I
LAWRENCE SPENADEL’ and ROBERT J. G O O D 2 Applied Science Research Laboratory, University of Cincinnati, Cincinnati, Ohio
Synthetic Rubber for Self-sealing Fuel Tanks
A (2, 78)
fuel tank can be made by use of a layer of natural rubber in the lining, separated from the fuel by a fuel-resistant-e.g., a nitrilerubber and a diffusion barrier such as nylon. When the wall is pierced by a projectile, the natural rubber at the edge of the wound is exposed to the fuel, absorbs it, and swells, closing the wound. Such tanks were of great value to military aircraft in World War I1 and Korean War. When natural rubber was unavailable, synthetics such as GR-S were used. But all commercial synthetics were inferior to natural rubber in rate of swelling in fuel (3-5, IC-74, ZO), and low-temperature properties were unfavorable. Earlier work (77) had produced a copolymer of isoprene and methylpentadiene with good swelling properties, but unsatisfactory in low-temperature behavior. A copolymer of butadiene-methylpentadiene met all requirements for a good synthetic sealant (79). Its application to self-sealing fuel tanks is new. SELF-SEALING
Preparation of Copolymers The copolymers were prepared by emulsion polymerization in clean crowncapped bottles at 50” C., rotated end over end at 30 r.p.m. for 10 to 26 hours, depending on monomer reactivity, ratio of weights, and desired conversion. Conversions varied from 70 to 9570. Less than 9201, conversion appeared to have little effect on the swelling ability of rubber. This was verified by gel tests.
Recipe Water, ml. Soap flakes, g. Daxad 11, g. NaOH (5.4% in water), ml. Cumene hydroperoxide, ml. Methylpentadiene,a g. Butadiene, g.
120.0
6.0 0.39 15.0 0.35 X 2/
r + y = 75g. a Commercial mixture of 85% 157, 4-methyl-1 ,a-pentadiene.
2- and
After reaction the bottles were removed and cooled. Phenyl-2-naphthylamine and hydroquinone were added. The latex was coagulated, thoroughly washed, and dried in vacuum at 45” C. for 24 hours to constant weight. The copolymer was then compounded with sulfur (0.025 to by0), mercaptobenzothiazole Present address, Esso Research and Engineering Co., Linden, N. J. Present address, Convair Scientific Research Laboratory, San Diego, Calif.
(0.5070), and tetramethylthiuram disulfide accelerator (o.7070), and presscured in a mold for 50 minutes at 1305” C. and 3000 p.s.i. Testing A synthetic sealant must have good swelling ability, low temperature resistance, and mechanical strength. The test for swelling rate was similar to Whitby’s (20). Strips of rubber were placed in a ~~~~ fuel mixture (70y0i s o - o ~ t a n e -toluene) and after a definite time, the swollen strips were weighed. A torsion stiffness tester ( 8 ) was constructed to test low temperature performance. Ten-second readings of the twist of the rubber were taken. By plotting twist against temperature and extrapolating the sharply rising portion of the curve back to the axis, the value for the “freezing temperature” of the rubber could be obtained. Mechanical strength (Goodyear Tire and Rubber Go. test) was evaluated by drilling a 9/~4-inch hole in the sample, and allowing the fuel mixture to flow through the hole until it sealed. Ability to support a 1-foot head of fuel for 24 hours, and subsequently a 4-foot head for a few minutes, indicated satisfactory mechanical strength.
Choice of Polymer System Three-dimensional-Le., cross-linked -polymers swell rapidly at first and then more slowly, with no well defined end value. Scott (15) has shown that swelling can be divided into two fairly definite processes: initial rapid swelling, and secondary swelling. Secondary swelling (the “increment”) proceeds a t a slow and fairly constant rate, and is attributed to rupture of primary chemical bonds. There is more interest here in the initial rate than equilibrium swelling. If a projectile pierces a fuel tank, the wound should close rapidly, before too much fuel is lost. I n general, rate of swelling is a monotonic function of the equilibrium swelling, and equilibrium swelling can be taken as one measure of the rate. I n choosing a rubber that will exhibit maximum swelling, elastomers were considered whose cohesive energy densities are similar to that of the fuel mixture (see table). Examination of these data leads to expectation of optimum swelling for values slightly lower than that of polymethylpentadiene. The average number of methyl side chains should be
6“ (Cal./
Polymer or Solvent CC.)”2 Natural rubber 8.35 (18) Polybutadiene 8.45 ( 1 6 ) 7.75 (f7) Polymethylpentadiene 8.90 (0) Toluene Iso-octane 6.85 (0) Fuel mixture (707, isooctane-30% toluene) 7.2b a 6 = square root of cohesive energy density, Calculated.
slightly less than in polymethylpentadiene; but for optimum low-temperature properties, the number of side chains should be kept to a minimum (6-8)only slightly greater than in polybutadiene. For optimum freedom in varying chain composition, butadiene (B) and meth)lpentadiene (MPD) were chosen as comonomers to be used in various ratios. Other variables were kept constant, and swelling and low-temperature stiffening were tested on various copolymers, to find the optimum ratio. Copolymers varying in weight ratio from 1 : 9 to 9: 1 were prepared by the same recipe, formulation, and cure (1.570 sulfur).
Experimental Results Swelling as a Function of Monomer Ratio. T h e swelling of each sample was measured (Table I and Figure 1). Maximum swelling occurs in the range of 20 to 30y0butadiene.
Effects of Varying Sulfur Content. The swelling, Q, increased rapidly with a decrease in sulfur content of the rubber (Figure 2). The line indicated for natural rubber was obtained from a sample formulated by the Goodyear Tire and Rubber Co. for use in the sealant layer. Samples of 5/5 B/MPD formulated with 0.25% sulfur had greater rate of swelling in the fuel mixture than the natural rubber, although equilibrium swelling was somewhat less. Swelling Results. The theory of mixing two liquids, which gives good results for the solution of polymers in solvents, was applied by Flory and Rehner (7) to the swelling of cross-linked polymers. They obtained the formula:
where V, = molal volume of solvent, $+ = volume fraction of polymer at swelling equilibrium, and p = a parameter derived from the entropy and heat of mixing such that P = ICs
+ [VO‘,(6O- 6,)21/RT
VOL. 51, NO. 8
AUGUST 1959
(2)
935
16
0.34
14
0.31
j - - -I - -
t Natural
12
i
Rubber
\
0.2s
10 0.23
8
0.22
e 6
0
0.19 4
0.16 2
0.13 0 0.10
Figure 1. Maximum swelling of butadiene-methylpentadiene polymers occurs in range 20 to 30% butadiene Table I. Swelling Values of ButadieneMethylpentadiene Copolymers Q’, M1. Fuel c0p01yReaction Imbibed/ G. % Time, mer ratio B / l I P D Conversion HI. Poly. 88.6 88.0 95.5 78.9 92.0 95.8 86.2 86.2 85.5 69.4 76.4 76.4
10 11 16 12.5 17 19 19 19 21 23 26 26
2.26 2.44 3.64 6.00 6.00 4.82 9.00 9.08 14.82 13.90 6.70 5.73
p c = average molal volume of chdins between cross linkages. 8, = square root of the cohesive energy density for the solvent, and 6, = square root of the cohesive energy density for the polymer. The parameter, p s , is an empirical correction term, of the order of 0.3 for most elastomer solutions. An expression (73) for Q (ml. of fuel imbibed per ml. of polymer) at equilibrium has been given : In (1
and
-
+ wh2 + + VA’/S v, qT
Q
=
(1
-
&)/&
=
(3) (4)
where Q ’ = Q/(density of polymer). Because lib’, is a measuie of the number of cross links. if each cross link contains the same number of sulfur atoms, l / V c should be proportional to the amounl of sulfur added. A plot of 1/V, us. per cent siilfur showed this relation to be approximately true, but. at 6% sulfur content the value of 11 V , falls well above the line. This may be due to a variation in the structure of the cross links. T h e decrease in the number of cross links makes it mechanically possible for polymer chains to expand to a greater degree, resulting in a greater equilib-
936
+ a L - - L 0
% Sulfur Added
0
1
2
U
3
4
U 5
i 6
8 Sulfur Added
Figure 2 .
Swelling increased rapidly with decrease in sulfur content
rium swelling (Figure 2, 23). A decrease in cross links probably also increases the number of “holes” in the network, making it easier for the solvent molecules to gain entrance within the bulk, and thus increasing the initial swelling rate (Figure 2, A). Low Temperature Testing. The freezing temperature, Tn was measured foi the rubbers tested for swelling (Tables I and 111). Freezing temperatures comparable to that of natural rubber are obtained in the range of 515 to ? / 3 B/MPD. It was concluded that the 5/5 butadiene methylpentadiene copolymer had the best compromise of properties for a synthetic sealant. O n this basis the 5/5 copolymer was chosen for tests of variation of sulfur content described above. Mechanical Strength Tests. All samples were of satisfactory mechanical strength, including the 5/5 B/MPD sample of 0.25% sulfur content Acknowledgment T h e authors thank the Naval Bureau of Aeronautics for funds that made this investigation possible, under contract K o a s 53-1084-c; ;Milton Orchin for valuable advice; and John K. Blatchford for assistance in the experimental \\rork.
Literature Cited (1) Flory, P. J., Rehner, J., J . Chem. Phys. 11, 521 (1943).
Table II. Molal Volume between Cross Links of 5/5 B/MPD a t Equilibrium as a Function of Sulfur Content a = Sulfur Added, 1/vc wt. yo Q 9r v, x 105
INDUSTRIAL AND E N G I N E E R I N G CHEMISTRY
0.25
0.50 1.0 1.5 3.0 6.0
17.0 0.056 620,000 8.2 0.109 140,000 6.4 0.134 83,000 5.5 0.153 57,000 3.7 0.214 22,500 2.0 0.331 6,300
0.2 0.7 1.2 1.8 4.5 16.0
Table Ill. Rubber
Freezing Temperatures Freezing Freezing Temp., Temp., “C. Rubber O C .
Natural - 62.0 ButadieneMPD-isoMPD ( C o d . ) prene 9/1 -19.5 6/4 - 60.5 Butadiene5/5 -55.0 MPD 4/6 -50.5 10/0 9/1 812
7/3
-80.0 -72.5 -67.0 -62.0
3/7 2/8 1/9 0/10
-38.5 -29.0 -23.5 - 5.0
(2) Frolich, P. K., L.S. Patent 2,497,123 (Feb. 14, 1950). (3) Gee, G., TTans. Faraday SOC.38, 418 (1942). (4) Zbid., 42,33 (1946). ( 5 ) Zbid.,p. 585. (6) Gehman, S . D., IND.ENC. CHEM. 44.730 (1952). (7) Gehman, S.’D., Jones, P. J., Wilkinson, C. S . , Woodford, Ibid., 42, 475 (1950). (8) Gehman, S . D., Woodford, D. E., Wilkinson, C. S., Zbbid., 39, 1108 (1947). (9) Hildebrand, J. H., Scott, R . L., “Solubility of Nonelectrolytes,” Chap. XX, Reinhold, New York, 1950. (10) Rostler, F. S., others, Rubber Aze 58, 585 (1946). (11) Zbid., 59,299 (1946). (12) Zbid., 60, 57 (1946). (13) Zbid., 61, 59 (1947). (14) Salomon, G., Van Amerongen,. G J., J . Polymer Sei. 2, 355 (1947). (15) Scott, J. R., “Physical Chemistry of High Polymeric Systems,” J. Mark, A. V. Tobolsky, eds., p. 275, Interscience, New York, 1950. (16) Scott, J. R., Magat, M., J . Polymer Sci. 4,555 (1949). (17) Sievert, J. A., M. S. thesis, University of Cincinnati, 1953. (18) U. S. Dept. Defense, Specification MIL-T-5578A (April 27, 1951). (19) Whitby, G. S., “Synthetic Rubber,” p. 689, Wiley, New York, 1954. (20) Whitby, G. S., Evans, A. B., Pasternack, D. S., Trans. Faraday SOC.38, 269 (1 942).
RECEIVED for review May 14, 1957 ACCEPTED April 3, 1959 Submitted by Lawrence Spenadel in partial fulfillment of requirements for M.S. degree, University of Cincinnati, 1954.
