It’s been ages since I posted a True Lab Story, mostly because I’ve been too busy to do anything really dumb. I had a good day for True Lab Stories yesterday, though, so here’s a tale of something idiotic I did, or, rather, had my students do.

I have a student working on a project to put anti-reflection coatings on some diode lasers, which will help improve their performance in various ways that don’t really matter here. This requires the deposition of a very even layer of material that’s a couple hundred nanometers thick, which we do using a vacuum evaporator. This consists of a bell jar attached to a couple of vacuum pumps, to get all the air out of the system, and a metal tray that we heat to extremely high temperatures by pumping a hundred amps of current through it. We mount the lasers above the tray, place the coating material into the tray, and fire it up. The coating material melts, some of it boils off under vacuum, and rises up to coat everything above the tray, including the diode.

The tray in this case is made of Molybdenum, which has an extremely high melting point– 2600 C. The coating material is really expensive sand, with a melting point of 1700 C, so everything works out nicely.

So, my student loads the molybdenum tray with sand (really, it’s more like little glass pellets), and fires up the vacuum system. Only the ion gauge won’t turn on, as if the pressure hasn’t gotten low enough for it to read. Now, on the one hand, it’s entirely possible that the system has a leak in it; on the other hand, that gauge has been kind of flaky lately, so it might be that the system pumped out just fine, and the gauge is just acting up.

“Give it another half hour,” I said, “and if the gauge still won’t read, fire it up anyway. It won’t hurt anything.” In retrospect, those have the unmistakable ring of Famous Last Words.

I went off to do some other stuff, and came back an hour later. “How did it go?” I asked.

“Catastrophe.” came the response. When he turned on the heater, some coating took place, all right– all the parts close to the hot tray became coated with a thick layer of velvety looking black soot. The glass pellets in the tray didn’t actually melt, but the tray itself warped and bubbled and discolored.

It turns out, after a bit of Googling, that while molybdenum ha a very high melting point, it makes a compound called molybdenum oxide (tri-oxide, really) that has a much lower melting point– 795 C. How do you make this, you ask? By heating a piece of molybdenum metal in an atmosphere containing oxygen. Say, the inside of a bell jar that hasn’t been pumped out well enough…

This shouldn’t really have come as a surprise to me, of course, since I was at Yale for the DeMille group’s experiments in high-temperature chemistry. They were trying to make a non-magnetic vacuum oven to heat a sample of lead oxide for an EDM experiment, and banged one together out of quartz. The first time they heated it up in vacuum, the whole thing melted into a small lump at the bottom of their chamber– it turns out that the way you make lead crystal is to heat a mixture of, you guessed it, quartz glass and lead oxide.

And yet, I completely disregarded the possibility that funky things would happen at high temperatures with my own attempts to heat glass in vacuum. I’m such an idiot sometimes, it’s a wonder they let me near students.

Anyway, the soot wiped off the inside of the chamber easily enough, and we’ve got more of the molybdenum trays (I’ll be ordering another batch today, though), so no permanent damage was done. And you can bet we’ll be sure to get the vacuum nailed down before we try this again…


  1. #1 Jonathan Vos Post
    June 27, 2007

    Holy Moly!

  2. #2 patrick leclair
    June 27, 2007

    Wouldn’t have helped you in this case, but a nice reference for what sort of tray to use for what sort of stuff …

    Phrases like “alloys and wets” and “bad for vacuum” saved my bacon many times.

  3. #3 John Novak
    June 27, 2007

    …And here I was expecting a melted bell jar. We’ve actually done the equivalent here with a fiberglass thermal insulating box, and a temperature regulating thermocouple that went bad and engaged in positive feedback.

    And by “we” for a change, I do not mean “me.” My involvement started with, “Sniff. Sniff. Hey, that smells like burning….” and ended with me and another guy almost collapsed in laughter looking for the nearest digital camera. The culprit may think he’s lived that down, but now I need to go mock him about it.

  4. #4 Carl Brannen
    June 27, 2007

    While researching photographic glass at Corning, Stookey left a kiln on overnight with the temperature too high by 1000C. He expected his glass to be in a pool at the bottom of the furnace, destroying it, but instead it was still solid. He had invented glass ceramics, now used in all kinds of high temperature applications like spacecraft windows and ovens.

    I can fully understand the inclination to not believe readouts. I was once convinced that a kiln filled with porcelain wasn’t really at 2250 F. So I opened it up to look to see if all the filaments were on. The “black body” radiation was blinding, I couldn’t see anything but white hot light. The asbestos gloves I was wearing caught on fire.

  5. #5 Lab Rat
    June 28, 2007

    Heh, you gotta love high temperature synthesis work. I was trying out a high temperature (>700 C) reaction involving some boron thin films, and thinking that I’d save my advisor some money, I deposited the boron films on silicon wafers instead of the more expensive sapphire substrates. Of course the reaction didn’t work, and when I brought it up to the boss, he just went, “well yeah, you’ve been making borosilicate glass…”

  6. #6 SW
    November 20, 2007

    Very often when trying to evaporate SiO2 it will disassociate into SiO and O, unless you provide an O over-pressure. Sputtering or CVD works a lot better if you can swing it. The optical properties of resistively evaporated SiO2 are often disappointing.

New comments have been temporarily disabled. Please check back soon.