Large scale three-dimensional simulations for thick SU-8 lithography process based on a full hash fast marching method

Comprehensive Simulations for Ultraviolet Lithography Process of Thick SU-8 Photoresist

“…Because the wavelength of the 365 nm UV light is rather long, optical diffraction from masks will affect the precision of the lithography pattern. Furthermore, the reflection and refraction at the interface between the air gap (or compensation materials) and the SU-8, and the SU-8 and the substrate will also affect the precision of the lithography pattern [ 36 , 40 , 47 ]. Currently, aerial image simulations are based on two main categories of models.…”

“…One category is based on scalar diffractive theory [ 29 , 30 , 31 , 38 , 41 , 47 , 59 ] and the other is based on rigorous electromagnetic field theory [ 60 , 61 , 62 , 63 , 64 , 65 , 66 , 67 ]. The former can be further divided into models based on the Fresnel diffraction [ 27 , 28 , 29 , 30 , 31 , 59 ], Fresnel–Kirchhoff diffraction [ 38 , 40 , 41 , 47 ], diffractive angular spectrum theory [ 33 ], and so on. While the latter can be further divided into models based on the finite difference time domain (FDTD) method [ 60 , 61 , 62 ] and waveguide method [ 63 , 64 , 65 , 66 , 67 ].…”

“…To deal with this issue, Zhu et al presented the “mask shifting approach” to handle the diffraction and refraction effects simultaneously [ 40 ]. To further reduce the light intensity distribution calculation time, Zhou et al presented a method using an adaptable element size in x , y , and z directions for the whole calculation domain [ 47 ]. Using this method, the calculation time for 3D simulation can be reduced from 2 h to about 16 min for a moderately complex mask shape.…”

“…Liu et al presented an enhanced Dill model using a set of modeling parameters, taking into account of the effects caused by light diffraction or the scattering in the AZ4562 thick photoresist (MicroChemicals GmbH, Ulm, Germany) and the change of photoresist refractive index in the exposure process [ 71 ]. Similarly, Zhou et al developed an improved Dill model describing the nonlinear effects in the SU-8 photoresist [ 45 ], to simulate the exposure kinetics during the exposure process by the following [ 38 , 47 ]: where the SU-8 was divided into n layers, and α i was the absorption coefficient for the i th layer. I 0 ( x , y , z ) was the light intensity.…”

“…However, lithography process simulations of thick photoresist such as SU-8 cannot be accurately implemented by the software specified for thin photoresist, because there are some differences between these two lithography processes [ 28 , 29 ], such as nonlinear factors, the special exposure method for thick SU-8 lithography process, and so on. To overcome this problem, some specified researches for the UV lithography simulations of thick photoresists have been reported during the past several years [ 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 , 48 ].…”