Tim Trial, Carlisle SynTec Inc., Carlisle, PA Ross Robertson, Carmel, IN and Brian Gish, Carlisle SynTec Inc., Carlisle, PA ABSTRACT Single-ply rubber membranes, based upon ethylene-propylene-diene terpolymers (EPDM), have been used as roofing membranes for over 30 years. A summary of the physical properties of aged field samples was reported in 1991 . Under the auspices of the EPDM Roofing Association, this subject will be revisited and extended to include the performance of aged EPDM membranes in place for the past 20 to 25 years. Roof systems by various manufacturers will be sampled in a variety of locations. An inde¬ pendent laboratory will evaluate the samples to determined elongation, tensile strength, and tear resistance of the aged membrane. The impact of the results on the actual field performance and long-term stability of EPDM roofing membranes in var¬ ious climates and system configurations will be discussed. Tim Trial Tim Trial joined Carlisle SynTec Inc. in 2003 as a polymer scientist. He is involved in the develop¬ ment of formulations for rubber membrane and flashing products for roofing and waterproofing. Trial has authored five papers and received three patents. He earned a B.S. from Texas State University in 1991 and a Ph.D. in chemistry from The University of Southern Mississippi in 1996. Trial, Robertson, and Gish — 101 EPDM ROOF MEMBRANES: LONG TERM PERFORMANCE REVISITED INTRODUCTION Ethylene-propylene-diene-terpolymer (EPDM) mem¬ branes were introduced into the single-ply roofing market in the mid 1960s. EPDM continues to be the number one roofing membrane of architects, roof consultants, and contractors for both new construction and replacement roofing projects. 1 Due to the chemical structure of EPDM, the material is inherently ozone resistant as com¬ pared to other rubber materials.2 Additionally, the materi¬ al is intrinsically resistant to most acids and bases (alka¬ li). EPDM is mixed into complex rubber formulations to improve the physical and mechanical properties of the base material. Carbon black is added to provide rein¬ forcement, which serves to enhance the physical and mechanical properties of the formulation, and to improve UV resistance. Processing oils improve processability and the flexibility of the membrane at low temperatures. Curatives are mixed in the formulation to facilitate vul¬ canization or cross-linking of the rubber. Chemical cross¬ linking serves to improve the heat and solvent resistance, increase the hardness, tensile strength, modulus, and tear strength of the compound. Since EPDM polymers are hydrocarbons, proprietary ingredients can be incorporat¬ ed into the rubber formulation to impart fire resistance. The long-term performance of a roofing material is dependent upon its resistance to the combined effects of water, UV radiation, ozone, heat, and thermal cycling. Additionally, the design of the roof system and site loca¬ tion can serve to exploit or diminish the impact of the environmental factors. In 1985, Strong and Puse reviewed several EPDM aging and weathering studies? EPDM samples (80 mils thick) were aged for 10 years in a tropical environment. The samples were relatively unaffected by the humid conditions. After ten years, a 25% decrease in tensile strength and ultimate elongation were observed. Samples (40 mils thick) exposed 10-15 years in subtropical Florida conditions displayed a 25 % reduction in tensile strength after 15 years, and a 50 % decrease in ultimate elongation after 10 years. The tensile strength and per¬ cent elongation of exposed and ballasted rooftop samples aged seven years were compared at various test tempera¬ tures (-20 to 80°C). The results indicate a relatively small decrease in tensile strength and a greater loss in ultimate elongation over the test temperature range for the exposed sample. Gish and Lusardi studied the aged properties of 45 in situ roof samples cut from various roof systems ranging in age from five to 17 years.4 In general, an increase in tensile strength, tear resistance, brittleness temperature, Shore A hardness, and a reduction in ultimate elongation as compared to unaged samples were observed. No difference was observed in the glass transition temperature or appearance of the aged membrane as compared to the unaged samples. The aged properties of 8 to 9-year-old ballasted membranes were comparable to those of exposed (adhered) membranes, with the excep¬ tion of ultimate elongation. Exposed membranes suffered the greatest decrease in ultimate elongation (25 to 40% reduction for 5 to 12-year samples, 54% for the 17-year samples). A less pronounced reduction in the ultimate elongation (12 to 40% for 5 tolO-year samples) was observed for the ballasted membranes. Eighty-seven per¬ cent of the samples were observed to exceed the ASTM and MRCA ME-20 specifications for new membranes? 6 All samples were observed to exceed the ASTM and MRCA ME-20 requirements for heat-aged membranes. The purpose of this study is to revisit and extend the Gish/Lusardi study to include performance of aged EPDM membranes, which have been in service for the past 16 to 26 years. Field samples of ballasted and exposed membranes have been collected by two major manufacturers of EPDM roofing membrane and submit¬ ted to an independent laboratory for testing. In this paper, we report on the findings.
Samples of EPDM membrane (approx. 6″ x 4.5′) were cut from existing (in service) roofs. A one-square¬ inch sample was cut from the sample membrane for opti¬ cal studies. The substrate and adhesive from exposed samples were carefully removed. A total of 33 samples representing nine states were obtained from commercial roofing systems (not experimental). Roof systems include: fully adhered (22), ballasted (10), and non-pene¬ trating, mechanically fastened (1). The samples were submitted to Architectural Testing Inc., York, Pennsylvania, an independent, third-party lab¬ oratory. The samples were tested according to ASTM Trial, Robertson, and Gish — 103 D412 and D624 for tensile strength, ultimate elongation, and tear resistance.7-8 Median values for all test results are reported. Optical studies were per¬ formed on an American Optical Binocular Stereomicroscope, Model Forty at 10 x magnifica¬ tion. Optical micrographs were recorded at 190 x magnification utilizing a Zeiss Stemi 2000 stereomicroscope equipped with a COHU High Performance CCD Camera, Model Number 4915- 51001/0000. Five discrete mea¬ surements were obtained, and the median value is reported. Photographs were recorded with an Olympus Camedia C-700 Figure 1 – Tensile strengths of ballasted samples. Digital Camera. Physical Property Test Results Physical property test results are separated into two categories: results for exposed membranes and ballasted membranes. Exposed membranes were obtained primarily from fully adhered roof systems with one non-penetrating mechanically fas¬ tened system. The physical prop¬ erty results are presented visually in Figures 1 to 6. Each bar repre¬ sents a specific roof. Actual val¬ ues are presented in Tables 1 and 2. Tensile Strength The tensile strength for bal¬ lasted membranes ranged from 10.8 to 14.9 MPa (1560 to 2160 psi) (Figure 1) and from 9.4 to 13.5 MPa (1350 to 1950 psi) for exposed membranes (see Figure 2). The ASTM D 4637 minimum requirement for new sheets is 9.0 MPa (1305 psi) and 8.3 MPa (1205 psi) for heat aged samples. MRCA ME-20 requires new membranes to meet a minimum of 6.0 MPa (850 psi), and 5.5 MPa (800 psi) for aged membranes. All samples were observed to meet the minimum ASTM and MRCA specifications for new membrane, and well exceed the requirements for heat-aged samples. Trial, Robertson, and Gish — 104 Ultimate Elongation The ultimate elongation val¬ ues observed for ballasted mem¬ branes ranged from 290 to 370% (Figure 3) and from 150 to 320% (Figure 4) for exposed mem¬ branes. The ASTM D 4637 mini¬ mum requirement for new sheet is 300% and 200% for heat-aged samples. MRCA ME-20 requires new membranes to meet a mini¬ mum of 250%, and 200% for aged membranes. All ballasted samples were found to meet the minimum ASTM and MRCA for new and aged membranes. Most exposed membranes examined did not meet the minimum ASTM (21/23) and MRCA ME-20 (17/23) requirements for new membranes. Twelve samples were observed to exceed the minimum require¬ ments for heat-aged sam- 70.0- 18 20 21 21 21 22 23 23 23 23 Age (years) ——– Typical Membrane Property, 4th Quarter, 1983 ——— —— – ASTM D 4637 Requirement (New Membrane) — . — – — ASTM D 4637 Requirement (Heat Aged Membrane) Figure 3 – Tear resistances of ballasted samples. pies. Tear Resistance Tear resistance values for ballasted membranes ranged from 45 .8 to 65 .0 kN/m (261.7 to 371.2 Ibf/in) and from 38.1 to 50.5 kN/m (217.7 to 288.2 Ibf/in) for exposed mem¬ branes. The ASTM D 4637 minimum requirement for new sheet is 26.3 kN/m (150 Ibf/in) and 21 .9 kN/m (125 Ibf/in) for heat-aged samples. MRCA ME-20 requires a minimum of 21.0 kN/m (120 Ibf/in) for Sample Location (State) Age (years) Gauge (mils) Tensile Strength (MPa) Elongation (%) Tear Resistance (kN/m) ASTMD 4637 New (Heat Aged) 40 (40) 9.0 (8.3) 300 (200) 26.3 (21.9) 1 GA 18 45 13.7 340 49.2 2 VA 20 45 13.7 330 50.9 3 MI 21 45 12.1 300 45.8 4 IL 21 45 14.0 360 56.1 5 OH 21 45 14.1 360 52.4 6 OH 22 45 13.3 370 62.0 7 MI 23 45 12.9 340 54.9 8 OH 23 45 13.2 360 55.4 9 IA 23 45 14.2 330 65.0 10 OH 23 45 14.9 290 60.4 new membranes. There is Table 1 — Physical properties of ballasted membranes. no MRCA ME-20 require¬ ment for aged membrane. All samples were observed to meet the minimum ASTM and MRCA ME-20 requirements for new membranes and all ASTM requirements for heat-aged samples Optical Studies and Measurements Weathering resistance was assessed by visual inspec tion according to ASTM D 4637. No crazing (network of fine surface cracks) was observed for ballasted membranes, regardless of age. About one-half (12 out of 23) of the exposed membranes Trial, Robertson, and Gish — 105 Table 2 — Physical properties of exposed membranes. Sample Location State) Age (years) Gauge (mils) Tensile Strength (MPa) Elongation (%) Tear Resistance (kN/m) ASTMD 4637 New (Heat Aged) 40 (40) 9.0 (8.3) 300 (200) 26.3 (21.9) 11 IA 16 60 13.5 270 47.6 12 WI 17 45 9.7 180 48.8 13 FL 17 60 10.1 150 41.6 14 WI 17 45 10.5 230 46.3 15 FL 17 60 11.2 180 38.1 16 MI 17 60 11.4 220 45.0 17 MI 17 60 11.5 320 50.5 18 MI 17 60 11.7 290 45.7 19 MI 17 60 12.9 300 49.0 20 WI 18 60 10.4 190 44.9 21 MI 18 60 11.0 240 44.2 22 OH 18 60 12.4 240 47.4 23 GA 19 45 9.5 150 44.7 24 FL 19 60 11.1 180 42.4 25 WI 19 60 11.4 270 45.4 26 MI 20 60 9.4 160 41.3 27 MI 20 60 10.2 180 40.8 28 MI 20 60 10.9 170 43.2 29 WI 20 60 11.3 190 42.2 30 OH 21 60 10.9 230 48.5 31 OH 22 60 9.7 180 48.5 32 PA 24 60 10.7 290 45.0 33 PA 26 60 12.4 230 46.2 Figure 7 – Surface photograph of new membrane, exposed membrane (26 years old), and ballasted membrane (18 years old). exhibited some degree of crazing, when examined under magnification. Crazing was not observed without the aid of magnifi¬ cation. (See Figure 7) In order to measure the width of any observed crazing, optical micro¬ graphs of the membrane samples were obtained. The results can be found in Table 3. The width of the measured artifacts ranged from 0.00059 to 0.0017 inches. DISCUSSION OF RESULTS As can be seen by examination of Figure 1, the tensile strength for the ballasted membranes remained relativity con¬ stant without regard to the age of the membrane. The ultimate elongation and tear resistance values (Figures 2 and 3 respec¬ tively) displayed the same trends, i.e., no significant deterioration of the physi¬ cal properties has been observed in up to 23 years of service life. These results are comparable to those obtained by Gish and Lusardi in 1991 ,4 The same general trend is observed for tensile strengths (Figure 4) and tear resistance values (Figure 6) for the exposed membranes. However, a significant decrease in the ultimate elongation (Figure 5) was observed. Although no significant decrease in the physical properties of the exposed membrane as compared to the ballasted Trial, Robertson, and Gish — 106 membrane was observed in the Gish/Lusardi study, the results observed in the current study are consistent with the accelerated weathering model presented in the Gish/Lusardi paper. As per the model, a decrease in the ultimate elongation is observed after 3,000 hours of accelerated weathering, while the tensile strength displays a slight decrease.4 It is attractive to attempt to compare the aged results to those obtained for newly manufactured membranes; however, such comparisons are difficult for a variety of reasons. The formulations used today are not necessarily comparable to those produced 17 to 25 years ago. The industry has seen the develop¬ ment of new polymerization catalysts used to produce EPDM. These catalysts allow for the design of poly¬ mers specifically engineered for the single-ply roofing industry. Additionally, advances in the technology of other ingredients have resulted in the improvement of membrane properties. These advances in polymer and ingredient technology have resulted in subtle changes in the physical properties and processability of the resulting rubber formulations. For these reasons, the best comparison are data gleaned from the Gish/Lusardi study, which reports typical original properties for the fourth quarter of 1983.4 The typical membrane properties are displayed in Table 4. The tensile strength and tear resistance of ballasted and exposed membranes in this study are consistent with or slightly higher than the 1983 data. The increase in the tensile strength and tear resistance are likely due to post vulcanization cross-linking at rooftop tempera¬ tures. The decrease in ultimate elongation is consistent with the accelerated weathering model dis¬ cussed above. Additionally, the decrease in ultimate elongation is in agreement with the data reviewed by Strong and Puse as discussed in the introduction to this paper.3 CONCLUSIONS The results obtained in this study confirm the outstanding field aging performance of EPDM membranes. The tensile strength and tear resistance data obtained Sample Location (State) Age (years) Crazing Size (inches) 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 IA WI FL WI FL MI MI MI MI WI MI OH GA FL WI MI MI MI WI OH OH PA PA 16 17 17 17 17 17 17 17 17 18 18 18 19 19 19 20 20 20 20 21 22 24 26 No No No No No Yes No No No Yes Yes Yes No No No Yes Yes Yes No Yes Yes Yes Yes 0.00092 0.00092 0.00092 0.00084 0.00076 0.0009 0.00076 0.00059 0.0011 0.00092 0.0017 Table 3 — Weathering resistance of exposed membranes. ——– Typical Membrane Property, 4th Quarter, 1983 — — ASTM D 4637 Requirement (New Membrane) — ■ ■ — ASTM D 4637 Requirement (Heat Aged Membrane) Figure 4 – Tensile strength of exposed samples. Trial, Robertson, and Gish — 107 for both ballasted and exposed roofs exceed the ASTM D 4637 specifications’ new and heat aged membranes after 17 to 26 years of service life.5-6 Although a number of samples cut from exposed roofs displayed ultimate elongations below the ASTM D 4637 specifications for heat aged membrane, it is important to remember these EPDM mem¬ branes can still stretch to almost twice their dimensions. Additionally, in all roof systems sampled, the membranes were found to be watertight and functional. The tensile strength and tear resistance of the exposed systems are consis¬ tent with respect to age. The decrease in the ultimate elongation and surface crazing are expected observations after long-term exposure to weathering. For black EPDM, we believe that the reduction in elongation is primarily due to heat aging and the observed crazing results from exposure to the UV portion of solar radiation. The physical properties of the ballasted membranes remain constant with respect to age. The data obtained in this study indicates the membranes will contin¬ ue to function and remain watertight for years to come. ACKNOWLEDGMENTS The authors gratefully acknowledge the EPDM Roofing Association for the funding of this paper. Also acknowledged are the contributions from our colleagues at Carlisle SynTec Incorporated and Firestone Building Products who have assisted with the work that has resulted in this paper. We especially wish to thank Becky Over and Glenn Fitting of Carlisle SynTec Incorporated for their contributions. 600 500 -400 c 16 17 17 17 17 17 17 17 17 18 18 18 19 19 19 20 20 20 20 21 22 24 26 Age (years) ——– Typical Membrane Property, 4th Quarter, 1983 ASTM D 4637 Requirement (New Membrane) — ■ —— ■ — ASTM D 4637 Requirement (Heat Aged Membrane) Figure 5 – Ultimate elongations of exposed samples. REFERENCES 1. Building Design and Construction and the National Roofing Contractors Association (NRCA) 2002-2003 mar¬ ket surveys, 2003. 2. Basic Elastomer Technology, K. C. Baranwal and H. L. Stephens eds., p. 42, The Rubber Division, American Chemical Society, The University of Akron, Akron, OH, 2001. 3. Strong, A. G., Puse, J. W., Proc. 2nd Inti. Sympo- Age (years) Typical Membrane Property, 4th Quarter, 1983 ASTM D 4637 Requirement (New Membrane) ASTM D 4637 Requirement (Heat Aged Membrane) sium on Roofing Figure 6 – Tear resistances of exposed samples. Trial, Robertson, and Gish — 108 Technology, Chicago, IL, 1985, pp 376 – 381. 4. Gish, B. D., Lusardi, K. P., Proc. 3rd Inti. Symposium on Roofing Technology, Montreal, CA, 1991, pp 159-166. 5. Annual Book of ASTM Standards, Volume 04.04, American Society for Testing and Materials, Philadelphia, PA., 2003. Table 4 — Typical original membrane properties, 4th quarter, 1983.4 Property Membrane Mean Typical Range Tensile Strength (MPa) 11.7 11.4-12.2 Ultimate Elongation (%) 500 450 – 550 Tear Resistance (kN/m) 40.3 36.8 – 42.0 6. Midwest Roofing Contractors Association, Recommended Performance Criteria for Elastomeric Single-Ply Roof Membrane Systems, Kansas City, MO., 1982. 7. Annual Book of ASTM Standards , Volume 09.01 , American Society for Testing and Materials, Philadelphia, PA., 2003. 8. Annual Book of ASTM Standards, Volume 09.02, American Society for Testing and Materials, Philadelphia, PA., 2003. Trial, Robertson, and Gish — 109
