In the past, I have synthesized Phenyl-2-Propanone from Phenylacetic acid and Lead(II)acetate.

As the reaction sits, its a bitch. Reaction temp for the method using Calcium Acetate is real high, even compared to the Lead(II)acetate method (dunno about the use of Barium Acetate), the latter of which is fairly toxic due to its water solubility and there's the expense of Lead(II)acetate to contend with at 3-4 Kg per Kg of phenylacetic acid used. The yeilds are low, for the most part. Even when Lead(II)acetate is increased to 4:1 with PAA the yields dont get over 40%, to my knowledge. The 'aftermath' of the reaction is a big 'clunker' of moon rock looking crap that fills half of the reaction flask and sets up almost like cement. That hard to get out, too. Most 'lead cooks' I have talked to said they just bought lots of glass and smashed them up when done.

However, after much experimentation, brain crunching thought and some lengthy research, I managed to address most all these issues and, in fact, employed little innovations here and there until finally this reaction was brought into the 20th Century as a player. Firstly, the reaction is better ran in Stainless Steel (Aluminum works fine also) so that the damn clinker can be easily removed. Its improved by also using a large capacity propane burner with the steel reactor. The reactor can be quite easily fabricated from a suitable sized pressure cooker that uses a metal to metal seal or the synth rubber seal provided can be replaced with a hand cut hi-temp silicone gasket (good for 6-10 runs). It is essentially a distillation process, so a condenser is made using a hand bent inner stainless steel tube with opposing compression fittings forming the end pieces that also butt and seal the outer jacket, etc. Nipples were easy, as well. The distillation head is a goodly sized 90° steel pipe fitting with (you guessed it) another compression fitting sized to accept the slightly bent inner tube (for 15° slope). Since one cannot see in the steel reactor, a compression fitting seals a thermocouple probe that extends just into the headspace (or through the top of the dist head) to monitor vapor temp. Another similar probe extends well into the reaction mass. Since airborn lead fumes can be pretty toxic, the system should be closed. So, a receiver adapter is fabricated between the condenser output and a good sized separatory funnel. The fume take-off could be lead to a suitable fume trap, but this reaction apparantly suffers most of its yeild losses due to the brutal conditions in the reaction environment. Thermal decomposition seems the culprit. Swim reasoned that the desired ketone was, in large part, getting destroyed because it couldn't get out of the reaction fast enough. The molten lava conditions in the pot have the product trapped in a super dense muck that boils like a tar pit. At the end, it can only more or less 'sweat' the remaining product off of the solidifying post reaction mess.

Preliminary experiments showed that the reaction itself initiated at a lower temp than that which was sustained during the roast. So, the first modification was to run it under a partial vacuum (5" Hg suction on a standard gauge). This did wonders for the reaction. Rapid heating to approx 185°C, then turn the gas down (vac from the start) so the initiation point is smooth (otherwise it'll puke into receiver) at about 195°C or so. A special gauge is placed in the vac. line: a tall water column type U-tube manometer made with clear vac tubing and filled with colored water for ease of viewing. A mark is placed next to the water level corresponding to 5" Hg vac on the gauge. This was a big help because when the rxn kicks, two thing immediately happen: CO₂ is evolved in rapidly increasing amounts (built in speed-o-meter), which causes the system pressure to rise when it tries to over-run the vac pump. At the same time, pot temp will rapidly rise (aided by the pressure increase). However, if the operator is prepared, he simply adjusts the vac. orifice valve to compensate (could use automatic control; never did) and hold the vacuum set-point. Heat can be backed off a bit, but the vacuum helps cool the pot contents enough to bring the pot temp down to a bit over 215-225°C, where it runs smoothly, at a vacuum, to actually provide very high yeilds in the 70% range for a practiced operator using a 4 fold excess of Lead(II)acetate.

Onset of reaction is also accompanied by the appearance of a moderate to thick, white smoke, which shows up in the receiver (sep funnel) and is rich in lead compounds. Therefore, best to use one of those solid teflon vac pumps like KNF makes, or similar. Swix used a large chiller/circulator for antifreeze/water condenser coolant, set at about 10°C.

The only problem left to solve was the Lead(II)acetate toxicity and $75.00/Kg price (plus limited avail in bulk, etc.). This was addressed by innovating the reaction. The Lead(II)acetate toxicity was partially solved by not using it (sorta). Instead, the PAA was mixed with an amount of glacial acetic acid calculated to yield 4Kg of Lead(II)acetate when mixed with the proper amount of litharge (PbO) right in the reaction vessel, just around 30 minutes prior to the run. Litharge is hardly water sol and its toxicity is therefore less an issue. It is also cheap as dog shit and available from any mining supply Co. where it sees extensive use. The cheapest grade works as well as the best grade but only costs about $2.00/lb in 50 or 100 lb bags. This reaction, properly operated is no joke. Using a 20 liter stainless steel pressure cooker about a gallon of phenylacetone will come pissing out into a lye/water solution thats placed in the separatory funnel within the course of about 1½ hrs! Considering another 2-3 hrs involved in set-up and supplying the heat necessary to kick the rxn, then getting the last little cup or so of the 'squeeze' before the ketone becomes rather red in color and is mostly high boilers and dibenzyl ketone, its no slouch. The ketone can be used without a second distillation, only washing it with base, etc, especially if the reaction was a particularly smooth one, but Swix doesn't advise such. It's not a tricky distillation, especially if you dry it well before distillation and use a good, extra-fine capillary tube to avoid bumping. Then, a column isn't even really necessary, since the fore-run boils quite low and the cut-off between the phenylacetone and the poly-aromatic high-boiling by-product is pronounced. This system does a much more honest pre-distillation than the method available to the public.