Index Home About Blog
From: "Barry L. Ornitz" <>
Subject: Re: heatsink insulator ?   Beryllium Oxide dangers
Date: 14 Dec 1998

Roy Lewallen wrote in message <>...
>Thank you for the information. This agrees with what I had read when
>previously researching the topic. (I looked into it before specifying
>BeO as substrate material for hybrid circuits I've designed.) Several
>people, however, also posted cautions about getting BeO powder into
>cuts, and I believe even contacting the skin, which aren't hazards I
>was aware of and aren't included in your source. Perhaps one of the
>people posting these cautions could also let us know the source of
>their information.


I think I was one of those people.  My information was from National
Beryllia, a division of General Ceramics, when I talked to them about
making some beryllia windows for a microwave instrument I was building.
National Beryllia is the raw material supplier for most companies making
products using beryllium oxide.  NGK is a Japanese ceramics company that
also produces raw beryllia.

What most people do not understand is that MSDS information varies by
manufacturer.  Some companies downplay the safety issues and other
emphasize them.  It all depends on what their profit motive is,
unfortunately.  For beryllium oxide, most MSDS information is for the fired
ceramic form which is pretty inert.  This is not the case for fine
beryllium oxide dust.

I just found about 100 MSDS listing for materials contain BeO.  Most were
by vacuum tube manufacturers or heatsink companies.  These were naturally
the ones least likely to include the more heinous effects of beryllium
oxide, as their products were generally the fired ceramic.

The carcinogenicity of beryllium oxide is an interesting topic in itself.
ACGIH (American Council of Government and Industrial Hygienists)  and NIOSH
(Nat. Inst. Occupational Safety & Health) and a number of other safety
agencies list it as cancerous.  OHSA (Occupational Health and Safety
Administration) does not.  The permissible exposure levels are different
for each agency too.  Some states have their own standards too.  Having
worked with chemicals and their safety for a number of years, all I can
suggest is that you take the conservative route and ALWAYS handle such
conflicting information by taking the lowest concentration as your
standard, unless you know with certainty that the data was flawed - why,
and by how much.

Just to show some random information from the MSDS information I found,
here are a few quotes:

Litton Industries:  "may be fatal is inhaled, swallowed, or absorbed
through the skin."
Thermalloy:  "no hazard associated with handling manufactured forms."
General Ceramics, National Beryllia Div.:  "SKIN CONTACT: PROPERLY CLEAN

Unfortunately my 6 volumes of RTECS (Registry of Toxic Effects of Chemical
Substances) are packed away.  These list the references on which studies
the various exposure levels are based.

My personal feeling is that beryllium oxide insulators are not the sort of
things that hams should be using.  In fact, I feel that many companies
using beryllium oxide in commercial applications handle it far too
frequently without the proper precautions.  Fortunately boron nitride is
rapidly replacing beryllia in many industrial applications.

        Barry L. Ornitz     WA4VZQ

From: "Barry L. Ornitz" <>
Subject: Re: heatsink insulator ?
Date: 11 Dec 1998

Don wrote in message <>...
>Has anyone seen any data or general guidelines for how much voltage a
>heat sink insulator can safely hold off? I have an assortment here, and
>there are several types: Some grey colored ones that are like rubber,
>very flexible. Some yellow ones that seem like a plastic film, also,
>some clear ones that might be the same material. And of course, some of
>the old mica types.

The gray ones are usually silicone or fiberglass impregnated silicone.
Their voltage rating is not very high.  Berquist is the company generally
associated with these.  These have the worst loss tangent  of the group if
the insulator is going to be used at high frequencies.  Their thermal
conductivity is not the best either but they avoid the mess of using
silicone grease.

The yellow (or brownish orange) ones are probably Kapton (DuPont's version
of polyimide polymer).  They are quite strong physically, rated for high
temperatures, and have good electrical properties through the low microwave

The old standby, ruby mica, is naturally rated for the highest temperatures
(well above the temperature that the typical semiconductor fails).  If the
surface is not damaged, it likely has the highest voltage rating.  The RF
dielectric properties are excellent.

With the last two types, you will need silicone grease to insure a good
thermal contact.    As Roy, W7EL, pointed out, you will need to insure the
mating surfaces are very clean and very flat.  If the semiconductor device
will be at high voltage with respect to the normally grounded heatsink,
consider using the Nylon insulating covers for the device.  These generally
snap on and provide lots of protection.

Beryllium oxide insulators are available at an extremely high cost.  They
are an excellent electrical insulator with a thermal conductivity greater
than aluminum.  BUT, if broken, they are EXTREMELY TOXIC.  If the dust is
breathed, ingested, or gets into a cut on the skin you can expect no end of
health problems.  Beryllium oxide is probably as good as mica in its RF

Aluminum oxide is normally not used as a heatsink insulator because it has
such poor thermal conductivity.  However, anodized aluminum IS used.  In
this case the oxide coating is so thin that its thermal properties are not
too important.  These are only useful at low voltages.

Beryllium oxide is normally white.  I have never seen it in pink, although
it may have been produced that way for some special application.  Many
electrical insulators made by Heany Ceramics are pink.  In fact, this is
their trademark color.  I think most of these are Alsimag ceramics though.

Boron nitride is now being used for a thermal conductor / electrical
insulator.  I do not know if it is commercially available for heatsink use.

     73,  Barry L. Ornitz     WA4VZQ

Index Home About Blog