Beruflich Dokumente
Kultur Dokumente
SURFACE CHEMISTRY
CAMBRIDGE
UNIVERSITY PRESS
LONDON Fetter Lane
:
NEW TOEK
The Macmlllan Co.
MADBAS
Macmillan and
Co., Ltd.
TORONTO
The
Macmillan Co.
Canada, Ltd.
of
TOKYO
Maruzen-Kabusliiki-Kaisha
AN INTRODUCTION TO
SURFACE CHEMISTRY
By
IN PHYSICAL CHEMISTRY
CAMBRIDGE UNIVERSITY
CAMBRIDGE
AT THE UNIVERSITY PRESS
1926
EMERSON.
^^
^^
IN GREAT BBITAIN
PREFACE
By
Professor F. G.
DONNAN
THE
surfaces,
"
lies
PREFACE
VI
Since this
is also
and
colloid
advancing
field of
physico-chemical science.
F. G. D.
UNIVERSITY COLLEGE,
LONDON.
April, 1926.
INTEODUOTIO'N
between the properties of matter when
in bulk and in the colloidal state were first described by Thomas
essential differences
JL
clearly enunciated
thermodynamic
for his
stimulating
ERIC
K.
RIDEAL
CONTENTS
OHAP.
I
II
....
PAGE
1
'
31
III
63
IV
LIQUID-LIQUID INTERFACES
95
V THE
VI
VII
GAS-SOLID INTERFACE
123
166
.
EMULSIONS
IX
257
INDEX OF NAMES
INDEX OF SUBJECTS
208
SUSPENSIONS AND
....
298
330
.
334
CHAPTER
The
rapidly but continuously one into the other, so that the boundary
would have to be considered as a more or less arbitrary mathe-
lamina of
its
new
any way
it
be
altered,
primarily
is
size,
it were surrounded by an
tendency to contract. In a soap-film the force
of contraction may be measured directly by suspending from a
balance-pan a light square framework hanging in a vertical plane
so that the lower horizontal bar is immersed in a soap solution and
the part between the surface and the upper bar is covered with a
The
The
pull
is
tension and
is
force exerted
named
the surface
on a straight
line of
we
usual unit
weight per
is
mm.
The
The
mgm.
is
film
is"
calculated
We
The numerical
is
same
clearly the
The dimensions
as
r]\TogQ~|
.
i^p
j2
minimum, the
surface
must contract
to a
minimum
area:
we
are
surface
cally,
is
is
a quantity of heat
da
T-j^,
cLJ.
is
which
to
-
compensate
fresh surface is
this quantity is
T-j^:
CuJ.
We
are as a rule
known at the
Total Sur-
free
when for
12
i,
The
measure-
ments
in the case where both the bulk phases are mobile such as at
liquid-gas or liquid-liquid interfaces the value of the interfacial
surface tension or free surface energy is readily determined.
The methods of measurement of the surface energy of liquids
may be
The
static
(1)
The rise
(2)
The weight
(3)
liquid.
(4)
The measurement
or bubbles.
(5)
The
of a liquid.
for thin
insoluble films on
liquid surfaces.
The
(8)
Vibrating jets of
(9)
Vibrating drops.
(10)
ripple method.
fluid.
usually
that between each of
fluids,
and Thom-
.D
Let A represent the solid, B, G the two fluids, EG, GD, GF the
two solid-fluid and the fluid-fluid interfaces respectively, the line
GF forming an angle a with ED. This angle is called the angle of
contact of the system. Then since FQ- represents an equilibrium
of the system in this position
configuration the potential energy
must be a minimum,
an infinitesimal displacement of GF
an alteration in the energy of the system.
so that
If
(T
A c,
total
<TBC>
where BA
is
B instead
of with
but this
(7,
<?AC
total
<?AB
now
energy change
= &BC cos
in contact with
is zero,
hence
Hence the
interfaces
direct measurement.
The determination
difficulty
and very
different results
observers for the angle of contact between the same pair of fluids at
a given solid surface. The results have been found to differ
according
to whether the fluid has been in previous contact with the solid and
the length of time during which contact has occurred. Thus Quincke
found the angle of contact of mercury-air on a glass plate to be
initially 148 55' but this value fell after two days to 137 14'. Ablett
investigated the angle of contact between water and a horizontally
revolving cylinder of wax half immersed in the liquid. With the
drum at rest a was found to be 104 34' + 5' at 10*4 C. When the
mean
before.
it follows
that
o-wax . air
ANGLE OF CONTACT
to within one part in 500 it may be concluded that sin 6 < -g^ or
< 7' of arc. This is true of most pure wetting liquids for which
a- by two such methods are available.
By various
?A\
measurement methods Magie (Phil. Mag. xxvi. IQQ, 1888),
Langmuir (Trans. Farad. Soc. XV. 62, 1920), Anderson and Bowen
(Phil. Mag. xxxi. 143, 285, 1916) and Richards and Carver (J.A.C.S.
accurate values of
direct
surface
a liquid with zero angle of contact must therefore wet the tube to
some distance above the visible meniscus. It is an experimental
fact that unless care
values for
cr
should similarly fail to wet a containing tube until well below the
meniscus: the column of liquid will then be shielded by a gas film
from the glass tube until this film is squeezed out by the hydro-
methods employed
tensions of liquids
*
See also
the surface
is
given.
N. E. Adam and G.
8
4.
The
method.
capillary rise
Of the methods
of
When
"a>
oi. a
respectively.
we have
Now
>
(c72
cr 2
<?i-z
- cr
2-77T
1>2 )
cos a
<TI
= g 7rr*h (pi
Bh
where a
7,
is
^cTx
pg ) Sh,
'
cos a
~gr(pi-p g y
If the tube is not infinitesimal in radius, the calculation
more
we
becomes
difficult,
shall
imagined.
2o-cosa -
,,
= r (fi + r
._
0'
where a2 =
I
/"
\Pi
ui
fi
finds
O
pg)
r
practical purpose if
This formula
is
V2
accurate enough
O for
any
V
> 6.
between
r-^
^-y
9(pi~P0)1>
10
rise for
The
capillary rise h,
is
of
by surface
employ so
tension.
and one
and
2g- cos
and
whence
To
2<r cos
11
k
6i
with warm aqua regia or chromic acid, and finally steaming for
about half an hour. The apparatus should therefore be designed
to facilitate this thorough cleaning. These conditions of course
apply equally to any method of measuring surface tensions.
11
may be
section
The
first term
may now be negative.
In Ramsay's experiments the forms of apparatus used were
capable of sustaining pressures up to 100 atmospheres. The wide
and narrow tubes were concentric: the wide tube was therefore
annular in shape, and the allowance for the capillary rise in it
becomes difficult to calculate. Ramsay did not make a sufficient
allowance for the rise in the annular tube and in consequence all
his values, and those of later workers who have adopted his figures
Sugden
a particular liquid, and assumes that the rise in the two tubes
will be the same for all other temperatures and liquids. By this
HarMns
12
may reduce
ment
means
of a cathetometer or found
by a
slight modification of
Ferguson's method.
Again the case of two parallel plates dipping vertically or inclined at a slight angle at a short distance apart in the liquid
tension
is
approximately
g
where d
is
^ dhg (p p g )
e
difficulties
glass,
and
The method
of
the weight of a drop formed slowly and allowed to drop from a tip
ground plane which is with a sharp circular edge, has frequently been
employed owing to the simplicity of the method. The experimental
difficulties to
results
it is
many
13
Tate (Phil. Mag. xxvn. 176, 1864) discovered empirically that the
weight of a drop of liquid falling from the end of a tube was proportional to the radius of the tube and the surface tension of the
liquid.
W=irrcr
'
constant
"
above
4.
is
due to the fact that the actual dean extremely complicated dynamical
of course
is
and that
W=
f-a where
2-Trrcr J
to
a?
The
v
be approximately a cubic function of CL
/Y*
approaches
being unity when
0'6
falling to a
minimum
of about
14
is
developed, where
a-1} <r z
maximum
From
With the
15
The weight
of a
The
From a graph constructed from the table above we find that the
tube radius required to give a drop of water of the same characv
teristic ratio
-,
~
01239
also <r-
HO
= 72-80
'
0-998 whence
'
cr
Many
and p w A
xigfj
Hg
-*_i-
f-'TT
TT f\
utilised
in practice.
number
formed by a standard
fluid.
of drops formed
This method
is
is
Bubble pressure.
By reversing the position of liquid and gas assumed in the preceding section we obtain the bubble pressure method. The theory
corresponds closely with that of the drop weight and has been developed by Cantor, Feustel and Schrbdinger (Ann. d. Physik, XLVI.
413, 1915).
The equation
derived
is
2 rp
lr
it
16
where
is
the
is
of the tube.
carefully
worked out
we
shall see (pp. 57-58) from that against the pure vapour.
This method undoubtedly has the widest application of any yet
employed, and is at the same time accurate. Among other workers
of
a-
~ \g + 27TCT
2ir
~X^
where \
is
DYNAMIC METHODS
If the time of vibration of the tuning fork be
17
r,
then
T=
V
,
whence
far
bulk
liquid.
When
a liquid interface
is
unstable until the surface phase has acquired its correct excess
or deficit of solute by diffusion from or into the bulk of the solution.
is
is
as
method the
surface
is
may
vibration
another.
tion, a
and forms a
From a determination of the wave length of the vibrameasurement which can be made with some accuracy
by Bohr.
2
18
"
"
axes of the
jet,
'"max.
^min.
'
p the density of the liquid and <r its surface tension, p 1 the density
of the gas above the liquid, A, the wave-length, 77 the viscosity of
the liquid, Bohr showed that
-*
-\
/"
\r
'
The term
-(
{_
94 \n
^ U>/
"\
r l/ f
^-r
h
pva-^TT
/r^-ri
\ 2
/rnvv \ 4
"
I
however even in 23
. . .
'/'"
'
"1
pV&
/n
"
7T
cane sugar
solution
is
Static
The
for
(Zeit.
static
to values
methods appear
existed for a
much shorter
method due
to
19
should throw light on the important question of the rate of attainment of equilibrium, in the surface phase.
7.
The
The
is
may be
obtained.
surface layer.
surface
Since
-5
is
As examples
of the
free surface
dependent
22
20
the surface of water (Ch. ill) came to the conclusion that such
molecules were orientated in the surface with their hydrocarbon
chains in the gas phase and their polar carboxyl or hydroxyl groups
in the water phase. If we assume a similar orientation of the surface
film of a homogeneous liquid and that the surface energy is defined
by the nature of the surface film alone the experimental facts may
be interpreted at least in a qualitative way. For hydrocarbons will
expose the methyl groups at the ends of the chains to the air
whilst fatty acids and alcohols will likewise possess a surface layer of
CH
bulk of the
CO OH
and
We thus
arrive at a hypothesis as to the structure of the surface phase for pure liquids, that it consists of a layer
of orientated molecules with the active portions of the molecule
liquid.
of actual films.
21
example we must assume" on this hypothesis a layer of orientated molecules on the surface of the liquid.
The surface molecules however are continually evaporating and recondensing on the surface. The rate of evaporation and condensation may readily be determined with the aid of the Herz-Knudsen
In the case of water
for
KP
where p
moment
total surface
energy
is
is
determined
is
COOCH
replaced in the
,
0001, CONH,.
CH
energy of benzene and substituted derivatives came to the conclusion that this molecule lay as a flat ring on the surface, but by the
introduction of a polar group the ring was tilted on edge causing
a rise in the surface energy, according to Harkins by permitting
the residual valencies of the carbon atoms in the ring to come into
play.
The natural period of a molecule as determined from both the thermal conand from the natural infra-red vibration frequency is of the order of
10~12 seconds.
ductivity
22
X$Y-^X$Y
X(f>
the
having progressive residual affinities,
residual affinity of the group X lying between F3 and F4 the
molecules should orientate themselves in the following manner:
the groups
first
those of
progressively increase whilst
identical.
Further
if
we take a
series
F^Fj,
F <F F $F
2
none of these
On
series
groups are all in the vapour phase the hydrogen being in the liquid,
a conclusion quite opposite to that usually obtained. In the second
23
and third
maxima.
We must conclude that the surface tension
of a pure liquid is
determined by properties of the molecule as a whole together with
its foundation rather than by the external groups only.
8.
The
it is
should nob be the case, for a molecule proceeding from the interior
to the surface requires the supply of sufficient energy to remove
the non-polar end of the molecule from the liquid to the vapour
attached by the polar end to the bulk liquid
process of superficial vaporisation entails the
separation of the active group from the liquid.
Clearly equality between the two stages in the process of
phase,
phase.
it
being
still
The second
evaporation from the interior of the liquid should only obtain for
perfectly symmetrical molecules with increasing asymmetry, the
second step or the detachment of the polar end of the molecule
should become more important and the energy associated with
;
24
methyl
9.
alcohol.
The
influence of temperature
of liquids.
The
rises.
rr
_Q_
at
= o-
or
a- t
where
K=
(1
CTO
- at),
EOTVOS LAW
25
and
critical
point
T are
Van
8 constants.
their experi-
the equation
'V-FJ-
Hence
^ df^ 2
and -j~ vanishes with
<r
Experiment however
where n varies somewhat for different liquids but is in the neighbourhood of 1'21 as determined by Ferguson (Trans. Farad. Soc.
xix. 408, 1923).
10. Eotvos'
Law.
The
Eotvos that the molar free surface energy of any liquid should be
proportional to the distance from its critical temperature and to a-.
= RT, the
universal constant, a "law" analogous to the gas law
product
PV
PV being the free molar volume energy of the gas.
Corre-
spondingly in the case of the liquid- vapour surface the molar free
energy is proportional to the product of a- and the surface occupied
(M\*
j
where
is
Instead of
where
is
T we
write
T and
obtain
2'2.
26
and modified
it
to
a-
f
j
= K (T
8),
the
number
of
somewhat below
critical point.
To avoid the
we may
differentiate
^
or assuming that
is
(i.e.
for
non-
associated liquids)
K:=M dT
Katayama (Tokio
( fff>
')'
5.
the
full
EOTVOS LAW
K of Edtvos
value
2-2,
27
J0Y
aJ.
Yl
\pj
is
mean
coefficient of the
found to be in the
neighbourhood of 2'2 this" group included the saturated hydrocarbons, liquids which of all others would be expected to exist as
:
2-0.)
grounds,
coefficients of expansion and
specific inductive capacity, indicate
that these substances are
probably associated to some extent,
coefficients of molecular
first group by
replacing M,
the formula weight,
by M, where a is the degree of association and
aM therefore the mean molecular weight. In this procedure it is
assumed that the average
is the
composition of the surface
layer
same
as that of the
liquid in bulk, for otherwise the value of a
obtained can apply only to the surface
layer. In order to decide
this question,
Ramsay and Aston examined the surface tensions at
different temperatures of mixtures of
organic liquids of different
molecular weights and
for the mixture
compared the values of
assumption
justified.
Unfortunately, however, the examples chosen are not sufficient to
prove the case, for in no example do the surface tensions of the
liquids differ by more than six dynes per centimetre, so that little
deviation from a simple mixture rule could be
If instead
expected.
of two organic
liquids with not very different surface tensions the
had
been
carried out with water and an alcohol, a
experiment
different conclusion
simple liquid
28
and
its
the other.
the
Temp.
...
15 C.
55 C.
95 0.
o-
...
48-21
49-22
mean
assumed.
EOTVOS LAW
29
the molecules
lie flat
and
to the surface,
much below
fall
if
this value.
In the
first
case the
calculated
and
At
still
be
1'4, so
approached.
low
is
K will be found.
It
would be
of
much
whose
molecules
may
values of
unaffected
what we should expect of these bodies if the three hydrocarbon chains lay parallel.
The free acids have a much smaller K and since the ratio of length to diameter in
their case must be even greater than for the glycerides we must conclude that there
is far less orientation, but that the shape of the molecule allows more molecul es to
come into a given space of surface in a random disposition than would be the case
for a molecule of the
of regular shape.
B'lore
18G
30
The Parachor.
Sugden
where
c is a characteristic
The parachor
P is
*Jc
c(pi-
M or
P=
^- M
(Pi
is
/>)*
J.
'
'
the substances.
J.U
Pv)
0=
4-8,
treble
= 46-6,
= 23'2,
= 61,
ring
bond
H = 17-1,
double bond
O = 20'0,
six
= 54-3,
Br = 68-0,
I = 91-0,
Cl
ester
60'0.
CHAPTER
II
action occurs and the surface phase will be poorer in solute than
the bulk phase. This enrichment or impoverishment of the surface
The
Gibbs' equation
may be regarded
as the
fundamental basis
for
the
For any
in energy
infinitesimal
may be
is
written
(1),
32
where
W
W
does not
geneous or not, so long as
When this
intersect a phase boundary.
condition is not fulfilled, a term of the form
<rdA must be added to the right-hand side
of equation (1) for each interface intersected
and also the second term will have to be
M
resolved into parts if any of these interfaces
are appreciably curved.
Let us consider the case in which there are two phases separated
is within the surface
by an interface of which only an area
interface is flat.
We
obtain
The non-homogeneous
molecule thick and
in this respect
it is
drawn somewhere
in this
it.
parallel to
V and V"
By assuming that the entropies S', S" in ergs per c.c. per 0. and
the densities of each component p/, p 2 PI", pz", etc. in mols. per c.c.
which obtain in the bulk of the two phases remain constant right
up to the dividing surface, we can assign a portion of dU to each
'
of the phases
i.
In a similar way an increment of entropy and quantities of components must be present in the interface
dm,
where ^
is
it follows
+ f^d(AT^ + f^d(AP } +
2
...... (3).
33
that
).
J. ) r\
(A
/j, a
*)
the surface
They must
is
On differentiation we obtain
= ffdA -f Ada- + Td (ASj) + ASi dT + ^dAT, +
+ 2 dAT 2 + AT d(* +
2
fj,
Subtracting from
(3)
we
...... (4).
obtain
dtr^-SidT-^dfr-Ttdfr-....
For a binary system at constant temperature
da-
I\ d/j,!
this reduces to
T z dp%.
non-homogeneous
zero
we
region,
obtain
dor
r2
cZjU, 2
=
fj,
whence
For
-da-
lJ.
+ RT loge C,
= RT.T.d log
or
r=~
dcr
j^^
'
oge c
the concentration
all non-ideal and concentrated solutions
may
3
34
"
"
activity
fo
= p + RT logefo,
T = - ^m
RTd
whence
or
7-
logefc
RTd log P
e
where
of the solute, or
1
do-
if
surface
may be
It
may be
homogeneous phase.
The
field
is
very
wide, for
it
composition.
2.
The experimental
Since
EXPERIMENTAL VERIFICATION OF
GIBBS'
35
THEOREM
r=
n x
2
grm. mols per cm.
determined
results
were obtained
A similar
values of
unsatisfactory
is
surprisingly good
on the whole.
32
36
earlier experiments
by W.
C.
the ad-
8 and caffeine
sorptions of potassium chloride, BaCl 2 CuCl a
were determined at the interface of water-paraffin oil and for sodium
,
AgN0
was
to
Donnan and Barker a similar plan has also been used by Patrick
who measured the adsorption of HgNOg, salicylic acid, picric acid,
Of these cases only
neofuchsin, etc., at a mercury interface.
:
aniline
and
caffeine
every other substance showed an adsorption considerably too great, up to one hundred times the calculated value.
It is worth devoting some space to consider the causes of these
Gibbs' theorem
In the
whole divergence.
much more
That
is
than
it
to say that
-^
appeared to
and therefore
also
37
be.
titrated electrometrically.
evidence for Gibbs'
Although absolutely conclusive quantitative
truth
is
from
theorem
so
many
as a valid principle.
3.
We would
material
capillary active
of the solution will fall rapidly.
the solute and the surface tension
38
and subsequently approaches almost asymptotiWith the aid of the Gibbs equation we may
curve, where T is the surface
readily determine the form of the T,
excess. The surface concentration may be assumed equal to F if we
falls
rapidly at
first
and mixtures
-1
Pure water
of
3
-4
Mol Fraction
-5
-6
of Ethyl Alcohol
Pure alcohol
Again
value of
For
no
39
Thermodynamics).
-1
Pure water
'3
-4
>5
Mol Fraction
'6
.9
TO
Pure pyridine
of Pyridine
of these curves (Proa. Roy. Soc. A, oix. 57,, 1925). The two examples
depicted in the curves are (1) ethyl alcohol water mixtures from
/c
which
40
1-0
1-5
2-0
Log1Q Phenol
T,
0-5
0-0
Activity
ideal solution
da-
r==
It will be noted that there exist
two
curves.
(2)
The
r,
Ghimie
et
de Physique,
vm.
ix.
For comparison
is
inserted the F,
41
N curve,
Lack of
The
N curves
cr,
p. 36.
activity of which can be varied by the addition of various concentrations of sodium chloride present interesting features (J.G.S.
that whilst the slopes of the curves for different salt concentrations
are identical indicating that for identical bulk activities of phenol
the surface concentrations are likewise identical, yet the surface
tensions are
of the phenol alone even when this layer is tightly packed, but are
affected by the nature of the
underlying liquid. (From the value
of Fa, A. is found to be 23'S A. for phenol on water, in close agree-
follow the
as
same course
an asymptotic
Fmax-
fall
is
in the
value.
obtained at about
(Phil Mag.
and a and
/3
42
From
(1)
we obtain by
differentiation
dlogN
and
dcr
r==
leads to the expression
r=
2-303 JRT'
as pointed out
by Langmuir
maximum
v.
form:
o'o-o-
where
and
this equation
_.
n_
are constants.
If
fK + i
N be
large
compared with
An
limit
ZSQSRTc
tensions; thus
is positively
is
ca.
phase
is
43
the bulk phase and the increase in surface tension thus produced
is nearly proportional to the salt concentration.
In rare cases
maximum
mixture
<r, JV"
We
curve and the "activity" of the solute in the solution; it does not
however give us any information as to the thickness of the phase,
beyond an indication that the liquid above the geometrical dividing
surface
is
uniform in composition and does not acquire the characa phase in bulk. In the next chapter our attention will
teristics of
films are
in
all
is
how far we
are justified in
it is
evidently important
cases both for pure liquids and for solutions the Gibbs layer
considered as similar in character, viz. unimolecular in
may be
orientated
as orientation requires.
Even more
44
Eb'tvb's
constant
for
unsymmetrical
also
with the depth. The top layer alone may be said to be formed by
the operation of chemical forces.
Thus in the case of concentrated solutions of ethyl alcohol in
water the subsequent decrease in T with increasing values of
it is
45
4.
were
definitely orientated.
suggested that in all cases for both soluble
He
and insoluble
substances and
for
we
=a
sq.
obtain
A = p as
the area
by one
to
46
we
of
r max
for alcohol
water mixtures
one respect, however, such soluble films differ from the insoluble
films which we shall have occasion to examine in the next chapter;
the surface tension of solutions which according to the Gibbs adsorption equation
~
have a value of
T = 2 x 10~
RTdN ...........................
10
would not
affect
much
as one-
tenth of a dyne.
Soluble substances appear to be capable of occupying enormous
areas of the water surface and to be still capable of affecting the
surface tension.
F=ais
-a-,
values of
F or
combined with
(1) gives
or
FA = RT.
The resemblance between this equation and the equation P V= RT
for perfect gases led
Traube
by a solute
is
due
47
to the
thermal
at the surface.
The
agitation of the solute molecules adsorbed
observed surface tension is on this view the resultant of two forces,
cr
-far
investigated.
We
F = JcN
equation
If however
we
to
behave
to be infinitely small
For at
them.
all
FA
obtained
FA
-^ over
by
PV for a real
when -~~
F may
a range of values of
conveniently be
manner.
Since
whence
The
dF=Tdfj
RT RTdF
2-303.R2 d log
'
48
dp
= 2-303RTd Iog
10 c,
FA =
whence
d log
Rf dl^J"
The data
of
v.
FA.*
the values of
^-^
for
determining
one in which the Gibbs film is established; only as a first approximay the ideal dynamic surface tension of a solution, i.e.
of a solution without a Gibbs film, be regarded as equal to the
mation
* For weak
electrolytes the degree of dissociation found by the freezing point
method agrees with that found from conductivity within the limits of experimental
error and Ostwald's dilution law is obeyed.
When
water
given by
is
jnm
number
(Z/ia
N-,,
=
= -ffdn
i
2
Integrating this
The
of
-^
__flt[c(l + a)]
^
55-51,
djj^=RT
--
is
^ m rdc adc
= ET\
+
+
I
&
~]
da
\.
.J
we obtain
definite integral
be evaluated
may
if
A bg/
/ ~fc
thus reduce to
f
for
the higher fatty acids determined
.
by
49
is
faithfully reflected.
22
F in
24
26
28
SO
PVPb
to represent the behaviour of gases,
an equation in which 6 is a
constant for any one gas and represents the least volume the
gas
can occupy at these temperatures. On the other hand % varies'
both with the nature of the gas and with the temperature,
always
approaching unity for very high temperatures and becoming
50
A corresponding equation
FA_FB
RT
in which
is
0'3 A.
This value
water-benzene interface.
is
The
also given
values of
?i-butyric acid at a
for the iso-acids are
by
consistently a little larger than those for the normal acids. 'The
values of sc show that these acids resemble gases whose critical
The
iso-
51
when
It
branched.
is
as for
butyric acid at
interesting to note the large value of
Immersion of the hydrocarbon chains
a water-benzene interface.
eliminating
log
F+
an equation which is found to, give results which agree remarkably well with the precise surface tension measurements of
von Szyszkowski.
Adsorption of
electrolytes.
uonelectrolytes.
considered,
ions adsorbed.
the case of potassium iodide, for example, for very small values of
the Traube relationship
FA
minimum,
as is the case
where molecules
42
52
is
rapidly, indicating that a repulsive action of great magnitude
becoming operative as the superficial concentration increases. The
nature and structure of this double layer so produced will be
Surface Pressures.
surface films
C.
^S 7r
'
= 2*3,
whilst -~
has this
when the
value
"
"
surface pressure at a mercury interface, the surface tension of
which has been lowered 30 dynes by cane sugar, is equivalent to
an osmotic pressure of 440 atmospheres.
The
ratio
P/F
absolute units.
is
Under high
30'5 c.c. per mol. w-butyric acid at a water-air interbe compared to nitrogen at
C. between the pressures
400 1500 atmospheres, since x 0'73 for both substances. B for
= 1*49 x 10 s sq. cm. per mol., hence
butyric acid
nitrogen
face
is
may
P
P
1-49x10
30=5
Aon
1A*,
'*
82xl(
53
constitution.
it is
in equilibrium,
may
**
-*
where Q
it
phase
'
From a knowledge
of the surface
It is interesting to
Langmuir
series of solvents of
for
54
We
OH -group
2
in a hydrocarbon chain is
of
its position in
5.
a chain.
Negative Adsorption.
and in
inorganic salts raise the surface tension of water
accordance with the thermodynamic considerations implied in the
Gibbs equation, the surface concentration of solutions must be less
The
The
of such solutions
is
scanty.
of Stocker (vibrating jet) (Zeit. Phys. Chem. xciv, 149, 1920), who
compared his values with tkose of Grabowski and Pann (Diss.
rise)
where
The
"water
be
The comparison
of the values of
marked
variations exist.
NEGATIVE ADSORPTION
55
It
is
3
interesting to observe, that with both CdCl2 and
is initially concave to the concentration axis and that
AgN0
the curve
we
can, as
Langmuir has
indicated, obtain
some idea
as to the
be r
and the bulk concentration G, the quantity of solute which has left
the film, on the hypothesis of a unimolecular film of
pure solvent
size of
p ._
_,
is
^"
'
dp
1000
The
of r
is
thus
= 4 JL
j~for
the thickness of a
From
56
^^
by dissolved salts
= 4*4 A. and A
obtains T
(J.
22
The
salt as
F would appear
indicated
by the
The
the
to vary
following
sq.
cm. and
salt,
57
must be
by solution of
salts in
the liquid, a
The adsorption
of gases
on
liquid, surfaces.
Very
little
accurate information
The
58
r==
where dp
= 2'303RTd Iog
10
~^?
P.
we
F = -j
which are in
fair
the mercury surface in the form of a thick layer, the free surface
of which would possess the properties of the free surface of the
liquid in bulk,
is
within relatively wide limits. Iredale has shown that these values
do not correspond to the tensions mercury/liquid, liquid/air,
but are due to a condensed film, a secondary film on the mercury
Whilst the transition from the primary film to the thick
surface.
'
is little
doubt
59
from the data that the change is definite, or that the secondary
film must be built up from a great number of layers before we
arrive at a film with the properties akin to those on the free
This continuous change in the P, cr curves
surface of the liquid.
obtained by Iredale near the
critical
point
sat
appears to be
mixtures near the salting out point, and may be taken as evidence
for a progressive thickening of the capillary layer before it acquires
the properties of a bulk phase.
That adsorption occurs with great ease on mercury surfaces is
vibrating jet
for
Harkins
Similar
(
Wied.
Ann. LXYL 499, 1898) and Harkins and Ewing (J.A.C.S. xui.
2539, 1920) and Hogness (J.A.O.S. XLIII. 1621, 1921). The former
obtained values for
<r
admitted.
Iredale obtained the following valuesin dry air at various pressures
60
According to
contact with
Iredale
its
own vapour
Some
on liquid surfaces
cr^
VA > VB
where
The
o-j/ is
+ VAB>
O-B
= ^2-05,
= 28-17,
cr
AB =
ax'
34'68,
= 27-90,
indicating that benzene will readily spread on water. It must however be noted that benzene may likewise spread on surfaces by a
process of evaporation and condensation (see p. 66).
It is noted that if a layer of benzene be poured on to a limited
collects into
a number of
flat
lenses each
I.
5wsrevws?yxxxvS(X
61
The
is
independent of
its
area beyond
certain small limits of lens area, but the limiting continuous film
thickness h' is determined by the nature of the vessel (being
we
shall
character and
is
information
long before the vapour pressure above the solution is equal to that
of the pure alcohol (since rmax> is obtained in this case at 0'3 molar)
or acid.
Thus for these substances to obtain equality of vapour
pressure between a surface and bulk phase, layers more than one
molecule thick must be necessary.
(1)
secondary
See p. 97.
62
spreading liquid.
If the film be primary or unimolecular in character the vapour
pressure is directly related to the surface tension of the solution,
for the
amount
adsorbed or 6
of surface covered 6
KT.
is
proportional to the
amount
fv*
-j
~ vKT,
Out
where
At
is
is
bare.
From
the covered surface 0'443 x 10~ 9 gm. mols are present. Thus the
7
life of a benzene molecule on the
mercury surface is 4*9 x 10~ seconds.
Over a
free
benzene surface,
if
4-7
x 10~ 8 seconds.
CHAPTER
III
Introduction.
substances on liquids.
If a drop of oil be placed upon the surface of a liquid in which
insoluble there will exist at the moment of contact an unstable
it is
system in which three surface forces are operative at three interfaces as indicated in the diagram.
777777777-x
0/2
We
will define
by
cr^ <rz
drop.
On
ar^
+ er =
12
<7 aj
the liquid-
o- 12
it is
evident that
no decrease in the
by spreading
free
energy
or contraction of the
if
"i
"i2
>
"a
a decrease in free energy of the system can take place on contraction of the drop exposing a greater liquid-air interface for the
second liquid. The edges of the drop then contract and the flattened
Tt
an expression
first
obtained by
Neumann.
04
A third possibility
Under
oil will
liquid.
The
when the
oil
spreads
cr12
and may be
evidently o-2
{cr^ +
taken as a measure of the tendency of the oil to spread, a point first
noted by
W.
B.
sq.
cm.
is
Hardy and
to
name
for
a number of liquids
which spread upon water and for liquids which do not spread
o- J2
but form lenses, in this latter case the value of o-2
is
{a-^ +
}
evidently negative.
section.
oil continues to
spread until the rate of superbalanced by the rate of return of molecules from
the surface film into the lens. The oil of the lens
may be regarded
ficial
solution
is
65
"ia
O"oily water-
The
whence we obtain
for
"oily
water;
or
"2
we have noted
0^2
~ o"a = ~
&>
F the
drop is
This view of the process of spreading and the conditions of equilibrium possesses certain advantages because many solids such as
the fatty esters and higher fatty acids spread upon water and their
spreading coefficients can only be determined by a direct measure
of the film pressure which exists when equilibrium is attained.
Further investigation has shown that the spreading coefficient
varies with the temperature, a point which we shall refer to later (p. 8*7).
coefficients for
C.
66
solution in what
it
may
2.
films.
If a spreading oil be dissolved in a volatile solvent such as benzene and a definite quantity be placed upon clean water in a dish
layer of molecules.
The
for the
67
this film
thick.
tion that the density of the oil in such films is the same as that
of oil in bulk. If this were not the case, but if an oil film, for
example, had only half the density of ordinary oil, the value calculated for the thickness wo aid be doubled, and we should either
have to abandon the unimolecular theory or make the further
approximately "symmetrical.
Devaux also advanced the important theory that the characteristics of the solid, liquid and gaseous states of matter are retained
so long as one continuous layer of molecules remains unbroken.
to be described.
is readily
distinguished from another. In certain
properties, a solid film of unimolecular thickness resembles quantitatively a three-dimensional solid mass of the same substance,
The compressibility of
appears to be the same as for the material in bulk, but the meltingpoint, which is not a vectorial property, is very different. If the
refractive index of a film could be measured, we should expect to
find that in the plane of the surface it was the same as for an
direction.
solid film in two dimensions,
although so closely
analogous to its three-dimensional equivalent, must therefore be
regarded as a distinct phase.
52
68
about 15. At the lower steady value the film behaved like a rigid
solid, so that particles of talc floating on it could not be blown
about; at the upper value, on the other hand, floating particles
were completely mobile. Labrouste, in fact, observed a somewhat
indefinite two-dimensional fusion or volatilisation.
maximum
69
as wide as the trough and just touching its surface. This paper
acted as a barrier for the oil film. At the narrow spaces between
the barrier and the trough a jet of air played with sufficient force
to prevent leakage of oil but not to disturb the barrier sensibly.
The trough was provided also with other barriers of paper which
could be moved along the trough, pushing before them the film of
oil or any impurity. To carry out an experiment the space on one
was swept clear of impurity, leaving pure
was formed a layer of oil of known
volume and of an area controlled by the paper barriers. While
the area of the oiled surface was altered the position of the
hanging barrier was kept constant by adjusting the weights in the
opposite balance-pan. What was measured, therefore, was the
difference between the surface tension of pure and of contaminated
water. This measurement has the advantage of being direct, inside of the floating barrier
temperature.
portance.
x 10 23
number
calculated the
for the
The
on
at 16 C.
figure
compressive force
(i.e.
is
21'8 A.
particles of dust to
is
move about
freely.
As compression
is
applied
70
by the moving barrier the area falls, following the curved line QS.
The lowest part of this line appeared to be asymptotic to the axis
of area. At 8 dust particles on the surface lost their mobility as
if
now
is
70
60
50
40
30
20
10
20
10
21
From 8
such ceases to
exist.
higher compressions, like ordinary solids which melt with an increase of volume.
near
71
We have
cular.
own experiments
of his
oil
minimum
surface tension
Devaux
however pointed out that the ratio was not so great as two to one
in any case, and explained the phenomenon by supposing the
molecules to be capable of looser or tighter packing.
The theories offered so far gave no clue to the origin of the
unimolecular films. Langmuir however pointed out that if the
molecules in the film are regularly orientated on the surface of
the water owing to the attraction between the water and the
active groups of the organic liquid, these groups e.g.
COOH
CH
OH,
are dissolved as
it
"
CH
groups are distributed through the mass, it is natural that the oil
should prefer to remain in globule, rather than spread out as
a layer two molecules deep. In the mass of liquid the carboxyl
groups probably tend to gather into clusters, because of the
groups for one another than for the
hydrocarbon chains."
3.
Molecular Size.
72
The lengths
by dividing the
cross-section
volume
The
many
We
The
number
much, but
by
side,
occupying as
much
if
they were in
MOLECULAR
that the molecule
work
is
73
SIZE
itself.
In view of later
We
should
and
Miiller
(J.O.S.
cxxm.
2043,
De
Becker and Jahnke, Z. Phys. Chem. xcix. 242, 1923), who have
shown that two types of chain formation occur in carbon com-
pounds in accordance with our preconceived views on the tetrahedral orientation of the valencies of the carbon atom.
of the planes in these
diagram
two types
is
The spacing
74
is
all
Oriented
capable of wide application.
probability govern the behaviour under
thin films upon water. The force which attaches the carboxyl or
other "active" group of the long chain compound to the water is
of the same kind as that holding to a metallic nucleus of copper,
cobalt or platinum, the
ammonia
or
amine in a
so-called coordinated
salts,
such as the
We
or
may
well
by the
of
to be built up.
The orientation theory has made possible a great advance in our
knowledge of molecular architecture, and Langmuir's work on this
by Adam.
The
MOLECTJLAE SIZE
75
axis, in
or less.
Thus for cetyl alcohol
= 2O7 Aoctadecyl urea 26'3. The quantity
found for the intersection of the steeper line with the axis of area
is characteristic of all the hydrocarbon derivatives considered.
active group
it
is
21
7,
for
was the
cross-section
effective
of the
CH
chain
the
is
its length.
NH
NH
It will
water and
salt solutions
76
in which the
common
though
the
somewhat
21 '8 A.,
greater value for the area,
at which it cuts the axis of area,
be due to some other form of
may
compound with
FD represent
by the area
The area
when unconstrained
of
the head
is
2O
21
Sq..
22 23 24 25
26
A. per molecule.
The
so far investigated
head groups
MOLECULAR
77
SIZE
AB
atoms or
O4
per cent,
have been found. If we take the latter value, and calculate the
surface compressibility of palmitic acid on the assumption that
the length of the molecule is 21'IA., we find that a force of
78
The
substituted acetamides
ie
H NHCOCH
33
and C17
H NH GO GH
85
Remarkable results were obtained with pentaerythritol tetraThis symmetrical body gave a film of exceptional
rigidity, so that it withstood a force at one end of 5*6 dynes per cm.
without any support at the other end. When first put on the
area was 100 A. per molecule and was reducible by compression
to 80 A., or about four times the area of a single closely packed
The four chains must therefore lie parallel, and two of
chain.
them must be bent back through a large angle.
palmitate.
by
local thickening
under the
where the
strain: it
an
definitely
crystalline
form.
EXPANDED FILMS
with an
air-jet,
is
79
discontinued the crystal
phase
On the other hand the polar groups which have been removed
forcibly from the water surface may be heavily solvated and the
structure of the ejected material may approximate to that of a
partly crystallised soap film possessing but little tendency towards
adhesion with a water surface.
5.
Expanded Films.
commencing
to increase
it
becomes
Adam
GJ -
on N/fOO HCl.
O onN/10 HCl
D = mean of 6
20
30
shown
obna. in
40
50
expansion
80
being in this case 28*5 C., whilst the area of the expanded film at
this
value is ca. 41 A. Langmuir in his investigations on the
down
for
a number of acids at
by
Adam
with
It will be noted
an acid
CH
numbers
EXPANDED FILMS
81
is
15
10
l
A U
30
40
35
45
Area per molecule.
50
55
60
The temperature
film
is
at
impossible.
The apparent
similarity of the F,
chain
B.
increases.
s. c.
82
The
is
In an attempt
F(A
.5)
co
for
expanded
very large, so large in fact for long chain compounds that practically
every collision between the orientated molecules on the surface is
an
inelastic collision.
On
this view
expanded
films
and being
when a
another.
on liquid
of non-volatile substances
surfaces.
Adam
At a
(Proc. Boy. Soc. ox. 423, 1926) has recently obtained the following values
for
83
rises until
slow enough to push out the film already formed without any
is reached.
until the equilibrium value
An examination of the rate of such compression for
where
is
is
expanded
dn
* It
is somewhat remarkable that surface solution only occurs at the edge of the
crystal exposed to water-air interface and solution does not occur from the parts of
the crystal immersed in the water.
6-2
o4
is
dt
= kF
T
k
or
dn dri^ _
~dt~ ~dt~
_
~
by the equation
dff
~~dt'
whence
dt
T=25C.
85
it
acid at 25
differential tension
crystal surfac'e.
If the expansion rate
trough where the expanding area touches the side of the trough
the film would, if rigid, undergo compression in this direction and
continue to expand at the normal rate over the other portion of
full surface. Actually the film is not rigid and an augmentation
"
"
in the velocity of spread in the wave front is noted.
Thus, if a
in a circular trough in .a
water
on
acid
be
of
oleic
placed
drop
the
86
as the
7.
may
Again,
if
two
solubility
The heats
8.
<r
= F might be regarded as
liquid or
number
of solid
if
by
As
typical
87
fol-
is
We
solid
At 17
C.,
however,
88
commences
(Fig. p. 87)
system
to pack
may be
tension.
The point
considered as the equilibrium point of the
From
we may
acid, or as a
latter
case there
is
an additional
restraint on the
system
in
At
invariant point.
At
the crystal melts and there are present three phases
solid crystal, liquid lens and film.
The system is accordingly
:
nonvariant.
The F, T curve beyond this point is again linear, but the surface
pressure decreases rapidly with elevation of the temperature until
the point B is reached where an abrupt change in the
slope of the
curve is noted. In many cases such as in the long chain esters
the film
found to be condensed at
and expanded at B, thus
expansion interval of Labrouste and Adam. For acids
on the other hand with the exception of stearic, the film is already
expanded at the melting point and the portion of the curve
is
AB is the
AB
is
missing.
Since the equilibrium pressure
tensions
~oil
F is
+ ^interface
89
with a
of existing in two
water
interface
coefficient, at
is
by
AB
or completely
point.
From
with a knowledge
points
it is
change from
Thus
OA
i~
AB
+ 0'553
C.
^=-0-67
dF =
-0-153
dT
BO
= -S dT + AdF for
1
the film,
90
Since
P is constant vdP =
=S
dT
~A~ = ~TA
>
heat involved in the spreading of the film from the liquid lens
From data on
91
surface
followed
Hardy, Report
Soap
We
and
to
films.
to one another
It is thus
chains and the polar groups held together with water as a sandwich
between the orientated hydrocarbon chains. These elementary
leaflets
may
may
slip over
examined by Shearer
one another with great
(see
ease,
p.
73).
The
thus providing
films.
The
Dewar who showed (Proc. R, I. XXII. 193, 1917) that the rate of
penetration varied with both the film thickness and the excess
few of the values obtained for hydrogen
pressure in the bubble,
when
is
92
sq.
found to be invert
is
= 0-705,
^.t
at
where x
sq.
cm. and
t is
the tim<
days.
BLACK BUBBLE.
The
film.
through unimolecular
fi
LUBRICATION OF SURFACES
93
10.
Lubrication of surfaces.
The
CO OH
or
as -the
OH, yet a secondary film consisting of a
This
is generally built up.
relatively thick multimolecular layer
of time to adjust
secondary film takes a relatively long period
that molecular orientation is proitself and it
appears probable
a rate much slower than for the
ceeding in the layers of the oil at
a load on a lubricated
underlying unimolecular layer. On placing
surface the load floats on the multimolecular layer which can be
until presumably a bimolecular
squeezed out by increasing the load
with the molecules orientated towards the metal and load
layer
94
interface
is left.
The
multimole-
where N"
is
is
the
^=
number
&
-d-o(JV-2),
/A
in themselves exert
no
effect
on the
lubri-
CHAPTEE IV
LIQUID-LIQUID INTERFACES
1
Introduction.
interfacial energy
The study
of the interfacial liquid-liquid phase however is cornseveral factors, of which the chief is the mutual soluNo two liquids are completely immiscible
bility of the liquids.
even in such extreme cases as water and mercury or water and
plicated
by
On
interface and consequently no interfacial energy.
addition of a solute to one of the liquids a partition of the solute
between all three phases, the two liquids and the interfacial phase,
boundary
an extended
interfacial
96
LIQUID-LIQUID INTERFACES
2.
Antonow's Rule.
but the surface tension of the solvent may, as we have already seen,
nevertheless be reduced very much. The following examples* may
be quoted in support of Antonow's rule.
The agreement
by Eeynolds
is very good.
Equally good
in the case of benzene and
but
initially
was
sometimes considerably:
mixture the surface tensions of the two phases were respectively
*
ANTONOW
97
RULE
and 31 '5, and the interfacial tension fell from 41 '0 after five
minutes contact to 15'1 after standing together for a day. In
such cases a chemical change may be assumed, although it is not
73-0
litre.)
for
by consideration
hi
P,
(2)
Hence
cra
o~2
-}
^gpzhz
gpzhz
o-j
gpih-i'
= cr +
12
pi)>
LIQUID-LIQUID INTERFACES
98
Antonow's rule is thus only exact when the two mutually saturated
liquids possess the same density. The observed values of o- lz should
crl
in general be slightly less than those determined from <ra
In
the case of oleic acid floating on water Devaux obtained a lens
.
We
if
film, for then we shall have two surfaces whose combined free
energy equals that of the original surface. If on the other hand
the liquid of higher tension were to distil over and form a thick
film on the other liquid the sum of the surface
energies would
have increased by <TA + <T AB -a-s = %<TAB for each unit of
surface;
the phase of lower surface tension must therefore remain uncon-
3,
and
solubility.
closely
related series of
compounds, the interfacial tension increases as the
Thus at 20 the inter-
Propyl
Isobutyl
alcohol
M
Isoamyl
4-4
Octyl
8-5
and the
Isobutyl alcohol
342-7
359*0
'
INFLUENCE OF TEMPEEATUEE
The
99
is
cr
&A',
Then by Antonow's
Now
if
much from
cr
rule,
<r^
may however be
Influence of temperature.
The
temperature.
facial tensions
72
LIQUID-LIQUID INTERFACES
100
temperature =1 68 C.)
(Mean temperature
(Critical solution
and water.
C.)
phase at a
certain,
critical point
&-0.
Since however
F=
da-
.,
>
or'
TTJS
\dN) A
\dfj,/A
. ,.
it follows
.
,
that for
INFLUENCE OF TEMPERATURE
to
assume a
finite
value
-v^.
must
also
be
zero.
101
Thus
as pointed
tension
the case for surface tensions of liquids near their critical point,
but
would be of considerable
interest to have data for the
surface and interfacial tenIt
approximately by the formula a = 211 + 014/7 (600 - ), though conmake the figures somewhat uncertain.
JZeit.
LIQUID-LIQUID INTERFACES
102
In bringing together the surfaces of two liquids to form a liquidliquid interface heat is either evolved or absorbed during the
process. This heat change may be termed with Harkins the heat
of adhesion. The sum total of the surface energies of the two
surfaces before contact less that of the interface after contact
dT +
_/pQ^i
o2
if
the
_rpd
G
for
following interfaces:
ca.
25 A. Thus
a gm. molecule of octyl alcohol would occupy some 1"51 x 10 sq. cms.
and the heat evolved would thus be of the order of 6000 calories
T
where \
mdo-yt}
dT
isobutyl
alcohol
and chemical
surface.
constitution.
(J.A.G.S.
We
solutions. If
two
INFLUENCE OF TEMPEEATURE
103
is
thus numerically
a is similarly denned as the work required to pull apart a composite bar consisting of half of one liquid and half of the other, at
the place of junction.
a is
two
when
the
liquids are
first liquid
and
Thus
for
water
for octane
for octane-water
for octyl alcohol
for octyl alcohol-water
'8
4t3'8
W
W
=
a=
c
sq.
cm.
55*1
91*8
LIQUID-LIQUID INTERFACES
104
and
The
W -W
a
as is exemplified
may be markedly
affected
by a
of the
H of the aqueous phase. As the solution becomes more
alkaline the carboxyl groups of the acid commence to dissociate
and the interfacial surface tension falls rapidly. (Hartridge and
Peters, Proc. Roy. Soc. A,
5.
01.
105
If the
component
The
work in either
former
may
direction.
If the original liquids are again partially miscible, and the added
in either the mutual solubility may be increased
component soluble
if so
be the
altogether.
It is logical to conclude with the case
tension
is
increased
by
its
is
possible.
If the interfacial
We
by the addition
of
most
another
adsorbed at the surface between the solid and the other liquid, that
LIQUID-LIQUID INTERFACES
106
is
solid.
The
solid
A. But
solid-liquid interfaces
for considering
the
water-benzene: the
maximum
morphine hydrochloride,
See p. 36.
107
for if
results.
On
is
may
as great
the solute.
An
and adding potassium carbonate until two phases were reached, one
and one poor in this salt.
rich
interfacial tension of
amalgams
salts,
of the
obtaining
LIQUID-LIQUID INTERFACES
108
KOH
is
reduced to 33
of its value
by the addition
of 0-0224
of
is still
more
amalgam-normal LiOl
it is
not
that the rapid fall in the value of the surface tension does not
begin at the very lowest concentrations of solute. The first addition
of alkali metal indeed produces little effect on the interfacial
tension, and there is a point of inflexion on the concentration
surface tension at its steepest part. This behaviour appears to be
characteristic of amalgams: the explanation is not clear and the
investigation,
Emulsions.
The emulsions
We
have noted
respect to the continuous and discontinuous phases.
that the interfacial energies of immiscible liquids such as water
and
computed values
for
relatively rapidly.
Dilute emulsions in
LIMITS OF EMULSIONS
are the effective precipitating agents.
109
The volumnar
ratio of the
to the
tension.
7.
Limits of emulsions.
The
increased in number or size until point contact between neighbouring spheres was obtained, yield a volumnar ratio of disperse phase
to dispersion medium of 74 26 which would fix
automatically
the upper limits possible in emulsification. The experiments of
:
S.
that
agent must possess great mechanical strength but must at the same
time not be too brittle otherwise fracture may result and the
emulsion be broken.
rigid,
and
110
The emulsifying
agent.
surface tension of such liquid and thus iri accordance with the
Gibbs' equation go to the interface and there produce a film which
form a
solid
may
membrane
(Nugent,
66, 1922)
oil.
In the same
way the
emulsion.
The
stability of
an emulsion
is
EMULSIFYING AGENT
111
a benzene-water interface by determining the amount of soap extracted by emulsification of the benzene and thus increasing the
liquid-liquid interface, when the following figures were obtained.
(ibid.
to separate
and
it
was found on
analysis of the soap content of the aqueous lower layer and the
concentrated emulsion in the upper layer that the emulsion had
abstracted soap from the solution.
Owing to the fact that the
fatty acids are soluble in the oil the hydrolysis of the soaps and
the subsequent removal of the fatty acid in the oil
phase had to
Some
112
LIQUID-LIQUID INTERFACES
As a result of a number of experiments the mean areas of paraffinwater interface occupied by a molecule of the emulsifying soaps
were obtained as follows. For comparison Langmuir's values for
the actual areas of the fatty acids are appended whilst Adam's
head is 25'1 A.
It
COOH
unstable.
if less
The experimental
EMULSIFYING AGENT
113
Water
Paraffin*- Water
Paraffin
WaCer
Paraffin
the
that of the non-polar portion to which it is attached. In
converse is the case. It is clear that the ingress of more soap
molecules into the interface would necessitate an increase in the
the second.
of
LIQUID-LIQUID INTEKPACES
the emulsifying agent. Furthermore if we represent diagrammatically the cross-section of two soap molecules in juxtaposition to one
oil
it is
"""
is
dependent on three
end, the
----
Emulsion
EMULSIFYING AGENT
The
when
115
dissolved in water
COOCs,
COO/
-coo.
>Mg,
COO X
COO'
-COO
\:Zn,
COO'
Both the
on the head
stability
stability
sizes as is indicated
for
the
WATER
No
IN OIL.
the emulsion
with short hydrocarbon tails are very water soluble and the sodium
soaps soluble to a greater degree, interfacial adsorption is consequently small. Thus the concentration required to produce a
saturated film at the oil-water interface will be
correspondingly
This necessitates a high sodium ion concentration in the
greater.
82
116
LIQUID-LIQUID INTERFACES
aqueous phase which although weakly adsorbed by the electronegative oil drops may rise to those concentrations where discharge
of the emulsion droplets occurs before the protective film has been
formed.
A microscopic
stabilised with
case.
/rn\
this
Benzene
-Lampblack
is
more soluble
figures.
The
was-
EMULSIFYING AGENT
117
Although
all
are, as is indicated
by the lowering
The Phase
The
Inversion.
~~
fact 'that
")>
Ca
sodium
is less,
mixed soap
may be so adjusted as
LIQUID-LIQUID INTERFACES
118
first realised
sq. em., in
section of n
+ 2m
readily determined
hydrocarbon chains.
by the abrupt
oil
phase
is
been investigated by Bhatnagar* (J.C.S. cxvu. 542, 1920), but unfortunately no data on the interfacial adsorption of the mixed salts
are available as yet.
Some
Emulsion
10
c.c.
Emulsifying agent
water, 10
:
c.c. oil.
potassium
oleate.
Al" > Or"' > Ni" > Pb" > Ba" > Sn", Ca".
The explanation of stabilisation and phase inversion advanced
above has as been noted received support from the experimental
data on the effect of soaps and the salts of the fatty acids.
PHASE INVERSION
119
emulsion.
emulsion was at the inversion point since two emulsions, the one
the invert of the other, were always obtained. The work of Baur
(see p. 119) has demonstrated quite clearly that the inorganic
ions are oil soluble and that the difference between the
partition
ratios of cation and anion between oils and water
may give rise to
interfacial potential differences as well as interfacial concentration
differences of cation and anion on each side of the interface. The
difficulty of obtaining stable emulsions with a fat free paraffin and
inorganic soluble salts indicates however that these effects are re-
A careful examination
on benzene or
of the emulsifying
powers of inorganic salts
120
this connection.
We
.a
The
COONa + HCl
soaps are in
all
tion
by hydrolysis
the
oils,
and caustic
10.
soda.
By means
prepared;
emulsion
these
may
two phases.
brittle
of the dense
AGEING- OF EMULSIONS
121
case with gelatine (Nugent, Trans. Farad. Soc. XVII. 703, 1922)
where the gelatine film apparently gets thicker on standing.
Nugent noted
Nugent assumes
with age and that the caustic soda strips the film layer by layer. The
inhibition period is thus a measure of the thickness of the gelatine
layer which evidently on this view increases with age. After
72 hours ageing the inhibition period was found to be constant at
75 minutes, the adsorbed gelatine film now being in true equilibrium
with the gelatine in solution, this limiting inhibition period varies
The mechanism
sion with soda
is
by no means
On
emul-
clear.
122
LIQUID-LIQUID INTERFACES
NaOH ^ Na gelatinate + H
0,
-j-~k
or
= Kt.
7
at
The
is
concentration.
Iso-electric gelatine is however but
sparingly soluble in water and although the first layer of gelatine
may be produced by adsorption subsequent layers may result as a
by condensation on these
CHAPTER V
THE GAS-SOLID INTERFACE
1.
Adsorption.
The property
similar
brought into contact with a solid surface will be adsorbed into the
surface to- an extent which is dependent on several factors, the
nature of the gas and
gas in the solid to form solid solutions or compounds with the solid.
The term 'sorption has been proposed by McBain to include the
The
will
Crystals.
amount
and
if
section be
surface
Gibbs-Helmholtz equation
The
GAS-SOLID INTERFACE
124
but
may be
in composition.
surrounded by a
field
is
but
little
but that crystals develop those facets such that the interfacial
surface energy will, for equal values of the forces of crystallisation,
become a minimum.
125
We
Valeton (JPhys. Rev. xxi. 106, 1920), Niggli (Zeit. f. anorg. und
angew. G/iem. ox. 55, 1920) and Tertsch (Zeit. f. anorg. Ghem. cxxxvi.
205, 1924) have attempted to correlate the rate of growth of a
crystal face with the interplanar spacing. Thus in a face centred
cubic crystal of the rock salt type the interplanar spacing for the
1
in the respective ratios
^
V2 V3
:
r4fc
-~- whilst
planar distances, but we must ascribe different valencies or adsorptive powers to the planes themselves.
Tertsch assumes that these stand in the ratio
thus giving
or 1
V2
rise to velocities of
3,
growth
Between the
alternate Na'
and
01' planes,
growth
in the
of opposite sign round it. An ion
containing six groups
round it and
100 plane has only five groups of opposite sign
the six which may be regarded as
requires one more to complete
ions in the 110 and 111 planes
its coordination number, whilst
their sets.
ions respectively to complete
three
and
two
require
of atoms
note that the density of surface packing
we
Again
may
G-AS- SOLID
126
INTERFACE
may
We
more
closely
packed
atom
V2
r~
JU
ty
\/3
-^O
Tertsch
(loo. cit.)
makes
the somewhat arbitrary assumption that for these metals the surface
valencies vary in the order 100 110 111
4 5 3 giving mean
velocities of growth of 2 1'767 1-732.
:
for
the planes.
A crystal
'
127
Joio
(a a) are
The
The
Similar computations have been made for the edge energies of such
cubic crystals. Jones and Taylor
3.
(loo. cit.)
Agglomerates.
at low temperatures
1916) which,
if
GAS-SOLID INTEBFACE
128
is
surface field of force and a greater free surface energy than a crystal.
temperatures different surface energies and will consequently exhibit differences in adsorptive powers although smaller in magnitude
than in the case considered.
On
become con-
AGGLOMEEATES
129
The
is
accompanied
x=
(ss)
let*.
A preliminary
fall
Dunn
is
due
to a sinter-
s. c.
GAS-SOLID IOTEKPACE
130
for a short time. It
first
formed
pseudomorph
it is
Many
to
is
"activation"
by
alternate oxidation
given by the ideal diffusion equation and approaches the rate given
by the empyric logarithmic equation of Tammann
t
where a and
A
is
= ae ** - a,
1
b are constants.
likewise to be noted
Phase discontinuities.
PHASE DISCONTINUITIES
131
We
affected not
ever increasing boundary between dehydrated and hydrated salt. In a similar manner it has been found
that the reduction of copper oxide proceeds from a nucleus of copper
increasing in rate as the length of the superficial interface between
oxide and metal increases. Again Langmuir (loo. cit.) has shown that
autocatalytically at the
when
it is
"promoted"
surfaces.
the surface
is
92
132
5.
GAS-SOLID INTERFACE
Porous materials:
solid offer
pare ash free charcoal, whilst in addition the presence of the gases
hydrogen, oxygen and nitrogen present in chemically combined
forms is almost unavoidable.
If condensation of liquid in the
micropores of charcoal
when
Under these conditions the liquid filling the pores will be included
in the amount of vapour adsorbed by the charcoal and
give an
erroneous impression as to the true extent of
adsorption. At the
same time
of a layer
preliminary to
1906),
Homfray
(Zeit.
ADSORPTION ISOTHEEM
133
174, 641, 1910), Geddes (Ann. der Phys. xxix. 797, 1909),
Richardson (J.A.GJ3. xxxvm. 1828, 1917) and others.
(ibid.
>
The value of a is dependent on the units employed for measurement, whilst n is a characteristic not only of the adsorbing agent
but also of the adsorbate. Thus at 20 0. the approximate values
of - for a few easily liquefiable gases are as follows
= 0'84
ammonia by
charcoal at pressures
between 3-5 mm., whilst from Travers' (loo. cit.) data on the
the following variations
adsorption of carbon dioxide by charcoal
may
be noted
GAS-SOLID INTERFACE
134
n
= av
r
T ~~
*-
in which
we
replace
P by
^~
P
_ Ttm
~J~I
~~
~~"
~T
as,
the
'
sq. cm.,
we
obtain
-JL*Z = ap
W
RTdp
- da- =
or
On
when p =
0,
<r
o- Q ,
'
I_i
aJRTp*
dp.
is
o-
= naRTp n +
]_
c,
whence
^
CT O)
CTO
a-
= F = naRTp n
n log
= n A + log ^p
ADSORPTION ISOTHERM
135
is
reduced.
It is evident that according to this equation there is no saturation maximum for a definite area of adsorbing surface. Indefinite
Ghem. LXXIV. 689, 1910), Langmuir (loc. cit.), Taylor and Rideal
(Catalysis in Theory and Practice) indicate in a perfectly definite
manner that in. many cases an adsorption maximum or surface
is obtained. Arrhenius (Medd. Nobel Inst. II. 7, 1911)
was the first to introduce the concept of a definite saturation value
(5) which varied with the nature of the adsorbing agent and with
the adsorbate. On the hypothesis that the amount taken up for a
small alteration of the pressure depended on the relative extent
saturation
p,
,doc_s
dp
as
a
0-4343
= A/f0
0-4343 --p.
r
,
or
ks
The equation
is thus
superior to that of Freundlich in that the
existence of a surface saturation is realised as well as a physical
c.c. at N.T.P.
per gm. together with the a constant
der Waals' equation for the respective gases.
charcoal in
of the
Van
GAS-SOLID INTERFACE
136
Although there is some experimental justification for the conception of a saturation concentration yet this equation containing but one variable was found by Schmidt to conform to the
experimental results but very indifferently. Schmidt (Zeit. Phys.
Chem. LXXVIII. 667, 1912) made the additional hypothesis that the
adsorption coefficient k decreases in magnitude with the amount
of gas adsorbed in an exponential manner, or
7n
fi/
where
is
*~_ In
A/Q
""
J-iffi
C?
os
s
dp
an equation somewhat similar in form to that proposed by Arrhenius.
That these equations do not, in actuality, express the adsorption
isotherm over very wide ranges of pressure can be seen from the
following values of k calculated with the aid of the Arrhenius
equation from the experimental data of Titoff and Homfray:
(Ct>
fl
where
-1_,l_
W\
}S = Ke
f-*t
-w-y.
s
01,
JVI
- is
is
ADSORPTION ISOTHERM
137
somewhat
x^ap
proportional
pressure
obtained an equation log - = & where & and k,
it
condensation and
on the surface may vary from a small
fraction of a second to an
In those cases where the strength of
is
energy
from shocks, provided by bombarding gas molecules and pulses
from the solid adsorbate, to equal the latent heat of evaporation,
GAS-SOLID INTEEFACB
138
the
mean
is greatest.
By
striking a solid,
is
unity.
On
where
/LL
gm. mols.
is
T the
K.
temperature in
may
is
almost
vacuum.
If the surface be partly covered with 'the gas the impinging
molecules will strike and adhere to the vacant portions.
At equilibrium a fraction &' of the surface will be covered with
adsorbed gas and a fraction
will still be bare.
We
thus obtain
+ & = 1.
The
unit area.
will
If
At equilibrium the
be equal or
we *
put - = t then &
v
-^
^
+ tfi
ADSORPTION ISOTHERM
139
If
&
where
j~- x,
the
TT
jtLence
os
N!
= -~
to,
r-
-.
abp
*-
+ ap
which is Langmuir's equation for the adsorption
isotherm, where a and 6 are constants, a being proportional to the
length of life of the adsorbed molecule on the surface. At very
low pressures where ap is small compared with unity we obtain
iv (1
since
//,
t/j.)
= %>,
os
= abp,
obtained.
It will be noted that on
saturated
when
it is
~
1
It
+ ap
'
carbon monoxide,
is
adsorbed extremely
tenaciously
7.
On the existence
&'
of nraltimoleeular films.
of view
was adopted
firstly
by Faraday and
at a later date
by
GAS-SOLID INTERFACE
140
for gases at
A, cm. 191, 1923) in which the total quantity of gas adsorbed was
found to exceed the amount calculable on the assumption of the
existence of a unimolecular film.
Some
On the other hand Carver (J.A.G.S. XLV. 63, 1923) obtained but a
unimolecular film for toluene on a glass surface after
prolonged
outgassing of the glass over, however, a very restricted range of
In
those
cases
where
the
unimolecular
character
of the
pressures.
film is exceeded alternative
hypotheses may be advanced in
We
explanation.
may with Langmuir adopt the view that these
cases are really exceptional, since these
particular gases may actually
be full of cracks and
go into solution, and the solids examined
may
fissures of
submicroscopic dimensions, thus leading to a very considerable underestimation of the available surface. Another factor
141
will in reality
be extremely unhomogeneous
in respect to its adsorbing power, the free atoms, the corners and
edges of individual crystals and the various crystal faces all
differing in the
A gas but
monoxide which
is relatively
strongly adsorbed will adhere to all
parts of the surface. Thus the saturation value of a gas or the
availability of a surface will vary in accordance with the system
chosen.
We
find for
such as copper varies with the nature of the gas. For ethylene
and hydrogen the ratio of the saturation capacities at
C. on
active reduced copper
at 450 0. the ratio
is
saturation capacities
marked
of the surface
10~4 There are thus areas of the highly reactive localised patches
which can adsorb and cause reaction in these gases, there are other
less active areas which can adsorb both gases but fail to bring them
into reaction and there exist still less active areas which can adsorb
only ethylene but not hydrogen. On analogy with the copper surface
it is probable that a small residual area exists on which neither
ethylene nor hydrogen is adsorbed at a pressure of 760 mm. and at
C. One of the most interesting cases which await examination
.
142
GAS-SOLID INTERFACE
where
and
and
- fC ~
Again
if
we obtain
t/*8b
so
and
/j,
will
be obtained.
life
laws of probability.
thus be b t c 2Tt and
of the adsorbate
is t
will
be governed by the
a particular t will
for
Hence we obtain
^r
ftjaSb
=^?
N J1+
and
M
ret*
M
e^*,
tfi
x,
curve
may be
may
obtained.
likewise be
it will also
catalytically
number
oxalic acid
for
is
Vavon and Husson (C.R. CLXXV. 2*77, 1922) have likewise clearly
demonstrated the existence of patches of varying activity in
platinum black by showing that the progressive addition of a poison
such as carbon disulphide will first inhibit .the catalytic hydrogenation of substances more difficult to hydrogenate suchasacetophenone
without appreciably affecting the activity in respect to hydrogenation of nitrobenzene which is readily reduced.
threshold value for the work functions and suggest that in these
cases also there are small patches of the surface associated with a
144
G-AS-SOLID INTEKFACE
is still
We
Similar conclusions
Hardy on
lubrication.
may be drawn
molecules
We
HEAT OP ADSORPTION
145
films
8.
When
an adsorbing surface is exposed to a gas or vapour adsorption will take place, being accompanied by the absorption or
evolution of heat. Such thermal changes have already been noted
in the extension and contraction of surface films of liquids.
Although the direct determination of the surface energy of solid
surfaces presents
there
no doubt.
is
occurs.
u
it is
rr
= "- T
dcr
dT>
we obtain the
value of
T j^.
Also
we can integrate
the above expression. This is evidently possible if we can determine
0.
the value of T at which a
By analogy with the liquid-air
solid-gas interface is obtainable provided that
surface
<r
and
-7
be possible.
There are three isothermal heats of adsorption of gases capable
of being defined
s. c.
layer.
10
GAS-SOLID INTERFACE
146
is
(&)
The beat
(c)
The heat
with a
fall
of adsorption at constant
is
in pressure.
The thermodynamical
relationship between these heats of adsorption has been developed in an exact form by Williams (Proc.
Roy. Soc. Edin. xxxvui. 24, 1918) who has deduced the following
values
(a)
From
we
of adsorption
A,y.
dU=
\ T=
pdV + <rdA,
\ydx
fdU\
(dV\ +
-j\dxjgT
,p
\CLOGjgT
where g
is
the
number of gins,
TT
,
= - fdU\
+*
JpV-(^ n
T
cr
fdA\
-j-
\aCC/T
of adsorbent adsorbing at
(dV\ fdp\ +
\dpj \docj T
a*)P\-J-}-{j}
/r
ff
and
(dA\
-I"
\dxJ T
\dxJg
where n is the number of mols. of gas in the system. The third term
represents the work done by the gaseous phase in expansion and
may in
(b)
general be neglected.
isopiestic isothermal heat of adsorption
The
(dU\
^pT=~(-j-)
v
_
\dxjgT
where
,ir ,(
+pV(-j\dsc
das'
chamber
*--'
dU \
HEAT OF ADSORPTION
In actual practice the heat
effect
measured
147
will
\dx and
be
-'!
which corresponds
to the
we
obtain
Clapeyron equation.
close
G.
equation for the heat of adsorption of ammonia by charcoal at
\ is calculated and determined in calories per c.c. of ammonia (at
N.T.P.) adsorbed, so is measured in c.c. per gm. at N.T.P.
Data
for
Favre, 1894; Langmuir; Eideal, J.G.S. cxxr. 309, 1922; Foresti, Gag. Ghim.
and Burns, J.A.C.S. 1923.
102
148
GAS-SOLID INTERFACE
Soc.
contact has noted that the potential required to break down the
between platinum, tungsten and carbon films varies
loose contact
unaffected
by
filament.
sq.
cm.
may be
unimolecular
expressed by
the equation
3007
ergs,
where
desorption.
HEAT OF DESOKPTIOST
HEAT OP ADSOBPTION
149
effect
Lamb and
The
vapours
and
with various
tifee
GAS-SOLID INTERFACE
150
The
schaum has been
likewise
ammonia by meer-
XIX. 21, 1883) over an extended range of pressures with the follow-
ing results
In this case
approached at high
heat of wetting
for various
we
are in all
We
10 22
atoms and that the saturation maximum of ammonia is ca. 158 c.c,
at N.T.P. or 4 x 10 21 atoms; the adsorption maximum of bromine
10 millimols or 6 x 10 21 atoms. Still larger quantities of other
gases, e.g.
S0 2
are adsorbed.
platinum.
and
regarded as
some 40 mm. these
behaving
up
have
determined the pressure exerted
investigators
by a limited
amount of water vapour in contact with
glass and platinum surfaces
over a range of temperatures. In all cases the
pressures observed
vapour may
as a perfect gas
to pressures of
if all
HEAT OF ADSORPTION
151
be deter-
Two
series of determinations of
below.
ADSORPTION ON GLASS.
ADSORPTION ON PLATINUM.
9.
The
GAS-SOLID INTEBFACE
152
layers permits us to
faces
From
(Rideal
geneous gas
be drawn in respect
Practice) a certain number of conclusions may
to the valency of the adsorbate and the number of elementary
adsorbate occupies or adheres
spaces on the crystal lattice which the
to. If we consider a unimolecular reaction to occur catalytically at
n
the rate of evaporation of the unchanged product will be v6
Provided the chemical reaction occur but slowly we obtain
(1)
a/i
= i/0 n
m adjacent
change
will
be given by
(2)
Hence from
(1)
and
^=
z,2
(2)
m.
dec
The velocity will thus vary as the m/n power of the gas pressure.
In the case of the dissociation of hydrogen molecules into atoms
at the surface of a hot metal wire Langmuir found m/n
^, whilst
for the recombination of atoms at a surface m/n was found to be 2.
Thus if we insert the values m I and n = 2 in the first case
and m = 2, n = 1 in the second, so that the hydrogen atom is held
on by one elementary space and the hydrogen molecule by two
elementary spaces, the conception of a uni- and divalent union is
readily realised. In this way a "valency" of two for carbon monoxide,
sulphur trioxide and ammonia on platinum two for the oxygen molecule and one for carbon monoxide on quartz can be established.
;
153
Another method
adsorbed molecule
catalysts.
lies
that small quantities of sulphur poison, so far as hydrogen adsorpand catalytic hydrogenation are concerned, palladium black.
tion
The
compounds has
likewise been
CO union
group WOO. At higher temperatures however the
becomes weaker and the gas may be desorbed unchanged. Again
he noted that small quantities of oxygen inhibit by selective
adsorption the dissociation of hydrogen at the surface of a hot
tungsten wire but the adsorbed oxygen is not as readily reduced
WO
the
At
to
is
0=0
>
0.
At high temperatures,
as is well
known,
154
GAS-SOLID INTBEFAOE
oxygen
is
superficial
orientation
affected.
by a
containing
o.
550, 1922, A,
ci.
487,
155
that this group was the only one taking part in the processes
occurring at the surface of the copper.
10.
We
mean
where
equal to the latent heat of adsorption at that particular temperature. If the mean length of life of such a molecule
is
or -
^
= re
r -A
dt
It is possible that
is
the
mean length
number
is
pressure temperature curve for a unimolecular film, but investigations on this point would prove of great interest.
If we assume an equal probability for all adsorbed molecules to
evaporate we can readily calculate the average life of a molecule
156
GAS-SOLID INTERFACE
life
at high temperatures.
life of carbon
monoxide on platinum
op(l-0),
where p
is
The
rate
of the
157
amount adsorbed
adsorbed molecules
may be computed.
The most
is
more
reactive.
In considering homogeneous
The
The active
or
and probably
molecules in a higher quantum
and
is
moving molecules.
N^Ne~^.
In the case of the adsorbed molecules a certain number of these as
we have seen are capable of re-evaporating. If Ev be the energy
GAS-SOLID INTERFACE
158
_JBv
time will be
RT
For reaction,
nickel,
where
E^
is
e.g.
the relative
Evidently the catalytic efficiency will depend upon
We
Thus
in the reaction
->2H
the hydrogen molecule must be excited to a definite level to promote disruption, the quantity of energy to be supplied is some
96,400 calories per gm. molecule.
On
a catalyst
reactions
series of
of the adsorption
compound
HX
2
to
two
possible
159
molecule, (6)
2
2 into HXHX.
HX
Thus we
for
each
H ^2H,
2
at all
compounds
X H ;2HX
2
(1)
result in the relatively high concentration of HX or adsorbed atomic hydrogen, which although not so reactive as gaseous
may
The catalytic
far as
efficiency of
the hydrogen
is
a metal
for
concerned on
(1),
firstly
equilibrium is attained and thirdly the reducing power of the adsorbed atomic hydrogen as measured by the decrease in free energy
occurring during the process of adsorption or conversely the energy
from X 2 2
of excitation required to form
HX
11.
An
quantum state.
The active molecule must
GAS-SOLID INTERFACE
160
Reaction
^ CO +EtOOCOHK
COOEt.
may
E cals.
Solvent
Nitrobenzene
Tetrachlorethene
alcohol
Propyl alcohol
Benzyl alcohol
per
gm. mol.
Amyl
...
. . .
28,530
30,390
33,190
33,750
35,920
161
bromacetophenone.
We
may be made
in a
were obtained.
Assuming that all the surfaces are equally uniform and equally
covered under the pressures (11'5 mm.) and at the temperature
C. at which the reaction is carried out it is evident that the
11
162
G-AS- SOLID
INTERFACE
number of
and the
HCOOH-C0
all
cases exhibit a
163
similar uniformity and thus not only a constant area of surface but
also a constant area of catalytically active surface,
12.
The
effect of
Maxted (J.O.S. cxxi. 1760, 1922), who has examined the effect
of progressive additions of poisons to catalysts such as platinum,
came to the conclusion that the retardation in catalytic activity
was proportional
to the
amount
of poison adsorbed.
A critical
ex-
is
gases and effect catalysis; less active areas adsorb both gases but do
not promote catalysis and still less active areas adsorb ethylene but
not hydrogen. Diagrammatically we may represent our catalyst
surface as consisting of three types of surface.
112
GAS-SOLID INTEEPAOE
164
From
prepared copper catalyst are (1) ca. O'Ol %, (2) 40/ (3) 60 %.
Bodenstein and Ohlmer's (Zeit. Phys. Chem. LIII. 160, 1905)
investigation on the catalytic oxidation of carbon monoxide at the
,
reaction rate
is
accordingly given
dx _
dt
by
O3
CO
'
On the crystalline quartz on the other hand the gases are but feebly
adsorbed and the rate is governed by the probability of reaction
between oxygen atoms and carbon monoxide on adjoining elementary
spaces or
The
catalytic operations the possibilities of which have been demonstrated by Vavon and Husson (C.R. CLXXV. 277, 1922)
(p. 143).
Acetophenone does not undergo hydrogenation with such ease
as cyclohexene. In the
of
active
black
both
presence
these substances
may
platinum
be hydrogenated, but if the very active areas
the
addition of carbon disulphide
by
acetophenone
is
The hydrogenation
165
C2
CH CHOCH
3
+CO.
An
EI -^Kg
1*54 approximately;
CHAPTER
VI
We
freshly formed
o-sZ
this value
by the measure-
ment
and
others.
On
167
Sw
for dilute solutions,
RT log 5- = Bw RTlog -
i.e.
lig
SwRTlog-cs
-^
p
~--[Bw
r.
or
RTlog
s
where r > 0'0002 cm. is 15'83 millimols per litre. From these data
cr gZ = 1100
ergs per sq. cm.; in the case of barium sulphate <rs was
found to be 1300.
The more recent experiments of Dundon have indicated that
i
owing to the
hydrate.
Dundon
salts.
LIQUID-SOLID INTERFACE
168
At equilibrium
a dynamic interchange between the ions in solution and
those on the surface. In the vaporisation of an ionised salt, such
also the ions of the salt in solution (see p. 185).
there
is
of the ions
by
the same, but by the removal of a barium ion from the space
and its replacement by a radium ion. In a similar manner
lattice
crystals of silver chloride will adsorb silver ions into the space lattice
of the crystal
presumably where the chlorine ions are exposed.
maybe
calculated
if it
where
e is
electric
charge
an
to
K,
the
s.i.c.
of the
Zeit. v. 91,
medium and
1909; Knapp,
The
169
'
or
"-*
pr
r
'
STTKT*
_s"i /
~"
The displacement
is
-i
'
167r/ sZ
167r/ecr
formed
(see p. 210)
which
The
between
& and
energies
by
crig,
cr
energy liquid-solid by
<r te ,
simul-
o-fftXrig
+ a-ig*.
Similarly a second liquid (2) will displace another liquid from the
surface if
(ii)
<r ls
2 i}.
rare occurrence.
*
We
This displacement only takes place by a process of evaporation and condensation not by spreading, see p. 91.
LIQUID-SOLID INTERFACE
170
water from the surface of metallic sulphides but that the converse
takes place at the surface of powdered quartz, a phenomenon which
is
whilst the two liquids are brought into contact with the solid. The effect of preferential
is
rigid
(1)
If
<T SI
particles
and
(2)
(i)
(ii)
all
("i)
the solid
finally if
two liquids, since both liquids are adsorbed by the solid and
exist in the adsorption layer as a homogeneous mixture (iii). (See
of the
p. 105,
Ch.
iv.)
with ether, chloroform and benzene, but are removed from the
water by preferential wetting in the case of butyl and amyl alcohols,
whilst the reverse holds true in the case of aqueous
suspensions of
arsenious sulphide.
HEAT OF ADSORPTION
which
affects the
171
aqueous phase.
4.
It has already
vapours
data of Gaudechon.
A more
areas of solids
range
is
it offers
a promising method of
172
had occasion
<r
and Lt -j^ = 0.
21 =0"-i
-7
r=o
a,JL
&*-
between the
a-
= u + aT*
-~o
is
markedly
different to that
p. 19).
per
sq.
is
cited
HEAT OF ADSOEPTION
48, 1918)
computes
for
173
Eng. Ohem.
XL
sq.
metres,
sq.
Lamb
Ewing (J.A.C.S. XLIII. 1798, 1921) less than 120 sq. metres.
The most exhaustive examination of the surface areas of various
forms of charcoal has been made by Paneth and Radu (Ber. LVII.
1221, 1924) who, on the assumption that a saturation maximum
obtained when the surface of charcoal is covered with a unimo-
is
attributed to pene-
surface.
of silk
Nitro-silk
Cupro-cellulose
Specific surface
1820
456
Type
Acetic silk
28
LIQUID-SOLID INTEEFACE
174
5.
The determination
by immersion
in
and Ewing
liquids
in
till
k.
solid
was
Svenska
originally
Vet. Akad.
Handl.
Coolidge
(J.A.0.8. XLII. 1146, 1920) considered that the net heat of adsorption was due to compression alone and that with charcoal the
compressive force was substantially the same for all liquids, viz.
similar conclusion was arrived at
37,000 atmospheres.
by
may be
equation.
represented by the
charcoal -i
we can assume
may be
represented
by a term \ /
i
<
x
^2
y/ ^2
175
Such an
assumption however
on
it
is
state, of v2
If V Q
solid calculated
without
Hence
v
v
or
+ av = v + av
z
= VQ +
Oi
~v
a ).
= 0-511,
= 0-75 c.c.,
va
CC1 S H = 0-461,
a CC1 8
H = 0-85
c.c.,
^5 _
0-75
^.9 = o-05.
58,000 atmospheres
This hypothesis of a compressional force of great magnitude on
properties
change
a solid-liquid system.
transition from a solid- vapour to
176
It
LIQUID-SOLID INTERFACE
is
We
employed as immersion
small.
and Ewing.
Charcoal
Density
Coconut
2-12
Sugar
2-26
2'28
Graphite
The
removed by the
atmospheres.
An
these,
two hypo-
knowledge
NON-ELEOTKOLYTES
177
which water
is
If in
calculated.
of the occluded
ments.
Non-electrolytes.
Such
selective
adsorption.
As a rule the solutes
removal
is
by
solid
vapour
==
"b
vapour
"ab>
where a and b are two phases in equilibrium (in his limited definition two mutually saturated liquids) of which a has the higher
surface tension against vapour.
Now Antonow's rule appears to be valid for
s. c.
12
178
LIQUID-SOLID INTERFACE
Tims
""charcoal, water
"cliarcoal,
Now
KOI solution ~
rw
"c air
"c air
"H 2 O air
"*
.(ii).
"c
0"e
KOI
aq.
air
we have
assumed crcliarcoalj water is equivalent to crc arcoal saturated with water, water
as demanded by Antonow's rule. There is probably a not inconi!
p. 58).
possible explanation of
mercury saturated
may be used as a
in water are positively
(cf.
This difference
why some
solutes
179
NON-ELECTKO LYTES
is
which period secondary chemical action may take place. For 'comor
parative purposes adsorption as distinguished from absorption
sorption (J. W. McBain, Phil. Mag. xvm. 6, 1909) is considered to
take place rapidly in solutions as well as in gases (see p. 123).
The early experimental data of Freundlich (Kapillarchemie)
indicated that the adsorption of a solute from a solution followed
the general laws governing the adsorption of vapours by solids, i.e.
where x
is
the
constants.
The
diverse manner, as
f
{
||
1T
is
exemplified
by the
following:
122
LIQUID-SOLID INTERFACE
180
Freundlichj Kapillarcliemie.
7.
The
who found
that the
amount
H OH, CH C000 H
C H CH CH OOOG H
Blood charcoal
...
C013 H, C 2
Sugar charcoal
...
CC1 3H,
3,
5,
C 6 H6) CeHgOHg,
6H6
2 H 5 OH.
,
Although
Freundlich
acids,
(ibid. p.
who found
That the solvent plays an important par't in the adsorption process is evident from the various cases of "negative" adsorption
(Gore, Ohem. News, LXIX. 23, 1894; Lagergren, Bihang till k.
Vet. AJcad. Handl. n. 244; Trouton, B.A. Reports, 328,
1911; Gustafson, Z&it. Phys. Ohem. xci. 385, 1916) in which
Svenska
181
of decreasing as would be
instead
anticipated.
may
pass through
all
'
when
If in dilute solution
c is
proportional to concentration or
=u
wo =
where
k'
may be
c is
nearly unity
a similar type
7 ,
*?7J
or
or
If
i*
w
it
and
c,
(1
c)
u ir*t f*
-
a relationship of
,.,
\
Whilst in concen-
we have
w = (k'~v)(l-G),
w =
r/
is positive near
follows that near
T! n
is
c,
w}
(k
~ kc,
or
p.
182.
182
LIQUID-SOLID INTERFACE
Osaka (Mem.
iodide.
(Zeit. Physik. Chem. cxi. 405, 1916) has attempted
extend the conception of Williams that solvent and solute are
both adsorbed with the aid of .the Freundlich isotherm in the
Gustafson
to
following way.
In general there will be
u grams
of solute and
w grams of solvent
183
u = Jcc n
1
and
Woo
w = k'c n
',
by
an adsorbing agent is a matter of some difficulty. The most general
method suggested by Arrhenius (see Williams, loc. cit.; Osaka,
Mem. Coll. Sci. Kyoto, VI. 257, 1915) is to leave the dry adsorbent in
a desiccator exposed to the vapour of the pure solvent or solution
Thus there
is
LIQUID-SOLID INTERFACE
184
volatile
at
a higher temperature
method
for
material.
In some cases
it
maximum
OT,
saturation
of silver ions
was found that 5'5 and 8 5 to 9 mgm. of silver ions were adsorbed
by a square metre of metallic silver and gold respectively, such a
-
It
surface concentration
is
practically unimolecular.
The
adsorption
maximum
ADSOEPTIOKT OF IONS
185
8.
The Adsorption
of Ions..
With charcoal as adsorbing agent Michaelis and his coworkers (Lachs and Michaelis, Kail. Zeit. ix. 275, 1911 Rona and
Michaelis, Biochem. Zeit. xciv. 240, 1919; xcvii. 85, 103, 1920 and
Hartleben, Biochem. Zeit. cxv. 46, 1921) have found the following
salts.
NH
4 ",
K', Na'
ANIONS.
Organic colour acids
OH', ON',
S", T,
N0
3 ',
Br', 01',
and anions the more mobile hydrogen and hydroxyl ions are most
readily adsorbed, whilst in the case of the metallic ions the influence
primarily of the valency of the ion and both the position of the
With
it is
adsorbed.
LIQUID-SOLID INTERFACE
186
where in general a 4=
'
d and
=}=.
'
The
On
ion concentration
Identical
maximal values
were found
and hydroxyl
HN0 H
,
187
may
take place; thus the silicates adsorb methylene blue hydrobut an equivalent of calcium ions replaces the methylene
chloride,
calcium silicate
+ methylene
blue
HC1
>
methylene blue
+ calcium
silicate
chloride.
earth have been replaced by hydrogen ions from the earth, according
to Michaelis not directly as such but by resulting from the liberation
of calcium ions, the calcium chloride formed then undergoing
hydrolysis.
Bancroft
(p.
and the adsorbing agent selectively adsorbs the acid H'X' the
hydrolysis will evidently proceed further until equilibrium is
attained. In the case of Fuller's earth it may be considered that
the earth selectively adsorbs hydrochloric acid leaving an alkaline
solution of caustic soda. This view of promoted hydrolysis
by adsorption is at variance with the hypothesis of ionic interchange and
can only be tested by careful analysis of the solution to find out
whether ionic interchange has taken place or no.
Michaelis and
Rona
(loo.
cit.)
the mechanism
is
hydrolysis.
Bartell and Miller (J.A.C.8. XLVI. 1866, 1922, LXV. 1106,
1923)
on the other hand in their investigations on the adsorption by clean
LIQUID-SOLID INTEEFAOE
188
is left acid.
The
operations of dyeing.
The
The fact that not only may the different surfaces of a crystal
have different surface energies due to an alteration in the closeness
of the packing of the molecules on the planes but also that the
molecules at the corners and edges of a crystal cannot adhere so
strongly as those in the interior of the plane surfaces give us, as
Gibbs (Scientific Papers, p. 325) has pointed out, a great deal of
solution.
The
conditions
molecules per
c.c.
and we consider a
mean
free
path of
N^ molecules, then
energy
-JL
N' rE=a
1_
J~
VTT
fl
\
H=\
-r,
&TJ
Vv J^_
WT
ti>-L
The number
is
189
equal to this
is
the
N =N
s s,
Ns /=
= ^ = -l
<j>v
VTTJ
JB=
A=
is
The molecules at the surface of a crystal move only in a vibramanner about an equilibrium position perpendicular to
the crystal surface; the number escaping per second will evidently
be proportional to (a) the number on the surface layer, (&) the
number having at any instant a kinetic energy greater than X per
gram molecule, and (c) the number of times on an average the outward movement is repeated per second (assuming that the normal
tional
TT
On
where
integration
if
RT
RT
-^
this reduces to
2i
*,
where
A' =
190
LIQUID-SOLID INTERFACE
In a saturated
and deposition
exists, or
or
Writing
0=
~! x 1000, where ^V
n = Wt-Wi
loge tf
-^
lo
= the
Avogadro number
I W,
'
&IoooW
Wz
liquid).
necessarily follows that a certain small but finite degree of supersaturation is necessary to permit a crystal to
grow.
and
which
o-fiZ
is least,
191
We
sion of Gibbs:
"The
effect of dissolving
a crystal
is
therefore to
of crystals
(see
Again a
this
way
also
exposed edges
may
dissolve
plane surfaces.
11.
The
by the
early experiments of
dx
by the equation
.
=&(-*),
Nernst
(ibid.
689, 1897),
56,
1904),
LIQUID-SOLID INTEEFACE
192
by the Fick
dt
layer.
to the
mechanism
in
acetic,
As examples may be
cited the following values for the solution rate of small rods of
magnesia rotating at a constant speed in the solvent:
The values
The
is
lie
water.
We may
193
The
must be regarded
as migrating
acid.
Fischer
rotating cathode:
Dushman (Jour. Phys. Ohem. XIV. 885, 1910) found $ = '040 mm.
under similar conditions, whilst Brunner (loc. cit.\ found that Sn%
was approximately constant.
'at
and
fats
R. S. 0.
1922)
who
by aqueous
alkali.
13
LIQUID-SOLID INTERFACE
194
12.
In the
of a nucleus once
it is
formed.
Nucleus formation.
greater than that of large ones, hence before a nucleus can act
as a centre from which spontaneous crystallisation may proceed it
is
Adopting
Langmuir's hypothesis of inelastic collision each contact between
two molecules of solute in a solution results in an adhesion for a
definite although short period of time, the
periods being in all
somewhat longer than in the case of gases. In
probability
general
the problem of adherence on collision in solution is
complicated by
the fact that both ions of
electrolytes and molecules of non-electro-
necessary
postulate either
contact between unsolvated molecules or
ions, or a movement of
the solute particles towards one another within
their solvated
A somewhat
similar
agitation by
which intermolecular contacts
are produced; at the same
'time
the same
agitation is the operative
process by which rupture is
caused between
adhering molecules. Thus there will be a maximum
RATE OP CRYSTALLISATION
195
The
existence of auch a
of solutions by
Rate of
13.
Tammann
maximum
(Knstallisieren
und Solimdzen,
p. 151).
crystallisation.
growth
face, e.g.
the faces
It
wan
many
crystal,
may bo regarded
as
an ana-
vin.
expression.
where
is
(7*
The
LXVH. 470, 1909, Lxvm. 104, 1909, LXXIIL 685, 1910, LXXV. 710
1.911, LXXDC, 71, 1912), Le Blanc (ibid. LXXVH. 614, 1911), and
others.
132
LIQUID-SOLID INTEEFAOE
196
The
crystallisation.
The
as iron in acids is
as
is
exemplified by the following figures for the retardation caused
by the addition of small quantities of gum acacia added to
normal
The
addition agent)
from attack.
insoluble
is
An
many
cases of passivity.
is
marked and
definite.
xui.
17, 1.902) in his investigathe lowering of the crystallisation velocity of fused benzoGli&m,.
tions on
phenono in tho presence of various substances came
to the con-
olectrolytically deposited
Tho effect of such addition agents, which are strongly adsorbed,
not limited to an alteration, in the velocity of crystalline
is in
general
affects the crystal habit. The adsorbed
growth, but frequently
material lowers the tntorfaeial surface energy of the crystal arid
LIQUID-SOLID INTERFACE
198
thus
too
Z&it.
few cases have been recorded (Dreyer, Z&it. Phys. Ghem. XLVin.
487, 1904) where the effect of the addition agent on the crystallisation velocity is to be attributed to an alteration in the diffusion
constant of the solute in the solution, thus the
velocity of crystaland not decreased by the addition
of
methyl
15.
The
effect of
dispersing a solid in a liquid is to produce a solid
disperse phase in a liquid dispersion medium. Such colloidal
systems are termed "sols" dispersoids, suspensoids or suspensions.
Suspensions may be produced by condensation from soluble salts
in solution or
by the disintegration or dispersion of a massive solid
We
METHODS OF DISPERSION
199
stability.
16.
Methods of dispersion.
ag-
investigated
in detail
Colloid
Chemistry, 1918)
where he adopts Freundlich's interpretation of the mechanism of
the process. On the adsorption of a liquid by a solid the surface
somewhat
difficult to
may be
the molecules in the crystal is less than the adhesion of the water
to the molecules constituting the surface of the crystal. As the
solution.
200
LIQUID-SOLID INTERFACE
The
The
interfacial surface
Kolloid. Chem. XIV. 65, 1914) has noted however that processes of
mechanical disintegration of metals are greatly facilitated by the
addition of gelatine. The metal particle which has adsorbed the
gelatine is now less sensitive to precipitation by electrolytes and
has in fact acquired properties more akin to an emulsion than to
a suspension and is said to be "protected."
method
power of these
stabi-
PEOTECTION OF SUSPENSIONS
Using
colloids.
201
The coagulation
a
taining 0-0058 / of gold to occur on the addition of 1 m.l. of
ten per cent, sodium chloride solution.
The gold numbers of the various protective colloids vary greatly
offers
may be
cited.
STARCHES.
ALBUMEN.
White of egg
Albumin
Wheat
0-08
0-1
Dextrin
0-3
Ovomucoid
0-020-05
0-040-08
Crystalline albumin
Alkali al bumin
0-00140-14
Globulin
starch.
...
46
612
...
Potato starch
...
25
28
Gum arable
Gum tragacanth...
0-5
Gelatine
...
0-0050-01
Isinglass
...
O'Ol
Casein
0-01
Na oleate
Na stearate
0-41
at
6 C.
100 0.
Stannic acid
0-02
10
0-001
00
electrolytes.
202
LIQUID-SOLID INTERFACE
lytes is
18.
filtration.
Two methods
of dispersion
by
electrical
Davy noted
xi. 521,
disintegration
of a lead cathode in alkaline solution is due to the formation of a
lead sodium alloy which
reacts
with
the
water yieldsubsequently
ing a fine black dispersed lead suspension. This electrical dispersion
203
still
such as
effected.
some forty
usually sufficient to effect the dispersion of most metals
platinum, gold, palladium, iridium and silver. Many metals
may be
fall
of
when
and insertion of a condenser in parallel with the arc gap and reducing the current density of the arc Svedberg noted a great
improvement in the purity of the metals dispersed in this way but
decomposition of the medium was still marked when highly
reactive metals such as the alkalis, or catalytic metals such as
"
"
hard
cobalt, nickel or platinum were employed, whilst several
metals,
e.g.
this treatment.
coil
as source
electrodes a great
sols was achieved.
The mechanism
of colloid production
204
LIQUID-SOLID INTEBFACE
of definite dispersity but one that is not necessarily stable. The processes of suspension formation by methods of condensation on the
other hand are more complicated, owing to the fact that unless the
resulting colloidal suspension possesses at least some degree of
stability the process of condensation once set in operation will not
cease but proceed until the transformation to the macrocrystalline
structure
is complete.
conditions of effecting condensation so as to
produce a
colloidal solution have been
investigated in detail by von Weimarn
The
colloidal solution
and
may
forming
dependent on the formation of small
which
as
has
been
crystals
pointed out are produced in solutions
in which nucleus formation
readily obtams,but where the subsequent
of
the
minute
nuclear crystals is
growth
relatively small.
colloidal solutions are thus
The
is
METHODS OE CONDENSATION
amount
and on
its
205
solubility,
thus
if
is
initial
dn
at
-nS
=R
or
s
R^Ps =
Tr
U,
~ *"*
~~-Sx
5
dt
These two equations thus define, according to von Weimarn, both
the nature of the disperse phase produced by condensation and the
of the disperse in the form produced at the moment of separa-
life
tion.
number
to the formation of a
If
on
will be highly
disperse phase separating for large values of
or suspension
dispersed, conditions favourable for true dispersion
formation.
Weimarn
206
LIQUID-SOLID INTERFACE
supension, a gel
is
when prepared
in this
METHODS OF CONDENSATION"
207
by
by
Methods
of condensation in
little specific
surface energy.
colloids are
which protective
em-
of
suspensions.
CHAPTER
VII
The
a homogeneous
throughout the body of
but the possibility of obtaining an electromotive
electric potential
is
constant,
phase
from the terminals of a
force
of homogeneous
potential
observed that the surface
Warburg (Wied. Ann. XLL 1, 1890)
tension of the interface between mercury and dilute acid decreases
as the amount of the corresponding mercury salt present in the
solution increases.
by McLewis
(Zeit.
T=
^7n~7i
RTdlogc
would o
give the
"
"
pure
r
of undissociated
adsorption
r
is
taining the
Hence
salt,
_ _ da _
dor
96-5xl0 10 dF"
209
is
down by Gibbs.
The Helmholtz electric boundary
as laid
2.
layer.
moment by the
expression
F=47rJf xSOO
(1)
may be
expressed
more fully in terms of p, the volume density of charge at a distance
SB from an
imaginary plane drawn somewhere in the non-homogeneous region between the phases, thus
sq.
M=$pa;Sx.
(2)
He
from the upper to the lower electrode, thus completing the electrical
circuit.
210
height
electrode surface is a
measure of the
maximum
value.
throughout the bulk of the electrolyte. Hence without serious error the
V
E.M.F.
be
to
the
difference in the values
applied
may
equated
of V at the two electrodes; and since V at the
"unpolarisable"
electrode remains unaltered the variation of a with V is observed
at the cathode.
With the aid of thermo-dynamic
reasoning,
relation
,
dor
dV
=-ex
10*,
.(3)
where
to the metal
and
211
e is
the
between
it
and the
electrolyte unaltered.
3.
moment
He
called
surface-tension
F=0
layer
since
by (1)
given by
F=Fmax
by Lippmann's equation,
.
'
.(4)
47T6
we
curve.
be
in-
versely proportional to the square of the area they occupy and thus
inversely proportional to the fourth power of their distance apart *,
a
a,
142
212
S.
Lippmann
therefore
Gouy (Ann.
in general
de Phys. vi.
Fmax
following table
Fj/^Tre
5,
is
At the beginning of the present century, when molecular magnitudes became more exactly known, it became apparent that S was
smaller than an atomic diameter. It was therefore suggested that
the charges of the double-layer must be regarded as separated by
a dielectric medium *.
Equations
evaluation of
(1)
and
M all
213
curve.
If
is zero at the maximum of the
electro-capillary curve, it is
evidently possible to find the absolute potential difference at a
standard electrode by using it as the anode of a capillary electrometer and observing the E.M.F. which must be applied to reach the
maximum
experiments,
Nernst's theory
is
correct
Nevertheless, for
N/10 KC1
an applied E.M.F. which differs very little from 0'53 for the decinormal calomel or 0'48 for the normal calomel electrode, and it has
been several times suggested that the simple Helmholtz theory is
which
applicable in such cases, but that for electrolytes such as KI
are in
Gouy
many
organic compounds
maximum
Later
;
Kriiger's theory
(Zeit. JElektrochemie, XIX.
normal
KN0
capillary curve
that for this solution alone
for
is
214
K N0
is
less
than those
for
some other
arbitrary.
5. Experimental verification of Lippmann's equation.
Lippmann's method of deducing equation (3) has been criticised
on the ground that it ignores the "residual" current which continues to traverse a capillary electrometer after the surface charge
on the cathode has taken up the value corresponding to the E.M.F.
applied.
is
an irreversible
dynamic treatment.
The very
always
(Phil.
valid.
Mag.
VI.
of
.215
Lippmann's equation
is
On
the form
da-
r=
96'5xl0 10
This,
!7*
= 96-5 xlOT.
and the charge per unit area has been traced by Schofield*, by using
is
which
it
is
in contact
when
it
its ions,
metallic ions
vironment," but
if it is
an ion the
* Phil.
Mag.
i.
phase
216
must
= 96-5
+k
I0
'
in the electrolyte
and
is
w
only independent of c when the latter is small, 96*5 x I0 n is the
7
charge on a grm. ion in coulombs x 10~ n is the valency of the
"
metallic ions, and km embraces both the concentration and en,
V= (ETlog
is
evidently independent of
c-[km - &J)/96'5 x
10 10
c.
Hence,
In
Were it supposed that at an interface the density of each component changed abruptly at an imaginary geometrical surface,
being constant on either side into the interior of each phase, the
amount of each component actually present in the system would
differ from that calculated on this basis by an amount depending
on the position chosen for this dividing surface. By a suitable
any one component can be
and in the present system it is convenient to so
that there is no surface excess or deficit of metallic ions
made
to vanish,
choose
it,
at the interface, since these are the one constituent present in both
phases. In accordance with this convention all the metallic ions in
e.g.
will differ and consequently some of them will pass across the
boundary or compensating adjustment of the surface will take place.
*
Butler, Trans. Faraday Sac. xix. 729, 1924, and Heyrovsky, Eoy. Soc. Proc. A,
on. 627, 1923, have given expressions relating V to the molecular properties of the
metal and the electrolytes. Their treatments may be regarded as attempts to
ls
m and k e in terms of the ionisation potential of the metal
of hydration of the ions In the solvent respectively.
evaluate
may be conveniently
217
transfer.
is
given by
two homogeneous phases, a particular molecular and ionic arrangement must be established in the intervening non-homogeneous layer
in order that the conditions of chemical and electrical
equilibrium
may be simultaneously obeyed.
The Nernst ionic transfer of ions to the metallic phase will give
a surface deficiency of electrons (the other component of the
phase), similarly their removal will cause a surface excess of electrons.
Since there can be no net volume charge the surface density of the
rise to
or
more
of these.
dpK St
"
is
likewise zero
of the
in
all
218
pelectrons __
=~
^"
=~
__
9 6'5
to the
^
10W
the
is
equal to
,
7-^-=
r- of the slope
r of the corresponding'
& electro-
96*5 x 10 10
t-
219
7.
The
structure of the
boundary
layer.
ions were kept separate from the surface or whether the charges
should be considered as separated by a dielectric medium.
From Helmholtz's equation it is possible to calculate the equi-
220
Whilst the
moment
electric
it is
moment
We
that since the apparent molecular electric moment varies with the
concentration that Helmholtz's conception of a rigid double layer
is
due to
M NaCl
is
and that
On
this
of uniform concentration at
is
its
surface
is
approached
which tend
to
Gouy
ideal solution of
calculated the distance of the " centre
and
calculated values
is
con-
221
of the
to
the extent of
0*15 x 10~ 10 grm. ions per sq. cm. The solution contains 10~4 grm.
ions per c.c., hence the slice thickness which must be denuded is
1*5
x lO"7 cms., since the tailing off is gradual in character this thick-
ness
is
minimum
probably steeper over the inner double layer and falling off more
gently across the outer or diffuse layer. In some cases at least it
is
is
may
outer layer
is
layer
Distance from
phase boundary
222
8.
Motion due
The
phenomenon termed
solid will
and on impressing an
electric
solid, if this
endosmose;
be kept stationary, a
been applied.
Electric endosmose.
ELECTEIC ENDOSMOSE
223
I.
855,
1882).
tube
Vi
there
is
.
cnargfe is
a
z
dV
-=
dec
We
is
The curved
the liquid.
is 2vrr
and the
electric
_dV
dr'
.
4nrq
hence
IT**
dV\
\~~dr)'
dV _
where
77
is
the viscosity.
du
224
Eliminating q we obtain
du
R dV ^Tr
dx
The
electric current
G is by Ohm's
_d_VA
dx
where
law equal to
'
TT
JbLence
r-
dr
This equation
is
wall
If the charges in the liquid forming part of the double layer are
in close proximity to the tube wall there will be no charge in the
bulk of the
liquid,
i.e.
or
j~
0,
The
where
is
= U say. Hence
= KA
-r
be um A
move
to.
ELECTRIC ENDOSMOSE
225
_2g(F -F,)J:
1
A
On
As has been
indicated
attributed
to
the selective
or
first
Zeit Phys, Chem. LXXIX. 385, 1912; Bethe and Toropoff, Zeit.
Phys. Chem. LXXXVlii. 686, 1914, LXXXIX. 59*7, 1915; Glixelli,
Bull, de I'Acad. des Sciences de Cracovie, A, 102, 1917; Gyemant,
Koll. Zeit. xxvin. 103, 1921). The general conclusions in regard
to the influence of various electrolytes on electric endosmose
may
be briefly summarised as follows
:
solutions hydroxyl ions are adsorbed and the liquid moves to the
cathode being positive relative to the negatively charged diaphragm.
hydroxyl ions
equal ionic adsorption has taken place, the double layer potential
or electrokinetic potential being at this point zero and the dia-
phragm
is
B.S.C.
226
thickness.
Gyemant (Koll
ELECTRIC ENDOSMOSE
227
diaphragms are found in weakly alkaline solutions, of the acid diaphragms in acid solution, as indicated by the following data:
it is
152
228
charge will be diminished and the velocity of electric endosmose reduced as is shown by the following data obtained by Pen-in (loc. cit.):
may
take place,
e.g.
229
ELECTRIC CATAPHORESIS
The attainment
of a definite
maximum
for
endosmotic now
this transverse
maximum
rate of
of naphthalene.
10.
Electric eataphoresis.
ment
is
layer potential
migration.
and equal
to
G both
and
Q-
being parallel to
as.
At
the motion
is
differentiating
<
=-
Vas,
slow
a
it
is
irrotational
potential
function
at
great
distance
230
is
~, where fB
/3
The
force
is
the coefficient of
friction.
is
motion
o<ty__ dV
PdS~~ q dfi'
This condition
is satisfied if
we
take
~r
=-
<
Hence
v
But
and
the
v is equal
everywhere.
y
also.
is
Hence
field.
if c
-j-
this part
we
obtain
and therefore
The
<?
%~
~jr~
>
where S
is
is
du
v
whence
-y
= KRG 57
<,
4-77-77
in an emulsion,
= 0-05
Now
McLewis
volt.
t=
5
^
K
'
ELECTBIC CATAPHORESIS
If r be the radius of the particle
'
or a
a
e
-.
and
,
and
,,
231
Se
jzrKr
For
E.S.U.
for
difference
the product
rjv
232
In the following
11
fall
sec.
(v)
aqueous solution.
demonstrated
brought
oxygen
The experiments
233
of
is
number
information
may be
obtained by
measurement of the
E.M.F. of a cell
of the type
Hg Hg2 Cl 2%-K01
|
solution
Air solution
B
gap
I
Hg.
234:
(Zeit.
Phys.
Ghent.,
xix.
625,
1896) and
narrow
of a
vertical glass tube down the inner surface of which one liquid is
made to flow, the other liquid emerges from a fine glass tip in the
form of a narrow jet down the axis of the tube. The two solutions
In spite of this apparent latitude, however, the following conclusions may be drawn from the measurements, in which distilled
water or dilute sulphuric acid was used as reference
liquid
(solution B).
Electrolytes.
KOH
With the
and
exception of the alkaline hydroxides NaOH,
possibly the fluorides KF, NaF, the direction of the E.M.F. indicates
that all monovalent anions are repelled from the free surface of
waters less than the cations
Li",
Na', K",Cs*.
slight differences.
the fact that the surface tension of
water
is
raised
by the
lie
CNS' > Cl O/ > T > Mn O/ > NO/ > ON 0' > Br',
BrO/>CN >Cl
/
>OH
/
,
F,
thus the extent to which the anions are repelled from the surface
with this
we
it is
In the case of
that the
CNS'
235
KCNS
is
falls
as
we
actually lowered so
is
may
be
Cl
it
the free surface possessing a net negative charge.
The monobasic acids all show a more marked E.M.F than the
KOH
monium
compound.
It ma,y be noted that similar lyotropic series for the anions have
been observed at other interfaces in addition to the air- water
interface.
adsorption
is
p.
CJNS'>r>NO,',
and Gouy
Br'>Cl'>S0 ",
4
at a neutral
(see p. 212) obtained the adsorption
Br'
mercury
236
Ions such as S", HS', T and ON' are probably somewhat displaced
owing to the "chemical" attraction they exert for mercury.
with a radio active material is used. The ionisation of the air round
the gas jet or radioactive material is sufficient to enable accurate
determinations of the potential difference between this air electrode
and the liquid. On account of the somewhat marked convection
currents caused by the flame the radioactive deposit on the platinum,
wire is to be preferred as air electrode. Frumkin has extended this
method to a variety of capillary active organic compounds such as
the fatty acids and alcohols including the long chain insoluble compounds examined by Langmuir and Adam. Although this method
has not yet been developed the results already obtained indicate
that this is a powerful weapon with which the vexed question of
the polarity of organic compounds might be attacked.
Non-electrolytes.
the few cases where the electromotive force can be accurately com-
and n-caproic
^.-valeric
the electric
observed potential
obtain
found that
whence
/u
F= 0'349 volts or
be
T we
it is
difference.
F=4rfl>,
= j-
^.
237
0'758 molar.
is
1B
of the
= 0'305 x 10~18 as compared with
butyric acid molecule is //,
a value of ca. /i = 2'0 x 10~18 calculated from the refractive index
7?)
(r5
where
is
Gouy
at a
for
8,
a value in close
(K
interface.
concentrations
of
for
of a
238
Water
itself in
maybe imagined
is
to orientate
represented as follows:
R R
R
+0
R
C+
falls
and
effect of
go over to a negative value, the total negative
- +
+
a C0 2 Na must thus exceed that of a C0 2 group. A representation
in a manner somewhat as follows would indicate a mechanism by
may even
R
0+
(-0
R
mm
--
H 0-}-
{0
0}-
Na +
thus
we
obtain
Vm
its effect
must be obtained.
the
239
normal
on the surface tension of water
for
effect of
01
01-
01
0+
+
{0
0}-
and
volatile to give
any
definite values.
The
iodo and
brom
acetic acids
gave similar
derivatives.
moment
effect
on the
of this molecule.
its
orientation as follows:
RRR
RRR
H+
R
OH-
Y Y
OH-
where
The
may be
OH which
240
In a
cell of the
type
M MX MX M
i
c2
V3
V2
algebraic
y-^+F.+ F,.
The two
electrode potentials
Nernst conception
and
expression
T
is
^F
l0
^'
M ion in the
thus
Fl
first electrolyte,
+F
due to Nernst
by the
RT,
Fl==
where
and
cell is
F,
M"
X'
iv.
129, 1889).
M'
X'
MX
c,.
solutions.
A fraction
u+v
where u
is
the mobility
J of the cation and v the
is
carried
by the cation
and a
fraction
241
fit
f\\
u+v byJ
the anion.
The
fraction
u+v
of cation
is
the fraction
is
The net
u+v
U+V
ca
RT-
or
The
electrical
perfectly reversible
u+v
Cz
log
F
For large values of the difference between the mobility of the
cation and that of the anion, the diffusion potential V2 will by no
means be negligible. Both hydrogen and hydroxyl ions possess
exceptionally large mobilities in aqueous solution and thus diffusion
potentials between solutions of different concentrations of both
acids and alkalis may assume large values as indicated by the
NH N0
16
242
X X
1
M' N'
and
X'
7'
xxxi. 363, 1907); Henderson (Zeit. Phys. Ghem. Lix. 118, 1907;
and Bjerrum (Zeit.f. EleUrochem. XVIL 391, 1911).
If the two electrolytes of concentration G and c 2 are monovalent
and the ionic mobilities u1} Vi and us v% respectively, then
RT
T-
where f
is
log^-logf
^ + lo
In those
cases where
CI
GZ
we
find
+ Vi
1
,
==
rr~
Jf
*0g
U%-f- V%
E7~
J?
-\
AJQO
==
n~ log %~
JP
A/2
As an
243
of only one kind of cation and one anion, but where the junctions
are diffuse and a relatively gradual conversion from one pure
electrolyte to the other occurs, the transport ratio will be fluctuating
from one extreme value to the other and the above equation of
Planck developed on the assumption of a sharp and well-defined
rigid.
(lac. cit.)
have
u2
V~
2
Cl
>
\
which undergoes
still
further simplification
when
GI
F
- W - <>! -Vz)
RT
MI +
(%!
^_
--^i
F fa u H- fa WB) iQQr& u + v*
= c2
O37
r o
"*
...........
-'---
""^r"" 7
2)
'-" --'
._-^-~
z]
equimobile
ion.
by Lamb
and Larson (J.A.Q.S. XLII. 229, 1920) by allowing one liquid to flow
against the other, whilst Bjerrum (Zeit. f. Elektroohem. Lin. 428,
1905), has obtained a relatively diffuse junction by the insertion
of a sand diaphragm between the two electrolytes.
similar
phenomenon
is
to
the
162
244
An
System
where CQH
is
is
for
which zero
CT>m
fff,.
where the
Since
suffixes
V=
a and
JLlJL
6 refer to the
at equilibrium
two media.
~^ p
r\
\Jfth
Ha
.
= -~ ^^
^016
^0016
sodium
chloride.
245
and
finite
equal to
= __
log
nff
where
L is
2i
QHL
a constant.
a
,
which
is
the limiting
<
NaJT
KOI
sat.
in
water
e.g.
NaJT
NaCl
in oil
in oil
NaOl
KC1
in water
sat.
Calomel
roi
Calomel
in water
In each
cell
J Cl
in oil
NaCl
in
oil
NaCl
Calomel
in water
||
giving rise to
if
The
phases be
Cj C 2 o a c 4
we
obtain
-V
VK K- V
K,
a
246
In one experiment
solvent,
of Beutner, guaiacol
assumed
an increase of
With
1'2 gemhos.
10
^ aniline
hydrochloride, the
conductivity rose to 10'3 gemhos. Inserting these values as representative of the concentrations of the chlorine ions in the various
solutions
we obtain
"
01
c.
.'.
'
10-3-01
10-2
1-2
1-2
= 8-5,
volt,
A closer agreement
volt.
no correction
has been made for the concentration potential b and the conductivity
rafcio of the two solutions is
only an approximate measure of the
is
The
is
clearly
The
marked:
Beutner
Haber
(loo. cit.),
Oremer
(loo. cit.),
247
cells
concentrations
an electrolyte
MX in two different
ca ,
cz
Water
MX
whilst the
minimum
In general the
value will be
E.M.F. of
Water
Oil
MX
F=0.
and at
c2
be
C 2)0 ii,
V=~
8
F log-
1
.
^^.
c2 c1)0ii
When
is
is
02,011
= 0^ Ou,
i.e.
of the
common
ion
^a.oil
^2, oil
is zero.
oil,
oil
KC1 + C
On
/OH
/OH
H \CO OH -* CH \CO OK + HOI.
4
two acids
salicylic
and hydro-
248
chloric are
phases the
verified
increasing
E.M.F. of the cell will gradually
fall,
an anticipation
by experiment.
may
own
methylamine
its
to nitrobenzene.
is
Calomel
MX
inH
non-aqueous
phase
in
which the
E.M.F. actually
non-aqueous
phase
II
measured
MX
in
Calomel
is
24:9
C6 H5 CHO
Oresol
-0-13 volt
Phenol
+0-13
Guaiacol
+O08
0*07
Acetophenone
The
oil
MX
of the type
MY
Oil
inHoO
cells
MX
and
auu
inH2
inH
Oil
NX
inH2 O
of
Utilising potassium chloride in the left-hand compartment
the cell and substituting for potassium chloride other potassium
salts in the
cell rises as
the
right-hand compartment of the cell, the E.M.F. of
the anion becomes more oil soluble. Beutner (Z&it.
Meotrochem. XXIY.
oil solubility
94,
of the anions
SO/'
01'
Br'
SON',
Mg".
Oa"
Ba"
Na'
was
K\
have
Hartridge and Peters (Proo. Roy. Soc. A, oi. 348, 1922)
measured the interfacial tension of liquid fatty acids and their
solutions in
Oh. ix).
250
occurring abruptly from one area to the other in the neighbourhood of H = 5 where the interfacial surface tension shows the first
signs of change.
with
oleic acid
For the
buffer solution
PH = 5*6
solution of palmitic
with the strength does not exceed the experimental error, but in
the more alkaline solution we must conclude that either sodium
ions or phosphate ions (or both) are positively adsorbed according
to the equation
^PH constant =
r and
IVa'
^Na* ~
FpOj," <^P0 4
"
(1)
= - rH
d/AH-
~ IW d/%a
(2)
GLASS ELECTEODE
this value causes
cm., whilst
dp
253
m &RTx S'303 x
-
41 dynes
per
2.
From
we assume Fpo4 =
We
charge.
- 5-5 > + FH
The
then FH-
+ TNa =
if
there be no volume
thus obtain
difference
2-303
RTx<2 = - FNa
2-303
RT x 2.
between
4-1
FNa
FNa =
.
2-303
FH-
RT x 2 = - FH
= 4-5 x 10 n grm.
PH = ^
2-303
RT x 2 = 5-0,
equivalents per
sq.
cm.
OH
CO
COONa groups. It
groups have been converted into
must, however, be observed that the method does not permit us to
discover whether or not the groups are ionised.
The
14.
glass electrode.
E.M.F. is
obtained equal to
-^ x
The
glass is
ions in
is
96-5
detail
by
xlO
10
The behaviour
more
Ann.
252
XL. 244, 1890), Borelius (Ann. Phys. XLV. 929, 1914), Freundlich
and Eona (Site. Preuss. Akad. der Wiss. 397, 1920), ibid, and
(Zeit. Phys. Chem. cxvi. 401, 1925), Hughes (J.A.O.S. XLIV.
2860, 1922), Horovitz (Zeit, f. Physik, XV. 368, 1923), Cremer
(Zeit. f. BioL xxix. 562, 1908; Zeit. Phys. Chem. cxv. 424,
Ettisch
1925).
It is found that in alkaline solutions
silver
to
may
When equilibrium is attained between all the electromotively active ions in the glass and solution the formal relation-
electrodes.
ship
must be
- KI.H-
= RT log CW
x 10 10
- KgNa - K Wft
96-5
x 10 10
obeyed.
and sodium
of ionic interchange
is
occurring hydrated
silicates
15.
Membrane
shared in
common with
glass
by the naturally
the zeolites.
potentials.
shown
may
permeable
MEMBRANE POTENTIALS
There
exist a
number
of
membranes which
253
are permeable to
some
ions
and not to
differences
of
in structure
adsorption, or in
the capillaries.
The
electrolyte
is,
medium
only, the
We
}f
Na"
CP
01'
Na*
Na*
or*
Initial
Final
^
and
254
If c2
is
""
we
(K
obtain
G\
'
C2
OS
The sodium chloride will thus distribute itself on either side of the
membrane in a ratio which is a function of the concentrations of
both
salts.
The
colloid
with an
H,0
so
JL
Initial
fl!
Na*
Na"
a;
H'
OH'
so
R'
Final
As
255
MEMBRANE POTENTIALS
If the ionisation constant of
xa_
The value
Kw
of
Kw
Kw
x,
/V\
100
(X), or
x~
or
= 0-01,
(a?)
c-i
is
e.g. for c
Kw -
then
(PI
or
water be
The
='05%
insertion of the
membrane
1911;
Garner,
ibid.
The magnitude
consideration of the
when
equilibrium is obtained.
If nSn mols, of Na" are transferred from one side to the other and
(1
ri)
The
electrical
if
and
solution, is
We
work done,
work done
(Fa
ca
ao
cc
= .X
say,
J
:>"
256
We
The
-,
hence X=c2
for a small c2
membrane
is
accordingly
cx ratio.
The magnitude
of this
membrane
potential
is
discerned from
and
ferrocyanide membrane.
It is evident that increasing the quantity of diffusible electrolyte
in the solution decreases the membrane potential, a phenomenon
observed in the case of Procter and Wilson (J.O.S. oix. 1317, 1916)
for tannin
III.
607;
ill.
691;
in. 827; iv. 33; IT. 97, 1921) (see p. 304) in the case of gelatine.
CHAPTER
VIII
AND EMULSIONS
In discussing the various methods by which dispersions such as
suspensions or emulsions are formed we noted that the stability of
such systems was not necessarily determined by the fact that they
a microscope, or
if
particles
by
may
first
noted
that
B.B.O.
17
MI
258
proved that
116, 1867), Dancer (Trans. Mane. Phil, Soc. 162, 1868), Jevons
(ibid. 78, 1870),
That the Brownian agitation was the result of molecular bombardment was suspected by Carbonelle in 1874, Deslaux (Jour. Hoy.
Microscopic JSoc. 1877) and byBodaszewski (Ghem. Centralblatt, XVII.
709, 1881) who first observed similar irregular zigzag motions in
the particles of smoke from burning paper and in the fog particles
of ammonium chloride, but this point of view was first stated in a
definite form
VII.
561, 1881;
CM.
CIX.
pended
visible molecules
were in experimental agreement with those that would be anticipated from an application of the kinetic theory making the basic
assumption that the suspended particle may be regarded for kinetic
purposes as a large molecule possessing a mean kinetic energy equal
mean
to the
which
it is
2.
The
In a
259
manner which
in the
The
Atomes).
Consider
Sa?
the
equilibrium
in
cylinder of suspension of
density p l in a suspension medium
of density p a of unit cross-section
vertical
and height
are
,o?
x there
volume and
If at a height
at a height x 4- Bos, n
in osmotic pressures
"
+ 8n
RT
RT.
..
cross-section is
p a )#.
RT
RT
or
where n and
pt)
are the
number
g (ID -ok),
172
260
may be
prepared.
The determination
of the
By
fall
determined.
moving through a
force
F is given by
where r
is
gamboge emulsions
'9
or
On
is
introduced.
acidulating a
volume can
readily be ascertained.
a sufficient number of
gamboge particles on
a glass coverslip the length of a
string of particles could be
(c)
By precipitating
DISTEIBUTIOlSr
261
measured and the number in the string counted, from which the
radius could be deduced.
As typical of the variations in particle radius determined by these
three methods the following experimental values of Perrin for
The
cited:
was determined by Perrin with the aid of a microscope and micrometer focussing arrangement. The height of the column under the
at various depths was
microscope was O'l mm. and the number
counted with the aid of the eye. The following results are typical
of such determinations.
8
Kadius of gamboge particle 2'6 x lO" cm.:
(i)
(ii)
B
Badius of gamboge particle 1'06 x 10~ cm.:
of
an atmospheric
262
criticism of Perrin's
another.
term
to
Writing
Dm
this in the
form
A8n = (GriBn^) 8$ we
C
n_
obtain
_j_.
B+lce A
Thus
so,
to
_ V(pi-p\g-
kez
not in fact contain charges all of one sign but is electrically neutral
as a whole, further, that a
charged colloidal particle with its attached
of Perrin's treatment
concentrated solutions.
Porter makes use of the Sackur-Porter osmotic law for concentrated solutions
nET
263
where
b corresponds to the b
proportional to
together with its water of solvation, thus
is
dn
If the osmotic pressure be a function of the hydrostatic pressure
of the solution p and of the numerical concentration we have the
mathematical identity
dir
dn
/C?TT\
/drr\
\dpj n
\dnjp dec dp
due
'
'
dp
xm.
Soc.
A,
Li.
dp
dir
_ _
a- is
the shrinkage.
s is
dp
--sp-
also
dp
doo
Also
g
*
cr
Hence
(1)
crcr
.__.
From
where
p2
dpo
=
JL
dp
where
and
(2)
d'TT
dn
g dn
doo
=-_.__
/2)
^ /
we obtain
,
>'
ua
u.
.'.
s~-a-
= c(p
u)=i nm
= Jen (1
-y-
thus
Pl
Or
l~
u),
"brif,
gNm(
/
/c-
'
where
wnere
(pi
-u)
on
ari ^
nas a P
m^
f inflection for
-^
as
to the value
The
by
Porter:
whilst values calculated with the aid of the above equation were,
3
=
putting 6 10-2 x lO^cm. k- 121,
,
3.
of translation.
F is
influence of a force
F
If there are n particles in a unit volume of liquid then
^
nF
will
in a solution will be
TT
,,,.,,
but for dilute solutions
,
IT
= RTn
TT-
nF
hence
or oir
= RTSn
^r~
RTS
D the
quantity moving across unit area under the influence of a concentration gradient
ax
will
be
D -7dx
diffusion will
265
or
We
obtain
where
is
a constant
= 0-815
free
path of a
or
_ a2
a;
=E
Tt
Qirrjrd}+^,
where
Then
is
large
number
term $x vanishes.
Replacing
d^v
u,
-5
by z we obtain
mdz
-^-j
2i(Lt
Now %m& is
the
mean
.,
RT
m^=-j,
The
RT
mdz
and
rim
z=~
When
RT
from which by integration we obtain
.
Tt
to
Independently of Einstein, Smoluchowski arrived at a similar solution
the problem of the mean displacement of a particle. He obtained
_32j5
4.
Tt
Not
siderations,
to
an unequal
(loc, cit.)
referred to a parameter
by the
following equation
FLUCTUATIONS IN CONCENTRATION
if
-^
or
267
RTt
x =
a torque of
moment / be
will
be
&>
Hence
RTt
Thus
r"
where r
5.
is
Fluctuations in concentration.
it will
contain at
any instant a
certain
number
of particles
U{
- n\
/ mean
g=
when n
is
large
and when n
is
v^
small
____
where k
is
number, and
is
equal to n
when n
number when n
is
is
a whole number.
not a whole
If Boyle's law does not hold exactly, the above equations suffer
the following transformations.
For
large values of n, S
= *//~2
V
~Q
7TW
TT
>
for
small values of n,
Sg
ET
ft
2<A'
6.
Experimental verification.
On
by the expression
were
obtained.
it will oscillate
about
EXPERIMENTAL VERIFICATION
269
given time interval will thus vary with the interval chosen; from
the relationship
_.
RT
the velocities
for different
stant values of
t,
j-
fT
and \l
i.e.
for con-
io
30 seconds
6'7
60
9-3
90
11-8
120
microns
13-95
whilst the square roots of the times are proportional to the numbers
and 13-4.
Svedberg obtained the following values for the displacement of
gold particles over varying time periods, confirming the anticipated
6-7, 9-46, 11-6
relationship
x = K*/t.
(6)
He
obtained
B
By observing the velocities of particles of radii 5 x 10~ and
2*13 x 10~5 cms. respectively for equal time intervals the square of
the velocities were found to vary inversely as the radii, thus with
found
whence
T
A
1:1-7 and
:1'6.
773
271
EXPERIMENTAL VERIFICATION
x in given
Svedberg's method of examination of the displacements
time intervals is liable to be somewhat inaccurate. Nevertheless, as the following data indicate, good agreement between
the experimental figures and the theoretical expectations was
obtained.
f)
is
c
approxi-
mately constant.
In Einstein's formula x
of J. 2
by experiment.
Svedberg has likewise examined
v.
GOLD SUSPENSION.
No. of particles per 384^3 =1'4.
-=0-418
-=0-408
k\
-=1-024
Po
Two
MBBOURY SUSPENSION.
No. of particles per 384/i 3 =0'5.
^=0-479
n
=
0-450
Electrolytic coagulation.
been noted that for substances which do not yield ions in
solution, such as charcoal, the charge is always produced by the
preferential adsorption of an ion from the solution, e.g. the
It has
ion.
it is
hydrogen
For substances which yield ions in solution, e.g. the hydroxides,
that
the
evident
charge can be accounted for by superficial
(Fe (OH)*).
* {Fe (OH),
or
In some cases
(x
. 1}
Fe (OH) 2 }+ + OH',
M HSiO
may be
s }'
+ H\
partial
results in the formation of a negatively
charged colloid
2
S3
H S|
HO
"
'
H O.S
8
ELECTEOLYTIC COAGULATION
273
is
movement
size
will
'
E.
8. 0.
18
274:
The normal
opposite sign will move away, thus the sides of the two particles in
close proximity to one another will exhibit opposite polarity.
On the assumption of a somewhat mobile double layer, electrostatic attraction
ment
may
effect,
between particles may occur, due to this displaceif the total net charge be not zero, i.e. coagulation
even
is
reached.
The data
of
electrokinetic potential,
volt,
is
for gold.
Zsigmondy
The primary
suspension
is
action of
primary
first
rises
maximum
ELECTROLYTIC COAGULATION
At
275
we may assume
this concentration
The
been mentioned, every collision is not effective in producing coaguand the rate of precipitation is greatly affected by the con-
lation
40
353025J?
20-
ra
,9
15-
T~
r-
20
40
'
60
80
100
120
Time in seconds
182
thus
will
v.
Smoluchowski
(PhysiJcal. Zeit.
Zeit.
the process.
If
be the
and
number
after a
P P P
time
the
sol
and
tertiary particles
P
= the diffusion coefficient and r
where /3 = 87rP Dr in which
the radius of the particle (2r being taken as the mean distance
apart of two particles when within one another's sphere of attrac-
tion).
Also
Pl
'
(1
The
half
life
+ /30
of the suspension,
i.e.
'
for
is
toto>
given by
277
ELECTROLYTIC COAGULATION
TT= I'D
'
0-75-
0-5
0-25-
number
sions.
The
=0-27 x 1010
~7
cm.
by Westgren and
Reitstbtter
Phys. Chem. XOIL 750, 1918) with gold suspensions somewhat coarser than those employed by Zsigmondy and by Kruyt
and von Arkel (Reo. d. Trav. Chim. d. Pays-Bas, xxxix. 656, 1920)
(Zeit.
for selenium.
number
is
We
v.
Smoluchowski
by an
(loc. cit.)
expression r
where
as
life
is
=
period T
the
number
of suc-
cessful collisions.
This expression for the slow coagulation has in fact been confirmed by Westgren (Arkf. Matem. Astron. och Fys. xm. 14,
1918)
for coarse gold
suspensions but the investigations by Paine (Kolloidchem. Beihefte, iv. 24, 1912) on copper
suspensions of Miyazawa
XXXIIL 1179, 1912); Ishizaka (Zeit. Phys. Ohem.
(J.CJ3. Tokio,
Gann (Kolloidchem.
1916) on aluminium hydroxide and Lottermoser
xv. 145, 1914) on tungstic acid and
Wiegner (Koll.
vm.
(Koll. Zeit.
65,
vm. 227, 1911) and Galecki (Zeit. Anorg. Ohem. LXXIV. 174,
1912) on gold have shown that the rate of coagulation over the
sensitive range of
electrolyte concentration is not a simple reaction
of the first order as
the above considerations but
postulated
Zeit.
by
behaves as
of the type
ELECTROLYTIC OOAG-TJLATION
279
The
KC1 per
litre)
may be
cited.
Steiner (Zeit.
we
mode
of
mechanism
nature of the
&
medium. After
where
is
the
s.l.c.
of the
will
&
be larger and
where
r' is
is
accompanied by a
potential.
8.
pension or emulsion
will
containing
electrolyte
electrolytes
containing a
common
cation necessary to
produce
281
is
e.g.
Na S0
2
a reaction
Colloid
*. Colloid
2H'S0 4"+2Na'+201'.
will
or precipitate will
to give the original suspension electric neutrality.
negative
arsenious
sulphide
by the
following table:
independent
of
any
i.e.
electrical neutrality,
definite
stoichioraetric
and
ratio
is
of
reactants.
measured
Gta,mi(Kottoidchem. Beihefte,vin. 127, 1916) has likewise
the adsorption of various precipitating anions by colloidal aluminium
hydroxide and found that although the liminal concentrations
is
necessary for precipitation are very different yet precipitation
caused on adsorption of equivalent concentrations as indicated by
the following
figures.
amounts adsorbed
ratio
3m :%m:m,
is
1}
2>
necessary
Thus on the
some 50-60
of the
283
is
adsorbed.
may promote
gum
copper ions will be adsorbed and the solution will be left feebly
a phenomenon noted by Spring (Rec. d. Trav. Chim. d. Paysacid.
Bas, iv. 215, 1900).
and Matula (Kolloid. Zeit. XXL 49, 1917) in the case of sulphate
ions displacing the less adsorbable chloride ions on ferric oxide.
284 CONDITIONS
03?
AND EMULSIONS
STABILITY IN SUSPENSIONS
9. Lyotropic series.
In the adsorption of electrolytes by charcoal
it
this,
The
series.
the lyotropic
determining the
were primarily the
factors
series,
Prakt. Chem. xxv. 431, 1883; xxvn. 370, 1883; xxxn. 390, 1884),
Linder and Picton (J.G.8. LXI. LXVII. LXXI.), Freundlich (Zeit.
Phys. Ohem. LXXIIL LXXIX. LXXX. LXXXIII. LXXXV. LXXXVL),
1899),
10.
The
effect of valency
much more
coagulative concentrations 1
mately 30.
CC
i.e.
Schulze
104
63
49
810
863
810
coagulation
result of a series of independent operations
primarily dependent
on an adsorptive process due to diffusion; the manner in which the
285
suspension will not produce any effect although the final dilutions
may be
identical.
Again,
it is
extremely
Small
precipitation occurs
The
to
addition of small quantities of electrolyte and allowing the suspension to stand overnight. The degree of coarseness of the control is
immaterial, provided precipitation does not take place. The concentration time curves of various electrolytes can readily be compared
in this manner.
Burton and Bishop (Jour. Phys. Chem. June, 1922) and Kruyt
(Koll Zeit. xxii. 81, 1918 ; Rec, d. Trav. Chim. d. Pays-Bas, xxxix.
656,1920; XL. 249,1921)have found that theratio electrolyte colloid
is not the only governing factor in the production of coagulation ;
:
may be
it
The valency
glance at
adsorption of various ions on the surface of charcoal.
the table indicates that in a definite series, such as the alkali metals
or the halides (Pappada, Koll. Zeit. IV. 56, 1909) the electropositive
nature of the element is an important factor.
11. Ionic mobility and adsorbing power.
There exists also a relationship between the adsorbing power and
the ionic mobility, a point developed by Mukherjee (Trans. Farad.
Soc. XVI. 103, 1921) embracing the following assumptions.
ion adsorbed on the surface of a suspension will draw near to
An
it
have a valency n-^ and w2 is the valency of the opposite charged ions
in contact with the solid and
in the liquid of dielectric constant
separated from the former by a distance OB, we obtain
where
e is
_JE
where
e kff,
probability of an ion remaining fixed is given by 1
k is the Boltzmann gas constant = 1*372 x 10~16 erg/degree,
T the
absolute temperature.
The
There
will thus
be at any definite
its
If
number
maximum
Nm of valency n
is
IONIC MOBILITY
If
we
Denote by
maximum
287
adsorbed when
evaporation per
evidently
K&Nn e-&
and Nm
the number of ions adsorbed
(iii)
where
JTj is
a constant
6i
per unit
(iv)
The
the
to
ions.
RT
-jr-,
The former
is
field is proportional to
K (1 ~
z
for
#1) n>
where
is
may
write
a constant. In com-
negligible.
where
and (vi)
equations (iv)
is
given by
we
^2
N,n
Since
z,
k and
^O fl-^0 J(l~0
1
(a)
where
1 ),
-^n Ue
kT
by adsorption we have
electrical adsorbability of an ion of opposite
charge
dependent on the product n U or its valency and its mobility.
Thus for a negative surface Mukherjee gives the following order
Evidently the
is
of adsorption
Th > Al > Ba > Sr > Ca > Mg > H > Cs > Rb > K > Na > Li
forming a lyotropic
series,
be
Br>N0 >SCN>
C0 > S0 > OH > I > 01 > OOOH > F.
3
Elissafoff (Zeit.
or
(a)
and
(6)
+ 1 -1 = 0.
K C02*+ ezK
^=1
and putting nz =
1 for the
adsorption of
+0 -l = 0.
a
IONIC MOBILITY
The
show
289
Elissafoff.
Electrolyte NaCI, K
Electrolyte
BaCl 2
= -0076.
= -0163.
The two factors of valency and ionic mobility are however not
the only ones to be considered in the phenomenon of adsorption,
since the lyotropic series developed from these considerations alone
should remain invariable in all cases of coagulation. This is far from
being the case the order varying in a marked manner with the nature
of the suspension. In addition, no adequate explanation of the
teristic
following
may be
cited
19
Colloid
Sulphur
...
Platinum
...
Silver
...
Mastic
...
Reference
Ba,Sr>Ca>Al>Mg,Cs>Eb>K>Ni
Cd,
Al,
Zn>Na>NH 4 >Li>H
Jtori.xi.il.
...
H>Ba>Mg>Li,
Na,
1,263, 1912
Ca>H>Cs
>Eb>K>Na>Li
ferric oxide
Kolloide
Al>Ba>Sr>Ca>H>Cs>Rb
>K>Na>Li
Al>Hg2 >H>Ba, Ca>Mg>Ag
Hydrous
Der
Pb>Ba>Ag>K, Fa
Albumen
Ode*n,
1911
Th>0u, Zn>Ca>Mg>Li>K
218
when
12.
There
exists
suspension and
an equilibrium at the
its solution
between a charged
dynamic interchange of
interface
consisting of a
= a Cc
evident that
if
n and
m are not
291
somewhat
less
If one adds to a
On
The diminution
by the addition
polyvalent ions,
of electrolytes
192
292 CONDITIONS OF
STABILITY IN SUSPENSIONS
AND EMULSIONS
and
(Biochem.
such
as
1919) to
substances
similar precipitation
by ethyl
the
precipitation by electrolytes
is
affected.
The
by
(Amer. Jour, of
Sci.
who observed
that
the
the adsorption
powdered
from
For a
ferric
293
(Biockem. Zeit. LXXXI. 87, 1917) have noted similar effects on this
of many other capillary active nonsuspension on the addition
and thymol.
electrolytes such as the urethanes, camphor
is
(p.
230)
^KRG
4s7rt)
obtained for a
ferric
hydroxide
sol
The
Although an
double layer
may be
again
and tetravalent cations as noted
not explicable on this hypothesis.
:
is
di-
If the non-electrolytes are in fact selectively adsorbed it necessarily follows that the ions are displaced from the colloid surface;
is
cut
down
by the
salt
295
in sensitiveness will
down an increase
were determined.
14.
As has
m, 1916) and
of these
"
"
colloids will
protective
in small quantities.
when employed
the
U numbers
According to Gann
(loc.
cit.)
may be
cited:
substances
is
Thus
"gelatine H'
COO-
-000
COOH
-NH +
-NH +
"*~
NH OH
Alkaline
Acid
4-7
contact with gold more acid than H 4 7 the positive gelatine ions
will be adsorbed, neutralise the charge and precipitate the gold,
On
result
The
isoelectric point of
an amphoteric electrolyte
297
is
likewise
Kermack and Wright (Bwchem, J". xvii. 635, 1923) have shown
that gelatine at a H of 4-7 exerts but little protective power on a
negative gum benzoin sol, acid gelatine precipitates the colloid in
on gold, but
its
was not
it.
Similar
effect of casein
established.
CHAPTER IX
GELS
1.
Introduction.
particles of a true gel which possesses but little mechanical resistance and are readily broken up by mechanical or thermal agitation.
The
structure of gels.
theories
proposed
(2)
High number
of particles
per unit volume.
STRTJCTUBE OF GELS
(3)
Small difference in
299
and
liquid.
(4)
1 and 4.
was considered that the gels were homogeneous
or one-phase systems. Evidence in
support of this hypothesis has
been advanced by Pauli (Kolloidchem, d. Eiweisskorper), Katz
(5)
At one time
it
(Roll. Zeit. ix. 1, 1917), Procter (J.C.8. ov. 303, 1914) and Loeb
(Jour. gen. Physiol in. 827, 1921; IV. 73, 97, 351, 1921-1922).
The investigations of Procter and Loeb were confined chiefly to
the chemical behaviour of gelatine (see p. 304) and to the phenothis gel, a subject investigated systematically
menon of swellings in
many
and
fatigue,
basis.
others,
The ultimate
mechanism of
gela-
elastic
phases
is for
many
still
The
gels that
all
of
them
such as
diffusivity,
refractive index
and
electrical
800. OLI. 183, 1861; Arrhenius, Ofvers. Stoclch. AJcad. vi. 121,
1885;
300
GELS
that there
is
sol to
the gel.
The weight of
ex-
perimental evidence
3.
Two-phase
is
assume a polyhedral
or
honeycomb
structure.
emulsion
will
rigidity and
acquire a considerable
may
gelatine
example, consisted of a solution of gelatine in water in a
solution of water in gelatine, the two liquids
possessing a definite
interfacial surface tension. If the two liquids are
perfectly homogel, for
geneous these gels would differ from the emulsion gels only in that
the latter possess a third substance which
to the interface
goes
diminishing the interfacial surface tension, whilst in these no such
extraneous stabilising agent is present, In a
gelatine gel for
example, on increasing the gelatine concentration in the water, the
water in gelatine phase grows at the expense of the
gelatine in
SOLID-LIQUID
301
agent,
by Bancroft (Applied
with a
may
4.
still
The
solid-liquid,
Hardy
within
it a
disperse phase of gelatine in water. Support to this
hypothesis has been given by Freundlich (Kapillarchemie), by
Fischer (Soaps and Proteins, 1921) and by Anderson (Zeit. Phys,
difficulty in these
assumptions
in water
is
the former
The
Yet on
mobile.
phase.
is
is
solid
302
if
furthermore somewhat
difficult to
relatively non-compressible
5.
Gel
and
due
embedded in a
medium.
Fibrils.
These two
which the mobile liquid phase penetrates. The fibrils in the case
of gelatine would, according to Bancroft, consist of a viscous water
in gelatine solution, and, according to Hardy, a solid solution of
water in gelatine.
The
fibrillar
Biitschli
who
at first
been noted
(Phys.
Similar
fibrils
have
medium
GEL FIBRILS
303
and the
medium and
That the
Flade
(Z&it.
found to be amorphous.
The
The
by
spiral
movements
as a result of Brownian
bombardment. It appears
formed is the governing
'
304
GELS
The
6.
structure of gelatine.
line.
of
O'l//.
He obtained in a 0'4
diameter.
The ultramiscroscopic
fibril
of gelatine thus
Scherrer's examination of
multiple twinning of minute crystals.
has revealed the fact that the crystal intergold sols, however,
ference lines are well developed when the particle contains as few
as eighty molecules arranged in the cube lattice, a spherite of
O'l//,
if
x. 32, 230,
phenanthrene
that
when
bromshown
the solid crystalline states at the melting point where the thermal
agitation of the molecules just balances the cohesive power, the
crystals frequently assuming the spherical form under the influence
of surface tension. These swarms are definitely composed of minute
crystals since they exhibit the property of double refraction and
EIGID
der Lingen, Verh.
nomenon
AND MOIST
GELS
305
Physih Oes. XV. 913, 1913), a pheexplicable on the hypothesis of the existence of orientated
d. Deutsch.
7.
But
little is
known
as to the influences
which
affect
the growth
and slender
tinisation.,
by a more
S
linkage by a
destroys the power of gela-
H H
group such as
or
CH CO
8
like-
s. o.
20
306
G-ELS
Farad. Soc, xrv. 10, 1921) for the composition of the sols and gels
of the inorganic colloidal hydroxides, e.g. zirconia produced by
the hydrolysis of zirconium oxychloride. By electropotentiometric
measurements of the hydrogen and chlorine ion concentrations of
bF8
__
complex zirconic acid [Zr(OH)4 ZrOCl sCl 2]H 2
whilst in a similar manner he was able to identify the following in
well
as
the
as
sols,
[>Fe(OH) 3 yFe"],
[A1(OH).C1 2 .A1(OH) 2 01],
r-
8.
The imbition
of liquids
by
-1 _L
J_
LSnOySn0 8 JK
gels.
The
assumed on the
fibrillar
rendered
The
visible
fibrils are
The
nature of the
and
fibrils,
dependent on the
the more moist
and more rigid the
fibrils
KIOTO GELS
307
9.
Rigid gels.
On
drying a rigid gel the aqueous phase enmeshed in the feltbegins to evaporate, the gel is sufficiently rigid as not
like fibrils
e.g.
Such a
with
filled
air.
tending the gel. Imbition for the elastic gels is actually a case
of adherence to the fibrils and is only caused by liquids possessing
this property, e.g. water in the case of gelatine and benzene and
similar nonpolar solvents in the case of rubber.
10,
gel.
may be
molybdenum and tungsten. Its preparation may readily be accomplished by the interaction of water glass (cct. 2 / SiO a ) and
hydrochloric acid, although to obviate the subsequent removal of
dialysable salts, hydrolysis of silicon tetrachloride or organo-silicon
compounds has been suggested. The initial solution of ortho-silicic
acid formed in this
manner
is
unstable and
is
rapidly converted
sol of
a rigid gel
reversible
and
is
affected
by a number of
factors,
on elevation
sol-gel transformation is very considerably increased
of the temperature (Fleming, Zeit, Pliys. Qhem. XLI. 493, 1912) and
also
by the presence
this relatively
weak
of various electrolytes.
acid
The
sol particles of
202
308
Decreasing H'
PH?
Isoelectric
Increasing H'
point
The unstable
sol
evidently possess a
maximum
stability at the
xxxm.
391, 1900;
xxxm.
11, 272,
salts
(Hardy,
are
required to
Zeit. Pliys.
Ghem.
Jtal.
c.c.
for the
time in
hours necessary to
effect
litre.
The anion as is
to
309
random
at
Kruyt
been proposed
for the
No
Ohem.
XXL
I.
Vetenskap. Nobelinst.
251, 1913)
where y x
is
medium and
centration.
application to
T)
formula proposed
is
= l~
tyfa
where 1000
is
Zeit.
7700
xxvn.
1,
1920)
<>
gated by Graham,
Adsorption, 1910),
310
NH
(/.
second point
"
of inversion,
when
sunk to
silica.
shown in the
0-7
12345
10 20
80
40
160
BA
BA
is
practically isopneumatic,
is
reached
DA
i.e.
is
311
increases the
vapour
Saturated
V.P. of
water
Yapompressure
Water content
of gel
of desiccation and the vapour pressure rises rapidly along the line
The vapour pressure curves cannot be interpreted on the
B'G'.
Bemmelen
fibrillar
the region
whilst the
mesh
impossible.
Along
fibrils is
BA
is
at
axis
relatively uniform.
is
fibril-
the
another takes place since the curve AB is not reversible
on the rewetting indicating a smaller mesh.
higher vapour pressure
interfibrillar
Nevertheless this contraction is but small and the
fill with air and the gel in conseto internal scattering of light at the
the water
spaces left vacant by
due
quence becomes opaque
all
the water
is
removed
when
312
The removal of this water is a perfectly reversibleprocess. After the adsorbed water is removed further dehydration
of the gel results in disruption of the fibrils, which presumably
consist of Si0 2 2
and the gel structure is definitely destroyed.
The hypothesis that the curve
represents a nearly, but not
removed.
AB
Bachmann found
a
by
AS
below
may
trarily
to a
arbi-
capillaries are
at
the
Phys. Ghem.
LXXXVIII. 191, 1914) has calculated these radii for a silica
gel with
the above assumptions from determinations of the
vapour pressure
with the aid of the equation
313
where p a and /o? are the densities of the vapour and liquid respectively, P and Pl the vapour pressures above a plane surface of the
liquid and the gel.
11.
COO-
OH'+P
HoOH
Some
roooH
rooo-i
I
+H':
LNH
LNHa+J
and
this
most readily at
and
as
lca and 7fy are the acid and basic dissociation constants of the
amphoteric protein molecule. If the total concentration be n then
where
ao
= n - CA - CK>
314
f
fVT
V*.
"
tl>
ffl
/W
_ ^j
__
_=
^
or
.r^
^H
^OH
l+^L +
\^TT
*
3
^*-^O!R
/vt
-=
and
In water as
p-+
p
^H ^OH
dispersion medium C H C O H
J.
-f-
^w or
is
is
evidently a
maximum when
or
&;
OH-
minimum
when
d%
^C H
hence
-^
Jc a
CH 2
= 7^-.
,kb_Q
kw
or
which
is
e.g.
hydrochloric acid to
an
The amount
e.g.
of
maximum
AHCi
315
In the
following curve are plotted the values obtained by
Manabe and Matula (Biochem. Zeit. LII. 369,
1913) for ox serum
albumin.
0-01
O05
0*06
n-EOl added
A similar point
will likewise
maximum
of
isoelectric
amphoteric protein.
Both
albumin and
isoelectric
side, these
maxima
maximum
dis-
+
sociation of the
hydrated
ions.
two
salts
AHJC1
and
NajA
into
the heavily
or
undissociated
salts, e.g.
maximum
salts
The
dissociation of the
316
Increase
in
OH
point
Maximum
dissociation of salt
Maximum
NaA5:Na'+A'
12.
The
dissociation of salt
AHCl^AH'
properties of gelatine.
chiefly glutin and salts as impurities, the protein in the gelatine undergoes gradual hydrolysis
to peptones and proteoses. Many of the physical characteristics
e.g.
is
apparently almost
PROPERTIES OF GELATINE
The
317
influence of the
(loc.
cit.)
and Bogue
It will be noticed that all the properties of the gel with the
exception of the turbidity and foam have minimum values at or
near the isoelectric point
H = 4*7, whilst these two attain their
maximum values at this point. Evidently as the data for the
particles
Biochem.
alcohol
form large
xxxm.
Zeit.
number
it is
318
13.
GELS
The
swelling of gelatine.
for the
heat of wetting:
rises the
swelling pressure and the
heat evolution decrease and at the same time the consistency of the
gel becomes less. A large portion of the water is readily removed
application of pressure, whilst dilute gels will on standing undergo contraction accompanied by a coarsening of the fibrils and
an exudation of the contained and adsorbed liquid, a phenomenon
frequently termed syneresis.
by
As
(Blitschli,
p. 22, 1896;
Uber den
Bachmann,
Zeit.
Ban
Anorg. Chem,
c. 1,
1907).
Gelatine,
however,
is
fluid.
Zsigmondy (Kolloid-
SWELLING OP GELATINE
319
bases as well as
salts
01'
>H
is
> SO/'.
The swelling
permeable
ions.
We
of
PH = 4'7,
HOI
~ZZ- gelatine
HOI ^3H
Interfibrillar liquid
GH
IF
as
01'
or
resulting
isoelectric gelatine.
Fibril
fibrils
fibrillar
be
>
2
i
On
and equilibrium
320
and bases on
The
studies of Pauli
and
and
HA^H'
however, have
revealed the fact that isohydric solutions of different acids do not
effect equal combination with the isoelectric protein; relatively
(loc. cit.)
his co-workers,
more acetic acid for example being combined than hydrochloric acid
in isohydric solutions. Again, both the actual position of these
maxima as well as the magnitudes of the viscosities observed vary
much
weak
is
probable
that the degree of solvation of the protein molecules and of the
protein salts must not be regarded as constant but that they vary
both with the nature of the salt and in the presence of neutral salts
potentially
one reactive
2 group in its molecule.
other proteins appear to be relatively more
complicated, thus
on the progressive addition of strong acids to
proteins such as
The
NH
SWELLING- OF GELATINE
321
These results are repeated in the case of alkalis, thus Pauli noted
that the caseinate ion obtained on treating isoelectric casein with
dilute soda
OJ -> Na
(C ) -* Na C^
Whether the complex salt [Na8 C"' casein] must be regarded as a
complex salt or an ionic micelle of the type investigated by McBain
Isoelectric casein
in his investigations on the soap solutions is not yet clearly established, whilst information on the acid and basic dissociation
14.
lacking.
The
The soaps
Soaps.
consist of the
tion of the boiling point of aqueous solutions of the salt does not
indefinitely with increasing concentration but rises to a
proceed
maximum and
boiling point.
homologous
K. S. 0.
series
which
may be
little
is
the
elevation of the
first
salt of
the
322
G-ELS
(Krafft and Stern, Ber. xxvii. 1747, 1894) or the lowering of the
vapour pressure (A. Smits, Zeit. Phys. Ghem. xxxn. 1584, 1899;
xxxix. 385, 1902; XLV. 608, 1903), whilst the osmotic pressure
methods are vitiated by the effects produced due to the appearance
of a
133,
is
property of a definite minimum. The values for the specific conductivity of potassium palmitate are (McBain and Martin, J,C,S,
cv. 957, 1914) plotted in the following curve.
98
'02-04-06 -08
The
-1
is not due to
pronounced hydrodetermination of the hydroxylion concentration
(loo. cit.)
both by means
SOAPS
triaeetonamine, the extent of hydrolysis
the following data relatively small.
323
is
as observable
from
POTASSIUM PALMITAEE.
NaHP
Taylor
HP.2NaP.
In order to account for the relatively high conductivity but low
osmotic pressures and the low hydroxylion concentration of soap
solutions, Bancroft assumed that the hydroxylions were adsorbed
colloidal potassium palmitate, this assumption, was,
however, proved to be incorrect by McBain who showed that on
the addition of caustic soda no adsorption of the added hydroxylions
took place. As an alternative explanation McBain suggested that
by the neutral
the conductivity was due to the part of the soap which existed in
fche colloidal state, a colloidal ion. The colloidal soap ion differs
from the ordinary charged colloidal particle in that not only is its
mobility great but the equivalent conductivity is high, the conductance, in fact, boing much greater than that of the simple ion of
the fatty acid from which it is derived and comparable to that of
a potassium ion. To account for the high mobility and equivalent
conductivity of the colloidal ion in an electric field one must
a
for the number
postulate that it possesses relatively large charge
of molecules forming the aggregate and one must consider that the
colloidal ion or ionic micelle (Nageli, Pflanzenphysiologische Untersucliunyan, Zurich, 1858)
is
212
324
will,
colloid.
composition
The
value of
least ten.
The
b the
fl
number
only
in elasticity and
rigidity, whilst the electrical conductivity, refractive
index, concentration of metallic ion and lowering of the
vapour
pressure are all identical, results to be anticipated on the fibrillar
theory. The gel as we have seen is fibrillar in nature and the con-
SOAPS
325
fibrils,
experiments of Langmuir (see p. 75) have indicated. The detergent action and emulsifying power of soaps is likewise due to this
cause. If we insert in an organic material containing hydrocarbon
chains sufficiently long so as to render their mutual lateral
adherence great enough to overcome in part the disintegration
due to the thermal agitation, polar groups of the type
COOX,
SO
X,
N(CH )J,
8
which are
sufficiently strong to
immerse the
somewhat
326
15.
The
colloidal dyes.
comparable to the
salts of
the
the polarity of the polar group appears likewise to affect the dispersion. Thus phenolphthalein undergoes practically complete dispersion in alkaline solution, i.e. when present as its sodium salt, but
is colloidal when present as the free acid.
In the soaps we note
also
COLLOIDAL DYES
327
is
the basis
number
of volumetric processes of estimation; thus the evaluation of tannin solutions may be accomplished with
night blue or
of eosine with night blue.
The mechanism of dyeing (see First Report on Colloid Chemistry,
B. Ass, 1917, for the complete re'sume' of the various theories) has
irreconcil-
K. K. Gewerbe-Museums
in Wien,
saponin
may be formed
at air-liquid
and
liquid-liquid interfaces it
improbable that any dyeing can result from such processes alone
as the fastness or the formation of solid films is due to subsequent
is
charged in water
and readily adsorb basic colours, on increasing the negative surface
328
G-ELS
wool and silk are like the proteins amphoteric in their nature and
can combine with both acid and basic dyes. Since the charge on
is as we have seen localised at certain points along the
where the polar portions of the molecule are present the
gel fibrils
fibril
still
As has
is
chiefly
closer union,
relatively unstable and dyeing will not be fast.
however, results in the neutralisation of the electric charge on the
dye ion and on the fibre and a process akin to coagulation takes
by adherence
to the
non-
Penetration in membranes.
Charged
PENETRATION IN MEMBRANES
329
The
different speeds.
fibrils
as
we have nqted
are hydrated in
whether
Such changes
fibrils.
may conveniently be
followed by observation
partitions.
Thus we
and by
opaque
^
find in
CO/
/
00<
>
j
,
NH
00<N
2
,
C0<^
>
not at
OH >
OHO.
more
all.
materials.
changed
slight changes in the hydrogen ion conneighbourhood of the isoelectric point of the
membrane permeability by
centration in the
membrane.
186
INDEX OF NAMES
Ablett
6, 7
Aokerberg 193
Adam
Adams
163
Allmand 255
Alty 233
Amagat 49
Antonow
Brown
Adolf 321
Allen 166
Anderson
Bradford 304
Bravais 124
7,
Arkel 277
Cain 160
Armstrong 164
Cameron 110
Cantor 15, 16
Cantori 258
Carbonelle 258
Carver 7, 140, 144
Gary 65, 84, 86
Cederberg 56
Chappuis 133, 171
Aston 27
Ayrton 234
Bachmann
Bakr 183
Baly 27
Bancroft 116, 178, 187, 199, 289, 301,
802, 323
Barker 34, 36
Barratt 302
Bartell 187
Bashforth 9
Baudouin 225
Baur 119, 249
Becker 73
Beebe 128
Berkeley 52
Bethe 225
Beutner 235, 243, 245, 246, 247, 249
Bhatnagar 118
Bichat 234
Billiter 292
Biltz 279
Bircumshaw
39
Bishop 285
Bjerrum 242, 243
Blake 292
Blondlot 234
Dancer 258
Daniells 310
Davis 180
Davy 202
De
Broglie 73
Debye 172
Dennet 196
DerMeulen 111
Deslaux 258
Devaux 66,
Dewar 91
Bodaszewski 258
Bodenstein 140, 142, 164
Donnan
Dorsey 17
Draper 114
Dreyer 198
Drucker 193
Duclaux 283
Boltzmann
Duhem
Boss 304
Dundon
Bowen
Dunn
318
31
166, 167
129, 130
INDEX OP NAMES
Bushman
193
Edland 161
Edser 23
Einstein 258, 265, 266, 269, 271, 309
Elissafoff 225, 288, 289
Ellis 229
Eotvos
25, 28, 29
Ettiseh 252
Euler 184, 208
Evans 140
154, 273,
Harkins
Exner
331
Guillaume 225
Gustafson 180, 182
Gusfcaver 142
Gyemant 225, 226
Gyorgy 292
Harned 56
Fajans 184
Faraday 139, 203, 268
Ferguson 11, 12, 25
Feustel 15
Fick 192
Fischer 193, 301
Fitzgerald 171
Flade 303
Fleming 307
Foder 184
Forch 16
179, 180, 197, 199, 222, 227, 252, 282,
283,, 284, 290, 293, 301, 309, 327
Friedrich 73
Friend 196, 197
214, 234, 236
Galecki 278
Garni 278, 279, 282, 295
Hofmann 105
Gans 237
Garner 255
Gaudeohon
Gauss 223
Helmholtz
Hempel 132
Frumkin
Harris 255
Hartleben 185
Hartley 52, 130
Hartridge 104, 249
Hatschek 309
Hedestrom 92
Hedges 262
171, 172
Hogness 59
Holmes 110
Homfray 132, 136
Geddes 133
George 140
Horovitz 252
Howard 176
Georgievics 327
Gernez 196
Howell 303
Gibbs
Goldschmidt 193
Gore 180
Gortner 303
Gouy
Grabowski 54
Graham
Green 256
Griffin 111
Grindley 73
Grosehuff 307
Griinmach 17
Gryns 328
Hughes 252
Hulett 37, 166, 167, 173, 174, 176
Husson 143, 164
Ihmori 140
Iredale 14, 15, 57, 58, 59, 60, 144
Ishizaka 278
Jager 16
Jahnke 73
Jevons 258
Johnson 242
Jones 166
Judd 196
Junck 171
Kammerlmgh Ormes
Katayama 26
Katz 299
8,
29
INDEX OF NAMES
332
Merriam 191
Meyer 59, 107,
Kendall 309
Kenrick 234, 236
Kermack 297
King
Miller 187
Kirchhoff 267
Knapp 168
Knecht 328
Knudsen
Koehler 172
Kohlransch 234
Konovalow 39
Krafft 322
Kriiger 213
277, 285, 292, 294, 309
Kiister 128
Kruyt
Lamb
78,
80
Langevin 265
Langrmiir 5, 7,
Laplace 23
Larson 243
Le Blanc 195
Ober 281
Leeuwenhoek 257
Lehmann
292, 304
Lenard
18, 232
Lewis G. N. 39, 41, 101, 208
Liebmann 101
Linder 229, 281, 284
Linebarger 16
Lippmann
Lohnstein 13
Lorenz 101, 200, 203
Lottermoser 278
Lomy
McBain
Macleod 30
168, 230
Manabe 315
197, 198
Marcelin 66, 71
Martin 322
Matthews 172
Matula 283, 315
Masted 163
Maxwell
Palmaer 214
Palmer 129, 154
Paneth 173, 185
Pann 54
Pappada
173
Manjuin 304
Marc 195, 196,
Naeken 191
Nageli 323
Nernst 105, 172, 191, 209, 213, 215, 216,
217, 218, 240, 243, 327
Neumann 63
Nicoll 160
Niggli 125
Norris 193
Norrisb. 161, 162
Noyes 191, 205
Nugent 110, 121
Parks 172
Parsons 119
Partington 166
Paschen 214
Patrick 36, 106, 208, 309
Pauli 283, 299, 305, 320, 321
Pawlow 128
Perrin 67, 91, 225, 228, 258, 259, 261,
262, 269
Perrot 13
Perry 234
Peters 76, 104, 249
Pettijohn 140
Pickardt 197
Pickering 109
Picton 229, 281, 284
Piper 73
INDEX OF NAMES
Plateau 2
Pockels 66
Porret 222
Porter 223, 229, 262, 263
Potts 116
PouiUet 171
Po.wis 229, 274
Sugden
Poynting 5
Procter 256, 299, 316, 319
Quincke
6, 12,
Tammann
Tate 13
25, 26, 27,
258
120
Eandall 39, 41
Eaoult 34, 53
Eayleigh 9, 13, 14, 16, 66,
218
Beinders 105, 106, 170, 198
Eeitstotter 277
Beuss 222
Beyohler 325
Beynolds 96
Eichards 7, 11
Eichardson 133, 143
Bona
333
Stern 322
Stewart 161
Stoeker 18, 54
Stockle 59
Stokes 260, 264
Strutt 157
Stflbel 303
Saam 193
Saokur 262
Saeck 202
Salmon 322
Scherrer 206, 303, 304
Schmidt 18, 59, 108, 135, 136, 180
Schofield 209, 215, 218
Schonfeld 184
Sohrodinger 15
Schulze 284
Seddig 270
Sentis 54
Shearer 73, 75, 91
Shields 25, 26, 27
Siedentopf 272
Smyth 237
Spangenherg 198
Speigel 321
Spring 191, 283
Stefan 23
Steiner 279
Valeton 125
Vallance 196
Valson 54
Volkmann 11, 12
Von Halban 160
Walden
Walton 196
Warburg 208,
215, 218
Wegelin 200
Weiser 295
Wenk
196
Wenzel 191
Westgren 277, 278
Whatmough 16,
Whetham 284
Whitney
38, 99
Whittaker 19
Wiedemann
Wiegner 278
Wiener 258
223, 225
INDEX OF NAMES
334
Wilberforce 218
Wrensky 39
Wright 297
183
Young 143
Wimpler 22
Zadwiski 40
Witt 327
see Pauli
Wolfgang-Pauli 272,
Worley 28
Worthington 13
INDEX OF SUBJECTS
Collapse of films 78
Colligative properties 38
Colour of films 129
electrolytes 213
Activation 130, 159
Abnormal
Activity 34, 38
Addition agents 196
Allotropes 128
Definitions 3
Amicrons 278
Amphoteric colloids
296, 313
Differential tension 69
Diffusion potential 240; rates 129
Dilution factor 285
Angle of contact 5, 1
Annealing 128
Antonow's rule 96
Arcs 203
Arsenious sulphide 290, 292
Association 18, 27, 30
Atomic volumes 126
Autocatalytic coagulation 279
Barium sulphate
20S, 206
Benzopurpurin 303
Brownian movement 267
rotation 266
translation 264
pressure 15, 16
Electric
esis
9, 10, 11
Casein 321
222
229,
Catalysts 159
Cleaning apparatus 10
Coagulation 273, 284
Coalescence 274
Cohesion 103
202;
effect of
238
INDEX OP SUBJECTS
335
Mirrors 130
Mixtures, spreading of 86
Mobility of double layer 274
Moist gels 305
Molar free surface energy 25
Molecular size 71
Moments of molecules 219, 236, 237
Multimolecular films 139, 150
Myristio acid 81
Foam 317
Fogs 258
Free energy
Negative adsorption 54
Nernst ionic transfer 217
triangle 63
Nonylic acid 35
Nucleus formation 194
Neumann
3,
Friction 93, 94
Gas adsorption 57
Gas laws, application
of 46-50
Gelatine 121, 122, 304, 313
Gelatinisation 316
Gel strength 317 ; structure 300 cont.
Gibbs' equation 31-36
Glass electrode 251
Globulin 313
Gold numbers 201; sols 295, 304
Greases 109, 302
Parachor 30
Paraffin water interface 36
'
Hardy's rule 284
Heat of adhesion 102; of adsorption
145, 149, 171
Hydrolysis 187, 283
Imbition 306
Inhibition 121
Parallel plates 12
Partition of ions 119
Pentaerythritol palmitate 78
Permeability of films 91, 92, 93
Phase boundary 32 ; discontinuity 130
inversion 117; rule 88
numbers
295
Ionic mobility, influence of 286
Iso-acids, area of 50
Isoelectric point 122, 273, 282, 314
as emulsifier 116
Latent heat of evaporation 23; surfaces
Lampblack
3; transformation 86
Lateral cohesion 51
Membrane formation
Poisons 163
Polar groups 20, 21
Polymerisation 28
Porous materials 132
Potential differences 208
Precipitation 273 by colloids 295
Pressures, equivalent 52
Primary films 58, 59, 60; particles 276,
;
277
Promoters 131
Properties in adsorbed layers 155
Protection 130, 200
Proteins 313
Badius of curvature 9, 10
Bate of crystallisation 195 ; evaporation
21
Besidual valencies 21, 22
Bigid gels 305, 307
Bipples 16
Salts, influence of 23
Saponine 35
Saturation of surfaces 135
interface 36
Metals, tension of 24
Metathesis 119
Micelles 29, 34, 323
336
INDEX OF SUBJECTS
Symmetry, molecular 29
and
Syneresis 318
Tate's law 13
98
Solution at plane surfaces 191
Solutions, tension of Oh. n
Spherical molecules 32
Spherites 305
Spreading 61, 63, 82, 90
Spreading coefficients 64, 65, 66
Stability of suspensions 257
Static tensions 4-16
Structure of surfaces 20
Substitution, influence of 21, 22
Superficial ionisation 272
Supersatu ration 205
Surface compression 174 ; energy 2
tension 3, 4, 166
pressures 52
valency 126
Suspensions 198
;
3,
19
Ultramicrons 306
Unimolecular films 45
Vegetable fibres 303
Vibrating jets 16, 17
Viscosity 309, 317
Volatile substances, spreading of 65
Volatilisation, surface 68
;
;
Thickness of layers 55
Total energy of surfaces
Trough apparatus 68
Turbidity 317
Swarms 304
CAMBBIDGE
24, 28
Thermodynamic concentration 34
PBINTED BY W. LEWIS,
M.A.,