1 Introduction

Mechanical ventilation is meant as a temporary intervention to assist in breathing [8]. When a patient is undergoing mechanical ventilation treatment, it is very common to deliver aerosolized medication. These medicines can help to keep the airways fit for oxygen delivery and carbon dioxide expulsion and can help to treat the underlying conditions. It is also being used for insulin delivery, pain management, cancer therapy, and nano-therapeutics [15]. There are several commonly used drug aerosol treatments such as bronchodilators, corticosteroids, aerosolized antibiotics, and ipratropium bromide [9, 25]. The goal of drug delivery is to ensure the medications described above traverses the path from their release points to the site that will maximize their medical effectiveness. Determining where a particle deposits hinges on understanding the mechanisms that cause deposition. Airflow and particle transport in the human airways has been studied for decades; a comprehensive review can be found [15]. However, all these studies conducted under normal breathing conditions for the particles released through human oral [16, 28], nasal passage [12], at the trachea inlet [4, 11] or at the alveolus [18]. Aerosols therapeutic for patients under mechanical ventilation are complex [6]. The amount of drug delivered during conventional mechanical ventilation (CMV) is known to depend on several parameters that depends on the ventilation mode, variations in the ventilation circuit, drug type, and above all on patient specific conditions [6, 10]. The understanding of the mechanisms influencing particle transport to upper respiratory tract is essential to optimize ventilation strategies for patients in intensive care units [6].

Although there are several in vitro and in vivo studies that explored drug delivery under CMV, those studies focused on the total amount of aerosol passed the ventilator circuit. Furthermore, the applied approaches in those studies are not capable to characterize the particle transport and deposition in the large-scale lung model [5, 6]. In addition, there are only a few CFD studies of drug delivery under CMV, which focus on reducing lost aerosol on the circuits or on the aerosol devices such as MDI, DPI and jet nebulizer [19–21]. In a companion paper [27], the flow structures under invasive CMV were investigated for different respiratory waveforms and the flow mechanisms that could have a role on aerosol deposition were identified. The present paper focuses on understanding the effect of ventilator waveforms and intubation (via the endotracheal tube: ETT) on particle transport inside the human airways. The current study was motivated by the lack of work on understanding the role of mechanical ventilation parameters on pulmonary drug delivery, which is of great importance to clinicians for formulating effective ventilation management techniques.

2 Methods

2.1 Numerical technique and waveform creation

Large eddy simulation (LES was used to model gas transport in the domain and the aerosolized drug transport was modeled using a Lagrangian particle tracking technique that was implemented through a user-defined FORTRAN routine. LES was chosen for its ability to directly resolve the large eddies and the effect of small scales are modeled using subgrid scale (SGS) model; which gives it a distinct advantage over traditional turbulent models such as Reynolds-averaged Navier–Stokes (RANS). The wall-adapting local eddy (WALE) SGS model is adopted for its ability to reproduce the laminar to turbulent transition process [24]. The waveforms considered are: (a) pressure-controlled constant (P-constant), (b) pressure-controlled sinusoidal (P-sine), (c) volume-controlled ramp (V-ramp), and (d) flow-controlled ascending ramp (F-ascending ramp). These waveforms were created using equation of motion that governs flow under invasive mechanical ventilation [1, 27]. The obtained velocity from the waveform, prior to CFD simulation, was imposed at the inlet of the ETT. Details of waveforms creation and CFD inlet condition are provided in a companion paper part I [27]. No-slip condition was assumed at the wall and mass flow rate fraction as a function of the inlet mass flow rate was employed at the outlets. The essential details of realistic 3-D geometry reconstruction; numerical technique and boundary conditions; mesh generation and convergence study as well as respiratory waveforms creation are described elsewhere [27]. Figure 1 shows the lung model and the subdivided zones for the particle transport study.

Fig. 1
figure 1

Lung model and subdivided zones for particle transport study

Full size image

2.2 Particle transport modeling

A Lagrangian tracking model was implemented to calculate one-way coupled trajectories of mono-dispersed aerosols during invasive CMV. The particles used in this study were assumed to be spheres with diameters (d p) having a uniform distribution between 0.05 and 10 μm [29] and a density (ρ p) of 2000 kg/m3 [7]. At the highest volumetric flow rate seen during inspiration, the Stokes numbers (\(St_{\text{k}} = \frac{{\rho_{\text{p}} d_{\text{p}} C_{\text{c}} u}}{18\mu L},\) where C c is Cunningham coefficient, u is the fluid velocity, µ is fluid dynamic viscosity and L is a characteristic length which is assumed to be the diameter of the ETT). The Lagrangian transport equations can be expressed as [12]:

$$m_{{{\text{p}},i}} {\text{d}}v/{\text{d}}t = F_{{{\text{D}},i}} + F_{{{\text{L}},i}} + F_{{{\text{B}},i}}$$
(1)

where x p is the position of the particle, m p is the mass of the particle, v is the particle velocity, F D is the drag force, F L is the Saffman lift force, and F B is the Brownian motion force. The drag force is modeled as [23]:

$$F_{{{\text{D}},i}} = \rho \pi C_{\text{D}} d_{\text{p}}^{2} |u - v|(u_{i} - v_{i} )/8C_{\text{c}} ,\quad {\text{where}}\,C_{\text{c}} = 1 + \lambda /d_{\text{p}} \left( {2.34 + 1.05{\text{e}}^{{ - 0.39d_{\text{p}} /\lambda }} } \right)$$
(2)

where ρ is the fluid density, u is the fluid velocity, and λ is the mean free path. The lift force for a particle close to a wall is defined as [3]:

$$F_{{{\text{L}},i}} = - \rho_{\text{f}} d_{\text{p}}^{2} \hat{n}_{i} /4m_{\text{p}} \left[ {u_{\text{s}}^{2} \cdot g\left( {\kappa ,\varLambda } \right)} \right]$$
(3)

\(\hat{n}\) is the unit wall normal vector out of the domain. The wall tangent slip velocity is described by, \(u_{\text{s}} = \vec{v} \cdot \hat{t} - \vec{u} \cdot \hat{t}\) where \(\hat{t}\) is the unit wall tangent vector in the direction of the particle velocity. The Brownian force term is modeled as [17]:

$$F_{{{\text{B}},i}} = \zeta_{i} /m_{\text{p}} \sqrt {2k_{\text{B}}^{2} T^{2} /\left( {\tilde{D}\Delta t} \right)} ;\quad \tilde{D} = k_{\text{B}} TC_{\text{c}} /3\pi \mu d_{\text{p}}$$
(4)

where k B is the Boltzmann constant, T is the absolute temperature of the fluid, Δt is the time step used for particle integration, \(\tilde{D}\) the Brownian diffusivity and \(\zeta\) is a vector of three randomly generated numbers each taken from a Gaussian distribution with zero mean and unit variance. A total number of 105 particles were injected over the inspiration period at the ETT inlet. The distance from ETT inlet, where the particles were injected, to the carina ridge is ~10 cm. The interaction between the aerosol generation and the ventilator was not considered. To check the effect of particles count on the deposition fraction, the total number of the injected particles was increased by factor of 1.5 and only resulted in <1 % difference in the deposition rates. The particles were tracked over one cycle. An approach that commonly used for long cycle duration (4 s) [11]. Inthavong et al. [11] in their study of particle deposition under different breathing condition reported that one cycle (2 s) was sufficient for statistical convergence.

The particle deposition fraction results from the current numerical technique were validated against available experimental data from Kim and Garcia [13] (Re = 679–5548) for frequencies of 16–50 cycles/min, and experimental study of Kim and Iglesias [14] (Re = 1132–3369) for constant flow. A single bifurcation geometry that matches the model used in both experimental studies [13, 14] was reconstructed. Simulations with inlet flow of Re = 1500 and 4000 and St k number ranging from 0.01 to 0.25 were run. The results were compared to the experimental data. Figure 2 shows that calculated deposition fraction fits well with experimental data validating our present numerical approach.

Fig. 2
figure 2

Comparison of particle deposition efficiency between numerical results and different experimental data of Kim and Garcia [13] and Kim and Iglesias [14]

Full size image

3 Results

3.1 Particle end status

The particle end states, as they occur over time, are shown for the four waveforms in Fig. 3. The number of particles entering the domain is directly related to the flow rate, so the rate of particles entering the domain follows a shape similar to the flow rate waveform, while the total number of particles that have entered the domain is similar to the volume waveform. An offset is observed between the entry rate curve and the deposition rate curve, and between the entry rate curve and the escape rate curve, primarily due to the difference in location between the sites of the events. Deposition usually occurs where the ETT jet impinges [27] on the wall of the right main bronchus or at one of the carinas. The offset for the deposition rate curve is less than the offset for the escape rate curve because of the shorter distance the particles need to travel before deposition as compared to the distance traveled for escape. Very little offset is observed between the deposition and escape rate curves (Fig. 3b, c) for the waveforms that maintain a low flow rate (i.e., P-sine and V-ramp). Deposition occurs much more quickly than escape in the waveforms with a higher flow rate (i.e., P-constant and F-ascending ramp) near the onset of inspiration because so many of the particles will deposit before reaching the outlets (see Fig. 3a, d).

Fig. 3
figure 3

Particle end status for different waveforms. Data points represent the rate of particles entering, escaping, or depositing. Lines represent the sum of all of the particles that have entered, escaped, or deposited. T normalized = t/T, where T is the total cycle time, 4 s. The particle rate was normalized by the maximum rate of 2.11 × 105 particle/s

Full size image

At peak expiration, the deposition rates in Fig. 3 are much lower than at peak inspiration, despite the high flow rates. This behavior is due to the decrease in potential impaction sites and the reduction in turbulence including the absence of the large recirculation zone in the right main bronchus. The reduction in the deposition rate along with the high flow rate causes a high escape rate in expiration. An additional factor that contributes to the high escape rate at peak expiration is the higher percentage of small particles that did not impact in inspiration. In the P-constant and V-ramp waveforms, the highest escape rate is reached in expiration because of the large population of particles in the domain just before expiration. The P-sine waveform also has a large population of particles just before expiration, and a high escape rate is still seen despite the slow flow acceleration.

3.2 Total deposition

The most commonly used parameter to characterize total particle deposition (DFT) is the deposition fraction, defined as the ratio of number of particles that have deposited in the whole tracheobronchial domain to the number of particles that have entered the ETT. The deposition fractions for the P-constant and F-ascending ramp were found to be similar at ~56 %. Because the V-ramp does not reach as high of a flow rate, the deposition fraction is slightly lower at ~53.85 %. The deposition fraction for the P-sine is significantly lower than the others (~39.88 %) due to the lower flow rates in inspiration.

The total deposition fractions for different particle diameters are shown in Fig. 4a. Deposition at the lowest particle diameters was similar between the different waveforms. The P-constant and F-ascending ramp showed very similar deposition fractions over the diameter range (0.05–10 μm) of particles used in our analysis. The P-sine showed a lower deposition fraction due to the decreased role of impaction, turbulent dispersion, and turbophoresis. The V-ramp had a lower deposition fraction than the P-constant and the F-ascending ramp for particles between 2 and 4 μm. Because of the consistent flow rate in the V-ramp case, larger particles never had a period of time when they could pass the particle deposition hot spots, due to a very low flow rate. This is different from the P-constant and F-ascending ramp cases where higher flow rates existed, but there were also long periods of very low flow rate that lowered the deposition fraction for particles larger than about 3 μm.

Fig. 4
figure 4

a Total deposition fraction of particle diameter ranging from 0.05 to 10 μm for each waveform. b Deposition fraction of particle inside left and right lungs for each waveform

Full size image

Figure 4b shows the total deposition fraction inside the left and right lungs. A drastic difference between the deposition fractions in both lungs was observed for all waveforms with majority of the particles being deposited inside right lung due to the presence of the high-speed ETT-jet and its impingement [27], which increases drastically the particles impaction inside the right lung segments. For P-constant and F-ascending ramp waveforms, the deposition fraction inside right lung was ~12 times that inside left lung. Similarly, the deposition fraction inside right lung was tenfold more than left lung when employing the V-ramp. The ratio of the total particle deposited in right to left lung was reduced to 6 when using P-sine. This is due to the low flow rate, which relatively reduces the impaction mechanism in the right lung and allows more particles to pass to left lung.

3.3 Local deposition

Local deposition is characterized quantitatively by local deposition fraction (DFL) which is defined as the number of particle deposited in selected zone to the total number of injected particles at ETT inlet. The domain was subdivided into five zones as follows: zone 1 for ETT, zone 2 for trachea, zone 3, which includes main bronchi and right and left upper lobes, zone 4 in the middle of the domain, and zone 5, which includes lower regions of both lungs (see Fig. 1). Figure 5a–f shows local deposition fractions of various particle size ranges in different zones for all waveforms. High deposition occurred in zone 3 due to impingement of the ETT-jet on the right bronchus [27]. For particles with size >4 μm, the deposition fraction occurs greatly in zone 3 and then in zones 1 and 4. The lowest deposition was observed to occur in zones 2 and 5. For particles <4 μm, especially for submicron particles, i.e., d p < 1 μm (Fig. 5a), relatively different behavior was observed. Besides the high deposition in zone 3, enhancement of the deposition was observed in the lower regions, i.e., zones 4 and 5. This is because of the reduction of the impaction effect on smaller particles. The enhancement of the deposition in the lower zones (beyond trachea) was also reported by Xi et al. [29] due to presence of laryngeal jet. In all zones but in zone 3, the deposition fraction has only marginal difference between waveforms. In zone 3, however, the DFL significantly varies from one waveform to another, depending on the particles size. Zone 3 includes main bronchi where the strongest events of the flow structures were observed to take place such as jet impingement [27]. This interprets the different behavior of the local depositions in this zone. For example, due to higher flow rate in case of using P-constant and F-ascending ramp, significantly higher deposition occurred in zone 3 than the case when using V-ramp and P-sine waveforms. However, this difference becomes smaller as the particle size increases due to the increase in impaction mechanism, which is always dominant for larger particles (see Fig. 5a–f).

Fig. 5
figure 5

Local deposition fraction in identified zones for each waveform and wide range of particle size. Zones are identified in the left panel

Full size image

3.4 Deposition pattern and concentration

Qualitative identifications of particle deposition patterns are presented. Several deposition hot spots are visible and defined in Fig. 6a. Key deposition hot spots are labeled as: (1) site of impingement and the site where the ETT jet impinges and deflects off of the wall of the right main bronchus; (2) site where the deflected jet impinges on the wall of the right main bronchus; (3) the walls surrounding the large rotating structure in the right main bronchus; (4) superior wall of the branch leading to the right upper lobe; (5) superior wall of the left main bronchi near the first bifurcation; (6) carinas of the right lung; and (7) carinas of the left lung. Figure 5b compares the deposition patterns of submicron particles (50 nm–1 µ) and larger particles (9–10 µ). In addition to the deposition at carinas, the particles of the large size (9–10 µ) heavily deposited and concentrated in the inner wall of the right bronchus (site 1 in Fig. 6a) and the branch leads to the right lower lobe (site 3 in Fig. 6a). As a result, large number of the particles was filtered out and only a marginal amount of particles successfully reached the left side of the lung. This is attributed to the enhancement of the impaction mechanism as a consequence of the high-speed ETT-jet impingement. The large swirling flow from right bronchus to left lung causes the velocity to skew toward the outer wall of the left bronchus [27]. As a result, the large particles that transport to the left side, greatly deposited at the outer wall of the left bronchus near the bifurcation site (site 5 in Fig. 6a). In contrast to the large particles that significantly affected by the impaction, the submicron particles (50 nm–1 μm) have relatively uniform deposition, as shown in Fig. 6b. The submicron particles are known to greatly be influenced by the turbulence dispersion and secondary flow rather than impaction mechanism; consequently, small particles tend to follow the flow trajectory. As a result, the particles diffusion is enhanced, which in turn leads to more spreading of the deposition. Furthermore, the deposition of small particle in the lower regions and left lung is significantly enhanced compared to the large size particle.

Fig. 6
figure 6

a Key deposition hot spots, b deposition pattern of different particle sizes

Full size image

Particle deposition locations colored by particle diameter are presented in Fig. 7 for each waveform. For the P-sine case, only particles with higher diameter (~greater than 3 μm) deposited at sites 1–5 (Fig. 6a) and most of the larger particles are filtered out before traveling through the left main bronchus. The P-constant and F-ascending ramp cases show a much greater range of particles deposited at sites 1–5, and most of the higher diameter particles are filtered out before reaching site 5 which represents the superior wall of the left main bronchi (near the first bifurcation). The V-ramp shows a larger range of particle sizes deposited at sites 1–5 than the P-sine but a smaller range of particle sizes deposited than the P-constant or F-ascending ramp. These results suggest that a higher peak inspiratory flow rate will cause particles to deposit at a shorter downstream distance and that this effect is more pronounced for larger particles.

Fig. 7
figure 7

Deposition pattern by particle diameter for each waveform

Full size image

Particle deposition locations colored by the time of deposition are analyzed for each waveform in Fig. 8. In general, particles did not deposit in the lower airways until later times and particles deposited in the left lung at a later times than the right lung because the flow would direct particles through the right lung before reaching the left lung because of ETT orientation [27]. The particles deposited most heavily at times near the time of peak inspiration because of the influence of impaction, turbulent dispersion, and turbophoresis. The time of peak inspiration for the P-constant and F-ascending ramp were sharp and occurred at 0.1 and 1.0 s, respectively, and the corresponding colors for these times are clearly dominant. The peak inspiration for the P-sine is smoother and occurred at 1.0 s so a wider range of colors can be seen, but they are still centered around the colors corresponding to 1.0 s. The V-ramp was at peak inspiration between 0.1 and 1.0 s so a wide range of colors are seen.

Fig. 8
figure 8

Deposition pattern at corresponding ventilation cycle time for each waveform

Full size image

The deposition enhancement factor (DEF) was computed for during the four waveforms. It is defined by Balshazy et al. [2]:

$${\text{DEF}} = \frac{{N_{i} /A_{i} }}{{N_{\text{T}} /A_{\text{T}} }}$$
(5)

where N i is the number of particles deposited on patch i, A i is the area of patch i, N T is the total number of particles deposited on the domain, and A T is the total area of the domain upon which particles could deposit. Balshazy et al. [2] have shown that DEF is highly dependent upon patch size used and that a larger patch size will result in a lower maximum DEF. In the current study, a new triangular surface mesh was used to facilitate the calculation of DEF. A patch was taken to be all of the faces adjacent to a single node, and the patches were allowed to overlap as needed. The average area of a patch used was 0.20 mm2, and the standard deviation of the patch area was 4.76 × 10−2 mm2. The values of DEF were calculated for each node on the surface. In general, higher values of DEF are present in the right lung compared to the left lung due to the increased deposition. The differences in DEF between the key deposition sites are more pronounced for the P-constant and F-ascending ramp than the P-sine or V-ramp waveforms. This occurs because, at higher flow rates, there are more particles depositing at the lower numbered sites and this leaves a smaller number of particles left to deposit at the higher number sites. At lower flow rates, the selection of particles reaching the different deposition sites is more similar. The maximum DEF was 313.3 for P-constant, 167.5 for P-sine, 190.9 for V-ramp and 400.1 for F-ascending ramp. These values are similar to those presented in the study by Xi et al. [29]. In general, there are much greater differences in the maximum DEFs between the waveforms than the overall deposition fractions. For example, the percent difference between the overall deposition fraction for the V-ramp and F-ascending ramp is 6.20 %, while the percent difference between the maximum DEFs is 70.78 %. This suggests that while the overall particle deposition is similar for the two waveforms, the particle deposition locations are much more concentrated for the flow control ascending ramp waveform than for the volume-controlled ramp waveform. This could have significance for deposited drug aerosols that may have negative side effects in large concentrations.

4 Discussion

Deposition fractions as a function of particle diameter (Fig. 4a) have been reported in other studies [22, 29, 31, 33]. Of these studies only one by Zhang et al. [33] used an unsteady inlet condition. Geometric differences such as curvature and size of the airways, shapes of the carinas, branching angles, overall size (area) of the domain, the presence of ETT and respiratory waveform can also account for differences in deposition fractions. Studies by Xi et al. [29] and Zhang et al. [31] included the larynx and the studies by Xi et al. [29] and Luo and Liu [22] used CT-scan-based geometry. The presence of a jet can alter fluid flow and particle deposition in the airways as well as differences in flow rates [22, 29]. Deposition fractions presented in Fig. 3a at lower diameters agree well with the study performed by Luo and Liu [22] when compared to the pressure-controlled sinusoidal waveform with similar peak inspiratory flow rate. The deposition fractions in this study are about three times higher than those reported by Luo and Liu [22] for other particle diameters. This is likely due to the presence of the ETT jet increasing the role of impaction [27]. Even when the laryngeal jet is included, as in the study by Zhang et al. [31], higher deposition fractions (~3 times higher) are seen in the current study. This is likely due to the closer proximity of the ETT jet to the carina of the first bifurcation and the absence of the oral airways filtering out a large percentage of the particles due to strong upstream deposition [31]. In general deposition fractions in the current study are higher than those reported in previous literature which is hypothesized to be caused by the unique characteristics of the ETT jet as well as geometric differences [27].

The deposition at the impingement sites of the ETT jet is caused by impaction similar to that reported by Xi and Longest [28]; with the impaction sites for the current work appearing in the right main bronchi. The deposition caused by the large rotating structure is not as concentrated as the deposition at the impingement site; however, this is clearly an area of enhanced deposition caused by turbulent dispersion and turbophoresis similar to the enhanced deposition around the recirculation zone in the trachea [29]. The deposition on superior wall of the branch leading to the right upper lobe and the superior wall of the left main bronchi near the first bifurcation are caused by impaction as the flow from the right main bronchi changes direction rapidly, while it is redirected into the left main bronchi. This phenomenon is unique to this study and has not been reported previously. Enhanced particle deposition at the carinas is an established result of inertial impaction and has been reported in several prior studies [16, 22, 29, 33].

The CT-scan-based studies showed much less deposition in the right main bronchus and a much greater deposition on the ventral side of the left main bronchus than the current study [16, 22, 29]. In the current study there was a much greater deposition in the right main bronchus compared to the CT-scan-based studies of the upper tracheobronchial region due to the impaction caused by the ETT jet and the large rotating structure in the right main bronchus along with the increased turbulence [16, 22, 29]. The deposition is much greater on the superior side and much less on the inferior side of the left main bronchus than in the CT-scan-based studies of the upper tracheobronchial region because of the way the air enters the left main bronchus [16, 22, 29].

DEFs have been reported by studies in the past using a steady inlet condition [26, 29–32]. The studies performed by Shi et al. [26], Zhang et al. [32], and Zhang and Kleinstreuer [30] were focused on nano-scale particles, and the study by Zhang and Kleinstreuer [30] was performed in a simplified model of the oral airways. Only the studies by Xi et al. [29] and Zhang et al. [31] included a jet structure and the first few bifurcations and the study by Xi et al. [29] was the only study with geometry based on CT scans. In previous studies the carinas were areas where there was significant particle deposition [26, 29, 31, 32]. In the current study, there was similar DEFs seen at the carinas; however, there were also large DEFs at site of impingement and the site where the ETT jet impinges and deflects off of the wall of the right main bronchus. Furthermore, at the site where the deflected jet impinges on the wall of the right main bronchus; the walls are surrounding the large rotating structure in the right main bronchus; and the superior wall of the branch leading to the right upper lobe (sites 1–4 in Fig. 6a).

Despite the valuable results from present simulations, there are several limitations that could be considered for a future research. The change of particle size due to mass transfer is an important factor that could affect particle deposition throughout the domain. In addition, varying airways’ resistance and compliance is thought to alter the respiratory mechanics and particle transport as consequences. The considered geometry was rigid, even though the wall cartilage rings support the upper airways, the wall deformation could reveal important results.

5 Conclusions

Drug delivery to the realistic upper tracheobronchial region with invasive ETT for a patient receiving CMV was investigated by varying respiratory waveforms. LES was used to model the turbulence under deferent CMV conditions. A user-defined FORTRAN routine was implemented to track the transported particles throughout the domain. From the quantitative and qualitative evaluation of particle deposition within the tracheobronchial model, the results may be summarized as follows:

  • It was found that particle impaction was linked to inspiratory flow rate with the rate of particle deposition generally following the inspiratory flow rate. Flow-controlled ascending ramp, volume-controlled ramp and pressure-controlled constant result in similar total deposition fraction and was ~43.65 % greater than using pressure-controlled sinusoidal waveform.

  • The presence of ETT jet and its orientation and impingement location demonstrated a significant role on the local and total deposition. As consequences of intubation orientation toward right bronchus, the total deposition inside right lung was drastically higher (~12 times) than that inside left lung. This ratio decreases 50 %, in case of using pressure-controlled sinusoidal waveform resulted in most balanced distribution of particles between the left and right lungs.

  • Unique key deposition locations were identified including: the site of impingement and deflection of the jet caused by the ETT on the wall of the right main bronchus, the site where the deflected jet impinges on the wall of the right main bronchus, the walls surrounding the large rotating structure in the right main bronchus, the superior wall of the branch leading to the right upper lobe, the superior wall of the left main bronchi near the first bifurcation, and the carinas of the airway bifurcations.

  • The size of the particles depositing at the key deposition locations is linked with peak inspiratory flow rate due to the effects of particle impaction, turbulent dispersion and turbophoresis. The time at which particles deposit is also typically matched with the time of peak inspiratory flow.

  • The differences in the DEFs and the overall deposition fractions give evidence of drug aerosol concentrations in key deposition sites, which may be significant for drugs with negative side effects in high concentrations.