Exposure Conditions and Characterization

The four custom-built sealant testing chambers employed in this study have the ability to independently control temperature (±0.2°C), relative humidity (RH) (±0.5 %), UV radiation, and cyclic movement. Because the deformation can be controlled, mechanical characterization tests can be performed without remov­ing the specimens from the chamber. A full description of the chamber design is documented elsewhere [20]. The temperature was controlled with a precision temperature regulator, humidity control was accomplished via proportional mixing of dry and saturated air, and a highly uniform flux of UV radiation was attained by attaching the chambers to an integrating sphere-based radiation source (simulated photodegradation via high energy radiant exposure

FIG. 1—Schematic illustration of the test specimen geometry used (not to scale).

Copyright by ASTM Int’l (all rights reserved); Tue May 6 12:07:08 EDT 2014 Downloaded/printed by

Rochester Institute Of Technology pursuant to License Agreement. No further reproductions authorized.

[SPHERE] [23]). The SPHERE is equipped with a microwave-powered lamp system consisting of six VPS/I600-60 lamp modules. Partially enclosing each light source is a dichroic mirror that removes almost all of the thermal radiant energy (i. e., visible and infrared radiation) from the beam while reflecting the spectra UV emissions into the SPHERE. Thus, without external heating, the temperature in the chamber is about 27° C ± 2°C. A cut-off filter is positioned between the light source assemblage and the SPHERE that prevents almost all of the radiation below 290 nm from entering the SPHERE. It should be noted that no attempt was made in this study to simulate the full spectrum of terres­trial solar radiation or the spectra power distribution of the UV portion of such. Thus, the sealants were exposed to an output in the spectral region between 290 nm and 450 nm and an irradiance of approximately 500 W/m2. A compari­son of the spectral power distribution of the SPHERE radiation source with the reference solar UV spectral distribution from ASTM G173-03 [24] is shown in Fig. 2. Hereinafter, the radiation is referred to as UV for simplicity.

The sealant specimen was attached between a fixed and a movable grip with a computer-controlled stepper motor and a transmission system providing precise movement control. Each chamber had two motors, with four specimen holders on each motor, for a total of eight specimen holders. Each specimen holder was attached to a hermetically sealed load cell with a capacity of ±113.4 kg. Two linear variable differential transformers (LVDTs), one for each motor, with a deflection range of ±6.35 mm were used to measure sealant movement. Data from load cells

FIG. 2—Comparison of the irradiance of SPHERE and the ASTM G173-03 solar spectrum.

Copyright by ASTM Int’l (all rights reserved); Tue May 6 12:07:08 EDT 2014 Downloaded/printed by

Rochester Institute Of Technology pursuant to License Agreement. No further reproductions authorized.

and LVDTs were fed directly into a Keithley 2701 ethemet-based data acquisition system. A custom LabVIEW program was written to collect the voltage measure­ments from the Keithley system every 15 s. The data were averaged once per mi­nute and appended to a tab-delimited database on a remote server.


There were four exposure variables: temperature, RH, cyclic movement, and UV radiation. In all cases, the exposure time was fixed at 1 month. The air tempera­ture was held at 30°C, 40°C, or 50°C. The RH was maintained at 0 %, 25 %, 50 %, or 75 %. Note that exposure to higher levels of RH, liquid water (which allows the possibility for the abstraction of components in the sealants), and freezing conditions is an important area that is not covered in the current study. The deformation involved cyclic movement in a triangular wave varying from 0 % strain to a prescribed maximum strain level at a rate of 38 min/cycle. The maximum strain level was 0 %, 8 %, 15 %, or 25 %. The UV radiation was either on or off. Of 96 possible conditions in the full factorial experimental design, 54 were investigated. These conditions generated a total of 312 data points, which were subsequently trimmed down to 293 after data cleaning to remove out-of-range or suspicious values resulting from faulty collection.

Prior to and after the exposure tests, the specimens were allowed to recover and the mechanical properties of each were characterized at room temperature. The specimens first were subjected to two loading-unloading-recovery cycles at a maximum strain of 26 %. This was followed by a stress relaxation measure­ment at a strain of 18 %. The strain history used is shown schematically in Fig. 3. The loading-unloading tests utilized a crosshead speed of 2.64 mm/min, so the total time under load was 150 s. To allow for viscoelastic recovery, the specimen was held at 0 % strain for 1500 s before the next step. The purpose of the two loading-unloading cycles was to quantify the Mullins effect and elimi­nate its influence in the subsequent characterization test. In stress relaxation measurements, the crosshead speed was 70 mm/min, which meant that the specimen reached the hold strain in just under 2 s. In order to allow for non­instantaneous loading, data points during the first 15 s were ignored.

Downloaded/printed by

Rochester Institute Of Technology pursuant to License Agreement. No further reproductions authorized.

From the stress relaxation data, an apparent modulus Ea was calculated using a relationship based on the statistical theory of rubberlike elasticity [25-27].


W and B = width and breadth of the sealant (Fig. 1), respectively,

L = load, t = time, and

k = extension ratio, which is given by

k = 1 + H (2)


A = crosshead displacement, and

H = undeformed height of the sealant.

From this information, an apparent modulus versus time curve is gener­ated. The magnitude and time dependence of this apparent modulus are related to the molecular structure of the sealant. If the changes in this modulus with ex­posure time are monitored in a degradation experiment, changes in the molecu­lar structure of the sealant can be estimated. Changes in the modulus over time also provide crucial information about how a sealant responds to the stresses imposed by the expansion and contraction of a structure over the diurnal cycle. A modulus ratio F was used to characterize the effect of the environment on Ea.

F Ea (t)

Ea, o(t)

in which Ea (t) and Ea 0 (t) are the apparent moduli before and after exposures, respectively.

The relative effects of the various environmental factors can also depend on the type of evaluation used as the criterion of failure. We believe that changes in the modulus are a clear indication that there are chemical and mechanical changes occurring in the sealant. Initially, these changes might be either detrimental or ben­eficial to sealant performance, but eventually, if the changes become large enough, the performance will likely deteriorate. For the particular material tested here, we have found that a decrease in the modulus is a precursor to cracking and debond­ing, which would allow moisture penetration (the usual definition of failure).