Composite Structures
Composite Structures
Composite Structures
Composite Structures
journal homepage: www.elsevier.com/locate/compstruct
a r t i c l e i n f o a b s t r a c t
Article history: Composite materials are used in several engineering applications, where can be exposed to a range of
Available online 18 July 2014 corrosive environments during their in-service life. Thus, it is necessary understanding the impact of a
corrosive environment in the working life of composite. According the authors’ knowledge, the effect
Keywords: of different commercial oils was not yet studied. Therefore, glass fibre/epoxy composites were subjected
Environmental degradation to oil immersion tests, using an universal multi-grade engine oil (15W-40) and an extra high performance
Mechanical properties hydraulic brake fluid (DOT 4), in order to study the effects of oil absorption behaviour on flexural and
Mechanical testing
impact strength properties of glass fibre/epoxy composites. Both solutions affect the flexural properties
Polymer–matrix composites (PMCs)
and the impact strength. However, for all tests performed, the automotive brake fluid promotes the low-
est values comparatively to the automotive engine oil.
Ó 2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.compstruct.2014.07.017
0263-8223/Ó 2014 Elsevier Ltd. All rights reserved.
2 A.M. Amaro et al. / Composite Structures 118 (2014) 1–8
temperature solutions or to higher exposure duration, the polyes- down to room temperature maintaining pressure and finally get
ter resin has lower modulus values than the bisphenol A epoxy the part out from the mould. The plates were manufactured in a
vinyl ester resin. However, in both resins, the average hardness useful size of 300 300 2.1 mm3.
increases after 2 weeks of exposition and, then, decreases after The specimens used in the experiments were cut from these
4 weeks exposure (but still higher than the unexposed). Finally, thin plates, using a diamond saw and a moving speed chosen to
they have concluded that under acid and higher temperature expo- reduce the heat in the specimen. The static three point bending
sures, the microstructure of polyester material degrades more rap- (3 PB) tests were performed using specimens cut nominally to
idly than the microstructure of bisphenol A epoxy vinyl ester resin, 100 14 2.1 mm3. Relatively to the impact tests, the samples
which can be justified by the increased surface roughness, cracks used were cut from those thin plates to square specimens with
and diffusion of sulphur into the cracks. Amaro et al. [13,14] com- 100 mm side and 2.1 mm thickness (100 100 2.1 mm3).
pared the effects of alkaline and acid solutions on glass/epoxy com- Finally, the specimens were completely submerged into two types
posites. Authors showed that the alkaline solution is more of oils, 15W-40 (an universal multi-grade engine oil) and DOT 4 (an
aggressive than the acid solution, promoting the lowest flexural extra high performance hydraulic brake fluid). The exposure time
properties. In terms of acid solutions, for example, the hydrochloric was 15 and 45 days at room temperature. An important remark
acid is responsible by the worst results [14]. The ultramicroinden- is the fact that both faces of the samples were exposed to the oil
tation showed a decreasing of the matrix mechanical properties solution, what it is not very typical in real conditions. Afterwards,
and the roughness was higher for the samples immersed in alka- they were cleaned with tissue paper.
line solutions [13,14]. The bending tests were performed according to ASTM D790-2,
In fact, the acid and alkaline solutions significantly affect the using a Shimadzu AG-10 universal testing machine equipped with
mechanical properties, but their effects are strongly dependent of a 5 kN load cell and TRAPEZIUM software at a displacement rate of
the exposure time [13–16], concentration [16–18] and tempera- 5 mm/min. All 3PB tests were also carried out at room tempera-
ture of the solution [15,17]. Relatively to the exposure time, Mah- ture, with a span of 34 mm and, for each condition, 5 specimens
moud and Tantawi [15] found a significant influence of this were used. Bending strength was calculated as the nominal stress
parameter on the flexural strength, hardness and Charpy impact at middle span section obtained using maximum value of the load.
resistance. For Stamenovic et al. [16], the changes observed on The nominal bending stress was calculated using Eq. (1):
the tensile properties are proportional to the exposure time (num- 3PL
ber of days in solution). Moreover, these authors found that ulti- r¼ 2
ð1Þ
2bh
mate tensile strength and modulus show significant decreases
with increasing the pH value. According Mortas et al. [17], the cor- where P is the load, L is the span length, b is the width and h is the
rosive environment decreases the impact strength and its effect is thickness of the specimen. The stiffness modulus was calculated by
highly dependent of the concentrations of solution. Finally, inde- the linear elastic bending beams theory relationship (Eq. (2)):
pendently of the nature of aggressive solutions and of their con- DP L3
centration, the fact is that if composites are exposed to acid E¼ ð2Þ
48Du I
solutions at higher temperatures, the exposure induces a decreas-
ing of the mechanical properties [15,17]. where I is the moment of inertia of the cross-section and DP, Du are,
In the literature there are several works about the effect of respectively, the load range and the flexural displacement range at
moisture, hygrothermal and aggressive solutions (alkaline and acid middle span for an interval in the linear region of the load versus
solutions) on the mechanical properties of composite materials. displacement plot. The stiffness modulus was obtained by linear
But, according the authors’ knowledge was not yet performed regression of the load–displacement curves considering the interval
any study about the effect of different commercial oils in the in the linear segment with a correlation factor greater than 95%.
mechanical properties of composite materials. Thus, the aim of this The low velocity impact tests were made using a drop weight-
work is to increase the knowledge of those materials’ degradation testing machine Instron-Ceast 9340. A hemispherical impactor
under the exposure of two different oil solutions: 15W-40 (an uni- with diameter of 10 mm and mass of 3.4 kg was used. The tests
versal multi-grade engine oil) and DOT 4 (an extra high perfor- were performed on circular section samples of 70 mm of diameter
mance hydraulic brake fluid). In this paper, the understanding of and the impactor stroke at the centre of the samples. The samples
degradation is evaluated in terms of flexural and impact strength were centrally supported by a specimen of 100 100 mm dimen-
mechanical properties. The bending tests were selected because, sions. The impact energy used for all impacts was 4 J, which corre-
according Banna et al. [12], are the most sensitive to changes of sponds to an impact velocity of 1.53 ms1. For each condition, five
exposure conditions and, on the other hand, impacts at low veloc- specimens were tested at room temperature. After impact tests, all
ity are very dangerous, because they affect dramatically the the specimens were inspected in order to evaluate the size and
mechanical performance of such materials [19–23] and, at same shape of the delaminations. As the glass-laminated plates are
time, the damages promoted are difficult to detect visually [24,25]. translucent, it is possible to obtain the image of the damage using
photography. Nevertheless, in order to achieve the best possible
definition of the damaged area, the plates were photographed in
2. Material and experimental procedure
counter-light using a powerful light source. Plates were framed
in a window so that all the light could fall upon them.
Composite laminates were prepared in the laboratory from
Based on BS EN ISO 62:1999 standard, the following procedure
glass fibre Prepreg TEXIPREGÒET443 (EE190 ET443 Glass Fabric
was used to obtain the oil absorption: the samples were placed in
PREPREG from SEAL, Legnano, Italy) and processed in agreement
an oven at 40 °C for 6 h, then cooled and weighed in order to obtain
with the manufacturer recommendations, using the autoclave/vac-
the dry weight (DW); afterwards, a series of samples were
uum-bag moulding process. The laminates were manufactured
immersed in the respective oils (15W-40 and DOT 4) and, periodi-
with the stacking sequence [02,902]2s. The processing setup con-
cally, weighted to obtain the current wet weight (CWW). The oil
sisted of several steps: make the hermetic bag and apply
absorption, in weight percentage (W%), was calculated using Eq. (3):
0.05 MPa vacuum; heat up to 125° C at a 3–5 °C/min rate; apply
a pressure of 0.5 MPa when a temperature of 120–125 °C is CWW DW
W% ¼ 100 ð3Þ
reached; maintaining pressure and temperature for 60 min; cool DW
A.M. Amaro et al. / Composite Structures 118 (2014) 1–8 3
After exposure, the roughness profiles were obtained by a Mitutoyo all conditions tested. Both curves are practically linear until the
equipment, model SJ-500. In order to obtain the roughness param- maximum load is reached, where – a sudden drop occurs. In order
eters of all samples, measures were made in different zones of sam- to explain the differences observed, the failure mechanisms are
ple. Simultaneous, the surface topology was also observed in a presented in Fig. 2. For all conditions tested, the damage starts
scanning electron microscope (SEM) and, all specimens, were sput- by the fracture of the fibres in the tensile surface following, poste-
tered coated with a 10 nm layer of gold prior to SEM observation. riorly, different damage mechanisms. Relatively to the control
The morphology was evaluated using Philips XL30 equipment. samples (Fig. 2a), after the fracture of the fibres in tensile, with
Finally, the hardness was evaluated by ultramicroindentation, using small delaminations around the broken fibres, occurs compressive
Fisherscope H100 equipment and a load of 500 mN. The hardness breakage of the longitudinal fibres in the pin load contact region.
values were corrected for the geometrical imperfections of the Vick- According to Reis et al. [27], this phenomenon is consequence of
ers indenter, the thermal drift of the equipment and the uncertainty the high compressive stress concentration in the pin load contact
in the zero position according with reference [26]. The hardness, H, region associated to the low compressive strength of the fibres.
is defined as the maximum applied load during the indentation test, For the samples immersed in the 15W-40 solution (Fig. 2b), asso-
Pmax, divided by the contact area of the indentation immediately ciated to the fracture of the fibres in the tensile surface occurs sig-
before unloading, AC. nificant delaminations around the broken fibres. Similar damage
can be found for the samples exposed to the DOT 4 solution
3. Results and discussion (Fig. 2c), however, in this case, extensive delaminations (cracks)
occurs between layers of the mid-thickness in tensile. It is evident
Tensile and bending tests were performed by Banna et al. [12] that these solutions promote matrix/fibre interface degradation,
in order to evaluate the degradation of composite materials under especially for the aggressive one, where this damage mechanism
aggressive solutions and concluded that the last ones (bending is more evident (Fig. 2c). These main damages are in agreement
tests) are the most sensitive. Therefore, Fig. 1 shows an example
of the load versus flexural displacement curves for the control sam-
ples and samples exposed at 15W-40 and DOT 4 solutions, consid-
ering different exposure time. These curves are representative of
(a)
1 mm
700
Control (a)
600 15W40 15 days
DOT4 15 days
500
400
Load [N]
300
(b)
200
1 mm
100
0
0 1 2 3 4
Displacement [mm]
700
Control (b)
600 DOT4 15 days
DOT4 45 days
500
(c)
Load [N]
400
1 mm
300
200
100
Cracks
0
0 0.5 1 1.5 2 2.5 3 3.5
Displacement [mm]
Fig. 1. Typical load–displacement curves: (a) Effect of the solution; (b) Effect of the Fig. 2. Failure mechanisms: (a) Control samples; (b) Samples exposed to 15W40
exposure time. during 45 days; (c) Samples exposed to DOT4 during 45 days.
4 A.M. Amaro et al. / Composite Structures 118 (2014) 1–8
Table 1
Effect of the solutions on bending properties.
Load [N]
15 days 810.2 21.9 27.2 2.23
45 days 764.6 21.5 25.3 2.27
Energy [J]
DOT4
15 days 798.9 14.9 26.1 2.70
45 days 735.9 16.3 23.4 2.87
(a)
50 µm
(b) (c)
50 µm 50 µm
(d) (e)
50 µm 50 µm
Fig. 5. SEM pictures for: (a) Control samples; (b) Samples exposed to 15W40 during 15 days; (c) Samples exposed to 15W40 during 45 days; (d) Samples exposed to DOT4
during 15 days; (e) Samples exposed to DOT4 during 45 days.
6 A.M. Amaro et al. / Composite Structures 118 (2014) 1–8
Table 4 Table 6
Statistics of the roughness measurements. Number of impacts to failure.
Samples/statistic values (lm) Exposure time (days) Samples Number of impacts to failure
0 15 45 Control 12
15W40
Control samples
15 days 11
Arithmetic average, Ra 2.07 ± 0.16 – –
45 days 11
Root mean square, Rq 2.52 ± 0.19 – –
Average peak to valley height, Rz 10.59 ± 0.29 – – DOT4
Core roughness depth, Rk 6.45 ± 0.17 – – 15 days 10
Reduced peak height, Rpk 2.49 ± 0.18 – – 45 days 8
Reduced valley depth, Rvk 1.22 ± 0.17 – –
15W40
Arithmetic average, Ra – 2.38 ± 0.16 2.71 ± 0.15
Root mean square, Rq – 3.39 ± 0.15 3.44 ± 0.16 of the load-penetration depth curves obtained. These curves agree
Average peak to valley height, Rz – 11.62 ± 0.21 13.22 ± 0.27 with other ones presented in the literature [13,14,26] and the
Core roughness depth, Rk – 6.34 ± 0.18 8.25 ± 0.16 important quantities in this loading–unloading cycle are maximum
Reduced peak height, Rpk – 6.78 ± 0.19 5.03 ± 0.12 load, maximum depth, final depth after unloading and the slope of
Reduced valley depth, Rvk – 2.19 ± 0.17 1.68 ± 0.16
the upper portion of the unloading curve known as the elastic con-
DOT4 tact stiffness. Table 5 presents the average values of hardness,
Arithmetic average, Ra – 2.88 ± 0.14 3.99 ± 0.15
indentation modulus (ER) and Young’s modulus (E) of the matrix.
Root mean square, Rq – 3.57 ± 0.15 5.51 ± 0.18
Average peak to valley height, Rz – 15.05 ± 0.28 24.01 ± 0.29 It is possible to conclude that the measured hardness decreases,
Core roughness depth, Rk – 9.58 ± 0.18 10.46 ± 0.17 independently of the solution and exposure time, in comparison
Reduced peak height, Rpk – 3.93 ± 0.19 3.53 ± 0.16 with the control samples (0.288 GPa). For example, after 45 days
Reduced valley depth, Rvk – 2.53 ± 0.17 11.61 ± 0.15
of exposition, the average value of hardness decreases around
10.8% and 11.5%, respectively, for universal multi-grade engine
oil (15W-40) and hydraulic brake fluid (DOT 4). Finally, the inden-
250 tation modulus (ER) and Young’s modulus of the matrix was
Control obtained according with Antunes et al. [26], where the indentation
15 days- 15W40 modulus (ER) is a function of the Young’s modulus (E) and the Pois-
200 son ratio (m) of the specimen and the indenter. For both parame-
45 days - 15W40
ters, a similar evolution to the hardness can be found. After
Indentation Load [mN]
15 days- DOT4
150 45 days of exposure a decrease about 23.2% and 33.0% occurs,
45 days- DOT4
respectively, for the 15W-40 solution and DOT 4 solution in com-
parison to the control samples. Therefore, this can explain also
100 the lower flexural properties observed associated to the matrix/
fibre interface degradation.
Finally, the effect of these solutions on laminates subjected to
50 multi impacts will be analysed. Table 6 shows the impact resis-
tance and the laminates are considered failed when full perforation
occurs. Full perforation is defined when the impactor completely
0 moves through the samples. The performance of the laminates to
0 2 4 6 8
repeated low velocity impacts is dependent of the solutions. After
Indentation depth [mm]
45 days of immersion, for example, the impact resistance decreases
Fig. 6. Schematic representation of the typical load-penetration depth curves. around 8.3% for the 15W-40 solution and 33.3% for the DOT 4 solu-
tion, when compared with the control samples. This was expected,
because the automotive brake fluid promotes major damages on
the exposition to the DOT 4 solution promotes the highest values, the laminates, in terms of matrix and interface matrix/fibre, as seen
where Rz increases around 42.1% and 126.7% after 15 days and previously. Fig. 7 shows the aspect of the samples (one impact
45 days of exposure time, respectively, relatively to the control before of the full penetration) for each condition tested, where is
samples. Therefore, besides the effect on the adhesion between evident the matrix/fibre interface degradation.
the fibres and matrix, DOT 4 solution promotes multiple cracks Fig. 8 presents the evolution of the elastic energy (restitution
on the matrix. energy) against the normalised impact numbers (N/Nf), where
The hardness was also evaluated by ultramicroindentation and the last impact it is not represented because occurs full penetration
Fig. 6 shows, for all conditions tested, a schematic representation (elastic energy = 0). N is the current number of impact and Nf is the
Table 5
Effect of the solutions on average values of hardness, indentation modulus and Young’s modulus of the polymeric matrix.
Samples Aver. hardness VH (GPa) Std. dev. (GPa) Aver. ER (MPa) Std. dev. (MPa) Aver. E (GPa) Std. dev. (GPa)
Control samples 0.288 0.01 8.50 1.59 7.83 1.47
15W40
15 days 0.258 0.01 6.73 0.66 6.13 0.62
45 days 0.257 0.01 6.55 0.85 6.01 0.80
DOT4
15 days 0.256 0.01 6.35 0.59 5.83 0.50
45 days 0.255 0.01 6.19 0.84 5.25 0.78
A.M. Amaro et al. / Composite Structures 118 (2014) 1–8 7
(a) (b)
(c)
Fig. 7. Aspect of the samples, one impact before of the full penetration, for: (a) Control samples; (b) Samples immersed in 15W40 solution during 15 days; (c) Samples
immersed in DOT4 solution during 15 days.
80 4. Conclusions
70 This work studied the flexural and low velocity impact response
60 of a glass fibre/epoxy composite after immersion in universal
multi-grade engine oil (15W-40) and in extra high performance
Elastic energy [%]
50
Elastic energy [%]
Acknowledgement
40 600
This research is sponsored by FEDER funds through the program
30 COMPETE – Programa Operacional Factores de Competitividade –
400
and by national funds through FCT – Fundação para a Ciência e a
20 Tecnologia, under the project PEst-C/EME/UI0285/2013.
Control samples
15W40 - 45 days
200
10 References
DOT4 - 45 days
0 0 [1] Gellert EP, Turley DM. Seawater immersion ageing of glass-fibre reinforced
0 0.2 0.4 0.6 0.8 1 polymer laminates for marine applications. Compos Part A-Appl Sci
1999;30:1259–65.
N/Nf [2] Imielinska K, Guillaumat L. The effect of water immersion ageing on low-
velocity impact behaviour of woven aramid–glass fibre/epoxy composites.
Fig. 9. Elastic energy and damaged area against number of impacts. Compos Sci Technol 2004;64:2271–8.
[3] Reis PNB, Silva AP, Santos P, Ferreira JAM. Hygrothermal effect on the impact
response of carbon composites with epoxy resin enhanced by nanoclays. Mech
Compos Mater 2013;49:429–36.
number of impacts to failure. The elastic energy decreases when [4] Kaushal S, Tankala K, Rao RMVGK. Some hygrothermal effects on the
the impact numbers increases and this tendency show to be very mechanical behaviour and fractography of glass-epoxy composites with a
dependent of the solution and exposure time. On the other hand, modified interface. J Mater Sci 1991;26:6293–9.
[5] Boukhoulda BF, Adda-Bedia E, Madani K. The effect of fiber orientation angle in
lower elastic energy means that the damage area is major and this composite materials on moisture absorption and material degradation after
relationship can be found in Fig. 9. hygrothermal ageing. Compos Struct 2006;74:406–18.
8 A.M. Amaro et al. / Composite Structures 118 (2014) 1–8
[6] Selzer R, Friedrich K. Mechanical properties and failure behavior of carbon [20] de Moura MFSF, Marques AT. Prediction of low velocity impact damage in
fibre-reinforced polymer composites under the influence of moisture. Compos carbon-epoxy laminates. Compos Part A-Appl Sci 2002;33:361–8.
Part A-Appl Sci 1997;28:595–604. [21] Amaro AM, de Moura MFSF, Reis PNB. Residual strength after low velocity
[7] Akay M, Kong Ah, Mud S, Stanley A. Influence of moisture on the thermal and impact in carbon–epoxy laminates. Mater Sci Forum 2006;514–516:624–8.
mechanical properties of autoclaved and oven-cured Kevlar-49/epoxy [22] Amaro AM, Reis PNB, de Moura MFSF. Delamination effect on bending
laminates. Compos Sci Technol 1997;57:565–71. behaviour in carbon–epoxy composites. Strain 2011;47:203–8.
[8] Reis PNB, Ferreira JAM, Costa JDM, Santos MJ. Fatigue performance of Kevlar/ [23] Reis PNB, Ferreira JAM, Antunes FV, Richardson MOW. Effect of interlayer
epoxy composites with filled matrix by cork powder. Fiber Polym delamination on mechanical behavior of carbon/epoxy laminates. J Compos
2012;13:1292–9. Mater 2009;43:2609–21.
[9] Tsai YI, Bosze EJ, Barjasteh E, Nutt SR. Influence of hygrothermal environment [24] Adams RD, Cawley PD. A review of defects types and non-destructive testing
on thermal and mechanical properties of carbon fiber/fiberglass hybrid techniques for composites and bonded joints. NDT Int 1998;21:208–22.
composites. Compos Sci Technol 2009;69:432–7. [25] Amaro AM, Reis PNB, de Moura MFSF, Santos JB. Damage detection on
[10] Aoki Y, Yamada K, Ishikawa T. Effect of hygrothermal condition on laminated composite materials using several NDT techniques. Insight
compression after impact strength of CFRP laminates. Compos Sci Technol 2012;54:14–20.
2008;68:1376–83. [26] Antunes JM, Cavaleiro A, Menezes LF, Simões MI, Fernandes JV. Ultra-
[11] Ray BC. Temperature effect during humid ageing on interfaces of glass and microhardness testing procedure with Vickers indenter. Surf Coat Technol
carbon fibers reinforced epoxy composites. J Colloid Interf Sci 2002;149:27–35.
2006;298:111–7. [27] Reis PNB, Ferreira JAM, Antunes FV, Costa JDM. Flexural behaviour of hybrid
[12] Banna MH, Shirokoff J, Molgaard J. Effects of two aqueous acid solutions on laminated composites. Compos Part A-Appl Sci 2007;38:1612–20.
polyester and bisphenol A epoxy vinyl ester resins. Mater Sci Eng A-Struct [28] Griffiths R, Ball A. An assessment of the properties and degradation behavior of
2011;528:2137–42. glass-fibre-reinforced polyester polymer concrete. Compos Sci Technol
[13] Amaro AM, Reis PNB, Neto MA, Louro C. Effects of alkaline and acid solutions 2000;60:2747–53.
on glass/epoxy composites. Polym Degrad Stabil 2013;98:853–62. [29] Aslan Z, Karakuzu R, Okutan B. The response of laminated composite plates
[14] Amaro AM, Reis PNB, Neto MA, Louro C. Effect of different acid solutions on under low-velocity impact loading. Compos Struct 2003;59:119–27.
glass/epoxy composites. J Reinf Plast Compos 2013;32:1018–29. [30] Hosur MV, Adbullah M, Jeelani S. Studies on the low-velocity impact response
[15] Mahmoud MK, Tantawi SH. Effect of strong acids on mechanical properties of of woven hybrid composites. Compos Struct 2005;67:253–62.
glass/polyester GRP pipe at normal and high temperatures. Polym Plast [31] Iqbal K, Khan S-U, Munir A, Kim J-K. Impact damage resistance of CFRP with
Technol Eng 2003;42:677–88. nanoclay-filled epoxy matrix. Compos Sci Technol 2009;69:1949–57.
[16] Stamenovic M, Putic S, Rakin M, Medjo B, Cikara D. Effect of alkaline and acid [32] Reis PNB, Ferreira JAM, Santos P, Richardson MOW, Santos JB. Impact response
solutions on the tensile properties of glass–polyester pipes. Mater Des of Kevlar composites with filled epoxy matrix. Compos Struct
2011;32:2456–61. 2012;94:3520–8.
[17] Mortas N, Erb O, Reis PNB, Ferreira JAM. Effect of corrosive solutions on [33] Reis PNB, Ferreira JAM, Zhang ZY, Benameur T, Richardson MOW. Impact
composites laminates subjected to low velocity impact loading. Compos Struct response of Kevlar composites with nanoclay enhanced epoxy matrix. Compos
2014;108:205–11. Part B-Eng 2013;46:7–14.
[18] Pai R, Kamath MS, Rao RMVGK. Acid resistance of glass fibre composites with [34] Amaro AM, Reis PNB, Magalhães AG, de Moura MFSF. The influence of the
different layup sequencing: part I – diffusion studies. J. Reinf Plast Compos boundary conditions on low velocity impact composites damage. Strain
1997;16:1002–12. 2011;47:e220–6.
[19] Davies GAO, Hitchings D, Zhou G. Impact damage and residual strengths of [35] de Moura MFSF, Gonçalves JPM. Modelling the interaction between matrix
woven fabric glass/polyester laminates. Compos Part A-Appl Sci 1996;27: cracking and delamination in carbon–epoxy laminates under low velocity
1147–56. impact. Compos Sci Technol 2004;64:1021–7.