A New Method for Monitoring Polymerization of `Pre-preg` Composite Material Using FTIR and Machine Learning

Adva Ziv Zehavi Cohen, iaf, tel nof, Israel (advazivair@gmail.com)
Edan shahmoon, Iaf, Tel Aviv, Israel
Faran Levy, Iaf, Tel Nof, Israel
Adina Itzhak, Iaf, Tel Nof, Israel
Shlomi Ben Haim, Iaf, Tel Nof, Israel
Yonatan Ezra, Iaf, Tel Nof, Israel
Yalfal Siyum, Iaf, Tel Nof, Israel
Moshe Rabaev, Iaf, Tel Nof, Israel

Prepregs are composite materials in which fibers are pre-impregnated within a resin matrix that is partially cured (B stage). Prepreg is commonly used because of its high strength characteristic, uniformity and repeatability compared to a typical manual wet lay-up composite. Epoxy resins are frequently used as a matrix in prepreg manufacturing. The matrix is a reactive resin, so the prepreg film is immediately cooled to -18ºC after the impregnation process in order to minimize further curing reaction. This curing reaction shortens the shelf life of the Prepreg material. Therefore, it is necessary to determine the curing stage of the resin during storage and the degree of curing at the end of the curing process.

FTIR allows the kind of nondestructive analysis that is needed. Since black carbon fibers absorb light on the infrared spectrum, there are quite a few difficulties in distinguishing spectra of different curing stages of the epoxy matrix. Because all spectra at different curing stages appeared the same to the eye, machine learning was implemented to predict the stages of curing. Prediction was successful with 97% efficiency. The combination of FTIR analysis and machine learning allows for the tracking of epoxy curing in prepreg composite material, which had been previously impossible.

---

Short Biography of Presenting Author

Dr. Adva Cohen is the head of the analytical laboratory at Depot 22, Materials Division, Israeli Air Force.