ML19115A239
| ML19115A239 | |
| Person / Time | |
|---|---|
| Site: | 07103052 |
| Issue date: | 02/07/2019 |
| From: | Brehm K, Ecker K, Kowalewsky H, Weise H Govt of Germany, Federal Institute for Materials Research and Testing, US Dept of Transportation, Pipeline & Hazardous Materials Safety Admin, US Dept of Transportation, Radioactive Materials Branch |
| To: | Office of Nuclear Material Safety and Safeguards |
| References | |
| E-53126 IAEA-SM-286/44P | |
| Download: ML19115A239 (10) | |
Text
Enclosure 5 to E-53126 Gas Permeation Report, IAEA-SM-286/44P
PERMEATION THROUGH ELASTOMERJC 0-RING SEALS*
K. BREHM, K.H. ECKER, H. KOWALEWSKY, H.-P. WEISE Federal Io titute for Materials Testing, Berlin (We t)
Abscract PERMEATION THROUGH ELASTOMERJC 0-RJNG SEALS.
IAEA-sM-286/44P Permeation of He, Ne, Ar, Kr, Xe and ~ lhrough 0-ring eals of fluorocarbon rubber (e.g. Vil.on), ethylene-propylene rubber (EPDMJ and silicone rubber (VMQ) has been measured Bl temperatures from 290 K up to !he destruction of the materials, at about 600 to 700 K. Test g pressures ~p up to 2 bar were applied. The quantity of the test gas Iha! penneated was dc!Celed using high sensitivity quadrupole mass spectrometry. To investigate the influence of seaJ geometry, permea-tion through. VMQ O*rings was measured as a function of their compression and compared with measurements of sheet geometry for the same material. The measured time dependence of the permea-tion gas flow i well described by theoretical predictions, when the flow through the 0-ring is considered as the sum of flows through a number of rectangular slabs approximating the shape of the compressed 0-riog. Equilibrium gas flow, the penneation coefficient, solubility, the diffusion. coefficient, the acliva-tion energy of diffusion and the heat of solution arc derived from the measurcmenlS and are compared with values given in the literature, where available.. By extrapolating the results for rare gases He to Xe, permeation parameters for Rn are estimated for the three elastomers under investigation.
- 1.
INTRODUCTION The containment of radioactive gases, including fi sion products, in transport or storage casks is primarily determined by the construction of the sealing system and the choice of the ealing material. Permeation of gaseous radionuclides through the gasket is a major a.ctivity leakage factor when using elastomeric seals. Permeation parameters are usually derived from measurements of the gas flow Q through a eal as a function of time t after the application of a pres ure gradient ~p of a test gas.
In most cases, experimental data are given for the permeation of gas through a sheet and for temperatures up to 400 K. To permit a realistic estimate of the gas release from typical containment systems with O~ring eals at elevated temperatures, we have extended our previous flat-geometry permeation measurements [l] in three ways.
Fir t, in order to relate parameters obtained from measurements using different eal geometries we have measured Q(t) for silicone rubber 0-rings of various compres ions and compared it with measurements on a beet of the same sealing
- Work upported by the Fede.ral Ministry of the Interior, Federal Republic of Germany.
359
360 BREHM et al.
material. Second, measurements were performed up to temperatures where the sealing failed in order to get an idea of the safety margins of a sealing system. Finally, from the range oft.est gases used (i.e. D2, He, Ne, Ar, Kr and Xe), an attempt is made to give a rough estimate of permeation parameters for the radioactive gases tritium and radon, important for design considerations of casks for nuclear wastes.
- 2.
EXPERIMENT Figure 1 is a diagram of the apparatus used for permeation measurements.
Sheet -
as well as 0-ring -
samples up to an outer diameter of 16 cm are mounted in an all-metal, sealed permeation cell, which is he1d in an oven at a temperature in the range between 20 and 500°C. Test gases up to 2 bar are applied to one side of the permeation cell, while the other side is kept at high vacuum by turbomolecular pumps. A quadrupole mass spectrometer (QMS) is used to measure the partial pressure of the test gas on the high-vacuum side. This pressure is proportional to the flow of the permeating gas. A system of two throttle valves permits the handling of a wide range of gas flows with high detection sensitivity, while still maintaining the total pressure in the QMS below 10- 6 mbar, necessary for reliable mass spectro-metry. The system is calibrated with known gas flows through a capillary (for details, see Ref. [2]). The control of the measurement procedure, as well as the data acquisi-tion and analysis, is fully automated using a personal computer with interfaces to the experiment based on the IEE-bus.
winchester
- floppy data lronsf*r-process contr ol lnt*rfoc*
FIG. I. Experimental setup for pennLation measwtments.
IAEA-SM-286("4P 361
-3 Ill '
..J - 4 L..
SJ -5 e
0
~ -6 0
..J l2 16 20 t(m i n)
-4
-5
-6
., -7
~
0
-8 C
-9
-10
-1 1
-12
. 003. 006. 009. 012. 0 1 5 l/t FIG. 2. Permeation gas flow of Ar through a silicone rub~r sheet as a function of ri~ and compared with theory.
- 3.
RESULTS AND DISCUSSION Figure 2(a) hows the gas flow Q as a function of time t for a typical permeation experiment on a silicone rubber sheet. The initial rising part of the curve can be well described by the 'Holstein approximation' [3]:
Q(t) = Q.., (2dNrrDt) exp (-d 2/(4 Dt)) for t < 0.4 d2/D (1)
The equilibrium gas flow (i.e. the asymptotic value for t -
oo) is represented by the con tant factor Qoo = A*P -~p/d (2)
Here A is the surface area of the sheet, d its thickness and p is the difference of partial pre ures of the test gas on the two sides of the sample. P and D are penneation and diffusion coefficients which are coupled by P = oi*D (3)
362 BREHM et al.
- 4 i/
0-rlng compression 23%
...J -s L
!1 One slab J f
,a a *4. l SJ e do*4.7 o, a
~
0
//
1 j
-6 Three slab c:n 111*3. 7 0
..J
- / I dl*4.9
-d,-
d2*3. I
---do---
-d,-
4 s
tCmtnl FIG. 3. Ptmteatio11 gas flow of Ar through a silicone rub~r 0-ring with diflft!nsions 153 x 5. 33 mm, oompressiOII 23% 1111d I bar pressure gradient, compared with theory (- * - * -
one slab; -- three s/JJb opproximmio11).
r 10.,...::,;:.:...._3::;o:.:.o_........:2:;o.:.0 ____
1:.;.0:..0 __
._!:c __ -=;.._
p 1.5 2,0 2,S J,O K"1 3.S 1000,., -----
FIG. 4. PmneaJion coefficients as fanctioru of 1emperai11u for EPDM.
~
p lAEA-SM-286/44P
T 400 300 200 100
- c 20
,o-io........_ ___..._ __
......._ _ _ ~---,--~
3.0 1<-1 J,S 1,S 2,0 2,5 1000 r-1 ----
FIG. 5. Pe~a,ion coefficients as funcrions of tempuaru.re for silicone rubber.
363 with a being the solubility of the gas in the sealing material. Rewriting Eq. (1), a Jinear relationship between In (Q(t).Jt) and Vt can be obtained:
In (Q(tWt) = ln (Q.., 2d/..firf>) - d2/4Dt (4)
For flat amples, the coefficients P, for permeation, and D, for diffusion, are determined by a least squares fit (Fig. 2(b)) of this relation to the experimental data points, with t < 0.4 d2/D. Since Q usually changes by several orders of magnitude during the early stages of an experiment, for the determination of D, in particular, a fit to (4) i preferred to a fit to (1), which would emphasize data points at high values of Q at long times. In such cases, where the equilibrium permeation flow Qoo was actually reached (i.e. for times t > 0.4 d2/D), P is determined from the experimental Qoo using Eq. (2).
For samples with 0-ring geometry, the shapes of the experimental permeation curves Q(t) deviate from the theoretical predictions for flat geometry (Eq. (1)). At short times, e pecially, a considerable deviation is observed (as depicted in Fig. 3) for the case of a silicone rubber 0-ring compressed by 23 %. The dashed--dotted I ine in Fig. 3 is a fit of the data points to Eq. (1) and can be considered as a first-step approximation which a umes the compressed 0-ring to be of rectangular shape.
When inserting the diffusion coefficient D from the flat geometry measurements of the rune material into the 'fitting' parameter di.JD, an effective thickness do for thl zero-order approximation can be derived. It is noted that for the present investigation
364 BREHM et al.
T 10** 400 300 200 100 Viton
,.s 2,0 2,S 3,5 1000 r-1 ----
FIG. 6. Pe~a1ion coefficients as functions of temperature for Viton.
of silicone rubber 0-.rings compressed by 17-34%, the values of do are smaller, by 10-20% as compared with the 'nominal thickness', i.e. the width of a rectangle with the same area as the cross-section of the uncompressed 0-ring.
As a second step, the gas Oow through the 0-ring as the sum of flows through rectangular slabs can be considered. The broken line in Fig. 3 is a fit to the experimentaJ data, assuming three slabs, and is a satisfactory description of the experiment. The insert in Fig. 3 shows the slab dimensions obtained from the fit compared with the theoreticaJ shape of the compressed 0-ring, which has been calcu-lated. from the theory of elasticity [4]. For experiments on rings compressed in the range of 17-34%, a similar good approximation of the theoretical 0-ring shape is achieved by the three-slab model. Thus, together with previous flat geometry measurements of Viton and ethylene-propylene rubber (EPDM) (Eq. (1)), we are now able to deduce permeation data from 0-ring measurements within this range of compressfons.
IAEA M-286/44P 365 TABLE I. ACTIVATION ENERGIES OF DIFFUSION E0 AND HEATS OF SOLUTION Hs FOR GASES IN THREE ELASTOMERlC SEALING MATERIALS (kcaJ/mot)*
Silicone rubber EPDM Viion Gas Eo Rs Bo Hs Eo Rs He 2.3 1.6 4.9 I. I 5.5 1.7 Ne 2.4 0.81 5.0 2.1 7.3 1.3 D2 2.6
-0.11 6.3 0.47 7.9 0.93 Ar 2.7
-0.24 7.3 0.39 u.s 0.43 Kr 2.9
- 0.85 7.8 0.83 13.2 0.06 Xe 3.0
- 1.2
- I calorie = 4. 184 J.
Permeation coefficients as functions of temperature are presented in Figs 4-6.
Similar figures (not shown here) have been obtained for the diffusion coefficient D.
Within the experimental error, most data lie on traight li.nes when log P or log D are plotted ver us 1rr (Arrhenius plot). Thu, for the range investigated here, the classical relationship P = D*a = Do exp (-Eo/Rn*ao exp (- Hs/Rn (5) holds with a single a.ctivation energy for diffusion E0 and one beat of olution H5
- Some deviation from thi behaviour are ob erved for the helium measurements of Viton and EPDM above 200°C. Since in the e cases we have also observed a change in the shapes of the permeation curves Q(t), it is at present not clear whether the deviation is due to a deformation of the 0-ring seal or due to a change in the permea-tion process. Values of Eo and Hs, as derived from the slope of straight lines fitted to the data up to 200°C, are given in Table I. For the range of gases investigated here, in general a decrease in Hs and an increase in B0 are observed with increasing i.z.e of the gas molecules for all three elastomers.
In Fig. 7 the logarithm of the permeation coefficient has been related to the square of the gas molecular diameter, as taken from Ref. [5].. Similar to the findings of Hammon et al. [6] for rare gas permeation in a number of polymers, we also ob erve a good straight line fit for the Ne, Ar, Kr and Xe data while the points for He and D2 are scattered. Extrapolating in Fig. 7 to Rn, approximate permeation coefficients of 8X 10- 9, 8x 10- 11 and 2.Sx 10- 14 m 2/s for Rn in silicone rubber, EPDM and Viton, respectively, were found.
366 BREHM et al.
V-9 S.1teone rubber Rn FIG. 7. Permeation coefficients as fancrions of the square of the gas molecular diameter.
t 10 a
. _........... ---:*t**'
SHicone rubber.-****
J i
.......-********::~:.::.:. :.~.= __ l
................... - v,,on
. /
/.:,*
,/
/
t'
/!,,,,
I'
/i ii
/I
/'
i
/
60 120 180 21.0 t-(min) 00 FIG. 8. Penneation gas flow of He through eulS/omeric O*ring seals as a function of time, during a heating-up aperimenI (see text).
IAEA-SM-286/44P 367 To investigate the rea1i tic behaviour of an elastomeric seal during a possible fire accident, the permeation gas flow for He as a function of time has been measured while the eal was heated up. Figure 8 hows typical experimental re ults, where the temperature incrca e with time (solid line) was chosen to simulate the tempera-ture at the seal of a TN-1300 pent fuel ca kin a fire at 800°C environment tempera-ture [7]. The gas flow hown in Fig. 8 are normalized to a pressure gradient of 1 bar helium across the seal. However, during this particular experiment a gas mixture of 0.5 bar helium and 0.5 bar air was applied for a more realistic approach. The seal failed at temperatures of 380-400°C, corresponding to a fire duration of about 4 h.
REFERENCES
[I) HEUMOS, K., KOWALEWSKY, H., WEISE, H.*P., in Packaging and Transponation of Radioactive Materials (PATRAM '80) (Proc. 6th Int. Symp. Berlin (West), 1980)
(H0BNER, H.W., Ed.), Bundesanstalt fur Materialpriifung, Berlin (West) (1980).
[2)
KOWALEWSKY, H., Tes Methods for the Assessment of the fotegrity of Containment Systems for Radioactive Material, Final Repon, Bundesanstalt filr Matcrialpn1fung, Berlin (West) (1983)
(in German).
(3]
HOLSTE!LN, T., Rep. WCAP-411-9-D, Westinghouse Electric Corporation, Pitisburgh, PA (1954).
[4]
CURRO J.G., SALAZAR, E.A., Rubber Chem. Tcchnol. 46 (1973) 530.
[5]
HJRSCHFELDER, J.O., CURTISS, C.F., BIRD, R.B., Molecular Theory of Gases and Liquids, Wiley, New York (1964).
[6]
HAMMON, H.O., ERNST, K., NEWTON, J.C., l. Appl. Polym. Sci. 21 (1977) 1989.
(7]
Proje.kt Sicherlteitsstudien E!ntsorgung (PSE), Vol. FB7, Bundesanstalt fiir Materialpriifung, Berlin (West) (1985).