Outdoor

sample dimensions for ten days. All samples, from both static tension and com­pression, were returned into the pre-exposure sample dimensions (dotted line) after treatment (Fig. 7).

Results of the stress-relaxation tests on the Sealant A samples exposed to tension are shown in Fig. 8. The data are similar to the dynamically strained data from Fig. 4. The 5 %-strain data are intermediate between the no exposure and exposure with no strain values. Again, these data have a similar curve shape. The 10 %-static strain with exposure data is virtually identical to the no-exposure, no strain data, suggesting that a 10 % static-tensile strain de­creases the modulus as much as the exposure increased the modulus. For val­ues of greater strain, such as 20 and 25 % tension, the rate in change of modu­lus decreases as can be seen in the shallower slope of the stress relaxation curve over time. The 25 % static-strain data, with a value below baseline, had an almost identical curve shape and value to the dynamically strained tension sealant. This finding suggests that a dynamic strain of 7 % tension produces the same amount of molecular change monitored by a change in the complex modulus similar to a 25 % static-strained sample.

The compressively strained Sealant A samples are shown in Fig. 9. In this figure, all values for the statically and compressively strained samples had a

Strain (%)

Подпись: FIG. 7—Width of Sealant A immediately after removal from static outdoor exposure on September 1 (before) and after ten days of forced return to the dimensions prior to exposure (after).

modulus greater than either the samples with no exposure or the dynamically strained (compression) samples. This result suggests that the static – compressive strain on the samples, up to 30 %, was insufficient to produce an effect on the modulus similar to the 7 % dynamic-compressive strain.

The Sealant C samples again had more dimensional stability when re­moved from static strain (Fig. 10). Again, the dimensional changes were pro­portional to the degree and direction of the static strain with the larger strains producing larger changes in sample dimensions. All of these Sealant C samples were able to return to their pre-exposure sample dimensions after ten days under original conditions.

The stress-relaxation data for the statically strained Sealant C samples are shown in Fig. 11. Consistent with the dynamically strained Sealant C samples, the modulus values decreased with the addition of either strain or exposure. Both the 5 and 10 % static tensile-strain values are intermediate between the zero strain and the dynamically strained values. The static strains of 20 and 25 % produced results similar to the dynamically strained Sealant C results. There was no apparent change in curve shape in these results, again consistent with the dynamic data.

The stress-relaxation data for the compressively strained Sealant C samples are shown in Fig. 12. The modulus values of 5 and 20 % compressive strain were comparable to the modulus value of exposure with no strain. The 25 % compressive strain modulus was analogous to the modulus value of the dy­namic compression for Sealant C. This finding suggests that a dynamic strain

1e+6

Подпись: Outdoor: Sample A Tension Подпись:image239"Подпись: CO D_ 8e+5

6e+5

3

T3

о

4e+5

CL

CL

< 2e+5

101

Подпись: 100
Подпись: 102 103 Time (s) Подпись: 104 Подпись: 105

0

FIG. 8—The stress-relaxation modulus for Sealant A shown for prior to any exposure (Baseline), after exposure with no strain (0 % strain), after exposure to static tension of several levels (5, 10, 20, 25 % strain). Also shown is the stress relaxation data for dynamic tension (Summer/tension) from Fig. 4. The points represent the mean values and the error bars are the relative standard uncertainty.

of 1 % compression produces a similar decrease in the modulus as a static compressive strain of 25 %. Again, the decrease in modulus is much smaller in compression than in tension.

For both Sealants A and C, a static strain of 25 % tension produced a similar change in the modulus value to the dynamic-strained samples with an average 1 % strain. This result suggests that even a small dynamic strain creates a much more aggressive strain compared with a long-term static strain. If the sealant had a viscous response on the order of days, it could relax applied strain in a matter of days; consequently, a static strain could be relaxed by lowering the internal stress within the sealant. A dynamic strain would be much more difficult for the sealant to relax by either molecular reorganization or network reformation due to a faster timescale of the applied strain in comparison with the viscous dissipation of the sealant.

Summary

Two sealant formulations (A and C) were exposed to outdoor weathering either with static tensile or compressive strain or with dynamic strain. The dynamic

101

image240
Подпись: 105

FIG. 9—The stress-relaxation modulus for Sealant A shown for prior to any exposure (Baseline), after exposure with no strain (0 % strain), after exposure to static compres­sion of several levels (5, 10, 20, 25 % strain). Also shown is the stress relaxation data for dynamic compression (Summer/compression) from Fig. 4. The points represent the mean values and the error bars are the relative standard uncertainty.

strain on the samples was produced by an instrument composed of two differ­ent materials with different coefficients of thermal expansion. The resulting changes in the complex modulus were quantified with stress-relaxation experi­ments. One of the formulations, Sealant A, exhibited significant dimensional changes in response to the applied strain. These dimensional changes were reversed by maintaining the samples at the pre-exposure sample dimensions for ten days. Sealant A also exhibited two apparent mechanisms that affected the modulus. Exposure with no strain produced an increase in the modulus. With the addition of strain to the exposure, either a reduction in the modulus was observed or a decrease in the modulus and a change in the curve shape were noted.

For Sealant C, a greater degree of dimensional stability was observed, but it also was able to return to pre-exposure sample dimensions within ten days under original conditions. The stress-relaxation data for Sealant C revealed a single mechanism that affected the complex modulus as all exposures either with or without strain exhibited a decrease in the modulus. The decrease in the

Подпись:1.4e+6 1.2e+6 1.0e+6

image241CO

8.0e+5

Подпись:image242
6.0e+5

g 4.0e+5

<

2.0e+5 0.0

100 101 102 103 104 105

Time (s)

FIG. 11 —The stress-relaxation modulus for Sealant C shown for prior to any exposure (Baseline), after exposure with no strain (0 % strain), after exposure to static tension of several levels (5, 10, 20, 25 % strain). Also shown is the stress relaxation data for dynamic tension (Summer/tension) from Fig. 6. The points represent the mean values and the error bars are the relative standard uncertainty.

Outdoor: Sample C – Compression

image243"1.4e+6

1.2e+6

CO

CL

1.0e+6

сл

3 8.0e+5

TO

Подпись: 4.0e+5 < 2.0e+5 Подпись: • 0 % Strain ▼ Summer/compression • 5 % Strain о 20 % Strain о 25 % Strain

6.0e+5

0.0

100 101 102 103 104 105

Time (s)

FIG. 12—The stress-relaxation modulus for Sealant C shown for prior to any exposure (Baseline), after exposure with no strain (0 % strain), after exposure to static compres­sion of several levels (5, 10, 20, 25 % strain). Also shown is the stress relaxation data for dynamic compression (Summer/compression) from Fig. 6. The points represent the mean values and the error bars are the relative standard uncertainty.

modulus was greater for the tensile-strained samples compared with the compressive-strained samples for both static and dynamic applied strains.

Acknowledgments

Support from an industry/government consortium entitled the Service Life Pre­diction of Sealant Materials, centered at National Institute of Standards and Technology (NIST) was critical to the completion of this project. Assistance from DAP, BASF, Dow Corning, Kaneka Texas, SIKA, Solvay, Tremco Wacker, and Rohm & Haas was greatly appreciated. The Department of Housing and Urban Development, through the Partnership for Advancing Technology for Housing program, provided essential financial support for these efforts.