September/October  2003

Independent’s Day
    an iga viewpoint

tim@iga.org

How Safe is Windshield Repair?

by Tim Smale

I am asked often about the safety of windshield repair. Through the years, as I have searched for information, I have contacted many sources, including the National Windshield Repair Association, and have found that while there are anecdotal references to the strength of repairs and proprietary studies by companies in the repair business, but little exists in the form of public studies. 

I also have found standards, such as the ANSI/SAE standard Z26.1, which addresses Safety Glazing Materials for Glazing Motor Vehicles and Motor Vehicle Equipment Operating on Land Highways. Among the nine durability and performance requirements for windshields stated in this standard, “Impact Test 12” is the most relevant when it comes to structural integrity of a windshield, whether it has been repaired or not. This test requires that the windshield be struck by a ˝-lb. steel sphere dropped from a height of 30 feet. Though small fragments of glass may leave the glass, the surface should be well-covered with tiny particles of tightly adhering glass. 

Other requirements of the Z26.1 specification are a second impact test, two tests for light stability, two tests for durability and tests for optics, penetration resistance and abrasion resistance. But these are standards and test guidelines; not test results themselves. (If studies about this topic do exist, I hope that a reader is able to share them with me.)

Motivated by curiosity, the need for a straight answer for our members and the absence of publicly available test results, the Independent Glass Association (IGA) recently commissioned a study of moisture content and adhesion levels of a dozen repaired windshields. 

It is important for anyone studying this issue to remain completely unbiased, and the IGA is certainly unbiased regarding windshield repair. Also, any study has its faults, and ours has two primary ones; a company that has some stake in the matter did it, and the sample size was quite small and certainly not valid statistically. But, the results may surprise you and will certainly raise the question—should further studies be performed?

From the Beginning
Windshields are assembled in clean rooms controlled to tight tolerances of moisture and relative humidity. The moisture content of the PVB interlayer has a large effect on the adhesion between the glass and the PVB. 

According to Mark Gold, applications development manager at St. Louis-based Solutia Inc., the PVB manufacturer which conducted the study, when a windshield breaks, initial absorption of moisture from the air into the PVB interlayer occurs within minutes of exposure. 

What follows are the actual procedures and results as reported to the IGA from Solutia. A dozen windshields of various makes and models that had been removed from vehicles were brought to Solutia by the IGA. 

Testing Procedure
Although some of these windshields were cracked in advance of this test work, the windshields were brought into Solutia’s test laboratory, where an undamaged portion of each was cracked using a standard glass cutting instrument to score a 2-inch long line on the outboard side. This score line was converted to a break through the outer ply of glass by tapping the glass. Several of the windshields arrived with bullseyes in them and these breaks were not altered in any way. The cracked portions of the windshields were left exposed to ambient air that was neither dried nor humidified for 14 days in Springfield, Mass., in April.

An outside, qualified windshield repair technician then was contracted to repair these windshields, selecting from the 12 broken windshields seven windshields that he felt could be repaired. These windshield repairs were conducted at the Solutia facility. Because the windshields were unsupported by the frame of the vehicle, only those windshields that had bullseyes were able to be repaired (otherwise, the scores would run when the necessary pressure to do a repair was applied to the unsupported 
windshield). 

Following the completion of the repaired point defects, the windshields were allowed to cure for 72 hours and then tested using standard laboratory test methods. The windshields were cut into approximately 1-square-foot sections centered at the point of the repair. They then were tested for moisture content of the interlayer at the point within 1/8 inch of the repair as well as at a point 6 inches away from the repair. Moisture measurements were done using a Perkin-Elmer Lambda 900 spectrophotometer equipped with a near-infrared light source that measured the absorption of energy by the water molecules in the PVB interlayer.

Following this test, each windshield was tested for glass adhesion to the interlayer within an approximate 40-square-inch area around the point of the repair. Adhesion testing was done using the laminating industry standard method of pummel adhesion. The pummel adhesion test consists of repeatedly impacting a piece of laminated glass and then comparing the amount of glass that remains adhered to the interlayer to a set of standards. Pummel adhesion is rated on a scale of 0 through 9, with 0 indicating a total loss of adhesion between the glass and the interlayer and 9 indicating a complete bond between the glass and interlayer. OEM manufacturers strive to achieve pummel adhesion levels in the range of 3 to 6 in order to assure both windshield durability and conformance to the Z26.1 code requirements.

Test Results
Moisture Levels: Testing of the moisture content of the repaired section of the windshields show that the moisture content increased in six of the seven tested specimens. As shown in Table 1, this moisture increase ranged from 5 percent to 95 percent of the original moisture content. 

Adhesion: Adhesion levels are stated in pummel adhesion units (PU), a 0-9 scale that indicates the relative amount of glass that remains on the interlayer. Adhesion levels below 2 are generally considered unacceptable for commercial vehicle windshields. Such adhesion levels, especially between the interior glass and the interlayer, would be insufficient to protect the occupants from injuries caused by glass that would become dislodged from the area on and near the repair, should the windshields be struck in the area of this low adhesion. The adhesion of glass to the PVB interlayer declined in all repaired samples in the area of the repair. This adhesion decline was found on both the impacted (and repaired) outside of the windshield as well as the unimpacted inside piece of glass. Table 2 summarizes the adhesion of these repaired windshields in the area of the repair.

Understanding the Results
Solutia summarized its study as follows: moisture infusion into the PVB interlayer of a broken windshield was shown to reduce the adhesion between the glass and the PVB in that region, thereby adding risk of injury from glass spalling (pieces of glass flying off the windshield) during subsequent impacts. Of the several broken windshields that were repaired, windshields number 3, 5 and 7 showed no remaining adhesion between the PVB and the glass in the areas at and surrounding the repair. These low-adhesion results are consistent with the higher levels of moisture content that were found. In one case, (windshield #2) there was minimal reduction in adhesion relative to the unrepaired section of the windshield, which was consistent with the lack of moisture found in that particular sample. This could be the result of the crack not penetrating the glass. 
Conclusions
The above test procedures and results were the findings of Solutia and warrant further comment. I think it would be interesting to perform similar tests using windshields that have undergone crack, rather than point, repairs. Such repairs would probably require that the repairs take place in the vehicle, though I would defer to experts, such as Ultra-Bond, to offer advice regarding long crack repairs and testing. Also, further studies would be interesting to actually measure moisture exposure and time prior to repair, as both variables affect the repair.

Everyone in the auto glass industry has a duty to protect consumers, and I know that we all take our roles very seriously. Though this study is by no means conclusive, my sincere hope is that this preliminary work by the IGA will result in more tests being made public that have already been performed, and any new studies will be similarly independent and unbiased. Our ultimate objective is to ensure that the public is safe above all else, and we as an industry have the facts to back it up.

Tim Smale is the chief executive officer of the Independent Glass Association, based in Idyllwild, Calif.

AGRR

© Copyright Key Communications Inc. All rights reserved. No reproduction of any type without expressed written permission.