For the first attempt, I made a form from several beer cans and some tin foil, wrapped it in sourdough to get an open-topped cube, with a lid, and baked in more foil to help it hold its shape.  After baking, the crucible preform was pyrolyzed in wrapped in several layers of aluminum foil, with flashing tape for extra seal.

After about six or seven hours in a charcoal grill, with charcoal piled around, under, and on top of foil packet, outgassing slowed considerably. I added a final layer of charcoal (I had to add charcoal throughout this process), closed the grill, and let it continue to pyrolyze overnight.  Grill and foil packet were still warm when I got home from work the next afternoon, although it was unclear if this was due to still-lit charcoal or sun.

Packet was removed from grill and allowed to cool in the shade until I could remove (most of ) the foil packet. At this point, project was placed on hold until the next day.

Finally, MAPP gas torch was applied through hole left in side of crucible, with the intent of melting the foil and beercan form inside so crucible could be used.  At this point, it was already apparent that the walls of the crucible were painfully thin, in some cases so thin holes had already formed before pyrolysis, but I had hopes the lower section might be salvageable -- no such luck.  All walls were too thin, and eventually cracked under the thermal stress (see photos under project files, and first photo in gallery).  However, before this happened, the beer cans began to soften collapse, and the foil burned off.  The lid remained intact, being considerably thicker than the walls.

LESSONS

- Dough flows too easily to be held in a flexible form like aluminum foil when walls are so thin.  A more rigid form might be used, or, I will bake a full loaf in a loose form (I know this works from previous experiments), and carve it into shape, "slipping" the interior with more dough or flour and water to make a smooth surface.

- The foil adhesive did not seem to materially contribute to the integrity of the packet once the adhesive burned off, so I will forgo it next time.