EVA Interlayers in Hardened Facades

Introduction

Explosions, whether accidental or intentional, can result in very high loads of extremely short duration. Protective design against blast loading requires hardening of building systems, including glazing. As explosions are generally considered to be an extreme load condition, glazing is typically designed to accommodate some permanent damage. The focus of design is generally to control costs while protecting building occupants and equipment.

It is common in protective design practice to utilize laminated glazing panels to mitigate the threat posed by hazardous glass shards resulting from an explosion. Polyvinyl butyral (PVB) is the most common film interlayer utilized in glazing panels designed to have blast resistance. As the demand for solar energy has increased over the past decade, the interlayer used in photovoltaic cells (i.e., ethylene-vinyl acetate (EVA)) has become widely available and is potentially an economically viable alternative to PVB, particularly where the edges of the laminated glazing panel are exposed to weather.

With the recently increased availability of EVA, it is desirable to understand the physical characteristics and behavior of EVA when exposed to blast loads. In addition, an understanding of the ability of widely available software to accurately, or at least conservatively, represent the performance of the material is critical for widespread implementation in a blast scenario. This paper summarizes a recently completed effort to assess the relative efficacy of EVA-laminated glazing panels to resist blast loads and the incorporation and comparison of that data into a common software tool utilized for designing glazing for blast loads.

Blast Load Analysis

Before the specific results of this study are discussed, a brief primer regarding blast loading will be beneficial. FEMA 427 [1] describes an explosion as “a rapid release of energy in the form of light, heat, sound and a shock wave”. This rapid release of energy generates a shock wave with extremely high pressures traveling at supersonic speeds. If an explosive is detonated at ground level, this wave will expand hemispherically from the point of origin. As the wave expands radially, the magnitude of the blast pressure decreases relative to the cube of the standoff distance. The two main factors contributing to the magnitude of a blast load upon reaching an object are the size of the charge being detonated and the distance of the object from the detonation.

Blast loads exhibit an extremely high pressure but are of a very short (milliseconds) duration. This quick but large load imparts an impulse to the structure which must then be resolved by the force resisting systems within the structure. In reality, most of the structural reaction to a blast load takes place after the load has passed. This is very different from other dynamic loads (such as wind and seismic) which act over significantly longer periods of time and have cyclical properties.

If a pressure history reading of a blast wave is taken at a given point, it will consist of a rapid rise in pressure (so rapid that it is usually considered to be instantaneous) followed by a rapid decrease in pressure (lasting only milliseconds, also known as the positive-phase duration of the blast). This is followed by a longer, but lesser magnitude, negative phase duration. This pattern is illustrated in Figure 1. This wave will interact with objects that it passes, imparting loads. If the object happens to be a large, flat, surface, perpendicular to the direction of travel of the wave (such as a window), it will impart a head-on, reflected load onto that surface.

Figure 1. Typical blast load incident pressure and reflected pressure history (from FEMA 427 [1])

Glazing Design

Given the severity typically associated with blast loads, designing façade components to resist the full load while remaining elastic (undamaged) can be cost prohibitive. Therefore, facades are typically designed to allow for some level of permanent damage, while protecting building occupants. In the case of glazing, the goal of protective design is to allow permanent damage while retaining fragments so the glazing does not become projectiles flying into occupied space.

When exposed to blast loads, annealed and heat strengthened glass will tend to break into large fragments with sharp edges. Tempered glass is stronger than annealed or heat strengthened glass and will fragment into much smaller pieces. In both scenarios, without some means of containment, these fragments will be projected into the occupied space at high speeds. At these velocities, any fragment will be hazardous to the occupants of the space. To prevent the creation of glass projectiles, it is common to utilize a high strength adhesive interlayer within laminated glazing panels.

Glazing Interlayers

The most common interlayer material used for blast resistant glazing is polyvinyl butyral (PVB). PVB is a clear resin that provides optical clarity, with strong adhesion to glass surfaces and inherent toughness and flexibility. It is readily available and well established as a blast resistant interlayer. Through decades of testing and practical application, the material behavior under blast loads is well documented and understood. However, other materials can potentially be utilized for this application.

Ethylene-vinyl acetate (EVA) is a thermoplastic polymer that is commonly produced as a film known for its relatively low cost, high adhesion strength, and high transparency. One common application of EVA film is as the adhesive interlayer used in photovoltaic modules in solar panels. In recent years, as production has increased, there has been an interest in expanding the use of EVA film into architectural applications. One such potential application is as the adhesive interlayer for laminated glazing panels.

EVA interlayers have only recently been put forth as a potential alternative to PVB interlayers in blast-related glazing applications. Therefore, information for commercial-off-the-shelf EVA interlayers has not been readily available to blast protection professionals responsible for designing glazing. Information concerning the mechanical properties of EVA interlayers, both in new and weathered states, and the dependence of these properties on load duration is essential. In addition, it is critical to understand the ability of EVA interlayers to hold on to glazing fragments resulting from blast loads and their relative resistance to tearing.

Development Process

A four-phase effort as shown in Figure 2 was used to investigate EVA’s efficacy as an interlayer in blast-resistant laminated glazing.

Figure 2. Approach Used to Investigate EVA-Laminated Glazing Panels for Blast Loads.

Phase 1 - Literature Review

An investigation was performed into existing literature to identify commercial off-the-shelf EVA materials. While little information was available regarding the behavior of EVA under blast loads, information regarding EVA interlayers utilized for architectural applications was available. This included documentation of:

• Film properties, including the historical use of EVA films, the influence of EVA’s chemistry on its material properties, how EVA film is manufactured, publicly available stress-strain data for architectural grade EVA films, and material properties for Safety Glazing Certification Council (SGCC) certified EVA films.

• Laminated glazing properties, including how EVA-laminated glazing panels are made and verified for quality, the structural performance of EVA-laminated glazing panels under both quasi-static and dynamic loading, the adhesion of EVA interlayers to other substrates, and the impact of weathering on EVA-laminated glazing panels.

Research into EVA interlayers yielded the following information:

• While non-cross-linked EVA is a soft, translucent, easily plastically deformable thermoplastic, the process of cross-linking (via the lamination process) transforms the EVA material into a transparent elastomeric material with enhanced mechanical and thermal stability that makes it well suited for use as a PV module encapsulant or an interlayer in a laminated glazing panel.

• The vinyl acetate content (VAC) has a measurable impact on the material properties of EVA films. Within the range of VACs found in architectural grade EVA films (typically 25 to 35 percent of its weight), certain material properties such as tear resistance can vary widely (Figure 3). It is not clear from the literature surveyed if cross-linking serves to minimize this tear resistance variation.

Figure 3. VAC impact on the mechanical properties of EVA [2]

• The lamination protocol used in the manufacture of EVA-laminated glazing panels varies for each EVA film product, layup, lamination method (i.e., vacuum or autoclave), and even the lamination device. There is no single lamination protocol that can be used in all circumstances.

• UV radiation exposure is the primary weathering effect of concern for EVA-laminated glazing panels. High temperatures and high humidity can serve to accelerate weathering in combination with UV exposure.

• In the context of blast loaded laminated glazing panels, cross-linked EVA’s primary benefits over PVB include its excellent adhesion to materials other than glass, its relative resistance to moisture absorption and thus delamination, its ability to be used in exposed edge laminated glazing panels, its simpler storage requirements, its relative thermal stability, and its superior elongation at break. Areas where PVB excels when compared with EVA include its higher strength and stiffness, which allows for smaller interlayer thicknesses to achieve the same measure of blast resistance; cost; and familiarity within the laminated glazing industry (including with its lamination protocol).

Phase 2A - Dynamic Material Characterization Testing

The first step in characterization was to perform lab testing of the various crosslinked EVA and PVB materials utilized for the research. This included over 120 tensile tests on film samples at various strain rates, temperatures, and manufactured with different lamination processes. Strain rate is a measure of the material strain over the time that strain is applied. The higher the strain rate, the faster the material is elongating. Understanding the response of materials at high strain rates is critical when evaluating responses to blast loads.

The objectives of testing were: (1) to corroborate manufacturer-reported material property information, (2) to assess the response of EVA and PVB films across the higher strain rates likely to be encountered by laminated glazing exposed to blast loads, (3) to assess the impact of temperature on the material properties of EVA and PVB films at both quasi-static and dynamic load rates, and (4) to assess the impact of the lamination process on the material properties of EVA and PVB films at both quasi-static and dynamic load rates.

The films tested included a PVB film (i.e., Saflex® Clear by Eastman Chemical Company) and four EVA films (i.e., SE-381TF by SWM International, evguard® by Folienwerk Wolfen GmbH, EVALAM Visual by Hornos Industriales Pujol, and EVALAYER by Interlayer Solutions). Films were tested in pre- and post-laminated states. The lamination protocol used for the post-laminated specimens were different for each film tested and were based on manufacturer’s recommendations.

The following observations were gleaned from the dynamic test results:

• The PVB film was characterized by higher peak engineering stresses but lower peak elongation strains when compared with those of the EVA films. This result was consistent with the material property data reported by PVB and EVA film manufacturers.

• The initial stiffness of the PVB film was lower than that of the EVA films at strain rates below 2 s^-1 but was markedly higher than that of the EVA films at strain rates greater than or equal to 2 s^-1.

• The small strain response of the four EVA films tested was remarkably consistent. For large strains, the responses of the EVA films diverged but generally followed a consistent pattern. EVA films with higher levels of cross-linking tended to have smaller elongations but higher tensile strengths. Thus, the level of cross-linking (i.e., as measured through gel rate testing) appears to have a notable impact on the large strain response of EVA films.

Phase 2B - Shock Tube Testing

Upon completion of dynamic material characterization (DMC), one PVB and two EVA films were selected for further investigation via 46 shock tube tests on 48-inch by 48-inch laminated glazing panels [3]. A shock tube is a testing apparatus which can be used to generate a variety of pressure history profiles, which provide a means of approximating the behavior of an explosive detonation. A typical glazing panel within the testing frame is shown in Figure 4. The shock tube utilized for this testing is shown in Figure 5.

Figure 4. Typical glass panel and frame for shock tube test

Figure 5. Shock tube utilized for testing

The objectives of this testing were: (1) to compare the dynamic response of EVA and PVB-laminated glazing panels under identical pre-conditions, (2) to assess the impact of glazing type (i.e., annealed, thermally tempered, and heat strengthened) on the dynamic response of EVA-laminated glazing panels, (3) and to assess the impact of glazing panel boundary restraint (i.e., mechanically captured, mechanically captured wet glazed) on the dynamic response of EVA-laminated glazing panels. Film product, glazing type, target blast load, and boundary restraint were varied and two specimens were tested for each laminated glazing layup tested. The first test for each combination used the target blast load. The second test modified the target blast load upwards or downwards based on the results of the first test.

Phase 3A - Pummel Adhesion Testing

To better understand the baseline adhesion performance of EVA-laminated glazing panels as compared to PVB-laminated glazing panels, a series of pummel adhesion tests were performed on the laminated panes of insulated glazing units (IGUs) with EVA (i.e., SE-381TF and evguard®) and PVB (i.e., Saflex® Clear – RA 41) films [4].

Pummel adhesion testing is used to measure the bond strength between glass and an interlayer. This provides an indication of the material’s resistance to delamination and impact failure. Pummel adhesion testing was performed in accordance with the semi-automatic methodology in ASTM C1908-21 with the pummels run on laminates conditioned to 0 °C. The laminated panels that underwent pummel adhesion testing were two pieces of nominal 0.25-inch thick annealed glass with a 0.03-inch thick interlayer. The same lamination protocols used for the DMC testing was used to laminate these samples. As three duplicate tests were performed for each combination, a total of 18 pummel tests were performed.

Representative photographs of the post-test state of the specimens are included in Figure 6. As can be inferred visually from Figure 6, both EVAs exhibited pummel ratings of 9 or 10 whereas the PVB exhibited pummel ratings between 4 and 7. Even though the tests were performed in the blind, the personnel performing the tests were easily able to distinguish the EVA from the PVB samples as the EVA interlayers were soft and rubbery following laminated panel freezing whereas the PVB interlayer was rigid. Instead of the glass popping off upon pummel impact as would be typical for a PVB interlayer, the glass was driven into the EVA interlayer, which effectively further reinforced the interlayer’s adhesion and potentially increased its pummel rating.

(a) Saflex® Clear (PVB).	(b) SE-381TF (EVA).	(c) evguard® (EVA).

Figure 6. Representative photographs of specimens following pummel adhesion testing

Phase 3B - Arena Blast Testing

The shock tube testing indicated that EVA film is predisposed to tearing when used with annealed glass. Two arena blast tests were subsequently performed to investigate if EVA-laminated glazing panels could be reliably designed to both prevent interlayer tearing under a design blast load and facilitate an interlayer tearing failure (rather than a bite failure) if overloaded. Whereas the shock tube tests involved laminated panel test articles, the arena blast tests involved Insulated Glazing Units (IGUs). Film product, daylight opening dimensions, interlayer thickness, and blast load were varied. Further detail regarding film thicknesses and glazing layups are described in “Construction Specification Guidance for EVA-Laminated Glazing Used in Blast Protection Designs: Results from Blast Tests of EVA-Laminated Glazing Panels” by Weaver and colleagues [4].

For each test, the IGUs were designed to crack and exhibit various levels of interlayer tearing while remaining in their frames (although it was anticipated that panels with torn interlayers could fall to the floor on the protected side). The typical glazing panel daylight opening was 63 inches high by 39 inches wide. Two of these test articles were placed in the Type 1 wall front (i.e., Figure 7a). The only difference between these two test articles was that one had 50% more interlayer than the other. Additional glazing panels with different size daylight openings were included to investigate the impact of aspect ratio on EVA-laminated glazing panel response (i.e., Figure 7b).

(a) Type 1.	(b) Type 2.

Figure 7. Wall front types included in arena blast testing

Varying levels of displacement and interlayer tearing in the EVA-laminated glazing panels were observed in the two blast tests [4]. For both types of EVA interlayer tested in the Type 1 wall front during Test 1, one EVA-laminated glazing panel cracked but remained in the frame while the other EVA-laminated glazing panel exhibited significant interlayer tearing and pieces of the panel were propelled into the room (Figure 8). As the only difference between the two EVA panels tested in each reaction structure was thickness of the interlayer, this test demonstrated the interlayer thickness level at which EVA-laminated glazing panels become a hazard to building occupants. The remainder of the glazing panel test articles were retained by their respective frames, although interlayer tearing was observed in several test articles that remained in their respective frames.

(a) PVB.	(b) EVA (SE-381TF shown).

Figure 8. Representative post-test photographs of Type 1 wall front following Test 1

Phase 4 – Analysis and Development of Technical Guidance Documents

GSA Performance Condition

Prior to discussion of analysis results, an understanding of evaluation of design window performance under blast loading is needed. A common means of defining glazing performance is through use of the performance conditions published by the GSA. The GSA Performance Condition (PC) descriptions are defined in GSA-TS01-2003 [5]. These descriptions are listed in Table 1 and illustrated in Figure 9. In general, a performance condition of 3b or better is desired for design of blast resistant glazing.

Table 1. GSA Performance Condition Descriptions Used to Evaluate Test Article Response.

Figure 9. Window Performance Characteristics

Each scenario described for the shock tube testing of Phase 2 was also evaluated using the Window Glazing Analysis Response and Design (WINGARD [6]) analysis software to assess the efficacy of the software to safely design EVA-laminated glazing panels for blast loads. WINGARD is a design tool developed by the GSA to provide a means of modeling window response to the effects of an explosion. Using the film material properties developed through DMT, WINGARD analyses were performed with the average blast pressure histories measured during testing. The WINGARD analysis results were then compared with the test results. For this research, several metrics were used to evaluate the relative efficacy of the WINGARD analyses to safely design EVA-laminated glazing panels for blast loading.

Performance Condition

For the PVB-laminated glazing panel tests, the WINGARD analysis always predicted an equivalent or worse GSA PC than that observed during the test. For the EVA-laminated glazing panel tests, the WINGARD analysis predicted an equivalent or worse GSA PC than that observed during the test in all but one case. However, this case was based on a WINGARD analysis feature which the software developer indicates is preliminary, which could have skewed the analysis results.

Deflection

In general, the average peak deflection for both types of interlayers favored overprediction of the deflection, rather than underprediction of the deflection, which is generally preferred from a design conservatism perspective. Examining the peak deflection results is instructive for comparing the relative blast performance of EVA and PVB interlayers. In one case, a PVB panel and an EVA panel exposed to roughly the same blast load, had the same glass type and thickness, and exhibited roughly the same peak deflection. The only difference between the two panels was the EVA panel had four times the amount of interlayer included in the PVB panel. Where only two or three times the amount of EVA was provided (with everything else being equal), the EVA-laminated glazing panel always exhibited greater peak deflections than the PVB-laminated glazing panel.

Required Bite Depth

Based on the WINGARD analysis, thirty-nine percent of the EVA-laminated glazing panel designs had insufficient bite. However, this was not borne out by testing, in which only 1 of the 20 PVB-laminated glazing panel blast tests showed such a result. Thus, it can be stated that the required bite depth prediction ability in WINGARD is somewhat conservative.

In many of the blast tests performed on EVA-laminated glazing panels, particularly those with annealed glass, EVA interlayer tearing rather than bite pullout was the ultimate failure mode. This is despite EVA having roughly three times the elongation potential of PVB. While further investigation into this phenomenon is warranted, it is possible that EVA’s high adhesion could be causing extreme stress localization at glass cracks, which ultimately could be promoting premature tearing. It is interesting to note that when the cracks were not sharp or irregular (in the thermally tempered specimens), EVA interlayer tearing was not observed.

Conclusions

Following research of existing data and four testing efforts (i.e., DMC, pummel adhesion, shock tube, and arena blast), several high-level conclusions were derived:

• EVA-laminated glazing panels with annealed glass are prone to tearing under blast loading conditions. This is despite the fact that EVA’s elongation potential is roughly three times that of PVB. One possible reason for this tearing could be EVA’s characteristically high adhesion, although further investigation into the mechanisms causing and means to prevent this tearing is warranted. The tearing seemed to be less likely when EVA was combined with tempered glass.

• The blast tests performed indicate that roughly four times the amount of EVA interlayer thickness is needed for EVA-laminated glazing panels to exhibit the same blast load performance as PVB-laminated glazing panels.

• The EVAs tested exhibited similar qualitative and quantitative responses for the dynamic loads considered as part of DMC and blast testing performed as part of this effort. This finding supports the potential for EVA material interchangeability from a blast response perspective, although EVA materials should undergo DMC testing to verify their dynamic material properties prior to inclusion in blast-resistant windows.

• In general, the default EVA film input parameters for WINGARD appear to well approximate the responses of EVA-laminated glazing panels observed in the blast testing performed. However, further investigation is warranted to better understand, predict, and evaluate the hazard posed by EVA interlayer tearing.

• It should be noted that, in cases where the actual blast loads exceed the designed strength of a laminated window, the entire glazing panel could become detached from the frame and thrown into occupied space. This results in substantial hazard to occupants. In this condition, interlayer tearing offers a potentially preferable ultimate limit state versus bite pullout for overloaded panels if the secondary fragments developed on account of tearing are prevented from being thrown into the protected space. Further investigation into potential means to accomplish this objective via frame retention measures or otherwise is of interest.

The results of this research indicate that EVA is a viable alternative to PVB in the application of blast resistant laminated glazing. The physical characteristics and behavior of EVA when exposed to blast loads were found sufficient to provide desired performance with reasonable glazing layups, and currently available software is shown to conservatively represent the performance of the material.

Ongoing Research

As a follow-on effort to the work described herein, the U.S. Department of State is exposing a portion of the PVB and EVA-laminated glazing panel layups involved in blast testing to natural weathering (including UV exposure) according the protocol defined in ANSI Z97.1 over a period of ten years. The intent is to repeat the blast tests on these weathered panels at the conclusion of this 10-year period to assess the level of blast load resistance degradation on account of weathering.

[1] Federal Emergency Management Agency, Primer for Design of Commercial Buildings to Mitigate Terrorist Attacks, 2003.

[2] H. Domininghaus, Kunststoffe - Eigenschaften und Anwendungen (Plastics - Properties and Applications, 5th Edition), Berlin: Springer-Verlag, 2012.

[3] M. J. Lowak, "Shock Tube Testing of EVA Laminated Glass," Baker Engineering and Risk Consultants, Inc., San Antonio, TX, 29 November 2022.

[4] M. K. Weaver, E. Kjolsing, l. Abdul Hadi and C. M. Newberry, "Construction Specification Guidance for EVA-Laminated Glazing Used in Blast Protection Designs: Results from Blast Tests of EVA-Laminated Glazing Panels," Karagozian & Case, Inc., Glendale, CA, TR-22-26, 29 November 2022.

[5] GSA-TS01-2003, "Standard Test Method for Glazing and Window Systems Subject to Dynamic Overpressure Loadings," U.S. General Services Administration, 1 January 2003.

[6] WINGARD Professional Edition (PE), Version 6.1, Applied Research Associates, Inc., September 2018.

[7] "Window Glazing Analysis Response & Design (WINGARD) - Technical Manual," Applied Research Associates, Inc., Vicksburg, MS (FOR OFFICIAL USE ONLY), March 2020.

[8] ASTM F2912-17, "Standard Specification for Glazing and Glazing Systems Subject to Airblast Loadings," ASTM International, West Conshohocken, PA, 2017.