Effects of variation in autoclave pressure,temperature,and vacuum-application time on porosity and mechanical properties of a carbon fiber/epoxy composite *

Abstract

This article presents results from investigation of the effects of variation in autoclave pressure, temperature, and vacuum-application time on porosity, hot/wet (H/W) and room temperature/dry (RT/D) short beam shear (SBS) strength, and failure mechanism of a commercial carbon fiber/epoxy prepreg, Cycom IM7/977-2 unidirectional tape. Fourteen cure cycles were designed to study a wide range of curing pressures, curing temperatures, and two different vacuum-application durations, including vacuum vented at recommended pressure and vacuum held throughout the cure cycle. The results showed that the SBS strength did not vary significantly over a relatively wide range of curing temperatures and pressures if vacuum was vented at recommended curing pressure; however, after a certain point, a decreasing trend in the SBS strength was observed by reducing the curing temperature and pressure. The C-scan images of panels cured with the vacuum held throughout the cure cycles revealed presence of a high-porosity cross-shaped defect at the center of the panels. The observed defect became larger as the curing pressure decreased. The correlation between the SBS strength and the void content was studied using theoretical models and experimental data. The investigation of the failure modes for each panel showed a change in both the H/W and the RT/D failure mechanism as a result of variation in curing temperature and pressure.

Keywords

Curing temperature curing pressure vacuum short beam shear strength (hot/wet and room temperature/dry)failure mode

Introduction

Final mechanical properties of polymer composite materials highly depend on their curing parameters in the autoclave or oven. Composite material manufacturers specify the proper curing condition with one or more manufacturer recommended cure (MRC) cycles. The MRC cycle should be closely followed to cure composite laminates properly and hence attain the best mechanical properties. The curing parameters specified in the MRC cycle for autoclave curing of composite materials are curing time, curing temperature, curing pressure, and vacuum pressure. It is desirable to ensure the quality of the cured laminates by proper control of the curing parameters; however, the precise control of curing parameters is often costly¹ since deviations from the prescribed curing parameters inevitably happen in production. These deviations may adversely affect the mechanical properties of composite materials. As such, it is important to know how much variation in the curing parameters is acceptable that will not adversely impact the mechanical properties of the cured laminates.

The effect of variation of curing parameters on final mechanical properties of composite materials has been studied extensively.^2–17 In these studies, various defects were often intentionally induced in the laminates by changing one or two curing parameters. Then, the effect of defect on the mechanical properties of interest was investigated and the correlation between the mechanical properties and variation of the curing parameters was reported.

The effect of curing temperature variation on the final properties of polymer composite materials has been studied more than other curing parameters. This is mainly because of the known influence of the state of curing on the mechanical properties of polymer composite materials. For example, if curing temperature is not sufficiently high, resin will not cure fully and, therefore, will not support the fibers completely resulting in reduction of the final mechanical properties of the laminate. Curing temperature also affects the glass transition temperature (T_g) and the yield stress of thermosetting resins.¹⁸ While some studies investigated the effect of isothermal curing temperature on the mechanical properties of composite materials,^2–5 others tried to improve the mechanical properties by designing proper curing cycles such as stepped cure cycles.^6,7

Some studies utilized curing pressure variation to induce voids in composite laminates.^8–14 Voids are one of the most harmful defects in polymer composite materials since they weaken the matrix-dominated mechanical properties such as shear and compressive strength. This is mainly because the voids reduce the load-bearing cross-sectional area of composite laminates. As a result, the stresses on selected fibers increase which, in turn, leads to the localized stress concentrations in the laminate.^19,20 Voids can also link up to make larger defects such as delamination. In addition, voids can reduce the fatigue life of the composite parts.²¹ Several factors could cause void formation in epoxy composite materials¹⁹ including entrapped air in the laminate during the layup,²² dissolved moisture in the resin,²³ and nucleation of void bubbles by chemical reactions releasing the volatiles.²⁴ The mechanism of void formation and growth in polymer composite materials has been studied extensively by Kardos et al.²⁵ They reported that voids tend to grow during cure when the void pressure exceeds that of resin while the resin is still in the liquid state. While the resin pressure and surface tension act against the void growth, increasing the curing temperature helps to grow the voids by increasing the void pressure.²³

The effect of curing pressure on porosity and mechanical properties of polymer composite materials has been studied over the years.^9,10,12,14 These studies suggested that decreasing the pressure increased the porosity which, in turn, reduced the mechanical properties, especially matrix dominated properties such as interlaminar shear strength (ILSS). Although the reported drop in mechanical properties due to pressure reduction in each study did not quite match the findings of the other studies, the observed trends were comparable. One study reported up to 50 percent reduction in interlaminar shear strength when the porosity reached 25% for a glass fiber reinforced polymer.¹⁵ Beside the curing pressure, vacuum pressure has proven to affect the porosity of the cured laminates.¹³

Two micromechanics-based equations have been developed by Murthy and Chamis^26,27 to state the correlation between ILSS and void content and fiber volume fraction of composite materials. The Murthy–Chamis equation for cylindrical voids is defined as:

{ILSS}_{r} = 1 - [4 V_{v} / 3.14 (1 - V_{fv})]^{1 / 2}

(1)

and the Murthy–Chamis equation for spherical voids is defined as:

{ILSS}_{r} = 1 - 0.785 [6 V_{v} / 3.14 (1 - V_{fv})]^{2 / 3}

(2)

where V_v is the void content, V_fv is the fiber volume fraction, and ILSS_r is the ratio of interlaminar shear strength of the laminates with voids to the interlaminar shear strength of void-free laminate.

The Murthy–Chamis equation for spherical voids was shown to be in good agreement with experimental data obtained for ILSS of AS4/PMR-15 UD laminates.⁸

Polymer composite materials are subject to different environmental conditions while in service. As such, it is important to study the effect of environmental factors such as moisture on the properties of composite materials. Although moisture adversely influences the mechanical properties of polymer composite materials, it is not possible to avoid moisture in service since humidity exists in the atmosphere everywhere.^16,28 When thermosetting resins are exposed to moisture at elevated temperatures, reversible or irreversible resin plasticization may occur due to water absorption. When combined with temperature variation, plasticization could cause considerable change in resin toughness, which consequently affects the mechanical properties of composite materials. Additionally, both temperature and humidity can create dimensional variation, which can induce stress in the laminates and, therefore, degrade the fiber–matrix interface.²⁸ Presence of voids in the composite laminates mainly affects the matrix-dominated properties such as interlaminar shear and flexural strength. Moreover, voids create passage for moisture ingression which, in turn, increases vulnerability of the composite laminates to the moisture. Voids also facilitate the passage of air and increase the possibility of fiber oxidization or fiber–matrix interface degradation.²⁹

The effect of environmental factors on mechanical properties of composite materials has been reported previously.^16,17,30,31 According to one study,³⁰ short beam shear (SBS) strength of carbon fiber/epoxy composite specimens reduced between 18% and 21%, depending on stacking sequence, after the specimens were exposed to hygrothermal conditioning. Another study reported the critical void content values below which the ILSS of carbon fiber/epoxy laminates was not affected significantly by environmental conditioning.¹⁶ A different study concluded that the environmental conditioning did not significantly affect the ILSS of the porous laminates fabricated from a carbon fiber/epoxy tape. However, the ILSS of the porous laminates fabricated from a carbon fiber/epoxy fabric were found to be more sensitive to the environmental conditioning.¹⁷

Considering the crucial effect of curing process and environmental factors on the final properties of laminated composite materials and noting the lack of comprehensive study of all major curing parameters, this study was performed to investigate the effect of curing parameters on RT/D and H/W mechanical properties of the 977-2 UD laminated composite. The effects of variation in curing parameters including curing temperature, curing pressure, and vacuum-application time on the DOC, T_g, void content, and thickness of the cured composite laminates, were studied. Also, the effects of intentionally induced defects on the SBS strength and SBS failure mechanism of the 977-2 UD laminates were investigated. The experiments were designed to achieve two research goals. First, studying the effect of a wide range of isothermal curing temperatures on DOC, T_g, SBS strength, and consequent failure mechanism of the 977-2 UD laminates. Second, studying the effect of a broad range of curing pressures along with two different vacuum application durations on void content, void distribution, SBS strength, and consequent failure mechanism of the 977-2 UD laminates. It should be noted that this paper, for the first time, reports the combined effect of vacuum application duration and curing pressure variation on properties of the 977-2 UD laminates. Moreover, performing H/W along with RT/D SBS testing provided the opportunity to investigate the effect of environmental conditioning on the mechanical properties of the 977-2 UD specimens with various degrees of cure and porosities.

Experimental

Material, fabrication, and curing

The material used in this study was a commercial carbon fiber/epoxy prepreg, Cycom IM7/977-2 unidirectional tape, manufactured by Cytec. It is a toughened epoxy resin reinforced by intermediate modulus unidirectional carbon fibers, which is designed for autoclave or press molding curing. The MRC cycle for this prepreg is isothermal cure at 177°C for 180 min.³²

The stacking sequence of [0/90₂/0]_4swas chosen to fabricate 14 balanced and symmetric laminates. Panels 1 to 5 were made using 430 × 430 mm² plies and panels 6 to 14 were made using 305 × 305 mm² plies. All of the laminates were laid up by hand. Debulking was performed for each set of four laid up plies for 5 min, and the final debulking of 32-ply laminates was done for 30 min.

The panels were cured in an autoclave using the cure cycles 1 to 14 shown in Table 1. Non-perforated Tedlar release film was used to avoid pulling too much resin out of the panels during cure. The cure cycles were chosen to cover a wide range of curing temperatures, curing pressures, and duration of applying vacuum. Cure cycles 1 to 5 were designed to study the effect of isothermal curing temperature variation. Cure cycles 6 to 8 were designed to investigate the effect of curing pressure variation. Finally, cure cycles 9 to 14 were designed to study the combined effect of curing pressure variation and maintaining vacuum throughout the cure cycle.

Table 1.

Designed cure cycles for this study (cure cycle 2 is the MRC cycle for the IM7/977-2 UD prepreg)

Cure cycle (panel) number	Ramp up (°C/min)	Dwell temperature (°C)	Dwell time (min)	Ramp down (°C/min)	Cure pressure (kPag)	Vacuum (26 inHg)
1	2.8	182	180	2.8	586	Vent at 69 kPa
2	2.8	177	180	2.8	586	Vent at 69 kPa
3	2.8	171	180	2.8	586	Vent at 69 kPa
4	2.8	160	180	2.8	586	Vent at 69 kPa
5	2.8	149	180	2.8	586	Vent at 69 kPa
6	2.8	177	180	2.8	552	Vent at 69 kPa
7	2.8	177	180	2.8	276	Vent at 69 kPa
8	2.8	177	180	2.8	138	Vent at 69 kPa
9	2.8	177	180	2.8	690	Throughout
10	2.8	177	180	2.8	552	Throughout
11	2.8	177	180	2.8	414	Throughout
12	2.8	177	180	2.8	276	Throughout
13	2.8	177	180	2.8	138	Throughout
14	2.8	177	180	2.8	0	Throughout

Thermal analysis

Glass transition temperature (T_g) was obtained using a TA instruments Q2000 modulated differential scanning calorimeter (MDSC) according to ASTM E2602.³³ Samples of about 5–15 mg cut from the cured panels were encapsulated in Tzero hermetic aluminum pans. Then the samples were heated from 45°C up to 290°C at a rate of 1.5°C/min. The modulation amplitude and period were ±1°C and 60 s, respectively.

Degree of cure (DOC) was obtained using a TA instruments Q2000 differential scanning calorimeter (DSC). Samples of about 10 to 15 mg made with three layers of uncured prepreg were encapsulated in Tzero aluminum pans. Subsequently, the samples were heated in the DSC using the same temperature profiles as those used for curing the panels. DOC was calculated using the following equation:³⁴

α (t) = \frac{\int_{0}^{t} q d t}{H_{U}}

(3)

where α(t) is the degree of cure, q is the heat released from the sample from time 0 to time t during cure measured by DSC, and H_U is the ultimate heat of reaction obtained by the experiment.

Ultrasonic C-scan and thickness variation measurement

A through-thickness ultrasonic C-scan was utilized to obtain the porosity of the panels. A seven-axis ROMER arm coordinate measurement machine (CMM) was used to obtain the thickness variation of the panels.

Panel machining layout configuration

The machining layout for the panels with uniform thickness and porosity was designed to evenly distribute the SBS and acid digestion specimens across the panel. However, this layout could not be used for all panels since some panels had non-uniform thickness and porosity. For each of these panels, a customized machining layout was designed to obtain the SBS and void content specimens from different porosity levels. The size of each SBS specimen (width and length) in the layout was determined per ASTM D2344³⁵ based on the obtained thickness contour of the panel. The long axis of the SBS coupon was aligned with the 0° ply orientation.

Void content

The method used for measuring the density of most of the void content specimens was water immersion. However, this method could not be used for the specimens with visible surface voids. As such, the density of these specimens was obtained by direct measurement of the dimensions and the mass. Acid digestion method was used according to ASTM D3171³⁶ to obtain void content and fiber volume fraction for all the specimens. Sulfuric acid was used to remove the resin from void content specimens.

Short beam shear test

Short beam shear tests were performed according to ASTM D2344. This testing method was selected mainly because of its proven sensitivity to the defects in the laminates.³⁷ Moreover, SBS specimens are relatively small which makes it possible to cut a large number of test specimens out of a moderately sized panel. However, SBS test is not commonly used for design purposes due to the complexity of failure modes.³⁵

To perform H/W and RT/D testing, half of the coupons machined from each panel were conditioned and the other half remained unconditioned. For panels with nonuniform porosity, at least two SBS specimens were machined from each porosity level, one specimen was used for RT/D testing and the other specimen was used for H/W testing.

Dry/room temperature testing

RT/D tests were performed on five unconditioned specimens from each panel. The average value and standard deviation of SBS strength were reported for each panel. In addition, for panels with non-uniform porosity, the value of RT/D SBS strength for each specimen was reported and correlated with the specimen’s porosity.

Hot/wet testing

H/W test specimens were conditioned at 71°C with 85% relative humidity for 30 days according to ASTM D5229.³⁸ The testing was performed at 82°C on five specimens from each panel. The average value and standard deviation of SBS strength were reported for each panel. In addition, for panels with non-uniform porosity, the value of H/W SBS strength for each specimen was reported and correlated with the specimen’s porosity.

Correlation between void content, density and SBS strength

An exponential model was fitted to the experimental data correlating SBS strength to the void content and density. Moreover, the normalized experimental data were compared to the curves obtained using Equations (1) and (2) described in the introduction section.

Investigation of failure mechanism

In order to investigate any change in failure mode, the cross-section of failed SBS specimens (H/W and RT/D) was observed using a Zeiss optical microscope.

Results and discussion

Degree of cure and glass transition temperature

Glass transition temperature and degree of cure were obtained only for cure cycles 1 to 5. It was decided not to obtain the T_g and DOC for panels 6 to 14 since they all had the same temperature profile as cure cycle 2. Moreover, the variation in curing pressure and vacuum application duration in cure cycles 6 to 14 is expected to have minimal effect on the T_g and DOC.

The T_g and final DOC values for panels 1 to 5 are shown in Figure 1. Both the T_g and DOC showed a decreasing trend with reducing isothermal curing temperature. This drop is most pronounced for panel 5 which had the lowest curing temperature.

C-scan, thickness variation, and panel layout

C-scan images for panels 6 to 14 can be seen in Figure 2. C-scan images of panels 1 to 5 were similar to those of panels 6 and 7, and are not reported here. No defects were detected in C-scan images of panels 6 and 7 and a large, nebulous defect was detected at the center of panel 8. For panels 9 to 13, which were under vacuum throughout the cure, a cross-shaped defect was observed at the center of the panels. For panel 10, which was cured according to the MRC cycle, except for holding the vacuum throughout, the defect was distributed more evenly. It is interesting to note that the dimensions of the observed cross-shaped defect increased with decreasing curing pressure. Panel 14, which was cured with no pressure, had very little ultrasonic signal transmitted, indicating widespread defects throughout the panel.

No significant distortion or thickness variation was observed in panels 1 to 5. Therefore, only panels 6 to 14 were inspected using the CMM to obtain thickness variation across the panel. Figure 3 shows the average value and standard deviation of thickness for panels 6 to 14. On each panel, the maximum thickness was observed at the center. The thickness decreased gradually toward the edges in a circular pattern, as indicated in Figure 4.

As shown in Figure 3, thickness variation for panels 6 to 8 was less than that for panels 9 to 14, due mainly to the difference in vacuum application duration. This suggests the importance of proper vacuum application duration for fabrication of high quality laminates.

Due to the uniformity observed in panels 1 to 5, the same machining layout was used for these panels and the SBS and void content specimens were evenly distributed across each panel. However, panels 6 to 14 had nonuniform thickness and porosity. For each of these panels, a customized machining layout was designed to obtain the SBS and void content specimens from different porosity levels. The variation of both thickness and porosity was accounted for in designing the customized machining layouts for panels 6 to 14.

Figure 4 shows two images. The left image is the machining layout for panel 9 superimposed on top of its thickness variation contour and the right image is the C-scan of panel 9. As can be seen in the figure, the SBS and void content specimens were obtained from different spots including the cross-shaped defect. The thickness contours obtained with the CMM for panels 6 to 14 were utilized to calculate the dimensions (width and length) of SBS specimens for each panel. The machining layout used for panels 10 to 14 was similar to the one used for panel 9, except for some small changes in the size of SBS specimens to account for thickness variation.

In order to track the specimens, each specimen was labeled with a number or letters. SBS specimens (RT/D) were marked with numbers and void content specimens were marked as center (CE), middle (MI), corner (CO), and edge (ED). The porosity of SBS specimens was determined using the porosity of adjacent void content specimens. As an example, the void content of SBS specimen number one (SBS 1) in Figure 4 was calculated by taking the average of the void content of CE and MI specimens. Following the same pattern, the void content of SBS 2 was calculated by taking the average of the void content of MI and ED specimens. Also, the void content of MI was assigned to SBS 3, the void content of ED was assigned to SBS 4, and the void content of CO was assigned to SBS 5, SBS 6 and SBS 7. The same procedure was followed to assign porosity to H/W specimens. To obtain two sets of SBS specimens with similar range of porosity for H/W and RT/D testing for panels 8 to 14, the H/W specimens were taken from the upper left quarter of each panel and the RT/D specimens were taken from the lower right quarter of each panel.

Void content

The void content for selected specimens for panels 1 to 14 can be seen in Figure 5. The figure shows the void content for CE, MI, and CO specimens only. As expected, no significant void volume content was observed for panels 1 to 5. The results for panels 6 to 14 will be discussed in more detail accordingly.

Figure 6 illustrates void volume content, average density of cured prepreg, and average fiber volume fraction vs. curing pressure for panels 6 to 14. As can be seen in Figure 6(a), the void volume content increased significantly when the vacuum was maintained throughout the cure cycle. This increase was more significant for the specimens taken from the center of the panels. It is also noticeable that the density of the cured prepreg dropped as the curing pressure decreased (Figure 6(b)). The density drop was more pronounced for panels 9 to 14 for which the vacuum was held throughout the cure cycle. The trend observed for fiber volume fraction variation was quite similar to that of density (Figure 6(c)).

The observed trend of void content, density and fiber volume fraction variation with pressure drop is consistent for panels 6 to 14. Based on the C-scan results in Figure 2 and void content results in Figure 6, it can be concluded that, panels 9 and 10 had a narrow cross-shaped defect at the center, therefore, void content of CE specimen for these panels was higher than that of MI and CO specimens. Panels 11 and 12 had a larger cross-shaped defect at the center which created much higher void in the CE than MI and CO specimens. Void content of the CE specimen for panel 13 did not follow the observed trend for panels 9 to 12, but it is worth noting that void content of MI and CO specimens of panel 13 was much higher than those of panels 9 to 12. Finally, panel 14 had one large splotch so, CE, MI, and CO, all showed high void content values.

To further demonstrate the effect of vacuum application duration on porosity of the cured laminate, the values of void volume content obtained for three different specimens from panels 6 to 8 were compared with those obtained for similar specimens from panels 10, 12, and 13 (Figure 7). The figure shows that maintaining the vacuum throughout the cure cycle increased the average void volume content and induced a nonuniform porosity in the panels.

It can be concluded that for composite laminates fabricated with the 977-2 resin system, the vacuum should not be held throughout the cure cycle as it could result in a high porosity defect in the laminate. For the laminates fabricated with the 977-2 resin system, vacuum is vented once a certain curing pressure is reached at the very early stages of cure to avoid void formation due to the excessive resin bleed. If the vacuum is maintained at elevated curing temperatures while the resin is in liquid state, the resin may bleed until the applied curing pressure is only supported by the fibers. At this stage, the volatiles in the resin start boiling and hence pushing more resin away from the fibers which, in turn, results in creation of voids.

Short beam shear strength

Panels 9 to 14 had nonuniform porosity. As such, for these panels, in addition to the average SBS strength of each panel, the SBS strength of each individual coupon was reported. Figure 8 shows the average RT/D and H/W SBS strength for panels 1 to 14.

Room temperature/dry results

As it can be seen in Figure 8, no significant variation in the average SBS strength was observed for panels 1 to 4. However, statistical analysis showed that the average RT/D SBS strength of panel 5 was significantly less than that of the preceding panels. According to the experimental data, the average SBS strength of panel 5 was 10% less than that of panel 4.

The average RT/D SBS strength of panels 6 and 7 did not vary significantly although the curing pressure for panel 7 (276 kPa) was half of that for panel 6 (552 kPa). This suggests that reducing the curing pressure down to 276 kPa will not adversely affect the mechanical properties of the 977-2 UD laminates provided that the vacuum is vented properly. Nevertheless, the average RT/D SBS strength dropped after the curing pressure was reduced to 138 kPa for panel 8. Also, a general decreasing trend in the average SBS strength was observed for panels 10 to 14 for which the vacuum was applied throughout but the pressure was reduced from 552 kPa to 0 kPa. The most significant drop was for panel 14 which was cured with no curing pressure. It was also observed that for panels 6 to 8 and panels 10 to 14, the standard deviation of the average SBS strength increased with decreasing the curing pressure. The variation in SBS strength for panels 9 to 14 was caused by non-uniform porosity in the panels. As Figure 9 illustrates, in general, the specimens with lower porosity had higher SBS strength.

Hot/wet results

As shown in Figure 8, for each panel the average H/W SBS strength was lower than the corresponding average RT/D SBS strength.

No significant change was observed in the average SBS strength for panels 1 to 2. A decreasing trend could be observed for panels 2 to 5. Similar to the RT/D condition, the drop was more significant for panel 5. For panels 6 to 14, the observed trend of H/W SBS strength was similar to that of RT/D SBS strength.

The observed reduction in the H/W SBS strength of epoxy laminate could be related to several reasons. The water absorption can cause matrix plasticization which may reduce the glass transition temperature of the resin. Also, the combined effect of humidity and temperature may cause degradation of fiber–matrix interfacial bond.³⁹

Comparison

Table 2 contains the average values of RT/D and H/W SBS strength along with standard deviation and coefficient of variation for panels 1 to 14. The percentage reduction of the H/W SBS strength as compared to the RT/D SBS strength for each panel is also included in the table which signifies the effect of H/W conditioning.

Table 2.

The average room temperature/dry and hot/wet short beam shear strength along with standard deviation (STDEV) and coefficient of variation (COV) for panels 1 to 14

Cure cycle no.	Room temperature/Dry			Hot/Wet			Reduction (%)
Cure cycle no.	Strength (MPa)	STDEV (MPa)	COV (%)	Strength (MPa)	STDEV (MPa)	COV (%)	Reduction (%)
1	91.01	0.72	0.79	61.03	1.10	1.80	32.93
2	88.54	2.88	3.25	60.32	1.07	1.77	31.88
3	89.85	2.70	3.00	58.37	1.35	2.31	35.03
4	90.78	2.39	2.63	56.67	2.38	4.21	37.57
5	81.38	2.55	3.14	48.41	0.97	2.01	40.51
6	90.72	3.89	4.28	58.83	1.05	1.79	35.15
7	89.71	3.05	3.40	58.92	1.18	2.00	34.32
8	87.16	4.90	5.62	57.06	2.19	3.83	34.54
9	84.84	7.93	9.35	56.09	3.88	6.91	33.89
10	90.32	4.89	5.42	58.28	1.41	2.41	35.47
11	79.42	10.20	12.84	55.40	3.73	6.74	30.24
12	77.29	14.85	19.22	53.03	7.63	14.39	31.39
13	68.12	19.89	29.20	51.46	9.18	17.84	24.46
14	28.07	12.72	45.31	20.35	7.98	39.20	27.49

For panels 1 to 5, the reduction percentage increased by reducing the curing temperature and consequently decreasing the DOC. However, for panels 6 to 14, the reduction percentage decreased by reducing the curing pressure. Overall, it appears that the value of H/W SBS strength is more sensitive to the curing temperature variation rather than the curing pressure variation. One reason for this observation could be the detrimental effect of lowering curing temperature on the T_g which, in turn, adversely influences the yield stress of thermosetting resins especially at the elevated temperatures.¹⁸

Correlation between composite properties and SBS strength

The correlation between the void volume content and RT/D and H/W SBS strength can be seen in Figure 9. As the figure shows, the RT/D and H/W SBS strength decreased exponentially by increasing the void content. As such, an exponential model was fitted to the experimental data shown in the graph.

The correlation between the cured laminate density and RT/D SBS strength is shown in Figure 10. The results indicate that the RT/D and H/W SBS strength increased exponentially as the density of the cured laminate increased. Therefore, an exponential model was fitted to the experimental data.

Comparison between experimental data and theory

Figure 11 shows the experimental RT/D SBS strength data along with two curves obtained with theoretical Equations (1) and (2). The experimental data in the figure were normalized by dividing the individual SBS strength values by the average SBS strength of the void-free specimens. Figure 12 illustrates the normalized experimental H/W SBS strength data along with two curves obtained with theoretical Equations (1) and (2). As the figure shows, the trend of H/W SBS strength is similar to that of RT/D SBS strength.

The results shown in Figures 9 to 11 suggest that the experimental data agree well with the theoretical models if the void volume content is less than 1%. However, as the void volume content increase, the models gradually deviate from the experimental data. One of the main reasons for this deviation might be the presence of delaminations in the laminates, which were considered as voids in the measurements though they have more severe effect on the SBS strength. Overall, Equation (1) developed for cylindrical voids was found to be closer to the experimental data than Equation (2) developed for spherical voids.

Failure mechanism

Room temp/dry condition

SBS failure mode for all panels except panel 5 was compression-interlaminar shear. The failure mode for panel 5 was interlaminar shear. The failure modes for all panels were acceptable according to ASTM D2344. The typical failed SBS specimen for panels 2 and 5 can be seen in Figures 13 and 14, respectively.

The lighter stripes are 0° plies, with fibers running from left to right in the photo. The darker stripes are 90° plies, with fibers perpendicular to the page. For these specimens failure started under the loading nose and propagated toward the edges. Unlike panels 1 to 4, compression of 0° plies could not be observed around the loading nose. Moreover, failure propagated in 90° plies rather than 0° plies all across the sample. The difference between panel 5 and panels 1 to 4 could be related to the lower curing temperature of panel 5, and its resulting drop in DOC. Poorly cured matrix could not support the fibers properly so no fracture transverse to fibers had occurred.

The failure mode for most of the specimens of panels 6 to 14 was compression-interlaminar shear. However, some specimens did not show compression. Figure 15 shows the failed SBS specimens cut from the center of panel 6. Figures 16 and 17 show the failed SBS specimens from the corner, middle, and center of panels 13 and 14. These figures clearly show that the cracks in the specimens propagated across the thickness through the 90° plies. As such, voids played an important role in propagation of cracks. Also, elongation of voids is noticeable from corner to center as the void content increases. It is worth noting that the voids are visible in the 90° plies (dark stripes), which suggests that they form parallel to the fibers. This observation is in agreement with the findings of Huang and Talreja.⁴⁰ They reported that typical voids in unidirectional composites manufactured by autoclave process exhibited an elongated cylindrical, cigar-shaped pattern running along the fiber direction and, therefore, the voids were better approximated as cylinders with elliptical cross-section.

Hot/wet condition

The H/W SBS failure mode for panels 1 to 5 was interlaminar shear accompanied by inelastic deformation and slight compression. The failure modes for all of the specimens were acceptable according to ASTM D2344. The typical failed SBS specimen for panels 2 and 5 can be seen in Figures 18 and 19, respectively. For the H/W specimens, unlike the RT/D samples, no failure was observed at the edges. Failure in these specimens started under the loading nose and propagated into the middle section. The inelastic deformation, separation of layers, and compression became more pronounced for the specimens cured at lower temperatures. The number of delamination locations was also higher for cure cycle 4 and 5, which had the lower curing temperature. It seems that humidity and temperature reduced the resin mechanical properties even more which, in turn, created weaker bonds between layers.

The failure mode for panels 6 to 8 was compression-interlaminar shear. However, a slight inelastic deformation could be seen in most of the failed specimens. Figure 20 shows the failed SBS specimens cut from the center of the panels 6. The failure mode for panels 9 to 14 was compression-interlaminar shear accompanied by inelastic deformation for most of the specimens. Figures 21 and 22 show the failed SBS specimens from different locations of panels 13 and 14, respectively.

Cracks propagated across the thickness through the 90° plies. The inelastic deformation was more pronounced for specimens with higher void content. The highest inelastic deformation was observed for the center specimen of panel 14, Figure 22(c), which had the highest void content. Comparing Figure 22(c) to Figure 22(a), the amount of variation in thickness can also be noticed.

Conclusions

The results from investigation of the effects of curing temperature, curing pressure, and vacuum application duration on the thermal properties, porosity and mechanical properties of a commercial carbon fiber prepreg, Cycom IM7/977-2 unidirectional tape, were presented. The mechanical property of interest was room temperature/dry (RT/D) and hot/wet (H/W) short beam shear (SBS) strength.

Fourteen cure cycles were designed for this study. Cure cycles 1 to 5 were designed to study the effect of isothermal curing temperature variation, cure cycles 6 to 8 were designed to investigate the effect of autoclave pressure variation, and cure cycles 9 to 14 were designed to study the combined effect of autoclave pressure variation as well as maintaining vacuum throughout cure cycle.

It was observed that maintaining the vacuum throughout the cure cycle for panels 9 to 13 resulted in the formation of a cross-shaped high porosity defect in the panel. The dimension of the cross-shaped defect was increased by decreasing the curing pressure. Panel 14, which was cured with no pressure, had the most widespread defect. The cross-shaped defect observed in the C-scan coincided with observed thickness variations in the panels. The cross-shaped defect was more widespread for the panels with higher thickness variation. Panel 14 which was cured with no pressure had the highest thickness variation.

It was also found that the porosity and the SBS strength of the cured laminates for panels 1 to 4 did not vary significantly over a relatively wide range of curing temperatures (from 160 °C to 182 °C for RT/D strength and from 171°C to 182°C for H/W strength). Moreover, the porosity and the average SBS strength of panels 6 and 7 did not vary significantly although the curing pressure for panel 7 (276 kPa) was half of that for panel 6 (552 kPa). This suggests that the laminates that might be otherwise rejected due to curing temperature or curing pressure variations could still attain acceptable mechanical properties even if they were cured at temperatures or pressures lower than the process specifications. However, after a certain point, a decreasing trend in the average H/W and RT/D SBS strength was observed by reducing the curing temperature as well as the curing pressure. Comparison of the SBS strength showed that the H/W SBS strength is more sensitive to temperature variation than to curing pressure variation.

The SBS strength decreased exponentially with increasing void volume content. As such, an exponential model could fit the experimental data closely. The experimental data of the SBS strength vs. void volume content were also compared with Murthy–Chamis theoretical models. It was observed that, for void volume contents less than 1%, the experimental data agreed well with the theoretical models. However, as the void volume content increased, the models gradually deviated from the experimental data. Generally, Equation (1) developed for cylindrical voids was closer to the experimental data than Equation (2) developed for spherical voids. This was consistent with the observed voids in the failed specimens, which appeared to be aligned in the fiber direction.

Investigation of the failed SBS specimens indicated a change in both H/W and RT/D SBS failure modes with lowering curing temperature and curing pressure. However, the failure mode change was more dominant for the specimens cured at different curing temperatures.

Footnotes

*

The views expressed in this article do not necessarily represent the views of the agency or the United States.

Acknowledgment

The authors gratefully acknowledge financial support from the National Aeronautics and Space Administration (Grant No. NNX09AO58A).

References

Ciriscioli

Springer

Lee

. An expert system for autoclave curing of composites. J Compos Mater 1991; 25(12): 1542–1587.

Lee

S-Y

Springer

. Effects of cure on the mechanical properties of composites. J Compos Mater 1988; 22(1): 15–29.

Peters

PWM

. Influence of cure temperature and degree of surface treatment on transverse cracking and fibre/matrix bond strength in CFRP. J Compos Mater 1994; 28(6): 507–525.

Gernaat C, Alavi-Soltani S, Guzman M, Rodriguez A, Minaie B and Welch J. Correlation between viscoelastic and mechanical properties for and out-of-autoclave polymer composite. In: International SAMPE Technical Conference, Wichita, KS, USA, 2009.

Alavi-Soltani

Sabzevari

Koushyar

Minaie

. Thermal, rheological, and mechanical properties of a polymer composite cured at different isothermal cure temperatures. J Compos Mater 2011 (in press). DOI: 10.1177/0021998311415443.

Cao

Cameron

. The effect of curing conditions on the properties of silica modified glass fiber reinforced epoxy composite. J Reinf Plast Compos 2007; 26(1): 41–50.

White

Kim

. Staged curing of composite materials. Composites Part A 1996; 27(3): 219–227.

Bowles

Frimpong

. Void effects on the interlaminar shear strength of unidirectional graphite-fiber-reinforced composites. J Compos Mater 1992; 26(10): 1487–1509.

Tang

J-M

Lee

Springer

. Effects of cure pressure on resin flow, voids, and mechanical properties. J Compos Mater 1987; 21(5): 421–440.

10.

Boey

FYC

Lye

. Void reduction in autoclave processing of thermoset composites: part 1: high pressure effects on void reduction. Composites 1992; 23(4): 261–265.

11.

Liu

Zhang

B-M

Wang

D-F

Z-J

. Effects of cure cycles on void content and mechanical properties of composite laminates. Compos Struct 2006; 73(3): 303–309.

12.

Olivier

Cottu

Ferret

. Effects of cure cycle pressure and voids on some mechanical properties of carbon/epoxy laminates. Composites 1995; 26(7): 509–515.

13.

Boey

FYC

Lye

. Effects of vacuum and pressure in an autoclave curing process for a thermosetting fibre-reinforced composite. J Mater Process Technol 1990; 23(2): 121–131.

14.

Guo

Z-S

Liu

Zhang

B-M

. Critical void content for thermoset composite laminates. J Compos Mater 2009; 43(17): 1775–1790.

15.

Mouritz

. Ultrasonic and interlaminar properties of highly porous composites. J Compos Mater 2000; 34(3): 218–239.

16.

De Almeida

SFM

De Mas Santacreu

. Environmental effects in composite laminates with voids. Polym Polym Compos 1995; 3(3): 193–204.

17.

Costa

Rezende

De Almeida

SFM

. Strength of hygrothermally conditioned polymer composites with voids. J Compos Mater 2005; 39(21): 1943–1961.

18.

Bicerano

. Prediction of polymer properties, 3rd edn. New York: Marcel Dekker, Inc, 2002.

19.

Campbell

Mallow

Browning

. Porosity in carbon fiber composites: an overview of causes. J Adv Mater 1995; 26(4): 18–33.

20.

Wisnom

Reynolds

Gwilliam

. Reduction in interlaminar shear strength by discrete and distributed voids. Compos Sci Technol 1996; 56(1): 93–101.

21.

Muller de Almeida

Nogueira Neto ZdS . Effect of void content on the strength of composite laminates. Compos Struct 1994; 28(2): 139–148.

22.

Zhang

Sun

. Void formation model and measuring method of void formation condition during hot pressing process. Polym Compos 2010; 31(9): 1562–1571.

23.

Grunenfelder

Nutt

. Void formation in composite prepregs - effect of dissolved moisture. Compos Sci Technol 2010; 70(16): 2304–2309.

24.

White

Kim

. Staged curing of composite materials. Composites Part A 1996; 27(3): 219–227.

25.

Kardos

Dudukovic

McKague

Lehman

. Void formation and transport during composite laminate processing: an initial model framework. ASTM Special Technical Publication 1983; 797: 96–109.

26.

Murthy

PLN

Chamis

. Integrated composite analyzer (ICAN)-users and programmers manual. NASA Technical Paper 2515 1986.

27.

Chamis

Handler

Manderscheid

. Composite nanomechanics: a mechanictic properties prediction. NASA/TM—2007-214673 2007.

28.

Lee

. International encyclopedia of composites, Vol. 1–2. New York: VCH.

29.

Harper

Staab

Chen

. A note on the effects of voids upon the hygral and mechanical properties of AS4/3502 graphite/epoxy. J Compos Mater 1987; 21(3): 280–289.

30.

Botelho

Pardini

Rezende

. Hygrothermal effects on the shear properties of carbon fiber/epoxy composites. J Mater Sci 2006; 41(21): 7111–7118.

31.

Imaz

Rodriguez

Rubio

Mondragon

. Hydrothemal environment influence on water diffusion and mechanical behaviour of carbon fibre/epoxy laminates. J Mater Sci Lett 1991; 10(11): 662–665.

32.

http://www.cytec.com/engineered-materials/products/Datasheets/CYCOM%20977-2.pdf (accessed November 15, 2009).

33.

ASTM International. Standard test method for the assignment of the glass transition temperature by modulated temperature differential scanning calorimetry. ASTM Standard: E2602 - 09, 2009.

34.

Dusi

Lee

Ciriscioli

Springer

. Cure kinetics and viscosity of fiberite 976 resin. J Compos Mater 1987; 21(3): 243–261.

35.

ASTM International. Standard test method for short-beam strength of polymer matrix composite materials and their laminates. ASTM Standard: D 2344/D 2344M – 00, 2006.

36.

ASTM International. Standard test methods for constituent content of composite materials. ASTM Standard: D3171 - 09, 2009.

37.

Bank

Gentry

Barkatt

. Accelerated test methods to determine the long-term behavior of FRP composite structures: environmental effects. J Reinf Plast Compos 1995; 14(6): 559–587.

38.

ASTM International. Standard test methods for moisture absorption properties and equilibrium conditioning of polymeric matrix composite materials. ASTM Standard: D5229/D5229M - 92(2010) 2010.

39.

Pomies

Carlsson

. Analysis of modulus and strength of dry and wet thermoset and thermoplastic composites loaded in transverse tension. J Compos Mater 1994; 28(1): 22–35.

40.

Huang

Talreja

. Effects of void geometry on elastic properties of unidirectional fiber reinforced composites. Compos Sci Technol 2005; 65(13): 1964–1981.