Delamination analysis of carbon fiber composites under hygrothermal environment using acoustic emission

Abstract

Non-destructive test technique for monitoring delamination failure under complex load and environments is still not mature until now. The purpose of this paper is to study mixed-mode delamination failure properties of carbon fiber/epoxy composite laminates under hygrothermal environment using acoustic emission. Different water-saturated composite specimens with initial intralaminar and interlaminar defects are tested. Two loading modes including single-leg and over-leg three-point bending are applied under hygrothermal environment. By analyzing the responses of acoustic emission parameters including amplitude and energy, the effects of the hygrothermal environment, layup pattern and initial defect on the delamination behaviors of composite specimens are studied. Besides, different failure modes are observed through scanning electron microscope. Quantitative acoustic emission results show hygrothermal environment and load mode affect the delamination properties of composites remarkably.

Keywords

Delamination carbon fiber/epoxy composite laminates hygrothermal environment acoustic emission

Introduction

Complicated failure mechanisms are distinct features of carbon fiber composites which generally include intralaminar fiber breakage, matrix cracking, fiber/matrix interface debonding and interlaminar delamination.^1,2 In particular, delamination decreases the stiffness, strength and integrity of composite structures largely. Furthermore, complex environments such as moisture and high temperature accelerate the delamination failure due to performance deterioration or ageing of polymer materials.³ In general, moisture increases the flexibility and ductility of resin/adhesive layers,^4–7 but decreases their elastic moduli and strengths.⁸ It is essential to understand how hygrothermal environment affects the mechanical properties of composites.

A lot of research has already been performed to address the hygrothermal degradation mechanisms of composites. The flexural properties for unidirectional composites were found to be fiber-dominated before ageing, where brittle failure with fiber breakage was observed.⁹ For interlaminar shear test, reduction of the strength of a short composite beam due to water ageing indicating interface degradation was observed.¹⁰ By orthogonal test, the solution temperature was shown to be the most important factor affecting the mechanical properties of composites compared with the moisture concentration and test period.¹¹ Currently, multidirectional composites by optimizing the layup pattern are specially designed to achieve favorable mechanical properties. It was shown that moisture reduces thermal residual stresses in composites due to matrix swelling and increases the static strengths of optimized composite structures.¹² However, the strength degradation of water-saturated composite specimens leads to damage initiation in cross-ply laminates at a low stress level compared to dry specimens, which arises from progressive interface degradation due to the sensitivity of adhesive layers to water attack when the water molecule attempts to invade the interface.¹³ Furthermore, the initiation and growth of voids for porous glass/polyester composites expand the interface regions, facilitating the invasion of water into the composites.¹⁴ Subsequent moisture diffusion and swelling decrease the load-bearing capability and fracture toughness of the matrix, especially at high temperature.¹⁵

A primary work is to seek advanced non-destructive test techniques to gain deep insight into the failure mechanisms and damage evolution behaviors of composites under hygrothermal environment. Acoustic emission (AE) has become an important test technique because it can dynamically monitor the damage evolution behaviors of composites.¹⁶ Various failure mechanisms above can be identified by analyzing the responses of AE characteristic parameters including amplitude and energy.^17–20 Dominating frequency distribution can be also obtained by Fast Fourier Transform.^21–23 In addition, AE wave propagation and velocity evolution of glass/epoxy composites in the fiber orientation were studied,²⁴ and damage evolution properties in carbon fiber composites were evaluated.^25,26 In particular, delamination initiation and growth^27–30 and interactions between matrix cracking and delamination³⁰ of composites were studied using AE. However, there is little research on the delamination failure mechanisms and degradation mechanisms of composites under hygrothermal environment using AE.

This paper performs three-point bending experiment and AE test on carbon fiber/epoxy composite laminates with initial intralaminar and interlaminar defects under hygrothermal environment. The originality lies in the identification and analysis of dominant delamination failure mechanisms of composite specimens using AE signals under hygrothermal environment. By analyzing the response of AE amplitude and energy, the effects of the hygrothermal environment, load mode, layup pattern and initial defect on the delamination behaviors of composites are investigated. This work provides a technique support for studying the delamination mechanisms of fiber-reinforced polymer composites under hygrothermal environment.

Experimental procedures

Material specimens

Material parameters for T700/BA9916 composites are listed in Table 1. Composite laminates include two layup patterns: [0°]₁₆//[0°]₁₆ and [15°/−15°]_4s//[15°/−15°]_4s (“//” denotes the initial delamination). Currently, single-leg bending (SLB) has become a general method to study the delamination behaviors of composites.^31–33 Recently, Szekrenyes and Uj³⁴ further proposed over-leg bending (OLB) loading mode. In this research, two loading modes: SLB and OLB are comparatively used, as shown in Figure 1. Geometry sizes and layups for two composite specimens are listed in Table 2. Each specimen includes 32 plies, and the thickness of each ply is about 0.15 mm. The fiber volume fraction is 61%. A Teflon film is inserted at the middle plane of specimens to make an initial delamination crack. Before AE test, the specimens are soaked in 70℃ distilled water until moisture saturation. By comparison, the composite specimens under room temperature are taken as the regular specimens.

Table 1.

Elastic parameters of T700/BA9916 composites.

$E_{11}$ (GPa)	$E_{22}$ (GPa)	$E_{33}$ (GPa)	$G_{12}$ (GPa)	$G_{13}$ (GPa)	$v_{12}$	$v_{13}$
116	8.9	8.9	4.15	4.15	0.27	0.27

where E, G and v denote the Young’s modulus, shear modulus and Poisson's ratio, respectively.

Table 2.

Geometry sizes and layup patterns for two composite specimens (mm).

Number	Length×width ×thickness	Initial crack length	Truncation length	Layup patterns
①	150 × 20 × 4.8	50	25	[0°]₁₆//[0°]₁₆
②	150 × 20 × 4.8	50	25	[15°/−15°]_4s//[15°/−15°]_4s

where the subscript “s” denotes the symmetry and (“//”) denotes the delaminated interface.

Three-point bending test and AE monitoring

Three-point bending test is performed in a UTM5000 electronic test machine, as shown in Figure 2. A 48-channel AE equipment is used, and two AE sensors are placed on two ends of composite specimens. The AE equipment is composed of a transducer, a 2/4/6-AST preamplifier and a PAC Samos-48 AE apparatus. After more than 10-day soak of dry specimens in water, specimens are weighed each day. If the weight difference within two consecutive days is smaller than 0.02%, the moisture-saturated specimens are obtained,³⁵ which are then placed in a high-temperature chamber under 70℃ temperature and 100% relative humidity for 10 min and loaded up to collapse. The load speed is 2 mm/min. AE signals are detected by the sensors and enhanced by the preamplifier which employs enlarging circuits with a frequency range 50 kHz–2 MkHz. The gain selector of the preamplifier is set to be 40 dB in order to filter the noise. Timing parameters are: peak definition time = 50 µs, hit definition time = 200 µs and hit lockout time = 300 µs. The response distributions of AE parameters including amplitude and energy are recorded.

Results and discussion

Mechanical properties of composite specimens

Figure 3(a) shows the load–displacement curves for the specimen-① using the SLB and OLB tests under room temperature and hygrothermal environments, respectively. Figure 3(b) shows the corresponding load–displacement curves for the specimen-②. From Figure 3, each curve has two local peak points. The first peak point corresponds to the critical value for the delamination initiation, similar to the conclusion by Szekrenyes and Uj.³⁴ The second peak point represents the collapse of structures. It is shown that the critical and maximum loads for the hygrothermal specimen are smaller than those for the regular specimen. This indicates that the hygrothermal environment has an adverse effect on the load-bearing capability of specimens. The load-bearing capability of specimens using the OLB test is higher than that using the SLB test, but the softening stage using the SLB test is more distinct than that using the OLB test, which shows the crack propagation using the OLB test is more stable.³⁴ By comparing Figure 3(a) and (b), the initial delamination load for the specimen-② is larger than that for the specimen-①, which shows stronger delamination resistance for the specimen-② than the specimen-① no matter under room temperature or hygrothermal environment.

Delamination behaviors of composite specimens

By analyzing the distributions of AE characteristic parameters including the amplitude and energy, the delamination failure behaviors and load-bearing abilities of hygrothermal specimens using the SLB and OLB tests are studied. The waveform AE signal parameters are shown in Figure 4, including mainly three characteristic parameters: energy, counting and amplitude.³⁶ Energy and amplitude are used to analyze the failure mechanisms of composites. Figure 5 shows the load/energy–displacement distributions, the amplitude–time distribution and the amplitude–location distribution for the [0°]₁₆//[0°]₁₆ specimen for the SLB test under hygrothermal environment. As shown in Figure 5(a), an abrupt change of the load corresponds to the AE peak energy, which means AE energy is high, while the load decreases steeply. The test was stopped manually, and the displacement is constant at the transition line. According to Figure 5(a) and (b), the failure process can be divided into three stages. At the first stage (displacement 0–2.5 mm), the delamination crack starts to propagate when the load increases rapidly to the critical value 295 N, which is determined by detecting the sudden drop of load and the appearance of more AE energy. Before the initial delamination, low AE energy is mixed with 40–60 db matrix cracking, 50–70 db interface debonding, 60–80 db delamination and little 80–100 db fiber breakage signals.²⁰ At the second stage (displacement 2.5–11.3 mm), the strain energy exceeds the critical value, and the crack starts to propagate where the load decreases to 104 N. AE energy increases abruptly to 31,500 aJ (aJ is AE energy unit that is defined by voltage×time) because of more energy dissipation due to the dominant delamination growth. As shown in Figure 6(a), different failure modes at the crack tip are observed by scanning electron microscope (SEM). At this stage, there are signals with a wide range of amplitudes represented by matrix cracking, interface debonding, interlaminar delamination and fiber breakage. According to the amplitude–location distribution shown in Figure 5(c), there are two sites with dense amplitude signals: the load end and the crack tip. Because the load end is fixed, the delamination crack tip is identified by another location with dense amplitude. When the delamination crack propagates to the 45 mm location as shown in Figure 5(c), the slope of the load–displacement curve is smaller than that at the first stage because of stiffness degradation due to delamination.^21,26 At the third stage (after 11.3 mm displacement), the specimen collapses after reaching the maximum load, where both the load and AE energy reach the maximum values 582.1 N and 34,500 aJ, respectively. More signals for fiber breakage appear, but those for interface debonding and delamination decrease. By comparison, Figure 7 shows that AE energy signals are fewer for the regular specimen at this stage, represented by matrix cracking, interface debonding and delamination. From the whole delamination process, there are peaks for 80–100 db signals at the location 57–78 mm near the load end. These signals appear during the middle and later periods of the stable crack propagation. As shown in Figure 6(b), the SEM observation results validate more fiber breakage at the load end for the SLB test. However, the 80–100 mm location corresponds to the fast crack propagation with few AE signals. In addition, there are 40–60 db matrix cracking and 50–70 db interface debonding signals at the location 10–50 mm.

Figure 8 shows the load/energy–displacement distributions, the amplitude–time distribution and the amplitude–location distribution for the [0°]₁₆//[0°]₁₆ specimen for the OLB test under hygrothermal environment. Similar to the SLB test, the OLB test can be also divided into three stages including elastic deformation, stable crack propagation and final collapse. Because of different delamination fracture mode-ratios for the SLB and OLB tests, there are fewer signals including matrix cracking, interface debonding and delamination for the OLB test than those for the SLB test at the first stage. As shown in Figure 9, there are different failure modes observed by SEM at the crack tip. At the second stage in Figure 8(a), the crack propagates to approximately the 67 mm location. AE energy for the delamination of the hygrothermal specimen is more than that of the regular specimen, as shown in Figure 10. At the third stage in Figure 8(a), the load is relatively smooth according to the results of the OLB test, different from the sharp drop according to the SLB test. The 40–60 db signals for matrix cracking are dense, but there are few other signals. There is some difference between the amplitude–location curves in Figures 8(c) and 5(c). Because of slow delamination crack propagation for the OLB test, the signals appear mainly at the right location 70 mm, indicating the crack propagation for the OLB test is more stable and easier to be controlled than that for the SLB test.³⁴ In addition, the maximum load for the OLB test is larger than that for the SLB test.

Figure 11 shows the load/energy–displacement distributions, the amplitude–time distribution and the amplitude–location distribution for the [15°/−15°]_4s//[15°/−15°]_4s specimen for the SLB test under hygrothermal environment. The test for the [15°/−15°]_4s//[15°/−15°]_4s specimens are also divided into three stages, similar to the [0°]₁₆//[0°]₁₆ specimen using the SLB test. When the load increases to the critical value 421 N for the [15°/−15°]_4s//[15°/−15°]_4s specimen that is higher than 295 N for the [0°]₁₆//[0°]₁₆ specimen, delamination starts to appear. This shows that the [15°/−15°]_4s//[15°/−15°]_4s specimen exhibits higher delamination resistance under hygrothermal environment. At the second stage, the load drops slightly to 289.4 N and then continues to increase unstably, different from the [0°]₁₆//[0°]₁₆ specimen using the SLB test. The AE peak energy for the SLB test under hygrothermal environment is 38,000 aJ for the [15°/−15°]_4s//[15°/−15°]_4s specimen, larger than 32,500 aJ for the [0°]₁₆//[0°]₁₆ specimen at the critical load. As shown in Figure 12, AE peak energy under hygrothermal environment is much higher than that under regular environment. As shown in Figure 13, different failure modes are observed by SEM at the crack tip. According to Figure 11(c), the crack propagates to approximately the 55 mm location for the [15°/−15°]_4s//[15°/−15°]_4s specimen using the SLB test, shorter than that for the [0°]₁₆//[0°]₁₆ specimen using the SLB test. Because the longitudinal strength for the [15°/−15°]_4s//[15°/−15°]_4s specimen is weaker than that for the [0°]₁₆//[0°]₁₆ specimen, AE energy for the [15°/−15°]_4s//[15°/−15°]_4s specimen is much larger than that for the [0°]₁₆//[0°]₁₆ specimen. More 80–100 dB signals appear for the [15°/−15°]_4s//[15°/−15°]_4s specimen, indicating more fiber breakage during crack propagation. In addition, fiber breakage for the hygrothermal specimen is also more than that for the regular specimen. At the third stage, there are more AE energy and more 80–100 db signals for fiber breakage, similar to the [0°]₁₆//[0°]₁₆ specimen.

Figure 14 shows the load/energy–displacement distributions, the amplitude–time distribution and the amplitude–location distribution for the [15°/−15°]_4s//[15°/–15°]_4s specimen for the OLB test under hygrothermal environment. The AE signals using the OLB test are much denser than those using the SLB test at the first stage. Besides, there are more AE signals and energy for the [15°/−15°]_4s//[15°/–15°]_4s specimen than those for the [0°]₁₆//[0°]₁₆ specimen. Different failure modes are observed by SEM at the crack tip, as shown in Figure 15. At the second stage, the energy and amplitude with time for the [15°/−15°]_4s//[15°/−15°]_4s specimen are similar, which are much higher than those for the [0°]₁₆//[0°]₁₆ specimen because 80–100 db signals increase, indicating more fiber breakage in the [15°/−15°]_4s//[15°/−15°]_4s specimen. The crack propagates to approximately the 70 mm location according to Figure 14(c). For all specimens, the crack propagation length for the OLB test is smaller than that for the SLB test. At the third stage, there are only matrix cracking, interface debonding with low energy for the OLB test, different from those for the SLB test. According to Figure 14(c), high-decibel amplitude signals (80–100 db) at the location 70–100 mm appear at the second stage. Compared with the SLB test, AE signals for the OLB test are more uniform at the location 65–100 mm. Because of different layup patterns, only the axial force for the [15°/−15°]_4s//[15°/−15°]_4s specimen resists the bending load. Thus, AE energy and amplitude for the [15°/−15°]_4s//[15°/−15°]_4s specimen for the OLB test are larger than those for the [0°]₁₆//[0°]₁₆ specimen for the OLB test. By comparing Figure 16 with Figure 14(c), 50–80 db signals for the [15°/–15°]_4s//[15°/–15°]_4s specimen at the location 10–60 mm for the OLB test under hygrothermal environment are much denser, indicating hygrothermal environment increases the delamination rate and decreases the delamination crack resistance.

Concluding remarks

In this paper, delamination behaviors of carbon fiber composite laminates with different layup patterns and initial defects using the SLB and OLB tests under regular and hygrothermal environments are studied by analyzing the AE responses in terms of the amplitude, energy and AE events. From the theoretical analysis on the AE signals, the following conclusions are obtained:

The evolving delamination crack tip can be judged by identifying the AE amplitude–location distributions. Compared with the regular environment, the hygrothermal environment leads to much severer delamination growth represented by more AE amplitude signals and high-decibel fiber breakage signals.

Delamination shear resistance for angle-ply laminates is larger than that for unidirectional laminates. From the AE amplitude–location distributions, the delamination resistance for the [15°/−15°]_4s//[15°/–15°]_4s specimen is stronger than that for the [0°]₁₆//[0°]₁₆ specimen, but the load-bearing capability for the former is weaker than that for the latter.

From the load–displacement curves, initial delamination appears more early for the SLB test than for the OLB test, but the load-bearing capability for the SLB test is slightly smaller than that for the OLB test. In addition, from the AE amplitude–location distributions, the delamination crack propagation is slower and more stable for the OLB test than that for the SLB test, represented by smoother amplitude signals for the OLB test than those for the SLB test.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Pengfei Liu would sincerely like to thank the support of the National Natural Science Funding of China (no. 51375435), the National Key Fundamental Research and Development Project of China (no. 2015CB057603), Aerospace Support Technology Funding (no. GFJG-112108-E11402) and Aerospace Science and Technology Innovation Funding (no. GFJG-112108-E81504).

References

Dharmawan

Simpson

Herszberg

. Mixed mode fracture toughness of GFRP composites. Compos Struct 2006; 75: 328–338.

Pereira

de Morais

. Mixed mode I+II interlaminar fracture of glass/epoxy multidirectional laminates-Part 2: experiments. Compos Sci Technol 2006; 66: 1896–1902.

Asp

. The effects of moisture and temperature on the interlaminar delamination toughness of a carbon/epoxy composite. Compos Sci Technol 1998; 58: 967–977.

Abdel Wahab

Crocombe

Beevers

. Coupled stress-diffusion analysis for durability study in adhesively bonded joints. Inter J Adhes Adhes 2002; 22: 61–73.

El Yagoubi

Lubineau

Roger

. A fully coupled diffusion-reaction scheme for moisture sorption-desorption in an anhydride-cured epoxy resin. Polymer 2012; 53: 5582–5595.

El Yagoubi

Lubineau

Saghir

. Thermomechanical and hygroelastic properties of an epoxy system under humid and cold-warm cycling conditions. Polym Degrad Stabil 2014; 99: 146–155.

El Yagoubi

Lubineau

Traidia

. Monitoring and simulations of hydrolysis in epoxy matrix composites during hygrothermal aging. Compos Part A 2015; 68: 184–192.

Ashcroft

Abdel Wahab

Crocombe

. The effect of environment on the fatigue of bonded composite joints. Part 1: testing and fractography. Compos Part A 2001; 32: 45–58.

Browning CE, Husman GE and Whitney JM. Moisture effects in epoxy matrix composites. In: Composite materials: testing and design, West Conshohocken, PA, ASTM STP 617, January 1977.

10.

Gillat O and Broutman LJ. Effect of an external stress on moisture diffusion and degradation in a graphite-reinforced epoxy laminates. In: Advanced composite materials-environmental effects. Philadelphia, PA: ASTM STP 658, January 1978.

11.

Zhang

. Study of the influence of immersion on the carbon fiber/epoxy composites. J Reinf Plast Compos 2008; 27: 659–666.

12.

Telford

Katnam

Young

. The effect of moisture ingress on through-thickness residual stresses in unsymmetric composite laminates: a combined experimental-numerical analysis. Compos Struct 2014; 107: 502–511.

13.

Bowditch

. The durability of adhesive joints in the presence of water. Inter J Adhes Adhes 1996; 16: 73–79.

14.

Zhang

. Effects of voids on residual tensile strength after impact of hygrothermal conditioned CFRP laminates. Compos Struct 2013; 95: 322–327.

15.

Derrien

Gilormini

. The effect of moisture-induced swelling on the absorption capacity of transversely isotropic elastic polymer-matrix composites. Int J Solids Struct 2009; 46: 1547–1553.

16.

De Rosa

Santulli

Sarasinia

. Acoustic emission for monitoring the mechanical behaviour of natural fibre composites: a literature review. Compos Part A 2009; 40: 1456–1469.

17.

Fuwa

Harris

Bunsell

. Acoustic emission during cyclic loading of carbon-fibre-reinforced plastics. J Phys D Appl Phys 1975; 8: 1460–1471.

18.

Berthelot

Rhazi

. Acoustic emission in carbon fibre composites. Compos Sci Technol 1990; 37: 411–428.

19.

De Groot

Wijnen

PAM

Janssen

. Real-time frequency determination of acoustic emission for different fracture mechanisms in carbon/epoxy composite. Compos Sci Technol 1995; 55: 405–412.

20.

Liu

Chu

Liu

. A study on the failure mechanisms of carbon fiber/epoxy composite laminates using acoustic emission. Mater Des 2012; 37: 228–235.

21.

Bak

Kalaichelvan

Arumugam

. A novel approach for classification of failure modes in single lap joints using acoustic emission data. J Compos Mater 2014; 48: 3003–3017.

22.

Oskouei

Ahmadi

. Acoustic emission characteristics of mode I delamination in glass/polyester composites. J Compos Mater 2010; 44: 793–807.

23.

Arumugam

Adhithya Plato Sidharth

Santulli

. Failure modes characterization of impacted carbon fiber reinforced plastics laminates under compression loading using acoustic emission. J Compos Mater 2014; 48: 3457–3468.

24.

Mechraoui

Laksimi

Benmedakhene

. Reliability of damage mechanism localisation by acoustic emission on glass epoxy composite material plate. Compos Struct 2012; 94: 1483–1494.

25.

Chou

Mouritz

Bannister

. Acoustic emission analysis of composite pressure vessels under constant and cyclic pressure. Compos Part A 2015; 70: 111–120.

26.

Assarar

Bentahar

El Mahi

. Monitoring of damage mechanisms of sandwich composites using acoustic emission. Inter J Damage Mech 2015; 24: 787–804.

27.

Fotouhi

Pashmforoush

Ahmadi

. Monitoring the initiation and growth of delamination in composite materials using acoustic emission under quasi-static three-point bending test. J Reinf Plast Compos 2011; 30: 1481–1493.

28.

de Morais

de Mourab

Marquesb

. Mode-I interlaminar fracture of carbon/epoxy cross-ply composites. Compos Sci Technol 2002; 62: 679–686.

29.

Silversides

Maslouhi

LaPlante

. Acoustic emission monitoring of interlaminar delamination onset in carbon fiber composites. Struct Health Monit 2013; 12: 126–140.

30.

Scholey

Paul

Michael

. Quantitative experimental measurements of matrix cracking and delamination using acoustic emission. Compos Part A 2010; 41: 612–623.

31.

Szekrényes

. Beam and finite element analysis of quasi-unidirectional composite SLB and ELS specimens. Compos Sci Technol 2004; 64: 2393–2406.

32.

Davidson

Bansal

Bing

. Geometrically nonlinear determination of energy release rate and mode ratio in single leg bending tests. J Reinf Plast Compos 2009; 28: 1881–1901.

33.

de Moura

MFSF

Goncalves

JPM

Fernandez

. Fatigue/fracture characterization of composite bonded joints under mode I, mode II and mixed-mode I+II. Compos Struct 2016; 139: 62–67.

34.

Szekrenyes

. Over-leg bending test for mixed-mode I/II interlaminar fracture in composite laminates. Inter J Damage Mech 2007; 16: 5–33.

35.

Quino

El Yagoubi

Lubineau

. Characterizing the toughness of an epoxy resin after wet aging using compact tension specimens with non-uniform moisture content. Polym Degrad and Stabil 2014; 109: 319–326.

36.

Liu GM. Nondestructive inspection technology. China, Beijing: National Defence Industry Press (In Chinese), 2006.