Multiscale Simulation Of Nanocontact In MEMS

Micro Electro Mechanical Systems (MEMSs), as a type of integrated microdevices or systems, combine electronic and mechanical components. It can execute thefunctions such as actuation, sensation and control individually or simultaneously.Different from the traditional mechanical systems, the surface effects dominate inMEMS, so adhesion and fiction are the key factors to determine the performance andreliability of MEMS. Adhesion is a main failure mode and is also responsible for loweligibility in MEMS. Since the study of adhesive contact has an important significancein the design and application of MEMS, the nanoscale adhesive contact between Niindenter and single crystal Cu substrate is investigated by using quasicontinuum (QC)method in this dissertation.Firstly, a smooth surface nanocontact model is built using QC method, and themicroscopic deformation mechanism is analyzed during contact process.“Pile-up”around the contact edge is observed in the Cu substrate during the contact process. Thesimulated results indicate that the “Lomer-Cottrel locks” observed during nanocontactprocess act as obstacles to the dislocation motion in the Cu substrate beneath the Niindenter, which leads elastic deformation dominantly in the Cu substrate duringnanocontact process. The contact force and the contact radius, as well as the contactstress distribution during the nanocotact process are investigated in detail. In addition,the classical contact theories such as Hertz, Johnson-Kendall-Roberts (JKR) andMaugis-Dugdale (M-D) theory have been simply reviewed, and the applicability ofthese theories for nanocontact problems are verified. The comprehensive analysis showsthat the M-D theory can precisely describe the relation between the contact force andthe contact radius during nanocontact process. The stress distribution curve obtainedfrom the M-D theory basically agrees with that obtained from the QC method. Due tothe adhesion effect, a small irregular tension zone adjacent to the non-local regionunderneath the indenter is observed in the QC simulation.Secondly, using the QC method, the single asperity contact process is simulated tostudy the microscopic deformation mechanism, in which the asperity and the indenterare of the same order of magnitude. The responds of contact force and contact area verse indenter displacement are investigated. The small and large force drops areobserved in the curve of contact force vs. displacement, corresponding to differentdeformation mechanisms. New dislocations formed in substrate at small force dropstages, while the earlier formed dislocations are destroyed and large amount of atommigration are observed in substrate at large force drop stages. The comparative analysisshows that the asperity can reduce the adhesive effect between indent and substrateeffectively. Due to the existence of large plastic deformation and atomic migrant in thesingle asperity during the contact process which is not considered in M-D theory, thecontact radius vs. displacement curve obtained from the M-D theory deviates seriouslyfrom the corresponding QC curve, which indicates that the M-D theory can not be usedin the single asperity contact problem.Lastly, two multi-asperity nanocontact models with different indenter initiallocation are established based on QC method to investigate the impact of indenter initiallocation on nanocontact process. The indenter locates directly above the middle asperityin one model (=0), while in the other model the indenter locates just above the valleybetween the left two asperities (=-L). The simulation results reveal that thedeformation mechanisms in the two multi-asperity contact models are different. In the=0model, the twinning deformation dominate the whole contact process. In the=-Lmodel, many Lomer-Cottrel locks which are interval distributed, are generated duringthe contact process, inhibiting the occurrence of twinning deformation. In addition,unsymmetrical deformations are observed in the two multi-asperity nanocontact modelsduring the contact process. The relation between nanohardness and indenterdisplacement for the smooth surface model and the two multi-asperity nanocontactmodels are examined together based on the QC method. Finally, the Oliver-Pharrmethod is introduced to estimate the nanohardness at the maximum displacement for thethree contact models, and the estimated results are compared with the QC results. Theanalysis shows that the nanohardness obtained from the Oliver-Pharr method is about30%higher than that obtained from the QC method for the smooth surface contactmodel, but the Oliver-Pharr method can well estimate the nanohardness with about5%error for the two multi-asperity nanocontact models. It also shows that, regardless of themethod used, the nanohardness of=-L model is about20%higher than the=0modelat the maximum displacement.