Mechanical Properties
Mechanical Properties
ABSTRACT
In the early 1960's the first high strength glass fibers, S-glass, were developed in joint work
between Owens Corning and the US Air Force. Today high strength glasses with various batch
formulations are manufactured commercially by several companies in every major economic
region. The utility of high strength glass fiber compositions are compared by physical,
mechanical, electrical, thermal, and optical properties. The influence of filament diameter, over
the range 9 - 26 microns, and size chemistry on composite mechanical performance is described,
in epoxy and vinyl/polyester resins. Data on high strength glass composites reinforced with UD
fiber and fabrics are reported from –55°C to 80°C.
In the early 1960s high strength glass fibers, S-glass, were first used in the joint work between
Owens Corning and the United States Air Force for use in strategic missiles and other military
end-uses. In the period 1966-1968 Owens Corning introduced high strength glass
reinforcements for commercial applications under the trade name S-2 glass®. The two fibers
share the same glass composition; the primary difference in the two product families was the
sizing. The original S-glass product range has now been entirely replaced by S-2 glass rovings.
S-2 glass fibers are a magnesium alumino-silicate produced for textile substrates or
reinforcements in composite applications. RH glass was introduced in 1968 by Vetrotex, and
remained in production until the early 1990’s. More recently yarns and assembled rovings
belonging to the S-glass family have been manufactured in Japan under the names T glass and U
*S-2 GLASS and ZenTron are registered trademarks of Advanced Glassfiber Yarns LLC.
*ADVANTEX® is a registered trademark of Owens Corning Inc.
glass. A comparison of the chemical composition and physical, mechanical and other properties
of S-glass, E-glass and various other commercially available glass fibers has been published
elsewhere[1]. S-2 glass fibers are manufactured commercially in slivers of 11 to 66 tex, with
filament diameters of 5 to 13 microns. S-2 glass rovings are assembled from these slivers, and
are available up to 2050 tex. In 1997 a new single ended roving process for high strength glass
was introduced commercially, and these S-glass rovings are sold under the trade name ZenTron.
Unlike the S-2 glass assembled rovings, the single ended rovings are made in a one step process,
which typically results in rovings with a linear density in the range of 660-2050 tex, and filament
diameters in the range of 13 to 25 microns. Whereas S-2 glass products are largely used in
defense and aerospace, the ZenTron products have found applications in the automotive (2) and
sporting goods industries.
2.1 Introduction The following glass types account for the vast majority of glass fibers
produced commercially for use as reinforcements:
Advantex®. Calcium aluminosilicate glass used as a boron free substitute for E-glass, providing
E-CR Glass acid resistance with the reinforcing characteristics of E-glass. Introduced in 1997.
AR Glass. Alkali resistant glasses composed of alkali zirconium silicates, developed for use in
use in cement and concrete. Mainly used for GRC (Glass reinforced Cement) roofing and
sheeting panels.
C glass. Calcium borosilicate glasses used for their exceptional stability in acidic environments.
The resistance to water is poor, and so C glass is rarely used as a reinforcement in North
America or Europe, but finds application in surface mats, glass flakes for coatings and acid-
resistant surface cloth.
D glass. Borosilicate glasses with low dielectric constant used for radomes and other specialty
applications requiring permeability to electromagnetic waves.
E glass. Alumina-calcium-borosilicate glasses with a maximum alkali content of 2% w/w used
as a general purpose reinforcement where strength and high electrical resistivity are required.
Together with Advantex, accounts for over 90% of all GRP reinforcements produced today.
Introduced in 1939.
E-CR Glass. Calcium aluminosilicate glasses with a maximum alkali content of 2% w/w used
where acid corrosion resistivity is required in addition to strength and electrical resistivity.
Introduced in 1980.
R-glass. Calcium aluminosilicate glasses with higher strength and temperature resistance than
E-CR Glass or Advantex.
S-glass. Magnesium aluminosilicate glasses used for textile substrates or reinforcement in
composite structural applications, which require high strength, modulus and durability under
conditions of extreme temperature of corrosive environments. S-2 glass, T glass, U glass and
ZenTron are examples of current commercial products. Introduced for commercial use in 1966.
TABLE 1-Composition ranges for glass fibers
Continuous glass fibers were first manufactured in commercial quantities by Owens Corning
Textile Products in the 1930’s for high temperature electrical applications. High strength glasses
differ from E-glass and Advantex products in several aspects, most importantly the higher
softening point temperature and the much more modest scale of production. As a consequence it
is necessary to use specially built ultra high temperature (UHT) furnaces for continuous
commercial production, as shown in figures 1 and 2.
Raw materials such as silicates, limestone or metallic oxides are finely ground and then blended
together in a bulk quantity, called the “batch”. More than half the mix is silica sand, the basic
building block of any glass fiber. The blended mix is then conveyed to a small furnace. The
temperature is sufficiently high for the sand and other ingredients to dissolve into molten glass.
The walls of the furnace are protected with tailor made liners in order to withstand the very high
temperatures necessary to allow the glass to flow and form fibers. The molten glass flows
through a platinum/rhodium bushing – a block of metal with precisely drilled holes to allow the
glass to run through, and then through individual bushing tips and orifices ranging from 0.75 to 2
mm diameter. High strength glasses are typically processed at about 120°C higher than E-glass
in the melter and hundreds of degrees higher at the bushing. The glass is then rapidly quenched
and attenuated (drawn down) in air and water to prevent crystallization, forming fine fibers
ranging from 5 to 26 µm in diameter. Experimental filaments over 35 µm have been produced in
Huntingdon, PA.
Mechanical winders pull the fibers at lineal velocities up to 60 meters/second over an applicator
which coats the fiber with an aqueous mixture of chemicals called a “sizing”. This protects the
filaments from abraiding each other during further handling, aids processing, and provides
compatibility and adhesion of the fiber with the matrix or resin.
The greater temperatures involved in producing high strength glass present formidable technical
challenges to commercial production, and few companies globally are able to make a
consistently high quality product. The platinum/rhodium bushings have a hard, short life. The
bushings need to be designed to allow a flat temperature profile across all orifices, such that all
filaments of glass have the same diameter. As a consequence, the process is intrinsically more
expensive than the larger scale lower temperature E-glass production from conventional ceramic
lined furnaces.
In most cases the strands, which are typically 5-13µm diameter filaments aligned in 11 to 80 tex
slivers, are wound onto a “forming cake”. The forming package is an intermediate product that
is cured in an oven to drive off the moisture in the size, and then typically processed in one of
two ways before being shipped to the customer. To produce a twisted yarn, the forming cake is
unwound, twisted and wound onto a bobbin. To produce a multi-end roving package, several
forming cakes are simultaneously unwound, then wound again onto a new package with typically
5 to 30 ends of glass strand. These “conventional” roving doffs or spools are external take off
packages similar to carbon fiber or aramid fiber spools, and are used for prepregging, filament
winding or weaving. The conventional roving/yarn process is illustrated in figure 1.
The ZenTron roving is unique in the world of high strength glass as the winder collects all the
strands into a single end roving package. This process is illustrated in figure 2. The roving has a
filament diameter in the range 13 – 26µm and a tex from 660 to 2400 g/km. This necessitates an
S -2 G lass ® M anu factu ring
facturing
M u lti-e n d R o v in g
U H T fu rn ac e
F o rm in g P a c k a g e s S q u a r e -e d g e
B a tc h P acka ge
H ouse B u s h in g
(n 00 filam e n ts ) D ry /c ure
4 8 h o u r s u rg e
order of magnitude increase in bushing size and design complexity, as thousands rather than
hundreds of holes are needed. The packages may be internal take off, which is typical of E-glass
rovings, or external take off, which is typical in the advanced composites
Z en T ro n ® R ov
o v ing
in g M an u factu rin g
U H T fu rn a c e
B a tc h
Ho use B u s h in g (n ,0 0 0 fila m e n ts )
S q u a re -E d g e d P a c k ag e
(w ith /w ith o u t tu b e )
1 3 µ -66 0 te x -> 2 6 µ-2 4 0 0 tex
D r y/c ure
P a c k a n d S h ip
.
Figure 2 Schematic of single end glass roving manufacturing
industry. The packages are dried in an oven and are then ready to be shipped to the customer
without further processing.
2.3 Properties of High Strength Glass Fiber.
These have been described in detail by Hartman, Greenwood and Miller(1). The S-glass family
is represented commercially by S-2 glass and ZenTron from the US, and T glass and U glass
from Japan. All three have similar compositions conforming to S glass in Table 1. U glass is
available only as a RFL coated cord for use in synchronous (timing) rubber belts.
R glass is produced in Europe, and is intermediate in performance between Advantex or EC-R
glass and the S glass family.
3.1 Introduction As far as composite performance is concerned, the fiber matrix interface is at
least as critical as the properties of the reinforcing fiber. An E glass fiber with the appropriate
adhesion and interface will outperform an S glass or carbon fiber with an unsuitable fiber-matrix
interface. This is especially true of thermosetting resin systems in the presence of water.
3.2 Current Industrial Practice A “size” or coating is applied to glass filaments during the
extrusion process and prior to the filaments being gathered into a strand and wound into a
package. The sizing provides a number of properties and functions to the glass. One important
function is to prevent defects and fracture of the filaments during formation. Bare glass is very
abrasive to itself; it takes a coating of size to be able to wind a continuous strand into a package
during forming. Size is typically an aqueous solution or emulsion that is applied by passing the
filaments over a rolling applicator. The applicator can be made from such materials as a rubber
belt or graphite cylinder. By dipping the applicator into a bath of sizing a thin layer is deposited
on the applicator surface and then onto each individual glass filament
This surface treatment also plays an important role in downstream functions of the glass fibers.
Textile sizes, typically made from starch/oil and polyvinyl alcohol, have low coefficients of
friction and provide good protection during subsequent processes of weaving, braiding, or
knitting. The weaver will then typically remove the size by heat cleaning and apply a finish
compatible with the matrix resin used in the final product.
Reinforcement sizes are used when the size is not thermally removed from the glass and provides
a good chemical bond between the glass and matrix resin. This chemical bonding is typically
provided by the addition of silanes. Organofunctional silanes contain a silanol group on one end
of the molecule which reacts with the glass surface and a variety of other functional groups on
the same molecule which are capable of reacting or interacting with the matrix resin. The only
other coupling agents that have found commercial application are methlyacrylate-chrome
complexes such as Volan A®, and although technically effective the current trend is to replace
these with organofunctional silanes due to environmental considerations.
Film formers and lubricants are also used in reinforcement size formulations and have a variety
of purposes, these include the following: to give a desired level of strand integrity or filament to
filament adhesion within a fiber bundle, abrasion resistance to prevent broken filaments and fuzz
during processing, and compatibility with the matrix resin to provide desired wet out of resin into
the fiber bundle. The various components of sizes as well as their function are given in table 3.
TABLE 3- Glass fiber size formulation
Ingredient Category Typical Chemistry Role
Film Formers Epoxies, polyesters, polyvinyl Fiber protection, strand
acetate, polyvinyl alcohol, integrity, strand wetting and
polyolefins, polyurethanes, solubility
starch, etc.
Lubricants Imidazolines, alkanolamides, Fiber protection, strand
mineral oil, polyethylene integrity, lower surface
glycol fatty acid esters, friction, improve fiber forming
vegetable oils, etc
Emulsifiers Ethoxylated fatty acids, Make film formers and
ethoxylated alkylphenols, other lubricants water compatible
ethylene oxide derivatives, etc.
Coupling Agents Silanes, titanates, zirconates, Provide glass to resin and film
chrome complexes. former bonding
Other Additives Metal halides, quaternary Antistatic agents, pH control,
ammonium compounds, acids, nucleating agents
and bases.
4.2 Filament wound pressure vessels using roving in epoxy resin. This process, either for
hand-held anti-tank rocket launchers or metal lined hoop and fully wrapped pressure cylinders
(e.g. types NGV2-2, NGV2-3) account for about 15-20% of high strength glass sales worldwide.
Safety is of paramount importance to success and the introduction of any cost saving measures in
the design or manufacture of the cylinders must not preclude this consideration. Impact
performance and resistance to stress corrosion of the composite components are as important as
strength in this application. Carbon fibers are poor in the first respect, and excel in the latter; the
inverse is true of GRP in general. Figure 3 illustrates data comparing the performance of
various glass fibers, courtesy of Powertech in BC, Canada. The % of original ultimate tensile
strength (UTS) vs. time after exposure to an acid solution (19% v/v Sulfuric acid). All fiber tows
were impregnated with Shell Epon 8132 epoxy/Ancamide 506 hardener cured in ambient
conditions for 16 hours followed by a post cure of 2 hours at 100°C.
For practical purposes, the performance of the S-2 glass® and RH glass rovings is identical, and
as they are both members of the S-glass family as shown in Table 1, any differences in corrosion
performance would be unexpected, given suitable sizing. The R-glass and EC-R glass also give
very similar performance in terms of % degradation over time. Although not shown as a
separate line, Advantex has been measured under the same conditions and found to give the same
performance as E-CR or R glass. E glass is not considered suitable for reinforcement where
acid is likely to be encountered.
10000
1000
100
E glass
Time hours
R glass
10 E-CR glass
RH glass
S-2 glass
1
% UTS 70 60 50 40 30
0.1
0.01
4.3.1 933-AA-750 in Hexcel 913 epoxy resin. A screening program based on Airbus Industrie
test methods was performed with S-2 glass and Hexcel 913 epoxy resin. This resin cures at
125°C and is used for primary aircraft structures and helicopter rotor blades. 913 has been well
established since the 1970’s, and a large database exists with many types of fiber. The results
are summarized in table 4. These data were generated in the laboratories of Tenax Fibers GmbH
& Co. KG in Wuppertal, Germany.
TABLE 4- Summary of screening program results, 125°C cure epoxy UD tape
4.3.2 E glass and S-2 glass fabrics in Hexcel 913 epoxy resin. In order to provide a back to
back comparison, a smaller program was also conducted in the same resin using both 6781 and
7581 style fabrics. Both are 8 harness satin weaves, 295 gsm, using 9µ 66 tex twisted glass
yarns. The former uses SCG75 1/0 1.0z 636 S-2 glass® yarn, the latter ECG75 1/0 1.0z E glass
yarn. The strength results are shown in figure 4. The improvements in flexural and tensile
strength obtained in using S-glass are a more modest 30-35%, notably less than in the case of UD
tape. There was only about 7-10% increase in moduli, and no change in compressive
performance. The most plausible explanation is that the “crimp” effect or the non-alignment of
filaments in the fabric negates the higher intrinsic performance of the S-glass. Improved flame
and heat resistance, and impact performance would be more compelling drivers for using S-2
glass® bi-axial fabrics.
900.0
800.0
700.0
600.0
500.0
S-2 glass 6781
MPa
E-glass 7581
400.0
300.0
200.0
100.0
0.0
Tensile Compressive Flexural
Strength, ambient
Figure 4 Comparison of S-2 glass® and E glass 8HS fabric reinforcements in epoxy resin.
4.3 Polyester based SMC. Although thousands of tonnes per year of polyester based sheet
molding compound are used by the automotive industry and others, high strength glass has never
been used in commercial volumes. Figure 5 shows a comparison of the performance of E-glass
and S-2 glass structural SMC panels, based on 65% w/w glass content, using randomly
orientated fibers of either 25mm or 50mm length.
Structural SMC Strength SMC Shear + Impact
500 35
30
400
S-2 2" 25
300 S-2 1" 20
MPa
200 E 2" 15
E 1" 10
100 5
0 0
UTS Comp. Flexural SBS - MPa Notched Izod
J/cm
30cm x 46cm panels were molded using a constant resin formulation based on AOC E987
polyester resin and constant molding conditions. The S-2 glass was 365-AA-250 9µ 2000tex
roving. The E glass was Owens Corning 973C roving. Testing was to ASTM D639, D695,
D790, D5379 and D256 for the tensile, compressive strength, flexural, shear and impact strength
respectively. The conclusions drawn are as follows:
a) The longer fibers improve tensile, flexural, and impact but not compressive strengths.
b) The S-glass gave about 32-45% greater tensile strength than E-glass, and about 13-30%
greater flexural performance. In the case of flexural performance it is clearly more cost
effective to increase the fiber length than to change the glass chemistry.
c) The S-glass improved the impact performance by over 60%, as measured by notched Izod.
d) The SBS data indicate that the 365 size may not give such good bonding to Polyester as the
973C, and further optimisation may be required.
4.4 Influence of filament diameter. There are two ways in which this may influence the
performance of a composite. The strength of the filament itself may be affected. Hillermeier (3)
found that a decrease in fiber diameter increased the tensile strength and in some cases the
moduli of polyester, PAN, rayon and E glass fibers. In the case of glass, the tensile strength
increased by a factor of 1.5 as the diameter was reduced from 13 to 3.8 µm. Other early work on
E glass (4) also indicated that the tensile strength of glass fibers varies inversely to the diameter.
On the other hand there are reports (5-7) that under certain circumstances at least the tensile
strength of E-glass fibers may be independent of filament diameter. Pabler and Bruckner (6)
found that several drawing factors in the fiber manufacturing process could influence the
strength of fibers, and Thomas (6) reports no change in tensile strength of E-glass over the
filament diameter range of 5 to 20 microns, under laboratory conditions. The most likely
explanation is that the strength is dependent on surface flaws, and under industrial manufacturing
conditions finer filament diameters will lead to greater strength.
The second way in which filament diameter will influence the mechanical properties of the
composite is by altering the geometry, such that the surface to volume ratio increases as the fiber
diameter decreases, assuming constant fiber volume fraction. Thus a greater surface area is
available for bonding at the interface. The superior compression performance of Boron fibers
(typically 100 µm diameter) compared to both glass and carbon is often attributed to the greater
buckling resistance of a fatter filament.
This section of the present study was initiated as part of a program to increase compression
strength in epoxy composites used in rotorcraft prop-rotor retention yokes and prop-rotor blades.
The baseline system used 449-AA-250 S-2 glass roving in Cytec-Fiberite E773 epoxy resin.
The roving is a conventional roving assembled from 30 ends of 66 tex slivers, sized with a
standard epoxy compatible silane blend. Standard ZenTron 721B-AB-750 roving of 13µm
filament diameter was plied 3 times to give the same overall yield as the S-2 glass, and special
production runs of 18 µm and 25µm ZenTron® were made using the same 721B size chemistry.
There appears to be a useful increase in compression strength with filament diameter. The
compression and tensile moduli are little affected, as one would expect. The surprisingly good
tensile performance of the 18µm and 25µm filaments is more difficult to explain. It is possible
that the 721B size gives better results than 449 in the E773 epoxy resin although this was not
obvious from the short beam ILSS; also single end rovings may allow better filament alignment
than multi end rovings. More work is needed to build on these preliminary results.
1800
1600
1400 9µ 2000 tex 30 end
1200 BHTI data
13µ 2000 tex 3 end
1000
MPa
600
25µ 2000 tex single end
400
200
0
Comp. Strength Tensile Strength
Figure 6 Effect of roving filament diameter on UD tape strength, normalized to 60% Fv.
50.5
50
49.5 9µ 2000 tex 30 end
49 BHTI data
48.5 13µ 2000 tex 3 end
GPa
48
18µ 1320 tex single end
47.5
47 25µ 2000 tex single end
46.5
46
45.5
Comp Modulus Tensile Modulus
Figure 7 Effect of roving filament diameter on UD tape modulus, normalized to 60% Fv.
ACKNOWLEDGEMENTS
The authors thank the Canadian Gas Research Institute (GRI) and Joe Wong of Powertech Labs
Inc. for sharing data generated for under GRI contract 5094-290-3100. We are also grateful to
Terry Wilson of AGY LLC for his contributions in adapting the ZenTron® product for filament
diameter trials.
REFERENCES
1. D.R. Hartman, M.E. Greenwood, D.M. Miller, High Strength Glass Fibers, Owens Corning
Inc., Technical paper ref. 1-Pl-19025-A, July 1996. Reprinted by AGY LLC as Pub. No.
LIT-2001-011 (05/01) May 2001.
2. Kinsella, M.A. et al., 2000, Innovative Automotive Applications Based on ZenTron® High
Strength Glass Fiber, Proc. 45th International SAMPE Symposium and Exhibition – Science
of Advanced Materials and Process Engineering Series, Vol. 45, Long Beach CA, May 21—
25, 2000. SAMPE pp 1979-1994
3. K. Hillermeier, Melliand Textilber., 50, No. 1, 26-28, (1969)
4. B.F. Jones and R.G. Duncan, J. Mater. Sci., 6, 289-293, (1971)
5. J.A. Burgman and E.M. Hunia, Glass Technology, 11, No. 6, 147, (1979).
6. W.F. Thomas, Glass Technology, 12, No. 2 42 (1971).
7. G. Pabler and R. Bruckner, Glastochn. Ber., 54, No. 3, 52 (1981)
BIOGRAPHIES
Michael A. Kinsella is employed by Advanced Glassfiber Yarns LLC where he is the market manager for the
aerospace and automotive industries. Dr. Kinsella holds a B.Sc. in Chemical Engineering from Nottingham
University and a M.Sc. and Ph.D. from the Loughborough University of Technology in the UK. His current address
is Advanced Glassfiber Yarns LLC, 29700 SW Jackson Way, Wilsonville, OR 97070 Telephone (503) 582 8244.
Dennis Murray is employed by Advanced Glassfiber Yarns LLC where he is the product engineer for high strength
glass fiber rovings and chopped strands. Mr. Murray has 15 years experience in the manufacture and development
of glass fiber products and holds a B.S. in Mechanical Engineering and Applied Mathematics from California
Polytechnic in Pomona. His current address is Advanced Glassfiber Yarns LLC, Huntingdon Plant, 1200
Susquehanna Ave., Huntingdon, PA 16652.1946 Telephone (814) 641 8304.
David Crane is employed by Bell Helicopter Textron Incorporated where he is a materials and processes engineer
for the Non-metallic Materials Engineering Laboratory Group. Mr. Crane has 20 years experience specializing in
composite materials R&D, test, evaluation, design support and manufacturing support for applications on rotorcraft
rotating equipment and airframe structure. He holds BS and MS degrees in Mechanical Engineering from the
University of Wyoming. His current address is P.O. Box 482, Fort Worth, TX 76101 Telephone: (817) 280 5404.
John Mancinelli is employed by Advanced Glassfiber Yarns LLC where he is the product engineer for high strength
glass fiber yarns. Mr. Mancinelli has 13 years experience in the manufacture and development of glass fiber
products and holds a B.S. in Industrial Engineering from Pennsylvania State University. His current address is
Advanced Glassfiber Yarns LLC, Huntingdon Plant, 1200 Susquehanna Ave., Huntingdon, PA 16652.1946
Telephone (814) 641 8319.
Mark Kranjc is employed by Advanced Glassfiber Yarns LLC in the Central Technology Organization where he is
the sizing development chemist for high strength glass fiber yarns and rovings. Dr. Kranjc has 9 years of experience
in the manufacture and development of glass fiber products. He holds a B.S. in Chemical Engineering from the
University of Cincinnati and a Ph.D. in Polymer Science from the University of Akron. His current address is
Advanced Glassfiber Yarns LLC, WHQ – CTO, 2558 Wagener Road, Aiken, SC 29801 Telephone (803) 643-1520