United States Environmental Protection Agency Hazardous Waste Engineering Research Laboratory Cincinnati, OH 45268 Research and Development EPA/600/S2-88/040 Aug. 1988 &EPA Project Summary Loading Point Puncturability Analysis of Geosynthetic Liner Materials Daren L. Laine, Michael P. Miklas, Jr., and Charles H. Parr Geomembrane liner performance was examined in laboratory tests subjecting polyvinyl chloride (PVC), chlorosulfonated polyethylene (CSPE), and high density poly- ethylene (HOPE) materials, in two thicknesses each, to varying pres- sures, temperatures, and point loads. Loads were induced by placing the geomembrane material over trun- cated rigid epoxy cones used as loading points for 9.5, 19.0, and 24.5 mm heights above the sand subgrade while arranged in three- cone clusters and applying a hy- drostatic load to the top side of the liner. Constant hydrostatic loads of 17.93 kPa at 23°C and 50°C were applied during a one-year test. HOPE material measuring 1.5 mm thick failed for the loading point height of 25.4 mm above the subgrade. After 365 days, the load- ing pressure was increased to 60.03 kPa for an additional 30 days. Failures were induced in 1.5-mm HOPE for loading point heights of 19.0 and 25.4 mm and in 2.5-mm HOPE for loading point heights of 25.4 mm. HOPE with a thickness of 1.5-mm failed for a loading point height of 19.0 mm with a 1.5-mm geotextile placed between the HOPE and the loading point at 17.94-kPa pressure and ambient temperature. HOPE with a thickness of 2.5 mm overlaying a 3.8-mm geotextile failed under 60.03-kPa pressure for a 25.4-mm loading height at the high temperature test condition. No materials failed when overlaid upon a 5.3-mm geotextile. Transient pressure loading test without geotextile support exhibited failures caused by the maximum pressure load attained. The test results indicate that moderate economic benefit may be gained by allowing particles project- ing up to 25.4 mm above the subgrade to remain in a finished surface. Finished installation cost reductions of up to 28 percent per hectare could result if the largest particles in the subgrade were comparable with the tested sizes. Geomembrane material performance was improved with the addition of geotextiles, indicating a positive cost-benefit advantage when a ge- otextile underlay is used. This Project Summary was developed by EPA's Hazardous Waste Engineering Research Laboratory, Cincinnati, OH, to announce key findings of the research project that is fully documented in a separate report of the same title (see Project Report ordering information at back). Introduction Geosynthetic materials are used to provide reinforcement, drainage, filtration, or liquid migration barriers as a part of geotechnical engineering projects. A geomembrane is a class of geosynthetic material composed of an impervious synthetic polymer, elastomer, or plastomer placed as a liner in landfills or surface impoundments to prevent the migration of primary or secondary leachates into the ground. Many of these geomembrane materials can be purchased either unremforced or ------- reinforced with scrim. The typical thickness range of geomembrane materials is from 0.5 mm (20 mils) to 3.1 mm (120 mils). One of the most important factors for liner performance is the quality and stability of the supporting subgrade over which the geomembrane is installed. Because geomembranes are not designed to directly support the overburden load caused by water, soil, or solid-waste material placed on the liner, the nature of the liner subgrade becomes extremely critical. Knowledge of the interactions between the geomembrane and the subgrade is vital to the design of waste storage installations. This technical report is in direct response to SHWRD program objectives to understand geomembrane performance in service environments and the related factors which may cause liner failure. The results of this testing program provide valuable data for specifying geomembrane installations for solid- and liquid-waste disposal facilities and suggest economic guidelines for geo- membranes, geocomposites, and sub- grade preparation. Research Approach and Procedures The objectives of this research program were: (1) to develop an under- standing of the interactions between geosynthetic materials and the sup- porting subgrade in order to determine what factors may lead to premature and/or long-term geomembrane fail- ures; and (2) to determine the economic benefits of alternative liner installations for landfills or surface impoundments which may lead to technical guidelines for the specification of the geosynthetic materials and subgrade preparation requirements. Preliminary Tests Preliminary tests were performed to select the design and shape of the loading points to be used in this program. The criteria for selecting a given load point design were: (1) Load point represented natural load points in texture and angularity; (2) Load point had an easily reproducible shape; and (3) Load point failed geomembranes at a fifty percent failure rate in pre- tests. Figure 1 shows a detailed cross-section design of the load point that was used in this testing program. In addition to selecting the loading point design, tests were conducted to select the initial test conditions which would simulate worst-case field conditions, including pressure loading, temperature, and the height above the plane of the subgrade that the load points protrude. A test matrix was developed and approved. To ensure that the materials tested in this program met or exceeded the physical property specifications stated by their manufacturers, ASTM physical property tests were conducted on each of the subject geomembrane materials. Table 1 lists the ASTM physical property tests that were used to evaluate the geomembrane materials. Test Equipment Long-term hydrostatic puncture- resistance tests were conducted on PVC, CSPE, and HOPE geomembrane material, using two different thicknesses for each type of material. These tests were conducted in 36 round plastic pressure test vessels. The vessels were capable of withstanding pressures of up to 103.5 kPa (15 psig) and temperatures of 50°C (123°F). Figure 2 shows the design detail of the pressure test vessel. A pressure-regulated nitrogen-over- water system was used to force the geomembrane material onto a supporting sand subgrade. Eighteen of the pressure test vessels were connected to a closed-loop hot water circulating system for elevated temperature testing. Three artificial load points were placed in each test vessel to simulate an irregular subgrade surface. The 36 pressure 13.20cm -8.25 cm - Figure 1. Cross-section of the conical loading point. vessels were monitored for temperatu pressure, and geomembrane failure. constant pressure was maintained each of the test vessels by 2 two-sta pressure regulators connected to 17,180-kPa (2,460-psi) nitrogen g bottle. Individual vessel temperature v\ measured by dedicated thermocoup wired to a rotary-style thermocouj switch. Temperature readings we manually recorded at periodic interv from a digital display located on t temperature controller box. Leaks in 1 liner test specimens were detected magnetic float switches located at t base of each pressure vessel. Thirty- indicator lights, mounted on ti temperature controller unit, were wired the individual magnetic float switches. I leak occurred in the geomembrane, wal would flow through the hole and woi accumulate, activating the float swit and indicator light. Approximately 10 of water were required to activate tl switch and indicator light. Daily logs we recorded for each of the 36 pressure te vessels. Economic Analysis Cost data were collected on each the three types of polymeric materic tested and on three thicknesses geotextiles. The majority of cost da were collected by telephone interviev with specialists involved on a day-t day basis with geomembrane tin installation estimates. Publications we used primarily for verification ar corroboration. The cost data were used conjunction with the pressure test data establish the relative expenses fi installing the different types of materia over differing subgrades. Charts ar tables were created to present the da with clarity The analysis results provic a comparison of the costs associate with different combinations of lin< materials and subgrade conditions. matrix was developed to facilitate co comparisons between the various lin< combinations and subgrade finishes. Results and Discussions Laboratory tests were conducted determine the failure mechanisms < geomembrane materials placed over i irregular subgrade and subjected to constant hydrostatic load of 17.93 kPa fi 365 days. The failures during the lov pressure tests are shown in Table 2. Tr 1 5-mm (60-mil) HOPE was the onl tested synthetic material to fail to th limit of water leakage in less than 36 days when subjected to hydrostatic loac ------- Table 1. ASTM Tests on Subject Geomembrane Material Properly Specific gravity Tear strength Tensile properties including breaking factor Hydrostatic resistance Puncture resistance FML material HOPE (1.5 mm & 2.5 mm) PVC (0.5 mm & 1.0 mm) HOPE (1.5 mm & 2.5 mm) PVC (0.5 mm & 1.0 mm) CSPE (0.9 mm & 1.1 mm) PVC (0.5 mm & 1.0 mm) HOPE (1.5 mm & 2.5 mm) CSPE (0 9 mm & 1. 1 mm) HOPE, PVC, and CSPE (2 thicknesses each) HOPE, PVC, and CSPE (2 thicknesses each) ASTM test D792, Method A 0792, Method A 01 004, Die C D751, Method Ba D882, Method A or B 0638 075), Method Ab 075? (Mullen) FTMS 101C, Method 2065 Temperature °C 23 23 23, 50 23, 50 23, 50 23, 50 23, 50 23 23, 50 a With 8-in. x 8-m. test specimens. b Grab method of 17.94 kPa (2.6 psig) with load point heights of 25.40 mm (1 0 in.). The 1.5- mm (60-mil) HOPE material failed after (2) 148 days of continuous testing at 17.93 kPa (2.6 psig). Figure 3 shows Sample 13, 1.5-mm HOPE, which failed. For verification purposes, another specimen of 1 5-mm HOPE was tested and failed (3) after 60 days at ambient temperature, 1794 kPa, and 25.4-mm load height. The failed verification sample is shown in (4) Figure 4. Polyethylene is one of several plastics which exhibit a marked plastic flow and necking at high load levels. Under some conditions, the failure elongation is more than ten times (1000 (5) percent strain). In the prominent failure of 1.5-mm HOPE, Sample No. 13, the following interpreted sequence of events occurred: (1) As the sheet was pressed down on the cone, a general stretching of the (6) material occurred within a radial distance of approximately 50.8 mm around the cone. Near the apex, a localized stress concentration caused elongation of the polyethylene into the necking region of behavior in the direction of the line of symmetry of the cone. At this stage, the elongation in the direction perpendicular to the line of symmetry remained small. The elongation continued until necking ended. At this point, the wall had thinned to about 0.25 mm. This corresponds with a strain of about 550 percent if the lateral strains are still small. Because of the flow during necking, considerable alignment of polymer chains caused stiffening, which strengthened the polymer in the flow direction but reduced the polymer strength in the transverse direction. Continued downward tension over the cone increased the elongation Water Outlet Thermocouple - Air ^ In/Out 50. SO cm Air Pressure Gauge ABS Vessel Water Inlet 25.40cm Double Neoprene Gaskets Geomembrane Specimen Fine Sand - Cone 35° W/ 45° Truncation at Top Drain Valve ~*BS Vessel - Float Switch Figure 2. Cross-section drawing of the hydrostatic pressure vessel. and stress in the transverse direction. Because the strength in this direction was compromised, fractures occurred. This failure mode is a complex process involving a nonlinear thermo- viscoelastic material with an intricate loading history. Loading conditions changed from stress-type to dis- placement-type loading during the test. The stress and displacement fields are two-dimensional and vary with time in each direction. The failure of the 1.5- mm HOPE appears to be a normal failure in that the loading simply extended the material beyond its capability. Characteristically, materials which are biaxially stressed fail at strain levels considerably lower than the failure strain limit in uniaxial tests. SEM examination confirmed that the 1.5-mm HOPE experienced a simple overload failure caused by the stress concentration at the apex of the point loading cone. The thickness of the unstressed portions of the HOPE test specimens were found to vary significantly. Six thickness readings on Sample No. 13 varied from 1.40 mm to 1.62 mm, with an average thickness of 1.47 mm. The thickness of Sample No. 15 varied from 1.47 mm to 1.70 mm, with an average of 1.65 mm. Commercial sheeting of the type tested has a specified thickness tolerance of 10 percent or, in this case, 0.15 mm, so that, except for one point, the sheets were within tolerance. Note, however, that the bending stiffness of the sheet varies with the cube of the thickness. Therefore, based upon the measured average thicknesses, Sample No. 15 is 40 percent stiffer than Sample No. 13. This relative increase in stiffness may explain why Sample No. 15 failed to a lesser extent than Sample No. 13, even though the test temperature was higher. ------- Mate Type HOPE HOPE HOPE trial mm 1.5 1.5 1.5 Sample GeofexMe No. mm 24 15 13 Temp. "C 23 23 50 Press. kPa 17.94 17.94 17.94 Time days 60 148 148 Load height mm 25.40 25.40 25.40 Obviously, this thickness variation is a real-world effect and must be considered when assessing the behavior of geomembrane liners. In addition to the failures displayed by the HOPE material, partial failures were noted in the 0.9-mm and 1.1-mm CSPE tested at 17.94 kPa for 155 day with load point heights of 25.4 mm. A partial failure is defined as a failure that does not compromise the capacity of the liner material to prevent water flow through the membrane to the subgrade. After the partial failures were noted, the hydrostatic pressure tests were continued for an additional 210 days. At the end of this testing period, the pressure was increased to 60.03 kPa for an additional 30 days of testing. A microscopic study was made of the partial failure area of Sample No. 64, which was 1.1-mm CSPE material with scrim reinforcement. The scrim is made from rugged 10 x 10 1000 denier polyester yarn. The partial failure was directly over one of the load points that was 25.4 mm above the subgrade, and the material had been stressed to a maximum pressure of 60.03 kPa for 30 days, resulting in tearing of the lower side of the sheet and scrim breakage in two places. The breakage is shown in Figure 5. Note the buckled shape of the entire area, indicating that the CSPE and/or scrim had been stretched beyond its elastic limit. Figure 6 shows a close- up view of the broken fiber shown on the right in Figure 5. The fiber break itself is unremarkable, but the striated area above it, which appears to be almost fibrous in nature, is a lineation which commonly occurs in plastic flow areas of Figure 3. SEM of failure (21X); Sample No. 13 1.5-mm HOPE. Note: Striations (A) and edge of drawn region IB). polymers. A photograph of the s pattern is shown in Figure 7 roughened appearance of the C surface directly above the fibers indicate that the CSPE and scrim \ stretched beyond the elastic limit of CSPE material. After removal of the I the fibers snapped back, putting plastically deformed CSPE i compression. This compression rest in local buckling of the CSPE. The C material was overstressed whe hydrostatic load of 17.94 kPa was apr. to the material over a 25.4-mm I point. The applied force caused the Ic CSPE material layers to tear and internal scrim to break. At the completion of 365 day; low-pressure testing, the hydrost pressure was increased at a rate of kPa per hour until a pressure of 61 kPa was achieved. This test press was maintained constant for an additi 30 days. The materials that failed high-pressure testing are listed in T; 3. The results of these tests indicate the HOPE 1 5-mm and HOPE 2.5- materials failed at 23°C and 50°C cone heights of 19.0 mm and 25.4 above the subgrade. As previoi discussed for other field samples, th failures are normal events in which material was stressed beyond its ela limit until failure occurred. Of particular interest during accelerated high-pressure tests was accumulation of excessive water in pressure test vessel containing Sanr No. 76, PVC 0.5-mm, at 23°C witl cone height of 19.0 mm. The excess water accumulation led to speculai that the PVC liner material had leaks, a casual inspection of the mate identified no apparent punctures. Only after an intense inspection w five pin-size holes found. These he could only be seen when an intense I source cast illumination directly ber the hole. The holes were located near edges of the sheet and not near the o points; therefore, the areas were highly stressed. It is suggested that ------- Figure 4. View of verification test failure on HDP£ 1.5-mm Sample No. 24. holes might have been manufacturing imperfections in the sheet. Additional visual inspection of the entire liner surface revealed three rows of pits in the surface of the sheet. These pits reduced the load-carrying capacity of the PVC membrane. However, there was no evidence found during optical or SEM inspection that there was any connection between these pits in the material and the pinhole failures. Transient pressure loading tests were conducted by placing new samples of geomembrane material into the test Figure S. Scrim breakage on 1.1 -mm CSPE after loading at 25.4 mm (W2X). vessels and applying pressure at a rate of 55.2 kPa per hour until the final test pressure of 60.03 kPa was achieved. The final pressure of 60.03 kPa was held constant for 24 hours. Load point heights of 25.4 mm, 19.0 mm, and 9.5 mm were used at ambient- and high-tem- perature exposures. The results of these transient pressure loading tests are presented in Table 4. The CSPE 1.1-mm and HOPE 1.5-mm materials failed the test at a pressure of 60.03 kPa with a loading point height of 25.4 mm and a temperature of 50°C. Figure 8 shows the failure of the 1.1-mm CSPE material, and Figure 9 shows the failure of the 1.5-mm HOPE material. The HOPE 2.5-mm material failed at a temperature of 23°C at the final pressure of 60.03 kPa and a loading point height of 25.4 mm and is shown m Figure 10. Examination of the failure areas with an optical microscope indicated that the failures were normal, resulting from overstressmg the geomembrane material in the region of the point load. Tests using geotextile materials were conducted to evaluate the mechanical advantages of a composite liner configuration versus a geomembrane liner material without a protective underlining of geotextile material. A typical composite liner was modeled by placing the geotextile material between the load points and the geomembrane material. Those geomembranes which failed the short-term tests, accelerated tests, and the impact load tests were used in this test sequence. Continuous- filament polyester nonwoven needle- punched geotextiles having thicknesses of 1.5 mm (60 mil), 3.8 mm (150 mil), and 5.8 mm (230 mil) were tested. Woven geotextiles were not evaluated because of their limited use in current liner installation practice. The geomembrane materials tested were HOPE 1.5-mm, HOPE 2.5-mm, CSPE 1.1-mm, and CSPE 0.09-mm. Testing was conducted at 23 °C and 50 °C for a cone height of 25.4 mm. The results of the composite material tests are presented in Table 5. Tests at ambient temperature on the 1.5-mm HOPE liner material combined with a 1.5-mm geotextile at 60.03 kPa caused failure to occur at 210 days. The HOPE 2.5-mm material combined with a 3.8- mm geotextile failed at 50°C and 60.08 kPa after 60 days of testing. The failure mode of the HOPE, in both cases, was a normal failure caused by over-stress on the liner material at the loading point. ------- Figure 6. Close-up of broken scrim fibers 1.5 mm CSPE (SOX) The CSPE material did not fail when a geotextile was placed between the membrane and the load points, and there was no evidence of scrim separation such as that observed in test samples tested without a protective geotextile material. Tables 6 and 7 illustrate the probable costs associated with the installation of tested geomembranes and geotextile materials. Based upon the pressure loading test results, certain composite combination costs are not listed because the subject liner material Figure 7. Close-up of scrim pattern; 1.1 -mm CSPE (11.5X). failed under the given situations stal (Example: HOPE material of 1.5-1 thickness should not be used or subgrade in which particles exceec height of 3.5 mm above the plane of iubgrade preparation to varying level: fineness based on the maximum size; particles which are allowed to remair the surface plane of the subgrade. It is not recommended that any the liner installation options be utilizec lieu of others based solely on c considerations. The cost figures intended to quickly convey exper information concerning each option wl providing the reader with the relal confidence offered by an appropri subgrade finish. A given site nr "require" only a 0.5-mm PVC mate thickness over a 25.4-mm or It irregular subgrade, based on the t results and economic analysis; howev an economically astute owner might el to use a 3.8-mm geotextile at a gi\ facility and/or a finer finish of i subgrade. The geotextile would make liner system more secure and safe fr larger particle penetration by piec which, though absent from the mitic prepared surface, might work their v upwards through the subgrade A fn finish than 25.4-mm particle project would reduce the ultimate potential penetration and cause subsequent faili during the anticipated lengthy (usually years or more) life of the facility Conclusions and Recommendations Geocomposites (geomembrane p geotextile) have a mechanical advantc in improved performance over g« membranes alone The econonr advantage of using a composite syst< is associated with the improved p formance of the geomembrane in p venting a costly waste clean- operation. Further long-term testing required to evaluate the minimi subgrade bedding requirements I geosynthetic liners. New tests shoi emphasize investigations of polyethyle material and should incorporate testing geomembranes over other types irregular surfaces Geognds are synthetic, ope weave, high-flow capacity drainage m that are placed between tv geomembranes to provide mcreas drainage capacity in double-liner stallations. Geognds are used in place traditional natural drainage materials su as sand or gravel and are being us more widely in the liner industry; thi ------- Table 3. High Pressure Failure Data on Prestressed Geomembrane Material Type HOPE HOPE HOPE HOPE mm 1.5 1.5 2.5 2.5 Sample Geotext/le No. mm 6 17 39 32 Temp °C 23 50 50 23 Press. kPa 60.03 60.03 6003 6003 Time days 30 30 30 30 Load height mm 19.00 19.00 1900 2540 they should be included in future testing of synthetic liner configurations. Frequently, high density polythene material is used in the construction of brine storage impoundments and experimental solar ponds for electrical power generation. In both of these applications, elevated temperatures of greater than 95°C are encountered. Therefore, geomembrane testing should be extended to include temperature tests to at least 95°C. Additional tests should be devised to provide insight into the relationship between laboratory loading point geometries and natural loading points for more realistic failure analysis studies. Short-term high-pressure testing, us- ing various loading point shapes, should be done and the results compared with tests using natural loading points to establish practical differences on the mechanics of liner failure." Figure 8. Failed CSPE 1.1 -mm for 25.4-mm loading point. ------- Table 4 Transient Pressure Loading Failure Data Material Type Sample Geotextile Temp. Press. Time Load height No. mm °C kPa days mm CSPE HOPE HOPE 1.1 1.5 2.5 91 12 25 50 50 23 60.03 60.03 60.03 25.40 25.40 25.40 Figure 9. Failed 1.5-mm HLDPE for 25.4-mm loading point. Figure 10. Failed HOPE 2.5-mm for 25 4-mm loading point. ------- Material Samole Type HOPE HOPE Table 6. mm No. 1.5 < 2.5 40* Installation Cost Matrix for Materials including normal cost to install - for one hectare in unit quantity Geotextile Temp. mm °C 1.5 23 3.8 50 Press. Time kPa days 60.03 210 60.03 60 Tested Liners and Geotextiles (Costs in $1,000'S) Subgrade with up to 25.4 Subgrade with up to 19.0 mm projection above mm projection above plane installed plane installed Least Confidence Pond Depth in Meters to Pond Depth in Meters to Load height mm 2540 25.40 Subgrade with up to 9.5 mm protection above plane installed Greatest Confidence Pond Depth in Meters to Material (mm) 1 hec 100 hec 1.8 6.1 1.8 6 1 1.8 6.1 CSPE 0.9 W 1.5 GEO 38 GEO 5.3 GEO CSPE 1 1 W 1.5 GEO 3.8 GEO 53 GEO PVC 0.5 W 1 5 GEO 3.8 GEO 5 3 GEO PVC 1 0 W 1.5 GEO 3.8 GEO 5.3 GEO HOPE 1 5 W 1 5 GEO 38 GEO 5.3 GEO HDPE 2.5 W 1 5 GEO 3.8 GEO 5.3 GEO 82 97 99 106 91 106 109 116 32 42 49 57 52 59 69 77 82 91 99 106 94 106 111 119 59 67 72 79 69 77 82 89 22 32 35 42 37 42 49 57 54 62 67 74 62 67 74 94 10.4 10.4 10.4 104 10.4 10.4 10.4 10.4 10.4 10.4 10.4 10.4 104 104 10.4 104 NR 104 10.4 10.4 10.4 W 4 104 104 10.4 10.4 10.4 10.4 NR' 10.4 10.4 10.4 10.4 10.4 10.4 10.4 10.4 10.4 10.4 10.4 NR" NR 10.4 104 NR" 10.4 NR 10.4 16.1 16.1 16.1 16.1 16.1 16.1 16 1 161 16.1 16.1 16.1 161 16.1 16.1 16.1 16 1 16.1 16.1 16.1 16.1 16.1 16 1 16.1 16 1 16.1 16.1 16.1 16.1 16.1 16.1 16.1 16.1 16.1 16.1 16.1 16.1 16.1 16.1 16 1 16.1 NR 16.1 16.1 16.1 16.1 16 1 16.1 16.1 19.5 19.5 19.5 195 19.5 19.5 19.5 19.5 19.5 19.5 19.5 195 19.5 195 195 19.5 19.5 19.5 19.5 195 195 19.5 19.5 19.5 19.5 19.5 19.5 195 19.5 19.5 19.5 19.5 19.5 19.5 19.5 19.5 19.5 19.5 19.5 19.5 19.5 19.5 19.5 19.5 19.5 19.5 19.5 19.5 NR Not Recommended, liner NR* Not Recommended: liner NR" Not Recommended: liner W With geotextile. and/or geotextile/lmer combination failed at these conditions failed impact loading tests failed both above. ------- Table 7. Probable Costs Associated with Composite Liner Installation Thousands of Dollars 100 p 90 E- 80 ••j wo Hectares • One Hectare PVC 05 PVC Material MM HOPE 1.0 1.5 CSPE 1 1 GEO 1.5 GEO CSPE 0.9 HOPE 25 GEO 3.8 53 Daren L Laine, Michael P. Miklas, Jr., and Charles H. Parr are with Southwest Research Institute, San Antonio, TX, 78284. Charles J. Moench is the EPA Project Officer (see below). The complete report, entitled "Loading Point Pun durability Analysis of Geosynthetic Liner Materials," (Order No. PB 88-235 544/AS; Cost: $19.95, subject to change) will be available only from: National Technical Information Service 5285 Port Royal Road Springfield, VA 22161 Telephone: 703-487-4650 The EPA Project Officer can be contacted at: Hazardous Waste Engineering Research Laboratory U.S. Environmental Protection Agency Cincinnati, OH 45268 10 ------- m 3> _ 2 5 5 > 0) O jf ^ (D c in (D CO 8 C C <"> ft' Jo -j oo -< - n i- .ju M )-«] 3 « y> s :*• sr t"s <: o a o c 5 a C K O a •D m x i HI 01 T3 O CO 2" .8£ m> mm com 1 a ------- |