NIST logo
*

Building Joint Sealant Service Life Prediction

Summary:

The Building Joint Sealant Service Life Prediction Research  project has been initiated by BFRL to deliver critical predictive models for service life to the sealant industry (a significant part of the U.S. building industry) and continues BFRL's global leadership in service life prediction performance-based standards. In this project, high precision sealant mechanical property data is generated from laboratory and field exposures as a function of ultraviolet radiation (UV), temperature, moisture and mechanical loading for use in a reliability-based methodology for service life prediction. The laboratory data is used to develop models to predict and validate the mechanical properties observed in the outdoor exposure.

One of primary activities of this BFRL project is to organize and coordinate an industry-attended Sealant Consortium which meets every 6 months (typically at NIST) to share the latest results from the research, and from the analysis of data collected over the prior 6-month period.  The ITL/SED component in this Project has been to carry out such data analysis in conjunction with consortium meetings, to present the results of the data analysis to consortium members, and to make recommendations as to future experiment design data settings.  This ITL/SED data analysis & presentation cycle has occurred at all 4 of the consortia meetings hosted at NIST.

Description:

The primary building sealant problem is that no method exists to provide high quality data for elucidating the relative importance of the k = 4 major weathering factors:

   X1: temperature,
   X2: humidity,
   X3: ultra-violet radiation, and
   X4: mechanical loading

(and interactions) on the change in pre-failure sealant mechanical properties. A critical attribute of an effective sealant is the ability to span and seal gaps between dissimilar building materials.  These polymeric materials experience daily (~±7%) and yearly cycles (~±25%) of strain deformation.  As these materials are exposed to the weather, molecular changes occur that eventually prevent the sealant from responding to these imposed strains, leading to failure of the sealant.  Characterizing these molecular changes and attributing them to specific exposure factors and conditions will enable the development of models to predict in-service performance. Improvements in sealant material performance and durability will then enable the development of sealant products having longer service life, significantly-decreased maintenance/repair costs, and ultimately result in increased sustainability.

Measuring the mechanical properties of sealant materials is intrinsically challenging due to their inherent non-linear viscoelastic nature. Assessing the molecular changes in the sealant requires straining the sample and measuring the stress response to obtain the sealant modulus. The measured modulus is dependent on specific strain level, time required to impose the strain, and strain history. While a sealant experiences a strain event--either during  testing or within the in-service environment--it is also relaxing those imposed strains on time scales ranging from minutes to months.  As the sealant is  constantly changing and responding to applied strain, assigning change in modulus to the environmental factors is a considerable technical challenge.
Currently used in-field and in-lab methods to assess sealant durability do not have the capability to predict performance.  These methods generally fall into two categories:

   1. Threshold based methods (ASTM C719) which impose a series
      of environmental  exposures after which the sealant is visually
      evaluated, and

   2. Multi-year outdoor exposure tests evaluated with visual
       inspection.

Further, such visual evaluation fails to yield an understanding of the essentially non-linear viscoelastic  modulus, or tan appreciation of the molecular changes occurring during exposure that precede failure. Applied strain and the corresponding stress, which are critical components to  the sealant performance, are not monitored by these current methods.

The BFRL at NIST is perfectly attuned to attack this engineering problem because

   1. the problem is precisely aligned with BFRL's strategic goal of
       "Sustainable  Infrastructure Materials",
   2. BFRL has a unique, controllable,  in-lab weathering facility
       (the NIST SPHERE),
   3. BFRL has the personnel with the  experience, expertise, and
       familiarity with rheological characterization of  elastomeric
       materials, and
   4. BFRL has pre-existing strong relationships with the sealant
       industry.

The Statistical Engineering Division within NIST's ITL is perfectly attuned to collaborate with BFRL on this problem because

   1. It is SED's charter mission to be of ready-assistance to
       NIST mission-critical  projects
   2. ITL/SED has the experiment design, data analysis, and
       modeling expertise  to efficiently carry out and guide the
       statistical components of this problem.

The NIST/BFRL/ITL solution to this problem involves

   1. developing precise modulus-measurement instrumentation,
      characterization and exposure protocols,
   2. developing appropriate experimental sampling plans,
   3. developing multivariable databases,
   4. applying insightful statistical data analysis tools (primarily
       graphical), and
   5. developing quantitative mathematical models.

A series of experiments were carried out to assess the effects, the interactions, and the optimal settings of the 4 principal factors considered to be most critical to the aging of sealants.

Finally, by designing and executing stress-relaxation experiments at NIST to characterize the non-linear viscoelastic modulus of the sealant before and after exposure, and by carefully controlling and monitoring the above 4 factors, molecular changes within the sealant can be correlated to the specific exposure conditions.

Major Accomplishments:

Under the auspices of this project, an industry/government Sealant Consortium has been formed.  NIST/BFRL has hosted 4 meetings of this consortium--on a 6 month basis.

Experiment design recommendations have been given by ITL/SED for future test conditions and experiments.

Data analyses of executed NIST experiments have been carried out by ITL/SED:

    1. to determine the relative importance of the 4 factors:
        temperature, humidity, ultra-violet radiation, and
        imposed mechanical strain;
    2. to ferret out interactions among these 4 factors,
    3. to determine optimal settings of these factors,
    4. to assess the scope and robustness about factor conclusions,
    5. to provide unambiguous insight into the underlying structures
        and conditions affecting sealant performance and durability.
 

BFRL has demonstrated NIST Sphere experimental capability exposure temperatures down to 5 degrees C, and achieved significant reduction in the relative standard uncertainty of this device from 20% to less than 5% (industry-state-of-art is 5% rsu).

State-of-the-art knowledge has been gained on the effect of applied strain on the  water solubility of sealant. This insight helps relate the imposed strain and resulting stress to the chemical changes within the sealant.

A member of the Sealant Service Life Consortium has already adopted the practice of simultaneous exposure to multiple weathering factors in new product development  as pioneered by this project. Two consortium companies have asked for a commercially available version of the SPHERE to facilitate their own sealant testing.

Standards and Codes: Since BFRL is a continued active member of C24--Building Seals and Sealants, knowledge from this Sealant Service Life project has propagated to the larger industry via standards for characterizing building joint sealant. This standard has expected issue date in FY2010.




nistsphere

End Date:

ongoing

Lead Organizational Unit:

itl

Staff:

Chris White, Materials and Construction Research Division, BFRL
Don Hunston, Materials and Construction Research Division, BFRL
James Filliben, Statistical Engineering Division, ITL