SIMULATION OF NANOSCALE ROUGHNESS EVOLUTION OF SILICON SURFACES UNDER CHLORINE PLASMAS

Postgraduate Thesis uoadl:2701773 548 Read counter

Unit:
Κατεύθυνση Τεχνολογίας Ολοκληρωμένων Κυκλωμάτων
Πληροφορική
Deposit date:
2018-03-16
Year:
2018
Author:
Antoniou Iro-Maria
Supervisors info:
ΓΕΩΡΓΙΟΣ ΚΟΚΚΟΡΗΣ, ΣΥΝΕΡΓΑΖΟΜΕΝΟΣ ΕΡΕΥΝΗΤΗΣ, ΙΝΣΤΙΤΟΥΤΟ ΝΑΝΟΕΠΙΣΤΗΜΗΣ & ΝΑΝΟΤΕΧΝΟΛΟΓΙΑΣ, ΕΚΕΦΕ "ΔΗΜΟΚΡΙΤΟΣ"
Original Title:
SIMULATION OF NANOSCALE ROUGHNESS EVOLUTION OF SILICON SURFACES UNDER CHLORINE PLASMAS
Languages:
English
Translated title:
SIMULATION OF NANOSCALE ROUGHNESS EVOLUTION OF SILICON SURFACES UNDER CHLORINE PLASMAS
Summary:
A surface model for Si etching by Cl2 plasma is developed and coupled with a Monte Carlo (MC) modeling framework to predict the etching rate and the nanoscale roughness of Si surfaces. The surface model takes into account all etching mechanisms, i.e. ion-enhanced etching, physical sputtering, and pure chemical etching. However, under the conditions of the measurements and the pertinent calculations, ion-enhanced etching is dominant. Thus, a critical parameter for the accuracy of the calculations is the ion-enhanced etching yield (number of atoms of the substrate removed per incident ion).
The application of an etching yield expression from the literature to the MC framework leads to results which strongly deviate from etching rate measurements. The origin of the deviation is that the published expressions for the etching yield are phenomenological or “macroscopic”. The “macroscopic” etching yield captures the net effect of all surface processes, including redeposition of etching products and angle dependence of etching. This “macroscopic” etching yield is not suitable for a Monte Carlo framework which treats separately each surface process and not their net effect. For a Monte Carlo framework, a “nanoscopic” etching yield suitable to reproduce the “macroscopic” etching yield and rate is required. In this work, the “nanoscopic” etching yield of the dominant etching mechanism, namely ion-enhanced etching, is extracted by fitting the etching rate calculated by the MC framework to available measurements at ion energies of 100 eV. Then, then the results of the MC framework are compared to measurements of etching rate and roughness for different values of the ion energy (50 to 500 eV).
The MC framework reproduces well the experimentally measured dependence of the etching rate on the ion energy. The etching rate is very close to the measurements up to ion energy of 300 eV and is overestimated for ion energies greater than 350 eV. This overestimation may be due a) to the assumption of a fully chlorinated surface which may not be valid at high ion energies and b) to the change of the composition of the ion flux arriving on the surface which is not taken into account. Regarding the root mean square (rms) roughness, although the absolute values are overestimated, the behavior of rms roughness versus ion energy is captured when the redeposition of the etching products is intense. The origin of the overestimation may be an underestimation of the sticking probability of ions on the surface.
Main subject category:
Technology - Computer science
Keywords:
nanoscale surface roughness, Silicon, modeling, simulation, ion- enhanced etching, Monte Carlo
Index:
Yes
Number of index pages:
4
Contains images:
Yes
Number of references:
69
Number of pages:
46
Irw_Thesis_10_12032018_gkok2 (3)_reviewed.pdf (912 KB) Open in new window