DSR Geometry Impact on Asphalt Binder Linear Viscoelastic Properties

J.A. Rodrigues, G.S. Pinheiro, K.L. Vasconcelos, L.B. Bernucci

Department of Transportation Engineering, University of São Paulo, São Paulo, SP, Brazil
R.C.O. Romano, R.G. Pileggi
Department of Civil Construction Engineering, University of São Paulo, São Paulo, SP, Brazil

ABSTRACT: Asphalt binder is a rheological complex material with elastic and viscous components response,
according to the temperature and loading conditions, being its rheological characterization an important tool
to better predict the material performance. The Dynamic Shear Rheometer (DSR), using parallel-plates geom-
etry, is so far the most common oscillatory testing carried out for rheological characterizations in bituminous
materials. However, the rheological response needs to be characterized considering the stages of binder appli-
cation and use, as the material have different behavior for each phase during hot asphalt mixture production,
application and design life performance. So, the setup of rheological tests must be adequate to correctly repre-
sent the real conditions during each stage. The main purpose of this work was to compare the impact of dif-
ferent test configurations (using parallel-plate, Vane and DIN) in the response of a neat asphalt binder apply-
ing the strain sweep test at different temperatures.

1 INTRODUCTION specific temperatures (isothermal plot), or versus

temperature at specific loading times (isochronal
Rheology (from the Greek, “ρεω”, translated literal- plot) (Airey, 2002; Mezger, 2014). Furthermore, the
ly as “to flow” and “λογοσ”, that means “science”, rheometers present different measuring systems, or
hence, “the study of the flow”) consists in the sci- geometries, that can be used to perform the rheolog-
ence in which the relationship between an induced ical tests. The following sub-sections list distinct as-
stress/strain on a material and the resulting flow or pects of these measuring apparatus according to
deformation is studied (Airey, 2002; Tabilo- Mezger (2014).
Munizaga & Barbosa-Cánovas, 2005). Some materi-
als are rheologically and structurally complex, i.e. 1.1.1 Parallel-plate (PP) geometry
suspensions of particles, paint, blood, foodstuffs and The PP is composed by a statically inferior and a
some asphalt binders, such as polymer solutions, motion superior cylindrical plate (Fig. 1, on the left),
presenting a variety of behaviors. Specifically, for which cause shearing in the sample, producing a de-
asphalt binders, this behavior may vary from New- flection path and a deflection angle. Although, for a
tonian to highly viscoelastic material (one that ex- strict rheological property determination, sample ad-
hibits both elastic and viscous components response) hesion to the plates (avoiding any wall-slip effect)
according to temperature changes and loading condi- and homogeneous deformation throughout the entire
tions (Wood, 1958; Phan-Thien & Mai-Duy, 2017). shear gap must be established. Some advantages of
Empirical properties are still used as an indicative of the apparatus can be mentioned: short preparation
asphalt rheological characteristics, commonly ob- time, errors normally reduced compared to the cone-
tained through softening point and penetration tests. plate geometry, and simple cleaning process. But,
However, it is well known that fundamental tests, melt fracture, inhomogeneous flow, surface effects
based on known physics concepts and equations, can can be pointed as disadvantages; further, larger gap
determine the true properties of the materials, ena- heights, increase the temperature gradient in the
bling to define their behavior and predict perfor- sample (Mezger, 2014; Hung et al., 2015).
mance (Hraiki, 1972; Tabilo-Munizaga & Barbosa-
Cánovas, 2005).
1.1.2 DIN geometry
The DIN geometry consists of two concentric cylin-
1.1 Oscillatory test ders (CC) (Fig. 1, in the center), also known as “cup-
Oscillatory tests conducted in a small-strain range, and-bob”. It has the capability to prevent the materi-
using a dynamic shear rheometer (DSR), are im- al to flow off the shear gap (even for high shear
portant tools to determine and better understand the rates); can guarantee a better temperature control
asphalt viscoelastic properties. Those tests are oper- compared to parallel- and cone-plate geometries;
ated within the region of linear viscoelastic (LVE) and does not require trimming the sample (avoiding
response and this response can be presented in dif- any unexpected disruption). Nonetheless, a large
ferent ways, through the rheological parameters (i.e. amount of material is necessary, and it is more tim-
G*, G’, G’’, η*, δ, tanδ, etc.) versus loading time at
ing consuming to change, prepare and stabilize the There is no consensus on the use of different ge-
temperature (Mezger, 2014). ometries for the determination of asphalt binder rhe-
ological characterization, since several investiga-
1.1.3 Vane geometry tions have been conducted with different
The Vane geometry is composed by several rectan- conclusions with respect to the effect of geometry.
gular vanes connected to a cylinder (Fig. 1, on the The objective of this study was to evaluate different
right). The vane insertion into the sample causes far DSR geometries and their accuracy for the rheologi-
less disturbance which is a significant factor with cal characterization of the viscoelastic properties of
thixotropic systems, such as asphalt binders. How- asphalt binders and then try to identify which one
ever, the sample localized between the vane area would be the most suitable geometry for this pur-
might not be sheared, and there might occur inho- pose.
mogeneous “plastic” deformation behavior, if the
test is conducted at high rotational speeds (Mezger,
2014). 2 EXPERIMENTAL DESIGN

There is a diversity of geometries configurations that The rheological tests were performed on a neat
may led to different viscoelastic responses, due to asphalt binder, CAP 50/70 (penetration grade be-
the type of flow produced. Various authors from dis- tween 50- and 70- 10-1mm). Comparisons were
tinguished areas studied geometry types that might made between PP geometry using 1- and 3-mm
better represent the material characteristics and ap- gaps, and CC with two different rotor geometries:
plication. Hung et al. (2015) used DIN geometry to DIN and Vane. The CC includes a cup with radius of
investigate and compare traditional PP, at 40°C, 30.4 mm and both rotors have 42 mm height, with
65°C and 90°C, for quality control on projects that radius of 15mm (DIN) and 28 mm (Vane), and same
used asphalt rubber binders. Their results indicated 1 mm bottom gap. DSR tests were performed ac-
no significant difference between the two geome- cording to the ASSHTO T315 (2019) standard pro-
tries. Baumgardner & D’Angelo (2012) investigated cedure varying the temperature from 52°C to 88°C,
the ability of the DIN geometry to investigate rheo- applying 20 min. of conditioning time for each one,
logical characteristics of conventional and modified with 0.01% oscillation strain, keeping the frequency
binders. The authors concluded that the DIN geome- at 1.6 Hz. Figure 1 shows the testing plan resume.
try indicates similar results for Superpave parame- Five samples per each gap in PP geometry and three
ters, as well as Multiple Stress Creep Recovery samples per rotor type (DIN and Vane) in the CC
(MSCR) testing, performed at 64°C, which exhibited geometry were tested. At first, limiting value of the
the same peak strain in the creep portion of the linear viscoelastic region (LVR) material range in
curve, indicating that both geometries had the same terms of the shear strain were empirically found by
shear rate. But differences in the compliance results amplitude sweep test, according to ASTM D7175
were evident, reaching values up to 71,4% higher for (2015).
the DIN geometry, when compared to the PP for one
of the tested binders.
Polacco et al. (2003) tested the PP and DIN ge-
ometries, varying from 60°C up to 90°C using neat
and modified asphalt binders. It was observed that
PP geometry presents satisfactory responses until the
material approaches the liquid state, when the ten-
dency to flow out from the PP geometry and the low
torque values appears. Hence, both problems can be
solved with the DIN geometry, giving more reliable
data than PP in this case. Puga and Williams (2015)
investigated the influence of varying gap height and
geometry types, PP and CC, on asphalt rubber bind-
ers. The authors detected different responses when
compared the same material with higher G* values
from the DIN testing geometry compared to the ones
obtained with PP. Also, Rønholt et al. (2014) ob-
served that the elastic modulus (G’), obtained from
small amplitude oscillatory shear rheology tests, has
considerably dependence. For this tests, nine differ-
ent geometry types where used (including PP, DIN
and Vane) on the same material.
Figure 1 - Experimental design flowchart.
 the PP geometry indicate that the amount of material
sample can also influence the axial displacement as
the material response - expansion or retraction- dur-
ing the test. Hence, the axial forces were up to 10
3 RESULTS AND DISCUSSIONS times higher in CC than for PP for the same range of
temperature.
In order to better evaluate the efficiency and ac-
curacy of different testing geometries in rheological
characterization of asphalt binders, results from DSR
tests were analyzed using different graphs.
Figure 2 presents the dynamic shear modulus re- 1E+05
sults as function of temperature and difference be-
tween DIN results and others can be observed. Alt-
hough the exact cause of results divergence has not

Dynamic Shear Modulus (|G*|), Pa

1E+04
been conclusive identified, it’s important to point
that rheological response of the material depends on
many factors. One of them is the gap between the
cup inner surface and the rotor side, which can con- 1E+03
tribute for the occurrence of wall-slip effects. DIN
has 7.7 mm lateral gap and Vane has only 2.2 mm, PP 1 mm
more like PP gaps, then during DIN tests the flux of
1E+02 PP 3 mm
material inside the cup are less influenced by the
friction generated between the wall and the material DIN
inside the cup than VANE and PP. In addition, the VANE
second important factor is that each rheological in- 1E+01
strument affects differently the flow imposed to the 50 60 70 80 90
Temperature, C
material. Once physical conditions vary for each
rheological measuring system adopted, the sample Figure 2 - Dynamic Shear Modulus versus temperature.
properties may exhibit different responses. (Rønholt
et al., 2013).
The black space graph (Figure 3), that correlates
for each geometry with the average results of Dy- 1E+05
namic Shear Modulus (|G*|) and Phase Angle (δ),
shows a considerable difference among the results of T = 64 C
the different geometries. For the CC, using Vane or
Dynamic Shear Modulus (|G*|), Pa

Din geometries, the correlation between |G*| and δ 1E+04

follows the same pattern (but with distinct absolute

values), results followed the expected trend, where
the phase angle tends to 90° with the temperature in- 1E+03
crease. A different pattern was observed in the re-
sults obtained with the PP geometry (even without PP 1 mm
influence of gap position during the test). PP results
1E+02 PP 3 mm
showed an unexpected material behavior above
64°C: there was an inversion on the phase angle with DIN
the decrease in the dynamic shear modulus. VANE
Thus, the results obtained using the different ge- 1E+01
ometry cannot be assumed as effective material re- 84 85 86 87 88 89 90
sponse in a generalized way, since as already men- Phase angle,

tioned, the rheological response depends of the type Figure 3 - Black space diagram of the different geometries cor-
of the flow generated along the test, even using the relating the Dynamic Shear Modulus and phase angle.
same sample of the binder.
The first hypothesis for the conflicted results was
some error caused by the contact loss between bind- These results suggest better efficiency of CC ge-
er sample and the plates of the geometry. However, ometry (considering the neat asphalt binder evaluat-
according to the results presented in Figure 4, there ed), especially at higher temperatures when asphalt
was not an abrupt change in the axial force measured binder tends to flow. So, the confinement guaranteed
with the change of the temperature, indicating the by the cup and the shear condition applied by the
accuracy during the test. Additionally, the axial complete immersion of geometry inside the sample
force differences observed for 1mm and 3mm gap in can prevent the asphalt binder from draining at high
temperatures, assuring the effective material contact softening point. Therefore, temperatures below 50°C
with the rotors (DIN or Vane) responsible for meas- were not evaluated and they might be an issue and
uring the rheological response of material. limitation of its use for asphalt binder characteriza-
However, tests carried out using CC geometries tion in a broader range of temperatures.
need special care during sample conditioning time, In summary, the study identified that the correct
which can interfere in the homogeneity of the sam- geometry selection depends on the temperature
ple temperature. As the amount of asphalt binder range and applicability of the material, as these vari-
used in CC is around ten times greater than the PP ables yield the material to different stress state con-
samples, it requests longer conditioning times. In ditions, and consequently different rheological be-
contrast, temperature control of the sample is guar- haviors.
anteed due to the relatively large contact area at the
wall of the cup (Mezger, 2014).
5 ACKNOWLEDGMENTS

0,1 Supported by São Paulo Research Foundation

(FAPESP: 2017/25708-7; 2019/11354-4 and
2019/08415-1).
0,08

0,06 6 REFERENCES
Axial force N

Airey, G.D. 2002. Use of Black Diagrams to Identify Incon-

0,04
PP 1 mm sistencies in Rheological Data Road Materials and Pave-
PP 3 mm ment Design (n. 4) vol. 3: 403-424.
0,02
DIN
Baumgardner G. & D’Angelo J.A. 2012. Evaluation of New
Dynamic Shear Rheometer Testing Geometry for Perfor-
VANE
0 mance Testing of Crumb Rubber–Modified Binder. Journal
of the Transportation Research Board vol. 2293.
Hraiki S. 1974. Rheological properties of bituminous materials.
-0,02
50 60 70 80 90
Rheologica Acta vol. 13: 567-570.
Temperature C Hung S. et al. 2015. Comparison of Concentric Cylinder and
Figure 4 - Axial force of the different geometries. Parallel Plate Geometries for Asphalt Binder Testing with a
Dynamic Shear Rheometer. Journal of the Transportation
Research Board vol. 2505.
Mezger, T.G. (ed. 4) 2014. The Rheology Handbook. Hanover:
Vincentz Network.
4 CONCLUSIONS Phan-Thien N. & Mai-Duy N. (ed. 3) 2017 Understanding Vis-
coelasticity an Introduction to Rheology. Cham: Springer
The influence of the geometry used during DSR International Publishing.
tests and rheological characterization of asphalt Polacco G. et al. 2003. Dynamic Master Curves of Polymer
binder materials was clearly identified for a neat as- Modified Asphalt from Three Different Geometries. Ap-
phalt binder. Even though an asphalt binder of the plied Rheology vol. 13 (3): 118-124.
same origin and properties has been used for all the Puga K. (dissertation) 2015. Comprehensive Study on the Sus-
tests, rheological response of the material was con- tainable Technology of Asphalt Rubber for Hot Mix Asphalt
sidered dependent on the geometry of the test. Binders and Mixes. Ames: Iowa State University
Although PP geometry is the standard and most Rønholt S. et al. 2013. Small Deformation Rheology for Char-
commonly used worldwide in rheological character- acterization of Anhydrous Milk Fat/Rapeseed Oil Samples.
ization of asphalt materials, the results shown above Journal of Texture Studies (45): 20-29.
identified important limitations of the geometry for Tabilo-Munizaga G. & Barbosa-Cánovas G.V. 2005. Rheology
certain test conditions (clearly associated with the for the food industry. Journal of Food Engineering (67)
higher temperatures) which compromises the results 147-156.
obtained. The rheological properties, |G*| and δ, are Wood P.R. 1958. Rheology of Asphalts and Its Relation to Be-
associated with the type of flow generated during the havior of Paving Mixtures. Highway Research Board Bulle-
test. A representation of what the material will face tin (n. 192): 20-25.
during its application and service life would be the
most recommended. Two other geometries were test-
ed. DIN and Vane rotors combined with CC do not
have the same limitation (material flow out of the
geometry) identified in PP at temperatures above its