Comparative Assessment of Single- and Double-Interface Shear Tests of MSW Landfill Liners Using Large Direct Shear Equipment

Geosynthetics form an integral part of engineered municipal solid waste (MSW) landfill lining systems, as they provide cost saving and technical benefits. Their introduction in lining systems, however, has presented new potential interfaces for shear failure especially on bases and on side slopes of engineered MSW landfills. In the laboratory, interface shear strength parameters used in slope designs of landfill lining systems are determined through single-interface testing configurations as per ASTM D5321-20 and D6243-20 standards. However, single-interface testing configurations do not provide a clear understanding of shear strength transfers among the lining components and does not aid in pin-pointing the interface facilitating failure especially when a multi-layered lining system has been proposed. Multi-interface testing configurations present an alternative that may address such limitations. But there is minimum information available on this type of testing as it is not stipulated in any official laboratory testing standard. This study was conducted to provide an understanding of the potential use of multi-interface testing configurations of MSW landfill liners, through identifying similarities and dissimilarities to the single-interface testing configurations. These interface tests were conducted on geosynthetics and geomaterials from a proposed lining system of a new landfill cell in South Africa.


Introduction
Due to strict environmental regulations imposed by government authorities coupled with public concerns over solid waste disposal practices, the demand for engineered municipal solid waste (MSW) landfills incorporating the use of engineered lining systems as pollution barriers has evolved in most parts of the world, South Africa included [1].An engineered MSW landfill is a sensitive geo-environmental barrier system that achieves a maximum disposal capacity through increasing landfill's side-slope gradient while limiting leachate migration on the base to ground and surface water reserves [2].The lining systems used in these structures usually constitute of a multi-layered band of numerous types of geosynthetics and geomaterials that are project and site-specific.The arrangement of these multi-layered lining systems depends mostly on the choice of materials that will achieve the intended functions.
Since their inception, geosynthetics have been used extensively in different geotechnical projects fulfilling various functions including acting as pollution or contaminant barriers, reinforcement, separation, filtration, drainage, and protection [3].Geosynthetics also perform important roles in the lining systems of an engineered landfill.However, as several layers of geosynthetics and geomaterials are being incorporated in MSW landfills' lining systems, their interaction becomes pertinent when considering shear failure mobilisation [4][5][6][7][8].For landfill design engineers, the process of designing side-slope and basal lining systems, as well as setting strict construction quality assurance (CQA) requirements is of utmost importance in ensuring the integrity of the entire MSW landfill during its construction, service life, and decommissioning [7,9].To establish the interface shear strength properties in the laboratory, single-interface shear strength 59 Page 2 of 16 tests of either soil-geosynthetic or geosynthetic-geosynthetic are typically conducted following ASTM D5321-20 and ASTM D6243-20 standards [10,11].Although this approach has been widely accepted for several reasons, such as confidence in the determined results, it poses an uncertainty which could lead to an overestimation of the interface shear strength parameters due to specimen confinement on the shearing device [12].
Previous studies found that the measured shear strength parameters at peak and at large displacements (LD) for single-and multi-interface configurations are comparable at low normal stresses [13].Consequently, it was determined that the weakest/critical interface through single-interface configuration on the side-slope lining system was the same as that obtained in the multi-interface shear testing configuration [14].Another study found that, as the applied normal stress increased in the singleinterface shear testing, the weakest interface changed at both peak and LD [15].Conversely, when similar interfaces were assessed using double-interface configuration, peak strength envelopes were determined to be the same.Interface shear strength parameters also proved to be similar at peak.In contrast, however, at LD, interface shear strength parameters obtained were very dissimilar, with the single-interface testing configurations achieving higher values than the double-interface testing configurations [15].Observations from these studies confirmed that tests conducted under the multi-interface testing configurations were still ambiguous as it was difficult to identify the interface facilitating failure, even though there was a slight degree of comparability to the single-interface testing configurations.
Interface shear strength parameters of landfill lining components in the laboratory rely on many factors, including testing conditions, the expertise of a test conductor, material types with their properties and most of all, the proposed liner arrangement [13,15].Depending on the design, some landfill lining arrangements may not permit conducting interface shear tests of more than three landfill lining components concurrently, i.e., multi-interface testing configuration.A double-interface shear strength test involving testing only three landfill lining specimens adjacent to each other could be suitably employed in such cases.Based on a review of available literature, this study was conducted to further investigate the comparability of interface shear results obtained from single-and double-interface testing configurations through a case study of the proposed lining system for a new landfill cell in the Western Cape Province in South Africa.This study was envisioned to benefit landfill designers and other stakeholders in understanding the implications of different testing configurations in the laboratory when establishing interface shear strengths of landfill lining components.

Proposed Lining System
The proposed landfill cell was classified as a general waste landfill (documented as a Type 2 waste stream as per South African regulations) with a net positive water balance during winter months due to Cape Town's seasonal rainfall pattern, thus creating the potential to generate significant amounts of leachate at certain times of the year.It was planned to be built in a "piggyback" fashion on top of a pre-existing 8 m thick waste deposit constituting of municipal solid wastes previously dumped without an engineered lining system.A combined multi-layered basal and side-slope lining system mainly divided into four segments, i.e. primary and secondary leachate collection, leak detection, and base preparation layers was proposed, as shown in Fig. 1 [16].

Geosynthetics
Based on the above arrangement, the two nonwoven geotextiles used were Bidim ® A5 (GTX-1) and Bidim ® A7 (GTX-2) shown in Fig. 2a and b, respectively.These geotextiles were locally manufactured from 100% polyester, thus forming high-performance nonwoven geotextiles whose continuous filaments were bonded by needle-punching.Their properties are shown in Table 1 as per the manufacturer's datasheet.Particularly for this study as intended by the design engineer, these geotextiles performed separation and protection functions.Furthermore, a high strength woven geotextile-based geogrid known as RockGrid ® PC (GTX-3) shown in Fig. 2c and d was used to perform a reinforcement function to assist in bridging areas of differential settlement within the underlying pre-existing waste deposit.
Two smooth-sided geomembranes, i.e. 2.00 mm thick HDPE geomembrane (GMB-1) and 1.50 mm thick LLDPE geomembrane (GMB-2) were also used due to gentle slopes on the base and on the side slopes of the proposed landfill Fig. 1 A proposed combined basal and side-slope lining system [16] cell as well as conforming to the requirements by local regulations.The LLDPE geomembrane with more flexibility than the HDPE geomembrane, allowed for its easy placement and anchorage.The geomembranes used in the proposed lining system were manufactured through a calendered flatdie extrusion process.Figure 3c shows the surface outlook of the geomembranes used.
Lastly, the cuspated drain (CD) (see Fig. 3a and b) used was a durable, heat-formed, multiple-cuspated synthetic sheet that provided a high capacity and non-restricted drainage path that was compatible with GMB-1 and GMB-2.The cusps of the CD were symmetrical; thus, providing a high resistance to crushing when exposed to heavy loads.Sand infill within the cusps was included to prevent any long-term creep strains over time.Wedged between the two geomembranes, the CD formed part of the primary leak detection system.

Geomaterials
This study utilised three geomaterials, namely: sand, gravelly sand (GS), and leachate collection stone (LCS) with unified soil classification system (USCS) of poorly graded sand (SP), poorly graded sand with gravel (SP with gravel) and poorly graded gravel (GP), respectively.The GS and LCS were products of crushed building waste materials and were abundantly available in the old section of the landfill.The use of these geomaterials was motivated by their availability, the need to re-use building waste, and in the case of the LCS, the high permeability, allowing for easy leachate collection from the proposed landfill cell.The mechanical properties are presented in Table 2 with images of samples shown in Fig. 4. All grading assessment were conducted in the Geotechnical Engineering Laboratory at the University of Cape Town.

Single Interfaces
In the single-interface testing configuration, only two scenarios were assessed, i.e. soil-geosynthetic and geosynthetic-geosynthetic interfaces.In the soil-geosynthetic setup as shown in Fig. 5, the geosynthetic component was clamped on the lower shear box and soil was compacted to the target density in the upper shear box, whereas in the case of geosynthetic-geosynthetic setup as can be seen in Fig. 6, the respective geosynthetic specimens were clamped on the upper and lower shear boxes.

Double Interfaces
In the double-interface testing configurations, three scenarios existed, namely soil-geosynthetic-soil, soil-geosynthetic-geosynthetic and geosynthetic-geosynthetic-geosynthetic interfaces.The modification to the single-interface testing configuration was an addition of a third soil/geosynthetic specimen in the setup.For the soil-geosynthetic-soil setup as shown in Fig. 7, the soil was first compacted in the lower shear box followed by geosynthetic specimen placement on top of the lower box without any clamping mechanisms and finally, the second soil specimen compacted on the upper shear box.A similar approach was used for the soil-geosynthetic-geosynthetic setup, except that a geosynthetic specimen was clamped on top of the lower shear  box to replace the soil as can be seen in Fig. 8. Lastly, for a geosynthetic-geosynthetic-geosynthetic setup also shown in Fig. 9, two geosynthetics were placed and clamped on one of the ends of the upper and lower shear boxes and one geosynthetic specimen was placed freely in the middle.

Large Direct Shear Equipment
The equipment used in this study was ShearTrac-III Large Shear Apparatus (LSA), also shown in Fig. 10.The device was fully automated to conduct consolidated drained direct shear tests for soil-geosynthetic interactions.This equipment has capabilities to conduct interface shear tests as stipulated by standards such as ASTM D3080-04, D5321-20 and D6243-20 [17].The equipment constituted of two shear boxes, i.e. the upper static shear box (305 × 305 × 100 mm) and the lower mobile shear box (405 × 305 × 100 mm).The equipment also contained two load cells, i.e. horizontal load cell for shear forces and vertical load cell for normal forces accompanied by their displacement transducers.The ShearTrac-III LSA was also complemented by a computer unit that hosted similar equipment controls as the machine with additional capabilities of showing real-time progress of the test using graphical displays.

Testing Procedure and Program
After the samples had been loaded onto the ShearTrac-III LSA housing unit, the complementary software in the computer unit known as the DS program was opened.Test conditions such as the amount of normal stress, the shearing displacement rate, maximum horizontal displacement, and consolidation times of the tested specimens were inserted.Finally, the system monitor window was opened along with the calibration summary window.The vertical and horizontal load cells were correlated by copying the system monitor offset values to the calibration summary.Tests were run for the side-slope and basal lining components at normal stresses of 150, 300 and 450 kPa following the anticipated field loading conditions resulting from the proposed landfill cell.In addition, a shearing rate of 1.0 mm/ min was used for all the lining components as recommended by the ASTM D5321-20 standard for non-cohesive materials.This ensured that any excess pore pressure build-ups that would interfere with the interface shear results are nullified.This study used the shearing displacement of 75 mm interpreted as large displacement (LD) for all interface shear tests.
Furthermore, a 30 min consolidation time was adopted in this study [15,17].This time was deemed sufficient for specimens' gripping surfaces to engage fully and establish an intimate interaction [13].Fresh tap water was added to the basin within the LSA equipment to hydrate samples, and a 30 min hydration time was adopted to simulate full hydration conditions in the field at the respective normal applied stresses.
After completing the consolidation and hydration phases, the shearing phase was initiated.A gap of 1 to 2 mm was created between the upper and lower shear boxes by slightly raising the reaction beams.This was necessary to prevent friction between the two shear boxes, thereby, allowing the measured frictional properties to be at the intended surfaces of lining components [17].While the applied normal force remained constant, the shear force was applied through a horizontal load cell on the lower shear box to establish shear stress.This enabled the lower shear box to move horizontally relative to the upper shear box at a shearing rate of 1.0 mm/min.The DS software embedded in the computer unit recorded all crucial results, and tests were automatically terminated when the horizontal displacement of 75 mm was reached.A total of 45 single-and double-interface shear tests were conducted on the lining components for the proposed landfill cell to achieve the set-out objectives for this study as shown in Table 3.

Data Processing
After completing the interface shear tests, the recorded results were further analysed using Microsoft Excel.The analysis included interpreting results by plotting normal stress-shear stress relationships to establish Mohr-Coulomb failure envelopes.Interface shear strength parameters were derived using Eq.1: where τ is the peak or LD shear strength; c a is the apparent adhesion; σ n is the applied normal pressure; δ is the interface friction angle.

LCS | GTX-1 | GS Interfaces
In the single-interface configuration, the development of peak shear stresses was observed to be more prevalent (see Fig. 11a, b) than in the double-interface configuration as can be seen in Fig. 11c.For the case of the LCS | GTX-1 interface, for example, peak shear stresses were mobilised at lower horizontal displacements of less than 12.5 mm, while the GTX-1 | GS interface mobilised peak shear stresses at approximately 20 mm horizontal displacement.Contrary to single interfaces, the double-interface configuration constituting similar lining components, mobilised peak shear stresses at higher horizontal displacements approximately 70 mm closer to the maximum horizontal displacement accommodated by the LSA of 75 mm.Commonly to all (1) τ = c a + σ a tanδ Table 3 Testing configurations a GTX-3S1 = Geotextile 3 with rough surface (reinforcing staple fibres) b GTX-3S2 = Geotextile 3 with smooth surface Single interfaces Double interfaces interface configurations, the mobilised peak shear stresses were higher than the strengths mobilised at LD.In the double-interface configuration, the extent of soil ploughing or soil blistering in the lower shear box was observed to increase with increasing applied normal stresses [4,13].In addition, this scenario progressed throughout the test duration.This scenario was noted to happen when soil specimen was placed in the lower shear box comparable to soil failure under a conventional footing [4].In this study, soil ploughing/blistering may be defined as the escaping of soil mass through a frontal gap (100 mm) of the LSA between upper (300 × 300 mm) and lower (300 × 400 mm) shear boxes during the shearing phase because of constant application of vertical pressure.This scenario, therefore, would explain why peak stresses for the double-interface configuration were only mobilised at higher horizontal displacements.
In summary, the mobilised peak and LD shear stresses registered by the double-interface configuration were higher than those observed in both single-interface configurations.However, the horizontal displacements at which the peak shear stresses were mobilised were higher and closer to the maximum horizontal displacement offered by the LSA.

GS | GTX-2 | GMB-1 Interfaces
Peak shear stresses were mobilised at incredibly low horizontal displacements (~ 2 mm) for the GTX-2 | GMB-1 interface.The frictional resistance offered by the two materials was low, also evidenced by low peak shear stress values.From Fig. 12b, it can be observed that this interface made a sharp increase in the shear stress for all applied normal pressures immediately after the shearing phase was initiated.However, this peak subsequently declined after a small horizontal displacement to a residual condition indicating the strain-softening phenomenon [8,18].
In addition, the GS | GTX-2 interface mobilised peak shear stresses at higher horizontal displacements, especially for the applied normal pressures of 300 and 450 kPa.This interface's friction behaviour was driven by the crushed gravelly sand (GS) material.Its particle size distribution comprising 25% fine gravel penetrated the protection geotextile in all applied normal pressures, causing minor interlocks that shattered the fibres in the shear direction.This action increased the mobilised shear stress and its subsequent horizontal displacement at which it would be achieved.The peak shear stresses mobilised by both single-interface configurations were higher than those mobilised at LD, however, not by significant magnitudes.
For the double-interface configuration, i.e.GS | GTX-2 | GMB-1 interface, peak shear stresses with definitive peaks, as seen in Fig. 12c, were mobilised at low horizontal displacements.In addition, notably in Fig. 12, the peak shear stresses of a double interface were seen to be closer to those of the GTX-2 | GMB-1 interface but significantly lower than those of the GS | GTX-2 interface.This observation suggested that failure within the double interface occurred along the GTX-2 | GMB-1 interface attributed to its low frictional properties offered by the two geosynthetic-geosynthetic lining components.Upon specimen assessment after the test, it was observed that the GS persistently ruptured the GTX-2 fibres in the direction of shear; however, this effect did not proceed through to GMB-1 surface.Furthermore, this scenario confirmed that the GTX-2's thickness was sufficient to protect the GMB-1 against punctures from fine gravels contained in the GS.

GTX-2 | GMB-1 | CD Interfaces
Peak shear stresses mobilised by the GMB-1 | CD interface were higher than those mobilised by its counterpart single interface (GTX-2 | GMB-1).For instance, at an applied normal pressure of 450 kPa, the former mobilised a peak shear stress of 158 kPa compared to 93.8 kPa mobilised by the latter interface.GTX-2 | GMB-1 interface mobilised peak shear stresses at incredibly low horizontal displacements (less than 2.5 mm).This led to development of standard peak curves, as seen in Fig. 13a.Conversely, the GMB-1 | CD interface mobilised its peak strengths at higher horizontal displacements closer to the maximum horizontal displacement accommodated by the LSA (i.e.75 mm), thereby making this interface to not have definitive peaks as can be seen in Fig. 13b.Notably, this interface achieved similar strengths at LD, unlike the GTX-2 | GMB-1 interface, whose LD strengths were lower than those obtained at peak.The peak strengths for the double-interface GTX-2 | GMB-1 | CD were mobilised at lower horizontal displacements, most especially for the applied normal pressures of 300 and 450 kPa.The horizontal displacements at which peak shear stresses were mobilised were comparable to those of the GTX-2 | GMB-1, indicating that failure within the double-interface configuration occurred on the GTX-2 | GMB-1 interface.The peak strengths observed were between the two single-interface configurations.All lining components evaluated contributed to the strengths obtained under this configuration with no significant predominance by a specific interface.The lower peak stresses obtained for both testing configurations indicated that the three lining components exhibited low frictional and interfacial shear properties.In addition, after the interface shear of all lining components was completed, specimen assessment suggested that no damage was done to any of the specimens except the compression of GTX-2 at higher applied normal pressures attributed to its thickness.

GMB-1 | CD | GMB-2 Interfaces
CD | GMB-2 interface mobilised higher peak shear stresses than GMB-1 | CD interface, although both interfaces mobilised these peak shear stresses at higher horizontal displacements.The surfaces of the two geomembranes were similar, i.e. smooth geomembranes; therefore, the only difference that resulted in the observed mobilised shear stresses lies with the CD material.The cusps of the CD interfaced with the GMB-2 while its smooth side interfaced with GMB-1.This was the primary difference, apart from GMB-1 being an HDPE material (stiff) while GMB-2 being an LLDPE material (flexible).For both single-interface configurations, mobilised peak shear strengths were comparable to the strengths obtained at LD as seen in Fig. 14a, b.
The double-interface GMB-1 | CD | GMB-2 mobilised peak shear strengths between the two single-interface configurations, indicating that all lining components contributed to the observed strengths with no single interface being predominant.Despite this, CD | GMB-2 interface proved to be the strongest interface and contributed more to the peak shear strengths.The horizontal displacements at which peak strengths were mobilised for the double-interface configuration were higher than those of the single-interface configurations, except for the GMB-1 | CD interface at 150 and 300 kPa.The behaviour of the lining components through double-interface shear testing was still comparable to that of single interfaces, thus, no definitive peaks were observed from the shear stress-horizontal displacement relationships, as seen in Fig. 14c.

CD | GMB-2 | SAND Interfaces
For single-interface configurations, GMB-2 | SAND interface mobilised higher shear strengths at lower horizontal displacements (less than 10 mm) than CD | GMB-2 interface, which mobilised its peak strengths at higher horizontal displacements.This resulted in the GMB-2 | SAND interface having definitive peaks, as shown in Fig. 15b, especially for the applied normal pressures of 300 and 450 kPa.The higher peak strengths obtained from this interface were attributed to the sandy soil, which developed a more assertive frictional behaviour by polishing the GMB-2 and creating minor grooves.This behaviour was also observed and compared by other scholars, who described that the shear behaviour of smooth geomembranes against soil depends on the particle sizes and gradation of soils [8,14,19].In their studies, large and angular soils resulted in higher strengths due to interlocking mechanisms with the geomembrane.In addition, the LD strengths of the GMB-2 | SAND interface were low compared to the LD strengths of the CD | GMB-2 interface due to a consistent decline of strengths, also known as strainsoftening on the GMB-2 | SAND interface after achieving a peak condition.
For the double-interface configuration of the three lining components, the peak shear strengths were mobilised at higher horizontal displacements than those of the GMB-2 | SAND interface but lower than those of the CD | GMB-2 interface.The shear stress-horizontal displacement relationships were comparable to those of the single weak interface (CD | GMB-2) as can be seen in Fig. 15c.Interestingly, however, the determined peak shear strengths were lower than both single interfaces but within a tolerable limit of 5% to those obtained on the CD | GMB-2 interface.This indicated that failure within the double interface occurred on the CD | GMB-2 interface.Moreover, this observation was consistent with a similar study on double interface of three lining materials: the geomembrane, geocomposite and sand [19].
In that study, it was shown that shear displacement occurred along the geomembrane-geocomposite interface, thus confirming that failure within a double interface occurred there.

SAND | GTX-3 | SAND Interfaces
The peak shear stresses for both single-interface configurations were mobilised at erratically higher horizontal displacements, i.e. there was no consistent trend that can be deduced from the results.For example, at an applied normal pressure of 150 kPa, the SAND | GTX-3S1 interface mobilised peak shear stress at a horizontal displacement of 43.91 mm.In contrast, its counterpart SAND | GTX-3S2 interface mobilised peak shear stress at a horizontal displacement of 74.91 mm, which is also defined as large displacement (LD).Notably, on both interfaces, the mobilised peak shear stress was either comparable to or slightly higher than the mobilised shear stress at LD.A unique trend was observed on the SAND | GTX-3S2 (smooth side of the reinforcing geotextile).As the normal applied pressure increased, the mobilised peak shear stress increased, and the horizontal displacements at which the mobilised peak shear stress was attained increased.However, this trend was observed to have a minimal impact on the strengths obtained at LD, except for the applied normal pressure of 450 kPa, where the specimen was worn-out after reaching the peak, consequently decreasing its strength, as shown in Fig. 16b.The double-interface configuration for these lining components mobilised peak shear stresses comparable to those at LD, as can be seen in Fig. 16c.In addition, these mobilised strengths at peak and LD were higher than those mobilised by the single interfaces.In addition, there was evidence of soil ploughing through the gap between the upper and lower shear boxes during the double-interface testing.

At Peak
A summary of soil-geosynthetic and geosynthetic-geosynthetic interfaces tested under a single-interface configuration obtained at peak for the range of applied normal stresses is presented in Fig. 17a.The top and bottom shear stress-normal stress relationships represent strongest and weakest/critical interfaces, respectively.SAND | GTX-3S1 interface was the strongest interface for the applied normal pressures exceeding 150 kPa.In contrast, its counterpart, i.e. the SAND | GTX-3S2 interface, was the strongest at lower applied normal stresses.The GTX-2 | GMB-1 interface was observed to be the critical interface for the entire range of applied normal stresses.

At Large Displacement
Generally, the SAND | GTX-3S1 interface was observed to be the strongest interface at LD for applied normal stresses above 200 kPa, as seen in Fig. 17c.However, below the applied normal pressure of 200 kPa, SAND | GTX-3S2 was the strongest.The GTX-2 | GMB-1 interface was observed to be the weakest single interface at LD, having generated the lowest interface shear resistances at all applied normal stress ranges.This was consistent with the critical interface observed at peak for single-interface configurations.

At Peak
The peak strengths for the double-interface configurations are as shown in Fig. 17b.LCS | GTX-1 | GS interface was observed to be the strongest for the applied normal stresses beyond 160 kPa, while for lower applied normal stresses, SAND | GTX-3 | SAND interface was observed to be the strongest.The latter observation was in line with what was observed in the single-interface configurations, where both SAND | GTX-3S1 and SAND | GTX-3S2 interfaces were observed to be the strongest interfaces.The former, i.e.LCS | GTX-1 | GS interface, however, was observed to produce high shear resistances due to minimal crushing of the LCS layer, which comprised of stones in the top shear box of the LSA.This action enabled the development of a solid intimate/embedment interaction of the LCS with GS layers through the protection geotextile (GTX-1).In addition, the presence of a GS layer instead of a heavy-duty rigid steel substrate (as is the case for single-interface configuration) in the lower shear box maximised the contact area between LCS and the protection geotextile, allowing sufficient transfer of applied normal stresses among the lining components.
The GS | GTX-2 | GMB-1 interface was observed to be the weakest interface under this testing configuration.Failure within this lining set occurred on the GTX-2 | GMB-1 interface which was also perceived to be the critical interface under a single-interface testing configuration.This observation agreed with other scholars who established that critical interfaces for the side-slope and basal lining systems were similar for the materials assessed under single-and multilayer testing configurations [13,14].In addition, the results for the critical interface were also in line with other studies which identified geomembrane-geotextile combinations as the ultimate critical interface [6].In addition, other similar studies that showed a smooth geomembrane producing the least shear resistances compared to a textured geomembrane among the lining components [18,19].

At Large Displacement
As seen in Fig. 17d, the strongest interface was initially SAND | GTX-3 | SAND interface up to applied normal stress of 200 kPa, after which the strongest interface became LCS | GTX-1 | GS.These two strongest double interfaces observed at LD were comparable to those observed under peak strengths.Two interfaces, i.e.GS | GTX-2 | GMB-1 interface and GTX-2 | GMB-1 | CD interface, were observed to be the weakest.These produced similar shear resistances at LD conditions, as shown in Table 4.As observed in the singleinterface configurations at peak, it was emphasised through the observed shear stress-normal stress relationship that failure within these two interfaces took place along the GTX-2 | GMB-1 interface.As a result, the combination post-peak/ large displacement (LD) strength envelope would coincide with the critical interface as sliding occurred on this (a) ( interface.In addition, the interface shear strength parameters for the two double interfaces at LD conditions were observed to be similar, as shown in Table 4.

Soil-Geosynthetics and Soil-Geosynthetic-Soil Interactions
Two interface test series in the double-interface configuration involved soil-geosynthetic-soil interactions.The recorded interface shear strength parameters, however, were observed to vary across all sets.For instance, the angles of interface friction obtained in the single interfaces LCS | GTX-1 and GTX-1 | GS were lower than those determined in the double-interface LCS | GTX-1 | GS as can be seen in Table 4.However, the observed apparent adhesion values for the above-mentioned single interfaces were higher than those observed in their respective double-interface configuration.
The peak interface angle of friction recorded by the single interfaces SAND | GTX-3S1 (37.7°) and SAND | GTX-3S2 (32.2°) were higher to those obtained in the double-interface configuration of SAND | GTX-3 | SAND (31.1°).At large displacement (LD), the single-interface SAND | GTX-3S2 recorded the lowest angle of interface friction compared to the single-interface SAND | GTX-3S1 and the respective double-interface configuration of SAND | GTX-3 | SAND as shown in Table 4.With exception to the observed apparent adhesion value of 125 kPa on the single-interface SAND | GTX-3S2, other values of apparent adhesion at peak and LD for the single interfaces in this test setup were observed to be lower than those obtained for the respective doubleinterface configuration.The inconsistency in the observed shear strength parameters in this test setup may be attributed to soil ploughing or soil blistering in the lower shear box due to its size difference with the upper shear box.

Soil-Geosynthetic, Geosynthetic-Geosynthetic and Soil-Geosynthetic-Geosynthetic Interactions
There were two test series that involved soil-geosynthetic-geosynthetic interactions in the double-interface configuration.The first set involved GS | GTX-2 | GMB-1 double-interface configuration.At peak, the interface shear strength parameters of 13.5° and 4.5 kPa were obtained while at LD values of 9.4° and 6.8 kPa were obtained.In its respective single-interface configuration, the above-mentioned values were closer to those of the GTX-2 | GMB-1 interface with 11.3° and 4.3 kPa at peak as well as 10.2° and 2.4 kPa at LD.The close relationship of the observed values suggests that the failure within a double-interface configuration occurred on the GTX-2 | GMB-1 interface and certainly not on the GS | GTX-2 as this interface recorded higher interface shear strength properties as can be seen in Table 4.
The other set involved single interfaces CD | GMB-2 and GMB-2 | SAND with the former recording an angle of interface friction at peak of 24.0° and the latter recording 29.1°.The value recorded by CD | GMB-2 was closely comparable to that observed in the double-interface configuration of CD | GMB-2 | SAND of 24.5° at peak.The apparent adhesion values at peak, however, for the single-interface CD | GMB-2 were higher than both GMB-2 | SAND and the double-interface CD | GMB-2 | SAND which recorded no apparent adhesion values.This trend at peak was similarly observed at LD as the angle of interface friction for CD | GMB-2 (23.8°) was comparable to that of CD | GMB-2 | SAND (24°), while GMB-2 | SAND recorded the lowest value of interface friction angle at 22.7°.The apparent adhesion value of the CD | GMB-2 of 9.5 kPa was highest as compared to its counterpart single-interface GMB-2 | SAND and the double-interface CD | GMB-2 | SAND which both recorded no value for apparent adhesion.This setup suggested that the interface shear strengths of the weakest interface will prevail in the multi-interface configuration provided test conditions are kept the same.

Geosynthetic-Geosynthetic and Geosynthetic-Geosynthetic-Geosynthetic Interactions
For the geosynthetic-geosynthetic-geosynthetic test series, the following were observed.The double-interface configuration of GTX-2 | GMB-1 | CD recorded a peak angle of interface friction of 13.4° which was in-between those recorded for single interfaces of GTX-2 | GMB-1 (11.3°) and GMB-1 | CD (18.5°).However, the interface angle of friction recorded at LD of 9.3° was lower compared to that of GTX-2 | GMB-1 (10.2°) and GMB-1 | CD (18.4°).In terms of apparent adhesion, it was observed that the doubleinterface configuration recorded higher values compared to both single-interface configurations as can be seen in Table 4.In addition, in the double-interface GMB-1 | CD | GMB-2, the peak angle of interface friction of 20.3° was higher than GMB-1 | CD (18.5°) but lower than CD | GMB-2 (24.0°) with the apparent adhesion value for the double interface being higher than both single interfaces as shown in Table 4.The trend observed in the peak angles of interface friction was similarly observed at LD.However, the LD apparent adhesion value for the double interface (5.1 kPa) was observed to be lower than that of the single-interface CD | GMB-2 (9.5 kPa) but higher than that observed on the GMB-1 | CD since no value for apparent adhesion was recorded for this interface.

Design Application
Design application of the obtained results involved stability assessment to determine the lower bound factors of safety (FoS min ) of the side slope and base of the proposed MSW landfill cell lined by site and project-specific geosynthetics and geomaterials.A two-part wedge analysis depicting a translational failure mode was analysed using limit equilibrium method (LEM).This approach assumes that within a waste mass, there exists an active and passive wedge resulting into a two-part wedge system.The active wedge is usually lined and tends to initiate failure on the side slope.A passive wedge, however, provides resistance to the generated instability.Further details about this approach can be found in these studies [20][21][22][23].
The design parameters included typical geometric values for solid waste landfills and inherent solid waste parameters (see Table 5) as well as the obtained interface shear strength parameters through single-and double-interface configurations (see Table 4).The peak strengths were used at the base and large displacement (LD) strengths on the side slopes of the proposed landfill cell.This is because the critical interface could be different at both locations, i.e. at the base and at the side slope of the lining system since it is influenced by the variability of waste depth and placement routines.As a result, a generalised approach could lead to unconservative FoS min especially, if a landfill is lined with a multi-layered soil-geosynthetics components.
Tables 6 and 7 provide a summary of the FoS min obtained from this study on single-and double-interface testing configurations.From this assessment, the following were established: • The lowest FoS min value of 1.1 was obtained at a singleinterface GTX-2 | GMB-1.Therefore, it can be deduced that the GTX-2 | GMB-1 interface was the weakest as elaborated previously, and if failure were to occur, this interface would be the first to fail at both locations of the landfill, i.e. at the base and at the side slope.In addition, the highest value of FoS min observed was 4.3, with the base interface being GS | GTX-2 and the side-slope interface being SAND | GTX-3S1.These two interfaces were noted to be the strongest under the single-interface testing configuration as explained earlier in this study.• For a double interface's FoS min evaluation, it was observed that an interface with GTX-2 and GMB-1 combination had the lowest FoS min of 1.2.The highest FoS min value of 5.5 was obtained at the LCS | GTX-1 | GS interface, indicating that this interface was the strongest in both landfill locations, i.e. at the base and at the side slope.Another double interface that recorded a higher value of FoS min than those recorded by single-interface configurations was SAND | GTX-3 | SAND interface, with 4.5 at the side slope and 4.4 at the base.As a whole, the FoS min below 1.5 (the required FoS for permanent slopes) highlighted in red as can be seen in Table 6 for single interfaces indicates that the critical interface at the base could be different to that on the side slope.This suggests that at failure shear strength transfer plays a role which results into different interfaces failing concurrently at different landfill locations.However, in this study, this behaviour did not recur under double-interface configuration as can be seen in Table 7, where the failed interfaces had a combination of only GTX-2 and GMB-1 lining components.

Conclusions
This study involved a series of interface shear tests implemented using two testing configurations, i.e. single, and double interfaces.Interpretation of the results included assessing the tested specimens to pin-point failure modes and establishing an interface at which failure could have occurred, especially for a non-standardised double-interface testing configuration.The following conclusions were established: • Higher shear strengths at peak and LD were observed under single-interface configurations.This observation was attributed to specimen clamping in the LSA under a single-interface configuration compared to doubleinterface configurations.• For interfaces that involved geomaterials, i.e. soil and stone in their setups, higher interface friction angles and apparent adhesion parameters were observed at peak but declined at LD.However, interfaces that involved only geosynthetics in their setups, similar interface friction angles and apparent adhesion parameters at peak and LD were observed.• Double-interface configuration revealed that failure occurred along the same interface as in the single-interface configuration, provided that the shear plane was

Fig. 17
Fig. 17 Critical interface assessment: at peak a single interfaces and b double interfaces and at LD c single interfaces and d double interfaces

Table 1
Properties of geosynthetics a NW nonwoven geotextile b W woven geotextile/geogrid c PET Polyester, HDPE high-density polyethylene, LLDPE linear low-density polyethylene

Table 4
Summary of interface shear strength parameters δ p and δ LD are the peak and large displacement (LD) angles of interface friction c α-p and c α-LD are the peak and large displacement (LD) apparent adhesions

Table 6
Lower bound factors of safety for single-interface configurations

Table 7
Lower bound factors of safety for double-interface configurations Page 15 of 16 59the weakest.For instance, tests conducted on a doubleinterface GS | GTX-2 | GMB-1 produced interface shear strength parameters closely comparable to those of the single-interface GTX-2 | GMB-1.These interfaces were determined to be weakest/critical in this study.•Inthe FoS min evaluations, it was established that provided the interface recorded lowest values of interface shear strength parameters independent of the testing configuration, low factors of safety were expected.For instance, interface test configurations with a combination of GTX-2 and GMB-1 recorded lowest FoS min (1.1 to 1.2) irrespective of whether they were evaluated under single-or double-interface configurations.