Inclined Plane Shear Behaviour of a Recycled Construction and Demolition Material–Geocomposite Interface

Over the years, recycled construction and demolition (RC&D) materials have become useful and valuable construction materials. The applications of these materials have been increasing, and there are cases where they may be in contact with geosynthetics. This work, which is part of a broader investigation into the feasibility of replacing soils used in landfill final cover systems by RC&D materials, analysed the inclined plane shear behaviour of the interface between a fine-grained RC&D material and a drainage geocomposite. Inclined plane shear tests were performed with different vertical stresses and different compaction conditions (degree of compaction and water content) of the RC&D material. The friction angle at the RC&D material-geocomposite interface was estimated by two approaches: the standard and the one based on the Mohr–Coulomb failure envelope. The sliding mechanism of the RC&D material over the geocomposite was examined. For comparison with the RC&D material-geocomposite interface, the behaviour of the RC&D material under inclined plane shear movement was also characterised. The main findings of the work included: the friction angle (standard approach) at the RC&D material-geocomposite interface decreased by increasing the applied vertical stress and the RC&D material water content (with meaningful changes in the slinging mechanism in the latter case), and tended to increase by increasing the degree of compaction of the RC&D material; the behaviour of the RC&D material under inclined plane shear movement did not differ significantly from that of the RC&D material-geocomposite interface; the failure envelope approach was more conservative (smaller friction angles) than the standard approach.


Introduction
The construction sector is one of main producers of waste.In 2020, construction and demolition (C&D) waste represented more than a third of the waste produced in the European Union [1].C&D waste results from the construction, renovation, repair and demolition of buildings and infrastructure.Thus, its composition can be very diverse, including, for example, concrete, bricks, plastics, metals, wood or glass.Some C&D waste components have high value.Others, although of lesser value, can still be reprocessed into new products or materials.To enhance C&D waste recovery options, sorting procedures at source are essential.
To move towards a more sustainable construction, applications for C&D waste must be created, transforming these residues into valuable raw materials.Over the last years, possible applications of recycled C&D materials have been researched, including their use in the manufacture of binders [2,3], mortars [4,5] and concrete [6,7], or in road construction as filling material [8,9].In some applications (for example, in road construction [10] or structural embankments [11]), recycled C&D materials may be in contact with geosynthetics.Accordingly, studies can be found in the literature on the interface properties between recycled C&D materials and geosynthetics 88 Page 2 of 16 [12][13][14][15], as well as on the mechanical damage that these materials induce to geosynthetics [16,17].The effect of long-term exposure of geosynthetics to recycled C&D materials has also been evaluated [18].
The existing studies on the interface properties between recycled C&D materials and geosynthetics are mostly based on direct shear tests [11][12][13][14].Pullout tests have also been used [14,15].By contrast, studies based on inclined plane shear tests, if any, are scarce.These latter tests are often used to characterise the interface properties between soil and geosynthetics, especially under low normal stresses.
Inclined plane shear tests have been carried out on soilgeosynthetic [19][20][21][22][23][24] and geosynthetic-geosynthetic [25][26][27][28] interfaces.In Europe, the inclined plane shear test is regulated by EN ISO 12957-2 [29].This test method, and the respective data processing, is not consensual, with criticisms arising over the years.Gourc and Ramirez [30] showed that more information could be drawn from the standard method and proposed a dynamics-based interpretation of the sliding mechanism.Afterwards, Briançon et al. [31] proposed a new procedure for inclined plane shear tests of geosyntheticsthe "Force Procedure".The authors claimed that the new procedure was more representative of real-world conditions.Later, Carbone et al. [32] proposed the "Unified Inclined Plane Procedure".The main claims against EN ISO 12957-2 [29] include: the standard method uses a static analysis for dynamic conditions; the standard method is non-conservative, overestimating the friction angles; and the interface shear strength in inclined plane tests cannot be characterised by a single parameter, as proposed by the standard method [30][31][32].Despite all the criticism, the EN ISO 12957-2 [29] method remains current, having been confirmed in 2021.
This work, which is part of a wider research study aiming to assess the feasibility of replacing soils used in landfill final cover systems by alternative materials such as recycled C&D materials, focused on the analysis of the inclined plane shear behaviour of the interface between a fine-grained recycled C&D material (hereinafter designated as RC&D material) and a drainage geocomposite.Inclined plane shear tests were carried out under different vertical stresses, and considering different placement conditions of the RC&D material: degree of compaction and water content.The friction angle at the RC&D material-geocomposite interface was estimated by two approaches-the standard, which is described in EN ISO 12957-2 [29], and the one based on the Mohr-Coulomb failure envelope.The behaviour of the RC&D material under inclined plane shear movement was also characterised.The main goals of the work included: (1) determine the influence of vertical stress and RC&D material compaction conditions on the behaviour of the RC&D material-geocomposite interface; (2) compare the behaviour of the RC&D material-geocomposite interface with that of the RC&D material under inclined plane shear movement; and (3) compare the friction angles obtained by the standard and failure envelope approaches.

Recycled C&D Material
The RC&D material was supplied by a Portuguese recycling plant and was mostly made up of particles with equivalent diameter smaller than 10 mm (Fig. 1).It had two main sources: demolition and rehabilitation of buildings, and cleanup of illegally dumped C&D wastes.The non-selected C&D wastes were treated to produce the RC&D material, which included sorting (removal of, for example, metals, plastics and wood), crushing and separation by grain size.

Geocomposite
The drainage geocomposite (GCD) was formed by a polymeric core (a geonet) covered by a nonwoven geotextile on both sides (Fig. 2).The geonet was made from high-density polyethylene and the nonwoven geotextile from polypropylene.The geocomposite had a mass per unit area (EN ISO 9864 [43]) of 873 g/m 2 and a thickness at 2 kPa (EN ISO 9863-1 [44]) of 5.25 mm.Its tensile strength (EN ISO 10319 [45]) was 24.7 kN/m in the machine direction and 16.0 kN/m in the cross-machine direction.The nonwoven geotextile had a mass per unit area of about 120 g/m 2 .

Equipment and Procedure
The inclined plane shear tests were carried out in a prototype equipment developed at the Faculty of Engineering of the University of Porto.The equipment was formed by the following main components: a tilt table, a motorized tilting mechanism, an upper box (which contained two wedges with inclination 1(H):2(V) inside), a lower box, a rigid steel support (when needed), a rigid steel plate, roller runners, a load application mechanism, a load cell, an inclinometer, a displacement sensor, safety sensors, a control board, a data acquisition system and a structure of steel profiles (Fig. 3).
The upper box (sliding box), which was filled with a 60 mm thick layer of RC&D material during the tests, was 300 mm wide by 300 mm long and had a depth of 80 mm (internal dimensions).On top of the RC&D material, a rigid steel plate (covering the inner area of the upper box) was placed, on which the vertical force was applied by weights.The upper box was fitted with rollers, which rested on roller runners during the tests.
The lower box had internal dimensions of 510 mm in length, 350 mm in width and 80 mm in height.Two types of inclined plane shear tests were performed, in which the lower box was filled in differently: (1) with a rigid steel support (tests on the RC&D material-GCD interface) or (2) with RC&D material (tests on the RC&D material under inclined plane shear movement).The inclined plane shear tests on the RC&D material-GCD interface followed the basic principles described in EN ISO 12957-2 [29] (Fig. 4a).The geocomposite (flat specimen 600 mm long and 350 mm wide, free from folds and wrinkles) was installed over the rigid steel support, which was placed inside the lower box.The upper box was installed on top of the geocomposite, but not in contact with it, with a gap between them of about 0.5 mm (the rollers of the upper box rested on the roller runners).The next step was to fill the upper box with RC&D material, namely with a layer 60 mm high.Then, a vertical force was applied over the entire RC&D material area.The inclination speed of the tilt table, which was horizontal when the test started, was 0.7º per minute.The test ended when the upper box reached a displacement of 50 mm.Data collected included elapsed time, vertical force, displacement and inclination, with readings being taken every 4 s.
The RC&D material was used dry and with an optimal water content, in each case with two degrees of compaction (DC): 55% (loose material) and 70% (moderately dense material).Limitations on the maximum inclination of the equipment (about 45º) made it impossible to test the RC&D material with higher compaction.Three different vertical stresses (σ v ), namely 5, 10 and 25 kPa, were considered to simulate the normal stress levels commonly experienced in the intended application (final covers of landfills).For each combination of water content (w), degree of compaction and vertical stress, the inclined plane shear tests were performed in triplicate.Thus, a total of 36 tests were carried out on the RC&D material-GCD interface.The geocomposite was tested on the machine direction (direction parallel to the movement of the upper box).
The tests on the RC&D material under inclined plane shear movement were slightly different from those described above for the RC&D material-GCD interface.In this case, the lower box was filled with RC&D material (layer with a height of 80 mm) (Fig. 4b).Then, the upper box was placed over the lower box with its rollers resting on the roller runners, and, as before, filled with RC&D material (layer with a height of 60 mm).Finally, a vertical force was applied and the test started.Test speed, end of test criteria and data collected were the same as those described for the tests on the RC&D material-GCD interface.The RC&D material was used dry and with an optimal water content, but only with one degree of compaction: 70%.Vertical stresses of 5, 10 and 25 kPa were again considered.Tests were carried out in triplicate for the different combinations of water content and vertical stress, making a total of 18 tests on the RC&D material under inclined plane shear movement.
The experimental conditions of the inclined plane shear tests are going to be identified by codes, namely Dry-DC55, Dry-DC70, Opt-DC55 and Opt-DC70.In these codes, Dry and Opt represent the water content of the RC&D material (Opt stands for optimal water content), while DC55 and DC70 refer to its degree of compaction.In some cases, a number (5, 10 or 25) will be added to previous codes (for example, Dry-DC55-5).This number indicates the applied vertical stress in kPa.

Data Processing
Data processing was carried out in two ways: (1) by the standard approach, which is described in EN ISO 12957-2 [29] and (2) by the approach based on the failure envelope.The latter was used, for example, in Izgin and Wasti [19], Wasti and Özdüzgün [25] or Lopes et al. [23].
As defined in EN ISO 12957-2 [29], the slipping angle (β 50 ) is the angle at which the upper box reaches a displacement of 50 mm.To determine the friction angle (ϕ sg,50 when considering the RC&D material-GCD interface, or ϕ 50 when considering the shear strength of the RC&D material under inclined plane shear movement), it is necessary to determine the normal stress (σ n ) and the shear stress (τ) when the tilt table is inclined at angle β 50 (force diagram is shown in Fig. 3).These parameters can be calculated by Eqs. 1 and 2, respectively. (1) A Where: F v is the vertical force; A is the contact area, which was constant (0.09 m 2 ) in all tests; and f (β50) is the force required to prevent the upper box, when empty, from moving when the tilt table is inclined at angle β 50 .The friction angle (ϕ sg,50 or ϕ 50 , as defined above) was obtained based on Eq. 3.
The approach based on the failure envelope also used the normal and shear stresses determined by Eqs. 1 and 2, but in a different way.The Mohr-Coulomb failure criterion considers a linear relationship between τ and σ n , which can be represented by Eq. 4.
Where: c is cohesion and δ is the friction angle.To obtain the failure envelopes, plots of shear stress as function of normal stress were performed.Each plot included 9 data points, corresponding to the normal and shear stresses determined in triplicate for each vertical stress (5, 10 and 25 kPa) used in the inclined plane shear tests.The straight lines that best fit the data points were calculated.In those lines, tan δ corresponds to the slope of the failure envelope and c to the intercept with the τ axis.

Physical and Environmental Characterisation of the RC&D Material
The RC&D material was mainly made up of fine particles, as illustrated in Fig. 5.The values found for its D 10 (particle size corresponding to 10% passing), D 50 (particle size corresponding to 50% passing) and D max (maximum particle size) were respectively 0.008, 0.39 and 10 mm.The presence of such small particles made more difficult to determine the constituents of the RC&D material.The EN 933-11 [34] applies to coarse recycled aggregates, only allowing the classification of particles larger than 4 mm.Those particles represented only 13.7% by mass of the RC&D material.
The RC&D material resulted from the treatment of nonselected C&D wastes (as described in the 'Recycled C&D Material" section), corresponding to the fine grained fraction.Thus, it was expected to have a very heterogeneous composition.Within the particles larger than 4 mm, 51.1% were classified as R u (unbound aggregate, natural stone and hydraulically bound aggregate).R c (concrete, concrete products, mortar and concrete masonry units), R b (clay masonry units and calcium silicate masonry units) and R s (soils) were also present in significant amounts, respectively 21.9%, 16.1% and 9.5%.The percentages of R g (glass), R a The unclassified particles (86.3% by mass of the RC&D material) would also fall into the above categories.As previously mentioned, one of the sources of the RC&D material was the cleanup of illegally dumped C&D wastes.Therefore, many of the unclassified particles may be soil.
The Proctor compaction test (EN 13286-2 [35]) allowed to conclude that the RC&D material had a maximum dry unit weight of 18.8 kN/m 3 and an optimal water content of 12%.Therefore, RC&D material with a water content of 12% was used in the inclined plane shear tests (in addition to the dry RC&D material).
The use of RC&D materials in construction works requires that they have adequate environmental performance, not releasing excessive amounts of hazardous substances into the environment.Therefore, their leaching behaviour must be satisfactory, complying with the limits defined by the Council Decision 2003/33/EC [46] for the acceptance of waste at inert landfills.The results obtained in the chemical analysis of hazardous substances in the RC&D material leachate, which had a pH of 8.3 at 20 ºC, can be found in Table 1.Table 1 also includes the limits for acceptance at inert landfills (leaching at liquid to solid ratio of 10 L/kg) [46], which are shown in square brackets.
As can be seen in Table 1, the concentration of most hazardous substances was below the limits defined by the Council Decision 2003/33/EC [46] for the acceptance of waste at inert landfills.There were only two exceptions: the sulphate and TDS values.
The sulphate concentration was 2.3 times higher than the admissible value (1000 mg/kg).However, the Council Decision 2003/33/EC [46] includes a note on sulphate, mentioning that if its content is above the limit, the residue can still be considered as compliant with the acceptance criteria as long as the leaching does not exceed 6000 mg/kg at a liquid to solid ratio of 10 L/kg.Thus, taking into account this note, even exceeding the 1000 mg/kg limit, the sulphate concentration in the leachate can be considered acceptable.
The TDS concentration slightly exceeded the admissible value (4000 mg/kg), namely by about 13%.The Council Decision 2003/33/EC [46] has no tolerance for TDS (as it does for sulphate), but mentions in a note that its value can be used as an alternative to sulphate and chloride values, which may indicate that TDS determination is not mandatory.Furthermore, it should be noted that the TDS concentration was well below the admissible value for non-hazardous residues, which is 60,000 mg/kg.
Overall, even with the two exceptions, which may be considered of low relevance as discussed, the environmental performance of the RC&D material can be considered as acceptable.Thus, the excessive release of hazardous substances into the environment should not be a problem, enabling the use of the RC&D material in construction works.

RC&D Material-GCD Interface
The inclined plane shear test results (β 50 , σ n , τ and ϕ sg,50 ) obtained for the RC&D material-GCD interface can be found in Table 2. Within each experimental condition, the β 50 values (obtained in triplicate) were not very different from each other.The coefficients of variation associated with β 50 varied between 0.2% (Dry-DC70-25) and 5.0% (Dry-DC55-10).These values indicated a low scatter between the results, showing a relatively good repeatability.
The displacement-inclination curves obtained for the RC&D material-GCD interface are illustrated in Fig. 6, showing that the sliding mechanism had some differences between tests (the differences will be addressed later).In the curve identification codes, (I), (II) and (III) represent, respectively, replicate I, II and III.Most curves showed the typical three phases reported by Gourc and Ramirez [30]: phase 1 (static) -the upper box was motionless until reaching critical angle β 0 ; phase 2 (transitory) -after β 0 , the upper box moved as the inclination increased; phase 3 (non-stabilised sliding) -at β s (inclination angle for non-stabilised sliding), the upper box underwent non-stabilised sliding.The movement observed in phase 2 was of gradual sliding, mostly progressive.In some cases, for example Dry-DC55-25, a stick-slip sliding has occurred.As shown in Fig. 6, in most cases, β s corresponded, or was very close, to β 50 .
The average friction angles ϕ sg,50 obtained for the RC&D material-GCD interface according to the standard approach are shown in Table 3.As for β 50 , the scatter between the ϕ sg,50 values was low, with coefficients of variation ranging between 0.2% (Dry-DC70-25) and 4.8% (Dry-DC55-10).To quantitatively assess the differences between the displacement-inclination curves (shown in Fig. 6), two additional parameters were determined, whose average values can also be found in Table 3.These parameters were: (1) the initial friction angle (ϕ sg,1 ), defined for a small displacement of the upper box -namely 1 mm, as suggested by Pitanga et al. [22].The calculation of ϕ sg,1 was identical to that of ϕ sg,50 but used β 1 (inclination when the displacement of the upper box reached 1 mm) in Eqs. 1 and 2 instead of β 50 .In Eq. 2, f (β50) was also replaced by f (β1) , the latter parameter corresponding to the force required to prevent the upper box (when empty) from moving when the tilt table is inclined at angle β 1 ; and (2) the displacement of the upper box when the non-stabilised sliding started (d s ).The criterion for estimating d s was in accordance with Pinho-Lopes and Lopes [24].As can be seen in Fig. 6, when the RC&D material was used dry, there was a clear distinction between the transitory and non-stabilised sliding phases, making it relatively easy to estimate d s .When the RC&D material was compacted with optimal water content, the sliding mechanism tended to change (differences will be discussed later).In these cases, it was more difficult to distinguish the gradual sliding of the transitory phase from the non-stabilised sliding (sudden final sliding).This made estimating d s harder and probably also less accurate.
The scatter between the ϕ sg,1 values was higher than that observed for ϕ sg,50 , although overall still not excessively high.In most cases, the coefficients of variation were below 10%: between 1.1% (Dry-DC70-10) and 9.5% (Opt-DC55-25).In only two cases they were above 10%, namely in Opt-DC55-10 (10.6%) and Dry-DC55-5 (18.2%).The d s values had, in most cases, a high dispersion, as can be concluded from the high standard deviations shown in Table 3.
The coefficients of variation associated with d s were below 10% in only two cases: Dry-DC55-10 (4.3%) and Dry-DC70-10 (3.6%).The estimation of d s was not possible in some cases since the upper box kept sliding gradually until reaching a displacement of 50 mm.All those cases were observed when the RC&D material was compacted at its optimal water content.
In addition to the average values of ϕ sg,50 , ϕ sg,1 and d s (and their respective standard deviations), Table 3 also shows the difference, in degrees, between the friction angle ϕ sg,50 and the initial friction angle ϕ sg,1 .This difference can contribute to characterise the displacementinclination curve and, thus, the sliding mechanism.
The experimental program allowed evaluating the influence of three parameters on the behaviour of the RC&D material-GCD interface: vertical stress, and degree of compaction and water content of the RC&D material.This evaluation will be carried out in the following subsections, where the results presented in Tables 2 and 3 and Fig. 6 will be discussed in more detail.

Influence of Vertical Stress
As can be seen in Table 3, increasing the vertical stress resulted in smaller ϕ sg,50 at the RC&D material-GCD interface.This behaviour was observed in all experimental conditions: Dry-DC55, Dry-DC70, Opt-DC55 and Opt-DC70.When dry RC&D material was used, increasing the vertical stress from 5 to 25 kPa led to reductions in ϕ sg,50 of 9.4% (degree of compaction of 55%) and 15.7% (degree of compaction of 70%).Making an identical comparison for RC&D material compacted at its optimal water content, the reductions were not as pronounced-4.8%(degree of compaction of 55%) and 9.5% (degree of compaction of 70%).As illustrated, for both moisture conditions (dry and optimal), the percentage reductions in ϕ sg,50 resulting from increasing the vertical stress from 5 to 25 kPa were more marked when the RC&D material had a degree of compaction of 70% (compared to 55%).The behaviour of ϕ sg,1 with increasing vertical stress differed from that of ϕ sg,50 .In cases where the RC&D material had a degree of compaction of 55% (Dry-DC55 and Opt-DC55), the ϕ sg,1 did not change very markedly as the vertical stress increased.However, when the degree of compaction was increased to 70% (Dry-DC70 and Opt-DC70), a decrease in ϕ sg,1 was observed with increasing vertical stress.The d s values were not significantly influenced by the increase in vertical stress in any of the cases.The difference between ϕ sg,50 and ϕ sg,1 underwent some variation with increasing vertical stress.The variation was, in most cases, of relatively minor importance.
The dependency reported here between ϕ sg,50 and the vertical stress is in agreement with results of previous works.Lopes et al. [23] observed an identical behaviour for three soil-geosynthetic interfaces.The authors used a well-graded residual soil from granite with different water contents and three geosynthetics: a geocomposite, a geotextile and a geogrid.Focusing on the interface between a silty sand and different geosynthetics (geotextiles, reinforced geomats and a geomembrane), Pitanga et al. [22] reported that, for all geosynthetics, the friction angle decreased with the increase of normal stress.
The decrease of ϕ sg,50 with increasing vertical stress can be explained by Fig. 7.As the vertical stress increases, the normal stress at the interface also increases.If the shear strength for normal stress equal to zero is not null (or considering that the failure envelope for very low values of normal stress is not linear), it is easily comprehensible that ϕ sg,50 tends to decrease with the increase of vertical stress.

Influence of the Degree of Compaction of the RC&D Material
The increase of the degree of compaction of the RC&D material from 55 to 70% tended to cause an increase in ϕ sg,50 at the RC&D material-GCD interface.As can be observed in Table 3, there was only one exception: when the RC&D material was compacted at its optimal water content and tested under a vertical stress of 25 kPa.Under these conditions, different degrees of compaction of 55% and 70% resulted in very close ϕ sg,50 values.When dry RC&D material was used, increasing the degree of compaction from 55 to 70% resulted in increases in ϕ sg,50 of 13.9% (vertical stress of 5 kPa), 13.2% (vertical stress of 10 kPa) and 6.0% (vertical stress of 25 kPa).In cases where the RC&D material had optimal water content, when the degree of compaction increased, the ϕ sg,50 increased by 4.0% and 3.8% (vertical stresses of 5 and 10 kPa, respectively) and decreased by 1.1% (vertical stress of 25 kPa).Taking into account the dispersions associated with the ϕ sg,50 values being compared, the 1.1% variation in ϕ sg,50 can be considered meaningless.As can be easily concluded, the increase of the degree of compaction had a more pronounced impact in ϕ sg,50 (higher percentage differences) when the RC&D material was used dry.Additionally, for both moisture conditions, the percentage differences in ϕ sg,50 resulting from using more compacted RC&D material tended to decrease when higher vertical stresses were used.Indeed, as quantified above, the percentage differences between the ϕ sg,50 were maximum at a vertical stress of 5 kPa (although with very close percentage differences at 10 kPa) and minimum at a vertical stress of 25 kPa.
In addition to affecting the ϕ sg,50 , the use of more compacted RC&D material also influenced ϕ sg,1 and d s (Table 3).The ϕ sg,1 increased by changing the degree of compaction of the RC&D material from 55 to 70%, showing that the upper box started to move latter.For example, for Opt-DC55-5 and Opt-DC70-5, average β 1 values of 21.6º (ϕ sg,1 of 24.8º) and 33.1º (ϕ sg,1 of 37.4º) were respectively obtained.The results obtained for d s indicated that the displacement of the upper box before the non-stabilised sliding tended to be lower when the RC&D material was more compacted.Lower d s values associated with smaller differences between ϕ sg,50 and ϕ sg,1 showed that the transitory phase (gradual sliding) had a shorter duration when the RC&D material had a higher degree of compaction.This behaviour can be easily observed by comparing, for example, Fig. 6a and b.
When the RC&D material was used more compacted, the interlocking between the particles was more pronounced, which means that the resistance to be overcome for the sliding of the upper box was higher.Additionally, the use of a higher degree of compaction may have resulted in a higher pressure on the geocomposite, which exhibited some compressibility, particularly above the openings of the polymeric core-where the nonwoven geotextile had no support underneath (Fig. 2).These features may have contributed to increase friction at the RC&D material-GCD interface.

Influence of the Water Content of the RC&D Material
The ϕ sg,50 values of the RC&D material-GCD interface decreased by changing the RC&D material water content from dry to optimal (Table 3).When the RC&D material had a degree of compaction of 55%, increasing the water content resulted in ϕ sg,50 reductions of 6.9% (vertical stress of 5 kPa), 4.7% (vertical stress of 10 kPa) and 2.2% (vertical stress of 25 kPa).When the RC&D material was more compacted (70%), the reductions were 15.0%, 12.6% and 8.8% at vertical stresses of, respectively, 5, 10 and 25 kPa.Two conclusions can be drawn from these results: (1) the effect of increasing the water content was more pronounced when the RC&D material was more compacted-higher percentage reductions in ϕ sg,50 ; (2) the percentage reductions in ϕ sg,50 tended to be less pronounced as vertical stress increased.
Like ϕ sg,50 , ϕ sg,1 also decreased when the RC&D material water content changed from dry to optimal.However, and contrary to what was observed in ϕ sg,50 , the percentage reductions in ϕ sg,1 tended to be less pronounced when the RC&D material was more compacted-exception when a vertical stress of 25 kPa was used, where the percentage reductions were very close: 30.1% and 30.5% for, respectively, degrees of compaction of 55% and 70%.In most cases, the percentage reductions in ϕ sg,1 were more pronounced than those in ϕ sg,50 .Thus, changing the RC&D material water content from dry to optimal had higher impact on ϕ sg,1 than on ϕ sg,50 .For example, comparing Dry-DC55-25 and Opt-DC55-25 (the most extreme case), the reductions in ϕ sg,1 and ϕ sg,50 were, respectively, 30.1% and 2.2%.
The d s was also impacted by changing the water content of the RC&D material from dry to optimal.As can be seen in Table 3, d s increased significantly when the RC&D material was compacted at its optimal water content.In some cases, the upper box still had a gradual sliding movement when it reached the displacement of 50 mm, not allowing estimation of d s .In most cases, changing the RC&D material water content from dry to optimal resulted in greater differences between ϕ sg,50 and ϕ sg,1 , which proved the longer duration of the transitory phase.Overall, the presence of water facilitated the slippage at the interface between the RC&D material and the geocomposite.The changes in the sliding mechanism can be easily seen in Fig. 6.
The results obtained here are partially in line with the work of Lopes et al. [23].The authors tested soil-geosynthetic interfaces with three soil water contents: dry, halfoptimal and optimal.When soil water content was changed Page 11 of 16 88 from dry to optimal, ϕ sg,50 did not noticeably decrease-in most cases, it even slightly increased.The use of the failure envelope approach for the determination of friction angles resulted in a different outcome.In the latter approach, higher friction angles were obtained when the soil was used dry, which is in consonance with the behaviour reported here for the RC&D material-GCD interface.

RC&D Material Under Inclined Plane Shear Movement
The results obtained for the RC&D material under inclined plane shear movement will be presented similarly to those of the RC&D material-GCD interface.Thus, Table 4 discloses the results (β 50 , σ n , τ and ϕ 50 ) obtained for each replicate, while Table 5 includes the average values of ϕ 50 , ϕ 1 and d s , and also the difference between ϕ 50 and ϕ 1 .The displacement-inclination curves obtained for the RC&D material under inclined plane shear movement can be seen in Fig. 8. Replicates I, II and III are identified in the curve codes by (I), (II) and (III), respectively.
As in the RC&D material-GCD interface, the β 50 values had a relatively low scatter, with coefficients of variation ranging from 0.6 (Opt-DC70-10) to 5.9% (Dry-DC70-5).Accordingly, the dispersion between the ϕ 50 values was also low: coefficients of variation between 0.5% and 5.4%.The average ϕ 50 values obtained for the RC&D material under inclined plane shear movement were very close to the average ϕ sg,50 values found for the RC&D material-GCD interface.The differences varied between -3.0% (Dry-DC70-10) and + 3.6% (Opt-DC70-5), illustrating the extreme proximity between ϕ 50 and ϕ sg,50 .Taking standard deviations into account, these small differences can be considered meaningless.Thus, all conclusions previously reported for the variation of ϕ sg,50 at the RC&D material-GCD interface, particularly in relation to the influence of vertical stress and RC&D material water content, are also valid for the variation of ϕ 50 in the RC&D material under inclined plane shear movement.The displacement-inclination curves presented in Fig. 8 show that, as for the RC&D material-GCD interface, the sliding mechanism differed between tests.The most meaningful differences resulted again from changing the water content of the RC&D from dry to optimal.As illustrated in Table 5, ϕ 1 tended to vary like ϕ 50 , that is, it decreased with increasing vertical stress and RC&D material water content.The variability associated with ϕ 1 was not exaggeratedly high, with only one coefficient of variation value above 8%, namely 12.5% (Opt-DC70-5).The difference between ϕ 50 and ϕ 1 was less pronounced when the RC&D material was used dry (maximum difference: 4.5º) than when it had an optimal water content (minimum difference: 11.5º).In the latter case, the transitory phase lasted longer.The variation of vertical stress did not result in very significant changes in the difference between ϕ 50 and ϕ 1 .In addition to changes in ϕ 50 and ϕ 1 , the use of RC&D material with optimal water content (instead of dry) caused an increase in d s .This behaviour had already been observed before in the RC&D material-GCD interface.Overall, and as a conclusive note, the results obtained for the RC&D material under inclined plane shear movement were not very different from those obtained for the RC&D material-GCD interface.

Comparison of the Standard and Failure Envelope Approaches to Estimate the Friction Angle
The friction angles when the upper box reached a displacement of 50 mm were obtained in two ways: by the standard approach, which is based on EN ISO 12957-2 [29] (results already presented in the previous sections) and by the approach based on the failure envelope.Regarding the latter case, the plots of shear stress as function of normal stress, as well as the respective linear regression curves, are As can be seen, both in Figs. 9 and 10, there was a good linearity between the values of normal and shear stresses (included in Tables 2 and 4), with coefficients of determination (R 2 ) close to 1 (R 2 varied from 0.9969 to 0.9997).The adhesion/cohesion values (intercepts with the τ axis) were very small, between 0.3 and 1.4 kPa.The friction angles obtained from the failure envelopes are shown in Table 6.Additionally, they are compared with those obtained by the standard approach.The friction angles presented in Table 6 for the standard approach resulted from averaging the ϕ sg,50 or ϕ 50 obtained under vertical stresses of 5, 10 and 25 kPa.
The friction angles estimated from the failure envelopes were, in all cases, lower than those obtained by the standard approach, which is easy to understand based on the schematic representation shown in Fig. 7.For the RC&D material-GCD interface, the difference ranged from − 3.8% (Opt-DC55) to − 14.7% (Dry-DC70), corresponding to the lowest and highest adhesion, respectively (Fig. 9).In the case of the RC&D material under inclined plane shear movement, the differences were -16.4% (Dry-DC70) and − 12.7% (Opt-DC70).Even compared with the minimum values of ϕ sg,50 or ϕ 50 (obtained under a vertical stress of 25 kPa), the friction angles estimated by the failure envelopes were smaller.Thus, it can be assumed that the failure envelope approach provided a more conservative estimate of the friction angle than the standard approach.
Finally, since different results were obtained by the two approaches, it becomes relevant to assess how the RC&D material conditions influenced the friction angles estimated from the failure envelopes.Increasing the degree of compaction from 55 to 70% did not significantly influence the friction angle at the RC&D material-GCD interface-variations of + 2.5% and − 3.1% when the RC&D material was, respectively, used dry and with optimal water content.Still at the RC&D material-GCD interface, changing the RC&D material water content from dry to optimal had a negligible effect on the friction angle for the degree of compaction of 55% (reduction of 0.8%) but caused a reduction of 6.3% when the degree of compaction was 70%.For the RC&D material shear strength under inclined plane shear movement, the increase of the water content resulted in a 7.8% reduction in the friction angle.Overall, comparing both approaches, the influence of the RC&D material conditions on the friction angle was less noticeable when the failure envelop approach was used.

Conclusions
This work characterised the inclined plane shear behaviour of a RC&D material-GCD interface.The influence of vertical stress and RC&D material conditions (degree of compaction and water content) on the interface behaviour was evaluated.Two approaches were used to estimate the friction angles−the standard [29] and the failure envelope.Tests were also conducted to characterise the shear strength of the RC&D material under incline plane shear movement.The main conclusions of the work are the following: • The displacement-inclination curves obtained for the RC&D material-GCD interface exhibited, in most cases, three phases: (1) upper box motionless with increasing inclination; (2) upper box with gradual sliding as the inclination increased (transitory phase); and (3) upper box with non-stabilised sliding.• The increase of vertical stress resulted in smaller friction angles (standard approach) at the RC&D material-GCD interface.• The friction angle (standard approach) at the RC&D material-GCD interface tended to increase by increasing the degree of compaction of the RC&D material from 55 to 70%.• The change of the RC&D material water content from dry to optimal caused reductions in the friction angle (standard approach) at the RC&D material-GCD interface.These reductions were more pronounced when the RC&D material was denser and tended to be less pronounced with the increase of vertical stress.The sliding mechanism also underwent relevant changes, with the transitory phase (gradual sliding) becoming longer when RC&D material with optimal water content was used.• The behaviour of the RC&D material under inclined plane shear movement was very similar to that of the RC&D material-GCD interface.• Despite the EN ISO 12957-2 [29] indication for determining the friction angle for a displacement of the upper box of 50 mm, in many cases the mobilization of shear strength occurred for considerable smaller displacements.This behaviour tended to be more notorious when the RC&D material was used with optimal water content.• The friction angles obtained by the standard approach were, in all cases, higher than those estimated from the failure envelopes.Thus, the failure envelope approach provided a more conservative estimate of the friction angle.• The influence of the RC&D material conditions (degree of compaction and water content) on the friction angle was less notorious when it was estimated by the failure envelop approach.
Finally, it should be noted that the frictional properties are highly influenced by the materials that are in contact.Thus, different behaviours from those reported in this work may be found when other soils (or RC&D materials) and geosynthetics are tested.

Fig. 4
Fig. 4 Inclined plane shear test setups: a tests on the RC&D material-GCD interface; b tests on the RC&D material under inclined plane shear movement.1: upper box (filled with RC&D material); 2: lower box (with rigid steel support installed); 3: geocomposite; 4: lower box (filled with RC&D material).Note: the image is not to scale n tan (bituminous materials) and X (materials other than the above) were very small, totalling 1.4%.

Fig. 5
Fig. 5 Particle size distribution of the RC&D material

Fig. 6
Fig. 6 Displacement-inclination curves obtained for the RC&D material-GCD interface: a σ v of 5 kPa and DC of 55%; b σ v of 5 kPa and DC of 70%; c σ v of 10 kPa and DC of 55%; d σ v of 10 kPa and DC of 70%; e σ v of 25 kPa and DC of 55%; f σ v of 25 kPa and DC of 70%

Fig. 8
Fig. 8 Displacement-inclination curves obtained for the RC&D material under inclined plane shear movement: a σ v of 5 kPa; b σ v of 10 kPa; c σ v of 25 kPa

Table 1
Chemical analysis of hazardous substances in the RC&D material leachate

Table 2
Results obtained for the RC&D material-GCD interface: slipping angle β 50 , normal stress, shear stress and friction angle ϕ sg,50 (standard approach)

Table 3
Results obtained for the RC&D material-GCD interface: friction angle ϕ sg,50 (average value; standard approach), initial friction angle ϕ sg,1 (average value), difference between ϕ sg,50 and ϕ sg,1 , and d s (average value) Note: in brackets are the obtained standard deviations * Result from 2 replicates (1 replicate with d s > 50 mm) Fig. 7 Schematic representation of the two approaches to estimate the friction angle

Table 6
Comparison of the friction angles estimated by the standard and failure envelope approachesNote: the friction angles presented for the standard approach resulted from averaging the ϕ sg,50 or ϕ 50 obtained under vertical stresses of 5, 10 and 25 kPa