The Baltic Journal of Road and Bridge Engineering

In recent years adhesion testing of pavement layers has gained more and more importance throughout Europe. In a lot of projects the bonding properties are determined after construction and many countries have developed standard requirements which the obtained bonding values have to fulfil. Tue paper focuses on the question of long-term behaviour of honding properties and presents the results (lf an extensive Swiss study. The study compares the bonding properties determined with the Layer Parallel Direct Shear Test (LPDS) according to Leutner of 14 high volume roads from the years 199397 to the values for the same road sections determined 9-13 years later. In addition, a second study conducted on a circular track investigated lhe differences between the bonding properties in the wheel path and outside the wheel path. It could be shown that in most cases the bond improves due to compaction and settlement caused by the traffic. Problems arise, result ing in a reduction of shear forces, when the pavement shows distress phenomena. The difference hetween the honding properties in the wheel path and outside it could he demonstrated in the circular track study, whereas the differences on the road depend on many factors and oflen seem to be eliminated over the years.


Introduction
In recent years bonding testing has gained more and more importance throughout Europe (Choi et al. 2005, Diak.hate et al. 2006, Krnntcheva et al. 2006, Tschegg et al. 2007). In a lot of projects the bonding properties are determined after construction and many countries have developed standard requirements which the obtained bonding values are to fulfil. The question of how the originally determined bonding properties develop over time is often neglected and has not attracted enough attention of the Road Authorities etc.
Although different researchers (Canstrari et al. 2005, Rabiot 1996 showed that the bond improves with time, this finding was never investigated on a larger scale in a systematic way. In some cases (Stockert 200 l) the re-evaluation of the shear strength took place onl)' a year after construction and therefore it cannot consider long-term effects and influences. Furthermore, the improvement of bonding properties is stated as a general fact, but no reference to traffic figures is provided. This paper describes the results of an extensive Swiss study of the long-term behaviour of bonding properties launched by the Swiss Road Authorities (ASTRA). Based on the results directly after construction of more than 1000 cores of 20 different pavements, a decade later the ISSN 1822·427X print I ISSN 1822-4288 online http: //www.bjrbe.vgtu.lt long-term bonding properties of some of these pavements could be determined again. The study compares the bonding properties determined with the Layer Parallel Direct Shear Test (LPDS) according to Leutner of 14 high-volume roads from the years  to the values for the same road sections determined in 2006. Jn add ition, a second study of a circular truck investigated the differences between the bonding properties in the wheel path and outside it.

Testing
The Layer-Parallel Direct Shear (LPDS) test device (Fig. I) is an EMPA modified version of equipment developed in Germany by Leutner being more versatile in the geometry and more sophisticated in the clamping mechanism (Leutner, 1979). The modified LPDS test device fits into an ordinary servo-hydraulic Marshall testing machine and allows the testing of cores with a diameter of about 150 mm and rectangular specimens; no normal force can be applied .
One part of the core (up to the shear plane to be tested) is laid on a circular u-bearing and held with a well defined pressure by a semicircular pneumatic clamp. The other part, the core head in Fig. 1, remains unsuspended. Shear load is induced to the core head by a sen1icircular shear yoke with a defonnation rate of 50,8 n1n1/n1in, thus producing fracture within the pre-defined shear plane of 5 n1m width.
The speci1nens were conditioned in a cli1nate cha1nber for 8 h and all tests were conducted at 20 °C. Since 2000 the LPDS is incorporated into the Swiss Standard SN 67196.
From the LPDS test, the shear force Fas a function of shear deforn1ation w is obtained. Additional to the shear force, the inax slope fro1n the diagran1 of shear force ?versus shear deforn1ation w can be used to define the LPDS max shear "stitTness" value Su)DS,max as follows: \Vhere StPDS,max -n1ax shear "stiffness'', kN/n11n; b.F -derivative of force, kN; b. lV -derivative of deforn1ation, n1m.

Situation and material
In 1999, a research project FA 12/94 of the interlayer adhesion of new pave1nent layers was co1npleted. It was found that the direct LPDS shear test according to Leutner was a useful tool for inter-layer shear adhesion assessn1ent (Partl, . The question of long-term performance of the inter-layer shear bonding, in tern1s of the relationship between the bonding properties directly after construction and the bonding properties after son1e years of trafficking was only touched on in a li1nited way in this forn1er project. Therefore a follow-up project vvas conducted entitled "Long-tenn performance of the bonding properties -relation between test value after construction and long-tern1 perforn1ance" (Raab, Partl 2007). In the fran1ework of this project, all 14 re1naining sections were investigated again, after l 0 or nlore years of traffic, in   (Raab, Partl 2004} order to predict the long-tenn perfonnance of the bonding properties. As shown in Table l, the surface course thickness varied bet\veen 29 and 60 mm. Since the construction took place before the European Standards becan1e effective, all tenns are taken according to the old Swiss Standard SN 640431. 7 pavements had stone mastic asphalt (SMA) and 4 road sections had asphalt concrete (AC) surface courses. In addition, 3 coring sites with special surface courses, i.e. gussasphalt (mastic asphalt) (GA), hot-rolled asphalt (HR) and porous asphalt (drain asphalt= DRA) were included. With the exception of 3 pave1nents all surface courses consisted of aggregate with a n1ax size of 11 1n1n. The surface courses were placed either on hot 1nix base course (HMT) with a n1ax aggregate size of 16, 22 or 32 1nn1 or on asphalt concrete (AC) and mastix asphalt (MA) with a max aggregate size of 16 inm each.
For nearly all road sections, traffic data, e.g. the average daily traflic ADT (vpd) and the percentage of heavy vehicles(> 3,5 t) are given in Table 1. In the original investigation, the pavements were either totally new or a new surface course had been placed over existing layers. In 2 cases, pavements with new surface courses on unknown old base courses were investigated (coring sites 18 and 19).
TI1e coring for investigating the long-term performance was conducted a few meters away from the original coring site. For every road section 5 cores were taken inside and another S outside the wheel path. In addition to the shear force at 20 °C, the max shear stiffness was determined.

Test results
Figs 2-4 depict the LPDS test results at 20 °C for the differen t test sections. Each Figure represents the comparison of the shear force values immediately after construction but before traffic and the shear force values after several years of traffic. In case of the values immediately after construction and before traffic all bars are mean values of at least 7 specimens. In case of the values after several years of traffic the mean value of all 10 cores is given, as well as the mean value of 5 specimens in the wheel path and 5 specimens outside the wheel path. The standard deviation is also presented.
For the SMA surface courses in Fig. 2 all test sections were between 9 and 13 years old and therefore quite comparable. Except for one section (road section 3), which showed severe rutting in the wheel path, the pavement surfaces of all other sections were in good condition and did not reveal surface distresses.
Apart from section 3 and section 6, a slight (9-14 %) to great (28-45 %) improvement of the shear forces over time could be observed for all sections. Although the mean value after trafficking was low, il is remarkable that section 3 had shear forces outside the wheel path which were by 12,7 % higher than directly after construction.
Section 6 was the only coring site, where an overall deterioration of the shear forces (ca. 3 1 % ) over time could be found.
For the great majority of road sections (4 sections) there was only a small difference between the shear forces inside and outside the wheel path, whereas in two cases the values in the wheel path were higher and in two other cases the ones outside the wheel path were higher. One section (section 2) subjected to a heavy bus traffic showed a S kN lower shear force in the wheel path than outside the path.
The data for the roa<l sections with SMA surface courses showed that in the wheel path an improvement of the shear force could not always be expected, but even the opposite was true.
For the AC surface courses in Fig. 3 all test sections were between 9 and 13 years old, therefore in this respect quite comparable. All investigated road sections except one (section 15), which revealed cracks and aggregate Joss, showed an intact pavement surface.
The shear forces for all road sections with AC surface courses were 7 % and 23 % higher after several years. For the above-mentioned road section 15, where aggregate Joss and cracking occurred, the shear forces in the wheel path were by 35 % lower than those directly after construe-  tion, whereas the values outside the wheel path were by 33 % higher than before traffic. In contrast to the road sections with SMA surface courses. all the sections with AC surface courses showed an improvement of the values in the wheel path of 17-39 %noticeably higher than those outside the wheel path, where the increase was only 5-10 %. Fig. 4 shows the result for the HR, GA and ORA surface courses which were 13, 12 and 10 years old and showed no surface distresses. For 3 road sections an improvement of the bonding properties of 12,2 %, 6,9 % and 45,2 % could be observed. For the sections with GA and ORA no significant difference between the shear fo rces in and outside the wheel path could be fo und. For the HR section, the shear forces outside the wheel path were by 30 % higher than the values directly after construction, It is clear that the shearstiffness increased in the course of the years between 9-45 %. However, it can not be demonstrated in which way the position of the cores -in or outside the wheel path -had an influence on the shear stiffness, since in some cases the stiffness in the wheel path was higher and in others it was lower.
For sections 3 and 6, where decreasing shear force values had been fou nd, the increase of the shear stiffness was lower than in other sections.
The comparison of the situation immediately after construction with the situation after 9 lo 13 years of traffic reveals an increase in shear force between surface and base course of 7-45 %. Of 14 sections only 2 sections showed a decrease in the average shear force (4,5 and 32 %). While the pavement of one section (section 3) showed rutting and an increase of the average shear force outside the wheel path, no optical reason could be found for the decrease observed for th e second section (section 6).
As soon as pavements show surface distresses such as rutting or cracking, the average shear forces after some years of traffic decrease in and in some cases outside the wheel path.
The traffic figures given in Table I are quite different for all sections. This is not only true for the ADT but also for the percentage of heavy vehicles (trucks> 3,5 t). These percentages vary between 2,6-1 l %.
Regarding ADT and percentage of heavy vehicles, sections 6, l 0 and 15 are the most critical ones. The measured shear forces of these sections demonstrate that a high percentage of heavy vehicles over a long period of time can cause bonding problems: Section 15 reveals severe pavement distresses and decreased shear forces in the wheel path. For section 10 the shear forces outside the wheel path are lower than immediately after construction. On the other hand, Section 6 is the only section which shows a decrease of shear forces despite its intact pavement surface. In addition, section 3, with severe rutting and a deterioration of the bonding properties in the wheel path, has a very high percentage of heavy traffic, especiaUy for a city road. Although there are no traffic figures given for section 2, this section is located near a bus stop and is therefore highly frequented by buses.

Situation and material
In the second study, conducted on a circular test track, the differences of bonding properties in the wheel path and outside the wheel path were i.nvestigated.
In the framework of another research project, different materials have been placed in a circular track and were traffic over a certain time (Rabaiotti, Caprez 2006). Although for these materials no values directly after construction were available, the project was considered ideal for investigating bonding properties in and outside the wheel path, since, as opposed to the situation on a real road, traffic happens in a very defined, canalised manner. The material was placed in 5 sections with a length of20 m as shown in Fig. 6. The pavement details of the different sections are in Table 2.
As shown in Table 2, all surface courses consist of identical asphalt concrete, AC l l, with a penetration grade binder of 50/70. The hot mix base courses, AC T 22, differ from each other in binder type. For section 5, a high modulus binder base course, AC EME (enrobe a module eleve), was used. The exact description of the materials in (Rabaiotti, Caprez 2006) .   Fig. 7. Fig. 8 depicts the air void content determined on cores taken from the different section in and outside the wheel path.
With mean values between 36,6-45 kN the obtained interlayer shear forces were very high. The comparison between the different sections reveals that the results for sections 1 and 2 with the same materials correspond very well with shear force values of38 kN and 36 kN. Section 3 with a softer binder in the base course shows a slightly higher shear force of 42 kN, which is comparable to section 5, where a value of 43 kN was found . The high est value of 45 kN was received for section 4, where a polymer modified bitumen was used in the base course.
In general, the investigation demonstrates the differences in the shear forces in and outside the wheel path. This can be directly linked with the different comp action and air void content in these a reas (Fig. 8), which is on a circular track more uneven than on a straight road.

Conclusions
As the results of this extensive Swiss study on long-term behaviour of bonding properties, the following main conclusions can be drawn.
1. The traffic induced compaction and settlement after construction has a positive effect on bonding properties of carefully designed and m anufactured roads. The increase in shear resistance in terms of shear force between surface and base course can reach 40 % after 10 years. In terms of shear stiffness the in crease can go even up to 57 %. 2. Pavements showing distress phenomena, such as cracks and aggregate loss, have a high potential to suffer also a reduction of interlayer shear resistance, especially in the wheel path.
3. As long as the road does n ot show obvious surface distresses, differences between interlayer shear force and stiffness inside and outside the wheel path are generally not expected; however, they may occur in special cases where a local chan ge in perform ance, such as rutting and cracking, is clearly predominant.
4. As soon as pavements show surface distresses such as rutting or cracking, the average shear forces and shear stiffness values after some years of traffic decrease and in some cases also outside the wheel path. 5. A high percentage of heavy vehicles over a long period of time can cause bondin g problems. 6. Behaviour on roads can divert substantially from test pave1nent sections that are tested under extren1e conditions in stationary accelerated pavement testing facilities, such as circular test tracks. This is partly due to the different co1npactions and air voids in the track sections resulting from special construction conditions and equip-n1ent as well as the more canalized accelerated loading Vo/hich does not leave much potential for mobilizing healing effects.
This investigation provides a valuable quantitative practical input, e.g. for standardization purposes. It shows that interlayer shear properties are clearly subjected to change over the life-ti1ne of a pave111ent. Hence, further investigations will also have to focus on the continuous monitoring of the whole process of change, i.e. the development of interlayer properties over tilne as a function of traffic and climatic history as well as of the evolution of volumetric and chen1ical-physical properties. 1bis future task \Vill certainly not only be a challenge in 1nodeling but also in sensoring and n1011itoring techniques.