Self-Healing spacecraft

Senior Editor

Engineers are continuing to use more fiber-reinforced composites, especially for spacecraft structures. This includes solar arrays, optical benches, and antenna systems that must withstand the rigors of interstellar travel.

SPACE COMPOSITES
Composites used for spacecraft typically have a polymer matrix made from thermosetting (epoxies, phenolics, and polyimides) resins reinforced with aramid, carbon, or glass fibers. But there is a common weakness with most composite structures: They don't tolerate impacts, i.e., they have low fracture toughness and don't resist crack propagation.

Unlike metals that can absorb a significant amount of impact, thermoset composites absorb energy by elastically deforming or fracturing. Impact loads typically generate microcracks in the "through thickness" direction where damage tolerance in composites is at its lowest, explains Hugo Williams and Ian Bond, researchers at University of Bristol's Department of Aerospace Engineering, in the U.K.

For space composites, temperature extremes, as well as micrometeroid strikes — dust grains traveling at several kilometers/second — start small cracks. These cracks often develop on the opposite side of the surface that gets struck or deep within the material at the fiber/matrix interface, says the Bristol team. This so-called barely visible impact damage, or BVID, makes detection difficult or impossible. And once cracks form, the composite will likely lose strength, stiffness, and stability. BVID also leaves the structure prone to further damage. Over time, microcracks get larger, weakening the spacecraft until a catastrophic failure is inevitable.

To minimize BVID failures, aerospace engineers build in strain safety factors of 5% or more. But adding safety factors comes with a price: Thicker composite walls, for example, add weight and can degrade performance.

Researchers may be on the brink of developing composite laminates that will let spacecraft components automatically mend nicks, scrapes, and cracks. The trick, says the Bristol team, is to devise composites with mechanical microvascular systems. These engineered veins "bleed" adhesivelike polymers that cure to a hard structure inside microcracks before the cracks have a chance to grow into a more serious breach.

The Bristol team is working with laminate composites with micro/nanoencapsulated healing agents throughout the matrix. As a microcrack propagates through the matrix, it eventually encounters and breaks a microcapsule filled with polymer resin. Capillary forces push the released polymer along the open crack front. As the front moves through the structure, it encounters and breaks another microcapsule containing a curing agent or catalyst.

As the polymer and catalyst mix, the self-healing reaction or polymerization solidifies the mixture inside the crack, stopping its progression. Polymerizing the resin forms a structure that, according to the Bristol team, nearly restores the composite to its original strength. And as long as there are still unbroken microcapsules of both the polymer and its catalyst, the composite will still be able to repair future damage.

To build these biomimetic structures, the researchers took a somewhat counterintuitive approach. They construct laminates with hundreds of fragile, hollow glass-fiber (filament) reinforcements. The hollow filaments crack easily when the composite is damaged, releasing either polymer or catalyst.

The filaments were made at Bristol University. They sport a 60-m m OD and a 30-m m ID. This gives the filament a 50% hollow fraction. This so-called 50/60-filament configuration stems from previous work and makes it easy to fill filaments with repair agent. Plies made from hollow filaments used in proof-of-concept trials have estimated gross-volume fractions on par with plies used in conventional spacecraft composites.

Plies with the hollow filaments are placed between conventional solid-glass-fiber-reinforced plies. Researchers fill half of all the hollow filaments with polymer resin and the other half with its catalyst. For future work, the Bristol team would like to tailor filament diameters and hollow fractions to match specific threats the composite will likely see. For example, larger-diameter filaments may be best suited to repair impact damage, while smaller diameter filaments repair internal cracking.

There are two key benefits to self-healing superstructures on spacecraft. They could double spacecraft life, cutting mission costs in half. And doubling spacecraft life may let mission planners aim for destinations farther out in the Solar System.

The key to finding the right self-repair systems for space composites, the Bristol team contends, is to first define the environmental threats the spacecraft will likely encounter.

To characterize environmental threats, the team conducted lab and space tests. These included low-Earth orbit (LEO), circular low-Earth orbit (CLEO), and highly elliptical-orbit (HEO) experiments. The goal was to identify the synergistic effects that high vacuum, atomic oxygen, UV radiation, impacts with cosmic debris, and thermal cycling have on the material.

BIOMIMETIC COMPOSITES
Because environmental threats from space cause internal as well as external damage, the Bristol researchers investigated placing self-healing, hollow filament plies at different depths within the composite. Three locations were considered:

Surface mounted. This approach is applicable to repairing surface damage due to atomic oxygen and UV exposure as well as, micrometeoroid and orbital debris (MOD) impacts.

Intermediate. This position is one or two plies from the outer surface. Putting filaments with self-healing agents here may help repair surfaces as well as internal damage.

Midline. Filaments placed here repair internal damage due to thermal cycling and MOD impacts.

Two types of repair systems, single and two part, were also considered. Both systems are options for the Bristol team's mechanical vascular technique. A single-component resin when released from the filament will contact a hardener that is part of the matrix polymer. In contrast, two-part systems similar to epoxies with "A" and "B" components harden on reaction between the resin and its corresponding catalyst.

Selecting and delivering the repair agent is influenced by where the healing-ply is placed in the laminate structure.

To simplify testing and keep laminate properties the same between different samples during the proof-of-concept stage, the team chose a [0/90] laminate stack-up. Here, the laminate is built with alternating plies containing in the first layer fibers that ran longitudinally (0°) and a second layer where fibers ran in the perpendicular (90°) direction. The more common stacking sequencefor space composites, however, is what's called a quasiisotropic lay-up. Here alternating plies layer at differing angles in relation to the first. The stack-up consists of the fibers in the first layer running longitudinally, the second layer a 45° angle, the third running perpendicular to the first, and the final layer at 45°. This alternating [0/ 45/ 90/ 45] layup can repeat to build composite thickness.

Although a [0/90] layup is not representative of typical space composites, researchers contend it gives better insights into advantages and disadvantages of selfhealingplies at critical locations in the laminate. That's because varying the position of healing plies in a quasi-isotropic laminate could significantly alter properties due to angle plies moving relative to a neutral (longitudinal) axis.

The healing ply structure is made from the [0/90] ply configuration. The repair resin resides inside fibers running longitudinally and the catalyst is put in the adjacent ply with filaments running perpendicular. Test laminates, contain 12 plies in alternating layers. Four healing plies are sandwiched at various locations between standard, 913-epoxy prepeg, solid E-glass fiber plies.

Although there are a number of commercial resin systems designed to withstand space environments, none were available for initial proof-of-concept tests.

Bristol researchers therefore chose an existing room-temperature curing epoxy, Ampreg 20, from SP Systems, in the U.K. The two-part epoxy offered the best combination of utility, range of viscosities, recommended cure schedules, and cost. It features a 1,100-cps-viscosity base resin and a 406-cps-viscosity catalyst (at ambient).

The 90° hollow plies were infused with the catalyst at room temperature. The Ampreg 20 resin needed a warm-water bath (40°C) to lower viscosity and facilitate infusion of 0° oriented filaments. Optimum viscosity should be less then 500 cps.

In the first round of tests, the team needed to show that a modest (1,000-N) impact load using a round indenter would easily break hollow filaments at all three locations within the laminate: The 1,000-N load would mimic a MOD impact.

The first indenter tests on laminates with midline hollow plies left localized fiber failure on the front and large peanut-shaped delamination zones on the back. Damage was even greater with hollow plies in the intermediate position. The front had deeper indentations and more fiber failures, reported the Bristol team. And the back saw more delamination along with evidence of exposed fiber failures.

The greatest damage to front and rear faces came with healing plies on the surface. The 1,000-N load went through the first two plies of the front surface. On the back there was a small area of delamination similar to that suffered by midline samples.

With filament failures in all three locations, the next step was to pinpoint the right resin systems.

Although long-term stability of two-part resins is greater than monolithic systems, researchers deduced that activation of twopart resins in surface-mounted plies is more complicated and likely requires a highly volatile system that would outgas too easily and make dimensionally destabilized structures. A monolithic system would be better for repairing plies on outer surfaces, providing appropriate UV or vacuuminduced activation (curing) could be devised.

This narrowed the list of candidates to epoxy, anaerobic, and silicone-based systems. These resins, however, require UV radiation to cure. Monolithic repair resins must also remain stable despite the presence of atomic oxygen. If degradation from atomic oxygen is likely, silicone-based systems may be the only option. However, this will require development of a way to get silicone gels into low-viscosity silicone resins to ease filament loading.

For internal (midline and intermediate) ply placement, researchers assume atomic oxygen and UV will not be able to reach the repair resin. Although, in theory, exposure could take place if MOD penetrates the surface. In this case, internal and external healing plies would work to repair the structure.

Potential two-part systems include epoxy, cyanate, and silicone-based resins. The Bristol team believes catalyst driven activation is the best choice. It would be better for internal repair than resins with a hardener in the composite's polymer matrix.

Silicones have low stiffness and bond strength compared to other systems; and would not be able to restore the composite to its original strength and stiffness. But it has not been excluded, yet, by the Bristol team.

If, however, a spacecraft manufacturer mandates full restoration of laminate stiffness, then silicone must be rejected. Instead, the repair system would have to be developed from the same resin family as the composite's matrix. In the meantime, however, researchers are assuming silicones will have the strength to maintain dimensional tolerances.

Although initial proof-of-concept tests prove promising, the Bristol team still has a lot of work ahead. They still need to run mechanical tests to assess short and long-term efficiency of candidate resins. This includes evaluating outgassing and volatile emission data for resin systems.

Manufacture's data covers only cured resins. Other parameters to investigate include the influence vacuum has on resin-fiber affinity, and subsequent fracture toughness of the healed plies.

Long-term research goals for the Bristol team include:

• Fiber orientation (i.e., perhaps orienting self-healing filaments through the thickness — similar to recently developed Z-pinning used to boost composite reinforcement in critical areas.
• Developing methods to replenish the healing resin.
• Identifying methods for repeated healing, i.e., branched networks.
• Application of foaming adhesives to restore aerodynamic profiles on surfaces.
• Developing self-healing reinforcements.

MAKE CONTACT
Bristol University, Dept. of Aerospace Engineering, www.aero.bris.ac.uk

European Space Agency, Advanced Concepts Team, www.esa.int/gsp/ACT/index.html andwww.bionics2space.org

SP Systems, +44 (0) 1983 828000,spsystems.com