J. [Link].

,23, 657-678 (September14, 1972)

The Movement of Aerosol Particles

WILLIAM LICHT, Ph.D.*

PresentedDecember13, 1971, New York City

Synopsis--TheMOVEMENT of an AEROSOL PARTICLE, subsequentto the initial VE-

LOCITY imparted by its source,is governed primarily by the influence of GRAVITY,
DRAG, INERTIA, DIFFUSION, and ELECTROSTATIC CHARGES. The combined effect
of these forces determines the path a particle follows. If this path brings it sufficiently
near to a surface,it may collide with that surfaceand be depositedthereon. Of special
interesthere is the possibledepositionupon surfacesof the body, the hair, and the respira-
tory tract.
Each of theseforcesmay be characterizedby a dimensionless parameter which is readily
definedand calculated. From an examinationof the numerical valuesof theseparameters,
the relative importance of each factor may be judged for a given case. The likelihood of a
particlebeingdepositedupon a givensurfacemay alsobe estimated.
The critical variableswhich control the movement are found to be the sizeand density of
the individual particle, its initial velocity,and the velocityof the surroundingair. These
may be adjusted,within limits, so as to tend to promote desired depositionof the particles,
or to prevent an unwanted deposition.

GENERAL CONCEPTS

The widespreaduseof aerosoldispensers for a variety of cosmetic,

pharmaceutical,and related productscreatesa need for an understand-
ing of how aerosolparticlesmoveand what determineswhere they finally
cometo rest. Sucha particleissues from the atomizerasa sphericaldrop-
let of fixed sizeand density,traveling with a certain initial speedand di-
rection. Dependingupon the intendedpurposeof the product,the user
maywishthe particlesto behavein one of severalpossibleways:

* Department of Chemical and Nuclear Engineering, University of Cincinnati, Cincinnati,

Ohio 45221.

657

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 658 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

(a) to remainsuspended aslong aspossiblein a confinedspace,in or-

der, for example,to producea disinfectingor deodorizingeffect
upon the air;
(b) to collide with, and be retained with maximum efficiencyupon a
particularsurface,as,for example,surfacesof the bodyboth exter-
nal (skin,hair, eyes)andinternal(respiratory tract,throat)in order
to providecosmeticor therapeuticeffects,e.g.,perfume,hair spray,
deodorant,nasaldecongestant, throat spray,etc.;
(c) conversely,to avoid collision with surfacesof the body as in the
caseof an insecticidewhich might causeskin irritation;
(d) to becomeimpactedon the surfacesof instrumentswhich are de-
signedfor the samplingand analysisof aerosols,e.g., microscope
slides,impactorstages.
A studyof the principleswhich governthe movementof aerosolparticles
shouldbe of value in helping to achieveany one of theseobjectives.
The path followedby an aerosolparticle,subsequent to its generation
and re'_ease, dependsnot only upon its initial velocityand upon the ve-
locityof the air streamin which it rides,but alsoupon the actionof cer-
tain external forces. These may be listed as:
(a) drag, the resistanceto motion offeredby the air surroundingthe
particle;
(b) gravity (opposedby a negligibleamountof buoyancy);
(c) inertia, the resistanceoffered by the particle to an attempted
changein speedor direction;
(d) diffusion,due to bombardmentby air moleculesor to turbulence
in the air;
(e) electrostatic,
if the particleacquiresa charge.
Whether the particle remainssuspendedin still or moving air, settles
out, or impingesupon someneighboringsurfacedependsupon its path
or trajectory,asdeterminedby the net effectof whicheverof theseforces
may be acting. After collision upon a neighboring surface,the particle
may reboundor it may remain depositedupon that surface. To predict
the probabilityof depositionof a particleupon a surfacerequiresan un-
derstandingof how eachof theseforcesaffectits path.
Eachof the aboveforcesisreviewedbriefly below,in order to indicate
what is known about its effectupon aerosolmotion, and to indicate how
the relative importanceof eachforcemay be judged in a given setof cir-
[Link] is done basicallyby analyzingthe equationof New-
ton's Law of Motion:

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 MOVEMENT OF AEROSOL PARTICLES 659

massof particle X accelerationof particle =

sum of all forcesacting upon particle

asapplied to an individual particle. Out of this analysisa seriesof pa-

rametersis obtained,each of which is characteristicof one type of be-
havior. By a relativelysimplecalculationof eachparameterand inspec-
tion of its magxfitudea goodidea may be obtainedof the relative impor-
tanceof eachof the severalforces;how this may be doneis illustratedby
a number of examples.

IMPLICATIONS FOR APPLICATIONS

Certain useful conclusionsmay be drawn from a knowledgeof the

forcesgoverningaerosolmotion in a given case. In a specificsituation,a
quicksurveyis firstmadeto determinewhichforces(or modesof deposi-
tion) are most important. This is readily done by calculatingcertain
parameterswhich are developedbelow:
G, •p,Re, dr/D, Pe-•, K•u

and comparing the values obtained with those given in the examples.
Usually one of thesewill be much greater than all of the others, thus
identifyingthe predominantforce. As soonas the relative importance
of the severalforcesis known, a judgment may be made on how to pre-
vent, or how to fosterdeposition,whichever may be desired. This will
be done mainly by adjustingvaluesof the independentvariables: par-
ticle densityor, particle sized•,,air streamvelocityv0, sizeof deposition
surfaceD (alsoR and L), and possiblyelectricalchargeon particle Qv.
The mostimportant propertyof the material of the aerosolis its den-
sity. This, of course,is determinedby its composition,which may be
rather fixed by the nature of the material for a particular application.
All of the illustrative calculationsgiven below have been basedupon a
particledensityof 1 g/cma. For other valuesof density,the resultsgiven
may easilybe correctedaccordingto the way in which p entersinto each
parameteror equation. The effectwill be significantwhenevergravity
(•', us, G) or inertia (a,) is involved,but doesnot enter into diffusional
phenomena(Pe,:o).
The particle size is perhapsthe most important variable which is
within the control of the user of the aerosol. Someguidelinesmay be
deducedfrom the theoryof movementand summarizedas follows:
If it is desiredto maintain an aerosolsuspendedin a free space,the
particlesshouldbe lessthan 50 t* in diameterto minimize settling;

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 660 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

If it is desiredto depositan aerosolupon a free surfaceby impinge-

ment, the particlesshouldbe lessthan 50 • in diameter,but larger than
about 10 t,,in order to take advantageot5inertial effects;
Depositionupon small surfaces,suchas a human hair, will be aided
by directinterceptionif the aerosolparticleis of the sameorderof mag-
nitude as the surface dimension;
There is no advantageto usingvery fine particles,lessthan say1 • in
diameter,unlessit is desiredto havean aerosolpenetratedeeplyinto the
respiratorytract for medicinalpurposes (seebelow);
1t5it is desiredto protectagainstan aerosolenteringthe respiratory
tract, it will help to maintain particlesizegreaterthan 10 • in order to
utilizetheremovalmechanisms (inertia,gravity)in thenasalsystem.
Thesestatements applyto a materialhavinga densityof 1 g/cma. If
the value of p is much different from this, the statementsshould be modi-
fied accordingly.
In all cases
whereit isdesiredto minimizegravitational effects
and/or
to promoteinertial effectsit will be desirableto useaslargean air stream
[Link] shouldbe at leastof the orderot5100cm/sec.
Sucha speedwill alsominimizediffusionaleffects,even for very small
particles.
Depositionin the respiratoryand nasalsystemis ot5specialimpor-
tanceand hasbeenthe subjectot5detailedstudy. Formalizedschematic
representatationsot5the respiratorytract havebeendevisedby Findeisen
(1) and by Landahl (2). Theserepresentthe tract by a seriesof straight
andcurvedtubes,havingbranchingintersections andfinallyterminating
in small spheres. Typical averagelength and diameter dimensionsare
assignedto each different part, as well as a number count ot5each.
Then, from a knowledgeot5typical breathingrates,air speedvelocity
may be calculated. Using this scheme,calculationsmay be madeusing
the theory outlined above to estimatedeposition,or penetration,ot5
particlesot5varioussizesby the variousmechanisms in the severalpor-
tions of the system.
The broad results ot5 such theoretical calculations have been sub-
stantiated by experimental tests. General conclusionshave been sum-
inarizedby Hatch and Gross(g),andmaybe paraphrased asfollows:
1. Particlesgreaterthan 10 • are essentiallyall removedin the nasal
chamber(inertia,sedimentation) anddo not penetrateinto thelungs.
2. Particleslessthan 1 • are essentiallynot removed at all in the
upper respiratorytract.

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 MOVEMENT OF AEROSOL PARTICLES 661

3. The efficiency
of particleremovalis essentially
100% in the pul-
monaryair spaces,
for particlesdown to about 2 v.
4. As particlesizedecreases from 2 v to about0.5 v, removalin pul-
monaryair spaces decreases,but increasesagainfor sizesbelow 0.5 v due
to depositionby diffusion.*
5. Depositionin the upper respiratorytract is increasedwith faster
breathingfrequency,due to increased rate of air flow.
6. Depositiondeepin the lungsincreases with slow,deepbreathing,
asa large fractionof the tidal air getsin and there is a longer time of
transitin and out of the lungs.

DETAILED ANALYSIS

The list of forceswhich may act upon an aerosolparticlewill next be

reviewedin detail in order to show how to calculatethe important
quantitiesassociated with each. This in turn leadsto the ability to
judgewhichforce(or forces)may be controllingthe particlebehaviorin
a given case.
Throughoutthe followingit will be assumed that the aerosolpar-
ticlesare sphericalin shapehavinga diameterof lessthan roughly100v,
andare movingin dry air at 20øCand 1 atm pressure, viscosity• • 1.81
X 10-4 poises,densityp/: 1.20 X 10-a g/cm•.
Drag
Under the conditions assumed, the force of frictional resistance
(drag)offeredby still air, isgivenby Stokeslaw*
FD = 3•'d•,u•
For horizontal motion in still air in the absence of all other forces,
Newton'slaw leadsto thequantityr: pvdv"/18v whichisa basicproperty
of the particle-airsystemcalled the relaxation time. Eventually,the
drageffectreducesthe speedof the particleto zero. The distancetraveled

*•Particlesof 0.I v remaining in an alveolar sacof 0.03-cmdiameter for 1.2 sec(as assumed
by Findeisen)would depositto the extent of 60%, accordingto the diffusional equation cited
above. Particlesof 0.01t• would be completelydeposited.
t For very small particlesFv must be correctedby dividing by the Curmingham factor
C, which is approximately 1 + 0.16/dr in air at 20øC and 1 a•tm, where dvis in microns.
For particlesgenerallylarger than 100-t•diameter,Stokeslaw mustbe replacedby the useof
a drag coefficientrelationshipto expressthe drag force, but such larger particlesare not of
interest here.

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 662 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

horizontallyfrom the startwith initial speedUo,calledthe stoppingdis-

tance,is given by
pvdv2uo
x8 = •u0 .... (1)
18/•

When the air streamis moving,the dragforcewill dependupon the

velocityof the particlerelativeto that of the air. This relativevelocity
mustbe takenin a vectorsensein order to allow for the possibilitythat
the air and the particlemaybe movingin differentdirectionsaswell asat
differentspeeds,
at a particularinstant.
Gravity
Becausethe particlehasweight,there will alwaysbe sometendency
for it to movevertically. When this forceis incorporatedin Newton's
law, in addition to the alwayspresentdrag, the vertical componentof
motion in still air may be obtained. In free vertical fall startingfrom
rest, the particlewill accelerateuntil it reachesa maximum speed,
calledthe terminalvelocity,givenby
us = g• (2)
The combinedeffectof drag (both horizontaland vertical) and
gravitydeterminesthe trajectoryof a particlemoving freely in quiet
air, in the absenceof all other forces. If the particle is given an initial
horizontalvelocityat speedu0,the distancetraveledhorizontallyis given
by
x = xs(• - e-t/•) = u0•(1- e-t/•) (3)
If the vertical speedof the particle is initially zero, the distancetraveled
verticallyis

y=ust--usr(1--e
-t/•)
=usa-f;--(1--e-t/•)]
(4)
The horizontaland verticaldistances are determinedindependently,
and in combinationthey describethe tra,:•[Link] may be done in
generalized
dimensionlessformby expressing thevariablesin ratios:
x x
horizontal distance ratio 1 - •-t/, (5)
Xs I•OT

y y t
vertical distance ratio - (1 - e-t/v) (6)
us• g•2 •
time ratio = t/•'

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 MOVEMENT OF AEROSOL PARTICLES 663

X/xs
0 •) o5 tQ

Figure 1. Generaltrajectoryof particle in still air

Figure 1 showsa plot of this generalizedtrajectory. It is evident that

after a period of time equal to about 5 •, the particle hastraveledalmost
its full stoppingdistancehorizontally,and has practicallyattained its
full free settlingvelocityvertically. Clearly the value of ß is basicto
the effectsof dragand/or gravityupon [Link] determinesthe
values of xs and us.
For illustrative purposestheserelationshipsmay be applied to par-
ticles of practical interest. Considering the material to have p = 1
g/cma, Table I showsthe effectof variouscombinationsof particle size
and initial velocity. For other values of p, each number in the table is
multiplied by •. It is clearthat particlesof about 50 t* and larger will not
remain suspendedlong in air even though proiected initially with a
horizontalvelocityof 100cm/sec.
Table I

Motion in Still Air

Particle Stopping
Distance
(x,,cm) TerminalVertical
Distance
Size Relaxation Velocity Traveled
(d•, t•) Time (r, sec) u0 = 10 cm/secu0 = 100 cm/sec (u, cm/sec) after 1 sec
1ø 3.54 X 10-0 3.54 X 10 -s 3.54 X 10 -4 3.47 X 10 -a 3.47 X 10 -a
10 3.08 X 10 -4 3.08 X 10 -a 3.08 X 10 -2 3.02 X 10 -• 3.02 X 10 -•
50 7.70 X 10 -a 7.70 X 10 -20.770 7.55 7.55
100 3.08 X 10 -20.308 3.08 30.0 29.1

Cunninghamfactoris required (seefootnote,page661).

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 664 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

Gravity Deposition
The movementsdiscussed thus far have been thoseof a particlein
still air. Sufficientlyheavyparticlesmay settleout of a streamof mov-
ing air aswell. The precisetrajectoryof the particlewill dependupon
the velocitypattern,or profile,of the air stream.
The simplestcaseis that of a horizontal air flow with a uniform ve-
locity v0 at all levels,a so-called"plug" or "piston"flow suchas might
occuracrossa largechamberor room. It is assumed that the particleis
riding with the speedof the air stream,i.e., that thereis no slip between
particleand air, henceno drag, sothat ux --- v0 is the horizontalcompo-
nent of particlevelocity. It is further assumedthat the particleis mov-
ing verticallywith a velocitycomponentequal to its terminal settling
velocity,sothat uv= u•. The path of the particleis then a straightline
with a slopeequal to:
uv u8 distancetraveled vertically
ux v0 distancetraveled horizontally, in sameperiod of time
The fraction of aerosolsettlingout from an ini,tial height R, while flow-
ing a horizontaldistanceL, is givenby
u8 L grL
V- voR- voR
- G (7)
If R/L is equal to or lessthan the ratio of u•/vo, 100% removalof the
particlesby sedimentationwill occur.
Shouldthe air streamhavea vertical componentof velocity,alsouni-
formly distributed,this is simplyaddedalgebraicallyto the terminal set-
tling velocity,to give
u•=us+v• (8)
takinga positivevelocityto be downward. If the currentof air is rising
at vv = --u•, the particlewill remain suspendedindefinitelyat the same
level.
For an aerosolflowing with a streamof air inside a tube, the situation
is complicatedby the fact that the axial velocitydependsupon the radial
position. In the caseof laminar flow, there is a parabolicvelocityprofile
symmetricalabout the axisof the tube with a maximum velocity-atthe
center equal to twice the averagevelocity. In the caseof turbulent flow,
a limiting situation("completeturbulence")may be definedrather sim-
ply as that in which the turbulence keepsthe suspendedparticlesdis-
tributed rather uniformly over the whole crosssectionof the tube. All

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 MOVEMENT OF AEROSOL PAR•IICLES 665

Table II

R = 1• cm; L = 5 cm; v0 = 100 cm/sec

Particle
Size Removal
byDeposition
s
(dr, t•) G Plug Flow Laminar Flow Turbulent Flow
1 3.47 X 10
10 3.02 X 10-2 •3% •-•1% •-•2%
50 0.755 •-•75% •-•41% ,--•38%

• Equationsfor thesecalculationsare given by Fuchs (4).

of thesecasesare relatedto the samebasicparameterG --- (u•L/voR)

whichmight be calledthe "gravitydepositionparameter." To illustrate
its effect, consider a set of values: R • « cm, L -- 5 cm, and v0 = 100
cm/sec,for which G = u•/100. Gravity depositionin this casewill be
asgiven in Table II.
Gravity depositionthuswill contribute significantlyto the removal of
particleslarger than 10 g from a flowing stream. In particular, the ex-
ampleconditionschosenrepresentvery roughly the conditionsin a nasal
passage,when the flow would be laminar.
The data in Tables I and II showthat particlesof the order of 10 g in
diametermay easilybe kept suspended for long periodsof time by very
gentle air currents. Thus a spaceinsecticide,for example, should be
dispersedin particlesof lessthan 50 g in diameter for maximum effective-
ness.

Inertia

When a particle is moving with a stream of air, important inertial

effectsmay arise when that streamcarriesthe particle in the neighbor-
hood of an obstaclein its path. The streamof air will changeits direc-
tion in order to flow around the obstacle. But the particle, becauseof its
inertia, may not be able to maintain its positionin the streamlineof air
and thus may be brought into contactwith the obstacle. Suchcollisions
are referred to as"inertial impaction."
Examplesare shownin Fig. 2. Case(a) would representan obstacle
ot•either cylindrical or sphericalshape;case(b) depictsthe situation
aroundeither a diskor a flat strip;case(c) showsa 90ø bend in a pipe or
tube. The streamlinesof the air are representedby the solidlines. The
trajectoryof a particle, causedby its inertia, is representedby a dashed
line. N o effectof gravityis beingconsidered,for the present.
Whether a collisionoccursdependsin part upon the position of the
streamlinewhich the particle was following initially. Along lines such

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 666 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

Vo •.ff (o }
(b}

Figure2. Collisionby inertial impaction

as & collisionoccurs,but along •2 it doesnot. The possibilityof collision

alsodependsin part upon the pattern of the streamlineflow: at high
velocitiesthe'streamlinesdivergesuddenlycloseto the obstacle,while at
low velocitiesthe divergencecommences more graduallyat a greaterdis-
tance upstream. Collision also dependsupon the massof the particle
(e.g.,sizeand density)in additionto the sizeand shapeof the obstacle.
Somedistanceupstreamof the obstacle,it may be reasonableto as-
sumea uniform air velocityfield at v0,and to assumeno slip betweenthe
particlesand the air. However, as soonas the fluid streamlinesbegin to
diverge,they no longer all have the samevelocity. As the particle tra-
jectorydivertsfrom the initial streamline,it beginsto crossfluid stream-
lines. Hence,a drageffectcomesinto play whichnow mustdependupon
the vector differencebetween fluid and particle velocity as mentioned
earlier.
Careful study, both theoreticaland experimental, of this inherently
complexphenomenonhas revealedthat the likelihood of collisionde-
pendsupon two parameters:
voDps
(a) Reynold'snumber,Re - (dimensionless)

whichreflectsthe patternsof the streamlines

andthe effectof air speed;
Xs II oT
(b)Inertial
impaction
parameter,
• - D- D (dimensionless)

(alsocalled Stokesnumber, Stk), which is the ratio of the stoppingdis-

tanceof the particle,calculatedwith an initial velocityequal to that of
the upstreamair velocity,to the appropriatedimensionof the obstacle.
An "efficiencyof collision,"or collectionefficiency,is defined:
cross-sectional
area of fluid streamfrom which particlesare removed
cross-sectional
arearepresented
bydimension
D, projected
upstream

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 MOVEMENT OF AEROSOL PARTICLES 667

In the caseof symmetricalflow about a center line ot5the object, these

areasmaybe representedgeometricallyin a simplemanner. All particles
approachingthe obstaclealong a streamlinewhich is within a distanceb
of the center line will collide; those on streamlines outside this distance
will passby. Thus:
b for a cylindricalobstacle, (Fig. 2,a)
v- D orforaflatstrip (Fig.
2,b) (9)
• = (•)2 foraspherical
adiskobstacle,
orfor (Fig.
2,a) (10)
(Fig.
2,b)
The evidenceindicatesthat ,•dependsonly upon the valuesof _Reand
½,that is:
• = •q(Re,•)
The precisenature of this functionalrelationshiphas been determined
by numerical approximationmethods,and also experimentally,for a
few [Link] (5) foundthat for a fixedvalueof _Re,the relationship
maybe representedby a straightline on "log-normalprobability"graph,
__

plotting ,• on the probability scaleagainstX/q•on the logarithmic scale.

_

The X/½is directly proportionalto the particle size.

Two constantsare needed to place such a straight line. They are
--

usually given as •ul/2, the value of X/q•for which ,• -- 0.50, and • --

(•841/2/•ul/•),calledthe geometricstandarddeviation,wherein •841/•is the
value ol•X/• for which't -- 0.84. Table III summarizes thesenumbers
for variousobstacleshapes,as given by Whitby. It is to be noted that
only for a cylindricalshapedobstaclehasthe effectof Reynoldsnumber
been determined: The graph will be a family of parallel straightlines
with largerRe givinglarger,•. For eachof the other shapesonly an aver-
agepositionof the line may be drawn. Figure $ showssucha graph.
For a fixed Re, the degreeof particle collisiondue to inertia will be
controlledby the ratio void for a given particle size,and may be esti-
mated as follows. From Fig. $, obtain the value of ½ for whatever value
of • maybe of interest. Then vo/D -- q,/r. For example,for 99% colli-
sion by particlesof p _-• 1 g/cma the valuesin Table IV are obtained.
It is evident that sucha high degreeof collisionwill only occurat reason-
ablevelocities(v0)when the obstaclesize(D) is very small.
Severalapplicationsof suchcalculationsto situationsof practicalim-
portancemay be cited. For 99% collectionof an aerosolsampleby im-
pingementur)ona flat surfacesuchasa microscope slide,consideredasa
ribbon with D __--2 cm:

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 668• JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

Table III

Inertial Impaction Parametersfor Various Collector Target Shapes

Collector Shape ½•t1/2as

Jet on plate, round 0. 373 1.24

Jet on plate, rectangular 0.55 1.24

Cylinder
¾---•-•,-
O• ½ 1.253N/•-0. 0.2 < NR, < 150
1.65

•- 0.77 1.91
Sphere ¾o

Ribbon • /9 0.54 2.22

Vo

Trough
or
cup I •• 0.46
1.91

(impinger) ¾o
= ,,

•,.•-- /9 O.54 2.32

Rectangular
half
body ¾o t

Sweeping
bend• 0.71 1.80

Focusing
away ¾o =_ • tQ 0.14 2.9

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 MOVEMENT OF AEROSOL PARTICLES 669

v0 = 1.24 X 107crn/secfor 1-• particles

v0 = 5.72 X 10• cm/sec for 50-• particles
With movementat the lower of thesetwo speeds,for example,q,for the
1-vparticleswould be
vo.r 5.72 X 103 X 3.54 X 10 -6
• - D - 2 = 1.01X 10-•
andv • 0 (fromFig. 3). Hence,if it wereattemptedto collecta sample
of an aerosolcontaininga mixture of thesetwo sizesby impingement
upona slide,the samplecollectedwould not be at all representativeof
the mixture. While this is admittedly an extreme kind of example, it
pointsout vividly that a carefuladjustmentof air flow velocitywould
needto be madein order to approachroughlyequal samplingof all sizes
in a mixture.
Taking the diameterof a humanhair to be 100v, particlesof various
sizesapproachingat variousspeedswould collideas shownin Table V.
This is basedupon consideringthe hair asa long cylinderperpendicular
to the direction of flow, and obtaining q,from Fig. 3.

JET ON PLATE
Round

CYLINDERS
I00
Re = I0
Re= 0.2

0.1 a/2 1 10

,1,

Figure 3. Collectionefficiencyby inertial impaction

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 67O JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

Table IV

99% Inertial Impaction

voiD, Seconds

Particle Cylinder Cylinder

Size (Re = 0.2) (Re = 100)ø Ribbon Round Jet Impinger
(d•,, ix) •b99= 16 •bg•= 9 •b• = 22 •b•9= 0.36 •p• = 9
1 4.52 X 106 2.54 X 106 6.21 X 106 1.02 X 105 1.13 X 10
10 5.19 X 104 2.92 X 104 7.14 X 104 1.17 X 10s 1.30 X 10
50 2.08 X 10 a 1.17 X 10 a 2.86 X 10 a 46.7 5.19 X 10

Alsoappliesto a 90ø bend.

Table V

Impaction on Human Hair

p• = 1.0g/cma; D = 100n; Re = 0.0662v0

v0 = 10 cm/sec Re = 0.66 v0 = 100 cm/secRe = 6.62 v0 = 1000 cm/sec Re = 66.2

1 3.54 X 10-a ,'-•0 3.54 X 10-• ,-¾0 3.54 X

l0 3.08 X 10-' 5% 3.08 •-•80% 30.8 •-•100%
5O 7.70 90% 77.0 •-dO0 % 770 •-d O0%

Table VI

Impaction in 90ø-Bend

1 1.77 X 10 -a
10 1.54 X 10-• -•18%
50 3.85 •96%

It isevidentfrom thishow inertial impactionon the hairsin the nose,for

example,will contributeto a substantialremovalof particlesgreaterthan
10 t, and prevent their being carried into the lungs. Other mechanisms
of removalare alsooperative.
Further removalin the passages of the respiratorytract will alsooccur
by impactionwheneverthe passage bendsor branches. This effectmay
be illustratedby consideringa 90ø-bendin a tube having a diameter of
theorderof thesmallerbronchiof thelungs(say0.20 cm). For a velocity
of 100cm/sec,pp-- 1.0 g/cma, inertial impactionswould occurin accord
with q•z (100 X •-/0.20)• 500 •-. Table VI showsthe valuesasobtained
from Fig. 3.

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 MOVEMENT OF AEROSOL PARTICLES 671

While the effectwill be lessthan this in the caseof bendsof gentler

than 90ø-curvature,or at branching points, the contribution toward the
removalof coarserparticlesis againsignificant.
Whenever impaction or depositionoccursupon a dry surface,there
is a possibilityof particlesreboundingand becomingre-entrainedby the
relativelyhigh speedair streamsinvolved. In this case"impaction" can-
not be synonymous with "removal,"and the listed collectionefficiencies
will be too large. However,on moistor oily surfacesmostparticleswill
be retained. Surfacesused for samplecollection by impingement are
often coatedwith a light grease,or adhesive,for this reason.

Direct In terception
Whenever the sizeof the aerosolparticlesbeginsto approachthat of
the principal dimensionof an obstaclein the stream,collisionsmay oc-
cur to a significantdegreeby direct interception. This processincreases
removalabovethat occurringby inertial impactionand may lead to an
efficiencyof morethan 100%. A limiting casefor which f --> oomay be
imagined,in which the inertia of the particlesis sogreat that they con-
tinue to travel in straight lines. All particles whose center lines are
within the projectedarea of the obstaclemust thereforecollide with it.
But, in addition, all particleswhich are moving on a streamline that is
within a distanceof (dp/2 -+- D/2) of the axis of the obstaclewill also
collidewith it as shownin Fig. 4. The collectionefficiencywill then
be

r/=1d-(•) for
acylinder (1t)
r/ = 1 d- for a sphere

Formulas have been derived from which this added collision effect
may be calculatedfor flow around spheresand cylinders(6). In gen-
eral, the effectbecomesappreciablefor dp/D > 0.1 roughly. This could
frequentlybe the casewhere the obstacleis a human hair, D • 100 •.

Figure 4. Collisionby directinterception

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 672 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

Diffusion
Aerosol particlesare constantlybeing subjectedto numerous colli-
sions,"bombardment,"by moleculesof oxygen,nitrogen, and other
gasesin the air. Generally,the impact of suchbombardmentis negligi-
ble unlessthe aerosolparticleis verysmall. However,under certaincir-
cumstancesthis may bring about a significantmovement of the aerosol
particles. Suchmovementsare called diffusion.
If there is a heavier [Link] coming from the
sameside of all particles,there will be a generalmotion toward the op-
positedirection. This will be the casewhen there is a large temperature
gradient in the gas. Gasmoleculeson the higher temperaturesidewill
possess more kinetic energythan thoseon the lower temperatureside,
hencewill impart momentumto the particlein the directionof decreas-
ing temperature. The process is [Link] is responsible
for the fact that in the vicinity of very hot surfacesthe air will be prac-
tically dust-free.
A similar situationexistswhen aerosolparticlesare illuminated by
stronglight from one side. Absorptionof the light by the particlesets
up a thermal gradientwithin the particle. Gasmoleculescolliding with
the warmer side of the particle rebound with increasedkinetic energy
andagainsetup a net motionawayfrom the directionof the warmerside.
This motion iscalledphotophoresis.
Another exampleof what may be called unbalancedbombardment
occurswhen the aerosolparticlesare suspended in a gaswhich contains
a high concentrationgradientof one kind of gasmolecules. As these
diffusein the directionof decreasing concentration they alsomovethe
aerosolparticlesin the samedirection. This is especiallynoticeablein
the gaszoneat the interfaceof either an evaporatingliquid or a con-
densingvapor. In the former,aerosolparticlesare pushedawayfrom
the liquid surface,and in the latter casetowardthe liquid surface,by
the streamof [Link] process is calleddil•usiophoresis.
In the absenceof any of the unbalancedsituationssuchasjust de-
scribed,a small aerosolparticle may still be made to move around in
random fashionby molecularcollisions. This is called Brownian di[-
[usion. It is responsiblefor aerosolparticlesdiffusingin the direction
of a decreasing aerosolconcentration gradient,obeyingFick'slaw, in the
samemannerasmoleculardiffusionoccurs. During the randomwander-
ingsof an aerosolparticleit may collidewith a neighboringsurface.
This becomes anothermechanism by whichparticlesmay becomere-

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 MOVEMENT OF AEROSOL PARTICLES 673

tnovedfrom [Link] is only a significantfactor for particlesof

sizelessthan about 1 v.
: The mean displacementof a particlein a givendirection by Brownfan
diffusionin still air may be calculatedby Einstein'stheory as

AxB
=i•:Dt (12)
where f) is called the diffusioncoefficientof the particle and is given by
C

= (13)
where k: Boltzman'sconstant,and T = absolutetemperature. Fuchs
(7_)hasprepareda table of samplevaluesof •xx, for t = 1 sec,and com-
phred them with us. Excerptsare quoted in Table VII. For particle
sizeswhich are much smaller than the mean free path of the gasmole-
cules,another method of estimatingvaluesof .• is available from Lang-
muir (8).
Table VII

Brownfan Diffusion in Still Air

Air at23øC, 1 atm; p• = 1 g/cma; t = 1 sec

Particle
Size
(dv, la)
•D,cm•
sec
__

AxB,sec u,, cm/sec

10 2.38 X 10 -8 1.74 X 10 -4 3.02 X 10 -2
1 2.74 X 10 -7 5.90 X 10 -4 3.47 X 10 -a
0.1 6.82 X 10 -6 2.95 X 10 -a 8.64 X 10-a
0.01 5.24 X 10 -4 2.58 X 10 -• 6.63 X 10 -6

It is clear from Table VII that for particlessmaller than about 0.5 p,
the Brownfan motion exceedsthat of settling. This providesa rough
guidelinefor identifyingwhat is meant by a "small" particle in termsof
the importanceof diffusion.
When the particlesare suspendedin a moving air stream flowing
over a surface, the contribution which Brownfan diffusion makes toward
depositionon that surfacewill be lessenedas the velocity is increased.
The effectis judged by the magnitudeof the Pecletnumber, definedas
voD
(14)
Again, D is a characteristicdimensionof the surfaceas was used in the
caseof inertial impactionand listed in Table III. The smallerthe Peclet
number,the greaterthe role playedby diffusion.

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 674 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

A number of studieshavebeendoneon the importantcaseof deposi-

tion by diffusionon to the surfaceof a large cylinder,from a gasflowing
perpendicularto the axisof the cylinder. The pattern of streamlinesis
involvedjust asit wasin consideringinertia and direct interception. A
collectionefficiency,by diffusion,is alsodefinedin the samemanner. As
an exampleof thesestudies,the resultsof Ranz (9) may be quoted:
1 0.55•rRe 1/6

Note that the secondterm allows for the effect of the changingpattern
of streamlineswith the Reynoldsnumber. Table VIII givessomeeffi-
cienciescalculated•rom this equation,as applied to a human hair.
When thesecollectionefficienciesare comparedwith thosefor iner-
tial impaction(Table V), it is seenthat diffusioncontributespractically
nothing to the collectionunlessthe particlesare very small. But as the
particle size decreases,
collection by diffusion increaseswhile that due
to inertia [Link] someintermediate particle sizethere is a mini-
mum collection where neither diffusion nor inertia is effective.

Table VIII

Diffusion onto Human Hair

Pv -- 1.0g/cm3; D = 100/z; v0 = 10cm/sec; Re = 0.662

dv, tz Pe •D •z• (%) • (% Table V)

10 4.2 X 106 0.63 X 10-4 --•0 ,'-•5%
1 3.65 X 105 0.32 X 10 -3 •0 -,•0
0.1 1.47 X 104 0.28 X 10 -2 •0.3 ,--•0
0.01 1.85 X 102 0.065 •6.5 ,'--'0

Electrostatic Attraction

Aerosol particles seemto acquire electrostaticchargesrather easily

during generationor during flow through air. As they approachan
unchargedsurface,an image force is set up, which attractsthe charged
particlesto the surfaceand becomesstill another mechanismof deposi-
tion or removal.

For example,the caseof a sphericalchargedparticle flowing near a

cylindrical surfacehas been studied by Lundgren and Whitby (10).
They found that the "efficiencyof collection" could be calculatedfrom
an electrostaticparameterK• by
•/• = 1.5K•1/• (16)

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 MOVEMENT OF AEROSOL PARTICLES 675

where

(•- 1)
(,•q-1)QP2C
= 2%z)2dv0 (17)
Here

0• = electrostaticchargeon a particle,coulombs
•t -• dielectric constant of collector
•0-- permittivityof free space= 8.85 X 102'coulombs2/dY
[Link]•
The value of Qe may be expressedin termsof the number of electronson
the chargedparticle. Notice that
QP
va-- (18)
Dx/d•vo
Sometypicalexamplesare givenin Table IX.

Table IX

Collection by ElectrostaticAttraction
1.0 g/cma; D = 100 t•; v0 = 10 cm/sec
Collection Efficiency(Va)
Charge-6 electrons Charge-300 electrons
10 •0 --•0
• ,--o •%
0.• •0.•% ,--,3.s%
0.0• •0.2% •%

The effect is small unlessthe particlesare small and the charge is

large. It may easily becomemore significantthan diffusion however.
Unfortunately, the sizeof chargeis usually not known, nor is it usually
under the controlof the userof the aerosol. In somecasesit is necessary
to take stepsto removeelectrostaticchargesin order to avoid random,
erratic, and unexpectedeffectsupon deposition. One method of doing
this is to passthe aerosolthrough a neutralizing field such as may be
generatedby a radioactivesource.
A cloud of aerosolparticles,each chargedwith the same polarity,
will expandin volume at a nearly uniform speedin all directionsas the
particlesrepel eachother. This effectcould be usedto enhancedeposi-
tion in all directionson the insidesof a confining spacesuch as walls
and ceiling of a chamber,or surfacesof the respiratorytract.

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 676 JOURNAL•OF THE SOCIETY OF COSMETIC CHEMISTS

SUMMARY OF METHOD OF CALCULATION

Given an aerosolparticle of specifiedsize dr, and density o, together

with its initial velocity u0 and the velocity v0 of the air stream in which
it will be moving, two general casesmust be distinguished.
1. Air is essentiallystill, i.e., v0 < < u0,or v0 = 0.
ppdp2
(a) Calculate r-
18/1
Xs • HoT

Us = gr
k TC

3•-gdv

•a=i-•X)t(t=1sec)
_

(b) Compareu•with zxXB

If u• > zxxB,gravity controlsand trajectory o[ particle may
be determinedby usingFig. 1 (seealsoTable I).
--

I[ u• < zxx• diffusioncontrolsmotion and particle deposi-

tion may be estimated[or variouscasesby equationsgiven by
Fuchs(11) (seealsoTable VII).
2. Air motion is significant,i.e., Vo> Uo;assumeparticlewill rapidly
acquireair velocityof v0.
(a)Calculate
r(as
above)
andRd- vøDps
•'- (b) Using-thesequantitiesand the appropriatedimensionsof sur-
facesupon whichparticlemight deposit,calculate
' usL grL
G- voR-voR (gravity
parameter)
Xs

•b= • = -• (inertia
parameter)
dp
(interceptionparameter)

rod
(diffusionparameter)

Os- 1)

ø .(electrostati
Ku= 12•FeoDS•dvv c parameter)

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 MOVEMENT OF AEROSOL PARTICLES 677

(c) Comparethesevalueswith thosegivenasexamples

in Tables
II, IV, V, VI, VIII, and IX.
(d) From thesecomparisons,
judgewhich[orcehasthe predomi-
nant effect. A rough estimateo[ the efficiencyot•deposition
may be madesimplyby scanningthesetables. I[ desired,the
efficiencyo[ aerosoldeposition (or collision)upon the surface
in questionmaybe estimatedmore accuratelyfrom the equa-
tionsgiven or cited.

TABLE OF SYMBOLS

b (cm) = spacingbetween center line of flow and outermost

streamline within which all particles collide with
obstaclein path
C --Cunningham factor, correction to Stokes law for
small particles
d•,(cm) = diameter of sphericalaerosolparticle
D (cm) = characteristicdimensionof obstaclein path
5) (cm"/sec) = diffusion coefficient of particle due to BrownJan
motion
(dyne) = drag force upon particle moving relative to air
(cm/sec•) = accelerationdue to gravity, 980 cm/sec2
= dimensionlessparameter(general)for gravity deposi-
tion
k = Boltzman's constant
= dimensionless parameter for electrostaticattraction
z, (cm) = lengthof duct or tube, alsoa horizontaldimension
Q•,(coulombs) = electrostaticcharge on an aerosolparticle
= Peclet number, dimensionless
= radial positionin a tube
R (cm) = radius of a sphereor cylinder, alsovertical dimension
Re = Reynolds number, dimensionless
t (sec) = time
T(øK) = absolutetemperature
• ½m/se•) = speedof motion of particle
•o ½m/•e•) = initial speedof motion of particle in still air
• ½m/•) = terminal settlingvelocity of particle
= horizontal componentof velocity of particle
• ½m/•) = vertical componentof velocity of particle
•o ½m/•) = averagevelocityof fluid

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])
 678 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS

vr (cm/sec) = velocity of fluid stream in axial direction, at radial

positionr
% (cm/sec) horizontal componentof velocity of fluid streamline
•_ (cm/sec) vertical componentof velocityof fluid streamline
AxB (cm) mean displacementof particle by Brownian motion
x era) distancetraveled horizontally by particle
(cm) stoppingdistanceof particle in stream
y (cm) distancetraveled vertically by particle
geometric
standarddeviationin •/•p
dielectric constant of collector
•0 pmmittivity of free space
collectionefficiencyof particlesby obstacle
collectionefficiencyof particlesdue to diffusion
collection efficiency of particles due to electrostatic
force
• (dyne/cm sec) = viscosityof fluid
= inertial impactionparameter,dimensionless
0p(g/cma) = density of particle
0l (g/cma) = density of fluid
½½c) -- relaxation time of particle
(ReceivedNovember 15, 1971)

REFERENCES

(1) Findeisen,W., Arch. [Link].,286, 467 (1935).

(2) Landahl, H., Bull. Math. Biophys.,12, 43 (1950).
(3) Hatch, T. F., and Gross,P., Pulmonary D• position and Retention o[ Inhaled Aerosols,
Amer. Ind. Hyg. Assn.,AcademicPress,New York, N.Y., 1964, pp. 67-8.
(4) Fuchs,N. A., The Mechanicso[ Aerosols,The Macmillan Co., New York, N.Y., 1964, p.
112.
(5) Whitby, K. T., Calculationof the clean fractionalefficiencyof low media densityfilters,
ASHREA J., 7, 56-65 (Sept.1965).
(6) Fuchs,N. A., The Mechanicso! Aerosols,The MacmillanCo., New York, N.Y.. 1964,p.
165.
(7) lb"d., p. 184.
(8) Langmuir,I., Theoryof Filtration of Smohes, [Link]. No. 865 (1942).
(9) Ranz, W. E., Tech. Rep. No. 8, [Link]. Eng. Exp. Sta.,Jan. 1, 1953.
(10) Lundgren,D. A., and Whitby, K. T., The effectof particle electrostaticchargeon filtra-
tion by fibrousfilters,Ind. Eng. Chem.,[Link].,4, 345-9 (1965).
(11) Fuchs,N, A., The Mechanicsof Aerosols,The Macmillan Co., New York, N.Y., 1964,
pp. 192-204.

Purchased for the exclusive use of nofirst nolast (unknown)

From: SCC Media Library & Resource Center ([Link])