Journal of Microscopy, Vol. 254, Issue 2 2014, pp. 95–107

doi: 10.1111/jmi.12122

Received 12 June 2013; accepted 15 February 2014

Microscale investigation of thin film surface ageing of bitumen P.K. DAS, N. KRINGOS & B. BIRGISSON Division of Highway and Railway Engineering, KTH Royal Institute of Technology, Stockholm, Sweden

Key words. Atomic force microscopy, bitumen, DSC, oxidation, surface ageing, UV radiation.

Summary This paper investigates the mechanism of bitumen surface ageing, which was validated utilizing the atomic force microscopy and the differential scanning calorimetry. To validate the surface ageing, three different types of bitumen with different natural wax content were conditioned in four different modes: both ultraviolet and air, only ultraviolet, only air and without any exposure, for 15 and 30 days. From the atomic force microscopy investigation after 15 and 30 days of conditioning period, it was found that regardless the bitumen type, the percentage of microstructure on the surface reduced with the degree of exposure and time. Comparing all the four different exposures, it was observed that ultraviolet radiation caused more surface ageing than the oxidation. It was also found that the combined effect was not simply a summation or multiplication of the individual effects. The differential scanning calorimetry investigation showed that the amount of crystalline fractions in bitumen remain constant even after the systematic conditioning. Interestingly, during the cooling cycle, crystallization of wax molecules started earlier for the exposed specimens than the without exposed one. The analysis of the obtained results indicated that the ageing created a thin film upon the exposed surface, which acts as a barrier and creates difficulty for the wax induced microstructures to float up at the surface. From the differential scanning calorimetry analysis, it can be concluded that the ageing product induced impurities in the bitumen matrix, which acts as a promoter in the crystallization process.

Introduction and motivation Bitumen is an organic material that is widely used in the top layer of flexible pavements. Even though bitumen is only one component that makes up asphalt mixtures, it is the component that gives the material its desired viscoelastic nature. In addition to increased driving comfort and flexible maintenance, this viscoelastic behaviour plays a prominent role in Correspondence to: Prabir Kumar Das, Division of Highway and Railway Engineering, Room N233, KTH Royal Institute of Technology, Brinellv¨agen 23, 10044 Stockholm, Sweden. Tel: +46 8 790 8664; e-mail: [email protected]

 C 2014 The Authors C 2014 Royal Microscopical Society Journal of Microscopy 

many aspects of mixture durability, such as resistance towards thermally or traffic induced cracking and ravelling. Being an organic substance, with time its viscoelastic property can deteriorate due to exposure to the environment, which is known as age hardening. This ageing of bitumen in the asphalt mixture is considered to be mainly caused by the oxidation and ultraviolet (UV) radiation, where in this process pavement temperature plays the role of catalyst. Oxidative surface ageing is an irreversible chemical reaction between hydrocarbons of bitumen and available atmospheric oxygen. The UV radiation catalysed reaction occurs rapidly, which takes place within the top layer of exposed bitumen surface (Branthaver et al., 1993; Petersen, 1993). During both of the ageing process, carbonyl, sulfoxide and benzylic carbon groups are formed which increase the polarity of the host compounds and make them much more likely to associate with other polar compounds. As they form these associations, they create less soluble (in nalkane medium) hydrocarbons that increase the bitumen’s viscosity and elastic stiffness (Branthaver et al., 1993; Petersen, 1993; Durrieu et al., 2007). A long-term asphalt-ageing test is meant to accurately predict the changing behaviour of the material due to chemical and physical property changes in the bitumen, which occur after certain years in the asphalt pavement. To simulate longterm field aging in laboratory, different types of test methods are currently available (Airey, 2003), focusing on accelerated ageing on bitumen, loose asphalt mixture, or on compacted asphalt specimen. These accelerated ageing tests, by assuming time-temperature superposition principle, are generally conducted at artificially severe conditions, for example at temperatures higher than pavement service temperature and at pressures higher than ambient pressure. Moreover, the influence of sunlight on bitumen ageing is ignored in these laboratory simulations though bitumen is known to be a good light absorber – particularly for UV radiation (Dickinson et al., 1958; Durrieu et al., 2007; Lu et al., 2008). It has been found that aging of bitumen at a higher temperature and pressure is fundamentally different from aging at a lower temperature and pressure is more representative of the physical process when kept at ambient conditions (Branthaver et al., 1993; Al-Azri et al., 2006; Lu et al., 2008). So, an investigation on

96

P.K. DAS ET AL.

accelerated test residue bitumen might not give the true information of what happens due to ageing in the field and could lead to erroneous conclusions and subsequent choices. On the other hand, one cannot investigate the material for the actual time span, so some extrapolation will need to be made. For this reason, it is important to be as close as possible to the physical process and develop a fundamental understanding of its mechanism. To understand the actual changes of bitumen properties due to the slow ageing process, in this research a controlled test environment is developed in which both the oxidation and the UV radiation are taken into account. After the conditioning procedures, the samples need to be characterized for their changes due to the ageing processes. Given recent advances, the atomic force microscopy (AFM) has been employed in this study. AFM is an imaging tool that can deliver the surface topography, stiffness and tackiness information of materials (Bellitto, 2012). It is capable of measuring surfaces at atomic and molecular resolutions as compared to the resolution limit of optical microscopy of about 200 nm. Over the past 10 years, this technique also has been applied to bitumen to investigate the topography and micromechanical properties of bituminous surfaces (Lesueur et al., 1996; Loeber et al., 1996; J¨ager et al., 2004; Masson et al., 2006; de Moraes et al., 2010; Schmets et al., 2010; Pauli et al., 2011; Allen et al., 2012; Fischer et al., 2012; Das et al., 2013). From these, it was established that despite its seemingly homogeneous character, bitumen is in fact heterogeneous at the micro- to nanometre length scale (Lesueur et al., 1996; Schmets et al., 2010; Das et al., 2013). One specific focus point of several of these studies has been the appearance of structures, named ‘bees’ (given their black and yellow stripes under the AFM). The earliest studies mainly explained the appearance of these structures on the ‘asphaltene’ phase in bitumen (i.e. the phase that is not soluble in n-heptane and is often referred to as the more heavy and polar molecules in the mixture; Loeber et al., 1996; J¨ager et al., 2004). Recent studies have, however, conclusively shown that the bee structures are related to the presence of crystalline fractions in the material, often referred to as the waxes (de Moraes et al., 2010; Schmets et al., 2010; Pauli et al., 2011; Das et al., 2013; Soenen et al., 2013). Obviously, the presence of such microstructures will dictate, if not heavily influence, the macroscale properties of bitumen, such as stiffness, viscoelasticity, plasticity, adhesion, fracture and healing characteristics. From the many AFM studies on bitumen types of various chemical origins, it has also become clear that the bitumen microstructures and their kinetic properties are not uniform among bitumen types and are strongly influenced by both chemistry and mechanical history of the material (Fischer et al., 2012; Das et al., 2013; Soenen et al., 2013). Having a better understanding of how these microstructures are evolved with ageing and related to the resulting mechanical response thus opens a large potential for tailoring and enhancing the long-term properties of asphalts.

Since AFM is only applicable for the surface investigation, to study the surface ageing effect on bulk crystalline molecule properties, differential scanning calorimetry (DSC) has also been employed. In this paper, a hypothesis regarding the mechanism of ageing has been developed and validated utilizing the AFM and the DSC, where the evolutions of microstructures on the bitumen surface due to the oxidation and/or UV radiation were studied.

Thin film surface ageing If bitumen is exposed to air or sunlight, the combination of oxidation and UV radiation may create a thin film of less soluble hydrocarbon molecules upon the exposed surface. This thin film may act as a semipermeable barrier, which creates difficulty for the wax-induced microstructures to float up at the surface. Since the film may become thicker or denser with time and increasing degree of exposure, thus the percentage of microstructure seen at the surface should decrease as shown in Figure 1. The micromechanical properties (such as relative stiffness or adhesion) at the surface should change due to surface ageing, which in its turn, should result in an increase of stiffness and reduce tackiness. In this study, AFM is utilized to provide information of changes of the microstructures and the micromechanical properties at the surface due to oxidation and/or UV radiation. A comparison between aged and unaged specimen utilizing the DSC is used to detect any changes in behaviour and percentage of crystalline fractions in the bulk.

Materials, experimental plan and methods Materials Based on the presence of natural wax in bitumen, three different types of 70/100 penetration grade bitumen from different sources were selected, with designations Bit-A, Bit-B and BitC. Table 1 lists the physical material properties according to European standards (EN), provided by the supplier. From Table 1 can be seen that Bit-A does not contain much waxes and Bit-B and Bit-C contain natural wax. Depending on the method, the measured wax content in bitumen can be quite variable. It is, therefore, important to note which method was used, when reporting the wax content. In this study, the DSC method has been utilized to investigate the microcrystalline fractions in bitumen and while calculation a constant melting enthalpy of 121 J g–1 was used a reference (Lu & Redelius, 2006). Specimen preparation. Given the fact that surface ageing is the main objective of this study, sample preparation has to be given extreme consideration to avoid any bias. Using  C 2014 The Authors C 2014 Royal Microscopical Society, 254, 95–107 Journal of Microscopy 

MICROSCALE INVESTIGATION OF BITUMEN SURFACE AGEING

97

Fig. 1. Hypothesis on microstructures evolution due to thin film surface ageing. Table 1. Bitumen physical properties.

Specification Bit-A Bit-B Bit-C

Pen EN1426 (dmm)

Soft. point EN1427 (°C)

Dyn. viscosity @ 60°C EN12596 (Pa s)

91 86 82

46.2 46.4 45.8

168 96 94

Kin. viscosity @ 135°C EN12595 (mm2 s–1 )

Fraass breaking point (°C) EN12593 (°C)

Wax content EN12606 (%)

DSC wax content

TLC-FID wax content

a* (%)

b* (%)

342 181 124

–15 –16 –7

0.3 1.7 ND

0 6.2 5.4

0 1.9 ND

ND, not determined; a*, ASTM D 3418-12; b*, Lu & Redelius (2006).

carefully placed on a rectangular aluminium plate (10 mm × 10 mm × 1 mm) and kept for 5 min on a hot plate at the same temperature to allow the bitumen to spread out and create a smooth surface. In the preparation process, the specimens were left horizontally and covered to prevent dust pick up. During the preparation specimens’ surfaces were exposed to the air. Since all specimens went through the same preparation process, it was assumed that oxidation during specimen preparation can be ignored or is at least constant amongst all samples. Experimental scheme Fig. 2. Experimental scheme.

material from the top of a can is not recommended as more aging occurs at the surface. Therefore the bitumen was first mixed properly by stirring at 120°C to avoid such biases. For AFM scanning, approximately 30 mg of hot liquid bitumen was  C 2014 The Authors C 2014 Royal Microscopical Society, 254, 95–107 Journal of Microscopy 

Right after the specimen preparation, specimens were exposed to the following four different levels of conditions: both UV and air, only UV but no air, only air but no UV and neither UV nor air, for 15 and 30 days, in a temperature (20°C) controlled environment. One specimen from each type of bitumen was kept unconditioned to give a reference measure.

98

P.K. DAS ET AL.

and micromechanical properties. The 15 and 30 days conditioned specimens were investigated by using AFM and the percentage of observed microstructures at the surface compared to investigate the ageing propensity. Thereafter, the 30 days conditioned specimens were investigated further, utilizing DSC and compared to study any change of the crystalline fractions due to the conditioning. More details on AFM and DSC will be given in the following sections. The schematic of the experimental plan is shown in Figure 2.

Atomic force microscopy analyses

Fig. 3. The force curve evaluation procedure (PeakForce QNM – Bruker).

Oxidative ageing conditioning. Since argon is one of the noble gas and heavier than the ambient air, it was used to create an oxygen free (no air) environment. To create oxygen free environments, specimens were placed in glass containers in which argon gas was introduced via an inlet. In this process, argon gas was allowed to spread in the container until the air was pushed out completely leaving for an airtight environment. UV conditioning. The UV spectrum that reaches the earth mainly contains UV-A type radiation since the other ranges of the spectrum almost entirely get absorbed by the ozone layer and the atmosphere. In this study, commercial UV light has been used to resemble the UV-A radiation. For the nonUV conditioning, specimens were placed inside containers, carefully wrapped in multiple layers of aluminium foil and kept in a dark place. After 24 h, the unconditioned bitumen specimens were scanned with the AFM to investigate the topography feature

The AFM force mapping was performed on a Bruker Multimode 8 AFM in PeakForce QNM (Quantitative Nanomechanical property Mapping) mode with a Nanoscope V controller and the software Nanoscope 8.15. The resulting force curves were evaluated using Nanoscope Analysis 1.30. The used cantilever model was ScanAsyst Air (Bruker) with a nominal spring constant of 0.4 N m–1 and a nominal tip radius of 2 nm. Force mapping was performed as follows; the sample was oscillated in the z-direction at 2 kHz at the same time as the sample was moved in the x–y direction at a rate of 0.5 kHz. In each contact between the cantilever and the substrate, a force curve was obtained and evaluated online according to Figure 3. During the calculation, the poison ratio was assumed as 0.5. For this study topography, elastic modulus and adhesive force values were evaluated over 20 μm × 20 μm scans at 38% ambient humidity and 20°C temperature. Depending on the type, source, time–temperature history of bitumen, microstructures may be observed (cf. Fig. 4A) on the bitumen surface from the AFM scans (Das et al., 2013; Soenen et al., 2013). To quantify these microstructures the open source Gwyddion 2.3 software was utilized. For this, at first the AFM topography image was analysed using a Laplacian of Gaussian edge detection technique (Marr & Hildreth, 1980) as shown in Figure 4(B). Once the microstructures were marked then the marked grains were extracted (cf. Fig. 4C)

Fig. 4. Quantification of microstructures.  C 2014 The Authors C 2014 Royal Microscopical Society, 254, 95–107 Journal of Microscopy 

MICROSCALE INVESTIGATION OF BITUMEN SURFACE AGEING

99

Fig. 5. Visual inspection of bitumen B after 30 days of exposures.

and thereafter, used for pixel area quantification to calculate the percentage of microstructure present in the frame. For each specimen, AFM scans were conducted at three different locations and the average percentage from these was used as representative average of the percentage of microstructure present at that specimen surface. The corresponding precisions for bitumen A, B and C were found as 0.4%, 0.8% and 1.0%, respectively.

Differential scanning calorimetry analyses To investigate any change in the crystalline fractions of specimens due to the previously described conditioning procedures, the DSC analysis was conducted on the 30 days conditioned specimens. The crystallization or melting of waxes in bitumen causes an energy change. Both heating and cooling cycle can be used to investigate the presence of wax fraction in bitumen. The exothermic and endothermic effect during the cooling and heating scan generally represents the wax crystallization and wax melting, respectively. The tests were performed using a METTLER instrument (model METTLER TOLEDO DSC 1) equipped with a refrigerated cooling system. Approximately 15 mg of bitumen was scraped from each AFM specimens into the DSC sample pans, which were sealed with lids with a small hole to prevent pressure build up and possible deformation of the cups during the tests. The sample pan was then placed horizontally on a hot plate at 80°C for 10 min to let the bitumen touch the pan bottom and also level out the surface. Afterwards, the sample pan was placed in DSC furnace and cooled to –70°C and kept at this temperature for 5 min. Data was recorded during heating from –70°C to +110°C and cooling from +110°C to –70°C, where the heating and cooling rate was 10°C min–1 . During the test and sample preparation the bitumen melted completely due to the heating, which makes it reasonable to assume that the hypothesized surface aged thin layer mixed with the lower  C 2014 The Authors C 2014 Royal Microscopical Society, 254, 95–107 Journal of Microscopy 

layers, resulting in a homogenized bulk measurement. Thus in the cooling cycle, any changes in crystalline fractions due to the conditioning can be observed in the exothermic events.

Results and discussion To have a visual inspection of the specimen surface, the optical image (top) and digital camera image under AFM (bottom) of 30 days exposed specimens were taken (Fig. 5). It can be seen that with the increasing degree of exposure the surface distorting increased, which might be caused due to the surface ageing. The maximum effect can be observed in Figure 5(A) since this was exposed to both UV and air. Figure 5(D), which shows the sample that was exposed to neither UV nor air, that is, no ageing, does not show any surface distortion as this one. All three bitumen showed exactly the same trend, therefore, only bitumen B specimen’s images are shown in the figure. To investigate the surface ageing further, the results obtained from the AFM and the DSC are discussed in the following sections.

Surface topography Figure 6 shows the typical topography images with their computed percentages of microstructure of the three unconditioned bitumen specimens. Typical bee structures can be seen for bitumen B and C since these two bitumen contain natural wax (cf. Table 1). Bitumen A shows 0.5% of nanostructures with maximum height of 10 nm, which indicates a super flat surface. This topography information of the unconditioned specimens can be used as a benchmark, to detect any change due to the systematic conditioning. The AFM topography images of the bitumen A under four different levels of exposures are shown in Figure 7. In this figure, the top row images were aged over 15 days and images in the bottom row over 30 days. Since bitumen A is nonwaxy bitumen, no bee-shaped microstructures are present on

100

P.K. DAS ET AL.

Fig. 6. Topography images (20 μm × 20 μm) of unconditioned bitumen: (A) Bit-A, (B) Bit-B and (C) Bit-C.

Fig. 7. Topography images of Bit-A with four different levels of exposures after 15 and 30 days.

the Figures 7(A)–(H). However, 0.3% of nanostructures with maximum heights of 5 nm can be observed in (C) and (D), which may be due to crystallization effects of some molecules that are too few to show any visible endothermic or exothermic effect in DSC analysis (Das et al., 2013). Notably, the same specimens after 30 days of conditioning did not show the presence of any nanostructures. In Figures 7(A), (B), (E) and (F), the black spots representing microholes on the surface. Interestingly, only those specimens that were ex-

posed to UV showed such microholes, leading to the conclusion that those microholes are most likely caused by UV radiation. Figures 8 and 9 show the topography images of bitumen B and C, respectively, which was systematically conditioned to four different levels over 15 and 30 days. To investigate the microstructure evolution due to increased exposure intensity, all the exposed specimens were compared with the reference specimens (unexposed one).

 C 2014 The Authors C 2014 Royal Microscopical Society, 254, 95–107 Journal of Microscopy 

MICROSCALE INVESTIGATION OF BITUMEN SURFACE AGEING

101

Fig. 8. Topography images of Bit-B with four different levels of exposures after 15 and 30 days.

In the case of bitumen B (cf. Fig. 8), after 15 and 30 days, oxidation only caused a reduction of microstructures by 21% and 38%, respectively. The UV radiation caused a reduction of 100%, and the combined effect (both UV and air) caused a reduction of 79% and 86%, respectively. Similarly for bitumen C (cf. Fig. 9), after 15 and 30 days, oxidation only caused a reduction of microstructures by 21% and 52%, respectively. The UV radiation caused a reduction of 56% and 76%, and the combined effect caused a reduction of 76% and 87%, respectively. Comparing the images between 15 and 30 days of exposure (i.e. (A) vs. (E), (B) vs. (F) and (C) vs. (G)), it can be observed that the percentage of microstructure also decreases. In case of unaged specimens, that is, (D) vs. (H), however, the percentage of microstructure increases with time. From this analysis, it was observed that the visible microstructures at the surface decrease with an increasing degree of exposure and time. In the case of unexposed specimens, with time the percentage of microstructure at the surface increases. This confirms the posed hypothesis that the thin film at the specimen surface caused due to ageing may create a barrier, which restricts the microstructures to float towards the surface and if there is no such barrier at the surface, the percentage of microstructure increases with time.  C 2014 The Authors C 2014 Royal Microscopical Society, 254, 95–107 Journal of Microscopy 

Micromechanical properties From the AFM-QNM force mapping, the elastic modulus and adhesive force measurements were also made. In Figure 10, a typical example of the investigated bitumen’s elastic modulus and adhesion is shown in the left and right images, respectively. The top images are representing the three-dimensional distribution of micromechanical properties over the surface while the middle and bottom graphs are the normal distribution of measured values over the section 1 and 2, respectively. Section 1 represents the bee-shaped microstructures with their surrounding regions and section 2 represents the smooth matrix. In the figure, the elastic modulus and adhesion increases when the colour of the image goes from dark to light. It can be seen that the elastic modulus is higher in section 1 than section 2, which means the stiffness measured in the bee region is higher than the smooth matrix. In case of adhesive forces, the region surrounded by the bee showed lower value than the smooth matrix. Similar phenomenon was observed for all the bitumen investigated in this study. Since bitumen is complex mixture of hydrocarbons, it has a wide range of melting temperature. From previous studies (e.g. Das et al., 2013), it was found that the micromechanical properties of the microstructures and the matrix are very temperature dependent.

102

P.K. DAS ET AL.

Fig. 9. Topography images of Bit-C with four different levels of exposures after 15 and 30 days.

One needs to be careful about the temperature during the AFM scanning, as a small variance in temperature may affect these properties significantly. In the presently shown AFM scans, the samples were kept under 20°C in a temperature controlled environment. Using the AFM-QNM force mapping data, the changes in the elastic modulus due to different types of exposure or ageing history have been investigated, shown in Figure 11. The corresponding measured modulus for bitumen A, B and C is shown in the Figures 11(A)–(C), respectively. It can be seen that the modulus of microstructures is always higher than the matrix at the measured temperature. The consequences of the matrix modulus due to the exposures effect on ageing compared with unaged bitumen have been summarized in Table 2. Compared with the unaged specimen, for bitumen A, the modulus of matrix increases by 1.3 times due to oxidation only, 5.7 times due to UV radiation only and 25.9 times due to the combined effect (both UV and air). Similarly, the modulus of the matrix of bitumen B increases 1.7 times for oxidation only, 3.2 times for UV radiation only and 13.7 times for the combined effect. For bitumen C, these values were 1.2, 2.6 and 7.3 times, respectively.

From the Figure 11 and Table 2, one may observe that regardless the bitumen type the modulus increases with the level of exposure and the highest value is always found after the combined exposure. More interestingly though, it should be noted that the specimens exposed to air always showed a lower modulus increase than the UV exposed one, which indicates that UV radiation hardens the bitumen surface more than oxidation. This is a very important observation, especially given that currently used standardized laboratory ageing tests ignore the effect of UV radiation entirely. For the three types of bitumen, the evolution of adhesive force or tackiness due to different types of exposure or ageing history are also investigated as depicted in Figure 12. The corresponding measured adhesive force for bitumen A, B and C is shown in the Figures 12(A)–(C), respectively. The consequences of the matrix modulus due to the exposures effect on ageing compared with unaged bitumen have been summarized in Table 3. It can be observed for bitumen A, oxidation only reduces the adhesive force of the matrix by 11%, UV radiation only caused a reduction of 63% and the combined effect caused a reduction of 42%. For bitumen B, oxidation only reduces the adhesive force of the matrix by 16%, 71%

 C 2014 The Authors C 2014 Royal Microscopical Society, 254, 95–107 Journal of Microscopy 

MICROSCALE INVESTIGATION OF BITUMEN SURFACE AGEING

103

Fig. 10. Histogram of (A) elastic modulus and (B) adhesive force measured by AFM-QNM force mapping.

due to UV radiation only, and 57% due to the combined effect. In case of bitumen C, these values were 22%, 57% and 81%, respectively. This analysis shows that, with the ageing, the adhesion of the bitumen specimens reduced. This finding supports the traditional ageing concept that the tackiness reduces with ageing of bitumen, which may cause adhesive failure of the bond between bitumen and aggregates. In agreement with the previous result (cf. Fig. 10), irrespective to bitumen type, the adhesive force of the microstructures is found to be lower than the matrix. Again, one may observe that the combined effect on adhesive force is not simply a summation or multiplication of the individual effects, as found in case of elastic modulus. Based on the obtained data (cf. Tables 2 and 3), in the following a discussion is given on how the total ageing effect could be based on the individual effects. Assuming that the unaged bitumen generic property E0 is affected by a total ageing factor χ = f (UV, air, T , t), which is a function of the individual ageing processes as described in this paper, temperature and

 C 2014 The Authors C 2014 Royal Microscopical Society, 254, 95–107 Journal of Microscopy 

exposure time. It should be possible to define a unique function for χ based on its materials properties, for example: Eχ = χ E0,

(1)

where Eχ is the generic property of aged bitumen and χ = β α , χUV , T , t). f (χair Assuming, for purpose of demonstration, a simplified function as: β

α χ = χair + χU V

(2)

in which α and β should be dependent on the material properties, time and temperature; and should be highly related to the fundamental mechanism which controls the ageing process. Based on a regression analyses, a best fit could be found for the tested bitumen. Nevertheless given the fact that this would not relate to the nature of the bitumen, nor contribute to functions that can be extrapolated, here merely the comment is made that more research is needed to allow for the determination of such functions. But that this would be the way forward in

104

P.K. DAS ET AL.

Fig. 11. Measured modulus at different level of exposure.

Table 2. Analysis of the elastic modulus due to different exposures. Elastic modulus (MPa) at exposure of Bitumen A B C

Normalizing with respect to unaged bitumen (d)

Air (a)

UV (b)

Combined (c)

W/o exposure (d)

Air (a/d)

UV (b/d)

Combined (c/d)

90 160 104

400 304 234

1810 1300 660

70 95 90

1.3 1.7 1.2

5.7 3.2 2.6

25.9 13.7 7.3

adding fundamental knowledge and meaningful characterization of the ageing propensity of bitumen into asphalt mix design. Heat flow characteristics From the aforementioned AFM topography and force mapping analysis, it has been observed that there is a certainly change on the bitumen surface, caused by the different ageing history. To investigate any change in crystalline fractions due to this thin film surface ageing, DSC analysis has been conducted. As described earlier, DSC is a thermoanalytic method that can determine physical changes in bitumen associated with heat exchange. If there are any crystalline molecules present in the bitumen, during the heating scan these molecules need more heat to melt. Once the bitumen absorbs extra heat and becomes cooler than the furnace, causes the flow curve point down which is known as endothermic event. The reverse logic applies to exothermic events where energy is released.

The heat flow characteristics of all the specimens with 30 days of exposures have been investigated. As expected from the material physical properties (cf. Table 1), both endothermic and exothermic effects were observed for bitumen B and C specimens whereas, bitumen A did not show any enthalpy changes. The correspondent zones to the endothermic and exothermic event for bitumen B are zoomed into in Figures 13(A) and (B). In the figure, the areas under the wide endothermic peaks represent the enthalpy absorbed due to melting of wax. One may observe that all the four curves in Figure 13(A) have almost the same starting–ending points, and can be superimposed into one curve. This indicates that there is no change in enthalpy due to the different exposures, which means the percentage of wax in the bitumen remained constant after 30 days of exposure. Since the percentage of wax remains constant, then the total area under the exothermic peaks in Figure 13(B) should also have a constant value. It can be observed, however, that during the cooling the shape and the starting–ending point of the exothermic peak somehow  C 2014 The Authors C 2014 Royal Microscopical Society, 254, 95–107 Journal of Microscopy 

MICROSCALE INVESTIGATION OF BITUMEN SURFACE AGEING

105

Fig. 12. Measured adhesive force at different level of exposure. Table 3. Analysis of the adhesive force due to different exposures. Adhesive force (nN) at exposure of

Normalizing with respect to unaged bitumen (d)

Bitumen

Air (a)

UV (b)

Combined (c)

W/o exposure (d)

Air (a/d)

UV (b/d)

Combined (c/d)

A B C

13.50 13.00 12.70

5.60 4.50 7.00

8.70 6.70 3.10

15.10 15.50 16.30

0.89 0.84 0.78

0.37 0.29 0.43

0.58 0.43 0.19

Fig. 13. Heat flow analysis of different levels of exposure for Bit-B.

differ from each other. The starting point of exothermic event represents the starting of crystallization process. It can be observed that the crystallization starts earlier for aged (i.e. only UV, only air and both exposed) specimens than the unaged one. In the crystallization process, any presence of impurities  C 2014 The Authors C 2014 Royal Microscopical Society, 254, 95–107 Journal of Microscopy 

may act as a promoter, which may accelerate the process. The less soluble hydrocarbons produced from the chemical reaction due to the exposures, may introduce such impurities in the bitumen matrix so that the crystallization process starts earlier than the without exposed one. Identical phenomenon was

106

P.K. DAS ET AL.

also observed in case of bitumen C specimens. Since bitumen A specimens did not show any exothermic and endothermic effect, no such phenomenon can be observed.

Microstructure evolution due to ageing From the aforementioned results and discussions it should be noted that with ageing the percentage of bee-shaped microstructures reduce on the surface, yet the total amount of wax in the specimens remains constant. From previous AFM studies (Huang & Pauli, 2008; Wu et al., 2009; Zhang et al., 2012) on PAV aged bitumen, however, an increase of beeshaped microstructure compared with unaged bitumen was reported. This has lead some researchers to conclude that ageing increases the amount of bees in bitumen, which was in line with the earlier belief that the crystalline fraction increases as bitumen ages. The hydrocarbons in bitumen react with atmospheric oxygen, where the reaction type and rate depends on the surrounding temperature and pressure. Thus, the type of chemical reactions may vary with the change of pressure and temperature. In the accelerated ageing simulations, especially the PAV test, bitumen gets exposed to extremely high pressure and elevated temperatures for a short time. There is the possibility that the product after the accelerated aging test may differ significantly from the slow but long ageing process with different temperature and pressure history (field ageing). Thus, ‘ageing’ from a PAV test is indeed a completely different process from ageing in the field.

From the experiments it was found that the micromechanical properties such as stiffness and tackiness of the specimens changed with the different level of exposures, increasing its overall stiffness and losing its tackiness. Comparing all four different exposure types, it was concluded that UV radiation caused more surface ageing than the oxidation. Thus, UV radiation itself has a great impact on surface ageing although in most of the laboratory ageing simulation tests ignore the effect of UV radiation. The combined exposure showed, however, not only the most stiffening effect and the lowest adhesive force since this was the most extreme exposure case, but it was observed that this combined effect is not simply a summation or multiplication of the individual effects. Though the thermo-analytical investigation showed that the crystalline fractions in the bitumen remain constant even after been exposed to air, crystallization of wax molecules did start earlier for the exposed specimens. This implies that the chemical reaction product due to the exposures creates impurities in material and causes the acceleration of the crystallization process, which may partly explain the observation of the heightened ageing in the case of the combined conditioning. Based on the research shown in this paper it is recommended to further identify the fundamental mechanisms and material properties that control the ageing process. This could lead to detailed ageing functions that can be used to accurately account for the ageing process in asphalt pavements and material design. This, in return, can give the asphalt community the tools to move away from the currently used accelerated laboratory tests that can lead to erroneous conclusions.

Conclusions

Acknowledgements

A hypothesis regarding the mechanism of bitumen surface ageing has been presented and validated utilizing AFM and DSC. To validate the surface ageing, three different types of bitumen with different wax contents were exposed to four different types of conditioning: both UV and air, only UV, only air and without any exposure, for 15 and 30 days. Regardless the bitumen type, the percentage of microstructure on the surface reduced with the ageing. For the extreme exposure (both UV and air), the quantity was lowest and with time it became even lower. Since it was previously (Das et al., 2013) concluded that the microstructures are related to the wax content, one would expect the wax content to also become less. From the DSC results, however, it was proven that this is not the case, as the wax content stayed in fact the same. This observation implies that a thin film forms at the specimen surface caused due ageing thus creating a barrier, which restricts the microstructures to float towards the surface. For the unconditioned specimens, the percentage of microstructure increased with time, as there is no such thin film formed due to UV radiation or oxidation, allowing thus the waxes to float up towards the surface.

This research was sponsored by Swedish Road Administration and KTH Royal Institute of Technology, Sweden. The authors acknowledge M˚ans Collin and Per Redelius for many valuable discussions around the table. Nynas A.B. is thanked for providing the bitumen. References Airey, G.D. (2003) State of the art report on ageing test methods for bituminous pavement materials. Int. J. Pavem. Eng. 4(3), 165–176. Al-Azri, N., Jung, S., Lunsford, K.M., Ferry, A., Bullin, J.A., Davison, R.R., Glover, C.J. (2006) Binder oxidative aging in Texas pavements: hardening rates, hardening susceptibilities, and impact of pavement depth. Transp. Res. Rec. 1962(1), 12–20. Allen, R.G., Little, D.N. & Bhasin, A. (2012) Structural characterization of micromechanical properties in asphalt using atomic force microscopy. J. Mater. Civil. Eng. 24(10), 1317–1327. ASTM D 3418–12. Standard Test Method for Transition Temperatures and Enthalpies of Fusion and Crystallization of Polymers by Differential Scanning Calorimetry. Bellitto, V.J. ed. (2012) Atomic Force Microscopy – Imaging, Measuring and Manipulating Surfaces at the Atomic Scale. InTech, Croatia.  C 2014 The Authors C 2014 Royal Microscopical Society, 254, 95–107 Journal of Microscopy 

MICROSCALE INVESTIGATION OF BITUMEN SURFACE AGEING

Branthaver, J.F., Petersen, J.C., Robertson, R.E., et al. (1993) Binder Characterization and Evaluation, Vol. 2: Chemistry. SHRP-A-368. Das, P.K., Kringos, N., Wallqvist, V. & Birgisson, B. (2013) Micromechanical investigation of phase separation in bitumen by combining AFM with DSC results. Road Mater. Pavement Des. 14(S1), 25–37. de Moraes, M.B., Pereira, R.B., Simao, R.A. & Leite, L.F.M. (2010) High temperature AFM study of CAP 30/45 pen grade bitumen. J. Microsc. 239(1), 46–53. Dickinson, E.J., Nicholas, J.H. & Boas-Traube, S. (1958) Physical factors affecting the absorption of oxygen by the films of bituminous road binders. J. Appl. Chem. 8, 673–687. Durrieu, F., Farcas, F. & Mouillet, V. (2007) The influence of UV aging of a tyrene/butadiene/styrene modified bitumen: comparison between laboratory and on site aging. Fuel 86, 1446–1451. Fischer, H.R., Stadler, H. & Erina, N. (2012) Quantitative temperaturedepending mapping of mechanical properties of bitumen at the nanoscale using the AFM operated with PeakForce TappingTM mode. J. Microsc. 250(3), 210–217. Huang, S.C. & Pauli, A.T. (2008) Particle size effect of crumb rubber on rheology and morphology of asphalt binders with long-term aging. Road Mater. Pavement Des. 9(1), 73–95. J¨ager, A., Lackner, R., Eisenmenger-Sittner, C. & Blab, R. (2004) Identification of four material phases in bitumen by atomic force microscopy. Road Mater. Pavement Des. 5, 9–24. Lesueur, D., Gerard, J.F., Claudy, P., L´etoff´e, J.M., Planche, J.P. & Martin, D. (1996) A structure-related model to describe asphalt linear viscoelasticity. J. Rheol. 40(5), 813–836. Loeber, L., Sutton, O., Morel, J., Valleton, J.M. & Muller, G. (1996) New direct observations of asphalts and asphalt binders by scanning electron microscopy and atomic force microscopy. J. Microsc. 182(1), 32–39.

 C 2014 The Authors C 2014 Royal Microscopical Society, 254, 95–107 Journal of Microscopy 

107

Lu, X. & Redelius, P. (2006) Compositional and structural characterization of waxes isolated from bitumens. Energy Fuels 20, 653–60. Lu, X., Talon, Y. & Redelius, P. (2008) Ageing of bituminous binders – laboratory tests and field data. In Proceedings of the 4th Eurasphalt and Eurobitume congress. European Asphalt Pavement Association (EAPA), Copenhagen, Denmark, pp. 1–12. Marr, D. & Hildreth, E. (1980) Theory of edge detection. Proc. R. Soc. Lond., B, Biol. Sci. 207, 187–217. Masson, J.F., Leblond, V., Margeson, J. & Bundalo-Perc, S. (2006) Bitumen morphologies by phase detection atomic force microscopy. J. Microsc. 221, 17–29. Pauli, A.T., Grimes, R.W., Beemer, A.G., Turner, T.F. & Branthaver, J.F. (2011) Morphology of asphalts, asphalt fractions and model wax-doped asphalts studied by atomic force microscopy. Int. J. Pavem. Eng. 12(4), 291–309. Petersen, J.C. (1993) Asphalt oxidation – an overview including a new model for oxidation proposing that physicochemical factors dominate the oxidation kinematics. Fuel Sci. Technol. Int. 11, 57–87. Schmets, A., Kringos, N., Pauli, A.T., Redelius, P. & Scarpas, T. (2010) Wax induced phase separation in bitumen. Int. J. Pavem. Eng. 11(6), 555–563. Soenen, H., Besamusca, J., Fischer, H.R., et al. (2013) Laboratory investigation of bitumen based on round robin DSC and AFM tests. Mater. Struct. doi: 10.1617/s11527–013–0123–4 Wu, S., Pang, L., Mo, L., Chen, Y. & Zhu, G. (2009) Influence of aging on the evolution of structure, morphology and rheology of base and SBS modified bitumen. Constr. Build. Mater. 23(2), 1005–1010. Zhang, H.L., Yu, J.Y., Feng, Z.G. & Xue, L.H. (2012) Effect of aging on the morphology of bitumen by atomic force microscopy. J. Microsc. 246(1), 11–19.