This page was printed from

Characterization of textured geomembranes predictive of interface properties

June 1st, 2009 / By: / Feature, Geomembranes

Currently under development is a new application of technology for textured geomembrane sheet characterization.


Currently under development is a new application of technology for textured geomembrane sheet characterization. We believe this method of characterization will demonstrate a strong correlation to interface shear testing with a more rapid testing response time and potentially much greater testing accuracy.

Textured geomembranes have been a large part of geomembrane design for about 20 years. A glance at the high-density polyethylene (HDPE) geomembrane data in the annual Geosynthetics Specifiers Guide shows that a large number of geomembranes commonly specified are textured. But a complete understanding of the interface friction between a textured geomembrane and either soil or another geosynthetic has not been easy.

The manufacturing processes for textured geomembranes, such as coextrusion with nitrogen gas, produce an unavoidable randomness in the peaks and valleys of the texturing. This randomness does not negatively affect the performance of these materials, but it has made them tough to characterize. This difficulty affects a design engineer’s understanding of how to best predict shear strength/interface properties. Thus, how to characterize textured geomembranes has been the subject of understandable debate among specifiers, testing experts, standards committees, and manufacturers. A new technological tool may help resolve these disputes by providing more accurate data faster.

With the application of new equipment and transferring technology from other fields, we are now able to produce and analyze thorough, three-dimensional data of textured geomembrane quickly. This data has been demonstrated to correlate well to low-load (veneer cover application) direct shear measurements and can also reduce the time for a direct shear performance report from days or weeks to minutes.

This is an important advance, one with the potential to truly influence the field because the development of materials (with their expanded functionality and better performance) and the strenuous requirements on engineering designs have tended to outpace the standards of practice of testing and materials characterization. This has certainly been true of textured geomembranes.

Previous methods have failed to give us as close a look at the textured sheet as the field would like. For example, consider the use of probes to discover a material sample’s asperity height, by which we determine the maximum variation in height between the peaks and valleys of the geomembrane’s surface texture. While the tips of probes can be standardized, the inherent variability in texturing methods may mean that the standard probe tips are not suited to measuring the particular sample strips that have been sent off for analysis.

Three-dimensional characterization (Figure 1) through the utilization of new technology may provide the best correlation yet for direct shear/interface friction.Figure 1 | Three dimensional topological characterization of a textured geomembrane


The prediction of the shear strength/interface properties of geosynthetic materials has been a concern since the beginning of geosynthetic use on slopes.

The true initiation of this issue, particularly as related to textured geomembranes, stems from one of the largest and earliest failures involving a geomembrane sheet that occurred at the Kettleman Hills (Calif.) landfill in 19881. This failure precipitated a dramatic increase in use of a new variety of geomembrane sheet—i.e., textured or surface-roughened geomembrane.

The initial version of a textured sheet was defined by what it was not and that was: not smooth. Time and requirements for more advances have led to much more definition and characterization. Each further advance in characterization (i.e., tilt table to shear boxes) has led to more daring engineering, projects with slopes designed to be “as steep as possible” to maximize landfill space and to improve the cost-effectiveness of both the site design and overall project performance.

The statement of the problem and related issues is relatively simple: Civil engineers want to be able to predict the interface strengths of geosynthetic-to-geosynthetic and of geosynthetic-to-soil interfaces2.

Manufacturers want to demonstrate the consistency and capabilities of their products in a timely fashion and in a method the engineering community3 can use. And the industry as a whole (owners, laboratories, institutes, regulators) wants these predictions and demonstrations consistent, reliable, and accurate.

As with most complex and multifaceted problems, the execution is extremely complicated. The civil engineer is plagued by a huge variation in predictions. Soil type, compaction levels, overburden pressure, moisture levels, seat time, variations in material performance, and many other factors combine to make nearly every project a new event requiring “ground up” calculations4. The manufacturers test the asperity height of a textured sheet5, but this clearly is not sufficient to satisfy customers.

Essentially, the industry cannot make concise and accurate predictions6. GAI-LAP data collected by Dr. George Koerner suggests that the direct shear testing protocol produces variation in excess of 15%. The performance of materials is not in doubt; but for a data-intensive design field, this disparity in characterization is bothersome.

So what can we do to improve our situation?

We have moved from the overly simple tilt table to a direct shear box7. The use of the shear box adds a wider range of test parameters and allows for more control of most variables.

However, the shear box is still plagued by several problems. One, the overall variability from laboratory to laboratory has not yet reached realistically acceptable levels. Two, the test is expensive — e.g., several thousand dollars per test is an average cost. Third, it is not timely. If you do not own a shear box8, it is a two- to three-week test from conception to final report. Even if you do own your own box8, most common test protocols take four or five days to complete.

(In this regard, the authors highly recommend the Designer’s Forum article by Richard Thiel, “A technical note regarding interpretation of cohesion (or adhesion) and friction angle in direct shear tests,” in the April/May issue of Geosynthetics).

All geomembrane manufacturers have made substantial efforts to address and bring progress to these issues. But in the end, it is not something they can control.

Thus, we have no good9 correlation between the asperity height, sheet variety, and the interface/shear properties. (That is to say, we have not found a good relationship that works across a wide range of materials and conditions.) In fact, we have seen some cases where the higher asperity heights actually produce interface shear results that are lower that those demonstrated by lower asperity materials. Clearly, more definition is needed than the simple peak/valley measurement generated by asperity height.

Not surprisingly, this avenue of thinking has been explored in the past. In the late 1990s, GSE supplied partial funding toward a research effort by Dr. Joseph Dove, then of Georgia Tech, now on staff at Virginia Tech10, to expand the characterization of textured geomembrane sheet into two dimensions. This was accomplished with the aid of a new apparatus that used a stylus on a floating arm to track the continuity of asperity heights as a sample was drawn across the apparatus.

This process was a significant advancement in the degree of characterization and the results have been published in several papers referenced herein. Unfortunately, there was no direct linkage of the two-dimensional characterizations to the interface shear results11.

The proposed technique of optical characterization

The goal is to improve both the characterization of the textured geomembrane and the correlation of this characterization to interface shear results. The characterization of the geomembrane has improved by moving from a one-dimensional mechanical methodology12 to a two- and three-dimensional optical characterization. This process was developed by improving on the two-dimensional stylus profilometry as mentioned above13.

The photograph (Figure 2) illustrates an overview of the device. Figure 2 | An optical scanning apparatus

Figure 3 gives a close-up view of the scanning beam in operation. Figure 3 | Scanning beam directly downward to a textured geomembrane surface

This new technology expands the test area, in normal operations providing the asperity height (and much more) characterization of an area up to 300mm2 (46.5in.2) per sample. Most important is the degree of completeness of this characterization. It consists of a noninvasive optical scan of the material’s surface. The data output is then used to generate a topological characterization of the entire tested section of materials.

In short, much more than asperity height can now be known about the sample. The shape, size, and distribution (density) of the peaks can be known. In an important step, the valleys are also topologically characterized. The “fill volume” can be easily calculated and thus one can understand and isolate the topological characteristics of the portion of the sample above the plane of the sheet.

Preliminary test results

Figures 4–9 demonstrate the capability of the software to generate data by altering the threshold of textured surface. Figure 4 is at the base thickness of the geomembrane and the light blue coloring represents the textured peaks.Figure 4 | 3-D topological characterization of a textured geomembrane (zero-plane threshold).

As the threshold is increased in height, one can see that the area the peaks cover decreases. Figure 5 | 3-D topological characterization of a textured geomembrane (10-mil threshold). Figure 6 | 3-D topological characterization of a textured geomembrane (15-mil threshold). Figure 7 | 3-D topological characterization of a textured geomembrane (20-mil threshold). Figure 8 | 3-D topological characterization of a textured geomembrane (25-mil threshold). Figure 9 | 3-D topological characterization of a textured geomembrane (30-mil threshold).

We have begun initial research regarding parameters such as: height, functional, spatial, hybrid, bearing ratio, and other volume-related parameters. Several hundred parameters are captured with each scan. A key question is “what are the parameters that correlate most directly with interface shear testing and how exact is the correlation?”

To that end, we have reviewed several parameters and conducted correlation calculations with a range of direct shear results. A portion of those correlations is illustrated in Figures 10–17. Clearly, one of the most-studied interfaces is that of a geotextile to a textured geomembrane in a landfill cover application. For that reason, in addition to the large volume of data available, we used that interface as some of the initial, primary correlation schemes.

Figures 10–13 illustrate some of the work that has been done. In all figures, the x-axis consists of different rolls of textured geomembrane produced on a variety of manufacturing lines during a range of time and raw materials. In all charts the primary y-axis (left side) are the results of a direct shear test with a overburden pressure run in accordance with ASTM 5321.

In Figure 10, the secondary y-axis (right side) charts the filling quantity (volume) of the valleys14 — the empty space below a threshold for a rangeof thresholds. Figure 10 | Chart of direct shear vs. filling quantity—cavities (0.15, 0.20 and 0.25mm threshold) by sample.

In Figure 11, the secondary y-axis (right side) charts the bearing area of the peaks15 — the area of the peaks above a 0.15mm threshold. Figure 11 | Chart of direct shear vs. bearing area (0.15mm threshold) by sample.

In Figure 12, the secondary y-axis (right side) charts the volume of the peaks16 — the volume of the peaks above a 0.20mm threshold. Figure 12 | Chart of direct shear vs. peak volume (0.20mm threshold) by sample.

In Figure 13, the secondary y-axis (right side) charts a portion of the bearing curve of the peaks17. Figure 13 | Chart of direct shear vs. bearing curve by sample.

The correlations for each of the comparisons are somewhat self-evident, however, a slightly better method18 is the plotting of the two measured values19 directly against each other and making a best fit line to demonstrate the consistency. This is done20 in Figures 14–17. Figure 14 | Chart of direct shear vs. filling quantity—cavities (0.15, 0.20 and 0.25mm threshold) direct comparison. Figure 15 | Chart of direct shear vs. bearing area (0.15mm threshold) direct comparison with trend line. Figure 16 | Chart of direct shear vs. peak volume (0.20mm threshold) direct comparison. Figure 17 | Chart of direct shear vs. bearing curve direct comparison with trend line.

Summary and recommendations

Clearly there is more work to be done here. It is our intention to extend this body of work to address an improved and more-defined relationship between three-dimensional surface topography and the interface shear performance of the characterized samples and, further, to broaden the range of interface shear conditions and characteristics and assure that the presumed correlations extend across as broad a range of interface conditions as possible. And still further, yet more immediate, to utilize this technology to reduce the variation in textured geomembrane sheet and to optimize the structure and performance of textured geomembrane for the requirements of civil engineering usages.


One of the real advantages of conference attendance and participation is the dialogue that occurs, including the valuable, Q-and-A that often follows presentations. After the original presentation of this material at the Geosynthetics-2009 conference in Salt Lake City, several excellent questions were asked by audience members. A handful of these are reconstructed here. Readers of Geosynthetics are invited to submit further questions to the magazine or to contact the authors directly.

What is the cost of the device?

Approximately $100,000, but the cost fluctuates depending on software selections.

What is the area of the scan?

Scans measure a 34-in.2 zone but the scan area can be altered to fit targeted investigations .

Can you pick up orientation effects?

This is not easy and, largely, not possible. Because these new scans produce three-dimensional sample data, they do not have a “directional lean.”

How long does it take to run and analyze a scan?

Scan times are approximately 20 minutes and automated analysis takes about five minutes. This is a significant improvement, given the weeks we have experienced for testing and analysis on textured geomembrane.

What does sPpk (mm) actually mean?

To be sure, this is a difficult concept to discuss concisely. It is a parameter that provides a good correlation with direct shear/interface friction response and comes out of what is known as the Abbot-Firestone Curve or the bearing-area curve (BAC). The BAC describes the surface texture of an object and plots the bearing-length ratio at different heights. From a mathematical perspective, it helps us understand the probability of cumulative distribution for real-value random variables. In the case of textured geomembranes for which absolute uniformity of texture is not possible, it provides characterization insight. Three-dimensional analysis helps get us closer to this parameter and, thus, the good correlation we need for direct shear/interface friction.

Boyd Ramsey is vice president, Technical Sales/Products Division, and Jimmy Youngblood is technical support manager, both with GSE Lining Technology Inc. in Houston, Texas.

This article is based on a paper originally presented at the GRI-22 conference Feb. 27, 2009, in Salt Lake City.


(1) R. John Byrne, J. Kendall, and S. Brown, “Cause and mechanism of failure Kettleman Hills Landfill B-19, Phase IA,”

(2) The construction of this sentence and its similarity to the title of GRI report #30 reference is not accidental.

(3) Read customers.

(4) The play on words is obvious.

(5) You’all must understand that there is no asperity height knob on the control panel of any geomembrane manufacturer’s production lines. Asperity height is an outcome of several factors, all of which affect other performance characteristics of the geomembrane sheet.

(6) Currently, the direct shear testing protocol produces variation in excess of 15%; Dr. George Koerner – GAI LAP data.

(7) Not enough can be said about the efforts of Rob Swan in this regard. I have listed a link to a list of papers and documents he has published as well as his master thesis in the references of this paper. He really does deserve credit as one of the “fathers” of this particular niche of civil engineering.

(8) Full disclosure – GSE owns its own shear box; it was originally built by Rob Swan.

(9) “Good” defined as a relationship that works across a wide range of materials and conditions.

(10) Again, some personal credit is due here as well. It is difficult for me to see how this work reaches the current state of progress without the efforts of Dr. Dove.

(11) Despite a sincere effort to do so.

(12) Asperity height.

(13) And in fact, the technology is so new that the ISO standards are still being developed for this new topological characterization technology.

(14) This is important! It is the valley’s volume (empty space) that is being calculated, not the peaks.

(15) This is again important! It is the surface area of the peaks (above threshold) that is being calculated.

(16) This is yet again important! It is the volume of the peaks (above threshold) that is being calculated.

(17) A bearing curve is a somewhat abstract mathematical representation of the roughness of the material, which discounts the extreme upper and lower portions of the peaks.

(18) Regards to Gary Kolbasuk.

(19) Direct shear values and the corresponding selected texture characterization parameter.

(20) With wide variation in the results (get it! variation in charting the variation!): Valley volume and peak volume appear to be irrelevant to direct shear; Bearing area has some correlation and the Bearing Curve results demonstrate the most consistent trend.


ASTM D5321, Standard test method for determining the coefficient of soil and geosynthetic or geosynthetic and geosynthetic friction by the direct shear method. American Society for Testing and Materials, West Conshohocken, Pa., USA, 2008.

ASTM D7466, Standard test method for measuring the asperity height of textured geomembranes. American Society for Testing and Materials, West Conshohocken, Pa., USA, 2008.

Dove, J.E. and Frost, J.D., “A method for measuring geomembrane surface roughness,” Geosynthetics International, Vol. 3, No. 3, pp. 369-392, 1996.

Dove, J.E., Frost, J.D. and Dove, P.M., “Geomembrane microtopography by atomic force microscopy”, Geosynthetics International, Vol. 3, No. 2, pp. 227-245, 1996.

Dove, J.E., Frost, J.D., Bachus, R.C., Han, J., “The influence of geomembrane surface roughness on interface strength.” Proceedings of Geosynthetics ’97 Conference, IFAI, Long Beach, Calif., USA, Vol. 2, pp. 863-876, 1997.

Dove, J.E., and Harpring, J.C. “Geometric and spatial parameters for analysis of geomembrane/soil interface behavior.” Proceedings, Geosynthetics ‘99, IFAI, Boston, Mass., I, pp. 575-588, 1999.

Giroud, J.P., “Quantitative analysis of the impact of adhesion between geomembranes and geotextiles on the stability of soil-geosynthetic systems on slopes.” J.P. Giroud Inc., Ocean Ridge, Fla., USA, pp. 14, 2004.

GRI Report No. 30, “Direct shear database of geosynthetic-to-geosynthetic and geosynthetic-to-soil interfaces.” Geosynthetic Institute, Folsom, Pa., USA, 2005.

GRI Test Method GM12, (Mod. 1: Oct., 1998) Standard Test Method for “Asperity measurement of textured geomembranes using a depth gage.” Geosynthetic Institute, Folsom, Pa., USA, 1998.

Hsuan, Y. G. and Koerner, R. M., “Rationale and background for the GRI-GM13 specification for HDPE geomembranes.” Proceedings of Geosynthetics ’99, IFAI, St. Paul, Minn., pp. 385-400, 1999.

Koerner, G.R., “Geosynthetic Institute’s Efforts In Accreditation and Certification.” Koerner Symposium 2004, Geosynthetic Institute, Folsom, Pa., USA, 2004.

Koerner, R.M., “Designing with Geosynthetics” 5th Edition, 2006, by Prentice Hall Inc. Upper Saddle River, N.J., USA, 2006.

Koerner, R. M. and Soong, T.-Y., “Assessment of ten landfill failures using 2-D and 3-D stability analysis procedures,” Terzaghi Lecture at Technical University, Vienna, 40 pp., Feb. 1999.

Martin, J. P. and Koerner, R. M., “Geotechnical design considerations for geomembrane lined slopes: Part I – Slope stability,” International Journal Geotextiles and Geomembranes, Vol. 2, No. 2, , pp. 299-321, 1985.

Ojeshina, Anthony O., “A new high friction HDPE geomembrane,” Geotextiles and Geomembranes, Volume 10, Issues 5-6, pp. 433-441.

Richardson, G.N., Thiel, R.S., “Interface shear strength: Part I – Geomembrane considerations.” Geotechnical Fabrics Report, Vol. 19, No. 5, pp. 14–19, 2001.

Swan, R.H. Jr., “Frictional behavior of high performance geosynthetics.” MSCE Thesis, Drexel University, Philadelphia, Pa., 175 pp., 1987 Oh, what the hell, we should probably add everything that Rob Swan has ever published; a somewhat aged link/listing is here:

Vaid, Y.P., Rinne, N., “Geomembrane coefficients of interface friction.” Geosynthetics International 2 (1), 309–325, 1995.

Leave a Reply

Your email address will not be published. Required fields are marked *

Comments are moderated and will show up after being approved.