It might be the unit system. Your system of uMKS may be giving the solver problems due to small mass density values in Kilograms/cubic micron. I would like to propose a different unit system: uM (1e-6 M) pKg(1e-12 Kg) us (1e-6 s) uN (1e-6 N). With this new system the solver quickly finds 100 modes and the frequencies match the values you were getting (except they are now in MHz.) Use of the MKS system also works.