The Science of the Total Environment 272 2001.

127 135

Laboratory assessment of exible thin-lm membranes as a passive barrier to radon gas diffusion
W.Z. Daoud, K.J. RenkenU
Uni ersity of Wisconsin-Milwaukee, Mechanical Engineering Department, Radon Reduction Technology Laboratory, 3200 N. Cramer Street, Milwaukee, WI 53211, USA

Abstract This paper presents the experimental results of utilizing a exible thin-lm membrane as a passive barrier to radon gas diffusion. Nine commercially available membranes of various compositions and thicknesses were evaluated as retardant to radon gas diffusion. The radon gas concentration ratios across the thin-lm membranes alone and in combination with an adjacent concrete sample effective diffusion coefcient. were measured in a laboratory system with state-of-the-art instrumentation. An 8.89-cm diameter, 10.2-cm thick concrete sample of standard composition wrc s 0.5 and cement:sand:gravel s 1:2:4. was used to simulate a basement and slab-on-grade foundation typical of Wisconsin. The radon gas transport characteristics of this concrete sample porosity, permeability and diffusion. are documented. The experimentation has identied two superior exible thin-lm membranes that may be employed as effective barriers to radon gas diffusion. These include: Polyethylene Naphthalate 7.62 = 10y5 m. and Polyethylene Terephthalate Glycol, PETG 7.62 = 10y5 and 1.27 = 10y 4 m. which had average diffusion coefcients, D, of 4.10 = 10y14 and 1.66 = 10y14 m2 sy1 , respectively. Measurements of the effective membranerconcrete diffusion coefcient yielded a further average reduction in D of 98% for the Polyethylene Naphthalate and 96% for the PETG. Details of the experimental set-ups and procedures are described. The results of this investigation have shown that the application of an effective thin-lm membrane adjacent to an intact concrete slab can signicantly reduce the diffusion of radon gas entry. Therefore, the employment of a exible thin-lm membrane should be considered as a viable radon reduction technology method for residential new construction. 2001 Elsevier Science B.V. All rights reserved.
Keywords: Radon; Diffusion; Films; Membrane; Concrete

Corresponding author. Tel.: q1-414-229-5755; fax: q1-414-229-6958. E-mail address: renken@uwm.edu K.J. Renken.. 0048-9697r01r$ - see front matter 2001 Elsevier Science B.V. All rights reserved. PII: S 0 0 4 8 - 9 6 9 7 0 1 . 0 0 6 7 6 - 3

128

W.Z. Daoud, K.J. Renken r The Science of the Total En ironment 272 (2001) 127 135

1. Introduction It has been known for sometime that convective ow is the main driving force behind elevated radon levels in homes and buildings due to the reduction of pressure relative to the surrounding soil. Diffusion is usually considered the second major driving force and can sometimes result in high indoor radon concentrations. Therefore, barriers that can retard both mechanisms will increase the resistance of the building against radon penetration Nowak and Song, 1990.. Recently, passive mitigation methods have proved efcient in preventing radon entry into buildings. One method is the application of thinlm membranes at new residential constructions to prevent radon gas penetration. In this paper, an accurate laboratory method to measure the effective diffusion coefcient of membranes in combination with a concrete sample is presented. The method and system used to characterize the transport properties of thin-lm membranes and concrete samples utilized the apparatus described by Daoud 1998. and Maas and Renken 1997.. Several materials have been investigated to identify efcient membranes as barriers. These included previously tested materials as well as commercially available membranes that had potential as a superior radon gas barrier. After conducting a literature search, it was discovered that few studies have addressed the issue of radon penetration through thin membranes and even less have discussed radon diffusion through membranes in combination with concrete, which exemplies the real application. Pohly Rulling et al. 1980., Jha et al. 1982., Hafez and Somogyi 1986., Mosley 1996., Nielson et al. 1996., Perry and Snoddy 1996. and Sanchez et al. 1996. have addressed the measurement of radon resistance of thin lm membranes. The basic theoretical and practical laboratory testing in these studies consists of a radon source chamber, a test specimen and a collection chamber for penetrated radon. The present study uses a steady state, one-dimensional transport model that allows for reasonable test run times to measure the radon gas diffusion coefcient through exible,

thin-lm membranes adjacent to a typical composition intact concrete slab.

2. Background The diffusion of radon gas through concrete samples utilizes Ficks Law and the diffusion equation Veith, 1991., J s yDC D 2 C y C s C t 1. 2.

where, J is the radon ux through concrete per unit cross sectional area, D is the diffusion coefcient, C is the concentration of radon per unit volume of air, and is the decay constant for radon. Since one-dimensional ow through the sample is maintained, neglecting the decay term and assuming steady state conditions, Eq. 1. reduces to D dC s const s J dx 3.

By sampling the chamber concentrations at the beginning and end of the diffusion run the average radon concentration gradient and the radon ux can be calculated and then used in Eq. 3. to calculate the radon gas diffusion coefcient. The decay term was neglected in Eq. 2., but this was compensated for by calculating the amount of radon that would have been present in the chambers if decay would not have occurred. Fig. 1 is a schematic of the diffusion apparatus used in this investigation. It consists of two chambers: a radon source chamber and a collection chamber. The radon source chamber was connected to a radon source loop, which contained a commercially available passive radon gas source to build-up the radon gas concentration. The collection chamber on the rear end of the sample was used to collect the radon diffused through the sample. By monitoring the radon levels in both chambers, the radon ux and concentration gradient was calculated. The volume of the radon

W.Z. Daoud, K.J. Renken r The Science of the Total En ironment 272 (2001) 127 135

129

Fig. 1. Diffusion apparatus used to measure both concrete and concrete-membrane diffusion coefcients.

source chamber was approximately 0.008 m3 to minimize the dilution effect during sampling of the chamber. These two chambers were separated by the thin-lm membrane andror concrete sample. Several sensors were used to monitor the conditions under which the diffusion experiment was performed. These sensors included two relative humidity sensors, three pressure sensors and two thermistors. The entire apparatus was contained in an environmental chamber, which controlled the temperature and relative humidity. Continuous radon monitors were used to measure the radon concentrations in both chambers. These monitors utilize a scintillation cell and a photomultiplier tube to count the number of alpha emissions given off by the radon gas present. All sensors including the radon monitors were conTable 1 Membranes tested for diffusion coefcient a Commercial name KALADEX Film Gard PS8010 PS8010 PT6200 PT6211 ULTROS ULTROS VALOX
a

nected to a modern PC-data acquisition system, which was controlled by a real-time display data acquisition software package.

3. Methodology Flexible thin-lm-membranes are dened as solid sheeting materials e.g., plastic, polyethylene, polyester, etc... Table 1 lists the thin-lm membranes and the corresponding thicknesses that have been solicited from US companies who specialize in the manufacturing of membranes. The concrete samples used in this investigation simulated typical Wisconsin poured-concrete basement foundations. The ingredients for our concrete included cement, sand, stones and water.

Material type Polyethylene Naphthalate Polyethylene Polyester Polyester Polyether Polyether PETG PETG Polyester

Average thickness m.a 7.62 = 10 1.52 = 10y4 2.54 = 10y5 7.62 = 10y5 2.54 = 10y5 7.62 = 10y5 7.62 = 10y5 1.27 = 10y4 1.27 = 10y4
y5

Manufacturer DuPont Carlisle Plastics Deereld Urethane Deereld Urethane Deereld Urethane Deereld Urethane Lustro Plastics Lustro Plastics GE Plastic

Experimental uncertainty of the thickness is "0.00127 mm.

130

W.Z. Daoud, K.J. Renken r The Science of the Total En ironment 272 (2001) 127 135

Pea gravel was used to obtain a homogeneous mixture with respect to the small size of the samples. An average value of water to cement ratio of 0.50 was used and the average ratio of the concrete ingredients i.e., cement:sand:gravel. was 1:2:4 Hool, 1918.. Cylindrical aluminum holders were used to cast and hold the concrete samples. The cylindrical shape was chosen to ensure that one-dimensional transport was applied. The samples were 10.1 cm in thickness and 8.9 cm in diameter. The concrete samples were removed from their holders 24 h after casting and placed in a high humidity chamber for 30 days. After curing, the samples were allowed to dry at ambient conditions for 1 week. The samples were then placed back into the cylindrical holders and the edges were sealed with a laboratory proven cementitious epoxy such that one-dimensional diffusion of radon gas was prevalent. The concrete samples were placed into the diffusion apparatus to determine the radon gas diffusion coefcient. The procedure for the diffusion measurement is as follows. The chambers were exhausted at rst and the radon gas was then allowed to circulate in the chamber. After reaching steady state conditions, two calibrated continuous radon monitor units were connected to the two chambers. The collection chamber was exhausted to ensure a zero level of radon gas. The PC-data acquisition system collected data for

at least 4 h. Then, the radon continuous monitor unit, which was connected to the source chamber, was replaced to eliminate errors in the radon concentration readings, and the program and the detectors were reinitiated again. This procedure resulted in two text les created by the controlling data acquisition program. These two les were retrieved by a spreadsheet software package to calculate diffusion coefcients. After testing the concrete samples for diffusion, the diffusion apparatus was prepared for the thin-lm membrane tests. Both ends of the sealing surfaces were cleaned to remove the sticky residue left by sealing material. After cleaning, new sealing gaskets were placed on both ends of the collection chamber and source chamber. These rubber gaskets were 15.2 = 15.2 cm and 0.3 cm thick. A hole of 10.1 cm in diameter was fabricated using a cutting edge to ensure the full frontal facial area of the sample was exposed to the radon gas. A schematic of the modied setup is shown in Fig. 2. A lm-membrane sample of 12.7 = 12.7 cm was selected and visually examined to verify that it was free of any apparent defects or foreign materials. In the same fashion as the concrete sample diffusion tests, the membrane sample was placed between the two gaskets in the apparatus, the system was sealed-off and tested for leakage and the diffusion test runs were initiated. The nal set of diffusion tests measured the effective diffusion

Fig. 2. Schematic of the gasket-membranerconcrete conguration.

W.Z. Daoud, K.J. Renken r The Science of the Total En ironment 272 (2001) 127 135 Table 2 Concrete test results Characteristics Porosity, Permeability, k Diffusion coefcient, D Values 7.1 13.7% 6.51 = 10y1 1 1.51 = 10y 12 cm2 1.12 = 10y8 m2 sy 1

131

Table 3 Comparison of radon gas diffusion coefcients for concrete D m2 sy 1 . 1.69 3.08. = 10 2.30 3.70. = 10y9 0.46 1.8. = 10y8 1.84 3.83. = 10y8 4.38 14.7. = 10y8 2.06 3.93. = 10y8 0.97 1.22. = 10y8
y8

Study Culot et al. 1976. Leung et al. 1994. Rogers et al. 1994. Snoddy 1994. Renken and Rosenberg 1995. Maas and Renken 1997. Present study

coefcient of the combination of thin-lm membranerconcrete sample. During these tests, the order of the thin-lm membrane and concrete sample was reversed. The thin-lm membrane front facial area was rst exposed to the radon gas with the concrete sample adjacent to the collection chamber. Then, the concrete sample was exposed to the Source Chamber with the thin-lm membrane on its rear. Diffusion test runs were between 12 and 76 h. 4. Experimental results A summary of the experimental results for this study are represented in Tables 2 6. More specically, the concrete radiological transport characteristics, the average diffusion coefcient for selected thin-lm membranes, the effective diffusion coefcient for the membrane concrete combination and the average percent reduction in the diffusion coefcient of the membrane concrete combination in comparison to the concrete are documented. 4.1. Concrete test results The physical properties of the concrete mix
Table 4 Average diffusion coefcient of the thin-lm membranes Commercial name KALADEX Film Gard PS8010 PS8010 PT6200 PT6211 ULTROS ULTROS VALOX Material type

used in this research that affect radon gas transport, namely porosity, permeability coefcient and diffusion coefcient are listed in Table 2. These properties dependent on the type of the aggregate used, waterrcement ratios and curing conditions. 4.1.1. Concrete porosity The concrete porosity measurements used the procedure of Maas and Renken 1997.. The concrete porosity, of the samples used in this study varied from 7.1 to 13.7% with an experimental uncertainty of 3% Daoud, 1998.. It has been reported Auxier, 1973. that humidity has an effect on porosity and consequently on the diffusion of radon through concrete. In this investigation, the humidity level was approximately zero. Auxier 1973. found that with increasing the relative humidity of concrete from 60 to 80%, radon transport decreased by 30%. This work is the only quantitative study of the effect of humidity on radon exhalation from building materials Colle et al., 1981..

Average thickness m. 7.62 = 10y5 1.52 = 10y4 2.54 = 10y5 7.62 = 10y5 2.54 = 10y5 7.62 = 10y5 7.62 = 10y5 1.27 = 10y4 1.27 = 10y4

Average diffusion coefcient m2 sy 1 . 4.10 = 10y 14 1.64 = 10y 11 1.10 = 10y 12 1.03 = 10y 12 6.29 = 10y 12 7.19 = 10y 12 2.51 = 10y 14 1.66 = 10y 14 2.30 = 10y 12

Polyethylene naphthalate Polyethylene Polyester Polyester Polyether Polyether PETG PETG Polyester

132

W.Z. Daoud, K.J. Renken r The Science of the Total En ironment 272 (2001) 127 135

Table 5 Comparison of radon diffusion coefcients for thin-lm membranes a Material Jha et al. 1982. m2 sy 1 . Hafez and Somogyi Nielson et al. Perry and Snoddy Mosley 1996. 1986. 1996. 1996. m2 sy 1 . 2 y1 2 y1 2 y1 m s . m s . m s . 1.9 = 10y 10 6. 1.43 = 10y 10 5. Present research m2 sy 1 .

Natural 6.36 = 10y1 0 rubber Cellulose 1.24 = 10y1 1 nitrate Cellulose acetate Polyvinyl5.00 = 10y1 1 chloride Polyethylene Polyethylene naphthalate Polyethylene terephthalate Polyethylene terephthalate glycol Polyester 1.95 = 10y1 3

2. 0.6. 7.50 = 10y1 3 1. 21. 5.8 = 10y 13 6.1 = 10y1 3 7.8 = 10y1 2 0.4. 0.4. 2.8.

3.36 = 10y 11 9.7 = 10y 11 6.

8.81 = 10y 12 6. 1.64 = 10y 11 6. 4.1 = 10y1 4 3.

3.0 = 10y1 3 0.5. 1.7 = 10y1 4 2.5 = 10y1 4 1.1 = 10y 12 2.3 = 10y1 2 1.0 = 10y1 2 6.3 = 10y1 2 7.2 = 10y1 2 5. 3. 1. 5. 3. 3. 1.

0.6.

4.3 = 10y 11 0.5.

Polyether Poly-carbonate Mylar

a

3.82 = 10y1 3 1. 8.36 = 10y1 4 0.7.

2.4 = 10y 12 0.6. 5.5 = 10y1 3 0.6.

Numbers between brackets are the thickness of the thin-lm membranes in units of mil 1 mil s 2.54 = 10y5 m..

4.1.2. Air permeability of concrete Experimental runs were carried out on the concrete samples at pressure differences of 15 100 Pa. The average permeability coefcient for the concrete batch ranged from 6.51 = 10y1 1 to 1.51 = 10y1 2 cm2 with standard deviations,
Table 6 Summary of the concreterthin-lm membrane diffusion test results Commercial name KALADEX Film Gard PS8010 PS8010 PT6200 PT6211 ULTROS ULTROS VALOX Material type Average thickness m. 7.62 = 10y5 1.52 = 10y4 2.54 = 10y5 7.62 = 10y5 2.54 = 10y5 7.62 = 10y5 7.62 = 10y5 1.27 = 10y4 1.27 = 10y4

of "0.46 = 10y1 1 and "0.26 = 10y1 2 cm2 at differential pressures of 15 and 100 Pa, respectively. As detailed in Daoud 1998., the method used to calculate the permeability coefcient maintained a steady-state pressure difference throughout the test. All permeability tests found in the

Density g cmy 3 . 1.36 0.92 1.21 1.21 1.14 1.14 1.27 1.27 1.34

Average Deff . m2 sy 1 . 1.52 = 10y 10 5.79 = 10y 10 8.26 = 10y 10 2.02 = 10y 09 8.02 = 10y 10 1.09 = 10y 09 5.05 = 10y 10 3.26 = 10y 10 1.90 = 10y 10

Average % reduction in D 98.3 91.7 83.4 71.0 87.3 89.4 95.4 96.1 96.6

Polyethylene Polyethylene Polyester Polyester Polyether Polyether PETG PETG Polyester

W.Z. Daoud, K.J. Renken r The Science of the Total En ironment 272 (2001) 127 135

133

literature used a leak method where the pressure difference dropped during the test according to the permeability of the concrete sample. 4.1.3. Diffusion in concrete The average diffusion coefcient of the concrete mix used in this research was 1.12 = 10y8 m2 sy1 with a standard deviation of 0.31 = 10y8 m2 sy1 . The initial time to reach steady state was based on the work of Maas and Renken 1997.. Table 3 compares the radon gas diffusion coefcient through concrete of the present study to previous investigations. As indicated, the results of this research are similar to the other published works. 4.2. Thin-lm membrane diffusion test results Table 4 reports the radon gas diffusion coefcient of the thin-lm membrane samples. Two materials that have been identied as a superior barrier to radon penetration are Polyethylene Naphathalate KALADEX, 7.62 = 10y5 m. with an average diffusion coefcient of 4.1 = 10y1 4 m2 sy 1 and Polyethylene Terephthalate Glycol ULTROS, 7.62 = 10y5 and 1.27 = 10y 4 m. with average diffusion coefcients of 2.51 = 10y1 4 and 1.66 = 10y1 4 m2 sy1 , respectively. The experimental uncertainty of the diffusion coefcient was estimated to be approximately "10% Daoud, 1998.. Comparing the present results to published values shows some differ-

ences especially with materials of the same type as shown in Table 5. These differences are expected and can be related to differences in material density, chemical composition, manufacturing, etc. A comparison of the polyethylene material to previous studies showed our results were comparable. As shown in Table 4, the new materials that were tested showed superior resistance to radon penetration. Diffusion experiments used variable radon gas concentrations that ranged from 36 000 432 000 Bq my3 . The thickness of the thin-lm membrane has a minimal effect on the resistance of radon penetration. 4.3. Concrete r thin-lm membrane diffusion test results Table 6 summarizes the concreterthin-lm membrane combination diffusion test results. Under the conditions of the test, Polyethylene Terephthalate glycol ULTROS, 7.62 = 10y5 and 1.27 = 10y4 m. and polyester VALOX, 1.27 = 10y4 m. were partially effective in resisting the radon penetration with average effective diffusion coefcients of 5.05 = 10y1 0 , 3.26 = 10y1 0 and 1.90 = 10y1 0 m2 sy1 , respectively. As shown in Fig. 3, Polyethylene Naphthalate KALADEX, 7.62 = 10y5 m. samples were more effective in slowing diffusion. This average effective diffusion coefcient was 1.52 = 10y1 0 m2 sy1 which resulted in a percent reduction in the diffusion coefcient as compared to an intact concrete sample of 98.3%

Fig. 3. Effective diffusion coefcient for each thin-lm membrane sample.

134

W.Z. Daoud, K.J. Renken r The Science of the Total En ironment 272 (2001) 127 135

Fig. 4. Percentage reduction in the diffusion coefcient of concrete.

as illustrated in Fig. 4. The results also showed that changing the direction of ow has minimal effect on the value of diffusion coefcient. Other materials that showed excellent retardation to radon gas penetration were ULTROS PETG, 7.62 = 10y5 m., ULTROS PETG, 1.27 = 10y4 m. and VALOX Polyester, 1.27 = 10y4 m. and resulted in an average reduction in the diffusion coefcient of the concrete of 95.4, 96.1 and 96.6%, respectively. The relationship between the material density reported by the manufacturer and the reduction in the gas diffusion coefcient is also examined in Table 6. Here, we see that the density of these materials has a signicant effect on their ability to retard radon penetration. Polyethylene Naphthalate KALADEX, 7.62 = 10y 5 m. which showed the highest resistant to radon penetration has the highest density at 1.36 g cmy3 . Similarly, VALOX and ULTROS also have relatively high densities compared to other materials of 1.34 and 1.27 g cmy3 , respectively.

lene Naphthalate KALADEX. has a high potentiality to retard radon penetration with a diffusion coefcient of 4.1 = 10y1 4 m2 sy1 . This conclusion was also supported by the results found by overlaying this material to the surface of a concrete sample and determining the effective diffusion coefcient of the combined system. This resulted in an average effective diffusion coefcient of 1.52 = 10y1 0 m2 sy1 with an average radon penetration reduction of 98.3%. Therefore, thin-lm membrane utilization should be considered as an effective method of reducing radon gas diffusion in residential construction.

Acknowledgements The authors would like to thank Dr Conrad V. Weiffenbach, State of Wisconsin Department of Health and Family Services for his suggestions. We would also like to thank the following companies for their donation of the thin-lm membranes: GE Plastic, Lustro Plastics, Deereld Urethane and DuPont. Special thanks to Mr Greg Barske, ME Instrument Maker for his ideas and fabrication of the test systems. References
Auxier JA. Contribution of natural terrestrial sources to the total radiation dose in man. PhD Thesis, Georgia Institute of Technology, Atlanta, Georgia, 1973. Colle R, Rubin RJ, Knab LI, Hutchinson JMR. Radon transport through and exhalation from building materials: a

5. Conclusions Diffusion coefcient measurements were performed on exible thin-lm membranes, concrete and on thin-lm membranes in combination with concrete. Measurements have shown that thin-lm membranes are useful tools in radon-resistant construction. The laboratory experiments with the thin-lm membranes determined that Polyethy-

W.Z. Daoud, K.J. Renken r The Science of the Total En ironment 272 (2001) 127 135 review and assessment. National Bureau of Standards, Technical Note 1139. Washington, DC, 1981. Culot MVJ, Olson HG, Schiager KJ. Effective diffusion coefcient of radon in concrete, theory and method for eld measurements. Health Phys 1976;30:263 270. Daoud WZ. Assessment of exible thin-lm membranes as a radon-resistant construction technique. Master thesis, University of Wisconsin, Milwaukee, Wisconsin, 1998. Hafez A, Somogyi G. Determination of radon and thoron permeability through some plastics by track technique. Nucl Tracks 1986;12:697 700. Hool GA. Concrete engineers handbook. New York: McGraw-Hill, 1918. Jha G, Raghavayya M, Padmanabhan H. Radon permeability of some membranes. Health Phys 1982;42:732 735. Leung JKC, Ng C, Tso MW. Radon release from building materials in Hong-Kong. Health Phys 1994;67:378 384. Maas JJ, Renken KJ. Laboratory assessment of cementitious coatings as a barrier to radon gas entry. In: The 1997 AARST International Radon Symposium, 2 5 November 1997, Cincinnati, OH, 1997; Session III-Mitigation: 1.1 1.13. Mosley R. Description of a method for measuring the diffusion coefcient of thin-lms to 222 Rn using a total alpha detector. In: The 1996 AARST International Radon Symposium, 29 September 2 October 1996, Haines City, FL, 1996; Session II-Radon Measurement: 2.1 2.12. Nielson KK, Holt RB, Rogers VC. Residential radon resistant construction feature selection system. EPA-600rR-96-005. US EPA, Research Triangle Park, NC 27711; 1996. Nowak M, Song BH. Evaluating radon resistance of lms and

135

sealants using peruorocarbon tracer gases. In: The 1990 International Symposium on Radon and Radon Reduction Technology, 19 23 February 1990, Atlanta, Georgia, 1990; Vol. V: VIII-4. Perry R, Snoddy R. A method for testing the diffusion coefcient of polymer lms. In: The 1996 AARST International Radon Symposium, 29 September 2 October 1996, Haines City, FL, 1996; Session III Radon Mitigation: 1.1 1.9. Pohl y Rulling J, Steinhausler S, Pohl E. Investigation of the suitability of various materials as 222 Rn diffusion barriers. Health Phys 1980;39:229 301. Renken KJ, Rosenberg T. Laboratory measurements of the transport of radon gas through concrete samples. Health Phys 1995;6:800 808. Rogers VC, Nielson KK, Lehto MA, Holt RB. Radon generation and transport through concrete foundations. EPA600rR-94-175. Research Triangle Park, NC: US EPA, 1994. Sanchez DC, Minga R, Sloan C. An evaluation of indoor radon reductions possible with the use of diffusion resistant exible construction membranes. In: The 1996 AARST International Radon Symposium, 29 September 2 October 1996, Haines City, FL, 1996; Session III Radon Mitigation: 3.1 3.13. Snoddy R. Laboratory assessment of the permeability and diffusion characteristics of Florida concretes, phase I, methods development and testing. EPA-600rR-94-053. Research Triangle Park, NC: US EPA, 1994. Veith WR. Diffusion in and through polymers. New York: Hanse Publishers, 1991.