RIE Plasma Exhaust Line Condenses HF Acid

Table of Contents

Purpose

At the request of a Techniques RIE Plasma system user, John Goodnight examined the pumping difficulty symptoms on this system and discovered a foul, acrid smell emanating from the pump exhaust. Examination of the use logs showed that there had been recent high usage of the system for SU-8 photo resist etching at high pressures and high power for long times, accumulating to about 30 hours of use over the past month. Further examination of the liquid in the exhaust line showed that it was a very strong acid, with a pH on the order of zero. The process used a CF4 plasma. This paper investigates the chemistry of this process and identifies the effluent and the liquid found in the exhaust line.

Observations

Process

The process used to remove the SU-8 photoresist is a long plasma etch using CF4 at 300 mTorr in a 300 Watt plasma, using a Fomblin charged rotary vacuum pump. The SU-8 sample film is very thick, (several hundred microns), so it requires a very long time to remove this material. An example time would be about 8 hours of continuous plasma.

Observations

This process produces quite a bit of effluent material which condenses as a dry, flakey gray solid in the vacuum chamber as well as a black deposit that condenses on the inside of the chamber lid. In the SS pumping line, there is a greenish solid crust that appears to be wet with something else. In the pump exhaust line, a clear liquid of a watery consistency condenses. We measured its ph and found it to be between zero and one. The pump line also appears to have a sooty black, wet solid encrusted in it.

Other processes conducted in this plasma system are Si etch also using CF4 and O2 as well as photoresist and e-beam resist ashing with a pure O2 plasma at lower pressures and plasma power levels (50 to 150 Watts at ~100 mTorr). These processes are generally much shorter than the long SU-8 runs, taking on the order of 1 – 3 minutes.

The following photos show the effluent condensates in the pumping line and pump exhaust line. Note the black deposit, the clear liquid condensate, and the green corrosion.

Figure 1. Greenish deposit in exhaust line at the juncture with the blue plastic exhaust line where the acidic liquid was found.
Figure 2. Acidic liquid in the blue plastic exhaust line.
Figure 3. Pump exhaust configuration. The pump is connected to the RIE chamber with a SS flex line and the exhaust is run through a SS pipe with a SS flex line joining it to the blue plastic line running to the scrubbed CR exhaust.
Figure 4. Results of the ph test on the strongly acrid smelling clear liquid in the exhaust line at the blue exhaust line junction. This is a very strong acid, with a ph on the order of zero or 1.

Comments

Since this plasma is a CF4 plasma attacking an organic photoresist, we need to know what the chemistry produces in the exhaust line. This relates to the safety procedures used in cleaning up the mess after a long period of SU-8 processing. The pump is charged with Fomblin oil which can handle O2 processes and is usually benign to reactions with effluent gases.

Examination of the composition of SU-8 from the MSDS sheets shows that it is composed of the following:

INGREDIENTS: Gamma Butyrolactone (CAS: 96-48-0); 22-60% Mixed Triarylsulfonium/ Hexafluoroantimonate Salt;

(CAS: 89452-37-9)/(CAS: 71449-78-0); 1-5% Propylene Carbonate (CAS: 108-32-7); 1-5%

Epoxy Resin (CAS: 28906-96-9); 35-75%

DECOMPOSITION PRODUCTS: Carbon Monoxide, Aldehydes, Acids, Hydrogen Fluoride Gas.

Since volatile HF gas may be formed in the decomposition of this material, and we found a strong acidic liquid condensing in the exhaust line, we need to alert all technicians working on this system to be extremely careful and conscious of safety procedures and be sure to wear personal protective clothing. This could be quite hazardous to the health of the technician cleaning the pumping system.