Interface Roughness Parameters and Shear Strength

The interlayer bond strength between binder and wearing course and several possible treatments of enhancing the contact surface roughness and the interlocking are investigated. For this purpose, conventional methods, such as shear tests, but also laser image acquisition of the binder upper surface have been used. The mechanical outcomes of a shear test device and the binder surface roughness parameters, have been compared looking for a relation between the shear performance and the surface characteristics. The comparison between the roughness average and the root mean square of the profile heights with the maximum shear stress shows the achievement of the same strength level for treatments with similar roughness parameters, as proved by the statistical analysis. Furthermore, the comparison between the roughness parameter kurtosis and the maximum height of the profile with the slope of the response curve before the peak and residual shear stress, demonstrates a better locking for more high peaks.


Introduction
Texture is a feature that plays an important role in most physical phenomena.In the field of road pavement, surface texture influences tire-road interactions, such as noise, wet friction, rolling resistance and tire wear (Loprencipe & Cantisani, 2013).Significant factors are shape, grading and size of the aggregates.Several measurement methods are employed: Sand patch, British pendulum test, Outflow meter and electro-optic methods.A large effort of research has been devoted to those aspects.
Texture seems to be important also for interfaces between layers forming the structure of a road pavement, with a direct influence on the bond resistance to the shear stresses caused either by the vertical loading applied by rolling wheels and in occasion of braking, accelerating and turning maneuvers.Therefore, focusing on the possible use of the most recent texture detection methods for the better comprehension of interface behavior seems to be of interest at the present time.In this paper, a novel application of the non-contact methodologies is proposed with the aim to evaluate the macrotexture of the upper binder surface in order to find a correlation between the roughness characteristics and the interface shear performance.The interface represents the contact plane between this investigated binder surface and the upper wearing coarse layer built on its top.
For this purpose a 3D laser scanning has been employed: it is an emerging technology that allows the acquisition of a great quantity of point clouds in few minutes.The application ranges from design, industrial quality control, surveying to scanning of buildings.The aim of this study is to apply this technique in order to acquire the contact surface shape and achieve roughness parameters.

Background
The importance of the interface between pavement layers and the negative influence of the lack of bond on the performance of the infrastructure, have been understood since the first 1970ies.The first achievement of most researches is the design and the development of new testing methodologies, able to characterize in situ or in lab the shear strength performance of the interface.Different devices and testing procedures have been presented, divided in several types: Torque Test; Tensile Test (Pull-off Test); Pure Direct Shear Test and Direct Shear Test with Normal Stress Applied, but the lack of standards on these procedures allows the diffusion of ever new prototypes, typically used to better investigate the different factors affecting shear interface strength.The comparison between the proposed devices has shown that the shear test is a good and effective method for testing the interlayer bond of asphalt pavements (Raab et al., 2009).
The pure direct shear test represents the most common equipment working through the application of a shear displacement and the record of the shear strength.The first one, known as Leutner Shear Test (Leutner, 1979), applies a constant shear displacement until the achievement of the failure of the investigated contact plane between the layers.In the following researches, comparable devices have been developed, such as the LPDS (Layer Parallel Direct Shear Test) used by Raab et al. (2004) and the shear test used by Molenaar et al. (1986).
Different working scheme belongs to the pure direct shear test known as LCB (Laboratorio de Camino de Barcelona) and presented by Miro et al. (2005) where a particular system of two supports holds the specimen and the shear load is applied at a constant deformation rate.The short distance between the two supports leads to the elimination of the bending moment.
To investigate the effect of the normal stress on the interface shear strength, direct shear test with normal load has been used by other researchers.From the first device of this type considered in the road pavement research of Uzan et al. (1978), direct shear boxes with the addition of a normal load have been developed as the ASTRA (Ancona Shear Testing Research and Analysis) (Canestrari et al., 2005) apparatus and the device presented by Chen et al. (2010).Recently, the Sapienza University of Rome designed and developed the SHSTM (Sapienza Horizontal Shear Test Machine), a machine able to perform tests on double-layer asphalt specimens 100 mm diameter and 120 mm high, 60 mm for each layer.The working scheme derives from the Miro device (Miro et al., 2005) but includes some modification in the support configuration in order to allow the application of a compression load during the test eliminating some specimen rotational problems that could develop in the previous scheme if a normal load is added.D'Andrea and Tozzo (2012) verified the statistic reliability of the device and investigated the relationship between several failure mechanisms and different state of imposed stress.
The experimental results obtained by the researchers have highlighted a strong dependence between layers interlocking and macrotexture of the binder surface, granulometric size distribution and compaction conditions.Several factors concerning the surface, such as cleanness, dryness, roughness and age have been considered in the experimental plan to define the interaction between different surface characteristics of the underlying layer and the adhesion properties (Raab et al., 2009;Tashman et al., 2006).This result is also confirmed by Santagata et al. (2008) that showed an increasing interlayer shear resistance for higher macro-texture.Sholar et al. (2004) found that coarse graded HMA mixes had higher shear strength compared to fine graded mixes and observed significant resistance improvement for milled surface.Furthermore, for milled sections, they noticed that using tack coat the increasing of the shear strength at the interface was not effective.The finding of a correlation between the interface shear strength and the surface features was confirmed by D'Andrea and Tozzo (2013) testing the interface shear strength of double layer asphalt specimens where the surface of the binder was treated in different ways to improve the roughness.West et al. (2005) also proved higher bond strengths for fine-graded mixtures than for the coarse-graded ones, considering the results of laboratory fabricated specimens and field experiments.The significant interactions of the mix type with other variables e.g.tack coat application rate, materials used, and testing temperature, was also evaluated with contradictory trends in some cases.
In the present paper an analysis based on visual methods is proposed with the aim to investigate the effect of roughness features of the binder upper surface on the interlayer shear strength when on the binder is built the wearing coarse and the double-layer system is tested.
With the introduction of emerging 3D laser imaging technology, the potential of high resolution 3D surface for the pavement analysis has been investigated by Wang et al. (2012).In this study, pavement texture and distress survey have been explored with a high-speed 3D imaging sensor system in order to model the pavement characteristics.They found a quantitative and useful correlation between the outcomes of the texture indicators and the different surface textures.
The use of laser inertial road profiler macrotexture measuring device has been also considered by several researchers (Flintsch et al., 2003;McGhee et al., 2003;McGhee et al., 2004), looking to possible applications as detecting and measuring methods for segregation and non-uniformities in HMA.Various parameters were computed to characterize the different surface texture, e.g.measure of contrast, correlation energy and homogeneity.
Furthermore, for quality assurance purposes, high speed laser profilers have been identified as promising tool for surface texture measurements in the state of Virginia.In the recent years, in the Virginia Smart Road has been tested a method for the final inspection process of asphalt pavements based on the automatic capture and analysis of newly constructed roads (de León Izeppi et al., 2007).From the captured images, various parameters for the characterization of the visual texture and a road profile with the non-uniformities location have been developed.They also implemented imaging techniques in several highway applications such as pavement crack detection, aggregate shape properties and characterization, pavement distress surveys.

Objective
In this study, different interfaces between asphalt and wearing course of asphalt pavement are investigated: different sizes of limestone chip coat, toothed compactor roller footprint, oily and smoothing surface and also no treatment.The purpose is to find a correlation between the interface strength and the parameters obtained through the surface characterization of the binder surface, acquired with a laser scanner.As study parameters, the maximum shear stress, the corresponding displacement and the slope of the response curve before the peak are selected.The surface data analysis is performed according to ISO standard 4287 and 13565, typically used for mechanical part characterization.The Roughness Average (Ra), the Root Mean Square Roughness (Rq), the Maximum Height of the Profile (Rt) and the roughness parameters skewness (Rsk) and kurtosis (Rku) are evaluated.

Specimen Preparation
Double-layer specimens including binder and wearing course have been prepared with bitumen PEN 60-70 and two different aggregate mixtures, according to the Italian standards.A Marshall Compactor has been used to prepare the samples in cylindrical molds, 130 mm high and 100 mm interior diameter, applying the compaction blows only to the upper surfaces.
According to previous research (D 'andrea & Tozzo, 2012), seventy-five blows have been assigned for both the layers because this level of compaction ensures the achievement of the density commonly required in Italian public work specifications as the 98% of the average bulk density (EN 12697) of the same mixture compacted in standard Marshall conditions (75 blows on upper and bottom side of a 63 mm high specimen).The mixtures characteristics are summarized in Table 1.
The upper binder surface has been treated in different ways: • Chipping: immediately after the binder compaction, limestone chip coat has been placed and inserted into the hot surface with the application of three compaction blows; two different chip coat sizes have been tested, using 5-10 mm and 10-20 mm limestone granules; • Smoothing: after cooling the binder specimens at room temperature, the upper surface has been rounded by tape grinding, usually referred as "sanding"; • De-bitumening: after cooling the binder specimens at room temperature, the interface has been left for 24 hours in a trichloroethylene bath, 1 mm depth to keep the surface clean by the bitumen; • Toothed compactor roller: on hot binder surface, three screws have been beaten by three compaction blows to simulate the passage of a toothed wheel; Laser scanner acquisitions have been carried out on all the treated binder surfaces.Moreover, a no treatment specimen has been acquired as term of comparison.
After the wearing course preparation, the samples, 120 mm high, have been left to cool at room temperature for more than 24 hours before performing the shear test.

Results
The outcom of maximu and residu corresponds to two different interlocking mechanism between the layers, as confirmed by two different strength levels.Also, the roughness parameter kurtosis and the maximum height of the profile are presented in relation with the slope of the shear response curve before the peak and the residual shear strength, respectively.The plots show the achievement of stronger resistance parameters for higher texture peaks, probably due to a better locking between the superposed layers.

Table 2
ce treatment; b) C erature of 20 °, slope of the r ; c) De-bitume in Table2in t e before the pe