Proc. Nat. Acad. Sci.

USA
Vol. 68, No. 7, pp. 1569-1571, July 1971

Effect of Gravity on Osmotic Equilibria

(osmotic pressure/buoyancy)

P. F. SCHOLANDER AND M. PEREZ

Scripps Institution of Oceanography, University of California, San Diego, La Jolla, Calif. 92037:
and Instituto De Pesquisas Da Marinha, Brasil
Contributed by P. F. Scholander, May 13, 1971

ABSTRACT Experiments with colloidal suspensions of tight film rests loosely on the surface, exerting no force on it.
oil and iron oxide at equilibrium show that the buoyancy When the column is turned upside down, rigidity of the cup
of the suspended particles is additive to the osmotic pres-
sure and that the relation remains with the free surface and cohesion of the fluid always maintain zero (ambient) fluid
and not with the membrane. The experiments contribute pressure at the surface, irrespective of the water tension in the
to the general concept that osmotic pressure is caused by solution.
the dispersal pressure of solute molecules and that the The bottom of the osmometer is connected with a capillary
osmotic interaction with the water at equilibrium is due
solely to a coupling at the free surface. E, which is lined up with the axis of rotation, and a simple
water monometer (not shown in Fig. 1) balances the pressure.
When a semipermeable membrane stops the advance of dif- The meniscus is contrasted with a black and white strip of
fusing solute molecules, it becomes stretched by the dispersal paper behind it and is read by a travelling microscope. All ma-
pressure. Similarly, when the solute is stopped by a free sur- nometer readings are corrected for a capillary blank E. When
face, balance of forces maintains the water under a negative the system is charged with water and inverted, one observes a
pressure that is equal and opposite to the dispersal pressure. slight displacement of the meniscus due to twist in the con-
This lowers the vapor pressure according to Poynting's rela- necting tube, but the balancing pressage remains the same,
tion (1) and is the essence of osmotic phenomena (2, 3). From within a fraction of a millimeter.
this point of view it follows that any factor which influences Experiments with light colloidal suspensions
the dispersal pressure in addition to nRT will change the col- When "soluble oil" (Standard Oil Co.), density 0.92, is mixed
ligative properties. Such effects may be induced by electro- with water it forms a milky suspension which shows vigorous
static, magnetic, and gravitational forces, and they are com- Brownian motion and does not visibly separate at 20 X g.
mon in colloidal solutions through interaction between the
particles themselves. Thus, when a colloidal suspension loses
water the dispersal pressure greatly accelerates by simple
mechanical crowding. Similarly, when the particles are re-
pelled by charges or undergo conformational changes that
involve their size, it will have a sensitive readout in the water
tension.
We demonstrated recently that the osmotic pressure of a col-
loidal magnetic solution relates to the free surface, not to the
membrane. A magnetic field which pulled the colloidal parti-
cles toward the free surface increased the osmotic pressure,
whereas a field pulling the particles toward the membrane
lessened or even eliminated it. It was shown that the magnetic E
C --C
vector was directly additive to the osmotic pressure (4). We
report here on the effect of acceleration on colloidal suspen-
sions with either positive or negative buoyancy.
METHOD
The osmometer (Fig. 1) consists of a dialyzing membrane
which is stretch-mounted (5) at the bottom of a 21-cm deep
cup A. The rigidity of this mounting permits measurements of
A
a few centimeters of water above or below ambient with an
accuracy of ± 1 mm of water. Thermal disturbances are small
because of the minute volume between membrane and menis-
cus. The cup is mounted in a pivot arrangement B so it can be FIG. 1. Pivoting osmometer. Right: upright position. Left:
inverted around a horizontal axis C-C, which is exactly in line inverted position, drawn displaced to the left. A, location of
with the free surface. The cup is filled to the rim and the membrane; B, pivot arrangement; C--C, axis of rotations; D,
surface is covered with a thin film of Saran wrap, which is Saran film clamped loosely on the surface; E, meniscus read with
clamped down by a ring and a rubber band D. This vapor- travelling microscope. C, left, connection to manometer.
1569
1570 Physics: Scholander and Perez Proc. Nat. Acad. Sci. USA 68 (1071)
CM
lw 9 CM
-4

-5 rI
2 Iw
o -41N~~~~~~~~
40r
I I I
-6 n , U I n U n U
-2 o 20 40 60
MINUTES
FIG. 4. Gravity effect on osmotic pressure of colloidal iron
oxide suspension. Ordinate: true balancing pressure in centi-
-3 meter of water. U cup upright, n cup inverted. Buoyancy
U _~ * I Ue p p difference, 11 mm.
_
A eo
4 L __ a *J L.0_ X
40

_ g z
u
5a
_
n ,
r----m
u ,
I----n
n , u
~r-e- -
, n pears. The cup is returned to the upright position, the manom-
eter is reset to Pi and balanced for 10 min, and the sequence is
repeated.
RESULTS
)_I I_
-IC In Fig. 2 are given the results with three different suspensions.
llI It will be seen that the osmotic pressure is greater in the
e I_@ L an@-
, u
@_
upright position than in the inverted position. With the free
-
l , , p ll
U fl surface facing up, the buoyant force of the suspended particles
adds to the thermal dispersal pressure and lowers the hydro-
0 20 40 60
MINUTES static pressure of the water accordingly. With the surface
FIG. 2. Gravity effect on osmotic pressure of colloidal oil facing down, the buoyant force subtracts from the dispersal
suspensions. Concentrations of oil by volume: 20, 33.3, and 50%. pressure and the partial pressure of the water becomes less
U surface up, c horizontal, and n down. Ordinate: true balanc- negative. With the cup in the horizontal position (middle
ing pressure in centimeters of water. Buoyancy difference in panel), the buoyant forces directed toward or away from the
upper panel is 7 mm. free surface is zero, leaving the osmotic pressure such as it
would be in an orbiting laboratory. A shallow layer of sus-
Six different suspensions were used, the proportion of oil pension over the membrane gave a similar figure.
being varied between 20 and 50% by volume. They were all It will be seen from Fig. 3 that there is a near proportional-
sonicated. Columns of two of these suspensions, 50 cm tall, ity between the buoyant force and the concentration of oil in
were thermally insulated and left overnight. When samples the suspensions, and it is hardly a coincidence that the buoy-
from top and bottom were dried at 40'C they left the same ancy of the suspended particles is close to that which one may
amount of oil, within a few per cent. The osmotic pressure calculate from bulk oil contained in a cylinder twice the
varied with the degree of sonication and showed a slow de- height of the cup. So, as in previous, magnetic experiments, it
terioration over days, but neither of these shortcomings are seems that a known experimental vector adds directly to the
detrimental to our buoyancy experiments. Stirring proved osmotic pressure by coupling at the free surface.
unnecessary. Experiments with a heavy colloidal suspension
Procedure We are using the same colloidal solution as described in our
The osmometer cup is charged with suspension and the Saran magnetic experiments (4). The suspended material is a heavy
film is clamped on, air bubbles being avoided. The balancing iron oxide which gives a density of 1.08 to the suspension. As
pressure P1 is determined by adjustment of the manometer expected, it was found (Fig. 4) that the buoyancy effect was
until no drift of meniscus E is discernible. After the reading opposite to that in the oil suspensions. Thus, when the osmom-
has been steady for 10 min, the cup is inverted and the hair eter was turned with the surface facing down, the osmotic
line is moved back to the meniscus. The new level P2 is found pressure increased, i.e., the buoyancy vector added to the
by compensating, minute by minute, the drift until it disap- thermal dispersal pressure. Again we see that the osmotic
effect relates to the free surface and not to the membrane.
MM
20
H.2O
2 , . | ' I
CONCLUSIONS
Experiments with colloidal suspensions of particles with posi-
tive or negative buoyancy demonstrate that the osmotic
10 _ _ pressure is always uniquely related to solute pressure acting
against the free surface and bears no direct relation to forces
acting on the membrane. This is supported by magnetic ex-
0 10 20 30 40 50 periments (4); the other side of the coin, namely gravitational
% OIL IN SUSPENSION independence of the solvent has been separately verified (2).
FIG. 3. Buoyant force in 14 different suspensions. Dashed In accord with the classical studies by Perrin on the
line is the buoyancy calculated from the sum of all oil particles Brownian motion (6) we may conclude that osmosis in
suspended beneath the surface of a 42-cm-deep column, with the general is driven by the dispersal pressure of solute molecules
density of the oil 0.924. which, through impact with the free surface, lowers the hydro-
Proc: Nat. Acad. Sci. USA 68 (1971)
static pressure of the solvent. Translational solute-solvent
interaction (drag) belongs uniquely with osmotic flux and dis-
appears when the system arrives at equilibrium, whether or
not this entails gravitational or other concentration gradients.
This study was supported by National Institutes of Health
Grant No. Gm-10521 and National Science Foundation Grant
G-24831. We express our appreciation for stimulating discussions,
in particular with Drs. Y. C. Fung, H. T. Hammel, A. R. Hargens,
A. Yayanos, and Mr. Bent Schmidt-Nielsen.
Gravity and Osmotic Equilibria 1571
1. Poynting, J. H., Phil. Mag., 5, 12 (1881).
2. Scholander, P. F., Microvasc. Res., 2, 1 (1971).
3. Scholander, P. F., Topics in the Study of Life: The Bio Source
Book (New York, 1971), pp. 138-147.
4. Scholander, P. F., and M. Perez, Proc. Nat. Acad. Sci. USA,
68, 1093 (1971).
5 Hargens, A. R., and P. F. Scholander, Microvasc. Res., 1, 417
(1969).
6. Perrin, M. J., Ann. Chim. Phys., Ser. 8 (1909), Transl. by
F. Taylor, London (1910).