Challenging Glass 6 - Conference on Architectural and Structural Applications of Glass

6 Louter, Bos, Belis, Veer, Nijsse (Eds.), Delft University of Technology, May 2018.
Copyright © with the authors. All rights reserved.
ISBN 978-94-6366-044-0, https://doi.org/10.7480/cgc.6.2132

Silicone Spandrel Glass Coatings:

Mitigating Glass Breakage Risk from
Thermal and Other Stresses
Chris Fronsoe a, Timothy Krytenberg a, Kris Vockler a, John Swanson a, Lawrence Carbary b, Chris Barry c,
Stephane Hoffman d, George Torok d, Scott Norville e, Samir Blanchet e
a
ICD High Performance Coatings, USA, [email protected]
b
Dow Corning Corporation, USA
c
Chris Barry Glass Consultant, USA
d
Morrison Hershfield, Canada
e
Texas Tech University, USA

Curtain wall design commonly uses insulating glass units for vision and spandrel glazing to provide better visual
harmonization of building façade glass. Risks with this design approach include higher thermal stresses, especially when
low-emissivity coatings are used on insulating glass units in spandrel areas. Ceramic enamel frit – commonly used to
opacify spandrel glass – is known to induce a bending strength reduction of up to 50%. The ability of ceramic enamel
frit coated glass to resist thermal stress is similarly reduced. Multiple incidences of thermal stress related fracture have
occurred with heat-strengthened, ceramic enamel frit opacified spandrel glass. An increased chance of spontaneous
breakage, by nickel sulfide inclusions, may occur if ceramic enamel frit opacified spandrel glass is fully-tempered to
withstand the thermal stresses that it is exposed to. Silicone spandrel glass coatings have been examined as a solution to
prevent the strength reduction in heat-treated glass when ceramic enamel frit is applied as an opacifier. Four-point
bending tests were used to investigate the flexural strength of coated heat-strengthened and fully-tempered glass. Ball
drop testing was used to investigate the impact resistance of coated fully-tempered glass. Test data shows silicone
coatings have no adverse effect on the flexural strength or impact resistance of the substrate and, in some instances,
improve it. Therefore, silicone opacifiers do not reduce the resistance to thermal stress of heat-treated glass. These
coatings also provide fallout protection in accordance with ASTM C1048 (ASTM, 2012). This suggests using a silicone
opacifier on heat-treated spandrel glass could greatly reduce the risk of fracture resulting from thermally induced tensile
stress, flexural stress, and impact related glass breakage in addition to reducing risk of injury from fallout if breakage
occurs.

Keywords: Glass Strength, Silicone Coatings, Insulating Glass Units, Glass Weakness, Ceramic Enamel Frit,
Architectural Glass, Uniform Lateral Load, Four-point Bending, Ball Drop

1. Introduction
In the world of façade glass, there are two types of glass: vision, and spandrel. Vision glass is transparent to provide
viewing areas for occupants and daylight for the interior. Spandrel glass areas, where slab ends, vents, and mechanical
elements reside, is opaque. Since the 1990’s, the amount of spandrel glass on a building has fluctuated between 15-
30% of a building's total glass area, as dictated by design trends. Recent energy codes, where the baseline building is
governed by lower glazing ratios, has resulted in increased spandrel areas of glazed facades. As a sizable area of a
building’s exterior, spandrel glass not only contributes to visual appeal, but a building’s energy use since insulating
glass (IG) units improve the energy performance of the spandrel area.

Insulating glass units have been used in commercial and residential high-rise construction for many years to increase
HVAC energy efficiency, and more recently, in the case of spandrel areas, to provide greater color and visual
harmonization between vision and spandrel areas. As vision glass improves – greater light transmission with lower
reflectance, plus increased energy performance – with the use of low emissivity (low-e) coatings it can become
difficult to visually harmonize façade glass unless the spandrel areas use the same glass configuration.

However, there is a risk associated with using IG units in a closed spandrel area. Most importantly, known to the
industry, is thermal stress will increase, particularly in the opacified inboard glass lite. This is primarily due to heat
gain that accumulates in the center of the inboard lite as the sun’s rays strike it. The increased heat differential between
the inboard glass center and its edges, often exacerbated with the use of low-e coatings or by dark colors in a spandrel
cavity (Mognato and Barbieri, 2013), leads to thermal stresses which can result in fracture. The accepted rule is that
0.34 MPa (50 psi) hoop stress is created around the edge of a lite of glass for every 0.56°C (1°F) increase in
temperature of the exposed area. Therefore, a center-to-edge temperature difference (ΔT) of 56°C (100°F) will create
34.5 MPa (5,000 psi) thermal stress.
 6
Challenging Glass 6
Issues from thermal stress are further aggravated by the use of ceramic enamel frit as a spandrel opacifier. In recent
years, there has been increased awareness in the industry that ceramic enamel frit, used as a spandrel coating, lowers
the flexural strength of both heat-strengthened (HS) and fully-tempered (FT) glass. The degree of weakening depends
on glass coverage, the colors used, and even the formulation of the ceramic enamel frit itself (Maniatis and Elstner,
2016). There are several newly published works that discuss the weakness of various heat-treated (HT) products with
an applied ceramic enamel frit opacifier; “full coverage black ceramic enamel…reduced the load resistance (LR) of
FT glass and HS glass by approximately factors of 2.0” (Natividad et al., in press) and 37.5% reduction of strength
(Krampe, 2014).

In summary: laboratory tests have shown significant flexural strength reductions in both new and artificially weathered
glass, in both HS and FT glass, when fully covered with a ceramic enamel frit opacifier. Strength reductions, of
approximately 50% to 20% have been measured in mean strength and in the 8 per 1000 probability of breakage
strength, respectively. Bergers, et al. (2016) attributed the difference between mean strength values and design
strength values to the fact that samples with ceramic enamel frit opacifiers have a much lower coefficient of variation
(CoV) than clear samples. A lower CoV signifies the difference between values for design load resistance is less than
the differences between mean load resistance.

European product standards EN 12150-1 (2015) & EN 1863-2 (2004) address reductions in flexural strength of HT –
HT referring to either HS and/or FT – glasses resulting from the application of ceramic enamel frit. These standards
reduce the minimum surface flexural strengths of HT glass from 120 N/mm2 (17400 psi) to 75 N/mm2 (10900 psi)
and from 70 N/mm2 (10200 psi) to 45 N/mm2 (6500 psi) for FT and HS glass with ceramic enamel frit, respectively.

The American standard, ASTM E 1300 “Standard Practice for Determining Load Resistance of Glass in Buildings,”
uses conservative 2X and 4X multiplying factors for the strength of HS and FT glass, when compared to annealed
glass (ASTM, 2016). Since no field-breakage has been reported from flexural stress in enameled HT glass, there has
been little immediate incentive to change the published uniform load strength values in ASTM E 1300-16 (2016).
Currently, there is no standard available to determine the resistance of glass to thermal loads in insulating glass units.

However, there have been a significant number of thermal stress breakages in HS IG spandrel units with ceramic
enamel frit (Barry and Norville, 2015). In all reported cases, the fracture origin is typically located between 13 mm
(0.5”) to 25 mm (1”) in from the cut edge of the glass (Fig. 1). Significantly, the fracture origin has always occurred
on the glass surface in contact with the ceramic enamel frit and never on the uncoated glass surface (Fig. 2). This
clearly indicates a relatively large reduction of the tensile strength of HS, in-service, glass with an application of
ceramic enamel frit.

Figure 1. Ceramic enamel frit opacifiers on spandrel glass with the Figure 2. Ceramic enamel frit opacifier on spandrel glass with the
thermal stress fractures beginning away from the glass edge. thermal stress fracture located where frit material interfaces with
glass.

In response to the thermal stress failures in the field, FT glass lites have been used in spandrel areas to mitigate
thermally induced breakage. However, this mitigation strategy increases the risk of spontaneous breakage from nickel
sulfide inclusions in FT glass, unless the glass is heat soaked, with a consequent cost increase. Also, if a FT inboard
 Silicone Spandrel Glass Coatings: Mitigating Glass Breakage Risk from Thermal and Other Stresses
lite of an IG unit breaks and goes unnoticed, the LR of the IG unit may decrease to where additional fracture may
occur allowing glass shards to fall from the spandrel cavity.

As such, there is an increasing need for a spandrel opacifier that does not weaken the glass, but ideally, increases its
strength and offers fallout protection. Silicone coatings were examined as a solution to the strength reduction issue
created by applying ceramic enamel frit to HT glass as a spandrel opacifier. Investigations were performed using four-
point bending and ball drop test methods.

2. Experiment Methods

2.1 Four-Point Bending Test

Six samples of HS and six samples of FT 102 x 305 x 6 mm (4 x 12 x ¼ in.) flat glass beams were obtained, with each
sample comprised of at least 30 specimens (Table 1).

Table 1: Samples tested via four-point bending.

Sample Sample Size Glass Type Coating
1 34 HS Clear (Uncoated)
2 33 HS OPACI-COAT-300®
3 30 HS OPACI-COAT-500®
4 33 FT Clear (Uncoated)
5 33 FT OPACI-COAT-300®
6 30 FT OPACI-COAT-500®

Specimens were coated using OPACI-COAT-300®, a water borne silicone elastomer, or with OPACI-COAT-500®,
a 100% solids silicone elastomer. All coatings were applied to the air side of the specimen – the side most commonly
coated with opacifiers in spandrel glass fabrication. A total of 193 specimens were tested. All coatings were black.
OPACI-COAT-300® was applied via spray gun to 330 μm (13 mils) wet film thickness (WFT). OPACI-COAT-500®
was applied via roll coat to 150 μm (6 mils) WFT.

Specimens were tested in four-point bending using an MTS machine to provide load at a uniformly increasing rate as
per ASTM C 1161-13 (ASTM, 2013). The MTS machine was certified by the American Association for Laboratory
Accreditation (ASLA Cert. No. 11455.01) for the basis of the ISO/IEC 17025 international standards for calibration
laboratories. During loading, a data acquisition (DAQ) system captured the load-time history from inception of loading
to fracture for each specimen. Traditional beam theory was used to convert load-time histories to stress-time histories.
Next, using equation 1 (Beason, 1980), each stress-time history was converted to a three-second equivalent fracture
stress (σ3). Finally, each value of (σ3), was used to determine values for equivalent fracture load (P3) from traditional
beam theory.

1�
𝑡𝑡𝑓𝑓 16
∫0 (𝜎𝜎(𝑡𝑡)−𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅)16 𝑑𝑑𝑑𝑑
𝜎𝜎3 = � � + 𝑅𝑅𝐶𝐶𝑆𝑆𝑆𝑆 (1)
3 𝑠𝑠𝑠𝑠𝑠𝑠

Where,

• 𝜎𝜎3 denotes the 3-second equivalent fracture stress

• 𝑡𝑡𝑓𝑓 denotes the time of fracture
• 𝜎𝜎(𝑡𝑡) denotes stress at time, t
• RCSS denotes the minimum observed residual compressive surface stress in a sample

All specimens were installed in the four-point testing mechanism with the float glass air side (coated side) facing
down (Fig. 3). This orientation induces tensile stresses in the float glass air side and compressive stresses in the float
glass tin side during testing. Glass will break under tensile stress. This is true whether glass is stressed mechanically
or thermally. The loading supports spanned 254 mm (10 in.) where each glass beam specimen was placed center to
center (c-c) allowing 63.5 mm (2.5 in.) between each support and loading point. The four-point bending creates
uniform stresses between the inside loading supports’ 127mm (5 in.) span.
 6
Challenging Glass 6

Figure 3. Schematic diagram of four-point bending test.

Once installed each test consisted of loading a beam specimen at a rate of 2.54 mm per min (0.1 inches per min), an
equivalent loading rate of 445 N per min (100 lbs per min), while the DAQ system recorded the load and time histories
at a sampling rate of 10 Hz. Each test concluded with the fracture of the beam specimen, followed by inspection and
measurement.

Using the Maximum Likelihood Estimation (MLE), the σ3 values were used to fit a three parameter Weibull
distribution to obtain cumulative distribution functions (CDFs) for the P3 using equation 2:

𝑃𝑃𝑓𝑓 = 1 − exp(−𝐵𝐵) (2)

𝐵𝐵 = 𝑘𝑘𝑆𝑆𝑚𝑚
𝑆𝑆𝑚𝑚 = 𝜎𝜎3 𝑚𝑚 𝐴𝐴𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠

Where,
• 𝒎𝒎 and 𝒌𝒌 denote statistical parameters
• 𝑩𝑩 denotes a risk function,
• 𝑺𝑺𝒎𝒎 denotes an equivalent plate area that depends on the magnitude of the tensile stress and the Weibull
parameter m
• 𝑨𝑨𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔 denotes the surface area

The CDF’s allow direct comparison concerning the LR of clear (uncoated) and coated beam specimens of the same
glass type. In this work, LR for the glass beams is defined as the constant applied force with 3-second duration that
leads to a probability of breakage equal to or less than 8 per 1000. Work by others (Bergers, et al., 2016) has indicated
a reasonable correspondence between results from four-point bending tests and full-scale test of rectangular lites.

2.2 Ball Drop Test

Four samples of FT 305 x 305 x 3 mm (12 x 12 x 1/8 in.) flat glass beams were obtained, with each sample comprised
of at least 30 specimens (Table 2)

Table 2: Samples tested via ball drop.

Sample Sample Size Glass Type Coating
1 34 FT Clear (Uncoated)
2 34 FT OPACI-COAT-300®
3 31 FT OPACI-COAT-500®
4 33 FT Ceramic Enamel Frit

Specimens were coated using OPACI-COAT-300®, a water-borne silicone elastomer, OPACI-COAT-500®, a 100%
solids silicone elastomer, or with a ceramic enamel frit. All coatings were applied to the air side of the specimen. A
total of 132 specimens were tested. All coatings were black. OPACI-COAT-300® was applied via spray gun to 330
 Silicone Spandrel Glass Coatings: Mitigating Glass Breakage Risk from Thermal and Other Stresses
μm (13 mils) WFT. OPACI-COAT-500® was applied via roll coat to 150 μm (6 mils) WFT. Ceramic enamel frit was
applied via screen print to 38 μm (1.5 mils) WFT.

Specimens were tested in a ball drop impact test frame following the parameters from a GANA Specification (GANA
76-12-10a, 2008), using a 50 mm (2 in.) diameter steel ball with 535 g mass weighing approximately 5.25 N (1.18
lbs).

Samples were loaded into the test frame, float glass air side (coated side) facing down. Impact height was increased
until fracture occurred, at which point drop height was recorded and the specimen inspected.

3. Results and Discussion

3.1. Four-Point Bending Test Results

All specimen fracture origins lay between the load points on the beam specimens, that is, within the area of constant
bending moment and flexural stress. No fracture origin was located on the edge of a specimen.

Below is a summary of the statistical results for the samples. Researchers used the minimum RCSS for each sample
in calculations described (Table 3).

Table 3. Sample statistics displaying mean P3, standard deviation, coefficient of variation (CoV), and minimal residual compressive surface stress
of all samples
Sample Mean P3 (kN) Standard Deviation (kN) CoV (%) Minimum RCSS (MPa)
1 2.19 0.23 10.5 50.2
2 2.35 0.16 6.67 52.0
3 2.36 0.17 7.26 52.9
4 2.75 0.15 5.48 104
5 2.74 0.16 5.76 104
6 2.77 0.11 3.81 82.6

The CoVs for samples of HS glass beams coated with OPACI-COAT-300® and OPACI-COAT-500® were relatively
small. The small values of CoV led to large and similar values of the statistical parameter m = 17 and m = 18 for
Sample 2 and Sample 3, respectively, which were significantly larger than m = 10 for Sample 1. One can make a
similar observation concerning values of m and the CoVs for Samples 4, 5, and 6.

Figures 4 and 5 display the empirical values and CDFs for the P3 values for HS and FT samples respectively while
Figures 6 and 7 present the lower portion of the CDFs for HS and FT samples respectively. The solid blue curved
lines in each graph represents the strength expected from HT glass when bought from a typical glass fabricator. Orange
and red perforated, curved lines in each graph represents HT glass strength when coated with the OPACI-COAT-
300®, and OPACI-COAT-500® silicone opacifiers.
 6
Challenging Glass 6

Figure 4. Cumulative distribution function for the probability of failure for clear HS, OPACI-COAT-300®,
and OPACI-COAT-500® specimens.

Figure 5. Cumulative distribution function for the probability of failure for clear FT, OPACI-COAT-300®,
and OPACI-COAT-500® specimens.
Silicone Spandrel Glass Coatings: Mitigating Glass Breakage Risk from Thermal and Other Stresses

Figure 6. Fitted cumulative distribution function for the probabilities of failure less than or equal to 10 per 1000 for clear HS,
OPACI-COAT-300®, and OPACI-COAT-500® specimens.

Figure 7. Fitted cumulative distribution function for the probabilities of failure less than or equal to 10 per 1000 for clear FT,
OPACI-COAT-300®, and OPACI-COAT-500® specimens.
 6
Challenging Glass 6
The cumulative distribution functions in Figures 4 and 6 indicate that both OPACI-COAT-300® and OPACI-COAT-
500® increased the strength of HS glass a statistically significant amount. This is demonstrated in mean P3 values and
most prominently at a probability of breakage less than or equal to 8 per 1000. This data is summarized in Table 4.

In Table 4, Samples 2 and 3 are HS glass specimens coated with OPACI-COAT-300®, and OPACI-COAT-500®
respectively, while samples 5 and 6 are FT glass specimens coated with OPACI-COAT-300®, and OPACI-COAT-
500® respectively.

Table 4. Change in P3 and load resistance (LR) relative to the associated clear samples.
Sample Δ P3 (%) Δ Load Resistance (LR) (%)
2 7.28 28.50
3 7.66 30.90
5 -0.16 -1.20
6 0.75 8.36

Fully-tempered specimens coated with OPACI-COAT® silicone coatings however do not demonstrate a statistically
significant change in either P3 or load resistance relative to uncoated specimens

3.2. Ball Drop Test Results

Table 5 summarizes the statistical results for the samples while Figure 8 displays the mean fracture height of all
samples. Table 6 shows the statistical values associated with mean fracture heights relative to those of uncoated
Sample 1. Figure 9 displays the percent change in fracture height of the sample means relative to that of uncoated
Sample 1.

Table 5. Sample statistics displaying mean fracture height, standard deviation, and coefficient of variation (CoV) of all samples.
Sample Mean Fracture Height Standard Deviation CoV
(m) (ft) (m) (ft) (%)
1 1.42 4.65 0.36 1.17 25.3
2 2.13 7.00 0.72 2.36 33.7
3 2.06 6.76 0.84 2.74 40.6
4 0.44 1.43 0.05 0.17 11.8

2,50 60% 51% 45%

2.13 2.06
2,00 40%
Δ Fracture Height (%)
Break Height (m)

1.42 20%
1,50
0%
1,00 -20% % Change
0.44
0,50 -40%

0,00 -60%
1 -80% -69%

Uncoated OC300 OC500 Ceramic Frit OC300 OC500 Ceramic Frit

Figure 8. Mean fracture height of coated and uncoated specimens. Figure 9. Percent change in fracture height of coated specimens
relative to uncoated glass specimens.

Table 6. Sample statistics displaying change in mean fracture height relative to uncoated Sample 1, and associated P-value.
Sample Δ Height Δ Height Δ Height P-Value
(m) (ft) (%)
2 0.72 2.35 50.6 7.9(10-5)
3 0.64 2.11 45.5 22.2(10-3)
4 -0.98 -3.21 -69.2 6.1(10-14)
 Silicone Spandrel Glass Coatings: Mitigating Glass Breakage Risk from Thermal and Other Stresses

Specimens coated with OPACI-COAT-300® demonstrated a mean increase in fracture height of 0.72 m (2.35 ft.)
compared to uncoated glass, while specimens coated with OPACI-COAT-500® displayed a similar mean increase in
fracture height of 0.64 m (2.11 ft.). This corresponds to increases of 50.6% and 45.5% respectively. Specimens coated
with ceramic enamel frit demonstrated a reduction in break height of 0.98 m (3.2 ft.), or -69.2% compared to uncoated
glass.

4. Conclusion and Next Steps

Four-point bending test results indicate OPACI-COAT® silicone coatings have a positive impact on the flexural
strength of heat-strengthened glass, and have little to no impact on the strength of fully-tempered glass. Furthermore,
ball drop tests displayed a large increase in impact resistance for OPACI-COAT® coated fully-tempered glass. Since
four-point bending is generally considered a more precise flexural strength testing methodology those results should
be given greater weight.

It is clear silicone opacifiers, such as OPACI-COAT® silicone coatings, certainly do not decrease the load resistance
or thermal stress control capability of heat treated glass, unlike ceramic enamel frit. As such, silicone opacifiers should
be considered instead of ceramic enamel frit as a spandrel opacifier when glass breakage from thermal or bending
induced tensile stresses, or from impact loads, is a concern.

Further testing should be conducted to explore the applications where increased load resistance and/or impact
resistance of silicone coated heat-strengthened glass compared to uncoated glass may be important.

References
ASTM “Standard Specification for Heat Strengthened and Fully Tempered Flat Glass.” ASTM C 1048-12e1, West Conshohocken, PA, (2012).
ASTM “Standard Test Method for Flexural Strengths of Advanced Ceramics at Ambient Temperature.” ASTM C 1161-13, West Conshohocken,
PA, (2013).
ASTM “Standard Practice for Determining the Load Resistance of Glass in Buildings.” ASTM E 1300-16, West Conshohocken, PA, (2016).
Barry, Chris, Norville, H. Scott, “Unexpected Breakage in Ceramic Enameled (Frit) HS IG Spandrels.”
IGMA Winter Conference, FL, (Feb 5, 2015).
Beason, William L., “A Failure Prediction Model for Window Glass.” Institute for Disaster Research, Texas Tech University, Lubbock, (NTIS
Accession No. PB81-148421), (May, 1980).
Bergers, Matt, Natividad, Kayla, Morse, Stephen M., Norville, H. Scott, “Full Scale Tests of Heat Strengthened Glass with Ceramic Frit.” Glass
Structures & Engineering; Springer: Switzerland, Vol. 1, Issue 1, (2016): p. 261.
EN 1863-2 “Glass in Building – Heat Strengthened Soda Lime Silicate Glass – Part 2: Evaluation of Conformity/Product Standard.” BSI, Brussels,
(2004).
EN 12150-1 “Glass in Building – Thermally Toughened Soda Lime Silicate Safety Glass.” BSI, Brussels, (2015).
GANA 76-12-10a “Fully Tempered Glass Uses Requiring Strength and Resistance to Temperature – Rev #1.” Topeka, KS, (2008).
Krampe, Philipp. “The Strength of Enamelled Glass.” In Challenging Glass 4 & COST Action TU0905 Final Conference; Louter, C., Ed.; Taylor
& Francis: London, (2014): p. 691.
Maniatis, Iris, and Elstner, Michael. “Investigation on the Mechanical Strength of Enamelled Glass.” Glass Structures & Engineering; Springer:
Switzerland, Vol. 1, Issue 1, (2016): p. 277.
Mognato, Ennio, and Barbieri, Alessandra. “The Breakage of Glass – Thermal Shock and Nickel Sulfide Inclusion.” COST Action TU0905, Mid-
term Conference on Structural Glass; Mocibob, D. Louter, C., Ed.; CRC Press: Boca Raton, (2013): p. 155.
Natividad, Kayla, Morse, Stephen M., and Norville, H. Scott. “Tests of Heat Treated Glass with Full Coverage Ceramic Frit.” J. of Architectural
Engineering, ASCE, (In Press).
6
Challenging Glass 6