Fluorocarbon Removal
 Warning: | This page has not been released yet. |
| About this Process | |
|---|---|
| Process Details | |
| Equipment | Fluorocarbon Removal |
| Technology | Cleaning |
| Fluorocarbon Removal | |
|---|---|
| |center|300px][File:110101.jpg|thumb]] | |
| Technology Details | |
| Other Names | S5T2 Furnace used for Burn Off step |
| Technology | Cleaning |
| Equipment | Solvent Bench 14, RCA Bench 81, Tempress S5T2 - Phosphorous Anneal Oxidation, Acid Bench 02, RCA Bench 81, PFC Bench 01 |
Introduction
Effective removal of fluorocarbon residues has proven challenging. This is a concern in some systems where fluorocarbons are considered a contaminant, e.g., the LPCVD furnaces. Fluorocarbon residues commonly appear after a sample has been treated in a plasma tool that uses fluorine chemistry, such as silicon DRIE systems. At the LNF, a methodical investigation was performed to find the best approach to decontaminate samples with fluorocarbon residues.
Procedure
Considering that the samples contain only silicon compounds and have only been treated in semi-clean or CMOS LNF tools, you can follow the protocol described below.
1. Remove organic residues using PRS-2000 in the Solvent Bench 14 per the SOP
2. Perform an RCA cleaning in the RCA Bench 81 per the SOP
3. Bake the samples at 850C for 8 hours in an O2 environment in the furnace S5T2
4. Perform an HF cleaning using 1:10 HFːDI water for at least 1 min in the Acid Bench 12
5. Perform a piranha cleaning for 10 min in the Acid Bench 12
6. Perform an RCA cleaning in the RCA Bench 81 per the SOP
Characterization
It is challenging to measure residual fluorocarbons located inside deep features. For example, EDP techniques lack the needed resolution to distinguish such remaining materials definitively. Also, ICP-MS is limited to the exposed sample surface. Considering all these challenges, the LNF characterization was based on a destructive approach. Silicon samples were used for the testing—the test involved exposing the samples to an XeF2 silicon etching environment, followed by SEM inspection. If the gas attacked the silicon, it was assumed the surface was fluorocarbon-free. Following this protocol, clear results were obtained. Many samples were prepared and tested. Next, we will summarize the results.
- The control samples (i.e., samples with no furnace treatment) consistently show the worst results.
- Extended oxygen furnace treatment didn't improve results; the results were detrimental. The formation of passivation layers, such as thick oxide or carbide layers, could explain this.
- The best results were obtained with eight hours of oxygen treatment at 850 C.
