The 12th International Conference of the Slovenian Society for Non- Destructive Testing »Application of Contemporary Non-Destructive Testing in Engineering«

September 4-6, 2013, Portorož, Slovenia

EFFECT OF MOISTURE ABSORPTION ON MECHANICAL PROPERTIES OF POLYESTER COMPOSITES EVALUATED

WITH DESTRUCTIVE AND NONDESTRUCTIVE TESTS

Roman Šturm1, Raimond Grimberg2, Janez Grum1

1 University of Ljubljana, Faculty of Mechanical Engineering, Aškerčeva 6, 1000 Ljubljana, Slovenia, roman.sturm@fs.uni-lj.si; janez.grum@fs.uni-lj.si

2 Nondestructive Testing Department, National Institute of R&D for Technical Physics, Iasi, Romania, grimberg@phys-iasi.ro

ABSTRACT In this research work the effect of moisture absorption on the mechanical properties of glass reinforced polyester composites is described. Composite resins were produced with two different production processes. Mechanical properties of composite materials were measured with DMA destruction tests. According to DMA the dependency by temperature of real component of the complex elastic modulus (E’), imaginary component of the complex elastic modulus (E”), as well as tan (δ) being traced. For a more efficient employment of the composite materials, the compliance tensor was obtained by means of nondestructive tests using ultrasound. A method for generation and reception of Lamb waves in plates of composite materials is represented by the using of air-coupling low frequency ultrasound transducers in pitch–catch configuration. The results of nondestructive measurements proposed in this paper are in good concordance with those obtained by DMA destructive tests. Key words: Glass Reinforced Polyester Composite, Moisture Absorption, Elastic Modulus, Ultrasound Testing 1. Introduction The mechanical properties of a material determine manufacturability, performance, and longevity; thus, knowing mechanical properties is essential for making good design decisions. Polymers are exceptionally complex materials – mechanical properties depend on chemistry, processing, and thermo-mechanical history. Mechanical properties also depend on volume constraints. Thus, in order to gain useful information for making sound decisions when designing with polymer composites, mechanical property measurements should be made on a relevant sample in a relevant context. Glass fiber reinforced polyester (GRP) is the most “popular” composite. The matrix is based on cured thermosetting resin. Its first main civilian application was for building boats. GRP uses also include hot tubs, pipes for drinking water and sewers, office plant display containers and flat roof systems. The mechanical performances of GRP composites depends on the fiber strength and modulus, the matrix strength and chemical stability, and the effectiveness of interface bonding between the matrix and fiber to enable stress

189

mailto:roman.sturm@fs.uni-lj.simailto:janez.grum@fs.uni-lj.simailto:grimberg@phys-iasi.ro

transfer [1,2]. Fiber-reinforced composite materials offer a combination of strength and modulus that are either comparable to or better than many pure materials [3]. But thermoplastic show hydrophobic nature [4]. The performance of GRP pipes is critical in many engineering applications when subjected to a combination of high temperature/high humidity environments. Diffusion of water or aqueous fluid into GRP pipes may lead to changes in thermo-physical, mechanical, and chemical characteristics. Many of these changes can result in degradation of the material’s performance. In order to properly predict the service lives of the GRP pipe, one must understand the mechanisms that govern these changes. Water absorption by resin may cause both reversible and irreversible changes of resin, including hydrolysis [5], plasticization [6,7], micro-cracking [8], and even glass transition temperature. For the reversible process, the mechanical properties can usually be recovered by drying, and for the irreversible case, the mechanical properties are permanently altered. Many studies have shown that temperature and environment are the most critical factors in reducing the strength of GRP materials. Generally, the higher the temperature of the environment and the longer the exposure time is, the larger the decrease in strength and modulus of GRP will be [9]. In those applications in which GRP composites are subject to a heat and mechanical load, it is essential to determinate the elastic properties using nondestructive methods, the best methods being the ultrasound methods [10,11,12]. 2. Experiment 2.1. Studied samples Samples of GRP plates having as reinforcement 6 sheets of ravings with 250± 50 gm-2 density and matrix from different types of unsaturated Orthophthalic polyester resins, made by Helios, Slovenia, have been taken into study. The characteristics of studied GRP samples are presented in Table 1.

Table 1: Studied GRP samples.

Sample name

Matrix Fiber volume ratio

Density [kg/m3]

Observations Production process

7201 COLPOLY 7201 0.43±0.005 1550±20 medium reactivity resin

In two steps

7524 COLPOLY 7524 0.43±0.005 1530±20 Chemical resistance

In situ

7243 COLPOLY 7243 0.57±0.005 1410±20 Preaccelerated thixotropic

In two steps

2.2. Water Absorption Test The effect of water absorption on GRP composites was investigated. Initially, DMA measurements on non-immersed samples and immersed samples have been effectuated on no conditioned samples. The obtained results were inconclusive. From this reason, the measurements were redone, this time on conditioned samples, the water that might be in the initial samples being eliminated. The conditioning has been made maintaining all the samples for 5 days into a drying oven at 500C. From each type of composite, one sample was weighted in order to have the initial mass. This sample has been marked. The immersion was done in distilled water. The immersion periods were 3, 6, 9, 12, 20, 40 and 100 days. From mistake, the sample from 7524 composite immersed 100 days is missing.

190

After each period of immersion, two of each type of samples was taken out, one of samples was used for determination of adsorbed water and the second one has served at DMA measurements. The adsorbed water was determined by ratio after each period.

00

(%) 100ttM MM

M−

= ⋅ (1)

Where M0 is the mass of the no immersed sample after conditioning and Mt is the mass of the same sample immersed during t time. In Figure 1 is presented the dependency of Mt (%) by the immersion time for the three types of composites taken into study. As reviewed by Weitsman, water absorption behavior of composite materials can be categorized into several types [13]. The curves in Figure 1 denote linear Fickian behavior, where the moisture weight gain gradually attains equilibrium after a rapid initial take off. In order to determine the water diffusion coefficient in composite, the data from Figure 1 were converted so that the abscise of the graphic in Figure 2 shall be √𝑡 expressed in √𝑠𝑒𝑐.

Fig. 1: Water uptake v.s immersion time. Fig. 2: Water uptake vs √𝑖𝑚𝑚𝑒𝑟𝑠𝑖𝑜𝑛 𝑡𝑖𝑚𝑒 in √𝑠. Considering that the studied composites absorb the water after the law of Fick:

[%] 4[%]

tM DtM h π∞

= (2)

where M∞ represents the maximum of water quantity that diffuses in composite, being evaluated after the curves from Figure 2, and h is the thickness of the sample. The values for diffusion coefficients are given in Table 2.

Table 2: Diffusion coefficient of moisture in different composites.

Composite D [m2s-1] 7201 1.81x10-12

7524 1.5355x10-12 7543 9.55x10-13

2.3. DMA measurements The measurement of elastic modulus along the three directions was made with a Dynamic Mechanical Analyzer DMA 242C – Netzsch Germany with a 3 points bending fixture and using the analysis software Proteus v.4.8.5. The measurements were effectuated at 1Hz frequency. Polymers are often employed in products because of their ability to both store and damp energy. The complex modulus E* is a phase vector which incorporates both capacities:

E* = E' + E'' (3)

0

0,1

0,2

0,3

0,4

0,5

0,6

0 500 1000 1500 2000 2500

Wat

er U

ptak

e (%

)

Time (h)

720175247543

0

0,1

0,2

0,3

0,4

0,5

0,6

0 500 1000 1500 2000 2500 3000

Wat

er U

ptak

e (%

)

√(Time) [√(s)]

7201

7524

7543

191

The real part (E') of the complex modulus is called the storage modulus because it quantifies the material's ability to store energy elastically. In materials with insignificant damping, the storage modulus is equivalent to Young's modulus. The imaginary part of the complex modulus (E'') is called the loss modulus, because it quantifies the material's ability to damp out energy. The dimensionless loss factor tan (δ) is independent of contact energy, because it is the ratio of the loss to the storage modulus:

tan(δ) = 𝐸′

𝐸′′ (4)

The mechanical performance of a composite material depends strongly on its glass transition temperature, Tg. Above this temperature the mechanical properties decrease rapidly. Therefore, it is necessary to check the influence of moisture on glass transition temperature of the GRP specimen. The glass transition temperature of the samples was determined with a DMA instrument. Both dry and immersed specimens were tested to determine the aging produced by water immersion. Figures 3,4 and 5 present the DMA results for dry and wet GRP specimens. The storage modulus, loss modulus, and loss factor are taken as ordinate while the temperature is taken as abscissa. Some observation can be made: first, there are some changes for storage modulus, loss modulus, and loss factor, i.e., peak value of loss factor. At low temperatures (0 to 40°C) the polymer is stiff/frozen and has a high storage modulus E' and low modulus E''. The chains are frozen in fixed positions because insufficient energy for translational and rotational motions of the polymer segments is available. As the temperature increases, the polymer obtains sufficient thermal energy to enable its chains to move more freely. At temperatures larger than 100°C the storage modulus decreases to about 500 MPa. The increasing of E’ after 3 days of immersion is due to the fact that water acts as plasticizer, initially, and then, once with the increasing of the immersion time, E’ decreases (Fig. 3). In analogy with this process a transition occurs at temperatures between 60 to 100°C and the loss modulus E'' goes through a maximum. This region is referred to as the glass rubber transition or glass transition Tg and can be qualitatively interpreted as the onset of large scale conformational rearrangements of the polymer chain backbone. The temperature at which this peak maximum is observed is conventionally defined as Tg. Figure 5 shows the loss factor tan (δ) of a dry and wet specimens. The peak value of tan (δ) is considered as glass transition temperature. For each immersion time and each composite, the glass transition temperature Tg and the activation energy for glass transition Eg in the basis of Arrhenius rule were determined. One can see clearly that moisture absorption by the specimen results in changing Tg, which indicates degradation process in the material (Fig. 6). It can be observed also that the water act initially as plasticizer. But after 100 days of immersion in water, glass transition temperature is almost the same as it is of non-immersed specimens (Fig. 6). In Figure 7 is presented also glass transition activation energy versus immersion time for all three composites.

Fig. 3: The dependency by temperature of storage modulus E’.

192

Fig. 4: The dependency by temperature of loss modulus E’’.

Fig. 5: The dependency by temperature of loss factor tan (δ).

Fig. 6: Glass transition temperature vs. immersion time.

Fig. 7: Glass transition activation energy vs. immersion time.

80

85

90

95

100

105

110

0 500 1000 1500 2000 2500Gla

ss tr

ansi

tion

tem

pera

ture

(°

C)

Time (h)

7201

7524

7543

140

180

220

260

300

0 500 1000 1500 2000 2500

Gla

ss tr

ansi

tion

activ

atio

n en

ergy

(KJ/

mol

)

Time (h)

7201

7524

7543

193

2.4. The principle of the ultrasound measurement method For the measurement of propagation speed of ultrasound longitudinal and transversal waves, we have used the equipment composed by a Pulser Receiver 5073 PR Panametrics NDT USA, set-up on impulse–echo mode; the signal received being visualized on Digital Oscilloscope LeCroy WaveRunner 64Xi with 600 MHz passband and 10 GS/s the sampling rate. The transducer G5KB–GE USA was used for generation of longitudinal waves with ZG-F gel Krautkramer-Germany for coupling and MB4Y-GE USA for the transversal waves using honeybee as coupling. The transducers were placed on a delay block made from Plexiglas with 20 mm thickness. The measurement for the phase velocity of Lamb waves, modes A0 and S0 were made using air-coupling US transducers, pitch-catch configuration, NCG100S25–UltranGroup USA, having the central frequency 100 kHz. The incident angle was calculated according to eq. (5), considering cair=340m/s and cs having the value obtained by determination of transversal waves propagation speed. The value of θ was 10.20. The incident angle, θ, is chosen so the SH0 modes horizontally polarized shear waves shall appear, that is:

arcsinz

air

s

cc

θ

=

(5)

Where cair is the ultrasound speed in air at the experiment temperature. The determination of phase velocity of fundamental modes S0 and A0 was made using the phase spectrum method [14]. The phase velocity of Lamb waves is given by:

2p

fLc πϕ

=∆

(6)

Where ϕ∆ is the difference in the phase spectrum of two signals that were collected with a different distance between them of L and f is the frequency. For a more simple visualization of modes and to determine the group velocity of Lamb wave, we can use the method of “spectrogram” [15] which represents a time-varying spectral representation that shows how the spectral density of a signal delivered by the US reception transducer varies with time. The principle scheme of the equipment is presented in Figure 8a, and in Figure 8b is presented the physical realization.

a) b)

Fig. 8: The experimental set-up for determination of Lamb waves propagation speed using air-

coupled US transducer: a) Basic scheme; b) Photo. The studied GRP samples are all, transversally isotropic as it has been experimentally demonstrated. In the case of transverse isotropy only five elastic coefficients can be independent [10]. More precisely, the isotropy normal to Z direction (Figure 9) requirement involves that

Aε σ= ⋅ (7) or

194

1 0 0 0

1 0 0 0

1 0 0 0

10 0 0 0 0

10 0 0 0 02

10 0 0 0 02

xy xz

x x x

xy xzx xx x xy yxz xz

z zx x z

xy xyxy

yz yzx

xz xz

xz

xz

E E E

E E E

E E E

E

G

G

υ υ

υ υε σε συ υε σε συε σ

ε σ

− −

− − − − = +

(8)

Where ε is the strain vector, σ is the stress vector, and A represents the tensor of elastic compliance.

Fig. 9: The coordinate system attached to GRP plate.

The five coefficients which characterize the material are

- two Young (or extensional) moduli Ex and Ez - two shear moduli Gxz - two Poisson’s ratios xzυ and yzυ

The Young modulus Ez, shear modulus Gxz and Poisson’s ratio xzυ can be determined by measuring the velocities of longitudinal waves

zpc and of transversal waves zsc that propagate along Z direction [12]:

( )( )( )

11 1 2z

z

z xzp

xz xz

xzs

Ec

Gc

υρ υ υ

ρ

−=

+ −

=

(9)

According to basic relationship [10,11,16] we can express for the fundamental symmetric mode, S0, the phase speed as:

( )0 21x

sxy

Ecρ υ

=−

(10)

The phase speed for the fundamental anti-symmetric mode, A0, can be obtained after a few algebraic calculus:

0

1/41/2

2p

AD

ch

ωρ

=

(11)

Where Dp is the flexural rigidity of the plate: ( )( )

383 2x x x

px x

hD

µ λ µλ µ

+=

+ (12)

195

and xλ and xµ are Lamé coefficients:

( )( )1 1 2x xy

xxy xy

x xy

E

G

υλ

υ υ

µ

=+ −

=

(13)

3. Experimental results By mean of destructive procedures, the mechanical descriptors that define the compliance tensor from eq. (8) were determined. These data are for non-immersed GRP specimens presented in Table 3 as average value of 5 samples from the same product. In addition, the glass transition temperature and the Arrhenius activation energy of this transition were determined [17].

Table 3: Mechanical properties determined by DMA.

Composite Ex=Ey [GPa]

Ez [GPa]

υxy Tg

[0C] Activation energy

[kJ/mol] 7201 9.1 8.2 0.2 82.9 149.6 7524 10.5 9.65 0.2 110.7 286.2 7243 8.4 8.2 0.2 105.7 296.4

Ez, υxz and Gxz were determined from the measurement of propagation speed of the ultrasound along the direction Z (according to Figure 9), the other mechanical properties being determined from the propagation speed of the Lamb waves, the fundamental modes S0 and A0. In Figure 10a is presented the signal received by the reception transducer at the examination of 7201 composite, where the modes A0, S0 and SH0 can be distinguished. For a more reliable identification, the amplitude spectrum has been traced, being presented in Figure 10b. In Figure 10c is presented the corresponding spectrogram. The signal is presented in time-domain (Figure 10a), frequency domain (Figure 10b) and as spectrogram (Figure 10c). Using spectrum phase method, the phase velocity of A0 and S0 were determined, and, using the eqs. (10) and (11), the other measures from the compliance tensor form eq. (8) can be determined. The data of ultrasound measurements for all three types of composites are synthetized in Table 4.

a) b)

196

c)

Fig. 10: The experimental results: a) The signal received by the reception transducer at the examination of 7201l composite; b) the amplitude spectrum; c) the spectrogram.

Table 4: Mechanical properties evaluated with ultrasound measurements.

Composite Cp

[ms-1] Cs

[ms-1] CA0

[ms-1] CS0

[ms-1] Ex=Ey [GPa]

Ez [GPa]

υxy υxz Gxy [GPa]

Gxz [GPa]

7201 2261 1550 912 1980 8.9 7.9 0.2 0.19 3.7 3.6 7524 2844 1603 1094 2230 10 9.5 0.2 0.27 4.2 3.0 7243 2868 1512 1036 2250 8.3 8.4 0.21 0.31 4.2 3.2

Comparing the data from Table 3 and 4, it can be observed a very good correlation between the mechanical measures that define the compliance tensor, determined by destructive and non-destructive procedures. 4. Conclusions Water absorption behavior of three types of GRP composite was recorded, and the results can be concluded as below:

- Destructive tests showed good mechanical results for composite 7524, where polyester is produced in-situ. In-situ production means cheaper process in comparison with similar composite 7201. Glass transition temperature is 10°C higher in the case of 7524 composite comparing to 7201 composite.

- A method for generation and reception of Lamb waves in plates of composite materials is represented by the using of air-coupling low frequency ultrasound transducers in pitch –catch configuration.

- The results of nondestructive measurements proposed in this paper are in good concordance with those obtained by classical destructive tests.

Acknowledgements This paper is partially supported by The National Authority for Scientific Research of Romania (ANCS) under Romanian Slovenian Bilateral Cooperation Program-Contract No.371/2009 – RO-SI and Slovenian Research Agency for the Republic of Slovenia (SRA) under Romanian Slovenian Bilateral Cooperation Program-Contract No.: BI-RO/10-11-002.

197

5. References 1. Mayer R., Design with Reinforced Plastics, Springer, Berlin, 1993. 2. Erden S., Sever K., Seki Y., Sarikanat M.: Enhancement of the Mechanical Properties of

Glass/Polyester Composite via Matrix Modification, Fibers and Polymers, Vol. 11, 2010, 732-737.

3. Rowell R.M. et al.:Utilization of Natural Fibers in Plastic Composites, Problems and Opportunities, Journal of Thermoplastic Composites Materials, Vol. 15, No. 4, 1997, 281-300.

4. Kim H.S., Lee B.H., Choi S.W., Kim S. Kim H.J.: The Effect of Types of Maleic Anhydride-Grafted Poly- propylene (MAPP) on the interfacial Adhesion Properties of Bio-Flour-Filled Polypropylene Composites, Composites Part A, Vol. 38, 2007, 1473-1482.

5. Apicella A., Migliaresi L., Nicolais L., Iaccarino and Roccotelli S.: The Water Ageing of Unsaturated Polyester-based Composites: Influence of Resin Chemical Structure, Composites, Vol. 14, 1983, 387–392.

6. Apicella A., Nicolais L.: Effect of Water on the Properties of Epoxy Matrix and Composite, Advances in Polymer Science, Springer-Verlag, Berlin, Vol. 72, 1986, 69–77.

7. Antoon M.K., Koenig J.L.: Irreversible Effects of Moisture on the Epoxy Matrix in Glass-reinforced Composites, Journal of Polymer Science: Polymer Physics Edition, Vol. 19, 1981, 197–212.

8. Talreja R.J.: Transverse Cracking and Stiffness Reduction in Composite Laminates, J. Compos. Mat., Vol. 19, 1985, 355.

9. Ray B.C.: Temperature Effect During Humid Ageing on Interfaces of Glass and Carbon Fibers Reinforced Epoxy Composites, J. Colloid and Interface Science, Vol. 298, No. 1, 2006, 111–117.

10. Lemaitre J., Chaboche J.L.: Mechanics of Solid Materials, Cambridge University Press, NY, 1994.

11. Schmerr L.W. Jr.: Fundamentals of Ultrasonic Nondestructive Evaluation – A Modeling Approach, Plenum Press, NY, 1998.

12. Krautkramer J., Krautkramer H.: Ultrasonic Testing of Materials, 4th ed., Springer Verlag, 1990-

13. Weitsman Y.: Fluids Effects in Polymeric Composites – An Overview, In: Progress in Durability Analysis of Composite Systems: Proceedings of the Third International Conference on Progress in Durability Analysis of Composite Syst., Rotterdam, 1998.

14. Sachse W., Poo, Y.H.: On the determination of phase and group velocities of dispersive waves in solids, J.Acoust.Soc.Am., Vol. 49, 1978, 4320-4327.

15. Boashash B.: Time-frequency signal analysis and processing – A comprehensive reference, Elsevier, Oxford, 2003.

16. Grum J., Šturm R., Barsanescu P., Savin A., Steigmann R., Grimberg R.: Ultrasound Examination of Fibre-Glass Polyester Composite, 12th International Symposium on Nondestructive Characterization of Materials, 2010, Blacksburg, Virginia, USA

17. Menard K.P.: Dynamic Mechanical Analysis – A Practical Introduction, CRC Press, 1999.

198