Sealant specimens for two formulations (A and C) were fabricated by our in­dustrial partners (see Acknowledgments) according to ASTM C719 a commonly referenced industrial standard. Sealant A exhibits a reversible stable dimen-


FIG. 1—Illustration of the ASTM C719 sealant sample size employed in these studies.

sional change when exposed to applied strain, while Sealant C does not exhibit a dimensional change when exposed to applied strain. The specimen geometry is shown in Fig. 1. Fabrication involved bonding a sealant sample (W = 50.8 mm, B =12.7 mm, and H =12.7 mm) between two aluminum blocks with dimensions of 76.0 by 12.7 by 12.7 mm. The fabricated specimens were subjected to tension tests by pulling the metal beams apart in a direction per­pendicular to their longitudinal axis and to compression tests by pushing them together in the opposite direction. One limitation of this geometrical system is that the sealant is constrained at the point where it attaches to the metal bars so that it cannot deform laterally (perpendicular to the imposed load) when it is either stretched or compressed. Consequently, the deformations are not uni­form throughout the sample and the modulus calculated from these tests will be somewhat different from that obtained in a tension experiment. Neverthe­less, the wide acceptance of this geometrical system makes it a useful starting place, and as long as an identical geometry is used for all tests, comparisons should be meaningful.

The stress-relaxation experiments were conducted on either a Model 1125 Instron machine with computer control[20] or a custom load-displacement appa­ratus described elsewhere [9]. The main difference between these two machines for the purposes of this study is that the custom apparatus can characterize up to 30 samples simultaneously. The sealants were soft and thus a rigid loading hxture was used so that the crosshead displacement could be obtained as sample deformation with no correction for machine compliance. Deformation is defined in terms of the extension ratio, X = {1 + (Д /H)}, where Д is the dis­placement and H is the sample height. The stress, a, is defined as {LX/ WB}, where L is the load and W and B are the sample width and thickness, respec­tively. These expressions assume that the samples were incompressible, which is generally an acceptable assumption for elastomers. The specimens were loaded rapidly (100 cm/min, cross-head speed) up to a given extension ratio, X, and maintained that value while the load was monitored as a function of time.

Подпись: Ea(t, X) Подпись: 3L(t) WB(X - X-2) Подпись: (1)

The time required to load a specimen was less than 1 s, so the hrst data point was not obtained until after 5 s to avoid transient effects associated with load­ing. An apparent modulus, Ea, was calculated using a relationship based on a statistical theory of rubber-like elasticity [8,10].

The modulus is termed apparent because of the assumptions used in the deri­vation of the equation and the complications discussed above with regard to the constraints on the specimen.

The standard uncertainties for the measurement of specimen dimensions and displacements were ±0.5 mm and ±0.01 mm, respectively. In all but one case, the standard uncertainties in the stress measurement and modulus deter­mination were ±2.7 % and ±5 %, respectively. For experiments that were com­pleted in minutes, such as the stress-strain tests, the uncertainty in the load measurement produced by zero drift was ±0.1 N. The stress relaxation tests, however, required measurements over 10 to 20 h and, under these conditions, the uncertainty in the load measurement associated with zero drift increased to ±0.4 N.