JOURNAL OF T H E

AMERICAN CHEMICAL SOCIETY

(Registered in U. S. Patent Office) (Copyright, 1950,by the American Chemical Society)

VOLUME72 NOVEMBER 17, 1950 NUMBER

11

[CONTRIBUTION FROM THE DEPARTMENT

OF CHEMISTRY, COLUMBIA UNIVERSITY ]

Theory, Production and Mechanism of Formation of Monodispersed Hydrosols

BY VICTORK. LAMERAND ROBERTH. DINEGAR
Introduction is again one of self-nucleation. It is the purpose
The preparation of a colloidal dispersion by of this paper to summarize the unpublished the-
condensation from initially supersaturated solu- ory’ for the reaction of acid and thiosulfate; to
tion is a well-known procedure. This method is show by an appropriate application of this theory
being reinvestigatedl in This Laboratory in con- how monodispersed sols can also be prepared by a
junction with modern theories of phase transi- dilution method; to present a mechanism for their
tions2J3to develop a theory which will account for formation involving the relationship between the
the production of monodispersed colloids. initial concentration of sulfur and the amount of
The preparation of a colloid which was suffi- water added; and to establish the growth rates
ciently monodispersed to exhibit higher order and the monodispersity of the preparations.
Tyndall spectra was achieved in 1941 in This Lab- Theory of Nucleation.-Consider the prin-
oratory* for various oil aerosols by the use of for- ciples involved‘ in the preparation of a monodis-
eign nuclei. The method was extended in 1943 to persed sulfur hydrosol by the slow decomposition
sulfur hydrols6 by the careful control of the con- of dilute sodium thiosulfate in dilute hydrochloric
centration of hydrochloric acid and sodium thio- acid. The reaction 2HSz03- = 2HS03- Sa +
sulfate but without the addition of nuclei. produces molecularly dissolved sulfur, bisulfite ion
The addition of sulfur, dissolved in an organic and also a certain amount of polythionate depend-
solvent such as ethanol or acetone, to a large vol- ing upon the initial concentrations. 7 s 9 10,11.12
9

ume of a miscible non-solvent like water has al- Schematically Fig. 1 shows that the concentra-
ways resulted in a polydispersed preparation, tion of sulfur builds up during stage I a t a rate
commonly called “milk of sulfur.’16 The process which varies as (Naz$zOa)’’X and (HCl)’’; until a
(1) V. K.LaMer, Paper read before the Division of Physical and
critical concentration is reached.* On reaching
Inorganic Chemistry, American Chemical Society, New York, N. Y., this reproducible metastable condition the system
September, 1947. V. K. LaMer and R. H. Dinegar, Paper read becomes heterogeneous by a process of self-nuclea-
before the Division of Colloid Chemistry of the American Chemical tion yielding droplets of supercooled liquid lambda
Society, Atlantic City, N. J., September, 1949.
(2) M. Volmer, “Kinetik der Phasenbildung,” Edwards Bros., sulfur. The rate of nucleation, as judged by the
Ann Arbor, Michigan, 1945. reproducibility of the timing and the concentra-
(3) J. J. Frenkel, “Kinetic Theory of Liquids,” Oxford Univer- tion of sulfur a t which the system becomes hetero-
sity Press, 1946, Chapter 7. geneous, is so exceedingly sensitive to an increase
(4) V. K. LaMer and D. Sinclair, N. D. R. C.-0. S. R. D. Report
57, July 1, 1941-see Sinclair and LaMer, Chemical Reuiews, 44, 245 in the concentration of dissolved sulfur that the
(1949). rate becomes effectively infinite. For practical
(5) V. K. LaMer and M. D. Barnes, J. Colloid Science, 1, 71-79 purposes it may be considered a critical phenom-
(1946); I. Johnson and V. K. LaMer, THISJ O U R N A L , 69, 1184 enon. 1 ~ 2 This critical concentration consequently
(1947).
( 6 ) A. W. Thomas, “Colloid Chemistry,” McGrow-Hill Co., 1934, (7) LaMer and Kenyon, J. Colloid Science, S, 257 (1947).
p. 135; R. Audubert, in Alexander, “Colloid Chemistry,” Chemical (8)E. M. Zaiser and V. K. LaMer, ibid., 5, 571 (1948).
Catalog Co., New York, N. Y., 1926, Vol. I, p. 354, mentions a (9) H.Bassett and R . G. Durant, J. Ckem. Soc., 1401 (1927).
thesis b y Boutaric (Paris, 1918) in which i t is pointed out t h a t (IO) C.K. Jablcynski and 2. Warszawska-Rytel, Bull. Soc. Ckim.
large spheres transmit blue light better than red and on p. 361 France, 59, 409 (1926).
discusses the problem of particle growth in a preliminary way which (11) A. F. Holleman, Rcc. frao. chim., 14,71 (189).
proves t o be in accord with the treatment we present. (12) F.Prakke and E. Stiasny, ibid., 13, 615 (1933).
4847
4818 VICTOR K. LAMERAND ROBERT
H. DINEGAR Vol. 72

then decreasing the solubility (increasing the

chemical potential) either by lowering the tem-
perature or by adding a miscible non-solvent for
sulfur, for example, the addition of water to sulfur
dissolved in an organic solvent such as acetone or
alcohol.
XI’ one starts with a given concentration of sul-
fur in alcohol or acetone and adds water, the sys-
tem remains homogeneous until the critical addi-
tion is reached. By adjusting the initial concen-
tration of sulfur and the amount of water added
one should be able to produce a supersaturation of
sulfur that will meet the conditions described
above, i. e., produce just one brief outburst of
nuclei, leaving enough dissolved sulfur to produce
diffusional growth. A monodispersed prepara-
Fig. l.-%hematic representation of the concentration tion should result. Schematically, but not to
of molecularly dissolved sulfur before and after nucleation scale, this is shown in Fig. 2. The relationship
as a function of time. between initial concentration and the amount of
represents a limiting state of supersaturation and water necessary to start rapid nucleation is la-
is labelled C,, on the figures. The appearance of belled the “critical curve.”
nuclei in stage I1 partially relieves the supersatu-
ration, a t least rapidly enough that the rate of
nucleation falls almost immediately to zero, SUPERSATURATION
when the initial concentrations are in the range
CN.B~O~ and CHCI= 0.002 AI. Consequently, by
CURVE \ \ \ , \
controlling the initial concentrations of acid and SoLualLlTY~,

thiosulfate, the period in which repetitive nuclea-

tion can occur is made so short that a monodis-
persed colloid results by the subsequent uniform
growth on the existing nuclei. The supersatura-
tion level a t which the rate of formation of nuclei,
R,, becomes practically zero is shown at a small
but at present indeterminate concentration below
the point where R, is effectively infinite.
When the concentration falls below R,
system is still supersaturated, consequently the
-
0 the 1

VOLUME Ha0
FRACTlON -
process of growth of stable nuclei to form discrete
particles can proceed by diffusion of molecularly Fig. 2 -Schematic representation of the critical super-
dissolved sulfur to the nuclei. The concentration saturation and equilibrium saturation concentration of dis-
of molecularly dissolved sulfur, in this region (111), solved sulfur in an organic solvent-water mixture as a
represents a balance between the rate of produc- function of the volume fraction of water.
tion by the chemical reaction and the removal by Experimental Results.-The experimentally determined
diffusion to the nuclear sinks. The final rate of critical curve for sulfur dissolved in 100% ethanol is shown
growth is accordingly controlled by the rate of the as in Fig. 3, where the log of the final concentration is plotted
a function of the partial molal volume fraction of water.
chemical reaction. The sulfur used was purified by the method of Bacon and
If the initial concentrations are not dilute ( > Fanellils and dissolved in 100% c. P. ethyl alcohol (%%D
ca. 0.01 M> the rate of production of sulfur by the 1.3600).
chemical reaction will become so rapid that the point Two methods of noting the phase transition or critical

exceed the concentration at which Rn - were employed in the construction of this curve-
concentration of dissolved sulfur will continually a visual method and a turbidimetric analysis. In the
0. A first the diluted sulfur-ethanol solutions were titrated drop-
continual cascading of nuclei in addition to growth wise with distilled water with continuous stirring at room
(-25 ”) . The end-point in the titration was
then ensues. Obviously, then, the size of any temperature
taken to be that point a t which the further addition of one
given particle will depend upon when it was drop of water produced an immediate large increase in the
formed. Non-uniform growth and a polydis- turbidity of the solution. The end-point was observed
persed sol is the final result in this case. visually, the solution being illuminated by approximately
There is nothing in the foregoing treatment that parallel white light, oriented 90 degrees to the observer.
When the sulfur solution was titrated rapidly with water
limits the mechanism by which the value of C,, is three regions of different optical properties were noted
reached. Supersaturation can be achieved not depending on the initial concentration of sulfur. If the
only by building up the concentration of dissolved initial concentration of sulfur in alcohol was less than 1 X
sulfur in a medium of constant solubility-but also lo-’ g. at. S/1. no sharp end-point was observed-only a
starting with a certain Concentration of sulfur and 13) Hacon and Fanelli. I n d . Eng Chcm , 84, 1043 (1942).
Nov., 1950 THEORY,
PRODUCTION AND O F MONODISPERSED
MECHANISM HYDROSOLS 4849

slowly increasing turbidity was noted. If the initial con-

centrations were between 2.5 X lo-‘ and 2.9 X 10-8 g.
at. S/1. the appearance of a sharp end-point with a rapidly
increasing turbidity was observed. In this case, however,
the Tyndall beam, initially of a bluish white color, trans-
formed in a few minutes into the higher order Tyndall
spectra of angular red and green colors characteristic of a
monodispersed s y ~ t e m . ~The
, ~ number of red spectral
bands observed was 1 to 2 in the lower concentrations and
as many as 6 in the upper concentration range. The par-
ticle radii produced were, accordingly, 0.2 to 0.45 micron.
With initial concentrations of sulfur greater than 3 x
lo-* g. at. S/1. the end-pomt of the titrations was sharp and
a rapid increase in turbidity occurred but the color change
was from a bluish white Tyndall beam to a white opalescent
solution indicating polydispersity. No higher order “yn-
VOLUME FRLCTlON ”10-
dall spectra were observed. The opalescent color was
soon followed by the formation of large particles that Fig. 3.-Experimental critical supersaturation concen-
scintillated in the illuminating beam indicating the ap- tration of sulfur in ethanol-water mixtures as a function of
pearance of a crystalline precipitate. The amount of the volume fraction of water: 0,visual; a, spectrophotom-
crystalline material was very large when the initial con-
centration of sulfur was 5.79 and 9.37 X 10-3 g. at. S/1. eter.
Back titration with alcohol showed that the sulfur appear- these initial and secondary increases in turbidity are shown
ing in the colloidal form could be easily redissolved by the below for the run starting with 20 ml. of ethanol srdfur
subsequent addition of one or two milliliters of ethanol. concentration equal t o 5.88 X g. at. SA.
Since the end-point in the titration was chosen to be that is the partial molal volume fraction of water; V,,O (ini-
of rapidly increasing turbidity it is not surprising that only
in certain ranges of concentration was this criterion of a
tial) = 0.41; v& (secondary) = 0.59.
The end-points accompanying slow addition were not
sharp end-point met. The time rate of increase of tur- particularly sharp and no monodispersed sol was produced
bidity is dependent not only on the increase of particle in this manner.
number but particularly upon the scattering coefficient
which varies from the f 4 to -2.2 power of the radius The slow increase in turbidity before the point
of the particle.l4 Hence, one would expect in those re- of maximum condensation is predictable theo-
gions where both nucleation and appreciable growth retically. It exhibits analogies to the light scat-
occur, that a more rapid increase in turbidity will occur
than in the regions where nuclei production only is taking tering observed on concentrating cationic deter-
place. Diffuse end-points correspond to those regions gent solutions as reported by Debye. l6,I6
where only nucleation and little growth occurs, since the Debye has shown that it is a natural conse-
formation of nuclei has exhausted the necessary super- quence of the law of mass action for equilibrium
saturation. In the monodispersed region nucleation fol-
lowed by slow growth occurs with the added restriction between simple molecules and their aggregates
that no secondary nucleations take place. At high con- that a critical concentration exists-above which
centrations of sulfur, both nucleation and growth occur the predominant form is the aggregate or nucleus
rapidly but here there is such a high degree of supersatura- and below which the system is composed mainly,
tion a t the endpoint that Rn does not fall to zero rapidly
enough to avoid repetitive nucleation. This is the cus- but not entirely, of the molecular species. If one
tomary situation leading to the production of “white recognizes that the addition of water releases sul-
milk of sulfur.” fur by decreasing the solubility in the medium
The appearance of increasing amounts of crystalline sul- and that the formation of the nucleus is the aggre-
fur in the same solutions which initially showed a white
opalescent colloid (polydispersed amorphous particles) gation of such molecules for which the over-all
is attributable to the relatively high concentration of dis- equilibrium n ( S ) i3 (S), holds, it can be shown
solved sulfur. These concentrations not only produced that small numbers of small particles, capable of
large numbers of amorphous droplets along with repetitive scattering light should appear a t volume fractions
nucleation, but also an appreciable number of crystalline
nuclei. The crystalline nuclei and supercooled liquid
particles then collide by a mechanism similar t o that of
coagulation, the crystal nuclei crystallizing the super-
of water below that for which the given amount of
sulfur is a t a critical concentration R, ob. The-
rate of increase of the number of particles with in-
cooled liquid droplets upon contact. The higher the in- creasing amounts of water is small below the criti-
itial concentration of sulfur the greater are all these effects
until it becomes impossible to prepare a sol that is stable cal point leading to a corresponding small rate of
for any appreciable length of time. At concentrations in increase of turbidity. At the critical particle
the range > 10-2 g. a t . S/l., stable sols could not be pre- formation point, however, the rate of increase of
pared-the dispersed droplets of any sol so prepared crys- number becomes exceedingly large and the addi-
tallized in a very few minutes.
If the titration was carried out by a slow addition of tional growth of the particles by diffusion leads to
water, a different result was obtained. The first additions a very rapid increase in turbidity.
produced no visible effect but with increasing amounts the Solutions prepared in a similar manner were
solution slowly became turbid. The turbidity of the solu- subjected to a turbidimetric analysis. This con-
tion then seemed to remain almost constant on further sisted of adding known amounts of water to sul-
addition of water until the volume fraction of water in the
mixture corresponded to that obtained when the water fur alcohol solutions and after several hours in a
was added rapidly. A few milliliters beyond this point the thermostat a t 25’ (10.02’) the resulting mixtures
whole solution became very turbid and opaque even to were examined in a Beckman spectrophotome-
indirect light. The amounts of water necessary t o cause ter, model DU, using 10-cm. quartz cells. The
(14) V. K. LaMer, J . Phys. Colloid Chcm., 52, 65 (1948), and ref. 4, (15) Debye, Annals of N . Y.Acad. of Sciences, 51, 575 (1949).
Chem. &os. (fig. 5 ) . (16) Debye, J . CoZZ. Sci., 8, 407 (1948).
.IS% VICTOR K. LAMERAND ROBERT11. D I N E G A R 1'01. i?

optical density of each sample was measured as a it is to be expected on this basis that a sulfur con-
function of wave length a t 250 A. intervals be- centration of less than 1.5 X lo4 g. at. S/1. could
tween 4000 and 6000 A. The optical density as a not result in any perceptible turbidity a t any vol-
function of the volume fraction of water in the al- ume fraction of water attainable. Extrapolation
cohol-water solution for t k e e runs is shown in of the critical curve obtained by the present
Fig. 4,for wave length 5000 A. method gives 1 X lom6g. at. S/l. for the critical
concentration of sulfur in pure water. This figure
12 is in satisfactory agreement with the value given
by Zaiser16 when it is recognized that her deter-
10 mination was made spectrophotometrically as-
suming the extinction coefficient of sulfur in pure
water to be the average of those of a number of or-
ganic solvents.
5 Rate of Growth and the DiiTusion Coefficient of Ss:
(1) Experimental.-In order to study the rate of growth of
monodispersed sols produced by dilution with non-solvent
of sulfur in ethanol and acetone, another method of pre-
paring the sols was devised. It consisted of adding to the
1 sulfur-ethanol or sulfur-acetone solution by an hypoder-
mic syringe the amount of water needed to cause nuclea-
i tion. This technique afforded more instantaneous and
2 homogeneous inking. The heat of mixing, however,
raised the temperature to about 31 '--enough t o prevent
s" 0.5 immediate nucleation. The mixture was chilled in 90
seconds with continuous stirring to approximately 29 '.
Nucleation took place between these temperatures and the
particles grew uniformly producing monodisperse prepara-
tions. The slight rise in initial temperature prevented
0.1 iiucleation until all the water had been added thus insuring
homogeneity and the liberation of sulfur in equivalent
amounts a t all points in the solution. The solution was
then put in a constant temperature room a t 25". The
maximum particle size was reached in about 25 minutes
0.05 during which time thz temperature of the mixture had de-
creased only about 2 . Very little, if any, change in size
occurred as the sol reached equilibrium a t 25" some 70
minutes later, indicating that the temperature coefficient
of the solubility of sulfur is too small to be noted by the
0.00 accuracy of the method uwd, for particles of size greater
than 0.4 fi radius.
0.2 0.4 0.6 08 1.0 The method used for determining the size of the particles
Volume fraction H20. was developed by LaMer and Sinclair for aerosols4 and
applied to hydrosols by Johnson and LaMer.'? It utilizes
Fig. 4 -Optical density (log 10,'I) of sulfur dissolved the change in angular position, 6, of the red bands of the
in ethanol-water mixtures as a function of the volume higher order Tyndall spectra with particle size. The
fraction of water Initial concentration of sulfur gr. a t 11.: angular positions of the red bands were determined as a
A, 37.5 B, 588 C, 1.18 (IO)-'. function of time by viewing the sol itlumindted by ap-
proximately parallel white light through a telescope on
Log Io/I W. volume fraction of water showed a which wds mounted a Polaroid H filter No. 330, oriented
so as to pass only the scattered light whose electric vector
region in which a few small particles were pro- is perpendicular to the plane determined by the light
duced followed by another region in which the op- source, particle, and observer. 6 is the angle of observa-
tical density increased tremendously on addition tion measured from the backward direction of incident
of a small increment of water (see Initial C = light propagation. The calibration'" has been determined
by transmission methods, theoretical calculation, the rate
5.88 X lo-* g. at. S/l.). In this region rapid par- of deposition of sulfur hydrosols in aqueous solution and
ticle formation is accompanied by rapid growth. rechecked by Kerker and LaXerls for dependence of ra-
The critical points obtained by this method fell dius upon time of reaction.
on the same curve as those obtained by the visual The number of particles was determined from the wave
length and intensity value of the minimum in the optical
method and are shown by appropriate symbols in density us. wave length transmission curve measured by
Fig, 3. The curve could not be extended by using a model DU Beckman spectrophotometer, using 1O-cm.
this method to volume fractions of water greater quartz cells. The value of the optical density at the mini-
than 0.8. Initial sulfur concentrations as lorn as mum in the transmission curve is related to the number of
particles, n, having a radius, I, by the relationshipls~aO
1.5 X 10" g. at. S/l., were tried and no change in
turbidity was noticed even on tenfold additions of 2 3 log (10/1) min = (rr2nZ)-'R, =: 2.08
water volume fractions. This corresponds to a where 2rr/h' min. = 6 8 for X' > 3200 A., X' = wave
final sulfur concentration of less than 1.5 X 10-s length of the light in the medium, 1 = cell length = 10 cm.
g. at. S/1. Since it has been reported previously* (17) 1. Johnson and V. K. LaMer, THIS JOURNAL, 69, 1184 (1947);
that a sulfur concentration of 3.2 X lo4 g. at. S/1. (18) M.Rerker and V. K.LaMer, ibid., 72, 3516 (1960).
(IS) M. D. Barnes, A. S. Kenyon, E. M. Zaiser nnd V. R.Lahler,
must be generated by the reaction of acid and thio- J . Colloid Science, a, 349 (1947).
sulfate in water for light scattering to be observed, (20) Kenyon and LaMer, {bid , 8, 183 (1949).
Nnv., 1950 THEORY,
PRODUCTION
AND MECIIANISM
OF MONODISPERSED
HYDROSOLS 4551

0.110
0.105
0.100
$
CT“ 0.090
s” 0.085
0.075
0.070 L- ~~ ~

400 450 500 550 600 650 700 750

Wave length in millimicrons.
5 10 20 30 40 Fig. 6 . 4 p t i c a l density (log Io/l) us. wave length for
Time in minutes e. two monodispersed sulfur-organic solvent-water sols :
Fig. 5.-Angular position of maximum red orders as a 8, alcohol sol, plotting increment = +O.Ol; I = 0.36
function of time for a monodispersed sol prepared by the micron, n = 1.93(10)6 cc.-l; 0, acetone sol; I , 0.43
addition of water to sulfur dissolved in ethanol. micron; n, 1.5e(10)6cc.-l.

The graph of theta ws. time in Fig. 5 is very similar t o problem has been treated by R e i ~ s . ~ ~Special-
,~*
those obtained previously for sols made by the acid de- izing this treatment to the case at hand, one may
composition of thiosulfate, except that the number and derive on theoretical grounds alone, an equation
position of the red orders reaches a constant value signify- giving the radius of the particles as a function of
ing the exhaustion of the liberated sulfur. The optical
density ws. wave length curve shows the characteristic the time after formation of nuclei.
“minimum” (Fig. 6). The numbers of particles per ml. Consider the situation within a very short inter-
calculated from these sols are in the same range as those val of time after mixing the solutions. Stable
obtained by acid decomposition, i. e., 1 to 4 X 106 per particles of super-cooled liquid sulfur have formed
ml. The final radius, number of particles per ml., along
with the initial concentration of sulfur and volume of in the midst of a mixture of organic solvent and
water added for fifteen runs for the solvents ethanol and water that is supersaturated with respect to dis-
acetone are shown in Table I. solved sulfur. Further deposition of sulfur on the
The sign of the electric charge on the particles of sols particles is a natural process and occurs spontane-
prepared by dilution methods is either positive or negative
depending upon the pH of the solution. For alcohol- ously, The final size to which the particles will
water sols the iso-electric point is a t PH 5.8 and for ace- grow and the rate a t which they will grow depends
tone-water sols, at pH 6.0, both sols being positive a t a upon three variables.
PH lower than the isoelectric point. The sign of the The first is the number of nuclei which are to
charge is reversible to titration with either HCl or NaOH.
The electrophoretic mobility of a n alcohol-water sol a t grow. This number is assumed constant through-
pH 6.9 of radius 0.5 p was -1.3 * 0.3 p/sec./volt/cm. out the growth and equal to the number of parti-
This corresponds t o a charge of 1500 electrons per particle, cles present at any time. Since the total amount
or a surface density of charge equal to 23 abs. e. s. u. cm. -2. of sulfur to be deposited is distributed homogene-
These electrophoretic experiments were performed jointly
with Robert H. Smellie, using apparatus he developed. ously throughout the solution one must propor-
He finds that traces of acetic acid produced by oxidation of tion l / n times this amount to each nucleus, where
the alcohol are responsible for the positive charges on the n equals the number of particles per ml. Then
sols below pH 5.8. each particle is influenced only as to growth by an
Theory of Growth.-The growth rates of sols amount of1 sulfur contained in a volume, spheri-
produced by the dilution method represent the dif- cally symmetric about the origin, of radius equal to
fusion of a substance in a thermodynamically un- lz and volume 4/3nh3. This “particulate volume”
stable state onto the surface of a nucleus. The associated with each nucleus equals l / n . Since
diffusing species are almost certainly S g rings,21,22 no material outside of this volume influences the
uncontaminated in the case of the dilution method growth of any one given particle, h is the radius of
by any other substances such as polythionates. an impermeable shell.
It is possible to gain information concerning the The second variable influencing the growth of
diffusion coefficient of SSfrom such measurements. the particles is the total amount of diffusible sul-
The results will be approximate, since the theory fur. Provided there are no other sources of dif-
used is not exact. In addition, certain approxi- fusable sulfur, this value is given by the difference
mations and assumptions concerning the experi- between the amount of sulfur originally in the
mental data are necessary. mixture and the solubility of supercooled liquid
The mathematical formulation of the diffusion sulfur in the medium, e. g., (ess- Cs). Unfortu-
(21) W. H. Latimer and J. Hildebrand, “Reference Book of Inor-
nately C, is not known independently. The dif-
ganic Chemistry,” The Macmillan Co.. New York, N. Y., Chapter (23) H.Reiss, Trans. N . Y. Acad. of Scicirccs, Series 11. Vol 11,
XII. No. 7 (1949).
(22) L. Pauling, Proc. Not. Acod. Sciences, 86, no. 9 (1949). (24) H. Reiss and V. K. LaMer, J . Chcm. Physics, 18, 1 (1950).
4852 VICTORK. LAMERAND ROBERT
H. DINEGAR Vol. 72

ference (Css - C,) = COis, however, obtainable Integration of equation 3 gives

from the size and number of particles after diffu-
sional growth has ceased. Only COwill enter into
the growth equation and its value may be esti-
.W [ = - 4m
1
Q ( t ) ] : 4- QWdt

mated. From condition (2), f(0) equals CSs. From equa-

Actually, COis a function of the temperature tion (l),Q(0)equals 0
since C, is dependent on the temperature. The
temperature coefficient of solution of sulfur in the .‘. f(t) = C,, - -
4rhD
1
Q(t) + 4-
3
rh2 Q(t)dt
medium has been estimated theoretically (calcu- or from equation ( 2 )
lations based on the change of mole fraction with
temperature as related to the heat of fusion) as
1.9% per degree. Experiments substantiate this
value approximately. This change of solubility
with temperature requires that the final Co be cor-
1
since h > > r , -
a
i
< < -1r
rected for the small temperature drop during
growth, by writing Co(t) = C’o(0) +
f ( t ) . The
value of Co(t)a t any time is obtained by construct-
:. C(7,t) = c., + 4rDr
-Qt
+ 4rh3
-3 8
Q(t)dt (4)
ing an average cooling curve and adding to C O ( O ) Since the particle is increasing with size as ma-
the correction f ( t ) for the sulfur liberated by the terial is deposited on its surface, -Q(t) must be
drop in temperature up to the time t. The effect proportional to the rate of increase of volume,
of the temperature drop is the same as if there were i. e.
a reaction generating sulfur in addition t o the
amount originally set free by the addition of wa-
ter. The equation taking into account C, as a Depending on the units used the proportionality
function of time (temperature) is identical in form constant varies. For Q(t) in g. set.-' and x in
with the equation considering C, as a constant but cm., ?c = radius of the particle
with a source function k present in the system
(cf. ref. 24). -Q(t) = 4 r p x 2 d x l d t (5)
The third variable is the diffusion coefficient, D , p = density of diffusing sulfur. Equation (4)
of sulfur in the medium. D in turn depends. in- and ( 5 ) then give
versely on the viscosity and directly on the tem-
perature. These temperature coefficients must C(7,t) = Ce, - PD- .x2 d x / d t ( l / r ) -
also be neglected. I n addition their direction of
change is opposite to that of (CSs- Cs) = Co and
3;,?rh3 (I 4 ~ dx~ dt x ( 62) ~ ~

since the product DCOwill always be dealt with,

Y

Evaluating
the variations will tend to cancel one another.
Consider the flux, Q, through any spherical dx
x2 17dt
shell of radius, r , concentric with spheres of radii x
and It (r equals radius of the nucleus). Assuming with x ( 0 ) = 0, and for r =x we obtain
Fick’s law; this flux is given by px dx px3
Q(t) = -4rrZD dcldr (1)
C ( x , 1) = c,, - Ddt--F (7)

subject to the boundary conditions: (1) a t the From condition (1) C(x,t) = C s ( t ) ; therefore the
surface of the particle the value of the concentra- final equation becomes
tion of sulfur is the saturation value; i. e., C(x,t)
equals C, ( t ) ; (2) a t any point in the system, ini- ‘w)
dt
= [C,
BD
- C, i t ) ] --
p
- 2h2-0 x 3 (8)
tially, the concentration of sulfur is the supersatu-
ration value, i. e., C(r,O) equals Css; ( 3 ) a t the This equation may now be integrated25in closed
impermeable shell a t h: (dc/bt),=h = 3Q(t)/ form to obtain the analytical expression between
4nh3. Integrating (1) with respect to r gives x and t only if C, is considered independent of t
Thus the final result for this special case is
C(r, t ) = Q ( t ) / 4 r r D +f(t) (2)
A t time t equals 0, equation (2) becomes
C(r,O) = Q ( 0 ) l B r r D 4f ( 0 )
but, since (dc/&)i=o = 0, Q(0) = 0, and C(r,o)
equalsf(0) equals C,, by condition ( 2 ) .
Boundary condition (3) demands that the time
derivative of equation 2 specialized to 7 equal h be
given by 3Q(t)/4nh3
Since it is possible to use equation (9) only for a
calculation of x versus t when COis constant and it
( 2 5 ) Pierce, “Short Tables of Iotegrals.” 3rd Edition, Eqn. 61.
Nov., 1950 PRODUCTION
THEORY, AND MECHANISM
OF MONODISPERSED
HYDROSOLS 4853

was desired to take into account the small depend- TABLE I1

ence of COon t as previously described, equation SULFUR I N ACETONE, 25-CC. SAMPLES USED
(8) was integrated numerically by the method of %
sulfur
Cauchy and LipsChitz.26 104 in
All quantities in this equation are experiment- C’SS HzO Radius 106 co cob C d
gr. at. added, (max.), 10 gr. at. loidal (csa -
ally measurable except one-the diffusion coef- No. S/1. cc. microns n/cc. S/1. form Co)
ficient, D. One may then choose values of D and 10 5.00 42.0 0.47 0.83 2 . 2 12 1.12
see if any value will give a theoretical curve of x 11 5.00 42.5 .43 1.15 2 . 4 13 1.13
vs. t that reproduces the experimental curve. 12 5.00 42.5 .49 0.85 2.6 14 1.14
The equation also depends upon the number of 13 5.00 42.5 .56 .70 3 . 2 17 1.17
particles, n. If one varies n experimentally and 14 5.00 43.0 .48 1.17 3 . 4 18 1.18
then finds that the same D must be used in each 15 5.00 44.0 .45 1.56 3 . 7 20 1.20
case to reproduce the experimental curve, the Av. 1.16
value of D so obtained takes on added signifi-
cance. I n addition, since D varies with the vis- NO. 9 11 12 13 14
cosity, it is important to see if sols prepared in Viscosity, c. p. 1.3
media of different viscosities would give growth 106D, sq. cm./sec. 5.9
curves that predict the same value of D when re- 106D0,sq. cm./sec. 7.7
ferred to a standard viscosity. to, min. all zero
The rate of growth curves are shown in Fig. 7 c a s = C’m (1 - v,,,
for sols prepared by the previous method from 10-6 cm.2 sec.-l for the acetone-water sols in
100% ethanol and acetone-the viscosities of good agreement with Do for substances of com-
these two media being significantly different. It parable molecular weight, which lie in the range 4
was found that the experimental curves could be to 7 X ann2 se~.-’.~’
reproduced by equation (8).
The values of the diffusion coefficient necessary / l I I I l 1 l l

are shown in Tables I and 11. The time, to, is the

time of formation of nuclei as predicted by the
theoretical equation. It is seen a t the bottom of
the tables that the starting times are within the
quenching periods with the majority of them being
f 0‘7
0.6 1 0
zero. The number of particles varies over a six-
fold range and the viscosity by a factor of 50%.
When referred t o a standard viscosity of 1 centi-
poise, the calculated value of DOis 7.2 X
cm.* sec.-l for the alcohol-water sols and 7.7 X
TABLEI
SULFUR
IN ETHANOL,
25-ML. SAMPLES
USED
% 0 5 10 15 20 25 30 35 40 45
Sulfur
104 in Time in minutes 4.
Cls0 Radius 105 co col- C d
No.
gr. at.
S/l.
H20
added,
cc.
(max.),
microns
106
nlcc.
gr. at. loidal
2/1. form
(Cm -
Co)
Fig. 7.-Growth rates of sulfur particles in terms of
radius (microns) us. time (minutes) prepared by adding
1 3.00 46.5 0.39 1.55 2.4 22 1.22 water to ethanol or acetone sulfur solutions. The circles
2 3.00 46.5 .30 3.18 2.2 21 1.21 represent experiment points, the solid line representing
3 3.00 46.5 .39 1.60 2.5 23 1.23 the theoretical calculation using diffusion coefficients:
4 3.00 46.5 .36 1.93 2.3 22 1.22 D = 3.5(10)-s cm.2 sec.-l for ethanol solutions (A);
5 5.88 36.0 .40 1.40 2.4 11 1.11 D = 5.9(10)-6 sec.-l for acetone solution (B); cf.
6 6.00 34.0 .37 3.05 4.0 16 1.16 Tables I and 11.
7 6.00 35.0 .47 1.00 2.7 12 1.12
8 6.00 38.0 .41 1.97 3.6 18 1.18 The sensitivity of the theoretical equation to
9 6.12 37.5 .37 4.40 5.8 24 1.24 the value of D chosen is seen in Fig. 8. In this
Av. 1.19 case the value of D giving the best agreement be-
tween experiment and theory is 6 X
No. 1 3 4 6 9 sec.-l. Using a value of D = 3 X cm.2
Viscosity, c. p. 2.1 sec.-l and 12 X set.-' give theoretical
106D, sq. cm./sec. 3.5 curves corresponding to the broken lines. It is
WDo, sq. cm./sec. 7.2 seen that variation by a factor of only 2 or ‘/2 in D
to, min. 1.5 0 0 2 1 gives theoretical curves that are markedly differ-
cs0 = C’, (1 - v,,o, ent not only in shape, but in the region of the
(26) Ford, “Differential Equations,” McGraw-Hill Book Co.,
most rapid change in slope.
New York, N. Y., 1933,p. 161. (27) “International Critical Tables,” Vol. V, pp. 63-76.
 RORRRT13. DINEGAR Vol. 72

ments of the particle to be formed-the ratio in-

0.7 -
1
creasing with an increase in both these effects.
To test these ideas and to obtain an estimate of
f 0.8 I- the importance of each, the supersaturation re-
:
2 0.5
quirements for the sparingly soluble sulfates of
barium, strontium and lead have been measured.
CJ
A comparison of barium sulfate with sulfur gives
'gc 0.4 an idea of the dependence of the supersaturation
.H
ratio upon the surface tension and the effects of
crystal orientation whereas a comparison of stron-
tium and lead with barium gives the magnitude
of the effects caused by solubility alone, since the
other effects should be approximately equal. The
_.I 1 I _ I - ! I 1 results are discussed in a following paper.
0 5 10 20 3 :10 35 41) 4.;
13 We should like to acknowledge the suggestion
Time in minutes +. and help of Professor Howard Reiss of Boston
University in calculating the diffusion coefficient
Fig. 8.--Effect of varying the value of D i n lo8 tin.* from our data.
sec.-l units employed in theoretical calculation for the
growth of the sulfur sol in acetone-water solution: A = 12; Summary and Conclusions
B = 6 ; C = 3; 0 = experimental values.
1. Monodispersed colloids exhibiting higher
General Considerations.-From the amount order Tyndall spectra have been prepared by pre-
of sulfur dissolved initially in the organic solvent cipitation of sulfur from ethanol and acetone solu-
and the amount appearing in the colloidal form, tions by the careful addition of water.
one can estimate the extent to which the initial 2 . The relationship between the concentra-
solution was supersaturated and calculate the tion of sulfur and the volume fraction of water a t
supersaturation ratio C,,/C, even though C, is not the critical point of sulfur-ethanol-water solu-
known independently. Where CO = (4nrapn)/ tions has been investigated.
96 = 0.262 nr3 gr. at./cme3,C, = (CSs- CO)and 3. Sols prepared by this dilution method have
C,,/C, = Css/(Css- CO). It is seen in Tables I been investigated as to dispersion of wave length
and I1 that approximately 10 to 20% supersatura- in their angular scattering of visible light and
tion is needed to produce a monodispersed prepa- their transmission as a function of wave length.
ration. Nucleation at this small degree of super- 4. The size, number and charge of the parti-
saturation is undoubtedly aided by the formation cles of such sols are in the same range as those pre-
of the minute cavities during the syringing and pared by the acid decomposition of sodium thio-
mixing of the solutions, but i t is also a direct con- sulfate.
sequence of the relatively high concentration of 5. The rates of growth of sols prepared by such
dissolved sulfur at the phase transition (-lo-* g. a dilution method have been investigated. The
at. S/l.) in organic solvent-water as compared growth rates obtained have been reproduced from
to the concentration (- lod g. a t S/l.) in water in previous theoretical considerations which allow
the acid-thiosulfate decomposition method where the estimation of the value of the diffusion coef-
the supersaturation value is higher. It is because ficient of S g in alcohol-water and acetone-water
nucleation can be accomplished at this small de- mixtures. The value of this diffusion coefficient
gree of supersaturation that a monodispersed sys- lies in the range '7-8 X cm.2/sec. for both
tem does result. Obviously, the older method, media. The application of this method for esti-
namely, that of pouring a small amount of sulfur mating dzusion coefficients is valid for any col-
concentrated in an organic solvent into a large loidal system that is characterized by a small par-
volume of HzO, should not, and never does, pro- ticle size distribution a t all stages of its growth.
duce a monodispersed colloidal dispersion. 6. A qualitative explanation is offered for the
The degree of supersaturation necessary for a necessary conditions under which monodispersed
phase transition to occur should depend not only colloids may be prepared for both the dilution and
on the total concentration of material present but acid decomposition of sodium thiosulfate methods.
also on the surface tension and orientation require- NEW YORK,N Y. RECEIVED
MARCH
29, 1950