This thread is dedicated to those who do desktop research and solely base theyre argument on theory. It is also dedicated to my good friend and
colleague Xtaldoc.
This subject is a favorite of mine. It’s a very nice (and hard) problem to tackle. Such an elegant and OTC mean of obtaining racemic
phenylpropanolamine (PPA) and yet so many obstacles and problems regarding proper work-up, understanding of the overall mechanism or the nature of the
byproduct formation. Not to mention a less than respectable reported molar yield of 15 % in average! I spent a lot of time studying and trying to
tweak the reaction protocol and finally I am very happy to report a 35 % molar yield of racemic PPA*HCl (based on alanine). I apologize to the
moderators for opening a new thread in advance but I figured the previous one was rather old and convoluted. Feel free to merge it with the old one if
its not worthy of a new thread.
This has been discussed over and over at many different sites and I am sure you are well aware of it. This is why I will solely base my post on the
Cycloknight's thread [1] as I believe it is the most complete and up to date on the subject. Furthermore, my improved work-up and yield is directly
inspired and based on CycloKnight’s write-up, the Yokoyama et al. article [2] and some comments that Nicodem made in the Cycloknight's thread [3].
Before I get to the actual protocol I need to make a few remarks which will make (I hope) the various steps in my procedure more transparent and
justifiable.
1-Lots of people are arguing that the only compound obtained in this procedure is 1,2-diphenylethanolamine (DPEA) quoting this study and affirming that no PPA is obtained at all. DPEA is indeed the major product but it is easily removed in the work-up. The reason the
authors are not isolating any PPA is mainly due to 1) the very little scale theyre working on (amount of PPA is almost insignificant at this scale),
2) they are omitting to concentrate the aqueous phase extracts in the work-up (the latter contains PPA*HCl). This will be re-explained later and will
prove by the same occasion that referring to that article, to conclude that PPA is not in fact produced, is incorrect.
2-The Yokoyama article mentions that the stoechiometry of the reaction involves two moles of benzaldehyde; one for the formation of PPA and one that
condense with the latter to form an oxazolidine. The latter can further be hydrolyzed in the work-up, to regenerate PPA and benzaldehyde, by refluxing
the whole mixture, after the condensation/decarboxylation is over. A large excess of benzaldehyde (6 equivalents) is also needed to maximize yields.
3-An important point was made by Nicodem in the CycloKnight’s thread regarding the possibility of an acid-base reaction between alanine and PPA. PPA
being a base (primary amine) and alanine having a carboxylic acid moiety, it is not impossible for the two to combine and form a salt that will surely precipitate in benzaldehyde. This topic was discussed and illustrated by me and Naf1 at The Vespiary
[4]. Knowing that in CycloKnights work-up all solids after the decarboxylation are assumed to be unreacted alalnine, and therefore discarded, it is
normal to wonder if theyre is not actual PPA*alaninate thrown away at the same time.
It is basically on those last two points that I believe my work-up of the reaction is superior, cleaner and brings a better yield of clean racemic
PPA*HCl. Heres the protocol:
In a 1 L RBF, 358 g of benzaldehyde (3.36 moles) and 50 g of finely crushed DL-alanine (0.56 moles) were mixed together with magnetic stirring. It was
attached to a simple distillation setup and heated on an oil bath at 140 celsius for 3 hours [5] (Figure-1). No more CO2 was evolved from the flask at
this point and the color of the mixture was a deep burgundy (almost black). The flask was removed from the oil bath and let cool on its own for 20
minutes. Afterwards, 100 mL of 20 % AcOH in Toluene were added to the flask and the latter was stirred, at RT, for 20 minutes [6]. When this induction
period was over, 400 mL of 15 % HCl (aq) was added to the flask and the whole mixture was refluxed gently on the oil bath for 3 hours. Afterwards, the
pH of the mixture was verified to be strongly acidic. If not, small portions of concentrated HCl (aq) were added until it was and the mixture was
refluxed for another 30 minutes. The flask was then removed from the oil bath, the mixture left to cool for a bit and then rapidly separated with a
separation funnel while still warm [7] (Figure-2). An orange/red aqueous phase and a dark black organic phase were obtained [8] (Figure-3).
The aqueous phase was washed 3 times with 150 mL of DCM [9] (Figure-4). Then the volume of the aqueous phase was reduced, with gentle heating and
stirring, until half of the volume was vaporized [10] (Figure-5). The remaining aqueous phase was washed 3 other times with 75 mL DCM. The aqueous
phase was then treated with small portions of NaHCO3 [11] until fizzing becomes less violent (Figure-6), then small portions of NaOH were added until
pH was strongly alkaline. A light brown-yellow oily layer with a strong and biting amine smell separated from the aqueous phase (Figure-7). The latter
was separated and the aqueous phase was extracted 4 times with 200 mL DCM [12]. All the NP fractions were pooled together, washed 2 times with 100 mL
BRINE and the DCM was stripped off to afford a viscous brown oil that partially turns to a solid once cooled [13].
The oil was dissolved in 5 times its volume of dry acetone. Small portions of an 1:3 HCl (muriatic)-isopropanol solution were added until pH 4 was
reached [14]. Large amounts of crystalline material precipitated during the addition (Figure-8 and 9). The flask was put in the freezer overnight, the
crystalline material was vacuum filtered and washed with small portions of cold and dry acetone. 33.2 g of racemic PPA*HCl were obtained (Figure-10).
Evaporation of the acetone liquor afforded another 3 g after appropriate washing and drying. Melting point, after 1 week in the dessicator, was
131-136 0C [15]. The product gave a positive test to Chen’s reagent and to the nitrous acid test [16]. It has a very distinctive smell reminding of
crude benzoic acid [17].
[1] http://www.sciencemadness.org/talk/viewthread.php?tid=5979
[2] https://www.thevespiary.org/talk/index.php?action=dlattach;t...
[3] http://www.sciencemadness.org/talk/viewthread.php?tid=5979&a...
[4] https://www.thevespiary.org/talk/index.php/topic,714.0.html
[5] No changes were made to CycloKnights protocol except longer time for decarboxylation and higher amount of benzaldehyde used.
[6] Glacial acetic acid being soluble in toluene, this step was performed in order to protonate the oxazolidine dissolved in the excess benzaldehyde
and make it more prone to hydrolysis latter on. I have no evidence if this step is necessary but I was inspired by the Yokohama et al. article were it
is used.
[7] The two phases separate very cleanly while still warm ( 50-55 0C). If the phases are left to cool at RT, the whole mixture turns into a nasty
emulsion that wont separate and heating needs to be reapply to separate the phases.
[8] The aqueous phase contains PPA salts, DPEA salts and other water soluble impurities. The NP layer contains a large amount of tar and lots of
unreacted benzaldehyde which can be vacuum-distilled later on by firstly washing a few times with diluted aqueous sodium carbonate the NP layer.
[9] This steps is really important as it removes most of the DPEA salts, chloroform is also equally good.
[10] This step is crucial as PPA freebase is soluble in water and will therefore allow one to maximize extraction efficiency once the freebase will be
released. An Erlenmeyer or a large flask with a neck is recommended here to minimize aerosol lost. Do not bring the solution to a boil either. The
temperature was 80-85 0C and it took roughly 5 hours for evaporation. Evaporation under vacuum would be of course ideal.
[11] The resulting aqueous solution being strongly acidic, the use of sodium bicarbonate allows for smoother neutralization of the excess acid (less
heat is generated). Also, a greater amount of salt will be present in the solution and will force the PPA freebase to separate more efficiently.
[12] Do not dry the DCM fraction with magnesium sulfate as a gelatinous addition complex is obtained.
[13] This is mostly due to the fact that the oil is a mixture of 4 different isomers with different melting points.
http://www.thevespiary.org/rhodium/Rhodium/chemistry/nor-pse...
[14] Gassing is not recommended here to claim the hydrochloride.
http://www.thevespiary.org/rhodium/Rhodium/chemistry/nor-pse...
[15] Reported melting point: 134-137°C (butanol-ether (50:50 v/v))
http://www.thevespiary.org/rhodium/Rhodium/chemistry/nor-pse...
[16] Chen’s reagent is specific for alpha-hydroxy amines such as (pseudo)ephedrine and PPA’s. Nitrous acid test indicates a primary amine.
[17] PHENYLPROPANOLAMINE HYDROCHLORIDE, ANALYTICAL PROFILES OF DRUG SUBSTANCES VOLUME 12, Isadore Kanfer, John M. Haigh, and Roslind Dowse
Please post any comments, critique and/or interrogations. I would love to discuss it with anyone interested in
optimizing yields furthermore or reproducing my results.- QD
|
. **
