Category Durability of Building and Construction Sealants and Adhesives

Standard laboratory tests were performed as part of a routine comparison of the two sealant binders and to augment the exterior and accelerated weathering data. Results are presented in Table 5 and Fig. 12.

Adhesion and Joint Movement Performance-The acrylic sealant has excel­lent wet and dry peel adhesion to mortar, with no adhesive failure and peel strengths substantially higher than the 5 lbf (22.2 N) required by ASTM c920- 05. The polyurethane sealant also passes ASTM C920-05 peel adhesion require­ments but with somewhat lower peel strengths. The acrylic sealant passes ±25 % ASTM C719-93 joint movement testing on concrete mortar with no adhesive or cohesive failure. The polyurethane sealant unexpectedly fails this test in the first room temperature cycle. Since the manufacturer’s technical data sheet (TDS) clearly indicates that this sealant passes ±25 % joint movement testing to mortar, the premature failure noted in this evaluation may be due to the fact that these sealants were tested without the use of primer. The adhesion and

joint movement properties in Table 5 and in the polyurethane TDS are consis­tent with the El Paso warehouse exposures where mortar adhesion is excellent and where the joint movement capabilities of the sealants are clearly adequate for the movement encountered.

Hardness and Tensile Properties—The acrylic sealant has substantially greater elongation than does the polyurethane sealant, a property generally associated with higher performance and greater joint movement capability. The measured hardness and stress values of the acrylic sealant are also higher than those of the polyurethane sealant but well within the ranges that are typical for

ASTM C920-05 Class 25 formulations. These measured differences are appar­ent in the field, where samples of acrylic sealant pulled from a joint are slightly harder and stiffer than pulled samples of the polyurethane sealant.

When comparing the tensile properties of sealants based on different chem­istries, it is essential to do so with an understanding of the differing viscoelastic natures of these chemistries. Acrylic sealants are typically more viscoelastic than are more heavily cross-linked sealants based on reactive chemistries such as silicones and polyurethanes. Because of this, the mechanical properties of acrylic sealants are more strain rate dependent than are those of sealants based on reactive chemistries. When deformed quickly (such as at the rates typically found in the laboratory), acrylic sealants are often harder and stiffer (with higher stress or modulus values) than their reactive chemistry counterparts. However, when deformed slowly (such as at the rates typically encountered outside), the properties of acrylic sealants fall in line with those of alternative chemistries.

The differences in the strain rate response of the tested acrylic and poly­urethane sealants are illustrated in Fig. 12, where stress at 25 % elongation (a measure of sealant stiffness) is plotted against the rate of tensile testing. Tensile measurements are generally done at rates of testing which are convenient for generating data in a timely manner. ASTM D412-06a, the commonly referenced “Standard Test Methods for Vulcanized Rubber and Thermoplastic Elastomers—Tension” [10] specifies that tensile testing be done at 20 in. (508 mm)/min. The author’s laboratory routinely uses 2 in. (51 mm)/min for tensile testing as a more reasonable compromise between timely data generation and real world deformation rates. However, both of these testing rates are several orders of magnitude greater than the rate of deformation in ASTM C719-93 joint movement testing (2 X 10-3 in.(5 X 10-2 mm)/min) and the rates of joint movement likely to be encountered in exterior low rise masonry buildings (2-5 X 10-4 in.(5-13 X 10-3 mm)/min) [11,12]. These high rates of tensile testing over emphasize the differences in mechanical properties between strain rate independent elastomeric sealants and more strain rate dependent vis­coelastic sealants.

When the rate of sealant testing is reduced from 2.0 in. (51 mm)/min to a more appropriate rate of 0.02 in (0.5 mm) /min, the stress of the acrylic sealant at 25 % elongation converges on that of the polyurethane sealant, minimizing the perceived differences between the two sealants.

Conclusions

The El Paso warehouse exposure provides a unique, side by side, comparison of a high performance acrylic sealant to a commercial two part polyurethane seal­ant. After 3 years of exterior exposure the polyurethane sealant continues to function as a sealant, with good adhesion and adequate joint movement capa­bility for the application. However, the polyurethane sealant exhibits consider­able crazing, chalking, and softening as the result of exposure, and the function of at least one sealant joint appears to have been compromised by crazing through to the underlying backer rod. Accelerated weathering data generally support these observations.

After 3 years of identical exposure, the acrylic sealant also continues to perform as a sealant, with good adhesion and adequate joint movement capa­bility. The acrylic sealant has comparable dirt pickup to the tested polyurethane sealant and better exterior durability (i. e., little crazing and no chalking or softening). Laboratory test results and accelerated weathering data support and confirm these results. The lack of plasticizer in the acrylic sealant formulation eliminates plasticizer migration and dirt pickup of coatings applied over the sealant joint. Feedback from the moisture-proofing contractor suggests that the properties of the wet acrylic sealant require minimal adjustment for optimal application and that the water cleanup of the acrylic sealant is a distinct ad­vantage from convenience, safety, and environmental points of view.

The data presented herein represent the results of 3 years of exterior and laboratory testing of a high performance acrylic sealant and a commercial two part polyurethane sealant. The results, of course, pertain to the sealants tested and are not necessarily representative of the performance of all high perfor­mance acrylic and polyurethane sealants. Certainly there are commercial ure­thane products on the market, which will out perform the product tested in this evaluation. However, the combined results of this side by side comparison clearly demonstrate the adhesion, durability, and aesthetic benefits of the tested acrylic sealant. While these results cannot necessarily be extrapolated to all high performance acrylic sealants, they do suggest that these products can be highly suitable for use in low rise industrial applications such as tilt-up ware­houses.

[1]FlackTech, Inc., 1708 Highway 11, Bldg. G, Landrum, SC 29356.

[2]SURA Instruments GmbH, Jena, Germany.

[3]DELO Industrial Adhesives, Windach, Germany.

[4]Rocatec™-Pre, Rocatec™-Plus, ESPE™ SIL: 3M Deutschland GmbH, Neuss, Germany.

[5]Plasmatreat GmbH, Steinhagen, Germany.

[6]LM = Low Modulus; 25: ±25% movement capability according to ISO 9047; G = Glazing applications.

[7]Test condition: irradiation energy= 180 W/m2 (300-400 nm), black panel temperature at 63 ° C, water spray for 18 min within 120 min weathering cycle, SUGA TEST INSTRU­MENTS SX-120.

[8]Test condition: black panel temperature at 63 ° C, water spray for 18 min within 120 min weathering cycle.

[9]All of the compared materials, silicone, polysulfide, and STPE, are commercial products of low modulus two-part sealant in Japan.

[10]Test condition: panel angle = 60°, direction = south, at Takasago, Hyogo, Japan.

[11]Test method of oil resistance: a piece of cured polymer (2 mm thickness) was immersed in certain oil at certain condition exhibited on Table 4; then the sample was taken out, swiped on the surface, and the weight change from the original was measured at room temperature.

[12]Silicone sealant was tested in the same condition for IRM 903 oil. After the test, bleed­ing out of the oil was observed on the surface of test piece. Such a phenomenon was not observed for STPA.

[13]JASS-8-2008 shows appropriate choice of sealant materials for each type of construc­tion application, and STPE and PU are not suitable for glass glazing use.

[14]This assumption neglects inertia dominated cases like bomb blast loading and related high speed phenomena.

[15]Fast curing mortar according to the manufacturer’s description.

[16]Nominal 50 mm/min.

[17]In addition, one should note that in the numerical model, the interfaces are geometric boundaries with stepped properties, while in nature, interfaces at least of the glass sur­faces show a different behavior on the micro-scale level [10].

[18]In this case, only load transfer along the symmetry axis is considered.

[19]An adequate flange length Lf is assumed.

[20]Certain commercial materials and equipment are identified in this paper in order to specify adequately the experimental procedure. In no case does such identification imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply necessarily that these items are the best available for the intended purpose.

[21]ISO 11600-F-25LM; ASTM C920, Type S, Grade NS, Class 35, Use NT, M, A, G, and 1; Canadian Specification CAN/CGSB-19.13-M87, Classification MCG-2-25-A-N.

[22]JIS A 1414-1: Water-Spray rate 0.4 L/(min-m2), Maximum test-pressure: 2303 Pa, Minimum test-pressure: 49 Pa.

[23]ASTM E331-00: Water-spray minimum rate of 3.4 L/(min/m2), Test-pressure of at least 137 Pa.

[24]Konica Minolta Sensing Americs, Inc., Ramsey, New Jersey 07446, USA. The Minolta Model CR-231 Chroma Meter color analyzer has a 25 mm diameter measuring area, 45° illumination angle, and 0° viewing angle. Illuminant: D65. Color measurement according to ISO 7724. Color-coordinates: CIELAB.

[25]Atlas Material Testing Technology, Chicago, Illinois 60613, USA. The Ci65A Xenon Weather-Ometer has a 6500 W water cooled xenon arc lamp and a total exposure area of 11 000 cm2.

[26]Q-Laboratory Corporation, Cleveland, Ohio, 44145, USA.

[27]PVC is the fractional volume of a pigment in the total volume solids of a dry paint film.

[28]Tinius Olsen Testing Machine Co., Inc., Horsham, PA 19044, USA.

To assess their usefulness in predicting real world weathering results, the seal­ants exposed in the El Paso warehouse exposure were also subjected to several accelerated weathering tests. Results are presented in Tables 2-4 and in Figs. 8-11.

Weathering of Plaques in Xenon Arc and Fluorescent UV Weathering Devices – The results of the accelerated weathering of sealant plaques exposed in the xenon arc and fluorescent UV weathering devices are detailed in Table 2 and in Figs. 8 and 9. In the fluorescent UV, the polyurethane sealant plaque starts to craze after approximately 500 h and exhibits severe crazing after 2200 h (Fig. 8). These crazes extend as far as 1 mm into the bulk of the specimen. The

acrylic sealant plaque, in contrast, exhibits no visible flaws after an extended exposure period of 6400 h in the fluorescent UV (Fig. 8). The polyurethane sealant in the xenon arc device looks similar to that in the fluorescent UV, exhibiting significant crazing after 2200 h. The acrylic sealant plaque exhibits no visible flaws in the xenon arc device after a comparable exposure period. Although there was no attempt to quantify crazing as a function of exposure time, the accelerated weathering results for both the acrylic and polyurethane sealants are consistent with the appearances of these two sealants after exterior exposure in El Paso. The early crazing of the polyurethane sealant plaque after accelerated weathering is similar to the crazing seen on the vertical joints of the polyurethane sealant after 3 years in El Paso. The deep crazing of the polyure­thane sealant plaque after longer periods of accelerated weathering is consis­tent with the appearance of the polyurethane sealant in the parapet joints after 3 years in El Paso.

Sealant Weathering Apparatus Cycles Hours Failure3 Visual Appearance Polyurethane Fluorescent UV 5 3916 0.35 in.2 (226 mm2) C at interface Fine surface crazing Acrylic Fluorescent UV 5 3916 None No flaws Polyurethane Xenon arc 5 3571 1.25 in.2 (806 mm2) C at interface Severe crazing; up to 3 mm deep Acrylic Xenon arc 5 3571 None Slight crazing aFailure is reported as the total area of cohesive plus adhesive failure over three specimens. C = cohesive failure.

The polyurethane sealant softens during accelerated weathering in both the fluorescent UV and the xenon arc. This phenomenon was not quantified due to the fact that the sealant plaques used for weathering are too thin for reliable hardness measurements. However, the softening is pronounced, and the weath­ered polyurethane sealant after 2200 h feels gummy and significantly less tough than when it was originally exposed. This observation is consistent with the softening of the polyurethane sealant in the parapet joints in El Paso. The hardness of the acrylic sealant appears to change little during accelerated weathering.

In addition to general weathering, sealant color change was also evaluated as a function of exposure time in the fluorescent UV (Fig. 9). During the first 1500 h of exposure the polyurethane sealant yellows and darkens, as measured by changes in L* and b*. After 1500 h this trend is reversed as the sealant starts to chalk and, as a consequence, lighten. The color change of the polyurethane sealant during accelerated weathering was not noted in the El Paso warehouse exposure presumably because it was obscured by early dirt pickup. However,

the chalking of the polyurethane sealant seen during accelerated weathering is consistent with the chalking observed in the field. The color of the acrylic seal­ant remains essentially unchanged after exposure in the fluorescent UV, and chalking after 2200 h is insignificant. This lack of chalking is likewise consis­tent with observations of the acrylic sealant after 3 years of exterior exposure in El Paso.

ASTM C1519-04 Durability—The results of ASTM C1519-04 durability test­ing are reported in Table 3 and in Fig. 10. After five cycles and 3916 h of weathering in the fluorescent UV, the ASTM C1519 polyurethane sealant H block specimens exhibit fine surface crazing and have a total of 0.35 in.2 (226 mm2) cohesive failure at the interface with the aluminum substrate. Under the same conditions, the acrylic sealant specimens exhibit no surface degradation and no adhesive or cohesive failure. After five cycles and 3517 h in the xenon arc apparatus, the ASTM C1519 polyurethane sealant H block speci­mens are severely crazed, with the crazes extending up to 3 mm into the bulk of the specimens. The specimens also have a total of 1.25 in.2 (806 mm2) cohe­sive failure at the substrate interface (Fig. 10). The acrylic sealant specimens exhibit slight surface crazing after xenon arc weathering but show no signs of adhesive or cohesive failure (Fig. 10).

The ASTM C1519 polyurethane sealant H block specimens weathered in the fluorescent UV (Table 3) show substantially less surface degradation after 3916 h of exposure than the plaques of polyurethane sealant weathered for 2200 h in the fluorescent UV (Table 2). The faster surface degradation of the polyurethane sealant plaques may be due to the thinner cross section of the plaques and to the greater impact of heat and moisture on bulk sealant prop­erties. Or it may be that the heat capacity of the H block specimens is higher than that of the plaques, resulting in less condensation during fluorescent UV exposure and less overall exposure to surface wetting. The greater degradation in the thin cross sections is consistent with the installer’s observations that some polyurethane sealants in the field tend to “burn-though” or degrade when applied in thin cross sections or over backer rod that has been installed at an insufficient depth.

The ASTM C1519 polyurethane sealant H block specimens weathered in the xenon arc apparatus (Table 3) generally look similar to the plaques of poly­urethane sealant weathered in the same device (Table 2). Both develop a dense network of cracks. However crack depths vary from roughly 1 mm in the ex­posed plaques to as much as the 3 mm in the ASTM C1519 H block specimens. The greater crack depths in the ASTM C1519 H block specimens are likely due to the longer exposure time (3571 h vs 2200 h) and to the repeated cycles of ±25 % joint movement that are part of the ASTM C1519 durability test.

The degradation of the ASTM C1519 polyurethane sealant H block speci­mens is significantly greater after weathering in the xenon arc device than it is after weathering in the fluorescent UV (Table 3). This may be due to the broader spectrum of irradiance of the xenon arc light source or to the fact that speci­mens in the xenon arc see prolonged exposure to significantly higher tempera­tures than they do in the fluorescent UV. A comparison of accelerated to exte­rior weathering of the polyurethane sealant indicates that testing according to ASTM C1519 in the xenon arc apparatus (Fig. 10) generally predicts the worst of the UV degradation seen in the El Paso exposure (Fig. 6).

The ASTM C1519 acrylic sealant H block specimens weathered in the fluo­rescent UV look similar after 3916 h of exposure to the plaque of acrylic sealant weathered for 6400 h in the fluorescent UV. Neither exhibits any signs of sur­face degradation and the accelerated thin film degradation of the polyurethane sealant does not happen to the acrylic sealant. The lack of degradation after both accelerated fluorescent UV exposures is consistent with the lack of degra­dation of the vertical acrylic sealant joints after 3 years of exterior exposure in El Paso.

The ASTM C1519 acrylic sealant H block specimens weathered in the xenon arc apparatus exhibit a slight amount of surface crazing, while the plaque of acrylic sealant weathered in the same device has none. This is likely due to the longer exposure of the ASTM C1519 specimens (3571 h vs 2200 h) and to the repeated cycles of ± 25 % joint movement. The fact that the ASTM C1519 acrylic sealant H block specimens craze slightly after xenon arc weath­ering but not after similar times of fluorescent UV weathering may be due to the full daylight spectrum of the xenon arc lamp or to a greater exposure to elevated temperatures. Testing according to ASTM C1519 in the xenon arc ap­paratus (Fig. 10) predicts the slightly greater crazing of the acrylic sealant in the horizontal parapet joints in El Paso (Fig. 7).

The ASTM C1519 polyurethane sealant H block specimens soften signifi­cantly after weathering in the xenon arc apparatus but do not appear to soften after weathering in the fluorescent UV. The softening after xenon arc exposure is consistent with that seen after thin plaque weathering and after exposure in the parapet joints in the El Paso warehouse. The fact that noticeable softening of the polyurethane sealant only occurs on specimens which exhibit substantial surface crazing suggests that there is a connection between these two phenom­ena.

Exterior Exposure in Static Joints—The results of the Spring House Farm exposure in static joints are summarized in Table 4 and in Fig. 11. The un­coated polyurethane sealant picks up relatively little dirt but crazes and discol­ors upon exposure. This discoloration is consistent with that seen in the fluo­rescent UV exposure—initial yellowing followed by whitening due to chalking. The uncoated acrylic sealant shows no signs of degradation but picks up more dirt than the polyurethane sealant. These dirt pickup observations differ from the El Paso data, which indicate that the two sealants are very similar in ap­pearance after 1, 2, and 3 years of exterior exposure. The reduced dirt pickup of the polyurethane sealant in the static joint relative to the exterior joints in El Paso may be due to the fact that the sealant exposed in the static joint was cured for several weeks prior to exposure, thus eliminating the early high tack, high dirt pickup phase of the polyurethane cure. The increased dirt pickup of the acrylic sealant in the static joint relative to the exterior joints in El Paso may be due to the different climate of the southeast PA exposure or to the fact that the static joints were exposed horizontally, which allows for less run-off and self-cleaning than do the vertical tilt-up joints in the El Paso exposure.

The EWC appears to be compatible with both the polyurethane and the acrylic sealants, with no apparent debonding at the coating/sealant interface. The EWC coated polyurethane sealant exhibits substantial dirt pickup. This, presumably, is due to plasticizer migration from the polyurethane sealant into the EWC. The EWC coated acrylic sealant exhibits very little dirt pickup due to the lack of plasticizer in the acrylic formulation.

Tilt-Up Warehouse Exposures

Early Observations-The applicator found that the acrylic sealant needed to be handled somewhat differently than the polyurethane sealants that he was used to working with. As mentioned earlier, the acrylic sealant sent to the job site was not optimized for application properties. The applicator, in fact, found that the acrylic sealant was lower in viscosity than the polyurethane sealant and had a shorter open time. The lower viscosity made the acrylic sealant easier to gun but more difficult to tool since it offered less resistance to the tooling imple­ments. The shorter open time meant that the acrylic sealant needed be tooled more frequently and after shorter application lengths. Since the applicator was aware that the acrylic sealant would shrink more than the polyurethane seal­ant, there was some attempt to apply the acrylic sealant at slightly greater application depths. However, this process was not optimized or quantified in this exposure series. At the end of each day, the applicator of the acrylic sealant used soap and water to remove residual sealant from his tools. The applicator of the polyurethane sealant used toluene.

After 24 h the polyurethane sealant was still tacky and glossy and had picked up a fair amount of dirt. The tack and gloss, however, decreased over the succeeding days. The acrylic sealant skinned to a smooth, nontacky surface in less than 1 h and picked up much less dirt than the polyurethane sealant in the first 24 h. From the pieces of sealant pulled from the joints, it was determined that the polyurethane sealant cured through in 5-7 days and that the acrylic sealant took closer to 30 days. From these pulled pieces it was also found that the cured acrylic sealant is slightly harder and stiffer to the touch than the cured polyurethane sealant.

Longer Term Observations—As determined visually and in comparative pho­tographs, the dirt pickup of the two sealants after 1, 2, and 3 years is compa­rable. Dirt pickup, as might be expected, varies from elevation to elevation. However, from a distance, and from any elevation, the acrylic and polyurethane sealants look similar. After 3 years and at a distance of 20 ft (6 m), the acrylic and polyurethane sealants on the east side of the building (Fig. 3) are indistin­guishable. Closer views of the two sealants on the north side of the building (Fig. 4) also reveal a similarity in overall appearance. On more detailed inspec­tion, however, differences between the two sealants become apparent. A close-up comparison of the two sealants on the east elevation of the warehouse (Fig. 5) reveals considerable crazing of the surface of the polyurethane sealant but no apparent crazing of the acrylic sealant. This is a trend seen throughout the warehouse exposure—after 3 years of exposure, the polyurethane sealant exhibits considerable crazing in all vertical joints in all elevations. The acrylic sealant, in contrast, exhibits no visible crazing on any of the vertical joints.

The polyurethane sealant also exhibits severe crazing in all vertical and horizontal parapet joints (Fig. 6). This crazing was readily apparent after 2 years of exposure and has become significantly more pronounced during the third year of exposure. In addition, the polyurethane sealant in one of the corner parapet joints has developed crazes that extend several millimeters into

the sealant bead and which may be compromising the functional performance of the sealant. The acrylic sealant, after 3 years of exposure, exhibits slight surface crazing or wrinkling in the horizontal joints at the top of the parapet (Fig. 7) but no visible degradation in any of the vertical parapet joints. This crazing or wrinkling was not present after 2 years of exposure. The greater degradation of the sealants in the parapet joints is due to the fact that these sealants are subjected to harsher exposure conditions than those installed at ground level. Sealants installed vertically around the interior of the parapet see reflected UV from the light colored roofing material; those installed in the hori­zontal joint at the top of the parapet see direct and continuous exposure to the sun.

The polyurethane sealant chalks markedly after 2 and 3 years of exterior exposure, as measured by transfer of white residue from the sealant surface to the inspector’s index finger. Chalking is particularly severe in the parapet joints. The acrylic sealant, in contrast, chalks little in either the vertical or the parapet joints. The polyurethane sealant also softens noticeably during exposure, par­ticularly in the parapet joints where the exposure conditions are most severe. The acrylic sealant does not change noticeably in hardness during exposure.

When sealant samples were pulled from representative joints to test for adhesion, both sealants pulled out easily due to substrate failure within the friable concrete mortar in the decorative aggregate surface layer. However, the mortar is holding up adequately for the joints experiencing movement, and no joint failures have been noted to date. Functional performance of both of the applied sealants appears to be intact, with the possible exception of the poly­urethane sealant in the parapet corner joint noted above.

To complement the exterior and accelerated exposures described above, three key ASTM C920-05 tests were run. These included adhesion, joint movement testing, and hardness. Adhesion to concrete mortar was tested according to ASTM C794-06 [8], with dry adhesion being tested after the initial 3 weeks of cure and wet adhesion after additional 1 week of water soak. Joint movement to concrete mortar, at ±25 %, was tested according to ASTM C719-93 (2005). Hardness was tested as per ASTM C661-06 [9].

Tensile properties were also measured as a general indicator of sealant performance. Samples for tensile testing were laid up as 1/8 in. (3.2 mm) thick wet plaques on PTFE foil covered aluminum panels and cured as below. Dumb­bell shaped specimens, with a gauge length of 0.725 in. (18.4 mm), were cut from the dried plaques and tested on a model H10K-S Tinius Olsen[28] tensile tester. Tensile testing was done at 2 in. (51 mm)/min and at 23±2 ° C and 50±5 % relative humidity. Elongation to break, maximum stress, and stress at 25 % elongation were measured and reported as the mean± standard deviation of three measurements. The stress at 25 % elongation was also measured at 0.2 in. (5.1 mm)/min and 0.02 in. (0.51 mm)/min to assess the strain rate sensitivity of the two sealants.

In the four laboratory tests described above, the acrylic specimens were cured for 1 week at 23±2 ° C and 50±5 % relative humidity, followed by 2 weeks at 50 ° C. The polyurethane specimens were cured for 3 weeks at 23±2 ° C and 50 ±5 % relative humidity.

In addition to the warehouse exposures described above, the durability of the two sealants was also evaluated using standard accelerated weathering proce­dures and ASTM C1519-04, Standard Practice for Evaluating Durability of Building Construction Sealants by Laboratory Accelerated Weathering Proce­dures.

Weathering of Plaques in Xenon Arc and Fluorescent UV Weathering Apparatus-Specimens for accelerated weathering were laid up as 5 X1^ X ^in.3 (127 X 38 X 3.2 mm3) thick wet plaques on 3 X 6 in.2 (76 X 52 mm2) aluminum panels. The specimens were cured for 3 days at 23±2 ° C, 50±5 % relative humidity, and then placed in either the xenon arc or fluorescent UV accelerated weathering apparatus, following the procedures described below. Changes in sealant surface appearance (e. g., crazing, pitting, and chalking) were monitored periodically over a minimum of 2000 h. Changes in sealant color, as measured by changes in L*a*b* [3], were measured as a function of time in the fluorescent UV apparatus using a Minolta CR-231 portable Chroma Meter color analyzer.[24]

Procedure for Exposure in Xenon Arc Light Apparatus-An Atlas Ci65A Xenon Weather-Ometer[25] was equipped with daylight filters conforming

to ASTM Practice G155 [4]. The exposure cycle was 102 min of light followed by a wet period of 18 min light with water spray. The irradiance was set to 0.51 W/(m2 ■ nm) at 340 nm and the chamber air temperature to 45 ° C. The uninsulated black panel temperature was measured at 68 ° C.

Procedure for Exposure in Fluorescent UV Apparatus-A QUV Accelerated Weathering Tester (model QUV/basic) from Q-Lab[26] was equipped with

fluorescent UVA-340 lamps that comply with the spectral power distri­bution specifications in ASTM Practice G154 [5]. The exposure cycle consisted of 8 h of UV exposure at an uninsulated black panel tempera­ture of 60 ° C, followed by 4 h of wetting by condensation at an uninsu­lated black panel temperature of 50 ° C. Irradiance was not controlled. ASTM C1519-04 [6] Durability-Three aluminum H block specimens, as de­scribed in Test Method C 719-93 [7], were made for each of the two sealants tested. The acrylic sealant specimens were cured for 1 week at 23±2 ° C, 50±5 % relative humidity, followed by 2 weeks at 50 ° C. The two part polyure­thane sealant was mixed immediately prior to sample preparation with a paddle mixer and following the manufacturer’s instructions. The resulting specimens were cured for 3 weeks at 23±2 ° C, 50±5 % relative humidity. Fol­lowing cure, the specimens were placed into either a xenon arc or fluorescent UV weathering apparatus. After 4 weeks the specimens were subjected to six room temperature cycles of ±25 % cyclic movement at a rate of 1/8 in. (3.2 mm)/h. They were then evaluated for overall appearance and amount of adhe­sive or cohesive failure. The cycle of weathering followed by joint movement is an ongoing process, and the test will continue until significant failures have occurred. Results to date, through a total of five cycles, are reported.

Exterior Exposure in Static Joints-Channels for static joint exposures were fabricated by nailing pine strips to a plywood base to form a series of 4 X 2 X 30 in.3 (19 X 13 X 762 mm3) channels. To prevent degradation of the sub­strate, the wood was first primed and then painted with a high quality exterior paint. After the paint was dry, the channels were filled with sealant. The seal­ants were tooled flat and flush with the tops of the channels and then cured for 4 weeks under ambient conditions. To assess sealant coatability and the appear­ance of overcoated sealants, half of each sealant was coated with a 38 pigment volume concentration (PVC)[27] all acrylic elastomeric wall coating (EWC). The coating was brush applied in two coats to the sealant and channel surfaces at a combined coating weight which resulted in a calculated final dry film thickness of 20 dry mils (0.5 dry mm). The first coat was dried for 24 h before the appli­cation of the second coat, and several additional days elapsed before the EWC coated channels were taken outside for exposure. The filled and coated chan­nels were exposed horizontally in a south-45° direction at the Spring House Farm in southeastern PA. After 1 year of exterior exposure the channels were brought back into the laboratory, and the coated and uncoated sealants were assessed for dirt pickup, crazing, and chalking.

A tilt-up warehouse in El Paso, TX, was renovated in the Spring of 2005. As part of this renovation, the failing 20 year old sealant used in the original construc­tion was removed and replaced with the acrylic and polyurethane sealants de­scribed above. To directly compare performance, these two sealants were ap­plied in alternating joints around all four elevations of the warehouse, for a total of 62 joints. The sealants were applied by a moisture-proofing contractor with over 60 years in the business using two applicators with over 48 combined years of experience in the field. One applicator installed all of the polyurethane sealant, and the other installed all of the acrylic sealant.

The warehouse was constructed of concrete tilt-up panels with a decorative aggregate surface finish (Figs. 1 and 2). The aggregate finish is made up of a large aggregate embedded in a soft friable concrete mortar. Because of the size of the aggregate, the aggregate layer is roughly 3/4 in. (19 mm) deep. Although it would have been preferable, from an adhesion point of view, to install the sealants against the underlying concrete panels, it was not possible to do so

because of the depth of the aggregate layer. In both the original installation and in the retrofit described below, the sealants were installed flush with the build­ing exterior and against the aggregate layer.

The joints for the replacement sealants were formed by cutting out the old sealant, widening the joints where necessary, and then smoothing the inner edges of the joints with a grinder (Fig. 2). The dust generated by the grinding process was cleaned off with a blower, and the joints were brushed off prior to the application of the sealant. The resulting joints varied from 1/2 in. (13 mm) to over 1.5 in. (38 mm) in width (Fig. 2). A closed cell backer rod was used for all sealant applications and was carefully inserted so that it remained convex and untwisted. The sealant was applied with a bulk loading gun. Since much of the application was done on a ladder, the applied joint lengths were generally limited to 4-6 ft (1.2—1.8 m) at a time. Each section of sealant was tooled immediately after application. Although tilt-up buildings in the El Paso area are often painted, the warehouse used for the current exposure was not top-coated due to its decorative aggregate finish.

During the initial construction of the warehouse, the tilt-up panels were welded together at built-in weld plates, theoretically limiting the movement of the panels relative to one another. However, movement indicators attached to representative joints at the roofline indicated that some joints exhibited sub­stantial movement while others exhibited none. Measured movements ranged from 14 to 22 % (total) of the initial joint widths, substantially below the 50 % total movement capabilities of the applied sealants. The contractor felt that the amount of movement measured in these joints, and the variable nature of the movement, was typical for tilt-up construction in the El Paso area.

Applicator comments about the use of acrylic sealants were noted through­out the installation process and reported below. To assess application quality, sealant adhesion, and sealant cure, small sections of sealant were periodically cut and pulled from representative joints. To assess aesthetics, durability, and functional performance, the sealants were visually inspected and photographed at irregular intervals over a period of 3 years. Sealant dirt pickup, gloss, and crazing were visually assessed and readily captured in photographs. Sealant chalking was gauged by rubbing the sealant surfaces with an index finger and assessing the amount of white residue transferred from the sealant to the finger. Sealant softening and tack were subjectively gauged by digging at the sealant joints with a finger nail and pressing with fingertips.

Sealants

A high performance acrylic sealant and a high performance polyurethane seal­ant, both conforming to the ASTM C920-05 Class 25 specification, were chosen for this evaluation.

The polyurethane sealant was specified by the moisture-proofing contactor as part of a commercial restoration project. A two part polyurethane was cho­sen due to the low humidity in El Paso and the extended cure times required for one part polyurethanes under these conditions. The specific product selected was not the contractor’s first choice-product selection was dictated by availabil­ity at the local distributor. However, the product selected is commonly available and widely used. The contractor has had extensive experience with this product

and has found that it crazes more than other commercially available polyure­thanes upon weathering. However, in his experience, this crazing has not led to complaints or call backs.

A laboratory prepared acrylic sealant, based on a commercially available binder, was chosen for comparison. A laboratory prepared sealant (Table 1) was used instead of a commercially available sealant so that the authors could con­trol the formulation and understand the relationship between formulation in­gredients and exterior performance. A plasticizer free formulation was chosen to minimize sealant dirt pickup. Since contractor application preferences were unknown at the time that the sealant was formulated, no attempt was made to optimize the viscosity, toolability, or open time of the formulated material that was sent to El Paso.

ABSTRACT: To demonstrate the suitability of high performance acrylic seal­ants to low rise industrial construction applications, a laboratory prepared high performance acrylic sealant was compared to a commonly used, com­mercially available two part polyurethane sealant. The centerpiece of this comparison is an exterior exposure in El Paso, TX, in which the two sealants were professionally installed in alternating joints around the perimeter of a tilt-up warehouse. The sealants were also subjected to a battery of labora­tory tests, including tensile testing, sealant specification testing, paintability, and accelerated weathering in both xenon arc and fluorescent UV devices. The 3 year El Paso exposure results, in combination with the laboratory, weathering, and application test results, demonstrate the performance ad­vantages of the high performance acrylic sealant and highlight its inherent suitability for use in low rise industrial applications such as tilt-up ware­houses.

KEYWORDS: acrylic sealant, polyurethane sealant, tilt-up, durability, crazing, accelerated weathering

Introduction

Perceptions of acrylic sealants in the construction industry have largely been formed by contractors’ experience with low-to-mid-performance formulations. Nonspec and ASTM C834-05 [1] compliant formulations were the first acrylic sealants to be introduced into the market and contractors collectively purchase huge volumes of these products every year for use in a variety of applications with minimal movement requirements. However, despite contractors’ lack of familiarity, high performance acrylic sealants with ±25 % joint movement ca­pability are widely available. These products fully comply with ASTM C920-05 [2], have the excellent weathering characteristics of acrylic chemistry, and have the added benefit of soap and water cleanup.

High performance polyurethane sealants are commonly used to seal exte­rior joints in low rise industrial buildings constructed of block, brick, and tilt-up panels. Although these sealants have been used for many years, they are not without problems. Commonly encountered issues include surface crazing and chalking caused by UV degradation, sealant burn-though in thin cross sec­tions, variable coatability, dirt pickup on coated joints due to plasticizer migra­tion, and the need for solvent cleanup.

ASTM C834-05 compliant acrylic sealants are frequently used to seal inte­rior joints in low rise tilt-up buildings. Although high performance acrylic seal­ants are readily available, they are rarely used to seal the exterior joints of these buildings. Despite the limitations listed above, contractors continue to use polyurethane sealants instead of high performance acrylic sealants. One of the reasons for this is that there is a general lack of knowledge about high perfor­mance acrylic sealants and how they compare to sealants based on alternative chemistries. Another is the lack of performance history of high performance acrylic sealants in commercial construction applications. To begin to fill these voids, a comprehensive study was undertaken to compare the performance of a high performance acrylic sealant to a commonly used two part polyurethane sealant.

A comparison was made between water leakage through deficient vertical and horizontal joints as shown in Fig. 19. The results reflect leakage rates of joints subjected to a water deposition rate of 4 L/(min-m2). The Y-axis provides the rates of leakage (L/min) across the horizontal joint; that of the X-axis for the vertical joint.

Results have been organized in terms of different crack lengths; cracks of 16 mm length are shown as circular data points, 8 mm as square points, and 4 mm as triangular points. The dotted lines delineate the outer boundary of the data and the oblique line joining points 0.0001 and 1 L/min on the plot indi­cates when the values of horizontal and vertical leakage rates are equal. A point falling beneath this line indicates that the leakage rate through the defect at the vertical joint is greater than the rate through the defect in the horizontal joint at the given test condition.

It is apparent from this plot that there can be substantial increases in leak­age rate of either vertical or horizontal joints and up to an order of magnitude difference.

The following was also evident:

• Overall, it is more likely that vertical joints will leak at higher rates than horizontal joints (ca. 59 %); as well, this was most prevalent at reduced water leakage rates (i. e., <0.005 L/min) where 81 % of the data points were those of the vertical joint having a greater leakage rate than that of the horizontal joint; on the other hand,

• At large crack openings (i. e., crack lengths of 8- and 16 mm, displace­ment of 1 and 2 mm) there is a greater chance (ca. 75%) that the rate of water leakage at the horizontal joint will be more severe than that of the vertical joint;

• Clearly the rate of water leakage depends on the nature of the crack opening (i. e., crack length and width); horizontal joints appear to be more susceptible to water leakage for joints having larger defect sizes.

Conclusions

1. For vertical joints evaluated in this study:

• There exists a linear relationship between crack width and joint dis-

Water leakage of vertical joint (L/min)

placement for cracks introduced in a sealant at the sealant-substrate interface; as well,

• Larger crack lengths induce greater crack widths and crack sizes in extended joints;

• The size and shape of the backer rod affects the nature of water leak­age across the vertical joint.

2. For both vertical and horizontal joints evaluated in this study:

• The crack length and joint displacement provide a multiplicative ef­fect on water leakage rates;

• If a crack exists in a sealed jointing system, even if the joint displace­ment is 0 mm, water may penetrate the opening at the crack;

• The higher the quantity of water deposition on or air pressure differ-

ential across the specimen, the greater the rate of water leakage of the jointing system.

3. Additionally, it may be suggested that if the crack length in a joint of an actual building is known or verified from a field inspection, an estimate of the rate of water leakage can be calculated by using the information given above and provided information is also given on the expected climate loads impinging on the fayade.

It should be borne in mind that estimates provided in this initial series of tests only offer a gross approximation of leakage across a deficient joint and are based on the limited number of tests and test variations. The movement of water through small openings will be affected by the tortuosity of the leakage path and the nature of the materials along which it flows. Hence, other factors such as the type of sealant, backer rod, and substrate material to which the jointing product is adhered may affect water entry. For example, a deficient joint of sealant installed on a concrete substrate is not likely to comport itself in exactly the same manner as suggested by results on the leakage through cracks reported in this study when considering the idealized test conditions. Nonethe­less, these studies offer some initial measure of the degree of water penetration at deficient joints—additional studies using the same approach would help elu­cidate the likely variations in leakage rate across a deficient joint that would arise given for example, different sealant and substrate materials or crack lo­cation and crack size.

Acknowledgments

That portion of this work on vertical joints was conduced over the course of a ten month visiting researcher work term at the Institute for Research in Con­struction (IRC), National Research Council Canada, in Ottawa. The authors are indebted to the Tokyo Institute of Technology, Japan, for having provided fund­ing to Dr. Miyauchi for his stay at the IRC, and to the IRC for their support to the research conducted by Dr. Miyauchi and Dr. Lacasse.

Evidence of water penetration at the horizontal joint is given in Fig. 15 and Fig. 16. A view of water entry along the interior side of the joint is shown in the photo of Fig. 14; water is seen to be pooling on the surface of the interior of the joint but ultimately made its way to the drainage opening. A photo (Fig. 16) at the underside of the joint at the crack location shows the path for water leakage through a crack opening of length 16 mm.

A summary of the results from water penetration tests on the horizontal joint is provided in Fig. 17. In this summary, the degree of water penetration is given in terms of water leakage (L/min) as a function of pressure difference across the test specimen (Pa) for ten test conditions. Results for water leakage for crack lengths of 4, 8, and 16 mm are given at joint displacements varying from 0 to 2 mm (10 % joint width). The range of scale for water leakage rate varies by three orders of magnitude, from a low of ca. 0.0018 L / min used for assessing water leakage across joints with no displacement, to a high of 0.2 L/min for joints having displacements of 2 mm.

Some key observations from water penetration tests on deficient horizontal joints are:

• Water leakage occurs when joints are "closed" (i. e., A = 0); even under low pressure differentials;

• Water leakage is pressure dependent; higher rates of leakage are ob­tained at higher pressure differences;

• A heightened degree of leakage can occur, up to ca. 1.6 L over a 10 min

500 1000 1500 2000

FIG. 17—Water penetration test results for horizontal joint. Variation in water leakage rates (L/min) in relation to pressure different across specimen (Pa) for joints having crack length deficiencies of 4, 8, and 1 mm and joint displacements (A) of 0, 2.5, 5, and 10 % joint width.

interval; this was estimated from the maximum leakage rates of greater than 0.16 L/min obtained for a crack length of 16 mm and 10 % joint opening at 1 kPa and 2 kPa driving pressures.

• Water leakage rates for a crack length of 16 mm are dependent on the crack opening size.

The final observation is more clearly evident from information provided in Fig. 18; the variation in water leakage as a function of pressure difference across specimen and joint displacements of 0, 2.5, 5, and 10 % are given for a joint having a crack length deficiency of 16 mm. The adjoining Fig. 18(b) pro­vides results for no displacement given that these are not readily apparent from that provided in Fig. 18(a). It is evident that as the crack length increases there is a corresponding increase in the rate of water leakage at the opening. For example, at the 1 kPa pressure difference, there is a ca. fifty-fold increase in water leakage rate between a joint displacement of 2.5 % (0.5 mm) and a closed joint (no displacement), and five-fold increase in leakage rate, for increases in joint displacement from 2.5 % to 5 % and from 5 % to 10 %, respectively.

Additionally it can been seen that for the smaller crack opening sizes (i. e., Д = 0, 2.5, 5 %; 0, 0.5 mm, 1 mm), rates of water entry increase with corre­sponding increases in pressure difference across the specimen; this suggests that the openings are completely occluded with water and the air pressure is driving water through these openings in increasing amounts and in proportion to the pressure difference. Whereas at the largest crack opening (Д =10%; 2 mm), the leakage rate reaches a maximum at 1 kPa pressure difference (0.162 L/min) and at 2 kPa there is only a small increase in leakage rate as compared to that obtained at 1 kPa (<2% to 0.165 L/min)). This suggests that at 1 kPa pressure level, the maximum leakage rate has been reached for the given water deposition rate and crack opening size; in this instance, the open­ing is no longer completely occluded with water hence air pressure cannot drive additional water through the opening and no additional rate of entry is possible at these test conditions. Such findings mirror those found for the ver­tical joint.

At water deposition rates at which the comparatively smaller openings are occluded, the larger openings are less readily filled but nonetheless this may occur intermittently given the erratic nature of water migration over openings. For larger openings, there are likely instances in which these openings will intermittently fill with water and thereafter, these water “plugs” would be ejected by the pressure differential across the opening.