Вы находитесь на странице: 1из 26

501

The Formation of Emulsions in Definable Fields of Flow


By G. I. T a y l o r , F.R .S., R oyal Society Y arrow Professor
(Received May 9, 1934)
[P lates 4 and 5]

The physical and chemical condition of emulsions of tw o fluids which do not


lix has been the subject of m any studies, b u t very little seems to be known
•out the mechanics of th e stirring processes which are used in m aking them .
rie conditions which govern th e breaking up of a je t of one fluid projected
ifco another have been studied b y Rayleigh* and others, b u t m ost of these
udies have been concerned w ith th e effect of surface tension or dynam ical
rces in making a cylindrical th read unstable so th a t it breaks into drops,
he mode of form ation of th e cylindrical th rea d has n o t been discussed,
s a rule in experim ental work it has been formed b y projecting one
juid into the other under pressure through a hole. I t seems th a t studies
f this kind which neglect th e disruptive effect of th e viscous drag of one fluid
i the other, though interesting in themselves, tell us very little about th e
anner in which two liquids can be stirred together to form an emulsion.
When one liquid is a t rest in another liquid of th e same density it assumes
lie form of a spherical drop. A ny m ovem ent of th e outer fluid (ap art from
are rotation or translation) will distort th e drop owing to th e dynam ical and
iscous forces which then act on its surface. Surface tension, however, will
md to keep the drop spherical. W hen th e drop is very small, or th e liquid
ery viscous, the stresses due to inertia will be small com pared w ith those
ue to viscosity.
Recently the present w riter made a rough theoretical estim ate, based on th e
ydrodynamical equations of a spherical drop in a shearing fluid, of th e
laximum size of drop which surface tension m ight be expected to hold together
gainst the disruptive forces due to th e viscous drags of the shearing fluid,
ince the drop must depart very m arkedly from the spherical form before it
ursts, this theoretical estim ate is unlikely to be of much value except as an
idication of the conditions under which m arked deviations from the spherical
orm began to occur. I t seemed w orth while therefore to make some exped­
ients on the deformation and bursting of a drop of one fluid in another under

* 6 Proc. Roy. Soc.,’ vol. 29, p. 71 (1879).


VOL. CXLVI.— A. 2 O
502 G. I. Taylor
controlled conditions m easuring th e interfacial tension of the two liquids
their viscosities and th e ra te of deform ation of th e outer fluid.
Among th e infinite v ariety of possible fields of flow two have been chosen
which can easily be produced in an actual fluid and a t the same time can be
represented b y simple m athem atical equations. The first is th a t represented
by
u = Cx, v = — Cy,

th e stream lines of which are rectangular hyperbolas. The second is that


represented by
u ' = cty', v' = 0. (2)

Four Roller ” Apparatus

To produce approxim ately th e field of flow represented by (1) the apparatus


represented in th e sketch of fig. 1 was constructed. F our brass cylinders
F IL L I N G HOLE 3*81 c m x 2 -3 9 cm diam eter were mounted at
th e corners of a square th e sides of which
were 3 • 18 cm. Their axles ran in brass bear­
ings fixed in two glass plates which formed
th e sides of a box th e internal dimensions of
which were 7*6 X 7*6 X 3*9 cm. The re­
m aining sides were brass and one of them was
pierced by a large hole through which the
ap p aratu s could be filled w ith liquid. The
cylinders were driven in th e directions in­
dicated by arrows in fig. 1 by means of two
vertical shafts and bevel wheels. These two
shafts were ro tated a t th e same speed but in
opposite directions by a m otor through reduc­
tion gears. The box was filled with golden
^DRIVING PULLEYS
syrup, illum inated by an electric lamp, and
F ig . 1— “ Four roller ” apparatus.
observed by means of a long focus camera
(magnification 2 •5) set w ith its axis on th e centre line in a direction parallel
to the axes of the cylinders.
Slight variations in the w ater content of the golden syrup caused streaks to
appear in th e image on th e cam era screen as soon as the apparatus was set in
motion. These streaks m ark out stream lines and the photograph, fig-
P late 4. shows th a t a t any rate in the centre of the field they are very
The Formation of Emulsions 503

>a!:i the rectangular hyperbolas which are stream lines of th e flow represented
bj(l).
’o compare th e velocity in th e field of flow actu ally produced w ith equation
I l a series of lines were ruled horizontally and vertically on th e screen of th e
“caiera a t distances ± 0 * 5 , ± 1 * 0 ? ± 1 * 5 ? ± 2 - 0 , and ± 2 - 5 cm from th e
Ucetre. The tim es a t which im ages of small particles in th e golden syrup
used successive lines were observed an d also th e tim e t of one revolution of
t l vertical shafts. In one set of such m easurem ents t w as 48 •7 seconds and
t l corresponding tim es for covering th e 0 •5 cm distances betw een successive
liis are given in Table I.
Table I
Distances, in cm, from centre Time, in secs Value o f C from formula (3)
■5 to 1-0 ................................................................. 7 0-0 9 9
■0 to 1-5 ................................................................. 3 -9 0-1 0 4
•5 to 2• 7 ................................................................. 2 -7 0-1 0 6
•0 to 2 -5 ................................................................. 2 -0 0-111

if the field of flow is represented by (1) th e velocity of a particle is


dt = Cx so th a t th e tim e of passage of a particle from x x to x 2 is
CL(log x 2 — log x x). H ence

C = (log x 2 — log x x) -r- (tim e from x x to x 2). (3)

The values of C corresponding w ith this form ula are given in Table I. I t
vd be seen th a t C is nearly co n stan t over th e range from 0*5 to 2*5 cm from
t* centre of the cam era screen which corresponds w ith th e range 0*2 to
) cm from the centre of th e field in th e ap p aratu s itself.
The values of C in any experim ent are proportional to th e speed of ro tatio n
the cylinder, i.e., to 1 / t . Taking th e average value of C in this set of m easure-
mts as 0 • 105 corresponding w ith t = 48 •7 it will be seen th a t in any other
penm ent for which t is m easured th e flow will be represented b y ( 1 ) provided

C = (48-7) (0• 105) T” 1 = 5• 1 / t . (4)


le number 5*1 is a dimensionless constant of th e apparatus.

“ Parallel Band ” Apparatus


To produce the flow represented by (2) two endless celluloid bands of cinema
n 35 mm wide were stretched between rollers, one of which in each case was
bed with pins to engage in the regularly spaced holes a t th e edge of th e film.
ie two bands could be driven a t any speed in either direction, th e ratio of

2 o 2
504 G. I. Taylor
their speeds being controlled b y a continuously variable gear. The bands and
rollers were contained in a glass sided box th e w idth of which was 3 mrp great®
th a n th a t of th e film. The ap p ara tu s is illu strated in th e sketch fig. 3.
W ith th is construction th e speed of th e b an d was definitely related to the
speed of th e rollers so th a t each revolution of a driving pulley moved the corre­
sponding endless b and th ro u g h 7-10 cm.
FILLING HOLE

FILM

,DR0 P

F ig . 3— “ P arallel b and ” apparatus.

The distance a p a rt of th e bands was 1*36 c m so th a t if the shearing motion


of th e liquid betw een th em is represented b y u' — y ’ being measure
perpendicular to th e bands,
7-10 (5)
a
1 -3 6
The Formation of Emulsions 505

w lire tx and t 2 are th e tim es of revolution of th e tw o driving pulleys. I f th e


eortant C in (1) is chosen to be equal to Joe an d if th e whole system (2), i.e.,
i eaxes of co-ordinates (xf, y r) is given a ro tatio n w ith angular velocity Joe
i if th e axes (x ', y f) are instan tan eo u sly a t 45° to th e axes (x, y) th e n a t th a t
m a jit the two fields of flow are identical. Effects therefore which depend
om on the instantaneous d istrib u tio n of velocity and are unaffected b y a
ro ition of th e whole system will be identical in th e tw o pieces of ap p ara tu s
w in they are operated a t corresponding speeds, i.e., so th a t C = Joe. On
th other hand, effects w hich do n o t depend only on th e instan tan eo u s dis-
t r ution of velocity b u t are d ependent on a sequence of such d istributions are
v€jr different in th e two.
!o illustrate th e significance of these rem arks consider th e effect of th e tw o
f a is of flow of a very viscous fluid on an elongated sym m etrical solid body,
so a prolate spheroid, placed w ith its long axis in th e plane xy. I f th is body
is>laced in the field of flow (1) w ith its long axis a t angle 0 to th e axis of x
;t surface will be subjected to exactly th e same stresses th a t would a ct on
t a the field of flow (2) if its long axis were placed a t 45° + 0 to th e axis of x .
P'e resultant effects of th e tw o fields of flow on th e m otion of th e body over a
>iod of tim e are, however, very different. I n th e field of flow (1) th e body
1 set itself perm anently w ith its long axis parallel to th e axis of x, whereas
h e flow (2) the body will continually roll over and over ro tatin g , a t a variable
ed, about an axis perpendicular to th e plane x'y'.

Calculation o f Sm all Deformations


Before describing th e experim ental results obtained w ith th e tw o forms of
3aratus we m ay see how far th eo ry can predict th e effects of th e tw o fields
fluid flow (1) and (2) on drops of another fluid im mersed in them .
Suppose th a t a spherical drop of a fluid of viscosity (j/ is im mersed in a fluid
viscosity (j. and th a t th e la tte r is caused to flow in th e velocity d istrib u tio n
>resented by (u = Ox, v = — Cy). I f th e flow is very slow th e viscous drag
1deform the drop only slightly from th e spherical form. This small deforma -
n will cause only a small change in th e distribution of stresses in either fluid,
^ordingly in the stress conditions which m ust be satisfied a t th e surface of
s drop, namely, (a) continuity of tangential stress, and (b)

T (ri-1 + r 2-1) = constant + p* — p 0, (6)


stresses may be reckoned as those which occur when th e drop is held
506 G. I. Taylor
spherical b y a distribution of norm al force a t th e surface. In eq u atio n (6) f
is th e interfacial surface tension between th e tw o liquids, p i and p 0 are the
normal pressures inside and outside th e drop, and r 2 are the principal radii
of curvature of th e deformed drop.
I t follows th a t th e analysis previously given b y th e present writer* for the
flow in th e neighbourhood of a spherical drop can be applied directly in the
present work. The value of — p 0 is therefore

1 V + 16H l t ^ £ ) constant,
Pt (x + fji \ a J (i)

Comparing (6) and (7) it will be seen th a t it is necessary to find the shape of the
nearly spherical drop for which th e variation in (r-f1 + r 2_1) is proportional to
(x2 — y 2) a r2. I t can be verified th a t for th e surface whose equation is

r = a + b (x2 — y 2) a r2 ( 8)
rj-1 + r 2-1 = 2a r1 + 4b (x2 — y 2) a - 4 ( ) 9
Combining (6), (7). and (9) and equating coefficients of th e variable part of the
pressure it will be seen th a t (8) represents th e deformed drop provided

iC 1 9 ^ + 16{jl _ 4T6 ( 10)


\x + p. a2

A convenient m ethod for expressing th e results of experim ent is to measure L,


the length of th e drop in th e direction of th e x axis, and B, the breadth in the
direction of th e y axis. These m easurem ents are connected with the constant
b of equation (8) b y th e form ula

L —B _ b
L + B a5
so th a t (10) becomes
L —B tp -f~ 16^ ( 12)
L + B 1 6 ^ '+ 16(x’
where
F = 2C[ia/T. (13)
F is non-dimensional.
I t will be noticed th a t over th e whole range of ratios \l j [X from 0 to a
(19p/ + 16fx)/(16(i/ + 16(x) varies only from 1*0 to 1*187, so th a t (L —®)/
(L + B) is nearly equal to F.

* “ The v iscosity o f a fluid containing sm all drops o f another fluid,” ‘ Proc. Roy.
A, vol. 138, p. 41 (1932). The expression here given in (7) is identical w ith equation (2
o f th a t paper excep t th a t 2C has been su bstituted f or a.
The Formation of Emulsions 507

or th e flow (u ' = a y \ v' = 0) th e d eform ation of th e d ro p a t slow ra te s


f ow should also be represen ted b y (12), b u t in t h a t case

F = [xaoc/T, (14)

,r th e long axis should lie in th e d irection m ak in g 45° w ith th e axis of x '


d ,, a t 45° to th e celluloid film as show n in fig. 3).

M ethod o f E xperim ent

n all th e experim ents to be described th e a p p a ra tu s w as filled w ith golden


up (which is a co n cen trated sugar solution) d ilu ted w ith a sm all q u a n tity
o water till th e viscosity w as betw een 50 an d 150 c.g.s. A n u m b er of liquids
ich do not m ix w ith w ater were used for th e d ro p so as to cover a large range
o values of \Lj\x. F o r low values of \l j\L a m ix tu re of carb o n tetrach lo rid e
vDh the paraffin oil sold as “ N ujol ” w as m ade u p to be of th e sam e d en sity
a the syrup and was coloured pu rp le w ith dissolved iodine. F o r values of p/
rther smaller th a n p. a lu b ricatin g oil sold as “ B B ” w as used. F o r values
oc p' equal to p a black lu b ricatin g oil w as used. F o r values of p ' considerably
eater th a n p a m ix tu re of coal ta r an d p itch w as m ade u p so th a t its viscosity
VkS about 2000 c.g.s. These values to g eth er w ith th e interfacial surface
isions are given in Table I I .

Table I I
n ’ (c.g.s.) T (c.g.s.)
CC14 and paraffin ........................... ....... 0 034 23
“ BB ” oil ...... ....... 60 17
Black lubricating oil .................. ....... 100 8
Tar-pitch m ixture ...................... ....... 2000 (23)

Before setting th e a p p ara tu s in m otion th e drop was introduced in to th e


rup by means of a p ip e tte w hich w as lowered th ro u g h th e hole in th e to p
the apparatus. The drop becam e spherical u nder th e influence of surface
insion and was photographed in th a t condition in order to m easure its radius,
ie apparatus was th e n set in m otion a t a slow speed and ad ju sted till th e
op was steady and stationary. A photograph was th e n taken. The speed
is then increased and ano th er p h otograph tak en , these operations being
peated till some lim iting condition such as th e bursting of th e drop was
lached.
Experim ents w ith the “ Four Roller ” A pparatus

A fter introducing th e d rop a t th e to p of th e a p p ara tu s th e la tter was slowly


set in m otion so as to carry th e drop to w ard s th e m iddle of th e field. After
prolonged runn in g th e drop g rad u ally to o k up a position in th e central hori­
zontal plane. I t was, how ever, highly un stab le so far as horizontal motion was
concerned tend in g to m ove off to th e rig h t or left if it got displaced from the
centre. This in sta b ility w as controlled b y v arying th e speed of the right- or
left-hand p air of rollers. I t has alread y been m entioned th a t th e two rollers
on th e rig h t w ere driv en b y one sh aft an d th e tw o on th e left by another.
E ach p air w as driv en b y a b elt w hich could be m ade to slip. W hen uncon­
trolled n eith er of th e belts slipped so th a t all four rollers ro ta te d a t the same
speed. I f th e d rop w as observed to get slightly off centre to th e right, the
rig h t-h an d p a ir of rollers w as re ta rd e d an d th e drop m oved back towards the
centre. A fter a little p ractice it w as possible to keep th e drop very close to the
centre of th e field.
The p h o to g rap h of fig. 4, P la te 4, shows a drop w hich h ad been maintained
in th is w ay sym m etrically in th e m iddle of th e field for a tim e which was long
enough to ensure th a t th e d rop an d surrounding fluid was in a steady state.
O ne-quarter of each of th e four rollers is shown in th e corners of th e photograph
an d th e direction of ro ta tio n of each is shown b y m eans of an arrow. Each of
th e photograph s ta k e n w ith th is a p p a ra tu s was sim ilar to th a t of fig. 4, but
in order to economize space only th e p a r t of th e field im m ediately surrounding
th e drop is shown in th e rem aining photographs.
= 0-0003— The sequence of p h otographs of fig. 5, P late 4, shows the
effect of th e field of flow represented b y equation (1) on drops of the CC14
an d paraffin m ix tu re ( ^ ' = 0 - 0 3 4 ) . A t each setting of th e apparatus the
tim e of revolution t o f th e rollers w as observed, th e ap p aratu s being un­
controlled during th e exposure, an d for sufficient tim e before it, to ensure
steady conditions. U sing eq u atio n (4), th e co n stan t C was found. The
diam eter 2a of th e u n d isto rted drop w as m easured on th e photograph and T
were m easured independently b y th e m ethods described in th e Appendix so
th a t th e non-dim ensional q u a n tity F = 2c[iaJT could be calculated.
The analysis of th e slightly d isto rted drop shows th a t th e shape of the drop
should depend on th e tw o non-dim ensional qu an tities £jl' / jjl and F. The shape
of th e drop can be defined b y th e ratio (L — B )/(L + B) where L is the greatest
length an d B th e b read th . F o r an y given value of jz'/ fx th e experimental
The Formation of Emulsions 509

ults can, therefore, be expressed b y m eans of a single curve giving


y —B )/(L + B ) for all values of F. In each case th e m axim um and m inim um
tJmeters of the photographs of distorted drops were m easured and
— B)/(L + B) was found. U nder each photograph of fig. 5 th e m easured
vlues of F and (L — B )/(L + B) are given and over each photograph th e
rlius a of the drop in centim etres.
In the first series of photographs (fig. 5) it will be seen th a t a t th e slowest
saed, F = 0 - 1 8 , the drop is only slightly distorted from th e spherical form,
ad the distortion, m easured by (L — B )/(L + B), is 0*15. In this drop
£./£ji = 0*034/100 = 0*00034 so th a t th e predicted relationship (12) is nearly
tactly F = (L — B)/(L + B). The experim ent therefore is in good agree-
rent with the theory.
In the second photograph of fig. 5 F = 0 * 2 8 and (L — B )/(L + B) = 0*26,
s th a t the theory still appears to represent th e experim ental conditions. In
te third photograph th e drop has developed into a form which is very far
r>m spherical, indeed th e ends have become pointed, so th a t th e th eo ry can
i) longer be applied. I t will be noticed, however, th a t F = 0 * 4 1 , (L — B )/
+ B) = 0 •44, so th a t th e theoretical relationship for small deform ations
i still nearly correct. The drop shown in photograph 3, fig. 5, was n o t in a
ily steady state, for after developing th e point shown in th e photograph a
in skin appeared to slip off its surface and th e ends of th e drop again became
unded. This condition persisted, as shown in th e fourth photograph, till
= 0*54. At F = 0 *65 (fifth photograph) th e ends of th e drop again became
>inted and these points rem ained as F increased up to th e highest speed a t
tiich the apparatus could be operated.
The sixth, seventh, and eighth photographs, fig. 5, show a pointed drop which
creases in length and decreases in thickness w ith increasing speed. Thus
1 B increases from 0*65 to 2*45 (L — B )/(L + B) increases from 0*51 to
87. I t was not possible to realize values of F higher th a n 2 •45 because the
ids of the drop got into th e region where the field of flow ceased to be even
pproximately th a t represented by (1). There was no sign, however, th a t
ie drop would have bu rst even if considerably higher speeds had been
fained.
The results described above, together w ith some others for which the drop
not here reproduced, are shown graphically in fig. 6 where values of F and
i_~ B )/(L + B) are plotted in a diagram. The theoretical relationship
= (L — B)/(L + B) applicable to small values of F is shown by means of
broken fine.
510 G. I. Taylor
[L l[L = 0*9—U sing black lu b ricatin g oil for w hich p/ = 100 c.g.s. in symp
for w hich pi = 110 c.g.s. th e value of p,'/[ l w as 0*91. The interfacial surface
tension T was found to be 8 -0 c.g.s. A drop of oil 0-144 cm diameter was
used and th e series of p h o to g rap h s show n in fig. 7, P la te 5, were obtained.
In th e first of these th e tim e o f rev o lu tio n of th e rollers was 97 seconds so that
from (4) and (13)
1 0 -2 (0 - 144 ) ( 110 )
(8 -0 ) ( 97 ) ’
The value of (L — B )/(L + B) o b tain ed b y m easuring th e photograph was
0-19 an d its th eo retical value from (12) is 1 -0 9 F = 0 - 2 3 . I t will be seen.

CALCULATED

ir r l * i*o F

F ig . 6— CCl4-paraffin. “ F ou r roller ” ap p aratu s, yi'/yL = 0 -0 0 0 3 .

therefore, th a t th e observed value 0 -19 is in fair agreem ent w ith its theoretical
value 0-23. The error is certain ly n o t g reater th a n th a t which arises from the
uncertainties involving in m easuring T.
The second p h o to g rap h shows th e dro p a t F = 0 - 3 0 w hen th e theoretical
value of (L — B )/(L + B) w as 1-09 (0-30) = 0-33. The observed value
was 0-29 so th a t again th e th e o ry is confirmed. In th e th ird photograph
F = 0 - 3 7 an d th e d rop h as becom e m uch elongated. The observed value ol
(L — B )/(L + B) = 0 - 5 4 is now considerably greater th a n the value
(1-09) (0-37) = 0 - 4 0 ob tain ed b y ex trap o latin g th e theory.
W hen th e speed reached th e p o in t a t w hich F = 0 - 3 9 th e drop began to
pull o u t into a th read -lik e form . This is shown in th e fo u rth and fifth photo
graphs of fig. 7 w hich were ta k e n while th e drop was bursting in this way
Shortly after ta k in g th e fifth p h o to g rap h th e ap p ara tu s was stopped. The
th rea d of oil w hich h ad seemed qu ite stable while th e ap p ara tu s was in motion
then gradually broke u p in to a num ber of small drops. The final appearanc
of th e oil in th is condition is shown in th e sixth photograph of fig. 7.
The Formation of Emulsions 511

These results are show n graphically in fig. 8. I t will be seen th a t w hen


fj p = 0-9 th e drop rem ains coherent so long as F <[ 0• 39 b u t b u rs t as soon
s F reaches th is value. This is ind icated b y th e broken line co n tin u atio n of
le curve.
jx'/jx = 20— To investig ate th e effect of high viscosity of th e d rop a m ix tu re
f ta r and pitch w as m ade w ith viscosity 2000 c.g.s. This h a d a d en sity n o t
ir from th a t of th e syrup, nam ely, 1 •40. I n th ese ex perim ents th e viscosity
f the syrup was 99 c.g.s. No w ay w as found for m aking a n in d ep en d en t

B U R ST S
HERE

CALCULATED
L -B
= 1-09 F

. F ig . 8— “ Four roller ” ap paratus. B la ck lu b rica tin g o il, = 0 -9 .

and reasonably accurate m easurem ent of th e interfacial surface tension.


On the other hand, if th e tr u th of th e th e o ry of slightly deform ed drops is
assumed equations (12) and (13) can be applied to find T.
A drop 0*16 cm radius was used. The first photograph, fig. 9, P late 4,
shows this drop in its undeform ed condition. The ap p ara tu s was set in m otion
so th a t the tim e of revolution t of th e rollers was 54 seconds and th e second
photograph of fig. 9 was taken. B y m easuring th is photograph it was found
th a t (L — B)/(L + B) = 0-15. Taking as 20 in (12) th e theoretical
512 G. I. Taylor
value of (L — B )/(L + B) is 1 -18F so th a t if th e th e o ry is correct F = 0-15
1-18 = 0 - 1 3 . R eferring to eq u atio n (13) it will be seen th a t when a = 016.
t — 54, jx = 99, th e n th e value of F is 0 • 13 p rovided T = 23. This value is
given in Table I I in b rack e ts an d is available for use in fu rth er experiments
w ith th e ta r-p itc h m ix tu re.
On increasing th e speed of th e rollers till t = 24 seconds th e drop was found
to be bursting. T he th ird p h o to g rap h of fig. 9, P la te 4, shows th e drop in the
a c t of b u rstin g . U sing T = 23 c.g.s. the cor­
responding valu e of F w as 0-28. I t appears,
1-0
therefore, th a t w hen u '/g . = 20 the drop
b u rsts for some v alu e of F rath e r less than
9
!
0-28.
0-8 X These resu lts are ind icated in th e diagram,
BU R STS fig. 10.
HERE
T ~ E xperim ents w ith the “ Parallel Band
0-6 1i . ...
1 A pparatus
L - B 1
8 T his a p p a ra tu s w as d riv en b y a motor
L + B 1 >
7 s th ro u g h various red u ctio n gears an d a variable
0-4 ____/ /
gear so th a t th e speed could be varied and
also, independently, th e ratio of th e speeds of
th e tw o bands. A drop of fluid was placed in
/ th e a p p a ra tu s an d photographed. The appara­
0-2 /
tu s w as th e n set in m otion. I f the drop
.CULATED
/T i- h ap p en ed to be placed exactly mid-way be­
/ 1l+| - M8F tw een th e b an d s it w ould rem ain a t rest when
OZ th e y b o th m oved a t th e same speed. If the
F d ro p w as n o t q u ite cen tral th e n it could be
p^ r ” - by ^ » *
of th e speeds of th e tw o bands.
W hen th e conditions h ad becom e stead y a p h o to g rap h was tak en and the
speeds of b o th b an d s m easured b y tim in g th e revolutions of th e driving rollers.
p//[x = 0 * 0 0 0 3 — A drop of th e sam e m ix tu re of CC14 and paraffin as that
used in th e “ four roller 55 a p p a ra tu s w as introduced. The first photograph
of fig. 11, P la te 5, shows th is dro p before startin g th e ap paratus. The dark
horizontal lines a t th e to p an d b o tto m are th e celluloid bands which in the
a p p a ra tu s itself were vertical. The arrow above th e whole series of pictures
shows th e direction of m otion of th e b an d w hich is a t th e to p of th e photograph.
The Formation of Emulsions 513

flie other b an d m oves in th e opposite direction. M easurem ent of th e first


lotograph showed th a t th e rad iu s of th e d ro p w as 0-157 cm.
The second p h o to g rap h of fig. 11 shows th e d ro p d isto rte d b y viscous d ra g ,
e tw o periods of rev o lu tio n of th e d riv in g rollers being = 192 seconds,
, = 233 seconds. I t will be seen th a t th e long axis of th e d isto rte d dro p lie
about 45° to th e band s, i.e., in th e d irectio n in w hich lines of p articles are
ongating a t th e g reatest ra te . I n th is ex p erim en t it w as fo u n d th a t
= 123 c.g.s., so t h a t from (5) an d (14)

F = (5*22) (0*157) (123) T -1 ( t u 1 + t 2- j ).

f T is ta k en as 23 c.g.s. th is becom es F = 4*4 ( t U 1 + ^ 2_1) so B = 0*04.


Measurement of th e p h o to g rap h gives (L — B )/(L + B) = 0*08 so th a t th e
leoretical relationship F = (L — B )/(L -f- B) is n o t fulfilled.
The th ird photo g rap h ta k e n w hen F = 0 - 1 0 gives (L — B )/(L + B) = 0 * 2 2
0 th a t again th e observed v alue of (L — B )/(L + B) is a b o u t tw ice as g reat
s the prediction. The consistency of th e error m akes it seem pro b ab le th a t
re surface tension of th e dro p in th e sy ru p w as considerably less d u rin g th is
eries of experim ents th a n it h a d been previously, p ro b ab ly owing to im p u rities
1 th e syrup. The sequence of d rops show n in th e eight p h o to g rap h s of fig. 11
aows th a t as th e speed of th e a p p a ra tu s is increased th e dro p elongates b u t
ioes not b u rst even a t th e h ighest speed obtainable. The highest value of F
h.own in fig. 11 is 2*30, b u t th is p h o to g ra p h w as o b ta in e d w ith a different
Uing of syrup from th e o th ers. One experim ent, how ever, w as tried w ith
very large drop (a = 0 -54) an d it w as found th a t th is drop w as still co-
Lerent a t F = 5-3 if T = 23 (or F = 11 if T = 11).
I t will be noticed th a t th e fifth p h o to g rap h of fig. 11 shows a th in streak
oming off from each end of th e drop. This was a tra n sie n t phenom enon an d is
vidently th e same as th a t previously described in connection w ith th e th ird
photograph of fig. 5, P la te 4.
The results obtained w ith tw o fillings of syrup are set o u t in fig. 12 as
4F irst series 55 and “ Second series.”
[l' I[l = 0*5—U sing fresh syrup an d a drop of “ B B ” oil th e results set fo rth
n Table I I I were obtained. T was m easured independently and found to be
l7 c.g.s. In th e first set of observations jjl was 135 c.g.s. while in th e second
t was 110 c.g.s. The second set is shown in th e photograph, fig. 13, P late 5.
t will be seen th a t for low values of F th e drop is, as predicted, an ellipse w ith
ts long axis a t 45° to th e celluloid bands. As th e speed of th e ap p aratu s is
514 Gr. I. Taylor

increased the drop becomes elongated till a t F = 1 •43 a steady motion ceases
to be possible and th e drop elongates into a thread-like form.

* 0-87

CALCULATED

it I 'I- of

P ig . 12— “ P arallel b a n d ” apparatus. CC14-paraffin, p '/p = 0 0003. • First series;


H Second series.

Table I I I — “ BB 55 oil drop in “ parallel band 55 apparatus, p ' = 60, T = 17


L - B
a (cm) V rj (secs) r 2 (secs) F
L+ B
0-140 110 62 51 0-17 0-17
0-167 135 157 828 0-05 0-07
0-123 135 99 87 0-11 0-10
0-123 135 83 85 0-12 0-19
0-123 135 24 25 0-42 0-45
0-101 135 9-5 10-5 0-84 0-60
0-100 135 7-0 8-0 1-12 0-81
0 • 100 135 5-5 6-2 1-43 burst

DROP HAS
BURST

CALCULATED

| k f =106

F ig . 14— “ Parallel band ” apparatus.

These results are set out graphically in fig. 14. For p '/ P- — 0 •5 the theoretical
relationship (12) is (L — B )/(L + B) = 1 -06F. This is represented in fig. 44
by a broken line. I t will be seen th a t the agreement is good.
The Formation of Emulsions 515

(j//{ji= 0*9—W ith the same black lubricating oil which had previously
en used in the “ four roller ” apparatus, the results given in Table IV were
otained. Some of the photographs are shown in fig. 15, P late 5. I t will be
Dticed th a t when F = 0-28 th e drop is not sym m etrical, one end being more
minted th an the other. This asym m etry is still more pronounced when F =
\ 55. I t always occurred when there was a large difference in density between
re drop and the syrup.

ible IV—Black lubricating oil in syrup, “ parallel band ” apparatus.


[x = 110, p/ = 100, T = 8 -0
L - B
a Tjl(secs) r 2 (secs) F
L -f-B
0-097 262 207 0-06 0-07
0-097 141 111 0-11 0-15
0-097 92 71 0-17 0-22
0-097 57 44 0-28 0-45
0-086 40 31 0-36 0-53
0-086 22 23 0-55 burst

DROP HAS
BURST

CALCULATED
L -B
* 1*09 F

F ig . 16— “ Parallel band 55 apparatus. Black lubricating oil. = 0 -9 .

The relationship between (L — B)/(L + B) and F is shown in fig. 16 and


e theoretical relationship for p// p. = 0 •9, namely, (L — B)/(L + B) = 1 •09F,
there shown by a broken line.
516 G. I. Taylor
fx'/fx = 20—Using th e pitch and ta r m ixture for which p/ was about 2000
c.g.s. the drop again became elliptical w ith its long axis a t 45° to the celluloid
bands when th e ap p aratu s was ru n a t very slow speeds. This is shown in the
second photograph of fig. 17, P late 5, tak en when F = 0-08.* As the speed
increased to F = 0*27, th ird photograph, fig. 17, th e drop became slightly
more elongated, b u t its long axis became more nearly parallel to the celluloid
bands. A t F = 1-13, fo u rth photograph, fig. 17, its deformation was only
slightly greater th a n a t F — 0 •27 and its long axis was still more nearly
parallel to th e bands. A t F = 1*69, fifth photograph, fig. 17, its shape and
orientation were alm ost identical w ith those a t F = 1-13.
In com paring these results w ith those shown in fig. 9, P late 4, which were
obtained w ith th e “ four roller 55 apparatus, a very striking difference will be
noticed. In th e “ four roller ” ap p aratu s th e drop burst a t a very low speed
represented by F = 0*28. I n th e “ parallel band ” apparatus the drop did
not bu rst however fast th e ap p aratu s was run. On th e contrary, it attained at
high speeds a constant condition in which (L — B ) /( L + B) was 0*26 and
the long axis was parallel to th e bands.
The explanation of th is phenom enon is simple. In th e “ four roller ”
apparatus, which produces th e field of flow (1), th e lines of particles which are
extending a t th e greatest rate, nam ely, those parallel to th e axis of x, remain
in th e direction of m axim um rate of elongation as long as th e flow continues.
The disruptive stress due to th e viscous drag of th e syrup is therefore always
tending to extend th e drop in th e same direction. As soon as this stress is
able to overcome th e cohesive effect of surface tension the drop bursts.
In th e “ parallel band ” ap p aratu s th e lines of particles which lie in the
direction of m axim um rate of elongation, namely, a t 45° to the bands, are
continually being ro tated aw ay from th a t position tow ards the line parallel
to the bands which is neither elongating nor contracting. After attaining
this position further ro tatio n brings th e line of particles into an orientation
where they are contracting. W hen th e drop is very viscous compared with the
surrounding medium it ro tates alm ost like a slightly plastic solid body, lines
of particles in it elongating slowly while th ey are w ithin 45° of the direction
of m axim um rate of elongation and contracting slowly while they are within
45° of the direction of m axim um rate of contraction. In this way the surface
of the drop attain s a perm anent position with its long axis parallel to the bands.
The tar-pitch m ixture rotates round inside this fixed envelope; in fact, accidenta

* The values o f F are found by assum ing T = 23 C.g.s., see Table II, and p. 50/.
?aylor Proc. Roy. Soc., A, vol. 146, Plate 4

F ig . 2 F ig . 4

a= 0-20 0*20 0-25 0-16 0-12

-M 1 ii
I * .^gjjjpK
| Ml <z>
1- ~ o
mum! . .1 > ^: ■

F = 018 0-28 0-41 0-54 0-65


L - B
T 1R= 0 15 0*26 0*44 0-54 0*51
F ig . 5a
0*25 0*25 0-25
^ «

0 -9 5 1 -4 0 2 -4 5

0 -6 9 0*79 0*87
+ B
F ig . 5b

a — 0 • 16 0*16 0*16


«

F = 0 013 0*28
L - B
0*15 B ursting
L + B ~ 0
F ig . 9
(Facing p. 516 )
Taylor Proc. Roy. Soc., A , vol. 146, Plate 5

oQO 12
o

C O
'rH CO CO

© o

r-1 qo co
t" CD '—I
o o o

it
>>

*oi»i
r-d
&
CD
QQ
P

5 2
© 6 i

2 £ 2

O S 2
© ©

2
o
II
pH pp pp 3
L -
I +
t-p h-p
Taylor Proc. Roy. Soc., A, vol. 146, Plate 5 (contd.)

0-26
0 -28
0-15
0-08

t''-
I— H IT
fH
—I
© ©

cc
d

© © o
II II II
PP e
l-P h-p
The Formation of Emulsions 517
41 unevennesses could be seen to move round th e contours of th e drops.
i the speed increases th e viscous stresses in th e surrounding medium and in
tl drop itself increase in th e same ratio. The stresses due to surface tension
<not increase so th a t ultim ately a t com paratively high speeds th e drop
a lines a form which is quite independent of surface tension.

Analytical Treatment o f very Viscous Drop


'he idea just p u t forward can be analysed m athem atically in th e case when
pis large compared w ith jx and th e speed is so great th a t the forces due to
siface tension can be neglected com pared w ith those due to viscosity.
referring to the treatm ent* previously given for finding th e distribution
o:3.ow inside and outside a drop held spherical by surface tension th e con­
: ons of continuity of velocity a t th e surface of th e drop m ust be satisfied
th at equations (8) and (9) of th e previous paper still hold. These are

JA _3 5B _3 = ifr A 2 (8 a )

2B_3 + Joe = tyA 2 + 2B 2. (9a )


b condition of continuity of tangential stress is

A _ 3 - 1 6 B _ 3 + « = ii-, (M A 2 + 4B 2). (17a )

electing the effect of surface tension th e condition of continuity of normal


ddss p rr is
V (Y — P ) = H*' (Y “ P ')
Ol'
[x ( 3A _3 + 24B _3 + a) = p/ ( — FfA 2 + 4B 2). (17b )
e solution o f eq u a tio n s (8 a ), (9a ), (17a ), (17b ) is

10B_3 = — 5a
2 [x -j~ 3 [a
(15)
5a [x
Bo
4 (2[x' + 3[x)
ferring to equation (5) of the previous paper (loc. tit., p. 43) the components
^velocity at the surface are

u — 2B2a cos </>sin 0, v = — 2Ba sin cf>sin 0, w = 0, (16)

ct The V iscosity of a Fluid Containing sm all drops o f another fluid,” 4 Proc. Roy. Soc.,’
Wol. 138, p. 41 (1932).

T)L. CXLVI.— A. 2 P
518 G. I. Taylor
w here
x = a cos cf>sin 6, y = a sin cf>sin 0.
W hen jjl' / jjl is large th ese are

u = £ —, aon cos cf>sin 0, v — — f- aoc sin d> sin 0. (17)

This expression gives th e in sta n ta n e o u s ra te of d efo rm atio n of th e drop referred


to P o lar co-ord in ates (0, </>) a t 45° to th e celluloid b an d s of th e apparatus.
I n th e c e n tral p lan e 0 = th e ra d ia l com p o n en t of velocity is

fa a cos 2<f>. (18)

I t h as a lre ad y been p o in ted o u t th a t th e co n d itio n of a dro p in the shearing


field (u 7 = a y , v f = 0) is id e n tica l w ith th a t in th e field (u — v = — fay)
w hen (x'y') is a t 45° to (xy) w hen th e la tte r is ro ta te d w ith angular velocity fa.
I f th e re su lt of th is r o ta tio n a n d th is ra te of d efo rm atio n is to keep the outer
surface of th e d ro p in a fixed p o sitio n in th e a p p a ra tu s th o u g h th e particles
m ove ro u n d on it, a n d if th is surface is n early spherical its cen tral section may
be rep resen ted b y th e e q u a tio n

r = {a1 + b f (<
)19)
/>
T he condition th a t th is su rface m a y rem a in fixed in space is th a t the com­
p o n e n t of velo city n o rm al to th e surface d u e to ro ta tin g it a t angular velocity
\ct shall be equ al to th e ra d ia l co m p o n en t of velocity of deform ation. Thus

bcnab = iaoc —f cos 2 <f>. (20)


2 dcf>J (x7 ^
The solution of (20) is
fc/ ( ^ ) = 4 —, Sin 2^>

an d th e eq u atio n to th e surface is th erefo re

r == a (l + i —, sin 2 cf>) .

R em em bering t h a t cf> = 0 is th e line a t 45° to th e celluloid bands of the apparatus


it will be seen th a t (21) in d icates :—
(а) th a t th e long axis of th e d ro p is parallel to th e bands.
(б) th e ratio (L — B )/(L + B) is equal to f ({jl/ jjl7),
and th is ra tio is in d ep en d en t of th e size of drop an d speed of th e apparatus.
The Formation of Emulsions 519

hotographs were ta k en w ith a view to verifying th e form ula (22). In


eai ease the value of (L — B )/(L + B) was ab o u t 0-25, while th e value o f
I [l was about 0*07, so th a t (22) was n o t verified q u an titativ ely , th ough th e
prdiction th a t th e drop will assum e a p erm an ent shape w ith its long axis
cpeallel to the bands was verified (see fig. 18).

Conclusions

t has been shown th a t w hen a drop is slightly disto rted from th e spherical
vn by the stresses in a viscous fluid which is in m otion round it th e deform a-
3i depends only on th e instantaneous conditions. The drop becom es
i ehgated in the direction along which lines of particles are elongating a t th e
latest rate. Thus in th e “ four roller ” ap p aratu s th e long axis of th e dro p
Lorizontal while in th e “ parallel b an d ” ap p aratu s it is a t 45° to th e bands.
Te shape of a slightly deform ed drop is in com plete agreem ent w ith a th eo ry
wich shows th a t
L — B = I V + 16(JLF
L -j~ B 16[l — }~ 16[x

the viscous drag on th e surface of th e drop increases th e drop elongates,


kfr its shape no longer depends on th e instantaneous conditions of flow. T hus
two types of apparatu s produce different effects. The ultim ate fate o f
drop as the speed of distortion of th e outer fluid increases depends also
y much on the ratio F or very small values of e.g., 0*0003, th e
>p remains coherent in spite of th e fact th a t it gets very long and narrow
-both types of apparatus.
\.s the viscosity of the drop increases th e speed necessary to burst it gets
with both kinds of flow : thus, w ith th e “ parallel b a n d ” apparatus, it
’sts when F = 1*2 approxim ately, for [i j\x = 0*5. F or [l' = jx it b ursts
:?F = 0 - 5 with the parallel band apparatus, and F = 0*4 w ith the “ four
ar(er ” apparatus.
2 p 2
520 G. I. Taylor
W hen \L ! [A h as increased to 20 th e tw o kinds of d isrupting field have very
different effects. I n th e “ four ro lle r” a p p a ra tu s th e drop b u rsts a t F =0-28
w hereas th e “ parallel b a n d ” a p p a ra tu s seems incapable of bursting the drop
even a t th e highest speeds a ttain a b le. A t high speeds in th e la tter apparatus
th e drop, in fact, a tta in s a co n sta n t shape w hich depends only on In this
condition th eo retical considerations lead to th e p red ictio n th a t

(L — B )/(L + B) = 5fx/4p/,

b u t th is form ula is n o t confirm ed b y th ese experim ents except th a t the drop


does assum e a c o n stan t shape.
I t is rem ark ab le th a t in th e “ parallel b an d ” a p p a ra tu s th e ease with which
th e drop can be b u rs t increases as th e viscosity of th e drop rises from small
values to be equal to th a t of th e su rro u n d in g m edium , b u t w hen the viscosity
of drop rises to be several tim es as g reat as th a t of th e o u ter fluid the viscous
drag of th e la tte r is incapable of b u rstin g it how ever big th e viscous stresses
m ay be. The drop m erely ro tates, rem aining nearly spherical.
F in ally th e m an n er in w hich th e drops b u rs t is of interest. The act of
b u rstin g is alw ays a n elongation to a th rea d lik e form . W hen this thread
breaks u p it degenerates in to d ro p s w hich are of th e order of 1 /100th of the
size of th e original drop. This seems to be rela ted to th e known fact that
w hen an em ulsion is form ed m echanically it contains drops which cover a
very large ran g e of sizes.

The experim ents here described were carried o u t in th e Cavendish Laboratory


th ro u g h th e kindness of L o rd R u th erfo rd , to w hom th e au th o r wishes to
express his th an k s.

S u m m a ry
T he disto rtio n of a drop of one fluid b y th e viscous forces associated with
certain m ath em atically definable fields of flow of an o th er fluid which surrounds
it is discussed. A n expression is found for sm all distortions from the spherical
form w hich occur a t slow speeds. I f L is th e g reatest, an d B th e least diameter,
(L — B )/(L + B) = F approxim ately, w here F is a non-dim ensional quantity?
proportional to th e speed of flow, w hich involves th e surface tension, viscosit) >
an d th e radiu s of th e drop.
A p p aratu s for producing in golden syrup tw o definable fields of flow wer
constructed an d th e ir effects on drops of various oily liquids were observ
and registered photographically.
The Formation of Emulsions 521
greem ent w ith th e o ry w as fo u n d in th e ra n g e o f low sp eed s w h ere a g re e m e n t
m ia t be ex p ected . A t h ig h e r sp eed s th e effect p ro d u c e d b y th e flow v a rie s
iTutly w ith [x'/[x, (x a n d [x' b ein g th e v isco sities o f th e s y ru p a n d d ro p .
□r [xr/[x = 0 •0003 th e d ro p e lo n g a te s a lm o s t in d e fin ite ly , b u t d o es n o t
f ust a t th e h ig h e st sp eed s a tta in a b le . F o r th e d ro p b u r s t a t
F = 1*4. F o r [x = fx' th e d ro p b u r s t a t a b o u t F = 0 -5 . F o r jx'/p. = 20 th e
arD b u rs t a t F == 0 -3 in o ne ty p e o f field, b u t in th e o th e r th e d ro p d id n o t
n is t even a t th e h ig h e st sp eed a tta in a b le .
he difficulty ex p erien ced in b u r s tin g d ro p s o f v isco u s flu id b y a d is ru p tiv e
f k t of flow in a s u rro u n d in g flu id o f c o n s id e ra b ly less v isc o s ity is sh o w n to b e
^qulitatively in acc o rd an ce w ith a th e o ry o f d ro p s in a la m in a r sh e a rin g field
wofflow w hen F a n d \x j ^x a re b o th large.
gA sording to th is th e o ry (L — B )/(L + B ) =
5 i/4 jx'.

A p p e n d ix

Measurement o f V isco sity— F o r th e oils, a n d


c t l syrup viscosity w as m e a su re d b y w eig h in g
i tl; a m o u n t of fluid flow ing in a g iv e n tim e
lough a glass tu b e . T h is m e th o d w as u n ­
: sizable for m e asu rin g th e v isc o sity o f th e
a h hly viscous m ix tu re o f t a r a n d p itc h , ac-
ecdingly th e a p p a ra tu s sh o w n in th e sk e tc h
o fig. 19 w as c o n stru c te d .
-TAR PITCH MIXTURE
^ brass cylinder A 1 • 27 cm d ia m e te r X 3 • 9 F i g . 19— V isc o m e te r fo r v e r y
• ci! long w as h u n g b y a fine steel w ire fro m v is c o u s liq u id s.
d torsion h ead B. A sm all m e ta l d ish con-
ling a w asher 1-2 m m th ic k w as p la ced below it a n d th e to rsio n h e a d
itly low ered till th e b o tto m o f th e c y lin d er w as ju s t in c o n ta c t w ith th e
3her. The w asher w as th e n re m o v e d a n d som e o f th e ta r- p itc h m ix tu re
3 dropped in to th e d ish so t h a t th e d e p th w as slig h tly m o re th a n 1 • 2 m m .
fe brass cylinder w as th e n care fu lly d ro p p e d o n to th e su rface o f th e
cture so t h a t it w as s e p a ra te d fro m th e d ish b y a disc o f th e m ix tu re
l m m th ic k X 1 • 27 cm d ia m e te r. A lig h t p o in te r C serv ed to m a rk th e
itio n of th e cylinder as it r o ta te d a b o u t its v e rtic a l axis.
f th e to rsio n h e a d w as r o ta te d su d d en ly th ro u g h a n angle 0O th e cy lin d er
uld begin to r o ta te in th e sam e d irectio n , th e r a te of ro ta tio n b eing p ro-
1rtional to 0o — 0 w here 0 is th e angle o f ro ta tio n of th e cy lin d er fro m its
522 G r. I. Taylor
in itia l position. I f th e to rsio n a l couple of th e w ire is K (0O — 0) and this is
equal to th e couple d u e to a c tio n of viscosity ;x' on th e base of the cylinder
th e n th e e q u atio n o f m o tio n is
,d 0 fa 27T
K (0O - 0)
^ dt J0 d ’ (23)

w here d is th e th ick n ess of th e disc of fluid w hich in th is case was 0 •12 cm


a n d 2a = 1 * 2 7 cm . T he rig h t-h a n d side of (23) is in c.g.s. u n its 2*12 p.' dtildt.
To find K th e cy lin d er w as allow ed to oscillate freely an d its period timed.
I n th is w ay it w as fo u n d th a t K = 2 - 1 1 X 102 in c.g.s. u nits. Hence (23)
becom es on in te g ra tio n
211 t
= 100 (24)
2 -12 pi'

I n m ak in g m e asu re m en ts it w as fo u n d co n v en ien t to tu r n th e torsion head


su d d en ly th ro u g h h a lf a tu r n so th a t 0O = n. T he tim e ta k e n for the cylinder
to tu r n th ro u g h I n an d 37r/4 w as observed an d it w as found, as w ould be expected,
t h a t th e tim e for 3rc/4 w as ex a c tly tw ice t h a t for |-7i.
U sing form ula (24) it will be seen th a t th e tim e corresponding with 0 = h
is p/ log2 2/100, w hile th a t corresponding w ith 0 = 37t/ 4 is log2 4/100.
T his a p p a ra tu s w as used in m ak in g a m ix tu re of ta r an d p itch of the desired
viscosity. I n th e m ix tu re chosen for some of th e ex p erim en ts the time for
0 = iy7u w as fo u n d to be 14 seconds w hile t h a t for 0 = 37i/4 was 28 seconds,
so th a t fji' = (100)(14)/log2 2 = 2 X 103 c.g.s.
M easurem ent o f Surface T en sio n — I n m easu rin g th e interfacial surface
ten sio n betw een tw o liq u id s one or b o th of w hich are v ery viscous it is advisable
to avoid th e use of a n y a p p a ra tu s co n tain in g an appreciable length of capillary
tu b e , for th is m a y u n d u ly p rolong th e tim e necessary for th e system to attain
equilibrium . I n th e p re se n t w ork th e surface ten sio n w as m easured by means
of th e a p p a ra tu s show n in fig. 20. This sk etch is self-explanatory. The syrup-
oil in terface w as a t th e end of a v ery sh o rt len g th of capillary tube whose
in te rn a l d ia m ete r w as 0*37 m m . T he sy ru p w as in an open glass-walled
box an d th e in terface could be observed th ro u g h a microscope. The oil was
in a U -tu b e an d th e pressure of th e air over th e air-oil surface, which ^ aS
large area, could be v aried a n d m easu red b y a m anom eter.
T he air pressure w as first a d ju ste d till th e syrup-oil interface appeared
in th e m icroscope (th e oil w etted th e glass) an d th e n m easured. The air
pressure w as th e n increased till it a tta in e d its m ax im u m stead y value when th
interface w as hem ispherical. T he difference betw een th e air pressure whe
The Formation of Emulsions 523
h interface is fiat and th e air pressure w hen it is hem ispherical w ith radius
4i j 2T/a.
Vith this ap p aratu s th e values given in Table I I were m easured.

OIL SYRUP INTERFACE

MICROSCOPE

C I3 J

r C L A S S BOX
W ATER

F ig . 20 — A pparatus u sed for m easu rin g surface ten sio n b etw een o il an d syrup.

D E S C R IP T IO N OF P L A T E S 4 A N D 5

O. 2— Stream lin es in “ four roller ” apparatus.


O. 4— Large drop in “ four roller ” ap paratu s.
5— Drops o f CCl4-paraffin m ixtu re in “ four r o lle r ” ap paratu s.
O- 7— Drops o f black lu b ricatin g o il in “ four roller ” apparatu s.
9— Drop o f tar-pitch m ixtu re in “ four roller ” ap paratu s,
k* 11— D rops o f CCl4-paraffin m ixtu re in “ parallel b and ” apparatus.
13— Drops o f “ B B ” oil in “ parallel band ” apparatus.
15—D rop o f black lu bricatin g oil in “ parallel band ” apparatus.
G . 17— Drop o f tar-pitch m ixtu re in “ parallel band ” apparatus.

Вам также может понравиться