Measuring the thickness of conformal coatings (CCs) and detecting defects in a nondestructive and noncontact fashion are important to ensure the reliability and performance of packaged electronic components and systems. In this study, we explore the application of terahertz (THz) time-of-flight tomography (TOFT), a noncontact and nondestructive method applied to conformal-coating thickness. We measure the time delay of THz pulses reflected (echoes) from fabricated structures covered by a CC to ascertain coating thickness. Orthogonal matching pursuit (OMP) is applied to deconvolve reflected THz signals from various interfaces, viz., air/CC and CC/substrate. By these means, we map the conformal-coating thickness distribution over significant areas. The CC thickness found over the metal area is similar to 64.3 mu m, while it is similar to 56.2 mu m over the metal-free area. Moreover, the Gaussian mixture model (GMM), a machine-learning algorithm, is used to detect a possible defect-rich region within the CC itself.
Terahertz Nondestructive Characterization of Conformal Coatings for Microelectronics Packaging
Calvelli, Serena;Ricci, Marco;Laureti, Stefano;
2024-01-01
Abstract
Measuring the thickness of conformal coatings (CCs) and detecting defects in a nondestructive and noncontact fashion are important to ensure the reliability and performance of packaged electronic components and systems. In this study, we explore the application of terahertz (THz) time-of-flight tomography (TOFT), a noncontact and nondestructive method applied to conformal-coating thickness. We measure the time delay of THz pulses reflected (echoes) from fabricated structures covered by a CC to ascertain coating thickness. Orthogonal matching pursuit (OMP) is applied to deconvolve reflected THz signals from various interfaces, viz., air/CC and CC/substrate. By these means, we map the conformal-coating thickness distribution over significant areas. The CC thickness found over the metal area is similar to 64.3 mu m, while it is similar to 56.2 mu m over the metal-free area. Moreover, the Gaussian mixture model (GMM), a machine-learning algorithm, is used to detect a possible defect-rich region within the CC itself.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.