Using Remendable Polymers for Aerospace Composite Structures

ABSTRACT

The work described in this research focuses on the development of a single-component remendable polymer suitable for integration into aerospace composite structures. The structural polymer can be used as a replacement matrix material in fiber-reinforced composites and allows for in-situ site-specific healing of delamination and matrix cracking when combined with a small volume fraction of heat-assisting materials such as magnetic particles. Whereas previous studies focused on the optimal volume fraction and composition of magnetic particles, and the healing of cracks in carbon-fiber composite coupons, current studies focus on the barriers to adopt this material in a aircraft manufacturing environment. These barriers include the (1) the extensive labor involved in producing a limited quantity of Mendomer (1-3 grams), (2) the evolution and entrapment of voids for all but exquisitely controlled environments, and (3) the low melting temperature (ca. 125°C) of the material when compared against high-temperature matrix systems. Proprietary steps have been formulated to increase raw material yield, reduce viscosity, and increase the glass transition temperature. Interests have also been motivated by reducing costs and adhering to conventional composite fabrication techniques. Research has also involved the investigation of automated damage detection to locate the site of damage for further healing, followed by automated healing since the ultimate goal of this research is to develop an autonomously healing composite system. However, a remendable composite is in itself valuable where its successful incorporation will reduce the need for part replacement and maintenance as well as increase the longevity and reliability of the structure.


MOTIVATION

Microcracking, followed by delamination caused by aging, accidental impact, and other damage events can compromise the integrity of a composite structure and lead to required maintenance or in some cases structural failure. The ongoing work presented addresses this issue by suggesting Mendomer as the matrix in fiber composites [Duenas 2006, Duenas 2009B]. Remendable materials re-heal during the application of heat and when used as a composite matrix, can heal cracks in the composite to prevent delamination and further failure. Site-specific heating can be accomplished by embedding heat-assisting materials such as carbon-fiber for healing by resistive heating, or magnetic particles for healing by induction heating [Duenas 2010]. Embedded magnetic particles are suitable for non-conducting systems such as fiberglass composites. Maintenance cost savings are expected where rather than removing and replacing damaged components, remendable composites can be healed repeatedly in place.

COMPARISON TO CONVENTONAL SELF-HEALING MATERIALS

Passive self-healing material systems that use microcapsules are superior to remendable systems in their autonomous response to damage where in the case of remendable systems only active self-healing facilitates an autonomous response; active implies the presence of a system that is actively monitoring the composite for damage to be followed by remending. However, there are several advantages to remendable materials: (1) They have the capacity to heal repeatedly without significant change in material properties [Murphy 2008]. (2) Where as it was previously shown that remendable membranes can self-heal after bullet penetration [Duenas 2010], Mendomer may similarly self-heal during similar heat-generated damage events. (3) Mendomer also exhibits a shape memory effect (SME) where after deformation, followed by heating, the material returns to its original shape [Duenas 2006, Murphy 2008]. This observed SME may assist in self-healing where previously severed bonds return to their original positions during heating and facilitate re-linking of the polymer. The remendability and SME of Mendomer is attributed to its highly cross-linked properties where thermally reversible linkages are the result of multiple Diels-Alder (DA) cycloadditions. Mendomer is a Furan-Maleimide crosslinked solid that can be regarded as a thermoset that softens with temperature [Chen, Duenas 2006, Murphy 2008].

MENDOMER LIMITATIONS

To be sure, there are significant limitations to Mendomer that limit its commercial use and range of applications. Since 2005, there has been an effort to scale up the synthesis process to larger quantities (i.e., produce 1 kg product instead of 1 gram product) by increasing yield [Murphy, Westscott-Baker]. Additional barriers exist for using this material in a fiber-composite matrix in aerospace applications including the frequent evolution of voids during processing and their subsequent entrapment, and the low glass transition temperature of Mendomer (150° C [Duenas 2006]) and associated healing temperature when compared with the operational temperature of conventional fiber-composite matrix systems. Voids appear in neat Mendomer as well as when a low-volume fraction of magnetic particles are added. The persistence of voids appears sensitive to material age, sample geometry, quantity, and processing parameters–such as temperature profile and severity of temperature excursions and whether vacuum is used during preparation. To address the non-autonomous limit of the material, detection methods [Duenas 2009C] and automated healing systems have been investigated for a commercially available remendable material (Surlyn 8940) to demonstrate the concept.

VOID ENTRAPMENT AND REDUCTION

Because of the repeated appearance and entrapment of voids when processing Mendomer at the aircraft manufacturing facility at NextGen with limited oven temperature control and the absence of vacuum and an inert gas supply, the team visited a university (University of California, Santa Barbara, UCSB) to process the materials there. The two samples in Figure 1 were heated for 16 hours at 140°C in Argon in a preheated oven.

 Samples fabricated by NextGen at UCSB fabricated. (a) 5mm-diameter tube leaned on its side, (b) 7-mm diameter upright tube.

Figure 1. Samples fabricated by NextGen at UCSB fabricated. (a) 5mm-diameter tube leaned on its side, (b) 7-mm diameter upright tube.

Both were placed inside a beaker before placed into the oven. The sample in Figure 1a was prepared in a tube leaned on its side while the sample in Figure 1b was upright. Voids appeared in both samples, but fewer were visible in the 7-mm diameter sample. The samples were cured without the presence of a vacuum.

Mendomer was also cured independently at another university (University of California, San Diego, UCSD) as shown in Figure 2. Approximately 200mg of Mendomer was polymerized in a glass vial. The vial was heated to 175 °C in a silicon oil bath over a period of approximately 20 minutes. High vacuum was applied to the vial while the monomer melted to remove trapped bubbles. The final sample is a transparent orange solid with no observable voids.

Polymerization of Mendomer.

Figure 2. Polymerization of Mendomer. 

In an effort to attribute void content to material composition and the presence of impurities when compared, Differential Scanning Calorimetry (DSC) tests were performed independently at different entities (UCLA, Sade, Inc., and UCSB) and compared to previously published results. A subset of these results is provided in Table 1.

Table 1. Mendomer nronerties as gathered bv several DSC nlots.

DSC Plot

Tm (°C)

Enthalpy (J/g)

Tc (°C)

Enthalpy (J/g)

[Murphy 2008]

124.46°C

102.5

129.42°C

41.74

1

130.12

84.49

150.68

40.69

2

126.76

91.41

144.07

57.48

3

120.04

41.69

125.64

48.68

The differences in the melting temperature, enthalpy and polymerization temperature differed from published results, but were credited to systematic error and not attributed to the purity of Mendomer. The differences in DSC peaks are attributed to the difference in heating rates of Mendomer during the independent DSC analyses.

MENDOMER IMPROVEMENTS

The current synthetic method for the Thiele’s acid and the Mendomer 400 monomers only result in a yield of 53 percent and 38 percent respectively. This translates to a 20 percent overall yield for Mendomer 400. Obviously this level of yield is low; but effort recently conducted by one of our team members suggests that another method of synthesis could provide an overall yield close to 90 percent. This method would ultimately replace the current process and make the commercial production of the Mendomer family of resin a realty.

The basic technology behind the remending ability of the Mendomer is not in question. The ability to use the monomer in a manufacturing environment, however, must be addressed. Synthesis of a low viscosity version of the monomer is straight forward to facilitate in processes such as filament winding and resin transfer molding. A highly viscous version for hot melt use such as in facilitating carbon, glass and aramid fibers impregnation is also possible because the healing mechanism of the Mendomer is not based on its ability to melt such as the common thermoplastic material but the fundamental re-arrangement of the backbone by heat.

The current low glass transition temperature of Mendomer is part due to low crosslink density. Increasing the service temperature is commonly achieved by replacing the aliphatic moiety with phenyl radicals. For instance the 1, 4-butanediol could be replaced with the likes of alpha,alpha’-Dichloro-p-xylene which should provide a higher glass transition temperature. There numerous other ways to increase the Tg of the Mendomer that are being considered at this time.

PROGRESS TOWARDS ACTIVE SELF-HEALING

A feasibility study was conducted to investigate the potential of an automatically healable carbon-fiber composite structure. In this study, carbon-fiber composite coupons were fabricated with embedded disks made of a commercially available remendable ionomeric material due to the limited quantities of Mendomer available. The composite coupons were a sandwich structure consisting of 2×2 twill woven carbon fiber plies, a traditional 2-part resin epoxy (2/3 Unibond 1070 and 1/3 601 hardener), and Surlyn 8940 (ca. 0.33mm thick in a disk shape). This sandwich structure involved 3 iterations, totaling 4 carbon-fiber plies and three ionomeric remendable disks. The sandwich structure is shown below in Figure3a, and final composite panel in Figure 3b.

a) Sandwich layup for remendable carbon-fiber coupon and b)the final 4inchx4inch carbon fiber composite with embedded disk-shaped remendable material.

Figure 3. a) Sandwich layup for remendable carbon-fiber coupon and b)the final 4inchx4inch carbon fiber composite with embedded disk-shaped remendable material.

Characterization of the composite was implemented by use of a non-destructive evaluation (NDE), ultrasonic guided wave method using PZT transducers. The testing process included 1) characterizing the Surlyn embedded coupons after fabrication, 2) damaging the coupons, 3) re-assessing the coupons, 4) healing the samples via heat or induction heat, and 5) re-examining the coupons to determine the amount of healing. It was seen that after applying a 5 ft-lb point load and subsequent healing cycle, the ionomer-embedded composite panel regained an average of 10% of the material’s dynamic wave propagating properties. For this study, healing was initiated using direct heat as applied using a powder coating oven. Figure 4 shows the different signals obtained from a pure Surlyn sample using the ultrasonic guided wave method before damage, after damage, and after healing. The source wave propagated (not shown) for the test performed in Figure 4 had frequency of 250 MHz and amplitude of 10V.

Results of the wave propagation tests performed on carbon-fiber panels embedded with pure Surlyn disks: before damage, after damage, and after healing to determine the amount of signal recovery obtained from the healing cycle.

Figure 4. Results of the wave propagation tests performed on carbon-fiber panels embedded with pure Surlyn disks: before damage, after damage, and after healing to determine the amount of signal recovery obtained from the healing cycle.

The design metrics of this approach of nondestructively examining a carbon-fiber panel pre- and post-damage, along with those of the healing cycle are potentially automatable. To most conveniently accomplish this automation, the same system that identifies the presence of damage will also deliver healing to the location of damage. By converging the two functions of damage detection and panel healing in one package, locating and addressing damage are streamlined and maintenance turnover rate reduced.

CONCLUSIONS

As mentioned there are several advantages to Mendomer: (1) It has the capacity to heal repeatedly with minimal change to material properties, though the number of allowable re-healing events before material degradation must be further investigated and quantified. (2) Where as it was previously shown that remendable membranes can self-heal after bullet penetration, Mendomer may similarly self-heal during similar heat-generated damage events. (3) Mendomer exhibits shape memory properties where after deformation, followed by heating, the material returns to its original shape. The authors have observed a vice versa effect in shape-memory polymers (SMPs) where SMPs exhibit self-healing properties, but this must be further investigated. This observed SMP effect may assist in self-healing where previously severed bonds, return to their original positions during heating and facilitate re-linking of the polymer.

Several disadvantages were articulated including the limited availability of the material, evolution and entrapment of voids, the low glass transition temperature of Mendomer, and the need for self-healing autonomy. While proof that void-free neat Mendomer coupons can be fabricated, the limited size and required control during processing remain outside the manufacturing requirements of aircraft composite skin fabrication. Preliminary, but proprietary steps have been formulated to increase raw material yield, reduce viscosity to minimize void evolution, and increase the glass transition temperature but were only briefly discussed. The identification of heat-generating damage events such as during certain phases of projectile flight have been investigated, but were also not been discussed. The ultimate vision of using this material is the harnessing of energy from damage events and the direction of this energy towards healing. This could be solved by first solving the problem on the macroscale using an active self-healing system as the one discussed and then ultimately scaling this active self-healing system to the material level. Multiscale modeling is expected to assist in this ultimate goal. However, while currently not an autonomous system, remendable composites provide a valuable alternative to part replacement where rather than replacing entire parts, damaged areas can be left in place and healed in-situ.

Next post:

Previous post: