I've heard good reports from others on this process:

From EP0996587
...
Example XVIII

Monoammonium phosphate

When monoammonium phosphate was subjected to microwave radiation, the phosphate heated rapidly. Almost immediately after turning on the microwave oven, the monoammonium phosphate melted and began to boil. Analysis of the gas that was released showed it to be a mixture of both ammonia and water. Monoammonium phosphate is composed of 14.7% NH 3 and 61.69% P 2 O 5 . After microwave heating for 10 minutes, the glassy material that was formed contained 4.1% NH 3 and 77.2% P 2 O 5 . In the heating of monoammonium phosphate temperatures are reached that exceed the decomposition temperature of ammonia in the presence of phosphoric acid. Thus, ammonia is released, forming phosphoric acid. This is polymerized to polyphosphoric acid. The melting point of monoammonium phosphate is 190°C and this temperature was reached very quickly.

(That Patent is attached)

Monoammonium phosphate is a fertilizer, it is available in Farm-size qty's.

This was apparently (So I was told) tried in a home MW and as long as the starting material was cleaned (ie. recrystallized several times from Fertilizer grade starting material) there were no issues (apart from the loud cracking noise and razor sharp shards of glass from several containers that tried to cool too quickly), well, that and the clouds of NH3 that get real old real quick.

Given that, it was apparently also tried in an old flask - destructive distillation of the dried crystals gives ammonia quite nicely thank you (about 200-250C) as well as the polyphosphoric acid glass (Big tip - to clean it out of glassware, just use anything with some Iodine in it - even PVP-I2 - and excess water).

I've always wanted to try and powder this material and use it in the Tarbutton, et al, synthesis of POCl3 from Salt & P2O5, or in the MsOH/Polyphosphoric acid route to Anhydrides... Ah well (fuckin pigs )

PS I also provided the link to where Java uploaded Polyphosphoric acid in Organic Synthesis

bomex glass is your friend.  You can get a metric ton of that chinese glassware for loose buttons and bits of string.  Otherwise, nickel or porcelain might be your best bet.  I'm sure it will attack porcelain, but at a much slower rate.  When I was experimenting with molten NaOH electrolysis, it only made a bit of surface abrasion which was localized to the electrodes.  I imagine glass would have been gone.