Materials and Structures (2011) 44:1809–1820 DOI 10.1617/s11527-011-9739-4

ORIGINAL ARTICLE

Rheological characterization of wax-modified asphalt binders at high service temperatures Filippo Merusi • Felice Giuliani

Received: 12 October 2010 / Accepted: 13 April 2011 / Published online: 4 May 2011 Ó RILEM 2011

Abstract The research reported herein focuses on the rheological characterization of wax-modified asphalt binders used in warm mix asphalt (WMA) technology. Wax-modified asphalt binders were produced by adding controlled quantities of different types of wax to a 50/70 unmodified bitumen. Five different kinds of wax were used, including synthetic hydrocarbons Fischer–Tropsch wax, Montan waxes and amidic-modified waxes. All the blends were subjected to different rheometric tests to assess their mechanical response at high service temperatures. The viscous deformation mechanism was analyzed with reference to static and repetitive creep loading; it was found that the viscous deformation is strongly affected by the presence of both hydrocarbons and amidic-modified waxes. Wax-modified binders exhibited an effective improvement in intrinsic resistance to non-reversible deformation, even in high applied stress and cyclic loading conditions. It was observed that the chemical composition and the consequent physical characteristics of the wax are the most important factors regulating the final behavior of wax-modified asphalt binders at high service

F. Merusi (&)  F. Giuliani Department of Civil and Environmental Engineering, University of Parma, Viale G.P. Usberti, 181/A, 43124 Parma, Italy e-mail: [email protected] F. Giuliani e-mail: [email protected]

temperatures. The final contribution of the experience performed is related to the technical evaluation of wax-modified asphalt binders and to the general development of WMA technologies for pavement applications. Keywords Warm mix asphalt  Asphalt binders  Wax  Rheology  Damage behavior  Creep

1 Introduction Production, laying and compaction of hot mix asphalts (HMA) imply high temperatures during the whole process cycle, and high temperatures require high costs and intrinsic environmental problems. Several advanced techniques, aiming at producing warm asphalt mixtures (WMA), have therefore been developed in recent years. The main goal of warm mix asphalt (WMA) technologies is to allow a reduction in process temperatures, leading to a consequent decrease in cost, energy consumption and atmospheric emissions. Moreover, in addition to the immediate environmental and economic advantages, the development of the WMA technology can induce further technological improvements: as the production temperature is the same as the one adopted for the traditional HMA, warm mixtures allow longer haul distances and longer construction season [1]. Nowadays, warm asphalt technology includes a large group of very different technical solutions,

1810

including process technology modification, waterbased technologies and bitumen fluidifying organic additives [2]. In this last group, some promising technical solutions involve procedures based on mixture modification with additives able to foam bitumen, such as zeolite, or fluidify bitumen, such as waxes. The foaming process creates a volume expansion of the bitumen, resulting in an improvement of the aggregate coating at temperatures lower than those required for HMA manufacture [1, 2]. Completely different is the case of bitumen modification with fluidifying additives. In this case the additive must show specific properties in order to guarantee lower viscosity of bitumen at mix production temperatures without affecting bitumen consistency at pavement service temperatures. Wax is suitable because of its melting/crystallizing properties. For this reason, many technical experiences have recently focused on the use of commercial wax to develop effective warm mix asphalt technologies [3– 5]. However, waxes are not new in asphalt technology because they are natural constitutive components of all petroleum products, including bitumen [6, 7]. As a consequence, wax has long been present in bitumen and asphalt technical literature. Several studies have been carried out about wax in bitumen, and more precisely about the determination of the wax content [8, 9], its crystallization properties [10, 11], chemical structure [12] but also its influence on properties of bitumen and asphalt mixtures [13–16]. As for performance, it has been largely experienced that the presence of wax in bitumen can be associated to different side effects affecting pavement quality and durability. In particular, it was traditionally hypothesized that wax melting can soften bitumen at high service temperature, affecting the rutting resistance of the pavement, and that wax crystallizing can increase mixture stiffness and its sensitivity to fatigue and thermal cracking in case of low temperatures. Hence, nowadays we have two very different technical perspectives of wax in bitumen: from a traditional point of view, wax is to be eliminated from bitumen; on the contrary, from a WMA point of view, wax needs to be added to bitumen. The key point which allows us to interpret this apparent incoherence is intrinsically inherent in the definition of wax, or, rather, the fact that wax does not have a single univocal definition.

Materials and Structures (2011) 44:1809–1820

Petroleum waxes, which give rise to the ‘traditional’ point of view, include different types of wax and also aromatics and molecules with polar functional groups which may crystallize while cooling [11]. Petroleum waxes can be divided into paraffin waxes and microcrystalline waxes [9, 12]. Paraffin wax refers to n-alkanes with linear carbon backbone chains and low molecular weight (\C45). These characteristics of its chemical structure determine to the main physical properties of paraffinic waxes: a low melting point (generally \70°C) and the typical ordinate macro-crystalline structure in the process of cooling. Microcrystalline waxes are aliphatic hydrocarbon compounds with branched chains, therefore are chemically referable to iso-alkanes and cycloalkanes. Moreover, microcrystalline petroleum waxes have a specific distribution of the molecular weight with a prevalence of high molecular weight compounds with carbon backbone chain length [C45 [7]. Hence, the physical properties of microcrystalline waxes differ from those of paraffins, and a high melting point (generally [70°C) is expected [12, 17]. Furthermore, the crystallized form is different too, and microcrystalline wax generally crystallizes as small needles [10]. On the other hand, there are waxes proposed and commercialized as bitumen flow improvers to be used in WMA: Montan waxes, Fischer–Tropsch waxes and functionalized waxes compose this group. Montan wax is a combination of non-glyceride long chain carboxylic acid esters, free long-chain organic acids, long-chain alcohols, and other organic compounds with a complex structure. Montan waxes derive from fossilized vegetables and are obtained by solvent extraction of certain types of lignite and brown coal [2]. Differently, Fischer–Tropsch waxes are produced by synthesis process [2]. The resulting synthetic wax is a mixture of n-alkanes and iso-alkanes, where n-alkanes with carbon chain lengths ranging from C40 to C120 are predominant. FT waxes have a consequent high molecular weight distribution ([1,600 g/mol) and a fine microcrystalline structure at low temperatures [2]. If compared to paraffin waxes, FT waxes have a higher melting point (generally [90°C) due to the longer carbon chains and to the reduced brittleness at low temperatures (because of the smaller crystalline structure). Functionalized waxes are chemically modified hydrocarbon microcrystalline waxes. Functionalized waxes for WMA technology

Materials and Structures (2011) 44:1809–1820

generally include waxy amides produced by amidation of fatty acids. However, other modified waxes produced through saponification or esterification of hydrocarbons waxes can also be recorded as bitumen modifiers in specific literature and patents [18]. These kinds of wax can have very different physical properties depending on their different chemical structure. In particular, functionalized waxes generally have a melting point higher than 100°C and show specific crystallization properties. At present, different types of wax are therefore designed and proposed as bitumen modifiers in WMA. This fact is of the utmost importance in pavement technology since it must be taken into account that different waxes produce different effects on bitumen rheology and asphalt performance and durability. In addition, all WMA waxes are nominally different from petroleum waxes and therefore escape from the traditional concept of wax in bitumen. The studies carried out by Edwards et al. [19, 20] confirmed this fact. The effective contribution of WMA wax on bitumen physical properties cannot be therefore easily predicted. On the contrary, due to the extreme variability in wax origin, composition and physical properties, their role in bitumen modification is rather complex to define [21, 22]. Hence, starting from the supposition that the effects of a certain wax on the physical properties of a bitumen are mainly controlled by the chemical composition of the wax, the general objective of the present experience is to outline the relations between the characteristics of the wax and the mechanical response of wax-modified asphalt binders at high service temperatures. Moreover, the development of the experimental program also refers to the analysis of damage behavior of waxmodified asphalt binders and to the identification of their effective performance-related and technical qualification.

2 Experimental program 2.1 Materials Five different waxes were used, referred to in the text as: N (Montan wax by Romonta GmbH), A and B (mixes of Montan wax and high molecular weight hydrocarbons by Romonta GmbH), L (a mix of

1811

Montan wax and fatty acids derivatives by Clariant), and W (Fisher Tropsch wax by Sasol Wax). All waxes were then mixed with a 50/70 unmodified bitumen to produce bitumen/wax blends. The bitumen, referred to in the text as B, was derived from a vacuum refinery process and contained no additive at origin. All blends were laboratory-produced by adding a controlled quantity of the five types of waxes to the same base bitumen B. Two different percentages, w/w 3.0 and 6.0%, were used for each wax. The blend preparation involved a previously standardized methodology. A bitumen sample of 300 g ± 5 g was heated in a ventilated oven at 150°C ± 5°C for 30 min and subsequently placed on a heating plate. Pre-weighted waxes were carefully added to the bitumen. The blends then were mixed for 15 min at 150°C ± 5°C. Finally, the binders obtained were divided into suitable amounts to prepare samples for all the conventional characterizations. The procedure used to prepare the samples for the rheological measurements was based on EN 12594. Wax-modified binders were treated as PMB, thus reheating and homogenization were carefully carried out at controlled temperature (160°C ± 5°C) in order to obtain reproducible results [23]. Special attention was then paid to the thermal history and to the storage conditions of the test samples (1 h at 25°C ± 0.5°C) because of their influence on rheological measurements [24]. 2.2 Research approach The experimental program consisted of two phases. The first phase focused on the linear viscoelastic behavior of asphalt binders at pavement service temperatures and aimed at evaluating the SHRP indicators for wax-modified asphalt binders. According to the results of SHRP [25], Marasteanu and Anderson [26] and Airey et al. [27], the extension of the linear viscoelastic domain was checked by performing stress sweep tests, and the limit of the LVE response was identified as the stress amplitude where the complex shear modulus |G*| had decreased by 95% from its initial figure. Linear viscoelastic limits at 40°C were established to be located in correspondence of shear stress amplitude equal to 500 Pa for all wax-modified binders.

1812

Materials and Structures (2011) 44:1809–1820

Materials behavior was then evaluated in terms of their complex shear modulus (|G*|) and phase angle (d) defined by using stress-controlled frequency sweeps from 1 to 100 rad/s. Dynamic mechanical data were subsequently analyzed to evaluate the high and the intermediate binder specification temperatures according to the SHRP criteria [25]. In this case, the methods and equations proposed by Elseifi et al. [28] were used: jG*j  sin d ¼ A  ebT

ð1Þ

jG*j= sin d ¼ C  edT

ð2Þ

where A, C, b, d are regression coefficients, and T is the temperature. The second phase of the experimental program focused on investigating the effects of wax on the damage behavior of asphalt binders at high service temperatures. With this condition, damage behavior is directly associated to the material time-dependent response and, consequently, the resistance to damage is identified by the binder’s ability to contrast the propagation of a viscous flow. Damage behavior can therefore be described by separating delayed elastic phenomena from the effective residual, and nonreversible, deformation. Hence, according to previous experiences reported in scientific literature and with specific regard to the results of Bahia et al. [29] and Polacco et al. [30], the proposed experimental approach deals with the analysis of binder viscous deformation mechanism under static and repetitive creep loading. Creep and recovery measurements were carried out at 40°C in order to outline the characteristics of the creep response in wax-modified asphalt binders. Creep and recovery testing were initially performed using 10 s for the creep phase and 30 s for the recovery phase. In this phase of the experimental program, basic rheological principles were used to analyze and generalize the results obtained through the experimental data [31]. Rheometric conditions for repeated creep tests (RCT) were outlined according to Bahia et al. [29], and several cycles of creep and recovery were carried out by applying shear stress using the DSR in stresscontrolled mode. The creep and the recovery times were equal to 1 and 9 s respectively. Each test was repeated for different applied shear stresses ranging from 10 Pa to 1 9 103 Pa.

All rheometric measurements were carried out using a dynamic shear rheometer (DSR). A parallel plate testing system was used with 25.00 ± 0.01 mm diameter plates and 2.000 ± 0.001 mm testing gap. The choice of the testing geometry was based on the general testing conditions and according to the expected stiffness of the materials used. The specimens were placed on the bottom plate, squeezed out between the two plates and trimmed off the edge of the plates using a hot spatula. Once these operations had been executed, the gap was set so as to guarantee the correct geometry of each sample. The temperature during the DSR tests was controlled by means of a Peltier conditioning system with a maximum admitted deviation of ±0.01°C. Due to the influence of the temperature on the rheological measurements, the specimens were subjected to a 30-min thermal conditioning period before each test in order to grant constant temperature during the whole experiment. With the aim of avoiding mistakes due to the instrument’s sensitivity, the torque applied was higher than the minimum suggested by the instrument producer (min. torque = 0.5 lNm). 3 Results and analysis In order to provide elements for a general basic characterization of the waxes, it is useful to analyze their Fourier Transform Infrared (FTIR) spectra and their melting range defined by the Differential Scanning Calorimetry (DSC). Figure 1 shows the FTIR spectrum of wax W. In this case the characteristic spectrum of an ordinary hydrocarbon wax was detected. The same result was obtained for waxes B and N. Figure 2 shows the FTIR spectrum of wax L. In this case the differentiating element is the presence of the characteristic peaks of amide groups at about 1,550, 1,650 and 3,300 cm-1. The same result was obtained for wax A. Table 1 reports the melting temperatures (Tm) and melting enthalpies (DHm,w) of the pure waxes obtained with the Differential Scanning Calorimetry (DSC) analyses. The two values reported for wax A correspond to the melting of two well separated peaks respectively, while the two Tm reported for wax W correspond to iso and n-paraffins (in this case, a single global DH is reported, because the two melting peaks are partially overlapping). In the case of wax B,

Materials and Structures (2011) 44:1809–1820

1813

–

Fig. 1 FTIR spectra of wax W

Fig. 2 FTIR spectra of wax L Table 1 Results of DSC analyses Wax

Tm (°C)

DHm,p (J/g)

L

146.7

124.0

A

69.0

18.5

N

145.0 84.7

107.5 127.8

B

114.7

168.3

W

98.9

214.0

111.0

the calorimetric spectrum is composed of several peaks, therefore only the temperature of the maximum peak and single global melting enthalpy are reported. FTIR spectra and DSC analyses provide the following basic characterization: –

B, N, W, hydrocarbon waxes, with high melting temperature (waxes W and B) and low melting temperature (wax N).

A, L, amide-modified hydrocarbon waxes.

The results of the preliminary characterization of the blends are listed in Table 2. The ring and ball softening point procedure was carried out according to EN 1427. Penetration was measured at 25°C according to EN 1426 and the penetration index was consequently calculated. The results of oscillatory shear analyses are summarized in Table 3, where temperatures THS (referred to the condition |G*|/sind = 1.0 kPa) and temperatures TIS (referred to the condition |G*|sind = 5.0 MPa) are presented for the correct identification of the SHRP qualification of the binders. Other parameters presented in Table 3 are the regression coefficients referred to Eqs. 1 and 2, calculated according to Elseifi et al. [28]. It can be immediately observed that all waxes produce changes in both SHRP indicators. The effect on the SHRP intermediate temperature is confined to a growth of TIS ranging from 3 to 6°C for each binder. In this case no particular effect is outlined with regard to the type and content of wax. The slope of function G*sind = f(T), represented here by coefficient b, is not particularly influenced by the specific wax and, with the exception of binder B-6L, all binders have similar values of b. This fact agrees with the consideration that the influence of waxes on the behavior of the base bitumen at the intermediate service temperatures is relatively limited. The effect on THS is very different. In particular, wax seems to have an essential role in the determination of the superior limit temperature (THS), which tends to increase by several Celsius degrees, providing an important shift in the binders’ PG grade. Moreover, the increase in THS is now strongly dependent on the type of wax. The most important increases of THS are outlined for binders containing the amidic-modified waxes (L and A). This fact is certainly due to the specific thermal characteristics of these waxes: the presence of the amidic group in the molecule shifts their melting temperatures towards rather high values. In this case, the SHRP superior temperatures are so high that similar values cannot be easily obtained even by taking into account special polymer modifications. Furthermore, in this case the THS of the binders increases according to the wax content, whilst a rather different behavior is obtained for hydrocarbons waxes (W, B and N). For binders modified with

1814

Materials and Structures (2011) 44:1809–1820

Table 2 Conventional properties of asphalt binders

Bitumen/ wax blend

Additive type (-)

Additive content (%)

Pen. @ 25°C (dmm)

Soft. point (°C)

Pen. index (IP) (-)

B

–

–

55

50.7

-0.8

B-3W

W

3.0

33

78.4

?3.0

B-6W B-3L

W L

6.0 3.0

26 31

95.5 100.5

?4.6 ?5.5

B-6L

L

6.0

24

115.9

?6.3

B-3N

N

3.0

33

57.7

-0.4

B-6N

N

6.0

29

68.1

?1.2

B-3A

A

3.0

32

103.8

?5.9

B-6A

A

6.0

28

116.8

?6.7

B-3B

B

3.0

30

85.6

?3.8

B-6B

B

6.0

26

97.0

?4.7

Table 3 SHRP grading results Binder

Intermediate limit temperature -1

Superior limit temperature

b (°C )

TIS (°C)

C (Pa)

d (°C-1)

6.81 9 107

-0.157

16.6

2.37 9 108

-0.185

66.8

B-3W

8

1.34 9 10

-0.149

22.1

2.25 9 108

-0.156

79.0

B-6W

1.72 9 108

-0.151

23.4

3.65 9 108

-0.162

79.2

8

A (Pa) B

THS (°C)

B-3L

8

1.66 9 10

-0.173

20.3

7.86 9 10

-0.083

80.2

B-6L

5.86 9 107

-0.112

21.9

2.67 9 108

-0.145

90.1

B-3N

1.65 9 108

-0.174

20.1

1.07 9 108

-0.168

69.1

B-6N

1.63 9 108

-0.156

22.3

5.57 9 108

-0.182

72.5

B-3A

1.21 9 108

-0.159

20.0

1.78 9 108

-0.157

77.1

B-6A

1.09 9 108

-0.142

21.7

1.67 9 106

-0.071

104.0

B-3B B-6B

1.05 9 108 1.44 9 108

-0.158 -0.167

19.3 20.2

2.72 9 107 8.27 9 107

-0.138 -0.125

73.9 72.1

waxes W, B and N the increases in THS are not so evident: the THS of these binders does not exceed the value of 79.2°C obtained for B-6W, and there appears no dependence on wax content. In fact, none of the three binders modified with hydrocarbon waxes changes their THS when the wax content increases from 3.0 to 6.0%. So, in this case, a sort of threshold regarding wax content can be identified, located around 3.0% by weight, which, once exceeded, provides ‘saturation’ of the bitumen with no other effect on THS. We can now confirm that the addition of a certain amount of wax proposed for WMA manufacture has a certain influence on the performance qualification of the base bitumen. However, from the synthetic analysis offered by results in Table 3, it is not

possible to assess if the presence of wax corresponds to an effective improvement in damage resistance at high temperatures. The SHRP THS is in fact originated using the dissipated energy in reversible cyclic loading (DWc) which is connected with the delay of the time-dependent response through phase angle d. Therefore, in the case of binders with high delayed elasticity, part of the energy cannot be directly associated with non-reversible deformation (damage) but should be considered as partially reversible and not associated to damage. As a consequence, the SHRP indicator does not include an actual separation between the delayed elastic and the viscous components of the mechanical response. In other words, the recorded increase in SHRP THS occurred for wax modification could be actually

Materials and Structures (2011) 44:1809–1820

1815

related to an increase in resistance to damage at high service temperatures or, on the contrary, it could be merely an apparent effect, which may not be effectively connected to a real enhancement in binder damage resistance. An examination of the results of the creep tests can clarify this key point. The separation between delayed elasticity and damage can be effectively identified with the determination of the creep function J(t). The binder response to a single loading cycle at a constant temperature is here initially considered because, with this specific loading condition, and supposing that s0 ? 0 and t ? 0, it can be assumed that the deformation evolves in the linear viscoelastic domain; the canonical form of the creep compliance written for a viscoelastic liquid can be therefore used to describe the evolution of the strain c(t) associated to the constant stress s0: cðtÞ ¼ Je  r0 þ Jde  wðtÞ  r0 þ

t  r0 g0

Fig. 3 Results of creep test at 40°C and s0 = 10 Pa (Base bitumen B)

ð3Þ

where Je is the instantaneous elastic (or glassy) compliance, Jde is the delayed elastic compliance, w(t) is the creep function and g0 is the zero-shear (or Newtonian) viscosity [32]. Figure 3 illustrates the deformation mechanism of the base bitumen B, whilst Fig. 4 shows the trend obtained for binder B-3L. Both experimental data and fitting curves are reported herein. For the sake of brevity, creep and recovery results of other modified binders are not reported here in graphical form. However, the trend reported for B-3L can be considered as representative of all other wax-modified binders because no formal difference occurs among the trends obtained for other binders. It can be immediately observed that substantial changes occur in the creep response of the base bitumen after its modification with wax. As expected, base bitumen B exhibits the typical behavior of a pure bitumen in its Newtonian domain. The deformation grows with a linear trend during the entire creep interval and the experimental data show no relevant recovery. Actually, the maximum deformation (cmax in Fig. 3) and the residual (viscous) deformation (cr in Fig. 3) substantially assume the same value. This fact indicates that almost all the deformation energy is dissipated in internal friction, or is spent to deform the liquid matrix of the bitumen,

Fig. 4 Results of creep test at 40°C and s0 = 10 Pa (Binder B-3L)

and mainly produces a non-reversible flow. The analysis of the deformation mechanism performed with Eq. 3 clearly confirms this assumption: with reasonable approximation, the very reduced elasticity recorded in this case can be considered as referable only to an instantaneous component; hence, if no elastic contribution is developed during the flow, the creep function is w(t) = 0 for t [ 0 and according to Polacco et al. [30], J(t, T) reduces to its viscous approximation (Eq. 4). Jðt; TÞ ¼ Je0 ðTÞ þ

t g0 ðTÞ

ð4Þ

Substantially, in this case we have ordinary resistance to non-reversible deformation expressed by the viscosity coefficient g0 of the unique dashpot of the model, which also represents the zero-shear viscosity of the system. The situation is very different for wax-modified binders: it is clear that the mechanical response now includes the existence of a well-structured delayed elastic component which interests the whole temporal

1816

Materials and Structures (2011) 44:1809–1820

extension of the flow (Fig. 4). So, in this case the response to a step stress, essentially related to the material constitutive equations, is characterized by the existence of w(t) = 0 for t [ 0: Z þ1   t wðtÞ ¼ LðkÞ  1  ek  dk ð5Þ 0

where L(k) is the continuous retardation spectrum and k is the retardation time. The complete form of the creep function, written with reference to a discrete form of the generalized model including three Kelvin-Voight elements (coefficients Gi, ki with i = 1, 2, 3), a single spring (coefficient G0) and a single dashpot (coefficient g0), must be therefore used to describe the creep behavior of B-3L, otherwise a large approximation could unavoidably affect the modeling. cðtÞ ¼

3  t s0 X 1  t þ  1  e  ki  s 0 þ  s 0 g0 G0 i¼1 Gi

ð6Þ

This change in the analytical form of J(t, T) has some critical consequences. First of all, the enhanced resistance to non-reversible deformation is now certainly expected just because of the need to add the delayed elastic term to the analytical expression of J(t, T). In fact, from a physical point of view the existence of w(t) = 0 means that the deformation energy can be partially stored by the material during the flow. Intrinsic consequences of such ability are the possibility of recovering deformation as the load is removed but also the reduction of the amount of Table 4 Results of creep and recovery tests (T = 40°C, s0 = 10 Pa)

Binder

B

energy available for the non-reversible flow. Reflections of this change in the composition of the mechanical response, and in particular in the balance between the elastic and viscous contributions, appear in the macroscopic behavior of the binders and can be evidently detected by analyzing the value of cmax, or the strain measured at the end of the creep phase (t = 10 s) and cr, or the residual non-reversible strain recorded at the end of the recovery phase (t = 30 s). From a rheological point of view, a rather exhaustive interpretation of binder behavior after wax modification can be directly related to the eight constants of the model (Eq. 6) and, in particular, to the value of g0 that univocally identifies the total binder resistance to non-reversible deformation. In fact, according to the structure of the rheological model (Eq. 6), the integrity of the material resistance to non-reversible flow is still described by the parameter g0, now represented by the inverse of the first derivative (slope) of the function J(t) evaluated after the exhaustion of the delayed elastic phenomena (Merusi and Giuliani 2011). For this reason, values of g0 extrapolated from the modeling of the creep phase are reported in Table 4 for all binders. Table 4 includes g0 evaluated from the recovery phase as shown by Eq. 7, developed by supposing the consistency of the model in the recovery phase: t ! 1 ) c r ! c v ; g0 ¼

s0  tcreep cr

ð7Þ

where s0 and tcreep are the stress and the creep time respectively, while cr is the strain at the end of the g0

Deformation (experimental data) cmax (t = 10 s) (%)

cr (t = 30 s) (%)

cr/cmax (-)

Model fitting (Pa s)

Recovery phase (Pa s)

1.18

1.16

0.98

8.59 9 103

8.60 9 103

4

B-3W

0.17

0.11

0.65

7.09 9 10

9.01 9 104

B-3B

0.22

0.13

0.59

8.95 9 104

7.85 9 104

0.85

4

1.51 9 104

4

1.14 9 105

4

1.01 9 105

5

B-3N B-3A B-3L

0.78 0.21 0.24

0.66 0.09 0.10

0.41 0.42

1.37 9 10 8.79 9 10 8.52 9 10

B-6W

0.04

0.02

0.57

3.37 9 10

4.51 9 105

B-6B

0.07

0.02

0.37

3.17 9 105

4.07 9 105

0.59

4

6.01 9 104

6

1.51 9 106

6

1.86 9 106

B-6N B-6A B-6L

0.29 0.03 0.04

0.17 0.01 0.01

0.24 0.15

5.67 9 10 1.10 9 10 1.13 9 10

Materials and Structures (2011) 44:1809–1820

recovery phase, and cv is the residual viscous component of the strain in the model (cv = Jv 9 s0). In addition, Table 4 includes the ratio of the residual non-reversible strain cr to the total strain cmax recorded at the end of the creep phase. The ratio cr/cmax is considered here as an indicator of the total elasticity (instantaneous and delayed) of the material. Results in Table 4 provide elements for some general considerations. The first part of Table 4 includes base bitumen B and the binders modified with 3.0% of wax content. For those binders, g0 derived from the model agrees with g0 derived from the recovery phase. So, we are able to identify a univocal value of g0 without exceeding in approximation and a well-defined ranking of the binders can be outlined on the basis of g0 only. The enhancement in resistance to nonreversible deformation described above from an analytical point of view is therefore correctly quantified by the value of g0. In fact, g0 increases from 8.6 9 103 Pa s registered for B to the higher values recorded for wax-modified binders. At the same time, the ratio cr/cmax appears to be strongly reduced for all wax-modified binders. This fact indicates that the actual non-reversible deformation recorded at the end of the recovery phase covers a reduced part of the deformation mechanism. In the presence of wax, the delayed component of the mechanical response becomes prevalently elastic. It is interesting to observe that, in case of 3.0% wax content, there is no particular difference between the performances of the different wax-modified binders. With the only exception of B-3N, which does not provide a remarkable increase in g0, the remaining binders exhibit similar values of g0, ranging from 7 9 104 Pa s to 1 9 105 Pa s, supported by very similar values of the ratio cr/cmax. The analysis offered by the data reported in Table 4 becomes more interesting when the binders modified with the 6.0% of wax content are taken into consideration. In this case, the analysis of the effect of the type of wax still cannot be outlined by the general consideration reported above. On the contrary, it appears that each binder has its own specific performance, evidently regulated by the wax type. Consequently, we can now clearly distinguish binders modified by hydrocarbon waxes (W, N and B) and binders modified by amidic waxes (A and L). With more detail it is possible to outline the existence of

1817

three groups of binders defined with reference to their performances. The first group is represented by the B-6N alone. In this case g0 is still identified without approximation and it is equal to 6 9 104 Pa s and cr/cmax = 0.59. Substantially, 6.0% of wax N leads to the same level of resistance offered by 3.0% of other hydrocarbon waxes. A second group is identified by B-6W and B-6B; in this case g0 is [3 9 105 Pa s. The third group includes B-6A and B-6L, which show the highest g0 ([1 9 106 Pa s) and the lowest cr/cmax. This subdivision suggests some indications to address a more general discussion. From a first, very simplified analysis, the enhancement in binder resistance can be explained by considering that all waxes are below their melting transition when tested at 40°C. This means that, during the test, each binder contains a certain amount of crystals. The stiffening contribution (reduction in cmax) can therefore be explained simply by considering that wax crystals dispersed in the continuous bituminous matrix certainly provide the reduction of the material mobility with shear conditions. Differently, the change in elastic components connected with the change in the analytical form of J(t, T) remains of rather difficult definition. However, according to the observations of Lu et al. [10] and Giuliani and Merusi [21], it can be hypothesized that wax crystals are not dispersed in the bituminous matrix but they can associate forming a temporary network. Under this condition, when stress r0 ? 0 is applied, the bituminous matrix remains constrained by the presence of the crystal network and only a little part of the deformation energy is lost during a flow, whilst most of the energy can be stored in the crystal network and be subsequently released after removal of the stress. So, for a given category of waxes the resistance to non-reversible deformation of the binder seems to be directly correlated with the degree of residual crystallinity in the binder. The term ‘‘degree of residual cristallinity’’ is clearly used here to identify not only the concentration of crystals, but their nature too (i.e. chemical composition, dimension, shape, etc.). So, considering the three hydrocarbons waxes, we observe that wax N is the one with the lowest melting temperature and the lowest melting enthalpy (Table 1). This means that wax N has the shorter carbon chains among the three waxes considered and, consequently, it forms few macro-crystalline structures which cannot remain well connected in the

1818

bitumen but can, on the contrary, work as a filler. Probably for this reason binders B-3N and B-6N show the relative minor enhancement in resistance to non-reversible flow (minor g0). Wax B and W have very similar melting temperatures. It is therefore possible to assign the same residual cristallinity to both waxes and in fact binders modified with wax B and wax W show similar g0 and similar cr. On the other hand, there are the two amidic waxes. In this case, the high Tm cannot be directly linked to the length of the carbon chains. However, A and L have the highest Tm. So, it can be supposed that in this case the chemical origin of the wax is the most important factor and, consequently, that the crystallization of the amidic compounds could lead to a very fine and capillary structure. Figure 5 summarizes this concept from a qualitative point of view. Based on previous suppositions, all resistance to non-reversible deformation in wax-modified binders is associated to the presence of crystals. With this assumption, it immediately emerges that higher intensity of loading, or the presence of cyclic repetitions of loading, can perturb the continuity of the crystal structures, leading to a consequent decay in the binder ability of contrasting deformation. For a limit condition characterized by the repetition of loading at a very high stress level, the enhancement in performance could be therefore almost totally lost. Finally, it is interesting to analyze the mechanical response of wax-modified binders under conditions of repetitive loading in order to determine how the presence of cyclic shear stress can influence the damage resistance of wax-modified binders. Results of the repeated creep tests (RCT), performed

Materials and Structures (2011) 44:1809–1820

according to Bahia et al. [29], are reported for binders B-6N (Fig. 6) and B-6A (Fig. 7) in terms of evolution of the shear deformation c(t). Contrary to expectations, the trends shown in Figs. 6 and 7 reveal that the presence of cyclic shear loading does not provide any particular effect on the mechanical behavior of the wax-modified binders. The mechanical responses of the two modified binders are still characterized by a prevalent delayed elasticity, and in both cases a large part of the deformation is recovered at the end of each loading cycle. Moreover, the difference between the performances of the two binders seems to be more highlighted in presence of repetitive loading. In fact, the binder modified with the amidic wax shows a tendency to harden according to the number of loading cycles. For this reason, immediately after the exhaustion of the first few cycles, the rate of strain accumulation has a remarkable decrease and seems to identify the existence of a sub-horizontal asymptote. This phenomenon, typical of some elastomer-modified binders, can be related to a reorganization of the internal structure, and it is an indicator of a further

Fig. 6 Results of repeated creep test (B-6N, s0 = 1 9 103 Pa)

Fig. 5 Relation between the melting temperature of the pure waxes and g0 of the binders

Fig. 7 Results of repeated creep test (B-6A, s0 = 1 9 103 Pa)

Materials and Structures (2011) 44:1809–1820 Table 5 Results of repeated creep tests Binder

Accumulated strain cacc (%) s0 = 10 Pa

s0 = 1 9 103 Pa

6.27 9 10-1

5.61 9 101

B-3W

-2

1.10 9 10

4.51 9 100

B-3B

2.12 9 10-2

5.84 9 100

B-3N

-1

4.23 9 101

-2

5.72 9 100

-3

B

B-3A

4.06 9 10 1.77 9 10

B-3L

7.27 9 10

5.49 9 100

B-6W

–

2.62 9 10-1

B-6B

3.03 9 10-3

1.26 9 100

B-6N

-2

7.84 9 100

-3

1.64 9 10-1

-4

4.44 9 10-1

B-6A B-6L

6.48 9 10 8.02 9 10 5.65 9 10

resistance to flow. The absence of this peculiar effect in behavior of B-6N (Fig. 6), can still be related to the difference between the residual crystallinity of the two types of wax and, in particular, to absence in B-6N of a well-organized structure composed by many fine residual crystals. The accumulated strain (cacc) listed in Table 5 for all binders confirms the relations between the characteristics of the wax and the resistance to deformation of wax-modified asphalt binders. Moreover, the data recorded for two different stress levels (Table 5) seem to confirm that the amount of stress has a certain effect on the recorded resistance. In fact, a non-linear relationship between s0 and cacc can be supposed for some wax-modified binders and for these same binders the resistance is reduced as the stress increases. However, great differences persist between the accumulated strains of the different binders when high stress levels are considered. Therefore, the data in Table 5 seem to confirm that the reduction of non-reversible deformation recorded for wax-modified binders is due to the existence of an actual intrinsic resistance and can be almost partially preserved even for cyclic loadings at high stress amplitude.

4 Conclusions The mechanical behavior of wax-modified asphalt binders was studied at high service temperatures using different rheological tests, with specific attention paid

1819

to the analysis of binders’ damage behavior and resistance to non-reversible deformation. Rheological experiments performed in the linear viscoelastic domain highlighted that waxes have a strong influence on several aspects connected to the performance qualifications of asphalt binders. In particular, it was found that wax modification has an essential role in the determination of the SHRP superior limit temperature (THS), which tends to increase according to the wax content and type, providing an important shift in binders’ PG grade. The results of the creep and recovery tests have shown that the increase in THS is related to an effective change in binder rheological behavior. In fact, the mechanical response to a constant step stress, essentially related with the material constitutive modeling, changes after wax addition. In case of wax modification, the presence of a well-structured delayed elastic component which interests the whole temporal extension of the flow must be analytically described by the addition of the delayed elastic term in the compliance function of the base bitumen. Moreover, the analysis of the deformation mechanism (carried out with reference to the basic rheological models) has led to the definition of a strong enhancement in the resistance to non-reversible deformation, represented by the model parameter g0 and by the ratio cr/cmax. Such particular enhancement was then recorded for different loading conditions characterized by high stress amplitudes with cyclic repetitions. However, the physical characteristics of the wax were found to be of decisive importance in controlling the binder’s final behavior. In other words, we must assume that each wax, probably in consequence of its residual crystallinity characteristics, has its own specific influence on the binder’s behavior, and its effect cannot and must not be confused with the effect of another wax. Among the different waxes used, amidic-modified waxes always provided the most important and interesting contribution, whilst different levels of resistance were found for hydrocarbon waxes and they were mainly connected with the wax melting temperature. The final contribution of the experience is related to the possible identification of criteria and guidelines to address the technical selection of wax in the warm mix asphalt context, also with regard to the required performances and the issue of climatic implications.

1820

References 1. Hurley GC, Prowell BD (2006) Evaluation of potential process for use in warm mix asphalt. J Assoc Asphalt Paving Technol 75:42–90 2. D’Angelo J et al (2008) Warm-mix asphalt: European practice Report No FHWA-PL-08-007 February 2008 3. Diefenderfer SD, McGhee KK, Donaldson BM (2007) Installation of warm mix asphalt projects in Virginia. Report FHWA/VTRC 07-R25 Virginia Transportation Research Council 4. Hurley GC, Prowell BD (2005) Evaluation of SasobitÒ for use in warm mix asphalt NCAT. Report 05-06 National Center for Asphalt Technology (NCAT) Auburn University 5. CEDEX (Centro de Estudio y Experimentacion de Obras Publicas) (2004) Influencia de la utilizacion de betunes asfalticos modificados con Sasobit en la mezclas bituminosas. Technical report Madrid Spain 6. Warth AB (1956) The chemistry and technology of waxes. Reinhold, New York 7. Thanh NX, Hsieh M, Philp RP (1999) Waxes and asphaltenes in crude oils. Org Geochem 30:119–132 8. Lu X, Kalman B, Redelius PA (2007) New test method for determination of wax content in crude oils residues and bitumen. Fuel 87:1543–1551 9. Edwards Y, Isacsson U (2005) Wax in bitumen Part 1— classification and general aspects. Road Mater Pavement Des 6:281–309 10. Lu X, Langton M, Olofsson P, Redelius P (2005) Wax morphology in bitumen. J Mater Sci 40:1893–1900 11. Redelius P, Lu X, Isacsson U (2002) Non-classical wax in bitumen. Road Mater Pavement Des 3:7–21 12. Lu X, Redelius P (2006) Compositional and structural characterization of waxes isolated from bitumens. Energy Fuels 20:653–660 13. Chmbrion P, Bertau R, Ehrburger P (1996) Characterization of bitumen by differential scanning calorimetry. Fuel 75:144–148 14. Edwards Y, Isacsson U (2005) Wax in bitumen Part 2— characterization and effects. Road Mater Pavement Des 6:439–468 15. Edwards Y, Redelius P (2003) Rheological effects of waxes in bitumen. Energy Fuels 17:511–520 16. Ho SM, Klesken B, Zanzotto L (2003) Direct tension test—a useful tool to study the low temperature properties of wax-containing asphalt. In: Proceedings of the 82nd annual meeting of transportation research board, January 12–16, Paper No 03-3157 17. Petersson M, Gustafson I, Stading M (2008) Comparison of microstructural and physical properties of two petroleum waxes. J Mater Sci 43:1869–1879 18. Cowley LG, Fisher RG (2002) Use of a bitumen/wax composition. International Patent International Publication Number: WO 02/16499 A1

Materials and Structures (2011) 44:1809–1820 19. Edwards Y, Tasdemir Y, Isacsson U (2006) Rheological effects of commercial waxes and polyphosphoric acid in bitumen 160/220—low temperature performance. Fuel 85:989–997 20. Edwards Y, Tasdemir Y, Isacsson U (2007) Rheological effects of commercial waxes and polyphosphoric acid in bitumen 160/220—high and medium temperature performance. Construct Build Mater 21:1899–1908 21. Giuliani F, Merusi F (2009) Flow characteristics and viscosity function in asphalt binders modified by wax. Int J Pavement Res Technol 2:51–60 22. Merusi F, Caruso A, Roncella R, Giuliani F (2010) Moisture susceptibility and stripping resistance of asphalt mixtures modified with different synthetic wax. Transp Res Rec J Transp Res Board 2180:110–120 23. Anderson DA, Marasteanu MO, Mahoney JM, Stephens JE (2000) Factors affecting the variability in the SHRP binders tests. Proceedings of the 79th Annual Meeting of the Transportation Research Board January 2000 Paper No 00-1360 24. Soenen H, De Visscher J, Vanelstraete A, Redelius P (2006) Influence of thermal history on rheological properties of various bitumen. Rheol Acta 45:729–739 25. Anderson DA, Bahia HU, Cristensen DW (1994) Binder characterization and evaluation - Volume 3: physical properties. SHRP-A-369 Strategic Highways Research Program National Research Council Washington DC 26. Marasteanu MO, Anderson DA (2000) Establishing linear viscoelastic conditions for asphalt binders. Proceedings of the 79th Annual Meeting of the Transportation Research Board January 2000 Paper No 00-1374 27. Airey GD, Rahimzadeh B, Collop AC (2003) Viscoelastic linearity limits for bituminous materials. Mater Struct 36:643–647 28. Elseifi MA, Flintsch GW, Al-Qadi IL (2003) Quantitative effect of elastomeric modification on binder performance at intermediate and high temperatures. Journal of Materials in Civil Engineering 15:32–40 29. Bahia HU, Hanson DI, Zeng M, Zhai H, Khatri MA, Anderson RM (2001) Characterization of modified binders in Superpave mix design. NCHRP project 9-10 Transportation Research Board – National Research Council Washington DC 30. Polacco G, Stastna J, Zanzotto L (2008) Accumulated strain in polymer-modified asphalts. Rheol Acta 47: 491–498 31. Barnes HA, Hutton JF, Walters FRS (1989) An introduction to rheology. Elsevier Science Pubblisher Amsterdam, The Netherlands 32. Malkin AYa (1994) Rheology fundamentals. ChemTec Publishing, Toronto, Canada