Experimental characterization of surfaces and metal-polymer heterostacks of optical systems for the automotive sector

The main target of the project consisted in applying the thorough experimental methodology typical of surface physics and materials science to the characterization of materials to be exploited in the automotive lighting sector. Several aspects were considered, thus covering a variety of surface, materials, optical, and physical properties of polymer and metal-polymer reflectors and lenses. The development of a concept machine with the aim of automatically scanning defective outer lenses based on a computer vision system has also been validated. The algorithms of the inspection system, developed in order to avoid the human-eye control, have been fine-tuned to increase the performance and the reliability of the prototype machine. Following the experience on the lenses’ inspection, the attention has been focused on reflecting elements. A reflector is one of the many components of rear automotive lamps. Its purpose is both to reflect light generated by a source and to provide a selected aesthetic aspect. It is composed by metallic and protective thin films deposited over plastic substrates: in particular, the structural (thickness of the metal layer, surface roughness, composition…) and optical (specular and total reflectance, spectral dependency of the scattered beam …) properties have been accurately studied. Flat plastic substrates (made of polycarbonate -PC- and acrylonitrile butadiene styrene -ABS-) have been metallized in vacuo (10-4 mbar) through Physical Vapor Deposition (PVD) of aluminum with the subsequent deposition of a protector made of hexamethyldisiloxane (HMDSO) to prevent both chemical and mechanical damaging. The structural properties of metal-polymer heterostacks have been characterized by means of Atomic Force Microscopy (AFM), determining the surface roughness and the thickness of the diverse layers. Different metallization treatments were compared in a thorough characterization of the optical properties by measuring the spectral Bidirectional Reflectance Distribution Function (BRDF). The optical data, corroborated with profilometry measurements, evidenced an important contribution from the surface deformations induced by the mold production process. The most proper techniques to investigate defects forming at the metallized surfaces and interfaces of reflectors have been identified. By means of a combined approach exploiting Scanning Electron Microscopy (SEM) and synchrotron radiation based X-ray computed tomography (SR-CT) interesting results were obtained, yielding insight into the nature of the defects both at the surface and at the metal-polymer buried interfaces. Finally, an innovative and unique prototype machine has been designed and commissioned. The machine has a double aim: on one side, it automatically recognizes the defects arising on the reflectors’ surfaces by visual scanning, while on the other side it checks the reflectance of the metallized layer exploiting optical processes. In fact, by means of specifically designed algorithms, a standard camera is able to determine if the aluminum thin film has the correct reflectance properties, i.e. if the reflector can be considered compliant or not. Furthermore, exploiting Support Vector Machines (SVM) supervised learning models the system, after a deep and precise training, can recognize autonomously the defects arising on the reflectors and classify them. During this three-years-long project, detailed information regarding the nature and origin of materials defects has been obtained. Moreover, the thorough studies involving the structural and optical properties of metal-polymer heterostacks led to a more precise and realistic representation of the reflectors’ surfaces in simulation environments. Finally, the two prototype machines allow for a faster, cheaper, more objective and reliable inspection of the components, providing real time information essential to monitor the production chain and detect potential problems.