The Berlin 2 geomembrane workshop, held 26 September 2014 in Germany, gathered engineering experts from around the world and who represented multiple perspectives in geosynthetics. The goal was to develop a protocol for better anticipating the end of life–to define end of life–for an exposed, high-density polyethylene geomembrane that is actively in service.
Read more about the origin and context of the workshop.

Berlin 2 Geomembrane Workshop
Berlin 2 group photo. Front row (from left): Werner Müller, John Cowland, Mike Sadlier, Fred Gassner J.P. Giroud, Ian D. Peggs, Helmut Zanzinger, Andreas Wöhlecke, Daniela Robertson. Back row (from left): Daniel Tan, Hyun-Jin Koo; Chris Kelsey; Robert Kienzl; Mauricio Ossa; Catrin Tarnowski; Amr Ewais; Ana M. Noval; Marcus Grob; Montse Garcia Estopá, Sebastian Hausmann, Vera Olischläger, Franz-Georg Simon

The group’s technical notes serve as a final report is now available openly as a PDF download and the participants welcome field discussion.
We publish here an abridged portion of Sections 5.0 and 6.0 from the Final Report.
The workshop was organized by Ian D. Peggs (USA) and Helmut Zanzinger (Germany), with the assistance of Elizabeth Peggs (USA), Chris Kelsey (USA), Werner Müller (Germany) and Andreas Wöhlecke (Germany). Dr. JP Giroud served as facilitator and kept discussions on track. He also worked to identify, at the end of the day, three or four topics that could be researched before the next geomembrane workshop.
Invited participants were sent a skeleton proposed protocol and were requested to prepare and submit a 400 word “abstract” of a relevant topic prior to the workshop to stimulate discussion. The abstracts were circulated to all participants before the workshop. The proposal:
The day after the INTERNATIONAL GEOSYNTHETICS CONFERENCE in Berlin ends, I-CORP INTERNATIONAL (USA) and SKZ – Testing GmbH (GERMANY) are co-chairing a workshop to develop a protocol for the determination of remaining service life in in-service exposed HDPE Geomembranes. The objective is to determine when liners need to be replaced before a catastrophic failure occurs.
Pre-service lifetime estimates based on testing virgin materials at specific temperatures, in specific environments, cannot possibly reproduce all the synergisms that can occur in the field, but once a material has been in service for some time it should be more appropriate to assess remaining service life in that continuing environment.
The outline of a remaining service life assessment protocol follows and abstracts/papers will be invited from about 35 experts modifying that proposal or proposing another protocol. About 15 of these papers will be briefly presented to initiate extensive discussion. All papers and the discussion will be recorded and published electronically.
The skeleton of a proposed reference protocol is:

  • Definition of EoL
  • Measure OIT of exposed surface layer to assess stabilizer remaining
  • Measure carbonyl index to assess extent of surface oxidation
  • Measure stress cracking resistance of unnotched specimen or of melted/solidified plaque
  • Determination of critical carbonyl index at which SC is initiated
  • Assess impact of SCR on SC initiation after critical CI is reached

Consider surface properties vs. bulk properties. Consider use of thin film specimens. Consider novel testing techniques – impact resistance, autoclave oxidation, strain-hardening modulus.
Remember, we are seeking remaining service life of already installed material, not the expected life of virgin material.
BERLIN 2: One target. Multiple methodologies. Lively discussion.
One commonly heard comment was,
I have no test results, so I have nothing I can contribute to the discussion”.
Clearly, there are little data available, hence the reason for the Workshop. The objective of the discussion was to identify the data that are needed. Thus, forward-looking, out-of-the-box thinking, rather than hard data, was required. Many participants had difficulty with this.


There are several recurring questions that can be followed during the initial presentations and the subsequent discussion. They are:

  • Definition(s) of failure in individually measured parameters that individually or together can define EoL.
  • What combination of parameters is required to define EoL?
  • Monitoring changes in measured parameters and extrapolating them as a function of time to determine remaining service life, as opposed to taking a sample, measuring relevant parameters, then performing accelerated aging tests to determine time to EoL.
  • Perform tests relevant to the service environment of each facility, as opposed to doing standard tests applicable to all installations.
  • Focus on tests on past performance of the liner to extrapolate to EoL rather than focus on tests that show the future performance of the liner and time to EoL.
  • Test seams as well as liner material since most failures occur at seams.

Selected abstracts submitted by the invited experts follow:
6.1   Abstract of Mike Sadlier:
The obvious answer to the EOL question is when it “ceases to perform the required function” but that can mean different things in different situations. Much of this will depend on what is going on around a liner. A liner buried under tailings or waste and immobilized may lose properties but continue to reasonably perform its required function. However a pond liner or floating cover that loses properties will quickly cease to function.
The other important question is what HDPE are we talking about?  To me there are three very broad categories of HDPE with distinctly different properties and this brings into question whether our current definition of HDPE is adequate:

  • The “old time” HDPE with carbon black and little else had its difficulties with welding and stress cracking but it brought excellent chemical resistance and excellent UV resistance as can be seen on some older installations.
  • The “current standard” HDPE intended to meet GM13 is easier to work with and has better stress cracking properties along with what seems to be reasonable chemical resistance and good UV resistance.
  • The “new world” HDPE that shows much improved mechanical properties (e.g. multiaxial elongation at over 40%, NCTL approaching 1500 h). With good high-pressure OIT (HP-OIT) we can anticipate good UV performance but with mechanical properties similar to linear low-density polyethylene (LLDPE) what do we expect of chemical resistance?

We also need to know more about the real world of exposure of a HDPE or other GMB material which may well involve influences beyond those considered at design. Some examples include acidic ash from certain coal fired power stations, active radiation and other minerals in the general environment, especially at mining sites. There have been occasions when deterioration due to the intended exposure was minimal but other more general exposures have caused damage.
In assessing the performance of a GMB in service we are often limited by the capacity to extract samples from exposed locations and the capacity to generate a “benchmark” for the performance of unexposed material. Does material sampled from an anchor trench really represent unexposed material?  This is where archived materials and exposure coupons can enable a much better informed answer to the perennial question “how much time do we have left?”
6.2   Abstract of John Cowland:
The expected life of GMBs will depend on their intended applications, which can be quite different for different projects, such as mining projects, solid waste landfills, the containment of water and the containment of hazardous liquids. Expected life can also be affected by exposure to high and low temperatures, and UV light.
All too often, the design engineer is required to take responsibility for the future performance of a GMB manufactured with additives that the design engineer knows little about. Indeed, it is my experience from many factory inspections that some manufacturers do not know as much about the additives they have been sold to add to their GMBs as they should do to safeguard their own liability. I have experienced unfortunate cases where additives have been used to protect the GMB against exposure to acids where the project will expose the GMB to alkalis.
The skeleton of the proposed reference protocol is excellent, and in my view the workshop should also discuss how this protocol can help design engineers.
Finally, in my opinion, this workshop should be expanded to include all types of GMB, not just HDPE, or medium-density polyethylene (MDPE).

  • Abstract of Fred Gassner:

In response to the request for an opinion on life of HDPE GMB, I provide the following points from the point of view of a designer:

  • The operational or functional life of a GMB is related to when the effect of the GMB is materially different to what it was intended. It is my opinion this is when the seepage rate through the liner due to defects has changed by one order of magnitude compared to the design intent.
  • In my experience the degradation/failure of GMB liners occurs or starts at welds. A defect formed due to degradation is usually a long split along the edge of seam of the GMB and its seepage effect can be assessed similar to a wrinkle. The impact of failures along weld edges is most marked in ponds, where usually no or minor confinement exists over the GMB. The consequence of failures along seam edges in landfills may be less due to the confinement effect on the GMB.
  • Much analysis and research is aimed at intact GMB. Consideration should be given to the weld conditions and their durability as welding results in degradation of stabilizers/AO from the welding heating effect and hence is likely to be the first area of age failure of GMB liners.
  • Stress cracking can occur while the GMB still has a moderate level of OIT time. Probably worthwhile to get a better understanding of current stabilizer packages and their effect on SCR. Some GMBs plateau at certain OIT values, and is possibly linked to mobility of stabilizers molecules within the GMB sheet. So OIT values may no longer be a reliable way to assess the degradation status of the GMB.
  • The remaining service life of an exposed GMB can be assessed by a simple tensile test on coupons across the welds, compared to strength values achieved referenced from installation quality control records. These residual strength values can be compared to design strength requirements for wind uplift etc. Once values drop to unacceptable factors of safety this liner has reached its end of life, as a design wind effect is likely to result in large extent of liner seam separation at the panel seams. Failed seams along panel edges are likely to result in seepage rates similar to no liner being present, in most situations. This can be assessed on a case by case basis.
  • With advances in the GMB industry, is there a basis to develop a GMB with most of the chemical resistance of HDPE but with a resin that is not prone to stress cracking?  The majority of projects do not need the very high chemical resistance of HDPE.

6.4   Abstract of Catrin Tarnowski:
In the recent years the focus on the relevant long-term performance parameter of HDPE GMBs has changed: Whereas for several years the community talked on stress crack resistance nowadays the focus is getting back to oxidative stability. It seems that with having established the performance parameter for NCTL being > 300 h acc. to ASTM D5397, Appendix, that the problem of early failure is solved. Also considering that GMBs shall be installed in relaxed conditions, stress cracking should not be an issue.
But just for exposed applications we cannot exclude external stresses like wind action for example. When failures are reported the failure mechanism is a brittle one – stress cracking.
Thus not only the stage of oxidation, but also the stress crack performance related to the specific conditions of use has to be considered. The topic – when does a liner need to be replaced before a catastrophic failure does occur – is just for exposed applications a very challenging one. So what is the stress crack resistance value needed to overcome possible stress?
From the authors point of view a protocol addressing critical/relevant points in projects would be a starting point rather than knowing the GMB performance only. This might possibly be leaching, wind speeds, points and number of wind action, critical weld areas or bridging problems, initial performance of the liner, kind of raw materials used and various others.
With such a protocol one might find similarities in different projects related to the aging stage of a GMB and possible time to failure or necessary point of replacement. Thinking about failures which have occurred after short time period only, whereas in other cases HDPE GMBs are exposed for more than 30 years and still in service. Also a database on the properties of these GMBs itself might be a helpful tool.
6.5   Abstract of Ian Peggs:
Many studies have been done on the expected service life of a given as-manufactured HDPE GMB material in a given environment usually covered by soil. Lifetime is usually presented as time to reduce a given parameter, often tensile strength, to 50% of its original value. This is quite an arbitrary parameter. And most researchers will clarify that the results apply only to that specific HDPE formulation in that specific environment. Unfortunately no laboratory test environment can reproduce the field environment and synergisms of temperature (expansion/contraction), chemistry (environment/internal), and stresses (expansion/contraction, wind uplift), and UV exposure. And field failures rarely occur as a result of the material losing its tensile strength.
Most usually it is recognized EoL occurs as a result of thermal oxidation and UV photo-oxidation which causes the initiation of stress cracks in the exposed surface which ultimately propagate through the GMB by slow crack growth or rapid crack propagation. Since it is not possible to predict the propagation rate of a stress crack through a given thickness of GMB EoL is therefore considered to be when stress cracks first occur in an exposed surface. So, how can we assess the oxidation “degradation” rate in an exposed GMB, provided of course it will remain in the same environment?  Two or three samples taken after different service times will be compared to the uninstalled GMB. The following parameters will be measured for the following reasons, remembering that oxidation is a surface process:

  • HP-OIT to assess the rate of consumption of anti-oxidants in the stabilizer formulation. This will be done on a surface layer of the exposed liner, not on the complete thickness as is usually done.
  • When the HP-OIT effectively reaches a plateau in the surface layer, oxidation may not immediately occur. Therefore CI will be measured to determine the extent and rate of oxidation. The use of an attenuated total reflection (ATR) crystal in a Fourier transform infrared spectrometer (FTIR) will enable the CI of a surface layer to be measured.
  • Researchers in HDPE gas pipe have proposed there to be a critical CI at which SC is initiated. To determine this relationship in the exposed material the surface CI measured above will be compared to the unnotched SCR of a thin plaque prepared from the full thickness sample.
  • This will require data to be generated on the relationship between CI and unnotched SCR on variably oxidized specimens to determine the critical CI at which SC occurs on the GMB surface.
  • Thus the measured CI at different service times can be extrapolated to the time to critical CI and the initiation of SC, therefore the time to end of practical service life.

Knowing the time to EoL should allow the owner to plan for appropriate repairs or replacement of the liner and to avoid an unexpected catastrophic failure.
6.6   Abstract of Werner Müller:
The remaining service life of exposed HDPE GMBs depends (1) on the type and amount of AOs, which are still present, (2) on the loading conditions, i.e. to which extent tensile forces are regularly imposed onto the GMB due to thermal shrinkage at low temperatures, wind suction, deformations, indentations, etc., (3) the surface properties, i.e. occurrence of critical stress initiation points, deep scratches and grooves, local stress concentration at geometric discontinuities associated with seems, etc. Typically HDPEs are stabilized by a combination of phenolic and phosphitic AO. These AOs are primarily consumed by UV accelerated oxidation processes and hydrolytic degradation at the surface, by extraction and evaporation of the AO molecules and their degradation fragments from the surface. This will be discussed in detail. Two regimes of AO depletion have been found. Regime A: Accelerated loss of AOs due to exudation. Regime B: loss by normal diffusion. Since the AOs get lost at the surface, there is a concentration gradient between surface and bulk which drives the diffusion of AOs from the bulk to the surface. If AO concentration in the bulk becomes very low and concentration at the surface is therefore essentially zero, then a heterogeneous oxidation process starts at various spots on the surface, which are depleted from AOs (regime C). With tensile force present, cracks are initiated at these spots. Failure of the whole product will then be imminent. The depth profile of such AO concentration can be measured by OIT. From such measurements, one may deduce remaining service life. Nowadays, hindered amine stabilizers (HAS) are sometimes added to the “classical” stabilization. Since HAS are less extractable and not sensitive to hydrolysis, HAS may still be present even after extended loss of phenolic and phosphitic AO components. However, HAS do not prevent oxidation, even though, oxidative degradation is significantly decelerated. What happens to HAS stabilized HDPE GMBs?  To which extend is stress cracking initiation prevented?  How can a depth profile of HAS be determined, since neither Std-OIT nor HP-OIT are applicable?
6.7   Abstract of Amr Ewais:
There are a number of challenges to creating a protocol for assessing the remaining service life of exposed GMBs. First, it is anticipated that the GMB resistance required will differ from one application to another. Second, the rate degradation of the GMB would not only depend on the GMB properties (resin) but also on the exposure conditions. Third, there are difficulties associated with the techniques used to assess the remaining resistance of a GMB. For example, one of the techniques used to assess the GMB resistance is the stress crack resistance (SCR, ASTM D5397) of the GMB. To use the SCR to assess the end of life for the GMB, the following questions need to be answered: (a) is this a useful technique, and (b) if so, what is the critical SCR at which the liner will rupture and when will this critical SCR will be reached?  To address these questions, two studies were conducted to investigate the degradation of polyethylene GMBs under different loading conditions (demand). The first study, considered the degradation of different GMBs exposed to very different climatological conditions. It was shown that: (a) SCR may decrease significantly without any evidence of chemical degradation, and (b) the GMB installed on a 3H:1V, 22 m high, slope had not ruptured after 16 years despite a reduction in SCR to about 70 hours. In the second study, three sets of polyethylene GMB samples (14 different polyethylene GMB) were left exposed on a wooden rack inclined at 2H:1V in southern Ontario Canada. The first set of the GMB samples were exhumed from leachate lagoon after 14 years in service before being placed on the rack for an additional 17 years (a total of 31 years aging to date). The second set of GMB samples were taken from an initially un-aged GMB and have been exposed since 1997. The third set of GMB samples were taken from 12 different polyethylene GMBs of different thicknesses and resins and have been exposed since 2010. The temperature of the GMBs from different sets was monitored and it was found that: (a) there was no significant difference in the temperatures of the black GMBs regardless the difference in resin and thickness, (b) in Kingston, Canada, the temperature of the black GMB may exceed 70 °C in the summer where the temperature of the white GMB did not exceed 40 °C, and (c) minimum temperature of the GMB in winter may reach -20 °C. Although the GMB (from which the samples of the first set was taken) extensively ruptured in the lagoon during 14 years of service; intact samples taken from the lagoon and aged on the roof were observed to rupture after 12 years exposure on the roof (26 years of aging in total for these samples). This highlights the role of load-demand on the rupture of the GMB. Although the GMB samples in the second set still have a standard OIT (Std-OIT) of ~40 minutes (30% of the initial value) and a HP-OIT of 1200 minutes after being exposed to the elements for 16 years, several severe surface cracks have developed and the stress crack resistance has decreased from ~ 6000 hours to ~ 3000 hours (possibly due to surface aging). Comparing the depletion of the Std-OIT from the GMBs of third set in the second study, it was found that: (a) an increased HP-OIT correlated with a decreased depletion of Std-OIT, (b) thicker GMBs had slower Std-OIT depletion, (c) the depletion of AOs from high density and linear low density polyethylene GMBs depended more on the AO package than the polyethylene resin, and (d) the white GMB had the slowest Std-OIT depletion. The SCR measured for the black GMBs in the third set of samples decreased by 30 to 70% although they still had significant Std-OIT and HP-OIT; in contrast, the stress crack resistance of the white GMB has not decreased over three years of monitoring this GMB. The results of these studies highlight the challenges of assessing the EOL of a GMB and the need to better understand the different factors that may affect the time to GMB rupture.
6.8   Overview
Few participants addressed the immediate topic of developing an integrated testing protocol to assess remaining service life. Several addressed the need for durability in specific applications and several addressed different individual test methods that might be used as a part of an overall protocol but few addressed the integration of a number of tests. For instance, stress crack resistance test methods, and the data they generated were discussed but there was no discussion on how these data would or could be used to define a remaining service life. Nevertheless, it was interesting to see how the various experts interpreted the topic and the depth of their knowledge of HDPE GMBs.
More sections of the Berlin 2 Technical Notes (Final Report) will be published on Geosynthetica on November 13 and 15, 2018.