Developing next-generation composites talent

Any educational institution that has developed a specialized area of study — engineering, history, art, medicine — likely can trace its genesis back to a person who supplied a great deal of personal dedication and passion to help bring that specialization to life. More often than not, that person is an educator, and someone who sustained that dedication over many years, in the process drawing the students and acolytes who built the critical mass necessary to make the program self-sustaining.

The composites industry, which is itself relatively young, has had only a little time to develop such specialization at colleges and universities. Still, throughout the world, there is now a healthy handful of strong composites engineering programs that are turning students into composites manufacturing professionals.

One such program can be found at the University of Southern California (USC, Los Angeles, CA, US), where a dedicated educator and a generous, deep-pocketed alumnus have steadily built a highly respected composites engineering program. The generous alumnus was the late M.C. Gill, namesake of composites fabricator Gill Corp. (El Monte, CA, US) and endower of USC’s M.C. Gill Composites Center. The educator is Steve Nutt, director of the M.C. Gill Composites Center and, for 24 years, a professor of engineering at USC (Fig. 1). CW was invited to visit Nutt and the M.C. Gill Composites Center to learn more about the program and the work it is doing.

Nutt, sitting in his office in Vivian Hall within USC’s Viterbi School of Engineering, explains that the M.C. Gill Composites Center has become a popular destination for engineering students, attractive in part because it offers a hands-on program focused primarily on work with prepreg materials and liquid molding. Nutt, assisted by research professor Timotei Centea, lecturer Lessa Grunenfelder and research scientist Bo Jin, leads the research projects of 14 Ph.D. students, six Masters students and eight undergraduates who work in several labs spread out across the center.

The research done at the center, says Nutt, is very much student-led, but guided by faculty. Because of this, students are given much free rein to explore a variety of materials and processing technologies, employing a large variety of tools populating the center’s labs.

The center’s facilities include a layup room with cutting table, a glass-walled miniature autoclave (more on that below), a polishing lab (with an electron polishing machine), a stereoscopy lab and a large catch-all lab (Fig. 2) that includes five Instron testing machines, a digital image correlator, a Wabash hot press, an RTM tool, a small autoclave, a Radius Engineering RTM injector, a small CNC machine and freezers for prepreg storage. During CW’s visit, students were at work in each of these facilities, setting up the glass-walled autoclave, bagging a laminate and evaluating polishing technology.

Nutt, because of his long tenure at the school and the center, is closely associated with the program and its success. He says, however, that whatever success the program enjoys flows not from him, but from its students: “If I can take any credit for the success of this program,” Nutt concedes, “it’s for the ability to attract incredible talent.”

In-situ visualization

It would be difficult to summarize all of the student-led research being done at the M.C. Gill Composites Center, but there are two students whose work is notable and worth closer scrutiny. Both students are part of small teams in which individual researchers focus on different aspects of a common problem.

The first is Mark Anders, who is pursuing a Ph.D. in mechanical engineering. His work, part of a project co-led at USC by Nutt and Centea, and performed in collaboration with another Ph.D. student (Daniel Zebrine) as well as the University of Delaware, has focused on research involving the glass-walled miniature autoclave referenced above. The mini-autoclave is a staple of the center and has been used in several projects. It features an 11-by-11-inch heated aluminum tool plate with a 3-by-3-inch cavity cut from its center. This cavity is filled with a glass viewport through which a bottom-mounted video camera records resin, fiber and core material behavior during the cure process (Fig. 3). Visual observation through the autoclave’s viewport — which enables the “in-situ visualization” — offers insight into material behavior during cure that is, otherwise, very difficult to assess.

Anders says the premise of the team’s research was relatively simple: “What can happen to a honeycomb sandwich structure when a given set of materials, for a given temperature cycle, cures under various pressure conditions?” Anders and his colleagues at USC hypothesized that unmanaged gas pressure in the honeycomb cells can lead to defects in the bond line caused by volatile-release behavior of both the film adhesive and the prepreg resin. Further, by extension he thought that active and knowledge-driven management of vacuum bag and core pressures might reduce bond line void formation.

For his work, he built a partial sandwich structure laid over the autoclave viewport, comprising a 76-by-76-mm square of aramid honeycomb core (from Gill Corp.) topped by a film adhesive (from Henkel), topped by four plies of plain weave carbon fiber/epoxy prepreg laminate (from Hexcel), topped by a vacuum bag (Fig. 4). During cure, the camera, looking through the viewport and up into the honeycomb core cells (Fig. 5), captured resin and film adhesive behavior during the cure cycle. Anders evaluated material behavior under four conditions, labeled A, B, C and D:

• Case A featured a continuous adhesive film, a sealed core and a simple temperature/pressure cycle (1-hour room temperature hold followed by 377 kPa and temperature dwells at 110°C and 180°C).

• Case B featured a reticulated adhesive film, a sealed core and a simple temperature/pressure cycle.

• Case C featured a reticulated adhesive film and a simple cycle, plus, importantly, equilibration of the core and vacuum bag pressures throughout cure.

• Case D featured a reticulated adhesive film, equilibrated core/bag pressures and a three-stage pressure profile.

Anders reticulated the adhesive film by heating and then perforating the film in the middle of each honeycomb cell. Surface tension in the film then caused the adhesive to retract from the middle of the cell and agglomerate along the edge of the honeycomb. Discontinuity of the adhesive, the team discovered in Cases A and B, affected the pressure distribution throughout the core and laminate. In both cases, however, unwanted bubbles grew and remained in the cured structure, thereby increasing porosity and compromising the integrity of the finished part (Fig. 6).

This realization led Anders to Cases C and D, where he began to manipulate pressure inside and outside the vacuum bag and, eventually, to modify the pressure profile of the cure process to minimize bubble formation. In Case C, very low core pressures caused bubbles to grow and then burst, reducing the final porosity but also reducing the size of adhesive fillets (since the bursting process redistributed the adhesive onto the cells walls). In Case D, Anders says, the three-stage pressure profile allowed him not to evacuate all entrapped gases (as might be a reasonable goal), but to moderate and manage bubble formation. “Rather than get the gases out,” he says, “we dissolve them in the resin — keep bubbles from forming in the first place. This keeps porosity low.”

In the end, says Anders, the team’s work showed that bag and core pressure must be managed separately from consolidation pressure: “Basically, you need to set your bag and core pressure to whatever you need to avoid voids, then you set the autoclave pressure to what is required for compaction.” Lab-scale tools such as the glass-walled mini-autoclave provide the means to visually identify the conditions required to suppress defects, and enable knowledge-driven decisions about manufacturing processes.

Anders admits that he was somewhat surprised at how effective management of in-bag pressure was for minimizing voids. He was also surprised to find so little existing research on the subject. The earliest prior research he could find that addressed super-ambient in-bag pressure during cure was on p. 197 of a January 1984 paper, “Processing Science of Epoxy Resin Composites,” written by R.A. Brand et al of the General Dynamics Convair Div. (San Diego, CA, US) and published by US Air Force Wright Aeronautical Laboratories (Wright-Patterson Air Force Base, OH, US). The paper describes development of an “internally pressurized cure cycle” for the fabrication of an F-16 vertical tail skin, and reinforces much of what Anders discovered in his research.

Anders is now working on research to understand gas transport through prepreg relative to resin viscosity affects core pressure. For more, read Anders et al’s paper, published by Elsevier, titled “Process Diagnostics for Co-cure of Sandwich Structures Using In-situ Visualization” and Brand et al’s paper, “Processing Science of Epoxy Resin Composites.”

Dewetting prepreg

Addressing similar material behavior is Sarah Schechter, a fourth-year Ph.D. student conducting research on creating advanced prepregs using dewetting. Previous research performed at USC by Lessa Grunenfelder (now a lecturer at the school), Amy Dills, and Timotei Centea showed that prepregs with discontinuous resin distributions can reduce defect levels during out-of-autoclave cure compared to materials with continuous resin films. Resin discontinuity can be created by dewetting conventional prepregs, as in a process patented by Cytec (now Solvay Composite Materials, Alpharetta, GA, US). However, these previous approaches did not allow for the creation of discontinuous resin patterns in a controlled manner, as the patterns created were dependent on the fiber bed architecture. Schechter says she set out to create an efficient way to create discontinuous resin patterns independent of the fiber bed and, ultimately, allow for its application to any fiber bed.

The premise of USC’s approach is simple: Apply resin to fiber in a regular but discontinuous fashion such that, during cure, entrapped air has a z-direction escape path. The dewetted prepreg Schechter used is one embodiment of a family of materials commonly referred to as USCpreg; it was compared to a control prepreg, which featured continuous resin film (Fig. 7). Schechter dewetted the resin films used to produce the USCpreg with a handheld spike roller at three temperatures (89°C, 104°C and 119°C) over several time spans, ranging from 15 seconds to 8 minutes. The distance between the spikes on the roller was 3.2 mm. Laminates consisted of 16 plies of unidirectional carbon fiber tape, with each tape prepregged on both sides; laminates were cured under vacuum bag on metal tooling, or on a glass window in an oven, analogous to USC’s glass-walled autoclave. Materials were tested with and without edge-breathing dams. Schechter eventually settled on three prepreg types for her assessment: the control prepreg, with continuous film resin; a USCpreg dewetted at 104°C for 30 seconds (104-30); and a USCpreg dewetted at 104°C for 2 minutes (104-120).

Schechter’s hypothesis was simple: Dewetting leads to a discontinuous resin pattern, which creates additional pathways in the through-thickness direction for gases to evacuate, resulting in a finished part with superior properties (Fig. 8). She tested and evaluated pre-cure resin distribution, pre-cure microstructure, resin flow during cure using in-situ visualization, surface defects, bulk porosity and laminate structure. She evaluated the prepregs in optimal and sub-optimal molding conditions.

Details of test results Schechter generated can be found in her paper, “Polymer Film Dewetting for Fabrication of Out-of-Autoclave Prepreg With High Through-Thickness Permeability,” published by Elsevier in the journal Composites: Part A. In summary, however, what became clear is that the overall best-performing dewetted prepreg was 104-120. Consider, for instance, the bulk void content of laminates cured with sealed edges. The control prepreg had a bulk void content of 3.2%. The 104-30 dewetted prepreg had a bulk void content of 0.2-0.3%. The 104-120 dewetted prepreg had a bulk void content of 0.1% (Fig. 9). Schechter notes in her paper, “The insensitivity of dewetted prepregs to restricted in-plane air evacuation demonstrates that air evacuation occurred almost exclusively by breathe-out in the z-direction.” Further, Schechter notes, dewetted prepregs were more forgiving in sub-optimal molding conditions.

Collectively, these findings confirm that discontinuous prepregs can address two long-standing limitations of vacuum-bag only prepreg cure — namely, scaling challenges associated with reliance on edge breathing, and high-defect levels caused by non-ideal manufacturing conditions.

Schechter and others at USC are evaluating processes to more quickly and consistently fabricate discontinuous prepreg in an effort to identify potential commercialization options. Nutt says he has been reaching out to prepreggers and prepreg machinery manufacturers to evaluate dewetting options provided by them as well.

Looking back, Nutt credits M.C. Gill, a pioneer in the composites field, for recognizing the importance of manufacturing technology to American industry. “M.C. foresaw the need to train future generations of engineers in composites processing, and he wanted USC to become a leader in the effort to address that need,” Nutt says. “We see abundant opportunities for innovation and problem-solving in composites manufacturing, and we are gratified to be training some of the engineers who will solve these problems.”