Modeling and Laboratory Investigation of Tack Coats as Bituminous Pavement Interlayer

Abstract

The adhesive properties of tack coats between asphalt pavement layers are crucial for the pavement’s structural behavior. This study first involved numerical analyses to compare stress patterns, deformations, and displacements in the pavement structure under various geometric and mechanical conditions. A rational calculation method based on the theory of elastic multilayer systems was used to quantify the impact of layer properties such as thickness, stiffness modulus, and Poisson’s ratio on interlayer bonding. Three bonding conditions—Full Friction, Partial Bonding, and Full Debonding—were analyzed to understand the tack coat’s effect between the top two layers. The second phase involved characterizing the mechanical behavior of the interface through shear strength tests (Leutner shear test) on both laboratory-prepared specimens and samples from a 10-year-old highway. Specimens were prepared using a Roller Compactor and tested under different interface conditions: hot-on-hot (H/H), residual bitumen 200 g/m² (RB 200), and residual bitumen 400 g/m² (RB 400). The tests examined the bonding effects in terms of tangential force and shear displacement at failure, as well as the impact of vehicular traffic on rutting and fatigue failure. Finally, this study investigated the long-term aging effects of the binder on interlayer bonding and sought to correlate the results of numerical calculations with those of the laboratory tests.

1. Introduction and Literature Review

From a global perspective, transportation infrastructure plays a fundamental role in the movement of people and goods, connecting centers and contributing to economic and social exchange. In particular, over the past years, roads and highways have witnessed continuous growth, driven by the ever-increasing demand for transportation and by the advent of new technologies and construction methodologies that led to the creation of safer, more functional, and durable infrastructures. In recent years, scientific research conducted in the field of road pavements has paid significant attention to the poorly understood topic of the adhesion between the layers of asphalt pavement [1,2,3,4,5,6,7,8,9,10,11,12,13]. For instance, in 2002, Mohammad et al. [1], evaluated the use of tack coats in asphalt pavements through controlled laboratory shear tests to determine the optimum application rate. Four emulsions (CRS 2P, SS-1, CSS-1, and SS-1 h) and two asphalt binders (PG 64-22 and PG 76-22 M) were selected as tack coat materials. The results showed that CRS-2P emulsion at 0.09 L/m² provided the maximum interface shear strength at both 25 °C and 55 °C. In 2009, another study [5] examined the long-term behavior of bonding properties in the lower layers of the pavement, finding that bond can decrease significantly over time. Results from the layer parallel shear test (LPDS) on Swiss roads show that rehabilitations or older constructions do not exhibit bond improvement, with shear force values potentially decreasing by up to 50% after 9–13 years.

Chun et al. [9], evaluated the impact of different interlayer bonding conditions on pavement performance and service life using finite element analysis and full-scale field tests, highlighting the importance of accurate modeling. Results showed that improved interlayer bonding positively affects pavement strain responses, enhancing structural performance and service life.

A study conducted on pavements in Malaysia [14] revealed that one of the major factors contributing to the degradation of the pavement is the poor bonding between layers, especially when it occurs between the wearing and binder courses, as compared to what happens along the deeper interfaces. Furthermore, degradation is amplified when high horizontal forces act in addition to vertical loads. This phenomenon can occur, for example, at road intersections, curves, slopes, and highway junctions where vehicles commonly brake and accelerate. Considering the simultaneous loss of pavement functionality due to fatigue and the accumulation of permanent deformation, the authors have found that the reduction in service life caused by poor bonding between wearing and binder courses is approximately from 94 to 98%.

A significant decrease in pavement service life may be also observed where bonding is poor between the base and binder layers courses. From a study conducted on the performance of pavements in Poland [15], it was observed that a low level of bonding between the base and binder courses leads to a simultaneous increase in tensile deformation (ε_t) on the intrados of these layers, resulting in a reduced fatigue resistance of the pavement and service life. From numerical calculations, it was observed that fatigue-induced failure initiates at the base of the binder layer, because the tensile deformations are greater than those observed at the bottom of the base layer. In addition, a low level of bonding also increases vertical deformations (εz) on the subgrade’s extrados, increasing the likelihood of deep settlements leading to the subsequent surface rutting of the pavement. Hence, the need to design and build road pavements composed of well-connected layers to make a monolithic structure with an extended service life.

In 2014, the study on the behavior of bonded interfaces in bituminous pavements has highlighted interesting information regarding the physical and mechanical characterization of the pavement components [16]. In particular, it has been observed that the presence/absence of bonding coats, the type of bituminous emulsion, and the compaction method of the material play a fundamental role in the level of adhesion. The effect of each of these parameters on bonding has been evaluated in terms of tangential plain stress at failure following the Leutner shear test executed on laboratory-made samples as well as on cores extracted onsite. Specifically, the test results showed that choosing a bituminous emulsion composed of hard bitumen (50/70) provides a better connection between the wearing and binder courses compared to an equal amount of emulsion made out of soft bitumen (160/200). This result is justified by the fact that a bond coat made of soft bitumen generates a significant stiffness gradient between the bond coat itself and the layers of the pavement, which are made of a harder binder (35/50). The tangential stress value observed on specimens characterized by the presence of a soft residual bitumen at the interface is comparable to that one obtained on specimens without a bond coat. Furthermore, using a bond coat made of hard bitumen ensures better bonding than using none, except for the hot-on-hot laying technique. The latter promotes aggregate interlocking, thus improving the resistance to tangential shear. Under these conditions, in fact, the interlocking phenomenon at the interface becomes dominant compared to the tangential load causing the slipping of layers past each other. Often, the hot-on-hot technique allows the achievement of performances comparable to bond coats. In addition to the type and/or presence of the bond coat, a fundamental role in the connection between one layer to the other is played by the compaction method of the material. Laboratory tests have shown that compaction with a gyratory press guarantees the best bonding condition, as opposed to compaction of the same material with a static roller.

Laboratory compaction performed with a vibrating roller or Standard Compactor [17] leads to similar results. It has also been observed that the tangential resistance at the wearing and binder courses’ interface improves with an increase in compaction level of the surface layer. Greater compaction of the wearing course, in fact, promotes the interlocking of the aggregates at the interface, thus leading to an increase in the shear strength required for the specimen’s failure.

In the context outlined in the literature, the primary objective of the initial phase of this study is to assess, through a combination of numerical analyses and experimental investigations, the influence of the bonding between bituminous layers (in flexible road pavements) on their overall performance. More specifically, this research delves into the impact of bonding strength on the response to stresses induced by traffic, considering the continual growth in road transportation demands. The central focus of this investigation is on the bituminous layers of the pavement, where traffic-induced stresses are most pronounced and where maintaining exceptionally high performance is essential to avert impairments that could jeopardize road safety and driving comfort.

The research activity of the second part of this study refers to the materials, techniques, and regulatory requirements in force in Switzerland, where the analysis of bonding between pavement layers has gained importance in recent years and is the subject of study in many laboratories and research institutes. The Swiss standard VSS 40 430, 2022, which is used to verify the routine compliance of materials used in construction sites, currently prescribes a minimum value of tangential force at failure (measured at the interface between the layers through direct shear tests), which is indicative of the bonding strength [18]. However, the standard does not set a specific limit for the displacement at failure, permitting the material to reach the minimum tension regardless of the amount of relative slip between the layers. Through the investigations conducted, the impact of the level of bond between the bituminous layers on the durability of the superstructure is determined, as well as how the displacement at failure, which is obtained from mechanical characterization tests, controls the interface’s performance.

2. Goals and Methodology

The scope of this study is divided into two parts: the theoretical investigation on the effect of tack coat on the structural behavior of a road pavement, followed by laboratory investigations on the performance of the tack coat in double layers (asphalt lab-made specimens and asphalt cores taken from site).

To accomplish the first goal, the theoretical behavior of a typical pavement section of Swiss highways has been determined using a mechanistic method. In addressing this topic, a rational calculation method tailored specifically for flexible pavements is adopted. This method focuses only on the layers’ structural characteristics, making it a key reference tool for design and verification. The algorithm for the rational method is illustrated in Figure 1a.

The second goal is reached by conducting shear strength tests (Leutner shear test) on specimens both prepared in the laboratory and taken from an existing highway pavement. The Swiss standard procedures were adopted to characterize the mechanical behavior of the interface between the bituminous layers. Moreover, the mixtures used for preparing the specimens in the laboratory are those ones commonly specified for the construction of highway pavements in this country, which are designed to withstand significant and frequent loading cycles. The algorithm of laboratory investigations is presented in Figure 1b.

3. Rational Calculation Method

To investigate the bonding between the double layers (wearing course and binder course), a numerical analysis was conducted. This analysis compared stress, deformations, and displacements at selected points within the pavement structure under varying thickness and mechanical properties [19]. To achieve this, a rational calculation method based on the theory of elastic multilayer systems was employed, consisting of two main steps.

First, the modeling of the pavement to calculate its structural response to vehicle loads and environmental conditions has been outlined. According to the Elastic Multilayer Theory, the pavement structure is represented as layers extending infinitely radially but finite vertically, except for the subgrade, which extends infinitely in both directions. Each layer is characterized by its thickness (t), stiffness modulus (E), and Poisson’s ratio (ν).

Secondly, the prediction of the pavement’s long-term performance was carried out using semi-empirical models to correlate the structural response with progressive deterioration.

While the study primarily focused on examining the interface between bituminous pavement layers, the calculated mechanical values highlighted the critical role that bonding between layers plays in determining the pavement structure’s service life.

3.1. Pavement Structure

The modeled pavement has the typical layer stratigraphy of a Swiss highway pavement, as presented in Figure 2. It consists of four bonded asphalt courses and a granular mixed foundation layer (unbound), which is responsible for transmitting the stresses to the subgrade. Focusing on the bonded layers of the pavement, different types of mixtures are placed at different depths, each with specific characteristics and functions.

The Semi-Dense Asphalt (SDA) mixture is used for the wearing course, characterized by a noise-absorbing feature (to reduce tire–road noise). Additionally, modified bitumen of type PmB CH-E 45/80-65 is employed to further enhance the viscoelastic behavior of the mixture.

High modulus asphalt mixtures (EME, French: Enrobés à Module Elevé) are used for the base and the binder layers to provide high durability with heavy vehicular traffic, including trucks. The high stiffness mixtures reduce the pavement’s thickness and maintenance costs [20]. The stiffness of these mixtures is achieved with graded “hard” binders (15/25 or 10/20) and selected (high performance, maximum packed) aggregates. This, combined with the correct binder dosage, results in a low void content (1% to 6%), which optimizes stress transmission. However, the use of very hard binders requires special attention in achieving optimal compaction, which is why they are only justified for pavements subjected to high dynamic loads. Specifically, high modulus bituminous mixtures, as specified by the SN 640 431-1-NA standard [21], serve the dual purpose of increasing fatigue resistance and limiting the accumulation of permanent deformations in the pavement.

Finally, the pavement structure ends with a fourth bitumen-bound layer that serves as the hot mix subbase. It is composed of the AC F 22 90% RA mixture, which is characterized by its composition of 90% recycled bituminous aggregate by weight.

3.2. Scenarios Definition

3.2.1. Layers Characteristics

For a thorough analysis of the model, various scenarios, each of which represents a specific geometric and mechanical condition of the pavement, are considered. These scenarios are defined according to the parameters listed in

Table 1. Modeling parameters.

Layer Composition	Thickness Range According to VSS 40 430–40 436 [18] [mm]	Selected Thicknesses (t) [mm]	Stiffness Modulus (E) [MPa]			Poisson’s Ratio (ν) [-]
			T [°C] = 10	T [°C] = 20	T [°C] = 30
SDA 8–12 (Wearing Course)	25 to 40	30 and 40	4000	2500	700	0.35
AC EME 22 C1 (Binder Layer)	80 to 120	80, 90 and 100	14,000	8000	3000	0.35
AC EME 22 C2 (Base Layer)	80 to 120	100	17,000	12,000	6500	0.35
AC F 22 90% RA (Sub-base Layer)	60 to 150	100	15,000	10,000	5500	0.35
Granular Mixed Foundation	400	400	500	500	500	0.40
Subgrade	-	-	50	50	50	0.40

.

The thicknesses of the layers have been chosen in compliance with VSS 40 430–40 436 [18], which defines a reference range for each type of mixture. Based on this reference, thickness values are comparable with those commonly adopted in design practice, to keep the model as close as possible to real conditions. The 30 and 40 mm thick wearing courses and the 80, 90, and 100 mm thick binder layers are evaluated through this model. Since the stiffness moduli of the bituminous layers are dependent on temperature and loading frequency (at the same frequency, the higher the temperature, the lower the modulus) three different reference temperatures were selected.

The level of bonding between the bituminous layers is also considered as a further parameter to define the modeling scenarios, which can be defined by assigning a specific value to the Standard Spring Compliance (AK), with which the rational method models the behavior of the interface. This parameter is correlated to the horizontal reaction modulus (K) according to Equation (1).

$$AK=1/K$$

(1)

In this equation, K expressed in N/m³, or K [MPa/mm] = 1/(AK × 10⁻⁹).

It is worth emphasizing that this parameter varies only at the interface between the wearing course and binder layer, while it remains constant for all the others, where an intermediate condition of Partial Bonding is always considered (AK = 10⁻¹¹ [m³/N]). An exception is the Full Friction condition, which has been applied to all the interfaces of the road structure. Therefore, a total of three different bonding conditions, each of which corresponds to a specific value of the parameter AK, have been defined between the first two layers as presented in

Table 2. Level of bonding at the wear–binder interface.

Level of Bonding (Wear–Binder Interface)	AK [m3/N]
Full Friction	-
Partial Bonding	10−11
Full Debonding	10−7

.

It is important to specify that the Full Friction condition is purely ideal, where the layers are perfectly bonded to form a fully monolithic system. This eliminates relative displacements along the interface ensuring improved short-term and long-term performances. Also, it optimizes the overall mechanical behavior of the pavement, and, finally, it ensures extended durability. However, this ideal condition differs from what can actually be achieved in the field or in the laboratory.

The Partial Bonding condition (AK = 10⁻¹¹ m³/N), although defined as “partial”, is characterized by a horizontal reaction modulus (K) large enough to minimize displacements like for the Full Friction (complete bonding), but unlike the latter, it is actually reproducible in the lab or in the field.

Lastly, the extreme condition of Full Debonding (AK = 10⁻⁷ m³/N) is related to a scenario where the pavement’s layers work independently of each other, resulting in significant relative displacements along the interface. In practice, this extreme condition never occurs thanks to the roughness of the surface of the layers in direct contact with each other providing always some residual shear strength of the interface.

3.2.2. Load Configurations

Two different load configurations have been modeled, reflecting the actual behavior of vehicle wheels traveling on the pavement, as presented in Figure 3.

Load configuration 1 involves the application of a vertical load (F_v) of 40 kN on a single wheel corresponding to an axial weight of 80 kN (standard axle load) (Figure 3a). Load configuration 2 involves the application of both a vertical load of 40 kN and a horizontal load (F_h) of 20 kN on a single wheel (Figure 3b). In this scenario, the horizontal force, mobilized in the direction of travel, models the accelerating force of vehicles (positive as per the selected reference system). This force arises from the bonding phenomenon at the interface between the vehicle wheel and the road pavement, influenced by the surface characteristics of the wearing course. Therefore, the horizontal force can be defined by Equation (2).

$$F_h=f·F_v$$

(2)

In this equation, a coefficient of friction for tire–pavement (f) equal to 0.5 [22] has been considered.

3.2.3. Pavement Stress–Strain Calculation Domain and Reference Points

The applied loads are distributed on the surface of the pavement over a circular contact area whose radius is inversely proportional to the tire inflation pressure. This aspect is crucial as loads distributed on the surface can lead to significant variations in the stress state within the pavement structure. To this end, a pressure of 0.8 MPa (8 bar) was applied, resulting in a contact area with a radius of 126 mm, as shown in Figure 4a.

The coordinates of the stress–strain calculation points refer to a Cartesian coordinate system {x, y, z}, where the x-axis runs along the direction of travel, the z-axis defines the depth from the surface, and the y-axis completes the right-handed coordinate system.

These reference points are distributed both on the pavement surface, and at the interface between the first two layers of the pavement, as shown in Figure 4b. This approach allows us to track the variation in stress and strain along the x-coordinate, making it possible to identify the maximum values. These maximum values were then selected as the primary parameters for comparison in the subsequent analyses.

3.3. Model Analysis

Considering the above-mentioned combinations of thickness and mechanical properties, 108 different scenarios were analyzed, and the results from the scenarios related to the 30 mm thick wearing course and the 80 mm thick binder course are reported below as examples. To carry out the simulation, as mentioned before, a rational calculation method based on the theory of elastic multilayer systems was employed, implemented using BISAR^® 3.0 (Bitumen Structures Analysis in Roads), a software developed by Shell (Shell Global Solutions, based in The Hague, Netherlands) and widely utilized in professional practice for the design and verification of flexible and semi-rigid pavements. The simulation results are presented in Section 5.1.

4. Laboratory Investigation

The shear resistance between bitumen-bound pavement layers is evaluated using the Leutner shear test, as per the European standard EN 12697-48 [17]. Proposed in 1979 by Professor Rolf Leutner, this test is the primary method used in Germany and Switzerland to assess the bonding level between layers. It is preferred over other methods because it can test both laboratory-made specimens and field samples, and it is faster to perform. The Leutner test was utilized in the experimental phases of the current study.

The Leutner apparatus features two metal semi-rings: a fixed lower ring and a movable upper ring, which applies controlled displacement to the bituminous layer, as demonstrated in Figure 5. The test focuses on the shear behavior of the interface, which must be accurately positioned to reflect a condition of simple shear.

In Figure 5, A is base body, B is sample support, C is lower shear ring, D is upper shear ring, E is upper body, F is guiding bar, and 1 shows a range max of ±20 mm. According to EN 12697-48 standard, the test typically uses cores with a diameter of 150 ± 2 mm and requires at least four samples for a representative series. The interface must be perpendicular to the core’s axis, with a maximum deviation of 5 mm. The jaws holding the core must match or slightly exceed the core’s radius by up to 2 mm.

Before testing, cores must be conditioned at 20 ± 1 °C for at least 4 h. The testing procedure involves assembling the Leutner frame, securing the core, and aligning the layer interface. The core is marked to indicate traffic or compaction direction. The press is started to record force and displacement data until the core breaks. The force–displacement curve is saved, and the layers are carefully separated post-test.

The standard mandates that the test must be completed within 15 min of removing the core from a temperature-controlled environment. The key measurements taken during the test include shear force, displacement, maximum shear force, the slope of the elastic linear segment, peak displacement, and the energies pre- and post-peak.

4.1. Laboratory-Prepared Specimen Experimentation

Specimens used in the experimentation were prepared in the laboratory by coring binder course’s slabs. These slabs (500 × 180 mm module with a total height of 100 mm) were compacted using a Large Roller Compactor (Figure 6a). Each slab comprises a 40 mm thick surface layer (SDA 8–12) compacted over a 60 mm thick binder layer (AC EME 22 C1). This choice ensured continuity with the bituminous mixtures discussed in Section 3, maintaining consistency throughout the experimental phase. The study focused on analyzing three different interface conditions as the primary variables: no tack coat in Hot-on-Hot (H/H) conditions, tack coat with residual bitumen of 200 g/m² (RB 200), and tack coat with residual bitumen of 400 g/m² (RB 400). Additionally, three levels of dynamic loading were applied to the specimens: 0 cycles, 30,000 cycles, and 60,000 cycles.

The loading cycles were applied to the compacted slabs using a wheel tracker (Figure 6b) before coring the specimens. The program also included creating one slab for each combination of variables (n = 9). From each slab, 3 core samples with a diameter of 150 mm were extracted (Figure 6c), resulting in a total of 27 specimens. Each specimen was then subjected to shear testing using the Leutner shear test method (Figure 6d) to determine the shear strength of the interface under different imposed conditions. The results of laboratory-prepared specimens’ experimentation are presented in Section 5.2.

4.2. Field Sample Experimentation

Alongside the investigation on laboratory-prepared specimens, the interface behavior between bituminous layers in an operational pavement was analyzed. Leutner shear tests were conducted on 40 cores extracted from a section of the A2 highway, a vital corridor crossing Switzerland from North to South and linking Southern and Northern Europe. This route experiences substantial daily traffic from heavy commercial vehicles, subjecting the pavement to significant stresses.

Tests were conducted on samples from an unloaded portion of the road section (emergency lane), assuming that the mechanical stresses on the examined material have been negligible in service. The coring of the 40 samples was concentrated along the emergency lane (South–North) near a rest area and its acceleration and deceleration lanes, between the Lugano Nord and Lugano Sud exits (chainage 24 + 400 km to 24 + 980 km), as depicted in Figure 7.

The cores are distributed along 20 pairs of sections, with each pair spaced 20 m apart. Within each section, the two cores are spaced approximately 1 m apart (Figure 8a). The coring process utilized a core drilling machine, which extracts samples using a rigid cylindrical steel motor-driven shaft. A core barrel at the end of the shaft penetrates the layers to obtain samples for the Leutner test (ϕ150 mm, variable length), as shown in Figure 8b.

Data from a 2010 investigation identified the mixtures used for the surface and binder layers of a specific road section. The surface layer was constructed with an AC MR 11 mixture (macro-textured asphalt), characterized by a grain size distribution with a high percentage of aggregates between 8 and 11 mm in diameter, providing a highly frictional macro-texture suitable for a wearing course. The underlying binder layer used the HMT 22 mixture, now known as AC B 22.

The average thicknesses of the surface and binder layers were found to be 35 mm and 60 mm, respectively. The samples were divided into odd-numbered samples (longitudinal failure direction) and even-numbered samples (transverse failure direction) to identify an anisotropy coefficient linking failure values in the longitudinal and transverse directions. After conditioning at 20 °C, the samples underwent failure testing through the Leutner Shear Test. The results of field sample experimentation are presented in Section 5.3.

5. Results and Discussion

This section summarizes and discusses the results of both the modeling and the experimental phases, linking them through the common theme of bonding between bituminous layers and proposing new strategies to control performance. A reference value has been established for optimizing pavement mix designs based on the materials’ mechanical characteristics. Additionally, using experimental data, a maximum displacement value at failure is defined. This, combined with the minimum tangential stress requirement from the VSS 40 430 standard (Leutner shear test), enhances the overall requirements for the shear behavior of the materials.

5.1. Outcomes of the Rational Calculation Method

5.1.1. Interface Shear Stress

Figure 9 represents the values of shear stress (τ_xz) generated at the wear–binder interface resulting from the rational calculation method, under the selected model parameters (different bonding conditions, loading configurations (1 and 2, as defined above), stiffness modulus, temperature, etc.).

According to the results of loading configuration 1, it emerges that the shear stress τ_xz at the wear–binder interface reaches its maximum at 125 mm from the vertical axis of the wheel, both in the front and rear parts, as shown in Figure 9a. This is because the application of vertical loads only leads to the symmetry of stresses, which is consistent with the theory of elastic multilayers. The symmetry is with respect to the origin of the chosen reference system, for equilibrium and for the sign convention of the shear stresses.

Regarding loading configuration 2, the maximum shear stress is also located at 125 mm from the vertical axis of the wheel, but here, it is localized in the front part only. This is due to the introduction of a horizontal component of load on the surface, in addition to the vertical one, resulting in an asymmetrical stress distribution that varies depending on the intensity and direction of application of the shear force. Moreover, the abscissa of the maximum is 125 mm for both the conditions of Full Friction and Partial Bonding. In the condition of Full Debonding, this abscissa is 63 mm (half of the contact radius) as shown in Figure 9b.

It can be concluded that the asymmetry generated by the introduction of the horizontal component of load on the surface leads to a decrease in tension in the rear part of the wheel and to an increase in the front part, especially in the conditions of Full Friction and Partial Bonding. Thus, the maximum shear stress is always within the contact area and is localized near the edge, especially in the presence of a high bonding between the layers.

Therefore, it emerges that, for both loading configurations, the shear stress τ_xz takes on higher values as the bonding between the wearing and binder layers increases, that is, as the AK parameter becomes smaller (for the same temperature and thicknesses). In this regard, it is crucial to underline that the shear stress that occurs in the Full Debonding condition is three orders of magnitude smaller (two in the presence of horizontal load) compared to the stress that develops in the other two conditions with progressively higher levels of bonding.

5.1.2. Interface Horizontal Stress

Figure 10 represents the normal stress (σ_h,i) developed near the interface between the two bituminous layers as calculated from the rational method, and under the selected model parameters (different bonding conditions, loading configurations, stiffness modulus, temperature, etc.).

The normal stress σ_h,i (symmetric as well as the tangential stress) developed near the interface between the two bituminous layers reaches its maximum value at 150 mm from the wheel axis in those scenarios including the vertical load only. In scenarios involving both vertical and tangential loads, the absolute maximum value is reached at −150 mm (behind the wheel) in Full Friction and Partial Bonding conditions, and at −125 mm in the Full Debonding condition. This is true for all temperatures.

From the comparison between the two load configurations, it emerges that, in the presence of the horizontal force component, there is a significant variation in the normal stress (σ_h,i), which increases by an order of magnitude as compared to the vertical load only.

The results also show that in Full Debonding condition, tensile stresses arise at all temperatures and for both load configurations. In Full Friction and Partial Bonding, however, compression dominates at all temperatures when the pavement is subjected to the application of a vertical load of 40 kN only. While with the addition of the horizontal load, even if bonding is present, the stresses show positive values (tensile). In this regard, it is important to note that the increase in the value of the maximum horizontal stress refers not only to an increase in absolute tensile values but also to a decrease in compressive values. It is also evident that the Full Debonding condition is the most unfavorable since it maximizes, for the same load configuration, the value of the tensile stress at the interface, thus promoting cracking and the subsequent deterioration of the pavement.

From

Table 3. Percentage variations of the horizontal stress at the interface.

Bonding Condition	T [°C]	t binder layer = Variable, t wearing course = 40 mm		t binder layer = 100 mm, t wearing course = Variable
		D[σh,i] [%]		D[σh,i] [%]
		Loading Config. 1	Loading Config. 2	Loading Config. 1	Loading Config. 2
Full Friction	10	2.74	10.40	−5.49	−477.28
	20	2.39	15.22	−6.33	−196.23
	30	1.49	25.78	−17.48	−55.37
Partial Bonding	10	1.99	14.29	−0.13	−342.49
	20	1.78	81.65	−2.12	−98.85
	30	1.20	14.10	−14.28	−47.53
Full Debonding	10	−1.91	−0.25	2.83	−32.33
	20	−1.34	−0.21	2.09	−32.76
	30	−0.25	−0.12	−5.23	−33.12

, it is possible to highlight several observations regarding the sensitivity of the interface horizontal stress to the variation of the thickness of the bituminous layers of pavement.

The mean percentage variation of stress with increasing thickness of the layers (D[σ_h,i]) is considered as an indicator of sensitivity, which can be computed through Equation (3).

$$D\left[{\mathsf{\sigma }}_{h,i}\right](\%)=100·{\displaystyle \frac{{\mathsf{\sigma }}_{mean}}{{\mathsf{\sigma }}_{ref}}}\left(t_{wearingcourse,}{t}_{binder}\right)$$

(3)

It is evident from Table 3 that there is a greater sensitivity to a change in the thickness of the wearing course layer from 30 to 40 mm, particularly when both vertical and horizontal loads act simultaneously on the pavement. In this case, it can be observed that the stress at the interface decreases in all the modeled conditions of bonding and at all temperatures as the thickness of the wearing course increases. From Table 3, the high sensitivity of σ_h,i with increasing thickness of the binder layer at a temperature of 20 °C needs to be recognized in the condition of Partial Bonding (+81.65%). In the condition of Full Debonding, however, the stress decreases at all temperatures and in both loading configurations.

5.1.3. Surface Horizontal Stress

The horizontal normal stress developed at surface (σ_h,s) reaches its maximum value in tension at 138 mm from the center of the imprint contact area symmetrically for loading configuration 1, and asymmetrically for configuration 2. In the latter, the maximum horizontal stress is localized 138 mm from the wheel axis in the rear for each bonding condition. Figure 11 compares the behavior of the maximum horizontal stress (σ_h,s) at the surface for the two considered loading configurations.

From the graphs, it can be observed that the surface horizontal stress reaches the highest values in the Full Debonding condition, in both loading configurations. Moreover, it is evident that the stress values, for a given bonding condition, are greater when there is a horizontal force at the surface. In fact, as far as loading configuration 2 is concerned, the resulting stresses are tensile at the interface for every bonding condition and at all temperatures. On the other hand, with the first loading configuration both in Full Friction and Partial Bonding, the horizontal stress is tensile only at 30 °C, while at other temperatures, tension is never reached. This is different in Full Debonding, where tensile stresses are present at every temperature.

It is noteworthy that, when subjected to the vertical load only, the stress magnitude rises with temperature, irrespective of the bonding condition. However, when a horizontal force is present, the stress increases with temperature only in the Full Friction condition. In the other two cases, the stress reaches the peak value at 20 °C. Being this maximum value negligible, it is possible therefore to consider stress as constant with temperature under such conditions.

5.1.4. Interface Relative Sliding

The behavior of the relative sliding at the interface (ΔU_X) is evaluated at the point of maximum shear stress, where the sliding itself is also maximum (using the horizontal reaction modulus K). Firstly, it is important to highlight that this quantity is defined by Equation (4).

$${∆U}_x=U_x\left(wearingcourse\right)-U_x\left(binder\right)$$

(4)

In this equation, ΔU_X represents the magnitude of displacement at the interface of the upper layer (wearing course) relative to the underlying one (binder layer).

Figure 12 represents the values obtained for the relative sliding at the interface, under the adopted model parameters (different bonding conditions, loading configurations, stiffness modulus, temperature, etc.).

According to Figure 12, the displacement is higher as the bonding level decreases, at the same temperature. Indeed, in Full Friction, there is no relative sliding (ΔU_X).

Moreover, for the same bonding level, temperature, and layer geometry, larger relative sliding is generated by the horizontal force of 20 kN, in agreement with the increase in the shear stress τ_xz for this loading configuration. According to the results, it can be noted that:

• For the vertical load only, the values obtained in Full Debonding are one order of magnitude higher than those obtained in Partial Bonding;
• For the horizontal force in Partial Bonding, the values remain of the same order of magnitude as those obtained with the first loading configuration;
• With Full Debonding, on the other hand, relative displacement is 10 times larger compared to that obtained from the application of the vertical load only.

Therefore, it is confirmed that the presence of a horizontal load, such as that caused by the braking or the accelerating forces of traveling vehicles, can significantly impact the relative slippage between the layers, especially when the bonding at the interface is on the lower end of spectrum.

5.2. Results of Experimentation on Laboratory-Prepared Specimens

Figure 13 summarizes the results obtained from the Shear Bonding Tests (SBT), under three different load cycles and interface condition scenarios, in terms of maximum breaking force (F_SBT,max) and horizontal reaction modulus (K_SBT), which corresponds to the slope of elastic linear portion of the failure curve.

It was observed that the maximum breaking force (shear breaking force according to the Leutner test) increases with the load cycles under the same interface condition, indicating an improved bonding strength between the surface and underlying layers over time. This increase is likely due to post-compaction from vehicle traffic. However, after this period, the breaking force values of the cores show a decline over time.

The necessity of a tack coat when laying down bitumen pavement layers can be influenced by temperature. In hot-over-hot paving conditions, the bituminous materials are more fluid and have lower viscosity, which can potentially improve bonding between layers without a tack coat. However, this does not mean that a tack coat is unnecessary. Even in hot conditions, a tack coat can ensure a stronger and more reliable bond, minimizing the risk of layer separation under traffic loads. According to Graziani et al. [23], the interlayer shear stress (ISS) values are significantly improved with a tack coat at lower paving temperatures (5 °C), highlighting its importance in ensuring proper layer adhesion, although the improvement is less significant at higher paving temperatures (40 °C). In fact, in low paving temperatures, bituminous materials become stiffer and less adhesive, increasing the risk of poor bonding between layers. In such conditions, a tack coat is crucial to compensate for the reduced bonding capacity.

This trend suggests that the breaking force tends to increase with t load cycles initially, without a sudden decrease. This result is plausible given the differences between laboratory testing conditions and actual site conditions. Laboratory conditions cannot simulate the long-term aging of the binder caused by environmental exposure. Over time, this aging in actual pavements could lead to a gradual decrease in the interface’s resisting force. Additionally, a high number of vehicular loads, combined with aging, can cause rutting and fatigue failure, promoting delamination between pavement layers and thus reducing shear resistance.

The test results indicate that a residual bitumen content of 200 g/m² (RB 200) is optimal for bonding. Under the same load cycles, this condition has achieved the highest value of the breaking force. An exception is noted for specimens not subjected to load cycles, where an interlayer content of 400 g/m² yields a slightly higher shear force. However, the differences in breaking forces between these conditions are negligible.

For specimens subjected to cyclic traffic loading, the horizontal reaction modulus varied significantly, ranging from 0.45 MPa/mm to 1.72 MPa/mm. The mean values, as shown in Figure 13b, remain relatively constant to around 1 MPa/mm, with an overall mean value of 1.03 MPa/mm. This corresponds to a bonding parameter AK value of approximately 10⁻⁹ m³/N. As discussed in Section 3, this AK value indicates a Partial Bonding condition between the layers.

Although the K_SBT parameter indicates the bonding level between layers by defining the slope of the linear portion of the failure curve and the AK parameter used in rational calculations, it does not provide information about the peak force and displacement values from experimental tests. Mechanistic methods are based on simplifying assumptions that approximate the behavior of pavement materials to be elastic linear.

However, laboratory mechanical tests reveal the elastoplastic nature of bituminous mixtures and their interfaces. To further analyze failure behavior, the peak force and displacement values of the F-δ curve were considered to derive a new modulus value (K*), representing the average behavior of the interface at failure. This parameter is the slope of the line passing through the origin and tangent to the failure curve at its maximum point, as shown in Figure 14a.

Figure 14b presents a scatter plot of the peak force and displacement values from tested specimens under three different interface conditions. A linear regression was performed on these data, with the intercept at the origin, yielding a modulus K* value of 0.65 MPa/mm.

Figure 14b reveals significant dispersion in results due to the substantial variability in the K_SBT modulus and displacement, underscoring the variable nature of the K* modulus. This dispersion likely stems from the limited number of specimens and the suboptimal reliability of the chosen compaction method. Compaction using a Roller Compactor often struggles with achieving volumetric homogeneity in the produced slab, which, along with the manual brushing method used for spreading the interface material, could have significantly impacted the results’ variability.

Despite this dispersion, the K* value obtained from linear regression is considered a global reference parameter. Comparing the elastic modulus (K_SBT) with the peak modulus (K*), a slight deviation is observed, confirming that bituminous materials exhibit some ductility, where the mechanical behavior of the interface becomes plastic during failure.

5.3. Results of Experimentation on Field Samples

Figure 15 presents the average values of the test results conducted on all 40 core samples, in terms of maximum breaking force (F_SBT,max) and horizontal reaction modulus (K_SBT).

As seen in Figure 15, there are no significant differences observed between the values along the two different failure directions. In terms of peak force (F_SBT,max), values ranging from 32 kN (longitudinal direction) to 34 kN (transverse direction) were recorded, resulting in only a 6% difference between the two. Likewise, the value of horizontal reaction modulus (K_SBT), has a negligible 1% variation observed between the longitudinal and transverse failure conditions. The parameter stabilizes at a value of 1 MPa/mm, corresponding to a value of AK equal to 10⁻⁹ m³/N. Thus, once again, it is evident that the AK parameter takes a value comparable to that one found in the laboratory-prepared specimens.

Thanks to the distinct group of results, it is possible to relate the average values obtained to define a coefficient of anisotropy (C_a) that connects the data recorded along the two different cutting directions. To encompass the magnitude of the breaking force and the displacement, the average values of the accumulated energy during the test before reaching the peak condition (E_SBT) were used. In particular, the value of the coefficient C_a is defined by Equation (5).

$$C_a={\displaystyle \frac{E_L}{E_T}}$$

(5)

In this equation, E_L is the average pre-peak energy along the longitudinal direction; and E_T is the average pre-peak energy along the transverse direction.

Figure 16 shows that the accumulated energy before failure is similar in both directions: 60.95 Nm in the longitudinal direction and 64.20 Nm in the transverse direction, resulting in an anisotropy coefficient (C_a) of 0.95. This suggests that the failure behavior of the interface is not significantly affected by the direction of the tangential force. The C_a value does not indicate the influence of vehicular traffic, as the pavement section sampled was not subject to vehicle loads. However, it provides insight into the method of laying the surface layer over the underlying one. The near-unity anisotropy coefficient indicates that the surface and binder layers were laid in succession with a tack coat of bituminous emulsion, confirming that there was no milling of the wearing course and its reapplication. This is supported by the fracture surface characteristics of the lower layer, which lacks the longitudinal grooves typical of milled surfaces. According to the authors’ experience, a milled surface would show higher pre-peak energy values (or breaking force) in the transverse direction due to increased interlocking from the grooves, enhancing the overall shear strength of the interface.

5.4. Comparison between Modeling and Experimental Investigation

The numerical analysis discussed in Section 3, using a rational calculation method, provided valuable insights into the characterization of the connection between the bituminous layers of pavement. It was observed that the choice of bonding condition significantly influences the overall stress–strain state of the superstructure, highlighting the need to test the bond strength of bituminous materials for informed design decisions. The analysis revealed that the maximum tangential stress at the interface at 20 °C under the most severe condition is 0.40 MPa, which is significantly lower than the 1.43 MPa observed in laboratory-prepared specimens and below the Swiss standard VSS 40 430 requirement of 0.85 MPa.

However, the numerical modeling and experimental investigation do not align closely enough for a direct comparison of calculated and measured quantities. The numerical model assumes a constant load and an elastic linear, homogeneous, isotropic material, whereas the Leutner test involves a gradually applied tangential force until failure, where the material exhibits plastic behavior.

Despite these differences, the calculation software remains a valid tool for simulating the behavior of layered superstructures like road pavements. The Leutner test, while conservative, provides shear strength in the “plane shear” mode like in situ shear strength tests on soils. Although it cannot fully simulate the operational behavior of a pavement, the modeling of the superstructure helps in understanding the experimental results.

The AK parameter (or K_SBT) obtained from tests allows, finally, for the accurate and unequivocal identification and classification of the bonding level between layers.

5.5. Design Reference and Limit Displacement

While tack coats positively impact shear resistance, they are insufficient for achieving optimal bonding conditions. Modeling results indicate that the best bonding conditions enhance the superstructure’s durability. However, the experimental analysis suggests using the AK parameter obtained from specimens as a reference value in road design. This approach avoids overestimating pavement service life by assuming overly optimistic AK values or designing under (unrealistic) Full Friction conditions. Overestimation could lead to early repairs under budget constraints. Therefore, it is here recommended to use an AK value of 10⁻⁹ m³/N between the wearing and binder layers during the design phase to ensure the superstructure modeling closely aligns with the materials’ actual performance.

Once the AK parameter design value was determined, focus shifted to meeting the minimum requirements specified by Swiss standards for the shear performance of bituminous mixtures. The standard mandates a minimum tangential stress of 0.85 MPa at failure, which corresponds to a shear force of 15 kN on the cylindrical specimens of ϕ150 mm tested in the Leutner test. Notably, the standard does not specify requirements for failure displacement or other control parameters like the elastic modulus K_SBT. Consequently, two key values were compared for reference:

• The minimum tangential stress (τ_min) equals to 0.85 MPa;
• The peak modulus (K*) equals to 0.65 MPa/mm.

The comparison focused on K* and τ_min because both are critical to determining the interface failure condition. Although the peak modulus is an independent parameter, it strongly correlates with the previously chosen design parameter. Specifically, in this case, the AK value of 10⁻⁹ m³/N (corresponding to K_SBT = 1 MPa/mm) aligns with a K* value of 0.65 MPa/mm. Following the investigation, it is anticipated and required that Leutner tests conducted on a significant number of samples to control shear performance will yield average failure values aligning along the slope of the K* peak derived from linear regression of experimental results. Combining this requirement with the minimum failure stress defined in Figure 17 establishes a specific limit displacement value (δ_lim). This value is determined by the intersection of the peak line (K*) and the horizontal line τ = τ_min = 0.85 MPa, resulting in δ_lim equal to 1.30 mm in this case.

This value has dual significance. Firstly, it represents the maximum displacement value required by the failure test to meet both imposed requirements. Specifically, with K* set as a design reference, if δ_SBT,max is less than δ_lim, it indicates that the failure stress is below the minimum required. Secondly, δ_lim is the maximum displacement achievable under the minimum tangential stress. Exceeding this displacement for the same force would result in a K* value smaller than that determined by the investigation, thereby negatively impacting the material’s operational performance.

6. Conclusions

This study combined numerical modeling and experimental investigation to explore the bonding at the interface of bituminous layers in road pavements. Using a rational calculation method, the research analyzed how the road superstructure response varies with design parameters (stiffness, layer thickness, load configuration) and bonding conditions. Modeling the response of a Swiss highway with high traffic, this study demonstrated that the interlayer connection (AK parameter) and external load configuration significantly influence pavement performance. Poor bonding leads to increased displacements and tensile forces, negatively impacting durability. The study concludes that accurate pavement design must consider interlayer bonding.

The three different conditions (Full Friction, Partial Bonding, and Full Debonding) provide a comprehensive understanding of the extremes within which the actual performance of the pavement layers can vary. By exploring these boundary scenarios, the study can identify the worst-case and best-case performance outcomes and provide a range of expected behavior that helps in risk assessment and management. These extreme conditions serve as a foundation for future research by identifying key parameters that significantly influence pavement performance. Further studies can then focus on optimizing these parameters for better outcomes.

BISAR^® (Bitumen Structures Analysis in Roads), which is used for these simulations, is a widely recognized and validated software tool in the field of pavement engineering, frequently used by researchers and practitioners for its robust and reliable calculations. The material definitions and system assumptions within BISAR are based on extensive research and have been thoroughly validated in numerous studies (e.g., [24,25]).

The experimental investigation identified an optimal quantity of tack coat that enhances the connection between bituminous layers, as determined by the tangential force at failure from the Leutner test. Additionally, cyclic loading was found to improve the interlayer bond strength. The analysis of the horizontal reaction modulus (K) indicated that, regardless of the interface condition (tack coat presence/absence) and the number of loading cycles, the interface always exhibits Partial Bonding. This has led to defining a reference value of AK as 10⁻⁹ m³/N for design purposes. By examining the material’s failure behavior, a limit displacement value at peak condition was established, complementing the existing requirement of the minimum tangential force at failure for a more comprehensive design parameter.

Additionally, sampling from an inactive section of the Swiss A2 highway (i.e., emergency lane) revealed that the bonding condition remains partial, even in an environmentally aged and unloaded pavement. Furthermore, in this specific case, the direction of the tangential force causing specimen failure did not affect the Leutner test results.