding PC ta , we formally employed the bloodair partition coefficient estimated by the vapor pressure, octanol-water partition coefficient, and molecular weight. 35,36ll PC ta of water, nicotine, PG, and glycerol were estimated as greater than 1000.
The puff volume and droplet concentration were 50 cm 3 and 1 Â 10 9 droplets/cm 3 , respectively.The initial droplet diameter was 0.5 μm.The same as in the previous models, droplet diameter distribution was not accounted for in this revised model.The total mass of all constituents per puff was 3.7 mg.The initial aerosol temperature of the EC was set at 60 C (Table 3).The vapor pressures of water, nicotine, PG, and glycerol as a function of temperature were estimated using the Antoine equation.
The constituents and their proportions in CC aerosol were obtained from the chemical analysis of a 3R4F reference cigarette 37 (Table 2).We chose 10 of the major constituents in the 3R4F cigarette aerosol with medium and low vapor pressures (<5 kPa) in addition to water.The physicochemical properties of these constituents were estimated using the same method as for the EC.The puff volume and droplet initial diameter were as in the EC condition.The droplet concentration and total mass inhaled were set as 2.7 Â 10 10 droplets/cm 3 and 1.85 mg/puff: The aerosol temperature was set at 37 C (Table 3).Constituents whose PC ta was estimated as lower than 1000 are styrene and pyridine whose PC ta values were estimated as 34.0 and 228.9, respectively.

| RESULTS AND DISCUSSION
When users consume tobacco products, the puff is withdrawn, held in the mouth, and then transported into the deep lungs along with air, which dilutes the puff.9][20] To examine the effect of including these equations in the MPPD model, we ran the model calculation with various input conditions.We also input constituents of ECs and CCs into the revised model to confirm the feasibility of performing these calculations for tobacco products.

| Comparison of a new mixing model and idealistic mixture
When considering the puff concentration at the trachea inlet, simply averaging the dilution air with the entire puff is commonly used in current models. 19However, in a real-life situation, the dilution air pushes the puff into the trachea during which the puff concentration changes.To reflect this situation, we introduced a new mixing model in which the dilution air is mixed with the puff in the oral cavity (Equation ( 6)). Figure 3 a comparison of the average inlet concentration normalized by the average concentration at the end of the mouth hold between the new mixing model and the simple averaging model.The solution volume was also normalized by the volume of the oral cavity.Compared with the simple averaging model, the new mixing model calculated an increased puff concentration.The difference between the two models was maximum with a 0.6-1.2normalized dilution volume, corresponding to 30-60 mL of dilution air when the oral cavity volume is set as 50 mL.The two models converged with large volumes of dilution air.These results suggest that the simple averaging method underestimates the puff concentration.

| Puff mixing in the deep lung
In addition to accounting for mixing in the oral cavity, the revised model considers mixing in the alveolar region.Figure 4 shows the results of comparing the vapor uptake fraction in the alveolar region (i.e., airway generations 17-23)  with and without mixing.The vapor uptake fraction with mixing was less than that without mixing, reflecting that the airway cross-sectional area is wider than the duct cross-sectional area.As the vapor went deeper, the difference in uptake fraction with and without mixing was reduced because the vapor concentration decreased in the deep lung.
Vapor mixing was also employed for modeling the vapor released from droplets.

| Tissue-air partitioning at the lung tissue surface
Once gas comes into contact with lung tissue, it partitions between the tissue and air.This partitioning is important for simulating the actual exposure dose because constituents distributed in the air transfer to different areas of the tissue or are exhaled.The tissue-air partition coefficient PC ta ð Þ was added as an input parameter in the revised model.PC ta is a constituent-specific value, where PC ta = 1 means that a constituent is equally distributed between the tissue and air.A PC ta value less than or greater than 1 means that the constituent has a greater distribution in air or tissue, respectively.We performed a sensitivity analysis of the PC ta with nicotine as a representative constituent to evaluate the impact of this value when calculating the deposition and uptake when the PC ta was estimated as more than 1000.As expected, the uptake and deposition fraction when nicotine was predominantly in air PC ta ¼ 0:01 ð Þwas lower in the TB and F I G U R E 4 Vapor uptake fraction in the alveolar region with and without convective mixing.
pulmonary regions than when an equal distribution was assumed PC ta ¼ 1 ð Þ .In addition, by increasing the PC ta to 100, the deposition and uptake fraction in the TB region increased, whereas in the pulmonary region it was lower (Figure 5).This is because the amount of nicotine entering the pulmonary region was relatively low in the PC ta = 100 condition, as a larger amount of nicotine would be trapped in the TB region.
As such, varying the PC ta affected not only the amount of total deposition but also deposition in specific regions of the respiratory tract.The use of PC ta as a variable in deposition simulation is therefore important when calculating constituent-specific deposition trends.9][40] We employed one value of each constituent's PC ta for the whole airway to simplify the calculation in this study.

| Clearance from the lung tissue
Deposited gaseous chemical constituents are absorbed and removed from the airway surface and eventually eliminated by metabolization.Because the constituent concentration at the airway surface affects its uptake (Raoult's law), we introduced clearance rate-related coefficients into the calculation (Equation ( 7)).To examine the impact of the clearance rate on the deposition and uptake fraction, we ran the model for nicotine, PG, and glycerol with two different clearance rates (0.0001/s and 1/s).The deposition and uptake fractions of all constituents were found to be higher with a clearance rate of 0.0001/s than with a rate of 1/s (Figure 6).The change in dosimetry regardless of value of clearance rate was large in the order PG > nicotine > glycerol, which is consistent with the order of their vapor pressures.This result indicates that vapor uptake is promoted when the concentration at the tissue surface is low.The difference in deposition fraction between the 0.0001/s and 1/s calculations was less than 0.1%, so it was shown that the addition of the clearance rate calculation into the revised model had little effect on the simulation results.The major reason is that the model is to simulate one puff of aerosol.Potentially, models that account for accumulation from multiple puffs or daily use can have a large impact on values of clearance rate, 41 but it was not highly sensitive, at least when inputting representative EC constituents into this model.

| Deposition and uptake of tobacco aerosol constituents
We introduced several new equations into the model to account for the physicochemical properties and the dynamics of aerosol constituents.Figure 7 shows the dosimetry results for each area and generation of the respiratory tract when inputting these parameters, assuming that EC aerosols are composed of water, nicotine, PG, and glycerol.In the URT and TB regions, the deposition and uptake fractions were estimated to be in the order PG > nicotine > glycerol, which is consistent with the order of their vapor pressures (Table 1).However, the deposition and uptake fractions in the alveolar region did not show the same trends as in the URT and TB regions.In the alveolar region, especially for generations after 20, the deposition and uptake fraction of glycerol and nicotine were calculated to increase markedly.We believe that vapor uptake dominates in the URT and TB regions, whereas droplet deposition dominates in the alveolar region because the impaction and sedimentation of droplets tend to occur in complex airways, and the airways narrow as the airway depth increases. 42,43I G U R E 6 Effect of clearance rate on the deposition fraction in the alveolar region for each constituent.PG, propylene glycol.
F I G U R E 7 EC deposition and uptake fraction in (A) each airway region and (B) each airway generation.EC, electronic cigarette; PG, propylene glycol; TB, tracheobronchial; URT, upper respiratory tract.
We also calculated the deposition and uptake fraction of 10 representative constituents in CC aerosols (water, nicotine, PG, glycerol, triacetin, hydroquinone, catechol, phenol, pyridine, and styrene) in each airway region and airway generation (Figure 8).Unlike the constituents of EC aerosols (composed of only water, nicotine, PG, and glycerol), the calculated deposition and uptake fractions of the CC constituents in the URT and TB regions were not completely dependent on the vapor pressure (Table 2).Because the estimated activity coefficient corrects the vapor pressure, constituents with a high activity coefficient (such as PG and triacetin) deposit in the URT more than those with a high vapor pressure in this model.Catechol has a vapor pressure in the middle range among the CC constituents, and this compound was simulated to behave similarly to hydroquinone, which is a nonvolatile compound.This result was also owing to the activity coefficient; catechol had a low activity coefficient (<1) and its vapor pressure was suggested to be low, close to the level of a nonvolatile compound.Therefore, in addition to the vapor pressure, the value of the activity coefficient affected the airway dosimetry calculation results.
Nicotine is a constituent in both EC and CC aerosols.For this compound, we found considerable differences in the deposition fraction modeled at each site of the respiratory tract.When simulating EC aerosol (composed of only water, nicotine, PG, and glycerol), only 6.9% of nicotine was simulated to deposit in the URT.In contrast, it was approximately 26% when simulating CC aerosol composed of 10 constituents.In addition, in EC aerosol, the deposition and uptake fraction of nicotine was simulated to increase with lung depth and was maximum in the deepest lung (generations 21-23).In contrast, in CC aerosol, airway generation 9 had the maximum deposition and uptake fraction of nicotine.These differences resulted from differences in the aerosol properties of the EC and CC constituents, especially the difference in the activity coefficients.In this calculation, compared with EC aerosol, CC aerosol was set to contain a smaller number of droplets, and the constituents in CC aerosol could more easily evaporate.This results in smaller droplet diameters of CC aerosols; thus, nicotine evaporation is simulated to be higher for CC than for EC.Importantly, these calculations were completed within 3 min using a typical laptop computer.A versatile model must have a low computational load.Our model is suggested to enable the efficient evaluation of products with few constituents, such as ECs, as well as products with many constituents, such as CCs and heated tobacco products. 2,44he limitations of the model are that the calculation of deposition and uptake fraction are based only on theoretical aerosol dynamics.Because actual measurement of the deposition and uptake amount of chemicals in the respiratory tract is extremely difficult owing to the dynamic nature of chemical absorption and metabolism, combination with a physiologically based pharmacokinetics (PBPK) model would provide deeper insight into the model validity.For example, the pharmacokinetics of plasma nicotine profiles could be found in the literature, 45 thus the combination of PBPK model and the model output would be more realistic.In addition, the MPPD simulation method is only for the airway generation to generation.This method enables low computational load and whole-airway modeling, while this model cannot calculate heterogeneous aerosol deposition within the oral cavity or one airway generation as reported by Kuga, et al. [46][47][48] and Asgari, et al. 11 We need to choose an optimal modeling method fitting for the purpose.The MPPD model is chosen when a whole or deep airway aerosol simulation is needed.Feng et al. 12 and Kannan, et al. 49 reported that lung geometries are found to be different in each individual and the difference critically affects the dynamics of aerosols in the airway.We developed a model with a geometry from one individual; therefore, the use of multiple geometries in the model would strengthen model reliability.We believe that the model revised in this study is useful for the calculation of human equivalent concentration, which plays a key role for in vitro to in vivo extrapolation in toxicological assessments without animal studies.However, aerosol chemistry of tobacco products is complex, thus estimating the deposition fraction of all the constituents in an aerosol is not realistic.Therefore, the human equivalent concentration of an aerosol should be calculated with the representative constituents such as nicotine.Together with the difficulty underlying complete validation of the model, adaptation of an uncertainty factor is warranted when the model is applied in the toxicological assessment.

| CONCLUSIONS
We revised the MPPD model in which a new geometry of the lung was adopted, and two aerosol dynamics were introduced: mixing in the oral cavity and deep lung, and constituent partitioning between the tissue surface and air.The revised MPPD enabled us to calculate the deposition fraction of each chemical constituent of interest in each region and branching generation of the respiratory tract by inputting the physicochemical properties and aerosol properties.The model also accounts for tobacco product use behaviors.We ran the model under various conditions of assumed EC and CC aerosol chemical constituents and user behavior, and the simulations were generally consistent with the global understanding of aerosol physics.Moreover, the calculations in this model could be performed rapidly, even when accounting for multiple chemical constituents.Although limitations are that the modeling was based only on theoretical calculations without validation by actual data and the geometry is constructed from only one donor, we believe that our model will be useful for investigating the actual exposure concentrations of chemical constituents in tobacco product aerosols, and it could be applied in toxicological studies with an uncertainty factor.
DATA AVAILABILITY STATEMENT Data sharing not applicable to this article as no datasets were generated or analysed during the current study.

F
I G U R E 1 A new mixing model of the oral cavity.(A) Puff inhalation into the oral cavity, (B) mouth hold, (C), (D) dilution air is inhaled and mixed with the puff in the oral cavity, thus, the puff reaches the trachea at a different concentration.
Idealistic representation of the alveolar region surface.

F I G U R E 3
Comparison of the average vapor concentration in the puff entering the lower respiratory tract simulated by the two mixing models.Normalized dilution volume: dilution volume versus oral cavity volume, normalized average volume: vapor concentration at the trachea entrance versus vapor concentration at the end of the mouth hold.

F I G U R E 5
Influence of the tissue-air partition coefficient on nicotine uptake and deposition.PC ta , tissue-air partition coefficient; TB, tracheobronchial.

F
I G U R E 8 CC deposition and uptake fraction in (A) each airway region and (B) each airway generation.Pyridine and styrene were excluded from the deposition fraction results in each airway generation because they are solely deposited in the URT region.CC, conventional cigarette; PG, propylene glycol; TB, tracheobronchial; URT, upper respiratory tract.
Constituents with medium and low vapor pressures in conventional cigarette 3R4F at 37 C. Input aerosol properties of the electronic cigarette and conventional cigarette.
T A B L E 2