Corrosion Science 53 (2011) 426–431

Contents lists available at ScienceDirect

Corrosion Science
journal homepage: www.elsevier.com/locate/corsci

Degradation of basalt fibre and glass fibre/epoxy resin composites in seawater

Bin Wei b, Hailin Cao a,c,⇑, Shenhua Song b
a
Harbin Institute of Technology, Harbin 150001, China
b
Department of Materials Science and Engineering, Shenzhen Graduate School, Harbin Institute of Technology, Xili, Shenzhen 518055, China
c
Shenzhen Aerospace Tech-Innovation Institute, Shenzhen Key Laboratory of Composite Materials, Shenzhen 518057, China

a r t i c l e i n f o a b s t r a c t

Article history: Epoxy resins reinforced, respectively, by basalt fibres and glass fibres were treated with a seawater solu-
Received 31 May 2010 tion for different periods of time. Both the mass gain ratio and the strength maintenance ratio of the com-
Accepted 22 September 2010 posites were examined after the treatment. The fracture surfaces were characterized using scanning
Available online 29 September 2010
electron microscopy. The tensile and bending strengths of the seawater treated samples showed a
decreasing trend with treating time. In general, the anti-seawater corrosion property of the basalt fibre
Keywords: reinforced composites was almost the same as that of the glass fibre reinforced ones. Based on the exper-
A. Glass
imental results, possible corrosion mechanisms were explored, indicating that an effective lowering of
B. Erosion
B. SEM
the Fe2+ content in the basalt fibre could lead to a higher stability for the basalt fibre reinforced compos-
C. Interfaces ites in a seawater environment.
C. Rust Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction has advantages over glass and asbestos fibres in terms of environ-
mental cleanness [4–9].
High performance filaments, such as carbon fibre and glass As a reinforcing material for various composites, basalt fibres
fibre, are frequently used as the reinforcement of composites. have been studied extensively. The studies of basalt fibres as rein-
Property degradation of composites is unavoidable in some appli- forcements of polymer composites have mainly focused on the
cations. For example, some containers, vessels, tubes, off-shore polypropylene and epoxy resin matrix composites [10,11].
platforms and equipments in marine applications, may be corroded A fibre reinforced epoxy resin composite has a series of sound
after long-term service in a seawater environment. It is believed properties, such as high specific strength, high resistance to seawa-
that one of the obstacles preventing the extensive use of compos- ter corrosion, anti-adhesion of biological forms, and absorbing im-
ites is the lack of long-term durability and performance data when pact energy [12]. Therefore, a composite of this kind is a promising
serving in this environment. Consequently, it is necessary to material for applications in the shipping industry.
understand how the materials behave during long-term applica- In the past, many reports have revealed the chemical durability
tions [1,2]. of basalt fibres, and the good corrosion resistance of basalt fibres
Basalt fibre is a novel fibre that has been used in recent years. It [13–15]. However, until the present time the durability of the ba-
has high strength, excellent fibre/resin adhesion and ability to be salt fibre reinforced composite in seawater environment has hardly
easily processed using conventional processes and equipments. been investigated.
In addition, basalt fibres do not contain any other additives in a In the present work, artificial seawater was used to treat basalt
single producing process, which makes them have an additional fibre reinforced epoxy resin laminates, and their degradation
advantage in terms of cost. As known, basalt fibre has a higher ten- behavior was investigated. The laminates were produced using
sile strength than the E-glass fibre and its strain to failure is larger the hot-pressing process. A heat-resistant mold was first placed
than the carbon fibre. Furthermore, basalt fibre has high chemical on a curing press, after the middle part of the plastic plate was
stability and sound mechanical properties as well as being non- heated to a set temperature; the mold was pressed directly under
toxic and non-combustible [3]. The use of basalt fibre reinforced a fixed pressure. After a certain time, the sample with the same
composite could be extended to heat-insulation, filtering, and it shape as the hollow cavity of the mold was formed. In the sample
fabrication, the resin content was controlled to be around 35 vol.%.
The aim of the work was to investigate the direct effects of the arti-
⇑ Corresponding author at: Shenzhen Aerospace Tech-Innovation Institute,
ficial seawater on the performance of the basalt fibre reinforced
Shenzhen Key Laboratory of Composite Materials, Shenzhen 518057, China. Tel.: epoxy resin composites. After the seawater treatment, the mass
+86 0755 26727055; fax: +86 0755 26996829. change, tensile strength, and bending strength were measured. Fi-
E-mail address: caohl@hit.edu.cn (H. Cao). nally, the fracture characteristic of the composite was analyzed

0010-938X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.corsci.2010.09.053
 B. Wei et al. / Corrosion Science 53 (2011) 426–431 427

using scanning electron microscopy. In order to evaluate the prop- Table 2

erties of basalt fibres exactly, a comparative study on the mechan- Specifications of the basalt and glass fabrics.

ical performance of the composites reinforced both with basalt Fabric structure Basalt Plain weave Glass Plain weave
fibres and glass fibres was made. Yarn fineness (warp and weft, tex) 800 1200
Warp density (ends/10 mm) 8 8
Weft density (ends/10 mm) 8 8
2. Experimental procedures Fabric weight (g/m2) 460 680
Fabric thickness (mm) 0.38 0.43
Basalt fibres (produced by Shenzhen Academy of Aerospace Single yarn (tow) strength (N) 350 510
Single yarn elastic modulus (MPa) 90 75
Technology, China, one bundle has 2400 fibres) and glass fibres
Single yarn elongation at failure (%) 3.1 4.7
(Type E, produced by PPG Company, US, one bundle has 4000 fi-
bers) were used. The specific mass of the fibres was 800 tex for ba-
salt fibres with an average diameter of 13 lm and 1200 tex for
glass fibres with an average diameter of 12 lm. The chemical com-
positions of the fibres are listed in Table 1. 100

The specifications of the basalt and glass fibre fabrics used are
given in Table 2. E-51 epoxy resin with an epoxy value of 0.51
(from Wuxi resin factory, industrial grade) was introduced as the 75

)
matrix. Laminates were fabricated using the hot-pressing process.

Temperature (
First, the epoxy and the accelerator (2-Ethyl-4-methylilidazole)
were uniformly mixed together under the ultrasonic condition
50
according to the mass ratio of 100:7. Secondly, the prepregnate
was prepared in the form of a sheet shaped material which is pre-
impregnated and solidified in the middle level, and prepared using
the wet-winding processing (after bathing in the resin, the rein- 25
forcing material was directly winded on the core module). After
being thoroughly dried, the prepregnate was cut into small pieces
and placed in layers on the lower mold. The curing schedule is 0
shown in Fig. 1. The glass fibre reinforced composites were fabri- 0 2 4 6 8
cated in the same way. Time (h)
The composite with the same resin content were chosen and
Fig. 1. Curing procedure of epoxy composite with 2 MPa pressure applied at point A
machined into the samples for corrosion and mechanical evalua- and 5 MPa pressure applied at point B).
tions. The samples for tensile and bending tests were cut according
to the geometry and size shown in Fig. 2. Obviously, the shape of
the bending samples is similar to that of the tensile samples with
different sizes. Before the treatment in seawater, all the samples
c
were dried until a constant weight was attained. The artificial sea- h
water was prepared by mixing the sea salt with distilled water and
f
the salt concentration was controlled at 6 wt.% (about twice the
average concentration in ocean). A temperature of 25 °C was main-
tained throughout the treatment, which is the average tempera- l
ture near the southeast coast of China. In order to analyse the
mechanical property of composites, the surface energy analysis Size (mm) f l c h
of the original basalt and glass filaments of the fabrics was per- Tensile test 250 1.9-2.2
formed with Dynamic Contact Angle Meter and Tensiometer 150 0.5 25 0.5
(DCAT, Germany). Bending test 50 1.9-2.2
The samples were treated in the seawater for 10, 20, 30, 60 and 32 0.5 12.5 0.5
90 days, respectively. Mass gains of the samples were examined
using an electronic analytical balance with a precision of 1 mg. Fig. 2. Specifications of the samples for both tensile and bending measurements.
Tensile strengths of the samples before and after the treatment
were measured with a universal mechanical testing machine man-
ufactured by SANS, China. The gauge length of the sample was
Load
150 mm and a crosshead speed of 2 mm/min was employed in
the test. Ten samples were tested for each condition and the mean
Sample
value of the data points obtained was taken as the measured result.
Three-point-bending tests were carried out and the test jig is sche-
matically shown in Fig. 3. The gauge length of the sample was Pivot 2 Pivot 2
32 mm and the load was applied at the middle of the sample until
the sample was fractured. The fracture surfaces were examined Fig. 3. Schematic illustration of the bending test jig.

Table 1
Chemical compositions of the basalt and glass fibres (wt.%).

Composition SiO2 Al2O3 CaO MgO K2O Na2O FexOy TiO2 B2O3
Basalt fibre 48.39 14.7 7.7 4.7 1.6 3.0 15.3 3.8 –
Glass fibre 52.1–53.4 13.5–14.5 18.5–19.5 3.6–4.4 0–1 – 0–0.60 0–0.5 8.0–9.0
428 B. Wei et al. / Corrosion Science 53 (2011) 426–431

using scanning electron microscopy (Hitachi S-4700 field emission tained for each sample. The normal distribution method was used
gun scanning electron microscope (FEG-SEM) equipped with an to treat these data and the results are listed in Table 3. The surface
energy dispersive X-ray spectrometer (EDS, manufactured by tension of the basalt fibre is 58.75 mN/m, including 43.28 mN/m
EDAX Inc., US). polar component and 15.47 mN/m dispersive component. Com-
pared with the conventional glass fibre, both surface free energy
3. Results and discussion and polar component are higher for the basalt fibre. For the fibre
reinforced composite, the weak interface bonding is always the
3.1. Mass gain main factor of influencing the composite performance. The excel-
lent surface character of the basalt fibre may entitle a series of out-
Mass gain is one of the important parameters that characterise standing properties to its composite. Accordingly, the basalt fibre
the performance change of material in the whole degradation pro- has an enormous development potential of being a kind of rein-
cess. The degradation process includes physical process and chem- forcement material as mentioned by Deng et al. [18].
ical process. Mass gain (M) is calculated by
M1  M0 3.3. Tensile strength
M¼  100% ð1Þ
M0
The tensile strengths of the seawater treated samples are shown
where M0 is the mass at the initial state and M1 is the mass at the in Fig. 5. Ten samples were used for each data point in the figure.
treated state. Mass gain was measured using the tensile samples. The deceasing trend of the tensile strength with increasing treating
Since all the samples used in the mass gain measurement have time is seen for both basalt fibre reinforced plastic (BFRP) and glass
the same size, it is acceptable to use M to show some changes in fibre reinforced plastic (GFRP) samples. For the BFRP samples, dur-
the sample during degradation. Fig. 4 shows the mass gain of the ing the first 30-day treatment, there is an obvious drop in the ten-
composite as a function of treating time. As seen, the mass gain in- sile strength. As for the GFRP samples, only 10 day saturation was
creases quickly until 30 days and after that it does not change con- required to cause a significant tensile strength decrease. The initial
siderably for both composites. Also, the mass gain is somewhat strength of BFRP is higher than that of GFRP. There are two reasons
higher for the basalt fibre reinforced composite than for the glass fi- for the above phenomena: one is the strength of the basalt fibres
bre reinforced one. (2.6 GPa) is somewhat higher than that of the glass fibres
When the samples were immersed in the seawater, the water (2.5 GPa), and the other is the BFRP’s interface property is better
molecules would penetrate into the material. This could increase than the GFRP’s one. This is confirmed by the inter-layer shear
the sample mass. At the same time, some soluble compound would strength (ILSS) of BFRP being larger (63 MPa) than that of GFRP
be extracted into the seawater solution which could cause the (56 MPa). This is also consistent with the surface free energy anal-
mass loss. Accordingly, the results presented in Fig. 4 arise from ysis of the fibres. It is known that the epoxy matrix is inert to the
the combination of these two effects. One might imagine that at water environment. Voids and cracks of the resin allow moisture to
the beginning the moisture penetrates quickly into the material, penetrate the composites, promoting the breakdown of the matrix
leading to a sharp increase in the sample weight as shown in structure. When in the dissolved state, NaCl exists as cations and
Fig. 4. It is believed that the higher void content in the matrix anions. Ions would penetrate along with the water molecules into
and fibre–matrix interface would facilitate the above effect [16]. the composite, causing damage to the matrix, fibre and interface.
After 60-day seawater treatment, there were local small white These would deteriorate the material, resulting in a decreased ten-
spots emerging at the edges and corners of the samples; this could sile strength.
be attributed to the extraction of soluble materials [17].
Table 3
3.2. Surface energies of the fibres Surface free energies of fibres (mN/m).

Fibre Surface energy Dispersive component Polar component

The fibres with sizing were first heated for 3 h at 100 °C. Then it Basalt fibre 58.75 15.47 43.28
was cooled down to room temperature in a dessicator. Single fibre Glass fibre 40.01 37.28 2.73
infiltration was tested by DCAT at 25 °C. Fifty data points were ob-

0.5 BFRP 1800

GFRP BFRP
1700 GFRP
0.4
Mass gain ratio (%)

1600
Tensile strength (MPa)

0.3 1500

1400
0.2

1300
0.1
1200

0.0 1100
0 20 40 60 80 100 0 20 40 60 80 100
Immersion time (days) Immersion time (days)

Fig. 4. Weight gain rate as a function of immersion time (error bars represent the Fig. 5. Tensile strength of the seawater treated samples as a function of immersion
standard deviation). time (error bars represent the standard deviation).
 B. Wei et al. / Corrosion Science 53 (2011) 426–431 429

3.4. Bending strength of the sample has also changed from black to ivory-white, and
there are many small grooves appearing along the fibre direction
Fig. 6 illustrates the bending behavior of the treated samples. A (see Fig. 7b). In addition, the color of the seawater had changed
decreasing trend of the bending resistance with treating time can from colorless to pale yellow, and its pH value is somewhat in-
be noticed in the figure. It is well known that the stress state of creased from 7.1 to 7.4. SEM images of the cross section of the BFRP
the bending sample is very complex. The upper surface mainly sample after 90-day treatment are shown in Fig. 8. Clearly, there
bears a compressive stress while the lower surface bears a tensile are many cracks present at the interfaces between fibres and epoxy
stress and the middle of the sample bears a shear stress. As shown matrix (see Fig. 8a). It is seen from the high magnification image
in Fig. 6, at the early stage, the bending strength of BFRP is higher that the epoxy resin has peeled off from the basalt fibres in some
than that of GFRP. After 50-day seawater saturation, the BFRP’s places (see Fig. 8b).
bending strength is close to the GFRP’s bending strength. After that, Figs. 9 and 10 show the SEM fractographs of the GFRP and BFRP
the BFRP’s bending strength falls continuously to an even smaller bending samples. The fracture surface of the untreated GFRP sam-
value, and is lower than the GFRP’s bending strength. It is believed ple indicates that the fibres are broken neatly, showing the effec-
that, under the conditions used, the oxides of the alkaline metals in tive bonding of the matrix on the fibres. After 90-day seawater
the fibres are chemically unstable, and thus they can react with immersion, the fibre/matrix interface has been seriously damaged,
other ions. This could cause the composites to deteriorate, result- and thus the fibres can no longer be broken simultaneously so that
ing in a reduced bending resistance. the sample breaks and bursts into many thin strands (see Fig. 9b).
As for the GFRP sample, the case is similar to the BFRP sample with
3.5. Corrosion mechanism the residual resin on the fibres less than the BFRP sample and these
fibres break into even thinner strands (see Fig. 10b). EDS micro-
Surface changes of the BFRP sample after 90-day seawater treat- analysis, as shown in Table 4, indicate the composition change of
ment are shown in Fig. 7. The surface luster of the BFRP sample is the basalt fibre surface layer after 90-day treatment. The concen-
decreased substantially after being treated. The color of some parts trations of the elements in the fibre are decreased except Si, Ti
and Fe. These alkali metals’ diffusion behavior from the fibre into
the solution can also partially explain the pH value increase of
the seawater. The concentrations of Si and Ti are not apparently
180 decreased probably due to their stable chemical. The role of Fe in
BFRP
GFRP
the whole degradation process is discussed below.
Fig. 11 shows the schematic diagram of the BFRP corrosion
160 mechanism in seawater environment. There is a little difference
Bending strength (MPa)

in the degradation process between BFRP composites and GFRP

ones, although the composition of the basalt fibre is more complex
140 than that of the glass fibre. Especially, there is Fe element existing
in the former with mainly Fe2+ [15,19]. During immersion, H2O, O2,
CO2 molecules and Cl, Na+ ions can penetrate into the matrix
120
(including the resin and the interface) through channels and/or
voids and react with the resin and/or fibre. At the same time, there
are some components leached out from the fibre, including Ca, Mg,
100
Al, and K and so on. These leached elements may form a hydrated
layer at the interface [20]. As a result, the microstructure of the
0 20 40 60 80 100 composite is changed and the cracks are consequently formed,
Saturation time (days) leading to the sporadic breaking of fibres in the mechanical test.
This implies that the bonding effect of the matrix on the fibres
Fig. 6. Bending strengths of the seawater treated samples at different saturation
times (error bars represent the standard deviation). has been greatly reduced. Also the leaching of network modifying
elements plays a minor role in the mechanical property deteriora-
tion. The main difference in the degradation mechanism between
BFRP and GFRP originates from chemical reactions involving Cl
and Fe2+. This is suggested by the color of the sea water after treat-
ment, since iron(II) chloride is known to be yellow. The chemical
reactions involved are described as follows:

Fe2þ þ Cl ! ½FeCl complex ð2Þ


½FeCl complex þ OH ! FeðOHÞ2 þ Cl ð3Þ

FeðOHÞ2 þ O2 þ H2 O ! FeðOHÞ3 ! Fe2 O3  nH2 O ð4Þ

Thereinto, reaction (2) occurs under the condition of [21]

Cl =OH P 0:6 ð5Þ

Until the reaction product can be detected, the whole reaction

process above may take a long time. From these chemical reactions
along with Fig. 11, we can imagine that the final reaction product
Fig. 7. Surface changes of the BFRP samples (a) untreated and (b) treated for of rust (Fe2O3nH2O) at the interface can absorb more H2O mole-
90 days. cules. This will cause the matrix to expand, making the formation
430 B. Wei et al. / Corrosion Science 53 (2011) 426–431

Fig. 8. SEM images of the cross section of the BFRP sample after 90-day treatment. (a) low magnification and (b) high magnification).

Fig. 9. SEM images of the tensile fracture surfaces of the BFRP bending specimens (a) untreated and (b) treated for 90 days.

Fig. 10. SEM images of the tensile fracture surfaces of the GFRP bending specimens (a) untreated and (b) treated for 90 days.

Table 4
Surface compositions of the basalt fibres before and after 90-day seawater treatment
(wt.%).

Element O Si Na Mg Al K Ca Ti Fe
Before 36.4 29.8 1.65 3.3 9.4 1.5 5.6 2.4 9.9
After 34.1 32.4 1.7 3.2 9.1 1.45 5.4 2.5 10.1

of microcracks easy under a tensile stress and consequently result-

ing in the bending strength of BFRP being lower than that of GFRP.

4. Conclusions

The mass gain change of the BFRP and GFRP composites after
being immersed in seawater is a combination of two effects: water
absorption and soluble material extraction. At the initial stage the Fig. 11. Schematic diagram of the BFRP’s corrosion mechanism in seawater.
 B. Wei et al. / Corrosion Science 53 (2011) 426–431 431

water absorption is predominant, and at the late stage the soluble [2] Thomas K. Tsotsis, Scott Keller, Kyejune Le, Jason Bardisc, Jack Bishd, Aging of
polymeric composite specimens for 5000 h at elevated pressure and
material extraction may play a major role. The tensile and bending
temperature, Compos. Sci. Technol. 61 (2001) 75–86.
strengths decrease with increasing immersion time, indicating that [3] M. Berozashvili, Continuous reinforcing fibers are being offered for
the material has experienced some forms of physical damage and/ construction Civil engineering and other composites applications, Adv Mater
or irreversible chemical degradation. The penetrated water may Com News, Compos Worldwide 6 (2001) 5–6.
[4] F.N. Rabinovich, V.N. Zueva, L.V. Makeeva, Stability of basalt fibers in a medium
cause the matrix to swell and break, and the soluble material of hydrating cement, Keramika 12 (2001) 29–32.
extraction at the interface may damage the coherence between [5] S.L. Xiong, Y. Chen, Z.W. LI, J.L. Shi, Feasibility study on the use of basalt fibers
the matrix and the fibre. All these effects make the properties of in friction material, Fiber glass 6 (2005) 5–11.
[6] Bulent Ozturk, Fazli Arslan, Sultan Ozturk, Hot wear properties of ceramic and
the composite deteriorate. basalt fiber reinforced hybrid friction materials, Tribol. Int. 40 (2007) 37–48.
In general, it may be claimed that: [7] X. Wang, B. Hu, Y. Feng, F. Liang, J. Mo, J. Xiong, Y.Q. Low, Velocity impact
properties of 3D woven basalt/aramid hybrid composites, Compos. Sci.
Technol. 68 (2008) 444–450.
(1) Diffusion/infiltration is a fundamental driving force of all [8] Jongsung Sim, Cheolwoo Park, Do Young Moon, Characteristics of basalt fiber
corrosion damage. The chemical degradation caused by dif- as a strengthening material for concrete structures, Composites Part B 36
fusion/ infiltration is the main form of the matrix material (2005) 504–512.
[9] Jirı Militky, Vladim_ır Kovacic, Jitka Rubnerov, Influence of thermal treatment
corrosion damage. The diffusion of sodium chloride mole- on tensile failure of basalt fibers, Eng. Fract. Mech. 69 (2002) 1025–1033.
cules (chloride ions) around the composite toward the inside [10] T. Czigany, J. Vad, K. Poloskei, Basalt Fiber as a Reinforcement of Polymer
of the material is the main reason for corrosive degradation. Composites Periodica Polytechnica, Mech. Eng. 49 (2005) 3–14.
[11] T. Czigany, Basalt fiber Reinforced Hybrid Polymer Composites, Mater. Sci.
(2) The hydrolysis reaction in the composite has occurred,
Forum 473–474 (2005) 59–66.
including the hydrolysis of fibre, matrix, and their interface. [12] G.Z. Liu, D.H. Li, L.G. Huang, D.X. Zhang, Experiments on corrosion mechanism
This hydrolysis may break the molecular chain and reduce and performance evolution rule of GFRP in salt steam environment, FRP/CM 1
the curing degree of cross-linked network, resulting in the (2008) 35–40.
[13] G. Huang, Tensile behaviors of quartz Aramid and glass filaments after NaOH
changes of material performance. treatment, Mater. Des. 29 (2008) 1893–1896.
[14] C. Scheffler, T. Förster, E. Mäder, G. Heinrich, S. Hempel, V. Mechtcherine, Aging
On the whole, the chemical stability of BFRP and GFRP in seawa- of alkali-resistant glass and basalt fibers in alkaline solutions: evaluation of the
failure stress by Weibull distribution function, J. Non-Cryst. Solids 355 (2009)
ter is nearly the same. It is suggested that an effective lowering of 2588–2595.
the Fe2+ content in the basalt fibre could lead to a higher stability of [15] B. Wei, H.L. Cao, S.H. Song, Environmental resistance and mechanical
BFRP in seawater environment. performance of basalt and glass fibers, Mater. Sci. Eng., A 527 (2010) 4708–
4715.
[16] G. Huang, Behaviors of glass fiber/unsaturated polyester composites under
Acknowledgements seawater environment, Mater. Des. 30 (2009) 1337–1340.
[17] M.C. Wang, Z.G. Zhang, Y.B. Li, M. Li, Z.J. Sun, Chemical Durability and
Mechanical Properties of Alkali-proof Basalt Fiber and its Reinforced Epoxy
This work was financially supported by The National Natural Composites, J. Reinf. Plast. Compos. 27 (2008) 393–407.
Science Foundation of China (Grant No. 50503007). The authors [18] R. Deng, M. Li, Z.G. Zhang, J.J. Jia, Surface free energy of basalt and glass fibers
by contact angle method, J.Beijing Uni. Aeronau. Astronaut. 33 (2007) 1349–
would like to thank Professor Andrew Peter Baker for his help in 1352.
manuscript preparation. [19] Dorothee J.M. Burkhard, Torsten Scherer, The effect of initial oxidation state on
crystallization of basaltic glass, J. Non-Cryst. Solids 352 (2006) 3961–3969.
[20] Majbritt D. Lund, Yuan-Zheng Yue, Influences of chemical aging on the surface
References morphology and crystallization behavior of basaltic glass fibers, J. Non-Cryst.
Solids 354 (2008) 1151–1154.
[1] Y. Shan, K. Liao, Environmental fatigue of unidirectional glass-carbon fiber [21] P. Kumar Mehta, Paulo J. M. Monteiro, Concrete: Microstructure, Properties
reinforced hybrid composite, Compos. Part B 32 (2001) 355–363. and Materials, ISBN 978-7-5083-6457-5, 113.