Dear LAURENT, I worked with chemical, mechanical and chemical mechanical polishing of silicon. I assume that you have chemical mechanical process as it is the best concerning the planarity of the polished surfaces and the removal the damaged layer from the surface of the material. Normally the polishing material consists of an etching solution in which a very fine mechanical powder is suspended. Without the mechanical motion and pressure on the wafers the etch rate is very small. So, the mechanical motion with pressure result in friction which increases the surface temperature of the wafer. This activates the etch process and speeds up the reaction.

I observed as the speed of the polishing machine increases the etch rate increases.

It is possible that there is a chemical reaction running in your machine resulting in a volatile phosphor compound having smell. I assume that the polishing process goes faster in the new machine leading to appreciable vapor with smell.

If you smell unpleasant vapor or gas you must basically go safe and use fumehood.

In Principe for your safety it is a routine to fumehood the polishing and etching machines.

This is just a situation estimation. I hope is will be useful.

i wish you success.

Dear Zekry,

Thanks for your answer that made me realize I forgot to precise that the polishing process is purely mecanical. Indeed we use crystal powder discs of different grain sizes and the process is performed under a water flow.

The fumehood solution is possible but needs major changes in the room where we use the machine. That is why I was asking the question in the first place. But of course if this is a safety issue, we will do it.

Dear Laurant,

Is the new polishing machine operates at higher speeds than the previous one?

Not really...

We may use it for longer times than the previous one (we can for example polish a quarter wafer for several hours)

Excuse me my colleague

IS the pressure on the wafer greater?

Is the type of the Pad the same?

Are all parameters the same?

Thank you

The pressure on the wafer may be a slightly greater even though I have not a clear idea about that (with the previous machine the pressure where applied directly with the hand whlist with the new machine there is an adjustable string to do it)

The pad is almost the same but in the new machine it is maintained by a mechanical arm so the pressure might be more homogenous (in time and space).

This information on the effect of water on indiumphosphide may in addition to what i said help explain the emission of phosphine or any other bad smelling vapor during the polishing process. In addition to long time polishing may lead to pile up of emitted gas. HERE is the link;www.setonresourcecenter.com/.../htdocs/.../Indium%20Phosphide.pd..

wish you success.

Laurent,

It is usual to smell a 'garlic-like' odor when polishing indium phosphide. The smell is not phosphine gas, but you should always polish InP within a fume hood with proper extraction.

Yes it was also my opinion in the first comment. that One has to follow the safety precautions against any hazardous. However it remains always to explain the observations satisfactorily

Thank you.

OK, thank you two for the answers.

We launched the procedure to get a fumehood for the polishing machine. It will be better for everyone.

Laurent,

I work at a major semiconductor company as an EHS professional and have dealt with this isue. The answer is Yes. InP layers will off gas PH3 during CMP process. There are 2 theories; (1) Low pH slurries acts as a reducing agent (adding H to P). (2) metal particles (Al, Ti) actinig as a catalyst. At this time we beleive the Ph3 generation is pH dependent.

Hydrides have been found to be generated at levels that exceed occupational exposure limits during polish. Th method of detection was a MDA TLD-1 Toxic Gas Meter w/ a PH3 key and Hydride tape. The mechanism of this hydride generation is known to be a result of interaction of III-V films with low pH materials, but the mechanical slurry-polish action is thought to play a role in the generation as well. Assessments performed on the polisher have shown that hydrides are generated at significant levels during wafer polish, but that hydride generation ceases when slurry is washed away (during rinse), and that hydride generation is non-detectable at all other times. Assessments during polish have shown the tool enclosure is effective at containing phosphine levels to non-detectable levels outside of the tool (well below the action required level defined as 25% of the occupation exposure limits). It is known however, that CMP tools are not specifically designed to contain hazardous atmospheres, so there is potential for this containment to differ significantly from installation to installation.

Thank you Kevin Martin for your decisive answer of Lauraent question. According to my experience and based on logical and physical facts i estimated as indicated previously in my answers the formation of phosphine gas during the polishing. Now it is experimentally proved.