VISCOSITY MEASUREMENTS OF ASPHALT-RUBBER BINDERS
R.A. Jimenez University of Arizona
Report ATIT-1
For Presentation at the
National Seminar on Asphalt-Rubber October 30-31, 1989
Kansas City, Missouri
Arizona Traffic and Transportation Institute College of Engineering and Mines
The University of Arizona Tucson, Arizona 85721
July 1989
120
VISCOSITY MEASUREMENTS OF ASPHALT-RUBBER BINDERS
Abstract
Early unpublished work is presented on the measurement for
viscosity of asphalt-rubber blends. The viscometers used were
selected with the thought to minimize "side friction" effects on
the flow of asphalt-rubber. The viscometers were of the falling
coaxial cylinder type and the forced flow Schweyer Rheometer. Most
of the work done was for evaluation of the Schweyer device for
determining effects on rheometer and rubber variables on measured
viscosity. These variables were (a) diameter of flow-tube, (b)
amount of rubber in the blend, (c) gradation of rubber particles,
(d) type of rubber, and (e) test temperature. The shear stress
shear rate relationship was taken to be a power function ( T =
I i 0) and viscosity ( T / i ) was calculated at a shear rate of 1
sec-1 and therefore equal to I the intercept at 10° on log-log
coordinate axes. The data indicated that:
1. the viscosity be calculated at a shear rate of 1,
2. flow tubes of 9.70 rom and 12.70 rom yielded the same
values of viscosity,
3. the flow tube of 2.43 rom (a standard) was not considered
adequate for asphalt-rubber,
4. the 25C viscosity of one type of rubber blend increased
as the rubber content increased from 20 to 25 to 30
percent,
121
5. there was not a direct relationship between viscosity and
range of particle size,
6. the viscosity-temperature susceptibility of one rubber
blend was greater than for the other, and
7. the Schweyer Rheometer was considered a good device for
comparing rheological properties of asphalt-rubber
blends.
122
INTRODUCTION
Asphal t-rubber was developed and patented by C. H. McDonald
during the early 1960s. The use. of asphalt-rubber (A-R) has
increased dramatically since then for highway construction, as a
binder, waterproofing layer, or a strain attenuating layer to
minimize reflection cracking. Public use of A-R requires that
there be specifications of its properties, one of those being the
rheological (flow) characteristics of A-R blends. The material
reported here is concerned with early and unpublished work that was
done for measuring the viscosity of asphalt-rubber blends.
Asphalt-Rubber
The asphalt-rubbers used in the work done were mixtures of
asphalt and fine grindings from rubber tires. The amount of rubber
in the early mixtures varied from 20 to 30 percent by total weight
of the asphalt rubber blends. These mixtures were different from
the old rubberized asphalts in that synthetic rubber (SBR) was used
instead of natural rubber (latex), the amounts of rubber mixed with
asphalt was much greater than the 3 to 5 percent for latex, and the
maximum particle size of the SBR was about 1.2 mm (0.05 in.) and
for the natural rubber about 0.07 rom (0.003 in.).
The rubber and asphalt are mixed at temperatures between 138
to 190 C (280 to 375 F) for some specified period of time. (In our
work, 30 minutes at 191 C (375 F). The dispersion of the rubber
123
particles to produce the desired improvements in the asphalt may
be affected by the following:
1. Mixing temperature--usually detrimental if held too long
above .216 c (420 F) [1].
2. Duration of mixing time, the effect is also dependent on
temperature; however, the effect becomes constant after
a minimum time [2].
3. stirring shear--break down of rubber if too high [1].
4. Particle size and its distribution.
5. Type and quantity of rubber.
6. Amount of aromatic (cyclic) component in the asphalt [1].
There is somewhat general agreement that rubber is not soluble in
asphalt and that under optimum conditions a specific particle size
may increase in volume (swell) by a factor of up to 5 for natural
rubber [2] and by a factor of up to 3 for synthetic rubber [3].
124
VISCOSITY TESTS
Materials
The materials used in this work were furnished by the Arizona
Department of Transportation (ADOT). At that time, ADOT was doing
considerable research and construction with asphalt-rubber. As a
consequence, no specification tests were performed on the materials
other than those used to quantify the variables of the program
being reported.
Asphalt
The two grades of asphalt used were of AR-1000 and AR-4000;
however, different batches of AR-1000 were used for work done at
different periods of time. The asphalts were assumed to contain
a sufficient amount of an aromatic component to react with the
rubber. An oil extender (an aromatic compound) was included to be
mixed with one of the rubber blends.
Rubber
The rubber particles used in the A-R mixtures were identified
with the letters TP for one source and G for the second source.
Gradations of the rubber types are shown in Table 1. In Table 2
are listed the gradation characteristics of coefficient of
uniformity and coefficient of curvature for the rubber sizes and
their combinations. The data in the tables show that the largest
particles were less than 1.2 mm (0.05 in.) and generally larger
125
Table 1 - Gradation of Rubber Particles
sieve size #8 #16 #30 #50 #100
Opening, in. 0.094 0.047 0.023 0.012 0.006
Opening, mm 2.4 1.2 0.58 0.30 0.15
Percent Passing sieve
TP-044 100 98 20 2 0
TP-027 100 95 23 12
TP-0165 100 75 20
G-274 100 72 32 12
Table 2 - Gradation of Parameter Values of Cu and Co for Blends of Various Rubber Particle sizes
Rubber Blends
G-274
TP-044
TP-027
TP-044 + TP-027, (1:1)
TP-044 + TP-027 + TP-0165, (1:1:1)
TP-0165
coefficient of uniformity
---------, coefficient of curvature
126
C * u
3.6
1.8
3.0
2.7
3.2
2.5
C ** o
1.3
1.1
1.6
1.2
0.9
1.2
#200
0.003
0.08
0
1
5
than 0.15 rom (0.006 in.). The G-274 rubber particle sizes had the
largest range while the TP rubber particles were more of a one
sized distribution.
127
Viscometers
Important characteristics of asphaltic binders are the
rheological properties and how they influence their usage and
performance. These materials are often specified with reference
to viscosity values and durability. During the initial study for
measuring viscosity, attention was given to the usual viscometers
available to us for making such measurements. In consideration of
the swelling of the rubber particles, the Saybolt viscosimeter was
discounted because of the size of orifice and also the Brookfield
because of the possible alignment and separation of particles by
the rotating spindle. Our first determinations were performed with
a home-made falling coaxial cylinder viscometer and later on with
a purchased Schweyer Rheometer.
Falling Coaxial Cylinder
A schematic diagram of the viscometer is shown in Figure 1.
As can be seen, the device was composed of an outer cylinder, an
inner coaxial cylinder, and the cylinders were separated by an
annulus of the sample. The shear stress was obtained at the inner
surface of the annulus and the velocity of the inner cylinder was
determined by following its rate of displacement with a
cathetometer for low viscosity materials on an extensometer dial
gage for high viscosity ones. A derivation of an equation for the
calculation of viscosity is given by Jimenez in Reference 4.
The dimensions of the viscometer were established in
consideration of the swollen maximum particle size of rubber. The
128
I-4--Weight L-..f'\n:-'t-..J
\j~- Coaxial Cylinder (falling)
/]...011-- 0 ute r Cylinder
Figure 1. Schematic of Falling Coaxial Cylinder Viscometer
129
Stand
largest particle was 1..2 nun (0.05 in.) and if it increased in
volume by a factor of 3, its diameter would be 3.6 mm (0.15 in.).
It was assumed that to have free flow through a tube, the tube
diameter should be 3 times the maximum particle size. On that
basis a tube diameter of about 12 .. 7 nun (0.50 in.) would be needed
or an annulus thickness of 6.85 nun (0.25 in.) for the viscometer.
In comparing results obtained with annuli of various
thicknesses and a constant height of 25.4 nun (1.0 in.), it was
established that the 6.85 X 25.4 nun (0.25 X 1.0 in.) annulus was
satisfactory for making the viscosity measurements.
Schweyer Rheometer
The device was obtained through funds from the National
Science Foundation and purchased from the Cannon Instrument
Company. The basics of the device are shown in Figure 2 and the
determination for viscosity is as follows:
The sample (asphalt) is placed in the sample tube and forced
into the test tube which has the desired capillary diameter. A
controlled air pressure is held in the air cylinder and the force
on the piston is transferred to the plunger for applying a constant
stress on the asphalt. Knowing the diameters of the sample tube
and test tube the stress on the sample (asphalt) in the capillary
can be calculated. The velocity of the plunger is obtained with
the LVDT connected to a strip chart recorder. Descriptions and
operations of the rheometer are given in Reference 5 and the manual
furnished by Cannon.
130
AIR CYLINDER
PISTON
PLUNGER
I
SAMPLE TU~E
TEST T UBE (CAPILLA RY)
II I !I H
DS ,I I
~ ~ LS
~ ··l ~ t -'I-
-II- DT
/-( ~ ) PRESSURE GAGE
.
r---{\~ ~
LVDT
I
I TES DENOTES T MATERIAL
Figure 2. Schematic of Schweyer Rheometer
131
Examination of the diagram and conversations with SChweyer
indicated the feasibility of using the sample tube as the measuring
tube and to also increase its diameter to 12.7 rom (0.50 in.).
Measurements for viscosity were made with tubes of 2.43, 9.70, and
12.7 rom in diameter (0.09, 0.38, and 0.50 in.). The 2.43 rom (0.09
in.) tube was the largest size supplied with the device.
Work Programs
The work reported here was done over three different periods
of time; however, the tables shown below are for the investigations
performed with the SChweyer Rheometer. The first viscosity
measurements were obtained with the falling coaxial cylinder
viscometer and the results are presented for comparison with
viscosity values resulting from use of the rheometer.
Table 3 lists the tests that were performed with the
Schweyer Rheometer. Although the two phases are shown under one
table the testing periods and asphalt cements of AR-1000 were
different.
Phase 1 of Table 3 shows the main variables to be
investigated for responses with the rheometer were the test tube
diameter and also the particle sizes of the rubber.
In Phase 2 of the table it is seen that the principal
variables were gradation of the rubber particles, the amount and
type of rubber.
132
..... tH tH
r-~
Rubber Bl end Gradation of Rubber
Test Temperature 15 DC
Test Tube Size 9.70 12.7 Diameter, mm
Type of Rubber
Type of Asphalt
Test Temperature DC
Gradation TP-044 of Rubber
Amount of Rubber, 20 25 % BTW
Table 3. Summary of Tests Performed
Phase I - Rubber Content of 25 Percent BTW
TP-Rubber and AR-I000 Asphalt
TP-044 TP-0165
25 35 25 35
2.43 9.70 12.71 2•43 9.70 12.7 2.431 9.701 12.7 2.43 9.70 12.7
Phase II - Test Tube Diameter of 9.70 mm
TP . ~ Gr274
AR-IOOO AR-4000 AR-4000 AR-1000
25 4 25 4 25 25
TP-044 +
TP-027 TP-044 TP-027 TP-044 TP-027 TP-044 TP-044 + + +
TP-027 TP-0165 TP-027 (I: 1) (1:1:1) (1:1)
30 25 25 25 25 25 25 25 20 10 20 20
RESULTS AND DISCUSSIONS
Results of the viscosity measurements made are presented and
were based assuming that the relationship ~etween shear stress and
shear rate, T vs. i , could be expressed as a power-law fluidity.
This expression was as follows:
; = I l' e ----------------------------- 1
The log-log coordinates of ; and 1'Plot as a straight line; I is
the stress value (intercept) where i' is equal to 1 (100) and c is
the geometric slope of the line.
Falling Coaxial Cylinder
Table 4 shows the viscosity values obtained for both AR-1000
and asphalt-rubber with 25 percent TP-044. The data suggest the
higher viscosity for the A-R blend at higher temperatures (but
< 600 C) and lower viscosity at lower temperatures (~ OOC). The
testing shear rates of the tests were selected at that time to
bracket a selected value of 5 X 10.2 sec· I •
Schweyer Rheometer
As indicated earlier, the work period and the AR-1000 were
not the same for the two phases.
134
Table 4 - Viscosity of AR-1000 Asphalt and Asphalt-Rubber by Falling Coaxial Cylinder Viscometer
Temperature °C. Viscosity, ( ~ = 1.0)104pa-sec
Temperature °C. Viscosity, ('Y = 1.0) 104Pa-sec
135
15 106
14 113
AR-1000
25 3.8
Asphalt-Rubber
25 8.6
35 1.3
40 2.2
Table 5 - Viscosity of Asphalt Rubber by SChweyer Rheometer Affected by Tube Diameter and Temperature
Viscosity (i' =1. 0) 104 Pa-sec
Temperature, ·C
Tube Diam mm
2.43 9.70
12.70
2.43 9.70
12.70
*Test tube diameter of 3.19 rom
136
15
46* 77 80
25
TP-044
4.0 8.0 8.0
TP-0165
2.4 8.2 8.3
35
1.1 5.1 4.8
1.3 4.9 5.1
Phase 1
The results of the tests listed in Table 3 as shown on Table
5 and depicted graphically on Figure 3. In Table 5 it is seen that
there is not much difference on the viscosity values obtained with
the 9.70 and 12.70 mID test tubes for either the TP-044 or TP-0165
rubber particles. The effects of temperature and test tube size
on viscosity are more easily visualized with the curves of Figure
3. The figure shows there was not much overlap of the rheograms
for the two tubes and also that the values for I or viscosity at
a shear rate of 1 sec-1 did not need extrapolation of the curves.
It is apparent that the viscosity obtained with the 2.43 mID
diameter tube were lower than those with the 9.70 mm one. This
result was contrary to first thoughts since one would expect the
swollen rubber particles with sizes of up to 4.5 mID to restrict
flow through the 2.43 diameter tube. The reason for this
difference is given as follows on a conjectural basis.
137
5
0105 a...
5
•
5 . -I "(, sec.
/0. _ 6.:0 - Dlo.-2.43mm
~""-,.c.. .... - I / I::i.
:/ V
/~
Figure 3. Typical Asphalt-Rubber Rheograms
138
It is noted that the viscosity of the asphalt AR-1000 as shown
in Table 4 was 3.8 x 104 Pa-sec and 4.2 x 104 Pa-sec shown in Table
8. These values are very close to the 40 x 104 Pa-sec shown for
the A-R measured with the 2.43 rom tube. This would seem to
indicate that asphalt only was being forced through the test tube.
For this to have occurred, the rubber particles had not formed a
structural skeleton which would have carried part 0 the plunger's
force. If so, then the hydrostatic pressure was not resisted by
the elasticity of the particles and piping of the asphalt was
allowed. It is emphasized that sufficient data are not now
available to support the above statements.
Examination of the data in Table 5 will show that the small
sized particles of TP-0165 in the A-R blend had lower viscosity
than the TP-044 blend when measured with the 2.43 rom tube, but had
essentially the same viscosity when determined with either of those
larger 9.70 or 12.70 mm tube. If the larger particles of the
TP-0165 rubber are assumed to swell to 1.5 mm (3 x 0.50 rom), they
should be able to flow through the 2.43 diameter test tube and the
viscosity should perhaps be slightly higher than that for the
straight asphalt (4.0 x 104 Pa-sec). At the present, the above
response differences of results due to tube size and rubber blends
cannot be explained.
Phase 2
The variables in this portion of those works were principally
amount of rubber and particle gradation. Also two types of rubber
139
and grades of asphalt were used. The test tube size had a diameter
of 9.70 mm which was the sample tube (see Figure 2) that was a
standard part of the rheometer.
For this portion of the report, visc?sity will be given as the
value of I ( 1 = 1. 0 sec-I); however, the T - t relationships
were developed from measured shear rates generally ranging from
0.01 to 0.10 sec-I when testing at 25C. Thus, in order to obtain
a value for I, the T - t curve had to be extrapolated over one
decade of shear rate. As a result of the extrapolation, the
viscosity values obtained in this phase should not be compared with
those obtained for similar mixtures of Phase 1.
In Table 6, an examination of the viscosity values shows that
the TP-044 yielded higher values than did the G-274 rubber. The
difference may be due to the greater amount of TP-044 and also that
the G-274 blend contained an extender oil which would serve to
reduce viscosity of the straight asphalt. The data also shows that
between the temperatures of 4 and 25 C, the increase in viscosity
(temperature susceptibility) of the G-274 was much greater than for
the TP-044 blend.
Table 7 has a listing of viscosity determined at 25 C for the
TP-044 and G-274 blends containing different amounts of rubber.
It can be seen that the viscosity of the TP-044 blend increased as
the amount of rubber increased. However, there was no significant
difference in viscosity for the G-274 blends for rubber content of
10 and 20 percent.
140
Table 6 - Viscosity of Asphalt Rubber Affected by Asphalt Grade and Rubber Type Measured with the Schweyer Rheometer at 4 and 25°C Tube Diameter of 9.70 mm
Asphalt Grade AR-I000 AR-4000
Rubber Type TP-044 G-274 TP-044 G-274
Amount of
Temp. , °c
Viscosity 104 Pa-sec
Rubber % 25 25 20 25 20
4 25 25 25 4
(1 =1. 0) , 480 28.6 18.5 64.9 5830
Table 7 - Viscosity of Asphalt-Rubber Affected by Rubber Content and Type Measured with the Schweyer Rheometer at 25° C
20
25
33.2
Type of A-R TP-044 + AR-I00 G-274 + AR-4000
Amount of Rubber, % BTW
Viscosity ( 1 =1. 0) , 104 Pa-sec
20
18.7
141
25
28.6
30 10 20
41.7 32.1 33.2
The effect of rubber particle gradation on viscosity is shown
in Table 8. The data indicates that with exception to the TP-027,
the viscosity increased as the coefficient of uniformity increased;
that is, the range of particle size increased. The reader is
reminded that the gradation (Cu ) was based on the original non
swollen particles. There is no assurance that all sizes swell by
the same factor and that the resulting values of Cu would be in the
same numerical order as for the dry rubber.
142
AR-I000 with 25
TP-044
TP-044 +
TP-027
TP-044 +
Table 8 - Viscosity of Asphalt-Rubber Affected by Gradation of Rubber Cu Measured with the Schweyer Rheometer at 25° C
Asphalt Cu of viscosity (1 = 1. 0) percent Dry Rubber 10· Pa-sec
1.8 28.6
027 (1:1) 2.7 47.2
3.0 37.6
027 + 0165 (1:1:1) 3.2 57.5
No Rubber 4.2
143
CONCLUSIONS
An objective of the work reported was to determine a method
for measuring the viscosity of asphalt-rubber at ambient or reduced
temperatures. The problem contemplated was the interference to
flow of rubber particles that were assumed to swell to about 4rom
(0.16 in.) in size. The work reported was not a unified effort;
it was performed during three different periods of time and with
different asphalts.
limited, they are
conclusions.
However, even though the data are somewhat
considered sufficient for the following
1. The viscosity of the asphalt-rubbers used and others of
similar type can be determined through flow of the
material within tubes of effective diameters of at least
9.70 rom (3/8 inch).
2. The viscometers of the falling coaxial cylinder and
Schweyer Rheometer were adequate for determining
viscosity at ambient temperatures.
3. The viscosity of the TP blends increased as the amount
of rubber increased and also as the range (Cu ) of
particle sizes increased.
4. The viscosity of the G-274 blends showed the greater
susceptibility to temperature.
5. Reproducibility of viscosity value is considered to be
affected principally by the preparation of sample and
144
extrapolation of test data rather than by the test
device.
6. It is suggested that the viscosity of asphalt-rubber be
determined at a shear rate of 1 sec-1 for tempe~atures
ranging from 4-35 C (39-95 F).
145
ACKNOWLEDGEMENT
The majority of the tests were performed by students in the civil Engineering Department. The measurements for viscosity were made over a period of time spanning almost two years and so slight variations in techniques for making and testing asphalt-rubber are certain to have existed. We appreciate the help given and thank those students: W. White, S. Wilson, "and K. Stokes.
146
REFERENCES
1. Endres, H.A., "Latest Developments in Rubberized Asphalt," A
paper prese.nted at the Fourth Annual Highway Conference at the
University of the Pacific, Stockton, California, 1961.
2. Bituminous Materials in Road Construction, Roads Research
Laboratory, Her Majesty's Stationery office, London, 1962.
3. Green, E.L. and Tolouen, William J., "The Chemical and
Physical Properties of Asphalt-Rubber Mixtures," Report
ADOT-RS-14(1621, Arizona Department of Transportation, 1977.
4. Jimenez, R.A. and B.M. Gallaway, "Laboratory Measurements of
Service Connected Changes in Asphaltic Cement," Proceedings,
AAPT, Vol. 30, 1961.
5. Schweyer, H.E., L.L. smith and G.W. Fish, "A Constant Stress
Rheometer for Asphalt Cements," Proceedings, AAPT, Vol. 45,
1976.
147
"Viscosity Measurements of Asphalt Rubber Binders" by Dr. Rudy Jimene3, P.E. University of Ari30na, Tucson, AZ
Question: Was the surface area differential between the 044 and the 027 and 0165 taken into account.
Rudy: The one with the greatest amount of -che finer si3es would have -che greatest surface area_ l would be referring to the California factors in this sense but to answer your question they were not taken into account.
Question: Did you actually run or have run natural rubber content, tests on the rubber you were using.
Rudy: No, I just took the samples as they were sent to me.
Question: As SBR and as natural?
Rudy: Right, as they were sent to me.
Gale Page: Florida Dept. of Transportation, I don't have a question but I do have a comment. I am glad too see that you have used the concept of measuring engineering properties of a material at the tempera-cure, environment, and loading of interest. I believe that it is very important -chat the high or low temperature properties of the AC or the AR binder be measured at or near the temperature of interest which you have done rather than extrapolate results. This may not have the precision or result in the discrimination necessary, so that other indicator tests may be necessary until the technology catches up. But the point is that we should not lose sight of the goal of measuring basic engineering properties before accepting indicator test at high and low temperatures.
Rudy: Thank you Gale.
Don White: University of Ari30na-for the benefit of many people here, I might comment that I have done viscosity work on slurrys, and I consider rubber and asphalt as a slurry even though it is a swollen slurry with a lot of the asphalt in -che rubber. I have been measuring the viscosity up in the range of one to three million centipoises and we had our best success with a modified Inston-rheometer with larger orifices
148
so the slurry would pass through and would give us the pseudoplastic effects, that is, the shear thinning at high shear rates. These are the shear rates that you would find in extruders, maybe 1,000 to 10,000 reciprocal seconds. Then we also used, a Haake rotary viscometer for a portion of the work. Unfortunately, those will only go up to 100 reciprocal seconds. I pass that information to you, but I am wondering today for asphalt mixtures of rubber, what is the most accepted rheometer for obtaining good viscosity data?
Question/Rudy: Are you asking about the Asphalt-Rubber?
Answer/Don White:·
Answer/Rudy:
Don White:
Rudy:
Yes, and I presume that these viscocities are fairly low. 1000 centi-poises or not, or are they higher than that?
They are much higher than that, about 1,000,000 poises at 77°F., and if you go to a lower temperature, they are going to be higher than even 10,000,000 poises.
OK! 50 what viscometers are people using today for Asphalt-Rubber?
This is ten-year old data and I know that they have been measuring viscosity at the higher temperature of 375 F. Jim, would you like to take a crack at that one?
Question/Jim Chehovits, Crafco, Inc:
I really don't know if anyone is measuring viscosity at o the lower temperatures. By low, I mean about 77 F. or below
today. But I think from what we have seen if we look at the data over the last 15 years, probably one of the most appropriate types of viscometer used for the lower temperature area in a relatively solid type range and your looking at the high viscosities is probably the sliding plate type. They are quite a bit more simple than a 5chweyer and would do the best job. Viscosities today are mostly only measured at high temperatures for applica-tion characteristics using either a Brookfield or a Haake as Rudy mentioned earlier.
Question: Does anybody know if at -the operating temperature
149
Answer/Rudy:
like 3500 F., or whatever. Does anyone know whether the mixtures are shear thinning, are they psuedo plastic in nature. In other words, at the higher shear rate, do you get a lower viscosity?
Yes. In figure #3, of the rheogram, note that the slopes of the lines are less than 1:1 (450
).
I do not know personally what they have been doing, as I have stayed away from it for all this time but I know that for control purposes, somebody has suggested a large orifice is tabled but that is only for I might say: "rule of thumb" comparison of con"trol in the field. Maybe the producers here can answer that and contractors could answer that question better than I can.
150