5967_09_p460-473 12/6/05 1:28 PM Page 460

JOURNAL OF AEROSOL MEDICINE

Volume 18, Number 4, 2005
Mary Ann Liebert, Inc.
Pp. 460473

Experimental Measurements of Particle Deposition in

Three Proximal Lung Bifurcation Models with an
Idealized Mouth-Throat

Y. ZHANG and W.H. FINLAY

ABSTRACT

In this paper, particle deposition in three idealized proximal lung bifurcation models with
an idealized mouth-throat were investigated experimentally. These bifurcation models in-
cluded (1) a small symmetric bifurcation, (2) an intermediate asymmetric bifurcation, and (3)
a large symmetric bifurcation. An idealized mouth-throat geometry (the Alberta geometry)
was used as the inlet to these bifurcation models. Monodisperse aerosol particles of DEHS
(di-2-ethylhexyl-sebecate) oil with mass median diameters in the range of 2.57.5
m were
employed at steady flow rates of 3090 L/min. Particle deposition measurements were con-
ducted by gravimetry. The results show that particle deposition in the mouth-throat and tra-
chea accounts for the major portion of total deposition in the entire models used, and parti-
cle deposition fraction in the proximal lung bifurcations is lower compared with that
deposited in the regions upstream (the mouth-throat and the trachea). Total particle deposi-
tion efficiency increases with increasing either inertial parameter or Stokes number. Total
particle deposition varies appreciably from model to model. The laryngeal jet is the key fac-
tor dominating particle deposition within the trachea. An effect of Reynolds number on par-
ticle deposition efficiency in the trachea is observed. In addition, particle deposition in the
bifurcation region is influenced little by the upstream flow condition, and therefore the ef-
fect of the laryngeal jet on deposition seemingly does not propagate to the bifurcations down-
stream.

Key words: particle deposition, mouth-throat, trachea, proximal lung bifurcation,

laryngeal jet

INTRODUCTION monodisperse aerosols deposited in the tracheo-

bronchial (TB) tree have been performed with

D ETERMINATION OF AEROSOL DEPOSITION

human respiratory tract is an important
in the

topic related to the effects of pollutant aerosols

on the human body and the delivery efficiency
of pharmaceutical agents for treating lung dis-
eases.(1) In vivo experimental measurements of
iron oxide particles(2,3) and polystyrene parti-
cles.(4,5) The mouthpieces used in these studies
were straight tubes with diameters of 1025 mm.
Significant variation of particle deposition effi-
ciency in the tracheobronchial (TB) tree was ob-
served in these studies. However, in vivo experi-

Department of Mechanical Engineering, University of Alberta, Edmonton, Alberta, Canada T6G 2G8.

460
5967_09_p460-473 12/6/05 1:28 PM Page 461
PARTICLE DEPOSITION IN PROXIMAL LUNG BIFURCATION MODELS
ments are not always practical due to cost and ra-
diological exposure concerns. At present, since
the resolution of radionuclide imaging method
does not allow accurate discrimination of indi-
vidual generations of the entire TB tree, in vivo
experimental data cannot be mapped onto the
generational level throughout the entire lung.(6)
As an alternative, in vitro experiments in physi-
cal models of human airways are often used to
obtain further information relating to particle de-
position. The in vitro experimental studies on par-
ticle deposition in single bifurcation models,(7,8)
multiple generation airways,(911) and human air-
way casts(2,1216) have demonstrated that aerosol
deposition occurs mainly near the bifurcation re-
gion and particle deposition fraction increases
with increasing Stokes number. Inertial im-
paction is the dominant deposition mechanism
for particles a few microns in diameter at typical
inhalation flow rates. The effects of branching an-
gle, the branching pattern (symmetry or asym-
metry) and the ratio of the daughter to parent
tube are minor.
However, most existing in vitro experimental
studies mentioned above were conducted in iso-
lated segments of bifurcation models or human
airway casts with an idealized inlet condition that
does not typically include a mouth-throat geom-
etry. In reality, the complicated inlet velocity pro-
file and turbulent flow produced by regions up-
stream will propagate into the TB tree. The effect
of the laryngeal jet on intra-tracheal particle de-
position has been studied experimentally and nu-
merically by several researchers.(2,1719) These
studies showed that neglecting upstream condi-
tions would underestimate the particle deposi-
tion in the upper region of the trachea. Further-
more, Stapleton et al.(20) found that at moderate
to high inhalation rates, the flow in the larynx and
downstream was turbulent, which could be ex-
pected to cause complex aerosol transport and
particle deposition within the trachea and se-
quentially bifurcating airways. Therefore, the up-
stream flow characteristics of the trachea (i.e. the
mouth-throat) should not be ignored in experi-
ments relating to fluid flow patterns and aerosol
deposition within the human proximal lung bi-
furcations (defined as the conducting airways
from the lower trachea to the end of the first sev-
eral generations of the lung).
Following advances in computational fluid dy-
namics (CFD), many numerical simulations of
461
particle deposition in the different generations of
the lung have been performed, for example, by
Oldham et al.,(21) in generations G3G5 of the
lung; Comer et al.(22) in double bifurcating airway
models for both in-plane and 90 out-of-plane
configurations; Zhang et al.(23) for the effects of
curved inlet tubes; Zhang and Kleinstreuer(24)
for the effect of particle inlet distributions; and
Zhang et al.(25) for the effect of particle release po-
sition. These theoretical and numerical studies in-
dicated that many factors, including branching
angle, branching geometry, inlet velocity profile,
secondary flow, and even inlet particle distribu-
tion, are important for predicting particle depo-
sition. However, the experimental results have
shown that deposition efficiency is solely depen-
dent on Stokes number even in a wide variation
of branching angle and diameter ratio of daugh-
ter airway to parent airway.(7,8,14,26) It is not clear
what causes these differences between experi-
mental and theoretical results. Even within
the experimental studies, discrepancies are also
found, possibly due to differences in methods
used to produce the models,(1) test aerosols and
procedures to determine particle deposition.
Therefore, our understanding of aerosol deposi-
tion in the human respiratory tract is far from
complete.
In the present work, we investigated particle
deposition within several different idealized
proximal lung bifurcation geometries using iden-
tical measurement and model fabrication tech-
niques, so that it is possible to minimize the ef-
fects of these factors on the experimental results.
Monodisperse particles deposited in three ideal-
ized proximal lung bifurcation models contain-
ing an idealized mouth-throat geometry were
measured for a range of particle sizes and steady
flow rates. The major factors that influence par-
ticle deposition efficiency in bifurcating airways
were determined. The goal of this study is to ex-
amine the effect of upstream flow (mouth-throat)
on particle deposition in the proximal lung bi-
furcations and variation of particle deposition
with different configurations of the bifurcations.
In addition, a relatively higher flow rate of 90
L/min is included, which is seldom used in pre-
vious studies. Information from this study will be
useful for understanding of aerosol deposition
mechanism within bifurcation regions and de-
veloping physical models for particle deposition
prediction.
5967_09_p460-473 12/6/05 1:28 PM Page 462
462
MATERIALS AND METHODS
Physical models and manufacture
Following advances in measuring techniques,
more and more details of the human airways
have been discriminated. As a result, a number of
human lung models have been proposed by dif-
ferent researchers. The symmetric lung model
(Weibel A) proposed by Weibel(27) has been used
widely in experiments and numerical modeling.
According to this model, the human lung is as-
sumed to be a dichotomous branching system
with 24 generations (G0G23). The conductive
airways include generations G0G16, while gen-
erations G17G19 (respiratory bronchioles) and
generations G20G23 (alveolar) belong to the tran-
sitory zone and alveoli zone. Horsfield et al.(28,29)
proposed another asymmetric model for the tra-
cheobronchial tree, which adopts an asymmetric
branching pattern and so gives a more realistic re-
semblance to actual human airways. Recently,
based on the analysis of morphometric data avail-
able from the literature, Finlay et al.(30) proposed
a new lung geometric model with relatively larger
diameter for the tracheo-bronchial generations.
This model was created using the airway data of
Phillips et al.(31) and the alveolar data given by
Haefeli-Bleuer and Weibel.(32) A symmetrically
branching pattern of lung is used, denoting the
trachea as generation G0, the conducting airways
as generations G0G14, and the alveolar airways
as generations G15G23. The major differences be-
tween the WB and FB model lie in the different
diameter and lengths of the TB generations.
In this study, we chose to create three idealized
proximal lung bifurcation models based on the
data proposed in these papers and some direct ob-
servations of computed tomography (CT) scans
of human subjects. The first model, namely,
WB, conforms quite closely to Weibels model
(Weibel A),(27) and the basic configuration con-
TABLE 1. DIMENSIONS OF EACH GENERATION
WB
Generation Length (mm) Diameter (mm)
0 120 16
1 47.6 12.2 36 (right), 48 (left)
2 19 8.3
3 7.6 5.6
ZHANG AND FINLAY
sists of symmetric bifurcations with a branching
angle of 70. The second model, called IB, uses
the same asymmetrically branching patterns as
the Horsfields model,(28,29) except for some mod-
ifications on the branching angle and outlet di-
ameters based on our direct observations from
CT scans. The third model, FB, completely fol-
lows the dimensions proposed by Finlay et al.(30)
with the branching angle of 70. The primary geo-
metric dimensions of all three models are sum-
marized in Table 1. A front view of the three
bifurcation models including an idealized mouth-
throat is shown in Figure 1. Each airway model
is divided into four sub-regions: the mouth-throat
(including the oral cavity, pharynx and larynx),
the trachea, the lung bifurcations (representing
the next two generations of the human lung, that
is, G1 and G2), and the distal bifurcation (G3) and
filter adapters. Airways beyond the generation
G3 have diameter less than 5 mm so that it is
difficult to produce solid wall pieces within this
size range accurately using our stereo-lithogra-
phy machine. In order to generate an outlet ve-
locity profile for each branch similar to that seen
in an actual lung bifurcation, distal bifurcations
are fabricated following the configuration of the
corresponding generation G3.
An idealized biological mouth-throat geometry
(the Alberta geometry) originally developed by
Stapleton et al.(20) is adopted here, and a detailed
description of the design of this geometry can be
found in several previous studies.(20,33,34) For sim-
plicity, a fixed dimensional larynx is used for this
mouth-throat. The cross-section of the larynx is
assumed to be an ellipse with a major axis of 1.2
cm and a minor axis of 1.0 cm (with cross-sec-
tional area of 95 mm2), which is about half-way
between the maximum openings seen during in-
halation and exhalation.(20) This idealized bio-
logical mouth-throat geometry has been verified
to be a good representation of realistic human
mouth throats.(3537) A straight tube with inner
OF THE LUNG BIFURCATION MODELS
IB FB
Length (mm) Diameter (mm) Length (mm) Diameter (mm)
120 16 124.56 18.1
12 36.14 14.14
20, 25 7.5, 5.5, 8, 8, 7.5 28.62 11.15
10 5, 5.5 22.81 8.85
5967_09_p460-473 12/6/05 1:28 PM Page 463

PARTICLE DEPOSITION IN PROXIMAL LUNG BIFURCATION MODELS 463

FIG. 1. Schematics of the three proximal lung bifurcation models with an idealized mouth-throat.

diameter of 17.3 mm and length of 1000 mm is Aerosol generation and particle deposition
used as the inlet to the oral cavity. Due to the measurements
highly curved three-dimensional surfaces of the
Monodisperse aerosol particles of DEHS (di-2-
airway geometries, the proximal lung bifurcation
ethylhexyl sebacate) oil were generated using a
and the idealized mouth-throat geometry used
controlled heterogeneous condensation Aerosol
here were designed using a CAD software pack-
Generator (CMAG, model 3475, Topas, Ger-
age Pro/Engineer (PTC, Needham, MA). The
many). Particle sizes with mass median diame-
model was designed to consist of several com-
ters of 2.57.5
m (SD
1.1) were used in this
ponents (i.e., the mouth-throat, the trachea, the
study. The size and dispersity of aerosol particles
proximal lung bifurcations and the distal bifur-
were monitored using a Mach II Aerosizer (TSI,
cation). Each component is made in two halves
St. Paul, MN). In order to acquire data rapidly, a
split along the central plane for each sub-region.
gravimetric method was used to determine total
In order to mimic a realistic shape of the carina,
and regional deposition. In this method, each
the sharp carinal ridges were rounded smoothly
component and filter was simply weighed before
with curved surfaces.
and after aerosol testing and the collected mass
The solid hollow models were produced using
was determined as the net mass of particle de-
Stereo-Lithography (model FDM 8000, Stratasys,
position within the specific part. This methodol-
Eden Prairie, MN), a three-dimensional printing
ogy was cross-checked with data obtained by
process that produces solid models in acryloni-
gamma scintigraphy and indicated good agree-
trilebutadiene styrene (ABS) plastic. The resultant
ment.(37)
models surfaces were coated with an airtight
epoxy layer, which was found to have high chem-
Experimental procedure
ical and physical resistivity.(38) The fluorocarbon
coating EGC-1700 (3M, St. Paul, MN) was applied A schematic of the experimental setup is shown
on the inner walls of the entire models to increase in Figure 2. The experimental system consists of
liquid surface contact angle in order to reduce the aerosol generator, the mixing chamber, the
surface wetting, which otherwise can decrease airway model, the filters, flow-meters with ad-
the accuracy of deposition measurements.(37) justment valves and a vacuum pump. To avoid
Since DEHS oil particles are liquid, no additional undesired effects of dense aerosol on particle de-
liquid coating is necessary to prevent bounce on position,(6) the original aerosol produced by the
the inner surface of model. aerosol generator is diluted by extra ambient air
5967_09_p460-473 12/6/05 1:28 PM Page 464

464 ZHANG AND FINLAY

FIG. 2. Experimental setup.

in the mixing chamber. The connections of dif-
ferent components are wrapped with parafilm to
create an airtight seal. Particles passing through
the airway models are collected by Marquest
Respirgard filters (Marquest Medical Products,
Inc., Boulder, CO). Two filters are connected se-
rially at each outlet of the bifurcations to avoid
possible leakage of the oil deposited on the filter
film. Three steady flow rates of 30, 60, and 90
L/min are used in the present study. For the sym-
metric models, even flow distribution in each
branch of generation G2 is used (i.e., 25% of the
total flow rate for each branch). For the asym-
metric model, the tracheal flow is distributed
among the five lobar bronchi according to the
data reported by Horsfield et al.(29) The exact per-
centages of flow rate through each lobar branch
are listed in Table 2.
Regarding flow patterns, whether a flow is
laminar or turbulent is governed by the ratio of
the inertial force of the fluid to the viscous force
of the fluid. This ratio is generally characterized
by the Reynolds number (Re), which is usually
defined as(39):
UD 4Q
Re     (1)


D
where U is a flow velocity, Q is the flow rate,

is the fluid dynamic viscosity, and D is a char-
acteristic length representing the geometry of
the flow (typically the diameter of the internal
channel). For flow through straight circular
tubes, the flow is considered laminar when Re is
less than 2300, turbulent when Re is great than
4000, and transitional between those values.(39)
These Re values are different for flow in differ-
ent geometries and vary with upstream turbu-
lence level.
When fluid flow changes direction suddenly,
particles carried by the fluid may be collected on
the surface by inertial impaction. The ratio of the
stopping distance to a characteristic dimension is
defined as the Stokes number (Stk). Stokes num-
ber is a measure of the importance of inertial ef-

TABLE 2. DISTRIBUTION OF FLOW RATE IN EACH BRANCH OF THE BIFURCATION MODELS

Right upper Right middle Right lower Left upper Left lower Total
Model (%) (%) (%) (%) (%) (%)

Data from the literature(29) 21 9 25 20 25 100

IB 20 10 25 20 25 100
WB 25 25 25 25 100
FB 25 25 25 25 100
5967_09_p460-473 12/6/05 1:28 PM Page 465
PARTICLE DEPOSITION IN PROXIMAL LUNG BIFURCATION MODELS
fects in determining particle trajectories. Usually,
the Stokes number is given by(39):
Ccpdp2U
Stk   (2)
18
D
where Cc is the Cunningham slip correction. For
larger particles, the slip correction factor is very
close to unity and varies within a small range
(1.031.07) for particles with diameters of 2.57.5

m, and thus can be neglected in formula (2) for
our purposes. Here, p is the particle density, dp
is the particle diameter, and
is the dynamic vis-
cosity of fluid. U is the velocity obtained by di-
viding the flow rate by the cross sectional area of
the trachea, and D is the characteristic length.
In particle deposition studies, the inertial para-
meter dp2Q (g
m2s1) is another convenient pa-
rameter for normalizing impaction dominant de-
position data when flow rate and particle density
are different, where  (g/cm3 ) is the particle den-
sity, dp (
m) is the particle diameter and Q (cm/sec)
is the flow rate. Although the inertial parameter
does not take into account the inter-subject vari-
ability of the geometries of human airways, it is
still widely used for presentations of particle de-
position data, especially for in vivo data, due to
the difficulty of determining an adequate charac-
teristic length for the realistic human airways.
The weight of each component is determined by
an analytical balance (Sartorius 1207MP2, Ger-
many) with a resolution of 0.0001 g. The measured
increase in weight is used to determine the mass of
particles deposited within the correspondent parts.
Generation G3 is not included in the presentation
of particle deposition in order to avoid the possible
effect of flow disturbances from the open ends of
terminal airways on the particle deposition. Since
the components of each model could be disassem-
bled into individual parts and weighed separately,
the total and regional deposition could be obtained
simultaneously from one test.
The total deposition efficiency (t), the absolute
deposition efficiency (a), and the relative depo-
sition efficiency (r) are given by:
mp
t (%)    100 (3)
mp  mf
mr
a (%)    100 (4)
mp  mf
mr
r (%)    100 (5)
min
465
where mp is the mass of particles deposited in the
all parts of the solid walled models, mr is the
mass of particles deposited in the specific region,
mf is the mass of particles collected in the down-
stream parts (including filters and adapter), and
min is the mass of particles entering the specific
region.
RESULTS
Total particle deposition in three airway models
A summary of total deposition efficiency as a
function of an inertial parameter dp2Q (g
m2s1)
in three idealized airway bifurcation models is
plotted in Figure 3. As indicated, total deposition
efficiency (t) increases with increasing values of
the inertial parameter dp2Q (g
m2s1) in the
three models, and differences in particle deposi-
tion within the three models are not significant
over the range of inertial parameter. The present
work, as well as previous publications,(711) sug-
gests that for the conducting airways up to the
end of the generation G2 of the lung, inertial de-
position is the dominant deposition mechanism.
Figure 3 shows that total particle deposition effi-
ciency is not sensitive to differences in the geome-
tries used here.
Regional particle deposition in the mouth-throat
Particle deposition efficiency in the mouth-
throat from the above tests is plotted in Figure 4.
Due to the high convection speed in the laryngeal
opening into trachea, it is supposed that differ-
ences of the flow patterns downstream caused by
variations of bifurcating pattern have negligible
effect on upstream particle deposition. As shown,
there are virtually no differences in particle de-
position among any of the models studied, de-
spite distinguishable geometric differences in se-
quentially bifurcating airways for each model.
This result demonstrates that, as expected, the
flow characteristics downstream of the mouth-
throat cause no effect on total particle deposition
in the mouth-throat.
Regional particle deposition in the trachea
Airflow in the trachea is complicated due to the
effects of complex flow and geometry upstream
and downstream. Absolute deposition efficiency
(a) in the trachea as a function of the inertial pa-
5967_09_p460-473 12/6/05 1:28 PM Page 466

466 ZHANG AND FINLAY

FIG. 3. Total deposition efficiency in the three models versus inertial parameter. Each point represents the average
of three repeats and error bars refer to standard deviation.

rameter dp2Q is shown in Figure 5. Particle de-
position efficiency increases with the inertial pa-
rameter at a slower rate than in the mouth-throat.
A peak exists for tracheal deposition within each
model. The peak may be explained as follows:
with increases of the inertial parameter, enhanced
particle deposition in regions upstream (i.e., the
mouth-throat) will greatly reduce the absolute
amount of particles entering the trachea. This will
partially counteract the enhanced quantity of par-
ticles deposition in the downstream airways by
inertial impaction with increasing inertial para-
meter, so that the net effect is decreased deposi-
tion with increasing inertial parameter.
Relative deposition efficiency (r) for particles
entering the trachea is plotted as a function of
Stokes number (Stk) in Figure 6 with the diameter
of the trachea as the characteristic length. It can be
seen that deposition efficiency (r) values increase
monotonically with increasing Stokes number,
and particle deposition efficiency in the FB model
is higher than that in either WB or IB models.

FIG. 4. Deposition efficiency in the mouth-throat versus inertial parameter. Each point represents the average of
three repeats, and error bars refer to standard deviation.
5967_09_p460-473 12/6/05 1:28 PM Page 467

PARTICLE DEPOSITION IN PROXIMAL LUNG BIFURCATION MODELS 467

FIG. 5. Deposition efficiency in the trachea versus inertial parameter. Each point represents the average of three re-
peats, and error bars refer to standard deviation.

Two major factors influence particle deposition
in the trachea, namely, local flow pattern and in-
let velocity profile. Due to the significantly larger
cross section area of the trachea (i.e., 257 mm2 for
FB model) compared with the constriction of the
larynx (95 mm2), the magnitude of the inlet ve-
locity to the trachea is much larger than the mean
velocity in the trachea. Therefore, it may be more
reasonable to use a dimension upstream as a char-
acteristic length to calculate the Stokes number.
As shown in Figure 7, the relative deposition ef-
ficiency (r) for particles entering the trachea is
plotted as a function of Stokes number (Stk) again,
and the characteristic length used here is the hy-
draulic diameter of the glottis Dh (10.887 mm),
which is given by
4S
Dh   (6)
L
where S is the cross sectional area of the glottal
aperture, and L is the perimeter of the cross sec-
tion. As shown, deposition data falls within a

FIG. 6. Deposition efficiency for particles entering the trachea versus Stokes number. Each point represents the av-
erage of three repeats, and error bars refer to standard deviation.
5967_09_p460-473 12/6/05 1:28 PM Page 468

468 ZHANG AND FINLAY

FIG. 7. Deposition efficiency for particles entering the trachea versus Stokes number. Each point represents the av-
erage of three repeats, and error bars refer to standard deviation.

more narrow data range and a linear increase of
deposition with increasing Stokes number is seen.
Since the identical mouth-throat is used for these
three bifurcation models, the single collapsed
curve suggests that the upstream laryngeal con-
dition is the dominant factor in determining the
particle deposition in the trachea, and not the lo-
cal dimensions of the trachea.
As seen in the experiment by Grgic et al.,(36) an
effect of Reynolds number on deposition could
be present. As shown in Figure 8, tracheal depo-
sition fraction in the FB model is plotted for the
three Reynolds number of 2276, 4553, and 6830,
corresponding to flow rates of 30, 60, and 90
L/min, respectively, and the characteristic length
used here is the diameter of trachea. A slight in-
crease of deposition efficiency is observed with
increasing Reynolds number. This may be due to
a steepening of velocity profiles near the wall
with increased Reynolds number leading to in-

FIG. 8. Deposition efficiency in the trachea (FB model) with the different Reynolds numbers versus Stokes number.
Each point represents the average of three repeats, and error bars refer to standard deviation.
5967_09_p460-473 12/6/05 1:28 PM Page 469
PARTICLE DEPOSITION IN PROXIMAL LUNG BIFURCATION MODELS
creased impaction velocities near walls.(34) Alter-
natively, enhanced turbulent mixing could play
a partial role in promoting particle deposition at
higher Reynolds number.
Previous deposition tests in an idealized mouth-
throat model have indicated a similar Reynolds
number effect on deposition.(36) Following the
empirical Reynolds number correction suggested
by Grgic et al.,(36) Stokes number (Stk) is multi-
plied by Reynolds number (Re) to the power of
0.37. With this correction, deposition data and a
function obtained by a least squares curve fit are
shown in Figure 9. It is found that particle depo-
sition in the trachea collapses more closely onto
a single curve. (Note that Re in Figure 9 refers
to the flow Reynolds number. Particle Reynolds
number [Rep] would of course be 1 and is in
the usual Stokes flow regime.)
Regional particle deposition in the proximal
lung bifurcations
Absolute and relative particle deposition ef-
ficiencies in the lung bifurcations (G1 and G2)
are shown in Figures 10 and 11. Deposition
efficiencies are plotted as a function of both in-
ertial parameter (dp2Q) and Stokes number
(Stk), where the diameter of the trachea is used
as the characteristic length. The results indicate
that absolute deposition efficiency in the FB
model is lower than that in either the WB model
FIG. 9. The data of Figure 8 is replotted with x-axis, indicating a Reynolds number correction. Each point repre-
sents the average of three repeats, and error bars refer to standard deviation.
469
or the IB model. This is due to the larger cross
sectional area of the FB model, which reduces
the magnitude of the velocity, and decreases
the probability of particle deposition by iner-
tial impaction. However, the plot with Stokes
number in Figure 11 shows that the dimension-
less deposition data collapses close to a single
curve.
Gas transport under the combined influence of
jet induced turbulence and oscillatory flow have
been examined experimentally by Kamm et al.(40)
They divided the region influenced by the jet
into two zones: a near field and a far field. In the
near field, the jet is treated as a confined jet. In
the far field, axial mass transport is increased
due to the turbulence, which decays in strength
away from the jet. The experiments showed that
the presence of a turbulence jet led to gas trans-
port enhancement within a region extending up
to 20 tube diameters. Recently, Corcoran and
Chigier(41) measured the flow field in the tracheal
sections of a cadaver based larynx-trachea model
using Phase Doppler Interferometry and ob-
served high levels of turbulence intensity near
the anterior tracheal walls within one diameter
downstream of the larynx. These turbulence lev-
els were still significant up to four diameters
downstream. However, it is not clear whether
this effect of the laryngeal jet continues into the
bifurcation downstream and affects particle de-
position there. In order to investigate this ef-
5967_09_p460-473 12/6/05 1:28 PM Page 470

470 ZHANG AND FINLAY

FIG. 10. Deposition efficiency for particles entering the proximal lung bifurcation as a function of inertial parame-
ter. Each point represents the average of three repeats, and error bars refer to standard deviation.

fect, particle deposition in the FB model with DISCUSSION

and without the mouth-throat is compared. As
shown in Figure 12, the difference between par-
ticle deposition in the bifurcations between the
two cases is minor. Therefore, at a constant flow
rate, the effect of the laryngeal jet seemingly does
not propagate into bifurcations downstream or
at least, does not significantly affect particle de-
position there.
Since a major test of in vitro data in physical
models is to obtain deposition data similar to data
from realistic human subjects, it is useful to com-
pare our experimental data directly with in vivo
data. However, in vivo deposition data in the first
several generations of the lung is sparse. Although
in vivo particle deposition data in the TB tree is

FIG. 11. Deposition efficiency for particles entering the proximal lung bifurcation versus Stokes number. Each point
represents the average of three repeats, and error bars refer to standard deviation.
5967_09_p460-473 12/6/05 1:28 PM Page 471

PARTICLE DEPOSITION IN PROXIMAL LUNG BIFURCATION MODELS 471

FIG. 12. Effect of the laryngeal jet on deposition for particles entering the proximal lung bifurcations. Each point
represents the average of three repeats, and error bars refer to standard deviation.

available, no data is present for each generation
of the lung due to the resolution limitations of ra-
dionuclide imaging method. However, there are
still several experimental data sets suitable for
comparison. In Figure 13, the tracheo-bronchial
tree deposition of ferric oxide aerosols obtained
from in vivo human tests(2,13,42) and polystyrene
latex fluorescent particles in a human airway cast
with mouth-throat(16) is plotted as a function of
inertial parameter. Note that data in Figure 13 is
relative deposition (see Equation 5) and so allows
comparison of our data to that shown for the en-
tire TB tree. Significant scatter can be seen in the
in vivo data. Possible reasons for the scatter in-
clude differences in breath pattern and variability
of airway geometry. It can be seen from Figure 13
that the present deposition data lies within the
range of several sets of in vivo data.

FIG. 13. Comparison of the present data with tracheo-bronchial deposition in humans in vivo(2,13,42) and in a human
airway cast.(16)
5967_09_p460-473 12/6/05 1:28 PM Page 472
472
Alternatively, a number of empirical correla-
tions are available for comparison of deposition
within the bifurcations. Among these correla-
tions, only the equations of Chan and Lipp-
mann(2) and ICRP(3) would be expected to repre-
sent typical values of deposition in an actual lung
since they were obtained from the casts of air-
ways that included the effects of a larynx and
multiple bifurcations. The predictions of deposi-
tion in generation G1 plus G2 based on the above
two correlations are plotted in Figure 11, where
the Stokes number is calculated based on the di-
mensions of the FB model. As indicated in Fig-
ure 11, the present data in the three proximal lung
bifurcation models is lower than that obtained
from previous empirical correlations. A possible
explanation for this discrepancy lies in the fact
that morphological features on the wall of the hu-
man airway casts could change the flow patterns
in the bifurcations and influence deposition, but
these morphological features are neglected in the
present idealized models.
CONCLUSION
Particle deposition in three idealized proximal
lung bifurcation models with an idealized mouth-
throat has been investigated for a range of typi-
cal flow rates and particle sizes of interest in in-
haled pharmaceutical aerosol applications. The
results demonstrate that particles entering the
mouth are mainly intercepted by the mouth-
throat and trachea with increasing both flow rate
and particle size. The laryngeal jet is the key fac-
tor dominating particle deposition within the tra-
chea. Slightly enhanced particle deposition in the
trachea is observed with increasing Reynolds
number. Particle deposition fraction within the
proximal lung bifurcations G1 and G2 collapses
nearly onto a single curve for all three geometries
when a proper Stokes number is used, and de-
position efficiency in the proximal lung bifurca-
tions is not significantly influenced by the laryn-
geal jet upstream, at least in this configuration of
the idealized mouth-throat geometry.
ACKNOWLEDGMENTS
The financial support of the Alberta Ingenuity
Fund and NSERC is gratefully acknowledged.
ZHANG AND FINLAY
REFERENCES
1. Kelly, J.T., B. Asgharian, J.S. Kimbell, et al. 2004. Par-
ticle deposition in human nasal airway replicas man-
ufactured by different methods. Part I. Inertial regime
particles. Aerosol Sci. Technol. 38:10631071.
2. Chan, T.L., and M. Lippmann, 1980. Experimental
measurements and empirical modeling of the re-
gional deposition of inhaled particles in humans. Am.
Ind. Hyg. Assoc. J. 41:399409.
3. ICRP. 1994. Human respiratory tract model for radio-
logical protection. pp. 4448; 231299. In Annals of the
ICRP. ICRP Publication 66. Elsevier, New York.
4. Emmett, P.C., R.J. Aitken, and W.J. Hannan. 1982.
Measurements of the total and regional deposition of
inhaled particles in the human respiratory tract. J.
Aerosol Sci. 13:549560.
5. Foord, N., A. Black, and M. Walsh. 1978. Regional de-
position of 2.57.5
m diameter inhaled particles in
healthy male nonsmokers. J. Aerosol Sci. 9:343357.
6. Finlay, W.H. 2001. The Mechanics of Inhaled Phar-
maceutical Aerosols: An Introduction. Academic
Press, London.
7. Kim, C.S., and A.J. Iglesias. 1989. Deposition of in-
haled particles in bifurcating airway models: I. inspi-
ratory deposition. J. Aerosol Med. 2:214.
8. Kim, C.S., D.M. Fisher, D.J. Lutz, et al. 1994. Particle
deposition in bifurcating airway models with varying
airway geometry. J. Aerosol Sci. 25:567581.
9. Ferron, G.A. 1977. Deposition of polydisperse aerosols
in two glass models representing the upper human
airways. J. Aerosol Sci. 8:409427.
10. Martonen, T.B. 1983. Measurement of particle dose
distribution in a model of a human larynx and tra-
cheo-bronchial tree. J. Aerosol Sci. 14:1122.
11. Kim, C.S., and D.M. Fisher. 1999. Deposition charac-
teristics of aerosol particles in sequentially bifurcat-
ing airway models. Aerosol Sci. Technol. 31:198220.
12. Schlesinger, R.B., and M. Lippmann. 1972. Particle de-
position in casts of the human upper tracheobronchial
tree. Am. Ind. Hyg. Assoc. J. 33:237251.
13. Schlesinger, R.B., and M. Lippmann. 1976. Particle de-
position in the trachea: in vivo and in hollow casts.
Thorax 31:678684.
14. Schlesinger, R.B., D.E. Bohning, T.L. Chan, et al. 1977.
Particle deposition in a hollow cast of the human tra-
cheobronchial tree. J. Aerosol Sci. 8:429445.
15. Schlesinger, R.B., J.L. Gurman, and M. Lippmann.
1982. Particle deposition within bronchial airways:
comparisons using constant and cyclic inspiratory
flows. Ann. Occup. Hyg. 26:4764.
16. Cheng, Y.S., Y. Zhou, and B.T. Chen. 1999. Particle de-
position in a cast of human oral airways. Aerosol Sci.
Technol. 31:286300.
17. Scherer, P.W., L.H. Schendalman, N.M. Greene, et al.
1975. Measurement of axial diffusivities in a model of
the bronchial airways. J. Appl. Physiol. 38:719723.
18. Katz, I.M., and T.B. Martonen. 1996. Flow patterns in
three-dimensional laryngeal models. J. Aerosol Med.
9:501511.
5967_09_p460-473 12/6/05 1:28 PM Page 473
PARTICLE DEPOSITION IN PROXIMAL LUNG BIFURCATION MODELS
19. Katz, I.M., B.M. Davis, and T.B. Martonen. 1999. A
numerical study of particle motion within the human
larynx and trachea. J. Aerosol Sci. 30:173183.
20. Stapleton, K.W., E. Guentsch, M.K. Hoskinson, et al.
2000. On the suitability of k- turbulence modeling
for aerosol deposition in the mouth and throat: a
comparison with experiment. J. Aerosol Sci. 31(6):739
749.
21. Oldham, M.J., R.F. Phalen, and T. Heistracher. 2000.
Computational fluid dynamic predictions and exper-
imental results for particle deposition in an airway
model. Aerosol Sci. Technol. 32:6171.
22. Comer, J.K., C. Kleinstreuer, and S. Hyun. 2000. Aero-
sol transport and deposition in sequentially bifurcat-
ing airways. ASME J. Biomech. Eng. 122:152158.
23. Zhang, Z., C. Kleinstreuer, and C.S. Kim. 2001. Effects
of curved inlet tubes on air flow and particle deposi-
tion in bifurcating lung models. J. Biomech. 34:659669.
24. Zhang, Z., and C. Kleinstreuer. 2001. Effect of parti-
cle inlet distributions on deposition in a triple bifur-
cation lung airway model. J. Aerosol Med. 14:1329.
25. Zhang, Z., C. Kleinstreuer, and S.K. Chong. 2002.
Aerosol deposition efficiencies and upstream release
positions for different inhalation modes in an upper
bronchial airway model. Aerosol Sci. Technol. 36:828844.
26. Chan, T.L., and R.M. Schreck. 1980. Effect of the la-
ryngeal jet on particle deposition in the human tra-
chea and upper bronchial airways. J. Aerosol Sci. 11:
447459.
27. Weibel, E.R. 1963. Morphometry of the Human Lung.
Academic Press, New York.
28. Horsfield, K., and G. Cumming. 1968. Morphology of
the bronchial tree in man. J. Appl. Physiol. 24:373383.
29. Horsfield, K., G. Dart, and D.E. Olson. 1971. Models
of the human bronchial tree. J. Appl. Physiol. 31:207
217.
30. Finlay, W.H., C.F. Lange, M. King, et al. 2000. Lung
delivery of aerosolized dextran. Am. J. Respir. Crit.
Care Med. 161:9197.
31. Phillips, C.G., S.R. Kaye, and R.C. Schroter. 1994. A
diameter-based reconstruction of the branching pat-
tern of the human bronchial tree. Part I. Description
and application. Respir. Physiol. 98:193217.
32. Haefeli-Bleuer, B., and E.R. Weibel. 1988. Morphometry
of the human pulmonary acinus. Anat. Rec. 220:401
414.
33. Brancatisano, T., P.W. Collett, and L.A. Engel. 1983.
Respiratory movements of the vocal cords. J. Appl.
Physiol. 54:12691276.
473
34. Heenan, A.F., E.A. Matida, A. Pollard, et al. 2003. Ex-
perimental measurements and computational model-
ing of the flow in an idealized extrathoracic airway.
Exp. Fluids 35:7084.
35. DeHaan, W.H., and W.H. Finlay. 2001. In vitro mono-
disperse aerosol deposition in a mouth and throat
with six different inhalation devices. J. Aerosol Med.
14:361367.
36. Grgic, B., W.H. Finlay, P.K.P. Burnell, et al. 2004. In
vitro inter-subject and intra-subject deposition mea-
surements in realistic mouth-throat geometries. J.
Aerosol Sci. 35:10251040.
37. Grgic, B., W.H. Finlay, and A.F. Heenan. 2004. Re-
gional aerosol deposition and flow measurements in
an idealized mouth and throat. J. Aerosol Sci. 35:2132.
38. DeHaan, W.H., and W.H. Finlay. 2004. Predicting ex-
trathoracic deposition from dry powder inhalers. J.
Aerosol Sci. 35:309331.
39. Willeke, K., and P.A. Baron. 1993. Aerosol Measure-
ment. Van Nostrand Reinhold, New York.
40. Kamm, R.D., E.T. Bullister, and C. Keramids. 1986.
The effect of a turbulent jet on gas transport during
oscillatory flow. ASME J. Biomech. Eng. 108:266272.
41. Corcoran, T.E., and N. Chigier. 2000. Characterization
of the laryngeal jet using phase doppler interferome-
try. J. Aerosol Med. 13:125137.
42. Stahlhofen, W., J. Gebhart, and J. Heyder. 1980. Ex-
perimental determination of the regional deposition
of aerosol particles in the human respiratory tract.
Am. Ind. Hyg. Assoc. J. 41:385398a.
Received on January 24, 2005
in final form, June 3, 2005
Reviewed by:
George A. Ferron, Ph.D.
Akira Tsuda, Ph.D.
Address reprint requests to:
Dr. Warren H. Finlay, Ph.D., P.Eng
4-9 Mechanical Engineering Building
University of Alberta
Edmonton, Alberta, Canada, T6G 2G8
E-mail: warren.finlay@ualberta.ca