RHEOLOGICAL PROPERTIES O F BITUMENS 149

REFERENCES
(1)
. . LANGMUIR. I.: J. Am. Chem. SOC.39. 1848 (1917).
(2) NELLENSTEYN, F. J. : Bereiding en constitutie'van asphalt (Manufacture and
constitution of asphaltic bitumen); Dissertatie Technische Hoogeschool,
Delft, 1923.
(3) PFEIFFER, J. PH.: De Ingenieur 64, P4-P10 (1939).
(4) PFEIFFER, J. PH.: De Ingenieur 64, Mk41-Mk47 (1939).
( 5 ) PEFIFFER,J. PH.,AND DOORMAAL, P. M.VAN:J. Inst. Petroleum Tech. 22, 414
(1936).
16) SAAL,R. X , J., AND LABOUT, J. W. A.: J Phys. Chem. 43, 149 (1939).

RHEOLOGICAL PROPERTIES OF ASPHALTIC BITUMENS'

R. N. J. SAAL AND J. W. A. LABOUT
Laboratoraum N . V . de Bataafsche Petroleum Maatschappij, Amsterdam, Holland

Received August 7, 1959

I. INTRODUCTION

In numerous investigations, the rheological properties of asphaltic

bitumens have been found to vary widely in elastic deformability and
thixotropy (5, 6). From the point of view of their thixotropic properties,
they may be divided arbitrarily into two groups-those of the sol type
and those of the gel type (3). Elasticity may be due to the elastic de-
formability either of the separate micelles or of a structure built up by
coherent micelles; it may therefore occur both in bitumens of the sol type
and in those of the gel type (4).
The wide and continuous range of chemical composition observed in
all asphaltic bitumens makes it likely that free (sol) micelles or isolated
small agglomerates of micelles, which may themselves be considered free
micelles, may be present in bitumens of the gel type as well as in those of
the sol type. Among the technical asphaltic bitumens, therefore, many
representatives of mixed gel-sol types may be expected.
In order to be able to judge in how far the rheological properties of
asphaltic bitumens may be explained by regarding them either as sols
or as gels, a close study was made of the rheological properties of two quite
different bitumens, not only the deformation under constant shearing
stress being investigated, but also the elastic recovery or the relaxation of
the stress. The measurements were carried out in a concentric rotation
viscometer with a conical bottom.
Presented a t the Sixteenth Colloid Symposium, held a t Stanford University,
California, July 6-8, 1939.
150 . R . N . J. a u L AND J. w. A. LABOUT

The chief characteristics of the bitumens used in this investigation are

given in table 1. The rheological measurements were made at 35”C.,a
temperature at which the deformations could be accurately observed. On
the strength of the penetration index bitumen A would seem to belong to
the gel type, bitumen B to the sol type (3). The low C/H ratio of the
asphaltenes should be expected to give both of these bitumens a high
elastic deformability (4).
The following rheological measurements were made: (1) The change in
deformation with time under constant shearing stress (subsequently re-
ferred to here as “deformation”). (2) The degree of elastic recovery
after various deformations. (3) The relaxation of the internal stress
with time after various deformations. The internal stress will be referred
to here &s the relaxation stress. (4) The recovery with time after defor-
TABLE 1
Chief chamten’slics of bitumens A and B

~~ ~ ~

Penetration at 25°C.. ............................. 64 17

R & B melting point, “C.. ........................ 68 62.5
Penetration index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 -0.7
Insoluble in eO/M gasoline, per cent by weight. . . . 25.0 25.3
C/H ratio of asphaltenes. . . . . . . . . . . . . . . . . . . . . . . . . 0.90 0.85
Penetration at 35°C.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 50

mation and partial relaxation of the stress. (5) The relaxation of the
internal stress with time after deformation and partial recovery.
11. THEORETICAL CONSIDERATIONS

For mathematical analysis asphaltic bitumens or colloidal systems in

general may be typified as (1) pure sols in which each micelle is independ-
ently mobile, (2) pure gels in which all the micelles are interconnected to
form a structure, or (9) gel-sols in which part of the micelles form a
structure, but others are independently mobile.
Using a rough theoretical model somewhat different from that of Burgers
(l),flow equations for the mixed gel-sol type may be derived. Figure 1
represents in two dimensions a unit of volume containing one free micelle.
The dimensions of the free micelle, the elasticity of which is indicated by
GI,are given by p and q. The deformation of the element is indicated by
s, that of the free micelle by s’. Further, there is an elastic element of the
structure Gp, to which, for the sake of simplicity, no vulume has been
sssigned. Rotations of the free micelle are left out of consideration. The
theoretical analyses of the five types of rheological measurements reported
are here indicated briefly.
 RHEOLOGICAL PROPERTIES OF BITUMENS 151

1. Deformation
For the deformation of this model system the following equations may
be set up:
ds

which says that the stress resisting deformation of the element is equal to
the sum of the viscous stress in compartments I1 and I and the elastic
stress Gas; and
-GI = -.!
s’
Q 1--q dt
(- -;ii)
ds - ds’

which says that the elastic stress in the micelle is equal t o the viscous
stress in compartment I above it.

I I
I
Liquid ‘ O f
Viscositk q
I

v’ ,’ I = --G
Fme Micelle
4
GI
P I+

E’IO. 1.

When solved for the case of a particular value of the constant 7 these
equations give

in which
XI and X I = -3{(1 -q + pq)a + a ) f $7
Here

and
152 R. N. J. SAAL AND J. W. A. LABOUT

As elements with different properties must be combined to obtain a

real material, the dependence of the deformation on time becomes more
complicated. The integral values of the deformation, however, probably
will be proportional to the stress whenever time is taken to be constant
making
s = 7fl(t) (14
At large deformations real structures will be destroyed and the applica-
bility of equation l a is, therefore, limited.

2. Recovery
Recovery phenomena may be derived from differential equations
similar to those used above:

and

If these differential equations are solved for the case in which recovery
follows immediately upon a deformation with shearing strPss 70 during
time to, an equation of the same type as equation 1 is obtained, which can
be reduced to
se = 70(fl(tO + t) - .fi(t)} (2)
where se is the part of the total recovery se0 which the bitumen still has
to cover a t the end of the time of recovery t. Referred to the previous
deformation so or to the total recovery ss0, equation 2 becomes

3. Relaxation
Relaxation is possible only with free micelles. Therefore thc formula
cannot be of type 1 or 2. If the relaxation follows immediately after a
deformation during t o under 7 0 , then

=~~s,, + ?EY (e'l'o - e's'o)"'' (3)

where
6 = (1 - p) (1 - q)a
 RHEOLOGICAL PROPERTIE8 OF BITUMENS 153

I n more general forms this equation is

7,/r0 = Gzfl(to) - fdto)e-" (34
or even
78/70 = Gzfi(t0) - .f3s(tO,t) (3b)
Therefore relative relaxation stress-time curves must coincide with
equal times of previous deformation, and it should further be possible to
make relaxation curves with differing to, but with equal SO, coincide by
shifting them in the direction of the time axis.

4 Recovery ajter partial relaxation

If,after deformation s!, a certain relaxation to the stress ~~0 takes plhce,
followed by recovery, a recovery curve will have to be found coinciding
with one after a deformation of so under a smaller shearing stress, of such
a magnitude that its relaxation stress at the first moment equals rR0.

5. Relaxation after partial recovery

In deformations to so by r0during to and partial recovery to s, during t,
followed by relaxation, the value of the elastic deformation of the free
micelle after the partial recovery must first be calculated. With large
enough recoveries this deformation can be negative, in which case during
the relaxation following, the stress will rise.
For this case the following formula can be obtained:
7 B = TR(t0 + te) - 781, (4)
This equation says that the relaxation stress in question is equal to the
difference between the relaxation stresses after deformations with the
+
same shearing stress during times (to te) and t,, if the three values of r g
are taken after equal relaxation time.
The above formulas are derived for mixed gel-sol type systems and
therefore also comprise the separate cases of the pure sol and pure gel
type. Since the particular bitumens we examined showed pronounced
relaxation which occurs only with free micelles, only the equations for the
pure sol type need to be deduced.
(1) Deformation in connection with time is then found to be
154 H. N. J. SAAL AND J. W. A. LABOUT

(8) The recovery is represented by

and the total recovery is therefore given by

Seo = ~0f3(t0)

so that formula 5a can also be written as

s - s0
. = rAt
Furthermore, formula 6 may be reduced to

so that for the sol type we must always refer the recovery to the total
recovery.
(3) For the relaxation

so that the generalized form

78 = TO.f4(tO, t)

becomes analogous to equation 3b.

(4) When after deformation partial relaxation is followed by recovery,
the relationships are the same for the sol type as for the more general
gel-sol type.
(5) When after deformation partial recovery is followed by relaxation,
for the pure sol type no rise in the relaxation stress is possible but only a
regular decrease to zero. It is not quite certain, however, that this will
continue to apply with a more generalized model.
Before ascertaining in how far measurements are in agreement with
the formulas, the following remarks have to be made: Owing to the
continuous transition in chemical nature between the components that
form the bitumen, it cannot be expected that there should be a sharp
distinction between free and bound micelles in the mixed gel-sol type.
Continuous transition may also be expected in the force with which the
micelles are bound together. If loosely bound micelles are present, the
theories given above may not fully apply.
The theory is based on the contrast between free and bound micelles.
A similar mathematical analysis might be based on differences in the rate
of mutual binding of micelles.
 RHEOLOGICAL PROPERTIES OF BITUMENS 155

111. MEASUREMENTS
I. Deformation
In figures 2 and 3 curves for deformations under constant shearing
stress have been drawn for bitumen A and bitumen B by plotting the
logarithm of the deformation (log s) against the logarithm of the time of
deformation (log t ) , s being expressed in centimeters qf shear per centi-
meter of layer thickness and t in seconds. It appears that with small
deformations the distances between the lines at the same times are ap-
proximately equal to the logarithms of the ratios of the shearing stresses
applied, which is in agreement with formula 1 or formula 5.

FIG. 2. Relation between deformation and time. Bitumen A

With greater deformations this rule no longer holds for bitumen A.

When the rate of deformation for different values of t is calculated, it
further appears for curves DI and DZfor bitumen A that the rate of defor-
mation increases with the time for greater deformations, which in the
graph is clear from the slope of the curves. This is in agreement with the
assumption that bitumen A contains a skeleton which is broken down a t a
certain deformation. It has further appeared that after a sufficient time
of rest (1 t o 3 days) the curve can be completely reproduced, whereas this
is not possible a t once, which indicates that this bitumen has the property
of self-healing or thixotropy. Beyond the deformation a t which structure
is temporarily destroyed the curves cannot be derived from the model.
156 R. M. J. BAAL AND J. W. A. LABOUT

For small deformations the curves for bitumen A are practically straight
and can therefore be represented by
s = Tat" (8)
Hence in formula la, fl(t) = at", where a = 1.38 X 10-6 and n = 0.38.
Curves for bitumen B have been found to display a distinct curvature
a t small deformations, but after great deformations the slope becomes
practically 1, so that ds/dt is constant. Hence there are here no indications
for skeleton breakdown.

FIQ.3. Relation between deformation and time. Bitumen B

2. Recovery
In figures 2 and 3 are also plotted the curves el, e2, and e3,which indicate
the total elastic recovery observed on releasing the external stress im-
mediately after deformation during the period indicated on the abscissa.
The distance between the corresponding lines D and e therefore gives for
each time of deformation that part of the deformation which is permanent.
It is clear that for bitumen A the total recovery after small deformation is
complete, whereas this is not true in the case of great deformations. This
makes it probable that in the case of small deformation bitumen A must
be considered to have a coherent structure. In the case of bitumen B such
a structure seems to be absent from the start.
With bitumen A the e lines display a distinct maximum near the defor-
 RHEOLOGICAL PROPERTIES OF BITUMENS 157

mation at which the D lines display the greatest curvature, indicating that
after breakdown of the skeleton the total recovery becomes smaller.
The order of magnitude at which serious breakdown occurs corresponds
with that calculated by Kratky (2).
With bitumen B also the e lines afford indications of a breakdown either
of a weak structure or of structures in small agglomerates, although for
this asphalt the D lines do not disclose this property. In any case the
behavior of bitumen B seems to approach that of the sol type, and for
this reason we have further plotted, in figure 3, log (s - sea) against log
time (curves D'), obtaining straight lines of slopes very close to 1 which
approximately satisfy equation 5b. The secondary effects which account
for the small deviations have been treated in previous publications (5).
It may further be noted that also in the case of bitumen A, log (s - sea)
plotted against log time gives practically a straight line but the slope of this
line is not always 1, and it is therefore described by the equation
s - S.O = bt" (9)
where m may be > 1 and can to a certain extent be used as a measure for
thixotropy (4).
In figure 2 the first parts of the e lines coincide with the D lines and can
therefore be represented, for these or even greater deformations, by the
formula
Seo = Toat; (10)
In figure 3 the first part of the e lines, unlike that of the D lines, is
straight and hence can also be represented by formula 10, where a =
0.49 X and n = 0.50.
It is important to define the conditions under which no skeleton break-
down occurs. I t was found that immediately after one experiment good
duplicate results could be obtained as long as the previous deformation
did not exceed s = 0.6 to 1.0 a t most.
For bitumens A and B the rates of recovery are given in figures 4 and 5
by plotting se/se0against time. This permits comparison of the results
with the theoretical recovery equation (2a), into which it has now become
possible to introduce the specific form of the function as given by equations
8 and 10, so that we have

In figures 4 and 5 there is a regular displacement of the lines dependent

upon the duration of the preceding deformation as shown on the graphs,
and apparently independent of other conditions. Quantitatively there is
reasonable agreement, points of equal t / t o which according to equation 11
158 R. N. J. S M L AND J. W. A. LABOUT

should have equal values of s./seO1 and which in a few cases have been
connected by lines, displaying differences between maxima and minima
of 0.05. Direct calculation of s,/scp by means of formula 11 gives values
for bitumen A that are somewhat lower than those found experimentally,
-
1"

FIG.4. ltelation between S./S.O and time after various deformations at different
shearing stresses. Bitumen A. Times on curves refer to duration of previous def-
jrmation. Horizontal dotted lines connect points of equal t/to.

NO. t +O

ec& dynar per rg.cm.

0.30 0.30 30 s,m
0.75 0.64 240 58,m
1.27 0.87 600 s,m
27.5 0.615 6Ooo 58,OOo
0.255 0.245 120 28,700
0.55 0.46 600 28,700
0.87 0.625 1500 28,700
0.135 0.12 360 7,080
0.375 0.255 2400 7,080

the maximum difference being 0.10; for bitumen B the equation applies
within the limits of error of our experiments. There therefore appears to
be satisfactory agreement with the model.
With values of so greater than 1, on the contrary, no agreement is found,
probably owing t o breakdown of structure.
 RHEOLOGICAL PROPERTIES OF BITUMENS 159

FIG.5. Relation between se/sCo and time after various deformations a t different
shearing stresses. Bitumen B. Times on curves refer to duration of previous def-
ormation. Horizontal dotted lines connect points of equal t l t a .

NO. CAICULATED

SeUWIds !gnu per sq.cm.

1 0.090 0.071 5 66,700
2 0.24 0.146 20 66,700
3 0.382 0.21 40 66,700
4 0.69 0.32 90 66,700
5 1.51 0.465 240 66,700
6 3.37 0.473 600 66,700
7 45.4 0.38 7740 66,700
8 0.161 0.094 30 33,000
9 0.438 0.195 120 33,000
10 2.40 0.356 900 33,000
11 15.8 0.282 6ooo 33,000
12 0.225 0.064 270 8,130
13 0.945 0.128 1500 8,130
14 4.22 0.131 7200 8,130
15 7.5 0.135 14100 8,130

S. Relaxation
In figures 6 and 7 the relaxation stress after various previous deforma-
tions under different shearing stresses have been plotted against time as
160 R. N. J. SAAL AND J. W. A. LABOUT

Q/TO. Here, too, the phenomenon discussed in the previous section is

observed: up to deformations of SO = about 1 the curves fall in the expected
order in accordance with the theoretical relaxation equation (3b), but a t
greater deformations these curves again lie in reversed order.
A second possibility of checking the relaxation equation is to compare
the relaxation curves at equal values of so but different values of to, since
by displacing them in a 7R-t diagram in the direction of the time axis
it should be possible to make them coincide in accordance with equation 3.
A few cases have been illustrated in figures 6 and 7, which must now be read
as ~ - diagrams
t according t o the scale division a t the right-hand side
of the diagrams. In these figures relaxation curves for a low shearing
stress have been laid with their first points on a curve of higher shearing
stress with the same previous deformation so. The agreement proves to
be satisfactory.
4. Recovery after partial relaxation
In figure 8 the curves A and D are ordinary recovery curves of bitumen
A, while curves E to G represent recoveries after partial relaxation. In
all cases so is about 1. It appears that the curves are of the same shape
for both groups. The relaxation stress a t the beginning of recovery agrees
with that predicted by the model. The conditions of the deformation in
experiment B are practically the same as those employed in the experiment
represented by curve 14 of figure 6, for which by extrapolation a relaxation
stress at t = 0 is obtained of about 26,000 to 28,000 dynes per square
centimeter. It will be seen that this recovery curve B coincides fairly
well with curve F , representing an experiment where a higher stress was
released to 29,000 dynes per square centimeter before recovery started.
Similar results were obtained using bitumen B.
5. Relaxation after partial recovery
This type of experiment has little accuracy, since calculation of the
results by means of the model involves a difference between two measure-
ments of relaxation stress.
In figure 9 is given the result of a single experiment using bitumen A.
In agreement with the model it is found that the relaxation stress can rise
a t first.
Calculation of the relaxation stresses by means of equation 4, using the
data of figure 6, gives reasonable agreement with the observations indi-
cated by figure 9.
The gradual decrease of the stress shown by the bitumen after longer
relaxation time is not in agreement with the model, for according to the
model the relaxation stress must approach a constant value Gzs,. Here
again the discrepancy may be attributed to skeleton breakdown, and
better agreement would be expected at lower deformations.
 RHEOLOGICAL PROPERTIES O F DITUMENS 161

FIG.6. Relation between m / r Oand time after various deformations a t different,

shearing stresses. Bitumen A . Times on curves refer to duration of previous
deformation.

NO. 'a

srconda dunu pcr aq.cm.

1 0.176 5 58,000
2 0.203 10 58,000
3 0.225 15 58,000
4 0.247 20 58,000
5 0.262 30 58,m
6 0.53 147 58,000
7 0.75 242 58,000
8 1.27 546 58,OOo
9 3.26 1559 58,000
10 27.2 6660 58,000
11 0.083 5 28,700
12 0.247 120 28,700
13 0.53 600 28,700
14 0.90 1500 28,700
15 3.86 7200 28,700
16 16.9 21720 28,700
17 0.13 499 7,080
18 1.50 24000 7,080

In the case of bitumen B similar phenomena are observed, as shown by

figure 10. Here, too, a maximum occurs in the relaxation stress-time
curve after which, however, the relaxation stress decreases in the normal
162 R. S . J. SAAI, AND J. W. A . LAROIJT

L.
r.

04
-+ESP*& f l d - m ' l P c m p d h e x p ~
0 ...e . - 46: .. .. c ..

02

FIG.7. Relation between T R / T ~and time after various deformations a t different

shearing stresses. Bitumen B. Times on curves refer to duration of previous
deformation.

NO.

aecondi dunw per S Q . C ~ .

1 0.097 5 66,700
2 0.24 23 66,700
3 0.375 42.4 F.6,700
4 0.68 92.4 66,700
5 1.50 242 66,700
6 3.34 603 66,700
7 8.4 1481 66,700
8 44.2 7525 66,700
9 0.052 10 33,000
10 0.24 58 33,000
11 0.68 208 33,000
12 2.24 900 33,000
13 16.1 6Ooo 33,000
14 40.3 15600 33,000
15 0.24 280 8,130
16 2.68 5100 8,130
17 7.5 13780 8,130

way. These results indicate a weak structure for bitumen B as contrasted

to a stronger structure for bitumen A.
From the foregoing discussions it has become clear that there is a
satisfactory agreement between material and model.
 RHEOLOGICAL PROPERTIEB OF BITUMENS 163

From the data given it appears that in the two bitumens examined there
must be a certain amount of coherence between the micelles. In the case

., .. ..
. . . . . . . . .

I" ..con&!

FIG.8. Relation between elastic recovery and time after partial relaxation.
Bitumen A

I
8
3-

FIG. 9. Changes in relaxation stress after partial elastic recovery. Bitumen A

of bitumen A this follows from the complete recovery after small deforma-
tions, from the maximum in the curve for the total recovery and from the
higher rate of deformation, recovery, and relaxation at high deformations,
164 R. N. J. S M L AND J. W. A. LABOUT

the last four factors pointing to breakdown of structure. In the case of

bitumen B only the maximum in the curve for the total recovery and the
higher rate of recovery and relaxation at high deformations were observed.
At deformations small enough to show no serious breakdown of the skele-
ton, a satisfactory agreement with the model system, representing a mixed
gel-sol system, was found in the five different ways of investigation used.
From this we conclude that the model in the main represents the constitu-
tion of the bitumen.
I

6
3
r'

I I I II I I I I I I I I I I I I I I I I I I

4
0 1 0 0 . 0 0
400 .oo ma IOW I1w

t. +(. (*cord.)

FIQ. 10. Changes in relaxation stress after partial elastic recovery. Bitumen B

As bitumen A showed complete recovery at small deformation and

pronounced skeleton breakdown at larger deformations, this bitumen
must possess a marked gel structure. As bitumen B showed incomplete
recovery and skeleton breakdown to only a slight degree, this product must
be considered as containing either free agglomerates of micelles or very
slight coherence throughout the material and is therefore more in the
nature of a sol. This is in agreement with the theories formerly given by
Pfeiffer and van Doormaal (3).
It will not be possible, however, always to account for still more compli-
cated treatment. The following experiment may serve as an instance:
The bitumen was deformed to s = 1, then by an opposite stress to s =
-0.4, and then released. There was at first recovery to s = +0.15, when
 RHEOLOGICAL PROPERTIES OF BITUMENS 165

the movement was reversed. After rcaching s = +0.10 it was again

reversed, reaching a final position of s = $0.12. To explain such phe-
nomena it will be necessary to assume coupling of differently formed
elementary systems (1).
1V. SUMMARY

1. Rheological measurements in a conicylindrical rotation viscometer

were carried out on two asphaltic bitumens of different types, employing
methods of investigation involving deformations under constant stress,
elastic recovery, relaxation, and combinations of these.
2. The experimental results were found to be in satisfactory agreement
with the properties of a theoretical model representing the simplest iorm
of a mixed gel-sol system.
3. It is concluded, in accordance with the views of Pfeiffer and van
Doormaal, that asphaltic bitumens must be considered such mixed gel-sol
systems and that the degrees of structure exhibited can vary widely with
composition.

The authors are indebted to +heManagement of the N. V. de Bataafsche

Petroleum Maatschappij for their permission to publish this paper and to
their collaborators in these investigations.
REFERENCES
(1) BURGERS,J. M.:First Report on Viscosity and Plasticity. Academy of
Sciences, Amsterdam (1935).
(2) KRATKY, 0.:Kolloid-Z. 70, 14 (1935).
(3) PFEIFFER,J. Pa., AND DOORMAAL, P. M. VAN: J. Inst. Petroleum Tech. 22,414
(1936).
(4) PFEIFFER,J. PH.,AND SAAL,R. N . J.: J. Phys. Chem. 43, 139 (1939).
(5) SAAL,R. N . J . : J. Inst. Petroleum Tech. 19, 176 (1933);Proc. World Petroleum
Congr., London, 1939, 11, 515.
(6) TRAXLER, R. N.,AND COLLABORATORS, a . 0 . : J. Applied Phys. 8, 291 (1937); J.
Phys. Chem. 40, 1133 (1936).